{"paper_id":"63d02a44-0be2-4046-bc98-e253819d33b5","body_text":"1 \n \n \n \n \n \n \nDisruption of the mRNA m6A writer complex triggers \nautoimmunity in Arabidopsis \n \n \n \n \n \n \n \n \n \n \n \n \n \nCarey L Metheringham1,^, Anjil K Srivastava1,^, Peter Thorpe1,^, Ankita Maji1, Matthew T \nParker1,#, Geoffrey J Barton1 and Gordon G Simpson1,* \n \n \n1School of Life Sciences, University of Dundee, Dundee, UK. \n \n# Current Address: Max Planck Institute for Plant Breeding Research, Cologne, Germany. \n^ Contributed equally \n* Correspondence: g.g.simpson@dundee.ac.uk \n \n \n \n \n \n \n \nKeywords autoimmunity, m6A, poly(A) tail, TIR domain, nanopore  \n \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted February 23, 2025. ; https://doi.org/10.1101/2025.02.18.638636doi: bioRxiv preprint \n\n 2 \nABSTRACT \n \nDistinguishing self from non -self is crucial to direct immune responses against \npathogens. Unmodiﬁed RNAs stimulate human innate immunity, but RNA modiﬁcations \nsuppress this response. mRNA m6A modiﬁcation is essential for Arabidopsis thaliana \nviability. However, the molecular basis of the impact of mRNA m6A depletion is poorly \nunderstood. Here, we show that disruption of the Arabidopsis mRNA m6A writer \ncomplex triggers autoimmunity. Most gene expression changes in m6A writer complex \nvir-1 mutants grown at 17 °C are explained by defence gene activation and are \nsuppressed at 27 °C, consistent with the established temperature sensitivity of \nArabidopsis immunity.  Accordingly, w e found enhanced pathogen resistance and \nincreased premature cell death  in vir-1 mutants at 17°C but not 2 7°C. Global \ntemperature-sensitive mRNA poly(A) tail length changes accompany these \nphenotypes. Our results demonstrate that autoimmunity is a major phenotype of mRNA \nm6A writer complex mutants,  which has important implications for interpreting this \nmodiﬁcation’s role. Furthermore, we open the broader question of whether unmodiﬁed \nRNA triggers immune signalling in plants.  \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted February 23, 2025. ; https://doi.org/10.1101/2025.02.18.638636doi: bioRxiv preprint \n\n 3 \nINTRODUCTION \n \nDistinguishing self from non-self is crucial to ensuring organisms speciﬁcally target immune \nresponses against pathogen infection. RNA modiﬁcations provide one layer by which this \ndistinction is made in humans (Freund et al, 2019; Karikó et al, 2005). Katalin Karikó, Drew \nWeissman and colleagues revealed that unmodiﬁed RNA stimulates the mammalian innate \nimmunity system by activating the Toll-like receptors (TLRs)  TL3, TL7 and TL8 , but \nincorporating modiﬁed nucleosides into RNA ablated this activity (Karikó et al , 2005) . \nConsistent with this, modiﬁed nucleosides have been critical in mRNA  therapeutics \ndevelopment, and the ﬁrst mRNA -based vaccines for COVID -19 were based on 1 -methyl \npseudouridine-containing mRNA (Karikó, 2021) . A broader set of factors beyond TLRs  \nfunction in RNA sensing in humans using RNA structure (including modiﬁcations), localisation \nand availability to distinguish self from non -self (Bartok & Hartmann, 2020). Precision in this \nprocess is important because chronic activation of nucleic acid sensing pathways in humans \nis associated with autoimmune and autoinﬂammatory conditions (Junt & Barchet, 2015). \nThe most abundant internal modiﬁcation of mRNA is the methylation of adenosine at \nthe N6 position (m6A) (Murakami & Jaffrey, 2022). Null mutations that eliminate the activity of \nthe corresponding N6 methyladenosine methyltransferase METTL3 are embryonically lethal \nin mouse, demonstrating that this modiﬁcation can play essential roles in biology (Geula et al, \n2015). A complex of proteins functions with METTL3 to modify mRNA. Orthologs of the human \nwriter complex components, METTL3, METTL14, VIRILIZER, ZC3H13, WTAP and HAKAI, \nare conserved in Arabidopsis and required for mRNA m6A modiﬁcation (Růžička et al, 2017; \nZhang et al, 2022; Shen et al, 2016). Arabidopsis null mutations in each of these components \n(except the HAKAI and ZC3H13 orthologs) are not viable, and where they exist, hypomorphic \nalleles have pleiotropic developmental defects (Růžička et al, 2017; Shen et al, 2016; Wong \net al, 2023; Zhang et al, 2022). In humans and Arabidopsis, m6A is predominantly written into \nthe terminal exon of mRNA in a preferred context characterised by the DRm6ACH consensus \nsequence (Parker et al, 2020; Murakami & Jaffrey, 2022).  \nReader proteins recognise RNA m 6A modiﬁcatio ns and ultimately inﬂuence mRNA \nprocessing and fate (Zaccara & Jaffrey, 2024, 2020) . The best -characterised m6A reader \nproteins have a YTH domain, which binds m6A through a cage of aromatic amino acids. Plants \nand apicomplexans are unique with respect to m6A readers because the conserved CPSF30 \ncomponent of the cleavage and polyadenylation complex, which binds the AAUAAA poly(A) \nsignal (Chan et al, 2014; Schönemann et al, 2014), has a YTH domain (Stevens et al, 2018). \nConsistent with this, a major impact of m6A loss on pre-mRNA processing in Arabidopsis writer \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted February 23, 2025. ; https://doi.org/10.1101/2025.02.18.638636doi: bioRxiv preprint \n\n 4 \ncomplex mutants is disrupted poly(A) site usage (Parker et al, 2020, 2022; Wong et al, 2023). \nNot all m6A effects are mediated by YTH reader domains. m6A can affect RNA structure (Roost \net al, 2015; Kierzek & Kierzek, 2003)  and, through an m6A switch mechanism, inﬂuence the \nassociation of speciﬁc RNA -binding proteins with transcripts in an m 6A-dependent manner \n(Korn et al, 2021; Wu et al, 2018; Liu et al, 2015). \n A multilayered innate immunity system mediates defence against pathogens in \nﬂowering plants (Locci & Parker, 2024). The ﬁrst layer consists of t rans-membrane receptor \nproteins called pattern-recognition receptors (PRRs) that detect pathogens in the external \nenvironment and signal an immune response known as pattern-triggered immunity (PTI). The \nsecond layer comprises networks of proteins that detect pathogen effectors and their activity \ninside plant cells and is known as effector-triggered immunity (ETI). ETI is mainly mediated by \nnucleotide binding/leucine -rich repeat (NLR) receptors. Cross-talk between PTI and ETI \npotentiates the immune response (Yuan et al, 2021; Ngou et al, 2021; Tian et al, 2021; Pruitt \net al, 2021). The diverse immune receptors converge on shared signalling complexes, such \nas those containing EDS1, to promote an immune response  (Locci & Parker, 2024). In cells \ninfected by pathogen s, this can trigger programmed cell death , called the hypersensitive \nresponse. Immune responses in neighbouring cells are also activated , but the gene \nexpression pattern differs from those in infected cells (Tang et al, 2023; Jacob et al, 2023). A \nconcentration gradient of the  hormone salicylic acid (SA) between the sites of infection and \nneighbouring cells controls  the hypersensitive response  and the massive expressi on of  \ndefence genes such as PATHOGEN RESPONSIVE 1 (PR1) in surrounding cells and systemic \nacquired resistance in distal tissues (Betsuyaku et al, 2018; Fu et al, 2012b; Zeier, 2021).  \n PRRs and NLRs are encoded by some of the largest and most rapidly evolving gene \nfamilies in plants (Barragan & Weigel, 2021). This diversity corresponds to selective pressure \nnot only for pathogen defence but also to dampen immune responses in the absence of \ninfection, reﬂecting trade -offs between the beneﬁts of disease resistance and the costs of \nsustained immune responses on development. Like humans, plants can develop autoimmune \nconditions (Alcázar & Parker, 2011; van Wersch et al, 2016; Freh et al, 2022; Wan et al, 2021). \nIn Arabidopsis, these manifest as compromised development and premature cell death, visible \nas leaf lesions. Some of the clearest examples of autoimmunity emerge in crosses between \ndifferent Arabidopsis accessions (Bomblies et al, 2007; Wan et al, 2021). This phenomenon, \nobserved in the ﬁrst or later generations of plant hybrids, is called hybrid necrosis because of \nthe severe pleiotropic symptoms that compromise development and viability  (Bomblies & \nWeigel, 2007) . A recurring explanation for hybrid nec rosis is  simple non -compatible \ninteractions between speciﬁc NLRs or other defence genes that activate immune response \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted February 23, 2025. ; https://doi.org/10.1101/2025.02.18.638636doi: bioRxiv preprint \n\n 5 \npathways (Wan et al, 2021). Naturally occurring genetic variation  (Todesco et al, 2010) or \ninduced mutations  (van Wersch et al , 2016; Freh et al , 2022)  also trigger Arabidopsis \nautoimmunity. For example, gain of function mutations that either stabilise the expression or \nautoactivate NLRs can cause autoimmunity (Zhang et al, 2003; Bonardi et al, 2012), and so \ncan the disruption of signal ling pathways that are either involved in or sensed by  defence \nresponses (van Wersch et al, 2016). A characteristic of Arabidopsis autoimmunity is that it can \nbe suppressed by elevated temperature or relatively high humidity , and this property has \nfacilitated the study of autoimmune genotypes (Alcázar & Parker, 2011). \n Although the Arabidopsis mRNA m6A writer complex is essential for viability, the gene \nexpression changes that explain this are unknown. Here, we asked what groups of genes are \naffected when the mRNA m6A writer complex  is disrupted. We discovered that immune \nresponse genes comprise the major class of altered gene expression, but consistent with the \ntemperature-sensitive nature of Arabidopsis immunity, this response was suppressed when \nplants were grown at elevated temperatures . Furthermore, Arabidopsis mRNA m6A writer \ncomplex mutants display temperature-sensitive increased resistance to pathogen infection \nand increased levels of premature cell death. Therefore, autoimmunity is a major phenotype \nof Arabidopsis mRNA m6A writer complex mutants. In contrast to cases of hybrid necrosis and \nsome other autoimmune mutants, visible developmental defects of mRNA m6A writer complex \nmutants were not rescued by growth at elevated temperatures, revealing that the impact of \ndefective mRNA m6A modiﬁcation on autoimmunity and development is separable. Therefore, \nas with humans, RNA modiﬁcations in Arabidopsis may contribute to distinguishing self from \nnon-self. Our ﬁndings suggest that u ncovering how disruption of the mRNA m6A writer \ncomplex triggers defence gene expression is fundamental to understanding this RNA \nmodiﬁcation's role in plant biology.  \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted February 23, 2025. ; https://doi.org/10.1101/2025.02.18.638636doi: bioRxiv preprint \n\n 6 \nRESULTS \n \nImmune gene expression is activated in mRNA m6A writer complex mutants \nTo understand the roles of mRNA m6A in Arabidopsis, we asked what groups of genes were \nmost affected by the loss of function of the m 6A writer complex protein , VIRILIZER. We \npreviously characterised gene expression changes in vir-1 mutants using a combination of \nIllumina RNA sequencing (RNA-seq) and Oxford Nanopore Technologies direct RNA \nsequencing (ONT DRS) (Parker et al, 2020). We analysed three genotypes with different VIR \nactivity: a wild-type Col-0 control, a hypomorphic Arabidopsis vir-1 mutant defective in VIR \nfunction, and a complementation line expressing VIR fused to Green Fluorescent Protein \n(GFP) (VIR complemented; VIRc) that partly restores VIR activity in the vir-1 mutant \nbackground (Růžička et al, 2017; Parker et al, 2020). For each genotype grown in sterile \nconditions at 2 2°C, we sequenced RNA puriﬁed from seedlings of  at least  six biological \nreplicates with Illumina RNA-seq and four with ONT DRS (Supplementary ﬁle 1).  \nUsing the Illumina RNA-seq data, we identiﬁed differentially expressed genes between \nvir-1 and WT Col-0 by ﬁtting a quasi-likelihood model in edgeR (Chen et al, 2016) (threshold: \nadj.p < 0.001, log2FC > 2.0). We found 806 genes signiﬁcantly upregulated in vir-1 compared \nto Col-0 and 349 genes signiﬁcantly downregulated (Supplementary Table 1). We examined \nGO (gene ontology) term distribution among the differentially expressed genes using gProﬁler \n(Kolberg et al, 2023). The most signiﬁcantly enriched GO terms were related to response to \nexternal stimuli and defence. For example, 92 of the 597 upregulated genes with GO term \nannotation were annotated with the biological process ‘defense response’ (GO:0006952), a \nsigniﬁcant enrichment ( adjusted p = 1.32x10 -8) compared to the background of all genes  \n(Figure 1A, Supplementary Table 2). \nTo examine the global expression trends for defence-related genes, in a different way, \nwe identiﬁed 1033 genes that included ‘defence’ or ‘defense ’ in their TAIR annotation \ndescription (Reiser et al, 2024). Of these genes, 86 were differentially expressed in at least \none condition. Plotting the zero-centred fold change of these genes shows the extent of \nexpression recovery in the VIRc complementation line (Figure 1B). For example, expression \nof the defence marker gene PR1 (AT2G14610) was 311-fold (8.28 log2FC) higher in vir-1 than \nCol-0 but restored to similar levels as Col -0 in the VIRc complement ation line. The \nupregulation of PR1 in vir-1 is also detected in the orthogonal ONT DRS data (Figure 1C-E). \nWe next asked if other m6A writer mutants had elevated PR1 expression. We analysed ONT \nDRS data of ﬁp37-4 mutants that disrupt the Arabidopsis m6A writer complex  ortholog of \nWTAP (Parker et al, 2022). Genes which are signiﬁcantly upregulated in ﬁp37-4 mutants are \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted February 23, 2025. ; https://doi.org/10.1101/2025.02.18.638636doi: bioRxiv preprint \n\n 7 \nsigniﬁcantly enriched for GO terms related to defen ce (Supplementary Table 3). Like vir-1 \nmutants, ﬁp37-4 mutants have elevated PR1 expression (Figure 1F). In contrast, PR1 is not \nupregulated in ﬁona1 mutants that dis rupt the Arabidopsis N6 methyladenosine \nmethyltransferase METTL16 ortholog that adds m6A to U6 snRNA (Parker et al, 2022) (Figure \n1C). We conclude that a phenotype of Arabidopsis mRNA m6A writer complex mutants is the \nupregulation of genes involved in defence signalling. \n \nLoss of the mRNA m6A writer complex triggers temperature-sensitive autoimmunity  \nThe upregulation of defence response genes  in vir-1 raised the question of whether \nautoimmunity might explain  the gene expression changes and pleiotropic developmental \nphenotypes of mRNA m6A writer complex mutants. A hallmark of Arabidopsis autoimmunity is \nits temperature  sensitivity, with autoimmune phenotypes suppressed at elevated ambient \ntemperatures (Alcázar & Parker, 2011) . Therefore, to address whether the gene expression \nchanges we detected in vir-1 mutants reﬂected an autoimmune response, we compared the \ngene expression proﬁles of vir-1 and WT Col-0 seedlings grown in sterile conditions at 17°C \nand 27 °C. We used a combination of Illumina RNA -seq and ONT DRS to analyse  gene \nexpression. We performed four biological replicates with each RNA sequencing technology, \ngenotype and temperature treatment . The resultant sequencing statistics are detailed in \nSupplementary ﬁle 1. \nWe analysed the Illumina RNA-seq data using a quasi-likelihood model (glmQLFit) in \nedgeR (Chen et al, 2016) and identiﬁed genes differentially expressed between vir-1 and WT \nCol-0 at each temperature.  This revealed 1215 genes which were signiﬁcantly upregulated \n(adj.p < 0.001, log2FC > 2.0) in vir-1 at 17°C (Supplementary Table 4). Remarkably, 91% of \nthese genes (1103) were not signiﬁcantly upregulated in vir-1 at 27°C (Figure 2A), revealing \nthat most gene expression changes in vir-1 are temperature-sensitive. Principal component \nanalysis (PCA) separates the biological replicates by genotype and temperature . The ﬁrst \ncomponent, which explains 40% of the variance, captures gene expression changes speciﬁc \nto vir-1 at 17°C. In contrast, vir-1 and Col-0 are indistinguishable at 27°C in this component \n(Figure 2B). Likewise, correlation matrix analysis reveals that the gene expression features of \nvir-1 mutants grown at 1 7°C are the most distinct among  all the datasets (Figure 2C). We \nfound elevated expression of PR1 in vir-1 grown at 17°C, just as we had previously seen at \n22°C, but PR1 expression was at similar levels to WT Col-0 in vir-1 grown at 27 °C (Figure \n2D). We also detected this differential PR1 expression pattern with orthogonal ONT DRS data \n(Supplementary Figure 2A) and RT-qPCR (Supplementary Figure 2B).  \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted February 23, 2025. ; https://doi.org/10.1101/2025.02.18.638636doi: bioRxiv preprint \n\n 8 \nWe used a generalised linear model (GLM) to model all conditions simultaneously and \nidentify differential gene expression speciﬁc to vir-1 at 17°C. The GLM design contrasts vir-1 \nat 17°C minus the average of Col-0 at 17°C and 27°C and vir-1 at 27°C. This model identiﬁed \n931 genes with signiﬁcantly increased expression (adj.p < 0.001, log2FC>2.0) in vir-1 at 17°C \ncompared to the other conditions (Supplementary Figure 2C, Supplementary Table 5). GO-\nterm analysis revealed that t he biological processes most signiﬁcantly enriched in genes \nupregulated in vir-1 at 17°C were related to defence responses (Figure 2E, Supplementary \nTable 6). The defence annotation GO  terms were similar in describing the gene expression \nchanges previously detected in vir-1 grown at 2 2°C (Supplementary Figure 2D). As an \northogonal approach, we analysed protein domain enrichment using DAVID (Sherman et al, \n2022) (Supplementary Table 6). The most signiﬁcantly enriched protein domains  in genes \nupregulated in vir-1 at 17 °C were RLP23-like, which is found in receptor-like kinases that \nfunction as PRR proteins (15-fold enrichment) and Defensin_plant, a domain found in highly \nexpressed marker proteins of defence ( 10-fold enrichment). Next, we  used a different \napproach to ask how defence gene expression was affected by temperature using the GLM \nanalysis. We plotted the zero-centred log2FC normalised gene expression of 136 genes with \ndefence/defense included in their TAIR annotation  in vir-1 at 17°C compared to other \nconditions. This analysis reveals that the expression level of most genes with a TAIR \ndefence/defense annotation in vir-1 at 17 °C is at WT Col-0 levels when vir-1 mutants are \ngrown at 27°C (Figure 2F).  \nTo determine whether the major defence gene expression differences observed in vir-\n1 at 17 °C compared to all other conditions might be explained by overlooked pathogen \ncontamination of our experimental material , we examined our RNA -seq data for non -\nArabidopsis sequences . We used a ll vir-1 Illumina RNA-sea data to produce a de novo  \ntranscriptome assembly and searched the resulting contigs against the GenBank NR (non-\nredundant) database using BLASTP (Camacho et al, 2009). No signiﬁcant enrichment of plant \npathogen sequences was found in vir-1 17°C samples (Supplementary Table 7).  \nIn summary, by exploiting  the established temperature sensitivity of Arabidopsis \nimmunity, we discovered that the major annotation terms associated with the upregulated \ngenes in vir-1 mutants at 17°C are related to defence. Therefore, at the gene expression level \nand strikingly dependent on temperature, we conclude that autoimmunity is a major phenotype \nof the Arabidopsis mRNA m6A writer complex mutant, vir-1. \n \nGenes that function in diverse aspects of immunity are upregulated in vir-1 mutants \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted February 23, 2025. ; https://doi.org/10.1101/2025.02.18.638636doi: bioRxiv preprint \n\n 9 \nGiven the global trend of defence gene activation in vir-1 mutants, we next analysed individual \ngene expression changes to understand what type of defence gene s were affected. \nExpression of mRNA encoding the master defence transcription factor SARD1  (Sun et al, \n2015; Wang et al, 2011) was upregulated (AT1G73805: log2FC 4.0) in vir-1 at 17°C but not \n27°C, consistent with the established temperature sensitivity of SARD1 transcription (Kim et \nal, 2022) (Figure 3A and Supplementary Figure 3A). The expression of mRNA encoding the \nFLS2 receptor, which detects the bacterial ﬂagellin ﬂg22 peptide (Gómez-Gómez & Boller, \n2000) and is one of the best-characterised Arabidopsis PRR proteins, was upregulated in vir-\n1 at 17°C (AT5G46330: log2FC 3.0) but not at 27°C (vir_27 vs Col-0_27: log2FC 1.3) (Figure \n3B and Supplementary Figure 3B). We detected the upregulation of 15 annotated NLRs - 13 \nTIR-NLRs, one C C-NLR (LOV1) and one RPW8-NLR (HR4) at 17 °C but not 27 °C \n(Supplementary Table 8). For example, the TIR -only TX0 was upregulated (AT1G57630: \nlog2FC 4.1) (Figure 3C and Supplementary Figure 3C ). TX0 can hydrolyse nucleic acids, \nparticularly RNA, and synthesise 2’,3’,-cAMP/cGMP molecules that ultimately signal cell death \nin the hypersensitive response (Yu et al, 2022).  \nWe next asked whether genes previously associated with Arabidopsis autoimmunity \nwere misregulated in vir-1. The TIR-NLR RPS6 (AT5G46470) is recurrently associated with \nautoimmunity. For example, the extreme phenotypes of Arabidopsis nonsense-mediated RNA \ndecay ( NMD) and mitogen -activated kinase mutants have been attributed to  RPS6 \n(Gloggnitzer et al , 2014; Takagi et al , 2020) , although  the mechanisms involved are not \nunderstood (Gloggnitzer et al, 2014; Takagi et al, 2020; Parker et al, 2021b). RPS6 expression \nis not signiﬁcantly altered in vir-1 RNA-seq data, but ONT DRS analysis indicates that the TIR-\nonly gene located downstream of the RPS6 locus is upregulated (Parker et al , 2021b)  \n(Supplementary Figure 3D). The TIR-NLR SNC1 has been used as a model to understand \nautoimmune signalling (Zhu et al, 2010); SNC1 was not signiﬁcantly upregulated in vir-1 at \n17°C (log2FC 1.22), but SIDEKICK3, a TIR-NLR required for SNC1-mediated autoimmunity \n(Dong et al , 2018) , is one of the most upregulated genes (log2FC 8.32) in vir-1 at 17 °C \n(Supplementary Table 8). Finally, we examined ACD6, which encodes a multipass  \ntransmembrane protein with intracellular ankyrin repeats , that mediates a trade -off between \ngrowth and defence (Chen et al, 2023). First identiﬁed in lab-based mutant screens (Rate et \nal, 1999; Lu et al, 2003), high-activity ACD6 alleles are frequently found in natural Arabidopsis \naccessions (Świadek et al, 2017; Zhu et al, 2018; Todesco et al, 2010). ACD6 is upregulated \nin vir-1 at 17°C (log2FC 6.27), but the expression level is similar to WT Col-0 in vir-1 grown at \n27°C (Figure 3D and Supplementary Figure 3E). We asked whether the changes in gene \nexpression between vir-1 and acd6 mutants were related by re-analysing a recently published \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted February 23, 2025. ; https://doi.org/10.1101/2025.02.18.638636doi: bioRxiv preprint \n\n 10 \nacd6-1 Illumina RNA-seq dataset (Fabian et al, 2023). We found a subset of differentially \nexpressed genes overlap between these two mutants; 205 upregulated genes and 21 down-\nregulated genes in common (Figure 3E) . However, 931 genes are uniquely upregulated in \nacd6-1 and 701 genes are uniquely upregulated in vir-1 at these thresholds (adj.p < 0.001, \nlog2FC>2.0), indicating that the misregulation of  ACD6 alone does not simply explain  vir-1 \nautoimmunity gene expression phenotypes. \n A group of ﬂowering time genes paralleled the expression of immune response genes. \nThe ﬂoral pathway integrator, FT, (Kinoshita & Richter, 2020) is upregulated in vir-1 at 17°C \nbut not 27°C (Supplementary Table 8). In addition, the expression of a group of genes known \nto function downstream of FT in ﬂoral development, including FUL, SOC1, SEP3, SPL4, SPL5, \nAGL19 and AGL24 phenocopied defence gene expression patterns (Supplementary Table 8, \nSupplementary Figures 3F-M).  \nIn summary, diverse genes attributed to the different ETI and PTI layers of Arabidopsis \ninnate immunity were upregulated in vir-1 mutants at 17°C but not 27°C, together with mRNA \nencoding the SARD1 master defence transcription factor that controls SA-dependent and SA-\nindependent defence responses. \n \nvir-1 mutants exhibit temperature -sensitive pathogen resistance and localised cell \ndeath  \nA characteristic of autoimmunity is that  plants show enhanced resistance to pathogen s \nbecause defence gene expression is already upregulated  prior to infection. We, therefore, \nasked whether the global changes in gene expression detected in vir-1 resulted in a functional \nimpact on  pathogen infection. We examined the susceptibility of WT Col -0, vir-1 and ﬂs2 \n(ﬂagellin sensing 2) mutants to the biotrophic pathogenic bacterium Pseudomonas syringae \npv. tomato (Pto) DC3000. The ﬂs2c mutant allele (Zipfel et al, 2004) lacks the receptor for \nbacterial ﬂagellin and is susceptible to infection, so is a positive control for infection in these \nexperiments. We ﬂooded seedlings grown at 17°C, 21°C and 27°C with P. syringae pv. tomato \n(Pst) DC3000. We found that ﬂs2c mutants were more susceptible than Col -0 to infection at \nall temperatures , consistent with previous reports  (Zipfel et al , 2004)  (Figure 4A, \nSupplementary Figure 4 , Supplementary Table 9 ). In contrast, vir-1 mutants were more \nresistant to infection than Col -0 when grown at 17°C and 21°C, but there was no signiﬁcant \ndifference in infection levels between Col-0 and vir-1 plants grown at 27 °C. Therefore, the \ntemperature-sensitive patterns of defence gene expression detected in vir-1 convert to a \ncorresponding change in immunity. \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted February 23, 2025. ; https://doi.org/10.1101/2025.02.18.638636doi: bioRxiv preprint \n\n 11 \n Localised premature cell death is a phenotype of plant immune responses to infection, \nand leaf lesions are a feature of some autoimmune genotypes (van Wersch et al, 2016; Freh \net al , 2022) . To investigate whether vir-1 mutants exhibited elevated levels of cell death \ncompared to WT Col-0 in the absence of pathogen infection , we stained seedlings grown in \nsterile conditions with the vitality marker  trypan blue (TB). We recorded microscopy images \nand quantiﬁed the levels of detectable TB staining for 10 individual leaves of each genotype \ngrown in each condition using ImageJ (Schneider et al, 2012). We detected the highest levels \nof cell death in vir-1 grown at 17°C (Figure 4B and 4C, Supplementary Table 10). However, \nat 27°C, cell death patterns in vir-1 were at negligible levels, comparable to those detected in \nWT Col-0 at 17°C and 27°C. We, therefore, conclude that vir-1 mutants show elevated levels \nof premature cell death at 17°C when immune response pathways are autoactivated. \n Overall, vir-1 mutants’ temperature-sensitive response to two orthogonal analyses of \nautoimmunity - enhanced pathogen resistance and increased premature cell death  - is \nconsistent with the global patterns of gene expression changes we detect at the RNA level. \nThese ﬁndings suggest that the mRNA m6A writer complex is required to dampen defence \npathway signalling to prevent autoimmunity in the absence of pathogen s but not for the  \ndefence responses  that suppress P. syringae  pv. tomato ( Pto) DC3000 infection. We, \ntherefore, conclude that vir-1 mutants have a temperature-sensitive autoimmune phenotype. \n \nmRNA m6A levels in vir-1 mutants are not temperature-sensitive \nAn ethyl methanesulfonate-induced 5’ splice site mutation in intron 5 (G+1 to A) causes the \nvir-1 allele, resulting in cryptic splicing events within exon 5 that disrupt the VIR open reading \nframe (Růžička et al, 2017). Since mutations that disrupt splic e sites can be temperature -\nsensitive (Sablowski & Meyerowitz, 1998) , we asked if the expression of VIR mRNA was \nrestored at 27°C. However, we found no evidence from our RNA-seq data to support this idea \n(Supplementary 5A and 5B). We next asked whether mRNA m6A levels in vir-1 mutants were \nrestored to wild -type levels  at 27°C. We ﬁrst used liquid chromatography -tandem mass \nspectrometry (LC-MS/MS) to analyse the m 6A/A (adenosine) ratio in poly(A)+ RNA puriﬁed \nfrom Col-0 and vir-1 grown at 17°C and 27°C. The level of poly(A)+ RNA m6A modiﬁcation in \nthe hypomorphic vir-1 allele was reduced to approximately 10% of that detected in Col-0 at \nboth 17°C and 27°C (Figure 5A, Supplementary Table 11), consistent with previous reports of \nreduced m6A levels in vir-1 mutants (Růžička et al, 2017; Parker et al, 2020). In addition to \nLC-MS/MS, we used the ONT DRS data to examine mRNA m6A levels. We have previously \nmapped m6A in ONT DRS data using the Differr and Yanocomp programs, which depend upon \ncomparing WT and mutant genotypes (Parker et al , 2020, 2021a) . We supplemented \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted February 23, 2025. ; https://doi.org/10.1101/2025.02.18.638636doi: bioRxiv preprint \n\n 12 \nYanocomp analysis with m6Anet, a neural-network-based method that can call read-level m6A \nstoichiometry without genotype comparison (Hendra et al, 2022). We found no restoration of \nm6A levels identiﬁed by m6Anet (Figure 5B) or Yanocomp analysis (Supplementary Figure \n5C) in vir-1 at 27°C compared to 17 °C. We conclude that the suppression of immune gene \nexpression detected in vir-1 at 27°C is not explained by a n accompanying change in mRNA \nm6A modiﬁcation. \n \nThe visible developmental defects of vir-1 mutants are separable from autoimmune \ngene expression \nArabidopsis autoimmune mutants characteristically show developmental defects that can be \nrescued by growth at elevated temperatures (Alcázar & Parker, 2011). We asked whether the \nimpact of autoimmunity might explain the developmental phenotypes of vir-1. We examined \nthe development of Col-0 and vir-1 mutant plants from germination to ﬂowering and seed-set \nat 17 °C and 27 °C. We included acd6-1 as a positive control for an autoimmune mutant \ncompromised in development at 17 °C, which appears more like WT Col-0 when grown at \n27°C. We found that the short stature and lack of apical dominance characterising the visible \ndevelopmental phenotypes of vir-1 mutants were not rescued by growth at 27°C (Figure 6A). \nHowever, we could replicate the previously reported developmental rescue of acd6-1 mutants \nat 27 °C compared to 17 °C (Supplementary Figure 6A). We, therefore, conclude that the \nimpact that loss of the mRNA m 6A writer complex has on development and defence gene \nexpression programmes is separable.  \nNext, we asked if we could detect gene expression changes that might underpin \ndevelopmental change in vir-1 mutants. We identiﬁed signiﬁcant gene expression differences \nconsistently affected by the vir-1 mutation across all our datasets, irrespective of temperature \n(adj.p < 0.001, log2FC>2.0) (Supplementary Table 12). The ﬂowering time control gene, FLC, \nwas the only gene signiﬁcantly downregulated across all our vir-1 experimental conditions. \nOnly 58 genes were consistently upregulated. However, the most enriched GO term biological \nprocesses for these genes were  “defence response ”, “systemic acquired  resistance”, and  \n“defence response to other organism” (Supplementary Figure 6B, Supplementary Table 13), \nindicating that defence gene upregulation remains an important vir-1 gene expression \nphenotype. For example, RNA encoding the ﬂavin-dependent monooxygenase, FMO1, was \nupregulated in vir-1 relative to WT Col-0 in all tested conditions (Supplementary Figure 6C). \nFMO1 is a critical regulator of systemic acquired resistance to pathogen infection (Hartmann \net al, 2018). We also ident iﬁed consistent upregulation of the AtNUDT24 nudix hydrolase. \nAtNUDT24 is uncharacterised, but other members of this protein family function to modify \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted February 23, 2025. ; https://doi.org/10.1101/2025.02.18.638636doi: bioRxiv preprint \n\n 13 \nsignalling nucleotides in plant defence  (Yu et al , 2022; Bartsch et al , 2006)  or in RNA \ndecapping and hydrolysis, among other roles (Laudenbach et al, 2021; Carreras-Puigvert et \nal, 2017).  \nOverall, we conclude that although defence gene activation in vir-1 mutants is \nsuppressed at 27°C, visible developmental defects are not, and autoimmune gene expression \nprogrammes remain a component of the vir-1 mutant phenotype.  \n \nAltered poly(A) tail length distributions are a temperature-sensitive phenotype of vir-1 \nmutants \nTo understand how the loss of mRNA m 6A might trigger an autoimmune response , we next \nasked if RNA processing was affected in vir-1 mutants grown at different temperatures . \nConsistent with our previous reports (Parker et al , 2020, 2022) , we found that genetic \ndisruption of the mRNA m6A writer complex resulted in a global shift to proximal poly(A) site \nusage in diverse mRNAs (Figure 7A). The scale of this effect was similar in vir-1 mutants \ngrown at 17°C and 27°C. Altered 3’ end formation may trigger  autoimmunity, bu t the \nsubsequent signalling might be suppressed at 27°C, so we addressed this question differently. \nIf alternative poly(A) site usage triggered the autoimmune response, it might be mediated by \nthe CPSF30-YTH m6A reader. We examined gene expression changes in cpsf30-yth mutants \nusing ONT DRS . The cpsf30-yth mutants express a truncated protein comprising the N -\nterminal Zinc-Finger domains that bind the AAUAAA poly(A) signal,  but lack the  C-terminal \nYTH domain. We found PR1 is not upregulated in cpsf30-yth, although it is upregulated in \nﬁp37-4 mutants analysed alongside here as a positive control  (Figure 7B). Therefore, this \ncombination of data does not provide evidence that altered poly(A) site usage triggers the \nautoimmunity phenotypes of Arabidopsis mRNA m6A writer complex mutants.  \n The cytoplasmic YTH reader domain proteins likely mediate speciﬁc impacts of m6A \non RNA fate (Brodersen & Arribas-Hernández, 2024). The most abundant of these are ECT2, \n3 and 4  (Arribas-Hernández et al, 2018). We analysed a triple mutant in each gene, te234 \n(Flores-Téllez et al, 2023) using ONT DRS. We found no evidence that PR1 was upregulated \nin te234. In contrast, PR1 upregulation was again detected in ﬁp37-4 mutants included here \nas a positive control (Figure 7C).  \nFinally, we asked if poly(A) tail length was altered in vir-1 mutants. We have previously \nreported a change in poly(A) tail length proﬁles at speciﬁc genes in vir-1 (Parker et al, 2020), \nand we asked if this was a more widespread phenotype. We used the ONT software Dorado \nto estimate transcript poly(A) tail length in our ONT DRS data. This analysis reveals a \ncharacteristic periodicity in estimated lengths of read poly(A) tails that likely reﬂects the \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted February 23, 2025. ; https://doi.org/10.1101/2025.02.18.638636doi: bioRxiv preprint \n\n 14 \nfootprint of binding of multiple  poly(A) binding proteins  (PABPs) (Baer & Kornberg, 1983; \nPassmore & Coller, 2022). We found that the distribution of estimated poly(A) tail lengths was \nmarkedly different in vir-1 compared to WT Col -0 - at 17°C; relatively fewer transcripts with \nshort poly(A) tails and more with longer poly(A) tails are detected in vir-1 compared to WT-\nCol-0 (Figure 7D). At 27°C, vir-1 mutant mRNA poly(A) tail length proﬁles are different again, \nwith a distribution more enriched in short poly(A) tails compared to WT Col -0 (Figure 7D). In \ncontrast, the 30 nt poly(A) tail of Saccharomyces cerevisiae ENOLASE II RNA, used here as \na spike-in calibration standard during ONT DRS library preparation , is consistent across all \nsamples, with an estimated median tail length of 33 nt (Supplementary Figure 7A). \nChanges in the global distributions of poly(A) tail length may be partly due to \ndifferences in the sets of genes expressed at detectable levels in the different conditions. To \nexamine the differences in tail length per gene and exclude genes that were only  expressed \nin single conditions, we used the Wasserstein distance metric to quantify shifts in mean per \ngene tail length distributions between conditions (Parker et al, 2021b). This analysis identiﬁed \na shift towards longer mean poly(A) tails in vir-1 at 17°C compared to Col-0 and a shift towards \nshorter mean poly(A) tails in vir-1 at 27 °C, consistent with the different distributions of \nestimated poly(A) tails lengths (Figure 7E, Supplementary Table 14). Temperature-sensitive \nchanges in mean poly(A) tail length in Col -0 were modest by comparison (Figure  7F). We \nasked if the shift in mean poly(A) tail length was directly associated with the loss of m 6A \nmodiﬁcation. We found that while genes with m6Anet-predicted m6A sites had shorter poly(A) \ntails, the temperature -dependent differences in poly(A) tail length distribution were seen in \ngenes predicted to be either m6A-modiﬁed or non-modiﬁed (Supplementary Figure 7B).  \nIn summary, our ﬁndings reveal global changes in poly(A) tail length distributions as \nthe primary temperature-sensitive mRNA processing phenotype of vir-1 mutants. This global \nphenotype is not restricted to genes that function in defence responses or to those transcripts \npredicted to have lost m6A modiﬁcation.  \n \n  \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted February 23, 2025. ; https://doi.org/10.1101/2025.02.18.638636doi: bioRxiv preprint \n\n 15 \nDISCUSSION \n \nWe have discovered that disruption of the mRNA m 6A modiﬁcation complex triggers \nautoimmunity in Arabidopsis. Using a combination of Illumina RNA -seq and ONT DRS, we \nreveal, in unprecedented detail, the temperature -sensitive changes in RNA expression, \nmodiﬁcation and processing caused by mRNA m 6A writer complex depletion. At 17°C, most \ngene expression class changes in vir-1 mRNA m6A writer complex mutants are explained by \ndefence gene activation. By exploiting the established temperature sensitivity of Arabidopsis \nimmunity, we found that the vast majority of these gene expression changes did not occur  \nwhen we grew the vir-1 mRNA m6A writer mutant at 27°C. We used orthogonal experimental \napproaches to examine the biological impact of this defence gene activation. We found \ntemperature-sensitive enhanced levels of premature cell death and resistance to P. syringae \nPst D3000 pathogen infection in vir-1 mutants consistent with the idea that these gene \nexpression changes convert into functional consequences for immunity. Not all gene \nexpression changes , or all developmental phenotypes of vir-1 mutants, were rescued by \nelevated temperature, demonstrating that the impacts of m 6A-dependent changes on gene \nexpression are separable. Together, these different lines of evidence all point to the disruption \nof the mRNA m6A writer complex triggering autoimmunity in Arabidopsis. \n The autoimmune phenotypes of Arabidopsis m 6A writer complex mutants have not \npreviously been described. However, while analysing the impact of mRNA m6A on plant \npathogen infection, it was recently reported that Arabidopsis mutants defective in METTL3, \nVIR and WTAP orthologs, and a line ectopically overexpressing the mRNA m6A demethylase \nALKBH10B, were all more resistant to infection by P. syringae  DC3000, P. syringae pv \nmaculicola and a fungal pathogen Botrytis cinerea (Prall et al, 2023). Consistent with this, \ngenes commonly upregulated in METTL3 ortholog mutant and ectopic ALKBH10B expression \nlines are enriched in GO term annotations for d efence signalling (Prall et al , 2023) . \nFurthermore, early microarray analysis of METTL3 ortholog mutants reported a general \nenrichment among the overexpressed genes for GO terms related to stress responses (Bodi \net al, 2012). These independent ﬁndings are consistent with our analysis of ﬁp37-4 and vir-1 \nand generalise the idea that autoimmunity is a major Arabidopsis mRNA m6A writer complex \nmutant phenotype. \nThese discoveries raise the key question of how disruption of mRNA m 6A writer \ncomplexes triggers autoimmunity. Our analysis does not explain how defence pathways are \nautoactivated in Arabidopsis mRNA m6A writer complex mutants. Remarkably, given its \nimportance, the mechanism by which modiﬁed RNAs ablate TLR signalling in humans is also \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted February 23, 2025. ; https://doi.org/10.1101/2025.02.18.638636doi: bioRxiv preprint \n\n 16 \nunknown. We ﬁrst considered whether m 6A readers might mediate the autoactivation of \ndefence gene expression. However, we found no evidence that either cpsf30-yth mutants or \nte234 mutants phenocopied the defence gene expression of vir-1 mutants. There are 13 m6A-\nreading YTH domain-containing protein s encoded in the Arabidopsis genome. Therefore, \napproaches that resolve redundancy in their function (Flores-Téllez et al, 2023) will be required \nto test an association with autoimmunity further . The major temperature -sensitive RNA \nphenotype we detected in vir-1 mutants was a global change in mRNA poly(A) tail length: we \nfound relatively fewer RNAs with short poly(A) tails (<40 nt) and relatively more with longer \npoly(A) tails (>100nt) compared to WT Col-0 at 17°C but this phenotype was reversed at 27°C. \nThis ﬁnding could indicate that the trimming of longer poly(A) tails  - an essential feature of \nnewly transcribed mRNAs (Passmore & Coller, 2022) - is defective and/or mRNAs with short \npoly(A) tails are more susceptible to decay in vir-1 mutants at 17°C. Poly(A) tails can reduce \ninnate immun e responses of human cells  to RNA  (Koski et al , 2004) . Furthermore, \nautoimmunity and pleiotropic developmental defects are phenotypes of Arabidopsis mutants \ndefective in the nuclear poly(A) polymerase, PAP1, that catalyses mRNA poly(A) tail formation \n(Vi et al, 2013). However, the inter-relatedness of the phenotypes we detect here is not yet \nclear. Poly(A) tail length changes were not restricted to transcripts with predicted m 6A sites, \nindicating this change is not a direct result  of m6A loss. Nor were poly(A) tail length changes \nrestricted to transcripts involved in defence functions. Increased poly(A) tail length is a stress \nresponse phenotype  of different species  (Tudek et al , 2021; Yamagishi et al , 2024)  and \npromotes stress granule  formation in humans  (Yamagishi et al , 2024; Tsai et al , 2025) . \nTherefore, the temperature -sensitive poly(A) tail length changes  we detect here may be a \npreviously unappreciated autoimmunity phenotype.  An aspect of RNA biology we have not \nexplored is whether changes in condensate association caused by loss of mRNA m6A might \ntrigger autoimmunity. mRNA m6A modiﬁcations can  inﬂuence separation into biomolecular \ncondensates such as human stress granules (Ries et al, 2023). Signiﬁcantly, the buffering of \nself RNA by condensates regulates human innate immune responses (Maharana et al, 2022). \nAnalysing other proteins more closely connected to Arabidopsis poly(A) tail processing and \nRNA fate (Sasse et al , 2024)  could help unravel connections between  RNA modiﬁcation, \npoly(A) tail length, altered RNA homeostasis and the causes or consequences of \nautoimmunity.  \nDifferent receptors detect RNA as a molecular signature of pathogen infection in \nhumans, and RNA's availability, localisation, and structure (including sequence and \nmodiﬁcation) are essential criteria for distinguishing self and non -self (Schlee & Hartmann, \n2016; Bartok & Hartmann, 2020) . For example, uridine mononucleotides  and di or \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted February 23, 2025. ; https://doi.org/10.1101/2025.02.18.638636doi: bioRxiv preprint \n\n 17 \ntrinucleotides are bound on different sites of TLR8 in monocyte endosomes in a manner that \ndepends on upstream RNAse 2 and T2 processing of pathogen RNA (Ostendorf et al, 2020). \nRNAs puriﬁed from P. syringae DC3000 bacteria or transcribed in vitro and inﬁltrated into \nArabidopsis leaves activate innate immunity, demonstrating that non -self RNAs can trigger \nimmune responses in Arabidopsis (Lee et al, 2016). Aside from the RNAi machinery, which \nplays crucial roles in viral defence in plants (Lopez-Gomollon & Baulcombe, 2022), receptors \nthat recognise pathogen RNAs are poorly characterised in plants. However, plant TIR domains \nhave recently been found to hydrolyse RNA  (Yu et al, 2022). Compared to DNA, RNA is the \npreferred substrate for TIR domains to synthesise 2’,3’ ,-cAMP/cGMP molecules that signal \ncell death in the hypersensitive response (Yu et al, 2022). Crucially, a mutation that disrupts \nthis synthetase activity is sufﬁcient to block cell death signalling (Yu et al, 2022). TIR domains \nare frequently found in Arabidopsis NLR proteins and can also be expressed as TIR-only \nproteins (Essuman et al, 2022) encoded by TIR -only genes or the widespread prematurely \nterminated transcripts of NLR genes (Parker et al, 2021b). We found that mRNAs encoding \nTIR domain proteins, including TX0, for which RNA hydrolysing activity has been \ndemonstrated (Yu et al, 2022), were signiﬁcantly upregulated in vir-1 mutants. The natural \nRNA substrates of Arabidopsis TIR domains are unknown. An important question is whether \nTIR domains sense non -self RNAs or perturbed RNA homeostasis that indicates pathogen \nactivity. The nudix hydrolase family member NUDT7 acts as a phosphodiesterase to modify \n2’3’-cAMP/cGMP and thus modulates signalling through EDS1 (Yu et al, 2022). Notably, one \nof the most consistently upregulated genes in vir-1 is the uncharacterised nudix hydrolase, \nAtNUDT24.  \nIt may not be the loss of mRNA m6A itself that triggers autoimmunity. ETI functions to \ndetect pathogen activity that disrupts host cell proteins or processes and activates an immune \nresponse. In this way, NLRs act as guards, with the effector-targeted host cell proteins or \nprocesses being guardees (Dangl & Jones, 2001; Van der Biezen & Jones, 1998; Remick et \nal, 2023). It is possible that the mRNA m6A writer complex is a guardee and that disruption of \nthe writer complex, rather than the absence of mRNA m 6A, is detected and triggers \nautoimmune signalling. Therefore, defence signalling pathways in vir-1 mutants may directly \ndetect non-modiﬁed RNA, a disrupted mRNA m6A writer complex, poly(A) tail perturbation, or \nchanged RNA homeostasis resulting from decay  or condensate association . However, the \nconnection between disrupted m6A writer complex and autoimmunity may be even more \nindirect. Loss of m 6A is associated with diverse changes in gene expression and pleiotropic \ndevelopmental changes. Therefore, if the changes in gene expression, RNA processing, or \ndevelopment found in mRNA m6A writer complex mutants phenocopy features of pathogen \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted February 23, 2025. ; https://doi.org/10.1101/2025.02.18.638636doi: bioRxiv preprint \n\n 18 \ninfection, they may indirectly trigger immune pathway activation. Consequently, understanding \nhow immune gene expression is activated is crucial to understanding the directness by which \nmRNA m6A impacts these diverse RNA and developmental phenotypes.  \nOur ﬁndings, therefore, have practical implications for studying the impact of mRNA \nm6A on plant biology. Since most gene expression changes in mRNA m6A writer complex \nmutants at lower ambient temperatures are caused by autoactivation of defence gene \nexpression, the interpretation of the effects of mRNA m6A will be complicated by indirect \nchanges that vary according to environmental conditions (such as temperature), which may \ndiffer between studies. Mutating defence signalling hubs in mRNA m6A writer complex mutant \nbackgrounds might suppress autoimmunity but not necessarily  comprehensively block \nautoimmune signalling. Our global transcriptome analysis only provides snapshots of gene \nexpression changes in  Arabidopsis seedlings . However, u nderstanding how mRNA m 6A \ndirectly inﬂuences mRNA processing and fate and, hence, development and autoimmunity will \nrequire alternative experimental approaches . Determining immediate gene expression \nchanges following mRNA m6A writer complex shutdown using, for example,  proteolysis \ntargeting chimaeras (PROTACS) (Békés et al, 2022) in deﬁned cell types and developmental \ncontexts may help us understand the direct roles of mRNA m6A.  \nIn conclusion, our study establishes a new conceptual framework for analysing the \nimpact of mRNA m6A on plant biology. The molecular basis of the events that trigger mRNA \nm6A writer complex-dependent autoimmunity is unknown, but uncovering this should lead to \nfundamental insights into the role of mRNA m6A in plant biology and how defence gene \nsignalling occurs.  \n  \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted February 23, 2025. ; https://doi.org/10.1101/2025.02.18.638636doi: bioRxiv preprint \n\n 19 \nMATERIALS and METHODS \n \nPlant Material  \nWild-type Arabidopsis thaliana accession Col-0 and te234 (Arribas-Hernández et al, 2020) \nwere obtained from the Nottingham Arabidopsis Stock Centre. The vir-1 and VIR -\ncomplemented (VIR::GFP -VIR) lines (Růžička et al , 2017)  were from K. R ůžička, Brno , \nCzechia; acd6-1 was from J. Greenberg, Chicago, USA; ﬁp37-4 was from R. Fray, \nNottingham, UK; ﬂs2c (SAIL_691C4) (Zipfel et al, 2004) was from P. Hemsley, Dundee, UK; \ncpsf30-yth (GK477H04) was from Ł. Szewc, Poznań, Poland.  \n \nPlant Growth Conditions \nSeeds were sown on MS10 medium plates, stratiﬁed at 4°C for 2 days, germinated in a \ncontrolled environment at 22°C under 16 hr light/8 hr dark conditions and harvested  for RNA \npuriﬁcation 14 days after transfer to 22°C.  For temperature assays, plant growth chambers \nwere set to either 17°C or 27 °C, with all other conditions the same as above. Seedlings were \nharvested 14 days after germination during the ﬁrst two hours of the light period following an \n8-hour dark phase. Four-week-old plants were used for phenotyping the adult plants at 17°C \nor 27 °C. \n \nTrypan Blue Staining \nTrypan blue staining was performed on leaves from 4-week-old plants of WT Col-0 and vir-1 \ngrown at 17°C and 27°C. Leaves were stained in a solution of Tris-EDTA equilibrated phenol \n(pH 8) (25%), glycerol (25% v/v), lactic acid (25% v/v) with trypan blue ( 10mg/ml). Leaves \nwere treated with staining solution for 10 minutes at 95 °C then incubated overnight at room \ntemperature. Leaves were destained in chloral hydrate solution twice, for 4 h and overnight. \nStained leaves were imaged under a Zeiss histology microscope at 10x magniﬁcation. Images \nwere imported into ImageJ (Schneider et al, 2012), and the total stained area was measured \nin pixels, with the stained area expressed as a percentage of the total leaf area. Data collected \nfrom 10 leaves per condition was plotted, and a two-way ANOVA test with post hoc Turkey’s \nHSD tests was used to assess the effects of genotype and temperature and their interaction \non the percentage of leaf stained.  \n \nGenotyping of Arabidopsis  \nIndividual mutants and progeny from crosses were genotyped by PCR analysis of puriﬁed \nDNA using a combination of derived cleaved ampliﬁed polymorphic sequence (dCAPS) \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted February 23, 2025. ; https://doi.org/10.1101/2025.02.18.638636doi: bioRxiv preprint \n\n 20 \nmarkers (Michaels & Amasino, 1998), T-DNA insertion markers and sequencing. Genotyping \nprimer pairs used in the study are listed in Supplementary Table 15. \n \nPathogenesis Assays \nArabidopsis Col-0, vir-1 and ﬂs2c seedlings were treated with Pseudomonas syringae pv \ntomato (Pto) DC3000 using ﬂood inoculation. Bacteria were cultured on MG agar \nsupplemented with rifampicin at 28 °C for 24-48 h. Four-week-old seedlings ﬂood-inoculated \nwith a bacterial suspension of P. syringae DC3000 (5x106 Colony-Forming Units (CFU)/ml) \ncontaining 0.025% Silwet L-77. Sterilised seedlings were grown on half-strength MS agar for \n2-3 weeks at 17°C, 21°C and 27°C before inoculation. The bacterial suspension was applied \nto the Arabidopsis seedlings, and plates were incubated at room temperature for 2-3 minutes. \nExcess liquid was drained, and seedlings were maintained at growing temperatures for three \ndays. Bacterial growth was calculated using serial dilution of material from three seedlings per \nplate and recorded as CFU/mg. The signiﬁcance of changes in bacterial growth in the differing \nconditions was tested using ANOVA. The experiment was repeated to conﬁrm results. \n \nRNA Isolation \nTotal RNA was isolated using the RNeasy Plant Mini kit (Qiagen) and treated with RNAse-free \nDNase (Promega-M6101). RNA concentration and integrity were measured using a NanoDrop \none spectrophotometer and Agilent 4150 Tapestation. \n \nGene expression analysis by RT-qPCR \nTotal RNA was extracted from 14-day-old seedlings. Total RNAs were treated with RNAse -\nfree DNase (Promega-M6101). First-strand cDNAs were synthesised using SuperScript™ III \nReverse Transcriptase (Thermo Fisher Scientiﬁc -12574026). qPCR was carried out on a \nLightCycler® 96 Instrument using Brilliant III Ultra -Fast SYBR Green qRT -PCR Master Mix \n(Agilent-600886). Three biological replicates (independently harvested samples) with three \ntechnical replicates for each were analysed. Relative expression levels were determined using \nthe 2 −ΔΔCT method. Arabidopsis UBQ10 (AT4G05320) was used as internal control. qPCR \nprimer pairs are listed in Supplementary Table 15.  \n \nPreparation of libraries for Illumina RNA-sequencing  \nIllumina RNA-seq libraries were prepared by Genewiz (Azenta LifeScience) using NEB Next \nUltra Directional  Library Prep Kit according to the manufacturer’s instructions. Paired -end \nsequencing with a read length of 150bp was carried out on the Illumina NovaSeq X following \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted February 23, 2025. ; https://doi.org/10.1101/2025.02.18.638636doi: bioRxiv preprint \n\n 21 \nthe manufacturer’s instructions. Raw sequence data was converted to fastq and de -\nmultiplexed using Illumina bcl2fastq version 2.20.  \n \nProcessing of Illumina RNA-seq data and differential gene expression analysis  \nQuality assessment of RNA-seq reads was performed using Fast QC (Andrews, 2017). For \ndigital expression, the Salmon index was built using Arabidopsis Araport11 transcript  \nannotations (Cheng et al , 2017) . Transcript and gene -level counts were estimated using \nSalmon (with gtf option and Araport 11 annotation) (version 1.9.0) (Krishnakumar et al, 2015). \nDifferential expression analysis was performed in edgeR (version 4.2.2) using a quasi -\nlikelihood generalised linear model (glmQLFit). Annotation of genes of interest, categorising \nthem as defence, ﬂowering or other and returning GO annotation with further annotation was \nperformed using custom scripts: \nhttps://github.com/bartongroup/PT_Arabidopsis_names_to_annot. To visualise dimensional \nreduction in the context of RNA -seq quality control , PCA plots were created using ggplot2 \n(version 3.5.1). Correlation and  heatmap plots were generated with the ptr script from the \nTrinity RNAseq package (version 2.15.2) (Grabherr et al, 2011). GO enrichment heatmap were \nmade using msbio (metascape)(v3.5.20240901) (Zhou et al, 2019). \nFunctional enrichment analysis was performed using a combination of Goseq (version \n1.42.0) (Young et al, 2010) and g:Proﬁler (version e111_eg58_p18_f463989d) with the g:SCS \nmultiple testing correction method and a signiﬁcance threshold < 0.05 (Kolberg et al, 2023). \nDomain enrichment analysis was performed in DAVID  (Sherman et al, 2022) using a FDR \nsigniﬁcance threshold of < 0.05.  \nTo determine whether pathogen contamination was present in the vir-1 RNA-seq \nsamples, reads were mapped to the TAIR10 genome, and from the resulting bam ﬁle, \nunmapped reads were returned using STAR (version 2.7.11b)  (Dobin et al, 2013). BBnorm \n(October 19, 2017) was then used to normalise the “unmapped” reads with the following \nsetting “ target=75 min=3 ” (Bushnell, 2022) . The normalised reads were assembled using \nTrinity (Grabherr et al , 2011)  (version 2.15.2) with –trimmomatic –no_normalise. The \ntranscriptome assembly was processed using cd-hit-est (version 4.8.1) (-c 0.90 -n 8 -T 24 -M \n0) (Fu et al, 2012a) to reduce redundancy at 90%. The resulting ﬁnal transcriptome assembly \nwas then searched against Genbank NR with Diamond-BLASTP using Diamond (version \nv2.0.5.143 ) (Buchﬁnk et al , 2021) . The diamond BLASTP output was post -taxonomically \nannotated using \nhttps://github.com/peterthorpe5/public_scripts/tree/master/Diamond_BLAST_add_taxonomic\n_info. The ﬁnal taxonomically assigned BLAST output was then interrogated for the presence \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted February 23, 2025. ; https://doi.org/10.1101/2025.02.18.638636doi: bioRxiv preprint \n\n 22 \nof plant pathogens, as  deﬁned here ( https://phytopathdb.org/pathogens_eg/). Digital \nexpression per condition to this assembl y was estimated using Salmon  (Patro et al, 2017), \nand differential expression analysis was performed as described above.  \n \nONT DRS Library Preparation \nTotal RNA was isolated , as detailed above. Poly(A)+ RNA was puriﬁed from approximately \n100µg of total RNA using the Dynabeads mRNA puriﬁcation kit (Thermo Fisher Scientiﬁc) \nfollowing the manufacturer’s instructions. The quality and quantity of mRNA were assessed \nusing the Nanodrop one spectrophotometer (Thermo Fisher Scientiﬁc) and Tape station 4150 \n(Agilent Technologies). ONT DRS libraries were prepared from 100ng poly(A)+ RNA for the \nWT Col-0 - vir-1 comparison at 17°C and 27°C. All other ONT DRS libraries were prepared \nfrom 500ng poly(A)+ RNA. Libraries were made using the Direct RNA sequencing kit (SQK -\nRNA002; Oxford Nanopore Technologies) according to the manufacturer’s instructions. The \npoly(T) adapter was ligated to the mRNA using T4 DNA ligase (New England Biolabs) in the \nQuick Ligase reaction buffer (New Eng land Biolabs) for 15 min at room temperature. First -\nstrand cDNA was synthesised by SuperScript III Reverse Transcriptase (Thermo Fisher \nScientiﬁc) using the oligo(dT)  adapter. The RNA –cDNA hybrid was then puriﬁed using \nAgencourt RNAClean XP magnetic beads (Beckman Coulter). The sequencing adapter was \nligated to the mRNA using T4 DNA ligase (New England Biolabs) in the Quick Ligase reaction \nbuffer (New England Biolabs) for 15 min at room temperature followed by a second puriﬁcation \nstep using Agencourt beads (as described above). Libraries were loaded onto R9 version \nSpotON Flow Cells (Oxford Nanopore Technologies) and sequenced using a GridION device \nat the Tayside Centre for Genomic Analysis, School of Medicine, University of Dundee, for a \n48-hour runtime. Four biological replicates were performed for each genotype.  \n \nONT DRS mapping \nReads were basecalled using Dorado version 0.5.3 (Oxford Nanopore Technologies) using \nthe rna002_70bps_hac@v3 high accuracy model. Reads were aligned to the Araport11 \ntranscriptome (Cheng et al, 2017) and the TAIR10 Arabidopsis genome (Lamesch et al, 2012) \nusing minimap2 version 2.17 (Li, 2018)  conﬁgured for spliced alignment . The following \nparameters were used for both alignments: --end-seed-pen=15 for end seed penalties, -A1, -\nB1, -O2,32, -E1,0 and -C9 to tune alignment scoring. For genomic alignment, splice junction \ninformation was incorporated using the --junc-bed parameter, which referenced the annotated \nintrons BED ﬁle. A junction bonus of 10 (--junc-bonus=10) was applied to prioritise alignments \nthat utili sed known splice junctions, increasing alignment accuracy for spliced reads . The \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted February 23, 2025. ; https://doi.org/10.1101/2025.02.18.638636doi: bioRxiv preprint \n\n 23 \nspliced alignment was optimi sed with parameters -k14, -uf, -w5, --splice, and -g2000, along \nwith a maximum intron size of 200,000 ( -G200000), --splice-ﬂank=yes for spliced alignment \nﬂanking detection, and -z200 for seeding thresholds. Alignments were converted to BAM ﬁles \nand indexed using samtools version 1.18. \n \nPrediction of m6A in ONT DRS data using m6Anet \nEvent information was extracted from raw signal data and transcriptome alignments using the \nf5c implementation of eventalign with event scaling (Gamaarachchi et al, 2020). Aligned event \nﬁles were processed using m6anet (Hendra et al, 2022). Data preparation and inference was \nperformed using the pretrained Arabidopsis model with the default read probability threshold \nof <0.0033. Predicted sites of modiﬁcation were ﬁltered using the recommended probability -\nmodiﬁed threshold of >0.9  (Hendra et al , 2022). The distribution of predicted modiﬁcation \nratios for all sites passing this threshold was plotted for each condition. \n \nAnalysis of poly(A) site usage in ONT DRS data \nDifferential 3’ analysis was performed on bam ﬁles using the d3pendr tool  as described \npreviously (Parker et al, 2021b). The statistical signiﬁcance of the 3’ shift was assessed by \npermuting read alignments between the control and treatment distributions to determine the \nmaximum distance achieved by random sampling.  \n \nEstimation of poly(A) tail length in ONT DRS data \nThe length of poly(A) tails per read was estimated using “--no-trim --estimate-poly-a” with the \nfollowing model: rna002_70bps_hac@v3 in Dorado (version 0.5.3). Reads were mapped to \nthe Araport 11 transcriptome using minimap2 (see above), and the resulting bam ﬁle was used \nto generate a read-to transcript table for further interrogation. Differences in mean poly(A) tail \nlength per gene between conditions were calculated as previously described (Parker et al, \n2020). In brief , poly(A) tail lengths were aggregated by gene ID , and where genes were \npresent with at least ten reads in both conditions, the distributions of poly(A) tail lengths were \ncompared using the Wasserstein distance. Signiﬁcance was assessed using a permutation \ntest with 999 bootstraps.  Genes were classed as m6A-modiﬁed if they had a t least one site \nabove the probability-modiﬁed threshold of >0.9 in at least one Col-0 sample.  \n \nGene tracks \nGene track ﬁgures were generated using Matplotlib (versio n 3.9.2) from normalised bigwig \nﬁles of Illumina RNA-Seq coverage and pooled bam ﬁles of reads per condition. For tracks \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted February 23, 2025. ; https://doi.org/10.1101/2025.02.18.638636doi: bioRxiv preprint \n\n 24 \nwith >100 ONT DRS reads per condition, a random subsample of 100 reads per track w as \nplotted. \n \nm6A liquid chromatography with tandem mass spectrometry \nm6A analysis using tandem liquid chromatography-mass spectroscopy (LC /MS-MS) was \nperformed as previously described (Parker et al, 2020, 2021b). LC/MS-MS was carried out by \nthe FingerPrints Proteomics facility at the University of Dundee. A two-way ANOVA test was \nused to assess the effects of genotype and temperature  and their interaction on the ratio of \nm6A to A.  \n \nCode and data availability  \nIllumina FASTQ and ONT FAST5 ﬁles for the 17ºC and 27ºC dataset are deposited in the ENA \nunder accession code PRJEB85795. ONT FAST5 for the te234 and cpsf30-yth datasets are \ndeposited in the ENA under accession codes PRJEB85859 and PRJEB85860 respectively. \nSource data for LC-MS, ﬂood inoculation and image staining are available in supporting ﬁles. \nSource code, R notebooks and Snakemake (Mölder et al, 2021) pipelines are available at \ngithub.com/bartongroup/m6a_arabidopsis_autoimmunity.  \n \nACKNOWLEDGEMENTS \nWe thank Prof. Steven Spoel (University of Edinburgh) for Pseudomonas syringae pv. Tomato \n(Pto) DC3000. We thank Dr. Martin Balcerowicz (University of Dundee) for providing access \nto temperature-controlled environment cabinets and Dr. Rachel Taylor (University of Leeds) \nfor temperature-controlled plant growth.  We are grateful to  Katie Dempsey, whose \nexperiments led us to investigate the temperature sensitivity of vir-1 mutants. We thank Dr. \nMartin Balcerowicz, Prof. Brendan Davies and Dr. Davide Bulgarelli for helpful comments on \nthe manuscript. We thank the University of Dundee HPC and Research Computing at the \nJames Hutton Institute for providing computational resources and technical support through \nthe BBSRC-funded “UK’s Crop Diversity Bioinformatics HPC” (BB/S019669/1 and \nBB/X019683/1). This work was supported by awards from the BBSRC (BB/V010662/1  and \nBB/M010060/1 to GGS and GJB; BB/W007673/1 to GGS ). The FingerPrints Proteomics \nfacility at the University of Dundee is supported by a Wellcome Trust Technology Platform \nAward (097945/B/11/Z). \n \nCompeting Interests \nThe authors declare no competing interests.  \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted February 23, 2025. ; https://doi.org/10.1101/2025.02.18.638636doi: bioRxiv preprint \n\n 25 \nFIGURES \n \nFigure 1: Immune gene expression is activated in m6A writer complex mutants.  \nA) Top 10 gene ontology terms enriched in the set of 597 genes signiﬁcantly upregulated (FDR \n< 0.001, log2FC > 2.0) in vir-1 mutants compared to Col-0 WT at 20°C. B) Heatmap showing \nthe TMM-FPKM normalised log2 centred fold expression between vir-1, Col-0 and VIRc for all \ngenes with TAIR annotations including the term  ‘defense/defence’. PR1 (AT2G14610) is \nhighlighted with an arrow. C) Normalised log2 counts per million of PR1 (AT2G14610) in Col-\n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted February 23, 2025. ; https://doi.org/10.1101/2025.02.18.638636doi: bioRxiv preprint \n\n 26 \n0 (n = 7), vir-1 (n = 6) and VIRc (n = 6) in Illumina RNA-seq. Boxes represent the interquartile \nrange of the logged values.  D) Normalised log2 counts per million of PR1 (AT2G14610) in \nCol-0, Col-0, vir-1 and VIRc in ONT DRS reads (n= 4 samples per genotype). E) Upregulation \nof PR1 (AT2G14610) in vir-1 at 20°C, shown by a gene track of Illumina RNA -seq and \ndownsampled ONT DRS reads. C) Normalised log2 counts per million of PR1 (AT2G14610) \nin Col-0, ﬁp37-4 and ﬁo-1 at 20°C in ONT DRS, showing the upregulation of PR1 in ﬁp37-4   \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted February 23, 2025. ; https://doi.org/10.1101/2025.02.18.638636doi: bioRxiv preprint \n\n 27 \n \n \nFigure 2: Loss of m6A triggers a temperature-sensitive autoimmune response.  \nA) Upset plot showing the overlap in signiﬁcantly upregulated and downregulated genes  \n(log2FC +/- 2.0 FDR < 0.001) in vir-1 at 17°C compared to Col-0 at 17°C and vir-1 at 27ºC \ncompared to Col-0 at 27ºC. B) Principal component analysis showing the clustering of samples \nby experimental condition (including genotype) and temperature . C) Correlation matrix and \nhierarchical clustering of expression proﬁles for each condition. The clustering shows that \ngene expression patterns are distinct for all conditions. In addition, biological replicates within \nconditions cluster together. However, the gene expression patterns detected in vir-1 separate \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted February 23, 2025. ; https://doi.org/10.1101/2025.02.18.638636doi: bioRxiv preprint \n\n 28 \nas the most different of all  possible comparisons . D) Signiﬁcant upregulation of PR1 \n(AT2G14610) in vir-1 at 17°C, shown by a boxplot of normalised expression (log 2 counts per \nmillion) in Illumina RNA-seq data (n= 4 samples per genotype).  E) Most enriched GO terms \namong genes signiﬁcantly upregulated (FDR < 0.001, log2FC<2.0) in vir-1 at 17°C compared \nto the average of: Col-0 17°C, 27°C and vir-1 at 27°C. Source data available in Supplementary \nTable 6. F) Heatmap showing the TMM -FPMK normalised log2 centred fold change for all \ndifferentially regulated genes in vir-1 at 17 °C (log2FC +/ - 2.0 FDR < 0.001) with TAIR \nannotations including ‘defense/defence’ for conditions vir-1 at 17°C, 27°C, Col-0 at 17°C and \n27°C. PR1 (AT2G14610) is highlighted with an arrow. \n \n  \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted February 23, 2025. ; https://doi.org/10.1101/2025.02.18.638636doi: bioRxiv preprint \n\n 29 \n \nFigure 3: Defence genes that function in diverse aspects of immunity are upregulated \nin vir-1 mutants \nA) Signiﬁcant upregulation of SARD1 (AT1G73805) in vir-1 at 17°C, shown by a boxplot of \nnormalised expression (log2 counts per million) in Illumina RNA-seq data (n = 4 samples per \ncondition). B) Signiﬁcant upregulation of FLS2 (AT5G46330) in vir-1 at 17°C, shown by a \nboxplot of normalised expression (log 2 counts per million) in Illumina RNA -seq data (n = 4 \nsamples per condition).  C) Signiﬁcant u pregulation of TX0 (AT1G57630) in vir-1 at 17°C, \nshown by a boxplot of normalised expression (log 2 counts per million) in Illumina RNA -seq \ndata (n = 4 samples per condition). D) Signiﬁcant upregulation of ACD6 (AT4G14400) in vir-1 \nat 17°C, shown by a boxplot of normalised expression (log 2 counts per million) in Illumina \nRNA-seq data (n = 4 samples per condition).  E) Upset plot showing modest overlap in \ndifferentially upregulated genes between vir-1 at 17°C and previously published acd6-1 RNA-\nseq data (Fabian et al, 2023).  \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted February 23, 2025. ; https://doi.org/10.1101/2025.02.18.638636doi: bioRxiv preprint \n\n 30 \n \nFigure 4: vir-1 mutants exhibit temperature-sensitive pathogen resistance and localised \ncell death.  \nA) Four-week-old Col-0 WT , vir-1 and ﬂs2c seedlings ﬂood-inoculated with a bacterial \nsuspension of Pst DC3000 (5x106 CFU/ml) and 0.025% v/v Silwet L-77. Bacterial populations \nwere quantiﬁed at 3 days post inoculation (dpi) (n = 3 per condition). One way ANOVA tests \non each genotype revealed a signiﬁcant effect of temperature in the vir-1 genotype (F = 37.23, \np = 0.0 358) which was not present in Col-0 WT.  Source data available in Supplementary \nTable 9. This experimental analysis was replicated independently in Supplementary Figure 4.  \nB) Trypan blue staining of Col-0 WT and vir-1 mutant leaves imaged with a Zeiss histology \nmicroscope at 10x magniﬁcation. C) Estimation of trypan blue staining patterns using ImageJ \n(n = 10 per condition). Two-way ANOVA revealed signiﬁcant effects of temperature (F = 22.34, \np = 3.45 × 10⁻⁵), genotype (F = 22.18, p = 3.63 × 10⁻⁵), and their interaction (F = 22.16, p = \n3.66 × 10 ⁻⁵). Post hoc comparisons using Tukey’s HSD test indicated that vir-1 at 17°C \nsigniﬁcantly differed from all other conditions ( p < 0.0001) . Source data available in \nSupplementary Table 10. \n  \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted February 23, 2025. ; https://doi.org/10.1101/2025.02.18.638636doi: bioRxiv preprint \n\n 31 \n \n \nFigure 5: m6A levels in vir-1 mutants are not temperature-sensitive. \nA) LC-MS/MS analysis showing the signiﬁcant effect of genotype on the m6A/A ratio and the \nlack of signiﬁcant interaction between genotype and temperature on m 6A levels  (two-way \nANOVA; p<0.001) (n = 3 per condition). Source data available in Supplementary Table 11. B) \nDensity distribution of the  ratio of m6A modiﬁcation per site for all sites with probability \nmodiﬁcation > 0.9 predicted by m6Anet. Individual replicates are plotted as solid lines, with \nthe combined density of a condition (genotype and temperature) plotted as a dashed line.  \n3545 sites were predicted to have an m6A modiﬁed site in at least one Col-0 sample, compared \nto only 327 in vir-1.  \n  \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted February 23, 2025. ; https://doi.org/10.1101/2025.02.18.638636doi: bioRxiv preprint \n\n 32 \n \n \nFigure 6: The primary visible developmental defects of vir-1 mutants are not rescued \nby growth at 27°C \nA) Developmental phenotype of Arabidopsis WT Col-0 and mutant vir-1 grown at 17°C and \n27°C. Plants are 28 days old and were grown at the indicated temperatures throughout their \ndevelopment following a 2-day stratiﬁcation treatment. \n \n  \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted February 23, 2025. ; https://doi.org/10.1101/2025.02.18.638636doi: bioRxiv preprint \n\n 33 \n \nFigure 7: Poly(A) tail length distributions of vir-1 are disrupted in a temperature \nsensitive manner \nA) Shifts towards upstream (promoter -proximal) poly(A) site usage are detected in vir-1 \nmutants at 17°C and 27°C compared to Col -0 WT. Source data available in Supplementary \nTable 14. B) Normalised expression (log 2 counts per million) of PR1 (AT2G14610) in ONT \nDRS analysis of cpsf30-yth mutants at 20°C (n = 4 per condition). C) Normalised expression \n(log 2 counts per million) of PR1 (AT2G14610) in ONT DRS analysis of te234 triple mutants \nat 20°C ( n = 4 per condition). D) Density distribution of poly(A) tail lengths in Col-0 and vir-1 \nat 17°C and 27°C. The distribution of each replicate is plotted individually. E) Histograms \ndepicting the distribution of Wasserstein distance metric for signiﬁcant changes in mean \npoly(A) tail length per gene between Col-0 and vir-1 at 17°C and 27°C, and between Col-0 at \n27°C and 17°C. At 17°C, 13 genes have signiﬁcantly shorter poly(A) tails in vir-1, while 7,894 \ngenes have signiﬁcantly longer tails. At 27°C, 6,669 genes displayed signiﬁcantly shorter \npoly(A) tails in vir-1, whereas 7 genes had signiﬁcantly longer tails. In Col-0 there are 1,425 \ngenes with signiﬁcantly shorter mean poly(A) tails at 17ºC compared to 27ºC. These ﬁndings \nare derived from data pooled across all replicates.  \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted February 23, 2025. ; https://doi.org/10.1101/2025.02.18.638636doi: bioRxiv preprint \n\n 34 \n \nSupplementary Figure 2 – Linked to Figure 2 \nA) Normalised logged counts per million of PR1 (AT2G14610) in Col -0 and vir-1 17°C and \n27ºC (n = 3-4 per condition). B) RT-qPCR showing the upregulation of PR1 in vir-1 at 17°C (n \n= 3 per condition). C) Volcano plot showing the log2 fold change and adjusted p -value of \ndifferential gene expression in vir-1 at 17°C contrasted to the average expression in vir-1 at \n27°C and Col-0 at 17°C and 27°C. Genes with log2FC > 2 and p < 0.001 are coloured in red, \ngenes which only pass the p-value threshold are coloured in black, and genes which only pass \nthe log2FC threshold are coloured in blue. Non-signiﬁcant changes (NS) are coloured in grey. \nSource data available in Supplementary Table 5. D) Overlap in enriched GO terms between \ngenes upregulated at 17°C contrasted to the average expression in vir-1 at 27°C and Col-0 at \n17°C and 27°C, and genes which were signiﬁcantly upregulated in vir-1 at 22ºC contrasted to \nCol-0 at 22ºC. \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted February 23, 2025. ; https://doi.org/10.1101/2025.02.18.638636doi: bioRxiv preprint \n\n 35 \n \nSupplementary Figure 3: Linked to Figure 3 \nA) Upregulation of SARD1 (AT1G73805) in vir-1 at 17°C, shown by a boxplot of normalised \nexpression (log 2 counts per million) in ONT DRS data (n = 3 -4 samples per condition). B) \nUpregulation of FLS2 (AT5G46330) in vir-1 at 17°C, shown by a boxplot of normalised \nexpression (log 2 counts per million) in ONT DRS data (n = 3 -4 samples per condition). C) \nUpregulation of TX10 (AT1G57630) in vir-1 at 17°C, shown by a boxplot of normalised \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted February 23, 2025. ; https://doi.org/10.1101/2025.02.18.638636doi: bioRxiv preprint \n\n 36 \nexpression (log 2 counts per million) in ONT DRS data (n = 3 -4 samples per condition). D)  \nGene track of ONT DRS data showing the upregulation of a novel TIR domain-containing gene \n(annotated as Novel gene) downstream of RPS6 (AT5G46470) in vir-1 at 17 °C (n = 3 -4 \nsamples per condition). E) Upregulation of ACD6 (AT4G14400) in vir-1 at 17°C, shown by a \nboxplot of normalised expression (log 2 counts per million) in ONT DRS data (n = 4 samples \nper condition). F-M) Boxplots showing the normalised log 2 counts per million (as produced \nby edgeR) for the ﬂowering genes; FT (AT1G65480), FUL (AT5G60910), SOC1 \n(AT2G45660), SEP3 (AT1G24260), SPL4 (AT1G53160), SPL5 (AT3G15270), AGL19 \n(AT4G22950) and AGL24 (AT4G24540), in Illumina RNA-seq of vir-1 and Col-0 at 17ºC and \n27ºC (n = 4 samples per condition). \n \n  \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted February 23, 2025. ; https://doi.org/10.1101/2025.02.18.638636doi: bioRxiv preprint \n\n 37 \n \nSupplementary Figure 4: Linked to Figure 4 \nFour-week-old Col-0 WT , vir-1 and ﬂs2c seedlings ﬂood-inoculated with a bacterial \nsuspension of Pst DC3000 (5x106 CFU/ml) and 0.025% v/v Silwet L-77. Bacterial populations \nwere quantiﬁed at 3 days post inoculation (dpi) (n = 3 per condition). One way ANOVA tests \non each genotype revealed a signiﬁcant effect of temperature in the vir-1 genotype (F = 23.02, \np = 0.00197) which was not present in Col -0 WT or ﬂs2. Source data available in \nSupplementary Table 9. This experimental analysis represents an independent replication of \nthe experiment presented in Figure 4A. \n \n  \n17°C 20°C 27°C\nCol−0 fls2c vir−1 Col−0 fls2c vir−1 Col−0 fls2c vir−1\n1e+06\n1e+08\n1e+10\nGenotype\nCFU/mg\ngenotype\nCol−0\nfls2c\nvir−1\n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted February 23, 2025. ; https://doi.org/10.1101/2025.02.18.638636doi: bioRxiv preprint \n\n 38 \n \nSupplementary Figure 5: Linked to Figure 5 \nA) Gene track showing the expression of VIRILIZER in Illumina RNA-seq VIR expression in \nvir-1 mutants at 17°C and 27°C. B) Magniﬁed portion of the VIRILIZER gene track showing \nthe combined coverage and alignment of Illumina RNA-seq around the EMS point mutation in \nvir-1. The vir-1 mutation affects the 5’ splice site of intron 5 (G+1 to A), which leads to the \nactivation of cryptic 5’ splice sites upstream in exon 5 detected only in vir-1 (denoted by an \narrow). No suppression of this cryptic splicing is found at 27 °C. Aligned reads were \nsubsampled to 200 reads  per condition. C) Density distribution of the ratio of modiﬁcation  \npredicted by Yanocomp, for modiﬁcations with an FDR < 0.05. Predicted modiﬁcation ratios \nfor vir-1 and Col-0 at 17°C were obtained by comparisons of vir-1 at 17°C and Col-0 at 17°C. \nPredicted modiﬁcation ratios for vir-1 and Col-0 at 27°C were obtained by comparing vir-1 at \n27°C and Col-0 at 27°C.  \n  \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted February 23, 2025. ; https://doi.org/10.1101/2025.02.18.638636doi: bioRxiv preprint \n\n 39 \n \n \nSupplementary Figure 6: Linked to Figure 6 \nA) Phenotype of Col-0, vir-1, ﬁp37-4, cpsf30-yth, te234 and acd6-1 grown at 17ºC and 27ºC. \nB) Gene ontology biological process terms enriched in the 58 genes consistently signiﬁcantly \nupregulated (log2FC +/- 2.0 FDR < 0.001) in vir-1 across 17°C, 20°C and 27°C compared to \nCol-0 and VIRc. Source data available in Supplementary Table 13. C) Upregulation of FMO1 \n(AT1G19250) in vir-1 at both 17°C and 27°C in Illumina RNA-seq and ONT DRS data, shown \nby gene tracks and boxplots of normalised expression (log 2 counts per million). \n \n \n  \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted February 23, 2025. ; https://doi.org/10.1101/2025.02.18.638636doi: bioRxiv preprint \n\n 40 \n \nSupplementary Figure 7: Linked to Figure 7 \nA) Density distribution of poly(A) tail lengths of Saccharomyces cerevisiae ENOLASE II \nspike-in sequences in Col-0 and vir-1 at 17°C and 27°C. ENOLASE II transcripts with a \npoly(A) tail length of 30 nt are included as the RNA calibration standard during ONT DRS \nlibrary preparation. B) Density distribution of poly(A) tails in Col-0 and vir-1 at 17°C and \n27°C, divided into those belonging to genes with a predicted m6A modiﬁcation in Col-0 and \nthose with no predicted modiﬁcation. The distribution of poly(A) tails is plotted individually for \neach replicate. \n \n \n \n  \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted February 23, 2025. ; https://doi.org/10.1101/2025.02.18.638636doi: bioRxiv preprint \n\n 41 \n \nREFERENCES \n \nAlcázar R & Parker JE (2011) The impact of temperature on balancing immune \nresponsiveness and growth in Arabidopsis. Trends Plant Sci 16: 666–675 \nAndrews S (20 17). FastQC: A quality control tool for high throughput sequence data. v3. \nBabrahamBioinformatics. https://www.bioinformatics.babraham.ac.uk/projects/fastqc/. \nArribas-Hernández L, Bressendorff S, Hansen MH, Poulsen C, Erdmann S & Brodersen P \n(2018) An m6A -YTH module controls developmental timing and morphogenesis in \nArabidopsis. Plant Cell 30: 952–967 \nArribas-Hernández L, Simonini S, Hansen MH, Paredes EB, Bressendorff S, Dong Y, \nØstergaard L & Brodersen P (2020) Recurrent requirement for the m6A -\nECT2/ECT3/ECT4 axis in the control of cell proliferation during plant organogenesis. \nDevelopment 147: dev189134 \nBaer BW & Kornberg RD (1983) The protein responsible for the repeating structure of \ncytoplasmic poly(A)-ribonucleoprotein. J Cell Biol 96: 717–721 \nBarragan AC & Weigel D (2021) Plant NLR diversity: the known unknowns of pan-NLRomes. \nPlant Cell 33: 814–831 \nBartok E & Hartmann G (2020) Immune sensing mechanisms that discriminate self from \naltered self and foreign nucleic acids. Immunity 53: 54–77 \nBartsch M, Gobbato E, Bednarek P, Debey S, Schultze JL, Bautor J & Parker JE (2006) \nSalicylic acid -independent ENHANCED DISEASE SUSCEPTIBILITY1 signaling in \nArabidopsis immunity and cell death is regulated by the monooxygenase FMO1 and \nthe Nudix hydrolase NUDT7. Plant Cell 18: 1038–1051 \nBékés M, Langley DR & Crews CM (2022) PROTAC targeted protein degraders: the past is \nprologue. Nat Rev Drug Discov 21: 181–200 \nBetsuyaku S, Katou S, Takebayashi Y, Sakakibara H, Nomura N & Fukuda H (2018) Salicylic \nacid and jasmonic acid pathways are activated in spatially different domains around \nthe infection site during effector-triggered immunity in Arabidopsis thaliana. Plant Cell \nPhysiol 59: 8–16 \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted February 23, 2025. ; https://doi.org/10.1101/2025.02.18.638636doi: bioRxiv preprint \n\n 42 \nBodi Z, Zhong S, Mehra S, Song J, Graham N, Li H, May S & Fray RG (2012) Adenosine \nmethylation in Arabidopsis mRNA is associated with the 3’ end and reduced levels \ncause developmental defects. Front Plant Sci 3: 48 \nBomblies K, Lempe J, Epple P, Warthmann N, Lanz C, Dangl JL & Weigel D (2007) \nAutoimmune response as a mechanism for a Dobzhansky -Muller-type incompatibility \nsyndrome in plants. PLoS Biol 5: e236 \nBomblies K & Weigel D (2007) Hybrid necrosis: autoimmunity as a potential gene-ﬂow barrier \nin plant species. Nat Rev Genet 8: 382–393 \nBonardi V, Cherkis K, Nishimura MT & Dangl JL (2012) A new eye on NLR proteins: focused \non clarity or diffused by complexity? Curr Opin Immunol 24: 41–50 \nBrodersen P & Arribas -Hernández L (2024) The m6A -YTH regulatory system in plants: A \nstatus. Curr Opin Plant Biol 82: 102650 \nBuchﬁnk B, Reuter K & Drost H -G (2021) Sensitive protein alignments at tree -of-life scale \nusing DIAMOND. Nat Methods 18: 366–368 \nBushnell B (2022) BBMap. Retrieved from https://sourceforge.net/projects/bbmap \nCamacho C, Coulouris G, Avagyan V, Ma N, Papadopoulos J, Bealer K & Madden TL (2009) \nBLAST+: architecture and applications. BMC Bioinformatics 10: 421 \nCarreras-Puigvert J, Zitnik M, Jemth A -S, Carter M, Unterlass JE, Hallström B, Loseva O, \nKarem Z, Calderón-Montaño JM, Lindskog C, et al (2017) A comprehensive structural, \nbiochemical and biological proﬁling of the human NUDIX hydrolase family. Nat \nCommun 8: 1541 \nChan SL, Huppertz I, Yao C, Weng L, Moresco JJ, Yates JR 3rd, Ule J, Manley JL & Shi Y \n(2014) CPSF30 and Wdr33 directly bind to AAUAAA in mammalian mRNA 3’ \nprocessing. Genes Dev 28: 2370–2380 \nChen J, Li L, Kim JH, Neuhäuser B, Wang M, Thelen M, Hilleary R, Chi Y, Wei L, \nVenkataramani K, et al  (2023) Small proteins modulate ion -channel-like ACD6 to \nregulate immunity in Arabidopsis thaliana. Mol Cell 83: 4386-4397.e9 \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted February 23, 2025. ; https://doi.org/10.1101/2025.02.18.638636doi: bioRxiv preprint \n\n 43 \nChen Y, Lun ATL & Smyth GK (2016) From reads to genes to pathways: differential expression \nanalysis of RNA -Seq experiments using Rsubread and the edgeR quasi -likelihood \npipeline. F1000Res 5: 1438 \nCheng C-Y, Krishnakumar V, Chan AP, Thibaud -Nissen F, Schobel S & Town CD (2017) \nAraport11: a complete reannotation of the Arabidopsis thaliana reference genome. \nPlant J 89: 789–804 \nDangl JL & Jones JD (2001) Plant pathogens and integrated defence responses to infection. \nNature 411: 826–833 \nDobin A, Davis CA, Schlesinger F, Drenkow J, Zaleski C, Jha S, Batut P, Chaisson M & \nGingeras TR (2013) STAR: ultrafast universal RNA-seq aligner. Bioinformatics 29: 15–\n21 \nDong OX, Ao K, Xu F, Johnson KCM, Wu Y, Li L, Xia S, Liu Y, Huang Y, Rodriguez E, et al \n(2018) Individual components of paired typical NLR immune receptors are regulated \nby distinct E3 ligases. Nat Plants 4: 699–710 \nEssuman K, Milbrandt J, Dangl JL & Nishimura MT (2022) Shared TIR enzymatic functions \nregulate cell death and immunity across the tree of life. Science 377: eabo0001 \nFabian M, Gao M, Zhang X-N, Shi J, Vrydagh L, Kim S-H, Patel P, Hu AR & Lu H (2023) The \nﬂowering time regulator FLK controls pathogen defense in Arabidopsis thaliana. Plant \nPhysiol 191: 2461–2474 \nFlores-Téllez D, Tankmar MD, von Bülow S, Chen J, Lindorff -Larsen K, Brodersen P & \nArribas-Hernández L (2023) Insights into the conservation and diversiﬁcation of the \nmolecular functions of YTHDF proteins. PLoS Genet 19: e1010980 \nFreh M, Gao J, Petersen M & Panstruga R (2022) Plant autoimmunity -fresh insights into an \nold phenomenon. Plant Physiol 188: 1419–1434 \nFreund I, Eigenbrod T, Helm M & Dalpke AH (2019) RNA modiﬁcations modulate activation of \ninnate Toll-like receptors. Genes 10: 92 \nFu L, Niu B, Zhu Z, Wu S & Li W (2012a) CD -HIT: accelerated for clustering the next -\ngeneration sequencing data. Bioinformatics 28: 3150–3152 \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted February 23, 2025. ; https://doi.org/10.1101/2025.02.18.638636doi: bioRxiv preprint \n\n 44 \nFu ZQ, Yan S, Saleh A, Wang W, Ruble J, Oka N, Mohan R, Spoel SH, Tada Y, Zheng N, et \nal (2012b) NPR3 and NPR4 are receptors for the immune signal salicylic acid in plants. \nNature 486: 228–232 \nGamaarachchi H, Lam CW, Jayatilaka G, Samarakoon H, Simpson JT, Smith MA & \nParameswaran S (2020) GPU accelerated adaptive banded event alignment for rapid \ncomparative nanopore signal analysis. BMC Bioinformatics 21: 343 \nGeula S, Moshitch -Moshkovitz S, Dominissini D, Mansour AA, Kol N, Salmon -Divon M, \nHershkovitz V, Peer E, Mor N, Manor YS, et al  (2015) m 6 A mRNA methylation \nfacilitates resolution of naïve pluripotency toward differentiation. Science 347: 1002–\n1006 \nGloggnitzer J, Akimcheva S, Srinivasan A, Kusenda B, Riehs N, Stampﬂ H, Bautor J, Dekrout \nB, Jonak C, Jiménez -Gómez JM, et al  (2014) Nonsense -mediated mRNA decay \nmodulates immune receptor levels to regulate plant antibacterial defense. Cell Host \nMicrobe 16: 376–390 \nGómez-Gómez L & Boller T (2000) FLS2: an LRR receptor -like kinase involved in the \nperception of the bacterial elicitor ﬂagellin in Arabidopsis. Mol Cell 5: 1003–1011 \nGrabherr MG, Haas BJ, Yassour M, Levin JZ, Thompson DA, Amit I, Adiconis X, Fan L, \nRaychowdhury R, Zeng Q, et al (2011) Full-length transcriptome assembly from RNA-\nSeq data without a reference genome. Nat Biotechnol 29: 644–652 \nHartmann M, Zeier T, Bernsdorff F, Reichel -Deland V, Kim D, Hohmann M, Scholten N, \nSchuck S, Bräutigam A, Hölzel T, et al (2018) Flavin monooxygenase-generated N-\nhydroxypipecolic acid is a critical element of plant systemic immunity. Cell 173: 456-\n469.e16 \nHendra C, Pratanwanich PN, Wan YK, Goh WSS, Thiery A & Göke J (2022) Detection of m6A \nfrom direct RNA sequencing using a multiple instance learning framework. Nat \nMethods 19: 1590–1598 \nJacob P, Hige J, Song L, Bayless A, Russ D, Bonardi V, El Kasmi F, Wünsch L, Yang Y, \nFitzpatrick CR, et al (2023) Broader functions of TIR domains in Arabidopsis immunity. \nProc Natl Acad Sci U S A 120: e2220921120 \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted February 23, 2025. ; https://doi.org/10.1101/2025.02.18.638636doi: bioRxiv preprint \n\n 45 \nJunt T & Barchet W (2015) Translating nucleic acid-sensing pathways into therapies. Nat Rev \nImmunol 15: 529–544 \nKarikó K (2021) Modiﬁed uridines are the key to a successful message. Nat Rev Immunol 21: \n619 \nKarikó K, Buckstein M, Ni H & Weissman D (2005) Suppression of RNA recognition by Toll -\nlike receptors: the impact of nucleoside modiﬁcation and the evolutionary origin of \nRNA. Immunity 23: 165–175 \nKierzek E & Kierzek R (2003) The thermodynamic stability of RNA duplexes and hairpins \ncontaining N6 -alkyladenosines and 2 -methylthio-N6-alkyladenosines. Nucleic Acids \nRes 31: 4472–4480 \nKim JH, Castroverde CDM, Huang S, Li C, Hilleary R, Seroka A, Sohrabi R, Medina -Yerena \nD, Huot B, Wang J, et al (2022) Increasing the resilience of plant immunity to a warming \nclimate. Nature 607: 339–344 \nKinoshita A & Richter R (2020) Genetic and molecular basis of ﬂoral induction in Arabidopsis \nthaliana. J Exp Bot 71: 2490–2504 \nKolberg L, Raudvere U, Kuzmin I, Adler P, Vilo J & Peterson H (2023) g:Proﬁler-interoperable \nweb service for functional enrichment analysis and gene identiﬁer mapping (2023 \nupdate). Nucleic Acids Res 51: W207–W212 \nKorn SM, Ulshöfer CJ, Schneider T & Schlundt A (2021) Structures and target RNA \npreferences of the RNA-binding protein family of IGF2BPs: An overview. Structure 29: \n787–803 \nKoski GK, Karikó K, Xu S, Weissman D, Cohen PA & Czerniecki BJ (2004) Cutting edge: \ninnate immune system discriminates between RNA containing bacterial versus \neukaryotic structural features that prime for high-level IL-12 secretion by dendritic cells. \nJ Immunol 172: 3989–3993 \nKrishnakumar V, Hanlon MR, Contrino S, Ferlanti ES, Karamycheva S, Kim M, Rosen BD, \nCheng C-Y, Moreira W, Mock SA, et al (2015) Araport: the Arabidopsis information \nportal. Nucleic Acids Res 43: D1003-9 \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted February 23, 2025. ; https://doi.org/10.1101/2025.02.18.638636doi: bioRxiv preprint \n\n 46 \nLamesch P, Berardini TZ, Li D, Swarbreck D, Wilks C, Sasidharan R, Muller R, Dreher K, \nAlexander DL, Garcia -Hernandez M, et al  (2012) The Arabidopsis Information \nResource (TAIR): improved gene annotation and new tools. Nucleic Acids Res  40: \nD1202-10 \nLaudenbach BT, Krey K, Emslander Q, Andersen LL, Reim A, Scaturro P, Mundigl S, Dächert \nC, Manske K, Moser M, et al (2021) NUDT2 initiates viral RNA degradation by removal \nof 5’-phosphates. Nat Commun 12: 6918 \nLee B, Park Y-S, Lee S, Song GC & Ryu C-M (2016) Bacterial RNAs activate innate immunity \nin Arabidopsis. New Phytol 209: 785–797 \nLi H (2018) Minimap2: pairwise alignment for nucleotide sequences. Bioinformatics 34: 3094–\n3100 \nLiu N, Dai Q, Zheng G, He C, Parisien M & Pan T (2015) N(6) -methyladenosine-dependent \nRNA structural switches regulate RNA-protein interactions. Nature 518: 560–564 \nLocci F & Parker JE (2024) Plant NLR immunity activation and execution: a biochemical \nperspective. Open Biol 14: 230387 \nLopez-Gomollon S & Baulcombe DC (2022) Roles of RNA silencing in viral and non-viral plant \nimmunity and in the crosstalk between disease resistance systems. Nat Rev Mol Cell \nBiol 23: 645–662 \nLu H, Rate DN, Song JT & Greenberg JT (2003) ACD6, a novel ankyrin protein, is a regulator \nand an effector of salicylic acid signaling in the Arabidopsis defense response. Plant \nCell 15: 2408–2420 \nMaharana S, Kretschmer S, Hunger S, Yan X, Kuster D, Traikov S, Zillinger T, Gentzel M, \nElangovan S, Dasgupta P, et al (2022) SAMHD1 controls innate immunity by regulating \ncondensation of immunogenic self RNA. Mol Cell 82: 3712-3728.e10 \nMichaels SD & Amasino RM (1998) A robust method for detecting single-nucleotide changes \nas polymorphic markers by PCR. Plant J 14: 381–385 \nMölder F, Jablonski KP, Letcher B, Hall MB, Tomkins -Tinch CH, Sochat V, Forster J, Lee S, \nTwardziok SO, Kanitz A, et al  (2021) Sustainable data analysis with Snakemake. \nF1000Res 10: 33 \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted February 23, 2025. ; https://doi.org/10.1101/2025.02.18.638636doi: bioRxiv preprint \n\n 47 \nMurakami S & Jaffrey SR (2022) Hidden codes in mRNA: Control of gene expression by m6A. \nMol Cell 82: 2236–2251 \nNgou BPM, Ahn H -K, Ding P & Jones JDG (2021) Mutual potentiation of plant immunity by \ncell-surface and intracellular receptors. Nature 592: 110–115 \nOstendorf T, Zillinger T, Andryka K, Schlee -Guimaraes TM, Schmitz S, Marx S, Bayrak K, \nLinke R, Salgert S, Wegner J, et al (2020) Immune sensing of synthetic, bacterial, and \nprotozoan RNA by toll -like receptor 8 requires coordinated processing by RNase T2 \nand RNase 2. Immunity 52: 591-605.e6 \nParker MT, Barton GJ & Simpson GG (2021a) Yanocomp: robust prediction of m6A \nmodiﬁcations in individual nanopore direct RNA reads. bioRxiv: 2021.06.15.448494 \nParker MT, Knop K, Sherwood AV, Schurch NJ, Mackinnon K, Gould PD, Hall AJ, Barton GJ \n& Simpson GG (2020) Nanopore direct RNA sequencing maps the complexity of \nArabidopsis mRNA processing and m6A modiﬁcation. Elife 9: e49658 \nParker MT, Knop K, Zacharaki V, Sherwood AV, Tomé D, Yu X, Martin PG, Beynon J, \nMichaels SD, Barton GJ, et al (2021b) Widespread premature transcription termination \nof Arabidopsis thaliana NLR genes by the spen protein FPA. Elife 10: e65537 \nParker MT, Soanes BK, Kusakina J, Larrieu A, Knop K, Joy N, Breidenbach F, Sherwood AV, \nBarton GJ, Fica SM, et al (2022) m6A modiﬁcation of U6 snRNA modulates usage of \ntwo major classes of pre-mRNA 5’ splice site. Elife 11: e78808 \nPassmore LA & Coller J (2022) Roles of mRNA poly(A) tails in regulation of eukaryotic gene \nexpression. Nat Rev Mol Cell Biol 23: 93–106 \nPatro R, Duggal G, Love MI, Irizarry RA & Kingsford C (2017) Salmon provides fast and bias-\naware quantiﬁcation of transcript expression. Nat Methods 14: 417–419 \nPrall W, Sheikh AH, Bazin J, Bigeard J, Almeida -Trapp M, Crespi M, Hirt H & Gregory BD \n(2023) Pathogen-induced m6A dynamics affect plant immunity. Plant Cell 35: 4155–\n4172 \nPruitt RN, Locci F, Wanke F, Zhang L, Saile SC, Joe A, Karelina D, Hua C, Fröhlich K, Wan \nW-L, et al  (2021) The EDS1 –PAD4–ADR1 node mediates Arabidopsis pattern -\ntriggered immunity. Nature 598: 495–499 \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted February 23, 2025. ; https://doi.org/10.1101/2025.02.18.638636doi: bioRxiv preprint \n\n 48 \nRate DN, Cuenca JV, Bowman GR, Guttman DS & Greenberg JT (1999) The gain-of-function \nArabidopsis acd6 mutant reveals novel regulation and function of the salicylic acid \nsignaling pathway in controlling cell death, defenses, and cell growth. Plant Cell 11: \n1695–1708 \nReiser L, Bakker E, Subramaniam S, Chen X, Sawant S, Khosa K, Prithvi T & Berardini TZ \n(2024) The Arabidopsis Information Resource in 2024. Genetics 227: 1–11 \nRemick BC, Gaidt MM & Vance RE (2023) Effector -triggered immunity. Annu Rev Immunol \n41: 453–481 \nRies RJ, Pickering BF, Poh HX, Namkoong S & Jaffrey SR (2023) m6A governs length -\ndependent enrichment of mRNAs in stress granules. Nat Struct Mol Biol  30: 1525–\n1535 \nRoost C, Lynch SR, Batista PJ, Qu K, Chang HY & Kool ET (2015) Structure and \nthermodynamics of N6-methyladenosine in RNA: a spring-loaded base modiﬁcation. J \nAm Chem Soc 137: 2107–2115 \nRůžička K, Zhang M, Campilho A, Bodi Z, Kashif M, Saleh M, Eeckhout D, El-Showk S, Li H, \nZhong S, et al (2017) Identiﬁcation of factors required for m6 A mRNA methylation in \nArabidopsis reveals a role for the conserved E3 ubiquitin ligase HAKAI. New Phytol \n215: 157–172 \nSablowski RW & Meyerowitz EM (1998) Temperature-sensitive splicing in the ﬂoral homeotic \nmutant apetala3-1. Plant Cell 10: 1453–1463 \nSasse A, Ray D, Laverty KU, Tam CL, Albu M, Zheng H, Lyudovyk O, Dalal T, Nie K, Magis \nC, et al  (2024) Reconstructing the sequence speciﬁcities of RNA -binding proteins \nacross eukaryotes. bioRxiv 2024.10.15.618476 \nSchlee M & Hartmann G (2016) Discriminating self from non-self in nucleic acid sensing. Nat \nRev Immunol 16: 566–580 \nSchneider CA, Rasband WS & Eliceiri KW (2012) NIH Image to ImageJ: 25 years of image \nanalysis. Nat Methods 9: 671–675 \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted February 23, 2025. ; https://doi.org/10.1101/2025.02.18.638636doi: bioRxiv preprint \n\n 49 \nSchönemann L, Kühn U, Martin G, Schäfer P, Gruber AR, Keller W, Zavolan M & Wahle E \n(2014) Reconstitution of CPSF active in polyadenylation: recognition of the \npolyadenylation signal by WDR33. Genes Dev 28: 2381–2393 \nShen L, Liang Z, Gu X, Chen Y, Teo ZWN, Hou X, Cai WM, Dedon PC, Liu L & Yu H (2016) \nN(6)-Methyladenosine RNA modiﬁcation regulates shoot stem cell fate in Arabidopsis. \nDev Cell 38: 186–200 \nSherman BT, Hao M, Qiu J, Jiao X, Baseler MW, Lane HC, Imamichi T & Chang W (2022) \nDAVID: a web server for functional enrichment analysis and functional annotation of \ngene lists (2021 update). Nucleic Acids Res 50: W216–W221 \nStevens AT, Howe DK & Hunt AG (2018) Characterization of mRNA polyadenylation in the \napicomplexa. PLoS One 13: e0203317 \nSun T, Zhang Y, Li Y, Zhang Q, Ding Y & Zhang Y (2015) ChIP -seq reveals broad roles of \nSARD1 and CBP60g in regulating plant immunity. Nat Commun 6: 10159 \nŚwiadek M, Proost S, Sieh D, Yu J, Todesco M, Jorzig C, Rodriguez Cubillos AE, Plötner B, \nNikoloski Z, Chae E, et al (2017) Novel allelic variants in ACD6 cause hybrid necrosis \nin local collection of Arabidopsis thaliana. New Phytol 213: 900–915 \nTakagi M, Iwamoto N, Kubo Y, Morimoto T, Takagi H, Takahashi F, Nishiuchi T, Tanaka K, \nTaji T, Kaminaka H, et al (2020) Arabidopsis SMN2/HEN2, encoding DEAD-box RNA \nhelicase, governs proper expression of the resistance gene SMN1/RPS6 and is \ninvolved in dwarf, autoimmune phenotypes of mekk1 and mpk4 mutants. Plant Cell \nPhysiol 61: 1507–1516 \nTang B, Feng L, Hulin MT, Ding P & Ma W (2023) Cell -type-speciﬁc responses to fungal \ninfection in plants revealed by single-cell transcriptomics. Cell Host Microbe 31: 1732-\n1747.e5 \nTian H, Wu Z, Chen S, Ao K, Huang W, Yaghmaiean H, Sun T, Xu F, Zhang Y, Wang S, et al \n(2021) Activation of TIR signalling boosts pattern-triggered immunity. Nature 598: 500–\n503 \nTodesco M, Balasubramanian S, Hu TT, Traw MB, Horton M, Epple P, Kuhns C, Sureshkumar \nS, Schwartz C, Lanz C, et al (2010) Natural allelic variation underlying a major ﬁtness \ntrade-off in Arabidopsis thaliana. Nature 465: 632–636 \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted February 23, 2025. ; https://doi.org/10.1101/2025.02.18.638636doi: bioRxiv preprint \n\n 50 \nTsai Y -C, Uechi H, Millar SR, Kadijk E, Schreiber KJ, Minkovich A, Huang JQ, Zhang Z, \nBarsacchi R, Möbius C, et al (2025) Dynamic proximal interactomics and chemical \ngenetic screening reveal CCR4-NOT sequestration in stress granules as a mechanism \nfor transcript stabilization. bioRxiv 2025.02.04.636058 \nTudek A, Krawczyk PS, Mroczek S, Tomecki R, Turtola M, Matylla -Kulińska K, Jensen TH & \nDziembowski A (2021) Global view on the metabolism of RNA poly(A) tails in yeast \nSaccharomyces cerevisiae. Nat Commun 12: 4951 \nVan der Biezen EA & Jones JD (1998) Plant disease -resistance proteins and the gene -for-\ngene concept. Trends Biochem Sci 23: 454–456 \nVi SL, Trost G, Lange P, Czesnick H, Rao N, Lieber D, Laux T, Gray WM, Manley JL, Groth \nD, et al  (2013) Target speciﬁcity among canonical nuclear poly(A) polymerases in \nplants modulates organ growth and pathogen response. Proc Natl Acad Sci U S A 110: \n13994–13999 \nWan W-L, Kim S -T, Castel B, Charoennit N & Chae E (2021) Genetics of autoimmunity in \nplants: an evolutionary genetics perspective. New Phytol 229: 1215–1233 \nWang L, Tsuda K, Truman W, Sato M, Nguyen LV, Katagiri F & Glazebrook J (2011) CBP60g \nand SARD1 play partially redundant critical roles in salicylic acid signaling: Role of \nCBP60 proteins in salicylic acid signaling. Plant J 67: 1029–1041 \nvan Wersch R, Li X & Zhang Y (2016) Mighty dwarfs: Arabidopsis autoimmune mutants and \ntheir usages in genetic dissection of plant immunity. Front Plant Sci 7: 1717 \nWong CE, Zhang S, Xu T, Zhang Y, Teo ZWN, Yan A, Shen L & Yu H (2023) Shaping the \nlandscape of N6-methyladenosine RNA methylation in Arabidopsis. Plant Physiol 191: \n2045–2063 \nWu B, Su S, Patil DP, Liu H, Gan J, Jaffrey SR & Ma J (2018) Molecular basis for the speciﬁc \nand multivariant recognitions of RNA substrates by human hnRNP A2/B1. Nat \nCommun 9: 420 \nYamagishi R, Inagaki H, Suzuki J, Hosoda N, Sugiyama H, Tomita K, Hotta T & Hoshino S -I \n(2024) Concerted action of ataxin -2 and PABPC1 -bound mRNA poly(A) tail in the \nformation of stress granules. Nucleic Acids Res 52: 9193–9209 \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted February 23, 2025. ; https://doi.org/10.1101/2025.02.18.638636doi: bioRxiv preprint \n\n 51 \nYoung MD, Wakeﬁeld MJ, Smyth GK & Oshlack A (2010) Gene ontology analysis for RNA -\nseq: accounting for selection bias. Genome Biol 11: R14 \nYu D, Song W, Tan EYJ, Liu L, Cao Y, Jirschitzka J, Li E, Logemann E, Xu C, Huang S, et al \n(2022) TIR domains of plant immune receptors are 2’,3’ -cAMP/cGMP synthetases \nmediating cell death. Cell 185: 2370-2386.e18 \nYuan M, Jiang Z, Bi G, Nomura K, Liu M, Wang Y, Cai B, Zhou J-M, He SY & Xin X-F (2021) \nPattern-recognition receptors are required for NLR -mediated plant immunity. Nature \n592: 105–109 \nZaccara S & Jaffrey S (2024) Understanding the redundant functions of the m6A -binding \nYTHDF proteins. RNA 30: 468–481 \nZaccara S & Jaffrey SR (2020) A uniﬁed model for the function of YTHDF proteins in regulating \nm6A-modiﬁed mRNA. Cell 181: 1582-1595.e18 \nZeier J (2021) Metabolic regulation of systemic acquired resistance. Curr Opin Plant Biol 62: \n102050 \nZhang M, Bodi Z, Mackinnon K, Zhong S, Archer N, Mongan N, Simpson G & Fray R (2022) \nTwo zinc ﬁnger proteins with functions in m6A writing interact with HAKAI. Nat \nCommun 13: 1127 \nZhang Y, Goritschnig S, Dong X & Li X (2003) A gain -of-function mutation in a plant disease \nresistance gene leads to constitutive activation of downstream signal transduction \npathways in suppressor of npr1-1, constitutive 1. Plant Cell 15: 2636–2646 \nZhou Y, Zhou B, Pache L, Chang M, Khodabakhshi AH, Tanaseichuk O, Benner C & Chanda \nSK (2019) Metascape provides a biologist -oriented resource for the analysis of \nsystems-level datasets. Nat Commun 10: 1523 \nZhu W, Zaidem M, Van de Weyer A-L, Gutaker RM, Chae E, Kim S-T, Bemm F, Li L, Todesco \nM, Schwab R, et al (2018) Modulation of ACD6 dependent hyperimmunity by natural \nalleles of an Arabidopsis thaliana NLR resistance gene. PLoS Genet 14: e1007628 \nZhu Y, Qian W & Hua J (2010) Temperature modulates plant defense responses through NB-\nLRR proteins. PLoS Pathog 6: e1000844 \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted February 23, 2025. ; https://doi.org/10.1101/2025.02.18.638636doi: bioRxiv preprint \n\n 52 \nZipfel C, Robatzek S, Navarro L, Oakeley EJ, Jones JDG, Felix G & Boller T (2004) Bacterial \ndisease resistance in Arabidopsis through ﬂagellin perception. Nature 428: 764–767 \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted February 23, 2025. ; https://doi.org/10.1101/2025.02.18.638636doi: bioRxiv preprint","source_license":"CC-BY-4.0","license_restricted":false}