Arabidopsis translation initiation factor binding protein CBE1 negatively regulates accumulation of the NADPH oxidase respiratory burst oxidase homolog D

Cell surface pattern recognition receptors sense invading pathogens by binding microbial or endogenous elicitors to activate plant immunity. These responses are under tight control to avoid excessive or untimely activation of cellular responses, which may otherwise be detrimental to host cells. How this fine-tuning is accomplished is an area of active study. We previously described a suppressor screen that identified Arabidopsis thaliana mutants with regained immune signaling in the immunodeficient genetic background bak1-5, which we named modifier of bak1-5 (mob) mutants. Here, we report that bak1-5 mob7 mutant restores elicitor-induced signaling. Using a combination of map-based cloning and whole-genome resequencing, we identified MOB7 as conserved binding of eIF4E1 (CBE1), a plant-specific protein that interacts with the highly conserved eukaryotic translation initiation factor eIF4E1. Our data demonstrate that CBE1 regulates the accumulation of respiratory burst oxidase homolog D, the NADPH oxidase responsible for elicitor-induced apoplastic reactive oxygen species production. Furthermore, several mRNA decapping and translation initiation factors colocalize with CBE1 and similarly regulate immune signaling. This study thus identifies a novel regulator of immune signaling and provides new insights into reactive oxygen species regulation, potentially through translational control, during plant stress responses.

Cell surface pattern recognition receptors sense invading pathogens by binding microbial or endogenous elicitors to activate plant immunity. These responses are under tight control to avoid excessive or untimely activation of cellular responses, which may otherwise be detrimental to host cells. How this fine-tuning is accomplished is an area of active study. We previously described a suppressor screen that identified Arabidopsis thaliana mutants with regained immune signaling in the immunodeficient genetic background bak1-5, which we named modifier of bak1-5 (mob) mutants. Here, we report that bak1-5 mob7 mutant restores elicitor-induced signaling. Using a combination of map-based cloning and whole-genome resequencing, we identified MOB7 as conserved binding of eIF4E1 (CBE1), a plant-specific protein that interacts with the highly conserved eukaryotic translation initiation factor eIF4E1. Our data demonstrate that CBE1 regulates the accumulation of respiratory burst oxidase homolog D, the NADPH oxidase responsible for elicitor-induced apoplastic reactive oxygen species production. Furthermore, several mRNA decapping and translation initiation factors colocalize with CBE1 and similarly regulate immune signaling. This study thus identifies a novel regulator of immune signaling and provides new insights into reactive oxygen species regulation, potentially through translational control, during plant stress responses.
The restriction of invading organisms is governed by passive and active defenses, which are effective against all types of plant pathogens and pests, including viruses, insects, nematodes, and parasitic plants (1). On the cell surface, conserved microbial molecules called pathogen-or microbe-associated molecular patterns or plant-derived damage-associated molecular patterns and phytocytokines (hereafter, generally referred to as elicitors) are recognized by pattern recognition receptors (PRRs) (2,3). For example, in Arabidopsis thaliana (hereafter Arabidopsis), the PRRs flagellin sensing 2 (FLS2), EF-TU receptor, and PEP1 receptor 1 and PEP1 receptor 2 recognize bacterial flagellin (and its cognate ligand, flg22), bacterial EF-Tu (and its cognate ligand, elf18), and endogenous Atpep1 and related peptides, respectively (4)(5)(6). These PRRs interact with the common coreceptor brassinosteroid insensitive 1-associated kinase 1 (BAK1) in a ligand-dependent manner (7)(8)(9). Following heterodimerization, numerous cell signaling events are initiated, including activation of receptorlike cytoplasmic kinases, production of apoplastic reactive oxygen species (ROS) catalyzed by the NADPH oxidase respiratory burst oxidase homolog D (RBOHD), altered ion fluxes, activation of calcium-dependent protein kinases, mitogen-activated protein kinase (MAPK) cascades, callose deposition, and large-scale transcriptional programming (10,11). To maintain immune homeostasis, plants use multiple strategies to adjust the amplitude and duration of immune responses (11). These include limiting the ability of PRRs to recruit their cognate coreceptors, regulation of signaling initiation and amplitude at the level of PRR complexes (i.e., post-translational modifications, protein turnover), monitoring of cytoplasmic signal-transducing pathways, and control of transcriptional reprogramming (11).
To identify loci involved in plant immunity, we previously conducted a forward genetic screen in the immunodeficient mutant bak1-5, called the modifier of bak1-5 (mob) screen (12). This ethyl methanesulfonate (EMS)-induced suppressor screen of bak1-5 phenotypes identified 10 mutants in nine allelic groups, named mob1 to mob10, with partially restored elicitorinduced ROS production (12)(13)(14). Through this suppressor screen, novel regulators of immune signaling have been discovered. MOB1 and MOB2 encode calcium-dependent protein kinase 28, which negatively regulates immune signaling by controlling the accumulation of the receptor-like cytoplasmic kinase botrytis-induced kinase 1, a central kinase involved in immune signaling downstream of multiple PRRs (12,15,16). MOB4 encodes constitutive active defense 1 (13). Constitutive active defense 1 is involved in immunity at different levels by controlling programmed cell death and regulating the homeostasis of the phyllosphere microbial community (17,18). MOB6 corresponds to site-1 protease, which controls the maturation of the endogenous rapid alkalinization factor 23 peptide to regulate immune signaling via the receptor kinase FERONIA (14,19,20). Hence, we predict that the identification of remaining MOB genes will continue to unravel mechanisms of immune regulation.
Here, we report that MOB7 corresponds to conserved binding of eIF4E1 (CBE1), a plant-specific protein that associates with the 5 0 mRNA cap (21) and the translation initiation factor eIF4E1 (22). We show that CBE1 colocalizes with ribonucleoprotein complexes and that cbe1 and other translational regulator mutants display enhanced accumulation of RBOHD protein, resulting in enhanced antibacterial immunity and ROS production, possibly through translational control of RBOHD protein levels.

Results
The mob7 mutation rescues bak1-5 immunodeficiency In the present study, we describe and characterize the mob7 mutation. First, we confirmed that the mob7 mutation was maintained in the M 5 generation, as bak1-5 mob7 suppressor mutants displayed partially restored ROS (H 2 O 2 ) production in seedlings upon treatment with the elicitors elf18 and flg22 (Fig. 1A). In addition, the mob7 mutation increased ROS production in adult leaves upon elicitation with elf18, Atpep1, and chitin; however, no difference was observed with flg22 (Figs. 1B and S1, A-D). Despite partially rescuing the ROS phenotype quantitatively, the mob7 mutation did not restore the delayed peak of ROS burst observed in bak1-5 (Fig. S1, B and E). However, the delayed response observed in bak1-5 is thought to be due to the compensation by other SERKs (23), which might not be as active as SERK3/BAK1 in immune signaling. This phenotype suggests a role of CBE1 downstream of the SERKs.
A late immune output triggered by several elicitors is the inhibition of seedling growth (10). While seedling growth inhibition is largely blocked in the bak1-5 mutant (9, 23), it was restored in suppressor mutant bak1-5 mob7 upon prolonged exposure with elf18, flg22, or Atpep1, while mock-treated seedlings grew similar to wildtype (WT) Col-0 (Figs. 1C and S1F). This sensitivity to flg22 of the bak1-5 mob7 mutant during seedling development, which was not observed in adult leaves to induce a partial regain of ROS production compared to bak1-5, is likely due to the different expression level of FLS2 at various developmental stage (24,25), and different growth conditions as some hormones regulate FLS2 expression and consequently flg22-triggered responses (26)(27)(28)(29). Furthermore, immunity to the hypovirulent bacterial strain Pseudomonas syringae pathovar tomato (Pto) DC3000 CORwas restored in bak1-5 mob7 suppressor mutants compared to bak1-5 (Fig. 1D). Altogether, these results show that mob7 partially restores immunity in bak1-5.

Identification of MOB7 as CBE1
Using the elicitor-induced ROS phenotype of mob7 and map-based cloning of the F 2 population from the outcross of bak1-5 mob7 (Col-0 ecotype) with Ler-0, linkage analysis revealed three regions of interest (Fig. S2). Whole-genome resequencing of bulked F 2:3 segregants that rescued seedling growth inhibition upon 1 μM Atpep1 treatment identified a single nucleotide polymorphism in AT4G01290, a gene that encodes CBE1 (Fig. 2A). The G to A transition is located at the last nucleotide of the third exon (Fig. 2B), which leads to a premature stop codon. This results in reduced CBE1 expression (Fig. S3, B and C). In addition, transient expression of eGFP-CBE1 mob7 in Nicotiana benthamiana revealed a truncated protein with an apparent molecular weight of 44 kDa, while GFP-CBE1 migrated at 137 kDa (Figs. 2C and S3A). The discrepancy of size observed and additional bands may be caused by yet unknown posttranslational modifications of CBE1 (Fig. 2C). It is possible that the premature stop codon in mob7 is recognized by the nonsense-mediated mRNA decay (NMD) machinery, which links premature translation termination to mRNA degradation (30).

CBE1 is a negative regulator of elicitor-induced ROS production and immunity
While mutation of CBE1 results in increased ROS production induced by various elicitors (Figs. 3A and S4A) and enhanced immunity to Pto DC3000 COR - (Fig. 3C), we did not observe any difference in seedling growth inhibition or MAPK activation between different cbe1 alleles and Col-0 (Figs. 3, D and E and S4B). Given the apparent specific impact of cbe1 mutations on ROS production, we tested whether transcripts and/or protein levels for the NADPH oxidase RBOHD were affected. Interestingly, while no significant reproducible difference could be observed at the transcript level ( Fig. 3F; ref. (22)), RBOHD protein accumulation was higher in cbe1 mutants, while unchanged in bak1-5 (Figs. 3G and S5A and S6A).To further investigate this phenotype, we analyzed RBOHD transcript and protein stability. RNA abundance of RBOHD was stable in Col-0 and cbe1 mutants after treatment with the transcription inhibitor cordycepin (Fig. S5B). These results suggest that CBE1 regulates RBOHD translationally or post-translationally, which could thus explain the effect on ROS production and immunity. Moreover, higher elicitor-induced ROS production in cbe1 mutants was phenocopied by overexpressing RBOHD in WT and bak1-5 (Fig. S6B).

CBE1 colocalizes with ribonucleoprotein complexes
CBE1 is known to interact with the translation initiation factors eIF4E and eIFiso4E, which localize to ribonucleoprotein complexes associated with the 5 0 cap of mRNA transcripts (22). We were therefore interested to investigate the subcellular localization of CBE1. When transiently expressed in N. benthamiana, CBE1-GFP displays a nucleocytoplasmic subcellular distribution, additionally localizing to distinct cytoplasmic foci (Fig. 4A). Comparatively, while CBE1 mob7 -GFP similarly localizes to the cytoplasm and nucleus, localization in cytoplasmic foci was not apparent (Fig. 4B). To investigate the localization of CBE1 within cytoplasmic foci, colocalization was measured using Pearson correlation coefficient with different ribonucleoprotein complex markers (31). Active translation is located within polysomes while processing bodies (P-bodies) and stress granules are generally associated with decay and storage of mRNA, respectively (32). To differentiate those different subcomplexes, we used relevant marker proteins. Associated with P-bodies, decapping 1 (DCP1) (33) is a member of the decapping complex, which is responsible for removal of the 5 0 cap, while up-frameshift suppressor 1 (34) is a factor of NMD. Although generally associated with active translation within polysomes, the translation initiation factor eIF4E (35) and poly(A) binding protein 2 (36) also localize to stress granules, together with the RNA-binding proteins (RBPs) oligouridylate-binding protein 1b (36) and RNA-binding protein 47C (35). We observed the highest colocalization correlation between CBE1 and DCP1 as well as partial colocalization between CBE1 and up-frameshift suppressor 1 (Figs. 4C and S7A). To a lesser extent, CBE1 also colocalized with polysome and stress granule markers eIF4E, oligouridylate-binding protein 1b, RNA-binding protein 47C, and poly(A) binding protein 2 (Figs. 4C and S7). The localization of CBE1 into these compartments in N. benthamiana was not influenced by flg22 treatment (Fig. S7B, Videos S1 and S2). This indicates that CBE1 constitutively colocalizes with ribonucleoprotein complexes and suggests a role for CBE1 in P-bodies.

AtCBE1 negatively regulates RBOHD
analyzed together with the single mutants pat1-1 and summ2-8. Similar to cbe1-1, eif4e1 and pat1 mutants, and to a lesser extent eif4g, showed a similar ROS phenotype upon elicitor treatment as observed in cbe1 (Fig. 5A). Accordingly, eif4e1 and pat1-1 mutants also displayed increased RBOHD protein levels similar to cbe1 (Fig. 5B), suggesting that RBOHD levels may be regulated by these factors.

Discussion
Immune signaling relies on tight regulation to allow a proportional and timely response (11,39). Here, we report that CBE1 contributes to RBOHD protein accumulation and consequently elicitor-induced ROS production and antibacterial immunity. Similarly, mutants of the decapping factor PAT1 and the translation initiation factor eIF4E phenocopy  cbe1. Overall, this suggests that CBE1, PAT1, and eIF4E regulate RBOHD levels translationally and thereby affect elicitor-induced ROS production. Translational regulation of plant immunity has recently been proposed, as elicitor perception induces global translational reprogramming (40)(41)(42) and remodeling of the cellular RNA-binding proteome (43). Notably, some of these RBPs control transcripts encoding important immune signaling components. For example, alternative splicing targets genes encoding PRRs, kinases, transcription factors, and leucine-rich repeat receptors (44)(45)(46)(47)(48)(49)(50)(51). In addition, the decapping and deadenylation protein complex as well as NMD factors have been shown to regulate stress-responsive transcripts (52)(53)(54)(55)(56)(57). Accordingly, these changes at the level of RBPs and transcripts contribute to plant immune responses against viruses (which depend on host translation) and other pathogens (43,56,57). ROS play an important role for biological processes such as plant development and responses to abiotic and biotic stresses but are also extremely reactive and toxic at high levels, making their regulated production critical to homeostasis (58). Finetuning of ROS accumulation happens at different levels in space and time (58), including post-translational modification of NAPDH oxidases. For instance, the most highly expressed NAPDH oxidase, RBOHD, is actively regulated to fine-tune ROS production to permit growth, signaling, and development while avoiding toxicity at high level (58)(59)(60)(61)(62)(63). Recently, post-translational modifications through phosphorylation and ubiquitination of RBOHD were shown to regulate its accumulation during immunity (63). Our work here suggests that CBE1 and other translational regulators represent another layer of regulation of RBOHD protein accumulation; however, the exact underlying mechanistic details remain unknown. Nevertheless, this study further emphasizes the importance of regulating ROS production through modulation of RBOHD abundance. Investigating if CBE1 binds RBOHD transcripts directly or binds other transcripts whose products regulate RBOHD levels will be important to further understand the role AtCBE1 negatively regulates RBOHD of CBE1. To determine if this is part of a regulated attenuation mechanism, it will also be necessary to determine if RBOHD is under immune-induced translational control. Interestingly, recent results demonstrated that during immune signaling, RBOHD transcripts increased in the set of ribosome-loaded mRNAs (64). However, the role of CBE1 in that process is still unknown, and expressing CBE1 in plants and bacteria has proven challenging (22). Accordingly, we failed to generate stable Arabidopsis transgenic lines expressing epitope-tagged CBE1 despite multiple attempts (Table S2). This highlights the importance of generating novel tools to answer these questions in future studies.
Based on previous work showing the association between CBE1 and eIF4E1 (22), as well as the colocalization and mutant analysis presented here, we suggest that CBE1 might work together with decapping factor DCP1 and translation initiation factor eIF4E1 to regulate RBOHD protein level and consequently elicitor-induced ROS production and immunity. We found that mutants lacking initiation factor eif4e showed similar enhanced sensitivity to elf18 as cbe1, whereas mutants in other initiation factors (eif4iso4e and eifiso4g1 eifiso4g2) were indistinguishable from WT. These results are in accordance with the specificities of the different eIF isoforms, which bind the 5 0 mRNA cap with a range of affinities (65,66). We also observed enhanced elf18-induced ROS and RBOHD accumulation in pat1-1, which is surprising as eIF4E1 and PAT1 are predicted to function antagonistically. Indeed, eIF4E initiates recruitment of the initiation complex and subsequent recruitment of ribosomes, whereas PAT1 contributes to decapping, which initiates 5 0 -3 0 decay by exoribonucleases (38). In addition, CBE1 seems to localize predominantly to P-bodies, which are generally associated with mRNA decay (67). Interestingly, the number of P-bodies increases when Arabidopsis is treated with flg22 (38,56), suggesting a link between P-body-mediated mRNA stability and immunity. Yet, we could not observe any increase in CBE1 levels or colocalization with P-bodies in N. benthamiana upon flg22 treatment, which could however be due to heterologous overexpression. Given that CBE1 is a plant-specific and nonessential protein, it has been proposed to regulate targeted transcripts in a context-dependent manner (22), which could conceivably provide a fine-tuning mechanism to regulate gene expression. Further work is needed to understand how CBE1 functions in translation initiation and/or mRNA decay.
N. benthamiana plants were grown on soil as one plant per pot (8 × 8 cm) at 25 C during the day with 16 h light (120 μmol m −2. s −1 ) and at 22 C during the night (8 h). Relative humidity was maintained at 60%.

Map-based cloning and whole-genome sequencing
The bak1-5 mob7 mutant (in Col-0) was crossed to Ler-0. Fifty-six F 2 segregants were genotyped for bak1-5 using a dCAPS marker (Table S1). Homozygous bak1-5 segregants were phenotyped for elf18-induced ROS production as for mob7. Linkage analysis was performed using an array of genome-wide markers designed in-house or by the Arabidopsis Mapping Platform (Table S1) (72). For whole-genome sequencing, 440 F 2 plants from the cross bak1-5 mob7 with bak1-5 were scored for chitin-induced ROS production. One hundred thirty-three plants showed moderately increased and 93 plants highly increased ROS production. Out of these 93 plants, 70 were tested in the F 3 generation, and only 15 showed a confirmed phenotype to restore Atpep1-induced seedling growth inhibition in three experiments. Thirty seedlings from each of the positive F 3 parents were bulked and ground to a fine powder in liquid nitrogen and gDNA extracted. Ground tissues were equilibrated in buffer containing 50 mM Tris-HCl (pH 8.0), 200 mM NaCl, 2 mM EDTA for 30 min at 37 C with occasional mixing, and a further 20 min at 37 C with 0.2 mg/ ml RNase. Roughly 10 ng of genomic DNA was then extracted using a standard chloroform/phenol method and resuspended in TE buffer (10 mM Tris HCl pH 7.5; 1 mM EDTA pH 8). Prepared gDNA of pooled bak1-5 mob7 F 3 segregants, as well as bak1-5 as a reference (12), was submitted to The Beijing Genomics Institute (Hong Kong) for Illumina-adapted library preparation and paired-end sequencing using the High-Seq 2000 platform. The average coverage from Illumina sequencing of bak1-5 mob7 over the nuclear chromosomes was 15.79. Paired-end reads were aligned to the TAIR10 reference assembly using BWA v 0.6.1 with default settings (73). BAM files were generated using SAMTools v 0.1.8 (73), and single-nucleotide polymorphisms (SNPs) were called using the mpileup command. High-quality SNPs were obtained using the following filters: (a) Reads with mapping quality less than 20 were ignored; (b) SNP position had a minimum coverage of six and a maximum of 250; (c) the reference base must be known; and (d) SNPs were present in bak1-5 mob7 but not in the bak1-5 control. The resulting pileup files contained a list of SNPs and their genomic positions. SNPs unique to bak1-5 mob7 and not present in the bak1-5 control were identified. SNPs passing filters were analyzed on CandiSNP (74). Relevant SNPs were confirmed in the original bak1-5 mob7 mutant and backcrossed lines by Sanger sequencing of PCR amplicons.
Oxidative burst assay ROS production was measured as previously described (23). For the assay, either adult plants (4-to 6-week-old plants) or seedlings (2-week-old) were used. For adult plants, leaf discs (4-mm diameter) were collected using a biopsy punch and floated overnight on distilled, deionized water in a white 96well plate to recover from wounding. For ROS assays on whole seedlings, seedlings were grown on MS agar plates for 5 days before being transferred to MS liquid medium in transparent 96-well plates. After 8 days, seedlings were transferred to a white 96-well plate and allowed to recover overnight in sterile water. The water was then removed and AtCBE1 negatively regulates RBOHD replaced with elicitor solution containing 17 μg/ml luminol (Sigma-Aldrich), 100 μg/ml horseradish peroxidase (Sigma-Aldrich), and the indicated elicitor concentration. For seedlings, the hyperactive luminol derivative 0.5 μM L-012 (Fujifilm Wako Chemicals) was used instead of luminol. Luminescence was recorded over a 40-to 60-min period using a charge-coupled device camera (Photek Ltd).

Seedling growth inhibition assay
Seedling growth inhibition was performed as previously described (23). Sterilized and stratified seeds were sown on MS media and grown in controlled environment rooms with 16/ 8 h day/night cycle and constant temperature of 22 C. Fiveday-old seedlings were transferred into liquid MS with or without the indicated amount of elicitor. 10 to 12 days later, individual seedlings were gently dry-blotted and weighed using a precision scale (Sartorius).

MAP kinase phosphorylation assay
Phosphorylation of MAPKs was measured as previously described (78). Leaf discs (4-mm diameter) from adult plants (4-to 6-week-old plants) were cut in the evening and left overnight on the bench, floating in 6-well plates on distilled, deionized water. In the morning, the elicitor peptide was added to the desired concentration, and tissue was blotted dry and flash-frozen in liquid nitrogen for protein extraction at the indicated time points. MAPK phosphorylation was detected by Western blot using an antibody specific to the active phosphorylated form of the proteins (phospho-p44/42 MAPK). Fifteen leaf discs were used per condition.

Bacterial spray inoculation
Spray inoculations were performed as previously described (79). P. syringae pv. tomato (Pto) DC3000 wildtype and COR -(defective in production of the phytotoxin coronatine) strains (80) were grown in overnight culture in King's B medium supplemented with 50 μg/ml rifampicin, 50 μg/ml kanamycin, and 100 μg/ml spectinomycin and incubated at 28 C. Cells were harvested by centrifugation and pellets resuspended in 10 mM MgCl 2 to an A 600 of 0.2, corresponding to 1 × 10 8 colony forming units (CFU)/ml. Immediately before spraying, Silwet L-77 (Sigma Aldrich) was added to a final concentration of 0.04% (v/v). Four-to five-week-old plants were uniformly sprayed with the suspension and covered with a clear plastic lid for 3 days. Three leaf discs (4-mm diameter) were taken using a biopsy puncher from three respective leaves of one plant and ground in collection microtubes, containing one glass bead (3mm diameter) and 200 μl water, using a 2010 Geno/Grinder (SPEX) at 1500 rpm for 1.5 min. Ten microliters of serial dilutions from the extracts were plated on LB agar medium containing antibiotics and 25 μg/ml nystatin (Melford). Colonies were counted after incubation at 28 C for 1.5 to 2 days.
Transient expression in N. benthamiana N. benthamiana plants were used for transient transformation at 4-to 5-weeks postgermination. Agrobacterium tumefaciens GV3101 overnight cultures grown at 28 C in LB were harvested by centrifugation at 2500g and resuspended in buffer containing 10 mM MgCl 2 and 10 mM MES for 3 h at room temperature. A. tumefaciens-mediated transient transformation of N. benthamiana was performed by infiltrating leaves with A 600 = 0.2 of each construct together with the viral suppressor P19 (84) in a 1:1 (or 1:1:1) ratio. Samples were collected 2 to 3 days after infiltration.

Stable transformation of Arabidopsis
Transgenic Arabidopsis plants were generated using floral dip method (85). Briefly, flowering plants were dipped into a suspension culture of A. tumefaciens GV3101 carrying the indicated plasmid. Plants carrying a T-DNA insertion event were selected either on MS medium containing the appropriate selection or as soil-grown seedlings by spray application of Basta (Bayer Crop Science). T 1 seedlings resistant to selective marker on MS plate were transferred to soil to produce the next generation. T 2 resistance was monitored to find single insertion lines, while T 3 resistance was screened for homozygous mutants and expression of tagged lines verified by Western blot.

Confocal microscopy
N. benthamiana leaf discs (4-mm diameter) transiently overexpressing the indicated proteins were sampled at 2 to 3 dpi with water as the imaging medium. For elicitor treatment in N. benthamiana, leaf discs were harvested 3 dpi and incubated overnight in petri dishes containing water. The next day, leaf discs were transferred to microscopic slides containing 1 μM flg22 or water. Live-cell imaging employed a laserscanning Leica SP5 Confocal Microscope (Leica Microsystems) and 63x (glycerol immersion) objective. GFP was excited at 488 nm and emission detected between 496 and 536 nm (shown in green). YFP was excited at 514 nm and detected between 524 and 551 nm (shown in yellow). RFP derivatives (mRFP, mCherry, tag-RFP) were excited at 561 nm and detected between 571 and 635 nm (shown in magenta). Colocalization was performed using sequential channel analysis by calculating Pearson's coefficient (31,86) using the Coloc 2 plugin of ImageJ. Image analysis was performed with Fiji (87).

RNA extraction and qPCR analysis
Total RNA was extracted using Trizol reagent (Invitrogen) according to the manufacturer's instructions. RNA samples were treated with Turbo DNA-free DNase (Ambion) according to the manufacturer's instructions. RNA was quantified with a Nanodrop spectrophotometer (Thermo Fisher Scientific). cDNA was synthesized from RNA using Rever-tAid (Thermo Fisher Scientific) according to the manufacturer's instructions. Quantitative PCR was conducted following the MIQE guidelines (88) using a 7500 Real-Time PCR System (Applied Biosystems) and PowerUp SYBR Green Master Mix (Applied Biosystems) with cDNA diluted 1:20. The 2 −ΔCt method was used for the calculation of relative expression.

RNA stability assay
RNA stability was measured as previously described (89). Briefly, three leaf discs from different plants (5-week-old) were collected in 24-well plate with 0.5 ml sterile water. The next day cordycepin (Chengdu Biopurify Phytochemicals) was added to a final concentration of 0.6 mM and discs were sampled at 0, 30, 60, 90, or 120 min, blotted dry, and flash frozen.

Statistical analysis
Statistical analysis was performed using R (4.1.2) and Rstudio (2021.09.1) or GraphPad Prism (9.3). Based on Gaussian distribution, parametric or nonparametric tests were chosen and when n ≥ 30, normal distribution was assumed. Prior to multiple comparisons, ANOVA or Kruskal-Wallis test were performed to assess differences across groups. For multiple comparisons, Dunnett's and Dunn's tests were favored to compare multiple groups to one control group. Tests were realized on the overall set of replicates, and replicates were included only when positive and negative controls showed the expected results.

Data availability
All data are contained within the manuscript.
Supporting information-This article contains supporting information.