N 6-methyladenosine modifications stabilize phosphate starvation response–related mRNAs in plant adaptation to nutrient-deficient stress

Liu, Kai; Wang, Xiaojia; Wang, Jingyi; Wang, Shuman; Bai, Haiyan; Dong, Weiguo; Qiao, Lulu; Jin, Qiongli; Zhang, Zhonghui; Luo, Guan-Zheng; Wang, Zhiye

doi:10.1038/s41467-025-59331-y

Download PDF

Article
Open access
Published: 01 May 2025

N ⁶-methyladenosine modifications stabilize phosphate starvation response–related mRNAs in plant adaptation to nutrient-deficient stress

Nature Communications volume 16, Article number: 4093 (2025) Cite this article

4977 Accesses
1 Citations
3 Altmetric
Metrics details

Subjects

Abstract

N⁶-methyladenosine (m⁶A), an abundant internal mRNA modification, is induced by various stress conditions and post-transcriptionally regulates gene expression. However, how m⁶A modifications help plants respond to nutrient-deficiency stress remains unclear. Here, we profile high-confidence m⁶A modifications in Arabidopsis transcriptome-wide under normal and inorganic orthophosphate (Pi)–deficient conditions (−P). High-confidence m⁶A modifications are identified using synthetic modification-free RNA libraries for systematic calibration. Pi starvation induces widespread m⁶A modifications, mediated by the Pi starvation response (PSR) master regulator PHOSPHATE STARVATION RESPONSE1 (PHR1) and its family members. Many Pi starvation–induced (PSI) m⁶A modifications occur on PSR-related mRNAs, including PHR1. In addition, PHR1 proteins interact with the m⁶A writers MRNA ADENOSINE METHYLASE (MTA) and METHYLTRANSFERASE B (MTB) in nuclei under −P conditions. m⁶A modifications facilitate systemic PSR signaling, as reflected by the reduced Pi content and PSR signaling in a knockdown artificial miRNA line targeting MTA, which shows a global decrease in m⁶A. Transcriptome-wide mRNA decay analysis reveals that PSI-m⁶A increases the stability of PSR-related mRNAs, but not through alternative polyadenylation site shifts. Analysis of transgenic plants with mutations in m⁶A loci demonstrates that m⁶A stabilizes PHR1 transcripts via a positive feedback loop. Our findings indicate that PSI-m⁶A modifications facilitate PSR signaling by enhancing the stability of certain mRNAs, shedding light on the role of m⁶A modifications in nutrient stress responses in plants.

RNA m⁶A modification meets plant hormones

Article 28 March 2025

Quantitative profiling of m⁶A at single base resolution across the life cycle of rice and Arabidopsis

Article Open access 07 June 2024

N⁶-methyladenosine RNA modification regulates photosynthesis during photodamage in plants

Article Open access 02 December 2022

Introduction

The most prevalent internal covalent modification of eukaryotic messenger RNAs (mRNAs) is the methylation of adenines to produce N⁶-methyladenosine (m⁶A), which occurs predominantly near the 3′ ends^1,2,3. The m⁶A modification is reversibly added and removed by evolutionarily conserved methyltransferases (writers) and demethylases (erasers), respectively^4,5,6,7,8. In Arabidopsis (Arabidopsis thaliana), the methylation reaction is primarily catalyzed by a multi-protein complex comprising MRNA ADENOSINE METHYLASE (MTA; ortholog of human METTL3), METHYLTRANSFERASE B (MTB; ortholog of human METTL14), FKBP12 INTERACTING PROTEIN37 (FIP37; ortholog of human WTAP), VIRILIZER (VIR), and HAKAI^{9,10,11,12,13}. In a small subset of mRNAs, m⁶A modifications are also produced by another methyltransferase, FIONA1 (FIO1; ortholog of human METTL16)^14,15,16. The m⁶A methyl groups can be removed by the demethylases ALKBH9B and ALKBH10B (orthologs of human ALKBH5)^17,18,19.

The underlying mechanisms and biological functions of m⁶A have been studied by profiling this modification across the transcriptomes of plants at various developmental stages and under different growth conditions^2,3. Transcriptome-wide m⁶A mapping usually involves immunoprecipitation of methyl-RNA combined with high-throughput RNA sequencing (MeRIP-seq or m⁶A-seq); however, the fidelity of this method is compromised by the non-specific binding of the anti-m⁶A antibody^20,21,22. To improve the reliability of m⁶A-seq, Luo and colleagues developed a method for the systematic calibration of epitranscriptomic maps using synthetic modification-free RNA library controls that resembled the endogenous transcriptome in terms of sequence and expression diversity²². This synthetic RNA library approach serves as a more robust and rigorous control²² than previously employed calibration controls, such as methyltransferase gene knockout and in vitro demodification. Although this gold-standard negative control has been used in epitranscriptome mapping studies of animals and viruses^23,24, it has not previously been used for plants.

Environmental stress dynamically alters the m⁶A landscape^3,24,25, reflecting the important regulatory functions of this modification in the adaptation to stress. In plants, the dynamic of m⁶A modifications responds to several abiotic and biotic stresses, such as salt^26,27, drought^28,29, heat³⁰, cold³¹, excess light³², viral pathogens^33,34, and fungal diseases³⁵. However, the underlying functions of m⁶A in stress adaptation are not well understood. Several studies have demonstrated that stress causes dynamic m⁶A deposition on specific stress-responsive transcripts, which mainly affects their stability. For example, salt stress–induced m⁶A not only stabilizes transcripts involved in the salt-stress response²⁷, but also destabilizes transcripts involved in the negative regulation of this response²⁶. These contrasting effects on RNA stability were also observed in the m⁶A-mediated regulation of development^{10,17,36,37,38}, pointing to a complex relationship between m⁶A modification and RNA decay. Despite these insights, a transcriptome-wide RNA half-life assay with m⁶A profiling under normal and stress conditions has not yet been performed in plants.

Nutrient deficiency is a major limiting factor in global crop production^39,40; however, the regulation of m⁶A modification in the plant response to nutrition-deficiency stresses remains unclear. Phosphorus (P) is essential for all life on Earth^41,42,43. P is a key element for photosynthesis, an important component of macromolecules such as nucleic acids, and is involved in posttranslational modifications and the biosynthesis of essential metabolites⁴⁴. Plants take up P as orthophosphate (Pi) from the soil. Studies of the plant response to Pi starvation are ecologically and agriculturally important because soils in many parts of the world have low levels of available Pi due to its immobility, low solubility, or the general lack of this mineral^45,46.

To withstand environmental Pi limitation, plants have evolved a complex Pi-starvation response (PSR) comprising a series of developmental and physiological changes^47,48. The evolutionarily conserved PHOSPHATE STARVATION RESPONSE1 (PHR1) and the related PHR1-like (PHL) family, collectively termed PHRs, act as central transcriptional regulators of systemic PSR signaling^49,50,51,52. PHRs belong to the GARP transcription factor (TF) family and possess a MYB-like DNA-binding domain and a coiled-coil domain^49,50. Upon Pi-deficiency, PHR1 forms homodimers or heterodimers with PHL1 and directly binds to a pseudo-palindromic DNA sequence (GNATANC; P1BS) in the promoters of many Pi starvation–induced (PSI) genes to upregulate their expression^49,53. Transcriptome analysis revealed that PHRs also indirectly regulate the expression of some Pi starvation–decreased (PSD) genes⁵³, although the underlying mechanism remains elusive. Besides transcriptional reprogramming, PHRs regulate changes in chromatin accessibility⁵⁴, lipid remodeling⁵⁵, metabolic changes⁵⁶, and mycorrhizal symbiosis⁵⁷ in response to Pi starvation. However, it is not known whether PHRs also regulate m⁶A modification and thereby RNA stability under Pi starvation. As PHRs are fundamental regulators of the PSR, their activity is extensively regulated. PHR1 activity is inhibited through physical interactions with SPX (named after yeast Syg1 and Pho81 and mammalian XPR1) proteins in response to Pi status^{42,58,59,60,61}. Other types of posttranslational regulation of PHR activity were also reported, including protein degradation⁶², sumoylation⁶³, and phosphorylation⁶⁴. However, the regulation in PHR1 transcript levels remains uncharacterized.

Here, we performed m⁶A-seq with systematic calibration using synthetic modification-free RNA libraries and produced high-confidence m⁶A profiles in Arabidopsis under normal and Pi-deficient stress conditions. We determined that Pi starvation induces global m⁶A modifications, largely mediated by PHR proteins. Many PSI-m⁶A modifications are deposited in PSR-related transcripts, including PHR1 mRNA. Although PHRs do not regulate PSI-m⁶A modifications through direct transcriptional regulation, PHR1 interacts with the m⁶A writers MTA and MTB in nuclei under −P conditions. A genome-wide RNA decay assay demonstrated that PHR-mediated PSI-m⁶A modifications stabilize PSR-related mRNAs. PAS-seq 2 data indicated that PSI-m⁶A did not impact RNA stability via alternative polyadenylation site (APA) shifts. Furthermore, mutations in putative m⁶A-loci in the 3′ untranslated region (UTR) of PHR1 mRNA led to shortened RNA half-lives and partial phr1 complementation, confirming that m⁶A stabilizes PHR1 transcripts as a positive feedback mechanism to promote PSI signaling and stress adaptation upon Pi starvation.

Results

Transcriptome-wide profiling of high-confidence m⁶A modifications using systematic calibration with a synthetic modification-free RNA library

To explore the regulatory function of m⁶A in Pi homeostasis and signaling, we performed m⁶A-seq to examine the effect of Pi starvation on transcriptome-wide m⁶A levels. To correct for non-specific binding to the anti-m⁶A antibody during m⁶A-seq, we used a synthetic modification-free RNA library to calibrate the epitranscriptomic maps²² (Fig. 1a).

**Fig. 1: Transcriptome-wide high-confidence m⁶A profiling with systematic calibration.**

Following the protocol shown in Supplementary Fig. 1a²², we replicated the entire transcriptome (all mRNAs) using T7 RNA polymerase–mediated in vitro transcription (IVT) and unmodified nucleosides (referred to as IVT RNA) (Fig. 1b). We confirmed the absence of m⁶A modifications in the synthesized no-m⁶A IVT RNA using liquid chromatography–mass spectrometry (LC–MS/MS) (Fig. 1c). We next performed mRNA-seq, which showed that approximately 95% of the genes represented in the endogenous total mRNA were also represented in the no-m⁶A IVT RNA. Furthermore, the global gene expression levels in the no-m⁶A IVT RNA library were strongly correlated with those in the endogenous transcriptome (Pearson’s r = 0.954) (Fig. 1d). The read coverages across the transcripts were also similar between the IVT RNA and endogenous mRNA libraries (Fig. 1e). These findings indicate that we successfully generated a non-m⁶A modified IVT RNA library that closely resembled the endogenous transcriptome in terms of sequence representation and expression.

We generated two biological replicates of the m⁶A-seq and the corresponding mRNA-seq using endogenous mRNA and synthetic no-m⁶A IVT RNA from seedlings grown under normal (+P) and Pi-starvation (−P) conditions. We mapped the sequencing reads (approximately 20–28 million reads per sample) to the Arabidopsis TAIR10 reference genome and used the exomePeak pipeline to call the m⁶A peaks in the transcripts of all samples using the immunoprecipitation (IP) enrichment criteria (IP/input ≥ 2, FDR < 0.05). More than 14,000 peaks were identified in the endogenous mRNAs (Supplementary Fig. 1b). Most peaks (~ 90%) were represented in both biological replicates, indicating high reproducibility (Supplementary Fig. 1b). In total, 12,837 peaks and 13,379 common peaks were identified in the endogenous mRNA samples under +P and −P conditions, respectively (Supplementary Fig. 1b and Supplementary Data 1). Strikingly, numerous m⁶A-irrelevant peaks were identified in the corresponding synthetic no-m⁶A IVT RNA samples, indicating the non-specific binding of the anti-m⁶A antibodies (Supplementary Fig. 1c, Supplementary Data 1). The peaks in the endogenous mRNA libraries were mainly located around stop codons and in 3′ UTRs (Fig. 1f), which is consistent with a previously published m⁶A peak distribution^1,2,3. By contrast, the m⁶A-irrelevant peaks in the no-m⁶A IVT RNA libraries were mainly distributed in coding sequence (CDS) regions (Fig. 1f). Moreover, VAGGUA (V = A/C/G) and RGMRGR (R = A/G; M = A/C) motifs were identified in the m⁶A-irrelevant peaks (Fig. 1g), which are similar to the motifs identified in the IVT RNA m⁶A-seq library in a mammalian study²².

Next, we systematically calibrated the transcriptome-wide m⁶A modifications with the identified m⁶A-irrelevant peaks to remove false-positive m⁶A peaks from the endogenous mRNA m⁶A-seq libraries (Fig. 1h, see Methods). Approximately 2000 and 1500 false-positive peaks were removed from the +P and −P samples, respectively, accounting for 11–16% of the total peaks (Fig. 1h). Consequently, we identified 10,757 and 11,861 high-confidence m⁶A peaks associated with 10,420 and 11,496 genes in Col-0 (WT) seedlings under +P and −P conditions, respectively (Fig. 1h and Supplementary Data 1). We also confirmed the false-positive and high-confidence m⁶A peaks using the Integrative Genomic Viewer (IGV), finding that false-positive m⁶A loci exhibited similar peaks in the WT endogenous mRNA and IVT RNA IPs, whereas the high-confidence peaks were only present in the WT mRNA IPs (Fig. 1i and Supplementary Fig. 1d). Furthermore, the high-confidence m⁶A peaks exhibited stronger enrichment around stop codon and in 3′ UTR than did the uncalibrated peaks (Fig. 1j). In addition to the plant-specific UGUA motif and common RRACH (H = A/C/U) motif, we also identified a CAGR motif in the calibrated m⁶A peaks under both +P and −P conditions (Supplementary Fig. 1e).

Therefore, the use of systematic calibration with a synthetic modification-free RNA library allowed us to identify high-confidence m⁶A peaks under normal and Pi-deficient conditions in WT seedlings.

Pi starvation induces global changes in m⁶A levels

Under −P, the m⁶A levels across the transcriptome were dramatically increased in comparison with +P conditions (Fig. 2a, b). In addition, 1651 transcripts underwent de novo m⁶A modification in response to −P conditions, while only 575 m⁶A-modified transcripts occurred under +P conditions (Fig. 2c). We accurately identified the m⁶A peaks that differed between +P and −P conditions using the exomePeak algorithm, with the criteria of fold change ≥ 1.5 and FDR < 0.05, excluding m⁶A-irrelevant peaks. Pi starvation resulted in 4418 hypermethylated m⁶A peaks (PSI-m⁶A) across 4364 genes compared to only 323 hypomethylated m⁶A peaks (PSD-m⁶A) across 316 genes (Fig. 2d, Supplementary Data 1), indicating that Pi starvation enhances the production of m⁶A modifications across the transcriptome. Compared with other published sets of genes whose m⁶A profiles changed in response to treatment¹⁹, major m⁶A-hypermethylation (71.1%) and m⁶A-hypomethylation (98.4%) in response to Pi starvation are specific (Supplementary Fig. 2a).

**Fig. 2: Pi starvation induces global m⁶A modification.**

These PSI-m⁶A and PSD-m⁶A peaks were mainly located near stop codons and in 3′ UTRs, although a few peaks were located at CDSs and 5′UTRs (Supplementary Fig. 2b, c). The plant-specific UGUA motif was enriched within both PSI-m⁶A and PSD-m⁶A peaks at the 3′UTR (Supplementary Fig. 2d). Notably, an unreported CAGR motif was specifically enriched within PSI-m⁶A peaks at the 3′UTR, indicating a potential link of this motif with PSI-m⁶A function (Supplementary Fig. 2d). We performed a Gene Ontology (GO) analysis to investigate the PSI-m⁶A hypermethylated genes. In addition to functions in phosphorus metabolism, the PSI-m⁶A genes are involved in RNA stability, RNA splicing and 3′-end processing, rRNA processing, tRNA modification, and histone and DNA modification pathways, which are indicative of m⁶A-mediated regulations of RNA metabolism and transcription in response to Pi starvation (Fig. 2e). PSI-m⁶A genes are also involved in photosynthesis, defense responses, protein transmembrane transport, and protein lipidation (Fig. 2e). By contrast, PSD-m⁶A hypomethylated genes are involved in responses to stress, inorganic anion transport, and sulfur compound biosynthesis (Fig. 2f). These results suggest that m⁶A RNA modifications are involved in regulating multiple processes during the PSR, from transcription to RNA metabolism and ion homeostasis.

We then focused on the PSR-related genes. Given that Pi starvation triggers the altered expression of thousands of genes⁵³, we assessed whether the changes in m⁶A modification were associated with changes in the expression of PSR genes. Consistent with previous studies⁵³, our RNA-seq data showed that thousands of genes were upregulated (−P/+P ≥ 1.5, P < 0.05; PSI genes) or downregulated (+P/−P ≥ 1.5, P < 0.05; PSD genes) by Pi-deficiency (Supplementary Fig. 2e, Supplementary Data 2). Notably, approximately one-third of the PSI genes and more than 70% of the PSD genes contained m⁶A modifications (Supplementary Fig. 2f). Furthermore, a joint analysis of the m⁶A-seq and RNA-seq data revealed a large number of m⁶A-hypermethylated PSD genes (638 PSI-m⁶A & PSD genes, accounting for ~71% of the total m⁶A-altered PSR genes). These results suggest that m⁶A modification might regulate the RNA levels of PSD genes in response to Pi starvation (Fig. 2g and Supplementary Data 1).

Furthermore, the transcripts of several known Pi homeostasis and PSR-signaling genes harbored m⁶A modifications. These genes included the central PSR-signaling gene PHR1 and its homologs^49,53, the PSR-signaling repressor genes SPX2 and SPX4^58,65, the phosphate transporter trafficking facilitator gene PHOSPHATE TRANSPORTER TRAFFIC FACILITATOR1 (PHF1)^66,67, the Pi efflux transporter gene PHO1⁶⁸, the root tip low-phosphate sensor LOW PHOSPHATE ROOT2 (LPR2)⁶⁹, the diphosphoinositol pentakisphosphate kinase genes VIP1 HOMOLOG1 (VIH1) and VIH2⁷⁰, and the vacuolar Pi transporter genes PHOSPHATE TRANSPORTER 5;1 (PHT5;1) and PHT5;3^71,72 (Fig. 2h, i and Supplementary Fig. 2g). The GO results also showed PSR-related pathways represented by m⁶A-modified PSI genes, regardless of PSI-m⁶A (Supplementary Fig. 2h, i).

Together, the above data show that Pi starvation induces global changes in m⁶A levels. Many PSR-related transcripts were m⁶A modified, suggesting that m⁶A regulates Pi homeostasis and PSR signaling.

Increases in m⁶A levels in PSR are largely mediated by PHRs

PHR1 and its homologs are central TFs controlling multiple processes of the PSR^47,48, raising the question of whether PSI-m⁶A modifications are mediated by PHRs. To address this, we conducted m⁶A-seq using the phr1 phl1 double knockout mutant under normal (+P) and Pi-deficient (−P) conditions. We obtained high-quality and highly reproducible m⁶A-seq and corresponding mRNA-seq data (used as Input) (Fig. 3a, b and Supplementary Fig. 3a, b). After systematic calibration with the non-m⁶A IVT RNA libraries (Fig. 1a), approximately 11,000 high-confidence m⁶A peaks were identified in the phr1 phl1 double mutant under both +P and −P conditions (Supplementary Fig. 3c, d and Supplementary Data 1).

**Fig. 3: Pi starvation–induced m⁶A modifications are largely mediated by PHRs.**

Clustering and principal component analysis (PCA) revealed distinct clustering patterns in the m⁶A-seq data of WT and phr1 phl1 between +P and −P conditions, with the phr1 phl1 and WT clusters near each other under +P but distant both from each other and from the +P clusters under −P (Fig. 3a, b). These results highlight the dramatic difference in global m⁶A modifications in WT and phr1 phl1 plants under Pi starvation, but not under normal conditions. This finding was confirmed by transcriptome-wide differential m⁶A modification analysis using the exomePeak algorithm. Under +P, the transcriptome-wide m⁶A profiles of WT and phr1 phl1 were comparable (P = 0.88, Wilcox test) (Fig. 3c), with fewer than 1000 differentially methylated m⁶A peaks between the two genotypes (Supplementary Fig. 3e and Supplementary Data 1). Under Pi starvation, however, the loss of PHR1 and PHL1 diminished the global increase in PSI-m⁶A modifications observed in the WT (P < 2.2e-16, Wilcox test) (Fig. 3c). Under these conditions, the levels of 3265 m⁶A peaks decreased (hypo-m⁶A peaks, corresponding to 3194 hypo-m⁶A genes), whereas the levels of only 477 m⁶A peaks increased (hyper-m⁶A peaks, corresponding to 465 hyper-m⁶A genes) in phr1 phl1 compared to the WT (Fig. 3d and Supplementary Data 1). Furthermore, approximately 60% of these hypo-m⁶A peaks were represented in the PSI-m⁶A peaks (hereafter PHR-mediated PSI-m⁶A peaks) (Fig. 3e and Supplementary Data 1). Meanwhile, 70 hyper-m⁶A peaks were PSD-m⁶A peaks (hereafter, PHR-mediated PSD-m⁶A peaks) (Fig. 3f and Supplementary Data 1), suggesting that the PHRs regulate the Pi starvation–triggered m⁶A dynamics, mainly by mediating the induction of PSI-m⁶A modifications. We further confirmed that the m⁶A levels of these corresponding PHR-mediated PSI-m⁶A genes significantly increased in response to Pi starvation stress in the WT, but not in phr1 phl1 (Fig. 3g, h). Notably, m⁶A induction of both total PSI-m⁶A and non-PHR-mediated PSI-m⁶A genes was also significantly repressed in phr1 phl1 compared with the WT under −P conditions, indicating that PHRs mediated global PSI-m⁶A modifications (Supplementary Fig. 3f, g). Consistent with the GO results for the PSI-m⁶A genes (Fig. 2e), GO analysis of the PHR-mediated PSI-m⁶A genes and non-PHR-mediated PSI-m⁶A genes revealed their involvement in phosphorus metabolism, RNA splicing, RNA stability, chromatin remodeling, photosynthesis, and protein transport, among other functions (Supplementary Fig. 3h, i).

These results indicate that the PSI-m⁶A modifications are dependent on PHRs, suggesting a new role of PHR proteins in regulating the PSR via m⁶A modification.

PHRs do not mediate PSI-m⁶A modification through direct transcriptional regulation

The PHR TFs regulate the expression of many PSI genes⁵³; thus, we investigated whether the PHRs mediate the m⁶A levels of these PSI-m⁶A transcripts via the direct transcriptional regulation of these genes. We first compared the PHR-mediated PSI-m⁶A genes with direct target genes of PHR1 that were previously identified using chromatin immunoprecipitation (ChIP)-seq⁷³. Only 155 of the PHR-mediated PSI-m⁶A genes (accounting for 8% of the total) overlapped with the direct target genes of PHR1 (accounting for 7% of the total) (Supplementary Fig. 4a), indicating that the PHR1-mediated regulation of PSI-m⁶A modifications is mainly achieved via a mechanism other than direct transcriptional regulation. To further explore this idea, we compared our RNA-seq data from WT and phr1 phl1. More than half of the PSI genes (~ 51%) were repressed (classified as PHR-controlled PSI genes), while only ~15% of PSD genes were upregulated (classified as PHR-controlled PSD genes) in phr1 phl1 compared with the WT under −P conditions (Supplementary Fig. 4b, c and Supplementary Data 2). This is consistent with the previous finding that PHRs mainly transcriptionally regulate the induction of PSI genes under −P conditions⁵³. However, only 34 of the PHR-mediated PSI-m⁶A genes (accounting for 2% of the total) overlapped with the PHR-controlled PSI genes (accounting for 3% of the total) (Supplementary Fig. 4d), confirming the notion that PHRs do not transcriptionally regulate PHR-mediated PSI-m⁶A genes.

In addition, we measured the expression levels of genes encoding m⁶A writers (MTA, MTB, FIP37, VIR, HAKAI, and FIONA1) and erasers (ALKBH9B and ALKBH10B) in the WT and phr1 phl1 under +P and −P conditions. The expression of these genes was not greatly altered in response to Pi-deficiency or the mutation of PHR1 and PHL1 (Supplementary Fig. 4e). We further detected the protein levels of MTA in the WT and phr1 phl1 mutants under +P and −P but found no obvious changes in total proteins (Supplementary Fig. 4f). Besides, the amount of MTA and MTB proteins in nuclei was comparable with the WT between +P and −P conditions (Supplementary Fig. 4g).

Taken together, these data demonstrate that PHRs mediate PSI-m⁶A modifications, but not via either direct transcriptional regulation or protein abundance changes of m⁶A writers, pointing to the involvement of an unknown regulatory mechanism.

PHR1 interacts with MTA and MTB in nuclei under Pi starvation conditions

Several factors interact with m⁶A writers to regulate miRNA biogenesis^74,75, the circadian clock⁷⁶ and chlorophyll homeostasis⁷⁷. Using yeast two-hybrid (Y2H) assays, we observed the interactions of PHR1 with MTA and MTB, but not with FIP37, another component of the core m⁶A methyltransferase complex (MTC) (Fig. 4a). Meanwhile, SPX1 and SPX2, which are repressors of PHR activity, did not interact with any MTC component in the Y2H system (Fig. 4b). To confirm the in vivo interaction between PHR1 and MTA/MTB, we performed bimolecular fluorescence complementation (BiFC) assays (Fig. 4c). PHR1 interacted with MTA and MTB in the nucleus, and they formed speckles in some cells (Fig. 4c). These results are consistent with reports of condensate formation by the MTC^74,75. Furthermore, we validated the PHR1-MTA/MTB interaction in plants under −P conditions by nuclear Co-IP (Fig. 2d).

**Fig. 4: PHR1 interacts with MTA/MTB in the nucleus under –P conditions.**

These results elucidated the association of PHR1 with MTA and MTB under −P conditions in plants. Given the current data obtained under −P conditions in plants, it appears that PHR1, a key transcription factor in the PSR, plays a previously unrecognized role in regulating PSI-m⁶A under −P, possibly through interacting with MTA and MTB.

m⁶A RNA modification facilitates Pi homeostasis and systemic PSR signaling

To further decipher the roles of m⁶A modification in Pi homeostasis and PSR signaling, we developed a line hypomorphic for the m⁶A writer gene MTA using the artificial microRNA method (amiR-mta) (Fig. 5a and Supplementary Fig. 5a). The expression level of MTA in amiR-mta was repressed to ~25% that of WT plants (Fig. 5f and Supplementary Fig. 5b). We quantified the global m⁶A level in amiR-mta using LC–MS/MS, revealing a dramatic reduction in the m⁶A/A ratio compared with the WT (decreased to ~40%) under both +P and −P conditions (Supplementary Fig. 5c).

**Fig. 5: m⁶A modification promotes Pi homeostasis and systemic PSR signaling.**

The amiR-mta plants exhibited inhibited growth under +P conditions, with reduced biomass and shorter primary roots, which is consistent with previous findings (Fig. 5a–c)^10,13. Under −P stress, this growth inhibition of amiR-mta was deteriorated (Fig. 5a–c). We then investigated the Pi homeostasis- and PSR-related physiological traits of the WT and amiR-mta. Compared with the WT, the knockdown of MTA resulted in a significant reduction in the shoot Pi concentration under both +P and −P conditions, indicating that the m⁶A modification plays a positive role in Pi homeostasis (Fig. 5d). In addition, the inhibition of primary root growth is an important trait of the local PSR^47,69. The ratio of primary root length under −P vs. +P conditions was slightly smaller in amiR-mta than in the WT, suggesting that the m⁶A modification moderately represses the local PSR (Fig. 5e). These physiological phenotypes caused by the knockdown of the m⁶A writer gene MTA provide further evidence that m⁶A modifications play a regulation role in Pi homeostasis and the PSR.

We further explored the involvement of the m⁶A modification in PHR-mediated systemic PSR signaling. We used RT-qPCR to assess the steady-state transcript levels of several systemic PSR-signaling genes, including PHR1 and the downstream Pi transport- and Pi metabolism-related PSI genes. The steady-state transcript levels of PHR1, PHF1, PHO1, PHT1;1/PT1, and ACID PHOSPHATASE 5 (ACP5) showed moderate but significant reductions in amiR-mta compared with the WT under −P (Fig. 5f). This is consistent with the reduced Pi concentration of amiR-mta plants (Fig. 5d). To further investigate the enhanced functions of m⁶A modifications in PSR and Pi homeostasis in plants under −P conditions, we treated 7-day-old WT plants grown under normal conditions with the chemical STM2457, an inhibitor of m⁶A writers⁷⁸ for 7 days under −P stress. Compared with 14-day-old WT plants grown under normal conditions, both STM2457-treated and mock WT plants exhibited typical morphological responses under Pi deficiency, such as inhibited primary root growth and anthocyanin accumulation (Supplementary Fig. 5d). Notably, the Pi concentration in shoots was lower in STM2457-treated WT plants than in mock plants under −P conditions (Supplementary Fig. 5e). Moreover, the expression levels of PHR1 and its downstream PSR signaling genes ACP5 and PHF1 were lower in STM2457-treated WT plants compared with mock plants (Supplementary Fig. 5f). These results support the hypothesis that m⁶A modifications positively regulate Pi homeostasis and PHR-mediated systemic PSR signaling in plants under Pi-deficiency stress.

We next profiled the transcriptomes of amiR-mta plants under +P and −P conditions to confirm the positive role of m⁶A in PSR signaling. The knockdown of MTA dramatically altered the transcriptome, 1905 genes were upregulated and 1649 genes were downregulated in amiR-mta compared with the WT under +P conditions, and 2258 genes were upregulated and 1764 genes were downregulated in amiR-mta compared with WT under −P conditions (Supplementary Fig. 5g, h and Supplementary Data 2), suggesting that m⁶A has a pronounced effect on RNA expression. We then examined the effect of m⁶A on the abundance of PSR transcripts. Under +P conditions, only a small proportion of PSI and PSD genes were significantly upregulated (7.1% of PSI and 13.1% of PSD) or downregulated (13.1% of PSI and 4.6% of PSD) by the knockdown of MTA (Supplementary Fig. 5i). However, under −P conditions, MTA knockdown significantly upregulated a large proportion (30.8%) of PSD genes, while significantly downregulating a large proportion (21.4%) of PSI genes (Supplementary Fig. 5j). Further analysis showed that the PSI and PSD transcript levels were comparable between the WT and amiR-mta under +P conditions (Fig. 5g, h); however, under −P conditions, the upregulation of PSI genes and the downregulation of PSD genes were significantly attenuated in amiR-mta compared with the WT (Fig. 5g, h). These RT-qPCR and RNA-seq results provide further evidence that m⁶A facilitates systemic PSR signaling, including both PSI and PSD signaling.

Taken together, the physiological and molecular phenotypes of amiR-mta, with its globally reduced m⁶A levels, demonstrate that the m⁶A modification promotes Pi homeostasis and systemic PSR signaling.

Regulation of transcriptome-wide RNA stability in response to Pi limitation

The finding that the m⁶A RNA modification regulates RNA stability^1,2,3 and that it also alters the transcript abundance of PSR signaling genes (Fig. 5g, h) prompted us to investigate whether the m⁶A modification also facilitates PSR signaling via its effects on RNA stability. We carried out a genome-wide mRNA decay assay using WT and amiR-mta plants under +P and −P conditions. We infiltrated the seedlings with the transcriptional inhibitor cordycepin and collected samples after 0, 15, 30, 60, and 120 min for RNA-seq analysis (Fig. 6a, see “Methods”)⁷⁹. Three biological replicates were sequenced and generated highly reproducible data (Supplementary Fig. 6a). Following data normalization and the exclusion of low-abundance mRNAs, we used transcript levels to model the decay rates for each gene (see Methods)⁷⁹.

**Fig. 6: PSI-m⁶A feedback enhances the RNA stability of PSD genes.**

We first benchmarked the fidelity of our RNA decay data. The previously reported unstable gene AT4G25490⁷⁹ was confirmed to have a short half-life (t_1/2) in our analysis (t_1/2 = 4.44 min under +P and 6.86 min under −P). Next, we randomly chose six genes with short, medium, and long half-lives (two genes each) under both +P and −P conditions and confirmed their RNA stabilities using RT-qPCR (Supplementary Fig. 6b). All tested genes exhibited a similar RNA decay pattern in both transcriptome-wide RNA decay analysis and target-specific RT-qPCR (Supplementary Fig. 6b, c). We then examined the decay rates of PHOSPHATE2 (PHO2) transcripts, which are known to be cleaved by miR399 and are therefore thought to have a shorter half-life under −P than under +P^80,81. Indeed, the PHO2 transcripts exhibited a shorter half-life in the −P samples than in the +P samples (Fig. 6b). These results demonstrate the high fidelity and quality of our transcriptome-wide RNA decay data.

We then evaluated the transcriptome-wide changes in RNA stability mediated by Pi starvation. In the WT, the half-lives of approximate 56% of genes were altered under −P treatment compared with the +P control (Fig. 6c), with approximately 7000 genes having a prolonged half-life and 5000 genes having a shorter half-life under Pi starvation (Fig. 6c and Supplementary Data 3). These results point to an additional layer of PSR regulation through the RNA decay pathway. Of the PSR-signaling genes examined, approximately 54% of PSI genes and 64% of PSD genes showed altered RNA stability under Pi starvation (Fig. 6d). We noticed that more than half (~ 55%) of PSI genes with altered RNA stability had decreased half-lives, whereas the majority (~ 67%) of PSD genes with altered RNA stability had longer half-lives (Fig. 6d). The opposite trends in RNA levels and RNA stability in response to Pi starvation indicated the existence of the feedback regulation of RNA stability for certain PSR genes.

Feedback regulation of PSI-m⁶A promotes the stability of PSD gene transcripts

We next evaluated whether PSI-m⁶A modifications participate in regulating the stability of PSR transcripts. We compared RNA half-life data from WT and amiR-mta plants. Under +P conditions, the RNA half-life of the majority of genes was higher in amiR-mta compared with the WT, with 11,193 genes showing a longer half-life and 2965 genes showing a shorter half-life in the amiR-mta plants (Supplementary Fig. 7 and Supplementary Data 3). These results suggest that m⁶A plays dual roles in RNA stability and that m⁶A modifications mainly destabilize transcripts under normal conditions. Moreover, the half-lives of both m⁶A-methylated and non-m⁶A-methylated transcripts increased globally in amiR-mta under +P conditions, indicating that m⁶A negatively regulates global RNA stability in both direct and indirect manners (Supplementary Fig. 7d).

Unexpectedly, under −P conditions, the RNA half-life did not exhibit a dramatic global increase in amiR-mta compared to the WT (Supplementary Fig. 7a). A detailed analysis showed that 6804 and 5336 genes had increased and decreased half-lives, respectively (Fig. 6e and Supplementary Data 3). The positive effect of m⁶A on RNA stability appears to be enhanced under −P conditions. We examined the RNA stability of the −P-mediated m⁶A-hypermethylated PSR genes. The half-lives of m⁶A-hypermethylated PSD genes (PSI-m⁶A & PSD genes), but not m⁶A-hypermethylated PSI genes (PSI-m⁶A & PSI genes), significantly increased under −P compared to control conditions (+P), indicating that PSI-m⁶A-mediated feedback enhances the RNA stability of m⁶A-hypermethylated PSD transcripts (Fig. 6f). Furthermore, this increased RNA stability of −P-induced m⁶A-hypermethylated PSD genes was repressed in amiR-mta plants (with dramatically reduced m⁶A levels of global and PSI-m⁶A & PSD genes [Supplementary Fig. 8a, b]) compared with the WT under −P but not +P conditions, confirming the positive role of PSI-m⁶A in the RNA stability of PSD genes (Fig. 6g). Moreover, more than half of these m⁶A-hypermethylated PSD genes were also PHR-mediated PSI-m⁶A genes (Fig. 6h). We named these overlapping genes PHR & m⁶A-controlled PSD genes (Supplementary Data 1). The half-lives of the PHR & m⁶A-controlled PSD genes also increased in response to Pi starvation in the WT but not in amiR-mta plants (Fig. 6i), indicating that feedback regulation of the PHRs fine-tunes the expression of some PSD genes via m⁶A-mediated mRNA stabilization.

Taken together, our transcriptome-wide RNA stability profiling of WT and amiR-mta plants under +P and −P conditions, as well as the m⁶A landscape in miR-mta under −P conditions, revealed a new layer of PSR regulation at the level of RNA stability. Moreover, this analysis revealed that PHR-mediated Pi starvation–induced m⁶A feedback extends the RNA half-lives of PSD genes.

PSI-m⁶A does not facilitate RNA stability through APA shifts

It has been reported that m⁶A could affect RNA stability via APA shifts^26,82. To investigate whether PSI-m⁶A facilitates RNA stability via APA shifts, we performed PAS-seq 2⁸³ to profile APA sites on a transcriptome-wide level under +P and −P conditions in WT and amiR-mta seedlings. PCA revealed high reproducibility among three biological replicates of each sample (Supplementary Fig. 9a). Moreover, most poly(A) tags (PATs) aligned to the 3′UTRs of protein-coding genes (~ 70%) (Supplementary Fig. 9b). These results reflect the high quality of our PAS-seq 2 data.

Following previous studies^82,84, we assigned adjacent PATs within a 24-nt window to PAS clusters (PACs). In total, 71,753 PACs (corresponding to 17,968 genes) were detected in WT and amiR-mta under −P and +P conditions. Furthermore, we compared the PAC shift events between the WT and amiR-mta. Under both +P and −P conditions, the knockdown of MTA led to dramatic PAC shifts (Supplementary Fig. 9c). Specifically, 2463 significant distal-to-proximal shifts (hereafter d2p) and 461 significant proximal-to-distal shifts (hereafter p2d) were detected under +P conditions and 1992 significant d2p and 397 significant p2d events were detected under –P conditions in amiR-mta compared with the WT (Supplementary Fig. 9c and Supplementary Data 4). This observation is also similar to the results for the m⁶A reader mutant cpsf30-l (cleavage and polyadenylation specificity factor 30 long-isoform)⁸². Approximately 54% of d2p-shifted genes (total of 1222) and 14% of p2d-shifted genes (total of 657) in cpsf30-l are also shifted in amiR-mta (Supplementary Fig. 9d). These results support the role of m⁶A modifications in regulating APA.

Next, we investigated the −P-induced PAC shifts in WT. We identified 203 significant d2p and 189 significant p2d events under −P compared with +P in the WT, pointing to the potential regulation of APA shifts in response to Pi deficiency (Supplementary Fig. 9c and Supplementary Data 4). Approximately half of these PSR APA-shifted genes were accompanied by changes in RNA stability (Supplementary Fig. 9e). We compared these PSR t_1/2-altered and APA-shifted genes with PSR m⁶A-altered genes. We observed a non-significant overlap of PSI-m⁶A and PSD-m⁶A genes with PSR t_1/2-altered d2p-shifted and p2d-shifted genes (Supplementary Fig. 9f). These overlapping genes only accounted for ~1% of PSI-m⁶A genes, indicating that the effects of PSI-m⁶A modifications on RNA stability are not due mainly to APA shifts (Supplementary Fig. 9f). Meanwhile, the APA of only 18.5% of PSI-m⁶A & PSD genes was altered in amiR-mta compared with WT under −P conditions (Supplementary Fig. 9g). Moreover, APA shift and IGV analysis showed that the APA of PHR1 transcripts was not obviously shifted in the WT or amiR-mta under +P or −P conditions (Supplementary Fig. 9h).

Our results showed a potential regulation in APA shifts in response to Pi deficiency, but the PSI-m⁶A-induced increase in RNA stability is not due to shifts in APA.

PSI-m⁶A modification stabilizes PHR1 mRNA to facilitate PSI signaling

We noticed that the global half-life of the m⁶A-hypermethylated PSI transcripts was unaffected by either Pi starvation or the knockdown of MTA (Fig. 6f and Supplementary Fig. 10), whereas the global expression of PSI genes was reduced in amiR-mta (Fig. 5g). Given that PSI-m⁶A modifications occurred on the transcripts of PHR1, encoding a master regulator of PSI signaling (Fig. 2g–i) and that its half-life and m⁶A level was reduced in amiR-mta compared with the WT under −P (Fig. 7a and Supplementary Fig. 8c), we hypothesized that PSI-m⁶A stabilizes PHR1 mRNA to facilitate the expression of PSI genes under −P conditions.

**Fig. 7: PSI-m⁶A stabilizes *PHR1* transcripts to facilitate PSI signaling.**

To explore this hypothesis, we mutated putative m⁶A modification loci (mutation from A to U) in the 3′ UTR of PHR1 and expressed this mutant gene in the phr1 null mutant background (P_PHR1:gPHR1-FLAG-4XMYC-3UTR-MUT; phr1, hereafter PHR1-MUT), with the WT variant serving as the control (P_PHR1:gPHR1-FLAG-4XMYC-3UTR-WT; phr1, hereafter PHR1-WT) (Fig. 7b). The 3′ UTR of PHR1 contains seven putative m⁶A loci, including one RRACH motif, two plant-specific UGUA motif, and four newly identified CAGR motifs (Fig. 7b and Supplementary Fig. 1e). UGUA and CAGR motifs are also enriched in 3′UTR-located PSI-m⁶A genes (Supplementary Fig. 2d). Six of these seven m⁶A loci were confirmed by a SELECT kit (Supplementary Fig. 11a)⁸⁵. This is consistent with the results of single base–resolved m⁶A profiling, which revealed ~5.6 m⁶A loci per peak, and the fact that m⁶A loci tend to form clusters in short regions²⁴. This m⁶A cluster at the 3′ UTR of PHR1 also appears to be strongly regulated by m⁶A modification dynamics. The m⁶A modifications are indeed abolished by mutagenesis in the PHR1-MUT transgenic plants (Supplementary Fig. 11b). In addition, the A-to-U mutations did not affect the RNA structure, as demonstrated by the similar minimal free energy of PHR1-WT (ΔG = −102.36 kcal/mol) and PHR1-MUT (ΔG = −106.66 kcal/mol). Considering the variation in the expression of PHR1-FLAG-4XMYC (PHR1-FM) in the transgenic plants due to the different copy numbers and insertion sites of the vector, we randomly chose three lines for each transgene for further analysis.

Compared with PHR1-WT, the average steady-state level of PHR1-FM expression of the three lines was significantly reduced in PHR1-MUT plants, but that of the control gene Basta was not (Fig. 7c), indicating that m⁶A enhances the stability of PHR1 transcripts. Consistent with the expression of PHR1-FM (Fig. 7c), transformation with PHR-WT, but not PHR1-MUT, fully rescued the phr1 null mutant, which accumulates less anthocyanin under −P conditions than the WT⁴⁹. Indeed, the PHR1-MUT and phr1 plants exhibited greener leaves than the WT and PHR1-WT complementation lines under Pi starvation (Fig. 7d). The average anthocyanin content of the three lines was significantly higher in PHR1-WT than in PHR1-MUT under −P stress (Fig. 7e and Supplementary Fig. 11c). Furthermore, like the phr1 null mutant, the average expression levels of two PSI genes downstream of PHR1 were also significantly reduced in PHR1-MUT vs. PHR1-WT seedlings (Fig. 7f and Supplementary Fig. 11d). Finally, we carried out target-specific RNA half-life measurements using RT-qPCR to test the role of m⁶A in the stability of PHR1 mRNA. The RNA half-life of PHR1-FM was significantly reduced in PHR1-MUT compared to PHR1-WT under both +P and −P conditions (Fig. 7g and Supplementary Fig. 11e). The similar RNA decay pattern of the control gene was indicative of successful cordycepin treatment of both PHR1-WT and PHR1-MUT materials (Supplementary Fig. 11f). Taken together, these results provide evidence that m⁶A enhances the stability of PHR1 transcripts to facilitate PSI signaling.

In summary, the results of physiological and molecular analyses of m⁶A-mutated transgenic plants demonstrate that m⁶A enhances the stability of PHR1 transcripts to facilitate PSI signaling.

Discussion

There is mounting evidence that widespread m⁶A modifications play critical roles in the post-transcriptional regulation of RNA metabolism and physiological processes^1,2,3. Plants are frequently exposed to nutrient-deficient stress, leading to widespread changes in gene expression^39,40. However, the regulatory function of m⁶A in responses to these physiological stresses remains elusive. In this study, we examined the m⁶A-methylome, the transcriptome, genome-wide RNA decay, global profiling of APA sites, PHR1-MTA/MTB protein interactions, and the effects of mutating m⁶A loci to dissect the role of m⁶A-mediated regulation of RNA stability in facilitating systemic PSR signaling and Pi homeostasis under Pi-deficient stress. We demonstrated that PHR-mediated PSI-m⁶A modifications not only promote stability of some PSD transcripts, but also stabilize PHR1 mRNA to facilitate the expression of PSI genes via a feedback loop (Fig. 7h).

PHR1 and its family members are central TFs in the transcriptional regulation of the PSR^49,53. Here, we showed that PHRs mediate epitranscriptomic regulation during Pi homeostasis and the PSR. We conducted m⁶A-seq with systematic calibration using synthetic modification-free RNA libraries to remove false-positive peaks and obtained high-confidence m⁶A landscapes under normal and Pi-deficient stress conditions (Fig. 1). Pi starvation induced widespread m⁶A modifications, many of which were added to Pi homeostasis- and PSR-related transcripts (Fig. 2). We also showed that PHR proteins mediate PSI-m⁶A, as these global modifications were diminished in the phr1 phl1 double mutant under −P conditions (Fig. 3). However, the PHRs do not transcriptionally regulate the expression of either the PSI-m⁶A-marked mRNAs or the known m⁶A methyltransferases or demethylases (Supplementary Fig. 4), raising the interesting question of how PHRs mediate PSI-m⁶A modifications. Blue light induces the liquid–liquid phase separation of CRYPTOCHROME2 (CRY2) to co-condense the m⁶A writer proteins MTA, MTB, and FIP37, thereby promoting the addition of m⁶A to transcripts of the central circadian clock oscillator genes⁷⁶. CRY2 also co-condenses another m⁶A writer protein, FIO1, to promote the methylation of transcripts encoding multiple regulators of chlorophyll homeostasis⁷⁷. In addition, the histone modification H3K36me3 can guide co-transcriptional m⁶A modifications in mammals by interacting with METTL14⁸⁶. Analogously, H3K36me2 is highly correlated with m⁶A modifications in Arabidopsis⁸⁷. Although a limitation exists in that the interaction between PHR1 and MTA/MTB in plants under +P conditions remains uninvestigated, we observed that PHR1 interacts with MTA and MTB in nuclei under −P conditions (Fig. 4). It is plausible that PHRs might act as “m⁶A recruiters” to guide m⁶A writers to certain transcripts, facilitating PSR-related m⁶A methylation upon Pi starvation. These findings expand our understanding of the regulatory functions of PHRs beyond the transcriptional level to the epitranscriptional level.

The m⁶A modifications regulate almost all aspects of mRNA metabolism, including transcription, pre-mRNA processing, folding, translation, and decay^1,2,3. To date, the best-characterized role of m⁶A in various plant species is its influence on mRNA stability, which affects multiple developmental processes and stress responses^2,3,88. m⁶A exerts positive or negative effects on specific transcripts under certain distinct conditions^3,88. In some studies, the combination of m⁶A-seq and RNA-seq led to apparently paradoxical conclusions about the effect of m⁶A on RNA abundance^3,27,36,89; however, these studies did not directly and globally investigate the relationship between mRNA half-life and m⁶A modifications in plants. We performed transcriptome-wide RNA decay assays in the WT and a line in which the m⁶A-writer gene MTA is suppressed (amiR-mta) under normal and Pi-deficient stress conditions (Fig. 6 and Supplementary Fig. 7). Intriguingly, our results exhibited a distinct effect of m⁶A modifications on overall mRNA stability, in which m⁶A mainly destabilizes transcripts under normal growth conditions but stabilizes them under Pi-deficient conditions (Fig. 6 and Supplementary Fig. 7). Our results confirm the dual effects of m⁶A on mRNA stability, indicating that m⁶A-mediated RNA decay is dependent on the environment, as exemplified by +P and −P conditions.

We showed that PSI-m⁶A modifications mediate the regulation of RNA stability in response to Pi starvation. The effects of PSR were previously not well understood at the RNA level; however, our genome-wide RNA decay assay revealed the previously uncharacterized regulation of RNA stability under Pi-deficient stress, in which the half-lives of more than half of transcripts are altered by Pi starvation (Fig. 6c). Approximately 2000 PSD genes show decreased expression levels under Pi starvation (Supplementary Fig. 2e)⁵³, but it is unclear how this is regulated. Our results show that PHR-controlled PSI-m⁶A fine-tunes the expression of these PSD genes by stabilizing their transcripts (Figs. 5 and 6). Moreover, we demonstrated that m⁶A modifications stabilize PHR1 mRNA to facilitate PSI signaling (Fig. 7). Although the steady-state level of PHR1 transcript is comparable or moderately increased by −P⁴⁹, the mechanism by which its expression is maintained had been unknown. Our findings provide evidence for the post-transcriptional stabilization of PHR1 mRNA by m⁶A upon Pi starvation, promoting the downstream signaling pathway.

The above results raise an interesting question about the mechanism of PSI-m⁶A-mediated RNA decay under −P conditions. m⁶A reader proteins bind to the m⁶A modification and determine the fate of m⁶A-methylated transcripts in diverse RNA metabolic pathways. In mammals, the binding of the m⁶A reader YTHDF2 to m⁶A-modified transcripts promotes their decay^90,91, while the binding of other m⁶A readers (IGF2BP proteins) stabilizes m⁶A-modified transcripts⁹². In Arabidopsis, the m⁶A reader EVOLUTIONARILY CONSERVED C-TERMINAL REGION2 (ECT2) binds to and promotes the stability of transcripts related to trichome morphogenesis⁹³. The m⁶A reader CPSF30-L regulate the alternative polyadenylation of some nitrate signaling–related gene transcripts⁸⁴. First, our PAS-seq 2 data showed that the impact of PSI-m⁶A did not occur through APA shifts (Supplementary Fig. 9). Second, we did not find that PHR1 was the target of ECT2 in published FA-CLIP-seq data⁹³. In addition, a preprint manuscript indicates that the abundance of the m⁶A reader ECT4 is reduced under Pi-deficiency stress⁹⁴. We further detected the phenotype of ect2/3/4 under −P stress. Consist with a recent study⁹⁵, the primary root length was reduced in ect2/3/4 compared with the WT under −P treatment (Supplementary Fig. 12a, b). However, the fresh weight and Pi concentration in shoots were comparable between the WT and ect2/3/4 under −P conditions (Supplementary Fig. 12c, d). As expected, PHR1 expression and mRNA stability were unchanged in ect2/3/4 versus the WT under −P conditions (Supplementary Fig. 12e, f). These results indicated that m⁶A readers ECT2/3/4 do not participate in PSI-m⁶A-mediated PHR1 stabilization.

Our results suggest that unknown PSR m⁶A reader(s) modulate PSI-m⁶A-mediated mRNA stabilization. It would therefore be interesting to systematically investigate the activities of m⁶A readers upon Pi starvation.

In conclusion, we analyzed the m⁶A-methylome, transcriptome, PHR1-MTA/MTB protein interactions, global APA site profiling, and mRNA decay landscape (validated by m⁶A-loci mutagenesis) to systematically investigate the role of the m⁶A modification in the PSR. Our findings demonstrate that PHR-mediated stress-induced m⁶A modifications play a crucial role in maintaining Pi homeostasis and enhancing systemic PSR signaling by stabilizing PSR-related transcripts. These findings uncover how plants respond to nutrition-deficient stress at the epitranscriptome level.

Methods

Plant materials and growth conditions

Seeds of Arabidopsis (Arabidopsis thaliana) accession Columbia (Col-0) and the phr1 (SALK_067629C), phl1 (CS832612), and ect2/3/4 (CS2110132) T-DNA null mutants were obtained from the Arabidopsis Biological Resource Center (https://abrc.osu.edu/). The phr1 phl1 double mutant was obtained by crossing phr1 with the phl1 single mutant. To generate amiR-mta plants, Col-0 plants were transformed with the binary vector pBA-P_35S:amiR-mta by the floral dip method⁹⁶. Similarly, for P_PHR1:gPHR1-FLAG-4XMYC-3UTR-WT; phr1 and P_PHR1:gPHR1-FLAG-4XMYC-3UTR-MUT; phr1 plants, homozygous phr1 mutants were transformed with the binary vectors pBA002a-P_PHR1:gPHR1-FLAG-4XMYC-3UTR-WT or pBA002a-P_PHR1:gPHR1-FLAG-4XMYC-3UTR-MUT by the floral dip method⁹⁶.

The wild type (Col-0), mutants, and transgenic lines were grown under a 12-h-light/12-h-dark photoperiod at 22 °C at a relative humidity of ~50%. For Pi starvation treatment, seeds were germinated and grown on Murashige and Skoog (MS) agar medium supplemented with 1% sucrose for seven days. The seven-day-old plants were transferred from MS agar medium to fresh MS agar medium containing 1.25 mM Pi (+P) or 6.25 μM Pi (−P) and incubated for seven more days.

Vector construction

Most coding and genomic sequences were cloned into the pENTR/D-TOPO (ThermoFisher) vector and confirmed by Sanger sequencing. Most plant constructs were generated using the Gateway cloning system (ThermoFisher). The primers for all constructs are listed in Supplementary Data 5.

The pBA-P_35S:amiR-mta plasmid was constructed as follows: First, the sequence of artificial miRNA used to silence MTA (amiR-mta) was designed using the WMD3 website (http://wmd3.weigelworld.org)⁹⁷. The amiR-mta sequence was introduced into a pENTER vector containing the primary-miR159a (pri-miR159a) stem-loop sequence by PCR to obtain pENTR-pri-amiR-mta-159a. Finally, pri-amiR-mta-159a was transferred into pBA-P_35S:DC to yield pBA-P_35S:amiR-mta.

pBA002a-P_PHR1:gPHR1-FLAG-4XMYC-3UTR-WT and pBA002a-P_PHR1: gPHR1-FLAG-4XMYC-3UTR-MUT were constructed as follows: First, an approximately 2.4 kb promoter and 5′ UTR fragment of PHR1 was amplified from the Col-0 genome and cloned into AscI/XbaI-digested pBA002a-DC-FLAG-4XMYC (FM) to generate pBA002a-P_PHR1:DC-FM vector using MultiF Seamless Assembly Mix (ABclonal). The PHR1 genomic fragment without the stop codon was amplified from the Col-0 genome and cloned into the pENTR/D vector to create the pENTR-gPHR1 vector. gPHR1 was transferred into pBA002a-P_PHR1:DC-FM to yield the pBA002a-P_PHR1:gPHR1-FM vector. Finally, a 663 bp fragment of wild-type PHR1 after the stop codon was amplified from the Col-0 genome and cloned into the SacI-digested pBA002a-P_PHR1:gPHR1-FM vector using MultiF Seamless Assembly Mix (ABclonal) to yield pBA002a-P_PHR1:gPHR1-FLAG-4XMYC-3UTR-WT. The corresponding 3′ UTR of PHR1 with the m⁶A-loci mutated was synthesized and cloned into the SacI-digested pBA002a-P_PHR1:gPHR1-FM vector using MultiF Seamless Assembly Mix (ABclonal) to yield pBA002a-P_PHR1:gPHR1-FLAG-4XMYC-3UTR-MUT.

Y2H vectors were constructed as follows: full-length CDSs of PHR1, MTA, MTB, FIP37, SPX1, and SPX2 were amplified and cloned into the NotI/AscI-digested pENTR/D-TOPO vector using MultiF Seamless Assembly Mix (ABclonal) and confirmed by sequencing. The CDSs were inserted into pGADT7-GW and pGBKT7-GW vectors to generate AD-gene and BD-gene vectors by LR reaction, respectively.

The PHR1-YN BiFC vector was constructed as follows: the full-length CDS of PHR1 was amplified and cloned into SalI/SpeI-digested pCAMBIA-35S-nYFP vector to generate pCAMBIA-35S-PHR1-nYFP (PHR1-YN) using MultiF Seamless Assembly Mix (ABclonal).

Synthesis of modification-free IVT RNA

Modification-free IVT RNA was synthesized as previously described²². The oligonucleotides were synthesized at Genscript and are listed in Supplementary Data 5.

Total RNA was extracted from the corresponding Pi starvation-treated or control samples using TRIzol (ThermoFisher). Cellular mRNA was isolated from the samples using Dynabeads Oligo(dT)₂₅ (ThermoFisher) according to the manufacturer’s instructions. For first-strand cDNA synthesis, 100 ng of mRNA was hybridized with 1 µl of poly(T) primer (12 µM) at 75 °C for 3 min and 42 °C for 2 min. Subsequently, a mixture containing 2 µl of 5× first-strand buffer for SMARTScribe Reverse Transcriptase (Takara), 0.5 µl of 100 mM DTT, 0.5 µl of SUPERase-In RNase Inhibitor (ThermoFisher), 1 µl of TSO T7 primer (10 µM), 1 µl of 50× Advantage UltraPure PCR Deoxynucleotide Mix (Takara), and 1 µl of SMARTScribe Reverse Transcriptase (Takara) was added to the mRNA–poly(T) primer mixture. The reaction was incubated at 42 °C for 1.5 h and stopped by heating at 68 °C for 10 min.

For double-stranded cDNA synthesis, following the addition of 10 µl of 10× Advantage 2 PCR buffer (Takara), 1 µl of Advantage 2 PCR Polymerase Mix (Takara), 2 µl of 50× Advantage UltraPure PCR Deoxynucleotide Mix (Takara), 2 µl of T7 extension primer (10 µM), 2 µl of RNase H (NEB), and RNase-free water to a final volume of 100 µl, the primer extension reaction was conducted at 37 °C for 15 min, 95 °C for 2 min, 60 °C for 1 min, and 68 °C for 10 min. Double-stranded cDNA was purified using the AMPure XP beads (Beckman).

For T7 RNA polymerase-mediated transcription, in vitro transcription was performed at 37 °C for 2 h using a HiScribe T7 High Yield RNA Synthesis kit (NEB) with modification-free NTPs from Takara Bio (Cat# 6140). The reaction mixture was treated with TURBO DNase (ThermoFisher) at 37 °C for 30 min to remove the dsDNA and was purified using RNAClean XP beads (Beckman).

LC-MS/MS quantification of m⁶A

To quantify the contents of m⁶A RNA modifications in cellular mRNA from WT, amiR-mta plants, and IVT RNA, 100 ng of cellular mRNA or IVT RNA was digested with 1U Nuclease P1 (NEB) at 37 °C for 4 h, followed by the addition of 1U Shrimp Alkaline Phosphatase (NEB) and 100 mM MES (pH 6.5) and incubation at 37 °C for 4 h. The digested samples were centrifuged at 15,000 × g for 5 min, and the supernatants were subjected to LC-MS/MS analysis at the Institute of Botany, Chinese Academy of Sciences. The detailed method used for the LC-MS/MS analysis was described previously³⁶.

m⁶A-seq

m⁶A-seq was performed as previously described⁹⁸. In brief, approximately 50 µg of total RNA was extracted from Pi starvation-treated or untreated 14-day-old WT and phr1 phl1 seedlings using TRIzol reagent (ThermoFisher). Cellular mRNA was isolated from the samples using Dynabeads Oligo(dT)₂₅ (ThermoFisher) according to the manufacturer’s instructions. The cellular mRNA or IVT RNA was fragmented to approximately 100 nucleotide (nt) segments with a Magnesium RNA fragmentation module (NEB) and purified by ethanol precipitation. Ten percent of the fragmented RNA was kept as input. The concentration of the fragmented RNA was measured by Qubit (ThermoFisher).

m⁶A immunoprecipitation (IP) was performed using commercial anti-m⁶A antibody from NEB (Cat# E1610S). Fragmented RNA (200 ng) was incubated with 1 µl of anti-m⁶A antibody in 200 µl of IP buffer (10 mM Tris-HCl pH 7.5, 150 mM NaCl, 0.1% v/v Igepal [CA-6300] and 0.2 U/µl SUPERase-In RNase Inhibitor [ThermoFisher] in nuclease-free water) at 4 °C for 2 h on an orbital rotator. After adding 25 µl of Dynabeads Protein G magnetic beads (ThermoFisher), the m⁶A-IP mixtures were incubated at 4 °C for 2 h. The beads were washed twice successively with low-salt buffer (50 mM NaCl, 10 mM Tris-HCl pH 7.5, 0.1% Igepal CA-630 in nuclease-free water) and high-salt buffer (500 mM NaCl, 10 mM Tris-HCl pH 7.5, 0.1% Igepal CA-630 in nuclease-free water).

RNA eluted from the beads and non-IP fragments (input control) were used to prepare libraries with a Stranded mRNA-seq Lib Prep Module for Illumina (ABclonal). The libraries of two biological repeats for each sample were sequenced as 150-bp paired-end reads on a NovaSeq 6000 instrument at Novogene.

m⁶A-seq data analysis

Approximately 23 million paired-end 150 bp (PE150) reads were obtained for each sample. Raw reads were subjected to adapter trimming and filtering of low-quality reads (quality <20) using Cutadapt (v4.1) software⁹⁹. Basic quality control was conducted using FastQC (v0.11.9) (www.bioinformatics.babraham.ac.uk/projects/fastqc/), and reads <15 bp long were discarded. Quality-filtered reads were mapped to the Arabidopsis genome (TAIR10) using HISTA2 (v2.2.1)¹⁰⁰ with the parameter ‘–rna-strandness RF’. Mapped reads were visualized in Integrative Genomics Viewer (IGV)¹⁰¹ after RPKM (Reads Per Kilobase per Million mapped reads) normalization and used for the following analyses.

First, m⁶A peaks were called using the R package exomePeak (v2.16.0) (bioconductor.org/packages//2.13/bioc/html/exomePeak.html) using the enrichment criteria (IP/Input) ≥ 2 and FDR < 0.05. Systematic calibration using modification-free IVT RNA libraries was then performed as previously described²². Briefly, false-positive peaks were defined as the overlapping/common peaks sharing more than 50% fractions of peak regions in the endogenous mRNA library and modification-free IVT RNA library. BEDtools (v2.30.0) intersect¹⁰² was used to find the overlapping peaks. The calibrated high-confidence m⁶A peaks were defined as peaks that were specifically identified in both biological replicates in mRNA m⁶A-seq libraries but not in IVT RNA m⁶A libraries. The differential m⁶A peaks were then identified using exomePeak based on the criteria of fold-change ≥ 1.5 and FDR < 0.05. The differential m⁶A peaks that overlapped with known false-positive m⁶A peaks were removed. Peak annotation was performed using BEDtools (v2.30.0)¹⁰². HOMER (v4.11) software (homer.ucsd.edu/homer/motif) was used to identify m⁶A peak motifs in the calibrated high-confident m⁶A peaks and peak motifs in peaks from the IVT RNA libraries.

RNA-seq data analysis

The non-m⁶A-IP control/Input RNA-seq data were used for transcriptome analysis of WT and phr1 phl1 plants under +P and −P conditions. The RNA-seq data from the RNA decay assay at 0 min were used for transcriptome analysis of WT and amiR-mta plants. Approximately 21 to 34 million paired-end 150 bp (PE150) reads were obtained for each sample.

Raw reads were subjected to adapter trimming and filtering of low-quality reads (quality <20) using Cutadapt (v4.1) software⁹⁹. Basic quality control was conducted using FastQC (v0.11.9) (www.bioinformatics.babraham.ac.uk/projects/fastqc/). Filtered reads were mapped to the Arabidopsis genome (TAIR10) using HISTA2 (v2.2.1)¹⁰⁰. Only uniquely mapped reads were retained for read counting and differential gene expression analysis. The read counts for each gene were calculated using featureCounts¹⁰³ and normalized using DESeq2 (v1.38)¹⁰⁴. The differentially expressed genes were identified using DESeq2 (v1.38)¹⁰⁴.

Y2H assay

Y2H assay was performed as previously described¹⁰⁵. The Matchmaker Gold Y2H system (Clontech) was used. Different combinations of constructs were co-transformed into the Y2H Gold yeast strain (Clontech) and grown on synthetic defined medium plates under a screening condition. For each combination, a minimum of ten independent colonies were subjected to interaction testing. The corresponding colonies were picked up and spotted onto the selective plates in 1:10 serial dilutions for picture recording.

Bimolecular fluorescence complementation assay (BiFC) and fluorescence microscopy

The BiFC experiments were performed as previously described⁷⁵. All BiFC constructs were transformed into Agrobacterium tumefaciens strain GV3101. The resulting positive Agrobacterium colonies harboring each construct were cultured in Luria-Bertani medium containing 100 μg/mL kanamycin and 40 μM acetosyringone at 28 °C to an optical density at 600 nm (OD₆₀₀) = 1.0. Then, the Agrobacterium cells were resuspended with the infiltration buffer (10 mM MgCl₂ and 150 μM acetosyringone) and incubated at room temperature for 3 h without shaking. Next, equal volumes of the corresponding Agrobacterium cell suspensions were mixed and co-infiltrated into the leaves of 4-week-old N. benthamiana. The fluorescence signals (YFP emission at 514 nm) were observed at 48 h after infiltration with a Zeiss LSM900 confocal laser scanning microscope.

Nuclei isolation

Nuclei isolation was performed as described previously⁷⁴. In brief, ~1 g of ground powder was dissolved in 4 ml lysis buffer (20 mM Tris-HCl pH 7.5, 20 mM KCl, 2.5 mM MgCl₂, 25% [v/v] glycerol, 250 mM sucrose, 5 mM DTT, and 1× proteinase inhibitor cocktail without EDTA [Roche]). The homogenate was filtered through a double layer of Miracloth. The flow-through was centrifuged at 1500 × g for 10 min at 4 °C. After discarding the supernatant, the pellet was washed three times with 5 ml nuclear resuspension buffer (20 mM Tris-HCl pH 7.5, 2.5 mM MgCl₂, 25% [v/v] glycerol, and 0.2% [v/v] Triton X-100). The pellet was re-suspended in 500 μl of nuclear resuspension buffer 2 (20 mM Tris-HCl pH 7.5, 2.5 mM MgCl₂, 25% [v/v] glycerol, 250 mM sucrose, 5 mM β-ME, 0.5% [v/v] Triton X-100, and 1× proteinase inhibitor cocktail without EDTA [Roche]) and transferred to 500 μl nuclear resuspension buffer 3 (20 mM Tris-HCl pH 7.5, 10 mM MgCl₂, 1.7 M sucrose, 5 mM β-ME, 0.5% [v/v] Triton X-100, and 1× proteinase inhibitor cocktail without EDTA [Roche]). After centrifugation at 10,000 × g for 60 min at 4 °C, the isolated nuclei were subjected to immunoblotting or co-immunoprecipitation (Co-IP).

Co-IP

Co-IP was performed as described with some modifications^59,105. The isolated nuclei of PHR1-WT were lysed in 0.9 ml IP buffer (40 mM Tris-HCl pH 7.5, 5 mM MgCl₂, 150 mM NaCl, 100 nM ZnCl₂, 1 mM DTT, 1% [v/v] Triton X-100, 5% [v/v] glycerol, 1 mM PMSF, 50 μM NaCl, and 1× proteinase inhibitor cocktail without EDTA [Roche]). Then, the lysate was treated with Turbo DNase (ThermoFisher) for 2 h at 4 °C. After centrifugation at 10,000 × g for 10 min at 4 °C, the supernatants were transferred to a new 2 ml tube. Five percent of the supernatant served as input, and the remaining was immunoprecipitated with the anti-Myc antibody (Sigma, Cat# C3956) for 2 h at 4 °C. Next, Dynabeads protein G Magnetic beads (ThermoFisher) were used to pull down the PHR1 complexes. The beads with PHR1 complexes were washed four times with IP buffer before application of SDS loading buffer at 95 °C for 10 min.

Immunoblot analysis

Immunoblot analysis was performed as described¹⁰⁵. The ground powders of materials, isolated nuclei and immunoprecipitated beads were mixed with SDS loading buffer (80 mM Tris-HCl pH 6.8, 2% SDS, 10% glycerol, 0.1 M DTT, and 0.005% Bromophenol blue) and then incubated at 95 °C for 10 min. Next, the extracted proteins in the supernatant were subjected to immunoblots. The blots were probed with primary antibodies against Myc (Sigma, Cat# C3956, IB 1:3000), MTA (homemade from Z. Zhang’s group, IB 1:1000)^74,75, MTB (homemade from Z. Zhang’s group, IB 1:1000)^74,75, AGO1 (Agrisera, Cat# AS09 527, IB 1:2000), and Histone H3 (ABclonal, Cat# A17562, IB 1:5000). The secondary antibody was the HRP-conjugated goat anti-rabbit IgG (ThermoFisher, Cat# 31460).

PAS-seq 2 and data analysis

Poly(A) sequencing (PAS-seq) 2 was performed as described previously⁸³. In brief, total RNA was extracted from 14-day-old Arabidopsis seedlings using TRIzol™ reagent (ThermoFisher) according to the manufacturer’s instructions. Poly(A) RNA was enriched using Dynabeads Oligo(dT)₂₅ beads (ThermoFisher). The enriched RNA was fragmented using RNA Fragmentation Reagents (ThermoFisher, AM8740) at 70 °C for 5 min. Next, fragmented mRNAs were subjected to first-strand cDNA synthesis by SuperScript III reverse transcriptase (ThermoFisher) with the PAS-seq 2 oligo(dT) primer (GTGACTGGAGTTCCTTGGCACCCGAGAATTCCATTTTTTTTTTTTTTTTTTTTV, where V represents A/C/G) and a template-switch oligo (TSO) containing locked nucleic acid (LNA) (CTACACGACGCTCTTCCGATCTCATrGrG+G). The cDNA was purified with AMPure XP beads (Beckman) and amplified with Illumina TruSeq Universal adapter and RNA PCR indexed primer. Amplified products were gel extracted to isolate fragments of 185–225 bp. Purified fragments were sequenced on an Illumina sequencing platform.

Data analysis was performed as described in previous studies⁸⁴. In brief, raw reads were trimmed to remove adapters and filtered for low-quality bases (Q < 20) using Cutadapt (v4.1). The filtered reads were aligned to the Arabidopsis thaliana genome (TAIR10) using STAR (v2.7.10). Internal priming artifacts, defined as six or more continuous As or seven or more As in a 10-nt window downstream from the poly(A) junction, were discarded. Next, polyadenylation sites (PASs) located within a 24-nt window and with TPM ≥ 3 in at least two biological replicates were clustered into a PAS cluster (PAC). The PAS shift events between indicated samples were compared with Fisher’s exact test. A pair of PAS shifts with FDR < 0.01 was considered significant. For genes with more than two significantly shifted PACs, the two most dominant differential PACs were selected to represent the shift orientation of the given genes.

Gene Ontology (GO) term analysis

GO analysis was performed with TBtools (v2.019)¹⁰⁶.

Transcriptome-wide RNA decay analysis

The analysis of transcriptome-wide RNA decay was performed as described⁷⁹ (bioconductor.org/packages/release/bioc/html/RNAdecay.html) with some modifications. Pi starvation-treated (−P) or untreated control (+P) two-week-old Col-0 (WT) and amiR-mta seedlings were harvested and placed in incubation buffer containing 15 mM sucrose, 1 mM KCl, 1 mM PIPES pH 6.25, and 1 mM sodium citrate. The transcriptional inhibitor cordycepin was then added to the incubation buffer to a final concentration of 1 mM. Vacuum infiltration (~ 12 psi) was conducted for 1 min to facilitate sample penetration. Initial (0 min, T₀) samples were immediately harvested and flash-frozen in liquid nitrogen. Then, vacuum infiltration (~ 12 psi) was performed twice for 1 min each time. Subsequently, samples were collected at 15, 30, 60, and 120 min following the first vacuum release, flash-frozen in liquid nitrogen, and ground to a powder. Three biological replicates were conducted for each time point and each material. Total RNA was extracted from the samples using TRIzol (ThermoFisher) and used to prepare libraries with a Stranded mRNA-seq Lib Prep Module for Illumina (ABclonal). The libraries were sequenced as 150-bp paired-end reads on an NovaSeq 6000 instrument at Annoroad.

The raw sequencing reads from the RNA decay RNA-seq were subjected to adapter trimming and filtering of low-quality reads (quality <20) using Cutadapt (v4.1)⁹⁹. Clean reads were mapped to the Arabidopsis genome (TAIR10) using HISTA2 (v2.2.1)¹⁰⁰. Only uniquely mapped reads were retained for the following analyses. The featureCounts¹⁰³ was used to obtain raw read counts for all annotated genes. PCA was performed with plotPCA from the DESeq2 R package¹⁰⁴. Normalization and modeling of RNA decay data were carried out using the package ‘RNAdecay’ (bioconductor.org/packages/release/bioc/html/RNAdecay.html). Specifically, 30 stably and highly expressed genes in WT and amiR-mta under +P and −P conditions (“AT1G44575”, “AT1G30380”, “AT5G42980”, “AT5G65220”, “AT2G20260”, “AT4G01310”, “AT2G37660”, “AT3G01480”, “AT5G08650”, “ATCG00800”, “AT3G02730”, “AT5G59613”, “AT5G38520”, “AT3G11930”, “AT5G41700”, “AT5G03880”, “AT1G76200”, “AT2G44650”, “AT2G42210”, “AT5G16710”, “AT1G02920”, “AT4G20150”, “AT1G21065”, “AT2G21960”, “AT1G20850”, “AT2G24940”, “AT1G66670”, “AT1G15120”, “AT5G16400”, “AT3G12260”) were chosen for data normalization. ComplexHeatmap¹⁰⁷ was employed to draw the RNA decay heatmap, which was constructed using the mean relative abundance and half-life of each transcript (Supplementary Data 3).

Measurement of target-specific RNA decay

Cordycepin treatment was performed as described above (for the transcriptome-wide RNA decay assay). Seedlings were collected at 0 min, 1 h, and 4 h following the first vacuum release, frozen in liquid nitrogen, and ground to a powder. Total RNA was extracted from the samples for reverse transcription and quantification of the expression levels of various genes by quantitative reverse-transcription PCR. EF1α (AT1G07940) was used as an internal control. All primers used are listed in Supplementary Data 5.

Validation of m⁶A sites in target mRNAs

The PHR1 m⁶A sites were validated using the SELECT kit (EPIBIOTECH, Cat. No: R202106M-0224T)⁸⁵. Briefly, 1 μg total RNA of each sample was equally divided into two fractions: one was treated with the m⁶A demethylase FTO at 37 °C for 30 min, and the other was treated with FTO and EDTA under the same conditions. Probes were designed to target the region flanking the tested m⁶A sites for annealing and extension. The extended products were ligated using SELECT^TM ligase, followed by qPCR to validate the m⁶A modification. The probes and primers used are listed in Supplementary Data 5.

Gene-specific m⁶A-IP-qPCR

The gene-specific m⁶A-IP-qPCR was performed as described in a previous study¹⁹. In brief, approximately 200 ng isolated mRNAs was fragmented to 100–150 nt using the Magnesium RNA Fragmentation Module (NEB) and purified by ethanol precipitation. The commercial m⁶A probes (NEB, Cat# E1610S) and non-m⁶A probes (NEB, Cat# E1610S) were mixed into the eluted fragmented RNAs as positive and negative controls, respectively. Ten percent of the RNA mixture served as input sample, and the remaining was incubated with 1 µl anti-m⁶A antibody (NEB, Cat# E1610S) in 200 µl of IP buffer (10 mM Tris-HCl pH 7.5, 150 mM NaCl, 0.1% v/v Igepal [CA-6300], and 0.2 U/µl SUPERase-In RNase Inhibitor [ThermoFisher] in nuclease-free water) at 4 °C for 2 h on an orbital rotator. The m⁶A-containing RNA fragments were pulled down and eluted according to the above m⁶A-seq procedure. Both input and IP samples were analyzed by RT-qPCR using primers listed in Supplementary Data 5. The enrichment of m⁶A in each sample was calculated by normalizing the value of the amplification cycle (Cq) of the IP to that of the corresponding input. The relative enrichment of each m⁶A region in samples was normalized to that of PHR1-WT, where the ratio was arbitrarily set to 1.

Quantitative reverse-transcription PCR

Total RNA was extracted from two-week-old Pi starvation-treated seedlings using TRIzol reagent (ThermoFisher). RNA quality was checked by formaldehyde gel electrophoresis. The total RNA was treated with TURBO DNase (ThermoFisher), and 1 µg of DNase-treated RNA was subjected to reverse transcription with SuperScript II. Quantitative real-time PCR (qPCR) was performed using SYBR Green Master Mix (ThermoFisher). The 2^−ΔΔCT method was used to calculate gene expression levels. EF1α (AT1G07940) was used as an internal control. All primers used are listed in Supplementary Data 5.

Measuring phosphate (Pi) content

Measurements of Pi concentrations in plants were carried out as previously described⁵⁰. In brief, 10–20 mg of fresh plant samples were placed in 1.5 ml Eppendorf tubes, frozen in liquid nitrogen, and ground to a powder using a Tissuelyser-24 (JINGXIN). After adding 500 µl of 1% acetic acid, the sample was incubated in the dark at 4 °C for 30 min with a shaking at 15 min. The mixture was centrifuged at 4 °C, 12,000 × g for 10 min, and 90 µl of the supernatant was transferred to a new 1.5 ml Eppendorf tube. The sample was mixed with 210 µl of reaction buffer containing 0.42% ammonium molybdate and 10% ascorbic acid (v/v, 6:1). The sample was incubated at 37 °C for 1 h, and the optical density (OD) at 820 nm was measured in a Spark Multimode Microplate Reader (Tecan). A standard curve was constructed using different concentrations of KH₂PO₄ (0, 0.5, 1, 2.5, 5, 7.5, 10, 15, 25, 30, 50 nmol), and the Pi concentrations in shoots were calculated.

Measuring anthocyanin content

Anthocyanin content was measured as previously described⁵³. After measuring the fresh weight of each sample (typically 0.1–0.5 g), the sample was treated with 1 ml of anthocyanin extraction buffer composed of propanol/HCl/H₂O at a ratio of 18/1/81 (v/v/v). Following overnight incubation at room temperature in the dark, the samples were centrifuged at 12,000 × g for 20 min. The supernatants were diluted two-fold with extraction buffer, and their absorbance (A) was measured at 535 nm and 650 nm. The relative anthocyanin content was then determined using the formula: 2× (A₅₃₅–A₆₅₀) per gram of fresh weight.

Statistics and reproducibility

Primary root lengths were measured with ImageJ (https://imagej.net/software/fiji/). Details of statistical analysis, including the statistical tests used, exact n values for each measurement, and P values, are presented in the figure legends and figures. All experiments were independently repeated at least once and showed similar results. No statistical method was used to predetermine sample size. No data were excluded from the analyses.

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.

Data availability

All data are available in the main text or the supplementary materials. The raw sequencing data of m⁶A-seq, RNA-seq, transcriptome-wide RNA decay assay, and PAS-seq 2 generated in this study have been deposited in the Genome Sequence Archive (GSA) under accession code CRA016391. The processed results of m⁶A-seq, RNA-seq, transcriptome-wide RNA decay assay, and PAS-seq 2 are available in Supplementary Datas 1–4. Source Data are provided with this paper. Plasmids and transgenic plants generated in this study will be available from the corresponding author upon request. Source data are provided with this paper.

References

Zhao, B. S., Roundtree, I. A. & He, C. Post-transcriptional gene regulation by mRNA modifications. Nat. Rev. Mol. Cell Biol. 18, 31–42 (2017).
CAS PubMed Google Scholar
Shen, L., Ma, J., Li, P., Wu, Y. & Yu, H. Recent advances in the plant epitranscriptome. Genome Biol. 24, 43 (2023).
CAS PubMed PubMed Central Google Scholar
Tang, J., Chen, S. & Jia, G. Detection, regulation, and functions of RNA N(6)-methyladenosine modification in plants. Plant Commun. 4, 100546 (2023).
CAS PubMed PubMed Central Google Scholar
Zaccara, S., Ries, R. J. & Jaffrey, S. R. Reading, writing and erasing mRNA methylation. Nat. Rev. Mol. Cell Biol. 20, 608–624 (2019).
CAS PubMed Google Scholar
Sharma, B., Prall, W., Bhatia, G. & Gregory, B. D. The diversity and functions of plant RNA modifications: what we know and where we go from here. Annu. Rev. Plant Biol. 74, 53–85 (2023).
CAS PubMed Google Scholar
Liu, J. et al. A METTL3-METTL14 complex mediates mammalian nuclear RNA N6-adenosine methylation. Nat. Chem. Biol. 10, 93–95 (2014).
ADS CAS PubMed Google Scholar
Jia, G. et al. N6-methyladenosine in nuclear RNA is a major substrate of the obesity-associated FTO. Nat. Chem. Biol. 7, 885–887 (2011).
CAS PubMed PubMed Central Google Scholar
Zheng, G. et al. ALKBH5 is a mammalian RNA demethylase that impacts RNA metabolism and mouse fertility. Mol. Cell 49, 18–29 (2013).
CAS PubMed Google Scholar
Zhong, S. et al. MTA is an Arabidopsis messenger RNA adenosine methylase and interacts with a homolog of a sex-specific splicing factor. Plant Cell 20, 1278–1288 (2008).
CAS PubMed PubMed Central Google Scholar
Shen, L. et al. N(6)-methyladenosine RNA modification regulates shoot stem cell fate in Arabidopsis. Dev. Cell 38, 186–200 (2016).
CAS PubMed PubMed Central Google Scholar
Růžička, K. et al. Identification of factors required for m(6) A mRNA methylation in Arabidopsis reveals a role for the conserved E3 ubiquitin ligase HAKAI. New Phytol. 215, 157–172 (2017).
PubMed PubMed Central Google Scholar
Wong, C. E. et al. Shaping the landscape of N6-methyladenosine RNA methylation in Arabidopsis. Plant Physiol. 191, 2045–2063 (2023).
CAS PubMed PubMed Central Google Scholar
Shen, L. Functional interdependence of N6-methyladenosine methyltransferase complex subunits in Arabidopsis. Plant Cell 35, 1901–1916 (2023).
PubMed PubMed Central Google Scholar
Xu, T. et al. FIONA1-mediated m(6) a modification regulates the floral transition in Arabidopsis. Adv. Sci. 9, e2103628 (2022).
Google Scholar
Wang, C. et al. FIONA1 is an RNA N(6)-methyladenosine methyltransferase affecting Arabidopsis photomorphogenesis and flowering. Genome Biol. 23, 40 (2022).
CAS PubMed PubMed Central Google Scholar
Sun, B. et al. FIONA1-mediated methylation of the 3’UTR of FLC affects FLC transcript levels and flowering in Arabidopsis. PLoS Genet. 18, e1010386 (2022).
CAS PubMed PubMed Central Google Scholar
Duan, H. C. et al. ALKBH10B is an RNA N(6)-methyladenosine demethylase affecting Arabidopsis floral transition. Plant Cell 29, 2995–3011 (2017).
CAS PubMed PubMed Central Google Scholar
Martínez-Pérez, M. et al. Arabidopsis m(6)A demethylase activity modulates viral infection of a plant virus and the m(6)A abundance in its genomic RNAs. Proc. Natl. Acad. Sci. USA 114, 10755–10760 (2017).
ADS PubMed PubMed Central Google Scholar
Tang, J. et al. The RNA N(6) -methyladenosine demethylase ALKBH9B modulates ABA responses in Arabidopsis. J. Integr. Plant Biol. 64, 2361–2373 (2022).
CAS PubMed Google Scholar
Meyer, K. D. DART-seq: an antibody-free method for global m(6)A detection. Nat. Methods 16, 1275–1280 (2019).
CAS PubMed PubMed Central Google Scholar
Helm, M., Lyko, F. & Motorin, Y. Limited antibody specificity compromises epitranscriptomic analyses. Nat. Commun. 10, 5669 (2019).
ADS CAS PubMed PubMed Central Google Scholar
Zhang, Z. et al. Systematic calibration of epitranscriptomic maps using a synthetic modification-free RNA library. Nat. Methods 18, 1213–1222 (2021).
CAS PubMed Google Scholar
Baquero-Pérez, B. et al. N6-methyladenosine modification is not a general trait of viral RNA genomes. Nat. Commun. 15, 1964 (2024).
ADS PubMed PubMed Central Google Scholar
Liu, C. et al. Absolute quantification of single-base m(6)A methylation in the mammalian transcriptome using GLORI. Nat. Biotechnol. 41, 355–366 (2023).
CAS PubMed Google Scholar
Yue, H., Nie, X., Yan, Z. & Weining, S. N6-methyladenosine regulatory machinery in plants: composition, function and evolution. Plant Biotechnol. J. 17, 1194–1208 (2019).
PubMed PubMed Central Google Scholar
Hu, J. et al. N(6) -Methyladenosine mRNA methylation is important for salt stress tolerance in Arabidopsis. Plant J. 106, 1759–1775 (2021).
CAS PubMed Google Scholar
Anderson, S. J. et al. N(6)-methyladenosine inhibits local ribonucleolytic cleavage to stabilize mRNAs in Arabidopsis. Cell Rep. 25, 1146–1157.e1143 (2018).
CAS PubMed Google Scholar
Gao, Y. et al. Drought induces epitranscriptome and proteome changes in stem-differentiating xylem of Populus trichocarpa. Plant Physiol. 190, 459–479 (2022).
CAS PubMed PubMed Central Google Scholar
Hou, N. et al. MdMTA-mediated m(6) A modification enhances drought tolerance by promoting mRNA stability and translation efficiency of genes involved in lignin deposition and oxidative stress. New Phytol. 234, 1294–1314 (2022).
CAS PubMed Google Scholar
Liu, G., Wang, J. & Hou, X. Transcriptome-wide N(6)-methyladenosine (m(6)A) methylome profiling of heat stress in pak-choi (Brassica rapa ssp. chinensis). Plants (Basel) 9, 1080 (2020).
CAS PubMed Google Scholar
Govindan, G. et al. mRNA N(6) -methyladenosine is critical for cold tolerance in Arabidopsis. Plant J. 111, 1052–1068 (2022).
CAS PubMed PubMed Central Google Scholar
Zhang, M. et al. N(6)-methyladenosine RNA modification regulates photosynthesis during photodamage in plants. Nat. Commun. 13, 7441 (2022).
ADS CAS PubMed PubMed Central Google Scholar
Zhang, K. et al. The dynamics of N(6)-methyladenine RNA modification in interactions between rice and plant viruses. Genome Biol. 22, 189 (2021).
CAS PubMed PubMed Central Google Scholar
Zhang, T. et al. N6-methyladenosine RNA modification promotes viral genomic RNA stability and infection. Nat. Commun. 13, 6576 (2022).
ADS CAS PubMed PubMed Central Google Scholar
Guo, T. et al. The m(6) A reader MhYTP2 regulates MdMLO19 mRNA stability and antioxidant genes translation efficiency conferring powdery mildew resistance in apple. Plant Biotechnol. J. 20, 511–525 (2022).
CAS PubMed Google Scholar
Zhou, L., Tian, S. & Qin, G. RNA methylomes reveal the m(6)A-mediated regulation of DNA demethylase gene SlDML2 in tomato fruit ripening. Genome Biol. 20, 156 (2019).
PubMed PubMed Central Google Scholar
Cheng, P. et al. RNA N(6)-methyladenosine modification promotes auxin biosynthesis required for male meiosis in rice. Dev. Cell 57, 246–259.e244 (2022).
CAS PubMed Google Scholar
Zhou, L. et al. N(6)-methyladenosine RNA modification regulates strawberry fruit ripening in an ABA-dependent manner. Genome Biol. 22, 168 (2021).
CAS PubMed PubMed Central Google Scholar
Schachtman, D. P. & Shin, R. Nutrient sensing and signaling: NPKS. Annu. Rev. Plant Biol. 58, 47–69 (2007).
CAS PubMed Google Scholar
Bailey-Serres, J., Parker, J. E., Ainsworth, E. A., Oldroyd, G. E. D. & Schroeder, J. I. Genetic strategies for improving crop yields. Nature 575, 109–118 (2019).
ADS CAS PubMed PubMed Central Google Scholar
Qi, W., Baldwin, S. A., Muench, S. P. & Baker, A. Pi sensing and signalling: from prokaryotic to eukaryotic cells. Biochem. Soc. Trans. 44, 766–773 (2016).
CAS PubMed Google Scholar
Wild, R. et al. Control of eukaryotic phosphate homeostasis by inositol polyphosphate sensor domains. Science 352, 986–990 (2016).
ADS CAS PubMed Google Scholar
Xu, C. et al. A phosphate-sensing organelle regulates phosphate and tissue homeostasis. Nature 617, 798–806 (2023).
ADS CAS PubMed PubMed Central Google Scholar
Lambers, H. Phosphorus acquisition and utilization in plants. Annu. Rev. Plant Biol. 73, 17–42 (2022).
CAS PubMed Google Scholar
López-Arredondo, D. L., Leyva-González, M. A., González-Morales, S. I., López-Bucio, J. & Herrera-Estrella, L. Phosphate nutrition: improving low-phosphate tolerance in crops. Annu. Rev. Plant Biol. 65, 95–123 (2014).
PubMed Google Scholar
Cong, W. F., Suriyagoda, L. D. B. & Lambers, H. Tightening the phosphorus cycle through phosphorus-efficient crop genotypes. Trends Plant Sci. 25, 967–975 (2020).
CAS PubMed Google Scholar
Paz-Ares, J. et al. Plant adaptation to low phosphorus availability: core signaling, crosstalks, and applied implications. Mol. Plant 15, 104–124 (2022).
CAS PubMed Google Scholar
Yang, S. Y., Lin, W. Y., Hsiao, Y. M. & Chiou, T. J. Milestones in understanding transport, sensing, and signaling of the plant nutrient phosphorus. Plant Cell 36, 1504–1523 (2024).
PubMed PubMed Central Google Scholar
Rubio, V. et al. A conserved MYB transcription factor involved in phosphate starvation signaling both in vascular plants and in unicellular algae. Genes Dev. 15, 2122–2133 (2001).
CAS PubMed PubMed Central Google Scholar
Zhou, J. et al. OsPHR2 is involved in phosphate-starvation signaling and excessive phosphate accumulation in shoots of plants. Plant Physiol. 146, 1673–1686 (2008).
CAS PubMed PubMed Central Google Scholar
Guo, M. et al. Integrative comparison of the role of the PHOSPHATE RESPONSE1 subfamily in phosphate signaling and homeostasis in rice. Plant Physiol. 168, 1762–1776 (2015).
CAS PubMed PubMed Central Google Scholar
Wang, Z., Zheng, Z., Zhu, Y., Kong, S. & Liu, D. PHOSPHATE RESPONSE 1 family members act distinctly to regulate transcriptional responses to phosphate starvation. Plant Physiol. 191, 1324–1343 (2023).
CAS PubMed Google Scholar
Bustos, R. et al. A central regulatory system largely controls transcriptional activation and repression responses to phosphate starvation in Arabidopsis. PLoS Genet. 6, e1001102 (2010).
PubMed PubMed Central Google Scholar
Barragán-Rosillo, A. C. et al. Genome accessibility dynamics in response to phosphate limitation is controlled by the PHR1 family of transcription factors in Arabidopsis. Proc. Natl. Acad. Sci. USA 118, e2107558118 (2021).
PubMed PubMed Central Google Scholar
Pant, B. D. et al. The transcription factor PHR1 regulates lipid remodeling and triacylglycerol accumulation in Arabidopsis thaliana during phosphorus starvation. J. Exp. Bot. 66, 1907–1918 (2015).
CAS PubMed PubMed Central Google Scholar
Pant, B. D. et al. Identification of primary and secondary metabolites with phosphorus status-dependent abundance in Arabidopsis, and of the transcription factor PHR1 as a major regulator of metabolic changes during phosphorus limitation. Plant Cell Environ. 38, 172–187 (2015).
CAS PubMed Google Scholar
Shi, J. et al. A phosphate starvation response-centered network regulates mycorrhizal symbiosis. Cell 184, 5527–5540.e5518 (2021).
CAS PubMed Google Scholar
Puga, M. I. et al. SPX1 is a phosphate-dependent inhibitor of Phosphate Starvation Response 1 in Arabidopsis. Proc. Natl. Acad. Sci. USA 111, 14947–14952 (2014).
ADS CAS PubMed PubMed Central Google Scholar
Wang, Z. Y. et al. Rice SPX1 and SPX2 inhibit phosphate starvation responses through interacting with PHR2 in a phosphate-dependent manner. Proc. Natl. Acad. Sci. USA 111, 14953–14958 (2014).
ADS CAS PubMed PubMed Central Google Scholar
Lv, Q. et al. SPX4 negatively regulates phosphate signaling and homeostasis through its interaction with PHR2 in rice. Plant Cell 26, 1586–1597 (2014).
CAS PubMed PubMed Central Google Scholar
Dong, J. et al. Inositol pyrophosphate InsP(8) acts as an intracellular phosphate signal in Arabidopsis. Mol. Plant 12, 1463–1473 (2019).
CAS PubMed Google Scholar
Guo, M. et al. A reciprocal inhibitory module for Pi and iron signaling. Mol. Plant 15, 138–150 (2022).
CAS PubMed Google Scholar
Miura, K. et al. The Arabidopsis SUMO E3 ligase SIZ1 controls phosphate deficiency responses. Proc. Natl. Acad. Sci. USA 102, 7760–7765 (2005).
ADS CAS PubMed PubMed Central Google Scholar
Zhang, G. et al. Brassinosteroid-dependent phosphorylation of PHOSPHATE STARVATION RESPONSE2 reduces its DNA-binding ability in rice. Plant Cell 36, 2253–2271 (2024).
PubMed PubMed Central Google Scholar
Duan, K. et al. Characterization of a sub-family of Arabidopsis genes with the SPX domain reveals their diverse functions in plant tolerance to phosphorus starvation. Plant J. 54, 965–975 (2008).
CAS PubMed Google Scholar
González, E., Solano, R., Rubio, V., Leyva, A. & Paz-Ares, J. PHOSPHATE TRANSPORTER TRAFFIC FACILITATOR1 is a plant-specific SEC12-related protein that enables the endoplasmic reticulum exit of a high-affinity phosphate transporter in Arabidopsis. Plant Cell 17, 3500–3512 (2005).
PubMed PubMed Central Google Scholar
Bayle, V. et al. Arabidopsis thaliana high-affinity phosphate transporters exhibit multiple levels of posttranslational regulation. Plant Cell 23, 1523–1535 (2011).
CAS PubMed PubMed Central Google Scholar
Hamburger, D., Rezzonico, E., MacDonald-Comber Petétot, J., Somerville, C. & Poirier, Y. Identification and characterization of the Arabidopsis PHO1 gene involved in phosphate loading to the xylem. Plant Cell 14, 889–902 (2002).
CAS PubMed PubMed Central Google Scholar
Svistoonoff, S. et al. Root tip contact with low-phosphate media reprograms plant root architecture. Nat. Genet. 39, 792–796 (2007).
CAS PubMed Google Scholar
Zhu, J. et al. Two bifunctional inositol pyrophosphate kinases/phosphatases control plant phosphate homeostasis. Elife 8, e43582 (2019).
PubMed PubMed Central Google Scholar
Liu, J. et al. A vacuolar phosphate transporter essential for phosphate homeostasis in Arabidopsis. Proc. Natl. Acad. Sci. USA 112, E6571–E6578 (2015).
CAS PubMed PubMed Central Google Scholar
Liu, T. Y. et al. Identification of plant vacuolar transporters mediating phosphate storage. Nat. Commun. 7, 11095 (2016).
ADS CAS PubMed PubMed Central Google Scholar
Castrillo, G. et al. Root microbiota drive direct integration of phosphate stress and immunity. Nature 543, 513–518 (2017).
ADS CAS PubMed PubMed Central Google Scholar
Zhong, S. X. et al. SERRATE drives phase separation behaviours to regulate m6A modification and miRNA biogenesis. Nat. Cell Biol. 26, 2129–2143 (2024).
CAS PubMed PubMed Central Google Scholar
Bai, H. Y. et al. The METHYLTRANSFERASE B-SERRATE interaction mediates the reciprocal regulation of microRNA biogenesis and RNA m6A modification. J. Integr. Plant Biol. 66, 2613–2631 (2024).
CAS PubMed PubMed Central Google Scholar
Wang, X. et al. A photoregulatory mechanism of the circadian clock in Arabidopsis. Nat. Plants 7, 1397–1408 (2021).
CAS PubMed Google Scholar
Jiang, B. C. et al. Light-induced LLPS of the CRY2/SPA1/FIO1 complex regulating mRNA methylation and chlorophyll homeostasis in Arabidopsis. Nat. Plants 9, 2042–2058 (2023).
CAS PubMed PubMed Central Google Scholar
Yankova, E. et al. Small-molecule inhibition of METTL3 as a strategy against myeloid leukaemia. Nature 593, 597–601 (2021).
ADS CAS PubMed PubMed Central Google Scholar
Sorenson, R. S., Deshotel, M. J., Johnson, K., Adler, F. R. & Sieburth, L. E. Arabidopsis mRNA decay landscape arises from specialized RNA decay substrates, decapping-mediated feedback, and redundancy. Proc. Natl. Acad. Sci. USA 115, E1485–e1494 (2018).
ADS CAS PubMed PubMed Central Google Scholar
Lin, S. I. et al. Regulatory network of microRNA399 and PHO2 by systemic signaling. Plant Physiol. 147, 732–746 (2008).
CAS PubMed PubMed Central Google Scholar
Pant, B. D., Buhtz, A., Kehr, J. & Scheible, W. R. MicroRNA399 is a long-distance signal for the regulation of plant phosphate homeostasis. Plant J. 53, 731–738 (2008).
CAS PubMed PubMed Central Google Scholar
Song, P. Z. et al. Arabidopsis N6-methyladenosine reader CPSF30-L recognizes FUE signals to control polyadenylation site choice in liquid-like nuclear bodies. Mol. Plant 14, 571–587 (2021).
CAS PubMed Google Scholar
Yoon, Y., Soles, L. V. & Shi, Y. S. PAS-seq 2: a fast and sensitive method for global profiling of polyadenylated RNAs. mRNA 3’ End. Process. Metab. 655, 25–35 (2021).
CAS Google Scholar
Hou, Y. F. et al. CPSF30-L-mediated recognition of mRNA m6A modification controls alternative polyadenylation of nitrate signaling-related gene transcripts in Arabidopsis. Mol. Plant 14, 688–699 (2021).
CAS PubMed Google Scholar
Xiao, Y. et al. An elongation- and ligation-based qPCR amplification method for the radiolabeling-free detection of locus-specific N6-methyladenosine modification. Angew. Chem. Int. Ed. 57, 15995–16000 (2018).
CAS Google Scholar
Huang, H. et al. Histone H3 trimethylation at lysine 36 guides m(6)A RNA modification co-transcriptionally. Nature 567, 414–419 (2019).
ADS CAS PubMed PubMed Central Google Scholar
Shim, S., Lee, H. G., Lee, H. & Seo, P. J. H3K36me2 is highly correlated with m(6) A modifications in plants. J. Integr. Plant Biol. 62, 1455–1460 (2020).
CAS PubMed Google Scholar
Shao, Y., Wong, C. E., Shen, L. & Yu, H. N(6)-methyladenosine modification underlies messenger RNA metabolism and plant development. Curr. Opin. Plant Biol. 63, 102047 (2021).
CAS PubMed Google Scholar
Parker M. T. et al. Nanopore direct RNA sequencing maps the complexity of Arabidopsis mRNA processing and m(6)A modification. eLife. 9, e49658 (2020).
Wang, X. et al. N6-methyladenosine-dependent regulation of messenger RNA stability. Nature 505, 117–120 (2014).
ADS PubMed Google Scholar
Du, H. et al. YTHDF2 destabilizes m(6)A-containing RNA through direct recruitment of the CCR4-NOT deadenylase complex. Nat. Commun. 7, 12626 (2016).
ADS CAS PubMed PubMed Central Google Scholar
Huang, H. et al. Recognition of RNA N(6)-methyladenosine by IGF2BP proteins enhances mRNA stability and translation. Nat. Cell Biol. 20, 285–295 (2018).
CAS PubMed PubMed Central Google Scholar
Wei, L. H. et al. The m6A reader ECT2 controls trichome morphology by affecting mRNA stability in arabidopsis. Plant Cell 30, 968–985 (2018).
CAS PubMed PubMed Central Google Scholar
Reis, R., Deforges, J., Clua, J. & Poirier Y. Alternative promoter usage modulates miRNA-guided translation inhibition of a m6A reader in phosphate starvation. Preprint at bioRxiv. (2020).
Reis, R. S. et al. Phosphate deficiency alters transcript isoforms via alternative transcription start sites. Plant J. 120, 218–233 (2024).
CAS PubMed Google Scholar
Zhang, X., Henriques, R., Lin, S. S., Niu, Q. W. & Chua, N. H. Agrobacterium-mediated transformation of Arabidopsis thaliana using the floral dip method. Nat. Protoc. 1, 641–646 (2006).
CAS PubMed Google Scholar
Carbonell, A. et al. Highly specific gene silencing in a monocot species by artificial microRNAs derived from chimeric miRNA precursors. Plant J. 82, 1061–1075 (2015).
CAS PubMed PubMed Central Google Scholar
Yu, Q. et al. RNA demethylation increases the yield and biomass of rice and potato plants in field trials. Nat. Biotechnol. 39, 1581–1588 (2021).
CAS PubMed Google Scholar
Martin, M. Cutadapt removes adapter sequences from high-throughput sequencing reads. EMBnet. J. 17, 10–12 (2011).
Google Scholar
Kim, D., Paggi, J. M., Park, C., Bennett, C. & Salzberg, S. L. Graph-based genome alignment and genotyping with HISAT2 and HISAT-genotype. Nat. Biotechnol. 37, 907–915 (2019).
CAS PubMed PubMed Central Google Scholar
Thorvaldsdóttir, H., Robinson, J. T. & Mesirov, J. P. Integrative Genomics Viewer (IGV): high-performance genomics data visualization and exploration. Brief. Bioinform. 14, 178–192 (2013).
PubMed Google Scholar
Quinlan, A. R. & Hall, I. M. BEDTools: a flexible suite of utilities for comparing genomic features. Bioinformatics 26, 841–842 (2010).
CAS PubMed PubMed Central Google Scholar
Liao, Y., Smyth, G. K. & Shi, W. featureCounts: an efficient general purpose program for assigning sequence reads to genomic features. Bioinformatics 30, 923–930 (2014).
CAS PubMed Google Scholar
Love, M. I., Huber, W. & Anders, S. Moderated estimation of fold change and dispersion for RNA-seq data with DESeq2. Genome Biol. 15, 550 (2014).
PubMed PubMed Central Google Scholar
Wang, Z. Y. et al. SWI2/SNF2 ATPase CHR2 remodels pri-miRNAs via Serrate to impede miRNA production. Nature 557, 516–521 (2018).
ADS CAS PubMed Google Scholar
Chen, C. et al. TBtools: an integrative toolkit developed for interactive analyses of big biological data. Mol. Plant 13, 1194–1202 (2020).
CAS PubMed Google Scholar
Gu, Z., Eils, R. & Schlesner, M. Complex heatmaps reveal patterns and correlations in multidimensional genomic data. Bioinformatics 32, 2847–2849 (2016).
CAS PubMed Google Scholar

Download references

Acknowledgements

We thank Hang Su (Institute of Botany, Chinese Academy of Sciences) for the assistance with the LC-MS/MS assay. We thank Wei Yin (the Core Facilities, Zhejiang University School of Medicine) for the technical support for fluorescent imaging. This work was supported by grants from the Natural Science Foundation of Zhejiang Province, China (LR24C150001 and LZ22C150002), the National Natural Science Foundation of China (32170262), the National Key Research and Development Program of China (2021YFF1000402 and 2022YFD1401600) and the Fundamental Research Funds for the Central University (226202400102) to Z.W., the National Natural Science Foundation of China (32170593) and the Guangdong Provincial Pearl River Talent Plan (2019QN01N108) to Z.Z.

Author information

Authors and Affiliations

State Key Laboratory of Plant Environmental Resilience, College of Life Sciences, Zhejiang University, Hangzhou, Zhejiang, China
Kai Liu, Xiaojia Wang, Jingyi Wang, Shuman Wang, Weiguo Dong, Lulu Qiao, Qiongli Jin & Zhiye Wang
Guangdong Provincial Key Laboratory of Biotechnology for Plant Development, School of Life Science, South China Normal University, Guangzhou, China
Haiyan Bai & Zhonghui Zhang
MOE Key Laboratory of Gene Function and Regulation, Guangdong Province Key Laboratory of Pharmaceutical Functional Genes, State Key Laboratory of Biocontrol, School of Life Sciences, Sun Yat-sen University, Guangzhou, China
Guan-Zheng Luo
The First Affiliated Hospital, College of Medicine, Zhejiang University, Hangzhou, Zhejiang, China
Zhiye Wang

Authors

Kai Liu
View author publications
Search author on:PubMed Google Scholar
Xiaojia Wang
View author publications
Search author on:PubMed Google Scholar
Jingyi Wang
View author publications
Search author on:PubMed Google Scholar
Shuman Wang
View author publications
Search author on:PubMed Google Scholar
Haiyan Bai
View author publications
Search author on:PubMed Google Scholar
Weiguo Dong
View author publications
Search author on:PubMed Google Scholar
Lulu Qiao
View author publications
Search author on:PubMed Google Scholar
Qiongli Jin
View author publications
Search author on:PubMed Google Scholar
Zhonghui Zhang
View author publications
Search author on:PubMed Google Scholar
Guan-Zheng Luo
View author publications
Search author on:PubMed Google Scholar
Zhiye Wang
View author publications
Search author on:PubMed Google Scholar

Contributions

Z.W. conceived the project and designed the experiments; K.L., X.W., J.W., S.W., H.B., L.Q. performed the experiments; K.L. conducted the data analyses, with assistance from W.D. and Q.J.; G-Z.L., and Z.Z. provided intellectual support. Z.W. and K.L. wrote the manuscript with input from all authors.

Corresponding author

Correspondence to Zhiye Wang.

Ethics declarations

Competing interests

The authors declare no competing interests.

Peer review

Peer review information

Nature Communications thanks Guifang Jia and the other, anonymous, reviewers for their contribution to the peer review of this work. A peer review file is available.

Additional information

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary information

Supplementary Information

Description of Additional Supplementary Files

Supplementary Data 1

Supplementary Data 2

Supplementary Data 3

Supplementary Data 4

Supplementary Data 5

Reporting Summary

Transparent Peer Review file

Source data

Source Data

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/.

Reprints and permissions

About this article

Cite this article

Liu, K., Wang, X., Wang, J. et al. N ⁶-methyladenosine modifications stabilize phosphate starvation response–related mRNAs in plant adaptation to nutrient-deficient stress. Nat Commun 16, 4093 (2025). https://doi.org/10.1038/s41467-025-59331-y

Download citation

Received: 25 April 2024
Accepted: 16 April 2025
Published: 01 May 2025
DOI: https://doi.org/10.1038/s41467-025-59331-y