A ubiquitin carboxyl-terminal hydrolase facilitates H2A.Z removal in response to phosphate starvation

Abstract

The histone variant H2A.Z plays a critical role in the modulation of environmental responses, but the mechanisms regulating H2A.Z dynamics in response to stresses remain poorly understood in plants. Here we identify the ubiquitin carboxyl-terminal hydrolase UCH2 as a Pi-responsive chromatin regulator that promotes Pi starvation-induced (PSI) gene activation. We show that PHR1 recruits UCH2 to PSI gene promoters under Pi deficiency, where UCH2 deubiquitinates H2A.Z and promotes reduction in H2A.Z occupancy at PSI gene promoters. Loss of UCH2 impairs this reduction in H2A.Z occupancy, represses PSI gene induction, and reduces Pi accumulation. A catalytically inactive UCH2 mutant fails to complement these defects, and ubiquitination-deficient H2A.Z variants render plants insensitive to UCH2 loss. Our findings establish UCH2 as a critical chromatin regulator linking Pi sensing with H2A.Z dynamics and provide insights into how plants integrate chromatin-based regulation to maintain Pi homeostasis.

Under phosphate starvation, plants undergo transcriptional reprogramming. Here Meng et al. show that the PHR1 transcription factor recruits the UCH2 deubiquitinase to remove ubiquitin from the H2A.Z histone variant at target genes, promoting H2A.Z eviction and activating phosphate starvation responses.

Introduction

Phosphorus is an essential macronutrient for plant growth and development, required for fundamental cellular processes, including energy metabolism, membrane integrity, and nucleic acid synthesis¹. Despite its abundance in Earth’s crust, bioavailable phosphate (Pi) in most soils is severely limited due to its low mobility and strong adsorption to soil particles, thereby constraining plant growth and agricultural productivity worldwide². To cope with Pi scarcity, plants have evolved sophisticated molecular mechanisms that collectively constitute the phosphate starvation response (PSR)^3,4.

The PSR encompasses coordinated transcriptional reprogramming that enhances Pi acquisition, remobilization, and metabolic adaptation. Central to this response is the basic helix-loop-helix transcription factor PHR1 (PHOSPHATE STARVATION RESPONSE 1) and its homologs, which activate hundreds of Pi starvation-inducible (PSI) genes by binding to P1BS (PHR1-binding sequence) cis-elements within their promoters^5,6. PHR1 transcript levels remain weakly responsive to Pi starvation, while post-transcriptionally it undergoes SUMOylation⁷. Its activity is primarily regulated through protein-protein interactions with SPX proteins in a Pi-dependent manner⁸. SPX1 and SPX2 are Pi sensors that are directly induced by PHR1 during Pi starvation^9,10. When Pi is replenished, the Pi signal molecule InsP₈ binds to SPX1/2 and promotes their interaction with PHR1 to inhibit its activity^11–14.

In addition to transcription factors, chromatin structure plays a pivotal role in transcriptional regulation, with histone post-translational modifications and histone variants serving as key determinants of gene expression efficiency¹⁵. The nucleosome, the fundamental unit of chromatin, consists of ~147 base pairs of DNA wrapped around a histone octamer, which includes two copies each of the canonical histones H2A, H2B, H3, and H4^16,17. In addition to these canonical histones, eukaryotic cells incorporate various histone variants that confer distinct regulatory properties to chromatin. The histone variant H2A.Z, incorporated by the SWR1 chromatin remodeling complex, generally marks transcriptionally active or poised chromatin domains^18,19. Paradoxically, studies in Arabidopsis and rice demonstrated that H2A.Z occupancy at PSI gene loci is associated with transcriptional repression under Pi-sufficient (P+) conditions, whereas its depletion is linked to gene activation under Pi-deficient (P−) conditions^20–22.

The dynamic regulation of H2A.Z involves not only its incorporation and eviction but also post-translational modifications such as acetylation, methylation, and ubiquitination, which modulate its affinity for chromatin^23,24. Although acetylation of H2A.Z has been shown to facilitate the transition between stress responses and autotrophic growth^25,26, ubiquitin conjugation to H2A.Z modulates chromatin structure and transcriptional outcomes through multiple mechanisms²⁷. In Arabidopsis, the Polycomb-group protein BMI1 has been implicated in catalyzing H2A.Z ubiquitination, contributing to gene silencing under stress conditions²⁸. However, whether this ubiquitination is reversible and, if so, which enzymes mediate its nutrient-sensitive reduction in occupancy, remains largely unexplored.

In plants, deubiquitinating enzymes have been increasingly recognized for their roles in developmental transitions, immune responses, and environmental adaptation^29,30. Among these, the ubiquitin C-terminal hydrolase (UCH) family forms a distinct class of cysteine proteases involved in removing ubiquitin moieties from histones or regulatory proteins^31,32. In Arabidopsis, three UCH enzymes (UCH1, UCH2, UCH3) have been identified, with UCH1 and UCH2 implicated in auxin signaling and shoot architecture³³. However, whether UCH enzymes participate in nutrient stress responses and chromatin regulation remains unexplored. Given the critical role of chromatin remodeling in PSR gene regulation, we hypothesized that deubiquitinating activities might contribute to the dynamic control of H2A.Z states during Pi adaptation.

Here, we demonstrate that UCH2 functions as a Pi-responsive chromatin modifier that directly regulates H2A.Z dynamics to fine-tune PSI gene expression. Through comprehensive genetic, biochemical, and genomic approaches, we show that UCH2 is recruited to PSI gene loci in a PHR1-dependent manner, where it promotes H2A.Z deubiquitination. Our findings reveal a regulatory mechanism linking nutrient sensing to epigenetic control and establish UCH2 as an essential component of plant Pi homeostasis.

Results

PHR1 physically interacts with the ubiquitin C-terminal hydrolase UCH2

To identify additional PHR1-associated factors that regulate PSI gene expression, we performed immunoprecipitation followed by mass spectrometry in Arabidopsis phr1 mutant plants complemented with FLAG-tagged PHR1 driven by its native promoter. Along with the known PHR1-interacting proteins SPX1 and SPX3, a ubiquitin C-terminal hydrolase, UCH2, was identified (Supplementary Data 1). We next used yeast two-hybrid (Y2H) assays to determine whether PHR1 interacted directly with UCH2. The C-terminal region (209–409 AA) of PHR1 comprises a MYB DNA-binding domain and a coiled-coil dimerization domain, required for DNA binding and oligomerization, respectively. The N-terminal region (1–208 AA) lacks a defined conserved motif but is generally described as important for protein-protein interactions. Y2H results showed that UCH2 binds the N-terminal, but not the C-terminal, region of PHR1 (Fig. 1a). Since three UCH family proteins—UCH1, UCH2, and UCH3—exist in Arabidopsis, we also tested interactions between PHR1 and UCH1 or UCH3, and found that neither protein interacted with PHR1 in Y2H assays, whether with the N- or C-terminal regions (Supplementary Fig. 1). The interaction between PHR1 and UCH2 was confirmed using split-luciferase complementation assays in tobacco (Nicotiana benthamiana) leaves (Fig. 1b). Moreover, bimolecular fluorescence complementation (BiFC) assays revealed that co-expression of UCH2-nYFP and PHR1-cYFP in tobacco leaves produced a fluorescence signal in the nucleus, indicating that UCH2 interacts with PHR1 in the nucleus (Fig. 1c). The PHR1–UCH2 interaction was further confirmed by co-immunoprecipitation (Co-IP) assays using tobacco leaves transiently expressing UCH2-MYC together with PHR1-FLAG (Fig. 1d). To validate this interaction at the endogenous level, we performed Co-IP using pUCH2::UCH2–FLAG/uch2 seedlings. UCH2–FLAG specifically co-precipitated endogenous PHR1, whereas no PHR1 signal was detected in the Col-0 negative control (Fig. 1e and Supplementary Fig. 2). Together, these results confirm that UCH2 physically interacts with PHR1 in Arabidopsis.

Fig. 1. UCH2 interacts with the phosphate starvation response transcription factor PHR1.

a Yeast two-hybrid (Y2H) assay showing that UCH2 specifically interacts with the N-terminal domain of PHR1 (PHR1(N)). BD, GAL4 DNA-binding domain fusion; AD, GAL4 activation domain fusion. Yeast co-transformants were selected on SD-WL (Trp– Leu–) and tested for interaction on SD-WLHA (Trp– Leu– His– Ade–) media. b Split-luciferase complementation (Split-Luc) assay in N. benthamiana leaves showing in vivo interaction between UCH2 and PHR1. NLuc and CLuc denote N- and C-terminal fragments of luciferase fused to UCH2 and PHR1, respectively. c Bimolecular fluorescence complementation (BiFC) assay confirming nuclear interaction between UCH2 and PHR1 in N. benthamiana epidermal cells. Scale bar, 30 μm. d Co-immunoprecipitation (Co-IP) of UCH2 and PHR1 in N. benthamiana leaves. Total proteins were extracted from leaves co-expressing UCH2-MYC and PHR1-FLAG. UCH2 was immunoprecipitated with anti-MYC antibody, and co-precipitated PHR1 was detected by immunoblotting with anti-FLAG antibody. Source data are provided as a Source Data file. e A Co-immunoprecipitation (Co-IP) assay showing endogenous interaction between UCH2 and PHR1 in Arabidopsis. Total proteins were extracted from pUCH2::UCH2-FLAG/uch2 and Col-0 seedlings grown under P+ or P− conditions, immunoprecipitated with anti-FLAG beads, and the co-precipitated endogenous PHR1 was detected by immunoblotting with anti-PHR1 antibody. Source data are provided as a Source Data file.

UCH2 regulates PSI gene expression and Pi homeostasis

To determine the role of UCH2 in regulating plant responses to Pi starvation, we examined the phenotypes of the uch2 mutant. Under P− conditions, the uch2 mutant exhibited a marked reduction in anthocyanin accumulation compared to wild-type control (WT, Col-0) (Fig. 2a, b). Consistent with this impaired anthocyanin accumulation, the expression of F’3H, a critical anthocyanin biosynthesis gene, was repressed in the uch2 mutant (Fig. 2d). We then determined the expression of several PSI genes in the uch2 mutant using RT-qPCR. Under P+ conditions, the transcript levels of IPS1, PHT1, 4 and PS2 showed no significant difference between WT and uch2 mutant. However, the expression of these genes was repressed in uch2 compared to WT under P− conditions (Fig. 2d).

Fig. 2. UCH2 positively regulates phosphate starvation responses.

a Phenotypes of Col-0, uch2, phr1, and phr1 uch2 double mutants grown for 7 days under P+ or P− conditions. Representative images from three independent experiments with similar results are shown. b, c Quantification of anthocyanin levels (b) and Pi content (c) in the shoots of the indicated genotypes under P+ and P− conditions. Data are presented as means ± s.d. (n = 3 biological replicates). Statistical significance was assessed using one-way ANOVA followed by Tukey’s two-sided multiple-comparisons test. Different letters indicate statistically significant differences (adjusted P < 0.05). Source data and exact adjusted P values are provided as a Source Data file. d Relative expression levels of representative phosphate starvation-induced (PSI) genes in seedlings grown under P+ and P− conditions. Data are means ± s.d. (n = 3 biological replicates). For each gene, statistical significance was assessed using two-way ANOVA followed by Tukey’s two-sided all-pairwise multiple-comparisons test. Different letters indicate statistically significant differences (adjusted P < 0.05). Source data and exact adjusted P values are provided as a Source Data file. e, f Volcano plot of differentially expressed genes in Col-0 seedlings under P− versus P+ conditions (e) and in uch2 mutants compared to Col-0 under P− conditions (f). Differentially expressed genes were identified with DESeq2 using a two-sided Wald test; P values were adjusted for multiple comparisons using the Benjamini–Hochberg (false-discovery-rate) procedure, and genes with |log₂ fold change| ≥ 0.5 and adjusted P ≤ 0.05 were considered differentially expressed. g A Venn diagram illustrating the overlap between PSI genes (n = 2653) and genes downregulated in uch2 vs Col-0 under P− conditions (n = 1742). The intersection defines UCH2-dependent PSI genes (n = 937; one-sided hypergeometric test, P = 1.79 × 0⁻⁵²⁹). h Venn diagram showing the overlap between UCH2-dependent PSI genes (n = 937) and PHR1-regulated genes (defined from published phr1 transcriptome data; see “Methods”). The intersection defines PHR1-UCH2 co-regulated PSI genes (n = 302; one-sided hypergeometric test, P = 2.04 × 10⁻³⁵⁴).

To better understand UCH2 function, we conducted transcriptome profiling by RNA sequencing (RNA-seq) in Col-0 and uch2 grown under both P+ and P− conditions. This analysis revealed 4796 genes that were significantly affected by Pi starvation in WT, with 2653 genes upregulated and 2143 downregulated (Fig. 2e). The 2653 upregulated genes are hereafter referred to as PSI genes. Meanwhile, 1081 genes were upregulated and 572 downregulated in the uch2 mutant compared to WT under P+ conditions (Supplementary Fig. 4). Under P− conditions, 1742 genes were downregulated and 1487 upregulated in the uch2 mutant compared to WT. Of the downregulated genes, more than half (937/1742) were PSI genes induced by Pi starvation in WT (Fig. 2g). These 937 genes are hereafter referred to as UCH2-dependent PSI genes. As expected, GO analysis confirmed that UCH2-regulated PSI genes were highly enriched for biological processes associated with cellular response to Pi starvation (Supplementary Fig. 5a), validating the specificity of our gene set. Approximately one-third (302/937) of the UCH2-regulated PSI genes were also regulated by PHR1, based on re-analysis of a previously published phr1 RNA-seq dataset³⁴ (Fig. 2h). This overlapping set is hereafter referred to as PHR1-UCH2 co-regulated PSI genes. GO analysis revealed that the PHR1-UCH2 co-regulated PSI genes were highly enriched for biological processes associated with Pi ion transport (Supplementary Fig. 5b). Furthermore, Pi content was significantly decreased in the uch2 mutant under P− conditions, but not under P+ conditions (Fig. 2c). Together, these results indicate that UCH2 regulates PSI gene expression and Pi homeostasis under Pi deficiency.

UCH2 regulation of PSI genes depends on PHR1

Since PHR1 is a central regulator of PSI gene expression and interacts with UCH2, we examined the role of PHR1 in UCH2-regulated Pi starvation responses. Although both phr1 and uch2 mutants accumulated less anthocyanin than Col-0, introduction of the uch2 mutation into the phr1 background did not further decrease anthocyanin content in the phr1 uch2 double mutant, suggesting that UCH2 and PHR1 function in the same pathway (Fig. 2a, b and Supplementary Fig. 3). The expression of F3’H gene was also consistent with this effect (Fig. 2d). The transcript levels of several PSI genes in uch2 phr1 double mutant were similar to those in phr1, but lower than those in the uch2 single mutant (Fig. 2d). Consistent with PSI gene expression patterns, Pi content in the uch2 phr1 double mutant was also similar to that in the phr1 mutant, but lower than that in uch2 (Fig. 2c). These results indicate that UCH2 regulates PSI gene expression and Pi homeostasis in a PHR1-dependent manner. Given that uch2 mutants exhibit PSR defects only under P− conditions, we tested whether Pi availability modulates the UCH2-PHR1 interaction. Endogenous Co-IP assays showed that UCH2 co-precipitated PHR1 to a similar extent under both P+ and P− conditions (Fig. 1e), indicating that the UCH2-PHR1 interaction is constitutive and not regulated by Pi availability.

Enzymatic activity is essential for PSI gene expression regulation by UCH2

Since UCH2 is a deubiquitinase, we investigated whether its enzymatic activity is required for PSI gene expression regulation. Yang et al. demonstrated that Cys-82 is the catalytic residue of UCH2, and that the C82S substitution abolishes DUB enzymatic activity³³. We introduced this catalytically inactive UCH2 mutant (UCH2^mDUB), fused to a FLAG tag and driven by its native promoter, into the uch2 mutant background (Supplementary Fig. 6a, b). Immunoblot analysis confirmed that the C82S substitution did not compromise UCH2 protein stability under either P+ or P− conditions (Supplementary Fig. 6b). Phenotypic analysis revealed that under P− conditions, anthocyanin accumulation in the UCH2^mDUB transgenic lines was markedly reduced compared to Col-0 and comparable to that observed in the uch2 mutant, but significantly lower than in complementation lines expressing wild-type UCH2 (Fig. 3a, b). Similarly, the UCH2^mDUB lines exhibited reduced shoot Pi content under P− conditions, mirroring the uch2 mutant phenotype and indicating a failure to restore Pi homeostasis (Fig. 3c). At the transcriptional level, representative PSI genes failed to recover wild-type expression patterns in the UCH2^mDUB lines, consistent with the requirement for UCH2 enzymatic activity in their proper induction under P− conditions (Fig. 3d). These data collectively demonstrate that the deubiquitinase activity of UCH2 is essential for its role in mediating the PSR. Together with the genetic and physical interaction between UCH2 and PHR1, our findings reveal that UCH2 functions through its catalytic activity to regulate transcriptional responses to Pi deficiency and promote plant adaptation to nutrient stress.

Fig. 3. UCH2 deubiquitinase activity is required for phosphate starvation responses.

a Phenotypes of Arabidopsis thaliana wild-type (Col-0), uch2 mutant, and transgenic complementation lines expressing wild-type UCH2 (pUCH2::UCH2-FLAG/uch2) or the catalytically inactive mutant (pUCH2::UCH2^mDUB-FLAG/uch2, C82S substitution) under P+ and P− conditions. The experiment was repeated independently three times, with similar results. Scale bar, 2 mm. b Quantification of anthocyanin levels in seedlings of the indicated genotypes under P+ and P− conditions. Data represent means ± s.d. (n = 3 biological replicates). Statistical significance was assessed using one-way ANOVA followed by Tukey’s multiple-comparisons test. Different letters indicate statistically significant differences (adjusted P < 0.05). Source data and exact adjusted P values are provided as a Source Data file. c Pi content in seedlings of the indicated genotypes grown under P+ and P− conditions. Data represent means ± s.d. (n = 3 biological replicates). Statistical significance was assessed using one-way ANOVA followed by Tukey’s multiple-comparisons test. Different letters indicate statistically significant differences (adjusted P < 0.05). Source data and exact adjusted P values are provided as a Source Data file. d Relative expression levels of representative PSI genes (F3’H, IPS1, PS2 and PHT1;4) in seedlings grown under P+ and P− conditions, as measured by RT-qPCR. Transcript levels were normalized to UBC. Data are means ± s.d. (n = 3 biological replicates). For each gene, statistical significance was assessed using two-way ANOVA followed by Tukey’s two-sided all-pairwise multiple-comparisons test. Different letters indicate statistically significant differences (adjusted P < 0.05). Source data and exact adjusted P values are provided as a Source Data file.

PHR1 recruits UCH2 to chromatin under Pi starvation

To investigate the response of UCH2 to Pi starvation, we initially examined whether it was transcriptionally regulated by Pi deficiency. RT–qPCR analysis showed that UCH2 transcript levels remained unchanged under P− conditions and were not altered in phr1 mutants compared to Col-0 under either P+ or P− conditions (Supplementary Fig. 7a). To determine whether PHR1 might regulate UCH2 post-transcriptionally, we introduced the phr1 mutation into the pUCH2::UCH2–FLAG transgenic line. Immunoblotting of UCH2-FLAG revealed no change in UCH2 protein levels in the phr1 mutant (Supplementary Fig. 7b, c), indicating that PHR1 does not regulate UCH2 expression at either the transcriptional or protein level.

Conversely, we also examined whether UCH2 affects PHR1 expression. RT-qPCR and immunoblot analyses showed that PHR1 transcript and PHR1 protein levels were comparable between Col-0 and uch2 mutants under both P+ and P− conditions (Supplementary Fig. 7d, e), indicating that UCH2 does not regulate PHR1 at the transcriptional or protein level, consistent with UCH2 acting downstream of, or in parallel with, PHR1 in the phosphate starvation response.

To identify the genome-wide binding sites of UCH2, we performed Cleavage Under Targets and Tagmentation (CUT&Tag) assays, an enzyme-tethering strategy that provides high-resolution sequencing libraries for efficient profiling of diverse chromatin components. In these assays, an anti-FLAG antibody was used to profile UCH2-bound chromatin from seedlings of the UCH2-FLAG transgenic line grown under P+ and P− conditions. Genome-wide peak calling identified a set of UCH2 binding peaks in two biological replicates under P− conditions (Supplementary Data 2). UCH2-FLAG enrichment was increased at 4983 loci under P− compared to P+ conditions (Fig. 4a, left panel). These peaks were located mainly in promoter regions, with 67.50% within 1 kb and 19.56% within 1–2 kb upstream of transcription start sites (TSSs) (Fig. 4b, upper panel). These P−-induced UCH2 peaks were associated with 7233 genes, hereafter referred to as P−-induced UCH2 targets.

Fig. 4. PHR1 recruits UCH2 to PSI gene promoters.

a Heatmaps showing CUT&Tag profiling of UCH2-FLAG occupancy in the indicated genotypes. Left: Genomic regions showing significantly increased UCH2 enrichment under P− compared to P+ conditions in Col-0 (n = 4983 peaks), indicating Pi-responsive UCH2 recruitment. Right: Genomic regions displaying significantly decreased UCH2 binding in phr1 mutants relative to Col-0 under P− conditions (n = 4435 peaks), revealing PHR1-dependent UCH2 chromatin association. b Genomic distribution of UCH2-binding peaks. Top: Annotation of P−-induced UCH2 peaks in Col-0 showing predominant promoter localization. Bottom: Annotation of PHR1-dependent UCH2 peaks showing similar promoter enrichment. c Venn diagram showing overlap between P−-induced UCH2 binding genes in Col-0 (n = 7233; P−-induced UCH2 targets) and genes associated with reduced UCH2 binding peaks in phr1 under P− conditions (n = 6465). The intersection identifies PHR1-dependent UCH2 targets (n = 6143; one-sided hypergeometric test, P = 2.05 × 10⁻⁴⁴⁵⁶). d Metagene profiles of UCH2-FLAG occupancy across the 6143 PHR1-dependent UCH2 targets (defined in c) under P+ and P− conditions in Col-0 and phr1. e A scatter plot showing the correlation between PHR1-FLAG and UCH2-FLAG CUT&Tag signal intensities at PHR1-UCH2 co-bound genomic loci under P− conditions in Col-0. Each point represents a binding region, color-coded by signal intensity. Strong positive correlation (two-sided Pearson’s R = 0.620, P < 0.01) indicates co-recruitment of UCH2 and PHR1 at these loci. f Box plots showing UCH2-FLAG enrichment at PHR1 target loci versus PHR1 non-target loci under P− conditions. Center line, median; box, interquartile range (25th–75th percentiles); whiskers, minimum to maximum. Statistical significance was assessed using a two-sided unpaired Student’s t-test. g Genome browser tracks showing representative PSI genes. UCH2-FLAG and PHR1-FLAG signals are co-enriched at promoter regions under P− conditions in Col-0 (orange and purple tracks). UCH2 binding is diminished in phr1 mutants (tan tracks), confirming PHR1-dependent recruitment. Gene models are shown below each track. The Col-0 (anti-FLAG) track shows the signal from the negative control.

We also introduced the phr1 mutation into the UCH2-FLAG line and analyzed UCH2 binding peaks. The phr1 mutation decreased UCH2-FLAG enrichment at 4435 peaks, which were associated with 6465 genes (Fig. 4a, right panel). These peaks were also located mainly in promoter regions, with 68.57% within 1 kb and 18.54% within 1–2 kb upstream of TSSs (Fig. 4b, lower panel). Among the 7233 P−-induced UCH2 targets, 6143 (84.9%) were PHR1-dependent, hereafter referred to as PHR1-dependent UCH2 targets (Fig. 4c). At these genes, Pi starvation greatly increased UCH2 binding around the TSSs (Fig. 4d), and the phr1 mutation significantly decreased UCH2 binding under P− conditions (Fig. 4d). These results indicate that Pi starvation enhances UCH2 recruitment to chromatin in a largely PHR1-dependent manner. Among the PHR1-dependent UCH2 targets, 239 were PSI genes (Supplementary Fig. 8a), representing a core set of Pi-responsive targets that depend on PHR1 for UCH2 recruitment. These 239 genes are hereafter referred to as PHR1-UCH2 co-bound PSI genes. The change in UCH2-FLAG signal intensity at TSS regions indicated that UCH2 enrichment at these 239 PSI genes was significantly reduced in the phr1 mutant under P− conditions (Supplementary Fig. 8b).

To further delineate the spatial relationship between UCH2 and PHR1, we investigated their genome-wide binding patterns. Pearson correlation analysis between UCH2-FLAG and PHR1-FLAG coverage revealed a high degree of co-localization between UCH2 and PHR1 at PSI gene promoters under P− conditions (Fig. 4e). Analysis of UCH2-FLAG binding coverage at all PHR1 target loci and PHR1 non-target loci showed that PHR1 target loci had a higher density of UCH2-binding signal under P− conditions, and this preference was compromised by the phr1 mutation and Pi replenishment (Fig. 4f). Representative examples of the preferential binding of UCH2 to PHR1 targets are shown in Fig. 4g. These results indicate that PHR1 tethers UCH2 to PSI gene promoters under P− conditions.

UCH2 promotes a reduction in H2A.Z occupancy at PSI gene promoters

The histone variant H2A.Z acts as a transcriptional repressor, keeping PSI genes silenced under P+ conditions; however, H2A.Z is depleted under P− conditions^21,22. Defects in H2A.Z removal may compromise PSI gene expression, reminiscent of the uch2 mutant phenotype. Therefore, we performed CUT&Tag profiling using an endogenous H2A.Z antibody in Col-0 and uch2 seedlings grown under P+ and P− conditions (Supplementary Data 2). Genome-wide peak calling identified 4257 peaks in Col-0 where H2A.Z enrichment decreased under P− compared to P+ conditions (Fig. 5a, left panel). These peaks were located mainly in promoter regions, with 68.32% within 1 kb and 17.38% within 1–2 kb upstream of TSSs (Fig. 5b, upper panel). These peaks were associated with 7343 genes. We also identified 3486 peaks with increased H2A.Z enrichment in the uch2 mutant compared to Col-0 under P− conditions, which were associated with 7342 genes (Fig. 5a, right panel). These peaks were also located mainly in promoter regions, with 66.84% within 1 kb and 18.64% within 1–2 kb upstream of TSSs (Fig. 5b, lower panel). Notably, nearly all genes that showed reduced H2A.Z occupancy under P− conditions in Col-0 overlapped with genes where H2A.Z enrichment was increased by the uch2 mutation under P− conditions, suggesting that UCH2 is actively involved in reduction in H2A.Z occupancy in response to Pi starvation (Fig. 5c). Among these genes, 338 were PSI genes, accounting for more than one third of the 937 PSI genes examined (Fig. 5d). Of the 239 PSI genes where UCH2 enrichment was increased under P− conditions in a PHR1-dependent manner (Supplementary Fig. 8), UCH2 was essential for reduction in H2A.Z occupancy at most of these loci (Fig. 5e).

Fig. 5. UCH2 promotes a reduction in H2A.Z occupancy at PSI gene promoters.

a Heatmaps showing genome-wide H2A.Z CUT&Tag profiles centered on transcription start sites. Left: H2A.Z occupancy in Col-0 under P+ and P− conditions, showing decreased enrichment at 4257 loci under P− conditions. Right: H2A.Z occupancy in uch2 mutants compared to Col-0 under P− conditions, revealing increased retention at 3486 loci. b Genomic distribution of differentially enriched H2A.Z peaks. Top: Annotation of P− responsive H2A.Z peaks in Col-0 (peaks with decreased enrichment under P− vs P+; n = 4257). Bottom: Annotation of UCH2-dependent H2A.Z peaks (peaks with increased enrichment in uch2 vs Col-0 under P−; n = 3486). c Venn diagram showing significant overlap between genes with reduced H2A.Z occupancy under P− conditions in Col-0 (n = 7343) and genes with increased H2A.Z occupancy in uch2 mutant under P− conditions (n = 7342). The intersection defines UCH2-dependent H2A.Z reduction genes (n = 7342; one-sided hypergeometric test, P = 4.93 × 10⁻⁶⁹¹³). d Venn diagram showing overlap between UCH2-dependent H2A.Z reduction genes (n = 7342) and UCH2 positive-regulated PSI genes (n = 937). The intersection defines UCH2-dependent H2A.Z PSI genes (n = 338; one-sided hypergeometric test, P = 1.37 × 10⁻¹⁰). e Venn diagram showing the intersection between PHR1-UCH2 co-regulated PSI genes (the 239 PSI genes within PHR1-dependent UCH2 targets; see Supplementary Fig .7a) and UCH2-dependent H2A.Z PSI genes (n = 338). The overlap (n = 149; one-sided hypergeometric test, P = 1.14 × 10⁻²³⁴) identifies PSI genes jointly regulated at both the UCH2 recruitment and reduction in H2A.Z occupancy levels during Pi starvation. f Metagene profiles of H2A.Z occupancy at UCH2-dependent H2A.Z reduction genes (defined in c) under P+ and P− conditions. g Metagene profiles of H2A.Z occupancy at the 149 PHR1-UCH2 jointly regulated PSI genes (defined in e) under P+ and P− conditions. h A scatterplot showing positive correlation between UCH2 and H2A.Z CUT&Tag signal intensities at promoters of the 149 PHR1-UCH2 jointly regulated PSI genes under P− conditions (two-sided Pearson’s R = 0.363, P = 0.047). i Genome browser snapshots illustrating H2A.Z enrichment at representative PSI genes.

Genome-wide, H2A.Z enrichment near TSSs was markedly reduced in Col-0 under P− conditions compared to P+ conditions (Supplementary Fig. 9a), and a similar trend was observed at PSI gene loci (Supplementary Fig. 9b). Interestingly, under P− conditions, uch2 mutants displayed significantly higher H2A.Z levels than Col-0 at both the genome-wide level and at PSI gene regions, suggesting that UCH2 promotes reduction in H2A.Z levels under P− conditions.

To further dissect the relationship between UCH2 and H2A.Z, we compared genomic regions where H2A.Z changes in uch2 overlapped with PHR1-dependent UCH2 binding sites under P− conditions. Approximately half of the UCH2-regulated H2A.Z peaks were located within regions targeted by UCH2 in a PHR1-dependent manner (Supplementary Fig. 10). These results support a model in which PHR1-dependent recruitment of UCH2 contributes to a reduction in H2A.Z occupancy at specific genomic loci under P− conditions. Consistent with this model, H2A.Z levels at PSI genes affected by UCH2 under P− conditions followed the expected trend: H2A.Z was reduced under low-Pi stress in Col-0 but remained elevated in uch2 (Fig. 5f). A similar pattern was observed for H2A.Z enrichment at genes co-regulated by PHR1 and UCH2 under P− conditions (Fig. 5g).

We further examined the spatial relationship between UCH2 and H2A.Z binding at these co-regulated loci. Correlation analysis revealed a strong overlap in their binding profiles at promoter regions under P− conditions (Fig. 5h), suggesting functional coupling between UCH2 recruitment and the reduction in H2A.Z occupancy. At representative PSI genes such as PS2, IPS1, and PHT1;4, H2A.Z enrichment was significantly reduced in WT plants under P− conditions but remained elevated in uch2 mutants (Fig. 5i). These results provide direct evidence that UCH2 promotes reduction in H2A.Z occupancy at key Pi-responsive genes and that this process depends on PHR1 under P− conditions.

To assess whether UCH2-mediated reduction in H2A.Z occupancy is linked to changes in chromatin accessibility, we compared our H2A.Z CUT&Tag data with a published ATAC-seq dataset from Col-0 seedlings under P+ and P− conditions³⁵. Genes showing UCH2-dependent H2A.Z reduction under P− significantly overlapped with Pi starvation-induced accessible chromatin regions (967 shared genes; Supplementary Fig. 11a). Metagene analysis further revealed that these genes with UCH2-dependent reduction in H2A.Z occupancy exhibited markedly elevated chromatin accessibility at their TSSs under P− conditions (Supplementary Fig. 11b), supporting a model in which UCH2-mediated reduction in H2A.Z levels contributes to local chromatin opening during Pi starvation.

UCH2 controls H2A.Z ubiquitination to fine-tune Pi starvation gene expression

To investigate how UCH2 promotes reduction in H2A.Z occupancy at PSI genes, we examined H2A.Z protein levels in Col-0 and uch2 mutants under both P+ and P− conditions.

Although Pi starvation did not affect total H2A.Z levels, it significantly decreased the level of ubiquitinated H2A.Z in Col-0 (Fig. 6a). However, UCH2 mutation blocked this reduction in H2A.Z ubiquitination (Fig. 6a). As UCH2 is a ubiquitin C-terminal hydrolase, we tested whether UCH2 could directly regulate H2A.Z ubiquitination. When UCH2 was transiently expressed in tobacco leaves, the ubiquitination level of FLAG-tagged H2A.Z (encoded by HTA9) was decreased, whereas co-expression with a catalytically inactive UCH2 mutant failed to reduce ubiquitination (Fig. 6b). To further validate that UCH2 catalytic activity is required for H2A.Z deubiquitination in Arabidopsis, we examined H2A.Zub levels in two independent UCH2^mDUB transgenic lines. Unlike wild-type UCH2, which fully restored H2A.Z deubiquitination under P− conditions, both UCH2^mDUB lines retained H2A. Zub levels comparable to those in the uch2 mutant under both P+ and P− conditions (Supplementary Fig. 12). This provides direct in planta evidence that the C82 residue is essential for UCH2-mediated H2A.Z deubiquitination. Together, these results indicate that UCH2 regulates H2A.Z deubiquitination in an enzymatic activity-dependent manner.

Fig. 6. UCH2 regulates phosphate starvation responses through deubiquitination of H2A.Z.

a Immunoblot analysis of HTA9 ubiquitination levels in the indicated genotypes under P+ and P− conditions. Right: Densitometric quantification of the proportion of ubiquitinated H2A.Z [HTA9ub/(HTA9ub + HTA9)]. Data are means ± s.d. (n = 3 biologically independent experiments). Different letters indicate statistically significant differences (one-way ANOVA followed by Tukey’s test, P < 0.05). Source data and exact adjusted P values are provided as a Source Data file. b UCH2 deubiquitinates HTA9 in a catalytic activity-dependent manner. Total protein extracts were analysed by immunoblotting with anti-FLAG and anti-MYC antibodies. Right: Densitometric quantification of the proportion of ubiquitinated H2A.Z. Data are means ± s.d. (n = 3 biologically independent experiments). Different letters indicate statistically significant differences (one-way ANOVA followed by Tukey’s test, P < 0.05). Source data and exact adjusted P values are provided as a Source Data file. c Phenotypes of Col-0, uch2, and transgenic lines expressing 35S::HTA9^RR-FLAG in the hta9 hta11 or hta9 hta11 uch2 backgrounds, grown under P+ and P− conditions. d, e Quantification of anthocyanin levels (d) and Pi content (e) in the shoots of the indicated genotypes under P+ and P− conditions. Data are means ± s.d. (n = 4 biologically independent replicates). Statistical significance was assessed using one-way ANOVA followed by Tukey’s two-sided multiple-comparisons test. Different letters indicate statistically significant differences (adjusted P < 0.05). Source data and exact adjusted P values are provided as a Source Data file. f Relative expression of PSI genes in indicated genotypes under P+ and P− conditions. Transcript abundance was normalized to UBC. Data are means ± s.d. (n = 3 biologically independent replicates). For each gene, statistical significance was assessed using two-way ANOVA followed by Tukey’s two-sided all-pairwise multiple-comparisons test. Different letters indicate statistically significant differences (adjusted P < 0.05). Source data and exact adjusted P values are provided as a Source Data file. g Proposed model illustrating how UCH2 regulates H2A.Z dynamics to control PSR in Arabidopsis thaliana.

H2A.Z can be monoubiquitinated at lysines K129 and K132 by PRC1 components^28,36. To test whether these specific ubiquitination sites are targets of UCH2, we overexpressed a FLAG-tagged HTA9 construct (FLAG-HTA9^RR), in which K129 and K132 were mutated to arginine residues, in the hta9 hta11 double mutant (Supplementary Fig. 13a). UCH2 was then mutated in the FLAG-HTA9^RR transgenic line using CRISPR/Cas9 (Supplementary Fig. 13a). Western blot analysis showed that the FLAG-HTA9^RR transgene was similarly expressed in both WT and uch2 mutant under P+ and P− conditions (Supplementary Fig. 13b). The FLAG-HTA9^RR transgenic line accumulated more anthocyanin than Col-0 under P− conditions (Fig. 6c,d). Pi content was also significantly increased in the FLAG-HTA9^RR line compared to Col-0 under P− conditions (Fig. 6e). However, the uch2 mutation did not affect anthocyanin accumulation or Pi content in the FLAG-HTA9^RR line, in contrast to what was observed in the Col-0 background (Figs. 2a, b and 6c, d).

Although the uch2 mutation repressed anthocyanin accumulation, PSI gene induction, and Pi content in Col-0, it failed to do so when K129 and K132 were mutated in HTA9 (Fig. 6d–f). Consistent with these physiological changes, transcript analysis showed enhanced expression of PSI genes, including F3’H, IPS1, PHT1;4, and PS2, in the FLAG-HTA9^RR line under P− conditions (Fig. 6f). The elevated expression of PSI genes was not repressed by the uch2 mutation, reinforcing the notion that UCH2 deubiquitinates H2A.Z to enhance PSI gene induction in response to Pi starvation.

HTA9^RR occupancy is reduced under Pi starvation

To understand why FLAG-HTA9^RR plants exhibit enhanced PSR only under P− conditions, we performed CUT&Tag to determine HTA9^RR genome-wide occupancy (Supplementary Data 2). Under P+ conditions, HTA9^RR showed robust chromatin enrichment, confirming that ubiquitination at K129/K132 is not required for H2A.Z deposition. Under P− conditions, however, HTA9^RR occupancy was significantly reduced compared to P+ conditions (Supplementary Fig. 14). This indicates that the non-ubiquitylatable H2A.Z variant is more susceptible to Pi starvation-induced depletion. These findings provide a mechanistic explanation for the enhanced PSI gene activation and PSR phenotypes observed specifically under P− conditions in FLAG-HTA9^RR lines.

Discussion

Our study reveals a regulatory module in the PSR pathway involving the transcription factor PHR1 and the deubiquitinase UCH2, which together orchestrate transcriptional activation through chromatin remodeling, particularly via modulation of H2A.Z ubiquitination (Fig. 6g). This reveals a key mechanism by which nutrient signaling is coupled to histone variant dynamics, establishing a signal-responsive chromatin regulatory axis in Arabidopsis.

UCH2 functions as a signal-responsive chromatin modifier

Our endogenous Co-IP data showed that the UCH2–PHR1 interaction is constitutive and largely unaffected by Pi availability (Fig. 1e), suggesting that these two proteins form a stable complex under P+ conditions. Pi signaling, therefore, does not primarily regulate complex assembly, but controls the chromatin recruitment or local activity of the UCH2–PHR1 module. This enables plants to respond rapidly to Pi limitation by deploying a pre-assembled deubiquitinase complex to target loci, without requiring de novo protein–protein interactions.

A long-standing question in PSR research has been how transcription factors such as PHR1 interface with chromatin to execute context-specific transcription^34,35,37. Our data establish that PHR1 not only recognizes target promoters but also scaffolds UCH2 recruitment under P− conditions (Fig. 4). UCH2-dependent H2A.Z reduction sites significantly overlapped with Pi starvation-induced accessible chromatin regions³⁵ (Supplementary Fig. 11), providing direct evidence that UCH2-mediated reduction in H2A.Z occupancy contributes to chromatin opening during Pi limitation. This is further supported by CUT&Tag profiling, which shows Pi-dependent co-occupancy of UCH2 and PHR1 at core PSI loci, including IPS1 and PHT1;4. This spatial convergence suggests that UCH2 activity is temporally restricted to PHR1-bound chromatin, thus avoiding global chromatin disruption.

The signal-responsive nature of UCH2 recruitment distinguishes it from constitutive chromatin regulators. Unlike global histone deubiquitinases that maintain basal chromatin states, our data indicate that UCH2 functions as an inducible modifier coupling environmental Pi availability to chromatin structure. This targeted recruitment contrasts with the recently characterized PEAT complex in Arabidopsis, which contains UBP5 as an H2A deubiquitinase that functions genome-wide to counter Polycomb-mediated repression^38,39. While UBP5 acts broadly to maintain genome-wide H2A ubiquitination homeostasis, UCH2 appears to operate in a locus-specific, signal-dependent manner, representing a mode of chromatin regulation tailored to nutrient response pathways.

A limitation of the present study is the absence of a validated antibody specific to monoubiquitinated H2A.Z (H2A.Zub), which would enable direct genome-wide profiling of H2A.Zub occupancy across genotypes and growth conditions. Sequential ChIP (ChIP-reChIP) using anti-H2A.Z followed by anti-ubiquitin antibodies⁴⁰ offers a promising alternative approach to map H2A.Zub genomic distribution and should be pursued in future studies.

Enzymatic activity couples nutrient sensing to chromatin dynamics

Our study further reveals that the enzymatic function of UCH2 is essential for its role in the PSR. Genetic complementation with the catalytically inactive UCH2^mDUB (C82S) mutant failed to rescue PSI gene expression or anthocyanin accumulation under P− conditions (Fig. 3), confirming that H2A.Z deubiquitination—not merely protein localization or scaffolding—is necessary for transcriptional activation. Prior studies have shown that H2A.Z stabilizes repressive nucleosomes under stress^28,41, and here we extend this view by demonstrating that Pi starvation induces a reduction in H2A.Z occupancy at PSI promoters in a UCH2-dependent manner (Fig. 5). This defines H2A.Z deubiquitination as a reversible regulatory node gating gene accessibility in response to environmental input. The requirement for catalytic activity places UCH2 among plant histone deubiquitinases with demonstrated chromatin-modifying functions. In Arabidopsis, UBP26 deubiquitinates H2B to regulate flowering time and seed development⁴², while OTLD1, another H2B deubiquitinase, controls plant growth and cell size^43,44. OTU1 has been shown to regulate seed and organ size through H2B deubiquitination⁴⁵. UCH2 targets H2A.Z in a nutrient-responsive manner, a substrate specificity that differs from these H2B-directed deubiquitinases. This substrate specificity is noteworthy given that H2A monoubiquitination in plants is more widespread and functionally diverse than previously appreciated^46,47, often functioning independently of the classical PRC2-mediated H3K27me3 pathway that dominates animal systems⁴⁶.

A notable feature of our experimental design is the use of a heterologous Nicotiana benthamiana transient expression system to demonstrate UCH2-mediated H2A.Z deubiquitination (Fig. 6b). We acknowledge that this system does not recapitulate the genomic context or the Pi starvation signal present in Arabidopsis. In the native context, UCH2 has access to its chromatin-bound substrate H2A.Zub is spatially and temporally restricted: PHR1 recruits UCH2 to specific PSI gene loci under Pi deficiency, thereby confining deubiquitination to the appropriate genomic locations and physiological conditions. This targeting mechanism confers both locus specificity and signal-dependence to the deubiquitination reaction, explaining why uch2 mutants exhibit PSR defects specifically under P− conditions. In the N. benthamiana transient system, however, both UCH2 and HTA9-FLAG are co-expressed at high levels from strong constitutive promoters. Under these conditions, the elevated local concentration of UCH2 protein enables it to encounter and deubiquitinate HTA9-FLAG through mass-action kinetics, independently of PHR1-mediated chromatin targeting. This behavior is well-precedented for histone deubiquitinases: the Polycomb repressive deubiquitinase PR-DUB (BAP1/Calypso) can deubiquitinate monoubiquitinated free histones and peptide substrates indiscriminately in vitro, yet its exquisite nucleosome-dependent substrate specificity in vivo depends on the ASXL/ASX co-factor and chromatin-targeting factors such as FOXK2 for locus-specific recruitment^48–50. The tobacco transient assay therefore reports on the intrinsic enzymatic capacity of UCH2 toward H2A.Z as a substrate, rather than its regulated, PHR1-dependent chromatin targeting in Arabidopsis. These two aspects of UCH2 function are complementary: the tobacco assay establishes the biochemical competence of UCH2 to deubiquitinate H2A.Z, while the Arabidopsis CUT&Tag and genetic data establish that this activity is deployed in a locus-specific, Pi starvation-dependent manner in vivo.

H2A.Z ubiquitination as a regulatory switch

Our biochemical and epistasis analyses establish a functional hierarchy whereby UCH2-mediated H2A.Z deubiquitination derepresses PSI gene expression. Notably, H2A.Z ubiquitination levels were elevated in uch2 mutants even under low-Pi stress (Fig. 6a), and plants expressing a non-ubiquitylatable HTA9 variant were insensitive to UCH2 loss (Fig. 6c–f), establishing H2A.Z as a direct target of UCH2 in vivo. The ability of UCH2 to remove ubiquitin from H2A.Z in transient assays (Fig. 6b) further confirms this direct enzymatic relationship. These findings position histone variant ubiquitination as a reversible chromatin switch downstream of nutrient sensing.

In Arabidopsis, the Polycomb-group protein AtBMI1 catalyzes H2A.Z monoubiquitination at K129 and K132, and this modification is essential for H2A.Z-mediated transcriptional repression²⁸. Our work reveals the opposing activity: UCH2-mediated deubiquitination removes these marks to enable gene activation. This dynamic equilibrium between BMI1-mediated ubiquitination under nutrient-replete conditions and UCH2-mediated deubiquitination under Pi starvation provides a molecular explanation for the rapid transcriptional responses observed during low-Pi stress. This mechanism is conceptually analogous to the role of BAP1 (a UCH family member) in H2A deubiquitination in metazoans^51–53, suggesting ancient evolutionary origins for UCH-mediated chromatin regulation. However, the coupling of UCH activity to nutrient-responsive transcription factors appears to be a plant-specific innovation that may reflect the sessile nature of plants and their need for sophisticated environmental sensing systems.

CUT&Tag profiling of HTA9^RR showed deposition levels comparable to wild-type H2A.Z under P+ conditions, but a greater reduction in occupancy under P− conditions. This pattern suggests that monoubiquitination at K129/K132 is not required for H2A.Z chromatin incorporation, but instead acts as a retention signal during Pi starvation-induced remodeling. UCH2-mediated deubiquitination removes this mark to facilitate the reduction in H2A.Z occupancy at PSI loci. The absence of this ubiquitin anchor in HTA9^RR would therefore render H2A.Z more susceptible to eviction, leading to stronger PSI gene derepression observed in FLAG-HTA9^RR plants.

Distinct H2A.Z distribution at PSI genes

At PHR1-UCH2 co-regulated PSI genes, H2A.Z displays a bimodal enrichment pattern with peaks at both the TSS and TES (Fig. 5f, g), which is a distribution not observed genome-wide (Supplementary Fig. 9a). This pattern suggests that PSI genes possess a distinct chromatin architecture under P+ conditions. This may reflect H2A.Z-mediated gene looping between the TSS and TES, maintaining PSI genes in a repressed but poised state. Alternatively, TES-associated H2A.Z may be linked to co-transcriptional regulation, including antisense transcription or non-coding RNA production at loci such as IPS1. The coordinated reduction in H2A.Z occupancy at both regions by UCH2 under P− conditions would then allow synchronized de-repression of both transcription initiation and elongation. Future chromatin conformation capture and strand-specific transcriptome studies will be needed to distinguish between these possibilities.

Specificity and selectivity in histone deubiquitination

A key question arising from our work is how UCH2 achieves substrate specificity for H2A.Z rather than canonical H2A or other ubiquitinated histones. Several mechanisms may contribute to this selectivity. First, the interaction with PHR1 is spatially restricted, directing UCH2 specifically to PSI gene promoters enriched in H2A.Z. Second, H2A.Z itself possesses structural features that facilitate preferential recognition by UCH2. The C-terminal extension of H2A.Z differs from canonical H2A and contains the ubiquitination sites (K129/K132 in Arabidopsis) that serve as UCH2 substrates. Third, local chromatin features at PHR1 targets, including histone marks and remodeling factors, may help specify UCH2 activity.

The landscape of plant histone deubiquitinases shows clear functional diversification, shaped by both substrate preference and regulatory context^29,54. UBP26 and OTLD1 act mainly on H2B monoubiquitination^42–44, influencing processes from flowering time to organ growth, whereas UBP5 functions as a major genome-wide H2A deubiquitinase that counteracts Polycomb repression³⁹. OTLD1 requires recruitment by DNA-binding transcription factors such as LSH10 to access its chromatin targets⁵⁵, hinting that transcription factor-mediated guidance may be a broader principle for this enzyme family. In light of these observations, UCH2 nutrient-responsive specificity for H2A.Z adds an additional layer of regulatory complexity and suggests that other stimulus-coupled histone deubiquitinases may yet remain undiscovered.

Although our work establishes a clear PHR1–UCH2–H2A.Z regulatory axis in Arabidopsis, its conservation across crop species remains to be determined. Given the conserved functions of PHR1-like transcription factors in rice and maize^56–58, and the emerging functions of H2A.Z in nutrient signaling^20,22,59, this mechanism is likely to have broader relevance. Future work incorporating tissue-specific epigenomic profiling and comparative genomics will be essential to evaluate the evolutionary conservation of UCH2-mediated chromatin regulation and to uncover additional regulatory components.

From an agricultural perspective, the PHR1–UCH2–H2A.Z axis represents a promising target for improving crop phosphorus use efficiency. While traditional breeding and genetic engineering approaches have focused primarily on Pi transporters and root architecture, our work suggests an alternative strategy: modulating chromatin-based regulation of PSI genes to optimize the transcriptional response to Pi limitation. Targeted modifications of UCH2 activity or its recruitment efficiency could potentially accelerate or amplify PSI gene induction under P− conditions while preserving normal growth under P+ conditions.

In conclusion, our findings establish that PHR1 recruits UCH2 to deubiquitinate H2A.Z at PSI promoters under P− conditions, promoting reduction in H2A.Z occupancy and nucleosome destabilization, thereby enabling gene activation. This integrative module bridges nutrient signaling with chromatin remodeling (Fig. 6g), positioning H2A.Z ubiquitination as a dynamically regulated epigenetic mark in environmental adaptation. In addition to advancing our mechanistic understanding of the PSR, our study establishes a conceptual framework for signal-responsive chromatin regulation in plants. The PHR1–UCH2–H2A.Z module exemplifies how transcription factors act not only as DNA-binding proteins but also as platforms for recruiting chromatin-modifying enzymes, thereby coupling environmental cues to epigenetic reprogramming. Furthermore, this work offers a molecular framework for nutrient stress responses in plants and a conceptual foundation for harnessing chromatin-based regulation to improve crop resilience and phosphorus use efficiency in sustainable agriculture.

Methods

Plant materials and growth conditions

Arabidopsis thaliana ecotype Columbia-0 (Col-0) was used as the wild-type background. The uch2 mutant used in this study is a CRISPR/Cas9-generated loss-of-function allele kindly provided by Dr. Xinjian He (National Institute of Biological Sciences, Beijing, China). The allele carries a single-nucleotide insertion within the second exon of UCH2, causing a frameshift and premature stop codon that truncates the UCH2 protein within the catalytic domain, resulting in complete loss of protein function⁶⁰. The phr1 mutant (SALK_067629) was obtained from the Arabidopsis Biological Resource Center (ABRC). The phr1uch2 double mutant was generated by crossing the two single mutants and confirmed by PCR-based genotyping. Plants were grown on half-strength Murashige and Skoog (½ MS) agar medium supplemented with 1.0% (w/v) sucrose and solidified with 1.1% (w/v) agar, under a 16 h light / 8 h dark photoperiod at 23 °C. The P+ (Pi-sufficient) and P− (Pi-deficient) media contained 625 µM and 50 µM KH₂PO₄, respectively.

Generation of mutants and transgenic plants

To generate UCH2-FLAG and PHR1-FLAG constructs, genomic fragments of UCH2 and PHR1, including their native promoters, were amplified and cloned into a modified pCAMBIA1300-FLAG vector. The resulting constructs were introduced into the uch2 and phr1 mutant backgrounds, respectively, via the floral dip method using Agrobacterium tumefaciens strain GV3101. To obtain UCH2-FLAG in the phr1 background, the UCH2-FLAG transgenic line was crossed with the phr1 mutant, and homozygous plants were selected in the F2 generation. A catalytically inactive variant, UCH2^mDUB-FLAG (C82S), was generated by site-directed mutagenesis and transformed into the uch2 mutant background using Agrobacterium-mediated transformation.

For the HTA9^RR-FLAG construct, the HTA9 cDNA was mutated to substitute lysines K129 and K132 with arginine (R), driven by the CaMV 35S promoter, and cloned into the modified pCAMBIA1300-FLAG vector, then transformed into the hta9 hta11 double mutant. The hta9 hta11 double mutant was kindly provided by Yuda Fang (Shanghai Jiao Tong University, Shanghai, China). Subsequently, CRISPR/Cas9-mediated mutagenesis was used to disrupt UCH2 in the 35S::HTA9^RR-FLAG/hta9 hta11 background. Primer sequences used for vector construction are listed in Supplementary Table 1.

Yeast two-hybrid assay

For yeast two-hybrid (Y2H) assays, the full-length, N-terminal, and C-terminal fragments of PHR1 were amplified and fused either to the GAL4 activation domain (AD, pGADT7) or the GAL4 DNA-binding domain (BD, pGBKT7). Coding sequences of UCH1, UCH2, and UCH3 were cloned into the corresponding reciprocal vectors. Constructs were co-transformed into Saccharomyces cerevisiae strain AH109 using the lithium acetate method. Transformants were selected on SD/-Leu/-Trp medium, and protein–protein interactions were assessed on SD/-Leu/-Trp/-His/-Ade medium. Primer sequences used for plasmid construction are provided in Supplementary Table 1.

Split-luciferase complementation assay

For split-luciferase (Split-Luc) complementation assays, the coding sequence of UCH2 was cloned into the nLuc vector, and PHR1 was cloned into the cLuc vector. The resulting constructs were introduced into Agrobacterium tumefaciens strain GV3101 and co-infiltrated into N. benthamiana leaves at an optical density (OD₆₀₀) of 0.6. Forty-eight hours after infiltration, 1 mM D-luciferin in 0.01% (v/v) Triton X-100 was sprayed onto the leaves, and luminescence images were captured using a plant in vivo imager (PyLoN1300B, Princeton Instruments). Primer sequences used for vector construction are provided in Supplementary Table 1.

BiFC assays

For bimolecular fluorescence complementation (BiFC), PHR1 and UCH2 coding sequences were fused to the N- or C-terminal halves of YFP in pSPYNE/pSPYCE vectors. Constructs were transiently expressed in N. benthamiana leaves through Agrobacterium-mediated transformation. Fluorescence was examined 48 h after infiltration using a Zeiss LSM 880 confocal microscope. Primer sequences used for construct generation are listed in Supplementary Table 1.

Co-immunoprecipitation (Co-IP)

For co-immunoprecipitation (Co-IP) assays, Agrobacterium tumefaciens strains carrying PHR1-FLAG and UCH2-MYC were co-infiltrated into N. benthamiana leaves. After 48 h, total proteins were extracted from 0.5 g of leaf tissue with 3 ml of immunoprecipitation (IP) buffer (50 mM Tris–HCl, pH 7.5, 150 mM NaCl, 10% glycerol, 0.5% NP-40, 1 mM EDTA, and protease inhibitor cocktail). Homogenates were filtered through two layers of Miracloth and centrifuged at 13,000 × g for 15 min at 4 °C. The resulting supernatant was incubated with 10 μL of anti-FLAG M2 magnetic beads (Sigma-Aldrich, M8823, lot SLBJ3044V) for 3 h at 4 °C with gentle rotation. Beads were washed five times with ice-cold IP buffer, and bound proteins were eluted by boiling in 50 μl of SDS loading buffer for 10 min. Immunoprecipitated proteins were analyzed by immunoblotting using anti-FLAG (Sigma-Aldrich, F1804, clone M2, lot SLBH1183V; 1:1000) and anti-MYC (ABclonal, AE070, clone ARC5004-12, lot 6100012056; 1:4000) antibodies. Primer sequences are listed in Supplementary Table 1.

For endogenous Co-IP assays in Arabidopsis, pUCH2::UCH2–FLAG/uch2 and Col-0 seedlings were grown under P+ or P− conditions for 7 days. Total proteins were extracted from 1 g of seedling tissue using the same IP buffer, immunoprecipitated with 10 μL of anti-FLAG M2 magnetic beads (Sigma-Aldrich, M8823, lot SLBJ3044V), and subjected to immunoblot analysis with anti-FLAG (Sigma-Aldrich, F1804, clone M2, lot SLBH1183V; 1:1000) and anti-PHR1 (1:1000) antibodies. The anti-PHR1 antibody was a kind gift from Prof. Dong Liu at Tsinghua University. It was raised against the N-terminal region of Arabidopsis PHR1 (residues 1–224)⁶¹. The specificity of this antibody was confirmed by immunoblotting using phr1 null mutant material (Supplementary Fig. 2).

Immunoprecipitation–mass spectrometry (IP–MS)

For immunoprecipitation (IP) coupled with mass spectrometry (MS), transgenic Arabidopsis seedlings expressing pPHR1::PHR1-FLAG/phr1 were grown under P− conditions for 7 days. Total proteins were extracted from 1 g of seedling tissue with 5 ml of IP buffer (50 mM Tris–HCl, pH 7.5, 150 mM NaCl, 10% glycerol, 0.5% NP-40, 1 mM EDTA, and protease inhibitor cocktail). Homogenates were filtered through two layers of Miracloth and centrifuged at 13,000 × g for 15 min at 4 °C. The supernatant was incubated with 10 μL of anti-FLAG M2 magnetic beads (Sigma-Aldrich, M8823, lot SLBJ3044V) for 4 h at 4 °C with gentle rotation. Beads were washed five times with ice-cold IP buffer, and bound proteins were eluted by boiling in 50 μl of SDS loading buffer for 10 min. The immunoprecipitated proteins were resolved on SDS–PAGE gels and subjected to in-gel trypsin digestion, followed by LC–MS/MS analysis on an Orbitrap Fusion Lumos mass spectrometer (Thermo Fisher Scientific). Peptide identification was performed against the TAIR10 protein database using the Mascot search engine. Search parameters were: trypsin/P with up to two missed cleavages; monoisotopic masses; carbamidomethyl (C) as a fixed modification; acetyl (protein N‑terminus), oxidation (M) and phospho (S/T and Y) as variable modifications; a precursor mass tolerance of ±20 ppm; and a fragment mass tolerance of ±0.2 Da. Peptide matches were accepted at a significance threshold of P < 0.05 with an ion score cut‑off of 15; the corresponding decoy‑database false‑discovery rates were 3.0% for matches above the identity threshold and 5.6% for matches above the homology‑or‑identity threshold. The search identified 1421 protein groups. This identification was based on a single IP–MS run (n = 1) without a separate control immunoprecipitation. The complete Mascot search report is provided as Supplementary Data 1.

Anthocyanin measurement and phosphate content determination

For phosphate (Pi) content determination, tissues from 10-day-old seedlings were collected in 1.5 ml tubes, snap-frozen in liquid nitrogen, and ground to a fine powder using metal beads. Pi levels were quantified using a phosphomolybdate colorimetric assay. Ground tissue was extracted with 250 μl of 1% glacial acetic acid, vortexed, subjected to three freeze–thaw cycles in liquid nitrogen, and centrifuged at 13,000 rpm for 1 min at room temperature. 50 μl of supernatant was mixed with 250 μl H₂O and 700 μl of freshly prepared reaction mixture (100 μl of 10% ascorbic acid and 600 μl of 0.42% ammonium molybdate tetrahydrate in 1 N H₂SO₄). After incubation at 45 °C for 20 min, absorbance was measured at 820 nm and Pi concentration was calculated as nmol Pi per mg fresh weight⁶².

For anthocyanin measurement, pigments were extracted in methanol containing 1% (v/v) HCl. Chlorophyll was removed by mixing with an equal volume of chloroform, and the absorbance of the aqueous phase was recorded. Relative anthocyanin content was calculated as (A_530– A ₆₅₇) per gram fresh weight⁶³.

Quantitative RT–PCR

Total RNA was extracted using the AFTSpin Complex Plant Fast RNA Extraction Kit (RK30122; ABclonal) and treated with DNase I following the manufacturer’s instructions. First-strand cDNA was synthesized from 1 μg of total RNA using the HiScript IV All-in-One Ultra RT SuperMix (R433; Vazyme Biotech Co., Ltd.). Quantitative PCR was performed on a CFX96 Real-Time PCR System (Bio-Rad) with the 2× Universal SYBR Green Fast qPCR Mix (RK21203; ABclonal). Transcript levels were normalized to UBC, and relative expression was calculated using the 2^–ΔΔCt method. Primer sequences used for RT–qPCR are listed in Supplementary Table 1.

RNA-seq and transcriptome analysis

Seedlings were grown under P+ or P− conditions for 7 days, and total RNA was extracted using the AFTSpin Complex Plant Fast RNA Extraction Kit (RK30122; ABclonal). RNA-seq libraries were prepared using the Fast RNA-seq Library Prep Kit V2 (RK20306; ABclonal) and sequenced on the MGI platform to generate paired-end reads. Raw reads were trimmed with TrimGalore (version 0.6.10), aligned to the Arabidopsis thaliana TAIR10 reference genome using HISAT2 (version 2.1.0), and quantified with featureCounts. Differentially expressed genes were identified with DESeq2 (thresholds: |log₂FC| ≥ 0.5 and FDR ≤ 0.05). Gene Ontology enrichment analysis was performed using the enrichGO function in the clusterProfiler package (v4.16.4).

Western blotting

For immunoblot analysis of UCH2-FLAG, UCH2^mDUB-FLAG, HTA9^RR-FLAG, HTA9^RR/uch2-FLAG, and H2A.Z, seedling tissues were ground to a fine powder in liquid nitrogen and extracted in 5% (w/v) SDS. Extracts were boiled at 95 °C for 5 min and centrifuged at 13,000 × g for 10 min. The supernatant was mixed with 5× SDS loading buffer and resolved on 12.5% SDS–PAGE gels. Proteins were detected by immunoblotting with the following primary antibodies: anti-FLAG (Sigma-Aldrich, F1804, clone M2, lot SLBH1183V; 1:1000), anti-H2A.Z (Agrisera, AS10 718, lot 2011; 1:1000), anti-PHR1 (1:1000; described above) and anti-H3 (ABclonal, A2348, lot 5500032947; 1:5000). HRP-conjugated goat anti-mouse/rabbit IgG (Abmart, M21003, lot 324677; 1:10,000) was used as the secondary antibody.

CUT&Tag and data analysis

Cleavage Under Targets and Tagmentation (CUT&Tag) assays were performed using the NovoNGS CUT&Tag High-Sensitivity Kit (N259; Novoprotein) according to the manufacturer’s instructions. Nuclei were isolated from seedlings grown under P+ or P− conditions for 7 days. The isolated nuclei were immobilized on 10 μL concanavalin A beads (N259; Novoprotein) and incubated with primary antibodies: anti-FLAG (Sigma-Aldrich, F1804, clone M2, lot SLBH1183V; 1 μg per sample) for UCH2-FLAG, PHR1-FLAG, UCH2-FLAG/phr1 and HTA9^RR-FLAG samples (as a negative control, CUT&Tag using the anti-FLAG antibody was also performed on non-transgenic Col-0 seedlings grown under P− conditions; no significant enrichment over background was detected in the Col-0 control) and anti-H2A.Z (Agrisera, AS10 718, lot 2011; 1 μg per sample). After washing, the samples were incubated with a rabbit anti-mouse IgG secondary antibody (Abcam, ab46540, lot 0351218; 1:100), followed by pA/G-Tn5 transposase (supplied in kit N259). Tn5 was activated by adding 0.5 μL 1 M Mg²⁺ in 50 μL ChiTag buffer at 37 °C for 1 h. Reactions were stopped by adding 1 μL 10% SDS and 1 μL of proteinase K (supplied in kit N259) and incubated at 55 °C for 10 min. Tagmented DNA was purified using Tagment DNA Extract Beads (N245; Novoprotein), and sequencing libraries were prepared with the NovoNGS CUT&Tag Kit (N259; Novoprotein) according to the manufacturer’s instructions.

Sequencing libraries were subjected to paired-end sequencing on an Illumina NovaSeq 6000 platform at Genergy Biotechnology Co., Ltd. (Shanghai, China). Raw reads were processed with TrimGalore (version 0.6.10) to remove adapter sequences and low-quality bases, and subsequently aligned to the Arabidopsis thaliana TAIR10 reference genome using Bowtie2 with the --very-sensitive parameter. Peak calling was performed with MACS2 (q < 0.05; --nomodel --call-summits). Normalized coverage tracks were generated from deduplicated BAM files using deepTools bamCoverage with --normalizeUsing CPM (Counts Per Million) to account for differences in sequencing depth across samples. For differential binding analysis, BAM files and MACS2 narrowPeak files were imported into DiffBind (v3.16.0), and read counts within consensus peak regions were obtained using dba.count with summits = 250. Differential enrichment between sample groups was assessed using the edgeR method, with significantly differentially bound peaks defined at FDR < 0.05. Peaks were annotated using ChIPseeker (v1.42.1). Heatmaps and metagene profiles were generated with deepTools computeMatrix and plotHeatmap, and peak coverage was calculated using bedtools. Pearson correlations were computed using the cor.test function in R (v4.4.0). Genome-wide signal tracks were visualized in IGV.

H2A.Z ubiquitination assay

For ubiquitination assays in Nicotiana benthamiana, constructs expressing HTA9–FLAG together with UCH2-MYC, UCH2^mDUB-MYC (C82S), or pCAMBIA1300-MYC vector were co-infiltrated into 4-week-old N. benthamiana leaves via Agrobacterium tumefaciens strain GV3101 (OD₆₀₀ = 0.6) using the syringe infiltration method. Leaf tissue was harvested 48 h post-infiltration, ground in liquid nitrogen, and homogenized in protein extraction buffer (50 mM Tris-HCl, pH 7.5, 150 mM NaCl, 10% glycerol, 0.5% NP-40, 1 mM EDTA, and protease inhibitor cocktail). The extracts were centrifuged at 13,000 × g for 10 min at 4 °C, and the supernatants were mixed with 5× SDS loading buffer and boiled for 10 min. Proteins were resolved on 12.5% SDS-PAGE gels and analyzed by immunoblotting using anti-FLAG (Sigma-Aldrich, F1804, clone M2, lot SLBH1183V; 1:1000) to detect HTA9-FLAG and anti-MYC (ABclonal, AE070, clone ARC5004-12, lot 6100012056; 1:4000) to detect UCH2 and its variants. Primer sequences used for construct generation are provided in Supplementary Table 1.

Statistics and reproducibility

Data are presented as mean ± s.d. unless otherwise stated. Statistical significance was assessed using two-sided Student’s t-tests for pairwise comparisons or one-way ANOVA followed by Tukey’s or Dunnett’s multiple-comparisons test, as indicated in the corresponding figure legends. For Dunnett’s tests, the indicated reference genotype was used as the control group. Multiple-comparison-adjusted P values were used to determine statistical significance, with adjusted P < 0.05 considered significant. All statistical analyses were performed using GraphPad Prism 9. For all bar plots, n refers to biologically independent replicates or independent experiments, as specified in the figure legends. Venn diagram overlaps were assessed using the hypergeometric test. Exact P values are provided in the Source Data where appropriate. The plant materials, plasmids and other reagents generated in this study are available from the corresponding author upon reasonable request.

Reporting summary

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

Author contributions

M.L. conceived and supervised the project. F.M., X.X., and G.M. performed most of the experiments and analysed the data with M.L. Q.W., Y.Z., H.C., X.L., and J.J. assisted with the experiments and discussed the results. F.M., V.S., and M.L. wrote the manuscript. All authors read and approved the final version of the manuscript.

Peer review

Peer review information

Nature Communications thanks the anonymous reviewers for their contribution to the peer review of this work. A peer review is available.

Funding

This work was supported by the National Natural Science Foundation of China (32570313 and U24A20385 to M.L.) and by the Interdisciplinary Research Project of Hangzhou Normal University (2025JCXK01 to M.L.).

Data availability

The CUT&Tag and RNA-seq sequencing data generated in this study have been deposited in the NCBI database under BioProject accession PRJNA1353614. Previously published ATAC-seq data reused in this study are available in the NCBI database under accession PRJNA717818. The mass spectrometry proteomics identifications supporting the identification of PHR1-associated proteins are provided as the Mascot database-search report (Supplementary Data 1). Source data are provided with this paper.

Competing interests

The authors declare no competing interests.

Supplementary information

The online version contains supplementary material available at https://doi.org/10.1038/s41467-026-77882-6.