Protection of telomeres 1b safeguards the Arabidopsis genome by regulating ROS homeostasis

Abstract

Telomeres safeguard the genome, acting as sentinels of oxidative stress and preventing chromosome ends from eliciting a DNA damage response. PROTECTION OF TELOMERES 1 (POT1) is a highly conserved telomere protein, essential for chromosome integrity and telomeric DNA replication. Arabidopsis thaliana encodes two divergent POT1 paralogs: AtPOT1a stimulates telomerase activity, but AtPOT1b function is unknown. Here we show that AtPOT1b modulates reactive oxygen species (ROS) homeostasis. Oxidative stress induces AtPOT1b expression and telomeric accumulation, while AtPOT1b inactivation elevates ROS, increases telomeric and genome-wide oxidation, and causes stochastic telomere length changes. To address how AtPOT1b controls ROS, we report its localization in nuclei and peroxisomes, and association with catalases and peroxidases that enhance ROS scavenging. Impairing AtPOT1b-CAT2 interaction increases ROS accumulation and telomeric oxidation. Moss or human POT1 rescues ROS overaccumulation in Arabidopsis pot1b mutants, but not telomere deficiency in pot1a pot1b mutants, supporting a conserved role for POT1 in modulating ROS homeostasis and genome stability, distinct from canonical telomeric functions.

The authors uncover a noncanonical role for Arabidopsis POT1b in maintaining genome stability through ROS regulation, establishing an unexpected connection between telomere biology and cellular oxidative stress.

Introduction

Telomeres are the terminal caps of linear chromosomes, which in somatic cells are naturally depleted after each cycle of DNA replication. However, in cells with long-term proliferative capacity (stem and germ cells and most cancer cells), telomere length is maintained, primarily via telomerase. Telomerase functions with shelterin (TRF1, TRF2, RAP1, TIN2, TPP1, and POT1 in mammals), a telomere-binding complex that blocks chromosome ends from inciting a DNA damage response^1,2. Telomere protection is postulated to reflect a genetic clock for lifespan, as it determines cellular proliferative capacity and is linked to aging in certain species³.

Reactive oxygen species (ROS), natural by-products of oxygen metabolism, act as signaling molecules in development but can over-accumulate under stress, leading to oxidation of DNA and other biomolecules^4,5. DNA oxidation is mutagenic and contributes to genome instability^6–8. Thus, ROS accumulation contributes to a metabolic clock for lifespan^4,5.

Telomeres are proposed to integrate genetic and environmental cues as sentinels of oxidative stress^6–8. While telomere attrition in response to oxidative damage is well documented⁶, human telomerase reverse transcriptase (TERT) and shelterin component TIN2 accumulate in mitochondria, promoting mitochondrial function^9–12. Moreover, mitochondrial-targeted damage specifically triggers telomere dysfunction, supporting retrograde signaling between mitochondria and telomeres¹³. TRF1, TRF2, and POT1 stimulate base-excision repair, the primary pathway for removing oxidized DNA bases 8-oxo guanine (8-oxoG) and thymine glycol (Tg)^14,15. In addition, TRF1 stabilization in response to ROS is postulated to protect telomeres from oxidative damage¹⁶. Despite these intriguing findings, the role of telomere-associated proteins in the oxidative stress response remains unclear, as their essential telomeric roles make it challenging to define potential redox functions.

POT1 is the most conserved member of shelterin, crucial for telomere length regulation and the prevention of an ATR-dependent DNA damage response^17–22. Human POT1 (hPOT1) interacts with telomeric DNA via two N-terminal oligonucleotide/oligosaccharide (OB) folds, is recruited to telomeres via TPP1, and protects the telomeric double-stranded-single-stranded DNA junction^19,21,22. hPOT1 also regulates telomerase by enhancing enzyme processivity or blocking access to chromosome ends^18,20. Loss or mutation of hPOT1 disrupts telomere length regulation and leads to chromosomal defects, including fragile telomeres, end-to-end chromosome fusions and micronuclei^17–19.

Most species harbor a single POT1 gene, but in cases of POT1 duplication, functional divergence or sub-functionalization occurs^23–27. For example, murine POT1a inhibits ATR activation and modulates telomere length, whereas POT1b controls chromosomal resection during telomere replication^2,23. While most plant species possess a single copy POT1^28,29 Arabidopsis thaliana carries three POT1-related genes: AtPOT1a, AtPOT1b, and AtPOT1c³⁰. AtPOT1c was silenced soon after gene duplication³⁰, but AtPOT1a and AtPOT1b are transcriptionally active, highly divergent paralogs that share only 52% amino acid sequence similarity. Unlike the human and fission yeast POT1 proteins, AtPOT1a is not a stable component of telomeres³¹, but instead serves as a positive regulator of telomerase, interacting with TERT and stimulating its repeat addition processivity³². Plants lacking AtPOT1a suffer progressive telomere shortening in successive generations³¹, similar to tert mutants^33,34.

Less is known about AtPOT1b. Overexpression of its first OB-fold (OB1) triggers massive chromatin bridges during anaphase, implicating AtPOT1b in genome stability²⁶. Unexpectedly, however, a single amino acid substitution in the DNA-binding pocket of OB1 impairs AtPOT1b interaction with telomeric DNA³⁵. Perhaps as a consequence, AtPOT1b fails to complement the telomere maintenance defect associated with AtPOT1a deficiency²⁷. Evolutionary analyses show the Brassicaceae POT1a lineage is under positive selection to reinforce its interaction with the CTC1/STN1/TEN1 (CST) telomeric DNA replication complex²⁷. In contrast, POT1b is not subjected to the same selective pressure, arguing that its function(s) are ancestral.

Here, we exploit the natural separation-of-function in AtPOT1a and AtPOT1b paralogs to explore telomere-independent functions of POT1. We report that AtPOT1b is a telomeric protein that is both regulated by and regulates ROS levels. Dually localized in nuclei and peroxisomes, AtPOT1b engages ROS scavenging enzymes, particularly catalase, and disruption of this interaction triggers telomere oxidation and elevated ROS. We further show that ROS regulation is an evolutionarily conserved feature of POT1, promoting genome integrity through redox homeostasis.

Results

AtPOT1a and AtPOT1b act synergistically to maintain genome integrity

We explored AtPOT1b function by generating a CRISPR-Cas9 null allele in wild type (WT) Col-0 with an 88 bp deletion in exon 2 (pot1b), which dramatically reduced AtPOT1b mRNA accumulation (Supplementary Fig. 1a, b). The pot1b mutant appeared phenotypically normal under standard growth conditions (Fig. 1a and Supplementary Fig. 2a–c), and bulk telomere length measurements assessed by Terminal Restriction Fragment (TRF) analysis over successive generations revealed no statistically significant difference compared to WT (Supplementary Fig. 1c, d). The 3’ single-strand extension on the telomere (G-overhang) was also unchanged (Supplementary Fig. 1e). We assessed telomere integrity by testing for end-to-end chromosome fusions using Telomere Fusion PCR (TF-PCR). Controls with stn1 mutants³⁶ showed multiple TF-PCR products, but none were detected in either the second (G2) or fourth (G4) generation pot1b (Supplementary Fig. 1f). Thus, AtPOT1b does not significantly contribute to conventional telomere maintenance or chromosome end protection.

Fig. 1. POT1a and POT1b synergistically maintain genome stability.

a Growth and silique development in 3- or 6-week-old WT, third generation (G3) pot1a, pot1b, pot1a pot1b double mutants, and complementation lines (p35S::FLAG-Myc (FM)-POT1a/pot1a pot1b; pUBQ10::FM-POT1b/pot1a pot1b; pUBQ10::FM-PpPOT1/pot1a pot1b). Scale bars, 1 cm (rosettes) and 5 mm (siliques). b Telomere length in leaves of 3-week-old WT, G3 pot1a, pot1b, double mutants, and complementation lines (p35S::FM-POT1a/pot1a pot1b; pUBQ10::FM-POT1b/pot1a pot1b; pUBQ10::FM-PpPOT1/pot1a pot1b). Data points represent average telomere length from individual chromosome arms assessed by Primer-Extension Telomere Repeat Amplification (PETRA; see Supplementary Fig. 2g, h). Data shown as violin plots (n = 3 chromosome arms per genotype). One-way ANOVA with Tukey’s test. c Telomerase activity measured by Telomeric Repeat Amplification Protocol (TRAP) in inflorescences from 5-week-old WT, G3 pot1a, pot1b and double mutants. Relative activity is indicated below each lane (biological replicates 1 and 2). d Telomerase activity measured by qTRAP in inflorescences of 5-week-old WT, pot1a, pot1b, double mutants, and complementation lines (p35S::FM-POT1a/pot1a pot1b; pUBQ10::FM-POT1b/pot1a pot1b; pUBQ10::FM-PpPOT1/pot1a pot1b). Bars represent mean fold changes relative to WT ± SD (n = 2–3 biological replicates). One-way ANOVA with Tukey’s test. e Fluorescence in situ hybridization (FISH) of telomeres (red) and DAPI-stained chromosomes (blue) in inflorescences of 5-week-old WT, G3 pot1a, pot1b and double mutants. DAPI-only mitotic spreads are shown below. Arrows indicate telomere fusions (yellow), diffuse chromatin (green), and micronuclei (orange). Scale bar, 2 µm. Right, quantification of mitotic cells with chromosome bridges and micronuclei. Bars represent mean ± SD (n = 3 biological replicates; 31–58 mitotic cells per genotype were analyzed per replicate). One-way ANOVA with Tukey’s test. f Root apical meristem (RAM) in 6-day-old seedlings of WT, pot1a, pot1b, and POT1b complementation lines (pPOT1b::FM-POT1b/pot1b), treated with 20 µM Zeocin for 6 h, and stained with propidium iodide. Scale bar, 50 µm. g RNA-seq analysis of 6-day-old WT and pot1b mutants seedlings treated with mock or Zeocin. Heatmap shows Gene Ontology (GO) for biological processes (BP), molecular functions (MF), cellular components (CC), and KEGG pathways (KEGG) in the indicated comparisons. Representative images (a, f) from three independent experiments with similar results.

To fully ablate POT1 function, we generated pot1a pot1b double mutants. First-generation (G1/F2) homozygous WT, pot1a and pot1b lines were phenotypically similar, while double mutants were smaller with decreased biomass in subsequent generations (Supplementary Fig. 2a–c). Compared to G3 single mutants, G3 double mutants exhibited dramatic developmental defects, including stunted growth, fragile aerial organs, smaller siliques and fewer seeds (Fig. 1a and Supplementary Fig. 2a–c). Complementation with constitutive promoter::FLAG-Myc-POT1a, introduced at G2, rescued the developmental defect of the double mutants in G3/T1 plants (Fig. 1a and Supplementary Fig. 2d). Although the POT1b complementation line shares the same G3 background as the phenotypically normal pot1a single mutant, developmental defects persisted (Fig. 1a), possibly due to cumulative damage from prolonged POT1a deficiency or a requirement for POT1a to support POT1b function. We conclude that AtPOT1a and AtPOT1b have partially overlapping but distinct roles in growth and development.

We explored the molecular basis of the double mutant phenotype by measuring telomere length on bulk and individual chromosome ends (Primer Extension Telomere Repeat Amplification, PETRA)³⁷. Telomeres in G3 pot1a mutants shortened below 2 kb, while telomeres in G3 pot1b mutants were comparable to WT (Fig. 1b and Supplementary Fig. 2e–h), between 2-5 kb. Telomere length in pot1a pot1b resembled pot1a mutants, averaging 1.5 kb (Fig. 1b and Supplementary Fig. 2e–h). However, ectopic expression of either AtPOT1b or AtPOT1a in G3 double mutants failed to restore the WT telomere range (Fig. 1b and Supplementary Fig. 2e–h), indicating that a single POT1 paralog is insufficient to rescue critically shortened telomeres. Telomerase activity measured by quantitative and conventional Telomere Repeat Amplification Protocol (TRAP) showed reduced activity in pot1a, but WT-like levels in pot1b (Fig. 1c, d). Telomerase activity was slightly higher in pot1a pot1b mutants than in pot1a, which may reflect variation or negative regulation of telomerase by AtPOT1b in the absence of AtPOT1a. Expression of AtPOT1a, but not AtPOT1b, partially rescued telomerase activity in double mutants (Fig. 1d and Supplementary Fig. 2i), arguing that AtPOT1a is a primary positive regulator of telomerase. Nevertheless, both POT1 proteins, likely along with other telomere-associated proteins, appear necessary for full telomerase function and normal telomere maintenance.

To further assess telomere integrity in POT1-deficient plants, we examined chromosome spreads in mitotically dividing cells (Fig. 1e). Consistent with telomere dysfunction, G3 pot1a mutants exhibited bridged chromatin in anaphase. Although no bridges were detected in G3 pot1b, DAPI staining appeared weaker and more diffuse compared to WT (Fig. 1e and Supplementary Fig. 3a), suggesting altered chromatin organization, though this remains to be validated. Strikingly, severe genome instability was evident in G3 double mutants, including multiple anaphase bridges, lagging chromosomes and micronuclei (Fig. 1e and Supplementary Fig. 3b). Telomere fluorescence in situ hybridization (FISH) to detect telomere end-joining events showed telomeric signals in bridged chromatin of pot1a mutants, but not in pot1a pot1b (Fig. 1e and Supplementary Fig. 3b). Potential TF-PCR products, mostly very faint, could be detected in only a small subset of the 14 double mutants tested (Supplementary Fig. 3c), implying that telomere deprotection is not a predominant driver of bridged chromatin. Supporting this conclusion, subtelomeric regions were intact in double mutants since PETRA products could be generated and telomere tracts were longer than the critical 1 kb length threshold for fusion (Fig. 1b and Supplementary Fig. 2g, h)³⁷. We conclude that AtPOT1a and AtPOT1b act synergistically in promoting genome stability, but likely not via the chromosome end protection mechanism previously described for other POT1 proteins.

We considered that genome instability in pot1a pot1b double mutants might be related to a defect in DNA repair, since several telomere-related mutants in A. thaliana are hypersensitive to DNA damage^38–40. G3 pot1b, but not pot1a, mutants were hypersensitive to the radiomimetic drug Zeocin, and this phenotype was rescued by AtPOT1b complementation (Fig. 1f). To assess DNA damage directly, we performed an alkaline comet assay⁴¹, quantifying single- and double-stranded breaks by scoring DNA in comet tails (DT) across generations. Since the comet tail reflects unrepaired breaks, statistically similar levels across genotypes indicate a DNA repair defect is not responsible for Zeocin hypersensitivity in pot1b mutants (Supplementary Fig. 1g).

Zeocin binds DNA and induces cleavage via reactive radicals and thus has been used as a proxy for genome accessibility^42,43. The hypersensitivity of pot1b mutants, together with the low-density DAPI signal in chromosomal spreads (Fig. 1e and Supplementary Fig. 3a), is consistent with increased genome accessibility. Strikingly, pot1a pot1b mutants showed even greater sensitivity to Zeocin than pot1b (Supplementary Fig. 3d), underscoring the synergistic contribution of both POT1 paralogs in promoting genome integrity.

To characterize the transcriptional landscape of pot1b in response to Zeocin treatment, we employed RNA-seq on WT and pot1b mutants under both mock and Zeocin conditions. Zeocin-treated pot1b and WT showed highly similar transcriptional responses (Fig. 1g), arguing that the DNA damage response is intact. Zeocin treatment did not substantially alter the pot1b transcriptome. In contrast, comparison of untreated pot1b with untreated WT revealed 479 differentially expressed genes (2-fold, FDR < 0.05) with 356 upregulated and 123 downregulated (Supplementary Fig. 1h). Gene ontology (GO) analyses indicated enrichment for stress response pathways, including oxidative and osmotic stress, and down regulation of growth hormone auxin responsive genes (Fig. 1g). These findings imply that AtPOT1b contributes to the stress response and growth processes rather than altering the Zeocin-induced DNA damage response.

AtPOT1a and AtPOT1b are differentially regulated by ROS and promote ROS homeostasis

We next analyzed publicly available mRNA-seq datasets from WT Col-0 using Genevestigator®⁴⁴, Klepikova expression atlas⁴⁵, and single-cell RNA-seq root atlas⁴⁶. We observed distinct transcription profiles for AtPOT1a and AtPOT1b during development and under different stress conditions (Fig. 2a and Supplementary Figs. 4a, b, 5). AtPOT1b mRNA peaks during germination, early root growth, and flower development, particularly in dried and stratified seeds, radicle tip, pistil and anther (Supplementary Fig. 4a, b). AtPOT1b mRNAs were also expressed at low levels in vegetative tissues. Single-cell RNA-seq⁴⁶ further revealed that AtPOT1b expression is restricted to a highly localized niche of young, dividing meristematic cells in the root tip (Supplementary Fig. 5). Conversely, AtPOT1a mRNA is broadly expressed (Supplementary Fig. 4a, b), with the highest accumulation in protoxylem cells (Supplementary Fig. 5).

Fig. 2. POT1a and POT1b are differentially regulated under stress conditions.

a Expression profiles of POT1a and POT1b under environmental stress using Genevestigator®. b Immunoblot of POT1a-FLAG and POT1b-FLAG using α-FLAG antibody in Col-0 Arabidopsis protoplasts with 0.5 µM methyl viologen (MV) treatment for the indicated times. Quantification values below the blot were normalized to actin. Rubisco (RBC) loading control is shown by CBB staining. c, d Phenotypes of 7-day-old G2/G3 pot1a and pot1b mutant seedlings grown on ½MS media with and without 0.5 µM MV. Representative images (c). Percentage of fully opened green cotyledons quantified from 24 seedlings per genotype per replicate (d). Data represent mean ± SD (n = 3 biological replicates). One-way ANOVA with Tukey’s test (d). e ROS accumulation in flowers, seeds and roots of WT, pot1b and POT1b overexpression line (pUBQ10::FM-POT1b/pot1b) visualized by 3,3’-diaminobenzidine (DAB) staining for flowers, 2’,7’-dichlorofluorescein diacetate (H₂DCFDA) staining for seeds, and Peroxy Orange 1 (PO1) staining for roots. The stigma is highlighted by the dashed red circle. f DAB staining of leaves from 3-week-old WT and G2/G4 pot1a and pot1b mutants. g Ascorbate peroxidase, catalase, and peroxidase activities in 6-day-old seedlings of WT, G2/G3 pot1a and pot1b mutants and POT1b complementation lines (pPOT1b::FM-POT1b/pot1b; L1 and L2). The cat2 mutant is a positive control for catalase activity. Bars represent mean ± SD (n = 3 biological replicates). One-way ANOVA with Tukey’s test. Representative blots (b) and images (e, f) from three independent experiments with similar results.

We examined pot1b-associated phenotypes in organs where AtPOT1b is expressed. Embryo viability analysis showed significantly decreased seed density in G4 pot1b plants (Supplementary Fig. 4c, d), a defect fully rescued by AtPOT1b expression (Supplementary Fig. 4c, d). Pollen viability was also compromised in pot1b across generations but remained steady and was uncoupled from telomere length (Supplementary Fig. 4e, f). In contrast, pot1a mutants displayed progressively worsening defects, including decreased silique size and seed output with no significant decrease in pollen viability (Supplementary Fig. 4c–f), likely reflecting pleiotropic effects from escalating telomere dysfunction.

Genevestigator® indicated that, unlike AtPOT1a, AtPOT1b expression increases under oxidative, biotic, and abiotic stresses, including necrotrophic fungi, drought and cold (Fig. 2a). Each of these stressors increases intracellular ROS, which modulates developmental transitions, including germination and dormancy in seeds, cell division and growth in roots, and pollen viability and fertilization in flowers⁵. Strikingly, these settings correspond to the highest AtPOT1b expression. Not only does AtPOT1b expression appear to be controlled by ROS, pot1b mutants are hypersensitive to oxidative stress. We measured AtPOT1a and AtPOT1b mRNA accumulation by qPCR following treatments known to induce oxidative stress, including osmotic stress and herbicide exposure^5,47. Mannitol and sodium chloride induced both AtPOT1 genes (Supplementary Fig. 6a, b). Conversely, methyl viologen (MV, paraquat)⁴⁷ upregulated AtPOT1b mRNA but not AtPOT1a (Supplementary Fig. 6a). AtPOT1b mRNA increased 2-fold by 6 h after treatment, while protein increased ~6-fold by 1 h and 70-fold by 6 h (Fig. 2b and Supplementary Fig. 6a). pot1b mutants were more sensitive to MV than WT (Fig. 2c, d). AtPOT1a is also implicated in ROS regulation since pot1a mutants were hypersensitive to MV treatment (Fig. 2c, d), and AtPOT1a protein increased following MV exposure, albeit to a lesser extent than AtPOT1b (Fig. 2b).

We measured ROS across development in plants expressing different levels of AtPOT1b. As a control, we used catalase 2 (cat2) mutants, which are deficient in the major catalase isoform responsible for hydrogen peroxide (H₂O₂) scavenging⁴⁸. H₂O₂ was visualized using 2’,7’-dichlorodihydrofluorescein diacetate (H₂DCFDA) in seeds, Peroxy Orange 1 (PO1) in roots, and 3,3’-diaminobenzidine (DAB) in leaves and flowers⁴⁹. Superoxide anion (O₂^•−) was evaluated in roots using dihydroethidium (DHE)⁴⁹. Compared to WT, pot1b mutants accumulated higher ROS (Fig. 2e, f and Supplementary Fig. 6c). Remarkably, constitutive AtPOT1b over-expression decreased ROS below WT levels in tissues where AtPOT1b is naturally expressed (dry seeds, root tips, and flowers) (Fig. 2e). This outcome was most evident in the stigma (Fig. 2e), an organ where ROS signaling is essential for pollen tube germination and fertilization⁵.

In pot1a mutants, ROS levels were also elevated (Fig. 2f). To determine whether this increase reflects ROS dysregulation or telomere dysfunction, we assessed ROS in G2 and G4 pot1a and pot1b leaves. In G2, pot1a mutants appeared WT with no ROS detected by DAB staining. However, in G4, when morphological defects became evident due to critical telomere shortening, ROS was elevated (Fig. 2f). These results imply that ROS accumulation in pot1a mutants reflects, at least in part, the pleiotropic effects of telomere failure. Consequently, we counted as direct effects of POT1 only those phenotypes apparent in early generations (G1 or G2) and ignored those that only appeared and/or worsened beyond G3.

Unlike pot1a, pot1b leaves in both G2 and G4 harbored elevated ROS (Fig. 2f). Since telomere shortening is not associated with advancing generations of pot1b (Supplementary Fig. 1c, d), ROS accumulation is uncoupled from telomere shortening. Furthermore, because AtPOT1b is not normally expressed in leaves, ROS dysregulation in this setting indicates AtPOT1b plays a role in global ROS homeostasis. Moreover, since ROS is more highly elevated in G2 pot1a pot1b double mutants than in G2 pot1b (Supplementary Fig. 6d), we postulate that ROS imbalance contributes to the severe morphological defects found in these plants.

To investigate whether ROS accumulation is caused by increased production or impaired scavenging, we monitored the major ROS-scavenging enzymes: catalase, ascorbate peroxidase and peroxidase⁴⁸. All three activities were reduced in pot1b mutants and rescued in the complementation lines (Fig. 2g), consistent with impaired ROS detoxification. Catalase and ascorbate peroxidase activities were also downregulated in G2 pot1a mutants (Fig. 2g), supporting a role for AtPOT1a in the oxidative stress response. We conclude Arabidopsis POT1 proteins, particularly AtPOT1b, limit ROS and support antioxidant activity to maintain redox balance.

AtPOT1b is a telomeric protein that regulates genomic and telomeric oxidation

To explore the molecular basis for how AtPOT1b facilitates redox homeostasis, we studied its subcellular localization using a super-folder (sf) split GFP system⁵⁰. In this system, GFP is divided into two fragments: the large sfGFP1-10 is targeted to specific subcellular compartments via a corresponding cellular localization signal (CLS), while the small sfGFP11 tags the protein of interest. Fluorescence is reconstituted when both fragments localize to the same compartment (Supplementary Fig. 7a). As expected, GFP fluorescence was reconstituted only when sfGFP1-10-CLS was co-expressed with mCherry-sfGFP11-CLS, but not with an untargeted mCherry-sfGFP11 (Supplementary Fig. 7a).

We transiently co-expressed sfGFP1-10-nuclear localization signals (NLS) with either AtPOT1a- or AtPOT1b-sfGFP11 in Nicotiana benthamiana. Both constructs reconstituted nuclear GFP fluorescence, indicating nuclear localization (Fig. 3a). In parallel, stable transgenic lines expressing sfGFP1-10-NLS and AtPOT1a/b protein-sfGFP11 under native promoters were generated. In root epidermal cells, AtPOT1a showed uniform nuclear distribution (Fig. 3b). Notably, AtPOT1b fluorescence was undetectable in plants grown with sucrose, but appeared in the nucleus under sucrose-free conditions (Fig. 3b). Consistent with prior reports that sucrose supports β‑oxidation during early seedling growth and acts as an antioxidant⁵¹, we observed increased ROS in WT under sucrose-free conditions (Supplementary Fig. 7b). We therefore interpret sucrose deprivation as imposing an oxidative challenge. In this context, AtPOT1b showed a punctate peri-nucleolar pattern (Fig. 3c), consistent with telomeric localization in A. thaliana⁵². Chromatin immunoprecipitation with pPOT1b::FLAG-Myc-POT1b lines using an α-Myc antibody confirmed telomeric enrichment. This association was enhanced in sucrose-deprived conditions (Fig. 3d). Whether this outcome reflects higher AtPOT1b abundance or increased telomere interaction under oxidative stress is unknown.

Fig. 3. POT1b is a telomeric protein that regulates nuclear ROS and genomic and telomeric oxidation.

a Split GFP assay in Nicotiana benthamiana leaves co-expressing POT1a-sfGFP11 or POT1b-sfGFP11 with sfGFP1-10-NLS. mCherry-sfGFP11-NLS with sfGFP1-10-NLS served as positive controls. Scale bar, 10 µm. b, c Confocal imaging of 4-day-old Arabidopsis roots expressing pPOT1a::Myc-sfGFP11-POT1a, pPOT1b::Myc-sfGFP11-POT1b, or pPOT1b::POT1b-Myc-sfGFP11 with pUBQ10::sfGFP1-10-NLS, grown without sucrose. c shows a higher magnification of pPOT1b:: POT1b-Myc-sfGFP11. Scale bars, 50 µm (b) and 5 µm (c). d Telomere ChIP with α-Myc antibody in 6-day-old WT and pUBQ10::FM-POT1b/pot1b grown with and without sucrose. Data represent % input (n = 4–6 biological replicates). Box-and-whisker plots show median (center line), 25th–75th percentiles (box), and minimum–maximum values (whiskers). One-way ANOVA with Tukey’s test. e Subcellular fractionation of 6-day-old pUBQ10::POT1b-Myc-sfGFP11/pot1b grown with and without sucrose. Total (T), cytoplasmic (C), and nuclear (N) fractions were immunoblotted with α-Myc, α-Actin (cytoplasmic marker), and α-Histone H3 (nuclear marker) antibodies. f Hoechst and H₂DCFDA staining in roots of 4-day-old WT, pot1b, pUBQ10::FM-POT1b/pot1b and cat2. Scale bar, 2 µm. g ELISA-based quantification of genomic 8-oxoG in 6-day-old WT, cat2, pot1a, pot1b, pot1a pot1b double mutant and pUBQ10::FM-POT1b/pot1b grown without sucrose. Violin plots show fold changes relative to WT (n = 4–13 biological replicates). Brown-Forsythe and Welch ANOVA with Dunnett’s multiple comparisons test (two-sided). h 8-oxoG measurements in chromosome bodies (chr) and telomeres (telo) from 6-day-old WT and pot1b mutants grown without sucrose. Violin plots show fold changes relative to WT chr (n = 5 biological replicates). One-way ANOVA with Tukey’s test. i Telomere length measured by PETRA at 5R in 6-day-old WT and pot1b mutants treated with MV. Mean ± SD for Δtelomere length relative to untreated controls. Lane removals indicated; full blots in Supplementary Fig. 8a. j PETRA in 3-week-old WT and pot1b mutants under mock, drought, and high-temperature. 1L and 5R were measured twice per condition (Supplementary Fig. 8d, e), yielding four data points. Violin plots show telomere length change relative to mock (n = 4). Two-way ANOVA with Dunnett’s test. Representative images (a-c, f) and blots (e, i) from three independent experiments with similar results.

Although AtPOT1a and AtPOT1b are differentially regulated by stress (Fig. 2a and Supplementary Fig. 6a) and throughout development (Supplementary Fig. 4a, b), they are co-expressed in meristematic tissues. Here, AtPOT1b expression peaks (Supplementary Fig. 5). Since meristems are active in DNA replication and require telomere function^53,54, we tested for AtPOT1a-AtPOT1b interaction. Subcellular localization, co-immunoprecipitation (Co-IP) and split luciferase assays not only revealed in vivo co-localization of the two proteins within the nucleus, but also suggested they assemble into the same complex (Supplementary Fig. 7c–e). To ask if AtPOT1b localization is ROS-regulated, we used cellular fractionation and western blotting in pot1b mutants expressing pUBQ10::POT1b-Myc-sfGFP11. We detected both intra and extranuclear AtPOT1b with increased total and nuclear-localized protein in the absence of sucrose (Fig. 3e).

We next asked whether pot1b mutants exhibit elevated intranuclear ROS using Hoechst staining to label chromatin and H₂DCFDA to detect ROS. Compared to WT, both cat2 and pot1b mutants displayed increased nuclear ROS (Fig. 3f). Reintroducing AtPOT1b restored nuclear ROS to WT. Strikingly, H₂O₂ treatment caused increased genome degradation in pot1b mutants relative to WT (Supplementary Fig. 7f), an observation consistent with a role for AtPOT1b in mitigating ROS-induced DNA damage.

Genome oxidation was assessed by measuring the relative abundance of genomic 8-oxo guanine (8-oxoG) (Fig. 3g and Supplementary Fig. 7g)⁵⁵. 8-oxoG levels were on average 3-fold higher than WT in pot1b mutants grown without sucrose. This phenotype was rescued by AtPOT1b complementation (Fig. 3g). Moreover, pot1a pot1b mutants showed a synergistic increase in genomic oxidation regardless of sucrose conditions (Fig. 3g and Supplementary Fig. 7g), supporting the notion that AtPOT1a and AtPOT1b cooperate to modulate genome oxidation.

Since telomeres proportionally accumulate more oxidative damage than chromosome bodies^7,55, we examined telomere oxidation in WT and pot1b mutants grown without sucrose. WT telomeres displayed 3-fold higher 8-oxoG than the rest of the genome (Fig. 3h), while pot1b telomeres showed a corresponding 5-fold greater increase (Fig. 3h). In vertebrates, telomere oxidation is associated with stochastic changes in telomere length^6,56. Likewise, we found telomere length was more variable in pot1b mutants under oxidative stress. MV induced a net decrease in telomere length, correlating with stress intensity (Fig. 3i and Supplementary Fig. 8a), while continuous light caused stochastic changes, increased or decreased length depending on the chromosome arm (Supplementary Fig. 8b, c). Interestingly, under drought or high temperature growth conditions, which both induce ROS⁵, pot1b mutants showed telomere shortening at specific arms (Fig. 3j and Supplementary Fig. 8d, e). These data imply that AtPOT1b promotes chromosomal stability by modulating telomere oxidation.

POT1b localizes to peroxisomes and supports peroxisome function

In addition to nuclear localization, split GFP experiments revealed that AtPOT1b, but not AtPOT1a, reconstituted detectable GFP signal in peroxisomes (Fig. 4a). Although the peroxisomal import route remains unknown, AtPOT1b, but not AtPOT1a, contains an N-terminal peroxisomal targeting signal 2 (PTS2)-like motif (RIQDAFKALHL), which resembles the canonical PTS2 consensus sequence (R[LIQ]X₅HL) (Supplementary Fig. 9a)⁵⁷. Peroxisomes are central hubs for redox biology and home to a multitude of antioxidant scavengers, with catalase accounting for 10–25% of peroxisomal proteins in Arabidopsis^57–59. Peroxisomes fuel germination and root development through β-oxidation of seed lipids, a process that is bypassed when sucrose is included in the media (standard germination conditions: ½MS containing 1% sucrose). Plants defective in peroxisomal function display a short-root phenotype in the absence of sucrose⁶⁰. Notably, similar to cat2 mutants, G2 pot1b mutants, but not G2 pot1a, exhibited a sucrose-dependent short-root phenotype (Fig. 4b and Supplementary Fig. 9b–d) that was rescued by AtPOT1b overexpression (Supplementary Fig. 9c, d). These phenotypes, coupled with the global reduction in ROS scavenging activities (Fig. 2g), are consistent with peroxisomal dysfunction. Zeocin hypersensitivity in pot1b mutants was also elevated in the absence of sucrose (Fig. 4c), suggesting that peroxisomal dysfunction may exacerbate DNA damage.

Fig. 4. POT1b localizes to peroxisomes and is required for peroxisome function.

a Split GFP assay in N. benthamiana co-expressing POT1a-sfGFP11 or POT1b-sfGFP11 with sfGFP1-10-peroxisomal targeting signal (SKL). mCherry-sfGFP11-SKL with sfGFP1-10-SKL served as controls. Scale bar, 10 µm. b Phenotypes of 7-day-old G2/G4 pot1a and pot1b mutants grown with sucrose and without sucrose. c RAM of 6-day-old seedlings grown with sucrose (Suc) and without sucrose (No suc), treated with 20 µM Zeocin as indicated then stained with PI. Scale bar, 50 µm. d Yeast two-hybrid (Y2H) assay in AH109 yeast cells expressing GAL4 DNA-binding domain (BD)-fused POT1a or POT1b and GAL4 activation domain (AD)-fused CAT2. Serial dilutions were spotted on selective SD media lacking leucine, tryptophan, histidine, and adenine (-LWHA). SD plates lacking leucine and tryptophan (-LW) served as controls. e Split luciferase assay in N. benthamiana co-expressing POT1a-nLuc, POT1b-nLuc, and CAT2-nLuc. RFP-nLuc and RFP-cLuc served as controls. cps, counts per second. f Co-immunoprecipitation (Co-IP) in protoplasts from 3-week-old Col-0 Arabidopsis expressing CAT2-HA, POT1a-FLAG, or POT1b-FLAG. Proteins were immunoprecipitated with α-FLAG beads and immunoblotted with α-FLAG and α-HA antibodies. Input (3rd and 5th panels) and RBC loading controls by CBB staining (4th and 6th panels) are shown. g Proximity labeling and quantitative MS in Arabidopsis transgenic lines expressing POT1b-TurboID (TbID)-HA in pot1b mutants (p35S::POT1b-TurboID-HA/pot1b) compared with control lines expressing only TbID-HA in Col-0 WT (p35S::TurboID-HA/Col-0) grown with and without sucrose for 6 days. Scatter plots show significantly enriched proteins (log₂FC > 1.5, p < 0.05) in red, as determined using a two-tailed unpaired Student’s t test (n = 3 biological replicates). Peroxidases highlighted in red with blue lines. Venn diagram shows the number of enriched proteins and overlap across conditions. h Gene ontology (GO) enrichment analysis of proteins identified from (g). Enrichment was performed using g:Profiler. Heatmap shows enriched cellular components, biological processes, and molecular functions with adjusted p-values. Enrichment significance was determined by a hypergeometric test with Benjamini–Hochberg correction. Red boxes highlight oxidative stress and peroxisome-associated terms. Representative images (a-d, e) and blots (f) from three independent experiments with similar results.

We sought AtPOT1b interaction partners to further investigate the role of AtPOT1b in promoting redox homeostasis. Yeast two-hybrid (Y2H) screening identified Catalase 2 (CAT2), one of three A. thaliana catalase isoforms, as an AtPOT1b interactor (Fig. 4d and Supplementary Table 1). The CAT2-AtPOT1b association was validated with a split luciferase assay, and Co-IP experiments in Arabidopsis protoplasts and N. benthamiana leaves (Fig. 4e, f and Supplementary Fig. 9e). AtPOT1a also bound CAT2 (Fig. 4e, f and Supplementary Fig. 9e). In addition, AtPOT1a and AtPOT1b interacted with CAT1, while AtPOT1b bound CAT3 (Supplementary Fig. 9f). These findings suggest AtPOT1-catalase interaction predates the Arabidopsis POT1 and CAT gene duplications.

We also employed TurboID-based (TbID) proximity labeling using p35S::POT1b-TbID-HA transgenic lines, with p35S::TbID-HA lines as controls. Plants were grown with and without sucrose prior to biotin treatment. Biotinylation and streptavidin pull-down were confirmed by western blot using streptavidin-HRP (Supplementary Fig. 9g, h). Quantitative mass spectrometry performed on plants grown without sucrose revealed enrichment of the peroxisomal import receptor PEX5, supporting AtPOT1b localization in peroxisomes (Fig. 4g, h). It also revealed multiple antioxidant enzymes, including class III secretory/apoplastic peroxidases (PER62, PER71, PER45, PER27, and PRX32) (Fig. 4g, h), which are annotated in TAIR and SUBA as cell wall–apoplastic peroxidases rather than bona fide peroxisomal proteins⁶¹. Co-IP in A. thaliana protoplasts confirmed AtPOT1b association with PER62 and PER71 after MV exposure and with PER45 under both MV-treated and untreated conditions (Supplementary Fig. 10a–c). We do not interpret enrichment of class III peroxidases as evidence of direct interaction within the apoplast. Instead, these proteins could be labeled via transient proximity to intracellular pools during biosynthesis and trafficking en route to secretion, rather than as mature extracellular enzymes. This interpretation is consistent with POT1b functioning within a broader ROS regulatory network. Notably, PEX5 enrichment and peroxisome-related GO terms (Fig. 4h) support AtPOT1b localization and function in peroxisomes and redox homeostasis.

Interestingly, TbID also revealed two telomere-related proteins: TRFL7, a putative shelterin component⁶² detected in both sucrose and non-sucrose conditions; and the telomerase accessory factor NAP57, homologous to dyskerin³², enriched under sucrose deprivation (Fig. 4g). The AtPOT1b-dyskerin interaction was confirmed by Co-IP, irrespective of oxidative stress (Supplementary Fig. 10f). AtPOT1b interaction with two distinct components of the telomerase RNP (dyskerin and AtPOT1a) and a putative telomere-associated protein provides plausible avenues for AtPOT1b recruitment to telomeres.

TbID experiments conducted with AtPOT1a using p35S::POT1a-TbID-HA showed similar associations with antioxidant enzymes. APX3 and GSTF6 were significantly enriched without sucrose, while PER57 was enriched with sucrose (Supplementary Fig. 11a, b). Co-IP with A. thaliana protoplasts indicated AtPOT1a interaction with PER71, PER32, and PER27, specifically after MV treatment, and with PER45 under both MV-treated and untreated conditions (Supplementary Fig. 10b–e). In addition, GO enrichment analysis uncovered telomere maintenance proteins and a variety of mRNA stability and organellar proteins (Supplementary Fig. 11c), underscoring the potential for AtPOT1a in broader cellular functions beyond telomerase regulation.

AtPOT1b cooperates with CAT2 to facilitate ROS regulation

To further investigate the mechanism by which AtPOT1b controls redox homeostasis, we examined the molecular interaction between this protein and catalase. Although AtPOT1b associates with multiple antioxidants, we focused on the AtPOT1b-CAT2 connection because: (1) this interaction was seen in multiple assays; (2) CAT2 is the most abundant catalase isoform in A. thaliana with high H₂O₂ scavenging capacity⁴⁸; and (3) both AtPOT1b and CAT2 localize to peroxisomes and nuclei⁵⁸. For CAT2, this dual localization is critical for redox balance and protection against oxidative damage⁴⁸. ROS levels measured by H₂DCFDA staining were restored to WT in pot1b mutant seeds overexpressing CAT2 (Fig. 5a); biomass and genome oxidation were similarly rescued in seedlings (Fig. 5b, c). In addition, DAB staining in pot1b, cat2, and pot1b cat2 double mutants revealed comparable ROS elevation (Supplementary Fig. 12a), raising the possibility of a cooperative role for AtPOT1b and CAT2 in redox regulation.

Fig. 5. POT1b cooperates with CAT2 to protect against genomic oxidation.

a H₂DCFDA staining of seeds from WT, pot1b, and p35S::CAT2-HA/pot1b (L1 and L2). Scale bar, 200 µm. b Phenotypes of 10-day-old WT, pot1b, and p35S::CAT2-HA/pot1b grown without sucrose. Fresh weight was quantified (54 seedlings per replicate). Data represent mean ± SD (n = 3 biological replicates). One-way ANOVA with Dunnett’s test. c ELISA-based quantification of genomic 8-oxoG in 6-day-old WT, pot1b, p35S::CAT2-HA/pot1b and cat2 grown without sucrose. Violin plots show fold changes relative to WT (n = 4–11 biological replicates). Brown-Forsythe and Welch ANOVA with Dunnett’s multiple comparisons test (two-sided). d Subcellular fractionation of catalase in WT and pUBQ10::FM-POT1b/pot1b grown with and without sucrose for 6 days. Cytoplasmic (C) and nuclear (N) fractions were immunoblotted with α-Myc, α-Catalase, α-Catalase2, α-Actin (cytoplasmic marker), and α-Histone H3 (nuclear marker) antibodies. e Telomere ChIP with α-catalase antibody in 6-day-old WT and pot1b grown with and without sucrose. Data represent % input (n = 4 biological replicates). Box-and-whisker plots show median (center line), 25th–75th percentiles (box), and minimum–maximum values (whiskers). One-way ANOVA with Tukey’s test. f Y2H domain mapping of AtPOT1-CAT2 interaction. Catalytic site in CAT2 D1 is highlighted (yellow). g Protein docking model of POT1b-CAT2 showing predicted interface residues. h Y2H validation of POT1b-CAT2 interface residues. AH109 yeast co-expressing AD-CAT2 and BD-POT1b (WT, R300A, K222A, K10A, and E3A) were serially diluted and spotted on selective SD media (-LW and -LWHA). i Growth phenotypes of 3-week-old WT and pot1b expressing POT1b^R300A and POT1b^K222A. Scale bar, 1 cm. j DAB staining of Arabidopsis leaves infiltrated with A. tumefaciens carrying pUBQ10::FM-POT1b, pUBQ10::FM-POT1b^K222A, p35S::FM-POT1a^WT, and p35S::FM-POT1a^E439A constructs. Negative controls include pUBQ10::FM-POT1b^R300A and p35S::FM-POT1a^Q378A, WT, and pUBQ10::FM-GUS/pot1b. k Telomeric 8-oxoG quantified by DNA immunoprecipitation using 8-oxoG antibody followed by telomeric qPCR in 3-week-old WT and pot1b expressing POT1b^R300A or POT1b^K222A. UBQ10 served as non-telomeric control. Bars represent % input ± SD (n = 4 biological replicates). One-way ANOVA with Tukey’s test. Representative images (a, h–j) and blots (d) from three independent experiments with similar results.

As with AtPOT1b, nuclear CAT2 was higher in plants grown without sucrose than with it (Fig. 5d). Importantly, nuclear catalase was further elevated when AtPOT1b was overexpressed (Fig. 5d), suggesting AtPOT1b may boost nuclear accumulation of CAT2. The antioxidant peroxiredoxin associates with mammalian telomeres⁶³, prompting us to test if catalase accumulates at plant telomeres. Strikingly, ChIP using α-catalase antibodies revealed telomere-associated catalase, which increased for plants grown without sucrose and decreased in pot1b mutants (Fig. 5e).

If the AtPOT1b-CAT2 interaction is important for regulating redox homeostasis, we predicted that impairing this association would elevate ROS. Y2H assays showed OB1 and OB2 of AtPOT1b interact with the D1 and D2 domains of CAT2, containing the catalase active site (54-70 aa) and heme-binding site (344-352 aa), respectively (Fig. 5f and Supplementary Fig. 12b–e). Structural modeling using ColabFold⁶⁴ and Cluspro 2.0⁶⁵ produced two high-scoring models. One shown in Fig. 5g predicts AtPOT1b^R300, AtPOT1b^K222, AtPOT1b^K10, and AtPOT1b^E3 residues as critical for CAT2 interaction. These residues were mutated to alanine (A). Y2H and Co-IP in protoplasts showed that AtPOT1b^K222A impaired CAT2 interaction (Fig. 5h and Supplementary Fig. 12f). To test functional relevance in planta, stable transgenic lines expressing AtPOT1b^R300A and AtPOT1b^K222A in pot1b mutants were generated. AtPOT1b^R300A lines were phenotypically normal, but AtPOT1b^K222A exhibited severe growth defects (Fig. 5i). Importantly, overexpression of GUS, AtPOT1b^WT, AtPOT1b^R300A, or AtPOT1b^K222A in WT Arabidopsis leaves, followed by DAB staining, showed increased ROS in AtPOT1b^K222A but not in the other constructs (Fig. 5j). Furthermore, 8-oxoG immunoprecipitation followed by telomeric qPCR revealed increased telomeric DNA oxidation in AtPOT1b^K222A but not in the UBQ10 control (Fig. 5k).

The AtPOT1a-CAT2 interaction was impaired using a similar approach. Unlike AtPOT1b, the C-terminal domain of AtPOT1a engages the D1 domain of CAT2 (Fig. 5f and Supplementary Fig. 12d, e, g, h). The structural model shown in Supplementary Fig. 12i predicts two critical residues for CAT2 interaction: AtPOT1a^E439 and AtPOT1a^Q378. Y2H and Co-IP in protoplasts demonstrated AtPOT1a^E439A abolished interaction with CAT2 (Supplementary Fig. 12j, k). Notably, stable transgenic lines expressing AtPOT1a^Q378 and AtPOT1a^E439 in the pot1a mutant background were phenotypically normal (Fig. 5i). Furthermore, overexpression of AtPOT1a^E439A slightly increased ROS accumulation in leaves (Fig. 5j), but not to the same extent as in AtPOT1b^K222A. These findings argue that AtPOT1-CAT2 interaction, particularly the AtPOT1b-CAT2 interaction, is important for ROS homeostasis, telomere integrity and plant growth.

Regulation of ROS homeostasis is an ancestral function of POT1

Finally, we examined the evolutionary origin of POT1-mediated redox regulation by first testing if AtPOT1a and AtPOT1b have overlapping roles in ROS homeostasis. AtPOT1a overexpression restored WT ROS levels in pot1b mutants (Fig. 6a), indicating functional substitution for AtPOT1b in this setting. We also examined conservation across species by complementation with single-copy POT1 genes from the moss Physcomitrium patens (PpPOT1; diverged 400 mya)²⁹ or humans (hPOT1; diverged 1.6 bya)¹⁷ (Supplementary Fig. 13a). Both restored ROS levels to WT in pot1b flowers or seeds (Fig. 6a and Supplementary Fig. 13b), demonstrating redox regulation is an ancient function for POT1 and predates land plants.

Fig. 6. Regulation of ROS homeostasis is an ancestral function of the plant POT1.

a H₂DCFDA and DAB staining of seeds and flowers from WT, pot1b, p35S::FM-POT1a/pot1b, pUBQ10::FM-POT1b/pot1b and pUBQ10::FM-PpPOT1/pot1b. Stigma outlined (dashed circles). Scale bars, 200 µm (seeds) and 2 mm (flowers). Y2H domain mapping of PpPOT1-CAT2 interaction. Yeast cells expressing BD-truncated PpPOT1 and AD-CAT2 (b) or BD-PpPOT1 and AD-truncated CAT2 (c) were serially diluted and spotted on -LW and -LWHA. d Schematic of POT1 and CAT2 interaction across A. thaliana (AtPOT1a and AtPOT1b), P. patens (PpPOT1), and human POT1 (hPOT1). Catalytic site in CAT2 D1 is highlighted (yellow). e Subcellular localization of PpPOT1-GFP in 4-day-old Arabidopsis. FM4-64 marks plasma membranes. GFP and FM4-64 signal intensities (arbitrary units, arb. unit) were measured along the dashed line. Dashed circle indicates nucleus. Scale bar, 5 µm. f Peroxidase and catalase activities in 6-day-old WT, pot1b, and pUBQ10::FM-PpPOT1/pot1b (L1 and L2). Data represent mean ± SD (n = 3–4 biological replicates). One-way ANOVA with Tukey’s test. g Phenotypes of 7-day-old WT, pot1b, cat2, pUBQ10::FM-POT1b/pot1b (L1 and L2), pUBQ10::FM-PpPOT1/pot1b, and pUBQ10::FM-GUS/pot1b grown with and without 0.5 µM MV. Percentage of green cotyledons from 36 seedlings per replicate (right). Data represent mean ± SD (n = 3). One-way ANOVA with Tukey’s test. h ELISA-based quantification of 8-oxoG in 6-day-old WT, pot1b, p35S::FM-POT1a/pot1b, pUBQ10::FM-POT1b/pot1b, and pUBQ10::FM-PpPOT1/pot1b grown without sucrose. Violin plots show fold changes relative to WT (n = 3–13 biological replicates). Brown-Forsythe and Welch ANOVA with Dunnett’s multiple comparisons test (two-sided). (i) Speculative model for AtPOT1b-mediated antioxidant enzyme recruitment and regulation of genome oxidation. AtPOT1a (P1a) and AtPOT1b (P1b) interact and associate with catalase (CAT) and potentially other antioxidant enzymes (AO), with AtPOT1b as the primary recruiter. AtPOT1b engages telomeres via AtPOT1a to form an antioxidant hub (gray cloud). Without AtPOT1b, AtPOT1a can partially mitigate genome oxidation. In POT1b^K222A, catalase recruitment is lost, but AtPOT1a binding remains, providing a decoy mechanism. In pot1a pot1b mutants, antioxidant recruitment fails, causing massive genome oxidation and chromosomal instability. Created in BioRender. Min, J.-H. (2026) https://BioRender.com/hme83mu. Representative images (a–c, e, g) from three independent experiments with similar results.

We further discovered that the POT1-catalase interaction is conserved. Y2H assays showed that PpPOT1 and hPOT1 interact with AtCAT2 through domains analogous to AtPOT1a and AtPOT1b, respectively (Fig. 6b–d and Supplementary Fig. 13c–f), consistent with convergent evolution. Like AtPOT1b, PpPOT1 is dually localized in the nucleus and cytoplasm (Fig. 6e). Peroxidase and catalase activities are increased in pot1b mutants overexpressing PpPOT1 (Fig. 6f). We also examined hypersensitivity to oxidative stress (0.5 µM MV) in these plants. While both pot1b and cat2 mutants showed decreased cotyledon greening relative to WT, this phenotype in pot1b mutants was fully rescued by AtPOT1b or PpPOT1, but not the GUS control (Fig. 6g). PpPOT1 expression also rescued the elevated genome oxidation in pot1b under sucrose deprivation (Fig. 6h), indicating the capacity to mitigate oxidative stress is conserved in PpPOT1. However, PpPOT1 expression did not fully rescue developmental phenotypes (Fig. 1a and Supplementary Fig. 2d), telomere length (Fig. 1b and Supplementary Fig. 2g, h) or telomerase activity (Fig. 1d and Supplementary Fig. 2i) in pot1a pot1b double mutants. We conclude that while the redox regulatory function of POT1 is conserved, its telomere-specific roles/interactions have diverged over evolutionary time.

Discussion

Cells preserve genome stability through a multitude of pathways, including DNA repair, telomere protection and ROS regulation^6–8. POT1 has been extensively studied as a telomere guardian, but here we report a role for POT1 in redox regulation. AtPOT1a evolved for telomere maintenance via telomerase³², while the data presented here reveal a primary function of AtPOT1b in redox balance. pot1b mutants show elevated ROS and stress-sensitive phenotypes, and although many stressors elevate ROS^4,5, including telomere dysfunction⁶, AtPOT1b is remarkable as its overexpression drives ROS below WT levels in multiple organs. In this study, we not only define AtPOT1b as a key modulator of redox homeostasis, but also provide insight into how its nuclear and peroxisomal localization, and antioxidant partners, could account for changes in ROS and genome stability.

Mammalian POT1 proteins are broadly expressed^17,66. Similarly, AtPOT1a mRNA can be detected in essentially all Arabidopsis cell types. Conversely, AtPOT1b expression is restricted to meristems and reproductive organs, niches where ROS signaling is critical for developmental transitions, especially reproduction⁵. In addition to transcriptional regulation, AtPOT1b protein abundance is markedly increased in response to oxidative damage, raising the possibility that ROS modulates AtPOT1b translation, stability or degradation, like other telomere-associated proteins that undergo turnover under stress conditions¹⁶.

Although AtPOT1b does not contribute to conventional telomere protection, it promotes genome integrity. Complete abrogation of AtPOT1 in pot1a pot1b double mutants triggers profound genome instability: mitotically dividing cells display extensive chromatin bridges, lagging chromosomes and micronuclei. We propose these defects primarily reflect massive genome oxidation rather than telomere dysfunction (see Fig. 6i). First, ROS over-accumulates in pot1b nuclei with 8-oxoG amassing at telomeres. This phenotype is exacerbated in pot1a pot1b double mutants. Second, genome oxidation induces chromosomal defects similar to those observed in pot1a pot1b mutants, including bridged chromatin and micronuclei^6,7. Third, oxidized telomeres in mammalian cells exhibit both shortening and lengthening depending on the degree of oxidation⁶⁷. pot1b mutants likewise display stochastic telomere length changes and increased genome instability under oxidative stress. Fourth, genome oxidation is associated with chromatin decompaction^7,68. We observed low-density DAPI signals and hypersensitivity to DNA-damaging agents in pot1b and pot1a pot1b double mutants^38,40. We therefore postulate that AtPOT1a and AtPOT1b synergistically defend against DNA oxidation. The capacity of pot1b mutants to display a strong redox phenotype without compromised telomere integrity provided a unique separation‑of‑function background to explore ancestral, telomere‑independent POT1 functions that would be difficult, if not impossible, to dissect in pot1a or hPOT1 mutants, where essential, telomeric roles dominate.

Although the mechanism for ROS regulation by POT1 is unknown, the dual localization of AtPOT1b in nuclei and peroxisomes may provide clues. Peroxisomes house crucial antioxidant enzymes and act as ROS metabolism hubs⁵⁹. We found that AtPOT1b interacts with multiple ROS scavengers, including catalases and peroxidases, and is required for their robust enzyme activity and for peroxisome function. Thus, one potential mechanism for AtPOT1b is the enhancement of peroxisomal ROS detoxification, which could indirectly protect the genome by preventing excess ROS from reaching the nucleus. A parallel function for AtPOT1b in Arabidopsis peroxisomes akin to mammalian TERT and TIN2 in mitochondria^9–12 would suggest a common role for organelle-localized telomere proteins.

While our current experiments do not test whether AtPOT1b directly alters CAT2 catalytic activity, we present a working hypothesis for AtPOT1b function that highlights CAT2 regulation as an important direction for future mechanistic studies. In response to elevated ROS, AtPOT1b is upregulated and accumulates at chromosome ends, where the genome is most susceptible to oxidation^7,55,67. Although AtPOT1b lacks the capacity to bind telomeric DNA³⁵, its binding partners include AtPOT1a and dyskerin, providing a means of engaging telomeres via telomerase³². Once telomere-bound, AtPOT1b could attract antioxidants, serving as a hub for modulating telomere oxidation (Fig. 6i). An intriguing, but untested proposition is that the stress-induced accumulation of AtPOT1b at telomeres reflects a condensate-like assembly driven by multivalent interactions with antioxidant binding partners. Given recent evidence of phase separation at mammalian telomeres^69,70, telomere-antioxidant assemblies in plants could be especially advantageous in response to adverse environments⁷¹. Although AtPOT1b plays a more prominent role in controlling DNA oxidation, in its absence, AtPOT1a can partially substitute (Fig. 6i). Notably, diminishing the CAT2-AtPOT1b interaction in the AtPOT1b^K222A line leads to ROS overaccumulation and a severe growth defect, reminiscent of A. thaliana P2ΔC mutants²⁶. These mutants overexpress a truncated allele of AtPOT1b lacking the C-terminus and exhibit severe genome instability. Because the P2ΔC deletion is predicted to disrupt CAT2 binding as with AtPOT1b^K222A, these variants may encode nonfunctional decoys that alter ROS scavenging via interactions with other antioxidants or AtPOT1a (Fig. 6i) to suppress responses that normally compensate for pot1b loss. Finally, we note that indirect (peroxisomal) and direct (telomeric) ROS mediation mechanisms are not mutually exclusive and could cooperate to defend against genome oxidation.

Our cross-species complementation experiments reveal that ROS regulation is an ancient trait of POT1; single-copy POT1 proteins from moss and humans engage catalase and restore ROS homeostasis in Arabidopsis pot1b mutants. Strikingly, these proteins cannot rescue the telomere-length and telomerase defects of plants lacking AtPOT1a and AtPOT1b. This observation is consistent with the divergent modes of DNA binding displayed by mammalian POT1 proteins²¹ and argues that telomere-specific functions of POT1 are free to evolve. In contrast, the conserved redox function of POT1 appears to represent an adaptive feature with POT1 acting as one key node in an integrated network essential for redox balance and genome stability.

Methods

Plant materials and growth conditions

A. thaliana wild type (WT) Col-0 and pot1a were reported previously³¹. A. thaliana T-DNA insertion line, cat2 (SALK_057998), pCAMBIA1380-pUBQ10::sfGFP1-10-NLS (CS69832), pCAMBIA1380-pUBQ10::sfGFP1-10-SKL (CS69835), were obtained from the Arabidopsis Biological Resource Center (ABRC). Gene editing line pot1b, pUBQ10::FLAG-Myc-POT1b/Col-0, p35S::FLAG-Myc-POT1a/pot1b, p35S::FLAG-Myc-POT1b/pot1b, pUBQ10::FLAG-Myc-POT1b/pot1b, pUBQ10::FLAG-Myc-PpPOT1/pot1b, p35S::CAT2-HA/pot1b, pUBQ10::FLAG-Myc-GUS/pot1b, pPOT1b::FLAG-Myc-POT1b/pot1b, p35S::FLAG-Myc-POT1a/pot1a pot1b, pUBQ10::FLAG-Myc-POT1b/pot1a pot1b, pUBQ10::FLAG-Myc-PpPOT1/pot1a pot1b, pPOT1a::Myc-sfGFP11-POT1a/pUBQ10::sfGFP1-10-NLS, pPOT1b::Myc-sfGFP11-POT1b/pUBQ10::sfGFP1-10-NLS, pPOT1b::POT1b-Myc-sfGFP11/pUBQ10::sfGFP1-10-NLS, p35S::TurboID-HA/Col-0, p35S::POT1b-TurboID-HA/pot1b, p35S::POT1a-TurboID-HA/pot1a, and p35S::PpPOT1-GFP/Col-0 transgenic plants were generated in this study (see below for details). cat2/pot1b was obtained by crossing. A. thaliana and N. benthamiana plants were grown in soil (Jolly Gardener C/20 or C/Gs, USA) under a 16 h light/8 h dark photoperiod of 75–100 μE m⁻² s⁻¹ light, 50% relative humidity, and at a constant temperature of 22 °C. Arabidopsis seeds were sterilized in 50% bleach with 0.1% Triton X-100, stratified for 2 days at 4 °C in the dark, germinated on half-strength Murashige and Skoog (½MS) containing 0.8% agar, 2.5 mM MES at pH 5.8, supplemented or not with 1% sucrose, and grown under the same conditions as above. For phenotypic analysis under oxidative stress, seedlings were grown on ½MS plates containing 0.5 µM methyl viologen (MV), and green cotyledons were examined 7 days after germination. For telomere length analysis under various stress conditions, 7-day-old seedlings were transferred to soil and grown under three specific conditions. Standard temperature of 22 °C with 95% soil water content (SWC), moderate drought at 22 °C with 35% SWC, and high temperature at 30 °C with SWC maintained above 80% to offset transpiration effects. Soil water content (SWC) was determined gravimetrically by defining 0% SWC as the weight of dry soil and 100% SWC as the weight after full saturation. Desired SWC levels (95% and 35%) were calculated based on these reference weights⁷². Plants in drought and high-temperature conditions were sampled after 2 weeks.

Plasmid construction and transgenic plant generation

pUBQ10::sfGFP1-10-NLS, pUBQ10::sfGFP1-10-SKL, pUBQ10::mCherry-Myc-sfGFP11-NLS, pUBQ10::mCherry-Myc-sfGFP11-SKL plasmids were obtained from Addgene. POT1b null allele, pot1bΔ88, in WT Col-0, was made using CRISPR-Cas9. Protospacers were cloned into the Streptococcus pyogenes Cas9 (SpCas9)-based system⁷³, and subjected to Agrobacterium tumefaciens GV3101-mediated transformation in WT plants. The T-DNA insertion encoding the SpCas9 was selected by germinating the seeds in ½MS plates containing glufosinate-ammonium (Basta, 30 μg/ml) and transferring only the Basta-resistant seedlings to the soil. Plants were genotyped for the deletion of 88 nucleotides in AtPOT1b and the Cas9 gene with primers in Supplementary Data 1. The cDNA or gDNA of AtPOT1a, AtPOT1b, PpPOT1, hPOT1, GUS, CAT1, and CAT3 were amplified with primers in Supplementary Data 1, cloned into an entry vector pENTR with a gateway system. AtPOT1a, AtPOT1b, PpPOT1, hPOT1, and GUS were subcloned into a plant gene expression vector pBA under the CaMV 35S or UBQ10 promoter with a gateway system to make binary constructs, pBA-p35S::FLAG-Myc-POT1a, pBA-p35S::FLAG-Myc-POT1b, pBA-pUBQ10::FLAG-Myc-POT1b, pBA-pUBQ10::FLAG-Myc-PpPOT1, pBA-pUBQ10::FLAG-Myc-hPOT1, and pBA-pUBQ10::FLAG-Myc-GUS. pBA-pPOT1b::FLAG-Myc-POT1b, pBA-UBQ10::POT1a-Myc-sfGFP11, and pBA-UBQ10::POT1b-Myc-sfGFP11 were obtained using the Gibson assembly method with primers in Supplementary Data 1 and a gateway system. POT1a^E439A, POT1a^Q378A, POT1b^K222A, and POT1b^R300A mutant variants were generated by site-directed mutagenesis with primers in Supplementary Data 1 using pENTR-POT1a and pENTR-POT1b as templates and were subcloned into a pBA vector under the UBQ10 promoter. Truncation variants, including POT1a OB1 + 2, POT1a OB1, POT1a C-terminus, POT1b OB1 + 2, POT1b OB1, POT1b OB2, POT1b C-terminus, PpPOT1 OB1, PpPOT1 OB2, PpPOT1 C-terminus, hPOT1 OB1 + 2, hPOT1 OB1, hPOT1 OB2, hPOT1 C-terminus, CAT2 D1, CAT2 D2, CAT2 D3 were amplified with primers in Supplementary Data 1 using the full-length AtPOT1a, AtPOT1b, PpPOT1, hPOT1, and CAT2 in a pENTR vector as templates. AtPOT1a, AtPOT1b, PpPOT1, hPOT1, CAT1, CAT2, CAT3, POT1a^E439A, POT1a^Q378A, POT1b^K222A, and POT1b^R300A mutant variants and truncation variants were subcloned into pGADT7 (AD) and pGBKT7 with a gateway system for yeast two-hybrid (Y2H). The cDNA or gDNA of AtPOT1a, AtPOT1b, CAT2, PER32, PER71, PER27, PER45, PER62, or NAP57 were amplified with primers in Supplementary Data 1 containing BamHI at the 5’-terminus and StuI or SmaI at the 3’-terminus and ligated into the pHBT vector under the 35S promoter for protoplast expression with the HA, FLAG, GFP, or mCherry epitope tags at the C-terminus using ClonExpress II One-Step Cloning Kit (Vazyme, China) according to the manufacturer’s protocols. Fragments of p35S::CAT2-HA were amplified with primers in Supplementary Data 1 and sub-cloned into pCB302 to obtain pCB302-p35S::CAT2-HA. gDNA of AtPOT1a and AtPOT1b was amplified with primers in Supplementary Data 1 containing BamHI at the 5’-terminus and SmaI at the 3’-terminus and ligated into the pHBT vector under the 35S promoter with the TurboID-HA epitope tags. Fragments of p35S::POT1a-TurboID-HA and p35S::POT1b-TurboID-HA were amplified and sub-cloned into pCB302 to obtain pCB302-p35S::POT1a-TurboID-HA and pCB302-p35S::POT1b-TurboID-HA. cDNA or gDNA of AtPOT1a, AtPOT1b, CAT2, and RFP were amplified with primers in Supplementary Data 1 containing BamHI at the 5’-terminus and SpeI at the 3’-terminus and ligated into the pHBT vector under control of the 35S promoter with the nLuc and cLuc epitope tags and the fragments of p35S::POT1a-nLuc, p35S::POT1b-nLuc, p35S::POT1b-cLuc, p35S::CAT2-cLuc, p35S::RFP-nLuc, and p35S:: RFP-cLuc were amplified with primers in Supplementary Data 1, and sub-cloned into pCAMBIA1300 to obtain pCAMBIA1300-p35S::POT1a-nLuc, pCAMBIA1300-p35S::POT1b-nLuc, pCAMBIA1300-p35S::POT1b-cLuc, pCAMBIA1300-p35S::CAT2-cLuc, pCAMBIA1300-p35S::RFP-nLuc, and pCAMBIA1300-p35S::RFP-cLuc.

Transgenic plants were generated using A. tumefaciens-mediated floral dipping. Transgenic plants were screened by Basta (30 μg/ml) for the pBA and pCB302 vectors and hygromycin (30 μg/ml) for pCAMBIA1380 and confirmed by immunoblotting (IB) for protein expression.

Terminal restriction fragment (TRF) analysis, primer extension telomere repeat amplification (PETRA), and telomere repeat amplification protocol (TRAP and qTRAP)

Genomic DNA was extracted from 3-week-old plants of WT, pot1a, pot1b, pot1a pot1b, p35S::FLAG-Myc-POT1a/pot1a pot1b, pUBQ10::FLAG-Myc-POT1b/pot1a pot1b, pUBQ10::FLAG-Myc-PpPOT1/pot1a pot1b, and pUBQ10::FLAG-Myc-PpPOT1/pot1b. Bulk telomere length was measured by TRF analysis⁷⁴. Genomic DNA was digested with MseI (NEB, USA) restriction enzyme and subjected to Southern blot hybridization using [³²P] 5’radiolabeled (TTTAGGG)₄ oligonucleotide as a probe³³. Single telomere analysis was performed using PETRA^37,40. Following primer extension of the telomeric G-overhang and PCR amplification with chromosome arm–specific subtelomeric primers (Supplementary Data 1), products were detected by Southern blot hybridization with the same telomeric probe. Telomere length on Southern blots was measured using the online tool WALTER⁷⁵. Telomerase activity was examined in inflorescences from 5-week-old plants of WT, pot1a, pot1b, pot1a pot1b, p35S::FLAG-Myc-POT1a/pot1a pot1b, pUBQ10::FLAG-Myc-POT1b/pot1a pot1b, and pUBQ10::FLAG-Myc-PpPOT1/pot1a pot1b by TRAP and qTRAP assays^32,74. For TRAP, total protein extracts were incubated in primer extension buffer (50 mM Tris-OAc pH 8.0, 50 mM KCl, 3 mM MgCl₂, 2 mM DTT, 1 mM spermidine) together with 0.66 µM forward primer (Supplementary Data 1) and dNTPs at 37 °C for 45 min to allow telomerase-mediated extension. Reactions were terminated by phenol:chloroform:isoamyl alcohol extraction, and the aqueous phase was recovered. Extended products were precipitated by the addition of 0.1 volume 3 M sodium acetate, 2.5 µl glycogen (10 mg ml⁻¹) and 2.5 volumes 100% ethanol, followed by overnight incubation at −20 °C. Samples were centrifuged at 15,000 rpm for 30 min at 4 °C, washed with 70% ethanol, centrifuged again at 15,000 rpm for 10 min at 4 °C, air-dried and resuspended in 10 µl nuclease-free water. The recovered extension products were used as templates for PCR amplification with 0.4 µM forward primer, 0.4 µM reverse primer (Supplementary Data 1), 66 nM [α-³²P]-dGTP and 2× GoTaq® Hot Start Colorless Master Mix (Promega). PCR products were resolved on 6% denaturing polyacrylamide gels and visualized using a Typhoon FLA 9500 phosphorimager. For qTRAP, 10.5 µl of diluted protein extract (4.8 ng µl⁻¹) was combined with 0.4 µM forward primer (Supplementary Data 1) and SYBR Green PCR master mix in a total reaction volume according to the manufacturer’s instructions. Primer extension was carried out at 30 °C for 45 min. Subsequently, 0.4 µM reverse primer (Supplementary Data 1) was added, and quantitative PCR was performed for 35 cycles (95 °C denaturation and 60 °C annealing/extension). Threshold cycle (Ct) values were determined using an iCycler iQ real-time PCR system. Telomerase activity was calculated relative to wild-type controls.

Fluorescence in situ hybridization (FISH) analysis of telomeres

Inflorescences from 5-week-old plants of WT, pot1a, pot1b, and pot1a pot1b were collected and fixed in an acetic acid/ethyl alcohol (3:1) solution for 2–3 h at room temperature, then stored at −20 °C. Pistils were dissected and washed twice in dH₂O for 30 sec each, followed by two washes for 5 min in 0.01 M citrate buffer (4.45 mM sodium citrate, pH 4.5, 5.55 mM citric acid). Pistils were incubated in an enzyme solution containing 1% cellulase (Sigma-Aldrich, USA) and 1% pectolyase (Sigma-Aldrich, USA) in 0.01 M citrate buffer at 37 °C for 60 min. After enzymatic digestion, the pistils were incubated in cold citrate buffer for 15 min, transferred to a slide with 70 µL of 60% acetic acid, incubated for 5 min, squashed under a cover slip, and frozen in liquid nitrogen for 1 min. The coverslip was removed, and slides were washed with fixatives before air drying for 20 min. For FISH, slides were dehydrated in graded ethanol (70%, 90%, and 100%) for 2 min each and dried for 20 min at room temperature. The slides were denatured at 95 °C for 7 min, immediately cooled on ice, and hybridized with a mix containing 25 ng of fluorescently labeled oligonucleotide probes (Supplementary Data 1) in a solution of 10% dextran sulfate, 50% formamide, and 2× SSC (20× SSC: 3 M NaCl, 0.3 M sodium citrate, pH 7.0) for 5 min in the dark. After hybridization, the slides were washed in 2× SSC at room temperature for 5 min, followed by a wash for 10 min in 0.1× SSC at 43 °C and another wash for 5 min in 2× SSC at room temperature. The slides were then dehydrated again in graded ethanol, dried for 20 min, and mounted with a DAPI-containing mounting medium (Biotium, USA). Finally, the slides were sealed with nail polish and allowed to dry for 20 min before microscopy. Images were captured at 100× oil immersion using a Nikon Ti microscope.

Propidium Iodide (PI) staining

Six-day-old seedlings of WT, pot1a, pot1b, and pPOT1b::FLAG-Myc-POT1b/pot1b grown on ½MS media supplemented or not with 1% sucrose were transferred to dH₂O (mock) or dH₂O containing 20 μM of Zeocin (Thermo Fisher, USA). Samples were wrapped in aluminum foil and left on a shaker for 4 h or 6 h. After incubation, seedlings were transferred to PI stain solution (10 μg/ml; Sigma-Aldrich, USA) dissolved in dH₂O. After 5 min, seedlings were washed in dH₂O, transferred to slides in a droplet of dH₂O, sealed with a cover slip, and imaged at 10× using a dsRED filter and brightfield of a Zeiss fluorescence microscope.

Comet assay

The comet assay was performed as outlined below³⁸. A comet assay kit (Trevigen) was used in protoplasts isolated from 6-day-old seedlings of WT (G2, G3, and G4), pot1a (G2, G3, and G4), and pot1b (G2, G3, and G4)⁷⁶. A concentration of 2 × 10⁵ cells/ml was used for the assay. Slides were run in an electrophoretic setup at 18 V for 10 min in complete darkness. After drying the agarose, slides were stained with PI stain solution (100 μg/ml), sealed with a cover slip, and imaged using a Zeiss fluorescence microscope at ×5 magnification with a dsRED filter. The parameters (Percentage DNA in tail, Tail Length, Tail Moment) were calculated using OpenComet software⁷⁷.

G-overhang and telomere fusion PCR (TF-PCR)

The G-overhang assay was performed based on published methods⁷⁸. Briefly, genomic DNA was extracted from 3-week-old WT, pot1a, pot1b, and ku70 plants. One hundred fifty micrograms of the extracted DNA was digested with 40 units of MseI (NEB, USA) and resolved on a 1% agarose gel for 16 h. The bottom half of the gel was subjected to a denatured Southern blot using a 4× Telo probe (5’TTTAGGGTTTAGGGTTTAGGGTTTAGGG3’). The top half of the gel was subjected to native in-gel hybridization. A telomeric C-rich radioactive probe (5’CCCTAAACCCTAAACCCTAAACCCTAAA3’) complementary to the G-overhang sequence was used to determine the G-overhang signal for each sample. Quantity One was used to calculate the radioactive signal on the gels (both in-gel and Southern blot membrane). The Southern blot signal was used as the loading control. The in-gel signal was first normalized to the loading control, and then everything was normalized to the WT signal. TF-PCR was performed using 2 µg of gDNA extracted from 3-week-old plants of WT, pot1b (G2 and G4), stn1, and pot1a pot1b following published methods³⁷. Chromosome fusion junctions were amplified by PCR using subtelomeric primers targeting the left arm of chromosome 1 (1L) and the right arm of chromosome 2 (2R) or chromosome 4 (4R) listed in Supplementary Data 1. PCR reactions were carried out in 20 μl volumes containing ExTaq buffer (TaKaRa), 125 μM dNTPs, 0.5 μM of each primer, and ExTaq polymerase with 100 ng genomic DNA. Amplification was performed with an initial denaturation at 96 °C for 2 min, followed by 34 cycles of 96 °C for 30 s, 55 °C for 1 min, and 72 °C for 1 min, with a final extension at 72 °C for 5 min. PCR products were resolved by agarose gel electrophoresis, transferred to nylon membranes, and hybridized with a [³²P] labeled telomeric probe to detect fusion events.

Total RNA extraction and qRT-PCR analysis

Six-day-old seedlings of WT, pot1b, pPOT1b::FLAG-Myc-POT1b/pot1b (L1 and L2) grown on ½MS media were treated with 0.5 µM MV, 400 mM mannitol, and 150 mM NaCl for 3 h and 6 h, and total RNA was extracted using TRIzol reagent (Invitrogen, USA). One microgram of total RNA was treated with RNase-free DNase I (NEB, USA) and reverse transcribed to synthesize the first-strand cDNA with SuperScript IV Reverse Transcriptase (Invitrogen, USA) and oligo (dT)₂₀ primer. Quantitative RT-PCR (RT-qPCR) was performed using the PowerUp SYBR Green master mix (Thermo Fisher, USA) with primers listed in Supplementary Data 1 in a Bio-Rad CFX96 Real-Time PCR System (Bio-Rad, USA). Expression of indicated genes was normalized to Actin2.

RNA-seq

Total RNA was extracted from 6-day-old WT and pot1b seedlings treated with and without 20 μM Zeocin (Thermo Fisher) for 6 h using TRIzol reagent. Libraries in biological triplicate were prepared using the Illumina TruSeq Stranded Total RNA with Ribo-Zero Plant kit and sequenced on a NextSeq 2000, yielding ~30 million paired-end 2 × 50 bp reads per sample. All data analysis was performed on the Galaxy Europe platform (https://usegalaxy.eu)⁷⁹. For each biological replicate, lane-level FASTQs were concatenated using the Concatenate datasets tool. Read quality was assessed using FastQC, and adapter and low-quality sequences were removed with the Trim Galore! tool. Trimmed reads were aligned to the A. thaliana TAIR10_v90 genome with HISAT2, producing BAM files. Gene-level read counts were obtained from BAMs using featureCounts. Using the limma-voom tool, counts were normalized by TMM, voom-transformed, and differential expression assessed by empirical Bayes linear modeling. Finally, gene ontology enrichment was performed with g:Profiler⁸⁰.

ROS detection (DAB, H₂DCFDA, and DHE)

For H₂O₂ detection in leaves and flowers using 3, 3′-diamino benzidine (DAB), the tissues of WT, pot1a, pot1b, pot1a pot1b, pUBQ10::FLAG-Myc-POT1b/WT, p35S::FLAG-Myc-POT1a/pot1b, pUBQ10::FLAG-Myc-POT1b/pot1b, pUBQ10::FLAG-Myc-PpPOT1/pot1b grown on soil for 3–5 weeks and transiently expressed WT leaves after infiltration were sampled and placed in a 0.1% DAB staining solution containing 0.05% Tween 20. DAB solution was taken up by leaves and flowers by vacuum infiltrating the tissues for 5 min. Samples were covered with aluminum foil and incubated on a shaker for 4–5 h at 80–100 rpm shaking speed. The samples were then carefully transferred to a bleaching solution (30% ethanol, 10% acetic acid, and 10% glycerol in dH₂O) and incubated in a 65 °C water bath for 30 min to bleach out the chlorophyll. After 30 min of incubation, the bleaching solution was replaced with a fresh bleaching solution and allowed to stand for 30 min. For general ROS detection in seeds using 2′,7′-dichlorofluorescein diacetate (H₂DCFDA), imbibed seeds of WT, pot1b, pUBQ10::FLAG-Myc-POT1b/WT, p35S::FLAG-Myc-POT1a/pot1b, pUBQ10::FLAG-Myc-POT1b/pot1b, pUBQ10::FLAG-Myc-PpPOT1/pot1b, p35S::CAT2-HA/pot1b, pUBQ10::FLAG-Myc-hPOT1/pot1b, pUBQ10::FLAG-Myc-GUS/pot1b in dH₂O for 3 days were incubated with 5 µM H₂DCFDA for 5 min, then washed by dH₂O. The fluorescence was investigated at 10× using a fluorescein filter and brightfield of a Zeiss fluorescence microscope. For H₂O₂ and O²⁻ detection in roots using peroxy orange1 (PO1) and dihydroethidium (DHE), respectively, 4-day-old seedlings of WT, pot1b, and cat2 were incubated with 40 µM PO1 or 40 µM DHE for 5 min and rinsed with dH₂O. The fluorescence was visualized with a Cy3 filter and a brightfield of a Zeiss fluorescence microscope. For ROS detection in the nucleus by H₂DCFDA and Hoechst under the confocal microscope, 5 µM H₂DCFDA and 1 µg/ml Hoechst solutions were treated with 4-day-old seedlings of WT, pot1b, pUBQ10::FLAG-Myc-POT1b/pot1b, and cat2 grown on ½MS without sucrose for 5 min. The fluorescent signal was detected 10 min later. Images were captured using a Leica SP8 confocal microscope with a 488 nm excitation and 500–550 nm emission for H₂DCFDA, and Hoechst fluorescence was detected with a 361 nm excitation and 486 nm emission.

Activity assays of ROS scavenging enzymes

One hundred milligrams of 6-day-old seedlings of WT, pot1a (G2 and G3), pot1b (G2 and G3), and pPOT1b::FLAG-Myc-POT1b/pot1b (L1 and L2) were ground in liquid nitrogen and homogenized in 0.5 ml of the extraction buffer (50 mM potassium phosphate buffer, pH 7.0, 2 mM EDTA, 1% polyvinylpyrrolidone, 1 mM phenylmethylsulfonyl fluoride). The homogenates were centrifuged at 15,000 × g for 20 min at 4 °C. The supernatant was used for the enzymatic assays. The ascorbate peroxidase activity was measured as described below⁸¹. The reaction mixture (50 mM potassium phosphate, 0.5 mM ascorbate, 0.2 mM hydrogen peroxide) was mixed with 10 μl of protein extract. The absorbance was recorded at 290 nm. The activity was calculated using the extinction coefficient of 2.8 mM⁻¹ cm⁻¹ and expressed as nmol H₂O₂ min⁻¹ mg⁻¹ protein. The peroxidase activity was measured using published methods⁸². Briefly, the reaction mixture (50 mM potassium phosphate, 3.2 mM guaiacol, 0.4 mM hydrogen peroxide) was mixed with 10 μl of protein extract. The absorbance was recorded at 470 nm. The activity was calculated using the extinction coefficient of 0.0266 μM⁻¹ cm⁻¹ and expressed as μmol tetra-guaiacol min⁻¹ mg⁻¹ protein. The catalase activity was measured as described⁸³. Briefly, the reaction mixture (50 mM potassium phosphate, 10 mM hydrogen peroxide) was mixed with 10 μl of protein extract. The absorbance was recorded at 240 nm. The activity was calculated using the extinction coefficient of 0.036 μM⁻¹ cm⁻¹ and expressed as μmol min⁻¹ mg⁻¹ protein.

Pollen viability assay

Pollen viability was performed as described below⁸⁴. Two milligrams per milliliter fluorescein diacetate (FDA) stock was prepared in acetone (1000×) and diluted to 0.5× in BK buffer (10 mM MOPS, pH 7.5, 0.127 mM Ca(NO₃)₂·4H₂O, 0.081 mM MgSO₄·7H₂O, 0.1 mM KNO₃, 15% sucrose). Microscopy slides were cleaned with ethanol before mounting pollen grains. Flowers of WT (G2, G3, and G4), pot1a (G2, G3, and G4), and pot1b (G2, G3, and G4) were brushed to the side of the slide, and the FDA-buffer mixture was dropped, followed by placing a coverslip on top. Slides containing pollen grains were imaged using a GFP filter (wavelength of blue light, 495 nm) on a Zeiss fluorescence microscope.

Subcellular localization and confocal microscopy

For the determination of AtPOT1a and AtPOT1b localization, 4-week-old N. benthamiana leaves were co-infiltrated with the indicated constructs. The fluorescence signal in the leaves was detected using Amersham Imager 600, and the signals in the cells were detected using a Zeiss fluorescence microscope. Transgenic plants expressing pPOT1a::Myc-sfGFP11-POT1a, pPOT1b::Myc-sfGFP11-POT1b, or pPOT1b::POT1b-Myc-sfGFP11 in the pUBQ10::sfGFP1-10-NLS were grown on ½MS without sucrose for 4 days. GFP fluorescence signals in the plants were examined using a Leica TCS SP8 laser scanning confocal microscope (Germany) with a 488 nm excitation and 490–530 nm emission. The pinhole was set at 1 Airy unit. Image analyses were performed using Leica Application Suite X (LAS X) software.

Protoplasts were transfected with p35S::POT1a-GFP and p35S::POT1b-mCherry constructs. p35S::PARP2-GFP was used as a nuclear protein control⁸⁵. Fluorescence signals in protoplasts were examined 12 h after transfection using the same confocal microscope. The excitation and emission wavelengths of GFP are the same as above. For the mCherry signal, the excitation and emission wavelengths are 588 nm and 590–620 nm, respectively.

Nuclear and cytoplasmic fractionation

Nuclear and cytoplasmic fractionation was performed as described⁸⁶. Briefly, 0.5 g of 6-day-old seedlings of WT and pUBQ10::POT1b-Myc-sfGFP11/pot1b grown on ½MS media with and without sucrose were frozen in liquid nitrogen, ground to fine powder and homogenized in lysis buffer (20 mM Tris-HCl, pH 7.4, 25% glycerol, 20 mM KCl, 2 mM EDTA, 2.5 mM MgCl₂, 250 mM sucrose, 1 mM DTT, 1 mM PMSF, and cOmplete™) on ice. The homogenate was sequentially filtered through a 40-μm nylon mesh. The nuclei were subjected to centrifugation at 1500 × g for 10 min at 4 °C, and the pellet of nuclei and supernatant of cytoplasmic fractions were collected. The pellet was washed four times with resuspension buffer (20 mM Tris-HCl, pH 7.4, 25% glycerol, 2.5 mM MgCl₂, 0.2% Triton X-100, 1 mM DTT, 1 mM PMSF, and cOmplete™) at 4 °C. The supernatant was centrifuged at 16,000 × g for 10 min and rescued, which was repeated once more. The nuclear pellet and cytoplasmic fractions were resuspended in fresh buffer with Laemmli sample buffer and heated at 95 °C for 10 min. The samples were centrifuged at 16,000 × g for 10 min. The supernatant was collected and then subjected to SDS-PAGE. After the run, the proteins were electrically transferred onto a PVDF membrane and then detected with an antibody, including α-Myc (Invitrogen, USA), α-CAT (Agrisera, Sweden), and α-CAT2 (PhytoAB, USA). α-Histone H3 (Agrisera, Sweden) and α-Actin (Agrisera, Sweden) antibodies were used for fractionation quality validation and internal control.

Telomere chromatin immunoprecipitation (ChIP)-qPCR

Telomere ChIP-qPCR was performed based on published methods⁸⁷. Briefly, 0.5 g of 6-day-old seedlings grown on ½MS media with and without sucrose were fixed in 1% formaldehyde with a vacuum for 10 min and quenched in a solution of 125 mM glycine with a vacuum for 5 min. The seedlings were washed three times with 40 ml dH₂O and then frozen in liquid nitrogen. Chromatin was isolated and resuspended in nuclear lysis buffer (30 mM Tris-HCl, pH 8.0, 10 mM EDTA, 1% SDS, and Roche protease inhibitor cocktail). Chromatin was fragmented to an average size of 200–500 bp using a Bioruptor (Diagenode). After centrifugation at 10,000 g for 5 min, the supernatant was diluted 10-fold with dilution buffer (1.1% Triton X-100, 1.2 mM EDTA, 16.7 mM Tris-HCl, pH 8.0, 167 mM NaCl). α-Myc (Sigma-Aldrich, USA) and α-CAT (Agrisera, Sweden) antibodies cross-linked with Dynabeads protein A magnetic beads (Thermo Fisher, USA) were used to immunoprecipitate the protein-DNA complex. After reverse cross-linking and digestion by proteinase K, the enriched telomeric DNA was measured by qPCR using the primers listed in Supplementary Data 1.

Genomic and telomeric oxidation (8-oxoG ELISA)

Genomic DNA was extracted from 6-day-old seedlings of WT, cat2, pot1a, pot1b, pot1a pot1b, p35S::FLAG-Myc-POT1a/pot1b, pUBQ10::FLAG-Myc-POT1b/pot1b, p35S::CAT2-HA/pot1b, and pUBQ10::FLAG-Myc-PpPOT1/pot1b grown on ½MS medium with and without sucrose. 8-oxoG was measured using the DNA Damage Competitive ELISA kit (Invitrogen, USA), following the manufacturer’s instructions. Genomic and telomeric enrichment coupled with 8-oxoG ELISA was performed as discussed below⁵⁵. For telomeric 8-oxoG detection, 50 μg of total genomic DNA was digested overnight with 30 U of MseI at 37 °C. Following digestion, large DNA fragments were enriched using solid phase reversible immobilization (SPRI) with SPRIselect magnetic beads (Beckman Coulter, USA) at a bead-to-sample ratio of 0.4, following the manufacturer’s left-side size selection protocol. Beads were incubated for 1 min to minimize retention of small DNA fragments. The enriched DNA fraction was subsequently subjected to competitive ELISA to quantify telomeric 8-oxoG levels.

Yeast two-hybrid (Y2H) assay

Y2H was performed using the Gold Yeast Two-Hybrid System according to the manufacturer’s manual (Clontech). Coding sequence DNA from different genes, combined with pGADT7 and pGBKT7, was co-transformed into the yeast strain AH109. The yeast transformants were simultaneously plated on medium minus Leu, Trp, His, Ade or minus Leu, His, Ade to screen the interactions and medium minus Leu, Trp as controls under 28 °C.

Protoplast isolation and co-immunoprecipitation (Co-IP) assays

Protoplast isolation and gene expression assays following published protocols⁷⁶. Briefly, for protoplast-based gene expression assays, 6–30 μg of DNA constructs indicated in the figures were transfected into 5 × 10⁴ protoplast cells of 3-week-old WT plants using the PEG-transfection method. The protoplasts were incubated at 25 °C for 8 h and treated with 0.5 μM methyl viologen (MV) for the indicated time in the figures. For protoplast-based Co-IP assays, protoplasts were collected by centrifugation and lysed in 300 μl IP buffer (50 mM Tris-HCl, 150 mM NaCl, pH 7.5, 5 mM EDTA, 0.5-1% Triton X-100, protease inhibitor cocktail, 2 mM DTT, 2 mM NaF, and 2 mM Na₃VO₃) by vortexing. After centrifugation at 10,000 × g for 5–10 min at 4 °C, 30 μl of supernatant was collected for input control, and 3 μl α-FLAG M2 magnetic beads (Sigma-Aldrich, USA) were added into the remaining supernatant and incubated at 4 °C for 1–3 h. Beads were collected and washed three times with washing buffer (50 mM Tris-HCl, 150 mM NaCl, pH 7.5, 5 mM EDTA, 0.5% Triton X-100) and once with 50 mM Tris-HCl, pH 7.5. Immunoprecipitates were analyzed by immunoblotting (IB) with the indicated antibodies. For N. benthamiana-based Co-IP assays, 4-week-old N. benthamiana leaves were co-infiltrated with the indicated constructs. Total protein was extracted with the IP buffer (40 mM Tris-HCl, pH 7.5, 150 mM NaCl, 5 mM MgCl₂, 2 mM EDTA, 0.5% Triton X-100, 5% glycerol, 2 mM PMSF, EDTA-free protease inhibitor mixture (Roche), and 5 mM DTT). Soluble proteins were cleared by centrifugation and then immunoprecipitated with α-Myc antibody (Sigma-Aldrich, USA) cross-linked with Dynabeads protein A magnetic beads (Thermo Fisher, USA). Co-IP assay was performed as described above.

Split luciferase complementation imaging (LCI) assay

4-week-old N. benthamiana leaves were infiltrated with an equal volume of combinations of A. tumefaciens carrying the derivative constructs of pCAMBIA-nLuc and pCAMBIA-cLuc. After infiltration, plants were cultured for 48 h under a 16 h light/8 h dark photoperiod. After infiltrating 1 mM luciferin, the chemiluminescence signals were captured by a NightSHADE LB 985 in vivo plant imaging system (Berthold, Germany) equipped with a CCD camera.

TurboID-based proximity labeling and mass spectrometry analysis

The TurboID-based proximity labeling was performed based on published methods⁸⁸. Briefly, 1 g of 6-day-old seedlings of p35S::POT1a-TurboID-HA/pot1a, p35S::POT1b-TurboID-HA/pot1b, and p35S::TurboID-HA/Col-0 (negative control), grown on ½MS medium with and without sucrose, were incubated in 250 µM biotin for 1 h, collected, and frozen in liquid nitrogen. Three independent biological replicates were performed per genotype and condition. Frozen material was ground to a fine powder, and proteins were extracted in 2.5 ml RIPA lysis buffer (50 mM Tris-HCl, pH 7.5, 500 mM NaCl, 1% NP40, 0.1% SDS, 0.5% sodium deoxycholate, 1 mM DTT, Protease inhibitor cocktail). After vortexing, the samples were centrifuged at 16,500 × g for 10 min, and the supernatant was run through the Zeba^TM Spin Desalting columns (Thermo Fisher, USA) to remove the excess biotin in lysates. A fraction of the lysate was removed from the desalted extracts and used to estimate the protein concentration. To enrich biotinylated proteins from the protein extracts, 80 μl of streptavidin-coated magnetic beads (Thermo Fisher, USA) were washed twice with RIPA lysis buffer, and the desalted lysates (~6 mg proteins) were then incubated with the equilibrated beads on a rotator overnight at 4 °C. The beads were sequentially washed once with 1 ml buffer I (2% SDS), once with buffer II (50 mM HEPES, pH 7.5, 500 mM NaCl, 1 mM EDTA, 0.1% deoxycholic acid, 1% Triton X-100), and once with buffer III (10 mM Tris-HCl, pH7.4, 250 mM LiCl, 1 mM EDTA, 0.1% deoxycholic acid, 1% NP40). To completely remove the potential detergent, the beads were washed twice in 50 mM Tris-HCl, pH 7.5, and six more times in 50 mM ammonium bicarbonate, pH 8.0. Finally, the beads were resuspended in 1 ml of 50 mM ammonium bicarbonate. Three biological replicates were performed. To confirm the enrichment of the biotinylated proteins, 10% of the suspension was taken for western blotting, and the remaining beads were frozen in liquid nitrogen and submitted to the UT Southwestern Proteomics Core for LC–MS/MS analysis. LC–MS/MS analysis was performed using either an Orbitrap Fusion Lumos or Q Exactive HF mass spectrometer (Thermo Fisher Scientific). Proteins were digested with trypsin prior to analysis. Raw data were processed using Proteome Discoverer version 3.0 (Thermo Fisher Scientific) and searched against the TAIR10 A. thaliana protein database (32,785 entries) using the Sequest HT node within Proteome Discoverer. Trypsin was specified as the protease with up to three missed cleavages allowed. Fixed modifications included carbamidomethylation of Cys, and variable modifications included oxidation of Met. Precursor ion mass tolerance was set to 10 ppm, and fragment ion mass tolerance was set to 0.6 Da for Orbitrap Fusion Lumos runs and 0.02 Da for Q Exactive HF runs. Peptide identifications were filtered using the Percolator node within Proteome Discoverer with a 1% false discovery rate (FDR) at the peptide level. Protein-level identifications were classified as High (1% FDR), Medium (5% FDR), or Low (>5% FDR). High- and medium-confidence proteins were retained for downstream analysis, including proteins supported by one or more unique peptides. Relative protein quantification was based on summed peptide peak intensities identified by the Minora Feature Detector node.

Transient expression in Arabidopsis

The Arabidopsis transient assays were performed as described⁸⁹ and outlined below. A. tumefaciens strain GV3101 carrying the indicated constructs was cultured in LB broth containing corresponding antibiotics and subcultured in Agrobacterium (AB) minimal medium containing 50 mM MES, pH 5.5, 2% glucose, and 200 μM acetosyringone for 12–16 h at 28 °C with shaking. The cells were collected and resuspended in infiltration buffer (½ AB medium, ¼MS medium, 25 mM MES, pH 5.5, 2% glucose, 200 μM acetosyringone, and freshly added 0.01% Triton X-100) at an OD₆₀₀ of 1.0, and the suspension was infiltrated into the three to four largest leaves of 3-week-old Arabidopsis plants. After infiltration, plants were cultured for 48 h under a 16 h light/8 h dark photoperiod and sampled.

8-oxoG DNA immunoprecipitation-qPCR

8-oxoG-enriched genomic fragments were obtained as described⁹⁰. Briefly, genomic DNA was extracted, and 4 μg of genomic DNA per immunoprecipitation was sonicated in 100 μl TE buffer using a Bioruptor (Diagenode). The fragmented DNA was denatured and immunoprecipitated overnight at 4 °C with 4 μl polyclonal antibodies against 8-oxoG (AB5830, Millipore) in a final volume of 500 μl IP buffer (110 mM NaH₂PO₄, 110 mM NaH₂PO₄, pH 7.4, 0.15 M NaCl, 0.05% Triton X-100, 100 mM Tris-HCl, pH 8.0, 0.5 M EDTA, pH 8.0). Then, 50 μl Dynabeads Protein G (10003D, ThermoFisher Scientific, previously blocked with 0.5% bovine serum albumin diluted in IP buffer for overnight) was added for 3 h at 4 °C, under constant rotation, and washed five times with 1 ml wash buffer (110 mM NaH₂PO₄, 110 mM Na₂HPO₄ pH 7.4, 0.15 M NaCl, 0.05% Triton X-100). The immunocomplexes were then disrupted by incubation in 110 μl elution buffer (100 mM Tris-HCl, pH 8.0, 10 mM EDTA, pH 8.0, 1% SDS, 0.5 mg/ml Proteinase K) overnight at 37 °C, and 1 h at 52 °C following additional incubation. 8-oxoG-enriched telomeric DNA was measured by qPCR using the primers listed in Supplementary Data 1⁵⁵.

Quantification and statistical analysis

The images of TRF and PETRA were quantified with WALTER⁷⁵. Statistical analyses were performed using GraphPad Prism 9. One-way, two-way, or Brown-Forsythe and Welch ANOVA was used depending on experimental design and group variance assumptions. Post hoc tests were selected based on comparison type and ANOVA model: Tukey’s test was used for pairwise comparisons, Dunnett’s test for comparing multiple groups to a single control. Two-way ANOVA followed by Šidák’s test was used to compare treatment effects within each genotype. Specific tests and the number of biologically independent replicates are shown in the figure legend. The p-values are indicated in the graphs.

Reporting summary

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

Author contributions

J.-H.M., C.C.-G., B.B.B., I.-C.Y., X.X., S.B., F.A.O.-M., D.L., J.S., P.H., J.Z., E.V.S., T.E.J., and D.E.S. conceptualized and designed the experiments; J.-H.M., C.C.-G., B.B.B., I.-C.Y., X.X., S.B., F.A.O.-M., D.L., E.C., D.C., M.Y., J.S., P.U., and C.P. performed experiments and analyzed data; J.-H.M., C.C.-G., B.B.B., A.T.W., E.V.S. and D.E.S. wrote and edited the paper.

Peer review

Peer review information

Nature Communications thanks Alison Baker, Peisheng Mao and the other anonymous reviewer(s) for their contribution to the peer review of this work. A peer review file is available.

Data availability

Raw RNA-seq data generated in this study were deposited in the NCBI Gene Expression Omnibus under the accession number GSE297308. Proteomics data generated in this study were deposited in the ProteomeXchange Consortium via the MassIVE repository under accession code PXD074350. 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 10.1038/s41467-026-70441-z.