RGF1 controls root meristem size through ROS signaling

Abstract

Stem cell niche and root meristem size are maintained by intercellular interactions and signaling networks of a peptide hormone, Root Meristem Growth Factor 1 (RGF1). How RGF1 regulates root meristem development is an essential question to understand stem cell function. Although five receptors of RGF1 have been identified, the downstream signaling mechanism remains unknown. Here, we report a series of signaling events following RGF1 action. The RGF1-receptor pathway controls the distribution of reactive oxygen species (ROS) along the developmental zones of the Arabidopsis root. We identify a novel transcription factor, RGF1 INDUCIBLE TRANSCRIPTION FACTOR 1 (RITF1) that plays a central role in mediating RGF1 signaling. Manipulating RITF1 expression leads to redistribution of ROS along the root developmental zones. Changes in ROS distribution, in turn, enhance the stability of the PLETHORA2 (PLT2) protein, a master regulator of root stem cells. Our study, thus, clearly depicts a signaling cascade initiated by RGF1, linking the RGF1 peptide to ROS regulatory mechanisms.

Roots encounter various environmental conditions and respond by altering their growth. Root growth arises through controlled cell division in the meristematic zone, equivalent to the transit amplifying zone in animals. After division, most cells increase their size in the elongation zone and mature in the differentiation zone. The size of these developmental zones is determined by intrinsic and extrinsic signals. Reactive oxygen species (ROS) are an intrinsic signal for establishing the size of the meristematic zone. Superoxide (O₂⁻) primarily accumulates in the meristematic zone, while hydrogen peroxide (H₂O₂) mainly accumulates in the differentiation zone^1,2. The balance between O₂⁻ and H₂O₂ modulates the transition from proliferation to differentiation².

The RGF1 peptide is an essential hormone in controlling the size of the meristematic zone both as an intrinsic and extrinsic signal ^3-5. RGF1 treatment increases the size of the meristematic zone, while the rgf1/2/3 triple mutant has a smaller meristematic zone³. Quintuple mutants of the rgf1 receptor (rgfr) lack most cells in the root meristem and are insensitive to RGF1^6-8. RGF1 signaling controls the stability of the PLETHORA (PLT) 1/2 proteins³, which are required for stem cell maintenance⁹. However, it is not known how RGF1 modulates the size of the meristematic zone and the stability of PLT1/2.

After RGF1 treatment, the meristematic zone-specific marker HIGH PLOIDY2 (HPY2)-GFP¹⁰ was detected in an enlarged area, correlating with a larger meristematic zone (Extended Data Fig. 1a-c), suggesting RGF1 controls gene expression primarily in the meristematic zone. Therefore, to identify target genes downstream of RGF1, we isolated the meristematic zone one hour after RGF1 treatment (Extended Data Fig. 1d). Since HPY2-GFP expression and the meristematic zone size were unchanged in this time period, we can exclude the possibility that an enlarged meristem is the reason for changes in RNA levels. RNA-seq profiling found 583 differentially expressed genes between RGF1 and mock treatment (Supplementary Table 1). Gene Ontology highly enriched categories included “glutathione transferase activity” and “oxidoreductase activity” (Extended Data Fig. 2 and Supplementary Table 2), suggesting RGF1 might signal through an ROS intermediate.

To examine the relationship between RGF1 and ROS signaling, we analyzed the distribution of O₂⁻ and H₂O₂ after RGF1 treatment. The specific indicator for H₂O₂, H₂O₂-3’-O-Acetyl-6’-O-pentafluorobenzenesulfonyl-2’-7’-difluorofluorescein-Ac (H₂O₂-BES-Ac)², exhibited lower fluorescence in the meristematic and elongation zones 24 h after RGF1 treatment (Fig. 1a and c). O₂⁻ signals were detected by nitro blue tetrazolium (NBT) staining¹ and were observed more broadly in the meristematic zone 24 h after RGF1 treatment (Fig. 1b and d). In the RGF1 receptor mutant rgfr1/2/3, whose meristematic zone is unchanged after RGF1 treatment (Fig.1e), levels of H₂O₂ and O₂⁻ were comparable between mock and RGF1 treatments (Fig. 1e-h).

Figure 1. Distribution of ROS levels upon RGF1 treatment.

(a) Confocal images of roots 24 h after treatment with mock and 20 nM RGF1. Propidium iodide (PI) staining (Red), H₂O₂-BES-Ac fluorescence (Green). (b) Roots stained with NBT 24 h after treatment with mock or 20 nM RGF1. (c) Quantification of H₂O₂-BES-Ac intensity in the meristematic zone (n = 6 independent roots, p < 0.003). (d) Quantification of NBT staining intensity (arbitrary units (A.U.)) in the meristematic zone (n = 7 independent roots, p < 0.001). (e) Confocal images of roots 24 h after treatment with mock or 20 nM RGF1 in wild type (Col-0) or rgfr 1/2/3. Staining as in (a). (f) Quantification of H₂O₂-BES-Ac staining intensity in the meristematic zone in wild type and rgfr 1/2/3 (n = 5 independent roots, *p < 0.03). (g) Roots stained with NBT 24 h after treatment with mock or 20 nM RGF1 in wild type or rgfr 1/2/3. (h) Quantification of NBT staining intensity (A.U. values) in the meristematic zone in wild type or rgfr 1/2/3 (n = 5 independent roots, *p < 0.001). White and blue arrowheads indicate junction between meristematic and elongation zones. Scale bar = 50 μm. Bar graphs represent mean. Error bars are ± SD. Dots indicate each data point. P values calculated by two-sided Student’s t-test.

To identify downstream factors in the RGF1/ROS signaling pathway, we combined our RGF1 transcriptome data with developmental zone-specific transcriptome data¹¹. Among genes that are both meristematic zone-specific and induced by RGF1, we identified the PLANT AT-RICH SEQUENCE and ZINC-BINDING TRANSCRIPTION FACTOR (PLATZ) FAMILY PROTEIN gene (AT2G12646) whose expression increased approximately 2-fold after 1 hour of RGF1 treatment (Fig. 2a). We named this gene, RGF1 INDUCIBLE TRANSCRIPTION FACTOR 1 (RITF1), and found that its expression is predominantly in the meristematic zone¹¹ (Fig. 2b). Quantitative RT-PCR showed RITF1 transcript abundance increased approximate 2-fold in wild type one hour after RGF1 treatment, and was maintained at 6 and 24 hours (Fig. 2c). By contrast, RITF1 expression in rgfr1/2/3 was unchanged upon RGF1 treatment (Fig. 2c). Expression of a construct with the RITF1 promoter driving the GFP coding sequence (pRITF1-GFP) mirrored our transcriptome analysis and increased in wild type following RGF1 treatment (Fig. 2b, d, and e). By contrast, pRITF1-GFP expression was very low and exhibited no change following RGF1 treatment in rgfr1/2/3 (Fig. 2d and e). These data indicate that RITF1 expression is regulated by the RGF1 pathway.

Figure 2. Expression of RITF1 and phenotype of RITF1 overexpression line.

(a) Expression of RITF1 in meristematic zone 1 h after 100 nM RGF1 treatment measured by RNA-seq (CPM, counts per million mapped reads) (n = 3 independent experiments, p < 0.01) (b) Expression of RITF1 in developmental zones measured by RNA-seq (FPKM, Fragments Per Kilobase of transcript per Million mapped reads). (c) Expression of RITF1 in meristematic zone of wild type and rgfr1/2/3 upon RGF1 treatment measured by qRT-PCR. (n = 3 independent experiments, *p<0.001, **p<0.002, and ***p<0.02). (d) Confocal images of pRITF1-GFP and PI stained root from wild type and the rgfr1/2/3 after RGF1 treatment. (e) Total intensity (arbitrary unit A.U.) of pRITF1-GFP expression in wild type and rgfr1/2/3 24 h after RGF1 treatment (n = 5 independent roots, *p < 0.001). Confocal images of roots stained with PI (f) and H₂O₂-BES-Ac (g) in Col-0 and XVE-RITF1 in Col-0 after mock or Estradiol treatment. (h) Light microscope images of NBT stained roots after mock or Estradiol treatment. (i) Number of cells in the meristematic zone in Col-0 and XVE-RITF1 after mock or Estradiol treatment. (n = 6 independent roots *p < 0.001). (j) Average intensity of BES-H₂O₂-Ac in differentiation zone after mock or Estradiol treatment (n = 6 independent roots, *p < 0.001). (k) Average intensity of NBT staining in differentiation zone after mock or Estradiol treatments (n = 7 independent roots, *p < 0.001). Scale bar = 50 μm. White and blue arrowheads indicate junction between meristematic and elongation zones and between elongation and differentiation zones. Bar graphs represent mean. Error bars are ± SD. Dots indicate each data point. P values calculated by two-sided Student’s t-test.

To understand its function, RITF1 was inducibly overexpressed using the estradiol inducible promoter system¹². After 24 h of β-estradiol treatment, the meristematic zone became enlarged and the number of cells increased (Fig. 2f and i) similar to that of RGF1 treated roots (Fig. 1a). We also found that H₂O₂ levels declined in all three developmental zones upon estradiol treatment (Fig. 2g and j) and enhanced O₂⁻ signals were observed in a broader area of the meristematic zone (Fig. 2h and k) with ectopic O₂⁻ signals in the elongation and differentiation zones (Fig. 2h). Altered ROS signals and an enlarged meristem suggest that RITF1 can modulate ROS signaling and root meristem size downstream of the RGF1 pathway. We also observed an earlier response to induction of RITF1 than to RGF1 treatment. A decrease in the BES-H₂O₂-Ac signal was detected 4 h after estradiol treatment (Extended Data Fig. 3a and b) as compared to no detectable change 4 h after RGF1 treatment in either the uninduced line or in wild type (Extended Data Fig. 3a and b). Changes in ROS signals were first observed at approximately 6 h after RGF1 treatment in those lines (Extended Data Fig. 4i, j, o, p, 5b, and c). If RITF1 functions downstream of the RGF1-receptor pathway, overexpression of RITF1 in rgfr1/2/3 should rescue root meristem defects and increase root meristem size. To test this hypothesis, we inducibly overexpressed RITF1 in rgfr1/2/3 and in wild type and observed an enhanced O₂⁻ signal and increased root meristem size in both (Fig. 3a-d). Finally, we examined two ritf1 mutant alleles. The ritf1-1 allele was generated using CRISPR/Cas9 and contains a frameshift mutation early in the coding sequence, rendering it unlikely to produce a functional RITF1 protein. The ritf1-2 allele has a T-DNA insertion in the intron, but still has low expression of full length RITF1 and is likely to produce low levels of a functional protein. The ritf1-1 mutant had a smaller meristem and lower root growth rate (Extended Data Fig. 6a and b) and was more resistant to RGF1 treatment than wild type or the weak allele, ritf1-2 (Extended Data Fig. 6b and c). Further, there was lower induction of the O₂⁻ signal in ritf1-1 after RGF1 treatment than wild type or ritf1-2 (Fig. 3e and f). Taken together, these results strongly suggest that RITF1 is a primary regulator of ROS signaling and root meristem size in the RGF1 signaling pathway.

Figure 3. ROS signals and meristem size of RITF1 overexpression lines in rgfr1/2/3.

(a) Light microscope images of NBT stained roots with or without XVE-RITF1 in Col-0 and rfgr1/2/3. (b) Total intensity of NBT staining in differentiation zone with or without XVE-RITF1 in Col-0 and rgfr1/2/3 24 h after mock or Estradiol treatments. (n = 8 independent roots, *p < 2.0E-05). (c) Confocal images of PI stained roots with or without XVE-RITF1 in Col-0 and rfgr1/2/3. (d) Percent increase (100% = number of cells in mock treatment) in number of cells in the meristematic zone 24 h after Estradiol treatment compared with mock in Col-0, XVE-RITF1 in Col-0, rgfr1/2/3, and XVE-RITF1 in rgfr1/2/3. (n = 6 independent roots, *p< 0.0001, **p< 0.001). (e) Light microscope images of roots of Col-0, ritf1-1, and ritf1-2 stained with NBT 24h after 5nM RGF1 treatment. Scale bar=50 μm. Blue arrowheads show junction between meristematic and elongation zones. (f) Quantification of NBT staining intensity (A.U.) in meristematic zone in Col-0, ritf1-1, and ritf1-2 after 5 nM RGF1 treatment. (n = 7 independent roots, *p < 2.40495E-05, **p < 0.025). Bar graphs represent mean. Error bars are ± SD. Dots indicate each data point. P values calculated by two-sided Student’s t-test.

To confirm post-translational regulation of PLT2, we compared transcriptional (pPLT2-CFP)¹³ and translational (gPLT2-YFP)¹³ fusion lines. At 24 h after RGF1 treatment, we observed broader localization of gPLT2-YFP (Extended Data Fig. 7b), while the localization and expression of pPLT2-CFP were comparable between Mock and RGF1 treatments, even though RGF1-treated roots had a larger meristematic zone (Extended Data Fig. 7a). The gPLT2-YFP signal decreased more gradually and was broadly localized in the larger meristematic zone after RGF1 treatment (Extended Data Fig. 7a-c). These results confirm that RGF1 regulates PLT2 post-translationally. PLT2 is a member of the APETALA2 /ETHYLENE RESPONSE FACTOR transcription factor family, which is reported to be regulated by oxidative post-translational modification^14-19. To determine if modifying the oxidative conditions can increase the stability of the PLT2 protein, we treated the gPLT2-YFP line with RGF1 and potassium iodide (KI), an H₂O₂ scavenger. We found that gPLT2-YFP was localized more broadly and meristem size increased compared to roots treated only with RGF1 (Fig. 4a-c). By contrast, increased H₂O₂ inhibited broad localization of gPLT2-YFP and reduced the increase of meristem size upon addition of RGF1 (Extended Data Fig. 8a-e). To decrease O₂⁻ levels, we used 500 nM diphenyleneiodonium (DPI), an NADPH oxidase inhibitor (Fig. 4d-f), resulting in slight inhibition of PLT2 stability and meristem size (Fig. 4d-f) with little effect on root meristem development. However, co-treatment of RGF1 and DPI dramatically reduced PLT2 stability and meristem size as compared to RGF1 treatment alone (Fig. 4d-f). Finally, we measured gPLT2-YFP, O₂⁻, and H₂O₂ in a time course (4-10 h) after RGF1 treatment. Broader localization of gPLT2-YFP and increased superoxide levels along with lower H₂O₂ signals at the distal end of the meristematic zone appeared 6h after treatment (Extended Data Fig. 4a-d, and 5a; 4i and j; 4o and p). At 8 and 10 h after treatment, expanded gPLT2-YFP expression and O₂⁻ signals correlated with declining H₂O₂ signals (Extended Data Fig. 4e-h, 4k-n, 4q-t, 5a-c). Taken together, these results indicate that ROS regulates PLT2 protein stability by modulating O₂⁻ and H₂O₂ levels.

Figure 4. Stability of PLT2 protein upon change in oxidation conditions.

(a and d). Confocal images of gPLT2-YFP 24 h after RGF1, KI (H₂O₂ scavenger), and DPI (NADPH oxidase inhibitor) treatment. (b and e) Localization of gPLT2-YFP upon RGF1 and KI treatment (n = 7 independent roots, *p<0.015) and RGF1 and DPI (n = 7 independent roots, *p<1.5E-05) treatment. (c and f) Meristem size upon RGF1 and KI (n = 7 independent roots, *p<0.002) and RGF1 and DPI (n = 7 independent roots, *p<3.0E-07) treatment. Bar graphs represent mean. Error bars are ± SD. Dots indicate each data point. P values calculated by two-sided Student’s t-test.

To further test the hypothesis that PLT2 protein stability is enhanced by ROS signaling produced by RITF1, we overexpressed RITF1 in the plt2 mutant. This produced an increase in the O₂⁻ signal (Extended Data Fig. 9a and b) but was unable to induce an increase in root meristem size (Extended Data Fig . 9c and d). Furthermore, we detected only a subtle change in root meristem size in plt2 as compared to wild type upon RGF1 treatment (Extended Data Fig. 10a and b). However, an elevated O₂⁻ signal was observed (Extended Data Fig. 10c and d). These results strongly suggest that ROS signals modulated by RITF1 enhance PLT2 stability. In summary, we identified a novel transcription factor, RITF1, induced by RGF1 in the meristematic zone. This factor controls ROS levels, which in turn, regulate PLT2 protein stability and meristem size. Overall, our data demonstrate a key role for the peptide hormone, RGF1, in regulating root growth via modulation of ROS levels, which control the transition from proliferation to differentiation.

Methods:

Plant materials and growth conditions

All Arabidopsis mutants and marker lines used in this research are in the Columbia-0 (Col-0) background. The T-DNA insertion line of plt2 (SALK_130119.20.25) was obtained from the Arabidopsis Biological Resource Center at Ohio State University. The T-DNA insertion was identified at 166 bp upstream of the transcription start site in the plt2 mutant. Seeds were surface sterilized using 50 % (vol/vol) bleach and 0.1% Tween 20 (Sigma) for 15 min and then rinsed five times with sterile water. All seeds were plated on standard MS media (1× Murashige and Skoog salt mixture, Caisson Laboratories), 0.5 g/L MES, 1% Sucrose, and 1% Agar (Difco) and adjusted to pH 5.7 with KOH. All plated seeds were stratified at 4 °C for 2 d before germination. Seedlings were grown on vertically positioned square plates in a Percival incubator with 16 h of daily illumination at 22 °C.

The ritf1 mutants

The ritf1-1 mutant was generated using the egg cell-specific controlled CRISPR/Cas9 system²⁰.

SgRNA sequences are as followed:

RITF1-sgRNA1: GGGATGTCCATACCATGAGA CGG

RITF1-sgRNA2: CCGTCTACCACAGTTGATCG AGG

RITF1-sgRNA3: GGCGAACTTGAAGGAGTCTA TGG

RITF1-sgRNA4: GACTTTCAGTTGAGTCCTCA TGG

The CRISPR construct was transformed into Col using the Agrobacterium-mediated floral dip method. The mutant was identified by direct sequencing of PCR products of the targets in the offspring in T1, 2, and 3 generations. The loss-of-function ritf1-1 allele contains an insertion of a cytosine at 74 bp after the transcription start site in the RITF1 gene (771 bp). The additional insertion of a cytosine results in a frameshift and creates many premature stop codons after the insertion. To exclude issues related to off-target mutations, the sequences of three potential off-target genes (At5g25170, At1g70110, At3g20640) including similar sequences of the target sites were confirmed by direct sequencing of PCR products in the offspring in the T1, 2, and 3 generations. We didn’t find any mutations in these genes. Further, we identified another independent CRISPR allele (ritf1-3). This allele contains an insertion of an adenine at 75 bp after the transcription start site in the RITF1 gene. The additional insertion of an adenine results in a frameshift and creates many premature stop codons after the insertion. Similar to ritf1-1, seedlings from ritf1-3 exhibited strong resistance to RGF1 peptide and didn’t increase O₂⁻ levels as compared to wild type or to the weak allele (ritf1-2) (Extended Data Fig. 6d and e). These results can exclude the possibility that off-target mutations cause the RGF1-resistant phenotype. In the ritf1-2 allele (SALK_081503C), the T-DNA insertion was identified at 787 bp downstream of the transcription start site (in the middle of the second intron) of the RITF1 gene. Even though the insertion disrupted an intron, a full-length transcript was weakly detected in this allele.

Detecting gPLT2-YFP and ROS signals

The seedlings of wild type and the rgfr1/2/3 mutant were grown for seven days on MS agar plates, then transferred to MS agar plates containing either water (mock) or 20 nM synthetic sulfated RGF1 peptide (Invitrogen). After RGF1 treatment, seedlings were stained for 2 min in a solution of 200 μM NBT in 20 mM phosphate buffer (pH 6.1) in the dark and rinsed twice with distilled water. For detection of hydrogen peroxide with BES-H₂O₂-Ac²¹, seedlings were incubated in 50 μM BES-H₂O₂-Ac (WAKO) for 30 min in the dark, then mounted in 10 mg/mL propidium iodide (PI) in water². Roots were observed using a 20 X objective with a Zeiss LSM 880 laser scanning confocal microscope. Excitation and detection windows were set as follows: BES-H₂O₂-Ac, excitation at 488nm and detection at 500-550 nm; PI staining, excitation at 561 nm and detection at 570-650 nm. Confocal images were processed, stitched, and analyzed using the Fiji package of ImageJ²². Maximum projection images were generated from about 30 z-section images of BES-H₂O₂-Ac staining. The average intensity of BES-H₂O₂-Ac in the meristematic zone was measured in 5 or 6 roots with three biological replicates. Images for NBT staining were obtained using a 10 X objective with a Leica DM 5000-B light microscope. The total intensities of NBT staining in the meristematic zone were measured in 10 roots with three biological replicates using the Fiji software package²².

For shorter time course experiments, seedlings of gPLT2-YFP¹³ were grown on MS agar plates for 7 days, then transferred to MS agar plates containing either water (mock) or 100 nM RGF1 peptide. At 4, 6, 8, and 10 h after transfer to RGF1 plates, images were taken with a confocal or light microscope after PI, NBT, and BES-H₂O₂-Ac staining, as described above.

Total RNA preparation, RNA amplification and library preparation for RNA-seq

The HYP2-GFP¹⁰ line was grown on MS plates for 7 days. HYP2-GFP seedlings were then transferred into liquid MS media and treated with water (mock) or 100 nM RGF1 peptide in 6-well-plates for 1 h. After treatment with mock or RGF1, the seedlings were taken out of liquid MS media and transferred onto a 2 % agarose plate. Using an ophthalmic scalpel (Feather), the meristematic zone of the seedlings was precisely dissected based on HYP2-GFP fluorescence as detected under a dissecting microscope (Axio Zoom, Zeiss). Total RNA was extracted from 20 root sections treated with mock or 100 nM RGF1 using the RNeasy Micro Kit (Qiagen). For each treatment, three replicates of the RNA extractions were performed. All total RNA samples were treated with DNase I during RNA extraction. RNA quality was examined using a 2100 Bioanalyzer (Agilent). The RNA Integrity Number (RIN) was over 9.0 in all samples. The concentration of total RNA was measured by a Qubit (Invitrogen) instrument. For each replicate, 50 ng total RNA was amplified using the Ovation RNA-seq System V2 (NuGEN). Following amplification, 3 μg of cDNA was fragmented using the Covaris S-Series System. 400 ng of the fragmented cDNA with an average size of 400 bp was used for library preparation using the Ovation Ultralow System V2 (NuGEN). Illumina sequencing was performed at the Duke Genome Sequencing Shared Resource. The libraries for three biological replicates of mock and RGF1 treated meristematic zones were sequenced on an Illumina HiSeq 2000 (100 base paired end reads).

Differential expression analysis following RGF1 peptide treatment

Illumina sequencing reads were mapped to the TAIR10 Arabidopsis genome using Tophat V2.1.1. The parameters used for mapping were: “-N 5 --read-gap-length 5 --read-edit-dist 5 --b2-sensitive -r 100 --mate-std-dev 150 -p 5 -i 5 -I 15000 --min-segment-intron 5 --max-segment-intron 15000 --library-type fr-unstranded”. To select properly mapped reads with unique mapping positions, only alignments with flag of 83, 99, 147 or 163 and a mapping quality score of 50 were kept for further analysis. Mapping positions of these reads were compared with the Araport11 genome annotation (https://www.araport.org/downloads/Araport11_Release_201606/annotation) using HTseq-count (v0.6.1) with parameters “--stranded=no --mode=intersection-nonempty”, which generated read count per gene. The raw read counts of miRNA, lncRNA and protein coding genes were then used as input into DESeq2 (v1.14.1) for differential gene expression analysis. Genes with an FDR-adjusted p-value less than or equal to 0.1 were regarded as differentially expressed between RGF treatment and mock. The enriched gene ontology (GO) groups among differentially expressed genes were identified using agriGO. The GO annotation downloaded from http://geneontology.org was used as input for agriGO. Enriched GO groups required an FDR-adjusted p-value of 0.01 or less and a minimum mapping entry of 10.

Expression analysis of RITF1 upon RGF1 treatment by qRT-PCR

To perform qRT-PCR, about 20 meristematic zones of wild type and the rgfr1/2/3 mutant were dissected at 1, 6, and 24h after RGF1 treatment as described above. cDNA was generated from ten micrograms of total RNA using SuperScript IV Reverse Transcriptase (Invitrogen). Three biological replicates and technical replicates were used for each experiment. Standard curves were run for the primer pairs of RITF1 5’-CAAGCCATGCCACACTCTAA-3’ and 5’- TTATCCGAGGAAGCTGAGGA-3’, and PROTEIN PHOSPHATASE 2A SUBUNIT A3 (PP2AA3, AT1G13320) as reference 5’-GGCCAAAATGATGCAATCTC-3’ and 5’- TGCGAAATACCGAACATCAA-3’. Expression of RITF1 was assayed by qRT-PCR on a LightCycler 480 (Roche) with SYBR-based detection, normalized to PP2AA3, and analyzed by the efficiency corrected quantification (ECQ) model.

Plasmid Constructs

The coding sequence and the promoter sequence (2121 bp) of the RITF1 gene (AT2G12646) were amplified using the Phusion High-Fidelity DNA polymerase (New England Biolabs) from a wild-type cDNA library and genomic DNA, respectively, then sub-cloned into the pENTR/D/TOPO vector (Invitrogen). The following primers were used for RITF1 amplification: 5’-CACCATGGGAATTCAGAAACCGG-3’ and 5’- TTAACAGAGAGGAGATCGTTG-3’ and RITF1 promoter amplification 5’- CACCGCATCATTTTATTATAACCCGA-3’ and 5’-GAGGACTCAACTGAAAGTCA-3’. The sequence of the RITF1 gene and the RITF1 promoter in pENTR/D/TOPO vector was confirmed by Sanger sequencing. The clones were recombined into the pMDC7 and pMDC204 vectors¹² using LR clonase II (Invitrogen) to fuse the estradiol inducible promoter (XVE)²³ and the coding region of the RITF1 gene and the RITF1 promoter and GFP with a C-terminus HDEL retention sequence.

Measurement of meristem size and detection of ROS signals after overexpression of RITF1

The XVE-RITF1 construct was transformed into wild type (Col). To measure meristem size and detect ROS signals, two independent lines of XVE-RITF1 and wild type were grown on MS media for seven days, then transferred to MS media containing DMSO (Mock) and 10 μM β-estradiol (Sigma). After 24 h with mock or estradiol treatment, meristem size and ROS signals were measured and detected in wild type and the XVE-RITF1 lines, as described above.

Expression of pRITF1-GFP in roots

The pRITF1-GFP construct was introduced into wild type (Col) and rgfr1/2/3. Two independent T3 lines in each background were grown for 7 days in MS media were treated with mock and 20 nM RGF1 peptide. 24 h after treatment, GFP signals were detected using a confocal laser scanning microscope, as described above.

Note:

UPB1 is not required for the RGF1-receptor pathway

We have previously reported that UPBEAT1 (UPB1) reduces H₂O₂ levels and controls meristem size by down-regulating peroxidase genes in the elongation zone². Our transcriptome analysis didn’t find significant changes in UPB1 expression upon RGF1 treatment (Supplementary Table 1 and 3). We did find elevated expression of 5 peroxidase genes (Supplementary Table 1), but these are not targets of UPB1², suggesting that RGF1 regulates meristem size independently of UPB1. To determine if the peroxidase genes upregulated by RGF1 play a role in meristem size control in the RGF1-signaling pathway, we overexpressed two of them (At5g39580 and At4g08780). In neither case did we observe a larger meristematic zone (data not shown).

Statistics and Reproducibility

Experiments were independently repeated three times with similar results.

Data Availability Statement

All RNA-seq data in this study have been deposited in NCBI GEO, with the accession identifier GSE108730.

Code Availability Statement

All code in this study is available upon request.

Extended Data

Extended Data Fig. 1. Expression of meristematic zone marker and transcriptome analysis upon RGF1 treatment.

Confocal images of HPY2-GFP are shown 24 h after treatment with (a) water (mock) or (b) 20 nM RGF1. Seedlings were grown on MS medium for seven days before treatment. (a and b left) Propidium iodide stained roots; (a and b right) GFP signals. White and blue arrowheads indicate junction between meristematic and elongation zones. Scale bar = 50 μm. (c) Area (μm²) of HPY2-GFP expression is shown. (n = 8 independent roots, p < 2.5E-07). Bar graphs represent mean. Error bars are ± SD. Dots indicate each data point. P values calculated by two-sided Student’s t-test. (d) Schematic of RNA extraction following RGF1 treatment.

Extended Data Fig. 2. Enriched GO categories upon RGF1 treatment.

The highly significant Enriched Gene Ontology (GO) categories within gene lists regulated by RGF1 (False Discovery Rate (FDR)-adjusted p<0.001). “glutathione transferase activity” (FDR-adjusted p= 4.1E-06, red) and “oxidoreductase activity” (FDR-adjusted p=0.00039, red) were highly significantly enriched. See also Extended Data Tables (Enriched GO categories upon RGF1 treatment). The p-value of GO enrichment analysis was based on Fisher’s exact test, with the sample size of all genes in the genome and Benjamini–Yekutieli FDR for multiple testing correction.

Extended Data Fig. 3. H₂O₂ levels after inducible overexpression of RITF1 and RGF1 treatment.

(a) Confocal images of H₂O₂-BES-Ac stained roots with or without XVE-RITF1 in wild type 4 h after treatment with water (mock), 10 μM Estradiol, or 100 nM RGF1. (b) Quantification of H₂O₂-BES-Ac intensity in the meristematic zone (n = 6 independent samples, *p < 0.001). Bar graphs represent mean. Error bars are ± SD. Dots indicate each data point. P values calculated by two-sided Student’s t-test.

Extended Data Fig. 4. Localization of gPLT2-YFP, NBT, and H₂O₂-BES-Ac staining after RGF1 treatment.

Localization of gPLT2-YFP (a-h), NBT, (i-n) H₂O₂-BES-Ac (o-t), 4 h after treatment with water (mock) (a) or 100 nM RGF1 (b), 6 h after treatment with water (mock) (c, i, and o) or 100 nM RGF1 (d, j, and p), 8 h after treatment with mock (e, k, and q), or 100 nM RGF1 (f, l, and r). 10 h after treatment with mock (g, m, and s) or 100 nM RGF1 (h, n, and t). Blue arrowheads indicate junction between the meristematic and elongation zones. Scale bar = 50 μm. Seedlings were grown on MS agar plates for seven days before treatment. Experiments were independently repeated three times with similar results.

Extended Data Fig. 5. Time course of gPLT2-YFP localization and NBT and H₂O₂-BES-Ac staining.

(a) Distance (μm) of gPLT2-YFP localization. (n = 5 independent roots, p < 5.7E-06) (b) Total intensity (arbitrary unit, A. U.) of NBT staining in the meristematic zone. (n = 9 independent roots, p < 0.001) (c) Average intensity of H₂O₂-BES-Ac staining in the elongation zone. (n = 5 independent roots, p < 0.003). Bar graphs represent mean. Error bars are ± SD. Dots indicate each data point. P values calculated by two-sided Student’s t-test. Experiments were independently repeated three times with similar results.

Extended Data Fig. 6. Phenotype of ritf1.

(a) Root growth (mm) of wild type (Col), ritf1-1 (CRISPR mutant) and ritf1-2 (SALK line) from 4 to 8 days after germination. (n = 21 independent roots) (b) Confocal images of roots of Col, ritf1-1 (CRISPR mutant), and ritf1-2 (Salk line) stained with PI. (c) Percent increase (100% = number of cells in mock treatment) in the number of cells in the meristematic zone of Col, ritf1-1, and ritf1-2 24 h after mock or 5 nM RGF1 treatment. (n = 7 independent roots, *p < 6.0E-06). (d) Light microscope images of roots of Col, ritf1-3, and ritf1-2 stained with NBT 24 h after 5 nM RGF1 treatment. Scale bar = 50 μm. Blue arrowheads show junction between the meristematic and elongation zones. (e) Quantification of NBT staining intensity (A.U.) in the meristematic zone in wild type, ritf1-3, and ritf1-2 after 5 nM RGF1 treatment. (n = 8 independent roots, *p< 0.003). All scale bars = 50 μm. Blue, and white arrowheads indicate the junction between the meristematic and elongation zones. Bar and line graphs represent mean. Error bars are ± SD. Dots indicate each data point. P values calculated by two-sided Student’s t-test.

Extended Data Fig. 7. Expression of pPLT2-CFP and gPLT2-YFP upon RGF1 treatment.

Confocal images of (a) pPLT2-CFP and (b) gPLT2-YFP 24h after 20 nM RGF1 treatment. PI staining (right images, a and b). Scale bar = 50 μm. Arrow heads show junction between the meristematic and elongation zones. (c). Extent (μm) of expression of gPLT2-YFP from QC cells. (n = 5 independent roots, p < 2.5E-13). Bar graphs represent mean. Error bars are ± SD. Dots indicate each data point. P values calculated by two-sided Student’s t-test.

Extended Data Fig. 8. Localization of PLT2 protein after RGF1 and H₂O₂ treatment.

(a-d) Confocal images of gPLT2-YFP 24h after treatment with mock, 20 nM RGF1, 20 nM RGF1 with 500 μM H₂O₂, and 500 μM H₂O₂. Seedlings of gPLT2-YFP were grown for seven days on MS agar plates before treatments. Scale bar = 50 μm. Arrow heads show the extent of gPLT2-YFP expression. (e) Extent (μm) of PLT2 protein localization as measured from the QC cells. (n = 6, *p < 0.0002). Bar graphs represent mean. Error bars are ± SD. Dots indicate each data point. P values calculated by two-sided Student’s t-test.

Extended Data Fig. 9. Phenotypes of overexpression of RITF1 in plt2.

(a) NBT stained roots with or without XVE-RITF1 in Col and plt2 24 h after treatment with mock or 10 μM Estradiol. (b) Quantification of NBT staining intensity in the differentiation zone with or without XVE-RITF1 in Col and plt2. (n = 8 independent roots, *p < 5.5E-06). (c) Confocal images of roots with or without XVE-RITF1 in Col and plt2 24 h after mock or 10 μM estradiol treatment stained with PI. (d) The number of cells in the meristematic zone with or without XVE-RITF1 in Col and plt2 24 h after mock or 10 μM estradiol treatment. (n = 7 independent roots, *p < 4.5E-05). All scale bars = 50 μm. White and blue arrowheads indicate junction between the meristematic and elongation zones. Bar graphs represent mean. Error bars are ± SD. Dots indicate each data point. P values calculated by two-sided Student’s t-test.

Extended Data Fig. 10. Phenotype of plt2 upon RGF1 treatment.

(a) Confocal images of PI stained roots in wild type (Col) and plt2 24 h after 20 nM RGF1 treatment. Scale bar = 50 μm. White arrow heads show junction between the meristematic and elongation zones. (b) Number of cells in the meristematic zone 24 h after 5 nM RGF1 treatment. (n = 8 independent roots, *p < 5E-07). (c) Light microscope images of roots of Col and plt2 stained with NBT. Seedlings were grown on MS agar plates for seven days before mock or 20 nM RGF1 treatment. (d) Total intensity of NBT staining in the differentiation zone of Col and plt2 24 h after mock or 20 nM RGF1 treatment. (n = 10 independent roots, *p < 0.001). All scale bars = 50 μm. White arrowheads show junction between the meristematic and elongation zones. Bar graphs represent mean. Error bars are ± SD. Dots indicate each data point. P values calculated by two-sided Student’s t-test.

Supplementary Material

Sup_Tab1

Supplementary Table1. Differential expression upon RGF1 treatment Expression of 583 genes is significantly altered in the RNA-seq data between RGF1 and mock treatment in three biological replicates (FDR-adjusted p value < 0.1). The differentially expressed genes were identified using the statistical library DEseq2, which models the read counts with a negative binomial distribution. The p-value is adjusted for multiple testing using Benjamini-Hochberg FDR procedure. The 3 biological replicates are typical for RNA-seq study, with both statistical and practical considerations.

Sup_Tab2

Supplementary Table2. Enriched GO categories upon RGF1 treatment List of significant Gene Ontology (GO) categories (FDR≤ 1) and AGI numbers in the GO categories. The p-value of GO enrichment analysis was based on Fisher’s exact test, with the sample size of all genes in the genome and Benjamini–Yekutieli FDR for multiple testing correction.

Sup_Tab3

Supplementary Table3. All gene expression upon RGF1 treatment Expression of all annotated Arabidopsis genes in mock or RGF1 treatment (Value is FPKM).