Membrane-associated NRPM proteins are novel suppressors of stomatal production in Arabidopsis

SUMMARY

In Arabidopsis, stomatal development and patterning require tightly regulated cell division and cell-fate differentiation that are controlled by key transcription factors and signaling molecules. To identify new regulators of stomatal development, we assay the transcriptomes of plants bearing enriched stomatal lineage cells that undergo active division. A member of the Novel Regulators at the Plasma Membrane (NRPM) family annotated as hydroxyproline-rich glycoproteins was identified to highly express in stomatal lineage cells. Overexpressing each of the four NRPM genes suppressed stomata formation, whilst the loss-of-function nrpm triple mutants generated severely overproduced stomata and abnormal patterning, mirroring those of the erecta receptor family and MAPKKK yoda null mutants. Manipulation of the subcellular localization of NRPM1 surprisingly revealed its regulatory roles as a peripheral membrane protein instead of a predicted cell-wall protein. Further functional characterization suggests that NRPMs function downstream of the EPF1/2 peptide ligands, and upstream of the YODA MAPK pathway. Genetic and cell biological analyses reveal that NRPM may promote the localization and function of the ERECTA receptor proteins at the cell surface. Therefore, we identify NRPM as a new class of signaling molecules at the plasma membrane to regulate many aspects of plant growth and development.

eTOC Blurb

External signaling pathways heavily influence plant development in Arabidopsis. Xue et al. identify the NRPM proteins as novel signaling molecules at the plasma membrane that suppress stomatal production. This underscores an unrecognized regulatory module between the membrane receptors and the MAPK module.

Graphical Abstract

INTRODUCTION

Stomata are microscopic pores in the epidermis of the aerial organs of a plant. Each stomatal pore is surrounded by a pair of guard cells that, through turgor-driven cell expansion, regulate gas exchange and water transpiration between plants and the environment. Stomatal development and patterning in the dicot model plant Arabidopsis have been well studied and provide an excellent system for studying cell polarity, cell division, cell-fate transition, and cell-fate differentiation.^1,2

The progression of stomatal differentiation in Arabidopsis requires successive activities of three basic helix-loop-helix (bHLH) transcription factors, SPEECHLESS (SPCH), MUTE and FAMA.^3,4 Another two bHLH transcription factors, SCREAM (SCRM) and SCRM2, redundantly form heterodimers with SPCH, MUTE, and FAMA, respectively, to drive the three successive steps of cell-fate transition in stomatal development.⁵ The mitogen-activated protein kinase (MAPK) cascades are critical signaling pathways that regulate diverse cellular processes by transmitting extracellular stimuli to intracellular machinery.^6–8 Stomatal development and patterning in Arabidopsis are also controlled by a canonical MAPK cascade, including the MAPK kinase kinase YODA (YDA), MAPK kinase 4 and 5 (MKK4/5), and MAPK 3 and 6 (MPK3/6).^9–11 Biochemical and genetic analyses certified a crucial role of this YDA-MKK4/5-MPK3/6 cascade in suppressing stomatal production through phosphorylation-mediated degradation of SPCH and SCRM.^4,12 In addition, recent studies revealed that SCRM as a scaffold brings SPCH approximately to MPK6, improving the efficiency of phosphorylation and subsequent degradation of SPCH.¹³

Plants have evolved the use of the secreted Epidermal Patterning Factor (EPF) peptides to prevent neighboring cells from entering the stomatal fate. The EPF peptide ligands are perceived by the cell surface leucine-rich repeat receptor-like proteins (LRR-RLPs) and kinases (RLKs), including TOO MANY MOUTHS (TMM), the ERECTA family (ER-f) and the Somatic Embryogenesis Receptor Kinase (SERK) family.^14–18 Activated RLP/RLK signaling triggers the downstream YDA-MKK4/5-MPK3/6 signaling cascade, which then regulates the substrate levels or activities, such as SPCH.^9–11 However, how the LRR-RLP/RLK complexes transmit the signal to the MAPK cascade is still unclear. In this study, based on the transcriptome profiling data, we identified a group of proteins annotated as hydroxyproline-rich glycoproteins (HRGPs) with the founding member highly expressed in the dividing stomatal lineage cells. Mutating three of the four closely related HRGP genes led to overly proliferative stomatal lineage cells, as well as broad defects in growth and development, to some extent resembling those of the yda and erecta family mutants.^9,19 In plants, HRPGs are known as a superfamily of plant cell wall proteins that function widely in plant growth and development.^20,21 However, how they function remains almost unknown. Here, we found that the Novel Regulators at Plasma Membrane (NRPM) family of these putative HRGPs function inside of the plasma membrane and negatively regulate stomatal production downstream of the peptide ligands EPF1/2, upstream of YDA, and partially through the ERECTA receptor.

RESULTS

RNA-seq-based identification of new regulators in stomatal ACD

In Arabidopsis, the bHLH transcription factor SPCH determines the initiation of stomatal lineage cells.²² Overexpression of a truncated SPCH variant, SPCHΔ49 that is insensitive to MAPK-triggered protein phosphorylation for degradation, produces excessive stomatal lineage cells in the leaf epidermis.⁴ Stomatal asymmetric cell division (ACD) is controlled by the scaffold protein BREAKING OF ASYMMETRY IN THE STOMATAL LINEAGE (BASL) that polarizes at the cortical region of the precursor cells to drive division orientation and specify distinct daughter-cell fates.^23–25 In the absence of BASL, the stomatal lineage divisions largely lost their physical and fate asymmetries.²³ To identify new regulators in stomatal ACD, we conducted a transcriptomic analysis using young plants overexpressing SPCHΔ49 (driven by the ubiquitous 35S promoter) that produce excessive stomatal lineage cells and overexpressing SPCHΔ49 in basl-2 mutants that produce a comparative amount of stomatal lineage cells but less asymmetric in divisions (Figure 1A). We expected that stomatal ACD regulators would show high expression levels in SPCHΔ49 but reduced levels in SPCHΔ49;basl-2 and wild-type (WT) plants.

Figure 1. Identification of genes differentially expressed in actively dividing stomatal lineage cells.

(A) Confocal images show epidermal cell patterns on cotyledons at 3 days post-germination (dpg) in wild type (WT), SPCHΔ49, and SPCHΔ49;basl-2. SPCHΔ49, overexpression of a phospho-deficient SPCH variant in wild type that induces overproduction of stomatal lineage cells. In SPCHΔ49;basl-2 mutants, these divisions are less asymmetric. Cell outlines were stained with propidium iodide (PI, magenta). Scale bar, 50 μm.

(B) Differentially expressed genes (DEG) in WT, SPCHΔ49, and SPCHΔ49;basl-2. Total RNAs were isolated from 12-dpg seedlings of indicated genotypes for transcriptomic analysis. Two biological replicates for WT and three biological replicates for SPCHΔ49 and SPCHΔ49;basl-2, respectively, were used for this experiment.

(C) Venn diagrams show DEGs down (left) and up (right) regulated when the two genotypes were compared as indicated.

(D) GO-term enrichments of up-regulated DEGs in SPCHΔ49 vs. SPCHΔ49;basl-2.

(E) Relative transcript levels of the 20 up-regulated DEGs at the intersection in (C) (SPCH excluded). The transcript levels of each gene in WT were normalized to 1.

Values are mean ± SEM (n = 2 or 3).

See also Figure S1 and Data S1.

To perform genome-wide RNA-seq analyses, total RNAs were extracted and purified from seedlings of WT, SPCHΔ49 and SPCHΔ49;basl-2 at 12 days post-germination (dpg). The heatmap displaying differentially expressed genes (DEGs) clustered with similar expression patterns suggested that the plants overexpressing SPCHΔ49 are most distinct from the WT plants and moderately different from SPCHΔ49;basl-2 (Figure 1B). Through pairwise comparison, a total of 1949 DEGs (TPM > 10, |b| ≥ 0.692, q < 0.05) were identified, and up-/down-regulated DEGs were plotted by Venn diagrams (Figure 1C; gene lists in Data S1). Genes known to regulate stomatal development, such as SPCH, MUTE, SCRM, and other regulators involved in signal transduction (the peptide ligands, membrane receptors, and cytoplasmic kinases), were highly enriched in the SPCHΔ49 overexpression plants (Figure S1A). As we were interested in the genes differentially expressed between SPCHΔ49 and SPCHΔ49;basl-2, the 172 up-regulated DEGs in SPCHΔ49 were subjected to the gene ontology (GO) analysis. We found that these genes are mostly enriched in the processes of cell cycle, stomatal complex development, cell division, cell wall organization, mitotic apparatus, asymmetric cell division, etc. (Figure 1D; Data S1). We selected candidate genes possibly involved in membrane dynamics, signal transduction, and cytoskeletal organization for promoter activity tests. Indeed, we found many promoters, such as that of AT4G26660 (encodes a kinesin-like protein) and AT1G53140 (encodes a dynamin-related protein), showed strong and specific activities in the stomatal lineage cells (Figures S1B and S1C). Furthermore, twenty intersection DEGs were commonly identified to show higher transcript levels among the three pairs of comparisons (Figure 1C). Among them, besides SPCH itself, genes encoding transcriptional factors WRKY50, NAC035, and NAC041, as well as known stomatal genes such as TMM, ETC3 (ENHANCER OF TRY AND CPC 3), and BHP (BLUE LIGHT-DEPENDENT H+-ATPASE PHOSPHORYLATION) were identified. The eight genes with unknown functions (labeled with AGI) were also isolated (Figure 1E). Particularly, AT4G25620 was found to be highly expressed in the actively dividing meristemoids in the scrm-D;mute mutants (Figure S1D).²⁶

We further screened the candidate genes by creating overexpression plants and/or the loss-of-function mutants (T-DNA insertional or CRISPR/Cas9-mediated mutagenesis). Among the tested candidates, AT4G25620 became the top priority because of the strong phenotypes in stomatal development associated with altered gene activity in Arabidopsis (see below).

NRPMs are expressed in the stomatal lineage cells

AT4G25620 encodes a hydroxyproline-rich glycoprotein (HRGP) with unknown functions and is active exclusively in the stomatal lineage cells (Figure 2A). AT4G25620 belongs to a four-member family (Figure 2B). We first examined the protein subcellular localization by transient expression in the Nicotiana benthamiana epidermal cells. Surprisingly, all four proteins are distributed along the plasma membrane (Figure S1E). This observation was in striking contrast to the original annotation that the HRGPs are usually located in the cell wall.²⁷ As their location was later determined to the PM (detailed below), AT4G25620 and the other family members were accordingly named as Novel Regulators at the Plasma Membrane (NRPM) from here on (Figure 2B). To characterize their functions, we first assayed the promoter activity of each NRPM gene. We found that, except for NRPM3 (AT1G63720), which appeared to be expressed in the mesophyll cells, all the other three NRPM promoters show activities in the epidermal cells (Figure S1F). Furthermore, compared to the broad expression pattern of NRPM2 (AT5G52430) in the leaf epidermis, NRPM1 and NRPM4 (AT1G76660) were more restricted to the stomatal lineage (Figure S1F). The endogenous protein localization of NRPMs was further assayed in Arabidopsis plants by expressing NRPM-YFP driven by the respective native promoter. Consistent with the transient protein-expression data, all three NRPM proteins were localized to the plasma membrane in Arabidopsis plants (Figure 2C). As NRPM3 is absent from the leaf epidermal cells (Figure 2C), we examined whether and how NRPM1/2/4 proteins are differentially expressed in the stomatal lineage cells. Interestingly, NRPM1-YFP is specifically expressed in the early stomatal lineage cells, i.e., the meristemoid mother cells (MMCs) and meristemoids (Figure 2D), contrastingly NRPM4-YFP is preferentially expressed in the late staged cells, such as guard mother cells (GMCs) and young guard cells (GC) (Figure 2D). On the other hand, NRPM2-YFP is expressed in almost all epidermal cells, except for the mature GCs (Figures 2C and 2D). Thus, despite their preferential cell-type expression patterns, all three genes, NRPM1, NRPM2, and NRPM4, could contribute to regulating stomatal development.

Figure 2. NRPMs are mainly expressed in the leaf epidermis.

(A) Activity of the NRPM1 promoter (AT4G25620) is detected high in stomatal lineage cells. Cell outlines were stained with PI (magenta). Scale bar, 20 μm.

(B) The phylogenetic tree of the Arabidopsis NRPM family. A Neighbour-joining tree with distance shows the evolutionary relationships between NRPM members. RGA (AT2G01570) serves as the outgroup to anchor the tree.

(C) Expression patterns of the four NRPM genes in 3-dpg cotyledons. NRPM-YFP (green) was driven by the respective native promoter. NRPM1 is active in meristemoid mother cells (MMC) and meristemoid; NRPM2 accumulates in all epidermal cells except for mature guard cells; no NRPM3 is expressed in epidermal cells; NRPM4 is present in all epidermal cells. Scale bars, 20 μm.

(D) Differential expression patterns of NRPM1/2/4-YFP protein (green) during stomatal development. Representative images for each stage were selected based on cell morphology. NRPM1 is active before guard mother cells (GMC). NRPM2 shows low to no expression in differentiating guard cells. NRPM4 shows elevated expression levels in GMC and young guard cells. Cell outlines were stained with PI. Scale bars, 20 μm. (E) Plasmolysis assays show NRPM1/2/4-YFP proteins (green) are associated with the plasma membrane that is detached from the cell wall after the seedlings were treated with 20% sucrose. Control: GFP-PIP2A (Plasma membrane Intrinsic Protein 2A), an integral membrane protein (green) Scale bars, 5 μm.

See also Figure S1.

NRPMs suppress stomatal production at the plasma membrane

Since no transmembrane domains were predicted by DeepTMHMM²⁸ and the proteins were annotated as HRGPs, the localization of NRPMs at the plasma membrane was a surprise. To confirm that NRPMs indeed associate with the plasma membrane, we conducted the plasmolysis assay, in which plants were treated with 20% (w/v) sucrose for 20 minutes. As shown in Figure 2E, the NRPM1/2/4-YFP signals retracted with the plasma membrane upon plasmolysis, but no significant signals were detected in the cell walls, suggesting that NRPMs are indeed peripheral membrane proteins.

To study their functions, we overexpressed each of the four NRPM genes under the control of the ubiquitous CaMV 35S promoter in WT Arabidopsis plants (Figure S2A). Interestingly, once overexpressed, all four NRPM members can suppress stomata development (Figures 3A–3C), as suggested by the Stomatal Index (% of stomatal number over total number of epidermal cells). In addition, overexpression of these NRPM genes caused general defects in plant growth and development, such as stunted stature and spiral rosette leaves (Figure 3A). To further test whether NRPMs function from outside or inside of the plasma membrane, we engineered the wild-type NRPM1 protein by adding a myristoylation (myr) lipid modification sequence²⁹ that helps to tether the target protein to the plasma membrane or adding the signal peptide (sp) of an Extensin protein³⁰ that directs the target protein to the cell wall (Figures 3D and sequence in S2B). As anticipated, myr-NRPM1 was found exclusively along the PM, whereas sp-NRPM1 was distributed to the endomembrane compartments and the cell walls (Figure 3D). Interestingly, myr-NRPM1-YFP maintained the ability to suppress stomata formation, whereas sp-NRPM1-YFP failed to induce this phenotype (Figure 3E), strongly suggesting that NRPMs may function at the inner side of the plasma membrane to regulate stomatal development.

Figure 3. Overexpression of NRPM suppresses stomata production at the membrane periphery.

(A) Representative images show that overexpression of NRPMs reduces stomatal production and disturbs plant growth and development. Upper panel: 3-week-old plants under normal growth conditions. Scale bars, 1 cm. Lower panel: stomatal development and patterning in the adaxial side of 3-dpg cotyledons. Stomatal guard cells and dividing cells are colored in cyan and pink, respectively. Two independent lines for each transgene were established, and expression levels are shown in Figure S2A. WT: Columbia-0 plants expressing ML1::YFP-RCI2A (the transmembrane protein RCI2A driven by the epidermal specific ML1 promoter). Scale bars, 20 μm.

(B) Quantitative PCR analyses show transcript levels of each NRPM in the overexpression plants shown in (A). The transcript levels of each NRPM in WT were normalized to 1. Values are mean ± SEM (n = 3).

(C) Quantification of stomatal index in 5-dpg cotyledons of indicated plants. Student’s unpaired t tests were used to calculate the two-tailed P value when compared with the wild type.

(D) Confocal images show the localization of two engineered NRPM1-YFP proteins (green). The addition of a myristoylation (myr) peptide sequence targets NRPM1 to the inner leaflet of the plasma membrane, whereas the addition of an Extensin signal peptide (sp) drives NRPM1 to the secretory pathway. Arrows mark the trans-Golgi network-like structures as anticipated. Scale bars, 20 μm.

(E) Quantification of stomatal index in 5-dpg cotyledons of the transgenic plants shown in (D). The myr-NRPM1-YFP suppresses stomata formation, whereas the apoplastic sp-NRPM1-YFP fails to do so. Student’s unpaired t tests were used to calculate two-tailed P values. Boxes in (C and E) show first and third quartiles with mean (square) and median (line). Whiskers extend to 1.5 × interquartile range (IQR) from the first and third quartile. n, number of cotyledons analyzed.

See also Figure S2.

Loss-of-function nrpm mutants show stomatal overproduction

To further characterize the NRPMs’ function in stomatal development, we collected individual NRPM T-DNA insertional mutants (Figure S3A). No visible phenotypes were observed in the single mutants under optimal growth conditions (Figure S3B). As the expression of NRPM3 was not detected in the epidermal cells, we generated double mutants among nrpm1, nrpm2, and nrpm4, in which nrpm1 and nrpm4 appeared to be null and nrpm2 was a knock-down mutant (Figure S3C). The double mutants, nrpm1;2 and nrpm2;4, were first isolated, but we did not detect observable stomatal or other growth defects (Figure S3D), except that nrpm2;4 mutants produce shorter siliques with fewer seeds that can be complemented by expressing NRPM2-YFP driven by the native promoter (Figure S3E). However, we failed to obtain a nrpm1;4 double mutant, although NRPM1 and NRPM4 are located at two different chromosomes. In fact, we found that any individuals with the genotype of nrpm1−/−;nrpm4−/+ or nrpm1−/+;nrpm4−/− were absent from the F2 population of nrpm1−/− crossed with nrpm4−/− (Table S1). Therefore, we speculated that missing more than three copies from these two NRPM genes (1 and 4) would cause embryonic lethality, which is further supported by seed abortion phenotype observed in the siliques of a nrpm1−/+;nrpm4−/+ plant (Figure S3F). Also, the promoter activities of both NRPM1 and NRPM4 were detected in developing seeds, i.e., the endosperm region around the micropylar (arrows in Figure S3G). Therefore, strong T-DNA insertional alleles of nrpm mutations, in particularly the combined effect of nrpm1 and nrpm4, could lead to detrimental failures in embryo development, which greatly interfered with our ability to assay their functional contribution in post-embryonic processes, including stomatal development.

To study the function of NRPM during stomatal development, we took advantage of the CRISPR/Cas9-mediated genome editing that allows the creation of weak alleles, such as short deletions or replacements. We assembled four guide RNAs in one construct with each guide RNA targeting one of the four NRPM genes in the exon region (Figure S3A) and transformed it into WT plants. Interestingly, we identified three categories of plants segregating from one mother plant that show 1) WT-looking, 2) mildly disturbed (crispr_nrpm-1–1), and 3) severely disturbed growth (crispr_nrpm-1–2), respectively (Figure 4A). Furthermore, the severity of overall growth phenotypes was mirrored by stomatal defects in these mutants (Figure 4B). Genotyping data demonstrated that we successfully generated mutations in NRPM1/2/4 but not in NRPM3 (summarized in Figures 4B and detailed in S4A). Specifically, both crispr_nrpm-1–1 (moderate) and crispr_nrpm-1–2 (strong) carry homozygous early terminations in both NRPM1 and NRPM2. Mutating NRPM4 appeared to be most influential to the phenotypes because the moderate mutant crispr_nrpm-1–1 harbors an in-frame replacement, resulting in 13 amino acids (AAs) replacement of the original 3 AAs, whereas the strong mutant crispr_nrpm-1–2 contains biallelic early terminations in NRPM4 (Figures 4B and S4A). Thus, results from our CRISPR/Cas9-mediated mutagenesis revealed the redundant role of NRPM1, 2, and 4 in suppressing stomatal production and enforcing stomatal patterning.

Figure 4. Mutations in NRPMs results in severely overproduced stomata.

(A) Mutating NRPM1/2/4 interferes with plant growth and development. Mutant plants (crispr_nrpm-1–1 moderate and crispr_nrpm-1–2 strong) were identified in the progenies of a single T1 transformant carrying CRISPR/Cas9-mediated mutagenesis of NRPM1/2/4. WT: Col-0 plants expressing ML1::mCherry-RCI2A (21-dpg). The crispr_nrpm plants were created in the same background. Scale bar, 1 cm.

(B) Stomatal phenotypes (upper panel) and genotypes (lower panel) of crispr_nrpm-1–1 and crispr_nrpm-1–2 mutants. Specific mutations are provided in Figure S5. Stomata guard cells and small dividing cells are colored in cyan and pink, respectively. Scale bars, 25 μm.

(C) Expression of CRISPR-resistant NRPM1/2/4 (rNRPM)-YFP transgene (green) largely rescued stomata defects in crispr_nrpm-1–1 plants. rNRPM-YFP carrying CRISPR-off-target synonymous mutations driven by the native promoter was introduced into crispr_nrpm-1–1. Confocal images with cell outlines marked by mCherry-RCI2A (magenta) were captured from 3-dpg cotyledons. Scale bars, 20 μm.

(D) Recovery of plant growth defects in crispr_nrpm-1–1 was observed by expressing the native promoter-driven rNRPM1/2/4. Representative images show 16-dpg plants. Scale bars, 1 cm.

(E) Quantification of stomatal index in the adaxial side of 5-dpg cotyledons of crispr_nrpm-1–1 and rNRPM1/2/4 complementation lines shown in (D). Boxes show the first and third quartile with mean (square) and median (line). Whiskers extend to 1.5 × IQR from first and third quartile. Letters (a, b, and c) indicate significantly different means at P < 0.05 (one-way ANOVA). n, number of cotyledons analyzed.

See also Figure S3, S4, Table S1.

Next, we performed complementation assays to confirm the functional relevance of NRPMs in plant development. Since we were unable to obtain CRISPR/Cas9-free individuals from the crispr_nrpm-1 population, we expressed the NRPM variants (rNRPMs) that carry synonymous mutations (Figure S4B) that are resistant to the CRISPR/Cas9-mediated mutagenesis in the crispr_nrpm-1–1 mutants. As expected, the rNRPM1/2/4-YFP protein fusions driven by their own promoter recapitulated the native protein localization at the PM (Figure 4C). Importantly, all three constructs efficiently complemented the crispr_nrpm-1–1 mutant phenotypes in stomatal development and overall growth (Figures 4D and 4E). Thus, NRPM1, 2, and 4 redundantly regulate many aspects of plant growth and development.

Functional connection between NRPM and ER-f in stomatal development

Among the T2 population of CRISPR/Cas9-generated nrpm mutants, we observed varying developmental defects reminiscent of sequentially knocking out the ER-f members (Figures 5A and S5A). Moreover, the phenotypic severity of nrpm mutants appeared to be associated with the mutation features of nrpm4, whilst nrpm1 and 2 were all early terminations (Figure S4A). For example, the weak mutant crispr_nrpm-4–2 showing phenotypes at the er105 level contains nrpm4−/+ (early termination, heterozygous), and the moderate mutant crispr_nrpm-1–3 showing phenotypes at the er105;erl2–1 level contains nrpm4−/− (missing 3-bp, homozygous). Furthermore, the strong mutant crispr_nrpm-4–1 containing nrpm4−/− (early termination, homozygous) is highly similar to the er105;erl1–2;erl2–1 triple mutant, both of which exhibit severe defects in both vegetative and reproductive growth (Figures 5A and S5A).

Figure 5. NRPMs promote the plasma-membrane localization of the ERECTA receptor.

(A) Developmental defects observed in the crispr-nrpm mutants mirror those identified in the erecta family mutants (24-dpg). Specific NRPM mutations in different crispr_nrpm alleles are provided in Figure S5. Scale bar, 1 cm.

(B) Stomatal phenotypes (5-dpg) generated by elevated expression of ERECTA and NRPM3 in the wild-type plants. Note, enhanced phenotype was observed in plants co-expressing ERECTA::ERECTA-YFP and 35S::NRPM3-mRFP. Stomatal guard cells and lineage cells are colored in cyan and pink, respectively. Scale bars, 50 μm. Histograms on the right show quantification of stomatal index in the adaxial side of 5-dpg cotyledons. Boxes show first and third quartile with mean (square) and median (line). Whiskers extend to 1.5 × IQR from the first and third quartile. Letters (a, b, and c) indicate significantly different means at P < 0.05 (one-way ANOVA). n, number of cotyledons analyzed.

(C) Confocal images show enhanced and lowered internalization of ERECTA-YFP (green) in crispr_nrpm mutants and NRPM1 overexpression plants (3-dpg), respectively. Yellow arrowheads: internalized signals of ERECTA-YFP. Note the difference in different genetic backgrounds. Cell outlines (magenta) were highlighted with mCherry-tagged PIP2A, RCI2A, or NRPM1 (driven by 35S), as indicated. Intensity profilings of YFP and mCherry were obtained along the lines drawn on the left, and internalized ERECTA-YFP signals are indicated with black arrows on the profiling. Scale bars, 20 μm.

(D-F) Co-expression of ERECTA-YFP (green) with mCherry-tagged endomembrane markers (magenta), including SYP61 (D, trans-Golgi network/early endosome, TGN/EE), RabG3f (E, prevacuolar compartments and tonoplast, PVC/V), and γ-TIP (F, vacuolar tonoplast, V). Co-localizations were indicated by white arrows. Scale bar, 10 μm.

(G) Representative confocal images show seedlings expressing ERECTA-YFP treated with 70 μM BFA for 1 hr. White arrows mark the association of ERECTA-YFP with BFA-bodies indicated by FM4–64. Histograms on the right show that more ERECTA-containing BFA bodies formed in crispr_nrpm-1–3 mutants. Boxes show first and third quartile with mean (line). Whiskers extend to 1 × SD from the first and third quartile. Student’s unpaired t tests were used to calculate two-tailed P values. n, number of cotyledons analyzed. Scale bars, 10 μm.

See also Figure S5, S6.

The phenotypic resemblance between nrpm and er-f mutants inspired us to evaluate if they are functionally connected. We generated transgenic plants expressing ERECTA-YFP driven by the native promoter in wild-type plants. We found that expressing an extra amount of ERECTA-YFP suppressed stomatal production in plants (Figure 5B), similar to the effect of NRPM overexpression (Figure 3A). When ERECTA::ERECTA-YFP was crossed with 35S::NRPM3-mRFP, we found that the stomatal phenotype was additive or synergistic, resulting in almost no stomata formation (Figure 5B). Thus, the NRPM proteins may facilitate the functions of ER-f receptor kinases at the plasma membrane and/or they function independently but converge on the same downstream regulator/s to suppress stomatal production.

Abnormal localization of ERECTA in nrpm mutants

It was recently reported that the subcellular dynamics of ER-f receptors are associated with their functions in specifying stomatal cell fate.^31,32 To test whether functions of NRPM are linked to the localization of ER-f receptor proteins, we examined ERECTA-YFP in crispr_nrpm-1–3 mutants that show moderate phenotypes. As previously reported, ERECTA-YFP in er105 ³¹ is predominantly localized at the plasma membrane with some internalization in leaf epidermal cells (Figure 5C, z-stacked images in Figure S5B). However, this typical localization pattern of ERECTA-YFP was altered when NRPMs’ activities were genetically modified (Figure 5C). ERECTA-YFP became more internalized in crispr_nrpm-1–3 mutants (middle row, Figure 5C and Figure S5B) but less so in NRPM1 overexpression plants (bottom row, Figure 5C), consistently supporting a positive role of NRPM in maintaining ERECTA-YFP at the plasma membrane. However, this role does not seem be widely applicable to other membrane receptors because the co-receptors of ERECTA, including ERL1 and SERK3, didn’t changes in the localization in crispr_nrpm-1–3 plants (Figures S5C and S5D).

To clarify the features of the ERECTA-YFP decorated membrane structures, we performed co-localization assay between ERECTA-YFP and mCherry-tagged endomembrane organelle markers, including ER-rk (CD3–959) for endoplasmic reticulum (ER), ³³ Man49-mCherry (CD3–967) for Golgi, ³³ mCherry-SYP61 for trans-Golgi network/early endosome (TGN/EE), ³⁴ RabG3f-mCherry for prevacuolar compartments (PVCs) and tonoplast, ³⁵ and γ-TIP-mCherry (CD3–975) for tonoplast³³ (Figures 5D–F, S6A and S6B). We found that ERECTA-YFP partially co-localized with the TGN/EE (Figure 5D), and largely co-localized with the markers of the plasma membrane, PVC, and tonoplast (Figures 5E and 5F). These observations were consistent with the recent discoveries suggesting that the ERECTA signaling is attenuated by U-box ubiquitin E3 ligases and ERL1 is internalized for vacuolar degradation. ^31,36 To evaluate whether NRPMs participate in the vacuolar degradation pathway to regulate ERECTA, we treated plants with Wortmannin, which disturbs the endosomal maturation to form the vacuole in plant cells. ³⁷ Indeed, we observed more prevacuolar fusions of ERECTA-YFP upon the Wortmannin treatment in wild-type plants, and this change was similarly induced for ERECTA-YFP in crispr_nrpm mutants (Figures S6C and S6D), indicating NRPMs may not participate in the PVC maturation process.

Functional homeostasis of the plasma-membrane proteins is regulated by endocytosis and endocytic recycling. In plants, Brefeldin A (BFA), an ARF-GEF inhibitor, can block endocytic recycling, resulting in the membrane proteins accumulating as the “BFA-bodies”.^38,39 To further investigate the effects of loss of NRPM on the subcellular dynamics of ERECTA, we used an endocytic tracer FM4–64 (8 μM) to stain the ERECTA-YFP seedlings for 40 mins followed by 70 μM BFA for 1 hour. Results showed that the BFA treatment induced ERECTA-YFP to aggregate with FM4–64 in the BFA bodies (Figure 5G), indicating the constant recycling of ERECTA-YFP to the PM. Furthermore, in crispr_nrpm-1–3 mutants, more ERECTA-containing BFA-bodies were formed (Figure 5G with quantification), supporting a positive role of NRPMs in facilitating ERECTA to return to the plasma membrane. Therefore, our studies suggest that the ERECTA receptor undergoes endocytosis, followed by endocytic recycling back to the plasma membrane and vacuolar targeting for degradation, and NRPMs may promote the endocytic recycling of ERECTA to the cell surface.

NRPMs function partially dependent on receptor signaling

The core signaling pathway regulating stomatal development involves the EPF peptide ligands, the perceiving ER-f receptors at the cell surface, and the YDA MAPK cascade in the cytoplasm, all of which are negative regulators of the key cell-fate transcription factors in stomatal production (Figure 6A). To test the genetic relationship of NRPMs with this pathway, we first overexpressed the peptide ligands EPF1/2 in the WT plants and crispr_nrpm-1–3 mutants, respectively (Figures S7A and S7B). Results showed that, differing from their effect of suppressing stomatal production in the WT plants, elevated expression of EPF1/2 failed to do so in the absence of NRPM1/2/4 (Figure 6B), indicating that NRPMs are downstream and required for the EPF-mediated signaling. This is also consistent with our results that NRPMs function inside the plasma membrane (Figures 3D and 3E), whereas the EPF peptides are apoplastic regulators.^14,15

Figure 6. NRPMs function downstream of the EPF ligands and receptor/co-receptor complexes.

(A) Graphic diagram describes the linear signaling pathway comprised of external peptide ligands (EPF), transmembrane receptors (TMM/ER-f, SERKs), cytoplasmic MAPK cascade (MAPKKK YDA-MAPKK 4/5-MAPK 3/6), and nuclear transcription factors (SPCH and SCRMs) in the regulation of Arabidopsis stomatal development. PM, plasma membrane; NE, nuclear envelope.

(B) crispr_nrpm-1–3 is epistatic to EPF1 and EPF2. Overexpression of EPF1 and EPF2 fails to suppress stomata formation in crispr_nrpm-1–3 plants as in the WT background. Expression levels of EFP1/EPF2 are provided in Figures S7A and S7B. Stomata guard cells and small dividing cells are colored in cyan and pink, respectively. Scale bars, 25 μm.

(C) Overexpression of myristoylated NRPM1 suppresses stomatal formation in tmm-1. 35S::myr-NRPM1-YFP reduced stomatal formation in tmm-1, indicating that the TMM receptor-like protein is not required for NRPM1 to function. Scale bars, 25 μm.

(D) Quantification of stomatal and lineage index in 5-dpg cotyledons of the genotypes as indicated. n, number of cotyledons analyzed.

(E) Stomatal phenotypes of indicated genotypes. crispr_nrpm-1–3;er105 shows additive stomata overproduction compared to er105 or crispr_nrpm-1–3 (see quantification data in Figure S7D). 35S::NRPM1-YFP suppressed stomatal formation in the loss-of-function ERECTA receptor family, i.e. er105 single, er105;erl2–2 double, and er105;erl1-c;erl2-c triple (c, crispr allele, mutations shown in Figure S7C). Scale bars, 25 μm.

(F) Quantification of the stomatal phenotypes observed in the genotypes shown in (E). Boxes in (D and F) show first and third quartiles with mean (square) and median (line). Whiskers extend to 1.5 × IQR from the first and third quartile. Student’s unpaired t tests were used to calculate two-tailed P values. n, number of cotyledons analyzed. The methods used for the generation of combinatory genotypes are described in detail in the Methods part (under “Creation of genetic materials”).

See also Figure S7.

The ERECTA receptor-like kinase family, together with the TMM receptor-like protein and the SERK kinases, plays pivotal roles in suppressing stomatal formation.^16–18 In the absence of TMM, the mutant plant produces more clustered stomatal guard cells.¹⁷ When the three ER-f members were progressively mutated, the stomatal overproduction phenotype became progressively severe.¹⁹ As we found overexpression of NRPM1 and myr-NRPM1 can suppress stomatal production (Figures 3C and 3E), we tested this regulation in the mutants carrying genetic lesions in the TMM or ER-f receptors. Results showed that myr-NRPM1 was capable of suppressing the stomatal overproduction phenotype in tmm-1 mutants (Figures 6C and 6D). Similarly, in the loss-of-function mutants, er105, er105;erl2–2, or the newly generated er105;erl1-c;erl2-c triple mutant (ERL1 and ERL2 were knocked out by CRISPR/Cas9 in er105, Figure S7C), NRPM1-YFP was capable of suppressing the production of stomatal lineage cells in all these genotypes (Figures 6E and 6F). In addition, when the mutations of crispr_nrpm-1–3 were combined with er105, we obtained an additive phenotype in stomatal overproduction (Figures 6E and S7D), suggesting the NRPM proteins may suppress stomatal production independent of these cell-surface receptors. However, the ability of NRPM1-YFP in suppressing stomatal lineage cells was indeed gradually attenuated with the addition of loss of ER-f members one by one (34.8% inhibition in er105, 20.1% inhibition in er105;erl2–2, and 17.3% inhibition in er-f triple mutant). Taken together, the genetic tests shown above indicate the NRPM proteins function downstream of the EPF1/2 and are likely partially dependent on ER-f kinases-mediated cell signaling in stomatal development. The fact that NRPM1 overexpression in er-f triple mutants maintained its activity in suppressing stomatal development also revealed its novel function independent of the ER-f receptors.

YDA is required for NRPMs to function

Downstream the cell-surface receptors, the canonical MAPK cascade, comprised of the MAPKK kinase (YDA), MAPK kinases (MKK4 and MKK5), and MAPKs (MPK3 and MPK6),^9,10 acts to phosphorylate the bHLH transcription factors SPCH and SCRM/ICE1 for degradation, thereby suppresses the stomatal formation.^13,40,41 When YDA is absent, the leaf epidermis is predominantly covered with stomatal lineage and guard cells in the super dwarf mutant plant.⁹ Interestingly, this strong growth phenotype was recapitulated in some individuals among the crispr_nrpm-1 population (Figure 7A). Inspired by this observation, we measured the MPK3/6 activity levels in crispr_nrpm mutants and plants overexpressing NRPM proteins. Results of immunoblotting by the p44/42 MAPK (Erk1/2) antibody indeed showed reduced MPK3/6 activities in crispr_nrpm and elevated activities in NRPM-overexpressed plants (Figures 7B and 7C), indicating a positive relationship between NRPM and the YDA MAPK pathway. To genetically test it, we overexpressed myr-NRPM1 in yda-3 null mutants and found that myr-NRPM failed to suppress stomatal lineage cells (Figure 7D), indicating that YDA is required for the NRPM-mediated function. On the other hand, elevated YDA activity (constitutively active YDA, YDA^CA driven by the TMM promoter) fully abolished stomatal lineage initiation and stomatal production in WT. In line with this phenotype, YDA^CA maintained the same activity in crispr_nrpm-1–3 (Figure 7E). Similarly, MKK5^CA10 or MPK6^CA42 recapitulated the effect of YDA^CA in crispr_nrpm-1–3 (Figures 7F and 7G), indicating that the YDA MAPK module functions downstream of NRPMs. In addition, the activity of SPCH transcription factor is enhanced in crispr_nrpm-1–3 plants (Figure S7E), which is consistent with the finding that NRPMs positively regulate the YDA MAPK pathway (Figure 7B). Finally, SPCH is absolutely required for stomatal lineage initiation, regardless of the presence or absence of NRPMs (Figure S7F). Thus, our biochemical and genetic data suggested that the NRPM proteins function upstream of the YDA MAPK cascade to suppress stomatal production in Arabidopsis.

Figure 7. NRPMs function upstream of the YDA MAPK pathway.

(A) Comparably growth phenotypes of a strong crispr_nrpm-1 mutant (segregating from the T2 population of nrpm1;2;4) with a yda-3 loss-of-function mutant (7-dpg). Scale bars, 1 mm.

(B) Representative immunoblotting data show NRPMs positively regulate the activities of MPK3/6 in vivo. Anti-p42/p44 and anti-MPK6 were used to show kinase activity and protein levels of MPK6, respectively. Total proteins were extracted from 5-dpg seedlings. (C) Quantification of MPK3/6 activity levels (p42/44) in plants with indicated genotypes in (B). The intensities of phosphorylated bands in WT were normalized to 1. Values are mean ± SD. n = 3 or 4 biological replicates.

(D-G) Confocal images show stomatal phenotypes in indicated genotypes. Images show the adaxial side of 3-dpg cotyledons. Stomata and small dividing cells are colored in cyan and pink, respectively. (D) YDA is required for the NRPM-mediated function. Overexpression of myr-NRPM1-YFP fails to suppress stomata formation in yda-3 mutant. (E-G) Activation of YDA-MKK4/5-MPK3/6 MAPK cascade in stomatal lineage cells suppresses stomata formation in crispr_nrpm-1–3 plants. Constitutively activated YDA (E), MKK5 (F) and MPK6 (G) driven by TMM promoter cause a stomataless phenotype. Scale bars, 20 μm. Detailed strategies used for the generation of combinatory genotypes are described in the “Methods” under “Creation of genetic materials”.

(H) A working model for how NRPMs function in stomata development in Arabidopsis. We propose that NRPMs associate the plasma membrane and function upstream of the cytoplasmic YDA MAPK pathway. Functions of NRPMs may promote the ERECTA receptor to stabilize and function at the plasma membrane directly or indirectly (?). NRPMs can also function independently of the ERECTA-family receptors via unknown regulatory mechanisms (?). Activities of NRPMs at the plasma membrane lead to elevated YDA-MKK4/5-MPK3/6 signaling activity, thus lowering the amount of SPCH/SCRMs and suppressing stomatal production.

DISCUSSION

Our work identified and characterized a non-canonical HRGPs family protein of NRPM, a group of novel negative regulators of stomatal production in Arabidopsis. HRGPs are a class of plant cell wall proteins that contain abundant hydroxyproline residues and are often heavily glycosylated.⁴³ HRGPs are divided into three major subfamilies based on their protein structure and glycosylation patterns, including arabinogalactan proteins, extensins, and proline-rich proteins.^43–45 Although the NRPMs were annotated as HRGPs, they contain unknown functional domains, therefore were not enlisted in the proposed 166 HRGPs in Arabidopsis.⁴³ We also present several lines of evidence to support that NRPMs are not cell wall proteins. (1) The plasmolysis assay demonstrates NRPMs are plasma membrane-associated proteins (Figure 2E). (2) The myr-NRPM1 but not sp-NRPM1 maintains the function to suppress stomata formation, suggesting that NRPMs play roles from the inner side of the plasma membrane (Figures 3D and 3E). (3) None of the NRPMs has an N-terminal endoplasmic reticulum signal sequence that is one typical structure for most cell wall HRGPs.^44,45 HRGPs have been reported to play roles during signal transduction.^46,47 For example, the extracellular leucine-rich repeat extensins LRX3/4/5 function with RAPID ALKALINIZATION FACTOR (RALF) peptides RALF22/23 and receptor-like kinase FERONIA to specify a signaling pathway to response salt stress.⁴⁷ But unlike LRXs, NRPMs neither have a leucine-rich repeat domain nor localize to the cell wall. Further research is needed to elucidate the precise mechanisms by which NRPMs associate with the inner plasma membrane and how this association contributes to signal transduction during stomatal development.

Interestingly, the nrpm mutant plants exhibiting different levels of stomata and growth defects were isolated from the same CRISPR/Cas9-edited T1 plant (Figure 4A). In the meanwhile, the nrpm mutant plants exhibiting comparable phenotypes but harboring different genotypes were identified from independent T1 plants (Figures 4A and 5A). These phenomena indicated that NRPMs (1/2/4) function redundantly and in a dose-dependent manner. A diverse array of weak/moderate nrpm alleles generated by CRISPR/Cas9 enabled us to overcome embryonic lethality and perform genetic manipulation to assay their genetic position and molecular mechanisms in different aspects of plant development and growth. In particular, the NRPM family functions upstream of YDA at the plasma membrane, leading to elevated activities of MPK3 and 6 that suppresses the differentiation of stomatal lineage cell (working model, see Figure 7H). This is based on the following key findings: (1) NRPM1 was isolated from ACD-enriched cell populations (Figure 1E) and active in stomatal lineage cells (Figures 2A, 2C, and 2D). (2) When NRPMs are elevated in expression, stomata formation is suppressed (Figure 3). On the other hand, loss of nrpm causes excessive and clustered stomata (Figures 4B–4D). (3) NRPMs suppress stomata formation downstream of EPF1/2 (Figure 6B) and are partially dependent on the cell-surface ER-f receptors (Figures 6E and 6F). (4) NRPMs activate MPK3 and 6 (Figures 7B and 7C), and their roles heavily rely on the YDA-MKK4/5-MPK3/6 cascade (Figures 7D–7G). Besides YDA (MAPKKK4), NRPMs may also through MAPKKK3 and MAPKK5, both of which regulate redundantly with YDA in plant immunity, growth and development. ⁴⁸

Importantly, our study identified an intricate genetic interaction between NRPM proteins and the ER-f receptor kinases in both vegetative and reproductive development. The main linkage that emerged from the phenotypic and cell biological analyses appears to support the positive role of NRPM proteins in maintaining the plasma-membrane homeostasis of the ERECTA receptor. This conclusion was based on the enhanced function and protein accumulation of ERECTA at the plasma membrane (Figures 5B and 5C), combined with the fact that ERECTA becomes more internalized in the crispr_nrpm-1–3 mutants (Figures 5C and 5G). A few recent reports started revealing that the subcellular dynamics of ER-f receptors are associated with their functions in specifying stomatal cell fate.^31,32 Yang et al. (2022) reported that the protein abundance of ERECTA is reduced by the loss of ERDJ3B chaperon protein that causes a damaged quality control system in the endomembrane reticulum.³² At the plasma membrane, ERECTA is activated after perceiving the EPF2 peptides, whilst its turnover is achieved by trans-phosphorylation of the co-receptor BAK1 on the PUB30/31 E3 ligases, which ubiquitinate ERECTA for eventual degradation, forming a self-organized negative feedback regulation.³⁶ Moreover, the co-receptor ERL1 undergoes constant internalization and recycling, demonstrating sophisticated trafficking behaviors related to the activities of upstream ligands and other receptors.³¹ Here, we identified that mutations in the NRPM genes caused more retention of ERECTA in the endomembrane system, mainly at the PVC and vacuolar compartments (Figures 5C–5F). The BFA treatment also revealed defects of returning ERECTA to the plasma membrane in crispr_nrpm-1–3 mutants, consistently suggesting that NRPMs help to maintain the plasma-membrane localization of ERECTA. Further investigations are necessary to address several outstanding questions. For example, whether the NRPM proteins interact with ERECTA directly or indirectly? Why do NRPMs regulate the ERECTA receptor more preferentially than the other co-receptors? Though, we should not rule out the possibility that the localization or activities of the other receptor kinase can be altered in strong or null nrpm mutants. Eventually, how broadly are NRPMs regulating the proteins at/in the plasma membrane?

Moreover, in the absence of ER-f receptors, NRPM1 maintains the ability to suppress stomatal formation, indicating the existence of uncharacterized pathways independent of the ER-f receptors that are regulated by NRPMs at the plasma membrane. As a group of newly identified membrane-associated proteins, NRPMs deserve a significant endeavor for further characterization, given the severe developmental defects observed in the nrpm mutants.

STAR★METHODS

RESOURCE AVAILABILITY

Lead contact

Further information and requests for resources and reagents should be directed to and will be fulfilled by the lead contact, Juan Dong (dong@waksman.rutgers.edu).

Materials availability

All plasmids and Arabidopsis thaliana lines generated in this study are available from the lead contact upon request.

Data and code availability

• Raw RNA-seq data have been deposited on NCBI’s Gene Expression Omnibus (GEO) under the project number GSE226455.

• No original code is reported in this study.

• Additional information required to reanalyze the data reported in this paper is available from the lead contact upon request.

EXPERIMENTAL MODEL AND SUBJECT DETAILS

Arabidopsis (Arabidopsis thaliana) plants are in the Columbia (Col-0) background as described in the key resources table. Arabidopsis and tobacco Nicotiana benthamiana were grown in controlled growth conditions, as described in method details.

KEY RESOURCES TABLE.

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Antibodies
Phospho-p44/42 MAPK (Erk1/2) (Thr202/Tyr204)	Cell Signaling Technology	Cat#9101; RRID: AB_331646
Anti-rabbit IgG, HRP-linked Antibody	Cell Signaling Technology	Cat#7074; RRID: AB_2099233
Bacterial and virus strains
Escherichia coli TOP10	N/A	N/A
Agrobacterium tumefaciens GV3101	N/A	N/A
Chemicals, peptides, and recombinant proteins
TRIzol	Invitrogen	Cat#15596018
SYBR Green Master Mix kit	Applied Biosystems	Cat#4309155
Protease Inhibitor Cocktail	Sigma-Aldrich	Cat#P9599
Propidium Iodide	Invitrogen	Cat#P3566
FM4-64	MedChemExpress	Cat#HY-103466
Brefeldin A	MedChemExpress	Cat# HY-16592
Wortmannin	Aladdin	Cat# W409260-1ml
Gateway LR Clonase II Enzyme mix	Invitrogen	Cat#11791100
Gateway BP Clonase II Enzyme mix	Invitrogen	Cat#11789020
Critical commercial assays
Plant RNeasy Mini Kit	QIAGEN	Cat#74904
TruSeq RNA Sample Prep Kit v2	Illumina	Cat#RS-122-2001
SuperScrip First-Strand Synthesis System	Invitrogen	Cat#11904018
Turbo DNA-free Kit	Invitrogen	Cat#AM1907
Deposited data
Raw and processed RNA-seq data	This study	GEO: GSE226455
Experimental models: Organisms/strains
Arabidopsis thaliana, Col-0 ecotype (WT)	N/A	N/A
Arabidopsis nrpm1	ABRC	SALK_015461
Arabidopsis nrpm2	ABRC	SALK_127646
Arabidopsis nrpm3	ABRC	WiscDsLox501H12
Arabidopsis nrpm4	ABRC	SALK_059180
Arabidopsis nrpm1;nrpm2	This study	N/A
Arabidopsis nrpm2;nrpm4	This study	N/A
Arabidopsis nrpm1−/+;nrpm4−/+	This study	N/A
Arabidopsis spch-4	ABRC	SALK_078595
Arabidopsis yda-3	ABRC	SALK_105078
Arabidopsis tmm-1	ABRC	CS6140
Arabidopsis erl2-2	ABRC	SALK_015275
Arabidopsis basl-2	ABRC	WiscDsLox264F02
Arabidopsis er105	Torii, et al.49	N/A
Arabidopsis er105;erl2-2	This study	N/A
Arabidopsis er105;erl1-c;erl2-c	This study	N/A
Arabidopsis er105;erl1-2;erl2-1−/+	Shpak, et al.16	N/A
Arabidopsis SPCH::YDACA-YFP	Lampard, et al.4	N/A
Arabidopsis 35S::SPCHΔ49	Lampard, et al.4	N/A
Arabidopsis 35S::SPCHΔ49 basl-2	This study	N/A
Arabidopsis SPCH::SPCH-CFP	Davies, et al.50	N/A
Arabidopsis ERECTA::ERECTA-YFP er105	Lee, et al.51	pJM284
Arabidopsis 35S::GFP-PIP2A	Cutler, et al.52	Q8
Arabidopsis ML1:mCherry-RCI2A	Lau, et al.53	N/A
Arabidopsis 35S::NRPM1-YFP	This study	N/A
Arabidopsis 35S::NRPM2-YFP	This study	N/A
Arabidopsis 35S::NRPM3-YFP	This study	N/A
Arabidopsis 35S::NRPM4-YFP	This study	N/A
Arabidopsis NRPM1::nucYFP	This study	N/A
Arabidopsis NRPM2::nucYFP	This study	N/A
Arabidopsis NRPM3::nucYFP	This study	N/A
Arabidopsis NRPM4::nucYFP	This study	N/A
Arabidopsis NRPM1::NRPM1-YFP	This study	N/A
Arabidopsis NRPM2::NRPM2-YFP	This study	N/A
Arabidopsis NRPM3::NRPM3-YFP	This study	N/A
Arabidopsis NRPM4::NRPM4-YFP	This study	N/A
Arabidopsis NRPM1::rNRPM1-YFP	This study	N/A
Arabidopsis NRPM2::rNRPM2-YFP	This study	N/A
Arabidopsis NRPM4::rNRPM4-YFP	This study	N/A
Arabidopsis 35S::myr-NRPM1-YFP	This study	N/A
Arabidopsis 35S::sp-NRPM1-YFP	This study	N/A
Arabidopsis UBQ::Cas9/crispr_nrpm4;3;2;1	This study	N/A
Arabidopsis UBQ::Cas9/crispr_erl	This study	N/A
Arabidopsis 35S::EPF1	This study	N/A
Arabidopsis 35S::EPF2	This study	N/A
Arabidopsis 35S::NRPM3-mRFP	This study	N/A
Arabidopsis 35S::NRPM1-mCherry er105	This study	N/A
Arabidopsis TMM::YDACA-YFP	This study	N/A
Arabidopsis TMM::MKK5CA-YFP	This study	N/A
Arabidopsis TMM::MPK6CA	This study	N/A
Arabidopsis ERECTA::ERECAT-YFP (3002)	This study	N/A
Arabidopsis ERL1::ERL1-YFP	This study	N/A
Arabidopsis SERK3::SERK3-YFP	This study	N/A
Arabidopsis ERL1::ERL1-YFP crispr_nrpm-1-3	This study	N/A
Arabidopsis SERK3::SERK3-YFP crispr_nrpm-1-3	This study	N/A
Arabidopsis AT4G26660::nucYFP	This study	N/A
Arabidopsis AT1G53140::nucYFP	This study	N/A
Arabidopsis NRPM2::NRPM2-YFP nrpm2;nrpm4	This study	N/A
Arabidopsis crispr_nrpm-1-1	This study	N/A
Arabidopsis crispr_nrpm-1-2	This study	N/A
Arabidopsis crispr_nrpm-1-3	This study	N/A
Arabidopsis crispr_nrpm-4-1	This study	N/A
Arabidopsis crispr_nrpm-4-2	This study	N/A
Arabidopsis crispr_nrpm-6-3	This study	N/A
Arabidopsis NRPM1::rNRPM1-YFP crispr_nrpm-1-1	This study	N/A
Arabidopsis NRPM2::rNRPM2-YFP crispr_nrpm-1-1	This study	N/A
Arabidopsis NRPM4::rNRPM4-YFP crispr_nrpm-1-1	This study	N/A
Arabidopsis ERECAT-YFP (3002) × NRPM3-mRFP	This study	N/A
Arabidopsis NRPM1-mCherry × ERECTA-YFP er105	This study	N/A
Arabidopsis ERECTA-YFP er105;crispr_nrpm-1-3	This study	N/A
Arabidopsis ERECTA-YFP (3002) crispr_nrpm-1-3	This study	N/A
Arabidopsis 35S::EPF1 crispr_nrpm-1-3	This study	N/A
Arabidopsis 35S::EPF2 crispr_nrpm-1-3	This study	N/A
Arabidopsis 35S::myr-NRPM1-YFP tmm-1	This study	N/A
Arabidopsis 35S::myr-NRPM1-YFP yda-3	This study	N/A
Arabidopsis er105;crispr_nrpm-1-3	This study	N/A
Arabidopsis 35S::NRPM1-YFP er105	This study	N/A
Arabidopsis 35S::NRPM1-YFP er105;erl2-2	This study	N/A
Arabidopsis 35S::NRPM1-YFP er105;erl1-c;erl2-c	This study	N/A
Arabidopsis TMM::YDACA-YFP crispr_nrpm-1-3	This study	N/A
Arabidopsis TMM::MKK5CA-YFP crispr_nrpm-1-3	This study	N/A
Arabidopsis TMM::MPK6CA crispr_nrpm-1-3	This study	N/A
Arabidopsis SPCH::SPCH-CFP crispr_nrpm-1-3	This study	N/A
Arabidopsis spch-4;crispr_nrpm-1-3	This study	N/A
Oligonucleotides
See Table S2 for Oligonucleotides
Recombinant DNA
35S::PIP2A-mCherry	ABRC	CD3-1007
pX-YFP-GW	ABRC	6530642502
pHGY	Kubo, et al.54	N/A
pH35GY	Kubo, et al.54	N/A
pBGYN	Kubo, et al.54	N/A
R4pGWB540	Nakagawa, et al.55	N/A
pGWB654	Nakamura, et al.56	N/A
pCAMBIA1300	Mcelroy, et al.57	N/A
pCAMBIA2300	Mcelroy, et al.57	N/A
pAtU6-sgRNA-pAtUBQ-Cas9	Mao, et al.58	N/A
TMM::MPK6CA	Guo, et al.59	N/A
pENTR/D-TOPO	Invitrogen	Cat#K240020
35S::NRPM1-YFP	This study	N/A
35S::NRPM2-YFP	This study	N/A
35S::NRPM3-YFP	This study	N/A
35S::NRPM4-YFP	This study	N/A
NRPM1::nucYFP	This study	N/A
NRPM2::nucYFP	This study	N/A
NRPM3::nucYFP	This study	N/A
NRPM4::nucYFP	This study	N/A
AT4G26660::nucYFP	This study	N/A
AT1G53140::nucYFP	This study	N/A
NRPM1::NRPM1-YFP	This study	N/A
NRPM2::NRPM2-YFP	This study	N/A
NRPM3::NRPM3-YFP	This study	N/A
NRPM4::NRPM4-YFP	This study	N/A
NRPM1::rNRPM1-YFP	This study	N/A
NRPM2::rNRPM2-YFP	This study	N/A
NRPM4::rNRPM4-YFP	This study	N/A
35S::myr-NRPM1-YFP	This study	N/A
35S::sp-NRPM1-YFP	This study	N/A
UBQ::Cas9/crispr_nrpm4;3;2;1	This study	N/A
UBQ::Cas9/crispr_erl	This study	N/A
35S::EPF1	This study	N/A
35S::EPF2	This study	N/A
35S::NRPM3-mRFP	This study	N/A
35S::NRPM1-mCherry	This study	N/A
TMM::YDA CA -YFP	This study	N/A
TMM::MKK5 CA -YFP	This study	N/A
ERECTA::ERECAT-YFP (3002)	This study	N/A
ERL1::ERL1-YFP	This study	N/A
SERK3::SERK3-YFP	This study	N/A
Software and algorithms
ImageJ	Schneider et al.60	https://imagej.nih.gov
OriginPro 2022 v.9.9.0.225 (SR1)	OriginLab Corporation	https://www.originlab.com/index.aspx?go=Support&pid=4440
PANTHER v.16.0	Mi et al.61	http://www.pantherdb.org
KALLISTO v.0.44.0	Bray et al.62	https://anaconda.org/bioconda/kallisto/files?version=0.44.0
SLEUTH	Pimentel et al.63	https://pachterlab.github.io/sleuth/about
GPLOTS	Warnes et al.64	https://rdrr.io/cran/gplots/
Other
StepOnePlus Real-Time PCR System	Applied Biosystems	Cat#4376598
Nanodrop spectrophotometer	Thermo Fisher	Cat#ND-ONEC-W
Agilent 2100 Bioanalyzer	Agilent Technologies	Cat#G2939BA

Plant materials and growth conditions

The Arabidopsis thaliana ecotype Columbia (Col-0) was used as the wild-type unless otherwise noted. Seeds were surface sterilized with 70% ethanol for 5 min, rinsed three times with autoclaved distilled water. After 3-day imbibition in the dark at 4 °C, seeds were sown on half-strength Murashige & Skoog medium with 1% agar (A1296, Sigma), pH 5.8. Seeds germinate in a growth chamber under 16-h light/8-h dark cycle with 100 μmol m⁻² s⁻¹ light intensity at 22 °C. The 7-day-old seedlings were transferred to the soil for further growth under the same conditions. The Nicotiana benthamiana plants were grown in a growth room at 25°C under 14-h light/10-h dark cycle. The plasmid CD3–1007 (35S::PIP2A-mCherry)³³ and T-DNA insertional lines nrpm1 (SALK_015461), nrpm2 (SALK_127646), nrpm3 (WiscDsLox501H12), nrpm4 (SALK_059180), erl2–2 (SALK_015275), and spch-4 (SALK_078595) were obtained from Arabidopsis Biological Resource Center (ABRC). basl-2 (WiscDsLox264F02),²³ ML1:mCherry-RCI2A,⁵³ 35S::GFP-PIP2A,⁵² TMM::MPK6^CA,⁵⁹ er105,⁴⁹ er105;erl1–2;erl2–1,¹⁶ ERECTA::ERECTA-YFP er105 (pJM284),⁵¹ SPCH::SPCH-CFP,⁵⁰ and 35S::SPCHΔ49⁴ were described previously.

METHOD DETAILS

Plasmid construction and plant transformation

NRPM coding sequence (CDS) were amplified using Q5 high fidelity DNA polymerase (NEB), and then were inserted into JW819-VENUS (pCAMBIA3300 backbone) by TA cloning to make 35S::NRPMs-VENUS. In order to facilitate the subsequent vector construction, we switched to LR Clonase II (Invitrogen)-based gateway cloning technology for more constructions. NRPM encoding regions with stop codons removed were cloned into pENTR/D-TOPO (Thermo Fisher Scientific) through NotI and AscI sites, and then the promoters were inserted at NotI site immediately upstream of the coding sequences. The plasmid pENTR/D-TOPO carrying the NRPM coding region were recombined into pX-YFP-GW to make 35S::NRPMs-YFP, or into pGWB654⁵⁶ to make 35S::NRPMs-mRFP. The plasmid pENTR/D-TOPO carrying the NRPMp::NRPM were recombined into pHGY⁵⁴ to generate NRPMp::NRPM-YFP. The CRISP/Cas9 resistant version of NRPMs (rNRPMs) were generated through two round PCR to introduce the synonymous mutation in the Cas9-gRNA recognition site. The same method was used to make NRPM::rNRPMs-YFP, ERECTA::ERECTA-YFP (3002), ERL1::ERL1-YFP, and SERK3::SERK3-YFP in pHGY backbone. Briefly, the genomic coding sequence of genes was amplified and cloned into pENTR/D-TOPO by NotI and AscI. Then promoter region was inserted by the NotI site before the coding sequence. NRPM promoters were cloned into pENTR/D-TOPO through NotI and AscI sites. The pENTR/D-TOPO carrying NRPM promoters were recombined into pBGYN⁵⁴ to make NRPMp::nucYFP.

To make 35S::myr-NRPM1-YFP and 35S::sp-NRPM1-YFP, the NRPM1 coding region was cloned into pENTR/D-TOPO via NotI and AscI, while introducing a KpnI site immediately upstream of the coding sequences. Then myr was inserted via NotI and KpnI sites to make pENTR/D-TOPO/myr-NRPM1. The myr was replaced by the signal peptide of Extensin 3 (AT1G21310) to make pENTR/D-TOPO/sp-NRPM1. Finally, through LR reaction, the entry vectors and pH35GY were recombined to obtain 35S::myr-NRPM1-YFP and 35S::sp-NRPM1-YFP.

To make 35S::EPF1/2, the genomic sequence of EPF1 and EPF2 were amplified and inserted into pCAMBIA1300/35S::MCS::NOS via BamHI and KpnI. The CaMV 35S promoter, NOS terminator, and multiple clone sites (MCS) between them have been pre-installed on the pCAMBIA1300⁵⁷ backbone by HindIII and EcoRI.

The TMM promoter was cloned into pDONR-P4-P1R (Thermo Fisher Scientific) via BP reaction. The pENTR/D-TOPO containing YDA^CA and MKK5^CA with stop codons removed were reported previously.¹¹ Double LR recombination reactions were performed to integrate pENTR/D-TOPO containing YDA^CA or MKK5^CA and pDONR-P4-P1R/TMMp into the R4pGWB540⁵⁵ to make TMM::YDA^CA-YFP and TMM::MKK5^CA-YFP, respectively.

To create CRISPR/Cas9-mediated mutagenesis in Arabidopsis, we adopted the system described in a previous study.⁵⁸ Following the introduction, the oligos NRPM4-CRI-F and NRPM4-CRI-R were phosphorylated by T4 PNK (NEB) and annealed in a thermocycler. The annealed oligos were cloned into the plasmid pAtU6-sgRNA-pAtUBQ-Cas9 via BbsI site. The same procedure was performed to create other three guide RNA cassettes. The chimeric pAtU6-NRPM4-pAtUBQ-Cas9 cassette was cut off and subcloned into pCAMBIA2300⁵⁷ through HindIII and EcoRI sites to get 2300/crispr_nrpm4. U6-NRPM3 was amplified by PCR and inserted into 2300/crispr_nrpm4 through KpnI and EcoRI sites to get 2300/crispr_nrpm4;3. Using same strategy, U6-NRPM2 were inserted at KpnI site to create the construct 2300/crispr_nrpm4;3;2 and finally U6-NRPM1 were inserted at EcoRI site to create the construct 2300/crispr_nrpm4;3;2;1 to knock-out the whole family. Same strategy as building 2300/crispr_nrpm4, one single guide RNA targeting both ERL1 and ERL2 was cloned into pCAMBIA1300 to make 1300/crispr_erl. See also Table S2 for the primers used in this study.

The plasmids were transformed into Agrobacterium tumefaciens GV3101, which were used for stable transformation in Arabidopsis and transient expression in N. benthamiana. Arabidopsis plants were transformed with the standard floral dipping method,⁶⁵ and the T0 seeds were subjected to antibiotic selection.

Creation of genetic materials

35S::SPCHΔ49 had been reported previously⁴ and was crossed with basl-2 mutant to obtain 35S::SPCHΔ49;basl-2. The construct containing CRISPR/Cas9 targeting NRPMs was introduced to Col-0 plants expressing ML1::mCherry-RCI2A (WT) to create crispr_nrpm mutants. Stable crispr_nrpm-1–1 and crispr_nrpm-1–3 with moderate fertility defects were isolated from the crispr_nrpm-1 population and used for further genetic analysis. The rNRPM constructs were introduced to crispr_nrpm-1–1 by dipping and the representative plants are shown in Figure 4. The mCherry-tagged endomembrane organelle markers were introduced to ERECTA::ERECTA-YFP;er105 by dipping except RabG3f-mCherry by crossing.

For genetic interaction assay, especially the combinations of overexpression cassettes with mutants, we follow this principle: at least two independent overexpression lines were chosen to cross with mutants, and then the desired plants were isolated from the F2 or F3 population by genotyping. Due to some technical difficulties, a few exceptions are described here. 1) 35S::EPF1/2 were introduced into crispr_nrpm-1–3 and WT, respectively. Several T1 plants in WT background showing stomata-less phenotypes at the early stage (5-dpg), indicating our constructs work. But among over 120 T1 seedlings in crispr_nrpm-1–3 background we examined, none shows stomata-less phenotypes. To ensure the same transgenic alleles in both WT and crispr_nrpm-1–3 background, several independent T1 plants in crispr_nrpm-1–3 background (father) were crossed with Col-0 (mother). The stomata-less seedlings at 3-dpg were isolated from the F1 population, then images were captured from the corresponding father lines and stomata-less F1 plants, which were represented in Figure 6B; 2) The material 35S::NRPM1-YFP er105;erl1-c;erl2-c was created by introducing CRISPR/Cas9 targeting ERL1/2 into 35S::NRPM1-YFP er105 background by dipping; 3) TMM::YDA^CA, TMM::MKK5^CA, and TMM::MPK6^CA were introduced to crispr_nrpm-1–3 by dipping because YDA^CA, MKK5^CA, and MPK6^CA cause stomata-less phenotype in WT background.

RNA extraction and library preparation for RNA-seq

Total RNA was isolated from 12-day-old Arabidopsis seedlings with the RNeasy Plant Mini Kit (Qiagen, Germany) following the manufacturer’s instructions. RNA quantification was measured with a Nanodrop spectrophotometer (Thermo Fisher Scientific, USA) and the integrity of RNA was examined with the Agilent 2100 Bioanalyzer (Agilent Technologies, USA). RNA samples with RNA integrity number more than 7.0 were used for library preparation. The RNA-seq libraries were prepared at the Core Facility for Genomics at Shanghai Center for Plant Stress Biology using the Illumina TruSeq RNA Sample Prep Kit v2 LS protocol (Illumina, USA) and then sequenced on Illumina HiSeq2000 platform.

Transcriptome analysis

The transcript levels were quantified using KALLISTO v.0.44.0⁶² by mapping to Arabidopsis (genome release version 11) primary transcript sequences. Differentially expressed genes (DEGs) were identified using SLEUTH⁶³ as those showing |b| ≥ 0.692 and q-value < 0.05 with transcript per million (TPM) reads counts over 10 in one group at least. Hierarchical clustering was performed using GPLOTS⁶⁴ in R based on the TPM data for DEGs. Gene ontology (GO) enrichment analysis was performed using PANTHER v.16.0⁶¹ with significant P-values (P < 0.05) to export the top GO terms.

Chemical treatment

All chemical treatments were performed on 3-dpg Arabidopsis seedlings. Treatment after a certain period of time as described in the main text, the seedlings were mounted for confocal imaging. For internalization assay, ERECTA-YFP seedlings were incubated with 8 μM FM4–64 for 40 mins before imaging. For BFA treatment, seedlings were stained with 8 μM FM4–64 for 40 mins, followed by 1-hr treatment with 70 μM BFA. For wortmannin treatment, FM4–64 staining was followed by treatment with 33 μM wortmannin for 3 hrs.

Plant cell imaging and image processing

Confocal images of plant cells expressing fluorescence-tagged proteins were taken on a Leica SP5 confocal microscope. The 3-day-old adaxial side of cotyledons was captured. Cell outlines were visualized by propidium iodide (PI, Invitrogen). Seedlings were stained in PI solution (1:100 diluted) for 12–15 min. Fluorescence was excited at 514 nm (YFP) and 594 nm (PI). Emissions were collected at 520–540nm (YFP) and 620–640 nm (PI). The confocal images were false-colored, and brightness/contrast were adjusted using either Adobe Photoshop 23.1.0 or ImageJ.⁶⁰ To present stomata patterning clearly, images with cell outlines decorated by PI or YFP signals were converted to black and white, and then stomata and small dividing cells were colored in cyan and pink, respectively.

Quantitative analysis of stomatal phenotypes in Arabidopsis

The adaxial cotyledons of 5-day-old Arabidopsis seedlings were stained with propidium iodide (Invitrogen), and images were captured using a Carl Zeiss Axio Scope A1 fluorescence microscope equipped with a ProgRes MF CCD camera (Jenoptik). The Stomata index or stomatal lineage cells index was calculated as the number of stomata or stomatal lineage cells versus the total number of epidermal cells.

Transient expression in N. benthamiana epidermal cells

For the method of transiently expressing genes in N. benthamiana, we made some modifications based on a previous report.⁶⁶ Agrobacterium tumefaciens GV3101 harboring the plasmids of 35S::NRPMs-YFP were cultured overnight in 10 ml Luria broth medium containing appropriate antibiotics at 30°C. Bacterial cells were collected at 10,000 rpm for 1 min and resuspended with 10 ml water to wash out the antibiotics, followed by one more time washing. Then cells were resuspended in the infiltration buffer (10 mM MgCl₂, 10 mM MES pH 5.7, 200 μm acetosyringone) and the cell concentration was adjusted to OD₆₀₀ = 0.8. Cells were placed in culture medium for 2 hours at room temperature prior to infiltration. By using a 1-ml syringe, the Agrobacterium solution was infiltrated into leaves through an incision. At third-day post-infiltration, the fluorescence on the epidermis was observed and images were captured on a Leica SP5 confocal microscope.

Real-time PCR

Total RNAs were extracted from 3-day-old seedlings using TRIzol (Invitrogen). DNA was digested with the Turbo DNA-free Kit (Invitrogen). The first-strand cDNAs were synthesized by the SuperScrip First-Strand Synthesis System (Invitrogen) using 1 μg total RNAs as templates in a volume of 20 μl. Real-time PCR was performed with a SYBR Green Master Mix kit (Applied Biosystems) and amplification was monitored on a StepOnePlus Real-Time PCR System (Applied Biosystems). The gene expression level was normalized to reference gene S18C (AT4G09800) using the ΔCT method. Primers are listed in Table S2.

Statistics and data presentation

All boxplots show the range from first to third quartile with mean (square) and median (line). Whiskers extend to 1.5 × IQR from first and third quartile. Data points outside the whiskers were considered outliers. Column bar plots show the mean with standard error of mean. Statistical analyses were performed using either OriginPro 2022 v.9.9.0.225 (SR1) or Excel. A two tailed t-test was used for comparisons between two groups. One-way ANOVA and Tukey’s method were used for multiple comparisons. The individual values were provided in the source data files.

Examination of MPK3/6 activity in Arabidopsis plants

To evaluate the MAPK activity, 5-day-old seedlings were frozen in liquid nitrogen and homogenized by pestles. Total proteins were extracted in a buffer containing 100 mM HEPES (pH7.5), 5 mM EDTA, 5 mM EGTA, 10 mM NaF, 10 mM Na₃VO₄, 50 mM β-glycerophosphate, 5% glycerol, 10 mM DTT, 1 mM PMSF and 1% protease inhibitor cocktail for plant cell extracts (Sigma-Aldrich). After centrifugation at 13,000 rpm for 15 min, supernatants were transferred to new tubes. Equal amounts of total protein were loaded and separated from each other based on their size by the electrophoresis on a 12% SDS-PAGE gel. The MAPK activity was determined by immunoblotting with the primary antibody against Phospho-p44/42 MAPK (Erk1/2) (Thr202/Tyr204) (1:1000, Cell Signaling Technology). Equal loading was indicated by immunoblot analysis with anti-MPK6 antibody.

Accession numbers

NRPM1 (AT4G25620), NRPM2 (AT5G52430), NRPM3 (AT1G63720), NRPM4 (AT1G76660), SPCH (AT5G53210), BASL (AT5G60880), EXT3 (AT1G21310), ERECTA (AT2G26330), ERL1 (AT5G62230), ERL2 (AT5G07180), SERK3 (AT4G33430), TMM (AT1G80080), YDA (AT1G63700), MKK5 (AT3G21220), MPK6 (AT2G43790), MPK3 (AT3G45640).

QUANTIFICATION AND STATISTICAL ANALYSIS

Differentially expressed genes (DEGs) were identified using SLEUTH.⁶³ Gene ontology enrichment analysis was performed using PANTHER v.16.0.⁶¹ Statistical analysis of quantitative data were performed using OriginPro 2022 v.9.9.0.225 (SR1) or Microsoft Excel. Statistical tests used, sample sizes (n) and their representations of each experiment are described in the corresponding figure legends. Significance was defined as P value < 0.05.

Supplementary Material

2

Data S1. RNAseq data analyses, related to Figure 1.

Highlights.

• NRPM1 is highly expressed during stomatal asymmetric cell division

• NRPMs are four novel stomatal repressors functioning inside the plasma membrane

• Mutating NRPM genes attenuates the plasma-membrane localization of ERECTA

• NRPMs function upstream of the MAPKKK YDA in stomatal development