Abstract The endoplasmic reticulum (ER) and the plasma membrane (PM) form ER–PM contact sites (EPCSs) that allow the ER and PM to exchange materials and information. Stress-induced disruption of protein folding triggers ER stress, and the cell initiates the unfolded protein response (UPR) to resist the stress. However, whether EPCSs play a role in ER stress in plants remains unclear. VESICLE-ASSOCIATED MEMBRANE PROTEIN (VAMP)-ASSOCIATED PROTEIN 27-1 (VAP27-1) functions in EPCS tethering and is encoded by a family of 10 genes (VAP27-1–10) in Arabidopsis thaliana. Here, we used CRISPR-Cas9-mediated genome editing to obtain a homozygous vap27-1 vap27-3 vap27-4 (vap27-1/3/4) triple mutant lacking three of the key VAP27 family members in Arabidopsis. The vap27-1/3/4 mutant exhibits defects in ER–PM connectivity and EPCS architecture, as well as excessive UPR signaling. We further showed that relocation of VAP27-1 to the PM mediates specific VAP27-1-related EPCS remodeling and expansion under ER stress. Moreover, the spatiotemporal dynamics of VAP27-1 at the PM increase ER–PM connectivity and enhance Arabidopsis resistance to ER stress. In addition, we revealed an important role for intracellular calcium homeostasis in the regulation of UPR signaling. Taken together, these results broaden our understanding of the molecular and cellular mechanisms of ER stress and UPR signaling in plants, providing additional clues for improving plant broad-spectrum resistance to different stresses. Key words: Arabidopsis, VAP27-1, ER–PM contact sites, ER stress, unfolded protein response __________________________________________________________________ This study reports that mutation of three key genes encoding VAP27 endoplasmic reticulum–plasma membrane (ER–PM) contact site (EPCS) tethers will cause defects in ER–PM connectivity and EPCS architecture, as well as excessive unfolded protein response signaling regulated by extracellular Ca^2+ influx. It further provides evidence demonstrating that VAP27-1-EPCSs play a crucial role in resistance to ER stress by modulating ER–PM contact architecture and connectivity through the molecular re-arrangement and dynamic variation of VAP27-1 at the PM. Introduction The interactions between organelles and the frequent exchange of materials and information between membrane-bound organelles have key roles in the maintenance of cellular homeostasis in eukaryotes ([41]Eisenberg-Bord et al., 2016; [42]Scorrano et al., 2019). Organelles with different functions transport functional proteins through classic vesicular transport and establish and maintain inter-organelle membrane contact sites (MCSs) to exchange lipids and bioactive small molecules. MCSs are dynamic and ubiquitous inter-organellar structures in which two organelles reside within 30 nm of each other without undergoing membrane fusion ([43]Scorrano et al., 2019). MCSs are critical regulatory hubs for crosstalk between organelles. Therefore, the study of organelle interaction and membrane contact is key for understanding cellular functions. The endoplasmic reticulum (ER) has the largest membrane surface area of any organelle within the cell and forms a network of ER tubules and cisternae that are interconnected like a spider’s web ([44]Stefano and Brandizzi, 2018). Tubular ER can extend and branch throughout the cytoplasm, forming MCSs with other organelles and the plasma membrane (PM) ([45]Stefano et al., 2014; [46]Wu et al., 2018). In early electron microscopy observations, researchers identified binding sites between the cortical ER (cER) and the PM, known as ER–PM contact sites (EPCSs) ([47]Hepler et al., 1990). More advanced optical microscopy techniques subsequently confirmed the presence of EPCSs by capturing the dynamics of ER tubules ([48]Griffing et al., 2014). There are almost no ribosomes at EPCSs, and ER–PM membrane fusion does not occur. However, the distance between the ER and PM ranges from 10 to 30 nm, making it possible to transport cargo and communicate signals between the ER and PM ([49]Saheki and De Camilli, 2017). The functions of EPCSs in calcium ion (Ca^2+) transport, lipid transport, stress responses, and other processes have been revealed in mammalian cells and yeast ([50]Stefan et al., 2013; [51]Balla et al., 2020). Four EPCS-tethering proteins have been identified in Arabidopsis: SYNAPTOTAGMIN1 (SYT1), VESICLE-ASSOCIATED MEMBRANE PROTEIN (VAMP)-ASSOCIATED PROTEIN 27-1 (VAP27-1), NETWORKED 3C (NET3C), and VAP-RELATED SUPPRESSOR OF TMM (VST1) ([52]Schapire et al., 2008; [53]Wang et al., 2014; [54]Ho et al., 2016). Interestingly, live imaging of EPCS tethers in Arabidopsis showed that SYT1-enriched EPCSs (S-EPCSs) are physically separated from VAP27-1-enriched EPCSs (V-EPCSs) ([55]Siao et al., 2016). This finding supports the idea that different tethering complexes result in different functions of the EPCSs ([56]Bayer et al., 2017). Recently, studies of EPCS-tethering proteins in Arabidopsis have revealed their essential roles in important cellular processes, including immune activity ([57]Kim et al., 2016), autophagy ([58]Ye et al., 2022), regulation of virus cell-to-cell movement ([59]Lewis and Lazarowitz, 2010; [60]Levy et al., 2015), stabilization of cER architecture ([61]Siao et al., 2016), Ca^2+-dependent resistance to abiotic stresses ([62]Schapire et al., 2008; [63]Yamazaki et al., 2008; [64]Sancho et al., 2015), endocytosis ([65]Stefano et al., 2018), and seed and root-hair development ([66]Wang et al., 2014, [67]2016). Plants live in dynamically changing environments that are often unfavorable for their growth and development. Many studies have found that various biotic and abiotic stresses cause significant increases in unfolded or misfolded proteins in the ER, thus resulting in ER stress ([68]Pastor-Cantizano et al., 2020). Although studies on stress perception by plant cells have focused mainly on ligand–receptor interactions that occur on the cell surface, ER stress is now widely recognized as an important response to stress conditions ([69]Zhu, 2016). When ER-stress sensors in plant cells perceive a stress stimulus, the plant cell initiates the unfolded protein response (UPR). The UPR includes increased transcription of downstream genes, such as those encoding molecular chaperones, to help proteins fold correctly or accelerate the degradation of misfolded proteins ([70]Liu and Howell, 2010; [71]Nagashima et al., 2011; [72]Iwata and Koizumi, 2012). In mammalian cells, the cER and its contact sites with the PM respond to ER stress ([73]van Vliet et al., 2017). However, the roles of plant EPCSs and their modes of regulation during stress remain unknown. Non-plant VAPs have generally been described as type II ER integral membrane proteins, acting as membrane tethers between ER and endosomes ([74]Alpy et al., 2013) as well as between the ER and PM at EPCSs ([75]Stefan et al., 2011). VAPs have diverse physiological functions, including lipid exchange and membrane trafficking ([76]Lev et al., 2008), but their biological roles are still being explored in plants. Here, we investigated the link between ER–PM connectivity and the dynamics and distribution of VAP27-1 in response to ER stress in Arabidopsis. High-precision cytological imaging and phenotypic evidence showed that VAP27-1-EGFP localizes in the ER and PM and is enriched in EPCSs. VAP27-1 regulated resistance to ER stress in Arabidopsis, and this regulation required its connection to the PM. Furthermore, non-invasive micro-test technology (NMT) and transcriptome sequencing showed excessive UPR signaling in the vap27-1 vap27-3 vap27-4 (vap27-1/3/4) triple mutant under ER stress, which was due to elevated intracellular Ca^2+ levels caused by increased Ca^2+ influx. More importantly, variable-angle total internal reflection fluorescence microscopy (VA-TIRFM) in conjunction with immunoblotting confirmed that VAP27-1 dynamics at the PM and changes in the subcellular distribution of VAP27-1 enhanced ER–PM connectivity. These results support the hypothesis that Arabidopsis VAP27-1 mediates ER–PM contact architecture and connectivity and is required for cellular resistance to ER stress. Results VAP27-1 localizes in the ER and PM and is highly enriched at EPCSs Prior to exploring the cellular and biological functions of Arabidopsis VAP27-1, we determined its endogenous subcellular localization in Arabidopsis on the basis of its previously observed localization after heterologous expression in Nicotiana benthamiana leaves driven by the 35S promoter ([77]Wang et al., 2014). Transgenic Arabidopsis expressing C-terminally EGFP-tagged VAP27-1 driven by its native promoter (VAP27-1-EGFP) were generated. Live-cell imaging using laser scanning confocal microscopy showed that VAP27-1-EGFP co-localized with the FM4-64-stained PM ([78]Supplemental Figure 1A and 1B). We compared the distribution of VAP27-1-EGFP with that of the PM marker protein GFP-Lti6a, which showed a continuous fluorescence signal at the PM upon confocal imaging. VAP27-1-EGFP displayed a non-uniform signal at the cell cortex in epidermal cells of the Arabidopsis cotyledon and hypocotyl ([79]Supplemental Figure 1C and 1D). To investigate the distribution of VAP27-1 at the ER, we used the ER marker ER-mCherry along with EGFP-MAPPER. MAPPER, a well-established and widely utilized EPCS marker, is commonly used in cytological studies in both plant and mammalian cells ([80]Chang et al., 2013; [81]Lee et al., 2019). We acquired transgenic plants co-expressing ER-mCherry and VAP27-1-EGFP as well as transgenic plants co-expressing VAP27-1-mCherry and EGFP-MAPPER. Live-cell imaging showed that VAP27-1 co-localized with both ER-mCherry and EGFP-MAPPER ([82]Figure 1A and 1B). Interestingly, VAP27-1-EGFP and ER-mCherry co-localized to hollow circles in the cytoplasm ([83]Figure 1A), a pattern that was not observed for EGFP-MAPPER ([84]Figure 1B). We also stained the nuclei of transgenic VAP27-1-mCherry plant cells using 4′,6-diamidino-2-phenylindole (DAPI). Merged images showed a hollow circular distribution of VAP27-1-mCherry wrapped around the DAPI-stained nucleus, indicating that VAP27-1 is located not only in the cell cortex but also in the ER surrounding the nucleus ([85]Figure 1C). Parallel PM and cell-component separation experiments showed that, similar to the EPCS marker EGFP-MAPPER, VAP27-1-EGFP was enriched in both the PM and organellar membrane, but it showed greater enrichment in the nucleus than EGFP-MAPPER ([86]Figure 1D). Figure 1. [87]Figure 1 [88]Open in a new tab VAP27-1-EGFP is unevenly distributed on ER–PM contact sites. (A) Co-localization of VAP27-1-EGFP and ER-mCherry in cotyledon epidermal cells of transgenic Arabidopsis seedlings. Gray value analysis along the yellow line segments is shown on the right. Cyan arrows indicate hollow circular structures inside the epidermal cells. Scale bars, 30 μm. (B) Co-localization of EGFP-MAPPER and VAP27-1-mCherry in cotyledon epidermal cells of transgenic Arabidopsis seedlings. Gray value analysis along the yellow line segments is shown on the right. Cyan arrows indicate hollow circular structures inside the epidermal cells. Magenta arrows represent the same area marked by the cyan arrows. Scale bars, 20 μm. (C) Confocal images of VAP27-1-mCherry in epidermal cells of Arabidopsis cotyledons. Transgenic Arabidopsis seedlings expressing VAP27-1-mCherry were incubated with DAPI to stain the nucleus. Gray value analysis along the yellow line segments is shown on the right. Scale bars, 25 μm. (D) Plasma membrane protein isolation and cell fractionation of transgenic Arabidopsis seedlings expressing VAP27-1-EGFP. Total proteins extracted from transgenic Arabidopsis seedlings expressing VAP27-1-EGFP were fractionated into plasma membrane (PM), organelle membrane (OM), cytosol (C), and nucleus (N). The EPCS marker EGFP-MAPPER, plasma membrane protein marker GFP-Lti6a, and soluble protein marker EGFP were used as controls. GFP and EGFP were immunoprecipitated using an anti-GFP antibody. Each experiment was repeated at least three times. (E–G) (E) Confocal images of VAP27-1-EGFP and ER-mCherry in cotyledon epidermal cells of transgenic Arabidopsis plants. Scale bars, 10 μm. Enlarged images in (F) and (G) were taken from the cyan-outlined area in (E). Three-dimensional luminance plots of VAP27-1-EGFP (F) and ER-mCherry (G) show varied fluorescence intensity analyzed with the Multi Measure plug-in for ImageJ. (H) TIRF-SIM images of cotyledon epidermal cells of transgenic Arabidopsis seedlings co-expressing VAP27-1-EGFP and mCherry-MAPPER. Gray value analysis along the yellow line segments is shown on the right. The blue arrows indicate areas where VAP27-1-EGFP has a fluorescent signal and mCherry-MAPPER has no detectable signal. Scale bars, 25 μm. To confirm the localization of VAP27-1 in the cortical ER, especially at the EPCSs, we compared the distribution pattern of VAP27-1-EGFP with that of ER-mCherry or mCherry-MAPPER in the cell cortex. VAP27-1 presented a network distribution similar to that of ER-mCherry, but it was enriched in puncta ([89]Figure 1E–1G). In addition, VAP27-1-EGFP extensively co-localized with mCherry-MAPPER at EPCSs, but mCherry-MAPPER did not exhibit a network-like distribution like that of VAP27-1-EGFP ([90]Figure 1H). Taken together, these results suggest that VAP27-1-EGFP localizes to the PM and ER, including the cortical ER and perinuclear ER, and is especially enriched in EPCSs. Lysine 59 and threonine 60 and 61 regulate VAP27-1 association with the PM To probe the dynamics of VAP27-1-associated EPCSs in the PM, we used VA-TIRFM to image VAP27-1-EGFP in live Arabidopsis cotyledon epidermal cells. The VAP27-1-EGFP fluorescent foci appeared as dispersed punctate structures at the PM ([91]Supplemental Figure 2A and 2B). Tracking the movement of these spots in real time using VA-TIRFM showed that VAP27-1-EGFP moved steadily at the PM over a 15-s time frame. Merged images revealed substantial co-localization of VAP27-1-EGFP at 0 s and 15 s, indicating that the lateral displacement of VAP27-1-associated EPCSs at the PM was minimal and that the motion state was relatively stable in the time range we tracked ([92]Figure 2A). Observing the trajectory of VAP27-1-EGFP spots in a 15-s time frame showed that these spots consistently exhibited restricted diffusion in a limited area ([93]Figure 2B). Figure 2. [94]Figure 2 [95]Open in a new tab Relatively stable dynamics of VAP27-1-EGFP at the PM and regulation of VAP27-1 connection to the PM by K59, T60, and T61. (A) VAP27-1-EGFP signal comparison between the initial (t = 0 s, in green pseudocolor) and final (t = 15 s, in magenta pseudocolor) frames during a 15-s time-lapse experiment. Scale bar, 5 μm. (B) Trajectories of VAP27-1-EGFP in cotyledon epidermal cells. An enlarged image of the region outlined in yellow is shown on the right. Scale bar, 1 μm. (C) PM protein isolation and cell fractionation of transgenic Arabidopsis seedlings expressing VAP27-1-EGFP, VAP27-1^K59N-EGFP, and VAP27-1^T60/61A-EGFP. Total proteins extracted from transgenic Arabidopsis seedlings were fractionated into plasma membrane (PM), organelle membrane (OM), cytosol (C), and nucleus (N). EGFP was immunoprecipitated using an anti-GFP antibody. Each experiment was repeated at least three times. (D) TIRF-SIM images of cotyledon epidermal cells of transgenic Arabidopsis seedlings expressing VAP27-1-EGFP, VAP27-1^K59N-EGFP, and VAP27-1^T60/61A-EGFP. Raw images were processed with the ImageJ lookup table “Green Fire Blue” to enhance signal visibility. Enlarged views of areas outlined in cyan are shown in the yellow boxes at the bottom left. Scale bars, 10 μm. Alignment of the conserved major sperm protein (MSP) domain of VAPs from various species indicated that a lysine residue at position 59 (K59) and two threonine residues at positions 60 (T60) and 61 (T61) in the Arabidopsis protein stabilize the non-specific electrostatic interactions between the MSP domain of VAP27-1 and an FFAT-like motif in its interacting proteins ([96]Neefjes and Cabukusta, 2021). Therefore, we mutated K59 to asparagine (K59N) and T60 and T61 to alanine (T60A and T61A) to explore the effect of the MSP–FFAT interaction on the binding of VAP27-1 to the PM. Transgenic Arabidopsis seedlings expressing VAP27-1^K59N-EGFP or VAP27-1^T60/61A-EGFP were obtained. Total proteins were extracted and separated by cell component. Compared with non-mutated VAP27-1-EGFP, VAP27-1^T60/61A-EGFP showed increased enrichment in the organelle membrane and nuclear components, indicating that its binding to the PM was weakened, which was not obviously observed in VAP27-1^K59N-EGFP ([97]Figure 2C). To assess the influence of the K59N, T60A, and T61A mutations on the distribution of VAP27-1 on the cell surface, we employed fast total internal reflection fluorescence (TIRF)-structured illumination microscopy (TIRF-SIM) at ultrahigh resolution to capture high-definition images of VAP27-1-EGFP, VAP27-1^K59N-EGFP, and VAP27-1^T60/61A-EGFP. The ImageJ lookup table “Green Fire Blue” was used to enhance visibility and image contrast. The results showed that native VAP27-1-EGFP was enriched at ER–PM contact sites, which mostly showed punctate green fluorescence signals ([98]Figure 2D). By contrast, the T60A and T61A mutations resulted in a loss of VAP27-1 localization to EPCSs, thus producing a relatively uniform distribution in the ER network. Although we did not observe a clear difference between the levels of VAP27-1^K59N-EGFP and native VAP27-1-EGFP in each cell component ([99]Figure 2C), the localization of VAP27-1^K59N-EGFP in the cell cortex remained similar to that of VAP27-1^T60/61A-EGFP ([100]Figure 2D). TIRF-SIM imaging of transgenic Arabidopsis seedlings expressing VAP27-1-EGFP, VAP27-1^K59N-EGFP, and VAP27-1^T60/61A-EGFP stained with FM4-64 revealed that the correlation between VAP27-1 and the PM was disrupted by the K59N and T60/61A point mutations ([101]Supplemental Figure 3). These results indicate that K59, T60, and T61 regulate the connection of VAP27-1 to the PM. VAP27-1 connection to the PM regulates plant resistance to ER stress in Arabidopsis Previous reports have shown that the three VAP27 members VAP27-1, VAP27-3, and VAP27-4 exhibit a highly similar domain composition and share sequence similarity. In addition, both VAP27-3 and VAP27-4 display a localization pattern similar to that observed for VAP27-1, which localizes to ER–PM contact puncta ([102]Wang et al., 2014, [103]2016). To explore the function of VAP27-1 and address the possibility of functional redundancy among the 10 Arabidopsis VAP homologs (VAP27-1–10), we used clustered regularly interspaced short palindromic repeat (CRISPR)/CRISPR-associated nuclease 9 (Cas9)-mediated genome editing to knock out the three VAP27 family members VAP27-1 (At3g60600), VAP27-3 (At2g45140), and VAP27-4 (At5g47180). For CRISPR-Cas9, we selected six 20-bp sequences with a protospacer adjacent motif at the 3′ end as single guide RNAs (sgRNAs) complementary to VAP27-1, VAP27-3, and VAP27-4. To reduce the risk of low editing efficiency, we used two sgRNA sequences for each gene ([104]Figure 3A). The first sgRNA was selected near the transcription start site to promote the introduction of frameshift mutations early in the coding sequence. Using a two-step assembly cloning strategy, we first inserted the sgRNA sequences targeting VAP27-1, VAP27-3, and VAP27-4 into sgRNA expression cassettes and then combined these cassettes with the Cas9 endonuclease coding sequence in a single binary vector, pYLCRISPR-Cas9P[ubi]-H ([105]Figure 3B). Transforming this construct into Arabidopsis via Agrobacterium tumefaciens–mediated transformation resulted in simultaneous expression of Cas9 and the six sgRNAs in the resulting transgenic plants. Figure 3. [106]Figure 3 [107]Open in a new tab Growth phenotypes of vap27-1/3/4 triple-mutant seedlings upon tunicamycin treatment. (A) Schematic illustrating the six single guide RNAs (sgRNAs; red lines) targeting the VAP27-1, VAP27-3, and VAP27-4 coding sequences. Blue boxes indicate exons; black lines indicate introns. The sgRNA sequences targeting VAP27-3 (T1 and T2), VAP27-1 (T3 and T4), and VAP27-4 (T5 and T6) are shown on the right. (B) Schematic of the single binary vector harboring the six sgRNA sequences and Cas9, which was used for Agrobacterium tumefaciens–mediated stable transformation of Arabidopsis. Three promoters from Arabidopsis (AtU3b, AtU3d, and AtU6-29) were used to drive the sgRNAs targeting VAP27-1, VAP27-3, and VAP27-4, respectively. The six sgRNA expression cassettes were sequentially ligated into the binary vector. The small colored boxes correspond to the target sequences according to the colors used in (A). (C) RT–qPCR analysis of relative VAP27-1 expression in WT, vap27-1/3/4, and four complementation lines shown in (D). Transcript levels were normalized to ACTIN2. Error bars represent mean ± SE. One-way ANOVA with Duncan’s post hoc test was performed. Different letters indicate significant differences at p < 0.05. (D) Phenotypes of 10-day-old WT, vap27-1/3/4, pVAP27-1::VAP27-1 (lines COM-6 and COM-8), pVAP27-1::VAP27-1^K59N, and pVAP27-1::VAP27-1^T60/61A complementation lines grown on ½ MS medium with 200 ng/ml TM; DMSO was used as a solvent control (CK). Scale bar, 2 cm. (E) Analysis of root length in the different genotypes, with or without TM treatment. Individual measurements are plotted as individual dots. Error bars represent mean ± SE (n ≥ 22). Data were analyzed by two-way ANOVA, followed by the LSD post hoc test. Different letters indicate significant differences at p < 0.05. (F) Six-day-old seedlings were grown vertically in ½ MS medium and transferred to a new culture plate containing 150 ng/ml TM. Root cells of WT and vap27-1/3/4 treated with 150 ng/ml TM for 36 h were stained with trypan blue. Scale bars, 500 μm. (G) Cell death severity was analyzed by quantifying the intensity of trypan blue staining, which was classified into three levels: faint (+), moderate (++), and strong (+++) (n ≥ 18). Examination of leaves from young T[1] seedlings revealed that VAP27-1, VAP27-3, and VAP27-4 were efficiently mutated at each desired target site. However, most of the mutations were chimeric. Therefore, we conducted detailed studies in the T[2] and T[3] generations to identify diverse possible mutation patterns. Several T[3] plants had mutations in both copies of VAP27-1, VAP27-3, and VAP27-4, including plants homozygous for the same alleles. All homozygous mutations were passed from the T[3] to the T[4] generation, including small insertions or deletions and even a long deletion from sgRNA T1 to sgRNA T2 in VAP27-3 ([108]Supplemental Figure 4). The homozygous vap27-1/3/4 triple mutant was used in subsequent experiments. To investigate whether VAP27-1-associated EPCSs contribute to ER-stress tolerance, we examined the tunicamycin (TM) susceptibility of the vap27-1/3/4 triple mutant. TM inhibits the N-glycosylation of proteins in the ER, thus inducing ER stress ([109]Bull and Thiede, 2012). TM treatment resulted in root growth defects in the wild-type (WT) seedlings, and these phenotypes were stronger in the vap27-1/3/4 triple mutant ([110]Figure 3D). In response to TM, vap27-1/3/4 root length was reduced to 58.0% ± 1.0% of that in the control treatment, whereas that of the WT was reduced to 70.0% ± 3.2%. These defects in vap27s were complemented in pVAP27-1::VAP27-1/vap27-1/3/4 (VAP27-1 COM) transgenic plants with two different levels of VAP27-1 expression ([111]Figure 3C). In VAP27-1 COM-6 and VAP27-1 COM-8, the root length upon TM treatment was reduced to 79.4% ± 2.5% and 75.4% ± 2.0% of that in CK, respectively. However, the expression of VAP27-1^K59N and VAP27-1^T60/61A, which led to impaired ER–PM connections, could not completely rescue the growth retardation exhibited by the vap27-1/3/4 mutant; TM treatment of these lines led to reductions in root length to 61.8% ± 1.8% and 54.0% ± 1.3%, respectively ([112]Figure 3C–3E). TM-triggered ER stress can induce cell death in rosette leaves and roots ([113]Mishiba et al., 2013). To evaluate root-cell viability in the vap27-1/3/4 triple mutant upon ER stress, we transferred 6-day-old WT and vap27-1/3/4 seedlings into medium containing 150 ng/ml TM for 36 h and assessed their root cells using trypan blue staining over the course of the TM treatment to determine the degree of cell death. The intensity of trypan blue staining was classified into three levels: faint, moderate, and strong. The cells of the vap27-1/3/4 mutant showed stronger trypan blue staining than those of the WT ([114]Figure 3F). Approximately 55.6% of WT seedlings showed faint staining (+), ∼44.4% showed moderate staining (++), and none showed strong staining. By contrast, ∼36.8% of vap27-1/3/4 seedlings showed strong staining. The proportions of moderate and faint staining in vap27-1/3/4 seedlings were 42.2% and 21.0%, respectively, indicating that the vap27-1/3/4 mutant had a higher degree of cell death than the WT in the root tip under ER stress ([115]Figure 3G). These observations indicate that the root growth defect of the vap27-1/3/4 mutant under TM-induced ER stress is partly due to loss of VAP27-1. The vap27-1/3/4 triple mutant displays enhanced UPR signaling upon ER stress ER stress induces the expression of numerous UPR-related genes ([116]Iwata and Koizumi, 2012). To gain insight into differences in gene expression patterns between the WT and the vap27-1/3/4 triple mutant under normal conditions and during ER stress, we treated WT seedlings and vap27-1/3/4 seedlings with dimethyl sulfoxide (DMSO) as a control or with 5 μg/ml TM for 3 h or 6 h to induce ER stress. We then performed RNA sequencing (RNA-seq) analysis to examine transcriptome changes in WT and vap27-1/3/4 seedlings before and after TM treatment. The samples containing vap27-1/3/4 seedlings were labeled “vap”, and control samples without TM treatment were labeled “CK”. Hundreds of unigenes were differentially expressed between each combination of two samples, including in vap CK vs. WT CK, vap + TM 3 h vs. WT + TM 3 h, and vap + TM 6 h vs. WT + TM 6 h ([117]Figure 4A). We performed Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analysis to determine the biological relevance of the upregulated genes. The KEGG analysis showed six upregulated biological pathways in vap CK vs. WT CK, including ribosome, protein processing in the ER, and plant–pathogen interaction, among others ([118]Figure 4B). Upregulation of protein processing in the ER, plant–pathogen interaction, and protein export pathways was observed in vap + TM 3 h vs. WT + TM 3 h ([119]Figure 4C). In addition, seven KEGG pathways, including ribosome, protein processing in the ER, and protein export, were simultaneously enriched in vap + TM 6 h vs. WT + TM 6 h ([120]Figure 4D). Notably, the protein processing in the ER pathway, which included multiple UPR-related genes, was enriched in each sample combination. The KEGG analysis also revealed 11 downregulated biological pathways in vap CK vs. WT CK, including plant hormone signal transduction, carbon metabolism, biosynthesis of amino acids, MAPK (mitogen-activated protein kinase) signaling pathway, and others. Downregulation of carbon metabolism, plant hormone signal transduction, biosynthesis of amino acids, and more pathways was also found in vap + TM 3 h vs. WT + TM 3 h. In addition, 10 KEGG pathways, including phenylpropanoid biosynthesis, carbon metabolism, biosynthesis of amino acids, and others, were simultaneously enriched in vap + TM 6 h vs. WT + TM 6 h. These pathways are mostly independent of the UPR ([121]Supplemental Figure 5). Figure 4. [122]Figure 4 [123]Open in a new tab Differences in gene expression between WT and the vap27-1/3/4 triple mutant before and after ER stress. (A) Venn diagram showing differentially expressed genes (DEGs) between vap27-1/3/4 and WT. The numbers represent upregulated or downregulated DEGs in three combinations: vap CK vs. WT CK, vap + TM 3 h vs. WT + TM 3 h, and vap + TM 6 h vs. WT + TM 6 h. (B–D) KEGG pathway enrichment analyses of upregulated genes in three combinations: vap CK vs. WT CK (B), vap + TM 3 h vs. WT + TM 3 h (C), and vap + TM 6 h vs. WT + TM 6 h (D). (E) Relative expression of UPR-related genes in vap27-1/3/4 and WT before and after TM treatment as measured by RNA-seq. All values at each time point represent the mean of three replicates. The biological functions of these genes are shown on the left. Relative gene expression values (log[2]) are shown, from the highest (red) to the lowest (blue). vap27-1/3/4 is abbreviated as “vap” in (A)–(E). (F) RT–qPCR analysis of selected UPR-related genes in vap27-1/3/4 and WT before and after TM treatment. Transcript levels were normalized to ACTIN2. Relative expression levels compared with the WT at 0 h are presented. Error bars represent mean ± SE. Data were analyzed by two-way ANOVA followed by the LSD post hoc test. Different letters indicate significant differences at p < 0.05. To narrow the focus of our analyses to genes related to the UPR, we performed a detailed expression analysis of 35 characterized UPR-related genes. Gene expression profiles showed that the mRNA levels of these genes increased significantly under TM treatment. Remarkably, most showed higher expression in the vap27-1/3/4 triple mutant than in WT, under both normal conditions and TM treatment ([124]Figure 4E). We subsequently used real-time quantitative PCR (RT–qPCR) to validate the expression profiles for 10 of the 35 UPR-related genes mentioned above (SUPPRESSOR OF SECRETION-DEFECTIVE 61 BETA [SEC61 BETA], HOMOLOG OF YEAST HRD1 [HRD1B], PDI-LIKE1-1 [PDIL1-1], AT3g60540, CALRETICULIN 1B [CRT1B], CO-FACTOR FOR NITRATE REDUCTASE AND XANTHINE DEHYDROGENASE1 [CNX1], BINDING PROTEIN3 [BIP3], ERDJ3B, SECRETION-ASSOCIATED RAS SUPER FAMILY 1 [SARA1A], and PDIL2-2). The gene expression trends were similar between the RNA-seq expression profiles ([125]Figure 4E) and the real-time PCR analysis ([126]Figure 4F). In summary, the vap27-1/3/4 triple mutant exhibits constitutive UPR signaling under normal conditions and stronger UPR signaling than the WT under ER stress. UPR intensity in the vap27-1/3/4 triple mutant is influenced by extracellular calcium influx In all eukaryotes, the Ca^2+ ion serves as a ubiquitous second messenger that triggers responses to abiotic and biotic stresses ([127]Kudla et al., 2018; [128]Dong et al., 2022; [129]Zeng et al., 2023). Upon the perception of a specific stimulus, the combined action of Ca^2+ influx and efflux mechanisms brings about a spatiotemporally defined increase in cytoplasmic Ca^2+ concentration. To explore Ca^2+ signaling and intracellular Ca^2+ levels in plant responses to ER stress, which is a universal abiotic stress, we visualized Ca^2+ in root-tip cells of WT and vap27-1/3/4 triple-mutant seedlings using Fluo4/AM. Compared with those of WT seedlings, the root-tip cells of the vap27-1/3/4 triple mutant showed a greater increase in intracellular Ca^2+ levels under ER stress triggered by a 6-h 5 μg/ml TM treatment ([130]Figure 5A and 5B). Figure 5. [131]Figure 5 [132]Open in a new tab Calcium signaling in Arabidopsis responses to ER stress. (A) Visualization of Ca^2+ levels after TM-induced ER stress. Fluo4/AM was used to stain Ca^2+ in root cells of 7-day-old WT and vap27-1/3/4 triple-mutant seedlings under control conditions (CK) or after 5 μg/ml TM treatment (+TM) for 6 h. Root tips were observed by confocal microscopy. An enlarged image of the region outlined in yellow is shown below. Pseudocolor images (blue-yellow-red palette) show the varied fluorescence intensity. Scale bars, 50 μm. (B) Quantification of average fluorescence intensities of the root tips observed in (A). Error bars represent mean ± SE. Statistical significance was determined using Student’s t-test (∗p < 0.05, ∗∗∗p < 0.001). (C) Mean Ca^2+ flux under normal conditions and following 5 μg/ml TM treatment for 6 h (n = 18). Error bars represent mean ± SE. Statistical significance was determined using Student’s t-test (∗p < 0.05). (D) RT–qPCR analysis of selected UPR-related genes in vap27-1/3/4 seedlings under DMSO, EGTA, TM, and combined EGTA + TM treatment. Error bars represent mean ± SE. Statistical significance was determined using Student’s t-test (∗∗∗p < 0.001). Next, we used NMT to measure Ca^2+ flux at the root meristem of WT and vap27-1/3/4 triple-mutant seedlings following 0 h and 6 h of 5 μg/ml TM treatment. There was a weak basal influx of Ca^2+ in both WT and triple-mutant roots under control conditions. However, compared with that in WT roots, the mean influx rate of extracellular Ca^2+ in the vap27-1/3/4 triple mutant increased significantly after a 6-h TM treatment ([133]Figure 5C). To determine whether this Ca^2+ influx was involved in UPR signaling triggered by ER stress in the vap27-1/3/4 triple mutant, we applied the calcium-selective chelator ethylenebis(oxyethylenenitrilo)tetraacetic acid (EGTA) to reduce extracellular Ca^2+ influx. We then examined the transcript levels of six classical UPR-related genes, SEC61 BETA, HRD1B, AT3G60540, ERDJ3B, and BIP3, in the vap27-1/3/4 triple mutant under DMSO, EGTA, TM, and EGTA + TM treatments. The results showed that treatment with 2.5 mM EGTA for 6 h did not trigger significant UPR-related gene expression in the vap27-1/3/4 triple mutant but did significantly inhibit the expression of the six UPR-related genes under ER stress triggered by TM treatment ([134]Figure 5D). These results suggest that extracellular Ca^2+ influx contributes to the regulation of UPR signal intensity in the mutant. Dynamic rearrangement of VAP27-1 regulates ER–PM connectivity in response to ER stress To explore how VAP27-1 functions in ER stress, we studied the response of VAP27-1 to ER stress. WT seedlings grown for 7 days were immersed in 5 μg/ml TM solution for 0, 2, 4, and 6 h to induce ER stress. We then assessed the transcript levels of VAP27-1 in response to ER stress by RT–qPCR and found that the transcript level of VAP27-1 increased slightly (about 14%) after 6 h of TM treatment ([135]Supplemental Figure 6A). In addition, immunoblotting analyses showed that the total VAP27-1-EGFP protein level did not change significantly with TM treatment ([136]Supplemental Figure 6B and 6C). On the basis of these results, we speculated that VAP27-1 likely responds to ER stress primarily through mechanisms unrelated to changes in its gene expression and protein accumulation. Therefore, our subsequent analyses focused on changes in the dynamic behavior of VAP27-1 at the PM and its subcellular localization and distribution during ER stress. We examined the velocities and diffusion coefficients of VAP27-1-EGFP at the PM under control and TM treatments. First, we treated the VAP27-1-EGFP transgenic seedlings with 0.1% DMSO as a control or with 5 μg/ml TM solution for 4 h to induce ER stress. Next, we performed VA-TIRFM to obtain real-time continuous imaging of single VAP27-1-EGFP particles on the PM surface for 15 s. We analyzed the collected time-series images using MATLAB software, focusing on the velocities and diffusion coefficients of VAP27-1-EGFP at the PM. In the control group, the distribution of velocities yielded a single population, and the Gaussian peak value (Ĝ) was 0.764 ± 0.010 μm/s. In the TM treatment, Ĝ decreased to 0.586 ± 0.017 μm/s, which was about 23.3% lower than that of the control ([137]Figure 6A). We then measured the diffusion coefficient of VAP27-1-EGFP at the PM. Under control conditions, the diffusion coefficient of VAP27-1-EGFP showed a bimodal distribution: larger diffusion coefficients (59.0%, Ĝ = 2.06 × 10^−2 μm^2/s) and smaller diffusion coefficients (41%, Ĝ = 2.01 × 10^−3 μm^2/s). TM treatment did not greatly alter the percentages of VAP27-1-EGFP that exhibited larger and smaller diffusion coefficients: the proportion of VAP27-1-EGFP that exhibited smaller diffusion coefficients increased from 41.0% to 42.5%, and the proportion that exhibited larger diffusion coefficients decreased from 59.0% to 57.5%. However, the Ĝ values for both groups fell after TM treatment, decreasing from 2.01 × 10^−3 μm^2/s to 0.44 × 10^−3 μm^2/s (a 78.1% reduction) for the smaller diffusion coefficients and from 2.06 × 10^−2 μm^2/s to 1.26 × 10^−2 μm^2/s (a 30.8% decrease) for the larger diffusion coefficients ([138]Figure 6B). Figure 6. [139]Figure 6 [140]Open in a new tab Morphology and quantification of ER and ER–PM connectivity in WT and vap27-1/3/4 under TM treatment. (A) Distribution of the particle velocity of VAP27-1-EGFP under control (CK) or 5 μg/ml TM treatment for 4 h. (B) Distribution of the diffusion coefficients of VAP27-1-EGFP under control or 5 μg/ml TM treatment for 4 h. (C) Effect of TM treatment on the localization of VAP27-1-EGFP at the PM. Before TIRF-SIM imaging, 7-day-old transgenic Arabidopsis seedlings were treated in liquid 0.1% DMSO (CK) or 5 μg/ml TM for 4 h. An enlarged image of the region outlined in yellow is displayed in the top right corner. The corresponding three-dimensional luminance plots are shown in the lower right corner. Scale bars, 12.5 μm. (D and E) Effect of TM treatments on the subcellular localization of VAP27-1-EGFP. Seven-day-old transgenic seedlings expressing VAP27-1-EGFP were treated in liquid 0.1% DMSO (CK) or 5 μg/ml TM for 4 h. Total proteins extracted from transgenic Arabidopsis seedlings were fractionated into plasma membrane (PM), organelle membrane (OM), cytosol (C), and nucleus (N). EGFP was immunoprecipitated using an anti-GFP antibody. Experiments were repeated independently at least three times. Protein expression was quantified by ImageJ measurement of band gray value. Relative protein expression levels under TM treatment compared to the CK are presented. Error bars represent mean ± SE. Statistical significance was determined using Student’s t-test (∗p < 0.05). (F) Effect of TM treatments on cortical localization of the EGFP-MAPPER EPCS marker stably expressed in either WT or vap27-1/3/4. Before TIRF-SIM imaging, 7-day-old seedlings were treated in liquid 0.1% DMSO (CK) or 5 μg/ml TM for 4 h. An enlarged image of the region outlined in yellow is presented in the top left corner. Three-dimensional luminance plots of the corresponding area are displayed in the top right corner. Scale bars, 25 μm. (G) Morphology of the ER in WT and vap27-1/3/4 before and after TM treatment observed by transmission electron microscopy. ER structures are labeled in yellow in the duplicate images. Scale bars, 1 μm. (H and I) Quantification of ER–PM connectivity expressed as the ratio of cER length/PM length (H) and distance from the ER to the PM (I) in WT and vap27-1/3/4 before and after TM treatment (n = 12). Error bars represent mean ± SE. One-way ANOVA with Duncan’s post hoc test was performed. Different letters indicate significant differences at p < 0.05. Besides the dynamic behavior of VAP27-1, we also observed expansion of VAP27-1-EGFP at the PM induced by TM treatment using TIRF-SIM ([141]Figure 6C). In parallel, we carried out biochemical experiments in which we analyzed VAP27-1-EGFP protein levels in total proteins extracted from transgenic seedlings expressing VAP27-1-EGFP following control and TM treatments. GFP-Lti6a was used as a control for successful membrane separation ([142]Figure 6D and 6E; [143]Supplemental Figure 7). Compared with that in the control treatment, VAP27-1-EGFP enrichment at the PM and nuclear membrane increased after the TM treatment, and VAP27-1-EGFP levels in the organelle membrane fraction decreased, indicating that its binding to the PM was enhanced by ER stress. To characterize the distribution of EPCSs on the cell surface and understand the influence of VAP27-1 on the redistribution of EPCSs upon ER stress, we used fast TIRF-SIM to image the EPCS marker EGFP-MAPPER under control and TM-treatment conditions in WT (EGFP-MAPPER/WT-4) and vap27-1/3/4 triple-mutant lines (EGFP-MAPPER/vap27-1/3/4-4), which exhibited similar EGFP-MAPPER transcription levels ([144]Supplemental Figure 8). Our results showed that EGFP-MAPPER puncta in the untreated vap27-1/3/4 line exhibited reduced size and intensity compared with those in the untreated WT. TM treatment had a broad effect on EPCS organization. Localization of the EGFP-MAPPER marker underwent a transition from an open beads-and-strings pattern to an expanded closed-reticule arrangement, accompanied by the emergence of some ER cisternae. Compared with the WT, the vap27-1/3/4 mutant exhibited a smaller magnitude of such alterations ([145]Figure 6F). To test whether the expansion of VAP27-1-associated EPCSs causes changes in the architecture of the ER, ER morphology and ER connectivity to the PM in root-tip cells were compared between the vap27-1/3/4 triple mutant and the WT under normal and ER-stress conditions. The morphology and distribution of the ER in seedling root-tip cells were observed by transmission electron microscopy after high-pressure freezing and freeze substitution (HPF-FS). To quantify ER architecture, we calculated the ratio of the total length of cortical ER segments to the length of the PM, as well as the closest distance of the ER to the PM, which reflect the abundance of ER membrane area and ER–PM connectivity to a certain extent. After image acquisition and analysis, we found that ER abundance in the vap27-1/3/4 triple mutant under normal conditions decreased to 43.6% of that in the WT and that the closest distance between the ER and PM increased by 36.2% compared with that in the WT. These data suggest defective ER morphology and reduced ER–PM connectivity in the vap27-1/3/4 triple mutant compared with the WT ([146]Figure 6G–6I). After TM treatment, the WT and vap27-1/3/4 triple mutant showed an increase in ER abundance and a decrease in the distance between the ER and the PM ([147]Figure 6G). However, ER abundance in the vap27-1/3/4 triple mutant decreased to 49.2% of that in the WT under TM treatment, and the closest distance between the ER and PM increased by 55.3% compared with that in the WT ([148]Figure 6G–6I). These results suggest that VAP27-1 promotes ER–PM connectivity under ER stress through subcellular relocalization and altered spatiotemporal dynamics at the PM. Discussion EPCS organization in mammalian cells has been described by a tethering signature model, suggesting that EPCSs are functionally diverse because of the existence of different tethering complexes ([149]Bayer et al., 2017). For example, E-Syt-EPCSs are involved in non-vesicular lipid transport, whereas STIM1-EPCSs are involved in store-operated Ca^2+ entry ([150]Fernandez-Busnadiego et al., 2015). A tethering arrangement model predicts that the differential molecular organization of a given tethering complex regulates EPCS structure and function ([151]Bayer et al., 2017). Although there is no STIM1 homolog in Arabidopsis, our results identified VAP27-1 as an ER–PM tether that is discontinuously distributed in the PM ([152]Figure 1). In addition, SYT1, a homolog of E-Syts, participates in a variety of biotic and abiotic stress responses as another ER–PM tether in Arabidopsis ([153]Schapire et al., 2008; [154]Lewis and Lazarowitz, 2010; [155]Sancho et al., 2015). Whole-mount immunocytochemistry and double immunogold labeling assays performed to examine the relationship between SYT1-enriched ER–PM binding sites (S-EPCSs) and VAP27-1-enriched ER–PM binding sites (V-EPCSs) revealed that SYT1 and VAP27-1 localized in different regions of the PM ([156]Siao et al., 2016). These results indicate that S-EPCSs and V-EPCSs represent independent ER–PM binding sites. The syt1 mutant possesses V-EPCSs, but these have altered stability, implying that although S-EPCSs and V-EPCSs localize to distinct regions, SYT1 is particularly important for V-EPCS stability ([157]Siao et al., 2016). Although it remains to be explored whether S-EPCS and V-EPCS functions overlap, these two types of EPCSs may perform different functions in Arabidopsis through a tethering signature model. Upon transient expression in N. benthamiana leaves, a VAP27-1-YFP (yellow fluorescent protein) fusion protein localized to the ER network and to immobile punctate structures. High VAP27-1 expression resulted in larger punctate structures with diameters of 0.25–1.2 μm ([158]Wang et al., 2014). When we observed transgenic Arabidopsis seedlings expressing C-terminally EGFP-tagged VAP27-1 driven by its native promoter, we found that VAP27-1-EGFP showed a beads-and-strings distribution in cotyledon epidermal cells. However, the bead-like puncta were much smaller than those observed in N. benthamiana ([159]Figure 1E and 1F). This size difference is probably due to differences between transient heterologous expression and stable expression of native proteins, indicating that the size of V-EPCSs can be affected by VAP27-1 expression level. A lysine residue at position 59 and two threonine residues at positions 60 and 61 in VAP27-1 stabilize the non-specific electrostatic interactions between its MSP domain and an FFAT-like motif in its interacting proteins ([160]Neefjes and Cabukusta, 2021). Our results showed that disruption of the MSP–FFAT interaction leads to defective VAP27-1 localization at the PM and even to the inability of VAP27-1 to function properly under ER stress ([161]Figures 2D, [162]3C, and 3E). We therefore suggest that VAP27-1 may directly link the ER and PM through interaction with unidentified PM proteins that contain an FFAT-like motif. Previous studies found that the spatiotemporal dynamics and stability of some PM proteins were regulated by the cell wall, membrane microdomains, and cytoskeleton ([163]Martiniere et al., 2012; [164]Lv et al., 2017). In the present study, we observed that fluorescent VAP27-1-EGFP spots showed restricted diffusion in a limited area, indicating that V-EPCSs are highly stable ([165]Figure 2A and 2B). This stability is most likely regulated by multiple factors. VAP27-1 co-localizes with microfilaments and microtubules, and microtubules limit the dynamics of VAP27-1 at the PM ([166]Wang et al., 2014). In addition, VAP27-1 tightly associates with the cell wall and moves more actively on the PM after enzymatic hydrolysis of the cell wall, indicating that the cell wall also limits the spatiotemporal dynamics of VAP27-1 at the PM ([167]Wang et al., 2016). Although we obtained several VAP27-1 T-DNA insertion lines from the Arabidopsis Biological Resource Center, we did not detect a reduction in VAP27-1 levels relative to the WT in any of them. Given the possible functional redundancy among VAP27 members, we generated the vap27-1/3/4 triple mutant by CRISPR-Cas9-mediated genome editing to enable us to investigate VAP27 function ([168]Figure 3A and 3B). Previous studies found that ER stress can affect plant growth and development, manifesting as smaller seedlings and shorter roots ([169]Watanabe and Lam, 2008; [170]Schott et al., 2010; [171]Kanehara et al., 2015). We found that roots of the vap27-1/3/4 triple mutant were shorter than those of the WT ([172]Figure 3C–3E), and cell death was more severe in the triple mutant than in the WT under ER stress ([173]Figure 3F and 3G). Notably, expression of VAP27-1^K59N and VAP27-1^T60/61A impaired the MSP–FFAT interaction and could not completely rescue the growth retardation of the vap27-1/3/4 triple mutant under ER stress, indicating that these three amino acid residues play a key role in regulating the VAP27-1–PM connection and plant resistance to ER stress. According to previous reports, ER stress results in the accumulation of misfolded and unfolded proteins in the ER. A variety of strategies exist to mitigate this stress, including the UPR, which ensures proper protein-folding fidelity, as well as mechanisms such as ER-associated degradation (ERAD) and ER-phagy, which are responsible for eliminating potentially harmful proteins ([174]Duan et al., 2023). In our study, we observed that the vap27-1/3/4 triple mutant displayed excessive UPR signaling. We also found increased expression of HRD1B, a conserved ERAD component in Arabidopsis, in the vap27-1/3/4 triple mutant, indicating an upregulation of ERAD in this mutant ([175]Su et al., 2011; [176]Chen et al., 2016; [177]Lin et al., 2019). Although the UPR and ERAD are mechanisms for restoring ER homeostasis, prolonged or irremediable ER stress can lead to hyperactivation of UPR signaling and the ERAD pathway. We therefore consider this excessive UPR signaling and enhancement of ERAD to be one of the main reasons for the impaired root growth observed in the triple mutant ([178]Yang et al., 2021). The much less abundant ER, as illustrated in [179]Figure 6G, could be another factor contributing to compromised root growth in the triple mutant. Recent studies have highlighted the essential role of VAP27-1 in endocytosis through its interaction with clathrin and lipids enriched in endocytic membranes ([180]Stefano et al., 2018). In addition, VAP27-1 acts as an essential component of the molecular architecture of the ER–autophagosomal membrane contact site, playing a role in the regulation of plant autophagy ([181]Ye et al., 2022). Thus, aberrant endocytosis and defective autophagy may also contribute to impaired root growth in the triple mutant. Further investigations are needed to explore these possibilities. Following the discovery of the ER-stress-sensitive phenotype in vap27-1/3/4, we focused on the strength of the UPR signal in the triple mutant during normal growth conditions and ER stress, as this may be the key cause of the root-length phenotype. Mutants of various genes show ER-stress-sensitive phenotypes ([182]Williams et al., 2010; [183]Chen and Brandizzi, 2012; [184]Mishiba et al., 2013; [185]Nawkar et al., 2017). Abnormal UPR signaling has been detected in such mutants, including weak, excessive, or chaotic signal strength, leading to decreased ER-stress resistance in Arabidopsis ([186]Chen and Brandizzi, 2012; [187]Nawkar et al., 2017; [188]Yu and Kanehara, 2020). RNA-seq analysis in previous studies showed that expression of UPR-related genes induced by ER stress in plant cells is usually associated with enhanced protein-folding activity, degradation of unfolded proteins, and the regulation of apoptosis ([189]Kamauchi et al., 2005). Our transcriptome data showed that upregulated differentially expressed genes in the vap27-1/3/4 triple mutant and the WT were mainly related to pathways associated with protein processing, export, and transport under ER stress ([190]Figure 4A–4D). We found that key ER-stress marker genes were upregulated in the vap27-1/3/4 triple mutant during the acute ER-stress response ([191]Figure 4E). These results, combined with the finding that cell death is more severe in vap27-1/3/4 than in the WT ([192]Figure 3F and 3G), led us to speculate that this cell death may be due to an overactive UPR. Such cell death due to overactive UPR signaling was also observed in agb1 mutants harboring mutations in a heterotrimeric G-protein β subunit. Although INOSITOL REQUIRING ENZYME1 (IRE1)-mediated upregulation of UPR-related gene expression in Arabidopsis is essential under ER stress, G-PROTEIN β SUBUNIT 1 (AGB1)-mediated downregulation of UPR-related gene expression is also crucial. Both pathways are independent but necessary, indicating that the intensity of the UPR signal must be accurately regulated to help plants survive instances of ER stress ([193]Chen and Brandizzi, 2012). Ca^2+ is a central second messenger that couples extracellular stimuli to characteristic intracellular responses and mediates a variety of biological processes ([194]Berridge et al., 2000; [195]Lee and Seo, 2021). Previous studies have found that the cytosolic Ca^2+ concentration ([Ca^2+][cyt]) increases rapidly upon exposure to various environmental stimuli, reaching micromolar levels ([196]Mehlmer et al., 2012; [197]Ma et al., 2019; [198]Dong et al., 2022). [Ca^2+][cyt] homeostasis under stress depends on precise regulation of Ca^2+ influx and efflux occurring at the PM and at the membranes of various subcellular compartments, including the ER ([199]Krebs et al., 2015; [200]Dong et al., 2022). Our results showed an increase in [Ca^2+][cyt] in root-tip cells after a 6-h TM treatment. Notably, the magnitude of this increase was significantly higher in the vap27-1/3/4 triple mutant than in the WT ([201]Figure 5A and 5B). The results of NMT assays revealed that intracellular calcium influx may partly contribute to the greater elevation of [Ca^2+][cyt] observed in the vap27-1/3/4 triple mutant ([202]Figure 5C). As a versatile second messenger, Ca^2+ is not only reciprocally connected with stress perception but also ensures downstream signal transduction ([203]Dong et al., 2022). Therefore, we speculated that the increased [Ca^2+][cyt] observed in the vap27-1/3/4 triple mutant was likely related to its excessive UPR signaling under ER stress relative to the WT. We subsequently found that disruption of intracellular calcium influx could inhibit UPR signaling in the vap27-1/3/4 triple mutant under ER stress ([204]Figure 5D). Whether VAP27-1 is an important ER–PM tether that interacts with Ca^2+ channels at the PM or ER, thereby participating in the regulation of intracellular calcium homeostasis, deserves further investigation. The tethering arrangement model proposes that the differential molecular organization of a given tethering complex at the PM regulates EPCS structure and function ([205]Bayer et al., 2017). Such rearrangement includes altered subcellular localization of EPCSs due to changes in the distribution of ER–PM tethers and variation in the three-dimensional nanostructure or dynamics of EPCSs due to molecular reorientation of ER–PM tethers. Evidence supporting the tethering arrangement model in mammalian cells was provided by the observation that perception of the depleted calcium pool in the ER by STIM1 triggers a change in its localization from a uniform distribution at the ER to an enrichment in EPCSs through molecular rearrangement, thus strengthening ER–PM connectivity and promoting the influx of Ca^2+ ions ([206]Idevall-Hagren et al., 2015). Salt stress induced by sodium chloride treatment in Arabidopsis also induces the expansion of SYT1-EPCS area, forming cisternae, which are likely to act as platforms for tight binding of the ER to the PM ([207]Lee et al., 2019). Here, our results showed that TM treatment had a broad effect on EPCS localization, which changed from an open beads-and-strings pattern to an expanded closed-reticule arrangement, accompanied by the appearance of some ER cisternae. VAP27-1 increased ER–PM connectivity under ER stress through subcellular relocalization and altered spatiotemporal dynamics at the PM ([208]Figure 6A–6F). This tethering arrangement model of V-EPCSs is similar to that of STIM1 and SYT1. In addition, the tethering arrangement model can also involve the dynamics and stability of EPCSs. The inter-membrane gaps and dynamics in both E-Syt-enriched EPCSs and STIM1-enriched EPCSs were controlled by the molecular rearrangement of these tethers in response to environmental signals (e.g., [Ca^2+]) ([209]Fernandez-Busnadiego et al., 2015). We also found that with the variation in subcellular distribution, the velocities and diffusion coefficients of VAP27-1-EGFP at the PM decreased under ER stress, indicating that the stability of V-EPCSs specifies their function as signal transduction platforms ([210]Figure 6A and 6B). In summary, EPCSs are emerging as membrane microdomains that are crucial for signal transduction and membrane remodeling ([211]Saheki and De Camilli, 2017). Understanding the adaptive response of dynamic cER arrangements to environmental stresses and the mechanisms by which EPCSs integrate signaling pathways and affect cell-fate decisions represent inspiring avenues for future research ([212]Bayer et al., 2017; [213]Wang et al., 2023). Our study suggests a regulatory model in which remodeling of Arabidopsis VAP27-1-associated EPCSs under ER stress is mediated by molecular rearrangement and dynamic changes in VAP27-1 at the PM, thereby enhancing plant resistance to ER stress ([214]Figure 7). In addition, we revealed a role of intracellular Ca^2+ homeostasis in ER-stress-induced UPR signaling. These results highlight the importance of further research on the functional relationships linking environmental stress and the specificity and plasticity of ER–PM contact to understand the cellular mechanisms that regulate plant responses to environmental stimuli. Figure 7. [215]Figure 7 [216]Open in a new tab Schematic model showing that VAP27-1 mediates ER–PM contact architecture and is required for cellular resistance to ER stress. Under normal conditions, VAP27-1-EGFP is localized to the PM and ER and is especially enriched in EPCSs. When Arabidopsis seedlings are subjected to ER stress, enrichment of VAP27-1-EGFP at the PM increases, while its presence in the ER correspondingly decreases. The PM-localized VAP27-1 contributes to the expansion of EPCSs. This relocalization of VAP27-1 and remodeling of EPCSs effectively enhance ER–PM connectivity. In the vap27-1/3/4 triple mutant, defective ER morphology and reduced ER–PM connectivity lead to a constitutive UPR signal under normal conditions, as well as a stronger UPR signal compared with the WT under ER stress. This excessive UPR signal observed in the vap27-1/3/4 triple mutant under ER stress depends on an increased influx of extracellular Ca^2+. Methods Plant material and growth conditions All Arabidopsis (Arabidopsis thaliana) WT, mutant, and transgenic plants in the current study were in the Columbia (Col-0) ecotype background. Half-strength Murashige and Skoog (½ MS) medium was used for plant culture. vap27-1/3/4 triple-mutant seeds were generated by CRISPR-Cas9. Seeds were surface sterilized and stratified at 4°C for 1–2 days in darkness. Plants were grown on ½ MS medium containing 1% sucrose under a 16-h light (23°C–25°C)/8-h dark (17°C–20°C) photoperiod. To test the sensitivity of seedlings to ER stress, various concentrations of TM dissolved in DMSO were added to ½ MS solid medium, and DMSO alone was used as the control (CK). Seeds were surface sterilized, vernalized at 4°C for 48 h, germinated on these media, and grown for 10 days. For root-length measurement, digital pictures of seedlings were obtained, and root length was measured using ImageJ software ([217]http://imagej.nih.gov/ij/). CRISPR-Cas9 target site selection and Cas9/sgRNA construct assembly The sequences of VAP27-1 (At3g60600), VAP27-3 (At2g45140), and VAP27-4 (At5g47180) were entered into an online tool for the design of target sgRNAs (targetDesign) from the web-based software package CRISPR-GE ([218]http://skl.scau.edu.cn/), which can identify CRISPR-Cas9 target sites within an input sequence ([219]Xie et al., 2017). Two available target sites were selected for each gene on the basis of their genomic locations, GC contents, and corresponding potential off-target sites. Corresponding target adaptors were designed for all six targets. The binary pYLCRIPSR-Cas9 multiplex genome-targeting vector system, which contained pYLCRISPR-Cas9P[ubi]-H carrying the CAS9 coding gene and three kinds of pYLsgRNA with sgRNA expression cassettes driven by AtU3b, AtU3d, and AtU6-29, respectively, was kindly provided by Professor Yaoguang Liu from South China Agricultural University. Six linearized sgRNA expression cassettes were obtained by ligation of the target adaptors to BsaI-digested pYLsgRNA followed by PCR. The recombined pYLCRISPR-Cas9P[ubi]-H construct carrying six sgRNA expression cassettes was generated using Golden Gate Cloning ([220]Ma and Liu, 2016). Gene editing and detection Genomic DNA extracted from Arabidopsis leaves was used as a template to amplify endogenous VAP27-1, VAP27-3, and VAP27-4 fragments by PCR. The primer pair VAP27-3-PCR-F: 5′-TGT CTT AGA TCC AAC TTC TTC CG-3′ and VAP27-3-PCR-R: 5′-ATC AGA AGC ATT CCC ATT ATC AGA G-3′ covered the region of target sites 1 and 2; the primer pair VAP27-1-PCR-F: 5′-TTG ATG TAG TTG AAT TGA AGA AGC A-3′ and VAP27-1-PCR-R: 5′-TTG ACG GTG ACA AAC TTG GAA GAA T-3′ covered the region of target sites 3 and 4; and the primer pair VAP27-4-PCR-F: 5′-CTT TGT CCA TCA CCA TTT GTT GC-3′ and VAP27-4-PCR-R: 5′-ATA AGC AGC TTC TTC CTC CCT C-3′ covered the region of target sites 5 and 6. PCR products were separated on a GoldView-stained agarose gel (1%), and the bands were recovered and identified by Sanger sequencing using sequencing primers for each target site: VAP27-3-T1-seq (5′-TTG TCT CTC TTT CTT TTC GTT ATT G-3′), VAP27-3-T2-seq (5′-TTT CTT TCT TTG GAT GTG TAG TTG A-3′), VAP27-1-T3-seq (5′-TTG ACG AAC AAG ACC GAC AAT AAT G-3′), VAP27-1-T4-seq (5′-ATG GAT GGT TAC AGT TTG GTG TTG G-3′), VAP27-4-T5-seq (5′-AAG CCT TTT CTC ATT CAC ATT TCA C-3′), and VAP27-4-T6-seq (5′-ATC TTT GGG TTG TAG ACG GGG TAA T-3′). All sequencing results were compared with the reference sequences of VAP27-1, VAP27-3, and VAP27-4 by alignment in DNAMAN (version 7.0). The allelic mutant sequences of target sites were determined using a web-based software tool (DSDecode) from CRISPR-GE ([221]Xie et al., 2017). Plasmid vector construction and transgenic plant production A 768-bp coding sequence (CDS) for VAP27-1 was amplified by PCR with the primers VAP27-1-F (5′-GCG GTA CCA TGA GTA ACA TCG ATC TGA TTG-3′) and VAP27-1-R (5′-GCT CTA GAT GTC CTC TTC ATA ATG TAT CCC AAA-3′). The PCR product was cut with KpnI/XbaI and cloned into the pCAMBIA1300-EGFP vector. Next, the PCR product of the 737-bp promoter sequence of VAP27-1, amplified with the primers pVAP27-1-F (5′-GCG AAT TCA TTT GTA ACT CTT TCT TCA CTG TAT-3′) and pVAP27-1-R (5′-GCG GTA CCC CCG ATC AGA TCG CCG GAG AT-3′), was digested with EcoRI/KpnI and inserted into the recombined pCAMBIA1300-EGFP vector containing the VAP27-1 CDS to generate the pVAP27-1::VAP27-1-EGFP construct. This construct was subsequently transformed into the WT background using the floral-dip method of A. tumefaciens–mediated transformation. For genomic complementation of the vap27-1/3/4 triple mutant, the PCR product of the CDS sequence for VAP27-1, VAP27-1^K59N, or VAP27-1^T60/61A was cloned into a modified pCAMBIA2300 vector containing the sequence of the VAP27-1 promoter. The recombined pVAP27-1::VAP27-1/VAP27-1^K59N/VAP27-1^T60/61A construct was subsequently transformed into the vap27-1/3/4 background by A. tumefaciens–mediated transformation. To generate GFP-MAPPER in the WT or vap27-1/3/4 background, we obtained a previously reported MAPPER sequence ([222]Chang et al., 2013) through gene synthesis and inserted it into a pBI121-EGFP vector. This construct was transformed into the WT or vap27-1/3/4 background by A. tumefaciens–mediated transformation using the floral-dip method. Cell death detection For cell death detection, 6-day-old seedlings grown on ½ MS solid medium were transferred onto ½ MS solid medium containing 150 ng/ml TM for 36 h. The roots of at least 18 seedlings per genotype were stained with 10 mg/ml trypan blue for 10 min. After several washes with distilled water, staining was observed by light microscopy. Only cells stained blue with trypan blue were counted as dead. The extent of cell death was calculated as the ratio of blue-stained area to root-tip area using ImageJ and was classified as faint (≤20%), moderate (20%–40%), or strong (≥40%). RNA-seq and data analysis Three biological replicates of WT and vap27-1/3/4 triple-mutant seeds were germinated on ½ MS solid medium and grown for up to 7 days. The seedlings were then placed in ½ MS liquid medium with 5 μg/ml TM for 3 or 6 h. Medium with 0.1% DMSO was used as the control. A portion of the seedlings were retained for subsequent RT–qPCR, and the remainder were quick-frozen in liquid nitrogen. RNA extraction, transcriptome sequencing, and data analysis were performed by Beijing Novogene Bioinformatics Technology. The RNA-seq data were deposited at the National Center for Biotechnology Information under BioProject accession number PRJNA1054668. RNA extraction and RT–qPCR For RT–qPCR, 1 μg of total RNA extracted with the RNAprep Pure Plant Kit (Tiangen Biotech, Beijing, China) was used for first-strand cDNA synthesis with One-Step Genomic DNA Removal and cDNA Synthesis SuperMix (Transgene Biotech). The reaction components and thermal cycling conditions were set according to the manual. The real-time system from Bio-Rad (USA) was used to perform the reactions. Each experiment was replicated four times. The RT–qPCR primers are listed in [223]Supplemental Table 1. Fluorescence staining and imaging of cell nuclei Cell nuclei in cotyledon epidermal cells were stained with DAPI at 0.5 μg/ml in 0.1% (v/v) Triton X-100 for 20 min. After three washes with ½ MS liquid medium, DAPI-stained nuclei were observed by laser scanning confocal microscopy (Leica SP8; excitation = 390 nm; emission = 460 nm). Laser scanning confocal microscopy and image analysis Arabidopsis seedlings were imaged with a confocal microscope (Olympus FluoView 1200) fitted with a water-immersion objective (Olympus UPLSAPO 60XW). mCherry and EGFP/FM4-64 were excited with 559-nm and 488-nm wavelengths, respectively, and fluorescence emission spectra were detected at 520–550 nm for EGFP and 560–640 nm for FM4-64/mCherry. Images were analyzed with the Olympus FV10-ASW software package and ImageJ software (NIH). Protein extraction and western blotting analysis For total protein extraction, 10-day-old seedlings of transgenic VAP27-1-EGFP lines were soaked in 5 μg/ml TM for 0, 2, 4, or 6 h. The treated seedlings were ground into a fine powder in liquid nitrogen and denatured in extraction buffer (125 mM Tris–HCl [pH 8.8], 1% SDS, 10% glycerine, and 50 mM Na[2]S[2]O[5]). For isolation of membrane proteins and proteins from different cell fractions, we used the Minute Plant Plasma Membrane Protein Isolation Kit (Invent Biotechnologies). Proteins were separated on a 10% SDS–polyacrylamide gel and transferred to a nitrocellulose membrane. A GFP antibody (Abcam, UK; 1:10 000 dilution) was used to detect the VAP27-1-EGFP fusion protein. Intracellular calcium visualization For intracellular calcium visualization of the root tips, 7-day-old Arabidopsis seedlings were pretreated with ½ MS liquid medium containing 0.1% DMSO or 5 μg/ml TM for 3 h. They were further incubated with the addition of 20 μM Fluo4/AM (Invitrogen, [224]F14201) and 0.02% (m/v) Pluronic F-127 at 4°C for 2 h, followed by a 1-h incubation at room temperature in darkness. After a wash with ½ MS liquid medium, Ca^2+ levels in seedling root tips were determined under a laser scanning confocal microscope (Leica TCS SP8) with excitation and emission wavelengths of 494 nm and 506–546 nm, respectively. ImageJ was used to subtract the background signal and quantify the mean fluorescence intensity. Net Ca^2+ flux was measured using NMT100-IM-XY (Xuyue, Beijing) as described previously ([225]Wang et al., 2019). Before testing, 7-day-old Arabidopsis seedlings were pretreated in 5 μg/ml TM solution for 0, 3, or 6 h. The Ca^2+ flux microsensor was then calibrated with low (0.05 mm CaCl[2], 0.05 mm KCl, pH 6.0) and high (0.5 mm CaCl[2], 0.5 mm KCl, pH 5.0) concentration correctors. Interference from TM and its solvent DMSO was eliminated. Seedlings were then transferred to the test solution (0.1 mM CaCl[2], 0.2 mM 2-(N-morpholino)ethanesulfonic acid, pH 6.0), with their roots fixed to keep them steady for 15 min. Net Ca^2+ flux of root-tip meristematic cells was then measured for 5 min. Net Ca^2+ flux was directly read by ImFluxes v2.0 (Younger USA, Amherst, MA, USA) in picomol cm^−2 s^−1. Mean efflux was indicated by positive values and mean influx by negative values. VA-TIRFM and single-particle fluorescence imaging analysis VA-TIRFM imaging was performed using an Olympus IX83 inverted microscope equipped with a total internal reflective fluorescence illuminator and a 100×/1.45 NA oil-immersion objective (Olympus) and a back-illuminated EMCCD camera (Andor iXon DV8897D-CS0-VP, Andor Technology, Belfast, UK) after filtering with two band-pass filters (525/545 nm and 609/654 nm). Each series of time-lapse images of single VAP27-1-EPCS particles was acquired with a 150-ms exposure time. EGFP signals were excited with 473-nm laser lines, and the fluorescence emission was collected with a BA510IF filter. Living leaf epidermal cells of 7-day-old seedlings were used for the analyses. The single-particle tracking analysis was performed as described previously ([226]Wang et al., 2015; [227]Cui et al., 2018a, [228]2018b). Sample preparation and observation for transmission electron microscopy The sample preparation process, including HPF-FS, embedding, and ultrathin sectioning, was completed on the platform of the National Protein Science Research (Beijing) Facility Tsinghua Base. Before testing, 10-day-old Arabidopsis seedlings were pretreated in 5 μg/ml TM solution for 0 h or 3 h. The Arabidopsis root tips were cut into lengths of 1–2 mm, placed on a sample carrier filled with 0.15 M sucrose, and frozen in a high-pressure freezer (Leica EM HPM-100). The frozen samples were transferred to liquid nitrogen–frozen substitution medium (0.5% glutaraldehyde and 0.1% uranyl acetate in anhydrous acetone) and placed in an Automatic Freeze Substitution System (Leica EM AFS2). The AFS2 was programmed to increase in temperature from −140°C to −80°C at a rate of 5°C/h, kept at −80°C for 3 days, then warmed to −20°C at a rate of 5°C/h. The samples were washed twice with pure acetone on ice for 20 min each, removed from acetone, and placed into new acetone. The samples were then embedded in Eponate 12 resin, and the blocks were trimmed and sliced. Leica UC7 and FC7 microtomes were used for ultrathin sectioning. Observations and photographs were obtained using a Hitachi HT7700 transmission electron microscope at the State Key Laboratory of Forest Genetics and Tree Breeding, Chinese Academy of Forestry. TIRF-SIM imaging TIRF-SIM was performed on 7-day-old Arabidopsis seedlings at physiological temperature. TIRF-SIM imaging was performed with a DeltaVision OMX SR imaging system equipped with a 60×/1.42 NA oil-immersion objective and 405-nm, 488-nm, 568-nm, and 640-nm diode lasers from Toptica. EGFP was excited at 488 nm, and fluorescence emission spectra were detected at 520–550 nm. Image acquisition was controlled by AcquireSR with the following parameters: TIRF-SIM for light path, sequential for image mode, 1024 × 1024 for image size, 1 × 1 for binning, and 272 MHz for readout speed. SI images were reconstructed by OMS SI Reconstruction. Quantification of ER–PM connectivity ER–PM connectivity was quantified by calculating the ratio of cortical ER length to PM length. The closest distance from the ER to the PM was also measured. Funding This work was supported by the National Natural Science Foundation of China (32170689, 91954202, 32030010), National Key Research and Development Program of China (2022YFF0712500), the Program of Introducing Talents of Discipline to Universities (111 Project, B13007), and Beijing Forestry University Outstanding Postgraduate Mentoring Team Building (YJSY-DSTD2022005). Author contributions Y.J. and J.L. directed the research. Y.M., Y.Z., and L.C. performed most of the experiments. J.Z. and Y.B. contributed to transcriptome sequencing and data analysis. X.Z., X.L., and Y.L. provided valuable suggestions on the research and manuscript. Y.M. wrote the manuscript. Y.Z., L.C., and Y.J. revised the manuscript. Acknowledgments