Calcium-dependent protein kinase 16 phosphorylates and activates the aquaporin PIP2;2 to regulate reversible flower opening in Gentiana scabra

Abstract

Flower opening is important for successful pollination in many plant species, and some species repeatedly open and close their flowers. This is thought to be due to turgor pressure changes caused by water influx/efflux, which depends on osmotic oscillations in the cells. In some ornamental plants, water-transporting aquaporins, also known as plasma membrane intrinsic proteins (PIPs), may play an important role in flower opening. However, the molecular mechanism(s) involved in corolla movement are largely unknown. Gentian (Gentiana spp.) flowers undergo reversible movement in response to temperature and light stimuli; using gentian as a model, we showed that the Gentiana scabra aquaporins GsPIP2;2 and GsPIP2;7 regulate repeated flower opening. In particular, phosphorylation of a C-terminal serine residue of GsPIP2;2 is important for its transport activity and relates closely to the flower re-opening rate. Furthermore, GsPIP2;2 is phosphorylated and activated by the calcium (Ca²⁺)-dependent protein kinase GsCPK16, which is activated by elevated cytosolic Ca²⁺ levels in response to temperature and light stimuli. We propose that GsCPK16-dependent phosphorylation and activation of GsPIP2;2 regulate gentian flower re-opening, with stimulus-induced Ca²⁺ signals acting as triggers.

Gentian flower re-opening is regulated by GsCPK16-dependent GsPIP2;2 phosphorylation and activation, with stimulus-induced calcium signals acting as triggers.

Introduction

In many species, flowers sense environmental conditions and are open only when conditions are suitable for reproduction (van Doorn and Van Meeteren, 2003; van Doorn and Kamdee, 2014). Although corolla opening appears to be an irreversible movement in some plant species, many species having bowl- and disc-shaped flowers exhibit reversible opening and closing movements in response to light and temperature (Claus, 1926; Stirton, 1983). van Doorn and Kamdee summarized the mechanisms of corolla movements and concluded that the direct driving force in flower opening is water transport into the petal cells and the subsequent changes in cell volume (van Doorn and Kamdee, 2014). These findings suggest that many flowers can control reversible corolla movements by appropriately regulating water transport in response to environmental stimuli; however, the underlying regulatory mechanisms of corolla movement are largely unknown.

Aquaporins, also known as plasma membrane intrinsic proteins (PIPs), are ubiquitous water- and/or small-solute-channel proteins (Abascal et al., 2014). In vascular plants, aquaporins are classified into at least five groups based on their structural characteristics (Deshmukh et al., 2015; Maurel et al., 2015). The PIP2 subfamily of aquaporins is abundant in the plasma membrane (PM) and play a central role in controlling water transport through the PM (Fetter et al., 2004; Sakurai et al., 2005, 2008). Many previous studies have suggested that PIP2s are involved in numerous water-related physiological processes (Afzal et al., 2016; Wang et al., 2020).

PIP2 function is regulated by various posttranslational modifications (Santoni, 2017; Nesverova and Törnroth-Horsefield, 2019). In many plant species, phospho-proteomic approaches have revealed phosphorylation of the penultimate and ultimate conserved serine residues in the PIP2 C-terminal tail (Wilder et al., 2008; Prak et al., 2008; Zhang et al., 2019). Biochemical and structural studies have proposed that phosphorylation of the penultimate conserved serine residue is important for conformational changes leading to channel opening and for interaction with regulatory proteins (Johansson et al., 1996; Törnroth-Horsefield et al., 2006; Nyblom et al., 2009; Prado et al., 2013, 2019). On the other hand, it has been suggested that phosphorylation of the ultimate conserved serine residue plays a vital role in controlling membrane trafficking (Prak et al., 2008). In addition, some studies have shown that the PIP2 phosphorylation state changes in response to environmental stimuli (Prak et al., 2008; Prado et al., 2013; Zhang et al., 2019), such as light and dehydration stress, and that PIP2 phosphorylation occurs in a calcium (Ca²⁺)-dependent manner (Johansson et al., 1996; Wilder et al., 2008). Therefore, plants may regulate water-related physiological events through phosphorylation-dependent PIP2 regulation via Ca²⁺ signals induced by external stimuli. However, the responsible protein kinase(s) has not yet been identified.

Although evidence supporting the role of aquaporins in corolla movement is limited, studies focusing on temperature-dependent reversible movements in tulip (Tulipa gesnerina) flowers have shown that the levels of Ser-phosphorylated putative aquaporin protein(s) increase during flower re-opening, and Ca²⁺ chelators inhibit their phosphorylation and flower re-opening (Azad et al., 2004). Based on this finding, Ca²⁺-dependent phosphorylation and PIP2 activation are considered important factors in achieving flower re-opening. However, there is no direct evidence that aquaporins are involved in the corolla movement.

Gentian (Gentiana spp.) is one of the first species to demonstrate repeated, reversible flower opening and closing movements that do not involve cell growth (Claus, 1926). This phenomenon protects pollen and ovules from rain and low temperatures (Bynum and Smith, 2001; Mu et al., 2010). Thus, in this study, we focused on reversible corolla movement, especially flower re-opening, and investigated its molecular mechanism using the Japanese Gentiana scabra breeding line “Bzc-1” as a model (Takahashi et al., 2020). Here, we showed that GsPIP2;2 and GsPIP2;7 directly regulate flower re-opening. In addition, we identified and characterized a Ca²⁺-dependent protein kinase, GsCPK16, responsible for phosphorylating and activating GsPIP2;2; we demonstrated that GsCPK16 is activated by light-induced increases in cytosolic Ca²⁺ levels and induces GsPIP2;2 phosphorylation, leading to flower re-opening. We provide evidence that flower re-opening is triggered by PIP2-mediated water transport and that PIP2 activity is regulated by CPK-dependent phosphorylation.

Results

Japanese gentian flowers move reversibly in response to temperature and light stimuli

Gentian flowers open and close within 30–60 min in response to temperature changes and light in the field, and these movements can also be reproduced in the laboratory (Bynum and Smith, 2001; Mu et al., 2010). Therefore, gentian flowers offer a good platform for elucidating reversible corolla movement mechanisms. Thus, we focused on the Gentian scabra breeding line “Bzc-1” (Takahashi et al., 2020).

Reversible corolla movements in Gentian sp. have only been reported in those native to Europe (Claus, 1926; Bynum and Smith, 2001; Mu et al., 2010). Accordingly, we investigated whether corolla movements were reversible in the gentian “Bzc-1.” In all experiments, we used the vegetatively propagated in vitro cultures of gentian “Bzc-1” to eliminate the effects of environmental and growth conditions. The gentian “Bzc-1” flowers closed within 30 min when exposed to cool temperatures (16°C; Figure 1, A and B). The closed flowers re-opened within 30 min after being exposed to normal growth temperature (22°C) and light, but re-opening was delayed under dark conditions (Figure 1, A and B). Light stimulation under cool-temperature stress conditions not only did not inhibit flower closure but also did not promote opening (Figure 1, C and D). Epidermal cell area measurements showed that adaxial epidermal cells changed in size in relation to corolla movement, but those on the abaxial side did not (Figure 1E). These results indicate that gentian reversible corolla movements are associated with temperature-induced reversible expansion and contraction of adaxial cells and that light stimulates re-opening. However, abaxial cell expansion was not observed, and we further examined the morphological features of the corolla. Analysis of the confocal microscopy image data revealed that the cell wall on the abaxial side contained high amounts of phenolic compounds, suggesting that differential cell volume changes may be due to differences in the composition and structure of the cell wall between the adaxial and abaxial sides (Figure 1F).

Figure 1.

Reversible opening and closing of gentian flowers. A–D, Gentiana corolla movements in response to temperature and light. Fully open flowers were incubated at 16°C for 30 min in the dark and then incubated at 22°C for 60 min in the light. Then, after incubating the flowers for at least 1 h to induce full opening, flowers were again incubated at 16°C and subsequently incubated at 22°C for 60 min in the dark. The corolla angle was measured every 30 min. Morphology (A), floral opening–closure angles (B). Flowers were incubated at 16°C for 30 min in the dark or light (C). After incubation at 16°C for 30 min in the dark, flowers were incubated at 16°C or 22°C in the light (D). Flower opening and closing movements were repeatedly investigated, and similar results were obtained in all tests. For biological replicates, flowers were collected and analyzed from four to five independent plants (n = 4–5). The data represent means ± sd from four to five biological replicates (n = 4–5). Data were analyzed using one-way ANOVA repeated measures with Tukey’s multiple comparisons test with a significance level of 5% (P < 0.05) (B) and Brown–Forsythe and Welch’s ANOVA test with Dunnett’s T3 multiple comparisons test with a significance level of 5% (P < 0.05) (C and D). E, Adaxial and abaxial surface of the corolla and their average cell area. The corolla was vertically cut into four to five segments, including the corolla robe, and then floated in the water. Samples were incubated under the same conditions as in Figure 1A. The epidermal cells on the adaxial and abaxial sides were observed, and the areas of epidermal cells were measured. The data represent the average ± sd of 10 cells from two independent experiments (n = 10). The results are from a Friedman’s test, followed by Dunn’s multiple comparisons test with a significance level of 5% (P < 0.05) (adaxial), and paired t test (abaxial). NS, not significant. F, Confocal laser scanning microscope image of the cross-sections of gentian corolla. Phenolic compounds and cellulose were visualized by autofluorescence (Ex 488 nm/Em 510–550 nm) or Direc Red 23 stain (Ex 543 nm/Em 600–660 nm). DIC, differential interference contrast. Scar bar, 100 μm. Gray values of autofluorescence and DirecRed23 fluorescence signals measured along a straight line in the upperpart. The intensity of the fluorescent signal was measured using ImageJ software.

Isolation of G. scabra homologs of Arabidopsis thaliana PIP2

To investigate the relationship between changes in cell volume and PIP2-mediated water transport in reversible corolla movements, we identified cDNAs encoding the PIP2 homologs in G. scabra. However, since the genome sequence of G. scabra has not been sequenced, we attempted to identify the AtPIP2 homolog using transcriptome assembly. We performed RNA sequencing (RNA-seq) on the corolla of gentian “Bzc-1” flowers (Supplemental Table S1). The RNA-seq reads were assembled de novo using Trinity (Grabherr et al., 2011), and a total of 119,167 (>200 bp) transcripts were detected. The coding regions of the assembled transcripts were predicted using TransDecoder (Haas et al., 2013). Subsequently, we used Cluster Database at High Identity with Tolerance (CD-HIT) (Fu et al., 2012) to reduce transcript redundancy and produce unique genes. Finally, 24,611 transcripts were detected. BUSCO analysis was performed to evaluate the completeness of the assembly against a dataset set of 1,614 core genes in embryophyta (Simão et al., 2015), resulting in 89% completeness in the 24,611 transcripts set (Supplemental Table S1).

The transcripts were annotated using the BLASTX search for the A. thaliana protein database (https://www.arabidopsis.org/). Gene expression levels were calculated from the fragments per kilobase of exon per million reads mapped (FPKM) values using the RSEM software within the Trinity package. These results revealed that homologs of Arabidopsis AtPIP2;1, AtPIP2;2, AtPIP2;5, and AtPIP2;7, were present and expressed in G. scabra (Supplemental Table S2). The two genes, homologous to Arabidopsis AtPIP2;2 and AtPIP2;7, were highly expressed, suggesting that these two genes encode the main water gateways in gentian corolla. Thus, we named them GsPIP2;2 and GsPIP2;7, respectively.

GsPIP2;1 and GsPIP2;7 gene silencing suppresses flower re-opening

To assess the role of GsPIP2;2 and GsPIP2;7 in gentian corolla movements, we produced GsPIP2;2 and GsPIP2;7 knockdown lines using a virus-induced gene silencing (VIGS) approach using Apple latent spherical virus (ALSV) silencing vectors (Supplemental Figure S1; Nakatsuka et al., 2015). Compared to the vector control lines, silencing of GsPIP2;2 or GsPIP2;7 resulted in the suppression of flower opening and significantly delayed flower re-opening after closure, but did not affect the flower closure process (Figure 2, A and B).

Figure 2.

Silencing of gentian GsPIP2;2 and GsPIP2;7 in gentian suppresses flower re-opening movement. A, Quantitative RT-PCR of GsPIP2;2 and GsPIP2;7 transcripts in corolla of vector control, ALSV:::GsPIP2;2- and ALSV:::GsPIP2;7-infected plants 30 days after infection. For biological replicates, total RNA was extracted and analyzed from the corollas of six independent infected plants (n = 6). The data represent means ± sd from six biological replicates (n = 6). The results are from a Brown–Forsythe and Welch’s ANOVA test, followed by Dunnett’s T3 multiple comparisons test with a significance level of 5% (P < 0.05). B, Corolla movements in response to temperature and light of vector control and ALSV:::GsPIP2;2- and ALSV:::GsPIP2;7-infected plants (∼60 days after infection). Morphology (upper), floral opening–closure angles (lower). The data represent the means ± sd of biological replicates obtained from six independent infected plants (n = 6). Data were analyzed using a repeated measures two-way ANOVA with Dunnett’s multiple comparisons test to compare the difference between vector control and ALSV-infected plants. Asterisks indicate statistically significant (**P < 0.01, ***P < 0.001) and ns, not significant. Scale bar = 1 cm. C, BiFC analysis of GsPIP2;2–GsPIP2;7 interaction in corolla protoplast. The N-terminal half of YFP fused GsPIP2;2 (GsPIP2;2-nYFP) and the C-terminal half of YFP fused GsPIP2;7 (GsPIP2;7-cYFP) were co-expressed in protoplasts. nYFP-GARU and AtGID1A-nYFP were used for negative control. Fluor, YFP fluorescence image; FM4–64 stain fluorescence image; Merge, merged image of Fluor and FM4–64 images; Scale bar = 10 μm.

In addition, we analyzed the interaction between GsPIP2;2 and GsPIP2;7 by bimolecular fluorescence complementation (BiFC) to investigate the functional redundancy of these genes. BiFC analysis showed yellow fluorescent protein (YFP) fluorescence on the PM (Figure 2C), indicating that GsPIP2;2 and GsPIP2;7 interacted on the PM. These results suggest that GsPIP2;2 and GsPIP2;7 coordinately regulate flower re-opening.

Flower re-opening requires Ca²⁺ ion-dependent GsPIP2 phosphorylation

Previous studies have shown that PIP2 activity is regulated by phosphorylation of the penultimate conserved serine residue in the C-terminal tail (Johansson et al., 1998; Qing et al., 2016). Multiple sequence alignments with well-characterized Arabidopsis PIP2s revealed that conserved Ser residues were also found in the C-terminal region of gentian GsPIP2;2 and GsPIP2;7 (Supplemental Figure S2A). Primary antibodies recognizing the phosphorylation of the penultimate conserved serine residue in the C-terminal tails of PIP2s are commercially unavailable; therefore, the anti-phospho Ser280 antibody, which was produced by co-author Dr. Shiratake to analyze the function of aquaporins in pear (Pyrus communis) fruits, was applied to gentian. To analyze endogenous GsPIP2 protein(s), we used a commercially available anti-PIP2 primary antibody. The specificity of the antibody was evaluated using gentian GsPIP2;1, 2;2, 2;5, and 2;7 proteins synthesized in a wheat cell-free translation system. A Phos-tag sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS–PAGE) assay revealed that GsPIP2;1, 2;2, and 2;7 were slightly phosphorylated by endogenous, unknown protein kinase(s) in the wheat germ extract (Supplemental Figure S2B). However, GsPIP2;5 was not phosphorylated (Supplemental Figure S2B). The anti-phospho Ser280 antibody specifically recognized phosphorylated GsPIP2;2 and slightly recognized GsPIP2;1 (Supplemental Figure S2B), and commercially available antibodies specifically recognized only the GsPIP2;2 protein (Supplemental Figure S2B). Moreover, we performed additional evaluations using GsPIP2;2 and corolla lysate as antigens, which confirmed that pretreatment of the antigen with lambda phosphatase reduced the signal derived from the anti-phospho Ser280 antibody below the detection limit. Furthermore, anti-phospho Ser280 antibody and anti-PIP2 primary antibody reactivity were competitively inhibited by phospho-peptides or nonphospho-peptides derived from the C-terminal tails of PcPIP2;2 and GsPIP2;2, respectively. (Supplemental Figure S2, C–H). Immunoblot analysis using ALSV-infected plants shown in Figure 2 revealed that band signals were reduced in GsPIP2;2 knockdown plants compared to GsPIP2;7 knockdown plants and control plants (Supplemental Figure S2I). To summarize these analyses, anti-phospho Ser280 antibody can detect mainly phosphorylation of the penultimate conserved serine residue in the C-terminal tail of endogenous GsPIP2;2 in gentian corolla.

Immunoblot analysis revealed that during flower development, Ser phosphorylation of GsPIP2;2 was detected only at developmental stage 4 (flower opening; Figure 3A). Furthermore, immunoblot analysis revealed that the Ser-phosphorylation level of GsPIP2;2 increased when a cool-temperature treatment was used to induce flower closing (dark at 16°C in Figure 3B; Supplemental Figure S3, A and B) but decreased rapidly following a return to normal growth temperature in the dark and light at 22°C for 15 min (Figure 3B). However, their phosphorylation levels increased at 30 min and maximized at 60 min after applying the normal growth temperature and light stimulation (light at 22°C for 30 and 60 min in Figure 3B; Supplemental Figure S3B).

Figure 3.

Flower re-opening movement is associated with Ca²⁺-dependent GsPIP2;2 phosphorylation. A, Immunoblot analysis of phosphorylated GsPIP2;2 at the flower development stage. Flower development stages were divided into four (left), and phosphorylated endogenous GsPIP2;2 (P-GsPIP2;2) were analyzed (right). B, Immunoblot analysis of phosphorylated GsPIP2;2 in gentian corolla. Each corolla was cut vertically into eight segments, including the corolla robe, and then floated in the water. Samples were incubated at 16°C for 30 min in the dark and then incubated at 22°C for 15–60 min in the light or dark. C, The relative phosphorylation level of GsPIP2;2 with phosphorylation at Ser280 and Ser283. Full-open flowers were incubated at 16°C for 30 min in the dark and then incubated at 22°C for 60 min in the light or dark. Membrane proteins were extracted for nano-LC/MS/MS analysis. The peak area of phosphopeptide was quantified and normalized based on the signal intensities of GsPIP2;2 in the immunoblot image. Relative phosphorylation level was calculated as the signal intensity ratio of the average value of Mock. The data represent the means ±  sd of biological replicates obtained from three independent corollas (n = 3). Significant differences (*P < 0.05; ns, not significant) were determined by Welch’s t test. D, Immunoblot analysis of phosphorylated GsPIP2;2 in the corolla of adaxial and abaxial sides. After incubation at 16°C for 30 min in the dark, flowers were incubated at 22°C for 60 min in the light or dark. The epidermal cell layer on the adaxial and abaxial sides was separated by peeling and analyzed by immunoblot analysis. E, Immunoblot analysis of phosphorylated GsPIP2;2 in the presence of 10-mM CaCl₂ and 10-mM EGTA. Full-open flowers were pretreatment with 10-mM EGTA and 10-mM CaCl₂ for 4 h. Then, flowers were incubated at 16°C for 30 min in the dark and then incubated at 22°C for 60 min in the light or dark, and the epidermal cell layer on the adaxial were peeled and analyzed. F, Flower opening and closure movement analysis in the presence of 10-mM EGTA and 10-mM CaCl_2. After pretreatment with 10-mM EGTA and 10-mM CaCl₂ for 4 h_, flowers were incubated at 16°C for 30 min in the dark and then incubated at 22°C for 60 min in the light or dark. Corolla angle was measured every 30 min. For biological replicates, flowers were collected and analyzed from four independent plants (n = 4). Flower opening and closing movements were repeatedly investigated and similar results were obtained in all tests. The data represent means ± sd from four biological replicates (n = 4). Data were analyzed using a repeated measures two-way ANOVA with Dunnett’s multiple comparisons test to compare the difference between Mock and each treatment. Asterisks indicate statistically significant (*P < 0.05, ***P < 0.001) and ns, not significant. G and H, GCaMP3 imaging in the adaxial cell of the adaxial corolla surface. GCaMP3, as a biosensor for Ca²⁺, was transiently expressed in the adaxial cell of the adaxial corolla surface by the bombardment method. Corolla was incubated at 16°C (G, upper) and then incubated at 22°C in the light (G, bottom; H, top) or dark (H, bottom). Confocal microscopy images were acquired at each time point presented. The fluorescence intensity of GCaMP3 was expressed as a pseudo-colored image (white indicates high and blue indicates low intensity, see the color code bar). The fluorescence intensity of cells was calculated by subtracting the background signal (F), and each plot shows the average of the relative values to the prestimulation (0 min, F0) values in five cells (n = 5). The data represent means ± sd from five cells (n = 5). Scale bar = 10 μm. Data were analyzed using a repeated measures two-way ANOVA with Bonferroni’s multiple comparisons test to compare the difference between light and dark treatment. Asterisks indicate statistically significant (**P < 0.01) and ns, not significant (H).

Next, we analyzed changes in endogenous GsPIP2;2 phosphorylation levels on the membrane during flower re-opening and re-closing using the phosphoproteome approach. Phosphoproteome analysis revealed that Ser280 and Ser283 at the C-terminus of GsPIP2;2 were phosphorylated (Supplemental Figure S4). In contrast, mono-phosphorylation of Ser280 or Ser283 and other phosphorylation sites was not detected. In addition, di-phosphorylation levels of GsPIP2;2 increased during flower re-opening (Figure 3C). These results suggest that the phosphorylation of the penultimate conserved serine residue (Ser280) and the ultimate conserved serine residue (Ser283) in GsPIP2;2 is associated with gentian corolla movement.

As shown in Figure 1E, epidermal cell expansion and contraction were observed only on the adaxial side but not on the abaxial side. Therefore, we next peeled apart the corolla and analyzed the GsPIP2;2 phosphorylation states on the two sides. To separate the adaxial and abaxial sides of the corolla, we used a modified version of the Tape-Arabidopsis sandwich method (Wu et al., 2009). After sandwiching the corolla with polyvinyl chloride tape, the epidermal cell layers on the adaxial and abaxial sides were peeled and harvested by carefully pulling the tape apart. Because the gentian corolla is soft and thick, it was possible to easily separate the adaxial and abaxial sides from the corolla. On the adaxial side, Ser-phosphorylation of GsPIP2;2 associated with reversible corolla movements (Figure 3D). In contrast, GsPIP2;2 phosphorylation increased during flower closing on the abaxial side, but the phosphorylation status did not change under subsequent conditions (Figure 3D).

Ca²⁺ ions are important intracellular secondary messengers (Dodd et al., 2010), and intracellular Ca²⁺ increases in response to external stimuli, such as light and temperature (Knight et al., 1996; Baum et al., 1999). Several studies have shown that PIP2-activating Ser-phosphorylation occurs in a Ca²⁺-dependent manner (Johansson et al., 1996; Wilder et al., 2008). Therefore, we investigated whether temperature- and light-dependent stimulation of GsPIP2;2 phosphorylation occurs in a Ca²⁺-dependent manner. Immunoblot analysis showed that GsPIP2;2 phosphorylation was suppressed by treatment with the Ca²⁺ chelator ethylene glycol tetra acetic (EGTA) acid, but recovered by co-treatment with exogenous CaCl₂ (Figure 3E). Furthermore, co-treatment with EGTA and CaCl₂ promoted phosphorylation and flower re-opening in the dark at 22°C compared to the mock treatment (Figure 3, E and F), whereas EGTA treatment significantly suppressed flower re-opening under both dark and light conditions at 22°C (Figure 3F). In contrast, EGTA and CaCl₂ treatments did not affect flower closure (Figure 3F).

Next, we analyzed the increase in cytosolic Ca²⁺ levels in response to environmental stimuli in the corolla epidermal cells using GCaMP3 as a Ca²⁺ biosensor (Tian et al., 2009). GCaMP3 was transiently expressed by using particle bombardment to deliver the construct to the adaxial side of the corolla and analyzed by confocal microscopy. The fluorescence intensity of G-GaMP3 increased within 5 min after cool-temperature treatment and then gradually decreased (Figure 3G). Subsequently, exposure to normal growth temperature and light stimulation induced a gradual increase in fluorescence intensity (Figure 3, G and H), which was light stimulus dependent (Figure 3H). These results suggest that Ser-phosphorylation of GsPIP2;2 was triggered by the elevation of intracellular Ca²⁺ levels induced by exposure to a cool temperature or light stimulation under normal growth temperature.

GsCPK16 activates GsPIP2;2 by phosphorylating the conserved Ser residue in a Ca²⁺ ion-dependent manner

Previous studies have suggested that phosphorylation of the C-terminal Ser residue(s) in ZmPIPs might be mediated by PM-associated Ca²⁺-dependent protein kinase(s), CPK(s) in Zea mays (Wilder et al., 2008; Qing et al., 2016). To identify the kinase(s) responsible for phosphorylating GsPIP2s, we focused on 39 contigs annotated as CPKs in the RNA-seq data (Supplemental Table S3). Of these, 21 contigs were estimated to have complete open reading frames (ORFs). Some CPKs contain a myristoylation motif at the N-terminus and are localized to the PM in a myristoylation-dependent manner (Hrabak et al., 2003). Thus, we performed a myristoylation prediction analysis based on the predicted amino acid sequences and found that six CPKs each contained a typical myristoylation motif (Supplemental Table S3). Finally, we selected four CPK candidates with putative myristoylation motifs that were highly expressed in the corolla and two CPK candidates that lacked a putative myristoylation motif. In total, six CPK candidate genes (GsCPK4, 13, 16, 28, 32, and 33) were isolated by reverse transcription PCR on gentian corolla RNA extracts and used for further analysis.

Using a wheat cell-free system (Madin et al., 2000; Takai et al., 2010; Nozawa et al., 2011), we synthesized the above six GsCPKs, GsPIP2;2, and GsPIP2;7 proteins. After confirming that the synthesized CPKs had kinase activity, they were used in the in vitro kinase assay (Supplemental Figure S5). In the Phos-tag SDS–PAGE assay, mobility-shifted bands of GsPIP2;2 and GsPIP2;7 were observed after incubation with GsCPK16 but not with the other GsCPKs (Figure 4A). This result suggests that GsCPK16 mediates the phosphorylation of GsPIP2s.

Figure 4.

Identification of GsCPK16 for activation of GsPIP2;2 by phosphorylation. A, In vitro kinase assay of GsPIP2;2 and GsPIP2;7 by GsCPK4, 13, 16, 28, 32, and 33. Phosphorylated GsPIP2;2 and GsPIP2;7 (P-GsPIP2) were analyzed by Phos-tag SDS–PAGE. Without protein kinase as a negative control (−). B, In vitro kinase assay of GsPIP2;2, GsPIP2;7 and their mutants. WT of GsPIP2;2 and GsPIP2;7 and their mutants GsPIP2;2^S280A, GsPIP2;7^S274A were incubated with WT or KD form of GsCPK16. Phosphorylated GsPIP2;2 (P-GsPIP2;2) were analyzed by Phos-tag SDS–PAGE. C, In vitro kinase assay of GsPIP2;2 and GsPIP2;7 using an anti-phospho-Ser280 antibody. WT of GsPIP2;2 and GsPIP2;7 were incubated with WT or KD form of GsCPK16. Phosphorylated GsPIP2;2 (P-GsPIP2;2) were detected by immunoblot analysis using anti-phospho-Ser280 antibody. D, In vitro kinase assay of GsPIP2;2 in the presence of CaCl₂ and EGTA. GsPIP2;2 proteins were incubated with GsCPK16 in the presence of 1- to 100-μM CaCl₂ and 100-μM EGTA. All synthesized GsPIP2s proteins were dephosphorylated by lambda protein phosphatase and purified prior to the in vitro kinase reaction with affinity-purified His-tagged GsCPKs (A–D). E, Subcellular localizations of GsPIP2;2-GFP, GsCPK16^WT-YFP, and GsCPK16^G2A-YFP in corolla protoplast. F, BiFC analysis of GsPIP2;2-GsCPK16 interaction in corolla protoplast. The N-terminal half of YFP fused GsPIP2;2 (nYFP-GsPIP2;2) and the C-terminal half of YFP fused GsCPK16 or GsCPK28 (GsCPK16-cYFP or GsCPK28-cYFP) were co-expressed in protoplasts. nYFP and cYFP were used for negative control. Fluor, GFP or YFP fluorescence image; FM4–64 stain fluorescence image; Merge, merged image of Fluor and FM4–64 images; DIC, differential interference contrast image. Signal intensity was measured along the white line in the direction of the arrowhead. The graphs were expressed as a relative with the maximum measured value as 1. Scale bar = 10 μm (E and F). G, In cell phosphorylation assay of endogenous GsPIP2;2 by transient expressed WT or KD form of GsCPK16. GsCPK16-V5 was expressed in corolla protoplast, and phosphorylated endogenous GsPIP2;2 proteins (P-GsPIP2;2) were analyzed by IB. Relative phosphorylation levels were calculated as relative values when the signal intensity ratio of phosphorylated GsPIP2;2 to total GsPIP2;2 protein in the mock was 1. H, cell-surface biotinylation assay. Protoplasts expressing GsPIP2;2 and GsPIP2;2^S280A proteins were incubated at 16°C for 30 min in the dark and then incubated at 22°C for 60 min in the light or dark. Cell surface proteins were loaded with biotin-SS-Sulfo-OSu and collected using STA beads. Biotinylated GsPIP2;2 proteins were detected by IB using anti-FLAG M2 antibody (top). Putative glycosylated proteins were indicated with an asterisk. The relative biotinylated GsPIP2;2 levels were calculated as a relative value when the signal intensity ratio of biotinyleated GsPIP2;2 to the total GsPIP2;2 of the mock was 1 (bottom). The data represent means ± sd from three biological replicates (n = 3). The results are from a Brown–Forsythe and Welch’s ANOVA test, followed by Dunnett’s T3 multiple comparisons test with a significance level of 5% (P < 0.05).

To identify the phosphorylation site(s) of GsPIP2s, we prepared the single amino acid substitution mutants GsPIP2;2^S280A and GsPIP2;7^S274A, in which a conserved Ser residue at the C-terminus was replaced with Ala. In addition, we prepared the kinase-dead (KD) form GsCPK16 (GsCPK16^KD, in which the conserved Lys115 residue in the ATP-binding sites was changed to Arg) (Supplemental Figures S5 and S6). The in vitro kinase assay of the mutants detected strong mobility-shifted bands of wild-type (WT) GsPIP2;2 and GsPIP2;7 in a GsCPK16 activity-dependent manner (Figure 4B; Supplemental Figure S5); however, the band for GsPIP2;2^S280A mutants was barely detectable (Figure 4B). In contrast, the mobility-shifted bands of GsPIP2;7^S274A were detected but the signal was not strong (Figure 4B). These results suggest that GsCPK16 mainly catalyzes the phosphorylation of conserved Ser residues that are presumed to be involved in the activation of GsPIP2.

In addition, we confirmed that site-specific GsPIP2;2 phosphorylation by GsCPK16 is recognized by the anti-phospho Ser280 antibody but barely recognized GsPIP2;7 phosphorylation (Figure 4C). Thus, we subsequently focused on GsPIP2;2 as a substrate for GsCPK16 in further analyses to clarify the biochemical properties of GsCPK16. The expected amino acid sequence of GsCPK16 contained Ca²⁺-binding motifs (alpha-helix (E)-loop-alpha-helix (F), called EF-hand) at the C-terminus and GsCPK16 could phosphorylate GsPIP2;2 only in the presence of Ca²⁺ (Figure 4D; Supplemental Figure S6). Furthermore, gene expression analysis revealed that the expression of GsCPK16 was induced in the late stage of flower development (stage 4; Supplemental Figure S7), which appeared to be consistent with GsPIP2 phosphorylation during flower development (Figure 3A).

GsCPK16 is mainly localized on the PM and catalyzes the phosphorylation of GsPIP2;2

Next, we examined the localization of GsPIP2;2 and GsCPK16 proteins in gentian corolla cells. Transient expression analyses showed that both GsPIP2;2-green fluorescent protein (GFP) and GsCPK16-YFP fluorescence overlapped with the fluorescence of the PM marker FM4–64, indicating that GsPIP2;2 and GsCPK16 were mainly localized on the PM in corolla cells (Figure 4E). In contrast, GsCPK16^G2A-GFP, a mutant of the putative myristoylation site (Gly2 to Ala), was mainly observed intracellularly (Figure 4E). Furthermore, BiFC analysis showed YFP fluorescence on the PM (Figure 4F), indicating that GsCPK16 interacts with GsPIP2;2 on the PM. In addition, phosphorylation of endogenous GsPIP2;2 increased with transient expression of WT GsCPK16, whereas GsCPK16^KD was expressed at the same level as the control (Mock; Figure 4G). These results suggest that GsCPK16 is anchored to the PM through its N-terminal myristoylation and catalyzes the phosphorylation of GsPIP2;2.

Phosphorylation of PIP2s has also been shown to be involved in intracellular localization (Prak et al., 2008). Therefore, using a combination of cell surface biotinylation assay and protoplast system, we investigated whether GsCPK16-dependent phosphorylation of GsPIP2;2 is involved in membrane trafficking. In protoplast cells expressing WT GsPIP2;2, the membrane abundance of GsPIP2;2 decreased significantly in response to cool temperature and subsequently increased in response to temperature increases (Figure 4H). A slight delay in the increase in membrane abundance of GsPIP2 appeared to have been observed under co-treatment with light stimulation, but quantitative analysis showed no significant difference (Figure 4H). Although WT GsPIP2;2 proteins were found in their unglycosylated (∼30 kDa) and putative glycosylated (∼32–35 kDa) forms, the GsPIP2;2^S280A mutant proteins were found only in the unglycosylated form (Figure 4H). However, the changes in the membrane distribution of GsPIP2;2^S280A mutant proteins were similar to those in the WT (Figure 4H), suggesting that phosphorylation at Ser280 does not affect GsPIP2;2 localization on the PM.

GsCPK16-dependent phosphorylation promotes GsPIP2;2 water transport activity

To investigate the function of GsCPK16-dependent phosphorylation of GsPIP2;2, we applied a spheroplast-based water channel activity assay (Azad et al., 2009). We used the Saccharomyces cerevisiae system instead of Pichia pastoris. We generated yeast strains that overexpressed GsPIP2;2 or several GsPIP2;2 mutants and used them to prepare spheroplast cells. Water channel activity was indicated as the rate of decrease in absorbance (optical density at 600 nm, OD_600 nm) that reflected the water influx and subsequent burst of yeast spheroplast cells under hypo-osmotic treatment. The absorbance of yeast spheroplasts overexpressing GsPIP2;2^WT rapidly decreased in response to hypo-osmotic treatment compared to control spheroplasts (Figure 5A), whereas it was alleviated by HgCl₂ treatment as aquaporin inhibitor (Supplemental Figure S8). Overexpression of a phosphorylation-mimic mutant GsPIP2;2^S280E (converted Ser280 to Glu) caused a rapid decrease in absorbance compared to overexpression of GsPIP2;2^WT and GsPIP2;2^S280A (Figure 5A). Similarly, co-expression of GsPIP2;2^WT and GsCPK16^WT caused a rapid decrease in absorbance compared to overexpression of GsPIP2;2^WT alone, but no significant difference was observed in cells co-expressing GsPIP2;2^WT with GsCPK16^KD or GsPIP2;2 ^S280A with GsCPK16^WT (Figure 5A). Furthermore, we calculated the rate constant of decrease in absorbance (OD_600 nm) by curve fitting the acquired data using a two-phase exponential decay function, resulting yeast expressing GsPIP2;2^S280E and co-expressing GsPIP2;2^WT and GsCPK16^WT showed a significantly higher rate constant value compared to other yeasts (Figure 5A). These results suggest that GsCPK16 phosphorylates the Ser280 residue of GsPIP2;2, resulting in activation gating of GsPIP2;2.

Figure 5.

Silencing of gentian GsCPK16 suppresses flower re-opening movement. A, Water transport assay in S. cerevisiae. GsPIP2;2^WT, GsPIP2;2^S280A or GsPIP2;2^S280E was expressed, and GsPIP2;2^WT-GsCPK16^WT, GsPIP2;2^WT-GsCPK16^KD or GsPIP2;2^S280A-GsCPK16^WT was co-expressed in yeast. Spheroplasts were prepared from these cell lines and treated with hypo-osmotic shock. Rupture of the spheroplasts due to water influx was monitored by measuring the decrease in OD_600 nm every 2 s. The optical absorbance was expressed as a relative value with the starting OD_600 nm value as 1.0. Data were obtained by six independent experiments and fitted to a two-phase exponential decay functions (top). Rate constant values (first-stage rate constant as k₁ and second-stage rate constant as k₂) for the fitted curves were calculated, and k₁ and k₂ were represented by relative values to the control (bottom). The data represent the means ± sd (n = 6). Data were analyzed using a Brown–Forsythe and Welch’s ANOVA test with Dunnett’s T3 multiple comparisons test with a significance level of 5% (P < 0.05). B, Quantitative RT–PCR of GsCPK16 in corolla of vector control, ALSV:::GsCPK16-infected plants 30 days after infection. Total RNA was extracted and analyzed from the corolla of six independent infected plants for biological replicates. The data represent means ± sd from six biological replicates (n = 6). Statistically significant changes compared with vector control plants were determined by a two-tailed unpaired t test (**P < 0.01). C, Corolla movements in response to temperature and light of vector control and ALSV:::GsCPK16-infected plants (∼60 days after infection). Morphology (top), floral opening–closure angles (bottom). The data represent the means ± sd of biological replicates obtained from six independent infected plants (n = 6). Data were analyzed using a repeated measures two-way ANOVA with Bonferroni’s multiple comparisons test to compare the difference between vector control and ALSV:::GsCPK16 plant. Asterisks indicate (*P < 0.05, **P < 0.01). ns, not significant. Scale bar = 1 cm. D, Immunoblot analysis of phosphorylated endogenous GsPIP2;2 in corolla of vector control and ALSV:::GsCPK16-infected plants. Corolla of ALSV-infected plants was vertically cut into four to five segments, including the corolla robe, and then floated in the water. Corolla segments were incubated at 16°C for 30 min in the dark and then incubated at 22°C for 60 min in the light or dark. Phosphorylated endogenous GsPIP2;2 proteins (P-GsPIP2;2) and endogenous GsPIP2;2 were analyzed by IB analysis (top). The signal value of phosphorylated endogenous GsPIP2;2 (P-GsPIP2;2) and endogenous GsPIP2;2 were measured by ImageJ software. Relative phosphorylation levels were calculated as relative values when the signal intensity ratio of phosphorylated GsPIP2;2 to total GsPIP2;2 protein in the mock of vector control was 1 (bottom). For biological replicates, one corolla was collected from each of the three independent infected plants and used in the experiment. The data represent the means ± sd of biological replicates obtained from three independent infected plants (n = 3). Significant differences (*P < 0.05; ns, not significant) were determined by unpaired t test with Welch’s correction.

GsCPK16 promotes flower re-opening through phosphorylation-dependent GsPIP2;2 activation

Next, we investigated whether GsCPK16-dependent phosphorylation and activation of GsPIP2;2 are involved in flower re-opening in gentian using a VIGS approach. From the results of RNA-seq analysis, we focused on 12 GsCPK genes with FPKM values greater than 20 and analyzed their gene expression levels in ALSV:::CPK16 infected plants by reverse transcription-quantitative PCR (RT-qPCR). The results showed that the expression level of the GsCPK16 gene alone was reduced in plants infected with ALSV:::GsCPK16 (Figure 5B; Supplemental Figure S9).

Compared to the vector control lines, silencing of GsCPK16 resulted in suppression of corolla expansion and a significant delay in flower re-opening but did not affect flower closure movements (Figure 5C). In vector control lines, the Ser-phosphorylation levels of GsPIP2;2 increased in response to cool temperatures and subsequently elevated temperatures and light stimuli (Figure 5D). In knockdown lines, the Ser-phosphorylation levels of GsPIP2;2 seemed similar to those of the vector control lines, but these levels barely increased in response to temperature and light stimuli (Figure 5D). Collectively, these results suggest that GsCPK16 regulates flower re-opening by catalyzing the phosphorylation-dependent activation of GsPIP2;2 in gentian plants.

Discussion

All of the results in this study indicate that GsCPK16 is activated by light-induced increases in cytosolic Ca²⁺ levels and induces GsPIP2;2 phosphorylation, thereby promoting flower re-opening. Based on our results, we proposed the biochemical and biological roles of GsPIP2;2 and GsCPK16 in controlling flower re-opening (Figure 6).

Figure 6.

A proposed model for gentian corolla re-opening in response to temperature and light. GsPIP2;2 and GsPIP2;7 interact with the PM and coordinately regulate flower re-opening. The shift from cool (16°C) to normal growth temperature (22°C) as a stimulus promotes solute influx into the cytoplasm, resulting in an increased intracellular osmotic pressure (magenta dotted arrows). In addition, the temperature shift stress induces dephosphorylation and re-localization of internalized GsPIP2;2 in the PM (magenta solid arrows). Thus, the temperature-dependent regulatory mechanism of solute balance and PIP2 density plausibly plays an important role in determining and promoting flower re-opening. Light stimulation induces elevation of intracellular Ca²⁺ levels (yellow solid arrows). Ca²⁺, by activating GsCPK16, promotes GsPIP2;2 phosphorylation and activation (solid yellow arrow). Therefore, the water transport activity of GsPIP2;2 increases, which promotes water influx and leads to rapid flower re-opening (thick black arrows). GsCPK16 possibly induces phosphorylation-dependent activation of GsPIP2;2 and GsPIP2;7 (yellow dotted arrow). In contrast, in the dark at normal growth temperature, the density of phosphorylated GsPIP2;2 proteins in the PM is low compared to that under light. Hence, water influx into the cells progresses gently, and thereby flower re-opening progresses more slowly in dark than in light (thin black arrows).

In this study, we first confirmed that the reversible movement of Japanese gentian flowers, similar to that of European gentian flowers as Gentiana kochiana, G. algida, and G. leucomelaena (Claus, 1926; Bynum and Smith, 2001; Mu et al., 2010), was primarily induced by temperature, but light stimulation promoted flower re-opening (Figure 1). Cell contraction and expansion occurred only on the adaxial side of the corolla, probably because of the difference in cell wall extensibility. Our studies further indicated that silencing GsPIP2;2 and GsPIP2;7 resulted in the suppression of flower re-opening, suggesting functional redundancy between the two genes. These results suggest that aquaporin-mediated rapid water transport on the adaxial side is essential for repeated flower opening without cell growth. Many previous studies have shown that PIP2 forms a heterotetramer with proteins of another subfamily, PIP1, and complex formation promotes permeability (Yaneff et al., 2015). Furthermore, protein–protein interaction analysis of membrane proteins in Arabidopsis revealed that PIP2 interacts with PIP1 and PIP2 family members (Lalonde et al., 2010). Similarly, the interaction between GsPIP2;2 and GsPIP2;7 was also observed in the PM of gentian corolla cells (Figure 2C). These findings suggest that GsPIP2;2 and GsPIP2;7 are not simply redundant genes; rather, the encoded proteins could function in a coordinated manner in gentian flower re-opening by forming a complex (Figure 6).

In this study, we found that the phosphorylation levels of the penultimate conserved Ser residues at the C-terminus of GsPIP2;2 increased during flower re-opening movements (Figure 3, A–C), and its phosphorylation was catalyzed by GsCPK16, which is activated by elevated cytosolic Ca²⁺ levels in response to light stimuli under normal growth temperature. Moreover, we demonstrated that GsCPK16-dependent phosphorylation of GsPIP2;2 could induce its activation (Figure 5). Therefore, these results suggest that the elevation of cytoplasmic Ca²⁺ levels and activation of GsCPK16 in response to light stimuli is one of the regulatory mechanisms of GsPIP2;2 activation to achieve rapid flower re-opening in gentian (solid yellow arrows in Figure 6). However, cool-temperature stress also induced an elevation in Ca²⁺ levels and the phosphorylation of GsPIP2;2, but the flowers did not open. Furthermore, under cool-temperature conditions, light stimulation did not induce flower re-opening. Gentian flower opening depended on temperature rather than light stimulation, suggesting that temperature and light play a role in determining and promoting flower re-opening, respectively.

Cellular water transport is regulated by PIP2 density and activity in the PM (Tyerman et al., 1999). Immunoblot analysis showed that the phosphorylation levels of total GsPIP2;2 were significantly increased at cool temperature, whereas proteome analysis did not show changes in their relative phosphorylation levels in the cell membrane. This difference may be because of decreased GsPIP2;2 abundance in the PM, which suggests that most of the phosphorylated GsPIP2;2 molecules are present in the cytoplasm during flower closure. Plants have the potential to selectively degrade phosphorylated PIP2 via autophagy or ubiquitin-dependent proteasomal degradation pathway and regulate intracellular water balance during environmental stress (Prak et al., 2008; Chen et al., 2021). However, it is not known whether internalized PIP2 is re-localized in the PM or degraded after stress release. In this study, we observed a rapid decrease in GsPIP2;2 phosphorylation levels and a slight increase in the GsPIP2;2 protein density in the PM on shifting from cool temperature to normal growth temperature (Figure 3B and 4H). In addition, relative phosphorylation levels of GsPIP2;2 in the cell membrane in the darkness at 22°C condition were not significantly different from those in the darkness at 16°C conditions, but appeared to be slightly lower (Figure 3C). These results suggest that some of the phosphorylated GsPIP2;2 molecules that became internalized owing to cool temperature stress were dephosphorylated and re-localized in the PM on shifting from cool temperature to normal temperature. Together, these results suggest that flower re-opening is regulated by GsPIP2;2 regulatory mechanisms in response to two environmental stimuli: (1) increased GsPIP2;2 density in the PM induced by temperature changes and (2) phosphorylation-dependent GsPIP2;2 activation by light stimulation (magenta and yellow arrows in Figure 6). In contrast, the C-terminal region of PIP2 interacts with transcription factors, 14-3-3 proteins, and Crinkler effectors of oomycete pathogens in a phosphorylation-dependent or independent manner (Prado et al., 2019; Zhang et al., 2019; Ai et al., 2021). These findings suggest that not only plants but also pathogens use plant aquaporins for survival and that aquaporins play important roles in various physiological and biological processes.

Because aquaporins are channel proteins, the speed and direction of water transport depends on the solute concentration on either side of the cell membrane. Therefore, the temperature change from 16°C to 22°C as a stimulus and the light stimulation probably causes an increase in osmotic pressure by inducing the influx of solute into the corolla cells (magenta dotted arrows in Figure 6). The functional link between photoreceptor phototropins and H⁺-ATPases may explain this complex relationship between temperature and light. H⁺-ATPases are sensitive to temperature, and their activity is significantly suppressed at cool temperatures (Dupont, 1989). Thus, the temperature change from 16°C to 22°C as a stimulus may cause the activation of H⁺-ATPase. Activated H⁺-ATPase promotes proton efflux and hyperpolarization of PMs, resulting in an influx of cations such as potassium and water (Inoue and Kinoshita, 2017). These temperature-dependent solute regulation mechanisms may play a role as determinants of flower re-opening. On the other hand, previous studies have shown that photoreceptor phototropins increase cytoplasmic Ca²⁺ levels and the phosphorylation and activation of H⁺-ATPase in response to light (Kinoshita and Shimazaki, 1999; Kinoshita et al., 2001; Harada et al., 2003; Merlot et al. 2007). Although it is not clear if phototropin functions in the gentian corolla, the movement of the Oenothera lamarkiana flower seems to be caused by the blue light, which is suitable for activating phototropins (Saito and Yamaki, 1967). These findings suggest that phototropin-mediated regulation of cytoplasmic Ca²⁺ and proton levels might be one of the mechanisms that regulate promotion of rapid resumption of gentian flowers (yellow and thick black arrows in Figure 6). However, it seems that aquaporin phosphorylation is not necessarily induced by light stimulation. In tulip, flower re-opening is thought to be due to phosphorylation and activation of aquaporin(s) induced by elevated cytoplasmic Ca²⁺ levels in response to temperature change as a stimulus (Azad et al., 2004). Thus, temperature-sensitive Ca²⁺ transporters, such as cyclic nucleotide-gated channels, may be primarily involved in the tulip corolla (Gao et al., 2012). In contrast, in Arabidopsis, light-dependent changes in leaf hydraulic conductivity are due to the phosphorylation-dependent functional regulation of AtPIP2;1 expressed in the veins (Prado et al., 2013). However, it has been shown that AtPIP2;1 phosphorylation and activation are induced in the dark. These findings suggest that some plants regulate the phosphorylation of aquaporins by stimulus-induced Ca²⁺ oscillations to regulate water balance adapted to environmental changes; however, environmental factors that cause Ca²⁺ oscillations could differ among plant species and organs.

Phosphorylation of both the penultimate and ultimate Ser residues in PIP2 is important for its activation; however, recent studies have suggested that their phosphorylation differently affect activity (Huang et al., 2020; Qiu et al., 2020). These inconsistencies might be due to differences in aquaporin types or assay systems, among other reasons. Although the two conserved Ser residues (Ser280 and Ser283) at the C-terminus of GsPIP2;2 are in close proximity, the mutant analysis showed that GsCPK16 mainly phosphorylates Ser280 (Figure 4B). GsCPK16-dependent phosphorylation of GsPIP2;2 leads to channel activation but does not appear to affect localization and trafficking behavior. However, phosphoproteome analysis revealed di-phosphorylation of Ser280 and Ser283 at the C-terminus of PIP2;2, and mono-phosphorylation was not detected. Therefore, our study suggests that the phosphorylation of Ser280 and Ser283 at the C-terminus of GsPIP2;2 is important for flower re-opening and that their phosphorylation is catalyzed by GsCPK16 and unknown protein kinase(s), respectively. However, several previous studies have suggested that phosphorylation of the two Ser residues at the C-terminus of aquaporin occurs through interdependent events resulting from either the progressive or partitioning function of protein kinase (Hoffert et al., 2006; Prak et al., 2008). These findings suggest that mutations in Ser280 can affect phosphorylation of Ser283 of GsPIP2;2, and GsCPK16 may potentially have the ability to catalyze the di-phosphorylation of both Ser280 and Ser283. Alternatively, phosphorylation of Ser280 by GsCPK16 may promote interdependent phosphorylation events of Ser283 by unknown protein kinase(s).

Previous studies have shown that flower opening is also regulated by phytohormones such as abscisic acid (ABA) and ethylene. Although ABA acts as a trigger molecule that induces stomatal closure, it also affects the rhythm of the flower opening and closing movements of Kalanchoë blossfeldiana (Schrempf, 1980). Grondin et al. showed that open stomata 1/Snf1-related protein kinase 2.6, an ABA signal regulator, induces activation by phosphorylating conserved Ser residues located in loop B of AtPIP2;1, thereby inducing stomatal closure (Grondin et al., 2015). However, its phosphorylation was not detected in the gentian corolla. Ethylene causes wilting due to water loss, primarily accelerating the senescence process. Qing et al. showed that the phosphorylation of two conserved Ser residues at the C-terminus of PIP2 is caused by unknown protein kinase(s) regulated by the ethylene signal (Qing et al., 2016). Furthermore, several previous studies have shown that ethylene is involved in the development and flower opening of rose corolla (Ma et al., 2008; Xue et al., 2008). However, the movement of the gentian corolla is unlikely to be regulated by changes in endogenous phytohormone levels because it is completed within 60 min.

In this study, we revealed that GsPIP2;2 and GsCPK16 play important roles in flower re-opening but could not provide evidence that they are actively involved in flower re-closure. In this study, we used the VIGS approach to analyze gene function, but it should be noted that the expression of the target genes was not completely silent. Even if the amounts of GsPIP2;2 and GsPIP2;7 proteins on the PM are reduced, it might be possible to rapidly transport the quantity of water needed to achieve flower closure. However, it is also possible that flower closure is regulated by other aquaporins expressed in the corolla. RNA-seq analysis revealed that the expression of GsPIP1;2 and GsPIP1;4 in the gentian corolla was higher than that in all GsPIP2 genes. Although PIP1 is thought to act as a modulator that regulates the activity of PIP2, some previous studies have revealed that several PIP1s also have water transport activity (Yaneff et al., 2015). These findings suggest that both PIP2 and PIP1 play significant roles in the motility of the gentian corolla. To prove this hypothesis, knocking out the target gene by genome editing would be effective. Recently, we successfully applied genome editing technology to specific gentian varieties (Tasaki et al., 2020). Genome editing of gentian PIP loci may help clarify the role of PIP2 and PIP1 in gentian corolla motility.

Based on our results, we conclude that gentian flower re-opening occurs due to water influx into the cells mediated by GsPIP2;2 and GsPIP2;2 activity and is positively regulated by GsCPK16-dependent phosphorylation. Our findings may contribute to understanding regulatory mechanisms underlying the reversible corolla movement in many plant species.

Materials and methods

Plant material

The G. scabra breeding line “Bzc-1” was used in this study. “Bzc-1” was developed at the Iwate Agricultural Research Center in Japan by cross breeding using Japanese gentian G. scabra and has been registered in the Plant Variety database (Application number 34732, Ministry of Agriculture, Forestry and Fisheries, Japan). The plants were maintained in vitro on solid Murashige and Skoog medium containing 3% (w/v) sucrose and 0.2% (w/v) gellan gum at 22°C under a 16-h light/8-h dark photoperiod with white fluorescent light (FL40SS W/37R white light bulb, Panasonic, Osaka, Japan; 50 μmol m⁻² s⁻¹) as described previously (Tasaki et al., 2016). Flower buds formed ∼2–3 months after the segment nodes were transplanted onto the medium and cultured.

Reversible flower opening and closing movement assay

For the flower opening and closing movement assay, fully opened flowers were detached from plants and placed in microtubes containing sterile water. The flowers were then sealed in a plant culture box to prevent them from desiccating. After placing the flowers in darkness or light for 30 min at 16°C, the flowers were incubated at 22°C for 30–120 min under either a continuous light intensity of 50 μmol m⁻² s⁻¹ (FL40SS W/37R white light bulb, Panasonic, Osaka, Japan) or darkness. The flower opening rate was calculated by measuring the angle from the center of the flower to the tip of the corolla. The epidermis of the corolla was observed with an opto-digital microscope (DSX500, Olympus Co., Tokyo, Japan), and the mean cell area was calculated from digital images using ImageJ software. For chemical treatments, detached fully opened flowers were pretreated with 10-mM CaCl₂ and 10-mM EGTA for 4 h under continuous light.

Plasmid construction

Total RNA was isolated from the gentian corolla using an RNeasy Plant Mini Kit (Qiagen, Hilden, Germany). To synthesize first-strand cDNA, 100 ng of total RNA was reverse transcribed using the ReverTra Ace qPCR RT Master Mix with gDNA Remover kit (TOYOBO, Osaka, Japan) according to the manufacturer’s instructions. Gentian GsPIP2;1, 2;2, 2;7 and GsCPK4, 13, 16, 28, 32, and 33 were amplified by PCR using synthesized cDNA as a template. GCaMP3 (accession number: ADJ53338.1) was synthesized by GeneArt Gene Synthesis (Thermo Fisher Scientific, Waltham, MA, USA). The DNA fragments of the ORF were cloned into pDONR221. After confirming the sequence, cell-free expression vectors and transient expression vectors were generated by LR clonase recombination. Mutagenesis was carried out using a PrimeSTAR Mutagenesis Basal Kit (Takara Bio, Shiga, Japan) according to the manufacturer’s instructions.

To construct yeast expression plasmids, DNA fragments of GsPIP2;2^WT-FLAG, GsPIP2;2^S280A-FLAG, GsPIP2;2^S280E-FLAG, GsCPK16^WT-V5, and GsCPK16^KD-V5 were amplified by PCR with primers harboring BglII, BamHI, or NotI sites using p35SΩ-GsPIP2;2^WT-3 × FLAG, p35SΩ-GsCPK16^WT-V5, and these mutants as templates. Amplified DNA fragments were cloned into pYES2/CT for GsPIP2;2 or pYES3/CT for GsCPK16, and the sequences were confirmed.

To construct the ALSV vector, the trigger fragments were amplified by PCR with primers harboring XhoI or BamHI sites (Supplemental Table S4) using GsPIP2;2, GsPIP2;7, or GsCPK16 in pDONR221 as the templates. Amplified fragments were cloned into the pEALSR2L5R5 vector (Li et al., 2004).

All genes were sequenced using an Applied Biosystems SeqStudio Genetic Analyser (Thermo Fisher Scientific, Waltham, MA, USA). All plasmid and primer sequences used in this study are listed in Supplemental Tables S4 and S5.

Library preparation and sequencing

Total RNA from the corolla of gentian “Bzc-1” flowers was isolated using the RNeasy Plant Mini Kit (Qiagen, Hilden, Germany). Total RNA was used to construct cDNA libraries using the NEBNext Ultra II DNA Library prep for Illumina (New England BioLabs, Ipswich, MA, USA) according to the manufacturer’s instructions. The extracted RNAs were sequenced on an Illumina HiSeq4000 platform. RNA sequencing data were deposited with DDBJ (accession no. DRA011818). In the quality control step, we filtered and discarded reads shorter than 50 bases and those with average read quality below 20 and trimmed the poly A and adapter sequences using FaQCs version 2.08 (Lo and Chain, 2014).

VIGS of GsPIP2;2, GsPIP2;7, and GsCPK16 in gentian plants

The generated ALSV vectors, ALSV:::GsPIP2;2, GsPIP2;7 or GsCPK16, and pEALSR1 were co-inoculated into plants grown for 1 month by the bombardment method using the PDS-1000/He particle delivery system (Bio-Rad, Hercules, CA, USA), as described previously (Nakatsuka et al., 2015). After bombardment, virus-infected plants were cultured for 1 month, and the proliferation of ALSV was confirmed by RT-PCR. Total RNA isolation from leaves and cDNA synthesis were performed using the method as described above (see “Plasmid construction”), and then PCR was performed with gene-specific primer pairs using the synthesized cDNA as a template. Amplified DNA fragments were analyzed by agarose gel electrophoresis. Plants infected with ALSV were subsequently cultured for 1 month and used for further analysis after flowering.

Gene expression analysis

Total RNA was isolated from the gentian corolla using an RNeasy Plant Mini Kit (Qiagen, Hilden, Germany). To synthesize first-strand cDNA, 100 ng of total RNA was reverse transcribed using the ReverTra Ace qPCR RT Master Mix with gDNA Remover kit (TOYOBO, Osaka, Japan) according to the manufacturer’s instructions. Using synthesized cDNA as a template, RT-qPCR analysis was carried out on a QuantStudio 3 system (Thermo Fisher Scientific, Waltham, MA, USA) with gene-specific primer pairs and Luna Universal qPCR Master Mix (New England BioLabs, Ipswich, MA, USA) used according to the manufacturer’s instructions. The acquired data were quantified using the ΔΔC_t method. The gentian ubiquitin gene was used as an internal control (Tasaki et al., 2020). All primer sequences used in this study are listed in Supplemental Table S4.

Immunoblot analysis of endogenous GsPIP2;2 in gentian corolla

For analysis of endogenous GsPIP2;2, corollas were taken at indicated intervals and frozen immediately in liquid nitrogen. In order to analyze the adaxial and abaxial sides separately, they were separated by peeling with polyvinyl chloride tape before freezing. For analysis of ALSV:::GsCPK16-infected plants, corollas were vertically cut into four to five segments, including the corolla robe. The corolla segments were floated on water and incubated under the indicated conditions. The samples were homogenized in a transmembrane protein extraction reagent (FIVEphoton Biochemicals, San-Diego, CA, USA) containing a phosphatase inhibitor cocktail (PhosSTOP, Roche, Basel, Switzerland) and a protease inhibitor cocktail (Roche, Basel, Switzerland). Homogenized samples were incubated on ice for 30 min and mixed vigorously every 15 min. After centrifugation at 15,000g for 10 min, the supernatant was collected, and the protein concentration was measured using the DC Protein Assay (Bio-Rad, Hercules, CA, USA). Total protein samples were incubated in SDS sample buffer at 70°C for 10 min and analyzed by immunoblotting (IB). Endogenous GsPIP2;2 was detected using an anti-PIP2-1-7 polyclonal rabbit antibody (1:3,000 dilution, Agrisera, Vännäs, Sweden). Phosphorylated GsPIP2;2 was detected using anti-phospho Ser280 rabbit polyclonal antibody (1:1,000 dilution) that was raised using synthetic peptide (AVKALG_pSFRSNPH) of P. communis PcPIP2;2 as antigens (Biologica, Nagoya, Japan). We confirmed that this antibody recognizes the phosphorylation of the penultimate conserved Ser residue (Ser280) of GsPIP2;2 (Supplemental Figure S2, and see “Validation of antibody”). Chemiluminescent detection was performed using Immobilon Western Chemiluminescent HRP Substrate (Millipore, Burlington, MA, USA) using the secondary donkey anti-rabbit IgG ECL antibody (1:20,000 dilution; GE Healthcare, Chicago, IL, USA). The membranes were visualized by Coomassie Brilliant Blue (CBB) G-250 staining.

Cell-free protein synthesis

In vitro transcription and wheat cell-free protein synthesis were performed using the WEPRO1240 expression kit (Cell-Free Sciences, Ehime, Japan) according to the manufacturer’s instructions. The translation reaction was performed in the bilayer mode (Takai et al., 2010) for 16 h at 16°C. For the synthesis of the GsPIP2s, azolectin liposomes were prepared and added to both the upper and bottom layers, as described previously (Nozawa et al., 2011). Aliquots were used for expression analysis and in vitro kinase assays.

Validation of antibody

For analysis of antibody specificity, GsPIP2;1, 2;2, 2;7-FLAG were analyzed by immunoblot analysis using anti-phospho Ser280 and anti-PIP2-1-7, and by Phos-Tag SDS–PAGE (see “In vitro kinase assays and in cell phosphorylation assays” below). For analysis of antibody specificity using lambda protein phosphatase, GsPIP2;1-FLAG, GsPIP2;2-FLAG, and total lysate prepared from gentian corolla were preincubated with lambda protein phosphatase (New England BioLabs, Ipswich, MA, USA) for 1 h at 30°C and then analyzed by immunoblot analysis. For the blocking peptide competition assay based on immunoblot analysis, we synthesized phospho-peptides and nonphopho-peptide derived from the C-terminal region of P. communis PcPIP2;2 (AVKALG_pSFRSNPH and VKALGSFRSNPH) and G. scabra GsPIP2;2 (AVKALG_pSFRSNPH and AVKALGSFRSNPH). The antigen proteins were separated by SDS-PAGE and then blotted onto immobilon-P polyvinylidene difluoride membrane (Merck, Kenilworth, NJ, USA). After blocking with 2% skim milk in Tris-buffered saline with 0.1% Tween 20 (TBS-T), the membrane was incubated with TBS-T containing an anti-phospho Ser280 antibody or an anti-PIP2-1-7 overnight at 4°C. Before incubation with membrane, an anti-phospho Ser280 antibody or an anti-PIP2-1-7 was added to 2 mL of TBS-T containing 10 μg/mL of synthetic peptide and preincubated for 1 h at room temperature. After washing with TBS-T, chemiluminescence was detected using the above method.

In vitro kinase assays and in cell phosphorylation assays

For in vitro kinase assays, synthesized gentian FLAG-tagged GsPIP2s/liposome were dephosphorylated with lambda protein phosphatase before in vitro kinase reaction. Dephosphorylation was carried out in a total volume of 150 μL containing 100 μL of FLAG-tagged GsPIP2s/liposome and 0.1 μL of lambda protein phosphatase (New England BioLabs, Ipswich, MA, USA) in 1× Protein Metallo Phosphatase buffer (New England BioLabs, Ipswich, MA, USA) with 1-mM MnCl₂ at 30°C for 60 min. Dephosphorylated FLAG-tagged GsPIP2s/liposome complex was collected by centrifugation at 20,000g for 10 min at 4°C and resultant pellet was washed with 150 μL of sterilized TBS 3 times. The pellet was resuspended in 100 μL of sterilized TBS and sonicated using Ultrasonic Disruptor UR-20P (Tomy Seiko, Tokyo, Japan) at 20°C for 5 min. His-tagged GsCPKs were purified using Ni-sepharose High-Performance beads (GE Healthcare, Chicago, IL, USA) as described previously (Nemoto et al., 2017). In vitro kinase assays were carried out in a total volume of 20 μL containing 10 μL of gentian GsPIP2;2-FLAG, GsPIP2;7-FLAG, and GsPIP2 mutants and 2 μL of His-GsCPKs (WT GsCPK3, 13, 16, 28, 32, and 33, and the KD form GsCPK16^KD were used) in kinase assay buffer (50-mM Tris–HCl [pH 7.6], 10-mM MgCl₂, 100-mM potassium acetate, 1-mM DTT, 100-μM ATP, and 0–100 μM CaCl₂) at 26°C for 60 min. To analyze GsCPK activity, 5 µg of myelin basic protein as a substrate was used in the in vitro kinase assay. For the Ca²⁺ chelating agent treatment, EGTA was added to a final concentration of 100 μM. After incubation, the reactions were terminated by the addition of 10 μL of 3× SDS sample buffer and incubated at 70°C for 10 min.

To analyze in cell phosphorylation analysis, protoplasts were isolated using the modified A. thaliana mesophyll protoplast method (Wu et al., 2009). After sandwiching the corolla with polyvinyl chloride tape, the epidermal cell layers on the adaxial and abaxial sides were peeled and harvested by carefully pulling the tape apart. Isolated protoplasts were resuspended in MaMg buffer (5-mM MES [pH 5.7], 0.4-M mannitol, and 15-mM MgCl₂). Ten micrograms of plasmid DNA of GsCPK16^WT-V5 or GsCPK16^KD-V5 was transfected into 2 × 10⁵ corolla protoplasts by PEG-methods (Wu et al., 2009) and then suspended in WI buffer (4-mM MES [pH 5.7], 0.5-M mannitol, and 20-mM KCl). After incubation for overnight in the light, protoplasts were harvested and incubated in 2× SDS sample buffer at 70°C for 10 min.

All samples were separated by SDS–PAGE or Phos-tag SDS–PAGE (50-μM Phos-tag Acrylamide AAL-107 [FUJIFILM Wako Pure Chemical, Osaka, Japan] and 50-μM MnCl₂) and analyzed by IB using anti-FLAG M2 antibody HPR (1:5,000 dilution; Sigma-Aldrich, St Louis, MO, USA), anti-V5-HRP antibody (1:10,000 dilution; Thermo Fisher Scientific, Waltham, MA, USA), anti-His-HRP antibody (1:10,000 dilution; FUJIFILM Wako Pure Chemical, Osaka, Japan), anti-PIP2-1-7 antibody (1:3,000 dilution; Agrisera, Vännäs, Sweden), and anti-phospho Ser280 antibody (1:1,000 dilution). Chemiluminescent detection was performed using Immobilon Western Chemiluminescent HRP Substrate (Millipore, Burlington, MA, USA). The membranes were visualized by CBB staining. The signal value was measured using ImageJ software.

Subcellular localization analysis

For subcellular localization analysis and BiFC assays, 10-μg plasmid DNA was transfected into 2 × 10⁵ corolla protoplasts as described above. After overnight incubation, protoplasts were incubated with 0.5% (v/v) FM4–64 (Thermo Fisher Scientific, Waltham, MA, USA) for 5 min, and fluorescence was observed using a FLUOVIEW FV1000 confocal laser scanning microscope (Olympus Co., Tokyo, Japan). Fluorescence images were acquired with the following settings: GFP and YFP (Ex 488 nm/Em 510–550 nm) and FM4–64 (Ex 543 nm/Em 600–660 nm). The signal intensities were measured from the displayed images using ImageJ software. The acquired value is presented as a relative value with a maximum value of 1. The experiments were performed at least 3 times using different plants. Over 10 protoplasts were analyzed for each biological replication, and finally over 30 protoplasts were analyzed. Similar localization was confirmed in all protoplasts, and the displayed data are representative images of them. All expression plasmids were listed in listed in Supplemental Table S5. Expression plasmid of p35SΩ-nYFP-GARU-NOST and p35SΩ-AtGID1A-nYFP-NOST was used for negative controls (Nemoto et al., 2017). The vector for BiFC analysis has a V5 epitope tag sequence inserted as a linker sequence between the Gateway cassette and nYFP or cYFP. The expressed protein was confirmed by immunoblot analysis with anti-V5-HRP antibody.

Cell-surface biotinylation assay

For the cell-surface biotinylation assay, 20-μg plasmid DNA of p35SΩ-GsPIP2;2^WT-3 × FLAG-NOST or p35SΩ-GsPIP2;2^S280A-3 × FLAG-NOST was transfected into 2 × 10⁵ corolla protoplasts as described above. After overnight incubation in the light, protoplasts were incubated at 16°C in the dark and then incubated at 22°C under continuous light or in the dark. Immediately after incubation, biotin-SS-Sulfo-OSu (DOJINDO, Kumamoto, Japan) was added at a concentration of 500 ng/mL, and the biotinylation reaction was carried out under the same conditions for 30 min. Protoplasts were collected and washed twice with quenching buffer (4-mM MES [pH 5.7], 0.5-M mannitol, 20-mM KCl, and 0.1-mM EDTA) buffer and then washed with WI buffer. Total protein was isolated using transmembrane protein extraction reagent (FIVEphoton Biochemicals, San-Diego, CA, USA) containing a phosphatase inhibitor cocktail (PhosSTOP; Roche, Basel, Switzerland), and a protease inhibitor cocktail (Roche, Basel, Switzerland) as described above. Biotinylated proteins were bound to Dynabeads M-280 streptavidin (STA; Thermo Fisher Scientific, Waltham, MA, USA) at 4°C overnight. The beads were washed 3 times with radioimmunoprecipitation assay buffer (50-mM Tris–HCl pH8.0, 150-mM NaCl, 1% NP-40, 0.5% sodium deoxycholate, 0.1% SDS). The beads were finally incubated in 2× SDS sample buffer at 70°C for 10 min, and supernatants were collected and analyzed by IB using anti-FLAG M2 antibody HPR.

Phosphoproteome analysis

For phosphoproteome analysis, full-open flowers were incubated at 16°C for 30 min in the dark and then incubated at 22°C for 60 min in the light or dark. To isolate crude membrane fraction, the samples were homogenized in a permeabilization buffer of Mem-PER Plus Membrane Protein Extraction Kit (Thermo Fisher Scientific, Waltham, MA, USA) containing a phosphatase inhibitor cocktail (PhosSTOP; Roche, Basel, Switzerland) and a protease inhibitor cocktail (Roche, Basel, Switzerland). Homogenized samples were incubated on ice for 10 min with constant mixing. After centrifugation at 16,000 × g for 15 min at 4°C, the pellet was collected and stored at −80°C for subsequent analysis.

Crude proteins extracted from the membrane fractions of the Gentiana flowers using the phase-transfer surfactant method (Masuda et al., 2008) were subjected to reductive alkylation, followed by successive digestion with Lys-C endopeptidase and trypsin as previously described (Niinae et al., 2021), and then desalted using a Stage tip (Rappsilber et al., 2007) with an SDB-XC Empore disk (3M). Fifty microgram of each digested sample was applied for hydroxy acid-modified metal oxide chromatography (Sugiyama et al., 2007) for phosphopeptide enrichment with slight modifications as described. (Sugiyama et al., 2019) The enriched phosphopeptides were isotopically labeled with TMT 16-plex (Thompson et al., 2019) (Thermo Fisher Scientific, Waltham, MA, USA) using a nanoscale solid-phase labeling method (Ogata et al., 2021).

The labeled samples were analyzed with nanoscale liquid chromatography coupled to tandem mass spectrometry (nano-LC/MS/MS), using an Orbitrap Fusion Lumos mass spectrometer (Thermo Fisher Scientific, Waltham, MA, USA) in data-dependent acquisition mode, coupled to Ultimate3000 RSLCnano system (CTC Analytics, Zwingen, Switzerland) and HTC-PAL autosampler (CTC). The raw MS data files were analyzed by MaxQuant version 1.6.17.0. Peptides and proteins were identified by means of automated database searching using Andromeda against Gentiana protein database derived from RNA-seq data with a false discovery rate of ˂1% at peptide spectrum match level. The peak area of each peptide in MS1 was quantified using MaxQuant and normalized based on the band intensities of GsPIP2;2 in IB. Details regarding the analytical conditions for nano-LC/MS/MS and data processing were the same as previously described (Niinae et al., 2021) except for the fact that minimum peptide length was set to 6 this time. The MS raw data and analysis files have been deposited to the ProteomeXchange Consortium (http://proteomecentral.proteomexchange.org) via the jPOST partner repository (http://jpostdb.org) with the data set identifier JPST001350, PXD029169 (Okuda et al., 2017).

Ca²⁺-imaging using G-CaMP3

The gentian corollas were cut and transferred so that the abaxial sides were in contact with the solid MS medium. The transient expression vector p35SΩ-GCaMP3-NOST was inoculated into the adaxial sides of the corolla by the bombardment method using the PDS-1000/He particle delivery system (Bio-Rad, Hercules, CA, USA), as described previously (Nakatsuka et al., 2015). After overnight incubation at 22°C under continuous light, samples were sandwiched between a slide glass and a cover glass and incubated at 16°C in the dark and then incubated at 22°C under continuous light or in the dark. The GCaMP3 signal was acquired every 5–10 min using a FLUOVIEW FV1000 confocal laser scanning microscope (Ex 488 nm/Em 510–550 nm) (Olympus Co., Tokyo, Japan) and analyzed using microscopy software. To analyze the fluorescence intensity, the same laser, pinhole, and gain settings were maintained between the processes. The fluorescence intensity of GCaMP3 (F) was measured using ImageJ software and subtracted from its background. The fluorescence intensity was calculated as a relative value when the fluorescence intensity measured from the cells before stimulation (F0) was one.

Fluorescence imaging of cell walls and cuticles

The gentian corolla was cut by hand to a thickness of ∼300 μm using a razor blade. The sections were fixed and cleared in fix solution (ethanol:acetic acid = 6:1) for 1 h. The sections were then washed twice with 70% ethanol and once with phosphate buffered saline (PBS). To visualize cellulose, the sections were stained with 0.05% (w/v) Direct Red 23 in PBS for 2 min and washed twice with PBS. After washing, stained sections were mounted with 50% glycerol/PBS solution and observed using a FLUOVIEW FV1000 confocal laser scanning microscope (autofluorescence, Ex 488 nm/Em 510–550 nm; Direct Red 23, Ex 543 nm/Em 600–660 nm). Fluorescence intensity was quantified using the microscopy software.

Water transport assay

Two types of plasmids, cDNA-containing or empty pYES2/CT and pYES3/CT (Thermo Fisher Scientific, Waltham, MA, USA), were combined and co-transformed into the S. cerevisiae strain INVSc-1 (Thermo Fisher Scientific, Waltham, MA, USA) using the LiAc/ssDNA/PEG method (Akada et al., 2000). Yeast cells were selected on synthetic complete (SC) medium (0.67% yeast nitrogen base without amino acids [BD Difco, Franklin Lakes, NJ, USA], 0.077% dropout supplements without uracil and tryptophan [Clontech, Mountain View, CA, USA]) with 2% glucose and 2% agar. All transformed yeast cells were cultured in SC medium with 2% raffinose at 30°C for 24 h. For protein expression, cultured cells were inoculated in SC medium with 2% galactose and 1% raffinose so that the optical density at 600 nm (OD₆₀₀ nm) was 0.4, and incubated at 30°C for 16 h. After centrification, cells were washed with 50-mM sodium phosphate buffer (pH 7.6) and resuspended in enzyme solution [1.2-M sorbitol, 50-mM sodium phosphate buffer (pH 7.6), 0.7-mM β-mercaptoethanol, and 10 U Zymolyase-20T (Nacalai Tesque, Kyoto, Japan)], and incubated for 1 h at 30°C. Following centrification, spheroplasts were washed and resuspended in assay buffer (1.2-M sorbitol and 50-mM sodium phosphate buffer [pH 7.6]). One hundred microliters of spheroplast solution were transferred to a spectrophotometer cuvette cell (10-mm path length), and the optical density at 600 nm was measured immediately using a NanoDrop 2000c (Thermo Fisher Scientific, Waltham, MA, USA) every 2 s after adding an equal volume of 50-mM sodium phosphate buffer (pH 7.6). The rate of decrease in OD_600 nm after hypotonic shock was defined as the water transport activity and expressed as a relative value with the starting OD_600 nm value of 1.0. For the inhibitory assay, spheroplasts were preincubated in assay buffer containing 200-μM HgCl₂ for 20 min. The data were obtained by three to six independent experiments and their rate constants were calculated by curve fitting with a two-phase exponential decay function applied using GraphPad Prism version 8 software. The graph displayed is the result of curve fitting by a two-phase exponential decay function from three to six trace records. First-stage rate constant as k₁ and the second-stage rate constant as k₂ for the fitted curves were calculated, and k₁ and k₂ are represented by relative values to the control.

Statistical analysis

The quantitative values obtained in the figures were analyzed using GraphPad Prism version 8 and Excel spreadsheets with the embedded basic statistical functions. Data sets were analysed using one-way ANOVA with Tukey’s multiple comparisons test, Brown–Forsythe and Welch’s ANOVA test with Dunnett’s T3 multiple comparisons test, Friedman’s test with Dunn’s multiple comparisons test, two-way ANOVA with Dunnett’s multiple comparisons test, two-way ANOVA with Bonferroni’s multiple comparisons test, Welch’s t test, and two-tailed unpaired t test. The results for statistical significance tests are included in the legend of each figure, and n values represent the number of independent experiments performed or the number of samples. Details of the statistical analysis results are shown in Supplemental Data Set 1.

Accession numbers

Raw Illumina data for Showtime Starlet were downloaded from the DDBJ DRA accession number DRA011818. Sequence data from this article have been deposited with the DDBJ/EMBL/GenBank Data Libraries under accession Nos LC627266–LC627275 and LC699353. The MS raw data and analysis files have been deposited to the ProteomeXchange Consortium (http://proteomecentral.proteomexchange.org) via the jPOST partner repository (http://jpostdb.org) with the data set identifier JPST001350, PXD029169.

Supplemental data

The following materials are available in the online version of this article.

Supplemental Figure S1. Sequence alignment of Gentian PIP2;1, PIP2;2, PIP2;5, and PIP2;7.

Supplemental Figure S2. Specificity analysis of anti-phospho Ser280 and anti-PIP2-1-7 antibody.

Supplemental Figure S3. Immunoblot analysis of phosphorylated GsPIP2;2 in gentian corolla.

Supplemental Figure S4. MS/MS spectrum of the phosphopeptides derived from the C-terminus of GsPIP2;2.

Supplemental Figure S5. Purification and activity analysis of recombinant GsCPKs.

Supplemental Figure S6. Amino acid sequence of GsCPK16.

Supplemental Figure S7. Analysis of GsCPK16, GsPIP2;1, and GsPIP2;7 expression levels in gentian corolla.

Supplemental Figure S8. Water transport assay in S. cerevisiae in the presence of HgCl₂.

Supplemental Figure S9. Quantitative RT-PCR of GsCPKs in corolla of vector control, ALSV:::GsCPK16-infected plants 30 days after infection.

Supplemental Table S1. Summary of de novo assembly of RNA-seq data.

Supplemental Table S2. Summary of PIPs genes expressed in gentian “Bzc-1” corolla from RNA-seq data.

Supplemental Table S3. Summary of CPK genes expressed in gentian “Bzc-1” corolla from RNA-seq data.

Supplemental Table S4. List of PCR primers.

Supplemental Table S5. List of plasmids for cell-free expression, transient expression, and ALSV-mediated VIGS.

Supplemental Data Set 1: Details of statistical analysis.

 

K.N. conceived the study, designed the experiments, and wrote the manuscript. K.N. performed almost all experiments and analyzed data; F.G. and A.W. did the plasmid construction and transformation. T.N. and N.S. performed phosphoproteome analysis. M.S. performed RNA-seq analysis. K.N. and K.S. antibody validation. M.N. critically read and reviewed the manuscript.

The authors responsible for distribution of materials integral to the findings presented in this article in accordance with the policy described in the Instructions for Authors (https://academic.oup.com/plcell) is: Keiichirou Nemoto (k-nemoto@ibrc.or.jp).