Light-stabilized GIL1 suppresses PIN3 activity to inhibit hypocotyl gravitropism

Expression pattern and subcellular localization of GIL1

To elucidate the expression pattern of GIL1, we constructed a pGIL1:GUS transgenic plant. In both darkness and red light, GIL1 was ubiquitously expressed in cotyledon, hypocotyl and root (Figure 1A, B). Further, we compared the expression of GIL1 under darkness, red light, far-red light and blue light using real-time polymerase chain reaction (PCR). Using 1.5×-fold change as the threshold (Darwish et al., 2022), the transcriptional levels of GIL1 after different light treatments were not obviously different compared to under darkness (Figure 1C). Compared to Col wild type, gil1-2 exhibited significantly better negative gravitropism when the seedlings were grown vertically in red light, but not in dark, far-red light, or blue light conditions (Figure S1). Additionally, gil1-2 showed an enhancement of negative hypocotyl gravitropism when exposed to far-red light pulses (Figure S2). The phenotypes of gil1-2 under red light and far-red light are consistent with previous results (Allen et al., 2006). Further studies were carried out mainly in red light, in which GIL1 showed the most obvious effect on gravitropism. GIL1-GFP driven under its native promoter was expressed in the gil1-2 mutant, and the transgene rescued the mutant phenotype under red light (Figure 1D, E). Previous studies have showed that mutation of red/far-red light receptors phyA and phyB resulted in better negative gravitropism of hypocotyls in red light (Poppe et al., 1996; Robson and Smith, 1996). We compared the gravitropic phenotype of hypocotyls in Col, gil1-2, pGIL1:GIL1-GFP/gil1-2, phyA, phyB and phyA phyB seedlings in parallel under red light conditions (Figure 1D, E). The gravitropic phenotype of gil1-2 is similar to that of phyA phyB double mutant, suggesting that GIL1 plays a crucial role in the RHO induced by red light.

Consistent with the β-glucuronidase (GUS) expression results, the fluorescence signal of GIL1-GFP (green fluorescent protein) was widely detected in cotyledon, hypocotyl and root (Figure S3). We stained pGIL1:GIL1-GFP/gil1-2 seedlings with the membrane dye FM4-64 and found that GIL1-GFP co-localized with FM4-64 on the plasma membrane of hypocotyl cells (Figure 1F). Due to the low expression of GIL1-GFP under its native promoter, we generated GIL1-GFP driven by heat shock promoter pHSP18.2, which significantly increased the GIL protein level (Figure 1G). We preformed cell fractionation experiments using pHSP18.2:GIL1-GFP/gil1-2 seedlings and found that GIL1-GFP was specifically detected in the membrane pellet (Figure 1H). Moreover, pHSP18.2:GIL1-GFP/gil1-2 seedlings were treated with the vesicle trafficking inhibitor, Brefeldin A (BFA), and GIL1-GFP fluorescence was found to be aggregated in BFA bodies (Figure 1I). When the BFA was washed out, GIL1-GFP signal in BFA bodies almost disappeared (Figure 1I). Together these results suggest that GIL1 is a membrane-localized protein expressed widely in plant organs, and that its subcellular localization is mediated by the BFA-sensitive secretion pathway.

GIL1 inhibits the negative gravitropism of hypocotyls in endodermal cells of the upper part of hypocotyls

To identify the localization of GIL1 essential for its function in inhibiting negative hypocotyl gravitropism, we generated transgenic lines expressing GIL1-GFP under the control of different promoters in the gil1-2 mutant background. Based on previous studies, pCAB3, pPKS4 and pCASP1 promoters were selected to drive GIL1-GFP, which are mainly expressed in the cotyledons, upper part of hypocotyls and lower part of hypocotyls, respectively (Endo et al., 2007; Schepens et al., 2008; Roppolo et al., 2011). Although the fluorescence of GIL1-GFP in these three transgenic lines was too low to be detected, we found that transformation with GIL1 driven by the PKS4 promoter complemented the gravitropic phenotype of gil1-2, whereas GIL1 driven by CAB3 and CASP1 promoters did not (Figure 2A, B). These results indicate that GIL1 expressed in the upper part of hypocotyls can inhibit the negative hypocotyl gravitropism under red light. Endodermal cells are essential for the gravitropism of hypocotyls, and scr and shr mutants showed defects in forming endodermal cells and lost the response to gravistimulation (Tasaka et al., 1999). To investigate the function of GIL1 in endodermal cells, we generated transgenic plants expressing pSCR:GIL1 in the gil1-2 mutant background. The pSCR:GIL1/gil1-2 plants exhibited reduced gravitropic hypocotyl responses compared to gil1-2 under red light (Figure 2C, D), indicating that GIL in endodermal cells can inhibit the negative gravitropism of hypocotyls. Together, our data suggest that the function of GIL1 during inhibition of hypocotyl gravitropism is dependent on its localization in endodermal cells of the upper part of hypocotyls.

GIL1 interacts with PIN3 and inhibits its activity to interrupt gravitropism

To identify the possible working mechanism of GIL1, we studied whether GIL1 regulates the development of amyloplasts, which are involved in gravity sensing (Caspar and Pickard, 1989; Kiss et al., 1989). Previous studies have shown that PIFs sustain the starch accumulation in amyloplasts, which is very important for the negative gravitropism of hypocotyls (Shin et al., 2009; Kim et al., 2011). In red light, epidermal phyB induced the degradation of PIF1 and inhibited the negative gravitropic growth of hypocotyls (Kim et al., 2016). To determine whether GIL1 also inhibits negative hypocotyl gravitropism through the suppression of starch accumulation, we compared the amyloplast morphology in the hypocotyls of Col and gil1-2 by transmission electron microscope (TEM). We found that the morphology of amyloplasts is similar in the endodermal cells of the hypocotyls of Col and gil1-2 growing in red light (Figure S4), which indicates that GIL1 may not affect hypocotyl negative gravitropism by mediating the development of amyloplasts.

Further, we investigated whether GIL1 inhibits the gravitropism via targeting factors mediating auxin distribution. As an auxin efflux carrier, PIN3 is essential for the asymmetric auxin distribution controlling negative hypocotyl gravitropism. PIN3 is localized on the plasma membrane, which is mediated by BFA-sensitive membrane protein recycling (Rakusová et al., 2011; Rakusová et al., 2016). Also, GIL1 and PIN3 could colocalize on the plasma membrane in the hypocotyls of Arabidopsis (Figure 3A). Based on these similar cellular characteristics, we tested the direct interaction between GIL1 and PIN3. Yeast two-hybrid assay showed that GIL1 interacted with the hydrophilic PIN3 loop (PIN3-HL) (Figure 3B). In addition, in in vitro pull-down assays, the maltose-binding protein (MBP)-tagged PIN3-HL was able to pull down glutathione S-transferase (GST)-tagged full-length GIL1, whereas MBP alone did not (Figure 3C). To confirm this interaction in vivo, we extracted membrane proteins from 4-day-old pPIN3:PIN3-GFP/pHSP18.2:GIL1-mCherry and pPIN3:PIN3-GFP seedlings growing under red light and carried out a co-immunoprecipitation (Co-IP) assay. The results revealed that PIN3-GFP associates with GIL1-mCherry in plants (Figure 3D). Moreover, we performed a firefly luciferase (LUC) complementation imaging (LCI) assay. Co-expression of GIL1-nLUC with cLUC-PIN3-HL in tobacco (Nicotiana benthamiana) leaves resulted in high LUC activity, whereas control combinations showed no signal (Figure 3E). These results indicate that GIL1 directly interacts with PIN3.

To study a possible regulatory impact of GIL1 on PIN3 activity, we expressed PIN3-RFP (red fluorescent protein) and GIL1-GFP independently and together in tobacco protoplasts by using Agrobacterium-mediated leaf transfection (Henrichs et al., 2012), and quantified efflux of radiolabeled indolylacetic acid (IAA). Benzoic acid (BA) served as the diffusion control, and its export was unaffected by PIN3 and GIL1 (Figure 3F). As anticipated, PIN3 exhibited notably heightened auxin export activity in comparison to the vector control (Figure 3G). The co-expression of GIL1 alongside PIN3 led to comparable IAA export levels as those observed in the vector control, suggesting that GIL1 significantly diminished PIN3-facilitated IAA export activity (Figure 3G).

To analyze the genetic relationship between GIL1 and PIN3, we crossed gil1-2 with pin3-4 and analyzed the negative hypocotyl gravitropism of the gil1-2 pin3-4 double mutant. Upon gravistimulation in red light, gil1-2 and pin3-4 showed increased and decreased hypocotyl gravitropic responses compared to wild type, respectively. The double mutant gil1-2 pin3-4 exhibited slower kinetics of hypocotyl bending compared to gil1-2 in red light (Figure 3H, I), indicating that mutation of PIN3 suppresses gravitropic response of the gil1 mutant. These data suggest that GIL1 may work upstream of PIN3 in mediating gravitropic responses in hypocotyls.

Light stabilizes GIL1 protein

Since the effects of GIL1 on the negative gravitropism of hypocotyls were various under different light conditions (Figures S1, S2), while its transcriptional levels did not show obvious change under various light conditions (Figure 1C), we investigated whether the abundance of GIL1 protein is regulated by light conditions. By comparing to under darkness, the abundance of GIL1 protein was dramatically increased in seedlings grown in red or far-red light, whereas slightly increased in blue light (Figure 4A, B). The abundance of GIL1 is consistent with the gravitropic phenotype of gil1 mutant under these light conditions. During darkness to red light transition, GIL1 was gradually accumulated with prolonged red light treatment (Figure 4C). Further, we used the heat shock promoter driven GIL1-GFP (pHSP18.2:GIL1-GFP) to study the protein stability of GIL1. After heat shock at 37°C, GIL1 protein abundance in pHSP18.2:GIL1-GFP/gil1-2 seedlings could be accumulated in both darkness and under red light. When the heat shocked seedlings were placed back to regular temperature (22°C) for 5 h, most of the GIL1 protein in seedlings grown under darkness was degraded, whereas GIL1 proteins in seedlings grown in red light remained stable (Figure 4D). These data suggest that GIL1 protein is unstable in the dark, whereas light can enhance its stability. Together, these data indicate that light stabilizes GIL1 protein to inhibit the negative gravitropism of hypocotyls.

Plant materials and growth conditions

The Arabidopsis thaliana gil1-2 (SALK_041346) and pin3-4 (SALK_005544) are in the Col ecotype. Seeds were sterilized by 10% NaClO for 10 min, washed in sterile distilled water at least three times, and cold-treated at 4°C for 3 to 4 d in darkness. The seeds were sown on Murashige and Skoog (MS) medium (4.4 g/L MS, 0.56 g/L 2-[N-morpholino]ethanesulfonic acid, and 8 g/L agar, pH 5.7–5.9), and treated in white light (70 μmol/m²/s) at 22°C for 12 h to induce germination. Then, then seedlings were grown in darkness, red light (5 μmol/m²/s), far-red light (5 μmol/m²/s) or blue light (5 μmol/m²/s). Seedling emergence tests were conducted according to the methods previously described (Allen et al., 2006). Seeds of Col or gil1-2 were spread on the surface of soil or covered with 1 cm nutrient soil. The pots were placed at 4°C for 4 d in darkness and then transferred to a greenhouse. Seedling emergence was analyzed after 1 week.

Generation of transgenic plants

To generate the GIL1-GFP expressing construct, the full-length coding sequence (CDS) of GIL1 (without stop codon) and native promoter of GIL1 (2052 bp) were cloned and inserted into the Xba I/Xho I and Sbf I/Xba I sites of pJim19-GFP (Sun et al., 2016), yielding pGIL1:GIL1-GFP. Further, the GIL1 promoter was replaced by HSP18.2 promoter (720 bp), CAB3 promoter (1537 bp), PKS4 promoter (1606 bp), and CASP1 promoter (1144 bp), respectively. For the construction of pSCR:GIL1, CDS of GIL1 was inserted into the Xho I/Afl II sites of pJim19 vector, and SCR promoter (1627 bp) was digested by Sbf I/Xho I and ligated into pJim19-GIL1. All the promoter sequences were amplified from genomic DNA. To generate stable transgenic plants, these constructs in Agrobacterium tumefaciens GV3101 were transformed into gil1-2 plants via the floral dip method (Clough and Bent, 1998). Transgenic plants were selected on MS medium containing 20 mg/L Basta. To obtain pPIN3:PIN3-GFP/pHSP18.2:GIL1-mCherry (abbr.:PIN3-GFP/GIL1-mCherry), the construct of pHSP18.2:GIL1-mCherry was transformed into pPIN3:PIN3-GFP plants (Zadnikova et al., 2010).

β-glucuronidase histochemical assay

β-glucuronidase activity assays were performed following the previous protocol (Li, 2011). In brief, plants were submerged in 90% acetone for 20 min on ice and subsequently washed by staining buffer (50 mmol/L sodium phosphate buffer pH 7.2, 0.2% Triton X-100, 2 mmol/L K₄[Fe(CN)₆] 3H₂O, 2 mmol/L K₃[Fe(CN)₆]). X-Gluc was added into staining buffer to a final concentration of 2 mmol/L. Plants were immersed, vacuum infiltrated in the staining buffer for 1 h, and then incubated at 37°C for 24 h. After staining, chlorophyll was removed by immersing the seedlings in 70% ethanol for 30 min.

Measurement of hypocotyl negative gravitropism

For vertically growing seedlings, the angles between hypocotyl growth orientation and negative gravity vector were measured by ImageJ software (https://imagej.nih.gov/ij/index.html). For gravistimulation assay, seedlings were grown vertically in darkness for 2 d, and then transferred into red light and rotated 90° to the horizontal position. The angles between hypocotyl growth orientation and horizontal direction were measured by ImageJ software. One-way analysis of variance was used to assess the statistical significance.

Confocal microscopy

Fluorescence of GFP tagged protein was excited with a 488 nm laser and emission signal was collected in the 500–550 nm range using a Zeiss LSM 710 confocal microscope or spinning-disk confocal microscopy (Andor Dragonfly 200). FM4-64 was used to label the plasma membrane, which was excited with a 514 nm laser and the emission signal was collected in the 580–640 nm range. Brefeldin A treatment and wash out was performed as previously described (Huang and Zhang, 2020).

Membrane protein extracts

Arabidopsis seedlings were ground into powder in liquid nitrogen and homogenized with 1× EB buffer (100 mmol/L Tris-HCl pH 7.5, 10 mmol/L ethylenediaminetetraacetic acid pH 8.0, 10 mmol/L ethyleneglycoltetraacetic acid pH 8.0, 5 mmol/L KCl, 25% sucrose, 5% glycerol, 20 mmol/L disodium β-glycerophosphate, 20 mmol/L Na₃VO₄, 50 mmol/L NaF) containing 1 mmol/L phenylmethylsulfonyl fluoride and 1× cocktail of protease inhibitors. After incubation on ice for 5 min, the mixture was centrifuged at 600 g for 3 min at 4°C. The supernatant was further isolated by ultracentrifugation at 55,000 g for 1 h at 4°C to obtain pellet microsomes. Microsomal membranes were then fully resolved by 1× RIPA buffer (Abcam, Cambridge, UK) containing 1 mmol/L phenylmethylsulfonyl fluoride and 1× cocktail of protease inhibitors. After centrifugation at 850 g for 5 min at 4°C, the supernatant was transferred into a new tube and 1× sodium dodecyl sulfate (SDS) was added as a loading buffer for immunoblots.

Yeast two-hybrid assay

The yeast two-hybrid assays were carried out as previously described (Serino et al., 1999). The GIL1 CDS was subcloned into the EcoR I/Xho I sites of pLexA vector, and PIN3 hydrophilic loop (PIN3-HL, 157–493 a.a.) was subcloned into the EcoR I/Xho I sites of pB42AD vector. The bait and prey constructs were co-transformed into yeast strain EGY48[p8op-lacZ] (CLONTECH, Kyoto, Japan), and clones containing both constructs were selected and β-Galactosidase activity was tested on medium containing X-gal.

In vitro pull-down assay

The EcoR I/Xho I DNA fragment encoding the full-length GIL1 CDS was cloned into the pGEX-4T-1 vector. PIN3-HL was digested by EcoR I/Sal I and ligated into the pMAL-c2x vector. GST-GIL1 (2 μg) was mixed with MBP or MBP-PIN3-HL fusion proteins (2 μg) in 1 mL binding buffer (50 mmol/L Tris-HCl, pH 7.5, 150 mmol/L NaCl, and 1% Triton X-100) containing 1 mmol/L phenylmethylsulfonyl fluoride and 1× cocktail of protease inhibitors. The mixture was incubated at 4°C for 2 h. The MBP resin was washed three times with binding buffer and then added into the mixture. Another incubation was performed for 2 h at 4°C. MBP resin was washed three times with binding buffer and boiled with 1× SDS loading buffer to elute the binding proteins. Immunoblots were performed with anti-MBP and anti-GST antibodies.

Co-immunoprecipitation assay

The microsomal membranes were resolved by 1× RIPA buffer (added with 20 mmol/L disodium β-glycerophosphate, 20 mmol/L Na₃VO₄, 50 mmol/L NaF, 1 mmol/L phenylmethylsulfonyl fluoride and 1× cocktail of protease inhibitors). After centrifugation (850 g , 5 min, 4°C), the supernatant was transferred into a new tube and 20 μL prewashed anti-RFP-conjugated beads (ChromoTek, Martinsried, Germany) were added and incubated at 4°C for 6 h under red light. The anti-RFP-conjugated beads can bind mCherry-fused proteins. Eluted proteins were analyzed by immunoblotting using anti-GFP and anti-mCherry antibodies. Immunoblots were performed as previously described (Dong et al., 2014).

Luciferase complementation imaging assay

The CDS of GIL1 was inserted into BamH I/Sal I sites of pCAMBIA-nLUC vector (Zhou et al., 2018). PIN3-HL was inserted into Kpn I/Sal I sites of pCAMBIA-cLUC vector (Zhou et al., 2018). Nucleotides ATG was added into BamH I/Sal I sites of pCAMBIA-nLUC vector using Gibson assemble kit, yielding translatable ATGnLUC construct. nLUC- and cLUC-fused constructs were transformed into Agrobacterium strain GV3101 and the paired constructs were injected into 4-week-old N. benthamiana leaves. The plants were subsequently grown in darkness for 24 h and then transferred into red light for another 24-h growth. The LUC activity was analyzed by LB985 (Berthold Technologies, Bad wildbad, Germany).

Auxin export assay

Simultaneous ³H-3-indolylacetic acid (³H-IAA; specific activity 20 Ci/mmol; American Radiolabeled Chemicals, Inc., St. Louis, MO) and ¹⁴C-benzoic acid (14C-BA; specific activity of 50 Ci/mmol; American Radiolabeled Chemicals, Inc., St. Louis, MO) export from mesophyll protoplasts was analyzed as previously described (Henrichs et al., 2012). In short, tobacco (Nicotiana benthamiana) mesophyll protoplasts were prepared 4 d after Agrobacterium-mediated transfection of combinations of indicated plasmids or plasmid combinations. Relative export from protoplasts was calculated from exported radioactivity into the supernatant as follows: ((radioactivity in the supernatant at time t = 15 min) − (radioactivity in the supernatant at time t = 0)) × (100%)/(radioactivity in the supernatant at t = 0 min); presented are mean values from four to six independent protoplast preparations derived from independent transfections.