The rice RING finger E3 ligase, OsHCI1, drives nuclear export of multiple substrate proteins and its heterogeneous overexpression enhances acquired thermotolerance.

Thermotolerance is very important for plant survival when plants are subjected to lethally high temperature. However, thus far little is known about the functions of RING E3 ligase in response to heat shock in plants. This study found that one rice gene encoding the RING finger protein was specifically induced by heat and cold stress treatments but not by salinity or dehydration and named it OsHCI1 (Oryza sativa heat and cold induced 1). Subcellular localization results showed that OsHCI1 was mainly associated with the Golgi apparatus and moved rapidly and extensively along the cytoskeleton. In contrast, OsHCI1 may have accumulated in the nucleus under high temperatures. OsHCI1 physically interacted with nuclear substrate proteins including a basic helix-loop-helix transcription factor. Transient co-overexpression of OsHCI1 and each of three nuclear proteins showed that their fluorescent signals moved into the cytoplasm as punctuate formations. Heterogeneous overexpression of OsHCI1 in Arabidopsis highly increased survival rate through acquired thermotolerance. It is proposed that OsHCI1 mediates nuclear-cytoplasmic trafficking of nuclear substrate proteins via monoubiquitination and drives an inactivation device for the nuclear proteins under heat shock.

Introduction

Extreme temperature is a major agricultural problem limiting crop yields worldwide. A transient increase in temperature, usually 10–15 °C above ambient, is generally considered heat shock or heat stress in living organisms, particularly in plants. Heat shock negatively affects plant growth, seed germination, photosynthesis, respiration, water relation, and membrane stability in plants (Wahid ). At the cellular and molecular level, heat shock leads to adverse outcomes in plant cell functions, including alterations in cellular composition of membrane fluidity and permeability, enzyme activity, metabolism, production of active oxygen species, and gene expression (Kampinga ; Alfonso ; Larkindale and Knight, 2002; Larkindale and Huang, 2004; Larkindale ). These alterations could cause reduced photosynthesis and carbon gain in plants, thereby leading to decreased growth and reproduction. For example, studies on the relationship between rice crop yields and temperature over the last two decades have demonstrated that grain yields decrease significantly by 10% for each 1 °C increase in the growing-season minimum temperature (Peng ).

Investigations into molecular mechanisms underlying thermoprotection have involved genetic and molecular approaches (Iba, 2002; Sung ; Ahuja ; Qin ). Plants generally possess basal and acquired thermotolerance by two heat tolerance mechanisms (Vierling, 1991). Basal thermotolerance is defined as an inherent ability to survive high temperatures, whereas acquired thermotolerance is the ability to tolerate an otherwise lethally high temperature after being pre-exposed to a sublethal increased temperature, mimicking an ‘immunization’ against high temperature. Once plants are exposed to high temperature, either basal, acquired, or both, thermotolerance mechanisms may be involved (Larkindale ).

One of the best-known mechanisms regarding acquired thermotolerance is the induction of heat shock proteins (HSPs; Vierling, 1991). HSPs are molecular chaperone stress-response proteins that protect organisms against various stresses, particularly high temperature. HSPs preserve structural and functional protein integrity by binding to proteins that have become denatured or misfolded as a result of heat shock (Perez ; Sarkar ). Plant adaptations to high temperature are not only HSP-based mechanisms but also other components such as phospholipids, the dehydration-responsive element binding protein 2A (DREB2A), and S-nitroglutathione reductase (GSNOR) (Ahuja ). For example, the heat stress transcription factor HsfA3 which is transcriptionally induced during heat shock by DREB2A, regulates the expression of an HSP-encoding gene (Schramm ). Furthermore, filamentous temperature-sensitive H11 protease and GSNOR activity contribute to plant adaptation to high temperatures (Chen ; Lee ).

Attachment of ubiquitin molecules (Ub, a small 76-amino acid protein) to target substrates for modification mediates a variety of cellular functions via the Ub/26S proteasome system in higher plants. In this pathway, the conjugation cascade subsequently requires three classes of enzymes: E1 (ubiquitin-activating enzyme), E2 (ubiquitin-conjugating enzyme), and E3 (ubiquitin ligase) (Vierstra, 2009). Approximately 5% of the Arabidopsis proteome is postulated to be involved in the Ub/26S proteasome pathway, and about 1300 genes are predicted to encode E3 ligase components in particular (Smalle and Vierstra, 2004). The E3 ligases specifically interact with target proteins to confer different fates by attachment of ubiquitin molecules. The substrate–ubiquitin structures determine the subcellular localization and different functions of many target proteins (Hicke and Dunn, 2003; Pickart, 2004; Roos-Mattjus and Sistonen, 2004). For example, attachment of single ubiquitin molecules to one or more lysines on target proteins, known as monoubiquitination or multimonoubiquitination, activates a variety of their functions, for example trafficking, subcellular localization, signal transduction, transcription regulation, and DNA repair (Deng ; Kaiser ; Hicke and Dunn, 2003; Wu ). In contrast, polyubiquitinated substrate proteins destined for degradation are usually targeted by the 26S proteasome (Roos-Mattjus and Sistonen, 2004; Vierstra, 2009).

The Ub/26S proteasome pathway is an important mechanism of tolerance against high temperature. For example, seedlings of Prosopis chilensis, which is a leguminous tree, are able to survive at 50 °C after germination at 35 °C (Medina and Cardemil, 1993). P. chilensis showed higher relative accumulation rates of free Ub, conjugated Ub, and HSP70 than cultivated Glycine max (soybean) under heat stress, suggesting that the ubiquitinated-proteolytic pathway is an important heat tolerance mechanism (Ortiz and Cardemil, 2001). In addition, small ubiquitin-like modifiers (SUMOs) that are ubiquitin-like polypeptides also attach to various target substrates and, thus, modify their cellular functions. In Arabidopsis, the findings that SUMO1/2 conjugates were highly accumulated by repeated heat shock, while HSP70-overexpressing plants showed fewer SUMO1/2 conjugates during heat shock, suggested that the accumulation of SUMO1/2 conjugates is relevant to thermotolerance (Kurepa ).

Plant single-subunit E3 ligases are generally classified into three groups based on the presence of the homologous E6-AP C-terminus, U-box, and RING domain (Smalle and Vierstra, 2004). Of these, the RING domain of the really interesting new gene was the first to be identified as a novel cysteine-rich sequence (Freemont ). The proteins harbouring a RING domain are believed to play E3 ligase for recognizing and ubiquitylation of substrate proteins. Subsequently, a number of RING E3s have been reported to play crucial roles in post-translational regulation of plant hormone signalling pathways, for example abscisic acid (ABA), and environmental stresses. For example, the RING E3 ligase ABI3-interacting protein 2 is a negative regulator of ABA signalling by promoting degradation of ABSCISIC ACID-INSENSITIVE 3 (ABI3; Zhang ). Another outstanding example is the KEEP ON GOING E3 ligase, which also regulates the protein level of ABI5, a basic domain/leucine zipper transcription factor, by 26S proteasomal degradation in an ABA-dependent manner (Stone ). The Arabidopsis RING E3 ligases DREB2A-interacting protein 1 and 2 negatively modulate the expression of drought stress-response genes (Qin ). Hot pepper RING membrane-anchor 1 homologue 1 (Rma1H1) functions as an E3 ligase plasma membrane aquaporin, PIP2;1, under water-deficient conditions (Lee ). In Arabidopsis, the high expression of osmotically responsive gene 1 (HOS1) harbouring a RING-like domain negatively regulates cold signal transduction (Lee ). Additionally, salt- and drought- induced ring finger 1 E3 ligase is believed to enhance salt stress-responsive ABA signalling (Zhang ). However, RING E3 ligase and its substrate proteins on heat shock response via ubiquitination still remain unknown in plants.

This study identified the molecular functions of a rice RING domain E3 ligase, OsHCI1 (Oryza sativa Heat and Cold Induced 1), which is highly induced under heat and cold stress conditions. Studies with a Golgi-localized OsHCI1-EYFP fusion protein showed that OsHCI1 dynamically moved from the cytoplasm to the nucleus along cytoskeletal tracts under heat shock conditions. To shed light on the molecular function of this gene, this study performed a yeast-two hybrid (Y2H) screen, a bimolecular fluorescence complementation (BiFC) assay, and an in vitro ubiquitination assay. The results demonstrated that OsHCI1 interacted with six substrate proteins and mediated subcellular trafficking of nuclear proteins to the cytoplasm via monoubiquitination. Furthermore, Arabidopsis overexpressing OsHCI1-EYFP exhibited a heat-tolerant phenotype, suggesting an important role of this protein in the regulation of heat-generated signals in plants.

Materials and methods

Plant materials and heat shock treatments

Seeds of rice (O. sativa L. cv. Donganbyeo) were grown on mesh supported in plastic containers with 1/2 Murashige and Skoog (MS) nutrient solution in a growth chamber (16/8 light/dark cycle at 25 °C with 70% relative humidity). Two-week-old seedlings were exposed to high salinity (250mM NaCl), dehydration, cold (4 °C), and heat (45 °C). The high-salinity and dehydration stress treatments were performed as described by Lim . Two-week-old seedlings were transferred to fresh MS nutrient solution with each of ABA (0.1mM), jasmonic acid (0.1mM), and salicylic acid (1mM). For ethylene treatments, seedlings were moved into air-tight plastic containers with fresh MS solution for the ethylene treatment. Ethylene gas (50 μl l–1) was injected into the plastic boxes using a syringe (Wuriyanghan ). Leaf tissues were sampled at 0, 1, 6, 12, 24, and 48h after the stress treatment. Healthy samples without stress treatment were harvested as controls at the same times. All leaf samples were ground using liquid nitrogen and immediately stored at –80 °C until total RNA extraction.

Dry seeds of A. thaliana ecotype Columbia were grown, and two constructs of 35S:EYFP (EV) and 35S:OsHCI1-EYFP were transformed via Agrobacterium tumefaciens GV3101 using the floral dip method (Zhang ). The assessment of segregation of kanamycin resistance in T3 transformants was conducted less than 1 month after harvest for the assay. Three independent OsHCI1-overexpressing lines and a control plant (35S:EYFP) were tested according to Larkindale to observe the heat shock effect. Transgenic Arabidopsis plants were grown on MS agar for 7 days, then dipped into water baths at either 38 or 45 °C, as appropriate. Basal thermotolerance treatments were performed by heating the plants in sealed plates at 45 °C for 1h. Acquired thermotolerance treatments were conducted by heating the plants initially to 38 °C for 90min, and then they were moved to a growth chamber (24 °C) for 120min before finally heating to 45 °C for 3h. Both heat shock treatments were performed in the dark. Heat-treated plants were recovered in a growth chamber at 24 °C for 5 days in the light.

To evaluate the expression patterns of six interacting protein genes with OsHCI1, rice plants were grown on MS agar for 14 days. Then basal or acquired heat treatments were performed as described above. Leaves were sampled at different time points using liquid nitrogen and immediately stored at –80 °C until total RNA extraction.

Rice protoplast isolation and transfection

Protoplasts were isolated from 2-week-old seedlings (Kim ). Seeds of rice were grown on 1/2 MS nutrient solution in a growth chamber (16/8 light/dark cycle at 25 °C with 70% relative humidity). Young leaves and sheaths were chopped and dipped in enzyme solution (0.5M mannitol, 1.5% cellulose RS (Yakult Honsa, Tokyo, Japan), 0.75% mecerozyme R10 (Yakult Honsa), 1mM CaCl2, and 0.1% BSA) with carbenicillin (100mg l–1). This mixture was incubated on a shaking incubator for 16h at room temperature then filtered through Miracloth. Protoplasts were pelleted by centrifugation for 4min at 300 g and resuspended in an equal volume of W5 solution (154mM NaCl, 125mM CaCl2, 5mM KCl, 5mM glucose, and 1.5mM MES, adjusted to pH 5.7) and incubated on ice for 5h. Protoplasts were centrifuged and resuspended in MMg solution (0.4M mannitol, 15mM MgCl2, and 4.7mM MES, adjusted to pH 5.7). Plasmid DNA (10 or 20 μg) was added to the protoplast solution and transfected with 40% polyethylene glycol (PEG) solution (40% PEG 4000, 0.4M mannitol, and 100mM Ca (NO3)2) for 20min at room temperature. W5 solution was added stepwise to dilute the PEG solution and discarded. Transfected protoplasts were incubated overnight at room temperature and then observed under confocal microscopy.

Gene expression study

Total RNA was extracted using TRIzol regent, according to the manufacturer’s protocol (Invitrogen, Carlsbad, CA, USA). First-strand cDNA synthesis from 500ng total RNA was conducted using a cDNA Synthesis kit (Takara-Bio, Ohtsu, Japan). Semi-quantitative reverse-transcription polymerase chain reaction (RT-PCR) was performed as described previously (Lim ). Gene-specific primers were designed using Primer-BLAST (NCBI, www.ncbi.nlm.nih.gov/tools/primer-blast/). Reliable genes such as OsSalT for salt stress (Claes ), OsbZIP23 for dehydration (Xiang ), LIP19 for cold (Shimizu ), and OsHsp90-1 for heat (Hu ) were used as positive controls to verify the stress treatments, respectively. Quantitative real-time PCR was performed with a Rotor-Gene Q (Qiagene, USA) by monitoring the SYBR Green fluorescence signal during DNA synthesis. The real-time PCR results were calculated by using the Delta-Delta CT Method (Livak and Schmittgen, 2001). Os18S-rRNA (Os09g00999) was used as an internal control (Kim ). Primers with restriction enzyme sites used in this study are listed in Supplementary Table S1 (available at JXB online).

Yeast two-hybrid (Y2H) screening and Y2H assays

A full-length coding sequence of OsHCI1 was amplified and cloned in-frame with the GAL4 DNA binding domain of the GBKT7-BD vector to generate the GAL4 DNA-BD fusion construct. A rice cDNA library was generated from 14-day-old seedlings treated with salt stress (250mM NaCl). Then, yeast transformation and library screening were conducted in accordance with the recommended procedures (Make Your Own ‘Mate & Plate’ Library System; Matchmaker Gold Yeast Two-Hybrid System; Yeastmaker Yeast Transformation System 2, Clontech, Palo Alto, CA, USA). The full-length OsHCI1 coding sequence was fused to the yeast GAL DNA-binding domain and used as a bait protein for screening. A rice cDNA library from salt-treated seedlings was fused to the yeast GAL4 activation domain as a prey protein. A total of 280 yeast transformants were selected on a synthetic defined medium lacking Leu and Trp supplemented with 40 μg ml–1 X-α-Gal and 70ng ml–1 aureobasidin A (AbA) (DDO/X/A) and repatched on synthetic defined medium lacking Ade, His, Leu, and Trp with 40 μg ml–1 X-α-Gal and 70ng ml–1 AbA (QDO/X/A).

Six full-length interaction partners were amplified by RT-PCR using primers listed in Supplementary Table S1 to confirm a positive interaction with OsHCI1. Each PCR product was digested with the appropriate restriction enzyme and introduced into the pGADT7 vector. These constructs with pGBKT7-OsHCI1 were co-transformed into the Y2H Gold yeast strain. Transformed yeast cells were separately grown onto synthetic defined/–Leu/–Trp and synthetic defined/–Ade/–His/–Lue/–Trp/X-α-Gal/Aba with 70ng ml–1 AbA for 5 days at 30 °C. All experiments were repeated three times.

Subcellular localization

Two fluorescence protein constructs were prepared for the subcellular localization assay. For the 35S:EYFP and 35S:DsRed2 constructs, the coding sequence of the EYFP and DsRed2 were amplified using a high-fidelity Pfu Turbo DNA polymerase (Stratagene, La Jolla, CA, USA) from pEYFP-C1 and pDsRed2-C1 (Clontech) as templates, respectively, with primers harbouring multiple cloning sites (Supplementary Table S1). The PCR products then were cloned into the pBIN35S binary vector under the control of the CaMV 35S promoter. The coding region of the full-length cDNA of OsHCI1 was amplified from rice cDNA with appropriate primer pairs and then inserted into the pBIN35S-EYFP vector between the XbaI and KpnI sites for the subcellular localization study. A single amino acid substitution (OsHCI1C172A) in the RING domain of OsHCI1 was generated using the QuikChange Site-Directed Mutagenesis kit (Stratagene) with the OsHCI1C172A-F and OsHCI1C172A-R primer pair. Additionally, full-length cDNAs of the six OsHCI1 interacting partners were cloned into the pBIN35S-DsRed2 vector with appropriate enzyme sites, respectively. The plasmid containing an organelle marker for the Golgi apparatus (Nelson ) was kindly provided by the Arabidopsis Biological Resource Center.

BiFC assay

The full-length OsHCI1 cDNAs and the six interacting partners were amplified by PCR using appropriate primers to generate BiFC constructs. PCR products were digested and then ligated into 35S-HA-SPYCE(M) and 35S-c-myc-SPYNE(R)173 vectors, respectively (Waadt ). The primers and restriction enzymes used for cloning are presented in Supplementary Table S1. A. tumefaciens GV3101 harbouring each construct was inoculated for 16h at 28 °C for transient expression. These cells were harvested and resuspended in infiltration buffer (10mM MES, 10mM MgCl2, 0.2mM acetosyringone, pH 5.6) to a final concentration at an optical density at 600nm of 0.5. Equal volumes of different combinations of the Agrobacterium strains were mixed and coinfiltrated into 5-week-old Nicotiana benthamiana leaves with a syringe. Infiltrated plants were placed at 25 °C for 3 days to detect YFP fluorescence.

In vitro ubiquitination assay

Full-length OsHCI1 cDNA was amplified by PCR with primer pairs (Supplementary Table S1). The amplicon was digested with NotI and BamHI and then ligated into a digested pMAL-c5X vector (New England BioLabs, Ipswich, MA, USA) with the same enzymes. Recombinant MBP-OsHCI1, MBP-OsHCI1C172A, and non-recombinant MBP (negative control) were expressed in Escherichia coli BL21 (DE3) pLysS (Promega, Madison, WI, USA), purified by affinity chromatography using amylose resin (New England BioLabs), and used for the in vitro self-ubiquitination assay. The full-length cDNAs of AtUBC10 and AtUBC11 were amplified and then introduced into the pET-28a(+) vector (Novagen, Gibbstown, NJ, USA) with a 6×His-tag. The fusion 6×His-tagged AtUBC10 and AtUBC11 were expressed in E. coli BL21 (DE3) pLysS and purified using the Ni-NTA Purification System (Invitrogen).

The in vitro self-ubiquitination assay was conducted as described previously by Hardtke with some modifications. Purified MBP-OsHCI1 (250ng) was mixed with 50ng yeast E1 (Boston Biochemicals, Cambridge, MA, USA), 250ng purified Arabidopsis E2 (AtUBC10 and AtUBC11), and 10 μg bovine ubiquitin (Sigma-Aldrich, St. Louis, MO, USA) incubated in ubiquitination reaction buffer (50mM Tris-HCl, pH 7.5, 10mM MgCl2, 0.05mM ZnCl2, 1mM ATP, 0.2mM DTT, 10mM phosphocreatine, and 0.1 unit of creatine kinase (Sigma-Aldrich)). After 3h incubation at 30 °C, the reaction was halted at different time points by adding 2×SDS sample buffer followed by 5min of boiling at 95 °C. Each reaction (10 μl) was analysed via 12% SDS-PAGE and then transferred to a nitrocellulose membrane. Immunoblot analyses were conducted using anti-ubiquitin antibody (Santa Cruz Biotechnology, Santa Cruz, CA, USA) with a secondary goat anti-rabbit IgG peroxidase antibody (Sigma-Aldrich). Detection was conducted using the chemiluminescent substrate SuperSignal West Pico (Thermo Scientific, Waltham, MA, USA) for HRP and imaged on X-ray film (Kodak, Rochester, NY, USA). To confirm that OsHCI1 mediated ubiquitination of the six interacting proteins, OsPGLU1, OsbHLH065, OsGRP1, and OsPOX1-His-Trx fusion proteins were affinity-purified, and 200ng purified protein was incubated together with purified MBP-OsHCI1 in the ubiquitination mixture for 3h. The mixture was then subjected to 10% SDS-PAGE and immunoblot analysis.

Confocal microscopy and imaging

Transformed tobacco leaves were cut 3–5 days after infection for microscopic analyses. Fluorescent images were obtained using a Multiphoton confocal laser scanning microscope (model LSM 510 META NLO and LSM 780 NLO, Carl Zeiss, Oberkochen, Germany) at the Korea Basic Science Institute, Chuncheon Center. Excitation/emission wavelengths were 514/535 590nm for EYFP and BiFC constructs and 543/565 615nm for the DsRed2 and mCherry construct. All images were acquired using either a C-Apochromat (×40/1.2 water immersion) objective. To prevent cross-talk between EYFP and mCherry (or DsRed2) signals, the spectral images were acquired using the lambda mode. Scanned images were captured as single optical sections or as a z-series of optical sections. Image processing was carried out using an LSM 5 Image Browser (Zeiss) and Photoshop 9.0 software (Adobe, Mountain View, CA, USA).

Results

OsHCI1 is upregulated by heat and cold

This research group previously defined expression diversity of members of the rice RING finger protein genes based on their expression profiles via in silico analysis (Lim ). Subsequently, in an effort to isolate RING finger protein gene(s) that play a critical role in extreme temperature, 48 RING finger protein genes were randomly selected and examined for their expression patterns via semi-quantitative RT-PCR (data not shown). Interestingly, one gene (Os10g30850) was highly induced at 1–48h after heat treatment (45 °C), whereas OsHsp90-1 used for validation of the stress treatment was similarly induced by the stress (Hu ) (Fig. 1A). Consequently, the gene was named O. sativa heat and cold inducible gene 1 (OsHCI1). The expression patterns of the gene were further examined against other abiotic stresses such as cold (4 °C), salinity, and dehydration (Fig. 1A). The gene was upregulated at 12–48h by cold stress, whereas LIP19 (Shimizu ) was induced at 1–48h. However, both the salinity and dehydration stresses exhibited no induction of the gene through 48h after the treatments. Two reliable stress-inducible genes were employed as quality control, OsSalT (Claes ) and OsbZIP23 (Xiang ) for salinity and dehydration, respectively. High induction of OsSalT and OsbZIP23 served as evidence that the plants had been subjected severe stresses, supporting no response of OsHCI1 to either stress. The transcript levels of OsHCI1 were further confirmed via quantitative real-time PCR, which revealed high expression patterns under heat and cold stresses but not under salt and drought stresses (Supplementary Fig. S1).

Fig. 1.

Expression levels of OsHCl1 in rice plants subjected to four abiotic stresses and four hormonal treatments. Rice seedlings were subjected to abiotic stresses and plant hormone treatments. (A), heat (45 °C), cold (4 °C), NaCl (250mM), and dehydration (dehydration on two pieces of tissues paper); OsHsp90-1, LIP19, OsSalT, and OsbZIP23 were used as reliable stress-inducible genes for each abiotic stress treatment, respectively. (B) 0.1mM abscisic acid (ABA), 0.1mM jasmonic acid (JA), 1mM salicylic acid (SA), 50 μl l–1 ethylene (Ethyl); OsSalT, OsPBZ1, OsPR1b, and OsERF3, were used as reliable genes for each hormonal treatment, respectively. C and T indicate untreated control and stress-treated samples, respectively. The experiments were performed with three biological replicates.

When plants are subjected to heat shock, phytohormones including ABA, salicylic acid, and ethylene act as key signals (Larkindale and Knight, 2002). Therefore, this study further examined phytohormonal regulation during OsHCI1 gene expression (Fig. 1B). Under 0.1mM ABA treatment, OsHCI1 was induced at 3h, the highest transcript level occurred at 12h, and then gradually decreased to 48h, whereas OsSalT exhibited an increase at 3h and then steady expression until 48h. In the case of jasmonic acid, OsPBZ1, which is inducible by hormone treatment (Lee ), showed a slight induction at 3h and then a subsequent increase up to 48h, whereas OsHCI1 exhibited a somewhat slight induction at 3–24h. In addition, OsHCI1 gene expression increased at 3h, reached its highest transcript level at 12h, and then showed no induction until 48h. However, OsPR1b (Agrawal ) was induced at 6h and gradually and slightly increased until 24h. Additionally, the transcription level of OsHCI1 under 50 μl l–1 ethylene treatment increased at 6h then reached its highest level at 12h. Collectively, the OsHCI1 expression patterns under phytohormonal treatments were induced gradually at 12h then its transcript levels decreased until 24h. These results indicate that OsHCI1 rapidly responds to hormone treatments.

Dynamics of OsHCI1-EYFP subcellular localization

It is generally believed that subcellular localization of a protein of interest is crucial to understand its cellular function. To examine subcellular localization of the OsHCI1 protein, this study constructed a binary vector harbouring the enhanced yellow fluorescence protein (EYFP) under the control of a CaMV 35S promoter. Transient expression of 35S:EYFP was diffuse in both the cytosol and the nucleus in tobacco epidermal cells (Fig. 2A, upper panel). This study further generated a 35S:OsHCI1-EYFP construct, which is transiently expressed in tobacco leaves. OsHCI1 fluorescence displayed a punctuate pattern; the fluorescence appeared to localize in the dispersed organization of Golgi stacks in most (about 93%) tobacco cells (Fig. 2A, lower panel). In contrast, about 7% of the transformed tobacco cells showed an additional reticulate fluorescence with a punctate pattern, which seemed to target endoplasmic reticulum network patterns (Supplementary Fig. S2A, C). To confirm whether the destination of the OsHCI1 protein alone was the Golgi apparatus, this study employed the G-rk-mCherry organelle marker localized to the Golgi body. Both constructs, OsHCI1-EYFP and G-rk-mCherry, were transiently co-expressed with p19 in tobacco cells (Nelson ). The OsHCI1-EYFP signal was closely overlapped by that of G-rk-mCherry (Fig. 2D), indicating that the final destination of OsHCI1 was the Golgi complex. The endoplasmic reticulum localization may represent newly synthesized OsHCI1-EYFP protein that has not yet been transported to the Golgi stack. Furthermore, the punctate patterns of OsHCI1-EYFP fluorescence were also displayed around the nuclear envelope (Fig. 2A lower panel and D). In addition, there was dynamic movement in which the OsHCI1-EYFP fluorescent signals moved rapidly and extensively along the cytoskeleton of leaf epidermis cells (Supplementary Fig. S2B and Supplementary Movie S1).

Fig. 2.

Subcellular localization of the OsHCI1-EYFP fusion protein. Each construct was transiently expressed in tobacco leaves and rice protoplasts. Images were captured and merged by single or z-series optical sections. Arrowheads indicate the position of the nucleus. (A–C) Transiently expressed 35S:EYFP and 35S:OsHCI1-EYFP fusion protein were expressed in tobacco leaves and incubated at 25 (A), 38 (B), or 45 °C (C) for 1h; nuclear staining was performed with Hoechst 33258. (D) Co-localization of OsHCI1-EYFP and the G-rk-mCherry-Golgi marker in tobacco leaves. (E) Transiently expressed 35S:EYFP and 35S:OsHCI1-EYFP fusion protein were expressed in rice protoplasts and incubated at 25, 38, or 45 °C for 15min. The green and red colours represent EYFP and chlorophyll autofluorescent signals, respectively; arrows indicate punctuate spots of OsHCI1-EYFP fluorescence; dotted line outlines the cell shape. (F) Quantification of OsHCI1-EGFP localization patterns under different temperatures. Protoplasts were counted based on their localization patterns: the Golgi-only pattern, the nucleus (NC)-only pattern, and the Golgi plus NC patterns. Sixty Golgi-localized rice protoplasts were counted at different temperatures.

The finding that the subcellular localization of fluorescently tagged fusion proteins is changed by environmental stress (von Arnim and Deng, 1994; Lee ) led to the question as to whether the Golgi localization of the OsHCI1 protein could be altered by heat shock. Thus, 35S:EYFP and 35S:OsHCI1-EYFP was transiently expressed in tobacco leaves, which were then incubated for 1h at 38 or 45 °C. Interestingly, strong OsHCI1-EYFP signals were found in the nucleus (Fig. 2B and C, lower panel). Because of a concern that heterogeneous expression of OsHCI1 caused protein mislocalization and functional diversity, the constructs were subsequently expressed in rice protoplasts, which were then incubated for 15min at 38 or 45 °C, resulting in a similar expression pattern compared to that of tobacco (Fig. 2E). These results support the previous finding (i.e. there is no significant difference in protein localization between tobacco and rice cells). Under moderate heat treatment (38 °C), approximately 55.0% of cells exhibited a nuclear localized pattern of the OsHCI1-EYFP protein and approximately 35.0% of cells displayed this pattern in both the Golgi and nucleus. However, approximately 10.0% of cells still showed only the Golgi-localized pattern of rice protoplasts (Fig. 2F). Similarly, approximately 66.6% and 21.6% of cells showed nuclear localization and both the Golgi and nucleus pattern, respectively, under severe heat treatment (45 °C). By contrast, approximately 11.6% of cells only displayed a Golgi-localized pattern at 45 °C.

Expression pattern and subcellular localization of proteins interacting with OsHCI1

A Y2H screen was performed to identify proteins that interact with OsHCI1. Twenty-four positive clones were selected, sequenced, and their α-galactosidase activity was measured (Supplementary Fig. S3). To confirm these positive interactions with OsHCI1, full-length coding sequences of the top six genes, which exhibited strong α-galactosidase activity, were cloned into GAL4 activation domain, respectively. Full-length OsHCI1 and each interacting protein were co-transformed into the Y2H Gold strain and grown on QDO/X/A medium (Supplementary Fig. S4). The six interacting protein genes were 20S proteasome subunit α7 (named OsPSA7, Os01g59600), periplasmic beta-glucosidase (OsBGLU1, Os03g53800), ethylene-responsive protein (OsbHLH065, Os04g41570, Li ), glycine-rich cell-wall structural protein (OsGRP1, Os05g02770), peroxidase (OsPOX1, Os07g48020), and 14-3-3 protein (Os14-3-3, Os11g34450).

This study also examined the expression patterns of the interacting partner genes with OsHCI1 under two different heat stresses via semi-quantitative RT-PCR with rice seedlings treated by basal or acquired heat shock treatments (Fig. 3). The results showed that the OsHCI1 transcript was highly induced by basal heat treatment (45 °C for 24h) and its transcript level was downregulated when seedlings were recovered at 24 °C for 2h. For acquired heat shock treatment, OsHCI1 was slightly induced by mild heat treatment at 38 °C for 90min and downregulated at 24 °C for 2h. The OsHCI1 transcript was highly accumulated by re-heat shock treatment at 45 °C for 24h and downregulated by 24 °C for 2h.

Fig. 3.

Expression patterns of the response of interacting protein genes with OsHCI1 under heat treatment. Two-week-old rice seedlings were exposed to basal (A) or acquired heat stress (B) and then placed to normal temperature for 2h. Each leaf sample was harvested at different time points.

Next, the expression patterns of the six interacting partner genes were evaluated under both heat shock conditions (Fig. 3). OsPSA7 and Os14-3-3 transcript levels were stable under both heat stress conditions. In contrast, OsPGLU1, OsbHLH065, OsGRP1, and OsPOX1 showed strikingly decreased transcript levels at 45 °C during the basal and acquired heat treatments. In addition, OsPGLU1 and OsbHLH065 displayed a slight decrease at 38 °C. Expression patterns of OsPGLU1, OsbHLH065, and OsGRP1 were likely to have a reverse correlation with that of OsHCI1. Three genes, OsPGLU1, OsbHLH065, and OsGRP1, were downregulated at 45 °C during basal and acquired heat treatments; their transcript levels were upregulated following the recovery period at 24 °C. These results suggest that heat shock results in high expression of the OsHCI1 transcript or protein, which can affect the transcript levels of its interacting genes.

This study questioned the subcellular localization of each OsHCI1 interaction partner that showed dynamic subcellular localization. The six interaction partners were tagged with DsRed2, and each construct was transiently expressed together with p19 in tobacco leaves. A series of DsRed2 signal z-stack images were captured and merged after 5 days of agro-infiltration. As shown in Fig. 4, DsRed2 fluorescent signals of the OsPGLU1, OsbHLH065, and OsGRP1 proteins were only associated with the nucleus, whereas OsPSA7-DsRed2 was found in both the cytoplasm and the nucleus. In contrast, OsPOX1-DsRed2 was observed in a punctuate/dot form pattern (Fig. 4E), and Os14-3-3-DsRed2 was localized to the cytoplasm and cytoskeleton (Fig. 4F).

Fig. 4.

Subcellular localization of six interacting proteins. The full-length OsPSA7 (A), OsPGLU1 (B), OsbHLH065 (C), OsGRP1 (D), OsPOX1 (E), Os14-3-3 (F), and empty EYFP (G) were tagged with DsRed2 and transiently expressed with p19 in Nicotiana leaves. Images were captured and merged by z-series optical sections after 5 days of agro-infiltration.

Subcellular localization of the complex of OsHCI1 and each interacting protein

BiFC technology was employed to visualize the interactions between OsHCI1 and each of the interaction partners in living cells (Waadt and Kudla, 2008). Full-length coding sequences of OsHCI1 and each of the six interacting protein genes were cloned into the 35S-HA-SPYCE(M) and 35S-c-myc-SPYNE(R)173 vectors, respectively. After 5 days of agro-infiltration, we observed YFP signals of all BiFC complex formations in tobacco cells. All of the YFP signals except that of OsPSA7 appeared to associate with the cytoplasm and nucleus (Fig. 5); however, the OsPGLU1-, OsbHLH065-, and OsGRP1-DsRed2 alone protein signals were detected only in the nucleus (Fig. 4B–D). In contrast, the OsHCI1 BiFC complex with OsPSA7 was localized to the cytoplasm with a punctuate complex (Fig. 5A).

Fig. 5.

BiFC assay for six substrate proteins confirms the interaction with OsHCI1 in living cells. Full-length OsPSA7 (A), OsPGLU1 (B), OsbHLH065 (C), OsGRP1 (D), OsPOX1 (E), and Os14-3-3 (F) were cloned into pSPYNE(R) and OsHCI1 was cloned into pSPYCE(M). Combinations of each construct and SPYNE(R):empty (G, negative control) with OsHCI1:SPYCE(M) were transiently expressed with p19 in Nicotiana leaves. Images were captured and merged by z-series optical sections after 5 days of agro-infiltration.

OsHCI1 functions as an E3 ligase and mediates ubiquitination of interacting proteins

OsHCI1 encoded a 246-amino acid protein with a predicted molecular mass of 28.8kDa and harboured a single RING-HC domain in its C-terminal region (Supplementary Fig. S5). It is generally believed that many proteins harbouring the RING-HC domain function are Ub E3 ligases (Stone ). An in vitro ubiquitination assay was used to test whether the OsHCI1 protein has E3 Ub ligase activity. A purified MBP-OsHCI1 fusion protein was mixed with ubiquitin, ATP, yeast E1 activating enzyme, and Arabidopsis E2 conjugating enzymes (AtUBC10 and AtUBC11) and then incubated at 30 °C for 3h. An immunoblot analysis with anti-Ub showed that ubiquitinated proteins were detected in the presence of all of these components (Fig. 6A). Furthermore, clearer ubiquitinated proteins were observed in the presence of the AtUBC10 enzyme but not AtUBC11 (Fig. 6A, lanes 6 and 7). In time-course experiments, MBP-OsHCI1 began to cause high-molecular-mass ubiquitinated ladders after 30min that gradually reached their highest level after 2h incubation (Fig. 6B). However, no ubiquitinated ladders were found at 0h. These results suggest that the OsHCI1 protein possesses E3 ligase activity in the presence of E1 and E2 enzymes.

Fig. 6.

OsHCI1 functions as an E3 ubiquitin ligase and mediates OsPGLU1, OsbHLH065, OsGRP1, and OsPOX1 protein ubiquitination in vitro. (A) E3 ligase activity of OsHCI1 in vitro. Maltose-binding protein-tagged OsHCI1 fusion protein was assayed for E3 ligase activity in the presence of yeast E1, Arabidopsis E2s (AtUBC10 and AtUBC11), and Ub. (B) MBP-OsHCI1 was incubated for the indicated time periods in the presence of yeast E1, E2 (AtUBC10), ATP, and Ub. Ubiquitinated proteins were detected by immunoblot analysis using an anti-Ub antibody. (C–F) OsHCI1 mediates the ubiquitination of OsGLU1 (C), OsbHLH065 (D), OsGRP1 (E), and OsPOX1 (F) proteins. The full-length OsGLU1, OsbHLH065, OsGRP1, and OsPOX1 genes were cloned into His and Trx tags pET-32a (+) vector (Novagen) and these purified fusion proteins were used as the substrate for the assay. Anti-Trx was used in the immunoblot analysis for detecting His-Trx-tagged substrate proteins.

The six interaction proteins were fused with His and Trx tags to determine whether OsHCI1 mediated ubiquitination of the six interacting proteins. The recombinant fusion proteins were expressed in E. coli BL21 (DE3) pLysS. However, His- and Trx-tagged OsPSA7 and Os14-3-3 fusion proteins were not expressed well in this E. coli system. Therefore, an in vitro ubiquitination assay was conducted with OsPGLU1, OsbHLH065, OsGRP1, and OsPOX1 as substrates. In the presence of E1, E2, and MBP-OsHCI1 as E3 ligases, an additional higher-molecular-weight band was detected by anti-Trx immunoblot analysis (Fig. 6C–F). Interestingly, nuclear-localized OsPGLU1, OsbHLH065, and OsGRP1 proteins had one additional ubiquitin monomer, whereas the OsPOX1 protein had polyubiquitinated chains on the original fusion protein bands. Collectively, the OsHCI1 protein was a functional E3 ligase and, mediated multiple substrate mono- and polyubiquitination.

OsHCI1 translocates nuclear substrate proteins into the cytoplasm

The findings that the Golgi-localized OsHCI1 protein relocated to the nucleus along the cytoskeleton under heat shock and that it mediated monoubiquitination of each of the three nuclear-localized substrates in an in vivo ubiquitination assay led to the hypothesis that OsHCI1 E3 translocates its substrate proteins for heat-stress regulation. To test this hypothesis, this study first investigated whether nuclear substrate proteins of OsHCI1 could be relocated by themselves under a heat shock condition in tobacco leaves. Nuclear localization of the OsPGLU1-, OsbHLH065-, and OsGRP1-DsRed2 signals was not significantly different between normal and heat shock conditions (Fig. 4 and Supplementary Fig. S6). Next, a single amino acid substitution (OsHCI1C172A) in the RING domain of OsHCI1 was generated to obtain a non-functional RING E3 ligase. MBP-OsHCI1C172A did not show self-ubiquitination activity in vitro (Supplementary Fig. S7). In addition, subcellular localization of OsHCI1C172A-EYFP was highly similar to that of wild-type OsHCI1-EYFP in tobacco leaves under normal and heat shock conditions (Supplementary Fig. S8). Combinations of each of the OsHCI1-EYFP, OsHCI1C172A-EYFP, and empty-EYFP constructs were transiently co-expressed with OsbHLH065-DsRed2 in tobacco leaves. The fluorescence signal of OsbHLH065-DsRed2 was detected in the cytoplasm and in the nucleus when co-expressed with OsHCI1-EFYP under heat shock and normal conditions (Fig. 7 and Supplementary Fig. S9). In contrast, no alterations in their subcellular localizations were observed when the OsHCI1C172A-EYFP or empty-EYFP construct was co-expressed (Fig. 7B and C). Furthermore, co-expression of OsPGLU1- and OsGRP1-DsRed2 with OsHCI1-EYFP showed the same patterns of fluorescence signals as OsbHLH065-DsRed2 in the cytoplasm and in the nucleus (Supplementary Fig. S10).

Fig. 7.

OsHCI1 protein mediates nuclear–cytoplasmic trafficking in tobacco leaves. Full-length OsbHLH065-DsRed2 fusion proteins were transiently co-expressed with wild-type OsHCI1-EYFP protein (A), OsHCI1C172A-EYFP protein (B), or empty-EYFP construct (C) in tobacco leaves. Tobacco leaves were incubated at 45 °C for 1h after 5 days of agro-infiltration. Images were captured and merged by z-series optical sections. Nuclear staining was performed with Hoechst 33258. Arrowheads indicate the position of nucleus.

This study questioned whether regulation of dynamic translocation under heat shock misleads through heterogeneous expression. In an effort to verify the mechanism in rice cells, the nuclear-localized OsbHLH065-DsRed2 was transformed in rice protoplasts. The fluorescent signal of OsbHLH065-DsRed2 was associated with the nucleus under normal conditions (Fig. 8A). However, the signal displayed both the nucleus and cytoplasm as punctuate formations under heat shock (Fig. 8B). Subsequently, OsbHLH065-DsRed2 and OsHCI1-EYFP were co-transformed into rice protoplasts and then by heat shock. The OsHCI1-EYFP fluorescence clearly moved from the cytoplasm to the nucleus, whereas the OsbHLH065-DsRed2 signal was displayed in both the nucleus and cytoplasm (Fig. 8C).

Fig. 8.

OsHCI1 protein mediates nuclear–cytoplasmic trafficking of the OsbHLH065 transcription factor in rice protoplasts. (A) Full-length OsbHLH065-DsRed2 was transfected into rice protoplasts. (B) Transformed protoplasts were incubated at 45 °C for 15min. (C) Full-length OsbHLH065-DsRed2 was transfected with OsHCI1-EYFP into the rice protoplast and the transformed protoplasts were incubated at 45 °C for 15min. Nuclear staining was performed with Hoechst 33258. Arrows indicate the nuclear-exported OsbHLH065.

OsHCI1-overexpressing Arabidopsis enhances heat shock tolerance

The distinct induction of OsHCI1 expression by temperature extremes and the dynamics of its subcellular translocation under heat treatment conditions suggest a crucial role of the gene in thermotolerance. To test this possibility, several independent Arabidopsis transgenic lines (T3) were developed with strong OsHCI1 gene expression and compared to plants without the gene (35S:EYFP), which served as controls (Fig. 9A). Plants were tested for basal heat treatment by heating directly to 45 °C for 1h, which resulted in no recovery (0%) in all tested control lines, whereas transgenic lines showed approximately 4–7% of survival rates at 5 days after treatment (Fig. 9B). For acquired heat treatment, the plants were subjected to heating to 38 °C for 90min and subsequently cooled for 2h at room temperature (24 °C). After pretreatment, plants were subjected to heating to 45 °C for 3h and then allowed to recover for 5 days at 24 °C (Fig. 9B). The OsHCI1-overexpressing lines showed strikingly high survival rates of approximately 55–65%; however, most control plants did not recover (Fig. 9C).

Fig. 9.

Thermotolerance phenotype of 35S:OsHCI1-EYFP. Seven-day-old seedlings of 35S:EYFP (empty vector, EV) and 35S:OsHCI1-EYFP T3 transgenic plants (three independent lines) were grown on agar in the light for 7d and heated to 38 °C for 90min, cooled 24 °C for 2h, then heated to 45 °C for 3h (acquired thermotolerance) or heated to 45 °C for 60min (basal thermotolerance). A, RT-PCR analysis of seven independent Col-0/35S:OsHCI1 T3 transgenic plants, control wild type (WT), and EV. (B) The phenotypes of EV and three independent OsHCI1-overexpressed plants were treated at various high temperatures. Images were captured 5 days after heat shock treatment. (C) Percentage of surviving plants relative to the control (EV) on the same plate was determined 5 days after heat shock. Bars indicate standard deviation from the mean over all experiments (n = 30). The experiments were performed with four biological replicates.

Discussion

This study’s findings regarding dynamic movement of OsHCI1 under heat shock, translocations of target proteins co-expressed with OsHCI1, and acquired thermotolerance via heterogeneous overexpression might provide some clues regarding a new molecular mechanism for the heat stress-regulated RING E3 ligase. RING E3 ligases have been recently reported as major players in plant responses to environmental stresses. For example, HOS1 RING E3 ligase is a negative regulator of plant cold responses by mediating degradation of ICE1, which binds the CBF promoter and induces its transcription (Dong ), and Rma1H1 RING E3 ligase functions in the downregulation of plasma membrane aquaporin levels as a response to drought stress (Lee ). However, the role of RING E3 ligases in the adaptation to heat shock in plants has remained largely unknown.

The finding that OsHCI1 gene expression patterns were specifically and somewhat rapidly increased by heat and cold stresses but not by salt and drought stresses indicates that the gene is associated closely with thermal stress in rice (Fig. 1A and Supplementary Fig. S1). Subcellular localization of OsHCI1 was mainly associated with the Golgi apparatus and these punctuate signals rapidly moved to the nucleus under heat shock (Fig. 2). Wild-type OsHCI1-EYFP expression effectively moved its nuclear target substrate proteins to the cytoplasm (Figs. 7 and 8 and Supplementary Fig. S10) and attachment of the ubiquitin molecule on the nuclear substrates by OsHCI1 fusion protein via in vitro ubiquitination assay might support this translocation of nuclear substrate proteins to the cytoplasm (Fig. 6). In addition, heterogeneous overexpression of OsHCI1 in Arabidopsis resulted in rising survival rates through acquired heat treatment (Fig. 9). These results suggest that the OsHCI1 E3 ligase might function in the heat shock response in plants.

A hypothesis regarding E3 ligase translocation for functional activation might be postulated by several findings. For example, the COP1 RING E3 ligase is localized to the nucleus in the dark but translocates to the cytoplasm under light signals (von Arnim and Deng, 1994; Deng ). Similarly, the Arabidopsis HOS1 protein exhibits nucleo-cytoplasmic partitioning in response to cold stimuli (Lee ). Recently, two alternative splicing forms of Arabidopsis XBAT35 RING E3 ligase have been reported that display dual targeting of this E3 ligase to the nuclear and cytoplasmic compartment, suggesting a novel player in ethylene-mediated regulation of the apical hook curvature (Carvalho ). The OsHCI1 protein, whose localization is confined to the Golgi stack under control conditions, accumulated in the nucleus in response to heat shock (Fig. 2). In addition, OsHCI1 protein interacts with substrate proteins localized in both the nucleus and the cytoplasm and relocates nuclear substrate proteins to the cytoplasm (Figs. 4, 5 and 7). These findings suggest that the nucleo-cytoplasmic partitioning of E3 ligases is an extensive regulatory mechanism to control cellular responses to environmental stimuli. However, the OsHCI1 protein can also interact with its substrate proteins and relocates them to the cytoplasm under both normal and heat shock conditions (Fig. 5 and Supplementary Fig. S9). It is possible that overexpression of OsHCI1 might lead to interaction with its nuclear proteins under normal conditions. However, further studies are necessary to test this possibility.

Plants and other organisms have the intrinsic ability to acquire thermotolerance for survival under lethally high temperatures. It is generally known that the ability accelerate transcription and translation of HSPs and decrease normal protein synthesis (Vierling, 1991; Barnabas ). Thus, translational modifications of transcription factors might be necessary to decrease synthesis of normal proteins under heat shock. A number of studies regarding the transcriptional regulation of targets via post-translational modification of transcript factors, such as ABI3, ABI5, DREB2A, and ICE1 by E3 ligases and 26 proteasomes have been reported (Zhang ; Dong ; Stone ; Qin ). Similarly, this study provides evidence to support that OsHCI1 interacts with multiple substrates including the OsbHLH065 transcription factor with a basic helix-loop-helix transcription factor, which is highly downregulated by heat shock treatment (Fig. 3). Interestingly, the transcription levels of three nuclear-targeted partner genes displayed a significant decrease following overexpression of OsHCI1 in rice protoplasts (Supplementary Fig. S11). These results lead to the hypothesis that OsHCI1 plays a crucial role in the thermotolerance mechanism via post-translational modifications. Significant future work on target protein degradation by OsHCI1 via the 26S proteasome is warranted.

A large body of evidence demonstrates the role of E3 ligase in the differential control of mono- versus polyubiquitination of target proteins. For example, ubiquitination by Mdm2, an oncogenic E3 ligase, causes two alternative p53 fates depending on Mdm2 levels. When Mdm2 levels are high, Mdm2 drives p53 degradation via polyubiquitination, whereas low levels promote p53 nuclear exclusion via monoubiquitination (Li ). Human Nedd4-1, an E3 ligase, catalyses monoubiquitination of hDCNL1, which drives its nuclear export (Wu ). The current study observed that OsHCI1 drives two different ubiquitination types depending on the target proteins (Fig. 6). In addition, co-expression of each of three nuclear-localized proteins and wild-type OsHCI1 promoted nuclear export of target proteins to the cytoplasm, while non-functional OsHCI1C172A did not affect (Fig. 7). Collectively, the findings suggest that OsHCI1 may mediate a nuclear–cytoplasmic translocation of nuclear target substrates via monoubiquitination, demonstrating an inactivation device of nuclear proteins in this compartment under heat shock (Li ). An alternative hypothesis is that translocation of the target proteins drives another cellular program to mediate thermotolerance mechanisms in plant cells (Mihara ). However, much work is needed to rule out this hypothesis. Why OsHCI1 drives different ubiquitination processes depending on the target protein localization is a mystery. A simple hypothesis may be that different fates of the target proteins exist under heat shock.

The finding that heterogenous OsHCI1 overexpression in Arabidopsis enhances heat shock tolerance suggests that this gene is involved in acquired thermotolerance (Fig. 9). An outstanding report suggested that the protection mechanism against heat-induced oxidative damage involves phytohormones such as ABA, salicylic acid, and ethylene in Arabidopsis (Larkindale and Knight, 2002). As shown in Fig. 1B, phytohormone treatment (i.e. ABA) causes a rapid increase in OsHCI1 transcripts, which suggests that the gene is related to the ABA-dependent pathway involved in temperature stress responses (Yamaguchi-Shinozaki and Shinozaki, 2006). Furthermore, induction of OsHCI1 by salicylic acid and ethylene treatments might be consistent with the previously reported relationship among salicylic acid, ethylene, and thermotolerance (Dat ; Wang and Li, 2006). This study tested whether OsHCI1 is related to the ABA-dependent pathway involved in acquired thermotolerance. However, constitutive expression of OsHCI1 did not confer sensitivity or insensitivity to ABA during seed germination, cotyledon greening, or root growth (Supplementary Fig. S12), suggesting that the OsHCI1 E3 Ub ligase is involved in the ubiquitination of unidentified proteins, which might function in the heat response in transgenic Arabidopsis plants in a ABA-independent manner.

This study demonstrated the specific expression patterns of the OsHCI1 transcript and dynamic movement of OsHCI1-EYFP under normal and heat shock conditions. In addition, OsHCI1 functions as an E3 ligase that mediated ubiquitination of substrate proteins in vitro. OsHCI1-overexpressing Arabidopsis showed higher tolerance than control plants under heat shock conditions. These results demonstrate that accumulation of the OsHCI1 RING E3 ligase by heat shock mediates nuclear–cytoplasmic trafficking of nuclear substrate proteins via monoubiquitination to improve heat tolerance as an inactivation mechanism. The results are an excellent example of the post-translational regulation of the heat tolerance mechanism via the Ub/26S proteasome system in plant cells.

Supplementary material

Supplementary data are available at JXB online.

Supplementary Table S1. Primer list.

Supplementary Fig. S1. Quantitative real-time PCR analysis of OsHCI1 in 2-week-old rice plants subjected to heat, cold, NaCl, and dehydration.

Supplementary Fig. S2. Subcellular localization of the OsHCI1-EYFP fusion protein.

Supplementary Fig. S3. Positive clones from yeast two-hybrid screening.

Supplementary Fig. S4. Identification of OsHCI1 interaction with six proteins.

Supplementary Fig. S5. Sequence analysis of OsHCI1.

Supplementary Fig. S6. Subcellular localization of nuclear localized OsPGLU-, OsbHLH065-, and OsGRP1-DsRed2 fusion proteins under heat shock.

Supplementary Fig. S7. The ubiquitination reaction contains E1, E2 (Arabidopsis UBC10), MBP-OsCTR1, Ub, and ATP.

Supplementary Fig. S8. Subcellular localization of the OsHCI1C172A-EYFP fusion protein.

Supplementary Fig. S9. OsHCI1 protein mediates nuclear–cytoplasmic trafficking of OsbHLH065 at normal temperature.

Supplementary Fig. S10. OsHCI1 protein mediates nuclear–cytoplasmic trafficking in tobacco leaves.

Supplementary Fig. S11. Expression patterns of interacting protein genes in overexpression OsHCI1-EYFP in rice protoplast.

Supplementary Fig. S12. Phenotypes of 35S:EYFP and 35S:OsHCI1-EYFP plants in response to different concentrations of ABA during seed germination and seedling growth.

Supplementary Movie S1. Dynamic movement of the Golgi-localized OsHCI1-EYFP fusion protein along the actin cytoskeleton.