Introduction

Unfavorable environmental conditions, such as extremely high temperatures, high salinity, and drought, are currently a major problem causing serious plant growth and development defects, thereby leading to a reduction in crop production worldwide. For example, the total increase in average temperature between the periods of 1850–1900 and 2003–2012 ranged from a minimum of 0.72 °C to a maximum of 0.85 °C because of the climate change caused by global warming (Stocker et al. 2013). High temperatures are known to affect pollination and photosynthesis during plant developmental and ultimately have a significant effect on crop yield (Song et al. 2014; Hatfield and Prueger 2015). In addition, drought stress causes a reduction in leaf size, stem expansion, root growth, and crop yield by disrupting plant growth. The yields of maize (39.3%) and wheat (20.6%) were reduced when the water content of soil was reduced to approximately 40% (Daryanto et al. 2016). Similarly, in dried soil (− 20 kPa), rice yield was reduced to 22.6% compared to that in continuously flooded soil (Carrijo et al. 2017). In nature, plants are frequently exposed to a combination of drought and heat stresses, which affects diverse physiological functions, nutrient circulation and absorption, and availability of water to plants, leading to more damage than when exposed to either stress alone (Fahad et al. 2017). Similar to drought and heat stresses, salt stress is one of the major severe abiotic factors affecting crop growth and productivity. The negative effects of soil salinity are initially reduced osmotic potential and then restricted water uptake by roots, causing nutrient imbalance in plants (Oliveira et al. 2013). These osmotic stresses, including drought coupled with heat and salinity, can lead to loss of turgor, disorganized plant cell membranes, protein denaturation or loss of activity, and excessive production of reactive oxygen species (ROS) (Farooq et al. 2009).

In plants exposed to abiotic stress, the posttranslational regulation of target proteins is one of the key physiological processes (Krasensky and Jonak 2012; Stone 2014). The strategies to overcome stress in plants lead to changes in intracellular metabolic processes, such as protein synthesis and degradation and regulation of transcription factors. The ubiquitin–proteasome system (UPS) is known to alter the cellular protein content required for sustained growth, development, and adaptation by protein degradation to changing environmental conditions (Sharma et al. 2016). As a central component of the ubiquitination pathway, approximately 470 and 425 potential REALLY INTERESTING NEW GENE (RING) E3 ligases have been identified from the model plants Arabidopsis thaliana and rice, respectively (Vierstra 2003; Sophia et al. 2005; Lim et al. 2013). Generally, RING-type E3s recognize and bind to their specific substrates, and ubiquitin is transferred to the substrate proteins by RING E3s. Finally, the targeted proteins with a poly-ubiquitin chain are degraded via the 26S proteasome. A host of studies have documented the key roles of RING E3s in plant developmental processes and responses to abiotic stresses, including cold, drought, salinity, and heat. In Arabidopsis, the Salt and Drought-Induced RING Finger Protein 1 (SDIR) gene is upregulated under drought and salt stress conditions (Zhang et al. 2007). The overexpression of SDIR1 caused abscisic acid (ABA) hypersensitivity and salt hypersensitivity in germination and enhanced ABA-induced stomatal closing and drought tolerance. Similarly, the plants overexpressing Arabidopsis ABA-Insensitive RING Protein 4 (AtAIRP4) exhibited ABA-mediated drought tolerance, but were hypersensitive to salt and osmotic stresses. Park et al. (2018) reported that the Oryza sativa Salt-, ABA- and Drought-Induced RING Finger Protein 1 (OsSADR1) gene might act as a regulator of abiotic stress responses. The Arabidopsis plants overexpressing OsSADR1 showed sensitive seed germination and seedling growth phenotypes under salt and mannitol stress conditions.

Many abiotic stress-inducible genes are known to be controlled by ABA or other signaling pathways, suggesting that both ABA-dependent and ABA-independent processes are involved in the regulation of stress-responsive gene expression (Yamaguchi-Shinozaki and Shinozaki 2005; Yoshida et al. 2010, 2014). Stress-inducible transcription factors (TFs) play important roles in the regulation of stress response and tolerance (Nakashima et al. 2018). Abiotic stresses induce the production and/or activation of TFs, leading to the expression of target stress-inducible genes. In general, the accumulation of these target proteins could enhance stress tolerance in plants. Typically, the TF Abscisic acid-Responsive Element Binding Protein 1 (AREB1) involved in the ABA-dependent pathway binds to a conserved cis-element called the ABA-responsive element in the promoter of an ABA-inducible gene to regulate its expression (Sakuma et al. 2006a, b). By contrast, it is known that the TFs Dehydration-Responsive Element-Binding Protein 1A and 2A (DREB1A and DREB2A) involved in ABA-independent pathways could bind to the promoter region of the target genes containing the essential cis-elements with the core sequence (A/GCCGAC), which is called the dehydration-responsive element, leading to the activation of a protection mechanism in plants exposed to stress (Nakashima et al. 2018).

It is generally believed that the activity of proteases is stimulated by abiotic stresses in plants. Hundreds of proteases have been divided into different families based on their evolutionary relationships in plants. Among these proteases, papain-like cysteine proteases (PLCPs) and legumain-like cysteine proteases (LLCPs) are two well-known types of cysteine proteases that can be inhibited by small proteins called cystatins (Wang et al. 2018a, b). Previous studies have shown that there are 33 and 30 protein sequences of PLCPs and 5 and 4 protein sequences of LLCPs in rice and Arabidopsis, respectively (Beers et al. 2004; Nakaune et al. 2005). Plant cysteine proteases are involved in physiological processes, such as senescence, seed germination, and abiotic stress responses (Khanna-Chopra et al. 1999; Liu et al. 2018). Among these processes, senescence can be modulated by endogenous and exogenous factors, including plant hormones, darkness, abiotic stressors (cold, heat, drought, and salt), and wounding (Parrot et al. 2010; Martínez et al. 2012). In rice, the expression pattern of 38 cysteine proteases was analyzed, showing that most of them were regulated by plant hormones and in response to diverse stresses, including cold, drought, and salt (Wang et al. 2018a, b).

In the glyoxalase system, two enzymes, glyoxalase I (GLYI) and glyoxalase II (GLYII), have been reported to catalyze the detoxification of methylglyoxal (MG) to non-toxic d-lactate using reduced glutathione (GSH) (Zeng et al. 2016; Sankaranarayanan et al. 2017). MG is a well-known inhibitor of growth and development in plants. It induces oxidative stress in cells through increased ROS generation and leads to harmful effects in plants (Hoque et al. 2012). Plants have evolved the glyoxalase detoxification pathway as a cellular defense against MG and oxidative stress. Similar to cysteine proteases, the glyoxalase enzymes are believed to play key roles in plant stress responses, and their overexpression leads to significant tolerance to various abiotic stresses, including salt, drought, and heavy metal (Smits and Johnson 1981; Singla-Pareek et al. 2003; Zeng et al. 2016). Both cysteine proteases and glyoxalases correlate with the production of ROS in response to abiotic stresses; these two different pathways are necessary to support plant survival during abiotic stresses (Hoque et al. 2012; Sueldo and van der Hoorn 2017). Many previous studies have shown that E3 ligases regulate substrate proteins via the 26S proteasome pathway, but the mechanisms of E3 ligase-mediated regulation of cysteine proteases and glyoxalases are still unknown.

Previously, the expression patterns of 44 Oryza sativa RING finger proteins (OsRFPs), which were randomly selected based on the domain analysis, were examined under salt treatment conditions (Hwang et al. 2016). In the present study, we characterized the rice RING finger protein gene, O. sativa Drought-, Heat-, and Salt-Induced RING Finger Protein 1 (OsDHSRP1; Os02g05692), whose transcripts are highly induced under abiotic stress conditions, including high salinity. We attempted to determine the microtubule (MT)-targeted localization of OsDHSRP1 and verified its E3 ligase activity via an in vitro ubiquitination assay. The functional interactions of two interacting proteins (OsACP1, Os02g27030; OsGLYI-11.2, Os08g09250) with OsDHSRP1 were identified using the yeast two-hybrid (Y2H) system. The interactions between OsDHSRP1 and substrate proteins were confirmed through bimolecular fluorescence complementation (BiFC), in vitro pull-down, and Y2H assays. We identified the regulation of substrates by E3 ligase OsDHSRP1 using an in vitro ubiquitination assay and a cell-free degradation assay. Lastly, the Arabidopsis 35S::OsDHSRP1 transgenic lines showed sensitive phenotypes under heat (acquired thermal tolerance), salt, mannitol, and polyethylene glycol (PEG) stresses. However, they showed no significant sensitivity under ABA treatment. These results suggest that the OsDHSRP1 E3 ligase may act as a negative regulator of abiotic stresses, such as drought, heat, and salinity, through an ABA-independent pathway.

Materials and methods

Gene expression analysis

Rice plants (Oryza sativa ‘Donganbye’) were grown on mesh in plastic containers kept in a growth chamber maintained under a 16/8-h light/dark photoperiod at 25 °C with 70% relative humidity for 2 weeks. Seedlings were subjected to abiotic stresses (200 mM NaCl, drought, or heat at 45 °C) or 1 µM ABA exposure for 0, 6, 12, and 24 h, and the root samples were quickly stored at − 80 °C. To evaluate the relative expression levels of genes, the harvested root samples were frozen in liquid nitrogen and stored at − 80 °C. Total RNA was extracted with TRIzol® (Invitrogen Life Technologies, Carlsbad, CA, USA) according to the manufacturer’s protocol. cDNA synthesis from 1 μg of total RNA was performed using the PrimeScript™ 1st-Strand cDNA Synthesis Kit (Takara-Bio, Ohtsu, Japan). The synthesized cDNA was used as a template for quantitative reverse transcription-polymerase chain reaction (RT-qPCR). The primers, which were designed using Beacon DesignerTM (PRIMER Biosoft, Palo Alto, CA, USA), were synthesized by a commercial service (Macrogen, Seoul, Korea). The RT-qPCR was performed using a CFX ConnectTM Real-Time PCR Detection System (BIO-RADTM) as described by Kim et al. (2019). The reaction mixture contained cDNA (10 ng), 10 μL of SYBR® Green TOP real qPCR 2X PreMIX (EnzynomicsTM, Daejeon, Korea), and 10 pmol of each primer in accordance with the manufacturer’s protocol. The amplification conditions were as follows: initial denaturation at 95 °C for 10 s, followed by 45 cycles of denaturation at 95 °C for 10 s, annealing at the appropriate annealing temperature (Tm) and time for each primer pair, and extension at 70 °C for 30 s. The OsActinII (Os03g50885) and AtUBC21 (AT5g25760) genes were used as internal controls in rice and Arabidopsis, respectively. Statistically significant differences in gene expression between the treated and control samples were verified using the two-tailed Student’s t-test.

Recombinant protein purification and in vitro ubiquitination assay

The full-length OsDHSRP1 (Os02g05692) gene was cloned into the pMAL-c5X vector (New England BioLabs, Ipswich, MA, USA). To generate a point mutation, a single amino acid substitution (C556A), in the RING finger domain, the maltose binding protein (MBP)-OsDHSRP1 plasmid DNA was amplified using Pfu Turbo DNA polymerase (Stratagene, La Jolla, CA, USA) and cloned into a pMAL-c5X vector. The recombinant plasmid DNAs, wild-type MBP-OsDHSRP1 and mutated MBP-OsDHSRP1C556A, were transformed into Escherichia coli strain BL21 (DE3) pLysS (Promega, Madison, WI, USA). The recombinant proteins were purified by affinity chromatography using an amylose resin (New England BioLabs). The amplified AtUBC10 fragment (E2) was cloned into the 6X His-tagged pET-28a (+) vector (Novagen, Gibbstown, NJ, USA) and then was purified using the Ni–NTA Purification System (Invitrogen Life Technologies).

For in vitro ubiquitination, the purified MBP-OsDHSRP1 was mixed with human E1 (Sigma-Aldrich, St Louis, MO, USA), purified 6X His-tagged AtUBC10, and bovine ubiquitin (Sigma-Aldrich) and then incubated in ubiquitination reaction buffer (1 M Tris–HCl, pH 7.5; 40 mM ATP; 100 mM MgCl2; 40 mM DTT) for 3 h at 30 °C (Kim et al. 2019). After incubation, the reaction mixtures were mixed with 5X sodium dodecyl sulfate (SDS) sample buffer and heated at 100 °C for 5 min. Finally, the heated samples were separated on 8.0% SDS–polyacrylamide gel electrophoresis (PAGE) gel and then transferred to a nitrocellulose membrane. Immunoblot analyses were performed using a primary anti-Ub (Sigma-Aldrich) or an anti-MBP antibody (New England biolabs) with a secondary goat anti-rabbit IgG peroxidase antibody (Sigma-Aldrich). The horseradish peroxidase chemiluminescence of the labeled antibodies was observed using the WESTSAVE STAR™ substrate (AB Frontier, Seoul, Korea), and photographs were obtained using the ChemiDoc™ XRS + imaging system (Bio-Rad, Hercules, CA, USA). To confirm the OsDHSRP1-mediated ubiquitination of the OsACP1 and OsGLYI-11.2 proteins, the His-Trx OsACP1- and OsGLYI-11.2 proteins were purified by the Ni–NTA Purification System (Invitrogen Life Technologies). Each purified protein was incubated with E1, E2, and the MBP-tagged wild-type or mutated OsDHSRP1 for 3 h and was detected via immunoblotting analysis using the anti-Trx antibody (Sigma-Aldrich).

Confocal imaging and drug treatments

To identify the subcellular localization of OsDHSRP1 and the partner proteins the full-length genes were cloned into YFP vector according to Lim et al (2013). For the BiFC analysis their full-length clones were inserted into 35S:HA-SPYCE(M) and 35S:c-Myc-SPYNE(R) vectors, respectively. In order to isolate rice protoplast, rice seedlings were grown on half-strength MS nutrient solution in a growth chamber for 14 days. Leaves of the rice seedlings were harvested and dipped in an enzyme solution as described by Zhang et al. (2011). Each of the plasmids was transfected into rice protoplasts using polyethylene glycol (PEG)-mediated methods (Zhang et al. 2011). Cytoskeleton-affecting drugs such as LatB (10 mM, Sigma-Aldrich) and oryzalin (20 mM, ChemService, PA, USA) were dissolved with dimethyl sulfoxide (DMSO) (Yabuuchi et al. 2015). To depolymerize the actin and microtubule (MT), the transformed protoplasts were submerged in water with a final concentration of 25 μM LatB and 5 μM oryzalin at 25 °C for 1 h, respectively (Yabuuchi et al. 2015). Dimethyl sulfoxide (DMSO) was added and incubated as control at 25 °C for 1 h. Protoplasts were observed using a confocal laser scanning microscope (model LSM 510 META NLO, Carl Zeiss, Oberkochen, Germany) at the Korea Basic Science Institute, Chuncheon Center.

Yeast-two hybrid screening the substrate proteins

To screen the substrate proteins of OsDHSRP1, the Y2H system was employed. A rice cDNA library was used as prey, as described by Chapagain et al. (2018). Yeast transformation and library screening were conducted in accordance with the commercial procedures (Make Your Own ‘Mate & Plate’ Library System; Matchmaker Gold Yeast Two-Hybrid System; Yeastmaker Yeast Transformation System 2, Clontech, Palo Alto, CA, USA). A full-length coding sequence of OsDHSRP1 was cloned into the pGBKT7-BD (DNA-binding domain) vector and used as a bait protein. Subsequently, the BD-OsDHSRP1-transformed yeast strain was mated with the salt-treated rice library yeast strains (Chapagain et al. 2018). Thereafter, positive interactions were screened on the double dropout (DDO) medium lacking leucine and tryptophan and supplemented with 40 mg mL−1 X-α-gal and 42 ng mL−1 aureobasidin A (AbA) (DDO/X/A). The positive clones were picked and repatched on the synthetic defined medium lacking adenine, histidine, leucine, and tryptophan, with 40 mg mL−1 X-α-gal and 42 ng mL−1 AbA (QDO/X/A) in accordance with the recommended commercial procedures (Clontech).

To identify protein–protein interactions, the full-length coding sequences of OsACP1 (Os02g27030, O. sativa abiotic stress-induced cysteine protease 1) and OsGLYI-11.2 (Os08g09250, O. sativa glyoxalase) (Mustafiz et al. 2014) were inserted into the pGADT7-AD vector and co-transformed with the pGBKT7-OsDHSRP1 into the Y2H Gold strain. The co-transformed yeast cells were incubated on the DDO medium for 5 days at 30 °C, and then each of the clones was grown in the liquid DDO medium adjusted to OD600 = 1.0 and spotted onto the DDO and quadruple dropout medium (QDO)/X/A media.

In vitro pull down assay

In order to confirm protein–protein interactions, an in vitro pull-down assay was performed using the recombinant His-Trx-tagged OsACP1 and OsGLYI-11.2 and MBP-tagged OsDHSRP1 proteins, using a Pull-Down Assay Kit (Thermo Fisher Scientific, Waltham, MA, USA) according to the manufacturer’s manual. The eluted proteins were denatured by boiling with 5X SDS sample buffer. Thereafter, the sample proteins were separated on 10% SDS–PAGE gel and transferred to a nitrocellulose membrane. The signals were detected with an anti-Trx (Sigma-Aldrich) or anti-MBP antibody (New England Biolabs).

In vitro cell free degradation assay

An in vitro cell-free degradation assay was performed as described by García-Cano et al. (2014). Both OsACP1 and OsGLYI-11.2 genes were cloned into a pET32a (+)-His-Trx vector and transformed into the E. coli strain BL21 (DE3) pLysS (Promega). The expressed Nus-His-tagged OsDHSRP1 and His-Trx-tagged OsACP1 and OsGLYI-11.2 proteins were purified using the Ni–NTA Purification System (Invitrogen Life Technologies). The rice seedlings grown in a growth chamber for 7 d were harvested and ground in liquid nitrogen. Total protein was extracted from the ground rice powder using a Plant Total Protein Extraction Kit (Sigma-Aldrich). Additionally, MG132, a 26S proteasome inhibitor was added to a final concentration of 40 μM. The reactions were stopped by boiling in 5X SDS sample buffer and subsequently analyzed by western blotting.

Arabidopsis transformation

Arabidopsis thaliana ecotype Columbia-0 (Col-0) seeds were used as control plants. To generate overexpressing transgenic plants, the 35S::OsDHSRP1-YFP plasmid was transformed into Agrobacterium tumefaciens GV3101 using the floral dip method (Zhang et al. 2006). The transformed (T1) seedlings were selected for germination on half-strength MS agar plates containing 50 mg L−1 kanamycin and 50 mg L−1 carbenicillin.

In vitro stress assays

The T3 seeds of the transgenic plants (lines #1, #2, and #3) and control seeds (35S::YFP) were tested to determine the heat shock effect, as described by Kim et al. (2019). The transgenic Arabidopsis plants were grown on half-strength MS agar for 7 d and then dipped into water baths at either 38 or 45 °C. The acquired thermal tolerance treatment was performed by heating the plants initially to 38 °C for 90 min and then moving them to a growth chamber maintained at 24 °C for 120 min, and then again heating them to 45 °C for 3 h. The basal thermal tolerance treatments were performed by heating the plants in sealed plates at 45 °C for 1 h. Both acquired and basal thermal tolerance treatments were processed under the dark conditions. The heat-treated plants were recovered in a growth chamber at 24 °C under a 16/8 h light/dark photoperiod and survival scored 7 d later. Plants that were completely bleached were counted as dead, while the green plants were scored as survived. A germination assay under heat stress was performed by placing the cold-stratified seeds in PCR tubes and subjecting them to heat shock for 1 h using a C1000 Thermal Cycler (Bio-Rad, Hercules, CA, USA), in which the temperatures were programmed in the range of 47.0–51.6 °C. Following heat shock, the seeds were placed on half-strength MS medium for 7 days, and the germination rate was scored for 7 days after treatment. In addition, germination rate between OsDHSRP1-overexpressing lines and control plants was evaluated under ABA (0.1, 1, and 5 μM), PEG (− 0.25 and − 0.5 MPa), salt (0, 100, 150, and 200 mM) and mannitol (0, 100, and 200 mM) for 7 days.

For performing the root length assay, the 3-days-old seedlings growing in half-strength MS medium were transferred to the culture dishes containing different concentrations of ABA, NaCl, mannitol, or PEG as needed. The PEG-infused agar plates were prepared according to Verslues et al. (2006).

Soil treatments

For abiotic stresses phenotypic analysis, the T3 seeds were plated in plant pot for 2 weeks. The drought stress experiment, fourteen-day-old plants subjected to water deficit for 7 days. Photographs were taken after re-watering for 5 days. For the heat stress test, fourteen-day-old plants were subjected to heat condition at 38/32 °C (day/night) with a 14 h photoperiod for 60 h and then recovered control condition at 24/18 °C for 7 days. Photographs were taken after heat treatment and recovery for 7 days. Fourteen-day-old plants were treated with 200 mM NaCl in water every other day for 3 weeks. Photographs were taken after the salt treatment and recovery for 12 days. Each experiment was carried out at least three times.

Chlorophyll determination

Chlorophylls a and b were extracted with 4 mL 99% acetone and 2 mL ethanol (2:1 v/v) from 0.5 g leaves of the plants grown under drought, heat, and salt stresses and were kept in the freezer in the dark for 30 min. The content of chlorophyll a and b was calculated by measuring the absorbance at the wavelengths of 663 nm and 645 nm, respectively, according to Pérez-Patricio et al. (2018).

Methylglyoxal (MG) level determination

Ten-day-old seedlings of control and transgenic plants were grown on half-strength MS agar and treated with distilled water (control), 150 mM NaCl, 20% PEG, 300 mM mannitol, or heat (42 °C) for 1 d before determining their MG content. The extract was prepared from approximately 300 mg seedlings using 3 mL of 0.5 M perchloric acid. One milliliter of the total reaction mixture (250 μL of 7.2 mM 1,2-diaminobenzene and 100 μL of 5 M perchloric acid) was mixed with 650 μL of supernatant, which was added last, and the absorbance of the derivative was read at 336 nm, according to Yadav et al. (2005).

Results

OsDHSRP1 transcriptional level is highly upregulated under heat treatment

To test the gene expression patterns of OsDHSRP1 in response to abiotic stresses, such as salinity (200 mM NaCl), drought, and heat (45 °C) as well as A.

BA (1 µM), we used the stress-treated samples after 0, 6, 12, and 24 h (Fig. 1). The relative expression levels of the three marker genes (OsSALT for drought and ABA, OsNAC10 for salt, and OsHSP90-1 for heat) were significantly higher than those in the non-treated samples, indicating the suitability of the treated samples used for the analysis of OsDHSRP1 mRNA levels (Fig. S1). The transcript levels of OsDHSRP1 were upregulated by approximately 6.3-fold after 12 h of salt stress treatment with 200 mM NaCl. In addition, the OsDHSRP1 transcript levels were upregulated by approximately 7.8-fold, after 12 h of drought stress treatment. Interestingly, the transcript levels of OsDHSRP1 increased gradually, up to approximately 24-fold higher in the heat-treated samples (45 °C) than in the control samples (non-treated samples). The OsDHSRP1 gene expression was higher at 6 h (approximately 2.7-fold) but decreased gradually at 24 h (approximately 1.5-fold lower) in the ABA-treated samples (1 µM) than in the control samples. These results suggest that the OsDHSRP1 gene is induced in response to various abiotic stresses.

Fig. 1
figure 1

Expression patterns of the OsDHSRP1 gene in response to abiotic stresses. The relative expression of OsDHSRP1 in rice roots under different abiotic stresses [ABA (1 µM), drought, heat (45 °C), and NaCl (200 mM)] was examined by RT-qPCR, using three biological replicates per time point (n = 9).The OsACTINII (Os03g50885) gene was used as an internal control. The asterisks ** and *** indicate statistically significant differences at P < 0.01, and P < 0.001, respectively, according to the two-tailed Student’s t-test

Full size image

OsDHSRP1 encodes a RING-H2 type E3 ligase

Domain analysis of OsDHSRP1 with 580 amino acids was performed using the rice annotation database of the Institute for Genomic Research (TIGR, https://rice.plantbiology.msu.edu/). To compare the RING domain structure of OsDHSRP1, we retrieved its orthologs in Arabidopsis (AT4G31450 and AT5G10650), Brachypodium distachyon (BRADI3G04020), Zea mays (GRMZM2G392320), and Sorghum bicolor (Sb04g003650 and Sb10g028510) from the rice annotation database. The alignment of their putative amino acid sequences with the amino acid sequence of OsDHSRP1 indicated that the sequences corresponding to the OsDHSRP1 RING-H2-type domain were highly conserved (Fig. 2a for the domain, S2 for the full-length sequence). To determine whether the RING-H2-type domain at the C terminus of the OsDHSRP1 protein harbors the E3 ligase activity (Fig. 2b), we performed an in vitro ubiquitination assay. To identify the E3 ligase activity of the OsDHSRP1 protein, the purified MBP-OsDHSRP1 protein was incubated with E1, E2, and ubiquitin and then subjected to immunoblotting with an anti-ubiquitin or MBP antibody. The immunoblotting results clearly showed the presence of poly-ubiquitinated chains (Fig. 2b, lane 3). However, no poly-ubiquitinated products were detected in the immunoblot assays when E1, E2, or E3 was individually excluded (Fig. 2b, lanes 1–2). In addition, these bands were not detected after incubation with the mutated MBP-OsDHSRP1C556A protein (Fig. 2b, lane 4). These results clearly indicated that OsDHSRP1 is a functional E3 ligase.

Fig. 2
figure 2

The multiple amino acid sequence alignment of the RING-H2 domain of the OsDHSRP1 and its in vitro ubiquitination assay and subcellular localization. a Multiple amino acid alignment of RING-H2 domain between the OsDHSRP1 and orthologs from Arabidopsis (AT4G31450 and AT5G10650), Brachypodium distachyon (BRADI3G04020), Zea mays (GRMZM2G392320), and Sorghum bicolor (Sb04g003650 and Sb10g028510). b In vitro ubiquitin assay of the wild-type OsDHSRP1 and mutated OsDHSRP1C556A proteins. The purified MBP-OsDHSRP1 and MBP-OsDHSRP1C556A proteins were incubated with E1 (human), E2 (Arabidopsis, UBC10), and ubiquitin for 3 h, respectively. The poly-ubiquitin chains were observed via immunoblotting using the Ub antibody. The presence of MBP-OsDHSRP1 protein was confirmed by western blotting using anti-MBP antibody (lower panel). c Subcellular localizations of the 35S::OsDHSRP1-YFP proteins. To confirm the microtubule signals of 35S::OsDHSRP1-YFP, 10 mM of actin depolymerization drug latrunculin B (Lat B) and 20 mM of microtubule inhibitor (oryzalin) were applied to 35S::OsDHSRP1-YFP expressed in rice protoplasts. The constructed vectors was transiently expressed in rice protoplasts for 16 h

Full size image

OsDHSRP1 localizes to the microtubule cytoskeleton

To identify the subcellular localization of OsDHSRP1, we constructed the recombinant protein 35S::OsDHSRP1-YFP and then transfected it into rice protoplasts. The 35S::OsDHSRP1-YFP fluorescence appeared only in the cytoskeletal structure (Fig. 2c), whereas the 35S::YFP fluorescence was localized to both the cytosol and nucleus of rice protoplasts (Fig. S3). Plant cells consist of actin filaments and microtubules (MTs), and two different types of linear proteinaceous polymers are well known for their conserved structures (Kost and Chua 2002). Previous studies have shown that two cytoskeleton drugs, latrunculin (Lat) B and oryzalin, could depolymerize actin filaments and MTs, respectively (Cai et al. 2010). To identify whether the subcellular localization of the OsDHSRP1 protein is associated with the MT or actin, the transformed protoplasts were treated using 25 μM LatB or 5 μM oryzalin, respectively (Fig. 2c). The results of the Lat B treatment showed that the fluorescence signals of 35S::OsDHSRP1-YFP were still observed in the cytoskeletal structures, whereas the signals of OsDHSRP1 were dispersed in the cytosol because of the oryzalin treatment causing MT depolymerization. However, no significant change was observed in the expression of 35S::YFP in the latB- or oryzalin-treated rice protoplasts (Fig. S3). These results implied that the OsDHSRP1 protein is associated with the MTs in rice protoplasts.

Interaction of OsDHSRP1 with two substrate proteins

A RING E3 ligase can interact with its substrate proteins and lead to their degradation via the ubiquitin–proteasome system (Smalle and Vierstra 2004; Vierstra 2009). We investigated the OsDHSRP1-interacting proteins in the Y2H system. BD-OsDHSRP1 was constructed to screen a prey cDNA library. As a result of Y2H screening, a total of five positive clones were selected (Fig. S4). Among them, two rice proteins with strong β-galactosidase activity were selected for further study. These two positive proteins encoded cysteine proteinase 1 precursor (Os02g27030) and glyoxalase (Os08g09250). To date, the cysteine proteinase 1 precursor (Os02g27030) protein has not been characterized, whereas glyoxalase (Os08g09250) has been named OsGLYI-11.2 by Mustafiz et al. (2014). Therefore, we named cysteine proteinase 1 precursor as O. sativa abiotic stress-induced cysteine protease 1 (OsACP1).

To identify the subcellular localization of the interacting proteins, the full-length cDNAs of OsACP1 and OsGLYI-11.2 were cloned into the YFP-tagging vector (35S::YFP). The fluorescence signals were localized to the cytoplasm in both 35S::OsACP1-YFP and 35S::OsGLYI-11.2-YFP transgenic plants (Fig. 3a). However, the signal of 35S::YFP was observed in the cytosol (Fig. 3a). To verify the physical interaction of OsDHSRP1 with OsACP1 and OsGLYI-11.21, a BiFC analysis was performed (Fig. 3b). The BiFC fluorescence signals of the OsDHSRP1/OsACP1 and OsDHSRP1/OsGLYI-11.2 complexes were localized to the cytoplasm. For further confirmation of their physical interaction, OsDHSRP1 and its two interacting proteins were cloned to the yeast GAL4 DNA-binding domain (BD) and GAL4-AD, respectively, and were co-transformed into the Y2H Gold strain. Subsequently, the co-transformed yeast cells (OsDHSRP1/OsACP1 and OsDHSRP1/OsGLYI-11.2) were separately spotted onto the SD/-Leu/-Trp (DDO) and SD/-Ade/-His/-Leu/-Trp/X-α-Gal/AbA (QDO/X/A) plates (Fig. 3c). The yeast colonies co-transformed with OsDHSRP1 and OsACP1 or OsGLYI-11.2 presented a strong α-galactosidase activity. However, no activity was detected when a non-interactive protein (BD-Lam) was co-transformed with OsACP1 or OsGLYI-11.2. To identify the protein–protein interactions of OsDHSRP1 with OsACP1 or OsGLYI-11.2, we constructed each of the His-Trx-tagged OsACP1 and OsGLYI-11.2. Thereafter, the MBP-OsDHSRP1 and His-Trx-OsACP1 or -OsGLYI-11.2 proteins were expressed and purified and co-incubated with the HisPur Cobalt resin. Finally, we confirmed that the bound proteins were eluted from the resins, using immunoblot assays with an anti-MBP or anti-Trx antibody (Fig. 3d). An in vitro pull-down assay demonstrated that both interacting proteins were pulled down by His-Trx-tagged OsDHSPR1 protein using HisPur Cobalt resin, but MBP alone was not pulled down by His-Trx-tagged OsACP1 or OsGLYI-11.2 (Fig. S5). These results support that both proteins serve as substrate proteins for OsDHSPR1. The transcription levels of OsACP1 and OsGLYI-11.2 were evaluated in heat- and salt-treated rice roots at different time points (0, 1, 6 12, 24, and 48 h) using RT-qPCR (Fig. 3e–f). The expression of OsACP1 was the highest (7.8-fold for salt and 15.0-fold for heat) at 48 and 12 h. The OsGLYI-11.2 transcripts were gradually upregulated up to 48 h (1.9-fold) after salt treatment and were highly induced at 12 h (14.8-fold) after heat treatment. In addition, the OsSALT and OsHSP90-1 genes showed the highest expression levels under salt stress at 48 h and under heat at 1 h, respectively.

Fig. 3
figure 3

Identification of proteins interacting with the OsDHSRP1: their subcellular localizations, physical interactions, in vitro pull-down assay and expressions. a Full-length sequences of the OsACP1 and OsGLYI-11.2 were fused into the YFP vector and transfected into rice protoplasts. b BiFC assay of the two interacting proteins with OsDHSRP1. Each of the interacting proteins was cloned into pSPYNE(R), and OsDHSRP1 was cloned into pSPYCE(M). The cloned interacting genes were co-transfected with OsDHSRP1 into rice protoplasts. c Y2H assay for the in vivo interaction of the OsDHSRP1 with interacting proteins. The full-length OsDHSRP1 was cloned into pGBKT7-BD, and the interacting genes (OsGLYI-11.2 and OsACP1) were cloned into pGADT7-AD. Each of the constructed vectors was co-transformed into the Y2H Gold yeast strain and grown on DDO/A (left) and QDO/X/A (right) media. Positive control, co-transformation of BD-53 (encoding the Gal4-BD fused with murine p53)/T-AD (encoding the Gal4-AD fused with SV40 large T-antigen); Negative control, co-transformation of BD-Lam (encoding the Gal4-BD fused with lamin)/AD-T. d In vitro pull-down assay of OsACP1 and OsGLYI-11.2 with OsDHSRP1. Anti-Trx and anti-MBP antibodies were used for immunoblotting analyses. (e–f) Relative expression levels of the OsACP1 and OsGLYI-11.2 genes in rice roots exposed to salt stress (200 mM NaCl) (e) and to heat stress (45 °C) (f), compared to the control roots (0 h). RT-qPCR was performed with three biological replicates per time point (n = 9). The asterisks *, **, and *** indicate statistically significant differences at P < 0.05, P < 0.01, and P < 0.001, respectively, according to a two-tailed Student’s t-test

Full size image

OsDHSRP1 E3 ligase ubiquitinates two substrate proteins and leads to degradation of the proteins via the 26S proteasome

To examine whether the OsDHSRP1 E3 ligase ubiquitinates the two interacting proteins and degrades them proteolytically via the 26S proteasome system, we first performed an in vitro ubiquitination assay of both proteins with the OsDHSRP1 protein (Fig. 4a). As a result, we observed poly-ubiquitinated substrate (OsACP1 and OsGLYI11.2) bands when the two substrate proteins were incubated with E1, E2, and OsDHSRP1 E3 ligase. However, no poly-ubiquitinated substrate bands were found in the lane without E1 or with the mutated OsDHSRP1C556A protein.

Fig. 4
figure 4

In vitro ubiquitination assay and in vitro degradation assays of two interacting proteins regulated by the OsDHSRP1. a In vitro ubiquitination assay of the extracted OsACP1- and OsGLYI-11.2-His-Trx proteins with MBP-tagged OsDHSRP1 E3 ligase. Each interacting substrate protein was incubated with E1, E2, and the MBP-tagged wild type or mutated (C556A) OsDHSRP1 for 3 h. The anti-Trx antibody was used for immunoblotting analyses. The intensity of Coomassie Blue (CBB) staining was used as a loading control of the purified OsACP1- and OsGLYI-11.2-His-Trx proteins. b In vitro protein degradation assay of the purified His-Trx-OsACP1 and OsGLYI-11.2 proteins via the 26S proteasome system. MG132 was employed to inhibit the 26S proteasome pathway. Anti-MBP or anti-Trx antibody was used for immunoblotting analyses. The Ponceau S staining of the Rubisco protein was used as a loading control

Full size image

Subsequently, an in vitro protein degradation assay was performed with the purified MBP-tagged OsDHSRP1, His-Trx-tagged OsACP1, and OsGLYI-11.2 proteins. We observed the protein levels of OsACP1 and OsGLYI-11.2 in the presence or absence of the MBP-OsDHSRP1 protein at different time points (0 and 3 h) (Fig. 4b). As a result, a clear reduction in substrate protein levels was found at 3 h after incubation. However, no degradation of the two substrate proteins occurred in the absence of OsDHSRP1. In addition, no protein degradation of OsACP1 and OsGLYI-11.2 was detected in the lanes with MG132-treated protein and mutated OsDHSRP1 C556A protein, even after 3 h of incubation (Fig. 4b). These results support that the interacting proteins, OsACP1 and OsGLYI-11.2, are ubiquitinated by the OsDHSRP1 E3 ligase and then proteolytically degraded via the 26S proteasome-mediated system.

Phenotype analysis of the OsDHSRP1-overexpressing Arabidopsis plants under abiotic stresses

To identify the molecular function of OsDHSRP1 against the ABA, drought, heat, mannitol, and salt stress treatments, we generated the OsDHSRP1-overexpressing transgenic Arabidopsis plants. Compared to 35S::YFP (expressed in control plants), OsDHSRP1 was highly expressed in three independent 35S::OsDHSRP1-YFP lines (line #1, #2, and #3) (Fig. 5a). To identify plant phenotypic effects under drought stress in soil, the 14-days-old OsDHSRP1-overexpressing and control plants were exposed to water deficit conditions by withholding irrigation for 7 days and then allowing recovery with water irrigation for 5 days. The OsDHSRP1-overexpressing plants showed significantly lower survival rates, by threefold, compared to those of the control plants (Fig. 5b). Furthermore, the chlorophyll content of the transgenic plants was 1.3-fold lower than that of the control plants after recovery (Fig. 5b). Subsequently, to identify the effect of the 35S::OsDHSRP1 under water deficit conditions, the plants were treated with − 0.25 MPa (control) and − 0.5 MPa (transgenic lines) PEG (Fig. S6). PEG treatment resulted in significantly lower germination ratios of the transgenic plants compared to the control plant, while no significant difference was found for the control treatment (Fig. S6a). Similarly, the root lengths of the transgenic plants were shorter than those of control plants grown in the PEG-treated medium, even though no difference was observed between the control and transgenic plants grown in the non-treated medium (Fig. S6b).

Fig. 5
figure 5

Phenotypic effects of the OsDHSRP1-overexpressing Arabidopsis plants in response to abiotic stresses. a RT-PCR analysis of three independent transgenic (35S::OsDHSRP1-YFP) and control (35S::YFP) plants. b Decreased water tolerance of 35S::OsDHSRP1-YFP plants. The 14-days-old plants were grown without water for 7 days and recovered by irrigation with water for 5 days. Survival rates after 5-days recovery were calculated. c Heat sensitivity of transgenic plants. The 14-days-old plants were treated at 38/32 °C (day/night) for 60 h and recovered at 24/18 °C for 7 days. Survival rates after 7-days recovery were calculated. d Salt sensitivity of the 35S::OsDHSRP1-YFP transgenic plants. The 14-days-old plants were treated with 200 mM NaCl for 3 weeks. Survival rates after 14-days recovery were calculated. The survival rates are presented as mean ± SD from three biological replications (n = 30). Total chlorophyll content was measured in 35S::OsDHSRP1-YFP and control (35S::YFP) plants after recovery. The values are means ± SD from three biological replications (50 mg of leaves from each replicate). The asterisks * and ** indicate statistically significant differences at P < 0.05 and P < 0.01, respectively, according to a two-tailed Student’s t-test

Full size image

Subsequently, the phenotypic effects of the OsDHSRP1-overexpressing plants against heat stress were evaluated. To identify phenotypic effects under heat stress in soil, the 14-days-old OsDHSRP1-overexpressing and control plants were exposed to heat treatment at 38/32 °C (day/night) for 60 h (Fig. 5c). The survival rates of the transgenic lines after 7 days were significantly lower by threefold compared to those of the control plants. The chlorophyll content was also investigated in the heat-treated adult seedlings of the OsDHSRP1-overexpressing and control plants. The chlorophyll content of the transgenic plants was 1.77-fold lower than that of the control plants after recovery. Similarly, the seeds of the control and three independent transgenic lines were subjected to heat stress with a temperature range of 47.0–49.8 °C, and their germination rates were recorded at 7 days after treatment (Fig. S7a). All three transgenic plants exhibited a threefold decrease in seed germination compared to the control plants under different temperatures. To identify the thermo-tolerance of transgenic plants, the OsDHSRP1-overexpressing and control plants were subjected to heat stress at 45 °C for 1 h. This resulted in a recovery rate of 0% (basal thermal tolerance) for both transgenic and control plants [Fig. S7b, (c)], whereas all plants survived when exposed to a temperature of 38 °C for 1 h [Fig. S7b, (b)]. We next tested the acquired thermal tolerance of both transgenic and control plants, which were subjected to a temperature of 38 °C for 90 min and then cooled to room temperature (24 °C) for 2 h, followed by reheating to 45 °C for 3 h. The control plants presented a higher recovery rate of 45.3% than the OsDHSRP1-overexpressing lines (average 23.6%) (Fig. S7b, (d), and bottom).

To compare the phenotypes of the OsDHSRP1-overexpressing and control plants under salt stress, all transgenic plants showed similar growth rates under normal salt condition (Fig. 5D). However, when exposed to 200 mM NaCl, the OsDHSRP1-overexpressing plants showed approximately 3.21-fold lower recovery ratio compared to that of the control plants (Fig. 5d). The chlorophyll content of the transgenic plants was 1.65-fold lower than that of the control plants after recovery. The transgenic lines were grown on half-MS medium containing 0, 100, 150, and 200 mM NaCl. Under normal conditions, no significant differences in germination rate or root length were observed between the control and transgenic plants (Fig. S8). However, germination rates of the OsDHSRP1-overexpressing plants were significantly lower with increasing salt concentrations (1.18-fold for 100 mM NaCl; 2.31-fold for 150 mM; and 1.08-fold for 200 mM after recovery for 7 days) compared to those of the control plants (Fig. S8A). Furthermore, the transgenic plants exposed to the treatments of 150 and 200 mM NaCl had significantly shorter root lengths (1.94-fold for 150 mM NaCl and 1.24-fold for 200 mM NaCl) than those of the control plants. However, the root lengths measured after 100 mM NaCl treatment showed no significant difference between the transgenic and control plants (Fig. S8B). Under normal salt conditions, all transgenic plants showed similar growth rates (Fig S8B).

Similarly, mannitol, which lowers the water potential of the medium, has been known to induce osmotic stress in plants (Claeys et al. 2014). The 35S::OsDHSRP1-overexpressing plants were treated with various concentrations of mannitol (0, 100, 200, or 300 mM) (Fig S9). Under mannitol stress, germination rates were significantly lower with increasing mannitol concentrations compared those of the control plants (Fig S9a). In addition, the OsDHSRP1-overexpressing lines showed shorter root lengths than those of the control plants, and their root lengths decreased with increasing mannitol concentrations (0, 100, or 200 mM) (Fig. S9b). Furthermore, ABA accumulation is considered to play an important role in the regulation of osmotic stresses, including salt and drought (Ma et al. 2019). The relative expression of the OsDHSRP1 gene was induced by ABA (Fig. 1), and we observed no significant differences in germination rate and root length between the transgenic and control plants under various ABA concentration treatments (Fig. S10). These results suggest that the OsDHSRP1-overexpressing plants exhibit reduced tolerance to stresses, including drought, heat, and salt but not ABA.

Transcriptional analysis of stress responsive genes in in OsHSPR1-overexpressing plants

To identify the transcriptional changes of several stress-responsive genes caused by OsDHSRP1 overexpression, we performed RT-qPCR of OsDHSRP1-overexpressing plants under drought (air drying), heat (45 °C), and salt (200 mM NaCl) conditions. To evaluate the expression of drought-responsive genes, we selected three drought-responsive genes, including AtDREB2A (AT5G05410), AtDREB2B (AT3G11020), and an unnamed gene (AT1G69870) (Sakuma et al. 2006a, b; Wang et al. 2018a, b). The expressions of these genes were lower in the OsDHSRP1-overexpressing plants than in the control plants, especially 12 h after air-drying treatment (Fig. 6a). In addition, we selected one heat shock transcription factor A3 gene (HsfA3; AT5G03720) and three heat shock protein genes (HSP17.4, AT3G46230; HSP18.2, AT5G59720; and HSP20, AT1G53540) (Li et al. 2017; Sewelam et al. 2019). The expression levels of the above four genes were significantly lower in the three OsDHSRP1-overexpressing lines at different time points in comparison with those in the control plants (Fig. 6b). For example, high differences in the expression levels of all four genes at 2 h after heat treatment were found between the transgenic and control plants. In addition, significant differences in their expression levels were observed under non-treatment conditions between the transgenic and control plants. In addition, one NAC (for no apical meristem; NAM, Arabidopsis thaliana activating factor 1–2; ATAF1–2, and cup-shaped cotyledon; CUC2)-domain transcription factor (NAC2; AT5G39610) and transcriptional regulators of two Salt Overly Sensitive (SOS) genes (SOS1, AT2G01980; SOS2, AT1G01030), which are known to be related to salt stress (Halfter et al. 2000; Shi et al. 2000; Balazadeh et al. 2010), were examined for salt stress response. The transcription levels of all three genes were significantly lower in the OsDHSRP1-overexpressing plants than in the control plants under salt treatment or non-treatment conditions (Fig. 6c). However, the expression levels of the ABA-responsive genes, such as RD20A (AT5G52310), MYB2 (AT2G47190) and bZIP (AT5G49450), showed no significant difference between the transgenic and control plants (Fig. S11). These results suggest that OsDHSRP1 overexpression negatively regulates the expression levels of the genes responsive to abiotic stresses, such as drought, heat, and salt, but not ABA.

Fig. 6
figure 6

Stress-responsive gene expression patterns in OsDHSRP1-overexpressing plants under drought (air-dried), heat (45 °C), and salt (200 mM NaCl) stress. a The expression levels of the dehydration-responsive element-binding protein 2A (AtDREB2A; AT5G05410) and 2B (AtDREB2B; AT3G11020) genes and one drought-induced gene (AT1G69870) in response to drought stress were analyzed by RT-qPCR. b The expression levels of one heat shock transcription factor A3 gene (AtHsfA3; AT5G03720) and three heat shock protein genes (AtHSP17.4; AT3G46230, AtHSP18.2; AT5G59720, and AtHSP20; AT1G53540) in response to heat stress were analyzed by RT-qPCR. c NAC domain transcription factor (AtNAC6; AT1g01720) and two Salt Overly Sensitive (SOS, AtSOS1; AT2g01980, AtSOS2; AT1g01030) genes in response to salt stress were analyzed by RT-qPCR, using three biological replicates per time point (n = 9). The expression level of each gene in control plants was defined as 1.0. The asterisks *, **, and *** indicate statistically significant differences at P < 0.05, P < 0.01, and P < 0.001, respectively, according to a two-tailed Student’s t-test

Full size image

Interaction of Arabidopsis orthologs of the OsDHSRP1-interacting proteins

Our phenotype studies raised a question whether OsDHSRP1-overexpression regulated the expression of the Arabidopsis orthologs of the OsACP1 and OsGLYI-11.2. In order to examine the hypothesis that the Arabidopsis orthologs of the interacting proteins might interact with the OsDHSRP1 protein, we first identified the Arabidopsis orthologous genes and aligned the full-length amino acid sequences of rice and Arabidopsis homologous proteins (Fig. S12). The protein products of the orthologous genes OsACP1 and AtACP1 showed 70% similarity and those of OsGLYI-11.2 and AtGLYI-11.2 showed 72% similarity (Fig. S12c). The in vitro ubiquitination assays showed poly-ubiquitinated substrate bands in the lanes with E1, E2, MBP-OsHSPR1, and orthologous proteins, whereas no poly-ubiquitinated substrate bands were observed in the lanes without E1 or with OsDHSRP1C556A (Fig. S13a). Subsequently, an in vitro pull-down assay showed that each of the interacting proteins was pulled down from the HisPur Cobalt resin by the His-Trx-tagged OsDHSRP1 protein (Fig. S13b). However, no empty MBP were pulled down by the His-Trx-tagged AtACP1 or AtGLYI-11.2 (Fig. S13c). These results support the notion that OsDHSRP1-overexpression regulates the expression of orthologous proteins that affect sensitive responses under abiotic stresses in transgenic Arabidopsis.

Methylglyoxal (MG) level in transgenic lines

The glyoxalase system is involved in the detoxification of MG. Mustafiz et al. (2014) reported that the heterologous expression of OsGLYI-11.2 lowers the MG content. No significant difference was observed in MG content between the control and transgenic plants under normal conditions (Fig. 7a). When exposed to 150 mM NaCl, heat (42 °C), 20% PEG, and 300 mM mannitol, the OsDHSRP1-overexpressing transgenic plants showed a higher accumulation of MG than that by the control plants [1.51-fold for 150 mM NaCl, 1.56-fold for heat (42 °C), 1.24-fold for 20% PEG, and 1.10-fold for 300 mM mannitol (Fig. 7a)].

Fig. 7
figure 7

Methylglyoxal (MG) content in transgenic lines and schematic representation of the OsDHSRP1 under abiotic stress. a The MG content in control (35S::YFP) and transgenic plants (35S::OsDHSRP1-YFP #1, #2, and #3) under normal and stress conditions [150 mM NaCl, heat (42 °C), 20% PEG, and 300 mM mannitol]. The asterisks * and ** indicate statistically significant differences at P < 0.05 and P < 0.01, respectively, with respect to the control plants. b During stress, the OsDHSRP1 gene was highly induced. The OsDHSRP1 protein targeted OsGLYI-11.2 as a substrate protein, which in turn was degraded via the 26S ubiquitin–proteasome pathway. These series of processes lead to the inhibition of the glyoxalase pathway. Finally, the increased levels of MG and reactive oxygen species (ROS) cause decreased stress tolerance

Full size image

Discussion

As sessile organisms, plants encounter adverse environmental conditions resulting from abiotic stresses, including drought, heat, and salt, which limit their growth and development. Among the various abiotic stresses, drought, salt, and temperature impose osmotic stress that can lead to reduced turgor pressure and a decrease in tissue water content (Farooq et al. 2009). Furthermore, these stresses adversely affect cellular structures by obstructing cell elongation, limiting cell division, and impairing key physical functions (Larcher 2003; Farooq et al. 2009). Rice plants are sensitive to various abiotic stresses, which adversely affect their physiological or biological status and overall vegetative, reproductive, and grain metabolisms (Ren et al. 2005; Cruz et al. 2013). Therefore, future studies are needed to understand the mechanisms of abiotic stress response in rice.

Increasing numbers of studies have shown the key roles of the RING-type E3 ligases under various environmental stresses; these roles include seed germination, root development, light signal transduction, ABA signaling transduction, and responses to abiotic stress (Shu and Yang 2017). For example, the OsSDIR1 protein is positively regulated in response to drought and salinity, and its overexpression resulted in enhanced tolerance to drought (Gao et al. 2011). Moreover, Li et al. (2011) demonstrated that the Arabidopsis RING-H2 E3 ligase RHA2a positively regulates drought response in an ABA-dependent manner. The rice RING E3 ligase O. sativa drought-induced SINA protein 1 (OsDIS1) was induced by drought treatment and plays a negative role in drought stress tolerance (Ning et al. 2011). The present study suggests that the OsDHSRP1 gene encodes a 588-amino acid protein called RING-H2-type E3 ligase. By analyzing the protein motifs in OsDHSRP1, a C3H2C3-RING domain was found at its C-terminal side (Fig. 2A). The OsDHSRP1 protein exhibited in vitro E3 ligase activity and was localized in plant MTs (Fig. 2B–C), which are key cytoskeletal elements involved in various cellular functions, such as cell division and cell expansion (Nogales 2001; Sedbrook 2004; Hashimoto 2015). In addition, the MT functions in plant cells are modulated in response to developmental and environmental cues (Hashimoto 2015). For example, an increase in the expression of Microtubule-Associated Stress Protein 1 (MASP1) has been reported during stress treatments in plants, and the MASP1-overexpressing plants could control MT stability, plant growth, and drought response; these functions were dependent on phosphorylation (Bhaskara et al. 2017). In addition, O. sativa Microtubule-Associated RING Finger Protein 1 (OsMAR1) was localized along MTs, where it decreased the germination rate and root length under salt stress by regulating the expression of the cytosolic protein O. sativa chymotrypsin protease inhibitor 2 (OCPI2) via 26S proteasome activity (Park et al. 2017). In another study, O. sativa microtubule-associated RING finger protein 1 (OsRMT1) was found to be associated with MTs, and the OsRMT1 E3 ligase targeted substrate proteins via MTs to MTs, nucleus, or both (Lim et al. 2015). Our findings demonstrated that OsDHSRP1 is clearly associated with MTs (Fig. 2c). Interestingly, this MT-localized OsDHSRP1 protein interacted with the cytosolic proteins OsACP1 and OsGLYI-11.2 (Fig. 3a) and led to protein degradation via the ubiquitin–proteasome system. The MT-localized OsRMT1 has also been shown to functionally interact with various substrate proteins localized to both the cytosol and nucleus (Lim et al. 2015). However, none of these proteins was localized to MTs. Further studies will be necessary to elucidate the molecular mechanisms underlying protein–protein interactions in different subcellular organelles.

Many studies have shown that the E3 ligases interacting with substrates and ubiquitinated substrates are degraded through the 26S proteasome system (Lourenço et al. 2013; Zhang et al. 2017; Park et al. 2018; Kim et al. 2019). We found two substrate proteins OsGLYI-11.2 and OsACP1, by using the Y2H system, in vitro pull-down assay, and BiFC analysis (Fig. 3a–d). The genes encoding these two proteins were upregulated in response to abiotic stresses, such as drought and salt (Fig. 3e–f). These target proteins were not localized in MTs, but in the cytoplasm (Fig. 3a). Interestingly, when the MT-localized OsDHSRP1 and each of its substrate proteins were co-expressed in rice protoplasts, the fluorescence signals were observed in the cytoplasm (Fig. 3b). For example, the interaction and regulation of OsMAR1 with cytoplasmic substrate protein, OCPI2, has been observed in the cytoplasm (Park et al. 2017). Similarly, Lim et al. (2015) reported that MT-localized OsRMT1 interacts with cytoplasmic substrate proteins, such as OsSALT, OsCPA1, and OsPB1. Therefore, the OsDHSRP1 protein might be translocated from MTs to the cytoplasm for targeting substrate proteins. In addition, the levels of the OsACP1 and OsGLYI-11.2 proteins decreased when they were incubated with the OsDHSRP1 protein (Fig. 4b). Our findings support the hypothesis that OsDHSRP1 ubiquitinates two substrate proteins, leading to a decrease in their levels through the 26S proteasomal degradation pathway.

In our study, OsDHSRP1 was highly upregulated by abiotic stresses, such as ABA, drought, heat, and NaCl (Fig. 1). These results suggest that OsDHSRP1 could be induced by various abiotic stresses, including ABA, drought, heat, and NaCl, and might possibly trigger the response of downstream target genes in the early stage of abiotic stresses. In heat stress response, the accumulation of heat shock proteins (Hsps) plays an important role in acquired thermal tolerance in plants (Baniwal et al. 2004; Wang et al. 2004; Nakamoto and Vígh 2007; Schramm et al. 2008). Furthermore, the major regulators of the expression of heat stress response genes are known as heat stress transcription factors (Hsfs) (Baniwal et al. 2004; Miller and Mittler 2006; Schramm et al. 2008). The relative expression level of OsHSP18.6 was highly induced by diverse stresses, including drought, salt, cold, and especially heat. Compared to the control plants, OsHSP18.6-overexpressing plants exhibit increased thermal tolerance with greater biomass, lower levels of malondialdehyde, and higher activity of catalase and superoxide dismutase in response to heat (Wang et al. 2015). In addition, TaHSF3 is highly upregulated by heat in wheat seedlings, and compared to wild-type plants, Arabidopsis plants overexpressing TaHSF3 show higher heat tolerance (Zhang et al. 2013). The overexpression of OsDHSRP1 reduced thermo-tolerance in Arabidopsis, besides lowering the survival rate and chlorophyll content (Fig. 5c), indicating that the overexpression of OsDHSRP1 reduced acquired thermal tolerance in Arabidopsis by lowering its ability to recover. In this study, the expression of one Hsf (AtHsfA3) and three Hsps (AtHsp17.4, AtHsp18.2, and AtHsp20) was lower in the OsDHSRP1-overexpressing plants than in the control plants under normal and heat stress conditions (Fig. 6b). These results were consistent with the chlorophyll content of the OsDHSRP1-overexpressing plants.

The NAC transcription factor and SOS signaling are well-known salt stress response pathways. NACs are members of a plant-specific transcription factor family that contains the highly conserved NAC domain (Aida et al. 1997; Olsen et al. 2005). The AtNAC2 gene has been found to be induced by salt stress, with its overexpression resulting in an increase in the number of lateral roots under salt stress (He et al. 2005). The SOS signaling pathway mainly mediates ion homeostasis, and the overexpression of SOS genes has been found to enhance salt tolerance in Arabidopsis (Halfter et al. 2000; Shi et al. 2000; Yang et al. 2009). In this study, three salt-induced genes, AtNAC2, AtSOS1, and AtSOS2, were downregulated in response to the overexpression of OsDHSRP1 in plants growing under normal or salinity conditions (Fig. 6c). These results are consistent with the reduced salt tolerance of the OsDHSRP1-overexpressing plants (Fig. 5d).

The DREB/C-repeat element binding factors have been reported as important transcription factors in the drought and salt stress signaling pathways (Nakashima et al. 2000; Sakuma et al. 2006a, b). The two DREB2-type proteins, DREB2A and DREB2B, have been considered as major transcription factors that function under drought and high-salt stress conditions (Nakashima et al. 2000; Sakuma et al. 2002). Sakuma et al. (2006a, b) reported that AT1G69870 is one of the 21 genes upregulated by the overexpression of DREB2A under drought stress treatment. AtDREB2A, AtDREB2B, and AT1G69870 were selected for further study, as the expressions of these genes were induced by drought. However, the expressions of these stress genes were downregulated in the OsDHSRP1-overexpressing plants (Fig. 6a), indicating that the overexpression of OsDHSRP1 reduces drought tolerance by downregulating stress-responsive genes (Fig. 5b).

Plant cysteine proteases are key enzymes involved in the regulation of programmed cell death (PCD) and play important roles in ensuring cell survival and maintaining protein homeostasis in plants (Solomon et al. 1999; Kidrič and Kos 2014). PLCPs, one of the two types of cysteine proteases (PLCPs and LLCPs), have been reported to be associated with functions such as seed germination, growth, development, aging, and immune and stress responses (Liu et al. 2018). Especially, in response to multiple environmental stresses, such as salt, PEG, and cold, the wheat PLCP gene (TaCP) was upregulated, and the transgenic Arabidopsis plants overexpressing this gene showed drought tolerance (Zang et al. 2010). In addition, Wang et al. (2018a, b) reported that rice cysteine proteases (OsCPs; PLCPs) are induced in response to abiotic stresses, such as cold, drought, and salt, in rice plants. LLCPs are a group of Asn-specific proteinases related to seed development and embryogenesis in plants (Hara-Nishimura 1995; Hiraiwa et al. 1999; Wang et al. 2019). In addition, plant glyoxalase I, which is a component of the glyoxalase system, is known to detoxify MG. For example, the overexpression of OsGLYI resulted in improved tolerance to NaCl, ZnCl2, and mannitol (Zeng et al. 2016). Mustafiz et al. (2014) reported that the heterologous expression of OsGLYI-11.2 exhibits improved adaptation to mannitol, MG, or salt. The levels of MG were higher in the OsDHSRP1-overexpressing plants than in the control plants after stress treatments, such as heat, salt, PEG, and mannitol (Fig. 7a). Under non-stress conditions, all transgenic plants showed similar MG levels. These observations suggest that, as an E3 ligase, the OsDHSRP1 protein can degrade the substrate protein OsGLYI-11.2, thereby leading to a decrease in stress tolerance (Fig. 7b). Interestingly, the expression of the genes encoding these two interacting proteins was upregulated under heat and salt treatment (Fig. 3e–f). Our results raise two questions: (1) why are the relative expression levels of the genes encoding OsDHSRP1 and its interacting proteins increased under stress conditions? and (2) why OsDHSRP1 regulates the two substrate proteins required for stress tolerance via the 26S proteasome system? All intracellular and extracellular proteins are continuously being hydrolyzed to amino acids and replaced by newly synthesized proteins in a process called turning over. The continual destruction of cellular proteins contributes to protein homeostasis. For example, individual proteins within the nucleus, cytosol, endoplasmic reticulum, and mitochondria are degraded in widely differing intervals, ranging from minutes to weeks (Mitch 1996). Cells possess multiple proteolytic systems and complex regulatory mechanisms, and selective proteolysis is very important in the proteolytic process (Flick and Kaiser 2012). In the proteolytic process, excessive breakdown of protein needs to be prevented, and the ubiquitin–proteasome system is one of the pathways with selective targeting. In this study, the two substrates of OsDHSRP1 might be highly induced to adapt to adverse environmental conditions; however, because the rates of protein synthesis and degradation must be balanced, the OsDHSRP1 protein might also be highly induced as an E3 ligase. Taken together, we suggest that the excessive degradation of abiotic stress-induced substrates, OsACP1 and OsGLYI-11.2, might lead to a sensitivity response in OsDHSRP1-overexpressing plants under stress treatments, such as drought, high temperature, and high salinity.

In the present study, we found that the expression of OsDHSRP1 was highly induced by abiotic stresses, such as drought, heat, and salt. Similarly, the expression of its two partner genes was highly increased under heat and salt treatment. The two substrate proteins were ubiquitinated and degraded by OsDHSRP1 via the 26S proteasomal system. The OsDHSRP1-overexpressing plants showed hypersensitive responses under drought, high temperature, and high salt stress conditions, and the levels of the two ortholog Arabidopsis substrate proteins, AtACP1 and AtGLYI-11.2, were regulated by the E3 ligase OsDHSRP1. Collectively, OsDHSRP1 might act as a negative regulator of abiotic stress-responsive genes by regulating the expression of stress-induced proteins. Although plant responses to abiotic stresses are relatively well known, the plant signaling mechanisms underlying multiple co-occurring stresses are still fragmentary. Therefore, we suggest that further studies of the loss of function of the OsDHSRP1 gene using genome editing tools, such as CRISPR-Cas9, can reveal a wide range of abiotic stress tolerance responses and that these results might provide valuable information relating to adaptation and regulation under abiotic stress via the ubiquitin–proteasome system in plant cells.