The Arabidopsis U-box E3 ubiquitin ligase PUB30 negatively regulates salt tolerance by facilitating BRI1 kinase inhibitor 1 (BKI1) degradation

2.1 Plant materials, growth conditions and stress treatments

Arabidopsis thaliana ecotype Columbia (Col-0) was used as the wild type. T-DNA inserted mutants, pub30-1 (Salk_012549), pub30-3 (Salk_073739), and pub31 (Salk_054774), were obtained from the Arabidopsis Biological Resource Center (ABRC, Ohio State University, Columbus, OH, USA). Seeds were sterilized with 20% bleach and 0.1% Triton-X100 for 10 min and washed 5 times with sterilized water, and then sown on plastic Petri plates containing half-strength Murashige–Skoog (MS) medium (Duchefa Biochemie, Haarlem, The Netherlands) including 1% sucrose and 0.3% phytogel (Sigma-Aldrich, St Louis, MO, USA) with or without 150 mM NaCl. The seeds were laid in the refrigerator at 4 °C for 3 days and transferred to the incubator at 23 °C with continuous light with an intensity of 120 μmol m⁻² s⁻¹.

2.2 Real-time quantitative reverse transcription polymerase chain reaction (RT-PCR)

Total RNA was isolated from 10-day-old seedlings sown on 1/2 MS medium using the RNAprep Pure Plant Kit (Tiangen, Beijing, China) following the manufacturer's instructions. Total RNA (2 μg) was reverse-transcribed using Transcript All-in-one First-Strand cDNA Synthesis Supermix for qPCR Kit (Transgene, Beijing, China) and equal quantities of reverse-transcribed products were used for the real-time quantitative RT-PCR reactions. Quantitative RT-PCRs were performed using a SYBR Premix Ex Taq™ RT-PCR kit (Takara, Shiga, Japan) following the manufacturer's instructions with the primers listed in Table S1.

2.3 Yeast two-hybrid assay

In the initial library screening step, full-length PUB30 was used as a bait to screen a cDNA library made from 12-day-old Arabidopsis seedlings with the Matchmaker Library Construction & Screening Kit (Clontech). To confirm the interaction between PUB30 and BKI1, the full-length coding sequence of PUB30 and PUB31 were amplified and cloned into the pGBKT7 (BD-PUB30 and BD-PUB31) and BKI1 was inserted into pGADT7 (AD-BKI1). The polymerase chain reaction (PCR) primers used for plasmid construction are listed in Table S2. The combinations of BD-PUB30 and AD-BKI1, BD-PUB31 and AD-BKI1, BD-PUB30 and AD, BD-PUB31 and AD, and BD and AD-BKI1 were transformed into yeast strain AH109. Yeast synthetic drop-out media lacking leucine (Leu) and tryptophan (Trp; Coolaber, Beijing, China) was used to identify transformed colonies. Three days after transformation, selective colonies were grown in –Leu and −Trp liquid media for 24 h. Then, cell suspension was pipetted onto yeast synthetic drop-out media lacking histidine (His), Leu, and Trp (Coolaber, Beijing, China) with 2.5 mM 3-amino-1,2,4-triazole (3-AT, Sigma-Aldrich, St Louis, MO, USA).

2.4 Pull-down assay

All of the PCR primers used for plasmid construction are listed in Table S2. In order to express the PUB30, PUB31, and BKI1 proteins in Escherichia coli DE3 cells, the coding sequence of PUB30, PUB31, and BKI1 were amplified and individually cloned into pCold-MBP, pCold-MBP, and pCold-GST vectors with infusion cloning kit (Takara, Shiga, Japan). The glutathione S-transferase (GST) and maltose binding protein (MBP) tags were fused to the N-terminus of target proteins (pCold-MBP-PUB30, pCold-MBP-PUB31, and pCold-GST-BKI1).

In pull-down assay, 2 μg of GST-BKI1 protein was incubated with 10 μg of MBP-PUB30 or MBP-PUB31 that was attached onto the agarose beads for 30 min at room temperature in 100 μl of binding buffer (20 mM Tris–HCl, pH 7.2, 10 mM MgCl₂, 2 mM DTT, and 5% NP40). After 5 times washing with binding buffer, the beads were boiled with sodium dodecyl sulfate loading buffer, and samples were separated on sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) and detected with 1:3,000 GST or 1:5,000 MBP antibody (CWBIO, Beijing, China). The horseradish peroxidase-conjugated anti-mouse IgG (H + L; 1:3,000 dilution, CWBIO, Beijing, China) was used as the second antibody. The western blots were developed with enhanced chemiluminescence reagent (Coolaber, Beijing, China).

2.5 Co-immunoprecipitation assay

The full-length coding sequence of PUB30, PUB31, and BKI1 were amplified and individually cloned into the vector 1307-FLAG and 1307-Myc (Zhao et al., 2016). The primers used for plasmid construction are listed in Table S2. The combinations of 1307-FLAG-PUB30 and 1307-Myc-BKI1 and 1307-FLAG-PUB31 and 1307-Myc-BKI1 were transformed to Arabidopsis mesophyll protoplast according to the standard protocols (Sheen, 2001). The transfected protoplasts were incubated overnight at 23 °C and homogenized in immunoprecipitation buffer containing 50 mM Tris–HCl, pH 7.5, 150 mM NaCl, 2 mM DTT, 0.1% Triton, and 1% protease inhibitor cocktail (Roche, Indianapolis, IN, USA). The sample was centrifuged at 12,000 rpm for 10 min at 4 °C, and the supernatant was incubated with agarose-conjugated anti-Myc antibody (Sigma-Aldrich, St Louis, MO, USA) for 3 hr at 4 °C. After 5 times washing with immunoprecipitation buffer, the protein were separated on 10% SDS-PAGE gel and subjected to immunoblotting using anti-Myc antibody (CWBIO, Beijing, China) and anti-FLAG antibody (CWBIO, Beijing, China).

2.6 In vitro self-ubiquitination assay

In vitro self-ubiquitination assay was carried out according to the method reported by Xie et al. (2002). MBP-PUB30 (500 ng) was immobilized on amylose resin beads. UBA1, UBC8, and UBQ14 were individually cloned into the pET28a vector to express E1, E2, and ubiquitin. The primers used for plasmid construction are listed in Table S2. Ubiquitination assays were carried out by adding 200 ng UBA1 (E1), 100 ng UBC8 (E2), and 2 μg ubiquitin in the buffer containing 50 mM Tris–HCl, pH 7.4, 2 mM ATP, 5 mM MgCl₂, and 2 mM DTT. The reaction mix (30 μl) was incubated for 2 hr at 30 °C. The reaction products were separated on 6% SDS-PAGE gel and subjected to immunoblot using anti-MBP antibody (1:10,000 dilution, CWBIO, Beijing, China) and anti-ubiquitin (anti-Ub) antibody (1:3,000 dilution, Santa Cruz, CA, USA).

2.7 In vivo ubiquitination assay

The in vivo ubiquitination assay was carried out according to the reported method (Irigoyen et al., 2014). Briefly, the wild-type and transgenic plants overexpressing 1307-Myc-BKI1 were grown on MS medium for 10 days and then treated with 50 μM MG132 for 6 hr. Fresh seedlings (0.2 g) were grinded in liquid nitrogen, and the powder was solubilized in 600 μl BI buffer containing 50 mM Tris–HCl, pH 7.5, 20 mM NaCl, 0.1% NP-40 and 5 mM ATP, 1 mM PMSF, 50 μM MG132, 10 nM Ub aldehyde (Sigma-Aldrich, St Louis, MO, USA), and 10 mM N-ethylmaleimide (Coolaber, Beijng, China). After 10 min of centrifugation at 12,000 rpm at 4 °C, the supernatants were incubated with prewashed p62-agarose (Enzo Life Sciences, Farmingdale, USA) for 3 hr, and then the agarose was washed twice with BI buffer and once with BII buffer (supplemented with 200 mM NaCl in BI buffer). The proteins were separated on 10% SDS-PAGE gel and subjected to immunoblotting using anti-Myc antibody (CWBIO, Beijing, China) and anti-Ub antibody (Santa Cruz, CA, USA).

2.8 In vitro cell-free degradation assay

Ten-day-old seedlings of pub30-1 and wild type were harvested and grinded into fine powder in liquid nitrogen. Total proteins were subsequently extracted in degradation buffer containing 25 mM Tris–HCl, pH 7.5, 10 mM NaCl, 10 mM MgCl₂, 4 mM PMSF, 5 mM DTT, and 10 mM ATP (F. Wang et al., 2009). After 10-min centrifugation at 17,000 g, the supernatants were adjusted to the equal concentration with degradation buffer. Purified GST-BKI1 protein (100 ng) was incubated with 50 μl total protein extracts at 23 °C. The products were subjected to western blotting with anti-GST antibody (CWBIO, Beijing, China).

2.9 In vivo degradation assay

Ten-day-old transgenic seedlings overexpressing Myc-BKI1 were pretreated with MG132 for 16 hr and then washed 5 times before being transferred to liquid MS medium containing 200 mM cycloheximide (Sigma-Aldrich, St Louis, MO, USA) to block de novo protein synthesis. The samples were harvested after 0 min, 90 min, and 180 min, and the seedlings were grinded in liquid nitrogen and solubilized with extraction buffer (100 mM Tris–HCl, pH 7.5, 300 mM NaCl, 2 mM EDTA, pH 8.0, 1% TrionX-100, 10% glycerol, and a protease inhibitor cocktail). After 10-min centrifugation at 12,000 rpm at 4 °C, the supernatant was incubated with 5 × sodium dodecyl sulfate loading buffer and separated by 10% SDS-PAGE gel and analysed by immunoblotting with anti-Myc antibody at 1:3,000 dilution (CWBIO, Beijing, China).

3.1 The pub30 mutants are more tolerant to salinity stress during seed germination stage

Hwang et al. (2015) reported that the Arabidopsis pub30 mutant was more tolerant to salt stress in the germination stage compared with the wild type. However, little is known about the underlying mechanism for the salt-tolerance phenotype of the pub30 mutant. Here, we aimed to elucidate the potential mechanism using two T-DNA insertion mutants, that is, pub30-1 (Salk_012549) and pub30-3 (Salk_073739; Figure 1a). Real-time PCR analysis showed that the full length transcript of PUB30 could not be detected in pub30-1 and pub30-3 mutants, suggesting that pub30-1 and pub30-3 are knockout mutants (Figure 1b). To investigate the function of PUB30 in salt stress responses, seeds of the wild type and the two pub30 knockout mutants were germinated on 1/2 strength MS medium with or without 150 mM NaCl (Figure 1c). The untreated wild-type and mutants seeds grew normally with 100% germination rate (Figure 1c and 1d). The radicle emergence rate of both wild type and mutants decreased with 150 mM NaCl treatment. The wild-type seeds were affected severely with only 22.66% emergence rate, whereas the mutants pub30-1 and pub30-3 recorded 78.75% and 81.35% radicle emergence rate, respectively (Figure 1c and 1d). A similar trend was observed when the wild-type and pub30-1 mutant seeds germinated on MS medium containing two other concentrations of NaCl (100 and 125 mM; Figure S1). These results suggest that the pub30 mutants are more tolerant to salt stress in seed germination.

3.2 Functional antagonism between two highly homologous proteins PUB30 and PUB31 in response to salt stress

According to homologous sequence alignment, we found PUB30 shares 81% identity with PUB31 (Figure 2a). Both proteins contain a U-box domain that is essential for E3 ubiquitin ligase activity at the N-terminal region and several Armadillo (ARM)-repeat motifs that mediate protein–protein interaction at the C-terminal region (Figure 2a). In order to determine whether PUB31 has functional redundancy with PUB30, we obtained the pub31 mutant (Salk_054774) from ABRC. Quantitative RT-PCR results showed that the full length transcript of PUB31 could not be detected in the pub31 mutant (Figure S2). The pub30-1 pub31 double mutant was generated and used to observe the phenotype. There was no significant difference among the wild type, pub30-1, pub31, and pub30-1 pub31 double mutant during seed germination when they grow on normal MS medium for 12 days. However, when they grow on the MS medium containing 150 mM NaCl for 12 days, the pub31 mutant is a little sensitive to salt stress compared with the wild type, in contrast to the salt-tolerant phenotype of the pub30-1 mutant. The double mutant pub30-1 pub31 behaved in between the pub30-1 and pub31 single mutant for salt stress response and resembled the pub31 mutant phenotype in radicle emergence rate (Figure 2b and 2c). In the presence of 150 mM NaCl, the radicle emergence rate of the wild type, pub30-1, pub31, and the pub30-1 pub31 double mutant is 40.01%, 77.39%, 18.88%, and 24.80%, respectively (Figure 2c). These results suggest that PUB31 has a functional antagonism with PUB30 in salinity stress response.

3.3 Expression pattern and subcellular localization of PUB30

To examine the expression pattern of PUB30, real-time quantitative RT-PCR was used to detect the expression of PUB30 in different tissues, including root, stem, cauline leaf, rosette leaf, and silique. Relative higher expression level was found in root and stem, lower level in cauline and rosette leaves (Figure 3a). To explore the subcellular localization of the PUB30 protein, we constructed two plant expression vectors. In the GFP-PUB30 vector, green fluorescent protein (GFP) was fused to the N-terminus of PUB30 genomic sequence driven by the PUB30 native promoter. In the PUB30-YFP vector, yellow fluorescent protein (YFP) was fused to the C-terminus of PUB30 driven by a 35S promoter. These two vectors were transiently expressed in tobacco leaves. After incubation for 5 days, we found that these two vectors showed similar localization pattern where PUB30 was extensively localized in the cytoplasm, nucleus, and plasma membrane (Figure 3b and 3c). Additionally, the transgenic seeds harboring GFP-PUB30 in the pub30-1 mutant background were used to detect the radicle emergence on the MS medium containing 150 mM NaCl. The results showed that radicle emergence rate of two independent transgenic lines was reduced compared with that of the pub30-1 mutant, implying that the GFP-PUB30 fusion protein is functional in complementing the salt-tolerant phenotype of the pub30-1 mutant (Figure S3).

3.4 PUB30 is a functional U-box E3 ubiquitin ligase

To examine whether PUB30 is a functional E3 Ub ligase, we produced a full-length Arabidopsis PUB30 protein fused with MBP in E. coli and subsequently affinity purified from the soluble fraction. In vitro self-ubiquitination assays were performed in the presence of Arabidopsis E1 (UBA1), Arabidopsis E2 (UBC8), and His-tagged Ub. Polyubiquitination was detected by anti-Ub or anti-MBP antibody only when E1, E2, and MBP-PUB30 were all present. When any of them were omitted in the assay, polyubiquitination was not detected (Figure 4). This result suggests that PUB30 possesses E3 Ub ligase activity in vitro.

3.5 PUB30 interacts with BKI1 in vitro and in vivo

To investigate whether PUB30 affects the salt stress response via ubiquitination of target protein, we performed yeast two-hybrid assay to screen its target using PUB30 as bait. Some positive clones were isolated from the cDNA library made from 12-day-old Arabidopsis seedlings. One of these clones encoded the full length BKI1 protein. The interaction between PUB30 and BKI1 was confirmed by cotransformation of AD-BKI1 and BD-PUB30 into the yeast cells (Figure 5a). Although PUB30 has highest identity (81%) with PUB31 in Arabidopsis, they have functional antagonism in response to salt stress (Figure 2). In order to investigate whether PUB31 also interacts with BKI1, AD-BKI1 and BD-PUB31 were cotransformed into the yeast cells. The result showed that PUB31 does not interacts with BKI1 although PUB30 does (Figure 5a). These results suggest that BKI1 may be the target of PUB30 in salt tolerance mechanism.

To further confirm the interaction between PUB30 and BKI1, we performed pull-down assay. PUB30 and PUB31 cDNA were cloned into pCOLD-MBP vector (Takara) containing a MBP tag, individually. MBP-PUB30 and MBP-PUB31 fusion proteins were purified from E. coli strain BL21 (DE3) with amylose resin (BioLabs). BKI1 cDNA was inserted into pCOLD-GST vector (Takara) with a GST tag. GST-BKI1 was purified with Glutathione Sepharose™4B (GE Healthcare) and eluted with reduced GST. In vitro pull-down assays showed that GST-BKI1 was pulled down by MBP-PUB30 but not by MBP-PUB31 (Figure 5b). This result further confirmed that PUB30 specifically interacts with BKI1.

To confirm the interaction between PUB30 and BKI1 in vivo, FLAG tag was translationally fused to the N-terminal region of PUB30 or PUB31 and 6 × Myc tag was translationally fused to the N-terminal region of BKI1. These gene fusions were driven by a double 35S promoter. The resulting vector Myc-BKI1 was cotransfected with FLAG-PUB30 or FLAG-PUB31 constructs into the wild-type leaf protoplasts. Total extracted protein was used to detect the expression of Myc-BKI1 and FLAG-PUB30 or Myc-BKI1 and FLAG-PUB31 in each transformation with anti-Myc antibody or anti-FLAG antibody (Figure 5c). Myc-BKI1 protein was then immunoprecipitated with anti-Myc affinity gel. The result showed that only FLAG-PUB30 is pulled-down by Myc-BKI1 and FLAG-PUB31 is not (Figure 5c). This result suggests that PUB30 specifically interacts with BKI1 in vivo.

3.6 BKI1 protein is ubiquitinated and degraded by the proteasome

U-box E3 Ub ligase specifically recognizes substrate proteins and target them for degradation by the 26S proteasome by attaching a polyubiquitin chain. Thus, we first examined whether BKI1 is degraded by the 26S proteasome. The Arabidopsis seedlings expressing Myc-BKI1 were treated with 50 μM proteasome inhibitor MG132 for different time. The total protein was extracted to detect the protein level of Myc-BKI1. The result showed that the level of Myc-BKI1 protein increased gradually as the incubation time with MG132 was prolonged (Figure 6a). Serine hydroxymethyltransferase 1, which is degraded in a 26S proteasome-dependent manner, was used as a positive control of MG132 treatment (Zhou et al., 2012). This suggests that BKI1 protein could be degraded by the 26S proteasome.

Then to check whether BKI1 protein is ubiquitinated in vivo, we used a p62-agarose matrix that is capable of binding ubiquitinated proteins (Irigoyen et al., 2014) to enrich the ubiquitinated proteins from the transgenic plants overexpressing Myc-BKI1. When the p62 resin-enriched protein fraction was detected using anti-Myc antibodies, multiple high molecular mass bands above the Myc-BKI1 band were revealed in the transgenic plants but not in the wild type, indicating that Myc-BKI1 is modified by polyubiquitin chains in planta (Figure 6b).

3.7 PUB30 mediates BKI1 protein degradation

To determine whether PUB30 is required for BKI1 degradation, we used in vitro cell-free system to detect the degradation of GST-BKI1 protein expressed in E. coli. The total protein was extracted from Col-0 and the pub30-1 seedlings and incubated with GST-BKI1 protein at 23 °C for 0, 20, 40, and 60 min. Subsequently, immunoblot assay was performed with anti-GST antibody. Initially, there was no difference in the protein level of GST-BKI1 between the wild-type and the pub30-1 mutant extract. However, the protein level of GST-BKI1 was reduced more rapidly in the wild-type than in the pub30-1 mutant extract as the incubation time was increased (Figure 7a and 7b).

To further confirm whether BKI1 protein is targeted by PUB30 for ubiquitin-mediated degradation in vivo and to compare the BKI1 protein dynamics in the wild-type and the pub30-1 mutant, we performed an in vivo degradation assay. In this assay, the wild-type and the pub30-1 mutants were transformed with Myc-BKI1. Two transgenic lines showing comparable transcript levels of Myc-BKI1 were selected from both the wild-type and the pub30-1 mutant genotypes (Figure 8a). To increase the starting level of the Myc-BKI1 protein, these two transgenic lines were pretreated with MG132 for 16 hr as described previously (Jung et al., 2015) and then washed 5 times before being transferred to liquid MS medium containing 200 mM cycloheximide for 90 or 180 min. Total protein was isolated from the two transgenic lines and used to detect the degradation of Myc-BKI1. The immunoblot assays showed that degradation of the Myc-BKI1 protein was slowed down in the pub30-1 mutant compared with the wild type (Figure 8b and 8c). All these results suggest that BKI1 undergoes 26S proteasome-dependent degradation, and PUB30 is involved in this process.

3.8 BKI1 functions in salinity stress responses

In order to functionally characterize BKI1 in seed germination under salt stress, the transgenic seeds harboring Myc-BKI1 were sown on 1/2 MS medium with or without 150 mM NaCl for 9 days. In contrast to the wild-type, transgenic plants overexpressing BKI1 showed higher radicle emergence rate on MS medium containing 150 mM NaCl, that is, 34.42%, 87.58%, and 11.43% for transgenic line Col-0 #4, Col-0 #7, and the wild type, respectively (Figure 9a and 9c). The radicle emergence rate was positively correlated with the transcript level of BKI1 in transgenic plants Col-0 #7, Col-0 #4, and the wild type (Figure 9b).

In addition, we also used the bki1 loss-of-function mutant (Jiang et al., 2015) to characterize its function in seed germination under salt stress. When the wild-type and the bki1 mutant were grown on normal MS medium, no significant difference in seed germination was observed (Figure 9d). The radicle emergence of the wild-type and the bki1 mutant grown on MS medium was almost 100% (Figure 9e). However, the bki1 mutant displayed salt-hypersensitive phenotype when germinated on MS medium containing 100 mM NaCl (Figure 9d). The fraction of seeds with radicle emergence was only 55% for the bki1 mutant, compared with 85.8% for the wild type (Figure 9e). These results suggest that BKI1 is a positive regulator of the salt-tolerance phenotype.