The Arabidopsis thaliana plants used in this study were of Colombia-0 (Col-0) background. The seeds were surface-sterilized and pre-germinated on Murashige and Skoog (MS)-salts agar plates at 22°C under a photoperiod of 16 h light/8 h dark. The SALK T-DNA insertion lines were obtained from the Arabidopsis Biological Resource Center (ABRC, http://abrc.osu.edu/). The double mutants were generated by crosses of the corresponding single mutants.

Total RNAs were extracted from seedlings using an RNAprep Pure Plant Kit (Tiangen, Beijing, China) according to the manufacturer’s instruction. The first-strand cDNA was synthesized from 5 μg of total RNAs using the SuperScript III First-Strand Synthesis System (Invitrogen, CA, USA) as described by the manufacturer’s instruction. The Real-time PCR assays were run using 2×Power SYBR Green PCR Master Mix (Applied Biosystems, www.appliedbiosystems.com) and the gene-specific primers (S1 Table) on an ABI 7500 Real-time instrument (Applied Biosystems, www.appliedbiosystems.com). The PCR program was set as 95°C for 10 min, followed by 40 cycles of 95°C for 15 s and 60°C for 1 min. The gene expression levels in three biological replicates were normalized to that of ACTIN2/8 in the same cDNA samples tested. The relative expression levels were calculated with the ΔCt (threshold cycle) method.

The coding DNA sequence (CDS) of TMS1 DnaJ Domain (DnaJ) was amplified by polymerase chain reaction (PCR) using the gene-specific primer pairs (S1 Table). The CDS of the mutated TMS1 DnaJ domain (DnaJ-mt) was generated by the overlap PCR method [23] using the gene-specific primer pairs (S1 Table). The amplified CDS fragments were purified and digested with EcoRI and XhoI, and cloned into the expression vector pGEX4T-1. The resulting constructs were introduced into E. coli strain BL21 (TransGen Biotech, Beijing, China) to produce recombinant TMS1 DnaJ domain proteins with a GST tag (GST-DnaJ) and the mutated TMS1 DnaJ domain with a GST tag (GST-DnaJ-mt). Purification of the GST-fusion proteins was as described previously [7]. After the GST fusion proteins were purified, the proteins were dialyzed using the dialyzing buffer (20 mM Hepes, pH 6.8, 75 mM KOAC, 250 mM sorbitol, 5 mM MgOAC₂, 10% glycerol) overnight for further experiments.

The full-length CDSs encoding the BiP1 and BiP3 proteins without the signal peptides were amplified by PCR using the gene-specific primer pairs (S1 Table), respectively. The amplified CDS fragments of BiP1 and BiP3 were purified and digested with EcoRI and XhoI and cloned into the expression vector pET30a (+) to generate the fusion protein-expressing constructs BiP1-His and BiP3-His, respectively. The expression, purification and dialyzation of the fusion proteins with an amino-terminal His⁺⁶ tag from E. coli strain BL21 (TransGen Biotech, Beijing, China) were as described previously [24].

The in vitro GST Pull-down assays were performed as described by Corsi and Schekman [24] with a small modification. Glutathione Sepharose 4B beads (GE healthcare, 17-0756-01, Sweden) were equilibrated with 1 mL binding buffer (20 mM Hepes, pH 6.8, 100 mM KCl, 5 mM MgCl₂, 0.1% NP-40, 2% [vol/vol] glycerol, 1 mM DTT, 0.2 mM AEBSF, and 1 mM EDTA) before used. 30 μg GST-DnaJ, GST-DnaJ-TB or GST proteins were added into 50 μL of 50% Glutathione Sepharose 4B bead suspension in the binding buffer. Then the volume was increased to 1 mL with the binding buffer. The reaction tubes were rotated at 4°C for 1 h. The unbound proteins were removed by centrifugation at 3000 g for 2 min at 4°C followed by two washes with 500 μL binding buffer. The purified BiP1 (30 μg) or BiP3 (30 μg) and ATP (1 mM) were then added in a final volume of 1 mL, and then were rotated at 4°C for 2 h. The unbound proteins were removed by a series of four washes with 1 mL binding buffer mentioned above. The proteins binding to the beads were solubilized in Laemmli sample buffer (BioRad Labs, Hercules, CA) by boiling, and separated by SDS-PAGE. The in-put GST, GST-DnaJ and GST-DnaJ-mt proteins were examined by Western blot using the anti-GST antibody (Sigma, A5838, USA). The BiP1-His and BiP3-His proteins were detected by Western blot using the anti-His antibody (Sigma, A5588, USA).

The ATPase enzyme activity assay kit from Innova Biosciences (Innova Biosciences, Cambridge, UK) was used to perform in vitro ATPase enzyme activity assays. According to the manufacturer’s instruction, in each reaction, 10 μg BiP1-His or BiP3-His proteins were mixed with an different amount of the GST-DnaJ or GST-DnaJ-mt or naked GST proteins in 100 mL assay buffer (pH 6.8) supplied with 5 mM MgCl₂ and 1 mM ATP, respectively. The concentration of the GST-DnaJ or GST-DnaJ-mt or naked GST proteins was increased in a gradient, while the concentration of BiP1-His or BiP3-His proteins was kept constant. The reaction mixtures were incubated at 25°C for 45 min, which was determined to be in the linear range of the reactions. ATP hydrolysis (%) was measured by OD₆₃₀ values using a plate reader (Bio-Rad labs, Hercules, CA) as described by the manufacturer’s instruction. The ATPase enzyme activities were calculated using the standard curve generated by the concentration gradients of inorganic phosphate (Pi).

The full-length CDSs of BiP1, BiP3 and TMS1 (without the signal peptides) and CDS of TMS1 DnaJ domain were amplified by PCR using gene-specific primer pairs (S1 Table). The amplified BiP1 and BiP3 DNA fragments were purified and digested with KpnI and SalI, and inserted into 35S::Nluc vector. The amplified DNA fragments were purified, digested with KpnI and BamHI, and inserted into 35S::Cluc vector. Agrobacteria-mediated transient expression, CCD imaging and LUC activity measurement were carried out as described by Chen et al. [25].

Previous studies have demonstrated that the expression of TMS1 can be induced by heat shock and tunicamycin [7, 8]. To further investigate the involvement of TMS1 in unfolded protein response in endoplasmic reticulum and in cytoplasm, we performed the DTT and AZC induction assays. DTT is a reductant agent that can prevent formation of disulfide bonds. Treatment of DTT can induce unfolded protein response (UPR) in endoplasmic reticulum [26]. AZC is a proline analog. Treatment of AZC can lead to unfolded protein response in endoplasmic reticulum and in cytoplasm (CPR) [27]. As shown in Fig 1A, when the 10-day-old wild-type seedlings were treated with 37°C for 1 h, 5 mM DTT for 2 h and 10 mM AZC for 3 h, the TMS1 mRNA levels were increased 724-, 34- and 330-folds, respectively, compared to those in the responding control seedlings grown at 22°C or treated with H₂O or DMSO. The results indicated that the expression of TMS1 was induced not only by heat shock, but also by DTT and AZC treatments, indicating that TMS1 is related to UPR and CPR.

The transcription factors bZIP28 and bZIP60 have been identified as the important UPR regulators for mitigating the ER stress in Arabidopsis thaliana [26, 28–32]. Previous study showed that the expression of BiP3, a marker gene for UPR, was reduced in the bzip28 bzip60 double mutant [33]. Among the Arabidopsis BiP1, BiP2 and BiP3, BiP1 and BiP2 are located closely to each other. Furthermore, their sequence similarity reaches to 99% [13, 14]. Therefore, only BiP1 and BiP3 were selected for the following assays. We measured and compared the expression of TMS1, BiP1 and BiP3 in the bzip28 bzip60 double mutant.

The bzip28 bzip60 double mutant used in this assay was generated using the SALK T-DNA insertion lines of bZIP28 (SALK_132285C) and bZIP60 (SALK_050203C) from ABRC (S1 Fig). The 10-day-old wild type and bzip28 bzip60 double mutant seedlings were treated with 5 mM DTT for 2 h. Meanwhile, H₂O was used as a negative control. As shown in Fig 1B, in wild type seedlings, the expression of the BiP1/2, BiP3 and TMS1 was increased 9-, 500- and 34-folds when treated with 5 mM DTT, respectively, compared to those in the control seedlings, whereas in bzip28 bzip60 seedlings, the expression of the BiP1/2, BiP3 and TMS1 was increased 1-, 3- and 2-folds when treated with 5 mM DTT, respectively. Therefore, the expression of the BiP1/2, BiP3 and TMS1 was significantly reduced in the bzip28 bzip60 double mutant, suggesting that TMS1 may participate in the bZIP28/bZIP60-related signal pathway.

Firefly Luciferase Complementary imaging (LCI) Assay [25] was performed to verify if the TMS1 acts as a DnaJ-like partner and interacts with the BiPs. The CLuc-TMS1 and Cluc-DnaJ were paired with the BiP1-Nluc and BiP3-Nluc, respectively and coexpressed in N. benthamiana leaves using Agrobacteria-mediated transient expression method [25]. CLuc-TMS1/Nluc, CLuc-DNAJ/Nluc, Cluc/BiP1-NLuc and Cluc/BiP3-Nluc pairs were used as the negative controls. Cluc-SLAC1/GHR1-Nluc pair was used as a positive control [34]. As shown in Fig 2A, coexpression of the positive control pair Cluc-SLAC1/GHR1-Nluc and the test pair CLuc-TMS1/BiP1-Nluc exhibited strong LUC signals as revealed using a low-light imaging system after adding luciferin, the substrate for firefly LUC. In contrast, the negative control pairs CLuc-TMS1/Nluc and Cluc/BiP1-NLuc showed only the background level LUC activity. This result indicated that TMS1 could interact with BiP1 in LCI assay system. Similarly, TMS1 also could interact with BiP3 as demonstrated by the same assay system (Fig 2B). Furthermore, the DnaJ domain of TMS1 also could interact with BiP1 and BiP3 in the same system, respectively (Fig 2C and 2D).

To further verify the binding of TMS1 DnaJ domain to the BiPs, we performed the in vitro GST pull-down experiments. The CDS of the TMS1 DnaJ domain containing 54 amino acid residues (the 27–80 aa in TMS1) (Fig 3A) was fused with the GST CDS to construct the expression vector GST-DnaJ. Studies showed that a conserved three-amino acid motif HPD in the DnaJ domains is essential for the interaction of DnaJ proteins with BiPs [24, 35, 36]. Therefore, a parallel mutant fusion protein expression vector GST-DnaJ-mt was constructed for assessing the specificity of the DnaJ domain interaction with the BiPs, in which the HPD motif was mutated to QPD (Fig 3B). The GST-expressing vector GST was constructed to produce the GST protein, which was used as a negative control. The CDSs of BiP1 and BiP3 without signal peptides were cloned into pET30a (+) vector to construct the BiP-expressing vectors, BiP1-His and BiP3-His, respectively. All the constructs were introduced into the E. coli strain BL21 (DE3) to produce and purify the proteins for pull-down and ATPase enzyme activity assays (see Materials and Methods). The quality of the purified proteins was measured by SDS-PAGE fraction and Coomassie brilliant blue (CBB) staining (Fig 3C). As shown in Fig 3D and 3E, in presence of 1 mM ATPs, the GST-DnaJ fusion protein could bind to the BiP1 and BiP3 proteins, respectively. In contrast, the GST-DnaJ could not bind to the BiP1 or BiP3 in the absence of ATP (Fig 3D and 3E). Furthermore, the GST-DnaJ-mt and GST, the negative control proteins, also did not bind to the BiP1 and BiP3, either in the presence or absence of ATPs (Fig 3D and 3E). These results indicated that the TMS1 DnaJ domain could bind to the BiP1 and BiP3 in an ATP-dependent manner.

The ATPase enzyme activity assays were performed to analyze the correlation of the ATPase enzyme activities of the BiP proteins with the concentration of the GST-DnaJ. The ATPase enzyme activities of the BiP proteins were measured by monitoring the final phosphate concentration in the reaction systems using the plate reader after incubated for 45 minutes (see Materials and Methods). As shown in Fig 4A and 4B, in a constant concentration of the BiP proteins, the ATPase enzyme activities of the BiP proteins were enhanced by increasing the concentration of GST-DnaJ. In contrast, increasing of GST and GST-DnaJ-mt concentration did not stimulate the ATPase enzyme activities of the BiP proteins. Therefore, the TMS1 DnaJ domain could stimulate the ATPase enzyme activities of BiP1 and BiP3.

Both TMS1 (ERdj3A) and ERdj3B are orthologs of yeast Scj1p [8]. Orthologs of ERdj3B are found in human, mouse, fruit fly, nematode and plant genomes. However, the orthologs of TMS1 are only found in plant genomes [8]. Study has shown that ERdj3B, SDF2 and BiPs can form a complex to play roles in plant immunity, but TMS1 does not interact with SDF2 in Y₂H system [15]. To further study the relationship between TMS1 and ERdj3B, their expression patterns were compared. The expression of both TMS1 and ERdj3B could be induced by heat shock, DTT, and AZC in seedlings, but the expression levels of ERdj3B were much lower than those of TMS1 (Fig 5A). Both TMS1 and ERdj3B were expressed in seedlings, roots, stems, leaves, flowers, inflorescences, pollen grains and siliques from wild type plants. However, the expression levels of TMS1 were different in the various tissues, with the highest level in pollen grains and the lowest level in stems (Fig 5B), showing a pattern different from that of ERdj3B. Furthermore, when treated various wild type plant tissues by 37°C heat shock for 1 hour, the expression of TMS1 in various tissues was increased more efficiently than that of ERdj3B (Fig 5C). In addition, the expression of ERdj3B was slightly downregulated in tms1-1 (Fig 5D). Taken together, TMS1 and ERdj3B had different expression patterns.