Quantitative Proteomic Analysis of ER Stress Response Reveals both Common and Specific Features in Two Contrasting Ecotypes of Arabidopsis thaliana

Accumulation of unfolded and misfolded proteins in endoplasmic reticulum (ER) elicits a well-conserved response called the unfolded protein response (UPR), which triggers the upregulation of downstream genes involved in protein folding, vesicle trafficking, and ER-associated degradation (ERAD). Although dynamic transcriptomic responses and the underlying major transcriptional regulators in ER stress response in Arabidopsis have been well established, the proteome changes induced by ER stress have not been reported in Arabidopsis. In the current study, we found that the Arabidopsis Landsberg erecta (Ler) ecotype was more sensitive to ER stress than the Columbia (Col) ecotype. Quantitative mass spectrometry analysis with Tandem Mass Tag (TMT) isobaric labeling showed that, in total, 7439 and 7035 proteins were identified from Col and Ler seedlings, with 88 and 113 differentially regulated (FC > 1.3 or <0.7, p < 0.05) proteins by ER stress in Col and Ler, respectively. Among them, 40 proteins were commonly upregulated in Col and Ler, among which 10 were not upregulated in bzip28 bzip60 double mutant (Col background) plants. Of the 19 specifically upregulated proteins in Col, as compared with that in Ler, components in ERAD, N-glycosylation, vesicle trafficking, and molecular chaperones were represented. Quantitative RT-PCR showed that transcripts of eight out of 19 proteins were not upregulated (FC > 1.3 or <0.7, p < 0.05) by ER stress in Col ecotype, while transcripts of 11 out of 19 proteins were upregulated by ER stress in both ecotypes with no obvious differences in fold change between Col and Ler. Our results experimentally demonstrated the robust ER stress response at the proteome level in plants and revealed differentially regulated proteins that may contribute to the differed ER stress sensitivity between Col and Ler ecotypes in Arabidopsis.


Introduction
In eukaryotic cells, protein folding machineries in the secretory pathways are well maintained to ensure the accuracy of protein folding and timely elimination of misfolded proteins accumulated in the endoplasmic reticulum (ER) [1]. When unfolded and misfolded proteins are built-up in ER during plant development and its adaptation to environmental stresses, a well-conserved response called the unfolded protein response (UPR), is elicited to sense transduce signals in the ER and regulate gene expression in the nucleus, enhancing the protein folding capacity of ER to cope with the protein folding demands [2].

TMT-Based Proteomic Analysis of ER Stress Responses in Col and Ler
To understand the ER stress sensitive phenotype of Ler, first, we checked the expression of three UPR downstream marker genes (ERDJ3A, CNX1, and PDI5) in Col and Ler plants under ER stress conditions, and the results showed that the ER stress-induced upregulation of ERDJ3A, CNX1, and PDI5 was quite similar between Col and Ler at 8 h, the fold induction of ERDJ3A was even much higher in Ler than that in Col (Figure 2), which could not sufficiently explain the observed difference in ER stress sensitivity between these two ecotypes. Although proteomic analysis of ER stress responses has been carried out in human and mouse cells [34,35], there has been no such proteomic study reported in plants. Therefore, we employed TMT-based quantitative proteomics to compare the ER stress responses in Col and Ler plants. In total, 34,799, 31,024 and 35,109 peptides, corresponding to 7439, 7046, and 7502 proteins, were identified in Col, Ler, and DM plants, respectively, among which 7155, 6750, and 7236 proteins contained quantifiable information in at least one of the three group of plants ( Figure 3). Principal components analysis (PCA) results showed good reproducibility of data for the control (CK) and ER stress treatment (TM) samples in each group ( Figure S1). Volcano plot and protein abundance data showed the differential expression of proteins between CK and TM in each group of plants ( Figure 4A-C and Table 1). According to the variance analysis with p < 0.05 and 1.3-or 0.7-fold change, a commonly used stringent parameter setting for TMT-based comparative proteomics [36][37][38], in total, 59, 82, and 111 quantified proteins were considered to be upregulated, and 29, 31, and 57 proteins were considered to be downregulated proteins at the protein level by ER stress in Col, Ler, and DM plants, respectively ( Figure 4D-F, Table S1). The enrichment of gene ontology (GO) analysis showed that similar biological process (BP) or molecular function (MF) were enriched in Col and Ler, in which GO terms related to protein folding and degradation were commonly found ( Figures S2 and S3). GO terms of cellular component (CC) related to the secretory pathways were enriched in Col, Ler, and DM plants ( Figure S4). These results were consistent with previous transcriptomic analysis that the expression level of molecular chaperones, protein disulfide isomerases, and ER-associated degradation proteins was affected by ER stress in Arabidopsis [9,24,39,40]. responses has been carried out in human and mouse cells [34,35], there has been no such proteomic study reported in plants . Therefore, we employed TMT-based quantitative proteomics to compare  the ER stress responses in Col and Ler plants. In total, 34,799, 31,024 and 35,109 peptides, corresponding to 7439, 7046, and 7502 proteins, were identified in Col, Ler, and DM plants, respectively, among which 7155, 6750, and 7236 proteins contained quantifiable information in at least one of the three group of plants ( Figure 3). Principal components analysis (PCA) results showed good reproducibility of data for the control (CK) and ER stress treatment (TM) samples in each group ( Figure S1). Volcano plot and protein abundance data showed the differential expression of proteins between CK and TM in each group of plants ( Figure 4A-C and Table 1). According to the variance analysis with p < 0.05 and 1.3-or 0.7-fold change, a commonly used stringent parameter setting for TMT-based comparative proteomics [36][37][38], in total, 59, 82, and 111 quantified proteins were considered to be upregulated, and 29, 31, and 57 proteins were considered to be downregulated proteins at the protein level by ER stress in Col, Ler, and DM plants, respectively ( Figure 4D-F, Table  S1). The enrichment of gene ontology (GO) analysis showed that similar biological process (BP) or molecular function (MF) were enriched in Col and Ler, in which GO terms related to protein folding and degradation were commonly found ( Figures S2 and S3). GO terms of cellular component (CC) related to the secretory pathways were enriched in Col, Ler, and DM plants ( Figure S4). These results were consistent with previous transcriptomic analysis that the expression level of molecular chaperones, protein disulfide isomerases, and ER-associated degradation proteins was affected by ER stress in Arabidopsis [9,24,39,40].   Nine-day-old Col and Ler seedlings grown on 1 2 MS plates were treated without or with 5 µg/mL tunicamycin (TM) for 8 h or 12 h, and total RNA was extracted for quantitative RT-PCR (RT-qPCR). Fold change is the expression level of genes in TM-treated plants relative to that in non-treated (CK) plants, both of which were normalized to that of the control ACTIN. Error bars represent SE (n = 3). Asterisks indicate significance levels as compared with two ecotypes in t-test. (*, p < 0.05; **, p < 0.01).

Differentially Regulated Proteins by ER Stress between Col and Ler
In order to understand the observed differential ER stress sensitivity between Col and Ler, we compared the differentially upregulated or downregulated proteins (p < 0.05, FC > 1.3, or FC < 0.7) among Col, Ler, and DM plants. Venn diagrams showed that there were 40 proteins that were commonly upregulated in both Col and Ler, among which 30 proteins were not upregulated at the protein level in DM plants ( Figure 5A and Table S2). These 40 commonly upregulated proteins were enriched in GO terms associated with the secretory subcellular locations ( Figure 5C), which was consistent with their role in UPR. There were 19 proteins that were specifically upregulated in Col plants as compared with that in Ler plants, and 16 of them were not upregulated at the protein level in DM plants ( Figure 5A, Tables 2 and 3, and Table S2). In contrast, there were 24 proteins that were specifically downregulated at the protein level in Col plants as compared with that in Ler plants ( Figure 5B, Tables 4 and 5,  and Table S2), many of which were related to growth and stress responses.    Among the 19 specifically upregulated proteins in Col, five proteins (FLA1, UTR1, UGT71C2, AT3G51000, and AT5G19250) are related to glyco-modification; five proteins (CRT1A, ERO1, FLOT1, SEC62, and AT3G16990) are involved in ER protein folding and trafficking; two proteins (EBS7 and OS9) are important components in ER-associated degradation (ERAD); two proteins (TGA6 and NAC091) are transcription factors (Table 2).
Calnexin (CNX) and calreticulin (CRT) are lectins that recognize oligosaccharide side chains on glycoproteins and form CNX/CRT protein-folding cycles to serve as a quality control system for successful protein folding in ER [41]. Defects in all three Arabidopsis CRT genes affected protein folding and conferred high sensitivity to drought [42]. During CNX/CRT cycles, continuous glucose trimming by glucosidase II (GCSII) and UDP-glucose-dependent re-glucosylation of unfolded glycoproteins by UDP-glucose:glycoprotein glucosyltransferase (UGGT) takes place [43]. Nucleotide sugar transporter is required in the CNX/CRT cycle to transport UDP-glucose from the cytosol to the ER lumen, which is controlled by two proteins UTR1 and UTR3 in Arabidopsis [44,45]. In the CNX/CRT cages, forming and reforming disulfide bonds through repeated oxidation and reduction catalyzed by protein disulfide isomerase (PDI) and other thiodisulfide oxidoreductases are important for correct protein folding in the ER, and the formation of protein disulfide bonds requires oxidizing equivalents, which are supplied by ER oxidoreductin 1 (ERO1) [46,47]. CRT1A, UTR1, and ERO1 are more accumulated at the protein level in Col than that in Ler under ER stress conditions, suggesting that ER protein folding is more efficient in Col than that in Ler plants. Autophagy, another quality control system in plants, selectively targets and degrades damaged organelles (e.g., ER) and misfolded proteins inside the organelles. SEC62, originally considered to be a constituent of the translocon complex regulating protein import in the mammalian ER, also functions as an ER-resident autophagy receptor to selectively deliver ER components to the autolysosomal system for clearance to maintain ER homeostasis [48]. Recent studies have confirmed the role of Arabidopsis SEC62 in ER-phagy under ER stress in which SEC62 likely functioned as an ER-phagy receptor in plants [49,50]. The accumulation of SEC62 at the protein level under ER stress conditions is only detected in Col, which is consistent with the notion that that ER-phagy in Col is more activated than that in Ler plants.
ERAD is also an ER-mediated protein quality control system that timely recognizes and removes misfolded proteins from the ER to the cytosol for degradation via the 26S proteasome [51]. There are two major complexes of the ERAD system in plants, the HMG-CoA reductase degradation 1 (HRD1) complex and the degradation of alpha2 10 (DOA10) complex [1]. In the conserved HRD1 complex, Arabidopsis osteosarcoma-9 (AtOS9)/EMS bri1 suppressor6 (EBS6), and AtHRD3/EBS5 recognize presumably misfolded ERAD targets [52][53][54], while E2 conjugase UBC32 and E3 ligase HRD1 ubiquitinate and degrade ERAD substrates via the 26S proteasome [53,55]. EBS7 is a plant-specific ERAD component that interacts with and stabilizes HRD1a in the ER membranes, regulating protein stability of the misfolded clients BRI-9 and BRI-5 [56]. In the current proteomics study, the upregulation of EBS7 and OS9 proteins at the protein level was only detected in Col plants, suggesting that the ERAD machineries in Col plants are more efficient than that in Ler plants, which is correlated to the observed difference in ER stress sensitivity between two ecotypes.
UPR is an important component of immunity to host pathogens and of systemic acquired resistance (SAR). Mutant plants such as ire1a, ire1b, and bzip60 that are compromised in the UPR are more susceptible to bacterial pathogens and less able to establish SAR to the bacteria when treated with salicylic acid (SA) [57]. The bZIP transcription factor TGA6 functions redundantly with TGA2 and TGA5 in SA-mediated SAR [58,59]. The NAC transcription factor NAC091 (also known as TIP) is closely related to NAC062, a plant-specific membrane-associated transcription factor involved in UPR [60]. These two proteins were specifically upregulated at the protein level by ER stress in Col plants. In the future, it is worthwhile to investigate in detail whether TGA6 and/or NAC091 regulate UPR in plants.

Gene Expression Analysis of Differentially Upregulated Proteins by ER Stress between Col and Ler
In order to know whether the differentially regulated proteins by ER stress between Col and Ler are correlated to differential gene expression levels, quantitative RT-PCR was performed. The results showed that 11 genes out of 19 genes encoding specifically upregulated proteins in Col plants were upregulated at the transcription level while the remaining eight genes (AT1G65720, AT2G24980, AT3G12250, AT3G20920, AT5G19250, AT5G44120, AT5G55730, and AT5G64400) were not upregulated, even after 12 h of treatment at the transcription level by ER stress treatments (fold change >2, p < 0.05) ( Figure 6A,B). For the abovementioned 11 genes, none of the denes differed in terms of transcript upregulation ratio (fold change >1.3 or <0.7, p < 0.05) in plants between Col and Ler ( Figure 6A,B). The above results indicated that these proteins were regulated by ER stress after gene expression, possibly at translational or post-translation levels in two different Arabidopsis ecotypes. Except for AT5G64400, which was upregulated (fold change >2, p < 0.05) in Ler plants, there were seven out of eight genes, including SEC62 and TGA6, that were not upregulated or downregulated (fold change >2, p < 0.05) at the transcription level by ER stress in both ecotypes ( Figure 6A,B), suggesting that these encoding proteins were upregulated by ER stress only at the protein level. gene expression, possibly at translational or post-translation levels in two different Arabidopsis ecotypes. Except for AT5G64400, which was upregulated (fold change >2, p < 0.05) in Ler plants, there were seven out of eight genes, including SEC62 and TGA6, that were not upregulated or downregulated (fold change >2, p < 0.05) at the transcription level by ER stress in both ecotypes ( Figure 6A,B), suggesting that these encoding proteins were upregulated by ER stress only at the protein level.

Concluding Remarks
In conclusion, the current study reports the different ER stress sensitivities in two Arabidopsis ecotypes Col and Ler. Quantitative proteomics reveals the common and specific proteins that are differentially regulated by ER stress at the protein level in these two ecotypes. Among the 19 differentially upregulated proteins between Col and Ler, transcripts of 11 proteins were upregulated by ER stress at the gene expression level, however, the transcript upregulation ratios of them were not significantly different in plants between Col and Ler. These 11 proteins are involved in protein folding, vesicle trafficking, and protein degradation, and the relative higher ER stress-induced protein level in Col is correlated to the relative less ER stress sensitivity in Col as comparing with that in Ler. There are with identified proteins, of which the transcript level is not affected, but the protein level is upregulated by ER stress in Col but not in Ler, further demonstrating the complementary power of proteomics c in identifying new components in UPR in plants. Future genetic experiments are needed to further dissect the molecular mechanisms underlying elevated ER stress sensitivity in Ler ecotype in Arabidopsis.

Plant Materials and Endoplasmic Reticulum (ER) Stress Treatment
Wild-type Arabidopsis thaliana lines of Col (Col-0) and Ler (Ler-0) were used in this study. The bzip28 bzip60 double mutant (DM) plants were obtained by crossing the T-DNA mutants bzip28 (SALK_132285) with bzip60 (SALK_050203), both of which were in the Col background. For ER stress sensitivity assays, seeds were germinated on agar plates containing half-strength Murashige and Skoog (MS) salts, 1.2% sucrose, 0.05% MES, and different concentrations of tunicamycin (TM) (Sigma, CA, USA) as described in the text, pH 5.7. Seedlings were grown in an illuminated growth chamber at 22 • C for 9 days after stratification at 4 • C for 2 days and plant phenotypes were recorded. For proteomics analysis, 9-day-old seedlings grown on 1/2 MS plates were transferred to liquid MS medium plus 5 µg/mL TM for 12 h, and whole seedlings were harvested for protein extraction. For gene expression analysis, 9-day-old seedlings grown on 1/2 MS plates were transferred to liquid MS medium plus 5 µg/mL TM for 8 or 12 h, and whole seedlings were harvested for RNA extraction. There were three biological replicates for both proteomics and gene expression analysis.

Protein Extraction, Digestion, and TMT Labeling
TMT-based proteomics was done with commercial service available at PTM BIO, Hangzhou, China (http://www.ptm-biolab.com.cn/). Plant tissues were ground in liquid nitrogen and powders were transferred to 5 mL centrifuge tubes. After that, four volumes of lysis buffer (8 M urea, 1% Triton-100, 10 mM dithiothreitol, and 1% Protease Inhibitor Cocktail) was added, followed by sonication three times on ice using a high intensity ultrasonic processor (Scientz, Ningbo, China) and centrifugation at 20,000× g at 4 • C for 10 min. Finally, the supernatant was precipitated with cold 20% TCA for 2 h at −20 • C. The protein was dissolved in 8 M urea and the protein concentration was determined with a BCA kit, according to the manufacturer's instructions. For digestion, the protein solution was reduced (5 mM dithiothreitol, 30 min) at 56 • C and alkylated with iodoacetamide (11 mM, 15 min) at room temperature in the dark. Then, the protein sample was diluted by adding 100 mM triethylamonium bicarbonate (TEAB). Finally, trypsin was added at 1:50 trypsin-to-protein mass ratio for the first digestion overnight and 1:100 trypsin-to-protein mass ratio for a second 4 h digestion. For TMT labeling, peptide was desalted with Strata X C18 SPE column (Phenomenex, CA, USA) and vacuum dried. After that, peptide was reconstituted in 0.5 M TEAB and processed according to the manufacturer's protocol for the TMT kit. Briefly, one unit of TMT reagent was thawed and reconstituted in acetonitrile. Then, the peptide mixtures were incubated for 2 h at room temperature and pooled, desalted, and dried by vacuum centrifugation.

HPLC Fractionation and LC-MS/MS Analysis
HPLC fractionation was performed with high pH reverse-phase HPLC. Briefly, peptides were dissolved in 0.1% formic acid, and then directly loaded onto reversed-phase analytical columns (15 cm in length, 75 µm i.d.). The gradient was increased from 6% to 23% (0.1% formic acid in 98% acetonitrile) over 26 min, 23% to 35% in 8 min, 35% to 80% in 3 min, and then held at 80% for the last 3 min, all at a constant flow rate (400 nL/min) on an EASY-nLC 1000 UPLC system. The peptides were sent to NSI source followed by tandem mass spectrometry (MS/MS) in Q ExactiveTM Plus (Thermo, CA, USA) coupled online to the UPLC in the Orbitrap with routine setting parameters. The mass spectrometry proteomics data were deposited to the ProteomeXchange Consortium via the PRIDE partner repository with the dataset identifier PXD022630.

Database Search and Bioinformatics Analysis
The resulting MS/MS data were processed using Maxquant search engine (v.1.5.2.8). Tandem mass spectra were searched against the Uniprot database concatenated with reverse decoy database using Col or Ler sequences. Trypsin/P was chosen as the cleavage enzyme allowing up to 2 missing cleavages. The mass tolerance for precursor ions was set as 10 ppm in First search and 5 ppm in Main search, and the mass tolerance for fragment ions was set as 0.02 Da. Cys carbamidomethylation was specified as a fixed modification, and acetylation modification and oxidation on Met were specified as variable modifications. FDR was adjusted to <1% and minimum score for modified peptides was set to >40. Differential expression proteins were obtained based on fold changes and P values as indicated in the text. Gene ontology (GO) annotations were retrieved from the UniProt-GOA database (http://www.ebi.ac.uk/GOA/) and classified as three categories, i.e., biological process, cellular component and molecular function. For each category, a two-tailed Fisher's exact test was employed to test the enrichment of the differentially expressed protein against all identified proteins. The GO with a corrected p-value <0.05 is considered significant.

Quantitative Gene Expression Analysis
RNA from 9-day-old seedlings were extracted with an RNA Prep Pure Plant kit (Tiangen, Shanghai, China). For reverse transcription, 1 mg of total RNA with oligo (dT) and random 6 mer primers were used to synthesize cDNAs in a 20 µL reaction using a M-MLV reverse transcriptase kit (TaKaRa, Shanghai, China). Quantitative PCR (qPCR) was performed using the SuperReal PreMix Color kit (Tiangen) in a CFX96 real-time system (Bio-Rad, CA, USA). All the primers used are included in Table S3.