Global Ubiquitome Profiling Revealed the Roles of Ubiquitinated Proteins in Metabolic Pathways of Tea Leaves in Responding to Drought Stress

Drought stress often affects the expression of genes and proteins in tea plants. However, the global profiling of ubiquitinated (Kub) proteins in tea plants remains unearthed. Here, we performed the ubiquitome in tea leaves under drought stress using antibody-based affinity enrichment coupled with LC-MS/MS analysis. In total, 1,409 lysine Kub sites in 781 proteins were identified, of which 14 sites in 12 proteins were up-regulated and 123 sites in 91 proteins down-regulated under drought stress. The identified Kub proteins were mainly located in the cytosol (31%), chloroplast (27%) and nuclear (19%). Moreover, 5 conserved motifs in EKub, EXXXKub, KubD, KubE and KubA were extracted. Several Kub sites in ubiquitin-mediated proteolysis-related proteins, including RGLG2, UBC36, UEV1D, RPN10 and PSMC2, might affect protein degradation and DNA repair. Plenty of Kub proteins related to catechins biosynthesis, including PAL, CHS, CHI and F3H, were positively correlated with each other due to their co-expression and co-localization. Furthermore, some Kub proteins involved in carbohydrate and amino acid metabolism, including FBPase, FBA and GAD1, might promote sucrose, fructose and GABA accumulation in tea leaves under drought stress. Our study preliminarily revealed the global profiling of Kub proteins in metabolic pathways and provided an important resource for further study on the functions of Kub proteins in tea plants.

Ubiquitination can mark proteins for degradation via the proteasome, alter protein subcellular location, affect their activities, and promote or inhibit protein interactions [10][11][12] . In plants, ubiquitination was associated with DNA damage response, membrane transport and transcriptional regulation, as well took part in enzymatic activity regulation and stress responses 13,14 . However, the role of ubiquitination in tea plants remains unearthed.
To investigate the possible mechanisms of Kub proteins in tea plants under drought stress, we studied the ubiquitome using antibody-based affinity enrichment coupled with LC-MS/MS analysis. Then, we analyzed the GO, KEGG and PPI of identified Kub proteins. Our study preliminarily revealed the global profiling of Kub proteins in metabolic pathways and provided an important resource for further study on the functions of Kub proteins in tea plants

Results and Discussion
Physiological characterization of tea plants subjected to drought and detection of Kub proteins in tea leaves. To compare the impact of drought stress on physiological and plant responses, tea plants were exposed to drought from 0 h to 96 h. During drought stress, the morphology of tea leaves became wrinkled and shriveled, especially at 96 h. The REC, LWC, Fv/Fm of tea leaves were investigated (Fig. 1), the LWC and Fv/Fm were declined, while REC was increased in drought treatment (DT). The results indicated that drought stress really caused cell dehydration in tea leaves, resulting in damage of membrane and photosynthetic system at physiological level. In order to examine the expression patterns of the Kub proteins in tea leaves under drought stress, we performed Western Blotting Assay (Supplemental Fig. 1). The bands of multiple Kub proteins in control treatment (CK) and DT were detected and all proteins in DT were reduced by long induction of drought stress, suggesting that the Kub proteins in tea leaves were changed dynamically and the lysine-ubiquitinated peptides could be enriched by di-Gly-Lys-specific antibody.
The ubiquitome profiles and functional classification of the Kub proteins in tea leaves under drought stress. In total, 1,409 Lys Kub sites were identified in 781 proteins, among which 1,226 sites were accurately quantified in 703 proteins. From these, 123 sites in 91 proteins were down-regulated and 14 sites in 12 proteins were up-regulated at a threshold of 1.5 (p < 0.05; Supplemental Table 1 Table 2). These Kub sites appeared in CK and disappeared after drought stress, suggesting that they were negatively regulated by drought stress. Meanwhile, 61 Kub sites in 52 proteins were identified in DT, including mannitol dehydrogenase (Lys-182), aconitate hydratase (Lys-864), fructose-1,6-bisphosphatase (FBPase; Lys-272) and fructose-bisphosphate aldolase (FBA; Lys-323), glutamate decarboxylase 1 (GAD1; Lys-5) and glutamine synthetase cytosolic isozyme 1 (GLN1-1; Lys-64 and Lys-65). These Kub proteins mainly related to carbohydrate metabolism and amino acid metabolism. The global proteome data were also collected with the identification of 4789 proteins (Supplemental Table 3).
Subcellular localization of the identified Kub proteins was analyzed ( Fig. 2A). Most of the Kub proteins were distributed in the cytosol (31.75%), chloroplast (27.02%), and nuclear (19.59%). The subcellular localization of whole proteome was also characterized for comparison ( Fig. 2A). According to these data, the subcellular localization of Kub proteins and global proteome had few significant differences.
We also compared the GO classification of Kub proteins and global proteome, and found that both Kub proteins and global proteome showed similar patterns in cellular components, molecular functions and biological processes ( Fig. 2B-D). In the cellular component, the Kub proteins mainly participated in cell and membrane. In the molecular function, the majority of Kub proteins were involved in binding and catalytic activity. In the biological process, the most of Kub proteins were associated with metabolic process and cellular process. These results indicated that lysine ubiquitome and proteome had similar localization distribution and biological functions.

Motif analysis and protein interaction networks for Kub proteins in tea plant.
In total, 622 of the 1409 Kub identified peptides were contained in amino acid sequences from −10 to +10 positions surrounding the Kub lysine. All of them were classified into 5 conserved motifs, including EK ub , EXXXK ub , K ub D, K ub E and K ub A (K ub indicates the Kub lysine, and X indicates any amino acid), and they exhibited different abundances (Supplemental Table 6; Fig. 5). Among them, EK ub , EXXXK ub , K ub D and K ub E were reported as Kub motif in other published studies 15 , the K ub A was firstly reported in our study. Moreover, the Kub lysine motifs showed a strong preference for glutamic acid (E) in the −4, −1 and +1 positions, as well as for aspartic acid (D) and alanine (A) in the +1 position. Similar preference for amino acid residues, such as glutamic acid, aspartic acid and alanine, adjoining Kub Lys residues has been observed in petunia, wheat and rice [14][15][16] . These results indicated that different plants might share common conserved motifs surrounding Kub sites.
To predict relationships among the Kub proteins in different metabolic pathways, we generated protein-protein interaction (PPI) networks for CK-unique, DT-unique and their common proteins against the STRING database. Among CK-unique proteins, there were 34 Kub proteins mapped to the protein interaction www.nature.com/scientificreports www.nature.com/scientificreports/ networks (Supplemental Table 7; Fig. 6A). And they were clustered into 8 sub-networks. The most abundant sub-network (Cluster 1) consisted of 14 Kub proteins, including ubiquitin-activating enzyme E1 1 (UBA1), ubiquitin carboxyl-terminal hydrolase 12 (UBP12), DNA-directed RNA polymerase II subunit 1 (NRPB5) and ribulose bisphosphate carboxylase/oxygenase activase (RCA). The interactions of these Kub proteins were mediated by UBP12 which involved in ubiquitin process, indicating that ubiquitin might play an important role in their interactions. The second sub-network (Cluster 2) consisted of 5 catechins biosynthesis-related Kub proteins, including naringenin,2-oxoglutarate 3-dioxygenase (F3H), TT4, chalcone-flavonone isomerase 3 (CHIL) and chalcone-flavonone isomerase 2 (TT5), of which all of the proteins interacted with each other.

The expressions of Kub proteins involved in ubiquitin-proteasome system under drought stress.
To demonstrate the expression of Kub proteins involved in proteolysis, we illustrated the process of UPS (Fig. 7). The UPS selectively removes substrate proteins by labeling ubiquitin protein tags and the activity of a series of enzymes. In the results, 14 Kub sites in 11 UPS-related proteins of CK and DT were significantly down-regulated under drought stress, such as Lys-29 and Lys-135 in RGLG2, Lys-118, Lys-191 and Lys-144 in UBC36, and Lys-74 in 26S proteasome non-ATPase regulatory subunit 4 (RPN10). Moreover, 6 Kub sites were www.nature.com/scientificreports www.nature.com/scientificreports/ found in 6 CK-unique proteins, including Lys-13 in 26S proteasome regulatory subunit 7 (PSMC2) and Lys-136 in ubiquitin-conjugating enzyme E2 variant 1D (UEV1D).
Previous research indicated that RGLG2, the RING domain ubiquitin E3 ligase, was negatively regulated by drought stress in Arabidopsis and a single mutant seedling, of rglg2 exhibited a dehydration-tolerant phenotype 17 . Moreover, RGLG2 catalyzed the synthesis of Lys-63-linked multiubiquitin chains 18 . Meanwhile, the Ubc13-Uev heterodimer consisted of Ubc13 and Uev was also required for the formation of Lys-63 linked multiubiquitin chains 19 . Since the multiubiquitin chains were involved in several cellular processes, including signal transduction, stress response and DNA repair 20,21 , RGLG2, Ubc13 and Uev might be involved in these cellular processes. The down-regulation of Kub sites in RGLG2, UBC36 (homolog of Ubc13) and UEV1D (Uev enzyme variant) suggested that the formation of Lys-63 linked chains might be inhibited under drought stress. So, it is tempting to speculate that RGLG2, UBC36 and UEV1D may play important roles in signal transduction and DNA repair.
26S proteasome consisted of multiple protein components catalyzes ATP-dependent breakdown of proteins conjugated with ubiquitin. The proteasome participated in several biological processes, including cell cycle progression, apoptosis, or DNA damage repair 22 . RPN10, the subunit of 26S proteasome, acted as an ubiquitin acceptor subunit through ubiquitin interactions and selected ubiquitin-proteins for destruction in human 23 . In Arabidopsis, RPN10 increased 20S proteasome levels which degraded proteins into small peptides, and thus enhanced Ub-independent protein degradation 24 . However, there is no report about the RPN10 in plants under drought stress. The down-regulation of Kub site in RPN10 in our study indicated that drought stress, to some extent, might enhance Ub-independent protein degradation. In addition, PSMC2, another subunit of 26S proteasome, could translocate Kub target proteins into a proteolytic chamber and degrade them into peptides in human and animals (such as mouse, bovine and rat) 25 . However, the functions of PSMC2 in plants were seldom reported. In the present study, the Kub site in PSMC2 of tea leaves was down-regulated under drought stress. We speculate that the down-regulation of Kub site in RPN10 and PSMC2 under drought stress may facilitate degradation of Kub protein, as well participate in cell cycle progression, apoptosis, or DNA damage repair in tea leaves.

The expressions of Kub proteins related to catechins biosynthesis under drought stress.
To elucidate the influence of Kub proteins related to catechins biosynthesis under drought stress, we analyzed the expressions of Kub proteins involved in phenylpropanoid and flavonoid pathway (Fig. 8) PAL is the first and committed step in the phenylpropanoid pathway. It is involved in the biosynthesis of the polyphenol compounds. The ubiquitination of PAL in petunia negatively correlated with the expression of PAL under ethylene treatment 15 . The content of catechins decreased under drought stress, which was consistent with the expression of PAL 26 . PAL was up-regulated in slight drought and down-regulated in serious drought 6 . The data in our study showed that the contents of ECG, EGCG in tea leaves increased under drought stress (Supplemental Table 8), but the Kub sites in PAL were all down-regulated. Therefore, we speculate that the UPS degraded PAL by ubiquitination and the down-regulation of PAL may negatively regulated biosynthesis of catechins under drought stress.
The enzyme CHS catalyzes the condensation of 4-hydroxycinnamoyl CoA and malonyl-CoA to form chalcone, which is the substrate for CHI and convert to naringenin. CHS and CHI were the critical genes in regulating catechins contents in tea plants in response to drought 27 . The contents of EGCG and total catechins had significant positive correlations with CHS and CHI during the development of tea leaves 28 . Moreover, CHS and CHI were detected being acetylated and differentially accumulated in leaves of ' Anjin Baicha' (an albino tea cultivar), suggesting that this PTM (post-translational modification) may contribute to the abundance of flavonoid across the developmental stages 29 . However, little research has been devoted to the ubiquitination of CHS and CHI in www.nature.com/scientificreports www.nature.com/scientificreports/ tea plants under drought stress. In present study, the content of naringenin was decreased and the Kub sites in CHS and CHI were down-regulated in tea leaves under drought stress, suggesting that the expression of Kub sites in CHS and CHI might positively regulate in the biosynthesis of naringenin. But the content of EGCG in tea leaves was increased under drought stress. Therefore, we speculate that the biosynthesis of EGCG may negatively associate with the expression of Kub sites in CHS and CHI in response to drought. Furthermore, an evidence showed that CHS and CHI were co-localized at the endoplasmic reticulum and tonoplast in Arabidopsis and the expressions of CHS and CHI were consistent with the higher accumulation of flavonoids 30 . In prior research, CHIL (type IV CHI protein) co-expressed, co-localized, and interacted with CHI for flavonoid production in Arabidopsis 31 . Previous study showed that CHS and CHI interacted with F3H and assembled as a macromolecular complex to promote flavonoid production in Arabidopsis, and CHI was posttranslationally modified, which played a role in controlling the association of CHI with other flavonoid enzymes 32 . However, to our knowledge, there is no report about the role of Kub CHS and CHI in plants. In our results, CHS, CHI, CHIL and F3H were all ubiquitinated in tea leaves and all the Kub sites in these proteins were down-regulated under drought stress. The PPI analysis showed that the interactive relationships existed among all of the four Kub proteins, suggesting that the down-regulation of 4 proteins accordingly slowed the reaction down from malonyl-CoA to dihydrokaempferol under drought stress. Therefore, we speculate that ubiquitination might play a key role in the interaction of CHI, CHS, CHIL and F3H. Their common down-regulations might decrease the flavonoid production.

The expressions of Kub proteins related to carbohydrate and amino acid metabolism under drought stress.
To elucidate the Kub proteins participated in carbohydrate and amino acid metabolism induced by drought stress, we analyzed the DT-unique Kub proteins in tea leaves. There were 6 Kub sites in 6 DT-unique proteins related to carbohydrate metabolism, including Lys-182 in GAD1, Lys-864 in aconitate hydratase, Lys-323 in FBA, and Lys-272 in FBPase. These Kub proteins were mainly involved in fructose and sucrose metabolism. Moreover, there were 3 Kub sites in 2 DT-unique proteins related to glutamate metabolism, including Lys-5 in GAD1, and Lys-64, Lys-65 in GLN1-1.
As for FBPase, three different groups of FBPase have been identified in eukaryotes and bacteria 33 . FBPase increased the soluble sugar (sucrose, glucose, and fructose) levels in the leaves of Arabidopsis, suggesting that the simultaneous overexpression of FBPase enhanced source capacity and consequently led to growth enhancement in transgenic plants 34 . High levels of FBPase contributed to the conversion of hexose into sucrose in tobacco, indicating that the increased FBPase activity led to enhance the synthetic ability and translocation efficiency of sucrose from source leaves to sink leaves 35 . Moreover, several metabolic enzymes involved in the Calvin cycle in wheat, including FBA, Rubisco, FBPase, and GAP, were found to be ubiquitinated 16 . However, no work has been done on the ubiquitination of FBPase in plants under drought stress. In the present research, the Kub site in FBPase induced by drought stress might promote sucrose accumulation.
FBA is an important enzyme in plants which is involved in glycolysis and the Calvin cycle, and plays a significant role in different stress responses. The overexpression of FBA increased the expressions of other main enzymes in Calvin cycle, net photosynthetic rate under salt stress, suggesting that FBA controlled photosynthesis, carbon partitioning and plant growth in tomato 36 . And, the abundance of FBA was increased under drought stress in Eragrostis tef, suggesting that drought stress may function in fructose-6-phosphate generation 37 . Furthermore, FBA involved in glycolysis and photosynthesis was detected to be ubiquitinated in rice, and was implied have a close relationship with salt tolerance 14 . In our study, the inducible expression of Kub site of FBA by drought stress indicated that the Kub site of FBA might regulate fructose biosynthesis in tea leaves under drought stress.
In addition, GAD1, the main enzyme of GABA biosynthesis, was involved in feedback controls of Ca 2+ -permeable channels to fluctuate intracellular GABA levels in tobacco, suggesting that GAD1 activity linked www.nature.com/scientificreports www.nature.com/scientificreports/ with Ca 2+ -permeable channels relayed an extracellular GABA signal 38 . And, the up-regulation of GAD gene expression induced the increased level of GABA under chlorsulfuron treatments in wheat, suggesting that GABA molecule might act as a protective and metabolic signaling molecule in carbohydrate and amino acid metabolism in plants under herbicidal treatments 39 . Moreover, the GABA level was increased in the gad1/2 mutant of Arabidopsis seedling under drought stress, and GABA accumulation during drought regulated the stomatal opening and prevented loss of water 40 . To our knowledge, little attention has been paid on the ubiquitination of GAD1 in plants under drought stress. Our study showed that the Kub site of GAD1 in tea leaves was induced by drought stress, suggesting that the Kub GAD1 might involve the GABA metabolism in tea leaves under drought stress.

Conclusion
In this study, we performed a global profile of Kub proteins in tea leaves under drought stress. Our results revealed that a large number of Kub proteins in tea leaves were participated in metabolic pathways, including ubiquitin-mediated proteolysis, catechins biosynthesis, and carbohydrate and amino acid metabolism. Several Kub sites in ubiquitin-mediated proteolysis-related proteins were down-regulated under drought stress, including RGLG2, UBC36, UEV1D, RPN10 and PSMC2. These Kub proteins might affect protein degradation and DNA repair in tea leaves. A large number of Kub proteins were related to catechins biosynthesis, including PAL, CHS, CHI and F3H, suggesting that these proteins were positively correlated with each other due to their co-expression and co-localization. Furthermore, some Kub proteins involved in carbohydrate and amino acid metabolism, including FBPase, FBA and GAD1, were induced significantly by drought stress, suggesting that these Kub proteins might promote sucrose, fructose and GABA accumulation in tea leaves. Based on this study, we can conclude that these proteins are indeed modified by ubiquitin, but as to the molecular outcome for each of these events it remains to be determined. Our study preliminarily revealed the roles of Kub proteins in metabolic pathways and provided an important resource for the further study of Kub functions in tea plants in response to drought stress. Physiological experiments and protein extraction. Tea plants (three biological replicates) were collected for DT and CK at 0, 24, 48, 72 and 96 hours. The relative electrolytic conductivity (REC), leaf water content (LWC) and leaf maximum photochemical quantum yield of PS II (Fv/Fm) of different treatment were determined as described previously 14 .

Materials and Methods
Tea leaves (1 g/fw) were grinded into cell powders by liquid nitrogen and then transferred to a 5-mL centrifuge tube. After that, four volumes of lysis buffer (8 M urea, 1% Triton-100, 10 mM dithiothreitol and 1% Protease Inhibitor Cocktail, and 3 μM TSA and 50 mM NAM inhibitors) were added to the cell powders, followed by sonication on ice using a high intensity ultrasonic processor (Scientz). The remaining debris was removed by centrifugation at 20,000, 4 °C for 10 min. Finally, the protein was precipitated with cold Trichloroacetic acid (TCA) (supernatant/TCA, 17:3, v/v) at −20 °C for. After centrifugation at 12,000 g, 4 °C for 10 min, the supernatant was discarded. The remaining precipitate was washed with cold acetone for three times. The protein was redissolved in 8 M urea and the protein concentration was determined with BCA kit according to the manufacturer's instructions.
Trypsin digestion. For digestion, 4 mg of isolated protein solution was reduced with 5 mM dithiothreitol for 30 min at 56 °C and alkylated with 11 mM iodoacetamide for 15 min at room temperature in darkness. The protein sample was then diluted by adding 100 mM NH 4 HCO 3 to urea concentration less than 2 M. Finally, trypsin was added at 1:50 trypsin-to-protein mass ratios for the first digestion overnight and 1:100 trypsin-to-protein mass ratios for a second 4 h-digestion.
Tandem mass tag labeling. After trypsin digestion, peptide was desalted with a Strata X C18 SPE column (Phenomenex) and vacuum dried. Peptide was reconstituted in 0.5 M TEAB and processed according to the manufacturer's protocol for the six-plex Tandem Mass Tag (TMT) kit. Briefly, 1 unit of TMT reagent (defined as the amount of reagent required to label 100 mg of protein) was thawed and reconstituted in 24 mL of acetonitrile. The peptide mixtures were then incubated for 2 h at room temperature and pooled, desalted, and dried by vacuum centrifugation.
HPLC fractionation. The tryptic peptides were fractionated into fractions by high pH reverse-phase HPLC using Agilent 300 Extend C18 column (5 μm particles, 4.6 mm ID, 250 mm length). Briefly, peptides were first separated with a gradient of 8% to 32% acetonitrile (pH 9.0) over 60 min into 60 fractions. Then, the peptides were combined into 18 fractions and dried by vacuum centrifuging. www.nature.com/scientificreports www.nature.com/scientificreports/ LC-MS/MS analysis. The tryptic peptides were dissolved in 0.1% formic acids (solvent A), directly loaded onto a home-made reversed-phase analytical column (15-cm length, 75 μm i.d.). The gradient was comprised of an increase from 6% to 23% solvent B (0.1% formic acids in 98% acetonitrile) over 26 min, 23% to 35% in 8 min and climbing to 80% in 3 min then holding at 80% for the last 3 min, all at a constant flow rate of 400 nL/min on an EASY-nLC 1000 UPLC system.
The peptides were subjected to NSI source followed by tandem mass spectrometry (MS/MS) in orbitrap fusion ™ tribrid (Thermo) coupled online to the UPLC. Intact peptides were detected in the orbitrap at a resolution of 60,000. Peptides were selected for MS/MS using NCE setting as 35 and ion fragments and detected in the Ion Trap. A top speed data-dependent procedure that alternated between one MS scan followed by most intense MS/MS scan was applied for the precursor ions above threshold intensity greater than 5E3 in the MS survey scan with 30.0 s dynamic exclusion. The electrospray voltage applied was 2.0 kV. Automatic gain control (AGC) was used to prevent overfilling of the ion trap; 1E4 ions were accumulated for generation of MS/MS spectra. For MS scans, the m/z scan range was 350 to 1550. Fixed first mass was set as 100 m/z. The mass spectrometry proteomics data are available via ProteomeXchange with identifier PXD011688. The proteomics for TMT experiment corresponds to 1-18 raw files, representing 18 fractions. And the raw file of ubiquitination modification was named by the respective sample. Bioinformatics analysis. Bioinformatics analysis was performed according to previously described protocols 14 . The Gene Ontology (GO) annotation proteome was derived from the UniProt-GOA database (http:// www.ebi.ac.uk/GOA/) 41 . The lysine ubiquitination peptide ID was converted to a UniProt ID and then mapped to a GO ID. The Kub proteins were then further classified by GO annotation based on three categories: biological processes, molecular functions, and cellular components. A two-tailed Fisher's exact test was employed to test the enrichment of the differentially expressed protein against all identified proteins. The Kyoto Encyclopedia of Genes and Genomes (KEGG) database was used to annotate protein pathways 42 . The KEGG online service tool KAAS was used to annotate the proteins' KEGG database description. The annotation results were mapped on the KEGG pathway database using the KEGG online service tool KEGG Mapper. The domain annotation was performed with InterProScan on the InterPro domain database via web-based interfaces and services. WoLF PSORT was used for predicting the subcellular localization 43 . The CORUM database was used to annotate protein complexes.
Motif-X software was used to analyze the model of the sequences with amino acids in specific positions of ubiquityl-15-mers (seven amino acids upstream and downstream of the Kub site) in all of the protein sequences. In addition, all the database protein sequences were used as the background database, and the other parameters were set to the default values. The setting parameters for searching motifs using Motif-X software was "occurrences 20" and "the Bonferroni corrected p = 0.005".
All CK-unique, DT-unique and their common DEPs were selected for protein-protein interactions. First, the sequences of these DEPs were fetched and then upload to the STRING (http://string-db.org/). Next, the model species Arabidopsis thaliana was selected as a database. Through BLAST, the DEPs were matched to the Arabidopsis protein. STRING defines a metric called "confidence score" to define interaction confidence. The interactions that confidence score ≥0.4 were selected. Finally, interaction networks from STRING were visualized with Cytoscape 44 .