A Survey on Cell wall Proteins of C. Sinensis Leaf by Combining Cell Wall Proteomic and N-Glycoproteomic Strategy

Background: Camellia sinensis is an important economic crop with uoride over-accumulation in the leaves, which pose a serious threaten to human health due to its leave being used for making tea. Recently, our study found that cell wall proteins (CWPs) probably play a vital role in uoride accumulation/detoxication in C. sinensis. However, CWPs identication and characterization were lacking up to now in C. sinensis. Herein, we aimed at characterizing cell wall proteome of C. sinensis leaves, to develop more CWPs related to stress response. A strategy of combined cell wall proteome and N-glycoproteome were employed to investigate CWPs. CWPs were extracted by sequential salt buffers, while N-glycoproteins were enriched by hydrophilic interaction chromatography method using C. sinensis leaves as a material, afterwards all proteins were subjected to qualitative analysis via UPLC-MS/MS. Results: 501 and 195 CWPs were identied by cell wall proteomic and N-glycoproteomics proling, respectively, with 118 CWPs being in common. Notably, N-glycoproteome is a feasible method for CWPs identication and consequently enhance CWP coverage. Among identied CWPs, proteins acting on cell wall polysaccharides constitute the largest functional group with most of them possibly being involved in the remodeling of cell wall structure. The second abundant group encompass mainly various proteases, being considered to be related to CWPs turnover and maturation. Oxidoreductases represent the third abundance with most of them especially Class III peroxidases being known to be implicated in defense response. As expected, identied CWPs emphasized on plant cell wall formation and defense response. Conclusion: This was the rst large scale survey of CWPs by cell wall proteome and N-glycoproteome in C. sinensis. The results not only provides a database that will aid deep research on CWPs, but also improve the understanding underlying cell wall formation and defense response in this important economic specie.


Discussion
Identi cation and functional classi cation of identi ed CWPs Totally, 3618 ECWPs were identi ed in C. sinensis leaves by sequential salt extractions and UPLC-MS/MS. Among them, 501 ECWPs and 3079 ECWPs were considered to be CWPs and intracellular proteins, respectively, via multiple bioinformatics analysis. Notably, intracellular proteins represent 85.1% of ECWPs, indicating ECWPs were subjected to the contamination during ECWPs preparation. Similarity, high contamination of intracellular proteins was also detected d in sugarcane [25] and rice [29], accounting for 81.6% and 80.5%, respectively. However, the study of cell wall proteome was still rare in plant species and thus CWPs extraction need be improved. However, our study still led to an enlargement of CWPs and offer new knowledge in C. sinensis in spite of the high contamination of intracellular proteins.
At the same time, 262 N-glycoproteins were identi ed in the leaves of C. sinensis. As expected, most Nglycoproteins (195, 74.4%) were targeted into the cell wall/extracellular/plasma membrane and thus were assigned as CWPs. The result was in good accordance with that in tomato fruit [35] and Brachypodium distachyon leaf [45] which 65% and 60% of N-glycoproteins were reported to be located in the apoplast/cell wall/plasma membrane, respectively, demonstrating that N-glycoproteome is a feasible method to identify and characterize CWPs.
Taken together, 501 CWPs and 195 CWPs were identi ed by cell wall proteomic and N-glycoproteomic analysis, respectively, and 118 of which was in common. Excitingly, 25 new CWPs being absent in WallProtDB were assigned in this study (Additional le 3: Table S3). The result suggested cell wall proteome is more effective method than N-glycoproteome for CWPs identi cation. However, it should be noted that the use of N-glycoproteome enhanced CWP identi cation. As a result, the combined strategy of cell wall proteome and N-glycoproteome should be considered during CWP identi cation and characterization.
To obtain a global view of the biological processes in which the identi ed CWPs were involved, 578CWPs were divided into nine functional groups according to their functional domains in this study. Unsurprisingly, PACs (147, 25.4%), Ps (94, 16.3%) and ORs (62, 10.7%) represent top three functional groups. The function distribution of CWPs was in good concordance with that of A. thaliana rosettes and B. distachyon leaves (Additional le 5: Fig. S3). Notably, the proportion of PSs (9.7%) in C. sinensis was obviously higher than that of A. thaliana rosettes and B. distachyon leaves with 3.7% and 4.0%, respectively [12,20], which maybe account for the long lifecycle of the evergreen leaf in C. sinensis.
Possible substrates of most of GHs families were hemicelluloses (xyloglucan, xylans, glucomannans) and pectin (galactans, homogalacturonan). Out of GHs identi ed in this study, GH16, GH29, GH31 and GH65 potentially act on xyloglucans, GH10 and GH51 show possibly action on xylans, and GH28 and GH35 could hydrolyze homogalacturonan and galactans, respectively [46-48] (Additional le 6: Table  S5). Moreover, GH1, GH3 and GH5 possess broad substrates range, their enzymes are reported to be involved in the modi cation and/or breakdown of cell wall hemicelluloses and pectins [49,50], and also be implicated in ligni cation and secondary metabolism [51]. Identi cation of these GHs families suggested that hemicelluloses and pectins might undergo important structural changes in the leaves of C. sinensis. Furthermore, GH127 (DUF1680 domain protein), being characterized recently as a novel beta -L-arabinofuranosidase, might be implicated in the degradation of cell wall polysaccharides and hydroxyproline-rich glycoproteins [52], and GH9 was known to catalyze the endohydrolysis of cellulose.
Some identi ed GHs could participate in defense against pathogens and various stress. Chitin and beta − 1,3-or beta − 1,6-glucan are main components of cell walls of various fungi. GH17 acts as beta − 1,3glucanase, together with chitinases (GH18 and GH19) and GH20 that function as key hydrolyzed enzyme of chitin, have shown to possess antifungal activity by degrading their cell walls and then participate in defense against pathogens [47,53]. Intriguing, chitinases being in response to abiotic stress were also reported [54,43]. GH37, a non-reducing sugar, was identi ed to be a new CWP in this work without being documented as CWPs in WallprotKB. GH37 acts as a universal stabiliser of protein conformation, might contribute to various stress defense [55].
Several identi ed GHs including GH13, GH27 and GH32 might be implicated in mobilization, allocation and partitioning of storage reserves. GH13 is was associated with the hydrolysis of starch and glycogen to yield glucose and maltose [56], GH27 is one of three hydrolyzed enzymes of galactomannans as a cell wall storage polysaccharide [57], and GH32 as invertases is involved in long distance nutrient allocation and carbohydrate partitioning [58,59]. Additionally, a couple of GHs enzymes including GH3, GH18, GH19, GH35, GH38 and GH79 were known to be involved in post-translational modi cations (PTMs) of glycoproteins [32,47]. Here, GH3, GH35, GH38 and GH79 were veri ed as N-glycoproteins.
Collectively, identi ed GHs potentially give rise to complex cell wall carbohydrates remodeling, pathogen and stress response, mobilization and allocation of storage reserves as well as glycoproteins PTMs. The high number of GHs associated with cell wall metabolism and defense response were found in this work, which is consistent with published reports of sugarcane stems and leaves [26], B. distachyon grains [21], Saccharum o cinarum cell suspension [25]. The results might be attributed to sustainability remodeling during plant growth and development and terrestrial habit of plants. Trichome birefringence-like proteins and PNGase A are also two modi cation enzyme families of cell wall. The former was characterized as xylan acetyltransferases, and was believed to be implicated in the mediation of xylan O-acetylation, which being required for secondary wall deposition and pathogen resistance [64]. The latter is one of deglycosylation enzyme and has been considered to be involved in the release of N-glycans from glycopeptides generated by the proteolysis of denatured glycoproteins [65]. Laccases, like Class III PODs, are candidates for polymerizing monolignol unit into lignin, suggesting be required for cell wall ligni cation [73,74].
BBE-like proteins, act as monolignol oxidoreductases, may participate in the mobilization and oxidation of monolignols required for polymerization processes [75]. All in all, three high represented enzyme families in the class were considered to be involved in ligin production and subsequence the reinforcement of cell walls strength and rigidity, which favoring plant defense against adverse environmental factors.
Other CWPs related to redox processes including monocopper oxidase-like proteins (SKU5 and SKS1), blue copper proteins and ascorbate oxidases were identi ed, which probably play a role in both cell wall loosening, expansion and reticulation processes [24,76].
Five structure proteins were identi ed in present study including three leucine-rich repeat extensin-like protein (LRR-EXTs), non-classical arabinogalactan protein 31-like (AGP) and hydroxyproline-rich glycoprotein. LRR-EXTs have known to in uence mechanical properties of cell wall by their ability to form insolubilized, covalently crosslink to cell wall components [95], as well as function as perceive extracellular signals and indirectly relay into the cytoplasm to regulate plant growth and salt tolerance, thereby suggesting they are important for cell wall development, plant growth and stress tolerance [96].
Non-classical AGPs have both a proline-rich domain and a non-proline-rich domain, may be function in metal ion-binding, defense response and interact with pectin [97,98]. As for hydroxyproline-rich glycoprotein, which is an important structural components of plant cell walls and are thought to be implicated to structural integrity, cell-cell interaction and intercellular communication [99].
Several enzymes of CWPs inhibitor were also detected in this study. PMEIs that inhibited partly the activity of PMEs, adjust the degree of pectin methyl-esteri cation. PGIPs (polygalacturonase inhibitorlike) speci cally bind with polygalacturonases (GH28), thereby they can inhibit the hydrolyzation of pectin and then regulate pectin degradation, which can trigger defense against microbes and insects [100]. In summary, two couple of PMEIs and PME, PGIPs and PG occurred coincidentally and modulate precisely pectin metabolism. As for Cys proteinase inhibitor possess inhibitory activities against speci c Cys proteases, probably play a role in insect predation [101].

Identi ed CWPs emphasizing on plant cell wall formation and defense response
Under dynamically changing environmental conditions, plant grow and develop continuously, and always encounter variable stresses and deleterious attack of insects and microbes. To acclimate, plant cell walls that acting as the rst barrier change constantly, whereas CWPs play central roles in altering cell wall properties.
Doubtlessly, to meet normal growth and development, a large amount of CWPs could be triggered to adjust vigorously cell wall structure. Here, identi ed numerous CWPs related to PACs, mainly including GH1, GH3, GH5, GH9, GH10, GH16, GH28, GH29, GH31, GH35, GH51 and GH65, might contribute to the rearrangement of cell wall structure. In contrast, expansins probably lead to cell wall extension. Certainly, several CWPs associated with the formation and metabolism of secondary cell wall, like Class III PODs, BBEs, laccases, LTPs, GDSLs and DIRs, maybe favor to the reinforcement/modi cation of cell wall (Fig. 4).
Facing to adverse environment, C. sinensis, a terrestrial plant, have no ability to escape. Therefore, they have evolved in the context of altering cell wall properties for improved defense responses. Today, ample identi ed CWPs were potentially involved in various defense. GH17, GH18, GH19 and GH20 were reported to be involved mainly in against pathogens as well as abiotic stress by hydrolyzing chitin. Class III PODs, monocopper oxidase-like proteins, blue copper proteins and ascorbate oxidases were known to be implicated in respond to various biotic and abiotic stresses by redox reaction. LTPs, GDSLs and DIRs were also associated with defense response through the regulation of secondary cell wall. PGIPs and Cys proteinase inhibitor might function in improving protection against insects and pathogens [102] via inhibiting the activity of degradation enzymes of invaders. Likewise, BCPs, GLPs, cupins and Thaumatins also serve functions in defense response (Fig. 4).
To sense changed environment and the status of complex cell wall structures, plants have developed cell wall integrity-sensing pathway to transduce signals into cytoplasm. A number of sensors at the plasma membrane including RLKs and FLAs were identi ed in present study, which enable C. sinensis to coordinate the processes of the cell wall and the cytoplasm (Fig. 4).
In summary, a work model of identi ed CWPs were proposed (Fig. 4), which emphasizing on plant cell wall formation and defense response, and further making a bit explanation for plant internal activities during normal growth under natural environment.

Conclusions
In this experiment, a study of combined cell wall proteome and N-glycoproteome was performed to depict

Plant materials
The rst to fth leaves of 20 uniform 2-year-old cutting seedlings of the Echa 1 variety (Camellia sinensis cv. 'Echa 1') were collected from tea germplasm bank located in Wuhan city of Hubei province (China), then washed three times with Milli-Q water, grinded into ne power in liquid nitrogen immediately, and nally stored at -80 °C for further use.

Cell wall enrichment
Cell wall fraction was obtained from the leaves of Camellia sinensis using sequential washes as described by Printz et al [103] with slight modi cation. Brie y, 5g ne power of the leaves were homogenized with 3-fold volumes of 0.4 M sucrose buffer for 10 min, vortexed for 2 min, shaken overnight at 250 rpm at 4 °C, and then centrifuged. Subsequently, 0.6 M sucrose buffer was added into the precipitations, shaken for 30 min at 250 rpm at 4 °C and centrifuged. After that, 1 M sucrose buffer was added into the precipitations again, suspended and centrifuged. Finally, the precipitations were washed twice using 5 mM sodium acetate buffer. The nal precipitations were cell wall fraction (pellet).
Sucrose buffers contained 5 mM sodium acetate and 1% protease inhibitor cocktail (ApexBio), all buffers (pH 4.6) were precooled at 4°C and the centrifugation was operated at 1000 rpm for 15 min at 4 °C.

Cell wall protein extraction
CWPs were extracted successively using CaCl 2 , EGTA and LiCl according to the method reported by Printz et al [103]. Brie y, 0.2 M CaCl 2 buffer was rstly added into cell wall pellet, shaken for 30 min at 200 rpm at 4°C followed by centrifugation, then the supernatants were collected. This step was repeated once and the supernatants were pooled as CaCl 2 fractions. Afterwards, cell wall pellet was again mixed with 50 mM EGTA buffer followed by shaking for 1 h at 300 rpm at 37 °C, centrifugation and supernatants collection. This step was repeated twice and all supernatants were pooled as EGTA fractions. Cell wall pellet was nally resuspended in 3 M LiCl buffer, homogenized overnight at 250 rpm, 4 °C. After centrifugation, the supernatants were collected. The proteins were once again extracted from the pellet with 3 M LiCl buffer by shaking for 6 h at 250 rpm, 4 °C. The obtained supernatants were pooled and stored as LiCl fractions. Finally, CaCl 2 , EGTA and LiCl fractions were combined as extracted CWPs (ECWPs) fractions. All extraction buffers were precooled at 4°C and the centrifugation was performed for 15 min at 10000 rpm at 4 °C.

Whole protein extraction
Whole protein was extracted from C. sinensis leaves according to several published paper [43,104,105]. Brie y, about 0.5 g ne powder were rstly homogenized with 5 ml pre-cooled homogenization buffer [20 mM Tris-HCl (pH7.5), 250 mM sucrose, 10 mM EGTA, 1 mM PMSF, 1 mM DTT, 1 % (v/v) Triton], and then centrifuged at 12000 g for 20 min at 4 °C. The obtained supernatants were pooled and stored as whole protein fractions.

Protein precipitation and cleaning
According to our previous study [43], whole protein fractions and ECWPs fractions were precipitated severally by Tris-phenol (pH ≥ 8.0) and ammonium acetate. In brief, the fractions were mixed with equal volume of Tris-phenol, vortexed followed by centrifugation at 12000 g for 20 min at 4 °C. Afterwards, the phenol phases were transferred carefully into other tubes, mixed thoroughly with 5 volumes of 0.

HPLC fractionation
After tryptic digestion, the peptides from whole proteins and ECWPs were fractionated severally by the use of high pH reversed-phase HPLC (high-performance liquid chromatography) with Agilent 300 Extend C18 column (5 μm particles, 4.6 mm inner diameter, 250 mm length). Brie y, the digested peptides were rst separated into 60 fractions with a gradient of 8% to 32% acetonitrile (pH 9.0) over 60 min. Subsequently, the peptides were pooled into 4 fractions and dried by vacuum centrifugation for further use.

A nity enrichment of N-glycopeptides
To enrich N-glycosylation peptides, the dried peptides from whole proteins were rstly dissolved in 40 μL enrichment buffer (80% acetonitrile, 1% tri uoroacetic acid) and then loaded into HILIC microclumn to divide into glycopeptides and non-glycopeptides by centrifugation for 15 min at 4000 g. To remove nonspeci cally adsorbed peptides, HILIC microclumn was washed three times with enrichment buffer.
Subsequently, the bound peptides were eluted from the microclumn with 10% acetonitrile and then vacuum-dried. The lyophilized N-glycopeptides were reconstituted in 50 μL 50 mM NH 4 CO 3 buffer in heavy oxygen water and incubated with 2 μL PNGase F at 37 °C overnight. Finally, the resulting Nglycopeptides were desalted with C18 ZipTips (Millipore) according to the manufacturer's instructions and lyophilized for LC-MS/MS analysis.

UPLC-MS/MS analysis
For LC-MS/MS analysis, the peptides were rstly dissolved in subjected to solvent A (0.1% (v/v) formic acid and 2% acetonitrile) and then subjected to a gradient elution by EASY-nLC 1000 UPLC system.
Subsequently, the separated ECWPs peptides and deglycosylated peptides were injected into a nanoelectrospray ion source followed by MS/MS analysis in Q ExactiveTM and Orbitrap Fusion mass spectrometer (Thermo Fisher scienti c), respectively. Brie y, the applied electrospray voltage was 2.0 kV, the intact peptides and their secondary fragments were detected and analyzed by Orbitrap and a datadependent acquisition mode that automatically altered between MS scan and MS/MS scan was adopted.
For ECWPs peptides, which were detected at a resolution of 70,000 with m/s scan range of 350-1800 for full scan. After that, the 10 most intense parent ions per scan were selected for higher-energy collisional dissociation fragmentation (HCD) at 28% collision energy. The generated fragments were further analyzed at a resolution of 17,500 with a xed rst mass of 100 m/z. To improve the effective utilization rate of mass spectrometry, automatic gain control of 5E4, 30 s dynamic exclusion, 100 ms maximum inject and signal threshold of 20000 ions/s were applied. Likely, deglycosylated peptides were detected at a resolution of 60,000 with m/s scan range of 350-1550 for full scan. The 20 most intense parent ions per scan were selected for HCD at 35% collision energy, and then the resulting fragments were analyzed at a resolution of 15,000 with a xed rst mass of 100 m/z. Likewise, automatic gain control of 5E4, 15 s dynamic exclusion, 200 ms maximum inject and signal threshold of 5000 ions/s were used.

Database Search
The resulting raw MS/MS data was processed using MaxQuant search engine (v.1.5.2.8) with the following query parameters: ( ) tea tree genome database (Camellia_sinensis_4442 with 53512 sequences; [106]) concatenated with reverse decoy database and mass spectrometry contaminants database for MS/MS search; ( ) Trypsin/P for enzyme cleavage and 2 for missing cleavages; ( ) mass tolerance of 20 ppm and 5 ppm for peptide ions in rst search and main research, respectively, and 0.02 Da for fragment ions; ( ) 7 amino acid residues for minimum peptide length and 5 for maximum modi cation number in a peptide; ( ) Cysteine alkylation as xed modi cation; ( ) Variable modi cation: methionine oxidation and N-terminal acetylation of protein for ECWPs, and methionine oxidation and deamidation (NQ), asparagine deamidation ( 18 O for N-glycoproteins; ( ) FDR≤1% for protein identi cation and peptide-spectrum matches identi cation.

Figure 1
Experiment work ow using in this work. The extraction, precipitation, digestion, fractionation, and MS/MS and data analyses of ECWPs were operated according to blue arrow instruction. Likely, those of glycoproteins were operated according to red arrow instruction.    This is a list of supplementary les associated with this preprint. Click to download.