Comparative proteomic analysis of thick-walled ray formation of Haloxylon ammodendron in the Gurbantunggut Desert, China

The thick-walled ray cells have been reported in Haloxylon ammodendron for the rst time. This study measured the wall thickness of ray cells and performed a proteomic analysis of ray cell wall formation in the xylem of H. ammodendron using isobaric tags for relative and absolute quantitation. wall thickness of ray cells in ± 0.42 was signicantly lower than that in Shihezi were


Background
Haloxylon ammodendron is an important afforestation species in the arid desert region both in Asia and Africa. Previous studies have found that the cell wall of ray cells was obviously thickened in the xylem of H. ammodendron, and the wall thickness of the ray cell can be up to 2.85 µm ~ 3.08 µm, which is 3 ~ 6 times the thickness of axial parenchyma and is slightly higher than that of bre (2.64 µm ~ 2.97 µm) (Zhou and Gong 2017). Generally, ray tissue is composed of parenchyma cells, and the wall thickness of ray cells is much thinner than that of bre cells in most species (Plavcová and Jansen 2015). However, in parenchyma, thick-walled ray cells have been reported in some species, such as Melia azedarach and Symbolanthus macranthus (Carlquist and Grant 2005). The mechanical properties will be enhanced when tissues have thickened walls (Alves and Angyalossy-Alfonso 2002). Therefore, research on the regulatory proteins (genes) involved in the cell wall formation process will be important for both the environmental adaptability of the characteristic of xylem and improvement of wood quality in plantations.
The cell wall in higher plants consists mainly of cellulose, hemicellulose and lignin (Mellerowicz et al. 2001). The cell wall is mainly composed of polysaccharides. Both the primary wall and secondary wall contain cellulose and hemicellulose. The primary wall also contains pectin, enzymes and structural proteins; the secondary wall contains a small amount of proteins or pectin but generally contains lignin (Carpita and McCann 2000). The cell wall not only provides mechanical support and defence against pathogen invasion and nutrition stress but is also related to the physiological function of plant cells, such as material transport (Dhugga 2005). Moreover, photosynthates stored in plant secondary cell walls are important sources of bre materials and raw materials of biomass energy (Huang and Li 2016), which play an important role in human survival and development.
Recently, with the development of genomics and molecular genetics, research on plant cell wall formation has made good progress. Research reports have mainly focused on Arabidopsis thaliana (Taylor et al. 2003), Populus trichocarpa (Suzuki et al. 2006), Picea sitchensis (Bong) Carr (Fernandes et al. 2011), and Gossypium hirsutum (Pear et al. 1996). Many achievements have been made in the regulatory mechanisms of cellulose, hemicellulose, lignin and pectin biosynthesis. Studies have shown that glucan chains in cellulose are synthesized by the cellulose synthase complex (CSC) on the plasma membrane and secreted into the extracellular space. The CSC is composed of three different types of cellulose synthase (CESA) catalytic subunits (Gonneau et al. 2014). The main chains of hemicellulose are synthesized by cellulose synthase-like (CSL), except for the xylan chain, and Trigonella foenum-graecum galactosyltransferase (TfGalT) was shown to be responsible for the synthesis of galactosyl side chains (Scheller and Ulvskov 2010). The main synthesis sites of hemicellulose and pectin are the Golgi apparatus, and the nucleotide sugar, which is the substrate of polysaccharose, is catalysed by CSL protein and glycosyltransferase and is synthesized mainly in the cytoplasm (Bar-Peled and O'neill 2011).
The precursors needed for lignin biosynthesis are synthesized by the phenylpropanoid pathway, and deamination of phenylalanine to cinnamic acid initiates this process. All of these processes have been extensively studied (Mellerowicz et al. 2001). Although much work has been done in cell wall biosynthesis, the content and structure of cell wall components vary with species and tissues, which leads to diversity and complexity in cell wall composition (Burton et al. 2010). Therefore, the biosynthetic pathway of the cell wall in speci c tissues of different species still needs to be studied.
Thus, the proteomic characteristics of ray cell wall formation in H. ammodendron were studied. GO annotation combined with KEGG pathway enrichment and other bioinformatics methods was used to explore the differentially expressed proteins and metabolic pathways related to the ray cell wall formation of this plant.

Sample location and sampling
The climate and growth characteristics of H. ammodendron plantation in the sample area were described by Zhou and Gong (2017). At the end of June 2017, samples were taken in Jinghe (82°53′35″E,44°36′10″N) and Shihezi (86°14′44″E 45°00′34″N) Desert Research and Experimental Station, Shihezi University, in the Gurbantunggut Desert, Xinjiang, China. The identi cation of H. ammodendron was performed according to the morphological characteristics in website (http://www.iplant.cn/info/Haloxylon%20ammodendron) built by Institute of Botany, Chinese Academy of Sciences. Perennial branches of H. ammodendron with a diameter of approximately 1 cm were collected. The bark, phloem and cambium were scraped from the branches, and a blade sterilized with anhydrous ethanol was used to scrape the xylem. The samples were wrapped in aluminium foil, placed in liquid nitrogen and quickly cooled, brought back to the laboratory and stored in a -80 ℃ refrigerator.

Measurement of indexes
Observation and measurement of the wall thickness of ray cells Observations under a light microscope and measurements of the wall thickness of ray cells in H. ammodendron were performed as described in Zhou and Gong (2017).
Scanning electron microscopy (SEM): The xylem of H. ammodendron was cut into blocks of 1 cm x 1 cm x 1 cm, air dried and polished to a smooth surface. The samples were dehydrated using graded ethanol and dried using the liquid CO 2 critical point method. The sample was soaked in 98% H 2 SO 4 for 5 min. The sample was attached to conductive tape, metal spraying was performed for 200 s, and then, the sample was observed by SEM.

Proteomic analysis
Protein Extraction: Samples (1 ~ 2 g) with 10% PVPP were ground into powder in liquid nitrogen and then sonicated on ice for 5 min in lysis buffer 3 (8 M urea and 40 mM Tris-HCl containing 1 mM PMSF, 2 mM EDTA and 10 mM DTT, pH 8.5) with 5-fold volumes of samples. After centrifugation at 25,000 rpm at 4 ℃ for 20 min, the supernatant was treated by adding 5-fold volumes of 10% TCA/acetone with 10 mM DTT to precipitate the proteins at -20 ℃ for 2 h/overnight. The precipitation step was repeated with acetone alone until there was no colour in the supernatant. The proteins were air dried and resuspended in lysis buffer 3 (8 M urea and 40 mM Tris-HCl containing 10 mM DTT, 1 mM PMSF and 2 mM EDTA, pH 8.5). Ultrasonication on ice for 5 min (2 sec/3 sec) was used to improve the protein dissolution. After centrifugation, the supernatant was incubated at 56 ℃ for 1 h for reduction and alkylated by 55 mM iodoacetamide (IAM) in the dark for 45 min. Fivefold volumes of acetone to samples were used to precipitate the proteins at -20 ℃ for 2 h/overnight. Lysis buffer 3 was used to dissolve the proteins with sonication on ice for 5 min (2 sec/3 sec).
QC of Protein Extraction: We separately added 0, 2, 4, 6, 8, 10, 12, 14, 16 and 18 µL of BSA (0.2 µg/µL) solution to a 96-well plate and then added 20,18,16,14,12,10,8,6,4 and 2 µL of pure water to the corresponding wells. We also prepared serial dilutions (20 µL each well) of the unknown sample to be measured. Then, 180 µL of Coomassie Blue was added to each well and mixed. The absorbance of each standard and sample well was read at 595 nm. Each sample had at least two duplicates. The absorbance of the standards vs. their concentrations was plotted. The extinction coe cient and the concentrations of the unknown samples were calculated.
We mixed 15 ~ 30 µg proteins with loading buffer in a centrifuge tube and heated them at 95 ℃ for 5 min. Then, the samples were centrifuged at 25000 rpm for 5 min, and the supernatant was added to the sample wells in a 12% polyacrylamide gel. SDS-PAGE was performed at constant voltage at 120 V for 120 min. Then, the gel was stained with Coomassie Blue for 2 h, destaining solution (40% ethanol and 10% acetic acid) was added and the gel was placed on a shaker (exchange destaining solution 3 ~ 5 times, 30 min a time).
Protein Digestion: The protein solution (100 µg) with 8 M urea was diluted 4 times with 100 mM TEAB. Trypsin Gold (Promega, Madison, WI, USA) was used to digest the proteins at a ratio of protein:trypsin = 40:1 at 37 °C overnight. After trypsin digestion, the peptides were desalted with a Strata X C18 column (Phenomenex) and vacuum-dried according to the manufacturer's protocol.
Peptide Labelling: The peptides were dissolved in 30 µL of 0.5 M TEAB with vortexing. After the iTRAQ labelling reagents were at ambient temperature, they were transferred and mixed with the proper samples.
Peptide labelling was performed using the iTRAQ Reagent 8-plex Kit according to the manufacturer's protocol. The labelled peptides with different reagents were combined and desalted with a Strata X C18 column (Phenomenex) and vacuum-dried according to the manufacturer's protocol.
Peptide Fractionation: The peptides were separated on a Shimadzu LC-20AB HPLC Pump system coupled with a high pH RP column. The peptides were reconstituted with buffer A (5% ACN, pH 9.8) to 2 mL and loaded onto a column containing 5 µm particles (Phenomenex). The peptides were separated at a ow rate of 1 mL/min with a gradient of 5% buffer B (95% ACN, pH 9.8) for 10 min, 5% ~ 35% buffer B for 40 min, and 35% ~ 95% buffer B for 1 min. The system was then maintained in 95% buffer B for 3 min and decreased to 5% within 1 min before equilibrating with 5% buffer B for 10 min. Elution was monitored by measuring absorbance at 214 nm, and fractions were collected every 1 min. The eluted peptides were pooled as 20 fractions and vacuum dried.
HPLC: Each fraction was resuspended in buffer A (2% ACN and 0.1% FA in water) and centrifuged at 20,000 rpm for 10 min. The supernatant was loaded onto a C18 trap column at 5 µL/min for 8 min using an LC-20AD nano-HPLC instrument (Shimadzu, Kyoto, Japan) by the autosampler. Then, the peptides were eluted from the trap column and separated by an analytical C18 column (inner diameter 75 µm) packed in-house. The gradient was run at 300 nL/min starting from 8-35% of buffer B (2% H 2 O and 0.1% FA in ACN) in 35 min, increasing to 60% in 5 min, maintaining at 80% B for 5 min, and nally returning to 5% in 0.1 min, followed by equilibration for 10 min.
Mass Spectrometry Analysis: The peptides separated from nano-HPLC were subjected to tandem mass spectrometry Q EXACTIVE (Thermo Fisher Scienti c, San Jose, CA) for DDA (data-dependent acquisition) detection by nanoelectrospray ionization. The parameters for MS analysis are listed as follows: Bioinformatics: For peptide data analysis, raw mass data were processed using Mascot 2.3.02 (Matrix Science, London, UK) against a database (I-AZaGb004, 144985 sequences). The search parameters are shown in Table 1. Blast2GO software was used for the Gene Ontology (GO) analysis of differentially expressed proteins, and the protein functional categories were determined according to biological process, molecular function and cell component. An online database (http: www.genome.jp/kegg/) was used for the enrichment analysis of KEGG pathways for the differentially expressed proteins and detected the most signi cant pathways.

Results
Wall thickness of ray cells in the xylem of H. ammodendron The ray cell wall of the xylem in H. ammodendron showed an obviously thick wall structure (Fig. 1), and the wall thickness of H. ammodendron ray cells in Jinghe (2.85 ± 0.42 µm) was signi cantly lower than that in Shihezi (3.08 ± 0.44 µm) (P < 0.01).

Differentially expressed proteins
In total, 6767 peptides and 3076 proteins were identi ed with 1% FDR. Finally, repeat experiments de ned differentially expressed proteins with a 1.2-fold change (P < 0.05). A total of 795 and 421 proteins were identi ed as upregulated and downregulated in Shihezi, respectively.
GO annotation for the identi ed proteins The identi ed differentially expressed proteins were enriched in biological processes, cellular components and molecular functions using Gene Ontology. A total of 6,810 differentially expressed proteins functionally related to biological processes. The majority of functions belonged to the categories of metabolic process (24.92%), cellular process (20.73%), and single-organism process (18.43%) (Fig. 2).
There were 6,685 proteins associated with cellular component function. The largest numbers of differentially expressed proteins were found in cells (23.08%), cell parts (23.08%) and organelles (16.92%) (Fig. 3). A total of 3317 differentially expressed proteins correlated with molecular function. The largest numbers of differentially expressed proteins were related to catalytic activity (47.45%), binding (39.16%) and structural molecule activity (3.65%) (Fig. 4).

Pathway annotation for the identi ed proteins
To better understand the pathways of differentially expressed proteins during the formation of ray cell walls in the xylem of H. ammodendron, we performed pathway annotation analysis for differentially expressed proteins based on the KEGG database. The results showed that a total of 9 pathways were related to ray cell wall synthesis in the xylem of H. ammodendron, such as phenylpropanoid biosynthesis, photosynthesis, glycolysis/gluconeogenesis, carbon metabolism, starch and sucrose metabolism, and metabolic pathways (Table 2).

Discussion
The proteome related to ray cell wall formation in the xylem of H. ammodendron from the Gurbantunggut Desert was studied. A total of 3,076 proteins were identi ed, of which 795 proteins were signi cantly upregulated and 421 proteins were signi cantly downregulated in Shihezi, where the wall thickness of H. ammodendron ray cells was higher than that in Jinghe. Among the metabolic pathways involving differentially expressed proteins, phenylpropanoid biosynthesis, photosynthesis, glycolysis/gluconeogenesis, carbon metabolism, starch and sucrose metabolism, metabolic pathways, plant hormone signal transduction, cysteine and methionine metabolism as well as amino sugar and nucleotide sugar metabolism are related to ray cell wall synthesis (Table 2).
Metabolic pathways related to ray cell wall biosynthesis Studies have shown that in poplar, more than 1,600 genes encode carbohydrate-active enzymes, and more than 34 genes are involved in phenylpropanoid biosynthesis and lignin biosynthesis. These results indicate the complexity of both cell wall biosynthesis and secondary xylem development. Phenylpropanoid biosynthesis and lignin biosynthesis are the key pathways in cell wall biosynthesis (Geisler-Lee et al. 2006). Studies utilizing reverse genetics and Arabidopsis mutants showed that phenylpropanoid biosynthesis is involved in cell wall biosynthesis and that carbohydrate-active enzymes are involved in polysaccharide biosynthesis (Brown et al. 2009). The biosynthesis of phenylpropanoid and lignin monomers has been well studied in poplar and Arabidopsis (Boerjan et al. 2003;Vanholme et al. 2008). The phenylpropanoid biosynthesis pathway provided the precursors required for lignin biosynthesis, starting with phenylalanine for deamination to cinnamic acid. Lignin is a heterogeneous phenolic polymer found mainly in secondary thickened cell walls that plays a role in support and defending against diseases and insect pests. Due to its hydrophobicity, lignin provides impermeability for tubular molecules, allowing water and solutes to be transported in microtubule systems (Mellerowicz et al. 2001). In this paper, a total of 34 differentially expressed proteins were involved in phenylpropanoid biosynthesis ( Table 2). These differentially expressed proteins catalyse the biosynthesis of p-hydroxyphenyl lignin, guaiacyl lignin, 5-hydroxy-guaiacyl lignin and syringyl lignin (Fig. 5). The results suggested that phenylpropanoid biosynthesis may be involved in the ray cell wall biosynthesis of the xylem in H. ammodendron.
Some identi ed metabolic pathways, such as glycolysis/gluconeogenesis, carbon metabolism and photosynthesis, provide substrates or energy for cell wall biosynthesis. In starch and sucrose metabolism, sucrose degradation and metabolism provide high levels of energy and substrates during wood formation and participate in the biosynthetic process of both phenolic compounds and lignin during differentiation of xylem cambium (Hauch and Magel 1998). In this paper, a total of 36 differentially expressed proteins were involved in starch and sucrose metabolism ( Table 2). In metabolic pathways, glucose-1-phosphate was catalysed, and UDP-glucose was synthesized. In plants, the glycogen units used in polysaccharide synthesis of the cell wall are derived from UDP-glucose or GDP-mannose (Gibeaut 2000). In this study, a total of 378 differentially expressed proteins were involved in metabolic pathways. Carbon metabolism provides energy and intermediates for secondary wall synthesis. Glucose is used in glycolysis to provide energy for cell wall synthesis (Yang et al. 2007). A total of 71 and 39 differentially expressed proteins were involved in carbon metabolism and glycolysis/gluconeogenesis, respectively.
During photosynthesis, photosystem corresponds to NADP + reduction and the photophosphorylation cycle and consists of at least 8 peptides, and the main components are P700 chlorophyll α A1 and A2 apoproteins (Meng et al. 1988). PS I can also drive both cyclic and pseudocyclic electron transport (Brettel 1997). Cyclic and pseudocyclic electron transport provides the rst stable product of photosynthesis in plants, ATP NADPH. These compounds provide the cells with all the reduction activities and energy required for biosynthesis and energy consumption (Fork and Herbert 1993). In this paper, photosystem I P700 apoprotein A1 was shown to be involved in the process of electron transport in photosynthesis and was signi cantly upregulated in the ray cell wall of H. ammodendron in Shihezi with a higher wall thickness of ray cells.

Upregulated proteins related to ray cell wall biosynthesis
Among the upregulated differentially expressed proteins, a total of 55 proteins, including beta expansin EXPB2.1 (Mirabilis jalapa), glucan endo-1,3-beta-D-glucosidase (Beta vulgaris subsp. vulgaris), hypothetical protein JCGZ_24101 (Jatropha curcas), pectin acetylesterase family protein (Theobroma cacao), polyphenol oxidase (Spinacia oleracea), etc. are related to ray cell wall formation in the xylem of H. ammodendron. Among them, ve differentially expressed proteins were associated with cell wall loosening, the biosynthesis of cellulose and hemicellulose involved 46 differentially expressed proteins, and 4 differentially expressed proteins were involved in pectin biosynthesis (Table 3).  Table 3), which may be involved in the process of cell wall loosening.
Expansin is a protein that loosens the cell walls  Table 3). The xyloglucan endotransglycosylase/hydrolase 1, somatic embryogenesis receptor kinase 1-like precursor and kinase protein with tetratricopeptide repeat domain isoform 1 catalyse biochemical processes from brassinosteroid biosynthesis to cell elongation (Fig. 6, Table 4). Reversibly glycosylated polypeptide 1 is involved in the transport of sugar into the Golgi cavity and the biosynthesis of non-cellulose polysaccharides (Saxena and Brown 1999). In this study, reversibly glycosylated polypeptide 1 was involved in amino sugar and nucleotide sugar metabolism and was signi cantly upregulated in Shihezi (ratio 1.28, Table 3).
GDP-mannose-3′,5′-epimerase (GME) catalyses the conversion of GDP-D-mannose to GDP-L-galactose (Ma et al. 2014), which is a structural component of agar and cell wall polysaccharides (Siow et al. 2013). GME is involved in ascorbic acid biosynthesis in the Smirnoff-Wheeler pathway (Smirnoff and Wheeler 2000), and the rst step is GME catalysing the formation of GDP-D-mannose (Wolucka and Van  Montagu 2003). Ascorbic acid, an antioxidant and cofactor of enzymes, plays an important role in the photosynthesis and biosynthesis of cell wall components (Conklin and Barth 2004). In this paper, GME was shown to be involved in metabolic pathways and was upregulated in the ray cell walls of H. ammodendron in Shihezi (Table 3).
Beta-1,4-glucanase, encoded by the KORRIGAN gene, is involved in cellulose biosynthesis (Szyjanowicz et al. 2004). However, little is known about the functions of KORRIGAN EGase (KOR). In this experiment, beta-1,4-glucanase participated in starch and sucrose metabolism and was upregulated in the ray cell wall of H. ammodendron in Shihezi (Table 3).
In conclusion, during the process of ray cell wall biosynthesis in the xylem of H. ammodendron, it is assumed that the proteins (including predicted: nonspeci c lipid-transfer protein-like protein At5g64080- involved in cellulose, hemicellulose and pectin biosynthesis of the cell wall by providing components or energy. Finally, the proteins in phenylpropanoid biosynthesis promote the ligni cation of ray cell walls and complete the biosynthetic process of cell walls.
In this paper, it is notable that the upregulated proteins corresponded to signi cantly increased wall thickness of ray cells. The regulatory proteins or genes related to wall thickening of ray cells in H. ammodendron can be further explored to determine their functions, and the genes can be applied to the improvement of timber plantations, which is important for both identifying the effect of speci c xylem on the environment and improving the wood mechanical properties of timber plantations.
(2) A total of 795 upregulated proteins and 421 downregulated proteins were found in the ray cell wall of H. ammodendron. Phenylpropanoid biosynthesis, photosynthesis, glycolysis/gluconeogenesis, carbon metabolism, starch and sucrose metabolism, metabolic pathways, plant hormone signal transduction, cysteine and methionine metabolism as well as amino sugar and nucleotide sugar metabolism may be related to the formation of the ray cell wall of H. ammodendron.
(3) During the process of ray cell wall biosynthesis in the xylem of H. ammodendron, it is assumed that nonspeci c lipid-transfer protein-like proteins and beta expansin EXPB2.1 (M. jalapa) rst loosen the cell wall loosening, followed by extension as well as expansion, and the xyloglucan endotransglycosylase/hydrolase 1 cleaves and links the xyloglucan chains. Then, photosystem I P700 apoprotein A1, reversibly glycosylated polypeptide 1 and GDP-mannose-3′,5′-epimerase, etc. are involved in cellulose, hemicellulose and pectin biosynthesis of the cell wall by providing components or energy. Finally, the proteins involved in phenylpropanoid biosynthesis promote the ligni cation and complete the biosynthetic process of ray cell walls. Functional categorization of differentially expressed proteins based on "biological process" Figure 2 Functional categorization of differentially expressed proteins based on "cellular component" Figure 3 Functional categorization of differentially expressed proteins based on "molecular function"

Figure 4
Pathway of phenylpropanoid biosynthesis Figure 5 Anatomical characteristics of H. ammodendron rays a and b (SEM), Tangential section; c, Cross section.
The arrows show ray cell with thick-wall.