Cytosolic Acetyl-CoA Generated by ATP-Citrate Lyase Is Essential for Acetylation of Cell Wall Polysaccharides

Abstract

Plant cell wall polysaccharides, including xylan, glucomannan, xyloglucan and pectin, are often acetylated. Although a number of acetyltransferases responsible for the acetylation of some of these polysaccharides have been biochemically characterized, little is known about the source of acetyl donors and how acetyl donors are translocated into the Golgi, where these polysaccharides are synthesized. In this report, we investigated roles of ATP-citrate lyase (ACL) that generates cytosolic acetyl-CoA in cell wall polysaccharide acetylation and effects of simultaneous mutations of four Reduced Wall Acetylation (RWA) genes on acetyl-CoA transport into the Golgi in Arabidopsis thaliana. Expression analyses of genes involved in the generation of acetyl-CoA in different subcellular compartments showed that the expression of several ACL genes responsible for cytosolic acetyl-CoA synthesis was elevated in interfascicular fiber cells and induced by secondary wall-associated transcriptional activators. Simultaneous downregulation of the expression of ACL genes was demonstrated to result in a substantial decrease in the degree of xylan acetylation and a severe alteration in secondary wall structure in xylem vessels. In addition, the degree of acetylation of other cell wall polysaccharides, including glucomannan, xyloglucan and pectin, was also reduced. Moreover, Golgi-enriched membrane vesicles isolated from the rwa1/2/3/4 quadruple mutant were found to exhibit a drastic reduction in acetyl-CoA transport activity compared with the wild type. These findings indicate that cytosolic acetyl-CoA generated by ACL is essential for cell wall polysaccharide acetylation and RWAs are required for its transport from the cytosol into the Golgi.

Introduction

Plant cell walls constitute the bulk of the estimated 450 billion tons of plant carbon biomass in the biosphere (Bar-On et al. 2018). Plant cell wall-based biomass has been widely used in our daily life with applications ranging from lumber, furniture, paper, cardboard, textile and fuelwood to potential lignocellulosic biofuel production. To better tailor plant biomass for diverse uses, it is imperative to build a knowledge base of how plant cell walls are synthesized and assembled. The main constituents of plant cell walls are polysaccharides, including cellulose, hemicelluloses (xylan, mannan, xyloglucan and mixed-linkage glucan) and pectin. Among them, xylan, mannan, xyloglucan and pectin are O-acetylated. In particular, up to 60% of xylose (Xyl) residues in xylans from woody species may be esterified with acetyl groups. As a result, acetate could account for up to 6.7% (w/w) of wood biomass in poplar (Johnson et al. 2017). Although the acetylation of cell wall polysaccharides is important for their physiochemical properties, it impedes the use of lignocellulosic biomass for biofuel production (Pauly and Ramirez 2018). Understanding the biochemical mechanism controlling cell wall polysaccharide acetylation will potentially provide molecular tools to custom-design cell wall composition better suited for biofuel production.

Genetic and biochemical studies have implicated three groups of proteins, including DUF231-containing proteins, Altered Xyloglucan9 (AXY9) and Reduced Wall Acetylation (RWA) proteins, in cell wall polysaccharide acetylation. Members of the DUF231 family have been demonstrated to be O-acetyltransferases catalyzing the acetylation of xylan, mannan and xyloglucan. Xylan can be substituted with acetyl groups at O-2 or O-3 of Xyl residues, at both O-2 and O-3 of the same Xyl residue and at O-3 of Xyl residues that are substituted at O-2 with glucuronic acid (GlcA) (Zhong et al. 2019b). In Arabidopsis, genetic analysis revealed the involvement of nine DUF231 members in xylan O-acetylation. Among them, ESK1 (Eskimo1) plays a dominant role in 2-O- and 3-O-monoacetylation and 2,3-di-O-acetylation, and TBL32 (Trichome Birefringence-Like32) and TBL33 are specifically responsible for 3-O-acetylation of 2-O-GlcA-substituted Xyl residues (Xiong et al. 2013, Yuan et al. 2013, Yuan et al. 2016a, Yuan et al. 2016b, Yuan et al. 2016c, Zhong et al. 2017a). Further biochemical studies have established that these Arabidopsis DUF231 members and their orthologs from poplar and rice are xylan O-acetyltransferases (XOATs) catalyzing the regiospecific acetylation of xylan (Urbanowicz et al. 2014, Zhong et al. 2017a, Zhong et al. 2018a, Zhong et al. 2018b). The acetylation of mannan occurs at O-2 or O-3 of mannosyl residues and biochemical studies have demonstrated that a group of phylogenetically closely related DUF231 members are mannan O-acetyltransferases (MOATs) mediating mannan acetylation in different taxa of land plants (Zhong et al. 2018d, Zhong et al. 2019a). Xyloglucan can be O-acetylated on the side-chain galactose (Gal) residues for the XXXG-type and on the backbone glucose (Glc) residues for the XXGG-type (Pauly and Keegstra 2016). Two Arabidopsis DUF231 members, AXY4 and AXY4L, were genetically shown to be essential for the acetylation of the XXXG-type xyloglucan (Gille et al. 2011) and biochemically demonstrated together with their poplar orthologs to be xyloglucan O-acetyltransferases (XGOATs) mediating O-acetylation of fucosylated Gal residues on xyloglucan side chains (Zhong et al. 2018c). A grass-specific DUF231 protein from Brachypodium distachyon, BdXyBAT1 (Xyloglucan backbone acetylation1), was reported to be involved in O-acetylation of the backbone Glc residues in the XXGG-type xyloglucan, although its enzymatic activity has yet to be determined (Liu et al. 2016). Pectins (homogalacturonan and rhamnogalacturonan I) have acetyl moieties attached to galacturonic acid residues at O-2 and/or O-3 (Ishii 1997). Mutational analysis indicates an involvement of TBL10, a DUF231 member, in the acetylation of rhamnogalacturonan I, but its enzymatic activity remains to be studied (Stranne et al. 2018). All of these DUF231 proteins contain the conserved GDS and DXXH motifs that have been shown to be essential for the acetyltransferase activities of XOATs, MOATs and XGOATs (Zhong et al. 2017a, Zhong et al. 2018c, Zhong et al. 2018d). It was proposed that like the bacterial alginate O-acetyltransferase AlgX (Riley et al. 2013), the Ser residue in the GDS motif, together with the His and Asp residues in the DXXH motif of these OATs, likely forms a Ser–His–Asp catalytic triad required for their mechanism of action.

While the O-acetyltransferases in the DUF231 family are specific for a particular cell wall polysaccharide, the other two groups of proteins, AXY9 and RWAs, affect acetylation of more than one polysaccharide when their genes are mutated. Mutation of AXY9 causes a reduction in the acetylation of both xylan and xyloglucan (Schultink et al. 2015). Although recombinant AXY9 protein did not exhibit any acetyltransferase activity toward xylan, xyloglucan or mannan, it displayed a weak acetylesterase activity toward several pseudosubstrates (Zhong et al. 2018c). It was proposed that AXY9 might act as an intermediate transferring acetyl groups from an acetyl donor to the O-acetyltransferases responsible for the acetylation of xylan and xyloglucan. The expression of the four Arabidopsis RWA genes is closely associated with secondary wall-forming cells and their simultaneous mutations resulted in an overall reduction in the acetylation of not only xylan but also xyloglucan, mannan and pectin (Lee et al. 2011, Manabe et al. 2013). RWAs harbor multiple transmembrane helices and show sequence homology to the fungal Cas1p (capsule synthesis1 protein) that is required for the acetylation of the capsule polysaccharide glucuronoxylomannan (Janbon et al. 2001). However, unlike Cas1p, they lack a putative acetyltransferase domain and their exact functions are currently unknown.

Recombinant XOATs, MOATs and XGOATs can use acetyl-CoA as an acetyl donor to transfer acetyl groups onto their respective oligosaccharide acceptors in vitro (Urbanowicz et al. 2014, Zhong et al. 2017a, Zhong et al. 2018c, Zhong et al. 2018d), implying that acetyl-CoA could be an acetyl donor for cell wall polysaccharide acetylation in planta. Acetyl-CoA in plant cells is synthesized in four subcellular compartments, i.e. plastid, mitochondrion, cytosol and peroxisome, by distinct acetyl-CoA-generating enzymes including acetyl-CoA synthetase and plastid pyruvate dehydrogenase complex in plastids, mitochondria pyruvate dehydrogenase complex in mitochondria, ATP-citrate lyase (ACL) in the cytosol and 3-ketoacyl-CoA thiolase in peroxisomes (Xing and Poirier 2012). Because no acetyl-CoA-generating pathway is known to exist in the Golgi and the lipid membrane is impermeable to acetyl-CoA (Xing and Poirier 2012), acetyl-CoA must be transported into the Golgi if it is used for cell wall polysaccharide acetylation. Currently, little is known about the source of acetyl donor for cell wall polysaccharide acetylation and how it is transported into the Golgi. In this report, we provide genetic evidence indicating that cytosolic acetyl-CoA generated by ACL serves as acetyl donor for cell wall polysaccharide acetylation and its transport from the cytosol into the Golgi requires the multitransmembrane RWA proteins.

Results

Differential acetyl-donor promiscuity of XOAT1/ESK1, MOAT4 and XGOAT1/AXY4

The bacterial peptidoglycan O-acetyltransferase B (PatB), which is involved in bacterial cell wall polysaccharide acetylation, has been shown to exhibit promiscuity toward a variety of acetyl donor substrates including acetyl-CoA and several pseudosubstrates (Moynihan and Clarke 2014). To elucidate whether plant cell wall polysaccharide O-acetyltransferases also display acetyl-donor promiscuity, we tested the activities of recombinant XOAT1/ESK1, MOAT4 and XGOAT1/AXY4 (Supplementary Fig. S1), three representative O-acetyltransferases catalyzing the acetylation of xylan, mannan and xyloglucan, respectively, using several acetyl donor substrates. The recombinant proteins were incubated with their respective oligosaccharide acceptors and acetyl-CoA, p-nitrophenyl acetate (pNP-Ac), 4-methylumbelliferyl acetate (4MU-Ac) or acetylsalicylic acid (Ac-SA; Oi and Satomura 1967, Moynihan and Clarke 2014). Examination of the reaction products by ¹H-NMR spectroscopy revealed strong resonance signals attributed to acetyl groups in the XOAT1/ESK1-catalyzed reactions with all of the tested acetyl donor substrates (Fig. 1A, B), suggesting that XOAT1/ESK1 could use these substrates as acetyl donors to acetylate xylan. By contrast, in the MOAT4- and XGOAT1/AXY4-catalyzed reaction products, the acetyl signals were only evident when acetyl-CoA was used, and little signals were seen when pNP-Ac, 4MU-Ac and Ac-SA were used (Fig. 1C, D), implying that MOAT4- and XGOAT1/AXY4 could only effectively use acetyl-CoA but not pNP-Ac, 4MU-Ac or Ac-SA as an acetyl donor to acetylate mannan and xyloglucan, respectively. These results demonstrate that XOAT1/ESK1 exhibits a broad acetyl-donor promiscuity, whereas MOAT4 and XGOAT1/AXY4 have a strict acetyl-donor specificity.

Fig. 1

Assay of acetyltransferase (A to D) and acetylesterase (E and F) activities of recombinant XOAT1/ESK1, MOAT4 and XGOAT1/AXY4 using various acetyl donor substrates. (A) ¹H-NMR spectra of control reaction products showing the resonance region for acetyl groups. For each recombinant protein, the control reactions contained heat-inactivated protein, the respective oligosaccharide acceptor and the acetyl donor substrate (acetyl-CoA, pNP-Ac, 4MU-Ac or Ac-SA). Since all control reactions had the same pattern of spectra, only those for XOAT1/ESK1 were shown. (B) ¹H-NMR spectra of the XOAT1/ESK1-catalyzed reaction products showing the resonance region for acetyl groups. The reactions contained XOAT1/ESK1, xylohexaose and various acetyl donor substrates. (C) ¹H-NMR spectra of the MOAT4-catalyzed reaction products showing the resonance region for acetyl groups. The reactions contained MOAT4, mannohexaose and various acetyl donor substrates. (D) ¹H-NMR spectra of the XGOAT1/AXY4-catalyzed reaction products showing the resonance region for acetyl groups. The reactions contained XGOAT1/AXY4, xyloglucan oligomers and various acetyl donor substrates. (E) Time course of pNP release by incubating XOAT1/ESK1, MOAT4 or XGOAT1/AXY4 with pNP-Ac. (F) Time course of 4MU release by incubating XOAT1/ESK1, MOAT4 or XGOAT1/AXY4 with 4MU-Ac. Error bars denote SE of three biological replicates.

Although XGOAT1/AXY4 exhibited little O-acetyltransferase activity when pNP-Ac or 4MU-Ac were used as an acetyl donor (Fig. 1D), we previously found that it displayed a weak acetylesterase activity toward them (Zhong et al. 2018c). The acetyl esterase activity of bacterial peptidoglycan O-acetyltransferases has been considered as the initial step of the acetyltransferase reaction (Moynihan and Clarke 2014, Sychantha et al. 2017). We next examined whether XOAT1/ESK1 and MOAT4 also possessed such an acetylesterase activity. It was shown that like XGOAT1/AXY4, ESK1/XOAT1 and MOAT4 readily hydrolyzed pNP-Ac and 4MU-Ac, releasing pNP and 4MU, respectively (Fig. 1E, F). This finding indicates that although MOAT4 and XGOAT1/AXY4 were unable to use pNP-Ac or 4MU-Ac as an acetyl donor for their O-acetyltransferase activity, they display comparable acetylesterase activities as XOAT1/ESK1.

The Arabidopsis ACLA and ACLB genes are upregulated in secondary wall-forming cells and induced by secondary wall-associated NAC and MYB transcription factors

The finding that acetyl-CoA can be efficiently used as an acetyl donor by all three representative polysaccharide O-acetyltransferases implies that acetyl-CoA is most likely the acetyl donor for cell wall polysaccharide acetylation in plant cells. Since xylan, one of the major secondary wall components, is heavily acetylated, we reasoned that if acetyl-CoA is the acetyl donor for xylan acetylation, genes involved in the supply of acetyl-CoA should be highly induced during secondary wall biosynthesis. Therefore, we examined whether the expression of any genes involved in acetyl-CoA generation was associated with secondary wall biosynthesis in Arabidopsis. Acetyl-CoA is synthesized by a suite of enzymes in four distinct subcellular compartments, including plastid, mitochondrion, cytosol and peroxisome (Xing and Poirier 2012; Fig. 2A). Because acetyl-CoA participates in many essential biochemical reactions, as expected, most genes encoding the acetyl-CoA biosynthetic enzymes were ubiquitously expressed in various Arabidopsis organs (Supplementary Fig. S2). Comparison of the expression levels of these genes in laser-dissected interfascicular fiber cells and pith cells from Arabidopsis stems revealed that several ACL genes were highly upregulated in the secondary-wall containing fiber cells compared with the pith parenchyma cells (Fig. 2B). ACL is located in the cytosol and responsible for generating the cytosolic pool of acetyl-CoA by catalyzing the ATP- and CoA-dependent cleavage of citrate. It is a hetero-octamer consisting of four A-subunits (ACLA) and four B-subunits (ACLB) and in Arabidopsis, the ACLA subunit is encoded by three genes (ACLA-1, ACLA-2 and ACLA-3) and the ACLB subunit is encoded by two genes (ACLB-1 and ACLB-2; Fatland et al. 2002, Fatland et al. 2005).

Fig. 2

Quantitative PCR analysis of the expression of genes involved in acetyl-CoA production. (A) Diagram of acetyl-CoA generation in different subcellular compartments and the potential source of acetyl donor for cell wall polysaccharide acetylation. (B) Relative expression levels (shown as a heat map) of genes involved in acetyl-CoA production in laser-dissected interfascicular fiber cells (fiber/pith; shown as ratios of the expression level in fiber cells over that in pith cells) and in leaves of plants overexpressing SND1 (SND1-OE), VND7 (VND7-OE) and MYB46 (MYB46-OE) (shown as fold induction in the overexpressors over that in the wild type). The data were from three biological replicates. ACL, ATP-citrate lyase; ACS, acetyl-CoA synthetase; AXY9, altered xyloglucan9; KAT, 3-keto-acyl-CoA thiolase; MtPDHC, mitochondrial pyruvate dehydrogenase complex; OAT, O-acetyltransferase; PtPDHC, plastid pyruvate dehydrogenase complex; RWA, reduced wall acetylation.

To further corroborate the correlation of ACL expression and secondary wall biosynthesis, we examined whether the expression of ACL genes was induced by several secondary wall-associated master transcriptional switches, including SND1, VND7 and MYB46 (Fig. 2B). Overexpression of these master transcriptional switches has been shown to activate secondary wall biosynthetic genes, leading to ectopic secondary wall deposition in epidermis and mesophyll cells of leaves and concomitantly a curly leaf phenotype (Zhong and Ye 2015). Consistent with their upregulation in fiber cells, the ACLA-1, ACLA-2 and ACLB-2 genes were found to be highly induced by overexpression of SND1, VND7 and MYB46 in transgenic Arabidopsis plants (Fig. 2B). By contrast, none of the genes involved in acetyl-CoA generation in plastid, mitochondrion and peroxisome were significantly induced by SND1, VND7 and MYB46. These results demonstrated a close association of the expression of ACL genes responsible for generating the cytosolic acetyl-CoA pool with secondary wall biosynthesis.

RNAi downregulation of the expression of ACLA and ACLB in Arabidopsis leads to reduced acetylation of xylan

Having established the association of ACL expression with secondary wall biosynthesis, we next set out to examine the effect of downregulation of ACL expression on the acetylation of xylan, one of the major polysaccharides in secondary walls. To do this, we generated two sets of transgenic Arabidopsis RNAi plants, one with simultaneous downregulation of all three ACLA genes (ACLA-RNAi) and the other with simultaneous downregulation of both ACLB genes (ACLB-RNAi; Fig. 3 and Supplementary Fig. S3). Similar to the observed phenotypes of antisense ACLA-1 transgenic plants in a previous report (Fatland et al. 2005), downregulation of either ACLA or ACLB genes resulted in a significant reduction in plant growth, ranging from moderate to severe phenotypes. Plants with a severe phenotype did not yield enough materials for cell wall analysis, and thus only those with a moderate phenotype (Fig. 3A, B) were used for study of the effects of ACL downregulation on xylan acetylation. ACLA- and ACLB-RNAi plants often displayed wilted leaves in sunny afternoons when grown in the greenhouse (Fig. 3A) and their inflorescence stems had deformed xylem vessels (Fig. 3C–E), which was reminiscent of the Arabidopsis esk1 mutant defective in xylan acetylation (Yuan et al. 2013). Furthermore, transmission electron microscopy revealed that while the secondary walls in the xylem vessels of wild-type stems displayed a regular pattern of layered deposition (Fig. 4A), those in the xylem vessels of ACLA- and ACLB-RNAi stems were drastically disorganized (Fig. 4B, C). The altered secondary wall structure phenotype was nearly identical to that observed in the tbl33 esk1 mutant with a strong reduction in xylan acetylation (Fig. 4D; Yuan et al. 2016a), indicating a possible xylan acetylation defect caused by downregulation of ACL expression. Cell wall composition analysis revealed no significant alterations in the amounts of xylose and mannose in the ACLA- and ACLB-RNAi lines (Fig. 3F). An elevation in the amounts of Gal, Glc, rhamnose, arabinose and galacturonic acid was observed in ACLB-RNAi lines, suggesting an increase in the relative levels of xyloglucan and pectins in the cell walls of ACLB-RNAi lines compared with the wild type.

Fig. 3

Effects of RNAi downregulation of ACL expression on plant growth, vessel morphology and total cell wall acetylation. (A) and (B) ACLA- and ACLB-RNAi plants with a moderate phenotype showing wilting leaves (A) and retarded growth (A and B). (C–E) Cross sections of stems showing deformed xylem vessels (arrows) in the xylem bundles of ACLA- and ACLB-RNAi plants (D and E) compared with the wild type (C). Bars = 72 μm. (F) Glycosyl composition analysis of cell walls from the stems of ACLA- and ACLB-RNAi plants. Error bars denote SE of two biological replicates. (G) Reduction of total cell wall acetylation in the stems of ACLA- and ACLB-RNAi plants. Error bars denote SE of three biological replicates.

Fig. 4

Effect of RNAi downregulation of ACL expression on secondary wall structure in xylem vessels. Ultrathin sections of stems were cut, stained with lead citrate and uranyl acetate and then visualized for secondary walls in xylem vessels under a transmission electron microscope. (A) Secondary walls in the xylem vessels of the wild type showing a regular pattern of layered deposition. (B–D) secondary walls in the xylem vessels of ACLA-RNAi (B), ACLB-RNAi (C) and the tbl33 esk1 double mutant (D) showing their drastically disorganized structure. ve, vessel; xf, xylary fiber. Bars = 1.6 μm.

Measurement of the cell wall acetyl content in the stems revealed a large reduction in the ACLA- and ACLB-RNAi plants compared with the wild type; the acetyl content in the stem cell walls of ACLA- and ACLB-RNAi plants was reduced to 43% and 46%, respectively, of that of the wild type (Fig. 3G). Because the bulk of the acetylated cell wall polysaccharides in stems is xylan, downregulation of ACL expression most likely had a major impact on xylan acetylation. Therefore, we next analyzed xylan acetylation in the stems of ACLA- and ACLB-RNAi plants. Acetyl xylan was extracted with dimethyl sulfoxide (DMSO) and digested with xylanase into xylooligomers for structural characterization. Matrix-assisted laser desorption ionization-time-of-flight mass spectrometry (MALDI-TOF MS) demonstrated an altered pattern of acetylation in the acetyl xylooligomers from ACLA- and ACLB-RNAi plants compared with the wild type (Fig. 5A). While the acetyl xylooligomers from the wild type had predominant signal peaks corresponding to di- and tri-acetylated xylooligomers [Xyl₆(GlcA)(Ac)₂ at m/z 1,093, Xyl₆(MeGlcA)(Ac)₂ at m/z 1,107, Xyl₆(MeGlcA)(Ac)₃ at m/z 1,149, Xyl₇(GlcA)(Ac)₂ at m/z 1,225 and Xyl₇(GlcA)(Ac)₃ at m/z 1,267], those from ACLA- and ACLB-RNAi plants displayed highly elevated signal peaks corresponding to non-acetylated [Xyl₄(GlcA) at m/z 745 and Xyl₄(MeGlcA) at m/z 759] and mono-acetylated [Xyl₄(MeGlcA)(Ac) at m/z 801] xylotetramers. The increase of these shorter, less-acetylated xylooligomers in ACLA- and ACLB-RNAi plants indicates a reduction in the frequency of acetyl groups in xylan, which renders it more accessible to xylanase digestion. The degree of xylan acetylation in ACLA- and ACLB-RNAi plants was further analyzed by nuclear magnetic resonance (NMR) spectroscopy (Fig. 5B). ¹H-NMR spectra of xylooligomers from the wild type and the ACLA- and ACLB-RNAi plants showed resonances characteristic of acetyl groups around 2.22 ppm and those for the sugar groups between 3.0 and 5.5 ppm. Integration analysis demonstrated that the amount of acetyl groups in the xylan of ACLA- and ACLB-RNAi plants was reduced to 47% and 53%, respectively, of that of the wild type (Fig. 5B), indicating that the decrease in the overall cell wall acetyl groups in the stem of ACLA- and ACLB-RNAi plants is largely attributed to the reduction in xylan acetylation.

Fig. 5

Effect of RNAi downregulation of ACL expression on xylan acetylation. (A) MALDI-TOF MS analysis of xylooligomers released by xylanase digestion of xylan from the wild type, ACLA- and ACLB-RNAi plants. Ion peaks are marked with their mass and the identity of the corresponding xylooligomers are shown in the table. Xyl_n(GlcA)_n(Ac)_n represents a xylooligomer with n number of xylosyl residues substituted with n number of GlcA and n number of acetyl groups (Ac). (B) ¹H NMR spectra of xylan from the wild type, ACLA- and ACLB-RNAi plants. The degree of xylan acetylation (DS_AC) and its percentage relative to the wild type (in parentheses) are shown at the right of each spectrum. The resonances ranging from 3.0 to 5.5 ppm correspond to carbohydrate and those between 2.05 and 2.25 ppm to acetyl groups. The data shown are representatives of three biological replicates.

RNAi downregulation of the expression of ACLA and ACLB causes a reduction in the acetylation of glucomannan, xyloglucan and pectin

To find out whether downregulation of ACL expression also affected the acetylation of other cell wall polysaccharides, we isolated glucomannan, xyloglucan and pectin and then determined their degree of acetylation. MALDI-TOF MS revealed that the predominant ion signals of glucomannosyl oligomers released by mannanase digestion of wild-type glucomannan corresponded to mono-acetylated [(M/G)₃Ac at m/z 569 and (M/G)₄Ac at m/z 731]) and di-acetylated [(M/G)₃Ac₂ at m/z 611, (M/G)₄Ac₂ at m/z 773 and (M/G)₅Ac₂ at m/z 935] oligomers, but little ion signals for non-acetylated ones were observed except for glucomannosyl trimer at m/z 527 (Fig. 6A). Although the mono- and di-acetylated oligomers were still prevalent in glucomannosyl oligomers from ACLA- and ACLB-RNAi plants, it was evident that the ion signals corresponding to nonacetylated oligomers, including glucomannosyl trimer (m/z 527), tetramer (m/z 689), pentamer (m/z 851) and hexamer (m/z 1,013), were significantly elevated compared with the wild type (Fig. 6A), indicating that downregulation of ACL genes resulted in a decrease in the frequency of acetyl groups in glucomannan. ¹H-NMR spectroscopy further showed that the degree of glucomannan acetylation in ACLA- and ACLB-RNAi plants was reduced to 86% and 71%, respectively, of that in the wild type (Fig. 6B). A similar decrease in the acetylation of xyloglucan and pectin was also detected by ¹H-NMR spectroscopy. Specifically, the degree of xyloglucan acetylation in ACLA- and ACLB-RNAi plants was reduced to 78% and 72%, respectively, of that in the wild type (Fig. 7A), and that of pectin acetylation in ACLA- and ACLB-RNAi plants was reduced to 46% and 52%, respectively, of that in the wild type (Fig. 7B). These results demonstrate that downregulation of ACL expression has a broad negative impact on the acetylation of cell wall polysaccharides, including xylan, glucomannan, xyloglucan and pectin.

Fig. 6

Effect of RNAi downregulation of ACL expression on glucomannan acetylation. (A) MALDI-TOF MS analysis of glucomannosyl oligomers released by mannanase digestion of mannans from the wild type, ACLA- and ACLB-RNAi plants. Ion peaks are marked with their mass and the identity of the corresponding glucomannosyl oligomers. M/G denotes glucomannosyl oligomers. (M/G)₃, (M/G)₄, (M/G)₅ and (M/G)₆ represent glucomannosyl triose, tetraose, pentaose and hexaose, respectively. (M/G)_nAc and (M/G)_nAc₂ represent glucomannosyl oligomers decorated with one and two acetyl groups, respectively. (B) ¹H NMR spectra of glucomannan from the wild type, ACLA- and ACLB-RNAi plants. The resonances for carbohydrate (3.0–5.5 ppm) and acetyl groups (2.0–2.2 ppm) are marked at the top of the spectra. The degree of glucomannan acetylation (DS_AC) and its percentage relative to the wild type (in parentheses) are shown at the right of each spectrum. The data shown are representatives of three biological replicates.

Fig. 7

Effect of RNAi downregulation of ACL expression on the acetylation of xyloglucan (A) and pectin (B). The resonances for carbohydrate (3.0–5.5 ppm) and acetyl groups (2.0–2.3 ppm for xyloglucan and 1.9–2.3 for pectin) are marked at the top of the spectra. The degree of acetylation (DS_AC) in xyloglucan (A) and pectin (B) and its percentage relative to the wild type (in parentheses) are shown at the right of each spectrum. The data shown are representatives of three biological replicates.

Simultaneous mutations of the four RWA genes in Arabidopsis cause a drastic reduction in acetyl-CoA transport into Golgi-enriched membrane vesicles

The finding that ACL expression is essential for cell wall polysaccharide acetylation implies that the cytosolic acetyl-CoA pool generated by ACL is the source of acetyl donor for cell wall polysaccharide acetylation. Since acetyl-CoA is impermeable to lipid membranes (Xing and Poirier 2012), it needs to be transported via transporters from the cytosol into the Golgi where the acetylation of cell wall polysaccharides occurs. The Golgi-localized multitransmembrane proteins, RWAs, are candidates for such transporters based on previous studies showing that simultaneous mutations of the four Arabidopsis RWA genes caused defects in cell wall polysaccharide acetylation (Lee et al. 2011, Manabe et al. 2013). To test this hypothesis, we investigated the effect of the RWA mutations on the transport of acetyl-CoA into the Golgi. Golgi-enriched membrane vesicles were isolated from the stems of the wild type and the Arabidopsis rwa1/2/3/4 mutant and tested for acetyl-CoA transport activities. It was found that although Golgi-enriched vesicles from both the wild type and rwa1/2/3/4 displayed time-dependent acetyl-CoA transport activities when incubated with radiolabeled acetyl-CoA, the activity of rwa1/2/3/4 vesicles was reduced to about 30% of that of the wild type (Fig. 8A). In comparison, the activities of xylan xylosyltransferase and xylan methyltransferase were not significantly altered in rwa1/2/3/4 compared with the wild type (Fig. 8B). The finding that the RWA mutations reduced acetyl-CoA transport into Golgi-enriched membrane vesicles indicates that RWAs are essential for the transport of acetyl-CoA from the cytosol into the Golgi.

Fig. 8

Acetyl-CoA transport activities of Golgi-enriched membrane vesicles from the wild type and the rwa1/2/3/4 mutant. (A) Time course of acetyl-CoA transport into the membrane vesicles showing a much lower transport activity in the rwa1/2/3/4 mutant than the wild type. (B) Assay of xylan xylosyltransferase (left panel) and methyltransferase (right panel) activities in the membrane vesicles showing similar activities between the wild type and the rwa1/2/3/4 mutant. Golgi-enriched membrane vesicles isolated from the stems of the wild type and the rwa1/2/3/4 mutant were used for the activity assays. Error bars were SE of three biological replicates.

Discussion

There exist four distinct pools of acetyl-CoA in plant cells, i.e. plastid, mitochondrion, cytosol and peroxisome. The cytosolic acetyl-CoA pool is implicated in the acetylation of a variety of metabolites, including alkaloids, anthocyanins, isoprenoids and phenolics, and in the synthesis of various compounds such as waxes, sterols and rubber (Xing and Poirier 2012). Cytosolic acetyl-CoA in plant cells is generated by the heteromeric enzyme ACL composed of ACLA and ACLB subunits and antisense RNA downregulation of ACLA-1 in Arabidopsis was shown to cause a severe defect in plant growth and a reduction in the accumulation of several cytosolic acetyl-CoA derived products, including cuticular wax in stems and flavonoids in seeds (Fatland et al. 2005). Cytosolic acetyl-CoA was also reported to be essential for histone H3K27 acetylation as demonstrated by artificial microRNA downregulation of ACLA-1 (Chen et al. 2017). A number of cell wall polysaccharides are known to be acetylated. Although recombinant cell wall polysaccharide O-acetyltransferases, such as XOATs, MOATs and XGOATs, readily use acetyl-CoA as an acetyl donor in vitro, no genetic proof is available to indicate the source of the acetyl donor for cell wall polysaccharide acetylation in planta. Similarly, the identity of the acetyl donor for bacterial cell wall polysaccharide O-acetyltransferases, such as PatB (peptidoglycan O-acetyltransferase B), OatA (peptidoglycan O-acetyltransferase A) and AlgX (alginate O-acetyltransferase X), was not resolved as they exhibited acetyl donor promiscuity in vitro (Baker et al. 2014, Moynihan and Clarke 2014, Sychantha et al. 2017). In this report, we provide genetic evidence demonstrating that RNAi downregulation of ACL expression resulted in a significant reduction in the acetylation of xylan, glucomannan, xyloglucan and pectin, indicating that cytosolic acetyl-CoA generated by ACL serves as acetyl donor for cell wall polysaccharide acetylation (Fig. 2A).

It is interesting to find that although cytosolic acetyl-CoA generated by ACL is ubiquitously needed for diverse cellular functions, the expression of ACL genes was upregulated in secondary wall-forming cells and induced by secondary wall-associated transcriptional activators (Fig. 2B). The increased ACL expression in secondary wall-forming cells is most likely to provide a steady supply of acetyl-CoA to meet the high demand of acetyl donor for the acetylation of xylan, one of the major secondary wall polysaccharides. This is reminiscent of UDP-Xyl synthase genes, whose high-level expression in secondary wall-forming cells was needed to supply UDP-Xyl for xylan backbone elongation during secondary wall biosynthesis (Zhong et al. 2017b). Consistent with the association of increased ACL expression with secondary wall biosynthesis, RNAi downregulation of ACL expression led to a significant decrease in the degree of xylan acetylation. Xylanase digestion of xylan from the ACLA-RNAi and ACLB-RNAi plants generated an increased proportion of xylooligomers with fewer or no acetyl groups compared with the wild type (Fig. 5). This xylooligomer distribution pattern resembles that of xylan from the tbl33 esk1 mutant deficient in xylan acetylation (Yuan et al. 2016a). In addition, similar to the observed secondary wall defect in the tbl33 esk1 mutant (Yuan et al. 2016a), the xylem vessels of ACL-RNAi plants displayed a severe disorganization of secondary walls (Fig. 4). Although the altered secondary wall structure in ACL-RNAi plants was most likely due to xylan acetylation deficiency, a possibility that the reduction in pectin acetylesterification might also affect the secondary wall structure could not be excluded. A previous study showed that an alteration in pectin methylesterification affected the mechanical strength of stems (Hongo et al. 2012). RNAi downregulation of ACL expression also caused a reduction in the acetylation of glucomannan, xyloglucan and pectin albeit with a lesser effect on glucomannan and xyloglucan. It should be noted that downregulation of ACL expression resulted in plant growth retardation ranging from moderate to severe phenotypes. Since the severely retarded plants could not yield enough materials for cell wall analysis, only transgenic plants with a moderate phenotype were used in this study. Therefore, we only observed a partial loss of cell wall polysaccharide acetylation in these plants.

The bacterial O-acetyltransferases PatB and OatA and the human sialate O-acetyltransferase CASD1 have been shown to exhibit weak acetylesterase activities in vitro (Moynihan and Clarke 2014, Baumann et al. 2015, Sychantha et al. 2017). The acetyl esterase activity of these O-acetyltransferases was proposed to carry out the first step in the proposed ping-pong bi-bi mechanism of action, i.e. transferring the acetyl moiety to the Ser residue in the Ser–His–Asp catalytic triad to form a covalent acetyl-enzyme intermediate. Subsequently, the acetyl group from the acetyl-enzyme intermediate is transferred to the hydroxyl group of the acceptor sugar by the acetyltransferase activity (Moynihan and Clarke 2014, Baumann et al. 2015, Sychantha et al. 2017). Mutation of the catalytic Ser residue in CASD1 was shown to abolish the formation of the acetyl-enzyme intermediate and the subsequent acetyl transfer onto sialic acid (Baumann et al. 2015). Our finding that XOAT1/ESK1, MOAT4 and XGOAT1/AXY4 all possess weak acetylesterase activities indicate that plant cell wall polysaccharide O-acetyltransferases also likely employ a ping-pong bi-bi mechanism of action for polysaccharide O-acetylation. This is further supported by site-directed mutagenesis showing that mutation of Ser to Ala in the putative Ser–His–Asp catalytic triad of XOAT1/ESK1, MOAT1 and XGOAT1/AXY4 eliminated their ability to acetylate cell wall polysaccharides (Zhong et al. 2017a, Zhong et al. 2018c, Zhong et al. 2018d). It is intriguing that although XOAT1/ESK1, MOAT4 and XGOAT1/AXY4 all exhibited acetylesterase activities toward pNP-Ac and 4MU-Ac, only XOAT1 could efficiently use pNP-Ac and 4MU-Ac as acetyl donors to acetylate its acceptor (Fig. 1). It is possible that the catalytic site of XOAT1/ESK1 has a broad plasticity that can accommodate both the promiscuous acetyl donor and its acceptor, whereas binding of the promiscuous acetyl donor to the catalytic site of MOAT4 and XGOAT1/AXY4 may result in a conformational change rendering it unfit to bind the acceptor.

Our finding that Golgi-enriched membrane vesicles from the rwa1/2/3/4 quadruple mutant exhibited a significantly reduced acetyl-CoA transport activity compared with the wild-type provides genetic evidence for the involvement of the Golgi-localized RWAs in channeling acetyl-CoA from the cytosol into the Golgi lumen. A role of RWAs in acetyl-CoA transport is consistent with their predicted multiple transmembrane helices and their mutant chemotype showing a defect in the acetylation of multiple cell wall polysaccharides (Lee et al. 2011, Manabe et al. 2013). The requirement of RWAs in acetyl-CoA transport indicates that they may be functionally equivalent to the bacterial transmembrane proteins PatA and AlgI involved in bacterial cell wall polysaccharide acetylation. PatA and AlgI were proposed to transport the unidentified acetyl donor from the cytoplasm into the periplasm, which supplies acetyl groups to the respective O-acetyltransferases PatB and AlgX to acetylate their acceptors although the transport activities of PatA and AlgI have not yet been proven (Baker et al. 2014, Moynihan and Clarke 2014). It should be noted that although the rwa1/2/3/4 quadruple mutant had a significant reduction in acetyl-CoA transport activity compared with the wild type, it still retained about 30% of activity (Fig. 8). This could be due to an incomplete knockout of the functions of all the four RWA genes or the existence of an RWA-independent mechanism for acetyl-CoA transport. Nevertheless, this observation is consistent with early reports showing that only a partial reduction in the acetylation of cell wall polysaccharides was observed in the rwa1/2/3/4 quadruple mutant (Lee et al. 2011, Manabe et al. 2013).

In summary, we have demonstrated that RNAi downregulation of ACL expression leads to a significant reduction in cell wall polysaccharide acetylation and simultaneous mutations of the four Arabidopsis RWA genes result in a defect in acetyl-CoA transport into Golgi-enriched membrane vesicles. Our results indicate that the cytosolic acetyl-CoA pool generated by ACL provides the supply of acetyl donor for cell wall polysaccharide acetylation and RWAs are required for the transport of acetyl-CoA from the cytosol into the Golgi. Our findings provide further support to the biochemical mechanism of plant cell wall polysaccharide acetylation involving acetyl-CoA transporters and polysaccharide-specific O-acetyltransferases (Fig. 2A), which is analogous to bacterial cell wall polysaccharide acetylation.

Materials and Methods

Production of recombinant proteins in HEK293 cells

The expression constructs for production of recombinant proteins of XOAT1/ESK1, MOAT4 and XGOAT1/AXY4 in the mammalian HEK293 cells were generated as previously described (Zhong et al. 2017a, Zhong et al. 2018c, Zhong et al. 2018d). They were transfected into HEK293 cells using the Invitrogen (Waltham, MA, USA) FreeStyle 293 Expression System according to the manufacturer’s protocol. The transfected cells were cultured for 5 d and the culture media containing the secreted recombinant proteins were collected and passed through nickel resin columns for purification of His-tagged recombinant proteins. The purified proteins were verified by SDS polyacrylamide gel electrophoresis and Coomassie Blue staining.

Assay of acetyltransferase activities

Acetyltransferase activities were assayed by incubating the purified recombinant proteins with various acetyl donor substrates (acetyl-CoA, pNP-Ac, 4MU-Ac and Ac-SA) and the respective oligosaccharides (xylohexaose, mannohexaose and xyloglucan oligomers) as acceptors. The recombinant proteins (100 μg) were incubated in a reaction mixture containing 50 mM HEPES buffer (pH 7.0), acetyl donor substrate (1 mM) and oligosaccharides (200 μg) for 20 h at 37°C. After incubation, the reaction mixture was passed through Dowex 1X4 anion exchange resin and then subjected to MALDI-TOF MS and NMR analyses. For each recombinant protein, the reaction products from three biological replicates were analyzed. Xylohexaose and mannohexaose were purchased from Megazyme (Bray, Co. Wicklow, Ireland) and the xyloglucan oligomers were generated by endo-1,4-β-glucanase digestion of Arabidopsis xyloglucan as described previously (Zhong et al. 2017b). In brief, alcohol-insoluble Arabidopsis stem cell walls were incubated with endo-1,4-β-glucanase (E-CELTR; Megazyme) and the released xyloglucan fragments were purified by passing through a Sephadex G25 size exclusion column and then pooled and lyophilized.

Acetylesterase activity assay

The acetylesterase activities of recombinant XOAT1/ESK1, MOAT4 and XGOAT1/AXY4 were assayed using pNP-Ac and 4MU-Ac. The reaction mixture contained 50 mM sodium phosphate buffer (pH7.0) and 20 μg of recombinant protein. The reactions were initiated by addition of 0.5 mM pNP-Ac or 4MU-Ac and incubated at room temperature. The amount of products released [p-nitrophenol (pNP) from pNP-Ac and 4-methylumbelliferone (4MU) from 4MU-Ac] was monitored in real time for a duration of 15 min. pNP was measured spectrophotometrically at 405 nm and 4MU was measured for its fluorescence signals with an excitation of 365 nm and an emission of 450 nm (Weadge and Clarke 2006). Recombinant GXM1 (glucuronoxylan 4-O-methyltransferase1; Lee et al. 2012) was used as a control protein and the spontaneous hydrolysis of pNP-Ac and 4MU-Ac in the absence of an enzyme was also measured.

Generation of ACLA- and ACLB-RNAi lines

For generation of the ACL-RNAi constructs, the first 500-bp sequence starting from the start codon of each of the ACL cDNAs was utilized. As there are three ACLA genes and two ACLB genes in Arabidopsis (Fatland et al. 2005), we generated RNAi constructs to simultaneously inhibit the expression of all three ACLA genes (ACLA-1, ACLA-2 and ACLA-3; the construct was named ACLA-RNAi) or both ACLB genes (ACLB-1 and ACLB-2; the construct was named ACLB-RNAi). The corresponding DNA fragment containing the three ACLA sequences in tandem or the two ACLB sequences in tandem was ligated in opposite orientations on each side of the GUS gene in the binary vector pBI121 (Clontech, Mountain View, CA, USA). The inverted repeat expression cassette thus created had the GUS sequence as a spacer and was driven by the cauliflower mosaic virus 35S promoter. The ACLA- and ACLB-RNAi constructs were introduced into wild-type Arabidopsis by the Agrobacterium-mediated transformation and the first generation of transgenic plants were analyzed. The transgenic plants exhibited phenotypes ranging from mild to severe retardation of plant growth as previously reported (Fatland et al. 2005). Transgenic plants showing a moderate phenotype with bolting of inflorescence stems were chosen for further analysis. Stems from a pool of 80 independent transgenic plants were used as a biological replicate and three biological replicates were analyzed for each construct.

Gene expression analysis

Interfascicular fibers and pith cells were laser-dissected from elongating stems of 6-week-old Arabidopsis plants as previously described (Zhong et al. 2006). Transgenic Arabidopsis plants overexpressing SND1, VND7 or MYB46 were generated as previously described (Zhong et al. 2006, Zhong et al. 2007, Yamaguchi et al. 2010) and their rosette leaves showing a curly leaf phenotype were collected for RNA isolation. Inflorescence stems of 6-week-old wild-type Arabidopsis plants and 8-week-old transgenic ACLA- and ACLB-RNAi plants were used for RNA isolation. Total RNA was isolated using the Qiagen (Hilden, Germany) RNA isolation kit. After treatment with DNase, the RNA was converted into first-strand cDNA and used for real-time quantitative PCR analysis of gene expression with the QuantiTect SYBR Green PCR kit (Clontech). The expression level of each gene was calculated by normalizing its PCR threshold cycle number with that of the EF1α and UBQ10 reference genes. Three separate pools of isolated cells, leaves or stems for each sample were used as biological replicates for RNA isolation and subsequent real-time quantitative PCR analysis.

Histology

The bottom parts of inflorescence stems of 7-week-old wild type and 10-week-old transgenic ACL-RNAi plants were fixed and embedded in LR White resin as previously described (Burk et al. 2006). Stem sections with 1-μm thickness were cut with a microtome and stained with toluidine blue for anatomy. For transmission electron microscopy, stem sections with 85-nm thickness were cut, post-stained with lead citrate and uranyl acetate, and examined under a transmission electron microscope. Stem sections from at least five transgenic plants were examined and representative images were shown.

Cell wall isolation

Stems from wild-type Arabidopsis and transgenic plants were ground into powder in liquid N₂ and then extracted for alcohol-insoluble cell walls by homogenizing sequentially in 70% ethanol, 100% ethanol and 100% acetone (Zhong et al. 2005). The alcohol-insoluble cell walls were first incubated with 0.2 M ammonium oxalate to solubilize pectin and then extracted with DMSO (Teleman et al. 2002). The DMSO extracts were digested with endo-1,4-β-xylanase M6 (rumen microorganism; Megazyme), endo-1,4-β-mannanase (Bacillus circulans; Megazyme) or xyloglucanase-specific endo-1,4-β-glucanase (GH5; Megazyme) to generate xylosyl, mannosyl and xyloglucan oligomers, respectively, for structural analysis. For each sample, three separate pools of stems were used as biological replicates for cell walls isolation.

Glycosyl composition analysis

Glycosyl composition analysis was performed at the Complex Carbohydrate Research Center by combined gas chromatography/mass spectrometry (GC/MS) of the per-O-trimethylsilyl (TMS) derivatives of the monosaccharide methyl glycosides produced from the sample by acidic methanolysis as described previously by Weiss et al. (2012). In brief, the samples were heated with 500 μl of 1 M methanolic HCl (Sigma, St. Louis, MO, USA) in a sealed screw-top glass test tube for 18 h at 80°C. After cooling and removal of the solvent under a stream of nitrogen, methanol was added and the samples were again dried under nitrogen. This step was repeated twice more. The samples were derivatized with 200 μl Tri-Sil^® (Pierce) at 80°C for 30 min. The samples were then gently dried under nitrogen. The dried samples were dissolved in 2 ml hexane and centrifuged using a low speed desktop centrifuge (Fisher Scientific centrific model 228) and the supernatant was transferred to a clean test tube and dried down. The samples were resuspended in 100 μl hexane and transferred into Thermo Scientific GC vials containing a 300 μl target insert (Thermo Fisher Scientific, Waltham, MA, USA). GC/MS analysis of the TMS methyl glycosides was performed on an Agilent 7890A GC interfaced to a 5975C MSD, using a Supelco Equity-1 fused silica capillary column (30 m × 0.25 mm ID).

Quantitative measurement of cell wall acetyl content

The acetyl content in cell walls was determined by quantifying the amount of acetic acid released from cell walls upon treatment with NaOH according to Teleman et al. (2002). The alcohol-insoluble cell walls were first treated with NaOH and then neutralized with HCl, and the amount of acetic acid liberated in the extract was determined using an acetic acid assay kit (Megazyme).

Mass spectrometry

The xylosyl and mannosyl oligomers generated by enzymatic digestion of the DMSO extracts of cell walls were analyzed by MALDI-TOF MS using a Bruker Autoflex TOF mass spectrometer (Billerica, MA, USA) in reflection mode (Zhong et al. 2005). An average of 100 laser shots was collected for each spectrum.

¹H-NMR spectroscopy

Cell wall materials were analyzed for the degree of acetylation with a Varian Inova 500 MHz spectrometer (Varian Inc., Palo Alto, CA, USA). The NMR spectra were recorded with 512 transients. The resonances corresponding to the carbohydrate and acetyl regions in the NMR spectra were assigned based on the previously reported NMR spectra data for xylan (Teleman et al. 2002), glucomannans (Teleman et al. 2003), xyloglucan (York et al. 1988) and pectin (Tamaki et al. 2008).

Isolation of Golgi-enriched membrane vesicles and acetyl-CoA transport assay

Golgi-enriched membrane vesicles were isolated from stems of 6-week-old wild-type plants and the rwa1/2/3/4 quadruple mutant (Lee et al. 2011) by discontinuous density gradient centrifugation following the procedure by Leelavathi et al. (1970) and resuspended in 50 mM HEPES-KOH buffer (pH 7.0) and 250 mM sucrose. For acetyl-CoA transport assay, Golgi-enriched vesicles (containing 100 μg of proteins) were incubated in the reaction mixture (50 μl) with 50 mM HEPES-KOH (pH 7.0), 250 mM sucrose and [acetyl-1-¹⁴C]-CoA (0.1 μCi; American Radiolabeled Chemical). After incubation, the reaction mixture was passed through Dowex 1X4 anion exchange resin to remove free [acetyl-1-¹⁴C]-CoA and the membrane vesicles containing imported [acetyl-1-¹⁴C]-CoA were counted for the amount of radioactivity with a PerkinElmer scintillation counter (Waltham, MA, USA). Three biological replicates were performed for each sample.

Assay of xylan xylosyltransferase and methyltransferase activities

Xylan xylosyltransferase activity in Golgi-enriched membrane vesicles was assayed using UDP-[¹⁴C]Xyl as the xylosyl donor and xylohexaose as the acceptor according to Kuroyama and Tsumuraya (2001). The reaction products were separated from UDP-[¹⁴C]Xyl by paper chromatography according to Ishikawa et al. (2000) and counted for the amount of radioactivity with a PerkinElmer scintillation counter. Glucuronoxylan methyltransferase activity was assayed using S-[¹⁴C-methyl] adenosylmethionine as the methyl donor and GlcA-substituted xylotetraose as the acceptor according to Lee et al. (2012). The radiolabeled reaction products were separated from S-[¹⁴C-methyl] adenosylmethionine by Dowex 1X4 anion exchange chromatography and counted for the amount of radioactivity with a PerkinElmer scintillation counter. The assays were performed in three biological replicates.

Funding

The U.S. Department of Energy, Office of Science, Office of Basic Energy Sciences [grant no. DE-FG02-03ER15415]. Glycosyl composition analysis was supported in part by the Chemical Sciences, Geosciences and Biosciences Division, Office of Basic Energy Sciences, U.S. Department of Energy grant [DE-SC0015662] to P.A. at the Complex Carbohydrate Research Center.

Disclosures

The authors have no conflicts of interest to declare.