A Glucuronosyltransferase from Arabidopsis Thaliana Involved In

The Plant Journal (2013) 76, 1016–1029 doi: 10.1111/tpj.12353 A b–glucuronosyltransferase from Arabidopsis thaliana involved in biosynthesis of type II arabinogalactan has a role in cell elongation during seedling growth

Eva Knoch1,†, Adiphol Dilokpimol1,†, Theodora Tryfona2, Christian P. Poulsen1, Guangyan Xiong3, Jesper Harholt1, Bent L. Petersen1, Peter Ulvskov1, Masood Z. Hadi4, Toshihisa Kotake5, Yoichi Tsumuraya5, Markus Pauly3, Paul Dupree2 and Naomi Geshi1,* 1Department of Plant and Environmental Sciences, Faculty of Science, University of Copenhagen, Thorvaldsensvej 40, Frederiksberg C 1871, Denmark, 2Department of Biochemistry, University of Cambridge Cambridge CB2 1QW, UK, 3Energy Biosciences Building 212C, 2151 Berkeley Way, Berkeley, CA 94720-5230, USA, 4Joint BioEnergy Institute, Lawrence Berkeley National Laboratory, 717 Potter Street, Berkeley CA 94720, USA, and 5Division of Life Science, Graduate School of Science and Engineering, Saitama University, 255 Shimo–okubo, Sakura–ku, Saitama 338–8570, Japan

Received 11 September 2013; revised 30 September 2013; accepted 8 October 2013; published online 15 October 2013. *For correspondence (e-mail [email protected]). †These authors contributed equally to this work.

SUMMARY We have characterized a b–glucuronosyltransferase (AtGlcAT14A) from Arabidopsis thaliana that is involved in the biosynthesis of type II arabinogalactan (AG). This enzyme belongs to the Carbohydrate Active Enzyme database glycosyltransferase family 14 (GT14). The protein was localized to the Golgi apparatus when transiently expressed in Nicotiana benthamiana. The soluble catalytic domain expressed in Pichia pastoris transferred glucuronic acid (GlcA) to b–1,6–galactooligosaccharides with degrees of polymerization (DP) ranging from 3–11, and to b–1,3–galactooligosaccharides of DP5 and 7, indicating that the enzyme is a glucuronosyltransferase that modiﬁes both the b–1,6- and b–1,3-galactan present in type II AG. Two allelic T–DNA insertion mutant lines showed 20–35% enhanced cell elongation during seedling growth compared to wild-type. Analyses of AG isolated from the mutants revealed a reduction of GlcA substitution on Gal–b– 1,6–Gal and b–1,3–Gal, indicating an in vivo role of AtGlcAT14A in synthesis of those structures in type II AG. Moreover, a relative increase in the levels of 3-, 6- and 3,6-linked galactose (Gal) and reduced levels of 3-, 2- and 2,5-linked arabinose (Ara) were seen, suggesting that the mutation in AtGlcAT14A results in a relative increase of the longer and branched b–1,3- and b–1,6-galactans. This increase of galactosylation in the mutants is most likely caused by increased availability of the O6 position of Gal, which is a shared acceptor site for AtGlcAT14A and galactosyltransferases in synthesis of type II AG, and thus addition of GlcA may terminate Gal chain extension. We discuss a role for the glucuronosyltransferase in the biosynthesis of type II AG, with a biological role during seedling growth.

Keywords: glycosyltransferase family 14, glucuronosyltransferase, arabinogalactan protein, type II arabinogalactan, plant cell wall, Golgi apparatus, Arabidopsis thaliana.

INTRODUCTION

The arabinogalactan proteins (AGPs, AG proteins) belong processes has been reported, including somatic embryo- to a highly diverse class of glycoproteins present on cell genesis, cell–cell interactions and cell elongation (Seifert surfaces of plants (Seifert and Roberts, 2007; Ellis et al., and Roberts, 2007). Most of these studies involved use of 2010; Tan et al., 2012). AGPs consist mainly of glycans monoclonal antibodies raised against the AG polysaccha- (>90% w/w), and are synthesized by post-translational rides, and a temporal and spatial appearance of speciﬁc modiﬁcation in the secretory pathway. The importance of AG epitopes during development has been reported. the carbohydrate moieties of AGPs in various cellular However, as the precise epitope structure for most of the

1016 © 2013 The Authors The Plant Journal © 2013 John Wiley & Sons Ltd This is an open access article under the terms of the Creative Commons Attribution-NonCommercial-NoDerivs License, which permits use and distribution in any medium, provided the original work is properly cited, the use is non-commercial and no modiﬁcations or adaptations are made. b–glucuronosyltransferase involved in AGP biosynthesis 1017 antibodies is not known, the molecular function of the AG an extensive reduction in Ara moieties in beta-glucosyl glycan structures remains to be determined. The appear- Yariv-precipited AGPs (Gille et al., 2013). This alteration in ance of different AG glycans is probably controlled by a arabinosylation leads to shortened primary roots, but glycosylation process rather than redundancy of protein lateral root growth is not affected. cores, as a developmentally controlled pattern of glycosyl- We have characterized an Arabidopsis GT encoded by ation was observed on a single synthetic peptide At5g39990 that belongs to the GT14 family in the Carbo- expressed in Arabidopsis (Estevez et al., 2006). Therefore, hydrate Active Enzyme database (CAZy, www.cazy.org; elucidation of the enzymes involved in the biosynthesis of Cantarel et al., 2009). The GT14 family contains several AG glycans is expected to facilitate our understanding of mammalian GTs involved in protein glycosylation, e.g. the function of AGPs. b–1,6-N–acetyl glucosaminyltransferases catalyzing b–1,6- The structure of AG polysaccharides is very heteroge- linked N–acetylglucosaminylation in core 2 or I–branched neous even on a single peptide (Estevez et al., 2006; Tan O–glycosylation and protein O–b-xylosyltransferases et al., 2010), but commonly consists of a b–1,3-galactan (Bierhuizen et al., 1993; Yu et al., 2001; Wilson, 2002). In backbone with substitution at the O6 position with b–1,6- contrast, none of the putative plant GT14s [e.g. 11 from galactan side chains (type II AG, Tan et al., 2012); a model Arabidopsis, 12 from rice (Oryza sativa)] have been char- structure is shown in Figure 3(a). The side chains are acterized. Plant GT14s are phylogenetically related to the typically further substituted by arabinose (Ara) and less protein family containing Domain of Unknown Function frequently with other sugars such as glucuronic acid or 266 (DUF266; Ye et al., 2011), and a mutation in a 4–O–methyl glucuronic acid (collectively referred to as DUF266 protein in rice (brittle culm 10, BC10) caused a GlcA), rhamnose (Rha) and fucose (Fuc) (Tsumuraya et al., severe alteration in the mechanical strength of the stem 1988; Tan et al., 2010; Tryfona et al., 2010, 2012). The gly- and the AG quantity and structure (Zhou et al., 2009). cosylation of AGPs is catalyzed by glycosyltransferases The authors concluded that BC10 is probably a GT, but (GTs) that are located mainly in the Golgi apparatus. GTs involvement in the AG glycosylation pathway is not act in a regio- and stereo-speciﬁc manner (Lairson et al., clear. 2008), and it is expected that at least ten functionally dis- In this paper, we provide evidence for GlcAT activity of tinct GTs are required for the biosynthesis of type II AG. At5g39990 and its role in the biosynthesis of type II AG So far, two fucosyltransferases (AtFUT4 and AtFUT6; Wu structures and cell elongation during seedling growth. et al., 2010), two galactosyltransferases (AtGALT2; Basu et al., 2013; AtGALT31A; Geshi et al., 2013) and a putative RESULTS arabinosyltransferase (AtRAY1; Gille et al., 2013) from CAZy family GT14 Arabidopsis have been reported in the AG glycosylation pathway. AG fucosyltransferase activity was demonstrated Little is known about plant GTs in CAZy family GT14. As by gain of function of fucosylated AGs by heterologous plants do not have the same type of glycoconjugates pro- expression of Arabidopsis AtFUT4 and AtFUT6 in tobacco duced by mammalian GT14 enzymes (Bierhuizen et al., BY2 cells (Wu et al., 2010). Fucose on AGPs is important 1993; Yu et al., 2001; Wilson, 2002) and the plant GT14 for root development (Van Hengel and Roberts, 2002), but enzymes are phylogenetically distantly related to their a role for AtFUT4 and AtFUT6 in vivo remains to be eluci- mammalian counterparts (Figure 1), a distinct activity is dated. Galactosyltransferase (GalT) activity towards expected for the plant enzymes. We hypothesized that hydroxyproline in the synthetic peptide was demonstrated plant GT14 enzymes may be involved in the AG glycosyla- for Arabidopsis AtGALT2 expressed in Pichia pastoris tion pathway for three reasons: (i) plant GT14 members (Basu et al., 2013). The atgalt2 mutants demonstrated are related to DUF266 (Ye et al., 2011), and mutation in lower GalT activity and a reduced level of b–galactosyl one of the DUF266 proteins in rice (BC10) caused severe Yariv-precipited AGPs, but no apparent morphological phe- alteration in AG quantity and structure (Zhou et al., 2009), notype was reported (Basu et al., 2013). The GalT activity (ii) some GT14s are co-expressed with genes encoding the in elongation of the b1,6-galactan side chains of AG was protein backbone of AGPs (Showalter et al., 2010), and (iii) demonstrated by Arabidopsis AtGALT31A expressed in some GT14 genes are co-expressed with AtGALT31A Escherichia coli and Nicotiana benthamiana (Geshi et al., (At1g32930), whose heterologously expressed protein 2013). A mutation in AtGALT31A caused aberrant asym- demonstrated GT activity in elongation of the b–1,6-galac- metric formative divisions in the hypophysis during tan side chains of AG (Geshi et al., 2013). An Arabidopsis embryogenesis, and embryo development was arrested at gene, At5g39990 (indicated by an asterisk in Figure 1), is the globular stage, indicating an essential role for AG co-expressed with AtGALT31A during stem elongation glycan in the normal development of embryo (Geshi et al., (Figure S1), thus we selected the protein encoded by this 2013). Although GT activity of RAY1 has not yet been gene for further characterization regarding its involvement demonstrated, mutations in this GT family 77 gene led to in AG biosynthesis.

Figure 1. Phylogenetic relationship among selected plant proteins and characterized GT14 proteins. The tree includes plant sequences from Arabidopsis thaliana (At), Oryza sativa japonica group (Os/OJ), Triticum aestivum (Ta) and Populus tremula. The phylog- eny analysis was performed and the phylogenetic tree was created using MEGA5 software (Tamura et al., 2011). The asterisk indicates At5g39990 (AtGlcAT14A) described in this paper. The sequence relationship within Arabidopsis GT14 enzymes is shown in Table S1.

Recombinant expression of the catalytic domain encoded AtGlcAT14A also lacks the DXD motif and is fully active in by At5g39990 in Pichia pastoris the absence of Mn2+ or Mg2+, while high concentrations of 2+ 2+ Mn (>1mM)orMg (>10 mM) had an inhibitory effect For biochemical characterization, the catalytic domain of (Figure 2d). EDTA at >5mM concentrations also inhibited At5g39990 was expressed as a soluble secreted construct the enzyme activity, thus other divalent metal ions may be with an N–terminal FLAG tag in Pichia pastoris. The recom- involved in the enzyme activity. Apparently, family GT14 binant protein was purified by immunoprecipitation and enzymes including AtGlcAT14A have a different mode of analyzed by SDS–PAGE (Figure 2a). The purified protein substrate binding than most of the other GTs, which bind appeared as a smear, with the lowest apparent molecular the UDP moiety of the substrate by the DXD motif via Mn2+ size corresponding to 52 kDa, which was slightly higher or Mg2+ ions. than the expected size (47101.4 Da; Figure 2a, lane 8). Treatment with endoglycosidase H resulted in a shift of Characterization of GlcA incorporation sites for the protein size from 52 to 48 kDa, suggesting that the AtGlcAT14A recombinant protein was N–glycosylated (Figure 2a, lanes Next, we investigated the site of GlcA incorporation in 9 and 10). the AG acceptors. b–GlcA residues are present at the O6 position of Gal in b–1,6-galactan side chains as well as in Characterization of donor substrate specificity of the the b–1,3-galactan main chain of type II AG (Tryfona recombinant protein et al., 2010, 2012; Nie et al., 2013), thus we tested b–1,6- First, we tried to identify the donor substrate for the and b–1,3-galactooligosaccharides of various lengths recombinant protein by testing the GT activity towards labeled with 2–aminobenzoic acid (2AA) and analyzed the seven different NDP-[14C]-sugars. We used microsomes products using normal-phase HPLC (Figure 3). Using an prepared from N. benthamiana after expression of a syn- acceptor mixture of b–1,6-galactooligosaccharides of DP3– thetic peptide encoding a glycomodule for AG glycosyla- 11, incubation with empty pPICZaA vector (Figure 3a,e,g, tion as acceptor for the assay (Figure 2b; GAGP8–GFP; Xu i) and AtRGXT1 (Figure 3b) controls did not result any et al., 2005). Structural characterization of this GAGP8–GFP changes, and the acceptor mixture was detected as it is acceptor indicated the presence of various type II AG poly- (indicated by black arrowheads in Figure 3a,b). In con- saccharides (Geshi et al., 2013), and thus an appropriate trast, incubation with AtGlcAT14A resulted in additional mixture of AG acceptors for various AG GTs. We assessed peaks detected between b–1,6-galactooligosaccharides 14 the transfer of [ C]-sugar to the GAGP8–GFP acceptor by acceptors (indicated by white arrowheads and marked as immunoprecipitation of GAGP8–GFP and scintillation GP in Figure 3c). Further treatment using exo-b–glucuron- counting. Pichia culture broth containing the secreted idase reduced the level of those additional peaks and At5g39990 protein (Figure 2a, lane 6) showed GT activity increased the level of unmodified 2AA-labeled b–1,6-ga- specifically towards UDP-[14C]-glucuronic acid (UDP-[14C]- lactooligosaccharide peaks (compare black and white GlcA) (Figure 2b), but no activity was observed for the arrowheads in Figure 3d). The results indicate that Pichia culture broth harboring an empty pPICZaA vector as AtGlcAT14A transfers GlcA to 2AA-labeled b–1,6-galac- a control. This result indicates that the At5g39990 catalytic tooligosaccharides of DP3–11 via a b–linkage. Likewise, domain encodes a glucuronosyltransferase (GlcAT). We when 2AA-labeled b–1,3-galactooligosaccharide acceptors confirmed this finding using purified recombinant protein of DP3, 5 and 7 were tested, incubation with AtGlcAT14A (Figure 2a, lane 8), which showed the same activity, i.e. resulted in no modification onto b–1,3-galactooligosac- 14 transfer of [ C]-GlcA onto GAGP8–GFP acceptor (Fig- charide of DP3 (Figure 3f), but an additional peak ure 2c), while the empty pPICZaA vector and Arabidopsis appeared when b–1,3-galactooligosaccharides of DP5 and a–1,3-D–xylosyltransferase, which is involved in biosynthe- 7 were used as acceptors (Figure 3h,j). This result sis of pectic rhamnogalacturonan II (AtRGXT1, Figure S2; indicates that AtGlcAT14A transfers GlcA to 2AA-labeled Petersen et al., 2009) subjected to the same purification b–1,3-galactooligosaccharides of longer than DP5. Taken procedure did not show such activity (Figure 2c). Thus, we together, AtGlcAT14A is probably responsible for adding confirmed that At5g39990 encodes a GlcAT involved in AG GlcA to both the b–1,6-galactan side chain and the b–1,3- glycosylation, and we named this protein AtGlcAT14A galactan main chain of type II AGs (sites of action indi- (Arabidopsis thaliana GlcA transferase from family GT14). cated in Figure 4). Many GTs require divalent metal ions, commonly Mn2+ Subcellular localization of AtGlcAT14A or Mg2+, for catalytic activity (Lairson et al., 2008). How- ever, the mammalian GT14 enzymes do not have such a To investigate the subcellular localization of AtGlcAT14A, requirement, because their sequences lack the metal ion full-length AtGlcAT14A conjugated with monomeric CFP binding DXD motif (Williams et al., 1980; Bierhuizen and (AtGlcAT14A–mCer3, Figure 5a) and STtmd–YFP (a Golgi Fukuda, 1992; Yeh et al., 1999; Schwientek et al., 2000). marker comprising the sialyltransferase short cytoplasmic

(a)

(b) (c)

(d)

Figure 2. Enzyme activity of recombinant At5g39990 (AtGlcAT14A). (a) SDS–PAGE analysis of the recombinant At5g39990 catalytic domain expressed in Pichia pastoris; lane M, markers. Lanes 1–6, 10-fold concentrated Pichia culture broth (12 ll each) expressing empty pPICZaA vector (2.8 lg total protein, lanes 1 and 4), RGXT1 (6.8 lg total protein, lanes 2 and 5) or At5g39990 (3.9 lg total protein, lanes 3 and 6); proteins were stained using Coomassie brilliant blue (lanes 1–3) or analyzed by Western blotting (lanes 4–6). Lanes 7 and 8, Western blot of affinity- purified samples (10 ll of bead slurry each) for RGXT1 (114 ng protein, lane 7) or At5g39990 (77 ng protein, lane 8). Lanes 9 and 10, deglycosylation of the purified At5g39990 by endoglycosidase H: untreated samples (lane 9) and treated samples (lane 10). 14 (b) Donor substrate identification for recombinant At5g39990. Seven NDP-[ C]-sugars were tested for GT activity on GAGP8–GFP acceptor using tenfold concentrated Pichia culture broth expressing the empty pPICZaA vector [black bars, lane 4 in (a)] or At5g39990 [gray bars, lane 6 in (a)]. Incorporation of [14C]-sugar onto the acceptor was analyzed by immunoprecipitation using anti-GFP antibody followed by scintillation counting (n = 3). At5g39990 encodes a GlcAT that is involved in AG glycosylation, and was thus named AtGlcAT14A (Arabidopsis thaliana GlcA transferase from family GT14). (c) Glucuronosyltransferase activity using purified At5g39990 [38.5 ng protein, gray bar, lane 8 in (a)] in comparison with the empty pPICZaA vector (black bar) and purified RGXT1 [white bar, lane 7 in (a)]. The assay method was the same as described in (b) (n = 3). Purified RGXT1 is active as a xylosyltransferase (Figure S2). (d) Effect of divalent metal ions and EDTA on the catalytic activity of purified AtGlcAT14A. The assay method was the same as described in (b) (n = 2). tail and single transmembrane domain fused to YFP, fluorescence was detected in punctate vesicles that Figure 5b; Boevink et al., 1998) were transiently co- co-localized with STtmd–YFP (Figure 5c), indicating locali- expressed in N. benthamiana leaves. AtGlcAT14A–mCer3 zation of AtGlcAT14A in the Golgi apparatus.

(e)

(a)

(f)

(b)

(g)

(i) (d)

(j)

Figure 3. Glucuronosyltransferase activity using b-1,6- or b-1,3-galactooligosaccharides (b-1,6-Galmix-2AA or b-1,3-Galn-2AA) as acceptor. (a–d) Assay of GlcAT activity towards b-1,6-Galmix-2AA (DP3–11) performed using afﬁnity-puriﬁed materials from empty pPICZaA vector (a), RGXT1 (b) and At- GlcAT14A (c), and the reaction mixture of (c) treated with b-glucuronidase (d). For comparison, the chromatogram from (c) is shown as a gray line in (d).

(e–j) Assay using afﬁnity-puriﬁed materials from empty pPICZaA vector (e,g,i) and AtGlcAT14A (f,h, j) of GlcAT activity towards b-1,3-Gal3-2AA (e,f), b-1,3-Gal5- 2AA (g,h) and b-1,3-Gal7-2AA (i,j). DP, degree of polymerization; GP, glycosylated product; black arrowheads indicate b-1,6-Galmix-2AA or b-1,3-Galn-2AA used as acceptor; white arrowheads indicate glycosylated product on b-1,6-Galmix-2AA or b-1,3-Galn-2AA.

In silico transcriptomic database analysis using Gene- According to transcriptomics analysis using GeneCAT CAT (Mutwil et al., 2008) indicated that AtGlcAT14A is (http://genecat.mpg.de; Mutwil et al., 2008), Genevestigator co-expressed with AtGALT31A, an enzyme that is involved (http://www.genevestigator.com/gv/) and the Arabidopsis in the synthesis of b–1,6-galactan side chains in type II AGs eFP browser (Winter et al., 2007), AtGlcAT14A is highly (AtGALT31A; Figure S1). To determine whether the AtGl- expressed in root, the shoot apex and the shoot apex inflo- cAT14A and AtGALT31A proteins interact, we investigated rescence (Figure S1). Hence, we first investigated roots and plausible protein–protein interaction using the acceptor hypocotyls of the mutant seedlings grown in the dark. photobleaching FRET technique (Poulsen et al., 2013), but Interestingly, 5-day-old seedlings of atglcat14a–1 and -2 the result did not indicate molecular interactions between showed significantly increased elongation rates compared the two proteins (Figure S3). to wild-type in both roots (30 and 18%, respectively) and hypocotyls (33 and 23%, respectively) (Figure 7a). The Characterization of T–DNA insertion lines of AtGlcAT14A monosaccharide composition of AG extracts from both at- To investigate a role for AtGlcAT14A in vivo, we analyzed glcat14a–1 and -2 did not indicate significant changes in two allelic homozygous T–DNA insertion lines in Arabidop- GlcA content, but showed an approximately 12% increase sis (SALK_064313 and SALK_043905, designated atgl- in Gal and an approximately 12% decrease in Ara cat14a–1 and atglcat14a–2, respectively). The T–DNA compared to wild-type (Figure 7b). We further performed insertions are in the 4th and 2nd exons of atglcat14a–1 and glycosidic linkage analysis for the AG extracts, which atglcat14a–2, respectively (Figure 6a). RT–PCR analysis revealed an increase in 3-, 6- and 3,6-linked Gal and a using two sets of primers (annealing sites on the atgl- decrease in 2-, 3-, 5- and 2,5-linked Ara in the mutant AG cat14a sequence indicated in Figure 6a) indicated the pres- compared to wild-type (Table 1). The 3-, 6- and 3,6-linked ence of a truncated transcript in atglcat14a–1 and a lack of Gal are specific to the type II AG structure, demonstrating transcript in atglcat14a–2 (Figure 6b). little contamination by pectic rhamnogalacturonan I in the

Figure 4. Model structure of type II AG based on AG analyzed from Arabidopsis leaf (modiﬁed from Figure 7 in Tryfona et al., 2012;). Expected cleavage sites by hydrolases used in this study are indicated. The boxed GlcAp indicates the sites of GlcA transfer by AtGlcAT14A based on the in vitro enzyme assay shown in Figure 3.

(a) (b) (c)

Figure 5. Subcellular localization of AtGlcAT14A–mCer3 transiently expressed in N. benthamiana leaves. (a) AtGlcAT14A–mCer3, (b) STtmd–YFP, and (c) overlaid image of (a) and (b). The results indicate co-localization of AtGlcAT14A–mCer3 and STtmd–YFP in the Golgi apparatus. Scale bars = 10 lm.

AG extracts. The structures of 2-, 3- and 5-linked Ara are We anticipated the reduced level of GlcA in the mutants, commonly found as a part of type II AGs (Tan et al., 2004; but the level of GlcA in the isolated AG extracts was low, Tan et al., 2010; Tryfona et al., 2010, 2012), but the 2,5- and we did not detect significant differences by HPAEC- linked Ara detected in this study has not been reported PAD (high-performance anion exchange chromatography and may appear specifically in type II AG during seedling with pulsed amperometric detection) monosaccharide growth. The increase in 3,6- and 6-linkages of Gal in the analysis (Figure 7b) and immune labeling on the root sur- mutant AG may be the result of an increase in chain length face using LM2 antibody (Smallwood et al., 1996) (Figure in both b–1,6- and b–1,3-galactan. The relative decreases S4). Therefore, we performed a detailed analysis of AG detected in multiple linkages for Ara are most likely the side-chain structures from roots growing hydroponically result of the relative increase in Gal linkages, or GlcA may for 3 weeks, by treatment of AG extracts with specific be required for the addition of Ara. These results indicate hydrolases followed by polysaccharide analysis using car- that the mutation in AtGlcAT14A facilitates elongation of bohydrate gel electrophoresis (PACE). We chose mature both b–1,6- and b–1,3-galactans, and initiation of 6–branch- roots for the analysis to obtain enough material and also ing of b–1,3-galactan main chains in type II AG. to minimize the effects of redundant activities, because

(a) their substitutes at the O6 position with GlcA (GlcA-Gal, GlcA-Gal2, GlcA-Gal3) as these structures deﬁned by Tryfona et al. (2012) (Figure 8a). Quantiﬁcation of the oligosaccharides indicated that the ratios of GlcA-Gal/Gal

and GlcA-Gal2/Gal2 were approximately 25 and 30% less, (b) respectively, in atglcat14a-1 and -2, respectively (Figure 8b,

c). There may also be differences in the GlcA-Gal3/Gal3 ratio, but the intensity of the corresponding bands was too low to perform statistical analysis. These results indicate reduced GlcA substitution on the b–1,3-linked Gal in

the main chain and b–1,6-Gal2 in the side chains of the mutants compared to wild-type, and demonstrates the in vivo role of AtGlcAT14A in synthesis of those structures in type II AG (Figure 4). Figure 6. Analysis of AtGlcAT14A T-DNA insertion lines by RT-PCR. (a) Schematic drawing of the AtGlcAT14A (At5g39990) gene structure and predicted sites of T-DNA insertions in two independent lines: atglcat14a-1 DISCUSSION (SALK_064313) and atglcat14a-2 (SALK_043905). Exons are represented by thick gray boxes; introns and non-coding regions are represented by thin A novel b–GlcAT from the GT14 family is involved in gray lines. The expected T-DNA insertion sites for each mutant are indi- biosynthesis of type II AG cated. Black arrowheads indicate primer binding sites. (b) Semi-quantitative RT-PCR to analyze the presence of transcript in the The AG polysaccharides are very complex structures and mutant lines. UBC indicates ubiquitin, which was used as the reference therefore require a large number of GT enzymes for their gene. biosynthesis; however, little is known about the AG glycosylation pathway. Here, we demonstrated that recombi- AtGlcAT14A is highly expressed in roots while its closest nantly expressed AtGlcAT14A possesses GlcAT activity, homolog (At5g15050, 72% amino acid sequence identity, adding GlcA to both b–1,3- and b–1,6-galactooligosaccha- Figure 1) is expressed at a low level. No visible growth rides (Figure 3). Consistent with the enzyme activity shown changes were observed at this stage for the mutants used in vitro, the level of GlcA in the corresponding structures in the analysis. The HPAEC-PAD monosaccharide composi- was reduced in the two allelic T–DNA insertion mutant tion analysis of mature roots did not indicate signiﬁcant lines, indicating an in vivo role for AtGlcAT14A in adding differences between wild-type and atglcat14a–1 and -2 GlcA to both the b–1,3-galactan main chain and the b–1,6- (Figure S5), while PACE using treatments with exo-b–1,3- galactan side chain of type II AG. galactanase (Tsumuraya et al., 1990) followed by a–arabin- GTs that are classiﬁed in the same CAZy family often ofuranosidase (Takata et al., 2010) (cleavage sites indicated share the same retaining or inverting reaction mechanism, in Figure 4) revealed the presence of diverse oligosaccha- and often transfer a donor sugar to the same position on rides co-migrating with Ara, Gal, b–1,6-linked galactan and the acceptor substrate (Amado et al., 1999). Therefore, it is

(b)

(a)

Figure 7. Growth phenotype and monosaccharide composition of wild-type and mutant seedlings. (a) Length of hypocotyls and roots from seedlings grown in the dark and in the right for 5 days, respectively, was compared to the wild-type as 100% (n > 45). Asterisks indicate signiﬁcant differences compared with wild-type (Student’s t test, P < 0.01). (b) Monosaccharide composition (mol%) of AGP extracts from 5-day-old etiolated seedlings analyzed by HPAEC-PAD (n = 6 for wild-type, n = 5 for atglcat14a-1 and -2). X/M, mixture of Xyl and Man. Asterisks indicate signiﬁcant differences compared with wild-type (Student’s t test, P < 0.01).

Table 1 Glycosidic linkage analysis (% total peak area) for AG extracts from wild-type and atglcat14a mutant plants

T–Araf T–Fucp 2–Araf 3–Araf T–Galp 5–Araf 3–Galp 2,5–Araf 6–Galp 3,6–Galp Total

Wild-type 9.0 Æ 1.2 0.7 Æ 0.1 11.1 Æ 0.6 5.9 Æ 0.4 5.0 Æ 0.3 6.0 Æ 0.1 5.3 Æ 0.3 9.2 Æ 0.9 9.4 Æ 0.5 19.4 Æ 0.5 80.9 atglcat14a 9.3 Æ 0.5 0.8 Æ 0.1 7.2 Æ 0.1 3.5 Æ 0.1 5.7 Æ 0.3 5.1 Æ 0.1 8.3 Æ 0.1 5.4 Æ 0.1 14.1 Æ 0.3 24.1 Æ 0.3 83.5

Values are means Æ standard deviation. Values that are signiﬁcantly different between wild-type and atglcat14a shown in bold (P < 0.01, Student’s t test, n = 4). T-Araf: terminal arabinofuranose, T-Fucp: terminal fucopyranose, 2-Araf: 2-linked arabinofuranose, 3-Araf: 3-linked arabinofuranose, T-Galp: terminal galactopyranose, 5-Araf: 5-linked arabinofuranose, 3-Galp: 3-linked galactopyranose, 2,5-Araf: 2,5-linked arabinofuranose. 6-Galp: 6-linked galactopyranose, 3,6-Galp: 3,6-linked galactopyranose.

(a) (b)

(c)

Figure 8. Structural analysis of AG side chains. (a) AG side chains were released from wild-type, atglcat14a-1 and atglcat14a-2 AG extracts of mature roots by treatment with speciﬁc exo-b-1,3-galactanase followed by a-arabinofuranosidase (cleavage sites indicated in Figure 4), and separated by PACE. The migration positions of Ara, Gal, GlcA-(1?6)-b-Gal, GlcA-(1?

6)-b-Gal2, GlcA-(1?6)-b-Gal3 and b-1,6-galactooligosaccharides of DP 2–8 are shown. (b,c) Ratio of GlcA-substituted versus non-substituted Gal (GlcAGal/Gal) (b) and b-1,6-galactobiose (GlcAGal2/Gal2) (c) based on quantified intensity of corresponding spots in (a) (the positions indicated by arrows). Values are means Æ SD of three biological replicates. Asterisks indicate significant differences compared with wild-type (Student’s t test, P < 0.05). reasonable that AtGlcAT14A (b–1,6-GlcAT) and mammalian AtGlcAT14A required neither Mn2+ nor Mg2+ for its cata- b–1,6-GlcNAc transferases belong to the same CAZy family lytic activity (Figure 2d), which is also observed for several (GT14) as both enzymes catalyze an inverting reaction and mammalian GT14 enzymes (Williams et al., 1980; Bierhui- transfer a sugar residue onto the O6 position of the zen and Fukuda, 1992; Yeh et al., 1999; Schwientek et al., acceptor sugar. In plants, GUX from family GT8 has been 2000), and thus may be a common feature in GT14 family characterized biochemically as a GlcAT that adds GlcA to a members. They all lack the DXD motif that is present in b–1,4-xylan backbone via a–1,2 linkage (Mortimer et al., most other GT families and binds the nucleotide diphos- 2010; Lee et al., 2012; Rennie et al., 2012). As the GUX phate of the substrate via a divalent metal ion. The crystal enzyme has a retaining activity, AtGlcAT14A described in structure of a GT14 family member, murine b–1,6-GlcNAc this paper is the first example of a GlcAT with inverting transferase (C2GnT–L), indicated that R378 and K401 play a activity in plants. role in electrostatically stabilizing the nucleotide

© 2013 The Authors The Plant Journal © 2013 John Wiley & Sons Ltd, The Plant Journal, (2013), 76, 1016–1029 b–glucuronosyltransferase involved in AGP biosynthesis 1025 diphosphate leaving group instead of divalent metal ions 2008). Together with our results, these results indicate a (Pak et al., 2006). Plant GT14 sequences including AtGl- role for GlcAT(s) and GlcA hydrolase(s) towards type II AG cAT14A appear to lack the corresponding residue for R378 in cell elongation. However, the cell elongation phenotype (Figure S6), thus plant GT14 enzymes appear to interact was not directly related to a change in the particular chem- with the donor substrate in a slightly different way than ical composition of type II AG; we observed an increase in mammalian GT14 enzymes. Elucidation of the crystal struc- Gal and a decrease in Ara in the AG extracts of atglcat14a ture of AtGlcAT14A together with the substrate will provide mutants due to altering AG glycosylation in the Golgi dur- further insights regarding evolution of GT14 enzymes ing biosynthesis, while Eudes et al. (2008) observed a between mammals and plants. decrease of both Gal and Ara in the AG extracts of plants over-expressing AtGUS2, which most likely does not inter- Physiological role of AtGlcAT14A and GlcA substitution on fere with the biosynthesis of AGs but affects AG modifica- AGPs in vivo tion after deposition in the apoplast. However, both types We analyzed two independent T–DNA insertion mutant of mutant plants exhibited increased cell elongation in lines, and the results suggested that AtGlcAT14A has seedlings. In both cases, the changes in GlcA level were significant influence on cell elongation during seedling very minor, and Eudes et al. (2008) further reported a sub- growth (Figure 7a). In both AtGlcAT14A mutant lines, we stantial increase in xylose in AG extracts from the AtGUS2 detected a relative increase of Gal and a reduction of Ara over-expressing line, which was not the case for AG in AG extracts (Figure 7b). Further detailed structural extracts from atglcat14a mutants (Figure 7b). The cell-wall analysis by determining the glycosidic linkage showed a extensibility results from a complex and poorly understood relative increase of Gal in 3-, 6- and 3,6-linkages (Table 1), network of interactions between various cell-wall polymers indicating the presence of longer b–1,3- and b–1,6-galac- (Wolf et al., 2012) and an AGP covalently linked to pectin tans as well as 3,6-branched Gal in the mutants. The and arabinoxylan has been described (Tan et al., 2013). We increase in the number of 3,6- and 6-linkages for Gal is do not know to what extent such structures are present in probably the result of increased acceptor sites (O6 site of the cell walls, and little is known regarding the interactions b-1,3-linked Gal and O6 site of the terminal Gal of b–1,6- of type II AG with other polymers, but the cell elongation linked galactan) at which the GalT(s) may initiate and elon- results reported here may be a result of overall changes in gate b–1,6-galactan side chains, as addition of GlcA and cell-wall architecture and integrity caused by large struc- Gal occurs at the same acceptor sites (Figure 4). Neverthe- tural changes in AG and other interacting polymers. Alter- less, addition of GlcA appears to terminate elongation of natively, the process may be more specifically mediated by b–1,3- and b–1,6-galactans, as well as initiating 3,6-Gal AG and the intracellular Ca2+ signaling pathway as branches in type II AG in vivo. interruption of cell-wall integrity, such as binding of the The level of GlcA in type II AG is very low, and we were b–glucosyl Yariv reagent to AG, triggers a rapid increase in unable to detect an altered level of GlcA in the mutants intracellular Ca2+ (Roy et al., 1999; Pickard and Fujiki, using HPAEC-PAD monosaccharide composition analysis 2005), and AG directly binds and releases Ca2+ in a pH- or LM2 immunolabeling on the root surface in seedlings dependent way (Lamport and Varnai, 2013). The binding of (Figure 7b and Figures S4 and S5). Recent technical Ca2+ is considered to occur via GlcA residues in the side advances using PACE coupled with AG-specific hydrolases chains of AG, thus family GT14 enzymes may play a role in allowed us to obtain more detailed information for the Ca2+ signaling via biosynthesis of GlcA in type II AG. side-chain structures of type II AG. As shown in Figure 8, Our future research will focus on characterization of the the GlcA-Gal/Gal and GlcA-Gal2/Gal2 ratios were approxi- GT14 enzymes with regard to catalytic mechanisms and mately 25–30% less in the two AtGlcAT14A mutants, indi- their relationship to other GTs in the biosynthesis of type II cating a role for AtGlcAT14A in synthesis of these AG in order to understand the roles of biosynthetic structures in vivo. We did not observe a visible alteration enzymes in the final structures and properties of AG, such of growth in the 3-week-old mutants analyzed, but we are as a role of GlcA in regulating the intracellular Ca2+ signal- working on generation of the double mutant atgl- ing pathway. cat14a at5g15050 in order to investigate further physiological roles of AG GlcAT in plants. EXPERIMENTAL PROCEDURES Evidently, the mutation in AtGlcAT14A caused substan- Materials tial changes in type II AG structure in the mutants. Eudes et al. (2008) investigated an Arabidopsis b–glucuronidase UDP-a–D–glucose (UDP-Glc), UDP-a–D–xylose (UDP-Xyl), GDP-a–D– a– – a– – – (AtGUS2), whose endogenous substrate is considered mannose (GDP-Man), GDP- D fucose (GDP-Fuc), UDP- D N acetyl-D–glucosamine (UDP-GlcNAc), UDP-a–D–glucuronic acid mainly to be type II AG, and reported increased cell elon- (UDP-GlcA), UDP-a–D–galactose (UDP-Gal) and b–glucuronidase gation in seedlings of over-expressing AtGUS2 and the type II from Helix pomatia were purchased from Sigma Aldrich 14 14 opposite phenotype in a T–DNA insertion line (Eudes et al., (www.sigmaaldrich.com). UDP-a–D–[ C]-Glc, UDP-a–D–[ C]-Xyl,

14 14 14 GDP-a–D–[ C]-Man, GDP-a–D–[ C]-Fuc, UDP-a–D–[ C]-GlcNAc, For generation of the phylogenetic tree, the full-length amino 14 14 UDP-a–D–[ C]-GlcA and UDP-a–D–[ C]-Gal were purchased from acid sequence of model plants (see Figure 1) and characterized PerkinElmer (http://www.perkinelmer.com). The synthetic GT14 proteins were aligned using the T–Coffee multiple sequence glycomodule GAGP8–GFP construct, placed under the control of alignment tool (http://www.tcoffee.org/Projects/tcoffee/; Notre- the tobacco signal sequence in the pBI121 vector (Xu et al., 2005), dame et al., 2000), and the phylogenetic analysis was performed was kindly provided by Marcia Kielieszewski (Department of by neighbor-joining implemented in the Molecular Evolution and Chemistry and Biochemistry, Ohio University, Athens, OH). The Genetic Analysis software package version 5 (MEGA5; Tamura method for preparation of the microsomes after expression of et al., 2011) with complete deletion of gaps and the Poisson cor- GAGP8–GFP has been described previously (Geshi et al., 2013). rection distance of substitution rates. Statistical support for phylo- The mixture of b–1,6-galactooligosaccharides was prepared by genetic grouping was estimated by 1000 bootstrap re-samplings. sequential digestion of leaf AGP extract from the mur1 mutant using a–arabinofuranosidase, exo-b–1,3-galactanase and b–glucu- Cloning of the AtGlcAT14A catalytic domain for ronidase as described previously (Tryfona et al., 2010). b–1,3-ga- expression in Pichia pastoris lactooligosaccharides of DP3, 5 and 7 were prepared by partial – acid hydrolysis with 40 mM trifluoroacetic acid at 100°C for 1 h Total RNA was extracted from the shoot apex of A. thaliana (Col TM (Kitazawa et al., 2013). Labeling of oligosaccharides with 2AA was 0) using the Spectrum Plant Total RNA kit (Sigma-Aldrich), and performed as described previously (Tryfona et al., 2010). converted to cDNAs using the iScript cDNA synthesis kit (Bio–Rad, Chemically synthesized 4–O-methyl GlcA (Sixta et al., 2009) was www.bio-rad.com). The full-length At5g39990 (AtGlcAT14A) a kind gift from Paul Kosma (Division of Organic Chemistry, Uni- sequence was amplified by PCR from cDNA using the gene-spe- ′ ′ versity of Natural Resources and Applied Life Sciences, Vienna, cific forward primer 5 -ATGAAGAAATTGAGAAGCTATTAC-3 and ′ ′ Austria). Monomeric Cerulean3 (mCer3) in the pmCer3–C1 vector reverse primer 5 -TCACTTACACTGTTTTGATCGG-3 with partial ′ (Markwardt et al., 2011) was a kind gift from Mark A. Rizzo Gateway sites added to the 5 end of each primer. The resulting (Department of Physiology, University of Maryland, Baltimore, PCR product was re-amplified using Gateway site-specific primers, MD). Termamyl SC was a kind gift from Novozymes A/S (http:// and cloned into a Gateway entry pDONR221 vector (Life Technolo- www.novozymes.com). gies, www.lifetechnologies.com). This full-length cDNA was used as a template to amplify the soluble catalytic domain (amino acid PLANT MATERIALS residues 58-447) by PCR using forward primer 5′GGATCCTCAC- GTGACTTCACCAACCGGAGGAGT-3′ and reverse primer 5′ Arabidopsis T–DNA insertion lines atglcat14a–1 (SALK_064313) GCGGCCGCGAGCTCTCACTTACACTGTTTTGATC-3′ (BamHI, PmlI, and atglcat14a–2 (SALK_043905) were obtained from the Salk NotI and SacI restriction sites are underlined). The amplified DNA Institute (Alonso et al., 2003;). Arabidopsis ecotype Col–0 was was cloned into the modified pPICZaA vector (Life Technologies) used as the wild-type for comparison. Plants were grown in the containing an introduced N–terminal FLAG tag (MDYKDDDDK) greenhouse under an 8 h photoperiod at 20°C and 70% relative (Petersen et al., 2009) using PmlI and NotI. All constructs were humidity. Homozygous lines were identified by PCR using primer verified by sequencing at each cloning step. sets suggested by the Salk Institute (gene-specific primers for SALK_064313: 5′-ACCTTAAGGCATGTTGTGTGG-3′ and 5′-CCAACA Expression of secreted soluble AtGlcAT14A in Pichia GCATTCAAGCTTTTC-3′; gene-specific primers for SALK_043905: pastoris 5′-ATTGGTTCAATCTTCGCTTTG-3′ and 5′-TCAACCAATGAGAAAT GGAGC-3′;T–DNA left border primer: 5′-ATTTTGCCGATTTCGGAA The soluble catalytic domain of At5g39990 cloned into the modi- C-3′). fied pPICZaA vector was transformed into P. pastoris KM71 strain Plants were grown hydroponically as described by Lehmann (Petersen et al., 2009), which was grown as described by Dilokpi- et al. (2009). Plants were grown for 3 weeks under short-day con- mol et al. (2011). Each 100 ml of culture medium contained Æ ditions (8 h photoperiod at 20°C), and then transferred to long-day approximately 2.5 0.9 mg recombinant protein. The culture ° conditions (16 h photoperiod at 20°C). medium was harvested by centrifugation at 12 000 g at 4 C for 1 h and concentrated tenfold using Vivaspin 20 ultrafiltration To determine the cell elongation rate in mutant lines, seeds devices (10 000 molecular weight cut-off, Polyethersulfore; GE were surface-sterilized and plated onto solidified half-strength MS Healthcare, www.gelifesciences.com). The culture medium was medium in culture plates. After 4 days of stratification at 4°Cin replaced with PBS by repeating the concentration step in the pres- the dark, seeds were exposed to 4 h fluorescent white light at ence of PBS. For purification of the recombinant AtGlcAT14A, 20°C to synchronize germination before wrapping plates in alumi- 1 ml of buffer-exchanged material was treated with 50 ll anti-Flag num foil for growth in darkness for 5 days at 20°C. M2 affinity slurry containing 50% gel (Sigma-Aldrich) in the pres- Wild-type N. benthamiana was grown in the greenhouse at ence of 5 mM n–dodecyl b–D–maltoside, and incubated for 16 h at 28°C (day) and 18°C (night) with a 16 h photoperiod and used for 4°C with rotation. The gel was collected by centrifugation at 500 g transient expression. for 30 sec at 4°C followed by washing three times with PBS, pH ° Bioinformatics 7.0, by centrifugation at 500 g for 30 sec at 4 C, prior to further characterization and enzymatic assay. Each 1 ll anti-Flag M2 affin- Transmembrane helices were predicted using the TMHMM Server ity slurry contained 7.7 Æ 1.5 ng protein. version 2.0 (http://www.cbs.dtu.dk/services/TMHMM-2.0/; Sonnham- mer et al., 1998), and N–glycosylation sites were predicted using Protein analyses NetNGlyc 1.0 (http://www.cbs.dtu.dk/services/NetNGlyc/; Blom Protein concentration was determined by the Bradford assay et al., 2004). In silico expression analysis was performed using using Bio–Rad protein assay dye reagent. SDS–PAGE was per- GeneCAT (http://genecat.mpg.de; Mutwil et al., 2008), Genevesti- formed using Criterion XT pre-cast gels (12% Bis/Tris; Bio–Rad) gator (http://www.genevestigator.com/gv/) and the Arabidopsis and Western blotting was performed using the Criterion system eFP browser (Winter et al., 2007). (Bio–Rad). After blocking the membrane using PBS containing 5%

© 2013 The Authors The Plant Journal © 2013 John Wiley & Sons Ltd, The Plant Journal, (2013), 76, 1016–1029 b–glucuronosyltransferase involved in AGP biosynthesis 1027 w/v skimmed milk, FLAG-conjugated peptides were detected using GGGCGAGGAG-3′) and reverse primer 5′-CTCTAGGGACTAGTTAA monoclonal anti-Flag M2 antibody (Sigma-Aldrich) followed by TTAAGCGTAATCTGGAACATCGTATGGGTACTTGTACAGCTCGTCC horseradish peroxidise-conjugated anti-mouse immunoglobulins ATGCC-3′) (restriction sites AvrII and SpeI are underlined, the HA (Dako Cytomation, www.dako.com), and the signal was developed tag is in italics) and replaced with cyan fluorescent protein (CFP) by chemiluminesence (Supersignal West Dura extended duration in a pEarleyGate 102 Gateway vector (Earley et al., 2006) using substrate, Thermo Scientific, www.thermoscientific.com). Endo- AvrII and SpeI. Full-length At5g39990 cDNA cloned into glycosidase H treatment was performed according to the manu- pDONR221 (Life Technologies) was moved to the modified pEar- facturer’s instructions (New England BioLabs, www.neb.com). leyGate 102 vector by LR reaction (Life Technologies) and transformed into Agrobacterium tumefaciens strain C58C1 pGV3850. Enzyme assays Determination of the subcellular localization was performed as described by Harholt et al. (2012). The sialyltransferase short cyto- For the assay using N. benthamiana microsome-expressed plasmic tail and single transmembrane domain fused to YFP GAGP –GFP as the acceptor, the reaction was performed in the 8 (STtmd–YFP; Boevink et al., 1998) was used as a Golgi marker. presence of 0.1 mM NDP-sugar (containing 277.5 Bq of NDP- After 2 days of infiltration, leaves were analyzed using a Leica TCS [14C-]-sugar), 28 mM HEPES, 10 mM MnCl , pH 7.0, and 5 llof 2 SP5 inverted confocal laser scanning microscope (www.leica.com) either tenfold concentrated AtGlcAT14A in PBS or affinity-puri- with a 63 x water immersion objective (Numerical Sperture of fied AtGlcAT14A on bead slurry as the enzyme source and 5 ll 1.2). Excitation with an argon laser was performed at 458 and of GAGP –GFP (5 lg llÀ1) as the acceptor. Incubation was per- 8 514 nm, and emission was detected at 475–505 and 525–600 nm, formed at 22°C for 16 h, and samples were stored at À20°C for mCer3 and YFP, respectively. prior to analysis. The product formed on the GAGP8–GFP acceptor was collected by immunoprecipitation by incubating with RT-PCR 0.4 lg GFP antibody (Sigma-Aldrich) in the presence of 0.15 M NaCl and 5 mM n–dodecyl b–D–maltoside at 4°C for 16 h, fol- RNA was extracted from plant material ground in liquid nitrogen lowed by incubation with 10 ll protein G/agarose (approxi- using the SpectrumTM plant total RNA kit (Sigma-Aldrich). RT–PCR mately 50% v/v slurry in H2O, Sigma-Aldrich) at 4 °C for 1 h. was performed using cDNA synthesized using the iScript cDNA The presence of 14C-sugar in the immunoprecipitated material synthesis kit (Bio–Rad) and primers 5′-AGAGAGGAGCTTCATGG was analyzed by scintillation counting. AT-3′ (a), 5′-ACCGGTTTTGCTCTTTGTGATGCTT-3′ (b), 5′-CATCCCA For the assay using 2AA-labeled b–1,6-galactooligosaccharides ACAGCATTCAAGC-3′ (c) and 5′-AACTTTCTAGCAAACGGGGC-3′ as acceptor, the reaction was performed in the presence of 0.1 mM (d) (position of primers indicated in Figure 6a). Ubiquitin-specific ′ NDP-sugar, 28 mM HEPES, 10 mM MnCl2, pH 7.0, 5 ll affinity-puri- primers were used as control: LP, 5 -TCAAATGGACCGCTCTTATC- fied AtGlcAT14A as the enzyme source and 5 ll 2AA-labeled 3′; RP, 5′-CACAGACTGAAGCGTCCAAG-3′. b–1,6-galactooligosaccharides (10 lg containing approximately 1900 fluorescence signal per ll) as the acceptor. Treatment of the Structural analysis of AG products with 0.6 units of b–glucuronidase (type II, Sigma-Aldrich) An AG-enriched fraction was prepared as described by Tryfona ° was performed at 37 C for 1 h. The reaction mixtures were filtered et al. (2012) with modifications. Fresh frozen plant material was – (Ultrafree MC poly(vinylidene difluoride) membrane, pore size of ground to a fine powder using a Retsch tissue lyzer (Qiagen, l 0.22 m) and desalted using Superdex peptide HR 10/30 (GE www.qiagen.com). Four volumes of PBS were added to samples Healthcare, www.gehealthcare.com) prior to analysis by HPLC and heated to 100°C for 30 min. During boiling, samples were trea- – (Dionex, www.dionex.com) using a TSKgel Amide 80 column ted with 10 lg llÀ1 (1.2 kilo Novo units per lg) Termamyl SC l (3 m spherical silica particle, 4.6 mm inside diameter, 150 mm (Novozymes) to remove starch. One kilo Novo unit is defined as the length of the column; 3.2 mm inside diameter, 15 mm length of amount of enzyme that dextrinizes 5.26 g of starch dry substance the guard column; TOSOH, www.tosohbioscience.com) using a (Merck Amylum soluble, Merck Millopore, www.merckmillopore. gradient of acetonitrile (ACN)/50 mM ammonium formate, pH 4.3 com) per hour under standard conditions (as defined by Novo – (80% ACN, 12 min; 80 45% ACN, 45 min; 45% ACN, 15 min; Nordisk, www.novonordisk.com) for a–amylase determination 45–80% ACN, 5 min; 80% ACN, 20 min for re-equilibrating) at a 2+ (37 Æ 0.05°C, 0.3 mM Ca , pH 5.6) (Wu et al., 2006). After the reac- constant flow rate (0.5 ml per minute) at 25°C. tion, samples were cooled to room temperature, cell debris was The products were detected using a fluorescence detector removed by centrifugation at 4000 g, and the supernatant was b– b– (RF2000, Dionex). 2AA-labeled 1,6-galactose, 1,6-galactobiose mixed with 2.5 volumes of 96% v/v ethanol and kept at 4°C over- b– and 1,6-galactotriose were used as standards to identify the night. Precipitates were isolated by centrifugation (12 000 g, b– 2AA-labeled 1,6-galactooligosaccharides elution positions. 20 min, 4°C), and resuspended in 10 volumes of water by constant For the assay using b–1,3-galactooligosaccharides as acceptor, shaking at 4°C overnight. Monosaccharide composition analysis by the reaction was performed in the presence of 0.1 mM NDP-sugar, HPAEC-PAD was performed as described by Obro et al. (2004). The 16 mM McIlvaine buffer (McIlvaine, 1921), pH 5.0, 5 ll affinity-puri- PACE analysis was performed as described by Tryfona et al. (2010). fied AtGlcAT14A as the enzyme source, and 2.5 llof2mM b–1,3- Glycosidic linkage analysis was performed by GC/MS analysis of Gal3-, b–1,3-Gal5-orb–1,3-Gal7-2AA as the acceptor. The reaction their partially methylated alditol acetates derivatized as described mixtures were filtered through Ultrafree–MC poly(vinylidene by Ciucanu and Kerek (1984) with modifications by Ciucanu (2006). difluoride) membrane (pore size of 0.22 lm) prior to analysis The resulting partially methylated alditol acetates were separated using a TSKgel Amide–80 column by HPLC as described above using a gas chromatograph (Agilent 7890A, Agilent Technologies, with a modified ACN gradient (80% ACN, 12 min; 80–45% ACN, www.agilent.com) equipped with a Supelco SP2380 column 55 min; 45% ACN, 15 min; 45–80% ACN, 5 min; 80% ACN). (Sigma-Aldrich) and a mass spectrometer (Agilent 5975C). Peaks were identified based on their retention time and ion fragmentation Subcellular localization pattern, and assigned according to the Complex Carbohydrate Research Center’s partially methylated alditol acetate database Monomeric Cerulean3 (mCer3; Markwardt et al., 2011) was PCR- (http://www.ccrc.uga.edu/specdb/ms/pmaa/pframe.html). amplified using forward primer 5′-GGTGCCTAGGGTGGTGAGCAA

ACKNOWLEDGEMENTS Dilokpimol, A., Nakai, H., Gotfredsen, C.H., Baumann, M.J., Nakai, N., Abou Hachem, M. and Svensson, B. (2011) Recombinant production and char- We would like to thank Jack Egelund and Peter B. Jørgensen for acterisation of two related GH5 endo-b–1,4-mannanases from Aspergillus their support with the cloning work. This work was supported by nidulans FGSC A4 showing distinctly different transglycosylation capac- – the Faculty of Life Sciences, University of Copenhagen (E.K.), the ity. Biochim. Biophys. Acta, 1814, 1720 1729. Danish Council for Strategic Research, Food, Health and Welfare Earley, K.W., Haag, J.R., Pontes, O., Opper, K., Juehne, T., Song, K.M. and Pikaard, C.S. (2006) Gateway-compatible vectors for plant functional and the Danish Council for Independent Research, Technology genomics and proteomics. Plant J. 45, 616–629. and Production Sciences (A.D., C.P.P. and N.G.), the Villum Egelund, J., Petersen, B.L., Motawia, M.S., Damager, I., Faik, A., Olsen, C.E., Foundation’s Young Investigator Program (J.H.), and the US Ishii, T., Clausen, H., Ulvskov, P. and Geshi, N. (2006) Arabidopsis thali- Department of Energy (M.H.). Work performed by G.X. and M.P. ana RGXT1 and RGXT2 encode Golgi-localized (1,3)-a–D–xylosyltransfe- was funded by the Energy Biosciences Institute. Imaging data rases involved in the synthesis of pectic rhamnogalacturonan–II. Plant were collected at the Center for Advanced Bioimaging of Cell, 18, 2593–2607. Denmark, University of Copenhagen, Denmark. Ellis, M., Egelund, J., Schultz, C.J. and Bacic, A. (2010) Arabinogalactan proteins: key regulators at the cell surface? Plant Physiol. 153, 403–419. SUPPORTING INFORMATION Estevez, J.M., Kieliszewski, M.J., Khitrov, N. and Somerville, C. (2006) Char- acterization of synthetic hydroxyproline-rich proteoglycans with arabino- Additional Supporting Information may be found in the online ver- galactan protein and extensin motifs in Arabidopsis. Plant Physiol. 142, sion of this article. 458–470. Figure S1. Co-expression analysis of AtGALT31A and AtGlcAT14A. Eudes, A., Mouille, G., Thevenin, J., Goyallon, A., Minic, Z. and Jouanin, L. (2008) Purification, cloning and functional characterization of an endoge- Figure S2. Xylosyltransferase activity assay of RGXT1. nous b–glucuronidase in Arabidopsis thaliana. Plant Cell Physiol. 49, Figure S3. Localization and spectral FRET analysis for AtGAL31A– 1331–1341. mCer3 and AtGlcAT14A–YFP. Geshi, N., Johansen, J.N., Dilokpimol, A. et al. (2013) A galactosyltransfer- Figure S4. Surface labeling of lateral root tips from 8-day-old ase acting on arabinogalactan protein glycans is essential for embryo development in Arabidopsis. Plant J. 76, 128–137. seedlings using LM2 antibody. Gille, S., Sharma, V., Baidoo, E.E.K., Keasling, J.D., Scheller, H.V. and Pauly, Figure S5. Monosaccharide composition analysis of AG extracts M. (2013) Arabinosylation of a Yariv-precipitable cell wall polymer from roots of mature plants. impacts plant growth as exemplified by the Arabidopsis glycosyltransfer- Figure S6. Homology model of AtGlcAT14A, and amino acid ase mutant ray1. Mol. Plant, 6, 1369–1372. sequence alignment of GT14 enzymes. Harholt, J., Jensen, J.K., Verhertbruggen, Y. et al. (2012) ARAD proteins associated with pectic Arabinan biosynthesis form complexes when tran- Methods S1. Materials and methods for the supporting informa- siently overexpressed in planta. Planta, 236, 115–128. tion. Kitazawa, K., Tryfona, T., Yoshimi, Y. et al. (2013) b–galactosyl Yariv reagent Table S1. Amino acid sequence identity (%) among Arabidopsis binds to the b-1,3-galactan of arabinogalactan proteins. Plant Physiol. GT14 proteins. 161, 1117–1126. Lairson, L.L., Henrissat, B., Davies, G.J. and Withers, S.G. (2008) Glyco- syltransferases: structures, functions, and mechanisms. Annu. Rev. REFERENCES Biochem. 77, 521–555. Lamport, D.T.A. and Varnai, P. (2013) Periplasmic arabinogalactan glycopro- Alonso, J.M., Stepanova, A.N., Leisse, T.J. et al. (2003) Genome-wide inser- teins act as a calcium capacitor that regulates plant growth and develop- tional mutagenesis of Arabidopsis thaliana. Science, 301, 653–657. ment. New Phytol. 197,58–64. Amado, M., Almeida, R., Schwientek, T. and Clausen, H. (1999) Identification Lee, C., Teng, Q., Zhong, R.Q. and Ye, Z.H. (2012) Arabidopsis GUX proteins and characterization of large galactosyltransferase gene families: galac- are glucuronyltransferases responsible for the addition of glucuronic tosyltransferases for all functions. Biochim. Biophys. Acta, 1473,35–53. acid side chains onto xylan. Plant Cell Physiol. 53, 1204–1216. Basu, D., Liang, Y., Liu, X., Himmeldirk, K., Faik, A., Kieliszewski, M., Held, Lehmann, M., Schwarzländer, M., Obata, T. et al. (2009) The metabolic M. and Showalter, A.M. (2013) Functional identification of a hydroxypro- response of Arabidopsis roots to oxidative stress is distinct from that of line-O–galactosyltransferase specific for arabinogalactan protein biosyn- heterotrophic cells in culture and highlights a complex relationship thesis in Arabidopsis. J. Biol. Chem. 288, 10132–10143. between the levels of transcripts, metabolites, and flux. Mol. Plant, 2, Bierhuizen, M.F.A. and Fukuda, M. (1992) Expression cloning of a cDNA 390–406. encoding UDP-GlcNAc:Galb1–3-GalNAc–R (GlcNAc to GalNAc) b1–6Glc- Markwardt, M.L., Kremers, G.J., Kraft, C.A., Ray, K., Cranfill, P.J.C., Wilson, NAc transferase by gene transfer into CHO cells expressing polyoma K.A., Day, R.N., Wachter, R.M., Davidson, M.W. and Rizzo, M.A. (2011) large tumor antigen. Proc. Natl Acad. Sci. USA, 89, 9326–9330. An improved Cerulean fluorescent protein with enhanced brightness and Bierhuizen, M.F.A., Mattei, M.G. and Fukuda, M. (1993) Expression of the reduced reversible photoswitching. PLoS One, 6, e17896. developmental I antigen by a cloned human cDNA encoding a member McIlvaine, T.C. (1921) A buffer solution for colorimetric comparison. J. Biol. of a b–1,6-N–acetylglucosaminyltransferase gene family. Genes Dev. 7, Chem., 49, 183–186. 468–478. Mortimer, J.C., Miles, G.P., Brown, D.M. et al. (2010) Absence of branches Blom, N., Sicheritz-Ponten, T., Gupta, R., Gammeltoft, S. and Brunak, S. from xylan in Arabidopsis gux mutants reveals potential for simplification (2004) Prediction of post-translational glycosylation and phosphorylation of lignocellulosic biomass. Proc. Natl Acad. Sci. USA, 107, 17409–17414. of proteins from the amino acid sequence. Proteomics, 4, 1633–1649. Mutwil, M., Obro, J., Willats, W.G.T. and Persson, S. (2008) GeneCAT – Boevink, P., Oparka, K., Cruz, S.S., Martin, B., Betteridge, A. and Hawes, C. novel webtools that combine BLAST and co-expression analyses. Nucleic (1998) Stacks on tracks: the plant Golgi apparatus traffics on an actin/ER Acids Res. 36, W320–W326. network. Plant J. 15, 441–447. Nie, S.P., Wang, C., Cui, S.W., Wang, Q., Xie, M.Y. and Phillips, G.O. (2013) Cantarel, B.L., Coutinho, P.M., Rancurel, C., Bernard, T., Lombard, V. and A further amendment to the classical core structure of gum arabic Henrissat, B. (2009) The Carbohydrate-Active EnZymes database (CAZy): (Acacia senegal). Food Hydrocolloids, 31,42–48. an expert resource for glycogenomics. Nucleic Acids Res. 37, Notredame, C., Higgins, D.G. and Heringa, J. (2000) T-Coffee: a novel D233–D238. method for fast and accurate multiple sequence alignment. J. Mol. Biol. Ciucanu, I. (2006) Per-O–methylation reaction for structural analysis of car- 302, 205–217. bohydrates by mass spectrometry. Anal. Chim. Acta, 576, 147–155. Obro, J., Harholt, J., Scheller, H.V. and Orfila, C. (2004) Rhamnogalacturo- Ciucanu, I. and Kerek, F. (1984) A simple and rapid method for the perme- nan I in Solanum tuberosum tubers contains complex arabinogalactan thylation of carbohydrates. Carbohydr. Res. 131, 209–217. structures. Phytochemistry, 65, 1429–1438.

Pak, J.E., Arnoux, P., Zhou, S.H., Sivarajah, P., Satkunarajah, M., Xing, X.K. Tan, L., Showalter, A.M., Egelund, J., Hernandez-Sanchez, A., Doblin, M.S. and Rini, J.M. (2006) X–ray crystal structure of leukocyte type core 2 and Bacic, A. (2012) Arabinogalactan-proteins and the research chal- b1,6–N-acetylglucosaminyltransferase – evidence for a convergence of lenges for these enigmatic plant cell surface proteoglycans. Front. Plant metal ion-independent glycosyltransferase mechanism. J. Biol. Chem. Sci. 3, 140. 281, 26693–26701. Tan, L., Eberhard, S., Pattathil, S. et al. (2013) An Arabidopsis cell wall pro- Petersen, B.L., Egelund, J., Damager, I., Faber, K., Jensen, J.K., Yang, Z., teoglycan consists of pectin and arabinoxylan covalently linked to an Bennett, E.P., Scheller, H.V. and Ulvskov, P. (2009) Assay and heterolo- arabinogalactan protein. Plant Cell, 25, 270–287. gous expression in Pichia pastoris of plant cell wall type–II membrane Tryfona, T., Liang, H.C., Kotake, T. et al. (2010) Carbohydrate structural anchored glycosyltransferases. Glycoconj. J. 26, 1235–1246. analysis of wheat flour arabinogalactan protein. Carbohydr. Res. 345, Pickard, B.G. and Fujiki, M. (2005) Ca2+ pulsation in BY–2 cells and evidence 2648–2656. for control of mechanosensory Ca2+-selective channels by the plasma- Tryfona, T., Liang, H.C., Kotake, T., Tsumuraya, Y., Stephens, E. and lemmal reticulum. Funct. Plant Biol. 32, 863–879. Dupree, P. (2012) Structural characterization of Arabidopsis leaf arabino- Poulsen, C.P., Vereb, G., Geshi, N. and Schulz, A. (2013) Inhibition of cyto- galactan polysaccharides. Plant Physiol. 160, 653–666. plasmic streaming by cytochalasin D is superior to paraformaldehyde Tsumuraya, Y., Ogura, K., Hashimoto, Y., Mukoyama, H. and Yamamoto, S. fixation for measuring FRET between fluorescent protein-tagged Golgi (1988) Arabinogalactan-proteins from primary and mature roots of radish components. Cytometry A, 83, 830–838. (Raphanus sativus L). Plant Physiol. 86, 155–160. Rennie, E.A., Hansen, S.F., Baidoo, E.E.K., Hadi, M.Z., Keasling, J.D. and Tsumuraya, Y., Mochizuki, N., Hashimoto, Y. and Kovac, P. (1990) Purifica- Scheller, H.V. (2012) Three members of the Arabidopsis glycosyltransfer- tion of an exo-b-(1?3)-D–galactanase of Irpex lacteus (Polyporus ase family 8 are xylan glucuronosyltransferases. Plant Physiol. 159, tulipiferae) and its action on arabinogalactan-proteins. J. Biol. Chem. 1408–1417. 265, 7207–7215. Roy, S.J., Holdaway-Clarke, T.L., Hackett, G.R., Kunkel, J.G., Lord, E.M. and Van Hengel, A.J. and Roberts, K. (2002) Fucosylated arabinogalactan-pro- Hepler, P.K. (1999) Uncoupling secretion and tip growth in lily pollen teins are required for full root cell elongation in Arabidopsis. Plant J. 32, tubes: evidence for the role of calcium in exocytosis. Plant J. 19, 105–113. 379–386. Williams, D., Longmore, G., Matta, K.L. and Schachter, H. (1980) Mucin Schwientek, T., Yeh, J.C., Levery, S.B., Keck, B., Merkx, G., van Kessel, synthesis. II. Substrate specificity and product identification studies on A.G., Fukuda, M. and Clausen, H. (2000) Control of O–glycan branch for- canine submaxillary gland UDP-GlcNAc:Gal b1–3GalNAc (GlcNAc ? Gal- mation – molecular cloning and characterization of a novel thymus-asso- NAc) b6–N-acetylglucosaminyltransferase. J. Biol. Chem. 255, 1253–1261. ciated core 2 b1,6–N-acetylglucosaminyltransferase. J. Biol. Chem. 275, Wilson, I.B.H. (2002) Functional characterization of Drosophila melanogaster 11106–11113. peptide O–xylosyltransferase, the key enzyme for proteoglycan chain Seifert, G.J. and Roberts, K. (2007) The biology of arabinogalactan proteins. initiation and member of the core 2/I N–acetylglucosaminyltransferase Annu. Rev. Plant Biol. 58, 137–161. family. J. Biol. Chem. 277, 21207–21212. Showalter, A.M., Keppler, B., Lichtenberg, J., Gu, D.Z. and Welch, L.R. Winter, D., Vinegar, B., Nahal, H., Ammar, R., Wilson, G.V. and Provart, N.J. (2010) A bioinformatics approach to the identification, classification, and (2007) An ‘electronic Fluorescent Pictograph’ browser for exploring and analysis of hydroxyproline-rich glycoproteins. Plant Physiol. 153, analyzing large-scale biological data sets. PLoS One, 2, e718. 485–513. Wolf, S., Hématy, K. and Höfte, H. (2012) Growth Control and Cell Wall Sixta, G., Herok, W., Gruber, C., Weber, H.K., Sixta, H. and Kosma, P. (2009) Signaling in Plants. Annu. Rev. Plant Biol. 63, 381–407. Synthesis of 4–O–methyl glucuronic acid. Lenzinger Ber. 87, 11. Wu, X., Wang, D., Bean, S.R. and Wilson, J.P. (2006) Ethanol production Smallwood, M., Yates, E.A., Willats, W.G.T., Martin, H. and Knox, J.P. (1996) from pearl millet using Saccharomyces cerevisiae. Cereal Chem. 83, Immunochemical comparison of membrane-associated and secreted 127–131. arabinogalactan-proteins in rice and carrot. Planta, 198, 452–459. Wu, Y.Y., Williams, M., Bernard, S., Driouich, A., Showalter, A.M. and Faik, Sonnhammer, E.L., von Heijne, G. and Krogh, A. (1998) A hidden Markov A. (2010) Functional identification of two nonredundant Arabidopsis a model for predicting transmembrane helices in protein sequences. Proc. (1,2)fucosyltransferases specific to arabinogalactan proteins. J. Biol. Int. Conf. Intell. Syst. Mol. Biol. 6, 175–182. Chem. 285, 13638–13645. Takata, R., Tokita, K., Mori, S. et al. (2010) Degradation of carbohydrate Xu, J.F., Shpak, E., Gu, T.Y., Moo-Young, M. and Kieliszewski, M. (2005) moieties of arabinogalactan-proteins by glycoside hydrolases from Neu- Production of recombinant plant gum with tobacco cell culture in biore- rospora crassa. Carbohydr. Res. 345, 2516–2522. actor and gum characterization. Biotechnol. Bioeng. 90, 578–588. Tamura, K., Peterson, D., Peterson, N., Stecher, G., Nei, M. and Kumar, S. Ye, C.Y., Li, T., Tuskan, G.A., Tschaplinski, T.J. and Yang, X.H. (2011) Com- (2011) MEGA5: molecular evolutionary genetics analysis using maximum parative analysis of GT14/GT14-like gene family in Arabidopsis, Oryza, likelihood, evolutionary distance, and maximum parsimony methods. Populus, Sorghum and Vitis. Plant Sci. 181, 688–695. Mol. Biol. Evol. 28, 2731–2739. Yeh, J.C., Ong, E. and Fukuda, M. (1999) Molecular cloning and expression Tan, L., Qiu, F., Lamport, D.T. and Kieliszewski, M.J. (2004) Structure of a of a novel b–1,6–N-acetylglucosaminyltransferase that forms core 2, hydroxyproline (Hyp)-arabinogalactan polysaccharide from repetitive core 4, and I branches. J. Biol. Chem. 274, 3215–3221. Ala-Hyp expressed in transgenic Nicotiana tabacum. J. Biol. Chem. 279, Yu, L.C., Twu, Y.C., Chang, C.Y. and Lin, M. (2001) Molecular basis of the 13156–13165. adult i phenotype and the gene responsible for the expression of the Tan, L., Varnai, P., Lamport, D.T.A., Yuan, C.H., Xu, J.F., Qiu, F. and human blood group I antigen. Blood, 98, 3840–3845. Kieliszewski, M.J. (2010) Plant O–hydroxyproline arabinogalactans are Zhou, Y.H., Li, S.B., Qian, Q. et al. (2009) BC10, a DUF266-containing and composed of repeating trigalactosyl subunits with short bifurcated side Golgi-located type II membrane protein, is required for cell-wall biosyn- chains. J. Biol. Chem. 285, 24575–24583. thesis in rice (Oryza sativa L.). Plant J. 57, 446–462.