TY - JOUR
AU1 - Rocha, Joana
AU2 - Cicéron, Félix
AU3 - de Sanctis, Daniele
AU4 - Lelimousin, Mickael
AU5 - Chazalet, Valérie
AU6 - Lerouxel, Olivier
AU7 - Breton, Christelle
AB - Abstract The plant cell wall is a complex and dynamic network made mostly of cellulose, hemicelluloses, and pectins. Xyloglucan, the major hemicellulosic component in Arabidopsis thaliana, is biosynthesized in the Golgi apparatus by a series of glycan synthases and glycosyltransferases before export to the wall. A better understanding of the xyloglucan biosynthetic machinery will give clues toward engineering plants with improved wall properties or designing novel xyloglucan-based biomaterials. The xyloglucan-specific α2-fucosyltransferase FUT1 catalyzes the transfer of fucose from GDP-fucose to terminal galactosyl residues on xyloglucan side chains. Here, we present crystal structures of Arabidopsis FUT1 in its apoform and in a ternary complex with GDP and a xylo-oligosaccharide acceptor (named XLLG). Although FUT1 is clearly a member of the large GT-B fold family, like other fucosyltransferases of known structures, it contains a variant of the GT-B fold. In particular, it includes an extra C-terminal region that is part of the acceptor binding site. Our crystal structures support previous findings that FUT1 behaves as a functional dimer. Mutational studies and structure comparison with other fucosyltransferases suggest that FUT1 uses a SN2-like reaction mechanism similar to that of protein-O-fucosyltransferase 2. Thus, our results provide new insights into the mechanism of xyloglucan fucosylation in the Golgi. INTRODUCTION Plant cells are surrounded by a strong wall comprising complex polysaccharides and glycoproteins, which confer mechanical properties important for the shape and development of plants and for their adaptation to environmental changes. The primary cell wall consists mostly of cellulose, hemicelluloses, and pectins. These polymers represent one of the largest source of renewable carbohydrate on Earth and therefore are of considerable interest for the production of bioenergy and biomaterials (Burton and Fincher, 2014; Loqué et al., 2015). In addition, plant polysaccharides are important sources of dietary fiber with applications in human health (Koropatkin et al., 2012). Better knowledge of wall architecture and the fine structure of its constituents will undoubtedly lead to novel applications. Similarly, identifying the molecular players of the biosynthetic machinery will provide opportunities to manipulate cell wall components and improve plant performance for industrial applications. The hemicelluloses are used in numerous industrial applications (i.e., as food additives and in medicinal applications) (Pauly et al., 2013). Among hemicelluloses, xyloglucan (XyG) is found in various proportions in the primary cell walls of all angiosperms, but its fine structure depends on the species (Fry 1989; Hsieh and Harris, 2009). XyG is the most abundant hemicellulose in the primary wall of dicots (making up to 20 to 30% of the dry mass) (Scheller and Ulvskov, 2010). This polymer consists of a β-1,4-linked glucan backbone that can be substituted with a diverse array of glycosyl residues and with O-acetyl groups (Scheller and Ulvskov, 2010; Pauly and Keegstra, 2016). A one-letter code is used to describe the XyG side chains: e.g., G corresponds to the unbranched glucosyl residue, while X, L, and F correspond to glucosyl residues substituted with Xyl, Gal-Xyl, or Fuc-Gal-Xyl motifs, respectively (Fry et al., 1993). The glucan backbone is regularly substituted with α-d-xylosyl residues at the O6 position, but two different core motifs have been described: (1) the XXXG-type (three xylosyl residues per four glycosyl residues) in dicots and commelinid monocots and (2) the XXGGn core motif in Poaceae and Solanaceae plant species (Peña et al., 2008; Hsieh and Harris, 2009). In XyG of the XXXG-type, xylosyl residues can be substituted at O2 with β-d-Gal (L side chain) and the Gal residue further extended at O2 with α-l-Fuc (F side chain), thus forming the fucogalactoxyloglucan. Whether the structural diversity of XyG is related to specific functions is yet to be determined, but XyG is thought to play an essential structural role in cell wall mechanics (Park and Cosgrove, 2015). The most important enzymes involved in XyG biosynthesis belong to a large group of enzymes called glycosyltransferases (GTs). GTs catalyze the transfer of a sugar residue from an activated donor (usually a nucleotide-sugar) to an acceptor molecule. They are classified as either inverting or retaining enzymes, depending on whether the anomeric configuration of the transferred sugar is inverted or retained in the final product. Based on amino acid sequence similarities, GTs are currently classified into 97 families (designated GTX, X corresponding to the family number). The 3D structures of nucleotide-sugar-dependent GTs are relatively conserved; only two types of folds (and variants thereof), termed GT-A and GT-B, have been described to date. Most of the GTs involved in XyG biosynthesis in Arabidopsis thaliana have been identified (Pauly and Keegstra, 2016). The backbone is produced in the Golgi by one or more members of the cellulose synthase-like family C (CSLC) (Cocuron et al., 2007). Side chains are added by Golgi-resident GTs, including α6-xylosyltransferases (XXT1, XXT2, and XXT5), β2-galactosyltransferases (Mur3 and XLT2), and α2-fucosyltransferase (FUT1) (Perrin et al., 1999; Faik et al., 2002; Madson et al., 2003; Jensen et al., 2012; Cavalier and Keegstra, 2006; Culbertson et al., 2016). With the exception of CSLC proteins, which are multipass transmembrane proteins, all other GTs involved in XyG biosynthesis are type II membrane proteins consisting of a short N-terminal cytoplasmic tail, a single transmembrane domain followed by a stem region, and a large C-terminal catalytic domain facing the luminal side. In Arabidopsis, a single XyG:α2-fucosyltransferase (FUT1) was shown to add the terminal fucose residue (Perrin et al., 1999, 2003; Faik et al., 2000). According to the sequence-based classification of GTs (CAZy database, http://www.cazy.org/), FUT1 belongs to the inverting GT37 family. This family comprises only plant sequences and includes 10 Arabidopsis homologs (denoted as FUT1 to FUT10). FUT1 catalyzes the transfer of fucose from GDP-β-l-fucose to the second galactosyl residue in the XLLG and XXLG subunits, thereby producing XLFG and XXFG subunits, respectively. However, the enzyme does not add fucose to the first galactosyl residue of XLLG, indicating that it has a strong acceptor specificity. Moreover, although the shorter oligosaccharides XXLG and XLLG can function as inhibitors, they are poor substrates for the enzyme, which strongly prefers longer chains of XyG as acceptor (Faik et al., 2000; Cicéron et al., 2016). Previous analysis performed on pea (Pisum sativum) FUT1 indicated that it does not require divalent cations for activity and that the nucleotide-sugar and acceptor substrates associate with the enzyme in a random order (Faik et al., 2000). A truncated recombinant form of Arabidopsis FUT1, deleted of its cytosolic domain and transmembrane domain, has been successfully produced in the baculovirus/insect cell system. The resulting protein was active and behaved as a noncovalent dimer in solution (Cicéron et al., 2016). We report here the crystal structures of the catalytic domain of Arabidopsis FUT1, in its apoform and in a ternary complex with its product GDP and the branched XLLG nonsaccharide acceptor substrate. The structures showed a previously unreported variant of the classical GT-B folding. These data, together with mutant analysis, provide insights into the molecular mechanism by which this glycosyltransferase recognizes its donor and acceptor substrates and enhance our understanding of the XyG fucosylation process. This work establishes the structure of a glycosyltransferase involved in plant cell wall biosynthesis and provides the initial structure for the entire GT37 family. RESULTS AND DISCUSSION Overall Structure of FUT1 We report here the crystal structures of a soluble form of Arabidopsis FUT1 (deleted of its first 68 amino acids, thus excluding the cytosolic and transmembrane regions), in its apoform (2.1 Å resolution, referred to as apo-FUT1) and in complex with GDP and a xylo-oligosaccharide fragment (2.2 Å resolution, referred to as FUT1:GDP:XLLG) (Supplemental Tables 1 and 2). According to the DALI server (Holm and Rosenström, 2010), the closest structural homologs of FUT1 are fucosyltransferases belonging to different GT families: the human α6-fucosyltransferase FUT8 (Ihara et al., 2007) and the Bradyrhizobium sp WZ9 α6-fucosyltransferase NodZ (Brzezinski et al., 2007, 2012), both of which are members of the GT23 family; and GT65 Caenorhabditis elegans POFUT1 (Lira-Navarrete et al., 2011) and GT68 human POFUT2 (Chen et al., 2012), which are involved in protein O-fucosylation. All of these proteins are members of the GT-B superfamily of glycosyltransferases, which share the same global GT-B fold (Supplemental Figure 1). The canonical GT-B fold consists of two Rossmann-type α/β/α domains of similar size, designated as the N-domain and C-domain, separated by a large cleft where the active site is located, and stabilized by two long C-terminal α-helices (Breton et al., 2006). However, variants of the GT-B fold with insertions and deletions in the N-domain and/or the presence of add-on domains have been described (for a recent review, see Breton et al., 2012). The protein FUT1 exhibits several unusual features in both domains that prompt us to classify this GT37 fucosyltransferase as a GT-B fold variant. The FUT1 N-domain (residues 130 to 340) displays a core β-sheet (Nβ1 to Nβ4) organized in a noncanonical order and flanked by helical elements (Nα2 to Nα8) on one side and loops on the other side (Figure 1). Two disulfide bridges (C111-C216 and C146-C171) connect the more exposed protein chain elements to the domain core, including the stem and the Nα3 helix. The C-domain (residues 341 to 513) displays a central β-sheet (Cβ1 to Cβ6, strand 6 being antiparallel to the rest) surrounded by α-helices (Cα1 to Cα6), resembling a Rossmann fold. The C-domain is followed by a long protein segment (residues 514 to 558) containing two β-hairpins (Cβ7-Cβ8 and Cβ9-Cβ10) that lines the N-domain, and which can be regarded as an extra C-terminal domain. This region establishes several polar and hydrophobic contacts with residues of both N- and C-domains and most importantly, two disulfide bridges (C282-C530 and C521-C548) firmly anchor this region to the rest of the protein (Figure 1). Interestingly, this region also plays an important role in acceptor recognition (see below). The structural organization seen in FUT1 has not been described in other GT-B enzymes. Figure 1. Open in new tabDownload slide Overall Structure of FUT1. (A) Representation of the FUT1 monomer structure. The stem region is shown in red, the N-domain is in blue, the C-domain is in green, and the extra C-domain is in orange. The N-glycosylated asparagine residues and disulfide bonds are in stick representation colored gray for carbon, red for oxygen, blue for nitrogen, and yellow for sulfur atoms. (B) Topology diagram of the FUT1 structure, colored as in (A). The missing loop regions are represented in dashed lines. The N-glycosylated asparagine residues are represented as red stars, and disulfide bonds as yellow lines, with the residue annotation. Figure 1. Open in new tabDownload slide Overall Structure of FUT1. (A) Representation of the FUT1 monomer structure. The stem region is shown in red, the N-domain is in blue, the C-domain is in green, and the extra C-domain is in orange. The N-glycosylated asparagine residues and disulfide bonds are in stick representation colored gray for carbon, red for oxygen, blue for nitrogen, and yellow for sulfur atoms. (B) Topology diagram of the FUT1 structure, colored as in (A). The missing loop regions are represented in dashed lines. The N-glycosylated asparagine residues are represented as red stars, and disulfide bonds as yellow lines, with the residue annotation. We previously reported that the purified truncated FUT1 enzyme behaves as a noncovalent homodimer in solution (Cicéron et al., 2016). This is corroborated by our analysis of the crystal contents showing that four molecules in the asymmetric unit associate in two equivalent dimers designated as A-B and C-D (Supplemental Figure 2). The dimer formation implies a solvent occlusion of ∼1200 Å2, which corresponds to ∼6% solvent-accessible surface per monomer, as assessed with PISA (Protein Interfaces, Surfaces, and Assemblies; Krissinel and Henrick, 2007), and involves the C-domains of the two protein chains (Figure 2A). PISA analysis also indicated that the dimer observed in the crystals has physiological relevance, with a complex formation significance score of 0.8. The dimer interface comprises several polar and hydrophobic interactions between equivalent residues from the Cα1, Cα3, Cβ3, and Cα4 secondary structure elements of the two monomers (Figure 2B). In particular, W431 is almost completely buried in a hydrophobic cavity formed by the other protein chain in a plug-like interaction delimited by Cα1, Cα4, and Cβ3, whereas H356 stacks against its own symmetric residue. Figure 2. Open in new tabDownload slide Dimeric Functional Unit of FUT1. (A) Solvent-accessible surface representation of the FUT1 dimer. The color code for monomer A (left) is the same as in Figure 1, and the monomer B (right) is shown in gray with the visible portion of its stem region highlighted in pink. (B) Residues contact network at the dimer interface. Residues are in stick representation colored by element in green (chain A) or gray (chain B) for carbon, red for oxygen, and blue for nitrogen atoms. Residues and secondary structure elements annotated underlined green or italic gray for monomers A and B, respectively. Figure 2. Open in new tabDownload slide Dimeric Functional Unit of FUT1. (A) Solvent-accessible surface representation of the FUT1 dimer. The color code for monomer A (left) is the same as in Figure 1, and the monomer B (right) is shown in gray with the visible portion of its stem region highlighted in pink. (B) Residues contact network at the dimer interface. Residues are in stick representation colored by element in green (chain A) or gray (chain B) for carbon, red for oxygen, and blue for nitrogen atoms. Residues and secondary structure elements annotated underlined green or italic gray for monomers A and B, respectively. The FUT1 structure shows that 10 of the 11 cysteine residues present in the protein construct are involved in intramolecular bonds (Figure 1). Moreover, the unique free cysteine (C246) is buried in the protein chain and thus is not available to establish a covalent linkage between FUT1 monomers or other interacting proteins. Interestingly, using bimolecular fluorescence complementation assay and coimmunoprecipitation assays, Chou et al. (2015) showed that FUT1 can form disulfide-bonded homomers. The full-length sequence of FUT1 shows an extra cysteine residue (C48) located in the transmembrane helix; therefore, it is possible that an intermolecular disulfide bond is formed between two FUT1 transmembrane helices. Such covalent association has been described in other Golgi GTs (reviewed in Kellokumpu et al., 2016). The three-dimensional structure of FUT1 also reveals a flexible protein segment preceding the N-domain, corresponding to a portion of the stem region that connects the transmembrane domain (deleted in this protein construct) to the catalytic domain of FUT1. The stem region, which is solvent exposed, is perpendicular to the N- and C-domain interface where the initial Nα1 helix stacks against helical elements of the C-domain (Figure 1). Electrostatic surface analysis of the FUT1 dimer shows contrasted views with a prominent positively charged “membrane-facing” surface and a negatively charged “luminal side,” with the exception of the positive GDP/GDP-fucose binding pocket that is located on the luminal side (Figure 3). From these observations, one can hypothesize that the FUT1 dimer may interact through its membrane side with the negatively charged membrane bilayer. The FUT1 protein construct used in this study harbors two N-glycosylation sites, at positions N88 (in the stem region) and N504 (in the C-domain), which according to previous MALDI-TOF analysis are fully occupied (Cicéron et al., 2016). The two asparagine residues are solvent exposed but there is weak electron density near each Asn residue, indicating the likely attachment of N-glycan chains at these sites. The comparison between apo-FUT1 and FUT1:GDP:XLLG structures reveals no major conformational changes in the polypeptide chains. The two structures are almost identical and only subtle loop movements in the vicinity of the donor and acceptor binding pockets are observed. The Cβ1-Cα2 loop (V365-F375) moves ∼2 Å toward the donor binding site, whereas the Cβ5-Cβ6 loop (L499-C512) moves slightly (∼1 Å) toward the XLLG acceptor binding site, leading to rearrangement of several residue side chains such as R366 or R501 (see below). Figure 3. Open in new tabDownload slide Electrostatic Surface Representation of the FUT1 Dimer. Views from the luminal and membrane-facing sides are displayed. The dimer interface is indicated as a black dashed line and the colored arrows show the binding sites for the GDP donor (red) and XLLG acceptor (black) substrates. Figure 3. Open in new tabDownload slide Electrostatic Surface Representation of the FUT1 Dimer. Views from the luminal and membrane-facing sides are displayed. The dimer interface is indicated as a black dashed line and the colored arrows show the binding sites for the GDP donor (red) and XLLG acceptor (black) substrates. Donor Sugar Binding Site Despite extensive attempts to obtain FUT1 in complex with its donor substrate (GDP-fucose), our crystals always showed clear electron density only for the GDP moiety. This can be attributed either to the flexibility and/or disorder of fucose residue since it is exposed to the solvent or to the hydrolysis of the donor substrate. In all structures, the GDP ligand occupies the same position as in the higher resolution FUT1:GDP:XLLG structure (Figure 4A). The guanine is sandwiched between the side chains of F484 and L418, and the carboxylate group of E466 establishes two H-bonds with the N1 and N2 atoms of the guanine ring (Figure 4B). The ribose establishes two hydrogen bonds, with a water molecule and with H459 side chain. The nucleotide donor is firmly held in place mainly by the diphosphate moiety, which forms an intricate network of polar interactions with several protein residues. The α-phosphate interacts with the backbone amides from G183 and N184, which are the only residues from the N-domain that interact with the GDP moiety. The O3A atom of the phosphate group establishes two hydrogen bonds with T483 side chain and F484 main chain amide. The β-phosphate establishes a total of five hydrogen bonds with S482 and T483 side chains, with R366 NE and NH1 side chain atoms and a solvent water molecule. Figure 4. Open in new tabDownload slide Structure of the FUT1:GDP:XLLG Ternary Complex. (A) Surface representation of FUT1 (colored as in Figure 1) in complex with GDP and XLLG. The binding sites are located on the luminal side of FUT1, opposite to the stem region. (B) View of the GDP ligand in the FUT1 donor binding pocket. The GDP electron density Fo-Fc map at 2.2 Å is represented in green mesh contoured at 3.0σ. Ligand and protein residues in stick representation colored yellow (GDP) or gray (protein) for carbon atoms, blue for nitrogen, red for oxygen, and orange for phosphorous. Water molecules are represented as red spheres. (C) Schematic representation of XLLG, a xylo-nonsaccharide acceptor, showing the β1,4-glucan backbone with side chains made of xylose or xylose-galactose. FUT1 transfers a fucose residue to the OH group of Gal3 (dashed red box). (D) View of the XLLG acceptor binding pocket. Ligand and protein residues in stick representation are colored as in (B). Residues annotated in orange are from the extra C-domain. The XLLG Fo-Fc electron density map at 2.2 Å is represented in green mesh contoured at 3.0σ. Water molecules are represented as red spheres. Figure 4. Open in new tabDownload slide Structure of the FUT1:GDP:XLLG Ternary Complex. (A) Surface representation of FUT1 (colored as in Figure 1) in complex with GDP and XLLG. The binding sites are located on the luminal side of FUT1, opposite to the stem region. (B) View of the GDP ligand in the FUT1 donor binding pocket. The GDP electron density Fo-Fc map at 2.2 Å is represented in green mesh contoured at 3.0σ. Ligand and protein residues in stick representation colored yellow (GDP) or gray (protein) for carbon atoms, blue for nitrogen, red for oxygen, and orange for phosphorous. Water molecules are represented as red spheres. (C) Schematic representation of XLLG, a xylo-nonsaccharide acceptor, showing the β1,4-glucan backbone with side chains made of xylose or xylose-galactose. FUT1 transfers a fucose residue to the OH group of Gal3 (dashed red box). (D) View of the XLLG acceptor binding pocket. Ligand and protein residues in stick representation are colored as in (B). Residues annotated in orange are from the extra C-domain. The XLLG Fo-Fc electron density map at 2.2 Å is represented in green mesh contoured at 3.0σ. Water molecules are represented as red spheres. Even though FUT1 and its structural homologs FUT8, NodZ, POFUT1, and POFUT2 share <20% overall sequence identity, comparison of these structures revealed the same overall GDP-fucose binding mode with identical or similar residues in the same spatial location in their C-domains (Supplemental Figure 3). The most conserved residues belong to three peptide motifs (denoted as I, II, and III) that have been previously identified in fucosyltransferases belonging to five different CAZy families (GT11, GT23, GT37, GT65, and GT68) (Chazalet et al., 2001; Martinez-Duncker et al., 2003). The importance of some of these residues for activity has been demonstrated by mutational analyses for several enzymes. Of paramount importance is the invariant arginine residue (R366 in FUT1, motif I), located at the Cβ1-Cα2 loop in FUT1, which interacts in a conserved manner with the β-phosphate of GDP. Its mutation in alanine (or lysine) abolished enzyme activity of FUT8 (Takahashi et al., 2000), NodZ (Chazalet et al., 2001), POFUT1 (Lira-Navarrete et al., 2011), and POFUT2 (Chen et al., 2012), thus highlighting its crucial role in donor sugar binding. Four other amino acids are similarly positioned in the GDP/GDP-fucose binding pocket in all fucosyltransferase structures, corresponding to S482, T483, F484, and E466 in FUT1. The [482-S-T-F] sequence, located at the beginning of Cα5, is part of motif III, whose function seems to be the correct positioning of pyrophosphate and fucose moieties for the transfer reaction. The acidic position equivalent to E466 was also shown to be critical for activity in FUT8 and NodZ (Ihara et al., 2007; Chazalet et al., 2001). The most conserved residues in these peptide motifs are therefore key anchoring points for the guanine and diphosphate moieties (Supplemental Figure 3). It is worth noting that the three peptide motifs are also observed in other Arabidopsis FUT proteins of family GT37 (Figure 5A). However, the FUT3 sequence lacks the conserved arginine residue in motif I, and FUT9 lacks motif II. Figure 5. Open in new tabDownload slide Multiple Sequence Alignment of Arabidopsis FUT Protein Sequences of Family GT37. (A) Amino acid composition of conserved peptide motifs I, II, and III in FUT1 to FUT10 sequences. Residues in bold red in FUT1 are involved in GDP binding. FUT3 sequence lacks the conserved arginine variant in motif I. FUT9 lacks motif II. (B) All proteins except FUT3 contain the extra C-terminal domain with its four conserved cysteines (yellow shading). Residues in bold blue are involved in XLLG acceptor binding. FUT3 harbors a truncated extra C-domain. (C) Amino acid conservation of catalytic residues in FUT sequences. Figure 5. Open in new tabDownload slide Multiple Sequence Alignment of Arabidopsis FUT Protein Sequences of Family GT37. (A) Amino acid composition of conserved peptide motifs I, II, and III in FUT1 to FUT10 sequences. Residues in bold red in FUT1 are involved in GDP binding. FUT3 sequence lacks the conserved arginine variant in motif I. FUT9 lacks motif II. (B) All proteins except FUT3 contain the extra C-terminal domain with its four conserved cysteines (yellow shading). Residues in bold blue are involved in XLLG acceptor binding. FUT3 harbors a truncated extra C-domain. (C) Amino acid conservation of catalytic residues in FUT sequences. Acceptor Substrate Binding Site The ternary complex FUT1:GDP:XLLG provides insight into the acceptor binding site of the enzyme. We obtained this complex by soaking apo-FUT1 crystals with both ligands. Clear electron density for the bound GDP was observed in each of the four monomers in the asymmetric unit. By contrast, electron density for the bound XLLG molecule was visible in only two monomers (B and C). This could be attributed to crystallographic symmetry constraints. As shown in Supplemental Figure 4, access to the XLLG binding site in monomers A and D is obstructed by contacts with symmetry-related molecules. Therefore, it seems plausible to consider that both acceptor binding sites are accessible in the functional dimer. The oligosaccharide XLLG can be considered as one of the minimal acceptor structures for FUT1. It is composed of a β-1,4-d-glucan backbone with branches made of Xyl and Xyl-Gal. For clarity, all sugars have been labeled as indicated in Figure 4C. The FUT1 enzyme transfers a fucose selectively to the O2 position of Gal3 in XLLG, forming a α1-2 linkage in the resulting product XLFG. The XLLG binding site is solvent exposed and, like the donor binding pocket, it locates opposite to the stem region (Figure 4A). Several water-mediated and direct interactions are established between the XLLG ligand and FUT1 residues from the N-domain and extra C-terminal region. The oligosaccharide is held in place mainly by contacts with Xyl1 and Gal3 moieties (Figure 4D). Although most of the acceptor monosaccharides could be accurately modeled in the electron density maps of two FUT1 monomers, the Xyl2 and Gal2 moieties are solvent exposed and do not fit well in density, thus reflecting the flexibility for this part of the ligand. As illustrated in Figure 4C, these two residues point in opposite directions from Xyl1 and Xyl3-Gal3. The O3 and O4 atoms of the Xyl1 moiety interact with the K278 NZ atom. Furthermore, Xyl1 O4 is also involved in H-bond formation with P525 backbone amide, while the O6 atom interacts with the side chain of S524. The W553 residue stacks against the Glc1 ring and also hydrogen bonds to the O3 atom of Glc2 via a water molecule. The R501 side chain interacts both with Xyl2 O4 and Glc3 O2 atoms. This protein residue locates in the Cβ5-Cβ6 loop that moves upon substrate binding. The Gal2, Xyl3, and Glc4 sugar residues do not contact protein residues, directly or via solvent interactions. By contrast, all the hydroxyl groups of the Gal3 moiety are involved in protein and/or solvent interactions. The C6 hydroxyl establishes two hydrogen bonds with S524 OG atom and with N301 side chain amine. The C4 hydroxyl establishes three H-bonds, with the H523 side chain, the W481 main chain amide via a solvent molecule, and the Glc2 O2 atom, in an intramolecular water-mediated H-bonding. Two solvent-mediated contacts are established between the Gal3 C3 hydroxyl group and the main chain carbonyl of D300 and W481. Moreover, it hydrogen bonds with the N184 side chain that also interacts with the O2 atom of Gal3. Curiously, N184 tightly interacts with the S180 side chain, which is an outlier in the Ramachandran plot (S180 is also an outlier in the apo-FUT1 structure). The FUT1:GDP:XLLG structure provides insight into the acceptor specificity of FUT1. The data clearly demonstrate that FUT1 can accommodate both XLLG and XXLG oligosaccharide acceptors, as it was previously shown for the pea FUT1 enzyme (Faik et al., 2000), and explain why only Gal3 (but not Gal2) in XLLG/XXLG can be fucosylated, thus forming the XLFG and XXFG products. Previous studies highlighted the strong preference of FUT1 for longer chains of XyG as acceptor (Faik et al., 2000). One can anticipate from our structural data that a polymeric acceptor substrate, which is also a multivalent ligand, would make additional contacts with the FUT1 enzyme and/or that occupancy of both acceptor binding sites in the dimer by a single molecule would create a cluster glycoside effect, thus resulting in a much higher affinity of FUT1 for the XyG polymer (Figure 3). As shown in Figure 4D, the acceptor is held in place through contacts with amino acid residues of the N-domain and of the extra C-terminal region. As described above, the latter region, which is unique to FUT1, is characterized by the presence of two β-hairpins and four conserved cysteines that are engaged in disulfide bonds (Figure 1). In our structure, the sequence motif [523-H-S-P] appears to contribute to the correct anchoring of XLLG/XXLG XyG subunits. Interestingly, this extra C-terminal region is also present in the nine other Arabidopsis FUT sequences of family GT37 (Figure 5B). The extent of overall sequence identity to FUT1 polypeptide sequence ranges from ∼38 to 55% (Sarria et al., 2001). In Arabidopsis, FUT1 is currently considered the sole fucosyltransferase active on XyG. The precise biochemical function of the other FUT proteins remains largely unknown, except for the Arabidopsis FUT4 and FUT6 proteins, which were shown to be α2-fucosyltransferases acting on arabinogalactan proteins (Wu et al., 2010). If we hypothesize that the extra C-terminal region is similarly involved in acceptor binding in other FUT proteins, then three subgroups can be defined on the sole basis of their corresponding [523-H-S-P] sequences: FUT2 and FUT10 are the closest to FUT1 and can be grouped together; FUT4 to FUT7 form another subgroup; FUT8 and FUT9 can also be grouped, although FUT9 lacks the peptide motif II (Figure 5A). FUT3 appears as the most atypical member and may not even be a fucosyltransferase, since it lacks the invariant catalytic arginine residue and harbors a truncated C-terminal region. We anticipate that this grouping may help to decipher the precise biochemical function of the other plant FUT members. Interestingly, an Arabidopsis mutant named mur2 harboring a point mutation in the FUT1 sequence has been previously described (Vanzin et al., 2002). The mur2 mutation, D550N, which abolished enzyme activity as compared with the wild-type enzyme, is located in the extra C-terminal region and may thus affect the acceptor binding site. However, this mutation also creates an additional N-glycosylation site and one cannot exclude that occupancy of this site by a N-glycan would impair the correct folding of the extra C-terminal region and consequently the acceptor binding site. Proposed Reaction Mechanism for FUT1 Even though no FUT1:GDP-fucose structure is available at the moment, comparison of the FUT1:GDP:XLLG structure with the closest homologs POFUT1 (Lira-Navarrete et al., 2011) and POFUT2 (Chen et al., 2012), for which a complex with GDP-fucose is available, allowed modeling of the fucose residue in the binding site (Figure 6A). Due to the close proximity of Gal3 of the acceptor substrate and the pyrophosphate moiety of the GDP, and considering that the fucose anomeric carbon has to locate near the Gal3 C2 atom for the transfer reaction to occur, the positioning of the fucose ring was straightforward. The side chains of F368 and W481, together with the Glc3 and Glc4 acceptor sugar residues, create a cavity (just below Gal3) that can easily accommodate the fucose ring. The protein-ligand interactions observed for the guanosine moiety are essentially the same in both models (GDP versus GDP-fucose), but the position of the pyrophosphate groups is slightly different, and new H-bonds are established (Figure 6A; Supplemental Figure 5). In the GDP-fucose model, the α-phosphate now interacts with G183 main-chain amide and T483 side chain, while the β-phosphate only contacts the T483 side-chain and main-chain amide, and the R366 NE atom. The fucose moiety establishes two hydrogen bonds with the N184 side-chain and the acceptor Xyl3 O3 atom. Interestingly, the presence of the fucose induces a movement of Gal3 moiety upwards close to the D300 side chain (Supplemental Figure 5). Figure 6. Open in new tabDownload slide Proposed Reaction Mechanism for FUT1. (A) The active site of FUT1 in complex with GDP-fucose and XLLG. The fucose moiety was modeled based on the structure of the POFUT1 homolog. Protein residues and ligands are displayed as sticks, colored red for oxygen, blue for nitrogen, orange for phosphorous, and gray, green, or yellow for amino acid, GDP-fucose, or XLLG carbon atoms, respectively. Protein-ligand interactions are depicted as black dashes. (B) FUT1 activity assays of mutants compared with wild-type enzyme (average ± sd, n = 3). (C) Proposed reaction mechanism for FUT1. The enzyme reaction is proposed to occur through a single-displacement SN2 mechanism via an oxocarbenium ion-like transition state assisted by a catalytic base, D300. Figure 6. Open in new tabDownload slide Proposed Reaction Mechanism for FUT1. (A) The active site of FUT1 in complex with GDP-fucose and XLLG. The fucose moiety was modeled based on the structure of the POFUT1 homolog. Protein residues and ligands are displayed as sticks, colored red for oxygen, blue for nitrogen, orange for phosphorous, and gray, green, or yellow for amino acid, GDP-fucose, or XLLG carbon atoms, respectively. Protein-ligand interactions are depicted as black dashes. (B) FUT1 activity assays of mutants compared with wild-type enzyme (average ± sd, n = 3). (C) Proposed reaction mechanism for FUT1. The enzyme reaction is proposed to occur through a single-displacement SN2 mechanism via an oxocarbenium ion-like transition state assisted by a catalytic base, D300. FUT1 and its structural homologs are inverting glycosyltransferases. Inverting GT reactions are suggested to occur in a single displacement SN2 mechanism via an oxocarbenium ion-like transition state assisted by a catalytic base (Lairson et al., 2008; Breton et al., 2012). An aspartate, glutamate, or histidine usually serves as a catalytic base to deprotonate the nucleophile hydroxyl group of the acceptor substrate. In POFUT2, a conserved glutamate residue (E54) in the N-domain is appropriately positioned to be proposed as the catalytic base (Supplemental Figure 6A) (Chen et al., 2012). In addition, its mutation to alanine totally abolished enzyme activity, which is consistent with a role as a general base. Contrary to what was expected for an inverting GT, no catalytic base could be identified in the active site of POFUT1. Instead, mutagenesis data pointed to the catalytic importance of two amino acid residues (R240 and N43 in POFUT1), and a SN1-like mechanism has been proposed for this protein-O-fucosyltransferase 1 (Supplemental Figure 6B) (Lira-Navarrete et al., 2011). In this model, the cleavage of the glycosidic bond would take place first, leading to the formation of an intimate ion pair in the transition state. A similar mechanism, intermediary between SN1 and SN2 reaction, has been proposed for the human α3-fucosyltransferase V belonging to family GT11 (Murray et al., 1997). Exploration of the FUT1 catalytic pocket highlighted several residues of the N-domain that, in addition to the invariant arginine residue (R366), are potentially important for catalysis (Figure 6A). Mutagenesis studies revealed the critical importance of N184 (Figure 6B), a residue strictly conserved in other FUT members of family GT37 (Figure 5C), and which is similarly positioned as N43 in POFUT1 (Supplemental Figure 6B). Point mutation of D300 to alanine also resulted in complete loss of enzyme activity (Figure 6B). In the GDP-fucose model, this acidic position (observed only in FUT1 and FUT3 and replaced by an asparagine in other FUT proteins) is at 3.4 Å from the O2 atom of Gal3 and therefore may play a role as a general base. Point mutations of S180 and H271 residues severely impaired enzyme activity (the S180A and H271A mutants retained 6.8 and 2.3% activity, respectively; Figure 6B). The histidine residue is conserved in all FUT proteins except FUT3, whereas the serine residue is only present in FUT1 and FUT2. In addition, H271 closely interacts with D300 side chain and thus most likely participates in charge relay. The hydrogen bond network between D300, H271, S180, and N184 together with R366 contributes to create the catalytic pocket of FUT1 (Figure 6A). Particularly, S180 appears to be important for the correct positioning of N184 to establish hydrogen bonds with O2 and O3 of Gal3. Curiously, S180 perfectly superimposes with the proposed catalytic glutamate residue (E54) in POFUT2. Finally, it is clear that FUT1 shares common catalytic features with both POFUT1 and POFUT2 as illustrated in Supplemental Figure 6. The presence of a catalytic base (D300) correctly positioned to abstract the proton of the nucleophile hydroxyl group of the acceptor in the FUT1 active site is in favor of the general SN2 reaction of inverting GTs observed in POFUT2 (Figure 6C). However, future (mutagenesis) experiments will be necessary to gain further insights into substrate/acceptor binding behavior and reaction mechanism of FUT1. The structures reported here show how FUT1 recognizes its donor and acceptor substrates and provide information on the reaction mechanism and selectivity of the enzyme. They also suggest that FUT1 is organized as a functional dimer in the Golgi apparatus, which may explain its strong preference for polymeric acceptor substrates. Further studies exploring the dynamic properties of FUT1 with longer acceptor molecules are needed to progress in our understanding of the XyG fucosylation process. METHODS DNA Manipulations and Protein Production The DNA fragment encoding the stem and catalytic regions (corresponding to residues 69 to 558 of FUT1 protein sequence) was cloned into the pVT-Bac-His1 transfer vector, generating pVT-Bac-His1-FUT1Δ68, as previously described (Cicéron et al., 2016). This plasmid harbors the fut1 gene and is preceded by (His)6 and X-press tags used for purification purposes, originating the FUT1Δ68 protein, named simply FUT1 for clarity. The mutants were generated according to QuikChange II site-directed mutagenesis method (Agilent Technologies), using pVT-Bac-His1-FUT1Δ68 as template, and sense and antisense primers listed in Supplemental Table 3. Mutants were systematically checked by double-strand sequencing. Each construct was cotransfected with Baculogold AcNPV DNA into Sf9 insect cells (Baculovirus Expression Vector System; Pharmingen) and incubated for 72 h at 27°C. A high-titer virus stock solution (>108 pfu/mL) was used for the infection of High Five cells (Invitrogen). The cells were grown in Erlen flasks in the serum-free EXCELL405 medium (Sigma-Aldrich) at 27°C, in a shaking incubator at 120 rev/min, for 4 d. Cells and impurities were removed by centrifugation at 13,000g for 30 min and the medium supernatant containing secreted FUT1 was injected through a histidine tag purification column (cOmplete His-Tag; Roche), followed by washing in 25 mM HEPES buffer, pH 7.4, 500 mM NaCl, and 50 mM imidazole. Protein was eluted using an imidazole step gradient (up to 500 mM). Fractions containing FUT1 protein were pooled and concentrated, further purified by size-exclusion chromatography using a Superdex 200 10/300 GL (GE Healthcare) column and buffer exchanged to 25 mM HEPES buffer, pH 7.4, and 150 mM NaCl. The pure protein was concentrated up to 7 mg/mL and stored at −20°C until further use. FUT1 behaves as a dimer in solution and harbors two N-glycans as verified by size-exclusion chromatography, dynamic light scattering, and MALDI-TOF analyses. FUT1 Enzyme Activity Fucosyltransferase activity toward tamarind XyG was determined as follows. Enzyme assays were carried in 10 mM HEPES-KOH buffer, pH 7, in a final volume of 100 μL, containing FUT1 (20 nM), 30 μM unlabeled GDP-fucose, 1.5 μM (20 nCi, 45,000 cpm) GDP-[14C]-fucose, and xyloglucan acceptor (from tamarind seeds; Megazyme) at 2.5 mg/mL. Reactions were incubated at 30°C for 30 min and 120 min and terminated by addition in the reaction mix of 1 mL anion exchange AG 1x8 resin (1 g resin in 4 mL double distilled water; 200 to 400 mesh; Bio-Rad), which aims at trapping all nucleotides (GDP and unreacted GDP-fucose). Reaction mix was then centrifuged 2 min at 13,000g and 600 μL supernatant collected. This last step was repeated once and the two supernatants pooled and placed into a counting flask of 5 mL to measure radioactivity incorporation into the xyloglucan polymer. Scintillation liquid (3.8 mL) was added to the flask and flasks were counted for 1 min using a Tri-carb 1600TR Packard Bell instrument (Perkin-Elmer). Control reactions were performed in the absence of acceptor. Preparation and Characterization of XLLG The XLLG acceptor substrate used for soaking experiments was prepared as previously described (Cicéron et al., 2016). Briefly, 100 mg pure XyG from tamarind seeds in 1 mL 10 mM sodium acetate were treated with 5 units of EG II from Trichoderma reseii (Megazyme) for 18 h at 35°C. After heat inactivation of EG II (30 min at 70°C), the reaction mix was centrifuged for 5 min at 13,000g and applied to a syringe filter (0.2 µm from Nalgene). XyG-derived oligosaccharides were then loaded on a HW-40 gel permeation column (1 m × 15 mm), conditioned with 0.1 M ammonium carbonate. Refractive index of eluted solvent was monitored and 1-mL fractions were collected. Three main peaks were identified on the chromatogram and fractions from these peaks were characterized using MALDI-TOF (Autoflex; Bruker) in reflectron mode according to Lerouxel et al. (2002). Two microliters of fractions containing XyG-derived oligosaccharides were mixed in a 1:1 ratio using a dihydroxybenzoic acid matrix at 2 mg/mL (double distilled water/acetonitrile/trifluoracetic acid, 70/30/0.1%). After identification, fraction #23 containing pure XLLG (peak at m/z = 1409) (Supplemental Figure 7) was collected and lyophilized for experiments. Crystallization, Diffraction Data Collection, and Processing Native FUT1 crystallization and crystal characterization are described elsewhere (Rocha et al., 2016). Suitable x-ray diffracting crystals appeared within 3 to 5 days in 100 mM HEPES buffer solutions, pH 7.5 to 7.8, with either 300 to 500 mM NaCl and 18 to 20% PEG 8000 or 300 to 500 mM NaCH3COO and 20 to 22% PEG 4000. Crystals were cryoprotected in mother liquor solution supplemented with 15% (v/v) ethylene glycol, prior to flash-cooling and storage in liquid nitrogen. The FUT1-ligand complex crystals were obtained after native crystals were soaked in crystallization solutions containing either GDP-fucose or GDP + xyloglucan oligosaccharide acceptor (XLLG). The best FUT1:GDP-fucose protein-ligand data set was obtained by short incubation (∼3 min) of the FUT1 native crystal in a cryoprotective solution of 100 mM HEPES, pH 7.5, 250 mM NaCl, 18% PEG 8000, and 15% (v/v) ethylene glycol supplemented with 5 mM GDP-fucose. For the soaking experiments with GDP and XLLG ligands, the best data set was obtained after ∼25 min incubation of the native crystal in the above solution supplemented with 10 mM GDP and 7 mg⋅mL−1 XLLG oligosaccharide. Crystals were directly plunged and stored in liquid nitrogen. All diffraction data were measured at ID29 (de Sanctis et al., 2012) or ID23-2 (Flot et al., 2010) beamlines at the European Synchrotron Radiation Facility (ESRF, Grenoble, France). Diffraction images were processed with XDS (Kabsch, 2010), scaled and merged with AIMLESS (Evans and Murshudov, 2013), and the corresponding structure factors were calculated using cTRUNCATE (Padilla and Yeates, 2003) from the CCP4 suite (Winn et al., 2011). Structure Determination and Refinement The crystal structure of native enzyme (apo-FUT1) was determined as described previously (Rocha et al., 2016). Briefly, x-ray diffraction data up to 2.9 Å resolution was collected on a single tantalum bromide (Ta6Br12) derivatized crystal, and initial phases were determined at 4.0 Å using the 2W-MAD implemented method in SHARP (Bricogne et al., 2003). Phases were further improved with the program RESOLVE (Terwilliger, 2000) using the automatically implemented 4-fold noncrystallographic symmetry averaging, solvent flattening, and phase extension procedures up to 2.9 Å resolution. The RESOLVE output model was further inspected and edited in COOT (Emsley and Cowtan, 2004) and used as search template for molecular replacement, using a 2.1 Å resolution native data set with the program PHASER (McCoy et al., 2007). Iterative model building and refinement were carried with COOT and PHENIX (Adams et al., 2010) until model completion. The FUT1:GDP:XLLG structure was obtained after rigid body refinement with REFMAC5 (Murshudov et al., 2011) of the apo-FUT1 monomers against the 2.2 Å resolution FUT1:GDP:XLLG data set. Readily interpretable electron density maps were produced and clear well-defined residual density was visible for a GDP molecule nestled in all four molecules. Extra electron density that could accommodate the XLLG acceptor was also found in the residual maps for two protein monomers. The phases for the FUT1:GDP-fucose structure were also retrieved by rigid body refinement of apo-FUT1 with the 2.5 Å resolution protein-ligand data set. Inspection of the initial electron density maps revealed electron density that could accommodate only the GDP molecule. Since no density was found for the fucose, and the GDP ligand occupied the same position in the higher resolution FUT1:GDP:XLLG model, no further analysis of the FUT1:GDP-fucose structure was performed. Further steps of protein-ligand model refinement were performed with COOT and PHENIX. The molecular stereochemistry of FUT1 structures was evaluated using the program MOLPROBITY (Chen et al., 2010). Intermolecular contacts and crystallographic packing were retrieved with PISA (Krissinel and Henrick, 2007), and structural comparisons among the models were performed with SSM (Krissinel and Henrick, 2004). Images were prepared with PYMOL (The PyMOL Molecular Graphics System, version 1.7.4, Schrödinger LLC). Crystal Characterization and Validation of FUT1 Models Upon successive crystal optimizations and data collection sessions, data sets of apo-FUT1 and FUT1:GDP:XLLG ternary complex reached final resolutions of 2.1 and 2.2 Å, respectively. Detailed crystal and diffraction data statistics are given in Supplemental Table 1. Crystals belonged to space group P21 with similar cell dimensions and contained four independent molecules in the asymmetric unit, which were organized into two distinct dimeric structures designated as A-B and C-D (Supplemental Figure 2). Overall, the entire length of the peptide chains showed good electron density, except for the small missing segments located at the N- and the C-domains (Nα3-Nβ1, Nα5-Nβ3, Cα2-Cβ2, and Cβ3-Cα4 loops) or some residue side-chains exposed to the solvent that did not show well-defined density for all the atoms. Globally, the apo-FUT1 model contained 1809 out of the 2084 residues (4 chains × 31 residue tag plus 490 residues). In all four monomers, the N-terminal 31-residue tag was not visible and thus was not included in the model. Also, some loop regions (namely, S69-R79, D257-G260, V399-N406, and G450-K457 for chain A; S69-S94, E163-G167, D257-E259, H404-N406, and G450-K457 for chain B; S69-N93, T258-E259, E400-N406, and G450-K457 for chain C; S69-A83, D166-D168, I256-E259, H404-N406, and G450-K457 for chain D) were not visible in the electron density maps and therefore were omitted in the final model. When appropriate, ions and solvent molecules were added to the model. The apo-FUT1 was refined to R work/R free of 17.73/21.51% and the final model contained 96.7, 3.1, and 0.2% of the residues within the preferred, allowed, or outlier regions of the Ramachandran diagram (Ramachandran et al., 1963), respectively (see Supplemental Table 2 for refinement and model quality statistics). The FUT1:GDP:XLLG model contained 1801 out of the 2084 total residues. Similarly to the apo-FUT1 structure, some residues were not visible in electron density maps and were omitted from the model: the initial 31-residue tag for each chain, and S69-N91, D166-G167, D257-G260, V399-H404, and Q452-K456 for chain A; S69-S94, E163-G167, D257-E259, H404-N406, and T454-K456 for chain B; S69-N93, D257-E259, E400-N406, and Y451-K456 for chain C; S69-N93, D166-D168, I256-G260, H404-N406, and E455-K456 for chain D. Moreover, four GDP ligands (one per monomer) and two XLLG molecules (one per dimer) were modeled into well-defined electron density. Several solvent molecules and ions were also added to the model (see Supplemental Table 2 for further details). The FUT1:GDP:XLLG final model refined to R work/R free values of 17.90/21.49% contained 96.6, 3.2, and 0.2% of residues within the preferred, allowed, and outlier regions of the Ramachandran plot, respectively. Modeling of GDP-Fucose The structure of the GDP-fucose was built by superposition of the GDP structure observed in FUT1 and the structure of the GDP-fucose observed in POFUT1 (Lira-Navarrete et al., 2011). Only atoms of the diphosphate bridges were aligned to add the fucose ring to GDP, so that the GDP conformation and position were maintained in the complex. Only minimal modifications were made afterwards to maintain the features of the crystal structure. Hydrogens were added using the Schrödinger Suite (http://www.schrodinger.com). Steric clashes between the fucose ring and few surrounding residues of both the protein and the XyG acceptor were solved by energy minimization using the Sybyl-X suite (http://tripos.com). Standard parameters were used for energy calculation, including parameters from the TRIPOS force field. The final model was analyzed through comparison of the active sites of the four monomers (A, B, C, and D) in the AU. The active sites of monomers B and C (with a bound acceptor molecule) were considered for mechanistic consideration of FUT1 enzyme. Accession Numbers Sequence data from this article can be found the UniProt KB database under the following accession numbers: Q9SWH5, O81053, Q9CAZ1, Q9SJP2, Q9SJP4, Q9XI80, Q9XI81, Q9XI78, Q9XI77, and Q9SJP6 for FUT1 to FUT10, respectively. The structures of apo-FUT1 and FUT1:GDP:XLLG have been deposited in the Protein Data Bank (PDB) under accession numbers 5KOP and 5KOR, respectively. The PDB codes of POFUT1 and POFUT2 structures used in this work are 3ZY5 and 4AP6, respectively. Supplemental Data Supplemental Figure 1. Overall Architecture of the Classical GT-B Fold and Variants in Crystal Structures of Fucosyltransferases. Supplemental Figure 2. Asymmetric Unit Composition of FUT1 Crystals. Supplemental Figure 3. Structure-Based Sequence Alignment of the Most Conserved Regions in FUT1 and Its Closest Structural Homologs. Supplemental Figure 4. View of the Active Sites in Monomers A and B of Apo-Fut1. Supplemental Figure 5. Superimposition of the FUT1:GDP:XLLG Crystallographic Structure with the Modeled GDP-Fucose. Supplemental Figure 6. Comparison of the Active Sites of FUT1 with POFUT1 and POFUT2. Supplemental Figure 7. MALDI-TOF Analysis of XLLG Preparation Supplemental Table 1. Data Collection and Crystallographic Statistics for FUT1 Crystals. Supplemental Table 2. Refinement and Model Quality Statistics for FUT1 Structures. Supplemental Table 3. Oligonucleotides Used for PCR Mutagenesis. Acknowledgments We thank Ken Keegstra for his critical reading of the manuscript and insightful comments. This work was supported by the CNRS and the University Grenoble Alpes. F.C. received a PhD fellowship from UGA. Access to the ESRF Structural Biology beamlines is gratefully acknowledged. Glossary XyG xyloglucan GT glycosyltransferase AUTHOR CONTRIBUTIONS C.B. and O.L. conceived the project and designed research. J.R., F.C., D.d.S., M.L., V.C., and O.L. performed research. C.B., J.R., and O.L. discussed results. C.B. and J.R. wrote the manuscript. References 1. Adams , P.D. , et al. . ( 2010 ). PHENIX: a comprehensive Python-based system for macromolecular structure solution . Acta Crystallogr. D Biol. Crystallogr. 66 : 213 – 221 . Google Scholar Crossref Search ADS PubMed WorldCat 2. Breton , C. , Fournel-Gigleux , S. , Palcic , M.M. ( 2012 ). Recent structures, evolution and mechanisms of glycosyltransferases . Curr. Opin. Struct. Biol. 22 : 540 – 549 . Google Scholar Crossref Search ADS PubMed WorldCat 3. Breton , C. , Snajdrová , L. , Jeanneau , C. , Koca , J. , Imberty , A. ( 2006 ). Structures and mechanisms of glycosyltransferases . Glycobiology 16 : 29R – 37R . Google Scholar Crossref Search ADS PubMed WorldCat 4. Bricogne , G. , Vonrhein , C. , Flensburg , C. , Schiltz , M. , Paciorek , W. ( 2003 ). Generation, representation and flow of phase information in structure determination: recent developments in and around SHARP 2.0 . Acta Crystallogr. D Biol. Crystallogr. 59 : 2023 – 2030 . Google Scholar Crossref Search ADS PubMed WorldCat 5. Brzezinski , K. , Dauter , Z. , Jaskolski , M. ( 2012 ). Structures of NodZ α1,6-fucosyltransferase in complex with GDP and GDP-fucose . Acta Crystallogr. D Biol. Crystallogr. 68 : 160 – 168 . Google Scholar Crossref Search ADS PubMed WorldCat 6. Brzezinski , K. , Stepkowski , T. , Panjikar , S. , Bujacz , G. , Jaskolski , M. ( 2007 ). High-resolution structure of NodZ fucosyltransferase involved in the biosynthesis of the nodulation factor . Acta Biochim. Pol. 54 : 537 – 549 . Google Scholar Crossref Search ADS PubMed WorldCat 7. Burton , R.A. , Fincher , G.B. ( 2014 ). Plant cell wall engineering: applications in biofuel production and improved human health . Curr. Opin. Biotechnol. 26 : 79 – 84 . Google Scholar Crossref Search ADS PubMed WorldCat 8. Cavalier , D.M. , Keegstra , K. ( 2006 ). Two xyloglucan xylosyltransferases catalyze the addition of multiple xylosyl residues to cellohexaose . J. Biol. Chem. 281 : 34197 – 34207 . Google Scholar Crossref Search ADS PubMed WorldCat 9. Chazalet , V. , Uehara , K. , Geremia , R.A. , Breton , C. ( 2001 ). Identification of essential amino acids in the Azorhizobium caulinodans fucosyltransferase NodZ . J. Bacteriol. 183 : 7067 – 7075 . Google Scholar Crossref Search ADS PubMed WorldCat 10. Chen , C.-I. , Keusch , J.J. , Klein , D. , Hess , D. , Hofsteenge , J. , Gut , H. ( 2012 ). Structure of human POFUT2: insights into thrombospondin type 1 repeat fold and O-fucosylation . EMBO J. 31 : 3183 – 3197 . Google Scholar Crossref Search ADS PubMed WorldCat 11. Chen , V.B. , Arendall III, W.B. , Headd , J.J. , Keedy , D.A. , Immormino , R.M. , Kapral , G.J. , Murray , L.W. , Richardson , J.S. , Richardson , D.C. ( 2010 ). MolProbity: all-atom structure validation for macromolecular crystallography . Acta Crystallogr. D Biol. Crystallogr. 66 : 12 – 21 . Google Scholar Crossref Search ADS PubMed WorldCat 12. Chou , Y.H. , Pogorelko , G. , Young , Z.T. , Zabotina , O.A. ( 2015 ). Protein-protein interactions among xyloglucan-synthesizing enzymes and formation of Golgi-localized multiprotein complexes . Plant Cell Physiol. 56 : 255 – 267 . Google Scholar Crossref Search ADS PubMed WorldCat 13. Cicéron , F. , Rocha , J. , Kousar , S. , Hansen , S.F. , Chazalet , V. , Gillon , E. , Breton , C. , Lerouxel , O. ( 2016 ). Expression, purification and biochemical characterization of AtFUT1, a xyloglucan-specific fucosyltransferase from Arabidopsis thaliana. Biochimie pii : S0300-9084(16)30160-2 . Google Scholar OpenURL Placeholder Text WorldCat 14. Cocuron , J.C. , Lerouxel , O. , Drakakaki , G. , Alonso , A.P. , Liepman , A.H. , Keegstra , K. , Raikhel , N. , Wilkerson , C.G. ( 2007 ). A gene from the cellulose synthase-like C family encodes a beta-1,4 glucan synthase . Proc. Natl. Acad. Sci. USA 104 : 8550 – 8555 . Google Scholar Crossref Search ADS WorldCat 15. Culbertson , A.T. , Chou , Y.-H. , Smith , A.L. , Young , Z.T. , Tietze , A.A. , Cottaz , S. , Fauré , R. , Zabotina , O.A. ( 2016 ). Enzymatic Activity of Xyloglucan Xylosyltransferase 5 . Plant Physiol. 171 : 1893 – 1904 . Google Scholar Crossref Search ADS PubMed WorldCat 16. de Sanctis , D. , et al. . ( 2012 ). ID29: a high-intensity highly automated ESRF beamline for macromolecular crystallography experiments exploiting anomalous scattering . J. Synchrotron Radiat. 19 : 455 – 461 . Google Scholar Crossref Search ADS PubMed WorldCat 17. Emsley , P. , Cowtan , K. ( 2004 ). Coot: model-building tools for molecular graphics . Acta Crystallogr. D Biol. Crystallogr. 60 : 2126 – 2132 . Google Scholar Crossref Search ADS PubMed WorldCat 18. Evans , P.R. , Murshudov , G.N. ( 2013 ). How good are my data and what is the resolution? Acta Crystallogr. D Biol. Crystallogr. 69 : 1204 – 1214 . Google Scholar Crossref Search ADS PubMed WorldCat 19. Faik , A. , Bar-Peled , M. , DeRocher , A.E. , Zeng , W. , Perrin , R.M. , Wilkerson , C. , Raikhel , N.V. , Keegstra , K. ( 2000 ). Biochemical characterization and molecular cloning of an alpha-1,2-fucosyltransferase that catalyzes the last step of cell wall xyloglucan biosynthesis in pea . J. Biol. Chem. 275 : 15082 – 15089 . Google Scholar Crossref Search ADS PubMed WorldCat 20. Faik , A. , Price , N.J. , Raikhel , N.V. , Keegstra , K. ( 2002 ). An Arabidopsis gene encoding an alpha-xylosyltransferase involved in xyloglucan biosynthesis . Proc. Natl. Acad. Sci. USA 99 : 7797 – 7802 . Google Scholar Crossref Search ADS WorldCat 21. Flot , D. , et al. . ( 2010 ). The ID23-2 structural biology microfocus beamline at the ESRF . J. Synchrotron Radiat. 17 : 107 – 118 . Google Scholar Crossref Search ADS PubMed WorldCat 22. Fry , S.C. ( 1989 ). The structure and functions of xyloglucan . J. Exp. Bot. 40 : 1 – 11 . Google Scholar Crossref Search ADS WorldCat 23. Fry , S.C. , York , W.S. , Albersheim , P. ( 1993 ). An unambiguous nomenclature for xyloglucan-derived oligosaccharide . Physiol. Plant. 89 : 1 – 3 . Google Scholar Crossref Search ADS WorldCat 24. Holm , L. , Rosenström , P. ( 2010 ). Dali server: conservation mapping in 3D . Nucleic Acids Res. 38 : W545–W549 . Google Scholar Crossref Search ADS WorldCat 25. Hsieh , Y.S.Y. , Harris , P.J. ( 2009 ). Xyloglucans of monocotyledons have diverse structures . Mol. Plant 2 : 943 – 965 . Google Scholar Crossref Search ADS PubMed WorldCat 26. Ihara , H. , Ikeda , Y. , Toma , S. , Wang , X. , Suzuki , T. , Gu , J. , Miyoshi , E. , Tsukihara , T. , Honke , K. , Matsumoto , A. , Nakagawa , A. , Taniguchi , N. ( 2007 ). Crystal structure of mammalian alpha1,6-fucosyltransferase, FUT8 . Glycobiology 17 : 455 – 466 . Google Scholar Crossref Search ADS PubMed WorldCat 27. Jensen , J.K. , Schultink , A. , Keegstra , K. , Wilkerson , C.G. , Pauly , M. ( 2012 ). RNA-Seq analysis of developing nasturtium seeds (Tropaeolum majus): identification and characterization of an additional galactosyltransferase involved in xyloglucan biosynthesis . Mol. Plant 5 : 984 – 992 . Google Scholar Crossref Search ADS PubMed WorldCat 28. Kabsch , W. ( 2010 ). Integration, scaling, space-group assignment and post-refinement . Acta Crystallogr. D Biol. Crystallogr. 66 : 133 – 144 . Google Scholar Crossref Search ADS PubMed WorldCat 29. Kellokumpu , S. , Hassinen , A. , Glumoff , T. ( 2016 ). Glycosyltransferase complexes in eukaryotes: long-known, prevalent but still unrecognized . Cell. Mol. Life Sci. 73 : 305 – 325 . Google Scholar Crossref Search ADS PubMed WorldCat 30. Koropatkin , N.M. , Cameron , E.A. , Martens , E.C. ( 2012 ). How glycan metabolism shapes the human gut microbiota . Nat. Rev. Microbiol. 10 : 323 – 335 . Google Scholar Crossref Search ADS PubMed WorldCat 31. Krissinel , E. , Henrick , K. ( 2004 ). Secondary-structure matching (SSM), a new tool for fast protein structure alignment in three dimensions . Acta Crystallogr. D Biol. Crystallogr. 60 : 2256 – 2268 . Google Scholar Crossref Search ADS PubMed WorldCat 32. Krissinel , E. , Henrick , K. ( 2007 ). Inference of macromolecular assemblies from crystalline state . J. Mol. Biol. 372 : 774 – 797 . Google Scholar Crossref Search ADS PubMed WorldCat 33. Lairson , L.L. , Henrissat , B. , Davies , G.J. , Withers , S.G. ( 2008 ). Glycosyltransferases: structures, functions, and mechanisms . Annu. Rev. Biochem. 77 : 521 – 555 . Google Scholar Crossref Search ADS PubMed WorldCat 34. Lerouxel , O. , Choo , T.S. , Séveno , M. , Usadel , B. , Faye , L. , Lerouge , P. , Pauly , M. ( 2002 ). Rapid structural phenotyping of plant cell wall mutants by enzymatic oligosaccharide fingerprinting . Plant Physiol. 130 : 1754 – 1763 . Google Scholar Crossref Search ADS PubMed WorldCat 35. Lira-Navarrete , E. , Valero-González , J. , Villanueva , R. , Martínez-Júlvez , M. , Tejero , T. , Merino , P. , Panjikar , S. , Hurtado-Guerrero , R. ( 2011 ). Structural insights into the mechanism of protein O-fucosylation . PLoS One 6 : e25365 . Google Scholar Crossref Search ADS PubMed WorldCat 36. Loqué , D. , Scheller , H.V. , Pauly , M. ( 2015 ). Engineering of plant cell walls for enhanced biofuel production . Curr. Opin. Plant Biol. 25 : 151 – 161 . Google Scholar Crossref Search ADS PubMed WorldCat 37. Madson , M. , Dunand , C. , Li , X. , Verma , R. , Vanzin , G.F. , Caplan , J. , Shoue , D.A. , Carpita , N.C. , Reiter , W.D. ( 2003 ). The MUR3 gene of Arabidopsis encodes a xyloglucan galactosyltransferase that is evolutionarily related to animal exostosins . Plant Cell 15 : 1662 – 1670 . Google Scholar Crossref Search ADS PubMed WorldCat 38. Martinez-Duncker , I. , Mollicone , R. , Candelier , J.J. , Breton , C. , Oriol , R. ( 2003 ). A new superfamily of protein-O-fucosyltransferases, alpha2-fucosyltransferases, and alpha6-fucosyltransferases: phylogeny and identification of conserved peptide motifs . Glycobiology 13 : 1C – 5C . Google Scholar Crossref Search ADS PubMed WorldCat 39. McCoy , A.J. , Grosse-Kunstleve , R.W. , Adams , P.D. , Winn , M.D. , Storoni , L.C. , Read , R.J. ( 2007 ). Phaser crystallographic software . J. Appl. Cryst. 40 : 658 – 674 . Google Scholar Crossref Search ADS WorldCat 40. Murray , B.W. , Wittmann , V. , Burkart , M.D. , Hung , S.C. , Wong , C.H. ( 1997 ). Mechanism of human alpha-1,3-fucosyltransferase V: glycosidic cleavage occurs prior to nucleophilic attack . Biochemistry 36 : 823 – 831 . Google Scholar Crossref Search ADS PubMed WorldCat 41. Murshudov , G.N. , Skubák , P. , Lebedev , A.A. , Pannu , N.S. , Steiner , R.A. , Nicholls , R.A. , Winn , M.D. , Long , F. , Vagin , A.A. ( 2011 ). REFMAC5 for the refinement of macromolecular crystal structures . Acta Crystallogr. D Biol. Crystallogr. 67 : 355 – 367 . Google Scholar Crossref Search ADS PubMed WorldCat 42. Padilla , J.E. , Yeates , T.O. ( 2003 ). A statistic for local intensity differences: robustness to anisotropy and pseudo-centering and utility for detecting twinning . Acta Crystallogr. D Biol. Crystallogr. 59 : 1124 – 1130 . Google Scholar Crossref Search ADS PubMed WorldCat 43. Park , Y.B. , Cosgrove , D.J. ( 2015 ). Xyloglucan and its interactions with other components of the growing cell wall . Plant Cell Physiol. 56 : 180 – 194 . Google Scholar Crossref Search ADS PubMed WorldCat 44. Pauly , M. , Gille , S. , Liu , L. , Mansoori , N. , de Souza , A. , Schultink , A. , Xiong , G. ( 2013 ). Hemicellulose biosynthesis . Planta 238 : 627 – 642 . Google Scholar Crossref Search ADS PubMed WorldCat 45. Pauly , M. , Keegstra , K. ( 2016 ). Biosynthesis of the plant cell wall matrix polysaccharide xyloglucan . Annu. Rev. Plant Biol. 67 : 235 – 259 . Google Scholar Crossref Search ADS PubMed WorldCat 46. Peña , M.J. , Darvill , A.G. , Eberhard , S. , York , W.S. , O’Neill , M.A. ( 2008 ). Moss and liverwort xyloglucans contain galacturonic acid and are structurally distinct from the xyloglucans synthesized by hornworts and vascular plants . Glycobiology 18 : 891 – 904 . Google Scholar Crossref Search ADS PubMed WorldCat 47. Perrin , R.M. , DeRocher , A.E. , Bar-Peled , M. , Zeng , W. , Norambuena , L. , Orellana , A. , Raikhel , N.V. , Keegstra , K. ( 1999 ). Xyloglucan fucosyltransferase, an enzyme involved in plant cell wall biosynthesis . Science 284 : 1976 – 1979 . Google Scholar Crossref Search ADS PubMed WorldCat 48. Perrin , R.M. , Jia , Z. , Wagner , T.A. , O’Neill , M.A. , Sarria , R. , York , W.S. , Raikhel , N.V. , Keegstra , K. ( 2003 ). Analysis of xyloglucan fucosylation in Arabidopsis . Plant Physiol. 132 : 768 – 778 . Google Scholar Crossref Search ADS PubMed WorldCat 49. Ramachandran , G.N. , Ramakrishnan , C. , Sasisekharan , V. ( 1963 ). Stereochemistry of polypeptide chain configurations . J. Mol. Biol. 7 : 95 – 99 . Google Scholar Crossref Search ADS PubMed WorldCat 50. Rocha , J. , Cicéron , F. , Lerouxel , O. , Breton , C. , de Sanctis , D. ( 2016 ). The galactoside 2-alpha-l-fucosyltransferase FUT1 from Arabidopsis thaliana: crystallization and experimental MAD phasing . Acta Crystallogr. F Struct. Biol. Commun. 72 : 564 – 568 . OpenURL Placeholder Text WorldCat 51. Sarria , R. , Wagner , T.A. , O’Neill , M.A. , Faik , A. , Wilkerson , C.G. , Keegstra , K. , Raikhel , N.V. ( 2001 ). Characterization of a family of Arabidopsis genes related to xyloglucan fucosyltransferase1 . Plant Physiol. 127 : 1595 – 1606 . Google Scholar Crossref Search ADS PubMed WorldCat 52. Scheller , H.V. , Ulvskov , P. ( 2010 ). Hemicelluloses . Annu. Rev. Plant Biol. 61 : 263 – 289 . Google Scholar Crossref Search ADS PubMed WorldCat 53. Takahashi , T. , Ikeda , Y. , Tateishi , A. , Yamaguchi , Y. , Ishikawa , M. , Taniguchi , N. ( 2000 ). A sequence motif involved in the donor substrate binding by alpha1,6-fucosyltransferase: the role of the conserved arginine residues . Glycobiology 10 : 503 – 510 . Google Scholar Crossref Search ADS PubMed WorldCat 54. Terwilliger , T.C. ( 2000 ). Maximum-likelihood density modification . Acta Crystallogr. D Biol. Crystallogr. 56 : 965 – 972 . Google Scholar Crossref Search ADS PubMed WorldCat 55. Vanzin , G.F. , Madson , M. , Carpita , N.C. , Raikhel , N.V. , Keegstra , K. , Reiter , W.D. ( 2002 ). The mur2 mutant of Arabidopsis thaliana lacks fucosylated xyloglucan because of a lesion in fucosyltransferase AtFUT1 . Proc. Natl. Acad. Sci. USA 99 : 3340 – 3345 . Google Scholar Crossref Search ADS WorldCat 56. Winn , M.D. , et al. . ( 2011 ). Overview of the CCP4 suite and current developments . Acta Crystallogr. D Biol. Crystallogr. 67 : 235 – 242 . Google Scholar Crossref Search ADS PubMed WorldCat 57. Wu , Y. , Williams , M. , Bernard , S. , Driouich , A. , Showalter , A.M. , Faik , A. ( 2010 ). Functional identification of two nonredundant Arabidopsis alpha(1,2)fucosyltransferases specific to arabinogalactan proteins . J. Biol. Chem. 285 : 13638 – 13645 . Google Scholar Crossref Search ADS PubMed WorldCat Author notes 1 These authors contributed equally to this work. 2 Address correspondence to breton@cermav.cnrs.fr. The author responsible for distribution of materials integral to the findings presented in this article in accordance with the policy described in the Instructions for Authors (www.plantcell.org) is: Christelle Breton (breton@cermav.cnrs.fr). www.plantcell.org/cgi/doi/10.1105/tpc.16.00519 © 2016 American Society of Plant Biologists. All rights reserved. This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/open_access/funder_policies/chorus/standard_publication_model)
TI - Structure of Arabidopsis thaliana FUT1 Reveals a Variant of the GT-B Class Fold and Provides Insight into Xyloglucan Fucosylation
JF - The Plant Cell
DO - 10.1105/tpc.16.00519
DA - 2016-11-11
UR - https://www.deepdyve.com/lp/oxford-university-press/structure-of-arabidopsis-thaliana-fut1-reveals-a-variant-of-the-gt-b-FnIT70AVHc
SP - 2352
EP - 2364
VL - 28
IS - 10
DP - DeepDyve
ER -