Xyloglucan Fucosyltransferase, an Enzyme Involved in Plant Cell Wall Biosynthesis

See allHide authors and affiliations

Science  18 Jun 1999:
Vol. 284, Issue 5422, pp. 1976-1979
DOI: 10.1126/science.284.5422.1976


Cell walls are crucial for development, signal transduction, and disease resistance in plants. Cell walls are made of cellulose, hemicelluloses, and pectins. Xyloglucan (XG), the principal load-bearing hemicellulose of dicotyledonous plants, has a terminal fucosyl residue. A 60-kilodalton fucosyltransferase (FTase) that adds this residue was purified from pea epicotyls. Peptide sequence information from the pea FTase allowed the cloning of a homologous gene, AtFT1, from Arabidopsis. Antibodies raised against recombinant AtFTase immunoprecipitate FTase enzyme activity from solubilized Arabidopsis membrane proteins, andAtFT1 expressed in mammalian COS cells results in the presence of XG FTase activity in these cells.

In most multicellular organisms, cells are embedded in a complex extracellular matrix that keeps them together and influences the shape, development, and polarity of the cells they contact. Animal cells have such an extracellular matrix at their surface, but plants possess a distinct wall that encloses every cell. Many important differences between plants and animals with respect to nutrition, growth, reproduction, and defense mechanisms can be traced to the plant cell wall (1). Cell wall extensibility is a major determinant of plant growth (2). The biosynthesis of plant cell walls is very tightly regulated. Although an individual plant cell may expand its volume by nearly 20,000 times, its cell wall must maintain a uniform thickness and structure to prevent hemorrhaging of the cell through local defects (2). However, despite extensive descriptions of the chemical and physical structure of the plant cell wall, very little is known about its biosynthesis. One gene encoding a cell wall–synthesizing enzyme, cellulose synthase, has been cloned (3).

The flexible primary walls of young plant cells are mainly composed of cellulose microfibrils and matrix polysaccharides. Matrix polysaccharides include hemicelluloses that bind tightly but noncovalently to cellulose microfibrils, cross-linking them into a complex network. The hemicellulose xyloglucan (XG) makes up approximately 20% of the total cell wall in dicot and nongraminaceous monocot plants and forms a load-bearing network by associating to the surfaces of surrounding cellulose microfibrils through hydrogen bonds (4, 5). XG contains a β-1,4-glucan backbone decorated with side chains of xylose alone; xylose and galactose; and xylose, galactose, and fucose. The presence or absence of the terminal fucose residue may have structural and biological significance. Some models suggest that the presence or absence of this fucose residue will determine whether the xyloglucan conformation is planar and thus better able to bind to cellulose (6), though contradicting evidence has been described (7). XG networks may be modified by XG endotransglycosylase (XET), an enzyme that cleaves and rejoins adjacent XG chains. A recombinant XET demonstrated different activity rates for fucosylated versus nonfucosylated XG oligosaccharide acceptors, indicating that the fucosylation state may affect XET modification of the cell wall (8). In addition, oligosaccharides consisting of an XG nonasaccharide prevent auxin-promoted elongation of pea stems if these oligosaccharides contain fucose but not if they lack fucose (9). Thus, it is possible that XG fragments act as signaling molecules in vivo.

Most matrix polysaccharides are branched molecules modified by various sugars. These modifications are important because they allow heterogeneity in the shape of matrix polysaccharides and in the patterns of cross-links, resulting in a dynamic and porous cell wall. These polysaccharide modifications occur via glycosyltransferase reactions, many of which occur in the Golgi complex (10). Attempts to clone plant glycosyltransferases using sequences derived from bacterial or mammalian transferases have been unsuccessful (11). This is not entirely unexpected, for although Golgi glycosyltransferases often have similar general structural features, they rarely share extensive sequence similarity (12).

The terminal fucosyl residue on XG side chains is added by a fucosyltransferase (FTase). We purified enough of this FTase from pea epicotyls to determine partial amino acid sequences from the enzyme. Microsomes were prepared from the pea epicotyls, carbonate-washed to enrich for membrane proteins (13), and solubilized with nonionic detergent such as Triton X-100. A specific assay for this enzyme was developed using tamarind or nasturtium seed storage XG, which lack fucosyl residues, as acceptor molecules and radiolabeled guanosine diphosphate (GDP)–fucose as a donor (14, 15). GDP-agarose affinity chromatography, size exclusion chromatography, and anion exchange chromatography were used in conjunction with FTase activity assays to purify and detect the enzyme (Fig. 1) (16). It was possible to purify XG FTase 1400-fold after size exclusion chromatography, resulting in a total of 50 μg of protein containing 70 nanokats (nKat) (nanomoles of substrate incorporated into the product per second) of XG FTase activity. To confirm that the purified pea protein synthesizes an α-1,2 fucose: galactose linkage, carbohydrate analysis was performed on the product resulting from in vitro fucosylation of tamarind XG by purified FTase (Table 1) (17). Linkage analysis indicated that incubation of tamarind XG with purified FTase resulted in a decrease in the mole percentage of terminal galactose and the appearance of 2-galactose and terminal fucose, thus verifying the activity of the purified enzyme (Table 1). After biochemical purification and subsequent analysis, two polypeptides of approximately 65 and 60 kD in size were observed to copurify with XG FTase activity (Fig. 1). Limited peptide sequences were obtained from both proteins (18). The 65-kD peptide was identified as a homolog of BiP, a molecular chaperone usually localized to the endoplasmic reticulum. It remains unclear whether copurification of BiP with FTase represents an important interaction. Six peptides analyzed from the 60-kD protein were not significantly similar to proteins of known function in databases but did allow the identification of an Arabidopsis expressed sequence tag (EST), the sequence of which encoded four out of the six peptides, with amino acid identity ranging from 63 to 85% (18).

Figure 1

Biochemical purification of XG FTase from pea. (A) Silver-stained SDS-PAGE gels showing protein profiles from (i) carbonate-washed detergent-solubilized microsomes from pea epicotyls (Sol. Microsomes) or the fractions containing peaks of FTase activity from (ii) a GDP-agarose affinity column (GDP-affinity) or (iii) a size exclusion column (Size Excl.) (15). Numbers at left indicate sizes in kilodaltons. (B) Top: Silver-stained SDS-PAGE gel showing the protein profile from several fractions of an anion exchange column. Numbers at left indicate sizes in kilodaltons. Bottom: Total XG FTase activity for each fraction of the anion exchange column eluate. Bars align with the corresponding SDS-PAGE profiles above.

Table 1

Carbohydrate linkage analysis of tamarind XG before (tamarind XG) and after (fucosylated XG) incubation with purified pea FTase (17). Dashes indicate that no such linkage was detected.

View this table:

An analysis of this EST (number 191A6T7) indicated that it was not a full-length clone (19). The EST was used as a probe to screen an Arabidopsis cDNA library, and full-length cDNA clones were isolated (19). The cDNA clones contain a 1677-nucleotide open reading frame encoding a 63.5-kD protein and correspond to a region of the fully sequenced Arabidopsisbacterial artificial chromosome (BAC) T18E12 (20). The cDNA and the corresponding genomic clone have been designatedAtFT1. Analysis of the BAC indicates that there may be a second glycosyltransferase approximately 300 base pairs downstream fromAtFT1 that is 63% similar to AtFT1 at the amino acid level. In addition, two other Arabidopsis ESTs and three Arabidopsis genomic sequences that show significant similarity to AtFT1 have been observed in the databases (21). Thus, Arabidopsis may carry a family of FTases, each differentially regulated by such factors as environmental stress, tissue localization, or developmental stage, or specific to different acceptors.

To confirm the identity of AtFT1 as encoding a fucosyltransferase using XG as an acceptor, we prepared polyclonal antibodies directed against AtFT1 overexpressed inEschicheria coli and used them to immunoprecipitate proteins from carbonate-washed, detergent-solubilized Arabidopsisproteins (22, 23). The immunoprecipitated proteins were then assayed for XG FTase activity; 2.6-fold more FTase activity was correlated with pellets derived from immunoprecipitation reactions using immune antiserum rather than preimmune serum, thereby confirming that the Arabidopsis clone encodes a XG FTase (Fig. 2). In addition, a COS cell line expressingAtFT1 showed in vitro FTase activity that was 41 times higher than that of COS cells transformed with an empty vector (Fig. 2) (24). Taken together, these data indicate thatAtFT1 is involved in XG biosynthesis.

Figure 2

Confirmation of AtFTase activity. (A) Polyclonal antibodies to AtFT1 recognize an approximately 63-kD polypeptide in solubilized membrane proteins ofArabidopsis. Top: Left two lanes, immunoblot; right two lanes, Coomassie blue staining of immunoblot membrane. In both cases, lane 1 is Arabidopsis carbonate-washed, detergent-solubilized membrane proteins, and lane 2 is antigen (50 ng). Bottom: Antibodies to FTase immunoprecipitate more XG-specific FTase activity than does an equal volume of preimmune serum. The FTase activity of precipitated pellets is shown. This is an example similar to results seen in seven different replicates. (B) Full-length AtFT1 expressed in a Cos cell line shows XG-specific FTase activity. Activity is shown in the presence (+XG) or absence (–XG) of tamarind XG for untransformed Cos-7 cells, cells transformed with vector DNA (Cos-7 vector), cells transformed with vector containing AtFT1 (Cos-7 vector AtFT1), or solubilized pea Golgi vesicles (Pea Golgi). In graphs, error bars show ±1 SD; if no error bars are visible, SDs are contained within the width of the plot element.

Although AtFT1 has some structural characteristics common to other fucosyltransferases, it is quite divergent at the amino acid sequence level. Hydrophobicity plots predict that there may be an NH2-terminal transmembrane signal anchor sequence. In vitro translation in the presence of canine pancreatic microsomes followed by carbonate washing of the products indicates that the AtFT1translation product is a membrane protein (25). As with other glycosyltransferases, the COOH-terminal region is predicted to be largely hydrophilic. AtFT1 is not significantly similar to any other FTases from other organisms, although multiple sequence alignments have identified three motifs that appear to be conserved among all α1,2- FTases (26). One of these motifs, described previously, is found in all α1,2- and α1,6-FTases for which sequence data are known (27). Because these proteins have different acceptor molecules but share the same sugar nucleotide donor (GDP-fucose), it is possible that these regions are involved in GDP-fucose binding or conserved structural characteristics. Some small regions of similarity are observed between AtFT1 and NodZ, a fucosyltransferase in Rhizobium involved in the synthesis of nodulation factors.

The unique nature of this FTase will allow its use as a tool for identifying other plant-specific glycosyltransferases. Hundreds to thousands of different genes (28) are needed to synthesize the various polysaccharides that compose the cell wall. Substrate acceptors and assays remain unavailable for many of these enzymes. Identification of other carbohydrate transferases, perhaps by sequence similarity, could lead to tailored in vitro production of carbohydrates as well as an understanding of how the complex plant cell wall is biosynthesized.

  • * These authors contributed equally to this work.

  • Present address: Monsanto, St. Louis, MO 63198, USA.

  • To whom correspondence should be addressed at MSU-DOE Plant Research Laboratory, Michigan State University, East Lansing, MI 48824, USA. E-mail: nraikhel{at} (N.V.R.); keegstra{at} (K.K.)


View Abstract

Navigate This Article