A Barley Xyloglucan Xyloglucosyl Transferase Covalently Links Xyloglucan, Cellulosic Substrates, and (1,3;1,4)-β-D-Glucans*

Molecular interactions between wall polysaccharides, which include cellulose and a range of noncellulosic polysaccharides such as xyloglucans and (1,3;1,4)-β-d-glucans, are fundamental to cell wall properties. These interactions have been assumed to be noncovalent in nature in most cases. Here we show that a highly purified barley xyloglucan xyloglucosyl transferase HvXET5 (EC 2.4.1.207), a member of the GH16 group of glycoside hydrolases, catalyzes the in vitro formation of covalent linkages between xyloglucans and cellulosic substrates and between xyloglucans and (1,3;1,4)-β-d-glucans. The rate of covalent bond formation catalyzed by HvXET5 with hydroxyethylcellulose (HEC) is comparable with that on tamarind xyloglucan, whereas that with (1,3; 1,4)-β-d-glucan is significant but slower. Matrix-assisted laser desorption ionization time-of-flight mass spectrometric analyses showed that oligosaccharides released from the fluorescent HEC:xyloglucan conjugate by a specific (1,4)-β-dglucan endohydrolase consisted of xyloglucan substrate with one, two, or three glucosyl residues attached. Ancillary peaks contained hydroxyethyl substituents (m/z 45) and confirmed that the parent material consisted of HEC covalently linked with xyloglucan. Similarly, partial hydrolysis of the (1,3;1,4)-β-d-glucan:xyloglucan conjugate by a specific (1,3;1,4)-β-d-glucan endohydrolase revealed the presence of a series of fluorescent oligosaccharides that consisted of the fluorescent xyloglucan acceptor substrate linked covalently with 2-6 glucosyl residues. These findings raise the possibility that xyloglucan endo-transglucosylases could link different polysaccharides in vivo and hence influence cell wall strength, flexibility, and porosity.

The structural integrity of land plants is dependent, in large part, on the collective strength and flexibility of the walls that surround individual cells, whereas key elements of cellular function, such as water and nutrient exchange, depend upon the porosity of the walls. Plant cell walls are dynamic structures that are altered during cell division, growth, and differentiation to enable cells to adapt to changing functional requirements and to environmental and pathogen-induced stresses. In addition, walls are important for intercellular cohesion and cell-cell communication and must be selectively permeable to water, nutrients, and growth regulators.
The primary cell walls of vascular plants consist of cellulosic microfibrils that are embedded in a chemically complex matrix consisting mostly of polysaccharides but also containing structural proteins, enzymes, and phenolic acids (1,2). Xyloglucans and pectic polysaccharides are the major noncellulosic polysaccharides of primary walls from dicotyledonous plants, whereas in the Poales and related commelinoid monocots, including commercially important cereals and grasses, glucuronoarabinoxylans and (1,3;1,4)-␤-D-glucans are the predominant noncellulosic wall polysaccharides, and levels of pectic polysaccharides, glucomannans, and xyloglucans are relatively low (3). In addition, wall composition and the fine structures of component polysaccharides vary depending upon the growth phase, cell type, cell position, and local region within the wall (4,5). In some cell types, lignin is deposited throughout the wall during secondary thickening (6,7), and in response to pathogen attack, the rapid formation of a cross-linked protein network, together with the deposition of lignin and callose, can strengthen walls and create new physical barriers to invading microorganisms (8).
Although the types and abundance of polysaccharides in plant cell walls have been defined in detail, little information is available on the molecular interactions between constituent polysaccharides in the wall. In most wall models it is assumed that different polysaccharides are held in place through extensive intermolecular hydrogen bonding rather than through covalent interactions (1,9). However, there is recent circumstantial evidence (10) to support earlier suggestions (11)(12)(13) that xyloglucans might be covalently linked to pectic polysaccharides.
Extraction of HvXET5-Barley (Hordeum vulgare L., cv. Clipper) (2 kg dry weight) was surface-sterilized for 10 min in 0.1% (w/v) NaOCl, washed successively with tap water, 0.5 M NaCl, and sterile water, and steeped for 24 h in sterile water containing chloramphenicol (100 g/ml), neomycin (100 g/ml), penicillin G (100 units/ml), and nystatin (100 units/ml). Germinating grains were maintained at ϳ40% (w/w) moisture content by regular application of fresh antibiotic solution for 7 days at 21 Ϯ 2°C in the dark. Bacterial or fungal contamination of the grains was not evident at any stage during this period. The germinated grain and young seedlings were homogenized at 4°C in 2.0 volumes of homogenization buffer, pH 6, containing 0.1 M imidazole-HCl buffer, 1 M NaCl, 2 mM EDTA, 1 mM 2-mercaptoethanol, and 1 mM phenylmethylsulfonyl fluoride (buffer A), in a Waring Blender six times with 1-min intervals with intermittent cooling (2 min) on ice. The homogenate was held for 1 h at 4°C to extract proteins; insoluble material was removed by centrifugation (4000 ϫ g, 60 min, 4°C), and the extract filtered through Miracloth. The extract was precipitated to 90% with solid (NH 4 ) 2 SO 4 ; the precipitate was collected (8000 ϫ g, 45 min, 4°C) and resuspended in 4 liters of buffer A (without NaCl). The extract was stored in 0.5-liter aliquots at Ϫ20°C.
Purification of HvXET5-The HvXET5 enzyme was purified from extracts of 7-day-old barley seedlings using Sepharose Q, phenyl-Sepharose, chromatofocusing on PBE-94, and size-exclusion chromatography on Bio-Gel P-60 ( Table 1). The activity of HvXET5 during enzyme purification was determined radiometrically at 30°C in 100 mM succinate or ammonium acetate buffers, pH 6.0, containing 5 mM calcium chloride, 0.3% (w/v) TXG, and 3 H-labeled xyloglucan-derived saccharide heptaitol (XXXGol, specific radioactivity 83 MBq⅐mol Ϫ1 ) (14) with ϳ30,000 dpm per reaction mixture. The radioactivity incorporated in reaction products was counted on 2 ϫ 1-cm Whatman chromatography 3MM paper strips in plastic vials and using LSC6500 scintillation counter (Beckman Instruments, Fullerton, CA) with ϳ48% efficiency for tritium, using EcoLume scintillation fluid, and with 70% quenching. Enzyme activity of HvXET5 during purification is expressed in katals, where 1 katal represents 1 mol of product formed per s; specific activity is expressed in picokatals⅐mg Ϫ1 protein.
Isoelectric Focusing-The crude protein extracts and purified preparations were separated on a flatbed IEF apparatus (GE Healthcare) in 1-mm polyacrylamide gels using a pH gradient of 3.5-9.5. Pre-focused gels were run at 600 V for 30 min, followed by 800 V for a further 20 min. Proteins were detected with a Coomassie Brilliant Blue dye after the gels were fixed in 20% (w/v) trichloroacetic acid. Apparent pI values were estimated by reference to marker proteins with pI values of 4.45-9.6. Enzyme activity in gels was detected by overlaying the separation gels with a 1.5-mm 1.3% (w/v) aga-rose detection gel containing 0.2% (w/v) TXG and 5-10 M of SR-labeled xyloglucan-derived oligosaccharides XGO-SR (XXXG-SR, XXLG-SR, XLLG-SR molar ratios were 1:1.6:1.8) (19,20) in 0.1 M succinate buffer, pH 6, containing 5 mM calcium chloride. The separation gels contacted with detection gels were incubated for 1-5 h at 30°C, depending on the activity of the preparation under investigation. The detection gels were immediately fixed and de-stained in 60% (v/v) ethanol containing 5% (v/v) formic acid. The detection gel with fluorescent zones was evaluated under a UV lamp at 366 nm.
pH Optimum and Enzyme Stability-The effect of pH on the activity of HvXET5 was determined by incubating 1 nM HvXET5 at 30°C for 60 min in 50 mM citric acid, 100 mM sodium dihydrophosphate (McIlvaine) buffers, pH 4.0 -8.5, in the presence of 0.02% (w/v) bovine serum albumin. A comparison of succinate, ammonium acetate, or sodium phosphate buffers, each at 50 -200 mM, indicated that HvXET5 activity was unaffected by the ionic strength of these buffers. The thermal stability of HvXET5 was determined after 15 min of incubation at 0 -70°C. The freeze/thaw stability of HvXET5 at 1 nM concentration was determined after three cycles of freezing (Ϫ80°C) and thawing (4°C), each at duration of 3 min. Activity was subsequently measured at 30°C in 100 mM ammonium acetate buffer, pH 6, containing 5 mM calcium chloride, without and with the addition of 10% (v/v) glycerol. Enzyme activity was determined radiometrically as specified above, and expressed as % activity relative to maximal activity. Assays were performed in triplicate and standard errors of 8 -14% were observed.
Effects of Divalent Cations-The effect of Ca 2ϩ (as calcium chloride) and Mg 2ϩ (as magnesium sulfate), both in 0 -15 mM  concentration ranges, and of EDTA in a 0 -20 mM concentration range, were determined by incubating 1 nM HvXET5 in 100 mM succinate buffer, pH 6. Enzyme activities were determined radiometrically as specified above, in triplicate and with a standard error of 8 -10%. Substrate Specificities-The incubation mixtures contained 1.2% (w/v) soluble polysaccharides as donor substrates and as acceptor substrates either 23-27 M XGO-SR, SR-labeled cello-oligosaccharides (CEO-SR) (DP 2-8), or SR-labeled laminari-oligosaccharides (LAO-SR) (DP 2-6) (20) in 100 mM ammonium acetate buffer, pH 6, containing 5 mM calcium chloride and 0.5-1 nM purified HvXET5. The molar ratios of individual oligosaccharides in the two oligosaccharide-SR mixtures were 1:0.78:0.62:0.41:0.16:0.03:0.013 for C2-SR to C8-SR (SR-labeled cellobiose to cellooctaose), and 1.0:1.7:0.87:0.41:0.2 for L2-SR to L6-SR (SR-labeled laminaribiose to laminarihexaose). All incubations proceeded for 18 h at 30°C. Enzymes inactivated by boiling for 3 min served as controls. The efficiency of transfer of selected polysaccharides onto fluorescent acceptors was determined by size-exclusion HPLC. The SRlabeled oligosaccharides and polysaccharides were detected following HPLC by fluorescence detection (excitation 568 nm and emission 584 nm) or by evaporative light scattering detection (ELSD) at 568 nm, and by MALDI-TOF mass spectrometry analyses. Enzyme activities were determined by integrating peak areas, after subtracting background level obtained from boiled enzyme control reactions. Relative activities of HvXET5 are expressed as % of activity observed with TXG as a donor substrate and XGO-SR as an acceptor. In all instances assays were performed in duplicate with standard errors of 8 -12%. Detection limits of the fluorescence assays and ELSD were better than 0.1 pmol of XGO-SR or CEO-SR, or 1 ϫ 10 Ϫ5 % of the amount of XGO-SR or CEO-SR acceptors used in standard enzyme reactions, with a standard error of 6%. The efficiency of transfer of selected polysaccharides onto [ 3 H]XXXGol was further evaluated by ascending chromatography in 60% (v/v) ethanol on Whatman chromatography 3MM paper strips, and radioactivity in paper strips was determined by liquid scintillation counting as specified above.
HPLC Analysis-Native polysaccharides were fractionated by size-exclusion chromatography on either a P3000 or P4000 PolySep GFC columns (particle size not specified, 300 ϫ 7.   Corp., Waltham, MA) with a linear gradient of 23.45-26.65% aqueous acetonitrile at a flow rate of 0.2 ml/min. A model 1090 liquid chromatograph with diode-array detector, controlled by ChemStation software (Agilent Technologies, Palo Alto, CA), and fluorescence (model RF-10AXL, Shimadzu, Kyoto, Japan) and ELSD (model 800, Alltech Associates Inc., Deerfield, IL) connected in series to the 1090 DAD, were used for analyses of enzymic reactions. The eluant flow from the fluorescence detector to the ELSD was split in the ratio 5 (to collect) to 1 (ELSD). The ELSD was operated at 40°C and a nitrogen pressure of 1.5 bar, and the column temperature was 21°C. Size-exclusion HPLC of SR-labeled polysaccharides and oligosaccharides were carried out on a BioSep SEC S3000 column (5 m, 300 ϫ 7.8 mm); the eluant was 100 mM ammonium acetate in 20% (v/v) acetonitrile at a flow rate of 1.0 ml/min.
MALDI-TOF Mass Spectrometry Analyses-MALDI TOF spectra were acquired using a Bruker Ultraflex II MALDI-TOF/TOF mass spectrometer (Bruker Daltonik GmbH, Bremen, Germany) operating in a reflectron mode. Samples (1 l) dissolved in water were mixed with 1 l of a 5 g/liter solution of dihydroxybenzoic acid in 1% (v/v) phosphoric acid and spotted on a matt-steel target plate. External calibration was performed using peptide standards (Bruker Daltonik GmbH, Bremen, Germany), which were analyzed under the same conditions. Spectra were acquired using between 800 -3,000 laser shots. The ionization voltages were IS1 ϭ 25.0 kV, IS2 ϭ 21.7 kV, and lens ϭ 8.2 kV. The mass spectrometer was calibrated with XXXG-SR, XXLG-SR, and XLLG-SR that were purified by tandem normal phase and size-exclusion HPLC.

Purification of the Barley
HvXET5-The barley HvXET5 enzyme was purified ϳ3,000-fold from extracts of 7-day-old barley seedlings, using ammonium sulfate precipitation, ion exchange and hydrophobic chromatography, chromatofocusing, and sizeexclusion chromatography ( Table  1). The numbering of the barley XET isoenzymes is based upon the gene nomenclature proposed by Strohmeier et al. (23). The specific activity of the purified HvXET5 was 3,630 picokatals⅐mg Ϫ1 , which is about six times higher than the specific activity of a recombinant Populus XET PttXET16A (24) and close to that of a cauliflower XET purified by affinity chromatography on xyloglucan (25). It was necessary to use 1 M NaCl in the imidazole buffer for efficient enzyme extraction, presumably because the XET enzymes are tightly associated with cell wall material. The key steps during the purification of HvXET5 were chromatofocusing on PBE-94 and chromatography on Bio-Gel P-60, where several HvXET isoenzymes were separated from each other and from major contaminating proteins (Table 1). The enzymes bound strongly to phenyl-Sepharose and were not eluted by a 3-0 M linear gradient of NaCl. A 30 -70% linear gradient of ethylene glycol was required for elution, suggesting that the enzymes were highly hydrophobic. The HvXET5 isoenzyme also bound to Bio-Gel P-60 and 0.01% (v/v) Tween 20, and 0.2 M NaCl were required to elute the enzyme from the column.
The final purified enzyme preparation showed a single band of 34 kDa on SDS gels at high protein loadings (Fig. 1A), with no other protein species present. These data indicated that contaminating proteins accounted for less than 6 ng or 0.25% of the protein used for SDS-PAGE analysis. Furthermore, a single protein and activity band of pI 7.6 was detected on an isoelectric focusing gel and on a zymogram detection gel containing TXGand SR-labeled xyloglucan oligosaccharides XGO-SR (Fig. 1B). A single sequence was detected during NH 2 -terminal amino acid sequence analysis of the purified HvXET5 enzyme, where a 97% yield of a phenylthiohydantoin-alanine in the 1st cycle, normalized per mol of the total protein of HvXET5, was obtained. Another isoenzyme, HvXET6, was partially purified from 7-day-old barley seedlings but contained small amounts of other proteins and was not examined further.
pH Optimum and Enzyme Stability-The pH optimum of the purified HvXET5 was 6.0, and the temperature optimum varied between 28 and 30°C (Fig. 2). The enzyme also operated efficiently at 0°C, where it showed ϳ40% of the activity observed at 30°C. A similar tolerance to low temperature has also been observed with recombinant TCH4 XET enzyme from Arabidopsis (26). As for the freeze/thaw stability of HvXET5, the purified enzyme retained its activity for at least a year when stored at Ϫ20°C and did not lose activity after several freezethaw cycles. The addition of 10% (v/v) glycerol had no affect on HvXET activity after several freeze-thawing cycles.
The Effects of Divalent Cations-Effects of the divalent cations Ca 2ϩ and Mg 2ϩ were tested on the activity of HvXET5 under optimal conditions. Although Ca 2ϩ at concentrations between 5 and 15 mM stimulated HvXET5 activity by ϳ7-8%, Mg 2ϩ inhibited the activity of HvXET5 by 3-4% in the same concentration range. The chelating agent EDTA inhibited the activity of HvXET5 by ϳ30% in 2-20 mM concentration ranges (data not shown).
Substrate Specificity of the Purified HvXET5-Xyloglucan oligosaccharides (XGO-SR) and cello-oligosaccharides (CEO-SR) fluorescently labeled with sulforhodamine (SR) were used as acceptor substrates for the purified HvXET5. Transferase activity was observed when tamarind xyloglucan (TXG), hydroxyethylcellulose (HEC), sulfuric acid-swollen cellulose, and barley (1,3;1,4)-␤-D-glucan were used as donor polysaccharides (Table 2). No hydrolytic activity was detected with any of the donor substrates. The transfer of nonfluorescent donor polysaccharides onto fluorescent acceptors was determined by size-exclusion HPLC, where dramatic increases in molecular size of fluorescent material showed that the transfer reaction had occurred. In Fig. 3A the progressive formation over 24 h of high molecular mass, fluorescent HEC from unlabeled HEC donor substrate, and the fluorescent XGO-SR acceptor molecule can be seen. In Fig. 3B, the transfer reaction by HvXET5 is shown with the high molecular mass product of the reaction presented in Fig. 3A, whereby the fluorescent component XGO-SR of the high molecular mass HEC:XGO-SR material from Fig. 3A (shaded fractions) was removed from the reducing end of the polysaccharide and replaced with the nonfluorescent oligosaccharide XXXGol. As this occurred, low molecular mass fluorescent oligosaccharides XGO-SR were progressively released. A schematic representation of the transfer reaction shown in Fig. 3, A and B, is summarized in Fig. 3C.
The HvXET5 enzyme also catalyzes the transfer of TXG, celluloses, and (1,3;1,4)-␤-D-glucan onto fluorescently labeled CEO-SR, albeit at low levels ( Table 2). The fluorescence assay technique used in this study was sufficiently sensitive to confidently measure activities that were better than 0.1 pmol of XGO-SR or 1 ϫ 10 Ϫ5 % of the amount of XGO-SR acceptor used in standard enzyme reactions.
In control experiments, either the acceptor or donor substrates were omitted or the enzyme was inactivated by boiling. In no instance was any change in molecular size of fluorescent material observed. The absence of transferase activity, when either the donor or acceptor substrate was omitted, was particularly important, because it indicated that transglycosylation  Fig. 3 are indicated. a.u., arbitrary units. reactions that are often observed when polysaccharide hydrolases are incubated with high substrate concentrations (15) were not occurring. All experiments were performed at micromolar donor and acceptor concentrations, again to rule out the potential for transglycosylation reactions attributable to high substrate concentrations. However, it should be noted that the efficiency of such reactions would also depend upon overall substrate affinities and catalytic properties of the enzyme, together with substrate properties such as the reactivities of leaving groups.
Analysis of the Products of HvXET5 Action-To confirm the results obtained from the fluorescence assays, the high molecular mass HEC:XXXG-SR conjugate generated by incubation of the HvXET5 with XXXG-SR and nonfluorescent HEC (Fig. 4A) was partially hydrolyzed with a highly purified (1,4)-␤-D-glucan endohydrolase from T. reesei that contained no ␤-D-glucosidase or other contaminating activities. Three major fluorescent oligosaccharide fractions were released from the HEC: XXXG-SR (Fig. 4, A and B) during the partial hydrolysis with the (1,4)-␤-D-glucan endohydrolase. MALDI-TOF mass spectrometric analyses showed that these oligosaccharides had molecular masses corresponding to XXXG-SR with one, two, or three additional glucosyl residues attached (Fig. 4B). Furthermore, ancillary m/z peaks, which corresponded to Glc: XXXG-SR, Glc-Glc:XXXG-SR, and Glc-Glc-Glc:XXXG-SR containing hydroxyethyl substituents (m/z 45) from the donor substrate, were also detected in the spectra (Fig. 4, C and D) and confirmed that the parent material consisted of HEC covalently linked with the XXXG-SR.

DISCUSSION
Xyloglucans consist of a backbone of (1,4)-␤-D-glucan substituted with xylosyl, galactosyl, and fucosyl residues (28). The molecular sizes of xyloglucans can be altered after their deposition into the cell wall (29), and this process is likely to be mediated by a class of enzymes broadly known as xyloglucan endotransglycosylases/hydrolases (XTHs). However, enzymes within this group can have XET activity or both xyloglucan endotransglycosylase and xyloglucan endohydrolase activities (14,30,31). The XETs are abundant in the apoplastic space; they cleave the (1,4)-␤-D-glucan backbone of xyloglucans and, in the case of XETs, transfer the nonreducing fragment of the original substrate that remains bound to the enzyme directly onto the nonreducing terminus of another xyloglucan chain (14,31). The xyloglucan molecule that is cleaved by the enzyme initially is referred to as the donor substrate, whereas the xyloglucan chain to which the product of hydrolysis is transferred is known as the acceptor substrate. The transglycosylation activity of XETs can theoretically result in the disproportionation of xyloglucan molecules, such that some will increase in molecular mass, and others will decrease in molecular mass (14,31).
Sequences encoding XTHs are surprisingly abundant in barley EST data bases, given the relatively low levels of xyloglucans in walls of most barley tissues (23,32). There are at least 22 XTH genes in barley (23), about 30 in rice (33), about 40 in Populus trichocarpa (34), and about 33 in Arabidopsis (35). In an attempt to reconcile the relatively low abundance of xyloglucans in cell walls of barley against the large number of XTH genes and their high expression levels in many tissues of barley, Strohmeier et al. (23) suggested that some of the XTHs might be active on the more abundant matrix phase polysaccharides of cell walls in barley, namely the arabinoxylans and the (1,3; 1,4)-␤-D-glucans. A role for XTHs in the modification of highly abundant (1,3;1,4)-␤-D-glucans and arabinoxylans in walls of the commelinoid monocots would be consistent with the abrupt increase in molecular size of heteroxylans that has been observed in suspension-cultured maize cells following the deposition of the polysaccharide into the walls (29).
In an attempt to test suggestions that some barley XET enzymes could catalyze transfer of xyloglucan onto acceptors other than xyloglucans, an XET isoenzyme was purified from extracts of barley seedlings to a near monodisperse form. The difficulties encountered during the purification procedure of HvXET5 were largely attributable to the presence of numerous hydrophobic patches on the surface of the enzyme, as predicted from the three-dimensional structure of the Populus XET (36). However, after the HvXET5 enzyme was purified, its activity remained more or less constant for at least a year at Ϫ20°C. The difficulties with enzyme purification might also explain why so few XTHs have been purified from plant tissue extracts (24,38). Most of the known enzymic properties of XTHs have been determined following heterologous expression of the corresponding cDNAs (25, 26, 39 -41).
The HvXET5 isoenzyme that was purified in the present work catalyzed, in vitro, the formation of covalent linkages between celluloses such as chemically modified or paracrystalline HEC and sulfuric acid swollen cellulose or (1,3;1,4)-␤-Dglucans and xyloglucans (Figs. 3-6). The polysaccharides are linked from reducing to nonreducing ends of donor and acceptor substrates, respectively, rather than by cross-linking of the type observed between arabinoxylan chains through esterified hydroxycinnamic acids (7) or between pectic polysaccharides through borate (1,42). The results are consistent with recent data from Ait Mohand and Farkas (43), who showed heterotransglycosylating activity in unpurified extracts from nasturtium (Tropaeolum majus), but they were not able to unambiguously assign the enzyme responsible. The HvXET5 activity represents a non-Leloir type of biosynthetic reaction, insofar as the energy required for the formation of the new glycosidic linkage is provided from an existing glycosidic linkage rather than from a sugar nucleotide-activated donor. The data shown in Fig. 3A is particularly important with respect to the action pattern of the barley XET. The presence of fluorescent material of intermediate molecular mass, which is with a molecular mass between that of the starting HEC and the fluorescent acceptor substrate XGO-SR, indicates that the enzyme acts in an essentially stochastic manner. Conversely, when the HEC is tagged at its reducing terminus with the fluorescent XGO-SR, the absence of fluorescent products of intermediate sizes (Fig. 3B) indicates that the enzyme has a preference for binding and cleaving at the xylosylated XGO-SR tag, which is positioned at the reducing end of the HEC:XGO-SR conjugate. It should also be noted that during chemical modification of cellulose with hydroxyethyl groups, the HEC product is likely to be substituted primarily on the more reactive C-6 hydroxyethyl groups, perhaps in a block-wise fashion (44). If this were the case, the HEC substrate might represent a structural analog of xyloglucan. However, the barley (1,3;1,4)-␤-D-glucan clearly acts as a donor substrate, and cello-oligosaccharides act as acceptor substrates. We therefore believe that the potential for these hetero-transglycosylation reactions to occur in vivo warrants further investigation.
The substrate specificity of XET enzymes, which involves cleaving a (1,4)-␤-D-glucosyl linkage in the donor substrate before transfer to the nonreducing end of the acceptor substrate, would suggest that the HvXET5 re-forms a (1,4)-␤-linkage between the reducing end glucosyl residue of the donor polysaccharide, whether that be the HEC or the barley (1,3;1,4)-␤-D-glucan, and the nonreducing end of the XXXG-SR acceptor substrate. It is considered unlikely that the polymeric donor molecules would be attached to the xylosyl residues of the XXXG-SR acceptor substrate.
The rate of the reaction catalyzed by the HvXET5 enzyme described here with HEC is comparable with that on TXG ( Table 2). Values for the K m and k cat constants with TXG were 3 mg⅐ml Ϫ1 and 1 ϫ 10 Ϫ7 s Ϫ1 , respectively, and for the acceptor substrate XXXGol the values of K m and k cat were 69 ϫ 10 Ϫ6 M and 1.5 ϫ 10 Ϫ7 s Ϫ1 , respectively. 4 The rate of the reaction with (1,3;1,4)-␤-D-glucan is relatively slow ( Table 2), but it would be anticipated that contact between a large molecular mass donorenzyme complex and the nonreducing terminus of the acceptor substrate might not occur quickly. If the HvXET5 enzyme were 4 M. Hrmova and G. B. Fincher, unpublished data. to catalyze these reactions in vivo, cellulose, (1,3;1,4)-␤-D-glucans and xyloglucans in barley walls could potentially be linked to create a very large, continuous molecular network within the wall and would significantly alter the strength of walls, their porosity, and flexibility (Fig. 6). Potential accessibility and diffusion limitations in the cell wall environment could greatly reduce the catalytic rates in muro. The cell might compensate for this through the synthesis of relatively large amounts of stable enzyme, and this would be consistent not only with the high levels of XET mRNA transcripts that are found in plant cells (23 and data not shown) but also with the long term stability of the HvXET5 observed here. In any case, it has not yet been demonstrated that the ability of XETs to covalently link different polysaccharides has in muro significance. To do so would require the isolation of a short fragment of 10 or fewer glycosyl residues that can be clearly shown to originate from two distinct polysaccharide types (such as cellulose and xyloglucan). These "linkage regions" of two different polysaccharide types will only constitute a tiny proportion of total wall polysaccharides and have not been detectable using current technologies.
Emerging information on the remodeling of fungal cell walls during spore formation and under stress indicates that glycosylphosphatidylinositol-anchored transferase enzymes, some of which are members of family GH16, might also be involved in linking different polysaccharides such as ␤-D-glucans and chitin in the wall (45). There are also indications that pectic polysaccharides might be covalently linked with xyloglucans in plant cell walls (10 -13). However, the purified HvXET5 enzyme did not link polygalacturonan or ␤-D-galactans to xyloglucan, nor did the HvXET5 enzyme link arabinoxylans to xyloglucans, despite suggestions based on molecular modeling (23) that this was a possibility. However, there are multiple isoforms of XETs in plant cells (23,(33)(34)(35)46), and it remains possible that other isoforms might prefer different donor and acceptor substrate specificities. Preliminary data show that the partially purified HvXET6 enzyme has lower catalytic rates with celluloses and (1,3;1,4)-␤-D-glucan (not shown).
If covalent linkages between different polysaccharides do occur in plant cell walls, they will have important implications for wall rigidity, strength, and porosity. A thorough understanding of covalent linkages between wall polysaccharides would also provide opportunities to genetically manipulate agro-industrial processes such as paper production, food quality and texture, malting and brewing, bioethanol production, dietary fiber, and ruminant digestibility. We are now generating transgenic barley lines, in which selected HvXET genes have been up-and down-regulated, to further investigate this potential.