Biochemical Mechanisms of Synthesis of (1 -> 3), (1 -> 4) β-D--D-Glucans: Cellulose Synthase with an added Twist?

B. R. Urbanowicz, C. J. Rayon, N. C. Carpita, M. A. Tiné and M. S. Buckeridge

1 Introduction

Synthases of all polymers containing (1 -> 4) β-linked glucosyl, mannosyl and xylosyl residues have overcome a substrate-orientation problem in catalysis because this particular linkage requires that each of these sugar units be inverted nearly 180° with respect to its neighbors. We and others have proposed that this problem is solved by two modes of glycosyl transfer within a single catalytic subunit to generate disaccharide units, which maintain the proper orientation without rotation or re-orientation of the synthetic machinery in 3-dimensional space. A variant of the strict (1 -> 4) β-D-linkage structure is the mixed-linkage (1 -> 3),(1 -> 4) βD-glucan (β-glucan), a growth specific cell wall polysaccharide found in grasses and cereals and other members of the Poales. β-Glucan is composed primarily of cellotriosyl and cellotetraosyl units linked by single (1 -> 3) β-linkages. In reactions in vitro at high substrate concentration, a polymer composed of almost entirely cellotriosyl and cellopentosyl units is made. These results support a model in which three modes of glycosyl transfer occur within the synthase complex, but the generation of odd numbered units demands that they are connected by (1 -> 3) β-linkages and not (1 -> 4) β. We propose that a central part of the β-glucan synthase complex is derived from an ancestral cellulose synthase, and that an additional glycosyl transferase associates with it to generate these odd numbered cellodextrin units. In contrast to xyloglucan and pectin synthases, which are completely enclosed within the lumen of the Golgi apparatus, we provide evidence from limited proteolysis experiments that the catalytic domain of β-glucan synthase is oriented to the cytosolic side of the Golgi membrane and extrudes β-glucan through a channel into the lumen. Thus, the β-glucan synthase is the topological equivalent of cellulose synthase.

1.1 The Unique Twisted Structure of β-Glucan

Mixed-linkage (1 -> 3),(1 -> 4)β-D-glucan (β-glucan) is a plant cell wall polysaccharide specific to grasses and cereals that appears during cell expansion. In all cereal endosperm walls, the ratio of cellotriosyl and cellotetraosyl units is between 2 and 3. The polymer can be synthesized in vitro from enriched Golgi membrane fractions with UDP-Glc as a substrate. However, whereas a polymer can be made that is similar to that synthesized natively at substrate concentrations between 100 and 250 µM, at high substrate concentrations nearing saturation, β-glucan is mostly composed of cellotriosyl units and the next higher odd numbered cellodextrin.

1.2 Synthesis of β-Glucan at the Golgi Apparatus

As with all plant cell wall polysaccharides, the mechanism of synthesis of β-glucan is still not well understood. There is a steric problem associated with modeling (1 -> 4)β-linked polymers, such as cellulose and β-glucan, because each sugar unit is inverted 180° with respect to its neighbor. One proposed mechanism of (1 -> 4)β-D-glucan (cellulose) synthesis involves two glycosyl transferases operating opposite one another to add cellobiosyl units to the growing polymer. The synthase properties of β-glucan resemble those of cellulose more closely than those of other Golgi-associated polysaccharide synthase complexes, leading us to believe that β-glucan synthase may have derived from an ancestral cellulose synthase gene. The cellulose synthase genes (CesA) encode polypeptides of 110 kDa that synthesize (1 -> 4)β-D-glucosyl units. The proteins that are predicted to be encoded by these genes contain up to eight membrane-spanning domains and a large cytosolic active site with "U" motifs containing conserved aspartate residues and a QxxRW sequence. They are conserved in genes encoding several processive glycosyltransferases in which repeating β-glucosyl structures are synthesized.

A unique feature of CesA genes is a plant-specific region, formerly called the Hypervariable Regions (HVRs). However, these regions possess motifs of consecutive acidic and basic residues very well conserved in a sub-class-specific manner indicating a role in catalysis or regulation. Hence, we have now suggested that they be called Class-specific Regions (CSRs). A phylogenetic tree of the CesA gene family based on their CSR showed that CesA sequences from cereals are more closely related to each other than to those from dicot species. The diversity of polysaccharide structures encoded by the four U-motifs and CSR structures and the size of the CesA gene family suggest that some of them encode synthases for polysaccharides other than cellulose. These analyses suggest that (1 -> 3),(1 -> 4)β-D-glucan synthase is one such candidate.

The mixed-linkage (1 -> 3),(1 -> 4)β-D-glucan synthase possesses more similarities to cellulose synthase than do synthases of other non-cellulosic polymers bearing (1 right arrow] 4)β-linked backbones. Buckeridge et al. proposed a model of synthesis for β-glucan that builds off that for cellulose synthesis, which includes multiple sites of glycosyl transfer in the synthase complex and favoring a "three-site" model of substrate binding and glycosyl transfer that would add cellotriosyl units to the growing polymer. Proposed models for cellulose and β-glucan synthesis account for the synthesis of (1 right arrow] 3)β-[D]-glucan (callose) if there is a loss of all but one site of glycosyl transfer. Although this mode of synthesis has been indicated by our data, it is still not known how the β-glucan synthase complex is oriented topologically with respect to the Golgi membrane or how the components are organized. Muñoz et al. and Neckelmann and Orellana suggested that in pea Golgi a UDP-Glc transporter transports UDP-Glc into the lumen of the Golgi where Glc is then consumed in the synthesis of xyloglucans. They proposed further that all of the enzymes involved in xyloglucan synthesis reside entirely within the lumen of the Golgi. Taking a step further, if the β-glucan synthase borrowed a mechanism of synthesis directly from cellulose synthase, then its membrane topology would be expected to place the active site of UDP-Glc binding to the cytosolic face of the Golgi membrane and that, like cellulose synthase, the glucan product is extruded through a membrane channel into the lumen of the Golgi.

1.3 Metabolic Channeling of Glucose to Cellulose Synthase

Amor et al. reported that sucrose synthase (SuSy) from cotton fibers is tightly associated with cellulose synthase where it channels UDP-Glc from sucrose to cellulose synthase in the plasma membrane. In maize, SuSy is detected immunocytochemically in both plasma membrane-enriched and Golgi membrane-enriched fractions, whereas in soybean, the synthase is only detected in plasma membranes. If β-glucan synthase is derived from an ancestral cellulose synthase and possesses a similar mechanism of synthesis, then part of this mechanism would include the association of SuSy to control the supply of UDP-Glc to the active site.

Limited proteolysis studies of vesicle fractions have been a useful method to probe topology of many membrane-associated protein complexes. On the other hand, the use of detergents to solubilize β-glucan synthase, such as callose synthase, has been studied to understand the importance of the phospholipid environment in the activity of this membrane-associated synthase. The susceptibility of the β-glucan synthase complex to protease digestion and detergent treatments, indicated by loss of synthase activity, will help us determine whether the synthase complex is oriented on the cytosolic face of the Golgi membrane or, like xyloglucan and polygalacturonic acid synthases, fully contained within the lumen.

2 Experimental

2.1 Plant Material and Membrane Preparation

Maize (Zea mays L.) seeds were soaked in the dark overnight at 30°C in deionized water bubbled with air. They were then sown into moistened medium-grade vermiculite, and grown in darkness for 2 d. Fresh maize coleoptiles and etiolated shoots within were harvested into a chilled mortar and overlaid with an equal volume of ice cold homogenization buffer consisting of 100mM HEPES[BTP], pH 7.4, 20mM KCl, and 84% (w/v) sucrose. In addition, 1 g of activated charcoal per 10 g of plant material was sprinkled on coleoptiles before addition of homogenization buffer to absorb inhibitory phenolics released by the maize during mashing. After mashing, the homogenate was squeezed through nylon mesh (47 µm2 pores), and 20 ml of the homogenate was then pipetted into a 38.5 ml Ultraclear centrifuge tube (Beckman) and overlaid with 8 ml of 35% (w/v) sucrose in a gradient buffer containing 10mM HEPES[BTP], pH 7.2, 7 ml of 30% (w/v) sucrose in gradient buffer, and 4 ml of 9.5% (w/v) sucrose in gradient buffer. The sucrose gradient was centrifuged at 140 000 g in a swinging bucket rotor (Model SW28, Beckman) for 60 min, and the interface enriched in Golgi membranes (35%/30% interface) was removed with a Pasteur pipette and used directly for β-glucan synthase reactions.

2.2 Glucan Synthase Reactions

Reactions were performed with 500 µl (342 µg) of freshly isolated membranes in 4 ml borosilicate glass vials containing 500 µl of reaction buffer (RB), at a final concentration of 250 µM UDP-Glc (Sigma) in 10mM HEPES[BTP], pH 7.4, containing 1.08 M sucrose, 20mM KCl, 15mM MgCl2 and 1 µCi of UDP-[U-14C]-Glc (320 mCi/mmol; Amersham Pharmacia Biotech UK Limited), and placed at 30°C for 2 h. Reactions were stopped by addition of 3 ml of ethanol, and the suspensions were heated at 105°C for 5 min. The suspensions were then cooled to room temperature and centrifuged for 5 min at 10 000 g. The pellets were washed five times with 80% (v/v) ethanol and heated in a 105°C oven for 5 min, and then were shaken, cooled and recentrifuged. The washed pellets were dried under a stream of nitrogen gas at 45°C.

Reactions were performed the same as stated above with the addition of proteinase K (Sigma, EC 3.4.21.64) to the RB containing 250 µM UDP-Glc without radioactivity at concentrations indicated in a final reaction volume of 1 ml. At 30 min into the reaction, samples were placed on ice and 10 µl of 50mM PMSF (in ethanol) and 1 µCi of UDP- [U-14C]-Glc were added to the samples. They were then incubated for an additional 90 min at 30°C.

For the experiments with detergents, 500 µl of RB containing Triton X-100 or CHAPS were added to an equal volume of Golgi membranes to reach final concentrations up to 0.1% (v/v) of Triton X-100 and up to 0.6% (w/v) of CHAPS. The incubation was carried out at 4°C for 30 min. At 30 min, 1 µCi of UDP-[U-14C]-Glc was added to the samples, and they were incubated at 30°C for an additional 90 min.

2.3 Digestion with Bacillus subtilis (1 -> 3),(1->4)β-D-Glucan endo-4-glucanohydrolase (β-Glucan endohydrolase) as a Specific Assay for β-Glucan

A B. subtilis β-glucan endohydrolase selectively cleaves a (1 -> 4)β-D-glucosyl linkage only when it is preceded by a (1 -> 3)β-linkage, making it useful for quantifying the amount of polymer synthesized. The reaction products were resuspended in 100 µl of water and 50 µl of β-glucan endohydrolase in 20mM sodium acetate and 20mM NaCl, pH 5.5 were added. The products released from the digestion of purified β-glucan yield mostly cellobiosyl- and cellotriosyl-(1 -> 3)β-D-glucose. The samples were incubated for 3 h at 37°C and stopped by boiling for 2 min, cooled and microcentrifuged at 14 000 rpm for 5 min. The labelled oligosaccharides from β-glucan were separated by high performance anion-exchange chromatography pulsed amperometric detection (HPAEC-PAD), followed by liquid scintillation counting. The pellet was resuspended in 1 ml of water, and the radioactivity was determined by liquid scintillation spectroscopy.