Report

Sitosterol-β-glucoside as Primer for Cellulose Synthesis in Plants

See allHide authors and affiliations

Science  04 Jan 2002:
Vol. 295, Issue 5552, pp. 147-150
DOI: 10.1126/science.1064281

Abstract

Cellulose synthesis in plants requires β-1,4-glucan chain initiation, elongation, and termination. The process of chain elongation is likely to be distinct from the process of chain initiation. We demonstrate that a CesA glucosyltransferase initiates glucan polymerization by using sitosterol-β-glucoside (SG) as primer. Cotton fiber membranes synthesize sitosterol-cellodextrins (SCDs) from SG and uridine 5′-diphosphate–glucose (UDP-Glc) under conditions that also favor cellulose synthesis. The cellulase encoded by the Korrigan (Kor) gene, required for cellulose synthesis in plants, may function to cleave SG from the growing polymer chain.

Cellulose (β-1,4-glucan) microfibrils provide strength and flexibility to plant tissues and are also of great importance to wood, paper, textile, and chemical industries. Genetic evidence implicates plant CesAgenes, homologous to bacterial cellulose synthases, and theKorrigan (Kor) gene, a membrane-associated cellulase, in cellulose synthesis (1–5).

Herbicides that disrupt cellulose synthesis include 2,6-dichlorobenzonitrile (DCB), which acts in a way that is not understood (6), and isoxaben, which may interact directly or indirectly with certain CesA proteins (7). Another herbicide, thiatriazene-based CGA 325'625, inhibits synthesis of crystalline cellulose and causes accumulation of a noncrystalline cellulose, which, upon treatment with cellulase, releases CesA protein (8) and a small amount of sitosterol linked to glucose (9). Because hydrophobic glycosides can function as primers for other glycosyltransferases (10–14), sitosterol-glucoside (SG) may serve as a primer for glucan chain elongation, an idea supported by observations that it is synthesized on the inner face of plant plasma membranes (15), where cellulose synthesis occurs. An enzyme, UDP-Glc:sterol glucosyltransferase (SGT), which is responsible for synthesis of sterol-β-glucosides, has been found associated with plasma membranes in various plants, including cotton fibers (16). Diglucosyl and triglucosylsterols of unknown function have also been isolated from rice bran (17).

When cotton fiber membranes are incubated with 14C-labeled UDP-Glc in Tris buffer (18), two major compounds soluble in chloroform:methanol (C:M) are synthesized that migrate on thin-layer chromatography (TLC) as SG and SG acylated with palmitic acid (ASG) (Fig. 1A, lane 1). The glucose is in terminal linkage to sterol, >95% of which is sitosterol (19). Replacing Tris with Mops buffer leads to additional synthesis of the sterol cellodextrins (SCDs) S-cellobiose (SG2), S-cellotriose (SG3), and S-cellotetraose (SG4) [Fig. 1A, lane 2, and (20)]. Similarly, use of Mops favors synthesis of noncrystalline and crystalline cellulose over callose (β-1,3-glucan), and Ca2+ is required for synthesis of all glucans (Fig. 2B). Synthesis of SG, SCDs, and cellulose is highest in membranes isolated from fibers at the stage of active secondary-wall cellulose synthesis (Fig. 1C), when CesA and Korrigan protein levels are also highest (Fig. 1D).

Figure 1

Synthesis of SG, SCDs, and glucans by using cotton fiber membranes. CC, crystalline cellulose defined as glucan remaining insoluble after treatment with acetic-nitric reagent (28); NCC, noncrystalline cellulose defined as 4-linked glucan not resistant to acetic-nitric reagent. (A) Radioautogram of TLC separations of sterol derivatives synthesized from UDP-[14C]Glc (18). Lane 1, buffered with Tris-HCl; lane 2, buffered with NaMops; Glc2 refers to cellobiose. (B) Synthesis of glucans by fiber membranes. Callose is glucan in β-1,3 linkage (18). Lane 1, buffered with Tris; lane 2, buffered with Mops; lane 3, buffered with Mops and CaCl2 in reaction replaced with 5 mM EGTA. (C) Synthesis of SG, SCDs, and glucan as a function of age of fibers used to harvest membranes (200 μg per reaction). 10 D, stage of primary-wall cellulose synthesis; 24 D, stage of maximal cellulose synthesis; 17 D, transition stage from primary- to secondary-wall synthesis. (D) Level of CesA and Kor proteins in fiber membranes (10 μg membrane protein per lane) as a function of fiber age; detection by Western blotting with antibodies against GhCesA-1 and the Kor proteins (8).

Figure 2

(A) SG and UDP-Glc as substrates for SCD synthesis in fiber membranes. Reaction mixtures contained, for lane 1, [14C]SG only; lane 2, [14C]SG2, fiber membranes (200 μg protein), 2 mM CaCl2, 10 mM MgCl2; lane 3, as for lane 2 plus 1 mM unlabeled UDP-Glc. (B) C:M soluble products synthesized by yeast membranes with the use of [14C]SG and UDP-Glc as substrates. See (20) for details of yeast expression and (18) for assay conditions using Na-Mops, MgCl2, cellobiose, and CaCl2 as effectors. Yeast and/or fiber membranes (200 μg membrane protein for each) were used per assay. Lanes: Membranes from yeast expressing D1 (Asp312Ala) mutant form ofGhCesA-1 (lane 1); lane 2, expressing wild-typeGhCesA-1; lane 3, transformed with vector only. (C) Yeast membranes containing expressed GhCesA-1enhance synthesis of cellulose when incubated with fiber membranes. 2X CesA indicates that twice as many of the yeast membranes expressingGhCesA-1 were added to reaction.

Supplying a combination of [14C]SG with unlabeled UDP-Glc to fiber membranes results in conversion of some of the SG to SCDs (Fig. 2A), indicating that a glycosyltransferase can use UDP-Glc to form SCDs from SG. Addition of [14C]SCDs to membranes results in a labeled glucan product, supporting a role for SG as a primer [Web table 1 and (20)]. Kinetic studies also support a model in which SG is elongated to SCDs and subsequently to cellulose [Web fig. 1 and (20)].

To isolate the functions of CesA, an epitope-tagged cottonGhCesA-1 cDNA was expressed in yeast (20). Yeast membranes contain some endogenous β-glucosidase activity, but only those expressing wild-type GhCesA-1 are capable of synthesizing SG3 from SG (Fig. 2B). Yeast with vector only or expressing a form of GhCesA-1 in which the critical Asp312 (D1) residue at the active site (1) is mutated to Ala, do not carry out this reaction. No cellulose synthesis was detected with the use of membranes expressing GhCesA-1, but such membranes do enhance cellulose synthesis when mixed with cotton fiber membranes (Fig. 2C). Such enhancement might occur through provision of substrate SG3 synthesized by transgene yeast membranes with the use of sterols from the cotton membranes or from provision of other factors necessary for complete chain elongation and termination.

The requirement for Ca2+ for in vitro cellulose synthesis (Fig. 1) might be explained by the fact that the Kor cellulase requires Ca2+ for activity (20, 21). A possible role for this enzyme might be to cleave sterol from the growing glucan chain to allow further chain elongation. Our reaction mixtures also produce cellobiose and glucose [Fig. 1 and (20)] that might come from cellulase-mediated cleavage of SCDs. In cotton membranes, SG2 and SG3 can be partially cleaved to SG and glucose or cellobiose by an endogenous Ca2+-dependent activity, whereas a commercial endo-β-1,4-glucanase can only cleave SG3 (Fig. 3, A and B). When putatative Kor activity and cellulose synthesis are inhibited by EGTA, ability of cotton fiber membranes to synthesize cellulose can be restored by addition of very low levels of a Ca2+-independent endo-β-1,4-glucanase (Fig. 3C). [Higher levels of this glucanase or longer incubation times lead also to degradation of the glucan product (9).]

Figure 3

Cleavage of SCDs by a Ca2+-dependent cellulase and its role in cellulose synthesis. 14C-labeled sterol derivatives were synthesized, isolated, and sonicated into Mops buffer as described (18). All reactions contained NaMops, MgCl2, and cellobiose. (A) De- gradation of [14C]SG2 to SG and Glc by fiber membranes requires Ca2+. Lane 1, [14C]SG2 alone; lane 2, [14C]SG2 plus fiber membranes and 5 mM EGTA; lane 3, [14C]SG2, fiber membranes plus 2 mM CaCl2; lane 4, [14C]SG2 plus 5 mM EGTA and 1 unit purified EG (endo-1,4-β-glucanase, from Megazyme). (B) Degradation of [14C]SG3 to Glc and cellobiose by fiber membranes requires Ca2+. Lanes as for Fig. 3A. (C) EG restores ability of fiber membranes to synthesize cellulose from UDP-[14C]Glc in the absence of Ca2+. Lane 1, 5 mM EGTA; lane 2, 5 mM EGTA, 1 unit EG; lane 3, 2 mM CaCl2.

An additional connection between the synthesis of SCDs and cellulose comes from the observation that DCB inhibits synthesis of SG in vivo (Fig. 4A). In vitro, SG and SCD synthesis is inhibited by DCB but not by CGA 325'615 (Fig. 4B). Inhibition by DCB is effective only if fibers are pretreated with DCB, not when it is added directly to isolated membranes; presumably some alteration in these activities must occur in vivo with DCB that cannot at present be mimicked in vitro. Addition of SG reverses the effect of DCB (but not CGA 325'615) on cellulose synthesis in vivo (Fig. 4C), suggesting that DCB inhibits cellulose synthesis via inhibition of SG synthesis.

Figure 4

DCB affects synthesis of SG, SCDs, and cellulose. (A) Synthesis of SG in vivo is reduced in cultured cotton fibers incubated with U-[14C]Glc, as described (8), with or without 25 μM DCB. Material soluble in C:M was extracted from membranes, chromatographed, and subjected to radioautography (18). Amounts of the extract having equal counts per minute of radiolabel were loaded per lane. (B) Effects of in vivo herbicide treatments of fibers on capacity of membranes for in vitro synthesis of sterol derivatives. In vitro assays done as for Fig. 1, lane 2. (C) The inhibitory effect of DCB (25 μM), but not CGA 325'615 (20 nM), on synthesis of crystalline cellulose (28) in vivo can be reversed by addition of isolated cotton fiber SG (7 μM final concentration) to incubation medium during herbicide treatments. Free sterols (stigmasterol and sitosterol; from Supelco) did not show a similar effect (9). SG was suspended in 1% DMSO before addition to cotton culture medium (final DMSO concentration in controls and samples was 0.01%).

We suggest a model [Web fig. 2 (20)] in which SG serves as a primer for β-1,4-glucan chain elongation catalyzed by CesA proteins. Kor likely functions to cleave SG from SCDs, which seemingly allows chain elongation to proceed more efficiently. Because the Glc moiety of SG is attached via its reducing end to the sterol, subsequent elongation should also proceed from the nonreducing end as predicted (22). It remains to be determined whether all cellulose-producing eucaryotes will be found to use SG as primer; procaryotes, lacking such sterols, either must not use a primer or must use some other compound. Because the model assumes that SG serves as primer in other plants besides cotton, it leads to certain predictions that can be tested. One is that mutants of Korrigan should accumulate SCDs that could be the “lipid-linked cellodextrins” that are reported to accumulate in kor mutants (5). Other predictions are that at least a partial cause of the severe phenotypes observed in sitosterol- reduced transgenic or mutant plants of Arabidopsis (23, 24) may be due to impairment of cellulose synthesis or that any mutants in genes encoding SGTs may lead to a similar reduction in cellulose synthesis. Finally, because yeast membranes expressing GhCesA-1 can elongate SG to SG3 but cannot synthesize cellulose except in the presence of added fiber membranes, factors other than just GhCesA-1 protein and SG must also be required for efficient chain elongation. In view of the suggestion that more than one nonidentical CesA may be required for cellulose synthesis (7, 25, 26), GhCesA-2 is a likely candidate. The challenge now is to identify the remaining factors required to reconstitute cellulose synthesis in vitro.

  • * Present address: Plant Gene Expression Center, 800 Buchanan Street, Albany CA 94710–1198, USA.

  • Present address: National Institute of Agrobiological Resources, Kannondai 2-1-2, Tsukuba, Ibaraki, Japan 305-8602.

  • To whom correspondence should be addressed: E-mail: dpdelmer{at}ucdavis.edu

REFERENCES AND NOTES

View Abstract

Navigate This Article