Research CommentariesBOTANY

A Recipe for Cellulose

See allHide authors and affiliations

Science  30 Jan 1998:
Vol. 279, Issue 5351, pp. 672-673
DOI: 10.1126/science.279.5351.672

Cellulose [HN2], [HN3], [HN4], [HN5] is the world's most abundant biopolymer. These long paracrystalline cables, each containing several dozen glucose polymers, are spooled several times around plant cells to form the structural framework of the primary cell wall. Despite the efforts of many researchers since the mid-1960s, prolonged synthesis of cellulose has never been achieved in the test tube because the cellulose synthase machinery from flowering plants is so fragile. The enzyme polypeptide, too, has been elusive, even with use of affinity probes that mimic the substrate. In an age where tens of thousands of genes have been characterized, until very recently not a single gene encoding a higher plant polysaccharide synthase had been identified. This situation has now changed dramatically with a report by Arioli et al. on page 717 of this issue, in which a gene from the plant Arabidopsis is shown definitively to be responsible for cellulose synthesis (1). [HN6], [HN7]

Early progress in the identification of synthase genes came from studies in bacteria. [HN8] Acetobacter xylinum [HN9] and Agrobacterium tumefaciens [HN10] make cellulose that they extrude extracellularly and, unlike higher plant cellulose synthases, the bacterial synthases remain active when isolated. As a result, genes encoding polypeptides of the Acetobacter xylinum cellulose synthase were identified within an operon containing four genes (see figure) (2). Elation at this finding gave way to disappointment as the bacterial cellulose synthase clones, when used to probe plant libraries, failed to identify plant cellulose synthase homologs.

Two by two.

A model of cellulose synthesis that accounts for the 180° flip of each glucose unit with respect to its neighbor. Two glycosyl transferase activities operate cooperatively from opposite sites to add cellobiosyl units to the chain. Within three synthase genes (lower part) there are multiple domains for binding substrate (U-1 through U-4), UPD-Glc, and essential aspartyl residues (in red).

The next breakthrough came as a result of two studies. First, when the deduced amino acid sequences of suspected bacterial synthases were subjected to hydrophobic cluster analysis, Saxena and colleagues (3) discovered that the synthase genes and those of other nucleotide-sugar-requiring enzymes contain short sequences of exceptionally high conservation, which are thought to be critical for uridine 5′-diphosphate (UDP)-glucose (UDP-Glc) binding and catalysis (see figure). Further, synthases able to make β-D-(1Ø4)-glycosyl linkages contain an additional domain and a total of four sequences of high similarity. In the second study, Pear et al. (4) found two plant homologs containing all four of the UDP-Glc-binding sequences of the Acetobacter cellulose synthase gene by a search of sequences in a cDNA library [HN11] made from transcripts of cotton fibers taken at the onset of secondary wall cellulose formation. The plant CelA genes are highly expressed in fibers at the time of active secondary wall cellulose synthesis, encode polypeptides of about 110 kD, are predicted to have eight transmembrane domains, bind the substrate UDP-Glc, and contain two large domains unique to plants. After three decades, a prime candidate for a cellulose synthase gene had finally been identified.

Although the cotton CelA gene is a likely candidate for the catalytic cellulose synthase, the lack of in vitro synthesis of cellulose by the synthase has made direct proof of the function of the gene product difficult. In this issue, Arioli and his colleagues provide proof of function for a related homolog from Arabidopsis by complementation of a temperature-sensitive mutant. Several years ago, R. Williamson and his colleagues selected several Arabidopsis mutants in which the root tips swell at elevated temperatures. Actually, Williamson was using this strategy to identify microtubule mutants [HN12] that had lost the ability to direct the orientation of cellulose microfibrils. These kinds of cytoskeletal mutants are predicted to result in isodiametric expansion of the root cells instead of elongation. Not all of the swelling mutants turned out to be cytoskeletal mutants—several of them were unable to make cellulose microfibrils. A long chromosome walk to the defective gene of one such mutant revealed that it was indeed a homolog of the cotton and other suspected CelA genes.

Curiously, the defective cellulose synthase seems to be able to make the individual β-D-(1Ø4)-glucan chains but is unable to organize them into a paracrystalline array. Arioli and his colleagues correlate this defect with a disorganization of the “particle rosettes,” the plasma membrane complexes long suspected to be components of the cellulose biosynthetic machinery. The defective CelA gene maps to chromosome 4 and is located within several contiguous and overlapping yeast artificial chromosomes. Transformation of Arabidopsis with the wild-type CelA gene restores the normal phenotype. This complementation is the first evidence that a plant CelA gene functions in the formation of cellulose microfibrils.

The discovery of the CelA genes opened the door for the identification of many other cell-wall polysaccharide synthase genes. Cutler and Somerville (5), in an analysis of expressed sequence tags from the database, reported that Arabidopsis contains several homologs of cotton CelA, and that a number of other sequences, while not completely related to CelA, share significant identity to one or more of the suspected UDP-Glc-binding domains.

Why plants have so many different CelA and related genes is now a focus of investigation in a number of laboratories. The glucan chains in the cellulose microfibrils made in primary and secondary cell walls are different in their degree of polymerization, possibly as a result of different modes of catalysis and organization into paracrystalline arrays. Of the 40 or so cell types that plants make, almost all can be identified by unique features of their cell wall. The family of CelA genes may be required to encode the diverse cellulose synthases needed to build these specialized walls. For example, another Arabidopsis mutant was described in which the water-conducting cells form incomplete secondary walls (6), apparently as a result of defective cell-specific cellulose deposition. In addition, some of the related genes may well encode other glycosyltransferases, such as those that synthesize β-xylans, mannans, the backbone of xyloglucan, mixed-linkage β-glucans, or callose.

As the CelA genes were sequenced and the multiple binding sites for the substrate UDP-Glc were deduced, researchers began to rethink the biochemical mechanism of cellulose synthesis. The catalysis of this polymerization must overcome a steric problem because the β(1Ø4) linkage requires each glucose unit to be flipped nearly 180° with respect to its neighbors (see figure). To make such a linkage by addition of one sugar unit at a time, either the synthase or the growing chain must rotate 180°, or the sugar must be added in a constrained position and flip or flop alternately into the proper orientation by some other factor associated with the synthase. Because a cellulose microfibril is composed of several dozen chains, each arising from a synthase, such a drastic reorientation of enzyme or substrate would be unlikely. Saxena et al. (3) revived an old idea that the unit of addition is cellobiose, but with a new twist. With multiple UDP-Glc binding sites, Saxena and colleagues reasoned that two UDP-Glc-binding domains positioned 180° from each other make two simultaneous glycosyl transfers that add cellobiose units to the growing chain, circumventing the need for reorientation of synthase or chain. As it had not been established whether the glucosyl units were added to the reducing or nonreducing end of the chain, they also reasoned that retention of the UDP moiety at the reducing end kept the growing end of the chain activated. A similar two-site model was proposed by Carpita et al. (7), but in this model the two glucosyl units are simultaneously polymerized at the nonreducing end. Another version of this model of cellulose synthesis at the nonreducing end was presented recently, with the additional proposal that the glycosyl units form covalent intermediates with serine or threonine in the active site (8). However, Koyama et al. (9) showed by direct staining of the reducing ends and by microdiffraction electron crystallographic analysis that the glucose units are indeed added to the nonreducing end. They also invoke the two-site model for addition of glucosyl units, but the C-4 hydroxyl of the sugar to be added is activated not by a hydroxyl amino acid, but by aspartic acid—the very amino acid that hydrophobic cluster analysis of the original CelA genes had predicted to be essential in three of the four highly conserved domains (see figure).

What other polypeptides are involved in the synthesis of cellulose? The identification of a Zn-binding domain in the NH2-terminal region of CelA (10) not only indicates that the polypeptide interacts with another protein, but also provides a molecular means to identify a ligand. Although there is much yet to be learned about polysaccharide synthesis in plants, the long decades of disappointment are over, and the discovery of the first synthase genes marks a new age of progress in learning how plants make their cell walls.

HyperNotes Related Resources on the World Wide Web

General Hypernotes

Pedro's BioMolecular Research Tools is a collection of WWW links to information and services useful to molecular biologists. It provides links to molecular biology search and analysis tools; bibliographic, text, and Web search services; guides and tutorials; and biological and biochemical journals and newsletters.

The World Wide Web Virtual Library: Biosciences points to virtual library pages for Biomolecules, and Biochemistry and Molecular Biology. Each of these pages presents a long list of Web resources. The World Wide Web Virtual Library Biomolecules covers molecular sequence and structure databases, metabolic pathway databases, and other lists of Web resources. The World Wide Web Virtual Library: Biochemistry and Molecular Biology is a list of resources listed by provider.

The WWW Virtual Library: Forestry presents a list of Web resources for forestry and forestry products, including cellulose.

The MIT Biology Hypertextbook is a Web-based textbook developed for introductory biology courses at MIT. Sugars, a chapter of the textbook, describes the polymeric structure of cellulose.

The Virtual Cell is an interactive simulation of an electron micrograph of a plant cell.

Cell & Molecular Biology Online is a well-organized list of Web resources for cell and molecular biologists. For each resource, a brief description is provided.

CSUBIOWEB, the California State University Biological Sciences Web server, provides links to other Web sites on cell biology and molecular biology. The Dictionary of Cell Biology (London: Academic Press, 1995) defines operon and other terms used in this Perspective.

Numbered Hypernotes

1. Nicholas C. Carpita's Web page describes his research interests.

2. The Cellulose Electronic Network promotes communication among cellulose researchers. Cellulose Biosynthesis in Higher Plants is a review of cellulose research. Time Lapse Analysis of Cellulose Production by Acetobacter xylinum includes video clips depicting cellulose synthesis. Also available are directories of cellulose researchers and companies that process cellulose products and links to other Web sites related to cellulose.

3. The Macrogalleria at the University of Southern Mississippi presents information about polymers (natural and synthetic) and polymer science in a visually interesting way. The site includes movies and rotating molecular models, including models of cellulose.

4. Anselme Payen is a brief biography of the author of the first description of cellulose.

5. Cotton Up Close, developed by the United States Department of Agriculture, provides electron micrographs of cotton fibers and a model of the cellulose molecule.

6. The Arabidopsis thaliana Database project collects and disseminates information on the genome of Arabidopsis thaliana.

7. Arabinet is a list of Internet resources for Arabidopsis researchers.

8. The Bacterial Cell Wall is described in general terms in the Introduction to Clinical Microbiology.

9. A one-paragraph description of Acetobacter xylinum is presented by the Laboratory of R. Malcolm Brown Jr.

10. Agrobacterium tumefaciens is described in Agripedia, a multimedia encyclopedia at the University of Kentucky.

11. Cloning and Molecular Analysis of Genes by Phillip E. McClean describes cDNA cloning and cDNA libraries. These pages were developed for Intermediate Genetics, a course offered at North Dakota State University.

12. Microtubules are one of the subcellular elements described in Cellular Organelles by P. Vanderschaegen.

13. Department of Botany and Plant Pathology, Purdue University


Stay Connected to Science

Navigate This Article