Update on Mechanisms of Plant Cell Wall Biosynthesis: How Plants Make Cellulose and Other (1→4)-β- d -Glycans

The discovery of a gene that encodes a cotton (Gossypium hirsutum) cellulose synthase (Pear et al., 1996) revolutionized and invigorated the plant cell wall community to find the genes that encode the machinery of cell wall polysaccharide synthesis. The landscape was framed by the completion of the genome sequence of Arabidopsis (Arabidopsis thaliana; Arabidopsis Genome Initiative, 2000), which gave a complete gene inventory for a model plant species, but one with many genes yet to be annotated for function. An estimated 10% of the plant genome, about 2,500 genes, is devoted to construction, dynamic architecture, sensing functions, and metabolism of the plant cell wall. Based largely on prior discoveries of function in prokaryotic organisms, most of the tentatively annotated genes are organized into gene families for substrate generation, glycosyl transfer, targeting and trafficking, cell wall rearrangement, and modification by hydrolases, esterases, and lyases (Yong et al., 2005; Penning et al., 2009). However, the biochemical activities of most enzymes involved in glycosyl transfer within these families remain to be verified, and an additional 40% of the genome encodes genes whose functions are not known. As many of these proteins contain secretory signal peptides (Arabidopsis Genome Initiative, 2000), it is reasonable to infer that some have roles in cell wall construction. The Arabidopsis cellulose synthase/cellulose synthase-like (CesA/Csl) gene superfamily, which includes 10 CesA genes and 29 Csl genes in six distinct groups, was one of the first large families to be described (Richmond and Somerville, 2000), and comparative analyses of a reference dicot, Arabidopsis, with a reference grass, rice (Oryza sativa), revealed substantive differences in family structures, adding two groups not seen in the dicot genome (Hazen et al., 2002). Extension of these annotations to compare all cell wall-related gene families of the grasses with those of the dicots reveals some correlation of family structure with the differences between plants with type I walls and those of the grasses with type II walls (Fig. 1A; Penning et al., 2009). For CesA genes and certain Csl genes, establishment of specific function for the synthases they encode comes from the analysis of mutants lacking a particular function and, in some specific examples, by heterologous expression. However, genetic approaches alone do not inform us about the biological mechanism of synthesis. The knowledge gained from molecular genetic approaches now needs to be augmented by biochemical and cell biological approaches to achieve a greater understanding of proteins and their interactions within a synthase complex, their organization at membranes, and their dynamics. This Update focuses on the biochemical mechanisms of the synthesis of a single type of linkage, the (1→4)-β-d linkage, in which one sugar is inverted nearly 180° with respect to each neighboring sugar in the chain. This linkage presents a unique steric problem for processive catalysis that all living organisms have solved but we are still struggling to understand. The CesA/Csl gene superfamily. A, Of the 10 Arabidopsis CesA genes, at least three are coexpressed during primary wall formation, and mutations in each of them, AtCesA1 (RSW1; At4g32410), AtCesA6 (PRC1; At5g64740), and AtCesA3 (CEV1 and ELI1; At5g05170), result in cellulose deficiencies, indicating that each is essential for cellulose synthesis. The irx mutants AtCesA8 (IRX1; At4g18780), AtCesA7 (IRX3; At5g17420), and AtCesA4 (IRX5; At5g44030) are deficient in cellulose synthesis specifically in secondary walls. Seven additional subgroups were identified that are the likely synthases for noncellulosic polysaccharides with backbones of (1→4)-β-d-glycans. Whereas the CesA genes of Arabidopsis, rice, and maize appear to be orthologous, the Csl genes are divergent between dicots and grasses, species that make two distinct kinds of walls. From mutants and heterologous expression studies, members of the CslA group encode the synthases of (gluco)mannans, members of the CslC group are likely to encode the glucan backbone of xyloglucans, and the rice- and maize-only members of CslH and CslF encode the synthases of the mixed-linkage (1→3),(1→4)-β-d-glucans found only in grasses (after Penning et al., 2009). B, Domain model and class-specific regions (CSRs) for three CesAs known to function in primary cell wall cellulose synthesis. Two ZnF domains (in yellow) are found in the N terminus before the first membrane-spanning domain (in blue). Eight transmembrane helices, two upstream and six downstream of the catalytic domain, are predicted to interact to form a channel through which a single β-glucan chain is secreted to the cell wall. The large central catalytic domain contains four highly conserved “U motifs” of D, DxD, D, and QxxRW, important for substrate binding and catalysis. Once thought to be a hypervariable region (Pear et al., 1996), the class-specific regions are conserved among orthologs of the same subclade and vary in the number of upstream conserved Cys residues, the number of consecutive basic amino acids, Lys and Arg, and the number of consecutive acidic amino acids, Asp and Glu, downstream from the basic residues (after Carpita and Vergara, 1998; Vergara and Carpita, 2001). This Update reviews our present state of knowledge of the biochemical mechanisms of polysaccharide synthesis, including some classic discoveries, and presents an alternative hypothesis on the biochemical mechanisms and organization of complexes involved in synthase reactions that yield (1→4)-β-d linkages. In flowering plants, cellulose is a para-crystalline array of about two to three dozen (1→4)-β-d-glucan chains. Microfibrils of 36 glucan chains have a theoretical diameter of 3.8 nm, but x-ray scattering and NMR spectroscopy indicate that some microfibril diameters could be as small as 2.4 nm, or about two dozen chains (Kennedy et al., The are at the by complexes of that a single microfibril et al., and each of the six of the is to four to six of the glucan and to 36 chains are into a microfibril et al., 2002). In the found only on the of the are about in but only the membrane-spanning and domains (Fig. in of in the the catalytic domains of the cellulose synthases are estimated to be and into the and a that in many of the structure (Fig. and with cellulose synthesis in A, of the of the of with the of secondary wall of a The the of a single from a in as seen in be in each of the In these only the membrane-spanning domains and of the CesA proteins be B, structure the in and are at the terminus of a microfibril analysis of one of these the of with of to and synthase is an et al., and cellulose synthase genes in are to those of flowering plants et al., 2002). The amino of CesAs regions of with the CesA the four catalytic the D, DxD, D, that are highly conserved among those that kinds of et al., The plant CesA genes are predicted to encode of about each with a region and the four of the catalytic domain (Fig. for the functions of plant CesA genes in cellulose synthesis from Arabidopsis mutants of three of the CesA genes involved in primary wall the et al., the et al., 2000), and a with and and and et al., in the same and an of cellulose synthesis is in each The same was with the mutants and which a of as a result of cellulose during secondary cell wall et al., hypothesis is that the and AtCesA6 proteins to function in primary wall cellulose synthesis, the and AtCesA8 proteins to make secondary wall cellulose (Fig. with each of the a function in the of one CesA of the two into the et al., However, at least some of the are as by the of and primary wall by expression of a et al., and by the in the AtCesA3 et al., 2009). AtCesA1 is essential for cellulose synthesis et al., of AtCesA3 and et al., and et al., result in synthesis but not in et that the and proteins have functions with of three distinct CesA was in and by in by et with and AtCesA1 and AtCesA3 proteins in their mutants in and with catalytic and domains important for function et al., interactions of three distinct CesA in were by et al., some gave and of and AtCesA6 are et to those of et to that these three primary wall CesA proteins they that to gave an and that but not gave complexes et al., et a of complexes of CesA to about to with the of et only complexes of two of the CesAs are the is et al., 2009). an of CesAs was in et al., 2009). four of the biochemical mechanism by which cellulose is a and 2005; Somerville, et al., with only a of cellulose synthesis in with and et al., 2002). For cellulose and the synthesis by the of residues to the terminus of the glucan chain et al., et al., The hypothesis is that each CesA a single glucan chain. In the the form a channel through which a single glucan chain is (Fig. This of synthesis comes with a steric problem for synthesis. make a (1→4)-β-d linkage that each is 180° with respect to each the of sugar of the chain is of each (Fig. For the glycosyl transfer to the of catalysis within the the chain or the catalytic or amino between two to for the steric have that two or of glycosyl transfer within the catalytic complex, that are et al., et al., Carpita and Vergara, 1998; et al., et al., or that two to form two catalytic et al., Vergara and Carpita, 2001). In or are to the to that the (1→4)-β-d are of or terminus of the chain (Fig. for cellulose synthase and the steric problem of a A, The first model of for a single CesA was by CesA interact with to form the synthase The conserved region and class-specific region are from B, The steric problem of synthesis is in and of a single glycosyl in a (1→4)-β-d linkage of the of the chain or the of the synthase result in of the The to the steric problem is a catalytic of glycosyl transfer to form a to the of the of the chain. of an number of the as the as the chain is In the catalytic one of the is the of the which the of as the (after et al., 2001). the for a model of biochemical from of polysaccharide synthases indicate that a single is is an polysaccharide of of and and synthases in three distinct with I proteins to the the (Fig. and synthases have to contain activities in a single and et al., et al., a that the synthesis of a single only a single of A, I B, II synthases (after and The structure of a family sequence with a of the catalytic domain of a the the first of the and the of the residues in the of the group of a (Fig. and and that only a single for a substrate is within a single of of type glycosyl A, The synthase as a with a single binding for and B, The as a each with a single binding domain for of the two or domains From the of the I synthases and the it is that a single alone all the for synthesis. However, these still do not the steric problem of synthesis of a linkage of the cellulose glucan chains. In two that at least some synthases and of synthase form An structure to the synthase is the from which is predicted to contain a single In to the as a and each a (Fig. The through the which appear to be and, for were from the structure of the synthase and to our is the first by structure of a by but are For an contains two domains within a single each binding a or and it as a a of four domains (Fig. the is not the CesA catalytic domain through the of all for known of glycosyl and it is with the by and and et of a within which the catalytic domain of a single CesA the synthesis of of the chain within a single The II synthases contain two of but not the membrane-spanning domains (Fig. type of synthase two of the and the synthesis of the of by a single synthase is and However, a of two synthases is for synthesis to the that in which each is are still to make and and the I synthases have activities on a single not the that the of of single of synthase is for the four the steric problem one a channel of for the cellulose synthase is of to a (1→4)-β-d-glucan chain. The about channel was with respect to the synthases by and that certain for with the to the channel for However, interactions with a small number of domains be is still in synthases are about the of CesAs and contain those of a et al., 2001). and of and are et al., and all of contain a of to as many as transmembrane et al., et al., 1998; et al., et al., The of CesA complexes to to et that are involved in the into of the the region of as the which to domains that in a 2000), function in the organization of these into the et that CesAs are through the ZnF domain in a the first in the of CesAs into the discovery that a with the CesA ZnF domain in a to the that of the domains by the CesAs to which are to the et al., 2002). The of the chains into cellulose the et al., 2001). This be in the of that the by et al., 2002). ZnF domains of two CesAs as of the into the is all the interact to form a complete not the by et that CesA proteins and the ZnF domains were to of a single et that interactions by analysis do not the ZnF these that domains the of the CesA in reactions two to three dozen CesAs or are to form a of CesA domains within the catalytic The of CesA and the within the to the that region was not essential in catalysis (Pear et al., However, it is now that these regions are conserved and dicot species with a distinct subclade interactions through of region conserved Cys residues, of consecutive basic Lys and residues, and of acidic Asp and residues form the of a class-specific region (Fig. Vergara and Carpita, 2001). the catalytic we proteins only the catalytic domain of Arabidopsis and maize CesAs with and that and to by and the ZnF was in these through in the catalytic an of CesAs to form or the three basic of the single the steric problem is solved by synthesis and of of the of the of at the terminus of the a channel of membrane-spanning domains is with sugar and and the two ZnF domains for of the catalytic into (Fig. An is that of be CesAs to have a structure the and function II a CesA catalytic with four domains be of two (1→4)-β-d-glucan chains of are to the of and the of the ZnF in of the catalytic into catalytic hypothesis for cellulose A, catalytic model of two CesAs to form a that a single (1→4)-β-d-glucan chain. or of CesAs that which are to the of the glucan chain. in a channel of membrane-spanning to that of synthase and with B, in two ZnF domains that are now to two of complexes interact to one of the synthase a of for a et al., of in the and they be secreted to the to cellulose synthesis and proteins are to be for the of of synthesis, of of the and of chains et al., Somerville, et al., a of proteins that certain CesAs are at within the catalytic and domains et al., of some with amino that or that synthesis or interactions with the et al., In with an was found to be with a the of synthase et al., and was identified as synthase of was et al., and and that the of a mechanism to cellulose are from and on et al., was with CesA proteins in the et al., the of an of with cellulose synthases for of through a of However, a that all in not cellulose synthesis et al., 2009). of in of small but in cellulose et al., 2009). not appear to be for cellulose synthesis but by substrate at the of synthesis. et that from as of glucan chain with the the from the chain. and as they found that mutants of two synthases result in in but not in cellulose synthesis. However, the are in the but not the the a et al., a dynamic with the and the that and the cellulose synthase and that function with the of the many polysaccharides to it from as for CesA interactions et al., important to in interactions within a in of the of cellulose synthase complexes at the et al., and et al., 2009). the of the with the in As by and and these to on the of and cellulose through with and in the and in are still that of the of to the of cellulose synthesis, and these et al., and et al., 2009). of the same conserved domains of binding and catalysis as those of Csl genes were predicted to encode the synthases of noncellulosic with and xyloglucans, and the (1→3),(1→4)-β-d-glucans and Somerville, et al., 2002). For the most to be but are a is still an knowledge of most of the gene and interactions among to make specific among families at least one a glycosyl is a between the classic on the synthesis of these polysaccharides in and the discovery of genes the machinery that a The mixed-linkage is in the grasses Carpita, et al., certain et al., and et al., et al., but differences in the of their indicate that they by of For the grasses, glucan is not a of and but is of and in a of about by single et al., with a of are with the in the in the The synthesis of the in with and and Carpita, et al., et al., Whereas of result in and in et al., polysaccharides in are with of substrate et al., 2001). The of and are with substrate et al., indicating that the mechanism of synthesis of the is from synthesis of the that the of catalysis is on the et al., with specifically their to make the the synthesis of the and was of the to make is with in of the et al., a catalytic model are by cellulose synthase-like proteins and the by an additional that yet to be identified et al., and with the mixed-linkage of species for three of glycosyl transfer within the catalytic two from the cellulose synthase-like domain and a et al., is the of cellulose synthase at the or of the to make the for synthesis the to the et al., As the of cellulose synthase at the the synthase with cellulose are cellulose synthase and the synthase et al., to synthesis of the by of the complete to a single glycosyl et al., et al., of the to a single of glycosyl transfer make only et al., whose synthesis not the catalytic or 180° (Fig. it be that synthase in in of all et al., only from grasses, the only that make the make and Carpita, et al., The most for the synthesis of from a cellulose synthase comes from the of et that of a cellulose synthase of the the to make in Whereas some cellulose is in with cellulose synthase is a single gene in of the to make in as as (1→4)-β-d-glucan in from the that the single is for Two groups of Csl genes, CslF and which are found only in grasses (Hazen et al., have to expression of a rice CslF in Arabidopsis, a species that not make in small of the in the cell walls et al., However, greater of the result a CslH is coexpressed with a for CslH and CslF in the synthesis of the polysaccharide and that a catalytic the et al., 2009). an glycosyl is to make the Arabidopsis a This of by two distinct group members the that synthases of be by Csl genes of of the first cell wall polysaccharides to be in was An that is the substrate for (1→4)-β-d of cellulose and that is the substrate for and yet is still as the substrate for cellulose synthesis on most wall of biochemical with cotton in that is the substrate for cellulose synthesis and of and to in of into a and et al., The knowledge of as was in the discovery that a CslA gene encodes a synthase by expression of a species that large of as a cell wall et al., et that at least four members of the CslA group function in synthesis. However, of and in the heterologous expression et al., do not the of synthesis in and the CesA genes appear species, the Csl genes are not et al., 2009). In the CslA group is into three subgroups that are Arabidopsis or The CslA members as synthase genes et al., into the and the subgroups et al., 2009). The functions of these members of the grasses to be were among the first cell wall polysaccharides whose synthesis was in that from and are into polysaccharides and were by of et al., of with the are with small of and but and found that of each are to large the structure are The of the backbone and the three to make the structure are with an synthase for the glucan in the of two are a backbone is by the of the group with an et al., An is the of certain as to make in to et al., of with a which only at and backbone whose that the of by of a of and et al., This is an result for two the of these of the of the and to make a as many as residues be within the in to be to the the of group of these a was to the sugar of the with et al., a backbone with three residues a For many is with of at the two residues to the and is to the first a is In to the glucan backbone at least three of glycosyl and as many as are to all the of group of these a was to the sugar of the with et al., a backbone with three residues a For many is with of at the two residues to the and is to the first a is In to the glucan backbone at least three of glycosyl and as many as are to all the The of synthesis at the is still of that synthesis on for for backbone synthesis of in and is to for transfer within the et al., et that synthesis of the backbone is the of cellulose but the of all of and are within the of the Based on heterologous expression in et that CslC genes encode the synthases of the is to make of the with CslC is to the of (1→4)-β-d-glucan indicating that a of these proteins the synthase to of the to of function of is the of a CslC at the of the et al., 2009). members of the family are as the involved in synthesis, and these proteins are predicted to the et al., et al., is in in a et al., 2005; et al., are with a et al., but as those of species, have a structure with of et al., 1996), and the and species have of these two of et al., the in the are in that make from those that are the cellulose in the wall et al., 2009). The of a to the discovery of a gene it et al., the synthase a between the glucan synthases and the that it et al., the of the and be from the are to from et al., and from et al., et al., to in The synthesis of from with was first by and of two in and in the synthesis of with and et al., have that from grasses and of in a that and and and et al., et al., from a that the be involved et al., in the synthesis of it yet to be that Csl gene a In approaches yield genes as likely for the machinery for synthesis et al., and mutants with in synthesis, as et al., et al., et al., et al., 2005; et al., and et al., and and et al., do not a of the Csl gene synthesis to a et that mutants deficient in and were of a at the of the an deficient in an of the The or of the of the In and a is with some in the of have with nearly all of the et al., a model chains of are by the and these are the to make the polysaccharides and The and genes encode members of the group and genes encode members of the group that the and encode members of that are likely to encode the synthases of the backbones that are by a yet glycosyl et al., et al., family members encode and the members encode with respect to the with the linkage in the members make the members make or linkages. an is found in the groups are at the chain and and a synthesis of six consecutive residues in is and and appear to encode from but mutants of these genes have the chains et al., 2009). be to the of these groups a as an or chains are to make a a by the on the of and in the in synthesis of and et al., and a of synthesis and have that not be In their a type of mechanism is for the of the a mechanism to the synthesis described for (1→4)-β-d in are the of the and of the and and the between the of that is in the secondary and the with that is the of the primary cell walls of grasses and approaches of synthase complexes from a of and family members with a an that the and for of the into the polysaccharide et al., kinds of the to all the of synthase complexes whose activities are in the plant cell wall community is in biochemical functions of backbone synthases and the glycosyl that them, the biochemical of these complexes are still from are a many of polysaccharide synthesis in that to the differences between an and the function of a as a The differences between the of complexes in and in are by the of the synthases and glycosyl at the and membranes, which and are with cotton in that of the was essential to cellulose synthesis, and that of the the to that was to the of synthesis and This was to that synthesis in and In of cellulose synthesis could be through of the in living that not metabolism et al., and, in to cellulose synthesis at the it is of a that the synthesis of the mixed-linkage in in maize and Carpita, The and biochemical for the of these and on polysaccharide synthesis are still not do to the biochemical mechanism of synthesis, that not only the complexes but their to be to the synthesis of I and of for their of the and their many I for to the on structure and The in is by