Cellulose Synthase-Like CslF Genes Mediate the Synthesis of Cell Wall (1,3;1,4)-ß-d-Glucans

See allHide authors and affiliations

Science  31 Mar 2006:
Vol. 311, Issue 5769, pp. 1940-1942
DOI: 10.1126/science.1122975


A characteristic feature of grasses and commercially important cereals is the presence of (1,3;1,4)-β-d-glucans in their cell walls. We have used comparative genomics to link a major quantitative trait locus for (1,3;1,4)-β-d-glucan content in barley grain to a cluster of cellulose synthase–like CslF genes in rice. After insertion of rice CslF genes into Arabidopsis, we detected (1,3;1,4)-β-d-glucan in walls of transgenic plants using specific monoclonal antibodies and enzymatic analysis. Because wild-type Arabidopsis does not contain CslF genes or have (1,3;1,4)-β-d-glucans in its walls, these experiments provide direct, gain-of-function evidence for the participation of rice CslF genes in (1,3;1,4)-β-d-glucan biosynthesis.

Grasses, which include the common cereals, arguably represent the single most important group of plants for human societies worldwide. Foods prepared from rice (Oryza sativa), wheat (Triticum aestivum), sorghum (Sorghum bicolor), barley (Hordeum vulgare), the millets (Panicum miliaceum and Pennisetum americanum), and sugar cane (Saccharum officinarum) account for a high proportion of our daily caloric intake, and numerous forage and fodder grass species support the production of sheep, cattle, and other domesticated livestock. Maize (Zea mays) is also used widely for animal feed, and switchgrass (Panicum virgatum) and other perennial grasses are showing considerable promise as future biomass energy crops for North America (1).

In all cases, the noncellulosic polysaccharides of cell walls in the grasses, which are formally classified in the commelinoid monocotyledon group of land plants (2), are crucially linked to the grasses' widespread adoption, utility, and future potential in agricultural practice and energy production. In particular, walls of the grasses contain (1,3;1,4)-β-d-glucans, which are not present in walls of dicotyledons or most other monocotyledonous plants (24). The (1,3;1,4)-β-d-glucans have a structure that is unique in biological systems, insofar as the polysaccharide consists of an unbranched, unsubstituted chain containing a single type of monomeric unit, but with two distinct linkage types that are arranged in a nonrepeating, but nonrandom, fashion (5).

The (1,3;1,4)-β-d-glucans are components of dietary fiber that are highly beneficial in the prevention and treatment of serious human health conditions, including colorectal cancer, high serum cholesterol and cardiovascular disease, obesity, and non–insulin-dependent diabetes (6, 7). In contrast, (1,3;1,4)-β-d-glucans have antinutritive effects in monogastric animals, such as pigs and poultry (7), and are important in many cereal processing applications, including malting and brewing.

In the work described here, we have identified genes that mediate (1,3;1,4)-β-d-glucan synthesis in the grasses. The first clues to the identity of (1,3;1,4)-β-d-glucan synthase genes came from the discovery of cellulose synthase (CesA) genes by Pear et al. (8). Subsequent analyses of expressed sequence tag libraries and other gene sequence databases indicated that the CesA genes were members of a much larger superfamily of genes, which included both the CesA genes and the cellulose synthase–like (Csl) gene families (914) (fig. S1). Given the chemical similarities between cellulose and (1,3;1,4)-β-d-glucans, it appeared likely that genes encoding (1,3;1,4)-β-d-glucan synthases might be members of one of the Csl gene families (5).

The Csl gene families in most vascular plants are large and have been divided into subgroups, designated CslA to CslH (11) (fig. S1). In Arabidopsis, there are at least 30 known Csl genes and, in rice, at least 37 (11, 15). In contrast to the CesA genes, it has proved difficult to define the functions of the Csl genes. Dhugga et al. (16) first showed that a guar (Cyamopsis tetragonoloba) seed (1,4)-β-d-mannan synthase is encoded by a CslA gene, and Liepman et al. (17) recently confirmed that CslA family members from rice and Arabidopsis also encode (1,4)-β-d-mannan synthases. Thus, of the multiple Csl genes in plants, very few have been assigned a biological function.

Here, we have used comparative genomics to identify genes required for (1,3;1,4)-β-d-glucan synthesis in rice. Genetic mapping studies have revealed a high level of synteny, or conservation of genome structure, including gene order (co-linearity), in species of the common cereals (18). Because (1,3;1,4)-β-d-glucans are central determinants of the malting and brewing qualities of barley, quantitative trait loci (QTLs) of (1,3;1,4)-β-d-glucan contents of ungerminated barley grain have been investigated (19). One QTL that has a large effect on (1,3;1,4)-β-d-glucan content in ungerminated grain is located on barley chromosome 2H and is flanked by the Adh8 and ABG019 DNA markers (19) (Fig. 1). Because there was no sequence available for the ABG019 marker, we used the sequence of the Bmy2 marker, which is immediately adjacent to the ABG019 marker (19) (Fig. 1). Sequences from the Adh8 and Bmy2 markers enabled us to locate a syntenic region of about 3.5 megabases (Mb) on chromosome 7 of rice, in which synteny was confirmed by the presence of more than 10 markers common to both species. Examination of the rice genome sequence corresponding to this region revealed six Csl genes clustered in a region of about 118 kilobases (kb) (Fig. 1). The six rice genes are all classified in the CslF group and have been designated OsCslF1, OsCslF2, OsCslF3, OsCslF4, OsCslF8, and OsCslF9 by Hazen et al. (11). Besides two truncated OsCslF genes that might represent pseudogenes, including the gene designated OsCslF5 (11), five other genes of unknown function and four retrotransposon-like elements were detected in this 118-kb interval of rice chromosome 7. The other CslF genes, OsCslF6 and OsCslF7, are located on rice chromosomes 8 and 10, respectively (20). We are currently mapping the barley HvCslF genes and have so far shown that at least two of the genes map to the region of barley chromosome 2H defined by the (1,3;1,4)-β-d-glucan QTL shown in Fig. 1, close to the Bmy2 marker (20).

Fig. 1.

The CslF family was identified as the prime candidate for genes encoding (1,3;1,4)-β-d-glucan synthases. A major QTL for (1,3;1,4)-β-d-glucan content of ungerminated barley grains was identified on barley chromosome 2H by Han et al. (19) and the logarithm of the likelihood ratio for linkage (lod scores) shown here were derived from that work. Markers flanking the estimated position of the barley chromosome 2H QTL (Adh8, ABG019, and Bmy2), were used to identify a syntenic region of about 3.5 Mb on rice chromosome 7. Examination of the rice genome sequence in this region revealed a group of six OsCslF genes, close to the Bmy2 marker; the six genes were clustered within an interval of about 118 kb.

In this way, comparative genomics enabled us to identify members of the CslF group of genes as potential candidate genes for (1,3;1,4)-β-d-glucan synthases in cereals. It is noteworthy that the CslF group of genes is found only in monocotyledons (11, 13), consistent with the exclusive occurrence of (1,3;1,4)-β-d-glucans in the cell walls of grasses, cereals, and other members of the Poales (2). The possible role of the CslF genes in (1,3;1,4)-β-d-glucan synthesis was therefore tested by a gain-of-function approach in transgenic Arabidopsis plants. Arabidopsis walls contain no (1,3;1,4)-β-d-glucan, and no CslF genes are present in the Arabidopsis genome (11, 21). Thus, deposition of (1,3;1,4)-β-d-glucan into walls of Arabidopsis plants transformed with rice OsCslF genes would indicate that the gene products are required for (1,3;1,4)-β-d-glucan synthesis. This approach assumed and depended on the presence and availability in Arabidopsis of appropriate donor and acceptor substrates, precursor molecules, activators, intermediates, metal ions, cofactors, or any ancillary enzymes needed for (1,3;1,4)-β-d-glucan synthesis and deposition into the wall.

Accordingly, the full-length open reading frames of OsCslF2, OsCslF4, and OsCslF9 cDNAs were amplified from rice (cv. Nippon Bare) RNA preparations using the polymerase chain reaction (PCR), cloned into a binary vector behind the 35S promoter (fig. S2), and inserted into Agrobacterium tumefaciens, which was used to transform Arabidopsis by standard floral dip procedures. In case multiple OsCslF genes might be required for (1,3;1,4)-β-d-glucan synthesis, as observed for cellulose biosynthesis (22, 23), transformation was performed not only with single gene constructs, but also with combinations of the OsCslF genes. Southern hybridization analyses confirmed the presence of the transgenes, and transcript analyses showed that the 35S promoter was driving transcription in selected T1 and T2 Arabidopsis lines (figs. S5 and S6).

Transgenic Arabidopsis lines with high OsCslF transcript levels in leaves were chosen for further analysis, specifically with respect to the deposition of (1,3;1,4)-β-d-glucan in cell walls. In the first instance, immunocytochemical methods based on transmission electron microscopy and a monoclonal antibody raised against (1,3;1,4)-β-d-glucan (24) were used to screen for the presence of the polysaccharide in both T1 and T2 transgenic Arabidopsis lines. The antibody does not bind arabinoxylan, the (1,3)-β-d-glucan, callose, or cellodextrins, and inhibition studies show that it binds very weakly to (1,3;1,4)-β-d-oligoglucosides and xyloglucans, compared with polymeric (1,3;1,4)-β-d-glucan (24, 25). (1,3;1,4)-β-d-Glucan was detected in epidermal cell walls in three of nine T1 transgenic Arabidopsis plants examined, namely, lines A18 (Fig. 2, B and E), A28, and A29 (fig. S7, C and D). Within the epidermal layer, gold labeling in walls was not uniformly distributed. In cells where walls were labeled, the labeling was distributed all around the cell, but was often more intense in the outer periclinal wall of the epidermal cells, adjacent to the cuticle. In control experiments, preincubation of the antibody with (1,3;1,4)-β-d-glucan prevented subsequent antibody binding, so that no gold labeling was observed (24) (Fig. 2C). The A18 and A29 lines contained single copies of the OsCslF2 cDNA, whereas line A28 had two copies of the OsCslF4 cDNA (fig. S5) (26). The OsCslF transgenes were isolated by PCR from the three T1 lines and from the T2 line of A18 (Fig. 2D), completely sequenced, and shown to have no errors.

Fig. 2.

Immunoelectron microscopy with monoclonal antibodies against barley (1,3;1,4)-β-d-glucan and gold-labeled second stage antibodies show the presence of (1,3;1,4)-β-d-glucan in cell walls of the epidermal layer of leaves from Arabidopsis lines transformed with rice OsCslF genes. (A and B) Epidermal walls in leaves from 42-day-old Arabidopsis lines confirm the absence of (1,3;1,4)-β-d-glucan in wild-type (WT) Arabidopsis controls (A) and the presence of (1,3;1,4)-β-d-glucan in the T1 A18 line (B). (C and D) Gold labeling in the walls of A18 plants is not detectable after preincubation of the monoclonal antibody preparation with barley (1,3;1,4)-β-d-glucan (C), whereas the deposition of (1,3;1,4)-β-d-glucan is clearly evident in walls of plants from the T2 generation of the A18 Arabidopsis plant (D). (E and F) Leaf sections from the transgenic Arabidopsis line A18 carrying the rice OsCslF2 gene were treated before gold labeling with buffer (E) or with barley (1,3;1,4)-β-d-glucan endohydrolase isoenzyme EII (F). The abolition of labeling after the latter treatment confirmed that (1,3;1,4)-β-d-glucan was present in walls of the transgenic Arabidopsis lines. CW denotes cell wall, V, vacuole, and Ch, chloroplast.

Although extensive competition studies have demonstrated that the specificity of the monoclonal antibody against (1,3;1,4)-β-d-glucans is high (24), the presence of (1,3;1,4)-β-d-glucans in the walls of transgenic Arabidopsis lines was further examined by pretreatment of fixed leaf sections with a specific (1,3;1,4)-β-d-glucan endohydrolase before immunogold labeling (Fig. 2, E and F). Purified barley (1,3;1,4)-β-d-glucan endohydrolase isoenzyme EII (EC was obtained following heterologous expression of the corresponding cDNA in Escherichia coli (27). When leaf sections of the transgenic Arabidopsis line A18 were pre-incubated with the purified enzyme preparations before probing with the specific monoclonal antibodies, gold-labeling was essentially abolished, consistent with the enzymatic removal of (1,3;1,4)-β-d-glucans from walls of these lines (Fig. 2F). In control sections pretreated with buffer, subsequent labeling was unaffected (Fig. 2E). Again, this confirmed that the Arabidopsis lines transformed with the OsCslF2 gene contained (1,3;1,4)-β-d-glucan in their cell walls. We estimate that the (1,3;1,4)-β-d-glucan content of walls in the transgenic Arabidopsis lines is considerably less than 0.1% (wt/wt). Trethewey et al. (2) also concluded that the immunocytological procedure could detect levels of (1,3;1,4)-β-d-glucan corresponding to less than 0.1% of the wall, and noted the high sensitivity of the immunocytological method compared with enzymatic analyses in a survey of (1,3;1,4)-β-d-glucan content of walls in the Poales. Chemical characterization of such small amounts of (1,3;1,4)-β-d-glucan by methylation analysis was not possible against a background of high levels of cellulose and xyloglucans in the Arabidopsis walls and of callose in associated plasmodesmata.

The combined QTL, transcript, immunocytochemical, and enzymatic data presented here indicate that CslF genes are essential for (1,3;1,4)-β-d-glucan biosynthesis in grasses and cereals. However, the observations do not preclude a requirement for other enzymes, proteins, or cofactors in (1,3;1,4)-β-d-glucan synthesis. The generally low levels of (1,3;1,4)-β-d-glucan in walls of the transformed Arabidopsis plants, where OsCslF transcript levels were often high, would be consistent with limiting levels of other components that might be required for high-level synthesis of the polysaccharide or its transfer to the cell wall. Similarly, the preferential deposition of the (1,3;1,4)-β-d-glucan in the epidermal layers of the transgenic Arabidopsis lines, despite the fact that transgene expression was driven by the constitutive 35S promoter, could indicate that the epidermal cells contain ancillary factors that are not abundant in other cells of the leaf.

Supporting Online Material

Materials and Methods

SOM Text

Figs. S1 to S9

Tables S1 and S2


References and Notes

Stay Connected to Science

Navigate This Article