Analysis of the CSL gene family of rice: Cellulose synthase catalytic subunits are encoded by the large CESA family in all plants (e.g., Holland, N et al., Plant Physiol 123:1313, 2000). Plants also contain a large super-family of related genes, called "cellulose synthase-like", or CSL. Arabidopsis contains six families of CSL genes (CSLA, CSLB, CSLC, CSLD, CSLE, and CSLG). On several grounds it appears likely that CSL genes encode hemicellulose biosynthetic enzymes (Richmond, TA, Somerville, CR, Plant Physiol 124:495, 2000).
We have analyzed the CSL genes of cereals, particularly rice but also maize and sorghum (Hazen, S et al., Plant Physiol, in press, 2002; available on-line at http://www.prl.msu.edu/walton/research-cwb.htm). CSL genes are not highly represented in EST collections from any cereal. Our analysis of the rice CSL genes has been based on the Monsanto rice genome database (~50% complete), public genomic data, and the public EST collection (~100,000 ESTs). cDNAs corresponding to the ESTs were obtained mainly from the Rice Genome Project (RGP) in Japan and their sequences completed by us. Based on our experience, the rice EST collection at the Rice Genome Project in Japan (RGP) is of very high quality - virtually all of the clones that we have ordered have arrived quickly and been correct.
To date, we have identified 37 CSL genes in rice. Of these, we can reliably predict 23 full-length proteins using gene prediction software and manual alignments to the Arabidopsis Csl proteins. All of our sequences have been submitted to GenBank and are also available on-line at http://www.prl.msu.edu/walton/research-cwb.htm.
The CSL genes of rice and Arabidopsis have striking similarities as well as differences, which is consistent with the fact that the hemicelluloses of dicots and cereals are similar in some respects and different in others. For example, rice has Csl proteins that are closely related to the Arabidopsis CslA, CslC, CslD, and CslE families. Some CSL genes are physically linked in both species, e.g., the CSLB genes in Arabidopsis and four of seven CSLF genes in rice. On the other hand, rice has no members of the CSLG or CSLB families but instead has two new families, which we have named CSLF and CSLH (Hazen, S et al., 2002; and see TA Richmondís web page at: http://cellwall.stanford.edu).
As one way to test the function of the CSL genes, we are using Agrobacterium to transform rice calli with plasmid constructs designed to express double-stranded RNAs (RNAi), which has been shown in rice and other plants to result in post-transcriptional gene silencing of the endogenous message (Wesley SV et al., Plant J 27:581, 2001). We are now in the process of analyzing regenerated transgenic plants representing twelve CSL genes. Since the CSL mRNAs are not sufficiently abundant to analyze by RNA blotting, and RT-PCR cannot reliably detect quantitative changes in mRNA levels, our initial screening is being done at the whole plant level. To date, we have found that some RNAi CSL plants have no detectable phenotype whereas at least one, a plant designed to silence OsCSLC9, is dwarfed and partially sterile and has no or few root hairs.
Natural genetic variation and map-based candidate gene discovery: Hemicelluloses contribute to several complex phenotypic traits in cereals including pest resistance and digestibility of stover and silage in maize (Hedin, PA et al., J Chem Ecol 22:1655, 1996; Brice, RE, Morrison, IM, Carbohydr Res 101:93, 1982), nutritional quality and dough viscosity in wheat and rye (e.g., Vinkx, CJA, Delcour, JA, J Cereal Sci 24:1, 1996), and barley brewing clarity. Hemicellulose content has been shown to be under complex genetic control, and QTLs affecting hemicellulose composition have been identified in at least one case.
With the existing and emerging genetic resources available for cereals,
especially maize, it should be possible to identify QTLs affecting hemicellulose
composition and thereby identify the responsible genes. We found significant
(p < 0.05) differences in cell wall monosaccharide composition (arabinose,
xylose, and/or galactose) among eleven of eighteen seedling tissue types
tested and among six of eighteen juvenile tissue types. Ten inbreds were
analyzed. We have found that the inbred lines Mo17 and B73 are significantly
different for cell wall arabinose and xylose content. We can therefore
take advantage of the high-density genetic linkage map being developed
from a cross of B73 and Mo17 (called IBM) by the Missouri Maize Mapping
Project (http://www.cafnr.missouri.edu/mmp/) (Vuylsteke, M et al., Theor
Appl Genet 99:921, 1999; Davis, G et al. Maize Genetics Conference Abstracts
42:P79, 2000). We have analyzed the walls of a subset of the IBM population
with the intent to identify QTLs associated with the genetic variation
we have identified. Once the marker data is made available, these efforts
should reveal the number and impact of genes contributing to the variation.
A similar and unique study identified a major QTL associated with the ratio
of arabinose to xylose in wheat flour (Martinant, JP et al., Theor Appl
Genet 97:1069, 1998). Subsequent to mapping one or more QTLs, candidate
genes can be identified using comparative mapping databases. Completion
of the rice genome will allow the identification of candidate genes controlling
hemicellulose composition in rice based on synteny with maize.
Return to the MNL 76 On-Line Index
Return to the Maize Newsletter Index
Return to the Maize Genome Database Page