Dispatch
R213
Cellulose: How many cellulose synthases to make a plant?
Robyn M. Perrin
Many questions remain about the biosynthesis of
cellulose, the major plant cell wall component, not least
of which is why plants have so many genes for the
cellulose synthase catalytic subunit. Perhaps multiple
isoforms of cellulose synthase are needed in the same
cell for the formation of functional dimeric complexes.
Address: MSU-DOE Plant Research Laboratory, Room 110 Plant
Biology, Michigan State University, East Lansing, Michigan 48824, USA.
E-mail: [email protected]
Current Biology 2001, 11:R213–R216
0960-9822/01/$ – see front matter
© 2001 Elsevier Science Ltd. All rights reserved.
Cellulose, a homogenous polymer of β-(1,4) glucose, is the
most abundant biopolymer on Earth, with an estimated
180 billion tons produced annually in nature [1,2]. A great
deal of the world’s economy depends directly upon cellulose, as it is integral to the production of fibers for textiles
and paper products, dietary fiber and lumber for building
material. Cellulose is a remarkable structure that possesses a greater tensile strength than steel. Elucidating
how this material is formed is crucial for understanding
how plant morphology is controlled. The rigidity of cellulose is of direct consequence for plant growth, as the orientation of cellulose polymer deposition determines cell
shape, and thus tissue morphology, to a great extent.
There are, however, many outstanding biological questions in the area of cellulose biosynthesis.
The birth of cellulose occurs at the plasma membrane,
where a structure called the rosette terminal complex
— often called simply the rosette — synthesizes β-1,4
glucan chains using UDP–glucose as substrate (reviewed in
[2]). The rosette has a diameter of approximately 24 nm
and displays sixfold symmetry as observed by electron
microscopy [3,4]. In most higher plants, each rosette forms
a total of 36 cellulose chains. The chains coalesce into a
microfibril in a metastable crystalline form termed cellulose
I. The assumption of this crystalline form seems to be
dependent on a precise organization and orientation of the
rosette components. For example, a temperature-sensitive
Arabidopsis mutant called radial swelling 1 (rsw1) that has a
cell-swelling phenotype at non-permissive temperatures, as
a result of a defect in a cellulose synthase catalytic subunit,
also accumulates non-crystalline β-1,4 glucan [5]. Inspection of rsw1 rosettes shows that they are disorganized when
the plant is grown at elevated temperatures [5].
The lability of rosettes, and thus of cellulose synthase activity, has caused difficulty for biochemical approaches to
studying this enzyme. Cellulose synthase was first identified at the molecular level in the cellulose-producing bacteria Acetobacter xylinum and Agrobacterium tumefaciens
(reviewed in [2]). From the sequences of these enzymes,
motifs characteristic of cellulose synthases were recognized.
Candidate plant cellulose synthases were found by randomly sequencing cDNA libraries from developing cotton
fibers, collected at a developmental stage during which
a massive amount of cellulose is synthesized. Two candidate
genes were identified encoding proteins similar to the
bacterial enzymes, with sequence features characteristic of
cellulose synthases [6]. Eight transmembrane domains were
predicted for the cotton enzymes, two at the amino terminus
and six near the carboxyl terminus. The central soluble
region was capable of binding UDP–glucose in a Mg2+dependent manner, consistent with previous biochemical
studies [6]. Originally named CelA-1 and CelA-2, these genes
have recently been designated GhCesA-1 and GhCesA-2 [2].
In vivo evidence that the CesA genes do indeed code for
catalytic subunits of cellulose synthases rests largely upon
genetic data. The Arabidopsis rsw1 mutant described above
was shown to have a point mutation in a gene homologous
to the cotton CesA genes, causing cell wall abnormalities
throughout the plant when grown at elevated temperatures [5]. In the second example, a Arabidopsis mutant
identified on the basis of collapsed vascular (xylem) cell
phenotype was found also to have a mutation in a CesA
homolog [7]. The irx3 mutant lacks sufficient cellulose
deposition in vascular tissues to withstand the negative
pressure involved in water conductance, and thus its vasculature collapses. Furthermore, an elegant microscopy
study, which applied a freeze-fracture replica labelling
technique to plants for the first time, demonstrated that
antibodies against a CesA protein from cotton could label
rosettes of adzuki bean, a vascular plant [4].
With the sequencing of the Arabidopsis genome now
complete, a common observation within the plant biology
community is that there is much genomic redundancy.
Many genes appear to be present as members of families,
rather than as unique copies. This has also been the case
for cellulose synthase genes, but to an extreme. Arabidopsis appears to have a large superfamily of approximately 40
genes that bear similarity to the original CesA genes. This
superfamily has been divided into one family of ten ‘true’
CesA genes, and six families of cellulose synthase-like (Csl)
genes that are less closely related [8]. A web site maintained
by Richmond and Somerville (http:// cellwall.stanford.edu)
documents sequence data for cellulose synthase and
cellulose synthase-like genes.
R214
Current Biology Vol 11 No 6
Figure 1
(a)
Isoform A
Isoform B
(b)
CesA
isoform pool
Random distribution
into rosettes
(c)
Current Biology
Models for the distribution of CesA isoforms within a plant and
consequences for the structure of the cellulose synthase rosette.
(a) Different isoforms are expressed in different tissues of the plant
(blue, isoform A; yellow, isoform B). Rosettes contain only one type of
CesA isoform. (b) Different isoforms are expressed in the same cells
(green, mixture of isoforms A and B). Both isoforms are functionally
redundant and contribute to a general pool of catalytic subunits.
Cellulose synthase isoforms are randomly distributed into rosettes.
(c) Different isoforms are again expressed in the same cells, but the
isoforms are not redundant. One isoform of each type is required to
synthesize each glucan chain. Note that most cellulose microfibrils
produced by higher plants consist of 36 glucan chains, indicating that
a total of six cellulose synthase active sites would be present in each
of the six rosette subcomplexes. Only one pair per subcomplex is
shown here for simplicity.
Why do plants have many genes for cellulose synthase?
Two recent studies independently suggest a possible
explanation: that a functional catalytic enzyme may require
the presence of two CesA isoforms [9,10]. In both cases,
previously described Arabidopsis mutants were found to
have a defect in a CesA isoform, and the consequences
allow some conclusions to be drawn about the structure of
the cellulose synthase complex. The first study [9] examined a mutant named procuste1 (prc1) that was originally
described as having cell elongation defects in root and
dark-grown shoot (hypocotyl) tissues [11]. The cell elongation defects could be phenocopied using chemicals that
inhibit cellulose synthesis. Analysis of cell wall morphology in this mutant showed the presence of incomplete cell
wall ‘stubs’ in the affected tissues [9]. Such incomplete
cell walls have also been observed in cell division mutants,
but cell walls in embryonic tissues of prc1 mutant plants
were normal. This indicates that the mutant suffers a loss
of cell wall integrity during cell elongation, but is not
affected in cell division itself.
Fagard et al. [9] analyzed the cell walls in affected tissues of
prc1 mutant plants using a technique called Fourier transform infrared (FTIR) microspectroscopy, an approach that
has been applied to rapid analysis of cell wall composition
in as few as one cell [12]. The functional groups of cell wall
polymers determine the spectral features in this analysis, so
comparing spectra to one another allows cell wall differences between mutant and wild-type plants to be determined. Using this type of approach, it was possible to
distinguish prc1 cell walls from wild-type cell walls. The differences were due to peaks corresponding to purified cellulose and proteins, suggesting that prc1 cell walls in affected
tissues have a decrease in crystalline cellulose and a complementary increase in cell wall structural proteins. The
decrease in cellulose content was also confirmed by direct
chemical composition analysis and radiolabelling studies.
Map-based cloning showed that the affected gene in the
prc1 mutant is a CesA previously designated as AtCesA6 [9].
Interestingly, the prc1 mutant phenotype is very similar to
that caused by leaky alleles of rsw1, and a cross between
prc1-8 and rsw1-10 resulted in a double mutant that had a
shorter dark-grown hypocotyl length than either single
mutant. The expression patterns for the Rsw1/AtCesA1
and Prc1/AtCesA6 genes are very similar, although the
mutant phenotype for rsw1 is found in all cells of the
plant, whereas the phenotype for prc1 is restricted to
certain cell types in roots and dark-grown hypocotyls. This
led Fagard et al. [9] to suggest that the two CesA isoforms
might be working in concert, meaning that both forms
must be present in the same cells, or even in the same
complex, to function properly.
In the second study, Taylor et al. [10] analysed the irregular xylem locus irx1, which is distinct from the irx3 locus
mentioned above. As with irx3, the irx1 mutant also has a
collapsed xylem phenotype, apparently affecting the same
cell types in the same manner, as determined by ultrastructural studies [10]. The irx1 mutant also has a decrease
Dispatch
in cellulose content. Map-based cloning showed that the
affected gene is another CesA isoform, AtCesA8. Irx3/CesA7
and Irx1/CesA8 were found to have identical expression
patterns [10]. Because both CesA isoforms seem to be
acting on the same cells at the same time, Taylor et al. [10]
suggest that they also might be working in concert. Importantly, they tested for interaction between the two isoforms, by raising polyclonal antibodies specific to the two
proteins, Irx3/CesA7 and Irx1/CesA8, and preparing transgenic plants expressing histidine-tagged Irx1/CesA8 on an
irx1 mutant background. Solubilized membrane protein
from the transgenic plants was then applied to a nickel
affinity column; epitope-tagged Irx1/CesA8 did indeed
bind to the column and could be eluted with imidazole. In
addition, Irx3/CesA7 also could be detected in the eluted
fraction, indicating interaction between the two isoforms.
As a control, an unrelated protein did not co-elute with the
tagged Irx1/CesA8.
These studies [9,10] have a number of implications for the
composition of the cellulose synthase catalytic complex,
illustrated in Figure 1. If one considers all ten CesA
isoforms, one possibility is that different isomers are
expressed in different tissues or in response to various
conditions. Differential expression of some CesA genes has
indeed been observed (for example, see [13]). In this case,
a particular tissue might express one specific CesA isoform,
but the particular isoform expressed might vary between
tissues (Figure 1a). In cases where multiple CesA isoforms
are expressed within a given tissue, other possibilities
exist (Figure 1b). It is possible that the isoforms are functionally redundant and contribute to a general CesA subunit
pool; the subunits would then be randomly distributed
into rosettes. But if this were the case, one might expect
either a gene dosage effect or a dominant-negative effect
at the genetic level. Genetic analysis has shown that this is
not the case, indicating that — at least for the four CesA
genes considered in these studies — the isomers are not
functionally redundant and that the model illustrated in
Figure 1b is unlikely [9].
Another possible model is illustrated in Figure 1c: here,
both isoforms are required for function, and the proteins
may form a heterodimer within the rosette subunits. The
new data reported by Turner et al. [10] and Fagard et al.
[9] support this model. Furthermore, it is possible that the
models illustrated in Figure 1 parts (a) and (c) are not
mutually exclusive, in the sense that within the collection
of ten CesA isoforms, different subsets may work in
concert in different cell types. The possibility that the
functional cellulose synthase enzyme is a dimer has been
suggested previously, as it would explain how the glucose
units in cellulose, which are flipped nearly 180° in relation
to each other, might be attached without invoking the
need to rotate either the growing polymer or the catalytic
enzyme [14]. It remains theoretically possible that isoforms
R215
may work together as trimers, rather than dimers, with two
groups of three isoforms for each of the six subcomplexes
of the rosette. It has been noted, however, that there
might not be a need for a multimeric enzyme if each cellulose synthase molecule has at least two active sites, as has
been observed for hyaluronan synthase [15]. Even if not
strictly necessary for synthesis of a single glucan chain,
though, it is possible that multimerization might facilitate
organization of the rosette such that crystalline microfibrils
can form [15].
In conclusion, both genetic and biochemical evidence
supports the view that different CesA isoforms interact to
form a functional cellulose synthase enzyme. These studies
begin to show a clearer picture of the organization of the
cellulose synthase rosette machinery. But co-localization
studies are needed to confirm these results and the stoichiometry of the complex should be determined. And it is
almost certain that this is not the entire picture. As many
as twelve other polypeptides have been detected within
the rosette, and the identities of the other members have
not been determined [16]. The cellulose synthase story is
thus by no means completed.
Acknowledgements
The author wishes to thank Ken Keegstra and Douglas Luckie for critical
reading of the manuscript.
References
1. Amor Y, Haigler CH, Johnson S, Wainscott M, Delmer DP:
A membrane-associated form of sucrose synthase and its
potential role in synthesis of cellulose and callose in plants.
Proc Natl Acad Sci USA 1995, 92:9353-9357.
2. Delmer D: Cellulose biosynthesis: exciting times for a difficult field
of study. Annu Rev Plant Physiol Plant Mol Biol 1999, 50:245-276.
3. Chapman RL, Staehelin LA: Plasma membrane ‘rosettes’ in carrot
and sycamore suspension culture cells. J Ultrastruct Res 1985,
93:87-91.
4. Kimura S, Laosinchai W, Itoh T, Cui X, Linder CR, Brown RM:
Immunogold labeling of rosette terminal cellulose-synthesizing
complexes in the vascular plant Vigna angularis. Plant Cell 1999,
11:2075-2086.
5. Arioli T, Peng L, Betzner AS, Burn J, Wittke W, Herth W, Camilleri C,
Hofte H, Plazinski J, Birch R, et al.: Molecular analysis of cellulose
biosynthesis in Arabidopsis. Science 1998, 279:717-720.
6. Pear JR, Kawagoe Y, Schreckengost WE, Delmer DP, Stalker DM:
Higher plants contain homologs of the bacterial celA genes
encoding the catalytic subunit of cellulose synthase. Proc Natl
Acad Sci USA 1996, 93:12637-12642.
7. Taylor NG, Scheible WR, Cutler S, Somerville CR, Turner SR: The
irregular xylem3 locus of Arabidopsis encodes a cellulose
synthase required for secondary cell wall synthesis. Plant Cell
1999, 11:769-779.
8. Richmond TA, Somerville CR: The cellulose synthase superfamily.
Plant Physiol 2000, 124:495-498.
9. Fagard M, Desnos T, Desprez T, Goubet F, Refregier G, Mouille G,
McCann MC, Rayon C, Vernhettes S, Hofte H: PROCUSTE1
encodes a cellulose synthase required for normal cell elongation
specifically in roots and dark-grown hypocotyls of Arabidopsis.
Plant Cell 2000, 12:2409-2423.
10. Taylor NG, Laurie S, Turner SR: Multiple cellulose synthase
catalytic subunits are required for cellulose synthesis in
Arabidopsis. Plant Cell 2000, 12:2529-2539.
11. Desnos T, Orbovic V, Bellini C, Kronenberger J, Caboche M, Traas J,
Hofte H: Procuste1 mutants identify two distinct genetic pathways
controlling hypocotyl cell elongation, respectively in dark- and lightgrown Arabidopsis seedlings. Development 1996, 122:683-693.
R216
Current Biology Vol 11 No 6
12. McCann MC, Chen L, Roberts K, Kemsley EK, Sene C, Carpita NC,
Stacey NJ, Wilson RH: Infrared microspectroscopy: sampling
heterogeneity in plant cell wall composition and architecture.
Physiol Plant 1997, 100:729-738.
13. Holland N, Holland D, Helentjaris T, Dhugga KS,
Xoconostle-Cazares B, Delmer D: A comparative analysis of the
plant cellulose synthase (CesA) gene family. Plant Physiol 2000,
123:1313-1323.
14. Carpita NC, Vergara C: A recipe for cellulose. Science 1998,
279:672-673.
15. Saxena IM, Brown RM: Cellulose synthases and related enzymes.
Curr Opin Plant Biol 2000, 3:523-531.
16. Kudlicka K, Brown RM, Jr: Cellulose and callose biosynthesis in
→3)- and
higher plants. 1. Solubilization and separation of (1→
→4)-β
β-glucan synthase activities from mung bean. Plant Physiol
(1→
1997, 115:643-656.