Cell walls: the boundaries of plant development


The role of the extracellular matrix in the control of plant development: the 13th New Phytologist Symposium, London, UK, January 2005

The extracellular matrix, or cell wall, plays a diversity of important roles in higher plants. Besides providing skeletal support to the plant as a whole and providing a mechanical counterbalance to turgor pressure in plant cells, the wall is also key to many aspects of growth regulation and cell-to-cell interactions and signalling. From this point of view, it is not surprising that the extracellular matrix is also of consequence to many areas of plant development, and this formed the topic of the 13th New Phytologist Symposium. The Symposium brought together biochemists, physiologists, cell biologists and geneticists working on different experimental systems to focus on this topical issue – highlights from which are discussed in this report. The benefits of bringing different disciplines together was no more evident than at this meeting, with many stimulating discussions and debate taking place, from which we can look forward to exciting developments.

‘The cell wall is in essence a composite material built on a framework of cellulose microfibrils which have a greater tensile strength than steel’

Growth and expansion

Growth and expansion can be considered as distinct concepts in cell biological terms. Cell growth can be defined as the process by which the machinery of the cell is replicated, and this is intimately tied to the cell cycle or the process of endoreduplication. In the normal cell cycle, chromosomes are replicated before cell division, whilst endoreduplication involves chromosomal replication in the absence of division, resulting in polyploid cells; this process is common in plants. Both events lead to a doubling of nuclear material providing the potential increased production of ribosomal and other cell components, the difference being that in the normal cell cycle this material is partitioned into a new cell. In Arabidopsis, ploidy levels vary from 2C in meristematic and stomatal guard cells, for example, up to 24C in trichome cells. Plant cells can expand massively during their development, and some increase their volumes by 100 000 times or more. The final size that can be attained by plant cells appears closely tied to their ploidy number, with higher levels of ploidy leading to bigger cells (Sugimoto-Shirasu & Roberts, 2003). Whilst growth involves increases in the active components of the cell, that is to say the cytoplasm and its components, expansion can be thought of as an increase in cell volume, and in plants this can be dominated by vacuolar expansion.

Keith Roberts (John Innes Centre, Norwich, UK) described eight Arabidopsis mutants united in having smaller sizes than wild type in specific cell types and all of them carrying lesions in one of 4 proteins that make up an enzyme complex known as DNA topoisomerase 6 (Sugimoto-Shirasu et al., 2002). Crucially in all these mutants the lesions lead to the inability to reach a ploidy greater than 8C resulting in some cases in smaller trichomes, in some hairless roots, and in others shorter hypocotyls. Topoisomerase 6 is involved in disentangling replicated chromosomes by cutting and rejoining DNA at entangled points and this activity appears essential for progressing through the endocycle from 8C to 16C. Analysis of these mutants suggests a fairly linear relationship between ploidy levels, nuclear volume and cytoplasmic volume. Ultimately, this relationship extends to overall cell size and hence vacuolar volume although the relationship of ploidy level to cell volume does not appear strictly linear (Sugimoto-Shirasu & Roberts, 2003).

Hierarchies of control

There are many control points in biological systems, and in line with the words of a Bob Dylan song ‘They may call you Doctor or they may call you Chief, but you’re gonna have to serve somebody’ there are hierarchies of control involved in plant growth. Thus, although transcriptional programmes and hormonal pathways may instigate and mitigate growth, and although part of this control may be through the influence of cell ploidy and the cell cycle, in the end, if a plant cell is going to get bigger it has to increase its volume and this can only be achieved through cell expansion. Cell expansion in plants is dependent on turgor pressure, which provides the driving force for this process, and water uptake leading to cell expansion is dependent on the difference in water potential across the plasmamembrane that is generated by the presence of abundant solutes on the cellular side. The difference in osmotic potential across the cytoplasmic membrane is balanced by the generation of turgor pressure, which is contained by the strong cell wall that supports the cytoplasmic membrane. In a nongrowing situation, the wall is functionally rigid, but for a cell to grow there needs to be a relaxation event in the cell wall, which releases the counterbalance to turgor pressure, leading to a lowering of the cellular water potential. This lowering of cellular water potential then leads to a net movement of water into the cell, causing cell expansion. In theory, cell expansion can be regulated either through turgor pressure or changes in wall mechanical properties. In practice, in all but conditions of severe drought, growth is controlled through the properties of the wall (Cosgrove, 1993).

The plant cell wall is an outstanding material that combines impressive strength with remarkable extensibility. Plant cells typically generate around 0.5 MPa of turgor pressure, but in specialised cells such as stomatal guard cells these pressures can reach 5 MPa (50 times atmospheric pressure) and this is sustained by the counteraction of the relatively thin cell wall. During cell growth, not only must the wall be strong enough to bear the load imposed by turgor, but also be capable of undergoing controlled extension at rates that can exceed 10% h−1. One of the major goals of cell wall research is to understand how these characteristics are achieved.

The cell wall is in essence a composite material built on a framework of cellulose microfibrils. These microfibrils, which have a greater tensile strength than steel, are embedded in a matrix of other polymers. These are predominantly polysaccharides and fall into two general classes: hemicelluloses and pectins. Hemicelluloses can associate with cellulose microfibrils through extensive surface hydrogen bonding and are also sometimes called crosslinking glycans, as it is believed that they may form tethers between individual microfibrils to form a cohesive whole. Proteins (such as expansins) that induce wall extension generally do so by destabilising this cellulose–hemicellulose network (Cosgrove, 2000; Li et al., 2003).

Coextensive with this is another network formed by pectins, acidic polysaccharides dominated by the presence of galacturonic acid residues. The presence of carboxyl groups is a key feature of pectins, which associate with one another through electrostatic interactions. Many of the potential carboxyls in pectic polymers are masked by the presence of methyl esters in the nascent chains, and subsequent de-esterification modulates the properties of pectin in the wall. Keith Roberts described how examination of a wide range of Arabidopsis mutants exhibiting impaired hypocotyl elongation, using infrared spectroscopy, revealed that the pectins in all showed higher levels of unesterified carboxyls. The suggestion is that increased carboxyl groups lead to tighter pectin interactions in the wall and that this may be instrumental in limiting wall extension during growth. In support of this hypothesis, it was reported at the meeting that overexpression of an Arabidopsis pectin methylesterase gene on an inducible promoter led to dwarfing.

Hypocotyl expansion

Hypocotyl expansion has proved one of the most powerful model systems for examining the role of cell walls in growth, and considerable knowledge has been gained from the characterisation of mutants with impaired hypocotyls growth. Whilst a great deal has been achieved through the use of molecular genetics, Herman Höfte (INRA Versaille, France) showed the continuing importance of gaining a clear understanding of normal developmental events through careful observation. As part of the process involved in screening for plants with impaired hypocotyls growth, the Höfte group has meticulously examined the patterns of growth in wild-type Arabidopsis hypocotyls. Previous work of theirs has shown that hypocotyl elongation almost exclusively arises from the expansion of pre-existing cells laid down during Arabidopsis embryo development (Gendreau et al., 1996), as has been seen in other plants. The patterns of cell expansion in hypocotyls has also been characterised during growth both in the light and in the dark. In the dark, hypocotyl cells go through three phases of cell expansion. Initially, cells go through a phase of slow steady expansion; they then progress into a period of rapid elongation growth followed by a period in which growth slows and comes to a halt. The switch to rapid growth progresses almost as a wave of expansion, starting with cells at the base of the hypocotyl and spreading upwards towards the more apical cells. Recently, the Höfte group have described how cell walls in the slowly growing population are thickened early during germination and how, once the cells enter the phase of rapid elongation, the walls become progressively thinner, indicating that there are distinct differences in growth in these two phases (Refregier et al., 2004). Strikingly, the role of cellulose synthesis and deposition also appears quite distinct in these phases.

It is generally held, and logical, that cell wall extension cannot proceed without the synthesis of new cell wall material. This is obvious for long-term growth, where if the two are not coupled, the wall might become too thin to function properly. In line with this, interference in cellulose biosynthesis generally leads to the inhibition of cell expansion, as has been seen with cellulose synthase mutants and with the use of chemical inhibitors. Strikingly, it has been found that whilst treatment with the cellulose synthesis inhibitor isoxaben inhibits the slow phase of etiolated hypocotyl cell expansion, this treatment has no apparent impact on growth (although it did inhibit cellulose production) during the subsequent period of rapid cell expansion (Refregier et al., 2004), indicating that these two phases of growth were quite different in nature and that wall extension in the period of rapid elongation appears to be independent of cell wall synthesis. This suggests that wall integrity during this phase of growth may simply be maintained by rearrangements of existing polymers. In this context, it is fascinating to note that the orientation of microfibrils in these walls appears to undergo major changes during expansion, moving from a radial to an axial orientation during expansion. Because cellulose microfibrils are very long structures, such reorientation suggests that the wall must be in a partially fluid state during this phase of growth to allow their movement.

Division or expansion: which leads in differential growth?

The role of differential growth in the development of form in plants is an obvious one. For example, in the shoot apical meristem, leaf primordia first arise as small bulges on the meristematic flanks as the result of more rapid expansion of tissues in these areas. Whilst there has been extensive characterisation of the various transcriptional regulators that determine meristematic patterning, the actual mechanisms by which differential growth arise have seen less work. Andrew Fleming (University of Sheffield, UK) has been examining whether differential growth in development is led by cell division or by cell expansion. Clearly both components are involved in the process, but which is the leading role? To this end, tools have been deployed to manipulate cell division (localised overexpression of cell cycle regulatory genes) or cell expansion (localised ectopic expansin expression). The fields of growth on the flanks of a meristem are tiny (tens of microns), and to allow localised transgene induction with this level of spatial resolution, a tetracycline-inducible system has been deployed in combination with the application of tetracycline-loaded Sephadex beads. It was shown that localised induction of cyclin expression could induce localised increases of cell proliferation on the flanks of tobacco vegetative meristems, but that this had no developmental consequence in terms of the appearance of new primordia (Wyrykowska & Fleming, 2003). In contrast, locally induced expansin expression led to the appearance of ectopic leaf primordia, and eventually fully formed leaves, clearly demonstrating that locally induced tissue expansion appeared to be sufficient to initiate the entire process of leaf formation on the meristematic flanks (Pien et al., 2001). Similarly, locally induced expansin expression on the margins of primordial led to preferential enlargement of that margin in the mature leaf, whilst a similarly induced increase in local cell division eventually led to a reduction in the final size of the induced margin.

In the course of carefully studying leaf morphology, Andrew Fleming concluded that cells at the outer margin of the leaf lamina may play a key role in determining leaf growth and shape. Such cells are immensely long in the mature leaf and are characterised by the presence of particularly thick cell walls. It was shown that lesion in an Arabidopsis margin cell-specific gene (identified in an enhancer trap screen) led to a loss in cell identity for the margin cells, and that in their absence leaf development was severely impaired. Mutants in the yet-to-be-identified gene have small dense and dark-coloured leaves, and the locus has been named HEPATICA due to the liverwort-like appearance of the mutant plants.

Signalling: wax, shine and fiddleheads

Development in plants is coordinated through cell-to-cell communication and such signalling may be either symplastically or apoplastically transmitted. Evidence that cell wall components might serve as important developmental signals began to emerge in the 1990s. Paul McCabe (University College Dublin, Ireland) described work he carried out in Roger Pennell's laboratory (University College London) in which it was found that cells bearing a specific cell wall epitope produced a soluble signal necessary for suspension cells to form embryos. In this work, they showed that cells recognised by JIM8, a monoclonal antibody that binds to an arabinogalactan protein, produced a soluble signal that was sufficient to induce cells that did not carry this epitope to form embryos (McCabe et al., 1997). Current attempts to purify the soluble signal were described.

Many developmental signals are propagated through the apoplast and therefore must pass through cell walls. The role of the cell wall in signalling processes was highlighted in talks on stomatal development and on mutants that exhibit postgenital organ fusion. Julie Gray (University of Sheffield, UK) presented a talk on the Arabidopsis HIC gene, which when mutated leads to increased guard cell proliferation at high CO2 levels (Gray et al., 2000). The gene itself encodes a putative beta ketoacyl CoA synthase thought to be involved in producing very long chain fatty acids that play a role in cuticular waxes. Mutants in two other genes involved in wax biosynthesis, CER1 (a decarbonylase) and CER6 (a fatty acid elongase), exhibit greater numbers of stomatal pores (Aarts et al., 1995; Fiebig et al., 2000). These observations indicate a role for cuticular waxes in the control of stomatal proliferation during leaf development. A model was suggested whereby the cuticle controls the diffusion of a diffusible signal from guard cells which inhibits the development of stomatal pairs in close proximity.

There are a number of well-characterised systems where normal signalling across the cell wall and cuticle is involved, such as pollen–pistil interactions and carpel fusion, and there are a number of Arabidopsis mutants where ectopic organ fusion occurs, presumably because of altered signalling functions. One of the best characterised of these mutants is in a gene called FIDDLEHEAD, which Robert Pruitt's group, at Purdue University (West Lafayette, IN, USA), have shown to encodes a lipid biosynthetic enzyme (Pruitt et al., 2000) thought to be involved in epidermal wax biosynthesis. Organs at the shoot apex generally become fused in these plants, leading to the characteristic fiddlehead appearance of the plants. In addition, a number of other mutants showing ectopic organ fusion have been identified that carry lesions in genes thought to be involved in cuticular wax biosynthesis (Tanaka et al., 2001; Wellesen et al., 2001). It was suggested that these mutants all indicate that an intact cuticle is essential to maintain an inert, nonresponsive interface at the epidermal surface.

A more familiar role of the cuticle is in providing a waterproof and protective covering to the plant, roles fulfilled by cutins and waxes. Andy Pereira's group (Plant Research International, Wageningen, the Netherlands) have been studying the SHINE transcription factors of Arabidopsis, which were identified in an activation tagging screen by the shiny appearance of leaves of plants carrying a gain of function mutation in one of these genes. The shiny appearance of the leaves resulted from the increased deposition of waxes on the surface of the cuticle, with wax quantities having increased up to six-fold compared to wild type. Cloning of the gene identified it as one of a clade of three closely related AP2/EREBP transcription factors. Subsequent overexpression of the other two genes resulted in similar waxy phenotypes (Aharoni et al., 2004). Another group also independently showed that overexpression of these transcription factors led to a waxy phenotype (Broun et al., 2004).

At first glance, this appears as a relatively straightforward story. The regulation of cuticular wax biosynthesis is poorly understood, and this looks as if the transcriptional regulators have been overexpressed leading to increased wax formation. If so, this represents an excellent opportunity to study key components of this process. The story is, however, somewhat more complex. Logic would predict that because waxes form a hydrophobic barrier at the surface of leaves, this structure would be less permeable in the shine mutant than in wild type. In fact, it was shown that excised mutant leaves lose water more rapidly than wild type and also that chlorophyll was more readily extracted from the mutant than from wild type leaves, indicating enhanced permeability. Some clues to this apparent paradox came from comparative examination of wax components in mutant and wild type, which revealed alterations in composition as well as quantity of waxes in the mutants. This indicates that the overexpression of the SHINE genes does not increase all aspects of wax production proportionally and it may be that this change in wax composition is in part responsible for the increased permeability of the epidermis in these plants. Promoter/GUS reporter experiments showed that the SHN1 gene is normally predominantly expressed in areas of cell separation (abscission and dehiscence zones for example) and it was suggested that the transcription factor was involved in sealing the surfaces of such zones to make them less susceptible to pathogen ingress and water loss. Microarray analysis of the mutant plants were reported to reveal the induction of a number of transcripts of genes normally associated with abscission, suggesting that the SHN transcription factors may have a more general role in these processes beyond sealing the surfaces with cuticular waxes.

It was shown that the mutant plants showed a greater resistance to severe droughting than do wild type plants. In these experiments, water was withheld until the plants became wilted, at which time they were rewatered and examined for how well they recovered. The greater resistance of the SHN1 overexpressor mutants seems paradoxical, given that they showed more rapid water loss from excised leaves. However, the plants exhibited a number of phenotypic abnormalities, including reduced stomatal abundance, and it was suggested that this accounted for the greater apparent drought resistance.

A central theme running throughout this section of the meeting was the accumulating evidence that the cuticle has important roles in aspects of the control of cell fate. Both HIC and FDH encode members of the same gene family that encode putative beta ketoacyl CoA synthases that form part of the complex believed to be responsible for the synthesis of the very long chain fatty acid components of the cuticle. Along similar lines, SHN is a transcriptional activator involved in the control of wax biosynthesis, and shn overexpressors with abnormal cuticle composition exhibit developmental phenotypes such as reduced stomatal density.

Cellulosic cell walls are not land plant-exclusive

The possession of a cellulosic cell wall is not exclusive to land plants, and is a common feature among many algae, even the distantly related brown algae. The zygotes of the brown alga Fucus have provided a productive model for studying the development of polarity in plant cells (Brownlee et al., 2001). Polarity is established in the free-living single-celled zygotes of this species following fertilisation, with one pole giving rise to the thallus and the other to a rhizoid that eventually forms the holdfast. There then follows an asymmetric division, producing a smaller basal cell that gives rise to all rhizoid tissues and a larger apical cell that will give rise to the thallus. Before cell division occurs, the cell produces an outgrowth at the basal pole. Colin Brownlees’ group (Marine Biological Association, Plymouth, UK) have contributed greatly to this field of research, especially through the use of advanced optical methods leading to a sophisticated view of molecular mechanisms underlying the development of polarity and the generation of polarised growth in these zygotes. It was described how polarity could be determined by a range of different stimuli. Polarisation is initiated by perception of these stimuli and this is then translated to intracellular asymmetries that appear to be set up by changes in local levels of reactive oxygen species and calcium (Coelho et al., 2002). Outgrowth of the cell at the rhizoid pole appears to be established by the anchoring of actin microfibrils to the cytoplasmic membrane, as well as anchorage of the membrane to the wall. This presumably then establishes sites of preferential deposition of new wall material, and indeed the site of rhizoid outgrowth appears to be characterised by the presence of sulphated fucan polysaccharides (Shaw & Quatrano, 1996).

The cell wall is central to the development of macroscopic form in plants and provides the medium through which cell-to-cell and organ-to-organ communication is mediated. The importance of this structure in a whole range of processes in plants is gradually emerging, and this role is nowhere greater than in the control of growth and development, as was made clear from the range of talks presented at this lively meeting.