To gain information on the genes and proteins involved in cell wall elongation, it is important to dissect the different phases of the process. Arabidopsis hypocotyls are a suitable material, since almost no cell division occurs and only the cells present in the embryo undergo elongation [4-6]. Another advantage is that in etiolated hypocotyls, synthesis, addition, and reorganization of cell wall material occur in time-separate phases [4,7]. During the first 3 days after germination synthesis and deposition of cell wall material which result in cell wall thickening, are the main features. Through the following days, the hypocotyl will grow, mainly through extensive cell wall polymer disassembly and rearrangement.

Understanding the molecular mechanisms responsible for rearrangements of cell wall polymers during hypocotyl growth requires the identification of the proteins present in muro. On the one hand, previous studies identified gene families involved in rearrangements of cell wall components during cell elongation [8-11]. On the other hand plant cell wall proteomics emerged a few years ago, providing information on cell wall proteins (CWPs) present in different types of Arabidopsis cells including cell suspension cultures, roots, rosette leaves, and stems [12,13]. In most cases, limitations were encountered, e.g. the presence of intracellular contaminants [14] and a poor quality of separation of CWPs by two-dimensional electrophoresis (2D-E) [15]. The main families of CWPs were identified [16] on the basis on bioinformatics prediction of functional domains.

Important progress in plant cell wall proteomics was achieved by setting up a new separation method for CWPs. Separation of plant CWPs for proteomic purposes was difficult using 2D-E [12]. The window of protein separation is optimal for pIs between 3 and 10 and for molecular masses between 120 and 10 kDa. Since most CWPs are basic glycoproteins, they tend to migrate as a smear on the basic side of 2D-gels [15]. Alternative methods were proposed; they consisted in separation of CWPs into an acidic and a basic fraction by cation exchange chromatography followed by 2D-E and 1D-E respectively [20,21]. The new method includes a first step of separation by cation exchange chromatography at acidic pH, and a second step of separation by 1D-E. It gives more information on the physico-chemical properties of the proteins, allows comparative semi-quantification among different samples, as well as a better identification through MALDI-TOF MS. In the case of etiolated hypocotyls of Arabidopsis, it allowed the doubling of the number of proteins identified as compared to separation by 1D-E alone. In addition, since many CWPs can now be visualized, this work provided preparative tools for developing biochemical studies on CWPs, either for further purification or structural characterization.

Altogether, 137 CWPs were identified in this study among which 51 had not been previously identified through cell wall proteomics. This work also presents an overview of the dynamics of CWPs during cell elongation. Many differences were observed between elongating and fully-grown hypocotyls. When only the presence/absence of a CWP was considered, these changes concerned 53 out of the 137 identified CWPs (38%). When the proposed semi-quantification method was taken into account, this percentage increased to 63% (34 additional proteins). Changes in the same gene family can reflect the regulation of gene expression at different stages of development and/or differences in biological activity, as discussed below for XTHs, PGs, expansins, PMEs, and peroxidases. Proteins acting on carbohydrates are more numerous and more abundant in elongating hypocotyls than in fully-grown hypocotyls. This was to be expected since rearrangements of cell wall polysaccharides are very important during cell elongation [7]. The fact that proteases are more numerous and in higher amounts at 5 than at 11 days is more surprising. Nothing is known about their targets in cell walls. Are they contributing to release peptides involved in signaling [36]? Are they involved in protein maturation [13] or in protein degradation? Conversely, two protease inhibitors are much more abundant at 11 than at 5 days. Altogether, it seems that proteolytic activities are more important when elongation is active than during elongation arrest. Among oxido-reductases, five berberine-bridge enzymes were identified among which three were present only at 11 days. The role of such proteins in cell walls is still unknown. For proteins with interacting domains, some lectins and PME inhibitors are more abundant at 5 days. Among miscellaneous proteins, the amount of CWPs containing phosphatase domains was found to be higher at 5 than at 11 days. Such proteins were shown to be associated to the regeneration of protoplast cell walls [37] and pollen tube growth [38], but their precise roles are still unknown. A protein homologous to COBRA (AtCOBL10) was only found at 5 days. Although the function of AtCOBL10 is not known, it should be noted that COBRA was shown to participate in the orientation of cellulose microfibrils, and dark-grown hypocotyls of the cob-4 mutant have a 95% reduction in length compared to the wild-type [39]. AtCOBL10 may play such a role during the elongation of hypocotyl cells.

Many proteins expected to participate in cell wall extension, such as XTHs, expansins, PGs, PMEs and peroxidases [3,8,9,40] were found. But such proteins, i.e. same proteins or proteins of the same family were also found after completion of elongation. Several hypotheses can be proposed. Although many proteases were identified at both stages of development suggesting a regulation of CWPs by proteolytic degradation, these proteins can have a long half-life. However, it is probable that some of these proteins participate in the differentiation of tracheary elements, such as AtXTH32 which was only identified at 11 days. This XTH which belongs to the phylogenic group 3, like AtXTH31 and AtXTH33, has been assumed to have xyloglucan endo-hydrolysis activity [41]. AtXTH31-33 might be involved in the rearrangement of cell walls of differentiating vessels elements. Such elements can be observed using microscopy (not shown). In the same way, some expansins were found in differentiating tracheary elements [42]. Finally, at least PMEs and peroxidases were assumed to play a role both during the elongation process and elongation arrest. The enzymatic activity of PMEs may be modulated, depending on the pH of the extracellular matrix and on the structure of pectic homogalacturonans. They could have either a local activity favoring the enzymatic activity of endo-PGs thus producing fragments of pectin, or a processing activity leading to the de-esterification of stretches of GalA and to the formation of the so-called egg-boxes that tend to rigidify the pectic network [8]. Moreover, the degree of pectin methyl-esterification was shown to be positively correlated to hypocotyl growth [43]. The activity of peroxidases is also versatile [9]. During the hydroxylic cycle, peroxidases can produce reactive oxygen species that can break cell wall polysaccharides in a non-enzymatic way thus favoring cell wall extension [9,44]. On the contrary, during the peroxidative cycle, peroxidases can promote cross-linking of cell wall components such as structural proteins or lignins. In addition, members of most of these protein families were identified in apoplastic fluids of rosette leaves [21]. Since leaf cells are surrounded by mature walls, this can mean that those CWPs may play house-keeping roles.

Proteomics provides information about possible regulatory mechanisms of CWPs. As previously discussed [10], the presence of a protein does not mean that it has full biological activity. Proteins with putative enzymatic activities are numerous, but inhibitors of these activities are also present. This is the case for proteases (14) and protease inhibitors (7), PMEs (6) and PME inhibitors (3 PMEIs). Some PMEs have a pro-domain consisting of a PMEI. However, such domains are assumed to be cleaved during or just after protein export, since they were never found in purified PMEs [8]. In the same way, no peptide matching the PMEI domains were found during identification of PMEs by peptide mass mapping (data not shown). Other enzyme inhibitors are assumed to be involved in defense reaction, such as PG inhibiting proteins (3 PGIPs) and inhibitors of xyloglycan endoglucanases (3 XEGIPs). Indeed, some of them were shown to be specifically active against fungal enzymes [45]. Other regulatory mechanisms include variations in pH of the extracellular matrix that occur during growth arrest [45], physical contact between enzymes and their substrates [10] and proteolytic degradation.