**4. Concluding remarks**

protrusion occurs at the transition zone, which comprises the cells between the last proximal root hair cell in the radicle and the lower basal cells of the hypocotyl [69]. In *Arabidopsis* Col seeds that have been previously stratified, the radicle protrusion can initiate as early as 32 HAI. By this time, and immediately prior to the radicle emergence through the endosperm, the cells in the transition zone had incremented their size by 44% while the cells in the radicle 10% and in the hypocotyl 30%. By 40HAI, the radicle has already protruded and the elongated cells in the seedling have increased their size by 15% in the radicle, 52% in the hypocotyl, and 108% in the transition zone. Elongation is often accompanied by an increase in DNA content without

168 New Challenges in Seed Biology - Basic and Translational Research Driving Seed Technology

In the micropylar endosperm of tomato seeds, an important mobilization of protein bodies occurs, but it seems that there is no cell degradation as the radicle protrudes. Instead, a process similar to cell separation to allow radical protrusion was suggested [70]. A similar process was observed in celery seeds, where the radicle tip also seems to penetrate the micropylar endosperm by separating the endosperm cells, but, since the embryo needs to grow before germination is completed, cell degradation for storage mobilization occurs in the endosperm adjacent to the embryo [27]. The expression of *LeEXP8* and *LePG1* in the embryo elongation

Major reserve mobilization occurs once germination has concluded, and these reserves are utilized to feed the growing seedling rather than to fuel radicle protrusion. However, in cereal grains, the preparation for starch and oligosaccharide mobilization occurs within the first hours of germination [15]. In cereals, the endosperm is a nonliving storage tissue, and the endosperm cell walls protect its contents from enzymatic attack. Accordingly, the degrada‐ tion of cell walls is a limiting step in storage reserve mobilization that is induced by the GA

In most endospermic seeds this tissue is still living. Mannans in the endosperm cell walls of date palm (*Phoenix dactylifera*) and coffee are mobilized to support embryo development. It has been proposed that the mobilization of storage xyloglucans can be coupled to the growth rate of the seedling by transglycosylation. In legumes, endosperm galactomannans seems to function as reserves; they can constitute up to 30% of total seed dry weight. In fenugreek and *Schizolobium parahyba*, the cell walls of the endosperm are thickened with galactomannan and in some cases the cytoplasm is nonexistent. In *Tamarindus indica* and *Hymenaea courbaril*, reserve xyloglucans are stored between two primary walls and are degraded without hydrolyzing both walls [39]. During germination of celery seeds, the surrounding endosperm degrades leaving a small amount of un‐degraded polymers of the cell wall, except for the micropylar endosperm, in which only some protein bodies are mobilized and the rest of the cells persist until the radicle pushes through; once radicle protrusion has started, these micropylar cells

zone of tomato seeds has been reported at the onset of radicle protrusion [51, 68].

produced by the embryo (at the scutellum) and secreted to the aleurone layer [2].

*3.4.5. Cell wall participation in the mobilization of stored reserves*

subsequent mitosis (endoreduplication) [69].

are degraded [27].

In‐depth temporal screening of cell wall‐related transcripts and proteins has provided an important overview ofthe possible actors involved in the five stages during germination where wall modification is involved, as described above. In *rice*, a valuable integrative effort using ‐ omics approaches has been done to understand seed germination [23]. This analysis suggest‐ ed that the changes in transcript levels during early germination (3–12 HAI) drive the subsequent changes in the metabolome (12–24 HAI) of germinating seeds, supporting that most of the changes observed at the transcriptional level are related to the cellular processes involved in germination. Other authors have associated transcript abundance with specific seed compartments and some enzymatic activity assays, demonstrating the relevance of understanding tissue‐specific expression profiles [17, 38]. Much of the information available related to CWMPs still needs to be validated through enzymatic activity or in vivo interac‐ tion assays. Only about 121 (∼12%) of the total cell wall‐related genes are experimentally validated [48]. Also, many cell wall‐related proteins belong to families of unknown function. The ‐omics approach can be useful to propose hypothesis of wall‐modification complexes, whose activity could be regulated at several levels, and coordinated by unknown function proteins that could act as scaffolding proteins and direct this complex activity to specific polysaccharides. Cosgrove [44] proposed that CWMPs could be functionally classified into primary or secondary modifiers, but this idea has not being reflected in other studies. Following Cosgrove, the analysis of cell wall modification considering an alternating activi‐ ty of primary or secondary modifiers could facilitate the understanding of the dynamics of cell wall modification during seed germination. For instance, expansins could be primary modifiers as they affect cell wall loosening and extensibility, but they do notremove ortransfer polysaccharides into the wall during imbibition; other primary modifiers could be PMEs, as their activity precedes PGs and promotes cell expansion or cell separation, or the resulting exposed GAL residues can be cross‐linked with Ca2+ and promote wall stiffening. Secondary modifiers would include GHs and GTs that would act on exposed residues either promoting cell expansion, separation, or stiffening. In assays to study mucilage properties, sequential treatment with different hydrolases allows solubilization of other components, which are masked to the activity of other enzymes [55, 71]. The alternate perspective of primary and secondary modifiers could help in identifying potential interactions in silico and tested in vivo.

Spatial transcriptomic analyses that include the different seed compartments and the analy‐ sis of cell wall composition changes using specific antibodies for in situ localization of the different polysaccharide epitopes in seed tissues provide valuable information. Although *Arabidopsis* is the best‐known plant model, several authors demonstrate that comparative analysis allows higher resolution of tissue‐specific cell wall microdomains that are not achievable in *Arabidopsis*[8, 10, 60]. As an example, Lee et al. [59] describe the presence of LM13 epitopes in the inner and outer cell walls, but absent in the transverse cell walls of the endosperm in *Lepidium* seeds; in tobacco, which has a thicker endosperm than *Arabidopsis* or *Lepidium*, XGs were abundant in the embryo, and at the micropyle (rich in heteromannans), these polysaccharides were only present in the middle lamella and intercellular regions. Thus, the analysis of cell wall‐modification processes would benefit from the multispecies compar‐

ison of in situ localization of polysaccharide epitopes in seed tissues. The characterization of wall microdomains could be combined with the valuable information generated by ‐omics technologies, to propose new hypothesis of regulation and coordinated activity of CWMPs. Ultimately, the activity of these CWMPs must be confirmed by in situ localization, in vivo protein interactions, and enzymatic activity. By combining the resources available for model species with the selection of other plant systems with bigger‐easy‐to‐handle seeds, it could be possible to achieve a comprehensive view of seed‐compartment functions and regulation during germination. The endosperm role during germination is fundamental in endosper‐ mic seeds; however, in non‐endospermic seeds, this role must befall on either the embryo or the testa. Since the testa is a nonliving tissue, the radicle most certainly assumes part of this regulatory role, but a comparative analysis is needed to ascertain this supposition and to determine if some of the endosperm functions are developed by the testa while still in the maturation program. The occurrence of endospermic and non‐endospermic seeds within the same taxa is relatively common in legumes such as *soybean*, which could offer a model for analyzing transcriptomic differences within embryo compartments comparable to the differences described between the endosperm and the radicle.

ROS participation in germination is supported by several reports and transcriptomic profiles of germinating seeds [9, 14, 17]. However, the actual role of ROS and ROS‐related enzymes in promoting cell wall loosening needs to be further analyzed, since the physiological concen‐ trations of ROS during germination do not seem to be sufficient to induce wall extension, and attempts of increasing ROS concentration lead to wall breakage [5]. Müller et al. [72] de‐ scribe abnormal rupture of the micropylar endosperm of *Lepidium* seeds treated with H2O2, while the treatment with myrigalone A [73], which inhibits the hormone‐mediated accumu‐ lation of ROS during germination, also induces abnormal endosperm breakage. These observations further support the notion of ROS as a signaling agent that induces down‐ stream activation of CWMPs than inducing wall loosening on its own.
