In theory mutants affecting the development of specific tissue types within plant organs should provide a useful source of information regarding the relative contributions of different tissues to final leaf shape. In this scenario it is a cell-type rather than a cell lineage which is affected and thus, although some accommodation may take place between cell layers, incursions from different cell lineages should have no effect on the conclusions. A good example is the Arabidopsis reticulata mutant which shows a strong reduction in mesophyll cell density, apparently due to a defect in mesophyll cell proliferation during development (Gonzalez-Bayon et al. 2006). The paucity of mesophyll cells in the lamina gives mutant leaves a pale aspect, with veins showing up as darker lines. The mutant epidermal cell-layer appears phenotypically wild-type and, moreover, the overall form of the lamina is comparable to that in wild-type plants although slightly reduced in size. The ability of this mutant to produce a normal lamina despite its lack of mesophyll cells does provide an intriguing indication that the leaf epidermis could be the primary determinant of leaf shape, although, since the exact role of the RETICULATA protein is unclear, it is perhaps unwise to draw too many conclusions from this mutant. Additionally, the role of L3-derived vascular tissue in regulating leaf shape should not be underestimated.
Theoretically mutants which are thought to be directly compromised in specification of L1 identity might provide a more direct source of information regarding the role of the L1 in growth control. The HDZipIV-class home-odomain proteins ATML1 and PDF2 are expressed in the L1 in Arabidopsis meristems and organ primordia during shoot development (Lu et al. 1996; Sessions et al. 1999; Abe et al. 2003). They act redundantly to specify epidermal identity in shoot organs and maintain the meristematic L1 cell layer (Abe et al. 2003). They are thought to act by binding elements called L1 boxes in the promoters of targets, which include their own genes and genes involved in cuticle biosynthesis and epidermal differentiation such as FIDDLEHEAD (Abe et al. 2001). Double atml1 pdf mutants show characteristic defects which start during embryogenesis with abnormal cell division patterns in the apical part of the embryo including the meristem L1 layer and epidermis of the cotyledon primordia. Cotyledons frequently fail to form properly and are commonly fused. Although one or two leaf-like organs can form they are small, misshapen and usually consume the poorly maintained shoot apical meristem. If the meristem is not consumed it differentiates as vacuolarised cells. The surfaces of shoot organs are covered with cells which resemble mesophyll, with the exception of occasional stomatal clusters, which indicate that double mutants do retain some epidermal identity (Abe et al. 2003; Ingram, unpublished data). This may be provided by another related protein from the same family (Nakamura et al. 2006; Ingram, unpublished data). The inability to form morphologically normal organs when epidermal identity is compromised is an important indication of the role of the epidermis in organ growth and morphogenesis, and correlates with the observations obtained from ablation studies discussed above. However, the extreme nature of the organ phenotypes shown by these mutants, and the accompanying meristem degeneration, preclude their use in quantitative studies.
The fact that the defects observed in atml1/pdf2 double mutants do not affect early embryogenesis suggests three things. (1) Either epider-mal/protodermal identity is not entirely lost in these mutants, and enough is left to permit normal protoderm specification during early embryogenesis or (2) ATML1, PDF2 and related genes are required for epidermal differentiation in late embryogenesis and post germination, but not for the specification of protoderm identity during early embryogenesis. (3) That conclusions from chemical ablation studies suggesting that specification of the protoderm is necessary for early embryogenesis to proceed normally are wrong. Ongoing studies will distinguish between the three possibilities, although the phe-notypes associated with reduced function of the Arabidopsis AtDEK1 gene provides a tantalising indication that the first explanation could be correct. AtDEK1 is a likely membrane-bound cysteine protease that was first isolated by homology to the maize DEFECTIVE KERNEL1 (DEK1) gene (Lid et al. 2002, 2005; Johnson et al. 2005). Strong mutants in DEK1 in maize show defects in the specification of aleurone cell identity in the endosperm, and early embryo arrest. Weaker alleles show defects in leaf development, including aberrant specification of specific epidermal cell-types (Becraft et al. 2002). Null mutants in ADEK1 show early embryonic arrest and defects in the patterning of endosperm cell divisions in Arabidopsis. Null mutant embryos show aberrant cell division patterns and, notably, do not form a morphologically distinct protodermal cell layer (Johnson et al. 2005; Lid et al. 2005) and do not express epidermal markers, including ATML1 (Johnson et al. 2005). The extreme nature of the null mutant phenotype, and the unrestricted expression pattern of ADEK1 make it difficult to conclude whether the role of ADEK1 in young embryos is restricted to the protoderm. However, RNAi knockdown of ADEK1 in all cell-layers during late embryogenesis leads to the differentiation of mesophyll in the place of epidermal cells on the cotyledons, and a loss of meristematic activity in the SAM; phenotypes that are strongly reminiscent of those shown by atml1/pdf2 double mutants. This indicates that at least one of the roles of ADEK1 is the maintenance of epidermal identity in embryonic cells (Johnson et al. 2005). In plants with reduced ADEK1 expression due to RNAi, cotyledons appear narrower than in the wild-type, possibly due to aberrant lateral expansion of the lamina, and the SAM is usually lost so that no post-embryonic organs are produced.
Was this article helpful?