Adaxial-abaxial patterning

Abstract

Adaxial-abaxial patterning in lateral organ development is important for proper tissue differentiation and the entire shape of the organs. Although many transcriptional regulators are known to be involved in adaxial-abaxial patterning, the molecular mechanisms of the initial step of adaxial-abaxial polarity formation are still unclear. To determine these mechanisms, we have been analyzing the abaxial-specific expression of FILAMENTOUS FLOWER (FIL), which encodes a putative transcription factor. Recently, we found that the enf1 mutant, which has a mutation in the succinicsemialdehyde (SSA)-degrading enzyme, reduces the robustness of FIL expression patterning and has abnormally shaped leaves.¹ Here, we show that the transcriptomic data of enf1 provide more information on the relationship between SSA and adaxial-abaxial patterning, and we discuss the novel metabolic pathways of SSA production and the potential that the enf1 mutant represents a new tool in research on adaxial-abaxial polarity formation.

Adaxial-abaxial polarity is found in lateral organs, including leaves and floral organs. For example, adaxial mesophyll tissue (palisade tissue) is composed of uniform cylindrical cells arranged in a tightly packed manner, whereas abaxial mesophyll tissue (spongy tissue) is composed of more randomly arranged cells with large extracellular spaces. In the vascular bundle, xylem and phloem are differentiated on the adaxial side and on the abaxial side, respectively. The distributions of trichomes and stomata are different between the adaxial and abaxial epidermis. The adaxial-abaxial axis is also necessary for the lateral growth of lateral organs.^2,3

The transcriptional networks during adaxial-abaxial polarity development have been revealed.^4-6 In contrast to the large number of genes involved in adaxial-abaxial polarity development, the molecular mechanism of the initial step of adaxial-abaxial polarity formation is still unknown. Microsurgical experiments showed that leaf initial cells separated from the shoot apical meristem grew similar to a rod, with no or fewer adaxial tissue structures.^2,7-9 These reports have suggested that some signals from the shoot apical meristem promote the adaxial characteristics of the leaf primordia on the side near the shoot apical meristem.

To determine the molecular mechanism of the earliest step of adaxial-abaxial polarity formation, we have been investigating the regulatory mechanism of the expression of the FILAMENTOUS FLOWER (FIL) gene, which encodes one of the YABBY family proteins that binds DNA and interacts with corepressors.^1,10-14 FIL is expressed on the abaxial side of lateral organs^10,15 and their initial cells.^12,16 The loss-of-function mutant of FIL and its homolog YABBY3 (YAB3) has uniform mesophylls with little polarity and an abaxial epidermis with a mosaic of adaxial and abaxial tissue characteristics.¹⁵ In addition, the ectopic expression of FIL or YAB3 results in an adaxial epidermis with mosaic abaxial epidermal characteristics,^10,15 indicating that the abaxial-specific expression of FIL and YAB3 is necessary for the establishment of polarized tissue differentiation along the adaxial-abaxial axis. A 6 kbp sequence upstream of FIL is sufficient to drive abaxial-specific expression and contains at least two cis elements: an enhancer that drives the expression in lateral organs and a repressor that silences the expression on the adaxial side of the lateral organs.¹² Taken together, the adaxializing signals emanating from the shoot apical meristem activate the transcriptional repressors, which can bind the upstream sequence of FIL to silence expression on the adaxial side. Whether the adaxializing signals regulate the expression of FIL directly or indirectly is still not clear.

To identify the upstream regulators of the adaxial-specific expression of FIL, we isolated mutants in which the GFP expression pattern driven by the FIL upstream sequence (FILpro:GFP) is different from that of the wild type.¹ One of the mutants, enlarged fil expression domain1 (enf1), has a mutation in the gene encoding SUCCINIC SEMIALDEHYDE DEHYDROGENASE (SSADH).¹ The leaf primordia of enf1 show various patterning defects in the expression pattern of FILpro:GFP along the adaxial-abaxial axis, such as adaxialized leaves and abaxialized leaves.¹ The leaves of enf1 also show various patterns of shapes.¹ SSADH catalyzes the reduction of succinic semialdehyde (SSA) to succinic acid using NAD⁺ as a cofactor.¹⁷ A mutation in the GAMMA-AMINOBUTYRIC ACID TRANSAMINASE1 (GABAT1) gene, which encodes an enzyme catalyzing the conversion of gamma-aminobutyric acid (GABA) to SSA, suppresses the adaxial-abaxial patterning abnormality and the leaf shapes of enf1,¹ indicating that the increased and/or ectopic accumulation of SSA or its derivatives causes the adaxial-abaxial patterning phenotypes in enf1 plants. It has also been reported that a loss-of-function mutant of SSADH shows growth defects, an accumulation of reactive oxygen intermediates, and cell death in leaf lamina and that these phenotypes are also suppressed by a gabat1 mutation,^18,19 indicating that the excess or ectopic accumulation of SSA or its derivatives affects a large spectrum of plant development and cell states. Indeed, the exogenous application of SSA or its derivatives, including gamma-hydroxybutyric acid (GHB), causes abnormal leaf development and growth defects.^1,19

To gain further insight into the function of SSA and its derivatives in adaxial-abaxial polarity formation, we performed a transcriptome analysis using the shoot apex of wild-type and enf1 plants. We identified the differentially expressed genes (DEGs) in enf1 and compared them with another set of DEGs in the triple mutant of fil and its homologs yab3 and yab5.²⁰ We found that the group of genes that were increased in enf1 was enriched in genes that increased in the fil yab3 yab5 triple mutant and that the group of genes that were decreased in enf1 was enriched in genes that were decreased in the fil yab3 yab5 triple mutant (Fig. 1), suggesting that the increased and/or ectopic accumulation of SSA or its derivatives results in gene expression changes in a similar manner to the depletion of YABBY functions in vivo. These results suggest that SSA or its derivatives regulate the gene expression of the YABBY gene or YABBY protein functions.

Figure 1. Comparison between differentially expressed genes (DEGs) in enf1 and fil yab3 yab5. RNA from shoot apices of wild-type, enf1-1, and enf1-6 were extracted. The DEGs in enf1 were the genes that are differentially expressed in both enf1 alleles by emipirical bayes method compared with wild-type (false discovery rate < 0.05) after 75-percentile normalization. We used Agilent microarray, whereas Sarojam and colleagues used Affimetrix microarray.²⁰ So we only used the overlapped genes. The numbers of the increased (up) and decreased (down) expressed genes in enf1 were 386 and 388, respectively. Vertical axis represent the ratio of gene number of the higher (blue) or lower (red) expression in enf1 in the genes of higher (left) or lower (right) expression in fil yab3 yab5.

Although this evidence supports that SSA or its derivatives can influence adaxial-abaxial polarity, we did not know the roles of SSA or its derivatives at wild-type concentrations and distribution because we could only determine the situation in vivo with the overexpression and/or ectopic accumulation of SSA or its derivatives in these experiments using the enf1/ssadh mutant and exogenous application. To determine the native function of SSA and its derivatives, we should analyze a mutant in which SSA or its derivatives are not or less produced. Although GABAT1 is the only enzyme known to produce SSA in Arabidopsis, the gabat1 mutant still accumulates SSA at the same level as wild type,¹ indicating that other enzymes produce SSA in addition to GABAT1 in Arabidopsis. One candidate enzyme might be a gamma-hydroxybutyrate (GHB) dehydrogenase (GHBDH), which catalyzes the conversion from GHB to SSA.²¹ However, because the GHBDH1 protein was shown to be a SSA/glyoxylate reductase rather than GHBDH^22,23 and it is still unknown how GHB was produced other than from SSA in plants, it is difficult to further analyze GHBDH as an SSA-producing enzyme. There is also evidence for the existence of other pathways that produce SSA. The cotyledons of the enf1 mutant show variegation similar to rosette leaves (Fig. 2). Unlike rosette leaves, a gabat1 mutation does not suppress the enf1 cotyledon phenotype, suggesting that the abnormality in cotyledon development can be attributed to SSA-producing enzymes other than GABAT1. Identifying the unknown SSA-producing enzymes would aid in understanding the role of SSA under natural conditions.

Figure 2. Variegation of the cotyledons of wild type, enf1, gabat1 and enf1 gabat1. (A) Typical phenotypes of cotyledons. The enf1-1 and enf1-1 gabat1-1 cotyledons have white sectors, whereas wild-type and gabat1-1 cotyledons have no white sectors. Scale bars represent 1 mm. (B) Quantification of cotyledons with a white sector. Bars represent standard deviations from seven biological replicates. Different characters indicate statistically different by Welch’s t test (p < 0.05).

Regardless of whether native SSA and its derivatives function in leaf development, excess or ectopic SSA disturbs adaxial-abaxial polarity formation in leaves. Identifying the target molecules of excess or ectopic SSA will aid in further understanding the molecular mechanism establishing the adaxial-abaxial pattern in the leaves and other lateral organs.