The ULTRAPETALA1 trxG factor contributes to patterning the Arabidopsis adaxial-abaxial leaf polarity axis

Abstract

The SAND domain protein ULTRAPETALA1 (ULT1) functions as a trithorax group factor that regulates a variety of developmental processes in Arabidopsis. We have recently shown that ULT1 regulates developmental patterning in the gynoecia and leaves. ULT1 acts together with the KANADI1 (KAN1) transcription factor to pattern the apical-basal axis during gynoecium formation, whereas the 2 genes act antagonistically to pattern the adaxial-abaxial axis during both gynoecium and leaf formation. In particular, our data showed that ULT1 is necessary for the kan1 adaxialized organ phenotype. Here, we observe the internal structure of ult1, kan1 and ult1 kan1 rosette leaves to better understand the suppression of the kan1 adaxialized leaf polarity defect by ult1 mutations. Our results indicate that ULT1 and KAN1 act antagonistically to pattern the adaxial-abaxial axis in leaves by establishing the asymmetry of the internal cell layers.

The epigenetic regulation of gene expression is necessary for the correct deployment of developmental programs and for the maintenance of cell fates. Thousands of genes that control many essential biological processes are under epigenetic control in eukaryotes. Among the epigenetic regulators, Polycomb group (PcG) and trithorax group (trxG) proteins maintain the stable transcription patterns of key developmental regulatory genes by modifying histones to organize chromatin in an inactive or an active transcriptional state, respectively.¹ In Arabidopsis, ULTRAPETALA1 (ULT1) and the closely related ULT2 gene encode SAND domain proteins that function as trxG factors and promote transcriptional activity by limiting the deposition of repressive histone marks by PcG complexes.² ULT1 regulates a variety of developmental processes in Arabidopsis, such as the reproductive transition, shoot apical meristem maintenance, and floral meristem termination.^3-5 Recently, we demonstrated that the ULT1 and ULT2 trxG proteins also regulate developmental patterning in Arabidopsis gynoecia and leaves.^6,7

Arabidopsis leaves, like the evolutionary related gynoecia, are arranged along 3 axes of polarity: adaxial-abaxial, medial-lateral and proximal-distal. The establishment of adaxial-abaxial leaf polarity leads to the development of distinct tissues within the adaxial (top) and abaxial (bottom) domains. The adaxial surface of the leaf develops a thickened waxy cuticle that results in a glossy, dark green and trichome–rich epidermis, whereas the abaxial epidermis is matte, gray-green and, especially in the early leaves, trichome-poor. Internally, the water-conducting xylem tissue in the vasculature forms adaxially to the sugar-bearing phloem tissue. In addition, the adaxial region develops tightly packed layers of palisade mesophyll cells that optimize light capture, whereas the abaxial region contains loosely packed layers of spongy mesophyll cells as well as a higher density of stomatal pores that facilitate gas exchange and regulate transpiration to maximize photosynthesis.⁸

The establishment of adaxial-abaxial leaf polarity is controlled by the opposing activities of region-specific transcriptional regulators. Abaxial identity in Arabidopsis leaves is primarily conferred by the KANADI (KAN) and AUXIN RESPONSE FACTOR (ARF) families of transcription factors.^9-11 The KAN genes encode transcription factors containing a MYB-like GARP DNA binding domain. KAN1 is specifically expressed in the abaxial domain of young leaf primordia,¹⁰ where it negatively regulates the expression of adaxial identity determinants¹² and physically interacts with the ARF3 protein.¹³ KAN1 acts primarily as a transcriptional repressor that targets genes involved in auxin biosynthesis, transport and signaling in opposition to class III HOMEODOMAIN LEUCINE ZIPPER (HD-ZIPIII) transcription factors that confer adaxial leaf cell identity.¹⁴

We recently demonstrated that, although the 2 genes act together to pattern the apical-basal axis during gynoecium formation, ULT1 acts antagonistically to KAN1 to pattern the adaxial-abaxial axis during gynoecium and leaf development.⁷ At the morphological level, wild-type and ult1 rosette leaves both consist of a nearly flat lamina that grows outward along the interface between the adaxial and abaxial surfaces.^2,15 In contrast, the first-arising rosette leaves of kan1 seedlings emerge in an upright position and remain cupped throughout their development, due to their partial adaxialization.¹⁰ It has been shown that this phenotype is the result of a decrease in the cross-sectional area of the adaxial palisade mesophyll cells, coupled with an increase in the cross-sectional area, as well as the number, of abaxial spongy mesophyll cells.¹⁰ A striking feature of ult1 kan1 plants is the suppression of the kan1 adaxialized leaf phenotype, with the seedlings displaying flat leaves indistinguishable from those of wild-type.⁷

Here we dissect the ult1 kan1 leaf phenotype by observing the internal structure of the leaves to better understand the nature of this suppression. We collected individual rosette leaves from 10-day-old wild type and mutant plants grown at 21ºC under LD conditions (16 hours light, 8 hours dark). The third and fourth rosette leaves were used for analysis. Histological sectioning of Toluidine blue-stained tissues was performed as previously described,¹⁶ using 6-micron tissue sections that were visualized using a Zeiss Axiophot microscope. Our data show that, at the midvein, both wild-type and ult1–2 leaves contain layers of smaller, chloroplast-dense palisade mesophyll cells on the adaxial side, and layers of larger, more rounded spongy mesophyll cells on the abaxial side (Fig. 1A and 1B). Furthermore, the abaxial cells are arranged in a U-shape below the midvein whereas the adaxial cells form relatively straight rows above it. However, in kan1–12 seedlings the cells in the bottom region of the leaf are smaller and more chloroplast dense, resembling adaxial cells, and the U-shaped curvature is lost (Fig. 1C). In ult1–2 kan1–12 leaves the asymmetry is restored, with the palisade mesophyll cells on the adaxial side and the spongy mesophyll cells on the abaxial side closely resembling those of wild-type leaves (Fig. 1D). These results confirm that ULT1 activity is required for the kan1 adaxialized leaf polarity defect, and indicate that ULT1 and KAN1 function antagonistically to establish the asymmetry of the internal cell layers along the adaxial-abaxial leaf polarity axis.

Figure 1.

Transverse sections across the midvein of (A) Ler, (B) ult1–2, (C) kan1–12, and (D) ult1–2 kan1–12 rosette leaves. Scale bars, 100 μm.

The opposite roles of ULT1 and KAN1 in patterning the adaxial-abaxial leaf polarity axis is consistent with expression data showing that ULT1 but not KAN1 is expressed in the adaxial domain of developing leaves.^10,17 Also, it has been demonstrated that KAN1 promotes abaxial identity by directly repressing the transcription of ASYMMETRIC LEAVES 2 (AS2) in abaxial leaf tissue.¹⁸ This role of KAN1 as a negative regulator that establishes abaxial cell fate contrasts with the role of ULT1 as predominantly a transcriptional activator.^2,19 We hypothesize that ULT1 may promote adaxial leaf cell identity by maintaining the expression of one or more adaxial determinants that, when mis-expressed in kan1 leaves, confers the adaxialized phenotype. An attractive candidate is AS2, an adaxial identity gene that is ectopically expressed throughout kan1 kan2 leaves.¹⁸ Other candidates are the HD-ZIPIII transcription factor genes PHABULOSA (PHB), PHAVOLUTA (PHV) and/or REVOLUTA (REV). These genes are expressed in the adaxial domain of developing leaf primordia,^20,21 and PHV and REV are ectopically expressed throughout kan1 kan2 leaves.⁹ Further experiments to analyze the expression of these adaxial identity genes in ult1, kan1 and ult1 kan1 leaves should allow us to better understand through which genetic pathway ULT1 contributes to organizing adaxial-abaxial polarity in Arabidopsis leaves.

Funding

This work was funded by a grant from the National Science Foundation (IOS-1052050) to J.C.F.

Disclosure of Potential Conflicts of Interest

No potential conflicts of interest were disclosed.