Isolation of mutants with abnormal petal epidermal cell morphology

ABSTRACT

Plants consist of many different cell types with specific shapes optimized for their particular functions. For example, most flowering plants have conically shaped epidermal cells on the upper surface of their petals that are important for pollinator attraction. The control of cell morphology in organs such as roots and leaves has been extensively studied, but much less is known about the genes that promote conical expansion of petal epidermal cells. We have developed a technique to rapidly assay the morphology of conical petal epidermal cells, and we employed this method in an unbiased genetic screen to identify mutants that alter the development of these cells. Mutants isolated in this screen affected cell shape, cell size, cuticle synthesis, and cellular chirality. This approach allowed for the identification of novel cellular components that are critical for the morphology of conical petal cells, and demonstrates the usefulness of petal epidermal cells as a model system for studying cellular morphogenesis.

The adaxial epidermal cells on the petals of most flowering plants have a distinctive conical shape.¹ The shape of these conical petal epidermal cells affects wettability and light focusing and is important for pollinator attraction.^2-4 In Antirrhinum majus, the MYB-related transcription factor MIXTA is required for conical cell expansion, but relatively little is known about the downstream factors that directly regulate cellular properties to promote conical cell morphogenesis.

Like most flowering plants, Arabidopsis thaliana has conical epidermal cells covered by cuticular nanoridges⁵ (Fig 1A). Scanning electron microscopy (SEM) can be used to visualize the morphology of these cells at high resolution, but is not ideal for high-throughput analyses. To enable the identification of mutants affecting the conical shape of petal epidermal cells, we developed a new technique for assaying petal cell morphology. When petals are folded in half along the transverse axis, the adaxial epidermal cells along the crease protrude, displaying a profile view of their conical regions that is easily observed using light microscopy (Fig 1B). Using this approach, the conical shape of wild-type epidermal cells is obvious (Fig 1B), and abnormally shaped cells are readily apparent (Fig 1C-E). To rapidly determine the shape of conical cells, we folded petals in half on a piece of double-sided tape adhered to a microscope slide and then viewed them at 20X with brightfield optics.

Figure 1.

Arabidopsis petal epidermal cells are conical. A) Scanning electron microscopy (SEM) of wild-type Arabidopsis adaxial petal epidermal cells showing their conical shape and cuticular nanoridges. B-E) Folded petals observed at 20X magnification by light microscopy. B) Wild-type petal with conical cells. C-E) Mutants with skinnier, flatter, or more cylindrical cells are readily apparent. Panels C, D, and E correspond to the same isolates as Fig. 2 panels A, C, and E, respectively. Scale bars are 10 µm for all panels.

We used this assay to perform a genetic screen to identify mutants with abnormal conical petal cell morphology. We screened approximately 10,000 F₂ progeny of wild-type Ler Arabidopsis plants that had been mutagenized with ethyl methanesulfonate (EMS). For each isolate, the gross petal morphology was observed by eye, and then folded petals were examined. We identified 17 mutants with reproducible changes in adaxial epidermal cell morphology. Some of the mutants exhibited defects in petal morphology that were visible by eye, while many other mutants had superficially normal petals but showed obvious defects in epidermal cell morphology in our folded petal assay. For several loci only a single allele was isolated, but for other loci multiple alleles were isolated, suggesting that the screen was close to saturation.

Some mutants specifically affected the dimensions of the petal epidermal cell cones. For example, in two allelic mutants, the base of the conical region failed to extend to the cell margins, leading to narrower cones (Fig 2A). Another mutant had flatter more domelike petal epidermal cells (Fig 2B), and several other mutants had elongated conical cells.

Figure 2.

Mutants with abnormal conical petal epidermal cell morphology. SEM images of mutant adaxial petal epidermal cells. The gene affected is listed in the bottom right corner when known. Scale bars are 10 µm for all panels.

Five mutants were isolated with defects in both epidermal cell shape and in the production of cuticular nanoridges. Two allelic mutants had flatter conical cells and reduced nanoridges; the weaker allele had shorter cones and a partial decrease in nanoridge density (Fig 2C), while the stronger allele almost entirely lacked nanoridges (Fig 2D). These two allelic mutants affected the nanoridges, but were not permeable to toluidine blue, indicating an intact cuticle (data not shown). We also isolated three mutants with cylindrical epidermal cells that were much less pointed than the wild-type and entirely lacked nanoridges (Fig 2E, F). Because of similarities to previously reported mutants implicated in the production of cuticular nanoridges,⁶ we performed DNA sequencing of candidate loci which indicated that two of the isolates carried mutations in CYP77A6 and one had a mutation in GPAT6. Based on the cylindrical shape of these mutant cells, we suggest that the cuticle provides structural rigidity that allows the conical epidermal cells to restrict cell expansion to form the tip of the cone. This is a role normally thought to be carried out by cell walls, revealing a novel function for the cuticular ridges.

Five mutants displayed irregular cell size, with a mix of normal size and abnormally large epidermal cells (Fig 2G). Genetic mapping experiments indicated that four of these mutations are in the same interval and are likely to be allelic, while the fifth mutant had a relatively weak phenotype and was not studied further. This irregular cell size phenotype is highly similar to previously reported smt2/cvp1/frill1 and et2 mutants that exhibit ectopic endoreduplication of petal cells.^7-9 However, the genetic mapping experiments revealed that these four mutations are not alleles of SMT2/CVP1/FRILL1 or ET2. Thus, these mutants likely represent a novel regulator of endoreduplication and polyploidy in petals required to ensure a consistent petal cell size.

Finally, we identified a mutation in RHM1 that causes left-handed twisting of petal cells (Fig 2H), that we have described in depth elsewhere.¹⁰ Despite RHM1 being broadly expressed, a helical growth phenotype for rhm1 mutants had not been previously reported. Our discovery of a role for RHM1 in the regulation of cellular chirality demonstrates that conical petal cells can be used to identify new genes that control cell morphology. Petals have long been recognized as a powerful model for studying organogenesis,⁵ and we have now shown that petal epidermal cells are also a valuable system for investigating how specific cell types reproducibly achieve their characteristic morphologies.

Acknowledgments

This project was supported by grant MCB-1615387 from the National Science Foundation to V.F.I., and A.M.S. was supported in part through a Yale University Brown Fellowship.