Seeds of Arabidopsis plants expressing dominant-negative AtSKD1 under control of the GL2 promoter show a transparent testa phenotype and a mucilage defect

Abstract

We have recently shown that overexpression of dominant-negative AtSKD1 versions under control of the trichome and non-root-hair-cell specific GL2 promoter (GL2_pro) blocks trafficking of soluble cargo to the vacuole, resulting in its fragmentation and ultimately cell death. GL2_pro is also active in the Arabidopsis seeds. When we inspected seeds of the dominant-negative AtSKD1 variants we found two phenotypes. The seeds display a transparent testa phenotype caused by a lack of proanthocyanidin (PA) and do not possess seed coat mucilage. Both phenotypes could be connected by cell death induced by the overexpression of dominant-negative AtSKD1.

AAA ATPases are important regulators of a plethora of cellular functions such as peroxisome biogenesis, vesicle-mediated transport, control of cell divisions and gene expression. This variety is based on a common mechanism, the energy dependent unfolding, remodeling and disassembly of proteins and protein complexes.¹

Mammalian SKD1 and its yeast homolog VPS4 are AAA ATPases involved in the sorting of monoubiquitylated trans-membrane cargo to the lysosome/vacuole by dismantling the members of the endosomal sorting complex required for transport (ESCRT) complexes from the endosomal membrane.² The Arabidopsis SKD1 homolog, AtSKD1, has been characterized recently and has been shown to be an ortholog of SKD1/Vps4.^3,4 Mutations for all three ATPases are known that alone or in combination render them dominant-negative.^3–6 Overexpression of dominant-negative AtSKD1 is, however, lethal for Arabidopsis plants. Therefore, we have used the trichome and non-root-hair-cell-specific GL2_pro promoter to address the function of AtSKD1 in planta. Cells that express the dominant-negative versions show multiple nuclei, fragmented vacuoles and ultimately die. These phenotypes are most likely due to a block in vacuolar trafficking of soluble cargo that is instead secreted.⁴

GL2_pro is also active in the coat of developing Arabidopsis seeds. We therefore inspected seeds of lines overexpressing dominant-negative and wild-type AtSKD1.

Seeds of Dominant-Negative AtSKD1 Lines Display a Transparent Testa Phenotype and Do Not Contain PA

We have constructed three dominant-negative versions of AtSKD1 under GL2_pro control [in the following summarized as AtSKD1(dn)]: K178A that blocks ATP binding, E232Q that blocks ATP hydrolysis and a combination of both mutants (AQ). These versions and the AtSKD1(WT) control were stably transformed into Col-0 plants and the seeds of homozygous T3 plants inspected. All three dominant-negative versions had pale seeds (Fig. 1E–G) comparable to the transparent testa (tt) mutants ttg1,⁷(Fig. 1C) and tt1,⁸(Fig. 1D). Col-0 (Fig. 1A) and the AtSKD1(WT) controls (Fig. 1B), by contrast, showed normal brown seed coloration that is caused by oxidation of the proanthocyanidins in the vacuole of cells derived from the inner integument during seed development.⁹

Figure 1.

Transparent testa phenotype of AtSKD1(dn). Seeds are shown from (A) Col-0, (B) GL2_pro:AtSKD1(WT), (C) ttg1, (D) tt1, (E GL2_pro:AtSKD1(K178A), (F) GL2_pro:AtSKD1(E232Q) and (G) GL2_pro:AtSKD1(AQ).

To test whether the AtSKD1(dn) lines had a reduced PA content we performed a DMACA staining as described by Bogs and colleagues.¹⁰ As expected, we found that staining was absent in AtSKD1(dn) (Fig. 2E–G) and the tt mutant controls (Fig. 2C and G) whereas Col-0 and AtSKD1(WT) displayed the dark blue DMACA color indicating the presence of PA.

Figure 2.

AtSKD1(dn) seeds do not contain proanthocyanidins. DMACA stained seeds are shown from (A) Col-0, (B) GL2_pro:AtSKD1(WT), (C) ttg1, (D) tt1, (E) GL2_pro:AtSKD1(K178A), (F) GL2_pro:AtSKD1(E232Q) and (G) GL2_pro:AtSKD1(AQ).

Seeds of Dominant-Negative AtSKD1 Lines do not Possess Seed Coat Mucilage

In the process of the DMACA staining we realized that the AtSKD1(dn) seeds were clogging in the staining solution. This is indicative for the lack of seed coat mucilage. Therefore, we performed Ruthenium red staining¹¹ to check the presence of mucilage in dominant-negative and wild-type AtSKD1 lines. Col-0 plants and plants expressing the wild-type version of AtSKD1 showed the Ruthenium red stained halo that indicates the acidic pectins of the mucilage (Fig. 3A and B). The dominant-negative lines did not show any red staining confirming the lack of seed coat mucilage (Fig. 3C–E).

Figure 3.

ATSKD1(dn) seeds do not possess seed coat mucilage. Ruthenium red stained seeds are shown from (A) Col-0, (B) GL2_pro:AtSKD1(WT), (C) GL2_pro:AtSKD1(K178A), (D) GL2_pro:AtSKD1(E232Q) and (E) GL2_pro:AtSKD1(AQ).

How are the Seed Coat Phenotypes Related to AtSKD1's Cellular Functions?

The transparent testa phenotype of AtSKD1(dn) could be directly explained by the sorting defect we have described for AtSKD1(dn). The formation of the brown color requires transport of the PA precursors into the vacuole. This task is performed by transporters such as TT12,¹² and/or TT19.¹³ If transport is blocked by e.g., mislocalization of the transporter a tt phenotype is the consequence. This explanation is, however, not applicable for the mucilage defect. Here secretion is the predominant process. Secretion, however, is not affected in dominant-negative AtSKD1. Therefore, we consider a developmental defect the most likely scenario to explain both phenotypes by a common mechanism. As we have shown earlier, cells overexpressing dominant-negative AtSKD1 ultimately die as a consequence of vacuolar fragmentation.⁴ Perhaps, in the seeds of the AtSKD1(dn) lines cells expressing the transgene do so before they are able to fulfill their function to either accumulate PA or produce mucilage.

Addendum to: Shahriari M, Keshavaiah C, Scheuring D, Sabovljevic A, Pimpl P, Häusler RE, et al. The AAA-ATPase AtSKD1 contributes to vacuolar maintenance of A. thaliana. Plant J. 2010;64:71–85. doi: 10.1111/j.1365-313X.2010.04310.