The COMATOSE ATP-binding cassette transporter is required for full fertility in Arabidopsis

Abstract

COMATOSE (CTS) encodes a peroxisomal ATP-binding cassette transporter required not only for beta-oxidation of storage lipids during germination and establishment, but also for biosynthesis of jasmonic acid and conversion of indole butyric acid to indole acetic acid. cts mutants exhibited reduced fertilization, which was rescued by genetic complementation, but not by exogenous application of jasmonic acid or indole acetic acid. Reduced fertilization was also observed in thiolase (kat2-1) and peroxisomal acyl-Coenzyme A synthetase mutants (lacs6-1,lacs7-1), indicating a general role for beta-oxidation in fertility. Genetic analysis revealed reduced male transmission of cts alleles and both cts pollen germination and tube growth in vitro were impaired in the absence of an exogenous carbon source. Aniline blue staining of pollinated pistils demonstrated that pollen tube growth was affected only when both parents bore the cts mutation, indicating that expression of CTS in either male or female tissues was sufficient to support pollen tube growth in vivo. Accordingly, abundant peroxisomes were detected in a range of maternal tissues. Although gamma-aminobutyric acid levels were reduced in flowers of cts mutants, they were unchanged in kat2-1, suggesting that alterations in gamma-aminobutyric acid catabolism do not contribute to the reduced fertility phenotype through altered pollen tube targeting. Taken together, our data support an important role for beta-oxidation in fertility in Arabidopsis (Arabidopsis thaliana) and suggest that this pathway could play a role in the mobilization of lipids in both pollen and female tissues.

Figure 1.

Floral phenotype of cts mutants. A, Cumulative flowering on the primary inflorescence (n = 23–32). B, Photograph of partially dissected flowers (stages 13–15) from primary inflorescences of Arabidopsis, accession Ler, and the cts-1 mutant. C, Ratio of long stamen/pistil lengths in different genotypes, including cts-1 plants transformed with a CTS promoter-cDNA cassette (cts-1∷CTS). Values are means ± se. D, Photograph of partially dissected flowers (stage 13) of wild-type (Ler; Ws2), cts mutants, and cts plants treated with 10 μm NAA.

Figure 2.

Silique production in cts mutants and respective wild types. A, Total silique production. B, Total silique dry weight. C, Silique length. Black circles, Ler; white circles, cts-1. Black triangles, Ws2; white triangles, cts-2. Values are means ± se (n = 10–17).

Figure 3.

Fertilization in cts mutants. A, Percentage fertilized ovules and aborted embryos in siliques of Ler, cts-1, and cts-1 plants transformed with a CTS promoter-cDNA cassette (cts-1∷CTS). Siliques were harvested 10 DAF. Values are means ± se (n = 10–15 siliques). B, Percentage fertilized ovules and aborted embryos in siliques of wild type, Ws2, and β-oxidation mutant alleles: cts-2, kat2-1, lacs6-1,lacs7-1. The Ws4 wild-type behaved identically to Ws2 (data not shown). Values are means ± se (n = 11–18 siliques). Asterisks in A and B denote significant differences to wild type at 5%, as determined following ANOVA, using the lsds on the transformed data. C, Dissected siliques (harvested 10 DAF) showing unfertilized ovules (white arrowheads) and aborted embryos (red arrowheads) in cts mutants.

Figure 4.

Pollen tube growth in cts mutants and respective wild types. A, Pollen tube growth in mutant and wild-type pistils was determined in self- and cross-pollinated pistils. Pistils were removed 24 h after pollination and stained with aniline blue. Bar = 500 μm. B, Pistil lengths of Ler, cts-1, Ws2, and cts-2 flowers (stage 14). Values are means ± se (n = 22–28). Asterisk denotes that cts-2 is significantly different than Ws2 at 2%, as determined by ANOVA.

Figure 5.

Peroxisomes in reproductive tissues of Arabidopsis. Peroxisomes were visualized by confocal laser-scanning microscopy of plants expressing peroxisomally targeted enhanced GFP. A, Stage 12 flower showing GFP expression in petal, stamens (anther and filament), and pistil. Scale bar = 50 μm. B, Close-up of anther from stage 12 flower showing lack of expression in developing pollen grains (arrowheads) within the anther. Scale bar = 20 μm. C, Pollen germinated 4 h in vitro in Suc-containing medium, showing lack of expression. Scale bar = 10 μm. D, Dissected silique from stage 16 showing peroxisomes in the gynoecium wall and transmitting tissue. Scale bar = 50 μm. E, Higher magnification image showing peroxisomes within the transmitting tissue and funiculus. Scale bar = 10 μm. F, Peroxisomes in the fertilized ovule and funiculus from a stage 16 flower. Scale bar = 10 μm.

Figure 6.

GABA and GABA shunt metabolite content in flowers. A, Proposed scheme for GABA metabolism in Arabidopsis flowers. GABA is synthesized in the cytosol from Glu by Glu decarboxylase (GAD). The first step in GABA catabolism is catalyzed by mitochondrial GABA transaminase (GABA-T). In Arabidopsis flowers, this enzyme is encoded by POP2 and utilizes pyruvate (pyr) to yield Ala (ala) and succinic semialdehyde (SSA). Mutation of POP2 results in elevated GABA. In mitochondria, SSA undergoes oxidation to succinate, catalyzed by succinic semialdehyde dehydrogenase (SSADH). Alternatively, under hypoxia, SSA is reduced via the cytosolic enzyme, γ-hydroxybutyrate dehydrogenase (GHBDH; also known as SSA reductase), to yield GHB. It is possible that GHB, a short-chain hydroxy fatty acid, is metabolized further by β-oxidation, by analogy with mammalian systems (the dotted line indicates that this step is hypothetical). β-Oxidation of GHB is potentially inhibited in cts and kat2-1 mutants. B, Content of GABA, GHB, and succinate in Arabidopsis flowers. Values are means ± se (n = 4). A t test was performed comparing every mutant and its respective wild type (Ler or Ws2). Significant values (P < 0.05; critical t value 2.571) are indicated with an asterisk. Significant differences in GABA content were also found between the two Arabidopsis wild types.

Figure 7.

Pollen germination and tube growth in vitro. A, Pollen from stage 14 flowers was germinated in vitro in the presence or absence of a carbon source (2% Suc). Values are means ± se of three independent slides (approximately 800 grains/slide). B, Pollen tube growth in different genotypes: Ler (wild type), cts-1, and cts-1 transformed with bacterial artificial chromosome clone 159N1 (cts-1∷CTS); Ws2 (wild type), cts-2, kat2-1, and lacs6-1,lacs7-1. Pollen was germinated in vitro in the presence or absence of a carbon source (2% Suc) and pollen tube lengths of germinated pollen determined after 18 h. Values are means ± se (n = 50). Asterisks denote significant differences from wild type at 5%, as determined following ANOVA, using the lsds on the transformed data for A. Data presented in A and B are representative of several independent experiments.