The Plant Journal (2011) 67, 648–661
doi: 10.1111/j.1365-313X.2011.04624.x
PHOSPHATIDYLSERINE SYNTHASE1 is required for microspore development in Arabidopsis thaliana Yasuyo Yamaoka1, Yanbo Yu1, Junya Mizoi1,†, Yuki Fujiki1, Kyoko Saito2, Masahiro Nishijima2,‡, Youngsook Lee3 and Ikuo Nishida1,* 1 Laboratory of Plant Molecular Physiology, Graduate School of Science and Engineering, Saitama University, Shimo-Okubo 255, Sakura-Ku, Saitama 338-8570, Japan, 2 Department of Biochemistry and Cell Biology, National Institute of Infectious Diseases, 1-23-1 Toyama, Shinjuku-ku, Tokyo 162-8640, Japan, and 3 Division of Molecular Life Sciences, Pohang University of Science and Technology, Pohang 790-784, Korea Received 17 February 2011; revised 19 April 2011; accepted 3 May 2011; published online 24 June 2011. * For correspondence (fax +81 48 858 3384; e-mail
[email protected]). † Present address: Laboratory of Plant Molecular Physiology, Graduate School of Agricultural and Life Sciences, The University of Tokyo, Bunkyo-ku, Tokyo 113-8657, Japan. ‡ Present address: Showa Pharmaceutical University, 3-3165 Higashi-Tamagawagakuen, Machidashi, Tokyo 194-8543, Japan.
SUMMARY Phosphatidylserine (PS) has many important biological roles, but little is known about its role in plants, partly because of its low abundance. We show here that PS is enriched in Arabidopsis floral tissues and that genetic disruption of PS biosynthesis decreased heterozygote fertility due to inhibition of pollen maturation. At1g15110, designated PSS1, encodes a base-exchange-type PS synthase. Escherichia coli cells expressing PSS1 accumulated PS in the presence of L-serine at 23C. Promoter-GUS assays showed PSS1 expression in developing anther pollen and tapetum. A few seeds with pss1-1 and pss1-2 knockout alleles escaped embryonic lethality but developed into sterile dwarf mutant plants. These plants contained no PS, verifying that PSS1 is essential for PS biosynthesis. Reciprocal crossing revealed reduced pss1 transmission via male gametophytes, predicting a rate of 61.6% pss1-1 pollen defects in PSS1/pss1-1 plants. Alexander’s staining of inseparable qrt1-1 PSS1/pss1-1 quartets revealed a rate of 42% having three or four dead pollen grains, suggesting sporophytic pss1-1 cell death effects. Analysis with the nuclear stain 4¢,6-diamidino-2-phenylindole (DAPI) showed that all tetrads from PSS1/pss1-1 anthers retain their nuclei, whereas unicellular microspores were sometimes anucleate. Transgenic Arabidopsis expressing a GFP-LactC2 construct that binds PS revealed vesicular staining in tetrads and bicellular microspores and nuclear membrane staining in unicellular microspores. Hence, distribution and/or transport of PS across membranes were dynamically regulated in pollen microspores. However, among unicellular microspores from PSS1/pss1-2 GFP-LactC2 plants, all anucleate microspores showed little GFP-LactC2 fluorescence, suggesting that pss1-2 microspores are more sensitive to sporophytic defects or show partial gametophytic defects. Keywords: phosphatidylserine, cell death, microspores, Lact-C2, very-long-chain fatty acids, Arabidopsis.
INTRODUCTION Phosphatidylserine (PS) plays an important role in cell death signaling, vesicular trafficking, lipid–protein interactions, and membrane lipid metabolism (Vance, 2008). However, a comprehensive understanding of the roles of PS and its interaction with regulatory proteins requires further cellular and molecular studies (Leventis and Grinstein, 2010). In bacteria and yeast, PS is synthesized by a PS synthase (PSS) that utilizes CDP-diacylglycerol (CDP-DG) and L-serine (Ser) as substrates (CDP-DG-dependent PSS or CD-PSS; EC 648
2.7.8.8; Matsumoto, 1997). In Escherichia coli, PS is rapidly converted to phosphatidylethanolamine (PE) by PS decarboxylase (PSD), so that little PS accumulates in cells (Hawrot and Kennedy, 1975). However, E. coli psdts mutants accumulate high PS levels at non-permissive temperatures, leading to cell lysis (Hawrot and Kennedy, 1978). In mammals, PS constitutes 2–10% of cellular lipids and is synthesized by a calcium-dependent base-exchange-type PSS (BE-PSS) that catalyzes an exchange reaction between ª 2011 The Authors The Plant Journal ª 2011 Blackwell Publishing Ltd
Microspore defects in Arabidopsis pss1 mutants 649 an existing phospholipid head group and Ser (Hu¨bscher et al., 1959). In cultured Chinese hamster (Cricetulus griseus) ovary (CHO) cells, Cg-PSS1 and Cg-PSS2 primarily utilize phosphatidylcholine (PC) and PE, respectively, to synthesize PS in the endoplasmic reticulum (ER; Kuge and Nishijima, 2003). The PS is either converted to PE by mitochondrial PSD (Shiao et al., 1995; Achleitner et al., 1999) or transported to the Golgi and then plasma membranes via vesicular trafficking (Leventis and Grinstein, 2010). In mammals, PS is essential for cell growth (Kuge et al., 1986) and functions as a macrophage clearance signal during apoptosis (Fadok et al., 1992; Schlegel and Williamson, 2001). Phosphatidylserine is enriched in plasma membrane lipid raft domains (Pike et al., 2002). A number of physiologically important proteins interact with PS (Vance and Steenbergen, 2005), including annexin V (Swairjo et al., 1995), protein kinase C (Sutton and Sprang, 1998), and coagulation factor V (Macedo-Ribeiro et al., 1999), all of which have putative PS-binding domains (Leventis and Grinstein, 2010). Phosphatidylserine is a precursor to lysophosphatidylserine, a hypothetical lipid signaling mediator under various pathophysiological conditions (Bruni et al., 1988). In rat brain, PS contains high levels of docosahexaenoic acid (22:6 [n-3]; Kim et al., 2004). Phosphatidylserine may regulate membrane surface charges for recruiting cationic proteins like Rac1 and K-Ras (Yeung et al., 2008). The mechanism of PS biosynthesis varies between plant species. Activity of BE-PSS has been reported in the ER of castor bean (Ricinus communis) endosperms (Moore, 1975), whereas CD-PSS activity has been detected in spinach (Spinacia oleracea) leaves (Marshall and Kates, 1974). In leek (Allium porrum) and carrot (Daucus carota) tissues, both types of PSS activities have been identified (Vincent et al., 1999; Manoharan et al., 2000). In leek, CD-PSS activity is largely enriched in the ER and partly enriched in Golgi membranes, whereas BE-PSS activity is enriched in the plasma membrane, Golgi, and ER fractions (Vincent et al., 1999). The physiological significance of PS biosynthesis in plants remains unclear. Phosphatidylserine is a relatively minor plant cell lipid (Devaiah et al., 2006; Nakamura and Ohta, 2007). In mung bean (Vigna radiata), PS accounts for 3.1 and 4.3% of glycerolipids in plasma membranes and tonoplasts, respectively (Yoshida and Uemura, 1986). In many plant species, PS contains a relatively large proportion of very-long-chain fatty acids (VLCFAs; Murata et al., 1984; Bohn et al., 2001; Vincent et al., 2001; Devaiah et al., 2006). Phosphatidylserine is enriched in 70- to 80-nm-diameter ER-derived vesicles, and PS molecular species that contain VLCFAs (PS-VLCFAs) are proposed to stabilize such small vesicles (Sturbois-Balcerzak et al., 1999). Phosphatidylserine accumulates in carrot suspension cells and oat (Avena sativa) roots under sugar-starved (Manoharan et al., 2000) and drought-stressed (Larsson et al., 2006) conditions, respectively, and overexpression of wheat CD-PSS in trans-
genic tobacco plants induces cell death (Delhaize et al., 1999). Arabidopsis thaliana contains a single gene for a BE-PSS (At1g15110; EC 2.7.8; designated PSS1) that is homologous to mammalian enzymes. We show here that PSS1 is required for PS biosynthesis in Arabidopsis, using pss1-1 and pss1-2 knockout alleles isolated from T-DNA insertion and TILLING (Till et al., 2003) lines, respectively. Lipid changes associated with PSS1 expression in E. coli suggested that PE is the most plausible substrate for PSS1. We also used pss1-1 and pss1-2 knockout mutants to reveal an important role for PSS1 in microspore maturation. Dynamic PS distribution in nuclei and different vesicular compartments was also important during microspore maturation. RESULTS Phosphatidylserine is enriched in Arabidopsis non-photosynthetic tissues and contains VLCFAs Polar lipids from leaves, roots, and flowers were examined as described in Experimental Procedures. Phosphatidylserine comprised 1.6 0.3% (leaves), 6.3 0.7% (roots), and 8.7 2.2% (flowers) of non-plastidial phospholipids (Figure 1a). Figure 1(b) and Table S1 in the Supporting Information show the fatty acid compositions of PE, PC, and PS in flowers. Phosphatidylserine contained VLCFAs such as arachidic acid (20:0), eicosenoic acid (20:1 [n-9]), behenic acid (22:0), erucic acid (22:1 [n-9]), and lignoceric acid (24:0), together accounting for 34.7% of total fatty acids in PS. In contrast, VLCFAs constituted only 1.7% and 3.4% of total fatty acids in PC and PE, respectively. PSS1 encodes a BE-PSS A PSS1 cDNA clone was amplified from an Arabidopsis rosette cDNA pool and its nucleotide sequence was determined. The deduced PSS1 polypeptide sequence showed 27.8 and 31.8% identity to Cg-PSS1 and Cg-PSS2, respectively (Figure 2). PSS1 enzyme activity was demonstrated in E. coli M15 (pREP4) cells expressing a PSS1 protein with a C-terminal 6 · His tag (PSS1-6His). Expression was induced at 23C with 1.5 mM isopropyl-b-thiogalactopyranoside (IPTG) in LB medium containing 10 mM Ser, and was confirmed by immunoblotting using anti-6 · His antibodies (Figure 3a). Escherichia coli cells that expressed PSS1-6His (PSS1-6His cells) at 23C demonstrated arrested growth within 1 h of IPTG addition (Figure 3b). At 3 h, PS amounted to 2.0 1.0% of cellular glycerolipids, whereas PE levels decreased and cardiolipin levels increased compared with control cells (Figure 3c,d; Table S3). PSS1-6His cells formed aggregates by 6 h (Figure S1). No PS accumulated in cell cultures without Ser or in control cell cultures with Ser. However, when the membrane fractions of control cells were incubated with 0.2 mM [3-14C]Ser, the label was
ª 2011 The Authors The Plant Journal ª 2011 Blackwell Publishing Ltd, The Plant Journal, (2011), 67, 648–661
650 Yasuyo Yamaoka et al. pss1-1 and pss1-2 are knockout mutants of PSS1
Figure 1. Phospholipid analysis in different Arabidopsis tissues. (a) Composition of extraplastidial phospholipids in Arabidopsis leaves, roots, and flowers (see Table S1). Although phosphatidylglycerol (PG) constitutes a significant proportion in extraplastidial lipids from non-photosynthetic tissues, only phosphatidylcholine (PC), phosphatidylethanolamine (PE), phosphatidylinositol (PI) and phosphatidylserine (PS) were compared. (b) The PC, PE, and PS fatty acid composition in flowers (see Table S2).
incorporated into PS due to endogenous CD-PSS activity. In contrast, PSS1-6His cells had a 2.7-fold higher rate of incorporation (Figure 3e), suggesting that PSS1-6His expression promotes the conversion of PE and Ser to PS in these cells (see also the Discussion). PSS1 promoter activity is prominent in developing anthers Tissue-specific expression of PSS1 was examined in transgenic Arabidopsis expressing a GUS reporter construct with the PSS1 promoter (ProPSS1-GUS). The GUS staining was prominent in anthers and microspores at floral stages 9–12 (stages defined in Bowman et al., 1991) (Figure 4a–c), in tapetum at floral stages 7–11 (Figure 4d), in mature embryos 5 days after flowering (DAF; Figure 4e), in dry seed embryos (Figure 4f), and in trichomes, leaf veins, and root vasculature (Figure 4g). In contrast, GUS staining was not detected in female gametophytes. The RT-PCR analysis showed higher PSS1 transcript levels in floral tissues compared with stems and leaves (Figure 4h).
Figure 5a summarizes the structures of the pss1-1 and pss1-2 knockout alleles. pss1-1 and pss1-2 alleles were identified by PCR (Figure 5b) and by derived cleaved amplified polymorphic sequences (dCAPS; Figure 5c), respectively. Homozygous pss1-1 and pss1-2 seedlings were isolated among the F2 population at low frequencies of 4.3% (pss1-1; n = 163) and 5.5% (pss1-2; n = 36), but were infertile dwarfs (Figure S2). The pss1-1 allele contained a T-DNA insertion in exon 11 of the PSS1 RNA, between nucleotides +1887 and +1934 of the open reading frame (Figure 5a). No transcript corresponding to wild-type PSS1 was detected in pss1-1 mutants by RT-PCR (Figure 5d). The TILLING mutant pss1-2 carried a G929A substitution at the splice acceptor site of PSS1 intron 6. In this mutant, three unusual transcripts were identified by RT-PCR, all retaining intron 6 and having the same in-frame stop codon. These results suggested that pss1-1 and pss1-2 were null mutants. pss1-1 and pss1-2 mutants contained no PS, demonstrating that PSS1 is essential for biosynthesis of PS in Arabidopsis (Figure 5e). PSS1/pss1-1 and PSS1/pss1-2 plants were indistinguishable from wild-type plants when grown at 23C under continuous or 16-h light conditions. However, PS levels decreased to 4.2 1.0 and 4.5 0.7% in PSS1/pss1-1 and PSS1/pss1-2 flowers, respectively (Figure 5f), compared with 8.7 2.2% in the wild-type flowers (Figure 1a). By contrast, PE levels in PSS1/pss1-1 and PSS1/pss1-2 flowers were 31.0 2.4 and 27.8 2.9%, respectively, which were higher than 18.3 3.4% in the wild-type flowers. At 5 DAF, self-fertilized PSS1/pss1-1 and PSS1/pss1-2 plants developed siliques with 8.7 and 9.5% embryonic lethal seeds, respectively (Figure 5g, arrowheads). The lethal seeds were mostly arrested at the octant stage (Figure 5h), whereas normal seeds developed mature embryos (Figure 5i). A large number of unfertilized ovules were also found in the same siliques. Reciprocal crossing revealed that PSS1/pss11 and PSS1/pss1-2 flowers had decreased pss1-1 and pss1-2 transmission through male gametophytes, whereas female gametophyte transmission rates of PSS1 and pss1 were insignificant (Table 1). PSS1/pss1 anthers produce a significant proportion of dead pollen grains Scanning electron microscopy (SEM) revealed that 37.6% (160/425) of pollen grains in PSS1/pss1-1 anthers were shrunken or deformed (Figure 6a,b), whereas
