Genetic defects in the CDP-choline pathway for phosphatidylcholine biosynthesis cannot be transmitted to offspring via male gametophytes owing to interruption of autophagy-like processes required for pollen germination in Arabidopsis thaliana
Momoka Wada, Chiaki Kuga, Kimie Atsuzawa, Atsuko Miyagi, Toshiki Ishikawa, Masatoshi Yamaguchi, Yuki Fujiki, Yasuko Kaneko, Maki Kawai-Yamada, Ikuo Nishida
TL;DR
Genetic defects in phosphatidylcholine biosynthesis in plants are not passed to offspring through male gametes due to failed pollen germination.
Contribution
Shows that male transmission of defective genes is blocked by disrupted autophagy-like processes in pollen.
Findings
Mutant cct1-3 cct2-3 and cct1-3 cct2-5 seedlings are not viable.
Defective alleles are not transmitted via male gametophytes but are via female gametophytes.
Pollen germination failure is linked to disrupted autophagy and abnormal cellular structures.
Abstract
Phosphatidylcholine is a major plant membrane phospholipid that contributes to the biogenesis and desaturation of membrane lipids and storage lipids. Thus, to ensure reproductive capacity, any genetic defect that affects phosphatidylcholine biosynthesis must be eliminated before fertilization. In Arabidopsis thaliana, phosphatidylcholine biosynthesis depends on CCT1 and CCT2, both encoding CTP: phosphorylcholine cytidylyltransferase. Using A. thaliana T-DNA-tagged mutants, we demonstrate that neither cct1-3 cct2-3 nor cct1-3 cct2-5 seedlings are viable. Reciprocal crosses of cct2-3/CCT2 or cct2-5/CCT2 plants in the cct1-3 background revealed that neither cct2-3 nor cct2-5 was transmitted via cct1-3 male gametophytes, although each allele was transmitted via cct1-3 female gametophytes. Although all pollen grains on a pollen quartet from qrt1-1 cct1-3 cct2-5/CCT2 plants were viable, none…
Genes, proteins, chemicals, diseases, species, mutations and cell lines named across the full text — each resolved to its canonical identifier and authoritative record.
Click any figure to enlarge with its caption.
Figure 1
Figure 2
Figure 3
Figure 4
Figure 5
Figure 6
Figure 7
Figure 8- —Saitama University
Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Videos
No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.
Taxonomy
TopicsLipid metabolism and biosynthesis · Cellular transport and secretion · Plant Reproductive Biology
Introduction
Recent studies on gametogenesis have focused on how gametophytic selection has contributed to diverse evolutionary processes (Beaudry et al. 2020). In mammals, gametogenesis refers to a process that produces a haploid cell (n) from a diploid cell (2n) through meiosis. In plants, however, meiosis is not sufficient for the completion of gametogenesis: for maturation, both microspores and megaspores generally require mitosis and differentiation, which depend on the successful progression of certain fundamental processes such as biosynthesis of cellular components, polarized vacuolation, catabolism of storage materials, vesicular trafficking and membrane biogenesis (McCormick 1993; Skinner and Sundaresan 2018). Furthermore, autophagy is required for pollen maturation (Kurusu and Kuchitsu 2017) and germination (Fujiki et al. 2007; Qin et al. 2007). Some of these processes may be essential for sporophyte development. In this context, any mutation that is unfavorable for sporophytes must be eliminated during gametogenesis and/or before fertilization. Mutations in essential genes may have similar consequences such as the lethality of male and female gametophytes, which ensures that unfavorable mutant genes will not be transmitted to offspring. However, certain mutations are transmitted at different rates via male or female gametophytes (Bolaños–Villegas et al. 2015; El-Kasmi et al. 2011; Liu et al. 2021a). For example, functional loss of the canonical α-SNAP in A. thaliana results in gametophytic lethality by arresting first mitosis during gametogenesis, but reciprocal crosses revealed that transmission of the mutation via female gametophytes is leaky whereas that via male gametophytes is strictly prohibited (Liu et al. 2021a). Lipids are essential components of biomembranes, and some of the genes responsible for lipid biosynthesis are among the essential genes (Meinke et al. 2008). However, it is not well understood how mutations in lipid genes are eliminated before fertilization or if there is any difference in the rate of transmission of mutant lipid-associated genes via male or female gametophytes.
In plant cells, phosphatidylcholine (PC) is a major membrane phospholipid that serves as not only a biosynthetic precursor to galactolipids (Kobayashi el al. 2013) that are abundant in photosynthetic membranes but also as a substrate for fatty-acid desaturases that regulate membrane fluidity (Ohlrogge and Browse 1995). PC biosynthesis is upregulated by dehydration, exposure to a high-salt environment, or cold temperature (Inatsugi et al. 2002, 2009; Liu et al. 2021b; Xiao et al. 2025a). Thus, any failure of PC biosynthesis is anticipated to negatively impact sporophyte growth and development. In this context, any mutation that affects PC biosynthesis must be eliminated from the genome before fertilization or during the early stages of embryogenesis. However, no critical studies have been reported to evaluate the importance of PC biosynthesis in plants or to examine if and how mutations in PC biosynthesis could be eliminated before fertilization.
In Arabidopsis thaliana, PC biosynthesis occurs via the CDP-choline pathway, which is regulated by the genes CCT1 (AT2G32260) and CCT2 (AT4G15130), each of which encodes CTP: phosphorylcholine cytidylyltransferase (CCT, E.C. 2.7.7.15) (Choi et al. 1997; Inatsugi et al. 2002, 2009). CCT synthesizes CDP-choline, a precursor to the polar-head group of PC, from CTP and phosphorylcholine. Phosphorylcholine is synthesized from phosphorylethanolamine by phosphoethanolamine N-methyltransferases, and loss of these enzymes abolishes PC biosynthesis and is lethal (Chen et al. 2019; Liu et al. 2019). Amino alcohol phosphotransferases (AAPT; E.C. 2.7.8.2) then transfer the phosphorylcholine residue of CDP-choline to the sn-3 position of sn-1,2-diacylglycerol to produce PC. CCT1 is activated by phosphatidic acid, the production of which is controlled by PHOSPHATIDIC ACID PHOSPHOHYROLASE1 (PHA1) and PHA2 (Craddock et al. 2015). Recently, Xiao et al. (2025a) reported that, in the cct2-5 background, CCT1 regulates PC biosynthesis under normal conditions and root development under osmotic stress, with its phosphorylation state at S187 playing an important role in modulating its enzymatic activity and functions. Apart from the CDP-choline pathway, other eukaryotes like yeast can synthesize PC via triple methylation of phosphatidylethanolamine (PE) (Carman and Zeimets 1996; Vance and Ridgeway 1988). However, the A. thaliana genome does not have a gene responsible for the first methylation step (Chen et al. 2019; Keogh et al. 2009; Ohlrogge and Browse 1995), and phospholipid N-methyltransferase (PLMT) is responsible for the second and third methylation steps of PE biosynthesis. Interestingly, overexpression of PLMT increases PC content in planta and the triacylglycerol (TAG) level in seeds (Tan and Nakamura 2022). These results are consistent with a view that PE cannot be converted to PC and that PLMT-dependent PC biosynthesis depends on the CDP-choline pathway (Chen et al. 2019). However, it remains to be clarified if CCT1 can synthesize CDP-N-mono- and CDP-N-di-methylethanolamine so that AAPT can synthesize mono- and di-methyl-PE, respectively, both of which are substrates of PLMT for PC biosynthesis. Nonetheless, the PE methylation pathway towards PC does not complement the deficiency of the CDP-choline pathway towards PC in A. thaliana.
Several studies have revealed the roles of phospholipids in reproductive development of A. thaliana. For S-adenosyl-l-methionine: phosphoethanolamine N-methyltransferase (EC 2.1.1.103), which is required for phosphorylcholine biosynthesis, the temperature-sensitive mutant t365 exhibits male sterility (Mou et al. 2002). For CTP: phosphorylethanolamine cytidylyltransferase (PECT1; EC 2.7.7.14), which is required for PE biosynthesis, the null mutant pect1-6 is unable to proceed to embryo development beyond the early globular stage (Mizoi et al. 2006), whereas the pect1-4 mutant that retains 26% of the PECT1 activity causes a delay in anther development. Thus, downregulation of the CDP-ethanolamine pathway towards PE biosynthesis causes a delay in anther development, but disruption of the CDP-ethanolamine pathway does not completely abolish the development of male and female gametophytes. AAPT1 and AAPT2 are required for the final step in PC and PE biosynthesis. Disruption of both AAPT1 and AAPT2 abolishes the establishment of aapt1 aapt2 seeds, and 50% of pollen grains from aapt1 aapt2/AAPT2 or aapt1/AAPT1 aapt2 plants having the qrt1 background die, suggesting that disruption of both PC and PE biosynthesis induces the lethality of male gametophytes (Liu et al. 2015). Phosphatidic acid is the common precursor to glycerolipids, and glycerol-3-phosphate acyltransferase (GPAT; EC 2.3.2.15) catalyzes the first step of phosphatidic acid biosynthesis. GPAT9 is responsible for the production of phosphatidic acid that is necessary for cytoplasmic glycerolipids, and disruption of GPAT9 abolishes the establishment of gpat9 seeds (Shockey et al. 2016). Reciprocal crosses of gpat9-2/GPAT9 plants demonstrated that gpat9-2 is not transmissible via male gametophytes, and gpat9-2 pollen grains are unable to germinate in vitro in the qrt1 background. In the qrt1 background, gpat9-2 pollen grains are smaller than GPAT9 pollen grains (Shockey et al. 2016). Because disruption of phosphatidic acid biosynthesis inhibits triacylglycerol (TAG) biosynthesis, the inhibition of pollen germination is thought to be caused by TAG shortage (Shockey et al. 2016). Choline kinase is required for the biosynthesis of phosphorylcholine, the substrate of CCT, and the choline/ethanolamine kinase CEK4 is involved in embryo development (Lin et al. 2015). Finally, phosphatidylserine is involved in vesicular trafficking and required for normal progression of cell plate formation (Yamaoka et al. 2021). Disruption of phosphatidylserine biosynthesis affects embryo development and causes pollen lethality, but even the null mutant pss1 can escape pollen lethality and establish pss1 seedlings (Yamaoka et al. 2011).
Using T-DNA-tagged mutants of A. thaliana, designated cct1-3, cct2-3 and cct2-5, we investigated whether the double mutation cct1-3 cct2-3 or cct1-3 cct2-5 can be carried over to the F3 progeny. Our results show that neither cct1-3 cct2-3 nor cct1-3 cct2-5 seedlings are present in F3 progeny. We then investigated the mechanism by which the cct1-3 cct2-3 or cct1-3 cct2-5 mutation can be eliminated from the F3 seed population by conducting reciprocal crossing, Alexander’s pollen viability test, in vitro pollen germination test, and observation of ultrathin sections of germinating pollen grains by transmission electron microscopy (TEM). We conclude that cct1-3 cct2-3 and cct1-3 cct2-5 are strictly eliminated during pollen germination owing to PC shortage; this elimination inhibits the progression of autophagy required for pollen germination. We also showed that cct1-3 cct2-3 and cct1-3 cct2-5 are partly transmissible via female gametophytes. Our results suggest that plants utilize distinct transmission strategies between male and female gametophytes, at least in regards to mutant genes in PC biosynthesis: transmission of the mutant genes via male gametophytes is strictly prohibited by inhibition of pollen germination so that a wide dispersal of deleterious mutation among the progeny is prevented, whereas transmission of the same mutant genes via female gametophytes is partly permissive so that the background genome of the escaped mutant ovules can be rescued by fertilization with normal pollen.
Materials and methods
Plant materials
Arabidopsis thaliana (L.) Heynh. ecotype Columbia was obtained from Lehle seeds (Round Rock, TX, U.S.A., http://www.arabidopsis.com/). cct1-3 (GK-349C03-016244) was obtained from GABI-Kat (Genomanalyse im biologischen System Pflanze - Kölner Arabidopsis T-DNA lines) via the Arabidopsis Biological Resource Center (ABRC). cct2-3 (SK34804, CS1012739) was obtained from ABRC, whereas cct2-5 (SALK_200207, originally distributed by ABRC) was a gift from Peter Moffett (Fig. 1).
Fig. 1. Structures of the genes cct1-3, cct2-3 and cct2-5. cct1-3 contains a T-DNA insertion within exon 4, cct2-3 contains an insertion in intron 4, and cct2-5 contains an insertion in intron 7
Plant growth
A. thaliana seeds were sown on soil (Supermix A, Sakata, Kanagawa, Japan) packed in a stainless steel pan (185 mm W × 140 mm L × 30 mm D) or a plastic pot (60 mmϕ × 60 mm H). After a 2-day incubation at 4 ℃ under 100% relative humidity, pans (or pots) were incubated in a growth room regulated at 23 ℃ under a 16-h light/8-h dark photo regime at a photon flux density of 110 µmol m^–2^s^–1^.
A. thaliana transformation and selection for the transformants
A. thaliana was transformed by the floral dip method (Clough and Bent 1998) using Agrobacterium tumefaciens strain GV3101. Seeds were sterilized by immersion in 70% ethanol for 1 min then in a sterilizing mixture containing 5% (v/v) sodium hypochlorite and 0.02% (w/v) Triton X-100 for 5 min twice. After washing with sterilized water, seeds were aseptically sown on 1/2 MS plates containing a half-strength Murashige-Skoog salts (Company, City, Japan), 1× Gamborg B5 vitamins (Gamborg et al. 1968), 0.5% MES (pH 5.7), and 0.7% agar for plant growth (Company, City, Japan).
Genotyping by PCR
One cotyledon or an equivalent size of leaf was homogenized in 400 µl DNA extraction buffer containing 0.2 M Tris-HCl (pH 9.0), 0.4 M LiCl, 25 mM EDTA･2Na and 1% SDS in a round-bottomed 1.5-ml microtube (As one, Osaka, Japan) using a homogenizer pestle (CT1.5 3-325-0268, Kenis, Osaka, Japan). After centrifugation at 16,000×g for 1 min, a 300-µl portion of the supernatant was recovered for DNA precipitation with an equivalent volume of 2-propanol. Each DNA pellet was recovered by centrifugation twice at 16,000×g for 5 min, carefully eliminating the residual supernatant. After a 15-min evacuation under vacuum, DNA was dissolved in 100 µl of buffer (10 mM Tris-HCl pH 8.0, 1 mM EDTA･2Na) to make a template DNA solution.
PCR for genotyping was conducted using a Thermal cycler (2720, Applied Biosystems, Tokyo, Japan) under reaction conditions summarized in Table S1. Each reaction mixture contained 5 µl 2×Quick Taq^®^ HS DyeMix (TOYOBO, Tokyo, Japan), 1 µl template DNA solution, and 1 µl each of 10 pmol µl^–1^ primer solutions (Table S1).
Construction of cct mutants in the qrt1-1 background
qrt1-1 plants form inseparable pollen tetrads and hence are useful for meiotic segregation analysis of mutant phenotypes. To construct qrt1-1 cct1-3/CCT1 cct2-3/CCT2 and qrt1-1 cct1-3/CCT1 cct2-5/CCT2 plants, qrt1-1 plants (♂) were first crossed with cct1-3 cct2-3/CCT2 and cct1-3 cct2-5/CCT2 plants (♀) to obtain qrt1-1/QRT1 cct1-3/CCT1 cct2-3/CCT2 and qrt1-1/QRT1 cct1-3/CCT1 cct2-5/CCT2 plants, respectively, from which seedlings with the cct1-3/CCT1 cct2-3/CCT2 and cct1-3/CCT1 cct2-5/CCT2 genotypes were identified in the respective offspring by PCR. Finally, seedlings with qrt1-1 were identified under a scanning electron microscope (TM-1000, HITACHI, Tokyo, Japan).
Reciprocal crosses
Reciprocal crosses of cct2-3/CCT and cct2-5/CCT2 in the cct1-3 background were conducted by crossing cct1-3 vs. cct1-3 cct2-3/CCT2 and cct1-3 vs. cct1-3 cct2-5/CCT2, respectively. Segregation ratios of mutant alleles in the F1 offspring were determined by PCR.
Alexander’s staining to assess pollen survival
Pollen viability was determined by Alexander staining (Alexander 1969; Atlagić et al. 2012). On the day of flowering, open flowers were sampled in a 1.5-ml microtube and immediately submerged in Alexander’s solution. After 15 min, samples were washed with water, and anther and pollen grains were observed under a fluorescence microscope (LEICA DMR, Leica Microsystems, Tokyo, Japan).
Pollen germination in vitro
Pollen collected on the day of flowering was spread over a 1.5% agar plate made up with pollen germination medium containing 0.01% boric acid, 5 mM CaCl_2_, 5 mM KCl, 1 mM MgSO_4_, 10% sucrose, 10 µM brassinosteroid (a gift from Dr. Miho Ikeda at Saitama University) and 0.5% ethanol (Vogler et al. 2014) and incubated in a growth cabinet (BiOTRON LPH-240, NK Systems Limited, Tokyo, Japan) regulated at 23 ℃ in darkness. Pollen germination and pollen-tube elongation were observed under a fluorescence microscope (LEICA DMR, Leica Microsystems, Tokyo, Japan).
Observation of pollen morphology by scanning electron microscopy
On the day of anthesis, anthers were harvested, and pollen was spread over a glass slide. The morphology of the pollen surface was observed under a scanning electron microscope. Briefly, a specimen was set on a Hitachi M4 Aluminum Specimen Mount (Prod No. 16324) and observed with a scanning electron microscope under partial vacuum without coating in a scanning electron microscope (Hitachi TM-1000, Tokyo, Japan), according to the manufacturer’s manual (https://arf.berkeley.edu/files/attachments/equipment/SEM_Intructions_-_ARF_Hitachi_TM_1000.pdf).
DAPI staining for evaluation of mitosis in pollen
Four to five opened flowers sampled on the day of flowering were immersed in DAPI staining medium containing 0.4 µg ml^–1^ DAPI, 0.1% (w/v) Triton X-100, 1 mM EDTA･2Na and 100 mM sodium phosphate pH 7 (Schnedl et al. 1977). Pollen grains were collected by centrifugation and observed under a fluorescence microscope (LEICA DMR, Leica Microsystems, Tokyo, Japan).
Transmission electron microscopy
Samples for TEM were made according to Kaneko (2007). Pollen collected on the day of anthesis was spread over an agar plate made up with pollen germination medium (Vogler et al. 2014). After incubation for 30 min, pollen was collected by centrifugation and suspended in 0.05 M potassium phosphate buffer (pH 6.8) containing 2% glutaraldehyde and incubated for 4 h at room temperature and then overnight at 4 ℃ (prefixation). After washing six times with 0.05 M potassium phosphate buffer (pH 6.8), samples were incubated in 0.05 M potassium phosphate buffer (pH 6.8) containing 2% OsO_4_ for 2 h at ambient temperature (postfixation). Then, samples were washed once with 0.05 M potassium phosphate buffer (pH 6.8) and then subjected to sequential dehydration for 10 min with 10, 30, 50, 70, 85, 95 and 100% acetone. For this purpose, 100% acetone was prepared by storing over anhydrous sodium sulfite. After two additional washes with 100% acetone, samples were subjected to sequential incubations with 50, 75 and 100% Spurr’s resin solution and incubated overnight with 100% resin solution. After replacing the 100% resin solution with fresh resin solution, samples were incubated for 1 h, then placed in 1.5-ml microtubes and incubated at 70 ℃ for 8 h for resin solidification. Ultrathin sections were made and after staining with uranyl acetate and lead citrate subjected to TEM (Hitachi H-7500, at 80 kV) in the Comprehensive Analysis Center for Science at Saitama University.
Cloning of the pollen-specific promoter ACA9pro
A gene fragment with restriction enzyme tags for subcloning the pollen-specific promoter ACA9pro (Schiøtt et al. 2004) was amplified from A. thaliana genomic DNA by PCR (Table S1) using a KOD-Plus-Neo polymerase (TOYOBO). PCR products were subjected to 1% agarose gel electrophoresis, and a band for ACA9pro was purified using a gel extraction kit (Wizard SV Gel and PCR Clean-up System, Promega; or Gel/PCR Extraction Kit, FastGene). Then, purified ACA9pro was digested with XbaI (TaKaRa) and HindIII (TaKaRa), and digested samples were subjected to 1% agarose gel electrophoresis and bands were then purified as described above. The resultant XbaI–HindIII fragment of ACA9pro was subcloned into the XbaI–HindIII sites of pBluescriptⅡ SK(+) using a 2×Mighty Mix ligation kit (TaKaRa). The resultant plasmid was designated pBluescriptⅡ SK(+) ACA9pro.
Construction of CCT1 expression constructs under the control of ACApro9
Using the pBluescriptⅡ SK(+) ACA9pro as a template, the XbaI–HindIII ACA9pro fragment was amplified by PCR and then subcloned into the XbaI–HindIII sites of pPZP221 35Spro: CCT1cDNA-nosT (100 µg/ml spectinomycin) to create pPZP221 ACA9pro: CCT1cDNA-nosT. However, the resultant plasmid unexpectedly contained inverted ACA9pro (invACA9pro). Thus, it was designated pPZP221 invACApro9:CCT1cDNA-nosT (Fig. S1).
Transformation of Agrobacterium tumefaciens
A. tumefaciens GC3101 carrying antibiotic resistance for chloramphenicol, gentamycin, rifampicin and streptomycin was transformed by electroporation using 20 µl competent cells, which were prepared according to the method by Sean Weise (https://bmb.natsci.msu.edu/sites/_bmb/assets/File/Sharkey_lab/Agrobacterium%20Transformation%20and%20Competent%20Cell%20Preparation.pdf). Transformants were identified by colony PCR (Table S1).
Isolation of cct1-3 cct2-5/CCT2 plant lines expressing invACA9pro:CCT1cDNA-nosT
Primary shoots (~ 10 cm long) were cut off from cct1-3 cct2-5/CCT2 plants to promote secondary shoot regeneration. When the secondary shoots were ~ 15 cm long, all siliques and opened flowers were removed, and the remaining floral tips and rosette base were inoculated with A. tumefaciens culture according to the floral dip method. T1 seeds were selected on 1/2 MS agar plates containing 50 µg ml^–1^ gentamycin (FUJIFILM Wako Chemicals, Osaka, Japan) and 10 µg ml^–1^ meropenem trihydrate (FUJIFILM Wako Chemicals, Osaka, Japan), and cct1-3 cct2-5/CCT2 plant lines expressing invACA9pro: CCT1cDNA-nosT were identified by genotyping for cct1-3 and cct2-5. To demonstrate co-transmission of cct2-5 with the transgene invACA9pro: CCT1cDNA-nosT via male gametophytes, one of the lines, designated M21, expressing cct1-3 cct2-5/CCT2 invACA9pro: CCT1cDNA-nosT (+/–) was immediately used as a pollen source for pollination over the stigma of emasculated cct1-3 flowers.
Results
Assessment of segregation in the mutant alleles cct1-3, cct2-3 and cct2-5 in the F2 progeny
cct1-3 and cct2-5 plants were crossed in an attempt to isolate cct1-3 cct2-5 double mutants in the F2 progeny. However, no cct1-3 cct2-5 double mutant was obtained. Accordingly, offspring of a cct1-3 cct2-5/CCT2 F2 plant was further analyzed (Table 1). Again, no cct1-3 cct2-5 double mutant was obtained in the F3 progeny, suggesting that cct1-3 cct2-5 was eliminated during gametogenesis, fertilization, or early embryogenesis/seed development. Table 1 also shows that a segregation ratio was 1:0.77 for the cross between CCT2/CCT2 and cct2-5/CCT2 in the cct1-3 background, a ratio that differed significantly from the expected Mendelian ratio of 1:2 (p = 1.48 × 10^–4^ < 0.05). This result suggested that the seed yield of cct2-5/CCT2 was decreased by 62% in the cct1-3 background.
Table 1. Segregation analysis of offspring of a cct1-3 cct2-5/CCT2 plantNumber of offspringχ^2^ test for 1:2^a^Genotype cct1-3 CCT2
cct1-3 cct2-5/CCT2
cct1-3 cct2-5 χ^2^ p Theoretical ratio121Observed number79610Observed ratio1^b^0.77^b^014.401.48 × 10^–4 c^^a^ The χ^2^ test was conducted to test the null hypothesis that cct1-3 and cct1-3 cct2-5/CCT2 plants segregate from each other in the predicted 1:2 Mendelian ratio.^b^ The segregation ratio between CCT2 and cct2-5/CCT2 plants in the cct1-3 background was 1:0.77, which differed significantly from the predicted 1:2 Mendelian ratio, suggesting that the seed yield of cct2-5/CCT2 was decreased by 62% in the cct1-3 background.^c^ The p value is small enough (p < 0.05) to reject the null hypothesis.
Similar results were obtained from crossing experiments between cct1-3 and cct2-3 plants (Table S2). No cct1-3 cct2-3 double mutant was obtained from a cct1-3 cct2-3/CCT2 plant in the F3 progeny, and the segregation ratio of CCT2/CCT2 and cct2-3/CCT2 in the cct1-3 background was 1:0.77, which differed significantly from the expected 1:2 Mendelian ratio (p = 1.02 × 10^–4^ < 0.05). Because the T-DNA-tagged mutant cct2-3 carried a drug-resistance gene against ammonium glufosinate (BASTA), the unexpected segregation ratio of 1:0.77 between CCT/CCT2 and cct2-3/CCT2 plants in the cct1-3 background was also confirmed by a survival test on 1/2 MS agar plates containing 40 nmol ml^–1^ BASTA; among 94 seeds harvested from a cct1-3 cct2-3/CCT2 plant, 53 were dead (cct1-3 CCT2) and 41 survived (cct1-3 cct2-3/CCT2) (Fig. S2), again with a segregation ratio of 1:0.77.
We also conducted a segregation test for offspring of a cct2-3 cct1-3/CCT1 plant (Table 2). In accordance with results presented in Table S2, no cct1-3 cct2-3 double mutant was obtained from the cct2-3 cct1-3/CCT1 plant. However, the segregation ratio between CCT1/CCT1 and cct1-3/CCT1 in the cct2-3 background was 1:1.2, which did not differ significantly from the expected 1:2 Mendelian ratio (p = 0.055). These results suggested that cct2-3 cct1-3/CCT1 plants had greater sporophytic potency of seed nursery than cct1-3 cct2-3/CCT2 plants and that cct1-3/CCT1 retained more CCT activity than cct2-3/CCT2.
Table 2. Segregation analysis of offspring of a cct2-3 cct1-3/CCT1 plantNumber of offspringχ^2^ test for 1:2^a^Female CCT1 cct2-3
cct2-3 cct1-3/CCT1
cct1-3 cct2-3 χ^2^ p Theoretical ratio121Observed number29350Observed ratio11.205.820.055^b^^a^ The χ^2^ test was conducted to test the null hypothesis that cct1-3 and cct1-3 cct2-5/CCT2 plants segregate from each other in the predicted 1:2 Mendelian ratio.^b^ The p value is large enough (p > 0.05) to mistakenly reject the null hypothesis.
Reciprocal crossing of cct2-5/CCT2 and cct2-3/CCT2 plants in the cct1-3 background
To examine if cct1-3 cct2-5 is transmissible to offspring via male or female gametophytes, reciprocal crossing of cct2-5/CCT2 plants was conducted in the cct1-3 background (Table 3). When the stigma of emasculated cct1-3 flowers was pollinated with an anther of cct1-3 cct2-5/CCT2 flowers, only cct1-3 but no cct1-3 cct2-5/CCT2 seeds were recovered, indicating that cct2-5 was not transmissible via the male gametophyte in the cct1-3 background. In contrast, when the stigma of emasculated cct1-3 cct2-5/CCT2 flowers was pollinated with an anther of cct1-3 flowers, cct1-3 cct2-5/CCT2 seeds were recovered as well as cct1-3 seeds, and the segregation ratio between CCT2/CCT2 and cct2-5/CCT2 plants in the cct1-3 background (1:0.46) differed significantly from the predicted 1:1 Mendelian ratio (p = 2.39 × 10^–4^ < 0.05), suggesting that cct2-5 is partly transmissible via the female gametophyte in the cct1-3 background. Furthermore, a survival rate of 45.5% (100 × 30/66) was calculated for cct1-3 cct2-5 ovules, demonstrating that cct1-3 cct2-5 ovules were both permissive and fertile.
Table 3. Summary of reciprocal crossing of cct2-5/CCT2 plants in the cct1-3 backgroundReciprocal crossNumber of F1 plantsχ^2^ test for 1:1^a^FemaleMale cct1-3
CCT2/CCT2
cct1-3 cct2-5/CCT2χ^2^ p
cct1-3
CCT2/CCT2
cct1-3 cct2-5/CCT231031.002.58 × 10^–8b^ cct1-3 cct2-5/CCT2 cct1-3
CCT2/CCT2 663013.502.39 × 10^–4b^In this table, cct2-5 is boldfaced to emphasize transmission via gametophytes.^a^χ^2^ test was conducted to test the null hypothesis that CCT2/CCT2 and cct2-5/CCT2 plants segregate from each other in the predicted 1:1 Mendelian ratio.^b^The p values are small enough (p < 0.05) to reject the null hypothesis.
Similar results were obtained from reciprocal crossing experiments with cct2-3/CCT2 plants in the cct1-3 background (Table S3), in that no cct2-3 allele was transmitted via the male gametophyte in the cct1-3 background whereas it was partly transmitted via the female gametophyte.
Alexander’s test for pollen viability
We next examined whether cct1-3 cct2-3 and cct1-3 cct2-5 pollen could be established during pollen maturation. To assess the phenotype, we introduced qrt1-1 in the background, which constitutively produces four inseparable pollen grains called a pollen tetrad or a pollen quartet.
We first examined pollen viability by Alexander’s test. Alexander’s reagent (or acidic fuchsin) stains dead pollen grains green and live grains purple-red (Atlagić et al. 2012). Pollen grains in the anthers of qrt1-1 cct2-5 cct1-3/CCT1 (Fig. 2d) and qrt1-1 cct1-3 cct2-5/CCT2 (Fig. 2e) plants were all stained purple-red (viable) as were those in the anthers of qrt1-1 (Fig. 2a), qrt1-1 cct2-5 (Fig. 2b), and qrt1-1 cct1-3 (Fig. 2c) plants, and each pollen grain that was a part of a single pollen quartet was equally stained purple-red (Fig. 2a–e, red arrows). These data suggested that cct1-3 cct2-5 pollen grains would be dead after maturation.
Fig. 2. Alexander’s test for pollen viability. Pollen tetrads in anthers from qrt1-1 (a), qrt1-1 cct2-5 (b), qrt1-1 cct1-3 (c), qrt1-1 cct2-5 cct1-3/CCT1 (d) and qrt1-1 cct1-3 cct2-5/CCT2 (e) plants were stained with Alexander’s reagent. Arrows indicate pollen tetrads with four viable microspores. Bars = 1.0 mm
Examination of pollen grain shrinkage
Observation of pollen quartets under a scanning electron microscope revealed that shrunk pollen grains were included in some pollen quartets from qrt1-1 cct1-3 cct2-5/CCT2 and qrt1-1 cct2-5 cct1-3/CCT1 plants, but few were included in the quartets from qrt1-1, qrt1-1 cct1-3, qrt1-1 cct2-5 plants (Fig. S3a). Pollen grain shrinkage was carefully examined during in vitro pollen germination experiments, using qrt1-1, qrt1-1 cct1-3, qrt1-1 cct2-5, qrt1-1 cct1-3 cct2-5/CCT2 and qrt1-1 cct2-5 cct1-3/CCT1 plants (Fig. S3b). On the day of anthesis, harvested anthers were rubbed over the surface of agar plates containing pollen germination medium. After incubation at 23 ℃ for 15 min, the number of quartets having different numbers of shrunken pollen grains was counted under a fluorescence microscope. In qrt1-1, qrt1-1 cct2-5 and qrt1-1 cct1-3 plants, 88–97% of pollen quartets carried no shrunken pollen grains, and the remaining proportions of the pollen quartets contained only one or two shrunken pollen grains (Fig. S3c). Thus, the proportions of shrunken pollen grains in qrt1-1, qrt1-1 cct2-5 and qrt1-1 cct1-3 plants were calculated to be 2.6, 4.7, and 0.8%, respectively. By contrast, 83% of quartets from qrt1-1 cct2-5 cct1-3/CCT1 plants and 71% from qrt1-1 cct1-3 cct2-5/CCT2 plants contained no shrunken pollen grains (Fig. S3c). In these plants, no pollen quartets contained more than two shrunken pollen grains, suggesting that qrt1-1 cct2-5 CCT1 and qrt1-1 cct1-3 CCT2 pollen grains were almost intact in these plants. Thus, the proportions of shrunken pollen grains in qrt1-1 cct2-5 cct1-3/CCT1 and qrt1-1 cct1-3 cct2-5/CCT2 plants were calculated to be 6.5 and 13.3%, respectively, and the shrinkage rates of qrt1-1 cct1-3 cct2-5 pollen grains from qrt1-1 cct2-5 cct1-3/CCT1 and qrt1-1 cct1-3 cct2-5/CCT2 plants were calculated to be no more than 13.0 and 26.6%, respectively. Notably, the survival rate of qrt1-1 cct1-3 cct2-5 pollen grains was higher in the qrt1-1 cct2-5 cct1-3/CCT1 plants than in the qrt1-1 cct1-3 cct2-5/CCT2 plants, possibly reflecting higher CCT activity in the parental tissues of the former plants than the latter.
Evaluation of pollen maturation by DAPI staining
In A. thaliana, after pollen meiosis, the individual microspore initiates vacuole formation and then divides asymmetrically to produce a vegetative cell and a generative cell (pollen mitosis I). The generative cell then divides into two identical sperm cells (pollen mitosis II) and the vacuoles vanish, resulting in tricellular mature pollen grains (Bolaños-Villegas et al. 2010).
To examine if the two rounds of pollen mitosis proceeded normally to generate mature pollen, the nuclei were observed after DAPI staining. On the day of flowering, flowers were taken from qrt1-1, qrt1-1 cct1-3, qrt1-1 cct2-5, qrt1-1 cct1-3 cct2-5/CCT2, and qrt1-1 cct2-5 cct1-3/CCT1 plants and stained with DAPI (Fig. 3). In all plants examined, each pollen microspore on a pollen quartet contained three bright spots representing one vegetative nucleus (a less bright spot) and two generative nuclei (GN; two bright spots) (Fig. 3), suggesting that the two rounds of pollen mitosis had been completed. Thus, we concluded that cct1-3 cct2-5 microspores on a pollen quartet had matured completely.
Fig. 3DAPI staining of pollen grains. Pollen tetrads from qrt1-1(a), qrt1-1 cct2-5 (b), qrt1-1 cct1-3 (c), qrt1-1 cct2-5 cct1-3/CCT1 (d) and qrt1-1 cct1-3 cct2-5/CCT2 (e) plants were stained. Arrows indicate microspores with two bright fluorescing spots as well as one faint spot, implying completion of the two rounds of pollen mitosis. Bars = 25 μm
Examination of pollen germination in vitro
Pollen germination was examined on the same pollen germination medium as described above (Fig. 4a). After incubation at 23 ℃ for 4 h, the number of pollen quartets having different numbers of germinated pollen grains was counted under a fluorescence microscope. Among qrt1-1, qrt1-1 cct1-3 and qrt1-1 cct2-5 plants, almost 40% of pollen quartets had germinated in all four pollen grains (Fig. 4b). Thus, the single mutation cct1-3 or cct2-5 did not affect germination rates of pollen quartets from qrt1-1 cct1-3 and qrt1-1 cct2-5 plants. In pollen quartets from qrt1-1 cct2-5 cct1-3/CCT1 and qrt1-1 cct1-3 cct2-5/CCT2 plants, however, no pollen quartets contained more than two germinated microspores, suggesting that, in the pollen quartets from qrt1-1 cct1-3/CCT1 cct2-5 and qrt1-1 cct1-3 cct2-5/CCT2 plants, none of cct1-3 cct2-5 microspores had germinated in vitro.
Fig. 4. Observation of pollen germination in vitro. Anthers from qrt1-1, qrt1-1 cct2-5, qrt1-1 cct1-3, qrt1-1 cct2-5 cct1-3/CCT1 and qrt1-1 cct1-3 cct2-5/CCT2 plants were rubbed over the surface of agar plates containing pollen germination medium. After a 4-h incubation at 23 ℃, pollen quartets were observed under a fluorescence microscope. a Arrowheads denote germinated pollen grains. Bar = 10 μm. b The proportion of pollen quartets carrying different numbers of germinated pollen grains. In pollen quartets from qrt1-1 cct1-3 cct2-5/CCT2 and qrt1-1 cct2-5 cct1-3/CCT1 plants, no pollen quartets carried more than two germinated pollen grains
Ultrafine structures of pollen cells from wild-type, cct2-5 and cct1-3 plants
The germination rates of pollen grains from wild-type, cct2-5 and cct1-3 plants in the qrt1-1 background were 77.4, 72.3, and 66.0%, respectively. Observation of ultrathin sections of wild-type, cct2-5, and cct1-3 pollen cells at relatively low magnification showed that the preservation of normal appearance of the cytoplasm coincided with successful pollen germination (Fig. S9). Thus, to figure out which ultrafine structures could be important for the progression of normal pollen germination, we compared the ultrathin sections of pollen cells at higher magnifications.
On the day of anthesis, ultrathin sections for TEM were prepared from pollen grains from wild-type, cct2-5, and cct1-3 plants after 30 min of incubation in pollen germination medium. Results derived from images taken for all the sections of wild-type, cct2-5, and cct1-3 pollen cells are summarized in Figs. S4, S5, and S6, respectively, and the details of ultrafine structures were described in the supplementary text for the figures. We herein show the images taken for the sections of representative wild-type, cct2-5, and cct1-3 pollen cells in Fig. 5a, b, and c, respectively, and enlarged images of the respective pollen cells in Fig. 6a, b, and c.
Fig. 5TEM analysis of ultrathin sections of pollen grain cells from A. thaliana wild-type, cct2-5, cct1-3, cct2-5 cct1-3/CCT1, and cct1-3 cct2-5/CCT2 plants. Panels show photos of representative pollen grain cells from wild-type (a), cct2-5 (b), cct1-3 (c), cct2-5 cct1-3/CCT1 (d), cct1-3 cct2-5/CCT2 (e) plants. Panels a and b are the same as in Figs. S4p and S5t, respectively, and panels c, d, and e are modified from Figs. S6a, S7a, and S8a, respectively. Open arrowheads indicate cell septa for GN. GN, generative nuclei; LB, lipid body; M, mitochondrion; P, plastid; rE, rough endoplasmic reticulum; SGS, small granular structure; SGS*, swollen SGS; vSV, very small vacuolar structure. Bars = 1 μm
Fig. 6TEM analysis of typical ultrafine structures in the cytoplasm of pollen grain cells from A. thaliana wild-type, cct2-5, cct1-3, cct2-5 cct1-3/CCT1, cct1-3 cct2-5/CCT2 plants. Panels a, b, c, d, and e show enlarged photos of Fig. 5a, b, c, d, and e, respectively. LB, lipid body; M, mitochondrion; P, plastid; PVB, prevacuolar body; rE, rough endoplasmic reticulum; SGS, small granular structure; SGS*, swollen SGS; SVB, small vacuolar body; and vSV, very small vacuolar structure. Stars represent the site where the rough endoplasmic reticulum entraps other cellular components. Bars = 1 μm
The following features were collectively confirmed by viewing images taken for the sections of the wild-type pollen cells: (1) the wild-type pollen cells contained one vegetative nucleus and two GN as well as cellular septa enclosing the two GN [Figs. 5a (S4p), S4a (S4b), and S4g], consistent with the results of DAPI staining (Fig. 3a); (2) lipid bodies (LBs) were seen as dark, electron-dense round structures whose surface looked rather transparent compared with the inner bodies and, hence, may have had a membranous boundary (Figs. 6a, S4c, h, m, and q); (3) the cytoplasm contained numerous moderately electron-dense small granular structures (SGSs), which sometimes had a small, semi-transparent domain in the inner area (Figs. 6a, S4d, i, n, and r); (4) the cytoplasm had mitochondria and plastids (Figs. 5a, S4a, g, l, and p), and plastids often existed in contact with rough endoplasmic reticulum (rER; marked with rE in Figs. 5a, S4e, and s); (5) SGSs were sometimes entrapped by rER [Figs. 7a (S4l), g, j, and k]; (6) there were many, very small vacuolar structures of < 1 μm in diameter (vSVs), some of which contained structures similar to SGSs or LBs (Fig. 6a, black arrows; Figs. 7a, S4f, j, and k); (7) vSVs appeared to fuse with each other to generate a larger vacuolar body of 1 ~ 3 μm in diameter [Figs. 7a (S4l), S4g, and p], which we designated herein the small vacuolar body (SVB); (8) SVBs also ingested LBs for subsequent hydrolysis (series of photos shown in Fig. S4o), while fusing to each other created a larger vacuolar body (Fig. S4u); and (9) rER ran between vSVs or vacuoles to form a network (Figs. 6a and 7a, stars; Fig. S4f, k, and t). These results were consistent with a canonical view that pollen germination requires the development of vacuole to digest LBs and SGSs and control turgor pressure for pollen-tube budding and elongation.
Fig. 7. Developmental processes of SVBs from vSVs and PVBs are delayed in putative cct1-3 cct2-5 pollen cells compared with those in wild-type and cct2-5 pollen cells. Panels a and b show enlarged photos of wild-type (Fig. S4l) and cct2-5 (Fig. S5l) pollen cells, and panel c shows an enlarged photo of a putative cct1-3 cct2-5 pollen cell (Fig. S8a) from cct1-3 cct2-5/CCT2 plants. LB, lipid body; P, plastid; PVB, prevacuolar body; rE, rough endoplasmic reticulum; SGS, small granular structure; SGS*, swollen SGS; SVB, small vacuolar body; and vSV, very small vacuolar structure. The star in b represents the site where the rough endoplasmic reticulum entraps various cellular components including vSVs and SVBs, which are ingesting SGSs and LBs. Bars = 1 μm
The ultrafine structures of wild-type pollen cells, as described above, were also collectively assessed in images of cct2-5 pollen cells from a cct2-5 plant (Fig. S5). However, the following points should be noted. First, although LBs were also electron-dense and surrounded by a semi-transparent boundary (Figs. 6b, S5c, h, m, q, and u), they looked smaller in cct2-5 pollen cells than in the wild-type cells. Furthermore, the number of LBs in some sections was much less in cct2-5 pollen cells than that in the wild-type cells. On the other hand, the image of SGSs looked moderately electron-dense or “solid”, as seen in the wild-type cells. Because the germination rate of cct2-5 pollen cells (72.3%) was slightly lower than that of the wild-type pollen cells (77.4%), the decreased number and size of LBs in cct2-5 pollen cells compared with the wild-type pollen cells coincided with a slight decrease in pollen germination rate.
The ultrafine structures of cct2-5 pollen cells, as described above, were also collectively assessed in images of cct1-3 pollen cells from a cct1-3 plant, except some SGSs retained a moderately electron-dense or “solid” appearance [Figs. 5c (6c, S6a)] and sometimes had a small, semi-transparent domain (Figs. 6c, S6d, i, and n), and a few SGSs in some cells looked larger and “swollen” (Fig. 6c, SGS*; Fig. S6r). Because the germination rate of cct1-3 pollen grains from a cct1-3 plant (66%) was further decreased from that of cct2-5 pollen grains from a cct2-5 plant (72.3%), the occurrence of swollen SGSs in cct1-3 pollen cells coincided with the decrease in germination rate of cct1-3 pollen grains compared with that of cct2-5 pollen grains.
In summary, after 30 min of pollen incubation in vitro, smaller and fewer LBs were found in cct2-5 and cct1-3 pollen cells than in wild-type pollen cells, and these changes probably caused a ~ 5% decrease in pollen germination rate. The presence of moderately electron-dense or “solid” SGSs in the cytoplasm appeared to be a hallmark of pollen cells competent for germination; moreover, with increasing severity of the cct mutation, i.e., in the order of wild type < cct2-5 < cct1-3, SGSs deformed to yield SGS*. The occurrence of SGS* coincided with an additional ~ 5% decrease of pollen germination rate.
Evaluating abnormalities among cct1-3 cct2-5 pollen grains
cct1-3 cct2-5 pollen cells, especially those from cct1-3 cct2-5/CCT2 plants, were anticipated to have the most extreme phenotype considering both gametophytic and sporophytic genetic defects. After 30 min of incubation in pollen germination medium, the proportion of abnormal pollen cells from a cct1-3 cct2-5/CCT2 plant was 37.9% and 43.8%, respectively, as estimated by toluidine blue staining (Fig. S10; Fig. S12, middle) and TEM analysis (Fig. S11; Fig. S12, right). Because these values were less than 50%, i.e., reflecting the expected Mendelian segregation ratio of 1:1 for cct1-3 cct2-5 pollen cells from a cct1-3 cct2-5/CCT2 plant (Fig. S12, left), we concluded that all pollen cells displaying unusual ultrafine structures represented cct1-3 cct2-5 pollen cells. On the other hand, the germination rates of pollen grains from cct2-5 cct1-3/CCT1 and cct1-3 cct2-5/CCT2 plants in the qrt1-1 background were 32.0 and 30.8%, respectively, and these values were less than half the germination rates of pollen grains from a cct2-5 (72.3%) plant and a cct1-3 (66.0%) plant in the qrt1-1 background, respectively, demonstrating that more than half the pollen grains from cct2-5 cct1-3/CCT1 or cct1-3 cct2-5/CCT2 plants in the qrt1-1 background did not germinate in vitro. Thus, among the total pollen grains from cct2-5 cct1-3/CCT1 or cct1-3 cct2-5/CCT2 plants in the qrt1-1 background, although cct2-5 or cct1-3 interfered slightly with pollen germination, cct1-3 cct2-5 completely blocked pollen germination. Accordingly, we speculated that the unusual ultrafine structures found specifically in cct1-3 cct2-5 pollen grains contributed to the defect in pollen germination.
Ultrafine structures of cct1-3 cct2-5 pollen cells from cct2-5 cct1-3/CCT1 and cct1-3 cct2-5/CCT2 plants
On the day of anthesis, ultrathin sections of pollen grains were prepared for TEM via a 30-min incubation in pollen germination medium. All the results obtained from images of pollen grain cells from cct2-5 cct1-3/CCT1 and cct1-3 cct2-5/CCT2 plants are summarized in Figs. S7 and S8, respectively.
Ultrathin sections of pollen grains from cct2-5 cct1-3/CCT1 plants (Fig. S7) showed distinct features specific to cct1-3 cct2-5 pollen cells. First, cct1-3 cct2-5 pollen cells from a cct2-5 cct1-3/CCT1 plant contained extremely enlarged LBs together with small or tiny LBs [Figs. 5d (S7a, 6d), S7c, f, and h]. Regardless of their size, these LBs were electron-dense, and the greatest density was near the surface. However, they appeared to be further surrounded by a semi-transparent boundary, although the boundary was somewhat discontinuous. Second, the cct1-3 cct2-5 pollen cells contained numerous yet diffuse SGSs (Fig. 6d, SGS*; Fig. S7e and i) and a few “solid” SGSs (Figs. 6d and S7i). Third, the remnants of “swollen” SGSs could be seen within a membranous body having a diameter less than ~ 2 μm [Fig. 5d (6d), black arrows; Fig. S7d], which we referred to as a prevacuolar body (Fig. 6d, PVB). On the other hand, as seen in wild-type, cct2-5, and cct1-3 pollen cells, two GN [Figs. 5d (S7a, b) and S7g] and developing cellular septa [Figs. 5d (S7a, b) and S7g; open arrowheads] were evident, consistent with the DAPI staining results (Fig. 3d); moreover, rER was evident throughout the cytoplasm (Fig. 5d, rE; Fig. S7b, closed arrowheads) and between cellular components (Fig. 6d, star; Fig. S7g, closed arrowheads).
Some cct1-3 cct2-5 pollen cells from a cct2-5 cct1-3/CCT1 plant displayed the most extreme phenotype (Fig. 8a). These cells contained two types of PVBs having a diameter less than ~ 2 μm: one type of PVBs contained many undigested SGSs [Fig. 8a (S7s), boxes 2, 4, and 5], and another type of PVBs contained various cytoplasmic components [Fig. 8a (S7s), box 6], although some PVBs looked not completely closed. Each type of PVB were gathered on opposite sides of the cytoplasm, as if they were in the process of fusing to generate larger bodies (box 2, SGS). On the other hand, rER appeared to enclose SGSs, SGS*s, LBs, and other kinds of cellular components together to yield PVBs [Fig. 8a (S7s), box 6] or the rER entrapped several PVBs (Fig. 7c; Fig. S8m, arrow). However, fully developed SVBs were scarce [Fig. S7r (S7s)]. The occurrence of different types of PVBs suggested that the whole processes of SVB development from SGSs and/or PVBs might be delayed in cct1-3 cct2-5 pollen cells from a cct1-3 cct2-5/CCT2 plant. The limited development of SVBs might have resulted from a possible delay in the supply or vesicular delivery of degradative enzymes into vSVs and PVBs.
Fig. 8. Unusual ultrafine structures observed by TEM in cct1-3 cct2-5 pollen cells. Panel a shows images taken from 7 regions of a cct1-3 cct2-5 pollen cell from a cct2-5 cct1-3/CCT1 plant, shown in Fig. S7r (S7s). Box 1: some enlarged LBs were enclosed within a membrane boundary with discontinuous electron-dense staining. Boxes 2, 4, and 5: many SGSs were entrapped within PVBs of ~ 2 μm in diameter, within which entrapped SGSs looked diffused and undigested. Two types of PVBs could be seen: one containing the cytosol (box 2) might be a very swollen or fused form of SGS*, and the other containing different types of cytoplasmic components (box 6) was a typical PVB. Apart from these PVBs, the SGSs, enlarged LBs, and other cellular components were all entrapped within autophagosome-like bodies (boxes 3, 5, and 7). Panel b shows images from a cct1-3 cct2-5 pollen cell from a cct1-3 cct2-5/CCT2 plant, shown in Fig. S8p. Box 1: vSV or SVB appears to be ingesting a small LB. Boxes 2–4: LBs are in the process of being entrapped by membranous, autophagosome-like bodies, which were already ingesting other cellular components. Boxes 5 and 6: the endoplasmic reticulum or the autophagosome-like bodies are in the process of entrapping LBs and other cellular components, while the termini of the developing membranes displayed an unusual outward curvature as shown by asterisks. Bars = 1 μm
Apart from the two types of PVBs described above, i.e., one containing many SGSs and the other containing various cytoplasmic components other than SGSs, SGSs, SGS*s, enlarged LBs, and other cellular components were all together entrapped within a different type of membrane body, which we referred to as autophagosome-like bodies [Fig. 8a (S7s), boxes 3, 5, and 7]. The content of the autophagosome-like bodies remained undigested, possibly because there were no large vacuoles. Thus, the autophagic processes that normally operate in germinating wild-type pollen cells appeared to be suspended in cct1-3 cct2-5 pollen cells from a cct2-5 cct1-3/CCT1 plant.
The distinct features of cct1-3 cct2-5 pollen cells from a cct2-5 cct1-3/CCT1 plant, as described above, were also collectively recognized in ultrathin sections of cct1-3 cct2-5 pollen grains from a cct1-3 cct2-5/CCT2 plant. Briefly, extremely enlarged LBs co-existed with small ones [Figs. 5e (6e, S8a), S8b, f, and g]; almost all SGSs were swollen, and the two types of PVBs entrapping undigested forms of SGS*s [Figs. 7c (S8a), S8c, and k] and other cellular components (Figs. 7c, S8c, d, e, l, n, o, and l) appeared to be halted in their development into SVBs. Notably, a small proportion of cct1-3 cct2-5 pollen cells from a cct1-3 cct2-5/CCT2 plant displayed the most extremely deformed cytoplasmic structures [Figs. 8b (S8p), S8u, and S8aa]; as shown in Fig. 8b (S8p), although some LBs were incorporated into vSVs or developing SVBs (Fig. 8b, box 1; the same as Fig. S8r), extremely enlarged LBs were entrapped by autophagosome-like bodies (Fig. 8b, boxes 2–4; and Fig. S8q). However, it was uniquely seen in a cct1-3 cct2-5 pollen cell from a cct1-3 cct2-5/CCT2 plant that the autophagosome-like bodies entrapping various cellular components appeared to be incompletely closed and had unusual membrane openings, typically curling outward at both ends (Fig. 8b, boxes 5 and 6, asterisks; the same photos in Fig. S8s). Such features were never seen in the wild-type, cct2-5 and cct1-3 pollen cells or even in a cct1-3 cct2-5 pollen cell from a cct2-5 cct1-3/CCT1 plant. However, it is possible that the unusual membrane openings observed in the autophagosome-like bodies could have resulted from osmotic shock caused by chemical fixation.
In summary, cct1-3 cct2-5 pollen cells from cct2-5 cct1-3/CCT1 and cct1-3 cct2-5/CCT2 plants had two types of PVBs containing undigested SGSs or various kinds of cytoplasmic components. The former PVBs might have been derived from an extensively swollen form of SGS (Fig. 8a, box 2). Nonetheless, these results suggested that the normal process of vacuole development was substantially delayed or suspended in cct1-3 cct2-5 pollen cells. Moreover, the extremely enlarged LBs and other cytoplasmic components were entrapped within autophagosome-like bodies, and in the most extreme case (a cct1-3 cct2-5 pollen cell from a cct1-3 cct2-5/CCT2 plant) the enclosure by autophagosome membranes was incomplete. Because autophagic activity has been reported to be essential for pollen germination (Fujiki et al. 2007; Qin et al. 2007), the occurrence of autophagosome-like bodies enclosing undigested cellular components or even unclosed autophagosome-like bodies in some cct1-3 cct2-5 pollen cells suggested that the ultimate cellular process required for pollen germination was eventually suspended, possibly owing to PC shortage. Because PC is a bilayer-forming lipid, an extreme degree of PC depletion could cause an excess of PE, which is a non-bilayer lipid, which might cause outward curvature of membranes at the termini of developing autophagosomes. Finally, it should be noted that swollen SGSs were generated in cct1-3 cct2-5 pollen cells partly under the influence of sporophytic defects, i.e., a stronger sporophytic influence was recognized in cct1-3 cct2-5 pollen cells from a cct1-3 cct2-5/CCT2 plant than those from a cct2-5 cct1-3/CCT1 plant.
Complementation of the mutant phenotype by expression of a cDNA encoding CCT1
The results of our reciprocal crosses (Tables 3 and S3) showed that cct2-5 and cct2-3 cannot be transmitted via male gametophytes in the cct1-3 background, and such a defect was most likely caused by the lack of CCT1 expression in cct1-3 pollen. We therefore tried to complement this mutant phenotype by expressing CCT1 under the control of a pollen-specific promoter. However, our attempts to create a Ti-plasmid that could drive expression of a CCT1 cDNA under the control of the pollen-specific promoter ACA9pro (pPZP221_ACA9pro: CCT1cDNA-nosT) were unsuccessful. Inadvertently, however, we obtained an unusual construct that contained ACA9pro in the inverted orientation (invACA9pro-CCT1cDNA-nosT) as shown in Fig. S1. Interestingly, however, a transgenic cct1-3 cct2-5/CCT2 plant that contained an invACA9pro-CCT1cDNA-nosT construct (+/–) (designated line M21) was created, and pollination of emasculated cct1-3 flowers with pollen from the M21 plant resulted in seeds, some of which carried cct2-5 together with the invACA9pro-CCT1cDNA-nosT construct. Genotyping of seedlings revealed that cct2-5 was not found when the seedlings had no T-DNA; moreover, among the seedlings that had T-DNA, half of them carried cct2-5 (Table 4). Although the promoter construct was inverted and its pollen-specific expression was not verified, these results suggested that the expression of CCT1 cDNA was sufficient to rescue the defect of male gametophytes in cct2-5 transmission to offspring in the cct1-3 background. Thus, we concluded that co-transmission of invACA9pro: CCT1cDNA-nosT with cct2-5 was necessary and sufficient for the male gametophytic transmission of cct2-5 to offspring in the cct1-3 background.
Table 4. Pollination of emasculated cct1-3 flowers with pollen from cct1-3 cct2-5/CCT2 invACA9pro: CCT1cDNA-nosT (+/–) plants (T1, M21) aCrossingNumber of F1 plantsχ^2^ test for 1:1 ^b^Female cct1-3 MaleM21T-DNA(+/–) w/o cct2-5**T-DNA(+/–) with cct2-5Totalχ^2^ p Observed frequency78150.0670.796 ^c^Expected frequency7.57.515In this table, cct2-5 is boldfaced to emphasize transmission via male gametophytes.^a^ Among cct1-3 seedlings examined, 15 carried the T-DNA whereas 8 carried cct2-5. No cct2-5 was found in cct1-3 seedlings without the T-DNA.^b^ χ^2^ test was conducted to test the null hypothesis that cct1-3 individuals carrying T-DNA have CCT2/CCT2 and cct2-5/CCT2 in the predicted 1:1 Mendelian ratio (degree of freedom = 1).^c^ The p value was too large to mistakenly reject the null hypothesis.Thus, we concluded that the co-transmission of invACA9pro: CCT1cDNA is necessary and sufficient for transmission of cct2-5 from the pollen of cct1-3 cct2-5/CCT2 invACA9pro: CCT1cDNA-nosT (+/–) plants (T1, M21) to the ovule of cct1-3 plants.
Discussion
Because PC plays important roles in membrane biogenesis and temperature acclimation of sporophytes, we anticipated that genetic defects in PC biosynthesis would lead to a biological inferiority in mutant plants compared with wild-type plants. In A. thaliana, we showed that seeds or seedlings carrying cct1-3 cct2-3 or cct1-3 cct2-5 were inviable, so the transmission of a genetic trait negating the CDP-choline pathway to offspring was avoided. Because PC biosynthesis via the CDP-choline pathway is critical for successful fertilization, we first examined whether cct1-3 cct2-3 and cct1-3 cct2-5 could be transmitted to offspring via male or female gametophytes. Our reciprocal crossing of cct2-3/CCT2 and cct2-5/CCT2 plants in the cct1-3 background (Tables 3 and S3) showed that cct2-3 and cct2-5 could not be transmitted via male gametophytes in the cct1-3 background, whereas they can be transmitted via female gametophytes. We then examined whether cct1-3 cct2-3 and cct1-3 cct2-5 could be maintained or eliminated during gametogenesis. TEM (Fig. S3a), Alexander’s viability test (Fig. 2), and DAPI staining (Fig. 3) revealed that cct1-3 cct2-5 did not affect microspore maturation. However, pollen germination experiments in vitro showed that cct1-3 cct2-5 eventually prohibits pollen germination (Fig. 4b). By contrast, we deduced that cct1-3 cct2-5 partly allows maturation of megaspores and fertilization of female gametophytes (Tables 1 and S2). In the following sections, we discuss how cct1-3 cct2-5 inhibits pollen germination and if there is any biological relevance of differential transmission of mutant genes between male and female gametophytes.
For pollen-tube budding and elongation, pollen cells require energy and osmolytes as well as polar membrane lipids, and the vegetative cells of pollen grains accumulate an abundance of varied metabolic reserves (Hafidh et al. 2016; Pacini 1996; Pacini et al. 2006). Development of vacuoles may be important for catabolism of storage materials as well as accumulation of osmolytes. Energy also may be required for intracellular trafficking of pollen-tube materials to the budding point. In A. thaliana and other plants, the appearance, morphology, and functions of tonoplasts change dramatically during pollen development (Akita et al. 2021; Amela Galcía et al. 2002; Fabrice et al. 2017; Kuang and Musgrave 1996; Noguchi 2006; Owen and Makaroff 1995; Van Aelst et al. 1993; Vinckier et al. 2012; Yamamoto et al. 2003; Zhang et al. 2002). After chemical fixation of mature pollen grains, visualization of the cytoplasm of the vegetative cell has revealed many vacuoles having an average diameter of 300 nm as well as numerous small storage bodies that have both electron-dense and electron-transparent regions (Kuang and Musgrave 1996; Owen and Makaroff 1995; Van Aelst et al. 1993). On the other hand, Yamamoto et al. (2003) used A. thaliana pollen grains prepared by cryofixation to show that storage vacuoles form de novo from rER cisterna and predominate in pollen cells just before anthesis, whereas after germination the storage vacuoles begin to fuse and form an enlarged vacuole that contributes to the generation of turgor pressure to drive pollen-tube elongation. Furthermore, in pollen cells late after anthesis (when the stamen becomes higher than the pistil), lysosomal structures or lytic vacuoles appear that contain certain subcellular compartments, causing pollen microspores to undergo autolysis (Yamamoto et al. 2003). In A. thaliana pollen grains prepared by chemical fixation after a 30-min incubation in pollen germination medium, we confirmed that vacuole development occurs normally as described previously (Kuang and Musgrave 1996; Owen and Makaroff 1995; Van Aelst et al. 1993). For describing vacuoles at the initial and early stages of enlargement, however, we used the terms vSV and SVB, respectively: in cct2-5 and cct1-3 pollen grains, vSVs and SVBs with electron-transparent content were as prominent as in wild-type pollen grains (Figs. 5a–c and 7a and b), whereas in cc1-3 cct2-5 pollen grains from cct2-5 cct1-3/CCT1 and cct1-3 cct2-5/CCT2 plants, vSVs and SVBs with electron-transparent content were rare (Fig. 5d, e); moreover, some vacuoles contained undigested materials against the slightly electron-dense background (Fig. 5d, e). Furthermore, in cc1-3 cct2-5 pollen grains that were functionally compromised, PVBs enclosing various cellular contents were prominent (Figs. 7c and 8a). These results conform to our view that the digestion of stored materials is delayed in cc1-3 cct2-5 pollen cells.
LBs could serve as a source of the diacylglycerol backbone for polar lipids as well as a source of the osmolyte sucrose, to which glyoxisomal catabolism of TAGs might contribute. Our TEM analysis using pollen grains after a 30-min incubation in pollen germination medium showed that LBs in wild-type pollen cells had semi-transparent boundaries (Figs. 6a, S4c, h, m, and q), which is consistent with the fact that LB is covered with a phospholipid monolayer embedding oil-body proteins. Thus, semi-transparent boundaries of LBs may contribute to LB fusion with SVBs (Fig. S4o). LBs are then catabolized within SVBs (Figs. 7a, S4k, and o). Yamamoto et al. (2003) reported that LBs are surrounded by highly dilated rER in pollen cells just before anthesis. This process may be related to a recently identified process, termed lipophagy (or autophagic degradation of lipids) (Schulze et al. 2017; Tarique et al. 2019). Lipophagy includes two different mechanisms, namely macroautophagy and microautophagy. In macroautophagy, the degradation of LBs begins with their enclosure within autophagosomes, and subsequent fusion of the autophagosomes with vacuoles leads to lipid breakdown (Oku and Sakai 2018). In microautophagy, the degradation of LBs begins with their direct contact and engulfment/fusion with vacuoles, leading to lipid breakdown (Huang et al. 2019). Akita et al. (2021) reported that LB degradation in A. thaliana pollen is morphologically considered microautophagy, as shown in images of LBs that had fused with small vacuoles. Our work produced similar images as shown in Figs. S4o, S5o, and S6o, suggesting that lipid degradation by microautophagy occurs in wild-type, cct2-5, and cct1-3 pollen grains. By contrast, the images shown in Figs. S7e and S8n suggest that lipid degradation by microautophagy occurs less frequently or is delayed in cct1-3 cct2-5 pollen grains, which coincides with less frequent occurrence of vSVs and SVBs with electron-transparent content in cct1-3 cct2-5 pollen grains.
Another source of energy and osmolytes is SGSs [(Figs. 6a, S4d, i, n, and r), which may undergo autodegradation in germinating pollen cells as suggested by the presence of a small transparent area in the center, leading to the development of vSVs. However, SGSs were also found to be entrapped within membranous bodies to form PVBs (Figs. 6a, S4f, and k), in which SGSs were digested together with other cellular contents to form SVBs. The developing SVBs likely fuse with one another to form a larger vacuole (Fig. 7a). Thus, the development of SVBs is one of the important steps during pollen germination. Entrapment of LBs and other cellular components by the rER is another pathway to vacuole formation. Kuang and Musgrave (1996) reported that, in the mature pollen of A. thaliana, large numbers of small particles are scattered irregularly throughout the vegetative cytoplasm. These particles appear in pollen grains only when a flower is at anthesis, suggesting that they could be relevant to pollen germination. Furthermore, most of these particles contain electron-transparent areas and seem to be enveloped by a thin membrane. These features are very similar to what we observed for SGSs. Since 1996, however, little attention has been paid to the small particles in A. thaliana pollen grains. Our results not only enforce the previous view that the small particles or SGSs are involved in pollen germination but also provide new evidence that they might undergo autodegradation and/or entrapment by rER for producing energy and/or osmolytes required for pollen germination.
Compared with wild-type pollen cells (Figs. 5a, S4l, and p), we showed that SVBs become less prominent in cct2-5 pollen cells from a cct2-5 plant (Figs. 5b, S5g, and p) or cct1-3 pollen cells from a cct1-3 plant (Figs. 5c, S6l, and p). For cct1-3 pollen cells (Figs. 5c, S6l, and o), vSVs gathered closely to one another. These results suggested that fusion of vSVs into SVBs was delayed in cct2-5 and cct1-3 pollen cells compared with wild-type cells. However, we showed that such a delay only partially slowed the pollen germination rates of cct2-5 and cct1-3 pollen compared with wild-type pollen. In cct2-5 and cct1-3 pollen cells (Figs. 6c, S6a, g, l, and p), LBs were both smaller and less numerous than in wild-type pollen cells. This may be attributable to downregulation of LB biogenesis because PC serves as a precursor to TAG, the main component of LBs. However, because the semi-transparent boundaries of LBs were preserved in cct2-5 and cct1-3 pollen cells, we concluded that LBs in these cells could fuse with PVBs or SVBs for subsequent digestion of LB contents. Indeed, our pollen germination experiments showed that the decreased number and size of LBs in cct2-5 and cct1-3 pollen cells correlated with a decrease in pollen germination rate.
We showed that cct1-3 cct2-5 pollen cells from a cct2-5 cct1-3/CCT1 plant contain unusually enlarged LBs together with very small LBs (Figs. 6d, S7a, f, and S13b). The disproportionate sizes of LBs in the cct1-3 cct2-5 pollen cells may be similar to what has been reported for seed tissues of oleosin mutants (Shimada et al. 2008). Together with phospholipids, oleosins are required as an LB-surface material for sequestrating hydrophobic TAGs within the hydrophilic cytoplasm, and genetic downregulation of oleosin biogenesis causes fusion of LBs so that the mutant cells can save the surface materials. LBs have a phospholipid monolayer, mainly consisting of PC. Thus, the presence of enlarged LBs as well as the very small LBs in cct1-3 cct2-5 pollen cells from a cct2-5 cct1-3/CCT1 plant may have been a consequence of severe PC shortage. The surface of such enlarged LBs looked highly electron-dense or darker compared with LBs in wild-type, cct2-5, and cct1-3 pollen cells, which may have been caused by a shortage of oil-body surface phospholipids owing to the limited supply of PC. Some LBs retained a semi-transparent, yet fragmentary, boundary (Fig. S7h, closed arrowhead; Fig. 8a, box 1). The unusually enlarged LBs may have been too large to be incorporated into PVBs (Fig. S7b), and hence an alternative route for LB degradation would be required, as discussed below. By contrast, the smaller LBs may have been incorporated into PVBs (Fig. S7b, d). The cct1-3 cct2-5 pollen cells from a cct2-5 cct1-3/CCT1 plant contained many PVBs (Figs. S7d, l, q, and S13g) but few SVBs or vSVs.
The cct1-3 cct2-5 pollen cells from both cct2-5 cct1-3/CCT1 and cct1-3 cct2-5/CCT2 plants contained large, autophagosome-like bodies that entrapped undigested cellular contents such as the unusually enlarged LBs, swollen SGSs, and other cytoplasmic components (Fig. 8; Fig. S8aa, ac). This may reflect the attempt of such cells to facilitate the degradation of cellular components by autophagy. It has been reported that autophagy is essential for pollen germination, but under physiological conditions the turnover of such autophagosome-like bodies would occur too quickly for analysis by TEM. We also noticed that cct1-3 cct2-5 pollen cells from cct1-3 cct2-5/CCT2 plants contained incompletely closed autophagosome-like bodies (Fig. 8, boxes 5 and 6; Fig. S8s), suggesting that autophagy is required for pollen germination yet is interrupted owing to the shortage of PC. Thus, cct1-3 cct2-5 has formidable negative consequences for autophagy and, hence, the co-transmission of cct1-3 and cct2-5 via male gametophytes is eventually prohibited. However, future studies should examine whether the incompletely closed autophagosome-like bodies form de novo or represent an artifact of sample preparation. Alternatively, the occurrence of the large, autophagosome-like bodies that entrapped the undigested cellular contents could be a result of a programmed suicide process for eliminating unfavorable pollen grains; in this regard, Yamamoto et al. (2003) reported that pollen grains not participating in fertilization are eliminated by intracellular lytic bodies. Because we harvested A. thaliana pollen grains on the day of anthesis and prepared samples after a 30-min incubation in pollen germination medium, the timing of sample preparation was a few days earlier than that used by Yamamoto et al. (2003). Therefore, it seems unlikely that lysosomal lytic bodies enclosing various cellular contents had been created in our pollen grain samples at the harvesting time. Thus, the autophagosome-like body containing various cellular contents must have been created during incubation in pollen germination medium. Because similar structures were seen in several cct1-3 cct2-5 pollen grains from cct2-5 cct1-3/CCT1 and cct1-3 cct2-5/CCT2 plants (e.g., Figs. 8a and S8ac), we believe that the unusual autophagosome-like body was not derived from a degenerating pollen grain included in our pollen sample. However, it remains to be clarified whether the unusual autophagosome-like body is created via inhibition of normal autophagic processes required for pollen germination or induction of lytic processes during incubation by a suicide signal emitted in response to PC shortage. Regardless, both processes were found to be suspended by PC shortage in cct2-5 cct1-3 pollen grains.
Our reciprocal crossing of cct2-5/CCT2 and cct2-3/CCT2 plants in the cct1-3 background (Tables 3 and S3) revealed that simultaneous transmission of cct2-5 or cct2-3 alleles with cct1-3 via the female gametophyte is largely permissible—the survival rate of cct1-3 cct2-5 ovules was 45.5 and 71.4% in cct1-3 cct2-5/CCT2 and cct1-3 cct2-3/CCT2 plants, respectively. Thus, there are strategical differences between male and female gametophytes regarding the transmission of mutant genes: transmission of the defects in PC biosynthesis via male gametophytes is strictly prohibited because a wide dispersal of unfavorable traits by pollen is not beneficial to the population. By contrast, a portion of ovules carrying cct1-3 and cct2-5 (or cct2-3) is maintained so that transmission of the background genome can be ensured upon fertilization with normal pollen.
Shockey et al. (2016) reported that gpat9-2 affects phosphatidic acid biosynthesis in seeds and is also not transmissible via male gametophytes and that gpat9-2 pollen grains are unable to germinate in vitro in the qrt1 background. They concluded that because disruption of phosphatidic acid biosynthesis inhibits TAG biosynthesis, the inhibition of pollen germination is a consequence of TAG shortage (Shockey et al. 2016). Our results showed that cct1-3 cct2-5 pollen cells in the qrt1 background had a pollen-germination phenotype similar to that of gpat9-2, although cct1-3 cct2-5 pollen cells still contained TAGs and that TAG catabolism was inhibited. Furthermore, the autophagy processes that are required for pollen germination are probably inhibited or halted in cct1-3 cct2-5 pollen cells.
Recent studies have disclosed that CCT1 and CCT2 play distinct roles during A. thaliana development. In response to a pathogen(s), many effectors are delivered to the host cytoplasm where they can be recognized by specific nucleotide-binding, leucine-rich repeat proteins; the translational efficiency of CCT2 mRNA is upregulated upon activation of these proteins, and when challenged with Pseudomonas syringae DC3000 (AvrRpm1) (de Torres et al. 2003), CCT2 is required for optimal RPM1-mediated resistance (Meteignier et al. 2017). Yoshihara et al. (2022) reported that the mutant cct2-1 (Inatsugi et al. 2009) displays a very minor—although statistically significant—difference in gravitropism, but this was not the case for root elongation rate. Xiao et al. (2025a) reported that, in the cct2-5 background, CCT1 regulates PC biosynthesis under normal conditions and root development under osmotic stress, and its phosphorylation state at S187 plays an important role in modulating its enzymatic activity and functions. However, because CCT1 not only forms a homodimer but also interacts with its isoform CCT2 (Xiao et al. 2025b), the roles of CCT1 and CCT2 could be overlapping in some cases. We previously reported that residual CCT activity in rosette-leaf homogenates of cct1-1 and cct2-1 mutants accounted for 29.3 and 78.5%, respectively, of the total CCT activity in the homogenates of the wild-type (WS) ecotype (Inatsugi et al. 2009). The A. thaliana eFP browser predicts relatively higher expression of CCT2 (At4G15130) in pollen than in other tissues (https://bar.utoronto.ca/efp2/Arabidopsis/Arabidopsis_eFPBrowser2.html). Our results showed that disruption of both CCT1 and CCT2 abolished pollen germination. We also showed that cct1-3 cct2-5/CCT2 and cct1-3 cct2-3/CCT2 plants exhibited partly restricted development of cct1-3 cct2-5/CCT2 and cct1-3 cct2-3/CCT2 seeds, respectively (Tables 1 and S2), whereas cct2-3 cct1-3/CCT1 plants did not restrict the development of cct2-3 cct1-3/CCT1 seeds (Table 2). During seed development, cct1-3 cct2-5/CCT2 and cct1-3 cct2-3/CCT2 seeds contain one copy of CCT2 in the embryo and two copies of CCT2 in the endosperm. However, these copies of CCT2 were found to be insufficient for seed development. In contrast, in cct1-3 plants, in which seed development proceeds normally, developing cct1-3 seeds contain two copies of CCT2 in the embryo and three copies of CCT2 in the endosperm, and these copies of CCT2 were found to be sufficient for seed development. Similarly, in cct2-5 cct1-3/CCT1 plants, developing cct2-5 cct1-3/CCT1 seeds contain one copy of CCT1 in the embryo and two copies of CCT1 in the endosperm, and these copies of CCT1 were found to be sufficient for seed development. Thus, it seems likely that CCT1 and CCT2 differentially contribute to the establishment of A. thaliana seeds after fertilization.
Overall, our results demonstrate that genetic defects that disrupt the CDP-choline pathway towards PC biosynthesis are eliminated before fertilization, but only during pollen germination (Fig. S13). The ultrafine structures of the wild-type and mutant pollen cells after a 30-min incubation in pollen germination medium have been schematically represented in Figs. S14–S16. However, future studies should investigate whether complete disruption of the CDP-choline pathway might allow the establishment of A. thaliana seedlings. For this purpose, it will be essential to engineer a Ti-plasmid construct to specifically drive the expression of CCT1 cDNA in a pollen-specific manner.
Supplementary Information
Below is the link to the electronic supplementary material.
Supplementary Material 1
The reference list from the paper itself. Each links out to its DOI / PubMed record.
- 1Craddock CP, Adams N, Bryant FM, Kurup S, Eastmond PJ (2015) PHOSPHATIDIC ACID PHOSPHOHYDROLASE regulates phosphatidylcholine biosynthesis in Arabidopsis by phosphatidic acid-mediated activation of CTP:PHOSPHOCHOLINE CYTIDYLYLTRANSFERASE activity. Plant Cell 27: 1251–1264 10.1105/tpc.15.00037 PMC 455869825862304 · doi ↗ · pubmed ↗
- 2Keogh MR, Courtney PD, Kinney AJ, Dewey RE (2009) Functional characterization of phospholipid N-methyltransferases from Arabidopsis and soybean. J Biol Chem 284: 15439–1544710.1074/jbc.M 109.005991 PMC 270884119366698 · doi ↗ · pubmed ↗
- 3Qin G, Ma Z, Zhang L, Xing S, Hou X, Deng J, Liu J, Chen Z, Qu L-J, Gu H (2007) Arabidopsis At BECLIN 1/At Atg 6/At Vps 30 is essential for pollen germination and plant development. Cell Res 17: 249–26310.1038/cr.2007.717339883 · doi ↗ · pubmed ↗
- 4Vogler F, Schmalzl C, Englhart M, Bircheneder M, Sprunck S (2014) Brassinosteroids promote Arabidopsis pollen germination and growth. Plant Reprod 27: 153–16710.1007/s 00497-014-0247-x 25077683 · doi ↗ · pubmed ↗