Plant Mol Biol DOI 10.1007/s11103-013-0161-x

A novel chloroplast localized Rab GTPase protein CPRabA5e is involved in stress, development, thylakoid biogenesis and vesicle transport in Arabidopsis Sazzad Karim · Mohamed Alezzawi · Christel Garcia‑Petit · Katalin Solymosi · Nadir Zaman Khan · Emelie Lindquist · Peter Dahl · Stefan Hohmann · Henrik Aronsson 

Received: 23 August 2013 / Accepted: 3 December 2013 © Springer Science+Business Media Dordrecht 2013

Abstract  A novel Rab GTPase protein in Arabidopsis thaliana, CPRabA5e (CP = chloroplast localized) is located in chloroplasts and has a role in transport. Transient expression of CPRabA5e:EGFP fusion protein in tobacco (Nicotiana tabacum) leaves, and immunoblotting using Arabidopsis showed localization of CPRabA5e in chloroplasts (stroma and thylakoids). Ypt31/32 in the yeast Saccharomyces cerevisiae are involved in regulating vesicle transport, and CPRabA5e a close homolog of Ypt31/32, restores the growth of the ypt31Δ ypt32ts mutant at 37 °C in yeast complementation. Knockout mutants of CPRabA5e displayed delayed seed germination and growth arrest during oxidative stress. Ultrastructural studies revealed that after preincubation at 4 °C mutant chloroplasts contained larger plastoglobules, lower grana, and more vesicles close to the envelopes compared to wild type, and vesicle formation being enhanced under oxidative stress. This indicated Electronic supplementary material  The online version of this article (doi:10.1007/s11103-013-0161-x) contains supplementary material, which is available to authorized users. Sazzad Karim and Mohamed Alezzawi have contributed equally to this study. S. Karim · M. Alezzawi · C. Garcia‑Petit · N. Z. Khan · E. Lindquist · H. Aronsson (*)  Department of Biological and Environmental Sciences, University of Gothenburg, SE‑405 30 Gothenburg, Sweden e-mail: [email protected] K. Solymosi  Department of Plant Anatomy, Institute of Biology, Eötvös University, Pázmány P. s. 1/c, Budapest 1117, Hungary P. Dahl · S. Hohmann  Department of Chemistry and Molecular Biology, University of Gothenburg, SE‑412 96 Gothenburg, Sweden

altered thylakoid development and organization of the mutants. A yeast-two-hybrid screen with CPRabA5e as bait revealed 13 interacting partner proteins, mainly located in thylakoids and plastoglobules. These proteins are known or predicted to be involved in development, stress responses, and photosynthesis related processes, consistent with the stress phenotypes observed. The results observed suggest a role of CPRabA5e in transport to and from thylakoids, similar to cytosolic Rab proteins involved in vesicle transport. Keywords  Chloroplast · Plastoglobuli · Rab · Transport · Thylakoid · Vesicle

Introduction The plant Rab GTPases are small GTP binding proteins of 20–40 kDa belonging to the RAS superfamily. Like other GTPases Rabs have both GTP/GDP binding and GTP hydrolysis capability in order to work as GTP-bound active and GDP-bound inactive molecular switches (Takai et al. 2001; Agarwal et al. 2009). In Rab GTPase regulatory cycle, a Rab escort protein (REP) grabs the newly synthesized Rab and presents it to the geranylgeranyl transferase for geranylgeranylation on the two C-terminal cysteines of the Rab proteins. This enhances Rab hydrophobicity before targeting to the membrane (Leung et al. 2006; Schwartz et al. 2007). Geranylgeranylated Rabs in their GDP-binding conformation remain inactive and bound to the GDP dissociation inhibitor (GDI) to be stabilized in an inactive form (Stenmark 2009; Pfeffer 2012). GDI displacement factor (GDF) recognizes the Rab–GDI complex and catalyzes the dissociation from the GDI, thereby facilitating the association of the Rab to the appropriate membranes. At the specific membranes, GTP

13



binding, facilitated by regulatory enzymes called guanine nucleotide exchange factors (GEFs), activates Rabs which in turn activate several types of effectors. Upon the completion of an individual transport cycle, GTP hydrolysis by Rab GTPase is stimulated by the GTPase-activating proteins (GAPs) resulting in GDP bound Rab, which is recognized by GDI proteins to remain at the inactive state until the new transport cycle begins with the recruitment of Rab to a membrane (Ali and Seabra 2005; Nielsen et al. 2008; Stenmark 2009). The prime function of Rab proteins is their involvement in membrane transport among the organelles, through vesicle transport of cargo proteins to their destinations. With their effector proteins, Rabs work in a combinatorial way to regulate all stages of membrane transport within the organelles (Pfeffer 2001; Pfeffer and Aivazian 2004). Rabs interact with other regulatory and effector proteins to regulate the cycle of GDP/GTP binding and GTP hydrolysis. This modulates vesicle budding from the donor membrane, uncoating, movement through the cell, tethering and docking in the vicinity of the acceptor membrane, final delivery of cargo to the target membrane by membrane fusion, as well as regulating cargo sorting during coated vesicle transport (Rutherford and Moore 2002; Nielsen et al. 2008; Stenmark 2009; Angers and Merz 2011). However, the analysis of GTP hydrolysis deficient mutant shows that GTPase activity is not always essential for vesicle fusion (Novick and Zerial 1997). Rab proteins can act as well as molecular switches to regulate many ion channels differentially with their GTP or GDP bound states (Saxena and Kaur 2006). Every Rab has specific intracellular localization to govern explicit membrane and vesicle transport. However, this transport can be regulated by several Rabs indicating their redundancy (Leung et al. 2006; Nielsen et al. 2008). Among eukaryotes, yeast, human and Arabidopsis have about 11, 70 and 57 predicted Rab proteins, respectively. In Arabidopsis, the 57 Rabs are divided into eight (A–H) subclasses based on the sequence and functional similarities with yeast and mammalian orthologs (Pereira-Leal and Seabra 2001; Rutherford and Moore 2002; Vernoud et al. 2003; Pfeffer 2005; Stenmark 2009). It has been suggested that ancestral Rabs have been multiplied and diversified in higher plants to be functional as plant-specific paralogs. Out of the 57 Rabs in Arabidopsis, 26 belong to the subclass AtRabA and similar class distributions have been observed in other plant species (Nielsen et al. 2008). Conversely, in human and in yeast, this group of Rab proteins is present in a lower number such as in Schizosaccharomyces pombe (a single gene Ypt3), in Saccharomyces cerevisiae (two redundant genes Ypt31 and Ypt32) and in human (three genes Rab11A, Rab11B, and Rab25) (Chow et al. 2008). AtRabA5e (At1g05810), a member of the subclass AtRabA has been suggested to be

13

Plant Mol Biol

localized in the chloroplast because of its unique N-terminal extra stretches of amino acids representing a chloroplast transit peptide (Khan et al. 2013). Ypt31/32 designates two yeast Rabs Ypt31 and Ypt32, which are members of the yeast Ypt Rab family showing >80 % sequence identity to each other (Tsujimoto et al. 2012). Ypt31 and Ypt32 are involved in regulating vesicle transport in exocytosis and endocytosis and more precisely in trans Golgi trafficking and are required for recycling from the plasma membrane through early endosomes to the Golgi complexes (Chen et al. 2011). It is also suggested that Ypt31/32 GTPases work in the yeast exocytic pathway and are involved in Golgi-to-plasma membrane and endosome-to-Golgi transport (Jedd et al. 1997; Zou et al. 2012). Furthermore, it is suggested that the Ypt31/32 pathway regulates putative phospholipid translocases to promote formation of vesicles destined for the trans-Golgi network. Phospholipid translocases are thought to be implicated in the generation of phospholipid asymmetry in membrane bilayers (Furuta et al. 2007). While single deletion mutants of YPT31 or YPT32 are viable, double deletion is lethal and hence the protein is essential for yeast (Benli et al. 1996). Rab GTPases together with tethering factors and SNAREs (soluble N-ethylmaleimide-sensitive factor attachment protein receptor) play an important role during vesicle fusion. During the final stages of the fusion process between the vesicle and its target membrane for cargo delivery, the SNARE proteins together with the related tethering proteins such as transport protein particle (TRAPP) complexes play a central role. Vesicle tethering may have separate kinetic and thermodynamic elements and the tethering complexes have specific activator roles on specific Rabs through different mechanisms. In yeast, TRAPP II complex specific subunits function as GEFs for Ypt1 or Ypt31/32 (Whyte and Munro 2002; Sacher et al. 2008; Zou et al. 2012). The multiplicity of the Rabs confirms their various roles in the eukaryotic cell including membrane trafficking through vesicle transport, growth, development, signalling pathways, stress responses and defence mechanisms (Schwartz et al. 2007; Agarwal et al. 2009). In this paper, a novel chloroplast localized Rab GTPase CPRabA5e, which is an ortholog of Ypt31/32 involved in plant development, thylakoid biogenesis and stress response with a putative role in vesicle transport is reported.

Results CPRabA5e is a Rab GTPase and is highly homologous to yeast Rabs Ypt31/32 According to the classification of small GTPases, CPRabA5e (At1g05810) belongs to the AtRabA subfamily and

Plant Mol Biol

Fig. 1  CPRabA5e domains and GTPase activity. a CPRabA5e is a protein of 261 amino acids with a molecular weight of 28.8 kDa including at the N-terminus a 42 amino acid long chloroplast transit peptide (TP). It contains a GTPase domain including five regions of conserved motifs (G1, aa 62–69; G2, aa 84–92; G3, aa 109–115; G4, aa 167–171; G5, aa 196–220). The CC residues (aa 257–258) at the C-terminal end represent a geranylgeranylation motif, and a highly conserved motif YYRGA (aa 124–128, with unknown function) usually found in most Ypt/Rab proteins. b The mature version of CPRa-

bA5e without a transit peptide, CPRabA5e, containing the predicted GTPase domain and produced in E. coli, displays GTPase activity as detected using a malachite green assay. Measurements of inorganic phosphate (Pi) release during a 30 min GTPase assay, with 0–30 mM GTP and 1 μM of purified CPRabA5e. Average values from three experiments ±SD. c Non-linear fit of Michaelis–Menten to calculate the curve that best matches the observed values. The kcat was estimated to be 8.54 ± 1.29 min−1

has been previously named as AtRabA5e (Vernoud et al. 2003) although it has also been suggested to be an Arf1 homolog protein (Andersson and Sandelius 2004). However, due to its localization determined further in this paper, it is renamed to CPRabA5e (CP = chloroplast localized). This protein consists of 261 amino acids and contains a unique 42 amino acid long transit peptide for chloroplast localization as verified using TargetP (Andersson and Sandelius 2004; Emanuelsson et al. 2007). Analysis of CPRabA5e structural features revealed a conserved GTPase domain including G1, G3, G4 and G5 motifs essential for GTP binding and hydrolysis as well as a G2 motif for effector binding. CPRabA5e also contains a highly conserved RabF (Rab family) motif YYRGA found in most Ypt/Rab proteins, although the function of the motif is unknown. The crucial CC motif for geranylgeranylation of the C-terminal for Rab proteins to be prenylated is also present (Fig. 1a) (Borg and Poulsen 1994; Pereira-Leal and Seabra 2001; Takai et al. 2001; Leung et al. 2006; Agarwal et al. 2009; Wittinghofer and Vetter 2011). To demonstrate that CPRabA5e possesses a GTPase activity, the coding sequence of the mature CPRabA5e protein (i.e. without the transit peptide), containing the predicted GTPase domain was expressed in E. coli and purified. GTPase activity was determined by monitoring inorganic phosphate (Pi) release using a malachite green assay (Lanzetta et al. 1979). It was observed that for addition

of 1 μM purified CPRabA5e a steady increase over time of phosphate release occurred with increasing addition of GTP, where 25 mM of GTP had a phosphate release of ca. 148 μmol Pi (Fig. 1b), which was not observed for the boiled control sample (data not shown). Using a non-linear fit of Michaelis–Menten the estimation of the kcat value of CPRabA5e was 8.54 ± 1.59 min−1 (Fig. 1c). Thus, CPRabA5e is confirmed to be a GTPase. Homology search (Blast search) using the CPRabA5e amino acid sequence revealed 37 orthologs in other plant species. The closest orthologs in flowering plants were defined as having more than 75 % sequence identity and for moss Physcomitrella spp., Selaginella spp. and conifers ranged from 50 to 73 % sequence identity (Supplementary Table S1). A phylogenetic tree (Supplementary Figure S1) was generated using CPRabA5e, its predicted orthologs and putative protein sequences from different plant species such as: Arabidopsis thaliana; Arabidopsis lyrata subsp. lyrata; Brassica napus (rapeseed); Populus trichocarpa (black cottonwood); Ricinus communis (castor oil plant); Medicago truncatula (alfalfa); Lotus japonicus; Glycine max (soybean); Nicotiana tabacum (tobacco); Vitis vinifera (grapevine); Oryza brachyantha; Oryza sativa subsp. japonica (Japanese rice); Sorghum bicolor (sorghum); Hordeum vulgare var. distichum (barley), Zea mays (maize), Physcomitrella patens subsp. patens (moss), Selaginella moellendorffii (Selaginella) and conifers Picea sitchensis,

13



Plant Mol Biol

Pinus taeda, Pinus pinaster and Larix kaempferi. According to the branching pattern of the resulting phylogenetic tree, several other plant species do have potential CPRabA5e orthologs. Notably, in this homology search no other CPRabA5e orthologs were found to contain any transit peptide similar to the one in CPRabA5e, indicating their localization outside chloroplasts. Furthermore, searching for CPRabA5e homology in the yeast database (S. cerevisiae WU-BLAST2 Search, www.yeastgenome.org/cgi-bin/blast-sgd.pl) showed high homology to Rab Ypt31 (YER031C) with 56 % identity and 75 % similarity and to Rab Ypt32 (YGL210 W) with 54 % identity and 74 % similarity (Supplementary Figure S2). Since both Rab Ypt31 and Rab Ypt32 have roles in vesicle transport (Chen et al. 2011), CPRabA5e might also have a similar role. CPRabA5e complements Ypt31/32 but not Arf1 in yeast To demonstrate if CPRabA5e shared similar features with its yeast counterparts a complementation study was performed. However, in a previous study CPRabA5e was suggested to be an Arf1 ortholog (Andersson and Sandelius 2004). The small Ras-like GTPase ADP-ribosylation factor 1 (Arf1) plays a vital role in COPI coated vesicle biogenesis, mainly in the Golgi apparatus, but can also have other roles at various intracellular compartments by acting on different sets of effectors proteins (Popoff et al. 2011). Thus, to rule out the possibility of CPRabA5e being similar to Arf1 it was tested whether CPRabA5e could complement the lethality of the yeast arf1Δ arf2Δ double mutant (Takeuchi et al. 2002). The arf1Δ arf2Δ double mutant was kept viable with a URA3-based plasmid expressing yeast wild-type ARF1 (Takeuchi et al. 2002). The arf1Δ arf2Δ double mutant transformed with CPRabA5e and empty yeast expression vector pYX242 were plated on medium containing 5-fluoroorotic acid (FOA) to eliminate the yeast Arf1 plasmid (pURA3-Arf1) and tested for complementation (Fig. 2a, b). The colonies containing the empty vector and those expressing CPRabA5e did not grow on the FOA containing medium. Therefore, CPRabA5e is not able to complement the yeast arf1Δ arf2Δ mutant as a functional Arf1/Arf2 ortholog. A homology search of CPRabA5e sequences suggested that it is a close homolog of yeast Rabs Ypt31/32. To test the ability of CPRabA5e to restore the growth defect of a ypt31Δ ypt32ts temperature sensitive double mutant (Zou et al. 2012) at high temperature (37 °C), the cDNA corresponding to the CPRabA5e gene was cloned into the yeast expression vector pYX213 and transformed into the yeast

13

Fig. 2  CPRabA5e functionally replaces Ypt31/32 but not Arf1 in yeast. a Heterologous complementation of yeast arf1Δ arf2Δ double mutant with CPRabA5e and empty yeast expression vector pYX242 (VC). Yeast strains were plated on the medium containing 5-fluoroorotic acid (FOA) to eliminate the yeast Arf1 plasmid (pURA3Arf1). WT represents the wild-type yeast strain that contains functional Arf1 and Arf2. b Growth of all strains in the medium without FOA. Heterologous complementation of the ypt31Δ ypt32ts mutant strain by CPRabA5e at normal growth temperature (30 °C) (c) and at high temperature growth (37 °C) (d). The ypt31Δ ypt32ts mutant strain transformed with yeast Ypt32 (as positive control; PC), the same mutant strain transformed with empty pYX213 vector (empty vector control; VC) as well as the untransformed strain (negative control; NC). e The amino acid sequence from A. thaliana CPRabA5e is more similar to the two yeast (S. cerevisiae) Ypt proteins (Ypt31, Yer031c and Ypt32, Ygl210w) than the three yeast Arf proteins (Arf1, Ydl192w; Arf2, Ydl137w and Arf3, Yor094w). The evolutionary history was inferred using the Neighbor-Joining method and the tree is drawn to scale with branch lengths as the same units of the evolutionary distances used to infer the phylogenetic tree

Plant Mol Biol

ypt31Δ ypt32ts mutant. As a positive control we employed Ypt32. CPRabA5e could restore the growth of the ypt31Δ ypt32ts mutant at high temperature (37 °C), while the strain containing the empty vector pYX213 used as vector control and the mutant strain itself as negative control were unable to grow at 37 °C (Fig. 2c, d). Thus, CPRabA5e is a functional ortholog of Ypt31/32. In addition, when comparing the amino acid sequence of CPRabA5e with the three yeast Arf proteins (Arf1, Ydl192w; Arf2, Ydl137w; Arf3, Yor094w), and the two Ypt proteins (Ypt31, Yer031c; Ypt32, Ygl210w) it clearly shows that CPRabA5e is more similar to the Ypt proteins than the Arf proteins (Fig. 2e). CPRabA5e is localized in stroma and thylakoid membranes To experimentally determine the localization of CPRabA5e, its coding sequence was fused to the N-terminus of the enhanced green fluorescent protein (EGFP) and transiently expressed in tobacco epidermal cells under the control of the 35S cauliflower mosaic virus (CaMV) promoter. EGFP signal was detected in stomata guard cells and was concordant with chlorophyll autofluorescence (Fig. 3a–c), indicating the association of CPRabA5e:EGFP with chloroplasts and thereby confirming the prediction by Andersson and Sandelius (2004) also supported by TargetP. To investigate the subplastidic localization of CPRabA5e, an antibody produced against a synthetic peptide of CPRabA5e was used. Subfractions of the chloroplast (stroma, envelope and thylakoid membranes) were isolated and tested for the presence of CPRabA5e (Fig. 3d). CPRabA5e was found in thylakoid membranes and in the stromal fraction but not in envelope membrane. Different protein markers were used to verify the purity of the fractions. As expected, Lhcb1 was exclusively found in thylakoid membranes, Toc75 in the envelope membranes and Rubisco in the stromal fraction. CPRabA5e is expressed from germinated seeds to mature siliques and green tissues throughout development We examined the expression of the CPRabA5e gene using publicly available microarray data and the Genevestigator v3 (Zimmermann et al. 2004; Grennan 2006). Expression data from all high-quality ATH1 (22 k) arrays were analysed (Hruz et al. 2008). The developmental expression analysis revealed that CPRabA5e is mostly expressed at the stages of germinated seeds, ca. 11.3 Units, and seedlings, ca. 10.7 Units, followed by developed rosette and developed flowers (Fig. 4). Altogether the expression level was at medium with only expression in mature siliques being considered low (Fig. 4).

Fig. 3  Subcellular localization of CPRabA5e. Subcellular targeting of the CPRabA5e:EGFP fusion protein in 4-week-old tobacco leaves. a GFP fluorescence, b chlorophyll autofluorescence, and c merged channels. Scale bars 15 μm. d Detection of CPRabA5e with CPRabA5e specific antibody in 2-week-old isolated Arabidopsis chloroplasts. 5 μg of chloroplast were loaded. For the detection of CPRabA5e in chloroplast subfractions, 10 μg were loaded in each lane (stroma and thylakoid subfractions), except for the envelope fraction, which was 5 μg per lane. Chloroplast and subfraction proteins were analysed by SDS-PAGE and immunodetected with antibodies against CPRabA5e, the thylakoid protein Lhcb2, the envelope membrane protein Toc75 and the stromal protein Rubisco

Loss of CPRabA5e affects seed germination and results in susceptibility to oxidative stress Two knockout mutants for CPRabA5e, cprabA5e-1 and cprabA5e-2, were collected from RIKEN Arabidopsis transposon (Ds lines) insertion mutant lines for genotypic and phenotypic analysis (Fig. 5a). PCR with gene and Ds-T-DNA specific primers showed that both cprabA5e-1 and cprabA5e-2 had T-DNA insertions disrupting the CPRabA5e gene and were homozygous for this mutation (Fig. 5b). RT-PCR confirmed the absence of expression of CPRabA5e in both mutant lines (Fig. 5c), and no CPRabA5e protein was detected in the mutant lines after immunoblotting using the antibody against CPRabA5e (Fig. 5d). No visible phenotype was observed in the mutants compared to wild type for 2-week-old plants grown under long day condition on MS media (Fig. 6a), or for 4-week-old plants first grown on MS media for 2 weeks and thereafter being transferred to soil and grown two additional weeks in short day conditions (Fig. 6b). The chlorophyll content was similar among the mutants and wild-type plants with values around 1.1–1.2 nmol/mg (Fig. 6c). The mutants showed no

13



Fig. 4  CPRabA5e accumulates preferably in seed and seedling stage. Developmental expression profile of CPRabA5e according to the publicly available Affymetrix GeneChip microarray data. Data were retrieved using the Genevestigator analysis tool (www.genevestigator.com/gv/plant.jsp) and prepared with the MetaProfile Analysis tool using development representations in scatterplot format. Data from all high-quality ATH1 (22 k) arrays were analysed (Hruz et al. 2008)

difference from wild-type plants regarding the photochemical efficiency of the PSII i.e. the Fv/Fm value, which was found to be similar around 0.81 (Fig. 6d). Interestingly, it was observed that cprabA5e-1 and cprabA5e-2 seeds had a delayed germination under long day conditions on MS media compared to wild-type seeds (Fig. 7). After 48 h less than 10 % of the mutants seeds had germinated compared to ca. 60 % for wild-type seeds. The difference between the germination of the mutant and wild-type seeds diminished over time and after 6 days the germination efficiency of mutant seeds was almost 60 % of that of wild-type seeds, and after ca. 9–10 days the germination was homogenized in all lines, e.g. no difference could be observed. Some Rabs are found to be involved in oxidative stress responses in human and in plants (Li et al. 2010; Mazel et al. 2004). The cprabA5e-1 and cprabA5e-2 mutant lines showed growth arrest when grown in MS media containing 10 μM rose bengal (4,5,6,7-tetrachloro-2′,4′,5′,7′tetraiodofluorescein) known to generate intracellular

13

Plant Mol Biol

Fig. 5  Verification of two knockout lines of CPRabA5e. a Two RIKEN Arabidopsis transposon (Ds lines) insertion mutants of CPRabA5e, RATM11-6468-1_H (cprabA5e-1) and RATM53-2863-1 (cprabA5e-2) were collected with No-0 as wild-type (WT) background. The diagram shows the transposon insertion sites (triangles) of mutants to create the knockout mutants. Grey boxes represent two exons of CPRabA5e. b PCR analysis showed single PCR fragments from cprabA5e-1 and cprabA5e-2 genomic DNA while using CPRabA5e gene and inserted Ds-T-DNA specific primers but no fragments from upstream and downstream of the CPRabA5e gene specific primers suggesting homozygous condition in both lines for their knockout insertion. Corresponding fragment form the CPRabA5e related primers in wild-type genomic DNA was used as control. M indicates the DNA size marker. c RT-PCR analysis showed cprabA5e-1 and cprabA5e-2 lacking the expression of CPRabA5e while WT (No-0) showed the expression of CPRabA5e. Actin was detected in similar amounts in each lane, used as a control for equal loading. d CPRabA5e is detected in the chloroplast fraction of wild-type (WT) plants but not in CPRabA5e knockout mutants (cprabA5e-1 and cprabA5e-2). Western blot was performed using a CPRabA5e specific antibody and the name (CPRabA5e) points at the expected size of mature CPRabA5e being 24 kDa. Detection of CPRabA5e was performed in 5 μg of chloroplast per lane for all the lanes (samples)

singlet oxygen (Fig. 8). Compared to wild-type plants, both mutants were clearly more affected by the oxidative stress. Chloroplast ultrastructure of the cprabA5e mutants Chloroplast ultrastructure was basically similar in the 3-week-old cprabA5e mutant and wild-type plants under normal growth conditions at 20 °C, with plastoglobules

Plant Mol Biol

Fig. 6  Morphology and analysis of photosynthetic parameters of cprabA5e mutants. Homozygous cprabA5e-1 and cprabA5e-2 mutant plants were grown side-by-side together with wild-type (WT) plants on MS medium for 14 days. a Similar plants were grown in vitro for 14 days, and then transferred to soil and grown until they were 28 days old. b Representative plants are shown in each case. c Chlorophyll contents in 14-day-old, in vitro-grown plants of the indicated

Fig. 7  Seed germination. Rate of germination in CPRabA5e knockout mutant (cprabA5e-1 and cprabA5e-2) seeds and in wild-type (WT) seeds under normal growth condition. The values shown are means (±SD) derived from six independent measurements

being only slightly larger in the mutants than in the wildtype plants (Fig. 9a). The number of plastoglobules per plastid section varied between 13 and 16 and was not significantly different between the mutants and the wildtype plants, and was also not significantly influenced by any treatment. No vesicles were observed in the plastids and granum height was also similar (average number of appressed thylakoid membranes per grana being equal to 5 in all plants). When the plants were preincubated for 1 h at 4 °C before sampling, the size of plastoglobules increased when

genotypes were determined photometrically. Values shown are means (±SD) derived from five independent samples, each sample containing ten different plants. Units are nmol chlorophyll a  +  b per mg fresh weight. d Photosystem II photochemical efficiency (Fv/Fm) values were determined by measuring chlorophyll fluorescence in similar plants grown on soil for 4 weeks. The values shown are means (±SD) derived from nine independent measurements

Fig. 8  Oxidative stress in cprabA5e mutants. 4-week-old CPRabA5e knockout mutants (cprabA5e-1 and cprabA5e-2) and wild-type (WT) plants grown on MS media under long day (16 h) light conditions (−rs, left panel) and on MS media containing 10 μM rose bengal (+rs, right panel), which generates intracellular singlet oxygen to induce oxidative stress on plants

compared to those found at 20 °C, and they were significantly larger in the plastids of 3-week-old cprabA5e mutants than those of wild-type plants (average plastoglobule diameter: 360 vs. 150 nm, respectively) (Fig. 9a). At 4 °C, the plastoglobules of the mutants became more electron-transparent than those observed at 20 °C and in the wild-type plants (Fig. 9a). This may indicate altered molecular composition of the plastoglobules in the mutants at 4 °C. In addition, sometimes vesicles of 30–70 nm diameter

13



Plant Mol Biol

Fig. 9  Chloroplast ultrastructure of cprabA5e mutant and wildtype plants. a Transmission electron micrographs of chloroplasts of 3-week-old CPRabA5e knockout mutants (cprabA5e-1 and cprabA5e-2) and wild-type (WT) plants grown at 20 °C and incubated for 1 h at 20 °C or at 4 °C prior to sample collection. Scale bar 1 μm. b Transmission electron micrographs of chloroplasts of rose bengal

treated 3-week-old CPRabA5e knockout mutants (cprabA5e-1 and cprabA5e-2) and wild-type (WT) plants incubated for 1 h at 4 °C prior to sample collection. Vesicles are indicated by arrowheads and shown also in an inset for cprabA5e-2 (right panel). Scale bar 1 μm, except scale bar within the inset being 0.5 μm

could be observed in the stroma close to the envelope membrane in the chloroplasts of both mutants at 4 °C. These structures were almost absent in the wild-type plant chloroplasts. At 4 °C, granum height was significantly lower in the mutants than in the wild-type plants (4 vs. 5 appressed thylakoid membranes per grana on average). Under oxidative stress conditions (i.e. in the presence of rose bengal), the size of plastoglobules decreased in all plants when compared to unstressed plants at 4 °C and was slightly smaller in the mutants (average diameter 55 nm) than in the wild-type plants (average diameter 75 nm) (Fig. 9b; Table 1). Vesicles were more frequent in the plastids of all plants under the oxidative stress conditions when compared to normal growth conditions (no vesicles) and even low temperature preincubation at 4 °C, especially in the mutants, where vesicles were characteristic for almost all plastid profiles. The average number of appressed thylakoid membranes in grana (i.e. granum height) decreased

significantly in all plants from 4 to 5 appressed thylakoids present under normal growth conditions and in low temperature preincubated samples to 3–4 thylakoids under oxidative stress, and again grana were slightly but significantly lower in the mutants when compared to the wild-type plants (3 vs. 4 appressed thylakoid lamellae per grana on average, respectively, Table 1). Almost no starch was present in the plastids under normal growth conditions while starch grains were frequently observed in the plastids under oxidative stress conditions. To summarize, the CPRabA5e mutation affects the size of plastoglobules, granum height and vesicle accumulation in the chloroplast.

13

Identification of CPRabA5e interacting partner proteins through yeast two‑hybrid screening CPRabA5e in a GTP hydrolysis assay confirmed its function as an intrinsic GTPase (Fig. 1), and worked as a

Plant Mol Biol Table 1  Quantitative data related to chloroplast ultrastructure of 3-week-old CPRabA5e knockout mutants (cprabA5e-1 and cprabA5e-2) and wild-type (WT, No-0) plants under normal growth conditions, after low temperature preincubation and under oxidative stress conditions Treatment

Sample name

20 °C

Wild-type

4 °C

60 ± 15a a

Number of PGs per plastid section

Number of thylakoid lamellae per granum

13.8 ± 5.3a 15.7 ± 6.2a

4.9 ± 2.1a 5.3 ± 3.3a

cprabA5e-1

67 ± 19

cprabA5e-2 wild-type

121 ± 36b

12.8 ± 4.8a

4.8 ± 2.3a

b

116 ± 55 357 ± 100c

a

13.0 ± 4.9 15.0 ± 5.4a

4.9 ± 2.5a 4.4 ± 1.9b

170 ± 68d

12.8 ± 3.8a

4.1 ± 1.6b

a

cprabA5e-1

76 ± 32 54 ± 17e

a

14.9 ± 5.4 13.7 ± 4.9a

4.0 ± 1.3b 3.4 ± 1.5c

cprabA5e-2

53 ± 14e

15.4 ± 6.5a

3.1 ± 0.8c

cprabA5e-1 4 °C, rose bengal

Diameter of PG (nm)

cprabA5e-2 Wild-type

Plants were grown at 20 °C and incubated for 1 h at 20 °C or at 4 °C prior to sample collection. Rose bengal treated plants were grown 2 weeks on normal growth media and then 1 week with rose Bengal, thereafter incubated for 1 h at 4 °C prior to sample collection. Values shown are means (±SD) of average data from two or three leaves from different plants, and were measured on 50–90 randomly chosen plastid sections in each sample. The diameter of around 250–400 plastoglobules (PGs) were determined per sample, granum height (defined as the number of appressed thylakoid lamellae per one granum) was determined and averaged for 300–350 grana per sample. Different superscript letters indicate significant difference at P