SUMMARY
Parasite effector proteins target various host cell compartments to alter host processes and promote infection. How effectors cross membrane-rich interfaces to reach these compartments is a major question in effector biology. Growing evidence suggests that effectors use molecular mimicry to subvert host cell machinery for protein sorting. We recently identified CTP1 (chloroplast-targeted protein 1), a candidate effector from the poplar leaf rust fungus Melampsora larici-populina that carries a predicted transit peptide and accumulates in chloroplasts. Here, we show that the CTP1 transit peptide is necessary and sufficient for accumulation in the stroma of chloroplasts, and is cleaved after translocation. CTP1 is part of a Melampsora-specific family of polymorphic secreted proteins whose members translocate and are processed in chloroplasts in a N-terminal signal-dependent manner. Our findings reveal that fungi have evolved effector proteins that mimic plant-specific sorting signals to traffic within plant cells.
INTRODUCTION
Parasitic microbes deliver effector proteins into host cell compartments to manipulate a variety of processes and promote parasitic colonization (Dodds and Rathjen, 2010; Win et al., 2012). One driving theme in plant pathology is to uncover how effector proteins traffic to these compartments; most notably how they cross a multitude of membrane-bound biological interfaces. Bacterial effectors cross the host plasma membrane through a complex molecular syringe (or secretion system), whereas effectors of filamentous parasites, namely fungi and oomycetes, enter host cells via mechanisms that remain to be discovered (Petre and Kamoun, 2014). Once inside host cells, some effectors subsequently traffic to distinct compartments including nuclei, plasma membrane, endoplasmic reticulum, tonoplast, vesicles, chloroplasts and mitochondria (Lindeberg et al., 2012; Deslandes and Rivas, 2012; Hicks and Galan, 2013). To do so, several effectors possess domains that mimic host-targeting (or addressing) sequences. For example, effectors targeted to host nuclei through nuclear pores carry nuclear-localisation signals (NLS) (Schornack et al., 2010; Wirthmueller et al., 2014). However, how pathogens target effectors to membrane-bound organelles such as chloroplasts is poorly understood.
Chloroplasts are plant cell organelles composed of a double membrane envelope that surrounds a fluid called stroma, in which structures such as thylakoids and starch granules reside. The envelope can form elongated structures called stromules (stroma-containing tubules), which are thought to improve exchange of molecules with other compartments (Krenz et al., 2012). Chloroplasts selectively internalize proteins from the cytosol via the TOC/TIC (Translocon at Outer membrane of Chloroplasts/Translocon at Inner membrane of Chloroplasts) complex, by recognizing transit peptides that are present at the N-termini of proteins (Shi and Theg, 2013). Upon translocation, transit peptides are typically cleaved off by processing peptidases and mature proteins are released in the stroma (Texeira and Glaser 2013). Transit peptides are variable in term of amino acid sequence and length, and no consensus has been identified (Demarsy et al., 2014).
To date, only a few plant pathogen effectors have been reported to target chloroplasts (Jelenska et al., 2007; Rogriguez-Herva et al., 2012; Li et al., 2014; Petre et al., 2015). HopI1, HopN1, HopK1 and AvrRps4 are type-III effector proteins of the bacterial plant pathogen Pseudomonas syringae pv tomato that function as virulence factors by modulating chloroplast structures and functions. How HopI1 and HopN1 translocate into chloroplasts is unknown, as they lack a N-terminal transit peptide. In contrast, HopK1 and AvrRps4 carry N-terminal cleavable regions that are sufficient for import into chloroplasts, suggesting that these proteins carry transit peptides (Li et al., 2014). However, whereas both proteins accumulate in chloroplasts in stable A. thaliana transgenics, they diffuse in the cytosol and the nucleus when transiently expressed in N. benthamiana leaf cells (Li et al., 2014). Besides, several studies have reported AvrRps4 in the nucleus and cytosol rather than, or in addition to chloroplasts (Heidrich et al., 2011; Bhattacharjee et al., 2011; Sohn et al., 2012).
Rust fungi (Basidiomycetes, Pucciniales) are obligate biotrophic parasites of plants (Duplessis et al., 2015). Several rust species are devastating pathogens of crops, and a constant threat to agrosystems worldwide (Pennisi 2010). For instance, the leaf rust fungus Melampsora larici-populina triggers annual epidemics on cultivated poplars and dramatically reduces plantations yield (Gérard et al. 2006). We recently took advantage of the predicted effector complement of M. larici-populina to perform an expression screen of the most promising candidates in leaf cells of the model plant Nicotiana benthamiana. These analyses revealed that the effector MLP107772, renamed chloroplast-targeted protein 1 (CTP1) in the present study, accumulates in chloroplasts (Petre et al., 2015). Interestingly, CTP1 carries a predicted transit peptide of 83 amino acids at its N-terminus, which may explain its ability to translocate into chloroplasts (Petre et al., 2015).
In this study, we further exploited the N. benthamiana experimental system to study effector trafficking across the chloroplast envelope. We combined genetic transformation with live cell imaging, cell fractionation and biochemical analyses to demonstrate that CTP1 resides in the stroma of chloroplasts. We then showed that the predicted CTP1 transit peptide is necessary and sufficient for accumulation in chloroplasts, and is cleaved upon translocation. CTP1 is a member of a Melampsora-specific effector family. We also provide evidence that other members of the family also enter chloroplasts in a N-terminal-dependent manner. Our data indicate that rust fungi have evolved a class of effector proteins that translocate into chloroplasts by functionally mimicking plant-targeting sequences.
RESULTS
CTP1-mCherry accumulates in the stroma of chloroplasts and in mitochondria
We recently reported that the mature form of CTP1 fused to a GFP (CTP128-171-GFP) accumulates inside chloroplasts and mitochondria when it is heterologously expressed in N. benthamiana leaf cells (Petre et al., 2015). To test the robustness of this observation, we constructed a fusion protein with a different fluorescent tag (CTP128-171-mCherry) and assayed its localization in N. benthamiana leaf cells by confocal microscopy. CTP128-171-mCherry fluorescent signal overlapped with chlorophyll autofluorescence, but also labeled stromules and mitochondria-like bodies (Figure 1). Co-expression of CTP128-171-mCherry and ScCOX4-GFP, a marker of mitochondria (Nelson et al., 2007), showed overlapping signals, confirming the presence of CTP128-171-mCherry in mitochondria (Figure S1). The co-expression of CTP128-171-GFP and CTP128-171-mCherry revealed perfectly overlapping fluorescent signals, indicating that the nature of the fluorescent protein does not affect the localization. Saturation of the fluorescent signals revealed no accumulation in the cytosol and the nucleus, indicating that CTP128-171-fluorescent protein fusions are probably internalized into these organelles (Figure S2).
To unambiguously determine whether CTP1 translocates inside chloroplasts, we performed a series of co-localizations of CTP128-171-fluorescent protein fusions with various markers of sub-chloroplastic compartments. First, we co-expressed CTP128-171-GFP with NbpMSRA-mCherry, a marker of the stroma of chloroplasts (Lorrain et al., 2014a). In this assay, GFP and mCherry fluorescent signals overlapped perfectly, indicating that CTP1 was present in the stroma (Figure 2A). Second, co-expression of CTP128-171-GFP with a GBSSI-YFP fusion (a starch granule marker; Wang et al., 2013), allowed a better visualization of chloroplasts at high magnification. The uniform distribution of the GFP fluorescent signal around GBSSI-YFP labeled starch granules further confirmed the accumulation of CTP128- 171-GFP in the stroma (Figure 2B, Figure S2). Finally, we used cell fractionation to isolate chloroplasts from N. benthamiana leaves expressing CTP128-171-GFP. Confocal microscopy performed on purified organelles revealed a strong fluorescent signal in some chloroplasts, which was uniformly distributed around the starch granules (Figure 3). These experiments indicate that CTP1 enters the chloroplasts to accumulate within the stroma.
CTP1 carries a functional chloroplast transit peptide
We hypothesized that the predicted 83-amino-acid N-terminal transit peptide of CTP1 mediates trafficking to chloroplasts and mitochondria (Petre et al., 2015; Figure 4A). To test the functionality of this predicted transit peptide, we generated two truncations: CTP128-114 (the predicted transit peptide plus four amino acids), and CTP1107-171 (CTP1 without its predicted transit peptide). We fused the truncations with GFP (CTP128-114-GFP and CTP1107- 171-GFP) and determined their subcellular localization in N. benthamiana leaf cells. CTP128- 114-GFP signal overlapped with chlorophyll, but also showed a weak signal in the cytosol and the nucleus (Figure 4B). We used CTP128-171-mCherry as a marker of both the stroma and mitochondria, and co-expressed it with CTP128-114-GFP. Fluorescent signals overlapped in the stroma but not in mitochondria, suggesting that CTP128-114-GFP accumulates only in chloroplasts. Further observations performed on chloroplasts isolated from leaves expressing CTP128-114-GFP confirmed the presence of the fusion protein in the stroma (Figure 4C). On the other hand, CTP1107-171-GFP signal was not detected in chloroplasts or mitochondria, and showed a nucleocytosolic distribution (Figure 4D). This indicates that CTP1 lacking its transit peptide can no longer enter these organelles. In stable N. benthamiana transgenics, CTP128- 114-GFP fluorescent signal specifically labeled chloroplasts and stromules, validating the observations made in transient assays (Figure 4E). We conclude that CTP1 carries a functional transit peptide, which is necessary and sufficient to mediate trafficking to chloroplasts. However, this transit peptide is necessary but not sufficient for accumulation in mitochondria.
CTP1 is part of a Melampsora-specific family of modular and polymorphic proteins
CTP1 was first described as an orphan gene in M. larici-populina (Hacquard et al., 2012; Saunders et al., 2012). However, a more recent analysis reported that CTP1 displays sequence similarity to other secreted proteins of M. larici-populina (Cantu et al., 2013). To establish a list of proteins related to CTP1, we performed sequence similarity searches as well as protein and gene structure similarity analyses. We identified four proteins with sequence similarity to CTP1: Mlp52246 (hereafter CTP2), Mlp55690 and Mlp116083 in M.larici-populina, and Melli_sc2834 (hereafter CTP3) in the flax rust fungus Melampsora lini. The four proteins showed amino acid similarity to CTP1 that ranged from 16 to 28%. Although the sequence similarity was low, these five proteins display a similar modular protein structure. They all have a predicted signal peptide followed by a predicted transit peptide of variable length, as well as two to three repeated domains or modules, with the first one overlapping with the predicted transit peptide (Figure 5A). Also, gene structures matched protein organization, with the first exon encoding the signal peptide, and each of the other exons encoding a module. These modules are highly divergent (average paired amino acid identity below 30%, ranging from 15 to 100%) but present stretches of amino acids with similar properties and four conserved cysteine residues (Figure 5B). Interestingly, the phylogeny of the modules does not fully match the gene organization (Figure 5B), suggesting that the CTP family evolved potentially by exon shuffling. We conclude that CTP1 is part of a family of secreted, modular proteins that have probably emerged and diversified recently in Melampsora species.
CTP2 and CTP3, two proteins similar to CTP1, accumulate in chloroplasts
To determine whether CTP1-like proteins also target chloroplasts and mitochondria, we studied CTP2 and CTP3, two proteins representative of the diversity of the family, and which carry two and three modules, respectively. We generated fusions between the mature forms of CTP2 and CTP3 (i.e. without the signal peptide) with GFP or mCherry and expressed the fusion proteins in N. benthamiana. Confocal microscopy revealed that the two fusion proteins display a signal that overlapped with chlorophyll. CTP2 fluorescent signal accumulated in the stroma of the chloroplasts but also in the nucleus and cytosol (Figure 6A, Figure S4). CTP3 fluorescent signal distributed similarly to CTP2, but also labeled discrete areas of approximately 0.1 to 0.3 μm of diameter within the chloroplasts (Figure 6B, Figure S5A and S5B). A Populus trichocarpa Coproporphyrinogen III Oxidase-GFP fusion (PtCPO-GFP) was previously observed to accumulate in similar sub-chloroplastic structures in N. benthamiana (Lorrain et al., 2014b). Co-expression of CTP328-260-mCherry and PtCPO-GFP revealed overlapping signals, indicating that CTP328-260-fluorescent protein fusions reside in PtCPO-containing compartments whose nature remains to be determined (Figure S5C). Co-expression of CTP229-185- and CTP328-260-fluorescent protein fusions with mitochondria markers (ScCOX4-GFP or ScCOX4-mCherry) revealed no overlapping signals, suggesting that CTP2 and CTP3 do not accumulate in mitochondria (Figure S4, Figure S5A). We conclude that CTP2 and CTP3 translocate into chloroplasts when expressed in N. benthamiana cells.
CTP1, CTP2 and CTP3 are processed in chloroplasts
Upon translocation in chloroplasts and mitochondria, processing peptidases cleave off transit peptides (Teixeira and Glaser, 2013). To determine whether CTPs undergo processing in planta, we performed anti-GFP western blotting with total proteins isolated from benthamiana leaves expressing the CTP-GFP fusions. The band signals for all fusion proteins were smaller than their theoretical size, suggesting that they undergo cleavage in planta. Indeed, CTP128-171-GFP showed a unique band signal at 34 kDa, approximately 8 kDa below its expected size of 41.8 kDa (Figure 7), confirming previous observations (Petre et al., 2015). Both CTP229-185-GFP and CTP328-260-GFP displayed a band signal that coincided with their predicted molecular weight at 42.7 and 51.0 kDa, respectively, but also a second band signal that was 8 and 16 kDa smaller, respectively. The size of all these band signals coincide with the predicted molecular weight of the last repeat module of each protein fused to GFP, which suggests that cleavage events occur between modules A and B (CTP1 and CTP2) and between modules B and C (CTP3) (Figure 5).
To determine whether processing events occur specifically in chloroplasts, we generated CTP fusions that are expected to interfere with chloroplast translocation. Since the addition of N-terminal amino acid residues is known to abolish the functionality of transit peptides (Carrie et al., 2009), we constructed N-terminally GFP tagged fusion proteins (GFP-CTP128-171, GFP-CTP229-185 and GFP-CTP328-260) and expressed them in N. benthamiana leaves. We did not detect fluorescent signal in chloroplasts with the three fusion proteins, which accumulated only in the nucleus and the cytosol (Figure S6). Each of the three fusion protein yielded a single band signal around its predicted molecular weight with anti-GFP western blot analyses, demonstrating the absence of processing in the nucleus and the cytosol (Figure 7). From this set of experiments, we conclude that CTP1, CTP2 and CTP3 translocate in chloroplasts in a N-terminal signal-dependent manner, and that they undergo processing inside chloroplasts.
DISCUSSION
The N-terminus of the rust fungal protein CTP1 is necessary and sufficient to enter the chloroplast organelle of plants indicating that it mimics host transit peptides to subvert host cell machinery. Host mimicking domains are frequent in pathogen effectors (Hicks and Galan, 2013). However, how did a chloroplast transit peptide evolve in an organism that lacks chloroplasts? It should be noted that mitochondria and chloroplast transit peptides have similar properties (Carrie and Small 2013). The CTP1 transit peptide may have evolved from a mitochondrial transit peptide endogenous to the fungus. The observation that CTP1 can accumulate in mitochondria somewhat supports this hypothesis although the transit peptide does not fully account for this (Figure 4). On the other hand, the transit peptide may have evolved de novo. Transit peptides typically do not display conserved amino acid sequences, and their functionality is based on conserved physical properties (Bruce 2001). Therefore, they are more likely to evolve de novo compared with other domains whose function is based on strict amino acid sequence conservation (Tonkin et al., 2008). This possibility is perhaps more likely given that CTP1 does not share sequence similarity to any known protein sequence, and belongs to a polymorphic and lineage-specific family. Hence, the CTP1 transit peptide may have evolved de novo possibly by neofunctionalization after a duplication or recombination event.
Host-translocated effectors of oomycete and fungal plant pathogens usually carry host-targeting sequences just downstream of their signal peptide (Petre and Kamoun 2014). These targeting sequences have no known function in host cells. We have shown that the N-terminal region immediately downstream of CTP1 signal peptide functions as transit peptide in plant cells. This finding implies that, if CTP1 encodes a N-terminal host-targeting sequence, it overlaps with its transit peptide. It is also possible that CTP1 does not require a host-targeting sequence.
CTP1 accumulates in both chloroplasts and mitochondria. Although accumulation in mitochondria is relatively weak, CTP1 could very well target the two organelles. Many plant proteins have evolved to target and function in two cell compartments, and are called ‘dually-targeted’ proteins (Carrie and Whelan 2013). For instance, over a hundred plant proteins target both chloroplasts and mitochondria (Carrie and Small 2013). Some pathogen effectors have also been reported in multiple host cell compartments (Heidrich et al., 2011; Hicks and Galan, 2013; Li et al., 2014) and may also be dually-targeted proteins. However, it has to be noted that disparate locations observed by different authors could be artifactual, for instance due to the use of different tags or methods of localization. In rust fungi, several members of the RTP1 family have been immunolocalised in various compartments within the plant cell and at the host-pathogen interface (Kemen et al., 2005; Hacquard et al., 2012; Kemen et al., 2013).
CTP1, CTP2 and CTP3 display various patterns of accumulation in chloroplasts and mitochondria. Plant proteins evolve to form large families, whose members diversify to target and function in different cell compartments (Rouhier et al., 2006). The same holds true for effectors of plant parasites. For instance, several members of the Avrblb2 family of Phythophtora infestans accumulate in the nuclei, the cytosol or at the plasma membrane (Bozkurt et al., 2011; Zheng et al., 2014). Also, effector alleles can target different compartments (Leonelli et al., 2011; Caillaud et al., 2012). Whether effector variants carry similar or different functions in distinct cellular compartments remains to be determined. In plants, members of large protein families often carry similar biochemical activities, although they are performed with substrates or protein partners that are specific to a given compartment (Rouhier et al., 2006).
We have shown that three candidate effectors from Melampsora species accumulate in chloroplasts (and to some extent in mitochondria) in N. benthamiana leaf cells. Growing evidence indicates that effectors from unrelated parasites converge towards chloroplasts and mitochondria (Win et al., 2012; Hicks and Galan 2013; Li et al., 2014; Petre et al., 2015). This implies that manipulating chloroplasts and mitochondria is crucial for parasites to colonize host plants. Surprisingly, to date no oomycete effectors have been reported in these organelles, despite several large-scale subcellular localisation screens (Schornack et al., 2010; Caillaud et al., 2012; Stam et al., 2013; Win et al., unpub). However, all these studies used N-terminally tagged effectors, which may have impaired the proper translocation of the proteins into the organelles, resulting in false negatives.
The ability of CTPs to enter chloroplasts suggests that they have a function inside this organelle. What could be that function? CTPs display no sequence similarity to known proteins or protein domains, which prevents the prediction of a putative function based on their primary structure. The only similar feature among the CTPs is the four fully conserved cysteine residues within repeated modules. Such patterns of conserved cysteine residues have been observed previously for several families of secreted proteins in M. larici-populina (Duplessis et al., 2011; Hacquard et al., 2012). Repeated domains with conserved residues are common within effectors (Kloppholz et al., 2011; Mak et al., 2012; Chou et al., 2011) and can provide conserved folds that promote rapid evolution (Win et al., 2012). The cysteines could have different roles, for instance by being involved in intra- or extra-molecular disulfide bridges or by having a catalytic activity. Importantly, cysteine residues are also conserved in the repeats that overlap with transit peptides. It is tempting to speculate that repeats could be independently functional, and that the ones at the N-terminus may function as targeting sequences in addition to other activities. Importantly, CTPs are highly polymorphic and accumulate in various locations, which suggests that they may have different activities. One way to proceed further in elucidating CTP1’s function will be to evaluate the relevance of plant protein interactors identified previously (Petre et al., 2015).
EXPERIMENTAL PROCEDURES
Biological material and growth conditions
Escherichia coli (Subcloning Efficiency DH5α Competent Cells; Invitrogen, Carlsbad, California, USA) and Agrobacterium tumefaciens (electrocompetent strain GV3101) were conserved at -80°C and grown in L medium at 37°C and 28°C, respectively. N. benthamiana plants were grown in greenhouses at 22°C under 16/8h day/night conditions. The poplar hybrid Beaupré (P. trichocarpa × Populus deltoides) was propagated in greenhouse as previously described (Rinaldi et al., 2007). Urediniospores of Melampsora larici-populina (isolate 98AG31) were conserved at -80°C and propagated as previously described (Rinaldi et al., 2007). Beaupré leaf infections by M. larici-populina (isolate 98AG31, virulent on Beaupré) as well as M. larici-populina urediniospores germination in vitro were performed as previously described (Rinaldi et al., 2007).
Sequence analysis
Protein sequences were retrieved as follow: M. larici-populina (http://genomeportal.jgi-psf.org.html; Duplessis et al., 2011), M. lini (Nemri et al, 2014), Puccinia striiformis f sp tritici and Puccinia graminis f sp tritici (Cantu et al, 2013), N. benthamiana (http://solgenomics.net/; Bombarely et al., 2012) and P. trichocarpa (http://www.phytozome.net/poplar; Tuskan et al., 2006). Protein sequence analyses were performed with ClustalX and Jalview softwares. Protein targeting sequences were predicted with SignalP 3.0 (http://www.cbs.dtu.dk/services/SignalP-3.0/), TargetP 1.1 (http://www.cbs.dtu.dk/services/TargetP/), ChloroP 1.1 (http://www.cbs.dtu.dk/services/ChloroP/) and Wolf P Sort (http://psort.hgc.jp/). Protein parameters were calculated with Protparam (http://web.expasy.org/protparam/).
Cloning procedures and plasmids
The open reading frame (ORF) encoding CTP128-171 and CTP229-185 were amplified by polymerase chain reaction (PCR) using cDNA from Beaupré leaves infected by M. larici-populina (isolate 98AG31; Petre et al., 2012), with primers designed to generate BbsI AATG/ggTTCG compatible overhangs. The ORF of CTP328-260 was obtained through gene synthesis (Genewiz, London, UK), with codon optimization for plant expression and removal of internal BbsI and BsaI restriction sites. Sequences suitable for N-terminal tagging were generated by PCR amplifications with primers designed to insert BbsI AATG/GGCT compatible overhangs. Truncated sequences of CTP1 suitable for C-terminal tagging were generated by PCR amplifications with primers designed to insert BbsI AATG/ggTTCG compatible overhangs. DNA fragments were inserted into golden gate level 0 vectors pICSL01005 (BbsI compatible overhangs AATG/ggTTCG) or pICH41308 (BbsI compatible overhangs AATG/GGCT) following a digestion/ligation procedure (Weber et al., 2011; Engler et al., 2013; http://synbio.tsl.ac.uk/). Sequences of DNA fragments obtained by PCR were verified by sequencing. To generate C-terminally tagged proteins, the ORF of the protein of interest (BsaI compatible overhangs AATG/ggTTCG) was assembled with the ORF of a mCherry or a GFP (BsaI compatible overhangs ggTTCG/GCTT) into the golden gate level 1 binary vector pICH86988 (BsaI compatible overhangs AATG/GCTT; 35S promoter/OCS terminator). To generate N-terminally tagged proteins, the ORF of the protein of interest (BsaI compatible overhangs AATG/GCTT) was assembled with a 35S promoter (BsaI compatible overhangs GGAG/CCAT), the ORF of a mCherry or a GFP (BsaI compatible overhangs CCAT/AATG) and an octopine synthase (OCS) terminator (BsaI compatible overhangs GCTT/CGCT) into the golden gate level 1 vector pICH47742 (BsaI compatible overhangs GGAG/CGCT). Fusion protein sequences are shown in Table S1. Plasmids were multiplied using E. coli. Level 1 vectors were inserted in A. tumefaciens strain GV3101, and transformed bacteria were conserved at -80°C in 20% glycerol.
Expression of proteins in N. benthamiana
A. tumefaciens strain GV3101 was used to deliver T-DNA constructs into leaf cells of three-week-old N. benthamiana plants, following the agroinfiltration method previously described (Win et al., 2011). Briefly, overnight-grown bacterial cultures were adjusted at an OD600 of 0.2 into an infiltration buffer (10 mM MgCl2, 150 μM acetosyringone) one hour before leaf infiltration. For all co-transformations, A. tumefaciens strains were mixed in a 1:1 ratio in infiltration buffer to a final OD600 of 0.2. The leaves were collected two days after infiltration for further cell fractionation, protein isolation or microscopy. Stable N. benthamiana transgenics were obtained by A. tumefaciens mediated transformation (same strain as used for agroinfiltration), selected on Kanamycin and genotyped by PCR.
Chloroplasts isolation
Three agrotransformed N. benthamiana leaves were collected two days after infiltration, cut into 1 cm2 pieces and incubated in 50 ml cold Isolation Buffer (IB; 400 mM Sorbitol, 50 mM HEPES KOH 8M, 2 mM EDTA) for 15 min. Leaf pieces were homogenized with a PT 1200 C Polytron homogenizer (3 × 5 sec full speed), and the lysate was immediately filtrated onto doubled Miracloth to remove cell debris. The filtrate was collected in 50 mL conical tubes and centrifuged at 3,000 rpm for 5 min at 4°C. The pellet containing the organelles was resuspended in 1 mL of cold IB and chloroplasts were isolated by centrifugation on a Percoll gradient (2 mL 80 % v/v Percoll/IB; 6 mL 40 % v/v Percoll/IB in a 15 mL conical tube) at 4,500 rpm for 10 min at 4°C. The lower green band corresponding to intact chloroplasts was collected by pipetting and immediately used for further experiments.
Live-cell imaging by laser-scanning confocal microscopy
Small pieces of leaves (inferior face towards the objective) or solutions of isolated chloroplasts were mounted in water between a slide and a coverslip and immediately observed. Live-cell imaging was performed with a Leica DM6000B/TCS SP5 laser-scanning confocal microscope (Leica microsystems, Bucks, UK), using 10x (air) and 63x (water immersion) objectives. The GFP and the chlorophyll were excited at 488nm, the YFP was excited at 514 nm, and the mCherry was excited at 561 nm. Specific emission signals corresponding to the GFP, the YFP, the mCherry and the chlorophyll were collected between 505-525 nm, 530-550 nm, 580-620 nm and 680-710 nm, respectively, except where otherwise stated. Scanning was performed in sequential mode when needed. Image analysis was performed with Fiji (http://fiji.sc/Fiji).
Protein isolation and western blotting experiments
N. Benthamiana leaves were harvested two days after infiltration, frozen in liquid nitrogen and grinded into powder with mortar and pestle. Total protein extraction was performed by reducing and denaturing proteins form the leaf powder 10 minutes at 95°C in Laemmli buffer (1 M pH 6.8 Tris-HCL, 10 mM dithiothreitol [DTT], 2% SDS, 20% glycerol) in order to avoid in vitro non-specific degradation of the fusion proteins. Then, five to twenty microliters of isolated proteins were separated by 15% SDS-PAGE, and protein content was estimated by coomassie blue staining. Western blotting experiments were performed as previously described (Pégeot et al., 2014; Bozkurt et al., 2014). The following antibodies were used: rabbit anti-GFP (Invitrogen), goat anti-rabbit IRW800 (LI-COR Biosciences) and GFP (B2): sc-9996 (Santa-Cruz Biotechnology). Membrane revelation was carried out on an Odyssey infrared imager (LI-COR Biosciences) or on an ImageQuant LAS 4000 luminescent imager (GE Healthcare Life Sciences).
ACKNOWLEDGMENTS
We thank Y. Dagdas (TSL, Norwich, UK) and T. Bozkurt (Imperial College, London, UK) for critical reading of an early version of that manuscript. We thank P. Dodds (CSIRO, Canberra, Australia) for discussions. B. Petre was supported by an INRA Contrat Jeune Scientifique (CJS), by the European Union, in the framework of the Marie-Curie FP7 COFUND People Programme, through the award of an AgreenSkills’ fellowship (under grant agreement n° 267196) and by the laboratory of excellence (Labex) ARBRE, through the award of a mobility grant (12RW53). D. Saunders was supported by Leverhulme early career fellowship and a fellowship in computational biology at TGAC, in partnership with the John Innes Centre, and strategically supported by BBSRC. C. Lorrain is supported by an INRA CJS and by the Labex ARBRE. S. Duplessis is supported by the French National Research Agency through the Labex ARBRE (ANR-12-LABXARBRE-01) and the Young Scientist Grant POPRUST (ANR-2010-JCJC-1709-01). Research at TSL is supported by the Gatsby Charitable Foundation and the BBSRC.