Sub-cellular localization and complex formation by aminoacyl-tRNA synthetases in cyanobacteria: Evidence for interaction of membrane-anchored ValRS with ATP synthase

ORIGINAL RESEARCH published: 06 June 2016 doi: 10.3389/fmicb.2016.00857 Frontiers in Microbiology | www.frontiersin.org 1June 2016 | Volume 7 | Article 857 Edited by: Wendy Schluchter, University of New Orleans, USA Reviewed by: Conrad Mullineaux, Queen Mary University of London, UK Michael Ibba, Ohio State University, USA José Manuel García-Fernández, Universidad de Córdoba, Spain *Correspondence: Ignacio Luque [email protected] Specialty section: This article was submitted to Microbial Physiology and Metabolism, a section of the journal Frontiers in Microbiology Received: 26 April 2016 Accepted: 23 May 2016 Published: 06 June 2016 Citation: Santamaría-Gómez J, Ochoa de Alda JAG, Olmedo-Verd E, Bru-Martínez R and Luque I (2016) Sub-Cellular Localization and Complex Formation by Aminoacyl-tRNA Synthetases in Cyanobacteria: Evidence for Interaction of Membrane-Anchored ValRS with ATP Synthase. Front. Microbiol. 7:857. doi: 10.3389/fmicb.2016.00857 Sub-Cellular Localization and Complex Formation by Aminoacyl-tRNA Synthetases in Cyanobacteria: Evidence for Interaction of Membrane-Anchored ValRS with ATP Synthase Javier Santamaría-Gómez1, Jesús A. G. Ochoa de Alda2, Elvira Olmedo-Verd1, Roque Bru-Martínez3and Ignacio Luque1* 1Instituto de Bioquímica Vegetal y Fotosíntesis, Consejo Superior de Investigaciones Científicas and Universidad de Sevilla, Seville, Spain, 2Facultad de Formación del Profesorado, Universidad de Extremadura, Cáceres, Spain, 3Department of Agrochemistry and Biochemistry, Faculty of Science, University of Alicante, Alicante, Spain tRNAs are charged with cognate amino acids by aminoacyl-tRNA synthetases (aaRSs) and subsequently delivered to the ribosome to be used as substrates for gene translation. Whether aminoacyl-tRNAs are channeled to the ribosome by transit within translational complexes that avoid their diffusion in the cytoplasm is a matter of intense investigation in organisms of the three domains of life. In the cyanobacterium Anabaena sp. PCC 7120, the valyl-tRNA synthetase (ValRS) is anchored to thylakoid membranes by means of the CAAD domain. We have investigated whether in this organism ValRS could act as a hub for the nucleation of a translational complex by attracting other aaRSs to the membranes. Out of the 20 aaRSs, only ValRS was found to localize in thylakoid membranes whereas the other enzymes occupied the soluble portion of the cytoplasm. To investigate the basis for this asymmetric distribution of aaRSs, a global search for proteins interacting with the 20 aaRSs was conducted. The interaction between ValRS and the FoF1 ATP synthase complex here reported is of utmost interest and suggests a functional link between elements of the gene translation and energy production machineries. Keywords: aminoacyl-tRNA synthetases, membrane-anchoring, cyanobacteria, FoF1 ATP synthase, thylakoids, CAAD INTRODUCTION A hallmark of life is the reclusion of molecules in individual entities named cells with an uneven distribution of components within the internal space. Internal organization of cells is complex and it is optimized for functioning. Components are confined to a particular cell structure or compartment either because such environment is best suited for its function, to make possible their interaction with other components (i.e., substrates, regulators) or to avoid spurious interactions with other molecules (Yeates et al., 2008; Shapiro et al., 2009; Montero Llopis et al., 2010). Compared to eukaryotic cells, bacteria are structurally simple as they lack a nucleus and organelles. Cyanobacteria are unique among bacteria in the sense that they posses an intracytoplasmic system Santamaría-Gómez et al. Valyl-tRNA Synthetase-ATP Synthase Interaction of membrane saculi named thylakoids, which act as a physical support for large protein complexes involved in energy production by photosynthesis, including photosystems I and II, cytochrome b6fand ATP synthase (Liberton and Pakrasi, 2008). Recent investigations have demonstrated that in some cyanobacteria components of the gene translation machinery, namely aminoacyl-tRNA synthetases, are also anchored to thylakoid membranes (Olmedo-Verd et al., 2011). Gene translation in the cytoplasm of bacteria occurs at 70S ribosomes through pairing the anticodons of charged tRNAs to complementary codons of mRNAs. Charged tRNAs or aminoacyl-tRNAs are produced by aminoacyl-tRNA synthetases (aaRSs), a family of enzymes that catalyze the ATP-dependent esterification of amino acids to the 3′acceptor end of cognate tRNAs (Ibba and Soll, 2000). Cells generally count with a set of twenty aaRSs, each one specific for its cognate amino acid. According to structural and conservation criteria, aaRSs are partitioned in two classes with distinct phylogenetic origins (Eriani et al., 1990; Ribas de Pouplana and Schimmel, 2001). Once produced, aminoacyl-tRNAs form a complex with GTPactivated elongation factor-Tu that delivers them to the A site of the ribosome (Simonetti et al., 2009). A striking trait of species of the phylum Cyanobacteria observed in no other living organism is the presence of catalytically-active aaRSs anchored to membranes (Luque et al., 2008; Olmedo-Verd et al., 2011). Membrane anchoring was particularly well demonstrated for the Valyl-tRNA synthetase (ValRS) of Anabaena sp. PCC 7120 (hereafter Anabaena), which harbors an idiosyncratic domain, termed CAAD, with a putative C-terminal coiled-coil and two transmembrane helices, that mediates its binding to the lipid bilayer (Luque et al., 2008; Olmedo-Verd et al., 2011; Luque and Ochoa de Alda, 2014). This domain was also observed in the ValRS of some other cyanobacterial species. Interestingly, in few other species, CAAD was found inserted either in GluRS, IleRS, or LeuRS (Luque et al., 2008). CAAD is homologous to a family of standalone membrane proteins of cyanobacteria and plant chloroplast, named Curt1, or CURT1, respectively (Armbruster et al., 2013; Luque and Ochoa de Alda, 2014). Cyanobacteria and plant chloroplasts share a common phylogenetic origin, they both perform oxygenic photosynthesis and show functional and structural similarities. Thylakoid membranes are distributed in the stroma of chloroplasts or the cytoplasm of cyanobacteria, which are topologically equivalent. In chloroplasts, CURT1 proteins are integral membrane proteins that localize at the edges of thylakoid grana saculi and are responsible for the pronounced curvature of the membrane at this position (Armbruster et al., 2013). Cyanobacteria thylakoids do not pile up forming grana, however, Arabidopsis CURT1A was able to partially complement a curt1 mutant of the cyanobacterium Synechocystis sp. PCC 6803, indicating some degree of functional conservation (Armbruster et al., 2013). CAAD mediates the specific targeting of ValRS to thylakoid membranes in Anabaena (Olmedo-Verd et al., 2011). Consistent with the absence of any ValRS paralog in Anabaena, tRNAValcharging activity is confined to the membrane fractions of cell extracts. Artificial removal of CAAD from Anabaena ValRS rendered the enzyme soluble but had little impact on its catalytic parameters, indicating that CAAD’s function is mostly structural (Olmedo-Verd et al., 2011). Based on these data, it was proposed that in the other species where ValRS was also observed to contain CAAD, the enzyme would localize in thylakoid membranes. Furthermore, it was assumed that in the species where GluRS, IleRS, or LeuRS contained such domain, these enzymes would share such subcellular localization, but this has not been demonstrated. Conversely, aaRSs lacking CAAD in Anabaena and other species were all assumed to be soluble, which for the vast majority of them still awaits empirical demonstration. Worth to investigate is the possibility that despite the absence of an apparent membrane-anchoring domain, some of these aaRSs may be indirectly attached to membranes through interaction with membrane proteins. It has been shown that aaRSs form stable or transient complexes with other proteins in a variety of organism, a phenomenon that appears frequent in eukaryotes and archaea and less common in bacteria, where aaRSs are thought to be dispersed in the cytoplasm (Mirande et al., 1982; Hausmann and Ibba, 2008; Godinic-Mikulcic et al., 2011; Raina et al., 2012; Laporte et al., 2014). Notwithstanding the recent demonstration of the anchoring of the ValRS to thylakoid membranes in Anabaena, the information on the sub-cellular localization of aaRSs in cyanobacterial cells is limited and a global picture of the distribution of the full complement of aaRSs is lacking. This information is fundamental to understand how the initial steps of translation occur in these organisms and whether they are confined to a particular location. The aim of this work was to determine the sub-cellular localization for all aaRSs in Anabaena and to investigate the biological basis for such subcellular distribution. As an approach for this latter goal, we undertook the identification of interacting partners for each aaRSs. In this work we present compelling evidence on the interaction of the membrane-anchored ValRS with the FoF1-ATP synthase complex in Anabaena and we identify proteins interacting with soluble aaRSs in this organism. MATERIALS AND METHODS Organisms and Growth Conditions Anabaena sp. PCC 7120 and derivative strains were grown under standard growth conditions [30◦C, continuous illumination (75 µE m−2s−1) and bubbled with a mixture of CO2and air (1% v/v) in BG11 medium (Rippka, 1988) supplemented with 10 mM NaHCO3]. For the preparation of solid media Difco agar was added at a final concentration of 1% (w/v). When required, antibiotics were used at the following concentrations: neomycin, 10 µg ml−1for liquid media and 50 µg ml−1for solid media; streptomycin 2 µg ml−1for liquid media and 5 µg ml−1for solid media; spectinomycin 2 µg ml−1for liquid media and 5 µg ml−1 for solid media. In experiments where heterocysts were induced to differentiate, cells from bubbled cultures in BG11 medium were filtered, washed twice with BG110medium (similar to BG11 but lacking NaNO3), inoculated in BG110medium supplemented with 10 mM NaHCO3, and cultured for further 24 h at 30◦C Frontiers in Microbiology | www.frontiersin.org 2June 2016 | Volume 7 | Article 857 Santamaría-Gómez et al. Valyl-tRNA Synthetase-ATP Synthase Interaction under continuous illumination. When indicated, cultures were supplemented with L-methionine sulfoximine (MSX) at a final concentration 2.5 µM. Escherichia coli was cultured in LB medium supplemented with antibiotics at standard concentrations when necessary (Ausubel et al., 2010). Strain DH5αwas used for regular cloning and strain C41(DE3) was used for the expression of Anabaena ValRS::His and ValRS1C::His proteins under the control of the T7 promoter. Expression of the T7 RNA polymerase in C41(DE3) cells was induced by addition of IPTG (isopropyl β-D-1-thiogalactopyranoside) at a final concentration of 0.4 mM. Cell Fractionation Cyanobacterial cell fractionation and membrane preparation was performed following a procedure based on that described by Sobotka et al. (2008), with some modifications (OlmedoVerd et al., 2011). Cells from 600 to 700 ml cultures of cyanobacteria were harvested by centrifugation, washed with 50 mM Tris-HCl buffer [pH 7.5] and resuspended in buffer T (20 mM HEPES-NaOH [pH 7.5], 10 mM MgCl2, 5 mM CaCl2and 20% glycerol) at a ratio of 5 ml of buffer T per gram of cell pellet (wet weight). Cells were disrupted in a French press at 9000 psi in the presence of 1 mM PMSF (phenylmethylsulfonyl fluoride) and CompleteTM EDTA free, (Sigma) protease inhibitor cocktail. The cell extracts were centrifuged at 32,000 x g for 10 min at 4◦C to eliminate unbroken cells and cell debris. The supernatant (referred to as the “cell extract”) was ultracentrifuged at 100,000 x g for 1 h at 4◦C. To avoid contamination with the membrane fraction only the top portion (about 75% of the total volume) of the supernatant containing the soluble fraction was saved, while the bottom portion in direct contact with the pellet containing the membranes was discarded. The pellet containing the membranes was washed with buffer T and resuspended in the same buffer supplemented with 1% n-dodecyl-β-D-maltoside. In experiments where membrane purity was paramount, pellets were resuspended in buffer T and ultracentrifuged a second time at 100,000 x g for 1 h at 4◦C. The supernatant was discarded and the pellet was resuspended in the same buffer supplemented with 1% n-dodecyl-β-D-maltoside. The chlorophyll concentration in cyanobacterial cultures or cell fractions were determined as previously described (Mackinney, 1941), while the protein content was determined by the modified Lowry procedure (Markwell et al., 1978). Fractionation of IPTG-induced E. coli C41(DE3) cells was carried out by resuspension of cells in buffer A (0.05 M phosphate buffer [pH 8], 0.15 M NaCl, and 10% glycerol) and disruption in the French press at 12,000 psi in the presence of 1 mM PMSF. Extracts were subjected to a first centrifugation at 21,000 x g and the supernatant to a second centrifugation at 100,000 x g. The supernatant from the second centrifugation contained the soluble fraction. The pellet corresponding to the membrane fraction was washed extensively with buffer A and resuspended in buffer A supplemented with 1% n-dodecyl-β-Dmaltoside. The plasmid and strain construction is described in Table S5. The oligonucleotides used are listed in Table S6. Colorless-Native Page (CN-PAGE) Anabaena cells were cultured in BG11-Cu2+(BG11 medium omitting CuCl2) under standard laboratory conditions till they reached a chlorophyll concentration of 4–5 µg/ml. At this point, CuSO4was added at a final concentration of 1.5 µM to induce the expression of the petE promoter and cells were further cultured under the same conditions for 24 h. Cells from 700 ml cultures were harvested by filtration washed with thylakoid buffer (25 mM MES-NaOH pH 6.5, 5 mM CaCl2, 10 mM MgCl2, 20% glycerol) and resuspended in the same buffer at a ratio of 5 ml of buffer per gram of cells. Cells were disrupted after the addition of protease inhibitors (CompleteTM EDTA free, Sigma) by three passages through a French press at 9000 p.s.i. and extracts were fractionated according to Kopecna et al. (2012). The chlorophyll content of the membrane suspension was measured, digitonin was added at a ratio of 10 grams per gram of chlorophyll, and incubated at 4◦C for 30 min. Samples containing 5 µg of chlorophyll were first resolved in a 3–13% polyacrylamide gel containing 0.005% digitonin (Krause and Seelert, 2008). The gel was cut into portions corresponding to each lane incubated for 1 h in denaturing solution containing 2% SDS, 66 mM Na2CO3, and 0.67% 2-mercaptoethanol. Each gel piece was placed on a denaturing (8–20%) polyacrylamide gel and subjected to SDS-PAGE. Proteins were visualized by gel staining with Coomassie blue. In vivo Cross-Linking and Protein Purification Anabaena was cultured in BG11 or BG11-Cu2+medium under standard conditions to a chlorophyll concentration of 4–5 µg/ml. In those cases were induction of the petE promoter was needed CuSO4was added to a final concentration of 1.5 µM and cells were further cultured for 24 h. Cross-linking was performed as described (Staron et al., 2011). Briefly, a solution containing 4% formaldehyde in PBS buffer (10 mM Na2HPO4, 1.8 mM KH2PO4, 137 mM NaCl, 2.7 mM KCl) was added to cultures so that the final concentration of formaldehyde was 0.5% and were incubated under standard culture conditions for 20 min at room temperature. Cells from 80 ml cultures were harvested by centrifugation at 4500 x g for 5 min at 4◦C, washed with 100 mM Tris-HCl buffer pH 8 and resuspended in the same buffer at a ratio of 2 ml of buffer per gram of the cell pellet. Protease inhibitor (CompleteTM EDTA free, Sigma) was added and cells were disrupted by 10 passages through a French press at 18000 psi. Extracts were fractionated by centrifugation at 32000 x g for 15 min at 4◦C. The pellet containing unbroken cells and cell debris was discarded and the supernatant was ultracentrifugated at 100,000 x g for 1 h at 4◦C. The supernatant corresponding to the soluble fraction was carefully transferred to new tubes and the pellets corresponding to the membrane fraction was washed three times with 100 mM Tris-HCl buffer pH 8. Pellets were resuspended by addition of 500 µl of lysis buffer (150 mM NaCl, 1% Triton X-100, 50 mM, Tris-HCl pH 8, 10% glycerol) and incubation at 4◦C for 30 min. Purifications were performed using the µMACSTM Epitope Tag Protein Isolation kit (Miltenyi Biotec). Samples containing 7 mg of total protein Frontiers in Microbiology | www.frontiersin.org 3June 2016 | Volume 7 | Article 857 Santamaría-Gómez et al. Valyl-tRNA Synthetase-ATP Synthase Interaction were incubated with the corresponding magnetic beads-coupled antibody (anti-tag MicroBeads, Miltenyi) at a ratio of 50 µl of antibody per ml and mixtures were incubated for 1 h at 4◦C with gentle shaking. Inmunoprecipitations were performed following the instructions provided by the antibody provider. Crosslinking methylene bridges were removed at elution, by incubation for 5 min in a modified SDS-sample buffer (50 mM Tris HCl pH.6.8, 50 mM DTT, 1% SDS, 1 mM EDTA, 0.005% bromphenol blue, 10% glycerol) pre-heated at 95◦C. Eluted fractions were resolved by SDS-PAGE in 8–15% polyacrylamide gels. Enzymatic Assays Aminoacylation assays were performed as described (Francklyn et al., 2008). Reactions contained 100 mM HEPES [pH 7.2], 20 mM KCl, 30 mM MgCl2, 5 mM ATP, 0.1 mg ml−1BSA, 0.5 mM DTT, 20 µM [14C]-L-Val (250 Ci mol−1), 10 µM tRNA and 5– 15 nM enzyme. Reactions were incubated at 30◦C and stopped on filters soaked with 5% trichloroacetic acid, washed with 5% trichloroacetic acid and the radioactivity retained was counted by scintillation. Cellular ATP Content Determination ATP measurements were carried out using a commercial kit (ATP Biomass Kit HS, BioThema) based on luciferase activity following the instructions of the manufacturer. Briefly, 0.1 ng of chlorophyll of a cyanobacterial culture growing in standard conditions (around 25 µl) or incubated for 4 h in the presence of 2.5 µM MSX (L-methionine sulfoximine) was mixed with 50 µl of Extractant B/S and 400 µl of ATP Reagent HS. Light emission was measured in a scintillation counter (LS 6000, Beckman). Microscopy Anabaena filaments on solid medium were analyzed by confocal microscopy using a Leica HCX PLAN-APO 63X 1.4 NA oil immersion objective attached to a Leica TCS SP2 confocal laserscanning microscope. GFP was excited at 488 nm using an argon ion laser. Fluorescent emission was monitored by collection across windows of 500–540 nm (GFP imaging) and 630–700 nm (cyanobacterial autofluorescence). Bioinformatics Cyanobacterial species tree was inferred from concatenated small and large rRNA sequences (Ochoa de Alda et al., 2014). Data set selection, retrieval, concatenation, BMGE trimming (default, PAM100 matrix) and removal of constant sites were performed before bayesian phylogenetic reconstruction. GTR+4Ŵ+CAT (the evolutionary model that best fits dataset) was selected after carrying out a posterior predictive analysis of different alternatives using the ppred programme implemented in Phylobayes (Lartillot et al., 2009). Phylogenetic reconstruction was achieved using a parallelized version of phylobayes, MPI phylobayes (Lartillot et al., 2013), run at Cipres Gateway High Performance Computing Clusters (Miller et al., 2010). Convergence of two chains was checked with the bpcomp programme, whereby convergence was reached if the maxdiff value of the two chains was <0.1. Protein Identification by MS Protein bands were manually excised from Coomassie-stained gels and subjected to in-gel trypsin digestion (Shevchenko et al., 1996) in a Progest automatic in-gel protein digestor (Genomic Solutions, Cambridgeshire, UK) according to the manufacturer’s recommendations for Coomassie-stained samples. To remove dye and SDS impurities, gel plugs were extensively washed with 25 mM ammonium bicarbonate, in-gel reduced with 60 mM dithiothreitol, and S-alkylated with excess iodoacetamide followed by digestion with modified porcine trypsin (Promega, Madison WI) (1:100 wt/wt) at 37◦C for 6 h. Peptides were subjected to subsequent extractions in ammonium bicarbonate, 70% acetonitrile, and 1% formic acid, dried down by a Speed-Vac centrifugation and resuspended in 10 µL of 0.1% formic acid. An Agilent XCT plus ion trap mass spectrometer with a ChipCube interface fed by an Agilent 1100 series nanopump HPLC system was used for MS and MS/MS data acquisition. Samples were concentrated, desalted, and resolved using the ProtID-Chip-150 (II) (Agilent) as described (Martinez-Esteso et al., 2011). MS spectra were scanned at 26,000 m/z per second and MS/MS spectra at 8100 m/z per second in the range 300– 2200 m/z. The four most intense precursor ions in MS scans were selected for MS/MS and then passed to an active exclusion list released after 1 min. Each MS/MS spectra data set (ca. 1200 spectra/run) was processed to determine monoisotopic masses and charge states, to merge MS/MS spectra with the same precursor (1m/z<1.4 Da and chromatographic 1t<15 s) and to select high quality spectra with the Extraction tool of SpectrumMill Proteomics Workbench (Agilent). The reduced data set was searched against the Swissprot forward and reversed protein database without taxonomical restrictions in the identity mode with the MS/MS Search tool of SpectrumMill Proteomics Workbench using the following parameters: trypsin, up to 2 missed cleavages, fixed modification carbamidomethylation of cystein, variable modificacion oxidation of methionine, and a mass tolerance of 2.5 Da for the precursor and 0.7 Da for product ions. Peptide hits were validated first in the peptide mode and then in the protein mode according to the score settings recommended by the manufacturer. RESULTS Identification of Novel aaRSs Containing CAAD Previous work revealed 14 cyanobacterial species with CAAD inserted in either GluRS, ValRS, IleRS, or LeuRS (Olmedo-Verd et al., 2011). Analysis of newly sequenced genomes evidenced that these CAAD-containing aaRSs (hereafter aaRSsC) are widespread in the cyanobacterial radiation. Very interestingly, analysis of new genome sequences also revealed the presence of CAAD in other class I aaRSs, including ArgRS, MetRS and CysRS. Among the 279 sequenced cyanobacteria, 102 contained at least one aaRSC. ValRSC, the most widespread, was found in 79 species whereas IleRSCwas present in 11 species, GluRSC in 5, LeuRSCin 2, ArgRSCin 2, CysRSCin 1 and MetRSC in 1 (Figure 1A, Tables S1–S3). The absence of paralogs Frontiers in Microbiology | www.frontiersin.org 4June 2016 | Volume 7 | Article 857 Santamaría-Gómez et al. Valyl-tRNA Synthetase-ATP Synthase Interaction FIGURE 1 | CAAD in cyanobacterial aaRSs. (A) Phylogenetic tree of cyanobacterial species based on concatenated small and large rRNA (Continued) FIGURE 1 | Continued sequences. Representative species of distinct phylogenetic groups were chosen for the construction of the tree. The existence of aaRSs containing CAAD in particular species is indicated with colored boxes following the same color code as in (B). One or two asterisks indicate the insertion of CAAD at positions 750 or 820 in ValRS, respectively. Clades containing species with aaRSsCare indicated by brackets and roman numbers. (B) Diagram showing the position of the insertion of CAAD in distinct aaRSs. (C) Representation of bacterial aaRSs based on structures in the Protein Data Bank (PDB codes 1FFY, HVS, 2BYT, 1LI5, 2CT8, and 1F7U. The amino acids corresponding to the insertion point of CAAD in cyanobacterial aaRSs are highlighted in cyan. For the ValRS, the insertion points corresponding to positions 750 and 820 are indicated in cyan and green respectively. The diagrams for ValRS, IleRS, and LeuRS have been published elsewhere (Olmedo-Verd et al., 2011) and are shown here for the sake of comparison. in these species indicated that such enzymes are active for tRNA aminoacylation. Eight species contained two aaRSCin four different combinations (LeuRSC+CysRSC, ValRSC+ArgRSC GluRSC+MetRSC, GluRSC+IleRSC). The emerging picture was quite intriguing as two thirds of species contained no aaRSsC and one third possessed one or two aaRSsC(Figure 1A). Besides, no organism outside the cyanobacterial phylum was found to contain aaRSsC. In the new aaRSsCdetected in this study (ArgRSC, MetRSC, and CysRSC), CAAD is inserted at internal positions C-terminal to the catalytic domain (Figure 1B), which in the 3-D structure correspond to residues of the opposite surface to that interacting with tRNA (Figure 1C), similar to what is observed for ValRSC, LeuRSC, and IleRSC. In the particular case of ValRS, CAAD was previously found at two alternative insertion points (positions 750 and 820, approximately) in different species, suggesting that each insertion was the result of independent evolutionary events (Olmedo-Verd et al., 2011). In the new species here detected to contain ValRSC, CAAD is inserted at either of these positions and species with CAAD at the same position tend to cluster in the phylogenetic species tree (Figure 1A, clusters I to VII), supporting that in clustered species ValRSCis the result of a common insertion event. Phylogenetic analyses of CAAD lend support to this view (Figure S1). CAAD Functions as a Membrane-Targeting Domain in Cyanobacteria and in Heterologous Systems In order to check whether similar to ValRSCin Anabaena, other aaRSsCfrom different species also localize at the thylakoid membrane, genes encoding representatives of each of them were PCR amplified using genomic DNA from the corresponding species, fused to the ORF of the green fluorescent protein (GFP), cloned in a shuttle vector downstream of the petE inducible promoter and introduced in Anabaena by conjugation. Except for the construction expressing the GFP fusion of LeuRSCfrom Lyngbya sp. PCC 8106, all other constructions showed a strong toxicity in Anabaena and ex-conjugants could not be obtained despite multiple attempts, which is in line with previous findings about the barriers to horizontal transfer of genes involved in Frontiers in Microbiology | www.frontiersin.org 5June 2016 | Volume 7 | Article 857 Santamaría-Gómez et al. Valyl-tRNA Synthetase-ATP Synthase Interaction translation due to toxicity of their gene products (Sorek et al., 2007). For the GFP-LeuRSCfusion, fluorescence of GFP was monitored by confocal microscopy and it was observed to colocalize with the red fluorescence of the photosynthetic pigments, indicating a specific localization in the thylakoid membranes (Figure 2A, left; compare with controls of soluble and plasma membrane proteins in center and right panels). This reproduces the reported localization of the Anabaena ValRSC-GFP fusion and it is consistent with the observed confinement of tRNALeu charging activity in membrane fractions of Lyngbya cell extracts (Olmedo-Verd et al., 2011). These observations indicated that similar to the wellcharacterized ValRSCof Anabaena, other CAAD-containing aaRSs from distinct species would also localize in thylakoid membranes and in turn, that CAAD would function as a membrane-targeting domain in different cyanobacterial species. This raised the question of whether CAAD could mediate protein targeting to membranes independently of the host. To test this, the Anabaena ValRSCORF was fused to a His-tag (HisValRSC), expressed in E. coli C41(DE3) and its presence was analyzed in sub-cellular fractions. As observed in Figure 2B, His-ValRSCmostly localized in the membrane fraction of cells induced with IPTG. Consistent with this, membrane fractions of these cells exhibited tRNAVal aminoacylation activity, whereas membrane fractions of the recipient E. coli strain carrying the empty vector or a plasmid expressing His-ValRS1C(where CAAD was deleted) were virtually devoid of activity (Figure 2C). ValRS1Cwas actually found in the soluble fraction of E. coli cells Figure 2B. These results evidenced that CAAD can function as a membrane-targeting domain also in organisms where it does not exist, suggesting that the capacity for membrane targeting is inherent to this domain. It is worth mentioning that E. coli does not posses thylakoid membranes, so in this organism CAAD mediates protein targeting to the plasma membrane. Subcellular Localization of the 20 aaRSs in Anabaena The distribution of aaRSs in cyanobacterial cells was addressed. A possibility to be tested was whether aaRSs not containing CAAD could bind to membranes despite the absence of an apparent membrane-anchoring domain. We hypothesized that, if translation complexes existed, they were likely to accrue around membrane ValRSC, as a result of its restricted mobility. To investigate this, Anabaena strains were engineered to express each aaRSs fused to GFP from the petE inducible promoter and the subcellular localization of the fusion proteins was monitored by confocal fluorescence microscopy. Anabaena contains duplicated ThrRSs, named T1 and T2, (Napolitano et al., 2012; Rubio et al., 2015) and similar to many other bacteria, it lacks a GlnRS, Gln-tRNAGln being synthesized through the indirect pathway (Schon et al., 1986; Luque et al., 2008). For hetero-oligomeric aaRSs, namely GlyRS and PheRS, only one construct fusing the αsubunit to GFP was made. Thus, 20 different constructs were introduced in Anabaena. Consistently with published data, the GFPValRSCfusion showed a sub-cellular distribution that matched FIGURE 2 | CAAD possess an inherent membrane targeting ability. (A) Confocal fluorescent microscopy of Anabaena filaments expressing the LeuRS from Lyngbya sp. PCC 8106 fused to GFP (left), GFP (center), or the Anabaena ammonium transporter Amt1 fused to GFP (right). For each group of pictures, the left panel shows the fluorescence of GFP, the middle panel the red fluorescence of thylakoid membrane pigments and the right panel the merged picture of the former ones. (B) SDS-PAGE of fractionated extracts from E. coli C41(DE3) expressing Anabaena His-ValRSC(left) or His-ValRS1C (right). CE, whole cell extracts; SF, soluble fraction; MF, membrane fraction. White arrowheads point to the position of the His-ValRSCor His-ValRS1C.(C) Aminoacylation assay with membrane fractions from E. coli C41(DE3) cells expressing Anabaena His-ValRSCor His-ValRS1C. A control of the activity of membranes from E. coli C41(DE3) carrying the pET28b vector is shown. the autofluorescence of photosynthetic pigments, used here as an indicator of the position of thylakoid membranes (Figure 3, top panels). By contrast, the fluorescence of all other GFP fusions was detected in the central part of the cytoplasm, with a pattern complementary to that of thylakoid Frontiers in Microbiology | www.frontiersin.org 6June 2016 | Volume 7 | Article 857 Santamaría-Gómez et al. Valyl-tRNA Synthetase-ATP Synthase Interaction membranes (Figure 3). This signal was similar to that of the soluble GFP protein (Figure 2A) and indicates a sub-cellular distribution of these aaRSs in the soluble portion of the cytoplasm. In the absence of combined nitrogen, Anabaena filaments differentiate specialized cells named heterocysts, which are mostly devoted to nitrogen fixation and show a distinct ultrastructure and morphology. ValRSCwas shown to undergo a re-localization along the differentiation process to concentrate at the poles of mature heterocysts (Olmedo-Verd et al., 2011) (see also Figure S2, top panels). The behavior of the 20 GFP-aaRS fusions during differentiation was analyzed here. GFP-aaRSs not containing CAAD did not show any sign of re-localization, so that in mature heterocysts they showed a subcellular distribution similar to that in vegetative cells (Figure S2). Complex Formation by Membrane-Bound and Soluble aaRSs in Anabaena The asymmetric distribution of aaRSs in Anabaena, with 19 soluble and one membrane-bound aaRS was intriguing. We hypothesized that the membrane-bound aaRS could have acquired novel functionality, perhaps not related to its canonical aminoacylation function, that emanates from this new localization. This functionality could be exerted through interaction with other proteins, whose identity could shed light on this issue. On the other hand, soluble aaRSs could also interact with other proteins. To get an insight on this, a global analysis was conducted in search for proteins that interacted with any of the 20 aaRSs in Anabaena. The experimental approach consisted in the in vivo crosslinking of Anabaena cells expressing each GFP-aaRS fusion, followed by cell disruption and purification with anti-GFP antibodies coupled to magnetic beads. For the strains expressing the soluble GFP-aaRSs fusions, proteins were purified from the soluble fraction of cell extracts whereas for the strain expressing ValRSC-GFP, purifications were performed from whole cell extracts and membrane fractions (see below). Proteins from each purification were resolved by SDS-PAGE after crosslinking reversal (see Materials and Methods). For each of the soluble aaRSs, a major band coincident with the expected MW of the respective GFP fusion was observed in SDS-PAGE gels (Figure 4, white arrowheads) and both GFP and the corresponding aaRS were identified in such bands by mass spectrometry (MS). Differential bands (i.e., those that were absent from control lanes containing proteins purified from Anabaena wild-type cells or a strain expressing GFP) that putatively corresponded to proteins interacting with aaRSs, were selected for MS identification. Quite surprisingly, most of the 141 bands subjected to MS identification corresponded to putative proteolytic products of the corresponding aaRS (despite the rapid manipulation of samples at low temperature and the regular addition of a cocktail of protease inhibitors to all solutions). However, some proteins that corresponded to putative aaRS interactants were identified and are listed in Table S4. It is crucial to mention that GFP-α-GlyRS and GFP-α-PheRS copurified with their respective βsubunits (Figure 4, lanes 5 and FIGURE 3 | Confocal fluorescent microscopy of Anabaena filaments expressing the indicated fusion protein. Left panels show the green fluorescence of GFP; central panels, the red fluorescence of photosynthetic pigments and right panels the merged picture of the former ones. 22, black arrowheads), which validated this approach at least for the co-purification of proteins forming stable complexes. In the particular case of ValRSC-GFP, numerous bands corresponding to putative proteolytic products (labeled with Frontiers in Microbiology | www.frontiersin.org 7June 2016 | Volume 7 | Article 857 Santamaría-Gómez et al. Valyl-tRNA Synthetase-ATP Synthase Interaction FIGURE 4 | Purification of GFP-aaRS fusion proteins from Anabaena cells. Panels correspond to SDS-PAGE gels. Each lane contains fractions from the purification of a GFP-aaRS fusion protein (indicated at the top by the cognate amino acid of the aaRS) from Anabaena cells expressing it. Control lanes labeled as C or C’ carry purification fractions of Anabaena cells or Anabaena cells expressing GFP, respectively. White arrowheads point to the corresponding fusion protein. Black arrowheads point at the beta subunit of GlyRS (lane 5) and PheRS (lane 22). Marks at the left side of each panel indicate the position of the bands of the PageRulerTM Prestained Protein Ladder (Fermentas). white dots in Figure 5) were also observed in the electrophoreses, despite measures mentioned above were taken to avoid proteolysis. ValRSC-GFP was readily detected in purifications from whole cell extracts (Figure 5A, white arrowheads). Interestingly, specific bands (Figure 5A, black arrowheads), observed in the lane containing ValRSC-GFP that were absent in lanes containing fractions from control strains expressing ValRS1C-GFP or no fusion protein (Figure 5A, lanes 1 and 3, respectively) were identified as subunits α(AtpA) and β (AtpB) of the FoF1-ATP synthase. Further purifications were conducted using membrane fractions of the above-mentioned strains and again, subunits α(AtpA), β(AtpB) and, in some experiments, subunit γ(AtpC) of the ATP synthase complex were detected to co-purify specifically with ValRSC-GFP (Figure 5B, black arrowheads). Importantly, subunits of the ATP synthase did not co-purify with GFP-ValRS1Cfrom either whole-cell extracts (Figure 5A, lane 1) or membrane fractions (Figure 5B, lane 1), suggesting that CAAD would be directly involved in the interaction. To test this, membrane fractions from Anabaena strains engineered to express GFP-CAAD or GFP were used for purification experiments similar to those described above. Two bands identified as AtpA (α) and AtpB (β) were detected in the lane corresponding to the strain expressing GFP-CAAD (Figure 5C, lane 2) but not in the control lane corresponding to that expressing GFP (Figure 5C, lane 1), indicating that CAAD is directly involved in the interaction with these proteins. To rule out artifacts derived from the use of the GFP tag in the experiments above, membrane fractions of a strain engineered to express ValRSCfused to a C-terminal His tag were used for purification with anti-His tag antibodies coupled to magnetic beads after in vivo cross-linking. Proteins that copurified specifically with ValRSC-His were identified as AtpA (α), AtpB (β), and AtpC (γ) (Figure 5D, lane 2). Further evidence for this interaction was sought by reciprocal co-purification experiments. The atpA ORF, encoding the α subunit of the ATP synthase, was fused to the ORF encoding GFP and the resulting construct was introduced in wild-type Anabaena and in the strain expressing ValRSC-His. Cultures of these strains and control strains were subjected to in vivo crosslinking and purification with anti-GFP antibodies coupled to magnetic beads. Purified fractions were resolved by SDSPAGE after crosslinking reversal (Figure 6A). A number of bands were observed in lanes corresponding to the strains expressing AtpA-GFP in a wild-type background (lane 2) or a background expressing ValRSC-His (lane 3), but not in lanes 1 and 4 corresponding to control strains not expressing AtpA-GFP (lane 1, wild-type background; lane 4, ValRSC-His background). AtpAGFP was identified as the most intense band in lanes 2 and 3 by size and MS. Other prominent bands contained subunits β,γ,δ, b, b′, and εof the ATP synthase. Importantly, the intensity of the bands was consistent with the subunit stoichiometry of ATP synthase, indicating that AtpA-GFP had successfully integrated in the FoF1 complex. The presence of a minor band matching the size and identified by MS as the endogenous αsubunit (AtpA), indicated that some of the complexes contained both AtpA-GFP and AtpA subunits. Crucially, a specific band with a molecular weight of ca. 110 kDa only present in lanes 2 and 3 was identified by MS as ValRSC. A band of similar size was revealed by western blot with anti-His antibodies (Figure 6B, lane 3), corroborating the presence of ValRSC-His at this position. Frontiers in Microbiology | www.frontiersin.org 8June 2016 | Volume 7 | Article 857 Santamaría-Gómez et al. Valyl-tRNA Synthetase-ATP Synthase Interaction FIGURE 5 | Co-purification of Anabaena ValRSCwith subunits of the ATP synthase. (A) Whole cell extracts from Anabaena strains expressing the fusion protein indicated at the top of the panel were purified using anti-GFP antibodies and resolved by SDS-PAGE gel. Control lane labeled as “C” contains purification fractions of Anabaena cells not expressing any fusion protein. White arrowheads point to the corresponding fusion protein. Black arrowhead point to subunits of ATP synthase. White dots indicate proteolytic fragments of the GFP fusion proteins. Asterisks indicate the position of RubisCO, double asterisks indicate the position of the light chain of inmunoglobulins (B) and (C) Membrane fractions of Anabaena strains expressing the fusion protein indicated at the top of the panel were purified using anti-GFP antibodies and resolved by SDS-PAGE gel. Other details are like in (A).(D) Membrane fractions of Anabaena strains expressing the fusion protein indicated at the top of the panel were purified using anti-hexahistidine antibodies and resolved by SDS-PAGE gel. FIGURE 6 | Co-purification of Anabaena AtpA-GFP with other subunits of the ATP synthase and ValRSCor ValRSC-His. (A: Left), SDS-PAGE gel containing fractions purified with anti GFP-antibodies coupled to magnetic beads. Lanes contain purification fractions from Anabaena cells expressing the proteins indicated at the top. Control lane labeled as C carry purification fractions from Anabaena cells expressing no fusion protein. Proteins identified by mass spectrometry are indicated. An asterisk indicates the position of RuBisCO. (B: Right), western blot of a gel identical to that of the left panel using an antibody against hexahistidine tag. Results above strongly indicate that ValRSCinteracts with the ATP synthase complex in the thylakoid membranes of Anabaena. A technique extensively used for the identification and analysis of membrane protein complexes, including photosynthetic and respiratory complexes, is 2D-native electrophoresis that includes blue-native and colorless-native electrophoresis. In these gels, protein complexes are resolved in the first dimension in native conditions (in the presence of a mild detergent) and after denaturation, the components of each complex are separated in the second dimension by SDS-PAGE. Thus, in 2D-native gels the components of a particular protein complex distribute within a vertical line. To analyze complex formation by ValRSC in Anabaena, colorless-native (CN-native) electrophoresis was chosen for its better preservation of labile complexes (Wittig et al., 2007). To facilitate detection, membrane preparations of the Anabaena strain expressing ValRSC-GFP were utilized, so that its position in the gels could be ascertained by western Frontiers in Microbiology | www.frontiersin.org 9June 2016 | Volume 7 | Article 857