Structural and molecular comparison of bacterial and eukaryotic trigger factors

Abstract

A considerably small fraction of approximately 60-100 proteins of all chloroplast proteins are encoded by the plastid genome. Many of these proteins are major subunits of complexes with central functions within plastids. In comparison with other subcellular compartments and bacteria, many steps of chloroplast protein biogenesis are not well understood. We report here on the first study of chloroplast-localised trigger factor. In bacteria, this molecular chaperone is known to associate with translating ribosomes to facilitate the folding of newly synthesized proteins. Chloroplast trigger factors of the unicellular green algae Chlamydomonas reinhardtii and the vascular land plant Arabidopsis thaliana were characterized by biophysical and structural methods and compared to the Escherichia coli isoform. We show that chloroplast trigger factor is mainly monomeric and displays only moderate stability against thermal unfolding even under mild heat-stress conditions. The global shape and conformation of these proteins were determined in solution by small-angle X-ray scattering and subsequent ab initio modelling. As observed for bacteria, plastidic trigger factors have a dragon-like structure, albeit with slightly altered domain arrangement and flexibility. This structural conservation despite low amino acid sequence homology illustrates a remarkable evolutionary robustness of chaperone conformations across various kingdoms of life.

Figure 1

Evolutionary diversity of chloroplast trigger factor. (a) Alignment of trigger factor amino acid sequences from Escherichia coli (E.c.), Arabidopsis thaliana (A.t.), and Chlamydomonas reinhardtii (C.r.). Sequences were aligned by ClustalOmega and shaded using BoxShade (http://www.ch.embnet.org). Amino acids highlighted in black are perfectly conserved, similar residues are indicated by a grey background. Boxes indicate chloroplast transit-peptide cleavage sites as determined experimentally (green) and as predicted by TargetP and ChloroP (light red for AtTIG1)^, . Dark red box indicates putative ribosome-binding site and lines underneath the sequences indicate the three trigger factor domains according to the E. coli structure. (b) CrTIG1 and AtTIG1 are predicted to comprise the typical three-domain organization of the N-terminal ribosome-binding domain (red), the PPIase domain (turquois), and the C-terminal domain (blue). (c) Sequence homology comparison of EcTF, CrTIG1, and AtTIG1 according to http://imed.med.ucm.es/Tools/sias.html. (d) Phylogram based on amino acid sequence alignments of trigger factor from gram-negative bacteria (Escherichia coli, Vibrio cholerae, Thermotoga maritima), cyanobacteria (Nostoc sp. strain PCC 7120, Synechocystis sp. strain PCC 6803, Prochlorococcus marinus, Synechococcus sp. strain WH8102), the diatom (Phaeodactylum tricornutum), and the mature TIG1 sequences from red algae (Galdieria sulphuraria, Cyanidioschyzon merolae), green algae (Volvox carteri, Chlamydomonas reinhardtii, Coccomyxa subellipsoidea, Micromonas sp. strain RCC299, Ostreococcus lucimarinus), moss (Physcomitrella patens), and land plants (Zea mays, Setaria italica, Oryza sativa, Brachypodium distachyon, Populus trichocarpa, Brassica rapa, Arabidopsis thaliana, Capsella grandiflora). Bootstrap values are given next to the nodes. For sequence information see Supplementary Table S1; for phylogenetic comparison including truncated TIG2 variants see Supplementary Figure S1. The bar indicates branch length.

Figure 2

Intracellular localisation of TIG1. (a) EcTF, CrTIG1, and AtTIG1 were heterologously expressed in E. coli, purified via chitin affinity resins and size exclusion chromatography. 0.5 µg or 1 µg of each protein was separated by SDS-PAGE and stained with Coomassie. (b) 7.5 µg or 15 µg of soluble extracts from C. reinhardtii and A. thaliana was separated by SDS-PAGE next to 7.5 ng or 15 ng of purified TIG1 protein, transferred to nitrocellulose, and immunoblotted with antibodies against TIG1 (c). C. reinhardtii chloroplasts (CP) were isolated, lysed by hypo-osmotic shock, and separated into stroma (Str), low-density membranes (LM), and thylakoid membranes (Thyl). Mitochondria (Mito) were separated from the same strain. Whole cells (WC) and 7 µg of each fraction were separated by a 7.5–15% SDS-PAGE, transferred to nitrocellulose, and immunoblotted with antibodies against CrTIG1, chloroplast ribosomal protein PRPL1, HSP70B (stroma and membrane control), CF1β (thylakoid membrane associated), stromal CGE1, and mitochondrial carbonic anhydrase (CA). Note that LM might be over-represented compared to thylakoids. (d) Images of C. reinhardtii cells, stained with antibodies directed against CrTIG1, RbcL, and PRPL1 (FITC, green) and a FISH probe against 70 S ribosomes (red). The numbers of cells with the localisation patterns seen was 15 of 20 (75%) for CrTIG1, 16 of 20 (80%) for RbcL, 13 of 20 (65%) for PRPL1, and 20 of 20 (100%) for FISH. Antibody specificity is shown in Supplementary Figure S2. Scale bars represent 2 µm (e) 20 µg of soluble A. thaliana leaf extract or 20 µg of isolated chloroplasts was separated next to 7.5 ng purified AtTIG1 by SDS-PAGE, transferred to nitrocellulose, and immunoblotted with antibodies against AtTIG1, Actin, and RbcL. Immunoblots in (b, c) and (e) were cropped to the respective size of the displayed bands.

Figure 3

Secondary structure and thermal stability of TIG1s. (a) Left panels: CD spectra of 0.1 mg/mL of EcTF, CrTIG1, and AtTIG1 in 10 mM KCl, 20 mM Tris pH 7.5 buffer were recorded from 190 nm to 280 nm at 20 °C. Right panels: At the two minima at 208 nm and 222 nm, protein-unfolding curves were recorded from 20 °C to 94 °C at a heating rate of 1°/min. (b) Thermal stability of purified CrTIG1 was measured at 5 mg/mL with indicated buffers by stepwise heating of 1 °C/min over a temperature range of 20–95 °C. Protein-unfolding midpoints (T _m) were calculated from the first derivative of intrinsic protein fluorescence emission at 330 nm and 350 nm. (c) Accumulation of insoluble protein aggregates was observed upon cell exposure to various temperatures of heat stress. Protein aggregates and total lysates were separated by 12% SDS-PAGE and visualized by Coomassie staining. Gel slice of total lysates serves as input control. (d) Cropped images of immunoblots of samples shown in (c). show the accumulation of CrTIG1, Rubisco activase (RCA1), HSP90C, CPN20, and Cytochrome F in protein aggregates.

Figure 4

Purified chloroplast TIG1 is mainly monomeric. Determination of conformational states of EcTF, CrTIG1, and AtTIG1. 200 µg of protein was separated by SEC in a buffer containing 20 mM Tris pH 7.5 and 150 mM KCl and measured with a right-angle light scattering (RALS) detector combined with a refractive index (RI) detector. Here, the RI detector signal was used to follow the elution. Average molar masses are indicated in blue for each graph (right ordinate). For the full elution profile, see Supplementary Figure S3.

Figure 5

SAXS data from static measurements of trigger factors. Small-angle X-ray scattering data were collected for various concentrations of TIG1 proteins and merged. (a) and (d) Experimental SAXS profiles from merged scattering curves of CrTIG1 and AtTIG1 are indicated by small circles. Respectively, red curves represent the fit by GNOM, blue curves the best CRYSOL fit, grey curves the theoretical scattering of final DAMMIN models. Inset, Guinier plot of ln I(s) versus s ⁻² obtained from AUTORG. Data points used by AUTORG are labelled in blue. (b) and (e) Kratky plot s ² I(s) versus s ⁻¹ of CrTIG1 and AtTIG, respectively. (c) and (f) Corresponding p(r) function as calculated from the experimental scattering curves using GNOM. “au” is arbitrary units. For SAXS data on EcTIG and the FoXS fit, see Supplementary Figure S4.

Figure 6

SAXS data from HPLC-SAXS experiments of trigger factors. Small-angle X-ray scattering data were collected by size-exclusion high-performance liquid chromatography with a ENrich SEC 650 column online with small-angle X-ray scattering (HPLC-SAXS) with 400 µg of the respective protein. (a) and (d) Elution profiles of SEC-SAXS runs, represented by I(0) and R _g determined by AUTORG for each frame. Highlighted in grey are the frames used for averaging. (b) and (e) Experimental SAXS profiles from the averaged HPLC-SAXS frames of CrTIG1 and AtTIG1 are indicated by small circles. Respectively, red curves represent the fit by GNOM, blue curve the CRYSOL fit of the RaptorX models, and grey curves the theoretic scattering of final DAMMIN model. Inset: Guinier plot of ln I(s) versus s ⁻² obtained from AUTORG. Data points used by AUTORG are labelled in blue. (c) and (f) Corresponding p(r) function as calculated from the experimental scattering curves using GNOM. “au” is arbitrary units. For SAXS data on EcTF and the FoXS fit, see Supplementary Figure S5.

Figure 7

Global shape of the chloroplast trigger factor. Ab initio modelling from SAXS data. Shape topology and low-resolution bead models of CrTIG1 (a,b) and AtTIG1 (c,d) from static SAXS experiments (a,c) and HPLC-SAXS (b,d) were generated using AMBIMETER and DAMMIN based on 10 averaged GASBOR models for static or 20 DAMMIF models for HPLC-SAXS experiments, respectively. Bead models were superimposed by SUPALM on the RaptorX high-resolution model or the improved models by SREFLEX. Secondary structures were coloured in yellow (β-sheets), red (α-helix), and green (random coils). For different side views of the models, see Supplementary Figure S6.