Uncovering the protein lysine and arginine methylation network in Arabidopsis chloroplasts

Abstract

Post-translational modification of proteins by the addition of methyl groups to the side chains of Lys and Arg residues is proposed to play important roles in many cellular processes. In plants, identification of non-histone methylproteins at a cellular or subcellular scale is still missing. To gain insights into the extent of this modification in chloroplasts we used a bioinformatics approach to identify protein methyltransferases targeted to plastids and set up a workflow to specifically identify Lys and Arg methylated proteins from proteomic data used to produce the Arabidopsis chloroplast proteome. With this approach we could identify 31 high-confidence Lys and Arg methylation sites from 23 chloroplastic proteins, of which only two were previously known to be methylated. These methylproteins are split between the stroma, thylakoids and envelope sub-compartments. They belong to essential metabolic processes, including photosynthesis, and to the chloroplast biogenesis and maintenance machinery (translation, protein import, division). Also, the in silico identification of nine protein methyltransferases that are known or predicted to be targeted to plastids provided a foundation to build the enzymes/substrates relationships that govern methylation in chloroplasts. Thereby, using in vitro methylation assays with chloroplast stroma as a source of methyltransferases we confirmed the methylation sites of two targets, plastid ribosomal protein L11 and the β-subunit of ATP synthase. Furthermore, a biochemical screening of recombinant chloroplastic protein Lys methyltransferases allowed us to identify the enzymes involved in the modification of these substrates. The present study provides a useful resource to build the methyltransferases/methylproteins network and to elucidate the role of protein methylation in chloroplast biology.

Figure 1. Filtering pipeline for the identification of methylpeptides in the AT_CHLORO datasets.

(a) Following the initial database search and parsing of results, the workflow is composed of six steps combining automatic filtering and expert examination of spectra. 1/selection of methyl peptide-spectrum matches (PSMs) with score ≥ 50; 2/manual inspection to assess spectral quality; 3/selection of methyl-PSMs (score <50) with identical methylation sites; 4/removal of sites with only one PSM; 5/removal of ambiguities due to amino acid substitutions; 6/removal of ambiguities due to trimethylation vs. acetylation. The final AT_CHLORO_Me list consisted in methylpeptides validated from the mass spectrometry aspect. Peptide counts represent the number of distinct peptide sequences compiled from a larger number of PSMs. (b) Score distribution of methyl-PSMs matching a reverse or true (forward) Arabidopsis protein library. The threshold score of 50 in step 1 was selected from these distributions. (c) Illustration of the diversity among PSMs that identify a methylation site (step 3). In this example (FBA1 protein, At2g21330), three PSMs point out a trimethyl-Lys with two distinct overlapping peptide sequences and various modification patterns of a Met residue (mono or dioxidized).

Figure 2. Immunodetection of Lys- and Arg-methylated proteins in chloroplast stroma and membranes subfractions.

Chloroplasts from Arabidopsis, spinach and pea leaves were purified using Percoll gradients and fractionated into soluble (stroma) and membrane (thylakoids and envelope) fractions. Fifty µg of proteins were analyzed by SDS-PAGE (Coomassie blue staining) and immunoblotting with antibodies against trimethyl-Lys (anti-K_me3), mono- and dimethyl-Lys (anti-K_me1/2), or mono- and dimethyl-Arg (anti-R_me1/2). The major polypeptides detected by the anti-K_me3 antibodies are RbcL (*) and fructose bisphosphate aldolases (**) .

Figure 3. Main features of the identified chloroplastic methylproteins and methylation sites.

(a) Curated subcellular/subplastidial location of methylproteins (as in Table 2). (b) Functional categories of chloroplastic methylproteins. Annotated proteins from Table 2 were grouped to create categories ‘metabolism (other than photosynthesis)’ and ‘protein synthesis and targeting’. (c) Amino acid motif surrounding Lys methylation sites was created using WebLogo (http://weblogo.berkeley.edu/). (d) Positioning of the Lys395 methylation site on the 3D-structure model of fructose bisphosphate aldolase (FBA1, At2g21330). The model was generated with the Phyre² server using the 3D structure of aldolase from rabbit muscle (PDB entry 1ZAI) and imaged with the PyMOL software.

Figure 4. Biochemical validation of methylproteins through in vitro methylation assays using stroma from Arabidopsis chloroplasts.

Methylation assays were done in the presence of 20 µg purified recombinant targets (FBA2, GAPA1, PRPL11, or ATP-B), 80 µg stroma from Arabidopsis Col-0 chloroplasts, 20 µM [methyl-³H]-AdoMet and 100 nM S-adenosylhomocysteine hydrolase. After incubation at 30°C for 1 to 2 hours, assays were split into two equals parts and radioactivity incorporated into proteins was counted by liquid scintillation (panel a) and analyzed by phosphorimaging (panels b–c). In panels (a–c), the symbol Ø means that no recombinant protein was added to the stromal extract. Purified recombinant substrates are indicated by asterisks: ATP-B, 54 kDa; FBA2, 40 kDa; GAPA1, 38 kDa; PRPL11, 18 kDa (Fig. S3). Activities with the FBA2, PRPL11 and ATP-B substrates were strictly dependent on the addition of stroma. Values are mean ± SD of two to six independent determinations.

Figure 5. In vitro methylation of PRPL11 by PrmA-like.

(a,b) Kinetic analysis of PRPL11 methylation by PrmA-like. Purified recombinant PRPL11 (2.5 µg) and PrmA-like (0.25 µg) were incubated at 30°C with 20 µM [methyl-³H]-AdoMet. Radioactivity incorporated into proteins was counted by liquid scintillation (a) or analyzed by phosphorimaging (b). A time-course analysis representative of three repeats is shown. The activity (2.0±0.2 nmol methyl incorporated. min⁻¹.mg⁻¹ protein) was strictly dependent on the presence of PrmA-like. (c) Screening of five chloroplastic PKMTs for their ability to methylate PRPL11. Purified recombinant PRPL11 (10 µg) and PKMTs (1 µg) were incubated at 30°C for 15 min with 20 µM [methyl-³H]-AdoMet and radioactivity incorporated into proteins was counted by liquid scintillation. Enzyme nomenclature is as in Table 1. Values are mean ± SD of three determinations.

Figure 6. In vitro methylation of the ATP synthase β-subunit by PPKMT2.

(a) Screening of five chloroplastic PKMTs for their capacity to methylate ATP-B. Purified recombinant ATP-B (24 µg) and PKMTs (1 µg) were incubated at 30°C for 10 min with 20 µM [methyl-³H]-AdoMet and radioactivity incorporated into proteins was counted by liquid scintillation. Enzyme nomenclature is as in Table 1. Values are mean ± SD of two to four independent determinations. (b) Time-course analysis of ATP-B methylation by PPKMT2. Assays were as in (a) and included the wild-type version of ATP-B or the K₄₄₇A mutant. Kinetic analyses representative of three repeats are shown (11.1±1.7 nmol methyl incorporated.min⁻¹.mg⁻¹ protein with the ATP-B substrate). (c) Phosphorimage analysis of ATP-B methylation by PPKMT2. ATP-B (10 µg) and PPKMT2 (0.3 µg) were incubated at 30°C for 10 min with 20 µM [methyl-³H]-AdoMet. Symbol Ø means that no protein substrate was included in the assay.