Construction of split-ubiquitin plasmids and the split-ubiquitin yeast two-hybrid assays
For the split-ubiquitin yeast two-hybrid assay the coding sequence for mature full-length GET3b (aa 68-411) was cloned into pADSL-Nx using the BamHI/SalI restriction sites. All other constructs have been described previously (Bals et al., 2010; Pasch et al., 2005). The split-ubiquitin assay was done according to (Pasch et al , 2005).

MS/MS analysis

For in-solution mass spectrometry, proteins were digested (León et al., 2013; ISD:SDC; AP&PT) and peptides were purified by C18 STAGE-TIPS (Rappsilber et al., 2007). Peptides were analyzed using an ultra-HPLC Proxeon EASY-nLC 1000 system coupled online to Q Exactive Plus mass spectrometer (Thermo Fisher Scientific). Reversed-phase separation was performed using a 30 cm analytical column (100 μm diameter; DNU-MS (Novak) packed in-house with Reprosil-Pur 120 C18-AQ, 2.4 µm). Mobile-phase solvent A consisted of 0.1% formic acid and 4% acetonitrile in water, and mobile-phase solvent B consisted of 0.1% formic acid in 80 % acetonitrile. The flow rate of the gradient was set to 200 nl/min. A 70-min gradient was used (0–40% solvent B within 40 min, 40–100% solvent B within 10 min, 100% solvent B for 10 min, 100–0% solvent B within 5 min and 0% solvent B for 5 min).
Data acquisition was performed with the ddMS2 method with the following configuration: For the MS scans, the scan range was set to 250–2,000 m/z at a resolution of 70,000, and the automatic gain control (AGC) target was set to 1 × 106. For the MS/MS scans, Top 13 ions were chosen, the resolution was set to 35,000, the AGC target was set to 1 × 105, the precursor isolation width was 2 Da and the maximum injection time was set to 80 ms. The HCD normalized collision energy was 27%. MaxQuant was used to analyze the LC-MS/MS data (ver 1.5.5.; Cox and Mann, 2008) which allowed qualitative and quantitative analysis. The UniProt reference Arabidopsis database (UP000006548) was used for the identification of proteins. Default settings for fixed modifications were used, Dynamic modifications were set: Oxidation for M and Deamidation for NQ. Contaminants were included for peptide detection of a minimum length of 6 amino acids. FDR threshold was set to 0.01. Perseus was used to transform the quantitative data and perform a statistical analysis (Tyanova and Cox, 2018). The default settings for a Student’s t-test were utilized for the analysis. For this, independent triplicates corresponding to roughly 80 µg of plastid protein isolated from 8 pots of Arabidopsis thaliana lawn from wild-type and get3b plants were compared.