Proteomics

Sample preparation. Proteins from centrosome preparations were separated on 10% SDS–PAGE gels (Invitrogen) and stained with colloidal blue staining (LabSafe GEL BlueTM GBiosciences). Gel slices were excised (20 fractions) and proteins were reduced with 10 mM dithiothreitol before alkylation with 55 mM iodoacetamide. After washing and shrinking of the gel fractions with 100% MeCN, in-gel digestion was performed using recombinant endoproteinase rLys-C (Promega) overnight in 25 mM NH₄HCO₃ at 30 °C.

MS analysis. Peptides were extracted and analysed by nano-LC–MS/MS using an Ultimate 3000 system (Dionex S.A.) coupled to a LTQ-Orbitrap XL mass spectrometer (Thermo Fisher Scientific), as described⁴⁶. Samples were loaded on a C18 pre-column (300 μm inner diameter × 5 mm; Dionex) at 20 μl min⁻¹ in 5% MeCN and 0.1% TFA. After 3 min of desalting, the pre-column was switched on the C18 column (75 μm inner diameter × 15 or 50 cm, packed with C18 PepMap, 3 μm, 100 Å; LC Packings) equilibrated in solvent A (5% CH₃CN and 0.1% HCOOH). Bound peptides were eluted using a 97-min linear gradient (from 5 to 30% (v/v)) of solvent B (80% CH₃CN and 0.085% HCOOH) at a 150-nl min⁻¹ flow rate and oven temperature of 40 °C. Data-dependent acquisition was performed on the LTQ-Orbitrap mass spectrometer in the positive-ion mode. Survey MS scans were acquired in the Orbitrap on the 480–1,200 m/z range with the resolution set to a value of 60,000. Each scan was recalibrated in real time by co-injecting an internal standard from ambient air into the C-trap (lock mass option). The five most intense ions per survey scan were selected for collision-induced dissociation (CID) fragmentation and the resulting fragments were analysed in the linear trap (LTQ). Target ions already selected for MS/MS were dynamically excluded for 180 s.

Data analysis. Data were acquired using the Xcalibur software (v2.0.7) and the resulting spectra were analysed via the Mascot Software (v2.3) with Proteome Discoverer (v1.2, Thermo Scientific) using the SwissProt Mus musculus database. Carbamidomethylation of cysteine, oxidation of methionine, N-terminal acetylation and heavy ¹³C₆-lysine (Lys6) were set as variable modifications. We set specificity of trypsin digestion and allowed two missed cleavage sites and mass tolerances in MS, and MS/MS were set to 2 p.p.m. and 0.8 Da, respectively. The resulting Mascot result files were further processed using myProMS⁴⁷ (v3.0), allowing a maximum FDR of 1% by automatically filtering the Mascot score at the peptide level.

Protein quantification. For SILAC-based protein quantification, peptides XICs (extracted ion chromatograms) were retrieved from Proteome Discoverer. Scale normalization computed using the ‘package limma' from R was applied to compensate for mixing errors of the different SILAC cultures as described⁴⁸. Protein ratios were computed as the geometrical mean of related peptides. To estimate ratio significance, a t-test was performed with a Benjamini–Hochberg FDR control threshold set to 0.05. All quantified proteins have at least three peptides quantified (all peptides selected). Peptide intensity ratio outliers were removed when their value was too far from the median observed in the peptide intensity ratio set for a given protein. Protein quantification ratio outliers were not computed when the identified peptide number was too different between the two channels. Proteins displaying a minimal absolute fold change ≥10% that reaches statistical significance (adjusted P value of quantification ≤0.05) were considered as differentially associated with the centrosome of activated lymphocytes. This led to the selection of 835 proteins.

GO term enrichment analysis. Protein analysis by GO term enrichment was computed based on annotation only and did not take into account the relative abundance of the 835 proteins in resting and activated lymphocytes. The frequency of each GO was computed in the Mus musculus proteome (defined as the background, Slim Ontology file including all 21,283 mouse proteins) and compared with the set. We reported only the GO terms with the frequency statistically enriched in our protein set compared with the background. GO enrichment factors were computed with the GO::TermFinder⁴⁹ through myProMS. Briefly, to determine whether any GO term annotates a specified list of proteins at a frequency greater than the one expected by chance, GO::TermFinder calculates a P value using a hyper-geometric distribution. For multiple testing corrections, FDR was controlled and set to 1% (Benjamini–Hochberg). A P value was associated to each GO term individually. The FDR corresponds to the cutoff applied to the list of all the GO terms.