Protein extraction and analysis by LC-MS/MS

Fresh CRC tissues and para-tumor normal colorectal tissues (PN) were used for proteogenomic analysis. Three different parts of the same lesions for every sample were compared for data analysis and measurement of the variation caused by random biological effects. The samples were cut into small pieces (about 1 mm³) and rinsed in PBS to remove the blood. Then the tissues were homogenized in 4% SDS and 0.1 M DTT in 0.1 M Tris-HCl, pH 7.6 on ice, sonicated 10 times (80 w; 10 s sonication/5 s suspension), incubated for 3 min at 95 °C, and briefly sonicated. The protein concentrations of clarified lysates were determined using a fluorescence assay and then 200 μg of clarified lysates were proteolyzed on a 10 kDa filter (PALL Life Sciences, Shanghai, China) using the filter-aided sample preparation method [7]. The peptide samples were then desalted onto a solid-phase extraction cartridge. The lyophilized peptide mixture was re-suspended in water with 0.1% formic acid (v/v), and its content was estimated by ultraviolet light spectral density at 280 nm [8]. Then, 3 μg of the digest sample was analyzed by nano-liquid chromatography-tandem mass spectrometry on a LTQ Orbitrap Velos Pro mass spectrometer as previously described [9].

The acquired data from mass spectrometry runs were combined and searched against the UniProt Human database (05/2016, 153,652 entrys) using Maxquant software (version 1.3.0.5; http://maxquant.org/) as described [10]. Proteins were identified using the Andromeda peptide search engine integrated into the Maxquant platform. Trypsin-digested fragments were analyzed, allowing for a maximum of 2 missed cleavages. Carbamidomethyl cysteine was set as a fixed modification, with protein N-acetylation and methionine oxidation as variable modifications. Precursor ion tolerances were 20 ppm for first search and 6 ppm for a second search. The MS/MS peaks were de-isotoped and searched using a 20 ppm mass tolerance. The required minimum peptide length for identification was 7 amino acids, and the false discovery rate at the protein level, peptide level and site were set to 0.01. The normalized spectral protein intensity (label-free quantification) values were calculated for each protein group.

The Maxquant peptide and protein quantification result files were imported into Perseus software (version 1.5.1.6) to identify the differentially expressed proteins. After importing the quantitative data from ProteinGroups.txt into Perseus, a filtering criterion is set to keep the identified proteins with the quantified values of all ten reporter ions (no missing value) in the final identification list. The protein intensities are log₂-transformed and normalized by subtracting the median intensity in each column/sample. Principal component analysis is performed based on protein intensities to differentiate groups. Two-samples tests coupled with Benjamini–Hochberg (FDR cutoff of 0.05) correction are performed to identify the differentially expressed proteins.