The proteins secreted by human tissues (the secretome) are important for the basic understanding of human biology, but also for identification of potential targets for future diagnosis and therapy. Here, we present an annotation of all predicted secreted proteins (n=2,623) with information about their spatial distribution in the human body. A high-throughput mammalian cell factory was established to create a resource of recombinant full-length proteins. This resource was used for phenotypic assays involving β-cell dedifferentiation and for development of targeted proteomics assays. A comparison between host cells, including omics analysis, shows that many of the proteins that failed to be generated in CHO cells could be rescued in human HEK 293 cells. In conclusion, the human secretome has been mapped and characterized to facilitate further exploration of the human secretome.

Aliquots of the secretome protein products were diluted into a 96-well plate in a 50 mM ammonium bicarbonate (ABC) solution to a final concentration of 1.5 µM based on concentrations analyzed by absorbance measurements. All pipetting was done using a Tecan Freedom EVO system (Tecan Group Ltd., Männedorf, Switzerland). 15 pmol of each secretome protein product was transferred to a into a new 96-well plate using a CyBi-SELMA (Analytik Jena AG, Jena, Germany) and reduced (5mM DTT, 30 min at 56 °C), alkylated (2 chloroacetamide, 30 min in darkness) and cleaved with 200 ng of proteomics grade porcine trypsin (Sigma Aldrich) overnight at 37 °C. After digestion the samples were vacuum dried and stored at -20 °C awaiting LC-MS/MS analysis. Peptides from the previously prepared MS samples were pooled in sets of eight proteins and analyzed using the same LC-MS/MS setup as for the analysis of cell supernatants. Peptides corresponding to 1 pmol per protein were first loaded onto a trap column, washed 5 min at 5 μl/min with 100% of solvent A, separated on a 25 cm analytical column (Supplementary Table 10) using a linear 90 min gradient of 6-28% solvent B at a flow rate of 0.400 µl/min and analyzed using a Top5 method (Supplementary Table 12). Raw files were searched in MaxQuant version 1.5.2.8 against the Secretome protein sequences (Supplementary Table 13) with the same settings as described above and the msms files from the MaxQuant searches were used to build a spectral library in Skyline. Data for peptides present in the library was imported from the raw data files. All peptide peaks were manually verified and after removal of low quality fragment spectra, a new spectral library was exported excluding low specificity b2 and y2 fragments. Two different pools of plasma samples were digested essentially as described (https://doi.org/10.1101/472662). The DIA analysis was performed on the same instrument setup and LC-gradient as described above for spectral library generation, but two of the samples were analyzed using a 50 cm column (Supplementary Table 10) instead of on the 25 cm column previously used. For all samples a total of 1 µg of peptides was injected on the column and analyzed in DIA mode (Supplementary Table 14) with inclusion list first covered the 400-1,020 m/z range and then with a 10 m/z shift from 390-1,010 m/z before starting over again with a new MS1 scan. One sample was also injected 12 times and analyzed in DIA using small isolation windows of only 3 m/z with each injection covering a different 50 m/z range, but together covering the same range as the other DIA methods (Supplementary Table 15). Raw data files were imported in Skyline using the manually curated spectral library. Peaks with matching fragmentation spectra were integrated and MS2 peak intensities were exported for peptides that could be detected in the plasma samples.

Anonymized EDTA treated human plasma pools were digested as described by Edfors et al. (https://doi.org/10.1101/472662) and used for assessing the DIA assays.