Deciphering the tissue specificity of the protein secretory pathway in humans

Proteins that are components of the secretory machinery form a cellular pathway of paramount importance for physiological regulation, development and function of human tissues. Consistently, most secretory pathway components are ubiquitously expressed in all tissues. At the same time, recent studies identified that the largest fraction of tissue-specific proteins consists of secreted and membrane proteins and not intracellular proteins. This suggests that the secretory pathway is distinctively regulated in a tissue-specific fashion. However, a systematic analysis on how the protein secretory pathway is tuned in different tissues is lacking, and it is even largely unexplored if the secretome and membrane proteome differs in, for example, posttranslation modifications across tissues. Here, analyzing publically available transcriptome data across 30 human tissues, we discovered the expression level of key components previously categorized as housekeeping proteins were specifically overexpressed in a certain tissue compared with the average expression of their corresponding secretory pathway subsystem (e.g. protein folding). These extreme genes define an exceptional fine-tuning in specific subnetworks, which neatly differentiated for example the pancreas and liver from 30 other tissues. Moreover, the subnetwork expression tuning correlated with the nature and number of post translational modification sites in the pancreas or liver-specific secretome and membrane proteome. These patterns were recurrently observed also in other tissues, like the blood, the brain and the skeletal muscle. These findings conciliate both the housekeeping and tissue-specific nature of the protein secretory pathway, which we attribute to a fine-tuned regulation of defined subnetworks in order to support the diversity of secreted proteins and their modifications. . CC-BY-NC-ND 4.0 International license peer-reviewed) is the author/funder. It is made available under a The copyright holder for this preprint (which was not . http://dx.doi.org/10.1101/070870 doi: bioRxiv preprint first posted online Aug. 24, 2016;