The Deep Phosphoproteome: A Stem Cell Data Set

Researchers use human pluripotent stem cells (hPSCs) to model human development and disease, including research related to regenerative medicine and drug discovery. While large-scale hPSC analyses exist, deep characterization of pluripotency, lineage specification and reprogramming via proteomic analysis, especially using advanced technology like label-free quantification (LFQ), remains undone.

To this end, Singec et al. (2016) compared proteomic and phosphoproteomic changes between human embryonic stem cells (hESCs) and their multipotent neural stem cell derivatives (hNSCs).¹ They employed LFQ for deep analysis as well as a systems-level study of cell signaling pathways and protein families. The research team also mapped epigenetic proteins and compiled a large phosphoproteomics transcription factor data set with novel phosphorylation sites.

Logistically, the researchers cultivated pluripotent cells under feeder-free conditions and treated them for six days to suppress pluripotency and non-neural differentiation pathways. To show multipotency, they differentiated the hNSCs into mature cells from the three main neural lineages (neurons, astrocytes and oligodendrocytes) and measured neuronal activity as well as spontaneous synaptic currents, action potential generation and ion channel response.

Then, they digested phosphoproteins into peptides and performed liquid chromatography–electrospray ionization tandem mass spectrometry (LC-ESI-MS/MS) using an LTQ Orbitrap Velos mass spectrometer (Thermo Scientific) with LFQ to enhance quantitative reliability. They searched resultant MS/MS data against the International Protein Index human protein database, v. 3.73, via Sorcerer-SEQUEST. They relied on KEGG, Ingenuity, GeneGo, Cytoscape and Adobe Photoshop to analyze and process signaling pathways, protein networks and maps.

Singec et al. detected 12,904 total proteins (i.e., the phosphoproteome), including 1,752 proteins unique to hESCs (13.6%), 1,358 proteins unique to hNSCs (10.5%) and 9,794 common proteins (75.9%). Of these, most (9,864 proteins; 76.4%) were phosphorylated at almost 60,000 non-redundant sites. The hESC-unique proteins included many associated with pluripotency (ESRRB, UTF1, FOXO1, ALPI, TERT and TDGF1), while the hNSC-unique proteins contained neural markers (PAX6, PLZF, SOX5, SOX10, LHX1, SEZ6 and ARX). Enriched protein networks for hESCs included structural and regulatory proteins, while those enriched for hNSCs related to chromatin modification, neurogenesis, axonal guidance, cell adhesion, androgen receptor signaling and microtubules.

The data set also included 487 transcription factors, 416 associated with hESCs and 419 linked to hNSCs. Of the total detected transcription factors, 75.1% were phosphorylated. The team reported novel phosphorylation sites, including two on OCT4 (pT234 and pY291), two on NANOG (pS10 and pS52) and five on SOX2 (pY2, pT7, pT17 or pS18, pS257 and pS290). The authors highlight that these findings represent considerable movement toward comprehensive mapping of transcription factors and also underscore the importance (and understudied nature) of post-translational transcription factor regulation.

Other notable aspects of the data set included 479 protein kinases (92.5% detected in hESCs and hNSCs), 154 phosphatases (144 detected in hESCs and 139 detected in hNSCs), 415 identified biological pathways (including 81% and 89% coverage for transforming growth factor β and WNT, respectively) and 273 proteins associated with epigenetic changes and chromatin modifications (87% phosphorylated).

Finally, the research team produced functional experiments to show the utility of their data set. For this, they knocked down the secreted protein midkine (MDK), which they ultimately reported as upregulated during neural commitment as a promoter of neural specification. The researchers call for further investigation into the role MDK plays in PSC differentiation.

Overall, Singec et al. offered both their protocols and the resultant data set as resources for the scientific community, particularly for studies related to pluripotency, differentiation and reprogramming. These are freely available as curated tables, an interactive website and raw data files (PeptideAtlas database, Institute for Systems Biology).    

Reference
1. Singec, I., et al. (2016) “Quantitative analysis of human pluripotency and neural specification by in-depth (phospho)proteomic profiling,” Stem Cell Reports, 7(3) (pp. 527–542), doi: 10.1016/j.stemcr.2016.07.019.