Integrative omics connects N-glycoproteome-wide alterations with pathways and regulatory events in induced pluripotent stem cells

Abstract

Molecular-level differences ranging from genomes to proteomes, but not N-glycoproteomes, between human induced pluripotent stem cells (hiPSCs) and embryonic stem cells (hESCs) have been assessed to gain insights into cell reprogramming and induced pluripotency. Our multiplexed quantitative N-glycoproteomics study identified altered N-glycoproteins that significantly regulate cell adhesion processes in hiPSCs compared to hESCs. The integrative proteomics and functional network analyses of the altered N-glycoproteins revealed their significant interactions with known PluriNet (pluripotency-associated network) proteins. We found that these interactions potentially regulate various signaling pathways including focal adhesion, PI3K-Akt signaling, regulation of actin cytoskeleton, and spliceosome. Furthermore, the integrative transcriptomics analysis revealed that imperfectly reprogrammed subunits of the oligosaccharyltransferase (OST) and dolichol-phosphate-mannose synthase (DPM) complexes were potential candidate regulatory events for the altered N-glycoprotein levels. Together, the results of our study suggest that imperfect reprogramming of the protein complexes linked with the N-glycosylation process may result in N-glycoprotein alterations that affect induced pluripotency through their functional protein interactions.

Figure 1. Schematic workflow of the experimental design and summary of the N-glycoproteomics results.

(a) N-glycopeptides were enriched from nine cell lines (five hiPSCs, two hESCs, two parental somatic cells (SCs)) using the hydrazide chemistry approach, and were analyzed in triplicates using LTQ-Orbitrap mass spectrometry (MS). The raw MS files were processed by MaxQuant and the resulting N-glycoproteomics profiles were quantitatively compared using in silico and manual data analyses. Furthermore, the N-glycoproteomics profiles were integrated with the proteomics profiles (Gra1 and H9 Cells) and transcriptomics profiles (except for the Gar7 cells) to explore the biological significance of this study. (b) The Venn diagram represents the overlap between the N-glycoproteins identified in the nine cell lines from the three groups. In addition, the total numbers of N-glycoproteins, N-glycopeptides, and N-glycosites identified are shown. (c) The total numbers of N-glycoproteins and N-glycosites identified in each of the nine cell lines are shown. (d) The total numbers of N-glycoproteins and N-glycosites that were identified in any one of the cell lines, overlapped in 2 or more cell lines, and overlapped in 7 or more cell lines are shown.

Figure 2. The hiPSC type-specific N-glycoprotein alterations and the regulated biological processes and signaling pathways.

(a) The table shows the hiPSC type-specific N-glycoprotein alterations identified in 19 quantitative comparisons of hiPSCs and hESCs as well as PSCs (hiPSCs or hESCs) and SCs. The numbers of comparable and differentially regulated (≥2-fold, and CV ≤ 20%) N-glycoproteins and N-glycosite-specific events are shown. The most significant biological processes (b) cellular components (c) and signaling pathways (d) linked with the altered N-glycoproteins identified in hiPSCs compared to hESCs are shown. The biological processes and cellular components are Gene Ontology terms whereas the signaling pathways are KEGG terms. These terms were enriched with the significance (p-value) < 0.0001 and a false discovery rate (FDR) < 0.01 using the DAVID database.

Figure 3. Integrative proteomics-based validation of the hiPSC type-specific N-glycoprotein alterations.

(a) The integrative approach shows the validation of the hiPSC type-specific N-glycoprotein alterations (n = 62) identified in Gra1 cells (hiPSCs) compared to H9 cells (hESCs) using their corresponding quantitative proteomics data set. Among the 62 altered N-glycoproteins, 46 were calibrated using their protein expression levels, and 42 of 46 were identified as truly altered in Gra1 cells. (b) We show the accession numbers (IPI ids), symbols, fold changes before and after calibration, and site-specific N-glycopeptides of the 42 N-glycoproteins represented in Fig. 3a. The color bar represents the fold change values.

Figure 4. The protein interaction network analysis of hiPSC type-specifically altered N-glycoproteins and PluriNet proteins identifies the functional relationships and regulated signaling pathways.

(a) Functional relationships between the proteins identified as containing N-glycoprotein alterations in Gra1 hiPSCs and known PluriNet (pluripotency-associated network) proteins are shown. Altered N-glycoproteins are denoted with red (up-regulated) and green (down-regulated) asterisk (*), and the PluriNet proteins are shown without an asterisk. The altered N-glycoproteins that showed at least one direct and high confidence relationship with PluriNet proteins were considered to generate the subnetwork in the STRING database. The subnetwork matrices (coefficient clustering and mean shortest path (p < 0.001)) are significant compared to the 1,000 randomized networks. (b) The list of KEGG pathways that was enriched in subnetwork is shown, along with the corrected p-values.

Figure 5. The integrative transcriptomics analysis identifies that imperfect reprogramming of potential regulatory complexes may be responsible for the altered N-glycoprotein levels in hiPSCs.

(a) Relative gene expression levels of the OST and DPM complex subunits observed in two hESCs (H9 and NTU1), two hiPSCs (Gra1 and Gra2), and SCs (HGra). In addition, the relative expression levels of a housekeeping gene (GAPDH) in these cell lines are shown. The transcriptomics-based gene expression levels were obtained from two biological replicates. The average gene expression levels were derived from replicate analyses and the significant differences between hiPSCs and hESCs were defined using t-test. (b) Relative gene expression levels of the OST and DPM complex subunits observed in two hESCs (H9 and NTU1), two hiPSCs (CBF46 and CBF50), and SCs (HF). The other details are the same as described in Fig. 5a.

Figure 6. Potential implications of the current study for improving the somatic cell reprogramming and induced pluripotency of hiPSCs.