Volume 83, Issue 6, Pages 1052-1064 (June 2013) Molecular fingerprinting of the podocyte reveals novel gene and protein regulatory networks  Melanie Boerries, Florian Grahammer, Sven Eiselein, Moritz Buck, Charlotte Meyer, Markus Goedel, Wibke Bechtel, Stefan Zschiedrich, Dietmar Pfeifer, Denis Laloë, Christelle Arrondel, Sara Gonçalves, Marcus Krüger, Scott J. Harvey, Hauke Busch, Joern Dengjel, Tobias B. Huber  Kidney International  Volume 83, Issue 6, Pages 1052-1064 (June 2013) DOI: 10.1038/ki.2012.487 Copyright © 2013 International Society of Nephrology Terms and Conditions

Figure 1 Podocyte isolation and bioinformatic workflow. (a) High-efficiency podocyte harvest: podocytes from Gt(ROSA)26Sortm4(ACTB-tdTomato,-EGFP)Luo/J x hNPHS2Cre mice were isolated. Magnetic beads were perfused through the renal artery into the kidneys, and after mincing and digestion, glomerula could be isolated. A single cell suspension was obtained by further digestion and was subsequently used for fluorescence-activated cell sorting (FACS) analysis. Sorted podocytes and non-podocytes were used for transcriptome and proteome analysis. (b) Workflow of transcriptome, miRNA, and proteome analysis to obtain a general overview of podocyte networks. FITC, fluorescein isothiocyanate; GO, gene ontology; PE, phycoerythrin. Kidney International 2013 83, 1052-1064DOI: (10.1038/ki.2012.487) Copyright © 2013 International Society of Nephrology Terms and Conditions

Figure 2 Transcriptome analysis of podocytes and non-podocytes. (a) Principal component analysis of podocyte and non-podocyte transcriptome data (shown in green and red, respectively). The symbols represent each biological replicate. The dashed lines show corresponding biological samples. (b) Comparison of podocyte and non-podocyte gene expression. The heatmap depicts 5232 differentially regulated genes (moderated t-test, Benjamini–Hochberg corrected q-value <0.001). Rows have been hierarchically clustered using the Euclidian distance between expression value vectors. The most significant gene ontology terms obtained by conditional hypergeometric testing are shown on the left. Kidney International 2013 83, 1052-1064DOI: (10.1038/ki.2012.487) Copyright © 2013 International Society of Nephrology Terms and Conditions

Figure 3 Label-free quantification (LFQ) protein analysis of fluorescence-activated cell sorting (FACS)-sorted primary glomerular cells. (a) Experimental setup. Podocytes and non-podocytes from double transgenic mice were isolated by FACS. Cells were prepared as described in Figure 3a. (b) Volcano plot of protein intensity difference and significance of regulation (-log P-value) using a two-sided t-test. Protein intensity differences were identified in three biological replicates, each measured in two technical replicates. Significantly altered proteins are annotated as red squares (S0: 0.5; permutation-based false discovery rate 0.05). Black circles mark proteins that were also detected as significantly enriched in the SILAC-based experiments. (c) Significantly enriched and depleted gene ontology-term ‘cellular compartment’ of podocyte-specific proteins (Fishers’s exact test; P<0.01). (d) Correlation between SILAC-based and label-free quantification (LFQ) approach. Protein intensities of the LFQ approach were used to calculate ratios between both cell types and compared with SILAC ratios. A total of 611 proteins were detected and quantified by both methods. (e) Western blot analysis of significantly enriched proteins in the two cell types using α-tubulin as loading control. SDS–PAGE, sodium dodecyl sulfate–polyacrylamide gel electrophoresis. Kidney International 2013 83, 1052-1064DOI: (10.1038/ki.2012.487) Copyright © 2013 International Society of Nephrology Terms and Conditions

Figure 4 Comparison of fold differences of label-free quantification (LFQ) identified proteins with respective transcriptome differences. Significantly upregulated genes in podocytes and non-podocytes are shown as green and red dots, respectively. Yellow dots denote the upregulated proteins without a corresponding significant difference on the transcriptome level. Small black dots denote genes/proteins without significant differential regulation. Circles represent differentially regulated proteins identified with the label-free quantification approach. Linear regression is shown for podocytes (green) and non-podocytes (red). Slope and P-value of the Pearson correlation is shown in black. Kidney International 2013 83, 1052-1064DOI: (10.1038/ki.2012.487) Copyright © 2013 International Society of Nephrology Terms and Conditions

Figure 5 Gene set enrichment of gene ontology (GO)-term groups overrepresented in podocytes. Yellow and blue circles define GO terms (all ontologies) that are overrepresented in the label-free quantification proteome and transcriptome, respectively. The size of the circles corresponds to the group size. A connection between groups is drawn, if at least 20% of the genes of the smaller group are shared with the larger group. Titles for functional clusters have been added manually. Kidney International 2013 83, 1052-1064DOI: (10.1038/ki.2012.487) Copyright © 2013 International Society of Nephrology Terms and Conditions

Figure 6 Genome-wide quantitative study of alternative splicing probability based on exon expression. (a) Volcano plot of fold change difference and splicing probability (-log10 of P-value). Fold change denotes the difference in log2 gene expression values of podocytes and non-podocytes. Colored dots denote genes with a significant splicing probability (P<0.0001). Red and green dots denote differentially regulated genes in non-podocytes and podocytes (Benjamini–Hochberg corrected q-values <0.01, log2 fold difference >0.5). Yellow dots mark genes that are not differentially regulated, yet alternatively spliced. (b) Probe and probeset expression of sample genes Rapgef5 and Syne1 in podocytes (green), non-podocytes (red), and embryonic (E13.5, GSE17142/43/45) podocytes (blue). Vertical gray bars mark the probe and probeset borders. The gray slanted lines between the expression plots show the probe to probeset and probe to exon relationship. Exon locations along the chromosomal locations are marked below in orange. The known transcripts and the Affymetrix model on the exon chip are shown in blue and green, respectively. The star and dagger to the right of the Syne1 exon map mark the alternative transcripts used in non-podocytes/embryonic podocytes and podocytes, respectively. Kidney International 2013 83, 1052-1064DOI: (10.1038/ki.2012.487) Copyright © 2013 International Society of Nephrology Terms and Conditions