CD2AP regulates SUMOylation of CIN85 in podocytes

Abstract

Podocytes are highly differentiated and polarized epithelial cells located on the visceral side of the glomerulus. They form an indispensable component of the glomerular filter, the slit diaphragm, formed by several transmembrane proteins and adaptor molecules. Disruption of the slit diaphragm can lead to massive proteinuria and nephrotic syndrome in mice and humans. CD2AP is an adaptor protein that is important for the maintenance of the slit diaphragm. Together with its paralogue, CIN85, CD2AP belongs to a family of adaptor proteins that are primarily described as being involved in endocytosis and downregulation of receptor tyrosine kinase activity. We have shown that full-length CIN85 is upregulated in podocytes in the absence of CD2AP, whereas in wild-type cells, full-length CIN85 is not detectable. In this study, we show that full-length CIN85 is postranslationally modified by SUMOylation in wild-type podocytes. We can demonstrate that CIN85 is SUMOylated by SUMO-1, -2, and -3 and that SUMOylation is enhanced in the presence of CD2AP. Conversion of lysine 598 to arginine completely abolishes SUMOylation and leads to increased binding of CIN85 to nephrin. Our results indicate a novel role for CD2AP in regulating posttranslational modification of CIN85.

Fig 1

Phylogenetic and sequence analyses of CIN85 and CD2AP. (A) Phylogenetic tree of CD2AP and CIN85 genes in animals. A minimum evolution analysis rooted on the gene sequence from Hydra vulgaris (Hvu) strongly supports the hypothesis (bootstrap values are indicated at the respective nodes; nodes with values below 95% are collapsed into polytomies) that these two genes are the result of a gene duplication event (circle) early in the evolutionary history of vertebrates, before the split between chondrichthyians (Ler, the skate L. erinacea) and the bony fish. These two genes, presumably the result of a genome duplication events(s), are related to single-copy genes found throughout the invertebrates, including the freshwater cnidarian Hydra. Note that according to this phylogenetic analysis, this gene evolved before the evolution of podocytes, a cell type that characterizes the protostomes and deuterostomes, including vertebrates, but is not found in more basal animals, including cnidarians. Aca, the lizard Anolis carolinensis; Ame, the honeybee, Apis melifera; Bfl, the amphioxus Branchiostoma floridae; Cin, the sea squirt Ciona intestinalis; Dre, the zebrafish, Danio rerio; Dpu, the waterflea Daphnia pulex; Gga, the chicken, Gallus gallus; Hro, the leech Helobdella robusta; Hsa, Homo sapiens; Mmu, the mouse, Mus musculus; Sko, the hemichordate worm Saccoglossus kowalevskii; Spu, the purple sea urchin, Strongylocentrotus purpuratus; Tca, the beetle Tribolium castaneum; Tgu, the zebrafinch, Taeniopygia guttata; Xtr, the frog Xenopus laevis. (B) Protein domains and regions annotated according to human CIN85 (top) and CD2AP (bottom), including three SH3 domains (blue), a proline-rich region (red), and a predicted coiled-coil region (green).

Fig 2

Full-length CIN85 is posttranslationally modified in CD2AP^+/+ podocytes. (A) Protein lysates from two different clones of cultured differentiated CD2AP^+/+ and CD2AP^−/− podocytes were analyzed for CIN85, ubiquitin, CD2AP, and GAPDH expression by Western blotting (WB). The arrows show the masses of proteins and specific isoforms of CIN85. (Below) Densitometry summarizing the means and standard deviations of 200-, 130-, and 85-kDa CIN85/GAPDH obtained from three independent experiments. OD, optical density. *, P < 0.0005 compared to the wild type/knockout (WT/KO) by Student's t test; n.s., not significant. (B) PCR was performed on cDNA generated from cultured differentiated CD2AP^+/+ and CD2AP^−/− podocytes. Mouse-specific primers m1 (positioned in exon 1) and mmn (positioned in exon 9) were used for amplification of the 5′ region of CIN85 (lanes 1 and 2). Primer m6 (positioned in exon 19) and mc (positioned in exon 24) were used for amplification of the 3′ region of CIN85, including the coiled-coil (CC) domain (lanes 4 and 5). PCR primers m1, mmn, mc, and m6 were used as described previously (3, 10). Lanes 3 and 6, negative control. The arrows show amplified cDNAs (a, b, and c): a, 5′ region of CIN85; b, full-length 3′ region of CIN85; c, 3′ region of CIN85Δexon21. (Bottom) Hypoxanthin-phosphoribosyl transferase 1 (HPRT-1) as a loading control. (Below) Densitometry summarizing the means and standard deviations of amplified cDNAs a and b/c obtained from three independent experiments.

Fig 3

Modification of full-length CIN85 depends on expression of CD2AP. (A) pcDNA3.1, CD2AP^−/− podocytes were transiently transfected with myc-CD2AP or empty vector pcDNA3.1, and whole-cell lysates were immunoblotted for CIN85, Myc, and GAPDH. Densitometry summarizing the means and standard deviations of 130- and 85-kDa CIN85/GAPDH obtained from three independent experiments. *, P < 0.008 compared to pcDNA3.1/myc-CD2AP by Student's t test. (B) Undifferentiated CD2AP^+/+ podocytes were treated with 20 pmol of CD2AP siRNA or control siRNA (scrambled sequence); cultured for 0, 2, 4, and 8 days at 37°C for differentiation; and analyzed for CIN85, CD2AP, and GAPDH. (Right) Densitometry summarizing the means and standard deviations of 85-kDa CIN85/GAPDH obtained from three independent experiments. *, P < 0.0003 compared to CD2AP siRNA/control siRNA by Student's t test. (C) PCR was performed on cDNA generated from CD2AP^+/+ and CD2AP^−/− podocytes differentiated for 0, 2, 4, and 8 days. Amplified cDNAs show the full-length 3′ region of CIN85 (top) and HPRT-1 as a loading control.

Fig 4

SUMOylation of CIN85 in CD2AP^+/+ podocytes in vivo. (A) Fluorescence labeling demonstrates CIN85 (red) and CD2AP, SUMO-1 and -2/3, and Ubc9 (green) in renal cortex sections of 4-week-old CD2AP^+/+ and CD2AP^−/− mice. Colocalization of the indicated proteins resulted in yellow fluorescence in the merged pictures. (B) Whole lysates from isolated CD2AP^+/+ and CD2AP^−/− glomeruli were precipitated (IP) with SUMO-1 and -2/3 antibodies. The probes were blotted and analyzed for CIN85. For IgG control, glomerular lysates from CD2AP^+/+ mice were incubated with normal rabbit IgG instead of specific antibody. The lysates were analyzed for CIN85, CD2AP, and GAPDH as a control. (C) Fluorescence labeling of differentiated CD2AP^+/+ and CD2AP^−/− podocytes demonstrates CIN85 (green) and SUMO-1 and -2/3 and Ubc9 (red). Colocalization resulted in yellow fluorescence in the merged pictures. The arrows indicate colocalization in CD2AP^+/+ podocytes and membrane distribution of CIN85 in CD2AP^−/− podocytes. (D) Whole lysates from differentiated CD2AP^+/+ and CD2AP^−/− podocytes were precipitated (IP) with SUMO-1 and -2/3. The probes were blotted and analyzed for CIN85. For IgG control, lysates from CD2AP^+/+ podocytes were incubated with normal rabbit IgG instead of antibody. The lysates were analyzed for CIN85, CD2AP, and GAPDH as a control. The asterisks in panels B and D mark nonspecific IgG heavy-chain bands. (E) Subcellular fractionation of CD2AP^+/+ and CD2AP^−/− podocytes into membrane, cytosolic, and nuclear fractions showing expression of CIN85 and nephrin. As a control, membrane fractions were analyzed for Giα3, cytosolic fractions were analyzed for GAPDH, and nuclear fractions were analyzed for histone H1.

Fig 5

SUMOylation of CIN85 and of CD2AP in vitro. HEK 293 cells were cotransfected with plasmids expressing Flag-CIN85 (A) or Myc-CD2AP (B); untagged Sumo-1, -2, and -3; HA-Ubc9; and EGFP-tagged-SUMO-1. Whole-cell lysates were immunoblotted with anti-Flag, anti-Myc, or anti-HA antibody. (Below) Densitometry summarizing the means and standard deviations of SUMO-1-, -2-, and -3-modified/unmodified CIN85 or CD2AP (S1, SUMO-1; S2, SUMO-2; S3, SUMO-3) obtained from three independent experiments. *, P < 0.04 compared to S2/S3 by Student's t test. (C) HEK 293 cells were cotransfected with Flag-CIN85, Myc-CD2AP, untagged SUMO-3, and HA-Ubc9. Whole-cell lysates were immunoblotted with anti-Flag (top) or anti-Myc and anti-HA antibody (bottom). (Below) Densitometry summarizing the means and standard deviations of SUMO-3-modified/unmodified CIN85 (gray bars, without myc-CD2AP; white bars, with myc-CD2AP) obtained from three independent experiments. *, P < 0.1 by Student's t test; n.s., not significant. The arrows show expression of un-SUMOylated and SUMOylated protein.

Fig 6

Identification of the SUMO modification site in mouse CIN85. (A) Schematic representation of the putative SUMO site in murine CIN85, where lysine in position 598 is the putative acceptor amino acid. The consensus SUMOylation sequence is PKME, where P is a large hydrophobic amino acid, K is the SUMO-modified lysine residue, and E is an acidic residue. (B) HEK 293 cells were transfected with Flag-CIN85; Flag-CIN85^K598R; or untagged SUMO-1, -2, and -3 and were immunoblotted with anti-Flag antibody. The arrows show un-SUMOylated and SUMOylated protein. The immunoblot represents the results of three independent experiments. (C) HEK 293 cells were transfected with Flag-CIN85, Flag-CIN85^K598R, or pcDNA3.1 empty vector and GFP-nephrin. Whole-cell lysates were subjected to immunoprecipitation (IP) with Flag-coupled beads or, for the isotope-matched control, with normal mouse IgG and agarose beads and immunoblotted with anti-GFP or anti-Flag antibody. As a control, the lysates were immunoblotted with anti-Flag or anti-GFP. (Below) Densitometry summarizing the means and standard deviations of immunoprecipitated nephrin/nephrin in control lysates obtained from three independent experiments. *, P < 0.001 compared to Flag-CIN85/Flag-CIN85^K598R by Student's t test. (D) Podocytes were transfected with plasmids expressing GFP-tagged wild-type or mutant CIN85^K598R (green) and SV5-nephrin (red). After 48 h, the transfected cells were analyzed by fluorescence microscopy. The nucleus was stained with DAPI (blue). Overlap of CIN85^K598R and SV5-nephrin is indicated by the arrows.