Increased proteasome activity in human embryonic stem cells is regulated by PSMD11

Abstract

Embryonic stem cells can replicate continuously in the absence of senescence and, therefore, are immortal in culture. Although genome stability is essential for the survival of stem cells, proteome stability may have an equally important role in stem-cell identity and function. Furthermore, with the asymmetric divisions invoked by stem cells, the passage of damaged proteins to daughter cells could potentially destroy the resulting lineage of cells. Therefore, a firm understanding of how stem cells maintain their proteome is of central importance. Here we show that human embryonic stem cells (hESCs) exhibit high proteasome activity that is correlated with increased levels of the 19S proteasome subunit PSMD11 (known as RPN-6 in Caenorhabditis elegans) and a corresponding increased assembly of the 26S/30S proteasome. Ectopic expression of PSMD11 is sufficient to increase proteasome assembly and activity. FOXO4, an insulin/insulin-like growth factor-I (IGF-I) responsive transcription factor associated with long lifespan in invertebrates, regulates proteasome activity by modulating the expression of PSMD11 in hESCs. Proteasome inhibition in hESCs affects the expression of pluripotency markers and the levels of specific markers of the distinct germ layers. Our results suggest a new regulation of proteostasis in hESCs that links longevity and stress resistance in invertebrates to hESC function and identity.

Figure 1. Increased proteasome activity in hESCs and iPSCs

a, Chymotrypsin-like proteasome activity measured fluorometrically by digestion of the peptide Z-GGL-AMC. Proteasome activity (relative slope to H9 hESCs) represents the mean ± s.e.m. (H9 hESCs (n=11), NPCs (n=13), neurons (n=10), (P<0.00001)). b, Representative immunoblot of polyubiquitinylated protein levels. β-actin loading control. Total protein was visualized by Coomassie staining in a corresponding protein gel. c, Caspase-like (Z-LLE-AMC) proteasome activity (relative slope to H9 hESCs) represents the mean ± s.e.m. (n=5, (P<0.001)). d, Trypsin-like (Ac-RLR-AMC) proteasome activity (relative slope to H9 hESCs) represents the mean ± s.e.m. (n=5, (P<0.0001)). e, Proteasome activity (relative slope to H9 hESCs) represents the mean ± s.e.m. (H9 hESCs (n=6), 2 days of differentiation into trophoblasts (n=6), 5 days of differentiation (n=6), trophoblasts (n=7)). H9 hESCs lose their high proteasome activity in a progressive manner when they differentiate into trophoblasts (H9 hESCs vs 2 days of differentiation into trophoblasts (P<0.001), H9 hESCs vs 5 days of differentiation (P=7.7*10⁻⁷), H9 hESCs vs 8 days of differentiation (P<0.0001)). f, Proteasome activity (relative slope to H9 hESCs) represents the mean ± s.e.m. (n=3), (P<0.001). g, Chymotrypsin-like proteasome activity (relative slope to BJ fibroblasts) represents the mean ± s.e.m. (BJ fibroblasts (n=10), iPSC line 1 (n=10), iPSC line 2 (n=9), H9 hESCs (n=5)). iPSC lines derived from BJ fibroblast display increased proteasome activity compared to fibroblasts (P<0.0005) and no significant differences compared to H9 hESCs (iPSC line 1 vs H9 hESCs (P=0.11), iPSC line 2 vs H9 hESCs (P=0.29). Statistical comparisons in Figure 1 were made by Student’s t-test for unpaired samples.

Figure 2. Increased proteasome assembly in hESCs dependent upon PSMD11 expression

a, Chymotrypsin-like proteasome activity (relative slope to H9 hESCs + 0.025% SDS) represents the mean ± s.e.m. (H9 hESCs (n=6), NPCs (n=5), neurons (n=5), H9 hESCs + 0.025% SDS (n=5), NPCs + 0.025% SDS (n=4), neurons + 0.025% SDS (n=5)). Increased proteasome activity in hESCs (P<0.01). No significant differences were found among the different cells when SDS was added (H9 hESCs + 0.025% SDS vs NPCs + 0.025% SDS (P=0.25), H9 hESCs + 0.025% SDS vs neurons + 0.025% SDS (P=0.09)). 0.025% SDS was added to cell lysates 5 minutes prior to digestion assay. b, Graph (relative expression to H9 hESCs) represents the mean ± s.e.m. (H9 hESCs (n=10), NPCs (n=6), neurons (n=8), (P<0.00001)). c, Western blot analysis with antibodies to PSMD11 and PSMD1. β-actin loading control. d, Graph (relative expression to H9 hESCs) represents the mean ± s.e.m. (n=4, (P<0.05). e, Western blot analysis of PSMD11 in trophoblasts. β-actin loading control. f, Graph (relative expression to H9 hESCs) represents the mean ± s.e.m. (n=4, (P<0.05)). g, Western blot analysis of PSMD11 in fibroblasts. β-actin loading control. h, Graph (relative expression to BJ fibroblasts) represents the mean ± s.e.m. (BJ fibroblasts (n=10), iPSC line 1 (n=6), iPSC line 2 (n=6), (P<0.00001)). i, Western blot analysis of PSMD11 in iPSC lines. β-actin loading control. j, Proteasome activity (relative slope to LV-Non-targeting shRNA) represents the mean ± s.e.m. (Non-targeting shRNA (n=10), PSMD11 shRNA 1 (n=8), PSMD11 shRNA 2 (n=6), PSMC2 shRNA 1 (n=6), PSMC2 shRNA 2 (n=3), (P<0.01)). k, Native gel electrophoresis followed by western blot with alpha 1+2+3+5+6+7 (20S subunit) or PSMD2 (19S subunit) antibodies. l, Chymotrypsin-like proteasome activity (relative slope to GFP OE HEK293T cells) represents the mean ± s.e.m. (GFP OE (n=4), PSMD11 OE (n=5), (P<0.005)). m, n, Native gel electrophoresis followed by proteasome activity assay with chymotrypsin-like activity substrate LLVY-AMC and immunoblotting with PSMD1 (19S subunit) antibody. Extracts were resolved by SDS-PAGE and immunoblotting for analysis of PSMD11 levels and loading control. Statistical comparisons in Figure 2 were made by Student’s t-test for unpaired samples.

Figure 3. FOXO4 regulates proteasome activity in hESCs

a, Chymotrypsin-like proteasome activity in H9 hESCs transiently infected with lentiviruses (relative slope to non-infected cells) represents the mean ± s.e.m. (n=19, (P<0.00001)). b, Chymotrypsin-like proteasome activity (relative slope to GFP cells) represents the mean ± s.e.m. (GFP (n=7), 3′UTR FOXO4 shRNA 1(n=6), 3′UTR FOXO4 shRNA 2 (n=3), 3′UTR FOXO4 shRNA 3 (n=6)). Knockdown of FOXO4 decreases proteasome activity in stable infected H9 hESCs (GFP vs 3′UTR FOXO4 shRNA 1 (P<0.01), GFP vs 3′UTR FOXO4 shRNA 2(P<0.01), GFP vs 3′UTR FOXO4 shRNA 3 (P=4.5*10⁻⁸)). c, FOXO4 levels are down-regulated when H9 hESCs differentiate into NPCs, neurons ((H9 hESCs (n=6), NPCs (n=4), neurons (n=4)), trophoblasts (H9 hESCs (n=6), trophoblasts (n=6)) and fibroblasts (H9 hESCs (n=4), fibroblasts (n=4)). No significant differences were found between NPCs and neurons (P=0.43). Graphs represent the mean ± s.e.m. of the relative expression levels to H9 hESCs. FOXO4 levels are up-regulated when BJ fibroblasts are reprogrammed into iPSCs (graph represents the mean ± s.e.m. of the relative expression levels to BJ fibroblasts (BJ fibroblasts (n=8), iPSC line 1 (n=6), iPSC line 2 (n=6)). *(P<0.05), **(P<0.01), ***(P<0.001). d, Proteasome activity (relative slope to non-infected H9 hESCs) represents the mean ± s.e.m. (n=7). Transient overexpression of FOXO4 AAA up-regulates chymotrypsin-like proteasome activity in H9 hESCs (non-infected cells vs LV-FOXO4 OE cells (P=0.41), non-infected cells vs LV-FOXO4 AAA OE cells (P<0.05)). e, Proteasome activity (relative slope to GFP hESCs) represents the mean ± s.e.m. (n=4, 3′UTR FOXO4 shRNA 3 vs 3′UTR_3 FOXO4 shRNA + FOXO4 AAA (P<0.01)). f, Knockdown of FOXO4 decreases expression of PSMD11 in H9 hESCs (GFP vs FOXO4 shRNA (P<0.001), GFP vs 3′UTR FOXO4 shRNA 2 (P<0.05), GFP vs 3′UTR FOXO4 shRNA 3 (P<0.001)). Graph represents the mean ± s.e.m (LV-GFP (n=15), LV-FOXO4 shRNA (n=19), LV-3′UTR FOXO4 shRNA 2 (n=4), LV-3′UTR FOXO4 shRNA 3(n=5)). Stable overexpression of FOXO4 AAA mutant increases PSMD11 expression in H9 hESCs (P=0.69 GFP vs FOXO4 OE, P<0.01 GFP vs FOXO4 AAA OE). Data represent the mean ± s.e.m. of the relative expression levels to GFP hESCs (GFP (n=7), FOXO4 OE (n=8), FOXO4 AAA OE (n=7)). g, Western blot analysis of PSMD11 levels. β-actin loading control. h, Proteasome activity (relative slope to GFP H9 hESCs) represents the mean ± s.e.m. (n=4, GFP vs FOXO4 shRNA (P<0.01), GFP vs FOXO4 shRNA+ PSMD11 OE (P=0.50)). Statistical comparisons in Figure 3 were made by Student’s t-test for unpaired samples.

Figure 4. Acute proteasome inhibition affects pluripotency of hESCs

a, Real Time PCR analysis of pluripotency (OCT4, NANOG, SOX2, UTF1, DPPA4, DPPA2, ZFP42 and TERT), trophectodermal (CDX2), ectodermal (PAX6, FGF5), mesodermal (MSX1) and endodermal (AFP, GATA6, GATA4, Albumin) germ layer markers. Proteasome inhibition (62.5 nM MG-132 24h) in H9 hESCs induces a decrease in pluripotency markers and modified the levels of markers of the distinct germ cell and extraembryonic layers P-value: *(P<0.05), **(P<0.01), ***(P<0.001). Graph (relative expression to DMSO control H9 hESCs) represents the mean ± s.e.m. (DMSO (n=12), MG-132 (n=13). Statistical comparisons in Figure 4 were made by Student’s t-test for unpaired samples. b, Western blot analysis with antibodies to SOX2, PAX6, FGF5 and MSX1. β-actin loading control.