The ubiquitin-conjugating enzyme UBE2K determines neurogenic potential through histone H3 in human embryonic stem cells
Communications Biology volume 3, Article number: 262 (2020) Cite this article
Embryonic stem cells
Histones modulate gene expression by chromatin compaction, regulating numerous processes such as differentiation. However, the mechanisms underlying histone degradation remain elusive. Human embryonic stem cells (hESCs) have a unique chromatin architecture characterized by low levels of trimethylated histone H3 at lysine 9 (H3K9me3), a heterochromatin-associated modification. Here we assess the link between the intrinsic epigenetic landscape and ubiquitin-proteasome system of hESCs. We find that hESCs exhibit high expression of the ubiquitin-conjugating enzyme UBE2K. Loss of UBE2K upregulates the trimethyltransferase SETDB1, resulting in H3K9 trimethylation and repression of neurogenic genes during differentiation. Besides H3K9 trimethylation, UBE2K binds histone H3 to induce its polyubiquitination and degradation by the proteasome. Notably, ubc-20, the worm orthologue of UBE2K, also regulates histone H3 levels and H3K9 trimethylation in Caenorhabditis elegans germ cells. Thus, our results indicate that UBE2K crosses evolutionary boundaries to promote histone H3 degradation and reduce H3K9me3 repressive marks in immortal cells.
Embryonic stem cells (ESCs) can replicate indefinitely while retaining their potential to differentiate into all cell lineages 1 , 2 . A precisely coordinated network of transcriptional and epigenetic modifiers determines ESC identity. For instance, ESCs exhibit an endogenous transcriptional network that modulates their self-renewal and pluripotency, including transcription factors such as OCT4, NANOG and SOX2 3 . Likewise, chromatin modifiers regulate ESC pluripotency and differentiation 4 , 5 . In comparison with their differentiated counterparts, ESCs have a unique chromatin architecture such as fewer condensed heterochromatin foci, a pattern that facilitates dynamic reorganization of chromatin during development 6 . For instance, heterochromatin-associated histone modifications such as H3K9 trimethylation (H3K9me3) are usually reduced in ESCs 6 , 7 , 8 . Notably, regulators of histone modifications and chromatin compaction are required for ESC differentiation, including the polycomb repressive complex 1 (PRC1) 9 , 10 . In these lines, huntingtin protein (HTT) binds and inhibits SETDB1, a methyltransferase that specifically trimethylates lysine 9 of histone H3. Concomitantly, HTT maintains low levels of H3K9me3 in human ESCs (hESCs) 11 , 12 .
Although histones are key determinants of chromatin compaction, little is known about how their total levels are modulated. The ubiquitin-proteasome system is involved in the regulation of histone levels 13 , but the enzymes that catalyze their polyubiquitination for proteasome recognition have not been identified in multicellular organisms 14 . The attachment of ubiquitin is achieved through a sequential mechanism involving three classes of enzymes 15 . First, the ubiquitin-activating enzyme (E1) activates the C-terminal glycine residue of ubiquitin in an ATP-dependent manner. Activated ubiquitin is next transferred to a ubiquitin-conjugating enzyme (E2). In the third step, a ubiquitin ligase (E3) attaches ubiquitin from the E2 enzyme to the target protein. The same cascade can link additional molecules to the primary ubiquitin via internal ubiquitin lysines, forming a ubiquitin chain. Ubiquitination can affect numerous proteins in many manners: it can signal for their degradation through the proteasome or modulate their activity, intracellular localization, and interaction with other proteins. E2 enzymes are the main determinants for selection of the lysine to construct ubiquitin chains, which thereby control the cellular fate of the substrate. In humans, 35 E2 and over 600 E3 enzymes have been identified so far. Growing evidence indicates that ESCs not only exhibit increased proteasome activity, but also an intrinsic network of ubiquitin ligases 16 , 17 , 18 , 19 , 20 , 21 , 22 . As such, the ubiquitin-proteasome system has a central role in the immortality and cell fate decisions of pluripotent stem cells 16 , 17 , 18 , 19 , 20 , 21 , 22 .
Given their endogenous chromatin structure signature, ESCs could provide a novel paradigm to discover epigenetic regulatory mechanisms and their impact on differentiation. For this purpose, we ask whether the intrinsic ubiquitin-proteasome system of hESCs impinge on their epigenetic landscape. Notably, we find that the ubiquitin-conjugating enzyme E2 K (UBE2K), also known as huntingtin-interacting protein 2 (HIP2), is upregulated in hESCs compared with their differentiated counterparts. UBE2K reduces total histone H3 levels and trimethylation of H3K9, allowing for neurogenesis of hESCs. However, loss of UBE2K impairs the levels of SETDB1 and its regulator HTT, resulting in H3K9 trimethylation and repression of neurogenic genes during differentiation. Besides H3K9me3 regulation, UBE2K binds histone H3 and promotes its degradation primarily by 26S proteasomes, regulating total H3 protein amounts. Interestingly, UBE2K also impinges upon total H3 and H3K9me3 levels in the organismal model Caenorhabditis elegans, particularly in the immortal germline. Thus, our results provide a link between the ubiquitin-proteasome system and histone regulation in both ESCs and germ cells.
hESCs exhibit increased levels of UBE2K
To determine changes in the levels of E2 enzymes during differentiation, we analyzed available quantitative proteomics data 23 . We found 4 E2 enzymes (i.e., UBE2C, UBE2G1, UBE2K and UBE2O) increased in hESCs when compared with their neural progenitor cell (NPC) and neuronal counterparts (Supplementary Table 1 ). Notably, UBE2K markedly decreased during neural differentiation of hESCs (Supplementary Table 1 ), as we confirmed by western blot analysis (Fig. 1a ). In both NPCs and neurons, the decrease in the protein amount of UBE2K correlated with a reduction of the transcript levels (Fig. 1b ). UBE2K protein amounts also decreased during differentiation into either endoderm or mesoderm, indicating that this is not a specific phenomenon associated to neural differentiation (Fig. 1c, d ). We then asked whether high levels of UBE2K can be reprogrammed. Indeed, induced pluripotent stem cells (iPSCs) displayed increased levels of UBE2K compared with their parental fibroblasts (Fig. 1e, f ). Taken together, our results indicate that enhanced UBE2K protein expression is associated with pluripotency.
Fig. 1: The expression of UBE2K decreases during differentiation.
a Western blot analysis with antibody to UBE2K. The graph represents the UBE2K relative percentage values (corrected for β-actin loading control) to H9 hESCs (mean ± s.e.m. of three independent experiments). b Quantitative PCR (qPCR) analysis of UBE2K mRNA levels. Graph (relative expression to H9 hESCs) represents the mean ± s.e.m. of three biologically independent samples. c, d Western blot analysis with antibody to UBE2K. The graphs represent the UBE2K relative percentage values (corrected for β-actin) to H9 hESCs (mean ± s.e.m. of three independent experiments). e Western blot analysis of UBE2K levels comparing iPSCs with their parental HFF-1 fibroblasts. The graph represents the UBE2K relative percentage values (corrected for β-actin) to HFF-1 fibroblasts (mean ± s.e.m. of three independent experiments). f UBE2K mRNA relative levels to HFF1 fibroblasts represent the mean ± s.e.m. (n = 6 biologically independent samples). All the statistical comparisons were made by two-tailed Student’s t-test for unpaired samples. P value: *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001.
Full size image
With the essential role of histones in chromatin compaction and gene expression, it is of central interest to define not only the regulatory mechanisms of histone modification, but also total histone amounts. During spermatogenesis, histones are replaced by transition proteins and protamines in post-meiotic cells, a process mediated by the ubiquitin-proteasome system 47 . Moreover, somatic cells also exhibit replacement of histones at promoter regions or active gene bodies 48 . Importantly, an excess of histones impairs transcription, increases DNA-damage sensitivity, and promotes either chromosome aggregation or loss 49 . In yeast, excess histones are rapidly degraded by the proteasome in a pathway involving the E2 enzyme Ubc4/5 and E3 ligases such as Tom1 and Hel1 49 , 50 .
Despite their essential biological role, little is known about how histone levels are regulated in multicellular organisms. As the origin of multicellular organisms, pluripotent stem cells have stringent mechanisms to protect their genome 51 and proteome 18 . Since hESCs exhibit an endogenous epigenetic landscape with a more open and dynamic architecture of chromatin 52 , 53 , we hypothesized that these cells can also have intrinsic mechanisms to maintain the homeostasis of histone levels. For this purpose, we examined whether their enhanced ubiquitin-proteasome system is involved in histone regulation. Our approach led us to identify high levels of UBE2K as a determinant of histone H3 protein amounts, potentially via regulating its ubiquitination and proteasomal-mediated degradation. First, we confirmed UBE2K-mediated ubiquitination of H3 by in-vitro ubiquitination assays. In these experiments, UBE2K in combination with the E3 ligase RNF138 robustly induces the formation of ubiquitinated H3 with a molecular weight of ~24 kDa, which could indicate attachment of one ubiquitin. To a lesser extent, UBE2K-RNF138 also triggers the formation of ubiquitinated H3 with higher molecular weights. Among them, we found a ubiquitinated H3 form of ~48 kDa. According to the molecular weight, this band could correspond to H3 tagged with a chain of four ubiquitins, which is a mark for proteasomal recognition and degradation. However, in-vitro ubiquitination experiments alone cannot discard that the distinct ubiquitinated H3 forms detected in the assay might be caused by multi-monoubiquitination. Remarkably, immunoprecipitation of histone H3 supports that UBE2K overexpression triggers polyubiquitination of H3 in human cells, whereas we could not detect monoubiquitinated H3 in control or UBE2K overexpression conditions. In line with these findings, UBE2K overexpression results in degradation of histone H3, whereas the treatment with MG-132 proteasome inhibitor antagonizes these effects. Most importantly, a mild decrease in proteasome activity upon moderate knockdown of PSMD11, a key regulator of 26S proteasomes that degrade polyubiquitinated proteins, induces the strongest increase in total H3 levels when compared with activators of other proteasome assemblies involved in ubiquitin-independent proteolytic degradation 15 . For instance, even when PSMD11 and PSME4 knockdown induced a similar inhibition of total proteasome activity (~24%), only PSMD11 shRNA resulted in robust changes in histone H3 levels. Whereas these results support a prominent role of 26S proteasomes in H3 degradation, we cannot discard that other proteasome assemblies could also regulate H3 levels. For instance, it is possible that a stronger knockdown of PSME4 may induce a further reduction in proteasome activity, and eventually lead to changes in histone H3 levels. As a further evidence of a role of 26S proteasomes in H3 degradation, we found that PSMD11 knockdown blocks the degradation of H3 induced by UBE2K overexpression.
Besides total H3 levels, UBE2K also modulates H3K9 trimethylation, a histone modification associated with heterochromatin formation and gene repression. Our data indicate that UBE2K impinges upon H3K9 trimethylation by two mechanisms converging on SETDB1. First, we find that loss of UBE2K decreases the levels of HTT, an inhibitor of SETBD1 activity. In these lines, our results in C. elegans support that H3K9 trimethylation upon UBE2K knockdown requires the activity of trimethylases such as SETDB1. Besides changes in HTT, knockdown of UBE2K results in higher SETDB1 protein levels. Given the important role of HTT and SETDB1 in gene expression, it will be fascinating to define how UBE2K regulates their levels. Since UBE2K interacts with SETDB1, an intriguing hypothesis is that UBE2K directly modulates its ubiquitination and proteasomal degradation. Another possibility is that UBE2K regulates SETDB1 levels via its recruitment into histone complexes, preventing its degradation. To address these hypothesis, it will also be important to assess whether UBE2K-mediated ubiquitination and stability of H3 determines changes in the total protein amounts or interactions of key regulators of H3K9 trimethylation.
Since UBE2K particularly increased the H3K9me3 ratio, we focused on the impact of upregulated H3K9me3 in neurogenesis from hESCs. Importantly, loss of UBE2K not only increases global H3K9 trimethylation but also H3K9me3 enrichment in genes required for neural and neuronal differentiation such as GBX1, HES6 or distinct potassium voltage-gated channels. It is important to note that these changes do not have strong effects on hESCs and their commitment into early NPCs. However, H3K9me3 enrichment in the aforementioned genes diminishes their induction during differentiation into mature NPCs and neurons. Interestingly, a recent study found that low levels of H3K9me3 in hESCs allow for the establishment of compacted heterochromatin in specific protein-coding-genes at later stages, a critical process for cell differentiation 8 . Thus, UBE2K could contribute to maintaining low H3K9me3 levels in hESCs, defining their differentiation potential. In these lines, we observed that loss of UBE2K in hESCs does not affect the first steps of differentiation into early NPCs, but impairs their ability to differentiate into mature NPCs and neurons. We speculate that the H3K9me3-mediated repression of genes such as GBX1, HES6, KCNA3, KCNA5 could contribute to these phenotypes, including the lack of induction in distinct neural and neuronal markers during mid and late stages of differentiation. For instance, UBE2K knockdown does not affect the expression of the neural markers Nestin and SOX1 in early NPCs, but blocks their further induction during differentiation into mature NPCs. When compared with early NPCs, control mature NPCs also have increased levels of all the neuronal markers tested (i.e., AADC, TH+, MAP2, TUJ1, SYN1, NEUN, GABAR, NEFL, NEFM, DLG4). Importantly, loss of UBE2K in hESCs is sufficient to inhibit the induction of these factors during differentiation into mature NPCs, with the exception of TUJ1. Moreover, UBE2K knockdown also blocked the further induction of all the tested neuronal markers during terminal differentiation into neurons, including TUJ1. Despite the delay in TUJ1 alterations, our results indicate that UBE2K knockdown already has deleterious effects during differentiation into mature NPCs. Although our data suggest an important role of H3K9me3 changes in neuronal differentiation, it is uncertain whether H3K9 trimethylation or H3 stability is the key player in regulating the neurogenic potential of hESCs. For instance, dysregulated H3 levels may also contribute to diminishing neurogenic potential and even has a more prominent role in this phenotype relative to H3K9me3 changes. Whereas UBE2K knockdown does not significantly change the ratio of H3K4me3, H3K27me3, H3K27ac when normalized to total H3, it increases overall levels of these modifications. Thus, we cannot discard that upregulated H3K4me3, H3K27me3, H3K27ac amounts resulting from increased H3 levels also have a key role in regulating neurogenic potential. Another intriguing possibility is that UBE2K-mediated changes in H3 stability are the upstream signal to modulate H3K9 trimethylation. In this speculative model, changes in H3 levels could be considered as the key determinant of neurogenesis, because they would precede H3K9 trimethylation.
Remarkably, our experiments in adult C. elegans indicate that UBE2K knockdown particularly affects histone H3 and H3K9 trimethylation in the immortal germline, which generates gametes for reproduction. On the other hand, we observed that enhancing the levels of UBE2K in human cell lines is sufficient to mimic histone H3 modulation of pluripotent cells. Thus, our results indicate that a precise regulation of UBE2K levels could contribute to determining cell-type epigenetic landscapes.
hESC/iPSC culture and differentiation
Both H9 (WA09) and H1 (WA01) hESC lines were obtained from the WiCell Research Institute. The human iPSC line (ACS-1011) and their parental HFF-1 fibroblasts (SCRC-1041) were obtained from ATCC. hESCs/iPSCs were maintained on Geltrex-coated plates (ThermoFisher Scientific) using mTeSR1 media (Stem Cell Technologies). Undifferentiated hESC/iPSC colonies were passaged using a solution of dispase (2 mg ml−1), and scraping the colonies with a glass pipette. All the cell lines were tested for mycoplasma contamination at least once every three weeks and no mycoplasma contamination was detected. Research involving hESCs was performed with approval of the German Federal competent authority (Robert Koch Institute). The H9 and H1 hESCs used in this study were authenticated by short tandem repeat (STR) profile across 8 STR loci 17 .
Neural differentiation was performed following the monolayer culture method with STEMdiff Neural Induction Medium (Stem Cell Technologies) based on ref. 54 . Undifferentiated hESCs were rinsed once with PBS and then dissociated with Gentle Dissociation Reagent (Stem Cell Technologies) for 10 min. Then, we gently dislodged the cells and added 2 ml of Dulbecco’s Modified Eagle Medium (DMEM)-F12 (ThermoFisher Scientific) + 10 μM ROCK inhibitor (Abcam). Cells were then centrifuged at 300×g for 10 min and resuspended on STEMdiff Neural Induction Medium supplemented with 10 μM ROCK inhibitor. Cells were plated on polyornithine (15 μg ml−1)/laminin (10 μg ml−1)-coated plates at a density of 200,000 cells cm−2. As indicated in the corresponding figures, hESCs were cultured on neural induction medium for a total of 10 and 20 days to generate early NPCs (PAX6-positive cells) and mature NPCs with the ability to efficiently differentiate into neurons, respectively.
For neuronal differentiation, mature NPCs (20 days) were dissociated with Accutase (Stem Cell Technologies) and transferred to polyornithine/laminin-coated plates with neuronal differentiation medium Dulbecco’s Modified Eagle Medium (DMEM)/F12, B27, N2 (ThermoFisher Scientific), 1 μg ml−1 laminin (ThermoFisher Scientific), 20 ng ml−1 BDNF (Peprotech), 20 ng ml−1 GDNF (Peprotech), 1 mM dibutyryl-cyclic AMP (Sigma) and 200 nM ascorbic acid (Sigma) 22 . Cells were differentiated for 1 month, with weekly feeding of fresh neuronal differentiation medium. Endoderm differentiation of H9 hESCs was performed using STEMdiff Definitive Endoderm Kit (Stem Cell Technologies). For mesodermal differentiation, hESC colonies were dissociated with Accutase and single cells were seeded at a density of 50,000 cells cm−2 in mTeSR1 media (Stem Cell Technologies) supplemented with 10 μM ROCK inhibitor. One day after, we replaced mTeSR1 media with STEMdiff Mesoderm Induction Medium (Stem Cell Technologies). Then, cells were fed daily with STEMdiff Mesoderm Induction Medium and collected for experiments after 5 days.
Lentiviral infection of hESCs
Lentivirus (LV)-non-targeting shRNA control, LV-UBE2K shRNA #1 (TRCN0000237896), LV-UBE2K shRNA #2 (TRCN0000237893), LV-3′-UTR UBE2K shRNA (TRCN0000237895), LV-PSMD11 shRNA (TRCN0000003950), LV-PSME4 shRNA (TRCN0000158223) and LV-PSME3 shRNA (TRCN0000290094) in pLKO.1-puro vector were obtained from Mission shRNA (Sigma). To establish stable shRNA hESC lines, hESCs growing on Geltrex-coated plates were incubated with mTesR1 medium supplemented with 10 μM ROCK inhibitor for 1 h and individualized using Accutase. 50 000 cells were infected with 20 µl of concentrated lentivirus in the presence of 10 μM ROCK inhibitor for 1 h. Cell suspension was centrifuged to remove virus, passed through a mesh of 40 μM to obtain individual cells, and plated on a feeder layer of mitotically inactive mouse embryonic fibroblasts (MEFs) in hESC media (DMEM/F12, 20% knockout serum replacement (ThermoFisher Scientific), 0.1 mM non-essential amino acids, 1 mM L-glutamine, 10 ng ml−1 bFGF (Joint Protein Central) and β-mercaptoethanol) supplemented with 10 μM ROCK inhibitor. When small hESC colonies arose after a few days in culture, we performed 1 µg ml−1 puromycin selection during 2 days and colonies were manually passaged onto MEFs to generate stable knockdown hESC lines. Once stable knockdown hESC lines were established, the cells were maintained on Geltrex-coated plates using mTeSR1 media. To establish stable shRNA HEK293 lines, HEK293 cells (ATCC, #CRL-11268) were transduced with 5 µl of concentrated lentivirus and selected by adding puromycin at a concentration of 2 µg ml−1.
For the generation of UBE2K-overexpressing lentiviral constructs (UBE2K(OE)), human UBE2K complementary DNA was PCR-amplified and cloned into CD522A-1 pCDH cDNA Cloning Lentivector (System Biosciences) using NheI and BamHI. This construct was sequence verified and thereafter transfected into packaging cells to produce high titer lentiviruses. As a control, we generated lentiviral particles of CD522A-1 pCDH lacking UBE2K insert (empty vector). HEK293 cells were transduced with either empty vector or UBE2K(OE) lentiviruses and selected by adding puromycin at a concentration of 2 µg ml−1. To overexpress UBE2K in stable cell lines that were previously infected with lentivirus expressing shRNA and selected for puromycin resistance, we cloned human UBE2K into CD515B-1 pCDH cDNA Cloning Lentivector (System Biosciences), a vector that contains hygromycin resistance gene instead of puromycin resistance gene. As a control for these experiments, we generated lentiviral particles of CD515B-1 pCDH lacking UBE2K insert (empty vector). Cells were selected by adding hygromycin at a concentration of 100 μg ml−1.
We extracted RNA using RNAbee (Tel-Test Inc.). Libraries were then prepared using the TruSeq Stranded mRNA Library Prep Kit. Library preparation started with 1 µg total RNA. After selection with poly-T oligo-attached magnetic beads, mRNA was purified and fragmented using divalent cations under elevated temperature. The RNA fragments underwent reverse transcription with random primers followed by second strand cDNA synthesis with DNA Polymerase I and RNase H. After end repair and A-tailing, indexing adapters were ligated. Then, products were purified and amplified (20 µl template, 14 PCR cycles) to generate the final cDNA libraries. Following library validation and quantification (Agilent 2100 Bioanalyzer), we pooled equimolar amounts of library. Then, the pool was quantified by the Peqlab KAPA Library Quantification Kit and Applied Biosystems 7900HT Sequence Detection System. The pool was sequenced on an Illumina HiSeq 4000 sequencer with a paired- end (2 × 75 bp) protocol.
RNA-sequencing data were analyzed using a QuickNGS pipeline 55 . We performed basic read quality check and read statistics by using FastQC and SAMtools, respectively. The basic data processing consists of a splicing-aware alignment using Tophat2 56 followed by reference-guided transcriptome reassembly with Cufflinks2 57 , 58 . Read count means, fold-change (FC) and P values were calculated with DEseq2 59 , whereas gene expression for the individual samples was calculated with Cufflinks2 57 , 58 as FPKMs, using in both cases genomic annotation from the Ensembl database (version 87).
Sample preparation for quantitative proteomics and analysis
For the comparison between control and UBE2K shRNA H9 hESCs, label-free quantitative (LFQ) proteomics were performed as we described in ref. 12 . Cells were collected in urea buffer (8 M urea, 50 mM ammonium bicarbonate and 1× complete protease inhibitor mix with EDTA (Roche)), homogenized with a syringe and cleared using centrifugation (16,000 g, 20 min). Supernatants were reduced (1 mM DTT, 30 min), alkylated (5 mM iodoacetamide (IAA), 45 min) and digested with trypsin at a 1:100 w/w ratio after diluting the urea concentration to 2 M. After one day, samples were cleared (16,000 g, 20 min) and supernatant was acidified. Then, peptides were cleaned up using stage tip extraction 60 . The liquid chromatography tandem mass spectrometry (LC-MS/MS) equipment consisted of an EASY nLC 1000 coupled to the quadrupole based QExactive instrument (Thermo Scientific) via a nano-spray electroionization source. Peptides were separated on an in-house packed 50 cm column (1.9 µm C18 beads, Dr. Maisch) using a binary buffer system: (A) 0.1% formic acid, (B) 0.1 % formic acid in ACN. The content of buffer B was raised from 7 to 23% within 120 min and then increased to 45% within 10 min. Buffer B fraction was raised to 80% within 5 min and held for further 5 min after which it was decreased to 5% within 2 min and held there for further 3 min before the next sample was loaded onto the column. Eluting peptides were ionized by an applied voltage of 2.2 kV. The capillary temperature was 275 °C and the S-lens RF level was set to 60. MS1 spectra were acquired using a resolution of 70,000 (at 200 m/z), an Automatic Gain Control (AGC) target of 3e6 and a maximum injection time of 20 ms in a scan range of 300–1750 Th. In a data-dependent mode, the 10 most intense peaks were selected for isolation and fragmentation in the HCD cell using a normalized collision energy of 25 at an isolation window of 2.1 Th. Dynamic exclusion was enabled and set to 20 s. We used the following MS/MS scan properties: 17.500 resolution at 200 m/z, AGC target of 5e5 and maximum injection time of 60 ms. We analysed all LFQ proteomics data sets with MaxQuant software (release 22.214.171.124). We used the LFQ mode 61 and MaxQuant default settings for protein identification and LFQ quantification. Downstream analyses of LFQ values were performed with Perseus (v. 126.96.36.199) 62 .
Protein immunoprecipitation for interactome analysis
Protein immunoprecipitation for interactome assays were performed as we reported in ref. 63 . Cells were lysed in modified Radioimmunoprecipitation assay (RIPA) buffer (50 mM Tris-HCl (pH 7.4), 150 mM NaCl, 1 % IgPal, 0.25% sodium deoxycholate, 1 mM EDTA, 1 mM PMSF) supplemented with protease inhibitor (Roche). Lysates were centrifuged at 10,000 g for 10 min at 4 °C. Then, the supernatant was collected and incubated with anti-UBE2K antibody (Cell Signaling, #8226 1:50) for 30 min and subsequently with 100 µl Protein A magnetic beads (Miltenyi) for 1 h on the overhead shaker at 4 °C. As a control, the same amount of protein was incubated with anti-FLAG antibody (SIGMA, F7425, 1:100) in parallel. After the incubation with antibodies, supernatants were subjected to magnetic column purification followed by three washes using wash buffer 1 (50 mM Tris-HCl (pH 7.4), 150 mM NaCl, 5% glycerol and 0.05% IgPal). Next, columns were washed five times with wash buffer 2 (50 mM Tris-HCl (pH 7.4), 150 mM NaCl). After the washing steps, the pellet was incubated with 2x Laemmli buffer, boiled for 5 min and centrifuged 5 min at maximum speed. The supernatant was taken and loaded onto a sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS–PAGE) gel for western blot analysis.
Identification of ubiquitination events in histone H3
Histones were purified using a method that combines sulfuric acid extraction with ion exchange chromatography 64 , 65 . Briefly, HEK293 cells were treated with 5 µM MG-132 overnight to avoid proteasomal degradation, washed twice with ice-cold PBS and lysed by rotation in 6 ml of 0.1 M H2SO4 at 4 °C for 2 h. The lysate was centrifuged at 2200 g at 4 °C for 20 min. The pellet with non-soluble proteins and cell debris was discarded. The supernatant was neutralized in 1 M Tris-Hcl pH 8.0 buffer. Volume of lysate was made up to 15 ml by adding binding buffer (50 mM Tris pH 8.0, 0.5 M NaCl, 2 mM EDTA, 0.25 mM PMSF, 1 mM DTT). Sulfopropyl (SP)-Sepharose resin (S1799, Sigma) was packed into a column and pre-equilibrated with 10 volumes of binding buffer. 15 ml of lysate was passed through the column. The column was then washed with 10 volumes of binding buffer and 30 volumes of washing buffer (50 mM Tris-HCl, pH 8.0, 0.6 M NaCl, 2 mM EDTA, 0.25 mM PMSF, 1 mM DTT). Proteins were eluted with elution buffer (50 mM Tris-HCl, pH 8.0, 2 M NaCl, 2 mM EDTA, 0.25 mM PMSF 1 mM DTT) in ten fractions. Eluted proteins were precipitated overnight in 4% (vol/vol) PCA at 4 °C. The fractions were then centrifuged at 21,000 g at 4 °C for 45 min, and the resulting pellets were washed twice with 4% PCA, twice with 0.2% HCl in acetone and twice with acetone. The protein pellets were finally dissolved in milliQ water and protein concentration was determined by standard BCA method.
Then, 500 μg histone proteins were denatured by diluting in 8 M Urea (in 0.1 M Tris/HCl pH 8.5) in a final volume of 200 μl. Additionally, 2 μl of 1 M TCEP (final 10 mM) and 8 μl 0.5 M CAA (final 20 mM) were added and transferred to a Vivacon® 500 10 kDa cutoff filter and centrifuged at 14,000 g at 20 °C for 15 min. 200 μl of 50 mM ABC was then added and centrifuged again until half wet. This step was repeated twice. Partial digestion was then performed by adding 50 μl of 50 mM ABC with trypsin 1:2000 (trypsin: histone ratio) for 20 min at room temperature. Flow-through and retent were collected by centrifuging the filter units at 14,000 g at 20 °C for 10 min. Filters were then washed with 50 µl of 50 mM ABC and flow through was collected. The partially digested protein samples were later stage-tipped for label-free quantitative proteomics.
All samples were analyzed on a Q-Exactive Plus (Thermo Scientific) mass spectrometer that was coupled to an EASY nLC 1200 UPLC (Thermo Scientific). Peptides were loaded with solvent A (0.1% formic acid in water) onto an in-house packed analytical column (50 cm × 75 µm I.D., filled with 2.7 µm Poroshell EC120 C18, Agilent). Peptides were chromatographically separated at a constant flow rate of 250 nl min−1 using 150 min method: 5–30% solvent B (0.1% formic acid in 80% acetonitrile) within 119 min, 30–50% solvent B within 19 min, followed by washing and column equilibration. The mass spectrometer was operated in data-dependent acquisition mode. The MS1 survey scan was acquired from 300 to 1750 m/z at a resolution of 70,000. The top 10 most abundant peptides were subjected to higher collisional dissociation (HCD) fragmentation at a normalized collision energy of 27%. The AGC (automatic gain control) target was set to 5e5 charges. Product ions were detected in the Orbitrap at a resolution of 17,500.
All mass spectrometric raw data were processed with MaxQuant (version 188.8.131.52) using default parameters. Briefly, MS2 spectra were searched against the human Uniprot database with human histone sequences, including a list of common contaminants. False discovery rates (FDRs) on protein and peptide–spectrum match (PSM) level were estimated by the target-decoy approach to 0.01% (Protein FDR) and 0.01% (PSM FDR), respectively. The minimal peptide length was set to 7 amino acids and carbamidomethylation at cysteine residues was considered as a fixed modification. The maximum number of missed cleavages was set to 6. Variable modifications included oxidation (M), acetylation (protein N-term and K), methylation (KR) and diglycine (signature of ubiquitination) (K).
Re-immunoprecipitation of histone H3
Before immunoprecipitation, HEK293 cells were treated with 5 μM MG-132 for 16 h. Then, cells were lysed in modified RIPA buffer (50 mM Tris-HCl (pH 7.4), 150 mM NaCl, 0.25% sodium deoxycholate, 1% NP40, 1 mM PMSF, 1 mM EDTA, 1 mM NaF, 1 mM Na3VO) supplemented with protease inhibitor cocktail. Lysates were homogenized through a syringe needle (27 G) and centrifuged at 13,000 g for 15 min at 4 °C. The supernatant was collected and protein concentration was determined. Approximately 400 μg of protein were used for immunoprecipitation. After pre-clearing of the supernatant with Protein A agarose beads (Pierce), the samples were incubated overnight with 2 μg histone H3 antibody (Cell Signaling, #2650) at 4 °C. As a negative control, the same amount of protein was incubated with 2 μg anti-FLAG antibody (SIGMA, F7425) in parallel. Subsequently, samples were incubated with 30 µl of Protein A beads for 1 h at room temperature. Then, samples were centrifuged for 5 min at 5000 g and the pellet was washed three times with RIPA buffer. To elute histone H3, the beads were incubated with 50 µL 0.2 M glycine pH 2.6 (1:1) for 10 min with frequent agitation. The eluate was collected after centrifugation and immediately neutralized by adding an equal volume of Tris pH 8.0. Then, samples were diluted with RIPA buffer and used for re-immunoprecipitation with 2 μg anti-histone H3 or anti-FLAG control antibody. Subsequently, samples were incubated with 30 µl of Protein A beads for 1 h at room temperature. After this incubation, samples were centrifuged for 5 min at 5000 g and the pellet was washed three times with RIPA buffer. For the second elution, the beads were incubated with Laemmli Buffer, boiled for 5 min and centrifuged at 13,000 g for 5 min. The supernatant was then loaded onto a sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS–PAGE) gel for western blot analysis.
Chromatin Immunoprecipitation-sequencing (ChIP-seq)
ChIP-seq experiments were performed following the protocol described in ref. 12 . Cells were crosslinked with 1% formaldehyde for 10 min at room temperature. Crosslinked cells were quenched with 0.125 M glycine for 10 min at room temperature and scraped/transferred to a 15 ml conical tube on ice. Cells were then centrifuged for 5 min at 4 °C followed by two washing steps with 5 ml PBS 1×/1 mM PMSF 66 . H9 hESCs were resuspended sequentially in three different lysis buffers (lysis buffer 1: 50 mM Hepes, 140 mM NaCL, 10% glycerol, 1 mM EDTA, 0.25% TX-100, 0.5% NP-40, and protease inhibitor (Roche); lysis buffer 2: 10 mM Tris, 200 mM NaCL, 0.5 mM EGTA, 1 mM EDTA; lysis buffer 3: 10 mM Tris, 100 mM NaCL, 0.5 mM EGTA, 1 mM EDTA, 0.1% Na-Deoxycholate, 0.5% N-Lauroylsarcosine). Chromatin was sonicated for 20 cycles (30 s on, 45 s off) using Bioruptor (Diagenode). After sonication, the material was centrifuged at 16000 g for 3 min at 4 °C, with the supernatant representing the sonicated chromatin. 75 μl was not subject to immunoprecipitation and was used as total input control for the ChIP reactions. 750 μl were incubated with 10 μg of anti-H3K9me3 antibody (Abcam, #8898, reported suitable for ChIP) overnight at 4 °C. On day 2, magnetic Dynabeads G (Thermofisher) at 10x volume of H3K9me3 antibody were aliquoted into a new microtube. Magnetic beads were washed five times with cold RIPA wash buffer (50 mM Hepes, 1 mM EDTA, 500 mM LiCl, 1% NP-40, 0.7% Na-Deoxycholate). Then, beads were washed once with 1 ml TE + 50 mM NaCl on ice and samples were centrifuged at 950 g for 3 min at 4 °C. After removing all liquid from beads, elution buffer (50 mM Tris, 10 mM EDTA, 1% SDS) was added for 15 min at 65 °C. Finally, beads were centrifuged for 1 min at 1600 g and placed on magnet-holder to settle and the supernatant was transferred into a new tube. To reverse crosslinking, 3x volume of the elution buffer was added to the input and incubated together with the ChIP sample at 65 °C overnight. On day 3, 1x volume of TE buffer and 0.2 mg ml−1 RNase were added and incubated for 1 h at 37 °C. To digest proteins, 0.2 mg ml−1 Proteinase K was added and incubated for 2 h at 55 °C. Then, DNA was phenol-chloroform extracted at room temperature with 1x volume of 25:24:1 phenol-chloroform-isoamyl alcohol and centrifuged for 5 min to separate layers, followed by the addition of 1x volume chloroform. The DNA was transferred to a new tube for precipitation with 1/10 of NaOAc 3 M, 1 μl 20 mg ml−1 glycogen, 2x volume of ice-cold ethanol during 30 min at −80 °C. After centrifugation for 30 min at 4 °C, we removed the supernatant was removed and added 0.5 ml ice cold 70% ethanol followed by 5 min centrifugation. After removing the ethanol, the pellet was air-dried at room temperature and resuspended in 40 μl dH2O.
ChIP-seq libraries from ChIP and input DNA samples were prepared as previously described 67 . Libraries from H9 hESCs were sequenced with a 2 × 75 bp read length on Illumina HiSeq4000. For ChIP-seq, two biological replicates from independent experiments were analysed by QuickNGS (Next-Generation Sequencing) pipeline 55 . Quality check of sequencing data were performed with FastQC version 0.10.1 (Babraham Bioinformatics). ChIP-seq sequencing reads were mapped with Burrows-Wheeler Aligner (BWA) 68 to the Homo Sapiens genome (Ensembl database 87). For peak calling, the resulting Binary Alignment/Map (BAM) files were analyzed with MACS2 version 2.0.10 69 . The results comprise lists of statistically significant peaks compared with the respective input DNA controls. QuickNGS pipeline identifies all genes which are 10,000 bp up- or downstream from the MACS2 peaks. To identify differential read-enriched peak regions from ChIP-seq data between different conditions, we used bdgdiff module of MACS2. Data was uploaded into MySQL database. In addition, QuickNGS provides password-protected track hubs for the UCSC Genome Browser with hyperlinks for visualization.
hESCs and HEK293 cells were scraped from plates and lysed in protein cell lysis buffer (10 mM Tris-HCl, pH 7.4, 150 mM NaCl, 10 mM EDTA, 1% Triton X-100, 50 mM NaF, 0.1% SDS, 20 μg ml−1 Aprotinin, 2 mM sodium orthovanadate, 1 mM phenylmethylsulphonyl fluoride and protease inhibitor) by incubating samples for 10 min on ice and homogenization through 27-G syringe needle. Then, cell lysates were centrifuged at 10,000 g for 10 min at 4 °C and the supernatant was collected. Nematodes were lysed in protein lysis buffer (50 mM Tris-HCl, pH 7.8, 150 mM NaCl, 0.25% sodium deoxycholate, 1 mM EDTA and protease inhibitor) using a Precellys 24 homogenizer. Worm lysates were centrifuged at 10,000 rpm for 10 min at 4 °C and the supernatant was collected.
Protein concentrations were determined with BCA protein assay (Thermoscientific, Germany). Total protein was separated by SDS–PAGE, transferred to nitrocellulose membranes (Millipore, Germany) and immunoblotted. Western blot analysis was performed with anti-UBE2K (Cell Signaling, #8226, 1:1,000), anti-OCT4 (Stem Cell Technologies, #60093, 1:500), anti-SOX2 (Abcam, #97959, 1:1,000), anti-PAX6 (Stem Cell Technologies, #60094, 1:200), anti-Nestin (Stem Cell Technologies, #60091, 1:1,000), anti-MAP2 (Sigma, #1406, 1:1,000), anti-polyubiquitinylated conjugates (Enzo, PW8805-0500, 1:1,000), anti-ubiquitin (Merck Millipore, # 05-944, clone P4D1-A11, 1:1000), anti-H3K9me3 (Abcam, #8898, 1:1,000), anti-Histone H3 (Cell Signaling, #2650, 1:10,000), anti-H3K9me1 (Cell Signaling, #1418, 1:1,000), anti-H3K9me2 (Cell Signaling, #4658, 1:1,000), anti-H3K4me3 (Active Motif, #39916, 1:1,000), anti-H3K27me3 (Active Motif, #39155, 1:1,000), anti-H3K27ac (Active Motif, #39933, 1:1,000), anti-HTT (Cell Signaling, #5656, 1:1,000), anti-SETDB1 (Abcam, #107225, 1:500), anti-p53 (Cell Signaling, #9282, 1:2,000), anti-Histone H1 (Merck, 05-457, 1:1,000), anti-PSMD11 (Abcam, #99413, 1:1,000), anti-ß-actin (Abcam, #8226, 1:1,000) and α-tubulin (Sigma, T6199, 1:5,000).
In vitro ubiquitination assays
A concentration of 10 μg of purified human recombinant protein 6xHis::H3F3A (Prospec, PRO-1452) was mixed with 25 ng of E1 activating enzyme (Enzo Life Sciences, BML-UW9410-0050), 400 ng of UBE2K/E2-25K (R&D systems, E2-603), 2 μg of FLAG::ubiquitin (Sigma-Aldrich, U6253), ATP regeneration solution (Enzo Life Sciences, BML-EW9810-0100) and Ubiquitin Conjugation Reaction Buffer (Enzo Life Sciences, BML-KW9885-0001). When indicated in the corresponding figures, we also added 1 μg of hESC lysate or 0.8 μg of recombinant E3 enzymes RNF2 or RNF138. Samples were incubated at 30 °C for 1 h. The reaction was terminated by boiling for 5 min with SDS-sample buffer, and resolved by SDS-PAGE followed by immunoblotting using anti-6xHis antibody (QIAGEN, #34660, 1:1000) to monitor ubiquitination of histone H3F3A. RNF2 cloned into pRK5-FLAG plasmid was a gift from J. Zhang 70 . RNF138 cloned into the eGFP-C1 vector was a gift from M.J. Hendzel 71 .
26S proteasome fluorogenic peptidase assays
In vitro assay of proteasome activity was performed as previously described 72 . Cells were collected in proteasome activity assay buffer (50 mM Tris-HCl, pH7.5, 250 mM sucrose, 5 mM MgCl2, 2 mM ATP, 0.5 mM EDTA, and 1 mM dithiothreitol) and lysed by passing 10 times through a 27 G needle attached to a 1 ml syringe needle. Lysates were centrifuged at 10,000 g for 10 min at 4 °C. 25 μg of total protein were transferred to a 96-well microtiter plate (BD Falcon) and incubated with Suc-Leu-Leu-Val-Tyr-AMC (Enzo), a fluorogenic proteasome substrate to measure chymotrypsin-like activity. Fluorescence (380 nm excitation, 460 nm emission) was monitored on a microplate fluorometer (EnSpire, Perkin Elmer) every 5 min for 1 h at 37 °C.
Immunocytochemistry experiments were performed as we described in ref. 63 . Cells were fixed with paraformaldehyde (4% in PBS) for 30 min, followed by permeabilization (0.2 % Triton X-100 in PBS for 10 min) and blocking (3% BSA in 0.2% Triton X-100 in PBS for 10 min). Cells were incubated with primary antibody for 2 h at room temperature (Rabbit anti-H3K9me3 (Abcam, #8898, 1:500), Rabbit anti-PAX6 (Stem Cell Technologies, #60094, 1:300), Mouse anti-OCT4 (Stem Cell Technologies, #60093, 1:200), Mouse anti-Nestin (Stem Cell Technologies, #60091, 1:500), Rabbit anti-SOX1 (Stem Cell Technologies, #60095, 1:100), Mouse anti-MAP2 (Sigma, #1406, 1:200), Rabbit anti-H3K9me3 (Abcam, #8898, 1:500)). Cells were then washed with 0.2% Triton-X/PBS and incubated with secondary antibody Alexa Fluor 488 goat anti-mouse (ThermoFisher Scientific, A-11029, 1:500), Alexa Fluor 568 goat anti-rabbit (ThermoFisher Scientific, A-11011, 1:500), and 2 µg ml−1 Hoechst 33342 (Life Technologies, #1656104) for 1 h at room temperature. 0.2% Triton-X/PBS and distilled water wash were followed before we mounted the cover slips.
RNA isolation and quantitative RT–PCR
RNA extraction and quantitative RT–PCR were performed as described in ref. 63 . We extracted RNA using RNAbee (Tel-Test Inc.) and generated cDNA using qScript Flex cDNA synthesis kit (Quantabio). SybrGreen qPCR experiments were performed with a 1:20 dilution of cDNA using a CFC384 Real-Time System (Bio-Rad) following the manufacturer’s instructions. Data were analysed with the comparative 2ΔΔCt method using the geometric mean of ACTB and GAPDH as housekeeping genes. Supplementary Data 6 contains the sequences of all the primers used for this assay.
C. elegans strains and maintenance
C. elegans strains were grown and maintained on standard nematode growth media seeded with E. coli (OP50) 73 . Wild-type (N2), SS104 (glp-4(bn2)I), CF512 (fer-15(b26)II;fem-1(hc17)IV), MT13293 (met-2(n4256)III) were provided by the Caenorhabditis Genetics Center (CGC) (University of Minnesota), which is supported by the NIH Office of Research Infrastructure Programs (P40 OD010440).
For western blot experiments, synchronized animals were raised and fed OP50 E. coli at 25 °C until day 1 of adulthood. Then, worms were transferred onto plates with E. coli (HT115) containing either empty control vector (L4440) or expressing double-stranded RNAi. ubc-20 RNAi and ubc-22 RNAi constructs were obtained from the Vidal RNAi library and sequence verified. Wild-type and MT13293 worms were treated with 100 μg ml−1 5-fluoro-2′deoxyuridine (FUdR) to inhibit the proliferation of progeny.
Statistics and reproducibility
In each independent experiment using mammalian cells, biological replicates/wells were averaged for every condition. Then, the data of different independent experiments were averaged. Finally, we compared the average across conditions/groups. For C. elegans experiments, sample size determination was done according to standard C. elegans approaches. Sample sizes are indicated in the corresponding figure legends and supplementary information. Statistics were derived from at least n = 3. No data were excluded from the analyses. Statistical comparisons of western blot quantifications and qPCR data were performed with two-tailed Student’s t test for unpaired samples using GraphPad Prism 6.0. Statistically significant differences from proteomics data were determined with Perseus (v. 184.108.40.206) 62 after correction for multiple testing following the Benjamini–Hochberg procedure, which calculates false discovery rate (FDR) adjusted P values. Analysis of enriched Gene Ontology Biological Processes (GOBPs) were performed using Database for Annotation Visualization and Integrated Discovery (DAVID) 74 , that provides statistical comparisons with a modified Fisher’s exact test (EASE score). We used MACS2 to calculate fold changes and P values from ChIP-seq data, allowing for the identification of statistically significant differential enriched-peak regions between conditions 69 . For RNA-sequencing data, gene expression for the individual samples was calculated with Cufflinks2 as FPKMs 57 , 58 and fold-change and P values were calculated with DEseq2 59 .
Further information on research design is available in the Nature Research Reporting Summary linked to this article.
RNA-seq and ChIP-seq data have been deposited in Gene Expression Omnibus (GEO) under the accession code GSE146704. The mass spectrometry proteomics data showed in Supplementary Data 2 and Supplementary Fig. 12 have been deposited to the ProteomeXchange Consortium via the PRIDE 75 partner repository with the dataset identifiers PXD018625 and PXD018621, respectively. Uncropped images of western blots are presented in Supplementary Data 7 . All source data underlying the graphs presented in the main figures can be found as Supplementary Data 8 . All the other data are also available from the corresponding author upon request (DV).
Evans, M. J. & Kaufman, M. H. Establishment in culture of pluripotential cells from mouse embryos. Nature 292, 154–156 (1981).
The Deutsche Forschungsgemeinschaft (DFG) (VI742/1-2 and Germany’s Excellence Strategy-CECAD, EXC 2030-390661388), the Else Kröner-Fresenius-Stiftung (2015_A118) and the European Research Council (ERC Starting Grant-677427 StemProteostasis) supported this research. This work was also supported by ERC Consolidator Grant-616499 to T.H., the Foundation for Polish Science co-financed by the European Union under the European Regional Development Fund (POIR.04.04.00-00-5EAB/18-00) to W.P., and the Polish National Science Center (UMO-2016/23/B/NZ3/00753) to A.D. We thank L. Kurian for helpful discussions and comments on the manuscript. We thank the CECAD Proteomics Facility and the Cologne Center for Genomics (CCG) for their contribution and advice in proteomics and RNA sequencing experiments, respectively.
Cologne Excellence Cluster for Cellular Stress Responses in Aging-Associated Diseases (CECAD), University of Cologne, Joseph Stelzmann Strasse 26, 50931, Cologne, Germany
Peer Review File
Rights and permissions
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/ .
About this article
Cite this article
Fatima, A., Irmak, D., Noormohammadi, A. et al. The ubiquitin-conjugating enzyme UBE2K determines neurogenic potential through histone H3 in human embryonic stem cells. Commun Biol 3, 262 (2020). https://doi.org/10.1038/s42003-020-0984-3