SEARCH

SEARCH BY CITATION

Keywords:

  • Cell biology;
  • Induced pluripotent stem cells;
  • Artd1/Parp1;
  • Pluripotent stem cells;
  • Reprogramming;
  • Fgf4

Abstract

  1. Top of page
  2. A
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusions
  8. Acknowledgments
  9. Disclosure of Potential Conflicts of Interest
  10. References
  11. Supporting Information

The recently established reprogramming of somatic cells into induced pluripotent stem cells (iPSCs) by Takahashi and Yamanaka represents a valuable tool for future therapeutic applications. To date, the mechanisms underlying this process are still largely unknown. In particular, the mechanisms how the Yamanaka factors (Oct4, Sox2, Klf4, and c-Myc) directly drive reprogramming and which additional components are involved are still not yet understood. In this study, we aimed at analyzing the role of ADP-ribosyltransferase diphtheria toxin-like one (Artd1; formerly called poly(ADP-ribose) polymerase 1 [Parp1]) during reprogramming. We found that poly(ADP-ribosylation) (PARylation) of the reprogramming factor Sox2 by Artd1 plays an important role during the first days upon transduction with the reprogramming factors. A process that happens before Artd1 in conjunction with 10–11 translocation-2 (Tet2) mediates the histone modifications necessary for the establishment of an activated chromatin state at pluripotency loci (e.g., Nanog and Essrb) [Nature 2012;488:652–655]. Wild-type (WT) fibroblasts treated with an Artd1 inhibitor as well as fibroblasts deficient for Artd1 (Artd1−/−) show strongly decreased reprogramming capacity. Our data indicate that Artd1-mediated PARylation of Sox2 favors its binding to the fibroblast growth factor 4 (Fgf4) enhancer, thereby activating Fgf4 expression. The importance of Fgf4 during the first 4 days upon initiation of reprogramming was also highlighted by the observation that exogenous addition of Fgf4 was sufficient to restore the reprogramming capacity of Artd1−/− fibroblast to WT levels. In conclusion, our data clearly show that the interaction between Artd1 and Sox2 is crucial for the first steps of the reprogramming process and that early expression of Fgf4 (day 2 to day 4) is an essential component for the successful generation of iPSCs. Stem Cells 2013;31:2364–2373


Introduction

  1. Top of page
  2. A
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusions
  8. Acknowledgments
  9. Disclosure of Potential Conflicts of Interest
  10. References
  11. Supporting Information

The recent discovery of Takahashi and Yamanaka that it is possible to establish pluripotent stem cells, so called induced pluripotent stem cells (iPSCs), by reprogramming differentiated somatic cells [1] has opened new perspectives in the field of regenerative medicine. Therefore, many groups have worked on the refinement of reprogramming in order to optimize this technology for the use of iPSCs in clinical applications (reviewed in [2]). Nevertheless, the molecular mechanisms underlying the process of reprogramming are still largely unknown. An increased knowledge on how this process is driven and on the underlying mechanisms would lead to improved, more efficient reprogramming techniques.

Originally, reprogramming of mouse and human fibroblasts to iPSCs was performed by the retroviral-mediated introduction of the four transcription factors, Oct 4, Sox2, Klf4, and c-Myc (the so called Yamanaka factors [1]). How these factors directly drive the process of reprogramming and which additional components are involved still needs to be carefully analyzed. One of the Yamanaka factors, the transcription factor Sox2 (sex determining region Y-box 2), is a main player in maintaining pluripotency in embryonic stem cells (ESCs) [3]. Therefore, the regulation of Sox2 is most likely critical for the generation of iPSCs. One enzyme that has been demonstrated to post-translationally regulate Sox2 is the ADP-ribosyltransferase diphtheria toxin-like 1 (Artd1, formerly called Poly(ADP-ribose) polymerase 1/Parp1). Artd1 is a chromatin associated factor that catalyzes the covalent attachment of poly(ADP-ribose) (PAR) to itself and to other nuclear acceptor proteins [4, 5]. ADP-ribosylation plays an important role in numerous biological processes, such as maintenance of genomic stability, cell differentiation, cell death, replication, and transcriptional regulation [6, 7]. Different roles of Artd1 in the regulation and maintenance of pluripotency have also been previously described: deletion of Artd1 in ESCs globally affects gene expression patterns and Artd1 knockout ESCs differentiate into trophoblast derivatives [8, 9]. Furthermore, it was previously shown that Artd1-dependent PARylation of Sox2 induces its eviction from the Fgf4 (fibroblast growth factor 4) enhancer and thereby induces Fgf4 transcriptional activation [10]. In contrast, recently Lai et al. reported that in ESCs, Artd1 PARylates itself and thereby enhances its interaction with Sox2, which in turn prevents Sox2 from binding to Oct4/Sox2 enhancers [11]. Previous work indicates an important role of Artd1 during reprogramming. Artd1 knockout (Artd1−/−) fibroblasts exhibit impaired reprogramming capacity [11], but the mechanisms underlying this observation were not analyzed. In a recent work, Doege et al. describe a role of Artd1 in conjunction with 10–11 translocation-2 (Tet2) in mediating the histone modifications necessary for the establishment of an activated chromatin state at pluripotency loci [12].

In this study, we aimed at analyzing the role of Artd1 during reprogramming, paying special attention to its role in the first days upon transduction (days 0–4) of the cells with the Yamanaka factors. We found that PARylation of Sox2 by Artd1 between day 0 and day 4 plays an important role in the generation of iPSCs. Inhibition of the enzymatic activity of Artd1 during this time period in wild-type (WT) fibroblasts resulted in a strongly decreased reprogramming efficiency after retroviral-mediated transduction of the Yamanaka factors. The same could be observed when using fibroblasts deficient for Artd1 (Artd1−/−). Our data further show that Artd1-mediated PARylation of Sox2 is involved in the regulation of Fgf4 expression. The importance of Fgf4 during the first steps of reprogramming is also corroborated by our finding that addition of exogenous Fgf4 can rescue the reprogramming deficiency of the Artd1−/− cells.

In conclusion, our data clearly indicate a new role of Artd1 in regulating Fgf4 activity via Sox2 ADP-ribosylation during reprogramming and suggest a dual function of Artd1 during this process. Artd1 is essential for starting the Fgf4-mediated reprogramming process and later establishes the post-translation modification necessary for the activation of pluripotency genes [12].

Materials and Methods

  1. Top of page
  2. A
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusions
  8. Acknowledgments
  9. Disclosure of Potential Conflicts of Interest
  10. References
  11. Supporting Information

Reprogramming

Mouse embryonic fibroblasts (MEF) were isolated from 14.5-day-pregnant C57BL/6 mice and cultured in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% fetal bovine serum (PAA) and 1% l-glutamin/penicillin/streptomycin (10,000 U/ml penicillin G sodium; 10,000 µg/ml streptomycin sulfate; 29.2 mg/ml l-glutamine; 10 mM sodium citrate in 0.14% NaCl, Gibco, Invitrogen, Basel, Switzerland, www.invitrogen.com). The reprogramming of the MEFs was performed according to Yamanaka's protocol [13] using the pMXs retroviral vectors producing murine Oct4, Sox2, Klf4, and c-Myc (Addgene, cat. nos. 13366, 13367, 13370, and 13375). Two days after infection, MEFs were cultured in DMEM containing 15% fetal bovine serum, 1% l-glutamin/penicillin/streptomycin, 1× MEM nonessential amino acids (Gibco, Invitrogen, Basel, Switzerland, www.invitrogen.com), and 50 mM ß-mercaptoethanol (Gibco, Invitrogen, Basel, Switzerland, www.invitrogen.com) supplemented with 1,000 U/ml ESGRO murine Leukemia inhibitory factor (Millipore, Chemikon, Zug, Switzerland, www.millipore.com). FGF4 (Sigma, Buchs, Switzerland, www.sigmaaldrich.com/switzerland-schweiz.html) was added during the reprogramming process at 10 ng/ml unless stated otherwise, ABT-888 (Enzo Life Sciences, New York, www.enzolifesciences.com) at 10 µM and SU5402 (Millipore, Calbiochem, Zug, Switzerland, www.millipore.com) at 2 µM.

Immunofluorescence Staining iPSCs

For immunofluorescence staining, iPSCs derived from WT, Artd1−/− fibroblasts, and Artd1−/− fibroblasts reprogrammed in the presence of Fgf4 (Artd1−/−*) were grown on mitomycin C-treated MEFs and fixed in 4% paraformaldehyde. Then, iPSCs were incubated with primary antibodies against Oct4 (rabbit anti-Oct4, Santa Cruz Biotechnology, Santa Cruz, CA, www.scbt.com) and SSEA-1 (mouse anti-SSEA-1, Millipore). Secondary fluorescence-labeled antibodies were used for detection (goat anti-rabbit Alexa Fluor 594 and goat anti-mouse Alexa Fluor 488, Molecular Probes, Invitrogen, Basel, Switzerland, www.invitrogen.com). Nuclei of the cells were counterstained with 4′,6-diamidino-2-phenylindole (DAPI) (Roche, Basel, Switzerland, www.roche.ch).

Real-Time PCR

Total RNA was isolated using the RNeasy Mini Kit (QIAGEN, Venlo, Netherlands, www.qiagen.com) and 1 µg of total RNA was reverse transcribed with Oligo-dT primers (Invitrogen) and Superscript III (Invitrogen). Real-time PCR was performed in triplicates in a Rotor-Gene Q RG-6000 (QIAGEN) with Rotor-Gene SYBR green (QIAGEN) and analyzed with the Delta Ct-method. GAPDH was used for normalization. Error bars represent the SD of the mean of triplicate reactions. Primers are listed in Supporting Information Table S1.

In Vitro Differentiation

For monoculture neural and smooth muscle differentiation, iPSCs (WT, Artd1−/− and Artd1−/− cells reprogrammed in the presence of Fgf4 [Artd1−/−*]) were plated onto gelatinized 35 mm dishes. The iPSCs were cultivated for 10 days with neural differentiation medium (DMEM/F-12 [Gibco, Invitrogen, Basel, Switzerland, www.invitrogen.com], N2 [1:100, Gibco, Invitrogen, Basel, Switzerland, www.invitrogen.com], B27 [1:50, Gibco, Invitrogen, Basel, Switzerland, www.invitrogen.com], and 1% l-glutamin/penicillin/streptomycin) or smooth muscle differentiation medium (DMEM and 10% fetal bovine serum). At day 10, cells were fixed in 4% paraformaldehyde and stained for βIII-tubulin (Sigma) and smooth muscle actin (Sigma), respectively.

Western Blotting

Cells were collected in radioimmunoprecipitation RIPA buffer (50 mM Tris-HCl pH 8; 400 mM NaCl; 0.5% Nonidet P40; 1% Na-Deoxycholate; 0.1% SDS; Protease Inhibitor Cocktail Tablet, EDTA-free Roche, IN). Proteins were identified by SDS-PAGE (10% acrylamide) and Western blotting using the following antibodies: α-PAR ALX-210–890 (Enzo Life Sciences, New York, www.enzolifesciences.com), α-PARP-1/2 (H250) sc-7150, α-Sox2 15830 (Abcam, Cambridge, United Kingdom, www.abcam.com), α-Pcna (PC10) sc-56, α-tubulin T6199 (Sigma), IRDye 800CW anti-Rabbit, and IRDye 680RD anti-Mouse (LI-COR, Lincoln, Nebraska, www.licor.com). Images were acquired with an Odyssey Imaging System (LI-COR).

In Vitro Sox2 ADP-Ribosylation

HEK293 cells were seeded at a density of 2.7 × 106/150 mm dish and after overnight incubation transfected with pBluescript II and pCAG-HA-Sox2-IP (cat. no. 13459) vectors, respectively (using CaCl2 transfection). After 72 hours, cells were harvested and resuspended in NE buffer (50 mM Tris-HCl [pH 7.5], 0.15 M KCl, 5 mM MgCl2, 0.2 mM EDTA, 20% [vol/vol] Glycerol). After sonication at 4°C for 2 × 30-second, the cells were incubated with DNase (Fermentas, Thermo Fisher Scientific Waltham, Massachusetts, http://www.thermoscientificbio.com) for 30 minutes at 4°C. After DNA digestion, the cells were sonicated for 30 seconds and then centrifuged at 6,000 rpm for 10 minutes. The cleared lysate was subjected to immunoprecipitation overnight at 4°C using immobilized antibody against HA (ANTI-HA affinity gel, Sigma). Precipitates were washed three times with NE buffer and after centrifugation; the HA-Sox2 coupled beads were resuspended in reaction buffer (50 mM Tris-HCl pH8, 4 mM MgCl2, 0.25 mM dithiothreitol, 5 mM NaCl, 200 nM EcoRI linker). Recombinant human ARTD1 (10 pmol) and 32P-NAD+ (4 nmol) were added and the reactions were incubated for 15 minutes at 30°C. Empty pBlueScript II vector was used as a negative control and 0.1 µg of histone H1 (10223549001; Roche) as a positive control. Proteins were resolved by SDS-PAGE (10% acrylamide), exposed to x-ray film (Tx-RP) and analyzed (Typhoon imager, GE Healthcare Life Sciences, Switzerland, http://www.gelifesciences.com).

High Stringency Immunoprecipitation

At day 4, cells were collected in cold phosphate buffered saline and lysed in hypotonic buffer (5 mM Hepes pH 7.5; 85 mM KCl; 0.5% Nonidet P40; protease inhibitor Roche), and nuclei were spun down for 10 minutes at 8,000 rpm. Nuclei were then suspended in High Stringency Buffer (50 mM Tris-HCl pH 7.5; 0.4 M NaCl; 1% Nonidet P40; 0.4% Na-Dedoxycholate) followed by sonication and DNA digestion with DNaseI (Roche). Extracts were cleared by 10 minutes centrifugation at 14,000 rpm. Cleared nuclear extracts were diluted 1:2.7 in 50 mM Tris-HCl pH 7.5 and immunoprecipitations were carried out for 2 hours using 10H antibody or IgG as a negative control. Beads were then washed three times in the same buffer and lastly boiled in SDS loading buffer.

Immunofluorescence Staining During Reprogramming

WT and Artd1−/− fibroblasts were seeded on glass coverslips and reprogramming was induced as previously described. At the indicated time, cells were fixed in acetic acid/methanol (1:3) for 5 minutes on ice, blocked for 30 minutes in PBSMT (phosphate buffered saline containing 5% milk and 0.05% Tween-20), incubated with α-Sox2 (15830, Abcam, 1:200) and α-PAR (10H, Enzo Life Sciences, New York, www.enzolifesciences.com) (1:250) dissolved in PBSMT for 1 hour at room temperature, washed with phosphate buffered saline, incubated with secondary Alexa Fluor 488 α-rabbit (Invitrogen) and Cy3-IgG fraction monoclonal mouse anti-FITC antibodies, and embedded on microscopy slides with DAPI containing mounting medium VECTASHIELD. Images were acquired with a Leica SP5 microscope at the Centre for Microscopy and Image Analysis of the University of Zürich.

Sox2 Coimmunoprecipitation

WT and Artd1−/− fibroblasts were reprogrammed as previously described. At the indicated time, cells were harvested and lysed in hypotonic buffer (5 mM Hepes, 85 mM KCl, 0.5% Nonidet P40, Protease Inhibitor [Roche]). Nuclei were pelleted at 8,000 rpm for 10 minutes at 4°C and resuspended in immunoprecipitation buffer (50 mM Tris-HCl pH 7.5, 150 mM KCl, 5 mM MgCl2, 0.2 mM EDTA, 20% glycerol, protease inhibitor [Roche], 0.5 mM dithiothreitol), sonicated, and DNA was digested with DNaseI (Roche). Immunoprecipitations were carried out using 200 µg of nuclear proteins and 2 µg of α-Sox2 (15830, Abcam) for 2 hours at 4°C, followed by three washes in washing buffer (20 mM Tris-HCl pH 7.5, 100 mM KCl, 5 mM MgCl2, 0.2 mM EDTA, 10% glycerol, 0.1% Tween-20, protease inhibitor [Roche]), and eventually resuspended in 1× SDS-loading buffer.

Chromatin Immunoprecipitation

WT, Artd1−/−, and ABT-888 inhibited cells were crosslinked with 1% formaldehyde (Calbiochem). Chromatin was fragmented with the Bioruptor (Diagenode, Liège, Belgium, http://www.diagenode.com), incubated with specific antibodies, and collected with Protein A Agarose/salmon sperm DNA (Millipore). DNA was extracted and measured by real-time PCR using SYBR Green and Rotor-Gene 3000 (Corbett Life Science/QIAGEN). For Primer sequences see Supporting Information Table S1.

Results

  1. Top of page
  2. A
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusions
  8. Acknowledgments
  9. Disclosure of Potential Conflicts of Interest
  10. References
  11. Supporting Information

Artd1 Is Necessary for Successful Initiation of Reprogramming

MEFs were isolated from day 14 embryos obtained from homozygous breeding of WT and Artd1−/− mice, respectively. WT and Artd1−/− fibroblast were transduced with the Yamanaka factors as previously described [13] and the number of iPSC-colonies was assessed after 14 days of cultivation. The number of iPSC colonies obtained upon reprogramming of Artd1−/− fibroblasts was reduced by around 65% compared to WT fibroblasts (Fig. 1A). These observations are in agreement with recently published data [11] and indicate that Artd1 is required for the reprogramming of somatic cells to iPSCs. An even stronger reduction in the number of iPSC colonies (80%) was observed when the PARP-Inhibitor ABT-888, which mainly inhibits Artd1 and Artd2 [14], was applied. The small difference in the number of colonies between cells lacking Artd1 and PARP-inhibitor treated cells might indicate that the contribution of other ARTD family members than Artd1 during reprogramming is minimal. In addition, the ABT-888 inhibitor did not change the transcriptional levels of Artd1 and Artd2, indicating that the effects observed are exclusively due to the inhibition of the enzymatic activity (Supporting Information Fig. S1A, S1B). In order to define if ADP-ribosylation is necessary during the whole reprogramming process or only during a specific time window, we added the ABT-888 inhibitor starting from days 0, 2, 4, 6, and 8 after the transduction of WT fibroblasts with the Yamanaka factors. iPSC colony formation was strongly reduced when the inhibitor was added during the first 4 days of reprogramming, but unaffected if cells from later time points after viral infection were treated (Fig. 1B), indicating that Artd1 enzymatic activity is essential during the early phase of the reprogramming process. In order to determine the expression changes of Artd1 and Artd2 during the first 12 days of reprogramming, we performed quantitative real-time PCR analysis (Fig. 1C, 1D). The expression of both Artd1 and Artd2 constantly increased from day 2 to day 8 and dropped to levels similar to the untreated control cells by day 12. This increase in Artd1 expression was also observed at the protein level (Fig. 1E) and was in accordance with increased PARylation in fibroblasts at day 2, followed by a decrease at day 4 and a constant level in the following days (Fig. 1F). In summary, our data indicate an upregulation of Artd1 and Artd2 expression and of PARylation during the first days of reprogramming. Blocking of Artd1, either by genetic ablation or by applying a PARP-inhibitor, drastically decreases the efficiency of reprogramming.

image

Figure 1. Artd1 enzymatic activity is necessary for the initial steps of reprogramming. (A): Reprogramming efficiency in WT fibroblasts, Artd1−/− fibroblasts, and WT fibroblasts treated with ABT-888. (B): Artd1 enzymatic activity is essential during the early phase of the reprogramming process. ABT-888 was added at the indicated time points and the reprogramming efficiency was assessed. (C): Expression of Artd1 in WT cells during the reprogramming process. (D): Expression of Artd2 in WT cells during the reprogramming process. (E): Expression levels of Artd1 during reprogramming. Western blot for Artd1 on WT cells collected at the indicated time points during reprogramming. The smear of the Artd1 signal reflects Artd1 activity. Tubulin has been used as a loading control. Molecular size references in kilo Daltons are indicated. (F): PAR formation during reprogramming. Western blot for PAR on WT cells collected at the indicated time during reprogramming. Tubulin has been used as a loading control. Molecular size references in kilo Daltons are indicated. Dunnett's multiple comparison test. ns = not significant, **, p < .01; ***, p < .001; ****, p < .0001. Abbreviations: PAR, poly(ADP-ribose); WT, wild type.

Download figure to PowerPoint

Artd1 Is Responsible for Poly(ADP-Ribosylation) of Sox2 During Reprogramming

In ESCs, Artd1 was previously described to PARylate Sox2, thereby decreasing the association of Sox2 with the Fgf4 enhancer and inducing Fgf4 expression [10]. In order to test the capacity of Artd1 to PARylate Sox2, we transfected a HA-Sox2 expression vector in HEK293 cells and the recombinant protein was purified by immunoprecipitation. Upon in vitro incubation of the purified protein with recombinant human ARTD1 and radiolabeled NAD+, a signal at the predicted size of Sox2 was clearly detected, indicating that Sox2 is substrate of ARTD1 (Fig. 2A). To further prove that Sox2 is PARylated during reprogramming, WT, WT+ABT-888, and Artd1−/− day 4 cells were collected in high stringency ionic buffer in order to reduce cellular protein complexes. Cleared nuclear lysates were subsequently used to immunoprecipitate PARylated proteins with an anti-PAR antibody. PARylated proteins were resolved in SDS-PAGE, followed by Western blotting against Sox2 and Artd1. Sox2 as well as Artd1 were pulled down in WT extract but not in WT+ABT-888 or Artd1−/− extracts, indicating that Sox2 is a targeted by ADP-ribosylation in vivo. (Fig. 2B) In order to assess the expression pattern of Sox2 and the PARylation levels during the first phase of reprogramming, we performed Western blotting of Sox2 (Supporting Information Fig. S2) and monitored the localization of Sox2 and PAR by immunofluorescence in WT and Artd1−/− fibroblast at days 0, 2, 4, and 6 after viral infection with the Yamanaka factors (Fig. 2C and Supporting Information Fig. S3). In the nuclei of WT fibroblasts PARylation was detectable starting from day 2 post-transduction and was still present at day 6. Interestingly, Sox2 staining mainly colocalized with the PAR signal, suggesting that ADP-ribosylation of Sox2 occurs at the beginning of reprogramming. As expected, a PAR signal was not detectable in the Artd1−/− fibroblasts at any time point, thus supporting the idea that Artd1 is the major ADP-ribosyltransferase involved in the process.

image

Figure 2. Artd1 binds and post-translationally modifies Sox2 during reprogramming. (A): Trans ADP-ribosylation of HA-Sox2 by recARTD1. Recombinant human ARTD1 (10 pmol) was incubated with HA-Sox2 (lane 1), HA-empty (lane 2), or H1 as positive control (lane 3). Coomassie blue stained gel (left), autoradiography (middle), and Western blot for Sox2 bound to the beads (right) are shown. (B): Sox2 ADP-ribosylation during reprogramming. High stringent immunoprecipitation was carried on with either PAR antibody 10H or IgG control on nuclear extracts of WT, ABT-888 treated WT, and Artd1−/− cells collected at day 4. Western blots of Sox2 and Artd1 in immunoprecipitation samples and inputs are shown. (C): Immunofluorescence of Sox2 and PAR during reprogramming at day 4. WT and Artd1−/− cells were stained with DAPI, for Sox2 and for PAR. Signals in the relative channels are shown from left to right. Merge of the three channels is also reported. Scale bar = 10 µm. (D): Sox2/Artd1 interaction during reprogramming. Immunoprecipitation was carried on with either Sox2 antibody or IgG control on nuclear extracts of WT and Artd1−/− cells collected at the indicated time. Western blots of Sox2 and Artd1 in immunoprecipitation samples and inputs are shown. proliferating cell nuclear antigen (PCNA) was used as indicator of equal starting protein content for the immunoprecipitations. Abbreviations: DAPI, 4′,6-diamidino-2-phenylindole; ESCs, embryonic stem cells; Fibros, fibroblasts; WT, wild type.

Download figure to PowerPoint

In order to clarify if in vivo, during the first phases of reprogramming, Artd1 and Sox2 directly interact, WT and Artd1−/− fibroblasts were transduced with the Yamanaka factors and protein extracts were isolated at day 4 and day 8 upon transduction. Artd1/Sox2 coimmunoprecipitation was detectable at day 4 and day 8 in WT fibroblasts, clearly indicating that these two proteins interact in reprogramming fibroblasts. As expected, in Artd1−/− fibroblasts, no interaction was detectable (Fig. 2D). Taken together, our data show that within the first 8 days of reprogramming, Artd1 interacts with Sox2 and mediates its PARylation.

Artd1-Mediated ADP-Ribosylation of Sox2 Is Responsible for the Activation of Fgf4 Transcription

Although Sox2 has been reported to bind to the Fgf4 enhancer element, its effect on Fgf4 transcription remains controversial. Originally, it was described that Sox2 binding enhances Fgf4 transcription [15]. In contrast, a more recent study reported that in ESCs and in differentiating cells, Sox2 represses Fgf4 transcription and that ADP-ribosylation of Sox2 relives Fgf4 repression [10]. In addition, the work of Lai et al. showed that in ESCs, Sox2 positively regulates Fgf4 transcription and that auto-modified Artd1 interacts with Sox2, consequently inducing its release from the Fgf4 enhancer and repressing Fgf4 transcription [11]. To investigate the role of Sox2, Artd1, and ADP-ribosylation in the fine-tuning of Fgf4 transcription, we first analyzed the expression of Fgf4 upon initiation of reprogramming. In WT fibroblasts, expression of Fgf4 was detectable starting from day 2 and increased steadily until day 6. In contrast, in Artd1−/− as well as in WT fibroblasts treated with the ABT-888 inhibitor, expression was strongly reduced and delayed. This clearly indicates that the presence of Artd1 is necessary for the correct activation of Fgf4 transcription (Fig. 3A). In order to assess the binding capacity of Sox2 to the Fgf4 enhancer, we performed ChIPs using antibodies against Sox2 during reprogramming in WT, Artd1−/−, and in WT fibroblasts treated with ABT-888. As depicted in Figure 3B, during the reprogramming process, Sox2 is recruited to the Fgf4 enhancer and to other target sites such as the Nanog promoter. In ABT-888 inhibited fibroblasts, Sox2 recruitment is delayed, which is in agreement with the lower transcription of Fgf4 (Fig. 3A), indicating that PARylation positively influences the DNA binding capacity of Sox2. Strikingly, the delay of Sox2 recruitment in the presence of ABT-888 is phenocopied in Artd1−/− cells (Fig. 3B), confirming that Artd1 is mainly responsible for Sox2 ADP-ribosylation in reprogramming cells. Taken together, our data demonstrate that ADP-ribosylation of Sox2 strengthens the binding of Sox2 to its target sites and thereby stimulates the transcription of the corresponding target genes.

image

Figure 3. Artd1 activity is necessary for the binding of Sox2 to the Fgf4 enhancer and for driving Fgf4 expression. (A): Expression of Fgf4 during the reprogramming process in WT fibroblasts ± ABT-888 and Artd1−/− fibroblasts. (B): Recruitment of Sox2 to target genes: chromatin immunoprecipitation was carried on with either Sox2 antibody or IgG control in WT, Artd1−/−, and ABT-888 inhibited cells at the indicated time points. Recruitment of Sox2 to the Fgf4 enhancer, the Nanog promoter, and the unrelated Rps6 promoter is depicted. Values are expressed as enrichment over input signals. Abbreviation: WT, wild type.

Download figure to PowerPoint

Fgf4 Expression Is Crucial for the Initiation of Reprogramming

The reduced transcription of Fgf4 in Artd1−/− fibroblasts upon the initiation of reprogramming leads to the question if the impaired reprogramming efficiency of Artd1−/− cells is the direct consequence of the reduced Fgf4 levels or if other factors are involved. We therefore repeated the reprogramming experiments in WT and Artd1−/− fibroblast by adding 10 ng/ml and 25 ng/ml of exogenous Fgf4 to the cells. Addition of 10 ng/ml of Fgf4 was sufficient to restore the reprogramming efficiency to comparable levels as in WT cells (Fig. 4A), indicating that the phenotype observed in Artd1−/− cells is due to an insufficient expression of Fgf4 during the early stages of reprogramming. Interestingly, higher amounts of Fgf4 (25 ng/ml) or addition of exogenous Fgf4 to WT fibroblast rather impaired the reprogramming efficiency, suggesting that a tight control of autocrine Fgf4 production is essential for the initiation of reprogramming. The Artd1−/− iPSC colonies obtained upon addition of Fgf4 could be expanded for more than 10 passages and expressed the classic pluripotency genes. Furthermore, they were able to differentiate in vitro toward smooth muscles and neurons and showed no differences from WT and Artd1−/− iPSCs (Supporting Information Fig. S4).

image

Figure 4. Exogenous Fgf4 supplementation is sufficient to restore the reprogramming capacity in Artd1−/− fibroblasts and WT cells treated with ABT-888 inhibitor. (A): Reprogramming efficiency of Artd1−/− cells cultivated with exogenously added Fgf4. (B): Reprogramming efficiency of WT+ABT-888 cells in the presence or absence of Fgf4. WT fibroblasts were treated for the first 2, 4, or 6 days with ABT-888 alone or ABT-888 and 10 ng/ml Fgf4. (C): Reprogramming efficiency in WT cells and WT cells treated with SU5402, an inhibitor of Fgf receptor tyrosine kinase activity. (D): Schematic representation of Artd1-mediated PARylation of Sox2 and binding to the Fgf4 enhancer, which activates Fgf4 expression. Dunnett's multiple comparison test (A) and Student's t test (B, C). ns = not significant, *, p > .05; **, p < .01; ****, p < .0001. Abbreviation: WT, wild type.

Download figure to PowerPoint

The fact that Fgf4 transcription depends on ADP-ribosylation of Sox2 and the capacity of exogenous Fgf4 to restore the reprogramming efficiency in Artd1−/− fibroblasts suggests that the enzymatic activity of Artd1 is essential for the initiation of reprogramming. To test this, we reprogrammed WT fibroblasts and cultivated the cells for the first 2, 4, or 6 days with ABT-888 or with ABT-888 and 10 ng/ml Fgf4. The addition of Fgf4 for the first 2 or 4 days abolished the inhibitor effect of ABT-888 and significantly increased the reprogramming efficiency (Fig. 4B), indicating that the direct addition of Fgf4 compensates for the absence of ADP-ribosylation activity. This strengthens the observation that Artd1-mediated ADP-ribosylation of Sox2 is essential to modulate Fgf4 transcription during the initial phases of reprogramming. To further prove the importance of Fgf4, we performed Fgf4 knockdown experiments. The combination of viral transduction of the reprogramming factors with knockdown induced massive cell death in the fibroblast. This was not due to Fgf4 knockdown itself because scrambled controls had the same effect. We therefore decided to inhibit Fgf receptor tyrosine kinases with small chemical inhibitor SU5402 [16] in WT cells. The inhibition of Fgf receptor tyrosine kinases reduced the reprogramming efficiency of WT fibroblasts by around 50% (Fig. 4C). In summary, our results clearly identify Artd1-mediated ADP-ribosylation of Sox2 as an essential component for the correct activation of Fgf4 expression, which in turn plays a crucial role for the initiation of the reprogramming process.

Discussion

  1. Top of page
  2. A
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusions
  8. Acknowledgments
  9. Disclosure of Potential Conflicts of Interest
  10. References
  11. Supporting Information

Cell differentiation is normally an irreversible process and differentiated cells are not able to switch from one lineage to another. Cellular reprogramming to pluripotency therefore requires that pluripotency genes, which are inactive in differentiated cells, are reactivated. Exogenously introduced reprogramming factors must therefore bind and reactivate their target genes in tight collaboration with other endogenous factors, particularly epigenetic regulators.

In this work, we aimed at understanding the function of the well-known epigenetic regulator Artd1 during the first phase of reprogramming. Artd1 was previously shown to be necessary for proper differentiation of ESCs [8]. Even though Artd1 deficiency does not affect the growth of ESCs, its absence compromises cell survival and growth when ESCs are induced to differentiate [8, 10]. The molecular mechanisms underlying this observation are still largely unclear, but new studies indicate that Artd1 acts as a cofactor of Oct4 and Sox2 in ESCs by binding to the Fgf4 enhancer and thereby regulates Fgf4 expression [10, 11]. The major function of Fgf4 in pluripotent cells is to regulate the selection between the alternative fates of self-replication and lineage commitment during continuous proliferation. Autocrine production of Fgf4 is the major stimulus activating the Erk1/2 signaling cascade in naïve mouse ESCs [17, 18]. Inhibiting ERK and FGF activity with small chemical compounds prevents ESCs from differentiating without affecting the propagation of the undifferentiated ESCs [18, 19]. Similarly, ESCs lacking Fgf4 are resistant to neural and mesodermal induction, but are able to commit when FGF is provided exogenously [18].

During differentiation of ESCs, Artd1 was shown to directly interact with and to PARylate Sox2, leading to the dissociation and degradation of Sox2 from the Fgf4 enhancer. This releases Sox2 inhibition and induces Fgf4 gene transcription [10]. In the absence of activated Artd1, Sox2 cannot be ADP-ribosylated, augmenting its interaction with the Fgf4 enhancer and leading to a stabilization of Sox2 protein and a reduction in Fgf4 levels [10]. An alternative model suggests that Artd1 auto-PARylation enhances Sox2-Artd1 interactions and inhibits binding of Sox2 to the Oct4/Sox2 site at the Fgf4 enhancer. This process seems to be regulated by FGF/ERK signaling [11].

We first tested the reprogramming capacity of Artd1 knockout cells and found that the absence of Artd1 strongly reduces the number of iPSC colonies. We could also identify that the critical time period during which Artd1 activity is necessary are the first 2–4 days after transduction with the reprogramming factors. This is in agreement with the increased expression of Artd1 starting at day 2 after reprogramming and the concomitant increase of PARylation. The reduction of the reprogramming efficiency is mainly due to the lack of Artd1 activity and not of other members of the ARTD family, because the treatment of WT fibroblasts with ABT888, an inhibitor of poly(ADP-ribosyltransferases), mimics the genetic ablation of Artd1. We further observed a strong delay in Fgf4 expression upon the initiation of reprogramming in Artd1−/− fibroblasts or when WT fibroblasts are treated with ABT-888. Fgf4 expression occurs much earlier than the activation of transcription of other typical pluripotency markers such as Nanog, SSEA-1, or OCT-4. This observation is interesting because Fgf4 is typically expressed in pluripotent cells [18, 20, 21] and not in fibroblasts.

Based on previous studies [10, 11] indicating a role of Artd1 in modulating Sox2 activity in the context of Fgf4 regulation, we decided to analyze the capacity of Artd1 to modify Sox2 in vitro and in vivo. Our data demonstrate that human ARTD1 is able to PARylate Sox2 in vitro and strongly suggest that murine Artd1 mediates ADP-ribosylation of Sox2 in vivo in fibroblasts starting from day 2 during reprogramming. The role of Sox2 in the regulation of Fgf4 transcription is also highlighted by the observation that ADP-ribosylation of Sox2 increases its binding to the Fgf4 enhancer and leads to increased transcription. In summary, our data confirm that during the early phase of the reprogramming process, Artd1-mediated ADP-ribosylation of Sox2 is necessary for the binding of Sox2 to the Fgf4 enhancer and for inducing Fgf4 expression, which in turn is responsible for initiating the further events leading to the formation of iPSCs.

The importance of Fgf4 during the first phase of the reprogramming process is strengthened by the fact that Artd1−/− fibroblasts, which show a strongly reduced activation of Fgf4 upon reprogramming initiation, show a massive reduction in the number of iPSC colonies. The simple addition of Fgf4 during this time is sufficient to restore the reprogramming efficiency to comparable levels as in WT cells, indicating that Fgf4 is functionally the only factor regulated by Artd1 during the early phase of reprogramming (Fig. 4A). A similar effect can be observed when WT fibroblasts are cultivated in the presence of the ABT-888 inhibitor. Also in this case the simple addition of Fgf4 to the medium is sufficient to restore the reprogramming efficiency (Fig. 4B). The importance of Fgf4 during the early phase of reprogramming is also corroborated by the fact that WT cells treated with an inhibitor of Fgf receptor tyrosine kinases, which are normally activated upon the binding of Fgf, reduces the reprogramming efficiency by 50% (Fig. 4C).

Of interest, Artd1 in conjunction with 10–11 translocation-2 (Tet2) was recently shown to play an important role in the early stages of somatic cell reprogramming by mediating the histone modifications necessary for the establishment of an activated chromatin state at pluripotency loci (e.g., Nanog and Essrb) [12]. Furthermore, Artd1 induction promotes accessibility to the Oct4 reprogramming factor. Interestingly, pluripotency factors are detectable starting around days 10–12 of the reprogramming process [22-25], which would suggest that Artd1 might have two different functions. In the first phase (first week) of the reprogramming process, Artd1 is required for initiating the transcription of Fgf4. In a second step, Artd1 might be involved in promoting the accessibility of the reprogramming factors to the pluripotency gene promoters as shown by Doege et al. [12]. Since we did not observe an effect of PARP inhibitors on the reprogramming efficiency at this time point, it is fair to assume that this process is independent of ADP-ribosylation. Furthermore, our data demonstrate that even in the absence of Artd1, exogenously supplied Fgf4 permits reprogramming efficiencies as for WT fibroblasts, indicating that in the ARDT1−/− cells additional epigenetic modifiers must be interacting with Tet2 to mediate the histone modifications necessary for the activation of pluripotency genes.

Conclusions

  1. Top of page
  2. A
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusions
  8. Acknowledgments
  9. Disclosure of Potential Conflicts of Interest
  10. References
  11. Supporting Information

Our data indicate that PARylation of Sox2 by Artd1 plays an important role in the generation of iPSCs. Artd1-mediated PARylation of Sox2 favors its binding to the Fgf4 enhancer, thereby activating Fgf4 expression (Fig. 4D). Exogenous addition of Fgf4 during the first 4 days upon initiation of reprogramming was sufficient to restore the reprogramming capacity of Artd1 knockout fibroblast to WT levels, indicating that Fgf4 is an essential component for the correct initiation of the reprogramming process.

Acknowledgments

  1. Top of page
  2. A
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusions
  8. Acknowledgments
  9. Disclosure of Potential Conflicts of Interest
  10. References
  11. Supporting Information

We would like to thank Karolin Léger for her contribution in acquiring the confocal images. This work was sponsored by grants of the Swiss National Science Foundation (Grant 31003A-118361 to P.C.; Grant 323530-133905 to F.A.W.; Grant 310030B_138667 to M.O.H.).

References

  1. Top of page
  2. A
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusions
  8. Acknowledgments
  9. Disclosure of Potential Conflicts of Interest
  10. References
  11. Supporting Information

Supporting Information

  1. Top of page
  2. A
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusions
  8. Acknowledgments
  9. Disclosure of Potential Conflicts of Interest
  10. References
  11. Supporting Information

Additional Supporting Information may be found in the online version of this article.

FilenameFormatSizeDescription
stem1507-sup-0001-suppfig1.tif1153KSupporting Information Figure 1
stem1507-sup-0002-suppfig2.tif769KSupporting Information Figure 2
stem1507-sup-0003-suppfig3.tif4006KSupporting Information Figure 3
stem1507-sup-0004-suppfig4.tif5043KSupporting Information Figure 4
stem1507-sup-0005-supptab1.docx69KSupporting Information Table 1

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.