You are hereDecember 13, 2011 | Pluripotent Stem Cells
Two more “Friends” Added in the Pluripotency Network: DNA Repair and Splicing
By Stuart P. Atkinson
Although multiple studies over the last 5-10 years have begun to unravel the complex network which underlies the pluripotent state of embryonic stem cells (ESCs), bringing together transcription factors, small RNA species, chromatin and DNA methylation events, the specific molecular determinants of the pluripotent state are yet to be fully appreciated. However, two recent papers in Cell now reveal two previously unappreciated pluripotency-associated elements: the role of a DNA damage complex in transcriptional control and transcription factor splicing.
A DNA repair complex functions as an Oct4/Sox2 cofactor in embryonic stem cells.
In the first study, Robert Tjian and colleagues at the University of California, Berkeley, USA wanted to further understand the control of the pluripotent state by identifying factors that are required for the transcription of pluripotency-associated genes. This study utilised an in vitro reconstituted Nanog promoter assay, which reconstitutes Oct4- and Sox2-dependent transactivation to establish which additional factors are required for promoter activation (Fong et al).
The Nanog promoter construct contained four copies of the Nanog Oct-Sox-binding sites immediately upstream of the native Oct-Sox element found in the human Nanog promoter. Surprisingly, upon addition of Oct4 and Sox2 only a very weak activation of the Nanog promoter was detected suggesting that additional cofactors were required to potentiate a full activator-dependent response. Therefore, the authors developed a biochemical complementation assay to search for pluripotent stem cell-selective cofactors using fractionated extracts from NTERA-2 (NT2) pluripotent human embryonal carcinoma cells (ECCs), D3 mouse ESC (mESC) and HeLa cells. Of the fractions tested, the high salt phosphocellulose fraction (P1M) from ECCs and mESCs strongly potentiated transcription of the Nanog promoter in an Oct4-, Sox2- and TFIID-dependent manner. Additionally, they found that he CRSP/Mediator complex and transcriptional activators implicated in Nanog expression (Nanog, Sall4, Klf4 and Esrrb) did not replace or enhance Oct4/Sox2-dependent transcription of Nanog in vitro, and that Nanog only became potentiated when Sox2 and Oct4 were present suggesting that the cofactors present in the P1M fraction mediate the synergistic activation of Nanog by Oct4 and Sox2. Further purifications found that the specific activity lay in a protein complex of around 600 kDa with four major polypeptides and subsequent high-sensitivity mass spectrometry revealed all detectable constituents to be the Xeroderma pigmentosum group C (XPC)-RAD23B-Centrin 2 (CETN2) nucleotide excision repair (NER) complex. All three proteins were shown to be highly enriched in ESCs and ECC and in the NT2 P1M fraction compared to the HeLa P1M fraction.
Reconstitution of the heterotrimeric XPC-RAD23B-CETN2 complex from recombinant gene products allowed a detailed examination of these subunits roles in the activation of Nanog expression. A mutant DNA-binding-defective XPC did not affect Nanog expression in the reconstituted assay suggesting that its DNA binding (and repair) activity is dispensable and functionally separable from its transcriptional activity. XPCs interaction with TFIIH was also not required for Nanog activation, suggesting that XPC is most likely targeted to its binding sites via interactions with Oct4 and Sox2. Subsequent immunoprecipitation analysis in 293T cells confirmed that the XPC complex members co-immunoprecipitated with Oct4 and Sox2, providing a mechanism for potentiating Nanog transcription. However, in mESC extracts, a stable interaction between the NER complex subunits and Oct4/Sox2 could not be reproducibly detected. Subsequent, genome wide ChIP analysis using a RAD23B antibody, demonstrated that the XPC-complex binds specific sites across the genome and also co-occupies sites for Oct4 and Sox2 but not Nanog. Overall, this suggests that the XPC complex has a classical cofactor function rather than a DNA-damage repair function, as a role in DNA damage repair entails transient, non-specific DNA binding.
Loss of function studies in mESC were then utilised to further probe the role of this complex on gene expression and Nanog transcription. Combined knockdowns of the three complex members resulted in pronounced morphological abnormalities, decreased alkaline phosphatase (AP) activity, reduced proliferation rates and a 2- to 3-fold reduction in the mRNA level of Nanog, Fgf4, Zfp42, and Utf1. Upon long term XPC complex subunit depletion, the apoptosis of flattened, fibroblastic AP-negative cells surrounding collapsing mESC colonies was observed, altogether suggesting that loss of the XPC complex leads to the loss of the pluripotent state. However, loss of single components had only mild effects on Nanog expression with no overt defects in self-renewal, suggesting that there is some functional compensation from related genes. Next, this complex was shown to be involved in the reacquisition of pluripotency during somatic cell reprogramming. Downregulation of either XPC or RAD23B in Oct4-GFP mouse embryonic fibroblasts (MEFs) led to a dramatic reduction in the reprogramming efficiency, with a significant decrease in the number of AP-positive colonies, as well as a marked reduction in the percentage of partially and fully reprogrammed cells. Reprogramming efficiency using MEFs derived from XPC and RAD23B knockout (KO) mice was also highly compromised; overall suggesting that efficient reprogramming may require the XPC complex in conjunction with Oct4 and Sox2 to re-establish ESC specific gene expression programs.
Together, this study points towards an important role for the NER complex of XPC-RAD23B-CETN2 in pluripotent cells by acting as a cofactor to potentiate Sox2 and Oct4 mediated transcription, seemingly independent of its role in DNA damage repair. However, recent studies have shown that DNA damage pathways, such as Base Excision Repair (BER) (He et al, Hajkova et al) may play a role in the removal of DNA methylation after modification of a methylated cytosine by factors such as the TET family proteins (Ito et al). Therefore, studies of DNA methylation dynamics and also DNA damage repair dynamics at specific promoters targeted by the NER complex may yield some exciting results concerning the role of this complex in the pluripotent nature of ESCs.
An Alternative Splicing Switch Regulates Embryonic Stem Cell Pluripotency
The second study, from the laboratory of Benjamin J. Blencowe at the University of Toronto, Canada, investigated splicing events which occur during the differentiation of hESCs using microarray profiling (Gabut et al). The authors research was focused on a previously unidentified alternate splicing change in the FOXP1 gene, which took place during neural differentiation. FOXP1 (forkhead box P1) is a member of the large FOX family of transcription factors and is necessary for the proper development of the brain and lung in mammals. At day 10 of the neural differentiation protocol, analysis indicated that exon 18 of FOXP1 had increased inclusion compared to undifferentiated hESCs. However, upon PCR confirmation, the group detected two unexpected additional bands that were around 50 nucleotides and 170 nucleotides longer than expected for the transcript containing exon 18. This led to the discovery of an uncharacterised extra exon (exon 18b), which took the place of exon 18 in the 50 nucleotides transcript and the 170 nucleotides transcript contained both exons 18 and 18b. The dual exon transcript was rarely detected due to the introduction of a termination codon which likely elicits nonsense-mediated mRNA decay; however inclusion of exon 18b instead of exon 18 preserves the open reading frame but modifies the DNA binding forkhead domain. Analysis of exon 18b showed its efficient inclusion in undifferentiated hESCs and 2 day differentiated hESCs compared to the 10 day neurally-induced hESCs, consistent with a link between exon 18b and pluripotency. Indeed, exon 18b inclusion was highest in TRA1-81+ and SSEA3+ sorted hESCs suggesting that FOXP1 exon 18b (‘‘FOXP1-ES’’) is specific to self-renewing, pluripotent hESCs. It was then shown that the inclusion of exon 18b leads to the changes in DNA-binding specificities due to the modification of the forkhead domain. Through GST tagging of FOXP1 and FOXP1-ES and analysis with protein-binding microarrays, it was found that the canonical binding motif GTAAACAA was represented by the majority of the highest-scoring FOXP1- bound sequences whereas FOXP1-ES preferentially bound CGATACAA or closely related sequences. Additionally, the results revealed that FOXP1-ES binds a broader spectrum of sequences than FOXP1, although with apparent reduced affinity and collectively suggests that FOXP1-ES and FOXP1 direct different gene expression programs in ESCs.
To investigate whether FOXP1 and FOXP1-ES control different sets of genes, knockdowns were performed using specific siRNA pools targeting either exon 18 or exon 18b in undifferentiated H9 cells followed by RNA-Seq profiling. Knockdown of FOXP1 caused changes in expression of 153 genes, whereas FOXP1-ES knockdown caused changes in expression of 472 genes, 76 of which overlapped with the FOXP1-dependent gene set, with a higher proportion of genes upregulated upon loss of FOXP1-ES than FOXP1. Genes upregulated upon loss of FOXP1 were not correlated to any process but genes upregulated on FOXP1-ES loss were correlated to development, transmembrane receptor activity, and cell differentiation. Of the genes which decreased in expression upon FOXP1 or FOXP1-ES loss, many were related to early development and a subset of FOXP1-ES affected genes were involved in ESC pluripotency maintenance. Q-PCR confirmed that knockdown of FOXP1-ES resulted in at least a 2-fold decrease in the expression of the pluripotency genes (OCT4, NANOG, NR5A2, GDF3, and TDGF1) and a 2-fold or greater increase in expression of differentiation-associated genes (GAS1, HESX1, SFRP4, and WNT1), of which none were significantly changed upon FOXP1 loss. Overall, this suggests that FOXP1 and FOXP1-ES control distinct but overlapping sets of genes, with a substantially larger set of genes controlled by FOXP1-ES compared to FOXP1 in hESCs, FOXP1-ES predominantly acts to suppress gene expression and that the expression of FOXP1-ES in hESCs suppresses a large number of genes with important functions in cell differentiation and development, while promoting the expression of a specific subset of genes that support pluripotency.
To assess how FOXP1-ES and FOXP1 regulate target gene expression, ChIP was performed in hESCs using an antibody which immunoprecipitates both isoforms, followed by high-throughput sequencing, giving approximately 3,400 peaks across the genome. The resulting peaks were then assayed for the binding motifs associated with FOXP1 and FOXP1-ES. Motifs for FOXP1-ES and FOXP1 binding were highly enriched, but in contrast the CGATACA consensus and closely related sequences preferentially bound by FOXP1-ES in vitro did not appear to be widely utilized by this factor in vivo. Previous studies have revealed examples of transcription factors that preferentially bind lower-affinity sites in vivo (Jaeger et al and Rowan et al) and this property may be important to facilitate dynamic changes in transcriptional output mediated by FOXP1-ES and FOXP1 upon induction of ESC differentiation. Comparisons of the knockdown analysis and ChIP analysis provided a list of 116 candidate direct in vivo targets of these proteins, and the genes associated are associated with early development and cell differentiation, including OCT4 and NANOG. Further, a previously generated dataset on OCT4 binding sites (Kunarso et al) overlapped significantly with the FOXP1-ES and not with FOXP1, with the majority (26/33) of overlapping genes showing changes in the same direction upon knockdown of either factor collectively, suggesting that FOXP1-ES may regulate ESC self-renewal and pluripotency maintenance by directly controlling the expression of a subset of key pluripotency genes.
The FOXP1 gene is conserved in mouse, and exon 18 in humans relates directly to exon 16 in mouse. mFoxp1-ES was shown to be specifically expressed in mESCs and stimulates the expression of Oct4 and Nanog, suggesting that Foxp1-ES is required for mESC self-renewal and pluripotency. Forced expression of mFoxp1-ES abolished EB-mediated neural cell differentiation and cells maintained Oct4 expression. Knockdown of Foxp1 did not significantly impact proliferation, whereas knockdown of mFoxp1-ES reduced formation of mESC colonies by 3-fold and while mFoxp1-ES could compensate for mESC growth without LIF, Foxp1 could not. mFoxp1-ES-mESCs grown without LIF expressed Oct4, Nanog, and Nr5a2 at levels comparable to the parental mESCs, but displayed reduced levels of Sox2, Klf4, and Lifr and could also form teratomas which contained derivatives of all three germ cell types in vivo. Next, utilising a secondary mouse iPSC generation model which only requires the addition of Dox to mouse cells (in this case MEFs) to reactivate reprogramming factors Oct4, Sox2, Klf4 and C-Myc (oskm)), mFoxp1-ES was shown to be required for iPSC formation. Initial analysis of MEFs showed the presence of exon 16 only, which is progressively replaced with exon 16b during reprogramming. Knockdown of Foxp1-ES in MEFs resulted in a 5-fold reduction in reprogramming whereas knockdown of Foxp1 had little to no effect, while knockdown of Foxp1-ES at day 13 (of a 21 day protocol) also significantly reduced the proportion of reprogrammed cells. Additionally, while overexpression of Foxp1-ES with oskm did not substantially alter reprogramming, overexpression of Foxp1 completely blocked oskm mediated reprogramming.
Taken together, these data provide evidence that the alternative splicing-mediated Foxp1-ES expression is critical for efficient iPSC formation, as well as for the maintenance of ESC self-renewal and pluripotency. How this splicing is controlled and if this type of splicing event is widespread are surely the next questions to be addressed. Analysis such as this during other lineage specific differentiation protocols may also uncover other important genes that may also be regulated in such a way.
Fong YW, Inouye C, Yamaguchi T, Cattoglio C, Grubisic I, Tjian R.
A DNA repair complex functions as an oct4/sox2 cofactor in embryonic stem cells.
Cell. 2011 Sep 30;147(1):120-31.
Gabut M, Samavarchi-Tehrani P, Wang X, et al
An alternative splicing switch regulates embryonic stem cell pluripotency and reprogramming.
Cell. 2011 Sep 30;147(1):132-46.
Hajkova P, Jeffries SJ, Lee C, et al
Genome-wide reprogramming in the mouse germ line entails the base excision repair pathway.
Science. 2010 Jul 2;329(5987):78-82.
He YF, Li BZ, Li Z, et al
Tet-mediated formation of 5-carboxylcytosine and its excision by TDG in mammalian DNA.
Science. 2011 Sep 2;333(6047):1303-7.
Ito S, Shen L, Dai Q, et al
Tet proteins can convert 5-methylcytosine to 5-formylcytosine and 5-carboxylcytosine.
Science. 2011 Sep 2;333(6047):1300-3.
Jaeger SA, Chan ET, Berger MF, et al
Conservation and regulatory associations of a wide affinity range of mouse transcription factor binding sites.
Genomics. 2010 Apr;95(4):185-95.
Kunarso G, Chia NY, Jeyakani J, et al
Transposable elements have rewired the core regulatory network of human embryonic stem cells.
Nat Genet. 2010 Jul;42(7):631-4.
Rowan S, Siggers T, Lachke SA, et al
Precise temporal control of the eye regulatory gene Pax6 via enhancer-binding site affinity.
Genes Dev. 2010 May 15;24(10):980-5.
eNERgizing Pluripotent Gene Transcription
Splicing up Pluripotency