Proteomic analysis of Sox2-associated proteins during early stages of mouse embryonic stem cell differentiation identifies Sox21 as a novel regulator of stem cell fate

Abstract

Small increases in the levels of master regulators, such as Sox2, in embryonic stem cells (ESC) have been shown to promote their differentiation. However, the mechanism by which Sox2 controls the fate of ESC is poorly understood. In this study, we employed multidimensional protein identification technology and identified >60 nuclear proteins that associate with Sox2 early during ESC differentiation. Gene ontology analysis of Sox2-associated proteins indicates that they participate in a wide range of processes. Equally important, a significant number of the Sox2-associated proteins identified in this study have been shown previously to interact with Oct4, Nanog, Sall4, and Essrb. Moreover, we examined the impact of manipulating the expression of a Sox2-associated protein on the fate of ESC. Using ESC engineered for inducible expression of Sox21, we show that ectopic expression of Sox21 in ESC induces their differentiation into specific cell types, including those that express markers representative of neurectoderm and heart development. Collectively, these studies provide new insights into the range of molecular processes through which Sox2 is likely to influence the fate of ESC and provide further support for the conclusion that the expression of Sox proteins in ESC must be precisely regulated. Importantly, our studies also argue that Sox2, along with other pluripotency-associated transcription factors, is woven into highly interconnected regulatory networks that function at several levels to control the fate of ESC.

Figure 1

Preparation of samples for MudPIT analyses. (A) Flow chart describing the experimental outline for identification of Sox2-associated proteins from Dox-treated i-Sox2-ESC. Samples processed from untreated i-Sox2-ESC served as a negative control. (B) Western blot analysis to detect Flag-Sox2 in the immunoprecipitation eluates from Dox-treated i-Sox2-ESC. (C) Silver-staining of immunoprecipitation eluates from untreated and Dox-treated i-Sox2-ESC showing preferential pull-down of Sox2-associated proteins in Dox-treated i-Sox2-ESC.

Figure 2

Sox2-associated proteins identified by MudPIT analyses. (A) NSAF values of statistically significant (p < 0.05) Sox2-associated proteins identified only in Dox-treated i-Sox2-ESC, in all three MudPIT analyses (see also Supplemental Table 2A). (B) Fold enrichment (>9-fold) of statistically significant (p < 0.05) Sox2-associated proteins in Dox-treated i-Sox2-ESC over untreated i-Sox2-ESC from three MudPIT analyses (see also Supplemental Table 2B).

Figure 2

Sox2-associated proteins identified by MudPIT analyses. (A) NSAF values of statistically significant (p < 0.05) Sox2-associated proteins identified only in Dox-treated i-Sox2-ESC, in all three MudPIT analyses (see also Supplemental Table 2A). (B) Fold enrichment (>9-fold) of statistically significant (p < 0.05) Sox2-associated proteins in Dox-treated i-Sox2-ESC over untreated i-Sox2-ESC from three MudPIT analyses (see also Supplemental Table 2B).

Figure 3

Validation of MudPIT results by co-immunoprecipitation studies in i-Sox2-ESC and i-Sox21-ESC. (A) An antibody against Sox2 was used to co-immunoprecipitate Lin28, Sall4, HDAC1, and HDAC2 in i-Sox2-ESC treated with 4 μg/ml of Dox for 24 hours. (B) An antibody against Sox2 was used to co-immunoprecipitate Sox21 in i-Sox21-ESC treated with 500 ng/ml of Dox for 24 hours.

Figure 4

Integrated protein - protein and transcription factor - gene interactomes of Sox2-associated proteins. (A) Sox2-associated proteins presented in Figure 2 were used to generate the integrated interactome of Sox2-associated proteins and proteins that associate with Oct4, Nanog, Sall4, and Esrrb [28-31]. Nanog and Esrrb, which are detected with low frequency, are included to generate the interactome because of their previously established roles in ESC (Supplemental Table 3A). HDAC1 and HDAC2 which have enrichment values < 9-fold are also included because they have been shown previously to interact with Sox2 in ESC (Supplemental Table 3B). Ovals represent the bait proteins used to identify the interacting partners and smooth edged rectangles represent interacting proteins. Dashed lines represent previously reported interactions. (B) Transcription factor – gene interactome depicting binding of core ESC transcription factors Sox2, Oct4, and Nanog to genes encoding Sox2-associated proteins. We utilized previously published genome-wide binding data for Sox2, Oct4, and Nanog in mouse ESC to generate the interactome [2,8,9]. Genes are represented as smooth edged rectangles and proteins are represented as filled ovals. Arrows emanating from each protein refers to its binding to the indicated gene. (C) Venn diagram depicting the overlap of genes bound by Sox2, Oct4, and Nanog.

Figure 4

Integrated protein - protein and transcription factor - gene interactomes of Sox2-associated proteins. (A) Sox2-associated proteins presented in Figure 2 were used to generate the integrated interactome of Sox2-associated proteins and proteins that associate with Oct4, Nanog, Sall4, and Esrrb [28-31]. Nanog and Esrrb, which are detected with low frequency, are included to generate the interactome because of their previously established roles in ESC (Supplemental Table 3A). HDAC1 and HDAC2 which have enrichment values < 9-fold are also included because they have been shown previously to interact with Sox2 in ESC (Supplemental Table 3B). Ovals represent the bait proteins used to identify the interacting partners and smooth edged rectangles represent interacting proteins. Dashed lines represent previously reported interactions. (B) Transcription factor – gene interactome depicting binding of core ESC transcription factors Sox2, Oct4, and Nanog to genes encoding Sox2-associated proteins. We utilized previously published genome-wide binding data for Sox2, Oct4, and Nanog in mouse ESC to generate the interactome [2,8,9]. Genes are represented as smooth edged rectangles and proteins are represented as filled ovals. Arrows emanating from each protein refers to its binding to the indicated gene. (C) Venn diagram depicting the overlap of genes bound by Sox2, Oct4, and Nanog.

Figure 4

Integrated protein - protein and transcription factor - gene interactomes of Sox2-associated proteins. (A) Sox2-associated proteins presented in Figure 2 were used to generate the integrated interactome of Sox2-associated proteins and proteins that associate with Oct4, Nanog, Sall4, and Esrrb [28-31]. Nanog and Esrrb, which are detected with low frequency, are included to generate the interactome because of their previously established roles in ESC (Supplemental Table 3A). HDAC1 and HDAC2 which have enrichment values < 9-fold are also included because they have been shown previously to interact with Sox2 in ESC (Supplemental Table 3B). Ovals represent the bait proteins used to identify the interacting partners and smooth edged rectangles represent interacting proteins. Dashed lines represent previously reported interactions. (B) Transcription factor – gene interactome depicting binding of core ESC transcription factors Sox2, Oct4, and Nanog to genes encoding Sox2-associated proteins. We utilized previously published genome-wide binding data for Sox2, Oct4, and Nanog in mouse ESC to generate the interactome [2,8,9]. Genes are represented as smooth edged rectangles and proteins are represented as filled ovals. Arrows emanating from each protein refers to its binding to the indicated gene. (C) Venn diagram depicting the overlap of genes bound by Sox2, Oct4, and Nanog.

Figure 5

Dox-induced expression of Flag-Strep-tagged Sox21 triggers the differentiation of ESC. (A) The parental cell line, A2Lox.cre, was used to generate i-Sox21-ES cells. i-Sox21-ESC constitutively express the reverse-Tet transactivator (rtTA) from the Rosa26 promoter. In the presence of Dox, rtTA is able to bind the Tet-Responsive Element (TRE) and activate the expression of the integrated transgene, Flag-Strep-tagged Sox21. Positive integration in isolated clones was verified by genomic screening using published PCR primers LoxinF and LoxinR [72]. // represents plasmid backbone. (B) Western blot analysis showing the turn-on of FSSox21 expression using nuclear extracts from i-Sox21-ESC (clones 1 and 7) treated for 48 hours with increasing concentrations of Dox (ng/ml). An M2 anti-Flag primary antibody was used. (C) Untreated i-Sox21-ESC grown on pre-gelatinized plates in the presence of LIF (feeder-free) for 48 hours (top panels); i-Sox21-ESC treated with Dox for 48 hours at concentrations of 200 ng/ml (middle panels) and 500 ng/ml (bottom panels). (D) A2Lox.cre parental cells untreated and treated with 500 ng/ml Dox for 48 hours. Scale bar = 100μm.

Figure 5

Dox-induced expression of Flag-Strep-tagged Sox21 triggers the differentiation of ESC. (A) The parental cell line, A2Lox.cre, was used to generate i-Sox21-ES cells. i-Sox21-ESC constitutively express the reverse-Tet transactivator (rtTA) from the Rosa26 promoter. In the presence of Dox, rtTA is able to bind the Tet-Responsive Element (TRE) and activate the expression of the integrated transgene, Flag-Strep-tagged Sox21. Positive integration in isolated clones was verified by genomic screening using published PCR primers LoxinF and LoxinR [72]. // represents plasmid backbone. (B) Western blot analysis showing the turn-on of FSSox21 expression using nuclear extracts from i-Sox21-ESC (clones 1 and 7) treated for 48 hours with increasing concentrations of Dox (ng/ml). An M2 anti-Flag primary antibody was used. (C) Untreated i-Sox21-ESC grown on pre-gelatinized plates in the presence of LIF (feeder-free) for 48 hours (top panels); i-Sox21-ESC treated with Dox for 48 hours at concentrations of 200 ng/ml (middle panels) and 500 ng/ml (bottom panels). (D) A2Lox.cre parental cells untreated and treated with 500 ng/ml Dox for 48 hours. Scale bar = 100μm.

Figure 6

Ectopic expression of Sox21 in ESC disrupts their self-renewal and induces their differentiation. (A) The cloning efficiency of i-Sox21-ESC at a clonal density of 900 cells per 60-mm dish, in the absence or presence of Dox (200 ng/ml), was determined by counting the formation of colonies fixed after 96 hours. Colonies were categorized as either ESC, mixed, or differentiated based on morphology. A mixed colony was defined as a central compact colony of ES cells with a wide periphery of differentiated cells. (B) i-Sox21-ESC grown on pregelatinized plates in the presence of LIF (feeder-free) and treated with Dox at a concentration of 500 ng/ml for 48 hours. After the initial 48 hours, Dox was removed (top panel) or replenished (bottom panel) for an additional 48 hours. Numerous colonies were followed, but only a single representative colony for each condition is shown here, with the same two colonies photographed throughout. Scale bar = 100μm.

Figure 7

Quantitative RT-PCR (qRT-PCR) of cDNA prepared from the RNA of i-Sox21-ESC cultured in the absence or presence of Dox (500 ng/ml) for 48 hours. Results are shown as average differences in cycle threshold (Ct) values from Dox-treated i-Sox21-ESC compared to untreated i-Sox21-ESC and normalized to GAPDH expression levels. The Sox21-3'UTR primer set is specific to endogenous Sox21.