Dissecting Oct3/4-regulated gene networks in embryonic stem cells by expression profiling

Abstract

POU transcription factor Pou5f1 (Oct3/4) is required to maintain ES cells in an undifferentiated state. Here we show that global expression profiling of Oct3/4-manipulated ES cells delineates the downstream target genes of Oct3/4. Combined with data from genome-wide chromatin-immunoprecipitation (ChIP) assays, this analysis identifies not only primary downstream targets of Oct3/4, but also secondary or tertiary targets. Furthermore, the analysis also reveals that downstream target genes are regulated either positively or negatively by Oct3/4. Identification of a group of genes that show both activation and repression depending on Oct3/4 expression levels provides a possible mechanism for the requirement of appropriate Oct3/4 expression to maintain undifferentiated ES cells. As a proof-of-principle study, one of the downstream genes, Tcl1, has been analyzed in detail. We show that Oct3/4 binds to the promoter region of Tcl1 and activates its transcription. We also show that Tcl1 is involved in the regulation of proliferation, but not differentiation, in ES cells. These findings suggest that the global expression profiling of gene-manipulated ES cells can help to delineate the structure and dynamics of gene regulatory networks.

Figure 1. Global gene expression analysis of Oct3/4 manipulated ES cells.

(A) Experimental design of microarray analysis. Oct3/4-overexpressing ES cells differentiate to primitive endoderm and mesoderm, whereas Oct3/4-repressed ES cells differentiate towards Trophectoderm. (B)(C) K-means clusters of gene expression levels analyzed by TIGR MEV tool (http://www.tigr.org) and their landscape representations by MATLAB. (D) Comparison of genes grouped in each cluster of ZHTc6 cells and ZHBTc4 cells. The arrows (red and green) show highly correlated clusters between two cell lines (see Materials and Methods). (E) Diagram summarizing the unique mode of gene expression regulation by Oct3/4. Representative Genes identified as the direct downstream targets of Oct3/4 by the ChIP-on-chip assay or ChIP-PET assay are shown. Spp1 and Zfp42 (marked with asterisk) were not detected as a direct downstream target by either ChIP-on-chip or ChIP-PET assays, but have already been shown as such by more direct experimental methods (see the text).

Figure 2. Principal component analysis (PCA) and the expression pattern heatmap of known target genes for Oct3/4 and related genes.

(A) 2D-views of PCA for 2,757 genes that were identified as significantly differentially expressed during this time course. (B) The expression pattern and ES/TS specificity of each component of PCA was classified into 4 groups (Group I ∼ Group IV). (C) Known target genes for Oct3/4 and related genes. (D) Top 30 up- and down-regulated genes from 0 h v 24 h.

Figure 3. Delineation of Oct3/4-downstream target gene network.

(A) A list of genes that have been experimentally demonstrated as the direct downstream targets of Oct3/4. Yes: genes that were also detected by the global ChIP assays. No: genes that were not detected by the global ChIP assays. Mouse homologues of human genes identified in Boyer et al. were generated using NCBI HomoloGene Build 49. (B) Comparison of gene lists obtained by the expression profiling (list 1 plus list 2) and global ChIP assays.

Figure 4. (A) Primary targets of Oct3/4. Genes are separated into four groups: Genes detected by the current expression profiling, mouse ChIP-PET assay, and human ChIP-on-chip assay; previously known target genes that were also detected by the current expression profiling; Genes detected by the current expression profiling and mouse ChIP-PET assay; Genes detected by the current expression profiling and human ChIP-on-chip assay.

Genes classified as transcription factors by GO annotation were shown in bold. Genes that were down-regulated by the repression of Oct3/4 in ZHBTc4 cells were shown in blue. Genes that were up-regulated by the repression of Oct3/4 in ZHBTc4 cells were shown in red. (B) Representative GO categories that are enriched in the primary targets of Oct3/4 in a statistically significant manner (FDR>0.05).

Figure 5. Tcl1 is regulated by Oct3/4.

(A) Luciferase assay for 4 candidate Oct3/4 target genes and one known target gene (Zfp42). (B) Luciferase assay for deletion analysis of region upstream of Tcl1 gene. s1 and s2 are candidates for Oct3/4 binding sites, with sequence shown in (E). (C) EMSA for two candidate sites. The arrow indicates the band for Oct3/4/oligonucleotide binding. (D) ChIP assay of two target sites (see Experimental Procedures). (E) The sequence around s1 and s2 sites and the position of mutations for EMSA oligos (blue nucleotides).

Figure 6. Functional analysis of Tcl1 in ES cells.

(A) Expression level of Tcl1 gene effects on cell proliferation. The number of ES cells decreased when Tcl1 was knocked down by shRNA. (B) RT-PCR and Western blot analysis of ES cells with shRNA of Tcl1 gene. Tcl1 gene expression affected active Akt1 (p-Ser.473 Akt1), but not wild type of Akt1. (C) Reversibility of Tcl1 downregulation effect. The number of cells was rescued in the absence of Tet. (D) Photomicrographs of ES cell cultures. The number of ES cells decreased when Tcl1 was knocked down by shRNA. Cell proliferation was rescued in the absence of Tet. (E) Overexpression of Tcl1 does not rescue cell proliferation while Oct3/4 is repressed. (F) A model for the involvement of Oct3/4-Tcl1-Akt in ES cell proliferation.