Zfp322a Regulates mouse ES cell pluripotency and enhances reprogramming efficiency

Abstract

Embryonic stem (ES) cells derived from the inner cell mass (ICM) of blastocysts are characterised by their ability to self-renew and their potential to differentiate into many different cell types. Recent studies have shown that zinc finger proteins are crucial for maintaining pluripotent ES cells. Mouse zinc finger protein 322a (Zfp322a) is expressed in the ICM of early mouse embryos. However, little is known regarding the role of Zfp322a in the pluripotency maintenance of mouse ES cells. Here, we report that Zfp322a is required for mES cell identity since depletion of Zfp322a directs mES cells towards differentiation. Chromatin immunoprecipitation (ChIP) and dual-luciferase reporter assays revealed that Zfp322a binds to Pou5f1 and Nanog promoters and regulates their transcription. These data along with the results obtained from our ChIP-seq experiment showed that Zfp322a is an essential component of mES cell transcription regulatory network. Targets which are directly regulated by Zfp322a were identified by correlating the gene expression profile of Zfp322a RNAi-treated mES cells with the ChIP-seq results. These experiments revealed that Zfp322a inhibits mES cell differentiation by suppressing MAPK pathway. Additionally, Zfp322a is found to be a novel reprogramming factor that can replace Sox2 in the classical Yamanaka's factors (OSKM). It can be even used in combination with Yamanaka's factors and that addition leads to a higher reprogramming efficiency and to acceleration of the onset of the reprogramming process. Together, our results demonstrate that Zfp322a is a novel essential component of the transcription factor network which maintains the identity of mouse ES cells.

Figure 1. Zfp322a is required for ES pluripotency.

(A) Zfp322a mRNA level was decreased in mES cells cultured in LIF withdrawal ESC medium. The level of the Zfp322a and Pou5f1 mRNA were compared to control cells cultured in normal ESC medium and normalized against β-actin. (B) Zfp322A protein level showed a reduction as mES cells differentiated. β-actin served as loading control. (C) Zfp322a RNAi caused ES cell differentiation. AP staining was conducted on the fourth day of selection after the cells were transfected with Zfp322a shRNAs. Zfp322a RNAi cells displayed a lighter colour compared to the dark red colour of mock RNAi cells. (D) Pluripotency genes were down-regulated upon Zfp322a RNAi. ES cells transfected with empty pSUPER.puro vector were used as a control and gene expression levels were normalized against β-actin. (E) Zfp322a RNAi resulted in decreased Oct4, Nanog, Sox2 and Rex1 protein. β-actin served as loading control. (F) Quantification of the protein level changes. Protein levels of Zfp322a knocked-down cells were normalized against β-actin and compared to control RNAi cells using software ImageJ. (G) Representative flow cytometry results showed that fluorescence intensities of Oct4 and Nanog were repressed in Zfp322a depleted cells as compared to control cells. (H) Zfp322a depletion led to reductions of Oct4 and Nanog expression at cellular levels. The mean of fluorescence intensity was calculated using software flow software 2.5.0. Relative fluorescent intensities of Zfp322a RNAi cells were normalized against control knocked-down cells. Standard deviations were derived from three independent experiments. (I) Zfp322a RNAi caused up-regulation of lineage specific markers for endoderm, mesoderm and trophectoderm. Specific primers were used to check the respective gene expression levels by real-time PCR.

Figure 2. Changes of global gene expression upon Zfp322a knock-down in ES cells.

(A) Microarray heat map generated from relative gene expression levels. Zfp322a was knocked down in E14 cells and the cells were selected for 96 hours before whole genome cDNA microarray hybridization was performed. Duplicates were performed to ensure reproducibility of results. Relative highly expressed genes were shown in red and low expressed genes in green. 1574 genes showed an increased expression level of more than 1.5 fold after Zfp322a RNAi. Gene onthology analysis was performed relating to “biological process”. The enriched terms were classified into several function groups and listed in the figure. Many terms related to developmental processes were enriched. 904 genes expression level showed more than 1.5 fold reduction. Examples of down-regulated pluripotency-related genes upon Zfp322a knock-down in ES cells. Genes were selected according to their known functions in pluripotency or ES cells. Each selected gene was taken as individual tiles from the thumbnail-dendogram duplicates. (B) List of up-regulated MAPK pathway related genes upon Zfp322a RNAi. Genes were selected as they fell into the cluster “MAPK signaling pathway” according to gene ontology analysis for enriched KEGG pathways. Each selected gene was taken as individual tiles from the thumbnail-dendogram duplicates. (C) Phosphorylated ERK (p-ERK) level was elevated in Zfp322a depleted cells as compared to control cells, while the total ERK (t-ERK) level was not affected. β-actin served as a control for normalization.

Figure 3. Zfp322a positively regulates Oct4 and Nanog transcription.

(A) Zfp322a binds to Oct4 distal enhancer regions. Zfp322a ChIP DNA was analyzed by real-time PCR. Locations of primers used in qPCR were mapped to the Pou5f1 genomic region. (B) Zfp322a binds to Nanog proximal promoter, with a highest enrichment fold at TSS starting site. Locations of primers were pictured on mouse Nanog genomic region. Relative luciferase activities were down-regulated upon Zfp322a RNAi using Pou5f1 CR4-pSV40-Luc construct (C) and pNanog PP-Luc construct (E), but not Pou5f1 CR1-pSV40-Luc construct (D). Schematic structures of the constructs were presented. Empty pSUPER.puro vector were transfected in ES cells as a control RNAi. Renilla luciferase vector were transfected simultaneously and relative luciferase activities were normalized against Renilla luciferase activity.

Figure 4. Genomic-wide analysis of Zfp322a binding sites.

(A) Schematic definitions of locations of the putative Zfp322a binding sites relative to the nearest transcriptional unit. TSS referred to −1000 to +100 bp from 5′-end of annotated RNA. (B) Genomic distributions of Zfp322a binding loci. (C) Identification of genes that were predicted to be directly regulated by Zfp322a. The datasets from microarray analysis and ChIP-Seq targets were calculated for overlapping genes. The results revealed 1574 tentative genes that likely were activated and 223 tentative genes repressed directly by Zfp322a. (D) Predicted binding motifs for Zfp322a. Motifs were computationally determined based on the ChIP-Seq data. Three different motifs were identified, namely motif 1, motif 2, motif 3, with frequencies of 9%, 5%, 4% respectively. (E) Zfp322a can be integrated within ES cell transcription regulatory network. Shown was co-occurrence of transcription factors at the multiple binding loci. Colours in the heat map reflected the co-localization frequency of each pair of transcription factors (the darker the color was, the more frequently colocalized). All the transcription factors were clustered according to the colocalization frequency with other factors, which was calculated based on their co-occurrence at the same binding loci. (F) Zfp322a can interact with Oct4. Cell lysate of wild type ES cells were immunoprecipitated using either anti-ZNF322A antibody or anti-Oct4 antibody. Western blot was subsequently carried out with anti-Oct4 antibody or anti-ZNF322a antibody. Control IP was performed using anti-IgG antibody.

Figure 5. Zfp322a can enhance OSKM reprogramming and replace Sox2.

(A) Zfp322a enhanced reprogramming efficiency and accelerated the onset of reprogramming process. OSKM serves as control experiment. (B) The iPSCs generated from OSKM plus Zfp322a presented alkaline phosphatase activity. There were more AP stained colonies generated from OKSM+Zfp322a compare to OKSM. (C) The iPSCs expressed endogenous Oct4, Nanog, Sox2, Rex1 and SSEA-1, indicating that they were ES-cell like. Immunostaining using anti-Oct4, anti-Nanog anti-Sox2, anti-Rex1 and anti-SSEA-1 antibodies were performed with GFP⁺ iPSCs generated from OKSM+Zfp322a. (D) GFP⁺ iPSCs generated by OKSM+Zfp322a were able to express ectoderm, mesoderm and endoderm lineage markers in the EB formation assay. iPSCs were stained with anti-Nestin, anti-Gata4 and anti-alpha smooth muscle actin (SMA) antibodies and pictures were taken at 60× magnification. DAPI (blue) served as nucleus marker. (E) Zfp322a was able to replace Sox2, but not Oct4 or Klf4 in OSKM reprogramming process. Results from three independent experiments were presented. (F) iPSCs generated from OKM plus Zfp322a were positive with AP staining and more AP positive colonies were observed in OKM+Zfp322a as compared to OKSM. (G) iPSCs generated by OKM plus Zfp322a expressed pluripotency markers Oct4, Nanog, Sox2, Rex1 and SSEA-1. (H) iPSCs derived from OKM+Zfp322a could differentiate into ectoderm, mesoderm and endoderm lineages, which were showed by anti-Nestin, anti-Gata4, anti-SMA staining respectively.