INTRODUCTION

The pluripotency of embryonic stem cells appears to be under the regulation of multiple transcription factors1, 2, 3. Oct-4 is recognized as a critical factor in maintaining the self renewing ability of embryonic stem cells, perhaps through the regulation of a plethora of downstream genes both positively and negatively1, 2, 3. In addition, transcription factors STAT3 has also been implicated in regulating the stemness of stem cells2, 4. However, neither STAT3 nor Oct-4 is sufficient to maintain stem cell pluripotency5, 6. To this end, one novel homeoprotein named nanog has recently been identified as an intrinsic factor for ES cell self renewal6, 7, 8.

Like Oct-4, the expression of nanog mRNA is restricted to pluripotent stem cells and absent in differentiated ones7, 8. Nanog apparently operates in parallel with STAT3 and is sufficient to maintain stem cell pluripluripotency without gp130/STAT3 activation7, 8. It has been further suggested that nanog is the missing determinant for pluripotency of inner cell mass (ICM) and embryonic stem (ES) cells7. While Oct-4 functions to prevent the differentiation of ICM and ES cells into trophectoderm, nanog not only blocks the differentiation into endoderm independently but also actively maintains pluripotency7. Currently, the molecular mechanism through which nanog performs these two distinct function remains entirely unknown. Mitsui and colleagues suggest that nanog maintains self renewal of ES cells by transcriptional repression of those genes critical for differentiation7. Indeed, either gata4 or gata6 is elevated in nanog null ES cells7. Both gata4 or gata6 can trigger extraembryonic endoderm differentiation of ES cells transfected with either factor9. A direct link between nanog and gata6 became apparent when a nanog consensus sequence identified through SELEX selection was found in the enhancer region of gata67. Therefore, it is plausible that nanog maintains pluripotency by directly repressing downstream genes such as gata4 or gata67. Yet, little is known about the transactivating or repressing potential of nanog at the present.

Mouse nanog encodes a protein of 305 amino acid residues which can be roughly divided into three regions based on the position of the NK-2 type homeodomain7, 8. The 60 residue homeodomain is expectedto perform protein-protein interactions and DNA bindingas demonstrated for Oct-41. The region N-terminal to the homeodomain has 95 residues rich in Ser and Thr and acidic residues found in typical transactivators 7, 8, 10. On the other hand, the C-terminal region of 150 residues have no apparent transactivation motifs, but contains a prominent W-repeat7, 8. Therefore, it is plausible that the N-terminal domain of nanog maymediate trans-activation while the C-terminal domainmediates trans-criptional repression. While the repression function of a transcription factor is hard to establish, positive transactivation domains can be identified and characterized by the well-established Gal4 based reporter system10. To this end, we have performedfusion experiments between Gal4 DNA binding domainand individual ones from nanog. We report herethat both the N-and C-terminal domains have transactivation activities. To our surprise, despite thelack of any known features for transactivators, the C-terminal domain of nanog is a much stronger activator than its N-terminal counterpart.

MATERIALS AND METHODS

Effector constructs

Mouse Nanog cDNA was isolated by RT-PCR using high fidelity polymerase pfx (Invitrogen, CA) from P19 cells. The primerswere: accatgagtgtgggtcttcctggtcc for the 5′ end and tatttcacctggtggagtcacagagt for the 3′end. The mNanog cDNAs were inserted into the EcoRV site of modified pCR3.111. pCR3.1-Gal4DBD was created from a PCR fragment encoding the Gal4DNA binding domain(residues1 to 147). An EcoRV site was regenerated in the upstream or downstream of Gal4 DBD for inserting various domains of Nanog. A flag tag was fused N terminally or Cterminally to the Gal4 DBD for proteins detection. pCR3.1-NDGal4was created by inserting a PCR fragment encoding Nanog Ndomain(residue1-95) to the upstream EcoRV site of pCR3.1-Gal4DBD. pCR3.1-Gal4-CD and pCR3.1-Gal4-HD were generated by inserting a PCR fragment encoding nanog C domain(residues156-305) or the nanog homeodomain (residues 96-155) to the downstream EcoRV site of pCR3.1-Gal4 DBD respectively.

Cell lysates preparation and western blot analysis

HEK293T cell lysates were harvested 48h after transfection. Briefly, after washed with PBS, cells were lysed on ice with 250 μl of RIPA buffer(50 mM Tris-HCl, PH 7.5, 150 mM NaCl, 0.25% Sodium deoxycholate, 0.1% Nonidet P-40, 0.1% Triton X-100) for 10 min and cleared of debris by centrifugation at 15000 rpm for 15 min at 4°C. Cell lysates were electrophoresed through 10% SDSPAGE and electro-bloted to PVDF membranes (Millipore). The membranes were blocked in 5% non-fat milk and incubated with anti-flag antibody (1:2000, Sigma, MO). Alkaline phosphotaseconjugated anti-mouse antibody (1:2500, Pierce) was used as 2nd antibody and NBT/BCIP as substrates for detection as described 11.

Cell culture and reporter assays

HEK293T cells were cultured in DMEM(Invitrogen, CA) supplemented with 10% FBS(Hyclone, UT) and 100 mg/ml antibiotics (penicillin and streptomycin) as described11. For reporter assays, HEK293T cells were seeded in 12-well cell culture plate and transfected by calcium phosphate co-precipitation methods with 0.2ug of p5G-e1b-luciferase and 1ug of each expression vector. Transfection efficiencies were normalized by co-transfection with Reniila (Promega, WI) in each well and DNA concentrations were normalized by using pCR3.1 empty vector. Cells were harvested 36 h after transfection, and luciferase activity was measured by using Dual-luciferase Reporter Assay system (Promega, WI) according to the manufacture's instructions. Each transfection was carried out in duplicate and repeated at least twice.

RESULTS AND DISCUSSION

To test our hypothesis that nanog contains distinct transcription domains, we operationally divided its open reading frame into three subdomains: the 95 residue Nterminal domain (ND), the 60 residue homeodomain (HD) and the 150 residue C-terminal domain (CD) (Fig 1A). Since the cognate DNA binding sequence for nanog remains largely unknown, we employed the well-established Gal-4 fusion/reporter system to identify any transactivation domains in nanog as described previously 10. We fused the ND, HD and CD of nanog to gal-4 DNA binding domain as illustrated in Fig 1B. The resulting plasmids were confirmed by sequencing and the corresponding fusion proteins characterized by Western blotting with the anti-FLAG monoclonal antibody M2 (Fig 1C). The transactivation potentials of these fusions were then assessed using a reporter plasmid containing 5 copies of Gal-4 DNA binding sites upstream of the E1b TATA box. As shown in Fig 1D, both Gal4-ND and Gal4-CD plasmids transactivated the reporter efficiently, resulting over 35 or 240 folds of activation over the gal4 construct respectively in HEK293T cells. We also tested both plasmids in HeLa, NIH3T3, and P19 cells and observed transactivation activities for both plasmids (data not shown). The activation potential of ND is consistent with the fact that it is rich in acidic and Ser/Thr residues, hallmarks of transactivation domains 7, 8. On the other hand, we were surprised to detect robust activation function in CD given its lack of any discernable transactivation motifs. The fact that CD is almost seven times as active as ND suggests a dominant role for CD in mediating activation of downstream gene.

Figure 1
figure 1

Identification of two distinct transactivation domains in nanog. (A) Division of nanog into three distinct domains at the amino acid level. The homeodomain is marked in red. (B) Schematic illustrations of fusion constructs between Gal-4 DNA binding domain and the three subdomains of nanog. The FLAG tag is located at the C-terminal end for the ND-Gal4 fusion, but at the N-termini for the Gal4-CD and Gal4-HD. (C) Western blotting analysis of fusion constructs. HEK293T cells were transfected with control expression vector (lane 1), Gal4 DNA binding domain alone (lane 2), ND-Gal4 (lane 3), Gal4-CD (lane 4) and Gal4-HD (lane 5) as described in Materials and Methods. Cell lysates were fractionated, blotted and probed with anti- FLAG antibody. The * at the right of lane 3 depicts the expected position based on its calculated molecular weight. (D) Transactivation activities of fusion constructs. The reporter plasmid, p5G-E1b-luciferase (0.2 μg), were co-transfected with control vector (1 μg, lane 1), pCR3.1-Gal4DBD (1μg, lane 2), pCR3.1-ND-Gal4 (1 μg, lane 3), pCR3.1-Gal4-HD (1μg, lane 4) or pCR3.1-Gal4-CD (1ug, lane 5) to HEK293T cells in 12 well plates. The luciferase activity was assayed 36 h after transfection using the Dual-reporter assay system. The results were the average of at least two independent transfection experiments each with duplicates.

The data presented here establish for the first time that nanog possesses two independent transactivation domains. Oct-4, another homeodomain protein implicated in stem cell pluripotentcy1, 2, 12, 13, also contains two separate transactivation domains with no apparent similarity to the ones we defined in nanog. The divergent structures associated with these activation domains may confer the specificity of regulation for the target genes important for stem cell pluripotency. The ND of nanog is rich in Ser and Thr, suggesting that it may be regulated by phosphorylation. Indeed, western blotting analysis of ND-Gal4 indicates that its apparent molecular weight is 5 kDa larger than predicted, perhaps, as a result of phosphorylation at some of the Ser and Thr residues. More experiments are needed to map the sites of phosphorylation and identify the kinases responsible for the phosphorylations. On the other hand, the CD of nanog is entirely novel without any homology or similarity to other transactivation domains. Taken together, our data presented in this communication definitively prove that nanog contains two distinct transactivation domains with novel structural features. These findings may lead to the eventual elucidation of nanog's role in sustaining stem cell pluripotency at the molecular levels.