Author contributions: X.W.: conception and design, collection and/or assembly of data, data analysis and interpretation, manuscript writing; Y.Z.: provision of study material or patients, data analysis and interpretation; Z.X.: collection and/or assembly of data, data analysis and interpretation; B.C.: provision of study material or patients, data analysis and interpretation; Z.W.: collection and/or assembly of data; B.W.: data analysis and interpretation; J.Z.: data analysis and interpretation; J.H.: data analysis and interpretation; Y.G.: data analysis and interpretation; L.L.: provision of study material or patients; H.Z.: provision of study material or patients; W.Z.: data analysis and interpretation; H.L.: data analysis and interpretation; J.D.: conception and design, financial support, data analysis and interpretation, final approval of manuscript.
First published online in STEM CELLSExpress March 12, 2009
OCT4 is a pivotal transcription factor in maintaining the pluripotency and self-renewal capacities of embryonic stem (ES) cells. Human OCT4 can generate two isoforms by alternative splicing, termed OCT4A and OCT4B. OCT4A confers the stemness properties of ES cells, whereas the function of OCT4B is unknown. We present here the diverse protein products and a novel function of OCT4 gene. A single OCT4B mRNA can encode three isoforms by alternative translation initiation at AUG and CUG start codons, respectively. A putative internal ribosome entry site (IRES) has been identified in OCT4B mRNA accounting for the translation mechanism. The OCT4B-190 is upregulated under stress conditions and it may protect cell against apoptosis under stress. This work evokes the significance to distinguish the biological function of the protein products of OCT4. The OCT4 gene, by the regulation of alternative splicing and alternative translation initiation, may carry out more crucial roles in many biological events. STEM CELLS 2009;27:1265–1275
The transcription factor OCT4 (official symbol POU5F1, also known as OCT3, OCT3/4, OTF3, and OTF4) functions as a main regulator in maintaining the pluripotency and self-renewal capacities of embryonic stem (ES) cells [1, 2]. In human, OCT4 is expressed in totipotent and pluripotent stem cells throughout all stages from the unfertilized oocyte to the blastocyst . OCT4 is also highly expressed in human ES cells [4, 5], human embryonic germ (EG) cells , and human embryonic carcinoma (EC) cells . OCT4 regulates the expression of genes that maintain stem cell undifferentiated . The expression of OCT4 is downregulated during differentiation. Knockdown of its expression by RNA interference results in stem cell differentiation .
There are two isoforms of OCT4 generated by alternative splicing in human, termed OCT4A and OCT4B . During human preimplantation development, OCT4A and OCT4B display different temporal and spatial expression patterns. OCT4A is highly concentrated in the nucleus of the compacted embryos and blastocysts, whereas OCT4B expresses in the cytoplasm of all cells from the four-cell stage onward by indirect immunocytochemistry staining [11, 12]. Takeda et al. have indicated that both OCT4A and OCT4B mRNA express at low levels in all adult human tissues, such as heart, kidney, liver, placenta, spleen and islets, by RT-PCR . However, it is thought that some of the population may have no OCT4B expression because of a single nucleotide polymorphism (SNP) site in the OCT4B start codon (ATG → AGG) .
An earlier report  has shown that the OCT4 isoforms differ in DNA binding, transactivation, and abilities to confer self-renewal. And only OCT4A is responsible for the stemness properties, whereas OCT4B cannot sustain ES cell self-renewal. Atlasi et al.  have recently detected a novel OCT4 spliced variant, named OCT4B1. The expression of OCT4B1 is primarily in human ES and EC cells and is downregulated in differentiated cells. Indeed, the existence of variant isoforms of OCT4 raises further interests regarding the functions of OCT4 gene. However, the biological function of OCT4B remains unknown.
Alternative splicing, together with alternative translation initiation in a single mRNA contribute to the diversity of gene products [15–18]. It has been known that some eukaryotic cellular mRNAs can be translated via internal initiation by specific mRNA regions termed internal ribosome entry site (IRES) [17, 19, 20]. Remarkably, many IRES-containing mRNAs encode proteins that have important roles in development, differentiation, cell cycle progression, cell growth, cell apoptosis, and stress response [20–23]. The presence of IRES elements allows crucial survival factors to be transiently translated under stress conditions that require immediate changes in protein levels [24–27].
In this study, we have examined the OCT4A and OCT4B expression pattern in several cell lines. We have identified the alternative translation initiation at two AUG start codons and one CUG start codon on OCT4B mRNA by site-directed mutagenesis. A single OCT4B mRNA then can encode at least three protein isoforms: OCT4B-265, OCT4B-190, and OCT4B-164. We have demonstrated that the alternative translation is controlled by a putative IRES existing in OCT4B mRNA using the bicistronic expression vector. In addition, we have found that, under stress, the initiation of translation by this putative IRES element is maintained whereas the cap-dependent protein translation is greatly reduced. As a result, the endogenous expression of OCT4B-190 is upregulated under heat shock and oxidative stress treated cells. Over-expression of OCT4B-190 in HeLa cells has resulted in increasing the resistance to apoptosis induced by heat shock. Furthermore, OCT4B-190 relocalizes to the stress granules (SGs) in cells under stress, suggesting a significant role in the stress response.
MATERIALS AND METHODS
Total RNA Extract and Quantitative Real-Time RT-PCR
Total RNA was extracted from cell lines using Trizol (Invitrogen, Carlsbad, CA, http://www.invitrogen.com) reagent. Total RNA was treated with RNase-free DNase I (Invitrogen) to remove any DNA contamination. Reverse transcription (RT) was carried out by SuperScript III Reverse Transcriptase (Invitrogen). Real-time PCR for OCT4 isoforms was carried out using Power SYBR Green PCR Master Mix (Applied Biosystems, Foster City, CA, http://www.appliedbiosystems.com). The primers for quantitative real-time RT-PCR were shown, respectively, as follows: the OCT4A forward primer 5′-OA was 5′-CCCCTGGTGCCGTGAA-3′ and the reverse primer 3′-OA was 5′-GCAAATTGCTCGAGTTCTTTCTG-3′; the OCT4B forward primer 5′-OB was 5′-GTTAGGTGGGCAGCTTGGAA-3′ and the reverse primer 3′-OB was 5′-TGTGGCCCCAAGGAATAGTC-3′; the β-actin forward primer was 5′-ACCGAGCGCGGCTACAG-3′ and the reverse primer was 5′-CTTAATGTCACGCACGATTTCC-3′. Amplification of β-actin cDNA was analyzed for all of the samples as a normalizing control. Relative quantitation of the expression levels of OCT4 isoforms was analyzed by using the 2-ΔΔCT method and the PCR products were verified by sequencing analysis. The method of RT-PCR experiment was described in supporting information methods.
The human OCT4A and OCT4B cDNAs were obtained by RT-PCR from human EC cells (PA-1). PCR was done with Platinum Pfx DNA polymerase (Invitrogen).
Double-promoter plasmids (Figs. 2A, 2B and 3B–3D), bicistronic plasmids (Fig. 4A), bicistronic luciferase reporter plasmids (Fig. 6A), and plasmids expressing genes fused to green fluorescent protein (GFP) (Fig. 5A) were constructed as described in supporting information methods. All the plasmids were sequenced to verify their structure and open reading frames.
Cell Culture and DNA Transfections
HepG2, PA-1, HeLa, MCF-7, and ES cells were heat shock treated at 45°C for increasing periods of 0–2 hours. HepG2 and PA-1 cells were exposed to oxidative stress induced by 800 μM hydrogen peroxide (H2O2) for 0–24 hours or 0–9 hours.
HeLa cells or MCF-7 cells were transfected with Lipofectamine 2000 (Invitrogen). Cell lysates were prepared for Western blotting experiments and reporter gene analysis 24–48 hours after transfection. Luciferase activity was measured by using Dual-luciferase Reporter Assay System (Promega, Madison, WI, http://www.promega.com). To identify the protein expression and subcellular localization, cells were taken photos with Zeiss 200 inverted fluorescent microscope (Carl Zeiss, Jena, Germany, http://www.zeiss.com) or Zeiss confocal microscope (Zeiss-LSM 510 META) 24–48 hours after transfection. Stable transfected HeLa cell lines expressing GFP or OCT4B-190 were selected using 1,000 μg/ml G418. And three individual stable transfected cell clones were obtained for each plasmid.
Western Blotting Analysis
Anti-human OCT4 antibody (1:1,500, sc-8629, Santa Cruz Biotechnology Inc., Santa Cruz, CA, http://www.scbt.com) corresponds to the carboxy terminus of OCT4 and therefore recognizes both OCT4A and OCT4B. The anti-α-tubulin antibody (1:5,000; Sigma, St. Louis, MO, http://www.sigmaaldrich.com/) was used as a normalizing control. To verify whether the bands were specific for the OCT4A/B protein, the same volume of cell extracts was prepared for peptide/antigen competition examination. OCT4 antibody was preincubated with 8 μg blocking peptide (sc-8269 P, Santa Cruz Biotechnology Inc.). Then the two membranes were co-overexposed.
For each sample, approximately 1 × 106 cells were prepared for apoptosis assay using APO BRDU [TUNEL] kit (BioSource, Camarillo, CA, http://www.invitrogen.com/site/us/en/home/brands/BioSource.html), as instructed by the manufacturer. The cells were immediately analyzed on a FACS-LSR (Becton-Dickton, San Jose, CA) equipped with Celluest (Becton-Dickton) software.
Data were analyzed by Student's t test. A value of p < .05 was considered statistically significant.
Structure and Expression Pattern Analysis of Human OCT4A and OCT4B
As schematic structure shown in Figure 1A, OCT4A and OCT4B are generated from the same OCT4 gene by alternative splicing . They encode the protein isoforms OCT4A (360 amino acids) and OCT4B (265 amino acids), respectively. The two OCT4 protein isoforms are different in N-transactivation domain but identical in POU DNA-binding domain and C-transactivation domain.
It is important to distinguish the expression of OCT4A and OCT4B at both the mRNA and protein levels. We first examined OCT4 isoforms transcripts in several human cell lines by RT-PCR (data not shown) and quantitative real-time RT-PCR (Fig. 1C). To avoid the false-positive detection of OCT4 expression confused by OCT4 pseudogene [28, 29], the respective unique intron-spanning primers for OCT4A and OCT4B were used and the PCR products were verified by sequencing analysis. OCT4A was highly expressed in human EC cell lines PA-1 and NTERA-2. Relative to EC cells, only basal level transcription of OCT4A in human tumor cell lines (HepG2, HeLa, OS732, MCF-7) and normal cell lines (HEK293 and fibroblast) was detected. Comparatively very low levels of OCT4B expression was demonstrated in all human cell lines tested (Fig. 1C). Western blotting was used to detect the expression of OCT4A and OCT4B at protein level (Fig. 1D). OCT4A (48 kDa) was only highly expressed in human EC cell lines. A 23-kDa protein was detected in HepG2 cell line. The question of whether this 23-kDa protein is OCT4 isoform remained to be answered. Considering the presence of SNP site at the OCT4B (265 amino acids) start codon , we examined the DNA sequence of the OCT4B RT-PCR products containing the SNP site. Interestingly, the SNP site was “G” in HepG2 and HEK293 cells and it was “T” in other cell lines (Fig. 1E). It was predicated that a start codon ATG did not exist in the mRNA to encode the OCT4B protein containing 265 amino acids in HepG2 and HEK293 cells. Thus, the 23-kDa protein detected in HepG2 cells could attribute to three possibilities: (a) the 23-kDa protein was a product of OCT4A transcript; (b) the 23-kDa protein was encoded by other start codon on the OCT4B mRNA; (c) the 23-kDa protein might be a nonspecific band recognized by the anti-OCT4 antibody.
Identification of Protein Products Encoded by OCT4B
To investigate the 23-kDa protein detected in HepG2, we constructed the double-promoter expression plasmid modified from the plasmid pQCXIN (Fig. 2A). To test the first possibility mentioned earlier, the OCT4A expression vector was transiently transfected into HeLa cells and no 23-kDa protein band was detected by Western blotting using the anti-OCT4 antibody (data not shown). Then, we tested the second possibility. As the thymine/guanine (T/G) polymorphism located in the start codon (ATG) of OCT4B (265 amino acids), we constructed OCT4B expression vectors containing ATG or AGG at the translation initiation site, respectively (Fig. 2B). Unexpectedly, three protein bands were identified in HeLa cells transfected with vector DpQCXIN-OCT4B-265ATG (Fig. 2C, lane1). These three proteins had molecular masses of 35 kDa, 23 kDa, and 20 kDa. However, only two protein bands were identified in HeLa cells transfected with vector DpQCXIN-OCT4B-265AGG (Fig. 2C, lane 2). Compared with the bands in lane 1, the higher molecular weight protein (35 kDa) disappeared in lane 2. Although HeLa cells with mock-transfected (Fig. 2C, lane 3) and nontransfected (Fig. 2C, lane 4) did not synthesize the same protein bands. In addition, the 23-kDa protein expressed in transfected HeLa cells was identical to the endogenous protein expressed in human HepG2 cells (Fig. 2D). To verify whether the 35, 23, and 20 kDa protein bands were OCT4B proteins, we carried out competitive Western blotting experiments with anti-OCT4 antibody alone or after preincubation with corresponding synthetic blocking peptide. The data in Figure 2E showed that, though by overexposure, 35, 23, and 20 kDa protein bands were absent in the blot probed with neutralized antibody and some nonspecific bands were seen elsewhere (Fig. 2E, top). These results indicated that the endogenous 23 kDa protein in HepG2 and the exogenous 35, 23, and 20 kDa proteins in HeLa cells transfected with the OCT4B expression vectors were specific OCT4B proteins. At the same time, the results ruled out the third possibility mentioned earlier.
Identification of the OCT4B Translation Initiation Codons by Site-Directed Mutagenesis
The use of internal initiation of translation to express several proteins in a single eukaryotic cellular mRNA is an important cellular mechanism and contributes to the generation of protein diversity . We hypothesized that human OCT4B could use three initiation codons in a single mRNA to encode the three OCT4B protein isoforms. Thus, we carried out site-directed mutations to find out the alternative translation initiation sites of OCT4B.
First, we determined the start codon of the highest molecular weight form (35 kDa) of OCT4B. We found that different protein products were caused by the SNP site in the OCT4B (265 amino acids) start codon (ATG vs. AGG) (Figs. 2C, lanes 1 and 2; Fig. 3B, lanes 1 and 2). This indicated that the 35 kDa protein was initiated at ATG (at position 102–104) encoding 265 amino acids. It also suggested that the OCT4B-265 protein form would not be expressed in populations whose SNPs at this site were “AGG.”
Second, we investigated the start codon of the lowest molecular weight form (20 kDa) of OCT4B. Sequence analysis of the OCT4B cDNA showed that it contained another putative ATG translational initiation codon at position 405–407 in-frame (Fig. 3A). We then constructed expression vector DpQCXIN-OCT4B-164ATG which contains a deleted ORF of OCT4B starting at the second ATG (at position 405–407). HeLa cells transfected with DpQCXIN-OCT4B-164ATG only synthesized the 20-kDa protein form (Fig. 3B, lane 3). As expected, direct mutagenesis of this ATG at 405–407 (methionine) to GCT (alanine) abolished the synthesis of the 20-kDa protein band but not the 35 and 23 kDa protein bands (Fig. 3B, lane 4). Introduction of a TGA stop codon at 396–398 just upstream of ATG at 405–407 abolished both the 35 and 23 kDa OCT4B forms but not the 20 kDa form (Fig. 3B, lane 5). Thus, this result supported that the 20 kDa OCT4B form was synthesized at the second ATG start codon and it contained 164 amino acids.
Finally, we identified the start codon of the 23 kDa form of OCT4B. The size of this OCT4B isoform suggested that it could result from an initiation codon located between the first ATG at 102–104 and the second ATG at 405–407 of OCT4B mRNA. The absence of any other ATG start codon between the two ATG start codons forced us to consider other non-ATG start codons. Studies found that unusual initiation codons CTG and GTG could be used as putative alternative start codons [15, 31–33]. According to the nucleotide sequence of OCT4B cDNA (Fig. 3A), it appeared that CTG at nucleotide positions 204–206, 294–296, 327–329, 384–386, and GTG at 315–317 in-frame might be the candidate of the 23 kDa initiation codon. We constructed a subset of OCT4B expression vectors with site-directed mutations on those potential initiation condons (Fig. 3C). HeLa cells transfected with mutation vectors clearly showed that mutation of CTG at 204–206, 294–296, 384–386, and GTG at 315–317 did not abolish the synthesis of 23 kDa form of OCT4B (Fig. 3C, lanes 2–4, 6). However, mutation of CTG at position 327–329 exclusively abolished the synthesis of 23 kDa form of OCT4B (Fig. 3C, lane 5). This result demonstrated that the CTG at 327–329 encoding a new protein isoform of OCT4B containing 190 amino acids. Simultaneously mutagenizing any two of the three start codons in the OCT4B gene, only one form was synthesized in HeLa cells transfected with the mutation vectors (Fig. 3D). The double mutation experiments further confirmed the conclusion that OCT4B was able to use at least three initiation codons in a single mRNA encoding three OCT4B protein isoforms. In conclusion, a single OCT4B mRNA could encode at least three protein isoforms: OCT4B-265 (AUG-initiated isoform encoding 265 amino acids), OCT4B-164 (AUG-initiated isoform encoding 164 amino acids), and OCT4B-190 (CUG-initiated isoform encoding 190 amino acids).
Identification of the Putative IRES Element in OCT4B mRNA
IRES-mediated translation can be initiated in internal region of mRNA resulting in alternative initiation of translation and generating protein isoforms [24, 25]. To find out if the translation mechanism of OCT4B protein isoforms was resulted from the presence of the IRES element, bicistronic plasmids were constructed to identify the existence of IRES in OCT4B mRNA. As schematic structure shown in Figure 4A, these plasmids encode bicistronic mRNAs containing two tandem ORFs of RFP and GFP. Because of the stop codon introduced in the first ORF (expressing RFP protein), the second one (expressing GFP protein) cannot be translated unless an IRES element located between the two ORFs. Thus, these bicistronic plasmids enabled us to examine the existence of the OCT4B IRES.
MCF-7 cells were transiently transfected with bicistronic plasmids described in Figure 4A. As shown in Figure 4B, the plasmid RFP-GFP only expressing RFP protein was regarded as negative control (Fig. 4B, a–c). Both RFP and GFP proteins were expressed by the positive plasmid RFP-pQCXIN (IRES)-GFP (Fig. 4B, p–r). When nt 1–101 and nt 297–404 of OCT4B cDNA fragments were introduced in plasmids RFP-OCT4B (a)-GFP (Fig. 4B, d–f) or RFP-OCT4B (c)-GFP (Fig. 4B, m–o), respectively, GFP expression was not observed. However, nt 102–326 of OCT4B in RFP-OCT4B (b)-GFP drove the translation of GFP (Fig. 4B, g–i). As expected, the inverted fragment of nt 102–326 in RFP-OCT4B (b-REV)-GFP did not drive the translation of GFP (Fig. 4B, j–l). Thus, a 225-bp fragment of OCT4B at positions 102–326 between the first ATG start codon and the CTG start codon exhibited the IRES activity. Therefore, a putative IRES element was identified in OCT4B mRNA.
Subcellular Localization of OCT4 Isoforms in Cells
Most alternative initiation of translation regulates the subcellular localization of protein isoforms thus generates functional diversity [34, 35]. As the schematic structure analysis shown in Figure 5B, four protein isoforms of OCT4 was identical in C-termini. Although a putative nuclear localization signal (RKRKR)  remained in all these OCT4 isoforms, the subcellular localization of OCT4 isoforms might be different. We investigated the subcellular localization of OCT4A, OCT4B-265, OCT4B-190, and OCT4B-164.
MCF-7 cells were transiently transfected with plasmids expressing single OCT4 isoform fused with GFP proteins. Being a transcription factor, OCT4A-GFP protein clearly localized at the nucleus (Fig. 5C, c, d). However, OCT4B-265-GFP, OCT4B-190-GFP, and OCT4B-164-GFP proteins were all diffusely localized both at the nucleus and in the cytoplasm (Fig. 5C, e–j). GFP alone was found throughout the cells (Fig. 5C, a, b). In addition, many granules were found formed and scattered in some MCF-7 cells transfected with OCT4B-265-GFP or OCT4B-190-GFP.
OCT4B-190 Protein Expression when Cells Were Under Stress
Substantial evidence indicated that cellular IRES elements were predominantly found in specific mRNAs involved in the control of stress response and apoptosis induction. And IRES-dependent protein synthesis is often activated under conditions when the vast majority of cap-dependent protein translation is greatly reduced under stress conditions [24–27, 37]. To find out if the OCT4B-190 isoform was involved in cell stress and apoptosis, we first investigated whether the putative IRES element of OCT4B responds to the stress. As shown in Figure 6A and 6B, the firefly luciferase activity was apparently increased under heat shock when it was normalized to renilla luciferase activity. This suggested that, under stress, the initiation of translation by this putative IRES element of OCT4B was maintained whereas the cap-dependent protein translation was greatly reduced. Then, we analyzed the endogenous protein expression of OCT4B-190 under cell stress. As the endogenous expression of OCT4B-190 was detected in HepG2 cells (Figs. 1D, 2D), we tested the HepG2 cells first. When HepG2 cells were treated with heat shock at 45°C, OCT4B-190 expression was increased (Fig. 6C). The same result was obtained when HepG2 cells were exposed to 800 μM hydrogen peroxide (H2O2; Fig. 6D). To determine the OCT4B-190 expression activity in other human cells, we investigated the human EC cell line PA-1. OCT4B-190 expression in PA-1 cells was also upregulated under heat shock and oxidative stress (Fig. 6E, 6F). In addition, the expression of OCT4B-190 in human ES cells was also upregulated under stress conditions (Fig. 6G). Furthermore, the endogenous OCT4B-190 protein expression was also upregulated under hypoxia stress and genotoxic stress (data not shown) and was also detected upregulated under cell tress in HeLa and MCF-7 cells (supporting information Fig. S1A, S1B). Thus, we confirmed that IRES-mediated translation of OCT4B-190 was upregulated during cell stress.
The Antiapoptotic Function of OCT4B-190
To investigate the biological function of OCT4B-190 in response to cell stress, HeLa cells were stably transfected with plasmid DpQCXIN-OCT4B-190 expressing single OCT4B-190 isoform (Fig. 3D, lane 3). Untransfected HeLa cells and HeLa cells stably transfected with GFP were used as controls. The different clones were treated with heat shock at 45°C for 2 hours. Compared with untreated cells, the extents of apoptosis induced by heat shock in untransfected and mock-transfected HeLa cells were 15.5% and 19.4%, respectively. However, the extent of apoptosis induced by heat shock in HeLa cells expressing OCT4B-190 was as low as only 2.6% (Fig. 7A, 7B). Thus, OCT4B-190 showed an antiapoptotic function in response to cell stress.
The Relocalization of OCT4B-190 Induced by Heat Shock
OCT4B-190-GFP was found in many scattered granules in some transfected MCF-7 cells. It might be attributed to the relocalization of OCT4B-190 responding to cell stress. To test this hypothesis, we examined the subcellular localization of OCT4B-190 in heat-shock-treated and untreated MCF-7 cells after transient transfection with the vector expressing OCT4B-190-GFP fusion protein. Compared with untreated control cells, OCT4B-190-GFP distinctly relocalized to discrete granules induced by heat shock at 45°C for 3 hours (Fig. 7C).
It is well-established that in eukaryotic cells under stress conditions, some untranslated mRNA and associated proteins accumulates into discrete cytoplasmic foci termed SGs. SGs selectively recruited specific mRNA transcripts to regulate their stability and translation in response to a variety of stress conditions . We thus tested if OCT4B-190 relocalized to SGs upon heat shock by examining both OCT4B-190 and SG marker protein TIA-1  in MCF-7 cells. As shown in Figure 7D, OCT4B-190 fused to GFP and TIA-1 fused to RFP were coexpressed in MCF-7 cells. Coexpression of OCT4B-190 with TIA-1 did not alter OCT4B-190 localization in untreated transfected MCF-7 cells (Fig. 7D, g–i). In contrast, clear and robust colocalization of OCT4B-190 and TIA-1 was observed in cells exposed to heat shock (Fig. 7D, j–l). However, as a control, GFP did not localize to SGs induced by heat shock in MCF-7 cells coexpressed with TIA-1-RFP and GFP (Fig. 7D, a–f). The relocalization of OCT4B-190 suggested that OCT4B-190 may be involved in the stress response process.
Human ES cells are capable of indefinite self-renewal while maintaining the potential to differentiate into essentially all cell types, offering the possibility to meet many of the clinical demands for regenerative medicine [39, 40]. To achieve this potential, it is essential to understand the molecular mechanism regulating the pluripotency property and self-renewal of stem cells. OCT4 is regarded as a critical transcription factor in maintaining ES cells pluripotency and self-renewal capacities [1, 2]. Recently, the successful generation of induced pluripotent stem cells has shown that four genes (OCT4, SOX2, NANOG, LIN28 or OCT4, SOX2, MYC, KLF4) are sufficient to establish pluripotency in fully differentiated human somatic cells. Among these four genes, OCT4 and SOX2 appeared to be more essential [41, 42]. Thus, there is growing interest in understanding the characteristic functions of OCT4. But the confusion of OCT4A and OCT4B expression often leads to misleading results . Hence it is important to distinguish the expression of OCT4A and OCT4B both at the mRNA and protein levels. More importantly, the functional properties of OCT4A and OCT4B should be distinguished in particular.
In this study, we have shown that the different expression of OCT4A and OCT4B in several human cell lines. To further determine the OCT4 isoforms transcripts in these cell lines, we have determined the methylation status of the OCT4 promoter (supporting information Fig. S2). The distal enhancer region of OCT4 promoter is lowly methylated in PA-1 cells and relative highly methylated in HepG2, HeLa, OS732, MCF-7, HEK293, and fibroblast cells. The results are similar to the earlier studies . The apparent low amount of OCT4B compared with OCT4A in EC cells may attribute to the splicing mechanism, mRNA stability, and other unknown regulations.
Interestingly, the SNP site at the start codon of OCT4B is “G” in HepG2, and it suggests that HepG2 may have no OCT4B expression because of the absence of a start codon ATG (encoding 265 amino acids) in the mRNA of HepG2. However, we have identified that human OCT4B uses at least three initiation codons in a single mRNA encoding three OCT4B protein isoforms: OCT4B-265, OCT4B-190, and OCT4B-164. Thus, three OCT4B isoforms may be expressed in the population whose SNP is “T” at the OCT4B-265 start codon. Although two OCT4B isoforms may be expressed in the population whose SNP is “AGG” resulting in missing the OCT4B-265 isoform.
About 3–5% proportion of cellular mRNAs is translated by a cap-independent mechanism. And IRES-mediated translation is the only validated cap-independent translational mechanism in eukaryotic cells . We have identified the existence of a putative IRES element in the OCT4B mRNA. The OCT4B putative IRES (a 225-bp fragment) locates between the first ATG start codon (at 102–104) and the CTG start codon (at 327–329). Consequently, OCT4B-190 may be translated by IRES-mediated mechanism. As the 5′-UTR of OCT4B upstream of the first ATG start codon (at 102–104) does not have IRES activity; OCT4B-265 isoform may be translated by cap-dependent mechanism. Because of the absence of IRES between the CTG start codon (at 327–329) and the second ATG start condon (at 405–407), OCT4B-164 may be translated by the same IRES element used by OCT4B-190. This is similar to that found in other IRES-containing cellular mRNAs. Such as FGF-2, the longest FGF-2 isoform which initiates at the CTG codon nearest the 5′ end of the mRNA is exclusively expressed in a cap-dependent manner, whereas the other isoforms that initiate from three CTG and one ATG start codon are expressed under the control of a single IRES element [45, 46]. The c-Myc mRNA can be translated both by cap-dependent mechanism and IRES-dependent mechanism generating two isoforms alternative initiation at CUG and AUG codon . Diversely, the alternative translation initiation of VEGF mRNA is regulated by two independent IRESs . Some cellular IRES elements located within the coding region usually result in the production of several protein “isoforms” [19, 21, 48]. The IRES-mediated alternative translation initiation in a single OCT4B mRNA contributes to the diversity of OCT4 products. It may also endow OCT4 gene more complex and diverse biological functions.
In fact, many confirmed cellular IRES elements also hold cryptic promoter activity, such as HIF-1a, VEGF, c-myc, and XIAP . And we have inserted the putative IRES element of OCT4B into a promoterless vector and found in deed it has cryptic promoter activity (supporting information Fig. S3). However, this element of OCT4B does play a regulation role under stress as an IRES element at the level of translation but not as a promoter at the level of transcription because, under stress, the relative expression of firefly luciferase protein generated from the putative IRES of OCT4B is apparently increased (Fig. 6B), whereas the relative transcription of firefly luciferase mRNA is not increased (data not shown). Thus, we intend to think that this fragment works as an IRES element.
IRES-mediated translation represented a cellular “backup” pool leading to cell survival or cell death under unfavorable conditions and it confers a regulation mechanism of selective-translation at the posttranscriptional level to immediately respond to rapid changes [25, 27]. We have focused on the function of OCT4B-190 during stress response and apoptosis. It is well known that cell stress often induces cell apoptosis. Our result shows that under heat shock, the apoptosis rate of HeLa cells over-expressing OCT4B-190 is significantly lower than that of the control groups. Taken together, the endogenous expression of OCT4B-190 is at lower levels under normal conditions and at higher levels under stress conditions to protect against apoptosis. This functional identification of OCT4B-190 might enrich our understanding of antiapoptotic signals, especially in stress response mechanism.
In addition, we have identified the subcellular localization of OCT4 isoforms because the localization of a protein in the cells often associates with its functional properties. OCT4B-265, OCT4B-190, and OCT4B-164 proteins all diffusely localize both in the nucleus and cytoplasm. This is contrary to previous result of cytoplasm localization of OCT4B in human ES cells and preimplantation embryos [12, 13]. This inconsistency may be due to the possibility that OCT4B may have diverse functions resulting in different localization in different cells. The change in localization is closely correlated with the functional switch of a protein. Expectably, we have found that OCT4B-190 changed its localization from diffusing in the nucleus and cytoplasm to focusing in granules during stress response. The relocalization to SGs has confirmed the role of OCT4B-190 responding to cell stress.
Although three protein isoforms are generated from OCT4B mRNA, the endogenous protein isoforms may not simultaneously express in the same cell type. The protein expression of each of the isoforms may be accurately regulated, respectively. The OCT4B-265 and OCT4B-164 proteins, like OCT4B-190, are also partially in cytoplasmic; however, they may be endowed with different functions compared with OCT4B-190. We have only detected the endogenous expression of OCT4B-190 isoform and shown OCT4B-190 a biological function. The endogenous expression and functional characterization of OCT4B-265 and OCT4B-164 need to be further investigated.
It is well known that OCT4A is responsible for the stemness properties in human ES and EC cells [1, 13]; whereas OCT4B is shown to sustain HeLa cells survival in stress conditions by reducing cell apoptosis in this work. This discovery endows OCT4 gene a novel and significant function in stress resistance. The OCT4 gene, regulated by alternative RNA splicing and alternative translation initiation at the posttranscriptional level, may perform diverse functions in cell differentiation, cell growth, proliferation, the regulation of apoptosis, and so on. This work evokes the significance to distinguish the biological function of the protein products of OCT4 gene. There is growing interest in the investigation of OCT4 in ES cells, adult stem cells and cancer cells. Therefore, it requires more emphasis on the identifying and distinguishing the protein isoforms among the four protein products of OCT4 which would help us to dissect OCT4 functions indeed.
We have identified the alternative translation of OCT4B mediated by a putative IRES element in a single mRNA resulting in three protein isoforms: OCT4B-265, OCT4B-190, and OCT4B-164. Therefore, OCT4 gene encodes at least four protein isoforms. Here, OCT4B-190 is shown to have the significant role in cellular stress response. This finding expands our understanding of a pivotal biological function of OCT4 in stress resistance besides maintaining ES pluripotency and self-renewal capacities.
This work was supported by grants from the Ministry of Science and Technology of China (2006CB943601), NSFC (30688002; 30600304; 30800564), and Chinese Academy of Sciences (KSCX2-YW-R-133).
DISCLOSURE OF POTENTIAL CONFLICTS OF INTEREST
The authors indicate no potential conflicts of interest.