Dppa2 and Dppa4 Are Closely Linked SAP Motif Genes Restricted to Pluripotent Cells and the Germ Line


  • Joanna Maldonado-Saldivia,

    1. Gurdon Institute, Wellcome Trust/Cancer Research United Kingdom, Cambridge, United Kingdom
    Search for more papers by this author
  • Jocelyn van den Bergen,

    1. Australian Research Council Centre for Biotechnology and Development, Department of Paediatrics, University of Melbourne, Murdoch Children's Research Institute, Royal Children's Hospital, Parkville, Victoria, Australia
    Search for more papers by this author
  • Margarita Krouskos,

    1. Australian Research Council Centre for Biotechnology and Development, Department of Paediatrics, University of Melbourne, Murdoch Children's Research Institute, Royal Children's Hospital, Parkville, Victoria, Australia
    Search for more papers by this author
  • Mike Gilchrist,

    1. Gurdon Institute, Wellcome Trust/Cancer Research United Kingdom, Cambridge, United Kingdom
    Search for more papers by this author
  • Caroline Lee,

    1. Gurdon Institute, Wellcome Trust/Cancer Research United Kingdom, Cambridge, United Kingdom
    Search for more papers by this author
  • Ruili Li,

    1. Australian Research Council Centre for Biotechnology and Development, Department of Paediatrics, University of Melbourne, Murdoch Children's Research Institute, Royal Children's Hospital, Parkville, Victoria, Australia
    Search for more papers by this author
  • Andrew H. Sinclair,

    1. Australian Research Council Centre for Biotechnology and Development, Department of Paediatrics, University of Melbourne, Murdoch Children's Research Institute, Royal Children's Hospital, Parkville, Victoria, Australia
    Search for more papers by this author
  • M. Azim Surani,

    1. Gurdon Institute, Wellcome Trust/Cancer Research United Kingdom, Cambridge, United Kingdom
    Search for more papers by this author
  • Patrick S. Western Ph.D.

    Corresponding author
    1. Australian Research Council Centre for Biotechnology and Development, Department of Paediatrics, University of Melbourne, Murdoch Children's Research Institute, Royal Children's Hospital, Parkville, Victoria, Australia
    • ARC Centre for Biotechnology and Development, Department of Paediatrics, University of Melbourne, Murdoch Children's Research Institute, Royal Children's Hospital, Flemington Rd., Parkville, Victoria 3052, Australia. Telephone: 61-3-8341-6426; Fax: 61-3-8341-6429
    Search for more papers by this author


Despite the enormous medical potential of ESCs, the molecular mechanisms conferring the ability to differentiate into all cell types of the embryo remain elusive. We used an in silico approach to identify genes expressed exclusively in mouse preimplantation embryos and pluripotent cell lines. Two of these genes were developmental pluripotency-associated gene 2 (Dppa2) and Dppa4, which we show are closely linked genes encoding putative nuclear SAP domain proteins expressed in human and mouse pluripotent stem cells and germ cell tumor-derived embryonal carcinoma cells. In the mouse, these genes are transcribed in germinal vesicle-stage oocytes and throughout the cleavage stages of embryogenesis. They then become restricted to the pluripotent inner cell mass of blastocysts and are subsequently downregulated. After gastrulation, Dppa2 and Dppa4 are expressed only in the developing germ line, showing that these genes mark cells of the pluripotent cycle. In the germ line, both genes are downregulated as the germ cells commit to the oogenic pathway or soon after commitment to the spermatogenic pathway. We have observed similar germ line expression profiles for other pluripotent markers, and these results are consistent with the hypothesis that pluripotent markers must be downregulated during fetal germ line development, a process that may be required to facilitate appropriate germ line differentiation. The study of expression and function of pluripotent markers such as Dppa2 and Dppa4 is likely to unveil new aspects of the regulation of pluripotency and germ line development in mammals.


Cells of the cleaving mammalian embryo have the potential to differentiate into any type of cell, a property known as totipotency. During blastocyst formation, the cells of the inner cell mass (ICM) retain the ability to contribute all cell types of the embryo proper, whereas the outer trophectoderm (TE) layer gives rise only to extraembryonic tissues. The pluripotent quality of ICM cells can be captured in vitro by the derivation of ESCs, which are self-renewing cells that may be induced to differentiate both in vitro and in vivo. However, the molecular mechanisms that confer on early embryonic cells and ESCs the remarkable property of pluripotency remain poorly understood. Apart from transcription factors such as OCT4 [1], SOX2 [2], and NANOG [3, 4], very few factors have been shown to be involved specifically in the establishment and maintenance of cellular pluripotency.

Developmental potential of the ICM is lost as the embryo proceeds through gastrulation and cells commit to various somatic fates. This process involves the silencing of pluripotency genes (such as Oct4, Sox2, and Nanog [5, [6], [7]–8]) and the activation of somatic-specific programs. The specification of germ cells provides an important exception in this process, since germ cells are protected from taking on somatic characteristics and maintain expression of pluripotency genes. The germ line expression of these pluripotency regulators is consistent with the functional role of the germ line in generating a unique cell lineage, which can form gametes that combine to form the totipotent zygote. Although the functional activities of these genes are not understood, their regulation during germ cell development is strictly controlled. For example, at or soon after the stages of germ line sex determination, the expression of some of these markers becomes abolished or reduced [7, 8]. We aim to identify further components of the pluripotency/germ line cycle, and to this end, we developed a screen based on expressed sequence tag (EST) database mining. Using the huge collection of data compiled in EST databases, we undertook an in silico screen for genes expressed only in pluripotent cells by comparing gene representation in preimplantation embryos and pluripotent cell lines with representation in adult somatic tissues.

Among the genes found in our screen to be enriched in pluripotent cell types was developmental pluripotency-associated gene 2 (Dppa2), a poorly characterized gene. In addition, examination of the surrounding genomic region revealed that Dppa4, a gene with similar expression to Dppa2, lies within 16 kilobases (kb). Previously, in silico screening and reverse transcription-polymerase chain reaction (RT-PCR) studies identified Dppa2 and Dppa4 among a set of genes that exhibit preimplantation and ESC expression patterns similar to those of Oct4 and whose expression can be variable in cloned embryos [9]. Here, we present the first detailed analysis of Dppa2 and Dppa4, showing that these genes are closely linked, that they encode putative SAP domain proteins, and that their expression is restricted to pluripotent cells and the developing germ line. In addition, the expression patterns of these two genes in blastocysts and the germ line are similar to each other, although not identical, indicating that they may be differentially regulated. Together with other data [7, 8], the expression of Dppa2 and Dppa4 suggests that downregulation of pluripotent markers is required for differentiation of the early male and female germ lineages.

Materials and Methods

Mouse Strains, Embryonic Staging, and Dissections

The age of all mouse embryos was estimated using the day of the appearance of a vaginal plug as embryonic day 0.5 (E0.5). Inbred 129/Sv mice were used to obtain blastocysts for in vitro culture. For all whole-mount RNA in situ hybridization (WISH) experiments, embryos were obtained from hsd/ola × hsd/ola matings. Wild-type animals from the outbred MF1 strain were used for all other experiments. Collection of germinal vesicle (GV) oocytes, preimplantation embryos, and genital ridges was carried out following the instructions in Nagy et al. [10]. GV oocytes devoid of cumulus cells were selected using an inverted microscope.

Blastocyst Outgrowths

Expanded blastocysts (E4.5) were transferred into 1 drop (30 μl) of Dulbecco's modified Eagle's medium (DMEM) ESC medium containing leukemia inhibitory factor (LIF) on a Permanox eight-well Lab-Tek slide (Nalge Nunc International, Rochester, NY, http://www.nalgenunc.com) under oil and cultured for 2–4 days (37°C, 5% CO2). A minimum of four outgrowths were prepared and subjected to immunofluorescent staining, and the experiment was repeated at least three times to ensure confidence in the results.

Embryoid Bodies

A single-cell suspension of KES1 or W9.5 ESCs in DMEM without LIF was aliquoted on the inside lid of a 14-cm tissue culture dish (Bibby Sterilin, Staffordshire, U.K., http://www.bibby-sterilin.co.uk) in 20-μl drops, corresponding to approximately 600 undifferentiated cells per drop. The drops were incubated (37°C, 5% CO2) for 2 days to allow cell aggregation. The aggregates were then transferred to a nonadherent tissue culture dish containing DMEM without LIF for an additional 5 days. Finally, the embryoid bodies were seeded onto a gelatinized plate for 3 more days [11]. Samples were collected at days 0, 2, 4, 7, and 10, and RNA was prepared. Experiments were repeated for both KES1 and W9.5 ESCs, with similar results.

RNA Isolation, Reverse Transcription, Polymerase Chain Reaction, and Real-Time Polymerase Chain Reaction

Total RNA was isolated using Trizol reagent (Invitrogen, Carlsbad, CA, http://www.invitrogen.com) or RNeasy columns (Qiagen) and treated with DNase1 using DNA-free (Ambion, Austin, TX, http://www.ambion.com). One microgram of total RNA was reverse transcribed using a poly(T) primer and 1 μl of Superscript (Invitrogen), and 1 μl (50 ng of cDNA) of the reaction mixture was used in subsequent polymerase chain reaction (PCRs). All PCRs were carried out according to standard protocols. Real-time PCR was performed using the mouse Universal ProbeLibrary system and FastStart Taqman Probe Master mix with ROX (Roche Diagnostics, Basel, Switzerland, http://www.roche-applied-science.com). Fifty nanograms of cDNA was subjected to amplification using an ABI 7900 HT Real-Time PCR instrument. Samples were run in triplicate, and experiments were performed twice. All samples were normalized against Hypoxanthine guanine phosphoribosyl transferase (Hprt) (embryonic stem differentiation) or Mvh (gonad development) using the comparative CT method (ΔΔCT). Cycle parameters and primers for RT-PCR and the real-time PCR primer/Universal ProbeLibrary probe set combinations used are included in the supplemental online methods.


Blastocyst outgrowths were processed in a Permanox eight-well Lab-Tek slide (Nalge Nunc). Preimplantation embryos were processed in a 60-microwell dish (Nalge Nunc). Immunostaining of embryos was performed as described previously [12]. Embryonic gonads were fixed in 4% paraformaldehyde for 15 minutes and embedded in OCT, and 10-μm sections were cut. Adult testis was fixed in Bouin's solution and embedded in wax. Six-micrometer sections were cut, dewaxed, rehydrated, and permeabilized using 1% Triton X-100 for 10 minutes. Antigens were retrieved in 0.01 M citrate buffer (95°C, 15 minutes), and immunostaining was performed. Immunostaining of gonad sections was performed in phosphate-buffered saline containing 1% bovine serum albumin using the M.O.M. Staining Kit (Vector Laboratories, Burlingame, CA, http://www.vectorlabs.com). The DPPA2 antibody was diluted 1:300, germ cell nuclear antigen (GCNA) (a gift from G. Enders, University of Kansas Medical Center, Kansas City, KS) was diluted 1:10, and the OCT4 antibody (catalog number 611202; BD Biosciences, San Diego, http://www.bdbiosciences.com) was diluted 1:200. Primary antibody binding was detected using secondary antibodies conjugated to Alexa Fluor 488 (goat anti-mouse IgG green) or 594 (goat anti-rabbit IgG red) (Molecular Probes Inc., Eugene, OR, http://probes.invitrogen.com) diluted 1:500. DNA was stained with TOTO3 (Molecular Probes) or 4,6-diamidino-2-phenylindole. Embryos and cells were mounted on slides in Vectashield (Vector Laboratories) and examined using a Radiance 2000 confocal scanning laser microscope (Bio-Rad, Hercules, CA, http://www.bio-rad.com), and images were captured using Bio-Rad Lasersharp software. A minimum of 10 embryos were prepared for each stage and subjected to immunofluorescent staining, and the experiment was repeated. Immunofluorescent staining of developing gonads was performed on several sections from at least three samples.

Western Blotting

Protein extracts were prepared, separated on 12% polyacrylamide gels, and blotted onto Immobilon-P membranes (Millipore, Billerica, MA, http://www.millipore.com) for 1 hour at 100 V according to standard protocols. Primary antibodies were detected with anti-mouse/rabbit horseradish peroxidase (Amersham Biosciences, Piscataway, NJ, http://www.amersham.com) followed by enhanced chemiluminescence detection (Amersham).

Antibody Generation

Affinity-purified, rabbit polyclonal antibodies were prepared by Eurogentec (Seraing, Belgium, http://www.eurogentec.be) against a mouse DPPA2 peptide, RSRAKKNALPPNMPP. This epitope contains sequence that is present in DPPA2 but absent from DPPA4, thus enhancing the specificity of the antibody generated.

RNA Whole-Mount In Situ Hybridization

WISH using E9.5–E17.5 partial embryos or gonad-mesonephros complexes was performed as described [13]. WISH using blastocyst stage embryos was performed according to the peri-implantation stage protocol of Rossant [14]. Experiments were performed three times with at least 10 embryos. Sense and antisense probes used were generated from the following sequences: Dppa2 (AK010743), base pairs 180–533; Dppa4 (AK010753), 71–534; Oct4 (X52437), 526-1140; and Sox9 (AF421878), 671-1172.

Human Cell Sources

Human embryonic stem (hES) (HES2 and HES4) [15] and human embryonal carcinoma (hEC) (GCT27X1) [16] cell RNA was provided by Martin Pera (Australian Stem Cell Centre, Monash University, Clayton, VIC, Australia).


Digital Comparison of Mouse Preimplantation Embryos and Adult Tissues

We designed a bioinformatics program to rank genes in 13 UniGene pluripotent cell-type libraries from the highest to the lowest represented, provided the genes were absent from 14 adult somatic tissue libraries. Somatic libraries likely to include germline-specific genes (such as mid- and late-gestation whole embryos) were not selected, since many germ line-specific genes are also expressed in pluripotent cells. Genes highly represented in pluripotent cells but absent from a variety of adult cell types (Table 1) included Oct4 [17], Esg1 [18], Stella [19], and Rex-1 [20]. Among the first 50 entries, 31 corresponded to unknown genes, and the function of many of the 19 known genes had not been elucidated. A previously unidentified Riken clone, 2410088E07Rik, was ranked third in the screen and chosen for further study (Table 1). Further analysis showed that this sequence corresponded to the poorly characterized gene Dppa2.

Table Table 1.. Results of the in silico screen
original image

Genomic Structure of Dppa2 and Adjacent Region

Bioinformatic analysis reveals that Dppa2 is split into seven exons within 9.2 kb of sequence on chromosome 16B5. A detailed examination of this region revealed that a related gene, Dppa4, lies 16.2 kb upstream of Dppa2. Dppa4 spans 10.7 kb, and the exon-intron structure of Dppa4 appears similar to that of Dppa2. Importantly, the genomic arrangement of Dppa2 and Dppa4 has been conserved in humans, where these genes map to chromosome 3q (syntenic with mouse) (hDPPA2 [hDppa2] NM_138815.2 and hDPPA4 [hDppa4] NM_018189.2, respectively). The conserved genomic organization and close expression relationship suggest that Dppa2 and Dppa4 may be coregulated. Analysis of the DNA upstream of Dppa4 and Dppa2 was performed using the ECR Browser facility at http://www.dcode.org, and the results are summarized in supplemental online Figure 1. Use of the data in supplemental online Figure 1 in conjunction with http://www.dcode.org allows visualization of putative transcription factor binding sites in the Dppa4/Dppa2 upstream regions. Nine regions, each greater than 200 base pairs (bp), were at least 75% identical between mouse, human, and dog, and these regions contain binding sites for many transcription factors including POU domain proteins, SOX proteins, and orphan nuclear receptors. The significance of these binding sites remains unknown; however, it is of interest that no evolutionarily conserved regions more than 200 bp/75% identical were detected within the 16 kb of DNA separating Dppa4 and Dppa2. Interestingly, Morc, the gene nearest to Dppa2 and Dppa4, is expressed in the germ line and required during prophase 1 of mouse spermatogenesis [21, 22].

Considering the close genomic and expression relationship of mouse Dppa2 and Dppa4, we compared the predicted 301- and 298-amino acid protein sequences encoded by Dppa2 and Dppa4, respectively. Mouse DPPA2 and mouse DPPA4 share overall identity/similarity (32%/47%, respectively), whereas the mouse/human DPPA2/DPPA2 and DPPA4/DPPA4 proteins share 44%/61% and 52%/68% identity/similarity, respectively. DPPA2 contains a predicted DNA-binding SAP (SAF-A/B, Acinus, and protein inhibitor of activated STATs [PIAS]) domain (SMART domain prediction, European Molecular Biology Laboratory; [23]), within which four of five key functional residues are conserved in both the human and mouse orthologs (Fig. 1A). Despite the lack of a confidently predicted SAP domain within DPPA4 (the SMART score prediction falls just below the cutoff threshold), four of the five residues widely conserved in SAP domain proteins are also conserved in DPPA4 (Fig. 1A). This observation may be significant, as Dppa4 encodes at least two isoforms. Although the functional differences between these isoforms remain unknown, the region containing the putative SAP domain is differentially spliced between isoforms 1 (SAP+) and 2 (SAP). Both forms are expressed in ESCs and germ cells, but the SAP form is expressed at significantly lower levels than the full SAP+ form (J.v.d.B. and P.S.W., unpublished data). In addition to the putative SAP domain, there is a conserved 73-amino acid sequence of unknown function within both the mouse and human DPPA2 and DPPA4 C-terminal regions. Figure 1A summarizes the level of protein similarity for the mouse and human DPPA2 and DPPA4 protein sequences, the SAP domains, and the conserved C-terminal domains.

Figure Figure 1..

Dppa2 and Dppa4 encode evolutionarily conserved putative SAP domain-containing proteins expressed in pluripotent cells and the developing germ line. (A): Diagrammatic representation of the structure and sequence homology of the mouse and human DPPA2 and DPPA4 proteins. The aligned sequence shows the SAP domain of the mouse DPPA2 (Q9CWH0) and DPPA4 (Q8CCG4) sequences with similar/identical residues shaded in gray. The invariant residues, which define SAP domain function, are noted below the alignment. In each diagram, the numbers at the N- and C-terminal ends of the SAP and C-terminal domains refer to the amino acid residues of the mouse DPPA2 or DPPA4 protein sequences. (B): RT-PCR analysis of Dppa2 expression in ESCs, E13.5 embryos from which the gonads were removed, GV-stage oocytes, blastocyst, and E12.5 PGC samples. β-Actin served as a positive control. (C): RT-PCR analysis of Dppa2, Dppa4, and Oct4 expression in ES, EC, and EG cells and E13.5 embryonic head, trunk, and liver samples. Hprt served as a positive control. (D): RT-PCR analysis of Dppa2 and Dppa4 expression in postnatal days 1, 3, and 7 and adult testis samples and an adult ovary sample. Hprt serves as a positive control. (E): Real-time polymerase chain reaction (PCR) analysis of Dppa2, Dppa4, and Oct4 expression changes during ESC differentiation in embryoid bodies. All samples were normalized against Hprt expression, and expression levels are relative to the expression of each gene in undifferentiated ESCs (day 0 sample). It should be noted that this is not a comparison of the levels between Oct4, Dppa2, and Dppa4, but it is an analysis of the change in expression of each gene through time. For visualization purposes, an image of standard RT-PCR results after 30 cycles of PCR using the Dppa2, Dppa4, and Oct4 primers used in (C) and (D) has also been included. Abbreviations: aa, amino acid; C-TERM, C-terminal; E, embryonic day; EC, embryonal carcinoma; EG, embryonic germ; Emb, embryo; ES, embryonic stem; GV, germinal vesicle; Hprt, hypoxanthine guanine phosphoribosyl transferase; PGC, primordial germ cell; RT, reverse transcription.

Mouse Dppa2 and Dppa4 Are Expressed by Various Pluripotent Cell Types

The expression of Dppa2 and Dppa4 in pluripotent cells was assessed by various methods. Data gathered by NCBI suggested enriched expression of Dppa2 and Dppa4 in egg, embryonal carcinoma, and two-cell embryo libraries. Using RT-PCR, our experiments confirmed and extended previous observations [9], as Dppa2 expression was detected in ESCs, GV-stage oocytes, blastocysts, E12.5 primordial germ cells (PGCs), and embryonal carcinoma (EC) cells (Fig. 1B). Robust expression by GV oocytes strongly indicates that DPPA2 is present in the oocyte as a maternal protein. Likewise, when compared, strong expression of Dppa2, Dppa4, and Oct4 was detected in mouse ESCs, EG cells, and EC cells, but little or no expression was detected in somatic tissues of E13.5 embryos, confirming preferential expression of these genes in pluripotent cells (Fig. 1C). In addition, expression of Dppa2 and Dppa4 was not detected in adult somatic tissues (not shown), but expression of both genes was clearly detected in postnatal testis (days 1, 3, and 7 and adult) and in adult ovary (Fig. 1D). We then used quantitative real-time PCR normalized against Hprt to examine the expression of Dppa2 and Dppa4 as ESCs differentiated for 0, 2, 4, 7, and 10 days in embryoid bodies. At time 0, the Ct values (the number of PCR cycles taken to reach quantitative threshold levels) were 17.8, 21.8, and 23.1 for Oct4, Dppa2, and Dppa4, respectively, indicating that Oct4 is expressed more strongly than Dppa2, which is expressed more strongly than Dppa4 in ESCs. Strikingly, expression of both genes decreased sharply between days 0 and 4, and by day 4, expression was 5% of the level detected in ESCs at time 0. By contrast, the level of Oct4 expression was still 70% of the level expressed by ESCs at time 0. Also of significance was an initial sharp increase in the levels of Oct4 (1.7× time 0) and Dppa2 (2.9× time 0) at day 2. This may reflect a role for these two genes in the initial stages of ESC differentiation. In what may be a similar context, Oct4 overexpression in ESCs apparently results in differentiation [24]. By day 7, expression of all three genes had decreased to insignificant levels (0× day 0) (Fig. 1E).

Dppa2 and Dppa4 Expression Is Restricted to the ICM of Mouse Blastocysts

Expression of Dppa2 and Dppa4 has been reported in preimplantation stage mouse embryos; however, the tissue specificity of this expression remained undefined. We therefore used WISH to investigate whether Dppa2 and Dppa4 expression was limited to the inner cell mass in expanded blastocysts. Dppa2 expression was relatively easily detected and was limited to the ICM. However, although expression of Dppa4 was detected in the ICM in repeated experiments, it appeared to be expressed at a low level, which was more difficult to detect than Dppa2 (Fig. 2). This difference in expression level between Dppa2 and Dppa4 in blastocysts is of interest, as the opposite pattern was evident in developing germ cells.

Figure Figure 2..

Dppa2 and Dppa4 expression in developing blastocysts. Whole-mount RNA in situ hybridization using Dppa2, Dppa4, and Oct4 probes to detect expression in expanded blastocysts. A lateral view is shown in the top row, and a view from the proximal trophectoderm/inner cell mass (ICM) side is shown in the bottom row. Expression of all three genes appeared to be limited to the ICM, although Dppa4 was difficult to detect. Scale bar = 20 μm.

Mouse Dppa2 and Dppa4 Are Expressed by Developing Primordial Germ Cells

Because several pluripotency-specific genes are expressed in the developing germ line, we examined Dppa2 and Dppa4 expression in developing PGCs. Initially, examination of E12.5 partial embryos using WISH revealed that both Dppa2 and Dppa4 expression were detectable only in male and female gonads. In males, this expression was restricted to the developing testis cords and was therefore likely to be in Sertoli or germ cells. To distinguish these possibilities, Dppa2 and Dppa4 expression was examined in embryos from which the germ cells were depleted using busulphan treatment. Busulphan specifically depletes germ cells but leaves the somatic cells intact. By comparing Dppa2 and Dppa4 expression with Oct4 (germ cell-specific marker) and Sox9 (Sertoli cell-specific marker), we determined whether depletion of germ cells resulted in loss of Dppa2 and Dppa4 expression. In contrast to Sox9, but similar to Oct4, robust Dppa2 and Dppa4 expression was detected in untreated embryos but was reduced to almost undetectable levels in busulphan-treated gonads, confirming that both genes are expressed in germ cells (Fig. 3A). As shown later in this report, DPPA2 protein was also detected by immunohistochemical staining only in gonadal cells expressing OCT4.

Figure Figure 3..

WISH and real-time PCR analysis of Dppa2 and Dppa4 expression during germ cell development. (A): Dppa2 and Dppa4 expression was examined and compared with Oct4 and Sox9 in gonads of developing embryos treated with busulphan. Busulphan treatment results in depletion of germ cells but not somatic cells from the developing gonad. Oct4 is a germ cell-specific marker and acts as a positive control for the presence of germ cells in the gonad. Sox9 is expressed only in somatic cells of the gonad. With depletion of the germ cells from the developing gonads (+), Dppa2, Dppa4, and Oct4 expression was lost, indicating that Dppa2 and Dppa4 are specifically expressed by the germ cells of the gonad. As expected, Sox9 expression was unaffected by the busulphan treatment. (B): Expression of Oct4, Dppa2, and Dppa4 in E10.5 partial embryos and dissected E11.5 gonad mesonephros samples. At E10.5, Dppa2 expression was undetectable, whereas Dppa4 was detected in a substantially smaller number of germ cells than Oct4. (C): Expression of Dppa2 and Dppa4 in female E12.5–E14.5 gonad-mesonephros samples. (D): Dppa2 and Oct4 expression in female age-matched L and R gonads taken from the same embryo. (E): Dppa4 and Oct4 expression in female age-matched L and R gonads taken from the same embryo. Scale bars = 500 μm. (F): Real-time PCR analysis of Dppa2, Dppa4, and Oct4 expression changes in developing gonads isolated from E12.5–E17.5 female embryos. All samples were normalized against Mvh expression, and expression levels are relative to the expression of each gene at E12.5. (G): WISH analysis of Dppa2 and Dppa4 expression in male E12.5–E16.5 gonad-mesonephros samples. Scale bars = 500 μm. (H): Real-time PCR analysis of Dppa2, Dppa4, and Oct4 expression changes in developing gonads isolated from E12.5–E17.5 male embryos. All samples were normalized against Mvh expression, and expression levels are relative to the expression of each gene at E12.5. For (F) and (H), it should be noted that the comparison is not of the levels between Oct4, Dppa2, and Dppa4 but is an analysis of the change in expression of each gene through time. Abbreviations: E, embryonic day; L, left; PCR, polymerase chain reaction; R, right; WISH, whole-mount RNA in situ hybridization.

We next examined the temporal expression of Dppa2 and Dppa4 throughout PGC development. Despite detecting normal Oct4 expression in gastrulating and postgastrulation embryos, Dppa2 and Dppa4 expression was not detected in E6.5, E7.5, E8.5, or E9.5 embryos (not shown), indicating that the genes are not required for the specification of the germline. Consistent with this result, Dppa2 transcripts could not be detected in panels of single-cell cDNAs constructed from the nascent PGCs isolated from the proximal epiblast region of E7.5 embryos (not shown).

However, by E10.5, expression of Dppa4 was readily detected in primordial germ cells populating the developing gonads, whereas expression of Dppa2 was still not detectable. Although Dppa4 expression was readily detected in PGCs at this stage, the number of Dppa4-positive PGCs was significantly less than for Oct4, which is expressed by all PGCs at this developmental stage (Fig. 3B). In E11.5 and E12.5 gonads, both Dppa2 and Dppa4 expression could be readily detected in male and female PGCs (Fig. 3B); however, Dppa4 expression was consistently more easily detected than Dppa2, a pattern that was similar in all stages of PGC development where expression could be detected. Interestingly, this is opposite to the pattern observed in blastocysts, where Dppa2 was easier to detect than Dppa4. Combined, the temporal and expression level differences observed for Dppa2 and Dppa4 indicate that although these genes have similar expression patterns, their expression is likely to be controlled by at least some different genomic elements or other regulatory mechanisms (for example, mRNA stability).

In female E13.5 embryos (individuals from the same litters), Dppa2 and Dppa4 expression was downregulated in an anterior-posterior (A-P) wave consistent with the entry of the PGCs into meiosis. Consistent with this result, analysis using real-time PCR normalized to the germ cell-specific gene Mouse vasa homolog (Mvh) showed that in female gonads, downregulation of Oct4, Dppa2, and Dppa4 occurred primarily between E12.5 and E13.5 (Fig. 3F). As has been previously observed, Oct4 expression is also downregulated in this A-P pattern (Fig. 3D, 3E) [7, 25]. Using WISH, the A-P wave of Dppa4 downregulation was more difficult to detect than for Dppa2 or Oct4, as it appeared to occur rapidly and to significantly precede the Oct4 A-P wave (Fig. 3C). We therefore carefully compared the temporal downregulation of Dppa2 and Dppa4 with Oct4 in age-matched gonads from the same female E13–E13.5 embryos. By examining expression in gonad pairs where the left or right was subjected to Dppa2 or Dppa4 WISH and the other gonad to Oct4 WISH, we compared the relative time of downregulation of these two genes. Dppa2 appeared to be decreased at an equivalent time or very slightly prior to Oct4 (Fig. 3D). By contrast, downregulation of Dppa4 preceded that of Oct4 by several hours, so that by the time the Oct4 A-P wave was detected, Dppa4 was no longer expressed (Fig. 3E).

Using WISH to analyze male gonads, Dppa2 expression was detectable until E14.5 and Dppa4 until E16.5 (Fig. 3G). Quantitative analysis of this expression using real-time PCR (again normalized to Mvh) showed that Dppa2 expression decreased sharply between E12.5 (arbitrarily set at 1) and E13.5 (0.2 × E12.5 level) and Dppa4 steadily underwent a 60% decrease in expression level between E12.5 and E17.5 (Fig. 3H). Consistent with our previous observations [7] (but never before quantified), Oct4 expression decreased 60% between E12.5 and E14.5 and 80% by E17.5 (Fig. 3H). These expression patterns are similar to those observed for Esg1, Sox2, and Nanog in previous reports [7, 8]. Combined, these data indicate that the pluripotent markers Oct4, Nanog, Sox2, Esg1, Dppa2, and Dppa4 are all downregulated during commitment (in females) or soon after commitment (in males) to the female and male germ cell developmental pathways, indicating that pluripotent marker expression may be incompatible with normal differentiation of the male and female germ lineages.

DPPA2 Is a Nuclear Protein Expressed by Pluripotent and Primordial Germ Cells

These results show that, like Oct4 mRNA, Dppa2 and Dppa4 mRNA specifically marks pluripotent cell types, the inner cell mass and the developing primordial germ cells. The expression of the DPPA2 and DPPA4 proteins, however, remained unclear. Because Dppa2 and Dppa4 are predicted to encode similar putative DNA binding SAP domain proteins, we endeavored to assess the subcellular localization and expression pattern of the DPPA2 protein. To this end, we successfully raised a rabbit polyclonal antibody against a 15-amino acid peptide from the DPPA2 C terminus (RSRAKKNALPPNMPP), a region that is not present in DPPA4. The DPPA2 antibody specifically recognized a band of 39 kDa in ESC extracts by Western blotting, corresponding to the expected size of mouse DPPA2 (Fig. 4A). Analysis of EG cells and two EC cell lines revealed the presence of DPPA2 protein in EG and OKO EC cells, but surprisingly not in the P19 EC-cell extract (Fig. 4A). We recently observed a similar discrepancy between the human GCT27X1 EC line, which expresses Esg1, and the N-Tera2 cell line, which does not. These molecular differences may reflect important differences between types of embryonal carcinomas; for example, OKO cells are strongly pluripotent (Martin Pera, personal communication), whereas P19 cells exhibit more restricted multipotent characteristics [26]. Clearly, in future studies, it will be important to examine the functional significance of these differences using experiments.

Figure Figure 4..

Expression of DPPA2 protein was analyzed using a DPPA2-specific antibody (α-DPPA2). (A): By Western blot analysis, DPPA2 expression was detected in KES1 ES, 19G EG and OKO EC cells but not in P19 EC cells or in E13.5 PEFs. (B): Indirect immunofluorescent staining of preimplantation mouse embryos using α-DPPA2. Scale bar = 20 μm. (C): Indirect immunofluorescent staining of outgrowing inner cell mass/blastocyst cultures using α-DPPA2 and an OCT4-specific antibody (α-OCT4). Note the cytoplasmic staining of α-DPPA2 in the cells (arrows) immediately outside the perimeter of the α-OCT4-positive cell cluster. This DPPA2 staining pattern contrasts with the nuclear DPPA2 staining observed in OCT4-positive cells. Scale bar = 10 μm. For (B) and (C), α-DPPA2 was detected in the red channel (Alexa 594), α-OCT4 in green (Alexa 488), and DNA (Toto3) in the far red (blue) channel. (D): Sections of E13.5 developing testis showing DPPA2 expression in OCT4-positive germ cells. Scale bars = 25 μM (top) and 10 μM (bottom). (E): Sections of adult testis showing GCNA staining in spermatogonia. Despite the presence of Dppa2 mRNA (Fig. 1D), DPPA2 protein expression was not detected in adult testis. Scalebars = 100 μM (top) and 20 μM (bottom). For (D) and (E), α-DPPA2 was detected in the red channel (Alexa 594) and α-OCT4 in green (Alexa 488), and DNA was stained with DAPI. All images in (B–D) are single optical sections generated by scanning confocal microscopy. Abbreviations: DAPI, 4,6-diamidino-2-phenylindole; EC, embryonal carcinoma; GCNA, germ cell nuclear antigen; PEF, primary embryonic fibroblast.

Using immunocytochemistry, DPPA2 expression was examined in preimplantation embryos. Nuclear localization of the protein was detected from the two-cell stage onwards, with some cytoplasmic staining detectable until the four-cell stage. At all subsequent stages examined, DPPA2 appeared to colocalize with DNA. Expression of DPPA2 in the early blastocyst was detected in the nuclei of trophectoderm cells, albeit at a lower level than in ICM nuclei. At late blastocyst stage, DPPA2 expression had disappeared from the TE to become restricted to the ICM (Fig. 4B). Moreover, during derivation of ESCs via the in vitro culture of blastocysts, nuclear DPPA2 staining was detected specifically in OCT4-positive cells of the ICM outgrowth (Fig. 4C). Interestingly, cells in the region proximal to the OCT4-positive population were positive for DPPA2 protein in the cytoplasm but negative in the nucleus, suggesting that DPPA2 is exported from the nucleus prior to its degradation in the differentiating OCT4-negative cells (Fig. 4C). Consistent with this, DPPA2 protein was detected in the nucleus of OCT4-positive fetal germ cells (Fig. 4D). These observations corroborate the results of the WISH experiments and suggest that DPPA2 functions specifically in the nucleus of pluripotent cells and the developing germ line. Finally, DPPA2 protein was not detectable in adult testis samples despite strongly detectable GCNA staining (Fig. 4E). This is in contrast to the expression of Dppa2 RNA in these samples (Fig. 1D), indicating that Dppa2 message can be present in the absence of detectable protein, a result we also found for P19 EC cells (RT-PCR not shown; Fig. 4A).

Dppa2 and Dppa4 Are Expressed by Human Pluripotent Cells

The expression of hDPPA2 and hDPPA4 in human ESCs and EC cells was examined and compared with hOCT4 using RT-PCR (Fig. 5). After 30 PCR cycles, hDPPA2, hDPPA4, and hOCT4 could be strongly detected in both hES lines (HES-2 and HES-4) [15] and an hEC line (GCT27X1) [16]. These data confirm the presence of Dppa2 and Dppa4 expression in pluripotent cell types from two distantly related mammalian species.

Figure Figure 5..

Expression of hDPPA2, hDPPA4, and hOCT4 in human ES (HES-2 and HES-4), EC (GCT27X1), and embryonic kidney (293T) cells. Abbreviations: EC, embryonal carcinoma; ES, embryonic stem.


We have shown that the previously known but only partially characterized genes Dppa2 and Dppa4 are expressed only in pluripotent cells and the developing germ line and encode closely related SAP domain proteins. These data are strongly supportive of a role for Dppa2 and Dppa4 in maintaining cellular potency in the ICM, in pluripotent cell lines, and in the developing germ line, but elucidation of their possible function in this process lies in understanding the role of their SAP domains. This recently discovered 35 amino acid motif was named after three SAP-containing proteins, SAF-A/B, Acinus, and PIAS, and consists of a bipartite distribution of conserved hydrophobic, polar, and bulky residues separated by a region containing an invariant glycine [23]. In support of a role in pluripotency, Dppa2 and Dppa4 were originally identified as part of a gene set (which includes Oct4 and Esg1) aberrantly expressed in embryos generated by nuclear transfer [9]. Whether the misregulation of these genes is at the origin of the abnormal development of cloned embryos or a consequence of defective epigenetic mechanisms in these embryos remains to be investigated.

SAP domains have been identified in a variety of eukaryotic species and are thought to be involved in the modeling of nuclear architecture and in RNA metabolism. Interphase chromatin is arranged into topologically organized modules associated with distinct gene expression and replication activities. This level of epigenetic organization is due to chromatin regions known as scaffold and matrix attachment regions (S/MARs), which are located near the boundaries of actively transcribed genes. Interactions between S/MARs and the nuclear matrix affect the expression of adjacent genes. A role for SAP domain proteins in mediating these interactions, as well as RNA processing of nearby genes, has been postulated [27]. SAP proteins may organize the interphase chromosomes by binding to S/MARs while recruiting a set of factors responsible for elongation of transcription and pre-mRNA processing. Perhaps the best-characterized SAP proteins are the PIAS family members. PIAS 1, 3, X (α/β), and Y [28] modulate in a specific manner the transcriptional activities not only of STAT proteins but also of lymphoid enhancer factors, androgen receptor, c-Jun, p53, and SMADs [29], histone deacetylases, and Dnmt3a [30, 31]. Although the function of the SAP domain proteins DPPA2 and DPPA4 cannot be inferred from these examples, it seems likely that DPPA2 and DPPA4 operate through their SAP domains to regulate cellular plasticity in pluripotent cells and the germ line.

We observed that Dppa4 and Dppa2 are downregulated in female gonads as the germ cells enter meiosis. This may be significant for the following reasons. First, the downregulation of Dppa4 is a relatively early event, occurring before the coincident downregulation of Oct4 and upregulation of Scp3, suggesting that Dppa4 is an earlier marker of meiotic entry than either Oct4 or Scp3. Second, the downregulation of Dppa4 prior to Oct4 suggests that the downregulation of this gene is independent of the presence of OCT4 protein in germ cells. Finally, the similar A-P pattern of downregulation of all the pluripotent markers we have examined to date (Sox2, Esg1, Oct4, Dppa2, and Dppa4) suggests that there may be a global regulator that mediates the silencing of all these genes as the germ cells enter meiosis. This putative regulator may be a meiosis-inducing factor or may simply be required to silence the pluripotent markers, which may otherwise inhibit differentiation along the female germ cell developmental pathway. Two recent publications have shown that retinoic acid induction of Stimulated by retinoic acid gene 8 (Stra8) is required for meiotic entry in mice [32, 33]. A retinoic acid-induced protein, germ cell nuclear factor (GCNF) has also been shown to indirectly regulate pluripotency markers in vivo and in pluripotent stem cells [34, 35]. Therefore, it seems likely that retinoic acid (RA)/GCNF may also be required for the repression of pluripotency markers in premeiotic germ cells. Indeed, in support of this proposal, a study by Bowles et al. [33] observed a marked decrease in Oct4 expression in gonads treated with RA or a Cyp26b1 inhibitor (Cyp26b1 metabolizes RA).

The regulation of the pluripotent markers in the female germ line is of particular interest, as Sox2, Esg1, Nanog, SSEA1, Oct4 (J.v.d.B. and P.S.W., unpublished data), [7, 8] Dppa2, and Dppa4 are also downregulated in germ cells as they differentiate along the spermatogenic pathway. Therefore, this pattern is similar in both males and females as germ cells enter meiosis and mitotic arrest, respectively. Both these phases of germ line development involve progression from a relatively undifferentiated PGC to more differentiated (premeiotic and mitotically arrested, in males and females, respectively) germ cell types and exit from the cell cycle [36, [37], [38]–39]. Given that pluripotent markers inhibit differentiation in other cell types, it seems plausible that downregulation of some or all of these markers in the germ line is necessary to allow appropriate differentiation down the female and male pathways (possibly including effective cell cycle exit). Incomplete differentiation of the germ line may include inadequate establishment of necessary epigenetic silencing mechanisms during germ line differentiation, and this may eventually allow aberrant activation of pluripotent regulators at later postnatal stages. If this hypothesis is correct, it may have significant implications for the development of germ cell cancer in human patients, since many of these genes are aberrantly reactivated and overexpressed in embryonal carcinoma and seminoma, the undifferentiated stem cells of invasive germ cell-derived cancers [40]. Future work focusing on the normal role of the somatic environment and the regulation of pluripotent marker expression within the differentiating testis will be critical in understanding the reactivation and possible role of these genes in testicular cancer.

The characterization of Dppa2 and Dppa4 sequence and expression patterns identifies two potential chromatin regulators in the stem cell/germ cell pluripotency network. Control of chromatin architecture in the pluripotent nucleus is likely to be essential for cellular plasticity, and such control may be orchestrated not only by the combinatorial effects of histone-tail modifications, but also by the action of factors such as DPPA2 and DPPA4, via their SAP domains. Indeed, although the specific function for Dppa4 remains unclear, a recent study indicates that depletion of Dppa4 from ESCs leads to differentiation and abnormal stem cell maintenance [41]. Likewise, RNA metabolism and post-transcriptional modifications are likely to be as important as direct transcriptional regulation in maintaining broad developmental potency. Until now, studies on the molecular mechanisms of pluripotency have focused on the control of gene expression by transcription factors such as Oct4, Sox2, FoxD3, Stat3, and Nanog, but the molecular mechanisms through which cells proliferate in an undifferentiated state while maintaining a potential for differentiation are not yet fully understood. Further understanding of the mechanisms underlying pluripotency and germ line differentiation not only will enhance our ability to manipulate stem cells in vitro but is likely to lead to the elucidation of the mechanistic basis of germ cell carcinoma.


M.A.S. has acted as a consultant for CellCentric Ltd. within the past 2 years.


J.M.-S. and J.v.d.B. contributed equally to this work. We thank Martin Pera (Australian Stem Cell Centre, Monash University) for provision of the HES-2 and HES-4 human embryonic stem and GCT27X1 EC RNA. We also thank George Enders for the GCNA antibody and Craig Smith for helpful discussion and comments on the manuscript. This work was supported by the Australian Research Council through funding of the Centre for Biotechnology and Development (to A.H.S.) and by the Wellcome Trust through a Programme Grant to M.A.S.