NANOS2 interacts with the CCR4-NOT deadenylation complex and leads to suppression of specific RNAs

Abstract

Nanos is one of the evolutionarily conserved proteins implicated in germ cell development. We have previously shown that NANOS2 plays an important role in both the maintenance and sexual development of germ cells. However, the molecular mechanisms underlying these events have remained elusive. In our present study, we found that NANOS2 localizes to the P-bodies, known centers of RNA degradation that are abundantly accumulated in male gonocytes. We further identified by immunoprecipitation that the components of the CCR4-NOT deadenylation complex are NANOS2-interacting proteins and found that NANOS2 promotes the localization of CNOT proteins to P-bodies in vivo. We also elucidated that the NANOS2/CCR4-NOT complex has deadenylase activity in vitro, and that some of the RNAs implicated in meiosis interact with NANOS2 and are accumulated in its absence. Our current data thus indicate that the expression of these RNA molecules is normally suppressed via a NANOS2-mediated mechanism. We propose from our current findings that NANOS2-interacting RNAs may be recruited to P-bodies and degraded by the enzymes contained therein through NANOS2-mediated deadenylation.

In the mouse, the primordial germ cells (PGCs) are segregated from the somatic cell lineage at an early gastrulation stage (1). Although the PGCs are potent producers of both oogonia and spermatogonia, sexual differentiation is induced after their colonization of the embryonic gonads with somatic cells. However, the initial steps leading to diversification of these cells have long remained unsolved. Retinoic acid (RA) signaling has recently been identified as the initial trigger for feminization (2). RA molecules derived from the mesonephros trigger meiotic initiation in female gonocytes via the induction of the RA responsive gene Stra8, which is required for premeiotic replication (3). In contrast, male gonocytes are protected from exposure to RA by CYP26B1, an RA metabolizing enzyme produced from somatic cells, resulting in the suppression of meiosis up to E13.5 (4, 5). In addition, Nanos2 expression begins after E13.5 and is required for the maintenance and promotion of the male germ cell state (6).

Nanos is an evolutionarily conserved RNA-binding protein that is essential for germ cell development (7). In Drosophila, Nanos forms a complex with another RNA-binding protein, Pumilio, and represses the translation of the hunchback, cyclin B, and hid mRNAs thereby establishing embryonic polarity, mitotic quiescence, and suppression of apoptosis, respectively (8 –10). Three Nanos homologs, Nanos1–3, exist in the mouse, among which Nanos3 and Nanos2 are expressed in the germ cells and are required to protect these cells from undergoing apoptosis during migration and after colonization of the male gonads, respectively (11, 12). In addition, Nanos2 plays a key role during the sexual development of germ cells by suppressing meiosis and promoting male-type differentiation in the embryonic male gonads. Moreover, the forced expression of Nanos2 in female gonocytes can induce the suppression of meiosis and promotion of male-type gene expression (6). However, the molecular mechanisms underlying how this protein accomplishes such pleiotropic functions in the mouse germ cells remain unknown.

In our present study, we find that NANOS2 localizes to P-bodies, a central hub of RNA degradation (13, 14). We further identify components of the CCR4-NOT deadenylation complex as NANOS2-associated proteins in vivo, which can cleave poly(A) RNA in vitro. We also show that specific mRNAs interact with NANOS2, and thus propose that NANOS2 plays a role in recruiting the CCR4-NOT deadenylation complex to trigger the degradation of specific RNAs.

Results

NANOS2 Localizes at P-Bodies During Gonocyte Development.

To increase our understanding of the molecular mechanisms underlying the function of the NANOS2 protein, we first analyzed the cellular localization of this protein by immunostaining. Consistent with the results of our previous western analyses (15), NANOS2 protein was first detectable at E13.5 in the cytoplasm of male mouse gonocytes. This signal intensity increased until about E16.5 and then slightly decreased by E17.5. In addition, we found that some of the NANOS2 proteins formed discrete foci, the number of which gradually increased until E16.5 and then decreased thereafter (Fig. S1 A–F). Because Drosophila Vasa and Tudor are known to form cytoplasmic foci (16, 17), which are the polar granules in the germ plasm, we speculated that these NANOS2 foci might colocalize with the mouse homologs of Vasa, MVH (mouse vasa homolog) (18) and the Tudor protein TDRD1 (tudor domain containing 1) (19). However, these foci did not show any clear colocalization with NANOS2 (Fig. S2 A–F). We next tested the possibility that the NANOS2 foci might correspond to P-bodies, which are known to function as a center of RNA degradation. We thus conducted double-immunostaining using antibodies against the P-body components DCP2 and XRN1, an mRNA decapping enzyme and RNA exonuclease, respectively (13, 14). We were initially surprised to find that many P-bodies could be specifically observed only in germ cells and not in the somatic cells in E15.5 male gonads, and also that the NANOS2 foci clearly merged with those of DCP2 and XRN1 (Fig. 1 A–F) from E13.5 to E17.5 (Fig. S3 A–F). This suggests the possibility that NANOS2 may be involved in RNA degradation.

Fig. 1.

NANOS2 localizes to the P-bodies in male mouse gonocytes. (A–L) Sections prepared from wild-type E15.5 male gonads were double-stained with mouse anti-NANOS2 (green) (A and D) and either hDCP2 (B) or mXRN1 (E) antibodies (red staining in each case). Arrowheads indicate colocalization of NANOS2 and hDCP2 (C) or XRN1 (F). DNA was counterstained with DAPI (blue). (Scale bar in A, 20 μm for all panels.)

Nanos2 Functions in the Formation of P-Bodies.

We further examined the status of the P-bodies in the mouse gonads of both sexes by immunostaining of p54/RCK, a homolog of Drosophila Me31B and also a marker of these structures (20). Although the P-bodies seemed to be present in the same number and size in the gonocytes of both sexes at E12.5, they were gradually reduced and eventually lost by E14.5 in female gonocytes (Fig. S4 E and F). In contrast, the P-bodies become much larger in both number and size from E14.5, concomitant with the onset of NANOS2 expression, in male gonocytes (Fig. S4 A–D).

To further explore the role of NANOS2 in P-body formation, we examined the status of these structures in the absence of Nanos2. Although there were, somewhat unexpectedly, many P-bodies detected in both Nanos2 ^+/− and Nanos2 ^−/− male gonocytes at E13.5, their sizes became gradually larger, whereas their number became smaller, at the later stages of embryogenesis in the absence of Nanos2 (Fig. 2 A–D). This was also observed in Nanos2, Bax double-null male gonocytes (Fig. 2 E and F), where apoptotic cell death was suppressed, suggesting that apoptosis does not affect P-body status. This indicates that NANOS2 is not essential for the assembly of P-bodies but is required for the maintenance of their normal state. To further elucidate the functions of NANOS2 in P-body formation, we also examined the status of the P-bodies in NANOS2-expressing female gonocytes (6). Although they could not be detected in normal female gonocytes at E16.5, we found many P-bodies in NANOS2-expressing female cells and additionally observed that NANOS2 localizes at the P-bodies in these cells (Fig. 2 G–I). These data indicate that NANOS2 is sufficient to maintain the number of P-bodies when female gonocytes have acquired a male-type phenotype due to NANOS2 expression.

Fig. 2.

Functional role of NANOS2 during the formation of the P-bodies. (A–E) Male gonadal sections from Nanos2 ^+/− (A and C), Nanos2 ^−/− (B and D), and Nanos2 ^−/− Bax ^−/− (E) embryos at stages E13.5 (A and B), and E16.5 (C, D, and E) were immunostained with p54/RCK (green) and TRA98 (red) antibodies. (F) Average number of p54/RCK foci per male gonocyte at E16.5 was quantified in each picture using ImageJ software (National Institutes of Health) and a cell counter, with the foci of less than a 20 permission value excluded using Photoshop (Adobe). The data shown correspond to two to three pictures. (G–I) A female gonadal section from a NANOS2-expressing embryo at E16.5 was immunostained with anti-FLAG (green) (G) and anti-p54/RCK (red) (H) antibodies. DNA was counterstained using DAPI (blue). (Scale bar in A, 20μm for A–E and G–I.)

NANOS2 Interacts with the CCR4-NOT Deadenylation Complex and Regulates Its Localization.

To explore the molecular functions of NANOS2, we searched for proteins that interact with it. To this end, we prepared male gonadal extracts from Nanos2 ^+/− and Nanos2 ^−/− embryos at E14.5 and subjected them to immunoprecipitation with anti-NANOS2 antibodies. We found that two major bands of more than 200 kDa were exclusively precipitated from Nanos2 ^+/− gonads, and by mass spectrometric analysis identified these products as CNOT1, a component of the CCR4-NOT deadenylation complex (13) (Fig. 3A).

Fig. 3.

Interaction between NANOS2 and the CCR4-NOT deadenylation complex. (A) Proteins coimmunoprecipitated with NANOS2 from E14.5 wild-type (lane 1) and Nanos2 ^−/− (lane 2) male gonadal extracts using rabbit anti-NANOS2 antibodies. Arrowheads indicate CNOT1. *1, nonspecific band; *2, IgG polypeptide. (B) Immunoprecipitation–Western blot analyses of proteins from male gonadal extracts of wild-type and transgenic embryos expressing 3×FLAG-NANOS2. *3, IgG polypeptide from the anti-FLAG antibody. (C–E) Male gonadal sections from E15.5 embryos were immunostained with mouse NANOS2 (green) (C) and CNOT3 (red) (D) antibodies. Arrowheads in C–E indicate colocalization between NANOS2 and CNOT3. (F–K) Male gonadal sections from Nanos2 ^+/− (F–H) and Nanos2 ^−/− (I–K) embryos at E15.5 were immunostained with DCP1A (red) (G and J) and CNOT3 (green) (F and I) antibodies. DNA was labeled via DAPI counterstaining (blue). (L) Western blot analyses of proteins from the male gonads of Nanos2 ^+/− and Nanos2 ^−/− embryos at E15.5.

In further immunoprecipitation experiments, we used a transgenic mouse line expressing a FLAG-tagged NANOS2 under the direct control of the Nanos2 enhancer (15) (Fig. S5A), since we had confirmed that this fusion protein was functional (Fig. S5 B–F) and localized at the P-bodies (Fig. S5 G–I). Western analyses revealed that CNOT1 coprecipitates with FLAG-tagged NANOS2 (Fig. 3B , Upper), confirming the results of our mass spectrometric analysis. We also found that other components of the CCR4-NOT complex, CNOT3, CNOT6L/Ccr4b, CNOT7/Caf1a, and CNOT9/Rcd1 (13, 21), also coprecipitated with FLAG-tagged NANOS2, indicating that NANOS2 associates with the CCR4-NOT deadenylation complex in vivo. We additionally found that this interaction is independent of RNA, as the levels of coprecipitated CNOT proteins were not affected by treatments with RNase (Fig. 3B). Finally, these CNOT proteins were found to colocalize with NANOS2 in P-bodies (Fig. 3 C–E and Fig. S6 A–I), suggesting that this complex may play a role in the activities of these elements.

To better understand the physiological significance of its interaction with NANOS2, we investigated the localization of CCR4-NOT deadenylation complex in Nanos2 ^−/− male gonads by immunostaining CNOT proteins with DCP1A, another decapping enzyme and also a component of P-bodies (13, 14). Although CNOT3 was found to clearly localize to P-bodies in Nanos2 ^+/− male gonads (Fig. 3 F–H), we detected only weak signals for this protein in P-bodies in the absence of NANOS2 (Fig. 3 I–K) even though the levels of CNOT3 are not reduced in Nanos2 ^−/− male gonads (Fig. 3L). We obtained similar results for CNOT1 (Fig. S7). These data suggest that NANOS2 promotes the localization of the CCR4-NOT deadenylation complex to P-bodies, although a subpopulation of this complex still remains in these structures in the absence of NANOS2, possibly via a NANOS2-independent mechanism. Based on these findings and the fact that the CCR4-NOT deadenylation complex regulates the first step of mRNA degradation (22), we speculate that NANOS2 recruits this deadenylation complex to P-bodies where it promotes the degradation of RNAs.

Complex of NANOS2 and CCR4-NOT Deadenylation Complex in Male Germ Cells Retains Deadenylase Activity.

To address the critical question of whether NANOS2-interacting deadenylase actually has catalytic activity, we used NANOS2-overexpressing (NANOS2 O/E) adult testes to obtain sufficient amounts of this protein and thus overcome the limitations of using embryonic testis in biochemical analyses. In the testis of the postnatal mouse, NANOS2 is expressed in a small population of undifferentiated spermatogonia (23) and localizes to P-bodies (Fig. S8 A–C) as in the male gonocytes. This expression is subsequently lost as these cells differentiate. However, if FLAG-tagged NANOS2 is forcedly and continuously expressed in the spermatogonial population, the male mouse become infertile because the spermatogonia remain in an undifferentiated state in the testis, in which a large number of NANOS2-positive spermatogonia occupy the periphery of the seminiferous tubules (23). In addition, FLAG-tagged NANOS2 also localizes to the P-bodies in the spermatogonia in the manner similar to endogenous Nanos2 (Fig. S8 D–F). We prepared testis extracts from this mouse and performed immunoprecipitations with anti-FLAG antibodies and control IgG, and then subjected these immunoprecipitates to in vitro deadenylase assay (21) (Fig. 4A). As shown in Fig. 4B , cleavage of the poly(A) RNA substrate occurred only with NANOS2 immunoprecipitates, which also contains the CNOT6L and CNOT7 catalytic components of the deadenylation complex (Fig. 4C). These results lead us to propose that NANOS2 promotes the degradation of NANOS2-interacting mRNAs through the deadenylase activity of the CCR4-NOT complex.

Fig. 4.

The protein complex of NANOS2 and CCR4-NOT complex has in vitro deadenylase activity. (A) Schematic representation of the in vitro deadenylase assay method using NANOS2 over-expressing (O/E) testes. (B) FLAG-tagged NANOS2 was precipitated with anti-FLAG antibodies from the testis extracts of a 6-week-old NANOS2 O/E mouse and incubated with 5′-fluorescein isothiocyanate-labeled poly(A) RNA substrate for 0, 45, 90, and 180 min. Samples were then analyzed on a denaturing sequencing gel, as previously described (21) (G). (C) Western blot analyses reveaing that CNOT6L and CNOT7 are coprecipitated with FLAG-tagged NANOS2.

NANOS2 Interacts with Specific mRNAs and May Promote Their Degradation.

Based on our working hypothesis, we further speculated that (i) the NANOS2 complex should contain specific mRNAs that would be degraded via NANOS2-mediated deadenylation, such that (ii) the expression levels of these transcripts would be low in wild-type male gonocytes but up-regulated in the absence of NANOS2. To test these possibilities, RNAs that coprecipitated with FLAG-tagged NANOS2 were purified and subjected to RT-PCR. Because we had previously shown that male gonocytes could enter meiosis in the absence of NANOS2, it was plausible that mRNAs involved in meiosis might be directly suppressed through NANOS2-mediated RNA degradation. As expected, Sycp3, Stra8, Taf7l, Dazl, and Meisetz (3, 24 –27) transcripts that are implicated in meiosis were specifically detected only in the NANOS2 protein precipitates despite their very low expression in male gonads (Fig. 5 A and B). In contrast, the G3pdh, Dnmt3l and Dnmt3a mRNAs did not show specific accumulation in the NANOS2 precipitates although they are all highly expressed in male gonads. These data indicate that the mRNAs involved in meiosis specifically interact with NANOS2 in vivo.

Fig. 5.

NANOS2 interacts with specific mRNAs and may promote their degradation. (A) Male gonadal extracts from wild-type (wt) and transgenic (tg) mice expressing FLAG-NANOS2 at E15.5 were subjected to immunoprecipitation (IP) with FLAG antibodies. RNA precipitates were analyzed by semi-quantitative RT-PCR. (B) Quantification of each mRNA enrichment from a FLAG IP of tg extracts using real-time RT-PCR. Fold enrichment of each mRNA coprecipitated from tg compared with those from wt is indicated. Mean value of three independent QRT-PCR results is shown. (C–G) Expression profiling of the Sycp3 (C), Stra8 (D), Taf7l (E), Dazl (F), and Meisetz (G) genes in male gonads from Nanos2 ^+/− and Nanos2 ^−/− embryos at E13.5-E15.5 using the Affymetrix GeneChip System as previously described (43) (X-axis; embryonic stage, Y-axis; expression level, black bars; Nanos2 ^+/− embryos, white bars; Nanos2 ^−/− embryos).

We next investigated global changes in gene expression upon the loss of Nanos2 using comparative GeneChip analyses (Table S1). The resulting scatter plots showed that many genes become up- or down-regulated in Nanos2 ^−/− male gonads by E15.5 (Fig. S9 A–C). For example, we found that the genes highly expressed only in male gonocytes, such as Dnmt3l, Tdrd1 and Miwi2/Piwi-like 4 (19, 28, 29), are down-regulated in the Nanos2 ^−/− male gonads, whereas Figla, Lhx8 and Nobox, which have been shown to be essential only for oogenesis and not for spermatogenesis (30 –32), become accumulated in the Nanos2 ^−/− male gonads (6) (Fig. S9 D–I). These results suggest that male gonocytes cannot enter the male pathways and become feminized by the up-regulation of female-type genes. In addition, and consistent with the results of our immunoprecipitation assay, Sycp3, Stra8, Taf7l, Dazl, and Meisetzs mRNAs were also found to be up-regulated in Nanos2 ^−/− male gonads (Fig. 5 C–G). Our current findings thus indicate that NANOS2-interacting mRNAs become accumulated if NANOS2 is absent in male gonocytes, which in turn indicates that NANOS2 might be indirectly affecting the transcription of these genes, or that they are normally suppressed in wild-type male gonocytes through a NANOS2-directed mechanism, possibly a deadenylation pathway.

Discussion

Molecular Role of NANOS2.

In our current study, we show that the CCR4-NOT deadenylation complex is coprecipitated with NANOS2 from male gonadal extracts. This is the first evidence that the interaction between a Nanos homolog and the CCR4-NOT deadenylation complex exists in vivo, although it has been shown using a yeast two-hybrid system that Drosophila Nanos can directly and potently bind to NOT4, a component of the CCR4-NOT complex (33). Hence, as suggested previously by Kadyrova et al. for Drosophila Nanos, and as confirmed by our present analyses in vivo, the recruitment of the CCR4-NOT deadenylation complex to target mRNAs may be a conserved function of the Nanos proteins.

We also found that NANOS2 localizes to P-bodies in the male gonocytes and adult mouse spermatogonia. P-bodies are known to be a central hub of RNA degradation, in which decapping enzymes and exonucleases are also localized. However, emerging evidence in other systems suggests that P-bodies not only function to degrade RNAs but also to store mRNAs in a translationally quiescent state until needed (13). In addition, Drosophila Nanos promotes the deadenylation of poly(A) tail in hunchback mRNA and represses its translation without changing the mRNA level (34). We cannot therefore rule out the possibility that NANOS2 not only promotes the degradation of mRNAs involved in meiosis but also retains other transcripts at P-bodies to sequester them in a translationally inactive state during embryogenesis. These transcripts may be released from the P-bodies and translated to promote differentiation after birth as NANOS2 expression begins to disappear.

P-Body Formation in Male Mouse Gonocytes.

P-bodies have been well characterized in yeast and mammalian cultured cells, and the in vivo status of these foci has begun to be described recently also in worms and flies (35 –38). We found from our current analyses that P-bodies are specifically formed and/or maintained in the germ cells of male mouse embryonic gonads, whereas no such structures are detectable in somatic cells. Furthermore, female mouse gonocytes fail to maintain P-bodies at later stages of embryogenesis. We thus suggest that P-bodies play roles in cell-type specific differentiation during mouse development through RNA metabolism.

It has also been shown that P-bodies are dynamic structures and that their size and number reflects the status of the mRNA supply. If the transit of mRNAs into the P-bodies is inefficient, the size and number of these structures becomes extremely small. In contrast, they become larger when the mRNA decapping pathway is blocked (39, 40). Furthermore, it has been recently reported that deadenylation is required for P-body formation (41). Taking into account the data presented in these earlier reports and our current model, P-bodies would be expected to be small in Nanos2 ^−/− male gonocytes because the mRNA supply to these structures and subsequent deadenylation efficiency would be inhibited in the absence of NANOS2. However, we were surprised to find that the sizes of the P-bodies became larger in this biological context, although their number was decreased. These data thus indicate that male gonocytes have a unique program for P-body formation that occurs both in a NANOS2-dependent and -independent manner.

mRNAs Targeted by NANOS2.

We elucidated that the protein complex of NANOS2 and CCR4-NOT complex has deadenylase activity in vitro. We thus expected that the poly(A) tail lengths of NANOS2-interacting mRNAs would be maintained without NANOS2. To test this scenario, we assayed the poly(A) tail length of NANOS2-interacting mRNAs. However, we could not observe clear shortening of the poly(A) tail in wild-type male gonads, possibly because of their low abundance. New experimental systems will be required in the future to address this issue.

On the other hand, it was noteworthy that we identified Stra8 as a NANOS2-interacting mRNA because we have shown previously that Stra8 is up-regulated at the transcriptional level in Nanos2 ^−/− male gonocytes (6). These data together indicate that the suppression of Stra8 in male gonocytes is ensured at both the transcriptional and translational levels, suggesting the critical functional importance of suppressing this gene during male gonocyte development.

Materials and Methods

Mice.

Both the Nanos2 and Bax-knockout mouse lines and PCR methods used for the verification of each mutant allele have been previously described (11, 42). The NANOS2-expressing mouse line has also been described (23) The transgene containing 3×FLAG-tagged Nanos2 with the 3′-UTR under the control of Nanos2 enhancer (9.2 kb upstream sequence) was used for the production of the transgenic mouse line.

Histological Methods.

For immunostaining, mouse gonads of both sexes were directly embedded in O.C.T. compound (Sakura) and frozen in liquid nitrogen. After sectioning (8 μm), samples were stained according to standard procedures.

Immunoprecipitation.

Extracts of male gonads from E14.5 or E15.5 embryos were incubated with protein-A beads crosslinked with rabbit anti-NANOS2 antibody or anti-FLAG M2 affinity gel (Sigma).

In Vitro Deadenylase Assay.

The testis extracts from NANOS2-expressing mice were incubated with anti-FLAG M2 affinity gel or Mouse IgG-agarose (Sigma). After several washes, precipitates were then subjected to a deadenylase assay as previously described (21).

RT-PCR.

After synthesis of first-strand cDNAs with SuperScript III reverse transcriptase and (dT)₂₀ primer (Invitrogen), PCR analyses were carried out either using a regular or real-time protocol.

GeneChip Analysis.

Total RNAs were purified from cells corresponding to the male gonads of Nanos2-LacZ knock-in heterozygous and homozygous embryos, and analyzed using a GeneChip Mouse Genome 430 2.0 Array (Affymetrix).

Details of the methods and primer sequences used for each section are provided in SI Text.