Musashi interaction with poly(A) binding protein is required for activation of target mRNA translation

The Musashi family of mRNA translational regulators control both physiological and pathological stem cell self-renewal primarily by repressing target mRNAs that promote differentiation. In response to differentiation cues, Musashi can switch from a repressor to an activator of target mRNA translation. However, the molecular events that distinguish Musashi-mediated translational activation from repression are not understood. We have previously reported that Musashi function is required for the maturation of Xenopus oocytes, and specifically for translational activation of specific dormant maternal mRNAs. Here, we employed mass spectrometry to identify cellular factors necessary for Musashi-dependent mRNA translational activation. We report that Musashi1 needs to associate with the embryonic poly(A) binding protein (ePABP) reveals PABPC1 is present in immnoprecipitated Musashi1 complexes from both proliferating and from differentiated hiPSCs. No statistically significant difference in Musashi1:PABPC1 co-association was observed between proliferating and differentiated hiPSCs.

of this competition, it was suggested that the Musashi1:PABPC1 interaction prevented recruitment of the large 60S ribosomal subunit and subsequent cap-dependent translation of Musashi1 target mRNAs (27).
Our group was the first to report a critical role for Musashi in promoting activation, rather than repression, of target mRNA translation (28)(29)(30). During the transition of oocytes to fertilizable eggs in the frog, Xenopus laevis, gene transcription is suppressed and all new proteins necessary for oocyte maturation are translated from pre-existing maternal mRNAs (31,32). Translational activation of these dormant mRNAs involves release of repression and occurs in a sequential manner specified by three evolutionarily conserved mRNA sequence-specific translational regulators: Pumilio; Musashi; and the cytoplasmic polyadenylation element binding (CPEB1) (33,34). In response to progesterone-triggered maturation of immature oocytes, a signal transduction pathway initiates derepression of the Pumilio target mRNA, Ringo (35). Newly synthesized Ringo protein stimulates cyclin-dependent kinase (CDK) activity (36)(37)(38) which, in turn, initiates phosphorylation of Musashi and translational activation of Musashi target mRNAs, such as the mRNA encoding the MAP kinase kinase kinase, Mos (39). Translation of Mos and subsequent activation of MAP kinase signaling mediates a positive feedback amplification loop that drives robust Musashi activation, as well as activation of CPEBdependent late class mRNA translation and the all-or-none transition to commit to complete oocyte maturation (29,30,34).
Musashi-dependent mRNA translational activation has subsequently been reported in multiple mammalian systems (for example (6,24,(40)(41)(42)). The Musashi proteins are thus bifunctional and can switch from mediating repression to promoting activation of target mRNAs in response to appropriate extracellular stimuli (6,39,43,44). We have demonstrated that the switch in Musashi function is dependent upon the phosphorylation of two conserved serine residues in the C-terminal of the protein (39,43,44). The exact mechanism by which phosphorylation facilitates de-repression and promotes Musashi's ability to drive translation is unclear, but may involve altered interaction with, or altered function of, critical co-factors including PABPC1 (27), the microRNA-binding protein Lin28 which modulates let-7 miRNA biogenesis (45), or the poly (A) polymerase Germline Development 2 (GLD2) (46).
In this study, we employed mass spectrometry to identify Musashi-specific interacting proteins that are required for Musashi-dependent translational control.
We report identification and characterization of several members of the poly(A) binding protein family, namely the embryonic poly(A) binding protein (ePABP), which is functionally required for oocyte maturation and is the most abundant poly (A) binding protein in immature oocytes (47,48); PABP4, a newly identified member of the poly(A) binding protein family that contributes uniquely to aspects of early Xenopus embryo development (49); and PABPC1, although PABPC1 interaction was restricted to maturing oocytes. Contrary to the model in which interaction with poly (A) binding protein mediates repression of Musashi target mRNAs, we report that interaction with either ePABP or PABPC1 is necessary to promote Musashi target mRNA translational activation. Interestingly, we observed that Musashi is also associated with PABPC1 in a mammalian cancer cell line and in human stem cells under conditions of Musashi target mRNA translational activation and inhibited stem cell self-renewal. Together, our results indicate that the context-dependent interactions of Musashi with poly (A) binding proteins confer a differential ability to control target mRNA translation and cell fate.

Results
The Musashi interactome is dynamically regulated in response to progesterone stimulation. To better understand the molecular components that facilitate Musashi-dependent mRNA translational control, we sought to identify interacting partner proteins in both immature oocytes (where target mRNAs are repressed) and progesterone-stimulated oocytes (where target mRNAs are translationally activated). In three separate experiments, we microinjected immature Xenopus oocytes with RNA encoding either a GST-tagged form of Musashi1 to facilitate recovery of co-associated proteins after partial purification over glutathione sepharose or the GST moiety alone. Oocytes were incubated overnight to allow expression of the ectopic proteins, and then the GST-Musashi1 injected oocytes or control GST-injected oocytes were each split into two pools, one of which was left untreated (immature) and the other stimulated with progesterone. Following liquid chromatography-coupled tandem mass spectrometry, the spectral counts from identified proteins were normalized for protein size, giving us a Normalized Spectral Abundance Factor (NSAF) (50) and transformed to log2 with which to compare the relative abundance of each identified protein interacting with either GST-Musashi1 or GST (Supplemental Table 1).
Fifty proteins specifically interacted with the GST-Musashi1 protein but not the GST moiety alone ( Figure 1A). Of these, 40 were found to associate with Musashi1 in immature oocytes (Supplemental Table 2) while 29 associated with Musashi1 in progesterone stimulated oocytes (Supplemental Table 3). A comparison of the two lists revealed that 19 proteins remain associated with Musashi1 under either experimental condition (and represent a "core" interactome), 21 proteins preferentially associate with Musashi1 only in immature oocytes whereas 10 preferentially associate only in maturing oocytes (Supplemental Table 4). These observations suggest that the Musashi1-mRNA ribonucleoprotein (mRNP) complexes undergo dynamic remodeling in response to progesterone stimulation when Musashi target mRNAs are de-repressed and transition to a state of translational activation. Several known Musashi1 interacting proteins were detected in our analyses, including the cytoplasmic polyadenylation element binding protein (CPEB1) (51) in immature oocytes and PABPC1 (27) in maturing oocytes. The recovery of PABPC1 only in maturing oocytes likely reflects its reciprocal expression to ePABP, with PABPC1 being present at only low levels in immature oocytes and gradually increasing during maturation and early embryogenesis (48). Interestingly, we detected Musashi1 association with two additional members of the poly(A) binding protein family ( Figure 1A, the embryonic poly(A) binding protein, ePABP (also called PABPC1L) and PABP4 (also called inducible PABP or PABPC4)) in both immature and maturing oocytes. Gene set enrichment analysis revealed that the 50 identified coassociated proteins are primarily involved in RNA binding and translational regulation ( Figure 1B).

Musashi associates with the embryonic poly(A) binding protein (ePABP) in immature and maturing oocytes.
The persistence of ePABP and PABP4 interactions in Musashi mRNP complexes in maturing oocytes and indeed the progesterone-dependent recruitment of PABPC1 was unexpected since Musashi target mRNAs are translationally activated, not repressed, in maturing oocytes. Here, we have sought to determine the role and contribution of ePABP and PABPC1 to Musashi-dependent mRNA translational activation and oocyte maturation.
To confirm the interaction of ePABP with Musashi1 seen by mass spectrometry, immature oocytes were injected with RNA encoding GST-tagged Xenopus Musashi1 or the GST moiety alone and the ability of Musashi1 to associate specifically with endogenous ePABP was determined after partial purification over glutathione sepharose and western blotting. To ensure that any associations observed were due to protein:protein interactions rather than just co-occupancy of the same mRNA, all glutathione pulldown experiments included an RNase1 treatment step. Endogenous ePABP was found to co-associate in an RNA-independent manner with Musashi1 in both immature and progesterone-treated oocytes ( Figure 2A). When compared to levels of ePABP in the input lysates, less than 2% of the cellular ePABP was found in association with Musashi1. No co-association was observed with the GST moiety alone, indicating a specific interaction between Musashi1 and ePABP, verifying our mass spectrometry findings. Quantitation of ePABP recovery (normalized for the amount of GST-Musashi1 in each pulldown) revealed a 2-fold increase in ePABP association with Musashi1 after progesterone stimulation ( Figure 2B). This increase was observed in progesteronestimulated oocytes that had not yet completed germinal vesicle (nuclear) breakdown (GVBD (-)), as well as oocytes which had completed GVBD (+). We conclude that increased association of ePABP with Musashi1 is an early event during progesterone-stimulated oocyte maturation, occurring prior to GVBD.
Knockdown of ePABP attenuates Musashi-dependent MAP kinase signaling and blocks oocyte maturation. While it has been proposed that Musashi interaction with PABPC1 correlated with target mRNA repression in a mammalian cell line (27), Musashi has a unique role during the maturation of Xenopus oocytes where it mediates the translational activation of early class mRNAs prior to oocyte GVBD (29,30,44,52,53).
We therefore hypothesized that the interaction of ePABP with Musashi1 may be necessary to promote Musashi target mRNA translation during oocyte maturation. To directly assess the functional role and interdependence of the Musashi1:ePABP interaction, we first injected immature oocytes with Musashi1 and Musashi2 antisense oligonucleotides (Msi AS) which abolish progesteronedependent maturation (29,39,44) and then sought to rescue the deficit with either ectopic Musashi1 or ePABP expression. While ectopic Musashi1 was able to efficiently rescue maturation, ePABP expression was unable to effect any rescue, even at later time points ( Figure 3A). In a reciprocal experiment, immature oocytes were injected with antisense oligonucleotides that block ePABP translation (49) and the consequences on progesterone-stimulated maturation assessed.
Microinjection of the ePABP antisense oligonucleotides ablated progesterone-stimulated maturation( Figure 3B), a phenotype indistinguishable from Musashi antisense oligonucleotide injection ( Figure 3A). The block to maturation appeared to be specific, as it could be efficiently rescued by injection of ePABP mRNA containing nucleotide wobble (49) to prevent antisense targeting but maintain correct ePABP amino acid sequence integrity ( Figure 3B) with >90% rescue observed at later time points. By contrast, ectopic expression of Musashi1 was unable to effect any rescue of ePABP AS injected oocytes.
Taken together, these results suggest an interdependence of Musashi1 and ePABP for progesteronestimulated oocyte maturation.
Using protein lysates prepared from oocytes shown in Figure 3B, the disruption of the oocyte maturation signaling cascade following ePABP antisense injection was analyzed by western blotting ( Figure 3C). While having little effect on the basal levels of ePABP in immature oocytes, ePABP antisense oligonucleotides completely prevented the progesterone-stimulated increase in ePABP levels in maturing oocytes ( Figure 3C, upper panel). As a control for specificity, scrambled antisense oligonucleotides did not prevent the ePABP increase after progesterone stimulation. Ringo, an atypical CDK partner that is required for initial Musashi1 phosphorylation and activation (39), was translated normally despite ePABP antisense injection ( Figure 3C, Ringo). Consistent with Ringo protein translation and CDK activation, a low level of initial Musashi1 activating phosphorylation was observed in ePABP antisense-injected oocytes ( Figure 3C, pMsi1). However, the robust phosphorylation of Musashi1 seen in control (scrambled antisense injected) oocytes was blocked by ePABP antisense injection. Furthermore, downstream MAP kinase activation ( Figure 3C, pMAPK) was severely attenuated after ePABP antisense injection, suggesting a block of early class Mos mRNA translation, the primary activator of MAP kinase signaling in the oocyte and a known Musashi1 target mRNA. We thus position the effect of the ePABP antisense oligonucleotide block as acting immediately downstream of initial Musashi1 activation ( Figure 3D).
The observed signaling defects in ePABP antisense injected oocytes suggested that ePABP is required during Musashidirected early class mRNA translation, particularly for the early Mos mRNA. To test this hypothesis directly, we analyzed the polyadenylation status of the Mos and Nrp1b (which encodes Xenopus Musashi1) mRNAs, two established Musashi targets (28,30,53). Treatment with ePABP antisense attenuated progesterone-stimulated polyadenylation of both mRNAs compared to control oocytes ( Figure 4A-C, indicated by reduced PCR product size). This finding indicated that ePABP plays a role in polyadenylation of Musashi-dependent mRNAs, possibly through protection of the newly added poly(A) tail or modulation of the rate of poly(A) addition. We note that the extent of inhibition of polyadenylation was not as great as that seen in Musashi antisense oligonucleotide-injected oocytes. We have previously demonstrated that late class, CPE-dependent mRNA translational activation occurs at GVBD and requires prior activation of Musashi-dependent mRNA translation (29,30,52,53). The fact that we do not see CPE-dependent cyclin A1 polyadenylation in either Musashi or ePABP antisense injected oocytes ( Figure 4A and D) can thus be explained by the oocytes failing to properly engage the early class (Musashi-dependent) mRNA activation that is required to allow subsequent activation of late class mRNAs and progression to GVBD (54,55). Taken together, our results indicate that ePABP is necessary for robust Musashi-dependent mRNA polyadenylation and translational activation.
Deletion mapping reveals two distinct interaction sites for ePABP within the Musashi1 protein.
Having established a coincident requirement for both ePABP and Musashi for early mRNA polyadenylation and translational activation, we next sought to map the domain(s) within Musashi1 necessary for ePABP interaction. A series of Xenopus Musashi1 (XMsi1) deletion mutant constructs were generated and tested for interaction with endogenous ePABP ( Figure 5A). Two interaction domains were identified in Musashi1. The first encompasses amino acids 190-240, the previously characterized domain necessary for PABPC1 interaction (27). As can be seen in Figure 5C (left panel), expression of this domain alone (XMsi 190-240) was sufficient to retain significant interaction with endogenous ePABP, while further truncations within this region either dramatically attenuated ePABP interaction (XMsi1 190-230 and 190-220) or ablated interaction completely (XMsi1 210-240). A second, weaker binding site was identified within the N-terminal of Musashi1 spanning the end of the first RNA recognition motif (RRM1) and the start of RRM2 (amino acids 80-120). A comparison of these two interaction domains within Musashi1 revealed no obvious linear amino acid sequence homology.
We observed that while murine Musashi1 was also able to associate with ePABP to a similar degree as observed with Xenopus Musashi1 ( Figure 5B, mMsi1 WT), both the Xenopus Musashi2 ( Figure 7A and B) and murine Musashi2 proteins ( Figure 5C) exhibited a more than a 4-fold reduced interaction with ePABP. The N-terminal 80-120 amino acid domain is well conserved between Xenopus Musashi1 and Musashi2 (95% amino acid identity), between Xenopus Musashi1 and murine Musashi1 (100% identity) as well as between Xenopus Musashi2 and murine Musashi2 (97.5% identity). However, there is considerably more divergence in the C-terminal 190-240 domain between Xenopus Musashi1 and Xenopus Musashi2 (64% identity) as well as between murine Musashi1 and Musashi2 (64% identity). We propose that the reduced interaction of ePABP with Musashi2 when compared to Musashi1 is primarily due to isoform-specific sequence divergence within the C-terminal 190-240 interaction domain.
In a reciprocal set of experiments, we sought to map the Musashi1 binding sites within ePABP using yeast twohybrid analyses. Vertebrate PABPs consist of four RRMs and a C-terminal region containing a variable linker region and a highly conserved PABC domain ( Figure 6A). As mammalian Musashi1 was reported to repress translation by binding to PABPC1 RRMs 1-2 and disrupting the interaction of PABPC1 with eIF4G (27), we initially focused on this domain. Two RNA-binding proteins, MS2 coat protein (MS2) and iron regulatory protein (IRP)-1 were utilized as negative controls for specificity, and eIF4G served as a positive control. Surprisingly, neither the N or C-terminal fragments of Musashi1 interacted with RRMs 1-2 of PABPC1 or ePABP, and although full-length Musashi1 appeared to interact, this was RNA-dependent as point mutations in the RNP motifs of PABPC1 RRMs 1-2 that disrupt RNA-binding abrogated this interaction ( Figure  6B). In contrast, the interaction between eIF4G and the RNA binding mutant form of PABPC1 RRMs 1-2 was maintained ( Figure 6B). PABPC1 or ePABP RRMs 3-4 also failed to interact with either the N or C-terminal regions of Musashi1 (data not shown). The PABP C-terminal region does not bind RNA but has been characterized as a protein interaction domain. We observed that both fulllength Musashi1 and the C-terminal fragment of Musashi1 showed robust interaction with the C-terminus of both ePABP and PABPC1, with weaker interactions being observed with the N-terminal region of Musashi1 ( Figure  6C). These yeast two hybrid mapping experiments are in general agreement with the GST pull-down analyses ( Figure 5A-C). We conclude that the C-terminal domains of ePABP and PABP1 interact directly with the 190-240 domain (and to a lesser extent the 80-120 domain) of Musashi1.

Differential ePABP interaction correlates with functional differences between Musashi1 and Musashi2.
Our interaction mapping of ePABP binding sites within Musashi1 and Musashi2 revealed that ePABP binding to Musashi2 was significantly reduced compared to Musashi1 (Figures 5, 7A and 7B). To assess if the differential interaction with ePABP reflected any functional differences between Musashi1 and Musashi2, we assayed the ability of Musashi1 and Musashi2 to rescue oocytes that had been previously injected with Musashi antisense oligonucleotides. Given sufficient time, Musashi2 was eventually able to rescue maturation to the same extent as Musashi1. However, when assessed at the time when Musashi1 had effected maturation of 50% of the oocytes, Musashi2 was compromised in rescue ability ( Figure 7C). This rescue deficit was a consequence of the reduced rate at which Musashi2 was able to mediate progesteronestimulated maturation since Musashi2 was expressed to similar levels as the Musashi1 protein ( Figure 7D). We reasoned that the stronger interaction of Musashi1 with ePABP was necessary for the rapid endogenous mRNA translational activation and consequently determined the rate of maturation. In theory, an enhancement of ePABP interaction with Musashi2 would allow Musashi2 to function at the same rate as Musashi1 in the antisense rescue assay. To test this idea we generated a chimeric Musashi2 protein where amino acids 190-240 of Musashi2 were substituted with amino acids 190-240 from Musashi1, creating a full length Musashi2 protein encoding the Musashi1 ePABP interaction domain. The resulting chimeric Musashi2 protein showed a robust gain of function in our antisense rescue assay, and was functionally indistinguishable from the wild-type Musashi1 protein ( Figure 7C, Chimera). We propose that the relative ePABP interaction strength may contribute to the difference between the functional capacities of Musashi1 and Musashi2 protein isoforms to exert timely translational activation during Xenopus oocyte maturation.
Our mass spectrometry also detected interaction of PABPC1 with Musashi1, although this was restricted to maturing oocytes ( Figure 1A). We thus sought to confirm this interaction and determine if PABPC1 also interacted differentially with Musashi1 and Musashi2. We coexpressed GST-tagged Xenopus PABPC1 with either GFPtagged Musashi1 or Musashi2 and assessed their ability to co-associate in an RNA-independent manner. Consistent with the ePABP interaction results, PABPC1 was found to preferentially interact with Musashi1, and to a lesser extent with Musashi2 ( Figure 8A and 8B). We next determined if the PABPC1 isoform could function to promote progesterone-dependent Musashi target mRNA translation and maturation of oocytes. We expressed GST-tagged Xenopus PABPC1 in ePABP antisense oligonucleotide injected oocytes and assessed the ability of PABPC1 to exert rescue of progesterone-dependent maturation. Importantly, PABPC1 was able to efficiently rescue progesterone-stimulated oocyte maturation ( Figure 8C) with >90% rescue observed at later time points. At least in the context of the maturing oocyte, PABPC1 and ePABP appear to be functionally redundant as ectopic PABPC1 rescued ePABP antisense-treated oocytes at the same rate and to the same extent as ectopically expressed ePABP. We conclude that PABPC1, like ePABP, associates with Musashi and facilitates Musashi-dependent mRNA translational activation.

PABPC1 is associated with Musashi in mammalian cells under conditions that promote Musashi target mRNA translation.
We next sought to determine if endogenous Musashi associates with PABPC1 in mammalian cells under conditions of Musashi target mRNA de-repression and translation.
We have previously reported that differentiation of primary embryonic rat neuronal stem/progenitor cells or human SH-SY5Y neuroblastoma cells results in the de-repression and translation of Musashi target mRNAs (6,43). When proliferating cells are switched to media to promote differentiation, mammalian Musashi is subject to activating phosphorylation on the sites conserved with the Xenopus Musashi1 protein, resulting in translation of target mRNAs (43,44). An assessment of Musashi1 activating phosphorylation thus serves as an indicator of derepression and translation of target mRNAs.
We utilized the human SH-SY5Y neuroblastoma cell line and a human induced pluripotent stem cell line (iPSCs) to assess the interaction of PABPC1 under proliferating conditions where Musashi directs repression or under conditions where Musashi1 is phosphorylated to allow derepression and translation of target mRNAs through differentiation of SH-SY5Y cells or iPSCs along a neural progenitor lineage.
Two antibodies were found to immunoprecipitate human Musashi1.
One antibody (AbCam 52865) efficiently immunoprecipitated endogenous Musashi1 and weakly retained association of PABPC1 whereas a second antibody (AbCam 114107) immunoprecipitated both Musashi1 and Musashi2 and more effectively preserved association of PABPC1 ( Figure 9A). We utilized the AbCam 114107 antibody for all subsequent immunoprecipitations. When SH-SH5Y cells were cultured to promote differentiation, the differentiating SH-SY5Y cells exhibited a significant increase in Musashi1 activating phosphorylation ( Figure 9B, pMsi1) without significant change in overall levels of Musashi1 ( Figure 9B, Msi1). Notably, over three independent experiments, no statistically significant change in the levels of Musashi:PABPC1 interaction were observed when Musashi1 was immunoprecipitated from either proliferating or from differentiating SH-SY5Y cells ( Figure 9C). Similarly, the AbCam 114107 antibody immunoprecipitated Musashi1 and co-associated PABPC1 from either proliferating human iPSCs or iPSCs differentiated to adherent neural progenitor cells ( Figure  9D). When the iPSCs were differentiated to neural progenitor cells, a significant increase in activating Musashi1 phosphorylation was observed ( Figure 9E). Over three independent experiments, immunoprecipitation of endogenous Musashi1 revealed no significant changes in the level of co-associated PABPC1 in proliferating iPSCs (PSC) or in differentiated cell populations (Figures 8F). We conclude that PABPC1 interaction persists with Musashi1 under conditions of target mRNA de-repression and translation. These findings are consistent with PABPC1 interacting with Musashi1 to facilitate translation of Musashi-target mRNAs in differentiating cell contexts.

Discussion
In this study, we demonstrate that Musashi1 interacts with several members of the poly(A) binding protein family, including ePABP and PABPC1, to promote translational activation of target mRNAs. As such our data necessitate a fundamental revision of the model that describes a repressive-only function for the Musashi interaction with poly(A) binding proteins. Our mapping data indicate that Musashi1 possesses two domains for interaction with ePABP, while in reciprocal experiments Musashi1 was shown to interact with the C-terminal domain of both ePABP and PABPC1. Our data further indicate that differential interaction of Musashi1 or Musashi2 with poly(A) binding proteins may underlie differences in their functional properties.
We have previously shown that site-specific phosphorylation is required to facilitate de-repression and promote Musashi target mRNA translation, both in the oocyte system and mammalian cells (39,43,44). In the oocyte, the initial "trigger" phosphorylation of these activating sites occurs via early Ringo/CDK-mediated phosphorylation (39).
Full activating Musashi1 phosphorylation was shown to require feedback phosphorylation by MAP kinase signaling, a downstream effector of the Musashi target mRNA encoding Mos, a MAP kinase signaling activator (39). We found that full activating phosphorylation of Musashi1 is blocked in ePABP antisense-injected oocytes ( Figure 3C), although synthesis of Ringo protein occurs normally in ePABP antisense-injected oocytes and the initial Ringo/CDKmediated "trigger" phosphorylation of Musashi1 (39) seems unperturbed.
The failure to achieve full Musashi1 activation may be explained by the compromised polyadenylation of the Mos mRNA ( Figure 4) and the consequent failure to attain robust activation of MAP kinase and the MAP kinase-dependent feedback amplification of Musashi1 phosphorylation ( Figure 3D) (39). We conclude that the loss of ePABP interaction with Musashi resulted in attenuation of translation of Musashi target mRNAs, including the mRNA encoding Mos, and that this compromised downstream MAP kinase signaling.
Our mass spectrometry data suggest that the Musashi interactome undergoes dynamic regulation in response to progesterone stimulation.
Of note, the differential enrichment of two proteins, CPEB1 and PABPC1, exemplify the remodeling of Musashi mRNP complexes. We have previously reported that the interaction of Musashi1 with CPEB1 is specific, RNase1 insensitive (i.e. it does not simply occur via mRNA co-occupancy) and occurs via an indirect association (51). Here, we see the interaction of Musashi1 with CPEB1 is restricted to immature oocytes. The CPEB1 population undergoes significant degradation in response to progesterone stimulation (56), likely explaining the lack of a significant Musashi1 co-association in maturing oocytes. A failure to detect PABPC1 in the immature oocyte data set while observing association in maturing oocytes likely reflects the low abundance of this protein isoform in the immature oocyte and the subsequent progesterone-dependent accumulation of PABPC1 in the maturing oocyte (48).
It is noteworthy that approximately one third (16/50) of the proteins detected as Musashi1 partner proteins in the oocyte are have been previously identified as partner proteins of the human Musashi2 protein (57). Moreover, over half (9/16) belong to the "core" Musashi1 interactome that is shared between immature and maturing oocytes (Supplemental Table 4). We speculate that these proteins may play a conserved role in controlling the assembly and/or function of Musashi mRNP complexes on target mRNAs.
Consistent with repression of Musashi target mRNAs in immature oocytes, a number of the co-associating proteins are known to contribute to translational repression ( Figure  1A), including DDX6, LSM14A/B, PAT1 and EIF4ENIF1 ((58-61)). While association with PAT1 decreases in maturing oocytes, the other proteins remain associated and it will be interesting in future studies to determine how their activity is modified within Musashi mRNP complexes to allow progesterone-dependent target mRNA translation.
The increased interaction (approximately 2-fold) of ePABP with Musashi1 after progesterone-stimulation ( Figure 2), also supports the idea that Musashi interactions with partner proteins are dynamically regulated in response to extracellular cues. The mechanism underlying differential ePABP co-association is unclear but several non-exclusive possibilities may be considered. First, the increased ePABP interaction may represent ePABP recruitment to a previously unbound Musashi1 after progesterone stimulation. We note that Musashi1 protein increases in response to progesterone potentially providing a reservoir of available targets for ePABP interaction (28). Second, the two sites of interaction on Musashi1 may allow engagement of a second ePABP to a Musashi1 protein already bound to ePABP. Third, the increase may represent multimerization of ePABP bound to Musashi1.
We report a strong interaction of ePABP with Musashi Cterminal amino acids 190-240 as well as weaker interaction with Musashi N-terminal amino acids 80-120 ( Figure 5). Interestingly, three proteins have now been demonstrated to interact with the C-terminal amino acids 190-240 region of Musashi1: PABPC1 (27); GLD2 (46); and ePABP (this study). GLD2 interacts strongly with amino acids 190-220 of the Musashi1 protein (46) while this same region interacted poorly with ePABP ( Figure 5C). By contrast, ePABP interacted strongly with amino acids 190-240, interacted weakly with amino acids 190-230 and failed to interact with amino acids 210-240, indicating that ePABP has a larger footprint of interaction across this region of the Musashi1 protein ( Figure 5C). Whether GLD2 and either ePABP or PABPC1 can interact simultaneously to the same Musashi1 protein remains to be determined. Although antisense knockdown experiments indicate that all three proteins promote Musashi-dependent translational activation, their relative contributions will require generation of point mutations that selectively attenuate their interactions within the 190-240 region of Musashi1. With regard to the poly(A) binding proteins, our data indicate an attenuation of Musashi target mRNA polyadenylation in ePABP-depleted oocytes (Fig. 4). These findings support a role for ePABP in augmenting the ability of GLD2 to promote progesterone-dependent polyadenylation and/or protecting newly extended poly(A) tails to facilitate Musashi target mRNA translation.
Employing yeast two hybrids assays, Musashi1 was shown to interact directly with the C-terminal domains of ePABP and PABPC1 ( Figure 6). Contrary to a previous report (27), no RNA-independent interaction of Musashi1 with RRMs 1 and 2 of either ePABP or PABPC1 was observed. One possible explanation for the discrepancy between these findings is that the prior study employed coimmunoprecipitation in the presence of RNase A (a pyrimidine-specific endoribonuclease) which may not have fully degraded mRNAs in the sample, thus allowing recovery of Musashi1 and PABP due to mRNA transcript co-occupancy rather than protein:protein interaction. Indeed, in the yeast two hybrid assay, an interaction with full length Musashi1 and RRMs 1 and 2 was seen, but this was abolished when an RNA binding mutant form of the PABPC1 RRM 1 and 2 domain was employed. Of note, the RNA binding mutant form of the PABPC1 RRM 1 and 2 domain was still able to interact with eIF4G ( Figure 6). Since Musashi1 interacts with the C-terminal domain of PABPs and not the eIF4G-interacting RRMs 1 and 2, our data do not support a direct competition between Musashi1 and eIF4G for the same site of interaction on PABPs as proposed (27). Nonetheless, it is possible that Musashi1 interaction with the C-terminal domain of PABPC1 or ePABP mediates an indirect attenuation of eIF4G binding through a conformation change.
Irrespective of the precise mechanism of repression, our findings add to the current understanding of Musashidependent translational control and extend the existing model in several ways. First, Musashi1 interaction with PABPs is not simply limited to exerting repression. Musashi1 remains associated with ePABP and PABPC1 under conditions of Musashi target mRNA de-repression and translational activation (Figs 1, 2, and 9). Thus the regulatory switch from mRNA repression to translational activation does not require dissociation of PABPs, but rather stimulus-dependent modification of Musashi activity in the presence of retained PABP association. Second, association of PABPs with Musashi is required to promote target mRNA translational activation in response to appropriate differentiation cues. Like GLD2, the other characterized co-factor activator (46), PABPs remain constitutively associated with Musashi1 irrespective of whether target mRNAs are repressed or translated. Thus, the activity of the regulatory complex must be controlled in a stimulus-dependent manner. While this switch in regulatory function involves site-specific phosphorylation of the Musashi1 and Musashi2 proteins, it may also require additional modification of co-associated factors, perhaps including the PABPs (62,63) to facilitate formation of a regulatory complex capable of stimulating de-repression and translation of Musashi target mRNAs. Musashi1 and Musashi2 have been shown to act redundantly in several physiological systems, including neural stem cell self-renewal and Xenopus oocyte maturation, as well as pathologically in propagation of colon cancer stem cell function (23, 30,44,64). However, Musashi1 and Musashi2 have been reported to have distinct functions in pancreatic cells (65) and a differential ability to support Zika virus replication (42). We show here that Musashi1 and Musashi2 have overlapping but non-identical functions in the oocyte, as Musashi2 functions less effectively than the Musashi1 isoform in promoting maturation. Consistent with the idea of non-identical functions, several reports have demonstrated isoformspecific interaction with partner proteins. Musashi1, but not Musashi2, can interact with Lin28 (45) suggesting that regulation of miRNA biosynthesis may be a unique property of Musashi1. Similarly, mammalian GLD2 can interact with mammalian Musashi1 but not mammalian Musashi2, although the functional consequence of this altered GLD2 interaction has yet to be established (46). To add to this list of isoform-specific differences, we show here that ePABP and PABPC1 have dramatically reduced interaction with Musashi2 compared to Musashi1 and that this reduced interaction correlates with diminished capacity to promote cell cycle progression during oocyte maturation. Indeed, domain swap experiments revealed that substitution of Musashi2 amino acids 190-240 for the corresponding Musashi1 sequence recapitulated full activity to the chimeric Musashi2 protein (Figure 7). Our data support a model where differential interactions with partner proteins underlie functional distinctions between the Musashi1 and Musashi2 isoforms.
In summary, we provide evidence for a dynamic interaction of Musashi with members of the poly(A) binding protein family and demonstrate that poly(A) binding protein interaction with Musashi is necessary to promote target mRNA translation, in addition to mediating repression. These findings extend earlier models of Musashi-dependent translational control and provide new insights into the regulation of co-associated factor interactions as Musashi switches from a repressor to an activator of translation. Our results also highlight important Musashi isoform specific difference in partner protein interactions that can modulate the ability of Musashi1 and Musashi2 to control the translational output of target mRNAs.

Experimental Procedures
Oocyte Culture and Microinjections. Dumont Stage VI immature Xenopus laevis oocytes were isolated and cultured as described previously (66). Oocytes were microinjected using a Nanoject II Auto-Nanoliter Injector (Drummond Scientific). mRNA for oocyte injection was made by linearization of the plasmid and in vitro transcription using SP6 (Promega) or T7 (Invitrogen) RNA polymerase as appropriate. Oocytes were stimulated to mature with 2µg/ml progesterone. The appearance of a white spot on the animal pole was used to score the rate of oocyte maturation as it indicates germinal vesicle (nuclear) breakdown (GVBD). Where indicated, progesteronestimulated oocytes were segregated when 50% of the oocytes completed GVBD (GVBD50) into those that had not (-) or had (+) completed GVBD. In the event of ambiguous morphology, oocytes were fixed for 10 minutes in ice cold 10% trichloroacetic acid and dissected for the presence or absence of a germinal vesicle. Animal protocols were approved by the UAMS Institutional Animal Care and Use committee, in accordance with Federal regulations.

Oocyte Lysis and Sample Preparation.
For co-association experiments, oocytes were lysed in 10µl/oocyte of ice cold NP40 lysis buffer (1% NP40, 20mM Tris pH 8.0, 137mM NaCl, 10% glycerol, 2mM EDTA, 50mM NaF, 10mM Sodium Pyrophosphate, 1mM PMSF, 1x Protease Inhibitor (Thermo Fisher Scientific)). Yolk and cell debris were removed by centrifugation at 12,000+ rpm for 10 minutes in a refrigerated tabletop centrifuge at 4˚C. For each sample, half oocyte equivalents of lysate were prepared in NuPAGE sample loading buffer and electrophoresed through a 10% NuPAGE gel (Invitrogen). Where required, a portion of the lysate was transferred immediately following lysis to STAT-60 (Tel-Test, Inc) for RNA extraction using the manufacture's protocol followed by a subsequent purification by precipitation in 4M LiCl at -80˚C for 30 minutes and centrifugation at 12,000rpm for 10 minutes in a refrigerated tabletop centrifuge.

Pulldown and RNase Treatment.
Oocytes were injected with 57.5 ng of each in vitro transcribed mRNA and incubated for 16h at 18˚C. Lysates were prepared as described above. 300µl of oocyte lysate was added to 450µl ice cold NP40 lysis buffer and incubated with 50µl of 50% glutathione sepharose conjugated bead slurry (GE) at 4˚C for 6h with gentle rotation. Beads were then gently pelleted by centrifugation at 500 rpm for 5 minutes; the supernatant was removed and replaced with 500µl fresh NP40 lysis buffer and the beads inverted to mix and the buffer then removed. This process was repeated 3 times at 4˚C. On the third wash, 200U of RNase1 (Ambion) was added and incubated for 15 minutes at 37˚C. Following final centrifugation, all NP40 was removed and 50µl of NuPAGE sample loading buffer (Invitrogen) was added. Beads were incubated for 10 minutes at 70˚C, then crushed by centrifugation at 12,000 rpm for 10 minutes. Finally, 45µl of the sample was loaded per each lane of a 10% NuPAGE gel (Invitrogen) and electrophoresed.

Mammalian Cell Culture and Immunoprecipitation of endogenous Musashi.
The SH-SY5Y human neuroblastoma cell line (ATCC) was cultured in DMEM/F12 medium with 10% fetal bovine serum (Thermo Fisher Scientific) and human induced pluripotent stem cells (iPSCs, #KYOU-DXR0109B, ATCC) were cultured in Essential 8 Medium (Thermo Fisher Scientific), on vitronectin-coated dishes (Thermo Fisher Scientific) with 5% CO2, as per the supplier's protocol. Neuronal differentiation was induced in SH-SY5Y cells through 1 week culture in retinoic acidcontaining medium [DMEM/F12 with 1x B27 with vitamin A supplement (Thermo Fisher), further supplemented with 1 µM retinoic acid (Sigma)]. Differentiation to neural stem cells was induced in human iPSCs through 1 week culture in PSC Neural Induction Medium on vitronectin-coated dishes (Thermo Fisher Scientific). Neural progenitor cells exhibited greater than 10-fold loss of pluripotency gene expression (Nanog) and gain of neural progenitor gene expression (sox1), as per the suppliers protocol (Thermo Fisher Scientific). Cell lysate was prepared after rinsing the cells twice with cold DPBS (Thermo Fisher Scientific) followed by lysis with NP-40 lysis buffer (250-500 µl per 10 cm dish) and clearing through centrifugation. Musashi1 or Musashi1/2 were immunoprecipitated from cleared cell lysate (100-250 mg) with 10µg antibody [Anti-Musashi1 #ab52865 (AbCam) or Anti-Musashi2 #ab114107 (which also immunoprecipitates Musashi1, AbCam)] or Normal Rabbit IgG (control, #12-370, Millipore) and 20µl Protein A Agarose (#ab193254, AbCam) with overnight incubation at 4 o C with rocking. Immuno-complexes were washed twice with lysis buffer and dissociated at 100 o C, 10 min, with NuPAGE gel loading buffer (Thermo Fisher Scientific) with added 10 mM DTT and separated by electrophoresis (NuPAGE, Thermo Fisher Scientific).

Western Blotting.
Following electrophoresis, NuPAGE gels were transferred to a 0.2µm-pore-size nitrocellulose membrane (Protran; Midwest Scientific). The membrane was blocked with 5% non-fat dried milk or 1% BSA in TBST (20mM Tris pH 7.5, 150mM NaCl, 0.05% Tween20) for 60 min at room temperature, or overnight at 4˚C. Following incubation with primary antibody, filters were washed 3x10 minutes in TBST, incubated with horseradish peroxidase-conjugated secondary antibody, or Protein A HRP-conjugate (#12291, Cell Signaling) then washed 3x10 minutes in TBST. Blots were developed using enhanced chemiluminescence in a Fluorchem 8000 Advanced Imager (Alpha Innotech Corp.). Western blots were quantified using Fluorchem FC2 software (Alpha Innotech Corp.)
Polyadenylation Assays. cDNAs for polyadenylation assays were synthesized using RNA ligation-coupled PCR as described (52). The increase in PCR product length is specifically due to extension of the poly [A] tail (28,52,53). The same reverse primer P1' was used for all reactions and has the sequence: 5'-GCTTCAGATCAAGGTGACCTTTTT-3'.
The Mos forward primer has the sequence: 5'-GCAAGGATATGAAAAAAAGATTTC-3'. The Nrp1b (Xenopus Musashi1) primer has the sequence 5'-CAATACTGCAATGTACAATGTACTGC-3'. The Cyclin A1 primer has the sequence: 5'-CATTGAACTGCTTCATTTTCCCAG-3'. Deducting the size of the mRNA specific 3' UTR sequence prior to the site of poly [A] addition and the size of the ligated P1 DNA oligonucleotide from the mode of the PCR product size in immature oocytes, we deduce the Mos mRNA has ~60 adenylate tail; the Nrp1b mRNA has ~60 adenylate tail while the cyclin A1 mRNA has ~30 adenylate tail in control, immature oocytes. For Mos and Nrp1b the poly [A] tails are extended by approximately 60 and 70 adenylate residues, respectively, after 3 hours of progesterone stimulation (see Figure 4B-D).

Antisense Oligodeoxynucleotide Injections and Rescue.
by guest on November 4, 2020 http://www.jbc.org/ Downloaded from Antisense oligodeoxynucleotide 5'-CGGCTCCGGTTGCATTCATGTTTG -3' was used to target endogenous ePABP mRNA (49). Antisense oligodeoxynucleotides targeting Xenopus Musashi1 and Musashi2 have been described previously (29). Control oocytes were injected with randomized oligonucleotide with the sequence 5'-TAGAGAAGATAATCGTCATCTTA-3' (36). A total of 100ng of antisense oligonucleotides was injected for each condition and oocytes were incubated at 18 o C for 16 hours. For Musashi rescue assays, Musashi1/2 antisense injected oocytes were subsequently injected with 23ng RNA encoding wild type Musashi1, Musashi2 or the chimeric Musashi2 protein that contained the Musashi1 ePABP binding domain. The oocytes were then incubated for 1 hour at room temperature to allow expression of the protein, then stimulated to mature with progesterone. For rescue of ePABP antisense injected oocytes, oocytes were co-injected with 57.5ng of wobble Xenopus ePABP mRNA, which is resistant to the antisense oligos (49) or Xenopus PABPC1 mRNA which is not targeted by the ePABP oligonucleotides. Following overnight incubation, oocytes were stimulated with progesterone and the extent of GVBD scored.

Liquid Chromatography Coupled Tandem Mass Spectrometry.
We used mass spectrometry to identify Musashi1associated protein complexes. 240 oocytes were each injected with 115ng mRNA encoding GST-XMsi1 or GST alone. We estimate by western blot analysis that this results in a 5-10 fold overexpression of GST-XMsi1 to the endogenous Musashi1 protein.
Following overnight incubation, the oocytes from each injection were split into 2 separate pools and either left untreated or stimulated with progesterone.
All oocytes were collected when progesterone stimulated samples had reached GVBD50 and lysed in ice cold Tween 20 buffer (0.1% Tween 20, 20mM Tris pH 8.0, 137mM NaCl, 10% glycerol, 2mM EDTA, 50mM NaF, 10mM Sodium Pyrophosphate, 1mM PMSF, 1x Protease Inhibitor (Thermo Fisher Scientific)). A glutathione pulldown for each condition was then performed. 600µl of oocyte lysate was added to 250µl of 50% glutathione sepharose conjugated bead slurry (GE) at 4˚C for 6h with gentle rotation. Beads were then gently pelleted by centrifugation at 500 rpm for 5 minutes; the supernatant was removed and replaced with 1ml fresh Tween 20 lysis buffer. This process was repeated 3 times at 4˚C and following the final centrifugation, all lysis buffer was removed and 50µl of LDS sample loading buffer (Invitrogen) was added and incubated overnight at 4˚C for 6h with gentle rotation. Beads were incubated for 10 minutes at 70˚C, then crushed by centrifugation at 12,000 rpm for 10 minutes. Finally, 45µl of the sample was loaded per each lane of a 10% NuPAGE gel (Invitrogen) and electrophoresed. Following electrophoresis, the gel was fixed by 5 minute incubation with 10% acetic acid and 16% methanol at room temperature. Following incubation, fixative was replaced with deionized water and microwaved for 45s. The water was poured off and replaced with GelCode Blue (Pierce) and microwaved again for 30s to begin staining. The gel was then gently rocked for 30 minutes at room temperature. Finally, the GelCode Blue was replaced with 1% acetic acid and gently rocked for 1 hour. Protein gel bands were then excised and subjected to in-gel trypsin digestion. Gel slices were destained in 50% methanol (Thermo Fisher Scientific), 100mM ammonium bicarbonate (Sigma-Aldrich), followed by reduction in 10mM Tris [2-carboxyethyl]phosphine (Pierce) and alkylation in 50 mM iodoacetamide (Sigma-Aldrich). Gel slices were then dehydrated in acetonitrile (Fisher), followed by addition of 100 ng porcine sequencing grade modified trypsin (Promega) in 100 mM ammonium bicarbonate (Sigma-Aldrich) and incubation at 37oC for 12-16 hours. Peptide products were then acidified in 0.1% formic acid (Pierce). Tryptic peptides were separated by reverse phase Jupiter Proteo resin (Phenomenex) on a 100 x 0.075 mm column using a nanoAcquity UPLC system (Waters). Peptides were eluted using a 40 min gradient from 97:3 to 35:65 buffer A:B ratio. [Buffer A = 0.1% formic acid, 0.5% acetonitrile; buffer B = 0.1% formic acid, 75% acetonitrile.] Eluted peptides were ionized by electrospray (1.9 kV) followed by MS/MS analysis using collision-induced dissociation on an LTQ Orbitrap Velos mass spectrometer (Thermo). MS data were acquired using the FTMS analyzer in profile mode at a resolution of 60,000 over a range of 375 to 1500 m/z. MS/MS data were acquired for the top 15 peaks from each MS scan using the ion trap analyzer in centroid mode and normal mass range with a normalized collision energy of 35.0. Proteins were identified by searching against the UniprotKB database restricted to Homo sapiens (177,579 entries) using MaxQuant (version 1.5.3.8, Max Planck Institute) with a parent ion tolerance of 3 ppm and a fragment ion tolerance of 0.5 Da, fixed modification of carbamidomethyl on C, and variable modifications including oxidation on M, and Acetyl on peptide N-terminus.
Scaffold (Proteome Software) was used to verify MS/MS based peptide and protein identifications.
Peptide identifications were accepted if they could be established with less than 1.0% false discovery by the Scaffold Local FDR algorithm and contained at least 2 identified peptides. Protein probabilities were assigned by the Protein Prophet algorithm (68). The pulldown and MS experiments were repeated in triplicate and a p-value representing the t-test significance of an interaction with GST-Musashi1 versus the GST moiety alone after correction of assigned spectral hits for molecular weight of the target protein (log2 normalized spectral abundance factor, (50)) was derived for each identified protein. Fifty proteins were found to specifically interact with Musashi1 (Supplementary Table  1). The mass spectrometry proteomics data have been deposited to the ProteomeXchange Consortium via the PRIDE partner repository (69) with the dataset identifier PXD013585 and 10.6019/PXD013585. Gene set enrichment analysis (70) was used to query the molecular signatures database (MSigDB) for the top ten gene ontology molecular functions.
Yeast two hybrid assays (Y2H) Y2H analysis was performed in strain L40 as described in (71), with the strength of interaction being scored by the extent of blue color apparent within a fixed time.

Statistical Analysis.
All quantitated data are presented as the mean +/-S.E. Statistical significance was assessed by one way Analysis of Variance followed by the Bonferroni post hoc test or by Student's t-test when only two groups were compared. A probability of p < 0.05 was adopted for statistical significance.

Conflict of Interest
The authors declare that they have no conflicts of interest with the contents of this article. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.

Author Contributions
AMM conceived and designed the study and wrote the paper. MCM designed, performed and analyzed experiments in Figure 9A-F, contributed to the design of the study and wrote the paper. SDB, SGM and AJT designed, performed and analyzed the experiments shown in Figure 1. WAR and NKG designed, performed and analyzed the experiments shown in Figure 6, NKG also wrote the paper. CEC designed, performed and analyzed the experiments shown in Figures 2, 3, 4, 5, 7 and 8. LLH provided technical assistance and contributed to preparation of figures. GVC contributed to the design of the study and wrote the paper. All authors reviewed the results and approved the final version of the manuscript. The technique, strategy and where appropriate primer sequences used to generate the plasmid constructs utilized in this study are indicated. PCR fragment generation: Template was subjected to PCR amplification using the indicated primers. Resulting PCR fragments were then purified using agarose gel electrophoresis followed by clean up using a QIAquick Gel Extraction Kit (Qiagen). Next, the fragments and destination vector were digested using the indicated restriction enzymes and again purified and cleaned up using gel electrophoresis and the QIAquick kit. The fragment and vector were then ligated using the T7 Quick Ligase (New England Biolabs). Finally, the ligated fragment/vector was used to transform competent DH5-alpha E. coli. PCR-directed mutagenic deletion: Template was subjected to PCR amplification of the entire plasmid. Primer sequence "looped out" the desired sequence for deletion. Restriction fragment subcloning: Template DNA and destination vector were separately digested with the indicated restriction enzymes and the desired fragments isolated and purified using agarose gel electrophoresis followed by clean up using a QIAquick Gel Extraction Kit (Qiagen). The fragment and vector were then ligated using the T7 Quick Ligase (New England Biolabs). Finally, the ligated fragment/vector was used to transform competent DH5-alpha E. coli.  (A) Oocytes were injected with mRNA encoding GST-XMsi1 or the GST moiety alone. The injected oocytes were incubated overnight to express the introduced proteins. Following incubation, 1/2 of the GST and 1/2 of the GST-XMsi1 injected oocytes were stimulated to mature with progesterone and the rest left untreated (Immature). When 50% of oocytes have reached GVBD, protein lysates were prepared, partially purified over glutathione sepharose and subjected to mass spectrometry. The experiment was repeated on three separate occasions. A total of 50 proteins were identified that specifically interacted with Musashi1. Of these, 21 only interacted significantly with Musashi1 in immature oocytes, 10 only interacted significantly with Musashi1 in progesterone-stimulated oocytes, while 19 were common to both conditions. The identified proteins for each category are indicated. (B) Gene set enrichment analysis was performed on the 50 interacting proteins and the top ten significant gene ontology molecular functions are shown graphically.  Oocytes were injected with antisense oligonucleotides targeting ePABP (ePABP AS) or a scrambled control oligonucleotide (Con AS). Following overnight incubation, the ePABP ASinjected oocytes were re-injected with water (no rescue), RNA encoding wobble ePABP (which is not targeted by ePABP AS) or RNA encoding GST-XMsi1 and then stimulated to mature with progesterone. Maturation was scored when ~50% of wobble ePABP injected oocytes had reached GVBD. The combined data of 3 independent experiments is shown. * indicates p<0.05. (C) Oocytes were co-injected with RNA encoding GST-Musashi1 (to facilitate analysis of Musashi1 activation status) and either scramble or ePABP antisense oligonucleotides (AS) as indicated and western blot analyses of the indicated components of the oocyte signaling cascade were performed from untreated immature (I) or progesteronestimulated oocytes. For scramble-injected oocytes, oocytes were segregated at GVBD50 based on whether they had not (-) or had (+) completed GVBD. ePABP antisense injected oocytes did not mature in response to progesterone and so were harvested when scramble-injected oocytes reached GVBD50. Endogenous ePABP, Ringo, MAP kinase (MAPK) and tubulin (an internal loading control) were analyzed as indicated. The activation status of endogenous MAP kinase was assessed with a phosphorylation (activation)-specific antibody (pMAPK). Ectopically expressed, GST tagged Musashi1 was detected with GST antibodies (Msi1) and phospho-specific Musashi1 antibodies were used to assess Musashi1 activation status      Oocytes were co-injected with mRNA encoding GST-tagged Xenopus PABPC1 and GFP-tagged Xenopus Musashi1 (GFP-XMsi1), Xenopus Musashi2 (GFP-XMsi2) or the GFP moiety alone. Following overnight incubation, the oocytes were lysed, subjected to GST pulldown and PABPC1 co-association assessed by GFP western blotting. A strong interaction with Musashi1 is seen (arrowhead), but a greatly reduced interaction with Musashi2 is evident even after correction for recovered GST-PABPC1 in the samples. PABPC1 did not interact with the GFP moiety alone. (B) The ability of Musashi1 or Musashi2 to interact with PABPC1 (Panel F) was assessed in three independent experiments and the data summarized graphically. Association data was normalized to the levels of recovered GST-PABPC1 in each sample. * indicates p<0.05; ** indicates p<0.01. (C) Oocytes were injected with antisense oligonucleotides targeting ePABP. Following overnight incubation, the oocytes were reinjected with either water ((No rescue) or PABPC1 mRNA (PABPC1 rescue), then stimulated with progesterone and the extent maturation was scored (% GVBD). The graph shows the combined data from six independent experiments. ****, p<0.0001.