|Home | About | Journals | Submit | Contact Us | Français|
Zygote arrest (Zar) proteins are crucial for early embryonic development, but their molecular mechanism of action is unknown. The Translational Control Sequence (TCS) in the 3’ untranslated region (UTR) of the maternal mRNA, Wee1, mediates translational repression in immature Xenopus oocytes and translational activation in mature oocytes, but the protein that binds to the TCS and mediates translational control is not known. Here we show that Xenopus laevis Zar2 (encoded by Xzar2) binds to the TCS in maternal Wee1 mRNA and represses translation in immature oocytes. Using yeast 3 hybrid assays and electrophoretic mobility shift assays, Zar2 was shown to bind specifically to the TCS in the Wee1 3’UTR. RNA binding required the presence of Zn2+ and conserved cysteines in the C-terminal domain, suggesting that Zar2 contains a zinc finger. Consistent with regulating maternal mRNAs, Zar2 was present throughout oogenesis, and endogenous Zar2 co-immunoprecipitated endogenous Wee1 mRNA from immature oocytes, demonstrating the physiological significance of the protein-RNA interaction. Interestingly, Zar2 levels decreased during oocyte maturation. Dual luciferase reporter tethered assays showed that Zar2 repressed translation in immature oocytes. Translational repression was relieved during oocyte maturation and this coincided with degradation of Zar2 during maturation. This is the first report of a molecular function of zygote arrest proteins. These data show that Zar2 contains a zinc finger and is a trans-acting factor for the TCS in maternal mRNAs in immature Xenopus oocytes.
Maternal effect genes encode proteins or RNAs found in the cytoplasm of the oocyte that regulate development of the early embryo after fertilization and prior to zygotic genome activation (Farley and Ryder, 2008). Maternal effect genes were first characterized in Drosophila in the 1980s, but mammalian maternal effect genes were not described until 2000 (Li et al., 2010). One of the earliest mammalian maternal effect factors reported, Zygote arrest 1 (Zar1) (Wu et al., 2003a), was found in a subtractive hybridization screen to identify mammalian maternal acting genes. Female mice null for Zar1 are infertile because embryogenesis is blocked at the 1-cell stage. Since then, Zar1 has been identified in many vertebrate species from frogs to humans (Brevini et al., 2004; Michailidis et al., 2010; Uzbekova et al., 2006; Wu et al., 2003b). A sequence related to Zar1 has also been identified in many vertebrates, and called Zar2 (Xzar2 in Xenopus laevis) or Zar1-like/Zar1l (Hu et al., 2010; Michailidis et al., 2010; Misra et al., 2010; Nakajima et al., 2009; Sangiorgio et al., 2008).
Both zygote arrest proteins have been implicated in progression of embryogenesis. Zar1 has been implicated in completion of fertilization, activation of the zygotic genome and progression past the 1-cell stage in mouse embryos (Wu et al., 2003a), and Zar2 (aka Zar1l) has been implicated in epidermalization of Xenopus embryos and progression past the 2-cell stage in mouse embryos (Hu et al., 2010; Nakajima et al., 2009). Zar1 and Zar2 transcripts are predominantly expressed in ovary and testis, with the highest levels in immature (germinal vesicle) oocytes (Brevini et al., 2004; Hu et al., 2010; Michailidis et al., 2010; Pennetier et al., 2004; Sangiorgio et al., 2008; Uzbekova et al., 2006; Wu et al., 2003a; Wu et al., 2003b). Zar1 and Zar2 transcripts decline by the 2-cell stage in mouse embryos (Hu et al., 2010; Wu et al., 2003b), the 8-cell stage in pig embryos (Uzbekova et al., 2006), the blastocyst stage in bovine embryos (Brevini et al., 2004), and gastrulation in Xenopus embryos (Nakajima et al., 2009; Wu et al., 2003b). In the chick embryo, Zar1 and Zar2 transcripts are expressed for at least 7 days (Michailidis et al., 2010). Thus, Zar1 and Zar2 likely play a role in early zygotic development in many vertebrate species.
Zar1 and Zar2 share a highly conserved C-terminal domain with invariably conserved cysteines across all vertebrates sequenced to date, which has led to the suggestion that Zar1 and Zar2 contain zinc fingers that regulate transcription. Indeed, Zar1 is implicated in activation of the zygotic genome (Wu et al., 2003a), and Zar2 is implicated in RNA synthesis, histone methylation and expression of nuclear reprogramming factors (Hu et al., 2010). However, Zar1 and Zar2 have also been shown to localize to the cytoplasm in mouse oocytes and embryos, and specifically to P-bodies, suggesting a role for Zar proteins in RNA metabolism (Hu et al., 2010; Misra et al., 2010; Wu et al., 2003a). While it is clear that Zar1 and Zar2 have important roles in early development, their molecular mechanism of action is still unclear.
Translational control of maternal mRNAs is critical to the developing zygote prior to genomic activation (Farley and Ryder, 2008). A plethora of developmentally regulated cis elements in the 3’ untranslated regions of mRNAs determine where, when and to what extent an mRNA is translated (Colegrove-Otero et al., 2005; MacNicol and MacNicol, 2010). In vertebrate meiosis, mRNA translation is often correlated with elongation of the poly(A) tail in a process called cytoplasmic polyadenylation (Villalba et al., 2011). Polyadenylation and translation of maternal Wee1 mRNA occur after the first meiotic metaphase in maturing Xenopus oocytes (Charlesworth et al., 2000). The resulting Wee1 protein is implicated in the lengthening of the first embryonic mitotic cell cycle (Murakami et al., 1999; Murakami and Vande Woude, 1998) and in morphogenesis at gastrulation (Murakami et al., 2004). Wee1 mRNA translation is under the control of at least two different types of cis elements in the 3’ UTR, the cytoplasmic polyadenylation element (CPE) and the newly described Translational Control Sequence (TCS) (Charlesworth et al., 2000; Wang et al., 2008). Both CPEs and TCSs repress translation in immature oocytes and stimulate translation in mature oocytes. Whereas the mechanism of translational regulation by CPEs is well documented and is mediated by CPE-binding protein (CPEB) (Villalba et al., 2011), the mechanism of translational regulation by TCSs has yet to be described and the protein(s) that binds to the TCS has yet to be identified.
In a previous study (Charlesworth et al., 2006), we isolated two proteins in a screen to identify the trans-acting factor that bound to a new cis-element, the polyadenylation response element (Charlesworth et al., 2002), now referred to as the Musashi binding element (MBE) (Arumugam et al., 2010), in the 3’ UTR of the Mos mRNA. One protein, Musashi (C36), bound to the MBE, whereas the second protein (A8) did not. Here, we show that A8 is the C-terminal of Xenopus laevis Zar2 and that it binds to the TCS in the Mos and Wee1 mRNA 3’ UTRs via a zinc finger. Furthermore, we show that Zar2 represses translation in immature Xenopus oocytes. We propose that Zar2 is a trans-acting factor for the TCS.
The yeast screen has been described previously (Charlesworth et al., 2006). Yeast transformations and β-galactosidase expression colony lift and liquid culture assays were performed as described in Yeast Protocols Handbook (Clontech). Yeast strains and plasmids were kindly provided by Dr. Marvin Wickens, University of Wisconsin, Madison.
All restriction enzymes, Klenow fragment and Quick Ligase were obtained from New England Biolabs. DNA oligonucleotides were synthesized by Integrated DNA Technologies.
pACT2-A8 (C-Xzar2) was isolated from a Xenopus oocyte cDNA library (Clontech).
pIIIA MS2-2.1 Mos M1 48 has been described previously (Charlesworth et al., 2006). pIIIA MS2-2.1 Mos WT 48 was made from pIIIA MS2-2.1 Mos M1 48 using QuikChange. The disrupted CPE (TTTGGT) was changed back to wild type CPE (TTTtaT). pIIIA MS2-2.1 Mos ΔMBE was made from pIIIA MS2-2.1 Mos M1 48 using QuikChange (Stratagene). The disrupted CPE (TTTGGT) was changed back to TTTtaT and 20 nt of the MBE (ATC CAT ATG TGA ATA TAT AG) (Charlesworth et al., 2002) were deleted. pIIIA MS2-2.1 Mos ΔTCS was made from pIIIA MS2-2.1 Mos M1 48 using QuikChange by deleting the 7 nt TCS (TTTGTCT).
pIIIA MS2 IRE and pACT2 IRP have been described previously (Bernstein et al., 2002).
pIIIA 2-2.1 β-globin and pIIIA 2-2.1 β-globin/TCS: PCR primers were designed to amplify the last 81 bp of the β-globin and the β-globin/TCS 3’ UTR sequence with a 5’ XhoI site and a 3’ NarI site, using pGEMGST β-globin and pGEMGST β-globin/TCS (TTTGTCT). The β-globin 3’UTR was amplified using Platinum Pfx (Invitrogen) and the PCR product digested with XhoI and NarI, and ligated into XhoI/NarI digested pIIIA MS2-2.1.
pIIIA MS2-2.1 Wee ΔTCS: QuikChange mutagenesis was also employed to delete the TCS elements (ATTGTCT and ATTATCT) within the Wee1 3’ UTR to generate Wee1 ΔTCS.
(Requests for the above plasmids should be made to AMM.)
pENTR Xzar2b: The 5’ end of Xzar2b was cloned using First Choice RLM-RACE Kit (Ambion), according to manufacturer’s directions. 5’ RACE was performed on the mRNA from immature oocytes from two different frogs to ensure the full length 5’ end was identified. The inner 5’RACE primer was 5’–GTC GAC TGG CCA AGG GCT GCG CGA C and the outer primer was 5’–GGT AAT CCG TGG GCT CAG AAG CCT T. EST sequences AW644676, BJ614564 and AW637792 were used to verify the integrity of sequence we had identified and support the idea that we have the paralog of Xzar2 that was previously reported (Nakajima et al., 2009). Once the sequence from the 5’ end was confirmed, primers were designed to amplify the full length Xzar2b, forward primer 5’–CAC CAT GGC GGG CTT TAT GTA TGC GC, reverse primer 5’–TCA GAC GAT GTA CTT GTA GCT GTA AGT GTT GT, using the high fidelity Pfu (Stratagene) according to manufacturer’s directions. 5’ Primers had CACC (underlined) for directional cloning into pENTR (Invitrogen). The full length Xzar2b sequence has been submitted to Genbank (Genbank ID: JQ776638).
pVL1393-FLAG-C-Xzar2 for baculovirus protein expression: pVL1393 (Orbigen) was cut with BamHI/PstI. A duplex encoding 3xFLAG with BamHI and XbaI sites was synthesized (IDT). The C-terminal 159–307 amino acids were amplified from pENTR Xzar2b, forward primer 5’–GCA TCT AGA ATG GCG GGC TTT ATG TAT GCG and reverse primer, 5’–GCT CTG CAG TCA GAC GAT GTA CTT GTA GCT with XbaI and PstI sites (underlined). The 3xFLAG and C-Xzar2 were digested with the appropriate restriction enzymes and ligated into pVL1393.
pGEX 6P-3 Xzar2b, pGEX 6P-3 N-Xzar2 (amino acids 1–158), pGEX 6P-3 C-Xzar2 (amino acids 159–307) and cysteine mutants: Plasmids were made by inserting PCR products from template pENTR Xzar2b into pGEX 6P-3 vector (GE Lifesciences). The forward primer for full length and N-terminus was 5’–GAT CGG ATC CAT GGC GGG CTT TAT GTA T and the reverse primer for full length was 5–CTA GGT CGA CTC AGA CGA TGT ACT TGT A, and for the N-terminus was 5’–CTA GGT CGA CTC ACT CCT TCA GCG GCT GTG A, with BamHI and SalI sites (underlined). The forward primer for C-terminus was 5’–CGG GAT CCA GAG CGC CCT CCC CCG AG and the reverse primer was 5’–ATA AGA ATG CGG CCG CTC AGA CGA TGT ACT TGT AGC, with BamHI and NotI sites (underlined). The cysteine to alanine mutations were performed by QuikChange mutagenesis methods with the following codon changes: C215A, TGC→GCC; C242A, TGT→GCT; C259A, TGC→GCC; C267A, TGC→GCC; C287A, TGC→GCC.
pXen N-MS2: PCR primers were designed to amplify the MS2 coding sequence from pJC5 (kindly provided by Jeff Coller, Case Western Reserve University, OH) (Gray et al., 2000) with a 5’ NcoI site and a 3’ ClaI site, forward primer 5’–GAT CCC ATG GCT TCT AAC TTT ACT CAG TTC and reverse primer 5’–GAT CAT CGA TGC GTA GAT GCC GGA GTT TGC TGC. Digested PCR product was ligated into NcoI/ClaI digested pXen1, replacing the GST coding sequence.
pXen MS2-Xp54: PCR primers were designed to amplify the Xp54 coding sequence from Xp54 in the MSP vector (Minshall et al., 2001) (kindly provided by Nancy Standart, University of Cambridge, UK) with a 5’ KpnI site and a 3’ BamHI site, forward primer 5’–GAT CGG TAC CCA TGA GCA CCG and reverse primer 5’–GAT CGG ATC CTT AAG GTT TGT. Digested PCR product was ligated into KpnI/BamHI digested pXen N-MS2.
pXen N-Xzar2-MS2: pXen1 was digested with NcoI and ClaI to remove GST coding sequence, then treated with Klenow to blunt ends and self-ligated make pXen ΔGST. MS2 coding sequence was amplified from pJC5 using primers with a 5’ XmaI site and a 3’ XbaI site, forward primer 5’–CTA GCC CGG GCT ATG GCT TCT AAC TTT ACT CAG TTC and reverse primer 5’–GAT CTC TAG AGT TAG TAG ATG CCG GAG TTT GCT G. Digested PCR product was then ligated into XmaI/XbaI digested pXen ΔGST to make pXen C-MS2. Amino acids 1–158 of Xzar2b was amplified from pENTR-Xzar2b with a 5’ KpnI site and a 3’ BamHI site, forward primer 5’–GAT CGG TAC CAT GGC GGG CTT and reverse primer 5’–GAT CGG ATC CGC TCT CTT CAG, appropriately digested, and ligated into KpnI/BamHI digested pXen C-MS2.
pXen rluc: PCR primers were designed to amplify Renilla luciferase (rluc) coding sequence (plasmid kindly provided by Nancy Standart) with a 5’ NcoI site and a 3’ ClaI site, forward primer 5’–CAT GCC ATG GCT TCG AAA GTT TAT GAT CCA and reverse primer 5’–GAT CAT CGA TTT ATT GTT CAT TTT TGA GAA CTC G. Digested PCR product was then ligated into NcoI/ClaI digested pXen1, replacing the GST coding sequence. This plasmid was linearized with EcoRI prior to in vitro transcription.
pXen fluc: The firefly luciferase (fluc) coding sequence was amplified from JC18 (kindly provided by Jeff Coller) (Gray et al., 2000) using primers with a 5’ NcoI site and a 3’ XhoI site, forward primer 5’–GAT CCC ATG GAA GAC GCC AAA AAC ATA AAG and reverse primer 5’–GAT CCT CGA GTT ACA ATT TGG ACT TTC CGC C. The PCR product was inserted into pXen1 using NcoI/XhoI, replacing the GST coding sequence.
pXen fluc-2x-SL (pXen fluc with stem-loops): An Nde I site was inserted into the β-globin 3’UTR of pXen fluc, 59 nucleotides upstream of the polyadenylation hexanucleotide, using QuikChange site directed mutagenesis (Stratagene) to make pXen fluc-NdeI. A DNA duplex (IDT) containing the sequence of two MS2 stem-loops (2x-SL) (Bardwell and Wickens, 1990) with 5’ and 3’ NdeI sites was digested with NdeI and ligated into NdeI digested pXen fluc-NdeI. This plasmid was linearized with SacI prior to in vitro transcription.
All plasmids were sequenced to verify integrity using the University of Colorado Cancer Center DNA Sequencing and Analysis Core. Requests for the latter plasmids should be made to AC. For in vitro transcription, all plasmids were linearized with PstI unless otherwise noted. 5’ capped RNA was synthesized in vitro with SP6 mMessage mMachine transcription kit (Ambion). RNA quality was assessed using gel electrophoresis.
FLAG–tagged C-terminus Xenopus Zar2 (FLAG-C- Zar2) protein was prepared from baculovirus infected Sf9 insect cells. Infections and cell culture were performed by Lori Sherman, Protein Production/Mab/Tissue Culture Core Manger, University of Colorado Cancer Center. About 3 ml of pelleted cells were suspended in lysis buffer (50 mM Tris HCl pH 7.7, 150 mM KCl, 0.1% Triton X-100, complete protease inhibitors (Roche)) and lysed by sonicating 20 s twice at 30 % output (Virsonic 50, Virtis). After centrifugation at 14,000 ×g for 15 min at 4°C, supernatant was collected and incubated with 3 ml of anti-FLAG agarose (Sigma) overnight at 4°C. Beads were washed with lysis buffer supplemented with 0.35 M NaCl for 1.5 h and protein was eluted with 3xFLAG peptide overnight. After quality and quantity of eluate were checked by SDS-PAGE, protein was concentrated with Vivaspin2 columns (GE Healthcare) at 5000 ×g until a concentration of 0.4 mg/ml was reached.
GST-C-Zar2 protein was purified from E. coli BL21 (DE3) (Novagen). The protein expression was induced with 0.5 mM IPTG overnight at 25°C. Cell pellets were suspended into GST lysis buffer (PBS, 0.1% Triton X-100, Complete EDTA-free protease inhibitor (Roche)) and lysed by sonicating 20 s, 3 times at 30 % output (Virsonic 50, Virtis) on ice. After centrifugation, supernatant was incubated with glutathione sepharose beads (GE Healthcare) for 1.5 h at room temperature. Beads were washed with wash buffer (PBS, 0.5 M NaCl, 1 % Triton X-100) for 30 min at 4°C. GST-C- Zar2 was eluted from beads with elution buffer (20 mM reduced glutathione, 10 mM Tris HCl pH 8.0, 150 mM NaCl, 1 mM DTT, 0.1 % Triton X-100), dialysed against dialysis buffer (50 mM Tris-HCl pH 7.4, 300 mM NaCl, 1 mM DTT, 0.01% TritonX-100) using Slide-A-Lyzer 10,000 MWCO (Thermo Scientific), and concentrated with Vivaspin2 (GE Healthcare). For the Zn2+ requirement study, 1 mM EDTA was added to lysis and wash buffers at the time of purification, but not to elution or dialysis buffers.
RNA probes were the last 50 nt of the Wee1 3’UTR labeled at the 5’ end with Cy5 (Integrated DNA Technologies) and were heat denatured immediately prior to use. 80 fmol of probe was combined with 400–800 ng protein (or as described in figure legends) in a 20 µl reaction containing 10 mM HEPES pH 7.7, 100 mM KCl, 1 mM MgCl2, 10 mM DTT, 20 µM ZnCl2, 50 mg/ml tRNA, 0.1 mg/ml BSA, 5% glycerol, 0.25% NP40 and incubated 20 min at room temperature. 0.5 µl of heparin (200 µg/µl) was added for a further 20 min. 5 µl of 5× loading buffer (0.15 g/ml Ficoll 400, 0.25% Orange G, 1xTBE) was added to the binding reaction and 5 µl of this loaded onto a 6% RNA retardation gel (Novex, Invitrogen). Gels were run according to manufacturer’s instructions. For the Zn2+ requirement study, ZnCl2 was eliminated from the binding buffer. Anti-FLAG (Sigma) and anti-C-Zar2 antibodies used for the supershift assay were added for the last 10 min of the binding reaction. For competition assays, proteins were preincubated for 20 min at room temperature with unlabeled RNA (IDT). The gel was imaged directly with the Odyssey (LiCor) in the 700 nm channel.
Adult female Xenopus laevis were housed and sacrificed according to internationally recognized guidelines and with the approval of the University of Colorado Denver Institutional Animal Care and Use Committee. Xenopus laevis (Nasco) oocytes were isolated and cultured as has been described (Machaca and Haun, 2002). All incubations were carried out in 0.5X L-15 (MediaTech, Inc) with penicillin (100 µg/ml) and streptomycin (50 µg/ml). Dumont stage VI (Dumont, 1972) oocytes were selected and injected with 23 nl of the appropriate MS2 fusion mRNA: approximately 1 ng of MS2, 20 ng of MS2-Xp54, and 5 ng, 20 ng or 50 ng of N-Xzar2-MS2. Oocytes were incubated overnight (~16 hours) at 18°C, then injected with ~100 pg of fluc-2x-SL or fluc mRNA, together with ~5 pg of rluc mRNA as a loading control, in 23 nl of nuclease free water. Injections were performed using a Drummond NanoInject II microinjector, and media was supplemented with 2.5% Ficoll 400 during injections. Appropriate samples were then induced to mature with 2 µM progesterone and all samples were collected 2–3 hrs after progesterone treated samples underwent GVBD, as indicated by the appearance of a white spot. Immature samples were time-matched to progesterone-treated samples (~11 hours after luciferase injection).
Two pools of 5 oocytes were collected from each sample and lysed in duplicate in 50 µl/oocyte of Passive Lysis Buffer (Promega). A 10 µl portion of cleared lysate was analyzed using the Dual-Luciferase Reporter Assay System (DLR) (Promega) on a Synergy HT plate reader (BioTek). The ratio of firefly to Renilla luciferase activity (DLR ratio) was calculated for each experimental point. The mean DLR ratio was calculated for immature oocyte samples injected with MS2 alone and all immature samples were normalized to this value, resulting in MS2 alone values being arbitrarily set to 1 and all other values being expressed as a change in translation relative to MS2 alone. Progesterone-treated samples were similarly normalized to the mean value of progesterone-treated MS2 alone. This experiment was performed independently with the same parameters on multiple frogs (n = 4–5) and normalized results were pooled. One-way analysis of variance and post hoc Bonferroni multiple comparison test were performed using GraphPad Prism 5 software to analyze difference among the means of the normalized results. Differences with P-value less than 0.01 were considered statistically significant. Additionally, a Mann-Whitney U test was performed to analyze differences among medians of the normalized results with similar results.
Pools of 10 oocytes were collected from each sample and lysed in 10 µl/oocyte of NP-40 lysis buffer (1% Igepal CA-630, 20 mM Tris pH 8.0, 137 mM NaCl, 10% glycerol, 2 mM EDTA) supplemented with protease and phosphatase inhibitors (HALT, Pierce). 5 or 10 µl of cleared lysate (0.5 or 1 oocyte equivalent respectively) were loaded onto a NuPage 4–12% Bis-Tris polyacrylamide gel (Novex). For stage I to VI and in vitro matured oocytes, oocytes were lysed in NP-40 lysis buffer and total protein amount was measured by BCA protein assay kit (Pierce). 3.75 µg total protein per lysate was loaded onto the gel. Electrophoresis was performed using MOPS-SDS running buffer (Invitrogen), then transferred to 0.45 µm Immobilon-FL PVDF membrane (Millipore) using an XCell II Blot Module (Invitrogen) according to NuPage technical guide protocol. Membranes were probed with MS2 antibody (TetraCore), GST antibody (Santa Cruz Biotechnology), β-Tubulin antibody (Sigma), or Zar2 antibodies (Charlesworth). Secondary antibodies were 1:20,000 IR Goat anti-rabbit IR Dye 800 CW and Goat anti-mouse IR Dye 680LT (LiCor). Membranes were imaged on an Odyssey infrared imager and data was analyzed using Odyssey 2.1 software (LiCor).
Anti-C-Zar2 antibodies were raised in rabbit against peptides encoding C-terminus (amino acids 267 –286) of Zar2 and anti-N-Zar2 antibodies were raised against N-terminus (amino acids 29 – 44) and purified by peptide column (Proteintech). To test antibody specificity, 1.5 µg/ml of the immunizing peptide was incubated 1 h, 25°C with the primary antibody before adding to the transfer membrane.
Pools of 50 oocytes were lysed in 500 µl homogenization buffer (150 mM NaCl, 50 mM Tris pH 7.5, 0.5% NP40, 2% BSA) supplemented with 10 mM ribonucleoside vanadyl complex (NEB), HALT protease inhibitors, RNaseOUT (Invitrogen) and 1 mM DTT. Lysates were clarified by centrifugation 2× 5 min, 14,000 ×g, 4°C. 4 µg of antibody was added and incubated for 6 h, 4°C. 30 µl of a 1:1 slurry of protein A/G PLUS-agarose (Santa Cruz Biotechnology) was added and samples rotated 30 min. Agarose beads were washed twice with homogenization buffer and twice with NP40 lysis buffer. 5 oocyte equivalent of beads was used to analyze immunoprecipitation efficiency by western blot. RNA was extracted from the beads using RNA STAT-60 (Tel-Test) according to manufacturer’s directions.
The amount of immunoprecipitated RNA was determined by semi-quantitative PCR. cDNA was synthesized from 40 oocyte equivalents of RNA using iScript (Bio-Rad). 1 µl of cDNA was amplified using Platinum Taq (Invitrogen). Annealing was performed at 60°C. The following gene-specific primers and cycles were used: Mos (f) 5’–GAG AAT CAC AGT TCC ACA GCA ACC, Mos (r) 5’–AGA CAG TTC CCC CAA CAG AAG C, 30 cycles; Wee1a (f) 5’–TGC CGG AAG CAG ACA GAG TTG G, Wee1a (r) 5’–TTA GCG GCT TTC AAC TCC CTC TCA, 25 cycles; Protein Phosphatase Inhibitor 2 (PPI2) (f) 5’–CGT GTC ATT AGC AAG CCA GAG AC, PPI2 (r) 5’–GCA ATC AAG TGT CTG GCG AGT C, 40 cycles. The cycle number was determined by running 1/100 and 1/1000 oocyte equivalent of cDNA from total RNA and making sure there was a difference in the amount of PCR product between these two amounts of cDNA. Also, these standards were used to normalize the exposure of the gels so all the positive controls look the same and the immunoprecipitated mRNAs can be directly compared.
Total RNA was extracted from pools of 20 oocytes using Tri Reagent® Solution (Ambion) according to manufacturer’s directions. A LiCl precipitation step was included for further purification. cDNA was synthesized from 2 µg of total RNA using Superscript III (Invitrogen) and oligo dT(12–18) according to manufacturer’s directions. PCR was performed in duplicate (only one set of bands is shown) using One Taq™ Hot Start (New England BioLabs) according to manufacturer’s directions. Forward primer was fluc 5’-TCT TCC CGA CGA TGA CGC, reverse primer was β-globin 5’-AGA CTC CAT TCG GGT GTT CTT GAG G. Annealing was 56°C and the reaction proceeded for 26 cycles. Quantitative range was determined by using dilutions of the control cDNA.
A previous study screened an unfertilized Xenopus laevis egg library to identify proteins that bound to the MBE in the 3’ UTR of the Mos mRNA (Charlesworth et al., 2006). One of the proteins that was isolated was Musashi and we showed that Musashi bound to the MBE in the Mos 3’ UTR. Another protein, designated A8, was isolated in that screen, and we characterize A8 in this study. The screen was performed with the Mos M1 48 Mos UTR where the last 48 nucleotides of the Mos UTR with a mutation (M1) in the CPE (UUUAAU to UUUggU) were used to prevent recovery of CPEB. As A8 was recovered on Mos M1 48 this demonstrated that A8 did not bind to the CPE. The yeast three hybrid assay was further used to determine where A8 interacted with the Mos UTR. To show that A8 interaction was not an artifact of binding to the mutant CPE, the CPE was restored to make Mos WT 48. A8 still interacted with Mos WT 48 showing that the interaction was not an artifact of the mutant CPE. The interaction did not appear as strong for unknown reasons. Next we tested if A8 interacted with the MBE in the Mos 3’ UTR. The MBE was deleted from the Mos 3’ UTR in the RNA hybrid to make Mos ΔMBE as shown in Figure 1A. Removal of the MBE (Mos ΔMBE) did not prevent interaction with A8 (Fig. 1B), indicating that A8 does not interact with the MBE and likely interacts with a different element within the 3’ UTR of Mos mRNA. The only place left for this element to be located was 3’ of the polyadenylation hexanucleotide. Accordingly, when these seven nucleotides at the very 3’ end of the UTR were deleted to make Mos ΔTCS, interaction with A8 was abrogated. Sequence analysis revealed that these seven nucleotides were remarkably similar to the TCSs in the Wee1 3’ UTR. This suggested that Mos has a TCS (MacNicol and MacNicol, 2010) and that A8 was a TCS-interacting protein. The Wee1 3’ UTR has three maturation-type CPEs (Charlesworth et al., 2000) and two TCSs between the first two CPEs (Wang et al., 2008). To test if A8 interacted with other TCSs, the Wee1 3’ UTR was used in the yeast three hybrid assay, either with the TCSs intact (wild type, Wee WT) or with the TCSs deleted (Wee ΔTCS) (Fig. 1A). A8 interacted with Wee WT, but not when the TCSs were deleted (Wee ΔTCS), indicating that the TCS is necessary for interaction with A8 (Fig. 1B). It should be noted that the Wee ΔTCS construct still contains two intact CPEs, further showing that A8 does not interact with CPE sequences. To test if the TCSs are sufficient for interaction with A8, we inserted the Mos TCS into β-globin 3’ UTR (βg/TCS). A8 did not interact with wild type β-globin 3’ UTR (βg), but did interact with βg/TCS, as assessed by both colony lift assay (Fig. 1B) and liquid culture assay (Fig. 1C). These data demonstrate that a TCS is necessary and sufficient for interaction with A8.
BLAST® (NCBI) analysis showed that the amino acid sequence of A8 was similar to the entire C-terminal half (i.e. amino acids 159 – 307) of mouse Zygote arrest1 (Zar1) (Wu et al., 2003a). To ensure the most complete and accurate 5’ end of A8 was identified for analysis and cloned for mechanistic studies, 5’ RACE from two individual frogs was performed. The full-length sequence of A8 was 1128 nt, predicting a protein of 307 amino acids. Upon comparison, the full-length sequence of A8 was more similar to Xzar2 (Nakajima et al., 2009), which was identified during the timeframe of this study, than to Xenopus laevis Zar1 (Genbank ID: AY283176) (Wu et al., 2003b) (Fig. 2A). We propose we have cloned the paralog of Xzar2 arising from the pseudotetraploid nature of Xenopus laevis. Further, we suggest that Xzar2 (Genbank ID: AB190316) (Nakajima et al., 2009) be designated Xzar2a and our sequence from this point forward be designated as Xzar2b (Genbank ID: JQ776638). According to this proposed nomenclature, all experiments in this study were performed with Xzar2b. In addition to the observed C-terminal conservation between Zar1 and Zar2 proteins, an area of homology in the N-terminal was also observed. In the subsequent experiments, we show that the C-terminal half of Zar2 binds RNA and the N-terminal half regulates mRNA translation, as summarized in Fig. 2B.
The data in Figure 1 suggested that Zar2 might bind to maternal mRNAs that are regulated during oocyte maturation. To further this line of study, it was essential to determine if Zar2 protein is present in Xenopus oocytes. To do this we raised antibodies to peptides in the C-terminal and N-terminal domains. An antibody specific to a peptide in the C-terminal domain that had the most mismatches between Zar2 and Zar1 (amino acids 267–286) (Fig. 2A) was developed. To test that the antibody recognized Zar2, we expressed GST-Zar2 in immature oocytes. The C-terminal Zar2 antibody recognized a band at the same size as the GST antibody (Fig. 3A). To show specificity we blocked the antigen binding sites on the C-terminal Zar2 antibody by incubating with the immunizing peptide before western blot. This treatment prevented the C-terminal Zar2 antibody from recognizing GST-Zar2, demonstrating that the antibody was specific. Recognition of Zar2 and not Zar1 by the C-terminal antibody was verified by western blot of recombinant proteins (Fig. 3D). The C-terminal antibody is used for western blot. We also raised an antibody for immunoprecipitation against a peptide in the N-terminal of Zar2 that was predicted to be on the surface of the protein, amino acids 29–44 (Fig. 2A). These amino acids are not conserved at all between Zar2 and Zar1. We analyzed the N-terminal antibody in a similar fashion to the C-terminal antibody (Fig. 3B). The N-terminal antibody recognized GST-Zar2 and this recognition was blocked by the immunizing peptide. We use the N-terminal antibody mainly for immunoprecipitation (Fig. 6), however, it is also useful for western blot. To test if the antibodies could detect endogenous Zar2 we western blotted lysates from immature and progesterone-stimulated mature oocytes (Fig. 3C). Both Zar2 antibodies recognized a band that migrated about 39 kDa and for both antibodies this band was more intense in immature oocytes compared to mature oocytes. We conclude that endogenous Zar2, which has a predicted molecular weight of 35 kDa, migrates at 39 kDa on a polyacrylamide gel, similar to mouse Zar1 (Wu et al., 2003a). We have expressed different truncations of Zar2 with different tags in several eukaryotic cell types and Zar2 consistently migrated slightly larger than predicted (data not shown), which suggests post-translational modification. The C-terminal antibody recognized an additional band that migrated at 32 kDa (Fig. 3), however, the 32 kDa band was not detected by the N-terminal antibody (Fig. 3C), suggesting that amino acids 29–44 were not present and that the 32 kDa band is therefore not an unmodified version of Zar2. A similar smaller band was also seen with antibodies to mouse Zar1, and it was proposed that this might be a degradation product (Wu et al., 2003a). Although it has been shown that the N-terminal of Zar1 undergoes alternative splicing in several species (Uzbekova et al., 2006), isolation of the 5’ end of Xzar2 did not reveal any evidence of alternative 5’ ends in our hands. At this time, we do not know if the smaller band recognized by the C-terminal Zar2 antibody is a degradation product, a product of alternative splicing or a cross-reacting protein that is not Zar2. The C-terminal antibody also recognizes a band at 20 kDa. As this band follows the same pattern as Tubulin expression (Fig. 3E–G), we think it is an unrelated cross-reacting band.
Using the C-terminal antibody, Zar2 protein levels during oogenesis were characterized. The total amount of Zar2 per oocyte increased during stages I – IV of oogenesis (Dumont, 1972), as the oocytes grow in size. Zar2 reached maximum levels at stage IV, even though the oocyte continues to grow in size from stage IV to stage VI, as seen by the increase in Tubulin (Fig. 3E). However, it was stage I oocytes that had the greatest concentration of Zar2 relative to Tubulin (Fig. 3F). Interestingly, the 32 kDa band showed a similar profile. These data suggest that Zar2 might play a role in early oogenesis.
Next, we characterized Zar2 levels during meiotic maturation. Levels of Zar2 gradually declined during oocyte maturation (Fig. 3G). In mature oocytes, there was a 60 – 80% (n=5) decrease in the amount of endogenous Zar2 compared to immature oocytes. Interestingly, the 32 kDa band showed a similar profile. Thus Zar2 protein is present in immature oocytes, consistent with a role in maternal mRNA regulation.
As A8 (the C-terminal domain of Zar2) interacted with TCSs in the Wee1 and Mos 3’ UTRs (Fig. 1), we examined whether Zar2 directly binds to the TCS by use of electrophoretic mobility shift assays (EMSA). In these experiments 50 nt of the Wee1 3’ UTR, very close to the poly(A) tail, were used as the labeled probe (Fig. 4A). The cis elements are indicated. There are three CPEs and two TCSs. The TCSs overlap with the CPEs. The nucleotides of the TCS mutation (UU to gg) that disrupt polyadenylation and translation control (Wang et al., 2008) are marked in red. The C-terminal of Zar2 was FLAG-tagged (FLAG-C-Zar2) and expressed in Sf9 cells. The binding of FLAG-C-Zar2 to the Wee1 3’ UTR was determined. Figure 4B shows that a specific complex forms when FLAG-C-Zar2 is purified from baculovirus infected Sf9 cells but not with proteins mock purified from uninfected cells. Neither was a specific complex formed when the FLAG elution buffer was used in the EMSA. This shows that the specific complex was not formed from endogenous insect proteins or from FLAG peptides. The specific complex could be supershifted by antibodies to FLAG and Zar2, but not by antibodies to Tubulin, showing that the complex contained FLAG-C-Zar2 (Fig. 4B).
To show that Zar2 was targeting the TCS in the Wee1 3’ UTR, we used competition assays with unlabeled RNA specifying the same 50 nt as the labeled Wee1 probe. The labeled probe could be competed with a 50-fold excess of unlabeled Wee1 UTR, but it could not be competed when the TCSs in the unlabeled RNA were disrupted (red UU mutated to gg) (Fig. 4C). It should be noted that mutation of TCS1 also disrupts CPE1, however there are two other CPEs present in the unlabeled probe and neither of them competes with the labeled probe for Zar2 binding. These results indirectly show that Zar2 binds to the TCS. To directly show that Zar2 binds to the TCS, we used labeled probes where the TCSs had the same mutations as in the competition study. For these experiments, purified bacterially expressed Zar2 proteins were used. First, we verified that GST-tagged Zar2 expressed in E. coli binds to the Wee1 3’ UTR. Figure 4D shows that GST-C-Zar2 bound to the Wee1 3’ UTR, but GST alone or the N-terminal of Zar2 (GST-N-Zar2) did not bind, even though equivalent amounts of protein were used (Fig. 4D lower panel). Next, we asked if GST-C-Zar2 could bind to mutant Wee1 UTRs. When either TCS1 or TCS2 was disrupted, the binding of Zar2 was unchanged (Fig. 4E). However, when both TCSs were disrupted, binding of Zar2 to the Wee1 3’ UTR was markedly reduced. The number of CPEs was the same in the TCS1 mutant versus the TCS1&2 mutant showing that the change in binding was not due to Zar2 binding to the CPEs. Increasing amounts of Zar2 protein showed a dose response with respect to the amount of complex formed, verifying the specificity of this reaction. Because only one specific complex is seen, these data show that Zar2 can bind to either TCS, but two Zar2s do not bind to both TCSs at the same time.
The conserved C-terminus of Zar1 has been speculated to contain an atypical PHD domain, a fungal C6 domain, or a FYVE domain based on conserved cysteine residues (Sangiorgio et al., 2008; Uzbekova et al., 2006; Wu et al., 2003a). All of these domains are zinc fingers, but they are found in proteins with different functions, such as histone modification, phospholipid binding and DNA binding (Matthews and Sunde, 2002; Todd and Andrianopoulos, 1997). Also, it is not known which cysteines in Zar1 form the putative zinc finger(s). To test whether the C-terminal of Zar2 contained a zinc finger, we purified the protein from bacteria in the presence of EDTA to remove pre-incorporated Zn2+ and then performed the EMSA either with or without adding back Zn2+ to the binding buffer. EDTA treatment also removed Mg2+ that was essential for Zar2 RNA binding, but activity was recovered when Mg2+ was added back to the binding buffer (data not shown). In decreased Zn2+ conditions (but normal Mg2+ conditions), the binding of Zar2 to the Wee1 3’ UTR was dramatically reduced (Fig. 5A), demonstrating that Zar2 binds to RNA using a zinc finger. To substantiate this result, we mutated one cysteine in each conserved pair of cysteines illustrated in Figure 2. The ability of the mutant proteins to bind the Wee1 3’ UTR was assessed by EMSA. Mutating any of these four pairs of cysteines abolished binding of Zar2 to the Wee1 3’ UTR (Fig. 5B), even though equivalent amounts of protein were used (Fig. 5B, lower panel). These results demonstrate that Zar2 binding to RNA is dependent on the structural integrity of a zinc finger.
Our previous results assessed the in vitro interaction of recombinant Zar2 with synthetic RNA. To test whether these interactions actually occurred in the immature Xenopus oocyte, endogenous Zar2 was immunoprecipitated from immature oocytes (Fig. 6A) using the N-terminal Zar2 antibody (Figs. 2 and and3).3). RNA was extracted from the immunoprecipitate, and the presence of the endogenous Wee1 or Mos mRNA was assessed by semi-quantitative RT-PCR. The PCR conditions were optimized to be able to distinguish a 10-fold difference in total RNA (total 1/100 vs total 1/1000). Endogenous Wee1 mRNA co-precipitated with endogenous Zar2 indicating that Zar2 interacts with Wee1 mRNA in immature oocytes. More Wee1 mRNA was immunoprecipitated than is found in 1/100 of an oocyte (we estimate 1/30). As we originally found Zar2 because it interacted with the Mos 3’ UTR, we tested if endogenous Mos mRNA coprecipitated with Zar2. Figure 6B shows that Mos mRNA also co-precipitates with Zar2. The Mos mRNA that was immunoprecipitated was more than 1/1000 but less than 1/100 of an oocyte (we estimate 1/200). In contrast, the protein phosphatase inhibitor 2 mRNA, a negative control, (Charlesworth et al., 2006), showed much lower affinity and the amount of PPI2 mRNA that was immunoprecipitated was close to that found in 1/1000 of an oocyte. These data show that endogenous Zar2 preferentially co-immunoprecipitates with Wee1 and Mos mRNAs.
The TCS confers translational repression in immature oocytes and translational activation in meiotically maturing oocytes (Wang et al., 2008). If Zar2 is a trans-acting factor for the TCS, then it too should repress translation in immature oocytes and/or stimulate translation in maturing oocytes. To test this, we used the tethered assay, where viral MS2 coat protein fusions bind to MS2 stem-loops in the reporter RNA, thus tethering the MS2 coat proteins to the RNA ((Gray et al., 2000), reviewed recently in Minshall et al., 2010). This method separates effects on translation regulation from changes in RNA binding. The firefly luciferase reporter construct contained MS2 stem-loops (fluc-2x-SL), whereas Renilla luciferase did not contain MS2 stem-loops and was used as a loading control (Fig. 7A). Since the C-terminal domain of Zar2 is the RNA binding domain (Figs. 1 and and4),4), we hypothesized that the N-terminal contains the translational activation domain. Therefore, we replaced the C-terminal domain of Zar2 with MS2 coat protein (Fig. 7A). MS2-Xp54 was used as a positive control for translational repression (Minshall et al., 2001). Oocytes were injected with mRNA encoding the indicated MS2 fusion protein and incubated overnight to allow expression. Oocytes were then injected with the luciferase translation reporter constructs. Progesterone was added to half the oocytes and all oocytes were incubated until the progesterone-stimulated oocytes were at meiosis II. Oocytes were lysed and the amount of luciferase protein accumulated was measured. Firefly luciferase is reported relative to Renilla luciferase (Fig. 7B). Oocytes not expressing a fusion protein (−) showed no significant difference in translation of firefly luciferase mRNA from oocytes expressing MS2, verifying that the binding of MS2 to the reporter mRNA does not stimulate or repress translation, as expected (Gray et al., 2000). Also, MS2-Xp54 repressed translation of firefly luciferase by about 60% compared to MS2 alone (in agreement with Minshall et al., 2001). MS2 and MS2-Xp54 were expressed at equivalent levels (Fig. 7C). When N-Zar2-MS2 was tethered to the luciferase reporter, it repressed translation in immature oocytes by up to 33 ± 7.5 %. This decrease in translation was not due to destabilization of the firefly RNA reporter as assessed by semi-quantitative PCR (Fig. 7F). Increasing amounts of injected RNA resulted in increasing amounts of N-Zar2-MS2 fusion protein that accumulated (Fig. 7C) and increasing amounts of repression, showing that the observed repression is due specifically to the presence of the N-Zar2-MS2 fusion protein. At the highest dose tested, N-Zar2-MS2 protein expression was similar to MS2 and MS2-Xp54 (Fig. 7C). Data were analyzed using one-way ANOVA and differences in means with a p-value <0.01 (**) were considered significant. When 20 ng or 50 ng of N-Xzar2-MS2 RNA were injected, significant repression was observed. Additionally, Mann-Whitney U tests were performed to assess differences in median values with similar conclusions. Repression of translation was specific to N-Zar2-MS2 being tethered to the RNA, as no repression was seen when the stem-loops were omitted from the firefly luciferase reporter (fluc) (Fig. 7D), even though proteins were expressed at similar levels (Fig. 7E).
Next, we determined whether the translational control exerted by Zar2 changed during meiotic maturation of Xenopus oocytes. Accumulation of luciferase in MS2-, or MS2-Xp54-expressing oocytes, or oocytes not expressing fusion proteins (−), did not change during maturation (Fig. 7B). However, N-Zar2-MS2-mediated repression was relieved during oocyte maturation (Fig. 7B). This change was observed at all levels of protein expression and was determined to be significantly different (p-value <0.01)(**) using one-way ANOVA and Mann-Whitney U test. Because N-Zar2-MS2 binds to luciferase through MS2 coat protein:stem-loop interactions, RNA binding does not change during maturation. Rather, relief of repression could be due to loss of N-Zar2-MS2, as protein levels were markedly lower in progesterone treated samples (Fig. 7C). There was a 70 – 80% (n = 5) reduction in N-Zar2-MS2 in progesterone-treated oocytes, which is comparable to the reduction in endogenous Zar2 during oocyte maturation (Fig. 3G).
Taken together, these results show that Zar2 repressed translation in immature oocytes and that repression was relieved in maturing oocytes.
The closely related Zar1 and Zar2 proteins are implicated in early zygotic progression, but their mechanism of action is unknown. Histone modification, transcriptional regulation and RNA metabolism have all been proposed. Here we showed that Xenopus Zar2 binds maternal mRNAs and regulates translation.
The maternal Wee1 mRNA has been shown to contain two TCS cis-elements that repress translation in immature oocytes and activate translation in maturing oocytes (Wang et al., 2008). This study provides evidence that Zar2 is a trans-acting factor for the TCS. Several criteria are necessary to support this conclusion: Zar2 should bind to the TCS, repress translation in immature oocytes, and activate translation during meiosis. In support of these criteria we show the following pieces of evidence. (1) Zar2 associates with Wee1 and Mos mRNAs in vivo. In yeast cells, a Zar2 fusion protein interacts with Mos and Wee1 3’ UTRs (Fig. 1). The TCS was necessary and sufficient for Zar2 interaction with the RNA hybrid. Moreover, endogenous Wee1 and Mos mRNAs can be co-immunoprecipitated with endogenous Zar2 from immature oocytes (Fig. 6). (2) Zar2 binds directly to the TCS in vitro. Zar2 forms a specific complex with Wee1 RNA and labeled probe binding in the specific complex was competed by unlabeled RNA that contained a TCS (Fig. 4). Furthermore, when the TCSs were disrupted, the specific complex showed markedly reduced binding (Fig. 4E). (3) Zar2 represses translation in immature oocytes. This is an important criterion, as this is the translational activity that the TCS confers to RNA reporters (Charlesworth et al., 2000; Wang et al., 2008). In this study, we show that Zar2 repressed translation of a luciferase reporter in immature oocytes (Fig. 7). This translational repression is about 33% at the maximum Zar2 dose that was injected. This is comparable to the TCS in the Wee1 3’ UTR that contributes about 33% repression in immature oocytes (Wang et al., 2008). Thus, we propose that Zar2 is a trans-acting factor for the TCS.
The TCS also confers translational activation in maturing oocytes (Wang et al., 2008). The activation of mRNA translation by Zar2 during meiosis has not been established in this study. However, Zar2-mediated repression was relieved during meiosis. One reason for the relief of repression and not activation may be that N-Zar2-MS2 is degraded during meiosis so N-Zar2-MS2 could not repress, nor could it activate translation. Interestingly, the endogenous Zar2 protein is partially degraded during meiosis (Fig. 3), so degradation might be part of the mechanism of action of Zar2. Degradation of Zar2 might appear to be contradictory for a role in embryogenesis. However, it is known that approximately 80% of CPEB is degraded during meiosis (Mendez et al., 2002), and the remaining CPEB plays an important role in translating cyclin B1 mRNA during the first mitotic cell cycles (Groisman et al., 2000). Alternatively, the regulation of TCS-containing mRNAs may be controlled by more than one trans-acting factor. For example, Zar2 might repress translation of TCS-containing mRNAs during oogenesis. After degradation of Zar2 during meiosis, another protein may load onto the TCS-containing mRNA to activate translation. Precedent for protein switching in the regulation of translational control comes from analysis of CPEB4 protein, which replaces CPEB1 during Xenopus meiotic maturation (Igea and Mendez, 2010). We are initiating studies to assess TCS-directed translational activation during maturation.
Mouse Zar2 (Zar1l) co-localizes with P-body components, RAP55 (aka Lsm14a) and LIN28, to cytoplasmic foci in 2-cell embryos (Hu et al., 2010). This suggests that Zar2 is found in P-bodies, which are sites of mRNA translational repression and degradation (Balzer and Moss, 2007; Parker and Sheth, 2007; Yang et al., 2006). That Zar2 may be a component of P-bodies is consistent with our results demonstrating that Xenopus Zar2 binds mRNA and represses translation. The original proposal that Zar1 regulated transcription was based on the observation that in embryos from Zar1-null mice, the zygotic genome was not activated (Wu et al., 2003a). More specifically, the embryos did not activate transcription of Transcription Requiring Complex (TRC), a marker of embryonic genomic activation (Wu et al., 2003a). Consistent with this, is the observation that a dominant negative Zar2/Zar1l also disrupted general transcription (Hu et al., 2010). One explanation for these observations could be that zygotic transcription is activated by the protein product of a Zar2 target mRNA. However, it should be noted that although we have shown that Xenopus Zar2 binds to RNA, this does not necessarily mean that the Zar family of proteins are not transcriptional regulators as well as translational regulators. ChIP analysis of Zar2 in human breast cancer cells shows that Zar2 interacts with the BRCA2/Zar2 bidirectional promoter (Misra et al., 2010). Furthermore, dual function zinc fingers are not without precedent. Wilms’ Tumor 1 protein and TFIIIA contain zinc fingers that bind both RNA and DNA, and regulate gene expression both transcriptionally and post-transcriptionally (Hall, 2005; Morrison et al., 2008; Burdach et al., 2012).
Central to the predicted function of the Zar proteins are the conserved cysteines in the C-terminal domain. There are 12 conserved cysteines and different investigators have suggested that different cysteines may form different types of zinc fingers, however, these cysteines are not recognized by protein motif-predicting software. Zar proteins contain pairs of cysteines (C-X2-C) similar to zinc knuckles that are found in zinc fingers with interleaved pairs of cysteines/histidines, in a cross-brace topology, such as FYVE (C4-C4) (phospholipid binding), PHD (C4-HC3) (chromatin modification) and RING (C4-HC3) (ubiquitination) domains (Matthews and Sunde, 2002). However, motifs such as PHD domains and FYVE domains require specific amino acids relative to the cysteines, which Zar proteins do not have. Also, Zar proteins tend to have more amino acids between the cysteine pairs in the loops (22–33 aa) than PHD or FYVE domains (10–18 aa) (Bienz, 2006; Gillooly et al., 2001) (see Fig. 2 legend for putative loop positioning). Furthermore, loop 1 is separated from loop 2 by four amino acids in FYVE and PHD domains and thirteen in Zar proteins. Ring fingers do tend to have more amino acids in the loops, between the pairs of cysteines/histidines (30–40), similar to Zar proteins but the two loops in RING fingers are spaced only 2–3 amino acids apart (Budhidarmo et al., 2012). Zar proteins also have conserved cysteines and histidines that are not in closely spaced pairs, similar to other zinc fingers such as binuclear clusters and classical zinc fingers (DNA binding), CCCH proteins (mRNA stability, AU-rich element (ARE) binding protein), RANBP2 proteins (splicing) and the zinc finger of CPEB (maternal mRNA regulation) (Brown, 2005; Burdach et al., 2012; Hake et al., 1998; Hall, 2005; Kutateladze, 2006; Loughlin et al., 2009). However, the spacing of the cysteines and histidines that are not in CxxC pairs in Zar proteins do not match the spacing of any of these zinc fingers. Thus, the cysteines in Zar2 share similarities with all these types of zinc fingers, but lack various characteristics that would firmly place Zar2 in any one of these categories. We show here for the first time that a Zar protein requires Zn2+ for a biological function. We showed that Zar2 binding to the Wee1 RNA probe was markedly reduced in the absence of Zn2+ (Fig. 5A). Moreover, we identified four pairs of cysteines that are important for RNA binding (Fig. 5B), supporting predictions that Zar proteins contain zinc fingers. We further propose that Zar proteins have a new type of functional zinc finger motif.
Although Zar2 regulates translation during oocyte maturation, we have been unsuccessful in showing a role for Zar2 in oocyte maturation. However, it should be noted that Zar2 might not regulate meiosis because it is the downstream protein products of the target mRNAs of Zar2 that will determine the physiological role of Zar2. Although the Wee1 mRNA is translated during oocyte maturation (Charlesworth et al., 2000), the resulting Wee1 protein is implicated in the first embryonic mitotic cell cycle, and not meiosis (Murakami et al., 1999; Murakami and Vande Woude, 1998). Indeed, studies in mice with disrupted Zar proteins have shown embryonic defects around the 1- and 2-cell stage rather than meiotic defects (Hu et al., 2010; Wu et al., 2003a). Wee1 protein is also important at gastrulation (Murakami et al., 2004), which is later during Xenopus development. Moreover, in Xenopus embryos, it has been shown that Zar2 plays a role in production of secreted factors involved in epidermalization (Nakajima et al., 2009), which is also later in development. We propose that Zar2 represses TCS-containing maternal mRNAs until they are required for embryogenesis. Also, we expect that Zar2 regulates translation of maternal mRNAs other than Mos and Wee1.
Zar2 may also play a role repressing maternal mRNAs during oogenesis. The amount of Zar2 increases during oogenesis as the oocytes grow in size, reaching a maximum at stage IV (Fig. 3). However, the concentration of Zar2 relative to Tubulin is remarkably high in stage I oocytes compared to stage VI oocytes. Maternal mRNAs are repressed by multiple mechanisms throughout development (Colegrove-Otero et al., 2005), and during oogenesis it is known that the mechanism of CPEB-mediated repression changes as different partner proteins are expressed at different stages (Radford et al., 2008). Therefore, it is conceivable that Zar2 could provide yet another mechanism of translational repression early in oogenesis. Identification of additional Zar2 target mRNAs and their regulation during oogenesis and embryogenesis will help elucidate the physiological role of Zar2. 147 maternal Xenopus 3’UTRs were identified as being candidates for containing TCSs (Wang et al., 2008). Testing whether those mRNAs can interact with Zar2 will advance this goal.
In this report, we have shown that Xenopus Zar2 binds to the TCS in maternal mRNAs and represses translation in immature oocytes. Because of the areas of conservation in the C-terminal, we expect Zar1 to also bind RNA. There is also homology between Zar1 and Zar2 in the N-terminal domain and so we expect Zar1 to also regulate translation. Expression of Zar1 and Zar2 is normally confined to oocytes and early embryos. However, recently, it has been reported that Zar transcripts and Zar proteins are expressed in a variety of cancers. Zar2 is expressed in human breast cancer cells where it binds to the BRCA2 gene promoter and represses expression of BRCA2 during G0/G1 (Misra et al., 2010). In addition, the Zar1 gene is aberrantly methylated and over-expressed in malignant melanoma cells (Shinojima et al., 2010). The Zar1 gene is also hypermethylated in glioma cell lines and brain tumors (Watanabe et al., 2010). Understanding the molecular function of Zar proteins will help us understand their roles in early development and in pathological conditions.
We would like to thank Dr. Jeff Coller, Case Western Reserve University, OH; Dr. Nancy Standart, University of Cambridge, UK; and Dr. Marvin Wickens, University of Wisconsin, Madison for reagents. We also thank Lori Sherman, Protein Production/Mab/Tissue Culture Core Manger, University of Colorado Cancer Center for bacluovirus infection of Sf9 cells for protein purification. We would like to thank Dr. Mike Wunder for statistical advice, and Drs. Brad Stith and Aimee Bernard for critical reading of this manuscript (Dept. Integrative Biology, University of Colorado Denver). We would like to thank RNA Club at University of Colorado Denver Anschutz Medical Campus for intellectual and technical support. This work was funded in part by ACS RSG 0804401, NIH/NCRR RR20146, UAMS pilot study awards and University of Colorado Denver start up funds (to AC); and NIH grant R01 HD35688, ACS RPG 101279 and the Arkansas BioSciences Institute (to AMM). The DNA samples were sequenced by the University of Colorado Cancer Center DNA Sequencing and Analysis Core (http://DNASequencingCore.ucdenver.edu), which is supported by a NIH/NCI Cancer Center Core Support Grant (P30 CA046934).
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.