The RNA-binding protein DAZL functions as repressor and activator of mRNA translation during oocyte maturation

Abstract

Deleted in azoospermia-like (DAZL) is an RNA-binding protein critical for gamete development. In full-grown oocytes, the DAZL protein increases 4-fold during reentry into the meiotic cell cycle. Here, we have investigated the functional significance of this accumulation at a genome-wide level. Depletion of DAZL causes a block in maturation and widespread disruption in the pattern of ribosome loading on maternal transcripts. In addition to decreased translation, DAZL depletion also causes translational activation of a distinct subset of mRNAs both in quiescent and maturing oocytes, a function recapitulated with YFP-3'UTR reporters. DAZL binds to mRNAs whose translation is both repressed and activated during maturation. Injection of recombinant DAZL protein in DAZL-depleted oocytes rescues the translation and maturation to MII. Mutagenesis of putative DAZL-binding sites in these mRNAs mimics the effect of DAZL depletion. These findings demonstrate that DAZL regulates translation of maternal mRNAs, functioning both as the translational repressor and activator during oocyte maturation.

Fig. 1. Interference with Dazl mRNA translation depletes oocytes of the DAZL protein and inhibits translation of a specific downstream target.

a DAZL protein accumulation during the transition from the GV-to-MI stages of oocyte maturation. GV stage oocytes from wild type mice were cultured in vitro up to 8 h of maturation and used for Western blot analysis. 150 oocytes per lane was loaded on the gel and accumulation of α-tubulin was used as a loading control. The complete time course was performed once, while changes between the first and last time point confirmed in two additional independent experiments (see Supplementary Fig 1a). b Morpholino oligonucleotide (MO) down-regulation of DAZL protein. GV stage oocytes from RiboTag^fl/fl:Zp3-Cre:WT or RiboTag^fl/fl:Zp3-Cre:DAZL^+/− mice were injected with CON-MO or DAZL-MO as described in the methods. After 6 h, oocytes were collected and used for Western blot analysis. A representative experiment of the three performed is reported. c–e GV stage oocytes from RiboTag^fl/fl:Zp3-Cre or RiboTag^fl/fl:Zp3-Cre:Dazl^+/− mice were injected with CON-MO or DAZL-MO and preincubated overnight in 2 µM milrinone, then cultured in inhibitor-free medium for maturation. Oocytes were collected at 0 and 6 h for RiboTag IP followed by qPCR analysis. Ribosome loading of the endogenous Dazl mRNA (unpaired two tailed t-test, p = 0.0022) and Tex19.1 (unpaired two tailed t-test, p < 0.0001), but not CcnB1 (unpaired two tailed t-test, P = 0.8567) mRNA is disrupted after DAZL depletion. (GV: germinal vesicle; MI: Meiosis I. Each dot represents an independent biological sample collected from different experiments done in different days. f–h Input data for the immunoprecipitation reported in c–e. When N > 2, data are presented as Mean ± SEM and each dot represents an independent biological observation. Statistical significance among the different samples was evaluated with the Brown-Forsythe ANOVA test, assuming unequal SD. Calculated P values: Dazl = 0.6680, Tex19.1 = 0.9713, CcnB1 = 0.5404.

Fig. 2. Maternal mRNA loading onto ribosome is disrupted in oocytes depleted of DAZL.

a Comparison of transcriptomes of MO injected oocytes. Oocytes from RiboTag^fl/fl:Zp3-Cre:WT mice were injected with a CON-MO whereas oocytes from RiboTag^fl/fl:Zp3-Cre:Dazl^+/− mice were injected with a DAZL-MO. Oocytes were preincubated overnight in the presence of milrinone and the following morning were collected for RiboTag IP/RNA-Seq as described in the “Methods”. The average counts per million reads (CPM) for input (total transcripts) from two biological replicates is reported. b Comparison of transcripts recovered by RiboTag IP/RNA-Seq in Control and DAZL-MO injected oocytes (MI). GV stage oocytes from RiboTag^fl/fl:Zp3-Cre:WT or RiboTag^fl/fl:Zp3-Cre:Dazl^+/− mice were injected with CON-MO or DAZL-MO. After overnight preincubation with 2 µM milrinone, oocytes were cultured in inhibitor-free medium to allow reentry into the meiotic cell cycle. Oocytes were collected at 6 h for RiboTag IP and RNA-Seq analysis as detailed in the methods. Ribosome loading of significantly decreased transcripts (551 transcripts; FDR < 0.05, 2-fold change) is reported in blue, while ribosome loading of transcripts significantly increased (170 transcripts; FDR < 0.05, 2-fold change) after DAZL removal is reported in red. Transcripts not significantly changed are depicted in grey. Two independent biological samples were used in this analysis. c–e Effect of DAZL depletion on total RNA levels and ribosome loading of representative DAZL interacting targets. c, d GV stage oocytes from RiboTag^fl/fl:Zp3-Cre:WT or RiboTag^fl/fl:Zp3-Cre:Dazl^+/− mice were injected with CON-MO or DAZL-MO. The experimental conditions and data from analysis are as in b. The bar represents the mean and each dot represents the CPM from the two biological replicatesRNA-Seq data from supernatants (input) before IP (c) and RiboTag IP of CON-MO and DAZL-MO (d) are reported. e DAZL RIP-Chip of oocyte extracts of selected mRNAs is reported as the fold enrichment DAZL AB/IgG; Data are presented as mean values ±SEM of three biologically independent samples except for NSF1 where N = 2.

Fig. 3. RiboTag IP/qPCR confirms the presences of a subset of transcripts whose translation is upregulated and downregulated in oocytes depleted of DAZL.

Representative mRNA targets affected by DAZL removal. RiboTag/RNA-Seq data are reported in a down and b up. Ribosome loading DAZL-MO/CON-MO for the same transcripts assessed in independent biological replicates by RiboTag IP/qPCR is reported in c, down and d up. Data are reported as ratio DAZL KD/control. Dppa3 and CcnB1 are used as negative control in a, c. Gdf9 mRNA is used as negative control in b, d. Experimental conditions are identical to those described in Fig. 2b except for the qPCR analysis. Bars are the average of two or more observations. P values in panel C were calculated with the Bonferroni-Dunn method, with alpha = 0.05. Each gene was analyzed individually, without assuming a consistent SD. P values are the following: Dppa3 = 0.851918, Dazl = 0.015042, Tex19.1 = 0.035962, Btg4 = 0.069233, Rad51C = 0.059755, Ireb2 = 0.006635, Tcl1 = 0.032710, CcnB1 = 0.327854.

Fig. 4. DAZL physically interacts with transcripts whose translation is upregulated or downregulated during oocyte maturation.

a Comparison DAZL Rip-Chip and polysome array in oocytes. Changes in ribosome loading from 0 h (GV) to 16 h (MII stage) is reported in the y axes, while the interaction with DAZL assessed by RIP-Chip is reported in the X axes. Subset of transcripts whose translation increased from GV to MII stage are also enriched in DAZL immunoprecipitates of oocyte extracts (red); transcripts whose translation decreases during oocyte maturation and specifically immunoprecipitated by DAZL antibody (blue). The GV/MII ribosome loading data are those published in ref. . For DAZL RIP-Chip, oocytes were obtained from mice primed with PMSG and hCG. Oocyte lysates was immunoprecipited with DAZL-specific antibody or IgG and the mRNA recovered in the IP pellet measured by microarray hybridization. Each dot is the average of three independent biological replicates (b) Interaction of DAZL with selected targets upregulated or downregulated in oocytes after DAZL depletion. Oocytes were injected with a V5-tagged Dazl construct. After overnight incubation in milrinone, oocytes were extracted and an aliquot was used for WB (inset) or used for RIP/qPCR. Data are presented as bars with means ±SEM of n = 3 biologically independent experiments except for Btg4 where N = 2.

Fig. 5. The 3′ UTR of Oosp1, Obox5 and Cenpe recapitulates the effect of DAZL depletion on endogenous mRNA translation.

Oocytes were co-injected with ployadenylated mCherry mRNA and oligoadenylated YFP-Oosp1 3′UTR, YFP-Obox5 3′UTR, or YFP-Cenpe 3′UTR reporters with either CON-MO or DAZL-MO. Oocytes were then pre-incubated overnight to allow the mCherry signal to reach a plateau. At the end of the preincubation, oocytes were released in cilostamide-free medium for maturation and YFP and mCherry signal were recorded by time lapse microscopy. The YFP signal was analyzed using Metamorph software (version 7.8.13.0) corrected by the level of coinjected mCherry signal and normalized to the first recording of YFP/mCherry. Experiments were repeated at least three times and the data are the cumulative mean ± SEM of the three or more independent experiments. Individual oocyte YFP/mCherry signals were used to calculate the rate of translation of the reporters at the 0–2 h (prior to GVBD) and 4–8 h (after GVBD) as detailed in the Methods. The differences between rates of YFP-Oosp1 (b) YFP-Obox5 (d) and YFP-Cenpe (f) were analyzed by unpaired two tailed t test.

Fig. 6. The translation of the YFP-Oosp1 or YFP-Obox5 reporter in the DAZL-depleted oocytes is rescued by injection of DAZL protein.

a Human DAZL protein injection restores Oosp1 translation during oocyte maturation in MO injected oocytes. Oocytes were injected with mCherry mRNA and YFP-Oosp1 3′ UTR reporter with either CON-MO or DAZL-MO, with or without recombinant human DAZL protein, and incubated in cilostamide containing medium overnight to allow mCherry signal to reach a plateau. At the end of the preincubation, oocytes were released in cilostamide-free medium for maturation and YFP and mCherry signal recorded by time lapse microscopy. YFP signal were corrected by the level of co-injected mCherry signal and were normalized to the first time point. Experiments were repeated three times and the data are the mean ± SEM of three independent biological samples. b Rates of translation measured at 0–2 h and 6–10 h of maturation for the time courses reporter in a. Statistical analysis was done by unpaired two tailed t-test, p < 0.0001). c, d Similar experiments done the with Obox5 reporter. Data in d were analyzed by unpaired two tailed t-test, p < 0.0001). e Microinjection of a human DAZL protein rescues the meiotic block of oocytes injected with DAZL- MO. Bars are the mean ± SEM from three biologically independent experiments. Significance was calculated by unpaired two tailed t-test, p < 0.0001). f Representative Image of oocytes from the three conditions reported in e, at least n = 3 biologically independent experiments were performed. Oocytes maturation was scored by counting the number of oocytes with a polar body. Scale bar = 100 µm.

Fig. 7. DAZL depletion increases ribosome loading of Oosp1 and Obox5 endogenous transcripts and translation of the Oosp1 and Obox5 reporters in GV-arrested oocytes.

a, c, e GV stage oocytes from RiboTag^fl/fl:Zp3-Cre:WT or RiboTag^fl/fl:Zp3-Cre:Dazl^+/− mice were injected with CON-MO or DAZL-MO. Oocytes were pre-incubated overnight with 2 µM milrinone. In all, 0 h (GV stage) data from RiboTag IP/ RNA-Seq were used to calculate the Translational Efficiency (TE). TE is calculated as the ratio of the CPM for HA immunoprecipitated transcripts Oosp1 or Obox5 over the corresponding input at 0 h. The TEs are reported for Oosp1 (a) and Obox5 (c) or CcnB1 (e). Data are presented as mean of two biologically independent experiments for a, c, e. b, d, f GV stage oocytes were injected with mCherry-polyadenylated mRNA and YFP-3′UTR reporter for Oosp1 3′ UTR or Obox5 3′UTR with either CON-MO or DAZL-MO. Oocytes were pre-incubated overnight to allow mCherry signal to plateau, then further incubated in cilostamide-containing medium. YFP signal were corrected by the level of coinjected mCherry signal. Reporters injected are: b Oosp1 (unpaired two tailed t-test, p < 0.0001), d Obox5 (unpaired two tailed t-test, p = 0.0007) and f CcnB1 (unpaired two tailed t-test, p = 0.2274). Data are presented as mean values ± SEM of rates for each individual oocytes from n = 3 independent biologically experiments for b, d, f.

Fig. 8. Translation of Oosp1 and Obox5 reporter is dependent on the presence of a DAZL binding element.

Wild type Oosp1 3′UTR or Obox5 3′UTR as well as constructs with mutations in the DAZL binding site were injected along with mCherry-polyadenylated mRNA into GV stage oocytes. After overnight pre-incubation to allow mCherry signal to plateau, oocytes were released in cilostamide-free medium and signals recorded. YFP signal were corrected by the level of co-injected mCherry signal. Black symbols: control; Red symbols: mutant constructs. Experiments were repeated three times and the number of oocytes recorded is reported. a Scheme of the Oosp1 and Obox5 3′ UTR and position of the PAS, putative CPEB1 and DAZL-binding elements. Mutagenesis of the putative DAZL-binding element was performed as detailed in “Methods” section. A red oval represents the DAZL consensus sequence in the 3′UTR of Oosp1 and Obox5. A black cross indicates the mutated DAZL-binding consensus sequence. b, e The effect of DAZL-binding element mutation on Oosp1 (unpaired two tailed t-test, p = 0.0062) or Obox5 (unpaired two tailed t-test, p < 0.0001) translation in GV stage. Rates of reporter accumulation were calculated for each GV-arrested oocyte and plotted as individual dots. Mean and SEM values were calculated from three independent experiments. c, f Mutation of DAZL-binding element on 3′UTR of Oosp1 or Obox5 decreases translation of reporters during oocyte maturation. YFP signals were corrected by the co-injected mCherry signal. Each time point was normalized to the first YFP:mCherry ratio. d, g Rates of YFP-Oosp1 or YFP-Obox5 reporter accumulation compared to a wild type reporter calculated from c, f (unpaired two tailed t-test, p < 0.0001).

Fig. 9. DAZL positive or negative effects on translation depend on the 3′UTR context and involve an interaction with the CPEB1-dependent adenylation.

a Scheme summarizing the domain organization of the two long and short Oosp1 variant 3′UTRs expressed in mouse oocytes. PAS: Cleave and polyadenylation signal; Dazl: consensus DAZL binding sequence; CPE: CPEB1 binding consensus sequence. The mutated element is marked by an X. b, c Comparison of the effect of DAZL element mutagenesis on the translational activation of Oosp1 reporter during maturation. The Oosp1 long and short wild type (WT, no mutation) and mutant (ΔDazl) reporters were generated by fusing the two 3′UTRs to a YFP sequence and were injected in GV oocytes together with the mCherry constitutive reporter. After overninght recovery, oocytes were transferred to maturation medium and fluorescence recorded by time lapse microscopy. Each experiment was repeated three or more times in different days and oocyte recordings were combined. Each point is the Mean ± SEM of single oocyte measurements. The total number of oocytes used for each construct is reported among brackets. d Effect of single or combined CPE and DAZL binding element mutagenesis on polyadenylation and translational rate of the Oosp1-short 3′UTR reporter. Oligo- or poly-adenylated wild type and mutant reporters were generated as described in the Methods and injected in GV oocytes. After 3 h recovery, fluorescent recordings were monitored by time lapse microscopy in oocytes maintained in GV. Rates of reporter translation were calculated for each injected oocyte at the beginning (0–3 h.) and at the end (7–10 h.) of the incubation (see Methods for details). The average ± SEM from two experiments was calculated for each group of measurements and the statistical significance was calculated by T-test with Welch correction. To estimate the adenylation or deadenylation of the reporter, the ratio between 0–3 h and 7–10 h was calculated for each oocyte and the log2 average reported below the measurements (Upper panel: unpaired two tailed t-test, WT p = 0.6985, ΔCPE p < 0.0001, ΔDazl p = 0.3536, ΔCPE + ΔDazl p < 0.0001; Lower panel: unpaired two tailed t-test, WT p < 0.0001, ΔCPE p < 0.0001, ΔDazl p < 0.0001, ΔCPE + ΔDazl p = 0.0912). e Consequences of single or combined mutation of the CPE and DAZL element on the translation of the Oosp1-short reporter during oocyte maturation. The experimental design is as in c. Each point is the mean ± SEM of single oocyte measurements from three independent experiments. The total number of oocyte used in the experiment is reported among brackets.