Histone variant H3.3 is an essential maternal factor for oocyte reprogramming Duancheng Wena,b,c,1, Laura A. Banaszynskid,1, Ying Liua,b, Fuqiang Genga,b, Kyung-Min Nohd, Jenny Xiange, Olivier Elementof, Zev Rosenwaksc, C. David Allisd,2, and Shahin Rafiia,b,2 aDepartment of Genetic Medicine, Ansary Stem Cell Institute, bHoward Hughes Medical Institute, cRonald O. Perelman and Claudia Cohen Center for Reproductive Medicine, eGenomics Resources Core Facility, and fDepartment of Physiology, Weill Cornell Medical College, New York, NY 10065; and dLaboratory of Chromatin Biology and Epigenetics, The Rockefeller University, New York, NY 10065 Contributed by C. David Allis, April 11, 2014 (sent for review January 30, 2014) Mature oocyte cytoplasm can reprogram somatic cell nuclei to the importance of H3.3 in oocyte fertilization, we sought to determine pluripotent state through a series of sequential events including whether the H3.3 variant might also be a maternal “reprogramming protein exchange between the donor nucleus and ooplasm, factor” and whether it might play a specialized role during somatic chromatin remodeling, and pluripotency gene reactivation. Ma- cell reprogramming. ternal factors that are responsible for this reprogramming process Here, we show that maternal H3.3 is critical for the development remain largely unidentified. Here, we demonstrate that knock- of SCNT embryos and for the reactivation of many key pluri- down of histone variant H3.3 in mouse oocytes results in potency genes. We demonstrate maternal H3.3 remodeling of donor compromised reprogramming and down-regulation of key pluri- nuclear chromatin through replacement of donor nucleus-derived potency genes; and this compromised reprogramming for de- H3 with de novo synthesized H3.3 protein, with overall replacement velopmental potentials and transcription of pluripotency genes levels dependent on the identity of the donor nucleus. can be rescued by injecting exogenous H3.3 mRNA, but not H3.2 mRNA, into oocytes in somatic cell nuclear transfer embryos. We Results show that maternal H3.3, and not H3.3 in the donor nucleus, is Quantitative RT-PCR (RT-qPCR) demonstrated that the cyto- essential for successful reprogramming of somatic cell nucleus into plasm of mouse MII oocytes is abundant with h3f3a and h3f3b the pluripotent state. Furthermore, H3.3 is involved in this transcripts (H3.3A and H3.3B; Fig. 1A). These maternal tran- reprogramming process by remodeling the donor nuclear chroma- scripts become largely depleted by the first embryonic cleavage tin through replacement of donor nucleus-derived H3 with de (20 h after oocyte activation), whereas zygotic H3.3 is elevated novo synthesized maternal H3.3 protein. Our study shows that after the two-cell stage (Fig. 1B). We microinjected siRNAs H3.3 is a crucial maternal factor for oocyte reprogramming and against H3.3A and H3.3B (siH3.3; 10 μM) into MII oocytes by provides a practical model to directly dissect the oocyte for its using a fine capillary (3–5 μm in diameter), and found that ma- reprogramming capacity. ternal H3.3 mRNAs were significantly decreased and zygotic H3.3 was also suppressed up to 44 h after oocyte activation (Fig. BIOLOGY ioneering nuclear transfer experiments in amphibians have 1 C and D). RT-qPCR analysis of H3.3A, H3.3B, H3.1, and H3.2 DEVELOPMENTAL Prevealed that the cytoplasm of the egg is able to reprogram in the oocytes after siH3.3 injection demonstrated the specificity a differentiated nucleus to the embryonic state (1, 2). The suc- of H3.3 knockdown vs. the canonical histones (Fig. 1C). We next cess of somatic cell nuclear transfer (SCNT) to produce cloned tested a serial dilution of siH3.3 concentrations for knockdown animals using enucleated metaphase II (MII) oocytes (3, 4), and, efficiency (the oocyte poses an upper limit to the injected volume recently, the successful derivation of SCNT human embryonic that does not induce lysis; thus, the amount of siRNA that can be stem cells (5), have demonstrated that maternal factors in the mature ooplasm are capable and sufficient to reprogram a dif- Significance ferentiated cell nucleus to pluripotency. This process is known to involve a series of sequential events including protein exchange A differentiated cell nucleus can be reprogrammed into the between donor nucleus and ooplasm, donor nuclear chromatin pluripotent state by maternal factors in ooplasm; the factors remodeling, and pluripotency gene reactivation (6–12). How- that are responsible for this reprogramming process have not ever, maternal factors responsible for this reprogramming pro- yet been identified. In this paper, we show that histone variant cess and the underlying mechanism(s) remain largely unknown. H3.3 is one of the essential maternal factors involved in so- Thousands of different maternal proteins and mRNAs have matic nuclear reprogramming. Maternal H3.3, not H3.3 in the been found in mouse mature oocytes (13, 14), including variants donor chromatin, is required for development and the reac- of the core histone proteins that, along with DNA, constitute tivation of many key pluripotency genes in somatic cell nuclear nucleosomes. Accumulating evidence suggests that histone var- transfer (SCNT) embryos. H3.3 facilitates reprogramming by iants play important roles in chromatin remodeling and epige- remodeling the donor nuclear chromatin through replacement netic regulation orchestrating gene expression changes during of donor H3 in chromatin with de novo synthesized maternal reprogramming (12, 15, 16). In mammals, the histone variant H3.3 at the beginning of reprogramming in SCNT embryos. H3.3 is encoded by two different genes (h3f3a and h3f3b), whose translation results in an identical protein product (17, 18). Un- Author contributions: D.W., L.A.B., Z.R., C.D.A., and S.R. designed research; D.W., L.A.B., like canonical H3 histones that are expressed and incorporated F.G., and K.-M.N. performed research; D.W. and L.A.B. contributed new reagents/analytic into chromatin during S phase, expression of H3.3 is not cell tools; D.W., L.A.B., Y.L., J.X., and O.E. analyzed data; and D.W. and L.A.B. wrote the paper. cycle-regulated, and the variant is expressed in quiescent cells, The authors declare no conflict of interest. postmitotic cells, and proliferating cells throughout the whole Data deposition: The data reported in this paper have been deposited in the Gene Ex- cycle, enabling H3.3 deposition in a DNA synthesis-independent pression Omnibus (GEO) database, www.ncbi.nlm.nih.gov/geo (accession no. GSE56762). manner during and outside of S phase (19). It has been suggested 1D.W. and L.A.B. contributed equally to this work. that maternal H3.3 plays an important role in male pronucleus 2To whom correspondence may be addressed. E-mail: [email protected] or srafii@ formation and male genome epigenetic reprogramming; indeed, med.cornell.edu. maternal H3.3 is incorporated in the decondensing sperm nu- This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10. cleus as early as 1 h after fertilization (20–23). Considering the 1073/pnas.1406389111/-/DCSupplemental. www.pnas.org/cgi/doi/10.1073/pnas.1406389111 PNAS | May 20, 2014 | vol. 111 | no. 20 | 7325–7330 Downloaded by guest on September 23, 2021 Knockdown of maternal H3.3 in oocytes resulted in signifi- A 2.5 B 1.4 MII oocytes H3.3A cantly compromised embryonic development of parthenogenet- Enucleated oocytes 1.2 2.0 H3.3B 1.0 ically activated (PA) embryos, with arrest at the late morula or 1.5 0.8 early blastocyst stage; uninjected oocytes and oocytes subjected 0.6 to microinjection of an siRNA against luciferase served as con- 1.0 E A 0.4 trols (Fig. 1 , Fig. S2 , and Table S1). We next tested different 0.5 0.2 siH3.3 concentrations on the developmental potential of PA Relative levels to μ 0 unactivated oocytes 0 embryos, and found that injection of four siH3.3 at 4 M each mRNA relative to Gapdh H3.3A H3.3B 0 6 9 12 15 18 21 24 27 31 72 96 was the lowest concentration that significantly decreased blas- Hours after activation tocyst potential (Fig. S2B), whereas injection of 16 μM luciferase C 2.0 H3.3A D 1.2 siRNA did not significantly decrease the developmental potential of H3.3B H3.3A PA embryos (Fig. S2B). A similar phenotype in fertilized embryos 1.6 H3.1 1.0 H3.2 H3.3B was observed after maternal H3.3 knockdown in zygotes (25). 0.8 1.2 To examine whether exogenous H3 mRNAs (that are not the 0.6 0.8 target of siH3.3) could rescue compromised development, we 0.4 injected epitope-tagged H3.3-HA mRNA or H3.2-HA mRNA 0.4 0.2 A C D Relative levels to Relative levels to into oocytes after H3.3KD (Fig. 2 and Fig. S2 and ). We unactivated oocytes 0 unactivated oocytes 0 found that the number of blastocyst stage embryos was signifi- Ctl 4 µM 10 µM 0 6 9 15 21 24 27 44 72 96 H3.3KD Hours after activation cantly increased by injection of exogenous H3.3-HA mRNA, but 24h after activation not H3.2-HA mRNA (Fig. 1E, Fig. S2A, and Table S1), in- E dicating that this phenotype is the result of maternal H3.3 de- 100 pletion, not the reduction of total maternal histone H3 levels in H3.3KD embryos. 80 As the two H3.3 transcripts contain unique untranslated regions 60 containing distinct regulatory elements, we next tested whether 40 * * H3.3A and H3.3B are functionally equivalent during early em- 20 bryonic development. We found no difference in embryonic de- velopmental potential when only H3.3A or H3.3B was knocked 0 E A Ctl Luc KD KD AA BA H3.3 H3.2 down in oocytes (Fig.