Roles of NPM2 in Chromatin and Nucleolar Organization in Oocytes and Embryos

See allHide authors and affiliations

Science  25 Apr 2003:
Vol. 300, Issue 5619, pp. 633-636
DOI: 10.1126/science.1081813


Upon fertilization, remodeling of condensed maternal and paternal gamete DNA occurs to form the diploid genome. In Xenopus laevis, nucleoplasmin 2 (NPM2) decondenses sperm DNA in vitro. To study chromatin remodeling in vivo, we isolated mammalian NPM2 orthologs. Mouse NPM2 accumulates in oocyte nuclei and persists in preimplantation embryos. Npm2 knockout females have fertility defects owing to failed preimplantation embryo development. Although sperm DNA decondensation proceeds without NPM2, abnormalities are evident in oocyte and early embryonic nuclei. These defects include an absence of coalesced nucleolar structures and loss of heterochromatin and deacetylated histone H3 that normally circumscribe nucleoli in oocytes and early embryos, respectively. Thus, Npm2 is a maternal effect gene critical for nuclear and nucleolar organization and embryonic development.

A crucial step in zygotic development is the decondensation and reorganization of chromatin of male and female gametes, a process wholly dependent on factors produced during oogenesis. In vitro, Xenopus (x) NPM2 (NPL), an oocyte-specific nuclear protein, removes sperm protamines and facilitates nucleosome assembly and replication of the paternal genome (16). Although nucleoplasmin and related chaperones involved in chromatin remodeling are predicted to be conserved throughout evolution because oocytes can efficiently remodel heterologous sperm or somatic cell nuclei into pronuclei (79), mammalian NPM2 orthologs have until now eluded researchers.

To identify novel genes that function in oogenesis and as maternal effect genes, we engineered a polymerase chain reaction suppression-subtraction ovary library (10). Northern blot analysis with one sequence encoding the mouse (m) ortholog of NPM2 detected an ovary-specific mRNA transcript of 1.0 kb; in situ hybridization revealed expression limited to growing oocytes (Fig. 1, A and B). NPM2 protein is compartmentalized to the oocyte nucleus before germinal vesicle (nuclear membrane) breakdown (GVBD), but after GVBD, NPM2 redistributes into the cytoplasm (Fig. 1, C to E). After fertilization, NPM2 is found in both pronuclei, remains in the nuclei of early cleavage embryos through the eight-cell stage (Fig. 1, F and G), and is barely detectable in blastocysts.

Fig. 1.

Npm2 mRNA and NPM2 protein expression. (A) Northern blot analysis of mouse Npm2 and 18S rRNA (loading control). Br, brain; Lu, lung; He, heart; St, stomach; Sp, spleen; Li, liver; SI, small intestine; Ki, kidney; Te, testes; Ut, uterus; Ov, wild-type ovary; –/–, Gdf9/ ovary. (B and C) In situ hybridization (B) and immunohistochemistry (C) of ovaries detect Npm2 mRNA and protein (red staining), respectively, in all growing oocytes from the primary follicle (arrowhead) through antral follicle (double arrowhead) stages. (D to G) Immunofluorescence analysis of oocytes from 10-day-old mice (D) and adult pregnant mare serum gonadotropin (PMSG)-treated mice (E) shows that NPM2 protein is in the nucleus and excluded from the nucleolus (arrowhead). After fertilization, NPM2 localizes to pronuclei and polar bodies of one-cell zygotes (F) and persists through the eight-cell stage (G).

The 207–amino acid mNPM2 shares 39.5% identity with xNPM2. The subsequently identified human (h) and rat (r) NPM2 proteins have 61.4 and 81.6% identity with mNPM2, respectively (Fig. 2A). NPM2 orthologs share a conserved bipartite nuclear localization signal consisting of KR-(X)10-KKKK (5, 11), a stretch of negatively charged glutamic acid (E) and aspartic acid (D) residues implicated in binding protamines and histones, and many serines and threonines (28 in hNPM2, 36 in mNPM2, 37 in rNPM2, and 24 in xNPM2) that are putative phosphorylation sites.

Fig. 2.

Amino acid conservation of NPM2 and creation of Npm2 mutant mice. (A) Interspecies amino acid identity is highlighted in black. The conserved bipartite nuclear localization sequence is indicated by asterisks (*); a line is drawn over the acidic putative histone binding region. (B) An Npm2 targeting vector was constructed to delete exon 2 (including initiation ATG), exon 3, and the splice acceptor junction of exon 4 of the nine-exon Npm2 gene (see also figs. S1 to S3). B, Bam HI; Bg, Bgl II; P, Pst I; X, Xba I. (C and D) Southern blot analysis of a litter from Npm2+/ intercrosses indicated that homozygous null (–/–) mice were generated.

To study the roles of NPM2 in mammalian fertilization and early embryo development, we generated an Npm2-null allele using embryonic stem cell technology (12) (fig. S1 to S3; Fig. 2, B to D). F1 heterozygous mice (Npm2tm1Zuk/+; herein called Npm2+/) are viable and fertile. Genotype analysis of 230 F2 offspring from heterozygote intercrosses indicated a normal Mendelian ratio of 1:2:1 (54 wild-type, 124 Npm2+/, 52 Npm2/), and equal numbers of male and female homozygotes (Npm2/) were produced. Therefore, Npm2/ mice are viable and have normal sexual differentiation. Npm2 is not expressed in males, and Npm2/ male mice are fertile and normal.

Female Npm2/ mice are subfertile or infertile. Intercrosses of nine female and male Npm2+/ mice over 6 months resulted in a normal number of litters (n = 54; 1.00 ± 0.06 litters per month), with 8.98 ± 0.31 offspring per litter. In contrast, only 11 of 14 Npm2/ females became pregnant over this period, yielding 40 litters (0.48 ± 0.11 litters per month) with 2.65 ± 0.24 offspring per litter. Npm2/ ovaries demonstrated grossly normal oogenesis and folliculogenesis. Pharmacological superovulation resulted in similar numbers of eggs ovulated in Npm2/ and control females. In vitro, maturation and fertilization of Npm2-null eggs are apparently normal, but there is reduced cleavage to the two-cell stage (table S1). In vivo, fertilized Npm2-null eggs can be recovered 24 hours after human chorionic gonadotropin (hCG) treatment. However, mostly asynchronously fragmenting and dying embryos are found 45 to 55 hours after hCG treatment (Fig. 3, A, C to F), and unlike controls, few Npm2-null embryos can be cultured to the blastocyst stage (Fig. 3, B, G and H). Nuclei of these dying embryos are TUNEL (terminal deoxynucleotidyl transferase–mediated dUTP nick-end labeling)–positive (Fig. 3, K and L), although there is no evidence that DNA damage causes embryo loss, because one-cell embryos collected from null females 20 hours after hCG treatment exhibit TUNEL staining only within polar bodies (Fig. 3, I and J). All developmental defects occur when eggs are fertilized with wild-type spermatozoa, indicating that maternal NPM2 is crucial in early embryogenesis.

Fig. 3.

Egg analysis, culture, and fluorescent staining. (A and B) The percent cleavage of in vivo–fertilized embryos to various stages is shown after oviduct collection (A) and subsequent 24-hour culture (B). Times are given as hours after hCG treatment. (C to F) Most fertilized eggs from wild-type mice form two-cell and four-cell embryos (arrows) by 45 and 55 hours after hCG treatment, respectively, whereas few Npm2-null eggs cleave to form multicellular embryos, and even fewer form blastocysts (H) as compared with wild-type controls (G). Wild-type (I) and Npm2/ (J) fertilized eggs are TUNEL-negative, with the exception of their TUNEL-positive (red) polar bodies. (K and L) By 55 hours after hCG, DNA within fragmenting Npm2-null embryos stains TUNEL-positive. (M) Transcription-requiring complex (TRC) proteins were extracted from wild-type (WT) and Npm2-null (–/–) two-cell embryos after culture in 35S-labeled methionine. As a negative control, actinomycin D (ActD) inhibited transcription and TRC production.

To study these developmental defects further, we examined oocytes and eggs from Npm2/ females. Oocytes from PMSG-treated wild-type females exhibit an organization of heterochromatin surrounding the prominent nucleolus, termed the SN (surrounded nucleolus) configuration (Fig. 4C). The SN configuration is characteristic of advanced oocyte development, because SN oocytes are larger and are found in gonadotropin-dependent follicles. The condensation of chromatin correlates with transcriptional silencing, competence to resume meiosis, the appearance of M-phase characteristics, and postfertilization embryo developmental potential (1316). In contrast to wild-type oocytes, the DNA in Npm2-null oocytes is amorphous and diffuse with no condensation around the nucleolus (Fig. 4D). The loss of nucleolar clearing is also illustrated by immunofluorescence to detect acetylated histone H3 in these oocytes (Fig. 4, E and F), as well as the less mature non-SN oocytes isolated from 10-day-old untreated mice (Fig. 4, A and B). Immunofluorescence to localize the nucleolar protein fibrillarin demonstrates dispersed nucleolar-like bodies in Npm2-null oocytes as compared with the single organized nucleolus observed in controls (Fig. 4, G and H). These anomalies were observed in hundreds of oocytes examined from more than 30 Npm2/ females. Thus, NPM2 is essential for organization of oocyte nuclear and nucleolar domains and the compaction of oocyte chromatin during the final stages of oocyte development.

Fig. 4.

Analysis of wild-type (WT) and mutant oocytes and embryos. Immunofluorescence analysis of wild-type or Npm2-null oocytes (A to J), one-cell embryos (K to V), or eight-cell embryos (W to Z) was performed with the indicated antibodies. DNA was counterstained with DAPI (4′,6-diamidino-2-phenylindole) (A to L, O and P, and S to Z) or To-pro-3 (Q and R). Antigens are detected as magenta (A and B, E to H, K and L, O and P, S and T), orange (M and N), or green color (I and J, U to X). In (I and J) and (U and V), the mitotic spindle is green and metaphase chromosomes are blue. BrDU, bromodeoxyuridine.

Meiosis progresses essentially normally in the absence of NPM2, with no obvious defects in metaphase II arrest, spindle formation, chromosomal segregation, or extrusion of polar bodies (Fig. 4, I and J). Sperm DNA decondensation occurs normally without NPM2; fertilization is followed by the formation of both maternal and paternal pronuclei surrounded by nuclear envelopes (Fig. 4, K and L), and there is no persistent protamine 2B detectable in male pronuclei (17). DNA replication in the first S phase also proceeds without NPM2 (Fig. 4, M and N). However, as in Npm2/ oocytes, normal nucleoli are absent in Npm2-null one-cell embryos, and immunofluorescence to detect acetylated histone H3 in zygotes shows no nucleolar clearing as compared with controls (Fig. 4, O and P). Hypoacetylated histone H3, which is normally associated with compact chromatin rimming pronuclear nucleoli, was undetectable (Fig. 4, Q and R). Treatment with colcemid to inhibit spindle formation arrests both wild-type (n = 34) and Npm2-null (n = 28) one-cell embryos in metaphase, with condensed chromosomes staining for phosphorylated histone H3 [H3-P(Ser10)]; this indicates that the first mitosis initiates in essentially all cases (Fig. 4, S and T). Without colcemid treatment, spindle forms in wild-type zygotes 13 to 15 hours after pronuclear formation (Fig. 4U), and all zygotes complete mitosis by 19 hours. In contrast, Npm2-null zygotes are observed with metaphase spindle from 13 to 19 hours after pronuclear formation (Fig. 4V) and immediately preceding fragmentation, suggesting abnormal exit from the first mitosis. A few multicellular embryos lacking NPM2 can be recovered, and their nuclei contain somatic linker histone H1 (Fig. 4, W and X); however, their nuclei remain relatively amorphous until the blastocyst stage (Fig. 4, Y and Z). Thus, mammalian NPM2 is crucial for histone deacetylation and heterochromatin formation surrounding nucleoli in oocytes and early zygotes.

The cause of mitotic failure in Npm2-null zygotes and whether this represents an acute event remain unclear. Oocyte viability and the progression of meiosis may be resistant to insults capable of causing arrest in mitotic cell divisions (18). Despite disruption of nucleolar structure, there was no detectable alteration in 18S and 28S ribosomal RNA (rRNA) levels or changes in the absolute levels of protein synthesis in oocytes and one-cell embryos from Npm2-null females (figs. S4 and S5). Transcriptional deregulation may be involved, although neither chemical transcription inhibitors nor inhibitors of histone deacetylases block the first mitotic division in mouse embryos (19). Two-cell embryos lacking NPM2 synthesize the transcription-requiring complex (TRC) of proteins, indicating some zygotic gene transcription and translation (20), albeit at reduced levels (30%) as compared with wild-type two-cell embryos (Fig. 3M). NPM2 could function in the translational activation of specific maternal RNAs in early embryos, as has been proposed for xNPM2 (21). Because a few Npm2-null embryos develop to birth, potential compensatory mechanisms must exist. Transcriptional activation of paternal Npm2 is not involved because Npm2/ males can sire pups when mated with Npm2/ females. Despite the low homology in the primary amino acid sequences of NPM2 and other “ubiquitous” nucleoplasmins (2224), these more widely expressed nucleoplasmins are karyophilic, negatively charged proteins that may share functional redundancy with NPM2 in oocytes and developing embryos. In situ hybridization reveals that both nucleophosmin 1 (Npm1; B23) and nucleoplasmin 3 (Npm3) mRNAs are expressed in mouse oocytes (fig. S6), although the mechanisms by which these or other proteins may compensate for absence of Npm2 are not clear.

NPM2 represents one of a small number of oocyte-derived proteins with demonstrated functions in postfertilization development in mammals; others include MATER, oocyte DNMT1, HSF1, PMS2, and ZAR1 (2530). Of these maternal effect genes, only Npm2 and Zar1 are critical for the one-cell to two-cell transition. In the future, it will be important to define the functional domains of NPM2, investigate the roles of xNPM2 in post-fertilization development, and assess whether NPM2 gene mutations cause infertility in women.

Supporting Online Material

Materials and Methods

Figs. S1 to S6

Table S1


  • * These authors contributed equally to this work.

References and Notes

View Abstract

Navigate This Article