Epigenetic regulation of cell differentiation is surprisingly reversible, as demonstrated in many vertebrate species by somatic cell nuclear cloning, a procedure to create genetically identical animals by replacing egg nuclei with somatic cell nuclei (29
). The most striking evidence of this reversibility is the establishment of fertile mouse clones by using nuclei isolated from terminally differentiated lymphocytes and olfactory sensory neurons (24
). Whereas the success rate of mouse cloning is less than 5% (72
), some of the surviving mouse clones have unexpectedly normal gene expression profiles, as shown by proper expression of over 11,000 genes in the placentae and livers of newborn mouse clones (36
). Because no other experimental models with a comparable degree of genomic reversibility exist, with the exception of cell fusion between somatic cells and embryonic stem cells (20
), nuclear cloning provides a valuable opportunity for us to investigate the mechanisms of genome-wide epigenetic reprogramming activities that are important for the future of regeneration medicine. One of the key questions in nuclear cloning is whether a few general reprogramming factors exist that can nonspecifically affect multiple genes in addition to the obviously necessary gene-specific activators and suppressors. Currently, there is no evidence to support the existence of such general reprogramming factors in egg cytoplasm.
Massive nuclear swelling accompanied by global chromatin decondensation is one of the hallmarks of nuclear reprogramming observed in Xenopus laevis
). When somatic nuclei are injected into Xenopus
eggs (meiotic metaphase II), the nuclei swell up to 100-fold in volume within 1 hour, but they do not transcribe genes, reflecting physiological transcriptional silencing in eggs. When injected into oocytes (meiotic prophase), the nuclei swell more slowly, spending 3 days to accomplish the same 100-fold increase in volume (29
), but they remain transcriptionally active during this period. The swollen nuclei in oocytes tend to show more active transcription than those that have not swollen, suggesting that the chromatin decondensation is not merely a morphological event but is also closely linked with an increase in overall nuclear activity. Given the significance of subnuclear compartmentalization and chromosomal domains as regulatory mechanisms for a number of genes (15
), it is not surprising that the nuclear swelling and chromatin decondensation significantly impact the transcriptional status of the donor nuclei in oocyte cytoplasm. Nucleus-wide chromatin decondensation might facilitate reprogramming of the donor nuclei by derepressing condensed chromatin; however, there is a wide knowledge gap between chromatin decondensation at the microscopic level and derepression at the transcriptional level.
Nuclear swelling and chromatin decondensation in egg cytoplasm have mainly been studied in the more physiological context of sperm chromatin decondensation upon fertilization. Xenopus
sperm decondensation is induced by the acidic nuclear protein nucleoplasmin (Npm), which is expressed in oocytes and early embryos (9
). Npm was first purified from Xenopus
eggs as a molecular chaperon which helps load histone onto DNA during nucleosome assembly in vitro (43
). Through its histone-binding capacity, Npm also plays an important role in storing the maternally derived histone, especially histone H2A and H2B, to prepare for early development without zygotic transcription. Later, it was found that during sperm chromatin decondensation upon fertilization, Npm replaces sperm-specific basic proteins X and Y with egg histone H2A and H2B, resulting in assembly of somatic-type nucleosomes onto sperm DNA (59
). This phenomenon has been explained by interactions between the negatively charged Npm and the positively charged X and Y (32
). Hyperphosphorylation of Npm, which occurs during maturation of oocytes into eggs, facilitates the histone replacement on sperm DNA and sperm decondensation (44
), although exact phosphoamino acids have not been identified. Npm can also decondense erythrocyte nuclei by releasing the specialized linker histone H1o
in addition to partial removal of H1 (22
). Thus, chromatin decondensation by Npm has primarily been explained by histone exchanges through charge interactions between acidic Npm and histone or other basic proteins.
Npm consists of 200 or 196 (four residues deleted) amino acids and forms a pentamer in vivo. The Npm monomer is composed of two domains, the N-terminal core domain containing 120 amino acids and the C-terminal tail domain. The core domain has a small acidic amino acid cluster (acidic tract A1) and an eight-stranded β-barrel that forms a wedge shape which provides an interface for pentamer assembly (23
). The core domain binds to histone and is sufficient for in vitro nucleosomal assembly and sperm decondensation (3
). The tail domain contains two larger acidic tracts, A2 and A3, in addition to a bipartite nuclear localization signal. Homologous proteins have been isolated in Drosophila melanogaster
), mice (mNpm2) (10
), and humans (hNpm2) (10
). mNpm2 is composed of 207 amino acids and shares 39.5% identity with Xenopus
Npm at the amino acid level. Burns et al. found that sperm could be decondensed in mNpm2
knockout mouse oocytes, suggesting the existence of functionally redundant proteins in mouse oocytes (10
). The primary defects in the mNpm2
null mice were dispersed nucleoli and female infertility.
We previously established an in vitro nuclear reprogramming assay by combining Xenopus
egg extract and somatic cell nuclei to study the biochemistry of nuclear reprogramming in nuclear cloning (26
). In our current work we applied the assay to study widespread chromatin decondensation in somatic nuclei incubated in Xenopus
egg extract, expecting that the decondensation factor might be one of the general reprogramming factors described above. In this assay we used decondensation of mouse cell centromeres as a convenient indicator to monitor the chromatin decondensation activity. Mouse centromeres are composed of two types of heterochromatin, centromeric heterochromatin and pericentric heterochromatin, each with distinct DNA components (46
). Centromeric heterochromatin contains an approximately 300-kbp tandem repeat of the minor satellite sequence (120-bp repeating unit); pericentric heterochromatin contains an approximately 200- to 2,000-kbp repeat of the major satellite sequence (234-bp unit), depending on the chromosomes (39
). These repeated sequences are also characterized by methylated DNA, histone H3 with trimethylated Lys 9 (tm-H3K9) (57
), and association with heterochromatin proteins HP1α and HP1β through binding to methylated H3K9 (4
). Due to the highly condensed heterochromatin, mouse centromeres are clearly visible as DNA-dye-positive spots in interphase cells by using fluorescence microscopes. Such rich background information in centromeres makes them easy to detect and to characterize their decondensation in egg extract. For simplicity, we will refer to both centromeric and pericentric heterochromatin as centromeric heterochromatin in the following sections.
From our study based on the hypothesis that global decondensation of somatic cell chromatin in egg cytoplasm may facilitate reprogramming of the somatic cell genome, we found that Npm could decondense both euchromatin and centromeric heterochromatin, primarily in undifferentiated mouse nuclei. We show that this decondensation was not accompanied by histone release from DNA, unlike in sperm chromatin decondensation, but it was accompanied by a variety of epigenetic modifications. At the functional level, the chromatin decondensation facilitated new gene expression as shown by the nuclear transplantation into oocytes. This study provides new insight into the molecular and functional analyses of chromatin decondensation in the context of somatic cell nuclear cloning.