γ-Tubulin is a ubiquitous and highly conserved component of centrosomes in eukaryotic cells. Genetic and biochemical studies have demonstrated that γ-tubulin functions as part of a complex to nucleate microtubule polymerization from centrosomes. We show that, as in other organisms, Caenorhabditis elegans γ-tubulin is concentrated in centrosomes. To study centrosome dynamics in embryos, we generated transgenic worms that express GFP::γ-tubulin or GFP::β-tubulin in the maternal germ line and early embryos. Multiphoton microscopy of embryos produced by these worms revealed the time course of daughter centrosome appearance and growth and the differential behavior of centrosomes destined for germ line and somatic blastomeres. To study the role of γ-tubulin in nucleation and organization of spindle microtubules, we used RNA interference (RNAi) to deplete C. elegansembryos of γ-tubulin. γ-Tubulin (RNAi) embryos failed in chromosome segregation, but surprisingly, they contained extensive microtubule arrays. Moderately affected embryos contained bipolar spindles with dense and long astral microtubule arrays but with poorly organized kinetochore and interpolar microtubules. Severely affected embryos contained collapsed spindles with numerous long astral microtubules. Our results suggest that γ-tubulin is not absolutely required for microtubule nucleation in C. elegans but is required for the normal organization and function of kinetochore and interpolar microtubules.

The CCCH finger protein PIE-1 is a regulator of germ cell fate that segregates with the germ lineage in early embryos. At each asymmetric division, PIE-1 is inherited preferentially by the germline daughter and is excluded from the somatic daughter. We show that this asymmetry is regulated at the protein level by two complementary mechanisms. The first acts before cell division to enrich PIE-1 in the cytoplasm destined for the germline daughter. The second acts after cell division to eliminate any PIE-1 left in the somatic daughter. The latter mechanism depends on PIE-1's first CCCH finger (ZF1), which targets PIE-1 for degradation in somatic blastomeres. ZF1s in two other germline proteins, POS-1 and MEX-1, are also degraded in somatic blastomeres, suggesting that localized degradation also acts on these proteins to exclude them from somatic lineages.

Histone deacetylation plays an important role in methylated DNA silencing. Recent studies indicated that the methyl-CpG-binding protein, MBD2, is a component of the MeCP1 histone deacetylase complex. Interestingly, MBD2 is able to recruit the nucleosome remodeling and histone deacetylase, NuRD, to methylated DNA in vitro. To understand the relationship between the MeCP1 complex and the NuRD complex, we purified the MeCP1 complex to homogeneity and found that it contains 10 major polypeptides including MBD2 and all of the known NuRD components. Functional analysis of the purified MeCP1 complex revealed that it preferentially binds, remodels, and deacetylates methylated nucleosomes. Thus, our study defines the MeCP1 complex, and provides biochemical evidence linking nucleosome remodeling and histone deacetylation to methylated gene silencing.

In the early Caenorhabditis elegans embryo, maternally expressed PIE-1 protein is required in germ-line blastomeres to inhibit somatic differentiation, maintain an absence of mRNA transcription, and block phosphorylation of the RNA polymerase II large subunit (Pol II) carboxy-terminal domain (CTD). We have determined that PIE-1 can function as a transcriptional repressor in cell culture assays. By fusing PIE-1 sequences to the yeast GAL4 DNA-binding domain, we have identified a PIE-1 repression domain that appears to inhibit the transcriptional machinery directly. A sequence element that is required for this repressor activity is similar to the Pol II CTD heptapeptide repeat, suggesting that the PIE-1 repression domain might target a protein complex that can bind the CTD. An alteration of this sequence element that blocks repression also impairs the ability of a transgene to rescue a pie-1 mutation, suggesting that this repressor activity may be important for PIE-1 function in vivo.

Early embryonic germ cells in C. elegans and D. melanogaster fail to express many messenger RNAs expressed in somatic cells. In contrast, we find that ribosomal RNAs are expressed in both cell types. We show that this deficiency in mRNA production correlates with the absence of a specific phosphoepitope on the carboxy-terminal domain of RNA polymerase II. In both C. elegans and Drosophila embryos, this phosphoepitope appears in somatic nuclei coincident with the onset of embryonic transcription, but remains absent from germ cells until these cells associate with the gut primordium during gastrulation. In contrast, a second distinct RNA polymerase II phosphoepitope is present continuously in both somatic and germ cells. The germ-line-specific factor PIE-1 is required to block mRNA production in the germ lineage of early C. elegans embryos (Seydoux, G., Mello, C. C., Pettitt, J., Wood, W. B., Priess, J. R. and Fire, A. (1996) Nature 382, 713–716). We show here that PIE-1 is also required for the germ-line-specific pattern of RNA polymerase II phosphorylation. These observations link inhibition of mRNA production in embryonic germ cells to a specific modification in the phosphorylation pattern of RNA polymerase II and suggest that repression of RNA polymerase II activity may be part of an evolutionarily conserved mechanism that distinguishes germ line from soma during early embryogenesis. In addition, these studies also suggest that different phosphorylated isoforms of RNA polymerase II perform distinct functions.