Analysis and reconstruction of Xenopus ribosomal chromatin nucleosomes.

Abstract

The double-stranded DNA from staphylococcal(micrococcal)-nuclease-derived chromatin sub-units (nucleosomes) from transcriptionally active somatic cells of heterozygous mutant (0/+ nu) embryos of the amphibian Xenopus laevis has been analyzed. To determine whether the transcriptionally active ribosomal genes of these 0/+ nu embryos are selectively hydrolyzed by micrococcal nuclease cleavage and, more generally to determine whether any particular frequency class of DNA sequences was differentially removed by nuclease treatment of chromatin, DNA reassociation (cot) curves of heterozygote (0/+ nu) mutant subunit (nucleosome) DNA were compared with the cot curves obtained from whole 0/+ nu DNA. The results of these experiments clearly demonstrated that cleavage of the chromatin by micrococcal nuclease, to the point where about 75% of the DNA is present as fragments about 200 base pairs long, does not result in any preferential removal (hydrolysis) of either the ribosomal cistrons or any particular frequency class of gene sequences. However, these same reassociation curves also show that 0/+ nu DNA has a noticable decrease in sequence complexity in all of its frequency classes relative to that of whole 0/+ nu DNA. Further experiments demonstrated that this decreased complexity is probably due to two factors: (a) a change in the nucleotide base composition of the nucleosome DNA relative to that of whole DNA (base composition analyses indicate that the subunit DNA has about 10% fewer dA + dT residues than whole DNA); and (b) the cleavage by micrococcal nuclease of the base sequences between adjacent chromatin subunits within each of the frequency classes of DNA. Further evidence supporting the similarity of nucleosome and whole cellular DNA comes from DNA-driven cot reassociation experiments involving the renaturation of individually isolated frequency classes of DNA from nucleosomes (i.e., unique sequences and intermediate repetitive sequences) with an excess of whole (sheared) cellular DNA, which showed that the frequency classes of DNA within nucleosomes are indistinguishable, on kinetic grounds, from those of non-nuclease-treated DNA. Addition experiments involving solution hybridization reactions followed by hydroxyapatite analysis of nucleic acid duplex formation have indicated that nucleosome DNA derived from transcriptionally active cells contains fragments of DNA that anneal to both ribosomal RNAs and poly(A)-containing messenger RNAs. And, finally, the results from chromatin reconstitution experiments using isolated ribosomal DNA (rDNA) are presented that indicate that the association of nucleosomes with DNA in the formation of chromatin subunits is probably not base-sequence specific. All of these experiments are consistent with (but do not prove) the idea that transcriptionally active somatic cell ribosomal and unique-sequence genes can be partially, or perhaps transiently, associated with histones in the form of nucleosomes or chromatin subunits. These results are then discussed in terms of a dynamic model for the association of nucleosomes with transcriptionally active chromatin.