Fine structure and localization of the nuclear matrix in situ

Abstract

Under normal circumstances the nuclear matrix is not readily visible as a discrete structure in situ, as it is masked by chromatin and other nucleoplasmic components. The fibrous protein matrix as such is really evident only as a residual structure after bulk extraction of DNA and most proteins from isolated nuclei. The morphology of such residues is far from homogeneous, however, and usually includes a mixture of nuclear fragments, ghosts and membranes, in addition to matrices with a wide range of structural configurations. The precise architecture of the nuclear matrix in situ, therefore, is unclear and it is uncertain to what degree the morphology of isolated matrices is artifactually distorted and aggregated. We have attempted to clarify this situation by comparing the internal structure of isolated nuclei held in different ionic media, of purified nuclear matrices and of nuclei in intact tissues treated with α-amanitin. At low doses this transcriptional inhibitor induces transient structural changes in liver nuclei, including total condensation of chromatin and general nucleoplasmic clearing due to loss of soluble nuclear protein and RNP components. Under these conditions, a reticulate framework is revealed in the nucleoplasm which forms an ubiquitous and quite homogeneous matrix. The exposed matrix in situ stains intensely for protein but weakly for RNP, and consists of a network of fibers 10–15 nm thick which pervade all areas of the nucleus including chromatin. It is concluded that the protein framework in nuclei is far more extensive than indicated by studies with isolated matrices and that there are probably no truly matrix-free areas in situ, only compartments like the nucleolar matrix and clusters of linked RNP granules. It is also apparent that the architecture of purified matrices is grossly distorted during nuclear isolation and likely bears only limited resemblance to its actual conformation in intact cells. The molecular consequences of this fact should be fully assessed before highly specific functions are ascribed to the matrix constituents.