Intermediate filaments and gene regulation.

The biological role of intermediate filaments (IFs) of eukaryotic cells is still a matter of conjecture. On the basis of immunofluorescence and electron microscopic observations, they appear to play a cytoskeletal role in that they stabilize cellular structure and organize the distribution and interactions of intracellular organelles and components. The expression of a large number of cell type-specific and developmentally regulated subunit proteins is believed to provide multicellular organisms with different IF systems capable of differential interactions with the various substructures and components of their multiple, differentiated cells. However, the destruction of distinct IF systems by manipulation of cultured cells or by knock-out mutation of IF subunit proteins in transgenic mice exerts relatively little influence on cellular morphology and physiology and on development of mutant animals. In order to rationalize this dilemma, the cytoskeletal concept of IF function has been extended to purport that cytoplasmic (c) IFs and their subunit proteins also play fundamental roles in gene regulation. It is based on the in vitro capacity of cIF(protein)s to interact with guanine-rich, single-stranded DNA, supercoiled DNA and histones, as well as on their close structural relatedness to gene-regulatory DNA-binding and nuclear matrix proteins. Since cIF proteins do not possess classical nuclear localization signals, it is proposed that cIFs directly penetrate the double nuclear membrane, exploiting the amphiphilic, membrane-active character of their subunit proteins. Since they can establish metastable multisite contacts with nuclear matrix structures and/or chromatin areas containing highly repetitive DNA sequence elements at the nuclear periphery, they are supposed to participate in chromosome distribution and chromatin organization in interphase nuclei of differentiated cells. Owing to their different DNA-binding specificities, the various cIF systems may in this way specify different chromatin organizations and thus the expression of distinct sets of cell- or tissue-specific proteins. In support of this, different type III IFs have been shown to preferentially interact with guanine-rich, highly repetitive, double-stranded fragments of total genomic DNA, including chromosomal telomere sequences. Surprisingly, they also bound AT-rich, centromeric satellite DNA sequences with high efficiency. Since most of the affinity-isolated, non-telomeric and -centromeric DNA fragments contain regulatory elements that are normally located in 5'/3'-flanking and intron regions of genes, cIFs may activate gene expression or repress it as the result of telomeric and centromeric position effects. However, the nucleotide sequences of the cIF-bound, genomic DNA fragments also predict the involvement of cIF(protein)s in recombination and hence in evolutionary processes. Based on these observations, the initially observed minor effects of cIF protein knock-out mutations on the phenotype of transgenic mice may be interpreted as a redundancy phenomenon operating at the levels of the cytoskeleton and gene expression, whereas the capacity of the mutated animals to adapt to new environments via recombination processes may be severely disturbed and, as such, perceivable only after many generations of less favorable living conditions.