The attachments of chromatin loops to the nucleoskeleton.

It is widely assumed by cell biologists that chromatin is looped by attachment to some nuclear skeleton. 'Structural' attachments might be mediated through specific sequences; these would be attached in most cells in an organism, underlying the basic structure of the mitotic chromosome and persisting throughout interphase. 'Functional' attachments might also exist, perhaps if active polymerases are attached to the skeleton and replication and transcription occur as DNA is reeled through them. Cells of different tissues--and even cells of the same tissue--would have different attachments of this type. Problems associated with demonstrating these two kinds of attachment are discussed. We find little good evidence for 'structural' attachments and explore the idea that 'functional' attachments are the only kind that exist: 'functional' attachments involving active transcription units might be stable enough to organize chromatin during both interphase and mitosis, but 'dynamic' enough to allow duplication of attached sequences without disrupting loops.