Are the terminal domains in intermediate filaments organized as octameric complexes? Reevaluation of a recent suggestion.