Both microscopic and macroscopic models of DNA flexibility should lead to the same quantitative description of the elastic properties of the DNA helix. This belief is reinforced by the fact that essentially all experimental (solution) studies to date support the macroscopic, elastic model. The performance of microscopic models can therefore be checked by their ability to produce the correct macroscopic quantities (P and C). To most carefully address the influence of such factors as base sequence, DNA damage, and drug or protein interaction on the flexibility of DNA, methods are required that are most sensitive for DNA molecules of less than 500-1000 bp. The use of molecules in this size range will maximize the signal due to the structural alteration as well as facilitate the construction of DNA sequences of any desired arrangement. I have emphasized three such methods and summarized their strengths and weaknesses; however, their concurrent application to the determination of DNA flexibility provides an important check of self-consistency. These studies have indicated that the persistence length of DNA in buffers of moderate salt concentration is 450-500 A. Synthetic DNA is now readily available, and many procedures for the construction and cloning of DNA molecules of defined length and sequence (107-108a) are in common use. The availability of restriction fragments of precisely defined length has transformed the study of the physical (particularly hydrodynamic) properties of such molecules, since the hitherto pervasive problem of length polydispersity has been eliminated. Sheared, sonicated, or otherwise abused calf thymus (or other) DNAs should no longer be considered acceptable materials for physical studies. Many studies of bending and torsional fluctuations in DNA have been excluded from this discussion because the DNA samples used were not precisely defined. The torsional elastic constant of DNA has been fairly well established as approximately 3.0 x 10(-19) erg-cm, mainly through a combination of elegant theoretical and experimental studies of topoisomer distributions in circular DNA molecules. The other general approach to the determination of the torsional elastic constant, luminescence decay, is still burdened by the poor characterization of the DNA used in many of the experimental studies as well as by some continued theoretical uncertainties.(ABSTRACT TRUNCATED AT 400 WORDS)
