Evidence from light-scattering studies for a dimeric structure for deoxyribonucleic acid in solution.

The investigation of the size and shape of the sodium salt of deoxyribonucleic acid (DNA) in solution has attracted much attention both because of its inherent interest and its great biological importance. But the complications of its polyelectrolyte character and its very great size, both of which contribute to the marked intermolecular interactions, make the interpretation of the experimental results difficult. The least ambiguous method, light scattering, indicates that the nucleic acid derived from calf thymus is best regarded as a stiffened random coil with a molecular weight of about 6 x 106, and using the formulae appropriate to coiling polymer chains, Doty, Reichmann, Rice & Thomas (1954) have shown that the results from measurements of viscosity and sedimentation are consistent with this value. Crick & Watson (1953, 1954) have proposed recently that the analytical data for DNA and the X-ray diffraction diagrams by Franklin & Gosling (1953) obtained forDNA fibres can best be explained by a model in which two strands of nucleic acid intertwine to form a double helix held together by hydrogen bonds between the amino and hydroxyl groups of the purines and pyrimidines. X-ray diagrams indicating the presence of such structures have also been obtained from DNA in native materials (Wilkins, Stokes & Wilson, 1953). Since it is unlikely that such a specific structure can be formed in the process of producing the fibres, it follows on this hypothesis that DNA in solution must already be associated as a double molecule. To test this hypothesis we examined by light scattering the size and shape of DNA in concentrations of urea which we had previously shown to dissociate hydrogen-bonded aggregates of dyes (Alexander & Stacey, 1952) and ofpolyvinyl alcohol (Stacey & Alexander, 1954a). The dissociation by 4M urea of the haemoglobin molecule into two halves by the breaking of hydrogen bonds was clearly demonstrated in the ultracentrifuge by Steinhardt (1938). The results ofpreliminary experiments (Stacey & Alexander, 1954b) indicated that there was a dissociation of the DNA molecule, although it appeared to retain its linear dimensions. Greenstein & Jenrette (1941) had observed that urea decreases the viscosity ofsolution ofDNA, and this observation was confirmed by Conway & Butler (1952), who interpreted this change as a loss of the rigidity of the DNA molecule and also found indications for a halving of the molecular weight. The ionization of the amino groups of cytosine and adenine which are involved in the hydrogen bonds should also produce a dissociation of the twin spiral structure, and for this reason the effect of adding acid on the size and shape of DNA has also been examined. Cecil & Ogston (1948) found in the ultracentrifuge that in solutions of nucleic acid acidified to pH 3*5 a second heterogeneous component appeared with a lower sedimentation constant and increased rate ofboundary spreading, but Doty, Bunce & Reichmann (1953) have shown that it is possible to titrate DNA to pH 2-6 and back to pH 5-8 without degradation. From the change in the angular distribution of the light scattering they concluded that the well-known decrease in viscosity on acidification was due to the contraction of the coil to a much more compact structure. Pouyet & Sadron (1954) have challenged this interpretation because they find no change in the light scattering on adding acid to pH 3-8 and suggest DNA is rather more rigid. We have studied the light scattering of nucleic acid from herring sperm, which appears to have the same properties as the more extensively investigated DNA from calf thymus.