Stacking of short DNA induces the gyroid cubic-to-inverted hexagonal phase transition in lipid-DNA complexes

Lyotropic phases of amphiphiles are a prototypical example of self-assemblies. Their structure is generally determined by amphiphile shape and their phase transitions are primarily governed by composition. In this paper, we demonstrate a new paradigm for membrane shape control where the electrostatic coupling of charged membranes to short DNA (sDNA), with tunable temperature-dependent end-to-end stacking interactions, enables switching between the inverted gyroid cubic structure (QGII) and the inverted hexagonal phase (HCII). We investigated the structural shape transitions induced in the QGII phase upon complexation with a series of sDNAs (5, 11, 24, and 48 bp) with three types of end structure ("sticky" adenine (A)-thymine (T) (dAdT) overhangs, no overhang (blunt), and "nonsticky" dTdT overhangs) using synchrotron small-angle X-ray scattering. Very short 5 bp sDNA with dAdT overhangs and blunt ends induce coexistence of the QGII and the HCII phase, with the fraction of QGII increasing with temperature. Phase coexistence for blunt 5 bp sDNA is observed from 27°C to about 65°C, where the HCII phase disappears and the temperature dependence of the lattice spacing of the QGII phase indicates that the sDNA duplexes melt into single strands. The only other sDNA for which melting is observed is 5 bp sDNA with dTdT overhangs, which forms the QGII phase throughout the studied range of temperature (27°C to 85.2°C). The longer 11 bp sDNA forms coexisting QGII and HCII phases (with the fraction of QGII again increasing with temperature) only for "nonsticky" dTdT overhangs, while dAdT overhangs and blunt ends exclusively template the HCII phase. For 24 and 48 bp sDNAs the HCII phase replaces the QGII phase at all investigated temperatures, independent of sDNA end structure. Our work demonstrates how the combined effects of sDNA length and end structure (which determine the temperature-dependent stacking length) tune the phase behavior of the complexes. These findings are consistent with the hypothesis that sDNAs and sDNA stacks as long as or longer than the cubic unit cell length disfavor the highly curved channels present in the QGII phase, thus driving the QGII-to-HCII phase transition. As the temperature is increased, the breaking of stacks due to thermal fluctuations restores increasing percentages of the QGII phase. This journal is