STRETCHING MUST TWIST DNA

Recent single-molecule experiments indicate that a double-stranded DNA increases in length if put under tension greater than 10 piconewtons: beyond this point its conformation can no longer be described using a n inextensible wormlike-chain model. The simplest extensible wormlike chain with twist rigidity is considered as a model for DNA under tensi on, and it is found that the fact that DNA is chiral demands that stre tching be coupled to twisting at linear order in elastic theory; stret ching a DNA is thus a way to perturb its twist degrees of freedom. Non linearities can stabilize an ''overstretched'' DNA state in the model, allowing quantitative contact with experiment. Finally, it is argued that if its twist is initially highly perturbed, the interior of a str etched DNA will supercoil as it relaxes.