Unravelling the structural plasticity of stretched DNA under torsional constraint

Regions of the genome are often held under torsional constraint. Nevertheless, the influence of such constraint on DNA–protein interactions during genome metabolism is still poorly understood. Here using a combined optical tweezers and fluorescence microscope, we quantify and explain how torsional constraint influences the structural stability of DNA under applied tension. We provide direct evidence that concomitant basepair melting and helical unwinding can occur in torsionally constrained DNA at forces >∼50 pN. This striking result indicates that local changes in linking number can be absorbed by the rest of the DNA duplex. We also present compelling new evidence that an overwound DNA structure (likely P-DNA) is created (alongside underwound structures) at forces >∼110 pN. These findings substantiate previous theoretical predictions and highlight a remarkable structural plasticity of torsionally constrained DNA. Such plasticity may be required in vivo to absorb local changes in linking number in DNA held under torsional constraint.

data were recorded in a buffer of 20 mM tris pH 7.6. Errors are standard errors of the mean (SEM) (N~10).

Forces associated with T1-T4 as a function of divalent salt (MgCl 2 ) concentration.
Building upon our investigation into the effects of monovalent salt (NaCl) on the FD curves of torsionally-constrained DNA (Fig. 2), Supplementary Fig. 3 highlights how the forces associated with T1-T4 vary as a function of divalent salt (MgCl 2 ) concentration. As the MgCl 2 concentration is increased, the trends in FD behaviour of torsionally-constrained DNA are generally similar to those observed during the NaCl concentration sweep (Fig. 2).
Interestingly, however, the effects of MgCl 2 are typically ~100-fold higher than for NaCl. This is in accordance with previous studies that indicate that ~100-fold higher sodium concentration is required to induce the same effective DNA diameter as magnesium does (1). Using the parameters determined in Supplementary Fig. S5 (panels a-c), we calculated FD curves using the eWLC model (with 0.1 pN step-size). From these calculated curves, we then extracted the 2 nd derivative using a Savitzky-Golay (20 point) smoothing filter ( Supplementary Fig. 5d). Next, we determined the extension and force in the calculated FD curves at which T1 occurs as a function of NaCl concentration ( Supplementary Fig. 5, panels e and f, respectively). A similar procedure was performed for the MgCl 2 concentration sweep, displayed in Supplementary Fig. 6. As appreciated from Supplementary Figs. 5 and 6 (panels e and f), changes in the extension and force of torsionally-constrained DNA at T1 can largely be explained by the perturbations to Lp and S of the molecule as a function of ionic strength.

Ionic-strength dependency of T1. Panels a-c of Supplementary
Peak intensities of T1-T4 as a function of ionic strength. The intensity of each peak identified in the 2 nd derivative profiles of torsionally-constrained DNA ( Fig. 2a and Supplementary Fig. 3a) is essentially determined by the difference in slope of FD curves before and after each transition (T1-T4). By plotting the peak intensity of T1-T4 as a function of ionic strength ( Supplementary Fig. 7), we can determine how salt influences the change in pitch of FD curves during each transition.
Most notable here is the change in magnitude of the 2 nd derivative peak at T4.
Supplementary Fig. 7 reveals that the magnitude of this T4 peak is progressively larger as the ionic strength is increased. This change reflects two parameters: (i) the cooperativity of overstretching between T3 and T4 and (ii) the stiffness of the DNA molecule at forces > T4. As shown in Supplementary Fig. 8, the latter is largely constant as a function of either monovalent or divalent salt concentration. This indicates, therefore, that the increase in magnitude of the 2 nd derivative peak at T4 as a function of ionic strength is primarily associated with an increasingly flat (and thus cooperative) overstretching plateau between T3 and T4. This in turn has the effect of further decreasing the force at which T4 occurs at high salt concentrations (as observed in Fig. 2e and Supplementary Fig. 3e).
Phosphorylated end-cap #2 (10 l, 10 M) was then added to the above solution (in ~100fold excess to -DNA). The mixture was heated to 80°C for five minutes, and cooled rapidly on ice. T4 DNA ligase enzyme (4 l, 200 units Fermentas) was subsequently added to the solution and left to react at 16ºC for 12 hours. Once completed, the ligated DNA construct was purified and extracted by ethanol precipitation and stored in TE buffer.
Preparation of end-closed torsionally-constrained DNA. Once each end-cap was ligated to -DNA, the DNA molecule could be tethered between two streptavidin-coated beads via biotin-streptavidin bonds. The number of biotins which bind to a given bead is variable: in the majority of cases (typically ~80%), more than two biotins on each end of the DNA molecule bind to a bead. This renders the DNA molecule unable to change its total linking number and the molecule is thus torsionally-constrained, as highlighted in Supplementary   Fig. 2a.
Preparation of end-closed torsionally-relaxed DNA. Torsionally-relaxed DNA could be prepared using the DNA construct depicted in Supplementary Fig. 1, since in ~20% of cases only a single biotin moiety is successfully tethered to a bead on at least one end of the DNA molecule. As a result, the DNA molecule is free to rotate around its axis (via the single biotin-streptavidin bond) as tension is applied (5). This assay is illustrated schematically in Supplementary Fig. 2b. Since the DNA molecule is still end-capped (i.e. contains no nicks or free-ends), the formation of peeled single-stranded DNA during overstretching is topologically forbidden (5). This allowed us to compare torsionallyconstrained DNA with torsionally-relaxed DNA, while maintaining topological constraint ( Figs. 1 and 4).