Fig. 1: Applying turns to reconstituted chromatin fibers mimics the effect of torsional stress in vivo.

a Schematic summarizing the current understanding of the response of chromatin to transcription-generated torsional stress. Progressing RNA polymerase (RNAp) generates positive DNA supercoils, ahead of the transcription bubble, that destabilize nucleosomes (red). Negative supercoils that form behind RNAp facilitate nucleosome assembly. b Schematic of single-molecule magnetic tweezers used to apply tension and torsion to reconstituted chromatin. Fibers containing \(\sim\)30 regularly spaced nucleosomes (histone octamers, red, wrapped by DNA, grey) are flanked by \(\sim\)2 kb DNA handles for tethering to a paramagnetic bead (gold) and a glass slide (blue), respectively. The DNA handles include either multiple biotin or multiple digoxigenin moieties (triangles) in order to ensure rotational constraint. A pair of cubic magnets (red and green squares) exerts a constant stretching force on the paramagnetic bead, and hence on the tether. To apply rotations to the tethers, the magnets are rotated in either a positive (counter-clockwise) or negative (clockwise) direction. c Rotation-extension curves of DNA (grey) and a chromatin tether (red) under a force of 0.5 pN. At this force, the rotation curve of DNA handles alone (grey) is symmetric with respect to zero turns. The rotation curve of a chromatin tether (red) is asymmetric with respect to zero turns and exhibits a broad apex at positive turns. Note that this dataset obtained by rotating from negative to positive turns fully overlaps with a subsequent dataset obtained by rotating in the reverse direction (dark grey). Upon the addition of heparin to the same tether, the rotation–extension curve (blue) exhibits a maximal extension that exceeds that of the folded fiber by a factor of three. Furthermore, the apex of the curve is shifted by +34 turns, from which the negative linking number of the folded chromatin fiber is deduced (Lk = −1.4, see main text). Inset: the linking number of a chromatin fiber, quantified by the shift between rotation–extension curves of chromatin fibers versus bare DNA, is proportional to the number of assembled nucleosomes. The red line is a linear fit to the data with a slope of \(-\)1.35 (SE = 0.02) that corresponds with the population-averaged linking number per nucleosome (n = 14). Underlying rotation–extension curves are shown in Supplementary Fig. 1. Source data are provided as a Source Data file.