Figure 1.
SAXS profile of gH3 nucleosome core particle at 10mM KCl compared to CRYSOL scattering predictions.
In (A) we show the comparison between SAXS data for the gH3 nucleosome core particle (NCP) (dots) and CRYSOL predictions with folded tails and no DNA unwrapping (dashed, χ=7.6) or with ~20 bp of DNA unwrapped from one end of the nucleosome (solid, χ=4.7). The minima in the scattering prediction at approximately q=0.14 Å-1 for the NCP without unwrapping is deeper than the prediction for the NCP with unwrapping or the measured scattering profile. This shows that the data can only be described by the structure with the unwrapped DNA ends. In (B) we show the comparison between the CRYSOL predictions with DNA unwrapped and tails folded (solid, χ=4.7) or extended (dashed, χ=4.6). The differences between these signals are undetectable; both match the data (dots) comparably. Similarly, the differences between tail-extended and tail-folded models with the DNA fully wrapped are small (Figure S3 in Materials S1).
Figure 2.
SAXS profiles of all wild-type, gH3, and gH4 nucleosomes at two salt concentrations.
SAXS signals of the wild-type (black), gH3 (red), and gH4 (blue) constructs in 40 mM (A) and 200 mM (B) KCl are shown. The lower minima at q=0.14 Å-1 indicates less DNA unwrapping (see Figure 1). Note that in the 40 mM KCl signal, the signals are arranged in order of increasing DNA unwrapping from gH4 to WT to gH3; however the differences between gH4 and WT are minimal. For the 200 mM KCl SAXS data, the WT and gH4 constructs are indistinguishable at q=0.14 Å-1 within the noise of the SAXS signal, while the gH3 shows significant DNA unwrapping in comparison. These results were independent of signal matching region (Figure S5 in Materials S1).
Figure 3.
Goodness of fit between gH4 data at multiple salt concentrations and CRYSOL predictions of DNA-unwrapped nucleosomes.
Chi-squared value (goodness of fit) between CRYSOL predicted scattering and SAXS data for the gH4 construct at [KCl]=10mM (circles), 40mM (squares), 100mM (triangles), and 200mM (diamons) with various amounts of DNA basepairs unwrapped. For all salt concentrations, there appears to be two local minima in χ2, one at 10bp unwrapped and one at 20bp unwrapped. The global minimum (best fit) changes from the 20bp unwrapped construct at [KCl]=10mM to the 10bp unwrapped construct at [KCl]=40mM and 100mM. At [KCl]=200mM, there is again a strong shift towards the 10bp unwrapped construct; the 10bp and 20bp χ2 minimum are approximately the same.
Figure 4.
Plot of average energy transfer efficiency between the two end-labeled DNA in the nucleosome versus salt concentration.
Plot of average energy transfer efficiency (or FRET signal) between the two fluorescently labeled DNA ends in the nucleosome versus salt concentration for the wild-type (circles), gH3 (squares), and gH4 (triangles) constructs; the standard error is plotted and is seen to be approximately the size of the symbols. A larger transfer efficiency indicates a shorter distance between the DNA ends. The data show a minimum distance (maximum FRET) for all constructs between 50 and 100mM KCl indicating a minimum in the DNA unwrapping in this vicinity. The data also show an increase in end-to-end DNA distance in the gH3 nucleosomes (decreased FRET) compared to the wild-type indicating great DNA unwrapping with the removal of the H3 tails. The gH4 nucleosomes, meanwhile, shows decreased end-to-end DNA distance (increased FRET) compared to the wild-type nucleosomes, indicating a stabilization of the nucleosome through removal of the H4 tails. These data agree well with the SAXS analyses (within error of the larger SAXS uncertainties) in Figure 2 and Table 1.