DNA looping mediates nucleosome transfer

Abstract

Proper cell function requires preservation of the spatial organization of chromatin modifications. Maintenance of this epigenetic landscape necessitates the transfer of parental nucleosomes to newly replicated DNA, a process that is stringently regulated and intrinsically linked to replication fork dynamics. This creates a formidable setting from which to isolate the central mechanism of transfer. Here we utilized a minimal experimental system to track the fate of a single nucleosome following its displacement, and examined whether DNA mechanics itself, in the absence of any chaperones or assembly factors, may serve as a platform for the transfer process. We found that the nucleosome is passively transferred to available dsDNA as predicted by a simple physical model of DNA loop formation. These results demonstrate a fundamental role for DNA mechanics in mediating nucleosome transfer and preserving epigenetic integrity during replication.

Figure 1. Mechanical displacement of a single nucleosome.

(a) Experimental configuration. A single dsDNA molecule was mechanically unzipped using an optical trap (Supplementary Fig. 1A). The dsDNA contained a positioned nucleosome followed by a long naked DNA segment (Supplementary Fig. 2A, Supplementary Table 1 and Methods). (b) A representative unzipping trace. A force rise from the naked DNA baseline indicated the detection of a bound protein complex. A dashed vertical line indicates the dyad location of a nucleosome. N=121 traces. (c) Histogram of nucleosome transfer distance. A transfer distance was obtained from the first transfer event of each trace. The histogram was obtained by pooling data from 121 traces. The prediction (not a fit) from the DNA looping model is plotted for comparison. The resulting Pearson test gives a reduced χ² of 0.53 with a P value of 0.81 (Methods; Supplementary Fig. 4A).

Figure 2. Mechanical displacement of a single nucleosome in the presence of competitor DNA.

(a) Experimental configuration. A single dsDNA molecule was mechanically unzipped using an optical trap (Supplementary Fig. 1A). The dsDNA contained a positioned nucleosome followed by a long naked DNA segment (Supplementary Fig. 2A, Supplementary Table 1 and Methods). Linear competitor dsDNA of 2,987 bp was introduced into the chamber at varying concentrations immediately before mechanical disruption. N=121, 52, 57, 39, 38 and 92 traces for 0, 30, 50, 100, 200 and 300 ng μl⁻¹ competitor DNA concentrations, respectively. (b) Two example unzipping traces in the presence of 100 ng μl⁻¹ of competitor DNA. The top trace shows an absence of a transferred nucleosome to the downstream DNA, whereas the bottom trace shows the presence of a transferred nucleosome. (c) The probability of nucleosome transfer to downstream dsDNA as a function of competitor DNA concentration. Error bars represent 95% confidence intervals (Methods). A direct prediction (not a fit) based on DNA looping and a simple competitive binding relation (Methods) is shown for comparison.

Figure 3. Helicase displacement of a single nucleosome.

(a) Experimental configuration. A single dsDNA molecule was unwound by a T7 helicase as the two strands of the DNA were held under 12 pN of force by an optical trap, which assisted helicase unwinding but was insufficient to mechanically separate the dsDNA (Supplementary Fig. 1B). The nucleosomal DNA template is specified in Supplementary Fig. 2A, Supplementary Table 1 and Methods. (b) Representative helicase-unwinding traces on a nucleosomal (black) or naked (grey) template. Helicase unwinding was interrupted by discrete pauses along the DNA template. Dashed lines indicate the dyad locations of the initial positioned nucleosome and the transferred nucleosome. N=49 traces. (c) Histogram of nucleosome transfer distance. A transfer distance was obtained from the first transfer event of each trace as indicated by the arrow in Fig. 2b. The histogram was obtained by pooling data from 49 traces. The prediction (not a fit) from the DNA looping model is plotted for comparison. The resulting Pearson test gives a reduced χ² of 0.31 with a P value of 0.95 (Methods; Supplementary Fig. 4A).

Figure 4. Replisome displacement of a single nucleosome.

(a) Experimental configuration. Leading strand replication was carried out using the T7 replisome on a Cy5-labelled parental template containing a single nucleosome with minimal dsDNA downstream. (b) Nucleosome transfer after replication. The replication product was exonuclease III-digested and assayed on a denaturing gel (lane 7). Lane 1 is a ladder and lanes 2–6 are control experiments. Lanes digested with exonuclease III were loaded with five times as much sample as the other lanes to achieve more accurate quantification. N=4 replicates (Supplementary Fig. 7). (c) A line scan of lane 7 contained contributions from both the transferred nucleosome as well as background. In particular, a fraction of replisomes did not proceed past the nucleosome, and another fraction contained inactive replisome bound at the initial fork (see lane 6). The background in lane 6 is removed from lane 7 during subsequent analysis. The template schematic right of the line scan explains some features of the band positions in the gel and the line scans.

Figure 5. The passive nucleosome transfer model via DNA loop formation.

(a) Comparison of nucleosome transfer distance distributions as measured using three experimental approaches: mechanical fork progression (red; Fig. 1c); helicase unwinding (green; Fig. 2c); and leading strand replication (purple; Supplementary Fig. 7). Note that the peak near zero from the leading strand replication curve (purple) was background introduced by the fraction of reaction that did not proceed past the nucleosome as indicated in Fig. 3. The prediction (black, not a fit) from the DNA looping model is also shown for comparison. (b) A mechanistic model of passive nucleosome transfer mediated by DNA loop formation. When a replisome (purple) encounters a parental nucleosome (green) at the replication fork, a DNA loop forms in one of the daughter duplexes (red), bridging the nucleosome from its initial location to its new location and thus facilitating direct transfer to the daughter duplex. Nascent histones (yellow) are also deposited on the daughter strands by chaperones.