Fluidification of Entanglements by a DNA Bending Protein

In spite of the nanoscale and single-molecule insights into nucleoid associated proteins (NAPs), their role in modulating the mesoscale viscoelasticity of entangled DNA has been overlooked so far. By combining microrheology and molecular dynamics simulation, we find that the abundant NAP “integration host factor” (IHF) lowers the viscosity of entangled $\lambda \mathrm{DNA}$ 20-fold at physiological concentrations and stoichiometries. Our results suggest that IHF may play a previously unappreciated role in resolving DNA entanglements and in turn may be acting as a “genomic fluidizer” for bacterial genomes.

Figure 1

Molecular Dynamics simulations of kinked semiflexible polymers. (a) Snapshot of the simulation box ($M=50$ chains $N=1000$ beads long with persistence length $l_p=3\sigma$ at volume fraction $\rho =0.05$. IHF is modeled as static and stiff harmonic angles forcing 107° kinks randomly placed along the chains. See, also, SM Movies [22]. The inset shows an AFM image of a short DNA bound by IHF from Ref. [13]. (b) Normalized squared radius of gyration. (inset) Effective persistence length $l_p^{*}=3⟨R_g^2⟩/N$. (c) Entanglement length from primitive path analysis (PPA). Grey shaded area represents predicted $N_e^{*}(l_p^{*})$ with appropriate propagation of errors. (inset) Snapshots from PPA. (d) MSD of the center of mass of the chains. (e) Normalized diffusion coefficient. The fitted curve is $1+\kappa x$ with $\kappa =0.05$ (in units of number of IHF in a polymer of 1000 beads). (f) Relaxation time $\tau$ defined as $\mathrm{MSD}(\tau )\equiv ⟨R_g^2⟩$. The shaded area represents the values expected using the numerical values of $N_e$ measured in (c) with appropriate propagation of errors.

Figure 2

Entangled solutions of $\lambda \mathrm{DNA}$ are fluidized by IHF. (a) MSDs at different stoichiometries of IHF:DNAbp. The shaded area enveloping the curves is the standard error computed over ten movies in three independent experiments ($>100$ tracers total). (b) Box plot of the diffusion coefficient $D$ from fitting $\mathrm{MSD}=2Dt$ at large times. (Inset) Representative particle trajectories tracked over 2 minutes. (c) Normalized mean diffusion coefficient at increasing concentration of IHF. (d) Complex moduli $G^{\prime }$ (solid) and $G^{\prime \prime }$ (dashed) for the control and for 1 IHF every 100 bp. (e) Relaxation time $\tau =w_R^{-1}$ where $w_R$ is the crossover frequency at which $G^{\prime }(w_R)\equiv G^{\prime \prime }(w_R)$. (f) Elastic plateau obtained from the value of $G^{\prime }$ at 50 Hz. We have indicated the two biologically relevant stoichiometries in E. coli growing and stationary phases as “ECGP” and “ECSP.”

Figure 3

Effect of substrate length on IHF fluidification. (a) MSDs of passive tracers in dense solutions of $\lambda \mathrm{DNA}$ predigested with different restriction enzymes and before or after addition of 1 IHF every 80 bp. (b) Diffusion coefficients of the tracer particles extracted from the large time behavior as $\mathrm{MSD}=2Dt$. (c) Viscosity $\eta$ as a function of average number of entanglements per chain $⟨Z⟩$. (d) Normalized viscosity after ro before adding IHF:80 bp ${\eta }_{\mathrm{IHF}}/{\eta }_0$ plotted against average fragment length. In this figure, $\lambda \mathrm{DNA}=0$ cuts, $l=48.5\mathrm{kbp}$; $\lambda \mathrm{XhoI}=1$ cut, $l=24.2\mathrm{kbp}$; $\lambda \mathrm{BamHI}=5$ cuts, $l=8\mathrm{kbp}$; $\lambda \mathrm{PstI}=28$ cuts, $l=1.7\mathrm{kbp}$; $\lambda \mathrm{HaeIII}=149$ cuts, $l=323\mathrm{bp}$. $l$ is the average fragment length.