The search kinetics of a target inside the cell nucleus

The mean time required by a transcription factor (TF) or an enzyme to find a target in the nucleus is of prime importance for the initialization of transcription, gene activation or the start of DNA repair. We obtain new estimates for the mean search time when the TF or enzyme, confined to the cell nucleus, can switch from a one dimensional motion along the DNA and a free Brownian regime inside the crowded nucleus. We give analytical expressions for the mean time the particle stays bound to the DNA, $\tau_{DNA}$, and the mean time it diffuses freely, $\tau_{free}$. Contrary to previous results but in agreement with experimental data, we find a factor $\tau_{DNA} \approx 3.7 \tau_{free}$ for the Lac-I TF. The formula obtained for the time required to bind to a target site is found to be coherent with observed data. We also conclude that a higher DNA density leads to a more efficient search process.

💡 Research Summary

The paper presents a quantitative analysis of the time required for a transcription factor (TF) or a DNA‑repair enzyme to locate its specific target within the cell nucleus. Building on the classic Berg‑von Hippel framework, the authors model the search as an alternating process: a TF alternates between one‑dimensional sliding along DNA, where it is bound by a non‑specific attractive potential, and three‑dimensional free diffusion in the crowded nucleoplasm.

First, the non‑specific interaction between the TF and the DNA backbone is represented by a radial potential V(r)=−k r, with reflecting boundary at the DNA double‑helix radius R_int and an absorbing boundary at a larger radius R_ext (chosen as twice R_int). Solving the modified diffusion equation ∆u−∇V/(k_BT)∇u=−1/D in the annulus R_int<r<R_ext yields an analytical expression for the average residence time on DNA, τ_DNA. Using realistic parameters for the LacI repressor (diffusion constant D≈3 µm² s⁻¹, potential depth derived from electrostatic considerations), the model predicts τ_DNA≈5.7 ms, in excellent agreement with single‑molecule measurements (~5 ms).

Second, the authors estimate how many base pairs (bp) are scanned during a typical sliding episode. For a constant specific potential the maximal excursion of a one‑dimensional Brownian motion gives a baseline n₀≈2√(D τ_DNA)/l_bp. To incorporate the heterogeneous energy landscape of real DNA, they introduce a random potential with variance σ and nearest‑neighbor correlation ρ. Using results from random walk theory in random potentials, they derive a corrected expression (Eq. 13) that accounts for the probabilities of moving left or right at each site (p_i, q_i) and the mean dwell time per step (u_i). With σ≈2 k_BT and a modest positive correlation ρ≈0.02, the model yields n≈75 bp per binding event, close to the experimentally observed 85 bp for LacI.

Third, the free‑diffusion phase τ_free is analyzed by modeling the nucleus (or bacterial cell) as a sphere (or cylinder) containing DNA strands arranged on a square lattice. Each strand is approximated as a cylinder of radius ε≈15 nm, and the DNA volume fraction ρ_DNA is introduced. Solving the Laplace equation with absorbing conditions on the DNA cylinders and reflecting conditions on the outer boundaries leads to an explicit formula (Eq. 18) for τ_free that depends logarithmically on ρ_DNA. For E. coli parameters, τ_free≈1.5 ms. Importantly, τ_free decreases as DNA density increases, indicating that a more crowded DNA environment accelerates the search by reducing the mean free‑flight distance.

The total search time is then assembled using the classic expression T_S≈(τ_DNA+τ_free)·N_bp/n, where N_bp≈4.8×10⁶ is the total number of base pairs in the bacterial genome. Substituting the derived values gives T_S≈7 min 48 s, matching the measured average search time of 7 min 40 s for LacI. The analysis demonstrates that τ_DNA and τ_free are independent quantities, contrary to earlier assumptions that set them equal. Moreover, scaling the DNA density by a factor k reduces the overall search time by less than k, highlighting a non‑linear benefit of DNA compaction.

In summary, the study provides a rigorous, physics‑based framework that links microscopic parameters (diffusion constant, electrostatic potential depth, energy landscape roughness, DNA packing density) to macroscopic observables (binding residence times, number of bases scanned per slide, total search time). The agreement with single‑molecule experiments validates the model and suggests that cells can modulate transcriptional response and DNA‑repair kinetics by altering DNA organization and the non‑specific binding affinity of proteins. This work thus bridges stochastic biophysical theory and experimental genomics, offering predictive tools for understanding and engineering gene regulation dynamics.