Allostery through DNA drives phenotype switching

Allostery is the structural coupling between distant ligand-binding sites in biomolecules, producing cooperative binding and regulating activity in proteins and genetic circuits. Evidence has accumulated that DNA can also transmit allosteric signals, but whether natural promoters exploit long-range DNA-mediated allostery to regulate gene expression was unclear. The study focuses on phenotype switching to competence in Bacillus subtilis, controlled by the transcription factor ComK, whose stochastic threshold crossing leads to positive feedback and an all-or-none activation of target promoters. The mechanism by which ComK binds cooperatively to its natural promoters was unknown. The authors aim to test whether DNA transmits allosteric signals between distant ComK-binding elements (boxes) and to elucidate the physical mechanism and design principles underlying this cooperativity.

Prior work demonstrated DNA can participate in allosteric interactions, often using artificial promoter constructs or observing short-range effects. A-tracts (AT-rich tracts) are known to narrow the minor groove and induce DNA curvature, potentially influencing protein binding and communication between sites. Classical models include DNA looping, protein bridging via spacers, and elastic-coupling theories predicting spacer-length-dependent interactions modulated by DNA helical periodicity. However, whether natural promoters evolved long-range (>nanometres) DNA-mediated allostery with functional regulatory consequences remained largely unresolved.

• Promoter constructs: Natural ComK target promoters consist of two AT-rich boxes separated by a variable spacer (8, 18, or 31 bp in addAB, comG, and comK, respectively). The authors engineered labeled comG promoters (18 bp spacer) and designed additional promoters with spacers of 8, 14, 18, 24, and 31 bp. Single-box constructs (~43 bp) were also generated.
• Single-molecule FRET (smFRET): Site-specific donor/acceptor labeling within each box and within spacers mapped distances and monitored binding. FRET histograms were fit as superpositions of free and bound Gaussian populations to extract binding fractions across ComK titrations and derive Hill coefficients. Promoter curvature was probed at pH 7 vs pH 4 to assess dependence on DNA stiffness. End-to-end FRET was used to test for looping.
• ComK oligomeric state: Two-focus fluorescence correlation spectroscopy (2fFCS) with singly labeled ComK and quantitative size-exclusion chromatography assessed ComK oligomerization in solution, confirming a monomer at experimental concentrations.
• Cryo-electron microscopy (cryo-EM): Single-particle cryo-EM on ComK–DNA complexes with 8, 18, and 31 bp spacers provided 2D classes for binding stoichiometry and 3D reconstructions (tilted data) for structural arrangement. DNA curvature angles were computed from EM maps via spline fits.
• Binding models: A mechanistic Koshland–Nemethy–Filmer (KNF) model described stepwise binding with intra-box cooperativity (σ) and inter-box DNA-mediated coupling (J), yielding a coupling free energy Δg = −kBT ln J. Hill plots and model fits quantified cooperativity and coupling.
• Elastic-coupling theory: An elastic worm-like chain model incorporating helical periodicity (10.5 bp) was fit to spacer-length series to estimate the decay length ξ, effective tension f, and average change in bending per bp.
• Sequence perturbations: Spacer GC content was increased (to 55% and 72%), a 4-base mismatch was introduced centrally, and a nick was created in the spacer to test backbone integrity effects. Electrophoretic mobility shift assays (EMSA) qualitatively assessed binding across spacer lengths to determine periodic modulation.
• Multi-box promoters: Constructs with up to four boxes (8 bp spacing) tested range of allosteric propagation (nearest-neighbor vs longer-range).

• Cooperative binding requires both boxes: smFRET titrations on the natural comG promoter (18 bp spacer) gave Hill exponents n = 3.6 ± 0.1 (box 1) and n = 3.4 ± 0.3 (box 2). Single-box constructs yielded lower n: 1.7 ± 0.1 (box 1) and 1.5 ± 0.1 (box 2), indicating inter-box allosteric coupling is essential for high cooperativity.
• No DNA looping or spacer bridging: Cryo-EM 2D classes showed ComK density localized to boxes with unlooped DNA; end-to-end FRET showed no increase upon binding. Few particles had ComK bound to only one box even at cryo-EM conditions, consistent with all-or-none binding.
• Stoichiometry and arrangement: 3D reconstructions for 8 bp and 18 bp spacers revealed two ComK molecules per box (pseudo C2 symmetry), totaling four per promoter, with the two boxes oriented on the same DNA face. Protein density within a box was continuous, suggesting intra-box protein–protein contacts; no inter-box protein contacts were detected.
• DNA curvature mechanism: ComK binding reduces promoter DNA curvature on average by ~0.6°/bp from cryo-EM analysis. Modeling indicates ComK interacts with the minor groove of A-tracts, likely widening it and reducing curvature, transmitting mechanical changes through the spacer.
• Spacer-length dependence: Hill exponents and inter-box coupling free energies decreased with increasing spacer length. Coupling free energy Δg: 5.8 ± 0.9 kBT (8 bp), 4.5 ± 1.2 kBT (18 bp), 1.9 ± 0.1 kBT (31 bp). Designed spacer series showed oscillatory modulation with helical periodicity (~10.5 bp) and stronger coupling for symmetric (integer-turn) arrangements than antisymmetric (non-integer-turn) spacers.
• Elastic-coupling parameters: Fits yielded decay length ξ = 14 ± 8 bp, effective tension f ≈ 7 pN, and an average bending change of 1.1 ± 0.3°/bp upon ComK binding, consistent with cryo-EM curvature changes.
• Backbone integrity is critical: Increasing spacer GC content reduced cooperativity, but coupling persisted. A 4-base mismatch reduced cooperativity similarly, indicating continuous base stacking is not strictly required. Introducing a nick abolished inter-box allostery, reducing the Hill exponent to n = 1.8 ± 0.1 (near isolated-box values), implicating the intact DNA backbone in transmitting curvature-induced tension.
• Nearest-neighbor propagation: Adding more boxes (up to four with 8 bp spacing) did not increase the Hill exponent beyond the two-box promoter, indicating allostery propagates primarily between nearest neighbors.
• Functional distances: Effective DNA-mediated allostery operates across >6 nm (18 bp), significantly longer than previously reported short-range effects in many systems.

The study demonstrates that natural promoters exploit DNA-mediated allostery to generate high cooperativity in transcription factor (ComK) binding. The coupling arises from mechanical deformation: ComK binding reduces local DNA curvature at one box, which is transmitted through the spacer via the backbone, increasing affinity at the distant box. This long-range (>6 nm) allostery enables an all-or-none promoter occupancy within a narrow TF concentration range, consistent with the competence phenotype switch requiring thresholded, cooperative activation. The mechanistic dependence on spacer length, helical phasing, and sequence composition (GC content) highlights how promoter architecture encodes dose-response steepness. These findings link molecular-scale DNA mechanics to systems-level behavior, suggesting that tuning promoter mechanics can modulate noise filtering and the dynamics of genetic circuits, even in regimes with deterministic expression levels.

Natural ComK promoters in Bacillus subtilis harness DNA-mediated allostery to achieve strong cooperative binding without DNA looping or protein bridging across spacers. Mechanical signal transmission via reductions in DNA curvature, propagated through an intact backbone, couples distant binding boxes. Cooperativity and coupling strength are tunable by spacer length, helical phasing, and sequence features, with coupling energies up to ~6 kBT and a decay length of ~14 bp. These principles provide a design framework to fine-tune promoter dose-response curves and thereby engineer the dynamic behavior of gene regulatory networks. Future work could extend these concepts to other TF–promoter systems, quantify in vivo impacts on gene expression noise and switching, and explore rational promoter engineering leveraging spacer mechanics and sequence to program network dynamics.

• Structural resolution: Cryo-EM reconstructions were at moderate resolution with preferred orientations; for comG, ~10 bp at one end were not resolved, and comparisons of average bending angles to elastic models are qualitative.
• Surface effects: Evidence suggests ComK–DNA complexes may adsorb to the water–air interface during grid preparation, potentially affecting peripheral protein densities.
• Dynamics: smFRET indicates limited large-scale fluctuations, but small-amplitude motions cannot be excluded without more advanced analyses.
• In vitro conditions: Experiments required additives (e.g., L-arginine) to prevent ComK aggregation; while controls suggest DNA conformation and activity are retained, these conditions differ from the cellular environment.
• Backbone perturbations: The nicking experiment shows strong effects on cooperativity, but increased flexibility complicates quantitative comparison to intact duplexes.
• Generalizability: Findings are demonstrated for ComK promoters; broader applicability to other natural promoters requires further validation in vivo.