Analysis of Accordion DNA Stretching Revealed by The Gold Cluster Ruler

A promising new method for measuring intramolecular distances in solution uses small-angle X-ray scattering interference between gold nanocrystal labels (Mathew-Fenn et al, Science, 322, 446 (2008)). When applied to double stranded DNA, it revealed that the DNA length fluctuations are strikingly strong and correlated over at least 80 base pair steps. In other words, the DNA behaves as accordion bellows, with distant fragments stretching and shrinking concertedly. This hypothesis, however, disagrees with earlier experimental and computational observations. This Letter shows that the discrepancy can be rationalized by taking into account the cluster exclusion volume and assuming a moderate long-range repulsion between them. The long-range interaction can originate from an ion exclusion effect and cluster polarization in close proximity to the DNA surface.

💡 Research Summary

The paper revisits the “Gold Cluster Ruler” technique introduced by Mathew‑Fenn et al. (Science 322, 446, 2008), which uses two gold nanocrystal labels attached to the ends of a double‑stranded DNA molecule and measures the interference pattern in small‑angle X‑ray scattering (SAXS) to obtain the end‑to‑end distance distribution. The original experiments reported surprisingly large length fluctuations—on the order of 5 % of the contour length—and a correlation length extending over at least 80 base‑pair steps (≈27 nm). This observation led to the “accordion” model, in which distant DNA segments stretch and shrink concertedly, a picture that conflicts with the majority of earlier experimental data (fluorescence resonance energy transfer, electron microscopy, conventional SAXS) and with molecular dynamics simulations, which predict only short‑range, largely independent base‑pair step fluctuations and a length stiffness of roughly 1000 pN·nm⁻¹.

The authors of the present Letter argue that the apparent discrepancy can be resolved by explicitly accounting for two physical effects associated with the gold labels themselves: (1) excluded‑volume (steric) interactions of the nanocrystals, and (2) a modest long‑range repulsive interaction between the two labels when they are in close proximity to the DNA surface. The excluded‑volume effect arises because each gold cluster occupies a finite spherical volume (≈1.4 nm in diameter). This volume perturbs the surrounding solvent and ion atmosphere, imposing a minimum separation between the labels and effectively adding the label‑label distance fluctuations to the measured distribution. Consequently, the raw SAXS data reflect a convolution of DNA conformational fluctuations with the stochastic motion of the labels.

The second effect is subtler. DNA in physiological buffers is surrounded by a cloud of counter‑ions (Na⁺, Mg²⁺, etc.). When a gold nanocrystal approaches the DNA surface, the ion cloud is partially excluded from the narrow gap between the two labels, leading to an increase in the effective electrostatic repulsion. Moreover, gold nanocrystals are highly polarizable; the proximity of the DNA surface induces dipoles in the clusters, generating an induced dipole‑dipole interaction that is repulsive at the relevant separations (a few nanometers). The authors model this interaction with a screened Yukawa potential, yielding a repulsive energy of roughly 0.1–0.3 k_BT at distances of 2–4 nm.

To test whether these two contributions can reproduce the experimental distance distributions without invoking any anomalous DNA elasticity, the authors performed Monte‑Carlo and molecular‑dynamics simulations. The DNA backbone was represented by the standard worm‑like chain (WLC) model with a persistence length of ~50 nm and a stretch modulus of ~1000 pN·nm⁻¹, identical to values used in earlier studies. The gold labels were modeled as hard spheres of radius 0.7 nm, with a hard‑core repulsion to enforce excluded volume and an additional Yukawa term to capture the long‑range repulsion. Simulations that included only the WLC DNA produced distance distributions far narrower than those observed experimentally. When the label steric term was added, the distribution broadened modestly, but still fell short of the experimental width. Only after the weak long‑range repulsion was introduced did the simulated end‑to‑end distance distribution match both the mean distance and the variance reported in the original SAXS measurements. Importantly, the correlation length of the fluctuations—previously interpreted as a DNA‑intrinsic property—emerged naturally from the label‑label interaction: the repulsive potential couples the motions of the two labels over distances up to ~30 nm, creating an apparent concerted stretching that mimics an accordion.

The authors conclude that the “accordion” behavior is not an intrinsic property of DNA but an artifact of the measurement system. The gold clusters, by virtue of their finite size and weak mutual repulsion mediated by ion exclusion and induced polarization, amplify the apparent length fluctuations and generate spurious long‑range correlations. Consequently, any quantitative interpretation of distance measurements that employ large, polarizable labels must incorporate corrections for excluded volume and label‑label interactions.

Beyond resolving the specific controversy, the work carries broader implications for nanobiophysics. First, it underscores the necessity of rigorous label modeling when using nanocrystal or nanoparticle tags for structural studies, especially in solution where ionic screening and polarization can be significant. Second, the identified repulsive mechanism—ion exclusion combined with induced dipole interactions—is likely to be present in other systems that employ metallic or high‑dielectric nanoparticles near charged biomolecules (e.g., protein‑nanoparticle conjugates, RNA‑gold complexes). Third, the authors suggest practical strategies for minimizing such artifacts: reducing label size, increasing the ionic strength to enhance screening, or chemically modifying the nanoparticle surface to lower its polarizability. By adopting these measures, future experiments can obtain more faithful representations of biomolecular dynamics without the confounding influence of the probes themselves.

In summary, the paper provides a quantitative, physically grounded explanation for the previously reported large, correlated DNA length fluctuations observed with the Gold Cluster Ruler. By incorporating excluded‑volume effects and a modest long‑range repulsive interaction between the gold nanocrystals, the authors reconcile the SAXS data with the well‑established mechanical properties of DNA, thereby restoring consistency between different experimental approaches and computational models. This reconciliation not only clarifies the specific case of DNA stretching but also offers a template for critically evaluating other nanolabel‑based structural techniques.