Accuracy of direct gradient sensing by single cells

Many types of cells are able to accurately sense shallow gradients of chemicals across their diameters, allowing the cells to move towards or away from chemical sources. This chemotactic ability relies on the remarkable capacity of cells to infer gradients from particles randomly arriving at cell-surface receptors by diffusion. Whereas the physical limits of concentration sensing by cells have been explored, there is no theory for the physical limits of gradient sensing. Here, we derive such a theory, using as models a perfectly absorbing sphere and a perfectly monitoring sphere, which, respectively, infer gradients from the absorbed surface particle density or the positions of freely diffusing particles inside a spherical volume. We find that the perfectly absorbing sphere is superior to the perfectly monitoring sphere, both for concentration and gradient sensing, since previously observed particles are never remeasured. The superiority of the absorbing sphere helps explain the presence at the surfaces of cells of signal degrading enzymes, such as PDE for cAMP in Dictyostelium discoideum (Dicty) and BAR1 for mating factor alpha in Saccharomyces cerevisiae (budding yeast). Quantitatively, our theory compares favorably to recent measurements of Dicty moving up a cAMP gradient, suggesting these cells operate near the physical limits of gradient detection.

💡 Research Summary

This paper develops a quantitative theory for the physical limits of direct chemical gradient sensing by single cells, a problem that had previously received little theoretical attention despite extensive experimental evidence that many eukaryotic cells can detect extremely shallow concentration gradients across their diameters. The authors introduce two idealized spherical models that capture the essential physics of how a cell might acquire information from diffusing ligand molecules: (1) a perfectly absorbing sphere that removes each ligand that contacts its surface, thereby preventing any ligand from being measured more than once, and (2) a perfectly monitoring sphere that records the positions of all ligands inside its volume but allows ligands to diffuse freely in and out, so that the same ligand can be sampled repeatedly.

For concentration sensing, the absorbing sphere receives a steady particle flux J = 4πDa c (D is the diffusion constant, a the cell radius, c the ambient concentration). Over a measurement time T it absorbs N = 4πDa c T particles, which follow a Poisson distribution. The resulting relative variance is (δc/c)² = 1/(4πDa c T). In contrast, the monitoring sphere contains on average N ≈ (4/3)πa³c particles at any instant and can make roughly N_meas ≈ T D/a² independent measurements, giving (δc/c)² ≈ 3/(5πDa c T). The absorbing sphere therefore achieves about a 12 % reduction in uncertainty compared with the monitoring sphere.

Extending the analysis to gradient sensing, the authors consider a linear concentration field c(r) = c₀ + ∇c·r. For the absorbing sphere, the positions of absorbed particles are characterized by their polar angles θ_i. The optimal estimator for the z‑component of the gradient is c_z = (1/(4πDa²T)) ∑ cosθ_i. Using Poisson statistics and the independence of particle arrivals, they derive a variance (δc_z)² = c₀/(12πDa³T). Because the three Cartesian components of the gradient are statistically independent, the total gradient uncertainty becomes (δ|∇c|)² = 1/(4πDa c₀ T), which is numerically identical to the concentration‑sensing variance for the absorbing sphere.

For the monitoring sphere, the cell records the instantaneous positions of all interior particles. The optimal estimator involves the time‑averaged first moment m_z,T = (1/T)∫ dt ∑_inside z_i(t). The variance of this estimator depends on the autocorrelation function u(τ) = ⟨z(t)z(t+τ)⟩ of a single diffusing particle. By evaluating the double time integral and the spatial integral over the sphere, the authors obtain a correlation time τ_z = 2a²/(105D). Substituting this into the variance expression yields (δ|∇c|)² = 15/(7πDa c₀ T), which is roughly 2.5 times larger than the absorbing‑sphere result. Thus, the absorbing sphere is fundamentally superior for both concentration and gradient detection because it never re‑measures the same ligand.

To test the theory, the authors compare its predictions with experimental measurements of the Chemotactic Index of Dictyostelium discoideum cells moving toward a cAMP source created by a micropipette. Using realistic parameters (D ≈ 300 µm² s⁻¹, a ≈ 5 µm, T ≈ 3.2 s) the absorbing‑sphere model reproduces the observed dependence of the Chemotactic Index on distance from the source and on the local gradient magnitude, after a modest scaling to account for the maximal observed index (~0.9). The agreement indicates that Dictyostelium operates close to the diffusion‑limited physical limit of gradient sensing.

The authors argue that the superiority of the absorbing sphere provides a functional rationale for the presence of surface‑bound signal‑degrading enzymes such as phosphodiesterase (PDE) in Dictyostelium and Bar1 in budding yeast. By degrading ligands at the membrane, cells effectively implement an absorbing boundary, thereby maximizing the information extracted from each diffusing molecule.

In conclusion, the paper establishes a rigorous physical bound on the accuracy of direct gradient sensing, demonstrates that an absorbing boundary achieves this bound, and shows that real eukaryotic chemotactic cells appear to operate near this limit. The framework offers a valuable benchmark for future studies of cellular signaling fidelity, the design of synthetic chemotactic systems, and the interpretation of how biochemical network architectures may have evolved to approach fundamental physical constraints.