Testing the limits of gradient sensing

The ability to detect a chemical gradient is fundamental to many cellular processes. In multicellular organisms gradient sensing plays an important role in many physiological processes such as wound healing and development. Unicellular organisms use gradient sensing to move (chemotaxis) or grow (chemotropism) towards a favorable environment. Some cells are capable of detecting extremely shallow gradients, even in the presence of significant molecular-level noise. For example, yeast have been reported to detect pheromone gradients as shallow as 0.1 nM/μm. Noise reduction mechanisms, such as time-averaging and the internalization of pheromone molecules, have been proposed to explain how yeast cells filter fluctuations and detect shallow gradients. Here, we use a Particle-Based Reaction-Diffusion model of ligand-receptor dynamics to test the effectiveness of these mechanisms and to determine the limits of gradient sensing. In particular, we develop novel simulation methods for establishing chemical gradients that not only allow us to study gradient sensing under steady-state conditions, but also take into account transient effects as the gradient forms. Based on reported measurements of reaction rates, our results indicate neither time-averaging nor receptor endocytosis significantly improves the cell’s accuracy in detecting gradients over time scales associated with the initiation of polarized growth. Additionally, our results demonstrate the physical barrier of the cell membrane sharpens chemical gradients across the cell. While our studies are motivated by the mating response of yeast, we believe our results and simulation methods will find applications in many different contexts.Author summary: In order to survive, many organisms must not only be able to detect the presence of a chemical compound, but also in which direction that compound increases or decreases in concentration. For example, bacteria cells prefer to move towards areas with high sugar concentrations. The process by which cells determine the direction of a chemical gradient is called “Gradient Sensing”. Of particular interest is the gradient sensing capability of yeast cells. These cells have been observed detecting the direction of extremely shallow gradients, which produce only a 2% difference in the number of molecules across the cell. Because the molecular-level noise is much larger than this signal, it is unclear what noise-reduction mechanism the cell employs to reduce the noise and detect the signal. We developed a 3D computational simulation platform to calculate and study the exact positions of molecules during this process. Our platform utilizes High Performance Computing clusters and GPGPUs. We find that, of the two prevailing models in the literature, neither time-averaging nor receptor endocytosis sufficiently reduces molecular noise for yeast cells to reliably detect chemical gradients before they initiate polarized growth. This finding implies yeast must possess a mechanism for reorienting the direction of growth after cell polarization has occurred. We also find the cell membrane and similarly, any other physical barrier nearby the cell can improve the cell’s likelihood of detecting the gradient. Our simulation methods and results will be applicable in other areas of research.