Sensing their plasma membrane curvature allows migrating cells to circumvent obstacles

To navigate through tissues, migrating cells must balance persistent migration with changes in direction to circumvent obstacles. The protrusion at the leading edge of a moving cell is driven by actin polymerization and displays dynamic membrane patterns that constantly probe the environment. Here, we asked whether cells read out their membrane topography to decode their environment and decide if they should move ahead or turn away. To test this, we first parametrized the curvature of the plasma membrane of migrating immune cells to match microscopy data. Then, we created a theoretical model to explore what types of feedback between topography and protrusion could lead to the observed patterns. Our model predicts that negative coupling of positive (inward) curvature of the plasma membrane to actin polymerization would explain our data. To identify the putative positive curvature sensor, we screened for proteins with suitable membrane binding domains that are expressed during cell migration. We found that the BAR domain protein Snx33 localized to the leading edge, and destabilized it by inhibiting the major actin nucleation promoting factor WAVE2. This mechanism was indeed required for navigation, as Snx33 knockout cells fail to change direction when hitting inert or cellular obstructions but continued to migrate persistently. Our results show how cells can read out their surface topography to interpret their environment, allowing them to rapidly switch between persistent and exploratory migration in order to circumvent obstacles.