Fibronectin

fibronectin-coated PDMS surfaces were immunostained. Alternatively, rhodamine-labelled fluorescent fibronectin (20 μg mL⁻¹, Cytoskeleton Inc.) was used. The transversal cross-section of the topography was obtained from reslicing of a vertical confocal z stack (z step 0.37 μm, 60 × lense of AN 1.4 yelding an optical slice of 1 μm) and subsequent binarizing and skeletonizing the fibronectin signal (Supplementary Fig. S1A). An equation $y=f(x)$ was fitted to the cross-section by non-linear regression to describe the surface altitude as function of position along the transversal axis (Supplementary Fig. S1B). The curvature κ was then obtained at each point by $k=\frac{\ddot{y}}{{\left(1+{\dot{y}}^2\right)}^{\frac{3}{2}}}$ where the dot and double dot represent the first and second derivative of f(x), respectively. The local maxima and minima of κ over a single period of the topographical pattern correspond to the most convex and most concave points, respectively. Resulting values are shown in Supplementary Fig. S1C,E. Color-coded representations (Supplementary Fig. S1D) of the local curvature were produced by averaging κ between surfaces at each points and using binarized fibronectin staining as mask.

The micropatterned substrates consisting of 5 μm lines with 5 μm inter-line spacing were fabricated via standard photolithography techniques, according to previous protocols [25 (link)]. Briefly, the polydimethylsiloxane (PDMS) stamp was incubated with 50 μg/ml of fibronectin (Cytoskeleton, Denver, CO, USA) for 1 hour at room temperature. The stamp was then dried and put in contact with a PDMS coated coverslip for 15 minutes to imprint the pattern. Uncoated regions were blocked by immersing the micropatterned coverslips for 5 min in a 1% solution of Pluronic F-127 (Sigma-Aldrich, St. Louis, USA). Finally, the coverslips were three times washed with PBS and stored in PBS at 4°C until use.