Biosensors based on plasmonic nanostructures are widely used in various applications and benefit from numerous operational advantages. One type of application where nanostructured sensors provide unique value in comparison with, for instance, conventional surface plasmon resonance, is investigations of the influence of nanoscale geometry on biomolecular binding events. In this work, we show that plasmonic “nanowells” conformally coated with a continuous lipid bilayer can be used to detect the preferential binding of the insulin receptor tyrosine kinase substrate protein (IRSp53) to regions of negative surface curvature, i.e. the interior of the nanowells. Two different sensor architectures with and without an additional niobium oxide layer are compared for this purpose. In both cases, curvature preferential binding of IRSp53 (at around 0.025 nm-1 and higher) can be detected qualitatively. The high refractive index niobium oxide influences the near field distribution and makes the signature for bilayer formation less clear, but the contrast for accumulation at regions of negative curvature is slightly higher. This work shows the first example of analyzing preferential binding of an average-sized and biologically important protein to negative membrane curvature in a label-free manner and in real-time, illustrating a unique application for nanoplasmonic sensors.

We present a new grafting-to method for resistant “non-fouling” poly(ethylene glycol) brushes, which is based on grafting of polymers with reactive end groups in 0.9 M Na₂SO₄ at room temperature. The grafting process, the resulting brushes, and the resistance toward biomolecular adsorption are investigated by surface plasmon resonance, quartz crystal microbalance, and atomic force microscopy. We determine both grafting density and thickness independently and use narrow molecular weight distributions which result in well-defined brushes. High density (e.g., 0.4 coils per nm² for 10 kDa) and thick (40 nm for 20 kDa) brushes are readily achieved that suppress adsorption from complete serum (10× dilution, exposure for 50 min) by up to 99% on gold (down to 4 ng/cm² protein coverage). The brushes outperform oligo(ethylene glycol) monolayers prepared on the same surfaces and analyzed in the same manner. The brush heights are in agreement with calculations based on a simple model similar to the de Gennes “strongly stretched” brush, where the height is proportional to molecular weight. This result has so far generally been considered to be possible only for brushes prepared by grafting-from. Our results are consistent with the theory that the brushes act as kinetic barriers rather than efficient prevention of adsorption at equilibrium. We suggest that the free energy barrier for passing the brush depends on both monomer concentration and thickness. The extraordinary simplicity of the method and good inert properties of the brushes should make our results widely applicable in biointerface science.