Microfluidically Partitioned Dual Channels for Accurate Background Subtraction in Cellular Binding Studies by Surface Plasmon Resonance Microscopy

Unlike conventional surface plasmon resonance (SPR) using an antifouling film to anchor biomolecules and a reference channel for background subtraction, SPR microscopy for single-cell analysis uses a protein- or polypeptide-modified gold substrate to immobilize cells and a cell-free area as the reference. In this work, we show that such a substrate is prone to nonspecific adsorption (NSA) of species from the cell culture media, resulting in false background signals that cannot be correctly subtracted. To obtain accurate kinetic results, we patterned a dual-channel substrate using a microfluidic device, with one channel having poly-l-lysine deposited in situ onto a preformed polyethylene glycol (PEG) self-assembled monolayer for cell immobilization and the other channel remaining as PEG-covered for reference. The two 2.0 mm-wide channels are separated by a 75 μm barrier, and parts of the channels can be readily positioned into the field of view of an SPR microscope. The use of this dual-channel substrate for background subtraction is contrasted with the conventional approach through the following binding studies: (1) wheat germ agglutinin (WGA) attachment to the N-acetyl glucosamine and N-acetyl-neuraminic acid sites of glycans on HFF cells, and (2) the S1 protein of the COVID-19 virus conjugation with angiotensin-converting enzyme 2 (ACE2) on the HEK293 cells. Both studies revealed that interferences by NSA and the surface plasmon polariton wave diffracted by cells can be excluded with the dual-channel substrate, and the much smaller refractive index changes caused by the injected solutions can be correctly subtracted. Consequently, sensorgrams with higher signal-to-noise ratios and shapes predicted by the correct binding model can be obtained with accurate kinetic and affinity parameters that are more biologically relevant. The affinity between S1 protein and ACE2 is comparable to that measured with recombinant ACE2, yet the binding kinetics is different, suggesting that the cell membrane does impose a kinetic barrier to their interaction.