Search

The ability to determine the identity of specific proteins is a critical challenge in many areas of cellular and molecular biology, and in medical diagnostics. Here, we present a macine learning aided microfluidic protein characterisation strategy that within a few minutes generates a three-dimensional fingerprint of a protein sample indicative of its amino acid composition and size and, thereby, creates a unique signature for the protein. By acquiring such multidimensional fingerprints for a set of ten proteins and using machine learning approaches to classify the fingerprints, we demonstrate that this strategy allows proteins to be classified at a high accuracy, even though classification using a single dimension is not possible. Moreover, we show that the acquired fingerprints correlate with the amino acid content of the samples, which makes it is possible to identify proteins directly from their sequence without requiring any prior knowledge about the fingerprints. These findings suggest that such a multidimensional profiling strategy can lead to the development of a novel method for protein identification in a microfluidic format.

The biological function of proteins is dictated by the formation of supra-molecular complexes that act as the basic machinery of the cell. As such, measuring the properties of protein species in heterogeneous mixtures is of key importance for understanding the molecular basis of biological function. Here, we describe the combination of analytical microfluidic tools with liquid chromatography for multidimensional characterisation of biomolecules in complex mixtures in the solution phase. Following chromatographic separation, a small fraction of the flow-through is distributed to multiple microfluidic devices for analysis. The microfluidic device developed here allows the simultaneous determination of the hydrodynamic radius, electrophoretic mobility, effective molecular charge and isoelectric point of isolated protein species. We demonstrate the operation principle of this approach with a mixture of three unlabelled model proteins varying in size and charge. We further extend the analytical potential of the presented approach by analysing a mixture of interacting streptavidin with biotinylated BSA and fluorophores, which form a mixture of stable complexes with diverse biophysical properties and stoichiometries. The presented microfluidic device positioned in-line with liquid chromatography presents an advanced tool for characterising multidimensional physical properties of proteins in biological samples to further understand the assembly/disassembly mechanism of proteins and the nature of complex mixtures.

The ability to apply highly controlled electric fields within microfluidic devices is valuable as a basis for preparative and analytical processes. A challenge encountered in the context of such approaches in conductive media, including aqueous buffers, is the generation of electrolysis products at the electrode/liquid interface which can lead to contamination, perturb fluid flows and generally interfere with the measurement process. Here, we address this challenge by designing a single layer microfluidic device architecture where the electric potential is applied outside and downstream of the microfluidic device while the field is propagated back to the chip via the use of a co-flowing highly conductive electrolyte solution that forms a stable interface at the separation region of the device. The co-flowing electrolyte ensures that all the generated electrolysis products, including Joule heat and gaseous products, are flowed away from the chip without coming into contact with the analytes while the single layer fabrication process where all the structures are defined lithographically allows producing the devices in a simple yet highly reproducible manner. We demonstrate that by allowing stable and effective application of electric fields in excess of 100 V cm -1 , the described platform provides the basis for rapid separation of heterogeneous mixtures of proteins and protein complexes directly in their native buffers as well as for the simultaneous quantification of their charge states. We illustrate this by probing the interactions in a mixture of an amyloid forming protein, amyloid-β, and a molecular chaperone, Brichos, known to inhibit the process of amyloid formation. The availability of a platform for applying stable electric fields and its compatibility with single-layer soft-lithography processes opens up the possibility of separating and analysing a wide range of molecules on chip, including those with similar electrophoretic mobilities.