Non-specific protein-surface interactions in the context of particle based biosensors

Biosensors are compact devices that can be used at the point-of-care to measure low quantities of biomarkers in a complex body fluid like blood or saliva. Many biomarkers in the human body are protein molecules, for example the biomarkers that are used for the diagnosis of myocardial infarction. Sensitive detection of protein biomarkers in a biosensor is enabled by the use of molecules with a high specific affinity for the biomarker; typically these are antibodies. The basic concept of a particle-based biosensor is that the specific binding of antibodies to the protein biomarkers causes particle labels to become attached to a sensor surface, and thereby the amount of bound particles becomes a measure for the biomarker concentration in solution. The sensitivity and specificity of the biosensor are determined by the specific as well as the non-specific interactions in the assay. Non-specific interactions can for example induce particle binding without biomarkers being present, leading to false positive signals. Therefore it is important to reduce the non-specific interactions as much as possible and to try to understand the underlying mechanisms. Biosensor surfaces are often made of a polymeric material like polystyrene because of its excellent manufacturing and modification properties. There are a number of strategies to reduce non-specific binding on polymers, for example blocking of the surface with a protein like BSA or increasing the surface hydrophilicity by an oxidative treatment. Within the non-specific binding processes several physicochemical interactions play a role like electrostatic, van derWaals, hydrogen bonding, hydrophobic and steric or roughness related interactions; and on top of that protein conformations can change. However, the effects of the surface treatments on non-specific interactions with particles and single proteins have not been characterized and understood in detail. The first topic addressed in this work was to develop a model polystyrene surface with well controlled properties that could be varied to influence the non-specific interactions. We chose a spincoated polystyrene surface that was flat on the scale of proteins (as shown by AFM measurements, Rq <0.55 nm) to minimize the role of roughness in the interactions. The polymer surfaces were treated with UV/ozone in order to control the hydrophilicity of the surface. A direct relation was observed between treatment time and hydrophilicity, with an increase of surface oxygen content (measured by XPS, upto 24% from 300 s oxidation) causing a reduction of water contact angle. These UV/ozone treated polystyrene surfaces with a range of different hydrophilicities yield a platform for studying protein interactions. The first interaction we quantified was the non-specific association of proteincoated particles to the polystyrene surface. We described the interaction with an energy barrier for association that depends on the properties of both the surface and the solution. The experiments were performed with superparamagnetic particles; such particles are known to be suitable labels in integrated high-sensitivity biosensors due to the fact that the particles can be manipulated by magnetic fields. The particles consist of a polystyrene matrix filled with magnetic nanoparticles. In our experiments the particles were coated with myoglobin, a well-established cardiac biomarker. A new technique, the rotating particles probe, was used to quantify the fraction of unbound particles by measuring their response to a rotating magnetic field. To describe the non-specific binding process we propose a model with a distribution of energy barrier values and this model was shown to accurately fit the measured data. The extracted parameters signified high energy barrier values for binding on surfaces with long oxidation times and for solutions with high pH or low ionic strength. Both hydrophilicity and electrostatic interactions play an important role in the observed non-specific association. The energy barrier for association could be quantified by using the energy barrier spread from dissociation measurements: at physiological buffer conditions (150 mM) the energy barriers were found in the range 0–60 kBT. Next, we studied the dissociation of non-specifically bound particles. For this purpose we used magnetic tweezers and the same protein-coated particles as for the association measurements. Forces applied by the magnetic field gradient were calibrated by time-of-flight measurements. The force-induced dissociation measurements were performed by recording the number of bound particles as a function of time during the application of a constant force of 30 pN, 50 pN or 70 pN. The data appear not to obey the dissociation kinetics of bonds with a single energy barrier as in the standard Bell model. We show that the dissociation of the non-specific bonds can be modeled with a distribution of energy barrier values. The fits reveal that the energy barrier for unbinding decreases for increasing oxidation time (from 90 s to 300 s) from 38 kBT to 25 kBT with a constant spread of (7 4) kBT. So the association as well as the dissociation experiments show that the hydrophilicity of the surface is an important determinant for non-specific interactions; and both processes reveal a distribution of energy barriers rather than a single energy barrier. Finally, we zoomed in on the non-specific interaction between single proteins and a polystyrene surface. AFM tips were functionalized with single myoglobin molecules and force–distance curves were recorded. In the retraction curves clear steps were observed. These steps entail that the tip does not detach from the surface at once. The events appearing in between the steps have two different characteristics: either the force stays constant during retraction or the force increases like the stretching of a spring. There are two processes to which these events can be attributed: first, the pulling of polymer chains from the polystyrene surface and, second, disruption associated with the protein structure. The data reveal that the non-specific interaction between a single protein and a polymer surface can be stronger than the internal structure of the protein. These results provide new experimental approaches to study non-specific interactions between protein-coated particles and biosensor surfaces. We have learned that the interactions can be described by a generalized interaction potential that is characteristic for the properties of the surface and the composition of the solution. An important finding is that the energy barriers for association as well as for dissociation are given by a distribution rather than a single value or a set of values. Furthermore, first experiments on the non-specific binding of a single myoglobin molecule showed that the non-specific forces can be stronger than the internal protein structure. Overall, the experiments form a first step and foundation for the study of non-specific interactions between polymer surfaces and protein-coated particles as well as single protein molecules. Further research should focus on extensions to different proteins, surfaces and solution compositions, in order to study the validity range of the model descriptions; and on a stepwise increase of the complexity of the materials system with the aim to develop a complete understanding of the specific and non-specific interactions in biosensor assays.