Kinetics of Antibody Binding to Membranes of Living Bacteria Measured by a Photonic Crystal-Based Biosensor

Optical biosensors based on photonic crystal surface waves (PC SWs) offer a possibility to study binding interactions with living cells, overcoming the limitation of rather small evanescent field penetration depth into a sample medium that is characteristic for typical optical biosensors. Besides this, simultaneous excitation of s- and p-polarized surface waves with different penetration depths is realized here, permitting unambiguous separation of surface and volume contributions to the measured signal. PC-based biosensors do not require a bulk signal correction, compared to widely used surface plasmon resonance-based devices. We developed a chitosan-based protocol of PC chip functionalization for bacterial attachment and performed experiments on antibody binding to living bacteria measured in real time by the PCSW-based biosensor. Data analysis reveals specific binding and gives the value of the dissociation constant for monoclonal antibodies (IgG2b) against bacterial lipopolysaccharides equal to KD = 6.2 ± 3.4 nM. To our knowledge, this is a first demonstration of antibody-binding kinetics to living bacteria by a label-free optical biosensor.

. A single exponential binding model. The simple binding model is derived from the law that the velocity of a chemical reaction is proportional to the concentrations of the ligand and receptor [1]. It is a simple approximation of a binding event assuming that all receptors are equally available for binding and it does not allow partial binding. This model also assumes that neither ligand not receptor is modified after binding implying reversibility of the reaction.
A reversible receptor-ligand interaction can be described by the following reaction: Let Y be a concentration of receptor-ligand complexes at any time ( Ymax be a binding capacity of the receptor attached onto the surface. The rate of a complex formation (number of binding events per time) is proportional to the ligand concentration and amount of unoccupied receptor. Since not every ligand-receptor collision results in a complex formation, the association rate constant kon has to be introduced: If now L is a thickness of the adsorbed layer and Lmax is a maximum adlayer thickness proportional to the receptor concentration (when all sites are occupied), the net association rate defined as a time derivative of the thickness dL/dt is equal to the difference between the forward binding and dissociation process: After the integration, assuming that at t = 0 the adlayer thickness is zero, we obtain the expression for the simplest binding model described by a single exponential function: The adlayer thickness at the binding equilibrium obtained at infinite time is defined as follows: max From this derivation the observed rate constant kobs is an exponent in the exponential function and depends linearly on the ligand concentration. It is a measure of time needed to attain an equilibrium between association and dissociation processes during the complex formation. Thus the binding rate depends on association and dissociation rate constants, and linearly depends on the ligand concentration. A dissociation constant KD can be calculated from the series of experiments with a variable concentration from the linear fit of kobs as a ratio of koff/kon. If the simple binding model is not valid for a certain receptor-ligand system, the plot kobs against the ligand concentration will deviate from a straight line.
If the binding does not obey a linear dependence we may infer that the binding is more complex than it can be described the simple model. In this case conformational changes, covalent or hydrogen bonds formation or other phenomena may take place.

Viability Justification for the Bacteria
For the experiment, bacteria were harvested fresh after an overnight incubation, dispersed in PBS, and pumped through the flow cell. Viability of the bacteria left in the suspension after the experiment was confirmed by growing new bacteria from the culture used in the experiment.
Viability of bacteria attached to the PC chip was proved using a standard two-color fluorescence assay (LIVE/DEAD ® BacLight Bacterial Viability Kit, Molecular Probes) based on different permeability into a bacterial membrane. A mixture of SYTO ® 9 green-fluorescent nucleic acid stain and the red-fluorescent nucleic acid stain propidium iodide were used for the viability test. The SYTO 9 stain labels all bacteria that are incubated with it, whereas propidium iodide stains only bacteria with damaged membranes. After the kit application, living bacteria with intact membranes are stained in fluorescent green, whereas bacteria with damaged membranes appear red. The images show that only few bacteria are dead while almost all of them are alive. Since any freshly grown bacterial culture contains a small number of dead bacteria, the obtained results of fluorescence microscopy suggest that the technique used here preserves bacterial viability and does not result in bacterial death. A solution of polyclonal antibodies consists of several populations of heterogeneous antibodies with corresponding binding affinities and antigen epitope specificities within an antibody population. During an experiment with different antibody concentrations, a heterogeneous antibody sample is interacting with the bacteria as illustrated by different colors on Figure S3 (an antibody image size is associated with a different binding affinity). Due to the sample diversity, kinetic characteristics of each sample of different concentration are dependent on the presence of subpopulations of the antibodies in a polyclonal pool and, therefore, cannot be unified and described by a single binding constant.