QCM in the analysis of Protein Interactions

The sensor can be used for the direct, marker-free measurement of specific interactions between immobilized molecules and analytes in solution. Binding of a soluble analyte to the immobilized ligand causes a shift in the resonance frequency, and this signal can be recorded using a frequency counter with high resolution. This method, despite its existence for four decades, has only recently been developed for immunological measurements in a flowthrough system, as shown in Figures 1 and 2. The resulting frequency versus time curve is called a “sensorgram." As real-time measurements are performed, the sensorgrams are capable of deducing not only the equilibrium binding constants, but also the affinity rate constants. This methodology was recently applied to describe the binding kinetics of various phage-presented proteins, including recombinant antibodies.

In QCM, the ligand is immobilized as a monolayer on a flat gold surface. Immobilization can be achieved using various methods. The easiest way is absorption of the ligand on the gold surface. This method, however, does not allow regeneration of the ligand layer. As the gold reacts with thiols, yielding a stable, semi-covalent bond, proteins can be immobilized by the thiol groups of their cysteine residues. Alternatively, the sensor surface can be activated by using a thiol-containing bifunctional linker. Reproducible results were obtained with dithiobissuccinimidyl-propionate (DSP). DSP forms disulfide bonds to the gold surface and provides N-hydroxysuccinimide (NHS) groups that can react with the free amino groups on the ligand. If streptavidin is immobilized using DSP, biotinylated ligands can be conveniently coated. The availability of in vivo or in vitro systems for sequence-specific biotinylation should allow the generation of highly ordered ligand monolayers. Samples are applied to the immobilized ligand on the sensor surface by a continuous constant flow. A constant analyte concentration at any part of the flowthrough cell is thus provided, and diffusion effects can be neglected. Moreover, an optimized cell design helps to avoid mass transport effects. Viscosity effects, however, can lead to frequency shifts interfering with the specific signal. Therefore, the biosensor should be calibrated with a viscous solution to determine the time in which the viscosity change generates a signal. This time interval should be excluded from the sensorgram. A sample sensorgram is given in Figure 3. After a measurement, the sensor has to be regenerated to remove bound analyte from the sensor surface. For this regeneration step, elution buffers as used in affinity chromatography can be applied as long as they do not destroy the native structure of the ligand. A good overview regarding ligand stability is given in Harlow and Lane. It should be noted that the QCM method does not allow the measurement of true affinities. Because the ligand is immobilized, one degree of freedom is lost. Therefore, the measured affinities could be influenced by decreased mobility and sterical hindrance. Furthermore, avidity effects resulting from multivalent binding might affect the apparent affinity. On the other hand, partial denaturation of the immobilized ligand may decrease the apparent binding affinity. In the case of phage-presented derivatives, as tested by Hengerer et al. the affinity data obtained are only apparent, because the fraction of total phage carrying functional single-chain fragments of an antibody (scFv fragments) may not be 100%, and a fraction of surface-expressed scFv antibodies could be shed by proteolysis. In some phage-display systems, soluble scFv fragments may be present in the phage supernatants due to the existence of an amber codon between scFv and the virus coat protein. Finally, more than one fragment could be presented per phage. Thus, for the determination of true affinities an accurate determination of true molarities of scFv fragments would be required. Nevertheless, this method allows the ranking of a set of antibody clones from a phage-screening experiment according to their apparent affinities.

A QCM analysis in this protocol is broken down into the following steps: sensor cleaning, immobilization of the ligand on the sensor, sample preparation, system calibration, measurement procedure, optimization of regeneration buffer, and evaluation of sensorgram. Sensor cleaning and immobilization of ligand are performed before the sensor chip is mounted onto the AFFCo 2000 device. Sensor cleaning should be performed directly before immobilization. The sensor chip with the immobilized ligand can be stored for several days at 4°C in a wet chamber (storage conditions depend on the immobilized protein).