Optimising Protein-Protein and Protein-Small Molecule Kinetics Assays
Phil Buckle, Application Manager, Europe, ForteBio UK, Ltd.
Kinetic analysis of protein-protein and protein-small molecule interactions is a key application for real-time, label-free systems such as the Octet family of instruments. The Octet platform is currently utilized in various segments of the pharmaceutical and biotherapeutic drug development processes such as early discovery, process development, late-stage clinical trials and manufacturing/QC. Applications include quantification of protein concentrations in crude media, screening antibodies in crude supernatants, characterization of protein-protein interaction kinetics (ka, kd and KD), and screening small molecule fragment libraries.
In this discussion, tips and tricks for obtaining the best possible kinetic data will be presented. This will include optimisation of target immobilisation and analyte binding, along with regeneration strategies (if required) and sample plate and method design. The Octet analysis software options will also be discussed, giving details of the analysis models and their applicability to data interpretation.
Target Molecule Immobilisation
The most important consideration in choosing a strategy for target molecule immobilisation is the retention of biological activity.
Whilst direct covalent attachment results in stable, non-reversible target immobilisation, a number of factors need to be considered. Direct coupling is usually performed directly to the biosensor surface, such as with amine coupling. This has the potential for loss of target activity resulting from steric hindrance. In amine coupling procedures, it is difficult to restrict the number of sites on the target molecule which are linked to the biosensor surface. Potentially, any free lysine residue can be involved in the linking and if any are close to the analyte binding site, then loss of target activity can result.
The following requirements should be considered when performing amine coupling:
- The target must be pure, and must not contain any extraneous amines, or be diluted in any amine-containing buffer.
- The target must also be prepared in low salt buffer at a pH just below its pI value, to maintain a balance between creating enough charge to attract it to the biosensor on the one hand, but keep as much unprotonated lysine as possible on the other hand.
- If the target has been lyophilized from buffer, it may be necessary to perform a desalting step to reduce the buffer ionic strength.
For more information on amine coupling, refer to ForteBio’s Technical Note No. 7, Batch Immobilization of Protein Onto Amine Reactive Biosensors.
A site-directed approach is recommended to maximize surface activity. Favorable orientation of target on the biosensor surface can be achieved by the use of specific capture approaches aimed to couple the target via a known position or label. Additionally, steric hindrance can be minimized by the inclusion of chemical linkers.
Several oriented capture options are available on the Octet platform. The use of such a capture approach also enables immobilisation of targets from crude preparations.
A second important consideration in choosing an immobilisation technique is minimal dissociation of target from the biosensor. The most favourable in this respect are amine reactive and streptavidin biosensors, though others can be assessed for target capture stability.
The most common, and one of the most favored, capture systems employs biotin labeled target and streptavidin biosensors. This approach has a number of advantages over direct amine coupling:
- Biotinylation is performed in solution at neutral pH.
- The ratio of biotins per target molecule can be controlled.
- It is easy to incorporate a long chain linker into the biotin tag to minimize steric effects.
- The biotinylated target can be prepared in batches and re-used for multiple capture experiments. Refer to ForteBio’s Technical Note No. 10 Batch Immobilization of a Biotinylated Ligand onto Streptavidin Biosensors for more information.
Analyte can be prepared in most commonly used buffer systems, but it is important that the concentration is known. Any errors in this concentration will be carried through to errors in the calculated affinity.
We recommend ForteBio’s kinetics buffer (available as a 10X solution; ForteBio Part number 18-5032) which contains PBS + 0.1% BSA, 0.02% Tween20 and 0.05% sodium azide as a sample buffer unless the particular interaction calls for another buffer composition.
Anti-Human Fc Capture (AHC)
Human Fc tag
Anti-Murine Fc Capture (AMC)
Murine Fc tag
Table 1: Octet biosensors that allow site-directed immobilization of target.
Optimising the Kinetics Assay
Biotinylate the target using either NHS-PEG4-Biotin (Pierce, part no. 21329) or NHS-LC-LC-Biotin (Pierce, part no. 21343). Label at a molar ratio of about 1:1 biotin :target. It is important to remove unreacted biotin reagent prior to immobilisation. This is usually performed using desalt columns such as Zeba Micro Desalt spin columns (Pierce, part no. 89877). For more information on biotinylation of target proteins, refer to the following technical notes.
- Technical Note No. 6, Biotinylation of Protein for Immobilization onto Streptavidin K Biosensors
- Technical Note No. 11, Biotinylating Antibody in Stocks Containing Carrier Protein
- Technical Note No. 12, Biotinylating Very Small Quantities of Protein for Immobilization onto Streptavidin Biosensors
Initially, use the parallel processing capability of the Octet system to optimise the target concentration. Use a dilution series from 25 µg/mL to 10 µg/mL, aiming for biosensor saturation. Use the ‘extend current step’ and ‘skip to next step’ buttons during the experiment to optimise the loading times.
Ideally, the analyte dilution series should cover the affinity constants range of 10xKD to 0.1xKD. If the KD is unknown, begin with a high concentration and titrate down. For robust kinetics analysis, use at least 5 analyte concentrations, preferably in duplicate.
If the expected KD is in the nM range, an association of 15 minutes and a dissociation of 15–30 minutes may be sufficient to obtain kinetic constants with low error.
If the KD is < 1 nM, an association of 15–30 minutes and a dissociation of one hour or more may be necessary to obtain kinetic constants with low error.
If the KD is unknown, use the ‘extend current step’ and ‘skip to next step’ buttons during the experiment to optimise the analyte binding times.
The unique 8- or 16-channel simultaneous processing employed in Octet systems allows accurate kinetics constants to be measured without the need for regeneration at all. Still, in most cases, biosensors can be regenerated (removal of bound analyte) for re-use, which has favorable implications on assay costs.
For interactions of modest affinity (around 100 nM or lower), the analyte may often take a long time to fully dissociate. In such cases, if the biosensors are to be re-used, they should be regenerated. Capture biosensors such as Anti-Human Fc and Anti-Murine Fc Capture can be regenerated by removal of both analyte and target molecules. This is usually achieved using low pH, but should be optimised for each interacting pair. For more information on regenerating Anti-Human Fc and Anti-Murine Fc Capture biosensors, refer to the corresponding datasheet or technical note.
Amine reactive and Streptavidin biosensors are regenerated by removing only the analyte molecule, due to the irreversible nature of target immobilisation. For successful regeneration of biosensors, complete removal of bound analyte and retention of target activity are essential.
To optimise regeneration conditions, use the Octet system’s parallel processing to investigate 8 or 16 regeneration conditions in a single experiment. Figure 1 shows how the Octet system was used to assess 8 regeneration conditions, over 11 sequential analyte binding cycles (total assay time: 2 hours).
Using the grouping function in ForteBio’s Analysis software, rapid evaluation of each regeneration condition is possible, as shown in Figure 2.
Reference Wells and Biosensors
To obtain accurate protein-protein interaction kinetics, referencing-out the buffer effects are important. Include a reference sample well containing buffer in the sample plate in your experiment. For protein-small molecule interactions, include a reference biosensor to perform double referencing.
Reference biosensors should ideally include an immobilized non-active protein similar to the specific target protein. Avoid BSA as the non-active protein as it is prone to non-specific binding interactions. If no suitable non-active protein is available, block the active sites on the biosensor (for example, use biocytin to block Streptavidin and Super streptavidin biosensors).
For more information on protein-small molecule kinetics on the Octet platform, refer to ForteBio’s Technical Note No. 16, Small Molecule Binding Kinetics.
Setting up the Kinetics Method
A kinetics experiment requires baseline, association and dissociation steps to be set up in consecutive order. Octet data acquisition software simplifies assay set via experiment templates. Assay optimisation for pH scouting and regeneration scouting can be set up quickly using the templates.
Other factors to note are that the analyte concentrations should be entered in the software in MOLAR terms, and, to include at least one well of buffer (0 nM sample) for reference subtraction. Kinetic measurements under non-mass transport limiting conditions can typically be achieved using a shake speed of 1000 rpm.
Choosing a model for curve fitting in the Octet data analysis software should be done on the basis of an understanding of the chemistry underlying the binding interaction. As a general rule, use the simplest curve fitting model available (1:1) unless you have prior knowledge that the interaction is more complex. Before trying a curve fit model, get as much information about the interaction under study as possible, such as stoichiometry of binding, purity of ligand and heterogeneity of analyte.
Use other curve fit models in Octet data analysis software to assess experimental optimisation. For example, the 2:1 heterogeneous ligand binding model will show if the surface contains multiple affinity binding sites. If this model gives a good fit compared to 1:1, try to optimise the target loading level further, or consider oriented capture of the target on the biosensor.
The mass transport binding model will show if the analyte is diffusion-limited at the surface. If you suspect the presence of mass transport limitation, reduce the target immobilisation level and/or increase the shake speed can be increased up to 1500 rpm.
By careful optimisation of experimental parameters, kinetics analysis on the Octet systems will yield excellent kinetics data. Oriented, site-specific coupling of target molecule to the biosensor provides the best homogeneous surface. The optimal immobilization level of target on the biosensor should be arrived at by scouting target solution concentration and incubation time. Analyte concentration is ideally explored from 0.1xKD to 10xKD with multiple concentrations in-between. Use at least one reference sample and perform double-referencing for small molecule assays. Set up the experiment such that the interaction occurring at the biosensor follows a 1:1 binding model. Good luck with your experiments!