Molecular interaction is the decisive step in the vast majority of biological processes. Interactions between proteins, or between proteins and low molecular weight compounds, are therefore of interest for a wide range of pharmacological targets. Consequently about half of the assays used in drug discovery are molecular interaction assays performed in a cell-free environment.
The ideal molecular interaction assay should work in homogeneous phase without a separation step and have a high signal-to-noise ratio. One assay technology that satisfies these requirements is based on Fluorescence Resonance Energy Transfer (FRET), a technology that requires two interaction partners of which one is labeled with a donor fluorophore and the other is labeled with an acceptor fluorophore.
Transfer of energy from the excited donor to the acceptor is only seen when the two fluorophores are in close proximity, typically below 10 nm. Its distance dependence of 1/R6 makes FRET the assay of choice for measuring true molecular interactions. The quality of the signal can be further improved when assays are based on a long-lived donor fluorophore. This allows the use of gated detection systems that are insensitive to the short-lived fluorescence, which is intrinsic to samples, thus reducing the background and resulting in excellent signal-to-background ratios. These systems are known as time-resolved FRET or TR-FRET systems.
One such system is Homogeneous Time Resolved Fluorescence (HTRF) from Cisbio international. In HTRF, a rare earth cryptate works as a long-lived fluorescence donor, while either allophycocyanine (a fluorescent protein complex from blue algae) or a red fluorophore such as Cy5 or Dy647 (GE Healthcare, Dyomics) is used as an energy acceptor (Figure 1).
The quality of assays based on HTRF has led to interest in these tools, in particular for high-throughput screening. For a typical assay based on HTRF, the two interacting compounds need to be labeled with the donor and the acceptor respectively. One common way of doing this is by epitope tagging of the relevant proteins, e.g., by a FLAG-tag, a His-tag, or an HA-tag, and by using donor-labeled and acceptor-labeled antibodies for the corresponding protein tags.
While this increases assay complexity, it is a widely used approach. Tools for assay development as well as complete assays for particular targets, e.g., second messenger molecules, are commercially available.
An alternative tagging approach enables direct covalent labeling of the proteins of interest, eliminating the need for antibodies and reducing the assay complexity. The SNAP-tag technology from Covalys (www.covalys.com) consists of a protein tag that reacts covalently with a labeled substrate, forming a stably labeled fusion protein in the process. Substrates are derivatives of O6-benzylguanine (BG) that can carry a wide range of different labels. Substrates that carry a donor cryptate (BG-TBP) or an acceptor-fluorophore (BG-647) suitable for HTRF have recently become available. Use of the SNAP-tag reduces the size of the overall protein complex as compared to the use of one or even two antibodies. This results in high energy transfer efficiencies and also streamlines the development of HTRF assays.
As a model system, a protein-protein interaction depending on the presence of a low molecular weight compound was selectedthe high affinity interaction of FK-506-Binding Protein (FKBP) with the FKBP-Rapamycin-Binding domain of FRAP (FRB, Figure 2). This interaction is dependent on the presence of Rapamycin or similar compounds (Rapamycin from Calbiochem, other compounds from Ariad Pharmaceuticals).
Both proteins were expressed in E. coli as SNAP-tag fusions extended by a poly-histidine tag. Bacteria were collected by centrifugation, lysed, and the SNAP-tag fusion proteins were purified by Ni-NTA-chromatography. The purified protein was stored at -70C in the presence of 1 mM DTT. For labeling with the Cryptate donor, BG-TBP was added to the SNAP-tag fusion protein at a molar ratio of 2:1 and reacted for 1 hour at RT. Unbound BG-TBP was removed using a NAP5-column (GE Healthcare). For labeling of SNAP-tag fusion protein with the acceptor, BG-647 was added to the SNAP-tag fusion protein at a molar ratio of 2:1 and reacted for 1 hour at RT.
After 1 hour, unreacted BG-647 was removed using a Micro Bio-Spin column containing BioGel P30 (Bio-Rad). All assays were run in HTRF buffer (100mM phosphate, 50100mM KF, 0.5%BSA, 0.1%Triton). Time-resolved FRET was measured on a RUBYstar platereader (BMG Labtec). The concentration of the protein labeled with the BG-TBP was adjusted to give between 25,000 and 40,000 counts at 620 nm, corresponding to a maximum concentration of 12 nM for the labeled protein. All assay results are expressed as delta F%.
Figure 3 shows the titration of 8 nM donor-labeled SNAP-FKBP with 0.5 to 16 nM acceptor labeled SNAP-FRB in the presence of 1 M of the rapamycin-like drug A21967. The delta F% value already indicates considerable energy transfer at 0.5 nM and levels off between 8 nM and 16 nM SNAP-FRB at above 4,500 delta F%. This indicates on the one hand a strong interaction between the two proteins and on the other hand an effective energy transfer between the donor and the acceptor label. The maximum energy transfer values observed for the FKBP-FRB system were above 20,000 delta F% under optimized conditions, which indicates close proximity of donor and acceptor (data not shown).
The effect of a low molecular weight inhibitor of the FKBP-FRB interaction, Ascomycin, was tested subsequently. Donor-labeled SNAP-FKBP and acceptor-labeled SNAP-FRB were both used at 12 nM, and the Rapamycin analogue AP21967 was added to give half-maximal interaction (at 250 nM). The system showed a response of about 2000 delta F% under these conditions. Ascomycin was titrated in at concentrations from 1 nM to 100 M.
A sigmoidal displacement curve was obtained with a 50% inhibition at about 10 nM (Figure 4). For repetitive use of the assay a Z value above 0.7 was found, indicating a rugged system. Finally, SNAP-tag labeling works equally well in the presence of crude lysates, allowing the direct labeling and use in HTRF assays of delicate proteins that tend to lose binding activity during purification.
For the development of HTRF assays, the SNAP-tag presents a straightforward way to label proteins for measurement of protein-protein interactions. The ease of the labeling approach, obviating the need to use antibodies, and the possibility to work from complex samples make this a valuable tool for assays in HTS and beyond. This is supported by the high delta F% values and high Z value obtained in the FKBP-FRB model system.
The SNAP-tag technology is not limited to HTRF assays. The simple modification of tagged proteins by fluorescent benzylguanine substrates or affinity substrates allows the straightforward development of fluorescence-based microplate binding assays. The exquisite specificity of the SNAP-tag reaction allows labeling of SNAP-tag fusion proteins inside living cells, with the potential to do pulse chase labeling of one protein or selective labeling of receptor proteins in the cell membrane to follow receptor internalization. The SNAP-tag can also be used for the selective immobilization of proteins, with applications including pull-down assays, creation of protein arrays, and capture on surface plasmon resonance chips.