This protocol presents a technique for probing protein-protein interactions using glutathione-linked donor beads with GST-fused TPR-motif co-chaperones and acceptor beads coupled with an Hsp90-derived peptide. We have used this technique to screen small molecules to disrupt Hsp90-FKBP51 or Hsp90-FKBP52 interactions and identified potent and selective Hsp90-FKBP51 interaction inhibitors.
Targeting the heat shock protein 90 (Hsp90)-cochaperone interactions provides the possibility to specifically regulate Hsp90-dependent intracellular processes. The conserved MEEVD pentapeptide at the C-terminus of Hsp90 is responsible for the interaction with the tetratricopeptide repeat (TPR) motif of co-chaperones. FK506-binding protein (FKBP) 51 and FKBP52 are two similar TPR-motif co-chaperones involved in steroid hormone-dependent diseases with different functions. Therefore, identifying molecules specifically blocking interactions between Hsp90 and FKBP51 or FKBP52 provides a promising therapeutic potential for several human diseases. Here, we describe the protocol for an amplified luminescent proximity homogenous assay to probe interactions between Hsp90 and its partner co-chaperones FKBP51 and FKBP52. First, we have purified the TPR motif-containing proteins FKBP51 and FKBP52 in glutathione S-transferase (GST)-tagged form. Using the glutathione-linked donor beads with GST-fused TPR-motif proteins and the acceptor beads coupled with a 10-mer C-terminal peptide of Hsp90, we have probed protein-protein interactions in a homogeneous environment. We have used this assay to screen small molecules to disrupt Hsp90-FKBP51 or Hsp90-FKBP52 interactions and identified potent and selective Hsp90-FKBP51 interaction inhibitors.
Molecular chaperones contribute to protein homeostasis, including protein folding, transport, and degradation. They regulate several cellular processes and are linked to numerous diseases such as cancer and neurodegenerative diseases1. Heat shock protein 90 (Hsp90) is one of the most important chaperones whose function is dependent on conformational changes driven by ATP hydrolysis and binding with client proteins mediated by its co-chaperones2. Despite an obvious potential of Hsp90 as the therapeutic target, fine-tuning its function represents a big challenge. There are several Hsp90 inhibitors targeting the N-terminal ATP binding region, which have been evaluated in clinical trials, but none of them has been approved for marketing3. Due to the lack of a well-defined ligand-binding pocket4, targeting the C-terminal region of Hsp90 has had limited success4. Recently, disruption of Hsp90-cochaperone interactions by small molecules has been investigated as an alternative strategy5. Targeting the Hsp90-cochaperone interactions would not elicit general cell stress response and provides the possibility to specifically regulate various intracellular processes. The conserved MEEVD pentapeptide at the C-terminus of Hsp90 is responsible for the interaction with the tetratricopeptide repeat (TPR) motif of co-chaperones6. Of the 736 TPR motif-containing proteins annotated in the human protein database, ~20 different proteins interact with Hsp90 via this peptide7. Molecules competing for MEEVD peptide-binding would disrupt the interactions between Hsp90 and co-chaperones containing a TPR domain. The peptide binding site has similar tertiary structure but the overall homology between different TPR motif domains is relatively low7, providing an opportunity to identify molecules specifically capable of blocking interactions between Hsp90 and particular TPR-motif co-chaperones. Among these TPR-motif co-chaperones, FK506-binding protein (FKBP) 51 and FKBP52 are regulators of steroid hormone receptor (SHR) signaling and involved in several steroid hormone-dependent diseases including cancer, stress-related diseases, metabolic diseases, and Alzheimer's disease8. Although FKBP51 and FKBP52 share > 80% sequence similarity, their functions differ: FKBP52 is a positive regulator of SHR activity, while FKBP51 is a negative regulator in most cases8. Therefore, identifying molecules, specifically blocking interactions between Hsp90 and FKBP51 or FKBP52, provides a promising therapeutic potential for related diseases.
Amplified Luminescent Proximity Homogenous Assay (AlphaScreen) was first developed in 1994 by Ullman EF et al.9. Now it is widely used to detect different types of biological interactions, such as peptide10, protein11, DNA12, RNA13, and sugar14. In this technique, there are two kinds of beads (diameter 200 nm), one is the donor bead and the other is the acceptor bead. The biomolecules are immobilized onto these beads; their biological interactions bring donor and acceptor beads into proximity. At 680 nm, a photosensitizer in the donor bead illuminates and converts oxygen to singlet oxygen. Because the singlet oxygen has a short lifetime, it can only diffuse up to 200 nm. If the acceptor bead is in proximity, its thioxene derivative reacts with the singlet oxygen generating chemiluminescence at 370 nm. This energy further activates fluorophores in the same acceptor bead to emit light at 520-620 nm15. If the biological interactions are disrupted, the acceptor bead and donor bead cannot reach proximity, resulting in the singlet oxygen decay and low produced signal.
Here we describe a protocol using this technique for screening small molecules inhibiting interactions between Hsp90 and TPR co-chaperones, especially FKBP51 and FKBP52.The 10 amino acid long peptides corresponding to Hsp90 extreme C-terminus are attached to acceptor beads. Purified GST-tagged TPR co-chaperones interact with glutathione-linked donor beads. When the interaction between Hsp90-derived peptides and TPR-motif co-chaperones brings the beads together, an amplified signal is produced (Figure 1A). If the screened small molecules can inhibit the interactions between Hsp90 and TPR-motif co-chaperones, this amplified signal will be decreased (Figure 1B). Their IC50 can be calculated by quantitative measurement. This protocol can be extended to any chaperone – TPR-motif co-chaperone interactions of interest and is of great importance in the development of novel molecules, specifically blocking the interaction between Hsp90 and FKBP51 or FKBP52.
Figure 1: The basic principle of this assay. (A) Purified GST-FKBP51 interacts with glutathione-linked donor beads. The 10 amino acid long peptides corresponding to the extreme C-terminus of Hsp90 are attached to acceptor beads. The interaction between Hsp90-derived peptides and TPR domain of FKBP51 brings the donor and acceptor beads into proximity. At 680 nm, a photosensitizer in the donor bead illuminates and converts oxygen to singlet oxygen. The thioxene derivative on the acceptor bead reacts with the singlet oxygen and generates chemiluminescence at 370 nm. This energy further activates fluorophores in the same acceptor bead to emit light at 520-620 nm. (B) When small molecules inhibit the interactions between Hsp90 and FKBP51, the donor and acceptor beads cannot reach proximity. Then the singlet oxygen with short lifetime decays, and no detectable signal is produced. Please click here to view a larger version of this figure.
NOTE: An overview of this protocol is shown in Figure 2.
1. Expression and purification of GST-FKBP51 and GST-FKBP52 (Figure 2A)
2. Coupling of Hsp90 C-terminal peptide to the acceptor beads (Figure 2B)
3. The assay probing the interaction between GST-FKBP51 or GST-FKBP52 and Hsp90 C-terminal peptide, and inhibition with small molecular mass compounds (Figure 2C)
4. Data analysis
Figure 2: Schematic of this protocol. (A) Expression and purification of GST-FKBP51 and GST-FKBP52. (B) Coupling of Hsp90 C-terminal peptide to the acceptor beads. (C) The assay probing the interaction between GST-FKBP51 or GST-FKBP52 and Hsp90 C-terminal peptide. Inhibition with small molecular mass compounds. Created with BioRender.com Please click here to view a larger version of this figure.
In our assay, Z' factor and S/B ratio are 0.82 and 13.35, respectively (Figure 3A), demonstrating that our assay is robust and reliable for high-throughput screening. We then used it to screen small molecular mass compounds. Figure 3B presents dose-dependent inhibition of chaperone-cochaperone interactions with a selected small molecule (D10). The dose-response curves for D10 are generated by nonlinear regression analysis, based on which the values of IC50 are calculated. D10 shows dose-dependent inhibition both on Hsp90 – GST-FKBP51 and Hsp90 – GST-FKBP52 PPIs. But the values of IC50 are different: its IC50 for Hsp90 – GST-FKBP51 interactions is 65 nM, whereas, for Hsp90 – GST-FKBP52 interactions, complete inhibition was not achieved with the highest compound concentration (100 µM), its IC50 is estimated to be > 30 µM. These results provide evidence that selective inhibition of Hsp90-HKBP51 or Hsp90-FKBP52 PPIs with small molecules can be achieved (TPR domains of FKBP51 and FKBP52 have 60% sequence identity and > 80% sequence similarity), and this assay can be applied for this screening.
Figure 3: Assay analysis and results. (A) Z' factor and signal-to-background (S/B) ratio of this assay. Data represent signals of (○) positive control (Hsp90 C-terminal peptide, 30 µM) and (●) negative control (DMSO) from 48 wells. A Z' value of 0.82 and an S/B ratio of 13.35 were calculated from these two populations of data. (B) The inhibition of selected compound (D10) on interactions of Hsp90 with FKBP51 (●) or FKBP52 (○) in this assay. D10 inhibits Hsp90-FKBP51 or Hsp90-FKBP52 dose-dependently. Its IC50 is 65 nM for Hsp90-FKBP51 interaction but above 30 µM for Hsp90-FKBP52 interaction. Data are normalized to the control group and expressed as means ± SEM. Please click here to view a larger version of this figure.
Reaction component | Volume (µL) |
PCR buffer (5 x concentrate) | 4 |
Forward primer | 1 |
Reverse primer | 1 |
Plasmid | 0.5 |
dNTP mix (10 mM each) | 0.5 |
Phusion DNA polymerase | 0.5 |
Water (DNA grade) | 12.5 |
Total | 20 |
Table 1: PCR reaction set up for human FKBP51 and FKBP52 DNA amplification.
Stage | Temperature (°C) | Time | Cycles |
Initial denaturation | 94 | 3 min | 1 |
Denaturation | 94 | 30 sec | 35 |
Annealing | 56 | 30 sec | |
Extension | 72 | 1 min | |
Final extension | 72 | 5 min | 1 |
Note: Lid temperature is 105 °C. |
Table 2: Thermocyler conditions for human FKBP51 and FKBP52 DNA amplification.
Here we describe a protocol using the assay for screening small molecules inhibiting interactions between Hsp90 and TPR-motif co-chaperones, especially FKBP51 and FKBP52. Its high Z’ score (>0.8) demonstrates the robustness and reliability for a high-throughput format. Results can be obtained within one hour, and small amounts of beads, protein and compounds are required. Moreover, this protocol could easily be extended to any Hsp90/Hsp70 – TPR-motif co-chaperone interactions of interest. Several TPR-motif co-chaperones of Hsp90 have been implicated in various human disorders ranging from Alzheimer’s disease to autoimmune diseases, cancer, etc. The protocol described here provides an in vitro robust and inexpensive assay for high-throughput screening of small molecules inhibiting chaperone-cochaperone interactions of high medical importance.
In addition, drugs targeting TPR domains and inhibiting interaction with Hsp90 must be assessed not only for their affinity towards the target but also for their selectivity towards other TPR-motif proteins. The human genome encodes > 20 TPR motif proteins capable of interacting with the Hsp90 C-terminal peptide. This assay using glutathione beads and GST-tagged proteins allows assessment of the drug effect on multiple binding TPR partners. Our laboratory possesses a library of 20 human TPR proteins in their GST-tagged form and can obtain affinity and selectivity profiles for every small molecule tested.
It has been previously shown that PPI between Hsp90 and TPR co-chaperones is mediated by the C-terminal peptide of Hsp90; the deletion of Hsp90 C-terminus completely abolishes the binding of TPR-motif co-chaperones6. We have found that compounds efficient in our assay where the Hsp90 C-terminal peptide was used also were able to inhibit the interaction of full-length Hsp90 with GST-FKBP51/52 in a pull-down assay using glutathione sepharose beads (data not shown). Some of the selected compounds also show the binding affinity with FKBP51 by surface plasmon resonance technology, and their specific binding sites determined by co-crystallization are currently under investigation.
A critical step in our protocol is the order of addition; we first form a complex between a small molecule and its TPR target. If complexes between FKBP51 and Hsp90 C-terminal peptide have already been formed, longer incubation times and higher drug compound concentrations are required to break the PPIs. This is mostly due to the steric hindrance of bead limiting access of molecules to the site of interaction.
It is possible to covalently couple TPR proteins and Hsp90 C-terminal peptides to donor and acceptor beads, respectively, and directly probe PPIs. However, it is not recommended to covalently attach both binding partners to the beads due to the reduced movement of molecules in solution affecting the kinetics of the interactions. This also increases the risk of steric hindrance due to the bead size. Therefore, we choose acceptor beads coupled with Hsp90 C-terminal peptide and glutathione donor beads that are interacting with GST-tagged proteins in our assay, where multiple TPR partners can be assessed.
In this assay, it is probable that some identified compounds can be false-positive because of their interference with the assay technology, such as quenching singlet oxygen, quenching light, and scattering light. False-positive results can also be induced by disrupting the binding of GST tag with glutathione donor beads. These false-positive results can be avoided when screening molecules both on Hsp90-FKBP51 and Hsp90-FKBP52 PPIs to identify selective inhibitors. Other methods detecting PPIs should also be applied to verify the selected inhibitors.
The limitation of our assay is that it cannot be used to detect the interaction between Hsp90 and TPR-motif co-chaperones in crude biological samples, such as cell lysates. After the high-throughput screening, the inhibitions of selected compounds on Hsp90 – TPR-motif co-chaperones PPI in biological samples need to be verified by other methods, such as co-immunoprecipitation and proximity ligation assay.
The authors have nothing to disclose.
This study was supported by grants from Swedish Research Council (2018-02843), Brain Foundation (Fo 2019-0140), Foundation for Geriatric Diseases at Karolinska Institutet, Gunvor and Josef Anérs Foundation, Magnus Bergvalls Foundation, Gun and Bertil Stohnes Foundation, Tore Nilssons Foundation for medical research, Margaretha af Ugglas foundation and the Foundation for Old Servants.
384-well plates | Perkin Elmer | 6008350 | Assay volume 25 ml |
Amicon 10.000 MWCO centrifugation unit | Millipore | UFC901008 | Concentrate protein |
Ampicillin | Sigma | A0166 | Antibiotics |
Bacteria shaker Unimax 1010 | Heidolph | Culture bacteria | |
cDNA clones for human FKBP51 | Source BioScience | clone id: 5723416 | pCMV-SPORT6 vector |
cDNA clones for human FKBP52 | Source BioScience | clone id: 7474554 | pCMV-SPORT6 vector |
Chemically Competent E. coli | Invitrogen | C602003 | One Shot BL21 Star (DE3) |
Data analysis software | GraphPad Prism | 9.0.0 | Analysis data and make figures |
Data analysis software | Excel | Analysis data | |
DMSO | Supelco | 1.02952.1000 | Dilute compounds |
DPBS | Gibco | 14190-144 | Prepare solution |
EDTA | Calbiochem | 344504 | Prevent proteolysis during sonication |
Glutathione | Sigma | G-4251 | Elute GST-tagged proteins |
Glutathione donor beads | Perkin Elmer | 6765300 | Donor bead |
GST-trap column | Cytiva (GE Healthcare) | 17528201 | Purify GST-tagged proteins |
Isopropyl-β-D-thiogalactoside | Thermo Fisher Scientific | R0392 | Induce protein expression |
LB Broth (Miller) | Sigma | L3522 | Microbial growth medium |
PCR instrument | BIO-RAD | S1000 Thermal Cycler | Amplification/PCR |
PD-10 column | Cytiva (GE Healthcare) | 17085101 | Solution exchange |
pGEX-6P-1 vector | Cytiva (GE Healthcare) | 28954648 | Plasmid |
pGEX-6P-2 vector | Cytiva (GE Healthcare) | 28954650 | Plasmid |
Plate reader | Perkin Elmer | EnSpire 2300 Multilabel Reader | Read alpha plate |
Plate reader software | Perkin Elmer | EnSpire Manager | Plate reader software |
Plate reader software protocol | Perkin Elmer | Alpha 384-well Low volume | Use this protocol to read plate |
PMSF | Sigma | P7626 | Prevent proteolysis during sonication |
protease inhibitor cocktail | Sigma | S8830 | Prevent proteolysis during sonication |
Sodium azide | Sigma | S2002 | As a preservative |
Sodium cyanoborohydride (NaBH3CN) | Sigma | 156159 | Activates matrix for coupling |
Ten amino acid peptide NH2-EDASRMEEVD-COOH corresponding to amino acids 714-724 of human Hsp90 beta isoform | Peptide 2.0 inc | Synthesize Hsp90 C-terminal peptide | |
Test-Tube Rotator | LABINCO | Make end-over-end agitation | |
Tris-HCl | Sigma | 10708976001 | Block unreacted sites of acceptor beads |
Tween-20 | Sigma | P1379 | Prevent beads aggregation |
Ultra centrifuge Avanti J-20 XP | Beckman Coulter | Centrifuge to get bacteria cell pellets | |
Ultrasonic cell disruptor | Microson | Sonicate cells to release protein | |
Unconjugated acceptor beads | Perkin Elmer | 6762003 | Acceptor beads |
XCell SureLock Mini-Cell and XCell II Blot Module | Invitrogen | EI0002 | SDS-PAGE |