Multiplexed high-throughput screen for small molecule regulators of Bcl-2 family protein interactions, specifically Bim-Bcl-2
One arm of apoptosis is regulated by the balance of anti-apoptotic and pro-apoptotic Bcl-2 family members. In humans, six genes have been identified that encode anti-apoptotic proteins characterized by the presence of conserved motifs designated as three Bcl-2 homology (BH) regions, BH1, BH2, and BH3 [Reed, et al. 2004]. These domains form a hydrophobic cleft in tertiary structure. Pro-apoptotic more ..
BioActive Compounds: 9
University of New Mexico Assay Overview:
Assay Support: NIH 1X01 MH079850-01
HTS to identify small molecule regulators of Bcl-2 family protein interactions
PI: Larry Sklar, Ph.D.
Assay Implementatiion: Peter Simons Ph.D, Susan Young MS, Anna Waller Ph.D, Mark Carter MS
Dose Response Assay Background and Significance:
One arm of apoptosis is regulated by the balance of anti-apoptotic and pro-apoptotic Bcl-2 family members. In humans, six genes have been identified that encode anti-apoptotic proteins characterized by the presence of conserved motifs designated as three Bcl-2 homology (BH) regions, BH1, BH2, and BH3 [Reed, et al. 2004]. These domains form a hydrophobic cleft in tertiary structure. Pro-apoptotic family members contain BH3 regions that form an amphipathic helix, and these helices bind in the clefts of the anti-apoptotic proteins. Overexpression of pro-apoptotic BH3 peptides has been shown to increase apoptosis of leukemia cells using in vitro and animal model studies [Holinger, et al. 1999; Wang et al. 2000; Walensky et al. 2004]. The binding of fluorochrome-conjugated BH3 peptides (including Bim) to Bcl-2 family members thus provides the basis for construction of fluorescence-based assays amenable to flow cytometry high throughput screening for small molecule regulators of these interactions. This is a multiplexed assay to identify small molecule regulators of protein interactions between the BH3 peptide of Bim and the following six Bcl-2 family members: Bcl-XL, Bcl-W, Bcl-B, Bfl-1, and Mcl-1 and Bcl-2 (the eponymous founding member of the Bcl-2 family [Tsujimoto, et al. 1986]).
Each component of the multiplex assay consists of a glutathione labeled bead, a GST Bcl-fusion protein target (six total, supplied by project collaborator), and a fluorescent peptide probe, F-Bim (FITC-Ahx-DMRPEIWIAQELRRIGDEFNAYYAR-OH; Commonwealth Biotech, USA). Bead sets are coated with individual GST-conjugated Bcl-2 proteins in HPSMTB buffer (30mM HEPES, 100mM KCl, 20mM NaCl, 1mM MgCl(2), 0.01% Tween-20, 0.1% BSA) and incubated overnight at 4 degrees C.
The multiplex is constructed by using beads for each protein target that have been labeled with varying intensities of red color, so that each assay is built on a unique bead set, and each bead set is associated with a unique optical address. Beads are first washed in HPSMTB buffer for 20 minutes before adding the appropriate GST-Bcl fusion protein. The bead sets (ThermoFisher Scientific product numbers XPR-1687-XPR-1696), have similar size (~ 4 micron diameter) and are distinguished by distinct emission characteristics at 665 +/-10 nm with excitation at 635 nm. Thus, GST-Bfl-1 might be non-covalently coated onto red level 1 beads, GST-Bcl-XL onto red level 2 beads, etc. The 6 bead sets (each with bound protein) and uncoated beads (see below) are first centrifuged separately, then combined and centrifuged again, and finally diluted just before loading into 384-well plates, to minimize bead-protein dissociation before the assay begins.
The HTS assay was conducted in 384-well microplates in a total assay volume per well of 10.1 microliters (5 microliters of bead mixture, 0.1 microliters of test compound, and 5 microliters of 100 nM F-Bim in HPSMTB). Test compound concentration was 10 microM. Controls, which contained bead mixture and F-Bim but no test compound, were located in columns 1 and 2 on each plate. Plates were placed horizontal axis on rotators and incubated for 1-2 hours at 4 degrees C.
A glutathione-only bead set control (no associated GST-protein) was incorporated into each well as a fluorescence scavenger to determine inherent fluorescent properties (at 530 nm emission) of the test compounds. Specificity of F-Bim binding was determined with a Positive Control using a block of the F-Bim fluor with a non-fluoresceinated Bim peptide. The F-Bim blocking control was run daily as a separate single tube assay using Bim at 5 microM.
The primary screen was performed with 194,920 compounds at a compound concentration of 10 microM to identify 116 small molecule regulators of Bcl-2 family protein interactions, specifically Bim-Bcl-2 (AID 950). In the study reported here, the 116 compounds that satisfied the hit selection criterion in the primary screen (change in %Inhibition greater than 40%) were tested in a dose response format to confirm activity and determine potency. Additional compounds, which were actives from other Bcl targets, were also included in the dose response evaluations. Test compounds at 10 mM concentration in DMSO were serially diluted 1:3.16 eight times for a total of nine different test compound concentrations. Final compound dilutions in DMSO ranged from 1 microM to 10 mM. These dilutions were then diluted 1 to 100 to give an assay concentration range of 10 nanoM to 100 microM.
Sample acquisition and preliminary analysis is conducted with the HyperCyt(R) high throughput flow cytometry platform. The HyperCyt system interfaces a flow cytometer and autosampler for high-throughput microliter-volume sampling from 384-well microtiter plates [Kuckuck, et al. 2001]. The stream of particles is excited at 488 nm and 635 nM, and flow cytometric data of light scatter and fluorescence emission at 530 +/- 20 nm (FL1) and emission at 665 +/- 10 nm (FL8) are collected on a Cyan Flow Cytometer (Dako). Analysis of the time-resolved acquisition data file uses IDLeQuery software to merge the flow cytometry data files with compound worklist files generated by HyperSip software. The raw data are parsed in IDLeQuery to produce annotated fluorescence summary data for each well. The parsed data are then processed through an Excel template file constructed specifically for the assay to segregate data for each target and the fluorescence scavenger in the multiplex. Gating based on forward scatter (FS) and side scatter (SS) parameters is used to identify singlet bead populations. Gating based on FL8 emission distinguishes the beads coated with different proteins, and the green median fluorescence intensity (MFI) per bead population (well) is calculated.
In dose response experiments, the assay was performed without compound and with nine different concentrations of compound, from 10 nanoM to 100 microM, to produce a series of 9 data points. IDLeQuery calculates the median channel fluorescence (MCF) for each of these ligand concentrations, generating competition curves.
Ligand competition curves were fitted by Prism# software (GraphPad Software, Inc., San Diego, CA) using nonlinear least-squares regression in a sigmoidal dose response model with variable slope, also known as the four parameter logistic equation. Curve fit statistics were used to determine the concentration of added test compound competitor that inhibited fluorescent ligand binding by 50 percent (EC50, microM), the low and high boundaries of the 95% confidence interval of the EC50 estimate, the Hill Slope, and the correlation coefficient (r squared) indicative of goodness-of-fit.
Fluorescence of the compounds could be green, which would increase the apparent amount of F-Bim bound (signal). None of such artifacts occurred with the large dynamic range of this assay, and no effort was made to annotate them. Fluorescence of the compounds could be red, which would shift the bead sets out of their gates, which had been set using whole plate data. Again, none of such artifacts occurred (bead numbers decreased at the upper end of a few dose response experiments), and no effort was made to annotate them.
Compounds could also interfere with the binding of the GST fusion proteins to the GSH beads (potential false positives). Dose response experiments were performed with compound concentrations as above, but with GSH beads incubated overnight with 50 nM GST-green fluorescent protein (similar to GST-Bcl proteins), then diluted as usual in the dose response series, NOT including any F-Bim. These counterscreen data are reported in PubChem AID "Profiling Assay to determine GST-GSH interactions in multiplex bead-based assays (HPSMTB buffer)". Compounds deemed "False Positive" are noted such in PUBCHEM_ASSAYDATA_COMMENT and these compound were given PUBCHEM_ACTIVITY_SCORE of 0.
Compounds with EC50 less than 10 microM and magnitude of response greater than 40% (i.e., Bottom of sigmoidal curve < 0.6 * Top of sigmoidal curve, listed as FIT_PERCENT_SPAN) were said to be "Active", and thus given a PUBCHEM_ACTIVITY_SCORE. The PUBCHEM_ACTIVITY_SCORE was based on the following equation;
PUBCHEM_ACTIVITY_SCORE = 50 * ( 1 - EC50/10 microM) + 50 * (Response - 40)/40
In this assay active compounds have activity score range 9 - 100, and for inactive compounds the active score is 0.
Keywords: NIH Roadmap, NMMLSC, high throughput flow cytometry, Bcl, Bim, Bfl-1, Bcl-XL, Bcl-2, Bcl-W, Bcl-B, Mcl-1, multiplex, bead-based, screening, dose response
* Activity Concentration. ** Test Concentration.
Data Table (Concise)