Development of an HTS-Compatible Assay for Discovery of Melanoma-Related Microphthalmia Transcription Factor Disruptors Using AlphaScreen Technology

Proteins

Human MITF (residues 199–288) was constructed with an N-terminal 6xHis-tag with a tobacco etch virus (TEV) protease cleavage site (denoted as tev) in a pHisTEV vector and maltose binding protein (MBP) was fused before tev to increase the stability of MITF. 6xHis-MBP-tev-MITF (WT, R259stop, and Del217R, denoted as His-MITF, His-MITF_R259stop, and His-MITF_Del217R separately) was expressed in the bacterial strain BL21 (DE3) at 16 °C for 20 h and purified to homogeneity with a Ni-HiTrap affinity column. Further purification was performed through a size exclusion S200 column (GE Healthcare, Piscataway, NJ). Biotin-tagged MITF (denoted as Biotin-MITF) was obtained by TEV cleaving 6xHis-MBP-tev-_CGGMITF that contains a cysteine at the N-terminal of MITF. The purification of 6xHis-MBP-tev-_CGGMITF was done using the same method as purifying 6xHis-MBP-tev-MITF mentioned above. Following protease cleavage by TEV overnight, the nontagged _CGGMITF was further purified through a HiTrap SP Sepharose column and size exclusion S75 column (GE Healthcare). EZ-Link Iodoacetyl-PEG2-Biotin (Thermo Scientific, Rockford, IL) was used to label the _CGGMITF following the protocol indicated.

MITF Dimerization-Based AlphaScreen Assay Development and Characterization

The AlphaScreen histidine (nickel chelate) detection kit containing Ni-chelate acceptor beads and streptavidin donor beads was purchased from PerkinElmer (Waltham, MA).

All experiments were set up in 20 μL reaction volume in Optiplate-384HS plates (PerkinElmer) under green light conditions at room temperature. The reaction buffer contains 50 mM HEPES (pH 7.4), 250 mM NaCl, 0.1% Tween, and 0.1% bovine serum albumin (BSA). The AlphaScreen luminescence was read on a PerkinElmer EnVision 2104.

To optimize the suitable concentration of His-MITF and Biotin-MITF, cross-titration between these two proteins was performed, respectively (His-MITF [80 nM]—gradient Biotin-MITF [0–160 nM] titration; Biotin-MITF [80 nM]—gradient His-MITF [0–320 nM] titration). For both titrations, 5 μL of His-MITF was first incubated with AlphaScreen acceptor beads (chelated with nickel, final concentration 10 μg/mL) for 15 min. Five microliters of Biotin-MITF and 5 μL of AlphaScreen donor beads (conjugated with streptavidin, final concentration 10 μg/mL) were then added to the well. After adequate shaking by the MixMate (Eppendorf) with 1500 rpm for 2 min, the plates were incubated in the dark for 2 h before reading on a PerkinElmer EnVision 2104 using the AlphaScreen detection technology setup (680-nm excitation and 520-620 nm emission).

For the nontagged MITF titration, the gradient nontagged MITF/BSA (0–1280 nM) was incubated with His-MITF (80 nM) and AlphaScreen acceptor beads (final concentration 10 μg/mL) overnight before adding Biotin-MITF (80 nM) and AlphaScreen donor beads (final concentration 10 μg/mL).

For His-MITF WT/R259stop/Del217R—Biotin-MITF assay, a gradient concentration of His-MITF WT/R259stop/Del217R (0–250 nM) was implemented and incubated with AlphaScreen acceptor beads (final concentration 10 μg/mL) overnight before adding Biotin-MITF (80 nM) and AlphaScreen donor beads (final concentration 10 μg/mL).

Protocol for 384-Well Plate Assay

Pre-HTS experiments were set up in triplicate. First, His-MITF protein (final concentration 80 nM) and AlphaScreen acceptor beads (final concentration 10 μg/mL) were dispensed into the plate using the Aurora Discovery FRD (Beckman Coulter, Indianapolis, IN). One hundred nanoliters of test compounds (final concentration of ~15 μM, 0.5 % DMSO) were pinned into the plates using the Beckman Biomek NXP (Beckman Coulter), followed by incubation at 4 °C overnight. Then, Biotin-MITF (final concentration of 80 nM) and the AlphaScreen donor beads (final concentration of 10 μg/mL) were dispensed into the plate using the Aurora Discovery FRD. Plates were incubated for 2 h before reading on a PerkinElmer EnVision 2104 using the HTS AlphaScreen modality reading at a height of 0 mm, exciting for 0.04 s with an emission time of 0.09 s per well with cross talk optimization corrections applied.

Compound Library

The compounds tested in the assay validation stage were from the Sigma LOPAC1280 library containing 1280 proven pharmacologically active compounds. All data were normalized on a per-plate basis to their individual plate controls (N = 12 per plate). The percent activation for each compound was calculated as shown in the equation below, where Test Compound is defined as wells containing His-MITF, Ni-chelate AlphaScreen acceptor beads, streptavidin-conjugated AlphaScreen donor beads, and Biotin-MITF in the presence of the test compound. Positive Control is defined as wells containing the same formulation, but with the exception that nontagged MITF is used at 1 μM in place of the compounds. Low Control wells again are the same, but contain 0.5% DMSO.

% Activation = 100 × [1 – (Test Compound – Median Positive Control)/(Median Low Control – Median Positive Control)]

Since each compound was represented three times, we also calculated its average percent activity. The AlphaScreen compound concentration–response curve (CRC) assay was run under identical conditions to the HTS version in triplicate using a 10 mM source starting concentration and threefold serial dilutions for 10 points. The final assay compound concentrations were 15 μM at the highest dilution, progressing downward accordingly.

Assay Performance

To monitor assay sensitivity and evaluate the robustness of this assay, signal-to-background (S/B) ratios and the Z′ factor were calculated using the equations previously described. ³²

Data Analyses

All control EC₅₀ values were calculated by nonlinear regression analysis (sigmoidal dose–response with variable slope) using Prism software (GraphPad Software, Inc., San Diego, CA). Data analysis for the LOPAC1280 compound library was performed using internal informatics tools and Symyx Assay Explorer version 3.2 (Symyx Technologies, Inc., San Diego, CA). For CRC analysis involving the test compounds, we imported the data into the Scripps in-house database that allows ease of data handling for large numbers of compounds. Using database tools, percent activation was plotted against compound concentration. A four-parameter equation describing a sigmoidal dose–response curve was then fitted with an adjustable baseline using Assay Explorer software (Symyx Technologies). The reported EC₅₀ values were generated from fitted curves by solving for the X-intercept value at the 50% inhibition level of the Y-intercept value. In cases where the highest concentration tested (i.e., 15 μM) did not result in greater than 50% inhibition, the EC₅₀ was determined manually and labeled as greater than 15 μM.

Real-Time Kinetic Binding Assay

The interaction between 5′-biotinylated M-box DNA (biotin-CAGTCATGTGCT [sense]; AGCACATGACTG [antisense]) and HisMBP-MITF (6xHis-MBP-tev-MITF_WT) was tested using the ForteBio Octet Red system (ForteBio, Inc., Menlo Park, CA). The 5′-biotinylated M-box DNA was immobilized onto streptavidin biosensors (ForteBio), and the association (0–180 s) and dissociation (180–420 s) of the DNA molecule to HisMBP-MITF was monitored in binding buffer containing 1 × phosphate-buffered saline (PBS), 0.01% BSA, 0.002% Tween, and 1 mM MgCl₂. Hits from the pilot screen LOPAC (P2738, M7277, and D5564) were incubated with HisMBP-MITF for 8 h with a final concentration of 50 μM before the measurement.

General Strategy: Developing the MIDAS Assay to Monitor the Dimerization of MITF

To monitor the dimerization of MITF, an AlphaScreen assay was designed, which includes a His-MITF that is recognized by the AlphaScreen acceptor bead and Biotin-MITF that is recognized by the AlphaScreen donor beads ( Fig. 1A ). To determine the optimal reagent concentrations, an initial assay was carried out varying His-MITF and Biotin-MITF simultaneously from 0 to 320 nM using 10 μg/mL AlphaScreen acceptor beads and 10 μg/mL AlphaScreen donor beads ( Fig. 1B ). The optimal concentrations achieved were 80 nM Biotin-MITF and 80 nM His-MITF, at which the S/B ratio reaches 343 (the background signal defined as the signal generated from 10 µg/mL AlphaScreen acceptor beads and donor beads only).

OPEN IN VIEWER

Figure 1.

Design of MIDAS assay. (A) MIDAS assay for identifying MITF dimer disruptors. His-MITF and Biotin-MITF molecules are recognized by the AlphaScreen acceptor bead and donor bead, respectively. Dimerization of MITF leads to the association of the acceptor bead and donor bead in close distance and trigger fluorescence emission from the acceptor beads at 520–620 nm. An MITF disruptor compound would disrupt this signal. (B) Titration of increasing concentrations of His-MITF and Biotin-MITF simultaneously from 0 to 320 nM under 10 μg/mL AlphaScreen acceptor beads and 10 μg/mL AlphaScreen donor beads. The platform was achieved with 80 nM Biotin-MITF and 80 nM His-MITF with an S/B ratio of 343. (C) Identification of optimal His-MITF concentration. Increasing concentrations of His-MITF (0–320 nM) were incubated with a constant amount of 80 nM of Biotin-MITF. The platform was achieved with 80 nM His-MITF with an extreme S/B and Z′. (D) Identification of optimal Biotin-MITF concentration. Increasing concentrations of Biotin-MITF (0–160 nM) were incubated with a constant amount of 80 nM of His-MITF. The platform was achieved with 80 nM Biotin-MITF with extreme an S/B, and Z′ for the indicated concentrations. Data are represented as the mean ± SD of three independent experiments. The EC₅₀ values were calculated by nonlinear regression analysis using GraphPad Prism.

We then further optimized the concentration of Biotin-MITF and His-MITF through cross-titration of these two proteins respectively. First, 80 nM Biotin-MITF was added in the presence of titrated His-MITF (0–320 nM), and the plateau was reached with 80 nM His-MITF ( Fig. 1C ). The EC₅₀ value for the His-MITF was determined to be 25 nM, and the S/B and Z′ factor were 918 and 0.98, respectively ( Fig. 1C ). Then, 80 nM His-MITF was added in the presence of titrated Biotin-MITF (0–160 nM), the EC₅₀ value for the Biotin-MITF was determined as 19 nM, and the S/B and Z′ factor were 377 and 0.96, respectively ( Fig. 1D ). Thus, under the conditions of 80 nM His-MITF, 80 nM Biotin-MITF, 10 μg/mL AlphaScreen acceptor beads, and 10 μg/mL AlphaScreen donor beads, the MITF dimerization-based AlphaScreen (MIDAS) assay is robust and sensitive.

Specificity of the MIDAS Assay to Mirror the State of MITF

To check the specificity of this MIDAS assay, a competition assay was performed with nontagged MITF. Varying concentrations of nontagged MITF (which contains neither His tag nor Biotin tag) was added into the assay system, under the optimal conditions described above. A gradient of BSA was used as control protein. Nontagged MITF specifically blocked the AlphaScreen signal in a dose-dependent manner with an IC₅₀ of 26 nM, while BSA had no effect on the AlphaScreen signal ( Fig. 2A ), indicating the observed AlphaScreen signals are generated specifically by the dimer interaction between MITF monomers.

OPEN IN VIEWER

Figure 2.

The MIDAS assay reflects the biological function of MITF dimerization. (A) Nontagged MITF competition assay. Gradient increased nontagged MITF (contains neither His tag nor Biotin tag, gradient-increased BSA as control) was added into the MIDAS assay, with 80 nM His-MITF, 80 nM Biotin-MITF, 10 μg/mL AlphaScreen acceptor beads, and 10 μg/mL AlphaScreen donor beads. The IC₅₀ of nontagged MITF to inhibit MITF dimerization-mediated AlphaScreen signal was around 26 nM, while BSA had no effect. (B) Titration of MITF loss-of-function mutations in MIDAS assay. Two MITF loss-of-function mutations (His-MITF_R259stop and His-MITF_Del217R) were tested with gradient concentration (from 0 to 250 nM) by incubating with Biotin-MITF (WT, 80 nM).

To further check the physiological relevance of this MIDAS assay to MITF function, two MITF loss-of-function mutations, R259stop and Del217R, were tested. Both mutations, found in human Waardenburg syndrome disease, led to the loss of MITF-DNA binding and loss of transcriptional activity of MITF. MITF_R259stop forces the loss of the leucine zipper and thus cannot form dimers;^28,31 MITF_Del217R contains a mutation on the basic region and loses the DNA binding ability but maintains the dimer forming ability.^24,27 Increasing concentrations of His-MITF (WT/Del217R/R259stop, from 0 to 250 nM) were titrated against the assay under optimal conditions. His-MITF_R259stop, the mutant that disrupted MITF dimerization, generated no AlphaScreen signal. In contrast, His-MITF_Del217R, which has no effect on the MITF dimer interface, generated AlphaScreen signal as the His-MITF_WT ( Fig. 2B ). These results demonstrate that the MIDAS assay mirrors the biological feature of MITF dimerization.

LOPAC Screen

To assess the performance of the MIDAS assay for HTS, a pilot screen against the Sigma LOPAC1280 library was performed under semiautomated HTS conditions. A 20 μL reaction under optimized conditions was used for the LOPAC screen in 384 plates. Each plate included low control wells with signal that was generated by adding DMSO (0.5%) in the system and positive control wells with signal that was generated by adding nontagged MITF with a final concentration of 1 μM. These control signals were utilized to calculate the Z′ factor and S/B, and assess plate homogeneity and reproducibility. LOPAC compounds were tested in triplicate at a final concentration of 15 μM (0.5% DMSO final). A total of 14 plates were used for this assay, comprising four separate triplicate compounds plates and two DMSO plates, one at the beginning and one at the end of the run. As indicated from the positive and low control scatterplots, the entire LOPAC MIDAS screen demonstrated an excellent Z′ factor (0.81 ± 0.29) and a stable S/B ratio of 3.92 ± 0.16 ( Fig. 3A ; here the background signal was defined by low control that used 1 μM nontagged MITF to block the dimerization-based AlphaScreen signal). Further, a scatterplot of replicate measurements yielded r² = 0.971, indicating high reproducibility in assay data ( Fig. 3B ).

OPEN IN VIEWER

Figure 3.

LOPAC screening of the MIDAS assay. (A) Scatterplot analysis for the results from the high-throughput LOPAC screening. All data from 12 assay plates, including controls, are displayed. The LOPAC screening yielded a Z′ factor of 0.81 ± 0.29 and an S/B ratio of 3.92 ± 0.16. (B) Scatterplot of triplicate measurements yielded r² = 0.971, indicating high reproducibility of the assay. Representative graph of compound activity correlating plate replicas. The best-fit line has r² = 0.971, indicative of the high fidelity of the MIDAS assay’s hit identification.

Confirmation of Lead Hits in LOPAC Screen

An important parameter to assess for HTS readiness is the susceptibility of an assay to artifact, that is, compounds that interfere with the assay reagents or colored compounds that optically quench the AlphaScreen signal itself, resulting in a “false-positive” response. To exclude those false-positive hits, a counterscreen was carefully designed. This assay protocol uses a specific interaction between human LysRS and MSCp38, an interaction pair that we have previously characterized and is biologically independent of MITF. ³³ This LysRS-MSCp38 AlphaScreen assay is similarly formatted as the MIDAS assay. A His-tagged full-length LysRS (hereafter denoted as His-LysRS) with Ni-chelated AlphaScreen acceptor beads was first incubated with the test compounds, followed by the sequential addition of the biotin-labeled MSC p38 and the streptavidin-conjugated AlphaScreen donor beads ( Fig. 4A ). This counterassay is also highly specific because nontag p38 specifically suppresses the signal in a dose-dependent manner ( Fig. 4B ). Compound that causes a decrease in the MIDAS signal and the LysRS-MSCp38 AlphaScreen signal would represent a nonspecific, false-positive hit, whereas those only showing an effect on the MIDAS signal are potential true hits.

OPEN IN VIEWER

Figure 4.

LysRS-MSCp38 interaction-based counterscreen. (A) Titration of LysRS-MSCp38 interaction using AlphaScreen. Similar to the MIDAS assay, His-tagged LysRS (His-LysRS) and biotin-tagged MSCp38 (Biotin-MSCp38) are recognized by AlphaScreen acceptor bead and donor bead, respectively. Concentrations of His-LysRS and Biotin-MSCp38 were simultaneously increased from 0 to 50 nM with 10 μg/mL AlphaScreen acceptor beads and 10 μg/mL AlphaScreen donor beads. (B) Nontagged MSCp38 competition assay. Gradient-increased concentrations of nontagged p38 (which contain neither His tag nor Biotin tag) from 0 to 30 nM suppressed the AlphaScreen signal in a dose-dependent manner.

Following this criterion, three hit compounds were discovered, including P2738, D5564, and D7277 ( Fig. 5A ). These compounds significantly disrupted the MIDAS signal to the basal level with no effect on the LysRS-MSCp38 ( Fig. 5A ). Titration of these compounds indicated that they disrupted the MITF dimerization in a dose-dependent manner with an EC₅₀ around 1 μM ( Fig. 5A , B ). We further measure the ability of these compounds to disrupt MITF-DNA binding through real-time binding assay. Compounds D5564 and M7277 decreased the binding ability of MITF to sensor-conjugated M-box DNA, while compound P2738 totally disrupted the DNA binding ability ( Fig. 5C ), confirming the true efficacy of these compounds in inhibiting MITF activity.

OPEN IN VIEWER

Figure 5.

Confirmation of initial hits from LOPAC screening. (A) Data of the hit compounds from LOPAC screening. (B) MIDAS titration curves of hit compounds. P2738, D7277, and D5564 were titrated starting from 15 μM, with threefold serial dilution steps for 10 points, and disrupting the AlphaScreen signal in a dose-dependent manner at EC₅₀ values of 1, 1.2, and 0.72 μM, respectively. Each point on a curve represents the average of three results with error bars. (C) Inhibition of MITF-DNA interaction by hit compounds. All three compounds decreased the binding of MITF to its target DNA.