Fig 1.
TR-FRET screen for RIN1::ABL1 interaction inhibitors.
(A) Binding between RIN1 and ABL is initiated by a proline rich motif in RIN1 binding to the ABL-SH3 domain. ABL phosphorylates RIN1-Y36, which then binds the ABL-SH2 domain. For the assay, ABL was fused to eGFP and RIN1 was fused to a streptavidin binding peptide (SBP) that connects to a streptavidin-terbium complex. A 340 nm pulse excites terbium, which can transfer energy to excite the GFP acceptor if the fluorophores are in close proximity, reflecting RIN1::ABL binding. Predicted FRET inhibitor classes: 1. Orthosteric inhibitors, 2. Direct ABL kinase inhibitors and 3. Allosteric inhibitors. (B) RIN1::ABL binding was quantified as a FRET ratio: GFP emission at 520 nm to terbium emission at 486 nm. The negative control was donor and acceptor fluorophores only (no RIN1) and was normalized to 1.
Fig 2.
Validation of TR-FRET screening assay for detection of RIN1::ABL binding.
Binding was quantified as a ratio of GFP emission at 520 nm to terbium emission at 486 nm. The negative control was normalized to 1. Experiments were performed in quadruplicate. (A) ATP was omitted from the buffer mix to prevent RIN1 phosphorylation by ABL. *p = 3.0x10-6 (B) 10 μM or 100 μM imatinib was added to inhibit ABL kinase activity. *p = 0.001 (C) Added untagged ABL competed with ABL-eGFP (1:1 and 10:1 ratios were used). *p = 1.4x10-7
Table 1.
High throughput screen results.
Fig 3.
Representative data from primary HTS.
Example of 384-well plate data presented as a scatterplot. Each plate was screened with positive (purple) and negative (blue) controls consisting of streptavidin terbium and ABL-eGFP with and without RIN1-SBP, respectively. Screening compounds are represented in gray. The hit cutoff was a plate-based 3-standard deviation decrease in the FRET ratio.
Fig 4.
Elimination of off-target inhibitors.
(A) Two types of off-target inhibitors were predicted (red T-bars): compounds that quench GFP fluorescence, and biotin-like compounds that bind and sequester streptavidin. (B) Hit compounds were tested in triplicate dose-response curves with ABL-eGFP. Concentration response curve data from a representative GFP quencher (methyl 3,4,6-trihydroxy-5-oxo-5H-benzo[7]annulene-8-carboxylate) is shown. (C) Biotin mimics were identified by in silico screening for structural similarity to biotin. (D) Hit compounds were re-tested in the TR-FRET assay in triplicate 10-point dose-response curves to prioritize those with IC50<10 μM. Data from the representative compound CID 24512426 is shown.
Fig 5.
Theaflavin-3,3’-digallate is a direct ABL kinase inhibitor.
0.5 nM purified ABL-eGFP was mixed with 2 μM CRK and 10 μM test compound and incubated in kinase buffer for 30 minutes at 30°C. Samples were analyzed by immunoblot using anti-phosphotyrosine. All the hits were screened, and a representative immunoblot is shown here. Test compounds were as follows: DMSO (ctrl), imatinib (IM), CID 16312764 (1), CID 467320 (2), CID 7457610 (3), CID 6084 (4), CID 4844281 (5), IUPAC 1-(4-phenylphenyl)-2-[4-(pyrimidin-2-yl)piperazin-1-yl]ethan-1-ol (6), CID 9291491 (7), CID 3991439 (8) and CID 16350799 (9).
Fig 6.
MAPK1/3 phosphorylation is RIN1 and BCR-ABL1-dependent in K562 cells.
(A) K562 cells expressing a control vector, RIN1 shRNA or RIN1 over-expression construct were analyzed by immunoblot with anti-pMAPK1/3 and anti-MAPK1/3 antibodies (top) and anti-RIN1 and anti-tubulin antibodies (bottom). Band intensity was quantified using LI-COR Odyssey software. Phospho-MAPK1/3 signal intensities were normalized to total MAPK1/3 and RIN1 normalized to tubulin. Changes in expression or phosphorylation were calculated as fold-change compared to control vector (normalized to 1) and presented below each blot. The immunoblot shown is representative of three independent experiments. (B) K562 cells were treated with 1 μM imatinib (IM) for 1 and 4 hours at 37°C, then analyzed by immunoblot as in A. (C) K562 cells expressing control or RIN1 shRNA were analyzed by immunoblot as in A. The ratio of pMAPK/MAPK signal intensity is graphed. Results were averaged over five independent experiments. *p-value = 1x10-4
Fig 7.
Five compounds significantly decrease MAPK1 phosphorylation in K562 cells.
DMSO or compound was added to K562 cells at a final concentration of 1% or 10 μM, respectively, and incubated for 4 hours at 37°C before being lysed and analyzed by immunoblot with anti-pMAPK1/3 and anti-MAPK1/3. Each compound was tested with six biological replicates. Band intensity was quantified using LI-COR Odyssey, and the ratio of pMAPK/MAPK signal intensity is graphed for each compound and its DMSO control. Compounds are identified by their PubChem CID number.