Research ArticleMOLECULAR BIOLOGY

Rational discovery of antimetastatic agents targeting the intrinsically disordered region of MBD2

See allHide authors and affiliations

Science Advances  20 Nov 2019:
Vol. 5, no. 11, eaav9810
DOI: 10.1126/sciadv.aav9810
  • Fig. 1 In silico discovery of the MBD2 IDPR-targeting ligands.

    (A) Flow chart describing the computational process of ligand discovery. (B) Evaluation of the intrinsic disorder propensity of MBD2 (left) and c-Myc (right); disorder scores 1 and 0 mean fully disordered and fully ordered residues, respectively. Pink bars show positions of the determined DOT sites embedded in residues 360 to 393 for MBD2 and 395 to 430 for c-Myc. (C) Chemical structures of the top 10 compounds showing the most favorable binding to the MBD2 target site in the molecular docking screening of ZINC chemical library. (D) Representative structures of protein-ligand complexes obtained from the molecular docking results (original data file 1 for PDB coordinates): 10058-F4:c-Myc402 (top; control experiment), ABA:MBD2369 (middle), and APC:MBD2369 (bottom).

  • Fig. 2 Lead selection from hit compounds.

    (A) Computational analysis for off-target probabilities of the 10058-F4 (control experiment) and two selected lead compounds (ABA and APC). Max Tc and E value of the predicted binding are plotted for the n (number of potential targets predicted) off-target candidates yielded from SEA using 2060 human proteins in the database. See fig. S2 for the other hit compounds. (B) Cell migration inhibition by the hit compounds. The LM1 and HCT116 cancer cells were fixed and stained after 48 hours of Transwell migration in the presence of indicated concentrations of individual compounds, followed by counting the number of migrated cells (n = 2) to yield MI50 value.

  • Fig. 3 In silico analysis of the lead compound binding to target site.

    (A) Time-course alterations of the number of intermolecular contacts within 3 Å cutoff in MD simulations. (B) Heatmap describing the number of simulated compound-protein contacts during 50-ns trajectory for individual residues. Each value of a number of contacts was normalized by dividing it by the total number of contacts in each simulation. The already-known critical residues for PPI are shown in darker red. (C) Heatmap of the intermolecular heavy atom contacts between the lead compounds and target proteins during 50-ns trajectory. Number of contacts was normalized by the total number of contacts in each simulation. MBD2 N-terminal two residues, G and S, were from the NMR structure (PDB ID: 2L2L). MBD2 sequence starts from K360, after G, and S.

  • Fig. 4 Two lead compounds efficiently disrupt MBD2 interaction with p66α in vitro and in cell.

    (A) Inhibition of in vitro FRET dynamics of MBD2 interaction with p66α by ABA and APC. Relative mean FRET values for the corresponding ratios of chemical concentration over MBD2::p66α1–206 were plotted. See fig. S4A for the original data. n = 3. (B) Inhibition of FRET dynamics of MBD2 interaction with p66α by ABA and APC in cells. Quantified FRET activities of mock- and compound-treated samples were obtained, and the relative FRET ratios for compounds were calculated by FRETcomp/FRETmock (see Materials and Methods). See also fig. S4B for representative immunofluorescence microscopic photos of cells. n = 2. (C) Dose-dependent suppression of the endogenous MBD2-p66α association by the ABA and APC compounds revealed by in vivo co-IP. Relative fold changes of MBD2 interaction with p66α (right) were obtained by the quantification of immunoblots (left). Data (means ± SD) in (A) and (B) were analyzed using Student’s t test. Ab, antibody; IgG, immunoglobulin G.

  • Fig. 5 Induction of a MET on treatment of mesenchymal type of cancer cells with ABA or APC.

    (A) Representative images showing immunofluorescent signals for VIM or CDH1 (red) and 4′,6-diamidino-2-phenylindole (DAPI) (blue) in LM1 (left) and HCT116 (right) cells treated with 10 μM ABA or APC. Photo credit: S.H.S., Hanyang University. (B) Immunoblots showing the expression levels of EMT markers 48 hours after compound (10 μM) treatment. ACTB was used as a loading control. A.U., arbitrary units. (C) Effects on wound healing, estimated by the recovered surface areas of scraped cell monolayers, 48 hours after treatment with 10 μM ABA or APC. n = 4. (D) ABA and APC (10 μM) impact on cell migration (left) and invasion (right) represented by the number of migrated and Matrigel-invaded cells in Transwell plates 48 hours following compound treatments. n = 3. (E) Relative proliferation rates quantified by 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) assay after 2 days. Cells were treated with 10 μM ABA or APC. n = 2. (F) Cell cycle analysis by fluorescence-activated cell sorter (FACS). Cells were treated with 10 μM ABA or APC. n = 2. (G) Number of spheres counted by the naked eye after 5 days. Cells were treated with 10 μM ABA or APC. n = 3. (H) Representative cell population images for the stem-like CD44hi profile of the ABA- or APC-treated LM1 cells analyzed by FACS. Data from one experiment are shown as averages of two technical replicates. (I) Sensitivity to doxorubicin (left) and cisplatin (right) of the 10 μM ABA- or APC-treated cells quantified by MTT assay. n = 2. (J) Heatmap of mRNA-Seq data, which demonstrates similarity in gene expression between ABA- or APC-treated cells and MBD2 or p66α knockdown LM1 cells. Data (means ± SD) in (E) to (I) were analyzed using Student’s t test. **P < 0.01 and *P < 0.05.

  • Fig. 6 In vivo antimetastatic efficacy of ABA and APC.

    (A) Estimated volume (means ± SEM; P value for significance test by ANOVA) of original tumor developed during the experimental period with and without the drug administration. n = 8 for each group. (B) Body weights of mice monitored at the starting and ending point of experiment. (C) Effects of the compound administration on the xenograft tumor and its metastasis. Estimated tumor weights are presented for the original tumors, whereas the number of nodules developed by lung metastasis is plotted. (D) Representative photographs for lung nodules acquired 29 days after injection of the LM1 cells. Images of metastasized lung tissue sections illustrated by hematoxylin and eosin (H&E) staining and GFP immunohistochemistry (IHC). Yellow arrowhead represents the tumor nodule, and red dotted area indicates the tumor region. Numbers below the H&E-stained tissue sections indicate the average number of tumor nodules in all mice of the same group. Photo credit: M.Y.K. and S.C., Hanyang University. (E) Representative images of H&E-stained tissue sections for the major organs derived from the xenograft NOD-Prkdcscid IL2rg−/− (NPG) mice after completion of the metastasis inhibition tests with the ABA and APC administration (top). Histological scoring (tumor-bearing mice/total mice) for the H&E-stained major organs of the xenograft mice (bottom). Scale bars, 500 μm. Photo credit: M.Y.K. and S.C., Hanyang University. (F) CBC analysis of the ABA- and APC-treated xenograft mice. WBC, white blood cell count; RBC, red blood cell count; HGB, hemoglobin; HCT, hematocrit; MCV, mean corpuscular volume; MCH, mean corpuscular hemoglobin; MCHC, mean corpuscular hemoglobin concentration; RDW, red cell distribution width; PLT, platelet count; N.S., not significant. Data (means ± SD) in (B) to (D) and (G) were analyzed using Student’s t test.

Supplementary Materials

  • Supplementary material for this article is available at http://advances.sciencemag.org/cgi/content/full/5/11/eaav9810/DC1

    Supplementary Materials and Methods

    Fig. S1. Structural information on MBD2 and c-Myc.

    Fig. S2. SEA and cell migration analysis for the nine selected hit compounds targeting MBD2.

    Fig. S3. MD simulations of the selected compound-docked structures of MBD2 and c-Myc.

    Fig. S4. FRET dynamics of ABA and APC to the MBD2-p66α interaction.

    Fig. S5. Effects of ABA and APC on the expression of EMT markers and CSC properties in various breast and colon cancer cells.

    Table S1. Molecular docking result (H-bond, hydrogen bond; N/A, not available).

    Table S2. Selection of compound by in silico assessment of off-target probability by SEA analysis.

    Table S3. Backbone torsion angle variations (95% confidence interval) of the four key residues in the four different MD simulations of MBD2.

    Table S4. T test and P values on the backbone torsion angle summarized in table S3.

    Table S5. Primer sets for vector construction.

    Original data file S1. Figure 1D PDB files.

    References (4669)

  • Supplementary Materials

    The PDFset includes:

    • Supplementary Materials and Methods
    • Fig. S1. Structural information on MBD2 and c-Myc.
    • Fig. S2. SEA and cell migration analysis for the nine selected hit compounds targeting MBD2.
    • Fig. S3. MD simulations of the selected compound-docked structures of MBD2 and c-Myc.
    • Fig. S4. FRET dynamics of ABA and APC to the MBD2-p66α interaction.
    • Fig. S5. Effects of ABA and APC on the expression of EMT markers and CSC properties in various breast and colon cancer cells.
    • Table S1. Molecular docking result (H-bond, hydrogen bond; N/A, not available).
    • Table S2. Selection of compound by in silico assessment of off-target probability by SEA analysis.
    • Table S3. Backbone torsion angle variations (95% confidence interval) of the four key residues in the four different MD simulations of MBD2.
    • Table S4. T test and P values on the backbone torsion angle summarized in table S3.
    • Table S5. Primer sets for vector construction.
    • References (4669)

    Download PDF

    Other Supplementary Material for this manuscript includes the following:

    Files in this Data Supplement:

Stay Connected to Science Advances

Navigate This Article