* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
Download Supplementary Figures 1 - 5, Methods
G protein–coupled receptor wikipedia , lookup
Ancestral sequence reconstruction wikipedia , lookup
Lipid signaling wikipedia , lookup
Point mutation wikipedia , lookup
Biochemistry wikipedia , lookup
Signal transduction wikipedia , lookup
Proteolysis wikipedia , lookup
Biochemical cascade wikipedia , lookup
Multi-state modeling of biomolecules wikipedia , lookup
Metalloprotein wikipedia , lookup
Western blot wikipedia , lookup
Ligand binding assay wikipedia , lookup
Clinical neurochemistry wikipedia , lookup
Interactome wikipedia , lookup
Drug design wikipedia , lookup
Nuclear magnetic resonance spectroscopy of proteins wikipedia , lookup
Homology modeling wikipedia , lookup
Protein–protein interaction wikipedia , lookup
Supplementary data Material and Methods Immunoprecipitation, silver stain, Mass spectrometry and Immunoblotting. After incubation of HUVECs with 20 M GHCer for 2 hrs at 37°C, cells were washed three times with cold PBS and scraped from the dish in cold immunoprecipitation buffer (50 mM HEPES, pH 7.0/150 mM NaCl/ 10% glycerol/1.0% NP-40/1.5mM MgCl2/1mM EGTA/10 M sodium pyrophosphate/100 mM NaF/1 mM PMSF/protease inhibitor cocktail (Roche), and then sonicated in ice for 5 min. The lysates were centrifuged (12,000xg for 5 min), and supernatant was assayed for protein content. Aliquots (2mg/ml of protein) of samples were incubated at 4°C overnight with mAbVK9, anti-TRAX (Santa Cruz Biotechnology, Santa Cruz, CA) or isotype control antibodies (2g) and then mixed with protein A sepharose beads (GE) and incubated at 4°C for 1hrs. The immune complexes bound on the beads were washed three times with immunoprecipitation buffer, and the bound proteins were eluted by boiling in sample buffer for 10 min. Proteins were separated by SDS/PAGE and stained with silver according to standard procedure. In-gel digestion of a protein band with MW~ 35kDa found only in mAbVK9 treated cells was performed, followed by LTQ-FT analysis. In addition, SDS/PAGE gels were immunoblotted with the anti-TRAX or PLC1 antibody (Santa Curz). Bound lipids were extracted by chloroform/ methanol (1:2, vol/vol) and dried in a vacuum centrifuge followed by Dot-blot immunostaining. Dot-blot immunoblotting was performed by Bio-Dot microfiltration apparatus (Bio-Rad) using mAbVK9. Sorting of MCF-7 cells For inoculation into NSG mice, Globo-Hlow and Globo-Hhigh MCF-7 cells sorted by FACS Aria were suspended in serum-free MEM at a concentration of 5 × 106 cells in 100 μl with 2mg/ml matrigel, and injected subcutaneously (s.c.) into the mammary gland of 8-week-old female NSG mice. Tumor volumes were monitored with a caliper. The mice were sacrificed approximately 32 days after inoculation, and the tumors were removed and for further FACS and histologic analysis. Homology modeling of PLC1 protein The structure of human PLC1 protein was modeled using PLC3 (Protein Date Bank code: 4GNK; Chain E) (1), which is homologous to PLC1, as a template. First, the alignment of protein sequences of C-terminal domain of PLC3 with its homolog was made by Blastp (http://blast.ncbi.nlm.nih.gov). The 3D structure of PLC1 was then constructed using Modeller 9v4 (2), with the functions of the AUTOMODEL class in python scripts. The “Discrete Optimized Protein Energy” (DOPE) method (2) was used to select the best model from the 10 initially-generated models. Finally, the model was subjected to energy minimization using DeepView vers. 3.7 (3) with the GROMOS 43B1 force field, until the delta E between the two steps dropped below 0.05 KJ/mol. Prediction of PLC1-binding region on TRAX The potential PLC1-binding region on TRAX was predicted using protein-protein docking which was performed on GRAMM-X Server v.1.2.0 (http://vakser.bioinformatics.ku.edu/resources/gramm/grammx/) (4) to generate 40 potential PLC1-TRAX complexes. The inter-molecular interaction between PLC1 and TRAX was further estimated by HotLig package (5) to sort out the top 10 ranked PLC1-TRAX models. The prediction of PLC1-binding region on TRAX was based on binding frequency of a TRAX atom contacted by a PLC1 atom in these 10 models. According to the hydrophobic potential function in HotLig (5), hydrophobic effect occurs when the molecular surface distance between two atoms is < 5 Å. Hence, here we defined that two atoms are “binding” when their atomic surface distance is < 5 Å. The binding frequency (f) of a TRAX atom i was counted using the following equation: Binding frequency f i N binding (i) N total The Ntotal is the number of total models (i.e., Ntotal = 10), and the Nbinding(i) is the count of PLC1 models binding with atom i. The high binding frequency of an atom on TRAX means that it possesses high probability to interact with PLC1. Molecular docking of Globo-H against TRAX Molecular docking of Globo-H against TRAX was performed using Dock 5.1.1 (6) to generate possible protein-ligand binding modes. Subsequently, the HotLig (5) software was applied to analyze intermolecular interaction. To predict potential binding pocket for Globo-H, the cavities on TRAX were first detected using PscanMS (5); then these cavities were subjected to the Dock software for docking calculation. Kollman partial charges were applied to protein models for the force field calculation (1) in the Dock software. Energy-optimized three-dimensional coordinates of Globo-H were generated using Marvin 5.2.2 (7) and Balloon 0.6 (8) software. Additionally, the Gasteiger partial charges on atoms of Globo-H were calculated by OpenBabel 2.2.3 (9). The parameters for the Dock were set to generate 2000 orientations and 200 conformers per cycle of conformational search in the binding pocket with setting “anchor size” to 1. The docked conformers were then scored and ranked by HotLig(5) to predict molecular interaction between globo H and TRAX. The Fig.s for molecular modeling and interactions were rendered using Chimera (10) and Ligplot (11) software, respectively. The electrostatic potentials of protein were calculated using Delphi (12) with default parameters setting in the Chimera (10). Supplemental Result of computer modeling The TRAX protein is known to associate with the C-terminal region of PLC1, thereby inhibiting the activation of PLC1(13). The structure of human TRAX, as characterized by x-ray crystallography (PDB code: 3PJA) (14), consists of seven alpha helices (1–7 in Supplementary Fig. S2A). Since the structure of human PLC1 is still unknown, we first created a protein model of human PLC1 (Supplementary Fig. S2B) using methods of homology modeling, and then analyzed its molecular interaction with TRAX through protein-protein docking (see Methods). The C-terminal region of PLC1, which is reported to be involved in the interaction with TRAX (13), was predicted as a helix domain (Supplementary Fig. S3) by Psipred (http://bioinf.cs.ucl.ac.uk/psipred/). Therefore, we modeled the main C-terminal helix domain of PLC1 (A890–I1161) based on the structural information of its homolog, PLC3. The similarity of protein sequences between PLC1 and PLC with 34% of identities within the overlapping 272 amino acids (A890–I1161) of PLC1 (Supplementary Fig. S4). As shown in Supplementary Fig. S2B, the modeled structure of the C-terminal domain of PLC1 is composed of three helices. To predict the potential binding region for PLC1 on TRAX, protein-protein docking was performed as described in Methods. Among the top 10 ranked complex models, the binding frequencies of TRAX atoms bound by PLC1 (atomic surface distance < 5 Å) were calculated using eq. 1 described in Methods to show the most likely zone for binding of PLC1 (Fig. 5A–B). The top 1 ranked model of PLC1 binding to TRAX is shown in Fig. 5A; and the representative top 5 models were shown in Supplementary Fig. S5. The predicted PLC1-binding region predominantly distributed around a concave region formed by 4, 5 and 6 (i.e., the red and magenta area in Fig. 5B). The other side of TRAX (the 1-side) seems not be involved in binding of PLC1. To investigate whether Globo-H binds to TRAX, and whether the binding of Globo-H interferes with the interaction of PLC1 on TRAX, the binding mode of the TRAX-Globo-H complex was predicted through molecular docking studies (see Methods). Globo-H consists of a hydrophilic hexasaccharide and a hydrophobic ceramide, which is a conjugate of a sphingosine and a fatty acid (Fig. S2C). To predict the potential binding pocket for Globo-H, the cavities on TRAX were first detected using PscanMS (5); these cavities were then subjected to the Dock software(6) for sampling of potential conformers of Globo-H docking to the TRAX protein structure. Subsequently, these conformers were scored and ranked by HotLig (5) to predict the binding mode and molecular interactions between Globo-H and TRAX. Fig. 5C–5D show the predicted binding mode of Globo-H in complex with TRAX. Generally, the Globo-H molecule binds to the predicted PLC1-binding region on TRAX (Fig. 5C). In addition, the electrostatic interaction of TRAX with Globo-H was computed and evaluated using Delphi (Fig. 5D). The carbon chain of sphingosine binds to the hydrophobic groove formed by 4, 5 and 6, which coincides with the most likely binding region for PLC1 as described above (red and magenta in Fig. 5A–5C). In addition, the hexasaccharide of Globo-H interacts with a polar surface region (Fig. 5D), which also overlaps with part of the predicted PLC1-binding region (compared to Fig. 5C). On the other hand, the 16-carbon fatty acid of Globo-H binds to another hydrophobic pocket between 3 and 6, which is beside the PLC1-binding region. The detailed molecular interactions between Globo-H and TRAX were further illustrated by a schematic diagram in Fig. 5E. Overall, Globo-H might bind to TRAX through hydrophobic and electrostatic interactions, and competitively interfere with the binding of PLC1. Supplementary Fig.: Fig. S1: Globo-H expression on breast cancer cells and GHCer shed from MDA-MB-157 cells into MVs. The expression of Globo-H on (A) MCF-7 cells and (B) MDA-MB-157 cells was determined by flow cytometry with mAbVK9. (C) Microvesicles (MVs, 50g) harvested from the supernatants of MDA-MB-157 cells (MB-MVs) were incubated with HUVECs for 4h and examined for surface Globo-H. (D) Serial dilutions of MB-MVs, GHCer, and fresh medium were probed with mAbVK9. Caveolin-1 (Cav-1) is the marker for MVs. Serial dilutions of MVs, GHCer, and fresh medium were probed with mAbVK9. Caveolin-1 (Cav-1) is the marker for MVs. Fig. S2: Structure of TRAX, PLC1 and Globo-H. (A) The structure of TRAX consists of seven helices (PDB code: 3PJA). (B) Predicted structure of the C-terminal helix domain (A890–I1161) of human PLC1. (C) Structure of GHCer. GHCer consists of a hydrophilic hexasaccharide and a hydrophobic ceramide, which is a conjugate of sphingosine and fatty acid. Fig. S3: Secondary structure prediction of PLC1. The C-terminal region of PLC1 was predicted as a helix (light purple color) domain by Psipred; the yellow color indicates the sequences of -sheet. Fig. S4: Amino-acid sequence alignment of PLC1 and PLC3. The similarity of protein sequences between PLC1 and PLC3 was 58% with 34% identities within the overlapping 272 amino acids (A890–I1161) of PLC1. Fig. S5: Prediction of TRAX-PLC1 binding complexes. Reference in Supplemental data 1. Lyon AM, Dutta S, Boguth CA, Skiniotis G, Tesmer JJ. Full-length Galpha(q)-phospholipase C-beta3 structure reveals interfaces of the C-terminal coiled-coil domain. Nature structural & molecular biology. 2013;20:355-62. 2. Eswar N, Webb B, Marti-Renom MA, Madhusudhan MS, Eramian D, Shen MY, et al. Comparative protein structure modeling using Modeller. Current protocols in bioinformatics / editoral board, Andreas D Baxevanis 3. [et al]. 2006;Chapter 5:Unit 5 6. Guex N, Peitsch MC. SWISS-MODEL and the Swiss-PdbViewer: an environment for comparative protein modeling. Electrophoresis. 1997;18:2714-23. 4. Tovchigrechko A, Vakser IA. GRAMM-X public web server for protein-protein docking. Nucleic acids research. 2006;34:W310-4. 5. Wang SH, Wu YT, Kuo SC, Yu J. HotLig: a molecular surface-directed approach to scoring protein-ligand interactions. Journal of chemical information and modeling. 2013;53:2181-95. 6. Moustakas DT, Lang PT, Pegg S, Pettersen E, Kuntz ID, Brooijmans N, et al. Development and validation of a modular, extensible docking program: DOCK 5. Journal of computer-aided molecular design. 2006;20:601-19. 7. Marvin was used for drawing, displaying and characterizing chemical structures, substructures and reactions. ChemAxon (http://wwwchemaxoncom). 2009;Marvin 5.2.2. 8. Vainio MJ, Johnson MS. Generating conformer ensembles using a multiobjective genetic algorithm. Journal of chemical information and modeling. 2007;47:2462-74. 9. Guha R, Howard MT, Hutchison GR, Murray-Rust P, Rzepa H, Steinbeck C, et al. The Blue Obelisk-interoperability in chemical informatics. Journal of chemical information and modeling. 2006;46:991-8. 10. Pettersen EF, Goddard TD, Huang CC, Couch GS, Greenblatt DM, Meng EC, et al. UCSF Chimera--a visualization system for exploratory research and analysis. Journal of computational chemistry. 2004;25:1605-12. 11. Wallace AC, Laskowski RA, Thornton JM. LIGPLOT: a program to generate schematic diagrams of protein-ligand interactions. Protein engineering. 1995;8:127-34. 12. Rocchia W, Alexov E, Honig B. Extending the applicability of the nonlinear Poisson-Boltzmann equation: Multiple dielectric constants and multivalent ions. J Phys Chem B. 2001;105:6507-14. 13. Uversky VN, Aisiku OR, Runnels LW, Scarlata S. Identification of a Novel Binding Partner of Phospholipase Cβ1: Translin-Associated Factor X. PLoS ONE. 2010;5:e15001. 14. Ye X, Huang N, Liu Y, Paroo Z, Huerta C, Li P, et al. Structure of C3PO and mechanism of human RISC activation. Nature structural & molecular biology. 2011;18:650-7.