1. Introduction
Breast cancer is the most prevalent cancer for women both in the developed and the developing world. Approximately 30% of the women diagnosed with early-stage disease in turn progress to metastatic breast cancer, for which treatments with anti-breast cancer therapeutic agents are needed. Although many current anti-breast cancer therapies can alter tumor growth, in most cases the effect is not long-lasting. Cancer drug resistance is thought to reduce seriously the effectiveness of current anti-breast cancer therapies, which caused around 50% of all treated patients relapsed [1] -[3] . This indicates need for new agents, which are safer, more effective, and potentially able to extend the survival of breast cancer patients.
Antimycin A3, a mixture of the two nine-membered dilactones A3a and A3b isolated from Streptomyces sp., is an active agent that inhibits the electron transfer activity of ubiquinol-cytochrome c oxidoreductase and prevents the growth of human cancer cells (Figure 1). Antimycin A3 was also found to induce apoptosis of cancer cells by selectively killing the cancer cells that expressed high levels of anti-apoptotic Bcl-2 with IC50 of 50 µM on Hela cells [4] -[6] . While Bcl-2 is known to be over-expressed in 70% of breast cancer cells [7] , it is reasonable to expect antimycin A3 to induce apoptosis in those cells. Thus, it is also quite reasonable to expect its analogue to have a similar or higher anti-breast cancer activities. In this work, antimycin A3 analogues (Figure 2), are designed and subsequently simulated based on their interactions with receptor binding target Bcl-2 by a computational molecular docking approach. The top-ranked compounds showing stronger interaction, better affinity, as well as a greater inhibitory activity than antimycin A3 against breast tumor receptor binding target Bcl-2, may become lead compounds in our next synthesis project.
Studies on the structure-activity relationship of antimycin A3 by Miyoshi et al. in 1995 revealed that the nine- membered dilactone core in antimycin A3 was less effective for anticancer activity than 3-formamidosali-cylyl moiety [8] . Pettit et al. (2007) reported that respirantin which has an 18-membered polylactone core instead of nine-membered dilactone core in antimycin A3, showed stronger cytotoxicity than antimycin A3 on mouse leukimia P-388 cells and breast MCF-7 cells [9] (Figure 1). It has also been reported that the presence of hydroxyl groups in bioactive compounds significantly increase their biological activities due to the enhancement of its solubility in water, which is one of the important factors influencing the efficacy of drugs [10] . These facts suggested that, it is quite possible to design novel antimycin A3 analogues by replacing the nine-membered dilactone core in antimycin A3 with either an 18-membered polylactone core or polyhydro-xylated 18-membered polylactone core that contributes to the improvement of its anticancer activity.
In a recent study, we succeeded in synthesizing of novel polyhydroxylated 18-membered analogue of antimycin A3 (compound 14), which showed a potent anticancer activity against breast MDA-MB-231 cells [11] . It revealed that the polyhydroxylated 18-membered core was very important for anti-breast cancer activity. Therefore, in this work, the nine-membered dilactone core of antimycin A3 was replaced by an 18-membered tetralactone core in analogues 10, 11 and 12, and was replaced by polyhydroxylated 18-membered tetralactone core in 13 - 20. To study how the stereochemistry can affect the binding capability, we designed four hydroxyl groups with bottom facial stereochemistry on the 18-membered core in 13 - 16, and, in contrast to those analogues, with the top facial stereochemistry in 17 - 20. Subsequently, in order to increase the anticancer activity, we introduce two parts of 3-formamidosalicylyl moiety in 11, 14, and 18. To explore how the simple substitutions on 3- formamidosalicylyl moiety can influence the binding capability on receptor target Bcl-2, we substitute the existing hydroxyl group in formamidosalicylyl moiety with benzyloxy group in 10, 13, and 17. Whereas 12, 15, and 19 were designed by replacing the hydroxyl group in 3-formamidosalicylyl moiety with methoxy group. Furthermore, 3-formamidosalicylyl moiety was modified into 3-N-methylformamido-2-methoxy-benzoyl moiety in 16 and 20. In this work, we also investigated the interaction of some benzoic acid ring segments
![]()
Figure 1. Structure of antimycin A3 and respirantin.
![]()
Figure 2. Structure of designed compound.
(compound 3 - 9), the 18-membered tetralactone (1), and 18-membered tetraol (2) on receptor binding target Bcl-2 of breast cancer.
2. Methodology
In this research, we simulated some analogue compounds based on their interactions with Bcl-2 breast cancer, using computer software applications (Molecular method) [12] to determine the best compounds [13] . Analysis and screening were based on Gibbs Free energy (∆G) values, affinity, conformation of the structure, and hydrogen bonding interaction between compounds and the target proteins [14] .
2.1. Sequence Alignment and Homology Modelling
Target protein sequences were selected and downloaded from NCBI (http://www.ncbi.nlm.nih.gov/protein/133893254?report=fasta). The multiple sequence alignment method was based on the Clustal W2 program (www.ebi.ac.uk/Tools/clus talw2/index.html). Homology modeling was performed using the Swiss Model which can be accessed through http://www.swissmodel.expasy.org/SWISS-MODEL.html. Swiss model showed that Bcl-2 has structurally homologous to a target protein with template PDB code 1g5mA (target region 3-204, 88.00 % of sequence identity.
2.2. Structural Analysis of Target Protein
Validation of 3D structure from homology modeling was performed using the Protein Geometry program and superimposed using superpose program in MOE 2009.10 software. Based on superimposed the RMSD was calculated to find out structural similarity between template model mutated with 3D structure from homology modeling. Identification of catalytic site of protein target using site finder program in MOE 2009.10 software.
2.3. Optimization and Minimization of 3D Structure
Optimization and minimization of three-dimensional structure of the enzyme were conducted using the software of MOE 2009.10 with addition of hydrogen atoms. Protonation was employed with protonating 3D programs. Furthermore, partial charges and force field were employed with MMFF94x. Solvation of enzymes was performed in the form of a gas phase with a fixed charge, RMS gradient of 0.05 kcal/A0mol, and other parameters using the standard in MOE 2009.10 software.
2.4. Preparation of Compounds
Some antimycin A3 analogues were designed using ACD Labs software. With this software, The analogues were built into three-dimensional structures. The three-dimensional shape was obtained by storing the analogue in the 3D viewer in ACD Labs. Furthermore, the output format was changed into Molfile MDL Mol format using the software Vegazz to confirm for the docking process. Compounds were in the wash with compute program, adjustments were made with the compound partial charge and partial charge optimization using MMFF94 xforcefield. The conformation structure energy of compounds was minimized using the RMS gradient energy with 0.001 kcal/A˚mol. Other parameters were in accordance with the default setting in the software.
2.5. Molecular Docking
The docking process was begun with the docking preparation that was employed using a docking program from MOE 2009.10 software. Docking simulations were performed with the Compute-Simulation dock program. The placement method was conducted using a triangle matcher with 1,000,000 repetition energy readings for each position and other parameters were in accordance with the default settings in the MOE software. Furthermore, scoring functions used London DG, refinement of the configuration repetition force field with 1000 populations. The first repetition was done for 100 times and the second setting was conducted only for one of the best result.
3. Results and Discussion
The twenty designed compounds, including the analogue compounds, 18-membered polylactones, and simple benzoic acid ring segments, were simulated using molecular docking on target protein of Bcl-2 breast cancer. The results are displayed in Table 1. The top-ranked compounds were selected based on low ∆G binding energy, high pKi affinity, and number of hydrogen acceptor/hydrogen donors (hydrogen bonding interaction) to the catalytic site of Bcl-2 target protein.
As shown in Table 1, compared to antimycin A3 and respirantin, the 18-membered tetraol (2) exhibited higher binding energy, affinity, and hydrogen bond interaction on Bcl-2 breast cancer cells, indicating that tetraol 2 has a stronger inhibitory activity against receptor target Bcl-2. In contrast, all the series of benzoic acid ring segments, 3 - 9, showed a less number of hydrogen bonds than antimycin A3 and respirantin. Suggesting that the benzoic acid ring segment itself has low interaction to protein target of Bcl-2 breast cancer. The docking of the analogue compound 10 - 20, produced the five top-ranked compounds, namely, compounds 11, 14, 15, 16, and 20, which showed lower ∆G binding energy value and a higher number of hydrogen bonding interaction than the others compounds. The ∆G values of compounds 11, 14, 15, 16 and 20 are −16.4486, −15.9491, −15.0703, −17.1838 and −17.1553 kcal/mol, respectively, which are better than antimycin A3 and respirantin, with a ∆G value of −11.4295 kcal/mol.These results showed that, compared to antimycin A3, those five top-ranked compounds will form a more stable complex with Bcl-2, as well as, be better able to inhibit and reduce the activity of Bcl-2. The pKi value of the five top-ranked compounds are higher than antimycin A3, indicating that
![]()
Table 1. The properties of twenty designed compounds and antimycin A3 on the catalytic site of Bcl-2.
The blue color represents the top-ranked compound.
they have a higher affinity and interact effectively with the target Bcl-2. Moreover, all of those five top-ranked compounds have a number of hydrogen acceptor/hydrogen donor interactions more than antimycin A3, which demonstrated greater inhibitory activities on receptor target Bcl-2.
The catalytic site of the Bcl-2 breast cancer cells are Arg10, Glu11, Met14, Trp28, Asp29, Ala30, Gly31, Asp32, Val34, Glu46, Asn37, Asp168, and Ala171. If a compound interacts with the catalytic site of the protein target, it will reduce the activity of the target protein, and change the protein conformation. Generally, the interaction of the compound with the complex protein target is the hydrogen bond. The quantities of hydrogen bond interactions of the compound with the catalytic site of the target protein indicate its ability to inhibit the protein target. Figure 3 displays the ligand complex interaction of the five top-ranked compounds (11, 14, 15, 16, and 20) and antimycin A3 with the receptor target Bcl-2. As shown, all the five of top-ranked compounds could change the conformation of the receptor target cavity, and were able to enter the binding site of the receptor target Bcl-2. In addition, compared to antimycin A3, those five top-ranked compounds showed more hydrogen binding interaction against Bcl-2. Hydrogen bond interactions between amino acid residues of Bcl-2 breast cancer with 11, 14, 15, 16, 20 and antimycin A3 are summarized in Table 2.
As shown in Table 2, all the analogue compounds 11, 14, 15, 16 and 20 have a higher number of hydrogen bonds to the protein target Bcl-2 than that of the original antimycin A3. Compared to respirantin, compound 11, 14 and 20 have a higher number of hydrogen bonds to the Bcl-2. Both compound 11 and 14 which form four
![]()
Table 2. Hydrogen bond interactions of antimycin A3, respirantin, compound 11, 14, 15, 16, and 20 with Bcl-2.
The red color represents the catalytic site of Bcl-2.
![]()
![]()
(a) (b)
Figure 4. Conformation of compound 11 (a) and compound 14 (b) on catalyticsite of Bcl-2 breast cancer.
hydrogen bonds to the active site of Bcl-2, exhibited the strongest inhibitory activity. Compound 11 binds to the Bcl-2 catalytic site at the Glu11, Asp32, Asp168 and Arg10 residue whereas, 14 binds to the Bcl-2 catalytic site at the Glu11, Asp32 (two hydrogen bonds) and Glu46 residues. The three dimensional conformation of the two best analogues 11 and 14 on the catalytic site Bcl-2 are given in Figure 4. The docking results in Figure 4 revealed 11 which bears an 18-membered tetralactone core and two parts of 3-formamidosalicylyl moiety as a ligand, has more binding interaction, a more stable conformation and a stronger inhibitory activity on the catalytic site of Bcl-2 than antimycin A3. Similar to 11, compound 14 bearing the18-membered core with four hydroxyl groups at the bottom facial stereochemistry as a ligand, also showed stable conformation and strongly inhibited the activity of the Bcl-2 catalytic site. Consistent with the previous in-vitro assay [11] , the docking result of synthesized analogue 14 confirmed that introducing two parts of 3-formamidosalicylyl moiety and replacing the nine-membered dilactone core of antimycin A3 with the 18-membered tetraol core in 14 could remarkably increase its anti-breast cancer activity. Moreover, replacing the nine-membered dilactone core of antimycin A3 with the 18-membered tetralactone core in analogue 11, could also greatly improve its inhibitory activity against the receptor target Bcl-2 of breast cancer. Thus, compound 11 and 14 are promising candidates for new anti-breast cancer agents, and should be considered as the lead compounds in the next synthesis project.
4. Conclusion
In conclusion, we have simulated twenty designed compounds by molecular docking approach. Among them, the analogues 11 and 14 which have an 18-membered tetralactonecore and 18-membered tetraol core, respectively, demonstrated stronger inhibitory activity and greater interaction with amino acid residues in the catalytic site of Bcl-2 breast cancer compared to the original antimycin A3.
Acknowledgements
We wish to express our gratitute to Directorate of Research and Public Service University of Indonesia for the Research Grant, and to the Graduate School of Materials Science, Nara Institute of Science and Technology (NAIST), Japan, for International Research Collaboration Program (NAIST Global Initiative Program).
NOTES
*Corresponding author.