Computational Study of Anticancer Drug Resistance Caused by 10

Subscriber access provided by CORNELL UNIVERSITY LIBRARY

Article

A Computational Study of Anti-Cancer Drug Resistance caused by 10 Topisomerase I Mutations, Including 7 Camptothecin Analogs and Lucanthone Kelly Ann Mulholland, and Chun Wu J. Chem. Inf. Model., Just Accepted Manuscript • DOI: 10.1021/acs.jcim.6b00317 • Publication Date (Web): 26 Aug 2016 Downloaded from http://pubs.acs.org on August 27, 2016

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

Journal of Chemical Information and Modeling is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 55

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Journal of Chemical Information and Modeling

A Computational Study of Anti-Cancer Drug Resistance caused by 10 Topisomerase I Mutations, Including 7 Camptothecin Analogs and Lucanthone

Kelly Mulholland and Chun Wu*

College of Science and Mathematics, Rowan University, Glassboro, NJ, 08028 USA

*To whom correspondence should be addressed: [email protected]

1

ACS Paragon Plus Environment


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Abstract Although Camptothecin and its analogs as Topoisomerase I poisons can effectively treat cancers, serious drug resistance has been identified for this class of drugs. Recent computational studies have indicated that the mutations near the active binding site of the drug can significantly weaken the drug binding and cause drug resistance. However, only Topotecan and three mutations have been previously analyzed. Here we present a comprehensive binding study of 10 Topoisomerase I mutants (N722S, N722A, D533G, D533N, G503S, G717V, T729A, F361S, G363C, and R364H) and 8 poisons including 7 Camptothecin analogs as well as a new generation Topoisomerase I drug, Lucanthone. Utilizing Glide docking followed by MMGBSA calculations, we determined the binding energy for each complex. We examine the relative binding energy changes with reference to the wild type, which are linked to the degree of drug resistance. On this set of mutant complexes, Topotecan and Camptothecin showed much smaller binding energies than a set of new Camptothecin derivatives (Lurtotecan, SN38, Gimatecan, Exatecan and Belotecan) currently under clinical trials. We observed that Lucanthone exhibited comparable results to Topotecan and Camptothecin, indicating that it may serve as a promising candidate for future studies as a Topoisomerase I poison. Our docked results on Topotecan were also validated by a set of molecular dynamics simulations. In addition to a good agreement on the MMGBSA binding energy change, our simulation data also shows there is larger conformation fluctuation upon the mutations. These results may be utilized to further advancements of Topoisomerase I drugs that are resistant to mutations.

2


Page 2 of 55

Page 3 of 55

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Introduction Due to the limited rotation in cells, strand separation in DNA by polymerasehelicase complexes can induce supercoiling, creating tension at the replication fork. Nuclear Topoisomerase I (Top I) prevents the torsional stress of supercoiled DNA by nicking a strand of DNA resulting in a relaxed complex.1 This mechanism of action involves a key residue, Tyrosine 723, which induces nucleophilic attack on a DNA phosphodiesterase bond through an esterification reaction forming a covalent3’phosphotyrosine binary enzyme-DNA complex.2 This new bond is then attacked by the hydroxyl group on the DNA strand and the double-stranded DNA is released. When relegation does not occur, Top I will remain covalently attached to DNA, blocking the replication fork, and apoptosis will result.3 Camptothecin (Figure 1) and its analogs bind to this active site, decreasing rates of DNA relegation while increasing levels of Top IDNA covalent complexes, and as a result have been extensively studied as cancer therapeutics.4 To date, there are two Top I drugs approved by the FDA, and even more undergoing clinical development. These drugs have been widely used in treating smallcell lung cancer, ovarian cancer, and cervical cancer as second and third-line cancer therapeutics.5-7 However, point mutations have been identified as one of the major mechanisms that lead to serious drug resistance.8, 9 Camptothecin, originally isolated in 1966 by Wall and colleagues from the Chinese Camptothecin acuminate tree1, showed remarkable anticancer activity in preliminary clinical trials but also low solubility and severe adverse drug reaction.10 The efforts in creating more soluble Camptothecin analogs lead to two Top I drugs,

3



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Topotecan (Hycamtin) and Irinotecan (Camptostar), which were approved by the FDA in 1996 (Irenotecan is a prodrug that is converted into SN38, the actual Top I poison).11, 12

Figure 1. Topoisomerase I inhibitors. Camptothecin scaffold (A), non-Camptothecin drug Lucanthone (B), and Camptothecin derivatives with corresponding R groups in A (C). Hycamtin and Camptosar are widely used in treating small-cell lung cancer, ovarian cancer, and cervical cancer as second and third-line cancer therapeutics.5-7 In 2014, the Hycamtin oral capsule was developed and approved by FDA. New optimizations are also undergoing to develop more effective anti-cancer drugs with less adverse effects (Figure 1A-C).13 For example, Lurtotecan, developed in 1994 by Luzzio et al., exhibited much greater potency than Camptothecin as a Top I poison in cleavable complex enzyme assays and ex vivo cell cytotoxicity assays.14 Belotecan, a water soluble 4


Page 4 of 55

Page 5 of 55

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


derivative of Camptothcin, was recently approved for clinical use in South Korea. Gimatecan and Exatecan showed great progress in solubility and clinical tolerability, making them great candidates as potential oral capsules.15 Despite these advances, however, drug resistance has been prevalent in Campothecins.8, 9, 16

Point mutations in the Top I-DNA cleavage complex, identified in both mammalian

and yeast cell lines, are suggested to play a major role in drug resistance.17 Pommier and colleagues suggested eight mutations in the core domain and C-terminal domain (active binding site) which may lead to Camptothecin resistance (Figure 2).18 Figure 3 provides a visualization of these key residues relative to the Topotecan drug. Chrencik and colleagues provided a structural analysis on Topotecan bound to complexes F361S and N722S which shed even more light on the role of mutations in causing the drug resistance.19 As resistance remains problematic to Camptothecin derivatives, experts are seeking non-Camptothecin drugs with less resistance. For example, Lucanthone has been investigated as a promising anti-cancer therapeutic (Figure 1C). Due to its ability to intercalate into Top I as well as its unique scaffold, Lucanthone may exhibit lower drug resistance than the Camptothecin class.20 However, further evidence remains unreported.

Figure 2. Schematic diagram of mutated residue locations

5



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Molecular modeling and simulations have been recently used to decipher the mutation effects on drug binding affinity.21, 22 Pan and colleagues utilized this method to conduct a study on Topotecan resistance by three specific Top I mutations. They found mutations E428K, G503S, and D533G have significant influence on the binding affinities of Topotecan and could indeed be responsible for resistance seen in this drug.22 Sirikantaramas and colleagues used molecular dynamics (MD) simulations to explain the resistance mechanism found in camptothecin producing plants. This group used Topotecan -Top I complex to provide structural insight to the resistance mechanism.23 However, a comprehensive study of all major mutations on major Top I drugs has yet to be completed. In this study, we utilized homology modeling of mutant complexes, Glide extra precision (XP) docking and MMGBSA (molecular mechanics generalized Born with surface area term) calculations to understand the effects of ten key point mutations on binding affinity of eight Top I poisons. We compared these results to the wild type crystal complex, Top I-Topotecan (PDB: 1K4T), which provided structural insights on the drug resistance due to these mutations.1, 4 The MMGBSA binding energy change upon a point mutation was calculated for each poison in all mutant complexes and a good correlation between our MMGBSA data and the available experimental binding data was observed. Interestingly, we found that Topotecan and Camptothecin have much smaller binding energy decrease than a set of new Camptothecin derivatives (Lurtotecan, SN38, Gimatecan, Exatecan and Belotecan) currently under clinical trials. Lucanthone shows comparable results to Topotecan and Camptothecin, indicating it may exhibit the least drug resistance and is therefore a promising candidate for future studies as a Top I

6


Page 6 of 55

Page 7 of 55

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


poison. We investigated the binding nature by further decomposing the binding energy to evaluate the contribution of each energetic component. To validate our docking results on Topotecan, we also conducted a set of Molecular Dynamics (MD) simulations and found not only a good agreement on the MMGBSA binding energy change, but detailed conformation change upon the mutations as well. Although the computational cost of MD simulation is much higher than that of the single-conformation minimization approach, it can improve the performance of MM/GBSA by identifying the correct conformations of the complex due to a better sampling of conformation space. 24

7



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Methods Drug Preparation. The eight Top I drug structures were either obtained from the ZINC database25 or created using ChemDraw Professional 15.0. The drugs were prepared in three steps using Maestro Elements.26 First, hydrogen atoms were added based valence, then Epik (an empirical pka prediction program) was utilized to determine the charge of the molecule at pH 7.26 Finally, the geometry of each drug was optimized by minimizing the potential energy using the default parameters. Wild-type Top I structure and homology models of ten Top I mutants. The human wild type Top I crystal structure in complex with Topotecan, determined by Stalker et al. using X-ray diffraction method, was obtained from the RCSB Protein Data Bank database (PDB ID: 1K4T).4 There is no Mg2+ ion in this complex, consistent with the fact that Eukaryotic Top I enzymes do not require any metal cofactor unlike Prokaryotic enzymes which require magnesium and a single stranded segment of DNA.27 The complex was prepared using the protein preparation wizard of Maestro program in three steps: preprocessing, charge state determination and geometry optimization using default parameters for restrained minimization.26 Using this wild type complex as a template, the ten mutated complexes were generated by altering each of their corresponding amino acids to achieve the point mutation at the desired residue location. These mutant complexes were then prepared using the protein preparation wizard. Glide XP Docking. The Glide docking program with extra precision (XP) scoring was utilized in this docking study.28, 29 The XP scoring function and docking protocol have been developed to reproduce experimental binding affinities for a set of 198

8


Page 8 of 55

Page 9 of 55

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


complexes (RMSDs of 2.26 and 1.73 kcal/mol over all and well-docked drugs, respectively) and to yield quality enrichments for a set of fifteen screens of pharmaceutical importance. This algorithm performs a series of hierarchical searches for optimal drug pose within the binding site of a receptor. The first step involved a rough positioning and scoring followed by torsional energy optimization using the new and efficient OPLS3 non-bonded potential energy grid to endure potential poses.26 The pose conformations of the best candidate were refined once again using Monte Carlo sampling. The final docked pose was accomplished and given a docking score, which combines both empirical and force field based terms. After generating a receptor-grid file, each drug was docked to both the Top I wild type structure as well as all ten mutated homology models using Glide XP docking.29 The best drug pose was output and used in MMGBSA binding energy calculations. Molecular Dynamics System Setup. We constructed seven simulation systems from the crystal ternary complex (PDB ID: 1K4T), Topotecan bound to DNA-protein complexes (Top I: wild type, F361S, R364H, G503S, D533G, N722S, and T729A). Each system was built using SPC as water solvent model using orthorhombic solvent box with 8Å water buffer.30 The system was neutralized using Na+ ions and a salt concentration of 0.15 M NaCl. A recently developed OPLS3 force field was used to represent the ternary complex. The total number of atoms is about 72K for each system. Relaxation and Simulation Protocols. Desmond simulation package30 was used to run all simulations, systems were first relaxed using the default relaxation protocol. This relaxation protocol uses the following steps: 1) Brownian dynamics simulation under the NVT ensemble with temperature 10 K for 100 ps. 2) simulation under the

9



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

NVT ensemble with temperature of 10 K, small time steps and restraints on solute heavy atoms for 12 ps. 3) simulation under the NVT ensemble with temperature of 10 K and restraints on solute heavy atoms for 12 ps. 4) simulation under the NPT ensemble with temperature of 10 K, pressure of 1 bar and restraints on solute heavy atoms for 12 ps. and 5) simulation under NPT with temperature of 300 K, pressure of 1 bar and no restraints for 1.5 ns. After the relaxation, a 100 ns production run was conducted for each of the three systems under the NPT ensemble (300 K and 1 bar) using the default protocol. The temperature was controlled using the Nosé-Hoover chain coupling scheme with a coupling constant of 1.0 ps.31, 32 Pressure was controlled using the Martyna-TuckermanKlein chain coupling scheme33 with a coupling constant of 2.0 ps. M-SHAKE34 was applied to constrain all bonds connecting hydrogen atoms, enabling a 2.0 fs time step in the simulations. The k-space Gaussian split Ewald method35 was used to treat long-range electrostatic interactions under periodic boundary conditions (charge grid spacing of ~1.0 Å, and direct sum tolerance of 10–9). The cutoff distance for short-range non-bonded interactions was 9 Å, with the long-range van der Waals interactions based on a uniform density approximation. To reduce the computation, non-bonded forces were calculated using an r-RESPA integrator36 were updated every three steps. The trajectories were saved at 40.0 ps intervals for analysis. Simulation interaction diagram (SID) analysis. Desmond SID was utilized to analyze the behavior and interaction of proteins and drugs during the course of a simulation. These measures include root mean square deviation (RMSD) shown in Figure S85 and root mean square fluctuation (RMSF).

10


Page 10 of 55

Page 11 of 55

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Convergence of molecular dynamic simulations. To check for convergence of the MD simulations, we analyzed the protein Cα and drug RMSD plots for each trajectory (Figure S85). The limited fluctuations in the plots within last 50 ns indicate that the complex systems have reached a steady state. Trajectory Clustering Analysis. Desmond trajectory clustering was utilized to group complex structures from the simulations. Backbone RMSD matrix was used as structural similarity metric and the hierarchical clustering with average linkage was selected as the clustering method. The merging distance cutoff was set at 2.5Å. The centroid structure (the structure having the largest number of neighbors in the structural family) was utilized to represent each structural family. The centroid structures of populated structural families (>2% of total structure population) are shown in Figure S86 of the supporting material. MMGBSA Binding Energy Calculations. For docking systems, prior to Molecular Mechanism-General Born Surface Area (MMGBSA) binding calculation, the docked complexes were each minimized with restraints on heavy atoms using Maestro Protein Preparation Wizard. For MD systems, the frames in the last 50 ns (50 snapshots) of each system were used. OPLS3 force field37, VSGB 2.0 solvation model 38 and the default Prime procedure was used for the MMGBSA calculation. The default procedure consists of three steps: receptor alone (minimization), drug alone (minimization), receptor-drug complex (minimization). The default minimization method is based on an automatic method that automatically switches between different algorithms (i.e. Conjugated gradient minimization is used when the gradients are large, but switches to the truncated Newton method when they are sufficiently small). The RMSD change for

11



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

each docked complex can be viewed in Table S20. The binding energy was then calculated using Equation 1. This utilizes the molecular mechanics (MM) method to calculate drug-receptor interaction energies (GConformation, GGBELE, Gvdw and Glipo), with a Gaussian smooth dielectric constant functional method for the electrostatic part of the solvation energy (i.e. GB term) and solvent-accessible surface for the non-polar part of the solvation energy (Equation 2). Finally, the binding energy change of the mutant complex was calculated by using the wild type complex as the zero energy reference (Equation 3). Analysis of the decomposition of binding free energy values based on electrostatic, van der Waals, hydrophobic, and conformation interactions provides a more detailed understanding of the effect that resistance places on each complex. Twodimensional diagrams aided in visualizing these decomposition values. Eq 1:

∆𝐺 = 𝐺𝑐𝑜𝑚𝑝𝑙𝑒𝑥 − 𝐺𝑝𝑟𝑜+𝐷𝑁𝐴 − 𝐺𝑙𝑖𝑔

Eq 2: 𝐺 = 𝐺𝑐𝑜𝑛𝑓𝑜𝑟𝑚𝑎𝑡𝑖𝑜𝑛 + 𝐺𝐺𝐵𝐸𝐿𝐸 + 𝐺𝑣𝑑𝑤 + 𝐺𝑙𝑖𝑝𝑜 Eq 3:

∆∆𝐺 = ∆𝐺𝑀𝑢𝑡𝑎𝑛𝑡 − ∆𝐺𝑊𝑇

Note that because the solute conformational entropy is not included in our analysis, the MMGBSA binding energies can only be realistically used to rank drugs rather than provide an absolute binding free energy.29 Nonetheless, extensive and systematic benchmarking studies up to 1864 crystal complexes have shown that MMGBSA is a powerful tool in ranking ligands.24, 39-42

12


Page 12 of 55

Page 13 of 55

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Results Mutant homology model and Glide-docking results discovered subtle changes in drug binding pose as well as side-chain conformation within each complex. We built homology models for 10 mutants based on the wild type crystal structure (Figure 3) and docked each drug to the homology models.4 We then compared the binding poses of the crystal complex to our docked topotecan-Top I complex (Figure S1), and found a similar pose with subtle changes in side chains. Next, we performed MMGBSA calculations in which three geometry optimizations were conducted on a receptor only system, drug only system, the whole complex, and finally an energy calculation.26 The MMGBSA binding energies and its components are tabulated in the supporting document (Table S1-S8). To visualize the subtle binding pose changes of the drug within the mutant structure, we superimposed the structure of all complexes based on the sequence alignment. The aligned, 3-Dimensional (3D) representation of each drug when bound to the complexes can be seen in Figure 4.

13



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Figure 3. Crystal complex structure of Top I with Topotecan in red (PDB ID: 1K4T). The mutated residues are shown in blue, and the key catalytical residue, Y723, is shown in green.

The binding pose of drug in each drug-mutant complex is very similar to the crystal binding pose in Figure 3, in which the lactone ring of drug binds closely to the active Tyrosine 723 residue. An example on the crystal ligand (i.e Topotecan) is included in Figure S1 of the supporting material, the excellent overlap between the crystal pose and docked posed is indicated by small RMSD (1.0 Å) of heavy atoms. However, we noticed subtle conformation change for both the drug (Figure 3) and the side chains of the interacting residues with the drug (data are not shown). The 3D structure of top view and side view, 2D interaction diagram and ∆∆G Decomposition for each complex are included in the supporting document (Figure S5-S84). The interacting residues of each complex (mutant and wild type) with the drugs is also available and provides insight into which residues are present 5Å from the active binding site of each complex (Table S919). Again, subtle differences can be observed for the various drugs and mutants. These drug binding pose changes are also observed in the MD simulation results.

14


Page 14 of 55

Page 15 of 55

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Figure 4. Clustering of 8 Topoisomerase I poisons to WT and 10 Top I mutant complexes. For clarity, only WT Top I is shown in cartoon. For the mutated residues, side chains from WT and mutant are shown in different colors. For drugs, Oxygen atoms are displayed in red, nitrogen atoms are displayed in blue and carbon atoms are displayed in gray.

A good agreement between experimental binding affinity and calculated binding energy validated the computational methods in this study. To validate our computational results, we collected all available IC50 values for Camptothecin derivatives on wild type Top I (Table 1).6, 13, 14 Table 1. Experimental and predicted binding (free) energy of ligands to wild type Top I Ligand

IC50 (nM)

Experimental ∆G* (kcal/mol)

∆MMGBSA (kcal/mol)

-10.0 -9.7 -8.9 -8.7

-105.2 -101.1 -93.1 -83.4

Topotecan 50 LESN38 77 Camptothecin 300 Lurtotecan 416 *ΔG𝐵𝑖𝑛𝑑𝑖𝑛𝑔 = RT × ln(IC50 )

15



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 16 of 55

Encouragingly, there is a good agreement between experimental results and our predicted values, with a Pearson correlation coefficient of 0.9161 (Figure 5A). In addition, the IC50 fold change of Camptothecin in each of these mutated complexes has also been experimentally determined (Table 2).12, 16, 43-45 Table 2. Experimental and predicted binding free energy change of Camptothecin to Top I mutant complexes Experimental Experimental ∆∆G ∆∆MMGBSA FC* (kcal/mol) (kcal/mol) T729A 10 1.4 23.1 G503S 134 2.9 27.9 D533N 220 3.2 30.9 D533G 300 3.4 29.3 G717V 600 3.8 36.9 N722S 974 4.1 45.0 *FC: fold change of dissociation equilibrium constant in comparison to wild type Mutant

Our binding free energy data obtained from Glide docking shows almost the same ranking order in the fold change: N722A > G717V > D533N > G503S > T729A (Figure S2A). Figure 5B shows an excellent agreement between the experimental findings on Camptothecin resistance and our Glide docking calculations, giving a Pearson correlation coefficient of 0.7359. XP docking scores exhibited a lower correlation to the experimental values, so MMGBSA data was used throughout the study.

16


Page 17 of 55

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Figure 5. Correlation between experimental and predicted change in binding free energy of four drugs to WT Top I (Table 1) (A) Correlation between experimental and predicted binding free energy change of Camptothecin to six Top I Mutants (Table 2) (B) Docking results on Topotecan were validated with MD simulation results for wild type and six mutant complexes. Due to the high computational cost associated with MD simulations, we chose the wildtype and six mutants with the crystal drug (Topotecan) for 100 ns MD simulations to validate the docking results. The complexes we chosen by ranking the mutations based on docking results and selecting the point mutations that were most apportioned (Table S22). Protein and drug RMSD was calculated for each of the seven MD systems (Figure S85). The final protein RMSD for all of the systems was about 4Å, indicating that there was further structural relaxation. The systems reach a steady state at about 50 ns, so MMGBSA binding energy was calculated from the last 50 ns (Table 3).

Table 3. MMGBSA through docking and MMGBSA through MD simulations of Topotecan bound to wild type and six mutant complexes Dock MD Complex ΔG ΔΔG ΔG ΔΔG -105.2 0.0 -75.2 0.0 1K4T -95.4 9.8 14.5 R364H -60.7 -89.9 15.3 -60.0 15.3 F361S -80.9 24.3 -59.8 15.4 N722S -86.3 18.9 -58.3 16.9 T729A -85.5 19.7 -54.7 20.6 G503S -85.1 20.1 -53.6 21.6 D533G

The change in binding energy, ∆∆G, based on both docking and MD simulation has a favorable correlation (Figure 6). Although calculated ∆G based on MD results is systematically lower than MMGBSA based on docking results, both exhibit the same trend on ∆∆G (Table 3): mutant complexes are significantly weaker in binding than the 17



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

wild type complex. Additionally, complex D533G has the most effect on binding while complex R364H has the lowest effect.

Figure 6. A Correlation Between ΔΔG Through Docking and ΔΔG Through MD Simulations of Topotecan bound to wild type and six mutant complexes (Table 3)

Among the eight Top I Poisons, Camptothecin, Topotecan and Lucanthone exhibited the lowest mean change in the binding energy to the ten Top I mutants. We carried out the MMGBSA binding energy calculations of each docked drug in each complex. The data was organized appropriately to answer two questions: which drug exhibits the lowest effect due to the ten mutants and which mutant complex is responsible for causing the most effect to these eight drugs. To answer the first question, we grouped our ∆∆G data based on the drug (Table 4 and Figure S2) to determine which mutated residues are most energetically unfavorable for each drug.

18


Page 18 of 55

Page 19 of 55

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Table 4. Change in binding energy for each Topoisomerase poison (ΔΔG kcal/mol) Topotecan D533N (26.7) N722S/A (24.3) D533G (20.1) G503S (19.8) T729A (18.8) G363C (15.4) F361S (15.3) G717V (13.9) R364H (9.8)

Camptothecin F361S (38.8) D533G (29.3) N722S/A (21.8) G717V (16.9) D533N (16.7) G363C (15.5) G503S (11.1) T729A (10.7) R364H (9.1)

Lucanthone F361S (25.5) N722S/A (24.4) D533N (23.9) D533G (19.8) G363C (18.9) G503S (15.8) T729A (15.2) R364H (13.2) G717V (10.3)

Lurtotecan F361S (33.5) R364H (32.9) G503S (32.5) N722S/A (31.1) D533N (28.3) D533G (25.1) G717V (18.3) G363C (14.6) T729A (12.2)

LESN38 D533G (64.7) F361S (30.8) N722S/A (27.9) D533N (25.8) G503S (24.1) G717V (23.8) T729A (23.7) G363C (20.5) R364H (15.5)

Gimatecan D533N (50.4) D533G (50.1) T729A (33.7) G363C (32.3) N722S/A (32.1) G717V (31.1) G503S (21.8) R364H (18.1) F361S (13.8)

Exatecan R364H (53.4) N722S/A (41.0) D533N (39.9) G717V (37.1) G503S (34.9) G363C (33.6) F361S (18.8) T729A (17.1) D533G (10.7)

Belotecan G503S (49.7) D533N (47.8) G363C (45.8) T729A (44.4) D533G (43.8) F361S (31.3) N722S/A (31.1) F361S (31.0) R364H (27.9)

The mean change in MMGBSA binding energies and standard deviations over the ten mutants were calculated and ordered in Figure 7. The mean change in MMGBSA binding energy in the ten mutated complexes for Camptothecin was 19.2 kcal/mol. This predicted drug resistance was the second lowest following Topotecan, which exhibited a mean change in binding energy of 18.3 kcal/mol, indicating that this drug is the least affected by these mutations. Belotecan, with the highest change in binding energy at 38.9 kcal/mol, was most affected by the mutations. Clearly, Lucanthone, Lurtotecan, SN38, Gimatecan, Exatecan and Belotecan have higher mean binding energy change indicating that this group might have higher drug resistance when compared to Topotecan, Camptothecin and Lucantohone which exhibit a much lower mean binding energy change.

19



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Figure 7. Mean change in binding energy (∆∆G) for each Topoisomerase I poison. A) Topotecan B) Camptothecin C) Lucanthone D) Lurtotecan E) LESN38 F) Gimatecan G) Exatecan H) Belotecan

Among 10 Top I mutants, D533G exhibits the highest mean resistance and T729A the lowest resistance for these 8 drugs. To answer the second question, we clustered our ∆∆G data based on the mutant (Figure S3). The mean change in MMGBSA binding energy was calculated as well as the standard deviation and the center-to center distance (center mass of residue sidechain to center mass of drug) from the mutant residue to the drug (Figure 8). Complex T729A, with a mean change in binding energy of 22.0 kcal/mol, has the lowest effect on this group of drugs. Alternatively, Complex D533G, with a predicted value of 22.0 kcal/mol, has the most effect on this group of drugs. Interestingly, D533 is located in the lips region of the binding pocket of the Top I enzyme and forms direct contact with Topotecan in the wild type complex.46

20


Page 20 of 55

Page 21 of 55

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Figure 8. Mean binding energy change (∆∆G) for each mutant and center-to-center distance from mutated residue to ligand. 1) D533G 2) D533N 3) N722A 4) N722S 5) G363C 6) G503S 7) F361S 8) G717V 9) R364H 10) T729A Our data suggest a negative correlation between the mutation-drug distance and the binding energy decrease (Figure 8). We observed that the closer the mutated residue is to the active binding site, the more effect the mutation has on drug binding; and vice versa. There was one exception in this pattern, complex R364H, which is about 3 Å from the binding site and showed a lower change in the binding free energy. Interestingly, we discovered through the decomposition of each drug-R364H complex binding energy that electrostatic interactions generally contributed to the stabilization of the histidine in this mutation. Due to its lower polarity when compared to the wild type arginine residue, the histidine forms more hydrogen bonds with the surrounding residues (Figure S4). In

21



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

contrast, G363C with the similar mutation-drug distance followed the general trend and did not exhibit the same electrostatic stabilization. Complex SN38-D533G shows largest decrease of binding energy change and Lucanthone-G717V shows least decrease of binding energy change. Drug binding was affected by each complex after a thorough evaluation of both 2D interaction diagrams and MMGBSA binding energy of each of the 80 mutated complexes (see Figures S4-84). Upon further inspection of each mutant-drug complex two were particularly interesting. Complex SN38-D533G (Figure 9) exhibited the highest resistance with a binding free energy change (∆∆G) of 64.7 kcal/mol. Decomposition indicates a high electrostatic (GBELE) and van der Waals (VDW) contribution (Figure 9A). Significant contribution from van der Waals results in the loss of hydrogen bonding, which can be seen in the loss of hydrogen bonds to both a water molecule and Lys 532 which is not present in the SN38-D533G complex (Figure 9B-C).

22


Page 22 of 55

Page 23 of 55

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Figure 9. LESN38-D533G mutated complex. The decomposition of the ∆∆G (A), interacting residues with wildtype complex (B), and interacting residues with mutant complex (C). The Lucanthone-G717V Complex exhibits relatively low resistance when compared to the other 79 complexes (Figure 10A). With a binding free energy value of 10.3 kcal/mol, there are both hydrophobic and van der Waals contributions that may be responsible for the loss of π-π stacking between TGP11 and Ring A that is seen in the Lucanthone-WT complex, but not the Lucanthone-G717V complex (Figure 10B-C). This finding is especially interesting due to the fact that Lucanthone, a non-Camptothecin Top I poison, is less affected by each of the mutated complexes when compared to each Camptothecin analog. Furthermore, SN38, who’s prodrug was approved by the FDA in 1998 for the treatment of cancers in the colon and rectum, exhibits significant resistance to all of the mutated complexes.26 The decomposition of binding free energy along with two-dimensional diagrams obtained through molecular docking provides insight into the 23



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

resistance mechanisms of each of these complexes (Figures S4-83). Understanding the contributions to resistance allows for more accurate drug optimization and design.

Figure 10. Lucanthone-G717V mutated complex. The decomposition of the ∆∆G (A), interacting residues with wildtype complex (B), and interacting residues with mutant complex (C). Subtle conformation change and fluctuation observed from MD simulations on Topotecan bound to the wild type and six mutant complexes. The complexes, extracted from the trajectory of each complex, were categorized into structural families based on a clustering analysis as described in the method section. By setting a threshold of 2% population, one structural family was identified for the wild type and complexes D533G, F361S, and R364H and two structural families were identified for complexes G503S, N722S, T729A and G533N (Figure S86). Protein RMSF was calculated and 24


Page 24 of 55

Page 25 of 55

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


determined subtle changes in the protein structure comparing wild type to the mutant complexes (Figure 11).

A)

B)

Figure 11. A 3D representation of Topoisomerase I enzyme: Topotecan is shown in red, blue indicates core domain, yellow indicates N-terminal Domain, cyan indicates Linker, and green indicates C-terminal Domain (A) and protein RMSF of Topotecan bound to each mutated complex (B)

At the linker (residues 635 to 697), the greatest conformational change is observed with complex D533G RMSF of 7Å and the wild type RMSF of 4Å. Minor fluctuations are also observed throughout the structure, however the mutants exhibit the higher conformational change overall when compared to the wild type. Drug RMSF was also calculated to determine conformational changes in the drug (Topotecan) when bound to each complex (Figure 12).

25



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Figure 12. A 2-Dimensional representation of Topotecan atom numbers (A) and Drug RMSF of topotecan bound to each mutated complex (Atom Index Number corresponds directly to the atom numbers in A) (B)

Subtle changes are observed throughout the structure, however the largest conformational change is seen around the methyl sidechain (atom numbers 12 and 13). A structural comparison of the complex with the greatest change in binding energy, D533G, and the wild type can seen in Figure 13 (Figure S87 compares all mutant complexes). Here, the structual changes in the Top I linker, DNA and drug shifting is observed. It is clear that the drug shifting that appears in the docking results is also noticable in the MD simulations (For a complete list of interacting residues with Topotecan bound to each complex, see Table S22).

26


Page 26 of 55

Page 27 of 55

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Figure 13. Structural comparison of Top I (A), DNA (B), and Topotecan (C) of wild type and D533G complex as a result of MD simulations. Wild type complex is shown in red and D533N complex in blue. Topotecan bound to wild type is shown in yellow and in green when bound to D533N. (wild type compared to all mutant complexes can be seen in Figure S87)

27



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Discussion Molecular docking and simulations are powerful tools used to probe the mutation effects on drug resistance. Previous research conducted on the effects of mutations on Camptothecin has provided insight into the experimentally determined cellular resistance exhibited in these complexes.17-19, 46 Pan et al. has investigated the Topotecan binding affinity changes due to three point mutations (e.g. E418K, G503S, and D533G) using molecular dynamics simulations followed by MMGBSA binding energy calculations.22 The binding energy changes from their calculations correlate well with drug resistance data. Although these studies provide invaluable insights on the Topotecan drug resistance caused by three common point mutations, the effects that all known mutations have on other Top I drugs remains elusive. Given high computational cost of MD simulation, it is not trackable to use MD simulations to study the large number of drug-mutation complexes (8x10). In this study, we utilized glide XP docking followed by MMGBSA calculations to analyze the role ten mutations play on the resistance of Top I to Camptothecin, seven Camptothecin derivatives, and a non-Camptothecin, Lucanthone. To validate our calculations, we first inspected the correlation between our computed energies of binding to the known experimental data. Pearson’s correlation coefficient between our MMGBSA binding data and the experimental IC50 of four compounds to the wild type Top I is 0.916. Pearson’s correlation coefficient between our binding energy changes and the experimental IC50 change due to ten point mutations is 0.736. To further validate the docking results, we conducted MD simulations on Topotecan bound to the wild type and six mutant complexes. The drug binding poses from the docking were maintained with subtle changes. The MMGBSA binding energy

28


Page 28 of 55

Page 29 of 55

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


correlation coefficient between docking and MD simulations is reasonable (0.7355). Mutant complexes had larger structural fluctuation than WT, consistent with weaker binding energy than WT. Following these validations, we extended our analysis to eight drugs and ten mutations. By applying the same calculations to every drug-mutant complex, we determined that mutations indeed exhibit an effect on all the drugs studied. Interestingly, Camptothecin, Lucanthone and Topotecan are the least affected by mutations. The next-generation drug, Lucanthone, is expected to exhibit differences as it does not share a similar molecular scaffold to the Camptothecin analogs. However, it is surprising that SN38 (FDA-approved Camptosar) and the other Camptothecin analogs are significantly more effected than Camptothecin and Topotecan. Providing an alternative perspective, the mean change in binding free energy was organized according to mutant complexes to identify the mutation responsible for the most resistance (Figure 7). Interestingly, we observed that the furthest residue (T729A at 10+Å distance) from the drug is affected the least while the residue closest to the drug (D533G/N at 3Å distance) experienced the highest affect overall (Figure 7). This finding confirms that the location of the mutation indeed plays an important role in the functionality of the Top I-drug complex. Complex R364H was the only outlier, and when compared to the closest reside, complex G363C, we discovered that the histidine residue may be responsible for the lower binding energy because it forms more hydrogen bonds with the surrounding residues due to its low polarity (Figure S4).

29



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Conclusions Computational methods are becoming increasingly important in drug discovery as these methods economically provide detailed structural and energetic information that traditional drug development methods lack. By understanding drug-target interactions in detail, researchers are able to efficiently design promising drug candidates. In our study we utilized Glide XP docking and MMGBSA calculations to analyze resistance mechanisms of ten mutations surrounding the active binding site of Camptothecin, its analogs and a next-generation Top I poison, Lucanthone. To validate our calculations, we first correlated our predictions against known experimental data. Pearson’s correlation coefficient between our MMGBSA binding data and the experimental IC50 of four compounds to the wild type Top I is 0.9161. Pearson’s correlation coefficient between our binding energy changes and the experimental IC50 change due to ten point mutations is 0.7359. Our docked results on Topotecan were also validated by a set of molecular dynamics simulations. Our docking data shows that a binding energy decrease is present in every drug-mutant complex. However, the magnitude of the change depends on the specific mutation and the drug structure. Among the eight Top I poisons, Camptothecin, Topotecan and Lucanthone exhibit the lowest mean change in the binding energy of all the Top I mutants. Among the Top I mutants, D533G exhibits the highest mean resistance and T729A the lowest resistance for the eight drugs. Interestingly, we observed that the furthest residue (T729A at 10+Å distance) from the drug has the least resistance while the residue closest to the drug (D533G/N at 3Å distance) has the highest resistance overall. Our detailed binding information will be valuable for future Top I poison modifications. Our study on Top I poisons supports the use of docking approaches in assessing mutation

30


Page 30 of 55

Page 31 of 55

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


effects on drug resistance quickly and with reasonable accuracy for other anticancer/anti-virus drugs where point mutations play a major role in drug resistance.

31



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Supporting Information Tables summarizing the decomposition of MMGBSA binding energy for each Top I poison, tables containing amino acid residue interactions for each complex using both Glide docking and MD simulations, a table containing the RMSD for each drugmutant complex, plots of the change in binding energy of each mutant complexes, plots of the change in binding energy of each Top I poison, a table comparing mutated complexes G363C and R364H, a compilation of 3D and 2D diagrams and decomposition graphs of every mutated complex (80 total), a compilation of the structures with the most populated complex structure families as a result of clustering, and finally a structural comparison between the wild type and six mutant complexes provided by MD simulations. This information is available free of charge via the Internet at http://pubs.acs.org

32


Page 32 of 55

Page 33 of 55

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Acknowledgements This work was supported by Rowan Startup and SEED grant and the National Science Foundation under Grant NSF ACI‐1429467 and XSEDE MCB160004.

33



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

References 1. Pommier, Y., Drugging Topoisomerases: Lessons and Challenges. ACS Chem. Biol. 2013, 8, 82-95. 2. Eng, W.; Pandit, S.; Sternglanz, R., Mapping of the Active Site Tyrosine of Eukaryotic DNA Topoisomerase I. J. Biol. Chem. 1983, 264, 13373-13376. 3. Castelli, S.; Coletta, A.; D'Annessa, I.; Fiorani, P.; Tesauro, C.; Desideri, A., Interaction Between Natural Compounds and Human Topoisomerase I. Biol. Chem. 2012, 393, 1327-1340. 4. Staker, B. L.; Hjerrild, K.; Feese, M. D.; Behnke, C. A.; Burgin, A. B., Jr.; Stewart, L., The Mechanism of Topoisomerase I Poisoning by a Camptothecin Analog. Proc. Natl. Acad. Sci. USA 2002, 99, 15387-15392. 5. Chen, A. Y.; Chen, P. M. T.; Chen, Y.-J., DNA Topoisomerase I Drugs and Radiotherapy for Lung Cancer. J. Thorac. Dis. 2012, 4, 390-397. 6. Bruchim, I.; Jarchowsky-Dolberg, O.; Fishman, A., Advanced (Second) Line Chemotherapy in the Treatment of Patients with Recurrent Epithelial Ovarian Cancer. Eur. J. Obstet., Gynecol. Reprod. Biol. 2013, 166, 94-98. 7. Dezhenkova, L.; Tsvetkov, V.; Shtil, A., Topoisomerase I and II Inhibitors: Chemical Structure, Mechanisms of Action and Role in Cancer Chemotherapy. Russ. Chem. Rev. 2014, 83, 82-94. 8. Craig, S. B., Stability and Compatibility of Topotecan Hydrochloride for Injection with Common Infusion Solutions and Containers. J. Pharm. Biomed. Anal. 1997, 16, 199–205. 9. Hoki, Y.; Fujimori, A.; Yves, P., Differential Cytotoxicity of Clinically Important Camptothecin Derivatives in P-Glycoprotein-Overexpressing Cell Lines. Cancer Chemother. Pharmacol. 1997, 40, 433438. 10. Garcia-Carbonero, R.; Supko, J., Current Perspectives on the Clinical Experience, Pharmacology, and Continued Development of the Camptothecins. Clin. Cancer Res. 2002, 8, 641-661. 11. Riemsma, R.; Simons, J.; Bashir, Z.; Gooch, C.; Kleijnen, J., Systematic Review of Topotecan (Hycamtin) in Relapsed Small Cell Lung Cancer. BMC Cancer 2010, 10, 2-12. 12. Winokur, S., Network For Oncology Communication & Research. Oncologist 1997, 2, 181-195. 13. Sharkey, R.; McBride, W.; Cardillo, T.; Govindan, S.; Wang, Y.; Rossi, E.; Chang, C.-H.; Goldenberg, D., Enhanced Delivery of SN-38 to Human Tumor Xenografts with an Anti-Trop-2-SN-38 Antibody Conjugate (Sacituzumab Govitecan). Clin. Cancer Res. 2015, 21, 5131-5138. 14. Luzzio, M.; Besterman, J.; Emerson, D., Synthesis and Camptothecin Antitumor Activity of Novel Water Soluble Derivatives of as Specific Inhibitors of Topoisomerase I. J. Med. Chem. 1994, 38, 395-401. 15. Pommier, Y.; Leo, E.; Zhang, H.; Marchand, C., DNA Topoisomerases and Their Poisoning by Anticancer and Antibacterial Drugs. Chem. Biol. 2010, 17, 424-433.

34


Page 34 of 55

Page 35 of 55

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


16. Wang, L.; Ting, C.; Lo, C.; Su, J.; Mickley, L.; Fojo, A.; Whang-Peng, J.; Hwane, J., Identification of Mutations at DNA Topoisomerase I Responsible for Camptothecin Resistance. Cancer Res. 1997, 57, 1516-1522. 17. Sugimoto, Y.; Tsukahara, S.; Oh-hara, T., Decreased Expression of DNA Topoisomerase I in Camptothecin-Resistant Tumor Cell Lines as Determined by a Monoclonal Antibody. Cancer Res. 1990, 50, 6925-6930. 18. Pommier, Y., Topoisomerase I Inhibitors: Selectivity and Cellular Resistance. Drug Resist. Update 1997, 2, 307-318. 19. Chrencik, J.; Staker, B.; Burgin, A.; Pourquier, P.; Pommier, Y.; Stewart, L.; Redinbo, M., Mechanisms of Camptothecin Resistance by Human Topoisomerase I Mutations. J. Mol. Biol. 2004, 339, 773-784. 20. Bases, R.; Mendez, F., Topoisomerase Inhibition by Lucanthone, An Adjuvant In Radiation Therapy. Int. J. Radiat. Oncol. Biol. Phys. 1997, 37, 1133-1137. 21. Fan, Y.; Weinstein, J.; Kohn, K., Molecular Modeling Studies of the DNA-Topoisomerase I Ternary Cleavable Complex with Camptothecin. J. Med. Chem. 1998, 41, 2216-2226. 22. Pan, P.; Li, Y.; Yu, H.; Sun, H.; Hou, T., Molecular Principle of Topotecan Resistance by Topoisomerase I Mutations through Molecular Modeling Approaches. J. Chem. Inf. Model. 2013, 53, 997-1006. 23. Sirikantaramas, S.; Meeprasert, A.; Rungrotmongkol, T.; Fuji, H.; Hoshino, T.; Sudo, H.; Yamazaki, M.; Saito, K., Structural Insight of DNA Topoisomerases I From Camptothecin-Producing Plants Revealed by Molecular Dynamics Simulations. Phytochemistry 2015, 113, 50-56. 24. Hou, T.; Wang, J.; Li, Y.; Wang, W., Assessing the Performance of the MM/PBSA and MM/GBSA Methods. 1. The Accuracy of Binding Free Energy Calculations Based on Molecular Dynamics Simulations. J. Chem. Inf. Model. 2011, 51, 69-82. 25. Irwin, J.; Sterling, T.; Mysinger, M.; Bolstad, E.; Coleman, R., ZINC: A Free Tool to Discover Chemistry for Biology. J. Chem. Inf. Model. 2012, 52, 1757-1768. 26. Sastry, G. M.; Adzhigirey, M.; Day, T.; Annabhimoju, R.; Sherman, W., Protein and Ligand Preparation: Parameters, Protocols, and Influence on Virtual Screening Enrichments. J. Comput.-Aided Mol. Des. 2013, 27, 221-234. 27. Redinbo, M.; Stewart, L.; Kuhn, P.; Champoux, J.; Hol, W., Crystal Structures of Human Topoisomerase I in Covalent and Noncovalent Complexes with DNA. Science 1998, 279, 1504-1513. 28. Friesner, R.; Murphy, R.; Repasky, M., Glide: A New Approach for Rapid, Accurate Docking and Scoring. J.Med. Chem. 2004, 47, 1739-1749. 29. Friesner, R.; Murphy, R.; Repasky, M., Extra Precision Glide: Docking and Scoring Incorporating a Model of Hydrophobic Enclosure for Protein-Ligand Complexes. J. Med. Chem. 2006, 49, 6177-6196.

35



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

30. Shivakumar, D.; Williams, J.; Wu, Y.; Damm, W.; Shelley, J.; Sherman, W., Prediction of Absolute Solvation Free Energies Using Molecular Dynamics Free Energy Perturbation and the OPLS Force Field. J. Chem. Theory Comput. 2010, 11, 1509-1519. 31. Nose, S., A Unified Formulation of the Constant Temperature Molecular Dynamics mMethods. J. Chem. Phys. 1984, 81, 511-519. 32. Hoover, W., Canonical Dynamics: Equilibrium Phase-Space Distributions. Phys. Rev. A 1985, 31, 16951697. 33. Martyna, G.; Tobias, D.; Klein, M., Constant Pressure Molecular Dynamics Algorithms. J. Chem. Phys. 1994, 101, 4177-4183. 34. Kräutler, V.; van Gunsteren, W.; Hünenberger, P., A Fast SHAKE Algorithm to Solve Distance Constraint Equations for Small Molecules in Molecular Dynamics Simulations. J. Comput. Chem. 2001, 22, 501-508. 35. Shan, Y.; Klepeis, J.; Eastwood, M.; Dror, R.; Shaw, D., Gaussian Split Ewald: A Fast Ewald Mesh Method for Molecular Simulation. J. Chem. Phys. 2005, 122, 1-13. 36. Tuckerman, M.; Berne, B., Stochastic Molecular Dynamics in Systems with Multiple Time Scales and Memory Friction. J. Chem. Phys. 1991, 95, 4389-4396. 37. Harder, E.; Damm, W.; Maple, J.; Wu, C.; Reboul, M.; Xiang, J. Y.; Wang, L.; Lupyan, D.; Dahlgren, M.; Knight, J.; Kaus, J.; Cerutti, D.; Krilov, G.; Jorgensen, W.; Abel, R.; Friesner, R., OPLS3: A Force Field Providing Broad Coverage of Drug-like Small Molecules and Proteins. J. Chem. Theory Comput. 2016, 12, 281-296. 38. Li, J.; Abel, R.; Zhu, K.; Cao, Y.; Zhao, S.; Friesner, R., The VSGB 2.0 Model: A Next Generation Energy Model for High Resolution Protein Structure Modeling. Proteins 2011, 79, 2794-2812. 39. Kollman, P.; Massova, I.; Reyes, C.; Kuhn, B.; Huo, S.; Chong, L.; Lee, M.; Lee, T.; Duan, Y.; Wang, W.; Donini, O.; Cieplak, P.; Srinivasan, J.; Case, D.; Cheatham, T., Calculating Structures and Free Energies of Complex Molecules: Combining Molecular Mechanics and Continuum Models. Acc. Chem. Res. 2000, 33, 889-897. 40. Hou, T.; Wang, J.; Li, Y.; Wang, W., Assessing the Performance of the Molecular Mechanics/Poisson Boltzmann Surface Area and Molecular Mechanics/Generalized Born Surface Area Methods. II. The Accuracy of Ranking Poses Generated from Docking. J. Comput. Chem. 2010, 32, 866-877. 41. Xu, L.; Sun, H.; Li, Y.; Wang, J.; Hou, T., Assessing the Performance of MM/PBSA and MM/GBSA Methods. 3. The Impact of Force Fields and Ligand Charge Models. J. Phys. Chem. B. 2013, 117, 84088421. 42. Sun, H.; Li, Y.; Tian, S.; Xub, L.; Hou, T., Assessing the Performance of MM/PBSA and MM/GBSA Methods. 4. Accuracies of MM/PBSA and MM/GBSA Methodologies Evaluated by Various Simulation Protocols Using PDBbind Data Set. PCCP 2014, 16, 16719-16729.

36


Page 36 of 55

Page 37 of 55

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


43. Rubin, E.; Pantazis, P.; Toppmeyer, D., Identification of a Mutant Human Topoisomerase I with Intact Catalytic Activity and Resistance to 9-Nitro-Camptothecin. J. Biol. Chem. 1994, 269, 2433-2439. 44. Andoh, T.; Ishii, K.; Suzuki, Y.; Ikegami, Y.; Kusunoki, Y.; Takemoto, Y.; Okada, K., Characterization of a Mammalian Mutant with a Camptothecin-resistant DNA Topoisomerase I. Proc. Natl. Acad. Sci. 1987, 84, 5565-5569. 45. Fujimori, A.; Harker, W. G.; Kohlhagen, G.; Hoki, Y.; Pommier, Y., Mutation at the Catalytic Site of Topoisomerase I in CEM/C2, a Human Leukemia Cell Line Resistant to Camptothecin. Cancer Res. 1995, 55, 1339-1346. 46. Beck, D.; Abdelmalak, M.; Lv, W.; Reddy, P. V. N.; Tender, G.; O'Neill, E.; Agama, K.; Marchand, C.; Pommier, Y.; Cushman, M., Discovery of Ptent Idenoisoquinoline Topoisomerase I Poisons Lacking the 3Nitro Toxicophore. J. Med. Chem. 2015, 58, 3997-4015.

37



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

For Table of Contents Use Only A Computational Study of Anti-Cancer Drug Resistance caused by 10 Topisomerase I Mutations, Including 7 Camptothecin Analogs and Lucanthone

Kelly Mulholland and Chun Wu

38


Page 38 of 55

Page 39 of 55

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48


Table 1. Experimental and predicted binding (free) energy of drugs to wild type Top1.

Ligand

IC50 (nM)

Experimental ∆G* (kcal/mol)

∆MMGBSA (kcal/mol)

Topotecan

50

-10.0

-105.2

LESN38 Camptothecin Lurtotecan

77 300 416

-9.7 -8.9 -8.7

-101.1 -93.1 -83.4

*ΔG𝐵𝑖𝑛𝑑𝑖𝑛𝑔 = RT × ln⁡(IC50 )⁡



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48

Page 40 of 55

Table 2. Experimental and predicted binding free energy change of Camptothecin to Top 1 Mutants

Mutant T729A G503S D533N D533G G717V N722S

Experimental FC* 10 134 220 300 600 974

Experimental ∆∆G (kcal/mol) 1.4 2.9 3.2 3.4 3.8 4.1

*FC: fold change of dissociation equilibrium constant in comparison to wild type


∆∆MMGBSA(kcal/mol) 23.1 27.9 30.9 29.3 36.9 45.0

Page 41 of 55

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48


Table 3. MMGBSA through docking and MMGBSA through MD simulations of topotecan bound to wild type and six mutant complexes Dock Complex 1K4T R364H F361S N722S T729A G503S D533G

ΔG -105.2 -95.4 -89.9 -80.9 -86.3 -85.5 -85.1

MD ΔΔG 0.0 9.8 15.3 24.3 18.9 19.7 20.1

ΔG -75.2 -60.7 -60.0 -59.8 -58.3 -54.7 -53.6

ΔΔG 0.0 14.5 15.3 15.4 16.9 20.6 21.6



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48

Page 42 of 55

Table 4. Change in binding energy for each Topoisomerase poison (ΔΔG kcal/mol)

Topotecan

Camptothecin

Lucanthone

Lurtotecan

LESN38

Gimatecan

Exatecan

Belotecan

D533N (26.7)

F361S (38.8)

F361S (25.5)

F361S (33.5)

D533G (64.7)

D533N (50.4)

R364H (53.4)

G503S (49.7)

N722S/A (24.3)

D533G (29.3)

N722S/A (24.4)

R364H (32.9)

F361S (30.8)

D533G (50.1)

N722S/A (41.0)

D533N (47.8)

D533G (20.1)

N722S/A (21.8)

D533N (23.9)

G503S (32.5)

N722S/A (27.9)

T729A (33.7)

D533N (39.9)

G363C (45.8)

G503S (19.8)

G717V (16.9)

D533G (19.8)

N722S/A (31.1)

D533N (25.8)

G363C (32.3)

G717V (37.1)

T729A (44.4)

T729A (18.8)

D533N (16.7)

G363C (18.9)

D533N (28.3)

G503S (24.1)

N722S/A (32.1)

G503S (34.9)

D533G (43.8)

G363C (15.4)

G363C (15.5)

G503S (15.8)

D533G (25.1)

G717V (23.8)

G717V (31.1)

G363C (33.6)

F361S (31.3)

F361S (15.3)

G503S (11.1)

T729A (15.2)

G717V (18.3)

T729A (23.7)

G503S (21.8)

F361S (18.8)

N722S/A (31.1)

G717V (13.9)

T729A (10.7)

R364H (13.2)

G363C (14.6)

G363C (20.5)

R364H (18.1)

T729A (17.1)

F361S (31.0)

R364H (9.8)

R364H (9.1)

G717V (10.3)

T729A (12.2)

R364H (15.5)

F361S (13.8)

D533G (10.7)

R364H (27.9)


Page 43 of 55

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48


A) Camptothecin and Analogs

B) Lucanthone

C) Corresponding R Group in A

Inhibitors Topotecan Camptothecin LESN-38 Belotecan

R1 -H -H -H -H

Gimatecan Exatecan Lurtotecan

-H -H -F -CH3 c

a=

R2 -OH -H -OH -H

b=

R3 R4 -CH2NH(CH3)2 -H -H -H -H -CH2CH3 -H CHCH2NH2(CH3)2 -H a b -H d c=

d=

Figure 1. Topoisomerase I inhibitors. Camptothecin scaffold (A), non-Camptothecin drug Lucanthone (B), and Camptothecin derivatives with corresponding R groups in A (C).



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48

Figure 2. Schematic diagram of mutated residue locations


Page 44 of 55

Page 45 of 55

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48


A) Side View

B) Top View

Figure 3. Crystal complex structure of Top 1 with Topotecan in red (PDB ID: 1K4T). The mutating residues are shown in blue, and the key catalytical residue, Y723, is shown in green.



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48

Figure 4. Clustering of 8 Topoisomerase I inhibitors to WT and 10 Top 1 mutant complexes. For clarity, only WT Top 1 is shown in cartoon. For the mutating residues, side chains from WT and mutant are shown in different colors. For drugs, Oxygen atoms are displayd in red, nitrogen atoms are displayed in blue and carbon atoms are displayed in gray.


Page 46 of 55

Page 47 of 55

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48


A)

B)

Figure 5. Correlation Plots. A) Correlation Between Experimental and Predicted Change in Binding Free Energy of Four drugs to WT Top 1 (Table 1) B) Correlation Between Experimental and Predicted Binding Free Energy Change of Camptothecin to six Top 1 Mutants (Table 2).



25.0 ΔΔG Docking (kcal/mol)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48

20.0 15.0

10.0 5.0 R² = 0.7355 0.0 0.0

5.0

10.0

15.0

20.0

25.0

30.0

ΔΔG MDS (kcal/mol) Figure 6. A Correlation Between ΔΔG Through Docking and ΔΔG Through MD Simulations of topotecan bound to wild type and six mutant complexes (Table 3).


Page 48 of 55

Page 49 of 55

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48


Figure 7. Mean change in binding energy (∆∆G) for each Topoisomerase I inhibitor. A) Topotecan B) Camptothecin C) Lucanthone D) Lurtotecan E) LESN38 F) Gimatecan G) Exatecan H) Belotecan



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48

Figure 8. Mean binding energy change (∆∆G) for each mutant and center-to-center distance from mutated residue to drug. 1) D533G 2) D533N 3) N722A 4) N722S 5) G363C 6) G503S 7) F361S 8) G717V 9) R364H 10) T729A


Page 50 of 55

Page 51 of 55

A) Decomposition of ∆∆G

∆∆G (kcal/mol)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48


B) Wildtype

C) D533G

Figure 9. LESN38-D533G mutated complex. The decomposition of the ∆∆G (A), interacting residues with wildtype complex (B), and interacting residues with mutant complex (C).



A) Decomposition of ∆∆G

∆∆G (kcal/mol)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48

B) Wildtype

C) G717V

Figure 10. Lucanthone-G717V mutated complex. The decomposition of the ∆∆G (A), interacting residues with wildtype complex (B), and interacting residues with mutant complex (C).


Page 52 of 55

Page 53 of 55

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48


B)

A)

Figure 11. A 3D representation of Topoisomerase I enzyme: Topotecan is shown in red, blue indicates core domain, yellow indicates N-terminal Domain, cyan indicates Linker, and green indicates C-terminal Domain (A) and protein RMSF of Topotecan bound to each mutated complex (B)



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48

A)

B)

Figure 12. A 2D representation of Topotecan atom numbers (A) and Ligand RMSF of Topotecan bound to each mutated complex (Atom Index Number corresponds directly to the atom numbers in A) (B)


Page 54 of 55

Page 55 of 55

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48


A)

B)

C)

Figure 13. Structural comparison of Top1 (A), DNA (B), and Topotecan (C) of wild type and D533G complex as a result of MD simulations. Wild type complex is shown in red and D533N complex in blue. Topotecan bound to wild type is shown in yellow and in green when bound to D533N. (wild type compared to all mutant complexes can be seen in Figure S86)


Computational Study of Anticancer Drug Resistance Caused by 10

Recommend Documents