Article Cite This: Biochemistry XXXX, XXX, XXX−XXX
pubs.acs.org/biochemistry
Activity of Topotecan toward the DNA/Topoisomerase I Complex: A Theoretical Rationalization Semiha Kevser Bali,† Antoine Marion,*,‡ Ilke Ugur,§ Ayse Kumru Dikmenli,∥ Saron Catak,† and Viktorya Aviyente*,† †
Department Department § Department ∥ Department ‡
of of of of
Chemistry, Faculty of Arts and Sciences, Boğaziçi University, Bebek, 34342 Istanbul, Turkey Chemistry, Middle East Technical University, 06800 Ankara, Turkey Lifesciences, Technical University of Munich, 80333 Munich, Germany Chemistry and Chemical Biology, MacMaster University, Hamilton, Ontario L8S4L8, Canada
S Supporting Information *
ABSTRACT: Topotecan (TPT) is a nontoxic anticancer drug characterized by a pH-dependent lactone/carboxyl equilibrium. TPT acts on the covalently bonded DNA/topoisomerase I (DNA/TopoI) complex by intercalating between two DNA bases at the active site. This turns TopoI into a DNAdamaging agent and inhibits supercoil relaxation. Although only the lactone form of the drug is active and effectively inhibits TopoI, both forms have been co-crystallized at the same location within the DNA/TopoI complex. To gain further insights into the pH-dependent activity of TPT, the differences between two TPT:DNA/TopoI complexes presenting either the lactone (acidic pH) or the carboxyl (basic pH) form of TPT were studied by means of molecular dynamic simulations, quantum mechanical/molecular mechanical calculations, and topological analysis. We identified two specific amino acids that have a direct relationship with the activity of the drug, i.e., lysine 532 (K532) and asparagine 722 (N722). K532 forms a stable hydrogen bond bridge between TPT and DNA only when the drug is in its active lactone form. The presence of the active drug triggers the formation of an additional stable interaction between DNA and protein residues, where N722 acts as a bridge between the two fragments, thus increasing the binding affinity of DNA for TopoI and further slowing the release of DNA. Overall, our results provide a clear understanding of the activity of the TPT-like class of molecules and can help in the future design of new anticancer drugs targeting topoisomerase enzymes.
T
in a phosphotyrosine intermediate (so-called cleavable complex), and rotates the intact strand around the nicked one.6 After relaxation, the 5′-hydroxyl group on the nicked strand again attacks the 3′-phosphotyrosine, and the backbone integrity is recovered.7 Under normal conditions, the last esterification step is faster than the cleavage of the strand; however, during this step, the binary complex is also vulnerable to inhibitors.6,8 An inhibitor that prevents the attack of 5′-OH stabilizes the intermediate complex by intercalating within the DNA/TopoI complex; this results in an accumulation of the reaction intermediate leading to double-stranded DNA breaks both in vivo and in vitro.6,9−11 The interaction of the inhibitor with the protein and the flanking base pairs at the cleavage site helps to stabilize the cleavable complex. Furthermore, these interactions increase the likelihood of DNA breaks due to collisions with a replication fork during the S phase, which
opoisomerases are enzymes that play an important role in maintaining the topology of DNA by relaxing supercoils formed during replication and transcription.1 Supercoils are topological distortions that can affect the essential functions of DNA such as gene expression.2 The relaxation process differs depending on the type of topoisomerase. Type I topoisomerases covalently bind to DNA by creating a nick on only one strand, and type II topoisomerases break both strands of DNA.1 As depicted in Figure 1, human topoisomerase I (TopoI) is a monomeric protein that is composed of four domains: N-terminal domain (residues 1−214), core domain (residues 215−635), linker domain (residues 636−712), and C-terminal domain (residues 713−765).3,4 The core domain is further divided into three subdomains: subdomain I (residues 215−232 and 320−433), subdomain II (residues 233−319), and subdomain III (residues 434−635).3−5 Subdomains I and III are called “lip domains” and act as a clamp around the DNA.4 TopoI operates on DNA via a controlled rotation mechanism. The enzyme first introduces a nick on a single strand of DNA via a catalytic tyrosine residue (Y723), resulting © XXXX American Chemical Society
Received: December 27, 2017 Revised: January 23, 2018 Published: February 7, 2018 A
DOI: 10.1021/acs.biochem.7b01297 Biochemistry XXXX, XXX, XXX−XXX
Article
Biochemistry
GlaxoSmithKline) and irinotecan (Campostar, Pfizer), are more water-soluble and are currently being used to treat ovarian and small-cell lung cancers.20−22 Topotecan {TPT, 9[(dimethylamino)methyl]-10-hydroxy-camptothecin} was shown to be more stable than CPT and exhibits a lower toxicity.23,24 It was shown that the presence of TPT increases the half-life of the DNA/TopoI cleavable complex more than CPT does.25 As shown in Figure 2, TPT presents a
Figure 2. Lactone (TPT-closed) and carboxyl (TPT-open) equilibrium of TPT.
hydrolyzable lactone group at its E-ring extremity, resulting in a pH-dependent equilibrium between the lactone form (TPTclosed) and the carboxyl form (TPT-open). 26,27 The equilibrium between two forms is found at pH 6.28;28,29 the lactone form prevails at low acidic pH, while basic conditions favor the carboxyl form.29,30 Only the former, TPT-closed, is found to be active as an inhibitor of TopoI.18,31 In 2002, Staker et al.6 resolved the first crystal structure of human TopoI covalently bound to a 20 bp DNA and to one TPT molecule. Despite the fact that the lactone form is the active drug, the crystal structure revealed that both lactone and carboxyl forms can intercalate at the same site, as represented in the top and bottom panels of Figure 1, respectively.32 Mutation studies performed with CPT provided valuable information about residues that play a role in drug binding as well as in drug resistance. Among those, the following mutations were shown to cause CPT resistance: G363C,33,34 R364H,35,36 E418K,37 G503S,38,39 D533N,40 and D533G.41 Additionally, the T718A mutation mimicked the effect caused by CPT.34 T729A42 and N722S43 mutations resulted in adjustments in the binding site for the drug. Computational studies44,45 with anticancer drug TPT showed that many of these residues play an important role in the effect caused by TPT, as well. Staker el al.6 provided important insights regarding the interactions of TPT with surrounding residues, such as Asp533 and Asn722. A computational study, in which the binary complex (DNA/TopoI) and the TPT-bound ternary complex (TPT:DNA/TopoI) were compared, showed that TPT keeps the catalytic Lys532 from the DNA, preventing the relegation step.46 The conformational space of the linker domain becomes smaller upon binding of TPT due to the reduced flexibility of residues 633−643.46,47 Additionally, it was found that the lactone form of TPT forms direct stable
Figure 1. TPT:DNA/TopoI complex and representative snapshots from trajectories from this study. The domains of TopoI are shown with different colors: red for the N-terminal domain, cyan for the core domain, pink for the linker domain, and violet for the C-terminal domain. The top panel shows TPT-closed, the center panel the TPT:DNA/TopoI complex, and the bottom panel TPT-open.
ultimately leads to cell death.12−14 For the cleavable complex to encounter DNA polymerase or RNA polymerase enzymes, the inhibitor has to stay bound long enough and should not dissociate prematurely.6,12,15 Camptothecin (CPT) is a cytotoxic drug extracted from the bark of a tree, Camptotheca acuminata (family Nyssaceae),11,16 that inhibits the activity of the TopoI enzyme.17,18 CPT is found to be an anticancer agent; however, because of its low water solubility and adverse drug reaction,19 it is not suitable for clinical use. Derivatives of CPT, topotecan (Hycamine, B
DOI: 10.1021/acs.biochem.7b01297 Biochemistry XXXX, XXX, XXX−XXX
Article
Biochemistry
procedure.57,58 Parameters from GAFF, ff99bsc0, and ff14SB were used to model the ligands, the DNA, and the protein, respectively.59−62 Hydrogen atoms and atomic charges were automatically added using the tleap module of AmberTools, accordingly with the standard residue definitions in ff99bsc0 and ff14SB for DNA and protein residues, respectively. Ionizable residues in the protein were modeled in their standard protonation states, and the backbones of DNA terminal bases were set in their hydrolyzed form. Both complexes were equilibrated in a stepwise fashion from 0 to 300 K using a Langevin thermostat (collision frequency of 4.0 ps−1).63 First, the systems were heated to 20 K within 160 ps while all atoms were constrained (3.0 kcal mol−1 Å−2) with a constant volume, and then the constraint was applied to only backbone atoms (3.0 kcal mol−1 Å−2) as the temperature was gradually increased to 300 K within 1350 ps. The SHAKE algorithm and an integration time step of 1 fs were used during all simulations. The final equilibration was performed in the NPT ensemble using a Langevin thermostat and a Berendsen barostat (with default AMBER settings) for 500 ps at 300 K.63,64 Production runs were performed in the NPT ensemble using AMBER14 software with the CUDA-enabled GPUs version of pmemd.63−67 Coordinates were printed every 500 ps, and both simulations were performed until a total of 450 ns was reached. In addition, a second 300 ns simulation produced both systems (TPT-closed-2 and TPT-open-2) by setting different initial velocities after the systems had been heated. These parallel simulations were performed to assess the convergence of our results. As our conclusions for structural descriptors were found to be consistent for both simulations, computationally expensive interaction energy calculations were conducted for only one set of trajectories. Distance, hydrogen bond analysis, and root-mean-square deviation (RMSD) calculations were performed using the cpptraj module of AmberTools15.68 Default distance and angle cutoff values were used in hydrogen bond analysis. RMSDs of both complexes were calculated with respect to the crystal structure, and only backbone atoms were considered. Because of the fluctuations in RMSD during the first 30 ns, this part of the trajectories was systematically considered as equilibration time and neglected in all of the following analyses. QM/MM Interaction Energy Calculations. The energies of interaction between the drug and the DNA/TopoI complex were calculated using the hybrid QM/MM method. For this purpose, the system was separated into two regions: the QM region, which includes only the TPT molecule, and the surrounding DNA/TopoI complex described at the MM level. All interaction energy calculations were performed after all water molecules and Na+ ions had been removed. The QM site was treated at the M062X/6-31G(d) level of theory through the Amber/Gaussian09 interface as implemented in AmberTools15.58,69,70 The energy calculations for the MM part were performed using the same parameters as described in the previous subsection. Coulomb interactions between the QM and MM parts were accounted for by explicitly including the MM point charges in the QM calculations, while QM/MM van der Waals interactions were treated classically within the MM calculation. A total of 420 snapshots were extracted from MD simulations for each system (one every nanosecond). The nonrelaxed interaction energy (ΔE) was calculated as follows:
hydrogen bonds with Arg364 through its B and E rings.46 In another study, it has been shown that the intercalation site is stabilized by a hydrogen bonding network between DNA bases and the amino acid residues around the active site of the binary complex.8 Although the carboxyl form of TPT (TPT-open) slightly predominates at physiological pH, it has
