HM Medical Clinic

 

Type of article: original

Molecular and Biochemical Diagnosis (MBD)
Vol 1, No 2, 2014
Original Article
In Silico Studies on Fingolimod and Cladribine Binding to p53
Gene and Its Implication in Prediction of Their Carcinogenicity
Potential
Karim Mahnam1, Azadeh Hoghoughi1
1. Biology Department, Faculty of Science, Shahrekord University, Shahrekord, Iran Abstract
Background: New drugs namely; cladribine and fingolimodare known to be effective in
treatment of multiple sclerosis (MS). The interaction of these drugs with the promoter region
of the p53 gene may alter p53role in cancer progression. The aim of this study was to known
the interaction of these compounds with p53 gene.
Methods: Binding free energy of the cladribine, fingolimod and their modified drugs for the
p53 gene promoter were investigated using docking, 100 ns molecular dynamics simulations
and MM/PBSA calculation.
Results: The results showed that both cladribine and modified cladribine (replacing -OH on
carbon 3´ ribose sugar with -CH3 group) can bind the minor groove of p53 promoter, and
inhibit the binding of transcription factors and expressio n of p53. However, fingolimodand its
derivatives showed relatively weaker interaction with p53 promoter
Conclusions: Based on in silico studies we showed that the binding of cladribine to the p53
gene is stronger than that of fingolimod, hence it seems tha t the former drug can pose
potential carcinogenic effects. The binding power and carcinogenic effect of sm-fingolimod
(removing four carbons from its aliphatic tail) is more than that of fm-fingolimod (removing
one carbon from its aliphatic tail).
Keywords: Cladribine, Fingolimod, Molecular dynamics simulation, MM/PBSA, p53 gene
deficits, followed by progressive neurological Multiple sclerosis (MS) is a demyelinating deteriorations (Navikas et al., 1996). inflammatory disorder of the central nervous The first oral disease-modifying drug approved system (CNS) with autoimmune responses. food and drug administration (FDA) is The degree of axonal destruction is variable Fingolimod (Gilenya, Novartis) (Fig. 1A) to (Calabresi, 2004). The route of MS is highly postpone progression of physical disability in varied and unpredictable so that it may be sphingosine kinase to the active metabolite; fingolimod phosphate, which in turn blocks
*C orre sponding author. Karim Mahnam, PhD.
migration of lymphocytes from lymphnodes, Biology Department, Faculty of Science, Shahrekord University, P.O. Box: 8818634141,Shahrekord, Iran thereby reducing the number of lymphocytes T el., +98-038-4424402; Fax, +98-38-4424419 Email: [email protected] Molecular and Biochemical Diagnosis (MBD). Vol.1, No.2 (2014), 105-122 K. Mahnam et al.
in peripheral blood (Cohen et al., 2007). The incorporated into DNA, thereby causes down- possible mechanism of the therapeutic effect of regulation of cellular ribonucleotidereductase fingolimod in MS is through the reduction of and inhibit DNA synthesis (Foley et al., 2004). TATA element of the promoter is recognized (Francesca, 2007). by TATA binding protein (TBP). Foley et al showed that positions in the TATA sequence are most severely affected by cladribine incorporation (Foley et al., 2004). In general, drug targets are cytoplasmic proteins, membrane receptors or membrane- bound proteins, nuclear proteins, DNA etc. Small aromatic compounds can bind DNA by A: Covalent bond; through their functional
groups irreversibly attached to DNA, leading to inhibition of DNA synthesis processes and Figure 1. Structure of fingolimod (A) and cladribine (B).
cell death such as Cisplatin and Mitomycin Fingolimod has been associated with reduce (Elizondo-Riojas et al., 2001). heart rate (bradycardia) and usually fatal B: Non-covalent bond; by intercalation (such
infections such as cancer (Cohen et al., 2007). Another drug used to trea intominor groove binding such as; Netropsin, (HCL, leukemic reticuloendotheliosis) and Distamycin and into major groove binding; is cladribine (Leustatin, Litak and Movectro™) such as Norfloxacin (Neidle et al., 1987). The tumor suppressor p53 gene as an important tumor suppressor gene continually is transcribed to prevent cancer. P53 gene is the ibine). It is a and acts as most frequently mutated gene in human tumors suppressor of the immune system. Possible (Vogelstein et al., 2010). In some cancers, side effects of the cladribine include fever, transcription of the p53 gene is reduced (Bai et infection, anemia and cancer. CldAdo is taken al., 2006). up by cells, converted to 2-chloro-2´-deoxy The Molecular Mechanics/Poisson–Boltzmann Surface Area (MM-PBSA) method has been Molecular and Biochemical Diagnosis (MBD). Vol.1, No.2 (2014), 105-122 In silico fingolimod and cladribine binding to p53 gene
used to calculate relative free energies of DAPI sequences of DNA (Spacková et al., 2003). Currently the computational techniques are Promoter of p53 gene has 52 pair nucleotides. The widely applied in chemistry and biology sequence of the 5´ to 3´ strand of promoter of p53 ranging from the quantum mechanics of gene that was applied for this study was 5´- molecules to the dynamics of large complex molecular aggregates. Molecular interactions steer chemical reactions, phase transitions and (Reisman et al., 1993). 3D structure of p53 other physical phenomena and can be studied promoter was generated via 3D-Dart (3DNA- via molecular dynamics (MD) simulations, Driven DNA Analysis and Rebuilding Tools) showing the detailed motion of molecules or (haddock.science.uu.nl/ atoms as a function of time. The MD services/3DDART). Also, geometries of all simulations provide powerful links between ligands were obtained from Arguslab software the model equilibrium, minimal geometries of proteins and DNA and binding free energy of Lab.html) via molecular mechanics methods drugs (Karplus et al., 2005). The calculation of under MM+ force fields and used for docking relative binding free energies of ligands to a and MD simulation studies. The atomic receptor has been used for better understanding charges of all ligands were calculated with the of molecular interactions of proteins with Merz−Kol man electrostatic potential fitting procedure in the Gaussian quantum chemistry et al., 2005). package (Frisch et al., 1998). This was In our ongoing project, we have performed performed by means of a Hartee-Fock wave some theoretical studies to investigate the function obtained in a 6-31G* basis set for mechanism of binding of cladribine and compatibility with the partial charges from the fingolimodto promoter of p53 gene. In AMBER force field that was used for p53 addition, the effect of some structural modifications of these drugs in binding their free energy to promoter of p53 gene has been calculation was done using this command: HF/6-31G* Pop=MK IOp (6/33=2, 6/41=10, 6/42=17) et al., 2011). Cladribine was underlying carcinogenicity of cladribine or modified by replacing OH on carbon 3´ ribose Molecular and Biochemical Diagnosis (MBD). Vol.1, No.2 (2014), 105-122 K. Mahnam et al.
sugar with CH3 group. Modification of fingolimod was done by removing one carbon 2. Molecular dynamic simulations
fingolimod) from the aliphatic hydrocarbon Five molecular dynamics simulation of ligands tails. All images were generated with complexes with p53 promoter sequence were performed. The cycle time for each simulation was 20 ns. Then, one hundred ns MD Theoretical studies were done in three simulations were applied. MD simulation and following sections: molecular mechanic (MM) minimization were performed using GROMACS 4.5.3 package 1. Docking
under Amber99 force fields (Van der Spoel et Autodock 4 software was used for docking al., 2005; Berendsen et al., 1995; Hess et al., studies (Morris et al., 1998). The grid box size 2008 and Lindahl et al., 2001). Topologies of was set at 9090118 Å and spacing between ligands were generated by acpype/Antechamber grid points 0.375 angstrom. The p53 promoter based on a General Amber Force Field (GAFF) structures were fixed during docking, while the (Sousa et al., 2012). MD simulations were drugs were flexible. Grid searching was carried out in an NPT ensemble with periodic performed by a local search genetic algorithm boundary conditions. Van der Waals forces (LGA) to locate the ligands in the lowest were treated using a cut-off of 12 Å. The binding energy. Routine procedures and electrostatic interactions were calculated using default parameters were used in the docking the Particle-Mesh Ewald model with a 14 Å except dstep, tstep and qstep that were cut-off (Darden et al., 1993).The complexes considered 0.5 Å, 0.5°, 5° respectively were solvated by a layer of water of at least 12 (Majumdar et al., 2011). Å in all directions. The frequency to update the All ligands (cladribine, modified cladribine, neighbor list was 10 ps. MD simulation was accomplished in four steps for each system. In fingolimod) were docked on p53 promoter. the first step, the entire system was minimized Two hundred docking runs were performed for using the steepest descent followed by each docking. The best pose with the lowest conjugate gradient algorithms. In the second binding energy and the most populated step, the solvent and Na+ ions were allowed to conformation in each cluster was chosen as the evolve using minimization and molecular initial structure in the molecular dynamics dynamics in the NVT ensemble for 500 ps and Molecular and Biochemical Diagnosis (MBD). Vol.1, No.2 (2014), 105-122 In silico fingolimod and cladribine binding to p53 gene
in the NPT ensemble for 1000 ps at 100 K, energy of a DNA molecule to a ligand where the initial configuration of the structures molecule in a solution can be defined as: was kept fixed. In the third step, in order to ∆Gbinding=Gcomplex-(GDNA+Gligand) Eq.1 obtain equilibrium geometry at 300 K and 1 "A MD simulation is performed to generate a atm, the system was heated at a weak thermodynamically weighted ensemble of temperature coupling (τ = 0.1 ps) and pressure structures" (Kumari et al., 2014). The free coupling (τ = 0.5 ps). The Berendsenalgorithm energy term is calculated as an average over was chosen for thermostat and barostat in the considered structures: equilibration phase (Berendsen et al., 1984). To constrain the lengths of hydrogen-containing Total molecular mechanical energies EMM is bonds, the LINCS algorithm was used (Hess et calculated by using GROMACS utility with al., 1997). The temperature of the system was the AMBER99 force field. -T<SMM> is the then increased from 100 K to 300 K and the solute entropic contribution. Gsolvation represents velocities at each step were re-accredited the free energy of solvation and consists of two Maxwell-Boltzmann parts: Gpolar or GPB and nonpolar contributions, distribution at that temperature and equilibrated Gnonpolar. GPB is generated from the electrostatic for 200 ps. In the final (production) step, 20 ns MD simulations at 300 K with a time step of 2 (Massova et al., 1999). fs was performed for each complex and final In the current study, Gpolar was calculated using structures were obtained. The thermostat and APBS (Adaptive Poisson-Boltzmann barostat for production step were Nosé-Hoover Solver program) method (Baker et al., 2001) via the non-linearized Poisson Boltzmann (Berendsen et al., 1984). In all simulations, two equation. The non-polar contribution, Gnonpolar single strands of DNA were constrained to each was considered to be proportional to the other (Cheatham et al, 1998). Potential and solvent accessible surface area (SASA). kinetic energies and temperature at the last 5 ns In the MM/PBSA approximation and for were calculated using g_energy command of estimating Gfree-DNA and Gfree-ligand, snapshots Gromacs package. Other analyses were collected from the MD run for the DNA-ligand performed by using Gromacs package. complex were used. After equilibration, snapshots of complex, DNA and ligand 3. MM/PBSA calculation
(without water molecules) were taken every 50 As indicated by Kumari, the binding free ps for calculating the enthalpy. Molecular and Biochemical Diagnosis (MBD). Vol.1, No.2 ( 2014), 105-122 K. Mahnam et al.
Binding free energy calculations based on the have suggested that including corrections for MM/PBSA approach can be performed either changes in the configurational free energy of the according to the three trajectories method system lead to only a small improvement in the (TTM) or according to the single trajectory total. We decided to neglect the entropic term in method (STM). In our work, MM/PBSA our calculations. The last 5 nanosecond of the calculations were performed according to the MD simulations was considered for MM/PBSA STM protocol. A single trajectory run for the complex is required for this method, whereby The energy components EMM, Gpolar and Gnon- both the DNA and ligand structures are polar of each complex were calculated for 100 extracted directly from the complex structure snapshots extracted every 50 ps from the (Huo et al., 2002), thus zeroing out the Eint production trajectories at the last 5 ns. To term. In this case, the DNA and the ligands are calculate Gpolar, a box was generated using the assumed to behave similarly in the bound and extremes coordinates of the molecular complex in the free forms. in each dimension. A coarse-grid box (cfac =3) In the MM/PBSA approximation, EMM+Gsolv was obtained when the box expanded in each account for the enthalpy change is associated dimension by two-fold. A finer grid-box is with complex formation. The computational then placed within the coarse grid-box determination of binding free energies requires extending 50 Å (fadd=50) from the complex's the calculation of the entropic contributions to extremes coordinates in each direction. An complex formation including conformational ionic strength of 0.6 M NaCl with radii of 0.95 changes in the rotational, translational and and 1.81 Å, respectively for sodium and vibrational degrees of freedom of the solute. chloride ions was used during all Gpolar The MM/PBSA method was used by calculations. The values for vacuum (vdie) and g_mmpbsa command (Baker et al., 2001; Pronk solvent (sdie) dielectric constants were taken et al., 2013; Eisenhaber et al., 1995 and Kumari as 1 and 80 respectively. The solute (pdie) et al., 2014). In this module, entropic terms are dielectric constant was assigned a value of not included and therefore it is unable to give eight. Subsequently, the binding free energy of the absolute binding energy. Thus, it is proper each snapshot was calculated for each complex to calculate the relative binding energies for using a combination of Eq.1 and 2 without instance, to compare different ligands binds to entropic contributions in the binding energy the same receptor. In addition, the net entropic (Kumari et al., 2014 and Brown et al., 2009 contribution is often small, and multiple studies and Gohlke et al., 2004 and Kar et al., 2011 Molecular and Biochemical Diagnosis (MBD). Vol.1, No.2 (Autumn 2014), 105-122 In silico fingolimod and cladribine binding to p53 gene
and Bradshaw et al., 2011). p53 sequence are negative; so these drugs are able to bind the p53 promoter. Also, binding Results and Discussion
position of these ligands were mentioned. The 1. Docking
positions of all compounds were in the minor Investigation of the docking results in Table 1 groove of p53 promoter. The binding position shows that the binding free energy of cladribine, of cladribine and modified cladribine are 5´- fingolimod, modified cladribine (replacing OH T15T16G17-3´ nucleotide; and those of on carbon 3´ ribose sugar with CH3 group), the first and second modified fingolimod (removing one carbon or four carbons from the aliphatic modified fingolimod) to p53 promoter are 5´- hydrocarbon tail of fingolimod respectively) to G30T31T32T33T34-3´ nucleotides. Table 1.Van der Waals (VDW)contribution, Electrostatic contribution (Elec) and the lowest binding free energy
(L.B) of native and modified cladribine, fingolimod to p53 promoter, nucleotides15-34 are shown.
Compound
VDW + Hbond + desolvation
S equence of binding position
(kcal/mol)
(kcal/mol)
Cladribine
5´- T15T16G17-3´ Modified cladribine
5´-T15T16G17-3´ Fingolimod
5´-G30T31T32T33T34-3´ 5´-G30T31T32T33T34-3´ S m-fingolimod2
5´-G30T31T32T33T34-3´ 1. First modification of fingolimod (i.e. deleting one carbon of fingolimod tail). 2. Second modification of fingolimod (i.e. deleting four carbon of fingolimod tail). These sequences are the positions of binding of fingolimod are weaker to bind p53 promoter. transcription factors such as USF (upstream In all cases, Van der Waals (plus Hbond and stimulatory factor) or TFE3 (transcription desolvation) contributions are more negative factor E3) (Kim et al., 2008; Yasumoto et al., important than electrostatics 1994). Binding free energy of modified interactions (Table 1). cladribine to p53 promoter is lower than that for cladribine, it means that the binding of 2. Molecular dynamics simulation
modified cladribine is stronger than that for Table 2 shows the results of average potential cladribine but binding free energy of the first and kinetic energies, temperature, root mean and second modified fingolimod to p53 square deviation (RMSD) of p53 promoter and promoter are more than that for fingolimod, it ligands RMSD relative to initial positions means that the first and the second modified during the last 5 ns of 20 ns MD simulation. Molecular and Biochemical Diagnosis (MBD). Vol.1, No.2 (2014), 105-122 K. Mahnam et al.
There are small variations in potential and sufficient and stable under the simulation kinetic energy, temperature and RMSD of the conditions and thermal equilibrium of the p53 promoter during the last 5 ns of MD systems. By investigating the final structures simulation with a very low ratio of the total of 20 ns MD simulation it appeared that the energy drift to the average total energy (Table two strands of the p53 promoter remained 3). This shows that the simulations were together during 20 ns simulations. Table 2. The potential energy (P), kinetic energy (K) and temperature (T) and radius of gyration (Rg) and
RMSD of p53 promoter and drugs at complex during the last 5 ns of MD simulations.
RMS D of p53
Promoter at
promoter at
P (kcal/mol)
K (kcal/mol)
Cladribine
Modified cladribine
Fingolimod
S m-fingolimod2
1. Fm-Fingolimod:First modification of fingolimod (i.e. deleting one carbon of fingolimod tail); 2. Sm-Fingolimod: Second modification of fingolimod (i.e. deleting four carbon of fingolimod tail). *. Nanometer Table 3: The ratio of the total energy drift to average of total energy during 20 ns MD simulations of all species.
Ratio of the total energy drift to
S ystem name
average of total energy (10-5)
Cladribine
Modified cladribine
Fingolimod
S m-fingolimod2
1. Fm-Fingolimod: First modification of fingolimod (i.e. deleting one carbon of fingolimod tail) 2. Sm-Fingolimod: Second modification of fingolimod (i.e. deleting four carbon of fingolimod tail). Also, small RMSDs of ligand atoms during neighbors if the backbone RMSD between simulation relative to the starting position them was less than 0.2 nm. (Table 2) showed that the ligands reach to The middle structure of the most populated stable positions. structures obtained from clustering of trajectories To determine the relative populations of all during the last 5 ns of MD simulation showed conformations, the trajectories were clustered that cladribine and modified cladribine stay in using g_cluster command of the Gromacs the minor groove of p53 promoter in 5´- package. Two conformations were considered T16G17A18-3´ sequence however, fingolimod, Molecular and Biochemical Diagnosis (MBD). Vol.1, No.2 (2014), 105-122

In silico fingolimod and cladribine binding to p53 gene
the first and second modified fingolimod go away from their initial docking positions (Fig. 2).
Figure2.The middle structure of the most populated structures of drugs -DNA complex during the last 5 ns MD
simulation.The number of the nucleotides in double strandedp53 promoter was mentioned in Table 4. fm-
fingolimod: First modification of fingolimod,i.e. deleting one carbon of fingolimod tail. sm-fingolimod: Second
modification of fingolimod, i.e. deleting four carbon of fingolimod tail.
The number of nucleotides in double-stranded
fingolimod (belongs to 17.1 ns) and second p53 promoter has been indicated in Table 4. In modified fingolimod (belongs to 16.98 ns), no the middle structure of the most populated hydrogen bonds seen with p53 promoter. structures of cladribine (belongs to 19.6 ns) The average solvent accessible surface area and modified cladribine (belongs to 18.68 ns) (SASA) of the ligand atoms during the 20 ns MD in complex with p53 promoter, guanosine 17 simulation were calculated by g_sas command (H22 and N3 and O4´ atoms) and adenosine 89 and non-hydrogen atoms with SASA less than 10 (N3 atom) of double stranded p53 promoter, Å2 were determined. These atoms probably bind have hydrogen bonds with cladribine. In to the p53 promoter during MD simulation. The middle structure of the most populated results showed that cladribine bind the p53 structures of MD simulation of fingolimod promoter via its N2, N4, O1, C3, C2 and N3 atoms (these atoms were shown in Fig. 1A). Molecular and Biochemical Diagnosis (MBD). Vol.1, No.2 (2014), 105-122 K. Mahnam et al.
Table 4. The frequency of nucleotides in double-stranded p53 promoter.
Nucleotide Number
DNA strand direction:5´
DNA strand direction: 3´
Nucleotide Number
DNA strand direction: 3´
DNA strand direction: 5´
Molecular and Biochemical Diagnosis (MBD). Vol.1, No.2 (2014), 105-122 In silico fingolimod and cladribine binding to p53 gene
Modified cladribine bind the p53 promoter via Minimum distance between p53 promoter and its N2, N4, O1, C3, C7, C2, N3 and C8 atoms ligands and the number of contacts less than 0.6 (Fig. 1A). In fingolimod and first modification nm between p53 promoter and ligands during only, three atoms (i.e. C16, C1 and C4) and in the last five ns of MD simulations were also second modified fingolimod only, three atoms mentioned in Table 5. Figure 3 shows minimum (i.e. C13, C5 and C8) (Fig. 1B) have SASA distance between p53 promoter and ligands and less than 10 Å2. the number of contacts less than 0.6 nm Table 5 shows the average number of hydrogen between p53 promoter and ligands during the bonds between ligands and the p53 promoter. 20 ns MD simulation. Table 5. The average number of hydrogen bonds between ligands and p53 promoterand minimum distance
between them and number of contacts <0.6 nm between them during the last 5 ns of MD simulations
Average number of
Minimum distance
Number of contacts <0.6
hydrogen bonds between
between DNA and
nm between DNA and
DNA and drug
drug (nm)
Cladribine
Modified cladribie
Fingolimod
S m-fingolimod2
1. Fm-Fingolimod: First modification of fingolimod (i.e. deleting one carbon of fingolimod tail); 2. Sm-Fingolimod: Second modification of fingolimod (i.e. deleting four carbon of fingolimod tail). The maximum number of hydrogen bonds These results were confirmed by the minimum p53 promoter belongs to distance between ligands and p53 promoter cladribine and modified cladribine, and this and also the number of contacts between them parameter is similar in them. Then their (Fig. 3). In addition, first modified fingolimod interactions with p53 promoter are strong (fm-fingolomod) has the most minimum (Table 5). In addition, the number of distance and the least number of contacts with hydrogen bonds between fingolimod, first or p53 promoter among fingolimod and its second modified fingolimod are the same but derivatives. However, these parameters are lower than those between cladribine and more proper in second modification of modified cladribine. This means that the fingolimod (sm-fingolomod) and its interaction interaction of fingolimod and its derivates with p53 promoter is stronger relative to native with p53 promoter is weak. or first modified fingolimod (Table 5). Molecular and Biochemical Diagnosis (MBD). Vol.1, No.2 (2014), 105-122

K. Mahnam et al.

Figure 3. The minimum distance (A) and the number of contacts less than 0.6 nm between p53 promoter and
drugs (B) during 20 ns of MD simulations. Fm-fingolimod: First modification of fingolimod,i.e. deleting one
carbon of fingolimod tail. Sm-fingolimod: Second modification of fingolimod, i.e. deleting four carbon of
fingolimod tail.
3. Binding free energy results
snapshots during the last 5 ns of MD Table 6 shows binding free energy (ΔGb), Van simulation. Binding free energy of cladribine, der Waals and electrostatic energies of all modified cladribine and second modified ligands with p53 promoter obtained from 100 fingolimod to the p53 promoter is negative. Molecular and Biochemical Diagnosis (MBD). Vol.1, No.2 (2014), 105-122 In silico fingolimod and cladribine binding to p53 gene
This means that these drugs can bind the p53 promoter and through inhibition of the p53 fingolimod and first modified fingolimod to gene transcription probably induce cancer; p53 promoter is positive, so they may not bind then they can be supposedly carcinogen. the p53 promoter. Table 6. MM/PBSA binding free energies (kcal/mol) for ligand/DNA complexes during the last 5 ns of MD
simulation
Complex name
∆Eelec
∆Gpolar
∆Gbinding
Cladribine
Modified Cladribine
Fingolimod
S m-fingolimod2
1. Fm-Fingolimod: First modification of fingolimod (i.e. deleting one carbon of fingolimod t ail); 2. Sm-Fingolimod: Second modification of fingolimod (i.e. deleting four carbon of fingolimod tail). Abbreviations: ΔEelec = Electrostatic energy of interaction, ΔEvdw = Van der Waals energy of interaction. ∆Gpolar=polar solvation free energy, ∆Gnon-polar= Non-polar solvation free energy. Binding free energy of modified cladribine to consistent with visual inspection of the middle the p53 promoter is more positive and weaker structures of the most populated structures than native cladribine. The results obtained obtained from MD simulation (Fig. 2). from binding free energy (Table 6) and MM/PBSA results show that binding of docking (Table 1) for modified cladribine are cladribine to the p53 promoter is more opposite. Of course, results obtained from MD negative than fingolimod which means that simulation are more accurate than those from cladribine probably is a powerful inhibitor in dockings since water molecules and ions initiation of p53 gene transcription. This may explicitly present in molecular dynamics be due to the similarity of purine rings of simulation and MM/PBSA calculations, but in cladribine to adenosine. The results of dockings implicit solvent utilized and therefore MM/PBSA calculations shows that as compare water molecules and ions do not exist. This with the native fingolimod, if one carbon is suggests that MD simulation and MM/PBSA (Fm-Fingolimod), calculations are more accurate, and modified binding free energy (ΔGb) increases but it cladribine than to cladribine has a weaker decreases when four carbons (sm-fingolimod) interaction with p53 promoter. are removed (Table 6). These results are The negative binding free energy of the consistent with MD simulation (Table 5 and Fig. 3) but contrasted with docking results Molecular and Biochemical Diagnosis (MBD). Vol.1, No.2 (2014), 105-122 K. Mahnam et al.
(Table1). Reducing four carbons from the and more favorable for interactions of aliphatic tails of fingolimod increases binding fingolimod and its derivatives with p53 strength of fingolimod to the p53 promoter. promoter (Table 6). This suggests that the Then it is an inappropriate modification for fingolimod and it can be investigated through cladribine and fingolimod with p53 promoter empirical studies. There is a very good coordination between the average number of The number of the first ten nucleotides with hydrogen bonds during simulation and binding the most total energy contributions in binding free energy (Tables 5, 6). Also the differences of ligands to the p53 promoter were mentioned in the Van der Waals free and bound energies in Table 7. As seen 3´-A89C88A90T87-5´ or of all drugs during the last 5 ns MD simulation 5´-T16G17A18T19G20G21-3´ sequence has a were calculated. According to the MM/PBSA favorable interaction with cladribine however, results, the Van der Waals interactions are 5´-G17A18T19G20G21-3´ sequence has a more important (more negative) and more favorable interaction with modified cladribine favorable for interactions of cladribine and (Tables 6 and 7). Interactions of fingolimod modified cladribine with p53 promoter. and its derivatives are weak and interaction Electrostatic interactions are more important energies are below -1.1 kcal/mol (Table 7). Table 7.The first ten nucleotides that have the most total energy contribution in binding of drugs to p53
promoter (number of nucleotides are as mentioned in Table 4)

Cladribine
Modif ied cladribine
Fingolimod
Sm-f ingolimod2
Notes: 1.Fm-Fingolimod: First modification of fingolimod (i.e. deleting one carbon of fingolimod tail). 2. Sm-Fingolimod: Second modification of fingolimod (i.e. deleting four carbon of fingolimod tail). Num= Number of nucleotide in p53 promoter, Nuc=Nucleotide name, TE=Total energy of interaction each nucleotide with p53 promoter. Conclusions
in the binding of cladribine and fingolimod and In this in silico study we showed a difference some of their derivatives to the p53 promoter. Molecular and Biochemical Diagnosis (MBD). Vol.1, No.2 (2014), 105-122 In silico fingolimod and cladribine binding to p53 gene
This finding was confirmed by docking, 2-chloro-2´-deoxy triphosphate (CldATP) (Foley et al., 2004) and MM/PBSA methods. fingolimod phosphate (Cohen et al., 2007) on Based on the in silico studies it has been p53 gene promoter since they are produced by demonstrated that both cladribine and modified some enzymes in the cell. Moreover, the effect cladribine (replacing -OH on carbon 3´ ribose of these drugs on exons of p53 gene is worth sugar of adenosine with -CH3) can bind the minor groove of p53 promoter and may lead to conformational changes inp53 promoter. These drugs can cause qualitative changes in the p53 The authors are grateful to Dr. Rashmi Kumari for his help at installation of g_mmpbsa carcinogenesis. MD simulation and MM/PBSA module. The authors also wish to thank Miss calculations showed that by modification of Fateme Karimi for her assistance in preparing cladribine its interactions decreases and the modified cladribine may be less carcinogenic than cladribine, assuming that the former References
compound is a more favorable modification. HJC, Postma JPM, Van This phenomenon is explained by knowing the Gunsteren WF, Dinola A, Haak JR. 1984. increased cladribine size and steric prohibition Molecular dynamics with coupling to an with minor grove of p53 promoter. In addition, external bath. J ChemPhys 81:3684-3690. [2] BaiL, Zhu WG. 2006. p53: structure, function and therapeutic applications. J electrostatics interactions are more important Cancer Mol 2(4):141-153. for binding of cladribine to p53 promoter. [3] Berendsen HJC, Van der Spoel D, Van Removal of one carbon atom from the aliphatic Drunen R. 1995. GROMACS: A message - tails of fingolimod increased the binding free energy whereas binding free energy decreased by deletion of four carbon atoms. It is Communications 91:43-56. suggested that modifications in fingolimod or [4] Baker NA, Sept D, Joseph S, Holst MJ, cladribine structure may provide an interesting McCammon JA. 2001. Electrostatics of new direction for drug development. In the nanosystems: Application to microtubules future studies, it is suggested to investigate the Molecular and Biochemical Diagnosis (MBD). Vol.1, No.2 (2014), 105-122 K. Mahnam et al.
98(18): 10037-10041. Unrestrained 5 ns molecular dynamics [5] Brown SP, Muchmore SW. 2009. Large- simulation of a cisplatin-DNA 1, 2-GG adduct provides a rationale for the NMR Boltzmann surface area for routine physics- conformational flexibility at the platinum based scoring of protein-ligand complexes. binding site. Journal of molecular biology J Med Chem 52(10):3159-3165. [6] Bradshaw RT, Patel BH, Tate EW, [12] Eisenhaber F, Lijnzaad P, Argos P, Sander Leatherbarrow RJ, Gould IR. 2011. C, Scharf M. 1995. The Double Cube Lattice Method: Efficient Approaches to computational alanine scanning techniques Numerical Integration of Surface Area and forprobing a prototypical protein-protein Volume and to Dot Surface Contouring of interaction. Protein Eng Des Sel 24(1–2): Molecular Assemblies J Comp Chem B. 2007. Application of Interferon Beta-1b in Multiple Sclerosis American family physician 70:1935. [8] Cheatham TE, Srinivasan J, Case DA, [14] Foley TT, Hentosh P, Walters DE. 2004. Kollman PA. 1998. Molecular dynamics 2-Chloro-2'-deoxyadenosine: alteration of and continuum solvent studies of the DNA: TATA element binding protein stability of polyG-polyC and polyA-polyT (TBP) interactions. Journal of molecular DNA duplexes in solution. J Biomol Struct modeling 10:32-37 Dyn 16(2):265-280. [15] Frisch, M.J. et al., .1998. Gaussian 98 [9] Cohen B, Rieckmann P. 2007. Emerging (Gaussian, Inc., Pittsburgh, PA). oral therapies for multiple sclerosis. [16] Gohlke H, Case DA. 2004. Converging International journal of clinical practice free energy estimates: MM-PB(GB)SA studies on the protein-protein complex ras- [10] Darden T, York D, Pedersen L. 1993. raf. J ComputChem 25(2):238-250. Particle meshEwald - An N.log (N) [17] Hess B, Bekker H, Berendsen HJC, Fraaije method for Ewald sums in large systems. J JGEM. 1997. LINCS: a linear constraint ChemPhys98:10089-10092. solver for molecular simulations. J Comp [11] Elizondo-Riojas MA, Kozelka J. 2001. Chem18:1463-1472. Molecular and Biochemical Diagnosis (MBD). Vol.1, No.2 (2014), 105-122 In silico fingolimod and cladribine binding to p53 gene
[18] Hess B, Kutzner D, Lindahl E. 2008. [25] Lindahl E, Hess B, Van der Spoel D. 2001. GROMACS 3.0: a package for molecular efficient, load-balanced, and scalable simulation and trajectory analysis. Journal molecular simulation. J Chem Theory of Molecular Modeling 7:306-317. Comput 4:435-447. [26] Morris G, Goodsell D, Halliday R, Huey R, [19] Huo S, Massova I, Kollman PA. 2002. Hart W, Belew R, Olson A. 1998. Computational alanine scanning of the 1:1 Automated docking using a Lamarckian human growth hormone-receptor complex. genetic algorithm and an empirical binding J ComputChem 23(1):15-27. [20] Kar P, Lipowsky R, Knecht V. 2011. computational chemistry 19(14):1639-1662. Importance of polar solvation for cross- reactivity of antibody and its variants with Computational alanine scanning to probe steroids. J PhysChem B 115:7661-7669. [21] Kumari R, Kumar R, Consortium OSDD, approach to evaluate binding free energies. JACS 121(36):8133-8143. GROMACS Tool for High-Throughput [28] Majumdar R, Railkar R, Dighe RR. 2011. Calculations. J ChemInf Docking and free energy simulations to Model Article ASAP. predict conformational domains involved [22] Karplus M, Kuriyan J. 2005. Molecular in hCG-LH receptor interactions using ProcNatlAcadSci 102(19):6679–6685. Structure, Function, and Bioinformatics 79(11):3108-3122. [29] Navikas V, Link H. 1996. Review: YJ. 2011. Rotational viscosity calculation cytokines and the pathogenesis of multiple method for liquid crystal mixture using sclerosis. Journal of neuroscience research molecular dynamics 12(3):135-139. [30] Neidle S, Pearl LH, Skelly JV. 1987. DNA structure and perturbation by drug binding. 2008. Overexpression of USF increases Biochemical Journal 243(1):1-13. TGF-beta1 protein levels, but G1 phase [31C, WFV. 2005. arrest was not induced in FRTL-5 cells.
Free energies of ligand binding for structurally diverse compounds. Proc Natl Molecular and Biochemical Diagnosis (MBD). Vol.1, No.2 (2014), 105-122 K. Mahnam et al.
Acad Sci 102:6750-6754. [32] Pronk S, Pãll S, Schulz R, Larsson P, [33] Reisman D, Rotter V. 1993. The helix-loop- Bjelkmar P, Apostolov R, et. al. 2013. helix containing transcription factor USF GROMACS 4.5: a high-throughput and binds to and transactivates the promoter of highly parallel open source molecular the p53 tumor suppressor gene. Nucleic simulation toolkit. Bioinformatics 29 (7): acids research 21(2):345-350. Molecular and Biochemical Diagnosis (MBD). Vol.1, No.2 (2014), 105-122

Source: http://mbd.modares.ac.ir/article_13300_810c3375e528cd566e23f73d608629f9.pdf