Theoretical study of the interaction between a high-valent manganese porphyrin oxyl-(hydroxo)-Mn(IV)-TMPyP and double-stranded DNA

Theoretical Study of the Interaction between aHigh-Valent Manganese Porphyrin

Oxyl-(hydroxo)-Mn(IV)-TMPyP andDouble-Stranded DNA

PHILIPPE ARNAUD,1 KRYSTYNA ZAKRZEWSKA,2 BERNARD MEUNIER1

1Laboratoire de Chimie de Coordination du CNRS, 205 route de Narbonne,31077 Toulouse cedex 04, France

2Laboratoire de Biochimie Theorique, Institut de Biologie Physico-Chimique,13, rue Pierre et Marie Curie, 75005 Paris, France

Received 13 June 2002; Accepted 30 August 2002

Abstract: Cationic porphyrin derivatives such as meso-tetrakis(4-N-methylpyridinium)porphyrin, TMPyP, have beenshown to interact with double-stranded DNA. The manganese derivative, Mn(III)-TMPyP, activated by an oxygen donorlike potassium monopersulfate, provides an efficient DNA-cleaving system. Previous experimental work1 has shownthat DNA cleavage by the Mn(III)-TMPyP/KHSO5 system was due to an oxidative attack, within the minor groove ofB-DNA, at the C5� or C1� carbons of deoxyribose units. The aim of this study was to use molecular modeling toelucidate the specificity of the interactions between the transient active species oxyl-Mn(IV)-TMPyP and the DNAtarget. Geometric parameters, charges, and force field constants consistent with the AMBER 98 force field werecalculated by DFT methods. Molecular modeling (mechanics and dynamic simulations) were performed for oxyl-(hydroxo)-Mn(IV)-TMPyP bound in the minor groove of the dodecamer d(5�-TCGTCAAACCGC)-d(5�-GCGGTTT-GACGA). Geometry, interactions, and binding energy of the metalloporphyrin located at the A.T triplet region of thedodecamer were analyzed. These studies show no significant structural change of the DNA structure upon ligandbinding. Mobility of the metalloporphyrin in the minor groove was restrained by the formation of a hydrogen bondbetween the hydroxo ligand trans to the metal-oxyl and a DNA phosphate, restricting the access of the oxyl group tothe (pro-S) H atom at C5�.

© 2003 Wiley Periodicals, Inc. J Comput Chem 24: 797–805, 2003

Key words: manganese porphyrin; DFT; B-DNA; molecular modeling

Introduction

Bleomycin is an antitumor agent extracted from Streptomycesverticillus, which is able to cleave the DNA of cancer cells in thepresence of ferric ions, molecular oxygen, and a reducing agent.2–6

This molecule has been a paradigm for the design of metal com-plexes for probing structural variations in nucleic acids, identifi-cation of binding sites of DNA ligands, and for producing artificialDNA cleavers and potential chemotherapeutic agents. Good can-didates for this type of application are transition–metal porphyrincomplexes. Porphyrin derivatives have reasonable affinities forDNA and their binding mode depends on the nature of the meso-substituents on the porphyrin ring, DNA sequence, ionic strength,and the nature of the metal ion liganded by the porphyrin. Distinctsequence preferences in the case of H2TMPyP and its squareplanar complexes with ions such as CuII, and NiII, which do not

have axial ligands, have previously been rationalized by molecularmodeling studies.7,8 These results suggest intercalative binding at5�-CpG steps and nonintercalative (possibly minor groove) inter-actions at 5�TpA sequences. For the diaxially liganded porphyrinswith MnIII, FeIII, or CoIII, intercalation is precluded; consequently,they bind primarily to AT regions in a nonintercalative manner.Similarly, cationic complexes possessing bulky substituents on theporphyrin ring or at least one axially bound ligand are too thick forintercalation.

Manganese metalloporphyrin can be activated in a bleomycin-like fashion, with O2 and a reducing agent or, in the presence ofKHSO5,9–13 to generate a high valent metal-oxo species14,15 (Fig.

Correspondence to: P. Arnaud; e-mail: [email protected]

Contract/grant sponsor: CNRS

© 2003 Wiley Periodicals, Inc.

1) which is able to degrade single- or double-stranded DNA.11,12

Previous experimental analyses of the binding mode and the deg-radation of DNA by tetrakis-(4-N-methylpyridiniumyl)porphyrin-ato)manganese(V)-oxo) [oxo-Mn(V)-TMPyP] (Fig. 1) showed astrong preference for AT-rich regions.16–20 This selectivity wasexplained by electrostatic interactions of the cationic porphyrinwith the more negative potential in the minor groove of AT-richpolymers compared to GC-rich polymers. It is essential for therational drug design of antitumoral, antiviral agents, or otherpharmaceuticals based on DNA cleavers to understand at themolecular level all the different factors (steric, electronic, etc.)involved in their mode of interactions with DNA.

Studies of the highly specific cleavage induced by the high-valent oxo-Mn(V)-TMPyP complex on both 3�-sides of threecontiguous A.T base pairs (bp) (A.T triplets) of double-strandedoligonucleotides1 have shown that immediate DNA strand scissionoccurred through an oxidative C—H bond activation at C-5� of thesugar leading to 3�-phosphate and 5�-aldehyde termini at the sitesof breaks. These lesions were identified by the formation of fur-fural (FUR).

The homolysis of a C—H bond depends primarily on thecarbon substituents (from a thermodynamic point of view remov-ing a H-atom from a tertiary carbon is easier than from a secondarycarbon). For metal-oxo species interacting within the minor grooveof B-DNA, two tertiary C—H bonds are easily accessible: C4�—Hand C1�—H. The third tertiary C—H bond (C3�—H) is onlyaccessible from the major groove. C2� and C5� are secondarycarbon atoms.10 In both cases, one C—H bond is accessible fromthe minor groove the (pro-R) H atom at C2� and the (pro-S) Hatom at C5� while the second hydrogen at each site is accessiblefrom the major groove. The fact that the activation takes place ata secondary carbon C5� and not at C4� or C1� can have severalorigins. The H-atom abstraction step is probably controlled bystereoelectronic and steric parameters involving orbital orienta-tions, i.e., orientation of the metal-oxo species with respect withthe C—H bond itself.21 Another particularly interesting point isthat two single-strand breaks are observed at the 3�-ends of the A.Ttriplet.1 It is not clear whether they occur simultaneously or se-

quentially. It is generally assumed that the active species is anoxyl-Mn(IV)-TMPyP with one active oxygen.14,15 This wouldsuggest that the cuts are sequential, an opposite orientation of theoxygen being required for the cut on each strand. This view is inagreement with experiments for bleomycin showing that nicks onone strand of DNA increase cuts on the other strand.22

The aim of our work was to examine some of these questionsby molecular modeling. We started by studying the interactionbetween oxyl-(hydroxo)-Mn(IV)-TMPyP and a DNA dodecamerd(5�;-TCGTCAAACCGC)-d(5�-GCGGTTTGACGA), the se-quence used in the experimental cleavage studies.1 In this articlewe look at the conformational and dynamic aspects of the com-plexes with the porphyrin located in the DNA minor groove.Starting with a nonsymmetric ligand and a nonpalindromic basesequence we had to consider separately two cases corresponding tothe orientation of the metal-oxyl group towards one or the otherstrand of the DNA double helix.

Structural parameters, charges, and force field parameters fortwo spin states of manganese, the ground state of manganeseoxo-Mn(V)-TMPyP (A, Fig. 1) and the excited state of manganeseoxyl-Mn(IV)-TMPyP (B, Fig. 1) were obtained by ab initio cal-culations. These parameters were used in our molecular mechanicsstudy of the oxyl-Mn(IV)-TMPyP/dodecamer complex as well asin two MD simulations of the DNA–cleaver systems.

Methods

Structure, Charge, and Amber Force Field Parameters

Geometry, charge, and atomic parameters necessary for molecularmodeling with the Amber98 force field were obtained by DFTquantum mechanical calculations with the Gaussian 98 (releaseA.9) program.23 The DFT24,25 method and the B3LYP level26

were chosen in reference to other calculations on porphyrin-typemolecules.27,28 Two spin states of manganese were studied:Mn(V) � O (Mn-oxo), which corresponds to the ground state andMn(IV)-O. (Mn-oxyl) corresponding to an excited state.14,15 Inboth cases, oxo- or oxyl-Mn-TMPyP complexes were modeled intwo steps. In the first step, the porphyrin nucleus alone was used.In the second step, one of the N-methylpyridinium-4-yl pendantgroups was attached in the meso position of the porphyrin, inagreement with the crystallographic structure of Mn(III)-TMPyP.29 In the first step the 6-31G* basis set was used for H, C,N, and O atoms and the full geometry optimization and frequencycalculations were carried out. For the second step, the porphyrinwith the N-methylpyridinium-4-yl substituents, due to the size ofthe system, the basis set for H, C, N, and O atoms was reduced to3-21G*. In both steps, the DZVP230 basis set was used for man-ganese. For the complete molecule the geometry optimization andfrequency calculation were done only for oxo-Mn(V)-TMPyP casebecause the calculations on the porphyrin nucleus have shown thatthe structural differences between the fundamental and exitedstates were limited to an elongation of the Mn—O bond.

To obtain atomic charges of the ligand in a good balancewith the DNA atom charges the Merz–Kollman fitting algo-rithm31,32 was used. The procedure consists in least-squarefitting of atomic monopoles to reproduce the electrostatic po-

Figure 1. Schematic representation of the high-valent metal-oxo com-plex oxo-Mn(V)-TMPyP with a trans hydroxo ligand. This metal-oxospecies (A) can also exist as a metal-oxyl species (B) (see refs. 14 and15). The four pyridinium substituents at the meso position create fourpositive charges at the periphery of the molecule.

798 Arnaud, Zakrzewsak, and Meunier • Vol. 24, No. 7 • Journal of Computational Chemistry

tential at 3183 points defined by four shells of surfaces at 1.4,1.6, 1.8, and 2 times the Van der Waals radii. Esp fit chargeshave been shown to reproduce well conformation energies ofmolecules.33 The geometrical data, atomic charges, and forceconstants were then used to fully parameterize the oxyl-Mn(IV)-TMPyP molecule within the AMBER98 force field.These computations have led us to define new atom types forthe AMBER force field (Figs. 2 and 3), assigned according topublished guidelines33 and in reference to previous molecularmodeling results obtained by Guliaev.34

Molecular Mechanics

The DNA dodecamer d(5�-TCGTCAAACCGC)-d(5�-GCGGTTT-GACGA) was studied with the metalloporphyrin imbedded withinthe minor groove. In reference to previous experimental results1

that showed two sites for DNA breaks, two distinct orientations ofoxyl-Mn(IV)-TMPyP in the minor groove of an A.T triplet (Fig. 4)were considered. In the first, the metal-oxyl group was orientedtowards the H5� atom of the cytosine 9 (C9) of the 5� T-strand andin the second towards the H5� atom of the guanine 17 (G17) of the5� G-strand (see Fig. 4).

The complexes were energy optimized with the JUMNA (Junc-tion Minimization of Nucleic Acids) algorithm.35 This programprovides versatile tools for the generation of starting structures,conformational searching, and energy optimizations of complexesformed between nucleic acids and ligands. The description ofintramolecular flexibility by internal coordinates with fixed bondlengths is complemented by helicoidal variables that relate thetranslations and rotations of the nucleic acid bases and of the metalcomplexes with respect to a helical axis system. This featurefacilitates the creation of appropriate starting structures for energyminimization with different orientations of the metalloporphyrinrelative to the DNA molecule. The AMBER98 force filed andGeneralized Born method36 for the calculation of solvation energywere used in the potential energy calculation.

Molecular Dynamics in Water

The best complexes obtained by molecular mechanics were usedas starting structures for the simulations with molecular dynamicswith explicit water. For each orientations AMBER topology andparameter files for oxyl-Mn(IV)-TMPyP and for the B-DNA du-plex d(5�-TCGTCAAACCGC)-d(5�-GCGGTTTGACGA) weregenerated by using the xLeap module of AMBER 6.0.37 To neu-tralize the combined porphyrin/DNA charge, 18 sodium counteri-ons were placed along the phosphate bisectors and a rectangularbox was added providing at least 10 Å of water around any soluteatom (�4500 water molecules). To examine the influence of theporphyrin binding on the DNA conformation and dynamics weperformed also a 2-ns simulation for the free oligomer.

The simulation protocol involved a series of progressive energyminimizations followed by a 10-ps heating phase and a 25-psequilibration period before 2-ns data collection. During the mini-zation step, the water box was subjected to six rounds of minimi-zation during which harmonic restraints on the solutes and coun-terions were reduced from 100 kcal/mol/Å2 to 0.38,39 Theminimization step were followed by 35 ps of MD, during whichthe temperature was slowly raised from 0 to 300 K over 10 ps andwas maintained at 300 K with Berendsen temperature coupling.40

Figure 2. Atom types of oxyl-Mn(IV)-TMPyP.

Figure 3. Atom numbers of oxyl-Mn(IV)-TMPyP.

Figure 4. DNA duplex d(5�-TCGTCAAACCGC)-d(5�-GCGGTTT-GACGA).

Interaction between TMPyP and Double-Stranded DNA 799

During the heating step, constraints (25 kcal/mol/Å2) werereimposed on the solute with respect to its conformation at theend of minimization. For the remaining 25 ps of equilibration,constraints on the solute molecules were reduced by 5.0 kcal/mol/Å2 each round. Molecular dynamics (MD) simulationswere carried out by using the SANDER module of AMBER 6.0with SHAKE41 applied to all hydrogen atoms and 2-fs timesteps. Longe-range electrostatic forces were evaluated using theparticle mesh Ewald method.42,43 A 9 Å cutoff was applied tothe Lennard–Jones interactions. Constant pressure was main-tained with isotropic molecule based scaling. System coordi-nates were saved every 0.2 ps.

Three-dimensional structures and trajectories were visuallyinspected using the computer graphics program VMD.44 Root-mean-square (rmsd) deviations from the initial structures, in-teratomic distances and average structures from the trajectorieswere calculated using the CARNAL and PTRAJ modules ofAMBER.

DNA conformations were analyzed in terms of helical param-eters, widths, and depths of the grooves, sugar puckering states,and backbone torsion angles. The helical parameters refer to theglobal axis system of the DNA calculated by the CURVES pro-gram.45,46

All calculations were performed on the SGI Origin 2000 atToulouse University Computer Center, on DEC Alpha station atIBPC (Paris) and locally on SGI Octane workstations. Each MDsimulation in water took about 450 h of CPU time per ns.

Results and Discussion

Charge Distribution and Geometry of the Two Spin Statesof Oxo- or Oxyl-Mn-TMPyP

Figure 5 shows the B3LYP/3-21G* optimized geometry of oxyl-Mn(IV)-TMPyP. For both spin states, pyrrole rings were close toplanarity. The manganese atom was located in the plane containingthe four porphyrin nitrogens. The calculated bond lengths andangles in the rest of the molecule were in excellent agreement withcrystallographic and calculated data on typical metalloporphy-rins.27,29,47 The distances from the nitrogen atoms were similar(see Fig. 3 for atom type), Mn–N1 distance was 2.03 Å, Mn–N2:2.01 Å, Mn–N3: 2.02 Å , Mn–N4: 2.01 Å. The Mn–O distance was1.58 Å for the optimized Mn(V) � O and 1.82 Å for Mn(IV)–O �. This latter corresponding to a single bond distance. The distancesfor Mn–OH were similar for the ground and exited states of thehigh-valent manganese species with respectively 1.78 and 1.82 Å.The calculated Mulliken spin densities on the Mn(IV) and oxylatoms were 2.60 and 1.15, respectively, and were in agreementwith a Mn(IV)–O � electronic state. Due to the program limitationswe could not verify these values by the NBO analysis of spindensity in the basis set used. However, the NBO analysis per-formed for the smaller set STO-3G* gave very similar results withthe values 2.62 and 1.26 for the Mn and oxyl atoms, respectively.The total energy difference for the two spin states of manganesewas 5 kcal/mol. All these results are in good agreement with thoseobtained by Shaik and coll. for two spin states of Fe.14,15 Theresults of these calculations are resumed in Figure 6.

The exited state of manganese Mn(IV) is supposed to be theactive state involved in DNA cleavage via C—H bond hydroxy-lation.14,15 Therefore, the charge distribution used in the rest ofthis study corresponds to the excited state. The atomic monopolesobtained with the Merz–Kollman method are given in Table 1, andrefer to the atom numbering indicated in Figure 3. Except in thevicinity of manganese, these charges are similar to those obtainedin previous molecular modeling studies of TMPyP34 (this ligandhas four positive charges due to the four pyridinium substituents).

Figure 5. Structure of oxyl-Mn(IV)-TMPyP optimized using theB3LYP/3-21G* method. [Color figure can be viewed in the onlineissue, which is available at www.interscience.wiley.com.]

Figure 6. Schematic representation of the two spin states studied andthe main distance (blue) and energetic (black) results (see refs. 14 and15). [Color figure can be viewed in the online issue, which is availableat www.interscience.wiley.com.]


Additional AMBER Force-Field Parameters forOxyl-Mn(IV)TMPyP4�

Table 2 reports the AMBER atom types used and additionalparameters for bonds, bond angles, and torsions for oxyl-Mn(IV)-TMPyP. The energy minimized structure calculated with theseparameters within Amber 6.0 package resulted in a geometry withan RMSD of 0.19 Å with respect to the ab initio optimization.

Molecular Mechanics

Two models for oxyl-Mn(IV)-TMPyP interactions in duplex DNAembedded in the A.T triplet and differing only in the orientation ofthe oxyl-ligand with respect to the DNA strands, were docked andenergy-minimized using the JUMNA procedure. The binding en-ergies obtained are reported in Table 3.

Both orientations of the metalloporphyrin gave the same com-plexation energy. The formation of the complex DNA � metallo-porphyrin was strongly stabilized by the Lennard–Jones compo-nent, characterizing a good steric fit of the two molecules. As it canbe seen in Figure 7a and b, the oxyl-Mn(IV)-TMPyP moleculefitted snugly into the A/T region of the minor groove, with the twoinner N-methyl-pyridinium groups being almost parallel to thegroove floor. The other two N-methyl-pyridinium groups werepointing away from the minor groove. The porphyrin plane wasinclined with respect to the DNA helix axis at an average angle of50°, in agreement with linear dichroism measurements for othermetalloporphyrins.48 The metalloporphyrin binding site coveredroughly 3.5 base pairs including the three A.T base pairs. Thecenter of the metalloporphyrin, containing the oxyl-Mn(IV) spe-cies, was located close to the entrance of the groove. The oxyl-Mn(IV) was also close to one of the strands of the backbone and

in particular to its C4� and C5� atoms. Because the electronicdensity of the highest orbital of the oxyl-(hydroxo)-Mn(IV)-TMPyP is localized along the Mn-oxyl axis (Fig. 9), we calculatedthe angle between the Mn-oxyl vector and the correspondingH—C vector. In the C9 complex, the distance between the activeoxygen and C9–H5� was 2.52 Å, slightly shorter than its distanceto A8–H4� (2.66 Å). The angle between Mn-oxyl vector and thecorresponding C9–H5� vector was 127° and was similar (135°)with the T8–H4� vector. In the G17 orientation, these distances are2.63 Å with the average angle of 49° and 2.36 Å and the averageangle of 89° for G17–H5� and T18–H4�, respectively. Thesedistances and angles do not explain the H-atom abstraction pref-erence for H5� that has been demonstrated experimentally.1

To see how the dynamics of the system influences the dispo-sition of these different key atoms involved in the early steps of theDNA cleavage, we consequently performed simulations in explicitwater.

MD Simulation of Oxyl-Mn(IV)-TMPyP/DNA Complexes

MD trajectories 2 ns long were generated for the two orientationsof the ligand. These trajectories were sampled every 0.2 ps andexamined visually by using VMD. The overall stability of eachtrajectory was evaluated by calculating RMSD values of each0.2-ps snapshot with respect to the coordinates of the initial struc-tures. Plots of the average RMSD values as a function of time areshown in Figure 8a. After the initial heating to 300 K, the pro-gression of the rms deviation of the coordinates of the solutes withrespect to initial structures remained stable for both complexes.The absence of drifting to higher rms deviations suggests adequatesystem equilibration during the sampling time. In addition, all the

Table 1. Monopole Charges for Oxyl-Mn(IV)-TMPyP4�.

N1 0.26 C24 �0.34 H47 0.24 C70 0.41N2 0.14 H25 0.21 C48 0.34 C71 �0.36N3 �0.07 H26 0.23 C49 �0.28 H72 0.24N4 0.09 H27 0.20 H50 0.23 C73 0.06C5 0.18 H28 0.21 C51 0.01 H74 0.19C6 �0.23 H29 0.22 H52 0.21 N75 0.13C7 0.03 H30 0.22 N53 0.14 C76 �0.41C8 �0.03 H31 0.22 C54 �0.43 C77 �0.01C9 �0.17 H32 0.23 C55 0.00 H78 0.21C10 0.10 Mn33 0.00 H56 0.20 C79 �0.29C11 0.24 O34 �0.28 C57 �0.27 H80 0.24C12 �0.33 O35 �0.55 H58 0.23 H81 0.21C13 0.32 H36 0.35 C59 0.40 H82 0.21C14 0.20 C37 0.37 C60 �0.29 H83 0.21C15 �0.21 C38 �0.31 H61 0.23 H84 0.21C16 0.09 H39 0.23 C62 �0.01 H85 0.22C17 �0.24 C40 0.04 H63 0.21 H86 0.22C18 �0.21 H41 0.20 N64 0.14 H87 0.21C19 �0.27 N42 0.12 C65 �0.41 H88 0.21C20 �0.27 C43 �0.41 C66 0.04 H89 0.21C21 �0.25 C44 0.03 H67 0.20 H90 0.21C22 �0.31 H45 0.20 C68 �0.35 H91 0.21C23 �0.18 C46 �0.31 H69 0.24 H92 0.21


base pairs remained well stacked and internally hydrogen bondedin both duplexes studied for the whole length of the simulationwithout any external restrains.

The average distance between the porphyrin and the DNA masscenter was also monitored for each trajectory. The deviations ofthese values were all quite small, indicating that the metallopor-phyrin was strongly interacting with DNA in a stable local mini-

mum (Fig. 8b). The distance separating the porphyrin and DNAcenters of mass was similar in both cases, 10.6 � 0.49 Å and10.48 � 0.42 Å for the C9 and G17 complexes, respectively.

The steric arrangement for the homolytic reaction was analyzedduring the production step of the simulations. As for our molecularmechanics studies, we considered, first, the distances and relativeorientations between the oxyl entity and the corresponding hydro-gen H5� and H4�. For the C9 complex, the average distancebetween O34 and A8–H4� was 2.88 Å � 0.37, with an averageangle of 45° � 75, whereas the average distance between O34 andC9–H5� was slightly longer, 3.28 Å � 0.60 with a much smalleraverage angle of 21° � 19. For the G17 orientation, the average

Table 2. Additional Bonds, Bond Angles, Torsions, and ImproperTorsions for Oxyl-Mn(IV)-TMPyP Introduced to the AMBERForce-Field.

Bondsa Kr req

Mn-NP 50.0 2.040Mn-OX 200.0 1.828

Anglesd K� �eq

NP-Mn-OX 55.0 90.0NP-Mn-O35 55.0 90.0NP-Mn-NP 0.0 90.0CV-NP-Mn 30.0 126.5OX-Mn-O35 41.0 168.0Mn-O35-H36 35.0 126.4

Vn/2 � n

Specific torsionb

NP-Mn-O35-H36 2.50 180.0 1OX-Mn-O35-H36 2.50 180.0 1

Improper torsionb

X-X-CV-X 1.00 180.0 2X-X-NP-X 1.00 180.0 2X-X-CM-X 1.00 180.0 2

General torsionb

X-NP-Mn-X 0.00 180.0 2

aKr (kcal mol�1 Å�2), stretching force constant; req (Å), equilibrium bondlength.bVn/ 2 (kcal mol�1), torsional force constant; � (deg), phase; n, periodicity.cX means any atom.dK� (kcal mol�1 rad�2), bending constant; �eq (deg), equilibrium bondangle.

Table 3. Binding Energies (kcal/mol) of the Oxyl-Mn(IV)-TMPyP in theTwo Orientations with Respect to the Dodecamer.

Orientations

C9-H5� G17-H5�

E inter �69.7 �66.5E vdw �53.3 �49.0E elec �16.4 �17.5�E DNA 25.7 19.7�E liganda 10.0 11.6�E � E inter � �E DNA � �E liganda �34.5 �35.2

aIn the present table “ligand” stands for metalloporphyrin.

Figure 7. (a) Stereoview of the optimized oxyl-Mn(IV)-TMPyP/DNAcomplex with the oxyl oriented toward the H5� atom of the cytosine 9of the 5� T-strand. (b) Stereoview of the optimized oxyl-Mn(IV)-TMPyP/DNA complex with the oxyl oriented toward the H5� atom ofthe guanine 17 of the 5� G-strand. [Color figure can be viewed in theonline issue, which is available at www.interscience.wiley.com.]


distance of the oxyl from H5� was 3.72 � 0.66 with an averageangle 131° � 98, longer than the distance to H4�, which was 3.05Å � 0.46, with an average angle of 45° � 128.

The distances monitored during 2 ns of MD are shown inFigure 10a for the C9 complex and in Figure 10b for the G17orientation of the ligand.

These geometrical features did not explain the preference forthe H5� abstraction by the high-valent metal-oxo species, and evenif the angular arrangement seemed to privilege the H5� location forthe C9 complex this not true for the G17 orientation.

One of the possible reasons to explain the longer averagedistance of the metal-oxyl entity group from H5� than from the H4�is the formation of a hydrogen bond between the trans ligandhydroxo and a phosphate group on the “noncut” DNA strand. For

the C9 orientation, the phosphate anionic oxygen located withinthe minor groove side of G17 was involved and, for the oppositeorientation, that of A8. These hydrogen bonds were present for72% of the time with average distances between donor and accep-

Figure 8. (a) Time dependence of RMSD of atomic coordinates ofindicated oxyl-Mn(IV)-TMPyP/d(5�-TCGTCAAACCGC)-d(5�-GCG-GTTTGACGA) complexes relative to initial starting structures for 2ns of unrestrained MD. (b) Time dependence of distances between themass centers of oxyl-Mn(IV)-TMPyP and d(5�-TCGTCAAACCGC)-d(5�-GCGGTTTGACGA) for 2 ns of unrestrained MD. [Color figurecan be viewed in the online issue, which is available at www.interscience.wiley.com.]

Figure 9. Orbital representation of the excited state of oxyl-Mn(IV)-TMPyP. [Color figure can be viewed in the online issue, which isavailable at www.interscience.wiley.com.]

Figure 10. (a) Time dependence of distance between oxyl and C9/H5�(blue) and A8/H4� (green) during 2 ns of unrestrained MD. (b) Timedependence of distance between oxyl and G17/H5� (blue) and T18/H4�(green) during 2 ns of unrestrained MD. [Color figure can be viewedin the online issue, which is available at www.interscience.wiley.com.]


tor being respectively 2.87 Å � 0.3 for an average angle of 18° forthe C9 complex and 3.0 Å � 0.38 and an average angle of 23° forG17. No direct correlation was detected between the existence ofthese hydrogen bonds and the relative distances of the oxyl fromH5� and H4�, but this hydrogen bond nevertheless appears toreduce the mobility of the metalloporphyrin within the DNAgroove.

Complexation of DNA with oxyl-Mn(IV)-TMPyP induced onlyminor changes of DNA conformation. Figure 11 shows the trajectoryaveraged groove width along the oligomer sequence. In the free DNA,the minor groove is narrowed by roughly 2 Å in the A.T triplet regionwith respect to the rest of the sequence. The influence of the metal-loporphyrin was different for the two orientations. When the metal-oxo was oriented towards H5� of C9 on the 5� T-strand, the generaleffect is the opening up of the groove, and in the other orientation (i.e.,the metal-oxo pointing towards H5� of G17 of 5� G-strand) the grooveis narrower outside the binding zone (Fig. 11). Standard deviations forall structure were very weak, less than 0.03 Å, except for the G17–C5base pair of the G17 complex, for which the value was 0.36 Å.Changes in the DNA helicoidal parameters were very small, rarelyexceeding 10°. A few localized changes were observed in the DNAbackbone. They mainly involved the change of the � angle from transto gauche, with the corresponding sugar passing into the east confor-mational zone. For the C9 complex, these changes concern T20, A8,and T4 and, in the case of G17 orientation, only T18 is concerned.

Very minor disturbances of the DNA structure have alreadybeen observed for other minor groove binding ligands such asnetropsin, berenil and pentamidine.49,50 These ligands are, how-ever, much smaller, and it is remarkable that a ligand as big asoxyl-Mn(IV)-TMPyP fits so snugly within the minor groove. Itshould be noted that keeping DNA conformation unperturbed is animportant feature for structure-based drug design because it allowsto do most of the docking for a fixed conformation of DNA.

Conclusions

In this work we obtained atomic parameters consistent with theAmber force field for a manganese(IV) porphyrin complex withoxyl and hydroxo axial substituents. Ab initio (DFT level) calcu-lations for two degrees of oxidation of manganese have shown anelongation of the Mn(IV)-O � bond with respect to Mn(V) � O.These parameters have been applied to study the DNA complex-ation by oxyl-(hydroxo)-Mn(IV)-TMPyP. Two equienergeticstructures were obtained by molecular mechanics in which theoxyl ligand was oriented towards the H5� atom of the C9 and G17located at the two sides of the A.T triplet. For these two cases,molecular dynamics simulations were performed in explicit water.These 2-ns simulations did not show the dominance of conforma-tions with the oxyl group positioned in the vicinity of the C5�H,which would privilege the oxidative attack at this H-atom.

Our simulations have suggested that one of the reasons behindthese results might be the formation of a hydrogen bond betweenthe metalloporphyrin hydroxo ligand and an anionic phosphateoxygen located on the opposite strand. No direct correlation wasdetected between the existence of this hydrogen bond and therelative distance of the oxyl from the two candidates for theoxidative attack; but this hydrogen bond, present for 72% ofsimulation time could well inhibit the motion of the metallopor-phyrin within the minor groove.

This work raises a question concerning the nature of transientspecies able to abstract H5�: is the oxyl-(hydroxo)-manganese(IV)complex really the H-atom abstracting species or is the H5� ab-straction performed by a precursor of such high-valent metal-oxospecies? We are currently investigating this hypothesis.

Our dynamic simulations have also shown that a ligand as bigas an oxo-metalloporphyrin can be docked within the minorgroove of B-DNA double helix with only minor structural changes.This is a very encouraging feature for the drug design of minorgroove binders.

References

1. Pitie, M.; Pratviel, G.; Bernadou, J.; Meunier, B. Proc Natl Acad SciUSA 1992, 89, 3967.

2. Hecht, S.M. Acc Chem Res 1986, 19, 383.3. Stubbe, J.; Kozarich, J.W. Chem Rev 1987, 87, 1107.4. Sigman, D. S. Biochemistry 1990, 29, 9097.5. Moser, H. E.; Dervan, P. B. Science 1987, 238, 645.6. Barton, J. K. Science 1986, 233, 727.7. Ford, K. G.; Pearl, L. H.; Neidle, S. Nucleic Acids Res 1987, 15, 6553.8. Hui, X.; Gresh, N.; Pullman, B. Nucleic Acids Res 1990, 18, 1109.9. Stubbe, J.; Kozarich, J. W.; Wei, W.; Vanderwall, D. E. Acc Chem Res

1996, 29, 322.10. Pratviel, G.; Bernadou, J.; Meunier, B. Angew Chem Int Ed Engl

1995, 34, 746.11. Bernadou, J.; Fabiano, A. S.; Robert, A.; Meunier, B. J Am Chem Soc

1994, 116, 9375.12. Pitie, M.; Bernadou, J.; Meunier, B. J Am Chem Soc 1995, 117, 2935.13. Mestre, B.; Jakobs, A.; Pratviel, G.; Meunier, B. Biochemistry 1996,

35, 9140.14. Shaik, S.; Filatov, M.; Schroder, D.; Schwarz, H. Chem Eur J 1998, 4,

193.

Figure 11. Minor groove width along the sequence. Black: free DNA,Red: oxyl oriented towards the H5� of the cytosine 9 of the 5� T-strand,Green: oxyl oriented towards the H5� of the guanine 17 of the 5�G-strand. [Color figure can be viewed in the online issue, which isavailable at www.interscience.wiley.com.]


15. Schroder, D.; Schwarz, H.; Shaik, S. Struct Bond 2000, 97, 91.16. Dabrowiak, J. C.; Ward, B.; Goodisman, J. Biochemistry 1989, 28, 3314.17. Carvlin, M. J.; Mark, E.; Fiel, R.; Howard, J. C. Nucleic Acids Res

1983, 11, 6141.18. Pasternak, R. F.; Gibbs, E. S.; Villafranca, J. J. Biochemistry 1983, 22,

5409.19. Ward, B.; Skorobogaty, A.; Dabrowiak, J. C. Biochemistry 1986, 25,

6875.20. Ding, L.; Bernadou, J.; Meunier, B. Bioconjugate Chem 1991, 2, 201.21. Sorokin, A.; Robert, A.; Meunier, B. J Am Chem Soc 1993, 115, 7293.22. Keller, J. T.; Oppenheimer, N. J. J Biol Chem 1987, 262, 15144.23. Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb,

M. A.; Cheeseman, J. R.; Zakrzewski, V. G.; Montgomery, J. A., Jr.;Stratmann, R. E.; Burant, J. C.; Dapprich, S.; Millam, J. M.; Daniels,A. D.; Kudin, K. N.; Strain, M.C.; Farkas, O.; Tomasi, J.; Barone, V.;Cossi, M.; Cammi, R.; Mennucci, B.; Pomelli, C.; Adamo, C.; Clif-ford, S.; Ochterski, J.; Petersson, G. A.; Ayala, P. Y.; Cui, Q.; Moro-kuma, K.; Malick, D. K.; Rabuck, A. D.; Raghavachari, K.; Foresman,J. B.; Cioslowski, J.; Ortiz, J. V.; Baboul, A. G.; Stefanov, B. B.; Liu,G.; Liashenko, A.; Piskorz, P.; Komaromi, I.; Gomperts, R.; Martin,R. L.; Fox, D. J.; Keith, T.; Al-Laham, M. A.; Peng, C. Y.; Nanay-akkara, A.; Gonzalez, C.; Challacombe, M.; Gill, P. M. W.; Johnson,B. G.; Chen, W.; Wong, M. W.; Andres, J. L.; Head–Gordon, M.;Replogle, E. S.; Pople, J. A. Gaussian 98 (Revision A.7); Gaussian,Inc.: Pittsburgh, PA, 1998.

24. Hohenberg, P.C.; Kohn, W. Phys Rev B 1964, 136, 864.25. Kohn, W.; Sham, L. J. Phys Rev A 1965, 140, 1133.26. Becke, A. D. J Chem Phys 1993, 98, 5648.27. Ghosh, A. Acc Chem Res 1998, 31, 189.28. Kozlowski, P. M.; Rush, T. S., III; Jarzecki, A. A.; Zgierski, M. Z.;

Chase, B.; Piffat, C.; Ye, B.-H.; Li, X.-Y.; Pulay, P.; Spiro, T. G. JPhys Chem 1999, 103, 1357.

29. Prince, S.; Korber, F.; Cooke, P. R.; Smith, J. R. L.; Mazid, M. A. ActaCrystallogr C 1993, 49, 1158.

30. Godbout, N.; Salahub, D. R.; Andzelm, J.; Wimmer, E. Can J Chem1992 70, 560.

31. Besler, B. H.; Merz, K. M.; Kollman, P. A. J Comp Chem 1990, 11,431.

32. Singh, U. C.; Kollman, P. A. J Comp Chem 1984, 5, 129.33. Cornell, W. D.; Kollman, P. A.; Ferguson, D. M. J Am Chem Soc

1995, 117, 5179.34. Guliaev, A. B.; Leontis, N. B. Biochemistry 1999, 38, 15425.35. Lavery, R.; Zakrzewska, K.; Sklenar, H. Comput Phys Commun 1995,

91,135.36. Tsui, V.; Case, D. A. J Am Chem Soc 2000, 122, 2489.37. Perlman, D. A.; Case, D. A.; Caldwell, J. W.; Ross, W. S.; Cheatham,

T. E.; DeBolt, S.; Ferguson, D.; Seibel, G.; Kollman, P. A. ComputPhys Commun 1995, 91, 1.

38. Cheatham, T. E.; Miller, J. L.; Fox, T.; Darden, T. A.; Kollman, P. A.J Am Chem Soc 1995, 117, 4193.

39. Cheatham, T. E.; Kollman, P. E. J Mol Biol 1996, 259, 434.40. Berendsen, H. J. C.; Postma, J. P. M.; van Gunsteren, W. F.; DiNola,

A.; Haak, J. R. J Comput Phys 1984, 81, 3684.41. Rychaert, J. P.; Ciccotti, G.; Berendsen, H. J. C. J Comput Phys 1977,

23, 327.42. Darden, T. A; York, D.; Pedersen, L. G. J Chem Phys 1993, 98, 10089.43. Essmann, U.; Perera, L.; Berkowitz, M. L.; Darden, T.; Lee, H.;

Pedersen, L. G. J Chem Phys 1995, 103, 8577.44. Humphery, F. W.; Dalke, A.; Shulten, K. J Mol Graphics 1996, 14, 33.45. Lavery, R.; Sklenar,H. J Biomol Struct Dyn 1988, 6, 63.46. Lavery, R.; Sklenar, H. J Biomol Struct Dyn 1989, 6, 655.47. Ghosh, A.; Almlof, J.; Que, L. J Phys Chem 1994, 98, 5576.48. Sehlstedt, U.; Kim, S. K.; Carter, P.; Goodisman, J.; Vollano, J. F.;

Norden, B.; Dabrowiak, J. C. Biochemistry 1994, 33, 417.49. Brown, D. G.; Sanderson, M. R.; Garman, E.; Neidle, S. J Mol Biol

1992, 226, 481.50. Edwards, K. J.; Jenkins, T. C.; Neidle, S. Biochemistry 1992, 31, 7104.


Documents

Theoretical study of the interaction between a high-valent manganese porphyrin oxyl-(hydroxo)-Mn(IV)-TMPyP and double-stranded DNA