Investigations of the Binding of [Pt2(DTBPA)Cl2](II) and [Pt2(TPXA)Cl2](II) to DNA via Various Cross-Linking Modes

We have constructed models for a series of platinum-DNA adducts that represent the binding of two agents, [Pt2(DTBPA)Cl2](II) and [Pt2(TPXA)Cl2](II), to DNA via inter- and intra-strand cross-linking, and carried out molecular dynamics simulations and DNA conformational dynamics calculations. The effects of trans- and cis-configurations of the centers of these di-nuclear platinum agents, and of different bridging linkers, have been investigated on the conformational distortions of platinum-DNA adducts formed via inter- and intra-strand cross-links. The results demonstrate that the DNA conformational distortions for the various platinum-DNA adducts with differing cross-linking modes are greatly influenced by the difference between the platinum-platinum distance for the platinum agent and the platinum-bound N7–N7 distance for the DNA molecule, and by the flexibility of the bridging linkers in the platinum agent. However, the effects of trans/cis-configurations of the platinum-centers on the DNA conformational distortions in the platinum-DNA adducts depend on the inter- and intra-strand cross-linking modes. In addition, we discuss the relevance of DNA base motions, including opening, shift and roll, to the changes in the parameters of the DNA major and minor grooves caused by binding of the platinum agent.


Introduction
Platinum-based drugs, such as cisplatin, carboplatin and oxaliplatin, are agents that have been used effectively to treat various types of human cancer, including genitourinary, colorectal and non-small cell lung cancers [1][2][3][4][5]. Various investigations have revealed that DNA is most likely the biological target associated with the antitumor activity of these drugs [6,7].
The platinum at the center of these agents binds to nucleophilic sites of DNA (the N7 atoms of guanosine [G] and adenosine [A] bases), and these interactions result in distortions of the DNA conformation that are critical to the antitumor effects [8][9][10]. However, the use of these platinum-based agents has often been limited by the development of tumor resistance [11][12][13][14]. Therefore, considerable effort has been made to develop new generations of platinum agents that do not share these limitations [15][16][17]. Polynuclear platinum agents represent a novel class of antitumor metallodrugs, in which two or three platinum centers are bridged by polyamines of variable length. These agents bind to DNA by forming various long-range intra-strand and inter-strand cross-links [18][19][20]. This unique DNA-binding feature is believed to be responsible for distortion of the DNA conformation, enhanced cytotoxicity, and effective prevention of resistance to the antitumor drug [21][22][23].
The [Pt 2 (DTBPA)Cl 2 ](II) agent, with two trans-platinum centers and a large 14-membered ring linker (cyclam), and the [Pt 2 (TPXA)Cl 2 ](II) agent, with two cis-platinum centers and a small 6-membered ring linker (benzene), exhibit distinct structural features that may influence their DNA binding modes and DNA distortion properties. However, how the structures of these two platinum agents affect their interaction and binding with DNA is currently not fully understood at the molecular level. In the present work, we have used molecular dynamics simulations and DNA conformational dynamics analyses to study variations in the conformations of the [Pt 2 (DTBPA)Cl 2 ](II)-DNA and [Pt 2 (TPXA)Cl 2 ](II)-DNA adducts with different cross-linking modes. Six different DNA-adduct models for each platinum agent, with inter-/intra-strand cross-linking modes of 1-3, 1-4 and 1-5 base intervals, were investigated to examine the properties of the DNA distortions induced by the two platinum agents.

Results
The 12 platinum-DNA adduct models (six [Pt 2 (DTBPA)Cl 2 ](II)+DNA adducts: 26.28DTBPA-DNA, 4.7DTBPA-DNA, 3.7DTBPA-DNA, 7.28DTBPA-DNA, 6.28DTBPA-DNA and 7.26DTBPA-DNA; and six [Pt 2 (TPXA)Cl 2 ](II) + DNA adducts: 26.28TPXA-DNA, 4.7TPXA-DNA, 3.7TPXA-DNA, 7.28TPXA-DNA, 6.28TPXA-DNA and 7.26TPXA-DNA) were studied by performing 50-ns unrestrained molecular dynamics simulations. The root-mean-square deviation (RMSD) values for all the backbone atoms, referenced to the corresponding starting structures, over 12 trajectories for the various intra-/inter-strand cross-linking models of the platinum-DNA adduct, were examined to determine if each system had attained equilibrium. A small RMSD value for a simulation is often considered to indicate that the system is in a stable state, whereas large RMSD values suggest the occurrence of large conformational changes. Plots of the RMSDs of the 12 platinum-DNA system simulations over time are shown in Figure 1. The calculated conformational free energies of the platinum-DNA systems as a function of time, during the last 10 ns of the simulation, were analyzed [5,37]. The corresponding results for the 4.7DTBPXA-DNA, 6.28DTBPA-DNA, 4.7TPXA-DNA and 6.28TPXA-DNA models are shown in Figure A1 (a, b, c and d, respectively). It may be seen from Figures 1 and A1 that the platinum-DNA adduct systems reached equilibrium after 10 ns, and that their energies were found to be stable during the remainder of each simulation. Therefore, the trajectory analysis for the 12 systems extracted the equilibrated conformations between 10 ns and 50 ns of the simulation time, recording 20,000 snapshots at 2-ps time intervals for each trajectory. The analyses of the hydrogen bonds of the DNA base pairs and the DNA-groove parameters around the binding sites of the DNA were performed using the time-averaged structures for these models, and are shown in Table 1

Conformational Characteristics and Flexibility of [Pt 2 (DTBPA)Cl 2 ](II) and [Pt 2 (TPXA)Cl 2 ](II)
The [Pt 2 (DTBPA)Cl 2 ](II) agent consists of a large cyclam ring linker and two trans-platinum(II) centers, each of which is coordinated by one chloride and one nitrogen each from the cyclam ring acetamide and pyridyl groups. The [Pt 2 (TPXA)Cl 2 ](II) agent consists of a small benzene ring linker and two cis-platinum(II) centers, each of which is coordinated in a square-planar geometry by one chloride and three nitrogen atoms from each bis(2-pyridylmethyl)amine (BPA) moiety. The conformational characteristics of the DNA binding processes for these two platinum(II) antitumor agents are expected to differ, due to the different ligands and linkers of the platinum(II) centers [24,25]. In order to compare the conformational flexibility of [Pt 2 (DTBPA)Cl 2 ](II) and [Pt 2 (TPXA)Cl 2 ](II), the average platinum-platinum distances of these platinum-DNA adducts were evaluated before and after simulation, using the conformation overlap method. Figure 4 shows the full-overlap conformations of [Pt 2 (DTBPA)Cl 2 ](II) and [Pt 2 (TPXA)Cl 2 ](II) between the crystal structures and averaged structures during the 50-ns simulations. It may be seen from Figure 4 that the average platinum-platinum distance values for [Pt 2 (DTBPA)Cl 2 ](II) and [Pt 2 (TPXA)Cl 2 ](II) change, during simulation, from 9.70 Å and 9.07 Å in the crystal structures to 9.63 Å and 6.98 Å in the platinum-DNA adducts. Such small rangeability of the platinum-platinum distance for [Pt 2 (DTBPA)Cl 2 ](II), which has a large-sized linker, during the simulations demonstrates that it has a more rigid conformation than [Pt 2 (TPXA)Cl 2 ](II), which has a smaller-sized linker. The platinum atoms are shown as white in the crystal structures, and light cyan in the averaged structures.

Relationship between the Platinum-Platinum Distances in the Platinum(II) Agents and the N7-N7 Distances in the Pt 2 (DTBPA)-DNA and Pt 2 (TPXA)-DNA Adducts
To address the effects of the platinum-platinum distance and platinum-bound N7-N7 distance on the extent of the distortion of the DNA conformation in the platinum-DNA adduct, the overlaps of the averaged structures for intra-1,3-GG cross-linking 26.28DTBPA-DNA, intra-1,4-GG cross-linking 4.7DTBPA-DNA and intra-1,5-GG cross-linking 3.7DTBPA-DNA were compared with a bare DNA molecule ( Figure 5). The platinum-bound N7-N7 distance in the three models was used as a test parameter for the degree of the DNA conformational distortion. The platinum-bound N7-N7 distances in DNA and Pt-Pt distances in the platinum agents for the three models and for the B-DNA molecule are shown in Table 2. Of the three models, the platinum-platinum distance in the [Pt 2 (DTBPA)Cl 2 ](II) agent most closely corresponded to the platinum-bound N7-N7 distance in the intra-1,4-GG cross-linking 4.7DTBPA-DNA. It may be seen from Table 2 that the difference in the N7-N7 distance between the B-DNA molecule and the averaged platinum-bound adduct was 2.22 Å for the intra-1,3-GG cross-linking 26.28DTBPA-DNA, and 3.76 Å for the intra-1,5-GG cross-linking 3.7DTBPA-DNA adducts, but only 0.38 Å for the intra-1,4-GG cross-linking 4.7DTBPA-DNA. Moreover, as shown in Figure Table 1 and Figure 3). As would be expected, Figure 5 indicates visually that the distortion of the average DNA conformation, compared with the B-DNA molecule, was also larger for both 26.28DTBPA-DNA and 3.7DTBPA-DNA models than for 4.7DTBPA-DNA model. Therefore, large differences between the platinum-platinum distance in the platinum agent and the platinum-bound N7-N7 distance in the DNA molecule can cause substantial distortions of the DNA conformation during the process of platinum-DNA adduct formation. Similar results were observed for the [Pt 2 (TPXA)Cl 2 ](II) agent and the inter-strand cross-linking models, except for the fact that the effect on the conformationally-rigid [Pt 2 (DTBPA)Cl 2 ](II) agent was more obvious than that on [Pt 2 (TPXA)Cl 2 ](II), which has a more flexible conformation.  Table 2. Differences between the platinum-binding atoms of the base pairs (N7-N7) of bare DNA and the Pt 2 (DTBPA)-DNA adduct, before and after simulation (∆(Pt-N7)); and between the Pt-Pt distance in the [Pt 2 (DTBPA)Cl 2 ](II) agent and N7-N7 distance in bare DNA (∆(N7-N7)).

Relationship between the Flexibilities of the [Pt 2 (DTBPA)Cl 2 ](II) and [Pt 2 (TPXA)Cl 2 ](II) Agents and Their DNA Binding Interactions
Based on the fact that the [Pt 2 (DTBPA)Cl 2 ](II) agent has more rigidity than the [Pt 2 (TPXA)Cl 2 ](II) agent, a comparison was made of the DNA conformational distortions caused by the rigid [Pt 2 (DTBPA)Cl 2 ](II) agent and the flexible [Pt 2 (TPXA)Cl 2 ](II) agent, using the 4.7DTBPA-DNA and 4.7TPXA-DNA adducts. Figure 6a shows the overlaps of the averaged structures for these two models, which indicate that the distortion of the DNA conformation for the 4.7DTBPA-DNA model was larger than that for the 4.7TPXA-DNA model. The sum of the percentages of the hydrogen bonds destroyed at bases G4, A5 and A6, in comparison with the B-DNA molecule, was 202.05% for the 4.7DTBPA-DNA adduct, and only 120.28% for the 4.7TPXA-DNA adduct (see Figure 2a,c). The average deviation of the major groove width, depth and bend angle were also much larger for the 4.7DTBPA-DNA adduct ( Table 1 and Figure 3). Similar results for the other inter-/intra-strand cross-linking models were observed. Consequentially, the more rigid platinum adducts may lead to obvious distortion of the DNA backbone.

Relationship between the trans-[Pt 2 (DTBPA)Cl 2 ](II)/cis-[Pt 2 (TPXA)Cl 2 ](II) Configurations and the DNA Inter-/Intra-Strand Cross-Linking Modes
To investigate the effects of trans-[Pt 2 (DTBPA)Cl 2 ](II) and cis-[Pt 2 (TPXA)Cl 2 ](II) conformations on the distortion of the DNA conformation in the intra-/inter-strand cross-linking modes of the platinum-DNA adducts, comparisons were made of the DNA conformational distortions between the B-DNA molecule and each of the intra-1,4-GG cross-linking modes of the 4.7DTBPA-DNA and 4.7TPXA-DNA adducts, and the inter-1,4-GG cross-linking modes of the 6.28DTBPA-DNA and 6.28TPXA-DNA adducts. Figures 6b,c show the overlaps of the averaged structures for these four adducts, indicating that for the trans-type platinum agent, the distortion of the DNA conformation (in comparison with the B-DNA molecule) was larger for the inter-strand cross-linking 6.28DTBPA-DNA model than for the intra-strand cross-linking 4.7DTBPA-DNA model; in contrast, for the cis-type platinum agent, the DNA conformational distortion was larger for the intra-strand cross-linking 4.7TPXA-DNA model than for the inter-strand cross-linking 6.28TPXA-DNA model. The sum of the hydrogen bonds destroyed at the bases for trans-[Pt 2 (DTBPA)Cl 2 ](II) was 282.06% for the inter-strand cross-linking 6.28DTBPA-DNA adduct (A5, 10.91%; A6, 58.14%; G7, 48.16%; T8, 81.05%; C9, 78.43%; and A10, 5.37%), much larger than the value of 202.05% for the intra-strand cross-linking 4.7DTBPA-DNA adduct (at the G4, A5 and A6 bases) (see Figure 2a,b). In contrast, for the cis-[Pt 2 (TPXA)Cl 2 ](II) agent, the sum of the hydrogen bonds destroyed at bases G4, A5 and A6 was 120.28% for the intra-strand cross-linking 4.7TPXA-DNA adduct, much larger than the value of 55.04% at only one base (A10) for the inter-strand cross-linking 6.28TPXA-DNA adduct (see Figure 2c,d).
Similarly, for the trans-[Pt 2 (DTBPA)Cl 2 ](II) agent, the average deviations of the major groove width, depth and bend angle, in comparison with the B-DNA molecule, were larger for the inter-strand cross-linking 6.28DTBPA-DNA adduct ( Table 1 and Figure 3). Similar results were observed for other inter-/intra-strand cross-linking models of these two platinum agents. Taken together, these data suggest that the trans-type platinum agents facilitate greater DNA conformational distortion through an inter-strand rather than an intra-strand cross-linking mode, whereas the cis-type platinum agents create greater DNA conformational distortion through an intra-strand rather than an inter-strand cross-linking mode [38][39][40]. Furthermore, platinum agents with trans-type DNA cross-links have been shown to have high antitumor activity in cytotoxicity experiments [41]. The results further confirm that large differences between the platinum-platinum distance in the platinum agent and the platinum-bound N7-N7 distance in the DNA molecule can cause substantial distortions of the DNA conformation discussed above for the two [Pt 2 (DTBPA)Cl 2 ](II) and [Pt 2 (TPXA)Cl 2 ](II) agents. It can be seen from Table A1, Figures A3, A4 and A6 that the distortion of the DNA conformation for the rigid 4.7DPZM-DNA model with the deviation of DNA major groove width and depth of 9.65% and 30.06% was larger than that for the flexible 4.7DPZMCH2-DNA model with those of 5.08% and 13.35%, which confirms that the rigid platinum adduct may lead to more obvious distortion of the DNA backbone than the flexible platinum adduct. Moreover, for the trans-[Pt 2 (DPZM)Cl 2 ](II) agent, the distortion of the DNA conformation (in comparison with the B-DNA molecule) was larger for the inter-strand cross-linking 6.28DPZM-DNA model than for the intra-strand cross-linking 4.7DPZM-DNA model; in contrast, for the cis-[Pt 2 (DPZMCH2)Cl 2 ](II) agent, the DNA conformational distortion was larger for the intra-strand cross-linking 4.7DPZMCH2-DNA model than for the inter-strand cross-linking 6.28 DPZMCH2-DNA model, which further verifies that the trans-type platinum agents facilitate greater DNA conformational distortion through an inter-strand rather than an intra-strand cross-linking mode (see Table A1, Figures A3 and A6).

Principal Component Analysis of the Major Conformational Dynamics
Principal component analysis was used to investigate the trajectories from the corresponding simulations to examine the dominant dynamic motions of the DNA in the 12 Pt 2 (DTBPA)/Pt 2 (TPXA)-DNA adducts studied. The first three principal components (PC1, PC2 and PC3) described about 90% of the essential modes of the dynamics for these adducts (Table 3). It was found that the first three components of the conformational motions roughly corresponded to a superposition of bending, unwinding and twisting motions. Visual analyses of averaged DNA structures from the trajectories support the dominant motions identified by PCA. The overall centroid structures for these adducts have been analyzed, and are shown in Figure A7. It may be seen that compared with the undamaged B-DNA conformation, there were obvious differences between the various cross-linking models in the DNA conformations of the platinum-DNA adducts (including bending, unwinding and twisting motions). In particular, the PC1 occupancy time percentages for binding of [Pt 2 (DTBPA)Cl 2 ](II) to DNA were larger than those for [Pt 2 (TPXA)Cl 2 ](II), which would be expected to lead to larger DNA bending angles for the rigid [Pt 2 (DTBPA)Cl 2 ](II) agent than for the flexible [Pt 2 (TPXA)Cl 2 ](II) agent (with respect to the undamaged B-DNA molecule) (see Table 3). The unwinding motions, in the vicinity of the platinum sites, for the Pt 2 (DTBPA)/Pt 2 (TPXA)-DNA adducts also efficiently induce DNA groove conformational changes; the first three motions take place simultaneously and are irreversible.

DNA Conformational Dynamics in the Platinum-DNA Adducts
The [Pt 2 (DTBPA)Cl 2 ](II) and [Pt 2 (TPXA)Cl 2 ](II) agents disturb the DNA conformation through the covalent binding of the platinum centers to the N7 atoms of the DNA bases, and simultaneously, the locating of the corresponding large ligand of the agent to the major groove. The processes of DNA conformational dynamics result in changes in the DNA groove parameters. The frequency distributions of the DNA helical dynamics parameters have been analyzed, and are shown in Figure 7 for the 4.7DTBPA-DNA adduct, and in Figure A8 for the other 11 models of Pt 2 (DTBPA)/Pt 2 (TPXA)-DNA. In general, groove width and depth were measured mainly from the opening, twist and rise motions, and the shift motion, respectively, of the DNA base pair helixes or steps, as defined by CURVES analysis [43][44][45]. For example, for the 4.7DTBPA-DNA model (see Figure 7), the opening parameters of the DNA base pair helixes at connection sites G4:C33, A5:T32 and A6:T31 were changed ~−90°, ~−80° and ~−80°, respectively, away from the B-DNA molecule (compared with ~0° for the B-DNA molecule); the deviations in the twist angle values were ~20°, ~30° and ~−35° for the G4·A5, A5·A6 and A6·G7 base pair steps, with respect to the value of ~35° for the B-DNA molecule; and the rise value deviations were ~1.5 Å, ~−1 Å and ~5 Å for the G4·A5, A5·A6 and A6·G7 base pair steps, respectively, with respect to the value of ~3.5 Å for the B-DNA molecule. This results in a widening of the DNA major groove by 1.91 Å around base pairs G4:C33-G7:C30 (see Table 1 and Figure 8a), and the destruction of hydrogen bonds at base pairs G4:C33, A5:T32 and A6:T31 (see Figure 2a). The deviations in the shift values were ~8 Å, ~2 Å and ~−1 Å for the G4·A5, A5·A6 and A6·G7 base pair steps, respectively, compared with ~0 Å for the B-DNA molecule (see Figure 7); this results in a shoaling of the DNA major groove by 2.62 Å around base pairs G4:C33-G7:C30 (see Table 1 and Figure 8b). In addition, the DNA bend was measured by the twist, tilt and roll motions for DNA base pair steps. For example, for the 4.7DTBPA-DNA model (see Figure 7), the deviations in the values of the roll angles were ~−40°, ~−25°, and ~40° for the G4·A5, A5·A6 and A6·G7 base pair steps, with respect to ~0° for the B-DNA molecule; this contributes to a significant bending of the DNA helix toward the minor groove of 15.68°. Similar results were found for the other 11 models of Pt 2 (DTBPA)/Pt 2 (TPXA)-DNA (see Figure A9). Taken together, these data indicate that binding of the two types of platinum agent to the DNA molecule greatly affects the DNA groove parameters via dynamic changes in DNA conformation around the connected base pairs, leading to difficulties in DNA-repair as well as gene transcription.

Starting Structures
The starting structures of all the platinum-DNA adducts contained an 18-mer duplex DNA sequence, 5'-d(GAAGAAGTCACAAAATGT)-3', that was chosen on the basis of the breast/ovarian cancer-susceptibility gene, BRCA1 (Genbank #U14680) [10], and the [Pt 2 (DTBPA)Cl 2 ](II) or [Pt 2 (TPXA)Cl 2 ](II) agent selected from X-ray data [24,25]; see Chart 1. Previous studies have suggested that the two platinum center atoms may bind to the N7 atoms of the G/A bases of DNA, and might span 3-5 base pairs with intra-/inter-strand cross-links. Therefore, three types of intra-strand cross-link were constructed and simulated for [Pt 2 (DTBPA)Cl 2 ](II), with 1232 atoms: an intra-1,3-GG cross-linking mode (designated 26.28DTBPA-DNA), an intra-1,4-GG cross-linking mode (designated 4.7DTBPA-DNA), and an intra-1,5-AG cross-linking mode (designated 3.7DTBPA-DNA); and three types of inter-strand cross-link were constructed and simulated for [Pt 2 (DTBPA)Cl 2 ](II), with 1232 atoms: an inter-1,3-GG cross-linking mode (designated 7.28DTBPA-DNA), an inter-1,4-AG cross-linking mode (designated 6.28DTBPA-DNA), and an inter-1,5-GG cross-linking mode (designated 7.26DTBPA-DNA). The same protocol (with similar nomenclature) was applied to the [Pt 2 (TPXA)Cl 2 ](II) agent. Three of these 12 adduct models, 26.28TPXA-DNA, 4.7TPXA-DNA and 7.28TPXA-DNA, have been previously examined using denaturing polyacrylamide gels and MALDI-TOF MS analysis [25]. The initial coordinates of the 12 platinum(II)-DNA adducts used in our simulations were generated by docking the platinum agent in the major groove of the DNA. For each binding adduct, several different orientations of the platinum compound, coordinating with the major groove of DNA, were chosen as the starting structures for our molecular dynamics simulations, thereby providing a good test of whether our molecular dynamics simulations were capable of driving significantly distinct starting structures to a non-distinguishable one when the simulations reached equilibrium. Bump-checking was turned on to ensure that no overlapping atoms were produced during the docking process. To compare the differences between the DNA adduct conformations and an undamaged DNA molecule, a bare-DNA (B-DNA) molecule simulation was also performed, in which an idealized B-DNA was used as a starting structure for the simulation. Given that each strand of DNA has numerous phosphate groups, 32 Na + counterions were added to each system to achieve electroneutrality. The systems were explicitly solvated using the TIP3P water potential inside a box large enough to ensure that the solvent shell extended to 10 Å in all directions.

Force Field Parameter Preparation
The atoms for the studied platinum compound, excluding the platinum atoms themselves, were generated using the ANTECHAMBER module in the AMBER9 program [46]. The force field parameters and the RESP charges of the platinum compound were referenced from previous works [47][48][49][50][51]. Other force field parameters for the platinum compound were generated from the gaff force field in the AMBER9 program.

Molecular Dynamics Simulations and Trajectory Analysis
All molecular dynamics simulations were carried out using the AMBER9 package [46] with a classical AMBER parm99 [52,53] together with the parmbsc0 refinement [48,54] and gaff [55] force field parameters. Details of the molecular dynamics protocols are given in the Appendix.

Principal Component Analysis
Principal component analysis can be used to segregate large-scale correlated motions from random thermal fluctuations, thereby probing the essential dynamics of the system. Details of the analysis method are available in the Appendix.

DNA Groove Parameter Analyses of Trajectories
To investigate the distortion of the DNA, the frequency distributions (fraction of the time spent in each conformation) of the trajectories of the simulations for the models and a canonical B-DNA were calculated using the CURVES program [56]. To account for distortion of the whole DNA backbone, the overall bend, tilt and roll angles of the DNA time-averaged structures for the studied models were calculated from the CURVES outputs using MadBend [57]. Details of the calculation method are available in the Appendix.

Conclusions
Molecular dynamics simulations and DNA dynamics analyses for a series of [Pt 2 (DTBPA)Cl 2 ](II)-DNA and [Pt 2 (TPXA)Cl 2 ](II)-DNA adducts, with inter-/intra-strand cross-linking modes and 1-3, 1-4, 1-5 base intervals, were carried out to examine the distortions in the DNA double-helical structure. For each series of platinum agents, it was found that a larger difference between the platinum-platinum distance of the platinum agent and the platinum-bound N7-N7 distance of the DNA molecule was associated with a greater DNA conformational distortion. For the same cross-linking mode, the rigid [Pt 2 (DTBPA)Cl 2 ](II) agent induced a larger DNA conformational distortion than the flexible [Pt 2 (TPXA)Cl 2 ](II) agent. The trans-platinum configuration of [Pt 2 (DTBPA)Cl 2 ](II) may facilitate conformational distortion of the platinum-DNA adduct through inter-strand cross-linking modes, whereas the cis-platinum configuration of [Pt 2 (TPXA)Cl 2 ](II) may favor intra-strand cross-linking modes; this is consistent with previous studies [50][51][52]. DNA dynamics analysis revealed that the DNA base motions (opening, shifting, rising, rolling and twisting away from the DNA grooves) caused by the binding of the platinum agent induced widening of the DNA major and minor grooves and destruction of the hydrogen bonds of the DNA base pairs. These observations are validated for the other two trans-[Pt 2 (DPZM)Cl 2 ](II) and cis-[Pt 2 (DPZMCH2)Cl 2 ](II) agents binding to DNA. Our calculated results provide useful insights, in atomic detail, into how DNA conformations are affected by platinum agents containing different ligands and linkers, and may aid the experimental design of novel types of platinum-based anticancer agent.
can extract important dynamical features describing the collective and overall motions of the system [60][61][62][63]. The platinum-DNA adduct systems reached equilibrium after 10 ns; thus, the final 40 ns of the equilibrated MD simulations of the 20,000 trajectories for each of the platinum-DNA adducts was subjected to PCA analysis. The first three principal components described >90% of the essential modes of the dynamics for the platinum-adducted DNA complexes. The dynamics of the first three principal components were visualized using the Discovery Studio molecular graphics tool.

DNA Groove Parameter Analyses of the Trajectories
The PTRAJ module of the AMBER9 program was used to extract the produced conformations. These extracted snapshots were saved in the Protein Data Bank (PDB) format. Each nucleotide type was converted from the AMBER format to PDB format, and the resulting snapshots were submitted to the CURVES program [47]. The following CURVES parameters were extracted: global base-base parameters (shear, stretch, stagger, buckle, propeller and opening); and global inter-base-pair parameters (shift, slide, rise, tilt, roll and twist). Percentage occupancy distributions of the DNA helical parameters were calculated by normalizing the frequency distributions to 100%.
The overall bend, tilt and roll angles of the DNA for the averaged structures of these adducts were calculated from the CURVES output using MadBend (Department of Chemistry and Courant Institute of Mathematical Sciences, New York University and Howard Hughes Medical Institute, New York, NY, USA http://www.biomath.nyu.edu/index/software/Madbend/index.html), developed by Strahs and Schlick [48]. This method evaluates the DNA curvature by summing the projected components of local base-pair step tilt and roll angles after adjusting the helical twist. Bends in the helical axis defined by a negative roll angle indicate bending toward the minor groove, while bends defined by a positive roll angle correspond to bending toward the major groove [48].  Table A2. Differences between the platinum-binding atoms of the base pairs (N7-N7) of bare DNA and the Pt 2 (DPZM)-DNA adduct, before and after simulation (∆(Pt-N7)); and between the Pt-Pt distance in the Pt 2 (DPZM)Cl 2 agent and N7-N7 distance in bare DNA (∆(N7-N7)).