Abstract
Intercalation of small molecules between DNA base pairs affects DNA conformation, disrupting essential cellular processes, including replication, transcription, and repair. We investigated conformational changes in 18-mer DNA upon intercalation of doxorubicin, SYBR Gold, and YOYO-1 using extensive molecular dynamics simulations. Two main patterns for the intercalation were identified: RISE-type intercalation occurs between adjacent base pairs and extends the DNA helix with decreased twist angles, whereas BPEV-type intercalation proceeds through base-pair eversion without significant DNA extension. Kinetic analysis revealed that association rates for intercalation followed the order: first YO moiety (monointercalation) > SYBR Gold > doxorubicin > YOYO-1 (bis-intercalation). The free energy landscape showed that forces at DNA termini reached up to 117 pN during stretching. Notably, base pairs adjacent to intercalators were protected from strand separation, accompanied by additional helical unwinding. MM-GBSA/PBSA analysis revealed that the driving force for intercalation is the stacking energy, and the binding affinity was highest for minor-groove binding. Persistence length decreased with single-molecule binding but recovered with two molecules due to their electrostatic repulsion. Mechanical properties of intercalated DNA showed position dependence, demonstrating that multiple intercalation modes coexist in solution. The heterogeneous nature of intercalation explains why experimental measurements reflect ensemble averages, rather than single binding configurations.


Introduction
DNA intercalators are small organic molecules that are noncovalently inserted into the hydrophobic space between adjacent DNA base pairs of double-stranded DNA (dsDNA). The planar aromatic parts of the intercalators stabilize the dsDNA helix by new stacking interactions between them and adjacent base pairs. DNA intercalation is considered to affect the conformation of the dsDNA, resulting in the disruption of essential cellular processes including replication, transcription, and repair. ,
Intercalators are used for many purposes, such as DNA staining probes such as SYBR Gold (referred to as SYBR) and YOYO-1 [the oxazole yellow (YO) dimer, referred to as YOYO] and drugs such as doxorubicin (also called Adriamycin, referred to as DOXO). SYBR is a monointercalating cyanine dye. YOYO is a bis-intercalating dye that consists of two identical YO intercalative moieties connected by a linker and intercalates at two positions. DOXO, a monointercalator, is one of the anthracycline antitumor antibiotics, which are among the most effective drugs for cancer chemotherapy. ,
Intercalation can occur in two different ways: intercalation that occurs between adjacent base pairs and extends the DNA helix (we refer to this intercalation as RISE-type) and intercalation through base-pair eversion (we refer to this intercalation as BPEV-type). The latter is also known as intercalation through base-pair eversion or base-pair flipping out in molecular dynamics (MD) simulations, or insertion in an experiment.
Many single molecule measurements observed the DNA extension upon intercalation. Optical tweezer experiments and atomic force microscopy showed that intercalation of DOXO, , SYBR, , and YOYO ,, extends the DNA and unwinds the double helix by 11°, 19.1°, and 24° per molecule, respectively. However, it is difficult to analyze the precise location of the intercalation in the base pairs of the DNA at the atomic level by experiments.
Many computational calculations of the DNA with intercalators such as ethidium bromide, , copper complexes, daunomycin, − and DOXO ,, were performed to analyze the conformational changes in the adjacent base pairs at the intercalation site at the atomic level. However, comprehensive research comparing different intercalation mechanisms and their structural effects on DNA has remained limited.
To understand how the intercalation impacts DNA conformation at the atomic level, we carried out all-atom MD simulations of an 18-mer DNA with intercalators (DOXO, SYBR, and YOYO). An adaptively biased MD (ABMD) method was employed to facilitate intercalation by repeating the extension and contraction of the DNA. This generated a variety of intercalated DNA at different base pair positions and major/minor grooves. For representative conformations of the intercalated DNA, umbrella sampling simulations were carried out to obtain the free energy curves (FECs), which characterize the free energy landscape of the DNA intercalation for the extension of the DNA. The molecular mechanics Poisson–Boltzmann surface area (MM-GBSA/PBSA) approach was used to analyze the binding free energies of the intercalator–DNA complexes.
In this article, we report (1) diverse interaction modes between intercalators (DOXO, SYBR, and YOYO) and DNA base pairs, (2) free energy landscapes for DNA upon intercalation, (3) mechanical properties of intercalated DNA such as the stretch modulus and persistence length, and (4) energetic contributions to binding affinity through MM-GBSA/PBSA analysis. These findings elucidate how the intercalation mode determines the mechanical response of DNA, providing molecular-level insights into intercalator–DNA interactions.
Methods
The Force Field Parameters for the Intercalators
The force field parameters describing the SYBR, YOYO, and DOXO molecules (Figure ) were defined using the Antechamber package and the general Amber force field for organic molecules. These molecules were optimized at the HF/6-31G + level by the Gaussian 16 software. SYBR, YOYO, and DOXO were protonated so that the net charges for SYBR, YOYO, and DOXO were set at +2, +4, and +1, respectively. Discrete partial atomic charges were assigned to each atomic center in the molecules using the RESP methodology at the HF/6-31G* level.
1.

Structures of (a) DOXO, (b) SYBR, and (c) YOYO. The red group of atoms is defined as the intercalation part that intercalates into the dsDNA helix. The blue group of atoms is defined as the bulky out-of-plane substitute to distinguish whether it is located at the major or minor groove.
Construction of DNA
The initial structures of 18mer-DNA and 18mer-DNA with SYBR, DOXO, and YOYO were constructed using the nucleic acid builder module in the AMBER program (Table ). For SYBR and DOXO, which are monointercalators with relatively weak sequence specificity, we employed the 18mer-DNA sequence d(GCATGAACGAACGAACGC)2. This sequence was designed to include diverse base-step environments while maintaining overall B-DNA stability, thereby enabling a controlled comparison of intercalator-induced structural changes. In contrast, YOYO-1 is a bis-intercalator whose binding is expected to be more strongly constrained by the sequence context. Therefore, we selected the sequence d(GCGCGCGCTAGCGCGCGC)2, in which the (CGCTAGCG) motif is positioned in the center within a GC-rich environment. This is consistent with the crystal structure of a YOYO analogue (TOTO-1) bound to d(CGCTAGCG)2 (PDB code: 108d ). Details are given in the Supporting Information.
1. DNA Sequence for Intercalation.
| intercalator | DNA sequence | number of particles |
|---|---|---|
| DOXO and SYBR | d(GCATGAACGAACGAACGC)2 | 112,648 (DOXO) |
| 112,874 (SYBR) | ||
| 112,586 (free dsDNA) | ||
| YOYO | d(GCGCGCGCTAGCGCGCGC)2 | 112,646 (YOYO) |
| 112,202 (free dsDNA) |
We note that sequence selection may influence quantitative aspects of binding behavior; therefore, the present results should be interpreted within the chosen sequence contexts rather than as a comprehensive description of the sequence dependence.
MD Simulations of the Intercalator–DNA Complexes
After creating the 18mer-DNA structure, four intercalation molecules were placed around the center of the DNA, following the previous study by Galindo-Murillo and Cheatham. All simulations were carried out using an MD simulation program called GENESIS , with the AMBER bsc1 and ff99ions08 force fields for the DNA and ions, respectively. The MD simulations of the DNA and the DNAs with the intercalators were carried out in a cubic box ∼105 Å × 105 Å × 105 Å at a constant pressure of 1 bar and a temperature of 300 K. Each system comprised ∼113,000 atoms with ∼37,000 TIP3P water molecules. To neutralize the charges of the system, sodium ions were placed in the box, and then additional sodium and chloride ions were added at a concentration of 0.15 M NaCl. The summary of the procedure of the simulations is shown in Table S1. The dielectric constant used was 1.0, and the van der Waals interactions were evaluated with a cutoff radius of 9.0 Å and a pair-list distance of 11.0 Å. The PME method , was used for the electrostatic interactions with a direct space cutoff of 9.0 Å. The Bussi thermostat and barostat were utilized to control the temperature and pressure of the system. The SHAKE/RATTLE and SETTLE algorithms were used to constrain all the bond lengths involving hydrogen atoms. The Velocity Verlet integrator with a time step of 2 fs was used throughout the simulation to integrate the equations of motion. Details are given in Supporting Information.
ABMD Simulation to Generate the DNA Intercalation
It has been shown that the affinity constant of DNA intercalators increases exponentially with applied force. To facilitate the intercalation and generate a variety of conformations of intercalated DNA, the ABMD simulations of 24 replicas were independently carried out for an 18mer-DNA with intercalator molecules (Table S1) for 700 ns at a constant pressure of 1 atm and a temperature of 300 K.
The reaction coordinate d was defined as the distance between the two centers of mass of the base pairs at the ends of the DNA. The range of the reaction coordinate was set at 54.0 ≤ d ≤ 67.0 Å. High energy walls of 1000 kcal/mol were set at d = 54.0 and 67.0 Å to restrict the extension and contraction of the DNA within the range of the reaction coordinates. The resolution of the reaction coordinate, Δd, and the relaxation time for the free-energy profile, τ, were set at 1.0 Å and 1,000, 2500, 5000 ps, and infinite (conventional MD without a biasing potential), respectively. The conformation of the DNA and intercalators was stored every 1 ps for analysis.
ABMD Simulation to Estimate the Biasing Potential for the DNA Intercalation
After the ABMD simulations, six, six, and three models of distinct DNAs with the intercalator molecule at different sequences in the major or minor groove were selected for DOXO, SYBR, and YOYO, respectively. The systems in which the DNA was severely deformed due to other intercalation molecules or excessive extension of the DNA were not selected.
From the conformational ensembles in the DNA intercalation of each model, a series of 24 conformations were selected. To estimate the biasing potential for the DNA intercalation for the selected conformations in the models, another ABMD simulation was carried out for 30 ns per model. The range of the reaction coordinate for the ABMD simulations was set at 48.0 ≤ d ≤ 71.0 Å for the free dsDNA, 51.0 ≤ d ≤ 74.0 Å for intercalated DNA with DOXO and SYBR, 54.0 ≤ d ≤ 77.0 Å for intercalated DNA with one YOYO molecule, and 60.0 ≤ d ≤ 83.0 Å for intercalated DNA with two YOYO molecules. High energy potentials of 1000 kcal/mol were set at the ends of the reaction coordinate to confine the system within the range of the reaction coordinate. Δd and τ were set at 1.0 Å and 10,000 ps, respectively.
Umbrella Sampling Simulations for Refinement of the Biasing Potential and the Production Run
The umbrella potential for the i-th window used is V i (d) = k i (d – d i )2 + U bias (d), where d i is a fixed distance from d i = 48.0 to 71.0 Å for the free dsDNA, 51.0 to 74.0 Å for the intercalated DNA with one DOXO and SYBR molecule, 54.0 to 77.0 Å for the intercalated DNA with one YOYO molecule, and 60.0 to 83.0 Å for the intercalated DNA with two YOYO molecules, at the interval of 1.0 Å. k i is an arbitrary harmonic force constant, which was set at 0.20 kcal/mol/Å2. U bias is the biasing potential, which was obtained in the ABMD simulation and replaced with the new biasing potential obtained in the repeated 10 ns-umbrella sampling simulations. The WHAM algorithm was used to obtain the FEC. The last 450 ns of the trajectory was used for the analysis. In our study, the data for the window at d fix = 51.0 Å for DOXO model 1 and the window at d fix = 61.0 Å for SYBR model 1 were excluded from the analysis because of the state change from RISE-type to BPEV-type during the umbrella sampling simulations. The standard error of the mean of the free energies in the FEC was determined by calculating the FECs for five 90 ns nonoverlapping segments from 50 to 500 ns in the umbrella sampling simulations. Details are given in Supporting Information.
Conformational Entropy of the Base-Pair Steps
The conformational entropies of the base-pair steps were calculated using the quasiharmonic approximation as S conf = 0.5 k B ln det [1 + (k B Te 2/ℏ2)σ], where e is Euler’s number and ℏ is Planck’s constant divided by 2π. σ = < xx T> represents the mass-weighted covariance matrix, where x is the coordinates of the atoms in the base-pair step. Details are given in Supporting Information.
Definition of Intercalation: Two Distinct Types for Intercalation
In this study, we classified the observed intercalation mechanisms into two distinct types. The first, denoted as the RISE-type (axial-separation type), represents a distinct mechanism characterized by the axial separation of base pairs along the helical axis. The second corresponds to a base-pair eversion process (or base flipping) and is hereafter termed the BPEV-type (base-pair eversion type).
To distinguish between RISE-type and BPEV-type, Rise and Opening of the base-pair step parameters were analyzed using the DNA conformational analysis module in SCUBA based on the definition of X3DNA.
-
1.
If the value of Rise is greater than 5.0 Å at the DNA base-pair step of the i-th and j-th bases/(i + 1)-th and (j – 1)-th bases, then this conformation is a candidate for RISE-type (Figure a).
-
2.
If the value of Open is greater than 90° or less than −90° at the base pair of the i-th and j-th bases, then this conformation is a candidate for BPEV-type (Figure b).
2.

Schematic representations of (a) the RISE-type (extension of the base-pair stacking along the DNA axis) and (b) BPEV-type (opening of the base pairs toward the major or minor groove), (c) intercalator location at the major or minor groove, and (d) DNA bending. The planar aromatic part that intercalates in the dsDNA helix is shown in red. The bulky out-of-plane substituent, which is located at the major or minor groove, is shown in blue (see also Figure ).
These conditions could spontaneously happen even in a single DNA. To confirm whether intercalation occurred at the base pairs or not, the following quantities were monitored.
-
1.
The center of mass of the intercalative part and the bulky out-of-plane substituent of the intercalator are COM(intercalative) and COM(substituent), respectively.
-
2.
The centers of mass of all the base pairs were calculated as the COM (i,j) by calculating the COM of i-th and j-th bases in the base pair.
If the distance between COM(intercalative) and COM(i,j) and the distance between COM(intercalative) and COM(i + 1,j – 1) are less than 4.0 Å, then the candidate is defined as the RISE-type at the DNA base step of i-th and j-th bases/(i + 1)-th and (j – 1)-th bases.
If the distance between COM(intercalative) and COM(i – 1,j+1) and the distance between COM(intercalative) and COM(i + 1,j – 1) are less than 4.0 Å, then the candidate is defined as the BPEV-type at the base pair of i-th and j-th bases.
Definition of the Location of the Intercalators at the Major and Minor Grooves
The location of the intercalators at the major or minor groove was defined by monitoring the location of the other part from the intercalative part at the major or minor groove, as follows (see also Figure c);
-
1.
COM(RISE-type) is the COM of the i-th and j-th bases and (i + 1)-th and (j – 1)-th bases. COM(BPEV-type) is the COM of the base pair of (i – 1)-th and (j + 1)-th bases/(i + 1)-th and (j – 1)-th bases.
-
2.
INTER denotes the vector from COM(intercalative) to COM(RISE-type/BPEV-type).
-
3.
BASE denotes the vector from the COM of the ith base to the COM of the jth base.
-
4.
AXIS denotes the vector from COM(1,36) to COM(18,19) on the edge of the DNA in the system of the 18mer DNA.
Intercalation was defined to occur at the major groove when the value of (AXIS·(BASE × INTER)) was greater than 0 and at the minor groove when it was less than 0. In the case of the BPEV-type at the major and minor grooves, the base pairs flipped out toward these grooves.
The Direction of Bent DNA toward the Major and Minor Grooves
We monitored the direction of the DNA bending by comparing the direction of intercalation, INTER, and the vector from intercalator to the middle point of the DNA axis (see Figure d). If the dot product of these vectors is greater than 0, then the DNA bend is in the direction of the intercalator; if the intercalator is at the major and minor grooves, the DNA bends toward the major and minor grooves, respectively. If the dot product of these vectors is less than 0, then the DNA bends in the opposite direction of the location of the intercalator; if the intercalator is at the major and minor grooves, the DNA bends in the opposite direction.
Definition of Fraction of Intercalation
Figure S1 indicates that two intercalative molecules out of four in the system were substantially consumed to stack on the ends of the DNA, and the other two molecules can move relatively freely around the other base pairs from 2 to 17 and participate in the intercalation into the DNA. This phenomenon has been observed in other MD simulations. ,,
Consequently, the total possible number of intercalation sites was assumed to be two, and the number of intercalation sites was normalized by two. Thus, the fraction of intercalation is defined as 1.0 when two DOXO or SYBR molecules are intercalated throughout all of the 24 ABMD simulations or when four YO moieties of two YOYO molecules are intercalated throughout all of the 24 ABMD simulations.
Definition of the Occurrence Time of Stable Intercalation for the Kinetic Analysis
Due to structural fluctuations of the DNA, intercalation events are often transiently disrupted when evaluated at fine time resolution, as shown in Figure S2. Therefore, it should be noted that defining discrete intercalation events on the basis of instantaneous structural criteria can be ambiguous.
To analyze the kinetic rate of k on for the intercalation process, we defined the time of occurrence of stable intercalation. First, we monitored whether the condition of intercalation is satisfied from time = 0 to 700 ns. When the condition is satisfied at time = t, then we counted the event for 100 ns from t to t + 100 ns (or from t to the final simulation time of 700 ns if t is greater than 600 ns). If the event was observed at a rate of more than 50% in the range of 100 ns (or the same occurrence was observed after t = 600 ns), then we judged that intercalation stably occurred at time = t.
This definition excludes short-lived intercalation and identifies the stable intercalation states. As shown in Figure S2, once the system enters the intercalated state, it remains relatively stable, although intermittent intercalation persists. Accordingly, the time at which stable intercalation occurs for each trajectory was determined to calculate the kinetic rate k on through first-order exponential kinetic fitting.
Calculation of the Stretch Modulus
The extension of the DNA under a wide range of forces can be well described by an extensible worm-like chain (WLC) model
| 1 |
where d is the end-to-end distance of the DNA, L is the contour length of the DNA, F is the force applied to the DNA, L p is the persistence length, S is the stretch modulus of the molecule, k B is the Boltzmann constant, and T is the absolute temperature. At high forces, the second term of eq becomes small, allowing the approximation d = L 0 (1 + F/S), where L 0 was set at the contour length at the free energy minimum state in our study. Here, we assume that the fractional elongation for the contour length is the same as that for the end-to-end distance
| 2 |
where ΔL = L – L 0 is the extension from its original contour length, and Δd = d – d 0 is the extension from its free-energy minimum end-to-end distance. In this regime, the force–extension curve becomes relatively linear, and its slope corresponds to the stretch modulus S. Details are given in Supporting Information.
Calculation of the Persistence Length
The persistence length L p was estimated to characterize the bending rigidity of the DNA. Based on the WLC model, the bending persistence length l B of dsDNA is determined by fitting the bending angle distribution to a quadratic function of the bending angle , as follows
| 3 |
Here, p(θ, L 0) is an angular probability distribution, where θ is the bending angle, and L 0 is the average of the instantaneous contour length of , where l i is the distance between the centers of successive base pairs in a step, and N is the total number of base pairs. l i is calculated by using X3DNA. Details are given in Supporting Information.
Calculation of the Stretch Modulus from the Contour Length
Stretch modulus, S, is obtained from the contour length distribution as follows ,
| 4 |
where β = 1/k B T. From the slope of the graph of ln P(L) versus (L/L 0 – 1)2, the stretch modulus can be measured. To estimate the stretch modulus, trajectories within the elastic region in d were used in our study.
MM-GBSA/PBSA
MM-GBSA/PBSA is described as follows
| 5 |
where G is the calculated average free energy, and E vdw, E ele, and E in are the average molecular mechanical energy of the van der Waals interaction, the electrostatic interaction, and the internal energy such as the bond, angle, and torsion energy of the solute, respectively. G pol is the polar part of the solvation free energy. G nonpol is the nonpolar contribution to the solvation free energy. Salt concentration was set at 0.150 M. The entropic contribution (−TS), where S is the solute entropy and T is the temperature, was evaluated using the quasiharmonic analysis (QHA).
The binding free energy, ΔG bind, was calculated as
| 6 |
From the difference between the free energy of the DNA–intercalator complex, G complex, and the free energy of the DNA/intercalator, G DNA/intercalator, using the trajectory of the complex. Further, to account for the deformation energy of the DNA (G def), G def was calculated as the difference in energy between the free dsDNA and the intercalated DNA. The final corrected Gibbs binding free energy (ΔG corr) is given by
| 7 |
To facilitate the discussion of intercalation affinity with and without the entropy consideration, ΔG bind was first evaluated without the entropy term, which was subsequently added to G corr. In spite of the uncertainty in the accuracy of the absolute binding free energy in the MM-GBSA/PBSA method (for example, due to the possible overestimation of the stacking energy), it has often been suggested that it is quantitatively valid to discuss the relative stability of the two states from the difference in their free energies. , Details are given in Supporting Information.
Computational Time
The simulation times for (1) the ABMD simulation to generate the DNA intercalation, (2) the ABMD simulation to estimate the biasing potential for the extension of the DNA, (3) the umbrella sampling simulations to refine the biasing potential, and (4) the umbrella sampling simulation for the production run are summarized in Table . All the simulations, requiring a total time of 14.6 μs (3.1 μs for the ABMD simulations and 11.5 μs for the umbrella sampling simulations), were mainly performed on the Polaire supercomputer at Hokkaido University and the SGI8600 supercomputer at Japan Atomic Energy Agency.
2. Summary of MM-GBSA for DOXO, SYBR, and YOYO .
| intercalator | model | E def (VDW) | G def (GBTOT) | ΔG bind (GBTOT) | ΔE corr (VDW) | ΔG corr (GBTOT) | ΔG corr (GBTOT + QHA) | d fix [Å] |
|---|---|---|---|---|---|---|---|---|
| DOXO | 1 (RISE, major) | 21.0 | 10.1 | –35.0 | –27.5 | –24.9 | –29.2 | 60.0 (60.2) |
| 2 (RISE, major) | 26.0 | 21.1 | –53.0 | –27.2 | –31.9 | –34.2 | 61.0 (60.7) | |
| 3 (RISE, minor) | 23.9 | 15.5 | –58.6 | –38.7 | –43.1 | –43.3 | 61.0 (60.7) | |
| 4 (RISE, minor) | 22.8 | 19.9 | –60.9 | –40.5 | –41.0 | –41.0 | 60.0 (60.3) | |
| 5 (BPEV, major) | 29.3 | 36.0 | –59.9 | –27.9 | –24.0 | –25.5 | 57.0 (57.2) | |
| 6 (BPEV, minor) | 33.4 | 32.6 | –65.3 | –28.2 | –32.7 | –34.8 | 57.0 (57.1) | |
| SYBR | 1 (RISE, major) | 26.1 | 18.5 | –42.1 | –28.3 | –23.6 | –25.6 | 60.0 (60.3) |
| 2 (RISE, major) | 24.0 | 19.9 | –41.2 | –24.5 | –21.3 | –24.9 | 60.0 (60.1) | |
| 3 (RISE, minor) | 23.8 | 15.0 | –57.7 | –38.3 | –42.7 | –43.9 | 60.0 (60.3) | |
| 4 (RISE, minor) | 22.4 | 30.0 | –63.7 | –41.3 | –33.8 | –34.5 | 60.0 (60.0) | |
| 5 (BPEV, minor) | 30.0 | 32.3 | –60.1 | –36.0 | –22.9 | –22.9 | 56.0 (56.4) | |
| 6 (BPEV, minor) | 30.2 | 45.8 | –57.9 | –36.0 | –12.1 | –13.0 | 56.0 (55.9) | |
| YOYO | 1 (RISE, major) | 40.7 | 23.9 | –93.4 | –56.6 | –69.5 | –69.5 | 63.0 (63.0) |
| 2 (RISE, minor) | 44.3 | 31.5 | –99.6 | –50.2 | –68.1 | –70.2 | 63.0 (63.2) | |
| 3 (RISE, major/minor) | 88.4 | 80.4 | –39.1 | –102.4 | 41.3 | 41.3 | 70.0 (70.1) |
The units of the energies are kcal/mol. The values in parentheses for the column of d fix are d min.
Results
A Variety of Intercalated DNA Generated by ABMD Simulations
We considered three intercalators, DOXO, SYBR, and YOYO, and generated a variety of intercalated DNA conformations using the ABMD simulation for an 18mer-DNA with intercalator molecules for 700 ns at 1 atm pressure and 300 K (see Methods and Table S1), as well as a conventional MD simulation.
We classified the obtained conformations into four categories dependent on RISE or BPEV type at the major or minor grooves: RISE(major), RISE(minor), BPEV(major), and BPEV(minor) (Figure a,b).
3.

Fraction of intercalation for RISE-type and OPEN-type. The fraction of intercalation for RISE-type for (a1) DOXO, (b1) SYBR, and (c1) YOYO along the DNA base-pair step, and OPEN-type for (a2) DOXO and (b2) SYBR along the DNA base-pair, respectively. At each step, a bar for the average of tau = 1500 ps, 2500 ps, 5000 ps, and another bar for the conventional MD are shown from left to right. RISE-type in the major and minor grooves, and OPEN-type in the major and minor grooves are shown in blue, red, cyan, and magenta, respectively. The probability of OPEN-type at i-th and j-th base pair is shown at i-th DNA bp-step. The fraction of intercalation is normalized by 24, the number of replicas, and 2, the number of DOXO molecules that participate in intercalation (48 in total). This means that the fraction of intercalation is one if two molecules always intercalated in the DNA. The total fraction of intercalation from 1 to 17 is shown at 19 of the base-pair steps.
We defined the intercalation fraction such that 1.0 represents the state in which two of the four intercalator molecules are continuously involved in intercalation (see Methods). For DOXO, ABMD yielded significantly higher intercalation fractions compared with conventional MD: RISE(major): 5.37 × 10–2 vs 1.89 × 10–2, RISE(minor): 6.48 × 10–2 vs 0.0, BPEV(major): 1.26 × 10–2 vs 0.0, and BPEV(minor): 5.36 × 10–4 vs 0.0. Notably, BPEV-type conformations were completely absent in conventional MD (Figure a1,a2 and Table S2). Intercalation was facilitated by the ABMD simulation. RISE-type was observed 1.5 times more frequently than expected based on sequence compositions at the AC, GA, and AT steps, while it was less often at the AA, CA(=TG), and CG steps. We also analyzed the base pair composition at RISE-type sites. In the 18-mer DNA, A:T and C:G base pairs comprised 8 (44.4%) and 10 (55.6%) pairs, respectively. The observed intercalation distributions in ABMD (47.8% A:T, 52.1% C:G) and conventional MD (50.1% A:T, 49.9% C:G) simulations were consistent with this composition, indicating no significant sequence preference for either base pair type.
In contrast to the RISE-type interaction, BPEV-type at the A:T and G:C base pairs was 76.0% and 24.0% in the ABMD simulation, respectively, indicating that the A:T base pair more easily facilitates BPEV-type (Figure a2). This is probably due to the weaker hydrogen bonds in A:T pairs compared with G:C pairs. The 18-mer DNA contains three 5′-GAAC-3′ segments (base pairs 5–8, 9–12, and 13–16). BPEV-type intercalation occurred at the seventh and 11th A:T base pairs (the second A in each GAAC motif) for DOXO, and at the seventh, 11th, and 15th positions for SYBR, but was absent at the sixth, 10th, and 14th A:T base pairs (the first A in each motif). This indicates that BPEV-type intercalation preferentially occurs at A:T base pairs in the middle of 5′-AAC-3′ to 5′-GAA-3′. Notably, these BPEV-type conformations were not observed in conventional MD simulations.
For SYBR, intercalation is also facilitated by the ABMD simulation (Figure b1,b2 and Table S2). RISE-type intercalation occurred 1.5 times more frequently than expected at the AA, CA(=TG) steps, while it occurred less frequently at the GA and AT steps. Moreover, a moderate preference for A:T base pairs was observed in the SYBR intercalation. BPEV-type intercalation occurred predominantly at A:T base pairs (83.1%) compared with C:G base pairs (16.9%) in the ABMD simulation, similar to DOXO, but was not observed in conventional MD simulation. Details of the discussion with the values are in the Supporting Information.
For YOYO, bis-intercalation occurred in two steps: first, one YO moiety intercalated between DNA base pairs and, subsequently, the second YO moiety intercalated at a nearby site. Intercalation was observed 1.5 times more frequently than expected at the AG(=CT) and CG steps, while it was less frequent at the GC step (Figure c1 and Table S2). The first intercalation of the YO moieties clearly showed a strong preference for intercalating to the CG steps but little preference for intercalating to the GC steps (Figure c1). This is likely because the GC step exhibits stronger stacking interaction than the CG step, along with a greater number of interbase contacts, resulting in a more tightly packed structure and a higher energetic barrier for the initial intercalation step. We observed bis-intercalation events where the two moieties intercalated at sites separated by one base pair: specifically, at C6G7/C8T9, A10G11/C12G13, and C4G5/C6G7 steps. Notably, bis-intercalation did not occur at the C8T9/A10G11 step in the center of the 5′-CGCTAGCG-3′ sequence, despite this being the known intercalation site for TOTO in crystal structures [21]. No BPEV-type intercalation was observed for YOYO, even in the ABMD simulation. Details of the discussion with the values are given in the Supporting Information.
In summary, the ABMD simulation revealed distinct base-pair step and sequence preferences for each intercalator: RISE-type occurred preferentially at AC, GA, and AT steps for DOXO, AA and CA steps for SYBR, and AG and CG steps for YOYO. BPEV-type intercalation was predominantly observed at A:T base pairs for both DOXO and SYBR but was absent for YOYO even in ABMD simulations.
Intercalation Kinetics of DNA–Intercalator Complexes
We examined the intercalation kinetics under the DNA extension and contraction force with ABMD. Figure shows that k on for τ = 1500 ps is larger than that for τ = 2500 and 5000 ps except for YOYO (first intercalation). This suggests that more frequent extension–contraction cycles (shorter τ) generally enhance intercalation rates. Notably, k on values could not be reliably determined from conventional MD simulations due to insufficient intercalation events, except for YOYO monointercalation.
4.

Kinetics of intercalated DNA for (a) DOXO, (b) SYBR, (c) YOYO (first monointercalation), and (d) YOYO (second monointercalation). The plot shows the fraction of intercalation, exp (−k on t), with the kinetic rate of k on and the time of intercalation. The time of intercalation is defined in Methods. k on was estimated based on a simple two-state model for the reaction of an intercalator (I) with the DNA (D), i.e., [D] + [I] = [DI], where [I] has first order exponential kinetics as [I] t = [I]0 exp (−k on·t). The data in the graphs represent the intercalation events observed in the 24 MD simulations of 700 ns for each time of τ. For DOXO, the values of k on for τ = 1500, 2500, and 5000 ps with an asymptotic standard error were estimated to be 271 ± 40.8, 394 ± 62.7, and 325 ± 29.0 ns, respectively. For SYBR, they were estimated to be 204 ± 17.5, 238 ± 25.3, and 285 ± 17.2 ns–1, respectively. For the first monointercalation of a YO moiety of YOYO, the values of k on for τ = 1500, 2500, 5000 ps, and infinity were estimated to be 50.8 ± 1.38, 46.2 ± 1.11, 55.8 ± 0.60, and 162 ± 6.95 ns–1, respectively. For the second intercalation of the other YO moiety (bis-intercalation of YOYO), the values of k on for τ = 1500, 2500, and 5000 ps were estimated to be 434 ± 29.2, 807 ± 100.6, and 539 ± 82.5 ns–1, respectively.
The values of k on were calculated by converting the slope [s–1] in Figure to [M–1s–1] in Table S3 under the assumption that the concentration of intercalators in the MD systems is 2.970 mM based on the simulation box volume (see Methods). The k on values (×108 M–1s–1) for τ = 1500, 2500, and 5000 ps were 1.24, 0.86, and 1.04 (DOXO); 1.65, 1.41, and 1.18 (SYBR); and 6.63, 7.29, and 6.03 (YOYO first or monointercalation). For comparison, conventional MD yielded 2.08 × 108 M–1s–1 for YOYO monointercalation. The k on value for the monointercalation is higher than that of kon for SYBR and DOXO. This is reasonable because YOYO has two YO moieties that can independently attempt intercalation, effectively doubling the binding opportunity. However, the second intercalation step (k on values of 7.76 × 107, 4.17 × 107, and 6.25 × 107 M–1s–1 for τ = 1500, 2500, and 5000 ps, respectively) was approximately 1 order of magnitude slower than monointercalation, likely due to conformational constraints imposed by the first bound moiety and the requirement for proper spatial positioning of the linker.
Overall, our results showed that the values of kon were large in the order of the first YO moiety of YOYO (monointercalation) > SYBR > DOXO > second YO moiety of YOYO (bis-intercalation).
Free Energy Profiles of the Intercalated DNA for the Extension
We calculated free energy profiles for specific intercalated DNA models with intercalators at different base pair positions: six models each for DOXO (Figures a–c and S3A) and SYBR molecule (Figures d–f and S3B), two models for one YOYO, and two YOYO molecules (Figures g–i and S3C).
5.

Atomic models of the intercalated DNA for DOXO (a, b, and c for models 1, 3, and 5, respectively), SYBR (d, e, and f for models 1, 3, and 5, respectively), and YOYO (g, h, and i for models 1, 2, and 3, respectively). A snapshot of the conformation of the selected models during the umbrella sampling simulation is depicted. The other models and details of the conformations are given in Figure S3.
Representative conformations were selected from structurally stable DNA–intercalation complexes that preserve canonical base-pairing except at the intercalation site. These conformations were selected based on the following criteria. (1) Structures that maintained intact base-pairing during repeated extension and contraction cycles in the initial ABMD simulations (for creating a diverse set of intercalations) in the range of the reaction coordinate. (2) Structures that also preserved base-pairing in subsequent ABMD simulations (for extending the intercalated DNA) around the minimum pre-estimated FECs. By applying these criteria, we selected conformations that are robust against structural perturbations and thus represent stable intercalated states. This approach was adopted to avoid artifacts arising from transient or partially destabilized configurations. Nonintercalated ligand molecules were removed prior to the subsequent ABMD (for extending the intercalated DNA) and umbrella sampling simulations to focus on the free energy profile of the intercalated state.
First, we carried out ABMD simulations to extend and contract the intercalated DNA by ∼15 and ∼10 Å from the end-to-end distance of the intercalated DNA, respectively, and estimated preliminary FECs for extension as a function of the end-to-end distance d. These FECs were subsequently refined using umbrella sampling simulations (see Methods). Figure a–c shows FECs for RISE-type (for clarity, only models 1, 3, and 5 are shown for DOXO and SYBR; individual models are shown in Figure S4 and those with error in Figure S5). The lowest free energies were at d = 60.5 ± 0.23, 60.2 ± 0.13, 63.1 ± 0.10, and 70.1 Å for DOXO, SYBR, one-YOYO, and two-YOYO, respectively (see also Table S4), while the equilibrium end-to-end distance of free dsDNA (without intercalators) was 56.8 Å for the sequence used with DOXO/SYBR and 56.7 Å for the sequence used with YOYO (Figure d). This extension of approximately 3.4–3.7 Å per monointercalator is consistent with the insertion of one planar aromatic moiety between base pairs, corresponding to approximately one base-pair step height (∼3.2 Å). For YOYO, the extensions of 6.4 Å (one-YOYO) and 13.4 Å (two-YOYO) demonstrate additive contributions from each intercalated YOYO molecule. In contrast to RISE-type, the lowest free energies for BPEV-type occurred at d = 57.2 ± 0.05 (DOXO) and 56.2 ± 0.25 Å (SYBR), representing slight changes (0.4 Å and −0.6 Å, respectively) relative to free dsDNA. This indicates that in BPEV-type intercalation, the intercalator disrupts base stacking without significantly increasing the helical rise.
6.

FEC for the extension and contraction of the intercalated DNA. The FECs of the DNA intercalation against the end-to-end distance of the DNA, d (in units of Å), for (a) DOXO, (b) SYBR, (c) YOYO, and (d) the free dsDNA, respectively. The values of the free energies were set at zero at d = d min. For clarity, only models 1, 3, and 5 are shown for DOXO and SYBR. Individual FECs with error bars for all the models are shown in Figures S4 and S5.
Stretch Moduli Estimated from FECs
The FECs exhibit an approximately quadratic curvature near the minimum d min and approximately linear behavior at both sides. This indicates that the dynamics of the intercalated DNA are elastic around d min. The corresponding force–extension relationship, obtained as the derivative of the FECs with respect to d, is shown in Figure S6. In the quadratic region, we estimated the stretch modulus, which characterizes the elastic stiffness of the DNA (Figure a; individual model values in Figure S7a). For reference, the stretch moduli that were estimated from the contour length are shown in Figure S7d and Table S4. It should be noted that they do not necessarily match the values in our study.
7.

(a) The stretch modulus for the intercalated DNA; the stretch moduli for extension (in red) and contraction (in blue) of the intercalated DNA for DOXO, SYBR, YOYO, and the free dsDNA are shown. See also Table S4. Individual values for all the models are shown in Figure S7a. (b) Force at the first peak for the extension (in red) and contraction (in blue) and average force at the inelastic extension region (in green) and inelastic contraction region (gray) of the intercalated DNA; see also Table S4. Individual values for all the models are shown in Figure S7b. (c) The persistence lengths of the intercalated DNA; see also Table S4. Individual values for all the models are shown in Figure S7c.
Most of the intercalated models showed reduced stretch moduli compared with free dsDNA, indicating that intercalation generally increases DNA flexibility during extension. However, DOXO models 2 and 3 (both RISE-type) showed larger stretch moduli. These models had d min = 60.7 Å, compared with <60.3 Å for other DOXO models, suggesting that these DNAs were already near-maximally extended at equilibrium. Consequently, further extension would require overcoming additional structural constraints, resulting in higher apparent stiffness.
Overall, BPEV-type consistently exhibited lower stretch moduli than RISE-type for both extension and contraction, reflecting their more disrupted base-stacking arrangements. Compared with free dsDNA, the higher contraction modulus for two-YOYO indicates that two-YOYO significantly restricts DNA flexibility for contraction. The asymmetry between extension and contraction moduli reflects the different molecular mechanisms governing DNA stretching versus compression in the presence of intercalators. Details of the discussion with values are given in the Supporting Information.
Force during DNA Extension in the Elastic Regime
Figure S6 shows the force extension as a function of d. The force increased linearly from 0 pN to a first peak, marking the end of the elastic regime (Figure b). Details of the discussion with the values are given in the Supporting Information.
Notably, despite variations in stretch moduli among different models, the first peak forces remained relatively consistent within each intercalator type (DOXO and SYBR). This is because models with lower stretch moduli exhibit wider elastic ranges, while those with higher moduli show narrower ranges, such that the force × distance product at the elastic limit remains approximately constant. This suggests that intercalated DNA with lower stretch moduli can accommodate larger deformations before leaving the elastic regime, likely due to local structural rearrangements and stabilization of base pairs surrounding the intercalation site. The first peak forces for YOYO (>100 pN for both one-YOYO and two-YOYO intercalation) were notably higher than those for DOXO and SYBR (∼70–80 pN for RISE-type). This indicates that YOYO’s bis-intercalative structure with its flexible linker provides enhanced DNA stabilization compared with the monointercalators.
Force during DNA Overstretching in the Inelastic Regime
Beyond the first peak, the elastic response ceased, and forces began to fluctuate significantly. As shown in Figure S8, substantial structural deformation occurred in the inelastic region, including partial melting of the double helix structure. However, base pairs around the intercalation site remained paired throughout the overstretching process. This indicates that the intercalation sites are structurally stable and do not easily collapse during the extension. This is consistent with the experiments showing that intercalated YOYO molecules exhibit neither static reorientation nor alteration in their rotational dynamics during the force-induced DNA melting and that melting is confined to domains between adjacent YOYO molecules during the overstretching transition.
Figure S9 shows the base-pair step entropy, which is an indicator of local structural stability. As DNA extended, the entropy increased markedly at AT steps (positions 3, 6, 10, and 14), indicating that overstretching preferentially occurred at these sites. This is consistent with the experimental observation that the thermal melting of the DNA predominantly initiates at an AT step-rich sequence. The drop in force at the overstretch transition corresponds roughly to the changes in local base-pair entropy. Furthermore, Figure S8 reveals that bases can rearrange into ladder-stacking configurations (where bases from opposite strands stack parallel rather than paired), which maintain or even increase extension forces. Thus, force fluctuations arise from the dynamic interplay between base-pair melting (force drops) and the formation of alternative stacking arrangements (force recovery).
The averaged forces in the inelastic regime were estimated to be (in pN): DOXO: 60.6 ± 1.8 (RISE-type) and 53.6 ± 2.3 (BPEV-type), SYBR: 55.8 ± 3.2 (RISE-type) and 57.8 ± 1.4 (BPEV-type), and YOYO: 86.3 ± 1.6 (one-YOYO intercalation) and 96.4 (two-YOYO intercalation). Compared with free dsDNA (55.7 pN for DOXO/SYBR sequence and 65.3 pN for YOYO sequence), forces for DOXO- and SYBR-intercalated DNA remained similar to their respective free dsDNA, while YOYO-intercalated DNA showed significantly higher forces. This indicates that the intercalated DNA with YOYO is resistant to DNA overstretching. The elevated forces (>80 pN) for YOYO likely arise from the flexible linker connecting the two YO moieties, which mechanically couples the intercalated sites and resists DNA extension. This is particularly pronounced for bis-intercalated YOYO, where base pairs between the two intercalation sites are effectively clamped, requiring additional force to induce melting or structural rearrangement. These results demonstrate that while monointercalators (DOXO and SYBR) stabilize local structure without significantly affecting global overstretching mechanics, the bis-intercalator YOYO fundamentally alters DNA mechanical response in the inelastic regime.
Forces during DNA Compression and Bending
Figure b and Table S4 show the forces at the first peak during DNA contraction in the elastic regime (in piconewtons). Details of the discussion with the values are given in the Supporting Information. These contraction forces are lower in magnitude than the corresponding extension forces (78–117 pN, see previous section), indicating asymmetric mechanical responses. Importantly, the decrease in end-to-end distance d during compression results primarily from DNA bending rather than uniform helical compression, as visualized in Figure S10.
In the case of the RISE type at the major groove, the intercalator adopts a wedge-like geometry that facilitates bending away from the major groove, with the intercalator at the convex side of the bend. In the case of the RISE type at the minor groove, DNA preferentially bends toward the minor groove. The intercalator can be accommodated within the compressed minor groove, effectively stabilizing the bent conformation with the intercalator on its concave side. For the BPEV type, intercalators act as wedges that induce bending in the direction opposite to the intercalation site, regardless of groove location. This occurs because BPEV-type intercalation disrupts base-pairing without significantly extending the helix, creating an asymmetric hinge point.
In summary, intercalation not only reduces the forces required for DNA bending but also introduces directional preferences that depend on both the intercalation mode (RISE vs BPEV) and the groove location (major vs minor).
Persistence Length of Intercalated DNA
In this study, the persistence length of free dsDNA was estimated to be 61.2 ± 3.2 nm for both sequences used (DOXO/SYBR and YOYO, Figure c), which is in good agreement with previous MD simulations (∼30–60 nm ,, ) and experimental measurements (∼45–55 nm).
The persistence length of intercalated DNA with DOXO and SYBR was significantly reduced to 38.5 ± 6.9 nm (DOXO, RISE-type), 47.4 ± 5.0 nm (DOXO, BPEV-type), 34.2 ± 2.0 nm (SYBR, RISE-type), and 37.9 ± 3.0 nm (SYBR, BPEV-type), representing 37–77% of the free dsDNA value (individual model values in Figure S7c). For YOYO, one-YOYO reduced the persistence length to 40.8 ± 5.6 nm (67% of free DNA), while two-YOYO restored it to 63.2 nm, comparable to free dsDNA (61.2 nm). These results indicate that monointercalators (DOXO and SYBR) and one-YOYO increase DNA flexibility, while two-YOYO rigidifies the DNA. This will be discussed in the following subsection.
Local DNA Unwinding at Intercalation Sites
Intercalators adopt two distinct orientations relative to adjacent base pairs at intercalation sites (Table S4): parallel to the long axis of flanking base pairs or perpendicular to it. These orientations result in different local unwinding patterns. The unwinding values for all systems are summarized in Table S4 and Figure S11.
For DOXO, model 1 (parallel orientation) exhibited local DNA unwinding of −14.7° at the intercalated base step [Figure S11A(a)]. In contrast, models 2–4 (perpendicular orientation) showed a different pattern: unwinding occurred at base-pair steps adjacent to the intercalation site (−8.0 ± 2.2° per adjacent step), while the intercalated step itself was overwound by 3.2 ± 1.0° [Figure S11A(b–d)]. In the perpendicular orientation, base pairs flanking the intercalator exhibit significant buckling to wrap around the intercalator, as shown in Figures and S3. This adjacent-step unwinding pattern is consistent with the crystal structure of d(CGATCG)2 with two DOXO molecules intercalated at CG steps in perpendicular orientation (PDB code: 1d12 ). For BPEV-type models (5 and 6, parallel orientation), DNA unwinding was more pronounced at −30.4 ± 12.4° [Figure S11A(e,f)].
For SYBR, models 1 and 2 (parallel orientation) showed local DNA unwinding of −15.0 ± 1.5° at the intercalated base step [Figure S11B(a,b)], while models 3 and 4 (perpendicular orientation) showed unwinding at adjacent base-pair steps [−4.3 ± 1.4° per adjacent base-pair step, Figure S11B(c,d)]. BPEV-type models 5 and 6 (parallel orientation) showed unwinding of −27.1 ± 3.1° [Figure S11B(e,f)].
For YOYO, models 1 and 2 (parallel and perpendicular orientations) exhibited DNA unwinding of −14.5 ± 5.5° per one moiety of YOYO at the intercalated base steps [Figure S11C(a,b)]. For model 3 (two-YOYO), unwinding was −13.9 ± 5.8° per YO moiety [Figure S11C(c)], indicating additive contributions from each YOYO. This unwinding pattern is consistent with the crystal structure of d(5′-CGCTAGCG-3′)2 with TOTO (a YOYO analogue) at the minor groove in the parallel orientation. Notably, the orientation of the YO moiety at the minor groove in models 2 and 3 is inverted relative to the TOTO crystal structure, likely due to differences in DNA sequence context.
Global DNA Structural Parameters during Extension
We calculated average Rise values as a function of d (Figure ; for clarity, only models 1, 3, and 5 are shown for DOXO and SYBR; individual models are shown in Figure S12). At the free energy minima, RISE-type intercalation yielded Rise values of 3.62 ± 0.02 Å (DOXO, Figure a), 3.62 ± 0.02 Å (SYBR, Figure b), 3.79 ± 0.01 Å (one-YOYO), and 4.21 Å (two-YOYO, Figure c), compared to 3.39 Å and 3.43 Å for free dsDNA (DOXO/SYBR and YOYO sequences, respectively; Figure d). Summed across all 17 base-pair steps, these per-step increases correspond to total Rise changes of 3.91 Å, 3.91 Å, 6.12 Å, and 13.3 Å for DOXO, SYBR, one-YOYO, and two-YOYO, respectively. For BPEV-type, Rise values were 3.39 ± 0.01 Å (DOXO) and 3.35 ± 0.01 Å (SYBR), representing essentially no change (0.00 and −0.04 Å) relative to free DNA.
8.

Rise and Twist of the intercalated DNA along d. Rise and Twist of the intercalated DNA along the end-to-end distance of the DNA, d, for (a) DOXO, (b) SYBR, (c) YOYO, and (d) the free dsDNA, respectively. This value is averaged by 17 and 16 base-pair steps for RISE-type and Open-type, respectively. For clarity, only models 1, 3, and 5 are shown for DOXO and SYBR. Individual values for all the models are shown in Figure S12.
Average Twist values were also calculated as a function of d (Figure ). For RISE-type intercalation at the free energy minima, Twist values were 34.5 ± 0.2°, 34.4 ± 0.4°, 33.5 ± 0.2°, and 31.6° for DOXO, SYBR, mono-YOYO, and bis-YOYO, respectively, compared with 35.4° and 35.5° for free dsDNA (DOXO/SYBR and YOYO sequences, respectively). These correspond to per-step Twist reductions of −0.9°, −1.0°, −2.0°, and −3.9° for the four intercalated systems. Summed across all 17 base-pair steps, the total Twist changes were −15.3°, −17.0°, −34.0°, and −66.3° for DOXO, SYBR, one-YOYO and two-YOYO, respectively. This demonstrates the additive effect of two YOYO molecules. For BPEV type, Twist values were 33.2 ± 0.8° (DOXO) and 33.9 ± 0.1° (SYBR), showing per-step reductions of −2.2° and −1.5° relative to free DNA (35.4°). Across 16 base-pair steps, these correspond to total Twist changes of −35.2° and −24.0° for DOXO and SYBR, respectively.
Importantly, the relationship between Twist and end-to-end distance d was similar for free and intercalated DNA (Figure ), indicating that DNA unwinding during extension is an inherent mechanical property of B-form DNA, independent of the intercalator presence.
The total Twist changes for RISE-type intercalation were −15.3°, −17.0°, and −34.0° for DOXO, SYBR, and one-YOYO, respectively (ordered: DOXO < SYBR < YOYO). These values are in good agreement with single-molecule measurements: DOXO (−11°) < SYBR (−19.1°) < YOYO (−24°), supporting our simulation approach. Notably, BPEV-type induced greater total unwinding (−24° to −35°) than RISE-type, despite causing minimal helical extension, reflecting the fundamentally different structural perturbations of these intercalation modes.
Energetics of DNA Intercalation: MM-GBSA/PBSA Analysis: van der Waals Interactions Are the Major Driving Force for Intercalation
During intercalation, adjacent base pairs at the intercalation site must deform to accommodate the intercalator. This deformation requires energy to unstack base pairs and alter the sugar–phosphate backbone conformation for cavity formation (see eq in Methods). These deformation energies are offset by favorable stacking interactions between the intercalator and adjacent base pairs, as well as electrostatic interactions between the intercalator’s positive charges and the phosphate’s negative charges.
Table shows the contribution of the van der Waals interactions and solvation for intercalation, using the MM/GBSA and MM/PBSA methods. Details of values are given in Table S5. Here, we discuss the MM/GBSA results unless the MM/PBSA results show different trends.
For DOXO RISE-type, the deformation energy was G def(GBTOT) = 16.7 ± 4.3 kcal/mol, in which gas-phase van der Waals interaction energy (E def(VDW)) = 23.4 ± 1.8 kcal/mol (unfavorable), indicating that creating the intercalation cavity requires an energetic penalty. For BPEV-type, E def(VDW) (31.4 ± 2.0 kcal/mol) was larger than RISE-type because BPEV-type breaks two base-pair steps rather than one. Similarly, E def(ELE) (−28.3 ± 33.5 kcal/mol) was less favorable than RISE-type (−310.1 ± 44.6 kcal/mol) due to hydrogen bond breakage between the adjacent base pairs. The resulting G def(GBTOT) = 34.3 ± 1.7 kcal/mol was higher than RISE-type, indicating greater energetic cost for cavity formation.
For SYBR, RISE-type showed E def(VDW) = 24.0 ± 1.3 and G def(GBTOT) = 20.9 ± 5.6 kcal/mol, all lower than the corresponding BPEV-type values (30.1 ± 0.1 and 41.6 ± 4.3 kcal/mol, respectively), consistent with the DOXO pattern.
For one-YOYO, E def(VDW) (42.5 ± 1.8 kcal/mol) was about double that of DOXO and SYBR, reflecting the larger intercalating moiety. G def(GBTOT) = 27.7 ± 3.8 kcal/mol was comparable with SYBR RISE-type. For two-YOYO, E def(VDW) (88.4 kcal/mol) approximately doubled again, as expected for double of one-YOYO, while G def(GBTOT) = 80.4 kcal/mol more than doubled, indicating nonadditive effects from accommodating both molecules.
Next, we calculated binding free energies (ΔG corr) by adding G def to the raw binding energy (ΔG bind), which accounts for DNA deformation. For DOXO, ΔG corr(GBTOT) for RISE-type and BPEV-type were −35.2 ± 7.3 and −28.4 ± 4.4 kcal/mol, respectively. These negative values were large in magnitude because the magnitude of ΔG bind(GBTOT) was much larger than that of G def(GBTOT), indicating that the deformation energy was overcome by the favorable binding energy. These ΔG corr(GBTOT) values are close to the gas-phase van der Waals energies, ΔE corr(VDW) (−33.5 ± 6.1 kcal/mol for RISE-type and −28.1 ± 0.2 kcal/mol for BPEV-type), indicating that the van der Waals interactions are the major driving force for intercalation. The same pattern holds for SYBR and YOYO. For SYBR, ΔG corr(GBTOT) for RISE-type and BPEV-type were −30.4 ± 8.5 and −17.5 ± 5.4 kcal/mol, respectively, and the corresponding ΔE corr(VDW) were −33.1 ± 6.9 and −36.0 kcal/mol, respectively. For YOYO, ΔG corr(GBTOT) were −68.8 ± 0.7 and 41.3 kcal/mol for one-YOYO and two-YOYO, respectively, and ΔE corr(VDW) were −53.4 ± 3.2 and −102 kcal/mol, respectively.
Details of the discussion with electrostatic interactions are given in the Supporting Information.
Intercalation Preference: RISE-Type at Minor Groove > at Major Groove ≈ BPEV-Type
For DOXO and SYBR RISE-type intercalation, ΔG corr(GBTOT) values are less favorable at the major groove (DOXO/SYBR models 1 and 2) compared with the minor groove (models 3 and 4), indicating that the minor groove intercalation is thermodynamically preferred due to more favorable van der Waals interactions.
For BPEV-type intercalation, the deformation energy G def(GBTOT) was higher than that of RISE-type. However, ΔG corr(GBTOT) of BPEV-type was comparable to that of RISE-type at the major groove. Details of the discussion with the values are given in the Supporting Information. This is because the free energy of the DNA–intercalator complex, G complex(GBTOT) (in eq ), is more favorable for BPEV-type, likely due to stabilizing interactions between the intercalator and the disrupted (flipped-out) bases, which offset the higher deformation cost.
In addition to ΔG corr(GBTOT), ΔG corr(GBTOT + QHA) was analyzed by incorporating the conformational entropy of the complex using the quasiharmonic approximation (see Methods and Table S6). The entropic preference order was found to be RISE-type at the major groove > BPEV-type > RISE-type at the minor groove (except for YOYO). This trend can be attributed to differences in structural accommodation of the intercalator: tight binding in the RISE-type (at the minor groove) reduces conformational freedom, whereas more deformed binding modes exhibit increased flexibility.
Despite these differences, inclusion of the entropy term does not alter the thermodynamic preference order in the MM-GBSA analysis: RISE-type at minor groove (most favorable, lowest ΔG corr) > RISE-type at major groove ≈ BPEV-type. This is consistent with the predominance of minor-groove RISE-type conformations observed in our simulations. These results suggest that BPEV-type intercalation is thermodynamically competitive with major-groove RISE-type intercalation, despite requiring greater DNA deformation. In contrast, MM-PBSA results, including entropy, do not clearly distinguish among these intercalation modes, suggesting that additional sampling or more models may be required for a more definitive comparison. Overall, these results indicate that entropic contributions are non-negligible and play a crucial role in modulating sequence-dependent intercalation affinity.
Noncooperative Binding in Two-YOYO Intercalation
Although the van der Waals contribution for two-YOYO (ΔE(VDW) = 88.4 kcal/mol) is about double that of one-YOYO (42.5 ± 1.8 kcal/mol), the electrostatic contribution shows a striking deviation: ΔG(GBELE) = 128.7 kcal/mol for two-YOYO is far less favorable than double the one-YOYO value (−20.3 ± 5.2 kcal/mol). This indicates that the additional intercalation of YOYO is hindered not by stacking interactions, but rather by unfavorable electrostatic repulsion between the two positively charged YOYO molecules. Consequently, ΔG corr(GBTOT) of two-YOYO is substantially less favorable than twice the one-YOYO value, demonstrating that two-YOYO intercalation is thermodynamically suppressed by one-YOYO. This noncooperative binding contrasts sharply with previous MD simulations of bis-intercalating doxorubicin and daunomycin, , where ΔG corr for the second intercalation was more favorable than expected, indicating cooperative binding effects. ,
To confirm that noncooperativity arises from electrostatic repulsion between YOYO molecules, we analyzed ΔG corr for sequential intercalation steps: (1) first intercalation at the major groove and (2) the second intercalation at the minor groove and vice versa. The former yielded ΔG corr_first_major(GBTOT) = – 24.3 kcal/mol for the first step and 65.6 kcal/mol for the second, totaling ΔG corr(GBTOT) = 41.3 kcal/mol. The latter yielded −11.8 kcal/mol for the first and 53.0 kcal/mol for the second, again totaling 41.2 kcal/mol. These results demonstrate that while the first YOYO intercalation stabilizes DNA electrostatically, the second intercalation experiences reduced binding affinity due to electrostatic repulsion between the two positively charged intercalators. The asymmetry in ΔG corr(GBELE) between first and second intercalation steps directly reflects this intermolecular repulsion (Table ).
The noncooperative effect is further evidenced by DNA extension patterns: one-YOYO (Model 1) extends free DNA by 6.3 Å (from d min = 56.7 Å to 63.0 Å), while two-YOYO (Model 3) extends the already-intercalated DNA by 7.1 Å (from 63.0 Å to 70.1 Å). The greater extension required for the second intercalation (7.1 Å vs 6.3 Å) suggests increased structural strain, consistent with electrostatic repulsion forcing the two intercalators further apart.
In summary, electrostatic repulsion between positively charged YOYO molecules produces noncooperative bisintercalation, where the second binding event is thermodynamically less favorable than the first. This distinguishes YOYO from smaller intercalators such as doxorubicin that show cooperative binding, likely due to YOYO’s higher charge.
Discussion
DNA Sequence Preference of Intercalators: DOXO, SYBR, and YOYO
DOXO has been reported to prefer CG steps in solution, , yet experimental studies show comparable affinity for poly AT and poly CG sequences. , In the crystal structure of d(CGATCG)2 complexed with DOXO (PDB: 1d12 ), DOXO intercalates at the CG step in the minor groovea configuration rarely observed in our simulations. This discrepancy likely reflects differences between crystallization conditions and the solution environment modeled in our MD simulations. Importantly, intercalation can be affected not only by the adjacent base pairs, but also by those located downstream base pairs with which the DOXO amino group interacts. Kellogg et al. systematically investigated DOXO binding across 64 base-pair quartet combinations (CAxy, CGxy, TAxy, and TGxy, where x and y are A, C, G, or T) and demonstrated highly complex sequence specificity. These findings support our observation that DOXO exhibits moderate rather than strong sequence preferences.
In our analysis of ABMD-generated structures, we focused on base pairs directly at the intercalation site rather than including downstream quartets, as the latter would exponentially increase the parameter space. The quartet sequences in our models 1–4 (5′-G9A10A11C12-3′, 5′-A11C12G13A14-3′, 5′-A7C8G9A10-3′, and 5′-T23C24G25T26-3′, respectively) fall outside Kellogg et al.’s analyzed set. Nevertheless, models 3 and 4 (RISE-type at the minor groove) exhibited the lowest MM-GBSA/PBSA free energies, strongly suggesting these represent thermodynamically favorable intercalation modes.
To the best of our knowledge, no crystal structure of DNA–SYBR has been reported, nor have specific DNA sequence preferences been experimentally identified. This absence of strong sequence specificity in the literature is consistent with our simulation results, which show SYBR intercalating at multiple sequence contexts with comparable binding energies. The lack of strong sequence discrimination may contribute to SYBR’s utility as a general nucleic acid stain.
For YOYO, the crystal structure of d(5′-CGCTAGCG-3′)2 with TOTO (PDB: 108d ), a YOYO analogue, shows bis-intercalation at CT and AG steps in the minor groove. While we did not observe it at the C8T9/A10G11 steps, we did observe it at three alternative sites: C6G7/C8T9, A10G11/C12G13, and C4G5/C6G7. Notably, the AG (=CT) and CG steps where YOYO intercalates exhibit lower intrinsic Twist values in free B-DNA compared with GC steps [Figure S11C(d)]. This correlation suggests that YOYO preferentially intercalates at base-pair steps with inherently low Twist angles, which may require less energetic penalty for DNA unwinding during intercalation. In addition, such low-Twist steps can exhibit reduced stacking interactions due to decreased base overlap, especially in pyrimidine-purine steps like CG steps, thereby facilitating the separation between base-pairs and promoting intercalation. Together, these structural and energetic features make low-Twist steps particularly favorable for intercalation. The flexible linker connecting YOYO’s two YO moieties may allow adaptation to multiple sequence contexts while maintaining this preference for low-Twist and weakly stacked steps.
Intercalation Kinetics and Comparison with Experimental Data
Our calculated k on for YOYO bis-intercalation was substantially lower than for monointercalation by DOXO and SYBR, consistent with experimental measurements (Table S3): YOYO (3.3 × 105 M–1s–1 ) shows relatively slow association kinetics as compared with SYBR (2.1 × 106 – 1.2 × 107 M–1s–1 ) and DOXO (7.0 × 106 M–1s–1 ). The experimental k on for YO (3.2 × 107 M–1s–1) is about 100-fold faster than for YOYO (3.3 × 105 M–1s–1 ), indicating that the first YO moiety’s intercalation does not facilitate the second moiety’s binding. The linker connecting the two YO moieties likely restricts conformational freedom for achieving optimal positioning of the second moiety for intercalation. This geometric constraint is compounded by electrostatic repulsion between the positively charged moieties. This analysis of the noncooperative YO moiety’s intercalation is similar to our energetic analysis, which showed noncooperative binding in YOYO intercalation. Our simulations reproduce the experimental kinetic order: mono-YO > SYBR > DOXO > bis-YOYO, supporting the validity of our ABMD approach for capturing relative intercalation dynamics.
While our simulations correctly reproduce the qualitative kinetic order, the calculated k on values are 1–3 orders of magnitude faster than experimental measurements. Several factors contribute to this discrepancy. First, the small simulation box with periodic boundary conditions creates effective intercalator concentrations (2.970 mM) several orders of magnitude exceeding experimental conditions (typically μM–mM range), artificially accelerating encounter rates. Second, measured kinetics are highly sensitive to DNA length, intercalator concentration, salt type and concentration, pH, and solvent conditions, ,,, making direct quantitative comparison challenging. Finally, our operational definition of the intercalated state (see Methods) necessarily differs from experimental criteria. A rigorous comparison would require free energy barrier calculations and transition state analysis between groove-bound and intercalated states, , as well as substantially longer simulations to achieve full equilibration of intercalation, which lies beyond the scope of this study. Therefore, the calculated k on values should not be interpreted as quantitatively comparable to experimental rates. Instead, they provide a relative measure of intercalation kinetics under consistent simulation conditions. The comparative trends among different intercalators remain meaningful because all systems were simulated within the same computational system.
Despite these quantitative discrepancies, the consistent order of relative rates across all three intercalatorsand the correct prediction that YOYO bis-intercalation is slowestdemonstrates that our ABMD approach successfully captures the fundamental physical factors governing intercalation kinetics. The simulations provide mechanistic insights into why these kinetic differences arise, particularly the role of molecular size, charge distribution, and linker constraints in determining association rates.
Unwinding of the DNA by Intercalation
Intercalation induces DNA unwinding through two distinct mechanisms: unwinding at the intercalation site itself or unwinding at adjacent base-pair steps. The mechanism depends on the intercalator orientation. When DOXO and SYBR orient parallel to the long axis of flanking base pairs, unwinding occurs primarily at the intercalation site. When oriented perpendicular, unwinding occurs at adjacent base-pair steps instead (for YOYO, unwinding at the intercalation site predominates regardless of orientation).
Parallel orientation optimizes π–π stacking interactions between the intercalative part of the intercalator and adjacent base pairs, primarily through van der Waals interactions in the current force field. This concentrates on the structural perturbation at the intercalation site. Perpendicular orientation, in contrast, distributes the structural strain to adjacent steps, possibly to minimize backbone distortion at the intercalation site itself. The observed orientation and unwinding pattern likely reflect an energetic balance between maximizing base-intercalator stacking and minimizing backbone strain.
Increased fluctuations in base-pair step parameters at intercalation sites (Figure S11) indicate enhanced local structural flexibility, which expands the elastic range of d. When buckling occurs, this structural perturbation propagates to induce conformational changes in the neighboring base pairs.
Possible Interconversion between RISE-Type and BPEV-Type Conformations
Our MM-GBSA/PBSA analysis revealed comparable binding free energies for RISE-type at the major groove and BPEV-type conformations, suggesting thermodynamic accessibility of both states. Direct evidence for this interconversion appeared when DOXO model 1 and SYBR model 1 for the windows at d fix = 61.0 Å and 51.0 Å, respectively, spontaneously transitioned from RISE-type to BPEV-type during umbrella sampling, demonstrating that the energy barrier between these states is surmountable at physiological temperatures (these data were excluded from the umbrella sampling analysis). BPEV-type conformations have been observed in other MD simulations, , and the BPEV-type has been suggested as a transition state to the RISE-type in the intercalation. Together, these observations suggest a dynamic intercalation model in which BPEV-type and RISE-type conformations represent interconvertible states rather than distinct end points, with transitions governed by local energetic and structural factors.
Persistence Length Compared with Experiments
Recent single-molecule force–extension experiments on intercalated DNA employ low forces (<10 pN) to maintain constant intercalator occupancy during measurements. These experiments reveal that persistence length for SYBR and YOYO initially decreases at low fractional elongation (defined as the end-to-end extension normalized by the contour length) around 0.1 and then recovers toward free dsDNA values at higher elongations.
For SYBR, experimental measurements show that persistence length decreases at concentrations up to 0.16 μM and then increases at higher concentrations. In the low-concentration regime (0–0.16 μM), the contour length of 7.9-kb DNA increases from 2.4 to 2.9 μm, corresponding to a fractional elongation of (2.9–2.4)/2.4 = 0.208. In our simulation of 18-mer DNA with one SYBR molecule, the fractional elongation is estimated to be 1/17 = 0.059 (where 1 represents the extension from one intercalation event and 17 is the number of base-pair steps). As this value falls within the range of decreased persistence length in our simulation, our observation of reduced persistence length is consistent with experimental trends.
For YOYO, experiments similarly show decreased persistence length at low concentrations, followed by recovery to free DNA levels at higher concentrations. , Specifically, persistence length is minimized at a staining ratio (dye/bp) of ∼0.1 (fractional elongation ∼0.06) and recovers at a ratio of ∼0.2 (fractional elongation ∼0.35). Our simulations of 18mer-DNA yield fractional elongations of approximately 2/17 = 0.12 for one-YOYO and 4/17 = 0.24 for two-YOYO. These values bracket the experimental transition point, and our results, which show decreased persistence length for one-YOYO and recovery for two-YOYO, agree well with experimental data.
The nonmonotonic persistence length behavior observed for SYBR (+2 charge) and YOYO (+4 charge) likely arises from competing effects. At low intercalator concentrations, local structural disruption at intercalation sites introduces flexibility, decreasing the persistence length. At high concentrations, electrostatic repulsion between cationic intercalators distributed along the DNA produces a self-stiffening effect, mechanically restraining bending fluctuations and recovering the rigidity. However, the observation that the persistence length plateaus beyond a staining ratio of 0.4 despite stronger repulsion suggests that additional factors may operate at high concentrations such as changes in binding mode or saturation effects that limit further intercalation.
Unlike SYBR and YOYO, DOXO (+1 charge) shows monotonically increasing persistence length until fractional elongation of ∼0.23 in optical tweezer experiments. Previous MD simulations of DNA–daunomycin and DNA–DOXO complexes show that the second intercalation is more favorable than the first (binding free energy for two molecules is less than double that for one). This cooperativity promotes binding of multiple intercalators to DNA, resulting in uniform distribution along the helix. This accumulation progressively straightens the DNA helix and increases its persistence length. Testing this hypothesis would require extended simulations with longer DNA and multiple DOXO molecules to directly observe cooperative binding effects on global DNA structure.
Role of Linker Structure in Bis-Intercalators
It is helpful to understand YOYO’s mechanical behavior by comparing it with other bis-intercalators with different linker structures. Ditercalinium, another bis-intercalator with +4 charge, exhibits monotonically increasing persistence length, contrasting with YOYO’s plateau at high concentrations. The key structural difference lies in their linkers: YOYO possesses a long, highly flexible alkyl linker with quaternary amines and multiple rotatable bonds, whereas ditercalinium has a rigid linker comprising two piperidine rings. The rigid linker of ditercalinium, combined with its highly positive charge, likely prevents the conformational flexibility that would allow DNA bending, resulting in monotonic stiffening with increasing concentration.
Thiocoraline, an uncharged bis-intercalator, provides further evidence for the importance of linker rigidity: it increases stretch modulus while maintaining constant persistence length at high concentrations. Since thiocoraline is uncharged, its mechanical effects cannot arise from electrostatic repulsion. Like ditercalinium, thiocoraline possesses a rigid linkera large macrocycle with a disulfide bridge. This rigid linker likely constrains DNA conformational freedom, increasing axial stiffness (stretch modulus) while preventing the bending flexibility changes observed with YOYO’s flexible linker.
These comparisons demonstrate that linker structure is a critical determinant of bis-intercalator effects on DNA mechanics. Our simulations directly support this conclusion: we observed YOYO bis-intercalation with variable spacing between moietieseither one or three base pairs separating the two intercalation sites (Figure S13). This variability demonstrates that YOYO’s flexible linker can accommodate multiple binding configurations, enabling DNA bending motions that rigid linkers such as those in ditercalinium and thiocoraline would prohibit. Thus, YOYO’s flexible linker enables the nonmonotonic persistence length behavior we observed: at low concentrations, local disruption dominates; at high concentrations, electrostatic repulsion provides some stiffening, but the flexible linker cannot enforce the rigid DNA conformations imposed by ditercalinium or thiocoraline. This explains why YOYO’s persistence length plateaus rather than continuing to increase at high concentrations.
Additional Factors Affecting DNA Mechanical Properties
We examined the stretch modulus and persistence length of the DNA. We then considered factors such as base stacking, electrostatic interactions, and linker properties, which are thought to influence the persistence length or stretch modulus. However, several additional factors should be considered for a complete understanding of the intercalated DNA mechanics.
Our simulations employed physiological salt concentrations (150 mM NaCl), but the experimental conditions often vary. Ionic strength has complex, opposing effects on DNA mechanics: increasing ionic strength increases the stretch modulus while decreasing persistence length. MD simulations of free dsDNA reproduce these trends, suggesting that intercalated DNA likely shows a similar ionic-strength dependence. The balance between local disruption and electrostatic stiffening effects observed for YOYO and SYBR may shift significantly at different salt concentrations.
The binding mode may change with the intercalator concentration. The plateau in persistence length observed for YOYO , and SYBR at high concentrations may reflect changes in binding mode rather than simple saturation. For YOYO, proposals include a shift to major groove binding at high concentrations or preferential monointercalation over bis-intercalation. Similarly, DOXO may adopt alternative binding modes at high concentrations.
Our identification of BPEV-type conformations adds another layer of complexity: BPEV-type intercalation disrupts base-pairing without substantial DNA extension, potentially contributing to the mechanical properties observed at high intercalator concentrations. Future experimental interpretation should consider the possibility of mixed populations, including RISE-type, BPEV-type, and groove-bound states, each contributing differently to the overall DNA mechanics.
Computational Limitations
The definition of the intercalation state for kinetic analysis (see Methods) is inherently heuristic and depends on geometric criteria such as base-pair separation, stacking interactions, and intercalator orientation. As a result, the calculated intercalation fractions and kinetic rates are sensitive to the definition. Future work will focus on refining the definition, improving sampling with enhanced methods, and validating the results against those of experiments.
We also note that the selection of the representative configurations for the free-energy profiles and MM-GBSA/PBSA analysis may introduce a bias toward more stable conformations, which could influence the intercalation preferences estimated by these methods. More extensive sampling with a more diverse set of representative conformations will be necessary to obtain more robust quantitative estimates of the intercalation preferences.
Conclusion
The elastic properties of DNA, such as the stretch modulus and persistence length, play crucial roles in biological processes, including DNA packaging, replication, and transcription. Understanding how intercalators, including anticancer drugs, modulate these mechanical properties is essential for rational drug design and for interpreting single-molecule experiments.
Our MM-GBSA/PBSA analysis demonstrates that van der Waals (stacking) interactions drive intercalation. We identified two distinct modes: RISE-type, which extends the DNA contour length (∼3.4 Å per intercalator), and BPEV-type, which disrupts base-pairing without extending the DNA contour length. For DOXO and SYBR, minor groove binding is preferred over major groove binding, with the stability order: RISE-type (minor groove) > RISE-type (major groove) ≈ BPEV-type. Our kinetic predictions (mono-YO > SYBR > DOXO > bis-YOYO) match experimental trends, validating the ABMD approach.
RISE-type intercalation increased the DNA helical rise compared with free dsDNA, while BPEV-type intercalation showed essentially no change in rise relative to free DNA. Both intercalation modes induced DNA unwinding, with BPEV-type causing substantially greater unwinding than RISE-type, despite minimal helical extension. This reflects the distinct mechanisms of BPEV-type: disruption of base-pairing and stacking without increasing helical rise. Our calculated unwinding angles for RISE-type followed the order DOXO < SYBR < YOYO, in good agreement with experimental measurements.
Intercalation affects the persistence length through competing mechanisms: local structural disruption decreases stiffness at low concentrations, while electrostatic repulsion between cationic intercalators increases stiffness at high concentrations. This balance explains the nonmonotonic concentration dependence observed for SYBR (+2) and YOYO (+4). For YOYO, we observed noncooperative binding with the second intercalation electrostatically disfavored, unlike smaller intercalators. YOYO’s flexible linker enables conformational adaptability, unlike the rigid linkers of ditercalinium and thiocoraline.
All three intercalators exhibit moderate context-dependent sequence preferences. DOXO shows preferences based on base-pair composition, YOYO favors low-twist steps, while SYBR displays the broadest sequence tolerance. These preferences align well with their applications: DOXO as a targeted chemotherapeutic agent, YOYO as a fluorescent visualization tool, and SYBR as a general nucleic-acid stain.
This study provides a foundation for understanding DNA–intercalator mechanics by identifying distinct binding modes, elucidating the effects of linker flexibility, and demonstrating noncooperative binding in highly charged systemsinsights essential for future drug design and nanotechnology applications.
Supplementary Material
Acknowledgments
We would like to thank Yvonne Ishida for reading this manuscript carefully. This work was supported by the HPCI system provided by Hokkaido University (Project ID: hp220012 and hp230008) and JSPS KAKENHI (23K05726) [H.I.] and partially supported by the MEXT Quantum Leap Flagship Program (MEXT Q-LEAP, JPMXS0120330644), Research Support Project for Life Science and Drug Discovery (Basis for Supporting Innovative Drug Discovery and Life Science Research (BINDS)) from AMED (grant number JP26ama121024), and JSPS KAKENHI (26K00678) [H.K.]
Relevant conformations and trajectory files have been submitted to the Biological Structure Model Archive (BSM-Arc) under BSM-ID BSM00093 (https://bsma.pdbj.org/entry/93).
The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acs.jpcb.6c02297.
Detailed information about the methods for construction of DNA, MD simulations of the intercalator–DNA complexes, umbrella sampling simulations, conformational entropy of the base-pair steps, calculation of the stretch modulus, calculation of the persistence length, MM-GBSA/PBSA, details information about results for a variety of intercalated DNA generated by ABMD simulations, stretch moduli estimated from FECs, force during DNA extension in the elastic regime, and forces during DNA compression and bending (PDF)
H.I. and H.K. designed the research. H.I. carried out all simulations, analyzed the data, and wrote the article. H.K. assisted in analyzing the data and writing the article.
The authors declare no competing financial interest.
References
- Gago F.. Stacking interactions and intercalative DNA binding. Methods. 1998;14(3):277–292. doi: 10.1006/meth.1998.0584. [DOI] [PubMed] [Google Scholar]
- Mattioli R., Ilari A., Colotti B., Mosca L., Fazi F., Colotti G.. Doxorubicin and other anthracyclines in cancers: Activity, chemoresistance and its overcoming. Mol. Aspects Med. 2023;93:101205. doi: 10.1016/j.mam.2023.101205. [DOI] [PubMed] [Google Scholar]
- van der Zanden S. Y., Qiao X., Neefjes J.. New insights into the activities and toxicities of the old anticancer drug doxorubicin. FEBS J. 2021;288(21):6095–6111. doi: 10.1111/febs.15583. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Galindo-Murillo R., García-Ramos J. C., Ruiz-Azuara L., Cheatham T. E. 3rd, Cortés-Guzmán F.. Intercalation processes of copper complexes in DNA. Nucleic Acids Res. 2015;43(11):5364–5376. doi: 10.1093/nar/gkv467. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Lei H., Wang X., Wu C.. Early stage intercalation of doxorubicin to DNA fragments observed in molecular dynamics binding simulations. J. Mol. Graphics Modell. 2012;38:279–289. doi: 10.1016/j.jmgm.2012.05.006. [DOI] [PubMed] [Google Scholar]
- Cordier C., Pierre V. C., Barton J. K.. Insertion of a bulky rhodium complex into a DNA cytosine-cytosine mismatch: an NMR solution study. J. Am. Chem. Soc. 2007;129(40):12287–12295. doi: 10.1021/ja0739436. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Salerno D., Brogioli D., Cassina V., Turchi D., Beretta G. L., Seruggia D., Ziano R., Zunino F., Mantegazza F.. Magnetic tweezers measurements of the nanomechanical properties of DNA in the presence of drugs. Nucleic Acids Res. 2010;38(20):7089–7099. doi: 10.1093/nar/gkq597. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Silva E. F., Bazoni R. F., Ramos E. B., Rocha M. S.. DNA-doxorubicin interaction: New insights and peculiarities. Biopolymers. 2017;107(3):e22998. doi: 10.1002/bip.22998. [DOI] [PubMed] [Google Scholar]
- Biebricher A. S., Heller I., Roijmans R. F., Hoekstra T. P., Peterman E. J., Wuite G. J.. The impact of DNA intercalators on DNA and DNA-processing enzymes elucidated through force-dependent binding kinetics. Nat. Commun. 2015;6:7304. doi: 10.1038/ncomms8304. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Kolbeck P. J., Vanderlinden W., Gemmecker G., Gebhardt C., Lehmann M., Lak A., Nicolaus T., Cordes T., Lipfert J.. Molecular structure, DNA binding mode, photophysical properties and recommendations for use of SYBR Gold. Nucleic Acids Res. 2021;49(9):5143–5158. doi: 10.1093/nar/gkab265. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Kundukad B., Yan J., Doyle P. S.. Effect of YOYO-1 on the mechanical properties of DNA. Soft Matter. 2014;10(48):9721–9728. doi: 10.1039/C4SM02025A. [DOI] [PubMed] [Google Scholar]
- Günther K., Mertig M., Seidel R.. Mechanical and structural properties of YOYO-1 complexed DNA. Nucleic Acids Res. 2010;38(19):6526–6532. doi: 10.1093/nar/gkq434. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Galindo-Murillo R., Cheatham T. E.. Ethidium bromide interactions with DNA: an exploration of a classic DNA-ligand complex with unbiased molecular dynamics simulations. Nucleic Acids Res. 2021;49(7):3735–3747. doi: 10.1093/nar/gkab143. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Sahoo A. K., Bagchi B., Maiti P. K.. Understanding enhanced mechanical stability of DNA in the presence of intercalated anticancer drug: Implications for DNA associated processes. J. Chem. Phys. 2019;151(16):164902. doi: 10.1063/1.5117163. [DOI] [PubMed] [Google Scholar]
- Mukherjee A., Lavery R., Bagchi B., Hynes J. T.. On the molecular mechanism of drug intercalation into DNA: a simulation study of the intercalation pathway, free energy, and DNA structural changes. J. Am. Chem. Soc. 2008;130(30):9747–9755. doi: 10.1021/ja8001666. [DOI] [PubMed] [Google Scholar]
- Wilhelm M., Mukherjee A., Bouvier B., Zakrzewska K., Hynes J. T., Lavery R.. Multistep drug intercalation: molecular dynamics and free energy studies of the binding of daunomycin to DNA. J. Am. Chem. Soc. 2012;134(20):8588–8596. doi: 10.1021/ja301649k. [DOI] [PubMed] [Google Scholar]
- Trieb M., Rauch C., Wibowo F. R., Wellenzohn B., Liedl K. R.. Cooperative effects on the formation of intercalation sites. Nucleic Acids Res. 2004;32(15):4696–4703. doi: 10.1093/nar/gkh788. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Jawad B., Poudel L., Podgornik R., Ching W. Y.. Thermodynamic Dissection of the Intercalation Binding Process of Doxorubicin to dsDNA with Implications of Ionic and Solvent Effects. J. Phys. Chem. B. 2020;124(36):7803–7818. doi: 10.1021/acs.jpcb.0c05840. [DOI] [PubMed] [Google Scholar]
- Jawad B., Poudel L., Podgornik R., Steinmetz N. F., Ching W. Y.. Molecular mechanism and binding free energy of doxorubicin intercalation in DNA. Phys. Chem. Chem. Phys. 2019;21(7):3877–3893. doi: 10.1039/C8CP06776G. [DOI] [PubMed] [Google Scholar]
- Wang J., Wang W., Kollman P. A., Case D. A.. Automatic atom type and bond type perception in molecular mechanical calculations. J. Mol. Graphics Modell. 2006;25(2):247–260. doi: 10.1016/j.jmgm.2005.12.005. [DOI] [PubMed] [Google Scholar]
- Wang J., Wolf R. M., Caldwell J. W., Kollman P. A., Case D. A.. Development and testing of a general amber force field. J. Comput. Chem. 2004;25(9):1157–1174. doi: 10.1002/jcc.20035. [DOI] [PubMed] [Google Scholar]
- Frisch, M. J. ; Trucks, G. W. ; Schlegel, H. B. ; Scuseria, G. E. ; Robb, M. A. ; Cheeseman, J. R. ; Scalmani, G. ; Barone, V. ; Petersson, G. A. ; Nakatsuji, H. ; Li, X. ; Caricato, M. ; Marenich, A. V. ; Bloino, J. ; Janesko, B. G. ; Gomperts, R. ; Mennucci, B. ; Hratchian, H. P. ; Ortiz, J. V. ; Izmaylov, A. F. ; Sonnenberg, J. L. ; Williams; Ding, F. ; Lipparini, F. ; Egidi, F. ; Goings, J. ; Peng, B. ; Petrone, A. ; Henderson, T. ; Ranasinghe, D. ; Zakrzewski, V. G. ; Gao, J. ; Rega, N. ; Zheng, G. ; Liang, W. ; Hada, M. ; Ehara, M. ; Toyota, K. ; Fukuda, R. ; Hasegawa, J. ; Ishida, M. ; Nakajima, T. ; Honda, Y. ; Kitao, O. ; Nakai, H. ; Vreven, T. ; Throssell, K. ; Montgomery, J. A., Jr. ; Peralta, J. E. ; Ogliaro, F. ; Bearpark, M. J. ; Heyd, J. J. ; Brothers, E. N. ; Kudin, K. N. ; Staroverov, V. N. ; Keith, T. A. ; Kobayashi, R. ; Normand, J. ; Raghavachari, K. ; Rendell, A. P. ; Burant, J. C. ; Iyengar, S. S. ; Tomasi, J. ; Cossi, M. ; Millam, J. M. ; Klene, M. ; Adamo, C. ; Cammi, R. ; Ochterski, J. W. ; Martin, R. L. ; Morokuma, K. ; Farkas, O. ; Foresman, J. B. ; Fox, D. J. . Gaussian 16, Rev. C.01: Wallingford, CT, 2016. [Google Scholar]
- Bayly C. I., Cieplak P., Cornell W., Kollman P. A.. A well-behaved electrostatic potential based method using charge restraints for deriving atomic charges: the RESP model. J. Phys. Chem. 1993;97(40):10269–10280. doi: 10.1021/j100142a004. [DOI] [Google Scholar]
- Case D. A., Aktulga H. M., Belfon K., Cerutti D. S., Cisneros G. A., Cruzeiro V. W. D., Forouzesh N., Giese T. J., Götz A. W., Gohlke H., Izadi S., Kasavajhala K., Kaymak M. C., King E., Kurtzman T., Lee T.-S., Li P., Liu J., Luchko T., Luo R., Manathunga M., Machado M. R., Nguyen H. M., O’Hearn K. A., Onufriev A. V., Pan F., Pantano S., Qi R., Rahnamoun A., Risheh A., Schott-Verdugo S., Shajan A., Swails J., Wang J., Wei H., Wu X., Wu Y., Zhang S., Zhao S., Zhu Q., Cheatham T. E. III, Roe D. R., Roitberg A., Simmerling C., York D. M., Nagan M. C., Merz K. M. Jr.. AmberTools. J. Chem. Inf. Model. 2023;63(20):6183–6191. doi: 10.1021/acs.jcim.3c01153. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Galindo-Murillo R., Roe D. R., Cheatham T. E. 3rd. Convergence and reproducibility in molecular dynamics simulations of the DNA duplex d(GCACGAACGAACGAACGC) Biochim. Biophys. Acta Gen. Subj. 2015;1850(5):1041–1058. doi: 10.1016/j.bbagen.2014.09.007. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Spielmann H. P., Wemmer D. E., Jacobsen J. P.. Solution structure of a DNA complex with the fluorescent bis-intercalator TOTO determined by NMR spectroscopy. Biochemistry. 1995;34(27):8542–8553. doi: 10.1021/bi00027a004. [DOI] [PubMed] [Google Scholar]
- Jung J., Mori T., Kobayashi C., Matsunaga Y., Yoda T., Feig M., Sugita Y.. GENESIS: a hybrid-parallel and multi-scale molecular dynamics simulator with enhanced sampling algorithms for biomolecular and cellular simulations. Wiley Interdiscip. Rev.: Comput. Mol. Sci. 2015;5(4):310–323. doi: 10.1002/wcms.1220. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Kobayashi C., Jung J., Matsunaga Y., Mori T., Ando T., Tamura K., Kamiya M., Sugita Y.. GENESIS 1.1: A hybrid-parallel molecular dynamics simulator with enhanced sampling algorithms on multiple computational platforms. J. Comput. Chem. 2017;38(25):2193–2206. doi: 10.1002/jcc.24874. [DOI] [PubMed] [Google Scholar]
- Ivani I., Dans P. D., Noy A., Pérez A., Faustino I., Hospital A., Walther J., Andrio P., Goñi R., Balaceanu A., Portella G., Battistini F., Gelpí J. L., González C., Vendruscolo M., Laughton C. A., Harris S. A., Case D. A., Orozco M.. Parmbsc1: a refined force field for DNA simulations. Nat. Methods. 2016;13:55–58. doi: 10.1038/nmeth.3658. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Joung I. S., Cheatham T. E.. Determination of alkali and halide monovalent ion parameters for use in explicitly solvated biomolecular simulations. J. Phys. Chem. B. 2008;112:9020–9041. doi: 10.1021/jp8001614. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Jorgensen W. L., Chandrasekhar J., Madura J. D., Impey R. W., Klein M. L.. Comparison of simple potential functions for simulating liquid water. J. Chem. Phys. 1983;79:926–935. doi: 10.1063/1.445869. [DOI] [Google Scholar]
- Darden T., York D., Pedersen L.. Particle mesh Ewald: An N·log(N) method for Ewald sums in large systems. J. Chem. Phys. 1993;98(12):10089–10092. doi: 10.1063/1.464397. [DOI] [Google Scholar]
- Essmann U., Perera L., Berkowitz M. L., Darden T., Lee H., Pedersen L. G.. A smooth particle mesh Ewald method. J. Chem. Phys. 1995;103(19):8577–8593. doi: 10.1063/1.470117. [DOI] [Google Scholar]
- Bussi G., Donadio D., Parrinello M.. Canonical sampling through velocity rescaling. J. Chem. Phys. 2007;126(1):014101. doi: 10.1063/1.2408420. [DOI] [PubMed] [Google Scholar]
- Bussi G., Zykova-Timan T., Parrinello M.. Isothermal-isobaric molecular dynamics using stochastic velocity rescaling. J. Chem. Phys. 2009;130(7):074101. doi: 10.1063/1.3073889. [DOI] [PubMed] [Google Scholar]
- Ryckaert J.-P., Ciccotti G., Berendsen H. J. C.. Numerical integration of the cartesian equations of motion of a system with constraints: molecular dynamics of n-alkanes. J. Comput. Phys. 1977;23(3):327–341. doi: 10.1016/0021-9991(77)90098-5. [DOI] [Google Scholar]
- Andersen H. C.. Rattle: A “velocity” version of the shake algorithm for molecular dynamics calculations. J. Comput. Phys. 1983;52(1):24–34. doi: 10.1016/0021-9991(83)90014-1. [DOI] [Google Scholar]
- Miyamoto S., Kollman P. A.. Settle: An analytical version of the SHAKE and RATTLE algorithm for rigid water models. J. Comput. Chem. 1992;13(8):952–962. doi: 10.1002/jcc.540130805. [DOI] [Google Scholar]
- Babin V., Roland C., Sagui C.. Adaptively biased molecular dynamics of free energy calculations. J. Chem. Phys. 2008;128:134101. doi: 10.1063/1.2844595. [DOI] [PubMed] [Google Scholar]
- Kumar S., Bouzida D., Swendsen R. H., Kollman P. A., Rosenberg J. M.. The weighted histogram analysis method for free energy calculations on biomolecules. I. The method. J. Comput. Chem. 1992;13:1011–1021. doi: 10.1002/jcc.540130812. [DOI] [Google Scholar]
- Ishida H., Kono H.. Free Energy Landscape of H2A-H2B Displacement From Nucleosome. J. Mol. Biol. 2022;434(16):167707. doi: 10.1016/j.jmb.2022.167707. [DOI] [PubMed] [Google Scholar]
- Schlitter J.. Estimation of Absolute and Relative Entropies of Macromolecules Using the Covariance Matrix. Chem. Phys. Lett. 1993;215:617–621. doi: 10.1016/0009-2614(93)89366-P. [DOI] [Google Scholar]
- Ishida H., Kono H.. Torsional stress can regulate the unwrapping of two outer half superhelical turns of nucleosomal DNA. Proc. Natl. Acad. Sci. U. S. A. 2021;118(7):e2020452118. doi: 10.1073/pnas.2020452118. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Lu X. J., Olson W. K.. 3DNA: a software package for the analysis, rebuilding and visualization of three-dimensional nucleic acid structures. Nucleic Acids Res. 2003;31(17):5108–5121. doi: 10.1093/nar/gkg680. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Odijk T.. Stiff Chains and Filaments under Tension. Macromolecules. 1995;28(20):7016–7018. doi: 10.1021/ma00124a044. [DOI] [Google Scholar]
- Mogurampelly S., Nandy B., Netz R. R., Maiti P. K.. Elasticity of DNA and the effect of dendrimer binding. Eur. Phys. J. E: Soft Matter Biol. Phys. 2013;36(6):68. doi: 10.1140/epje/i2013-13068-1. [DOI] [PubMed] [Google Scholar]
- Garai A., Saurabh S., Lansac Y., Maiti P. K.. DNA Elasticity from Short DNA to Nucleosomal DNA. J. Phys. Chem. B. 2015;119(34):11146–11156. doi: 10.1021/acs.jpcb.5b03006. [DOI] [PubMed] [Google Scholar]
- Qiu D., Shenkin P. S., Hollinger F. P., Still W. C.. The GB/SA Continuum Model for Solvation. A Fast Analytical Method for the Calculation of Approximate Born Radii. J. Phys. Chem. A. 1997;101(16):3005–3014. doi: 10.1021/jp961992r. [DOI] [Google Scholar]
- Brown R. F., Andrews C. T., Elcock A. H.. Stacking Free Energies of All DNA and RNA Nucleoside Pairs and Dinucleoside-Monophosphates Computed Using Recently Revised AMBER Parameters and Compared with Experiment. J. Chem. Theory Comput. 2015;11(5):2315–2328. doi: 10.1021/ct501170h. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Ishida H., Matsumoto A.. Mechanism for verification of mismatched and homoduplex DNAs by nucleotides-bound MutS analyzed by molecular dynamics simulations. Proteins. 2016;84(9):1287–1303. doi: 10.1002/prot.25077. [DOI] [PubMed] [Google Scholar]
- Nozaki T., Onoda M., Habazaki M., Takeuchi Y., Ishida H., Sato Y., Kujirai T., Hanada K., Yamatsugu K., Kurumizaka H., Kimura H., Kono H., Kawashima S. A., Kanai M.. Designer Catalyst-Enabled Regiodivergent Histone Acetylation. J. Am. Chem. Soc. 2025;147(16):13732–13743. doi: 10.1021/jacs.5c01699. [DOI] [PubMed] [Google Scholar]
- van Mameren J., Vermeulen K., Wuite G. J. L., Peterman E. J. G.. A polarized view on DNA under tension. J. Chem. Phys. 2018;148(12):123306. doi: 10.1063/1.5004019. [DOI] [PubMed] [Google Scholar]
- Zeng Y., Montrichok A., Zocchi G.. Bubble nucleation and cooperativity in DNA melting. J. Mol. Biol. 2004;339(1):67–75. doi: 10.1016/j.jmb.2004.02.072. [DOI] [PubMed] [Google Scholar]
- Baumann C. G., Smith S. B., Bloomfield V. A., Bustamante C.. Ionic effects on the elasticity of single DNA molecules. Proc. Natl. Acad. Sci. U. S. A. 1997;94(12):6185–6190. doi: 10.1073/pnas.94.12.6185. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Frederick C. A., Williams L. D., Ughetto G., van der Marel G. A., van Boom J. H., Rich A., Wang A. H.. Structural comparison of anticancer drug-DNA complexes: adriamycin and daunomycin. Biochemistry. 1990;29(10):2538–2549. doi: 10.1021/bi00462a016. [DOI] [PubMed] [Google Scholar]
- Cullinane C., Phillips D. R.. Induction of stable transcriptional blockage sites by adriamycin: GpC specificity of apparent adriamycin-DNA adducts and dependence on iron(III) ions. Biochemistry. 1990;29(23):5638–5646. doi: 10.1021/bi00475a032. [DOI] [PubMed] [Google Scholar]
- DuVernay V. H. Jr., Pachter J. A., Crooke S. T.. Deoxyribonucleic acid binding studies on several new anthracycline antitumor antibiotics. Sequence preference and structure-activity relationships of marcellomycin and its analogues as compared to adriamycin. Biochemistry. 1979;18(18):4024–4030. doi: 10.1021/bi00585a028. [DOI] [PubMed] [Google Scholar]
- Pérez-Arnaiz C., Busto N., Leal J. M., García B.. New insights into the mechanism of the DNA/doxorubicin interaction. J. Phys. Chem. B. 2014;118(5):1288–1295. doi: 10.1021/jp411429g. [DOI] [PubMed] [Google Scholar]
- Airoldi M., Barone G., Gennaro G., Giuliani A. M., Giustini M.. Interaction of doxorubicin with polynucleotides. A spectroscopic study. Biochemistry. 2014;53(13):2197–2207. doi: 10.1021/bi401687v. [DOI] [PubMed] [Google Scholar]
- Kellogg G. E., Scarsdale J. N., Fornari F. A. Jr.. Identification and hydropathic characterization of structural features affecting sequence specificity for doxorubicin intercalation into DNA double-stranded polynucleotides. Nucleic Acids Res. 1998;26(20):4721–4732. doi: 10.1093/nar/26.20.4721. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Reuter M., Dryden D. T.. The kinetics of YOYO-1 intercalation into single molecules of double-stranded DNA. Biochem. Biophys. Res. Commun. 2010;403(2):225–229. doi: 10.1016/j.bbrc.2010.11.015. [DOI] [PubMed] [Google Scholar]
- Rizzo V., Sacchi N., Menozzi M.. Kinetic studies of anthracycline-DNA interaction by fluorescence stopped flow confirm a complex association mechanism. Biochemistry. 1989;28(1):274–282. doi: 10.1021/bi00427a038. [DOI] [PubMed] [Google Scholar]
- Berge T., Jenkins N. S., Hopkirk R. B., Waring M. J., Edwardson J. M., Henderson R. M.. Structural perturbations in DNA caused by bis-intercalation of ditercalinium visualised by atomic force microscopy. Nucleic Acids Res. 2002;30(13):2980–2986. doi: 10.1093/nar/gkf409. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Camunas-Soler J., Manosas M., Frutos S., Tulla-Puche J., Albericio F., Ritort F.. Single-molecule kinetics and footprinting of DNA bis-intercalation: the paradigmatic case of Thiocoraline. Nucleic Acids Res. 2015;43(5):2767–2779. doi: 10.1093/nar/gkv087. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Larsson A., Carlsson C., Jonsson M., Albinsson B.. Characterization of the Binding of the Fluorescent Dyes YO and YOYO to DNA by Polarized Light Spectroscopy. J. Am. Chem. Soc. 1994;116(19):8459–8465. doi: 10.1021/ja00098a004. [DOI] [Google Scholar]
- Kucharska K., Pilz M., Bielec K., Kalwarczyk T., Kuźma P., Hołyst R.. Two Intercalation Mechanisms of Oxazole Yellow Dimer (YOYO-1) into DNA. Molecules. 2021;26(12):3748. doi: 10.3390/molecules26123748. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
Relevant conformations and trajectory files have been submitted to the Biological Structure Model Archive (BSM-Arc) under BSM-ID BSM00093 (https://bsma.pdbj.org/entry/93).
