Electrical modulation properties of DNA drug molecules

Abstract

DNA drug molecules are not only widely used in gene therapy, but also play an important role in controlling the electrical properties of molecular electronics. Covalent binding, groove binding and intercalation are all important forms of drug–DNA interaction. But its applications are limited due to a lack of understanding of the electron transport mechanisms after different drug–DNA interaction modes. Here, we used a combination of density functional theory calculations and nonequilibrium Green’s function formulation with decoherence to study the effect of drug molecules on the charge transport property of DNA under three different binding modes. Conductance of DNA is found to decrease from 2.35E-5 G₀ to 1.95E-6 G₀ upon doxorubicin intercalation due to modifications of the density of states in the near-highest occupied molecular orbital region, δG = 1105.13%. Additionally, the conductance of DNA after cis-[Pt(NH3)2(py)Cl]+ covalent binding increases from 1.02E-6 G₀ to 5.25E-5 G₀, δG = 5047.06%. However, in the case of pentamidine groove binding, because there is no direct change in DNA molecular structure during drug binding, the conductance changes before and after drug binding is much smaller than in the two above cases, δG = 90.43%. Our theoretical calculations suggest that the conductance of DNA can be regulated by different drug molecules or switching the interaction modes between small molecules and DNA. This regulation opens new possibilities for their potential applications in controllable modulation of the electron transport property of DNA.

Introduction

DNA has been a subject of intense scientific research over the past 20 years for its’ great application potential in the field of molecular electronics (1–6). The interaction between DNA and other small molecules, especially drug small molecules, has been a hot topic because DNA is the carrier of genetic information and DNA drug molecules are widely used in the field of gene therapy, such as anticancer and antitumor (7–10). By studying DNA conductance, it is meaningful to find the relationship between drug concentration or drug binding efficiency and DNA conductivity, which can translate the process of drug–DNA interaction into changes in DNA conductance. More importantly, it provides a new idea for the controllable modulation of the electron transport property of DNA, which is crucial for the realization of DNA-based nanoelectronic devices (11,12). In addition to the huge biological impact of the interaction between DNA and small drug molecules, a few recent experimental studies have investigated the effect of molecular binding on DNA conductance (13,14).

Harashima et al. studied how a DNA-binding molecule, ethidium bromide (EB) or Hoechst 33258, affects the electron transport of a single DNA molecule and found that the single-molecule conductance remains almost unaffected by groove binding and was greatly affected by the intercalation combination (15). Wang et al. described a reliable single-molecule electrical biosensor for directly revealing the intrinsic effect of individual EB/ SYBR Green I intercalations on DNA charge transport based on DNA-functionalized molecular junctions, and after EB treatment, they observed an obvious decrease in device conductance (16). Simultaneously, there are also a few theoretical studies on the physics behind the DNA structural changes and the effect on the charge transport mechanism (17,18).

There are several forms of interaction between molecules and DNA; groove binding refers to the direct interaction between the drug molecule and the base pair edge of the major groove or minor groove of DNA. This binding force is the comprehensive result of hydrogen bonding, hydrophobic interaction, and π–π interaction between molecules and DNA groove bases. Intercalation binding is to embed planar or almost planar aromatic ring molecules between base pairs. The force of intercalation binding comes from the interaction between the drug molecular and DNA base and hydrophobic interaction, and it is one of the most important forms of drug molecules interacting with DNA. In addition to intercalation and groove binding, covalent forms also exist during the process of DNA interacting with small drug molecules; there is a direct valence bond connection between drug molecules and DNA, such as cisplatin. Cisplatin prevents cell proliferation by a covalent bond with DNA to form platinum (Pt)/DNA conjugate and destroy DNA structure (19). However, research on the effect of covalent binding on DNA conductance is scarce and the effects of many other drugs on DNA conductance also need to be studied. In this article, we select three kinds of drug molecules to study the changes in DNA conductance under the three different DNA drug molecular interaction modes of covalent binding, groove binding and intercalation. The drugs studied in this work are pentamidine, doxorubicin and cis-[Pt(NH3)2(py)Cl]+ (cDPCP). By employing density functional theory (DFT) calculations and nonequilibrium Green’s function formulation with decoherence, we theoretically studied the physics behind the difference in the charge transport properties of the three different interaction modes.

Results

Effect of doxorubicins intercalated on the electrical properties of DNA

We constructed and calculated four different structures to analyze the effect of intercalation on the electrical properties of DNA (Fig. 1A–D). All four structures have the same sequence. Figure 1A represents the bare B-form DNA structure and Figure 1B–D represent the complex structure after molecular intercalation. Figure 1B–D images are consistent in DNA structure except for the difference in doxorubicin number. Helical rise and twist are two useful structural parameters, which describe the translation and rotation between successive base pairs with respect to the helical axis. Curves+ is used here to get structural parameters of DNA and drug–DNA complex (20). The average helical rise of bare B-form DNA (Fig. 1A) is 3.38 Å and the average twist is 36.0°. With the intercalation of doxorubicin, the distance between DNA base pairs (Fig. 1B) increased significantly to 7.05 Å and the average twist reduced to 33.3°.

Figure 1

DNA structures under doxorubicin intercalated modes. (A) Structure of bare B-form DNA. (B) Two doxorubicins intercalated into DNA. (C) One doxorubicin intercalated into DNA. (D) No doxorubicin but the changes of DNA structure after intercalation was retained. (E) Energy levels and bandgap values for different intercalation cases, DOX represents doxorubicin.

Figure 1E shows the highest occupied molecular orbital (HOMO) and lowest unoccupied molecular orbital (LUMO) distribution of the four earlier discussed structures. After intercalation of two doxorubicins, the HOMO of the drug–DNA complex increased significantly from −4.32 eV to −3.36 eV, and the LUMO reduced from −1.50 eV to −2.10 eV. The intercalator markedly changed level distribution and induced two new HOMO and LUMO levels in the bandgap region. This change reduced the bandgap value from 2.82 eV to 1.26 eV compared to the bare double-stranded DNA. In the one intercalator case, doxorubicin also induced two new HUMO and LUMO levels in the bandgap region of DNA, but the change was smaller compared to the two intercalators case. The bandgap value reduced from 2.82 eV to 1.68 eV compared with the bare dsDNA. In the last case, we deleted the two intercalators, but there were no changes in the DNA structure (Fig. 1D). We found that in this case, the change in levels distribution is very small compared to the earlier discussed cases and the HOMO and LUMO are −4.47 eV and −1.56 eV, respectively, with a bandgap of 2.91 eV. Therefore, we conclude that the change in energy level mainly comes from the intercalation of drug molecules and the greater the number of molecules, the greater the change in energy level. There are more energy states available in the drug–intercalated DNA complex than in the bare DNA. In contrast, the change in DNA structure has little effect on DNA energy level distribution.

It is widely accepted that charge transport through DNA is dominated by the level closest to the electrodes’ chemical potentials. That is, electron transport through the DNA is indeed dominated by the HOMO (4,21,22). Therefore, we mainly investigated the conductance changes of DNA at HOMO. Figure 2A and B shows the conductance and transmission variation of DNA before and after doxorubicin intercalate, respectively. After intercalated two doxorubicins, the conductance decreases from 2.35E-5 G₀ to 1.95E-6 G₀, δG = 1105.13%. This is almost down an order of magnitude. The transmission of the two cases correlates well with the trend of the conductance properties, which decreases from 5.92E-5 to 2.23E-6 (Fig. 2B). We also plot the conductance and transmission as a function of energy including two doxorubicins, one doxorubicin and no drug molecule of the complex (Fig. 2C and D). We found that near the intercalator-induced energy levels, the transmission of one doxorubicin shows a 69.9% decrease compared to the two doxorubicins cases, and the no drug molecule case shows a 36.2% decrease compared to the two doxorubicins cases. Overall, the transmission decreased with the decrease in the number of intercalated molecules.

Figure 2

Effect of doxorubicins intercalated on the electrical properties of DNA. (A) The conductance for bare B-form DNA and two doxorubicins intercalated case, dash line is the position of HOMO energy level and the higher represents the B-form DNA, the lower represents two doxorubicins intercalated case. (B) The transmission in the vicinity of HOMO of two cases. (C) From top to bottom, curves represent the conductance for two doxorubicins, one doxorubicin, and no drug molecule of complex, respectively. (D) From top to bottom, curves represent the transmission plots for two doxorubicins, one doxorubicin, and no drug molecule of complex, respectively. Dash line represents the position of HOMO of two doxorubicins case. (E) The energy level distributions along the strand of bare B-form DNA. (F) The energy level distributions of doxorubicin–DNA complex, Dox represents the position of doxorubicin.

Two-dimensional density of states plot of DNA before and after doxorubicin intercalated

To understand why the doxorubicin–DNA complex shows such a decrease compared to bare DNA (Fig. 2A), we calculated the two-dimensional density of states (2D DOS) plot of the two earlier discussed cases; the plot shows the distribution of electrons as a function of both space and energy. It’s helpful to analyze the changes in DNA energy levels and the distribution of energy levels upon drug interaction. As can be seen from Fig. 2E and F, new energy levels are introduced after molecular intercalation. However, this makes the energy levels more separated, and HOMO is only distributed on a few bases and intercalator. In contrast, the HOMO of bare DNA is more delocalized, it covers almost the whole molecule so that it is more conducive to the transport of holes from DNA. This may be the reason for the lower conductance of the doxorubicin–DNA molecule.

Effect of covalent binding and groove binding on the electrical properties of DNA

So far, we have investigated the effect of intercalation on the electrical properties of DNA. For the remaining two drug–DNA interaction modes, we constructed and calculated drug–DNA complex and bare DNA structure to analyze the effect of drug binding on the electrical properties of DNA, respectively. Figure 3 shows the structures and energy levels distribution information of nucleic acids before and after covalent binding and groove binding. The average helical rise of the cDPCP–DNA complex is 3.33 Å and the average twist is 33.8°. After covalent binding, the bandgap value increased from 1.17 eV to 2.69 eV. In the groove-binding case, the average helical rise is 3.28 Å and the average twist is 37.2°. The bandgap value decreased from 2.01 eV to 1.65 eV.

Figure 3

DNA structures under two drug–DNA interaction modes. (A) Bare DNA with the sequence (CTCGTCT)₂. (B) cDPCP is bound to DNA (CTCGTCT)₂ through a Pt-N covalent bond. (C) Pentamidine is deleted from the drug–DNA complex compared with d. (D) Pentamidine linked with DNA (TATATATA)₂ by groove binding. (E) Energy levels and bandgap values for four different DNA cases, from left to right, it corresponds to the structures of (A–D).

Fig. 4A and B shows the conductance and transmission variation of DNA before and after cDPCP binding, respectively. The conductance before the covalent binding is 1.02E-6 G₀, and after the covalent binding of cDPCP, the conductivity increases to 5.25E-5 G₀, δG = 5047.06%. As shown in Figure 4B, the increased conductivity trends with the transmission increase from 1.10E-6 to 3.00E-5. The DNA conductance was significantly improved with the covalent binding of cDPCP. The physics behind the increase in the conductance upon drug covalent binding can be understood from the electronic DOS in the region close to the HOMO for the bare dsDNA and drug–DNA complex. The 2D DOS plot for the two cases is shown in Fig. 4C and D. The HOMO of bare dsDNA is −2.97 eV. Before binding, the HOMO is only distributed on a few bases at the end, and the levels of the HOMO and HOMO-1 are very discrete. This is not conducive to charge transport. After binding, the HOMO of the cDPCP–DNA complex decrease from −2.97 eV to −4.45 eV and becomes more delocalized. The energy levels around the HOMO are also more concentrated. So, that causes higher conductance.

Figure 4

Electrical properties of DNA before and after cDPCP covalent binding. (A) The conductance for bare DNA and cDPCP covalent binding case, dash line is the position of HOMO energy level and the lower represents the bare DNA, the higher represents covalent binding case. (B) The transmission in the vicinity of HOMO of two cases. (C) The 2D DOS plot of bare DNA. (D) The 2D DOS plot of cDPCP–DNA complex.

To compare the effects of drugs on the electrical properties of DNA, we investigated the groove-binding effects of the pentamidine–DNA complex (Fig. 3D). We deleted the pentamidine to construct bare dsDNA. Figure 5A represents the conductance before and after drug action. When pentamidine binds to DNA through groove binding, the conductance decrease from 1.79E-6 G₀ to 0.94E-6 G₀, δG = 90.43%. Compared with intercalation and covalent binding, the conductance change of groove bonding is much smaller and this is consistent with other studies (5,9), especially near the HOMO (−3.63 eV) of the drug–DNA complex. The conductance and transmission of the two structures are almost the same.

Figure 5

Conductance (A) and transmission plot (B) as a function of Fermi energy for groove-bonding case. Dash line and stars represent the position of HOMO energy level.

Discussion

In this article, we have investigated the effect of pentamidine-groove-binding, doxorubicin intercalation and cDPCP covalent binding on the charge transport properties of dsDNA. We found that the intercalation of doxorubicin induces two new HOMO and LUMO levels in the bandgap region of DNA, which leads to a bandgap decrease separating the energy levels and decreasing conductance. We also found that decreasing the concentration of intercalators slightly decreases the dsDNA transmission. In contrast, the covalent binding of cDPCP makes the energy levels more concentrated and delocalized, which improves the conductance of DNA significantly. However, in the case of bound pentamidine to DNA through groove binding, the conductance change of groove bonding is much smaller.

Figure 6

Crystal structures of three different drug–DNA complexes and the 2D diagram of three small DNA drug molecules: (A) two doxorubicins intercalated into DNA. (B) cDPCP is bound to DNA through a Pt-N covalent bond. (C) Pentamidine linked with DNA by groove binding.

Overall, we demonstrated the conductance of DNA by drug–DNA interaction and different interaction modes have different effects on the charge transport properties of dsDNA. The changes in DNA conductance can be used to judge whether there is drug–DNA interaction, in what form the drug binds to DNA or even the drug concentration so as to achieve the purpose of monitoring the process of DNA drug therapy. In addition, the binding efficiency of new drugs can be tested based on the change in DNA conductance. This will be useful in understanding a drug’s role in various cell functions and will eventually help develop various DNA drugs to treat numerous diseases. This also provides valuable insight for developing DNA electronic devices in nanoelectronics.

Materials and Methods

Crystal structures of three different drug–DNA complexes

The crystal structures of three different drug–DNA complexes are obtained from three publications (23–25). For the case of intercalation (PBD IDL 236D), the length of dsDNA in the complex is six bases (CGATCG)₂ and two doxorubicin (C₃₂H₃₄N₂O₁₂) molecules were intercalated between the upper and lower CG base pairs, respectively (Fig. 6A). To better understand the effect of intercalators on DNA conductance, GaussView was used to delete doxorubicins from the complexes. We also built base dsDNA in the B-form for the aforementioned sequences by using the Nucleic Acid Builder (NAB) tool (26). For the case of covalent binding (PDB ID: 3CO3), as shown in Figure 6B, cDPCP (C₅H₁₁N₃Pt-) is bonded to DNA (CTCGTCT)₂ by an N atom in G base, to form a Pt-N covalent bond. For a good comparison, we also used GaussView to delete the cDPCP to study the effect of drug molecules on DNA conductance in covalent binding modes and did not make any changes to the molecular structure. The third compound, pentamidine (C₁₉H₂₄N₄O₂) (PBD ID: 3EY0), is linked with (TATATATA)2 by groove binding (Fig. 6C). This binding force is the comprehensive result of hydrogen bonding, hydrophobic and electronic interaction between small molecules and DNA groove bases.

The theoretically electrical measurement model with nonequilibrium Green’s function

Figure 7 is a schematic of a single-molecule junction with an electrode, anchor and bridge components. As we constructed earlier, the bridge is a drug–DNA complex, the anchor is –(CH2)3-SH and the electrode is gold. DNA–drug complex is connected to the gold electrodes at both ends through -(CH₂)₃-SH. In our work, contact coupling between the molecule and the two electrodes is set to 100 meV to simulate the effect of the electrode. During the computational process of DFT, the polarizable continuum model is used to account for the water solvent effect. Furthermore, the total charge of this system is equal to the number of phosphate groups in the drug–DNA complex because we did not add any other counter ions.

Figure 7

The idealized sketch of DNA–drug complex-based electrical measurement system. DNA–drug complex is connected to the gold electrodes at both ends through two anchors.

In this study, we use a combination of DFT and nonequilibrium Green’s function approach including decoherence (27–30) to theoretically study the charge transport properties of three different interaction modes. We used Gaussian09 (31) to perform the DFT calculation to obtain the Hamiltonian and the overlap matrices of the DNA–drug complex, and for the functional and basis here we used B3LYP and 6-31G (d, p), respectively. Next, the two matrices are transformed to the orthogonalized Hamiltonian matrix by using the Löwdin transformation (32). Then, the nonequilibrium Green’s function formulation with decoherence is used to calculate several major electrical information for the bare DNA and drug–DNA complex. The retarded Green functional can be calculated by

$$\begin{equation} {G^r}_M(E)=\frac{1}{E{I}_{MM}-{H}_{MM}-{\sum^r}_L{\sum^r}_R{\sum}_B} \end{equation}$$

(1)

where E is the energy, I is the identity matrix, H is the Hamiltonian of the molecule, |${\sum}_L^r$| and |${\sum}_R^r$| are the self-energy matrices from the left (right) contact. Büttiker probes are used to introduce decoherence into the system; the decoherence rate is set to 10 meV, and |${\sum}_B$| represents the self-energy matrices generated by the Büttiker probes. The transmission probability can be calculated by

$$\begin{equation} {T}_{\mathrm{eff}}={\varGamma}_L{G}^r{\varGamma}_R{G}^a+\sum \limits_{i,j=1}^{N_b}{\varGamma}_L{G}^r{\varGamma}_i{G}^a{W}_{ij}^{-1}{\varGamma}_L{G}^r{\varGamma}_j{G}^a \end{equation}$$

(2)

where Γ_L(R) is the coupling between the device and the left or right contacts, G^r is the retarded Green functional, G^a is the advanced Green functional and N_b represents the number of Büttiker probes. The W_ij⁻¹ is the inverse of W_ij. W_ij = [(1 − R_ii)δ_ij − Γ_iGrΓ_jG_a(1 − δ_ij)], where R_ii is the reflection probability at probe i. Further, the conductance and DOS can be gained by

$$\begin{equation} G\left({E}_f\right)={G}_0\int dE{T}_{eff}(E)\frac{-\partial f\left(E-{E}_f\right)}{\partial E} \end{equation}$$

(3)

$$\begin{equation} D\left({E}_f\right)=-\frac{1}{\pi}\sum \limits_a\operatorname{Im}\left(G\left({E}_f\right)\right) \end{equation}$$

(4)

where E_f is the Fermi functional, T_eff is the transmission probability, G₀ is quantum conductance (G₀ = 2e²/h ≈ 7.75 × 10-5S), a is all the number of relevant atomic orbitals and Im(G(E_f)) represents the imaginary part of G(E_f).

In our previous publication, we defined a parameter to better understand the relative change in conductance upon mutation (33). Here also we define a new parameter, δG, which represents the influence of a drug molecule binding to DNA in each mode. In Eq. (5), G_MIN represents the smaller conductance between bare-DNA structure and drug-DNA complex and G_MAX represents the larger one:

$$\begin{equation} \delta G=\frac{G_{\mathrm{MAX}}-{G}_{\mathrm{MIN}}}{G_{\mathrm{MIN}}}\times 100 \end{equation}$$

(5)

Acknowledgements

We thank Kai Wang (Third Military Medical University) for useful comments.

Conflict of Interest statement. The authors declare that there are no conflicts of interest.

Funding

China Scholarship Council Study Program for Young Backbone Teachers (201707845017); Scientific and Technological Research Foundation of Chongqing Municipal Education Commission (KJQN201900643, KJQN201900630); Chongqing Natural Science Foundation of Chongqing Municipal Science and Technology Bureau (cstc2020jcyj-msxmX0550); Raised Fund Plan of National Natural Science Foundation of China of Chongqing University of Posts and Telecommunications (A2020-528); Research and training program for college students of Chongqing University of Posts and Telecommunications (A2021-54).

Conflict of Interest statement. The authors declare that there are no conflicts of interest.

Authors’ contributions

L.H. supervised conceptualization, methodology and software, writing, reviewing and editing original draft; Z.X. performed conceptualization, software, data curation and writing original draft; X.L. analyzed methodology; Z.Z. performed conceptualization; F.Q. supervised methodology and software and N.Z. supervised conceptualization and methodology.