Open Access
Add to BibSonomy
Abstract
Among the various types of interactions between biomolecules, electrostatic interactions dominate as these are long-range interactions and are often a generic first step in the recruitment of specific ligands. DNA, being a highly charged molecule, attracts a plethora of molecules. Interactions between DNA and proteins or small molecules shape the overall function of the cell. Various processes such as DNA replication, DNA repair, synthesis of mRNA, and packaging of DNA are mediated by interactions between protein molecules and DNA that are predominantly electrostatic. Here, we present a fluorescence resonance energy transfer (FRET)-based probe which can report on the electrostatic interactions between the negatively-charged DNA and positively-charged metal ions, oligopeptides, as well as DNA groove-binding drug molecules. The simplicity, sensitivity, and versatility of the DNA-based probe makes it suited for applications where specific protein-DNA interactions can be probed, and DNA-binding drugs can be discovered in high-throughput screens of compound libraries. This is particularly relevant given that some of the most potent antitumor and antimicrobial drugs associate with DNA electrostatically.
© 2021 Optical Society of America under the terms of the OSA Open Access Publishing Agreement
1. Introduction
Electrostatic interactions are of fundamental importance in biology and govern many inter- and intra-biomolecular interactions since almost all biomolecules are charged. The charge on biomolecules determines their structure, dynamics, and function [1,2]. Intra-molecular electrostatic interactions determine the structure and stability of proteins, RNA, and DNA. DNA, which carries our genetic information and the blue print for the synthesis of all proteins, is a highly charged molecule [3,4]. Each base-pair of DNA carries one-negative charge on its backbone at physiological conditions [5]. The electrostatic interactions arising due to positive and negative charges on proteins and DNA respectively, play a critical role in various DNA-protein interactions which control cellular activity. For example, the interaction between positively charged histone molecules with the negatively charged DNA is the first step in compacting the two-meter-long human DNA into the micron-sized nucleus. The first encounter between DNA and protein molecules is mediated by electrostatic interactions.
Due to the DNA’s role in transferring genetic information and its importance in cell proliferation, as well as its unique properties such as having a well-defined structure and being a highly charged molecule, DNA has been one of the obvious targets for drug development [6,7]. The vast majority of drug molecules, such as many intercalators and groove binders, bind to DNA non-covalently [8–11]. Electrostatic interaction between the drug molecule and DNA determines the strength of these contacts [11]. Therefore, many compounds that strongly bind DNA such as Netropsin and Distamycin A, carry net positive charge at physiological conditions [9,12]. A small molecule which binds to the DNA has the potential to alter cell function by interfering with the DNA-protein interactions, resulting in many such compounds being cytotoxic and used in cancer therapy or as antimicrobials and antiparasitic drugs.
The quest to contrive novel drugs has led to the development of various techniques and methods to discover and study DNA-binding compounds [13,14]. Here we present a FRET-based probe to detect binding between DNA and small molecules which are dominated by electrostatic interactions. The probe presented here can be potentially used in high-throughput screens of compounds for the discovery of drug molecules which target DNA.
2. Materials and methods
The sensor was made of a single strand of 59 DNA bases designed to fold into a tongs structure (Fig. 1(a)) having two double-stranded arms connected by a stretch of single-stranded DNA [15]. The sequence was synthesized and internally modified with the fluorescence resonance energy transfer (FRET) pair Cy3 and Cy5 that straddle the DNA sugar phosphate backbone at the 4th and 56th bases, respectively. The sensor was purified by HPLC and supplied by Integrated DNA Technologies, Skokie, IL. All other reagents such as oligo-Lysine and various compounds were acquired from Sigma-Aldrich/Merk, Germany and were used directly without any further purification.
All the experiments were conducted in a buffer containing 50 mM NaCl, 10 mM TRIS-HCL, pH 7.5, 0.2 mg/ml BSA, excluding the experiments where the response of the sensor is recorded at various concentrations of NaCl and MgCl2. A 100 µl solution was prepared of various concentrations of the DNA-binding compound (e.g drug, or lysine) to which a fixed amount of sensor (0.5 µl) was added to get the desired relative concentration of compound to DNA base-pairs. After adding the sensor to the experimental buffer, the solution was thoroughly mixed and incubated for 30 min. The resulting concentration of the sensor was 0.5 µM (or 12 µM of dsDNA base-pairs). The solutions were subsequently transferred into 96-well plates (Corning 3651) to record the emission spectra of the sensor. The emission spectra in the absence and presence of the DNA-binding molecules were measured at room temperature on a plate reader (BioTek, Synergy H1) by exciting the donor only (Cy3) at a wavelength of 500 nm. The emission spectra were recorded from 530 nm to 700 nm to straddle both donor and acceptor fluorescence. To calculate the FRET ratio, the average intensity in an 8 nm window around the Cy5 peak was divided by the average intensity around the peak of Cy3 fluorescence.
3. Results and discussion
3.1 The sensor responds to monovalent and divalent salts
To detect electrostatic interaction with DNA we took advantage of the repulsion between the negatively-charged dsDNA helices, which is reduced upon binding to positively-charged molecules. We designed a sensor using a single DNA oligonucleotide strand which folds into a tongs shape as shown in Fig. 1(a). The two rigid double stranded arms (12 base-pairs long) are linked by a softer single-stranded hinge. To detect the proximity of the two arms of the sensor, each arm was labelled with a fluorophore forming a FRET pair (See Materials and Methods and Fig. 1(a)). When suspended in buffer, the sensor takes a conformation where the hinge can bend, partially closing the sensor and bringing the fluorophores closer to each other. Any change in the net negative charge on the two arms as a result of binding of positively-charged ligands will decrease the effective repulsion between the two arms resulting in a decrease in the average distance between the two arms of the sensor. The extent of fluorescence resonance energy transfer (FRET) will then report on how close these two arms are, and therefore on the binding of ligands. The short arms of the sensor are an order of magnitude shorter than the persistence length of dsDNA at the conditions studied and remain rigid. The single-stranded hinge has a length comparable to the persistence length of ssDNA. Therefore, its rigidity will only minimally influence the response of the sensor. This is confirmed in our previous work [15]where we demonstrated that changing the hinge length of a similar sensor by a few bases minimally changed the FRET signal under identical conditions.
Fig. 1. The DNA-based sensor shows increased FRET in solutions of increasing concentration of mono- and divalent salts. (A) The single-stranded oligonucleotide is designed to hybridize, forming two stiff double-stranded arms separated by a single-stranded hinge. The hinge renders the sensor flexible in the middle allowing it to bend. As the negative charge on the dsDNA backbone is increasingly neutralized (screened) as function of increasing salt concentration in the buffer, the effective repulsion between the two double-stranded arms of the sensor decreases, bringing the two arms of sensor closer to each other which results in higher energy transfer between the Cy3 and Cy5 FRET pair labelling the sensor arms. (B) Fluorescence emission spectra of the sensor in the presence of increasing concentrations of NaCl. Emission spectra were obtained by exciting the donor at 500 nm and recording the emission spectra from 530 nm to 700 nm. This range includes the emission peaks of the donor Cy3 (570 nm) and the acceptor Cy5 (670 nm). The donor intensity decreases and the acceptor intensity increases as the concentration of NaCl increases. The FRET ratio (acceptor intensity divided by the donor intensity) increases as function of increasing concentrations of the monovalent salt NaCl (C), and the divalent salt MgCl2 (D). The concentration of the sensor was 0.5 µM in all cases.
The repulsion between the negatively charged phosphate backbone of DNA is the major force destabilizing the double helix and also affecting its secondary and tertiary structure [16]. Mono- and divalent cations such as Na+ and Mg2+ play an important role in stabilizing dsDNA and facilitating its folding into secondary and tertiary structures [1,4]. Cations act to screen the negative charges of the sugar-phosphate backbone from interacting with one another. To characterize the sensitivity of the sensor to the degree of screening of negative charge on the dsDNA, we studied it in media with increasing concentrations of the monovalent salt NaCl. We excited Cy3 at 500 nm and recorded the emission spectra encompassing Cy3 and Cy5 fluorescence (Fig. 1(B)). At 50 mM NaCl we measured a low Cy5/Cy3 fluorescence ratio of 1.5. With increasing NaCl concentration, Cy5 fluorescence increased at the expense of Cy3 fluorescence, resulting in an increased FRET ratio (Fig. 1(B, C)). As the concentration of NaCl increases in the buffer solution, the negative charge on the sensor is increasingly screened resulting into less repulsion between the two arms of the sensor. We observed a similar response of the sensor in the presence of the divalent salt MgCl2 (Fig. 1(D)). However, a much smaller concentration of the MgCl2 was needed to bring the sensor into a closed conformation. These findings are in line with the known fact that divalent salts such as MgCl2 are more effective in screening the charges on dsDNA when compared with monovalent salts [16,17] . These results clearly demonstrate that the sensor can efficiently detect electrostatic association of positively-charged ions (cations) with the negatively-charged dsDNA.
3.2 Detecting electrostatic binding of oligopeptides to DNA
Fig. 2. Response of the sensor in the presence of oligolysines of various lengths. (A-C) Emission spectra of the sensor in the absence and presence of various concentrations of trilysine (Lys-3), tetralysine (Lys-4), and pentalysine (Lys-5). As the concentration of the positively-charged oligolysines increases, more bind to the DNA sensor rendering it less negative and reducing the repulsion between its arms. This leads to an increased FRET Ratio (D-F). (G) The FRET ratio is plot as a function of the concentration of lysine residues as opposed to the concentration of oligolysine molecules shown in (A-F). This highlights the larger efficacy of pentalysine in screening the DNA charge. The response of the sensor to the increasing concentrations of oligolysines was recorded in the presence of 50 mM NaCl, 10 mM Tris-HCl pH 7.5, and 0.2 mg/ml BSA. The concentration of the sensor was 0.5 µM in all cases.
One of the important factors which governs DNA-protein binding is electrostatic interaction between the negatively-charged dsDNA and the positively-charged residues within a protein molecule. For example, the interaction between histones and dsDNA which takes place in the first step of DNA compaction within the nucleus involves interaction between dsDNA and Lysine residues on the histone molecules. Given that many proteins which interact with DNA contain Lysine residues, oligo-lysines have been widely used as a model system to understand the interaction between dsDNA and proteins [18]. Furthermore, poly-Lysine has been extensively used to compact DNA for gene delivery systems [19,20]. We therefore studied the response of our sensor to trilysine (Lys-3), tetralysine (Lys-4) and pentalysine (Lys-5) which carry +3, +4 and +5 positive charges, respectively. Figures 2(A-C) show the emission spectra of the sensor as function of increasing concentrations of these three oligo-lysines. As the concentration of the oligo-lysine increases the intensity of Cy3 decreases and the intensity of Cy5 increases indicating that the sensor arms come closer together, thus increasing the FRET ratio (Fig. 2(D-F)). This indicates that all three oligo-lysines bind to the sensor electrostatically and screen the negative charge on the dsDNA.
However, the efficiency of the oligo-lysine to bind to the dsDNA and screen the negative charge increases as function of its length; i.e the number of lysines forming the molecule. Pentalysine can more effectively screen the dsDNA charge when compared to trilysine as judged by the increase in the FRET ratio at a concentration of 1.2mM: for trilysine the change is ∼ 0.4 whereas it is 2.1 for pentalysine. Furthermore, the response of the sensor saturates at 0.6 mM pentalysine, while as much as 6 mM trilysine does not result in saturation.
A better way to demonstrate the increased screening efficacy with increased oligo-lysine valency is to compare the FRET ratio as function of the concentration of lysine residues (rather than oligo-lysine molecules). For example, the concentration of trilysine molecules used in this experiment ranges from 0 to 6 mM, which corresponds to a concentration of 0 to 18 mM lysine residues. Figure 2(G) shows the stark difference between the response of the sensor to pentalysine and trilysine, where the former rises sharply and saturates at ∼3 mM lysine residues per base-pair. Our findings are in line with the fact that lysine binding to the dsDNA is cooperative [21] and as the number of residues per oligolysine increases its efficacy to compact the dsDNA increases. Irina Nayvelt and co-authors have reported that pentalysine completely condensed lambda DNA while trilysine did not result in lambda DNA condensation even at elevated concentrations [18].
3.3 Detecting dsDNA-drug association
DNA is the presumed intracellular target of some of the most important clinical compounds used as antitumor, antifungal or antibacterial drugs [6,7,22]. Due to DNA’s well-defined structure it has been considered as one of the most important targets in target-based drug discovery programs [6]. Drug molecules bind DNA mainly in two ways: either by intercalation or by groove-binding. In the case of DNA-intercalation, a planar aromatic molecule inserts itself between two successive base-pairs of the DNA resulting in DNA-deformation and lengthening [7,8,23]. However, in the case of DNA-groove binding, drug molecules bind to the DNA externally by aligning either with the minor or major grooves of the DNA, resulting in less deformation of the DNA structure when compared to the DNA-intercalation. Nevertheless, for both intercalators and groove-binders, the strong binding to the DNA almost always includes an electrostatic component as the ligand will be protonated, which enables strong binding to the negatively-charged dsDNA.
We therefore tested the ability of the sensor to detect the binding of groove-binding drugs to DNA. To start with, we studied the antiviral antibiotic, Netrospin, which is known to associate with dsDNA via groove binding [24]. The emission spectra of the sensor at increasing Netropsin concentration are characteristic of the reduced repulsion between the sensor arms where Cy3 intensity decreased and Cy5 intensity increased as the concentration of Netropsin increased (Fig. 3(A)), resulting in an increased FRET Ratio (Fig. 3(B)). We further tested two other DNA groove-binding drugs, furamidine and pentamidine [22], both of which increased the FRET ratio in a concentration-dependent manner (Fig. 3(C, D)). These measurements were performed at physiological buffer conditions but at room temperature. However, we also confirmed that the sensor response is indistinguishable from that in Fig. 1(C) in the presence of NaCl and shows a slightly lower Netrospin affinity at 37°C (data not shown), consistent with previous reports [25]. The sensor can therefore be used to study the temperature-dependent affinity of the DNA-binding drugs.
Fig. 3. Detecting the electrostatic interaction between dsDNA and DNA groove-binding drug molecules. (A) Emission spectra of the sensor in the presence of increasing concentrations of the groove-binding drug Netropsin. As the number of Netropsin molecules per base-pair increases, they effectively screen the negative charge on the dsDNA bringing the arms closer together which reduces Cy3 intensity and increases Cy5 intensity resulting in an increased FRET Ratio (B). The FRET ratio also increases in with increased concentrations of the other groove-binding drugs Furamidine (C), and Pentamidine (D). The concentration of the sensor was 0.5 µM in all cases.
These results combined clearly demonstrate that the sensor presented here can report on a variety of molecular species which interact with the dsDNA electrostatically.
To rule out other scenarios by which the fluorescence signal is altered in the presence of the molecules and to further demonstrate that it is indeed the reduced repulsion between the dsDNA arms of the sensor upon ligand binding that results in the increased FRET ratio, we studied two other groove binders whose association with DNA can be modulated. Mithramycin and Chromomycin are antibiotic and antitumor agents which belong to the aureolic acid group [26–28]. They bind to dsDNA thereby blocking its function as a template for DNA and RNA polymerases. Mithramycin and Chromomycin are negatively charged molecules above pH 7 (pKa 7.0) and therefore their binding to DNA should be hindered under our buffer conditions (pH 7.5). Indeed, as the concentration of Mithramycin increases we measure no change in either Cy3 or Cy5 intensities (Fig. 4(A)). The FRET ratio in the presence of either Mithramycin or Chromomycin remains constant (Fig. 4(B)), demonstrating their inability to bind DNA under these conditions.
Fig. 4. Interaction of Mithramycin and Chromomycin with the sensor is mediated by divalent ions. (A) Emission spectra of the sensor in the presence of increasing concentrations of the groove-binding drug Mithramycin with 50 mM NaCl, 10 mM Tris-HCl pH 7.5 and 0.2 mg/ml BSA. Neither donor (Cy3) nor acceptor (Cy5) intensities change indicating that the drug does not interact with DNA under these buffer conditions. (B) The FRET ratio remains constant with increasing concentrations of both Mithramycin and Chromomycin which carry a negative charge under these buffer conditions. (C-D) Adding the divalent salt MgCl2 results in an increase in the FRET ratio as function of increasing groove-binder concentration. The initial dip is due to intrinsic fluorescence from the DNA-bound groove binders which is overwhelmed by the FRET signal at higher drug concentrations. The concentration of the sensor was 0.5 µM in all cases.
It is, however, reported that Mithramycin and Chromomycin bind to DNA in the presence of the divalent cation Mg2+ [26]. Mg2+ forms a complex with these drug molecules which then can bind to the negatively charged DNA. We therefore measured the response of the sensor at various concentrations of Mithramycin and Chromomycin in the presence of 20 mM MgCl2. At small drug concentrations, the FRET ratio drops and then increases as the drug concentration is increased (Fig. 4(C, D)). This initial drop in FRET ratio was not observed for any other ion or ligand used in this work and results from the intrinsic fluorescence of these drugs upon binding to DNA [29]. Their fluorescence coincides with the emission of Cy3 (∼ 530-580 nm), resulting in an apparent reduction of the measured FRET ratio. To measure the correct FRET ratio, one needs to subtract the background resulting from the drugs at each concentration. However, the recovery in the FRET ratio indicates that with increased concentration of the drugs, the arms of the DNA come closer together such that the FRET signal overwhelms drug fluorescence. The increase in the FRET ratio is larger for Chromomycin than for Mithramycin at the same concentration suggesting that it has a higher binding constant under similar experimental conditions, consistent with previous reports [30].
4. Conclusion and prospect
We have shown that the sensor we designed reports on a variety of species which interact electrostatically with DNA including ions, oligopeptides, and groove-binding drugs. While the response of the sensor is generic and is not specific to the identity of the ligand, the methodology we developed makes possible a range of applications that are specific. For instance, since many specific protein-DNA interactions are mediated by electrostatics, one could tune the sequence of the sensor for use in protein binding assays. Moreover, the simplicity of the sensor design and the sensitivity of its fluorescence readout, makes it well-suited for high-throughput screens of DNA-drug interactions. Given that DNA groove-binders are among the most effective antineoplastic, antimicrobial, and antiparasitic drugs, screening compound libraries for such interactions could be a rapid first step in drug discovery programs.
The prospect of using the sensor within living cells is also intriguing given that it could additionally report fluorescently on the entry, binding and release dynamics of groove-binding drugs in the cell, as we previously demonstrated for intercalating drugs [7]. However, given its geometry, the present sensor will additionally respond to macromolecular crowding which is inevitable in cells [15]. Separating the signal due to drug binding from that due to crowding can nevertheless be achieved by changing the position of the FRET pair along the sensor arms. This allows tuning the dynamic range of sensitivity of the sensor as we previously demonstrated, making it possible to detect changes in the FRET signal as the drug enters the cells and binds the sensor.
Funding
National Science Foundation (PHY-150520, PHY-1915119).
Acknowledgements
This work was supported in part by the National Science Foundation grants PHY-1505020 and PHY-1915119 to GTS. The research was partially carried out using the Core Technology Platform at New York University Abu Dhabi.
Disclosures
The authors declare no conflicts of interest
Data Availability
All data used in this paper are included in the figures and may be obtained from the corresponding author upon reasonable request.
References
1. A. G. Cherstvy, “Electrostatic interactions in biological DNA-related systems,” Phys. Chem. Chem. Phys. 13(21), 9942–9968 (2011). [CrossRef]
2. H. X. Zhou and X. Pang, “Electrostatic interactions in protein structure, folding, binding, and condensation,” Chem. Rev. 118(4), 1691–1741 (2018). [CrossRef]
3. W. M. Gelbart, R. F. Bruinsma, P. A. Pincus, and V. A. Parsegian, “DNA-inspired electrostatics,” Phys. Today 53(9), 38–44 (2000). [CrossRef]
4. J. Lipfert, S. Doniach, R. Das, and D. Herschlag, “Understanding nucleic acid-ion interactions,” Annu. Rev. Biochem. 83(1), 813–841 (2014). [CrossRef]
5. U. F. Keyser, B. N. Koeleman, S. Van Dorp, D. Krapf, R. M. M. Smeets, S. G. Lemay, N. H. Dekker, and C. Dekker, “Direct force measurements on DNA in a solid-state nanopore,” Nat. Phys. 2(7), 473–477 (2006). [CrossRef]
6. L. H. Hurley, “DNA and its associated processes as targets for cancer therapy,” Nat. Rev. Cancer 2(3), 188–200 (2002). [CrossRef]
7. C. U. Murade, S. Chaudhuri, I. Nabti, H. Fahs, F. S. M. Refai, X. Xie, Y. E. Pearson, K. C. Gunsalus, and G. T. Shubeita, “FRET-based probe for high-throughput DNA intercalator drug discovery and in vivo imaging,” ACS Sens. 6(6), 2233–2240 (2021). [CrossRef]
8. A. A. Almaqwashi, T. Paramanathan, I. Rouzina, and M. C. Williams, “Mechanisms of small molecule-DNA interactions probed by single-molecule force spectroscopy,” Nucleic Acids Res. 44(9), 3971–3988 (2016). [CrossRef]
9. S. M. Nelson, L. R. Ferguson, and W. A. Denny, “Non-covalent ligand/DNA interactions: minor groove binding agents,” Mutat. Res. 623(1-2), 24–40 (2007). [CrossRef]
10. S. U. Rehman, T. Sarwar, M. A. Husain, H. M. Ishqi, and M. Tabish, “Studying non-covalent drug-DNA interactions,” Arch. Biochem. Biophys. 576, 49–60 (2015). [CrossRef]
11. L. Strekowski and B. Wilson, “Noncovalent interactions with DNA: an overview,” Mutat. Res. 623(1-2), 3–13 (2007). [CrossRef]
12. A. Lansiaux, L. Dassonneville, M. Facompre, A. Kumar, C. E. Stephens, M. Bajic, F. Tanious, W. D. Wilson, D. W. Boykin, and C. Bailly, “Distribution of furamidine analogues in tumor cells: influence of the number of positive charges,” J. Med. Chem. 45(10), 1994–2002 (2002). [CrossRef]
13. M. M. Aleksic and V. Kapetanovic, “An overview of the optical and electrochemical methods for detection of DNA - drug interactions,” Acta. Chim. Slov. 61, 555–573 (2014).
14. A. Kellett, Z. Molphy, C. Slator, V. McKee, and N. P. Farrell, “Molecular methods for assessment of non-covalent metallodrug-DNA interactions,” Chem. Soc. Rev. 48(4), 971–988 (2019). [CrossRef]
15. C. U. Murade and G. T. Shubeita, “A molecular sensor reveals differences in macromolecular crowding between the cytoplasm and nucleoplasm,” ACS Sens. 4(7), 1835–1843 (2019). [CrossRef]
16. Z. J. Tan and S. J. Chen, “Nucleic acid helix stability: effects of salt concentration, cation valence and size, and chain length,” Biophys. J. 90(4), 1175–1190 (2006). [CrossRef]
17. A. X. Kohll, P. L. Antkowiak, W. D. Chen, B. H. Nguyen, W. J. Stark, L. Ceze, K. Strauss, and R. N. Grass, “Stabilizing synthetic DNA for long-term data storage with earth alkaline salts,” Chem. Commun. 56(25), 3613–3616 (2020). [CrossRef]
18. I. Nayvelt, T. Thomas, and T. J. Thomas, “Mechanistic differences in DNA nanoparticle formation in the presence of oligolysines and poly-L-lysine,” Biomacromolecules 8(2), 477–484 (2007). [CrossRef]
19. H. G. Hansma, R. Golan, W. Hsieh, C. P. Lollo, P. Mullen-Ley, and D. Kwoh, “DNA condensation for gene therapy as monitored by atomic force microscopy,” Nucleic. Acids Res. 26(10), 2481–2487 (1998). [CrossRef]
20. S. W. Kim, “Polylysine copolymers for gene delivery,” Cold Spring Harb Protoc. 2012(4), 433–438 (2012). [CrossRef]
21. G. Liu, M. Molas, G. A. Grossmann, M. Pasumarthy, J. C. Perales, M. J. Cooper, and R. W. Hanson, “Biological properties of poly-L-lysine-DNA complexes generated by cooperative binding of the polycation,” J. Biol. Chem. 276(37), 34379–34387 (2001). [CrossRef]
22. P. G. Baraldi, A. Bovero, F. Fruttarolo, D. Preti, M. A. Tabrizi, M. G. Pavani, and R. Romagnoli, “DNA minor groove binders as potential antitumor and antimicrobial agents,” Med. Res. Rev. 24(4), 475–528 (2004). [CrossRef]
23. L. S. Lerman, “Structural considerations in the interaction of DNA and acridines,” J. Mol. Biol. 3(1), 18–IN14 (1961). [CrossRef]
24. M. L. Kopka, C. Yoon, D. Goodsell, P. Pjura, and R. E. Dickerson, “The molecular origin of DNA-drug specificity in netropsin and distamycin,” Proc. Natl. Acad. Sci. U. S. A. 82(5), 1376–1380 (1985). [CrossRef]
25. S. Wang, A. Kumar, K. Aston, B. Nguyen, J. K. Bashkin, D. W. Boykin, and W. D. Wilson, “Different thermodynamic signatures for DNA minor groove binding with changes in salt concentration and temperature,” Chem. Commun. 49(76), 8543–8545 (2013). [CrossRef]
26. P. Aich, R. Sen, and D. Dasgupta, “Role of magnesium ion in the interaction between chromomycin A3 and DNA: binding of chromomycin A3-Mg2+ complexes with DNA,” Biochemistry 31(11), 2988–2997 (1992). [CrossRef]
27. E. Berman, S. C. Brown, T. L. James, and R. H. Shafer, “NMR studies of chromomycin A3 interaction with DNA,” Biochemistry 24(24), 6887–6893 (1985). [CrossRef]
28. G. Gause, “Olivomycin, Chromomycin and Mithramycin: Antibiotics!,” J. W. Corocan & F. E. Hahn, Eds (Springer Verlag, 1975), pp 197-202.
29. D. C. Ward, E. Reich, and I. H. Goldberg, “Base specificity in the interaction of polynucleotides with antibiotic drugs,” Science 149(3689), 1259–1263 (1965). [CrossRef]
30. F. Barcelo, M. Ortiz-Lombardia, M. Martorell, M. Oliver, C. Mendez, J. A. Salas, and J. Portugal, “DNA binding characteristics of mithramycin and chromomycin analogues obtained by combinatorial biosynthesis,” Biochemistry 49(49), 10543–10552 (2010). [CrossRef]
