Reverse Orientation Preference of Cyclic Pyrrole

5 days ago - These results indicate that the orientation preference of cPIPs can be controlled by the position of the amino groups on the γ-turn unit...
0 downloads 0 Views 2MB Size
Download PDF
Article pubs.acs.org/JACS

Cite This: J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Control of Forward/Reverse Orientation Preference of Cyclic Pyrrole−Imidazole Polyamides Yuki Hirose,† Sefan Asamitsu,†,# Toshikazu Bando,*,† and Hiroshi Sugiyama*,†,‡ †

Department of Chemistry, Graduate School of Science, Kyoto University, Sakyo, Kyoto 606-8502, Japan Institute for Integrated Cell-Material Science (WPI-iCeMS), Kyoto University, Sakyo, Kyoto 606-8501, Japan

‡

Downloaded via NOTTINGHAM TRENT UNIV on August 10, 2019 at 02:55:13 (UTC). See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles.

S Supporting Information *

ABSTRACT: Pyrrole−imidazole polyamides (PIPs) bind to predetermined double-stranded DNA sequences and selectively target a large variety of DNA sequences. Although the forward-binding (5′-3′/N−C) orientation, in which the N-terminus of PIPs faces the 5′-terminus of DNAs, is considered to be the main binding manner of PIPs, a reverse-binding (5′-3′/C−N) orientation, in which the C-terminus of PIPs faces the 3′-terminus of DNAs, sometimes causes unintended binding. Here, we synthesized optical or structural isomers of previously reported cyclic PIPs (cPIPs), which differ in the position of the amino groups in the γ-turn units, and we investigated their binding affinities both in the forward- and reverse-binding orientation. We show that cPIPs with (R)-α-amino-γ-turn units prefer the forward orientation as do hairpin PIPs. More importantly, we document for the first time the remarkable reversebinding preference of cPIPs with (S)-α-amino-γ-turns. These results indicate that the orientation preference of cPIPs can be controlled by the position of the amino groups on the γ-turn units, which may markedly increase the number of DNA sequences that can be targeted by PIPs.

■

5′ → 3′ direction of DNA strands,17 some recent studies demonstrated that PIPs sometimes prefer the reverse-binding orientation in which the C → N direction in PIPs is the same as the 5′ → 3′ direction of the DNAs.18−20 The existence of two binding orientations causes unintended binding of PIPs (Figure 1). Several attempts have been made to control orientation selectivity of PIPs and avoid unintended binding. Introducing conformationally flexible β-alanine moieties and (R)-β-aminoγ-turns was reported to render hPIPs that prefer the reversebinding orientation.19 Additionally, previous studies reported that (R)-α-amino-γ-turns enhanced the forward-orientation preference of hPIPs.19,21 Similarly, (S)-α-amino-γ-turns were used to enhance the reverse-orientation preference of hPIPs, but the forward orientation remained dominant.21 Despite such successful attempts in hPIPs, the orientation preference of cPIPs has not been studied extensively. Considering the superiority of cPIPs as mentioned earlier, comprehensive studies on the orientation selectivity of cPIPs are highly desirable to precisely design cPIPs target sequences with reduced unintended binding. In this study, we first improved the macrolactamization step during synthesis. Using this scheme, we readily synthesized several types of cPIPs that differ in the position (α or β) and chirality (R or S) of the amino groups in the γ-turn. Then we investigated cPIPs’ binding affinities both in the forward- and

INTRODUCTION The pyrrole−imidazole polyamides (PIPs) are a class of DNAbinding molecules, which were first proposed by Dickerson and co-workers1,2 and developed by Lown et al.3,4 and Dervan and co-workers. 5,6 PIPs are mainly composed of Nmethylpyrrole (Py) and N-methylimidazole (Im) connected by amide bonds. These molecules bind to the minor groove of double-stranded B-DNA according to unique base-recognition rules. Antiparallel pairing of Im opposite of Py (Im/Py) recognizes a G•C Watson−Crick base pair, whereas a Py/Py pair recognizes an A•T or T•A base pair. Taking advantage of their sequence selectivity and high binding affinity, many types of PIPs have been developed and are used as anticancer drugs,7 DNA fluorescent probes,8 and gene regulators9,10 both in vitro and in vivo. Previously, mainly hairpin PIPs (hPIPs) were studied, in which two arrangements of Py and Im are connected by a γaminobutyric acid turn (γ-turn) (Figure 1). However, cyclic PIPs (cPIPs) have also been described. Dervan and co-workers obtained cPIPs by introducing a second γ-turn at the C- and N-termini of hPIPs.11 Our group reported cPIPs with cysteine turn units.12 Particularly, cPIPs with two γ-turn units showed higher DNA-binding affinity than corresponding hPIPs, which enabled X-ray crystal structure analysis of the PIP−DNA binding complex.13,14 However, cPIPs have been less studied, mainly because they are difficult to synthesize because of inefficient intramolecular macrocyclization.15,16 Although PIPs prefer the forward-binding orientation, in which the N → C terminus direction in PIPs is the same as the © XXXX American Chemical Society

Received: May 22, 2019

A

DOI: 10.1021/jacs.9b05516 J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Article

Journal of the American Chemical Society

performed according to Figure 3. After completion of the machine-assisted, solid-phase synthesis using a Py-trityl resin, hPIPs with a C-terminus were cleaved from the resin using CF3CH(OH)CF3. The resulting crude powder (6−10) was dissolved in dimethylformamide (DMF) and cyclized by condensation using a combination of pentafluorophenyl diphenylphosphinate (FDPP) and N,N-diisopropylethylamine (DIEA). The reaction was completed within 18 h, which is significantly shorter than the previously reported cyclization time of 3 d.15 Furthermore, the equivalence of FDPP (3 equiv) and DIEA (6 equiv) is more reasonable than that of DPPA (50 equiv) and DIEA (200 equiv) in former cyclization.16 Importantly, we noted that this intramolecular cyclization did not result in high yields when acetonitrile was used instead of DMF, or DPPA was used instead of FDPP. Finally, the protective Boc or Cbz groups of 6−9 were deprotected using proper acids to yield the target compounds (1−5), which were purified by high-performance liquid chromatography (HPLC) (yield: 15−32%). We used these HPLC-purified samples in the following assays. Binding Affinities of Each cPIP. The binding affinities of each cPIP (1−5) were investigated using double-stranded DNA melting temperature (Tm) measurements and surface plasmon resonance (SPR) assays. For the Tm assay, we prepared two types of DNA oligomers: 5′-CCAGTACTGG3′/3′-GGTCATGACC-5′ for the forward-binding orientation and 5′-GGTCATGACC-3′/3′-CCAGTACTGG-5′ for the reverse-binding orientation. We also prepared two 5′biotinylated hDNA oligomers for the SPR assay: 5′-biotinCGCCAGTACTGGCTTTTGCCAGTACTGGCG-3′ for the forward binding and 5′-biotin-GCGGTCATGACCGTTTTCGGTCATGACCGC-3′ for the reverse binding. The bold bases above are the binding sites of cPIPs. Using Tm assays, we can measure the relative binding affinities through Tm and ΔTm values (ΔTm = Tm (DNA + cPIP) − Tm (DNA)). These values are shown in Table 1 (PIP: 2 equiv) and Table S1 (Supporting Information) (PIP: 1.5 equiv), and representative denaturing profiles are shown in Figure S1 (PIP: 2 equiv). All SPR sensorgrams and the rates of association (ka) and dissociation (kd) that are calculated by fitting with 1:1 binding models are provided in Figures S2−S6, and the dissociation constants (KD) are listed in Table 2. The KD values indicate a concrete binding affinity, and we therefore determined the orientation selectivity of cPIPs comparing the KD values of each binding orientation (forward/reverse specificity, Table 2). Overall, little difference between Tm values with 1.5 and 2 equiv of PIPs provided some evidence for 1:1 binding of cPIP to dsDNA. However, no clear transition profiles in the melting curves for 2 and 12’s forward binding (Figure S1) suggested that multiple-state phase changes caused by a more than 1:1 binding might be involved. To check the binding stoichiometry for 2, we additionally performed the fitting of the sensorgrams for 2 with a 2:1 binding model.22 The two KD values are calculated to be almost equal (1.0 and 0.94 nM, Table S2), suggestive of a 1:1 binding mode of 2 to dsDNAs within a measured concentration range (31.25−125 nM). Given that we used the higher PIP concentrations for Tm assays, the linear-like melting profiles may be ascribable to unspecific binding of 2 to the dsDNA sequence. The higher ΔTm and lower KD value of 1 with a forward sequence indicated that 1 strongly bound to the forward sequence (ΔTm = 44.9 ± 0.5 °C, KD = 9.8 × 10−11 M).

Figure 1. Chemical structure and ball-and-stick notation of a hPIP that was previously reported to favor the reverse-binding orientation (right) over the forward-binding orientation (left).

in the reverse-binding orientations based on the assumption that the dominant binding orientation can be controlled by changing the position of the amino groups in the γ-turn (Figure 2). Our findings clearly indicate that the position and chirality of the substituent in the γ-turn units have a drastic effect on cPIPs’ preference of binding orientation.

Figure 2. Chemical structures of five cPIPs that were synthesized in this study. 1−4: (R)-β-, (S)-β-, (R)-α-, and (S)-α-substituted γ-turns, respectively. 5: No amino group.

■

RESULTS AND DISCUSSION Design and Synthesis of cPIPs. First, we synthesized five types of cPIPs (Figure 2). On the basis of the structure of cPIP 1, which Dervan et al. used in previous studies, we prepared different types of cPIPs with amino groups that differ in their position and chirality in the γ-turn units and compared them with cPIPs without amino groups (2−5). Upon binding, amino groups in the γ-turn of PIPs are known to greatly affect the interaction between PIPs and DNAs. We therefore expected that the orientation preference could be inverted by inverting the chirality of amino groups. The synthesis of cPIPs was B

DOI: 10.1021/jacs.9b05516 J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Article

Journal of the American Chemical Society

Figure 3. Synthesis of compounds 1−5.

Table 1. Summary of Tm Measurements

a remarkable reverse specificity (0.02). These results for 3 and 4 clearly demonstrate that the forward/reverse specificity of cPIPs can be controlled by changing the chirality of α-amino groups in the γ-turn: the (R)-α-amino-γ-turn prefers the forward binding whereas the (S)-α-amino-γ-turn prefers the reverse binding. Considering that both the binding affinity and specificity of 5, which was prepared as a control, were shown to be lower (ΔTm = 23.6 ± 0.4 °C, KD = 2.0 × 10−9 M, specificity = 1.20), these observations clearly indicate that the remarkable forward/reverse preferences of 3/4 are attributable to the chirality of the amino groups in the γ-turn. Collectively, we found that the binding orientation of cPIPs was controlled by the position and chirality of the amino groups in the γ-turn. To validate the reverse specificity regulation rule of the (S)α-amino-γ-turn, we additionally synthesized compound 12 (Figure 4a), which was expected to exhibit a reverse preference because of the effect of the (S)-α-amino-γ-turn. Because the Py−Im sequence of 12 is opposite the sequences in 1−5, we used the sequence 5′-GGTCATGACC-3′, which was used for the reverse binding of 1−5, for the forward binding of 12, and the sequence 5′-CCAGTACTGG-3′, which was used for the forward binding of 1−5, for the reverse binding of 12 (Figure 4b). The synthesis of 12 was almost the same as that of 4 except that we used a Py−Im dimer−trityl resin at the solidphase-synthesis step instead of a Py−trityl resin. The Tm and ΔTm values are shown in Figure 4c (PIP: 2 equiv) and Table S1 (PIP: 1.5 equiv), and the representative denaturing profiles are shown in Figure S1 (PIP: 2 equiv). The KD values obtained by the SPR assay using 12 are shown in Figure 4d and the SPR sensorgrams and ka and kd values are shown in Figure S7. As expected, 12 showed a high binding affinity in the reverse-binding orientation (ΔTm = 36.0 ± 0.3 °C, KD = 1.3 × 10−10 M). More importantly, 12 showed an extremely high reverse specificity (0.0026). These results reinforced the binding orientation regulation rule advocated here, where the (S)-chirality of the α-amino groups at the γ-turn determined the binding preference of the cPIPs to the reverse orientation. Molecular Modeling Studies on Binding Orientation of Compound 4. To explain the high reverse specificity of cPIPs with (S)-α-amino-γ-turn units, we constructed molecular models of DNA/cPIP complexes for two binding orientations of compound 4 based on previously determined crystal structures of cPIPs.13,14 The γ-turn unit parts of the energyminimized complex structures are shown in Figure 5.

However, 1 also strongly bound to the reverse sequence (ΔTm = 36.4 ± 0.6 °C, KD = 1.2 × 10−10), showing low forward/ reverse specificity (1.22). Although the high binding affinity of 1 makes it certainly suitable for X-ray crystal structure analysis, its low orientation specificity reduces the sequence specificity toward the target sequences and, therefore, increases unintended binding, which limits its biological application. Furthermore, as we had expected, 2, which is an enantiomer of 1, bound strongly to the reverse sequence (ΔTm = 41.8 ± 0.6 °C, KD = 1.4 × 10−10 M). However, like in the case of 1 the forward/reverse specificity of 2 was almost equal (0.88). These results indicate that the chirality of the β-amino groups does not influence the orientation specificity greatly. On the other hand, cPIPs with the α-substituted γ-turn exhibited different binding properties. While 3 also showed a high binding affinity to the forward sequence (ΔTm = 40.0 ± 0.7 °C, KD = 7.1 × 10−10 M), like 1, it strongly preferred the forward-binding orientation (9.72), and it was obviously superior to 1 in terms of orientation specificity. This result was consistent with a previous report and may be due to the steric bulk of the amino groups in the minor groove.13,21 Significantly, 4, which is an enantiomer of 3, showed a very high binding affinity to the reverse sequence (ΔTm = 32.2 ± 0.7 °C, KD = 4.0 × 10−11 M), and it possessed C

DOI: 10.1021/jacs.9b05516 J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Article

Journal of the American Chemical Society Table 2. Summary of SPR Assaysa

KD (reverse)/KD (forward) [i]Determined by fitting with a 1:1 binding model with mass transfer. [ii] Determined by fitting with a 1:1 Langmuirbinding model. a

Figure 4. (a) Synthesis of compound 12. (b) Binding orientation of 12. (c) Tm and ΔTm values of 12. (d) Results of the SPR assay by using 12. [a] KD (reverse)/KD (forward). [i]Determined by a general fitting mode. [ii]Determined by fitting with a 1:1 binding model with mass transfer.

down and the α-amine is at the axial position (Figure 5d). In the conformation shown in Figure 5a−c, the amino group may be directed toward the wall or floor of the minor groove, which may cause a steric clash. However, when cPIPs with (S)-α-γ-turns take the reversebinding orientation, they can adopt the conformation shown in

As Dervan et al. mentioned,13,14 γ-turns can adopt four conformations in which the β-methylene faces up and the αamine is at the axial position (Figure 5a), the β-methylene faces down and the α-amine is at the equatorial position (Figure 5b), the β-methylene faces up and the α-amine is at the equatorial position (Figure 5c), and the β-methylene faces D

DOI: 10.1021/jacs.9b05516 J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society

■

Article

AUTHOR INFORMATION

Corresponding Authors

*[email protected] (T.B.) *[email protected] (H.S.) ORCID

Hiroshi Sugiyama: 0000-0001-8923-5946 Present Address #

Sefan Asamitsu: Department of Genomic Neurology, Institute of Molecular Embryology and Genetics, Kumamoto University, Kumamoto, Japan. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS This research was supported by AMED under Grant No. JP18am0301005 (Basic Science and Platform Technology Program for Innovative Biological Medicine), JP18am0101101 (Platform Project for Supporting Drug Discovery and Life Science Research (BINDS)), and JSPS KAKENHI (Grant No. JP16H06356 to H.S. and JP17J01932 to S.A.).

■ Figure 5. Minimized models of compound 4 in (a) the forwardbinding position with the β-methylene group directed up, (b) the forward-binding position with the β-methylene group directed down, (c) the reverse-binding position with the β-methylene group directed up, and (d) the reverse binding position with the β-methylene group directed down. The amine residues are colored in purple.

Figure 5d, where the steric interaction may be relieved by directing the amino group upward and out of the minor groove. This surmise suggests that the sterically favored conformation of the γ-turn of (S)-α-cPIPs might contribute to the preferred reverse orientation.

■

CONCLUSION In this study, using improved macrolactamization conditions, we readily synthesized cPIPs with different γ-turn units to examine their binding profiles in terms of their binding orientation preferences. Our results clearly demonstrated that cPIPs with a (R)-amino-γ-turn preferred a forward-binding mode, whereas the ones with a (S)-amino-γ-turn preferred a reverse-binding mode. This effect was extremely pronounced when the amino groups were located on the α-position of the γ-turn. These findings open an avenue to a precise design of the targeted sequences of cPIPs with reduced unintended binding in biological applications. Moreover, by designing a reverse orientation preferring PIPs, we may be able to target a larger array of DNA sequences, including 5′-XXXG-3′ or 5′GXG-3′, whose sequences were thought to be unable to be targetable by the forward-binding modes.23−25

■

REFERENCES

(1) Kopka, M. L.; Yoon, C.; Goodsell, D.; Pjura, P.; Dickerson, R. E. The molecular origin of DNA-drug specificity in netropsin and distamycin. Proc. Natl. Acad. Sci. U. S. A. 1985, 82, 1376. (2) Kopka, M. L.; Yoon, C.; Goodsell, D.; Pjura, P.; Dickerson, R. E. Binding of an antitumor drug to DNA. Netropsin and C-G-C-G-A-AT-T-BrC-G-C-G. J. Mol. Biol. 1985, 183, 553. (3) Lown, J. W.; Krowicki, K.; Bhat, U. G.; Skorobogaty, A.; Ward, B.; Dabrowiak, J. C. Molecular recognition between oligopeptides and nucleic acids: novel imidazole-containing oligopeptides related to netropsin that exhibit altered DNA sequence specificity. Biochemistry 1986, 25, 7408. (4) Lown, J. W. Lexitropsins in antiviral drug development. Antiviral Res. 1992, 17, 179. (5) Dervan, P. B.; Edelson, B. S. Recognition of the DNA minor groove by pyrrole-imidazole polyamides. Curr. Opin. Struct. Biol. 2003, 13, 284. (6) Dervan, P. B.; Bürli, R. W. Sequence-specific DNA recognition by polyamides. Curr. Opin. Chem. Biol. 1999, 3, 688. (7) Morita, K.; Suzuki, K.; Maeda, S.; Matsuo, A.; Mitsuda, Y.; Tokushige, C.; Kashiwazaki, G.; Taniguchi, J.; Maeda, R.; Noura, M.; Hirata, M.; Kataoka, T.; Yano, A.; Yamada, Y.; Kiyose, H.; Tokumasu, M.; Matsuo, H.; Tanaka, S.; Okuno, Y.; Muto, M.; Naka, K.; Ito, K.; Kitamura, T.; Kaneda, Y.; Liu, P. P.; Bando, T.; Adachi, S.; Sugiyama, H.; Kamikubo, Y. Genetic regulation of the RUNX transcription factor family has antitumor effects. J. Clin. Invest. 2017, 127, 2815. (8) Kawamoto, Y.; Sasaki, A.; Chandran, A.; Hashiya, K.; Ide, S.; Bando, T.; Maeshima, K.; Sugiyama, H. Targeting 24 bp within Telomere Repeat Sequences with Tandem Tetramer PyrroleImidazole Polyamide Probes. J. Am. Chem. Soc. 2016, 138, 14100. (9) Ohtsuki, A.; Kimura, M. T.; Minoshima, M.; Suzuki, T.; Ikeda, M.; Bando, T.; Nagase, H.; Shinohara, K.; Sugiyama, H. Synthesis and properties of PI polyamide−SAHA conjugate. Tetrahedron Lett. 2009, 50, 7288. (10) Taniguchi, J.; Feng, Y.; Pandian, G. N.; Hashiya, F.; Hidaka, T.; Hashiya, K.; Park, S.; Bando, T.; Ito, S.; Sugiyama, H. Biomimetic Artificial Epigenetic Code for Targeted Acetylation of Histones. J. Am. Chem. Soc. 2018, 140, 7108. (11) Cho, J.; Parks, M. E.; Dervan, P. B. Cyclic polyamides for recognition in the minor groove of DNA. Proc. Natl. Acad. Sci. U. S. A. 1995, 92, 10389. (12) Morinaga, H.; Bando, T.; Takagaki, T.; Yamamoto, M.; Hashiya, K.; Sugiyama, H. Cysteine Cyclic Pyrrole-Imidazole Polyamide for Sequence-Specific Recognition in the DNA Minor Groove. J. Am. Chem. Soc. 2011, 133, 18924.

ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/jacs.9b05516. Experimental procedures, representative Tm denaturing graphs, SPR sensorgrams, HPLC, mass, and 1H NMR charts (PDF) E

DOI: 10.1021/jacs.9b05516 J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Article

Journal of the American Chemical Society (13) Chenoweth, D. M.; Dervan, P. B. Allosteric modulation of DNA by small molecules. Proc. Natl. Acad. Sci. U. S. A. 2009, 106, 13175. (14) Chenoweth, D. M.; Dervan, P. B. Structural Basis for Cyclic PyIm Polyamide Allosteric Inhibition of Nuclear Receptor Binding. J. Am. Chem. Soc. 2010, 132, 14521. (15) Chenoweth, D. M.; Harki, D. A.; Phillips, J. W.; Dose, C.; Dervan, P. B. Cyclic Pyrrole-Imidazole Polyamides Targeted to the Androgen Response Element. J. Am. Chem. Soc. 2009, 131, 7182. (16) Li, B. C.; Montgomery, D. C.; Puckett, J. W.; Dervan, P. B. Synthesis of Cyclic Py-Im Polyamide Libraries. J. Org. Chem. 2013, 78, 124. (17) White, S.; Baird, E. E.; Dervan, P. B. Orientation Preferences of Pyrrole-Imidazole Polyamides in the Minor Groove of DNA. J. Am. Chem. Soc. 1997, 119, 8756. (18) Minoshima, M.; Bando, T.; Sasaki, S.; Fujimoto, J.; Sugiyama, H. Pyrrole-imidazole hairpin polyamides with high affinity at 5′− CGCG−3′ DNA sequence; influence of cytosine methylation on binding. Nucleic Acids Res. 2008, 36, 2889. (19) Meier, J. L.; Yu, A. S.; Korf, I.; Segal, D. J.; Dervan, P. B. Guiding the Design of Synthetic DNA-Binding Molecules with Massively Parallel Sequencing. J. Am. Chem. Soc. 2012, 134, 17814. (20) Kang, J. S.; Meier, J. L.; Dervan, P. B. Design of SequenceSpecific DNA Binding Molecules for DNA Methyltransferase Inhibition. J. Am. Chem. Soc. 2014, 136, 3687. (21) Herman, D. M.; Baird, E. E.; Dervan, P. B. Stereochemical Control of the DNA Binding Affinity, Sequence Specificity, and Orientation Preference of Chiral Hairpin Polyamides in the Minor Groove. J. Am. Chem. Soc. 1998, 120, 1382. (22) Asamitsu, S.; Li, Y.; Bando, T.; Sugiyama, H. Ligand-Mediated G-Quadruplex Induction in a Double- Stranded DNA Context by Cyclic Imidazole/Lysine Polyamide. ChemBioChem 2016, 17, 1317. (23) Swalley, S. E.; Baird, E. E.; Dervan, P. B. Discrimination of 5′GGGG-3′,5′-GCGC-3′, and 5′-GGCC-3′ Sequences in the Minor Groove of DNA by Eight-Ring Hairpin Polyamides. J. Am. Chem. Soc. 1997, 119, 6953. (24) Wang, C. C.; Ellervik, U.; Dervan, P. B. Expanding the Recognition of the Minor Groove of DNA by Incorporation of βAlanine in Hairpin Polyamides. Bioorg. Med. Chem. 2001, 9, 653. (25) Turner, J. M.; Swalley, S. E.; Baird, E. E.; Dervan, P. B. Aliphatic/Aromatic Amino Acid Pairings for Polyamide Recognition in the Minor Groove of DNA. J. Am. Chem. Soc. 1998, 120, 6219.

F

DOI: 10.1021/jacs.9b05516 J. Am. Chem. Soc. XXXX, XXX, XXX−XXX