Interaction between human telomeric G-quadruplexes characterized by single molecule magnetic tweezers*

Project supported by the National Natural Science Foundation of China (Grant Nos. 11474346 and 11774407), the Key Research Program of Frontier Sciences, Chinese Academy of Sciences (Grant No. QYZDB-SSW-SLH045), and the National Key Research and Development Program, China (Grant No. 2016YFA0301500).

Wang Yi-Zhou1, 3, Hou Xi-Miao2, Ju Hai-Peng1, Xiao Xue1, 3, Xi Xu-Guang2, Dou Shuo-Xing1, 3, Wang Peng-Ye1, 3, ‡, Li Wei1, §
Key Laboratory of Soft Matter Physics, Beijing National Laboratory for Condensed Matter Physics, Institute of Physics, Chinese Academy of Sciences, Beijing 100190, China
School of Life Sciences, Northwest A & F University, Yangling 712100, China
School of Physical Sciences, University of Chinese Academy of Sciences, Beijing 100049, China

 

† Corresponding author. E-mail: pywang@iphy.ac.cn weili007@iphy.ac.cn

Project supported by the National Natural Science Foundation of China (Grant Nos. 11474346 and 11774407), the Key Research Program of Frontier Sciences, Chinese Academy of Sciences (Grant No. QYZDB-SSW-SLH045), and the National Key Research and Development Program, China (Grant No. 2016YFA0301500).

Abstract

Human telomeric G-quadruplex plays a crucial role in regulating the genome stability. Despite extensive studies on structures and kinetics of monomeric G-quadruplex, the interaction between G-quadruplexes is still in debate. In this work, we employ magnetic tweezers to investigate the folding and unfolding kinetics of two contiguous G-quadruplexes in 100-mM K+ buffer. The interaction between G-quadruplexes and the consequent effect on the kinetics of G-quadruplex are revealed. The linker sequence between G-quadruplexes is further found to play an important role in the interaction between two G-quadruplexes. Our results provide a high-resolution insight into kinetics of multimeric G-quadruplexes and genome stability.

1. Introduction

In human cells, telomeres at the termini of chromosomes are composed of tandemly repeated sequence 5’-TTAGGG and terminate with a single-strand 3’-overhang of approximately 200 nucleotides.[1] These repetitive sequences form multiple G-quadruplexes,[2,3] which are stacked with two or more G-quartets containing four coplanar guanines through Hoogsteen hydrogen bonds.[4,5] The G-quadruplex array can protect the chromosomes from damaging and end-fusing,[6] playing crucial functions in regulating the chromosome stability, gene transcription, recombination, and replication.[710]

Extensive studies have been carried out on the structures and kinetics of monomeric G-quadruplex including DNA G-quadruplex[1113] and RNA G-quadruplex.[14] The highly polymorphic structures of G-quadruplex are revealed, including the parallel, anti-parallel, hybrid-1 and hybrid-2 structure.[1517] Recently, single molecular methods including optic tweezers, magnetic tweezers and FRET have been employed for investigating the folding kinetics of monomeric G-quadruplex.[1113,18,19] All these studies revealed the monomeric G-quadruplex folding and unfolding mechanism at the molecular level and gave deeper insights into the telomere maintenance and design of ligands for anti-cancer purposes.

However, whether the interaction between G-quadruplexes exists, the crucial question for understanding the higher order of multimeric G-quadruplexes, is still in debate. Thermodynamics assays indicated an unfavorable coupling free energy in the multimeric assembly of G-quadruplexes,[20,21] while molecular simulation studies[22] and photoreaction assays[23] suggested the existence of stable interaction between two hybrid-type G-quadruplexes. Recently, the mechanic unfolding assays by optic tweezers on the full length human telomeric overhang show most of G-quadruplexes are independent, while a minor population of higher-order interactions between G-quadruplexes are observed.[24] These controversial results indicate that the interaction between G-quadruplexes is far from being understood.

Single molecule methods provide a powerful tool to investigate the kinetics of G-quadruplex, which helps to elucidate whether the interaction between G-quadruplexes exists and how the interaction affects the kinetics of G-quadruplexes. In this work, we construct the simplest multimeric system with two G-quadruplexes in vitro and investigate the folding and unfolding kinetics of the two contiguous G-quadruplexes in 100-mM K+ buffer by using magnetic tweezers. In our experiment, two dinstinct unfolding modes are observed, implying the existence of the interaction between two G-quadruplexes. Interestingly, the interaction destablizes one G-quadruplex unit and stablizes the other. Besides, we also demonstrate that the TTA linker plays a crucial role in the interaction between G-quadruplexes. Our findings shed light on the stability and interaction mechanism of G-quadruplexes at human telomeres, which should be helpful for understanding the regulation of gemone stability.

2. Materials and methods
2.1. Preparation of DNA construct

The DNA construct contains three parts: a 699-bp double strand DNA handle with biotin modification, a single strand telomeric DNA sequence 5’-GGG(TTAGGG)7 (wtTel45) and a 2271-bp double strand DNA handle with digoxigenin modification. The contour length of the DNA construct is ∼ 1 μm. Two handles were prepared by PCR amplification of PBR322 with primers. The PCR products were purified by a QIAquick PCR Purification Kit (QIAGEN, Germany), and then digested by XbaI and KpnI (Ipswich, MA) respectively. The digested products were further purified using the QIAquick PCR Purification Kit, and the concentrations of handle products were determined with a UV spectrophotometer. To ligate single strand DNAs with handles, the ssDNA that can form one or two telomeric G-quadruplexes (blue) was annealed with flank 1 and flank 2 in 1:1:1 ratio by slowly cooling down from 95 °C to 25 °C, then ligated with two dsDNA handles in 1:1:1 ratio by T4 ligase at 16 °C overnight. Finally, Agarose gels were performed to check the final product, and the DNA constructs were stored at −20 °C.

2.2. Magnetic tweezers stretching experiment

Our single molecule manipulation experiments were performed by magnetic tweezers. As shown in Fig. 1(a), the two ends of the DNA construct are tethered via digoxigenin and anti-digoxigenin ligation to a glass coverslip and via biotin-streptavidin ligation to a 1-μm diameter Dynabeads (Invitrogen Norway). Two small NdFeB magnets on the DNA constructs are controlled to pull on the Dynabeads and thus stretch the DNA molecule. The real time position of the bead is monitored by a JAI Giga-Ethernet CCD camera at 60 Hz through an inverted microscope objective (Olympus 100 × 1.2, oil immersion). The extension (end to end distance) of the DNA construct is determined at nanometer resolution by analyzing the diffraction pattern of Dynabeads.[25] The 1-μm diameter polystyrene reference beads are stuck on the coverslip using nitrocellulose and tracked simultaneously to eliminate the shift of the device. To ensure that only one DNA construct was tethered to a bead, we rotated magnet pairs for ∼ 50 turns at ∼ 5 pN and checked whether the bead is rotation-constrained, namely, more than one DNA being tethered to the bead. Our magnetic tweezers were controlled by a home-built LabVIEW program. In the force ramp experiment, magnet pairs were moved at nonlinear step size to keep a constant loading rate. The step size can be calculated according to the force-magnet position relationship fitted with exponential function.[26,27] (see Appendix A “Force calibration”)

Fig. 1. (color online) (a) Schematic setup of magnetic tweezers (not to scale). The 45 nt wtTel45 sequence is tethered between a coverslip and a superparamagnetic bead with two 699 bp and 2271 bp dsDNA handles. (b) Typical trajectory with two distinct ruptures indicate two stable G-quadruplexes formed in wtTel45 sequence.
2.3. Surface treatment and flow cell

To bound the reference beads, the inner surface of coverslips was coated with ∼ 0.2% w/v nitrocellulose (Sigma-Aldrich) and 0.03% w/v polystyrene beads (ACMEmicrospheres) in alcohol and then heated at 150 °C for 5 min,[28] allowing the nitrocellulose solvent to evaporate and the polystyrene beads to melt onto the surface as reference beads. Afterwards, 10-mg/ml anti-digoxigenin was used to modify the surface overnight at 37 °C and then dealt with passivation buffer (10-mg/ml BSA, 1-mM EDTA, 10-mM pH 7.4 phosphate buffers, 10 mg/ml Pluronic F127 surfactant (Sigma-Aldrich), 3-mM NaN3) at room temperature for 4 h to avoid non-specific ligation between magnetic beads and the surface. In our experiments, the DNA constructs were diluted to 40 pM and blended with 10 × diluted Dynabeads MyOne in 1:1 volume ratio for 30 min. The DNA-bead mixture was injected into the flow cell and incubated for 30 min, then rinsing away the unconnected beads with TE buffer.

3. Results and discussion
3.1. Dimeric G-quadruplexes less stable than monomeric G-quadruplex

As shown in Fig. 1(a), the wildtype human telomeric G-quadruplex sequence (wtTel45) 5’-GGG(TTAGGG)7 with a lower 699-bp dsDNA handle and an upper 2271-bp dsDNA handle is tethered specifically between an antidigoxigenin coated coverslip and a streptavidin-coated paramagnetic bead (MyOne Invitrogen). Force exerted on the paramagnetic bead is tuned by adjusting the magnets vertically and the extension is traced at a nanometer resolution based on the diffraction image of the bead.[25] To demonstrate that our sample can form two contiguous G-quadruplexes, force jump experiments were performed. Figure 1(b) shows the force-jump experiments between 0.1 pN and 7 pN and two extension steps around 9 nm reveal two G-quadruplex ruptures at the high tension. According to other single molecule experiments on double strand DNA (dsDNA), conformation change cannot occur when the force is less than 60 pN.[2934] Repeated measurements and similar results reveal that two stable G-quadruplexes are formed in the wtTel45 sequence, which is consistent with the results previously reported by AFM and NMR.[35,36]

To further quantify the two G-quadruplex ruptures formed in wtTel45 sequence, we perform the force-ramp experiments from 0.7 pN to 8.3 pN at the loading rate r = 0.16 pN/s. As shown in Fig. 2(a), two sequential extension jumps in each trajectory correspond to the unfolding of two G-quadruplexes, which is consistent with the result of previous force-jump experiment (Fig. 1(b)). For clarity, the trajectories are shifted in extension. The first rupture event, which happens with co-existence of the other G-quadruplex, is a key target for the investigation of the interaction between the G-quadruplexes. We repeat the force-ramp measurement and derive the unfolding force distribution p for the first G-quadruplex rupture as shown in the upper panel of Fig. 2(b).

Fig. 2. (color online) Force ramp experiments of dimeric G-quadruplexes and mono G-quadruplex. (a) Typical force responses of dual G-quadruplexes in a stretching cycle (upper panel) with a loading rate of 16 pN/s (lower panel). (b) Unfolding force distribution of the first G-quadruplex in the dimeric G-quadruplexes fitted with Evans model (upper panel, N = 200) and monomeric G-quadruplex (lower panel, N = 198).

To compare the tension response of G-quadruplex in the presence and in the absence of another G-quadruplex, the same force-ramp measurements are performed on the monomeric G-quadruplex formed in wtTel21 sequence 5’-GGG(TTAGGG)3 (see Appendix A in Fig. S3). The corresponding unfolding force distribution p of monomeric G-quadruplex is shown in the lower panel in Fig. 2(b). Compared with the monomeric G-quadruplex, the first ruptured G-quadruplex in the wtTel45 sequence prefers to unfold at lower forces. The experiment shows that the mechanical stability of G-quadruplex is attenuated by the existence of another G-quadruplex, which suggests the existence of interaction between G-quadruplexes. Our results agree well with the reported results that dimeric G-quadruplexes melt before monomeric G-quadruplex in previous CD thermal analysis.[22,37]

In our force-ramp measurement, the force distribution in the force spectroscopy measurements of single molecules can be described by Evans’ model[38] as follows: where r is the loading rate, F is the external tension, kB is the Boltzmann constant, and T is the absolute temperature. The corresponding to the unfolding rate at F = 0 pN and Δxu corresponding to the distance between the folded state and the reaction energy barrier in the reaction coordinate, are treated as the fitting parameters. The best fitting parameters of the first unfolding G-quadruplex are s−1, Δxu = 2.1 ± 0.7 nm (mean ± SD), and the best fitting parameters of monomeric G-quadruplex are s−1, Δxu1 = 2.5 ± 0.8 nm. The first G-quadruplex in the wtTel45 sequence presents higher unfolding rate (at F = 0 pN) and lower unfolding force (Fig. 2(b)) than the monomeric G-quadruplex, which also indicates that the G-quadruplex in the wtTel45 sequence is less stable in the presence of another G-quadruplex due to the interaction between the G-quadruplexes. It should be noted that the reaction distance Δx determined in the force ramp experiment is a little larger than results in Refs. [39] and [40]. A possible explanation is that Bell’s model is no longer valid at low transition forces at such a slow loading rate, because it ignores the entropic elastic contribution in transition kinetics that is often significant at low forces.[41]

3.2. Distinct unfolding kinetics suggesting the interaction between two contiguous G-quadruplexes

To reveal the interaction between the two G-quadruplexes and to understand the consequence effect on the G-quadruplex, the force clamp measurement is performed by tracing the folding and unfolding kinetics of the dimeric G-quadruplexes at different forces. As shown in Fig. 3(a), the length of the samples are traced from 1.7 pN to 8 pN. At each tension, the measurement is sustained for fifteen min, which provides an adequate time to reveal the folding and unfolding kinetics of the G-quadruplex at a given tension. Two different unfolding processes are distinguished (the green and yellow regions for the first and second unfolding process respectively in Fig. 3(a)), which correspond to the sequential ruptures of the two formed G-quadruplexes in the wtTel45 sequence. The folding and unfolding transition of G-quadruplex are controlled by the external tension that tilts the equilibrium of G-quadruplex from folded state to unfolded state.

Fig. 3. (color online) Folding and unfolding kinetics of two G-quadruplexes in wtTel45 sequence. (a) Trajectory (upper panel) under different constant forces (lower panel) showing two ruptures of G-quadruplex. (b) Relationships between reaction rate ku, kf, and tension F. The equilibrium tension Feq can be read from the point where ku = kf. The error bars reflect a ∼ 10% error in force determination. (c) Different folding and unfolding kinetics of two G-quadruplex near Feq. For the first G-quadruplex, τ0 = 20.5 s (left panel) and for the second G-quadruplex, τ0 = 4.0 s (right panel).

The reversible hopping behavior provides a framework for unraveling the different kinetics of two G-quadruplexes formed in the wtTel45 sequence. At a constant tension near the unfolding tension, G-quadruplex hops between folded and unfolded state. The unfolding rate ku from folded to unfolded state can be described as an Arrhenius-like expression[31,32] where F is the external tension, kB is the Boltzmann constant, and T is the absolute temperature. A similar relationship for the folding rate kf can be calculated accordingly. The unfolded rate ku increases and the folded rate kf decreases with the increase of tension as shown in Fig. 3(b). At the equilibrium force Feq, G-quadruplex hops between the folded and unfolded states with the same rate ku = kf and the average dwelling time τ0 of the folded state is equal to that of the unfolded state. In Fig. 3(b), the equilibrium force Feq can be read from the point where the folded rate kf (red line) equals the unfolded rate ku (blue line). Near Feq, the folding and unfolding kinetics of the first and second G-quadruplexes are shown in Fig. 3(c), corresponding to the region marked with a black rectangle in Fig. 3(a). The values of average dwelling time τ0 are 20.5 s and 4.0 s for the first and second G-quadruplex, respectively (Fig. 3(c)). The average dwelling time of the first G-quadruplex, much larger than that of the second G-quadruplex at Feq, indicates the slower folding and unfolding kinetics at the equilibrium tension in the presence of another G-quadruplex. We perform the same measurement for the monomeric G-quadruplex, and the average dwelling time τ0 = 2.7 s at Feq is revealed, which is similar to that of the second G-quadruplex in the wtTel45 sequence (see Appendix A, Fig. S4 online). Unlike the monomeric G-quadruplex and the second unfolding G-quadruplex, the first unfolding G-quadruplex shows slower folding and unfolding kinetics, suggesting the existence of the interaction between G-quadruplexes.

By analyzing the different kinetics of the two G-quadruplexes formed in the wtTel45 sequence, we derive the free energy ΔG0, an important quantity describing the stability of G-quadruplex. The relationship between the equilibrium rate constant keq = ku/kf and external tension F can be described as where is the equilibrium rate constant at F = 0 pN, ΔG(F) = FΔx + Δ Φ (F) is the free energy cost for unfolding the G-quadruplex, and accounts for the force-dependent entropic free energy change.[18] To calculate Δ Φ (F) the ssDNA force-extension curve xssDNA(F) can be described by a phenomenological model[42] as follows: where the empirical parameters h = 0.34 nm, a1 = 0.21, a2 = 0.34, f1 = 0.0037 pN, f2 = 2.9 pN, and f3 = 8000 pN. The salt dependent parameter a3 = 2.1 ln ([K+]/0.0025)/ln(0.15/0.0025), where [K+] is the salt concentration. The project length along the force direction of folded G-quadruplex is given by where l ∼ 1.7 nm.[18] At Feq such that keq = ku/kf = 1, we can derive free energy difference ΔG0 by calculating the value of Δ Φ(Feq). From the equation above, the values of free energy ΔG0 for the first and second G-quadruplex in the wtTel45 sequence are obtained to be 2.9 kBT and 5.0 kBT respectively. For the monomeric G-quadruplex, the free energy ΔG0 is 3.6 kBT. The lower free energy of the first G-quadruplex reveals that the interaction between G-quadruplexes weakens the stability of G-quadruplex.

3.3. Effect of linker DNA on interaction between two G-quadruplexes

In the wtTel45 sequence, the two G-quadruplexes are connected via a TTA linker, which is suggested to play an important role in the interaction between two contiguous G-quadruplexes.[22] We construct the mutant Tel45 sequence by replacing the TTA linker with an AAA linker and perform the measurements as shown in Fig. 4. In a representative trajectory, two sequential reversible unfolding processes are observed, corresponding to the two G-quadruplexes formed in (Fig. 4(a)). By plotting the folding and unfolding rate constant versus force, we obtain the equilibrium forces 1.6 pN and 4.1 pN for the first and second unfolding G-quadruplex in the mutant Tel45 sequence respectively (Fig. 4(b)). Compared with G-quadruplexes formed in the wtTel45 sequence, the two G-quadruplexes in the mutant Tel45 sequence show a relatively small equilibrium force, implying two less stable G-quadruplexes in the mutant Tel45 sequence. Near Feq, the two G-quadruplexes hop between folded and unfolded state with a similar rate. The values of average dwelling time τ for the first and second G-quadruplex are 4.2 s and 3.6 s respectively (Fig. 4(c)), which are similar to that of the monomeric G-quadruplex, indicating similar folding and unfolding kinetics of the G-quadruplex in the mutant Tel45 sequence to those of the monomeric G-quadruplex. Our results demonstrate that the TTA linker plays a crucial role in the interaction between G-quadruplexes, which is consistent with previous simulation.[22]

Fig. 4. (color online) Folding and unfolding kinetics of two G-quadruplexes in mutant Tel45 sequence. (a) Trajectories (upper panel) under different constant tensions (lower panel) show two ruptures of G-quadruplex. (b) Relationships between rate constant ku, kf, and tension F. (c) Near Feq, similar folding and unfolding kinetics of G-quadruplex. For the first G-quadruplex, τ0 = 4.2 s (left panel), and for the second G-quadruplex, τ0 = 3.6 s (right panel).
4. Conclusion and perspectives

In this work, we investigate the interaction between two G-quadruplexes formed in the wtTel45 sequence by magnetic tweezers. In the presence of another G-quadruplex, G-quadruplex is less stable and unfolded at lower tension. The different folding and unfolding kinetics of the two G-quadruplexes formed in the wtTel45 sequence due to the interaction between the G-quadruplexes are further revealed. All the work indicates that the interaction between G-quadruplexes exists and the interaction affects the kinetics of the G-quadruplex. We further reveal that the linker between G-quadruplexes is an important factor for the interaction in the force clamp experiments for the mutant Tel45 sample; however, the reason for the interaction is complicated. We perform the force clamp experiments on the (TTAGGG)4 and GGG(TTAGGG)3TTA sequence, which form G-quadruplex with a TTA residue on 5’ and 3’ respectively (see Appendix A, Fig. S5 and Fig. S6 online). Both the two G-quadruplexes with TTA residue have similar unfolding kinetics to the monomeric G-quadruplex (τ0 = 2.7 s), suggesting that the interaction is involved in two contiguous G-quadruplexes besides the TTA linker. It has been demonstrated that a long single strand G-rich DNA can form a higher order structure containing consecutive G-quadruplex. Our findings reveal the mechanism of G-quadruplex interaction, which should be helpful in the in-depth understanding of the telomeric G-quadruplex stability and regulation in vivo.

Reference
[1] Moyzis R K Buckingham J M Cram L S Dani M Deaven L L Jones M D Meyne J Ratliff R L Wu J R 1988 Proc. Natl. Acad. Sci. USA 85 6622
[2] Biffi G Tannahill D McCafferty J Balasubramanian S 2013 Nat. Chem. 5 182
[3] Todd A K Johnston M Neidle S 2005 Nucleic Acids Res. 33 2901
[4] Kaur P Wu D Lin J Countryman P Bradford K C Erie D A Riehn R Opresko P L Wang H 2016 Sci. Rep. 6 20513
[5] Dai J Carver M Yang D 2008 Biochimie 90 1172
[6] Maizels N 2006 Nat. Struct. Mol. Biol. 13 1055
[7] Cahoon L A Seifert H S 2009 Science 325 764
[8] Rodriguez R Miller K M Forment J V Bradshaw C R Nikan M Britton S Oelschlaegel T Xhemalce B Balasubramanian S Jackson S P 2012 Nat. Chem. Biol. 8 301
[9] Siddiqui-Jain A Grand C L Bearss D J Hurley L H 2002 Proc. Natl. Acad. Sci. USA 99 11593
[10] Wang H Nora G J Ghodke H Opresko P L 2011 J. Biol. Chem. 286 7479
[11] Dhakal S Cui Y Koirala D Ghimire C Kushwaha S Yu Z Yangyuoru P M Mao H 2013 Nucleic Acids Res. 41 3915
[12] Gray R D Buscaglia R Chaires J B 2012 J. Am. Chem. Soc. 134 16834
[13] Li W Hou X-M M Wang P-Y Y Xi X G G Li M 2013 J. Am. Chem. Soc. 135 6423
[14] Laguerre A Hukezalie K Winckler P Katranji F Chanteloup G Pirrotta M Perrier-Cornet J M M Wong J M Monchaud D 2015 J. Am. Chem. Soc. 137 8521
[15] Wang Y Patel D J 1993 Structure 1 263
[16] Parkinson G N Lee M P Neidle S 2002 Nature 417 876
[17] Luu K N Phan A T Kuryavyi V Lacroix L Patel D J 2006 J. Am. Chem. Soc. 128 9963
[18] You H Zeng X Xu Y Lim C J Efremov A K Phan A T Yan J 2014 Nucleic Acids Res. 42 8789
[19] Ju H P Wang Y Z You J Hou X M Xi X G Dou S X Li W Wang P Y 2016 Acs. Omega. 1 244
[20] Petraccone L Spink C Trent J O Garbett N C Mekmaysy C S Giancola C Chaires J B 2011 J. Am. Chem. Soc. 133 20951
[21] Yu H Gu X Nakano S-i Miyoshi D Sugimoto N 2012 J. Am. Chem. Soc. 134 20060
[22] Petraccone L Trent J O Chaires J B 2008 J. Am. Chem. Soc. 130 16530
[23] Li Y Sugiyama H 2015 Chem. Commun. 51 8861
[24] Abraham Punnoose J Cui Y Koirala D Yangyuoru P M Ghimire C Shrestha P Mao H 2014 J. Am. Chem. Soc. 136 18062
[25] Lionnet T Allemand J F F Revyakin A Strick T R Saleh O A Bensimon D Croquette V 2012 Cold. Spring. Harb. Protoc. 2012 34
[26] Chen H Fu H Zhu X Cong P Nakamura F Yan J 2011 Biophys. J. 100 517
[27] Zhao X Zeng X Lu C Yan J 2017 Nanotechnology 28 414002
[28] Cnossen J P Dulin D Dekker N H 2014 Rev. Sci. Instrum. 85 103712
[29] Smith S B Cui Y Bustamante C 1996 Science 271 795
[30] Cluzel P Lebrun A Heller C Lavery R Viovy J L Chatenay D Caron F 1996 Science 271 792
[31] Fu H Chen H Marko J F Yan J 2010 Nucleic Acids Res. 38 5594
[32] Fu H Chen H Zhang X Qu Y Marko J F Yan J 2011 Nucleic Acids Res. 39 3473
[33] Zhang X Chen H Fu H Doyle P S Yan J 2012 Proc. Natl. Acad. Sci. USA 109 8103
[34] Zhang X Chen H Le S Rouzina I Doyle P S Yan J 2013 Proc. Natl. Acad. Sci. USA 110 3865
[35] Xu Y Ishizuka T Kurabayashi K Komiyama M 2009 Angew. Chem. Int. Ed. Engl. 48 7833
[36] Hänsel R Löhr F Trantirek L Dötsch V 2013 J. Am. Chem. Soc. 135 2816
[37] Cousins A R Ritson D Sharma P Stevens M F Moses J E Searle M S 2014 Chem. Commun. 50 15202
[38] Evans E Ritchie K 1997 Biophys. J. 72 1541
[39] Dhakal S Cui Y Koirala D Ghimire C Kushwaha S Yu Z Yangyuoru P M Mao H 2013 Nucleic Acids Res. 41 3915
[40] You H Zeng X Xu Y Lim C J Efremov A K Phan A T Yan J 2014 Nucleic Acids Res. 42 8789
[41] Yuan G Le S Yao M Qian H Zhou X Yan J Chen H 2017 Angew. Chem. Int. Ed. 56 5490
[42] Cocco S Yan J Léger J-F F Chatenay D Marko J F 2004 Phys. Rev. 70 11910
[43] Strick T R Allemand J F Bensimon D Bensimon A Croquette V 1996 Science 271 1835
[44] Chen H et al. 2011 Biophys. J. 100 517
[45] Lipfert J Hao X Dekker N H 2009 Biophys. J. 96 5040
[46] Aggarwal T Materassi D Davison R Hays T Salapaka M 2012 Cellular and Molecular Bioengineering 5 14