Fundamental aspects of the nucleic acid i-motif structures

S. Benabou ^a, A. Aviñó ^b, R. Eritja ^b, C. González ^c and R. Gargallo *^a
aDepartment of Analytical Chemistry, University of Barcelona, Martí i Franqués 1-11, E-08028 Barcelona, Spain. E-mail: raimon_gargallo@ub.edu
^bInstitute for Advanced Chemistry of Catalonia (IQAC-CSIC), CIBER-BBN Networking Centre on Bioengineering, Biomaterials and Nanomedicine, Jordi Girona 18-26, E-08034 Barcelona, Spain
^cInstitute of Physical Chemistry “Rocasolano”, CSIC, Serrano 119, E-28006 Madrid, Spain

First published on 29th April 2014

---

---

S. Benabou

Sanae Benabou is a PhD student at the Department of Analytical Chemistry (University of Barcelona). Her current research interest is focused on the study of chemical equilibria involving G-quadruplex and i-motif structures found in sequences near the promoter region of the N-myc gene by means of spectroscopic techniques and multivariate data analysis methods.

A. Aviñó

Anna Aviñó is a research associate at the Biomedical Research Networking Center in Bioengineering and Nanomedicine (CIBER-BBN). She obtained her PhD in Chemistry from the University of Barcelona in 1996. She then pursued postdoctoral studies at Rega Institute in Leuven and at EMBL in Heidelberg. She works at the Nucleic Acids Chemistry Group of the Institute for Advanced Chemistry of Catalunya. Her areas of interest include synthesis and structural studies of quadruplex, triplex and i-motif, development of RNA nanoconjugates to improve the inhibitory activity of interference RNAs and new intercalating compounds in double-stranded DNA.

R. Eritja

Ramon Eritja studied Chemistry and Pharmacy at University of Barcelona, receiving his PhD in Chemistry from the University of Barcelona in 1983. He then carried out postdoctoral studies in Dr Itakura’s and in Dr Caruther’s groups. In 1990 he became a group Leader at CSIC and in 1994 he transferred to EMBL as group Leader. He returned in 1999 at IQAC-CSIC leading the Nucleic Acid Chemistry Group. Since 2006 he has been a member of the Spanish Networking Center on Bioengineering, Biomaterials and Nanomedicine. In 2012 he was appointed director of the IQAC-CSIC. His research focuses on oligonucleotide synthesis for biomedical and nanotechnological applications.

C. González

Carlos González is full professor at the Rocasolano Institute of Physical Chemistry (IQFR) of the Spanish Council for Scientific Research (CSIC) in Madrid. With a background in Physics, his expertise area is the application of NMR methods to the study biomolecules. He has more than 25 years of experience in scientific research, with more than 100 peer-reviewed publications. His main scientific interest is the structural studies of nucleic acids and their complexes with other molecules.

R. Gargallo

Raimundo Gargallo is Senior Lecturer at the Department of Analytical Chemistry (University of Barcelona). He obtained his PhD in Chemical Sciences from the University of Barcelona (1997) and then carried out his postdoctoral research at the University of Innsbruck. He is co-author of sixty articles published in peer-reviewed international journals. His research is focused on the chemical equilibria involving G-quadruplex and i-motif structures. The effect of variables such as pH, temperature, ionic strength, ligands, and the complementary DNA strands on the conformational equilibria is studied by means of spectroscopic techniques and multivariate data analysis methods.

---

Introduction

Recent genomic research has revealed that around 98% of biological DNA is comprised of non-coding regions, such as regulatory elements, introns, repeat sequences and telomeres, among others. Many types of non-coding DNA sequences have important biological functions, including the transcriptional and translational regulation of protein-coding sequences. Other non-coding sequences have likely, but not yet determined, functions. These non-coding regions often comprise repetitive sequences that are patterns of nucleic acid sequences (DNA or RNA) that occur in multiple copies throughout the genome. The most common DNA structure under physiological conditions is B-DNA, which is a right-handed double helical structure with Watson–Crick base pairing. However, repetitive DNA sequences have the potential to fold into non-B DNA structures such as the hairpin or left-handed Z-form under certain experimental conditions. Since the non-B DNA-forming sequences may induce genetic instability and, consequently, may cause human diseases,¹ the molecular mechanism for their genetic instability has been investigated extensively.

G-rich sequences are able to fold into a non-B-DNA structure known as the G-quadruplex. In vitro studies have determined the conditions at which the G-quadruplex may be formed and the spatial arrangement of bases in it. Because the fact that, at physiological conditions of temperature, pH and ionic strength, G-quadruplex structures are stable, the formation of these structures in vivo must be expected. In fact, evidence for the formation of G-quadruplex structures in the genome of mammalian cells by using appropriate ligands in a cellular context has been described recently.^2–4

C-rich regions have also the ability to form a folded structure known as an i-motif (also named i-tetraplex or i-DNA).^5,6 This structure is the only known DNA structure that consists of parallel-stranded duplexes held together through intercalated base pairs. Other structures, such as B-DNA or G-quadruplex are spatial arrangements of DNA strands held together by means of stacked base pairs. Unlike the G-quadruplex, the formation of i-motif structures requires C base protonation to form the C·C⁺ base pair. In vitro studies have shown that i-motif structures are stable at pH values lower than 7 at the physiological conditions of temperature and ionic strength. However, its stability is very low at neutral and basic pH values. Because of this, the hypothetical role of the i-motif structure in biological processes is uncertain and, consequently, much more effort has been done in the study of other non-canonical structures such as the G-quadruplex. In revent years, however, the property of i-motif formation to narrow pH changes boosted the study of potential uses of this structure in the development of sensors and motors at the nanoscale.^7,8 Finally, the isolation of proteins that specifically bind to C-rich sequences prompted again the study and discussion about its hypothetical biological role.

Few works have been devoted to review the advances in knowledge of the structure and solution properties of i-motif structures, including their potential role in vivo. Gilbert and Feigon included a section about the i-motif in their review about multistranded DNA structures.⁹ The focus was mainly put on some striking features of a few structures previously determined by means of NMR or X-ray studies, like the parallel disposition of flanking base pairs on the i-motif core. The first review focused on the i-motif structure was published 14 years ago.¹⁰ In that work, the latest discoveries about this structure in human telomeric and centromeric sequences, as well as the isolation of proteins that bind specifically to C-rich sequences were discussed.

The role of non-B-DNA structures, such as G-quadruplex or i-motif structures, as potential regulatory elements in transcription has been discussed recently.¹¹ It has been proposed that negative supercoiling may favour the formation of single strands in local unwounded regions of DNA. These single strands may form non-B-DNA structures which could play a role in transcription regulation. This hypothesis is based on several evidences. First, the fact that most of the G-rich and C-rich sequences are in close proximity to the transcriptional start site suggest a role in transcription. In fact, ∼43% of promoter regions in genes have the potential to form at least one G-quadruplex structure and hence, one i-motif structure. Second, the existence of proteins and ligands that recognize specifically the G-rich or C-rich regions suggest a role in transcription modulation. Third, the relatively low diversity observed among G-rich sequences in promoter regions suggests a conserved biological role.

Finally, Choi and Majima dedicated a short chapter to the i-motif structure in their extensive review about non-B DNA structures.¹² The existence of different species (i-motif-like and classical i-motif) is commented on the light of previous works, as well as the slow folding and unfolding kinetics associated with the formation of i-motif structures. In a later work, the most recent applications of fluorescence spectroscopy to the study of i-motif folding/unfolding processes have been reviewed by these two authors.¹³ Special attention has been made to the application of these conformational processes in nanotechnology, like the design of nanomachines, sensors or logic gates, among others.

In this framework, the present work reviews the recent literature on fundamentals aspects of the i-motif structure, as well as the experimental methods used in its study. To our knowledge, recent studies dealing with the formation and stability of these structures have not been reviewed yet. Also, the latest results on the interaction with ligands and their potential presence in vivo are also reviewed. Advances in nanotechnological applications of i-motif structures, such as the formation of metal nanoparticles, the design or nanomachines or molecular motors, and the development of sensing devices have not been included in this review.

The i-motif structure

The building block of the i-motif structure is a base pair involving one neutral C and one protonated C at N3 (the C·C⁺ base pair) bonded by three hydrogen bonds (Fig. 1a). Based on NMR spectroscopy and theoretical calculations using the (C₃TA₂)₃C₃ sequence as model it has been proposed that N3⋯H⁺⋯N3 bonds may be described as hydrogen bonds with asymmetric double-well potentials rather than a symmetric hydrogen bond with a single-well potential.¹⁴ The formation of these hydrogen bonds produces a stronger base-pair interaction than the canonical G·C base pair, as denoted by in silico calculations.¹⁵ Hence, the base-pairing energy (BPE) for the proton-bound dimer of cytosine (C·C) is 169.7 kJ mol⁻¹, whereas the BPEs of the canonical Watson–Crick G·C and neutral C·C base pairs are only 96.6 and 68.0 kJ mol⁻¹, respectively.

The name, i-motif, refers to the fact that is probably the only nucleic acid structure where base pairs are intercalated.⁵ The i-motif structure may be formed from the spatial arrangement of C·C⁺ base pairs involving C tracts present in one nucleic acid strand, producing a so-called intramolecular i-motif. On the other hand, it may involve C tracts present in two or four independent nucleic acid strands, producing intermolecular i-motif structures. In any case, the C tracts are arranged spatially as a tetramer composed of two parallel-stranded duplexes that are interspersed in an anti-parallel way (Fig. 1b). The depicted intramolecular i-motif contains several C·C⁺ base pairs formed by the interaction of twelve Cs that are arranged in four tracts and into a single nucleic acid strand. In general, the hydrogen bonded to N3 in each C·C⁺ base pair is not equidistant from the two N3 atoms, but adopts a position with the largest distance to the hydrogen of the next C·C⁺ base pair.¹⁴

The interaction of consecutive base pairs is mainly provided by the stacking of the exocyclic carbonyl and amino groups that are oriented with opposed dipoles, and is not observed between the aromatic heterocycles of the bases (Fig. 1c). This apparent absence of stability provided by stacking interactions between the bases can be overcome by the formation of a systematic intermolecular C–H⋯O hydrogen bonding network between the deoxyribose sugar moieties of antiparallel backbones in the four-stranded molecule.^16,17 Overall, the base-pair distance is only 3.1 Å, a stacking much closer than in B-DNA (3.4 Å) and similar to A-DNA (2.9 Å). The helical twist between adjacent C·C⁺ pairs (12–16°) is also smaller than in the case of B-DNA (36°) or A-DNA (32.7°).¹⁸ This spatial arrangement produces the existence of two broad and flat major grooves and two extremely narrow minor grooves (Fig. 1b). According to the spatial arrangement of the C·C⁺ base pairs, i-motif structures have been classified into two classes: those with the terminal C·C⁺ base pair at the 3′ end (3′E intercalation topology), and those with the terminal base pair at the 5′ end (5′E intercalation topology) (Fig. 2).^19,20

In addition to the formation of three hydrogen bonds at each C·C⁺ base pair and the favourable interaction of consecutive pairs, the decrease of the negative charge on the backbone because of the protonation of C facilitates the association of the four strands. However, these considerations apply equally to duplex C·C⁺ structures and do not explain the formation of an intercalated structure. Part of the differential stability of the i-motif was initially explained in terms of close contacts between sugars in the narrow grooves of the structure.⁵ These could give rise to favourable van der Waals energies (as observed from molecular dynamics simulations) and, as they exclude water, may represent hydrophobic contacts between deoxyribose sugars. As shown below, this effect would help to explain the lower stability of RNA i-motifs compared to their DNA counterparts.²²

Sequences such as ACTC₃T₂CTC₂TCTCTA or TCTCTC₂TG₂TC₂TC₂ are examples of C-rich sequences that form duplex C·C⁺ structures lacking the intercalated nature of the i-motif.²³ In this case, the formation of parallel-stranded homoduplexes at pH 4–5.5 was observed, which were held by C·C⁺ base pairs. It was shown that the insertion of guanine or T into an oligomeric C sequence made the formation of the i-motif unfavourable.

Instrumental approaches to study the i-motif structures

There are different approaches to study the formation and stability of i-motif structures. The first studies demonstrated the formation of the i-motif structure by stretches of deoxycytidine at acid pH in NMR measurements. The study of tetrameric i-motif structures showed similar imino proton spectra, NOEs, and exchange properties that indicate that all the tetramers studied at that time had similar structures. It was also shown that the amino protons provide crucial information, particularly on the dynamical symmetry of the C·C⁺ pairs and on the lifetime of the open state, which is apparently at least 100 times longer than in Watson–Crick duplexes.^24–26 The imino protons involved in the C·C⁺ base pair show a characteristic NMR signal at ∼15 ppm, whereas signals between ∼12 and ∼14 ppm are related to Watson–Crick G·C and A·T base pairing. The presence of signals between ∼10 and ∼12 ppm is usually related to additional non-canonical base pairs, such as T·²⁷ or G·T mismatches.^28,29 These additional base pairs are usually located in the loop regions, but can be also found in the middle of C-tracts.³⁰ Tetrameric i-motifs formed by four short oligonucleotide strands normally render relatively simple NMR spectra and their structural characterization is, in general, straightforward. However, in the case of more complex (and presumably more relevant) sequences the assignment of the NMR spectra is much more challenging. Due to severe spectral overlap, many NMR studies could only be completed by introducing chemical modifications (usually methyl-cytosines²⁶) or site-specific isotopic labels.³¹

Apart from NMR, one of the first techniques used to unravel the i-motif structure was X-ray diffraction. Hence, the crystal structure of C₄ solved at 2.3 Å resolution revealed the formation of a four-stranded molecule composed of two intercalated duplexes.³² The information provided by X-ray crystallography may be significantly different from that obtained from solution-based techniques, such as NMR. Condensation of single molecules from solution into crystals represents a transition between distinct energetic states. In solution, the atomic interactions within the molecule dominate. In the crystalline state, however, a set of additional interactions are formed between molecules in close contact in the lattice, the packing interactions. This fact was shown by Berger and colleagues in the study of the crystal structures of C₃T, TA₂C₃, C₃A₂T, and A₂C₄.³³ These molecules showed intercalated C segments that were similar in their geometry, even though the sequences crystallized in different space groups.

UV molecular absorption spectroscopy may be a useful technique to observe the changes in protonation and stacking of the nitrogenated bases involved in the formation of the i-motif. In this region, the protonation of Cs produces hyperchromicity at wavelengths from 275 to 300 nm, as well a shift of the maximum wavelength from ∼262 (for neutral C) to ∼275 nm (for protonated C). Hence, monitoring of absorbance changes in 275–295 nm range is a nice way to study the stability of the i-motif as a function of temperature or pH changes. Due to the spectral characteristics of protonated and neutral Cs, the shape of the thermally-induced unfolding of i-motif structures monitored by molecular absorption spectroscopy is pH dependent. At pH values higher than the pK_a of C, the absorbance at 295 nm decreases upon unfolding of the i-motif, whereas the opposite behavior is observed at pH values lower than the pK_a of C. At pH values near the pK_a, no spectroscopic changes are observed upon unfolding.³⁴ In the case of melting experiments, it is possible to determine the T_m value and thermodynamic parameters such as ΔH and ΔS associated with the unfolding process. These parameters are calculated by using the van't Hoff equation, assuming a two-state process and values of ΔH and ΔS that are independent of temperature. It is possible to know the multimeric nature of the i-motif from the determination of melting temperatures. If the melting of an i-motif-forming oligonucleotide shows a concentration-independent profile, the i-motif will have intramolecular pairing, i.e., it is formed from the folding of only one, long strand. On the other hand, if the melting spectra show a concentration-dependent profile, a bi- or multi-molecular association must be considered. When the i-motif is folded intermolecularly, T_m increases with the concentration.

The simplest way to detect the formation of the i-motif structure is probably recording the circular dichroism (CD) spectrum because it shows two characteristic negative and positive bands at ∼265 and ∼285 nm, respectively. Manzini et al. used CD to demonstrate the pH-induced formation of C-rich sequences in the human telomere and in the promoter region of the k-ras gene.³⁵ The exact position of these bands is also dependent on the nature of bases located at the loops. The stability of the i-motif with temperature or pH changes is usually studied by monitoring ellipticity around 285 nm. As an example, the CD signal at this wavelength has been used to monitor the light-driven DNA conformational switch of the (C₃TA₂)₃C₃ sequence in the presence of molecular malachite green carbinol base, a light-induced hydroxide ion emitter.³⁶ A related CD technique, synchrotron radiation circular dichroism (SRCD), has been proposed for the study of i-motif structures.³⁷ The advantage of SRCD is the large available photon fluxes in the vacuum UV region where absorption is strong, which allows the exploration of excitation energy for electronic coupling over a broad wavelength region. By using this technique, SRCD spectra of simple strands of C at different pH values were measured. The protonation state of bases determined the folding motif and as a result the extent of electronic coupling between bases. Different electronic couplings were found depending on the protonation of bases: single strands of all-protonated bases display no coupling, i-motifs have electronic coupling within a hemi-protonated base-pair and likely also nearest-neighbour coupling along the strand, and single strands of all-neutral bases have nearest-neighbour couplings along the strand.

Molecular fluorescence-based techniques have been also used to study the structure and kinetics of i-motif structures. Because canonical nitrogenated bases show little intrinsic fluorescence, C-rich sequences under study are usually labelled with appropriate fluorescent (or quencher) ligands at the 5′ and/or 3′ ends. Under appropriate experimental conditions, the C-rich sequence folds, allowing a closer contact between both labels giving rise to changes in the fluorescence of the bonded ligands. Lee et al. described the attachment of a pyrene moiety to the 5′ end of a telomeric C-rich sequence. The high fluorescence of this molecule in the unfolded strand is strongly quenched upon stacking on the i-motif core.³⁸ The fluorescence almost disappears after addition of the complementary G-rich sequence and subsequent Watson–Crick duplex formation. Fluorescence resonance energy transfer (FRET) has been shown to be a sensitive approach, provided that appropriate fluorophores are chosen.^13,39,40 In this technique, the excited state energy of a fluorescent donor chromophore is transferred to an unexcited acceptor chromophore. This yields quenched donor and increased acceptor fluorescence, the efficiency being dependent on the distance and dynamics of the structure. Fluorophores such as FAM or TAMRA may be used for this purpose.⁴¹ Finally, it has been shown that DNA sequences functionalized with pyrene moieties at both the 5′ and 3′ ends show fluorescent properties that allow the study of the thermal stability of i-motif structures.^23,42 Another molecular fluorescence-based approach is the so called molecular beacons. Here, the fluorescence emitted by a label attached to one end of the DNA is quenched by a functional group attached to the other end. As an example, BODIPY and DABCYL attached at the 5′ and 3′ ends of an i-motif-forming sequence have been used as fluorophore and quencher, respectively.⁴³

Fluorescence correlation spectroscopy (FCS) may provide information about the diffusion of the molecule as a function of pH, as well as information about the intrachain contact formation. The typical FCS setup consists of a laser line that is reflected into a microscope objective by a dichroic mirror. The laser beam is focused in the sample, which contains fluorescent molecules in such high dilution, that only a few are within the focal spot (usually 1–100 molecules in one fL). When the particles cross the focal volume, they fluoresce. This light is collected by the same objective and, because it is red-shifted with respect to the excitation light, it passes the dichroic mirror reaching a detector. Conclusions on physical phenomena have to be extracted from the measured signals with appropriate previously proposed models. The parameters of interest are found after fitting the autocorrelation curve to modelled functional forms.⁴⁴

Time-resolved transient absorption and emission spectroscopies have been also used to study the formation of i-motif structures.⁴⁵ These techniques take advantage of the fact that the excited states of cytidine and deoxycytidine multimers decay orders of magnitude more slowly than excited states in the corresponding monomers. Hence, variation in nucleic acid secondary structure affects profoundly the photophysical properties. It has been described that C-rich DNA sequences at physiological conditions of pH and ionic strength have similar excited state dynamics as the acid form of C-rich RNA sequences.⁴⁵

In a recent work, the study of complex transition kinetics in G-quadruplex and i-motif structures has been reviewed.⁴⁹ The single-molecule techniques used to investigate the kinetics of non-B DNA structures are shown with a special emphasis on the current progress in the mechanical manipulation and ligand binding of non-B DNA structures including i-motifs. The three major techniques used for the mechanical unfolding of biomolecular structures (AFM, laser tweezers, and magnetic tweezers) are explained in detail. The use of laser-tweezers has been reported to investigate the structures formed in C-rich regions. Unlike traditional methods such as CD, NMR, UV/Vis, etc. this method can reveal bimolecular structures in a highly dynamic fashion. This method is sensitive in identifying small populations, such as intermediates that can be formed during a folding process. Thanks to laser-tweezers, an i-motif and a partially folded triple-helix-like structure have been shown to coexist in the C-rich human ILPR (insulin linked polymorphic region) oligonucleotide⁴⁶ (Fig. 3).

Br₂ protection experiments are carried out to identify the specific C residues of a sequence that are involved in base pairings and intercalation to form the i-motif structure. Bromine is known to react selectively with the double bond of the C within DNA, resulting in 5-bromodeoxycytidine. In particular, Br₂ reacts with the C residues in a single-stranded region 10-fold higher than those in the i-motif core. Hence, C residues in the loop regions are more reactive to Br₂ than other C residues involved in base pairing and intercalation, allowing the deduction of those specific C residues required for base pairing and intercalation in the i-motif structure.⁵⁰

Differential scanning calorimetry (DSC) is a technique that has been also used to determine mode-independent unfolding thermodynamics.^51,52 The main advantage over UV-monitored melting experiments carried out at just one wavelength is the unravelling of processes involving more than two steps, with the consequent determination of thermodynamic parameters such as the change in enthalpy. A disadvantage, however, is that the deconvolution of a DSC profile, i.e., the proposal of a given number of steps through which the folding process takes place, may be uncertain.

Separation techniques, such as non-denaturing polyacrylamide gel electrophoresis (PAGE), have been used extensively to demonstrate the mono or multimolecular nature of the i-motif formed.^34,35,53 Recently, PAGE has been used to resolve mixtures of dimeric, tetrameric and supramolecular structures formed from the tetramers.⁵⁴ Other separation techniques, such as size-exclusion chromatography (SEC) have been used to determine the number of strands involved in the formation of multimeric i-motif structures, such as T₂C₈T₂, TC₅ or (C₃TA₂)₃C₃, a fragment of the human telomere.^24,55 As the elution time depends on the hydrodynamic volume which, in turn, is a function of the number of residues and of the spatial structure, the SEC system must be calibrated using an appropriate set of standards of a known number of residues and structure before the unknown structures are analysed. Recently, SEC has been used to monitor with time the formation of dimers, tetramers and supramolecular structures from monomers.⁵⁶

Other techniques employed include Raman spectroscopy or the synchrotron small-angle X-ray scattering (SAXS). Raman spectroscopy has been used to detect the formation of i-motif structures as well as to quantify the protonation of C bases, both in solid and in solution phase.⁵⁷ Using SAXS, Jin et al. suggested that the conformation of i-motif DNA at mild acidic conditions is similar to that of the partially unfolded i-motif DNA rather than the fully folded i-motif.⁵⁸ Hence, the i-motif is structurally dynamic over a wide pH range, adopting multiple conformations ranging from the folded i-motif structure to a random coil conformation. In a later work, SAXS was also used to study the effect of C60 fullerene attached to a C-rich sequence into the duplex–intramolecular structures equilibrium.⁵⁹ Representative tridimensional models may be postulated based on the analysis of the experimental SAXS data. Transient IR absorption (TRIR) study has provided for the first time the identification of signature IR bands for the long-lived species found upon UV-excitation of the i-motif in C-rich DNA.⁶⁰ The TRIR spectrum for UV-excited i-motif possesses complex dynamics pointing to multiple decay processes, including possible charge transfer between packed hemi-protonated C bases. The slow decay is not due to simple protonation, but is rationalised in terms of the specific structural features of the i-motif. The most likely origin is charge transfer between closely packed C bases.

The use of mass spectrometry (MS) for the study of i-motif-forming structures has also been described. In the earlier work, Rosu et al. revealed the existence of i-motif structures in the ions produced by electrospray MS⁶¹ by using infrared multiphoton dissociation (IRMPD) and ion mobility (IM). The i-motif structure was detected because of its characteristic broadening of the IR band at 1650 cm⁻¹ due to the contribution of a large blue-shift of the NH₂ scissoring and NH bending modes of the Cs. For the intramolecular human telomeric C-rich strand, ion mobility revealed that the lower charge states are more compact and give IR spectra characteristic of intramolecular hydrogen bond preservation in the gas phase. It has been described as the formation of C·C⁺ base pairs in the gas phase from C derivatives.⁶² The hemiprotonated dimers were observed in chloroform solution upon treatment with strong acid and in the gas phase by ESI-MS.

Other techniques include sedimentation analysis.⁶³ It has been described that the C₈ sequence adopts an A-like DNA structure at pH 7 and room temperature, whereas longer sequences containing 24 or 28 cytidine nucleotides assembled into an i-motif below ∼15–20 °C. The fourfold reduction in axial ratio is consistent with the formation of a four-stranded i-motif configuration. The results obtained in this work suggested that pure cytidine-containing oligonucleotides can form monomeric i-motif structures without heterologous nucleotides in the loops, such as thymidine or adenosine. The use of analytical centrifugation has been reported recently.⁶⁴ As pH increases, the hydrodynamic radius of individual DNA chains in aqueous solutions prepared by being heat-treated suddenly increases while the molar mass is constant, indicating that the conformation changes from an i-motif to a random coil. Finally, competition dialysis has been shown to be a useful technique to screen the selectivity of ligand–DNA interactions based on the differences in the secondary structure, such as i-motif.⁶⁵

Multivariate analysis

Modern instruments and computers allow the simultaneous recording of multivariate data, i.e., a complete set of spectra as a function of an external variable (such as temperature or pH), and its appropriate storage. Multivariate analysis methods allow the recovery of information by unravelling these, a priori, complex data sets. In the case of melting experiments these methods not only allow the calculation of T_m or thermodynamic parameters, but also check the validity of the two-state process usually applied in the univariate analysis of these data. Multivariate analysis may be conducted in two different ways, depending on whether a physico-chemical model is initially proposed (hard-modelling approach) or not (soft-modelling approach).

For hard-modelling approaches, the proposed model depends on the nature of the process under study. Hence, for acid–base experiments the model will include a set of chemical equations describing the formation of the different acid–base species from the neutral species, together with approximate values for the stability. Kudrev et al. have studied the pH-induced formation of i-motif structures in several sequences based on a C-rich sequence found at the promoter region of the c-kit gene.⁶⁶ The spectra recorded along spectroscopically-monitored acid–base titrations were analysed by means of the hard-modelling-based matrix method. This enables the characterization of the protonation process in polymers in terms of the intrinsic protonation constant (K_in), and two additional parameters that model the cooperativity (ω_c) or anticooperativity (ω_a) of the process. Whereas the protonation and folding of mutated sequences containing only thymidines at the loops were successfully modelled, protonation and folding of the wild sequence, showing a greater base variability at the loops, could not be modelled. This fact was related to a larger conformational variability in the wild sequence that in those mutated ones. Recently, a hard-modelling method has been applied to study the formation of i-motif structures in a sequence of the n-myc gene⁶⁷ (Fig. 4). In this case, CD and molecular absorption data were analysed simultaneously. For melting experiments, the physico-chemical model is related to the thermodynamics of DNA unfolding, according to the van't Hoff equation.^68–70

Among the soft-modelling methods, singular value decomposition has been traditionally used to determine the number of transitions during a process, such as a melting process.⁷¹ Analysis of UV absorption spectra recorded along the thermally-induced unfolding of DNA structures by means of SVD allowed the detection of intermediates and a rough estimation of the concentration dependence for each conformation with temperature. An evolution of SVD is the multivariate curve resolution – alternating least squares (MCR-ALS) method, which has been used to study the acid–base equilibria of a variety of DNAs, including C-rich sequences.⁷² As SVD, MCR-ALS is a soft-modelling method that does not use any physico-chemical model in the calculation. The main advantages of MCR-ALS over previous methods are the inclusion of complementary information (such as non-negativity) as a mathematical constraint, and the possibility of carrying out a simultaneous analysis of data recorded along complementary experiments. Recently, this approach has been used to study the acid–base behaviour of oligonucleotides (CAC)₃, A₃C₆, and C₃A₃C₃.⁷³

Simulation-based methods

There have been many applications of computational methods to study the formation and stability of i-motif structures. In particular, the use of molecular dynamics (MD) to simulate the solution behaviour of i-motifs has also been described. In an early work, Gallego et al. used the AMBER force field to simulate molecular dynamics trajectories of i-motif and duplex structures of centromeric DNA.⁷⁴ The electrostatic energy components were calculated using a Poisson–Boltzmann model, and the nonpolar energy components were computed with a van der Waals function and/or a term dependent on the solvent-accessible surface area of the molecules. It was found that the i-motif was mainly stabilized by favourable electrostatic interactions between hydrogen bonded protonated and neutral Cs, and by non-polar forces including the hydrophobic effect and enhanced van der Waals contacts, Later, Malliavin et al. studied the role of the phosphodiester backbones on the stability of tetrameric i-motifs formed by C₂.⁷⁵ The 4 ns – long simulations were done for two different topologies (5′E and 3′E) and by using the program AMBER 6.0. Analysis of the trajectories showed that the main energetic difference between the two topologies is due to the van der Waals term. The comparison of the solvent-accessible area of each topology revealed that the sugar–sugar interactions accounted for the greater stability of the 3′E topology. This stresses the importance of the sugar–sugar contacts across the narrow groove which, enforcing the optimal backbone twisting, are essential to the base stacking and the i-motif stability. The hairpin structures formed by C-rich sequences at neutral pH values in explicit solvent have been also studied.⁷⁶ The GROMACS package and the modified AMBER force field were used. The results indicated that hairpin structures were the most stable configurations in the absence of protonated Cs. Recently, the structure and mechanical properties of i-motif nanowires based on a repeated TC₅ sequence were studied by means of MD simulations.⁷⁷ The i-motif nanowire seems to share similarities with structural proteins, as far as its tensile stiffness, but is closer to nucleic acids and flexible proteins, as far as its bending rigidity is concerned. Furthermore, thanks to its very thin cross section, the apparent tensile toughness is close to that of a metal.

The interaction of porphyrin ligands with the i-motif structure formed near the c-myc NHE-III₁ (nuclease hypersensitivity element III₁) has been studied in silico.⁷⁸ The NHE modelled consisted of single folded conformers of the polypurine intramolecular G-quadruplex and the polypyrimidine intramolecular i-motif structures, flanked by short duplex DNA sequences. The modelled i-motif structure used in that work was theoretical; being the central intercalated C core interactions based on NMR structural data obtained for a tetramolecular (A₂C₄)₄ i-motif. The loop structures were also in silico predictions of the c-myc i-motif loops. It was found that the cationic porphyrin compounds TMPyP2, TMPyP3, and TMPyP4 were all bound to both ends of the i-motif, one immediately adjacent to the thymidine triplet, near the interface between the G-quadruplex and i-motif, and the other at the opposite end of the i-motif. The analysis of the interaction energy in binding to both ends of the i-motif revealed that it was dominated by van der Waal's energy, with very little electrostatic component. This was explained as due to the aromatic and hydrophobic nature of the porphyrin compounds. The planar, aromatic porphyrin ring structure tends to form stacking interactions with multiple pairs of bases of the DNA, and the pyridinium rings can interact favourably with other bases or with the deoxyribose rings of the DNA.

Stability of the i-motif structure

The relative stability of nucleic acid folded structures is often measured in terms of their melting temperature (T_m). T_m is defined as the temperature of mid-transition from the folded structure to the unfolded species, often induced by heating of the nucleic acid sample. The transition is usually monitored using a spectroscopic technique, such as molecular absorption or circular dichroism. As with other nucleic acid structures, the stability of the i-motif depends on several important factors, such as the nucleotide sequence, ionic strength or temperature, among many others. Because the protonation of one of the C bases in the C·C⁺ base pairs is an absolute requirement for the formation of the i-motif structure, the pH of the medium plays a crucial role. As the pK_a value of C is around 4.6 (in pure water at 25 °C), it would be expected that the formation of i-motif structures occurs at pH values lower than 6.6, approximately.^34,52 Hence, at pH values from ∼4 to ∼7 and 25 °C, the C bases are partially protonated and the DNA folds into the closed i-motif structure. In this pH range, the stability of the i-motif is a linear function of pH.^34,79 The highest stability of the i-motif structures occurs at pH values close to the pK_a of C. At higher pH, the C bases deprotonate and the structure unfolds to a single-stranded form. On the other hand, if the pH value is too low (below 3, approximately) all the C bases are protonated and they cannot form the hydrogen bond pattern needed for the C·C⁺ base pair.⁷⁹ This is not the case for G·C and A·T Watson–Crick base pairs, whose stability is pH-independent in a broader pH range (from 2 to 10, approximately). At 37 °C and 150 mM NaCl, the i-motif is not formed at physiological pH (∼7.4) but, depending on the sequence, may be formed at ischemic pH (∼6.7).⁸⁰ Nevertheless, some i-motifs have been observed at low temperature and at neutral²⁸ or even slightly basic pH values.⁸¹ Recently, folding of i-motif structures in presence of silver cations at physiological pH has been reported.⁸²

The formation of the i-motif is strongly enthalpy driven at near-neutral pH, being a maximum value at pH 6–7, and dropping below pH 5.5.³⁴ As an example, using as a model a 31-mer sequence belonging to the c-myc gene, Mathur et al. studied the thermodynamic parameters associated with the formation of the i-motif.⁵³ At 20 °C, a ΔG of −10.4 ± 0.1 kcal mol⁻¹ was observed with favourable enthalpy (ΔH = −76.0 ± 1.0 kcal mol⁻¹) and unfavourable entropy (TΔS = −65.6 ± 0.9 kcal mol⁻¹) at pH 5.3 in 20 mM NaCl for i-motif folding. It is possible to determine the contribution of a C·C⁺ base pair to the thermodynamic parameters. Using a set of intramolecular i-motifs differing in the number and length of C tracts, Mergny et al. determined that each base pair decreases the ΔH by 10.5 ± 0.5 kcal mol⁻¹ and ΔS by 57 ± 1 cal mol⁻¹ K⁻¹ at pH 6 and 100 mM NaCl.³⁴ Using sequences that form both intra- and intermolecular i-motif structures, values for ΔH and ΔS around 7 kcal mol⁻¹ and 20 cal mol⁻¹ K⁻¹, respectively, at pH 5.8 and 100 M Na⁺ were determined,⁵² with these values being larger than those corresponding to the disruption of base pairs in the Watson–Crick duplex (∼4.5 kcal mol⁻¹ and ∼11 cal mol⁻¹ K⁻¹, respectively). From the NMR titration of TC₃, a free energy of −7.6 kJ mol⁻¹ per cytidine base pair for the formation of the tetramer from single strands was calculated.²⁴ Overall, the values determined for the ΔH and ΔS have been shown to be independent of the nature (monomeric or tetrameric) of the i-motif structure. Finally, in the case of the c-myc i-motif, the formation of each C·C⁺ base pair leads to a decrease in enthalpy of 9–12 kcal mol⁻¹.⁵³

Apart from pH, the fundamental factor affecting the structure and stability of the i-motif structure is the number of C·C⁺ base pairs present in the folded structure. For a given pH value, the stability of two different i-motifs differing in the number of C bases in their sequences would be higher for the one showing a higher number of C·C⁺ base pairs.⁸³ It has been shown that the intra versus intermolecular folding primarily may also depend on the length of the C-tracts.⁸⁴ Two stretches of six or fewer C residues favour the intermolecular folding of i-motifs, whereas longer C-tracts promote the formation of intramolecular i-motif structures with high thermal stability.

Bases other than C may be also accommodated into the i-motif core formed by C·C⁺ base pairs. Thymidines can form symmetrical T·T base pairs that are nearly isomorphic of C·C⁺ base pairs. The NMR-based study of the 9-mer sequence 5mC₂TCTCTC₂ showed the existence of two open/closed T3/T7 motifs, but the central thymidines form two long-lived T5·T5 pairs that are intercalated into the i-motif core.⁸⁵ Interestingly, whereas that sequence forms a tetrameric i-motif, the similar sequence 5mCCTCACTCC forms a dimer.²⁷ In this case, the dimeric i-motif is built by intercalation of two symmetrical hairpins held together by six symmetrical C·C⁺ pairs and an additional T·T pair. In summary, the formation of T·T base pairs in general stabilizes the i-motif structure because of extension of the i-motif core and decrease in flexibility.⁸⁶ Recently, the influence of methylation and halogenation of cytosine on the base-pairing energies (BPEs) in the i-motif has been studied.¹⁵ To address this, proton-bound heterodimers of cytosine and 5-methylcytosine, 5-fluorocytosine, 5-bromocytosine, and 5-iodocytosine were studied. All modifications were found to lower the BPE and therefore would tend to destabilize DNA i-motif conformations. However, the BPEs in these proton-bound heterodimers still significantly exceed those of the Watson–Crick G·C and neutral C·C base pairs, suggesting that C·C⁺ mismatches are still energetically favoured such that i-motif conformations are preserved.

Apart from the number of C bases, another key factor is the length and nature of loops. First, the length of the loops has been shown to be critical in determining the multimeric nature of the i-motif structure. Hence, C-rich sequences with just one base at the loops tend to form mixtures of mono and bimolecular i-motifs,⁸⁴ whereas longer loops favour the formation of intramolecular i-motifs.³⁴ It has been proposed that i-motif structures may be classified into two groups depending on the length of the loops.⁸⁷ In class I, the loop sizes are 2:3/4:2 with four, five or six C·C⁺ base pairs. In class II, the loop sizes are 6/8:2/5:6/7. In general, the midpoint of the pH-induced folding from the neutral strand for class II i-motif structures is slightly higher (∼6.6) than that for i-motif structures identified as class 1 (∼5.8–6.4). due to the presence of stabilizing effects into longer loops. In this sense, it has been described as an i-motif showing a 12-base long loop that is able to form a hairpin stabilized by Watson–Crick base pairs.⁶⁷ Interestingly, the potential to form a stem-loop structure in the long loop directs a particular and well stabilized DNA structure from diverse choices of high-ordered arrangements when bases at this loop were mutated to T.

The nature of the bases at the loops also influences strongly the structure and solution behavior of i-motif structures. Using NMR methods, it has been shown that protonation of A bases at the loops when the pH is lower than 4.6 produces a conformational change involving disruption of the i-motif core.⁸⁸ The studied sequence, a mutant of a fragment of the human centromeric satellite III, C₂AT₂C₂AT₂C₂T₃C₂, folds into a intramolecular i-motif structure. Protonation of adenine bases hinders the formation of an A·T base pair that extends the C·C⁺ core at pH higher than 4.6. The application of a multivariate approach to the study of the acid–base equilibria of C₃T₃ and C₃TA₂ sequences allowed the quantitation of the relative contribution of all acid–base species as function of pH.⁶⁹ It was observed that the adenine-containing sequence formed two i-motif structures, probably differing in the protonation state of adenines at the loops. Moreover, it has been observed that adenine-containing i-motif structures unfold at lower temperatures than non-mutated sequences.^34,69 T· base pairs can occur in the loops connecting the C-tracts.⁸⁹ Recently, using the telomeric i-motif sequence as model, it has been proposed that sterically demanding A bases at the double-loop side lead preferentially to the 5′E conformation.⁸⁶ In this sense, the systematic modification of the 4A loop region of the retinoblastoma gene with PyA fluorophore units allows discrimination of the fluorescence signals corresponding to structural dynamics from single-stranded to i-motif structures.⁹⁰ In terms of fluorescence signals, i-motif structures modified with fluorophores at the 1,2 and 1,4 positions of the 4A loop provide the most dramatic fluorescence changes at a single excitation wavelength upon conformational transitions from single-stranded to duplex to i-motif structures, respectively. Loop interactions are of particular relevance in the dimeric structure of d(TC₃GT₃C₂A)²⁹ and d(TCGT₃CGT₂).²⁸ In both cases, the 5′-GT₃-3′ loops form intermolecular GT base pairs that interact to each other through their minor groove side, forming a G:T:G:T tetrad.

The bases at the 5′ and 3′ ends of C-rich sequences able to form i-motif structures have also an influence on the stability and molecularity of the i-motif structures. The study of the C₃TA₂C₃ and C₃TA₂C₃TA₂ sequences by means of PAGE and thermal melting experiments allowed the proposal that the inclusion of the additional –TAA segment at the 3′-end produces a change in the molecularity of the structure. Hence, the 9-mer folds in a bimolecular structure, whereas the 12-mer sequence exists in two (bimolecular and tetramolecular) forms.⁹¹ The different arrangement of flanking bases may also influence the stability of the i-motif structure. Using NMR methods, it was shown that the A₂C₃ and C₃A₂ folded into different i-motif structures.⁹² Whereas the first sequence adopts a unique structure showing A·A pairs stacked to C·C⁺ base pairs, the second sequence, which adopts two distinct intercalation topologies (3′E and 5′E), clearly shows well stacked adenosine bases to the adjacent C·C⁺ base pairs. Using X-ray crystallography, it was shown that the i-motif formed by the AC3T sequence was further stabilized at one end by a three-base hydrogen-bonding network, in which two As and a T form four hydrogen bonds via a reverse Hoogsteen and an asymmetric A–A base pairing.⁹³

As many other nucleic acid structures, the ionic strength of the solution has a certain influence on the stability of the i-motif structure. As described above, nitrogenous bases are neutral throughout a wide pH range, from 5 to 9, approximately. The same occurs with the pentose sugar, which can only lose the proton at very alkaline pH values. However, at neutral pH values, DNA is a polyanion because of the negatively charged phosphate groups. The presence of appropriate counterions, such as sodium, potassium, and others balances the negative charge. Folded conformations may produce changes in the counterion atmosphere surrounding the phosphate backbone. Moreover, in the case of the i-motif, protonation at N3 produces the appearance of a positive charge at the base pairs. Therefore, variations in ionic strength can cause conformational changes and variations in the relative stability of i-motif structures. In general, it has been observed than an increase in ionic strength from 0 to 100 mM NaCl produces a destabilization of the structure.³⁴ Higher NaCl concentrations did not produce changes in T_m values. Using a C-rich sequence found near the promoter region of the c-myc gene, it was found that folding was induced by uptake of about two to three protons per mole of i-motif while a marginal (0.5–1 mol mol⁻¹) counterion uptake was observed.⁵³ Using CD spectroscopy, S1 nuclease digestion, and multivariate analysis it was shown later that two different i-motif conformations may be formed differing in the number of C·C⁺ base pairs.⁹⁴ The transition between both conformations is induced by a change in pH or ionic strength.

Apart from the canonical i-motif structure formed at acid pH values (∼4–6), the formation of partially folded structures at neutral and slightly acid pH values (∼7–6) has been proposed. Using FRET and FCS techniques in the bulk phase and at the single-molecule level, it has been proposed that the partially folded species coexist with the single-stranded structure at neutral pH and room temperature.⁹⁵

Interaction with ligands

The stability of folded structures of DNA may depend strongly on the interaction with inorganic and organic ligands. Hence, research is being done to find ligands which could modulate in vivo the stability of characteristic DNA structures, such as those G-quadruplexes proposed near the promoter region of several oncogenes.⁹⁶ In contrast with the G-quadruplex, few molecules have been reported to bind to i-motif structures (Fig. 5). The reasons may be the low stability of the i-motif at physiological conditions and the fact that the i-motif is a very compact structure, where planar ligands which usually stack on base pairs cannot be introduced easily. Therefore, non-specific electrostatic interactions throughout the phosphate groups or interactions with bases at the loops seem to be the most plausible interaction mechanisms.

The interaction of single-walled carbon nanotubes (SWNTs) with i-motif-forming sequences was firstly studied by molecular fluorescence and S1 nuclease digestion. The cleavage patterns showed that SWNTs bind to i-motif DNA at the end of the major groove.⁹⁷ It was proposed that SWNTs can stabilize selectively human telomeric C-rich DNA and induce i-motif DNA formation under physiological conditions or even at pH 8.0. The strong affinity of i-motif DNA for SWNTs has been also used to distinguish single- and multiwalled carbon nanotubes.⁹⁸ Later, the interaction of SWNTs with human telomeric i-motif DNA was shown to accelerate the S1 nuclease cleavage rate at the loop regions.⁹⁹ In a recent work, the use of SWNTs to inhibit telomerase activity through stabilization of the i-motif structure has been reported.¹⁰⁰ The persistence of the i-motif and the concomitant G-quadruplex eventually leads to telomere uncapping and displaces telomere-binding proteins from the telomere.

TMPyP4 is a porphyrin that has been used extensively as a model ligand to study the binding characteristics of G-quadruplex structures. Hence, the study of its interaction with the i-motif is the natural continuation of such studies. Moreover, the interaction of TMPyP4 with a C-rich sequence corresponding to the promoter region of the c-myc gene at pH 7.4 has been shown to have an inhibitory effect on NM23-H2 DNA-binding activity.¹⁰¹ From a biophysical point of view, Fedoroff et al. studied the interaction of TMPyP4 with a tetrameric i-motif formed by the A₂C₄ sequence.¹⁰² From the results of NMR and docking studies, the ligand was suggested to bind peripherally near the ends by a non-intercalative mechanism that is independent of the multimeric nature of the i-motif structure. Using a intramolecular i-motif based on the telomeric sequence, it was shown that the binding of the porphyrin at pH 5 does not alter the structure, the interaction being mostly electrostatic in nature.⁶⁹ It was proposed that each i-motif binds up to two TMPyP4 molecules in an independent way, the value of the binding constant for each binding site being ∼10⁶ M⁻¹.

The interaction of the ruthenium complex Ru(bpy)₂(dppz)]²⁺ (where dppz stands for dipyrido[3,2-a:20,30-c]phenazine) with C-rich sequences has been studied at pH 7 and pH 5, where the i-motif is already formed.¹⁰³ Only a weak interaction was observed at pH 5, probably of electrostatic nature. Other complexes involving central metal atoms, such as the terbium-amino acid complex [Tb₂(DL-HVal)₄-(H₂O)₈]Cl₆ have been studied. In this case, the interaction at pH 5.5 with the telomeric i-motif structure produces a slight destabilization of the structure.¹⁰⁴

Graphene quantum dot (GQDs) have been described as stabilizing ligands for the human telomeric sequence (C₃TA₂)₃C₃T.¹⁰⁵ GQDs are graphene sheets that possess novel chemical and physical properties. They can stabilize and induce the i-motif formation, as denoted by an increase in T_m values upon addition of the ligand. It seems that GQDs not only stabilize the i-motif, but they can also induce i-motif formation in alkaline or neutral pH. The mechanism of interaction has been explained in terms of an interaction with the internal TAA loop.

In a recent work, the interaction of phenanthroline derivatives with the i-motif structure formed by the C₃(A₂TC₃)₃ sequence has been studied.¹⁰⁶ These derivatives stabilize the structure of i-motif with an increase in melting temperature of ∼8 °C in the presence of 10 times excess of these compounds. Their binding stoichiometric ratio and constant were 1:1 and 2 × 10⁵ M⁻¹, respectively. Xu et al. studied the interaction of the hairpin structures formed by the sequences CGC₄TA₂C₃TA₂CTA₂C₃TGCG and A₂(C₃TA₂)₂C₃T₂C₃T₄ at pH 7 with the ligands doxorubicin and Hoechst 33258. These sequences may be used as a pH-driven C-rich DNA drug release device because, at neutral pH, they may bind the ligands whereas no interaction was observed at pH 5.0, at which the i-motif structure is formed.¹⁰⁷ Finally, the interaction of a known dye, crystal violet, with the tetrameric i-motif structure 5′-(AC₃T)₄-3′ has been studied recently.¹⁰⁸ The ligand was shown to bind externally to the terminus of the i-motif structure with a 1:1 stoichiometry and a 7 × 10⁵ M⁻¹ binding constant.

Apart from “canonical” aromatic and planar ligands, the interaction of i-motifs with other substances has been studied. Hence, Ren et al. reported the use of PMNT, a polythiophene derivative, to visualize by the naked eye the formation of an i-motif structure upon a pH change.¹⁰⁹ In response to the conformational switch of the i-motif, the conformation of PMNT that forms the complex changes correspondingly, resulting in a different colour. In a very recent work, Hurley and colleagues described the specific interaction of the IMC-48 ligand with an i-motif-forming sequence found at the bcl-2 gene. At pH 6.6 the i-motif is in equilibrium with a hairpin showing Watson–Crick base pairs. By choosing the appropriate ligand, this equilibrium may be shifted efficiently, with dramatic effects on the gene expression.¹¹⁰

The interaction of the linear carotenoid ligands crocin and crocetin, as well as the monoterpene aldehydes picrocrocin and safranal, with the i-motif structure formed by GC(TC₂)₃TC₂T(TC₃)₃ has been also studied.¹¹¹ Finally, the affinity of metal cations such as silver or palladium ions for i-motifs are being exploited for the preparation of DNA-mediated metal nanoclusters that may be used as fluorescent probes of gene sensors.¹¹² A review of studies dealing with the development and applications of these nanoclusters is beyond the scope of this work.

Effects of chemical modifications on the stability

The stability of the i-motif structure may be modulated by introducing appropriate modifications in one or more of its building blocks. Hence, modification of nucleosides has been proposed as a way to increase the stability of i-motif structures. As example, the effect of the incorporation of 5-methyl-cytosine, deoxyuracil or 5-(1-propynyl)-deoxyuracil into C-rich DNA sequences was studied in detail in one of the first works dealing with this subject.¹¹³ It was observed that oligonucleotides including 5-propynyl-deoxyuracil formed a stable i-motif which precluded triplex formation. Later, the influence of methylation or other modifications such as halogenation of cytosine on the base-pairing energies (BPEs) in the i-motif was studied.¹⁵ Proton-bound heterodimers of cytosine and 5-methylcytosine, 5-fluorocytosine, 5-bromocytosine, and 5-iodocytosine tend to destabilize DNA i-motif conformations. However, the BPEs in these proton-bound heterodimers still significantly exceed those of the Watson–Crick G·C and neutral C·C base pairs, suggesting that C·C⁺ mismatches are still energetically favoured such that i-motif conformations are preserved.

Inclusion of a single substitution of 2′-deoxy-2′-fluororibocytidine into the TC₅ sequence enhances the stability of the i-motif in sodium citrate buffer (50 mM, pH 4.2).¹¹⁴ Although multiple substitutions do not increase the stability further, they are relatively well tolerated. This is in contrast with the effect of inclusion of ribonucleotides in the same sequence, which is strongly destabilizing. On the other hand, inclusion of arabinonucleotides facilitates an increase in stability.¹¹⁵ The different effect provoked by the relative orientation of the –OH group in riboses and arabinoses supports the importance of sugar–sugar contacts in stabilizing i-motifs. In the case of 2′-deoxy-2′-fluoroarabinocytidine, alteration of sugar–sugar contacts gives rise to multiple effects, including the formation of alternative i-motifs.

The i-motif is also sensitive to alterations of the phosphate component of the backbone (Fig. 6). Some modifications, such as phosphoramidate oligonucleotides have been shown to prevent the folding of C-rich sequences into i-motif structures at acidic pH values but not into triple helices.¹¹³ Similarly, sequences showing methylphosphonate backbones do not form i-motif structures.^113,116 On the contrary, phosphorothioate oligodeoxynucleotides have been shown to form i-motif structures.^113,116 It was observed that the stability of phosphorothioate oligodeoxynucleotides (measured in terms of T_m values) was slightly lower than for the phosphodiester backbone. Moreover, at neutral pH, the phosphorothioate associates and dissociates nine times faster than a phosphodiester oligodeoxynucleotide of identical sequence. Kanehara et al. also studied the stabilizing role of the R_p and S_p configurations of the phosphorothioate backbone in a tetrameric i-motif.¹¹⁷ As for duplex DNA, it was observed that the S_p configuration produced a more stable structure than R_p. Later, it was suggested that the S_p configuration at phosphorus of the phosphorothioate linkage changes the sugar–phosphorothioate conformation and intermolecular interaction in the narrow groove, leading to the destabilization of the i-motif structure.¹¹⁸ The comparison of the NOESY spectra showed that intra-residual H6–H3′ and H2′′–H4′ NOE cross-peaks of the all S_p isomer are weaker than those of the all R_p isomer and PO-d(TC₄), indicating the change in the C3′-endo conformation and glycosidic bond angle. Another modification of the phosphate backbone involves the inclusion of 3′-S-phosphorothiolate linkages.^119,120 It has been shown that this incorporation stabilizes the i-motif with a minimum perturbation of the overall structure. This stabilization probably arises from the preference of the phosphorothiolate residues for the C3′-endo sugar pucker, which is universally observed in solution for the cytidine sugars in the i-motif. These results also strongly suggest that the destabilization of the i-motif by ribose sugars is due to the 2′-substituent and not the C3′-endo sugar pucker.

The formation of i-motif structures in DNA-modified backbones, such as locked nucleic acids (LNA) (Fig. 7) has also been studied. This is a conformationally constrained nucleic acid mimic with a 2′-O–4′-C methylene bridge that locks the nucleotide analogue in a C3′-endo sugar conformation. Generally, the introduction of LNA nucleotide monomers increases the thermal stability of DNA and RNA duplexes and triplexes. However, the in vivo potential use of LNA derivatives is still under study because they have been shown to produce hepatotoxicity in animals.¹²¹ The i-motif formation with LNA-modified TC₅ oligonucleotides at low pH values was initially identified by means of CD, UV, and NMR spectroscopy.¹²² Later, the effect of LNA nucleotides on the formation in modulating the formation and stability of the i-motif structure formed by a c-myc-based sequence was also studied.¹²³ Choosing the right mutations, it is possible to increase or decrease the stability, measured in terms of T_m values, of a wild sequence.

On the other hand, Pasternak and Wengel studied the influence of unlocked nucleic acid (UNA) monomers on thermodynamic stability of the 22-nucleotide human telomeric i-motif.¹²⁴ It was shown that UNA monomers modulate i-motif stability in a position-dependent manner, whereas no structural changes were observed in any case. In addition, 5-(pyren-1-yl)uracilyl UNA monomers were found to destabilize an i-motif structure at pH 5.2, both under molecular crowding and non-crowding conditions.¹²⁵

The formation of i-motif structures from peptide nucleic acids (PNA) has also been described. One of the first works used alanyl-PNA, which consists of a regular peptide backbone with an alternating configuration of the alanyl amino acids. This backbone has nucleobases connected in the β position of the side chain.¹²⁶ It was found that the octamer-(AlaC-AlaC)₄-Lys-NH₂ did not form any i-motif structure at pH 4.5. On the contrary, the H-(Gly-AlaC)₄-Lys-NH₂ sequence formed the i-motif. It was concluded that the formation of i-motif structures in PNA sequences required the inclusion of a glycine unit as every second amino acid thus providing a spacer that facilitates the intercalation of linear double strands. Using the DNA and PNA analogues d(TC₅) and p(TC₅), it has been shown that the PNA i-motif is stable over a pH range (∼4.1–4.5) narrower than in the case of the DNA i-motif (∼4.7–5.7). The difference was explained in terms of a response to pH to the polyanionic nature of the DNA backbone which probably shields pH changes more effectively than the neutral PNA backbone.¹²⁷ This narrower pH range of existence was also confirmed for the i-motif formed by p(TC₈) its stability being, however, slightly higher than that of the corresponding DNA i-motif.¹²⁸ The use of a peptide template has been described as a method for the design of i-motif structures that may be stable even at neutral pH values.¹²⁹ In this approach, a cyclic peptide scaffold is used as a topological template for directing the intramolecular assembly of anchored oligonucleotides into an i-motif topology.

A mixed strategy involves the formation of hybrid i-motif structures from a binary mixture of C-rich DNA (TC₅) and PNA (H₂N-TC₅-Lys-COOH) sequences.¹³⁰ ESI-MS confirmed the formation of a tetrameric species, composed of PNA–DNA heteroduplexes whereas NMR spectroscopy confirmed that PNA and DNA form a unique complex comprising five C·C⁺ base pairs per heteroduplex. H1′–H1′ NOEs show that both heteroduplexes are fully intercalated and that both DNA strands are disposed towards a narrow groove, invoking sugar–sugar interactions, as seen in DNA i-motifs. As in the case of the PNA i-motif, this hybrid i-motif showed enhanced thermal stability and intermediate pH dependence. This is likely to be because two of the negatively charged backbones in the DNA i-motif are replaced by neutral polyamide PNA backbones, thereby reducing the electrostatic repulsion associated with multi-stranded structures. Furthermore, the DNA i-motif is net negatively charged, the PNA i-motif is net positively charged and the hybrid is an electrically neutral complex.

Circular oligonucleotides possess many distinctive properties, when compared to their linear counterparts, such as higher DNA-binding affinity, greater sequence selectivity, enhanced resistance to degradation by exonuclease, and an ability to serve as efficient templates for DNA and RNA polymerase.¹³¹ In this sense, it has been shown that the i-motif can direct the sequence-specific formation of a phosphodiester linkage and thus represent a new type of structural template for constructing circular oligonucleotides.^131,132 A particular example are the nanometer-sized circles ranging in length from 36 to 60 nucleotides based on the C-rich human telomere repeat, (C₃TA₂)_n.¹³³ These cyclic DNAs may act as templates for synthesis of human telomere repeats in vitro. The circles were constructed successfully by the application of the A-protection strategy, which allows for cyclization/ligation with T4 DNA ligase. Thermal denaturation studies showed that at pH 5.0, all circles form folded structures with similar stability, while at pH 7.0 no melting transitions were seen. Recently, there has been reported a minimal i-motif structure formed by the cyclic sequence 〈pTCGT₂CGT₂〉. This sequence forms a dimeric structure stabilized by a unique C·C⁺ base pair capped at both ends by G:T:G:T tetrads.²⁸ Finally, Zhou et al. described the synthesis of a fluorescein-labelled circular i-motif structure. The most striking feature of their work is that the dye is not bonded to the usual 5′ or 3′ ends, but to the central loop. This is accomplished by covalent bonding to one of the bases at the loops not involved in the i-motif core.¹³⁴ It was shown that this cyclic oligonucleotide resists the hydrolysis by exonucleases and the fluorescent moiety is indeed present in the ligation product.

Yang et al. studied the effect of tetra(ethylene glycol) (EG₄) substitution at the loop region of intramolecular i-motif formed by (C₃T₂)₃C₃. In general, single substitution at all loops preserved the stability of the i-motif, while the triple substitution of three bases at one loop resulted in a significant decrease of stability, and the more EG₄ substitutions, the less stable the system. The substitution at the narrow groove results in a faster migration in agarose gel compared to those at other two wide grooves, suggesting that modification at different positions has a different extent of influence on the topology.¹³⁵ The influence of the insertion of a non-nucleotide pyrene moiety into the loop between two C-rich regions has also been studied.¹³⁶ The stability of the i-motif structures was measured at different pH values under non-crowding and crowding conditions (20% poly(ethylene glycol)). When (R)-3-((4-(1-pyrenylethynyl)benzyl)oxy)propane-1,2-diol (TINA) was inserted, the oligonucleotides still formed i-motif structures of similar stability (at pH 6.2 but not at pH 5.2) to those observed for the corresponding wild type oligonucleotide. Interestingly, incorporation of pyrrolic-modified porphyrines into C-rich sequences induces i-motif stabilization.¹³⁷ This is probably due to porphyrine–porphyrine interaction between the parallel oriented strands in the tetrameric i-motif.

Other modifications include the addition of C60 fullerene to both 5′ and 3′ ends of a telomeric C-rich sequence.⁵⁹ Upon addition of the G-rich complementary strand, it was found that fullerene shifted the pH-induced conformational transition between the i-motif and the duplex structure, possibly due to the hydrophobic interactions between the terminal fullerenes and between the terminal fullerenes and an internal TAA loop in the DNA strand. Finally, Robidoux et al. described the synthesis of branched parallel cytidine-rich oligonucleotides that are joined at the ends with a riboadenosine linker. It was shown that this molecular architecture enhances the stability of the resulting i-motif.¹³⁸ In addition, these branched constructions were used to explore the effect of arabino- vs. ribo-substitutions in the stability of the i-motif, showing that arabino-cytidines do not provoke the dramatic destabilizing effect of rC. The formation of branched i-motif structures incorporating 2-deoxy-5-propynylcytidine residues was also confirmed by temperature-dependent CD- and UV-spectra as well as by ion-exchange chromatography. The low pK_a-value of this nucleoside (pK_a ∼ 3.3) compared to cytidine (pK_a ∼ 4.5) required strong acidic conditions for i-motif formation. The immobilization of oligonucleotides incorporating multiple residues of that nucleoside on 15 nm gold nanoparticles generated DNA–gold nanoparticle conjugates which are able to aggregate into i-motif structures at pH 5.¹³⁹

Kinetics of i-motif formation

Appropriate sequence and environmental conditions (such as pH, temperature or ionic strength, among others) may favor the formation of thermodynamically stable i-motif structures. However, kinetic aspects such as the time needed to accomplish the folding of a given DNA sequence or the folding pathway must be taken into account. As the most-studied i-motif-forming sequence, the C-rich strand of telomere DNA was recently reported to exist transiently as a 5′ single-stranded overhang at the chromosome ends in the S-phase of replicating human cells.¹⁴⁰ The rates of folding and unfolding are crucial in determining the opportunity or possibility of the formation of the i-motif under in vivo situations within the narrow “time window”, during which the DNA is liberated transiently.¹⁴¹

As expected, measurements done at different pH values show that both the folding- and unfolding-rate constants (k_f and k_u) are strongly dependent on pH. Zhao et al. studied the variation of the rate constants with pH for a intramolecular i-motif structure by SPR.¹⁴¹ They observed that the folding and unfolding of (C₃TA₂)₄ occur on the time scale of minutes at pH ∼5. These values, which were larger than those obtained in solution,¹³² were explained as due to the immobilization of the DNA on the SPR chip. The results showed that promotion of i-motif formation by protons is achieved by a combination of increased k_f and decreased k_u, which result in a rapid increase in the folding equilibrium constant, K_F, with decreasing pH. The impact of pH on the kinetic aspects of i-motif folding is also reflected in the observation of irreversible melting/annealing curves. FRET studies revealed high reversibility of the pH-induced folding of i-motifs and multiphasic folding kinetics with folding and unfolding time constants on the order of minutes.¹⁴²

Intramolecular i-motifs form faster than those that are multimeric in nature. Hence, the formation times for intramolecular and tetrameric structures are on the order of seconds and hours, respectively. Liu and Balasubramanian observed that the folding and unfolding processes of the (C₃TA₂)₃C₃ i-motif are both completed in about 5s in a proton-fueled DNA nanomachine using the fluorescence spectroscopy.¹⁴³ However, the formation of dimeric (mC₂TCACTC₂)₂ and tetrameric (TC3)₄ i-motifs takes place in hours.^27,85

Most studies have been done on the formation pathway of i-motif structures. As example, the formation and dissociation rates of three [TC_n]₄ tetramers (n = 3, 4 and 5), their dissociation constants and the reaction orders for tetramer formation by NMR have been reported.¹⁴⁴ The experimental observations suggested that i-motif formation proceeds via sequential strand association into duplex and triplex intermediate species and that triplex formation is rate limiting. The proposed model is different to the one involving association of two preformed duplexes. In the case of the [TC₃]₄ i-motif structure, the reaction order was 3. Studying the folding kinetics of a dimeric i-motif, Canalia et al. found that the association rate of this structure is smaller than in the case of the Watson–Crick duplex. As oligonucleotide association in Watson–Crick duplexes is described by a model involving the formation of a correct nucleus of a few base pairs, followed by rapid ‘zipping’ into the fully paired duplex, the formation of an i-motif structure could involve additional steps.²⁷ These could be: (i) the formation of a hemiprotonated nucleus; (ii) adequate intercalation during the nucleus elongation of a third strand with appropriate orientation; and (iii) association, in parallel orientation with this stand, of a fourth strand that finally locks the strand assembly into a long-lived structure.

Lieblein et al., by means of NMR spectroscopy, investigated the kinetics of folding of the i-motif initiated by a pH-jump from pH 9 to pH 6. Under these conditions, folding follows a kinetic partitioning mechanism, where two conformations form in the first step with a rate constant on the order of 2 min⁻¹. Subsequent refolding of the kinetically favoured conformation to the thermodynamically more stable conformation is slow, with rate constants of the order of 10⁻³ min⁻¹. They proposed that the two conformations differ in the intercalation topology of the C·C⁺ base pairs and in the formation of T·T base pairs. At equilibrium, the closing C·C⁺ base pair can either be formed at the end of the C-rich strand (5′E) in the major conformation or at the 3′-end (3′E) in the minor conformation.²¹

Using the stopped-flow circular dichroism (SFCD) technique, the kinetics of the pH-induced folding and unfolding process of the i-motif formed by (C₃TA₂)₃C₃ have been also studied.¹⁴⁵ The results showed that the molecule can fold or unfold on a time scale of ∼100 ms when the solution pH is changed from 8 to 5, or vice versa. On the assumption of an irreversible folding or unfolding processes, theoretical models to decipher the respective kinetics were proposed, suggesting that the cooperativity of protons is crucial for both the folding and unfolding process. In the unfolding process, the cooperative neutralization of two protons (out of the total six protons in the i-motif core) is the only rate-limiting step. In the folding process, on the contrary, there exists a critical step in which three protons bind cooperatively to the DNA strand.

The influence of pH on the dynamics of intramolecular folding of the (C₃A₂T)₃C₃ sequence has been studied recently by means of FRET and FCS techniques.⁹⁵ The measured fluorescence decay profiles were explained as the result of a mixture of two different species (the unfolded strand and a partially folded strand) that evolve simultaneously to yield the i-motif structure at pH 4.8. Interestingly, it is proposed that the partially folded species as well as the single-stranded structure coexist at neutral pH, supporting that the partially folded species may exist substantially in vivo. Finally, Zhou et al. claimed the formation of i-motif structure by the telomeric sequence at neutral and even slightly basic pH values, at 4 °C.⁸¹ The kinetic data provided by CD and FRET were fitted to a single exponential describing first-order kinetics. Surprisingly, the determined time constants depended strongly on the spectroscopic data used for the calculation (214 s for CD data, and 493 s for fluorescence data). The difference was explained in terms of a delaying effect of the dyes attached covalently into the sequence in FRET measurements.

i-Motif versus intermolecular Watson–Crick duplex competition

In vivo, C-rich regions are usually accompanied by the complementary G-rich regions. It is expected, therefore, that there will be a thermodynamic equilibrium involving the Watson–Crick duplex structure and the potential tetraplex (G-quadruplex and i-motif) structures. The hypothetical role in vivo of both tetraplex structures must be closely related to the shift of the equilibrium from the Watson–Crick duplex. This competition has been extensively studied, mainly in the case of the human telomeric DNA, but also in the case of guanine- and C-rich sequences found in the promoter regions of several genes.

Phan and Mergny published one of the first works studying the competition between the tetraplex (G-quadruplex and i-motif) and the Watson–Crick duplex structures in the human telomere.¹⁴⁶ Using NMR and UV melting experiments, phase diagrams for the 1:1 mixtures of the AG₃(T₂AG₃)₃ and (C₃TA₂)₃C₃T sequences were proposed in 100 mM NaCl or KCl. It was observed that the tetraplex structures are predominant at pH lower than 5, approximately, in the presence of KCl. In a coetaneous work, Sugimoto et al. showed that for these two sequences, at pH 7, 100 mM NaCl and 0 °C, the duplex is the predominant species, the concentrations of the intramolecular structures being residual.¹⁴⁷ The binding constant of the two DNA strands in the presence of 10 mM Mg²⁺ at pH 7.0 was shown to be 5.3 × 10⁷ M⁻¹ at 20 °C, about 400 times larger than that in the presence of 100 mM Na⁺ at pH 5.5. Using multivariate data analysis methods, the relative concentrations of all species involved in the equilibrium between intra- and intermolecular species in the case of C and G-rich regions at the bcl-2,¹⁴⁸ c-kit¹⁴⁹ and n-myc⁶⁷ genes have been determined. In all these three cases, the concentration contribution of the tetraplex structures at pH 7 and 25 °C is below 10%.

The thermodynamic parameters behind the competition between the Watson–Crick duplex and the intramolecular structures have been also calculated. Lee et al. reported the measured (using ITC) and calculated (using Hess cycles) enthalpies for the reactions involving G-quadruplex structures and C-rich sequences.¹⁵⁰ The overall results showed that the favourable free energy terms for the interaction of DNA intramolecular complexes with its complementary strands are enthalpy driven. The main observation is that an intramolecular structure can be disrupted by a complementary strand, provided that its length and sequence is appropriate. However, the favourable free energy term of these targeting reactions may well be increased by improving the stability of the duplex products, by using longer single strands with complementary sequences and/or DNA intramolecular structures with loops containing a larger number of unpaired bases.

The kinetics of the duplex formation from the equimolar mixture of G-/C-rich complementary sequences has been also studied. Li et al. investigated this transition at both pH 7.0 and pH 5.5.¹⁵¹ Fitting to a single-exponential function gave an observed formation rate of 8 × 10⁻³ s⁻¹ at 20 °C in 10 mM Mg²⁺ buffer at pH 7.0, which was about 10 times the observed rate at pH 5.5 under the same conditions. Both of the observed rates increased as temperature rose, implying that the dissociation of the intramolecular structures was the rate-limiting step for the Watson–Crick duplex formation. On the basis of SPR measurements involving immobilized DNA, it has been proposed that pH affects to the association process involving the single strands, whereas it has little effect on the dissociation step of the Watson–Crick duplex.¹⁴¹ Hence, whereas the value for the association rate is ∼10⁶ M⁻¹ s⁻¹ in the pH range 4.8–7.0, the corresponding value for the dissociation rate is reduced from 10⁻³ s⁻¹ (at pH 4.8) to 10⁻⁴ s⁻¹ (at pH 7.0). Overall, these values correspond to association equilibrium constants around ∼10⁸–10⁹ M⁻¹, values higher than those obtained in solution.

Recently, it has been demonstrated that the formation of G-quadruplex and i-motif conformations destabilize directly the proximal duplex regions.¹⁵² It has been shown that the large diameter of these folded conformations is mitigated with increased distance from the duplex region: a spacing of five base pairs or more is sufficient to maintain duplex stability proximal to predicted G-quadruplex/i-motif-forming regions. This difference in stability reflects a stronger contribution from entropy than from enthalpy, although both kinetic and thermodynamic influences are present.

The effect of pH and cation on the structures and stabilities of the isolated sequences G₄T₄G₄ and C₄A₄C₄ and of their mixtures were studied at 5 °C and 100 mM NaCl.¹⁵³ In addition to the expected effect of pH, it was found that Ca²⁺ ions induced a parallel G-quadruplex structure and then inhibited the duplex formation at pH 6. Interestingly, however, the addition of Mg²⁺ to the equimolar mixture of G₃(T₂AG₃)₃ and 5′-(C₃TA₂)₃C₃ at pH 7 stabilizes the duplex and destabilizes the G-quadruplex.¹⁴⁷

To understand the structure of biomolecules in vivo, their properties studied in vitro are extrapolated to the in vivo condition, while the condition in a living cell is inherently molecularly crowded and a non-ideal solution contains various biomolecules. The effect of molecular crowding on the structure and stability of the telomeric G-rich and C-rich sequences has been studied. Cell-mimic crowding can increase i-motif stability at acid pH and cause dehydration.¹⁵⁴ These crowding conditions are achieved at pH 5.5 using concentrations of polyethylene glycol (PEG) 200 (average molecular weight of 200) ranging from 0 to 50% (w/w). However, crowding cannot induce i-motif formation at physiological pH. On the other hand, it was shown that, in crowding conditions, the 1:1 (G-rich:C-rich) mixture folds into the parallel G-quadruplex and i-motif but does not form the Watson–Crick duplex, as observed in absence of crowding conditions.¹⁵⁵ The ITC measurements indicated that the thermodynamic stability (ΔG^°₂₀) of the duplex formation between the G-rich and C-rich DNAs in the non-crowding condition was −10.2 kcal mol⁻¹, while only a small energy transfer was observed in the ITC measurements in the molecular crowding condition. Overall, these results suggested that the conditions of molecular crowding may prevent Watson–Crick duplex formation.

The competition between intramolecular tetraplex structures and intermolecular Watson–Crick has been used to construct a nanomachine that, depending on the pH of the medium, is able to bind or release the telomere-binding protein TRF 1, and to release small quadruplex-binding molecules to impede progress of the polymerase.⁴¹ In a similar approach, a 36-mer circular oligonucleotide containing both a lateral 25-mer i-motif structure and a 11-mer A-rich region was synthesized.⁴³ The A-rich region was able to form Watson–Crick interactions with the corresponding T-rich complementary strand. At neutral pH, the i-motif structure is not formed and the duplex predominates. At low pH values, the formation of the i-motif structure generates a backbone bending on the non C-rich segment that prevents the formation of the Watson–Crick duplex.

Supramolecular i-motif structures

C-Rich sequences showing self-associative properties may form supramolecular (or higher order) structures (sms) that may be potentially interesting from a nanotechnological point of view. Among the pioneer works in this field, Ghodke et al. presented a strategy to build 1D scaffolds by using i-motif structures.¹⁵⁶ The formation of these structures, known as I-wires, was induced from a highly concentrated solution of the monomeric C₇ or C₉ sequence by annealing at pH 5.5 from 90 °C to 25 °C followed by incubation at 4 °C. The growth propagates through non-Watson–Crick base-pairing and leads to nanowires more than 3 μm long. Recently, the structure and mechanical properties of DNA i-motif nanowires were studied by means of molecular dynamics simulations.⁷⁷

The formation of sms has been investigated in sequences containing two C stretches of unequal length (C_nXC_m).⁵⁴ These sequences may associate into a tetrameric i-motif with a core of C·C⁺ base pairs and two dangling non-intercalated strands of the shorter C stretch at each end (Fig. 8). These dangling ends allow the formation of the sms structures by interacting with other tetrameric i-motif structures. Interestingly, the formation of these sms structures competes with the formation of dimeric i-motif structures, and is dependent on pH, DNA concentration and temperature. Later, it was shown that C stretches as short as CC can link i-motif tetramers into sms.⁵⁶ By using NMR and SEC, it was observed that the sms grew in competition with i-motif tetramers or dimers and their formation rate was controlled by the availability of the building block involved in sms formation. The sms stability increased with the number of cytidines in the shorter C stretch. The comparison of the sms lifetimes with those of i-motif tetramers showed that incorporation of tetramers into a larger structure has a stabilizing effect. The presence of a single non-C residue in the oligocytidine sequences prevents the formation of structures with staggered strands, an effect certainly favorable to sms elongation into linear structures. Due to sugar backbone stretching induced by systematic intercalation, the i-motif is a stiff structure. Hence, association of i-motif tetramers into sms can potentially form unbendable rods with a particularly large persistence length.

Apart from i-motif sms, a mixed strategy involves the formation of DNA pillars from i-motif stems and Watson–Crick duplex branches.¹⁵⁷ The central stem has some overhanging structures that can enable hybridization with complementary units by Watson–Crick pairing and, thus, multiple i-motifs can join to form the pillar.

Complex DNA molecules containing parallel and antiparallel duplex elements as well as i-motif structures have been designed and synthesized recently.¹⁵⁸ Oligonucleotide duplexes with parallel orientation containing reverse Watson–Crick A·T base pairs and short C₂ tails were shown to be stabilized under slightly acidic conditions by C·C⁺ base pairs. Corresponding molecules with antiparallel orientation containing Watson–Crick A·T base pairs did not show this phenomenon. This chimeric DNA with parallel duplex elements and long C₅ tails at one or at both ends assemble to tetrameric i-motif structures, even at pH 6.8. Moreover, molecules with two terminal C₅ tails form multimeric assemblies which have the potential to form nanoscopic scaffolds. On the contrary, antiparallel hybrid molecules were only able to aggregate into multimeric assemblies at pH 6.0.

i-Motif RNA?

In general, oligoribonucleotides form less stable i-motif structures than the corresponding oligodeoxynucleotides. The first work that studied the potential formation of i-motif structures by oligoribonucleotides described a difference in T_m of 29 °C for the 18-mer DNA vs. the corresponding 18-mer RNA, indicating a lower stability of RNA i-motif structure. In fact, oligoribonucleotides were shown to adopt preferably a triplex structure, instead of the i-motif.¹⁵⁹ As uracil-substituted oligodeoxynucleotides did not destabilize the i-motif, it was deduced that the methyl group of T does not play a role in the stabilization of the motif and, consequently, the difference between RNA and DNA must arise from other sources, such as the presence/absence of the 2′-OH group, the sugar conformation or other steric hindrances. Collin et al., studying the relative thermal stabilities of DNA and mixed DNA/RNA tetrads,²² proposed an explanation for the inability of RNA to form a stable i-motif. As the glycosidic angles χ in the DNA i-motif are generally high anti-conformation and sugar puckers are mostly C3′-endo, which are also typical for RNA A-type duplexes, the steric hindrance between 2′-hydroxyls in the narrow groove should be most responsible for the absence of an RNA i-motif. This was shown by the positional dependence of i-motif stability in the studied dihydroxylated tetrads, by the absence of effect of a 2′-hydroxyl substitution in tetrads containing arabinose and by the complete intolerance of 2′-O-methyl modifications.

The spatial structure of an RNA i-motif was shown to be very similar to the corresponding DNA i-motif, it being difficult to decide whether the differences between them are significant.¹⁶⁰ As example, the r(UC₅) sequence forms two i-motif structures that differ by their intercalation topologies. The stacking topology of the main structure avoids one of the six 2′-OH/2′-OH repulsive contacts expected in a fully intercalated structure. The C3′-endo pucker of the RNA sugars and the orientation of the intercalated C·C⁺ pairs result in a modest widening of the narrow grooves at the steps where the hydroxyl groups are in close contact. Finally, as observed previously, the free energy of the RNA i-motif, on average −4 kJ mol⁻¹ per C·C⁺ pair, is half of the value found in DNA i-motif structures.

Hybrid i-motifs may be a way to increase the poor stability of RNA i-motifs. It has been shown that a hybrid consisting of two DNA strands and two RNA strands is formed faster than the corresponding DNA i-motif.¹⁶¹ However, the thermodynamically more stable structure corresponds to the DNA i-motif, as disproportion is observed after 5 days. A mixed strategy involves the formation of a hybrid i-motif from a binary mixture of C-rich RNA and PNA sequences.¹⁶²

i-Motifs in vivo?

The Watson–Crick duplex is thermodynamically more stable than the intramolecular structures at the physiological conditions of pH and temperature. However, it is known that nuclear processes such as transcription, replication, recombination and repair produce negative supercoiling. A way to reduce the stress produced by this negative supercoiling is the local unwinding of the double helix.¹¹ In these conditions, the formation of intramolecular structures, such as the G-quadruplex or i-motif, could be favoured over the Watson–Crick duplex. On the other hand, it seems that the requirement of an acidic pH value to maintain a stable i-motif structure is a barrier that cannot be overcame. Dysregulated pH is known to be an adaptive feature of most cancers, regardless of their tissue origin or genetic background. In normal differentiated adult cells, intracellular pH is generally lower (around 7.2) than the extracellular pH (around 7.4). However, cancer cells have a higher intracellular (around 7.4) and a lower extracellular pH (6.7–7.1).¹⁶³ In these conditions, C-rich sequences may adopt i-motif structures and modulate the formation of the other nucleic acid structures. It is also worth mentioning that some biological processes can provoke local acidification in the cell. For example, poly(ADP-ribose) polymerases (PARPs) produce 1 mol of proton and nicotinamide for each mol of NAD consumed. This reaction could cause temporary acidification allowing the transient formation of i-motif structures.

The crowding nature of the in vivo environment must be taken into account. In general, the effect induced by the crowded intracellular environment is defined by the excluded volume effect and dehydration effect. The role of these effects on the formation and stability of i-motif structures has been studied extensively. The results, however, are not conclusive; whereas some studies suggest the formation of the i-motif at neutral pH under crowding conditions,¹⁶⁴ other studies did not observe such formation.¹³ Rajendran et al. reported evidence that the molecular crowding induces the formation of the intramolecular i-motif structure formed by the 5′-CG₂(C₂T)_nCG₂-3′ sequence, where n = 4, 6, 8 and 10, at pH 7.0 and 122 mM Na⁺. 20% (w/w) PEG 2000 and 8000 were used as cosolutes because they may mimic the intracellular environment.¹⁶⁴ Recently, the same cosolute was used to study c-myc promoter sequence i-motifs. It has been proposed that the c-myc i-motif can exist as a stable structure at pH values as high as 6.7 in 40% w/w polyethylene glycols having molecular weights up to 12000 g mol⁻¹.¹⁶⁵

Despite the uncertain existence of i-motif structures in vivo, it has been found that the existence of C-rich sequences that could potentially form i-motif structures may be a source of errors in the polymerase chain reaction (PCR) amplification of DNA fragments.¹⁶⁶ Undetected, this phenomenon may produce systematic errors in genetic analyses that may lead to misdiagnoses in clinical settings and, in consequence, the authors propose that PCR products should be checked for G-quadruplexes and i-motifs to avoid the formation of allele dropout-causing secondary structures.

Telomeric DNA

Two of the most studied G-rich and C-rich sequences are those corresponding to the end of the human telomeres. This region of the genome, which plays an important role in cell replication, has been related to cancer diseases, as well as aging. The telomere is a superstructure that protects the telomere from degradation. It is known that the length of the telomere is shortened after each cell replication cycle. After a series of cycles, the telomere is so short that it cannot protect the telomere from degradation and the cell dies. However, the telomeric specific polymerase, telomerase, prevents the shortening of the telomere. Hence, a relatively high activity of the telomerase enzyme is detected in 80–90% of tumour cells.¹⁶⁷

In this context, the C-rich strand of human telomere DNA, which could form an i-motif structure, has been reported to exist transiently as a 5′ single-stranded overhang at the chromosome ends in the S-phase of replicating human cells.¹⁴⁰ This strand, with a repetitive sequence of 5′-(C₃TA₂)_n-3′, has been studied in depth since the early works on the i-motif structure^25,35,55 (Table 1). Depending on the experimental conditions and on the specific sequence studied, different structures may be observed. Hence, by using NMR methods, the formation of additional T·T or A·A base pairs that help to stabilize the structure has been observed. Slow conversion of 3′E and 5′E conformations has also been observed for a tetrameric structure based on a Tetrahymena sequence.⁹² In the case of the human telomere, it seems that intramolecular folding of long sequences favours the formation of the 5′E conformation.¹⁶⁸ It has been proposed that long telomere sequences may show different behavior to truncated sequences.^51,91 Hence, while the 9-nt sequence adopts a bimolecular i-motif structure, the double repeat (12-nt) sequence exists in two (bimolecular and tetramolecular) forms.

---

The hypothetical telomeric i-motif might be an important and interesting matter of study in biomedical research. Firstly, it has been proposed as an anti-cancer target in gene regulation processes due to its potential ability to inhibit telomerase and to stop the multiplication of cancer cells. Hence, it has been reported recently that the i-motif may interfere with RNA transcription.¹⁰⁰ Secondly, the i-motif structures have been proposed to play an important role in chromosome recognition.¹⁶⁹ In eukaryotes cells the sister chromatids remain together until they separate in mitosis to produce two daughter cells. Thanks to these structures, the two antiparallel ends of telomeres might associate.

Centromeric satellite DNA

The sequence TC₂CGT₃C₂A is part of a region of the human centromeric alpha-satellite DNA called the CENP-B box, the binding site of centromere protein B (CENP-B). At acidic pH, this sequence and the full TC₃GT₃C₂A₂CGA₂G CENP-B box strand all fold and dimerize in solution forming a stable bimolecular structure containing two GT₃ hairpin loops that interact through a G:T:G:T tetrad.²⁹ The stem region of the dimer is a four-stranded intercalated motif in which two hairpins associate in a head-to-head manner. These minor groove tetrads have been observed recently in other i-motif structures and may contribute to stabilize the C·C⁺ base-pairs.²⁸ Later, the folding of this sequence was studied by means of molecular dynamics simulations (see above). Another sequence, belonging to the human centromeric satellite III showed a pH-dependent intercalation topology, as discussed above.⁸⁸

C-rich regions in gene sequences

Apart from telomeric regions, potential i-motif-forming sequences have been found in more than 40% of gene controlling regions. Table 2 summarizes the studied sequences. It has been speculated that the formation of i-motif structures near the promoter regions of these genes could have a biological function, such as chromosomal translocation.⁹¹ As example, the expression of the c-myc gene produces a protein that activates telomerase, altering the activity of transcription factors and forcing cell growth forward.¹⁷⁰ The formation of different tumours may be caused by overexpression of the c-myc gene. Therefore, one of the first studied i-motif structures near promoter regions was that of c-myc.¹⁷¹ It was shown that a 33-base-long sequence adopts several different i-motif structures, its stability being remarkable at pH 7.¹⁷⁰ Since then, many additional studies have been performed on this or similar sequences. Dettler et al. studied a shorter sequence and found that it can adopt multiple “i-motif-like”, i-motif and single-stranded structures as a function of pH. The i-motif is predominant in the pH range near the pK_a of C, whereas the “i-motif-like” is the most significant species at higher pH values.¹⁷² Recent results on the sequence 5′-TC₄AC₂T₂C₄AC₃TC₄AC₃T-3′¹⁷³ indicate that the intercalative C·C⁺ base pairs are not always necessary for an intramolecular i-motif, a similar conclusion proposed for cyclic nucleotides.²⁸ It seems that the dynamic character of the i-motif formed by this sequence is intrinsic to it and appears to provide additional stability to the i-motif. This dynamic character may be the responsible for protein binding. The C-rich sequences in the c-myc gene have been also used to study the competition between intramolecular Watson–Crick vs. intermolecular (G-quadruplex and i-motif) structures. As commented above, the presence of the G-rich complementary sequence induces the formation of the Watson–Crick duplex at neutral pH values. However, it has been shown that mutation of C to T bases may shift the equilibrium.¹⁷⁴ On the other hand, it has been proposed that negative superhelicity may also shift the equilibrium, inducing the formation of the intramolecular G-quadruplex and i-motif structures, even at neutral pH values.¹⁷⁵

---

Another well-studied C-region is that found near the promoter region of the bcl-2 gene. The expression of this gene produces a protein that inhibits cell death and causes their survival. It has been demonstrated, through mutational studies coupled with bromine footprinting, that the sequence CAGC₄GCTC₃GC₅T₂C₂T C₃GCGC₃GC₄T may form an i-motif showing a 8:5:7 loop folding pattern.¹⁷⁶ The presence of these long, lateral loops, also observed in the n-myc gene (see below), raises a question about their biological significance or advantage in relation to the stability of the i-motif. A possible reason for the larger loop sizes in the bcl-2 promoter is that these are needed to provide additional stability, perhaps through the formation of capping structures. Another reason could lie in a role as recognition scaffolds for specific interaction with nuclear proteins or other molecules. A smaller sequence was also studied by means of spectroscopic techniques and multivariate data analysis methods.⁷⁹ The existence of a intermediate species between the i-motif (at pH near the pK_a of C) and the single strand (at neutral pH values) was postulated .⁷⁹ This is in accordance with the later proposal of the “i-motif-like” species for the c-myc gene mentioned above.

Guo et al. studied the formation of i-motif structures in 24-base long sequence belonging to the human vascular endothelial growth factor (VEGF).^177,178 The structure was proposed to have six C·C⁺ intercalated base pairs with a dynamic equilibrium between a 2:3:3 and a 2:3:2 loop configuration. At pH 8, bromine footprinting assays revealed that the studied sequence seems to form partially unfolded i-motif structures.

Interaction of C-rich sequences with proteins

The interaction of C-rich DNA strands with proteins has been a matter of study for many years because the geometry and charge distribution of i-motifs make it an attractive model for specific structural recognition of DNA by proteins.¹⁷⁹ However, it is not clear that the interaction of a protein with a given C-rich sequence is produced through the previous formation of the i-motif structure. Many of the published studies have been carried out at neutral or slightly basic pH values and the formation of the i-motif has not been demonstrated.

Poly-C-binding proteins (PCBP) are ubiquitous oligonucleotide-binding proteins in eukaryotic cells that play a fundamental role in the regulation of gene expression via interaction with C-rich oligonucleotides. The family consists of the archetypal hnRNP K (heterogeneous nuclear ribonucleoprotein K) and isoforms of αCP1 (also known as PCBP), including αCP1-4 and αCP-KL.¹⁸⁰ In addition to their more recognized ability to bind RNA, αCPs have also been shown to bind single stranded DNA. The closely related hnRNP K is established as a transcription factor, binding to the CT element in the promoter region of c-myc.¹⁸¹ The αCPs, on the other hand, have also been found to recognize the C-rich strand of human telomeric DNA with high affinity.¹⁷⁹ Of all the αCPs, αCP1 in particular, showed remarkable specificity for the telomeric (CCCTAA)_n repeat motif.¹⁸²

In one of the pioneering works, a 39 kDa polypeptide from Trypanosoma brucei was shown to have selective affinity for the C-rich strands of the telomere repeats.¹⁸³ Unfortunately, the protein was not identified at that time. In an almost parallel work, Marsich et al. showed that HeLa nuclear extracts contain a protein that binds with high specificity to the single-stranded (C₃TA₂)_n repeat. Electrophoretic mobility shift assays showed that the oligonucleotide (C₃TA₂)₃C₃T forms a stable complex with this protein when incubated at pH 8.¹⁸⁴ On the other hand, the interaction of a well-characterized protein, DDP1, homologous to the multi-KH domain proteins, with a 12-nt long C-rich sequence was also observed.¹⁸⁵ Lacroix et al. studied the interaction of two proteins (hnRNP K and ASF/SF2) with the telomeric C-rich sequence throughout a relatively wide pH range (6–9.2). Their results did not prove that the oligonucleotide, once bound to the protein, was still in the i-motif configuration. However, they pointed out to the fact that the protein binds to the unfolded sequence and is able to open the i-motif at acidic pH.¹⁷⁹ In a later work, Bandiera et al. studied the interaction between the C-rich telomeric sequence and a series of heterogeneous nuclear ribonucleoprotein subgroups (hnRNP). Once again, their research was done at experimental conditions where the i-motif structure is quite unstable, like pH values near 8. The work demonstrated that the interaction of the proteins takes place with single strands.¹⁸² Similar studies have showed the interaction of hnRNP proteins with a C-rich sequence that belongs to the human VEGF promoter.¹⁷⁸ In this case, the chosen protein was hnRNP K, which had been previously shown to bind C-rich sequences.^179,186 The results suggested that hnRNP K binds to the VEGF C-rich sequence in a different conformation than that of the i-motif structure. Very recently, the interaction of an i-motif structure near the promoter region of the bcl-2 gene with the hnRNP LL transcriptional factor has been studied.¹⁸⁷ As the protein unfolds the i-motif structure to yield a stable single-stranded DNA–protein complex, it is concluded that the i-motif may act as a molecular switch that controls gene expression. The interaction of appropriate i-motif-specific ligands may be a way to shift the equilibrium involving the formation of the DNA–protein complex, thus modulating gene expression.

The crystal structures of the KH1 domain of the PCBP-2 protein in complex with the sequences A₂C₃TA and (A₂C₃T)₂ were obtained.¹⁸⁸ The KH-domain has a specific fold in which a stable three-stranded antiparallel β-sheet is packed against three α-helices on one face. In both cases, the protein:DNA ratio is 2:1. Also, the structure does not show the formation of C·C⁺ base pairs, nor the folding of the C-rich strand into an i-motif structure. Using SPR, it has been demonstrated that the KH1 domain makes the most stable interactions with both RNA and DNA.¹⁸⁰ SPR experiments, with a series of poly-C-sequences, revealed that C is preferred at all four positions in the oligonucleotide binding cleft and that a C-tetrad binds KH1 with 10 times higher affinity than a C-triplet.

There has been also reported the association of proteins of Saccharomyces cerevisiae with a model sequence of the C-rich telomeric strand, (C₃ACA)₃C₃.¹⁸⁹ A gel retardation assay of the yeast protein extract, in conditions where the DNA fragment folds into an intramolecular i-motif, showed the formation of one major retarded band. Differentially bound proteins were identified as Imd2p, Imd3p and Imd4p. These similar proteins are analogs of the two human NAD-dependent inosine 5′-monophosphate dehydrogenases (IMPDH) which occur as tetramers. Most of the assays gave identical results at pH 8 than at pH 6, at which the C-rich sequence folds into an i-motif. However, the results did not allow confirmation that the protein–DNA interaction involves the formation of the i-motif structure. Finally, the interaction of C-rich DNA with proteins that recognize single stranded DNA may be also the basis of an analytical method for the detection of i-motif formation by SPR.¹⁹⁰ In this case, a single-stranded DNA binding protein from E. coli was used.

DNAzymes

Finally, another biological field where C-rich regions have been studied is that of DNAzymes,^191,192 which are catalytic nucleic acids. By choosing adequate cytosine-rich sequences active DNAzymes can be produced at either acidic or basic pH values. This pH-induced interchangeable activation and deactivation of a cation (preferably Mg²⁺)-dependent DNAzyme is due to the self-assembly of the i-motif structures.

Conclusions

The interest of DNA researchers in the i-motif structure seems to have decreased in past years due to the general thought that an acidic pH was necessary to stabilize the structure in vivo. However, interest in this structure has increased again. The reasons for this change could be the discovery of proteins that may bind C-rich sequences, despite the fact that it is not still clear whether the interaction takes place through the i-motif or through the single unstacked strand. On the other hand, the low pH of endosomes and the acid pH of the tumor microenvironment as a result of the active metabolism of cancer cells increases the interest in pH responsive systems for selective delivery and the i-motif is an interesting pH-sensitive DNA scaffold. In this sense, i-motif sequences have been used to cap mesoporous silica nanoparticles generating a versatile delivery devices that can open and close the pores by changing the pH.¹⁹³ Moreover, the i-motif structures have great potential applications in nanotechnology. The open/closed nature of the structure, together with the fact that the pH-range where this opening takes place may be modulated by choosing the right sequence, allows for new applications such as molecular switches, biosensors and nanomachines.¹³

Acknowledgements

We acknowledge funding from the Spanish government (CTQ2012-38616-C02-02 and CTQ-2010-20541-C03). Sanae Benabou thanks the Spanish Ministerio de Economía y Competitividad for a PhD grant.

Notes and references

1. A. Bacolla, M. Wojciechowska, B. Kosmider, J. E. Larson and R. D. Wells, DNA Repair, 2006, 5, 1161–1170 CrossRef CAS PubMed.
2. G. Biffi, D. Tannahill, J. McCafferty and S. Balasubramanian, Nat. Chem., 2013, 5, 182–186 CrossRef CAS PubMed.
3. G. Biffi, M. Di Antonio, D. Tannahill and S. Balasubramanian, Nat. Chem., 2013, 6, 75–80 CrossRef PubMed.
4. E. Y. N. Lam, D. Beraldi, D. Tannahill and S. Balasubramanian, Nat. Commun., 2013, 4, 1796 CrossRef PubMed.
5. K. Gehring, J. L. Leroy and M. Gueron, Nature, 1993, 363, 561–565 CrossRef CAS PubMed.
6. C. Kang, I. Berger, C. Lockshin, R. Ratliff, R. Moyzis and A. Rich, Proc. Natl. Acad. Sci. U. S. A., 1994, 91, 11636–11640 CrossRef CAS.
7. P. Alberti, A. Bourdoncle, B. Saccà, L. Lacroix and J. L. Mergny, Org. Biomol. Chem., 2006, 4, 3383–3391 CAS.
8. D. Liu and S. Balasubramanian, Angew. Chem., Int. Ed., 2003, 42, 5734–5736 CrossRef CAS PubMed.
9. D. E. Gilbert and J. Feigon, Curr. Opin. Struct. Biol., 1999, 9, 305–314 CrossRef CAS.
10. M. Gueron and J. L. Leroy, Curr. Opin. Struct. Biol., 2000, 10, 326–331 CrossRef CAS.
11. S. Kendrick and L. H. Hurley, Pure Appl. Chem., 2010, 82, 1609–1621 CrossRef CAS PubMed.
12. J. Choi and T. Majima, Chem. Soc. Rev., 2011, 40, 5893–5909 RSC.
13. J. Choi and T. Majima, Photochem. Photobiol., 2013, 89, 513–522 CrossRef CAS PubMed.
14. A. L. Lieblein, M. Kramer, A. Dreuw, B. Furtig and H. Schwalbe, Angew. Chem., Int. Ed., 2012, 51, 4067–4070 CrossRef CAS PubMed.
15. B. Yang and M. T. Rodgers, J. Am. Chem. Soc., 2014, 136, 282–290 CrossRef CAS PubMed.
16. I. Berger, M. Egli and A. Rich, Proc. Natl. Acad. Sci. U. S. A., 1996, 93, 12116–12121 CrossRef CAS.
17. J. L. Leroy, K. Snoussi and M. Gueron, Magn. Reson. Chem., 2001, 39, S171–S176 CrossRef CAS.
18. V. A. Bloomfield, D. M. Crothers and I. J. Tinoco, Nucleic acids: Structures, Properties, and Functions, University Science Books, Sausalito, CA, 2000 Search PubMed.
19. A. T. Phan and J. L. Leroy, J. Biomol. Struct. Dyn., 2000, 17, 245–251 Search PubMed.
20. L. Cai, L. Chen, S. Raghavan, A. Rich, R. Ratliff and R. Moyzis, Nucleic Acids Res., 1998, 26, 4696–4705 CrossRef CAS PubMed.
21. A. L. Lieblein, J. Buck, K. Schlepckow, B. Furtig and H. Schwalbe, Angew. Chem., Int. Ed., 2012, 51, 250–253 CrossRef CAS PubMed.
22. D. Collin and K. Gehring, J. Am. Chem. Soc., 1998, 120, 4069–4072 CrossRef CAS.
23. F. Geinguenaud, J. Liquier, M. G. Brevnov, O. V. Petrausken, Y. I. Alexeev, E. S. Gromova and E. Taillandier, Biochemistry, 2000, 39, 12650–12658 CrossRef CAS PubMed.
24. J. L. Leroy, K. Gehring, A. Kettani and M. Gueron, Biochemistry, 1993, 32, 6019–6031 CrossRef CAS.
25. S. Ahmed, A. Kintanar and E. Henderson, Nat. Struct. Biol., 1994, 1, 83–88 CrossRef CAS PubMed.
26. J. L. Leroy and M. Gueron, Structure, 1995, 3, 101–120 CrossRef CAS.
27. M. Canalia and J. L. Leroy, Nucleic Acids Res., 2005, 33, 5471–5481 CrossRef CAS PubMed.
28. N. Escaja, J. Viladoms, M. Garavás, A. Villasante, E. Pedroso and C. González, Nucleic Acids Res., 2012, 40, 11737–11747 CrossRef CAS PubMed.
29. J. Gallego, S. H. Chou and B. R. Reid, J. Mol. Biol., 1997, 273, 840–856 CrossRef CAS PubMed.
30. M. Canalia and J. L. Leroy, J. Am. Chem. Soc., 2009, 131, 12870–12871 CrossRef CAS PubMed.
31. J. Dai, A. Ambrus, L. H. Hurley and D. Yang, J. Am. Chem. Soc., 2009, 131, 6102–6104 CrossRef CAS PubMed.
32. L. Chen, L. Cai, X. Zhang and A. Rich, Biochemistry, 1994, 33, 13540–13546 CrossRef CAS.
33. I. Berger, L. Cai, L. Chen and A. Rich, Biopolymers, 1997, 44, 257–267 CrossRef CAS.
34. J. L. Mergny, L. Lacroix, X. Han, J. L. Leroy and C. Helene, J. Am. Chem. Soc., 1995, 117, 8887–8898 CrossRef CAS.
35. G. Manzini, N. Yathindra and L. E. Xodo, Nucleic Acids Res., 1994, 22, 4634–4640 CrossRef CAS PubMed.
36. H. Liu, Y. Xu, F. Li, Y. Yang, W. Wang, Y. Song and D. Liu, Angew. Chem., Int. Ed., 2007, 46, 2515–2517 CrossRef CAS PubMed.
37. A. I. S. Holm, L. M. Nielsen, B. Kohler, S. V. Hoffmann and S. B. Nielsen, Phys. Chem. Chem. Phys., 2010, 12, 3426–3430 RSC.
38. I. J. Lee, J. W. Yi and B. H. Kim, Chem. Commun., 2009, 5383–5385 RSC.
39. J. L. Mergny, Biochemistry, 1999, 38, 1573–1581 CrossRef CAS PubMed.
40. T. Simonsson and R. Sjoback, J. Biol. Chem., 1999, 274, 17379–17383 CrossRef CAS PubMed.
41. Y. Xu, Y. Hirao, Y. Nishimura and H. Sugiyama, Bioorg. Med. Chem., 2007, 15, 1275–1279 CrossRef CAS PubMed.
42. A. Dembska, P. Rzepecka and B. Juskowiak, J. Fluoresc., 2013, 23, 807–812 CrossRef CAS PubMed.
43. Y. Wang, X. Li, X. Liu and T. Li, Chem. Commun., 2007, 4369–4371 RSC.
44. J. Choi, S. Kim, T. Tachikawa, M. Fujitsuka and T. Majima, J. Am. Chem. Soc., 2011, 133, 16146–16153 CrossRef CAS PubMed.
45. B. Cohen, M. H. Larson and B. Kohler, Chem. Phys., 2008, 350, 165–174 CrossRef CAS PubMed.
46. S. Dhakal, J. L. Lafontaine, Z. Yu, D. Koirala and H. Mao, PLoS One, 2012, 7, e39271 CAS.
47. S. Dhakal, J. D. Schonhoft, D. Koirala, Z. Yu, S. Basu and H. Mao, J. Am. Chem. Soc., 2010, 132, 8991–8997 CrossRef CAS PubMed.
48. S. Dhakal, Z. Yu, R. Konik, Y. Cui, D. Koirala and H. Mao, Biophys. J., 2012, 102, 2575–2584 CrossRef CAS PubMed.
49. Z. Yu and H. Mao, Chem. Rec., 2013, 13, 102–116 CrossRef CAS PubMed.
50. K. Guo, A. Pourpak, K. Beetz-Rogers, V. Gokhale, D. Sun and L. H. Hurley, J. Am. Chem. Soc., 2007, 129, 10220–10228 CrossRef CAS PubMed.
51. M. Kaushik, N. Suehl and L. A. Marky, Biophys. Chem., 2007, 126, 154–164 CrossRef CAS PubMed.
52. J. Völker, H. H. Klump and K. J. Breslauer, Biopolymers, 2007, 86, 136–147 CrossRef PubMed.
53. V. Mathur, A. Verma, S. Maiti and S. Chowdhury, Biochem. Biophys. Res. Commun., 2004, 320, 1220–1227 CrossRef CAS PubMed.
54. A. Laisné, D. Pompon and J. L. Leroy, Nucleic Acids Res., 2010, 38, 3817–3826 CrossRef PubMed.
55. J. L. Leroy, M. Gueron, J. L. Mergny and C. Helene, Nucleic Acids Res., 1994, 22, 1600–1606 CrossRef CAS PubMed.
56. E. Guittet, D. Renciuk and J.-L. Leroy, Nucleic Acids Res., 2012, 40, 5162–5170 CrossRef CAS PubMed.
57. J. M. Benevides, C. Kang and G. J. Thomas Jr, Biochemistry, 1996, 35, 5747–5755 CrossRef CAS PubMed.
58. K. S. Jin, S. R. Shin, B. Ahn, Y. Rho, S. J. Kim and M. Ree, J. Phys. Chem. B, 2009, 113, 1852–1856 CrossRef CAS PubMed.
59. K. S. Jin, S. R. Shin, B. Ahn, S. Jin, Y. Rho, H. Kim, S. J. Kim and M. Ree, J. Phys. Chem. B, 2010, 114, 4783–4788 CrossRef CAS PubMed.
60. P. M. Keane, M. Wojdyla, G. W. Doorley, J. M. Kelly, A. W. Parker, I. P. Clark, G. M. Greetham, M. Towrie, L. M. Magno and S. J. Quinn, Chem. Commun., 2014, 50, 2990–2992 RSC.
61. F. Rosu, V. Gabelica, L. Joly, G. Grégoire and E. De Pauw, Phys. Chem. Chem. Phys., 2010, 12, 13448–13454 RSC.
62. A. R. Moehlig, K. E. Djernes, V. M. Krishnan and R. J. Hooley, Org. Lett., 2012, 14, 2560–2563 CrossRef CAS PubMed.
63. D. M. Hatters, L. Wilson, B. W. Atcliffe, T. D. Mulhern, N. Guzzo-Pernell and G. J. Howlett, Biophys. J., 2001, 81, 371–381 CrossRef CAS.
64. S. Wu, X. Wang, X. Ye and G. Zhang, J. Phys. Chem. B, 2013, 117, 11541–11547 CrossRef CAS PubMed.
65. J. B. Chaires, Top. Curr. Chem., 2005, 253, 33–53 CAS.
66. P. Bucek, R. Gargallo and A. Kudrev, Anal. Chim. Acta, 2010, 683, 69–77 CrossRef CAS PubMed.
67. S. Benabou, R. Ferreira, A. Aviñó, C. González, S. Lyonnais, M. Solà, R. Eritja, J. Jaumot and R. Gargallo, Biochim. Biophys. Acta, Gen. Subj., 2014, 1840, 41–52 CrossRef CAS PubMed.
68. N. J. Greenfield, Nat. Protoc., 2007, 1, 2527–2535 CrossRef PubMed.
69. S. Fernandez, R. Eritja, A. Aviño, J. Jaumot and R. Gargallo, Int. J. Biol. Macromol., 2011, 49, 729–736 CrossRef CAS PubMed.
70. R. D. Gray and J. B. Chaires, Curr. Protoc. Nucleic Acid Chem., 2011, 17-4 Search PubMed.
71. I. Haq, B. Z. Chowdhry and J. B. Chaires, Eur. Biophys. J., 1997, 26, 419–426 CrossRef CAS.
72. R. Gargallo, R. Tauler and A. Izquierdo-Ridorsa, Biopolymers, 1997, 42, 271–283 CrossRef CAS.
73. T. Vojtylova, D. Dospivova, O. Triskova, I. Pilarova, P. Lubal, M. Farkova, L. Trnkova and P. Taborsky, Chem. Pap., 2009, 63, 731–737 CrossRef CAS.
74. J. Gallego, E. B. Golden, D. E. Stanley and B. R. Reid, J. Mol. Biol., 1999, 285, 1039–1052 CrossRef CAS PubMed.
75. T. E. Malliavin, J. Gau, K. Snoussi and J. L. Leroy, Biophys. J., 2003, 84, 3838–3847 CrossRef CAS.
76. J. Smiatek, C. Chen, D. Liu and A. Heuer, J. Phys. Chem. B, 2011, 115, 13788–13795 CrossRef CAS PubMed.
77. R. P. Singh, R. Blossey and F. Cleri, Biophys. J., 2013, 105, 2820–2831 CrossRef CAS PubMed.
78. D. J. Cashman, R. Buscaglia, M. W. Freyer, J. Dettler, L. H. Hurley and E. A. Lewis, J. Mol. Model., 2008, 14, 93–101 CrossRef CAS PubMed.
79. N. Khan, A. Aviño, R. Tauler, C. Gonzalez, R. Eritja and R. Gargallo, Biochimie, 2007, 89, 1562–1572 CrossRef CAS PubMed.
80. R. A. Zager, B. A. Schimpf and D. J. Gmur, Circ. Res., 1993, 72, 837–846 CrossRef CAS.
81. J. Zhou, C. Wei, G. Jia, X. Wang, Z. Feng and C. Li, Mol. BioSyst., 2010, 6, 580–586 RSC.
82. H. A. Day, C. Huguin and Z. A. E. Waller, Chem. Commun., 2013, 49, 7696–7698 RSC.
83. P. Fojtik and M. Vorlickova, Nucleic Acids Res., 2001, 29, 4684–4690 CrossRef CAS PubMed.
84. T. Li and M. Famulok, J. Am. Chem. Soc., 2013, 135, 1593–1599 CrossRef CAS PubMed.
85. M. Canalia and J.-L. Leroy, J. Am. Chem. Soc., 2009, 131, 12870–12871 CrossRef CAS PubMed.
86. A. L. Lieblein, B. Fürtig and H. Schwalbe, ChemBioChem, 2013, 14, 1226–1230 CrossRef CAS PubMed.
87. T. A. Brooks, S. Kendrick and L. Hurley, FEBS J., 2010, 277, 3459–3469 CrossRef CAS PubMed.
88. S. Nonin-Lecomte and J. L. Leroy, J. Mol. Biol., 2001, 309, 491–506 CrossRef CAS PubMed.
89. X. Han, J. L. Leroy and M. Gueron, J. Mol. Biol., 1998, 278, 949–965 CrossRef CAS PubMed.
90. J. W. Park, Y. J. Seo and B. H. Kim, Chem. Commun., 2014, 50, 52–54 RSC.
91. M. Kaushik, M. Prasad, S. Kaushik, A. Singh and S. Kukreti, Biopolymers, 2010, 93, 150–160 CrossRef CAS PubMed.
92. N. Esmaili and J. L. Leroy, Nucleic Acids Res., 2005, 33, 213–224 CrossRef CAS PubMed.
93. J. Weil, T. Min, C. Yang, S. Wang, C. Sutherland, N. Sinha and C. Kang, Acta Crystallogr., Sect. D: Biol. Crystallogr., 1999, 55, 422–429 CrossRef CAS.
94. P. Kumar, A. Verma, S. Maiti, R. Gargallo and S. Chowdhury, Biochemistry, 2005, 44, 16426–16434 CrossRef CAS PubMed.
95. J. Choi, S. Kim, T. Tachikawa, M. Fujitsuka and T. Majima, J. Am. Chem. Soc., 2011, 133, 16146–16153 CrossRef CAS PubMed.
96. D. Monchaud and M.-P. Teulade-Fichou, Org. Biomol. Chem., 2008, 6, 627–636 CAS.
97. X. Li, Y. Peng, J. Ren and X. Qu, Proc. Natl. Acad. Sci. U. S. A., 2006, 103, 19658–19663 CrossRef CAS PubMed.
98. Y. Peng, X. Wang, Y. Xiao, L. Feng, C. Zhao, J. Ren and X. Qu, J. Am. Chem. Soc., 2009, 131, 13813–13818 CrossRef CAS PubMed.
99. Y. Peng, X. Li, J. Ren and X. Qu, Chem. Commun., 2007, 5176–5178 RSC.
100. Y. Chen, K. Qu, C. Zhao, L. Wu, J. Ren, J. Wang and X. Qu, Nat. Commun., 2012, 3, 1074 CrossRef PubMed.
101. T. S. Dexheimer, S. S. Carey, S. Zuohe, V. M. Gokhale, X. Hu, L. B. Murata, E. M. Maes, A. Weichsel, D. Sun, E. J. Meuillet, W. R. Montfort and L. H. Hurley, Mol. Cancer Ther., 2009, 8, 1363–1377 CrossRef CAS PubMed.
102. O. Y. Fedoroff, A. Rangan, V. V. Chemeris and L. H. Hurley, Biochemistry, 2000, 39, 15083–15090 CrossRef CAS PubMed.
103. S. Shi, X. Geng, J. Zhao, T. Yao, C. Wang, D. Yang, L. Zheng and L. Ji, Biochimie, 2010, 92, 370–377 CrossRef CAS PubMed.
104. H. Xu, H. Zhang and X. Qu, J. Inorg. Biochem., 2006, 100, 1646–1652 CrossRef CAS PubMed.
105. X. Chen, X. Zhou, T. Han, J. Wu, J. Zhang and S. Guo, ACS Nano, 2013, 7, 531–537 CrossRef CAS PubMed.
106. L. Wang, Y. Wu, T. Chen and C. Wei, Int. J. Biol. Macromol., 2013, 52, 1–8 CrossRef CAS PubMed.
107. C. Xu, C. Zhao, J. Ren and X. Qu, Chem. Commun., 2011, 47, 8043–8045 RSC.
108. D. L. Ma, M. H. T. Kwan, D. S. H. Chan, P. Lee, H. Yang, V. P. Y. Ma, L. P. Bai, Z. H. Jiang and C. H. Leung, Analyst, 2011, 136, 2692–2696 RSC.
109. X. Ren, F. He and Q. H. Xu, Chem. – Asian J., 2010, 5, 1094–1098 CrossRef CAS PubMed.
110. S. Kendrick, H.-J. Kang, M. Alam, M. Madathil, P. Agrawal, V. Gokhale, D. Yang, S. M. Hecht and L. H. Hurley, J. Am. Chem. Soc., 2014, 136, 4161 CrossRef CAS PubMed.
111. R. Hoshyar, S. Z. Bathaie, A. Kyani and M. F. Mousavi, Nucleosides, Nucleotides Nucleic Acids, 2012, 31, 801–812 CAS.
112. A. Latorre and Á. Somoza, ChemBioChem, 2012, 13, 951–958 CrossRef CAS PubMed.
113. L. Lacroix and J. L. Mergny, Arch. Biochem. Biophys., 2000, 381, 153–163 CrossRef CAS PubMed.
114. C. P. Fenna, V. J. Wilkinson, J. R. P. Arnold, R. Cosstick and J. Fisher, Chem. Commun., 2008, 3567–3569 RSC.
115. S. Robidoux and M. J. Damha, J. Biomol. Struct. Dyn., 1997, 15, 529–535 CAS.
116. J. L. Mergny and L. Lacroix, Nucleic Acids Res., 1998, 26, 4797–4803 CrossRef CAS PubMed.
117. H. Kanehara, M. Mizuguchi, K. Tajima, K. Kanaori and K. Makino, Biochemistry, 1997, 36, 1790–1797 CrossRef CAS PubMed.
118. K. Kanaori, S. Sakamoto, H. Yoshida, P. Guga, W. Stec, K. Tajima and K. Makino, Biochemistry, 2004, 43, 5672–5679 CrossRef CAS PubMed.
119. J. A. Brazier, J. Fisher and R. Cosstick, Angew. Chem., Int. Ed., 2006, 45, 114–117 CrossRef CAS PubMed.
120. R. Cosstick, J. Buckingham, J. Brazier and J. Fisher, Nucleosides, Nucleotides Nucleic Acids, 2007, 26, 555–558 CAS.
121. E. E. Swayze, A. M. Siwkowski, E. V. Wancewicz, M. T. Migawa, T. K. Wyrzykiewicz, G. Hung, B. P. Monia and C. F. Bennett, Nucleic Acids Res., 2006, 35, 687–700 CrossRef PubMed.
122. N. Kumar, J. T. Nielsen, S. Maiti and M. Petersen, Angew. Chem., Int. Ed., 2007, 46, 9220–9222 CrossRef CAS PubMed.
123. N. Kumar, M. Petersen and S. Maiti, Chem. Commun., 2009, 1532–1534 RSC.
124. A. Pasternak and J. Wengel, Bioorg. Med. Chem. Lett., 2011, 21, 752–755 CrossRef CAS PubMed.
125. P. Perlikova, K. K. Karlsen, E. B. Pedersen and J. Wengel, ChemBioChem, 2014, 15, 146–156 CrossRef CAS.
126. U. Diederichsen, Angew. Chem., Int. Ed., 1998, 37, 2273–2276 CrossRef CAS.
127. Y. Krishnan-Ghosh, E. Stephens and S. Balasubramanian, Chem. Commun., 2005, 5278–5280 RSC.
128. N. K. Sharma and K. N. Ganesh, Chem. Commun., 2005, 4330–4332 RSC.
129. R. Bonnet, P. Murat, N. Spinelli and E. Defrancq, Chem. Commun., 2012, 48, 5992–5994 RSC.
130. S. Modi, A. H. Wani and Y. Krishnan, Nucleic Acids Res., 2006, 34, 4354–4363 CrossRef CAS PubMed.
131. T. Li, D. Liu, J. Chen, A. H. F. Lee, J. Qi and A. S. C. Chan, J. Am. Chem. Soc., 2001, 123, 12901–12902 CrossRef CAS.
132. D. Liu, J. Chen, A. H. F. Lee, L. M. C. Chow, A. S. C. Chan and T. Li, Angew. Chem., Int. Ed., 2003, 42, 797–799 CrossRef CAS PubMed.
133. J. S. Hartig and E. T. Kool, Nucleic Acids Res., 2004, 32(19), e152 CrossRef PubMed.
134. T. Zhou, X. Li, M. T. T. Ng, Y. Wang, N. M. Quek, J. Luo, W. Yuan, C. H. Tan, H. Zeng and T. Li, Bioconjugate Chem., 2009, 20, 644–647 CrossRef CAS PubMed.
135. Y. Yang, Y. Sun, Y. Xing, T. Zhang, Z. Wang, Z. Yang and D. Liu, Macromolecules, 2012, 45, 2643–2647 CrossRef CAS.
136. A. A. El-Sayed, E. B. Pedersen and N. A. Khaireldin, Nucleosides, Nucleotides Nucleic Acids, 2012, 31, 872–879 CAS.
137. A. W. I. Stephenson, A. C. Partridge and V. V. Filichev, Chem.–Eur. J., 2011, 17, 6227–6238 CrossRef CAS PubMed.
138. S. Robidoux, R. Klinck, K. Gehring and M. J. Damha, J. Biomol. Struct. Dyn., 1997, 15, 517–527 CAS.
139. F. Seela, S. Budow and P. Leonard, Org. Biomol. Chem., 2007, 5, 1858–1872 CAS.
140. G. Cimino-Reale, E. Pascale, E. Alvino, G. Starace and E. D'Ambrosio, J. Biol. Chem., 2003, 278, 2136–2140 CrossRef CAS PubMed.
141. Y. Zhao, Z. X. Zeng, Z. Y. Kan, Y. H. Hao and Z. Tan, ChemBioChem, 2005, 6, 1957–1960 CrossRef CAS PubMed.
142. S. Modi, C. Nizak, S. Surana, S. Halder and Y. Krishnan, Nat. Nanotechnol., 2013, 8, 459–467 CrossRef CAS PubMed.
143. D. Liu and S. Balasubramanian, Angew. Chem., Int. Ed., 2003, 42, 5734–5736 CrossRef CAS PubMed.
144. J.-L. Leroy, Nucleic Acids Res., 2009, 37, 4127–4134 CrossRef CAS PubMed.
145. C. Chen, M. Li, Y. Xing, Y. Li, C. C. Joedecke, J. Jin, Z. Yang and D. Liu, Langmuir, 2012, 28, 17743–17748 CrossRef CAS PubMed.
146. A. T. Phan and J. L. Mergny, Nucleic Acids Res., 2002, 30, 4618–4625 CrossRef CAS PubMed.
147. W. Li, P. Wu, T. Ohmichi and N. Sugimoto, FEBS Lett., 2002, 526, 77–81 CrossRef CAS.
148. M. del Toro, P. Bucek, A. Aviñó, J. Jaumot, C. González, R. Eritja and R. Gargallo, Biochimie, 2009, 91, 894–902 CrossRef CAS PubMed.
149. P. Bucek, J. Jaumot, A. Aviñó, R. Eritja and R. Gargallo, Chem.–Eur. J., 2009, 15, 12663–12671 CrossRef CAS PubMed.
150. H. T. Lee, C. M. Olsen, L. Waters, H. Sukup and L. A. Marky, Biochimie, 2008, 90, 1052–1063 CrossRef CAS PubMed.
151. W. Li, D. Miyoshi, S. I. Nakano and N. Sugimoto, Biochemistry, 2003, 42, 11736–11744 CrossRef CAS PubMed.
152. S. L. B. Konig, J. L. Huppert, R. K. O. Sigel and A. C. Evans, Nucleic Acids Res., 2013, 41, 7453–7461 CrossRef PubMed.
153. D. Miyoshi, S. Matsumura, W. Li and N. Sugimoto, Nucleosides, Nucleotides Nucleic Acids, 2003, 22, 203–221 CAS.
154. C. Zhao, J. Ren and X. Qu, Chem.–Eur. J., 2008, 14, 5435–5439 CrossRef CAS PubMed.
155. D. Miyoshi, S. Matsumura, S. I. Nakano and N. Sugimoto, J. Am. Chem. Soc., 2004, 126, 165–169 CrossRef CAS PubMed.
156. H. B. Ghodke, R. Krishnan, K. Vignesh, G. V. P. Kumar, C. Narayana and Y. Krishnan, Angew. Chem., Int. Ed., 2007, 46, 2646–2649 CrossRef CAS PubMed.
157. Y. Yang, C. Zhou, T. Zhang, E. Cheng, Z. Yang and D. Liu, Small, 2012, 8, 552–556 CrossRef CAS PubMed.
158. H. Mei, S. Budow and F. Seela, Biomacromolecules, 2012, 13, 4196–4204 CAS.
159. L. Lacroix, J. L. Mergny, J. L. Leroy and C. Helene, Biochemistry, 1996, 35, 8715–8722 CrossRef CAS PubMed.
160. K. Snoussi, S. Nonin-Lecomte and J. L. Leroy, J. Mol. Biol., 2001, 309, 139–153 CrossRef CAS PubMed.
161. S. Chakraborty and Y. Krishnan, Biochimie, 2008, 90, 1088–1095 CrossRef CAS PubMed.
162. S. Chakraborty, S. Modi and Y. Krishnan, Chem. Commun., 2008, 70–72 RSC.
163. B. A. Webb, M. Chimenti, M. P. Jacobson and D. L. Barber, Nat. Rev. Cancer, 2011, 11, 671–677 CrossRef CAS PubMed.
164. A. Rajendran, S. I. Nakano and N. Sugimoto, Chem. Commun., 2010, 46, 1299–1301 RSC.
165. J. Cui, P. Waltman, V. Le and E. Lewis, Molecules, 2013, 18, 12751–12767 CrossRef CAS PubMed.
166. J. J. Wenzel, H. Rossmann, C. Fottner, S. Neuwirth, C. Neukirch, P. Lohse, J. K. Bickmann, T. Minnemann, T. J. Musholt, B. Schneider-Raetzke, M. M. Weber and K. J. Lackner, Clin. Chem., 2009, 55, 1361–1371 CAS.
167. J. L. Huppert, FEBS J., 2010, 277, 3452–3458 CrossRef CAS PubMed.
168. A. T. Phan, M. Gueron and J. L. Leroy, J. Mol. Biol., 2000, 299, 123–144 CrossRef CAS PubMed.
169. R. D. Wells, D. A. Collier, J. C. Hanvey, M. Shimizu and F. Wohlrab, FASEB J., 1988, 2, 2939–2949 CAS.
170. T. Simonsson, M. Pribylova and M. Vorlickova, Biochem. Biophys. Res. Commun., 2000, 278, 158–166 CrossRef CAS PubMed.
171. T. Simonsson, P. Pecinka and M. Kubista, Nucleic Acids Res., 1998, 26, 1167–1172 CrossRef CAS PubMed.
172. J. M. Dettler, R. Buscaglia, J. Cui, D. Cashman, M. Blynn and E. A. Lewis, Biophys. J., 2010, 99, 561–567 CrossRef CAS PubMed.
173. J. Dai, E. Hatzakis, L. H. Hurley and D. Yang, PLoS One, 2010, 5, e11647 Search PubMed.
174. K. Halder, V. Mathur, D. Chugh, A. Verma and S. Chowdhury, Biochem. Biophys. Res. Commun., 2005, 327, 49–56 CrossRef CAS PubMed.
175. D. Sun and L. H. Hurley, J. Med. Chem., 2009, 52, 2863–2874 CrossRef CAS PubMed.
176. S. Kendrick, Y. Akiyama, S. M. Hecht and L. H. Hurley, J. Am. Chem. Soc., 2009, 131, 17667–17676 CrossRef CAS PubMed.
177. K. Guo, V. Gokhale, L. H. Hurley and D. Sun, Nucleic Acids Res., 2008, 36, 4598–4608 CrossRef CAS PubMed.
178. D. J. Uribe, K. Guo, Y. J. Shin and D. Sun, Biochemistry, 2011, 50, 3796–3806 CrossRef CAS PubMed.
179. L. Lacroix, H. Lienard, E. Labourier, M. Djavaheri-Mergny, J. Lacoste, H. Leffers, J. Tazi, C. Helene and J. L. Mergny, Nucleic Acids Res., 2000, 28, 1564–1575 CrossRef CAS PubMed.
180. Y. M. K. Yoga, D. A. K. Traore, M. Sidiqi, C. Szeto, N. R. Pendini, A. Barker, P. J. Leedman, J. A. Wilce and M. C. J. Wilce, Nucleic Acids Res., 2012, 40, 5101–5114 CrossRef CAS PubMed.
181. E. F. Michelotti, G. A. Michelotti, A. I. Aronsohn and D. Levens, Mol. Cell. Biol., 1996, 16, 2350–2360 CAS.
182. A. Bandiera, G. Tell, E. Marsich, A. Scaloni, G. Pocsfalvi, A. Akindahunsi, L. Cesaratto and G. Manzini, Arch. Biochem. Biophys., 2003, 409, 305–314 CrossRef CAS.
183. J. E. Eid and B. Sollner-Webb, Mol. Cell. Biol., 1995, 15, 389–397 CAS.
184. E. Marsich, A. Piccini, L. E. Xodo and G. Manzini, Nucleic Acids Res., 1996, 24, 4029–4033 CrossRef CAS PubMed.
185. A. Cortés, D. Huertas, L. Fanti, S. Pimpinelli, F. X. Marsellach, B. Piña and F. Azorín, EMBO J., 1999, 18, 3820–3833 CrossRef PubMed.
186. S. Fenn, Z. Du, J. K. Lee, R. Tjhen, R. M. Stroud and T. L. James, Nucleic Acids Res., 2007, 35, 2651–2660 CrossRef CAS PubMed.
187. H.-J. Kang, S. Kendrick, S. M. Hecht and L. H. Hurley, J. Am. Chem. Soc., 2014, 136, 4172 CrossRef CAS PubMed.
188. Z. Du, J. K. Lee, R. Tjhen, S. Li, H. Pan, R. M. Stroud and T. L. James, J. Biol. Chem., 2005, 280, 38823–38830 CrossRef CAS PubMed.
189. J. F. Cornuel, A. Moraillon and M. Gueron, Biochimie, 2002, 84, 279–289 CrossRef CAS.
190. Z. X. Zeng, Y. Zhao, Y. H. Hao and Z. Tan, J. Mol. Recognit., 2005, 18, 267–271 CrossRef CAS PubMed.
191. C. Teller and I. Willner, Curr. Opin. Biotechnol., 2010, 21, 376–391 CrossRef CAS PubMed.
192. J. Elbaz, S. Shimron and I. Willner, Chem. Commun., 2010, 46, 1209–1211 RSC.
193. C. Chen, F. Pu, Z. Huang, Z. Liu, J. Ren and X. Qu, Nucleic Acids Res., 2011, 39, 1638–1644 CrossRef CAS PubMed.
194. K. Kanaori, N. Shibayama, K. Gohda, K. Tajima and K. Makino, Nucleic Acids Res., 2001, 29, 831–840 CrossRef CAS PubMed.
195. J. A. Brazier, A. Shah and G. D. Brown, Chem. Commun., 2012, 48, 10739–10741 RSC.
196. S. Saxena, A. Bansal and S. Kukreti, Arch. Biochem. Biophys., 2008, 471, 95–108 CrossRef CAS PubMed.
197. S. S. Pataskar, D. Dash and S. K. Brahmachari, J. Biomol. Struct. Dyn., 2001, 19, 307–313 CAS.
198. P. Catasti, X. Chen, L. L. Deaven, R. K. Moyzis, E. M. Bradbury and G. Gupta, J. Mol. Biol., 1997, 272, 369–382 CrossRef CAS PubMed.
199. V. V. Jolad, F. K. Murad, J. R. P. Arnold and J. Fisher, Org. Biomol. Chem., 2005, 3, 2234–2236 CAS.
200. Y. Xu and H. Sugiyama, Nucleic Acids Res., 2006, 34, 949–954 CrossRef CAS PubMed.

This journal is © The Royal Society of Chemistry 2014