Electrochemical Monitoring of the Reversible ... - ACS Publications

Feb 26, 2018 - differ from the classical Watson and Crick double helix.1 These so-called noncanonical DNA structures, such as G-quadruplexes and i-mot...
0 downloads 8 Views 995KB Size
Download PDF
Article pubs.acs.org/Langmuir

Cite This: Langmuir XXXX, XXX, XXX−XXX

Electrochemical Monitoring of the Reversible Folding of Surface-Immobilized DNA i‑Motifs Catherine Adam,† José Manuel Olmos,†,‡ and Thomas Doneux*,† †

Chimie Analytique et Chimie des Interfaces, Université libre de Bruxelles (ULB), Boulevard du Triomphe, 2, CP255, B-1050 Bruxelles, Belgium ‡ Departamento de Química Física, Facultad de Química, Regional Campus of International Excellence “Campus Mare Nostrum”, Universidad de Murcia, 30100 Murcia, Spain S Supporting Information *

ABSTRACT: Two cytosine (C) rich DNA sequences folding in i-motif upon protonation of C at low pH have been immobilized at gold electrodes to study the impact of the electrode|electrolyte interface on the stability of the noncanonical DNA secondary structure. The effects of the molecular composition and environment on the melting and folding of the structures immobilized at the gold surface have been compared to the properties of the DNA strands in solution. The DNA folding into i-motif upon protonation, both at the surface and in solution, results in a significant variation of the charge density which is monitored electrochemically through the electrostatic interactions between the DNA strand and the electroactive hexaammineruthenium(III). This method is shown to be sufficiently sensitive to distinguish hemiprotonated folded state and single strand unfolded state of i-motif. The pH of melting has been determined for both sequences in the bulk and at the gold|electrolyte interface. The results evidence a stabilizing effect of the interface on i-motif structure, whereby the pH of melting is higher for the sequences immobilized at the surface. The reversibility and precision of the electrochemical model described here allows a clear and simple characterization of DNA structures and does not require any labeling of the sequence.

■

INTRODUCTION Some highly repetitive DNA sequences, such as the human telomeric sequence, are known to form secondary structures that differ from the classical Watson and Crick double helix.1 These so-called noncanonical DNA structures, such as G-quadruplexes and i-motifs, have attracted a lot of interest in the recent decades.2,3 In addition to their biological relevance, their robustness and reversibility are well adapted to the development of nanotechnologies, for example, in the design of responsive materials or nanomachines.4,5 The stability of noncanonical structures strongly depends on the molecular environment and slight changes (pH, temperature, ionic strength, or ionic composition) lead to the formation (folding) or disruption (melting) of the structure. Their stability with respect to the physicochemical environment of the solution is being intensively studied.3,6,7 However, the behavior of DNA noncanonical structures at solid|liquid interfaces is poorly discussed8 in comparison to classical DNA double helix,9−12 but represents a critical feature for the conception of sensors or nanodevices, including electrochemical DNA sensors. In such platforms, where single-strand DNA is immobilized on the electrode surface, the structure of the interface and the effect of the electric field need to be well understood and controlled in order to design accurate analytical systems. The present work is focused on i-motifs (Scheme 1), which are noncanonical DNA structures resulting from the protonation of cytosines (C) and the formation of intercalated C−H+−C hemiprotonated base pairs. © XXXX American Chemical Society

Scheme 1. i-Motif Formation from C-Rich DNA Sequence

These structures can be formed between separated DNA strands or through intramolecular bonding in C-rich sequences (e.g., human telomeric sequence) leading to the folding of the strand on itself. The stability of i-motifs is directly linked to the protonation of cytosines and thus strongly depends on the pH. The pH of melting (pHm) is defined as the pH at the midpoint of the proton induced single strand to i-motif transition and is a measure of the proneness of a given sequence to fold into an i-motif. The pHm, together with the temperature of melting (Tm), are key indicators to characterize the stability of the i-motif structure. The effect of the interface on the pHm is discussed Received: November 29, 2017 Revised: February 9, 2018

A

DOI: 10.1021/acs.langmuir.7b04088 Langmuir XXXX, XXX, XXX−XXX

Article

Langmuir

monolayers using 4-mercaptobutan-1-ol as the diluent in one step by immersion overnight in 1−1 DNA thiolated sequence-mercaptobutanol solution (total concentration 20 μM) in phosphate buffer (1 M) pH 7.4.17 Electrodes were thoroughly rinsed using phosphate buffer (1 M) pH 7.4 then the buffer electrolyte before measurement. For the titration experiments, the gold electrodes were modified with 6-mercapto1-hexanol to avoid DNA adsorption.19 Electrochemistry. A three electrode setup connected to a PGSTAT30 (Autolab-Metrohm) potentiostat equipped with Scangen and a Frequency Response Analyzer modules was used for all voltammetric and electrochemical impedance spectroscopy measurements. To avoid conductivity limitation in low concentration electrolytes, a fourth electrode was used to avoid artifacts in the high frequency range. In this case, a platinum wire was connected and capacitively coupled to the Ag|AgCl, KCl(saturated) reference electrode.20 For cyclic voltammetry under convection (CVC), a constant flux of species was established by fixing the rotation speed of the magnetic stirrer during all the experiment. Frequency of 37 Hz and amplitude of 10 mV were used for ac voltammetry. The faradaic contribution of the imaginary electrode admittance of adsorbed hexaammineruthenium(III), ″ , is determined as the peak height measured from ac abbreviated as Yel,p voltammograms. All the solutions were purged with nitrogen to remove dissolved oxygen.

herein by comparing the values determined in bulk solution (pHm,sol) and those measured for i-motifs immobilized at the surface (pHm,surf). The pH of melting pHm,surf is herein defined as the pH value, in the bulk solution, at which 50% of the surface-immobilized DNA is in the folded state. To determine the (pHm,surf) values a method using the electrostatic interaction between the negatively charged DNA-backbone and hexaammineruthenium(III) is proposed. This method, well documented in the literature,13−17 is shown to be sensitive enough to allow the electrochemical monitoring of charge density upon protonation and folding of the DNA strand. Results presented here aim at differentiating DNA structures (folded vs unfolded) using electrochemistry without labeling and highlight the importance of considering interfacial phenomena in the use of immobilized oligonucleotides on electrode surfaces.

■

EXPERIMENTAL SECTION

Material. All solutions were prepared using ultrapure water (Milli-Q from Millipore). DNA sequences 5′-(CCC TAA)3 CCC T-3′ (human telomeric repeat, sequence A), 5′-(CCC TTT)3 CCC T-3′ (sequence B), 5′-TTT AAT ATT AAT ATT AAT TTA T-3′ (control sequence showing no folding into i-motif structure) and their thiolated equivalent (5′-HS-(CH2)6-(CCC TAA)3 CCC T-3′, 5′-HS-(CH2)6-(CCC TTT)3 CCC T-3′ and 5′-HS-(CH2)6-TTT AAT ATT AAT ATT AAT TTA T-3′) were purchased from Eurogentec (Belgium) at a HPLCpurified quality. DNA stock solutions were prepared in phosphate buffer pH 7.4 (10 mM) and stored at −20 °C. Hexaammineruthenium(III) chloride (>32.1% Ru, Alfa Aesar), 4-mercaptobutan-1-ol (>97%, Fluka), and 6-mercapto-1-hexanol (97%, Sigma-Aldrich) were used as received. Buffer mixtures were prepared from 2-(N-morpholino)ethanesulfonic acid (MES monohydrate >99.5%, Sigma-Aldrich), trishydroxymethylaminomethane (Tris, 99.9% Sigma), and acetic acid (Suprapur, Merck) and the pH was adjusted using KOH or HClO4 0.1 M solutions. The buffer mixture in the molar proportion of 2−1−1 of Tris-MES-acetic acid (hereafter TMA) was used to keep a constant ionic strength of 8.3 mM upon addition of acid or base.18 Spectroscopy. UV-melting experiments were performed using a SAFAS-Monaco UVmc2 spectrophotometer equipped with a Peltier temperature control unit. Quartz cuvettes (Hellma Analytics) of 1 cm optical path and a total volume of 500 μL were used for each experiment. A concentration of 4 μM of DNA in TMA was chosen for the study and ionic strength was kept similar to the electrochemical experimental medium (8.3 mM) for comparison. At pH around 6, spectra were recorded at 75 and 25 °C to calculate thermal differential spectra (TDS; see Figure S1), from which the wavelength of 240 nm was found appropriate to measure temperature melting curves (Abs. vs T, see Figure S2) of the DNA structure. Temperature melting curves were measured between 25 and 75 °C with a scan rate of 1 °C min−1. The fraction of i-motif (θ) was then calculated using the eq 1:

θ=

Abs240 (T ) − Abs240 (75°C) Abs240 (25°C) − Abs240 (75°C)

■

(1)

where Abs240(T) is the absorbance at 240 nm at the given temperature. Equivalent studies were performed as a function of pH. The pH differential spectra were obtained at 25 °C by subtracting absorbance spectra at low pH (pHi‑motif) to the spectra at high pH (pHmelted) and the fraction of i-motif (θ) is then calculated using eq 2: θ=

Abs240 (pH) − Abs240 (pH melted) Abs240 (pHi‐motif ) − Abs240 (pH melted)

RESULTS AND DISCUSSION

Folding in Solution. Intramolecular i-motifs result from the folding of particular DNA single strands upon the protonation of cytosines in sequence leading to the formation of hemiprotonated C−H+−C pairs.2 Their stability strongly depends on the molecular environment (pH, temperature, ionic strength, or ionic composition). Two i-motif forming sequences have been investigated: sequence A (5′-(CCC TAA)3 CCC T-3′) is the human telomeric sequence and sequence B (5′-(CCC TTT)3 CCC T-3′) is used as comparison, since the presence of different nucleotides in the loop is known to influence the stability of i-motifs.7 Spectroscopic studies of the two sequences have been carried out under pH and temperature control. Thermal and pH differential UV-spectra of both sequences (see Figure S1) present the typical i-motif signature with two positive peaks at 240 and 260 nm and one negative peak at 295 nm.21 This confirms that the two investigated sequences indeed fold into an i-motif in the lower pH- and temperature-ranges, and unfold at higher pH and temperatures. The fraction of i-motif, θ, as a function of pH or temperature has been calculated from the corresponding melting curves (Abs240 vs pH or vs T, respectively) using eqs 1 and 2 described in the experimental section and is represented for both sequences in Figure S2. The temperature and pH of melting (Tm,sol and pHm,sol), corresponding to the pH and T at which half of the DNA sequences are in the folded state (assuming a two state transition) are used to describe the stability of the motif in our experimental medium. The pHm,sol of sequence A and sequence B are 6.6 and 6.9, respectively. The stability of i-motif is influenced by the length and composition of the loops of the sequence.7,22 According to McKim et al.,7 sequence A (pHm,sol = 5.86) presents a lower stability than sequence B (pHm,sol = 6.15) because of the unstabilizing effect of bulky adenine present in the loop of the motif. In our results the same trend is observed with a lower value of pHm,sol for sequence A than sequence B and with an equivalent difference about 0.3 pH unit between the two sequences. Our values are about 0.7 pH unit higher than those reported by McKim et al.,7 which is explained by the lower ionic strength of our experimental medium, known to stabilize i-motif structures.23 This observation might seem counterintuitive compared to dsDNA, where a higher ionic strength favors the

(2)

where Abs240(pH) is the absorbance at 240 nm at the given pH. Preparation of the Electrodes. Polycrystalline gold disc electrodes were polished using 1.0 μM alumina-water slurry on polishing cloth and sonicated in water for 10 min. After electrochemical cleaning and characterization in HClO4 (0.1 M), the thiolated sequences A, B, and control were immobilized on gold electrodes as self-assembled B

DOI: 10.1021/acs.langmuir.7b04088 Langmuir XXXX, XXX, XXX−XXX

Article

Langmuir

Figure 1. Differential pulse voltammograms of 40 μM [Ru(NH3)6]3+ in TMA buffer mixture with successive additions of sequence A at pH 8.4 (a) and pH 5.3 (b). Evolution of the peak current intensities as a function of DNA concentration for sequence A at pH 8.4 (c) (data from a) and pH 5.3 (d) (data from b). Dotted black lines are guided to the eye to show difference in stoichiometry of interaction. Solid red lines are best fit to a dynamic equilibrium (see main text for details). Fitting parameters are reported in Table 1.

at pH = 5.3 (pHm,sol), the DNA strand is unprotonated and in the unfolded state, whereas

i tot = Ω·Deff ζ ·c

Deff = DRu,f cb = C

(4)

c − cb c + DRu,b b c c

K (c + sc DNA) + 1 −

(5)

[K (c + sc DNA) + 1]2 − 4K 2scc DNA 2K

(6) DOI: 10.1021/acs.langmuir.7b04088 Langmuir XXXX, XXX, XXX−XXX

Article

Langmuir

Table 1. Values of Parameters Extracted from the Fitting Procedures, Dynamic Model (from DPV Data in Figures 1 and S5) pH

sequence

DRu,fa (10−6 cm2 s−1)

DRu,bb (10−6 cm2 s−1)

8.4

A B A B control

6.51 6.51 6.51 6.51 6.51

1.88 1.52 1.43 1.27 1.41

5.3

Kc (105 L mol−1) 3.7 3.5 3.0 3.2 3.7

± ± ± ± ±

1.1 1.6 0.7 0.5 0.6

sc 7.3 7.3 3.9 4.4 6.2

± ± ± ± ±

R2c 0.5 0.7 0.2 0.1 0.2

99.3 99.5 99.8 99.9 99.9

The value was fixed during fitting and was calculated in our experimental conditions from CV at different scan rates (data not shown). bThese values were fixed during fitting and calculated from data at the end of titration. cK is the value of the intrinsic binding constant and s is the stoichiometry of interaction, and those values are extracted from the fitting procedures. R2 is the correlation coefficient, as provided by the non linear fit tool of the software. The deviations are those associated with the fitting. a

where itot is the peak current for DPV and CV measurements and the plateau current in CVC experiments, Ω is a constant coefficient depending on the employed electrochemical method and is detailed in our previous work,19 c is the total concentration of [Ru(NH3)6]3+, ζ = 1/2 for CV and DPV, ζ = 2/3 for CVC and Deff is the effective diffusion coefficient of hexaammineruthenium corresponding to the weighed contributions of bound (DRu,b) and free (DRu,f) species and is described in eq 5. The concentration of bound [Ru(NH3)6]3+, cb, is calculated using eq 6, where cDNA is the total concentration of DNA added to the electrolyte. Results of the model for DPV measurement are contained in Table 1 (see SI, Tables S1 and S2, for other methods). The diffusion coefficient of free and bound [Ru(NH3)6]3+ are estimated from cyclic votammetry at different scan rates at the beginning and at the end of the titration respectively (data not shown). Using those values (see Table 1), the experimental data were fitted and values of K and s were extracted. The value of the intrinsic binding constant K does not vary appreciably between sequences or with pH. For both sequences, the stoichiometry of interaction is close to 7 [Ru(NH3)6]3+ per DNA strand at high pH (pH 8.4), which points reasonably to a total compensation of charge when the motif is unfolded (1 [Ru(NH3)6]3+ compensates for 3 phosphate groups, and the sequences are 22-mer). A good correlation between the number of bases and the quantity of bound redox probe, which is a typical behavior for ssDNA, can thus be attested in our experimental medium at pH 8.4. At low pH (pH 5.3), a clear decrease of s (≈4) is observed for both sequences, indicating a lower amount of [Ru(NH3)6]3+ interacting with the DNA strand. Such a significant decrease does not occur for the negative control sequence which is not able to fold into an i-motif. The partial compensation of charge of the DNA strand, linked to the protonation of C and the formation of C−H+−C bond leads to the release of [Ru(NH3)6]3+. The folding of i-motif in our sequence leads to the formation of 6 C−H+−C bonds and hence to the compensation of six negative charges or, equivalently, a difference of two [Ru(NH3)6]3+ per strand. Accordingly, a total charge compensation should ideally correspond to a stoichiometry of 5.3 hexaammineruthenium(III) per strand at the low pH. The extracted stoichiometries are slightly smaller, a fact that has also been observed with other DNA secondary structures such as G-quadruplexes and could possibly be linked to steric hindrance or charge densities issues that would not occur for single strand, unfolded sequences. In spite of this slight deviation from the total charge compensation hypothesis, the method presented here is sensitive enough in order to follow the change in charge distribution occurring with the folding/unfolding of the i-motifs. Folding of i-Motif at the Surface. Interaction of [Ru(NH3)6]3+ with DNA has also been widely exploited to

study monolayers immobilized on gold electrodes and presents great advantages since the interaction is quantitative, reversible, and does not require any labeling of the DNA sequence. It is thus possible to extend the method presented here to study the folding of DNA i-motif structures immobilized on gold surfaces, which is valuable for the design of sensors. Similarly to the characterization of the DNA sequences in solution presented above, electrochemical methods based on the reduction of hexaammineruthenium(III) have been used to monitor the folding of DNA i-motifs at the interface. In Figure 2, cyclic

Figure 2. Cyclic voltammograms (50 mV s−1) (a) and ac voltammograms (b) of sequence B immobilized on gold electrode in 50 μM [Ru(NH3)6]3+ in TMA buffer mixture at pH 8.4 (black) and 5.5 (red).

voltammetry and ac voltammetry of the immobilized DNA i-motif of sequence B in the presence of hexaammineruthenium at two pH are presented (see Figure S9 for sequence A). The two pH, where the DNA strand is assumed to be folded (pH 5.5) or unfolded (pH 8.4), were chosen far enough from the pHm,sol determined from the results obtained in solution. In cyclic voltammetry (Figure 2a), two cathodic peaks are observed. The first peak at −0.13 V versus Ag|AgCl corresponds to the diffusion limited reduction of hexaammineruthenium(III) dissolved in the bulk solution. The second peak is observed at lower potential (−0.29 V) and is characteristic of a surface confined redox probe, corresponding to [Ru(NH3)6]3+ adsorbed on the DNA backbone.17,24 In ac voltammetry (Figure 2b), the imaginary part of admittance is plotted in order to minimize the contribution of dissolved [Ru(NH3)6]3+ versus the contribution D

DOI: 10.1021/acs.langmuir.7b04088 Langmuir XXXX, XXX, XXX−XXX

Article

Langmuir of electrostatically adsorbed [Ru(NH3)6]3+, allowing thus a more precise characterization of the surface phenomena. In both measurements the signal of the adsorbed hexaammineruthenium is consequently diminished at low pH, corresponding to the decrease of the amount of the redox probe interacting with DNA backbone and attributed to the formation of the i-motif structure. The decrease of the height of the adsorption peak between the two pHs is observed for both i-motif sequences but is not observed for the control sequence (see Figure S10). The electrochemical monitoring of the interaction between [Ru(NH3)6]3+ and the DNA backbone is thus sensitive enough to differentiate the two states (folded and unfolded) of the motif immobilized at the gold electrode. Figure 3 represents the height of the ac voltammetric peak (Y″el,p) measured in different solutions alternating the pH between pH 6 (folded state) and pH 9 (unfolded state). The conservation of the intensity after various cycles is a good indication of the stability of the monolayer and of the reversibility of the formation of the i-motifs at the interface, which are essential features for the use of DNA in biosensing application or nanotechnologies. The faradaic contribution of the imaginary admittance Y″el,p plotted in Figure 3 is directly related to the surface coverage of

the evolution of the amount of adsorbed hexaammineruthenium(III) in the intermediate pH range. Figure 4 represents the

Figure 4. ac voltammograms of immobilized sequence B on gold electrode in 50 μM [Ru(NH3)6]3+ in TMA buffer mixture at different ″ as a function of pH (b), red line is guided to pH (a). Peak height Yel,p the eye (see Figure S11 for sequence A).

variation of the ac voltammetry signal as a function of pH for sequence B (see Figure S11 for sequence A). For the two sequences, a clear increase is observed at the peak potential corresponding to adsorbed [Ru(NH3)6]3+ on the immobilized DNA strand while increasing the pH. In this experiment, careful attention has been paid to maintain a constant ionic strength even though potassium hydroxide is added directly in the electrochemical cell. This was ensured by the use of the TMA buffer mixture as described in the experimental part.18 The gradual increase of the quantity of adsorbed [Ru(NH3)6]3+ is attributed to the unfolding of the i-motif structure to single strand and allows the electrochemical monitoring of the structural change of the DNA structure. The fraction of i-motif is calculated from the ac voltammograms of Figure 4, by analogy to what is reported in solution, using eq 7:

Figure 3. Peak height in ac voltammogram (Yel,p ″ ) in 50 μM [Ru(NH3)6]3+ in TMA buffer mixture measured in two distinct solutions and alternating pH between 6 and 9 for sequences A (a) and B (b).

[Ru(NH3)6]3+ adsorbed on the DNA strand immobilized at the gold electrode. The DNA surface coverage on the gold electrode in our experimental conditions was calculated from charge integration (data not shown) and is equal to 1.0 (±0.3) × 10−11 mol cm−2 and 1.5 (±0.3) × 10−11 mol cm−2 for sequence A and sequence B, respectively, which explains the lower values of Y″el,pon Figure 3a compared to Figure 3b. In agreement with this trend in surface coverage, the peak potential of adsorbed [Ru(NH3)6]3+ is more negative for sequence B (see Figure 2) than for sequence A (see Figure S9).25 For both sequences, the surface coverage in hexaammineruthenium is lower in the folded state (pH 6) than in the unfolded state (pH 9), with the same relative decrease of the intensity. Assuming a total compensation of charge at pH 9 and thus a stoichiometry of interaction of 7.33 [Ru(NH3)6]3+/ DNA strand (theoretically calculated for the 22-mer DNA strand), the estimated relative stoichiometry at pH 6 is equivalent to 4.2 for sequence A and 4.5 for sequence B. The trend is thus similar to that observed for the sequence in solution described in the first part of this work. Interestingly, the variations of peak heights are very significant between the two extreme pH, opening the possibility to monitor

θ=

′′ ′′ Y el,p (pH) − Y el,p (pH melted) ′′ ′′ Y el,p (pHi‐motif ) − Y el,p (pH melted)

(7)

As can be seen in Figure 5, the pH at the midpoint of the proton induced single strand to i-motif transition at the gold| electrolyte interface (pHm,surf), corresponding to the bulk pH at which half of the immobilized DNA strands are in a folded state is about 7.2 and 7.7 for sequences A and B, respectively. This figure compares the data obtained in solution, using the spectroscopic methods described above, and the results obtained at the interface using our electrochemical method. The trend in the stability between sequence A and sequence B is maintained at the surface, with a higher pHm,surf for sequence B, in line with E

DOI: 10.1021/acs.langmuir.7b04088 Langmuir XXXX, XXX, XXX−XXX

Article

Langmuir

situation where the volume fraction occupied by biomacromolecules, small molecules, ions, or cosolutes is very high. The interfacial environment at the surface can possibly mimic such crowding conditions: it is characterized by a very high density of DNA strands, as compared to typical dilute solutions conditions and by local concentrations of anions and cations, including protons, that can be much larger than in the bulk. Incidentally, the surface coverage of DNA in the present work is so high that intermolecular association between strands cannot be ruled out. These various contributions are not mutually exclusive, and each of them might contribute to some extent to the observed stabilization effect at the surface. Disentangling them and providing quantitative evaluation of their impact would require extensive and systematic investigations that are beyond the scope of the present work. Nevertheless, this work demonstrates that, in the case of DNA C-rich sequences, the secondary structure can be modified at the surface of the electrode. This effect occurs around physiological pH reinforcing the idea that important considerations need to be made for the use of immobilized DNA in its different applications.

Figure 5. pH-melting curves for sequences A (red) and B (blue) in solution (from UV-melting experiment; see Figure S2; dashed lines) and immobilized at the surface of gold electrode (from ac voltammogramms and calculated from eq 7; full lines). The data at the surface correspond to the average of four independent experiments, and the shaded envelope represents the standard deviation of the experimental data.

■

the destabilizing effect of bulky adenine contained in the loop of the motif, as observed in solution.7 Of more interest is the fact that for both sequences, pHm,surf is higher than pHm,sol. This result points to a stabilizing effect of the interface on the i-motif structure which remains folded at or close to the physiological pH. The effect of the interface thus plays an important role on the properties of adsorbed probes and needs to be carefully addressed in the development of technologies. In the field of bioanalytics, such effects are often neglected and the properties of an immobilized probe are often assumed to be similar to the properties exhibited in the bulk solution. However, the results presented here highlight the non-negligible effect of the interface on the secondary structure of DNA which will directly impact the recognition of the target analyte. This stabilizing effect of the surface is likely linked to the different physicochemical environment experienced by the DNA at the interface, although its exact origin is difficult to pinpoint owing to the complexity of the system. Nevertheless, various possible contributions to the actual value of the measured pHm,surf can be identified. (i) The formation of i-motif being a protonationdriven phenomenon, the pKa of cytosine should have a direct influence on the overall stability of the folded structure. This pKa is not necessarily identical for DNA immobilized at the surface and for DNA dissolved in solution. It is for instance well-documented that the surface pKa of carboxylic acids immobilized through self-assembled monolayers differs significantly from the bulk values and is affected by parameters such as the ionic strength, the electrolyte composition or even the electric field.26−29 (ii) The local proton concentration and hence the pH can vary between the bulk and the surface, or even between the bulk and the close vicinity of the DNA strands. It was very recently reported that, for i-motifs dissolved in solution, the local pH in the immediate vicinity of the nucleic acid is significantly lower than the bulk pH.30 This fact is rationalized by a simple Boltzmann distribution effect, whereby an enrichment in cations, including protons, occurs close to the negatively charged DNA strands due to the existence of a net negative electrostatic potential.30 In the present case, it is very likely that the surface potential is also negative, owing to the high density of negatively charged immobilized DNA, so that a similar Boltzmann effect would also increase the local concentration of protons. (iii) In solution, it is known that the stability of i-motifs is strongly affected by molecular crowding,31−33 which describes a

SUMMARY AND CONCLUSIONS Two C-rich DNA sequences, including the human telomeric repeat, have been studied by spectroscopy and electrochemistry to compare the folding of i-motif structures in bulk solution and at the gold|electrolyte interface. The variations of charge density between the folded and unfolded states are large enough to impact the electrostatic interactions between the DNA strands and the electroactive hexaammineruthenium(III) cation. The electrochemical quantification of these interactions, conducted in solution and for surface-immobilized sequences, revealed a stabilizing effect of the surface on the i-motif structure around physiological pH. This work shows that the notion of molecular environment should include surfaces as an additional lever to control the stability of noncanonical DNA structures. This will be beneficial for the design of accurate DNA sensing devices and the use of oligonucleotides in nanotechnology. It also highlights the need of considering interfacial phenomena for immobilized DNA and opens the way to study the impact of an electric field on the folding/melting of noncanonical DNA structures.

■

ASSOCIATED CONTENT

* Supporting Information S

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.langmuir.7b04088. UV-melting experiments for sequences A and B and control sequence and data for the titration of [Ru(NH3)6]3+, CV, and ACV of immobilized sequence A and control sequence (PDF).

■

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Tel.: +32 2 650 35 80. Fax: +32 2 650 29 34. ORCID

Thomas Doneux: 0000-0002-9082-8826 Author Contributions

The manuscript was written through contributions of all authors. All authors have given approval to the final version of the manuscript. F

DOI: 10.1021/acs.langmuir.7b04088 Langmuir XXXX, XXX, XXX−XXX

Article

Langmuir Funding

(16) Steichen, M.; Doneux, T.; Buess-Herman, C. On the Adsorption of Hexaammineruthenium (III) at Anionic Self-Assembled Monolayers. Electrochim. Acta 2008, 53 (21), 6202−6208. (17) Doneux, T.; De Rache, A.; Triffaux, E.; Meunier, A.; Steichen, M.; Buess-Herman, C. Optimization of the Probe Coverage in DNA Biosensors by a One-Step Coadsorption Procedure. ChemElectroChem 2014, 1 (1), 147−157. (18) Ellis, K. J.; Morrison, J. F. Buffers of Constant Ionic Strength for Studying pH-Dependent Processes. Methods Enzymol. 1982, 87 (C), 405−426. (19) De Rache, A.; Buess-Herman, C.; Doneux, T. Electrochemical Square Scheme Analysis of Macromolecule-Electroactive Ligand Interactions, and Its Application to DNA Binding. J. Electroanal. Chem. 2015, 745, 44−55. (20) Mansfeld, F. Minimization of High-Frequency Phase Shifts in Impedance Measurements. J. Electrochem. Soc. 1988, 135 (4), 906− 907. (21) Mergny, J. L.; Li, J.; Lacroix, L.; Amrane, S.; Chaires, J. B. Thermal Difference Spectra: A Specific Signature for Nucleic Acid Structures. Nucleic Acids Res. 2005, 33 (16), 1−6. (22) Choi, J.; Kim, S.; Tachikawa, T.; Fujitsuka, M.; Majima, T. pHInduced Intramolecular Folding Dynamics of I-Motif DNA. J. Am. Chem. Soc. 2011, 133 (40), 16146−16153. (23) Völker, J.; Klump, H. H.; Breslauer, K. J. The Energetics of IDNA Tetraplex Structures Formed Intermolecularly The Energetics of I-DNA Tetraplex Structures Formed Intermolecularly by d(TC5). Biopolymers 2007, 86 (2), 136−147. (24) Steel, A. B.; Herne, T. M.; Tarlov, M. J. Electrochemical Quantitation of DNA Immobilized on Gold. Anal. Chem. 1998, 70 (22), 4670−4677. (25) Shen, G.; Tercero, N.; Gaspard, M. A.; Varughese, B.; Shepard, K.; Levicky, R. Charging Behavior of Single-Stranded DNA Polyelectrolyte Brushes. J. Am. Chem. Soc. 2006, 128 (26), 8427−8433. (26) Bain, C. D.; Whitesides, G. M. A Study by Contact Angle of the Acid-Base Behavior of Monolayers Containing ω-Mercaptocarboxylic Acids Adsorbed on Gold: An Example of Reactive Spreading. Langmuir 1989, 5 (6), 1370−1378. (27) Creager, S. E.; Clarke, J. Contact-Angle Titrations of Mixed ωMercaptoalkanoic Acid/Alkanethiol Monolayers on Gold. Reactive vs Nonreactive Spreading, and Chain Length Effects on Surface pKa Values. Langmuir 1994, 10 (10), 3675−3683. (28) Burgess, I.; Seivewright, B.; Lennox, R. B. Electric Field Driven Protonation/deprotonation of Self-Assembled Monolayers of AcidTerminated Thiols. Langmuir 2006, 22 (9), 4420−4428. (29) Ma, C.; Harris, J. M. Surface-Enhanced Raman Spectroscopy Investigation of the Potential-Dependent Acid-Base Chemistry of Silver-Immobilized 2-Mercaptobenzoic Acid. Langmuir 2011, 27 (7), 3527−3533. (30) Ren, W.; Zheng, K.; Liao, C.; Yang, J.; Zhao, J. Charge Evolution during the Unfolding of a Single I-Motif DNA. Phys. Chem. Chem. Phys. 2018, 20, 916−924. (31) Miyoshi, D.; Matsumura, S.; Nakano, S.; Sugimoto, N. Duplex Dissociation of Telomere DNAs Induced by Molecular Crowding. J. Am. Chem. Soc. 2004, 126 (1), 165−169. (32) Saxena, S.; Joshi, S.; Shankaraswamy, J.; Tyagi, S.; Kukreti, S. Magnesium and Molecular Crowding of the Cosolutes Stabilize the IMotif Structure at Physiological pH. Biopolymers 2017, 107 (7), 1−11. (33) Nakano, S. I.; Miyoshi, D.; Sugimoto, N. Effects of Molecular Crowding on the Structures, Interactions, and Functions of Nucleic Acids. Chem. Rev. 2014, 114 (5), 2733−2758.

This work was supported by the Fonds de la Recherche Scientifique-FNRS under Grant No. MIS F.4542.16. JMO thanks the Ministerio de Economiá y Competitividad for the grant received under the Project CTQ-2012−36700 cofunded by the European Regional Development Fund. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS The authors thank Dr. Eduardo Laborda (Universidad de Murcia, Spain) for the critical reading of the manuscript and helpful comments.

■

REFERENCES

(1) Phan, A. T.; Mergny, J.-L. Human Telomeric DNA: GQuadruplex, I-Motif and Watson-Crick Double Helix. Nucleic Acids Res. 2002, 30 (21), 4618−4625. (2) Choi, J.; Majima, T. Conformational Changes of Non-B DNA. Chem. Soc. Rev. 2011, 40 (12), 5893−5909. (3) Biver, T. Stabilisation of Non-Canonical Structures of Nucleic Acids by Metal Ions and Small Molecules. Coord. Chem. Rev. 2013, 257 (19−20), 2765−2783. (4) Dembska, A. The Analytical and Biomedical Potential of Cytosine-Rich Oligonucleotides: A Review. Anal. Chim. Acta 2016, 930, 1−12. (5) Dong, Y.; Yang, Z.; Liu, D. DNA Nanotechnology Based on IMotif Structures. Acc. Chem. Res. 2014, 47 (6), 1853−1860. (6) Nguyen, T.; Fraire, C.; Sheardy, R. D. Linking pH, Temperature and K+ Concentration for DNA I-Motif Formation. J. Phys. Chem. B 2017, 121 (33), 7872−7877. (7) McKim, M.; Buxton, A.; Johnson, C.; Metz, A.; Sheardy, R. D. Loop Sequence Context Influences the Formation and Stability of the I-Motif for DNA Oligomers of Sequence (CCCXXX)4, Where X = A And/or T, under Slightly Acidic Conditions. J. Phys. Chem. B 2016, 120 (31), 7652−7661. (8) Mao, Y.; Liu, D.; Wang, S.; Luo, S.; Wang, W.; Yang, Y.; Ouyang, Q.; Jiang, L. Alternating-Electric-Field-Enhanced Reversible Switching of DNA Nanocontainers with pH. Nucleic Acids Res. 2007, 35 (5), 2− 9. (9) Johnson, R. P.; Richardson, J. A.; Brown, T.; Bartlett, P. N. A Label-Free, Electrochemical SERS-Based Assay for Detection of DNA Hybridization and Discrimination of Mutations. J. Am. Chem. Soc. 2012, 134 (34), 14099−14107. (10) Heaton, R. J.; Peterson, A. W.; Georgiadis, R. M. Electrostatic Surface Plasmon Resonance: Direct Electric Field-Induced Hybridization and Denaturation in Monolayer Nucleic Acid Films and LabelFree Discrimination of Base Mismatches. Proc. Natl. Acad. Sci. U. S. A. 2001, 98 (7), 3701−3704. (11) Mahajan, S.; Richardson, J.; Brown, T.; Bartlett, P. N. SERSMelting : A New Method for Discriminating Mutations in DNA Sequences. J. Am. Chem. Soc. 2008, 130 (28), 15589−15601. (12) Kaiser, W.; Rant, U. Conformations of End-Tethered DNA Molecules on Gold Surfaces: Influences of Applied Electric Potential, Electrolyte Screening, and Temperature. J. Am. Chem. Soc. 2010, 132 (23), 7935−7945. (13) Yu, H. Z.; Luo, C. Y.; Sankar, C. G.; Sen, D. Voltammetric Procedure for Examining DNA-Modified Surfaces: Quantitation, Cationic Binding Activity, and Electron-Transfer Kinetics. Anal. Chem. 2003, 75 (15), 3902−3907. (14) Keighley, S. D.; Li, P.; Estrela, P.; Migliorato, P. Optimization of DNA Immobilization on Gold Electrodes for Label-Free Detection by Electrochemical Impedance Spectroscopy. Biosens. Bioelectron. 2008, 23 (8), 1291−1297. (15) De Rache, A.; Doneux, T.; Buess-Herman, C. Electrochemical Discrimination between G-Quadruplex and Duplex DNA. Anal. Chem. 2014, 86 (16), 8057−8065. G

DOI: 10.1021/acs.langmuir.7b04088 Langmuir XXXX, XXX, XXX−XXX