Thermodynamics and solvation dynamics of BIV TAR RNA–Tat peptide interaction

Abstract

The interaction of the trans-activation responsive (TAR) region of bovine immunodeficiency virus (BIV) RNA with the Tat peptide is known to play important role in viral replication. Despite being thoroughly studied through a structural point of view, the nature of binding between BIV TAR RNA and the BIV Tat peptide requires information related to its thermodynamics and the nature of hydration around the TAR–Tat complex. In this context, we carried out the thermodynamic study of binding of the Tat peptide to the BIV TAR RNA hairpin through different calorimetric and spectroscopic measurements. Fluorescence titration of 2-aminopurine tagged BIV TAR RNA with the Tat peptide gives their binding affinity. The isothermal titration calorimetric experiment reveals the enthalpy of binding between BIV TAR RNA and the Tat peptide to be largely exothermic with the value of −11.7 (SEM 0.2) kcal mol⁻¹. Solvation dynamics measurements of BIV TAR RNA having 2-AP located at the bulge region have been carried out in the absence and presence of the BIV Tat peptide using the time correlated single photon counting technique. The solvent cage around the Tat binding site of RNA appears to be more rigid in the presence of the Tat peptide as compared to the free RNA. The displacement of solvent and ions on RNA due to peptide binding influences the entropic contributions to the total binding energy.

---

Introduction

Bovine immunodeficiency virus (BIV) is a lentivirus which causes lymphocytosis, lymphadenopathy, progressive weakness, and central nervous system disorders in infected cattle.^1,2 Tat is one of the accessory proteins, among Rev, Vif, Vpy, Vpw, and Tmx, encoded by BIV which is essential for activation of transcription of viral RNA by binding to its trans-activation responsive (TAR) region.^3,4 This binding is essential for the replication of BIV and any interference with these interactions hampers the pathogenicity of BIV. As these interactions are the key determinants of the viral pathogenicity inside the host cell, thus small proteins serve as a nice model to study RNA–protein interaction and the results can easily be extrapolated to the full length RNA–protein complex. It is well known that the proteins containing the arginine rich motifs (ARMs) show high binding affinities towards their cognate RNA targets although they also show some non-specific binding to other RNA targets.^5–13 For example ARM from the BIV Tat protein (residues 65–81) shows high binding affinity towards BIV TAR RNA while it can also bind to human immunodeficiency virus HIV-1 TAR RNA non-specifically.^5–13 Understanding the molecular forces that govern the affinities and specificities of such interactions requires the integrated information related to the thermodynamic and hydration characteristics of the binding pockets. RNA–peptide interactions provide a simple and well-defined system to understand the principles involved in determining RNA recognition and thus could be utilized to develop improved drug/derivatives having better selectivity towards infection causing viral transcripts. There are many reports revealing the intricacies of the thermodynamic basis of complex formation of RNA and small molecules like aminoglycosides^14–16 and peptides.^17,18

The solution structure of BIV TAR RNA and the TAR–Tat complex is well characterized through NMR and molecular dynamics studies.^19,20 Tat binds with BIV TAR RNA in a β-hairpin loop structure in the major groove. The NMR structure reveals that the arginine rich motif (ARM) spanning from amino acids 65 to 81 in the BIV Tat protein can specifically bind to its cognate BIV TAR RNA.²¹ The U10 residue is located at the bulge region of the major groove in BIV TAR RNA which is found to possess high specificity of binding with the Tat peptide. The major structural change that takes place in RNA after binding to the peptide is the unstacking of the U10 residue, resulting in widening of the major groove. The BIV Tat peptide assumes a β hairpin structure in the complex.^19,20

Hydration of biomolecules such as DNA, RNA, and proteins plays an important role in their structure, conformation, and function. Several studies have revealed that there is significant amount of water molecules bound to the biomolecular surfaces which dictate their biological applications.^22–26 There have been many studies relating to structure and dynamics of hydration around DNA,^27–29 proteins,^23,30,31 and micelles,^32,33 however, the dynamics of hydration of RNA is still not well understood. The classical study on RNA hydration by Egli et al.³⁴ described the regular network of water molecules surrounding the RNA duplex based on the X-ray synchrotron diffraction data as systematic and well organized in both major and minor grooves. They found higher thermodynamic stability of RNA due to favorable enthalpy as compared to the entropy favored DNA duplex. There are several studies on RNA hydration using NMR, osmotic stress measurements and molecular dynamics simulations.^34–36 In the present study, we use the time resolved fluorescence technique to understand the dynamics of the hydration layer surrounding the BIV TAR RNA in the absence and presence of peptide binding. To our knowledge this is the first quantitative measurement of solvation dynamics of the RNA structure, however, there have been many studies on the solvation dynamics of DNA duplexes carried out up to femtosecond timescales. Comparative solvation behavior of DNA and RNA is, though, out of the scope of this study; our goal is to qualitatively establish the changes in dynamics of interfacial solvent molecules during formation of the RNA–peptide complex thus contributing to the better understanding of the RNA–peptide interactions.

The BIV TAR RNA–Tat system was chosen as a model system for this study as it is very well characterized in terms of its structure and functional aspects. A detailed analysis of the thermodynamic basis of their interactions would add to existing knowledge about these interactions. In spite of all the structural information available from solution structure and other biochemical studies, the thermodynamic and hydration basis of the recognition between BIV TAR RNA and the BIV Tat peptide is still not clear. Recently, Suryawanshi et al. have studied the thermodynamics and hydration changes occurring during HIV-1 TAR–Tat complex formation.¹⁷ It was demonstrated that the basic reason for binding interactions between the HIV-1 TAR RNA hairpin and the Tat peptide is the hydrogen bonding of arginine with the RNA bases. In another study, the binding of the Rev peptide to the HIV-1 RRE RNA is shown to be enthalpy driven as compared to the RSG-1.2 binding to the RRE RNA which is favored entropically.¹⁸ The present study is the next step towards a better understanding of the energetic basis of molecular interactions between RNA and peptides. Here we explore the thermodynamic and solvent dynamic parameters of BIV TAR RNA and Tat interactions through various calorimetric and spectroscopic measurements. UV melting and circular dichroism measurements are indicative of the structural changes occurring during the complex formation, while fluorescence titration experiments give the binding affinity. We also performed isothermal titration calorimetry to obtain the enthalpy and entropic energy contributions to the binding phenomenon. Steady state and time resolved fluorescence experiments have been performed to understand the changes in the stacking interactions as well as the solvent dynamics after TAR–Tat complex formation.

We also applied the time resolved fluorescence technique to understand the change in solvation dynamics around the binding site in BIV TAR RNA after binding to the Tat peptide. The results from quantum yield, stacking interactions and solvation dynamics measurements corroborate the entropy and enthalpy changes found in the BIV TAR–Tat complex. Moreover, this work can be considered as an extension of the dynamic profiling of RNA–peptide interactions. The significant difference in the thermodynamics of binding in the two model systems, HIV-1 TAR RNA–Tat¹⁷ and BIV TAR RNA–Tat, despite being structurally homologous has encouraged us to study the BIV TAR RNA–Tat peptide system in more detail.

Materials and methods

The HPLC grade 28 base BIV TAR RNA oligonucleotide was obtained from Sigma-Aldrich, Singapore. The 17-mer Tat peptide was obtained from Hysel India Pvt. Ltd, and was used without any further purification. The structure and sequence of RNA and peptide are shown in Fig. 1. The RNA concentration was measured by heating the oligonucleotide up to 85 °C and extrapolating the tangent of the upper baseline of absorbance at 260 nm up to 25 °C.³⁷ The molar extinction coefficient of BIV TAR RNA at 25 °C (262.8 × 10³ M⁻¹ cm⁻¹) was used for concentration determination. Fluorescence-labeled RNA containing 2-aminopurine (2-AP) incorporated at the 11^th position of the BIV TAR RNA from the 5′ end of the RNA strand was purchased from Sigma-Aldrich. All the reagents used here were commercially available analytical grade and were used without further purification. All reagents were prepared in Milli-Q autoclaved water. All the measurements were carried out in 10 mM sodium cacodylate buffer containing 0.1 mM EDTA and 100 mM NaCl salt at pH 7.5 (buffer A).

Fig. 1 (a) Sequence and secondary structure of BIV TAR RNA, numbers correspond to the nucleotide position in BIV mRNA. (b) Sequence of BIV Tat peptide, numbering corresponds to the positions in the intact protein.

Circular dichroism (CD) spectroscopy

CD spectra were recorded in a Jasco spectropolarimeter (model 715, Japan) equipped with a thermoelectrically controlled cell holder and a cuvette with a path length of 1 cm. The scan rate was 100 nm min⁻¹. The spectra were recorded between 200 and 325 nm at 25 °C with the samples containing 5 μM concentration of the RNA strand and 15 μM concentration of the Tat peptide in buffer A. The experiments were performed in triplicate with an accumulation of three and the standard error of mean (SEM) is reported in the results.

Temperature dependent UV spectroscopy (UV melting)

All UV experiments were carried out on a Cary 100 (Varian) spectrophotometer equipped with a thermoelectrically controlled cell holder. A quartz cell with a 1 cm path length was used for all the absorbance studies. The absorbance versus temperature profiles were measured at 260 nm. The samples were heated from 40 °C to 100 °C with a heating rate of 0.5 °C min⁻¹. The RNA concentration was taken to be 1 μM in the strand and the peptide concentration was kept at 3 μM (i.e. three times that of RNA concentration) in all the thermal denaturation studies. The RNA solutions were prepared in buffer A. The melting temperature (T_m) was determined as described previously.³⁸

Steady state fluorescence measurements

The fluorescence spectra of 2-aminopurine (2-AP) fluorescence-labeled BIV TAR RNA (BIV TAR-2AP) were measured using a Fluoromax 4 (Spex) spectrofluorometer equipped with a thermoelectrically-controlled cell holder (quartz cuvette 1 cm × 1 cm). Emission spectra were recorded over a wavelength range of 330 to 450 nm with the excitation wavelength kept at 310 nm. The excitation and emission slit widths were kept at 5 nm. The measurements were carried out in buffer A at 25 °C and pH 7.5. The fluorescence titration experiments were performed at 25 °C keeping the RNA concentration fixed and varying the peptide concentration to estimate the binding affinity between the RNA and the Tat peptide. The changes in the fluorescence intensity at 370 nm were monitored as a function of peptide concentration. At any given RNA/peptide concentration ratio, the observed fluorescence can be considered as the sum of weighted contributions from both free and peptide bound RNA, as described by the equation

F = (1 − α) F₀ + αF_b, (1)

where F is the observed fluorescence intensity at each titrant concentration; F₀ and F_b are the fluorescence intensities at initial and final states of titration; and α is the mole fraction of RNA in the bound form. Assuming 1 : 1 stoichiometry of binding, the association constant, K_a, between the RNA and the peptide is related to the total RNA concentration, [R]₀, and the added peptide concentration, [L]_t, through

α²[R]₀ − α([R]₀ + [L]_t + 1/K_a) + [L]_t = 0. (3)

Solving eqn (3),

Using eqn (1) and (4), we obtain

where

ΔF = F − F₀ (6)

and

ΔF_max = F_max − F₀. (7)

Fluorescence quantum yield values (ϕ) of 2-AP labeled RNA in its free and peptide bound form were determined using free r2-AP riboside as the standard at 25 °C. The buffer used for preparation of free 2-AP was taken similar to that of the sample solutions. The quantum yield of r2-AP is taken to be 0.68 at 25 °C as reported in the literature.³⁹

Isothermal titration calorimetry

The calorimetric titrations of the Tat peptide binding to the TAR RNA were carried out using a Microcal VP-ITC (MicroCal, Inc.; Northampton, MA) at 25 °C. The samples of TAR RNA and the Tat peptide were prepared in buffer A at pH 7.5. The isothermal sample chamber containing 1.5 mL of 10 μM RNA solution was titrated with 4 μL aliquots of 350 μM peptide through a 250 μL rotating syringe (350 rpm). The total number of injections were taken from 15–25 at different temperatures. The corresponding control experiment was also carried out in which 350 μM peptide solution was injected into the buffer solution. The initial delay prior to the first injection was kept as 200 s followed by 300 s delay before each injection. The duration of each injection was taken as 8.0 s. Each heat burst corresponds to a single injection of peptide to the RNA solution. The heat burst curve (μCal s⁻¹vs. s) thus obtained gives a measure of the heat generated (area under the heat curve) during each injection. The heat of dilution was subtracted from each value of heat associated with each injection to obtain the net enthalpy change for the TAR–Tat interaction.

Time resolved fluorescence measurements

Picosecond fluorescence decay measurements for BIV TAR-2AP in the free and bound form were carried out using the time-correlated single photon counting (TCSPC) method. A 310 nm nanosecond pulsed light emitting diode (EPLED) was used as the excitation source. The fluorescence signal was detected using photomultiplier tube (Hamamatsu R928P) (PMT) detectors. The excitation wavelength is taken as 310 nm for all the time-resolved spectral measurements. The instrument response function obtained to be 850 ps was detected at 310 nm excitation using a dilute solution of Ludox scatterer. The temperature was kept at 25 °C for all measurements. Decay curves were obtained from a collection of 5000 counts in the peak channel. Measurements were repeated three times.

Time-resolved fluorescence data were analyzed by a standard reconvolution procedure using nonlinear regression. The fluorescence intensity decay was fit to a sum of exponentials

Where the pre-exponential factors α_i are the amplitudes of each component, and τ_i are fluorescence lifetimes.

Time resolved emission spectra (TRES) were generated from the fluorescence decay curves collected at different wavelengths spanning the entire emission band of the probe. The curves were best fitted with a three-exponential decay function. The fitted decay curves, D(λ,t), were normalized with respect to the steady-state emission spectrum, S₀(λ), and were used to obtain the time resolved emission spectrum, S(λ,t), at a given time t using the expression⁴⁰

The TRES were used to obtain the spectral-shift correlation function C(t) using the method as described by Maroncelli and Fleming.⁴⁰ The time dependent fluorescence Stokes' shift obtained from TRES is used to construct the normalized solvent correlation function, C(t), defined as

where ν(0), ν(t) and ν(∞) are the frequencies of emission maxima (in cm⁻¹) at time zero, t, and ∞, respectively. The analysis of the solvation dynamics data was done by the procedure used by Maroncelli and Fleming.⁴⁰

The solvent correlation decay curves fitted using the biexponential function will give rise to the value of average solvation time 〈τ_s〉 estimated using the following relation

〈τ_s〉 = α_s1τ_s1 + α_s2τ_s2, (11)

where τ_s1 and τ_s2 are the solvation times with their respective contributions α_s1 and α_s2.

Results and discussion

Circular dichroism spectroscopy

Circular Dichroism (CD) spectroscopy measurements are carried out to detect any conformational changes occurring in the BIV TAR RNA after binding to the Tat peptide. Fig. 2 shows the CD spectra of BIV TAR RNA with and without binding to the Tat peptide. The BIV TAR RNA spectrum shows a strong peak at 265 nm, a strong negative peak at 210 nm and a weak negative peak at 240 nm, indicating the presence of the A-form. The intensity of the positive band at 265 nm slightly decreases with the difference of 4.34 (SEM 0.045) mdeg upon peptide binding indicating a decrease in the base stacking interactions. This result is consistent with the known solution structure of the TAR–Tat complex¹⁹ where the U10 residue is known to destack from the hairpin loop structure upon binding along with a minor distortion at the junction of the G11–C25 pair. The 240 and 210 nm negative bands become stronger with binding to the peptide indicating the stabilization of the A-form of RNA upon peptide binding. The average difference in the absolute value of the CD signal of BIV TAR RNA and the TAR–Tat complex was found to be 1.86 (SEM 0.070) and 4.11 (SEM 0.182) mdeg for the 240 and 210 nm wavelengths respectively. The difference in the CD signal between BIV TAR RNA and the TAR–Tat complex is very small. This result corroborates the solution structure of the TAR–Tat complex¹⁹ which suggests only minor distortions of the helical structure of the RNA upon binding to the peptide.

Fig. 2 Circular dichroism (CD) spectra of BIV TAR RNA (solid line) and the TAR–Tat complex (dashed line) at 25 °C. The RNA concentration is 5 μM and the peptide concentration is 15 μM in 10 mM sodium cacodylate buffer (pH 7.5) having 0.1 mM EDTA and 100 mM NaCl salt (buffer A).

UV melting measurements

The thermodynamic stability of secondary structures of BIV TAR RNA and the TAR–Tat complex is measured using thermal denaturation profiles of the samples. Fig. 3 shows the melting profiles of BIV TAR RNA and its complex with the Tat peptide in buffer A. The secondary structure of the RNA starts disrupting on heating due to the breakage of intra-helical hydrogen bonds resulting in lesser stacking interactions among bases. Thus the absorbance at 260 nm starts increasing with temperature. The absorbance profile eventually shows saturation at higher temperatures indicating a fully expanded RNA strand with no secondary structure remaining in the oligonucleotide. The midpoint of the melting profile (α fraction) called the melting temperature (T_m) is used to determine the stability of the oligonucleotide. The melting temperature of the RNA in the absence of peptide is obtained to be 78 °C. Upon binding to the peptide, the melting temperature of RNA increases to 87 °C, suggesting that the binding to the Tat peptide increases the thermal stability of the BIV TAR RNA. The higher stability of the TAR hairpin structure induced by Tat binding can be explained to be due to the formation of extra H-bonds and other non-covalent interactions between RNA and peptide. The extent of change in the T_m value correlates with the strength of stability,^15,16 which can indirectly be corroborated to the enthalpy changes occurring in the system during TAR RNA–Tat peptide binding.⁴¹

Fig. 3 UV melting profiles of BIV TAR RNA (square) and its complex with Tat peptide (circle) at 260 nm in the presence of buffer A as mentioned in the Fig. 2 caption. The RNA concentration is 1 μM and the peptide concentration is 3 μM.

Fluorescence titrations

The binding affinity, K_a, between BIV TAR RNA and the Tat peptide is determined by measuring the fluorescence spectrum of 2-aminopurine (2-AP) labeled BIV TAR RNA (BIV TAR-2AP) at varying concentrations of Tat peptide at 25 °C. The fluorescent base analogue of adenine i.e. 2-AP is incorporated at the 11^th position of the BIV TAR RNA where it is stacked between two adjacent bases (i.e. U10 and G11). The position of 2-AP is selected such that it is not engaged in any base-pairing interactions and thus does not induce significant perturbation in the bulge region of the RNA hairpin structure which is also the site for peptide binding. It is also ascertained that the presence of 2-AP does not affect the Tat binding affinity significantly. The Tat peptide is known to bind to the bulge region of the hairpin structure of BIV TAR RNA so it is expected to observe the changes in the fluorescent properties of the fluorophore due to binding effects. Fig. 4a presents the steady state fluorescence emission spectra of 28-mer BIV TAR-2AP in the absence and presence of the Tat peptide. The concentration of peptide is taken as 10 times that of the oligonucleotide concentration (0.1 μM). The fluorescence intensity of 2-AP located in BIV TAR RNA increases as the Tat peptide concentration increases indicating that the peptide is binding to the bulge region of BIV TAR RNA where 2-AP is located. An increase in the fluorescence intensity is expected as the non-radiative decay decreases upon binding and the radiative emission is enhanced and can also arise due to a decrease in stacking interactions and exposure of the fluorophore to the aqueous environment. The binding isotherm, shown in Fig. 4b, is obtained by plotting the change in the peak emission intensity (i.e. 370 nm) of 2-AP at 25 °C at varying concentrations of the Tat peptide. Analysis of the isotherm based on eqn (5) gives the binding affinity of 1.1 (SEM 0.1) × 10⁷ M⁻¹ between RNA and the Tat peptide in buffer A at 25 °C. Bayer et al.⁷ has reported the value of dissociation constant (1/K_a) for the BIV TAR RNA–Tat complex to be 60 nM¹³ using RNA gel shift assay. There are other reports that depict the dissociation constant of the TAR–Tat complex to be 1.3 and 8 nM based on mobility shift assays.⁹ Owing to the discrepancies in the literature value of K_d due to variations in the buffers and methods utilized, we measured the dissociation constant for the TAR–Tat system in buffer A through fluorescence titration experiments. The K_a value of 1.1 (SEM 0.1) × 10⁷ M⁻¹ obtained in our experiment conforms to the value of binding affinity reported by Bayer et al.⁷

Fig. 4 (a) Steady state fluorescence emission spectra of 28-mer BIV TAR RNA having 2-AP in the absence (solid line) and presence of the Tat peptide (dashed line) at 370 nm emission wavelength. The RNA concentration is 0.1 μM and the peptide concentration is 1 μM using the buffer A. (b) Fluorescence titration curve of BIV TAR-2AP obtained by plotting the change in fluorescence emission intensity (normalized with respect to the fluorescence intensity of free RNA, divided by total binding-induced change in fluorescence) at 370 nm at 25 °C at varying concentrations of the BIV Tat peptide.

Isothermal titration calorimetry

Isothermal titration calorimetry (ITC) was used to determine the enthalpy changes associated with the binding of the BIV Tat peptide to the BIV TAR RNA at 25 °C in buffer A (pH 7.5). Fig. 5 (top panel) shows the heat burst generated after addition of 350 mM BIV Tat peptide to the solution containing 10 mM BIV TAR RNA. Fig. 5 (bottom panel) shows the amount of heat released after each addition of peptide to the RNA solution plotted against the peptide to duplex molar ratio. Fitting the isotherm obtained in Fig. 5 (bottom panel) with the model for one set of binding sites^42,43 gives different binding parameters such as binding affinity, enthalpy and stoichiometry of the peptide binding to the RNA structure. The binding affinity, K_a, thus obtained is 9.8 (SEM 0.4) × 10⁶ M⁻¹ which slightly differs from the value obtained from fluorescence titrations of the BIV Tat peptide with BIV TAR RNA under similar buffer conditions. This discrepancy may arise due to difference in sensitivity of these techniques and different range of concentrations involved in the calculation of binding affinities. The stoichiometry of binding calculated from the fitting of the binding isotherm gives a value of n = 1 (SEM 0.03) which corroborates the results obtained from fluorescence titrations. The value of enthalpy change, ΔH = −11.7 (SEM 0.2) kcal mol⁻¹, obtained from the fitting results shows that binding of the BIV Tat peptide to BIV TAR RNA is exothermic in nature. The increase in the melting temperature due to binding of the Tat peptide to the BIV TAR RNA also suggests the stabilization of the complex as compared to free BIV TAR RNA. The value of Gibb's free energy of binding (ΔG) is calculated using the equation ΔG = −RT ln K_a and is obtained to be −9.5 (SEM 0.1) kcal mol⁻¹. Entropy changes upon binding can be calculated using the standard Gibb's equation, i.e. ΔG = ΔH − TΔS. The entropy change, TΔS, associated with this binding event is obtained to be a small but significantly unfavourable value of −2.2 (SEM 0.1) kcal mol⁻¹. Overall it may be concluded that the binding of BIV Tat with the BIV TAR RNA is largely enthalpy driven (due to the large favorable enthalpy value that overcomes the small unfavorable entropy value).

Fig. 5 Isothermal titration calorimetry (ITC) for binding of the BIV Tat peptide to the BIV TAR RNA using buffer A at 25 °C. The top plot is the baseline corrected experimental data. The bottom plot represents the molar heat released plotted against the peptide to RNA molar ratio.

The enthalpy change during the binding process is contributed by two opposing forces. Favorable enthalpy changes are associated with the hydrogen bond formation and van der Waals contacts, while the unfavorable enthalpy change is contributed through desolvation of polar groups.²⁹ Electrostatic contribution to the entropy changes can be studied by varying the salt concentrations in the buffer. In this case of the BIV TAR–Tat complex, we obtained a large favorable enthalpy change during binding suggesting formation of large number of hydrogen bonds and van der Waals interactions. The change in entropy during binding of Tat to TAR may originate from structural changes, conformationally labile bases, number of water molecules and ions released. The unfavorable entropy obtained here (−2.2 kcal mol⁻¹) could arise from the desolvation of the polar groups and release of Na⁺ ions around the binding pocket as well as decreased flexibility of BIV TAR RNA induced by the peptide-restricted conformational space as described for HIV-1 TAR RNA.⁴⁴

Tat peptide contains many positively charged arginine and lysine residues; the grand average of hydrophobicity is −2.19 which is obtained using http://web.expasy.org/protparam/html, suggesting the hydrophilic nature of the peptide. These highly charged amino acids have a tendency to be highly solvated in the aqueous solution. During complex formation few water molecules are removed from the interface of the RNA and peptide. The release of water molecules from the complex leads to increased and favourable entropy change. The remaining interfacial water molecules help in better packing of the RNA–peptide complex, and are stabilized due to the hydrogen bonded network between polar bases of RNA and charged amide groups of the peptide ligand, thus increasing the enthalpy of binding in such complexes. Our ITC results show a negative and unfavourable entropic change during the RNA–peptide complex formation, suggesting either the sequestration of water molecules around the RNA–peptide complex or more structured BIV TAR RNA and Tat peptide in the complex as compared to their free forms. NMR data^19,20 show that the Tat peptide assumes the hairpin structure while binding to the RNA as compared to its disordered native structure whereas the BIV TAR RNA undergo minimum structural organizations during Tat binding, thus forming a more structured TAR–Tat complex (unfavorable entropy). It is already known that the major groove of BIV TAR RNA expands a little to facilitate better penetration of the Tat peptide¹⁹ indicating an increase in entropy (favorable entropy). The presence of ligand inside the binding pocket of RNA leads to sequestration of water molecules that have a significantly lower entropy than bulk water. The increase in entropy due to release of water molecules into the bulk is insignificant as compared to the decreased entropy due to trapped water molecules between the ligand and the RNA binding pocket⁴⁵ leading to overall unfavourable entropy of binding. The unfavourable entropy of binding in the case of polyamphiphilic surfaces has also been suggested by a classical paper by Lemieux.⁴⁶ The desolvation of polar groups can be better understood through solvation dynamics studies discussed in the next section.

ITC experiments have been performed at different temperatures to calculate the change in heat capacity ΔC_p during the binding of peptide to RNA (data shown as Fig. S1, ESI†). At higher temperature the increase in the negative value of the enthalpy change suggests more favourable enthalpy of binding (Table S1, ESI†). The slope of enthalpy vs. temperature plot gives the total heat capacity change ΔC_p, which is obtained to be −363.0 (SEM 30.3) cal mol⁻¹ K⁻¹ (Fig. S2, ESI†). It is known that burial of non-polar surfaces results in the negative ΔC_p value whereas burial of polar surfaces results in the positive ΔC_p value.^47,48 With the measured negative sign of the ΔC_p value we can speculate the decrease in the solvent accessible surface area (SASA) for non-polar surfaces of the BIV TAR–Tat complex. The solution structure of the BIV TAR–Tat complex by Puglisi et al.¹⁹ also indicates that the methyl group of Thr makes direct hydrophobic contacts with the ribose ring of G22 and the side chain of Ile makes van der Waals contacts with the aromatic ring of the bulge nucleotide U10 thus showing burial of non-polar surfaces of RNA as well as peptide. For most of the RNA–protein/peptide interactions the ΔC_p value falls within the range of −100 to −550 cal mol⁻¹ K⁻¹.^49–51 The ΔC_p value of BIV TAR RNA–Tat peptide interactions lies in the range of the general RNA–peptide interaction limit.

Dynamics of BIV TAR RNA: time-resolved fluorescence measurements

The extent of increase in the steady state fluorescence intensity of 2-AP located in the BIV TAR RNA after binding to the Tat peptide can be quantified in terms of changes in the quantum yield (QY) values as well the stacking interactions. The QY of the free BIV TAR-2AP and Tat peptide bound BIV TAR-2AP (BIV TAR-2AP–Tat) are calculated using free riboside r2AP as the standard and their values are shown in Table 1. The QY of 2-AP is reported to be 0.68 at 25 °C in the literature.³⁹ Using the standard procedure, the QY of BIV TAR-2AP is obtained to be 0.10, while binding to the peptide leads to an increase in its QY up to 0.37. Pilch and coworkers⁵² have reported the QY of 2-AP located in 27-mer RNA fragment in the range of 0.03 to 0.07, which they calculated using quinine sulphate in 1 N H₂SO₄ as the standard. The QY of BIV TAR-2AP (0.10) is observed to be much higher in value than those reported by Pilch and coworkers⁵² (0.03 to 0.07) for small RNA oligonucleotides although the 2-AP is positioned at the bulge region of both 27-mer RNA (used by Pilch and coworkers) and 28-mer BIV TAR RNA. It is assumed that 2-AP is not engaged in any significant base pairing interactions in both the RNA models. The difference in the QY of 2-AP located in two RNA molecules may arise from differences in the extent of stacking interactions of 2-AP with the neighboring bases. The 2-AP in BIV TAR RNA is located at the 3-nucleotide bulge region and is bulged out of the hairpin loop structure and thus is significantly destacked in its native conformation as compared to the slight bulging of 2-AP in the 1-nucleotide bulge of the Pilch model of RNA structure. As a consequence of stronger bulging, 2-AP experiences more hydrophilic surroundings in BIV TAR RNA resulting in higher QY as compared to the RNA oligonucleotide used by Pilch and coworkers.

Table 1 Steady state and time resolved fluorescence parameters of 2-AP in BIV TAR RNA and its complex with Tat peptide at 25 °C

Sample	QY	fstacked	τ1/ns (α1)	τ2/ns (α2)	τ3/ns (α3)	τmean/ns	χ^2
QY is quantum yield, fstacked is fraction of stacked bases, τ1, τ2 and τ3 are lifetime components with their relative contributions as α1, α2, and α3, and τmean is the mean lifetime. All the calculated values were found to be within the 5% error range.
BIV TAR-2AP	0.10	0.67	1.04 (0.03)	3.85 (0.16)	7.66 (0.81)	6.85	1.024
BIV TAR 2AP–Tat complex	0.37	0.15	0.14 (0.01)	3.34 (0.10)	9.33 (0.89)	8.61	1.035

---

The extent of base stacking can be quantified using the expression, f_stacked = 1 − Φ_rel/τ_rel, where f_stacked is the fraction of intrahelical stacked base, Φ_rel = (Φ/Φ_r2AP) and τ_rel = (τ/τ_r2AP) are the QY and amplitude-weighted fluorescence lifetime of 2AP in BIV TAR-2AP and the BIV TAR-2AP–Tat complex relative to the free riboside (r2AP). The fluorescence lifetime of r2AP at 310 nm excitation and 370 nm emission wavelengths, is obtained to be 13.5 ns under the buffer. The fluorescence decay measurements of BIV TAR-2AP and the BIV TAR-2AP–Tat complex are best described as the sum of three exponentials and the decay parameters along with the mean lifetime values are listed in Table 1. The amplitude weighted mean lifetimes (τ_mean) of BIV TAR-2AP and the BIV TAR-2AP–Tat complex are obtained to be 6.85 ns and 8.61 ns, respectively. The mean lifetime value of 2-AP in oligonucleotides is mainly influenced by the stacking interactions as well as the exposure toward the solvent. An increase in the mean lifetime value of 2-AP in BIV TAR-2AP upon binding to the Tat peptide may result from a decrease in collisions of 2-AP with the solvent molecules or has effects from binding induced decrease in the stacking interactions with the neighboring bases.

Using the values of mean lifetime and the QY, the values of f_stacked for the BIV TAR-2AP and BIV TAR-2AP–Tat complex are obtained to be 0.67 and 0.15 respectively. The drastic decrease in the value of f_stacked clearly indicates the significant destacking of 2-AP upon peptide binding. It is interesting to note that the CD measurements also indicate the decrease in base stacking interactions in BIV TAR RNA upon Tat binding. The destacking is possible if 2-AP relatively stretches out of the loop structure of RNA upon binding with the peptide. Another factor that influences the fluorescence lifetime of 2-AP is the collision with the solvent molecules. To understand the nature of the solvent environment around the fluorophore, the solvent dynamics of BIV TAR-2AP in the absence and presence of the BIV Tat peptide is also studied.

Solvation dynamics

To understand the solvent behavior around the fluorophore, the solvation dynamics measurements have been carried out in 2-AP labeled BIV TAR RNA in the absence and presence of the BIV Tat peptide. The lifetime decay measurements of 2-AP located in BIV TAR RNA and the BIV–Tat complex at all the wavelengths spanning their steady state emission spectra show the presence of time dependent Stokes' shift. Fig. 6a and b show the fluorescence decay curves of BIV TAR-2AP and the BIV TAR-2AP–Tat complex at the blue end (340 nm), the red end (440 nm), and the intermediate wavelengths (350 and 370 nm) of their steady-state fluorescence spectra, respectively. The decay at the blue edge shows the faster decay component and that at the red edge of the steady-state spectrum shows the growth component. The wavelength dependent fluorescence decay of 2-AP in free and bound BIV TAR RNA as shown in Fig. 6a and b indicates the presence of time dependent Stokes' shift. The time resolved emission spectra (TRES) of the 2-AP located in BIV TAR RNA as well as the BIV TAR–Tat complex (shown in Fig. 7a and b) have been constructed using the steady state emission spectrum and the impulse function as described by Maroncelli and Fleming.⁴⁰ We calculated the TRES curves for the time spanning from 0.05 ns to 8 ns where we observe saturation in the spectral shift. The insets of Fig. 7a and b show an expanded view of the TRES curves to emphasize upon the observed time dependent spectral shift. The red shift in the peak maxima of the TRES curve with time is observed due to the effect of solvent relaxation during the excited state lifetime of the dye. The saturation of the TRES curve with time appears to reach faster in the case of the BIV TAR-2AP–Tat complex as compared to the free BIV TAR-2AP. Using the TRES data, the solvent correlation curve is constructed for BIV TAR-2AP and the BIV TAR-2AP–Tat complex as shown in Fig. 8. The C(t) function represents the temporal response of the relaxation of the solvent in response to the change in the dipole moment of the excited fluorophore. The solvation correlation curve describes the hydration environment around the 2-AP probe located in the BIV TAR RNA and TAR–Tat complex.

Fig. 6 Fluorescence decay curves of 2-AP in (a) BIV TAR-2AP and (b) BIV TAR-2AP–Tat complex at 340 nm (black), 350 nm (red), 370 nm (blue), and 440 nm (wine). The green line corresponds to the instrument response function (IRF).

Fig. 7 Time-resolved emission spectra (TRES) of (a) BIV TAR-2AP and (b) BIV TAR-2AP–Tat complex at 0.05 ns (black, square), 1 ns (red, circle), 3 ns (green, triangle), 6 ns (blue, diamond), and 8 ns (purple, star). The inset shows the expanded view of the graph to clearly emphasize upon the observed time dependent shift in the peak maxima of the spectra.

Fig. 8 Normalized spectral shift correlation function, C(t), for the probe 2-AP located in the BIV TAR RNA with (square) and without (circle) the BIV Tat peptide.

Table 2 shows the fitting parameters associated with the solvent correlation curve for BIV TAR-2AP and the BIV TAR-2AP–Tat complex as shown in Fig. 8. The solvation of BIV TAR-2AP shows a single exponential decay with the correlation time of 1.8 ns. It must be noticed that the time resolution of our TCSPC instrument corresponds to ∼850 ps that means we can measure the shortest possible time window up to 200 ps after the deconvolution analysis of the florescence decays, which implies that we are missing a significant amount of faster solvation components. In the presence of Tat peptide, the solvent correlation function of BIV TAR-2AP shows two exponential fitting with the values of τ_s1 and τ_s2 as 0.81 ns and 5.4 ns. In the present case, it can be noticed that the values of solvation correlation time (τ_s1 and τ_s2) are shorter than the mean lifetime value of the probe (τ_mean, Table 1) in both free and peptide bound BIV TAR-2AP. Thus, the value of emission maximum at infinite time, ν(∞), required to calculate the correlation C(t) curve, can be safely obtained from the steady-state fluorescence spectrum of the probe.

Table 2 Decay parameters (τ_s1 and τ_s2) of the solvation correlation function C(t) with their relative contributions (α₁ and α₂) for BIV TAR RNA and the TAR–Tat complex

Sample	τs1/ns (α1)	τs2/ns (α2)	τs/ns
The error associated with the solvation decay components is found to be 8%.
BIV TAR-2AP	1.8	—	1.8
BIV TAR-2AP–Tat complex	0.81 (0.33)	5.3 (0.67)	3.73

---

There have been many reports probing the solvation behavior of DNA using different techniques and different probes. Zewail and coworkers reported a biexponential hydration dynamics of the calf thymus DNA in the minor groove with femto-second resolution using the drug Hoechst 33258.²² The faster component of solvation with a time constant of 1.1–1.4 ps was ascribed to bulk water, while the slower component with a time constant around 20 ps was assigned to water molecules “ordered” at the DNA surface. Berg and coworkers²⁴ used three different techniques to monitor three time scales of solvation behavior in DNA, i.e. time-correlated single photon counting from 40 ps to 40 ns, fluorescence up-conversion from 1 to 150 ps, and transient absorption measurements from 40 fs to 120 ps to indicate the power-law kinetics of solvation dynamics using a 17-mer coumarin hybridized DNA fragment. The long solvation time constant (∼8.5 ns) observed in the genomic DNA using Hoescht-33258 has been assigned to the δ-relaxation of the DNA environment, which is attributed to the diffusion of counter ions along the DNA chain.⁵³ Although DNA hydration is extensively studied, the hydration characteristics of RNA are drastically ignored. The present study is an attempt to understand the hydration characteristics of a small RNA fragment with and without binding to its cognate peptide using the time-correlated single photon counting setup. Since, the lowest temporal resolution of our machine corresponds to ∼200 ps thus it is evident that we are missing out a significant amount of hydration characteristics of the solvent layer of RNA close to the probe. However the observed changes in the solvation time scales imply significant differences in the behavior of layers of solvent shells surrounding the probe located inside the RNA backbone structure.

In the case of BIV TAR RNA, the magnitude of 1.8 ns for the solvent correlation time should not correspond to the diffusion of counterions (as the value is significantly smaller than the 8.5 ns time constant observed in DNA⁵³). This value does not have contribution from the free ‘bulk like’ water molecules either since it is too slow as compared to the bulk water dynamics. Thus, we may assume that the observed solvation time (1.8 ns) of 2-AP in BIV TAR RNA arises possibly due to the hydration layer composed of water molecules hydrogen bonded to the RNA bases or due to the diffusion of water molecules between the bound hydration layer and bulk water.

The occurrence of bimodal solvation has been widely discussed in many cases of peptides, micellar interfacial layers as well as in DNA molecules and also among intermolecular interactions between proteins, DNA and micelles.^{22–25,30–33} The two types of solvent correlation terms are thought to arise due to mechanically trapped and thermodynamically bound water molecules in micelles^32,33 whereas in the case of peptide surfaces the source of two values of solvation correlation times is known to be the diffusion of water molecules between two shells of water molecules.³⁰ Zhong et al.³¹ also got two distinct water dynamics of ∼1–8 ps and 20–200 ps range corresponding to initial local relaxation and subsequent collective network restructuring around a protein surface. When the Tat peptide is bound to the internal loop structure of RNA containing the 2-AP probe we observe a bimodal solvation behavior corresponding to the solvation correlation time of 0.81 ns and 5.4 ns. The bimodal solvation correlation values suggest two kinds of water dynamics around the probe in the TAR–Tat complex. The τ_s1 value of 2-AP in bound RNA (0.81 ns) is lower than that of 2-AP in free BIV TAR RNA (1.8 ns), indicating the faster motion of one hydration layer around 2-AP in the presence of peptide compared to the free RNA. The solvation of free RNA is thought to arise either due to the hydration layer hydrogen bonded to the RNA bases or due to diffusion of bonded water molecules into the bulk water. The more dynamic hydration of peptide bound RNA may arise due to destacking of 2-AP out of the bulge portion of the RNA in the BIV TAR-2AP–Tat complex which is already validated by f_stacked results. Due to destacking, the probe experiences faster diffusion between the bound water layer and trapped water layer thus may give rise to a faster solvent correlation time of 0.81 ns. The slower solvation correlation time (5.4 ns) may arise due to the relatively restricted motion (sequestration) of the water molecules hydrogen bonded to the RNA in the presence of the peptide (i.e. thermodynamically bound water). The solvation of the probe 2-AP overall becomes slower when the Tat peptide binds to the BIV TAR RNA as is seen from the higher value of the average solvation correlation time of the probe in bound RNA as compared to the free RNA. It suggests that the hydration layer around the probe becomes more restricted after Tat binds to the BIV TAR RNA.

The average slower solvation behavior in the BIV TAR-2AP–Tat complex also implies an increase in the amount of the thermodynamically bound water as compared to the mechanically trapped water in the presence of peptide binding such that diffusion between the two kinds of water molecules becomes slow. It is interesting to correlate the observed unfavorable change in the value of entropy upon peptide binding (TΔS = −2.2 kcal mol⁻¹) as calculated from the isothermal titration calorimetry measurements with the disturbances occurring in the solvent structure around the probe upon Tat–TAR complex formation. The slower solvation in the Tat–TAR complex as compared to the free BIV TAR RNA suggests a more tightly bound solvent structure around the probe in bound RNA as compared to the free RNA. The negative value of entropy change also implies more restricted solvent environment around the probe in the case of bound RNA as compared to the free RNA. It may be possible that the water released (desolvation) during the binding process could be compensated by the tighter packing of the solvent structure around the TAR–Tat complex.

BIV TAR RNA–Tat and HIV-1 TAR RNA–Tat peptide are compared in the literature extensively due to their homologous structures.^53–58 In the present study we are performing thermodynamic analysis of the BIV TAR RNA–Tat peptide complex using UV melting, steady state fluorescence titrations, ITC experiments etc. as was done for the HIV-1 TAR RNA–Tat peptide system in our previous work.¹⁷ The QY, stacking interactions and solvation dynamic measurements corroborate the enthalpic and entropic finding of the thermodynamics of the BIV TAR RNA–Tat peptide interactions. The TAR region of BIV and HIV-1 transcripts are similar in structure except the 1-nt and 3-nt, respectively, sizes of their Tat binding bulge region. It is interesting to note that the binding of HIV-1 TAR RNA to its cognate Tat peptide is found to be both enthalpically and entropically favoured. The favourable enthalpy for HIV-1 TAR RNA–Tat interaction arises majorly from the hydrogen bonds while overcoming the unfavourable desolvation energy changes. The favourable entropy observed in the HIV-1 TAR–Tat system is due to the release of structured water and/or condensed counterions from the binding site of RNA upon peptide binding. In the case of the BIV TAR RNA–Tat system, we observed the Tat binding process to be enthalpically favorable and entropically unfavored. While the favourable enthalpy of the BIV TAR–Tat complex could arise from the strong hydrogen bond network between BIV TAR RNA and its cognate Tat peptide as seen in the HIV-1 TAR–Tat complex but the net unfavorable entropy suggests overcoming of favourable entropy due to release of water molecules by reorganization of water molecules/counterions around the BIV TAR RNA–Tat complex. The reason for the discrepancy between the entropy factor between the HIV-1 TAR RNA–Tat complex and the BIV TAR RNA–Tat complex may be due to difference in the size and flexibility of the bulge region of TAR RNA in both lentiviruses. A few studies have also reported that the larger and more flexible bulge region of HIV-1 and BIV TAR RNA leads to the increased binding affinity of Tat peptides.^54,55 A more detailed comparison between the structure and thermodynamics of their binding needs to be done to understand the better correlations of their structure, functions and the binding thermodynamics.

Conclusions

The thermodynamic and solvent dynamics profiling of the BIV Tat peptide binding to the BIV TAR RNA is carried out in this study. To understand the thermodynamic parameters of the peptide binding UV melting, CD spectroscopy, fluorescence titrations, and isothermal titration calorimetry measurements were performed. The UV melting study suggests that the RNA becomes more stable in the presence of the Tat peptide. CD measurements imply no deformation of the RNA structure in the presence of the peptide. The binding affinity between the BIV TAR RNA and the Tat peptide is obtained to be 1.1 (SEM 0.1) × 10⁷ M⁻¹ from the fluorescence titration experiments of RNA containing the 2-AP probe located at the bulge position. The enthalpy change during peptide binding is found to be highly exothermic with a value of −11.7 (SEM 0.2) kcal mol⁻¹. Using the binding affinity and the enthalpy values, the entropy change during the Tat binding process is calculated to be highly unfavorable with a negative value of −2.2 (SEM 0.1) kcal mol⁻¹. The solvation dynamics measurements reveal a single mode of solvent behavior around the binding pocket of the BIV TAR RNA while it shows bimodal solvent behavior of the Tat-bound TAR complex. Overall, the solvent behavior becomes slower for BIV TAR RNA in the presence of Tat, which corroborates the observed unfavorable entropy change during the peptide binding. This study would provide some additional information in a large data pool of molecular forces governing the RNA–peptide recognition and binding. Moreover, using such a database of molecular interactions and their origin, it would be easier to design a peptide as a potential drug having higher binding affinity for the target RNA. Molecular dynamics simulations for different RNA–peptide complexes would provide a complimentary approach along with calorimetric and spectroscopic methods towards understanding the relative contributions of various molecular forces governing these interactions.

Acknowledgements

This work was supported by Council of Scientific and Industrial Research (CSIR) (Project title: Comparative Genomics and Biology of non-coding RNA, grant number: NWP-0036), India. TG and SK acknowledge research fellowship from CSIR.

References

1. H. Egberink and M. C. Horzinek, Vet. Microbiol., 1992, 33, 311–331.
2. M. A. Gonda, D. G. Luther, S. E. Fong and G. J. Tobin, Virus Res., 1994, 32, 155–181.
3. J. Karn, M. J. Gait, M. J. Churcher, D. A. Mann, I. Mikaelian and C. Pritchard, Control of human immunodeficiency virus gene expression by the RNA-binding proteins Tat and Rev, in RNA-Protein Interactions, K. Nagai, I. W. Mattaj, IRL Press, New York, 1994.
4. A. D. Frankel, Using peptides to study RNA-protein recognition, in RNA-Protein Interactions, ed. K. Nagai and I. W. Mattaj, IRL Press, New York, 1994.
5. D. J. Kenan, C. C. Query and J. D. Keene, Trends Biochem. Sci., 1991, 16, 214–220.
6. I. W. Mattaj, Cell, 1993, 73, 837–840.
7. T. S. Bayer, L. N. Booth, S. M. Knudsen and A. D. Ellington, RNA, 2005, 11, 1848–1856.
8. F. Hamy, E. Felder, G. Heizmann, J. Lazdins, F. Aboulela, G. Varani, J. Karn and T. Klimkait, Proc. Natl. Acad. Sci. U. S. A., 1997, 94, 3548–3553.
9. C. A. Smith, V. Calabro and A. D. Frankel, Mol. Cell, 2000, 6, 1067–1076.
10. H.-Y. Mei, M. Cui, A. Heldsinger, S. Lemrow, J. Loo, K. Sannes-Lowery, L. Sharmeen and A. Czarnik, Biochemistry, 1998, 37, 14204–14212.
11. D. Lazinski, E. Grzadzelska and A. Das, Cell, 1989, 59, 207–218.
12. A. Davidsona, T. C. Leepera, Z. Athanassioua, K. P. Komisarskab, J. Karnc, J. A. Robinsonb and G. Varania, Proc. Natl. Acad. Sci. U. S. A., 2009, 106, 11931–11936.
13. D. E. Draper, J. Mol. Biol., 1999, 293, 255–270.
14. C. M. Barbieri, A. R. Srinivasan and D. S. Pilch, J. Am. Chem. Soc., 2004, 10, 14380–14388.
15. D. S. Pilch, M. Kaul, C. M. Barbieri and J. E. Kerrigan, Biopolymers, 2003, 70, 58–79.
16. M. Kaul and D. S. Pilch, Biochemistry, 2002, 41, 7695–7706.
17. H. Suryawanshi, H. Sabharwal and S. Maiti, J. Phys. Chem. B, 2010, 114, 11155–11163.
18. S. Kumar, D. Bose, H. Suryawanshi, H. Sabharwal, K. Mapa and S. Maiti, PLoS One, 2011, 6, e23300.
19. J. D. Puglisi, L. Chen, S. Blanchard and A. D. Frankel, Science, 1995, 270, 1200–1203.
20. X. M. Ye, R. A. Kumar and D. J. Patel, Chem. Biol., 1995, 2, 827–840.
21. L. Chen and A. D. Frankel, Biochemistry, 1994, 33, 2708–2715.
22. S. K. Pal, L. Zhao and A. H. Zewail, Proc. Natl. Acad. Sci. U. S. A., 2003, 100, 8113–8118.
23. S. K. Pal, J. Peon and A. H. Zewail, Proc. Natl. Acad. Sci. U. S. A., 2002, 99, 1763–1768.
24. D. Andretta, J. L. P. Lustres, S. A. Kovalenko, N. P. Ernsting, C. J. Murphy, R. S. Coleman and M. A. Berg, J. Am. Chem. Soc., 2005, 127, 7270–7271.
25. S. Pal, P. K. Maiti, B. Bagchi and J. T. Hynes, J. Phys. Chem. B, 2006, 110, 26396–26402.
26. V. Pande and L. Nilsson, Nucleic Acids Res., 2008, 36, 1508–1516.
27. T. Umehara, S. Kuwabara, S. Mashimo and S. Yagihara, Biopolymers, 1990, 30, 649–656.
28. E. Leipinsh, G. Otting and K. Wuthrich, Nucleic Acids Res., 1992, 20, 6549–6553.
29. I. Luque and E. Freire, Proteins: Struct., Funct., Genet., 2002, 49, 181–190.
30. N. Nandi and B. Bagchi, J. Phys. Chem. B, 1997, 101, 10954; N. Nandi and B. Bagchi, J. Phys. Chem. A, 1998, 102, 8217–8221.
31. L. Zhang, L. Wang, Y.-T. Kao, W. Qiu, Y. Yang, O. Okobiah and D. Zhong, Proc. Natl. Acad. Sci. U. S. A., 2007, 104, 18461–18466.
32. M. Kumbhakar, T. Goel, T. Mukherjee and H. Pal, J. Phys. Chem. B, 2005, 109, 18528–18534.
33. M. Kumbhakar, T. Goel, S. Nath, T. Mukherjee and H. Pal, J. Phys. Chem. B, 2006, 110, 25646–25655.
34. M. Egli, S. Portmann and N. Usman, Biochemistry, 1996, 35, 8489–8494.
35. E. Rozners and J. Moulder, Nucleic Acids Res., 2004, 32, 248–254.
36. C. M. Reyes, R. Nifosi, A. D. Frankel and P. A. Kollman, Biophys. J., 2001, 80, 2833–2842.
37. L. A. Marky, K. S. Blumenfeld, S. Kozlowski and K. J. Breslauer, Biopolymers, 1983, 22, 1247–1257.
38. L. A. Marky and K. J. Breslauer, Biopolymers, 1987, 26, 1601–1620.
39. D. C. Ward, E. Reich and L. Stryer, J. Biol. Chem., 1969, 244, 1228–1237.
40. M. Maroncelli and G. R. Fleming, J. Chem. Phys., 1987, 86, 6221–6239.
41. J. Petruska and M. F. Goodman, J. Biol. Chem., 1995, 270, 746–750.
42. T. Wiseman, S. Williston, J. F. Brandts and L. N. Lin, Anal. Biochem., 1989, 179, 131–137.
43. M. M. Pierce, C. S. Raman and B. T. Nall, Methods, 1999, 19, 213–221.
44. J. Lu, B. M. Kadakkuzha, L. Zhao, M. Fan, X. Qi and T. Xia, Biochemistry, 2011, 50, 5042–5057.
45. N. R. Syme, C. Dennis, A. Bronowska, G. C. Paesen and S. W. Homans, J. Am. Chem. Soc., 2010, 132, 8682–8689.
46. R. U. Lemieux, Acc. Chem. Res., 1996, 29, 373–380.
47. K. M. Armstrong and B. M. Baker, Biophys. J., 2007, 93, 597–609.
48. R. S. Spolar and M. T. Record Jr., Science, 1994, 263, 777–784.
49. J. B. Chaires, Biopolymers, 1997, 44, 201–215.
50. I. Haq, J. E. Ladbury, B. Z. Chowdhry, T. C. Jenkins and J. B. Chaires, J. Mol. Biol., 1997, 271, 244–257.
51. J. M. Sturtevant, Proc. Natl. Acad. Sci. U. S. A., 1997, 74, 2236–2240.
52. M. Kaul, C. M. Barbieri and D. S. Pilch, J. Am. Chem. Soc., 2004, 126, 3447–3453.
53. E. B. Brauns, M. L. Madaras, R. S. Coleman, C. J. Murphy and M. A. Berg, J. Am. Chem. Soc., 1999, 121, 11644–11649.
54. C. A. Smith, S. Crotty, Y. Harada and A. D. Frankel, Biochemistry, 1998, 37, 10808–10814.
55. B. Lustig, I. Bahar and R. L. Jernigan, Nucleic Acids Res., 1998, 26, 5212–5217.
56. B. Xie, M. A. Wainberg and A. D. Frankel, J. Virol., 2003, 77, 1984–1991.
57. H. P. Bogerd, H. L. Wiegand, P. D. Bieniasz and B. R. Cullen, J. Virol., 2000, 74, 4666–4671.
58. J. Srinivasan, F. Leclerc, W. Xu, A. D. Ellington and R. Cedergren, Folding Des., 1996, 1, 463–472.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c2mb25357g

---