Mechanical characterization of base analogue modified nucleic acids by force spectroscopy†

We use mechanical unfolding of single DNA hairpins with modified bases to accurately assess intra- and intermolecular forces in nucleic acids. As expected, the modification stabilizes the hybridized hairpin, but we also observe intriguing stacking interactions in the unfolded hairpin. Our study highlights the benefit of using base-modified nucleic acids in force-spectroscopy.


Optical Tweezers and Force-Distance Experiments
The experiments are performed using an in-house built optical tweezers instrument described in detail in Bosaeus et al. 2 Two counter-propagating 150 mW, 845 diode lasers were used to form a single trap within the microfluidic chamber mounted on a motorized stage. The fluidic chamber is divided into three channels; two of them are used to dispense the two different types of coated beads and the third channel contains the trap and the micropipette. The optical trap is used to capture one of the polystyrene coated beads and the other coated bead is immobilized by suction at the tip of the micropipette. Optical fibers are used to guide the laser beams and the position of the trap is measured by redirecting 5% of the light intensity onto a position sensitive detector (PSD). The remaining light is focused through a water-immersion objective lens (60X, NA 1.20) to form the trap. The light exiting the trap is collected by an identical objective lens and redirected to a PSD and photodiode to measure the forces acting on the trapped bead in all three dimensions. A quarterwave plates and polarizing beam splitters are used to redirect the light exiting the condensing objective lens to the force detector by turning the polarization of light 90⁰ relative to the light entering the objective and monitored individually. The forces are measured based on the conservation of light momentum. [2][3][4] The single-molecule mechanical (un)folding experiments were performed by attaching the biotinylated handle on the DNA hairpin to a streptavidin-coated bead and the digoxigenin handle to the anti-digoxigenin-coated bead. The streptavidincoated bead was held on the micropipette and the anti-digoxigenin bead was trapped in the optical trap, as shown in Fig.  1B. A tether containing a single DNA hairpin was established between the two beads, the force-distance experiments were performed by moving the trap at a constant velocity (100 nm/s) in a specific force range to follow unfolding and folding of the hairpins recurrently. Typical force-distance curves are shown in Fig. 1C and Fig. S1 and (un)folding forces are observed as sudden jumps in force and extracted using custom-made MATLAB programs. Data were recorded at a frequency of 1 kHz. All the experiments were carried out at a constant temperature of 23.5±1⁰C. Measurements were performed in a buffer containing 10mM Tris pH 7.4, 1mM EDTA and using two different NaCl salt concentrations of (1M and 50mM).

Free-Energy Calculation
Single-molecule force measurements can be used to determine the free-energy difference (ΔG 0 ) between the folded and unfolded states. Crook fluctuations theorem (CFT) 5,6 can be used to extract equilibrium free energies from non-equilibrium processes, such as mechanical (un)folding. The work done during unfolding and folding of the DNA hairpins was calculated by integrating the area below the force-distance curve (as shown in Fig. S4A), where the force transition occurs (λ 0 and λ 1 ) for each cycle. The calculated work done during unfolding (W) and recovered during folding (-W) was used to construct probability distributions. The crossing point of two work distributions is equal to the free energy differences (ΔG 0 ) as shown in Figure S4. The ΔG 0 measured contains contributions from the hairpin under study, the handles and the bead in the optical trap. 7,8 The free energy of formation of the hairpin (ΔG) can be estimated using the following equation, The terms ΔW st and ΔW handles are the work needed to stretch the hairpin and the handles, respectively. ΔW bead is the work needed to displace the bead in the optical trap. The work needed to stretch the molecule (ΔW st ) was obtained by subtracting the work needed to stretch the unfolded ssDNA using the inextensible worm-like chain (WLC) model 5 and the work needed to orient the hairpin. The hairpin is modelled as a freely jointed chain with a Kuhn and monomer length equal to the B-DNA hairpin diameter d=2.0 nm. The contribution from the handles (ΔW handles ) and the beads (ΔW bead ) are combined using the measured effective stiffness, assuming it to be constant in the range of forces (f 0 <f<f 1 ) studied. 5 The free energy of formation of the DNA hairpin (ΔG) was obtained using equation 1. The free energy calculations were done using custom-made MATLAB programs.

Thermal Melting and Circular Dichroism (CD) Measurements
Melting studies on the DNA hairpins (without handles) were performed using a Varian Cary 4000 spectrophotometer equipped with a programmable multicell temperature block. Hairpin concentrations were set to 2 µM using the absorbance at 260 nm in a buffer containing 10mM Tris pH 7.4, 1mM EDTA and two different NaCl salt concentrations of (5mM and 50mM). The absorption spectra were measured using a spectral bandwidth (SBW) of 2 nm and signal averaging time of 2 s. The samples were heated from 25⁰C to 95⁰C at a ramp rate of 1⁰C/min. The temperature was held at 95⁰C for 5 min and cooled to 5⁰C at the same rate. Absorption at 260 nm was measured with a temperature interval of 1⁰C for two consecutive cycles of heating and cooling (as shown in Fig. S2B). The melting temperatures (T m ) of the hairpins were determined as the maximum of the first derivative of the melting curves. The alpha curves (fraction of single strands in duplex state) were obtained by selecting the lower and upper linear baseline and calculation the fraction by normalizing fractional absorbance to the total absorbance change over entire temperature range. 9 The alpha curves were fitted with Boltzmann sigmoidal fitting function in the Origin software to check if the melting curves fit with a two-state model. Circular dichroism (CD) spectra were recorded for the hairpins at a concentration of 6 µM at 20⁰C using a Chirascan CD spectrometer (Applied Photophysics). The spectra were recorded from 210 to 475 nm at a SBW of 1 nm and averaged over three scans and then background corrected using a blank sample. The scan rate was set to 0.2 s per point with a step size of 1 nm.

Estimation of Thermodynamic Parameters
Thermodynamic parameters like enthalpy (ΔH) and entropy (ΔS) can be computed by combining the thermal measurements and the mechanical measurements at the same salt concentration (50mM NaCl). The free energy of formation (ΔG) between the two states was calculated as: ΔG is the free energy of formation calculated using CFT from the single-molecule (un)folding measurements at 23.5±1⁰C (experimental temperature, T). The melting temperature (T m ) is the temperature at which the unfolded and the folded state exist at equal probability (no difference in energy between the two states). So, equation 1 can at T m be written as, The entropy is calculated by combing equation 2 and 3 and it is then used to calculate the enthalpy using equation 2 or 3.
The enthalpies and entropies computed for the DNA hairpins (unmodified, 1-tC and 2-tC(stack) are shown in Table 2.
The enthalpy and entropy for the other hairpins (2-tC, 3-tC, Abasic, 2-tC(opp) can be estimated by computing ΔG at 50mM NaCl from ΔG at 1M NaCl values and combing with T m from measurements at 50mM NaCl. The free energy of formation exhibits a simple linear logarithmic dependence with salt concentration 10 as given by: where ΔG 0 is the free energy at salt concentration C (50mM), ΔG is the free energy at the reference condition (1M), and m (kcal mol -1 ) is a correction factor (0.11*number of bp). We assume that the contribution from the 4-nucleotide loop in the salt dependence to the overall free-energy of the structure is negligible, and we assume that the tC modifications have the same salt dependence as any of the other Watson-Crick bases. The entropy and enthalpy estimated after salt correction for the DNA hairpins are shown in Table S4.

Generalized Linear Mixed Models
Statistical significance of (un)folding forces obtained from unmodified and tC incorporated hairpins were analyzed using the generalized linear mixed model (GLMM) to mainly consider the random (stochastic) effects present in the experiments (Table S2 and Table S3). With a bead pair, multiple cycles of stretching and relaxing the hairpin were performed, yielding different (un)folding forces for each cycle. Subsequently, we change to different bead pairs and acquired additional data. The forces obtained within a set of bead pairs is dependent, whereas the forces between different sets of bead pairs are independent. The experiments were neither completely independent nor dependent, so to account for the heterogenous data, we used GLMM and tested the significance using mixed function of the apex package in the R software. 11,12              : Salt correction and estimation of enthalpy and entropy. a Free energy calculated using CFT from force-distance experiments at 1M NaCl. b Free energy at 5mM NaCl computed from 1M NaCl using salt correction represented as mean±standard error of mean. c Melting temperatures (T m , 5mM NaCl) reported as mean±standard deviation. d T m for 3-tC was obtained from measurement using a different instrument (old instrument broken) in which the T m for the unmodified was measured to be 71.2⁰C e Enthalpy and entropy calculated by combining ΔG 0 (5mM NaCl) and T m (5mM NaCl) and reported as mean±standard error of mean. The values reported in the parentheses implies to the error associated with the last significant figure.