Thermo-Programmed Synthetic DNA-Based Receptors

Herein, we present a generalizable and versatile strategy to engineer synthetic DNA ligand-binding devices that can be programmed to load and release a specific ligand at a defined temperature. We do so by re-engineering two model DNA-based receptors: a triplex-forming bivalent DNA-based receptor that recognizes a specific DNA sequence and an ATP-binding aptamer. The temperature at which these receptors load/release their ligands can be finely modulated by controlling the entropy associated with the linker connecting the two ligand-binding domains. The availability of a set of receptors with tunable and reversible temperature dependence allows achieving complex load/release behavior such as sustained ligand release over a wide temperature range. Similar programmable thermo-responsive synthetic ligand-binding devices can be of utility in applications such as drug delivery and production of smart materials.

3 entropic contribution (ΔS) for each variant (see Table SI3). Moreover, to precisely estimate the entropic contribution associated to the different poly(T) linkers, we have used the triplex-forming receptor with the shortest linker (i.e., four nucleotides) as our reference, and we have subtracted its entropy from the entropic values estimated for the other receptors.

Theoretical model for the triplex-forming DNA receptors
In this work we developed a theoretical model to predict the thermo-responsive properties of the triplex-forming DNA receptors ( Figure S3) using the enthalpic and entropic values (Table S3) estimated with the experimental procedure described in the previous section ( Figure 2C and Table S3). Below we give a brief description of the rationale behind this model.
The melting transition of the receptor-ligand complex can be described as: Where R is the receptor, L is the ligand, and K d is the dissociation constant. This equilibrium gradually shifts towards the released (unfolded) state when increasing the temperature. Exploiting the relationship between the equilibrium constant and the free energy (ΔG) we can describe the K d using the following equation: where ΔH is the enthalpy, ΔS is the entropy, T is the temperature, and R is the molar gas constant (R = 1,9872 cal·K -1 ·mol -1 ). Thus, exploiting ΔH and ΔS values experimentally obtained, the dissociation constant K d can be estimated as a function of the temperature. To do this, we consider the concentrations of the receptors and the ligand used in the melting experiments ( Figure S3) which are equimolar, leading to the following condition: Where C R is the concentration of the receptor, and C L is the ligand concentration. Exploiting this condition, we can describe the molar concentrations in function of the variable x defined as the fraction of the reacted ligand and receptor and the constant C: Using these conditions, we can express the K d as follow: Which can be converted to a second-degree equation and can be solved as follow: Then, substituting Eq. 2 into Eq. 8, we can describe the fraction of the reacted ligand and receptor in function of the thermodynamic values: Introducing in Eq. 9 the enthalpic and entropic values calculated through the melting experiments ( Figure 2C and Table S3) and the relative concentration (equimolar) of the receptor and the ligand, we estimated the value of the fraction x for each temperature which allow to simulate the melting curve for each receptor ( Figure S3)

Theoretical model and thermodynamic analysis for the ATP-binding aptamer
For the characterization of the ATP-binding aptamer variants, we have developed an ad-hoc theoretical model which is used to: 1) analyze the melting curves of the ATP aptamer receptors and precisely estimate the ΔH and ΔS values (Table S6); 2) predict the thermo-responsive properties of the aptamer receptors ( Figure S10). Below we give a brief description of the rationale behind this model.
The melting transition of the receptor-ligand complex can be described as: Where R is the aptamer receptor, L is the ATP ligand, and K d is the dissociation constant. This equilibrium gradually shifts towards the released (unfolded) state when increasing the temperature.
The K d can be described as function of the free energy (ΔG) as reported n Eq. 2 and can be estimated as a function of the temperature. To do this, we need to describe the molar concentrations as function of the variable y defined as the fraction of free receptor and the constants C R and C L , leading to the following conditions: Where C R is the concentration of the receptor, and C L is the ATP ligand concentration. Because the concentration of the ATP ligand is in large excess respect to the receptor, we can assume the molar concentration of the ligand as a constant. Using Eq. 11-13, we can describe the K d as function of the fraction of the free receptor as follow: (Eq. 13) This last equation can be related to the thermodynamics through Eq. 2, and we can describe the fraction of the free receptor in function of the thermodynamic values and temperature, as follow: (Eq. 14) The melting curves of ATP-binding receptors in the presence of excess ATP ligand ( Figure 4B) have been fitted to Eq. 14 by nonlinear least-squares fit method. This fitting procedure allows to estimate the enthalpy (H) and entropy (S) values associated to the loading and releasing process of ATP ligand for each aptamer receptor (Table S6). Since, as expected, the obtained H values were equal within the experimental errors (Table S6), to reduce the error of the estimated S values, we fitted again the experimental melting curves to the above equation by fixing H at its average value and leaving only S as an adjustable parameter (Table S6). The very good fit of the melting curves, calculated by Eq. 14 using the average H value and the S values from Table S6, to the experimental points is shown in Figure S10.

Ligand-release kinetics of DNA-based receptors
Ligand-release kinetics for synthetic DNA-based bivalent receptors have been carried out using two triplex-forming DNA receptors (4-nt and 60-nt) and the 11-nt DNA ligand 1 ( Figure 3A-B).
In the first step, we prepared two different stock solutions containing 10 µM of triplex-forming receptor and 1 µM of 11-nt DNA ligand in the working buffer (PBS buffer, 10 mM MgCl 2 , at pH 5.5). Before use, each solution was heated to 90°C for 5 min and then allowed to cool to room temperature for 1 h. Then, fresh working buffer has been added in a quartz cuvette to reach a final volume of 1 ml, and it has been heated at a specific temperature which has been selected to induce a controlled release of 11-nt DNA ligand 1. Specifically, for the 4-nt triplex-forming DNA receptor we selected 50°C, 55°C, 57°C, 58.5°C, 60°C and 70°C ( Figure 3A) while for the 60-nt triplexforming DNA receptor we used 40°C, 48°C, 50°C, 52°C, 55°C, 60°C and 70°C ( Figure 3B). We heated the cuvette for at least 10 min to ensure that the solution can properly reach the programmed temperature. After this, we started to collect continuously the fluorescence signal until we achieved a stable background signal, then, we added 10 µl of the stock solution containing the triplexforming DNA receptor and the labelled ligand into the cuvette (for a final concentration of triplex receptor of 100 nM and 10 nM of ligand), and we collected the fluorescence signal for 5 minutes until the signal reach a constant value. To estimate the kinetic constant of the release process at the different temperatures for the triplex-forming receptors, the obtained kinetic profiles have been fitted using the following mono exponential equation: where F 0 represent the raw fluorescence signal upon the addition of the oligonucleotides in the solution, b describes the growth/decay rate, k is the kinetic constant of the release process, and t is the time (min). In Table S4 are reported k values associated to each triplex-forming receptor.
Finally, the raw fluorescence signal has been converted to % ligand released by using the upper 8 and lower baselines estimated during the analysis of the melting curves ( Figure 2B).
Release kinetics for ATP-binding aptamers have been carried out using two variants (4-nt and 70nt; Figure S12). We followed the same procedure used for the triplex-forming receptors and described previously. Specifically, we prepared two different stock solutions containing 5 µM of ATP-binding aptamer and 30 mM of ATP, and we used as working buffer 100 mM Tris HCl and 10 mM MgCl 2 , 3 mM ATP at pH 6.5. For the 4-nt ATP-binding aptamer we selected as temperatures 30°C, 40°C, 55°C, 57°C, 59°C, 62°C, 65°C and 70°C ( Figure S7) while for the 70nt ATP-binding aptamer we used 30°C, 36°C, 40°C, 43°C, 45°C, 47°C, 59°C, and 70°C ( Figure   S12). Once the solution in the cuvette has reached the programmed temperature, we added 10 µl of the stock solution containing the labelled ATP-binding aptamer and the ATP into the cuvette (for a final concentration of aptamer of 500 nM and 3 mM of ATP ligand). Again, to estimate kinetic constants from kinetic profiles of ATP and the conversion of the raw fluorescence to % ligand release we used the same approach described previously. In Table S7 are reported kt (min -1 ) values associated to ATP binding aptamer variants.

4 Cyclic temperature-jump experiments
Cyclic temperature-jump experiments for triplex-forming DNA receptors ( Figure 3C programmed to execute two different temperature cycles. In the first one, the temperature was held for 30 seconds at 25°C before the fluorescence signal was recorded. After this step, the temperature was increased to 55°C (with a ramp of ⁓0.8°C/s), held for 30 seconds before the fluorescence signal was recorded. Then, the temperature was decreased to the initial one (with a ramp of ⁓0.5°C/s), held for 30 seconds and the fluorescence signal recorded. This was repeated for ten complete cycles. In the second cycle, the temperature was held for 30 seconds at 55°C and next increased to 65 °C. Again, this temperature gradient was repeated back and forward for ten complete cycles. Finally, the raw fluorescence signal was converted to % ligand release by using the upper and lower baselines, estimated through analysis of the melting curves performed in the PCR instrument using the two triplex-forming receptors with the 11-nt DNA ligand 1, and Eq. 1 in the main text.
To perform load/release cycle experiments for ATP-biding aptamers ( Figure 4E) we used the same 12    We tested two receptors (100 nM) with a loop of 4-nt (red curve) and 60-nt (blue curve) and a ligand of 11 mer (10 nM).   Figure 2C).          (Table S6), we precisely estimated the entropic content for each receptor finding a linear correlation between T50% values and the obtained entropy values ( Figure 4C).