Template effects of vesicles in dynamic covalent chemistry

Vesicle lipid bilayers have been employed as templates to modulate the product distribution in a dynamic covalent library of Michael adducts formed by mixing a Michael acceptor with thiols. In methanol solution, all possible Michael adducts were obtained in similar amounts. Addition of vesicles to the dynamic covalent library led to the formation of a single major product. The equilibrium constants for formation of the Michael adducts are similar for all of the thiols used in this experiment, and the effect of the vesicles on the composition of the library is attributed to the differential partitioning of the library members between the lipid bilayer and the aqueous solution. The results provide a quantitative approach for exploiting dynamic covalent chemistry within lipid bilayers.


Cyanoacetamide S1
Methyl cyanoacetate (0.79 ml, 9.0 mmol) and n-decylamine (1.54 ml, 7.5 mmol) were added together into a round bottom flask and stirred at room temperature. After 20 min, the precipitate was filtered and washed with cold diethyl ether then dried under vacuum. 25 ml dichloromethane were added, and the mixture was washed with 1N HCl, saturated bicarbonate and brine. The cyanoacetamide S1 product was obtained as a white solid  Figure S1 1 H NMR (400 MHz, 301 K, CDCl3) of cyanoacetamide S1. Figure S2 13 C NMR (100 MHz, 301 K, CDCl3) of cyanoacetamide S1.

Receptor 1
The cyanoacetamide S1 (150 mg, 0.67 mmol) and benzaldehyde (68 μL, 0.67 mmol) were added together to 15 mL ethanol solution in the presence of catalytic amount of piperidine. The reaction mixture was stirred for 12 h at 60°C, and the solvent was evaporated. The residue was purified by column chromatography on silica gel (CH2Cl2:P.Et. 1:1) to give 1 (159 mg, 75%) as a white solid.

LC-MS identification of 3a-3d Adducts
To 1.0 ml of a solution 0.15 mM of receptor 1 in methanol were added different aliquots of thiols 2a,2d from a methanol solution and they were allowed to equilibrate for 1 hour and the reaction was monitored by HPLC-ELSD and LC-MS. In the case of the vesicle systems, receptor 1 (0.15 mM) was embedded in a lipid bilayer (DOPC, 1.0 mM) at pH=7.2 in HEPES buffer. In this case, the thiols were added from a HEPES buffer solution at pH=7.2.
Scheme S1 Addition of thiols 2a-2d to 1 leads to formation of the corresponding Michael adducts 3a-3d.
The adducts 3a-3d formation in the vesicles system and in single phase was monitored by LC-MS and the MS trace with the corresponding MS peaks for each adducts are reported in Figure S8 in the case of vesicles. The adduct formed reported in Figure S8-S9 were obtained adding 1.0 eq of 2a, 1.8 eq. of 2b, 12 eq. of 2c and 15 eq. of 2d. The same solutions were also analyzed with HPLC-ELSD to allow further quantification in DCL formation experiments (see section S4.2). The adducts formation is reported in Figure S9.
The method employed for the different mixture analysis is reported in the general method paragraph.

HPLC-ELSD Analysis of the Dynamic Covalent Library
To quantify the formation of the different library components an ELSD detector was employed for the HPLC analysis and the method was developed as follow.
For each library component, it was calculated a detector response factor in methanol. Initially, a series of HPLC-ELSD spectra of compound 1 at known concentration and the corresponding detector response calibration curves of 1 were obtained. The results are showed in Figure S10. The slope obtained by this correlation represents the linear response factor of the detector to 1.

Figure S10
HPLC-ELSD spectra of compound 1 at different concentration (left) and corresponding response factor calibration curve, the response factor is given by the slope (right).
Then, it was possible to obtain the response factor for each adduct formed 3a-3d. To 1.0 ml of a solution 0.15 mM of receptor 1 in methanol were added different aliquots of thiols 2a,2d from a methanol solution and they were allowed to equilibrate for 1 hour . The reaction was monitored by HPLC-ELSD.
Since the concentration of 1 is linearly correlating with the ELSD integral area based on the calculated response factor, it was possible to obtain the concentration of each adduct by subtracting the concentration of the unreacted receptor 1 calculated after each thiol addition [1]n to the initial concentration of [1] After that, the response factor for each adduct was calculated by plotting the calculated concentration of the adduct [3a-3d] towards the integral area of the adduct formed.
As example the ELSD spectra for the formation of adduct 3b at different equivalent of thiol 2b added are shown in Figure S11. The response factor was calculated by plotting the calculated concentration of 3b towards the integral area of the adduct formed ( Figure S11). Adopting the same approach, the response factors for each adduct were obtained repeating the calibration twice.
From these data, it was possible to obtain the response factors for each adduct which are showed in Table S1 including their characteristic retention time. In addition, the relative response factors used for adducts integral correction are reported in the last column of Table S1 and they were obtained by dividing the response factor of each adduct for the response factor of 1. Retention times of each adducts 3a-3d and their relative response factors values were further employed for the identification and quantification of the library components in the complex mixtures in both studied systems.

S13
The inverse of the relative response factor were used in the mixed experiment. By multiplying the integral corresponding to each adduct for the (Relative R. F.) -1 we obtained the corrected area of each library component.
After that, the fraction of each library component was calculated by dividing the corrected area for the total area of the library components.

Dynamic Covalent Library Reversibility Assessment
General Procedure for DCL formation eq. for each thiol) and they were allowed to equilibrate for 1 hour and the reaction was monitored by HPLC-ELSD. After 1 hour 7.5 μL of the last thiol (20 mM) of the series were added and the reaction was monitored until a second equilibration was reached. In the case of vesicles, receptor 1 (0.15 mM) was embedded in a lipid bilayer (DOPC, 1mM) at pH=7.2 in HEPES buffer. In this case, the solution containing all the thiols was prepared in HEPES buffer at pH=7.2.
In Figure S12 is showed a schematic representation for one example. In this case, thiols 2a, 2b and 2c were added to receptor 1 when embedded in the vesicles and the system was allowed to equilibrate for 1 hour (1st step). Then thiol 2a was added to the system (2nd step). The fractions of the library members observed in this series of experiment are shown for the single phase system ( Figure S13) and in the vesicles system ( Figure   S14). Figure S12. a) Schematic representation of the reversibility experiment. Initially, thiols 2a, 2b and 2c were added to receptor 1 when embedded in the vesicles and the system was allowed to equilibrate for 1 hour (1st step). Then thiol 2a was added to the system (2nd step).    Figure S16b (the corresponding results obtained in methanol are shown in Figure S16a). The time course of equilibration in aqueous buffer with no vesicles present is reported in Figure S16c.

UV-Vis spectra of thiol addition in vesicles
Kinetics of thiols 2a-2d addition to conjugate acceptor 1 embedded in vesicles were monitored with UV-Vis spectroscopy. To vesicles formed by DOPC (1mM, HEPES buffer, pH=7.2) loaded with receptor 1 (15% loading, 0.15 mM) were added different equivalent of thiols 2a-2d in buffer from stock solutions at 20 mM concentration. UV Vis spectra of kinetic experiments related to the addition of 1.0 eq. of each thiol 2a-2d to the system is reported in Figure S17.

UV-Vis spectra of thiol addition in methanol
Kinetics of thiols 2a-2d addition to conjugate acceptor 1 in methanol solution were monitored with UV-Vis spectroscopy. To 1 (0.15 mM) were added different equivalent of thiols 2a-2d from a 60 mM methanol solution. UV Vis spectra of kinetic experiments related to the addition of 10 eq. of each thiol 2a-2d to the system is reported in Figure S18.

Kinetic profile of the adduct formation and binding constant determination
The formation of the adducts was monitored taking into account that the absorbance of 1 at 300 nm is quenched after addition of thiols, 1 therefore it was possible to calculate the concentration of adduct 3a-3d formed at the equilibrium considering the formula In which [1]0 is the initial concentration of the receptor embedded in the bilayer (0.15 mM), A0 is the absorbance at 300 nm of the system before the thiol addition, At is the absorbance of the system at 300 nm at the time when the equilibrium is reached, and Aves is the absorbance estimated from a linear fitting of UV-Vis spectra which consider only the absorbance of the vesicles solution at 300 nm. In the case of methanol, the term Aves becomes 0 due to the absence of absorbance of the adducts at 300 nm.
The kinetic trend for the formation of adducts 3a-3d at different equivalents of thiols added respect to 1 are reported using vesicles (Figure S19-S22) and in methanol solution ( Figure S23-S26). After the determination of adducts concentration at the equilibrium, it was possible to obtain the association constants (K) for the reversible adducts formation since the concentration in solution of all the species was known. K was calculated using the following formula In which the concentration of the adduct was calculated taking into account of the above mentioned formula, In Table S2-S5 are reported the concentration values for each species at the equilibrium when the adducts were formed in vesicles. In Table S6    30 eq. 2e 20 eq. 2e 10 eq. 2e 5.0 eq. 2e 1.0 eq. 2e