Supramolecular multivalency effects enhance imine formation in aqueous medium allowing for dynamic modification of enzymatic activity

Imine formation under physiological conditions represents a challenging reaction due to the strong propensity of aldimines to be hydrolyzed. Herein we disclose the remarkable effect of supramolecular multivalency on increasing imine stability. A family of reactive aldehydes was synthesized bearing supramolecularly-active sites within their structure. The imine formation activity for such aldehydes was evaluated and compared with model aldehydes. The reaction of the best-performing species – containing two carboxylate groups-with a set of amines showed a significant decrease in imine yields as the degree of supramolecular multivalency between sidechains decreased. The reversible conjugation of amino acid derivatives and small peptides was also assayed, with excellent selectivities for the imine formation at the Nα position even in substrates containing competing sites. Preliminary results on protein bioconjugation revealed that a model enzyme could be dynamically inhibited upon reaction with the aldehyde, with its native activity being recovered by displacing the imine bonds with a suitable chemical effector (i.e., acylhydrazide).


Materials and Methods
All amines, aldehydes and components of buffer solutions were obtained from commercial suppliers and used without further purification.Carbonic anhydrase from bovine erythrocytes (CA) was obtained from SigmaAldrich.No unexpected or unusually high safety hazards were encountered.
A Mettler Toledo SevenCompact pH Meter S220 was used to monitor the pH of the solutions, adjusted with either NaOD or DCl solutions as appropriate.All reactions were performed at room temperature, unless otherwise noted.Imination reactions were carried out in 50 mM deuterated phosphate buffer pD 7.2 unless otherwise indicated.The yields and product abundances were determined by relative integration of the signals in the 500 MHz 1 H-NMR spectra (Bruker Ascend Spectroscope Advance Neo-500 MHz; 500 MHz for 1 H and 125 MHz for 13 C{ 1 H}).An example of the procedure is illustrated in Fig. S16 and S17 in the characterization section.First, the spectrum was referenced, phased, and the baseline was corrected.Afterwards, the signals of interest were accurately integrated to minimise the residual error, and finally, the numerical values are processed (e.g., integration of the bisimino species is divided by two) for calculating product abundances.The error in integration is of 5%.The identification of the bisimine species was also confirmed by HRMS.All the reactions were analysed after 10 min, 1 h, 24 h, and 48 h to confirm that the equilibrium state was reached.The equilibrium was attained in all reactions after 10 min, except for the condensations between T and A2/A3 in which 1 h was needed.Data analysis was performed using MestReNova (version 14.2.3) and OriginPro (version 9.8.0.200).HRMS-electro-spray ionization (HRMS-ESI) mass spectra were recorded using a ThermoFisher Exactive Plus EMR Orbitrap mass spectrometer.MALDI analyses were performed on a MALDI AUTOFLEX SPEED Bruker using para-Nitroaniline PNA as the matrix.UV-VIS spectroscopy measurements were performed on a JASCO V-670 UV-VIS spectrophotometer.CD spectroscopy measurements were performed on a J-1500 CD spectrophotometer.

Catalytic experiments and CA inhibition
A method was properly designed for monitoring the kinetic profiles of the assayed hydrolysis reactions.All catalytic transformations were performed in D2O (pD 7.2, 50 mM phosphate buffer) at 335 K, and the concentrations of p-NP were determined using a calibration curve (absorbance at 400 nm).In a typical experiment (inhibition of CA using A8 in a 1:1 molar ratio), 10 μL of a 0.1 mM solution of CA were diluted with 970 μL of D2O (pD = 7.2, PBS, 50 mM).To this solution, 10 μL of a 0.1 mM solution (D2O, pD = 7.2, PBS, 50 mM) of A8 were added and the mixture was equilibrated for 10 min.Subsequently, 10 μL of a 10 mM solution of p-NPA (CH3CN) were added and the kinetic profile for the hydrolysis reaction was monitored over 20 min.

Titration experiments
The concentration of CA was maintained constant during the consecutive additions of A8.For the NMR titration, 0.3 mM solution of CA in 0.4 mL of D2O (pD = 7.2, PBS, 50 mM) was treated with increasing amounts of a 1 mL stock solution (pD = 7.2, PBS, 50 mM) of A8 (10 mM) that also contained 0.3 mM of CA.A similar procedure was followed for the CD titration experiment, using

S3
in this case a 0.001 mM solution of CA and a 0.1 mM stock solution of A8 (containing 0.001 mM of CA), both in D2O (pD = 7.2, PBS, 50 mM).

DFT calculations
DFT calculations were performed using Gaussian 09 (revision B.01) at the b3lyp level of theory using the 6-31g(d,p) basis set. 1,2All free energies (kcal/mol) were calculated taking into account the energies of the corresponding reagents and the molecule of water (as its H3O + ion, considering that the amino components were calculated as their ammonium derivatives) released in the condensation reaction.The DFT models were calculated at room temperature and using water as the solvent (PCM). 33D representations were designed using the Mercury software.It must be noted that the free-energy values are only suitable for comparison between systems, as the energy of the compounds will depend on the protonation degree of each of the species, and thus it might change according to both pKa of the corresponding components and the pH of the medium.For all calculations, the highest degree of protonation was considered for the compounds to maximize the effect of possible interactions between charged groups.

Docking analyses
SwissDock is an open source web server-based molecular docking engine, and it can be accessed at http://www.swissdock.ch/.This webserver is developed and maintained by the Swiss Institute of Bioinformatics, Lausanne, Switzerland. 4The molecular structure of CA was download from PDB database (pdb: 1v9e).The molecular structure of A8 was modelized at the b3lyp level of theory using the 6-31g(d,p) basis set.Both structures were uploaded on the SwissDock web server, and their interaction was elucidated using a blind and accurate docking for screening all possible recognition sites.Binding energies for the different interaction modes were recorded.The top-ranked pose was selected (ΔG = -8.15kcal/mol) and it was represented using Chimera 1.16 software (Fig. S10).

Fig. S17 .Fig. S18. 1 HS17Fig. S20 .S19Fig. S23 .S21Fig. S27 .
Fig.S17.Selected example for the product abundance quantification using 1 H-NMR spectrum (500 MHz, D2O + phosphate buffer 50mM, pD 7.2, 295 K).The crude corresponds to the condensation reaction between A7 and T (5 mM A7 and 50 mM T) after 10 mins of equilibration.It must be noted that the integration of the (A7)2T imine CH signal (8.4 ppm) was divided by two for the quantification of this species since there are two equivalent iminr protons per molecule.

. Main text supporting Figures, Schemes and Tables Scheme S1. Synthetic
route for the herein proposed water-soluble reactive aldehydes, viz.