Hydrogen Bonding in a l-Glutamine-Based Polyamidoamino Acid and its pH-Dependent Self-Ordered Coil Conformation

This paper reports on synthesis, acid–base properties, and self-structuring in water of a chiral polyamidoamino acid, M-l-Gln, obtained from the polyaddition of N,N′-methylenebisacrylamide with l-glutamine, with the potential of establishing hydrogen bonds through its prim-amide pendants. The M-l-Gln showed pH-responsive circular dichroism spectra, revealing ordered conformations. Structuring was nearly insensitive to ionic strength but sensitive to denaturing agents. The NMR diffusion studies were consistent with a population of unimolecular nanoparticles thus excluding aggregation. The M-l-Gln had the highest molecular weight and hydrodynamic radius among all polyamidoamino acids described. Possibly, transient hydrogen bonds between l-glutamine molecules and M-l-Gln growing chains facilitated the polyaddition reaction. Theoretical modeling showed that M-l-Gln assumed pH-dependent self-ordered coil conformations with main chain transoid arrangements reminiscent of the protein hairpin motif owing to intramolecular dipole moments and hydrogen bonds. The latter were most numerous at the isoelectric point (pH 4.5), where they mainly involved even topologically distant main chain amide N–H and side chain amide C=O brought to proximity by structuring. Hydrogen bonds at pH 4.5 were also suggested by variable temperature NMR. The 2D NOESY experiments at pH 4.5 confirmed the formation of compact structures through the analysis of the main chain/side chain hydrogen contacts, in line with MD simulations.

Pages S1-S9 Figure S1-S6 Figure S1. 1 H-NMR spectrum of M-L-Gln recorded at pH 4.5 in: panel (a) 9:1 H2O:D2O and panel (b) D2O using a Brüker Avance III 400 MHz instrument. For the sake of clarity, the chemical shift assignments are also reported in Table S1. Figure S2. 13 C-NMR spectrum of M-L-Gln recorded in D2O at pH 4.5 using a Brüker Avance 400 MHz instrument. For the sake of clarity, the chemical shift assignments are also reported in Table S1.   Table S2 for M-L-Gln. Panel (a): experimental, simulated and β corrected titrations; panel (b): distribution of charged species. Determination of β parameters for -COOH and tert-amine of M-L-Gln referred to the 1 st experiment of Table S2. Panel (c): calculation of β values from Equation (S1); panel (d): trends of the β-corrected pKa values versus α according to Equation (S1).  Tables S1-S2 Table S1. Chemical shift assignments of 1 H and 13 C of M-L-Gln and diffusion coefficients obtained by DOSY experiments. Table S2. pKa Values of M-L-Gln from different experiments.

S2
NMR characterization Figure S1. 1 H-NMR spectrum of M-L-Gln recorded at pH 4.5 in: panel (a) 9:1 H2O:D2O and panel (b) D2O using a Brüker Avance III 400 MHz instrument. In the (1a) spectrum the solvent signal has been suppressed by excitation sculpting sequence. For the sake of clarity, the chemical shift assignments are also reported in Table S1.
*  Figure S2. 13 C-NMR spectrum of M-L-Gln recorded in D2O at pH 4.5 using a Brüker Avance 400 MHz instrument. For the sake of clarity, the chemical shift assignments are also reported in Table S1. Figure S3. 1 H, 13 C-HSQC NMR spectrum of M-L-Gln recorded in 9:1 H2O:D2O at pH 4.5 using a Brüker Avance III 400 MHz instrument. Color code: CH2 red and CH blue.

S4
Determination of pKa and β parameter values and speciation curves pKa determination. The pKa1 (-COOH) and pKa2 (tert-amine) values of the ionizable functions of M-L-Gln were equivalent to the pH values at the half-equivalence points in the respective buffer zone of interest. The half-equivalence points were estimated as the pH values where half of the titrant volumes between consecutive inflections were added. The inflection points were in turn determined by numerically calculating the second derivative of the pH versus volume curves ( Figure S4a). β parameter determination. The β parameters of the generalized Henderson-Hasselbalch equation (Equation (1) in the manuscript, here reported as Equation (S1)) were determined, for both pKa1 (-COOH) and pKa2 (tert-amine), from Equation (S1) as the slope of the pH versus -log((1-α)/α) curve ( Figure S4c). The points near inflections approached the validity limit of the logarithmic function and were not considered. Figure S4d shows the trends of the β-corrected pKa values versus α according to Equation (S1).

Simulation of the titration curves.
Simulated titration curves were obtained following the De Levie approach 1 in order to iteratively refine pKa and β values to achieve the best fitting of the experimental data.
 Initial conditions: where: Simulated titration curves ( Figure S4a) were obtained from Equations (S10) and (S11) by introducing the values of pKa1 and pKa2 in the respective buffer regions of interest, corrected for β1 and β2. Calculation were carried out considering and constant throughout the whole titration experiment and equal to 0.1 M. Concentration fractions α and pKa values were refined iteratively to achieve the best fitting to the experimental points.
Determination of speciation diagrams. Speciation diagrams ( Figure S4b) were obtained by plotting the concentration fractions of the different ionic species as a function of pH (Equations (S12a-S12c)): With D and y as previously described, and where the Ka1 and Ka2 values were corrected for β1 and β2. Figure S4. Titration and speciation curves referred to the 1 st experiment of Table S2  DOSY spectra were recorded in D2O at pH 4.5 using a Brüker Avance 600 MHz, following the standard Bruker sequence with pre-saturation during relaxation delay for water suppression. The diffusion coefficient, D, was determined from Equation (S13): where f(g) is the intensity as function of g, g the magnetic field gradient strength, I0 the initial intensity,  the gyromagnetic ratio 4.258 10 3 Hz/G, δ and Δ the delays, in particular thelittle delta value (s) and thebig delta value (ms), D the diffusion coefficient.
The D values for each proton are reported in Table S1.