Binding Between Cyclohexanohemicucurbit[n]urils and Polar Organic Guests

Inherently chiral, barrel-shaped, macrocyclic hosts such as cyclohexanohemicucurbit[n]urils (cycHC[n]) bind zinc porphyrins and trifluoroacetic acid externally in halogenated solvents. In the current study, we tested a set of eighteen organic guests with various functional groups and polarity, namely, thiophenols, phenols, and carboxylic and sulfonic acids, to identify a preference toward hydrogen bond–donating molecules for homologous cycHC[6] and cycHC[8]. Guests were characterized by Hirshfeld partial charges on acidic hydrogens and their binding by 1H and 19F NMR titrations. Evaluation of association constants revealed the complexity of the system and indirectly proved an external binding with stoichiometry over 2:1 for both homologs. It was found that overall binding strength is influenced by the stoichiometry of the formed complexes, the partial atomic charge on the hydrogen atom of the hydrogen bond donor, and the bulkiness of the guest. Additionally, a study on the formation of complexes with halogen anions (Cl− and Br−) in methanol and chloroform, analyzed by 1H NMR, did not confirm complexation. The current study widens the scope of potential applications for host molecules by demonstrating the formation of hydrogen-bonded complexes with multisite hydrogen bond acceptors such as cycHC[6] and cycHC[8].

As an internal reference for the fluorine signal position in 9 F NMR of 9, we added hexafluorobenzene 19; however, it appeared that presence of reference inside the samples was not necessary as we did not observed fluctuations in reference  between individual spectra.

Anion binding with cycHC[6]
Binding of anions to cycHC[6] was tested in CD3OD-d4 (0.8 mM cycHC[6]) and CDCl3 (1.2 mM cycHC[6]) by addition of salt excess to the macrocycle solution. Tetrabutylammonium (TBA) chloride and bromide have been added as a solid compound. Specific excess of salt was determined from integration of NMR signals against known concentration of macrocycle. 1 H NMR was measured shortly before and after a salt addition and then after 18 h. Dissolution of weakly soluble cycHC[6] in methanol was achieved by employing repeatedly heating and sonification.

Screening of potential guests in chloroform
Chemical shifts changes of cycHC[8] (ca 2.5 mM) proton signals induced by addition of 0.5, 5 and 40 equiv of guests (1-9) were investigated by 1 H NMR in CDCl3. Guests were added as a solutions of known concentration (typically 300-400 mM).

Test of hydrogen bond formation
1 H NMR of pentafluorophenol (50 mM) in CDCl3 was measured before and after the addition of solid cycHC[8] (20 mg), which provided solution containing roughly 0.5 eq of macrocycle.
The titration data were fitted using NumPy (1.10.2) and SciPy (0.18.1) libraries of Python 3. The script allows for simultaneous fitting of several datasets, which significantly improves stability of the fit results. This is essential in the case of 1:3 and 1:4 binding calculations. The leastsq function (implementation of the Levenberg-Marquadt algorithm) was used to determine the parameter set including the association constants K1, K2, and K3 that minimizes 2 = ∑[ ( ) − ] 2 , where ( ) is the theoretical value of the traced quantity (chemical shift) for a given pair of the host and guest concentration = ( ℎ 0, , 0, ); is the corresponding experimental value. The concentrations of the free host and guest molecules as well as their complexes are calculated Supplementary Material S4 numerically (function scipy.fsolve), using the definition of the binding constants and the mass balance equations.

2
Structures of hosts and guest used in the study Scheme 1. Structures of macrocyclic hosts and small guest molecules used in the study.

4
Screening of potential guests (1-9) in chloroform Figure S3. 1 H NMR spectra of free cycHC[8] (2.5 mM) in CDCl3 and in the presence of guestst 1-9 (40 equiv). Rest of the macrocycle's signals did not shifted or was covered with signals of guest. Additional signal in spectra with guest 8 corresponds to -OH group proton. S9 5 NMR titrations, Job plots and the data evaluation

Cooperativity evaluation criterions for cycHC[n]
Let's define a stepwise association constant Ki for an interaction between a substrate (host) M having m identical, independent binding sites and a ligand (guest) L having 1 binding site (Ercolani, 2003).
Evaluation of stepwise association constants has to take in count the number of equal binding sites on same molecule of host (Ercolani, 2003;Thordarson, 2011). It can be derived what should be the ratio of stepwise association constants in systems with no cooperativity and this ratio can be used as a criterion to asses cooperativity. For a two-step binding it is generally know that non-cooperative system exhibits ratio of stepwise constants according equation K1 = 4 K2.
Here is a general equation for ratio between any two consecutive Ki in system MLm: If the ratio Ki+1/Ki is satisfied with determined association constants then the system is noncooperative or statistical. If the ratio value is higher than it would be for statistical system then we speak about positive cooperativity. If the ratio value is lower than it would be for statistical system then we speak about negative cooperativity.

Normalization of titration curves
For the normalization of titration curves was used following formula: Nx is normalized value of chemical shift at x equivalents of cycHC[n] x equiv is chemical shift of guest at x equivalents of cycHC[n] 0 equiv is chemical shift of guest at 0 equivalents of cycHC[n] y equiv is chemical shift of guest at X equivalents of cycHC[n], which is point for which we normalize a titration curve

NMR titration data for association constants evaluation
Following NMR data for trifluoroacetic acid 16 (Table S13-S17) and methanesulfonic acid 18 (Table S19-S22) were used for qualitative comparison between binding of trifluoroacetic acid 16 with cycHC[6] and cycHC[8] and for qualitative comparison between binding of cycHC[6] with guest 16 and 18. All of these data for 16 and 18 were also used for evaluation of first three apparent stepwise association constants with our 3:1 binding model (Ustrnul et al., 2019). Following constrains were used to improve a chance for a binding model to converge: (i) K1obs > K2obs > K3obs and (ii) and HG1, HG2, HG3 ≥ EXPmax − 5(EXPmax EXPmin); to prevent the chemical shift  of complexed guest (HGx) from diverging extremely from experimentally observed values.

Note:
Even if we would be adding a guest to the macrocycle, we could not easily distinguish between complexes of different stoichiometry, which would eventually lead to large errors in evaluating stepwise Ka. Moreover, the number of variables in the evaluation process, which has to be fitted simultaneously, grows exponentially with the increasing stoichiometry, hence obtaining reliable values of all six or eight stepwise association constants for cycHC[6] or cycHC[8] is very difficult. Determination of all Ka would require excessive amount of experimental data from wide range of concentrations, which can be often impossible to do, for example due to a limited solubility of host or guest.              . Results corresponds to data obtained at 2 mM concentration of 16 ( Figure S5).  Table S13, Table S15, Table S16, Table S17 are assigned with circles, the dotted lines are shown to guide the eye. Graph illustrates the change in the shape of titration curves related to roughly one order of magnitude difference in concentration of titrated compound (here 16). Significant difference is caused by different proportion of formed complex as a natural consequence of constant value of binding constants at all concentrations.    Figure S8. NMR titration data for trifluoroacetic acid 16 (2 mM, green) and methanesulfonic acid 18 (ca 0.4 mM, purple) in the presence of growing concentration of cycHC[6] normalized at 9 equiv. Experimental datapoints from Table S13, Table S19 are assigned with circles, the dotted lines are shown to guide the eye. Although, titration curve of 16 is more pronounced than curve of 18 the concentration of 18 is roughly five-times lower, which has strong impact on the titration curve shape (proportion of formed complex as seen in Figure S7). Hence, we can speculate that cycHC[6] binds 18 similarly or rather stronger than 16.

Evaluation of association constants with 1:1 and 2:1 binding model in Bindfit
Titration data for guests 6, 9, 10, 11, 12, 13, 14, 15 and 17 was not possible to evaluate with 3:1 binding model due to their weak binding strength, therefore, we had to restrict our binding model to fit only K1obs and K2obs for other guests. We employed free online tool Bindfit (supramolecular.org) for comparing the quality of fit with 2:1 and 1:1 binding models (Thordarson, 2011;Hibbert and Thordarson, 2016). Generally, the weakest binding guests 6, 11, 12 and 17 provided unrealistic values of K2obs and could be fitted only with 1:1 binding models. Guests 9, 10, 13, 14 and 15 provided miscellaneous result without any clear trend corresponding to their order in our qualitative comparison. Mostly, it was possible to fit their NMR data with 2:1 and also 1:1 model obtaining similar quality of fit. In some cases (9, 10), both models (2:1 and 1:1) provided reasonable fit; in other cases (13, 14, 15) it meant unrealistic values from 2:1 model and at the same time, regular sinusoidal distribution of residuals for 1:1 model, indicating it as an incorrect model. Overall, the binding between the guests and cycHC[6] cannot be compared in quantitative manner, as the data cannot be evaluated with the same binding model. Notes: a we used 2:1 NMR binding model with "full flavor", which do not assume any specific relation: (i) between chemical shifts of HG and HG2; (ii) between stepwise association constants (K1, K2). We present all obtained values from 2:1 binding model, however, the quality of fit varied from good to insufficient. b Although the value and error of K3obs are same (3:1 binding model), we obtained random distribution of residuals ( Figure S10) indicating reliability of overall fit. c Fitting of data from line 2 and 5 for 18.2 mM trifluoroacetic acid (16) by 2:1 model provided insufficient quality of fit with sinusoidal shape of residuals. d Data from line 6 and 7 were fitted simultaneously to improve quality of fit with 3:1 binding model. Table S23 can be found on following web addresses: