Anion-assisted amidinium exchange and metathesis

Dynamic covalent chemistry has become an invaluable tool for the design and preparation of adaptable yet robust molecular systems. Herein we explore the scope of a largely overlooked dynamic covalent reaction – amidinium exchange – and report on conditions that allow formal amidinium metathesis reactions.


LCMS reaction monitoring
All samples for LCMS analysis were prepared by diluting 1-3 µL of the analyzed solution (e.g., a reaction mixture) in 1 mL of LCMS-grade acetonitrile. This extent of dilution was sufficient to stop (or significantly slow down) dynamic covalent exchange in order to reliably quantify the analytes by HPLC. LCMS analysis of amidinium exchange and metathesis was performed on Shimadzu LCMS-2020 using Ascen-tis® C8 HPLC column (10 cm × 4.6 mm, particle size -3 µm) or Kinetex® C18 HPLC column (10 cm × 4.6 mm, particle size -2.6 m). Ascentis® C8 HPLC column was used for all samples containing N,N'-di(4methylbenzyl)formamidinium (1bb).
HPLC method. Isocratic elution was applied using: a) 94% mobile phase B and 6% mobile phase A at 50°C and flow rate 1.0 mL/min (for Kinetex® C18 HPLC column) and b) 96% mobile phase B and 4% mobile phase A at 50°C and flow rate 1.0 mL/min (for Ascentis® C8 HPLC column). Mobile phase A: 0.023 M HCO 2 NH 4 and 0.0019 M HCO 2 H in H 2 O. Mobile phase B: acetonitrile. HPLC monitoring was performed with a photodiode array detector at λ= 220 nm.

N,N'-di(4-methylbenzyl)formamidinium tetraphenylborate (1bb·BPh 4 )
Formamidinium tetraphenylborate (FA·BPh 4 ) (0.3 mmol, 1.0 eq.) was dissolved in MeCN (4 mL), and 4methylbenzylamine (0.8 mmol, 2.8 eq.) was added to the solution. The reaction mixture was stirred under reflux for 30 min and all volatiles were then removed under reduced pressure. The residue was re-dissolved in MeCN (0.75 mL) and PhMe (7.5 mL) was added to the mixture. The solution was kept at +4°C overnight and the resulting crystals were filtered off to afford the desired products 1bb·BPh 4 as colorless crystals (0.10 g, 0.17 mmol, 65%). For preparation of the analytically pure sample (without traces of 4-methylbenzylamine), the product was recrystallized twice by dissolving it in the minimal volume of MeCN/CH 2 Cl 2 mixture (1:1) and adding 5-10 fold volume of PhMe.
(q, J = 1. 3 Amidinium exchange: substrate and solvent scope 3.1 General procedure A 1.5 mL HPLC vial was charged with a formamidinium salt (40 µmol) * and 1,2,4,5-tetramethylbenzene (as an internal HPLC standard). The mixture was dissolved in LCMS-grade MeCN (800 µL) and benzylamine (2.0 eq., 80 µmol) was rapidly added to the stirred solution via a Hamilton TM syringe. The reaction progress was monitored by LCMS. Amounts of BnNH 2 , 1a and 1aa (in mol% with respect to the initial amount of the starting formamidinium salts) were determined using the internal standard (calibration curves from Figure S2 were used). * In case of formamidinium acetate, 48 µmol was used and all other reagent and solvent quantities were scaled proportionally. Internal standard -1,2,4,5-tetramethylbenzene. nst/nx -molar ratio between the standard and the calibrated compound; Ast/Ax -ratio of chromatographic peak areas of the standard and the calibrated compound. For benzylamine, calibration curves varied depending on the HPLC method (possibly, due to different content of an acidic buffer used, which affected the degree of protonation of BnNH2 and thus its molar absorptivity); therefore, the curve was updated whenever a new HPLC method was applied.

S4
3.2 Amidinium exchange with FA-BPh 4 : solvent scope        A stock solution of BnNH 2 in MeCN (56 mM; 0 -0.10 eq. with respect to the total amount of the amidinium species) was then added. When it was necessary, a stock solution of a tetrabutylammonium salt in MeCN (515 mM; 0 -1.0 eq. with respect to the total amount of the amidinium species) was subsequently added.
The total volume of the reaction mixture (400 µL) was adjusted with MeCN (in fact, this amount of solvent was added to the HPLC vial prior to addition of the reactants). The reaction progress was monitored by LCMS. The amount of 1aa was determined using the internal standard (the calibration curve from Fig. S2,C was applied). The amount of 1bb was determined from the HPLC peak area multiplied by a factor which equalized the area of the peaks of 1aa and 1bb at t = 0 h (the first HPLC measurement where product 1ab was not yet formed; equality of the amounts of 1aa and 1bb was assured by quantitative NMR of the used stock solutions). Amount of 1ab was determined from the HPLC peak area assuming that molar absorptivity of 1ab was a mean of the molar absorptivities of 1aa and 1bb. Relative amounts of 1aa, 1bb, and 1ab (mol%) were calculated with respect to the initial amount of either 1aa or 1bb (which were used as an equimolar mixture). S11 4.2 Effect of the primary amine amount on the metathesis rate Figure S13: Kinetic plots of the amidinium metathesis between 1aa and 1bb in the presence of different amounts of BnNH2: A) 0 mol%; B) 2.5 mol%; C) 5 mol%; D) 10 mol% (with respect to the total amount of the amidinium species. Reaction conditions: 1.0 eq. 1aa (10 µmol), 1.0 eq. 1bb (10 µmol), 0 -0.2 eq. BnNH2 (0-2 µmol), solvent -MeCN (total volume -400 µL), r.t. S12 Figure S14: Kinetic plots of the amidinium metathesis between 1aa and 1bb in the presence of 100 mol% NBu4OAc and varying amounts of BnNH2: A) 0 mol%; B) 2.5 mol%; C) 5 mol%; D) 10 mol% (all relative quantities are with respect to the total amount of the amidinium species). Reaction conditions: 1.0 eq. 1aa (10 µmol), 1.0 eq. 1bb (10 µmol), 0-0.2 eq. BnNH2 (0-2 µmol), 2.0 eq. NBu4OAc (20 µmol), solvent -MeCN (total volume -400 µL), r.t. Comment: the metathesis reaction containing 0.05% H2O (50 mol% with respect to the total amount of the amidinium species) did not take place in the absence of both BnNH2 and NBu4OAc. This disproves the hypothesis that the residual water from NBu4OAc was the sole reason for the observed metathesis in case of (A). It is, however, possible that AcOcatalyzes both the hydrolysis reaction and the amidinium exchange (see Fig. 3 in the main text). Therefore, we wondered if varying the water content in the reaction mixture while keeping the concentration of AcOfixed would affect the metathesis rate ( Figure S15). S13 Figure S15: Kinetic plots of the amidinium metathesis between 1aa and 1bb in the presence of 100 mol% NBu4OAc (with respect to the total amount of the amidinium species) and varying amounts of water. The lines are shown to guide the eye. Reaction conditions: 1.0 eq. 1aa (10 µmol), 1.0 eq. 1bb (10 µmol), 2.0 eq. NBu4OAc (20 µmol), solvent -MeCN/H2O (total volume -400 µL), r.t. Comment: Even 1% (v/v) of water significantly increased the rate of the metathesis. This result indicates that AcOmight facilitate hydrolysis of the starting materials which, in turn, leads to release of free benzylamines that drive the metathesis. Therefore, even without addition of extra BnNH2, the metathesis could take place in non-anhydrous MeCN in the presence of AcO - (Fig. S14,A). Large amounts of water (>7% v/v) reduced the observed rate enhancement, possibly, for the same reasons as for the amidinium exchange in aqueous or alcoholic solvents (see Table S1 and discussion in the main text).  are with respect to the total amount of the amidinium species). Reaction conditions: 1.0 eq. 1aa (10 µmol), 1.0 eq. 1bb (10 µmol), 0.1 eq. BnNH2 (1 µmol), 2.0 eq. NBu4 + salt (20 µmol), solvent -MeCN (total volume -400 µL), r.t.

Effect of anions on the metathesis rate
In case of metatheses in the presence of H2PO4 -(A) and Cl -(B), the solvent contained 20% and 10% (v/v) MeOH respectively (to increase solubility of the amidinium salts). The time when the amount of metathesis product 1ab reaches 50% from its equilibrium amount.