Selective binding of ReO 4– and PtCl 42– by a Pd 2 L 4 cage in water

Titration with KReO4..................................................................................................................................... 3 Titration with K2PtCl4 .................................................................................................................................... 4 Titration with cisplatin 5 .............................................................................................................................. 5 Titration with oxaliplatin 6 ........................................................................................................................... 6 Titration with nedaplatin 7 .......................................................................................................................... 7 Titration with biotin 8 .................................................................................................................................. 8 Titration with sodium chloride .................................................................................................................... 9 Titration with sodium acetate ................................................................................................................... 10 Titration with caffeine 9 ............................................................................................................................. 11 Titration with uridine 10 ............................................................................................................................ 12 Titration with cytidine 11 ........................................................................................................................... 13 Titration with phenylalanine 12 ................................................................................................................. 14 Titration with glutathione 13 ..................................................................................................................... 15 Stability study with histidine 14 ................................................................................................................. 16

Section S1. Materials and methods All solvents and chemicals were purchased from commercial suppliers and used without further purification. Cage 3 was synthesized as previously reported. [1] NMR Nuclear Magnetic Resonance (NMR) spectra were recorded on a Bruker DRX 500 operating at 298 K for NOE and 1 H NMR titration experiments.

Acidity measurements
The acidity was measured with a SI analytics Handylab 100 with the pH electrode Blueline 14 pH (pH 0-14; -5-100 °C; 3 mol/L KCl referenced; catalogue number 285129140) or the pH electrode ScienceLine micro N 6003 (pH 0-14; -5-100 °C; Ag/AgCl referenced; catalogue number 285105176). The pH meter was calibrated beforehand with Sigma-Aldrich/Merck buffered reference standards of pH 4.00 (red colourcoded; catalogue number B5020), pH 7.00 (yellow colour-coded; catalogue number B4770) and pH 10.00 (blue colour-coded; catalogue number B4895). The acidity in D 2 O was measured with the electrode calibrated in H 2 O and reported as pH* as detailed by Krężel and Bal. [2] Titration studies A 0.83 mM solution of 3 in D 2 O was prepared, pH was adjusted to pH 7.0-7.4, sonicated for 30 minutes and then used as such. [1] During the titration, known aliquots of the stock solution of titrant in D 2 O was added to an NMR tube containing 600 µL of this 0.83 mM solution of 3 in D 2 O and a 1 H NMR spectra was recorded.
Association constants (K a ) were determined by monitoring the change in chemical shift (Δδ) for a selected hydrogen resonance of 3 and fitting these shifts to a 1:1 binding model using a non-linear least squares fitting implemented in Excel or HypNMR. [3] Computations All models were generated manually and geometry optimized with Spartan 2016 without any (geometrical) constraints. Following an initial MMFF optimization, the resulting coordinates were subjected to a computation using density functional theory (DFT) at the ωB97X-D / 6-31G* level of theory with an explicit water solvation model as implemented in Spartan 2016. Energies in kcal•mol -1 are derived from the energies in hartrees by ignoring entropy using the simple subtraction of (E adduct -E anion -E cage ) x 627.509608. Section S2. NMR binding studies Titration with KReO 4 Figure S1: Top: 1 H NMR spectra and assignment of a binding study of cage 3 with KReO 4 in D 2 O at pH* 7.0. The guest stock solution concentration was 0.021 M. Initial concentration of host = 0.83 mM. The vertical red dashed lines were added as a guide to the eye. Bottom: HypNMR binding analysis following hydrogen signals d, f, a and c of the cage. The chemical shifts could be fitted to a 1:1 binding model (left) with K a = 434 (rsd = 0.6%) with a reasonable fit or r 2 = 0.99963 on all 120 data points. This could be improved to r 2 = 0.9993 by assuming a 1:3 binding model (right) with K a 1:1 = 434 M -1 and the next constants set to 10 M -1 . A 1:3 stoichiometry implies one strongly bound interior ReO 4 ] 2-and two more loosely associated [ReO 4 ] 2-complexes to further compensate the charges of the cage (likely on the cage's exterior). On this basis the 1:1 binding is assessed as 434 M -1 . The modelled species distributions is also shown as coloured lines with 'Host' = green and 'Host-Guest' = blue and 'Host-Guest 2 ' = brown.
Titration with K 2 PtCl 4 Figure S2: Top: 1 H NMR spectra and assignment of a binding study of cage 3 with K 2 PtCl4 in D 2 O at pH* 7.0. The guest stock solution concentration was 0.020 M. Initial concentration of host = 0.83 mM. The vertical red dashed lines were added as a guide to the eye. Bottom: HypNMR binding analysis following hydrogen signals a, c, d, f and g of the cage. The chemical shifts could be fitted to a 1:1 binding model (left) with K a = 6,901 (rsd = 2.1%) with a reasonable fit or r 2 = 0.9738 on all 90 data points. This could be improved to r 2 = 0.9959 by assuming a 1:3 binding model (right) with K a 1:1 = 31,600 M -1 (10 4.5 ) and the next constants set to 10 M -1 . A 1:3 stoichiometry implies one strongly bound interior [PtCl 4 ] 2and two more loosely associated [PtCl 4 ] 2-complexes to further compensate the charges of the cage (likely on the cage's exterior). On this basis the 1:1 binding is assessed as in the order of 10 4 M -1 . The modelled species distributions are also shown as coloured lines with 'Host' = green and 'Host-Guest' = blue.
Titration with cisplatin 5 Figure S3: 1 H NMR spectra and assignment of a binding study of cage 3 with cisplatin (5) in D 2 O at pH* 7.3. The guest stock solution concentration was 8.3 mM (due to solubility limitations). Initial concentration of host = 0.91 mM. The vertical red dashed lines were added as a guide to the eye. Bottom left: the very small shifts observed at the end of the titration could be modelled (not fitted) with HypNMR to a 1:2 model with stepwise constants of 1.3 and 2.5 M -1 . The modelled species distributions is also shown as coloured lines with 'Host' = green, 'Host-Guest' = blue, and 'Host-Guest 2 ' = brown. Due to the lack of saturation, we interpret these shifts as the onset of genuine 1:1 binding with an affinity close to the detection limit of about 3 M -1 and report such shifts as 'not binding' in the paper. That these shifts are very small is illustrated by the rescaled plot in the bottom right, which can be contrasted with the shifts observed with the strongly binding K 2 [PtCl 4 ] with Δδ max = ± 0.1 p.p.m. (see Figure S2).
Titration with oxaliplatin 6 Figure S4: 1 H NMR spectra and assignment of a binding study of cage 3 with oxaliplatin (6) in D 2 O at pH* 7.1. The guest stock solution concentration was 10.1 mM (due to solubility limitations). Initial concentration of host = 0.91 mM. The vertical red dashed lines were added as a guide to the eye. Bottom left: the very small shifts observed at the end of the titration could be modelled (not fitted) with HypNMR to a 1:2 model with stepwise constants of 1.3 and 2.5 M -1 . The modelled species distributions is also shown as coloured lines with 'Host' = green, 'Host-Guest' = blue, and 'Host-Guest 2 ' = brown. Due to the lack of saturation, we interpret these shifts as the onset of genuine 1:1 binding with an affinity close to the detection limit of about 3 M -1 and report such shifts as 'not binding' in the paper. That these shifts are very small is illustrated by the rescaled plot in the bottom right, which can be contrasted with the shifts observed with the strongly binding K 2 [PtCl 4 ] with Δδ max = ± 0.1 p.p.m. (see Figure S2).
Titration with nedaplatin 7 Figure S5: 1 H NMR spectra and assignment of a binding study of cage 3 with nedaplatin (7) in D 2 O at pH* 7.4. The guest stock solution concentration was 0.010 M. Initial concentration of host = 0.24 mM The vertical red dashed lines were added as a guide to the eye. Bottom left: the very small shifts observed at the end of the titration could be modelled (not fitted) with HypNMR to a 1:2 model with stepwise constants of 1.3 and 2.5 M -1 . The modelled species distributions is also shown as coloured lines with 'Host' = green, 'Host-Guest' = blue, and 'Host-Guest 2 ' = brown. Due to the lack of saturation, we interpret these shifts as the onset of genuine 1:1 binding with an affinity close to the detection limit of about 3 M -1 and report such shifts as 'not binding' in the paper. That these shifts are very small is illustrated by the rescaled plot in the bottom right, which can be contrasted with the shifts observed with the strongly binding K 2 [PtCl 4 ] with Δδ max = ± 0.1 p.p.m. (see Figure S2). Binding of up to four equivalents of chloride seems reasonable to compensate the charges of both palladium cations. Nevertheless, the exact stoichiometry remains uncertain, but from these models one can infer that 1:1 binding is in the order of 10-20 M -1 .