Controlled hierarchical self-assembly of networked coordination nanocapsules via the use of molecular chaperones

Supramolecular chaperones play an important role in directing the assembly of multiple protein subunits and redox-active metal ions into precise, complex and functional quaternary structures. Here we report that hydroxyl tailed C-alkylpyrogallol[4]arene ligands and redox-active MnII ions, with the assistance of proline chaperone molecules, can assemble into two-dimensional (2D) and/or three-dimensional (3D) networked nanocapsules. Dimensionality is controlled by coordination between the exterior of nanocapsule subunits, and endohedral functionalization within the 2D system is achieved via chaperone guest encapsulation. The tailoring of surface properties of nanocapsules via coordination chemistry is also shown as an effective method for the fine-tuning magnetic properties, and electrochemical and spectroscopic studies support that the nanocapsule is an effective homogeneous water-oxidation electrocatalyst, operating at pH 6.07 with an exceptionally low overpotential of 368 mV.


Materials and Methods.
All solvents and chemicals were purchased from Sigma-Aldrich or Fisher Laboratories and used without further purification. Notably, manganese (II) nitrate tetrahydrate crystals were stored in glovebox at 23 o C. The solvents were dried by 3Å molecular sieves for uses of solvothermal synthesis. All combinations of PgC 3 OH macrocycles and manganese salts were carried out using glovebox techniques at 23 o C. All pH measurements were performed using a Thermo Scientific Orion Star A111 Benchtop pH Meter. Powder X-ray Diffraction data was collected on a Bruker Apex II CCD diffractometer at room temperature using Cu (Kα) radiation Inco-tech Microfocus II (1.5406Å). Powder X-ray diffraction was measured on a Bruker X8 Prospector single crystal X-ray diffractometer equipped with an IμS microfocus Cu-Kα X-ray source (λ = 1.54106 Å, power = 40 kV, 0.65 mA). Dry samples were hand-ground into powder and loaded directly into the tubing. Data collection was performed with the area detector and X-ray source fixed, and the tubing containing the sample at a 90 o angle to the X-ray beam at a sampleto-detector distance of 8.00 cm. The samples were rotated 360 o along the axis of the tubing during collection. Each data set was composed of a series of 2-minute long scans across the 2-theta range of 2.5 to 40 o . Photographic data were reduced by integrating along a 77 o -wide sector from 2.5 to 35 o 2-theta in 0.02 o slices along 2-theta. Small angle X-ray scattering analyses was characterized with Xenocs SAXS equipment. The experiment was conducted in 1200s, under a power of 50 kV, 0.6 mA, using a PILATUS 100K detector and the wavelength was 0.15148. Positive ion MALDI (Matrix-Assisted Laser Desorption Ionization) TOF (Time of Flight) mass spectrometer measured on a Bruker Autoflex Speed MALDI TOF MS using dithranol as the matrix. Samples in water was combined with methanol containing dithranol molecules. FT-IR spectra were recorded at room temperature using a Thermo Nicolet Avatar 360 FTIR Spectrometer in the 400-4000 cm −1 range. Elemental analysis (EA) was performed using a European A3000 Elemental Analyzer. Thermogravimetric analysis (TGA) was performed using a TA Instruments Q50 TGA, with a Pt sample pan under 40 mL min -1 nitrogen purge. The sample was heated from room temperature to 800 °C at the rate of 20 °C/min. Differential scanning calorimetry (DSC) was performed using a TA Instruments Q1000@Mfg-dsc, with an Al hermetic sample pan under 40 mL min -1 nitrogen purge. The sample was heated from 40 °C to 600 °C at the rate of 10 °C/min. X-ray photoelectron spectroscopy (XPS) spectra were recorded using a Thermo Fisher Scientific Escalab 250. UV-visible (UV-Vis) spectra were measured using a Varian 50 BIO spectrophotometer. Crystals of 1 and 2 was suspended in 0.1 M aqueous acetate buffer at pH 6.07. The mixture was sonicated for 15 min at 45 °C to yield a yellow solution. The solution was filtered using a Whatman Puradisc 30 syringe filters (pore size 0.2 μm) and then subjected to UV-Vis analysis. Corresponding UV-Vis samples in acetate buffer (filtered S3 solution) were subjected to Dynamic light scattering (DLS) analysis with BECKMAN COULTER DelsaTM Nano C particle analyser. Scanning electron microscopy (SEM) images were obtained in field emission scanning electron microscope (FESEM; MERLIN Compact, Carl Zeiss) at an acceleration voltage of 200 kV.
Corresponding UV-Vis and DLS samples in acetate buffer (filtered solution) were drop-casted on a silicon wafer following naturally drying and then washed with Milli-Q ultrapure water.
Single crystal X-ray diffraction data for 1 were collected on a Bruker Apex II diffractometer equipped with a CCD area detector using Mo-Kα radiation from a fine-focus sealed source with a focusing collimator (Bruker Nano). Data for 2 were collected on Bruker D8 Venture diffractometer with a Photon 100 CMOS area detector using Mo-Kα radiation (λ = 0.71073 Å) from an IµS microfocus source (Bruker Nano, Inc., Madison, WI, USA). Crystals were cooled to 100 K under a cold stream of N 2 gas using a Cryostream 700 cryostat for 1 (Oxford Cryosystems) and a Cryostream 800 cryostat (Oxford Cryosystems, Oxford, UK) for 2. Hemispheres of unique data were collected using strategies of scans about the phi and omega axes. The Apex3 software suite was used for data collection, unit cell determination, data reduction, scaling, and absorption correction. 1 Compound 1 was solved and refined using SHELXL-2017 2 and SHELXT 3 as the interface. The ordered portion of the structure was refined anisotropically. Some of the propanol chains were so strongly affected by disorder that no clear interpretation of the difference map was possible; however each PgC 3 OH moiety has at least 1 fully ordered propanol chain which confirms the identity of the moiety. For the disordered chains, the closest reasonable difference map peaks were refined as propanol chains using distance and angle restraints. Atoms that gave clearly unrealistic geometries or displacement parameters were not given any riding hydrogen atoms. These atoms were left in the structure to help make the disordered solvent calculation more accurate and allow better visualization of the disordered regions. Compound 1 also showed difference map peaks inside the MONCs which closely resembled the expected geometry of a proline ligand, but attempts to refine this ligand failed to converge with a realistic geometry. Ultimately the entire proline moiety was modeled as a rigid group using coordinates from a published structure with a similar conformation 4 and using a single parameter to describe the atomic displacements for all ring carbon atoms. This refinement revealed that the proline ring was disordered over two conformations, both of which could be modeled with distance and angle restraints.
The difference map also indicated the presence of a second unique proline molecule in the cavity, but not all atoms could be located. This molecule was ultimately excluded from the model and treated with S4 a solvent mask. Olex2 v. 1.3.0 was used for model building and as an interface for SHELX. 5 PLATON SQUEEZE was used to implement solvent masks. 6 Literature coordinates were obtained from the Cambridge Crystallographic Data Center using the database searching software ConQuest V. 2.0. 5. 7 Compound 2 was solved by isomorphous replacement. The coordinates of the isomorphous Mg 2+ analog 8 were used as an initial model with all Mg sites replaced with Mn. The structure was refined to convergence by full matrix least squares refinement against F 2 using SHELXL-2017. 2 The diffraction data for the crystal was essentially negligible beyond 1.1 angstroms (R int > 50% for the 1.44-1.39 angstrom shell; average I/sigma at 1.10 is approximately 0.40). Due to the lower data-to-parameters ratio of the Mn 2+ model caused by the weak diffraction, some disordered lattice solvent molecules from the Mg 2+ model were removed using a solvent mask. The converged model (after solvent masking) has a GooF near 1 and a reasonably smooth residual difference map, both indicate the coordinates from the Mg 2+ model agree well with the Mn 2+ data. The presence of two lattice acetonitrile molecules from the Mg 2+ model that refine well further supports that the packing in these two structures is almost identical.
H atoms could not be located for O-H groups and were left out of the model but included in the formula.
The identities of the axial ligands bound the Mn ions are uncertain from the X-ray diffraction data, so only the coordinating O atoms were included in the formula. The formula assumes that Mn 2+ ions are charge balanced by deprotonated phenolic O-H groups.

Preparation and characterization of the HSSs crystals 1 and 2
Preparation of 1.      Hydrogen atoms, axial ligands and hydroxyl tail alkyl chains not involved in metal-ligand coordination to adjacent MONC subunits were removed for clarity. Proline molecules could not be explicitly modeled in the crystal structure, but they are essential for the synthesis.      Figure S19. Continuous 30 CVs of 1 (a) and 2 (b) (0.5 mM, 50 mV s -1 scan rate) in 0.1 M acetate buffer at pH 6.07 using FTO as the working electrode. The catalytic current does not increase over successive cyclic voltammetric (CV) scans. A crossover profile with a re-reduction wave in the reverse CV scan was not observed. These observations suggest that MONC subunit is a homogeneous catalyst for water oxidation. 14 S17 Figure S20. Bulk electrolysis at 1.79 V vs. NHE of 1 mM 1 (a) and 2 (b) in 0.1 M acetate buffer at pH 6.07 using FTO as the working electrode. For comparison, bulk electrolysis of the blank buffer is also performed. The catalytic current does not increase, as would be expected for a heterogeneous catalytic system.