Multi-stimuli Control over Assembly and Guest Binding in Metallo-supramolecular Hosts Based on Dithienylethene Photoswitches

It is difficult to assemble multi-component metallo-supramolecular architectures in a non-statistical fashion, which limits their development toward functional materials. Herein, we report a system of interconverting bowls and cages that are able to respond to various selective stimuli (light, ligands, anions), based on the self-assembly of a photochromic dithienylethene (DTE) ligand, La, with PdII cations. By combining the concept of “coordination sphere engineering”, relying on bulky quinoline donors, with reversible photoswitching between the ligand’s open (o-La) and closed (c-La) forms, a [Pd2(o-La)4] cage (o-C) and a [Pd2(c-La)3] bowl (c-B) were obtained, respectively. This structural rearrangement modulates the system’s guest uptake capabilities. Among three bis-sulfonate guests (G1, G2, and G3), the cage can encapsulate only the smallest (G1), while the bowl binds all of them. Bowl c-B was further used to synthesize a series of heteroleptic cages, [Pd2LA3LB], representing a motif never reported before. Additional ligands (Lc-f), with short or long arms, tune the cavity size, thus enabling or preventing guest uptake. Addition of Br–/Ag+ makes it possible to change the overall charge, again triggering guest uptake and release, as well as fourth ligand de-/recomplexation. In combination, site-selective introduction of functionality and application of external stimuli lead to an intricate system of hosts with different guest preferences. A high degree of complexity is achieved through cooperativity between only a few components.


General procedures
All chemicals, except otherwise specified, were obtained from commercial sources, and used without further purification. Perfluoro-1,2-bis(2-iodo-5-methylthien-4-yl)cyclopentene [S1] , L c , [S2] L d [S3] were prepared according to a literature procedure. The tetrabutylammonium salts of the guests (G1, G2 and G3) were synthesized according to a previously reported procedure. [S4] Corresponding starting materials were obtained commercially as sodium salt or free sulfonic acids in the highest available purity. Recycling gel permeation chromatography was performed on a JAI LC-9210 II NEXT GPC system equipped with Jaigel 1H and 2H columns in series using chloroform as the eluent (HPLC grade). NMR measurements were all conducted at 298 K on Avance-500, Avance-600 and Avance-700 instruments from Bruker and INOVA 500 MHz machine from Varian. High-resolution ESI mass spectrometric measurements were carried out on maXis ESI-TOF MS, Compact and ESI-timsTOF machines from Bruker. Irradiations at 313 nm were performed by placing a quartz NMR tube in a distance of 5 cm in front of a 300 W Hg arc lamp from LOT-Oriel equipped with a dichroic mirror and 313 nm bandpass filter. Irradiations at 617 nm were performed by placing a quartz NMR tube 2.5 cm in front of a LED irradiation apparatus (3x 1.4 W 617 nm Power LED, 25 nm FWHM) from Sahlmann Photonics, Kiel. Chiral HPLC was performed on an Agilent Technologies 1260 infinity HPLC system equipped with Daicel CHIRALPAK IC columns (250 x 4.6 mm and 250 x 10 mm) using a dichloromethane/hexane/methanol (39%/60%/1%) mixture as eluent for the separation of c-L a . UV-Vis spectra were recorded on an Agilent 8453 UV-Visible spectrophotometer. CD spectra were measured on an Applied Photophysics Chirascan circular dichroism spectrometer. All heteroleptic cage and c-B models were constructed using SPARTAN [S5] and were first optimized on a PM6 level of theory (no counter ions were included) without constrains. The resulting structures were than further refined by DFT calculations (B3LYP/LANL2DZ) using Gaussian 09 software. [S6] 2. Synthesis

Synthesis of
(1.1 g, 8.0 mmol) and degassed toluene/EtOH/H2O (30 mL V/V/V = 2:1:1) were combined in an oven-dried Schlenk tube and stirred at 90 °C for overnight. After cooling down to room temperature, CH2Cl2 was added, and the solution was washed with water and brine, dried over anhydrous MgSO4 and concentrated in vacuo.

Symmetry
Similar to our previous reported cage, [S1] there are 6 possible stereoisomers of the bowl-like complex c-B:

Absolute configuration
The open-form ligand o-L a (10 mg) was dissolved in CD2Cl2 (1 mL) and then irradiated with 313 nm UV light. The pure (R,R) and (S,S) ligand enantiomers were separated by preparative chiral HPLC using a Daicel IC column (250 x 10 mm) under strict exclusion of daylight. The purity of separated 1 (the first fraction of HPLC) and 2 (the second fraction of HPLC) enantiomers were checked using analytical HPLC using a Daicel IA column (250 x 4.6 mm). In order to determine the absolute stereochemistry of the enantiomers, CD spectra were calculated by TD-DFT methods at the B3LYP/6-31G(d) level of theory in the Gaussian 09 software [S2] (using keyword iop(9/40=2)).      1 mM, CH3CN);. Both of the calculated CD spectrum of c-L (R,R) and fraction 1 give negative cotton effect from 500 nm to 800 nm (this area shows the characteristic peak of closed form DTE derives), which means fraction 1 has R,R configuration, in contrast, fraction 2 is S,S configuration.

Computational studies
All models were constructed using SPARTAN and were first optimized on a PM6 level of theory (no counter ions were included) without constrains. The resulting structures were than further refined by DFT gas-phase calculations (B3LYP/LANL2DZ) using Gaussian 09 software.

Photoswitching
Irradiations at 313 nm were performed by placing a quartz NMR tube in a distance of 5 cm in front of a               showing the reversible capture/release G2 upon adding Ag + /Br -. Figure S118. 1 H NMR spectra (500 MHz, 298 K, acetonitrile-d3) of disassembly and re-assembly of heteroleptic cage. The first spectrum on the bottom shows the 1 H NMR spectrum of (R)-c-B-L e . Stacking up one by one, the following spectrum shows the mixture of releasing L e with 2.0 eq. of Br -. After addition of 2.0 eq. or even 10 eq. Ag + into the mixture, the disassembled sample recovered back to (R)-c-B-L e . The 1 H NMR spectrum on the top setting for reference is (R)-c-B-Br. Red color marks the signals of (R)-c-B-L e , and green color marks uncoordinated 'free' L e .

Data Collection
Suitable single crystals for X-ray structural analysis of o-C, c-C and G1@c-B were mounted at room temperature in NVH oil. Crystals were stored at cryogenic temperature in dry shippers, in which they were safely transported to macromolecular beamline P11 [S7] at the Petra III synchrotron, DESY, Germany. X-ray diffraction data was collected at 80(2) K on a single axis goniometer, equipped with an Oxford Cryostream 800 low temperature device and a Pilatus 6M fast detector. The data integration and reduction were performed with XDS [S8] . The structure was solved by direct methods. The structure model was refined against all data by full-matrix least-squares methods on F 2 with the program SHELXL2014 [S9] . The SQUEEZE [S10] method provided by the program Platon [S11] was used to improve the contrast of the electron density map the structure.
For G1@o-B-L f , the crystal was mounted at ambient temperature in NVH oil and measured on a Bruker D8 Venture diffractometer with an INCOATEC microfocus sealed tube, Iys 3.0 source and Bruker D8 Venture CMOS with Photon 100 detector. Data reduction was performed with SAINT v8.30C (Bruker, 2009a) out of the APEX 3 (Bruker, 2019) program package. SADABS [S12] (version 2014/4) was employed for the incident beam scaling, determination of the spherical harmonic coefficients, outlier rejection and determination of the error model parameters. All the structures were solved by direct methods with SHELXT [S13] . They were refined by full-matrix least-squares against F² using SHELXL2014 with the help of the SHELXle [S14] graphical user interface.
Suitable single crystals of o-L e were mounted in NVH oil on a nylon loop. X-ray diffraction data were collected on a Bruker d8 venture systems based on a kappa goniometer with Incoatec microfocus X-ray 7 8 9 10 11 12 ppm G1@o-B G1@o-B-L f + 1.0 eq. L f S65 sources (IµS 2.0), Incoatec QUAZAR mirror optics and a Photon 100 detector. The data were collected at 100 K crystal temperature (Oxford Cryosystems CRYOSTREAM 700), 50 kV and 600 µA and an appropriate 0.5° omega scan strategy. Data reduction was performed with SAINT v8.30C (Bruker, 2009a) out of the APEX II v2.2012.2 0 (Bruker, 2009b program package. SADABS [S12] (version 2014/4) was employed for the incident beam scaling, determination of the spherical harmonic coefficients, outlier rejection and determination of the error model parameters. All the structures were solved by direct methods with SHELXT [S13] . They were refined by full-matrix least-squares against F² using SHELXL2014 with the help of the SHELXle [S14] graphical user interface.  The same atom names are used for different disorder components, which only differ in the part numbers.
Therefore, the TABS keyword behind the ACTA instruction (Sheldrick, 2017) was employed to generate the CIF.
For o-C, c-C, G1@c-B and G1@o-B-L f , these structures were treated with SQUEEZE routine of the PLATON program, because the crystal lattice contains large voids filled with disordered solvent molecules.

Structural analysis
The compound o-C crystalized in the tetragonal space group P42 while compound c-C crystallizes in the P42/mcm space group and the asymmetric unit contains both possible enantiomers with 50:50 occupancy. S68 Å for the neighbor H, and 2.808 Å for the opposite one ( Figure S117). Finally, the difference in flexibility is expressed also in the intramolecular Pd···Pd distances, of 12.533(3) Å and 12.256(14) Å for o-C and c-C, respectively (Table S3).

Figure S121.
Side and top views of the coordination environment for a) o-C and b) c-C, with focus on the Hb···Hb distances (dotted red line) and the twisting between the quinoline planes and the plane created by Pd and the four coordinating N atoms (cyan surfaces); the angle between the planes is measured as the angle between the normal to the planes (black arrow); c) overlay between o-C (red structure) and c-C (blue structure) coordination environments. Table S3. Crystallographic analysis for o-C and c-C compared to previously reported c-C S15 and o-C S1,15 .
o-C* c-C o-C S1 o-C S15 c-C S15 *there are two values of H···H distances and twist angle due to symmetric reasons. § calculated as the angle between the normal to the plane created by Pd and the four coordinating N atoms and the normal to the quinoline plane or to the normal to the pyridine plane for the previously reported structures. S15,S16 ~ pyridine-pyridine outward-pointing protons.

S69
Compound G1@c-B crystalized in P-1 space group and the asymmetric unit contains both enantiomers with 50:50 occupancy. Each metal center is coordinated by three quinoline donor groups and one water molecule. G1 is accommodated in the concave well of the bowl between the two Pd atoms. The following table reports the distances between the oxygen atoms of the sulfonates, pointing inside the bowl cavity, and the neighboring Pd ion and protons Ha of all three ligands. The two disordered parts are reported separately.    Abs. at 590 nm(a. u.)

UV-vis spectroscopy
Cycles