Construction and Transmetalation of Isostructural D3-Symmetric M11L6 Porphyrin Cage Complexes with Three Distinct Metal Centers

概要

1. General introduction
　Coordination-driven self-assembly has been widely used to construct a variety of well-defined nano-sized architectures. In general, self-assembly of highly symmetric ligands and one kind of metal ions affords a symmetric structure (ex. tetrahedra, octahedra, cube, etc.) having chemically equivalent metal centers. On the other hand, only a few examples of low symmetry self-assembled structures with distinct metal centers have been reported so far.
　Our group has previously reported the construction of D3-symmetric Zn11(L0·H2O)6(H2O)18(OTf)22 (L0-Zn11) with three distinct metal centers including one kind of Zn(bpy)3 center (Zn1) and two kinds of Zn(bpy)2(H2O)2 centers (Zn2 & Zn3) by using a C4v-symmetric porphyrin ligand, L0. By using its unique D3-symmetric structure, unsymmetrical guest encapsulation was also achieved. However, the construction of M11L6 cage complex was only limited to L0-Zn11 mainly due to not only low solubility of L0 in organic solvents but also delicate energy balance of L0-Zn11 in some solvents, although the isostructural M11L6 cage complexes have great potential due to their unique M11L6 structural features and metal-dependent characters.
　To overcome this problem, in this work, I have newly synthesized the porphyrin ligand L1 having methyl groups at the 5’-position of 2,2’-bipyridyl moieties to enhance its coordination ability to metal ions due to the electron- donating character and solubility in organic solvents. I expected that this chemical modification would prevent the formation of insoluble intermediates to cause stoichiometric changes between metal ions and ligands.
　Firstly, I have synthesized two isostructural M11L6 cage complexes, Cd11(L1·CH3OH)6(H2O)18(OTf)22 (L1-Cd11) and Zn11(L1·CH3OH)6(H2O)18(OTf)22 (L1-Zn11), which are isostructural but have different properties depending on the different ionic radii. On the other hand, no discrete species were observed in the case of complexation between L0 and Cd(OTf)2 due to the lower binding constants for Cd(bpy)n than those of Zn(bpy)n. Moreover, a heterometallic cage complex Cd2Zn9(L1·X)6A18(OTf)22 (L1-Cd2Zn9) was successfully converted from L1-Cd11 by site-selective transmetalation at the three types of CdII centers, which is an excellent way to construct an isostructural heteronuclear cage compound bearing more than one distinct metal centers.

2. Constructions of isostructural M11L6 cage complexes
　A ZnII-porphyrin-based ligand, L1·H2O, was prepared from 5’-methyl-2,2’-bipyridyl-5- carboxyaldehyde, which was synthesized from 2- bromo-5-methylpyridine in 3 steps, and pyrrole via Adler synthesis followed by complexation with Zn(OAc)2·(H2O)2 in 6.6% yield. L1·H2O was well characterized by NMR, ESI-MS, and elemental analysis. The C4v symmetric structure of L1·H2O was confirmed by 1H NMR.
　Firstly, L1·H2O was mixed with 11/6 eq. Zn(OTf)2 in CDCl3/CD3OD/D2O = 10:10:1 (v/v/v) to prepare a [Zn11L16]22+ complex. After heating at 50 °C for 24 h, the crude product was purified by precipitation using Et2O to afford L1-Zn11 as a dark green solid in 78% yield. L1-Zn11 was characterized in solution by NMR, and ESI-MS, and in the solid state by elemental analysis and single crystal XRD analysis. The molecular structure was found to have a D3-symmetric [Zn11(L1·CH3OH)6(H2O)18]22+ skeleton with three distinct Zn centers (one kind of Zn(Mebpy)3 (Zn1) and two kinds of Zn(Mebpy)2(H2O)2 (Zn2 & Zn3)). On the other hand, in the solution state, neither CH3OH at the axial position of L1 nor H2O molecules bound to Zn(Mebpy)2 moieties were observed in 1H NMR and ESI-MS spectra due to their rapid exchange with another solvent molecule in solution. It should be noted that the structure framework of L1-Zn11 in solution state was well accorded with the crystal structure as shown in the inter-ligand NOE patterns.
　Then, ligand L1·H2O was mixed with 11/6 eq. of Cd(OTf)2·H2O in CDCl3/CD3OD/D2O = 10:10:1 (v/v/v) to construct a [Cd11L16]22+ complex. After heating at 50 °C for 17 h followed by precipitation using Et2O as a poor solvent, the desired product, L1-Cd11 was obtained as a purple solid in 70% yield. L1-Cd11 was characterized in solution by NMR, and ESI-MS, and in the solid state by elemental analysis and single crystal XRD analysis. The molecular structure was found to have a D3-symmetric [Cd11(L1·CH3OH)6(H2O)18]22+ skeleton with three distinct Cd centers (one kind of Cd(Mebpy)3 (Cd1) and two kinds of Cd(Mebpy)2(H2O)2 (Cd2 & Cd3)). On the other hand, in the solution state, neither CH3OH at the axial position of L1 nor H2O molecules bound to Cd(Mebpy)2 moieties were observed in 1H NMR and ESI-MS spectra due to their rapid exchange with another solvent molecule in solution. It should be noted that the structure framework of L1-Cd11 in solution state was well accorded with the crystal structure as shown in the inter-ligand NOE patterns. These results indicate that L1-Cd11 is isostructural with L1- Zn11.
　The molecular structure framework of L1-Zn11 looks very similar to that of L1-Cd11. However, the bond distances and angles are significantly different from each other as shown in the typical data. The Cd1-Cd1 distance (27.745 Å) is longer than the Zn1-Zn1 distance (27.455 Å). The angles of 㲃Cd2-Cd3-Cd2 (61.60°) are smaller than those of ∠Zn2-Zn3-Zn2 (64.33°). On the other hand, the angles of ∠Cd3-Cd2-Cd3 (100.07°) are larger than those of ∠Zn3-Zn2-Zn3 (98.31°). These results indicate that L1-Cd11 is slightly extended along the C3 axis associated with changes in pore size and shape, although the calculated volume of inner cavity (Connolly surface, probe radius: 4.0 Å) was almost the same as each other (L1-Zn11: 511 Å3, L1-Cd11: 509 Å3).

3. Transmetalation of [Cd11L6]22+
　Considering smaller binding constants of Cd(bpy)n moieties than Zn(bpy)n moieties, transmetalation of L1-Cd11 was examined by adding Zn(OTf)2. Here, not only complete transmetalation at all the CdII centers of L1-Cd11 to L1-Zn11, but also site- selective transmetalation at only Cd(Mebpy)2A2 sites was examined to obtain L1-Cd2Zn9.
　Firstly, to a solution of L1-Cd11 in CDCl3/CD3OD/D2O = 10:10:1 (v/v/v) was added, 11/6 eq. Zn(OTf)2, and the reaction mixture was heated at 50 °C being monitored by 1H NMR. In the 1H NMR spectra of the reaction mixture, the signals of L1-Cd11 gradually changed to that of L1- Zn11. After 72 h, the signals were almost identical with that of L1-Zn11, and the yield was estimated by NMR to be ca. 80%.
　Then, site-selective transmetalation of L1-Cd11 at Cd(Mebpy)2A2 to L1-Cd2Zn9 was examined using a solution of L1-Cd11 in CD3CN. To a solution of L1-Cd11 in CD3CN was added 9/6 eq. Zn(OTf)2, and the reaction temperature was kept at 25 °C being monitored by 1H NMR. After 60 min, new 32 signals appeared in the aromatic region and 4 singlet signals in the alkyl region which were different from those of L1-Zn11 and L1-Cd11. Each signal was fully assigned by 1H-1H COSY and 1H-1H NOESY, which indicated the formation of L1-Cd2Zn9 in the solution. The ESI-MS spectrum suggested the formation of [Cd2Zn9L16(OTf)22-n]n+ (n = 5–7) as a major product. The structure of L1-Cd2Zn9 was finally determined by single crystal XRD analysis as Cd2Zn9(L1·H2O)6(H2O)18(OTf)22 having Cd(Mebpy)3 centers and Zn(Mebpy)2(H2O)2 centers. The space group was L1-Cd2Zn9 was R3c (#161) which was different from those of L1-Zn11 and L1-Cd11 (R-3c, #167). In the crystal state, Cd2Zn9(L1·H2O)6(H2O)18(OTf)22 has C3-symmetry. In its asymmetric unit, two distinct L1 moieties and five distinct metal centers were observed. In addition, the axial ligands of ZnII-porphyrin moieties were also different from those of L1-Zn11 and L1-Cd11. In the structure of Cd2Zn9(L1·H2O)6(H2O)18(OTf)22, axial ligand H2O coordinating to a ZnII-porphyrin moiety from the outer surface of the [Cd2Zn9L16]22+ skeleton, which was the opposite direction of the axial ligand MeOH coordinating to L1-Zn11 and L1-Cd11. Among the crystal structures of L1-Zn11, L1-Cd11, and L1-Cd2Zn9, the distance between two M(Mebpy)3 moieties of L1-Cd2Zn9 was the longest (L1-Cd2Zn9: 28.533 Å, L1-Zn11: 27.455 Å, L1-Cd11: 27.745 Å). the angles between three M(Mbpy)2(H2O)2 moieties of L1-Cd2Zn9 were slightly hindered (L1-Cd2Zn9: 㲃Zn1- Zn3-Zn2 = 66.55° and 72.25°, 㲃Zn3-Zn1-Zn3 = 93.54°, 㲃Zn3-Zn2-Zn3 = 87.59°). By elongating along the C3 axis, the structural distortion of the whole Cd2Zn9L16 skeleton caused by different ionic radii of ZnII and CdII seemed to be dissolved. The calculated inner volume of L1-Cd2Zn9 (734 Å3) was significantly larger than those of L1-Zn11 (511 Å3) and L1-Cd11 (509 Å3).
　Thus, this method provides an excellent way to construct heteronuclear self-assembled cage compounds with low symmetry. Notably, site-selective transmetalation just by utilizing chemically different metal centers and different binding ability of metals. In this study, construction of a D3-symmetric M11L6 skeleton from C4v-symmetric porphyrin ligand is a key approach to create heterometallic low symmetric supramolecular complexes by using symmetric ligand.

4. Conclusions and perspectives
　In this study, I have synthesized the C4v-symmetric porphyrin ligand L1 having 5’-methyl-2,2’-bipyridyl moieties to construct a variety of isostructural M11L6 cage complexes having metal-dependent characters, and have succeeded in construction two isostructural D3-symmetric M11L6 cage complexes, L1-Zn11 and L1-Cd11, by complexation with Zn(OTf)2 or Cd(OTf)2. Both of their structures in the solution state and the crystal state were fully characterized by NMR analyses, and Single crystal XRD analyses. By using different binding constants of Zn(bpy)n and Cd(bpy)n, transmetalation from L1-Cd11 to L1-Zn11 were investigated by NMR. In addition, construction of heterometallic L1-Cd2Zn9 via site-selective transmetalation at Cd(Mebpy)2A2 to Zn(Mebpy)2A2 was indicated by NMR and ESI- MS. Site-selective transmetalation of L1-Cd11 would be applicable to construct a variety of M11L6 cage complexes including heterometallic complexes which have desired metal dependent characters and applications. These results would give a new insight into the development of elaborate supramolecular systems from simple, symmetrical building blocks