A six-component metallosupramolecular pentagon via self-sorting

Abstract

The six-component pentagon P1 with its five dynamic vertices was conceived on the basis of three different orthogonal metal complex units in a 1-fold completive self-sorting of four linear ligands and two metal ions without using directional bonding.

---

Nature ingeniously uses self-assembly and self-sorting¹ to orchestrate the correct spatial and functionally active arrangement of multiple building blocks in superstructures that are elementary for life.^1b For instance, both the storage and utilisation of a cell’s genetic information require a specific base sequence of DNA and thus an error-free base pairing (= self-sorting).² In comparison to this impressive accomplishment, artificial supramolecular self-assembly^1,3 is presently reaching its limits at three- to five-component nanoarchitectures^4,5 with only a single discrete structure being known composed of more components.⁶

Herein, we report on the de novo design (Schemes 1 and 2) and synthesis of the unprecedented six-component metallosupramolecular pentagon P1. So far, pentagons have been developed as two- or three-component pentametallacycles^7,8 predominantly based on the directional bonding⁹ approach rendering the pentagonal architecture a rather difficult target due to a lack of 108° angles at metal centres.^3a,7 In contrast, the 1-fold completive^1c (= integrative)^4b self-sorting approach presented here enforces the pentagonal architecture P1 simply due to the implementation of three different dynamic complexation units C1–C3 in combination with entropic optimisation (Schemes 1 and 2).

Scheme 1 (a) 3-Fold completive self-sorting of the orthogonal complexes C1–C3 from an eight-component library. (b) Chemical structure of complexes C4 and C5.

Scheme 2 Synthesis of six-component pentagon P1.

To construct the odd number of vertices in P1, we chose to implement one homoleptic C2 and two heteroleptic cornerstones C1 and C3, the latter complexation units being derived from the HETPHEN (h̲e̲t̲eroleptic bisp̲h̲e̲n̲anthroline complex) and HETTAP (h̲e̲t̲eroleptic t̲erpyridine a̲nd p̲henanthroline complex) tool box.¹⁰ As a key challenge, the dynamic homoleptic coordination centre C2 should be fully orthogonal¹¹ to C1 and C3, because otherwise detrimental cross-talk will generate unsolicited structures. To preevaluate the required self-sorting,¹ the archetypical ligands 1–6 representing the interacting termini at the cornerstones were assessed in combination with suitable metal ions (i.e. Cu⁺ and Zn²⁺ ions) (Scheme 1).

At the start, we established the 2-fold completive self-sorted formation of both C1 = [Cu(1)(7)]⁺ and C3 = [Zn(3)(4)]²⁺ as dynamic HETPHEN¹² and HETTAP complexes from a seven component library (see ESI,† Fig. S21), i.e. from 1 : 3 : 4 : 5 : 6 : Cu⁺ : Zn²⁺ = 1 : 1 : 1 : 1 : 1 : 1 : 1, in a similar fashion to what has been observed in a related library.⁶

Formation of complex C2 = [Cu(2)₂]PF₆ (Scheme 1) may seem problematic at first due to the front shielding of 2-ferrocenyl-9-mesityl-[1,10]-phenanthroline (2), but the surprisingly high association constant log β_C2 = 11.0 ± 0.35 should warrant clean preparation of C2 from a 2 : 1 mixture of 2 and [Cu(CH₃CN)₄]PF₆ in CD₂Cl₂. Indeed, C2 formed readily as evidenced by ESI-MS (electrospray ionisation mass spectrometry), multi-nuclear NMR data and single-crystal X-ray analysis (see ESI†). The latter reveals Cu⁺ in a distorted tetrahedral geometry with the planes of both ligands being almost perpendicular (θ_z = 79°).¹³ In C2, the Cu–N_phen bond distances are in the range of 2.051(5)–2.063(6) Å.

Valuable information about C2 in solution was extracted from the ¹H-NMR. It revealed that the mesityl (x-H, δ = 7.06 ppm) and ferrocenyl (α-H, δ = 5.19 ppm) protons being homotopic in ligand 2 are diastereotopic in C2 (see ESI,† Fig. S14) as indicated by the two sets at δ = 5.60 and 6.45 ppm (for mesityl, i.e. x and x′-H) and δ = 5.62 and 5.01 ppm (for ferrocenyl, α and α′-H).

After proving the clean formation of C2, we decided to evaluate 2-fold completive self-sorting^1c scenarios in presence of C2, i.e. the orthogonal formation of C1 + C2 and C2 + C3 pairs, as a prerequisite for the required 3-fold completive self-sorting (Scheme 1). At first, we surveyed the stoichiometry dependence of the complexation involving a mixture of Cu⁺ and ligands 1 & 2. For example, addition of 1.0 equiv. of [Cu(CH₃CN)₄]PF₆ to a 1 : 2 mixture of 1 and 2 in CD₂Cl₂ endowed clean formation of a 1 : 1 mixture of C2 and ligand 1 (see ESI,† Fig, S16). In contrast, an equimolar mixture of 1, 2 and [Cu(CH₃CN)₄]PF₆ yielded both C2 (ca. 30%) and C4 = [Cu(1)(2)]PF₆ (ca. 15%) (Scheme 1b),‡ suggesting that the complex of both shielded phenanthrolines 1 and 2 is not kinetically impeded, as often observed with other bulky phenanthrolines (see ESI,† Fig. S17).¹⁰ Presumably, the higher front strain in C4 = [Cu(1)(2)]PF₆ with regard to that in C2 drives the selective formation of the 1 + C2 pair over the alternative 2 + C4 pair.¹⁴

To verify the relative energetics of C2 and C4, we added the slim ligand 5 and [Cu(CH₃CN)₄]PF₆ (each 1 equiv.) to a mixture of C2 + 1 (1 : 1) furnishing C5 = [Cu(1)(5)]PF₆ (Scheme 1b) without interference with C2 (see ESI,† Fig. S18), while the alternative pair C4 + [Cu(2)(5)](PF₆) (1 : 1) is not observed. Further addition of 1 equiv. of p-toluidine (6) to a 1 : 1 mixture of C2 and C5 completed the [Cu(1)]⁺ assisted formation of the iminopyridine ligand 7 (= (5)(6)–H₂O),¹² thereby furnishing a mixture of C2 and C1 (1 : 1) demonstrating their required orthogonality¹¹ (Scheme 1, Fig. S19, ESI†).

To test the interference-free formation of C2 and C3 (Scheme 1), we added 1 equiv. of C2 to a 1 : 1 : 1 mixture of 3, 4 and Zn(OTf)₂ and refluxed for 2 h in CH₂Cl₂. The ¹H-NMR and ESI-MS analysis of the reaction mixture confirmed their orthogonality (see ESI,† Fig. S20). Based on our prior knowledge,⁶ we suggest that the observed selectivity is largely guided by the preferred coordination number of zinc(II) (i.e. six) and copper(I) ions (i.e. four).^14,15 Indeed, one more time the additional Zn⋯OMe interaction present in C3⁶ provides a suitable pseudo-octahedral geometry to the Zn²⁺ ions, thus enthalpically enforcing the observed HETTAP complex C3.¹⁴

Considering the above insights, we finally examined the required 3-fold completive self-sorting process^1c (Scheme 1) using ligands 1–6 as well as Cu⁺ and Zn²⁺ ions. To our delight, full orthogonality of the complexes C1–C3 was established through ¹H-NMR and ESI-MS data (see ESI,† Fig. S23 and S41), thus providing a sound basis for the requested orthogonality of the dynamic corners in P1 (Scheme 2). The observed selectivity is achieved by the precise amalgamation of stoichiometry, steric and electronic effects, π–π interactions, metal-ion coordination specifics and metal-templated reversible imine bond formation in a one-pot process.

Besides the orthogonal formation of five dynamic cornerstones, the clean synthesis of P1 also requires full positional control, with each of the five metal–ligand corners finding their unique location in P1 (Scheme 2). Accordingly, the three ditopic ligands 8–10 were designed and prepared (see ESI†).

Bearing in mind that the pair C2 + C5 is orthogonal as well (C5 = [Cu(1)(5)]PF₆, vide supra), we chose first to synthesise the pentagon P2 = [Zn₂Cu₃(8)₂(9)(10)₂](OTf)₄(PF₆)₃ as precursor and then to prepare P1via a post-self-assembly modification approach,¹⁶i.e., P2 → P1, in presence of p-toluidine (6) (P1 : 6 = 1 : 2; Scheme 3a, step-I). This approach also facilitates our characterisation of P1 (vide infra). A retrosynthetic analysis of P2 suggests that it can be viewed as a combination of the angular subunit A = [Cu(8)₂](PF₆) and the tweezer subunit T = [Zn₂(9)(10)₂](OTf)₄ linked together by two dynamic C5-type copper(I) complexation sites (Scheme 3a, step-II).¹² As a result, we first inspected the reaction between ligand 8 and [Cu(CH₃CN)₄]PF₆ (2 : 1) in CD₂Cl₂ at 25 °C that furnished a clear red solution of A. Characterisation of A was established from the ESI-MS spectrum that showed one major peak at m/z = 2392.2 Da, corresponding to [Cu(8)₂]⁺ (Fig. S42, ESI†). A ¹H-NMR analysis of the reaction mixture substantiated the proposed C2-type binding motif (see Schemes 1 and 3a) in A by showing two sets of diastereotopically different ferrocenyl (α-H) protons of ligand 8 (Scheme 2), appearing at δ = 5.03 and 5.61 ppm (cf. in C2δ = 5.01 and 5.62 ppm), see Fig. 1a. In contrast, other diagnostic resonances, e.g. y and y′-H of the 2,9-dimesitylphenanthroline cores appear at a similar region to that of free ligand 8 (y and y′-H in A: δ = 6.92 and 6.94 ppm, and in 8: δ = 6.96 and 6.98 ppm), thus excluding the possibility of an alternative C4-type (vide supra) binding motif in A.

Scheme 3 (a) Retrosynthesis of pentagon P1. (b) Cartoon representation of the three different stereoisomers of P1.

Fig. 1 Partial ¹H NMR spectrum for comparison (400 MHz, CD₂Cl₂, 298 K) of (a) T, (b) A, (c) P2 and (d) P1.

The reaction of ligands 9, 10 and Zn(OTf)₂ (1 : 2 : 2), carried out at reflux temperature for 2 h in CH₂Cl₂/CH₃CN = 4 : 1 to destroy erroneously formed [Zn(terpy)₂]²⁺ complexes,¹⁷ quantitatively produced the HETTAP based tweezer T (Scheme 3) that was characterised from ¹H-NMR, ¹H–¹H COSY NMR, and ESI-MS data (see ESI†). For example, the ESI-MS spectrum of the crude reaction mixture exhibited two major peaks at m/z = 872.5 and 1382.8 Da for [Zn₂(9)(10)₂](OTf)_n⁽⁴⁻ⁿ⁾⁺ with n = 1, 2, respectively, that clearly supported the characterisation of T. The formation of HETTAP complex units, i.e. [Zn(10_phenAr2)(9_terpy)]²⁺ at each dynamic corner of T was further confirmed by the characteristic upfield shifts of the protons at the phenanthroline (e.g. OCH₃: δ = 2.95 and 2.97 ppm, see Fig. 1b) and the terpyridine protons (e.g. a′-H: δ = 7.63 ppm) in T, as compared to those in free 10 (OCH₃: δ = 3.71 and 3.73 ppm) and 9 (a′-H: δ = 8.87 ppm).^5c Notably, the aldehyde protons in T experience no upfield shift in comparison with that in ligand 10 (e.g. d-H in T: δ = 10.02 ppm, and d-H in 10: δ = 10.05 ppm). Thus, the terminal picolinaldehyde units are available for extra functionalisation.

As conceived, the angular subunit A (1 equiv.) with its two free 2,9-dimesitylphenanthroline terminals, tweezer T (1 equiv.) with its two picolinaldehyde units, and 2 equiv. of [Cu(CH₃CN)₄]PF₆ were cleanly reacted to the five-component supramolecular pentagon P2 (Scheme 3a, step-II) after heating to reflux for 2 h in CH₂Cl₂ (see ESI†). The characterisation and purity of the pentametallacycle P2 was verified from ESI-MS, ¹H-NMR, ¹H–¹H COSY NMR, DOSY NMR and elemental analysis. For example, the ESI-MS spectrum of the reaction mixture exhibited three major peaks at m/z = 1057.6, 1358.5 and 1861.2 Da, for [Zn₂Cu₃(8)₂(9)(10)₂] (OTf)_n⁽⁷⁻ⁿ⁾⁺ with n = 2, 3 and 4, respectively, that clearly supported the full characterisation of P2, while a single diffusion coefficient at D = 3.8 × 10⁻¹⁰ m² s⁻¹ in the DOSY NMR provided evidence for its purity (see ESI,† Fig. S33 and S44).

A comparison among the ¹H-NMR spectra of A, T and P2 (see ESI,† Fig. S31, Table S1) demonstrates that all the abovementioned diagnostic peaks for A and T complexation units show up also in identical regions for P2, thus confirming the existence of both C3- and C2-type corners in P2 (see Fig. 1a–c). In addition, the significant upfield shifts of the mesityl protons in P2 (y and y′-H: δ = 6.50 and 6.58 ppm) as compared to those in A (y and y′-H: δ = 6.92 and 6.94 ppm) and of aldehyde protons (d-H: δ = 9.47 and 9.45 ppm) as compared to those in T (d-H: δ = 10.02 ppm) further support the formation of two C5-type complex units. The observed 1 : 19 ratio (see ESI†) of the aldehyde protons in P2 proposes the existence of two§ diastereomers (Scheme 3b, see ESI,† Fig. S30), due to the three stereogenic axes at copper(I) centres.

Finally, the two C5-type complex units in P2 were interrogated in a post-self-assembly functionalisation as indicated in Scheme 3, step-I. Indeed, the six-component pentametallacycle P1 with its two constitutionally dynamic imine sites (Scheme 2) was cleanly obtained upon addition of 2 equiv. of 6 to a solution of P2 in CD₂Cl₂, as evidenced by ESI-MS (m/z = 1093.2, 1403.1 and 1920.6 Da for [Zn₂Cu₃(8)₂(9)(11)₂] (OTf)_n⁽⁷⁻ⁿ⁾⁺ with n = 2, 3 and 4, respectively), ¹H-NMR (Fig. 1d), DOSY NMR (D = 3.2 × 10⁻¹⁰ m² s⁻¹) and elemental analysis (see ESI†). To our satisfaction, full integrative self-sorting (Scheme 2) was equally effective when we examined the formation of P1 from its precursor ligands 6, 8–10 and metal ions (Cu⁺ and Zn²⁺) at correct stoichiometric onset (see ESI†). MM⁺ force field computations on P1 and P2 provided some insight in their structure as scalene pentagons. Taking the metal–metal distance as a measure, the five corners of P2 are separated by 1.51, 1.68, 1.74, 1.74 and 1.76 nm in the energy minimised structure and by 1.51, 1.68, 1.74, 1.74 and 1.75 nm in P1 (see ESI†).

In summary, the present study describes the clean and 1-fold completive (integrative) self-sorted synthesis of the unprecedented five- and six-component supramolecular pentagons P1 & P2. The generality of the present approach, devoid of control through directional bonding, is currently under investigation for the construction of 3D structures.

We are indebted to the DFG and Universität Siegen for financial support and to Dr S. Pramanik/Universität Siegen for his help in the synthesis of ligands 2 and 16 (precursor of 8).

Notes and references

1. (a) K. Osowska and O. Š. Miljanić, Synlett, 2011, 1643; (b) M. M. Safont-Sempere, G. Fernández and F. Würthner, Chem. Rev., 2011, 111, 5784; (c) M. L. Saha and M. Schmittel, Org. Biomol. Chem., 2012, 10, 4651.
2. J. D. Watson and F. H. C. Crick, Nature, 1953, 171, 737.
3. (a) R. Chakrabarty, P. S. Mukherjee and P. J. Stang, Chem. Rev., 2011, 111, 6810; (b) P. J. Stang, J. Am. Chem. Soc., 2012, 134, 11829.
4. (a) N. Christinat, R. Scopelliti and K. Severin, Angew. Chem., Int. Ed., 2008, 47, 1848; (b) W. Jiang and C. A. Schalley, Proc. Natl. Acad. Sci. U. S. A., 2009, 106, 10425; (c) M. M. J. Smulders, A. Jiménez and J. R. Nitschke, Angew. Chem., Int. Ed., 2012, 51, 6681; (d) M. Á. A. García and N. Bampos, Org. Biomol. Chem., 2013, 11, 27; (e) S. Li, J. Huang, F. Zhou, T. R. Cook, X. Yan, Y. Ye, B. Zhu, B. Zheng and P. J. Stang, J. Am. Chem. Soc., 2014, 136, 5908.
5. (a) M. Schmittel, B. He and P. Mal, Org. Lett., 2008, 10, 2513; (b) M. Schmittel and K. Mahata, Inorg. Chem., 2009, 48, 822; (c) K. Mahata, M. L. Saha and M. Schmittel, J. Am. Chem. Soc., 2010, 132, 15933; (d) M. L. Saha, S. Pramanik and M. Schmittel, Chem. Commun., 2012, 48, 9459.
6. M. L. Saha and M. Schmittel, J. Am. Chem. Soc., 2013, 135, 17743.
7. (a) S.-H. Hwang, P. Wang, C. N. Moorefield, L. A. Godínez, J. Manríquez, E. Bustos and G. R. Newkome, Chem. Commun., 2005, 4672; (b) L. Zhao, K. Ghosh, Y. Zheng, M. M. Lyndon, T. I. Williams and P. J. Stang, Inorg. Chem., 2009, 48, 5590.
8. (a) B. Hasenknopf, J.-M. Lehn, N. Boumediene, A. Dupont-Gervais, A. Van Dorsselaer, B. Kneisel and D. Fenske, J. Am. Chem. Soc., 1997, 119, 10956; (b) C. S. Campos-Fernández, B. L. Schottel, H. T. Chifotides, J. K. Bera, J. Bacsa, J. M. Koomen, D. H. Russell and K. R. Dunbar, J. Am. Chem. Soc., 2005, 127, 12909; (c) J.-F. Ayme, J. E. Beves, D. A. Leigh, R. T. McBurney, K. Rissanen and D. Schultz, Nat. Chem., 2012, 4, 15.
9. S. R. Seidel and P. J. Stang, Acc. Chem. Res., 2002, 35, 972.
10. M. L. Saha, S. Neogi and M. Schmittel, Dalton Trans., 2014, 43, 3815.
11. M. L. Saha, S. De, S. Pramanik and M. Schmittel, Chem. Soc. Rev., 2013, 42, 6860.
12. M. Schmittel, M. L. Saha and J. Fan, Org. Lett., 2011, 13, 3916.
13. J. F. Dobson, B. E. Green, P. C. Healy, C. H. L. Kennard, C. Pakawatchai and A. H. White, Aust. J. Chem., 1984, 37, 649.
14. M. L. Saha, K. Mahata, D. Samanta, V. Kalsani, J. Fan, J. W. Bats and M. Schmittel, Dalton Trans., 2013, 42, 12840.
15. M. L. Saha, J. W. Bats and M. Schmittel, Org. Biomol. Chem., 2013, 11, 5592.
16. (a) J. A. Thomas, Chem. Soc. Rev., 2007, 36, 856; (b) Y.-R. Zheng, W.-J. Lan, M. Wang, T. R. Cook and P. J. Stang, J. Am. Chem. Soc., 2011, 133, 17045.
17. [Zn(terpy)₂]²⁺ forms in rivalry to the desired HETTAP complexes.

---

Footnotes

† Electronic supplementary information (ESI) available: Experimental procedures and spectroscopic data of all new ligands and complexes, solid state structure of C2. CCDC 1013251. For ESI and crystallographic data in CIF or other electronic format see DOI: 10.1039/c4cc05465b

‡ In the ¹H-NMR spectrum (see ESI†), we observed additional signals representing the free ligand 1 (ca. 30%) and [Cu(1)](PF₆) (ca. 25%). Thus, the mixture contains C2 : C4 : 1 : [Cu(1)](PF₆) = 30 : 15 : 30 : 25.

§ Considering the structures, the isomers (P*, M*, P*) and (P*, P*, M*) could be magnetically equivalent, thus one cannot exclude the formation of all three possible diastereomers.

---