Neutral versus polycationic coordination cages: a comparison regarding neutral guest inclusion

Abstract

A neutral self-assembled container synthesized from a concave π-extended tetrathiafulvalene (exTTF) ligand and the cis-Pd(dctfb)₂(cod) complex (dctfb = 3,5-dichloro-2,4,6-trifluorobenzene; cod = 1,5-cyclooctadiene) is described. This molecular host exhibits a good binding ability for fused polyaromatic substrates. The corresponding inclusion properties are compared with those of a previously described analogous octacationic cage, offering therefore the opportunity to address the effect of the cavity charge state over the binding of neutral molecules.

---

The preparation of molecular cages able to encapsulate ionic or neutral organic guests constitutes a major challenge for various applications ranging from selective depollution, reactivity in confined environments, or guest transport. In this context, the coordination-driven self-assembly strategy has been successfully used to build, in a straightforward way, more and more sophisticated metalla-cages,¹ including electro-active ones.² Being most often prepared by the reaction of a polypyridyl ligand (typically di-, tri- or tetra-pyridyl) with a metal cation, they generally correspond to polycationic cages.^1,3 On this basis, we showed recently that the cationic coordination-cage M₄L₂⁸⁺ (Scheme 1 and Fig. 2b) is able to bind the B₁₂F₁₂²⁻ guest in a reversible way through a redox-driven assembly/disassembly process.⁴ Host–guest interactions are by nature more challenging to address in the case of neutral guests, which renders the design of appropriate receptors more delicate. On these grounds, an efficient redox control of the binding mode^5,6 has been recently demonstrated from a tetracationic organic covalent receptor. When considering neutral guests, the polycationic character of the cages usually obtained through coordination-driven self-assembly may hamper both the kinetics and the thermodynamics of the binding, in particular because of the presence of highly competitive counter anions which obstruct the cavity.^1m,7 On this basis, the design of neutral receptors through the self-assembled metal-driven strategy appears promising. Examples of neutral discrete metalla-assemblies are known and are synthetized either from the reaction of an anionic donating ligand (typically bearing carboxylate groups) and a cationic metal center,⁸ or from the reaction between a neutral ligand and a neutral metallic precursor.^1m,7 Some data related to the binding ability of such a neutral coordination cage for neutral guests have already been provided.⁷ Nevertheless, to the best of our knowledge, no comparative experimental study is available which addresses in a quantitative manner the binding affinity of a neutral guest for a charged coordination cage versus a neutral one, whereas this issue has already been reported in the case of an ionic guest.⁹ Such a comparative study, which is of importance for a better understanding of intermolecular forces which govern the recognition phenomena, needs to define a proper cavity whose geometry and size are the same for both the neutral and cationic states.

Scheme 1 Synthesis of cage M₄L₂ and M₄L₂⁸⁺. (i) Acetone, rt, 48 h, 87%; (ii) nitromethane, 40 °C, 5 min, 83%.

We describe herein a metal-directed neutral self-assembled cage M₄L₂ constructed from a tetra-pyridyl exTTF ligand and a neutral cis-oriented square planar bis-aryl palladium complex (cis-Pd(dctfb)₂(cod)).¹⁰ The resulting assembly offers a unique opportunity to quantitatively compare the binding affinities of planar polyaromatic guests for two similar cavities (M₄L₂ and M₄L₂⁸⁺) (Scheme 1) which essentially differ in the charge state of the metal center and therefore in the presence, or not, of counter anions.¹¹

The tetrapyridyl-exTTF ligand L (Scheme 1) was synthesized in three steps from 2,3,6,7-tetrahydroxyanthracene-9,10-dione, in a 57% overall yield as previously described.⁴ The presence of the four peripheral triethylene glycol (TEG) chains on the anthracene core ensures the solubility of the resulting cages in various solvents. The self-assembly reaction of ligand L and complex cis-Pd(dctfb)₂(cod) was carried out for 48 h in acetone at room temperature. The reaction converged into a single discrete compound M₄L₂ (Scheme 1) that could be isolated in 87% yield via simple filtration.

The complex M₄L₂ presents a high field shift of the α and β pyridyl signals in the ¹H NMR spectrum compared to ligand L (Fig. 1a, b and Fig. S1, ESI†), as expected from the coordination to the metal center. ¹H DOSY NMR exhibits only one alignment of signals with an extracted D value of 3.22 × 10⁻¹⁰ m² s⁻¹ in CD₃NO₂/CDCl₃ (1/1) and 3.35 × 10⁻¹⁰ m² s⁻¹ in CDCl₃ (Fig. S3 and S4, ESI†) and confirms the formation of only one discrete species. The corresponding hydrodynamic radius¹² estimated from the Stokes–Einstein equation¹³ (T = 298 K) is of 11.4 Å, a value which is in accordance with the formation of a M₄L₂ species. ESI-FTICR mass spectrometry experiments carried out with M₄L₂ are silent due to the lack of charges. Taking advantage of the well-known ability of TEG chains to bind alkaline cations, an excess of KOTf (8 equivalents) was added to a solution of M₄L₂ (C = 2 × 10⁻³ M, CH₂Cl₂/CH₃NO₂ 5/5) before ionization. Thus, the generated ionic species could be studied and confirmed the stoichiometry of the complex, as illustrated by characteristic multi-charged peaks [M₄L₂·(KOTf)₄-4OTf]⁴⁺ (m/z = 1213.94) and [M₄L₂·(KOTf)₄-3OTf]³⁺ (m/z = 1668.58) as well as by the good accordance between the experimental and the theoretical isotopic patterns (Fig. S5, ESI†).

Fig. 1 ¹H NMR, downfield region (CD₃NO₂/CDCl₃ 5/5), related to M₄L₂ study (left) and M₄L₂⁸⁺ (right): (a) L, (b) M₄L₂, (b′) M₄L₂⁸⁺ (c) coronene, (d) ¹H NMR and corresponding ¹H DOSY NMR of a mixture of M₄L₂ and coronene (1/1, C = 2 × 10⁻³ M), (d′) ¹H NMR and corresponding ¹H DOSY NMR of a mixture of M₄L₂⁸⁺ and coronene (1/1, C = 2 × 10⁻³ M).

Single crystals of M₄L₂ could be grown by slow diffusion of acetonitrile in a chloroform solution and an XRD experiment confirmed the formation of a M₄L₂ cage (Fig. 2a and Fig. S30, ESI†) which exhibits an internal ovoid cavity of ca. 15 Å length over 13 Å width, close to that of M₄L₂⁸⁺ (ca. 15.5 Å × 11.5 Å) (Fig. 2b).¹⁴ This shape similitude offers a unique opportunity to compare the ability of these respectively neutral and charged metalla-cages for encapsulating a neutral guest, since they essentially differ by (i) their charge state on the metal and by (ii) the presence (M₄L₂⁸⁺) or absence (M₄L₂) of counter anions. In a previous work, we demonstrated that a polycationic cage, similar to M₄L₂⁸⁺ but devoid of peripheral TEG chains, allows the inclusion of perylene according to a 1 : 1 stoichiometry in nitromethane.¹⁵

Fig. 2 X-Ray crystal structures of the host structure M₄L₂ (a) and M₄L₂⁸⁺ (b). For M₄L₂⁸⁺, triflate anions located inside the cavity are represented (in grey). Colors used for both M₄L₂ and M₄L₂⁸⁺: exTTF skeleton (green), PEG chains (blue), Pd complex (orange).

We therefore studied the inclusion of different neutral fused polyaromatic guests into the ovoid cavities of M₄L₂ and M₄L₂⁸⁺ by ¹H NMR and ¹H DOSY NMR. A solvent mixture (CD₃NO₂/CDCl₃ 1/1) has to be used in order to ensure the solubility of all the species and to allow a comparative study. As an illustrative example of the respective behavior of the neutral and the charged cages upon encapsulation, the ¹H NMR binding study of coronene is shown in Fig. 1. Importantly, whereas only a slight difference is observed for M₄L₂⁸⁺ upon addition of one equivalent of the guest (Fig. 1b′ and d′), the ¹H NMR signals of metalla-cage M₄L₂ and of coronene are significantly high-field shifted (Fig. 1b–d). The binding is accompanied in the case of M₄L₂ by a significant decrease of the coronene diffusion coefficient D value (D = 4.21 × 10⁻¹⁰ m² s⁻¹vs. D_free = 12.40 × 10⁻¹⁰ m² s⁻¹) in the ¹H DOSY NMR spectrum (Fig. 1c and d). In contrast, the coronene D value remains almost unchanged in the presence of M₄L₂⁸⁺ (D = 11.40 × 10⁻¹⁰ m² s⁻¹vs. D_free = 12.40 × 10⁻¹⁰ m² s⁻¹). As already demonstrated for the charged cage, the formation of a 1/1 host–guest complex was also found via a Job Plot analysis for M₄L₂ (Fig. S32, ESI†). From these ¹H DOSY NMR data, K_a binding constants of 2.6 × 10⁴ (M₄L₂) and 63 (M₄L₂⁸⁺) could be calculated (see ESI† for details), meaning a ratio of ca. 400 in favor of the neutral cage. The complexation of other planar aromatic guests was studied and their respective binding constants are compiled in Table 1. As it can be noted from these values, the binding abilities increase with the guest size and, remarkably, the neutral M₄L₂ cage systematically presents a higher affinity (one to several orders of magnitude higher) for the tested neutral polyaromatic guests. These results clearly demonstrate the positive effect of the absence of charges and of anions on the periphery of the ovoid cavity, for the encapsulation of neutral guests.¹⁶ Increasing the proportion of nitromethane (a poor solvent for the guest under consideration) in the solvent mixture results in a significant increase of the binding constant (K_a = 1.1 × 10⁵ with M₄L₂ and coronene). Single crystals of the coronene⊂M₄L₂ inclusion complex could be grown at room temperature from the latter solvent mixture. The XRD experiment confirms the encapsulation of one coronene unit located in the middle of the cavity defined by the exTTF concave fragments containing both 1,3-dithiol rings (Fig. 3). It is worth noting that in order to maximize the host–guest interactions, the Pd–Pd distances are significantly reduced upon complexation related to the free cage M₄L₂, as can be seen from the cavity size which changes from 15 Å × 13 Å (Fig. 2a) to 16 Å × 10.5 Å in coronene⊂M₄L₂,¹⁴ illustrating the unexpectedly flexible character of the cavity. Binding studies of coronene with the M₄L₂⁸⁺ assembly subjected to a TfO⁻ anion exchange (i.e. BF₄⁻ and B₁₂F₁₂²⁻) were carried out and monitored by ¹H DOSY NMR in CD₃NO₂/CDCl₃ 8/2. K_a values of 200 and 58 were found for BF₄⁻ and B₁₂F₁₂²⁻ respectively (170 for TfO⁻) (Fig. S12, S27 and S28, ESI†) showing that the nature of the counter anion does influence the binding of the coronene guest and justify the superiority of the neutral host over the polycationic one. Finally, an inclusion control study was led by ¹⁹F DOSY NMR experiments with the B₁₂F₁₂²⁻ anion (Fig. S29–S31, ESI†). K_a values of 25 and >10⁶ were obtained for M₄L₂ and M₄L₂⁸⁺ respectively, confirming that the polycationic cage binds much more strongly anionic guests than the neutral one does.

Table 1 Calculated binding constants K_a for cages M₄L₂⁸⁺ and M₄L₂ towards planar polyaromatic guests as estimated from ¹H DOSY NMR experiments in CD₃NO₂/CDCl₃ at 298 K (C = 2 × 10⁻³ M)

Guest	Solvent (CD3NO2/CDCl3)	M4L2^8+	M4L2
a No binding could be detected under these conditions. b C = 10^−3 M.
Coronene	8/2	170^b	1.1 × 10^5
	5/5	63	2.6 × 10^4
Perylene	5/5	96	4.0 × 10^3
Triphenylene	5/5	14	2.8 × 10^2
Pyrene	5/5		1.7 × 10^2

---

Fig. 3 X-Ray crystal structure of coronene⊂M₄L₂ (a) side view, (b) top view. Colors used: exTTF skeleton (green), PEG chains (blue), Pd complex (orange), coronene (grey).

In summary, an electron-rich exTTF-based M₄L₂ neutral cage which exhibits a good ability to bind planar polyaromatic guests in a 1 : 1 stoichiometry was synthesized. This neutral assembly could be compared to the previously described analogous M₄L₂⁸⁺ system, which essentially differs in the charges on the metal and therefore by the presence of counter anions surrounding the cavity. This comparative experimental study affords first quantitative evidence of the superiority of neutral hosts for the binding of neutral guests (with K_a values which are one to several orders of magnitude higher in this study), an observation which is of importance for the rational design of strong host/guest systems in future. On these grounds, the design of new neutral metalla-cages appears highly relevant.⁷ In particular, the present work illustrates the promising perspectives offered by the square planar cis-Pd(dctfb)₂(cod) complex¹⁰ for preparing neutral cages upon interaction with various poly-pyridyl ligands, in a straightforward way and under mild conditions. Therefore, the scope of this approach can in principle be extended to most of the abundant poly-pyridyl ligands available in the literature and previously used to build polycationic cages.

This work has been supported by the ANR JCJC program (ANR-14-CE08-0001 BOMBER). The authors gratefully acknowledge the MENRT for a PhD grant (VC). They also acknowledge Dr I. Freuze and B. Siegler (PIAM, Univ. Angers), as well as Drs F. Aubriet and V. Carré (CNRS-Univ. Lorraine (TGE “FT-ICR”)) for their assistance in spectroscopic analyses and finally Drs P. Fertey and S. Ravy (Synchrotron Soleil – CRISTAL beamline – project 20130173).

Notes and references

1. (a) For recent reviews see: N. Ahmad, H. A. Younus, A. H. Chughtai and F. Verpoort, Chem. Soc. Rev., 2015, 44, 9; (b) A. J. McConnell, C. S. Wood, P. P. Neelakandan and J. R. Nitschke, Chem. Rev., 2015, 115, 7729; (c) S. Zarra, D. M. Wood, D. A. Roberts and J. R. Nitschke, Chem. Soc. Rev., 2015, 44, 419; (d) M. Han, D. M. Engelhard and G. H. Clever, Chem. Soc. Rev., 2014, 43, 1848; (e) S. Mukherjee and P. S. Mukherjee, Chem. Commun., 2014, 50, 2239; (f) T. R. Cook, V. Vajpayee, M. H. Lee, P. J. Stang and K.-W. Chi, Acc. Chem. Res., 2013, 46, 2464; (g) K. Harris, D. Fujita and M. Fujita, Chem. Commun., 2013, 49, 6703; (h) A. Mishra, S. C. Kang and K.-W. Chi, Eur. J. Inorg. Chem., 2013, 5222; (i) M. M. J. Smulders, I. A. Riddell, C. Browne and J. R. Nitschke, Chem. Soc. Rev., 2013, 42, 1728; (j) M. D. Ward and P. R. Raithby, Chem. Soc. Rev., 2013, 42, 1619; (k) H. Amouri, C. Desmarets and J. Moussa, Chem. Rev., 2012, 112, 2015; (l) T. R. Cook, Y.-R. Zheng and P. J. Stang, Chem. Rev., 2012, 113, 734; (m) R. Chakrabarty, P. S. Mukherjee and P. J. Stang, Chem. Rev., 2011, 111, 6810; (n) L. R. MacGillivray, Angew. Chem., Int. Ed., 2012, 51, 1110; (o) Y. Inokuma, M. Kawano and M. Fujita, Nat. Chem., 2011, 3, 349; (p) S. De, K. Mahata and M. Schmittel, Chem. Soc. Rev., 2010, 39, 1555; (q) P. Jin, S. J. Dalgarno and J. L. Atwood, Coord. Chem. Rev., 2010, 254, 1760; (r) Y.-F. Han, W.-G. Jia, W.-B. Yu and G.-X. Jin, Chem. Soc. Rev., 2009, 38, 3419; (s) B. H. Northrop, Y.-R. Zheng, K.-W. Chi and P. J. Stang, Acc. Chem. Res., 2009, 42, 1554; (t) P. J. Stang, J. Org. Chem., 2009, 74, 2; (u) B. Therrien, Eur. J. Inorg. Chem., 2009, 2445; (v) M. Yoshizawa, J. K. Klosterman and M. Fujita, Angew. Chem., Int. Ed., 2009, 48, 3418; (w) M. W. Cooke, D. Chartrand and G. S. Hanan, Coord. Chem. Rev., 2008, 252, 903; (x) S. J. Dalgarno, N. P. Power and J. L. Atwood, Coord. Chem. Rev., 2008, 252, 825; (y) B. H. Northrop, D. Chercka and P. J. Stang, Tetrahedron, 2008, 64, 11495.
2. V. Croué, S. Goeb and M. Sallé, Chem. Commun., 2015, 51, 7275.
3. For anionic metalla-cages, see also: (a) D. L. Caulder and K. N. Raymond, Acc. Chem. Res., 1999, 32, 975; (b) Z. J. Wang, C. J. Brown, R. G. Bergman, K. N. Raymond and F. D. Toste, J. Am. Chem. Soc., 2011, 133, 7358.
4. V. Croué, S. Goeb, G. Szalóki, M. Allain and M. Sallé, Angew. Chem., Int. Ed., 2016, 55, 1746.
5. M. Frasconi, I. R. Fernando, Y. Wu, Z. Liu, W.-G. Liu, S. M. Dyar, G. Barin, M. R. Wasielewski, W. A. Goddard and J. F. Stoddart, J. Am. Chem. Soc., 2015, 137, 11057.
6. For a redox-based switching of guest binding performed upon stimulating the guest, see e.g. W.-Y. Sun, T. Kusukawa and M. Fujita, J. Am. Chem. Soc., 2002, 124, 11570.
7. P. Thanasekaran, C.-H. Lee and K.-L. Lu, Coord. Chem. Rev., 2014, 280, 96.
8. N. Ahmad, A. H. Chughtai, H. A. Younus and F. Verpoort, Coord. Chem. Rev., 2014, 280, 1.
9. J. Kuwabara, C. L. Stern and C. A. Mirkin, J. Am. Chem. Soc., 2007, 129, 10074.
10. L. J. Marshall and J. de Mendoza, Org. Lett., 2013, 15, 1548.
11. Note in addition that, whereas M₄L₂ involves two dctfb (3,5-dichloro-2,4,6-trifluorobenzene) as Pd co-ligands, M₄L₂⁸⁺ involves dppf.
12. Calculated from the experiment led in a pure solvent (CDCl₃) (Fig. S3, ESI†), for which the viscosity of the medium is known.
13. (a) Y. Cohen, L. Avram and L. Frish, Angew. Chem., Int. Ed., 2005, 44, 520; (b) L. Avram and Y. Cohen, Chem. Soc. Rev., 2015, 44, 586.
14. The cavity size values are obtained by measuring the distance between (i) the opposite central anthracene rings on one hand and (ii) two opposite palladium atoms on the other hand.
15. S. Bivaud, S. Goeb, V. Croué, P. I. Dron, M. Allain and M. Sallé, J. Am. Chem. Soc., 2013, 135, 10018.
16. Performing the binding studies at a similar ionic strength for M₄L₂⁸⁺ and M₄L₂, by addition of 8 equivalents of tetrabutylammonium trifluoromethane-sulfonate in the latter case, does not change the result.

---

Footnote

† Electronic supplementary information (ESI) available: Detailed synthesis protocols, experimental methods, X-ray diffraction studies and additional spectroscopic data for ligand L and complexes M₄L₂ and M₄L₂⁸⁺. CCDC 1441971 and 1441826. For ESI and crystallographic data in CIF or other electronic format see DOI: 10.1039/c6cc04610j

---