A conformationally persistent pseudo-bicyclic guanidinium for anion coordination as stabilized by dual intramolecular hydrogen bonds

Abstract

The first example of a pseudo-bicyclic guanidinium ligand is reported. When bound to an anion, the N,N′-bis(2-pyridyl)guanidinium cation persistently adopts the planar α,α conformation featuring intramolecular N⋯H–N–H⋯N hydrogen bonds in the solid state, which facilitates crystallization of sulphate from aqueous mixtures of anions.

---

Guanidiniums are excellent oxoanion receptors¹ owing to their ability, like the related urea family of receptors,² to direct two hydrogen bonds in bidentate fashion along an oxoanion O–X–O edge.^3,4 In the case of guanidiniums, the presence of the cationic charge further provides for coulombic strengthening of the binding interaction as well as gives the ligand designer the means to build in charge-complementarity as an additional selectivity principle. One problematic issue with simple substituted guanidinium-based receptors is their innate conformational flexibility, which enables them to exist in several different conformations.⁵ Reflecting a generic challenge in ligand design,^6–8 restricting such conformational freedom is thus necessary to control the directionality and cooperativity of their N–H donor groups (Fig. 1). This has been accomplished very effectively by employing a bicyclic framework, which has since often been used in guanidinium based anion-receptor design.^9–12 By analogy, it occurred to us to ask whether intramolecular hydrogen bonding could be employed to achieve a pseudo bicyclic guanidinium frame and how such an arrangement would be reflected in the structure of the resulting anion complexes.

Fig. 1 The three major conformations of N,N′-disubstituted guanidinium cations. α and β refer to the orientation of the R group relative to the NH₂⁺ group.

A current research direction in our group is to employ simple guanidinium ligands for selective separation of oxoanions, such as sulphate, via crystallization.¹³ A major challenge with this approach is to identify guanidinium cations that form relatively insoluble sulphate salts for effective separation from water. This is a difficult proposition, as most guanidinium salts have high aqueous solubilities. For example, the solubility of guanidinium sulphate in water is about 10 M.¹² Nevertheless, some isolated examples of highly insoluble guanidinium sulphate salts are known, such as 2-aminoperimidine sulphate, or more recently, glyoxal bis(amidiniumhydrazone) sulfate.^12,14 One common structural feature in these guanidinium salts is the presence of a rigid and planar extended π cation that can stack favourably in the crystalline state.

Herein we describe the simple N,N′-bis(2-pyridyl)guanidinium ligand (1) that is designed to achieve a planar α,α conformation through the formation of a pseudo-bicyclic motif via intramolecular N⋯H–N–H⋯N hydrogen bonding (Fig. 2). This pseudo-bicyclic motif is found to persist across a series of crystalline guanidinium salts and facilitates the selective crystallization of sulfate from aqueous anion mixtures. The N,N′-bis(phenyl)guanidine (2), which cannot attain planarity due to the steric clashing of the aromatic-C2 and NH₂⁺ protons, serves as a control for comparison with 1.

Fig. 2 The N,N′-bis(2-pyridyl)guanidinium (1) can have two intramolecular hydrogen bonds that enhance conformational rigidity compared to bis(phenyl)guanidinium (2), which can undergo free rotation about the guanidinium C–N bonds.

Synthesis of 1 is achieved in two steps from 2-aminopyridine to give the free guanidine ligand (see ESI†). Addition of a stoichiometric amount of the corresponding acid, followed by vapor diffusion of ether into methanol, yielded the sulfate, chloride, and nitrate salts of 1. The single-crystal X-ray structures of these salts (Fig. 3) show that 1 persistently adopts the planar, pseudo-bicyclic α,α conformation throughout the series via the formation of two intramolecular N–H⋯N hydrogen bonds between the guanidinium NH₂ and the pyridine groups.¹⁵ The remaining two N–H hydrogen bond donors chelate the anion in either a 1 : 1 (Cl⁻ and NO₃⁻), or a 2 : 1 (SO₄²⁻) fashion. The sulphate anion also has four water molecules bound in the equatorial plane, which complete the 12 hydrogen bonds of the coordination sphere of sulphate. From the chloride structure, the preference of 1 for anion binding via the α,α conformation may be seen to extend beyond oxoanions, raising the question of the origin of the stability and generality of this conformer.

Fig. 3 Crystal structures of 1 bound to various anions. (a) Side view and top view of 1 bound to sulphate, which is additionally hydrogen bonding to four water molecules (water protons could not be located). (b) 1 bound to chloride. (c) 1 bound to nitrate.

In direct contrast, the previously reported crystal structures of 2 with a variety of anions showed this guanidinium cation is generally non-planar and lacks any conformational preference.^16,17 Accordingly, in the case of 2–NO₃⁻, the guanidinium binds nominally in an α,α conformation but is twisted largely out of plane. The sulfate salt of 2 has the guanidinium existing as a mixture of conformers with no clear conformational preference. A Cambridge Structural Database (CSD Ver. 1.17) search for 2 yielded 24 hits (excluding disorder and errors); the salts show no preference for any particular orientation, and both types of N–H groups function as hydrogen-bond donors to anions. Our ligand also compares well with previously reported structures of true bicyclic guanidiniums bound to various anions.^18–20 These bicyclic systems, like our 1-complexes, maintain the planar guanidinium group observed only in the α,α conformation. One distinct characteristic of the pseudo bicyclic structure seen in the 1-complexes is the planarity of the entire guanidinium molecule, while the aliphatic backbones of the true bicyclic systems are often twisted out of plane.

In order to gain a better understanding of the observed structures, we employed electronic-structure calculations and Natural Bond Orbital (NBO) analysis²¹ to assess the relative stabilities of the major conformations of 1 and 2 and their anion-binding preferences using the 1 : 1 complexes with nitrate as representative models. Binding energies for 1 : 1 1–NO₃⁻ complexes and the relative stabilities of the three major conformations of free 1 and 2 calculated at the ωB97X-D/6-311++G(3df,3pd) level of theory are given in Fig. 4 and Table 1, respectively.²² The calculations are in accord with the structural evidence showing that 1 prefers to bind anions in the α,α conformation, while 2 has no preference. In this regard, the calculated 3.7 kcal mol⁻¹ stabilization of 1–α,α-NO₃⁻ vs. 1–α,β-NO₃⁻ appears to be significant, especially in view of the RT ln(2) statistical (entropic) advantage enjoyed by the α,β conformation.

Fig. 4 Structures and binding energies (kcal mol⁻¹) for 1 : 1 nitrate anion–ligand complexes in the α,α and α,β binding conformations obtained after geometry optimization at the ωB97X-D/6-311++G(3df,3pd) level of theory. Binding energies are obtained with respect to a free ligand in the most stable α,β conformation.

Table 1 Relative stabilities of the three major conformations of free cationic ligands 1 and 2^a (kcal mol⁻¹)

Ligand	α,α	α,β	β,β
a Conformations are defined in Fig. 1. Relative energies are obtained at the ωB97X-D/6-311++G(3df,3pd) level. Zero-point energies and thermal corrections to enthalpy are included at the B3LYP/6-311++G(d,p) level.^23
1	2.20	0	15.7
2	0.34	0	4.50

---

The persistence of the α,α form of 1 in its anion complexes may be explained in that the α,α form delivers the strongest pair of hydrogen bond donor groups and thus the most stable complexes. Stronger hydrogen bonding occurs via the PyrN–H group vs. the NH₂ group due to the electron-withdrawing ability of the pyridyl substituent, making the NH proton ostensibly more acidic. This argument is supported by the NBO analysis,²¹ which quantifies hydrogen bonding strength by a leading two-electron intermolecular donor–acceptor interaction (n_B → σ^*_HA) between the lone pair n_B of the Lewis base B and the unfilled hydrogen antibonding orbital σ^*_HA of the Lewis acid AH.²⁴ We find (Table S3 of ESI†) that the leading donor–acceptor interaction for the PyrN–H⋯N hydrogen bond in 1–α,β (26.2 kcal mol⁻¹) is much stronger than that for the HN–H⋯N hydrogen bond (19.7 kcal mol⁻¹). Moreover, when the NH₂ group forms two hydrogen bonds in 1–α,α the leading n_N → σ^*_HNH donor–acceptor interaction (per bond) becomes even weaker (16.9 kcal mol⁻¹). Thus, in the absence of an external hydrogen bond acceptor, the stronger PyrN–H donors will favor a stronger interaction with the other pyridine N atom, stabilizing the α,β conformer. By contrast, in the presence of strongly coordinating anions, the stronger PyrN–H donors will favor a symmetric planar α,α conformation. Consistent with this notion, the HN–H⋯O hydrogen bond (1.660 Å) in 1–α,β-NO₃⁻ is substantially longer than the PyrN–H⋯O hydrogen bond (1.600 Å) in 1–α,β-NO₃⁻ and the HN–H⋯O (1.632 Å) hydrogen bond in 2–α,β-NO₃⁻. Similarly, in the free guanidinium 1, the two H–N–H donor hydrogen bond distances are calculated to be 1.882 Å for the α,α conformation and 1.847 Å (HN–H) and 1.794 Å (pyN–H) for the α,β conformation.

Unlike the conformational preferences upon anion coordination, the non-symmetric α,β conformer is the global minimum for both ligands in their unbound free state (Table 1). We note that the 2–α,α form is only marginally less stable than the global minimum, while 1–α,α is considerably less stable. Due to lack of any intramolecular hydrogen bonding, the π–π stacked β,β conformation of 1 (see ESI†) is greatly destabilized compared to the other forms, while the β,β conformation of 2 is only 4.5 kcal mol⁻¹ above the global minimum (Table 1) and can still be accessible under standard conditions, as evident from several crystal structures containing this conformation (such as 2₂–SO₄²⁻). While the symmetrical α,α form of 1 is thus not the energetically stable form of the free ligand, the theoretical study shows that the pseudo-bicyclic scaffold of 1, unlike 2, confers strong directionality upon coordination to an anion and restricts the number of conformations accessible at room temperature. Similar results are seen for the 1₂–SO₄²⁻ complex (see ESI†). The results thus elucidate the persistence of the α,α form of 1 on anion binding but naturally raise questions regarding the conformations that exist in the solution state and the attendant issue of preorganization in binding reactions. Such new questions entail aspects of solvation and entropy that are the subject of our ongoing investigations. What we can already see at this point, though, is that by comparison to the fused-ring bicyclic systems²⁵ a pseudo-bicyclic approach allows additional conformational freedom that must be taken into account.

The preference of 1 for a planar, conformationally locked structure, as observed in other guanidinium sulfate salts of low aqueous solubility, prompted us to evaluate the potential of 1 for sulfate separation by crystallization from water. Mixing equimolar aqueous solutions of 1–Cl⁻ and sodium sulphate resulted in the immediate formation of a white precipitate, which was identified as 1₂–SO₄²⁻(H₂O)₇ by single-crystal and powder X-ray diffraction. No precipitate formed with other anions, including I⁻, Cl⁻, and NO₃⁻ (Fig. 5). The gravimetrically measured solubility of 1₂–SO₄²⁻(H₂O)₇ at 20 °C is 2.5 mg mL⁻¹ (10 mM). On the other hand, no precipitate formed with sulphate or other anions when 2 was used instead. As well as the rigid and planar extended π stacks, intramolecular hydrogen bonding has been shown to increase the lipophilicity of a molecule, which may contribute in part to the observed insolubility of the 1₂–SO₄²⁻(H₂O)₇ complex.²⁶

Fig. 5 (A) 1 + NaI, (B) 1 + NaCl, (C) 1 + NaNO₃, (D) 1 + Na₂SO₄, (E) 2 + Na₂SO₄.

To examine any structural impact due to crystallization from water vs. the methanol/diethylether system used initially, a single crystal of 1₂–SO₄²⁻(H₂O)₇ was obtained via slow evaporation of a saturated solution from water. The X-ray structure showed the expected 2 : 1 sulphate binding, though with slightly different packing than that obtained from methanol/diethylether (Fig. 6). Notably, the guanidinium cation adopts, once again, a perfectly planar pseudo-bicyclic conformation, with the cations stacking along the crystallographic b axis with the shortest interplanar distance of 3.3 Å measured between the central C atom of guanidinium and the C2 carbon of the pyridine ring. The overall crystal is composed of alternating hydrophobic layers of stacked guanidinium cations and hydrophilic sulphate-water layers (Fig. 6). Powder X-ray diffraction of the precipitate from water matched well the simulated powder pattern from the single crystal, confirming that the bulk precipitate was indeed the 2 : 1 complex with sulphate, as determined by single-crystal diffraction.

Fig. 6 X-ray crystal structure of 1₂–SO₄²⁻(H₂O)₇, obtained by crystallization from water, showing alternating guanidinium stacks and sulphate-water layers.

This study demonstrates a persistent pseudo-bicyclic structure for the novel guanidinium receptor 1 in anion binding. Both X-ray crystallography and DFT calculations show that the guanidinium 1 prefers to bind anions in its planar α,α conformation as rigidified by N⋯H–N–H⋯N intramolecular hydrogen bonds. In direct contrast, guanidinium 2 cannot form such intramolecular hydrogen bonds and exhibits no consistent structural preference. The ability of 1 to selectively crystallize with sulphate suggests an immediate use in anion separation. Moreover, with its strong inherent binding, positive charge, and hydrogen-bond induced planarity, 1 represents a unique and promising scaffold for the design of selective oxoanion receptors. Current efforts in our research group are directed towards fully characterizing this receptor's ability to bind sulphate in solution, the nature of the intramolecular hydrogen bond in solution, potential alternatives to pyridine N-donors as intramolecular hydrogen bond receptors, as well as the synthesis and application of lipophilic derivatives for extraction.

Acknowledgements

This research was sponsored by the U.S. Department of Energy, Office of Science, Basic Energy Sciences, Chemical Sciences, Geosciences, and Biosciences Division. This research used resources of the National Energy Research Scientific Computing Center, a DOE Office of Science User Facility supported by the Office of Science of the U.S. Department of Energy under Contract No. DE-AC02-05CH11231.

Notes and references

1. F. P. Schmidtchen and M. Berger, Chem. Rev., 1997, 97(5), 1609–1646.
2. V. Blažek Bregović, N. Basarić and K. Mlinarić-Majerski, Coord. Chem. Rev., 2015, 295, 80–124.
3. B. A. Moyer, R. Custelcean, B. P. Hay, J. L. Sessler, K. Bowman-James, V. W. Day and S. Kang, Inorg. Chem., 2013, 52(7), 3473–3490.
4. P. Blondeau, M. Segura, R. Pérez-Fernández and J. de Mendoza, Chem. Soc. Rev., 2007, 36, 198–210.
5. C. Schmuck and H. Y. Kuchelmeister in Artificial Receptors for Chemical Sensors, ed. V. M. Mirsky and A. K. Yatsimirsky, Wiley-VCH, 2010, pp. 270–313.
6. D. J. Cram, J. Inclusion Phenom., 1988, 6, 397–413.
7. F. P. Schmidtchen, Coord. Chem. Rev., 2006, 250(23–24), 2918–2928.
8. P. A. Gale, N. Busschaert, C. J. E. Haynes, L. E. Karagiannidis and I. L. Kirby, Chem. Soc. Rev., 2014, 43(1), 205–241.
9. H. Kurzmeier and F. Schmidtchen, J. Org. Chem., 1990, 55, 749–3755.
10. J. Sánchez-Quesada, C. Seel, P. Prados and J. de Mendoza, J. Am. Chem. Soc., 1996, 118, 277–278.
11. K. Kazuya and Y. Inoue, J. Am. Chem. Soc., 2003, 125, 421.
12. D. Kneeland, K. Ariga, V. Lynch, C. Huang and E. Anslyn, J. Am. Chem. Soc., 1993, 115, 10042.
13. R. Custelcean, N. J. Williams and C. A. Seipp, Angew. Chem., Int. Ed., 2015, 54(36), 10525–10529.
14. W. I. Stephen, Anal. Chim. Acta, 1970, 50, 413.
15. C. Lee, W. Yang and R. G. Parr, Phys. Rev. B: Condens. Matter Mater. Phys., 1988, 37, 785–789.
16. J. A. Paixao, P. S. P. Silva, A. M. Beja, M. R. Silva and L. A. Vega, Acta Crystallogr., Sect. C: Cryst. Struct. Commun., 1998, 54, 805.
17. A. M. Beja, J. A. Paixao, P. S. P. Silva, L. A. Vega, E. M. Gomes and J. Z. Martin-Gil, Kristallografiya, 1998, 213, 655.
18. P. Leibnitz, G. Reck, H. J. Pietzsch and H. Spies, Forschungszent. Rossendorf, [Ber.] FZR, 2001, 311, 34.
19. A. Gleich, F. P. Schmidtchen, P. Mikulcik and G. Muller, Chem. Commun., 1990, 55.
20. U. Wild, P. Roquette, E. Kaifer, J. Mautz, O. Hubner, H. Wadepohl and H. J. Himme, Eur. J. Inorg. Chem., 2008, 1248.
21. A. E. Reed, L. A. Curtiss and F. Weinhold, Chem. Rev., 1988, 88, 899–926.
22. J. D. Chai and M. Head-Gordon, Phys. Chem. Chem. Phys., 2008, 10, 6615–6620.
23. A. D. Becke, Chem. Phys., 1993, 98, 5648–5652.
24. F. Weinhold and C. Landis, Valency and Bonding, Cambridge University Press, Cambridge, 2005.
25. V. D. Jadhav, E. Herdtweck and F. P. Schmidtchen, Chem.–Eur. J., 2008, 14, 6098–6107.
26. B. Kuhn, P. Mohr and M. Stahl, J. Med. Chem., 2010, 53, 2601.

---

Footnote

† Electronic supplementary information (ESI) available. CCDC 1404793–1404796. For ESI and crystallographic data in CIF or other electronic format see DOI: 10.1039/c5ra21864k

---