DOI: 10.1002/chem.201402752
Full Paper
& Supramolecular Chemistry
Halogen- and Hydrogen-Bonding Catenanes for Halide-Anion Recognition Lydia C. Gilday and Paul D. Beer*[a]
Abstract: Halogen-bonding (XB) interactions were exploited in the solution-phase assembly of anion-templated pseudorotaxanes between an isophthalamide-containing macrocycle and bromo- or iodo-functionalised pyridinium threading components. 1H NMR spectroscopic titration investigations demonstrated that such XB interpenetrated assemblies are more stable than analogous hydrogen bonding (HB) pseudorotaxanes. The stability of the anion-templated halogen-
Introduction The essential biological, medicinal, chemical, and industrial functions of negatively charged species, as well as the continuing harmful environmental impact of certain anions, is well documented.[1] Consequently, the recognition and sensing of target anions is a central goal of anion supramolecular chemistry, and through the imaginative manipulation of a variety of complementary non-covalent interactions, a plethora of efficient synthetic anion receptors have been developed.[2] However, the challenge to mimic the impressive levels of anion recognition demonstrated by biological systems still remains. Inspired by nature’s sulfate- and phosphate-binding proteins, which effect strong and selective anion complexation in elaborate organised binding domains shielded from competing water molecules,[3] we have employed anions to template the construction of mechanically bonded molecules.[4] These rotaxane and catenane interlocked host structures each contain a unique, three-dimensional cavity between the individual components leading to notable enhancement of anion binding affinities by complementary electrostatics and hydrogen-bonding (HB) interactions in competitive solvent media compared with the macrocycle and axle parent derivatives. Halogen bonding (XB) is the attractive, highly directional, non-covalent interaction between an electron-deficient halogen atom and a Lewis base.[5] In spite of their stringent directionality and strength comparable with ubiquitous HB, the exploitation of XB in solution-phase applications is only just be[a] Dr. L. C. Gilday, Prof. P. D. Beer Chemistry Research Laboratory, Department of Chemistry University of Oxford, Mansfield Road, Oxford, OX1 3TA (UK) Fax: (+ 44) 1865-272690 E-mail: [email protected]
bonded pseudorotaxane architectures was exploited in the preparation of new halogen-bonding interlocked catenane species through a Grubbs’ ring-closing metathesis (RCM) clipping methodology. The catenanes’ anion recognition properties in the competitive CDCl3/CD3OD 1:1 solvent mixture revealed selectivity for the heavier halides iodide and bromide over chloride and acetate.
ginning to be realised.[6] In an important step in advancing solution-phase anion recognition, we have prepared the first XB rotaxane, which through the combined use of both halogenand hydrogen-bonding interactions, binds iodide more strongly than either chloride or bromide anions,[6c] and a XB catenane, which is selective for chloride and bromide over a variety of other anions.[6d] Herein, we describe the anion-templated construction of new catenane host systems designed to exploit halogen- and hydrogen-bonding for anion recognition.[7] Figure 1 schematically shows the synthetic strategy employed to prepare the target halogen-/hydrogen-bonding catenanes. Anion-templated threading of an acyclic positively charged XB-donor component through a suitable isophthala-
Figure 1. Cartoon representation of orthogonal pseudorotaxane species, in which X is a halogen atom.
mide-containing macrocycle gives an orthogonal interpenetrative assembly stabilised by halogen- and hydrogen-bonding– anion interactions, which upon cyclisation will afford the new XB–HB catenanes.
Supporting information for this article is available on the WWW under http://dx.doi.org/10.1002/chem.201402752. Chem. Eur. J. 2014, 20, 8379 – 8385
8379
2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Full Paper Results and Discussion Anion-templated pseudorotaxane assembly Macrocycle 1[8] and functionalised pyridinium derivatives 2– 4·BF4 were prepared by using multistep synthetic procedures reported previously (Figure 2).[6e] Pseudorotaxane assembly between macrocycle 1 and potential pyridinium threading compounds 2–4·BF4 was studied initially using 1H NMR titration experiments in CDCl3, in the presence and absence of a halide-anion template (Scheme 1).
Figure 2. Structures of macrocycle 1 and threading components 2–4·BF4.
Aliquots of pyridinium threading components as their tetrafluoroborate salts were added to an equimolar solution of macrocycle 1 and tetrabutylammonium (TBA) halide salt in [D]chloroform. Addition of bromo- and iodopyridinium derivatives 2·BF4 and 3·BF4 to macrocycle 1 in the presence of bromide or iodide caused significant shifts in a number of signals in the respective 1H NMR spectrum (Figure 3). Downfield perturbations of the macrocycle cavity proton signal c and amide protons signal d are indicative of halide anion binding within the macrocycle’s isophthalamide binding cleft. Importantly, a diagnostic upfield shift and splitting of the signal corresponding to the macrocycle hydroquinone protons g and h was observed, which is consistent with aromatic donor–acceptor intermolecular interactions between the macrocycle’s electron-rich hydroquinone units and the electron-deficient halo-pyridinium group, indicative of pseudorotaxane formation. Also, a downfield shift in the signal corresponding to the N-methyl pyridi-
Scheme 1. Pseudorotaxane formation between macrocycle 1 and threading components 2–4·BF4 in the presence of TBA·A.
Figure 3. Partial 1H NMR spectra of: a) macrocycle 1; b) macrocycle 1 + Br ; c) pseudorotaxane; d) bromopyridinium thread 3·Br (500 MHz, CDCl3, 298 K). Chem. Eur. J. 2014, 20, 8379 – 8385
www.chemeurj.org
8380
2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Full Paper nium protons a was seen, which results from hydrogen bonding to the macrocycle’s polyether oxygen atoms. Upfield shifts in the signals corresponding to the halo-pyridinium thread’s protons b and g are a result of both aromatic donor–acceptor interactions and the bromide anion-template binding to the macrocycle’s isophthalamide domain. The observed chemical shift perturbations of the hydroquinone protons with ten equivalents of threading component revealed that, in general, bromide anion templation causes the largest magnitude of perturbation with relatively smaller changes with iodide or in
Macrocycle 1 + Br Macrocycle 1 + I
Macrocycle 1
larger magnitude for interpenetration with the XB halopyridinium derivatives than with the equivalent protic pyridinium species (Kapp = 1500 vs. 340 m1 with iodide), which demonstrates that halogen-bonding, halide-anion-templated pseudorotaxanes are more stable than their hydrogen-bonding counterparts. Of the interpenetrative assemblies stabilised by XB interactions, the iodopyridinium derivative is more favourable than the analogous bromopyridinium threading component (Kapp = 1500 and 780 m1 , respectively, with iodide). This is attributed to the greater halogen-bond-donor ability of the more polarisable iodine atom compared with bromine. It is noteworthy that pseudorotaxane assembly is stronger when a bromide anion template is employed instead of iodide: Kapp (2·A) = 2300 versus 780 m1. This suggests that the relatively smaller bromide anion is more suitably sized to bind to the macrocycle’s isophthalamide anion recognition cleft than the larger iodide anion.
Dd [ppm][a] 0.39
Kapp [m1][c] 2300
Dd [ppm][a] 0.19
Kapp [m1][c] 780
Dd [ppm][a] 0.16
Kapp [m1][c] 270
Synthesis of catenanes
0.22
3300[d]
0.18
1500
0.17
250[e]
0.31[b]
1000
0.15
340
0.13
220[e]
Table 1. Observed chemical shift perturbations (Dd) of hydroquinone protons and apparent association constants, Kapp (m1), for macrocycle 1 and threading components 2·BF4, 3·BF4 or 4·BF4 in the presence and absence of halide anions.
2·BF4 (X = Br) 3·BF4 (X = I) 4·BF4 (X = H)
[a] After ten equivalents. [b] After seven equivalents. [c] Apparent association constants were determined by monitoring the perturbations of the macrocycle hydroquinone protons upon addition of threading species to an equimolar mixture of macrocycle and TBA·A. Error < 10 %. [d] Error < 25 %. [e] Error < 15 %; CDCl3, 298 K.
the absence of a halide templating anion (Table 1). Monitoring the perturbations of the macrocycle’s hydroquinone protons as a function of concentration of threading component gave the titration curves shown in Figure S11 in the Supporting Information, and analysis of the titration data by using the winEQNMR2[9] computer program gave apparent association constants, Kapp,[10] for pseudorotaxane formation reported in Table 1. Importantly, in all cases, the most stable pseudorotaxane assemblies were found in the presence of bromide and iodide halide-anion templates. In the absence of a halide-anion template, the interpenetrative association with the halo- and proticfunctionalised pyridinium tetrafluoroborate salts are all much weaker and are of similar modest stability. The apparent association constants are of Chem. Eur. J. 2014, 20, 8379 – 8385
Encouraged by the favourable anion-templated association between macrocycle 1 and macrocycle precursor compounds 2·BF4 and 3·BF4, in the presence of either bromide or iodide respectively, the synthesis of new anion-templated, halogenbonding [2]catenanes using a Grubbs’-catalysed ring-closing metathesis (RCM) reaction was undertaken (Scheme 2). The required bromide and iodide salts of the vinyl-appended macrocycle precursor components were prepared by anion exchanging the non-coordinating tetrafluoroborate anion for the
Scheme 2. Synthesis of new anion-templated, halogen-bonding catenanes 5·Br and 6·I.
www.chemeurj.org
8381
2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Full Paper
Figure 4. Partial 1H NMR spectra of: a) bromopyridinium macrocycle precursor 2·Br; b) catenane 5·Br; and c) macrocycle 1 (500 MHz, CDCl3, 298 K).
appropriate halide-anion template. Macrocycle 1 and 1.1 equivalents of either 2·Br or 3·I in anhydrous dichloromethane were stirred at room temperature for four days in the presence of Grubbs’ second-generation catalyst (10 % by weight). Purification using repeated preparative thin-layer chromatography, the target catenanes 5·Br and 6·I were each isolated in 8 % yield.[11] Evidence of successful interlocked structure formation was provided by high-resolution electrospray mass spectrometry and 1H NMR spectroscopy. The cationic molecular-ion peaks of both catenanes were observed in the mass spectrum, which suggests the presence of a single interlocked species, and the high-resolution mass spectra revealed that the isotopic distribution of the catenanes 5·PF6 and 6·PF6 are in good agreement with the theoretical spectra (see Figures S5 and S10 in the Supporting Information). The 1H NMR spectrum of bromopyridinium-containing catenane 5·Br together with those for the component macrocycle and bis-vinyl-appended precursor 2·Br for comparison are shown in Figure 4. Important spectral changes, which are diagnostic of catenane formation, include the disappearance of the vinylic multiplet m and the convergence of multiplet l into a pseudo-singlet in the catenane spectrum, which are attributed to cyclisation of the acyclic precursor. Crucially, the upfield shift and splitting of the macrocycle hydroquinone protons g and h is indicative of aromatic donor–acceptor interactions between the electron-rich macrocycle hydroquinone units and the electron-poor pyridinium group, confirming the two components are interlocked. Two dimensional 1H–1H rotating-frame Overhauser effect spectroscopy (ROESY) NMR spectroscopy provided further evidence of the interlocked nature of the [2]catenane 5·Br (Figure 5). The 1H–1H ROESY spectrum displays several through space cross-coupling interactions between the two components of the catenane. Important interactions include 1 (b!g) and 2 (b!h), which confirm the electron-rich hydroquinone units and the electron-poor pyridinium moiety are close in Chem. Eur. J. 2014, 20, 8379 – 8385
www.chemeurj.org
space, as would be expected when the macrocycle components are interlocked with respect to one another. Interaction 3 (a!h) also demonstrates the hydroquinone and pyridinium aromatic units are in close proximity. To highlight the important role of the halide-anion template in aiding formation of the orthogonal interpenetrative assembly of the precursor components, the synthesis of the catenanes was also attempted with 2·BF4 and 3·BF4. Although peaks at m/z 1238.4 and 1286.4 corresponding to [2]catenanes 5·BF4 and 6·BF4 were detected by electrospray mass spectrometry, in spite of numerous attempts, the catenanes could not be isolated, which suggests they were formed in negligible yield in the absence of a halide template.[12]
Anion-recognition properties of catenanes To study the anion-recognition properties of the catenanes, the respective templating halide guests were exchanged for the non-coordinating hexafluorophosphate anion by repeatedly washing a dichloromethane solution of the catenane with aqueous ammonium hexafluorophosphate. The halide and acetate anion-binding properties of the catenanes 5·PF6 and 6·PF6 were investigated by using 1H NMR spectroscopic titration experiments with TBA anion salts in the competitive solvent mixture [D]chloroform/[D4]MeOH 1:1. Upon addition of halide anions to both catenane species, significant downfield chemical shift perturbations of the ortho-isophthalamide proton c and pyridinium proton b were observed with increasing anion concentration. By contrast, addition of acetate induced considerably smaller shifts (Ddppm = 0.03 ppm for 6·PF6), which indicated that no acetate complexation occurred in this solvent mixture. WinEQNMR analysis of the titration curves obtained for chloride, bromide, iodide, and acetate binding (Figure 6) gave 1:1 stoichiometric association constants reported in Table 2.
8382
2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Full Paper
Figure 5. 1H–1H ROESY NMR spectrum of catenane 5·Br. Assignable inter-component through-space interactions are highlighted (500 MHz, CDCl3, 298 K).
Table 2. Association constants, Ka [m1], for addition of various anions to catenanes 5·PF6 and 6·PF6. Anion
Catenane 5·PF6 Ka [m1][a]
Catenane 6·PF6 Ka [m1][a]
Cl Br I OAc
130 200 220 NB
160 230 340 NB
[a] All anions were added as their TBA salt. Association constants were determined by monitoring the perturbations of isophthalamide cavity proton c. Error < 10 %. NB: no evidence of binding; Ddppm too small to allow an accurate association constant to be obtained. CDCl3/CD3OD 1:1, 298 K.
Figure 6. Normalised titration curves for addition of various anions to catenanes 5·PF6 (top) and 6·PF6 (bottom) obtained by monitoring the macrocycle cavity proton c. CDCl3/CD3OD 1:1, 298 K. Square data points represent experimental data; continuous lines represent theoretical binding isotherms. Chem. Eur. J. 2014, 20, 8379 – 8385
www.chemeurj.org
The association constant values reveal both catenanes 5·PF6 and 6·PF6 exhibited halide-anion recognition, but do not bind the basic acetate anion, which suggests that the oxoanion is too large and of the wrong shape to penetrate the catenanes’ binding pocket. Both interlocked hosts complex halides chloride, bromide and iodide with moderate affinity in this competitive solvent system. The iodopyridinium-containing catenane 6·PF6 is a superior halide anion host to the bromopyridinium-functionalised catenane 5·PF6 (Ka = 340 vs. 220 m1 for iodide association), which reflects the greater halogen-bond-donor ability of the iodine atom compared with bromine. Interestingly, for both catenanes the strongest host–guest association is with iodide, followed by bromide and chloride. This is the reverse of what was expected on anion-basicity grounds and may be attributed to a combination of factors including an interlocked bind-
8383
2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Full Paper ing domain more suited to the larger halide anion, a higher penalty for desolvation of chloride than bromide or iodide in this competitive protic solvent mixture and, possibly, a greater degree of charge-transfer-type halogen-bonding interaction, between the respective catenane’s bromo- and iodopyridinium halogen-bond-donor motif and the larger halide anion.
yield (25.8 mg, 34.3 mmol). 1H NMR (300 MHz, CDCl3): d = 8.24 (s, 2 H; pyridinium-H3&H5), 6.84 (s, 8 H; hydroquinone-H), 5.87–6.00 (m, 2 H; C=CH), 5.17–5.34 (m, 4 H; C=CH2), 5.16 (s, 4 H; CH2O), 4.24 (m, 3 H; CH3), 3.99–4.14 ppm (m, 20 H; CH2); 13C{1H} NMR (75.5 MHz, CDCl3): d = 156.3, 153.2, 152.5, 143.5, 134.5, 129.0, 117.3, 115.6, 115.3, 72.3, 70.7, 68.5, 68.0, 67.4, 40.5 ppm; MS (ESI): m/z: calcd for C34H43BrNO8 : 672.2168 [MBr] + ; found: 672.2163.
Conclusion
Compound 3·I
Through the combination of halogen- and hydrogen-bonding non-covalent interactions, bromide and iodide anions have been demonstrated to template the formation of interpenetrative assemblies between an isophthalamide-containing macrocycle and bromo- or iodo-functionalised pyridinium threading components. Pseudorotaxane formation was more favourable when XB was employed—specifically, the more stable bromide- and iodide-templated interpenetrated assemblies were found with the iodopyridinium species, which reflects the greater halogen-bond-donor ability of the more polarisable iodine atom. The stability of the anion-templated XB pseudorotaxanes was exploited in the construction of two new XB catenane architectures, the halide- and acetate-anion recognition properties of which were probed in competitive solvent mixture. The strength of anion binding was found to be greater with the iodo-functionalised halogen-bond-donor catenane with both catenanes displaying anion affinities iodide > bromide > chloride @ acetate. The integration of XB donor components into interlocked structural frameworks is continuing in our laboratories.
Compound 3·BF4 (51.2 mg, 63.4 mmol) was dissolved in CH2Cl2 (30 mL) and washed with 1 m NH4I (aq. 10 10 mL), H2O (2 10 mL), dried over anhydrous MgSO4, filtered and the solvent removed in vacuo to give 3·I as a pale orange oily solid in 99 % yield (53.1 mg, 62.7 mmol). 1H NMR (300 MHz, CDCl3): d = 8.44 (s, 2 H; pyridiniumH3&H5), 6.85 (s, 8 H; hydroquinone-H), 5.87–6.00 (m, 2 H; C=CH), 5.19–5.34 (m, 4 H; C=CH2), 5.02 (s, 4 H; CH2O), 4.02–4.13 (m, 21 H; CH2 & CH3), 3.75–3.8 ppm (m, 4 H; CH2); 13C{1H} NMR (75.5 MHz, CDCl3): d = 154.2, 153.3, 152.6, 135.0, 134.5, 117.4, 115.7, 115.5, 72.3, 70.9, 68.6, 68.0, 67.6, 61.5, 53.4, 40.8 ppm; MS (ESI): m/z calcd for C34H43INO8 : 720.2028 [MI] + ; found 720.2040.
Experimental Section General considerations Commercially available reagents and solvents were used without further purification unless otherwise stated. When anhydrous solvents were used, they were purged with N2, passed through a MBraun MPSP-800 column and then used immediately. Et3N was distilled from, and stored over, KOH. H2O was de-ionised and microfiltered by using a Milli-Q Millipore machine. TBA salts and Grubbs’ II catalyst were stored in a vacuum desiccator prior to use. Column chromatography was performed on silica gel (particle size 40–63 mm), preparative TLC was performed on 20 20 cm plates, with a silica layer of thickness 1 mm. 1H, 13C, 19F and 31P NMR spectra were recorded on Varian Mercury-vx 300 spectrometer or on Varian Unity Plus or Bruker Avance III 500 instruments. Mass spectra were obtained by using a Micromass LCT Premier XE spectrometer and accurate masses were obtained to four decimal places by using a Bruker MicroTOF spectrometer. Macrocycle 1 and functionalised pyridinium threading components 2–4·BF4 were prepared according to reported procedures.
Synthetic procedures Compound 2·Br Compound 2·BF4 (31.8 mg, 41.8 mmol) was dissolved in CH2Cl2 (30 mL) and washed with 1 m NH4Br (aq. 10 10 mL), H2O (2 10 mL), dried over anhydrous MgSO4, filtered, and the solvent was removed in vacuo to give 2·Br as a pale orange oily solid in 82 % Chem. Eur. J. 2014, 20, 8379 – 8385
www.chemeurj.org
Catenane 5·Br Macrocycle 1 (50 mg, 80 mmol) and 2·Br (68.4 mg, 90.8 mmol) were dissolved in anhydrous CH2Cl2 (5 mL) and stirred at RT under N2 for 30 min. Grubbs’ II (7 mg, 10 % by wt) was added, and the reaction mixture was stirred at RT under N2 for 3 d, after which time more Grubbs’ II (3.5 mg) was added, and the reaction mixture was stirred for a further 24 h. The solvent was removed in vacuo, and the crude residue was purified by preparative thin layer chromatography (95:5 CH2Cl2/MeOH and 3:2 MeCN/CH2Cl2) to give 5·Br as a white solid in 8 % yield (9.0 mg, 6.8 mmol). 1H NMR (300 MHz, CDCl3): d = 9.30 (s, 1 H; isophthaloyl-H2), 8.84 (br s, 2 H; amide-NH), 8.33 (d, 3J = 7.6 Hz, 2 H; isophthaloyl-H4&H6), 8.03 (s, 2 H; pyridinium-H3&H5), 7.59 (t, 3J = 7.6 Hz, 1 H; isophthaloyl-H5), 6.86 (d, 3J = 8.9 Hz, 4 H; hydroquinone-H), 6.80 (d, 3J = 8.9 Hz, 4 H; hydroquinone-H), 6.31 (d, 3J = 8.8 Hz, 4 H; hydroquinone-H), 6.08 (d, 3J = 8.8 Hz, 4 H; hydroquinone-H), 5.73 (s, 2 H; alkene-CH), 4.53 (s, 3 H; NCH3), 4.26 (br s, 4 H; CH2), 4.16 (br s, 4 H; CH2), 4.07–4.09 (m, 4 H; CH2), 3.97–3.98 (m, 12 H; CH2), 3.72–3.74 (m, 4 H; CH2), 3.59–3.66 (m, 12 H; CH2), 3.46–3.50 ppm (m, 8 H; CH2); 13C{1H} NMR (75.5 MHz, CDCl3): d = 166.7, 153.6, 153.4, 153.2, 152.2, 151.3, 138.5, 133.8, 131.8, 130.0, 128.9, 128.9, 128.7, 124.6, 115.9, 115.5, 114.9, 114.4, 71.1, 71.1(sic), 70.5, 70.0, 69.7, 68.9, 68.3, 67.8, 66.5, 39.8, 29.7 ppm; MS (ESI): m/z calcd for C64H77BrN3O17: 1238.4431 [MBr] + ; found: 1238.4457.
Catenane 5·PF6 Compound 5·Br (8.50 mg, 6.16 mmol) was dissolved in CH2Cl2 (20 mL), washed with 0.1 m NH4PF6 (aq. 10 10 mL), H2O (2 10 mL), dried over MgSO4, filtered, and the solvent was removed in vacuo to give 5·PF6 as an off-white solid in 81 % yield (7.66 mg, 5.53 mmol). 1H NMR (300 MHz, CDCl3): d = 8.25 (d, 3J = 7.8 Hz, 2 H; isophthaloyl-H4&H6), 8.12 (s, 1 H; isophthaloyl-H2), 7.61 (t, 3J = 7.8 Hz, 1 H; isophthaloyl-H5), 7.48 (s, 2 H; pyridinium-H3&H5), 7.13 (s, 2 H; amide-NH), 6.79–6.83 (s, 8 H; hydroquinone-H), 6.35 (d, 3J = 8.8 Hz, 4 H; hydroquinone-H), 6.14 (d, 4 H; 3J = 8.8 Hz, hydroquinone-H), 5.78 (s, 2 H; alkene-CH), 4.66 (s, 3 H; NCH3), 4.22 (s, 4 H; CH2), 4.00–4.06 (m, 16 H; CH2), 3.90–3.91 (m, 4 H; CH2), 3.74–3.75 (m, 4 H; CH2), 3.63 (s, 4 H; CH2), 3.59 (s, 4 H; CH2), 3.50 (s, 8 H; CH2), 3.45 ppm (s, 4 H; CH2); 13C{1H} NMR (125.8 MHz, CDCl3): d = 166.7, 160.6, 153.9, 153.5, 153.2, 152.4, 151.5, 142.7, 140.4, 134.5, 131.7,
8384
2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Full Paper 129.3, 129.0, 115.7, 115.4, 114.9, 71.1, 70.5, 70.1, 70.1(sic), 69.1, 68.2, 68.1, 67.1, 66.7, 66.5, 50.9, 39.6, 29.7 ppm; 19F NMR (282.4 MHz, CDCl3): d = 71.1 ppm (d, J = 712 Hz, PF6); 31P NMR (202.4 MHz, CDCl3): d = 144.0 ppm (sept, J = 712 Hz, PF6); MS (ESI): m/z calcd for C64H77BrN3O17: 1238.4431 [MPF6] + ; found: 1238.4423.
Catenane 6·I Macrocycle 1 (71 mg, 120 mmol) and compound 3·I (92.5 mg, 109 mmol) were dissolved in anhydrous CH2Cl2 (5 mL) and stirred at RT under N2 for 30 min. Grubbs’ II (9.3 mg, 10 % by wt.) was added and the reaction mixture stirred at RT under N2 for 3 d, after which time more Grubbs’ II (4.5 mg) was added, and the reaction mixture was stirred for a further 24 h. The solvent was removed in vacuo, and the crude residue was purified by preparative thin-layer chromatography (95:5 CH2Cl2/MeOH and 3:2 MeCN/CH2Cl2) to give 6·I as a white solid in 8 % yield (13.5 mg, 9.44 mmol). 1H NMR (500 MHz, CDCl3): d = 9.02 (s, 1 H; isophthaloyl-H2), 8.44 (br s, 2 H; amide-NH), 8.29 (d, 3J = 7.6 Hz, 2 H; isophthaloyl-H4&H6), 7.95 (s, 2 H; pyridinium-H3&H5), 7.58 (t, 3J = 7.6 Hz, 1 H; isophthaloyl-H5), 6.83 (s, 8 H; hydroquinone-H), 6.43 (d, 3J = 8.8 Hz, 4 H; hydroquinone-H), 6.12 (d, 3J = 8.8 Hz, 4 H; hydroquinone-H), 4.67 (s, 4 H; CH2), 4.30– 4.32 (m, 4 H; CH2), 4.00–4.07 (m, 16 H; CH2), 3.79 (s, 3 H; NCH3) 3.73– 3.75 (m, 4 H; CH2), 3.60 (s, 4 H; CH2), 3.55 (s, 4 H; CH2), 3.46–3.47 (m, 4 H; CH2), 3.40 ppm (s, 4 H; CH2); 13C{1H} NMR (125.8 MHz, CDCl3): d = 166.8, 153.6, 153.3, 152.4, 151.8, 151.4, 135.2, 134.1, 131.9, 128.9, 124.4, 120.3, 115.8, 115.5, 115.0, 114.8, 71.19, 71.12, 70.5, 70.1, 69.8, 69.0, 68.3, 68.0, 67.0, 66.8, 66.7, 39.4, 29.7 ppm; MS (ESI): m/z calcd for C64H77IN3O17: 1286.4292 [MI] + ; found: 1286.4300.
Catenane 6·PF6 Compound 6·I (13.5 mg, 9.44 mmol) was dissolved in CH2Cl2 (20 mL), washed with 0.1 m NH4PF6 (aq. 10 10 mL), H2O (2 10 mL), dried over MgSO4, filtered, and the solvent was removed in vacuo to give 6·PF6 as an off-white solid in 67 % yield (9.13 mg, 6.37 mmol). 1H NMR (500 MHz, CDCl3): d = 8.23 (d, 3J = 7.6 Hz, 2 H; isophthaloyl-H4&H6), 8.15 (s, 1 H; isophthaloyl-H2), 7.63 (2 H, s, pyridinium-H3&H5), 7.60 (1 H, t, 3J = 7.6 Hz, isophthaloyl-H5), 7.13 (br s, 2 H; amide-NH), 6.79–6.88 (m, 8 H; hydroquinone-H), 6.41 (d, 3J = 8.7 Hz, 4 H; hydroquinone-H), 6.15 (d, 3J = 8.7 Hz, 4 H; hydroquinone-H), 6.79 (s, 2 H; alkene-CH), 4.68 (s, 4 H; CH2), 4.23 (s, 4 H; CH2), 4.01–4.07 (m, 16 H; CH2), 3.93–3.94 (m, 4 H; CH2), 3.86 (s, 3 H; NCH3), 3.74–3.76 (4 H, m, CH2), 3.62 (4 H, s, CH2), 3.57 (4 H, s, CH2), 3.48 (s, 4 H; CH2), 3.42 ppm (s, 4 H; CH2); 13C{1H} NMR (125.8 MHz, CDCl3): d = 166.6, 153.4, 153.2, 152.5, 151.5, 150.6, 134.9, 134.5, 131.7, 129.3, 129.0, 115.6, 115.5, 115.0, 114.9, 71.1, 70.5, 70.0, 69.9, 69.1, 68.2, 68.1, 66.9, 66.7, 66.4, 34.1, 29.0 ppm; 31P NMR (MHz, CDCl3): d = 144.0 ppm (sept, J = 716 Hz, PF6); MS (ESI): m/z calcd for C64H77IN3O17: 1286.4292 [MPF6] + ; found: 1286.4287.
Acknowledgements We thank the EPSRC and the Department of Chemistry, University of Oxford for funding. Keywords: anions · catenanes · halogen bonding · molecular recognition · supramolecular chemistry
Chem. Eur. J. 2014, 20, 8379 – 8385
www.chemeurj.org
[1] a) J. L. Sessler, P. A. Gale, W.-S. Cho, Anion Receptor Chemistry, The Royal Society of Chemistry, Cambridge, UK, 2006; b) A. Bianci, K. BowmanJames, E. Garcia-EspaÇa, Anion Coordination Chemistry, Wiley-VCH, Weinheim, 2011. [2] a) P. D. Beer, P. A. Gale, Angew. Chem. 2001, 113, 502 – 532; Angew. Chem. Int. Ed. 2001, 40, 486 – 516; b) J. M. Llinares, D. Powell, K. Bowman-James, Coord. Chem. Rev. 2003, 240, 57 – 75; c) E. J. O’Neil, B. D. Smith, Coord. Chem. Rev. 2006, 250, 3068 – 3080; d) R. Martnez-MÇez, F. Sancenn, Coord. Chem. Rev. 2006, 250, 3081 – 3093; e) T. Gunnlaugsson, M. Glynn, G. M. Tocci, P. E. Kruger, F. M. Pfeffer, Coord. Chem. Rev. 2006, 250, 3094 – 3117; f) C. R. Rice, Coord. Chem. Rev. 2006, 250, 3190 – 3199. [3] a) J. W. Pflugrath, F. A. Quiocho, Nature 1985, 314, 257 – 260; b) H. Luecke, F. A. Quiocho, Nature 1990, 347, 402 – 406. [4] a) M. D. Lankshear, P. D. Beer, Acc. Chem. Res. 2007, 40, 657 – 668; b) G. T. Spence, P. D. Beer, Acc. Chem. Res. 2013, 46, 571 – 586; c) N. H. Evans, E. S. H. Allinson, M. D. Lankshear, K.-Y. Ng, A. R. Cowley, C. J. Serpell, S. M. Santos, P. J. Costa, V. Felix, P. D. Beer, RSC Adv. RSC Advances 2011, 1, 995 – 1003; d) A. Caballero, F. Zapata, P. D. Beer, Coord. Chem. Rev. 2013, 257, 2434 – 2455. [5] a) P. Metrangolo, F. Meyer, T. Pilati, G. Resnati, G. Terraneo, Angew. Chem. 2008, 120, 6206 – 6220; Angew. Chem. Int. Ed. 2008, 47, 6114 – 6127; b) A. C. Legon, Phys. Chem. Chem. Phys. 2010, 12, 7736 – 7747. [6] a) T. M. Beale, M. G. Chudzinski, M. G. Sarwar, M. S. Taylor, Chem. Soc. Rev. 2013, 42, 1667 – 1680; b) C. J. Serpell, N. L. Kilah, P. J. Costa, V. Felix, P. D. Beer, Angew. Chem. 2010, 122, 5450 – 5454; Angew. Chem. Int. Ed. 2010, 49, 5322 – 5326; c) N. L. Kilah, M. D. Wise, C. J. Serpell, A. L. Thompson, N. G. White, K. E. Christensen, P. D. Beer, J. Am. Chem. Soc. 2010, 132, 11893 – 11895; d) A. Caballero, F. Zapata, N. G. White, P. J. Costa, V. Felix, P. D. Beer, Angew. Chem. 2012, 124, 1912 – 1916; Angew. Chem. Int. Ed. 2012, 51, 1876 – 1880; e) L. C. Gilday, T. Lang, A. Caballero, P. J. Costa, V. Flix, P. D. Beer, Angew. Chem. 2013, 125, 4452 – 4456; Angew. Chem. Int. Ed. 2013, 52, 4356 – 4360; f) F. Zapata, A. Caballero, N. G. White, T. D. W. Claridge, P. J. Costa, V. Felix, P. D. Beer, J. Am. Chem. Soc. 2012, 134, 11533 – 11541; g) L. C. Gilday, N. G. White, P. D. Beer, Dalton Trans. 2013, 42, 15766 – 15773; h) M. G. Sarwar, B. Dragisic´, E. Dimitrijevic´, M. S. Taylor, Chem. Eur. J. 2013, 19, 2050 – 2058. [7] There are relatively few examples of catenane host systems for anion recognition. See reference [3c], reference [5d] and a) A. Andrievsky, F. Ahuis, J. L. Sessler, F. Vogtle, D. Gudat, M. Moini, J. Am. Chem. Soc. 1998, 120, 9712 – 9713; b) M. R. Sambrook, P. D. Beer, J. A. Wisner, R. L. Paul, A. R. Cowley, J. Am. Chem. Soc. 2004, 126, 15364 – 15365; c) M. K. Chae, J. M. Suk, K. S. Jeong, Tetrahedron Lett. 2010, 51, 4240 – 4242; d) N. H. Evans, C. J. Serpell, P. D. Beer, Angew. Chem. 2011, 123, 2555 – 2558; Angew. Chem. Int. Ed. 2011, 50, 2507 – 2510; e) L. M. Hancock, L. C. Gilday, N. L. Kilah, C. J. Serpell, P. D. Beer, Chem. Commun. 2011, 47, 1725 – 1727; f) J.-J. Lee, J. M. Baumes, R. D. Connell, A. G. Oliver, B. D. Smith, Chem. Commun. 2011, 47, 7188 – 7190; g) N. G. White, P. D. Beer, Chem. Commun. 2012, 48, 8499 – 8501. [8] J. A. Wisner, P. D. Beer, M. G. B. Drew, Angew. Chem. 2001, 113, 3718 – 3721; Angew. Chem. Int. Ed. 2001, 40, 3606 – 3609. [9] M. J. Hynes, J. Chem. Soc. Dalton Trans. 1993, 311 – 312. [10] M. R. Sambrook, P. D. Beer, J. A. Wisner, R. L. Paul, A. R. Cowley, F. Szemes, M. G. B. Drew, J. Am. Chem. Soc. 2005, 127, 2292 – 2302. [11] Both cis and trans isomers of the Grubbs’ RCM macrocycle, in which anion binding affinities are assumed to be very similar, are present. [12] Synthesis of the related HB catenane was undertaken by using similar reaction conditions. Macrocycle 1 and 1.1 equivalents of 4·Cl in anhydrous dichloromethane were stirred at room temperature for four days in the presence of Grubbs’ second-generation catalyst (10 % by weight). A small amount of HB catenane was isolated in very low yield (ca. 2 %).
Received: March 24, 2014 Published online on May 30, 2014
8385
2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Full Paper
& Supramolecular Chemistry
Halogen- and Hydrogen-Bonding Catenanes for Halide-Anion Recognition Lydia C. Gilday and Paul D. Beer*[a]
Abstract: Halogen-bonding (XB) interactions were exploited in the solution-phase assembly of anion-templated pseudorotaxanes between an isophthalamide-containing macrocycle and bromo- or iodo-functionalised pyridinium threading components. 1H NMR spectroscopic titration investigations demonstrated that such XB interpenetrated assemblies are more stable than analogous hydrogen bonding (HB) pseudorotaxanes. The stability of the anion-templated halogen-
Introduction The essential biological, medicinal, chemical, and industrial functions of negatively charged species, as well as the continuing harmful environmental impact of certain anions, is well documented.[1] Consequently, the recognition and sensing of target anions is a central goal of anion supramolecular chemistry, and through the imaginative manipulation of a variety of complementary non-covalent interactions, a plethora of efficient synthetic anion receptors have been developed.[2] However, the challenge to mimic the impressive levels of anion recognition demonstrated by biological systems still remains. Inspired by nature’s sulfate- and phosphate-binding proteins, which effect strong and selective anion complexation in elaborate organised binding domains shielded from competing water molecules,[3] we have employed anions to template the construction of mechanically bonded molecules.[4] These rotaxane and catenane interlocked host structures each contain a unique, three-dimensional cavity between the individual components leading to notable enhancement of anion binding affinities by complementary electrostatics and hydrogen-bonding (HB) interactions in competitive solvent media compared with the macrocycle and axle parent derivatives. Halogen bonding (XB) is the attractive, highly directional, non-covalent interaction between an electron-deficient halogen atom and a Lewis base.[5] In spite of their stringent directionality and strength comparable with ubiquitous HB, the exploitation of XB in solution-phase applications is only just be[a] Dr. L. C. Gilday, Prof. P. D. Beer Chemistry Research Laboratory, Department of Chemistry University of Oxford, Mansfield Road, Oxford, OX1 3TA (UK) Fax: (+ 44) 1865-272690 E-mail: [email protected]
bonded pseudorotaxane architectures was exploited in the preparation of new halogen-bonding interlocked catenane species through a Grubbs’ ring-closing metathesis (RCM) clipping methodology. The catenanes’ anion recognition properties in the competitive CDCl3/CD3OD 1:1 solvent mixture revealed selectivity for the heavier halides iodide and bromide over chloride and acetate.
ginning to be realised.[6] In an important step in advancing solution-phase anion recognition, we have prepared the first XB rotaxane, which through the combined use of both halogenand hydrogen-bonding interactions, binds iodide more strongly than either chloride or bromide anions,[6c] and a XB catenane, which is selective for chloride and bromide over a variety of other anions.[6d] Herein, we describe the anion-templated construction of new catenane host systems designed to exploit halogen- and hydrogen-bonding for anion recognition.[7] Figure 1 schematically shows the synthetic strategy employed to prepare the target halogen-/hydrogen-bonding catenanes. Anion-templated threading of an acyclic positively charged XB-donor component through a suitable isophthala-
Figure 1. Cartoon representation of orthogonal pseudorotaxane species, in which X is a halogen atom.
mide-containing macrocycle gives an orthogonal interpenetrative assembly stabilised by halogen- and hydrogen-bonding– anion interactions, which upon cyclisation will afford the new XB–HB catenanes.
Supporting information for this article is available on the WWW under http://dx.doi.org/10.1002/chem.201402752. Chem. Eur. J. 2014, 20, 8379 – 8385
8379
2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Full Paper Results and Discussion Anion-templated pseudorotaxane assembly Macrocycle 1[8] and functionalised pyridinium derivatives 2– 4·BF4 were prepared by using multistep synthetic procedures reported previously (Figure 2).[6e] Pseudorotaxane assembly between macrocycle 1 and potential pyridinium threading compounds 2–4·BF4 was studied initially using 1H NMR titration experiments in CDCl3, in the presence and absence of a halide-anion template (Scheme 1).
Figure 2. Structures of macrocycle 1 and threading components 2–4·BF4.
Aliquots of pyridinium threading components as their tetrafluoroborate salts were added to an equimolar solution of macrocycle 1 and tetrabutylammonium (TBA) halide salt in [D]chloroform. Addition of bromo- and iodopyridinium derivatives 2·BF4 and 3·BF4 to macrocycle 1 in the presence of bromide or iodide caused significant shifts in a number of signals in the respective 1H NMR spectrum (Figure 3). Downfield perturbations of the macrocycle cavity proton signal c and amide protons signal d are indicative of halide anion binding within the macrocycle’s isophthalamide binding cleft. Importantly, a diagnostic upfield shift and splitting of the signal corresponding to the macrocycle hydroquinone protons g and h was observed, which is consistent with aromatic donor–acceptor intermolecular interactions between the macrocycle’s electron-rich hydroquinone units and the electron-deficient halo-pyridinium group, indicative of pseudorotaxane formation. Also, a downfield shift in the signal corresponding to the N-methyl pyridi-
Scheme 1. Pseudorotaxane formation between macrocycle 1 and threading components 2–4·BF4 in the presence of TBA·A.
Figure 3. Partial 1H NMR spectra of: a) macrocycle 1; b) macrocycle 1 + Br ; c) pseudorotaxane; d) bromopyridinium thread 3·Br (500 MHz, CDCl3, 298 K). Chem. Eur. J. 2014, 20, 8379 – 8385
www.chemeurj.org
8380
2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Full Paper nium protons a was seen, which results from hydrogen bonding to the macrocycle’s polyether oxygen atoms. Upfield shifts in the signals corresponding to the halo-pyridinium thread’s protons b and g are a result of both aromatic donor–acceptor interactions and the bromide anion-template binding to the macrocycle’s isophthalamide domain. The observed chemical shift perturbations of the hydroquinone protons with ten equivalents of threading component revealed that, in general, bromide anion templation causes the largest magnitude of perturbation with relatively smaller changes with iodide or in
Macrocycle 1 + Br Macrocycle 1 + I
Macrocycle 1
larger magnitude for interpenetration with the XB halopyridinium derivatives than with the equivalent protic pyridinium species (Kapp = 1500 vs. 340 m1 with iodide), which demonstrates that halogen-bonding, halide-anion-templated pseudorotaxanes are more stable than their hydrogen-bonding counterparts. Of the interpenetrative assemblies stabilised by XB interactions, the iodopyridinium derivative is more favourable than the analogous bromopyridinium threading component (Kapp = 1500 and 780 m1 , respectively, with iodide). This is attributed to the greater halogen-bond-donor ability of the more polarisable iodine atom compared with bromine. It is noteworthy that pseudorotaxane assembly is stronger when a bromide anion template is employed instead of iodide: Kapp (2·A) = 2300 versus 780 m1. This suggests that the relatively smaller bromide anion is more suitably sized to bind to the macrocycle’s isophthalamide anion recognition cleft than the larger iodide anion.
Dd [ppm][a] 0.39
Kapp [m1][c] 2300
Dd [ppm][a] 0.19
Kapp [m1][c] 780
Dd [ppm][a] 0.16
Kapp [m1][c] 270
Synthesis of catenanes
0.22
3300[d]
0.18
1500
0.17
250[e]
0.31[b]
1000
0.15
340
0.13
220[e]
Table 1. Observed chemical shift perturbations (Dd) of hydroquinone protons and apparent association constants, Kapp (m1), for macrocycle 1 and threading components 2·BF4, 3·BF4 or 4·BF4 in the presence and absence of halide anions.
2·BF4 (X = Br) 3·BF4 (X = I) 4·BF4 (X = H)
[a] After ten equivalents. [b] After seven equivalents. [c] Apparent association constants were determined by monitoring the perturbations of the macrocycle hydroquinone protons upon addition of threading species to an equimolar mixture of macrocycle and TBA·A. Error < 10 %. [d] Error < 25 %. [e] Error < 15 %; CDCl3, 298 K.
the absence of a halide templating anion (Table 1). Monitoring the perturbations of the macrocycle’s hydroquinone protons as a function of concentration of threading component gave the titration curves shown in Figure S11 in the Supporting Information, and analysis of the titration data by using the winEQNMR2[9] computer program gave apparent association constants, Kapp,[10] for pseudorotaxane formation reported in Table 1. Importantly, in all cases, the most stable pseudorotaxane assemblies were found in the presence of bromide and iodide halide-anion templates. In the absence of a halide-anion template, the interpenetrative association with the halo- and proticfunctionalised pyridinium tetrafluoroborate salts are all much weaker and are of similar modest stability. The apparent association constants are of Chem. Eur. J. 2014, 20, 8379 – 8385
Encouraged by the favourable anion-templated association between macrocycle 1 and macrocycle precursor compounds 2·BF4 and 3·BF4, in the presence of either bromide or iodide respectively, the synthesis of new anion-templated, halogenbonding [2]catenanes using a Grubbs’-catalysed ring-closing metathesis (RCM) reaction was undertaken (Scheme 2). The required bromide and iodide salts of the vinyl-appended macrocycle precursor components were prepared by anion exchanging the non-coordinating tetrafluoroborate anion for the
Scheme 2. Synthesis of new anion-templated, halogen-bonding catenanes 5·Br and 6·I.
www.chemeurj.org
8381
2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Full Paper
Figure 4. Partial 1H NMR spectra of: a) bromopyridinium macrocycle precursor 2·Br; b) catenane 5·Br; and c) macrocycle 1 (500 MHz, CDCl3, 298 K).
appropriate halide-anion template. Macrocycle 1 and 1.1 equivalents of either 2·Br or 3·I in anhydrous dichloromethane were stirred at room temperature for four days in the presence of Grubbs’ second-generation catalyst (10 % by weight). Purification using repeated preparative thin-layer chromatography, the target catenanes 5·Br and 6·I were each isolated in 8 % yield.[11] Evidence of successful interlocked structure formation was provided by high-resolution electrospray mass spectrometry and 1H NMR spectroscopy. The cationic molecular-ion peaks of both catenanes were observed in the mass spectrum, which suggests the presence of a single interlocked species, and the high-resolution mass spectra revealed that the isotopic distribution of the catenanes 5·PF6 and 6·PF6 are in good agreement with the theoretical spectra (see Figures S5 and S10 in the Supporting Information). The 1H NMR spectrum of bromopyridinium-containing catenane 5·Br together with those for the component macrocycle and bis-vinyl-appended precursor 2·Br for comparison are shown in Figure 4. Important spectral changes, which are diagnostic of catenane formation, include the disappearance of the vinylic multiplet m and the convergence of multiplet l into a pseudo-singlet in the catenane spectrum, which are attributed to cyclisation of the acyclic precursor. Crucially, the upfield shift and splitting of the macrocycle hydroquinone protons g and h is indicative of aromatic donor–acceptor interactions between the electron-rich macrocycle hydroquinone units and the electron-poor pyridinium group, confirming the two components are interlocked. Two dimensional 1H–1H rotating-frame Overhauser effect spectroscopy (ROESY) NMR spectroscopy provided further evidence of the interlocked nature of the [2]catenane 5·Br (Figure 5). The 1H–1H ROESY spectrum displays several through space cross-coupling interactions between the two components of the catenane. Important interactions include 1 (b!g) and 2 (b!h), which confirm the electron-rich hydroquinone units and the electron-poor pyridinium moiety are close in Chem. Eur. J. 2014, 20, 8379 – 8385
www.chemeurj.org
space, as would be expected when the macrocycle components are interlocked with respect to one another. Interaction 3 (a!h) also demonstrates the hydroquinone and pyridinium aromatic units are in close proximity. To highlight the important role of the halide-anion template in aiding formation of the orthogonal interpenetrative assembly of the precursor components, the synthesis of the catenanes was also attempted with 2·BF4 and 3·BF4. Although peaks at m/z 1238.4 and 1286.4 corresponding to [2]catenanes 5·BF4 and 6·BF4 were detected by electrospray mass spectrometry, in spite of numerous attempts, the catenanes could not be isolated, which suggests they were formed in negligible yield in the absence of a halide template.[12]
Anion-recognition properties of catenanes To study the anion-recognition properties of the catenanes, the respective templating halide guests were exchanged for the non-coordinating hexafluorophosphate anion by repeatedly washing a dichloromethane solution of the catenane with aqueous ammonium hexafluorophosphate. The halide and acetate anion-binding properties of the catenanes 5·PF6 and 6·PF6 were investigated by using 1H NMR spectroscopic titration experiments with TBA anion salts in the competitive solvent mixture [D]chloroform/[D4]MeOH 1:1. Upon addition of halide anions to both catenane species, significant downfield chemical shift perturbations of the ortho-isophthalamide proton c and pyridinium proton b were observed with increasing anion concentration. By contrast, addition of acetate induced considerably smaller shifts (Ddppm = 0.03 ppm for 6·PF6), which indicated that no acetate complexation occurred in this solvent mixture. WinEQNMR analysis of the titration curves obtained for chloride, bromide, iodide, and acetate binding (Figure 6) gave 1:1 stoichiometric association constants reported in Table 2.
8382
2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Full Paper
Figure 5. 1H–1H ROESY NMR spectrum of catenane 5·Br. Assignable inter-component through-space interactions are highlighted (500 MHz, CDCl3, 298 K).
Table 2. Association constants, Ka [m1], for addition of various anions to catenanes 5·PF6 and 6·PF6. Anion
Catenane 5·PF6 Ka [m1][a]
Catenane 6·PF6 Ka [m1][a]
Cl Br I OAc
130 200 220 NB
160 230 340 NB
[a] All anions were added as their TBA salt. Association constants were determined by monitoring the perturbations of isophthalamide cavity proton c. Error < 10 %. NB: no evidence of binding; Ddppm too small to allow an accurate association constant to be obtained. CDCl3/CD3OD 1:1, 298 K.
Figure 6. Normalised titration curves for addition of various anions to catenanes 5·PF6 (top) and 6·PF6 (bottom) obtained by monitoring the macrocycle cavity proton c. CDCl3/CD3OD 1:1, 298 K. Square data points represent experimental data; continuous lines represent theoretical binding isotherms. Chem. Eur. J. 2014, 20, 8379 – 8385
www.chemeurj.org
The association constant values reveal both catenanes 5·PF6 and 6·PF6 exhibited halide-anion recognition, but do not bind the basic acetate anion, which suggests that the oxoanion is too large and of the wrong shape to penetrate the catenanes’ binding pocket. Both interlocked hosts complex halides chloride, bromide and iodide with moderate affinity in this competitive solvent system. The iodopyridinium-containing catenane 6·PF6 is a superior halide anion host to the bromopyridinium-functionalised catenane 5·PF6 (Ka = 340 vs. 220 m1 for iodide association), which reflects the greater halogen-bond-donor ability of the iodine atom compared with bromine. Interestingly, for both catenanes the strongest host–guest association is with iodide, followed by bromide and chloride. This is the reverse of what was expected on anion-basicity grounds and may be attributed to a combination of factors including an interlocked bind-
8383
2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Full Paper ing domain more suited to the larger halide anion, a higher penalty for desolvation of chloride than bromide or iodide in this competitive protic solvent mixture and, possibly, a greater degree of charge-transfer-type halogen-bonding interaction, between the respective catenane’s bromo- and iodopyridinium halogen-bond-donor motif and the larger halide anion.
yield (25.8 mg, 34.3 mmol). 1H NMR (300 MHz, CDCl3): d = 8.24 (s, 2 H; pyridinium-H3&H5), 6.84 (s, 8 H; hydroquinone-H), 5.87–6.00 (m, 2 H; C=CH), 5.17–5.34 (m, 4 H; C=CH2), 5.16 (s, 4 H; CH2O), 4.24 (m, 3 H; CH3), 3.99–4.14 ppm (m, 20 H; CH2); 13C{1H} NMR (75.5 MHz, CDCl3): d = 156.3, 153.2, 152.5, 143.5, 134.5, 129.0, 117.3, 115.6, 115.3, 72.3, 70.7, 68.5, 68.0, 67.4, 40.5 ppm; MS (ESI): m/z: calcd for C34H43BrNO8 : 672.2168 [MBr] + ; found: 672.2163.
Conclusion
Compound 3·I
Through the combination of halogen- and hydrogen-bonding non-covalent interactions, bromide and iodide anions have been demonstrated to template the formation of interpenetrative assemblies between an isophthalamide-containing macrocycle and bromo- or iodo-functionalised pyridinium threading components. Pseudorotaxane formation was more favourable when XB was employed—specifically, the more stable bromide- and iodide-templated interpenetrated assemblies were found with the iodopyridinium species, which reflects the greater halogen-bond-donor ability of the more polarisable iodine atom. The stability of the anion-templated XB pseudorotaxanes was exploited in the construction of two new XB catenane architectures, the halide- and acetate-anion recognition properties of which were probed in competitive solvent mixture. The strength of anion binding was found to be greater with the iodo-functionalised halogen-bond-donor catenane with both catenanes displaying anion affinities iodide > bromide > chloride @ acetate. The integration of XB donor components into interlocked structural frameworks is continuing in our laboratories.
Compound 3·BF4 (51.2 mg, 63.4 mmol) was dissolved in CH2Cl2 (30 mL) and washed with 1 m NH4I (aq. 10 10 mL), H2O (2 10 mL), dried over anhydrous MgSO4, filtered and the solvent removed in vacuo to give 3·I as a pale orange oily solid in 99 % yield (53.1 mg, 62.7 mmol). 1H NMR (300 MHz, CDCl3): d = 8.44 (s, 2 H; pyridiniumH3&H5), 6.85 (s, 8 H; hydroquinone-H), 5.87–6.00 (m, 2 H; C=CH), 5.19–5.34 (m, 4 H; C=CH2), 5.02 (s, 4 H; CH2O), 4.02–4.13 (m, 21 H; CH2 & CH3), 3.75–3.8 ppm (m, 4 H; CH2); 13C{1H} NMR (75.5 MHz, CDCl3): d = 154.2, 153.3, 152.6, 135.0, 134.5, 117.4, 115.7, 115.5, 72.3, 70.9, 68.6, 68.0, 67.6, 61.5, 53.4, 40.8 ppm; MS (ESI): m/z calcd for C34H43INO8 : 720.2028 [MI] + ; found 720.2040.
Experimental Section General considerations Commercially available reagents and solvents were used without further purification unless otherwise stated. When anhydrous solvents were used, they were purged with N2, passed through a MBraun MPSP-800 column and then used immediately. Et3N was distilled from, and stored over, KOH. H2O was de-ionised and microfiltered by using a Milli-Q Millipore machine. TBA salts and Grubbs’ II catalyst were stored in a vacuum desiccator prior to use. Column chromatography was performed on silica gel (particle size 40–63 mm), preparative TLC was performed on 20 20 cm plates, with a silica layer of thickness 1 mm. 1H, 13C, 19F and 31P NMR spectra were recorded on Varian Mercury-vx 300 spectrometer or on Varian Unity Plus or Bruker Avance III 500 instruments. Mass spectra were obtained by using a Micromass LCT Premier XE spectrometer and accurate masses were obtained to four decimal places by using a Bruker MicroTOF spectrometer. Macrocycle 1 and functionalised pyridinium threading components 2–4·BF4 were prepared according to reported procedures.
Synthetic procedures Compound 2·Br Compound 2·BF4 (31.8 mg, 41.8 mmol) was dissolved in CH2Cl2 (30 mL) and washed with 1 m NH4Br (aq. 10 10 mL), H2O (2 10 mL), dried over anhydrous MgSO4, filtered, and the solvent was removed in vacuo to give 2·Br as a pale orange oily solid in 82 % Chem. Eur. J. 2014, 20, 8379 – 8385
www.chemeurj.org
Catenane 5·Br Macrocycle 1 (50 mg, 80 mmol) and 2·Br (68.4 mg, 90.8 mmol) were dissolved in anhydrous CH2Cl2 (5 mL) and stirred at RT under N2 for 30 min. Grubbs’ II (7 mg, 10 % by wt) was added, and the reaction mixture was stirred at RT under N2 for 3 d, after which time more Grubbs’ II (3.5 mg) was added, and the reaction mixture was stirred for a further 24 h. The solvent was removed in vacuo, and the crude residue was purified by preparative thin layer chromatography (95:5 CH2Cl2/MeOH and 3:2 MeCN/CH2Cl2) to give 5·Br as a white solid in 8 % yield (9.0 mg, 6.8 mmol). 1H NMR (300 MHz, CDCl3): d = 9.30 (s, 1 H; isophthaloyl-H2), 8.84 (br s, 2 H; amide-NH), 8.33 (d, 3J = 7.6 Hz, 2 H; isophthaloyl-H4&H6), 8.03 (s, 2 H; pyridinium-H3&H5), 7.59 (t, 3J = 7.6 Hz, 1 H; isophthaloyl-H5), 6.86 (d, 3J = 8.9 Hz, 4 H; hydroquinone-H), 6.80 (d, 3J = 8.9 Hz, 4 H; hydroquinone-H), 6.31 (d, 3J = 8.8 Hz, 4 H; hydroquinone-H), 6.08 (d, 3J = 8.8 Hz, 4 H; hydroquinone-H), 5.73 (s, 2 H; alkene-CH), 4.53 (s, 3 H; NCH3), 4.26 (br s, 4 H; CH2), 4.16 (br s, 4 H; CH2), 4.07–4.09 (m, 4 H; CH2), 3.97–3.98 (m, 12 H; CH2), 3.72–3.74 (m, 4 H; CH2), 3.59–3.66 (m, 12 H; CH2), 3.46–3.50 ppm (m, 8 H; CH2); 13C{1H} NMR (75.5 MHz, CDCl3): d = 166.7, 153.6, 153.4, 153.2, 152.2, 151.3, 138.5, 133.8, 131.8, 130.0, 128.9, 128.9, 128.7, 124.6, 115.9, 115.5, 114.9, 114.4, 71.1, 71.1(sic), 70.5, 70.0, 69.7, 68.9, 68.3, 67.8, 66.5, 39.8, 29.7 ppm; MS (ESI): m/z calcd for C64H77BrN3O17: 1238.4431 [MBr] + ; found: 1238.4457.
Catenane 5·PF6 Compound 5·Br (8.50 mg, 6.16 mmol) was dissolved in CH2Cl2 (20 mL), washed with 0.1 m NH4PF6 (aq. 10 10 mL), H2O (2 10 mL), dried over MgSO4, filtered, and the solvent was removed in vacuo to give 5·PF6 as an off-white solid in 81 % yield (7.66 mg, 5.53 mmol). 1H NMR (300 MHz, CDCl3): d = 8.25 (d, 3J = 7.8 Hz, 2 H; isophthaloyl-H4&H6), 8.12 (s, 1 H; isophthaloyl-H2), 7.61 (t, 3J = 7.8 Hz, 1 H; isophthaloyl-H5), 7.48 (s, 2 H; pyridinium-H3&H5), 7.13 (s, 2 H; amide-NH), 6.79–6.83 (s, 8 H; hydroquinone-H), 6.35 (d, 3J = 8.8 Hz, 4 H; hydroquinone-H), 6.14 (d, 4 H; 3J = 8.8 Hz, hydroquinone-H), 5.78 (s, 2 H; alkene-CH), 4.66 (s, 3 H; NCH3), 4.22 (s, 4 H; CH2), 4.00–4.06 (m, 16 H; CH2), 3.90–3.91 (m, 4 H; CH2), 3.74–3.75 (m, 4 H; CH2), 3.63 (s, 4 H; CH2), 3.59 (s, 4 H; CH2), 3.50 (s, 8 H; CH2), 3.45 ppm (s, 4 H; CH2); 13C{1H} NMR (125.8 MHz, CDCl3): d = 166.7, 160.6, 153.9, 153.5, 153.2, 152.4, 151.5, 142.7, 140.4, 134.5, 131.7,
8384
2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Full Paper 129.3, 129.0, 115.7, 115.4, 114.9, 71.1, 70.5, 70.1, 70.1(sic), 69.1, 68.2, 68.1, 67.1, 66.7, 66.5, 50.9, 39.6, 29.7 ppm; 19F NMR (282.4 MHz, CDCl3): d = 71.1 ppm (d, J = 712 Hz, PF6); 31P NMR (202.4 MHz, CDCl3): d = 144.0 ppm (sept, J = 712 Hz, PF6); MS (ESI): m/z calcd for C64H77BrN3O17: 1238.4431 [MPF6] + ; found: 1238.4423.
Catenane 6·I Macrocycle 1 (71 mg, 120 mmol) and compound 3·I (92.5 mg, 109 mmol) were dissolved in anhydrous CH2Cl2 (5 mL) and stirred at RT under N2 for 30 min. Grubbs’ II (9.3 mg, 10 % by wt.) was added and the reaction mixture stirred at RT under N2 for 3 d, after which time more Grubbs’ II (4.5 mg) was added, and the reaction mixture was stirred for a further 24 h. The solvent was removed in vacuo, and the crude residue was purified by preparative thin-layer chromatography (95:5 CH2Cl2/MeOH and 3:2 MeCN/CH2Cl2) to give 6·I as a white solid in 8 % yield (13.5 mg, 9.44 mmol). 1H NMR (500 MHz, CDCl3): d = 9.02 (s, 1 H; isophthaloyl-H2), 8.44 (br s, 2 H; amide-NH), 8.29 (d, 3J = 7.6 Hz, 2 H; isophthaloyl-H4&H6), 7.95 (s, 2 H; pyridinium-H3&H5), 7.58 (t, 3J = 7.6 Hz, 1 H; isophthaloyl-H5), 6.83 (s, 8 H; hydroquinone-H), 6.43 (d, 3J = 8.8 Hz, 4 H; hydroquinone-H), 6.12 (d, 3J = 8.8 Hz, 4 H; hydroquinone-H), 4.67 (s, 4 H; CH2), 4.30– 4.32 (m, 4 H; CH2), 4.00–4.07 (m, 16 H; CH2), 3.79 (s, 3 H; NCH3) 3.73– 3.75 (m, 4 H; CH2), 3.60 (s, 4 H; CH2), 3.55 (s, 4 H; CH2), 3.46–3.47 (m, 4 H; CH2), 3.40 ppm (s, 4 H; CH2); 13C{1H} NMR (125.8 MHz, CDCl3): d = 166.8, 153.6, 153.3, 152.4, 151.8, 151.4, 135.2, 134.1, 131.9, 128.9, 124.4, 120.3, 115.8, 115.5, 115.0, 114.8, 71.19, 71.12, 70.5, 70.1, 69.8, 69.0, 68.3, 68.0, 67.0, 66.8, 66.7, 39.4, 29.7 ppm; MS (ESI): m/z calcd for C64H77IN3O17: 1286.4292 [MI] + ; found: 1286.4300.
Catenane 6·PF6 Compound 6·I (13.5 mg, 9.44 mmol) was dissolved in CH2Cl2 (20 mL), washed with 0.1 m NH4PF6 (aq. 10 10 mL), H2O (2 10 mL), dried over MgSO4, filtered, and the solvent was removed in vacuo to give 6·PF6 as an off-white solid in 67 % yield (9.13 mg, 6.37 mmol). 1H NMR (500 MHz, CDCl3): d = 8.23 (d, 3J = 7.6 Hz, 2 H; isophthaloyl-H4&H6), 8.15 (s, 1 H; isophthaloyl-H2), 7.63 (2 H, s, pyridinium-H3&H5), 7.60 (1 H, t, 3J = 7.6 Hz, isophthaloyl-H5), 7.13 (br s, 2 H; amide-NH), 6.79–6.88 (m, 8 H; hydroquinone-H), 6.41 (d, 3J = 8.7 Hz, 4 H; hydroquinone-H), 6.15 (d, 3J = 8.7 Hz, 4 H; hydroquinone-H), 6.79 (s, 2 H; alkene-CH), 4.68 (s, 4 H; CH2), 4.23 (s, 4 H; CH2), 4.01–4.07 (m, 16 H; CH2), 3.93–3.94 (m, 4 H; CH2), 3.86 (s, 3 H; NCH3), 3.74–3.76 (4 H, m, CH2), 3.62 (4 H, s, CH2), 3.57 (4 H, s, CH2), 3.48 (s, 4 H; CH2), 3.42 ppm (s, 4 H; CH2); 13C{1H} NMR (125.8 MHz, CDCl3): d = 166.6, 153.4, 153.2, 152.5, 151.5, 150.6, 134.9, 134.5, 131.7, 129.3, 129.0, 115.6, 115.5, 115.0, 114.9, 71.1, 70.5, 70.0, 69.9, 69.1, 68.2, 68.1, 66.9, 66.7, 66.4, 34.1, 29.0 ppm; 31P NMR (MHz, CDCl3): d = 144.0 ppm (sept, J = 716 Hz, PF6); MS (ESI): m/z calcd for C64H77IN3O17: 1286.4292 [MPF6] + ; found: 1286.4287.
Acknowledgements We thank the EPSRC and the Department of Chemistry, University of Oxford for funding. Keywords: anions · catenanes · halogen bonding · molecular recognition · supramolecular chemistry
Chem. Eur. J. 2014, 20, 8379 – 8385
www.chemeurj.org
[1] a) J. L. Sessler, P. A. Gale, W.-S. Cho, Anion Receptor Chemistry, The Royal Society of Chemistry, Cambridge, UK, 2006; b) A. Bianci, K. BowmanJames, E. Garcia-EspaÇa, Anion Coordination Chemistry, Wiley-VCH, Weinheim, 2011. [2] a) P. D. Beer, P. A. Gale, Angew. Chem. 2001, 113, 502 – 532; Angew. Chem. Int. Ed. 2001, 40, 486 – 516; b) J. M. Llinares, D. Powell, K. Bowman-James, Coord. Chem. Rev. 2003, 240, 57 – 75; c) E. J. O’Neil, B. D. Smith, Coord. Chem. Rev. 2006, 250, 3068 – 3080; d) R. Martnez-MÇez, F. Sancenn, Coord. Chem. Rev. 2006, 250, 3081 – 3093; e) T. Gunnlaugsson, M. Glynn, G. M. Tocci, P. E. Kruger, F. M. Pfeffer, Coord. Chem. Rev. 2006, 250, 3094 – 3117; f) C. R. Rice, Coord. Chem. Rev. 2006, 250, 3190 – 3199. [3] a) J. W. Pflugrath, F. A. Quiocho, Nature 1985, 314, 257 – 260; b) H. Luecke, F. A. Quiocho, Nature 1990, 347, 402 – 406. [4] a) M. D. Lankshear, P. D. Beer, Acc. Chem. Res. 2007, 40, 657 – 668; b) G. T. Spence, P. D. Beer, Acc. Chem. Res. 2013, 46, 571 – 586; c) N. H. Evans, E. S. H. Allinson, M. D. Lankshear, K.-Y. Ng, A. R. Cowley, C. J. Serpell, S. M. Santos, P. J. Costa, V. Felix, P. D. Beer, RSC Adv. RSC Advances 2011, 1, 995 – 1003; d) A. Caballero, F. Zapata, P. D. Beer, Coord. Chem. Rev. 2013, 257, 2434 – 2455. [5] a) P. Metrangolo, F. Meyer, T. Pilati, G. Resnati, G. Terraneo, Angew. Chem. 2008, 120, 6206 – 6220; Angew. Chem. Int. Ed. 2008, 47, 6114 – 6127; b) A. C. Legon, Phys. Chem. Chem. Phys. 2010, 12, 7736 – 7747. [6] a) T. M. Beale, M. G. Chudzinski, M. G. Sarwar, M. S. Taylor, Chem. Soc. Rev. 2013, 42, 1667 – 1680; b) C. J. Serpell, N. L. Kilah, P. J. Costa, V. Felix, P. D. Beer, Angew. Chem. 2010, 122, 5450 – 5454; Angew. Chem. Int. Ed. 2010, 49, 5322 – 5326; c) N. L. Kilah, M. D. Wise, C. J. Serpell, A. L. Thompson, N. G. White, K. E. Christensen, P. D. Beer, J. Am. Chem. Soc. 2010, 132, 11893 – 11895; d) A. Caballero, F. Zapata, N. G. White, P. J. Costa, V. Felix, P. D. Beer, Angew. Chem. 2012, 124, 1912 – 1916; Angew. Chem. Int. Ed. 2012, 51, 1876 – 1880; e) L. C. Gilday, T. Lang, A. Caballero, P. J. Costa, V. Flix, P. D. Beer, Angew. Chem. 2013, 125, 4452 – 4456; Angew. Chem. Int. Ed. 2013, 52, 4356 – 4360; f) F. Zapata, A. Caballero, N. G. White, T. D. W. Claridge, P. J. Costa, V. Felix, P. D. Beer, J. Am. Chem. Soc. 2012, 134, 11533 – 11541; g) L. C. Gilday, N. G. White, P. D. Beer, Dalton Trans. 2013, 42, 15766 – 15773; h) M. G. Sarwar, B. Dragisic´, E. Dimitrijevic´, M. S. Taylor, Chem. Eur. J. 2013, 19, 2050 – 2058. [7] There are relatively few examples of catenane host systems for anion recognition. See reference [3c], reference [5d] and a) A. Andrievsky, F. Ahuis, J. L. Sessler, F. Vogtle, D. Gudat, M. Moini, J. Am. Chem. Soc. 1998, 120, 9712 – 9713; b) M. R. Sambrook, P. D. Beer, J. A. Wisner, R. L. Paul, A. R. Cowley, J. Am. Chem. Soc. 2004, 126, 15364 – 15365; c) M. K. Chae, J. M. Suk, K. S. Jeong, Tetrahedron Lett. 2010, 51, 4240 – 4242; d) N. H. Evans, C. J. Serpell, P. D. Beer, Angew. Chem. 2011, 123, 2555 – 2558; Angew. Chem. Int. Ed. 2011, 50, 2507 – 2510; e) L. M. Hancock, L. C. Gilday, N. L. Kilah, C. J. Serpell, P. D. Beer, Chem. Commun. 2011, 47, 1725 – 1727; f) J.-J. Lee, J. M. Baumes, R. D. Connell, A. G. Oliver, B. D. Smith, Chem. Commun. 2011, 47, 7188 – 7190; g) N. G. White, P. D. Beer, Chem. Commun. 2012, 48, 8499 – 8501. [8] J. A. Wisner, P. D. Beer, M. G. B. Drew, Angew. Chem. 2001, 113, 3718 – 3721; Angew. Chem. Int. Ed. 2001, 40, 3606 – 3609. [9] M. J. Hynes, J. Chem. Soc. Dalton Trans. 1993, 311 – 312. [10] M. R. Sambrook, P. D. Beer, J. A. Wisner, R. L. Paul, A. R. Cowley, F. Szemes, M. G. B. Drew, J. Am. Chem. Soc. 2005, 127, 2292 – 2302. [11] Both cis and trans isomers of the Grubbs’ RCM macrocycle, in which anion binding affinities are assumed to be very similar, are present. [12] Synthesis of the related HB catenane was undertaken by using similar reaction conditions. Macrocycle 1 and 1.1 equivalents of 4·Cl in anhydrous dichloromethane were stirred at room temperature for four days in the presence of Grubbs’ second-generation catalyst (10 % by weight). A small amount of HB catenane was isolated in very low yield (ca. 2 %).
Received: March 24, 2014 Published online on May 30, 2014
8385
2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
