Calix[4, 5]tetrolarenes: A New Family of Macrocycles - ACS Publications

Letter pubs.acs.org/OrgLett

Calix[4, 5]tetrolarenes: A New Family of Macrocycles Yossi Zafrani*,†,‡ and Yoram Cohen*,† †

School of Chemistry, The Sackler Faculty of Exact Sciences, Tel Aviv University, Ramat Aviv 69978, Tel Aviv, Israel The Department of Organic Chemistry, Israel Institute for Biological Research, Ness-Ziona 74000, Israel

‡

S Supporting Information *

ABSTRACT: The facile and efficient one-step synthesis and the full characterization of novel π-electron rich macrocycles, calix[n]tetrolarenes (n = 4, 5), are described. The tetramer and the much rarer sized pentamer were easily prepared by reaction of the commercially available, partially methylated 1,2,3,5-benzenetetrol with paraformaldehyde under TFA catalysis, with a total isolated yield of 73%. The reaction is solvent sensitive, and the number of methylated oxygens also affects the tetramer/pentamer distribution. The compounds formed may be considered as “chimeric” macrocycles of calixarene and pyrogallol[n]arene that may serve as new building blocks in host−guest and supramolecular chemistry.

A

One of the most puzzling issues that remains despite decades of study of the above-mentioned calix structures is the thermodynamic selectivity toward macrocycles with an even number of arene units (n = 4/6/8) rather than those having the odd numbers (n = 5/7).1 To date the most relevant synthesis of calix[5]arenes, for example, require harsh conditions in a one-step procedure13 or stepwise process;14 however, both approaches suffer from relatively low yields. Indeed, even today the formation of the classical calix[5]arenes remains significantly more challenging than the even numbered calix[n]arenes which have been commercially available for many years.15 Almost all variants of hydroxy/alkoxy-benzene derivatives have been used as starting materials for cyclization reactions; however, 1,2,3,5-tetrahydroxy/alkoxybenzenes have not. With two nonsubstituted meta positions (4 and 6) we reasoned that these compounds may lead to the corresponding electron-rich macrocycles with vase shapes (Figure 1). Moreover, these macrocycles should have a calixarene-like lower rim and pyrogallolarene-like upper rim and, therefore, may be considered as chimeric structures. Driven by motivation to discover novel macrocycles, especially those with more than four arene units, we wish to disclose our results on the chemistry of three per- and partially methylated 1,2,3,5-benzenetetrols when reacted with paraformaldehyde. One of these electron-rich arenes provides an extremely efficient means to obtain not only a novel tetrameric calixtetrolarene but also the much less accessible pentameric

rene-based macrocyclic receptors such as calix[n]arenes, pillar[n]arenes, and various related structures play pivotal roles in host−guest and supramolecular chemistry.1−4 Depending on their shape, size, and electronic properties, these fascinating cyclophanes are used in many scientific fields.5 Since the pioneering study by Gutshe et al. describing efficient synthesis and characterization of the predecessor calix[n]arenes (1) (Figure 1),6 efforts have been directed toward the preparation of new calixarene derivatives.4,5 It was only following the discovery of pillar[n]arenes by Ogoshi et al., in 2008,7 that we witnessed a renewed interest in the design of novel oxygenbased macrocyclic receptors. In recent years a series of new cyclic oligomers with different shapes, sizes, and host−guest properties have been reported.8−12 These include, for example, biphen[n]arenes,8 assar[n]arenes,9 hybrid[n]arenes,10 oxatube[n]arenes,11 and more.12 It appears that three elements dictate the shape and the physical and chemical properties of these arene-based macrocycles: the positions of the methylenic bridges, the number of oxygen atoms on the aromatic ring, and the locations of these oxygens. Thus, hydroxybenzene (phenol), 1,3-dihydroxybenzene (resorcinol) and 1,2,3-trihydroxybenzene (pyrogallol) derivatives lead to meta-bridge macrocycles the calix[n]arenes, resorcin[n]arenes, and pyrogallol[n]arenes, respectively all with a vase shape (Figure 1). On the other hand, 1,2-dihydroxybenzene derivatives lead to ortho-bridged bowl-shaped cyclotriveratrilene,3 while their 1,4-dihydroxy isomers lead to the para-bridged macrocycles, i.e. the pillar[n]arenes.2,7 © 2017 American Chemical Society

Received: May 21, 2017 Published: June 29, 2017 3719

DOI: 10.1021/acs.orglett.7b01511 Org. Lett. 2017, 19, 3719−3722

Letter

Organic Letters

Figure 1. Molecular structures of calix[n]arenes (1), pyrogallol[n]arenes (2), and calix[n]tetrolarenes.

macrocycle. Even in the very recent work by Neri’s group on the synthesis of large resorcin[n]arenes, the pentamer was obtained only in low yield relatively to its hexameric counterpart.16 To the best of our knowledge pyrogallol[5]arenes have never been prepared in practical yields. We tested commercial benzenetetrol derivatives and found that three compounds, 3,4,5-trimethoxy phenol (3a), 1,2,3,5tetramethoxybenzene (3b) and 1,4-dihydroxy-3,5-dimethoxybenzene (3c), may serve as relatively low cost reagents for our purpose. Noting the expected potential of the partially methylated macrocycles, we began our study with compound 3a. We found that reaction of 3a with paraformaldehyde using trifluoroacetic acid (TFA) as a catalyst and dichloroethane (DCE) as a solvent (rt, 72 h)17 yielded two cyclic products (with some acyclic oligomers) that could be isolated quite easily. Combined NMR and HRMS analyses and an X-ray diffraction study revealed that these two products were the dodecamethoxycalix[4]tetrolarene (4a) and the unexpected pentadecamethoxycalix[5]tetrolarene (5a) (see the scheme in Table 1, Figures 2, S1, and S2 in the Supporting Information).

Figure 2. 1H NMR spectra (500 MHz, 25 °C, CDCl3) of the crude mixture of 4a and 5a, isolated 4a, and isolated 5a.

DCE as solvent, shortening the reaction time to 24 h resulted in a decrease in the yield of the cyclic products (runs 1, 2). An interesting solvent effect was observed when DCE was replaced by other halogenated solvents such as dichloromethane (DCM) or chloroform. In DCM (run 4) nearly equal amounts of 4a and 5a were observed (33 and 31%, respectively), whereas in chloroform (run 3) a very complex mixture of products was obtained with only 4% and 14% of the cyclic tetramer and pentamer. In dichlorobenzene (DCB), which very recently was found to be the solvent of choice for the synthesis of large calix[n]resorcinarenes,16 the starting material failed to solubilize (run 5). In chlorocyclohexane (Cl-CyC6), a solvent that recently proved to be a template for the synthesis of pillar[6]arene,18 some preference for 5a was observed; however, this reaction was much slower and lead to a large amounts of linear oligomers (run 6). The effect of the reaction temperature on the products distribution suggest that the cyclopentamer 5a is the kinetic product while the cyclotetramer 4a is the thermodynamic product (runs 7 and 8). The reaction under kinetic control (low temperature, 4 °C) resulted in a large amount of acyclic oligomers and, therefore, cannot be used for efficient synthesis of 5a, due to difficulties in the workup and products separation. The reaction at higher temperature (70 °C, 2 h) afforded 4a and 5a as almost sole products, which could be easily separated by chromatography (run 7). As shown in Figure 2, the absence of any aromatic hydrogen and the appearance of only two sets of signals in the 1 H NMR spectrum indicate that, under these experimental conditions, 4a and 5a were the only NMR observable products. Note, however, that in chloroform or Cl-CyC6 at 70 °C some acyclic oligomers are observed in the crude mixtures (runs 9, 10). Therefore, DCE as solvent and 70 °C were optimal for the synthesis of both 4a and 5a which were obtained in 53% and 20% isolated yields, respectively. The 1H and 13C NMR spectra of 4a and 5a revealed that the two cyclic oligomers are flexible at room temperature (Figures 2, S1, and S2). In the 1H NMR

Table 1. Reaction of Fully and Partially Methylated Benzenetetrols 3a−c with Paraformaldehyde

P distributiona run

3

solvent

temp [°C]

time [h]

4 (%)

5 (%)

1 2 3 4 5 6 7 8 9 10 11 12

3a

DCE DCE CHCl3 DCM DCB Cl-CyC6 DCE DCE CHCl3 Cl-CyC6 DCE DCE

rt rt rt rt rt rt 70 4 70 70 rt rt

72 24 72 72 72 72 2 96 2 2 72 72

62 31 4 33 ndc 19 72(53)b 16 18 58 ndc ndc

20 27 14 31 ndc 25 22 (20)b 43 32 16 ∼22d ndc

3b 3c

a

Determined by integrating the 1H NMR signals of both the OH and the OMe groups. bIsolated yield, cNot determined, dEstimated by LC-MS analysis.

Selected experiments for the optimization of this reaction together with the outcome of the reactions with the related compounds 3b and 3c are tabulated in Table 1. When using 3720

DOI: 10.1021/acs.orglett.7b01511 Org. Lett. 2017, 19, 3719−3722

Letter

Organic Letters

8.90 ppm only slightly changed during the cooling processes. This behavior suggests a cone-type conformation for compound 4a. Interestingly, the energy barrier for the ring inversion of 4a was found to be 11.8 ± 0.2 kcal/mol, a much lower value compared to that of 4-tert-butyl calix[4]arene1b (15.7 kcal/mol). Inspection of the 1H NMR spectra of the CD2Cl2 solution of the pentameric macrocycle 5a at various temperatures revealed that this compound is also flexible. When this solution was cooled to −65 °C the single line at 3.82 ppm (attributed to the accidentally isochronous CH2, OCH3, and 2OCH3) transformed into seven broad signals (Figure 4). The slightly broad

spectrum of 5a all the methoxy groups and the bridged methylenes have exactly the same chemical shift at 3.82 ppm (Figure 2), as confirmed by the HSQC NMR analysis (Figure S4). Interestingly, the reaction with 1,2,3,5-tetramethoxybenzene (3b) at room temperature afford mainly the larger macrocycles (n = 5−9) but not the tetramer 4b (run 11, Figure S3). LC-MS analysis of the isolated macrocycles mixture (DCM− methanol−precipitated fraction, 57%) revealed that the pentamer 5b is the major product (ca. 40%). Heating a crude mixture of this reaction to 70 °C for 24 h did not lead to the tetrameric macrocycle 4b. More interestingly, these results are in line with those obtained by Stoddart and co-workers in the synthesis of the pillar-related macrocycles assararenes obtained from 1,2,4, 5-tetramethoxybenzene.9 These results suggest that with benzenetetrol derivatives the reaction is sensitive to the nature of the substituent at the 5-position. With hydroxyl at the 5-position (3a) the cyclotetramer 4a was the major product, whereas with methoxy at this position (3b) the cyclopentamer 5b and larger macrocycles are the predominant products. This may well be due to H-bonding interactions and/or steric effects. The same reaction with 1,4-dihydroxy-3,5-dimethoxybenzene (3c) could have led to very interesting macrocycles. Unfortunately, the reaction of 3c with paraformaldehyde in the presence of TFA at room temperature for 72 h gave a precipitate (Table 1, run 12), which was found to be insoluble in any of the common organic solvents tested. Reducing the reaction time to 1 h resulted in formation of a mixture of products that according to its 1H NMR and MS analyses contained both cyclotetramer 4c and cyclopentamer 5c together with other side products (ca. 35% crude). We assume that 3c, as a hydroquinone derivative, is sensitive to these opened-to-air reaction conditions. Next we proceeded to investigate the structures and the properties of the more promising macrocycles, i.e. the partially methylated calixtetrolarenes 4a and 5a. 1 H NMR spectra were collected over a range of temperatures for the CD2Cl2 solutions of 4a and 5a (Figures 3, 4). It appears reasonable to assume that both 4a and 5a would exhibit strong intramolecular hydrogen bonds between the hydroxyl groups at the lower rim, interactions that may result in cone structures. On the other hand, it is also reasonable to assume that the meta-methoxy groups could cause a steric hindrance in the upper rim. In a cone conformation the protons of the methylene bridges are not equivalent and are expected to exhibit an AB quartet. Indeed, when the solution of 4a was cooled to −65 °C the singlet at 3.88 ppm transformed to an AB pattern (Figure 3). The signals of the two types of methoxy groups at 3.76 and 3.74 ppm and the signal of the hydroxyl group at

Figure 4. 1H NMR spectra (500 MHz) of 5a in CD2Cl2 at various temperatures.

signal at 7.9 ppm that is attributed at room temperature to the five hydroxyls at the lower rim splits at −65 °C to three signals at 8.62 ppm (2OH), 7.10 ppm (2OH), and 6.54 (OH). These intramolecular hydrogen bonds may suggest a saddle conformation for the cyclopentamer 5a. Intramolecular hydrogen bonds can easily be identified by determining the Abraham solute hydrogen bond acidity factor (A) for a specific hydrogen.19 By measuring the chemical shifts of a hydrogen in CDCl3 vs DMSOd6, parameter A can be calculated through the equation A = 0.0065 + 0.133Δδ(DMSO−CDCl3). For a hydroxyl group, an A > 0.5 indicates that the OH does not participate in an intramolecular hydrogen bond. On the other hand A values of less than 0.1 indicate that the OH is part of an intramolecular hygrogen bond. We calculated the A parameters for the OH groups of 4a and 5a and found them to be low, i.e. −0.11 and 0.04, respectively, indicating that the hydroxyls are involved in significant intramolecular H-bonds. X-ray studies of the tetramer 4a showed that in the solid state this macrocycle adopts a pinched cone conformation (Figure 5).

Figure 5. Top and side views of the X-ray structure of 4a.

This is in line with the dynamic NMR analysis (that is consistent with both a cone and a pinched-cone conformation) and the A values computed for the hydroxyls of compound 4a in solutions. In the solid state the two parallel aromatic rings

Figure 3. 1H NMR spectra (500 MHz) of 4a in CD2Cl2 at various temperatures. 3721

DOI: 10.1021/acs.orglett.7b01511 Org. Lett. 2017, 19, 3719−3722

Organic Letters

■

ACKNOWLEDGMENTS This work was supported by Israel Science Foundation (ISF). Grant No. 804/13. We thank Dr. Sophia Lipstman (Tel-Aviv University) for solving the crystal structures.

(each has two OMe in the endo and one OMe in the exo conformation) are at a distance of 7.79 Å. The distance between the other two aromatic rings (each having one endo and two exo OMe groups) is 10.07 Å. One can argue that the conformation is a pinched cone and not a simple cone due to the steric repulsion between the meta-methoxy groups in the symmetrical cone conformation, as previously observed for octamethyltetrahydroxycalix[4]arene.20 The X-ray structure of 5a shows a much more complicated and distorted cyclopentameric structure (Figure 6). As suggested

■

by the dynamic NMR study and hydrogen-bond acidity analysis, all five hydroxyl groups of compounds 5a participate in intramolecular hydrogen bonds. However, while four hydroxyls are tightly bound to each other in a spiral-like arrangement, one hydroxyl lays in the opposite direction forming an intramolecular hydrogen bond with the methoxy group of the neighboring arene moiety. In this arrangement cyclopentamer 5a adopts a less sterically hindered structure, in which two adjacent arenes form two planes at an angle of 111.1° that lay against the other three aromatic rings, which adopt a zigzag array with angles of ∼115.1°. In conclusion we presented a simple and efficient one-step reaction for the synthesis of a new family of macrocycles, i.e. calix[4,5]tetrolarene derivatives, which can be regarded as pyrogallolarene−calixarene chimeras. To our satisfaction we were able to obtain relatively high isolated yields of the pentameric macrocycle 5a. We are currently studying the host−guest chemistry of 4a and 5a macrocycles and their specific alkylation/demethylation reactions with the aim of using them, inter alia, as building blocks for hydrogen-bonded capsules.21

ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.orglett.7b01511. General experimental procedures, 1H and 13C NMR spectra of 4a, 5a, and 5b; HSQC analysis of 5a; dynamic 1 H NMR spectra of 5b; X-ray data for 4a and 5a (PDF) Crystallographic data for 4a (CCDC 1542194) (CIF) Crystallographic data for 5a (CCDC 1542206) (CIF)

■

REFERENCES

(1) (a) Gutsche, C. D. Calixarenes an Introduction, 2nd ed.; the Royal Society of Chemistry: Cambridge, Thomas Graham House, 2008. (b) Böhmer, V. Angew. Chem., Int. Ed. Engl. 1995, 34, 713. (2) Recent reviews on pillararenes: (a) Ogoshi, T.; Yamagishi, T.-A.; Nakamoto, Y. Chem. Rev. 2016, 116, 7937. (b) Li, C. Chem. Commun. 2014, 50, 12420. (3) Selected reviews on cyclotriveratrylenes: (a) Hardie, M. J. Chem. Soc. Rev. 2010, 39, 516. (b) Collet, A. Tetrahedron 1987, 43, 5725. (4) For selected examples of calixarene related compounds, see: Calix[n]imidazolium derivatives: (a) Chun, Y.; Singh, N. J.; Hwang, I.C.; Lee, J. W.; Yu, S. U.; Kim, K. S. Nat. Commun. 2013, 4, 1797. Calixpyrrole analogues: (b) Lee, C.-H.; Miyaji, H.; Yoon, D.-W.; Sessler, J. L. Chem. Commun. 2008, 24. Resorcinarenes and pyrogallolarenes: (c) MacGillivray, L. R.; Atwood, J. L. Nature 1997, 389, 469. (d) Avram, L.; Cohen, Y.; Rebek, J., Jr. Chem. Commun. 2011, 47, 5368. Calixnaphthalenes: (e) Georghiou, P. E.; Li, Z.; Ashram, M.; Chowdhury, S.; Mizyed, S.; Tran, A. H.; Al-Saraierh, H.; Miller, D. O. Synlett 2005, 2005, 879. Azacalix[m]arene-[n]pyridines: (f) Wang, M.-X.; Zhang, X.-H.; Zheng, Q.-Y. Angew. Chem., Int. Ed. 2004, 43, 838. Calix[4]phloroglucinarene: (g) Ogoshi, T.; Kitajima, K.; Umeda, K.; Hiramitsu, S.; Kanai, S.; Fujinami, S.; Yamagishi, T.; Nakamoto, Y. Tetrahedron 2009, 65, 10644. (5) Neri, P., Sessler, J. L., Wang, M. X., Eds. Calixarenes and Beyond; Springer Int. Publishing: Switzerland, 2016. (6) Gutsche, G. D.; Muthukrishnan, R. J. Org. Chem. 1978, 43, 4905. (7) Ogoshi, T.; Kanai, S.; Fujinami, S.; Yamagishi, T.; Nakamoto, Y. J. Am. Chem. Soc. 2008, 130, 5022. (8) Chen, H.; Fan, J.; Hu, X.; Ma, J.; Wang, S.; Li, J.; Yu, Y.; Jia, X.; Li, C. Chem. Sci. 2015, 6, 197. Ma, J.; Meng, Q.; Hu, X.; Li, B.; Ma, S.; Hu, B.; Li, J.; Jia, X.; Li, C. Org. Lett. 2016, 18, 5740. (9) Schneebeli, S. T.; Cheng, C.; Hartlieb, K. J.; Strutt, N. L.; Sarjeant, A. A.; Stern, C. L.; Stoddart, J. F. Chem. - Eur. J. 2013, 19, 3860. (10) Boinski, T.; Cieszkowski, A.; Rosa, B.; Szumna, A. J. Org. Chem. 2015, 80, 3488. (11) Jia, F.; He, Z.; Yang, L.-P.; Pan, Z.-S.; Yi, M.; Jiang, R.-W.; Jiang, W. Chem. Sci. 2015, 6, 6731. (12) (a) Boinski, T.; Cieszkowski, A.; Rosa, B.; Lesniewska, B.; Szumna, A. New J. Chem. 2016, 40, 8892. (b) Zhou, J.; Yang, J.; Hua, B.; Shao, L.; Zhang, Z.; Yu, G. Chem. Commun. 2016, 52, 1622. (13) Stewart, D. R.; Gutsche, C. D. Org. Prep. Proced. Int. 1993, 25, 137. (14) Hirao, T.; Tosaka, M.; Yamago, S.; Haino, T. Chem. - Eur. J. 2014, 20, 16138. (15) Parisi, M. F.; Gattuso, G.; Notti, A.; Pisagatti, I.; Pappalardo, S. in “Calix[5]arene: from Capsules to Polymers” in ref 5. (16) Sala, P. D.; Gaeta, C.; Navarra, W.; Talotta, C.; De Rosa, M.; Brancatelli, G.; Geremia, S.; Capitelli, F.; Neri, P. J. Org. Chem. 2016, 81, 5726. (17) Boinski, T.; Szumna, A. Tetrahedron 2012, 68, 9419. (18) Ogoshi, T.; Ueshima, N.; Akutsu, T.; Yamafuji, D.; Furuta, T.; Sakakibara, F.; Yamagishi, T. Chem. Commun. 2014, 50, 5774. (19) Abraham, M. H.; Abraham, R. J.; Acree, W. E., Jr.; Aliev, A. E.; Leo, A. L.; Whaley, W. L. J. Org. Chem. 2014, 79, 11075. (20) Dahan, E.; Biali, S. E. J. Org. Chem. 1991, 56, 7269. (21) Selected examples of dimeric and larger capsules from our group: (a) Frish, L.; Matthews, S. E.; Bohmer, V.; Cohen, Y. J. Chem. Soc., Perkin Trans. 2 1999, 669. (b) Avram, L.; Cohen, Y. J. Am. Chem. Soc. 2002, 124, 15148. (c) Avram, L.; Cohen, Y. Org. Lett. 2002, 4, 4365. (d) Avram, L.; Cohen, Y. Org. Lett. 2003, 5, 3329. (e) Guralnik, V.; Avram, L.; Cohen, Y. Org. Lett. 2014, 16, 5592. (f) Yariv-Shoushan, S.; Cohen, Y. Org. Lett. 2016, 18, 936. (g) Avram, L.; Goldbourt, A.; Cohen, Y. Angew. Chem., Int. Ed. 2016, 55, 904.

Figure 6. Top and side views of the X-ray structure of 5a.

■

Letter

AUTHOR INFORMATION

Corresponding Authors

*E-mail: [email protected]. *E-mail: [email protected]. ORCID

Yossi Zafrani: 0000-0001-5977-528X Yoram Cohen: 0000-0002-9442-8547 Notes

The authors declare no competing financial interest. 3722

DOI: 10.1021/acs.orglett.7b01511 Org. Lett. 2017, 19, 3719−3722