Dissipative Assembly of Aqueous Carboxylic Acid Anhydrides Fueled

Article pubs.acs.org/JACS

Dissipative Assembly of Aqueous Carboxylic Acid Anhydrides Fueled by Carbodiimides Lasith S. Kariyawasam and C. Scott Hartley* Department of Chemistry & Biochemistry, Miami University, Oxford, Ohio 45056, United States S Supporting Information *

ABSTRACT: Biochemical systems make extensive use of chemically fueled processes (e.g., using ATP), but analogous abiotic systems remain rare. A key challenge is the identification of transformations that can be adapted to a range of applications and make use of readily available chemical fuels. In this context, the generation of transient covalent bonds is a fundamental tool for nonequilibrium systems chemistry. Here, we show that carbodiimides constitute a simple class of chemical fuels for dissipative assembly, taking advantage of their known reactivity to produce (hydrolytically unstable) anhydrides from carboxylic acids in water. Both aliphatic and aromatic anhydrides are formed on convenient time scales using the common, commercially available peptide coupling agent 1-ethyl-3-(3(dimethylamino)propyl)carbodiimide (EDC). An important feature of this reaction is that no part of the carbodiimide is incorporated into the transient species; that is, the fuel is decoupled from the structureand thus functionof the assembled state. We show that intramolecular anhydride formation of oligo(ethylene glycol) diacids gives macrocycles analogous to crown ethers, representing minimal examples of out-of-equilibrium supramolecular hosts. The kinetics and yields of macrocycle formation respond to cation guests, with the presence of matched cations decreasing their overall production.

■

INTRODUCTION The concepts of molecular self-assembly, referring to the spontaneous construction of larger chemical species from smaller building blocks,1 have enabled chemists to synthesize a remarkable diversity of molecular, supramolecular, and polymeric structures.2 We have become quite good at designing chemical systems at equilibrium and at engineering nonequilibrium kinetic traps.3 However, by their nature these systems are static:4 while they may be microscopically dynamic, their macroscopic properties are unchanging or respond only to external factors that affect the equilibrium condition. In contrast, while many elements of biological chemistry represent equilibrium states when considered in isolation (e.g., protein folding), biochemical systems are fundamentally dissipative, requiring a continuous input of energy that maintains function and is ultimately lost as heat. This consumption of energy enables complex functionality: dissipative systems are capable of responsive behavior that is impossible at equilibrium, including adaptation, self-healing, and replication.2,4 We focus here on chemical systems involving discrete molecular bond formation or geometry changes (as opposed to colloidal or macroscopic systems5−9). Photochemical processes form the basis for many out-of-equilibrium systems that have been developed in organic chemistry3 (e.g., Feringa’s molecular motors10). Conversely, in biology, chemical fuels (e.g., ATP) are ubiquitous; accordingly, many functional chemical systems have been demonstrated that repurpose biochemical machinery.11−15 Comparable fully abiotic systems remain rare:16,17 © 2017 American Chemical Society

while the design of equilibrium systems focuses on energy minima, and is enabled by simple tools in the form of various types of fast dynamic bonding (i.e., intermolecular interactions, dynamic covalent chemistry), nonequilibrium systems require a careful balancing of reaction rates and thus the consideration of more of their energy profiles.18 Our selection of basic tools for nonequilibrium covalent assembly is currently limited. Oscillating chemical systems have long been known, but are relatively complex to adapt.19,20 Systems that use the energy of chemical fuels simply to effect the formation of transient bonds are, in principle, more straightforward to use and broadly applicable. In seminal reports, Eelkema and van Esch demonstrated out-of-equilibrium supramolecular assembly fueled by the hydrolysis of methylating agents.21,22 Transient methylation of appropriate building blocks gave assembly of gels that exhibit functional behavior directly ascribed to their out-of-equilibrium state, including fuel-driven regeneration.22 In related examples, Leigh has demonstrated a catenane-based molecular motor system fueled by the decomposition of a chloroformate that acts as a dynamic stopper;23 Nitschke has demonstrated the transient disassembly of a supramolecular host;24 and Fyles has demonstrated the dissipative assembly of an artificial membrane transport system.25 Received: June 12, 2017 Published: August 4, 2017 11949

DOI: 10.1021/jacs.7b06099 J. Am. Chem. Soc. 2017, 139, 11949−11955

Article

Journal of the American Chemical Society

We became interested in the possibility that carbodiimides and related species could serve as useful chemical fuels for dissipative covalent systems chemistry, as shown in Figure 1c. Long used as reagents,28 carbodiimides are dehydrating agents that are kinetically stable in water. As such, they are ubiquitous as a means to activate carboxylic acids toward added nucleophiles (e.g., in peptide synthesis). As part of these reactions, carboxylic acid anhydrides are often implicated as intermediates;29 since anhydrides are themselves hydrolytically unstable, we reasoned that the coupling of carboxylic acids in water could comprise a simple demonstration of the chemistry in Figure 1b. There is also an interesting parallel with previous work on ketenes and ketenimines in systems related to molecular motors.30,31 Examples of the hydrolysis of carbodiimides in the presence of carboxylic acids are certainly known (e.g., in acetate buffers).32−36 The focus here is on the potential functionality of the anhydrides, as opposed to studies of carbodiimide decomposition. We show that the formation and decomposition of aqueous anhydrides fueled by carbodiimides do indeed occur on convenient time scales for both aliphatic and aryl carboxylic acids and are readily monitored by NMR spectroscopy. To demonstrate the usefulness of the method, we show that macrocycles can be synthesized from simple diacids. The products, analogues of crown ethers, represent minimal examples of transiently formed supramolecular hosts. We find counterintuitive effects of cation binding on the kinetics of macrocycle formation and destruction, highlighting the unusual properties of these out-of-equilibrium systems.

The controlled formation of an unstable covalent bond is key to each of these systems. In a sense, they exploit out-ofequilibrium analogues of dynamic covalent chemistry,26 and there is a need for reactions of this type that can be broadly applied.27 We can imagine, in general, two complementary strategies for fueled covalent bond formation. In the first, shown in Figure 1a, some part of a fuel (F1−F2) binds to the

■

RESULTS We began by focusing on two different model carboxylic acids, potassium sulfobenzoic acid (KSB-Ac) and methoxyethoxyacetic acid (MEA-Ac), which were treated with the common peptide coupling agent 1-ethyl-3-(3-(dimethylamino)propyl)carbodiimide (EDC),37 as shown in Figure 2a. The kinetics of these reactions should be pH-dependent, but standard buffer systems could unfortunately not be used, as they would crossreact with the EDC, anhydrides, or intermediates. Accordingly, the amount of EDC was chosen to keep the maximum conversion of the acids on the order of 50%, generating significant quantities of products while keeping pH changes manageable. In our first attempts, the reaction was carried out using just the carboxylic acids and EDC; however, under these conditions small amounts of two byproducts were observed that we tentatively assigned to the (expected) isomeric N-acyl ureas.38 The presence of a small amount of pyridine (2 mM) in the reaction mixtures suppresses the formation of these byproducts and was used in all subsequent experiments. Both systems are readily monitored by 1H NMR spectroscopy (in D2O), as shown in Figure 2b,c. Reactions of KSB-Ac occur at reasonable time scales at 298 K, whereas those of MEA-Ac are faster and were therefore carried out at 276 K. In both cases, the carboxylic acids are rapidly consumed with concomitant appearance of new species tentatively assigned as the corresponding anhydrides. These new species disappear over the course of minutes with the regeneration of the original acids. In the case of KSB-Ac, we also observed a small amount of another species at short reaction times that comprised moieties from both the acid and the EDC, which we believe to be an activated carboxylic acid intermediate (“Int”).33 In both cases, the EDC is cleanly converted to the corresponding urea.

Figure 1. (a) In the first type of fueled assembly, part of the fuel (F1) is incorporated into the transient state; (b) in the second type, only the energy content of the fuel is used. (c) The formation of aqueous anhydrides fueled by carbodiimides is a possible example of the second type.

substrate (M), giving a transient state (M−F1) that undergoes some further process. As it is part of the assembled species, the fuel-derived moiety F1 must be closely related to the function of the system. The examples listed above make use of this approach. Alternatively, as shown in Figure 1b, consumption of a fuel (F) could drive association of two components (M1 and M2) with no structural part of the fuel incorporated into the transient state (M1−M2). This mechanism is complementary to the first, but, to the best of our knowledge, has not yet been demonstrated. In this second type, because only the energy of the fuel is used, the function of the system is decoupled from the fuel structure. It follows that certain types of applications, requiring the assembly of complex components or intramolecular bond formation, should be easier to implement. Both general schemes in Figure 1a,b are, of course, simply catalytic cycles for the decomposition of the fuel, with the substrate (M) acting as a catalyst. It is likely that many suitable reactions are already known and simply must be identified. A priori, the ideal chemistry for dissipative assembly should possess several features: (1) The fuel itself should be readily available and easily handled. (2) The transformation should involve common functional groups that are easily synthesized. (3) Assembly and disassembly should occur on reasonable time scales and be easily monitored. 11950

DOI: 10.1021/jacs.7b06099 J. Am. Chem. Soc. 2017, 139, 11949−11955

Article

Journal of the American Chemical Society

Figure 2. (a) Both KSB-Ac and MEA-Ac undergo anhydride formation in water when treated with carbodiimide EDC. (b) 1H NMR monitoring of a representative EDC-fueled formation of KSB-An (298 K). The aromatic region of selected spectra are shown at right, with signals assigned to KSBAc, KSB-An, and an intermediate indicated. (c) 1H NMR monitoring of a representative EDC-fueled formation of MEA-An (276 K). Regions of selected spectra corresponding to the acid/anhydride methylene groups and the EDC/urea methyl groups are shown at right, with the appropriate signals indicated (the variation of the chemical shifts of some species is expected because of pH changes over the course of the experiment). In both (b) and (c), the solid lines correspond to fits to the kinetic model.

Several observations demonstrate that the transient species observed are indeed the carboxylic acid anhydrides KSB-An and MEA-An: (1) The time scales for the disappearance of the transient species are distinctly different from those for the disappearance/appearance of all EDC-related species, indicating that they derive from the starting acids only. This is obvious for the KSB-Ac system (Figure 2b), but also holds for the MEA-Ac system when the spectra are closely examined. (2) The deshielding of 1H NMR signals attributed to the anhydrides is consistent with expected substituent effects. This is further supported by independent synthesis of MEA-An (see Supporting Information). It is also noteworthy that the chemical shifts of the signals assigned to MEA-An are pHindependent, in contrast to those of MEA-Ac, as expected if the carboxylic acid group is consumed. (3) Importantly, IR spectra collected on both systems show the appearance and disappearance of peaks at approximately 1800 cm−1, characteristic of the symmetric carbonyl stretching mode of anhydrides, as shown in Figure 3 (the expected asymmetric stretching mode is presumably masked by bands assigned to the EDC and starting acids). (4) Monitoring of the pH of the reaction mixtures shows the rapid increases expected for consumption of the carboxylic acid groups, followed by returns to the starting values as the acids are regenerated (Figures S15, S16). In the future, similar pH changes could potentially be coupled to pHsensitive systems such as molecular switches.39

Figure 3. IR spectroscopy (transmission) of the formation/ decomposition of KSB-An (left) and MEA-An (right) in water after treatment with EDC (0.5 and 4 equiv, respectively). The KSB-An reaction was carried out at room temperature. The MEA-An reaction was carried out by mixing cold (ice bath) solutions and monitoring at room temperature (the inset shows the time evolution of the peak at 1835 cm−1).

The systems were further analyzed using the following simple kinetic model: k1

Ac + EDC → Int 11951

(1) DOI: 10.1021/jacs.7b06099 J. Am. Chem. Soc. 2017, 139, 11949−11955

Article

Journal of the American Chemical Society

Figure 4. (a) Oligo(ethylene glycol) diacids undergo anhydride formation when treated with EDC, giving a mixture of cyclic and oligomeric linear anhydrides. (b) 1H NMR monitoring of a representative EDC-fueled formation of TEG-Cy in the presence of 1.0 M NaCl (276 K). The downfield signal assigned to TEG-Cy is easily distinguished from the linear anhydrides. (c) A representative EDC-fueled formation of TEG-Cy in the presence of 1.0 M KCl (276 K). In both (b) and (c), solid lines represent fits to the kinetic model. kiAc

Int ⎯→ ⎯ Ac + Urea kiAn

Int + Ac ⎯→ ⎯ An + Urea k2

An → 2Ac

1b. They also provide a simple way of quantifying the reactions. In each case, the net anhydride produced (and destroyed), [An]net, is simply

(2) (3)

[An]net = k 2

(4)

where Ac and An represent the acid and anhydride species, respectively. The inclusion of eq 2, representing (unproductive) direct hydrolysis of the intermediate, was necessary in order to obtain reasonable fits to the experimental data. For the reactions of the MEA-Ac system, we assumed a steady state in the intermediate, as previously done for the reaction of acetic acid with carbodiimides;40 for the KSB-Ac system, because we observe a species (tentatively) assigned as an intermediate, we made no such assumption. As the goal was to describe only broad features of the systems, the models lack significant mechanistic detail: For example, they neglect the pH dependence of the rate constants; the complex, pH-dependent speciation of the EDC (note that these reactions should actually occur via the amidinium ion, not free carbodiimide33); and pyridine catalysis of both the EDC coupling and anhydride hydrolysis. Nevertheless, fits to the experimental data are very good (solid lines in Figure 2b,c). Although it is difficult to find direct comparisons in the literature, the rate constants for anhydride hydrolysis are broadly consistent with expectations based on well-known structure−property effects.41−43 The extracted parameters were well-reproduced over three replicates. The fits to the models show that the behavior of these systems is consistent with fueled assembly according to Figure

∫0

∞

[An] dt

(5)

(see discussion in the Supporting Information). Thus, while the net “yield” of the anhydrides is obviously 0% in the traditional sense, we can define a meaningful yield as the total anhydride produced relative to the total EDC used (i.e., the efficiency of the process). For the KSB-Ac and MEA-Ac systems, these yields are (81.4 ± 0.1)% and (25.1 ± 1.0)%, respectively. While these numbers are dependent on the starting conditions and are not directly comparable, it is clear that the aliphatic system is less efficient overall. We also tested the ability of the MEA-Ac system to tolerate multiple injections of EDC (Figure S20). The acid, once regenerated, can indeed be recoupled as expected, although the buildup of the urea byproduct becomes problematic. To be useful, the fueled chemistry must be coupled to some other functional behavior. A useful demonstration would be the transient assembly of supramolecular hosts: In principle, this could be used to create “catch and release” systems (e.g., for reusable sensing or to drive active transport of guests). As a proof-of-concept, we reasoned that simple carboxylic-acidterminated oligo(ethylene glycol)s could cyclize to give anhydride analogues of crown ethers. We focused on tetraethylene glycol and pentaethylene glycol diacids TEG-Ac and PEG-Ac,44 which were expected to give analogues of 18crown-6 and 21-crown-7, respectively, as shown in Figure 4a. 11952

DOI: 10.1021/jacs.7b06099 J. Am. Chem. Soc. 2017, 139, 11949−11955

Article

Journal of the American Chemical Society

Figure 5. Yields of linear and cyclic anhydrides from MEA-Ac, TEG-Ac, and PEG-Ac in the presence of 1.0 M LiCl, NaCl, KCl, or CsCl. Experiments were performed in triplicate; error bars correspond to the standard errors.

the DFT-optimized geometries of the K+ complexes of 18crown-6 and TEG-Cy suggested that the presence of the carbonyl groups would have only a small (but measurable) effect on the ability of the cyclic anhydrides to bind cations (see Supporting Information). As shown in Figure 5, the yields of MEA-An from monoacid MEA-Ac were found to be unaffected by the specific salt used (Figure 5, left). In contrast, it was immediately obvious from the NMR monitoring experiments that the presence of KCl has a dramatic effect on the production of TEG-Cy from the diacid TEG-Ac (Figure 4c). The net yields of TEG-Cy fall in the order Li+ > Na+ > Cs+ > K+ (Figure 5, center). This order is the exact opposite of that of the known cation affinities of 18-crown-6 in water (Li+ < Na+ < Cs+ < K+).45 The yields of the larger crown ether analogue PEG-Cy are less strongly affected by the presence of cations, but fall off gradually in the order Li+ > Na+ > K+ > Cs+ (Figure 5, right). Again, this trend is the opposite of the cation binding ability of 21-crown-7 in water (Li+ < Na+ < K+ < Cs+). The smaller overall effect on PEG-Cy compared to TEG-Cy is consistent with the generally lower affinities for these cations of 21-crown-7 compared to 18-crown-6. We were surprised that the presence of matched guest cations acts to suppress the yields of cyclic anhydrides, and that the yields of linear anhydrides are similarly suppressed. Because the identity of the cations has no effect on MEA-Ac and affects the two diacids differently, it is likely that the decreased yields are somehow related to host−guest binding in the analogous crown ethers: It is unlikely, for example, that coordination to the carbonyl groups is an important contributing factor. However, the lower yields cannot be a direct result of guest binding to TEG-Cy and PEG-Cy after they are formed. While accelerated hydrolysis of the macrocycles because of cation binding would shorten their lifetimes, it would not affect the total amount produced and thus not affect the yields (i.e., as in eq 5, larger hydrolysis rates constants compensate for smaller integrated concentrations). Indeed, the hydrolysis rate constants obtained from the kinetic fits suggest that decomposition of TEG-Cy and PEG-Cy is somewhat slower in the presence of their matched salts, suggesting (modest) stabilization through complexation (see Supporting Information). Binding of the macrocycles to cations also fails to explain the effect on the yields of the acyclic species. Instead, this appears to be a sort of “negative templation”; that is, the initial formation of the anhydrides is affected by recognition of their matched cations. Examination of the fit parameters suggests that the matched salts accelerate the direct hydrolysis of the activated carboxylic acid intermediates relative to anhydride formation (by analogy with eq 2 vs 3 in the MEA-

While these examples are very simple, they complement previous demonstrations of dissipative assembly21−25 for which comparable cyclization via intramolecular bond formation is not possible (i.e., because of the mechanism in Figure 1a). On treatment of either TEG-Ac or PEG-Ac with EDC, we observed the appearance of two 1H NMR singlets at similar chemical shifts to the one corresponding to the methylene groups adjacent to the anhydride in MEA-An (e.g., Figure 4b). We assigned the more upfield of the signals to oligomeric linear anhydrides and the more downfield of the signals to the cyclic anhydrides TEG-Cy or PEG-Cy on the basis of several observations: (1) The chemical shifts of the upfield signals are, in both cases, within 0.01 ppm of the signals in MEA-An, which bears a very close structural resemblance to the linear anhydrides. (2) DFT calculations (Figures S17, S18) show that deshielding is expected if the methylene protons are oriented syn to the carbonyl group, as expected for the cyclic anhydride. (3) Macrocycle TEG-Cy could be independently synthesized and characterized. Its chemical shift is indeed downfield of both MEA-An and separately synthesized linear anhydrides of TEG-Ac (in CDCl3, Figures S9−S11). (4) There are similar amounts of the species associated with the upfield signals in both the TEG-Ac and PEG-Ac systems, as expected for the linear oligomers, whereas there is less of the species associated with the downfield signals for PEG-Ac vs TEG-Ac, as expected for cyclization. In the absence of added salts, anhydride formation works well but hydrolysis is slightly too fast for convenient quantitative analysis (Figure S19). However, the reactions slow with increasing ionic strength and could be analyzed accurately in the presence of, for example, 1 M NaCl, as shown in Figure 4b. Modeling of the two systems by analogy with the monoacid systems (i.e., eqs 1−4) is complicated by the fact that the exact concentrations of TEG-Ac and PEG-Ac are not directly measurable by 1H NMR spectroscopy (i.e., their signals cannot be distinguished from those for the acid termini of oligomeric linear anhydrides). As discussed in detail in the Supporting Information, approximating the fraction of TEG-Ac and PEG-Ac by analogy with a simple condensation polymerization provides a simple model that fits the data well and allows calculation of yields (solid lines in Figure 4). As for MEA-An, we verified that TEG-Cy (and the corresponding linear anhydrides) can be repeatedly formed with multiple injections of EDC (Figure S21). Given that the defining property of crown ethers is their ability to bind cations, we were interested in testing how the presence of different salts would affect the assembly of TEGCy, PEG-Cy, and, for comparison, MEA-An. Comparison of 11953

DOI: 10.1021/jacs.7b06099 J. Am. Chem. Soc. 2017, 139, 11949−11955

Journal of the American Chemical Society

■

ACKNOWLEDGMENTS We thank the Donors of the American Chemical Society Petroleum Research Fund (56517-ND4) for support of this work.

Ac system). This effect would explain why the yields of both cyclic and linear anhydrides are affected since their formations have these intermediates in common. At this point, the molecular-level mechanism for this process is unclear. However, we note that, with one carboxylic acid group of TEG-Ac or PEG-Ac activated by EDC, coordination of the matched cation to the oligo(ethylene glycol) linker would bring the second carboxylic acid group into close proximity with the first, which could affect its reactivity.

■

CONCLUSIONS The results show that carbodiimide-fueled coupling of carboxylic acids is a promising, very simple strategy for dissipative systems chemistry. As a proof-of-principle, we have demonstrated the assembly of crown ether analogues as minimal examples of supramolecular hosts. Intriguingly, the kinetics of assembly are sensitive to the presence of guests for the hosts, with an apparent negative templation effect whereby matched cations suppress anhydride formation. This counterintuitive observation underscores the complexity of nonequilibrium assembly. Even in these simple systems the molecular-level details of these processes are complex, and further study will be required to fully understand their mechanisms and to improve overall efficiency. A particular disadvantage of the current system is that it is not yet possible to use excess fuel because byproduct formation increases when the concentration of available acid is very low. Nevertheless, this system offers several very useful properties. First, only the energy content of the fuel is used; it is not a structural component of either the assembled or disassembled states of the system. This complements existing examples of dissipative assembly, in particular by enabling intramolecular coupling, and (in principle) by allowing increased structural complexity in the coupling partners. Second, the fuel itself is a commercially available reagent and couples one of the most common functional groups in organic chemistry. Third, the chemistry occurs over convenient time scales (minutes) and, because it is partly based on simple hydrolysis reactions, should be readily tunable. Efforts to apply this reaction to molecular geometry changes and more complex assemblies are currently underway. ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/jacs.7b06099.

■

REFERENCES

(1) Whitesides, G. M.; Mathias, J. P.; Seto, C. T. Science 1991, 254, 1312−1319. (2) Whitesides, G. M.; Grzybowski, B. Science 2002, 295, 2418−2421. (3) Mattia, E.; Otto, S. Nat. Nanotechnol. 2015, 10, 111−119. (4) Fialkowski, M.; Bishop, K. J. M.; Klajn, R.; Smoukov, S. K.; Campbell, C. J.; Grzybowski, B. A. J. Phys. Chem. B 2006, 110, 2482− 2496. (5) Burns, M. M.; Fournier, J.-M.; Golovchenko, J. A. Science 1990, 249, 749−754. (6) Grzybowski, B. A.; Stone, H. A.; Whitesides, G. M. Nature 2000, 405, 1033−1036. (7) Wang, W.; Duan, W.; Ahmed, S.; Sen, A.; Mallouk, T. E. Acc. Chem. Res. 2015, 48, 1938−1946. (8) Dreyfus, R.; Baudry, J.; Roper, M. L.; Fermigier, M.; Stone, H. A.; Bibette, J. Nature 2005, 437, 862−865. (9) Dey, K. K.; Sen, A. J. Am. Chem. Soc. 2017, 139, 7666−7676. (10) Koumura, N.; Zijlstra, R. W. J.; van Delden, R. A.; Harada, N.; Feringa, B. L. Nature 1999, 401, 152−155. (11) Dhiman, S.; Jain, A.; George, S. J. Angew. Chem., Int. Ed. 2017, 56, 1329−1333. (12) Maiti, S.; Fortunati, I.; Ferrante, C.; Scrimin, P.; Prins, L. J. Nat. Chem. 2016, 8, 725−731. (13) Angulo-Pachón, C. A.; Miravet, J. F. Chem. Commun. 2016, 52, 5398−5401. (14) Pappas, C. G.; Sasselli, I. R.; Ulijn, R. V. Angew. Chem., Int. Ed. 2015, 54, 8119−8123. (15) Debnath, S.; Roy, S.; Ulijn, R. V. J. Am. Chem. Soc. 2013, 135, 16789−16792. (16) Fletcher, S. P.; Dumur, F.; Pollard, M. M.; Feringa, B. L. Science 2005, 310, 80−82. (17) Collins, B. S. L.; Kistemaker, J. C. M.; Otten, E.; Feringa, B. L. Nat. Chem. 2016, 8, 860−866. (18) Moberg, C. Acc. Chem. Res. 2016, 49, 2736−2745. (19) Lagzi, I.; Kowalczyk, B.; Wang, D.; Grzybowski, B. A. Angew. Chem., Int. Ed. 2010, 49, 8616−8619. (20) Tamate, R.; Akimoto, A. M.; Yoshida, R. Chem. Rec. 2016, 16, 1852−1867. (21) Boekhoven, J.; Brizard, A. M.; Kowlgi, K. N. K.; Koper, G. J. M.; Eelkema, R.; van Esch, J. H. Angew. Chem., Int. Ed. 2010, 49, 4825− 4828. (22) Boekhoven, J.; Hendriksen, W. E.; Koper, G. J. M.; Eelkema, R.; van Esch, J. H. Science 2015, 349, 1075−1079. (23) Wilson, M. R.; Solà, J.; Carlone, A.; Goldup, S. M.; Lebrasseur, N.; Leigh, D. A. Nature 2016, 534, 235−240. (24) Wood, C. S.; Browne, C.; Wood, D. M.; Nitschke, J. R. ACS Cent. Sci. 2015, 1, 504−509. (25) Dambenieks, A. K.; Vu, P. H. Q.; Fyles, T. M. Chem. Sci. 2014, 5, 3396−3403. (26) Rowan, S. J.; Cantrill, S. J.; Cousins, G. R. L.; Sanders, J. K. M.; Stoddart, J. F. Angew. Chem., Int. Ed. 2002, 41, 898−952. (27) Ashkenasy, G.; Hermans, T. M.; Otto, S.; Taylor, A. F. Chem. Soc. Rev. 2017, 46, 2543−2554. (28) Khorana, H. G. Chem. Rev. 1953, 53, 145−166. (29) Nakajima, N.; Ikada, Y. Bioconjugate Chem. 1995, 6, 123−130. (30) Kelly, T. R.; De Silva, H.; Silva, R. A. Nature 1999, 401, 150− 152. (31) Mock, W. L.; Ochwat, K. J. J. Phys. Org. Chem. 2003, 16, 175− 182. (32) Díaz Pérez, V. M.; Ortiz Mellet, C.; Fuentes, J.; García Fernández, J. M. Carbohydr. Res. 2000, 326, 161−175. (33) Ibrahim, I. T.; Williams, A. J. Am. Chem. Soc. 1978, 100, 7420− 7421.

■

■

Article

Supplemental figures and discussion referred to in the text, experimental procedures, NMR spectra, and computational data (PDF) Python script for kinetic fits (TXT)

AUTHOR INFORMATION

Corresponding Author

*[email protected] ORCID

C. Scott Hartley: 0000-0002-5997-6169 Notes

The authors declare no competing financial interest. 11954

DOI: 10.1021/jacs.7b06099 J. Am. Chem. Soc. 2017, 139, 11949−11955

Article

Journal of the American Chemical Society (34) Ibrahim, I. T.; Williams, A. J. Chem. Soc., Chem. Commun. 1980, 25−27. (35) Williams, A.; Ibrahim, I. T. J. Am. Chem. Soc. 1981, 103, 7090− 7095. (36) Ibrahim, I. T.; Williams, A. J. Chem. Soc., Perkin Trans. 2 1982, 1459−1466. (37) Sheehan, J. C.; Cruickshank, P. A.; Boshart, G. L. J. Org. Chem. 1961, 26, 2525−2528. (38) Williams, A.; Ibrahim, I. T. Chem. Rev. 1981, 81, 589−636. (39) Tatum, L. A.; Su, X.; Aprahamian, I. Acc. Chem. Res. 2014, 47, 2141−2149. (40) DeTar, D. F.; Silverstein, R. J. Am. Chem. Soc. 1966, 88, 1013− 1019. (41) Bunton, C. A.; Fuller, N. A.; Perry, S. G.; Shiner, V. J. Tetrahedron Lett. 1961, 2, 458−460. (42) Berliner, E.; Altschul, L. H. J. Am. Chem. Soc. 1952, 74, 4110− 4113. (43) Brandão, T. A. S.; Dal Magro, J.; Chiaradia, L. D.; Nascimento, M. d. G.; Nome, F.; Tato, J. V.; Yunes, R. A. J. Phys. Org. Chem. 2004, 17, 370−375. (44) Wittmann, V.; Takayama, S.; Gong, K. W.; Weitz-Schmidt, G.; Wong, C.-H. J. Org. Chem. 1998, 63, 5137−5143. (45) Izatt, R. M.; Pawlak, K.; Bradshaw, J. S.; Bruening, R. L. Chem. Rev. 1991, 91, 1721−2085.

11955

DOI: 10.1021/jacs.7b06099 J. Am. Chem. Soc. 2017, 139, 11949−11955