Synthesis of Carbohydrate Based Macrolactones and Their Applications as Receptors for Ion Recognition and Catalysis

Glycomacrolactones exhibit many interesting biological properties, and they are also important in molecular recognitions and for supramolecular chemistry. Therefore, it is important to be able to access glycomacrocycles with different sizes and functionality. A new series of carbohydrate-based macrocycles containing triazole and lactone moieties have been designed and synthesized. The synthesis features an intramolecular nucleophilic substitution reaction for the macrocyclization step. In this article, the effect of some common sulfonate leaving groups is evaluated for macrolactonization. Using tosylate gave good selectivity for monolactonization products with good yields. Fourteen different macrocycles have been synthesized and characterized, of which eleven macrocycles are from cyclization of the C1 to C6 positions of N-acetyl D-glucosamine derivatives and three others from C2 to C6 cyclization of functionalized D-glucosamine derivatives. These novel macrolactones have unique structures and demonstrate interesting anion binding properties, especially for chloride. The macrocycles containing two triazoles form complexes with copper sulfate, and they are effective ligands for copper sulfate mediated azide-alkyne cycloaddition reactions (CuAAC). In addition, several macrocycles show some selectivity for different alkynes.


Introduction
Macrocyclic compounds are important classes of molecules with many biological applications and useful for drug discovery [1,2]. Carbohydrate-based macrocyclic compounds have unique molecular architectures and many practical applications [3,4]. Naturally existing and synthetic carbohydrate-based macrocycles have been utilized in drug discovery, molecular recognition, and as advanced functional materials [1,3,[5][6][7]. Among the different classes of macrocycles, macrolactones are especially unique compounds that exhibit biological activities and often function as enzyme inhibitors; they also showed applications in molecular recognition for supramolecular chemistry. For example, macrolactones synthesized from cholesterol derivatives have shown utility for molecular recognitions [8], and a synthetic macrodilactone exhibited enantioselective recognition of amino alcohols, as well as metal ions [9]. Naturally existing carbohydrate-based macrolactones or macrolides often exhibit many desirable biological activities for drug development [10][11][12]. The structures of several macrolactones are shown in Figure 1. Erythromycin (1) is a sugar-containing macrolide antibiotic. Ipomoeassina F (2) and analogs are macrolactones containing a disaccharide unit, which exhibit anticancer activities and recently have been identified as protein-translocation inhibitors [5,13]. Glucolipsin A (3) and analogs are natural products containing sugar dilactones-they are inhibitors for dual specific phosphatase Cdc25A [14]. This compound was synthesized from the corresponding hydroxy acids through a macrodilactonization reaction mediated by 2-chloro-1,3-dimethylimidazolinium chloride.
Using a hexapyranose as the scaffold, a variety of synthetic macrolactones can be prepared depending on the cyclization patterns from different positions of the sugar These compounds contain either one or two triazole moieties and are expected have properties in binding to guest molecules or ions, resulting in useful structures an applications.

Results and Discussion
As shown in Scheme 1, the macrolactones with general structures 5A can be synth These compounds contain either one or two triazole moieties and are expected to have properties in binding to guest molecules or ions, resulting in useful structures and applications.

Results and Discussion
As shown in Scheme 1, the macrolactones with general structures 5A can be synthesized starting from the intermediate 3S, which was prepared from N-acetyl-D-glucosamine (NAG) [35]. By installing a suitable leaving group at the C-6 position and a nucleophile, such as a long-chain and triazole-containing carboxylate linked to the anomeric center, an intramolecular S N 2 reaction of the carboxylate with the leaving group is expected to form the macrolactones. The carboxylate intermediate can be prepared using click chemistry to include one or two triazole functional groups.
To test the feasibility of this route to synthesize the glycomacrolactones, several different sulfonate leaving groups towards the macrocyclization were evaluated to obtain suitable conditions for the cyclization reactions. Common aryl sulfonyl chlorides were selected for the derivatiazation of the C-6-primary hydroxyl group of S3-these include p-bromobenzene-, p-chlorobenzene-, p-tolyl-, and 1-naphthyl-sulfonyl chloride. Methanesulfonyl chloride was not selected, since it typically reacts with both the C-4 and C-6 hydroxyl groups readily [36]. The C-6 primary hydroxyl group of S3 was then converted to different sulfonate leaving groups selectively to afford 6a-6d. The intermediates 6a-6c were treated with benzoyl chloride to afford the corresponding dibenzoates 7, with both the C-3 and C-4 positions benzoylated. However, the naphthyl derivative 6d afforded mainly the C-3 monobenzoylation product, presumably due to steric hindrance. As shown in Scheme 2, treating 7c and 9 [35] with different alkynoic acids 10a-c utilizing click reaction gave the carboxylic acids 11a-12c. The intramolecular SN2 reaction was carried out using K2CO3 as the base in DMF at 75 °C to afford monotriazole macrolactones 13a-13c, 14a-14c with different ring sizes in excellent yields. The synthesis of bistriazole-containing macrolactones is shown in Scheme 3. Using a similar method and the triazole-containing alkynoic acids 10d-e, the desired macrocycles 16a-d were obtained in Scheme 1. Macrolactonization using different sulfonates as the leaving groups.
The 4-bromobenzenesulfonate 6a was converted to the corresponding dibenzoate 7a followed by a click reaction to afford the intermediate 8a. The intramolecular macrolac-tonization reaction led to the monomeric lactone LM28 in 72% yield, together with 20% dimeric lactonization product DLM28. The 4-chlorosulfonate 6b was converted to chloride during the benzoylation step to afford the chloride 7b. Using a similar protocol, compound 7b was converted the LM28 in 67% yield together with 17% DLM28; and the tosylate 8c led to 82% LM28 and no dimer was isolated. Based on these, the tosyl group was superior to other leaving groups, and this was selected to synthesize a series of other macrocycles containing monotriazole and bis-triazoles (Schemes 2 and 3).
As shown in Scheme 2, treating 7c and 9 [35] with different alkynoic acids 10a-c utilizing click reaction gave the carboxylic acids 11a-12c. The intramolecular S N 2 reaction was carried out using K 2 CO 3 as the base in DMF at 75 • C to afford monotriazole macrolactones 13a-13c, 14a-14c with different ring sizes in excellent yields. The synthesis of bis-triazole-containing macrolactones is shown in Scheme 3. Using a similar method and the triazole-containing alkynoic acids 10d-e, the desired macrocycles 16a-d were obtained in 78-90% yields. The concentrations of the cyclization step for both monotriazole and bistriazole-containing macrocycles were typically~10 mM or 0.01 M of the reactants, which are not super dilute conditions; therefore, the method should be practical for possible large-scale synthesis.
Besides using DMF as the solvent, for the intramolecular lactonization reaction, acetonitrile was also used as the solvent; similar macrolactonization results were typically obtained through the reactions that required a longer time to complete. The monotriazole derivative LM28 was obtained in 77% yield in acetonitrile (9 mM concentration 6 h). The bistriazole-based macrocycles were also prepared in acetonitrile overnight to afford the macrocycle products in 71-78% yields. The typical concentration for cyclization for small scale reactions was about 4-6 mM; though the yields were slightly lower than those reactions when DMF was used, the lower boiling point of acetonitrile makes it easier to remove during work up. In addition, in acetonitrile, no dimerization was observed, perhaps due to reduced nucleophilicities of the carboxylate in acetonitrile than DMF. Using the macrolactonization method developed above, several macrocycles with the general structure 5B can be synthesized. As shown in Scheme 4, the azido sugar derivative 17 [37] was converted to the key intermediate tosylate 20 in three steps. For the benzoylation step of compound 19, pyridine was found to be a more effective solvent. The yields of the dibenzoylated products improved from 74% in DCM/pyridine to 91% in pyridine. The intermediate azide 20 was treated with different alkyne-containing acids to generate the precursors for macrolactonizations. As a proof of principle, here, only two alkynyl acids 10c and 10d, were used. The intermediate 21 was converted to the monolactonziation product 22 in good yield under similar conditions. Which indicated that the macrolactonization methods are applicable to the synthesis of glycomacrolactones of different positions to the C-6.
The bistriazole derivative 23 was subjected to macrolactonization using different bases to obtain suitable conditions for monocyclization. Three different carbonate bases in anhydrous DMF were screened, and the results are shown in Table 1. It seems that the cations used can affect the ratio of the monomer versus the dimer; all conditions showed 100% conversion of the starting materials based on the disappearance of the tosylate signal. Using sodium and potassium carbonate, the monolactone 24 was obtained, and when using cesium carbonate, mainly the dimeric-lactone 25 was isolated. In 5% MeOH/DCM, the monolactone was slightly more polar than the dilactone, which showed an opposite trend to the C1-series. Using either Na 2 CO 3 or K 2 CO 3 , the monolactonization products were obtained in good yields for the C2 series macrocycles.
100% conversion of the starting materials based on the disappearance of the tosylate signal. Using sodium and potassium carbonate, the monolactone 24 was obtained, and when using cesium carbonate, mainly the dimeric-lactone 25 was isolated. In 5% MeOH/DCM, the monolactone was slightly more polar than the dilactone, which showed an opposite trend to the C1-series. Using either Na2CO3 or K2CO3, the monolactonization products were obtained in good yields for the C2 series macrocycles.

Anion Binding Studies of the Macrocycles with Tetrabutyl Ammonium Halides
As mentioned earlier, triazole derivatives in rigid macrocycles have shown strong binding to chloride ions from the C-H bond [38,39]. In order to understand the anion binding capacities of the macrocycles, several representative macrocycles, including the C1 series monotriazole LM28, bistriazole DM35, and C-2 series monotriazole 22 and bistriazole 24, were studied. The tetrabutylammonium halides (TBAX), including TBACl, TBABr, and TBAI, were used to study halide binding. The 1 H NMR spectra of the macrocycles with different amounts of halides were obtained, and possible anion recognitions were analyzed. The structures and 1 H NMR spectra of several representative macrocycles with TBACl are shown in

Anion Binding Studies of the Macrocycles with Tetrabutyl Ammonium Halides
As mentioned earlier, triazole derivatives in rigid macrocycles have shown strong binding to chloride ions from the C-H bond [38,39]. In order to understand the anion binding capacities of the macrocycles, several representative macrocycles, including the C1 series monotriazole LM28, bistriazole DM35, and C-2 series monotriazole 22 and bistriazole 24, were studied. The tetrabutylammonium halides (TBAX), including TBACl, TBABr, and TBAI, were used to study halide binding. The 1 H NMR spectra of the macrocycles with different amounts of halides were obtained, and possible anion recognitions were analyzed.   The 1 H NMR spectra of LM28 with TBACl are shown in Figure 3, Figures S1 and S2. The triazole C-H and the amide NH signals of LM28 showed significant chemical shift changes upon adding halides. From 0 to 10.0 equiv of TBACl, the NH showed the most significant downfield shift of 0.59 ppm-this indicated strong hydrogen bonding with the chloride. The triazole signal also shifted downfield, with 2.0 equiv of TBACl, the Ha moved downfield by 0.16 ppm; interestingly, this signal only shifted another 0.04 ppm when 10.0 equiv of chloride was added. The significant downfield change of triazole and NH signals with increasing amounts of chloride indicated that they were affected by forming hydrogen bonds with chloride and by anion-π interactions. Other protons were less affected by chloride additions and showed small upfield shifts. The anomeric proton shifted upfield by 0.04 ppm, H 3 by 0.05 ppm, and H 4 by 0.10 ppm. Without chloride, the Hb, Hc, and Hd are multiplets centered at δ 4.73, 4.57, and 4.39 ppm, respectively, and the Hb and Hc each showed doublet of triplet splitting pattern (dt). However, with 10.0 equiv of TBACl, the Hb and Hc merged to a pseudo triplet. These changes indicated that the chloride binding to the ring caused significant chemical and magnetic environment changes for the macrocycle protons. The anion binding did not affect the conformation of the sugar ring much, but did change the dihedral angles of anomeric methylene groups ( Figure 2). The 1 H NMR spectra of LM28 with TBACl are shown in Figures 3, S1 and S2. The triazole C-H and the amide NH signals of LM28 showed significant chemical shift changes upon adding halides. From 0 to 10.0 equiv of TBACl, the NH showed the most significant downfield shift of 0.59 ppm-this indicated strong hydrogen bonding with the chloride. The triazole signal also shifted downfield, with 2.0 equiv of TBACl, the Ha moved downfield by 0.16 ppm; interestingly, this signal only shifted another 0.04 ppm when 10.0 equiv of chloride was added. The significant downfield change of triazole and NH signals with increasing amounts of chloride indicated that they were affected by forming hydrogen bonds with chloride and by anion-π interactions. Other protons were less affected by chloride additions and showed small upfield shifts. The anomeric proton shifted upfield by 0.04 ppm, H3 by 0.05 ppm, and H4 by 0.10 ppm. Without chloride, the Hb, Hc, and Hd are multiplets centered at δ 4.73, 4.57, and 4.39 ppm, respectively, and the Hb and Hc each showed doublet of triplet splitting pattern (dt). However, with 10.0 equiv of TBACl, the Hb and Hc merged to a pseudo triplet. These changes indicated that the chloride binding to the ring caused significant chemical and magnetic environment changes for the macrocycle protons. The anion binding did not affect the conformation of the sugar ring much, but did change the dihedral angles of anomeric methylene groups ( Figure 2).  The 1 H NMR spectra of LM28 with TBABr and TBAI are shown in Figures S3-S8; the patterns are similar to those of chloride, but with much smaller chemical shift changes. From 0 to 5.0 equiv of TBABr, the triazole signal showed about 0.10 ppm downfield shift, and the NH signal shifted downfield by 0.19 ppm, the anomeric proton showed a small upfield shift of 0.03 ppm. These results showed that the bromide also forms hydrogen bonds with the triazole hydrogen, amide hydrogen cooperatively. But the bromide binding is apparently weaker than chloride. TBAI had even less influence on chemical shift changes ( Figures S6-S8). From 0 to 5.0 equiv of TBAI, the triazole and the NH signals only showed small downfield changes of 0.03 and 0.06 ppm, respectively. This indicated that iodide didn't form strong interactions with the triazole and the macrocycle.
The 1 H NMR spectra of DM35 with different TBAX are shown in Figures S9-S17. For 0 to 5.0 equiv of TBACl (Figures S9-S11), the amide NH signal showed a 0.14 ppm downfield shift, and the anomeric proton showed a similar upfield change of 0.04 ppm. The triazole signals showed a small downfield shift of 0.04 ppm when adding up to 2 equiv of TBACl ( Figure S10); however, not much change was observed after adding additional TBACl. This could be due to the conformation change caused by the anion binding with the sugar ring. A similar trend was observed for TBABr, as shown in Figures S12-S14, from 0 to 5 equiv of TBAB, the NH signal moved downfield by 0.04 ppm, and the anomeric proton showed an upfield change of 0.03 ppm. The triazole signals showed concentration dependence. Upon addition of one equiv of TBABr, the two triazole signals showed 0.08 ppm downfield change; after further addition of TBABr, the triazole signals did not move ( Figure S13). This indicates that possibly one bromide is hydrogen bonding with both triazole protons, and bromide formed stronger interactions than chloride did. TBAI did not cause significant chemical shift changes (Figures S15-S17).
The 1 H NMR spectra of the C-2 series macrocycles 22 are shown in Figure 4, Figures S18 and S19. The triazole signal showed a small downfield (0.03 ppm) change after adding one equiv of TBACl. From 0 to 10.0 equiv of TBACl, the amide NH showed a larger downfield change of 0.19 ppm, the other protons from the sugar ring also showed significant changes ( Figure 4). The anomeric proton showed an upfield shift of 0.07 ppm.
The protons at C-3 and C-4 positions became more resolved into two triplets, and the methylene Hb and Hc merged from two separate doublets to a pseudo quartet. This indicated the chemical environment of the two protons become very similar.
iodide didn't form strong interactions with the triazole and the macrocycle.
The 1 H NMR spectra of DM35 with different TBAX are shown in Figures S9-S17. For 0 to 5.0 equiv of TBACl ( Figures S9-11), the amide NH signal showed a 0.14 ppm downfield shift, and the anomeric proton showed a similar upfield change of 0.04 ppm. The triazole signals showed a small downfield shift of 0.04 ppm when adding up to 2 equiv of TBACl ( Figure S10); however, not much change was observed after adding additional TBACl. This could be due to the conformation change caused by the anion binding with the sugar ring. A similar trend was observed for TBABr, as shown in Figures S12-S14, from 0 to 5 equiv of TBAB, the NH signal moved downfield by 0.04 ppm, and the anomeric proton showed an upfield change of 0.03 ppm. The triazole signals showed concentration dependence. Upon addition of one equiv of TBABr, the two triazole signals showed 0.08 ppm downfield change; after further addition of TBABr, the triazole signals did not move ( Figure S13). This indicates that possibly one bromide is hydrogen bonding with both triazole protons, and bromide formed stronger interactions than chloride did. TBAI did not cause significant chemical shift changes (Figures S15-S17).
The 1 H NMR spectra of the C-2 series macrocycles 22 are shown in Figures 4, S18 and S19. The triazole signal showed a small downfield (0.03 ppm) change after adding one equiv of TBACl. From 0 to 10.0 equiv of TBACl, the amide NH showed a larger downfield change of 0.19 ppm, the other protons from the sugar ring also showed significant changes ( Figure 4). The anomeric proton showed an upfield shift of 0.07 ppm. The protons at C-3 and C-4 positions became more resolved into two triplets, and the methylene Hb and Hc merged from two separate doublets to a pseudo quartet. This indicated the chemical environment of the two protons become very similar. The C2-bistriazole macrocycle 24 binding to chloride was also studied, as shown in Figures 5 and S20-S22. The triazole, anomeric, NH, and many other signals changed significantly upon adding TBACl. From 0 to 10.0 equiv of TBACl, the triazole proton Ha at 7.49 ppm shifted downfield by 0.10 ppm; and Hb at 7.37 shows a small downfield (0.03 ppm) change, the amide NH shifted downfield by 0.31 ppm. This indicated that the amide participated in intermolecular hydrogen bonding significantly, and one of the triazole hydrogen atoms also formed a stronger hydrogen bond with chloride than the other. The H3 The C2-bistriazole macrocycle 24 binding to chloride was also studied, as shown in Figure 5 and Figures S20-S22. The triazole, anomeric, NH, and many other signals changed significantly upon adding TBACl. From 0 to 10.0 equiv of TBACl, the triazole proton Ha at 7.49 ppm shifted downfield by 0.10 ppm; and Hb at 7.37 shows a small downfield (0.03 ppm) change, the amide NH shifted downfield by 0.31 ppm. This indicated that the amide participated in intermolecular hydrogen bonding significantly, and one of the triazole hydrogen atoms also formed a stronger hydrogen bond with chloride than the other. The H 3 and H 4 appeared as a pseudo triplet; H 4 shifted upfield by 0.05 ppm, while H 3 stayed the same. The methylene Hc and Hd changed from a pseudo quartet to a broad singlet, indicating that they have a similar chemical and the magnetic environment with 5.0-10.0 equiv of TBACl. These shifts showed that the macrocycle can bind to chloride, and the binding of anions resulted in a conformational change of the macrocycle. and H4 appeared as a pseudo triplet; H4 shifted upfield by 0.05 ppm, while H3 stayed the same. The methylene Hc and Hd changed from a pseudo quartet to a broad singlet, indicating that they have a similar chemical and the magnetic environment with 5.0-10.0 equiv of TBACl. These shifts showed that the macrocycle can bind to chloride, and the binding of anions resulted in a conformational change of the macrocycle. The binding properties of the macrocycles LM28, DM35 with Cu 2+ were also studied using several macrolactones and 1 H NMR spectroscopy at variable temperatures. The 1 H NMR spectra of DM35 and its complex with CuSO4•5H2O are shown in Figures 6 and S23-S25. In the complex, the two sharp triazole singlets at 7.70 and 7.69 ppm disappeared or became broadened and shifted downfield to around 7.81 ppm. The triazole signals were very broadened due to the paramagnetic properties of Cu (II) binding to triazole [40]. Typically, protons within 0.9 nm of the Cu (II) were not observed due to fast paramagnetic relaxation, and the protons in the outer shell between 0.9-1.7 nm from the copper showed chemical shift changes and broad resonances [41]. The signals far away from Cu (II) were not affected as much ( Figure S23). This showed that the copper ion is in close vicinity to the triazoles. The 1 H NMR spectra of the copper complex with DM35 were also evaluated at different temperatures, from 30 to 60 °C (Figures 7, S24 and S25), the triazole signals showed upfield shift to a broad signal, but more defined peak at around 7.71 ppm, this is close to the chemical shifts of macrocycle without Cu (II), which indicates that the Cu (II) is more dissociated from the binding to the two triazole nitrogen atoms at elevated temperatures. The anomeric proton appeared broad singlet at 30 °C, at 4.79 ppm, and shifted to 4.81 to The binding properties of the macrocycles LM28, DM35 with Cu 2+ were also studied using several macrolactones and 1 H NMR spectroscopy at variable temperatures. The 1 H NMR spectra of DM35 and its complex with CuSO 4 ·5H 2 O are shown in Figure 6 and Figures S23-S25. In the complex, the two sharp triazole singlets at 7.70 and 7.69 ppm disappeared or became broadened and shifted downfield to around 7.81 ppm. The triazole signals were very broadened due to the paramagnetic properties of Cu (II) binding to triazole [40]. Typically, protons within 0.9 nm of the Cu (II) were not observed due to fast paramagnetic relaxation, and the protons in the outer shell between 0.9-1.7 nm from the copper showed chemical shift changes and broad resonances [41]. The signals far away from Cu (II) were not affected as much ( Figure S23). This showed that the copper ion is in close vicinity to the triazoles. and H4 appeared as a pseudo triplet; H4 shifted upfield by 0.05 ppm, while H3 stayed the same. The methylene Hc and Hd changed from a pseudo quartet to a broad singlet, indicating that they have a similar chemical and the magnetic environment with 5.0-10.0 equiv of TBACl. These shifts showed that the macrocycle can bind to chloride, and the binding of anions resulted in a conformational change of the macrocycle. The binding properties of the macrocycles LM28, DM35 with Cu 2+ were also studied using several macrolactones and 1 H NMR spectroscopy at variable temperatures. The 1 H NMR spectra of DM35 and its complex with CuSO4•5H2O are shown in Figures 6 and S23-S25. In the complex, the two sharp triazole singlets at 7.70 and 7.69 ppm disappeared or became broadened and shifted downfield to around 7.81 ppm. The triazole signals were very broadened due to the paramagnetic properties of Cu (II) binding to triazole [40]. Typically, protons within 0.9 nm of the Cu (II) were not observed due to fast paramagnetic relaxation, and the protons in the outer shell between 0.9-1.7 nm from the copper showed chemical shift changes and broad resonances [41]. The signals far away from Cu (II) were not affected as much ( Figure S23). This showed that the copper ion is in close vicinity to the triazoles. The 1 H NMR spectra of the copper complex with DM35 were also evaluated at different temperatures, from 30 to 60 °C (Figures 7, S24 and S25), the triazole signals showed upfield shift to a broad signal, but more defined peak at around 7.71 ppm, this is close to the chemical shifts of macrocycle without Cu (II), which indicates that the Cu (II) is more dissociated from the binding to the two triazole nitrogen atoms at elevated temperatures. The anomeric proton appeared broad singlet at 30 °C, at 4.79 ppm, and shifted to 4.81 to The 1 H NMR spectra of the copper complex with DM35 were also evaluated at different temperatures, from 30 to 60 • C (Figure 7, Figures S24 and S25), the triazole signals showed upfield shift to a broad signal, but more defined peak at around 7.71 ppm, this is close to the chemical shifts of macrocycle without Cu (II), which indicates that the Cu (II) is more dissociated from the binding to the two triazole nitrogen atoms at elevated temperatures. The anomeric proton appeared broad singlet at 30 • C, at 4.79 ppm, and shifted to 4.81 to a doublet ( Figure S24a). The amide NH signal shifted upfield at higher temperature indicates that amide participates in intermolecular hydrogen bonding. The process is reversible as when the same NMR sample was cooled from 60 to 30 • C ( Figure S25), the signals for triazoles and amide reverted to the original pattern. a doublet ( Figure S24a). The amide NH signal shifted upfield at higher temperature indicates that amide participates in intermolecular hydrogen bonding. The process is reversible as when the same NMR sample was cooled from 60 to 30 °C (Figure S25), the signals for triazoles and amide reverted to the original pattern. The 1 H NMR spectra of the complex of DM35 with Cu(OAc)2 are shown in Figures  S26-S28. These are similar to the copper sulfate complex, but showed more significant broadening; this indicated that the macrocycle formed a strong complex with the Cu (II) through interactions with the nitrogen of the triazole ring, and the copper ion is in close proximity with several other hydrogens. The complex of the monotriazole LM28 with CuSO4•5H2O ( Figure S29) showed a similar trend; the triazole signal appeared as a broad singlet, but still visible, unlike in the bistriazole complexes. The 1 H NMR spectra of the copper complex at different temperatures showed some chemical shift changes for the NH and triazole signals ( Figure S31). These results showed that for monotriazoles, such as LM28, the copper ion was not bonded to the triazole as tightly as comparing to bistriazoles. The complex of the C2 bistriazole derivative 24 with CuSO4•5H2O was also prepared and the 1 H NMR spectra showed similar trends (Figures S32-S34).

Effect of the Macrocycles as Ligands for CuAAC Reactions
Besides the binding studies, the effect of these macrocycles on CuAAC reactions was also studied as shown in Scheme 5. Using the anomeric sugar azide 26 as the substrate, a series of reactions were carried out with different alkynes, including the aliphatic alkyne 1-octyne, aromatic alkynes, such as phenylacetylene, 4-t-Bu-phenylacetylene, and 5-phenyl-1-pentyne. The results are summarized in Tables 2-5 and SI Tables S1-S8, Figures S35-S42. For phenylacetylene, when using 2.5 mol% macrocycle and copper sulfate as the catalyst (Table 2a), at 2 h, all C1-series macrocycles (MCs), including both monotriazole and bis-triazole derivatives, were able to accelerate the reaction to over 50% conversion versus 15% conversion without the macrocycles. The most effective macrocycles are the bistriazole macrocycles DM25 and DM35, which reached 100% conversion within 5 h. The reactions in the presence of the monotriazole derivatives LM34, LM36, LM26 reached over 75% conversion at 5 h and almost completed after 9 h. When the monotriazoles (Table 2b) were increased to 5.0 mol%, LM26 and LM34 were both effective in catalyzing the reac- The 1 H NMR spectra of the complex of DM35 with Cu(OAc) 2 are shown in Figures S26-S28. These are similar to the copper sulfate complex, but showed more significant broadening; this indicated that the macrocycle formed a strong complex with the Cu (II) through interactions with the nitrogen of the triazole ring, and the copper ion is in close proximity with several other hydrogens. The complex of the monotriazole LM28 with CuSO 4 ·5H 2 O ( Figure S29) showed a similar trend; the triazole signal appeared as a broad singlet, but still visible, unlike in the bistriazole complexes. The 1 H NMR spectra of the copper complex at different temperatures showed some chemical shift changes for the NH and triazole signals ( Figure S31). These results showed that for monotriazoles, such as LM28, the copper ion was not bonded to the triazole as tightly as comparing to bistriazoles. The complex of the C2 bistriazole derivative 24 with CuSO 4 ·5H 2 O was also prepared and the 1 H NMR spectra showed similar trends ( Figures S32-S34).

Effect of the Macrocycles as Ligands for CuAAC Reactions
Besides the binding studies, the effect of these macrocycles on CuAAC reactions was also studied as shown in Scheme 5. Using the anomeric sugar azide 26 as the substrate, a series of reactions were carried out with different alkynes, including the aliphatic alkyne 1-octyne, aromatic alkynes, such as phenylacetylene, 4-t-Bu-phenylacetylene, and 5-phenyl-1-pentyne. The results are summarized in Tables 2-5 and SI Tables S1-S8, Figures S35-S42. For phenylacetylene, when using 2.5 mol% macrocycle and copper sulfate as the catalyst (Table 2a), at 2 h, all C1-series macrocycles (MCs), including both monotriazole and bis-triazole derivatives, were able to accelerate the reaction to over 50% conversion versus 15% conversion without the macrocycles. The most effective macrocycles are the bistriazole macrocycles DM25 and DM35, which reached 100% conversion within 5 h. The reactions in the presence of the monotriazole derivatives LM34, LM36, LM26 reached over 75% conversion at 5 h and almost completed after 9 h. When the monotriazoles (Table 2b) were increased to 5.0 mol%, LM26 and LM34 were both effective in catalyzing the reactions, reaching almost full conversions at 5 h. The C2 series macrocycles were not as effective as the C-1 series; only the bistriazole derivative C2DML (24) showed some acceleration for the reaction. tions, reaching almost full conversions at 5 h. The C2 series macrocycles were not as effective as the C-1 series; only the bistriazole derivative C2DML (24) showed some acceleration for the reaction.   When using 1-octyne as the substrate, different macrocycles showed very different results. As shown in Table S3, with 2.5 mol% macrocycles, 1.2 equiv of 1-octyne and 0.1 equiv of CuSO4, only DM34 was effective at catalyzing the reaction, reaching 93% conversion at 2 h; the other macrocycles didn't show improvement over the control experiment. When increasing the 1-octyne from 1.2 to 1.5 equiv, the reaction completed within 1 h in the presence of 2.5 mol% of DM34. Increasing the loading of MCs for LM26, LM34, and LM36 to 5.0 mol% did not improve the conversions (Tables 3 and S4). The experiments confirmed that DM34 was particularly efficient at catalyzing the click reaction of 1-octyne.   When using 1-octyne as the substrate, different macrocycles showed very different results. As shown in Table S3, with 2.5 mol% macrocycles, 1.2 equiv of 1-octyne and 0.1 equiv of CuSO 4 , only DM34 was effective at catalyzing the reaction, reaching 93% conversion at 2 h; the other macrocycles didn't show improvement over the control experiment. When increasing the 1-octyne from 1.2 to 1.5 equiv, the reaction completed within 1 h in the presence of 2.5 mol% of DM34. Increasing the loading of MCs for LM26, LM34, and LM36 to 5.0 mol% did not improve the conversions (Table 3 and Table S4). The experiments confirmed that DM34 was particularly efficient at catalyzing the click reaction of 1-octyne. 5-Phenyl-1-pentyne was found to be much more reactive in comparison to the other alkynes screened. Several experiments were carried out to analyze the effect of the macrocycles with reduced copper loading. A few summarized results are shown in Table 4 and Tables S5 and S6. Without macrocycles, at 2 h the reaction of the alkyne with 26 reached 100% (0.1 equiv of CuSO 4 ) and 51% conversion when using 0.05 equiv copper sulfate. Further reduction of copper to 1 mol%, at 2 h only 5% conversion was observed; however, with added 2.5 mol% of DM35, the reaction reached 52% conversion. From these results, we selected either 2.0 or 5.0 mol% of copper to evaluate the MC ligands. As shown in Table 4, 1.0-2.5% mol of the bis-triazole macrocycles were effective at accelerating the reactions significantly, reaching 100% conversion at about 1-5 h. When using 2 mol% Cu(OAc) 2 as the catalyst, the control reaction reached 13% conversion, but the reaction with DM35 reached 100% conversion at 2 h (Table S5). For 4-t-butylphenylacetylene (Table 5 and Tables S7 and S8), using 0.1 equiv of copper sulfate, the reactions were slow; only DM25 and DM35 helped the reaction to full conversion at 20 h; when the CuSO 4 loading was increased, the reactions in the presence of the two MCs reached full conversions at 5 h. DM34 and DM24 were not that effective, giving similar conversion compared to the control experiment. DM25 was much more efficient than the isomer DM34-apparently, the positions of the triazoles in the macrocycle had an influence towards the catalysis of the reaction. The energy minimized conformations of two macrocycles DM25 and DM34 are shown in Figure 8 [35]. These two compounds have the same molecular weight and ring sizes-the only difference is that the triazole is located at a different position to the anomeric center. The two MCs adopted quite different conformations in which the triazoles rings are more parallel in DM25, but at a dihedral angle about 30 degrees to each other for DM34. The two triazole rings can adopt different conformations upon binding with copper, and possibly have a cooperative effect when both triazole rings are embedded in the macrocycles. The conformation difference of these MCs perhaps is correlating with the selectivity among different alkynes. DM25 is more selective for phenylacetylene; it is also more effective for t-Bu-phenylacetylene, while DM34 was the most effective MC for 1-octyne, but not as good as other MCs for phenylacetylenes. The interesting selectivity towards different alkynes may be useful in differentiating acetylenes when reacting with the azide; this could be useful for other selective reactions.  Condition a : Sugar azide (1.0 equiv), CuSO4•5H2O (0.1 equiv), p-t-butyl-phenylacetylene (1.5 equiv), EtOH/H2O (v/v 1:1, 2.0 mL), NaAsc (0.3 equiv); Condition b : The same as "a" except 0.2 equiv of CuSO4 ·H2O and 0.4 equiv of NaAsc were used.
The energy minimized conformations of two macrocycles DM25 and DM34 are shown in Figure 8 [35]. These two compounds have the same molecular weight and ring sizes-the only difference is that the triazole is located at a different position to the anomeric center. The two MCs adopted quite different conformations in which the triazoles rings are more parallel in DM25, but at a dihedral angle about 30 degrees to each other for DM34. The two triazole rings can adopt different conformations upon binding with copper, and possibly have a cooperative effect when both triazole rings are embedded in the macrocycles. The conformation difference of these MCs perhaps is correlating with the selectivity among different alkynes. DM25 is more selective for phenylacetylene; it is also more effective for t-Bu-phenylacetylene, while DM34 was the most effective MC for 1-octyne, but not as good as other MCs for phenylacetylenes. The interesting selectivity towards different alkynes may be useful in differentiating acetylenes when reacting with the azide; this could be useful for other selective reactions.

General Methods
All reactions were carried out under normal conditions, reagents and solvents were obtained commercially from Sigma-Aldrich, VWR, and Fisher and used directly without purifications. All reactions, unless otherwise noted, were carried out in oven-dried glassware under a nitrogen atmosphere. All purifications were conducted by flash chromatography using 230-400 mesh silica gel with a gradient of solvent systems. Thin-layer chromatography (TLC) analysis was performed with aluminum-backed TLC plates with UV and fluorescence indicator and visualized using UV lamp at 254 nm, then stained with

Experimental Section General Methods
All reactions were carried out under normal conditions, reagents and solvents were obtained commercially from Sigma-Aldrich, VWR, and Fisher and used directly without purifications. All reactions, unless otherwise noted, were carried out in oven-dried glassware under a nitrogen atmosphere. All purifications were conducted by flash chromatography using 230-400 mesh silica gel with a gradient of solvent systems. Thin-layer chromatography (TLC) analysis was performed with aluminum-backed TLC plates with UV and fluorescence indicator and visualized using UV lamp at 254 nm, then stained with PMA solution. 1 H NMR and proton-decoupled 13 C NMR spectra were obtained with Bruker 400 MHz NMR spectrometer in DMSO-d 6 , D 2 O, or CDCl 3 . The chemical shifts were reported using CDCl 3 /DMSO-d 6 as internal standard at 7.26/2.50 ppm and at 77.0/39.5 ppm, respectively. 2D NMR experiments (HSQC, COSY) were also conducted to assist the compound characterizations. Melting point measurements were carried out using a Fisher Jones melting point apparatus. The molecular mass was measured using LC-MS on an Agilent LC1260 system and 6120B Single Quad Mass Spectrometer or with Shimadzu LCMS-2020. HRMS data were obtained using positive electrospray ionization on a Bruker 12T APEX-Qe FTICR-MS with an Apollo II ion source.
For compounds synthesized by similar methods, the procedures for the first compound are included in detail. For the rest, only the amount used, purification method, and the characterization data are provided.
Synthesis of compound 6a. Compound S3 (200.0 mg, 0.69 mmol, 1.0 equiv) was added to a 50 mL round bottomed flask (RBF) with a drying tube and nitrogen balloon, pyridine (4.0 mL) was added and the flask was cooled to 0 • C, then 4-bromobenzenesulfonyl chloride (352.1 mg, 1.37 mmol, 2.0 equiv) was added and the mixture was stirred at 0 • C for 20 min and the ice bath was removed. The mixture was stirred at rt for 20 h, at which time 1 H NMR spectrum showed about 95% conversion. The reaction was stopped, and solvent was removed, the crude product was purified on silica gel using a gradient of dichloromethane (DCM) and methanol, from DCM up to 10% MeOH/DCM (R f = 0.31 in 10% MeOH/DCM). The desired product 6a was obtained as a colorless liquid (232.0 mg, 66%). 1  Synthesis of compound 7a. Compound 6a (100.0 mg, 0.2 mmol, 1.0 equiv) was added to a 50 mL round bottom flask with a drying tube and nitrogen balloon attached, the reaction flask was cooled to 0 • C, then dichloromethane (2.5 mL), pyridine (5.0 equiv) and benzoyl chloride (0.06 mL, 0.5 mmol, 2.5 equiv) were added to the solution. The reaction mixture was stirred at to 0 • C and then stirred at rt for about 12 h. At this time, TLC and 1 H NMR indicated full conversion to the product. The reaction was stopped, and solvent was removed under vacuum using a rotovap. The crude product was purified using flash chromatography on silica gel using a solvent gradient of ethyl acetate and hexane from 1:4 to 3:2 ratio. The desired product was obtained as a colorless viscous liquid (112.0 mg, 79% yield). (R f = 0.38 in 5% MeOH/DCM). 1 13    Synthesis of compound LM28 (13c). Compound 8c (100.0 mg, 0.12 mmol, 1.0 equiv), K 2 CO 3 (33.0 mg, 0.24 mmol, 2.0 equiv) and DMF (14 mL) were added to a 50 mL RBF. The reaction mixture was stirred at 70 • C for 3 h, at which time 1 H NMR spectrum and TLC indicated the full conversion of starting materials. The reaction was stopped and solvent was removed. The crude was purified via flash chromatography using an eluent of DCM to 5% MeOH/DCM to obtain the desired product as a white solid (65.0 mg, 82%), R f = 0.  Synthesis of compounds LM28 and DLM28. Compound 8a (100.0 mg, 0.11 mmol, 1.0 equiv), DMF (7.0 mL), and K 2 CO 3 (30.7 mg, 0.22 mmol, 2.0 equiv) were added to a 50 mL RBF. The reaction mixture was stirred at 75 • C for 5 h, the 1 H NMR and TLC samples showed complete conversion of the starting materials. The crude was purified by flash chromatography using DCM to 5% MeOH DCM to obtain the desired compound LM28 (53.0 mg, 0.080 mmol, 72%) along with some later fraction which was identified as the dimerization product DLM28 as a white solid (15.0 mg, 0.011 mmol, 20% based on starting material conversion). The chloro compound 8b was cyclized by similar conditions, using compound 8b (50.0 mg, 0.07 mmol, 1.0 equiv), DMF (7.0 mL), and K 2 CO 3 (19.8 mg, 0.14 mmol, 2.0 equiv). The desired compound LM28 (31.7 mg, 0.048 mmol, 67%) and  13  Synthesis of compound 10d. 3-azido propionic acid (200.0 mg, 1.74 mmol, 1.0 equiv) and 1,7-octadiyne (276.0 mg, 2.6 mmol, 1.5 equiv) were dissolved in t-BuOH: THF: H 2 O (v:v:v 1:1:1, 25.0 mL), then CuSO 4 ·5H 2 O (84.8 mg, 0.34 mmol, 0.2 equiv), NaAsc (134.7 mg, 0.68 mmol, 0.4 equiv) were added to the reaction mixture. The reaction was stirred at rt for 24 h, at which time the reaction was completed as indicated by 1 H NMR and TLC. The reaction was stopped, and solvent was removed using a rotovap, the residue was diluted with EtOAc and acidified using 0.1 N HCl (5.0 mL) followed by water wash. The organic layer was collected and dried over anhydrous Na 2 SO 4 and solvent was removed under vacuum to obtain the crude, which was further purified with flash chromatography using eluent of hexanes to 60% EtOAc/Hexanes to obtain the desired product as a yellowish solid Eleven macrolactones cyclized from C1 to C6, and three macrocycles cyclized from C2 to C6 were synthesized and characterized. The macrocycles showed interesting anion binding properties with tetrabutylammonium halides. TBACl showed the strongest complexation with the macrocycles. Moreover, several bistriazole-containing macrolactones also formed complexes with Cu (II) ions. As indicated by 1 H NMR spectra, the amide proton and triazole signals exhibited significant chemical shift changes, which indicated that the Cu (II) or anions are binding to the macrolactones in a specific manner and may be useful for molecular recognition. The applications of these macrocycles as ligands for copper ions for the cycloaddition reactions of azide and alkynes (AAC) were explored. Interestingly the bis-triazole-containing macrocycles were highly efficient ligands for accelerating the AAC reactions significantly, and the monotriazole-based macrolactones were not as effective. The rate accelerating effects of macrocycles were also selective towards different acetylenes. The methods of synthesizing these novel sugar-based macrocycles should apply to other carbohydrate derivatives.