Mechanistic Basis of the Cu(OAc)2 Catalyzed Azide-Ynamine (3 + 2) Cycloaddition Reaction

The Cu-catalyzed azide–alkyne cycloaddition (CuAAC) reaction is used as a ligation tool throughout chemical and biological sciences. Despite the pervasiveness of CuAAC, there is a need to develop more efficient methods to form 1,4-triazole ligated products with low loadings of Cu. In this paper, we disclose a mechanistic model for the ynamine-azide (3 + 2) cycloadditions catalyzed by copper(II) acetate. Using multinuclear nuclear magnetic resonance spectroscopy, electron paramagnetic resonance spectroscopy, and high-performance liquid chromatography analyses, a dual catalytic cycle is identified. First, the formation of a diyne species via Glaser–Hay coupling of a terminal ynamine forms a Cu(I) species competent to catalyze an ynamine-azide (3 + 2) cycloaddition. Second, the benzimidazole unit of the ynamine structure has multiple roles: assisting C–H activation, Cu coordination, and the formation of a postreaction resting state Cu complex after completion of the (3 + 2) cycloaddition. Finally, reactivation of the Cu resting state complex is shown by the addition of isotopically labeled ynamine and azide substrates to form a labeled 1,4-triazole product. This work provides a mechanistic basis for the use of mixed valency binuclear catalytic Cu species in conjunction with Cu-coordinating alkynes to afford superior reactivity in CuAAC reactions. Additionally, these data show how the CuAAC reaction kinetics can be modulated by changes to the alkyne substrate, which then has a predictable effect on the reaction mechanism.


INTRODUCTION
The Cu-catalyzed azide−alkyne (3 + 2) cycloaddition (CuAAC) is the most prominent class of "click" reactions. 1,2y virtue of fast reaction rates, 3 the synthetic tractability of installing alkyne (1) and azide (2) groups in small molecules and biomolecules, 4 and the regioselective formation of a 1,4triazole product (3), the CuAAC is used across all facets of medicinal chemistry, 5 chemical biology, 6−8 and in the material sciences (Figure 1a). 9,10he CuAAC reaction typically proceeds with a Cu(II) source, which is subsequently reduced by the addition of a reductant (e.g., sodium ascorbate). 2 Alternatively, the use of Cu(I) salts and stabilizing ligands obviates the addition of a reductant, thereby simplifying the operation of this reaction. 1 major limitation in the wider application of the CuAAC reaction in chemical biology workflows and the preparation of bioconjugates is the need for superstoichiometric quantities of Cu(I) for the reaction to proceed in aqueous buffers because the Cu typically coordinates to a variety of Lewis basic sites present in the solvent and in biomolecules.11,12 This invariably results in the onset of oxidative damage to proteins and nucleic acids due to the formation of Cu(I)mediated reactive oxygen species 13−15 and reduces the yield of bioconjugate products.16 Therefore, despite the broad uptake and development of the CuAAC platform, there is a distinct need to develop alternative CuAAC reactions, which produce a single ligated product, with low Cu loadings and without significantly compromising the reaction kinetics of the process.17 Previous mechanistic analyses of the catalytic cycle of the CuAAC reaction using conventional alkynes and a Cu(I) source have identified the formation of a Cu-acetylide species being rate-determining.18−23 A proposed mechanistic feature of the CuAAC is the formation of a binuclear Cu species, which is held together within a ligand architecture.18,19,22,24,25 These binuclear Cu complexes can potentially exist in a mixed valence state [i.e., Cu(I) and Cu(II)], possibly working in concert to form the Cu-acetylide species, facilitate azide ligation, and finally to initiate the (3 + 2) cycloaddition.19,25−27 While the in situ reduction of a Cu(II) precatalyst by sodium ascorbate (NaAsc) is the most widely used strategy to effect CuAAC reactions, an alternative approach is to access catalytically competent Cu(I) via a Cu(II)-catalyzed Glaser−Hay coupling of an alkyne (Figure 1b).In this case, a diyne product (e.g., 4) is formed along with a Cu(I) species, which catalyzes the (3 + 2) cycloaddition reaction in the presence of an azide (e.g., 2).28−31 The implication of diyne formation as a means to access a Cu(I) species suitable to catalyze the CuAAC reaction is that this requires coordination of Cu(II) with an alkyne substrate.12,25,27,32 At present, the molecular characteristics which influence alkyne reactivity as the sacrificial reductant have not been exploited in the context of CuAAC-mediated reaction chemistry.33 One approach to enhance the reactivity of an alkyne toward Cu is to conjugate a Lewis basic site with the triple bond.A pertinent example is the use of aromatic ynamines (e.g., 5) in CuAAC reactions.34−37 These alkyne surrogates exhibit superior reactivity relative to alkyl and aromatic alkynes with azides to form 1,4-triazoles (6) using catalytic Cu(OAc) 2 (Figure 1c).36,37 In earlier work, we highlighted a mechanistic divergence of the azide-ynamine (3 + 2) cycloaddition reaction in which the rate-determining step (RDS) shifts from Cu-acetylide formation to azide ligation of the Cu catalyst.35 However, these studies did not identify the role of Cu(OAc) 2 ; whether the Cu(II) was a precatalyst that could form Cu(I) in situ via a Glaser coupling to form a diyne (e.g., 4) or had other mechanistic significance. 25,33n this work, we present a mechanistic framework that demonstrates how the Cu catalyst influences ynamine reactivity in CuAAC reactions.A combination of heteronuclear nuclear magnetic resonance (NMR) and electron paramagnetic resonance (EPR) spectroscopy is used to elucidate an array of catalytic speciation derived from Cu(OAc) 2 , in which the oxidation states and interactions with the substrates, product, and solvent evolve over time.
These data show that the Cu catalyst in the (3 + 2) cycloaddition reaction displays dynamic behavior 23 through the stages of initiation and reaction promotion.Finally, we show that the triazole product (e.g., 6) forms a complex with Cu, resulting in a catalytic resting state that can be reactivated by further addition of ynamine and azide substrates.

RESULTS AND DISCUSSION
The objective of this work was to determine how the nature of the Cu catalyst influences the ynamine reactivity relative to those of cognate alkyne substrates.Our initial hypothesis was Cu(OAc) 2 played a dual catalytic role: (i) the paddlewheel architecture assists in the formation of a diyne species (e.g., 4) via a Glaser−Hay coupling of an ynamine (e.g., 5); and (ii) the resultant Cu(I) species catalyzes the formation of a triazole product in the presence of a corresponding azide (e.g., 2, Figure 1c).
2.1.Divergent Ynamine Reactivity Is Promoted by a Combination of Base and Cu Catalyst.Hydrogen− deuterium exchange (HDE) of a terminal alkyne (7) was investigated in either CD 3 CN or a 9:1 mixture of CD 3 CN: D 2 O as a function of a 5 mol % additive (Figure 2a).All of the additives were soluble under experimental conditions, apart from CuOAc that was used as a suspension.Only 6% HDE was observed in a 9:1 CD 3 CN:D 2 O mixture when ynamine 5 was used (Figure 2c).The fastest HDE was observed using 5 mol % Cu(OAc) 2 and either 5 mol % AcOH or NaOAc, forming 93% 5-D within 30 min (Figure 2c).Interestingly, HDE was slower when CuOAc was used (70% after 2 h), with a further reduction in the formation of 5-D observed in the presence of either CuSO 4 (37%) or CuPF 6 (13%).However, the rate of HDE increased when CuPF 6 + NaOAc (5 mol %) was used (47%), suggestive of a cooperative effect between a Cu source and acetate.When only 5 mol % NaOAc was used, the rate of HDE of 5 to 5-D (81% after 2 h) was almost as fast as when Cu(OAc) 2 was added.Finally, no enhancement in the rate of HDE was observed using 8 or 9 in the presence of any of these additives (Figures S13−S15), with the reaction mixture ≥pH 6 when 5 mol % NaOAc was added, regardless of the counterion (Figure S121).
The time course of the (3 + 2) cycloaddition to form 1,4triazole 10/10-D as a function of the Cu catalyst (5 mol %) and salt additive was then monitored by 1 H NMR spectroscopy in CD 3 CN:D 2 O 9:1 (Figure 2d).We identified that the use of deuterated solvents resulted in the formation of 10/10-D in a ratio of 1:9 using 5 mol % Cu(OAc) 2 (Section S3.7) Without the addition of a catalyst, no formation of 10/10-D was detected over 2 h.Cu(OAc) 2 was identified as the superior catalyst, with the formation of 10/10-D being complete (96%) after 26 min.
In comparison, the use of CuOAc resulted in slower formation of 10/10-D (complete ∼2 h) and neither CuSO 4 or CuPF 6 induced any detectable CuAAC (Figure 2e).However, the addition of NaOAc (5 mol %) to CuPF 6 (5 mol %) did result in the formation of 10/10-D, albeit at a rate that was slower than that effected by direct use of Cu(OAc) 2 .This could suggest a potential inhibitory effect of the PF 6 counterion.No conversion to 10/10-D was observed using either a mixture of 5 mol % of CuSO 4 and NaOAc (Figure 2e), or NaOAc.However, the addition of 5 mol % of AcOH or NaOAc to 5 mol % Cu(OAc) 2 or CuOAc enabled

Journal of the American Chemical Society
the CuAAC reaction to proceed (Figure 2e).Since CuOAc can form binuclear species and higher order aggregates, 38 this suggests that a preformed Cu(OAc) 2 /CuOAc species is optimal to catalyze (3 + 2) cycloadditions. 39lkynes 8 and 9 did not undergo the (3 + 2) cycloaddition with 2 using any of the Cu catalysts used with 5 as the alkyne surrogate when CD 3 CN:D 2 O (9:1) was used.Furthermore, triazole products 11/11-D and 12/12-D only formed when the reaction was conducted in pure CD 3 CN (Figures S19− S21).These results emphasize the divergent reactivity of 5 when the Cu-catalyzed (3 + 2) cycloaddition is performed in aqueous conditions.
Taken together, these results show that the HDE of 5 is predominantly promoted by the acetate ligand.Furthermore, the presence of the benzimidazole N3 in 5 is essential for HDE and the initiation of the Cu-catalyzed (3 + 2) cycloaddition reaction with 2.  25,40 and that the Cu(I) species formed in situ could then catalyze the (3 + 2) cycloaddition reaction with a Cu-coordinating azide.In the case of ynamine 5, benzimidazole N3 coordination by Cu(OAc) 2 enhances the reactivity.Formation of complex (13) on mixing TIPS-protected ynamine (14) with Cu-(OAc) 2 confirmed coordination at the N3 position at the apical site of the Cu(OAc) 2 paddlewheel structure (Figures 3a and S120).However, EPR measurements under a catalytic regime (10 mol % of Cu(OAc) 2 ) showed that 14 can desymmetrize the paddlewheel complex in solution, resulting in a strong signal indicative of a Cu(II) species (Section S4.1).This suggests that the N3 position of 14 can denucleate the Cu(OAc) 2 paddlewheel (present in 10 mol %), to form a mononuclear complex in MeCN, to deliver the catalytically active species. 41

Journal of the American Chemical Society
To further test the influence of N3 coordination in the (3 + 2) cycloaddition, a range of isotopically labeled ynamine (5) and benzyl azide (2) substrates and their corresponding triazole products were prepared, and then used to monitor the reaction by NMR spectroscopy (Figure 3b).
We first explored the influence of the N3 position in a TIPS protected ynamine with Cu(OAc) 2 and [(MeCN) 4 Cu]-PF 6 using 14, 14-15 N 2 and 14-13 C 2 -15 N 2 .The addition of 5 mol % Cu(OAc) 2 resulted in a broadening of 1 H signals, with the extent of broadening increasing in the order, H Me , H c , H d , H a (Figure 4a).This order reflects the increasing spatial and electronic proximity to the site of Cu(II) coordination, i.e., at N3. 42,43 The impact of the addition of 5 mol % [(MeCN) 4 Cu]PF 6 on the 1 H NMR spectrum of 14 was far less pronounced than with 5 mol % Cu(OAc) 2 , with only significant signal broadening detected at H a , Figure 4b.
1 H− 15 N HMBC experiments were then acquired with 5 mol % of both Cu catalysts using 14-15 N 2 (Figure 4c,d).In neither case were the N1 or N3 resonances detected.Finally, negligible changes in the chemical shifts of C α and C β in 14-13 C 2 -15 N 2 were detected after the addition of 5 mol % Cu(OAc) 2 (Figure 4e); this was also the case for [(MeCN) 4 Cu]PF 6 (Figure S29).While the addition of one equivalent Cu(OAc) 2 resulted in precipitation, the addition of one equivalent of [(MeCN) 4 Cu]PF 6 resulted in an upfield shift of C α and a concomitant downfield shift of the C β resonance, suggesting a weak, long-range interaction between 14 and a Cu(I) source, likely via N3 coordination (Figure 4f).
Having established that Cu coordinates at the N3 position of 14, the potential for a Cu catalyst to form diyne 15 via a Glaser−Hay coupling with 5 was explored (Figure 5a).The time course for the reaction of 5 in the presence of 5 mol % Cu(OAc) 2 revealed immediate broadening of the 1 H NMR signals of 5, and an associated downfield shift of H a (Figure 5b).Sharpening of the H a resonance accompanied by an upfield shift then occurred over 14 min.Accompanying the sharpening of this resonance was the presence of a new set of peaks associated with the formation of 15.This effect was not observed when 5 mol % [(MeCN) 4 Cu]PF 6 was used (Figure 5c). 1 H− 15 N HMBC spectra were acquired to monitor the Glaser−Hay coupling using 5-15 N 2 in the presence of 5 mol % Cu(OAc) 2 .Spectra showed the characteristic disappearance of the 15 N resonance for N3 after 5 min.In contrast to the resulting 1 H− 15 N HMBC spectrum of 14-15 N 2 on the addition of 5 mol % Cu(OAc) 2 (Figure 4c), only the N3 resonance disappeared when 5 mol % Cu(OAc) 2 was added to 5-15 N 2 (Figure 5d).After 19 h, both 15 N signals were detected alongside a set of new signals associated with diyne formation (15-15 N 2 , Figure 5d).In stark contrast, the addition of 5 mol % [(MeCN) 4 Cu]PF 6 to 5-15 N 2 , only changed the chemical shift of N3, detected at 233 K (Figure 5e).This data show that although both Cu catalysts coordinate to the N3 position of 5, only the Cu(II) species induces Glaser−Hay coupling of 5 to form 15 prior to entering the CuAAC catalytic cycle.In contrast to the intense Cu(II) signals detected by EPR spectroscopy when 14 was mixed with Cu(OAc) 2 , analysis of a mixture of 5 with 10 mol % Cu(OAc) 2 showed a weak signal, which decreased over time (Section S4.2).This is indicative of Cu(II) being reduced to a Cu(I) species.
Previous studies on the Glaser−Hay coupling of diynes have revealed the type of N-ligand used influences the amount of diyne formed, 28,44 with the benzimidazole N3 position known to be a Lewis basic site for both Cu(II) and Cu(I). 45,46We then explored if Cu loading influenced diyne (15) formation using 5 and catalytic Cu(OAc) 2 by monitoring the reaction by high-performance liquid chromatography (HPLC; 0.5−50 mol %, Figure 6a).
There was a distinct correlation between Cu(OAc) 2 loading and the formation of 15 when the catalyst loading range is 0.5−5 mol %, i.e., formation of 0.25% 15 using 0.5 mol % Cu(OAc) 2 , i.e., resulting in complete reduction of Cu(II) to a Cu(I) species (Section S6.4).Deviation from this correlation occurred when catalyst loading was in the 10−50 mol % range.These results demonstrate that Cu(OAc) 2 forms 15, thus producing a Cu(I) species, as supported by the reappearance of 15 N resonances and the concomitant sharpening of 1 H resonances after ∼15 min.
Using the HPLC assay, the influence of the water content on the rate of formation of 15 was explored (Figure 6c).This was an important factor to understand as water could compete with the benzimidazole N3 at the apical coordination sites of Cu(OAc) 2 , 28 thus slowing down the Glaser−Hay coupling, and possibly leading to the slow accumulation of Cu(I) to catalyze the CuAAC reaction. 47In MeCN containing up to 5% water, approximately 2.5% conversion of 15 proceeded in 15 min when using 5 mol % Cu(OAc) 2 .However, when water was present at higher concentrations, e.g., 10−20% H 2 O, the rate of Glaser−Hay coupling was substantially reduced (Figure 6c).techniques.First, we explored the formation of 10-15 N 4 using 5-15 N 2 and 2-15 N 2 (Figure 7b−e).Under these conditions, the formation of triazole 10-15 N 4 was complete within 6 h.However, a series of spectral changes was observed during the course of this reaction to form 10-15 N 4 (Figure 7b,c).These included a broadening and downfield shift of both the H a/b 1 H NMR signal in 10-15 N 4 and the H 2 O resonance.Between 6 and 10 h, the water resonance returned to its original chemical shift position (Figure 7c).We surmise that the apical site on the Cu(OAc) 2 paddlewheel structure is dynamic, with water competing with the N3 of 10-15 N 4 toward the end of the reaction as 5-15 N 4 is consumed.Further analysis of this phenomenon by 1 H− 15 N HMBC also showed spectral changes in the intensity of the N 1 −H a and N α -H e cross peaks over time (Figure 7d,e).For example, the formation of 10-15 N 4 was observed after 30 min, with the intensity of the N α -H e cross peak remaining constant after 7 h.
A second set of experiments were conducted to provide further insight into changes in the alkyne by the reaction of 5-13 C 2 -15 N 2 with 2-15 N 2 to form 10-13 C 2 -15 N 4 (Figure 7f,g).A distinct downfield chemical shift of the C β resonance was observed before returning to its original position after 12 h.This suggests a structural change occurring with the Cu catalyst after 6 h involving coordination at the N3 position of the triazole product (10-13 C 2 -15 N 4 ).For example, the formation of 10-15 N 4 was observed after 30 min, with the intensity of the N α -H e cross peak remaining constant after 7  13 C NMR spectra of 14-13 C 2 -15 N 2 (gray) and after the addition of 5 mol % Cu(OAc) 2 (black).(f) 13 C NMR spectra of 14-13 C 2 -15 N 2 (gray) and after the addition of 1.0 equiv of [(MeCN) 4 Cu]PF 6 to 14-13 C 2 -15 N 2 (black).All spectra using isotopically labeled compounds were acquired in CD 3 CN at an ynamine concentration of 15.5 mM (c−f), and 62 mM for natural abundant 5 (e.g., b, c).h.In contrast, the corresponding N1/N3−H a cross peak rapidly started to lose intensity after 4.7 h and completely disappeared after 12 h (Figure 7e).
EPR measurements under these reaction conditions show no signal for Cu(II) over the reaction time frame until 6 h.
After 6 h a Cu(II) signal emerges, which suggests either reoxidation 48,49 of a Cu(I) species to Cu(II) or a reorganization of an EPR active Cu species (Section S4.4).

Cu-Catalyst is in a Postreaction Resting State after Completion of the Ynamine-Azide (3 + 2)
Cycloaddition.The observation that triazole 10 forms a Cu complex by coordination at the apical site of Cu(OAc) 2 led us to determine whether this species is catalytically competent after completion of the (3 + 2) cycloaddition.This was interrogated by first running the (3 + 2) cycloaddition with unlabeled substrates, followed by the addition of a second equivalent of isotopically labeled ynamine (5-13 C 2 -15 N 2 ) and azide (2-15 N 2 ) after 15 h (Figure 8a).The time course of the (3 + 2) cycloaddition showed that the presence of 5-13 C 2 -15 N 2 resulted in a direct displacement of 10 (Figures 8b and S76).In addition, the second (3 + 2) cycloaddition resulted in a faster rate of formation of 10-13 C 2 -15 N 4 relative to the first, i.e., the formation of 10 (Figure 8c), which is likely due to a Cu(I) species present in a resting state with the apical face occupied by N3 coordination to 10.Using 1 H− 15 N HMBC, the spectral characteristics of the second (3 + 2) cycloaddition reaction were similar to the first.For example, the disappearance of the N 1 −H b cross peak of 5-13 C 2 -15 N 2 correlated with the appearance of the N α -H e cross peak corresponding to the formation of 10-13 C 2 -15 N 4 (Figure 8d).The dynamic behavior of the triazole product (10-13 C 2 -15 N 4 ) was observed by the initial appearance of the N 1 −H a which then gradually disappeared between 2 and 6 h after the addition of the isotopically labeled substrates.
A key spectroscopic difference between the first and second (3 + 2) cycloaddition was the H 2 O resonance (Figure 8b).A larger downfield shift of the H 2 O resonance (Δδ − 0.031) was detected for the first (3 + 2) cycloaddition, compared to Δδ − 0.015 for the second (Figure S79).Furthermore, the intensity of the H 2 O signal was significantly lower for the second (3 + 2) cycloaddition than the first.This could be due to H 2 O acting as a proton shuttle with acetate and the N3 benzimidazole position as 10 is formed (pK a H of Nmethylbenzimidazole ∼4.6). 50We surmise that the addition of 5-13 C 2 -15 N 2 and 2-15 N 2 in the second (3 + 2) cycloaddition results in a further bias in the equilibrium, reducing the magnitude of the chemical shift change of the H 2 O resonance.
We then explored the catalytic potential of the Cu species present at the end of the first (3 + 2) cycloaddition to undergo a second Glaser−Hay reaction (i.e., formation of 15-13 C 4 -15 N 4 ) after the addition of 2-15 N 2 and 5-13 C 2 -15 N 2 .Unexpectedly, a 1:0.5 ratio of 15 to 15-13 C 4 -15 N 4 (Figure 8e, further confirmed by HPLC analysis, Table S27) was present in the reaction mixture after 1 h.A 50% reduction in the amount of diyne formed.Reaction conditions: [5] 62 mM in MeCN.Initial rate of reaction over a 2.5 min time frame exploring the rate of formation of 15 as a function of (d) catalytic loading of Cu(OAc) 2 and (e) water content (0−20% water).Lines through the data are solely a guide to the eye.Finally, we explored whether the amount of diyne formed (i.e., 15-13 C 4 -15 N 4 ) after completion of the first (3 + 2) cycloaddition was influenced by the presence of 2-15 N 2 in the reaction mixture.This was interrogated by only adding 5-13 C 2 -15 N 2 after the formation of 10 (Figure 8f).Monitoring the reaction by 1 H NMR showed a direct displacement of 10 from the Cu complex as observed previously (Figure 8a).Mass spectrometric analysis of the reaction mixture 1 h after addition of 5-13 C 2 -15 N 2 revealed a 1:1 ratio of 15: 15-13 C 4 -15 N 4 was formed (Figure 8g,h).This is in contrast to the ratio of 1:0.5 formed in the presence of 2-15 N 2 (Figure 8e, further confirmed by HPLC analysis, Table S27).
This suggests that there is ∼2.5 mol % of Cu(II) and Cu(I) present after the first (3 + 2) cycloaddition, and this dynamic mixed valent system is then influenced by the presence of both ynamine and azide substrates.
2.5.Proposed Reaction Mechanism of the Cu-Catalyzed Ynamine-Azide (3 + 2) Cycloaddition.Our study has highlighted how the interplay of two reactions, the Glaser−Hay alkyne coupling and CuAAC, is mediated by the choice of Cu catalyst and the alkyne substrate.Previous mechanistic studies of the Glaser−Hay reaction with conventional alkynes have shown that diyne formation results in concomitant reduction of a binuclear Cu(II) catalyst to form a mixed valent Cu(II)/Cu(I) species. 51,52In this study, the formation of 2.5% diyne 15 when 5 mol % Cu(OAc) 2 is used suggests a Cu(II)-mediated C−H activation of 5 53 followed by a single turnover of the Cu catalyst (i.e., a 2 × 1 electron reduction).This is sufficient to produce a Cu(I) species which then catalyzes the ynamine-azide (3 + 2) cycloaddition reaction to form 10.
Zhu and co-workers identified the importance of the initial binuclear paddlewheel structure of Cu(OAc) 2 , and indeed the role played by the acetate ligands, for alkyne−azide (3 + 2) cycloadditions when a coordinating azide is used. 25,27onsistent with this work, we also show that the N3 position of ynamine 5 is situated in the apical position of Cu(OAc) 2 , displacing H 2 O, i.e., 17.After complexation, a Glaser−Hay reaction produces 15 and the concomitant reduction of 50% of the available Cu catalyst.A mixed valent Cu(I)/(II) species could enter the CuAAC catalytic cycle via the N3 coordination.Disproportionation of CuOAc to a mixed valent state has also been reported, which could suggest a mixture of Cu(I)−Cu(I) and Cu(II)−Cu(I) is present 19 and that either of these could catalyze the (3 + 2) cycloaddition. 39aken collectively, we propose the following mechanism for the ynamine-azide (3 + 2) cycloaddition reaction, which deviates from a conventional CuAAC (Figure 9). 18N3 coordination at the apical face of Cu(OAc) 2 by 5 forms Cu complex 17.Homocoupling of 5 produces diyne 15 and a Cu(I) complex such as 18 25,38,52 could enter the CuAAC catalytic cycle.N3 coordination of 5 produces complex 19 in which the Cu atoms could be mixed valent 19,25 and rapidly forms a binuclear Cu-acetylide species 20. 35Similar mixedvalence Cu(I)/Cu(II) complexes have previously been implicated as catalytically active species for CuAAC. 19,25inally, azide ligation of 2 followed by protodemetalation of the Cu-triazolide forms 10.The final protonation of the Cutriazolide could occur via dissociated acetic acid or by a buildup of 21 as the triazole species 10 forms.When performed in MeCN, protonation of the benzimidazole N3 position of 10 (e.g., 21; pK a H of N-methylbenzimidazole is 4.6) 50 could provide a potential proton shuttle, which would likely be in equilibrium with the acetate ligands associated with Cu(OAc) 2 .These species could then facilitate protodemetalation of the Cu-triazolide. 22revious, speciation studies have highlighted the potential for Cu(OAc) 2 to act as a precatalyst and can form higher order and catalytically competent aggregates, 48 which are highly dependent on water content and Lewis basic sites.Therefore, the proposed mechanism based on binuclear Cu complexes may be an oversimplification, with the system evolving into a series of multinuclear species as the reaction progresses. 25As highlighted by the changes in the chemical shift of the H 2 O (Figures 7b and 8b), H 2 O competes with the benzimidazole N3 of 5 as a ligand for the apical position of the Cu species.Upon completion of the reaction, 10 coordinates to the apical face of a Cu species post CuAAC, which could result in the formation of a postreaction resting complex such as 22, which is catalytically competent for both a Glaser−Hay coupling and CuAAC.

CONCLUSIONS
In summary, a comprehensive investigation of the ynamineazide (3 + 2) cycloaddition reaction has provided insight into how two Cu-catalyzed reactions work in concert to produce 1,4-triazole products.A combination of HPLC, 1 H/ 13 C/ 15 N NMR, EPR and MS analyses has identified how the molecular features of the ynamine substrate manifest in its unique and divergent reactivity relative to conventional alkyne−azide (3 + 2) cycloaddition reactions.The data set emphasizes the susceptibility of the CuAAC reaction to kinetic modification based on simple changes to substrates, i.e., by the inclusion of simple coordinating groups to the alkyne substrate.These changes have clear outcomes: (i) the onset of Glaser−Hay Cu(II) reduction pathway, (ii) increases in the complexity of speciation, and, despite this, (iii) improved CuAAC kinetics.These findings will enhance our understanding of the CuAAC reaction in general and extend the application of ynamine substrates as kinetically advantageous CuAAC partners.
Synthetic experimental protocols, compound characterization, NMR, EPR, and HPLC assay protocols, and MS and pH experimental data and protocols (PDF) Accession Codes CCDC 2274319−2274320 contain the supplementary crystallographic data for this paper.These data can be obtained free of charge via www.ccdc.cam.ac.uk/data_request/cif, or by emailing data_request@ccdc.cam.ac.uk, or

Figure 1 .
Figure 1.(a) Overview of the CuAAC reaction.(b) Accessing Cu(I) via Glaser−Hay coupling.(c) This work: Dynamic behavior of Cu in the ynamine-azide (3 + 2) cycloaddition through initiation and reaction promotion via coordination through interaction of the N3 benzimidazole position.
Is a Source of Cu(I), Promoted by Benzimidazole N3 Coordination.We sought to further understand how Cu(OAc) 2 could promote the Cu(I) catalyzed (3 + 2) cycloaddition between 5 and 2 to form 10. Kuang et al. have proposed that Cu(I) could be formed either during a Cu(OAc) 2 catalyzed Glaser−Hay reaction of an alkyne in MeCN or via a Cu-mediated oxidation of MeOH when used as a solvent,

Figure 2 .
Figure 2. (a) Reaction scheme of ynamine HDE experiments as a function of additive in 9:1 CD 3 CN:D 2 O (for full scope, see Figures S18− S21).(b) Time course of HDE as a function of the Cu source to form 5-D. (c) Time course of HDE as a function of additive and Cu source to form 5-D. (d) Time course of triazole formation (10/10-D) as a function of Cu source.(e) Time course of triazole formation (10/10-D) as a function of additive and Cu source.[5] = 62 mM.Additional added NaOAc was 5 mol %.Lines through data are drawn solely as guides to the eye.

Figure 6 .
Figure 6.Catalyst loading and water effect on (a) formation of 15 from 5 as determined by RP-HPLC.Monitoring the conversion to 15 as a function of Cu(OAc) 2 catalytic loading over 60 min as a function of (b) catalytic loading of Cu(OAc) 2 and (c) water content (0−20% water).Reaction conditions:[5] 62 mM in MeCN.Initial rate of reaction over a 2.5 min time frame exploring the rate of formation of 15 as a function of (d) catalytic loading of Cu(OAc) 2 and (e) water content (0−20% water).Lines through the data are solely a guide to the eye.

Figure 7 .
Figure7.continued of the reaction extracted from stacked 1D13 C data.(i) Asymmetric unit cell of complex 16 formed when 10 is mixed with Cu(OAc) 2 in a 2:1 stoichiometry.Lines through data are solely a guide to the eye.