Combined XRD-paramagnetic ¹³C NMR spectroscopy of 1,2,3-triazoles for revealing copper traces in a Huisgen click-chemistry cycloaddition. A model case

Introduction

[1,2,3]-triazoles are a very well-known class of nitrogen aromatic heterocyclic compounds, continuously widespread in different fields such as synthetic organic chemistry1, medicinal chemistry2, polymer science,3 and catalysis4, amongst others. Their 1,2,3,-triazo-aromatic scaffolds comprise stable compounds to even acidic/basic hydrolysis conditions or reductive/oxidative environments due to their high aromatic stabilization.5 Moreover, this framework is relatively resistant to metabolic degradation.6 In order for an exclusive and high yield of a stereospecific 1,4-disubstituted [1,2,3]-triazole regioisomer, a modular synthetic approach is used: the Huisgen azidealkyne 1,3-dipolar cycloaddition7 recently adapted to the “click chemistry” philosophy8, wherein said cycloaddition is carried out in mild-aqueous conditions to produce high-yield / stereospecific products, using straightforward purification methods (crystallisation or distillation). Some strategies to install highly hindered groups within the heterocyclic ring comprise: increasing the Cu(I) catalyst amount and/or implementing harsh reaction conditions (for instance: microwave radiation combined with longer reaction periods).

On one hand, adamantane coupled to specific pharmaceutical formulations has provided an excellent lipophilicity due to its absorption, distribution, metabolism, and excretion modulation effects9. Examples of molecules with biological activities that include any adamantyl scaffold in their structures are amino adamantane derivatives10, Rimantadine11, and Amantadine hydrochloride12. On the other hand, the “cyclopropyl fragment” is a versatile player that frequently appears in clinical drug studies. Some notable pharmaceutical properties that are influenced by the cyclopropane ring include, amongst others, intrinsic lipophilicity13, metabolic stability14, and binding to target proteins15. Some drugs that contain a cyclopropyl fragment are Trametinib (renal carcinoma drug), Volasertib B1/6727 (selective PLK1 inhibitor for acute myeloid leukemia) and Vaniprevir (hepatitis C virus). Given the proven effectiveness of both a 1,2,3-triazole core and adamantyl-cyclopropyl substituents in different pharmaceutical formulations, as for other applications stressed above, it is thought a natural step to test clinical advantages of 1-adamantyl-4-cyclopropyl-1,2,3-triazole scaffolds. A further advantage of 1,4-disubstituted 1,2,3-triazoles is the simplicity and flexibility of their use in derivatizations (for instance to produce 1,3,4-trisubstituted 1,2,3-triazolium salts, precursors of N-heterocyclic carbene metal complexes, used as organocatalysts for key SN regioselective arrangements or as ionic liquids 16, 17).

Despite the significant number of reports enhancing the synthetic advantages of a one-pot step to regiospecifically unite azides and terminal acetylenes to form 1,4-disbstituted 1,2,3-triazoles with the aim of Cu (I) catalyzed ligation1, 18,19, few have discussed the efficient removal of metal traces in Copper - Catalyzed Alkyne-Azide Cycloaddition (CuAAC) “click” reactions20. Presence of Cu (I) or its paramagnetic Cu (II) raw material could produce inaccurate conclusions in terms of molecular assignments. For instance, paramagnetic species (with unpaired electrons in metals’ d-shells) in proximity to organic centers will produce important changes in experimental ¹H and/ or ¹³C NMR spectra of said organic species. Major spectral modifications comprise temperature-dependent large line widths and/or chemical shift dispersions. Active electron paramagnetic spin relaxation broadens signals from nuclei near paramagnetic centers by a direct dipolar coupling interaction between the unpaired electron spin with nuclear spins in the vicinity 21, 22. Furthermore, the appearance of a Fermi-contact shift contribution in the isotropic chemical shift promotes shift dispersions of neighboring nuclear spins, an effect that nowadays can be straightforwardly described with the calculation of an electron-spin density interaction with molecular orbitals of organic centers, using Density Functional Theory approach 23, 24. However, solution-state 13C detected paramagnetic NMR spectroscopic studies of samples directly or indirectly affected by active electron-spin interactions have been scarcely reported25.

In contrast, when paramagnetism is present as copper traces in CuAAC processes, said presence could not necessarily be unambiguously assigned with X-Ray Diffraction (XRD) data, typically using copper K(α) radiationto collect data, due to known polychromatic radiation effects of some paramagnetic species present in the monocrystal when an intense beam of wavelength equal to 1.5418 Å is used.

In the present work, we report for the first time the synthesis of the 1,4-disubstituted 1,2,3-triazole: 1-((3s,5s,7s)-adamantan-1-yl)-4-cyclopropyl-1H-1,2,3-tri-azole, its treatment with ammonia solution as standard removal strategy of copper traces and, finally, its derivatization to 1-((3s,5s,7s)-adamantan-1-yl)-4-cyclopropyl-3-methyl-1H-[1,2,3]-triazol-3-ium triflate salt (Scheme 1). Both 1,4-disubstituted and 1,3,4-trisubstituted compounds were crystallized for exhaustive XRD structural analysis. Finally, 13C detected paramagnetic NMR spectroscopy of the 1,4-disubstituted molecule, its ammonia washed byproduct and the 1,3,4-trisubstituted triazole were carried out to confirm the presence of copper paramagnetic traces. Clinical advantages of novel triazole molecular architectures and their metal-related toxicity are out of the scope of the present study, and will be published elsewhere.

Scheme 1

Synthetic route for the present Huisgen “Click chemistry” reaction and experimental conditions to produce the 1,4-disubstituted triazole: 1-((3s,5s,7s)-adamantan-1-yl)-4-cyclopropyl-1H-1,2,3-triazole, followed by its treatment with ammonia solution as standard removal strategy of copper traces and final derivatization to produce the 1,3,4-trisubstituted triazole: 1-((3s,5s,7s)-adamantan-1-yl)-4-cyclopropyl-3-methyl-1H-[1,2,3]-triazol-3-ium triflate salt. As stated in reference 18, the one-pot 1,3-dipolar copper catalyzed cycloaddition promotes the Cu-addition in position 5 of the 1,2,3-triazol, whereas its presence and ineffective removal, could explain the paramagnetic effects discussed in the present work.

Results and discussion

The synthesis of 1,4-disubstituted and 1,3(methyl),4-tri-substituted 1,2,3-triazoles with adamantyl and cyclopropyl scaffolds as promising molecules with lipophilicity and metabolic stability are described (see Scheme 1 and synthetic details in Experimental, vide infra). First, the 1,4-disubstituted-1,2,3-triazole was prepared following standard Huisgen 1,3-cycloaddition copper catalyzed conditions, with expected high yields (above 80%). After carrying out standard purification methods for all molecules herein presented (Experimental, vide infra) 4, 16, 26, high quality monocrystals of : 1-((3s,5s,7s)-adamantan-1-yl)-4-cyclopropyl-1H-1,2,3-triazole (Figure 1) and 1-((3s,5s,7s)-adamantan-1-yl)-4-cyclopropyl-3-methyl-1H-[1,2,3]-triazol-3-ium triflate salt (Figure 2) were collected for XRD studies (see Crystallography section, vide infra). Key geometrical data of both crystallographic data is presented in Figure 3. The same purified samples as well as that obtained from the washing ammonia solution of 1-((3s,5s,7s)-adamantan-1-yl)-4-cyclopropyl-1H-1,2,3-triazole were dissolved in deuterated CDCl₃ for ¹³C-NMR studies (Figure 4).

Figure 1

XRD Ortep plot of 1-((3s,5s,7s)-adamantan-1-yl)-4-cyclopropyl-1H-1,2,3-triazole with 50% probability thermal ellipsoids for non-H atoms. Atomic labels, crystallographic data, fractional atomic coordinates, isotropic and atomic displacements as well as geometrical parameters are reported within the supplementary material. CCDC accession number: 1896662.

Figure 2

XRD Ortep plot of 1-((3s,5s,7s)-adamantan-1-yl)-4-cyclopropyl-3-methyl-1H-[1,2,3]-triazol-3-ium triflate salt with 50% probability thermal ellipsoids for non-H atoms. Atomic labels, crystallographic data, fractional atomic coordinates, isotropic and atomic displacements as well as geometrical parameters are reported within the supplementary material. CCDC accession number: 1896671.

Figure 3

Key geometrical parameters (Å, °) of 1-((3s,5s,7s)-adamantan-1-yl)-4-cyclopropyl-1H-1,2,3-triazole (blue histograms) and 1-((3s,5s,7s)-adamantan-1-yl)-4-cyclopropyl-3-methyl-1H-[1,2,3]-triazol-3-ium triflate salt (red histograms). Standard deviations of bonds [± 0.005-0.01, blue histograms]; [± 0.03-0.06, red histograms] and angles [± 0.31-2.8, blue histograms]; [± 1.9-4.6, red histograms] between asymmetric units in both crystal structures are represented as error bars.

Figure 4

¹³C-direct excitation NMR spectra of 1-((3s,5s,7s)-adamantan-1-yl)-4-cyclopropyl-1H-1,2,3-triazole (top), the ammonia washing by-product (middle) and 1-((3s,5s,7s)-adamantan-1-yl)-4-cyclopropyl-3-methyl-1H-[1,2,3]-triazol-3-ium triflate salt (bottom), carried out at exact experimental conditions.

The crystal structure of 1-((3s,5s,7s)-adamantan-1-yl)-4-cyclopropyl-1H-1,2,3-triazole (T= 100K) presents two independent molecules within its asymmetric unit of the defined Monoclinic P2₁ space group. The derivatized 1-((3s,5s,7s)-adamantan-1-yl)-4-cyclopropyl-3-methyl-1H-[1,2,3]-triazol-3-ium triflate salt (T = 298 K) crystal structure is, in contrast, built up with four crystallographic independent molecules within its triclinic space group. Comparisons between both crystal structures were possible by obtaining mean geometric parameters of both 1,4-disubstituted (Z=2) and 1,3,4-trisubstituted (Z=4) 1,2,3-triazoles, displayed in Figure 3 (depicting standard deviations as error bars within the histograms), highlighting the following observations:

1. Both 1,2,3-triazole pentacycles present expected bond N-N (1.33 – 1.34 Å), N-C (1.33 – 1.35 Å) and C-C ranges (1.36 Å), confirming the effective Cu catalytic cycloaddition of a Huisgen process.

2. An equivalent triazole – adamantyl distance of N₁-C_ADM of 1.49 Å between both crystal structures is as well within the expected ranges of substituent values of N-C bonds in triazoles.

3. The effect of methyl substitution in the triflate salt might be regarded as follows: elongation of the C₄-C_CP(i) bond distance (1.54 Å) and the distortion of the N₁-C₅-C₄ angle (99.7^o) of the triazole ring, with respect expected ranges (respectively C₄-C_CP(i) = 1.49 Å, N₁-C₅- C₄ = 106.9^o observed in 1-((3s,5s,7s)-adamantan-1-yl)-4-cyclopropyl-1H-1,2,3-triazole), mostly due to the ionic triazolium-triflate stabilization.

However, abnormalities within the high resolution crystal structure of 1-((3s,5s,7s)-adamantan-1-yl)-4-cyclo-propyl-1H-1,2,3-triazole have been observed at the level of torsional angles. Important deviations of the torsional angles between the triazole ring and its 1,4-substituents are observed between asymmetric molecules within the unit cell: C₅-N₁-C_ADM(i)-C_ADM(ii) = 96^o (I), C₅-N₁-C_ADM(i)-C_ADM(ii) = 122^o (II); N₁-C₄-C_CP(i)-C_CP(ii) = 53^o (I), N₁-C₄-C_CP(i)-C_CP(ii) = 39^o (II). Said torsional angles retain consistency and remain in higher agreement with respect to expected values, for its triflate triazolium salt counterpart: C₅-N₁-C_ADM(i)-C_ADM(ii) = 127^o (I), C₅-N₁-C_ADM(i)-C_ADM(ii) = 145^o (II), C₅-N₁-C_ADM(i)-C_ADM(ii) = 123^o (III), C₅-N₁-C_ADM(i)-C_ADM(ii) = 132^o (IV); N₁-C₄-C_CP(i)-C_CP(ii) = 0^o (I), N₁-C₄-C_CP(i)-C_CP(ii) = 18_o (II), N₁-C₄-C_CP(i)-C_CP(ii) = 20^o (III), N₁-C₄-C_CP(i)-C_CP(ii) = 24^o (IV).

Despite its higher symmetry, less asymmetric molecules per unit cell (Z=2), higher symmetry between molecules in the unit cell (standard deviations of bonds and angles respectively in the ranges of [± 0.005-0.01], [± 0.31-2.8]) and better discrepancy index (R_int = 0.031) with respect to the triazolium triflate crystal structure (Z= 4, standard deviations of bonds and angles respectively in the ranges of [± 0.03-0.06], [± 1.9-4.6] and R_int = 0.176), observed abnormalities of torsional angles between triazole ring and substituents in 1-((3s,5s,7s)-adamantan-1-yl)-4-cyclopropyl-1H-1,2,3-triazole, might suggest the presence of additional electronic density not observed with the present XRD experimental and instrumental conditions. Orthogonal methods shall shed light to explain said discrepancies.

Assuming the presence of copper catalyst traces, nuclear magnetic resonance chemical shift- and particularly ¹³C-line widths at half-height (ν_1/2) analysis, provide unequivocal evidence of the presence of paramagnetic effects. Figure 4 presents the 13C-NMR spectra of the 1,4-disubstituted-1,2,3-triazole, before (top) and after (middle) copper removal procedure with NH₄OH, as well as 1,3,4-trisubstituted-1,2,3-triazoleum salt (bottom). Chemical shift inspection of the CuAAC cycloaddition product prior to any ammonia washing procedure, reveals only the presence of adamantyl carbon spin systems. Cyclopropyl and triazole carbon resonances are not observed, strongly suggesting the presence of fast electron spin relaxation from paramagnetic Cu (II) that will broaden 13C signals from nuclei in close proximity to the paramagnetic center (See Scheme 1). As elsewhere discussed20, nonquantitative ammonia removal of copper traces produces a byproduct with a non-negligible quantity of paramagnetic copper, as revealed in the 13C-NMR spectrum of washed 1-((3s,5s,7s)-adamantan-1-yl)-4-cyclopropyl-1H-1,2,3-triazole (Figure 4, middle). Cyclopropyl NMR resonances also appear, but not those from the 1,2,3-triazole moiety. 13C-NMR chemical shift analysis of 1-((3s,5s,7s)-adamantan-1-yl)-4-cyclopropyl-3-methyl-1H-[1,2,3]-triazol-3-ium triflate salt reveals the presence of all adamantyl, cyclopropyl and 1,2,3-triazole spin systems, strongly suggesting the better copper removal with methyl triflate. However in order to conclude the effectiveness in removing paramagnetic effects of 1,2,3-triazoles, a deeper inspection must be done by line width analysis. First, 1H-NMR line widths (see Experimental) of all molecules herein presented reveal line-broadening effects in all cases that produce, in turn, non-exploitable spectra which are difficult to assign. 1-((3s,5s,7s)-adamantan-1-yl)-4-cyclopropyl-1H-1,2,3-triazole, its ammonia treated by-product and the 1,3,4-trisubstituted-1,2,3,triazolium salt present proton line widths in the range of, respectively: 20.3 Hz (only adamantyl broad resonance is observed), 6.25 – 13.28 Hz and 4.12 – 20.55 Hz.

In contrast, 13C line widths –less dependent to paramagnetic effects²²- reveal carbon line broadening of 1-((3s,5s,7s)-adamantan-1-yl)-4-cyclopropyl-1H-1,2,3-triazole, its ammonia treated by-product and the 1,3,4-trisubstituted-1,2,3,triazolium salt in the range of, respectively: c.a. 2-4 Hz, 1-2.5 Hz and 1-2 Hz (Figure 5). Proton and carbon line widths serve, in turn, to disentangle the presence of paramagnetic effects in all cases. However, a significant improvement of copper removal is done when methyl triflate is used.

Figure 5

¹³C-detected chemical shifts as a function of measured line widths at half height (ν_1/2) of 1-((3s,5s,7s)-adamantan-1-yl)-4-cyclopropyl-1H-1,2,3-triazole (magenta triangles), the ammonia washing by-product (red crosses) and 1-((3s,5s,7s)-adamantan-1-yl)-4-cyclopropyl-3-methyl-1H-[1,2,3]-triazol-3-ium triflate salt (blue circles).

Conclusions

This work presents a strategy to reveal the presence of paramagnetic copper (II) traces in a standard Huisgen CuAAC cycloaddition click-chemistry process, using as a model system the novel synthesis of 1-((3s,5s,7s)-adamantan-1-yl)-4-cyclopropyl-1H-1,2,3-triazole and its derivatized 1-((3s,5s,7s)-adamantan-1-yl)-4-cyclopropyl-3-methyl-1H-[1,2,3]-triazol-3-ium triflate salt. Chemical shift and line widths analysis in paramagnetic NMR spectra strongly suggest the presence of copper traces that mostly affect cyclopropyl and 1,2,3-triazole spin systems, due to their proximity to the paramagnetic center. Particularly a Cu-C₅ adduct is the most plausible possibility that could explain carbon line broadening of said ¹³C resonances in 1-((3s,5s,7s)-adamantan-1-yl)-4-cyclopropyl-1H-1,2,3-triazole and its ammonia washing solution product. This result, in turn, might explain the torsional angle abnormalities of 1-((3s,5s,7s)-adamantan-1-yl)-4-cyclopropyl-1H-1,2,3-triazole substituents, observed in XRD 3D-data carried out with a Cu K(α) radiation beam. Whereas close copper centers could promote said structural distortions, paramagnetic copper centers would not be directly observed by XRD when diffraction patterns are collected with a radiation beam of wavelength = 1.5418 Å, as commonly used in organic chemistry.

Experimental