One-pot Parallel Synthesis of 5-(Dialkylamino)tetrazoles

Abstract

Two protocols for the combinatorial synthesis of 5-(dialkylamino)tetrazoles were developed. The best success rate (67%) was shown by the method which used primary and secondary amines, 2,2,2-trifluoroethylthiocarbamate and sodium azide as the starting reagents. The key steps included the formation of unsymmetrical thiourea, subsequent alkylation with 1,3-propane sultone and cyclization with azide anion. A 559-member aminotetrazole library was synthesized by this approach; the overall readily accessible (REAL) chemical space covered by the method exceeded 7 million feasible compounds.

Graphical Abstract

INTRODUCTION.

According to the recent reports, tetrazole is the third most frequently occurring azole system in the structures of FDA approved drugs.^1,2 A broad range of biological activities of tetrazole derivatives, including antimicrobial, antiviral, antifungal, antitumor, anticonvulsant, antihypertensive properties are extensively reviewed in the literature.^3–6 However, the availability of diversely substituted tetrazoles for their application in medicinal and other industries is highly dependent on the capability of the currently used synthetic methods for both multigram preparation and microgram-scale parallel synthesis.^3,7,8 To date, the reported combinatorial approaches to tetrazole ring construction were limited to a small set of reactions. The most commonly utilized method involves Ugi–azide four-component reaction leading to (1H-tetrazol-5-yl)methanamines.^9–16 Other less commonly used methods include azide cycloaddition with electron-poor nitriles¹⁷ and reaction of peptide N-terminal amino group with arylisothiocyanates.¹⁸

Whereas 5-substituted tetrazoles are well-known bioisosteres of carboxylic acid function,¹⁹ 1,5-disubstituted derivatives can be used to mimic cis-configuration of the amide bond,^4,8 which is higher in energy to the commonly found trans-configuration (Figure 1, A). Recently, 1-alkyl-5-(alkylamino)tetrazole motif was envisaged as a potential replacement of alkyl urea moiety,²⁰ which is present in structures of dozens therapeutics.^21,22 1,5-Disubstituted tetrazoles are represented by a series of approved cephalosporin antibiotics (e.g. Cefmetazole,^23,24 Ceforanide^25,26) and antiplatelet therapeutic Cilostazol^27,28 (Figure 1, B). Therefore, access to the diverse library based on N-substituted 1-alkyl-5-aminotetrazole scaffold 1 became the prime objective of the current work.

Figure 1.

Tetrazoles as bioisosteres of carboxylic acid, amide and urea moieties (A). Representative drugs containing 1,5-disubstituted tetrazoles (B).

One of the options to obtain 1-substituted 5-aminotetrazoles is nucleophilic substitution at the carbon atom of tetrazole bearing a good leaving group, for example amination of 5-sulfonyltetrazoles 2 (Scheme 1, A).²⁰ Alternative methods include the reaction of NaN₃ with aminoiminomethanesulfonic acids 3a,²⁹ α-chloroformamidines 3b,³⁰ (benzotriazolyl)-carboximidamides 3c,³¹ thioureas 4,³² and substituted cyanamides 5,^33,34 the reaction of primary amines with highly reactive cyanogen azide (6),³⁵ and diazotization of aminoguanidines 7.³⁶ Because the aforementioned reagents are not readily accessible and some of them are difficult to handle, they are not suitable for combinatorial synthesis. Recently described one-pot azide–isocyanide denitrogenative coupling followed by cyclization of the resulting carbodiimide with TMSN₃ catalyzed by Pd(0) – Fe(III) (B) might be a better choice for the tetrazole parallel synthesis.³⁷ However, poor availability of isocyanide reagents dramatically limits the scope of this method. One-pot three-component procedure developed by Ponnuswamy and colleagues, which used isothiocyanates, amines and NaN₃ as the starting materials, might be a more plausible method (C).³⁸ Nevertheless, utilization of HgCl₂ for the desulfurization step is undesirable due to the toxicity of mercury contaminants.

Scheme 1.

Reported approaches towards 1-substituted 5-aminotetrazoles 1.

In this work, two different combinatorial approaches towards synthesis of 1-substituted 5-aminotetrazole derivatives 1 were elaborated (Scheme 2). Utility and scope of these methods were tested on a wide range of commercially available substrates. In addition to the synthesis of two 5-aminotetrazole libraries 8 and 9, the capacity of the developed protocols for the generation of large collections for HTS was evaluated.

Scheme 2.

Retrosynthetic analysis of the scaffold 1.

RESULTS AND DISCUSSION.

Our initial efforts were put to the synthesis of 1-substituted 5-aminotetrazoles 8 from isothiocyanates 10 and disubstituted amines 11 (isothiocyanate-based method, Scheme 3). The reaction sequence involved the formation of thioureas 12. In contrast to the previously mentioned work of Ponnuswamy and colleagues (Scheme 1, C),³⁸ our protocol did not involve desulfurization of 12 but instead included their conversion to S-alkylisothiourea intermediates 13. After the alkylation of 12 with 1,3-propane sultone 14, excess of 14 and acidic sulfonate was quenched by adding NEt₃. Finally, the resulting S-alkylisothioureas 13 underwent cyclization with NaN₃ leading to the target products 8, which were further purified by reverse-phase HPLC on the C₁₈ column. This reaction sequence can be considered as a modification of our previous method for the combinatorial synthesis of 3-amino-1,2,4-triazoles, where hydrazides were used as the nucleophilic reagents instead of NaN₃.³⁹.

Scheme 3.

Synthesis of the compounds 8 by the isothiocyanate-based method.

To validate the efficiency of the proposed method, various aliphatic isothiocyanates 10 (58 examples) and secondary aliphatic amines 11 (280 examples) were used. The representative substrates 10 and 11 are listed in Figure 2. As a result, 161 out of 430 experiments allowed obtaining the target tetrazole-containing products (37% success rate, up to 81% yield). For some derivatives, moderate reaction yields could be caused by volatile nature of isothiocyanates with low molecular weight. Besides, in several cases the presence of intermediates 12, 13 and the corresponding carbamates (formed by hydrolysis of 13) was indicated by LC–MS along with the target tetrazoles 8. It was also noticed that branching at the α-carbon atom of isothiocyanates 10 was not crucial for determining the reaction outcome. However, because of the modest success rate, it was difficult to evaluate the precise relationship between structure and reactivity of the reagents 10/11. Therefore, additional modifications of the protocol were further considered.

Figure 2.

Representative reagents 10, 11, and 15

We have suggested that low efficiency of the first step (i.e. thiourea 12 preparation) in the isothiocyanate-based method might be a possible reason for its modest success rate. Recently, our group had reported the efficient application of bis(2,2,2-trifluoroethyl)carbonate (BTC) and -thiocarbonate (BTTC) as reagents for preparation of unsymmetrical ureas^40,41 and thioureas,³⁹ respectively. Therefore, a stepwise reaction of amines 15 and 11 with BTTC was envisaged for the preparation of 12 (BTTC-based method, Scheme 4, A). The reagent is commercially available and can be also synthesized in one step from thiophosgene and trifluoroethanol (Scheme 4, B).

Scheme 4.

Synthesis of the compounds 9 by BTTC-based method.

To ensure regioselectivity of 5-aminotetrazole formation, primary and then – secondary amines were used sequentially in the reaction with BTTC. Both primary (15, 378 compounds) and secondary (11, 262 compounds) amines were randomly selected from our building block collection; their representative examples are given in Figure 2. It should be noted that only aliphatic amines were used, whereas less nucleophilic anilines and heteroaromatic amines were avoided according to their previously noted inefficiency.³⁹ Further transformations of generated in situ thioureas 12 followed those for the isothiocyanate-based method. As a result, a total of 559 target products out of 830 experiments were successfully synthesized (67% success rate, up to 92% yield).

It was found that amines 15 with either primary or secondary aliphatic substituents showed an almost equal success rate for the BTTC-based method (Table 1); amines with tertiary aliphatic groups were not tested for this procedure. In the case of secondary amines, representatives 11a–d had similar reactivity. More α-branched amines 11e demonstrated lower efficiency, although this result should be considered with precaution due to paucity of the corresponding reagents in our set. The presence of functional groups like alcohol, carboxylic acid ester, primary, secondary or tertiary amide, carbamate, sulfone and sulfonamide in amines 11 and 15 was found to be compatible with the reaction conditions. Considering electronic properties, it was found that electron-withdrawing substituents (e.g. di- and trifluoroethyl) in amine 15 were unfavorable for the discussed transformations. Additionally, an attempt to carry out the reaction sequence with O-alkylhydroxylamines was unsuccessful in all cases.

Table 1.

Influence of amines 11 and 15 substitution pattern on the reaction success rate.

	Substitution pattern	Success rate	Total number of experiments
15a		68%	685
15b		64%	138
11a		70%	131
11b		66%	38
11c		68%	598
11d		78%	40
11e		40%	15

A subset of selected products obtained by both discussed synthetic approaches is depicted in Figure 3. When comparing the two aforementioned methods, application of BTTC-based sequence not only improved the success rate of the approach but also extended the accessible pool of the starting reagents. The wide availability of primary amines from commercial sources compared to the corresponding isothiocyanates makes BTTC-based protocol a powerful tool for achieving a great variety of final products. Additionally, because the developed protocol is tolerant to a wide range of polar functional groups and produces a wide range of lead-like compounds (95% of the synthesized products had MW 200–350 Da and cLogP 1–3, Figure 4),⁴² it provides a new potent method for lead-oriented synthesis. The calculated cLogP drift value (i.e. the difference of the mean cLogP of a synthesized array and the mean cLogP of the designed array)⁴² of 0.017 units was also beneficial for BTTC-based approach.

Figure 3.

Representative 1-substituted 5-aminotetrazoles 8 and 9 derived products prepared by isothiocyanate-based (A) and BTTC-based (B) methods.

Figure 4.

MW and cLogP characteristics of the synthesized (filled bars) and designed arrays (hollowed dashed bars).

Expanding chemical space of synthetically feasible 5-aminotetrazoles.

The major role of parallel synthesis has been providing a great number of new structurally related compounds for facile application in biological, physicochemical and other investigations. Nowadays, well-elaborated combinatorial protocols can also provide a base for prediction of all theoretical structures from already known starting materials. In several cases, it is more practical to model potentially accessible virtual chemical space,⁴³ analyze its properties by in silico methods,⁴⁴ and pick only the most promising representatives before carrying out the synthetic part. This approach not only avoids the waste of resources on the preparation of redundant compounds, but also facilitates the discovery of better hits due to an increased pool of processed ligands by virtual high-throughput screening (vHTS).^45,46 For this reason, REAL database⁴⁷ (REAL = readily accessible) was developed by us to accumulate highly feasible chemical compounds which can be obtained from available reagents by well-validated combinatorial chemistry protocols. Recently, the structure-based docking of 170 million feasible compounds from this database against AmpC and the D₄ dopamine receptor allowed to find the most potent AmpC reversible inhibitor, as well as 158 thousand potential D4 ligands with sub-1-μM predicted affinity values.⁴⁸ Thus, a further objective of the current work was to demonstrate the extension of accessible chemical space by the developed method.

The current commercial availability and literature coverage of 5-aminotetrazoles was estimated by substructural searches in eMolecules⁴⁹ and PubChem⁵⁰ databases, which revealed 17,219 and 55,163 compounds, respectively (Figure 5). To estimate the number of 5-aminotetrazoles feasible by isothiocyanate-based and BTTC-based methods, the combinations of commercially available secondary amines 11 with isothiocyanates 10 or primary amines 15, respectively, were selected by in-house developed algorithms based on our previous data on the reactivity of the reagents.^39–41 Because the abundance of primary amines is significantly greater than amount of isothiocyanates, it was not surprising that the number of BTTC-derived products (7,097,307) was 20 times larger than that of isothiocyanate-derived (326,702). Additionally, because the probability of successful preparation of the target products by the BTTC-based method was 67%, these 7 million virtual structures can be considered as REAL substituted 5-aminotetrazoles.

Figure 5.

Substituted 5-aminotetrazoles in commercial, public, and the REAL databases.

The utility of the generated REAL database subset for vHTS is ensured by the high percentage of structures passing PAINS⁵¹ (99%) and Lilly⁵² (95%) filters. The database scaffold diversity was represented by 809,308 Bemis – Murcko frameworks and included 56% of singletons. These characteristics make the REAL 5-aminotetrazole database a useful tool for lead discovery and optimization programs.

CONCLUSIONS

Two approaches for the combinatorial synthesis of 5-(dialkylamino)tetrazoles commencing from aliphatic isothiocyanates and 2,2,2-trifluoroethylthiocarbamate (BTTC) were developed. Of these protocols, the BTTC-based method was found to be superior. This approach involved the sequential reaction of primary and then – secondary amines with 2,2,2-trifluoroethylthiocarbamate, subsequent alkylation of the resulting unsymmetrical thiourea with 1,3-propane sultone, and cyclization with azide anion. As a result, a 559-member 5-aminotetrazole library were obtained with up to 92% yield (67% success rate). Wide accessibility of the starting reagents from commercial sources, together with tolerance to a broad range of polar functional groups make the BTTC-based protocol a new potent method for lead-oriented library synthesis. Immense expansion of chemical space of synthetically feasible aminotetrazoles was demonstrated by generation of the REAL (readily accessible) database containing over 7 million structures, 95% of which satisfied common medicinal chemistry filters.

EXPERIMENTAL SECTION

General Methods.

The solvents were purified according to the standard procedures.⁵³ All starting materials were taken at Enamine Ltd. The success rate was calculated as the number of successful experiments divided by the number of failed experiments. Reverse phase column chromatography was performed using C₁₈-modified silica gel as a stationary phase. ¹H, ¹⁹F, and ¹³C NMR spectra were recorded at 500 MHz, 470, and 126 MHz, respectively. Chemical shifts are reported in ppm downfield from TMS as internal standards. Mass spectra were recorded on an LC–MS instrument with chemical ionization (CI). LC–MS data were acquired on an Agilent 1200 HPLC system equipped with DAD/ELSD/LCMS-6120 diode matrix and mass-selective detector. Melting points were measured on a MPA100 OptiMelt automated melting point system. Elemental analyses were performed at the Laboratory of Organic Analysis, Department of Chemistry, Taras Shevchenko National University of Kyiv.

General protocol for 5-aminotetrazole synthesis (isothiocyanate-based method).

Isothiocyanate 10 (1.00 mmol) and amine 11 (1.00 mmol) were added to a sealable vial containing CH₃CN (0.7 mL). If the amine 11 was used as a salt, i-Pr₂NEt (1.20 mmol) was also added. After sealing, the vial was heated at 100 °C for 10 min and shaken vigorously. The reaction mixture was heated at 100 °C until full dissolution, and then – for additional 30 min. 1,3-Propane sultone (2.00 mmol) was added to the solution, the reaction mixture was shaken at 50 °C overnight and then was heated at 100 °C for 16 h. Et₃N (3.00 mmol) was added to the resulting mixture and the reaction mixture was heated at 100 °C for 30 min. NaN₃ (1.10 mmol) was added, the mixture was shaken at rt overnight and then stirred at 50 °C for 12 h. The resulting reaction mixture was evaporated and further dissolved in CHCl₃ (3 mL). The organic layer was washed with H₂O (2×4 mL). The organic phase was evaporated and subjected to HPLC purification on a C₁₈-column using MeOH − H₂O as eluent to give 8 with >90% purity.

General protocol for 5-aminotetrazole synthesis (BTTC-based method).

Amine 15 (1.00 mmol) and 2,2,2-trifluoroethylthiocarbamate (1.00 mmol) were added to a sealable vial containing Et₃N (1.00 mmol) in CH₃CN (0.5 mL). If the amine 15 was used as a salt, Et₃N (1.20 mmol for mono- and 2.20 mmol for dihydrochlorides) was also added. After sealing, the vial was shaken at rt overnight and then heated at 100 °C for 3 h. Next, amine 11 (1.00 mmol) was added to the reaction mixture. If the amine 11 was used as a hydrochloride – additional Et₃N (1.10 mmol) was added. After heating at 100 °C for 7 h, the reaction mixture was evaporated under vacuum and the residue was dissolved in CH₃CN (0.7 mL). 1,3-Propane sultone (2.00 mmol) was added to the solution, the reaction mixture was shaken at 50 °C overnight and then heated at 100 °C for 16 h. Et₃N (3.00 mmol) was added, and the reaction mixture was heated at 100 °C for 30 min. NaN₃ (1.10 mmol) was added, the mixture was shaken at rt overnight and then stirred at 50 °C for 12 h. The resulting reaction mixture was evaporated and further dissolved in CHCl₃ (3 mL). The organic layer was washed with H₂O (2×4 mL). The organic phase was evaporated and subjected to HPLC purification on a C₁₈-column using MeOH − H₂O as eluent to give 9 with >90% purity.

ABBREVIATIONS

BTC

bis(2,2,2-trifluoroethyl)carbonate

BTTC

2,2,2-trifluoroethylthiocarbamate

REAL

readily accessible