Abstract

The transformation of 3-[(ethoxymethylene)amino]-1-methyl-1H-pyrazole-4-carbonitrile into the 14-membered macrocycle, 2,10-dimethyl-2,8,10,16-tetrahydrodipyrazolo[3,4-e:3',4'-l][1,2,4,8,9,11]hexaazacyclotetradecine-4,12-diamine, by the reaction with excess hydrazine under various conditions was studied in detail. The reaction proceeded through the initial formation of 4-imino-2-methyl-2,4-dihydro-5H-pyrazolo[3,4-d]pyrimidin-5-amine followed by dimerization to give the final macrocycle. A convenient synthesis of the latter starting from 4-imino-2-methyl-2,4-dihydro-5H-pyrazolo[3,4-d]pyrimidin-5-amine was developed. A plausible pathway for the macrocycle self-assembly is discussed. Some features of the structure and reactivity of the obtained macrocycle are outlined.

Introduction

The chemistry of polyazamacrocycles (PAMs) is currently one of the most rapidly developing areas of heterocyclic chemistry . The great interest in PAMs is primarily due to their ability to bind various cations, anions, and neutral molecules . In addition, some representatives of PAMs were found in various natural products and play an important role in living systems (e.g., vitamin B12, chlorophyll, metalloproteins, cyclic peptides, etc). PAMs themselves and their metal complexes exhibit various useful properties , particularly, they possess a wide range of biological activities and are used as contrast agents for magnetic resonance imaging, radiopharmaceuticals, sensors, NMR shift reagents, luminescent materials, catalysis, etc.

To date, a large variety of PAMs with various ring sizes, number and location of nitrogen atoms, levels of unsaturation, etc. have been prepared and studied. Nevertheless, the synthesis of novel PAMs with interesting properties is of great importance. Recently, we developed some approaches to 14-membered cyclic bis-semicarbazones and bis-thiosemicarbazone , namely 7,14-dimethyl-1,2,4,8,9,11-hexaazacyclotetradeca-7,14-diene-3,10-diones and -3,10-dithiones. The prepared compounds were able to chelate various metal cations through the N1, N4, N8, and N11 atoms . In continuation of our research on 1,2,4,8,9,11-hexaazamacrocycles, which are under-explored representatives of the PAMs family, we were particularly interested in the synthesis of more unsaturated and therefore more conformationally rigid compounds. Previously, the unintentional preparation of two polyunsaturated 1,2,4,8,9,11-hexaazamacrocycles fused with two benzene or two pyrazole rings has been reported . In particular, Dolzhenko et al. attempted to reproduce the synthesis of 4-imino-2-methyl-2,4-dihydro-5H-pyrazolo[3,4-d]pyrimidin-5-amine described by Baraldi et al. using the reaction of 3-[(ethoxymethylene)amino]-1-methyl-1H-pyrazole-4-carbonitrile with excess hydrazine hydrate in EtOH under reflux. However, a pyrazole-fused 1,2,4,8,9,11-hexaazamacrocycle was unexpectedly obtained instead . Since this type of macrocycle self-assembly seems to be very promising, we decided to reproduce the Dolzhenko’s procedure, then study the macrocyclization in detail, and extend this approach to the synthesis of other polyunsaturated 1,2,4,8,9,11-hexaazamacrocycles.

Herein, we report the detailed studies of the hydrazine-promoted transformation of 3-[(ethoxymethylene)amino]-1-methyl-1H-pyrazole-4-carbonitrile (4) or 4-imino-2-methyl-2,4-dihydro-5H-pyrazolo[3,4-d]pyrimidin-5-amine (8) into 2,10-dimethyl-2,8,10,16-tetrahydrodipyrazolo[3,4-e:3',4'-l][1,2,4,8,9,11]hexaazacyclotetradecine-4,12-diamine (5) under various conditions. Mechanistic aspects of the macrocyclization are also discussed. Some features of the structure and reactivity of the obtained macrocycle are outlined.

Results and Discussion

The readily available 3-amino-1-methyl-1H-pyrazole-4-carbonitrile (3) was used as the starting material. This compound was prepared according to the described regioselective method based on the reaction of malononitrile with triethyl orthoformate followed by subsequent treatment of the obtained dinitrile 2 with benzaldehyde methyl hydrazone in benzene, conc. aqueous HCl in EtOH, and NaOH in water (Scheme 1). The key intermediate of the macrocycle preparation, imidate 4, was synthesized using the reported procedure by refluxing a solution of aminopyrazole 3 in triethyl orthoformate.

Scheme 1: Synthesis of the key intermediate of the macrocycle preparation, 3-[(ethoxymethylene)amino]-1-methyl-1H-pyrazole-4-carbonitrile (4).

First, we studied the reaction of imidate 4 with hydrazine hydrate in EtOH under Dolzhenko’s conditions (4 equivalents of N2H4·H2O, concentration of 4 = 0.5 mmol/mL, reflux, 2 h) . The resulting precipitate was isolated by filtration and washed with EtOH. In contrast to the reported data, the yield of the obtained product was significantly lower and did not exceed 38%. Moreover, according to NMR spectroscopic data, the isolated product was a mixture of the desired macrocycle 5 and a noticeable amount of an impurity (Scheme 2) whose formation was not mentioned in the cited reference.

Scheme 2: Synthesis of macrocycle 5 by the reaction of imidate 4 with hydrazine hydrate.

The structure of the concomitant impurity was established using 1D and 2D NMR spectroscopy. The 1H NMR spectrum in DMSO-d6 shows the presence of two methylpyrazole moieties (singlet signals of two methyl groups at 3.63 and 3.70 ppm, singlet signals of two CH protons of pyrazole rings at 7.81 and 7.84 ppm), a H–N–C–H fragment with trans-orientation of protons (two doublets at 9.87 and 7.50 ppm, 3J = 11.2 Hz), four NH2 groups at 6.27, 5.72, 5.59, and 4.61 ppm (singlets). Signals of 11 different carbon atoms including 8 carbons of two methylpyrazole moieties were observed in the 13С NMR spectrum. Thus, we concluded that the impurity has bis-pyrazole structure 6. This structure was also confirmed by the 1H,13C-HSQC and 1H,13C-HMBС spectra, as well as by comparing the experimental carbon chemical shifts in DMSO-d6 with those calculated for 6 by the GIAO method at the PBE1PBE/6-311+G(2d,p) level of theory using the DFT B3LYP/6-311++G(d,p) optimized geometries (DMSO solution) and applying a multi-standard approach (see the Supporting Information File 1 for details). The high-resolution mass spectrum (ESI+) of a mixture of 5 and the impurity, in addition to a peak at m/z = 329.1696 [M + H]+ for compound 5, shows a peak at m/z = 319.1862 [M + H]+, consistent with the molecular formula of C11H18N12 for bis-pyrazole 6. According to NMR spectroscopic data, the amount of bis-pyrazole 6 in the crude product formed under above conditions was about 18 mol %.

The structure of macrocycle 5 was confirmed by comparing its 1H and 13C NMR spectra with those reported in ref. . It should be noted that the 1H and 13C NMR spectra of compound 5 in DMSO-d6 show only a half-number set of proton or carbon signals (five and six signals, respectively), thus indicating its C2-symmetric dimeric structure. The analysis of 2D NMR spectroscopic data provided additional evidence for the macrocycle 5 structure (see Supporting Information File 1). The high-resolution mass spectrum (ESI+) confirmed its chemical formula as C12H16N12.

Thus, we found that, in contrast to the reported data , the reaction between imidate 4 and hydrazine hydrate (4 equiv) in refluxing EtOH for 2 h afforded macrocycle 5 in a relatively low yield, along with an appreciable amount of the byproduct 6. This prompted us to optimize the reaction conditions varying hydrazine hydrate excess (from 3.1 to 4.3 equivalents), solvent (EtOH, MeOH, 1,4-dioxane, DME), reaction time (2 h and 6 h), and also using anhydrous hydrazine instead of hydrazine hydrate. However, all our attempts to improve both the yield and the purity of 5 failed. For example, prolonging the reaction time between 4 and N2H4∙H2O (4 equiv) in refluxing EtOH to 6 h resulted in an increase in the purity of the macrocycle (5/6 = 91:9), but simultaneously to a decrease in its yield to 25%. Reducing the amount of hydrazine hydrate to 3 equivalents (EtOH, reflux, 2 h) had a similar effect and gave an 85:15 mixture of 5 and 6 in an overall yield of 22%. In refluxing MeOH (3 equiv of N2H4∙H2O, 2 h), a mixture of 5 and 6 in a ratio of 89:11 was obtained in 31% overall yield. In aprotic solvents (1,4-dioxane or DME), the selectivity of the reaction dramatically decreased and a mixture of 5 and 6 along with significant amounts of various unidentified byproducts was formed. For example, the reaction of 4 with N2H4∙H2O (4.1 equiv) in refluxing 1,4-dioxane for 2 h, followed by evaporation of the volatiles under reduced pressure, afforded a complex mixture containing only 7 mol % of macrocycle 5 according to the 1H NMR spectrum with the addition of a weighted amount of succinimide as a reference. Analogously, only 2 mol % of 5 were detected under the above conditions (1,4-dioxane, reflux, 2 h) when anhydrous hydrazine (4.2 equiv) was used as a promoter.

A plausible pathway for the transformation of imidate 4 into macrocycle 5 is shown in Scheme 3. This pathway includes fast substitution of the ethoxy group by hydrazine to give the intermediate amidrazone 7 followed by its rapid conversion to pyrazolopyrimidine 8. Slow dimerization of compound 8 results in macrocycle 5.

Scheme 3: Plausible pathway for the transformation of imidate 4 into macrocycle 5.

The formation of 5 through pyrazolopyrimidine 8 is confirmed by the literature data that the reaction of 4 with N2H4∙H2O in EtOH to give 8 proceeds under much milder conditions than the reaction to afford 5 (rt and reflux, respectively). Based on this background, we assumed that the synthesis of macrocycle 5 could be carried out directly from 8. We also hoped that this would be especially useful from a preparative viewpoint, since pure pyrazolopyrimidine 8 can be easily obtained in any required quantities, in contrast to pure imidate 4.

Pyrazolopyrimidine 8 was prepared by the reaction of 4 with N2H4∙H2O in EtOH according to our modification of the described procedure using room temperature (without pre-cooling), a lower excess of N2H4∙H2O (1.6 equiv instead of 5 equiv) and a shorter reaction time (1 h instead of 5 h). The precipitated compound 8 was isolated by filtration in a 96% yield (Scheme 4).

Scheme 4: Synthesis of pyrazolopyrimidine 8 by the reaction of imidate 4 with hydrazine hydrate.

Previously, the structure of 8 was assigned based on 1H and 13C NMR spectroscopic data . However, these data are insufficient to distinguish compound 8 and its isomer 9 resulting from a Dimroth rearrangement that is known to proceed in 3-substituted 4-iminopyrimidine systems . Our analysis of 1H, 13C NMR, and 2D NMR spectra (DMSO-d6 solution) of the prepared product confirmed its structure as compound 8. For example, the 1H,13C-HMBC spectrum showed correlation of the NH2 protons with carbon C-6 (through three bonds), and the 1H,1H-NOESY experiment revealed a diagnostic NOE between the NH2 and H-6 protons. The structure 8 was also confirmed by comparing the experimental carbon chemical shifts of the prepared compound in DMSO-d6 with shifts calculated for 8 and 9 by the GIAO method at the PBE1PBE/6-311+G(2d,p) level of theory using the DFT B3LYP/6-311++G(d,p) optimized geometries (DMSO solution) and applying a multi-standard approach . The calculated shifts of sp2-atoms C-7a, C-4, C-6, C-3, and C-3a in (Z)-8 and the s-cis-conformer (with respect to the C4–N bond) of 9 were 156.6, 152.1, 151,9, 128.3, 105.6 ppm and 160.9, 158.9, 157.4, 123.3, 98.4 ppm, respectively. The corresponding experimental shifts (155.0, 151.8, 149.9, 128.4, 105.4 ppm) were in good agreement with the structure 8. It is noteworthy that the DFT B3LYP/6-311++G(d,p) calculations using the PCM solvation model showed that (Z)-8 was significantly less stable than the s-cis-conformer of 9 in DMSO solution (ΔG = 7.17 kcal/mol; 298 K, 1 atm).

As we proposed, pyrazolopyrimidine 8 undergoes dimerization to produce macrocycle 5 under certain conditions (Scheme 5).

Scheme 5: Synthesis of macrocycle 5 by the dimerization of pyrazolopyrimidine 8.

The dimerization of 8 was thoroughly studied varying promoter, its amount, solvent, substrate concentration, and reaction time (Table 1).

Table 1: Synthesis of macrocycle 5 by the dimerization of pyrazolopyrimidine 8.a

Entry Promoter (equiv) Solvent Conc. of 8
(mmol/mL) Reaction time (h) Isolated productsb Mass yield of products (%) Molar ratio of productsc Estimated yield of 5 (%)d 1 N2H4·H2O (1.02) EtOH 0.39 2 5 + 6 + 8 18 56:7:37 13 2 N2H4·H2O (3.11) EtOH 0.46 2 5 + 6 35 80:20 28 3 N2H4·H2O (6.29) EtOH 0.34 0.5 5 + 6 21 64:36 13 4 N2H4·H2O (1.90) THF 0.33 2 5 + 6 + 8 60 3:4:93 3 5 N2H4·H2O (2.10) dioxane 0.20 2 5 + 6e 46 89:11 ≈41 6 N2H4·H2O (3.09) dioxane 0.33 2 5 + 6 41 84:16 34 7 N2H4·H2O (2.03) dioxane 1.00 2 5 + 6f 47 76:24 – 8 N2H4·H2O (1.44) dioxane 0.33 2 5 + 6 + 8e 43 54:9:37 ≈29 9 N2H4·H2O (3.07) pyridine 0.50 1.5 5 + 6 18 86:14 15 10 N2H4·H2O (2.20) pyridine 0.33 1.5 5 + 6 16 89:11 14 11 N2H4·H2O (1.99) MeCN 0.33 2 5 + 6e 34 56:44 ≈19 12 N2H4·H2O (6.19) iPrOH 0.31 1.5 5 + 6 15 58:42 9 13 N2H4·H2O (1.46) MeOH 0.50 3 5 + 6 41 89:11 36 14 N2H4·H2O (2.00) MeOH 0.50 3 5 + 6 40 89:11 36 15 N2H4·H2O (6.04) MeOH 0.34 1 5 + 6 20 82:18 17 16 N2H4·H2O (6.05) EtOH/H2O 12:1 (v/v) 1.03 1 5 + 6 21 81:19 17 17 N2H4·H2O (3.07) dioxane/H2O 12:1 (v/v) 1.03 1 5 + 6 21 82:18 17 18 N2H4 (2.21) MeOH 0.50 2 5 + 6 38 87:13 33 19 MeNHNH2 (2.63) EtOH 0.38 2 5 + 6f 22 91:9 – 20 N2H4·H2O (1.51) MeOH 0.48 3 5 + 6 43 90:10 39 21 N2H4·H2O (1.51) MeOH 0.50 3 5 + 6 46 89:11 41 22 N2H4·H2O (3.04)+ TsOH·H2O (0.05) EtOH 0.50 1 5 + 6 + 10 38 66:4:30 29 23 N2H4·H2O (2.93)+ TsOH·H2O (0.05) dioxane 0.51 1 5 + 6 + 10 87 13:0.5:86.5 19 24 TsOH·H2O (0.10) dioxane 0.50 1 8 98 – 0 25 TsOH·H2O (0.10) MeCN 0.24 2 5 + 8f 91 0.8:99.2 <1 26 TsOH·H2O (0.10) MeOH 0.38 5.5 5 4 – 4 27 TsOH·H2O (0.10) EtOH 0.37 2 5f – – 12g

aLoadings of 8 were 50–141 mg (0.30–0.86 mmol) in entries 1–20, 23, 26, 27, 1.016 g (6.19 mmol) in entry 21, 4.039 g (24.60 mmol) in entry 22, 0.337 g (2.05 mmol) in entry 24, and 0.315 g (1.92 mmol) in entry 25. All reactions were carried out under refluxing conditions; bmethods of the product isolation: (i) filtration of the precipitate formed (for entries 1, 2, 4–8, 10, 11, 13–21, 24–26), (ii) evaporation of the volatiles under vacuum followed by treatment with water and filtration of the formed precipitate (for entries 3, 9, 12, 22, 23), (iii) evaporation of the volatiles under vacuum (for entry 27); caccording to 1H NMR spectrum of the crude product; dcalculated based on overall mass yields and molar ratios of the products; eplus a small amount of unidentified impurities; fplus a significant amount of unidentified impurities; gthe yield was estimated by 1H NMR spectrum for a mixture of the crude product with a weighed amount of succinimide as a reference.

First, we studied the dimerization of 8 promoted by hydrazine hydrate in EtOH under reflux (Table 1, entries 1–3). We found that the starting material was completely consumed in the presence of 3 equivalents of N2H4·H2O within 2 h and the precipitated solid was isolated by filtration. According to the 1H NMR spectrum, this crude product was a mixture of macrocycle 5 and bis-pyrazole 6 in a molar ratio of 80:20 (Table 1, entry 2). An increase in the amount of N2H4·H2O to 6.3 equivalents led to a faster conversion of 8, however, the amount of bis-pyrazole 6 in the isolated mixture increased to 36% (Table 1, entry 3). In contrast, reducing the amount of N2H4·H2O to 1 equivalent resulted in an increase in the 5:6 ratio to 89:11 and a decrease in the conversion of 8 to 63% after 2 h of reflux (Table 1, entry 1).

Next, we tested other protic (iPrOH, MeOH, EtOH/H2O, 1,4-dioxane/H2O) and aprotic (THF, 1,4-dioxane, pyridine, MeCN) solvents for the dimerization of 8 promoted by hydrazine hydrate to improve both yield and purity of 5 (Table 1, entries 4‒17). As can be seen from Table 1 the solvent had a dramatic effect on the outcome of the reaction. With THF, pyridine, MeCN, iPrOH, EtOH/H2O, or 1,4-dioxane/H2O either low conversion of 8 (Table 1, entry 4), or poor product yield (entries 9, 10, 12, 16, and 17), or low purity of 5 (entries 11 and 12) were observed. Using 1,4-dioxane with 3 equivalents of N2H4·H2O (reflux, 2 h), a mixture of 5 and 6 in a molar ratio of 84:16 was obtained, and the calculated yield of 5 was 34% (Table 1, entry 6). The best result was achieved by the reaction of 8 with 1.5–2 equivalents of N2H4·H2O in MeOH under reflux for 3 h (5/6 = 89:11, macrocycle calculated yield of 36%) (entries 13 and 14 in Table 1).

The experimental data described above were obtained using 50–141 mg of pyrazolopyrimidine 8. We demonstrated that, under the optimized conditions (1.5 equiv of N2H4·H2O, MeOH, reflux, 3 h), the reaction can be scaled up to gram quantities without any loss of efficiency and even with a noticeable increase in the macrocycle yield up to 41% (Table 1, entries 20 and 21). The extremely poor solubility of product 5 in most organic solvents allowed to purify it from all admixtures, including byproduct 6, by a single crystallization from boiling DMF. Thus, pure product 5 was prepared on a multi-gram scale in a 35% isolated yield.

Several other promoters were also tested to dimerize pyrazolopyrimidine 8. In particular, the reaction proceeded in MeOH (reflux, 2 h) in the presence of anhydrous N2H4 (2.2 equiv) afforded product 5, but in somewhat lower yield and purity compared with N2H4·H2O (entry 18 in Table 1). The dimerization of 8 in refluxing EtOH promoted by methylhydrazine gave a 91:9 mixture of 5 and 6 along with a significant amount of unidentified side-products (Table 1, entry 19). In the presence of TsOH·H2O (0.1 equiv) in refluxing MeOH or EtOH, macrocycle 5 was formed in unacceptably low yields (Table 1, entries 26 and 27), however, a complete conversion of the starting material was observed. In contrast, in aprotic solvents (1,4-dioxane, MeCN) in the presence of TsOH·H2O (0.1 equiv), the starting material remained intact (Table 1, entries 24 and 25). The use of N2H4·H2O (2.9–3.0 equiv) with a catalytic amount of TsOH·H2O (0.05 equiv) in refluxing EtOH or 1,4-dioxane resulted in the formation of mixtures of macrocycle 5, pyrazolyl-1,2,4-triazole 10, and a very small amount of bis-pyrazole 6 according to NMR data (Table 1, entries 22 and 23). It is noteworthy that triazole 10 was the major product in dioxane (10/5/6 = 86.5:13:0.5, Table 1, entry 23). Again, macrocycle 5 was separated from this mixture