Synthesis of Dihydropyrano[3,2-c]pyrazoles via Double Bond Migration and Ring-Closing Metathesis

Three types of pyrazole-fused heterobicycles, i.e., 1,5-, 1,7-, and 2,5-dihydropyrano[3,2-c]pyrazoles, were synthesized from 4-allyloxy-1H-pyrazoles. A sequence of the Claisen rearrangement of 4-allyloxy-1H-pyrazoles, ruthenium-hydride-catalyzed double bond migration, O-allylation, and ring-closing metathesis was employed in this study.


Introduction
The synthesis of substituted or functionalized pyrazoles has been studied extensively thus far because they show or are expected to show important and diverse bioactivities [1,2]. Celecoxib, a non-steroidal anti-inflammatory drug (NSAID), is a representative pyrazole-containing compound, which acts through selective cyclooxygenase (COX)-2 inhibition. Whereas the late-stage construction of a pyrazole ring through some cycloadditions of already-substituted components is the basis for most syntheses of substituted pyrazoles [3,4], direct functionalization of pyrazoles has not been investigated satisfactorily to date. As investigations on it seem rare, we have been interested in and studied the direct functionalization of pyrazoles through coupling reactions of halogenated analogues derived from commercially available pyrazole [5][6][7][8]. In addition, pyrazole-fused heterocycles have recently been synthesized for reasons similar to those described above or because of characteristic activities not seen in monocyclic substituted pyrazoles [9]. Many pyrazole-fused heterocyclic compounds possess unique and important biological activities [10]. Some examples of pyrano [2,3-c]pyrazoles [11][12][13][14], pyrano [3,2-c]pyrazoles [15,16], and furo [3,2-c]pyrazoles [17,18] are presented in Figure 1.
Then, the C4-hydroxyl groups in 4-hydroxy-5-(1-propenyl)-1H-pyrazoles 5a and 5b were treated with aqueous NaOH followed by alkenyl halides in order to prepare the RCM substrates 6a and 6b. Conversion of 5c and 5d, which have a substituent, to 6c and 6d using the same condition took a long time with poor yields. So, alternative transformation of 5c and 5d to 6c and 6d was carried out using K 2 CO 3 in acetone under MW irradiation, respectively. The reactions proceeded smoothly and the chemical yields of 6c and 6d are presented in Scheme 1a. In a separate experiment, compound 2e, which already contains a small part of 5e as noted above, was transformed directly to 6e through treatment with K 2 CO 3 and allyl bromide in acetone under MW irradiation in 63% yield, since the yield from 2e to 5e was not satisfactory. The yield of the MW-aided transformation of 2e to 6e was improved to 85% by applying acetone-water (9:1) as the solvent system (Scheme 1b). 8b (entry 10). When the substrate had an R′ substituent, different results were obtained, as shown by the following entries. Substrates 6d and 6e did not react at rt (entries 13 and 15, respectively).  (15) a. 60% of starting material 6c was recovered. b. 50% of 6e was recovered. c. Undesired 11e (2%) was obtained during the recovery of 6e (21%). d. 11e was obtained (4%). e. A small amount of 6a was detected in the NMR spectrum and was inseparable from 7a.
The MW-aided reaction (140 °C) of 6c afforded RCM product 7c as a minor product (24%) and 9c (45%) with an exomethylene moiety as the major product (entry 12). The structure of 9c was determined through the heteronuclear single quantum coherence (HSQC) correlations between a carbon signal at δ 107.2 ppm and two proton signals at δ 4.78 and 4.96 ppm. Generally, endo-cyclic alkene is considered to be more stable than the corresponding exo-alkene. But in this case, 7c is thought to be less stable than exo-diene 9c due to the strain caused by 6-membered endo-diene structure in the thermodynamic condition.
However, the same MW conditions applied to substrate 6d did not result in 7d, but dimeric 10d formed through intermolecular metathesis in 30% yield (entry 14). Mass spectrometry (MS) revealed that compound 10d had an m/z of 632 (M + ), which corresponds to C42H42N4O2. The 1 H nuclear magnetic resonance (NMR) spectrum of 10d suggested the presence of a =CHCH3 moiety through the signals at δ 6.29 (q, J = 7.1 Hz) and 1.51 ppm (d, J = 7.1 Hz) in a 1:3 integral ratio and the lack of an exomethylene from the starting 6d. These data suggest that the intermolecular metathesis product   (15) a. 60% of starting material 6c was recovered. b. 50% of 6e was recovered. c. Undesired 11e (2%) was obtained during the recovery of 6e (21%). d. 11e was obtained (4%). e. A small amount of 6a was detected in the NMR spectrum and was inseparable from 7a.
RCM substrates 6 were treated with 5 mol% Grubbs second-generation catalyst (Grubbs 2nd ) in CH 2 Cl 2 . The results of the RCM reactions are summarized in Table 1. With substrate 6a, reaction at rt afforded the desired RCM product 7a within 30 min (entry 2). A shorter reaction time also led to 7a, but with an inseparable trace amount of 6a (entry 1). In contrast, extended reaction times led to reduced product yields (entries 3 and 4). The MW-aided reaction was also examined in an attempt to reduce the reaction time (entries 5-7). In these trials, only 7a was formed and double bond migration product 8a could not be detected [23]. Moreover, higher temperatures above 100 • C reduced the reaction yield (entry 7). The optimal reaction conditions in entries 2 and 5 for substrate 6a were applied to the RCM of 6b and gave similar results producing 7b (entries 8 and 9, respectively). The MW reaction of 6b at a higher temperature of 140 • C led to partial double bond migration to produce 8b (entry 10). When the substrate had an R substituent, different results were obtained, as shown by the following entries. Substrates 6d and 6e did not react at rt (entries 13 and 15, respectively).
The MW-aided reaction (140 • C) of 6c afforded RCM product 7c as a minor product (24%) and 9c (45%) with an exomethylene moiety as the major product (entry 12). The structure of 9c was determined through the heteronuclear single quantum coherence (HSQC) correlations between a carbon signal at δ 107.2 ppm and two proton signals at δ 4.78 and 4.96 ppm. Generally, endo-cyclic alkene is considered to be more stable than the corresponding exo-alkene. But in this case, 7c is thought to be less stable than exo-diene 9c due to the strain caused by 6-membered endo-diene structure in the thermodynamic condition.
However, the same MW conditions applied to substrate 6d did not result in 7d, but dimeric 10d formed through intermolecular metathesis in 30% yield (entry 14). Mass spectrometry (MS) revealed that compound 10d had an m/z of 632 (M + ), which corresponds to C 42 H 42 N 4 O 2 . The 1 H nuclear magnetic resonance (NMR) spectrum of 10d suggested the presence of a =CHCH 3 moiety through the signals at δ 6.29 (q, J = 7.1 Hz) and 1.51 ppm (d, J = 7.1 Hz) in a 1:3 integral ratio and the lack of an exomethylene from the starting 6d. These data suggest that the intermolecular metathesis product 10d formed by expelling an ethylene molecule [339 (6d) × 2 − 28 (CH 2 =CH 2 ) = 632 ((M + ) for 10d)]. The presence of a bulky R substituent may lead to serious repulsion in the transition state for RCM. When the substrate had a methoxycarbonyl group as R , the results were confusing. The MW reaction of 6e at 140 • C gave a complex mixture and only 7e was isolated in 15% yield (entry 18). The MW reactions of 6e at lower temperatures (80 and 100 • C) gave 10e in similar yields (29% and 30%, respectively) with 7e as a minor product (entries 16 and 17). In both of these entries, 11e, which is a metathesis product of 6e and the Grubbs catalyst, was also isolated as a minor product. The structure of 11e was confirmed through detailed NMR analysis and an M + peak at m/z 388.1785 (C 24 H 24 N 2 O 3 ) in the high-resolution MS (HRMS) spectrum. However, our attention was focused on increasing the yields of 7e and decreasing the yields of 10e by increasing the reaction temperature (entries [16][17][18]. Then, we hypothesized that 10e transforms into 7e; 10e may be the initial product at lower reaction temperatures. Therefore, the MW reaction of pure 10e with Grubbs 2nd at 140 • C was examined independently in an attempt to observe the formation of 7e as the major product in the reaction mixture.

Synthesis of 1,7-Dihydropyrano[3,2-c]pyrazoles
We attempted to expand this methodology to the syntheses of different types of pyrazole-fused heterobicycles, i.e., 1,7-dihydropyrano[3,2-c]pyrazoles (8) and furo [3,2-c] pyrazoles (17), as illustrated in Scheme 2. In order to realize this, 4-O-vinylation was required. First, the 4-hydroxyl group of 2a was treated with 1,2-dichloroethane to obtain a pyrazole with a 2-chloroethoxy group at C4, 12 Cl . However, dehydrochlorination of 12 Cl did not occur under basic conditions. Then, 2-bromoethylation of the 4-hydroxyl group was examined, aimed at improving the leaving ability. Desired 5-allyl-4-(2-bromoethyl)oxy-1H-1-tritylpyrazole (12a) was smoothly prepared through the MW-aided reaction of 2a. The examination of the dehydrobromination of 12a is summarized in Table 2. Whereas treatment of 12a with t-BuOK in toluene resulted in no reaction (entry 1), application of THF-MeOH (4:1) led to the desired dehydrobromination (entries 2-5). The MW reaction at 100 • C for 30 min afforded only double bond migration product (E/Z)-5-allyl-4-vinyloxy-1H-1-tritylpyrazole (13a) but in 14% yield (entry 2). Increasing the reaction time to 60 min resulted in an inseparable mixture of 13a and 5-(1-propenyl)-4-vinyloxy-1H-1-tritylpyrazole (14a) in 19% combined yield (entry 3). A higher temperature of 130 • C resulted in only 14a in 30% yield (entry 4). A similar MW reaction at 80 • C produced 13a in a similar yield (entry 5). In these trials (entries 2-5), the chemical yields of desired 13a and 14a were not satisfactory. Close inspection of entries 4 and 5 led us to isolate and elucidate the structures of side product 15 (28% yield), which should have formed via S N 2 attack by a methoxide on 12a, and 16 (17% yield) (see footnotes of Table 2). To improve the chemical yields, inhibition of the S N 2 attack on 12a by a nucleophile formed from the solvent under basic conditions was required. Hence, t-BuOH was applied instead of MeOH as a co-solvent. Although the MW reaction at 80 • C afforded only a trace amount of desired product 13a (entry 6), the same reaction at 130 • C afforded only 13a in 87% yield (entry 7). Inspired by the result in entry 4, the MW reaction was attempted at a higher temperature of 180 • C and afforded 14a selectively in 67% yield (entry 8). Treatment of the N-benzyl derivative 12b with t-BuOK at 130 • C resulted in only 14b (72%) (entry 9). Then, the dehydrobromination of 12b was examined at a lower temperature (entry 10), but resulted in an inseparable mixture of 12b and 14b.  The RCM of prepared substrates 13a, 14a, and 14b were examined. Treatment of 13a with Grubbs 2nd (5 mol%) at rt gave the desired product 8a in 95% yield. However, the corresponding reactions of 14a and 14b did not afford the desired products 17a and 17b, even with MW assistance. Further examinations of 14a with alternative catalysts, such as the Grubbs 1st , Hoveyda-Grubbs, and Schrock catalysts, also did not lead to 17a. Our synthesis of 17 will be continued in a future study.

Materials and Methods
Infrared (IR) spectra were obtained using a Perkin Elmer 1720X FT-IR spectrometer (Perkin Elmer, Wattham, MA, USA). HRMS was performed using a JEOL JMS-700 (2) mass spectrometer (JEOL, Tokyo, Japan). NMR spectra were recorded at 27 • C using Agilent 300, 400-MR-DD2, and 600-DD2 spectrometers in CDCl 3 using tetramethylsilane (TMS) as the internal standard. Liquid column chromatography was conducted using silica gel BW127ZH (Fuji Silysia Chemical Ltd., Tokyo, Japn). Analytical and preparative thin layer chromatography (TLC) analyses were performed using pre-coated Merck glass plates (silica gel 60 F 254 ), and the compounds were visualized by dipping the plates in an ethanol solution of phosphomolybdic acid followed by heating (Merk & Co., Inc., Darmstadt, Germany). MW-assisted reactions were carried out using a Biotage Initiator ® (Basel, Switzerland). Anhydrous CH 2 CH 2 was purchased from Wako Pure Chemical Industries (Osaka, Japan).