|Home | About | Journals | Submit | Contact Us | Français|
2-(((3R,4R)-4-(Allyloxy)-1-benzylpyrrolidin-3-yl)methyl)-6-(2,5-dimethyl-1H-pyrrol-1-yl)-4-methylpyridine (2), a key intermediate for the preparation of novel neuronal nitric oxide synthase (nNOS) inhibitors, is synthesized using diisopropyl (R)-(+)-malate as the starting material. The key steps involve a Frater-Seebach diastereoselective alkylation and a fast intramolecular cyclization.
Selective inhibition of the neuronal isozyme of nitric oxide synthase (nNOS) has attracted significant interest as a novel strategy in developing therapeutics for the treatment of neurodegenerative diseases including Parkinson's disease, Alzheimer's disease, and Huntington's disease.1 In our continuous effort to design nNOS selective inhibitors, 2 we have developed a stereospecific pyrrolidine-based inhibitor (1, Figure 1), which showed great potency (Ki = 5 nM) and extremely high selectivity for nNOS over its closely related isoforms endothelial NOS (eNOS, 3800 fold) and inducible NOS (iNOS, 1200 fold).3 Recent animal tests demonstrated that 1 could lead to a remarkable reduction in neurological damage to rabbit fetuses under hypoxic conditions,4 making 1 a strong candidate as a new drug for the treatment of neurodegenerative diseases.
Despite this exciting discovery, current and future research related to 1 is impeded by its difficult synthesis. In particular, the chiral pyrrolidine fragment 2 (Figure 1), which was achieved by a seven-step procedure,3b-c suffered from major disadvantages e.g., expensive starting material, hard chromatographic purifications, and low overall yield (<2%).3b-c Moreover, the utilization of racemic starting materials requires extra chiral resolution step(s) using either HPLC or chiral auxiliaries,3c which dramatically reduce the yield and efficiency. Therefore, the development of an efficient route to 2 is a bottleneck to future investigations of inhibitor 1.
Herein, we report the development of a concise stereospecific synthesis of 2. Our initial plan was to use a disubstitution reaction on dimesylate 3 with benzylamine (Scheme 1).5 Dimesylated compound 3 could be derived from dialkyl malate (4) using a sequential allylation– reduction procedure. Stereospecific compound 4 could be achieved by the diastereoselective alkylation protocol developed by Frater et al6 and Seebach et al7 using dialkyl (R)-(+)-malate (5) and 2-(bromomethyl)-6-(2,5-dimethyl-1H-pyrrol-1-yl)-4-methylpyridine (6) as starting materials.
The synthesis of 6 began with 2-aminopyridine (7, Scheme 2). The amino functional group of 7 was protected using 2,5-hexanedione in the presence of p-toluenesulfonic acid (p-TsOH) to give 8 in high yields. The 2,5-dimethylpyrrole protecting group was selected for two reasons. First, this protecting group is known to be stable under a variety of reaction conditions and can be easily removed under mild conditions. 8 Second, the electron-donating property of the 2,5-dimethylpyrrole group increases the chelating ability of the pyridine nitrogen to the lithium ion, which favors regioselective deprotonation of the 2-methyl group on the pyridine ring.9 8 was treated with n-BuLi at 0 °C, and the resulting anion was quenched with chlorotrimethylsilane (TMSCl) at the same temperature to generate 9 exclusively.8a Finally, 9 was allowed to react with 1,2-dibromotetrafluoroethane in the presence of CsF to provide 6 in quantitative yields.10
Next, optimization of the conditions for the Frater-Seebach alkylation was investigated (Table 1). When using lithium diisopropylamide (LDA) as the base, we isolated only a trace amount of product using either 5a or 5b as the starting material (Table 1, entries 1 and 2). With lithium hexamethyldisilazide (LHMDS) as the base however, we could isolate products 4a and 4b in 23% and 56% yield, respectively, with high diastereoselectivity (Table 1, entries 3 and 4). We then improved the yield to 85% by changing the ratio between 5b and 6 (Table 1, entries 5 to 7).
Allylation of 4b via NaH and allylbromide yielded 10, which was reduced using LiAlH4 to generate diol 11 in excellent yields (Scheme 3). When 11 was submitted to a variety of mesylation conditions, however, the only products that could be detected were compounds 13 and 14, derived by intramolecular cyclizations from either the pyridinyl nitrogen atom (13)11 or the hydroxyl oxygen atom (14),12 respectively.
To avoid these intrinsic problems, a new synthetic route was designed around key intermediate dialdehyde 15 (Scheme 4), which can undergo a single-step reductive amination reaction to provide 2.13 We hoped that under reductive conditions, dialdehyde 15 could be generated from diisopropylester 10.
The results of the Dibal-H reduction of 10 are summarized in Table 2. When 3.5 equiv of Dibal-H were used at -78 °C for 2 h (Table 2, entry 1), three different products, aldehyde 16, alcohol 17, and semi-acetal 18, were isolated. 18 was the major product, but no dialdehyde 15 was detected. Next, fewer equiv of the reducing reagent were used. The data showed that either only aldehyde 16 (Table 2, entry 2), or 16 and 17 (Table 2, entries 3 and 4) were isolated from the reaction without any evidence of dialdehyde 15 formation. Additional reduction of aldehyde 16 using Dibal-H (1 equiv) yielded only alcohol 17, which, together with the previous Dibal-H reduction data, confirmed that dialdehyde 15 could not be generated by reduction of 10.
Even though dialdehyde 15 was not produced, we did successfully isolate aldehyde 16 in good yields after simple optimizations (Table 2, entry 4). We sought to prepare amine 20 from 16 in the hope that the additional amino group of 20 would compete with the aminopyridine nitrogen for cyclization, thus preventing the formation of 13 and yielding the desired compound 2.
As shown in Scheme 5, reductive amination of 16 with benzylamine in the presence of NaHB(OAc)3 provided amine 19 in excellent yields with complete retention of stereochemistry. Next, the isopropyl ester of 19 was reduced with LiAlH4 to generate primary alcohol 20 in good yields. We found that a one-pot procedure without purification of 19 improved the overall yield (83%).
Finally, compound 20 was treated with methylsulfonyl chloride (MsCl) in the presence of TEA. The intramolecular cyclization from the benzyl-protected amine is so fast that 2 was obtained in quantitative yields without formation of any other side products.
In summary, we developed an efficient and highly diastereoselective synthesis of the chiral pyrrolidine building block (2) for a novel nNOS inhibitor (1), employing as key steps a Frater-Seebach type alkylation and a fast intramolecular cyclization, which avoids the unwanted cyclization by the pyridine nitrogen. This method takes nine steps in total with an overall yield of 42%, which is >20-fold higher than previous strategies.3b,c The current method has also been utilized for gram-scale preparations of inhibitor 1.
We thank the National Institutes of Health (GM49725) for financial support of this research.