An Improved and Practical Method for Synthesizing of α-Sanshools and Spilanthol

An efficient and practical route for the synthesis of α-sanshools and spilanthol is described herein. Several modifications of an existing method enabled the preparation of the (2E,6Z,8E,10E)-tetraene precursor of hydroxy-α-sanshool in good yield. A highly selective Wittig reaction employing newly synthesized phosphonium salt with low deliquescence and long-term stability yielded the desired Z-form tetraene. This improved methodology was shown to be applicable to the efficient synthesis of α-sanshool and spilanthol.

The synthesis of 1 has been reported previously by two independent research groups. Igarashi and co-workers developed two stereoselective approaches to hydroxyl-α-sanshool synthesis, both employing several metal reagents and requiring precise operations Igarashi et al., 2012). Toy and co-workers constructed a (6Z,8E,10E)-conjugated triene precursor moiety with moderate selectivity (6Z:6E = 2:1) using the Wittig reaction; a pure stereoisomer was isolated by recrystallization (Wu et al., 2012). The purpose of the current study was to produce high-purity hydroxy-α-sanshool 1. Among the three existing synthesis methods, Toy's is the simplest due to the use of more conventional reagents and procedures. Our synthesis of 1 via Toy's method, however, proved difficult when following the literature, and resulted in reduced yields due to the instability or deliquescence of intermediate species. Therefore, we set out to enhance the general practicality and robustness of Toy's method of sanshool synthesis.

RESULTS AND DISCUSSION
Our synthesis of hydroxy-α-sanshool began with the oxidation of 4-bromobutan-1-ol with PCC, which was poorly reproducible on the gram scale. A more effective strategy was catalytic oxidation using commercially available AZADOL as the catalyst and sodium hypochlorite pentahydrate (NaClO·5H 2 O) as a co-oxidant (Scheme 1) (Okada et al., 2014). The desired 4-bromobutanal 2 was FIGURE 1 | Sanshool compounds. SCHEME 1 | Synthesis of ester 3. produced in 55% yield together with small amounts of 4-bromobutanoic acid. These results were reproducible even on the gram scale (Scheme 1, i). Other nitroxyl radical catalysts did not improve the yield of 2. Note that partial decomposition of 2 during purification resulting in moderate overall yields. Then, the Horner-Wadsworth-Emmons (HWE) reaction was conducted, resulting in ester 3 in 80% yield (Scheme 1, ii).
In an effort to improve the selectivity of the Wittig reaction (6Z:6E = 2:1), we converted ester 3 to its corresponding phosphonium salt 4a with PPh 3 according to Toy's synthesis method. However, this reaction suffered from low reproducibility due to the high deliquescence of 4a. We therefore evaluated several methods to create a phosphonium salt 4 with lower hygroscopicity (Scheme 2). First, ester 3 was hydrolyzed to carboxylic acid 5 and the phosphonium salt 4b was obtained in good yield by the reaction with PPh 3 . Unfortunately, 4b exhibited deliquescence similar to that of 4a. To determine the influence of the phosphonium salt counter anion on deliquescence, we prepared the iodonium salt 4c using the corresponding alkyl iodide 6. However, this also resulted in a compound with high deliquescence. We found that the combination of counter anion and functional group is important in determining the deliquescence of phosphonium salts, and obtained the nondeliquescent iodine salt 4d from the iodo ester 7.
The developed method was applied to the synthesis of the biologically active compound spilanthol (Sharma et al., 2011;Barbosa et al., 2016), also known as affinin, which contains a (2E,6Z,8E)-decatrienamide moiety (Scheme 4). Several synthetic methods for spilanthol have been reported (Crombie et al., 1963;Ikeda et al., 1984, Ikeda et al., 1987. A recent short step synthesis by Pastre provided high stereoselectivity, but suffered from a relatively low overall yield of 18% (Alonso et al., 2018). Our synthesis, starting from the Wittig reaction of the ylide generated from 4d and crotonaldehyde to afford ester 12, resulted in a 95% yield of the (2E,6Z,8E)-single stereoisomer. Saponification of 12 gave carboxylic acid 13 in 91% yield. Spilanthol was then synthesized in 84% yield using the coupling reaction employed in the α-sanshool synthesis. Thus, the efficient and stereoselective synthesis of spilanthol was achieved from 4-bromobutanol in six steps with an overall yield of 47%.

CONCLUSION
We developed a practical and reproducible method for the synthesis of hydroxy-α-sanshool and α-sanshool. Notably, modifications of the Wittig reaction using a newly synthesized, SCHEME 3 | Synthesis of hydroxy-α-sanshool 1 and α-sanshool 11. SCHEME 4 | Synthesis of spilanthol.
non-deliquescent phosphonium salt under low-temperature conditions succeeded in forming single stereoisomers of (2E,6Z,8E,10E)-tetraene and (2E,6Z,8E)-triene moieties in good yields. This method was shown to be applicable to the synthesis of spilanthol in six steps, resulting in an overall yield of 47%. Further studies on the synthesis of other sanshool derivatives are ongoing.

DATA AVAILABILITY STATEMENT
All datasets generated for this study are included in the article/Supplementary Material.

AUTHOR CONTRIBUTIONS
AN, KM, and KT performed the experiments. AN and TM wrote the manuscript. All authors designed the experiments and were involved in the data analysis. All authors designed the experiments, were involved in the data analysis, and have expressed approval of the final version of the manuscript.

FUNDING
This work was financially supported by JSPS KAKENHI Grant Nos. 19K16329 and 18K05132.