A Large Scale Formal Synthesis of CoQ10: Highly Stereoselective Friedel-Crafts Allylation Reaction of Tetramethoxytoluene with (E)-4-Chloro-2-methyl-1-phenylsulfonyl-2-butene in the Presence of Montmorillonite K-10

Abstract

Ubiquinone, as its name represents, exists ubiquitously in human body, particularly in the heart. It mediates the electron transfer process in mitochondria and also exerts strong antioxidant effect in its reduced form. In clinical trial, it showed beneficial effect on heart-related diseases such as myocardial infarction, angina, and other related symptoms to cause decreased mortality compared to the placebo group. These interesting biological effects have induced various synthetic efforts towards Coenzyme Q10 (CoQ10). Among them, two approaches seem practical for a large scale production in terms of the supply of the starting material, solanesol and their synthetic efficiency: for the construction of its complete carbon framework, Koo’s synthesis used the coupling of the sulfone 2 with solanesyl bromide (3) and Lipsultz group employed nickel catalyzed cross coupling reaction of 4 with vinyl aluminum reagent 5 as the key steps, respectively (Figure 1). Earlier in our synthetic study towards CoQ10, we investigated the reproducibility of the Koo’s synthetic procedure toward 2 obtained from the Friedel-Crafts reaction of (E)-4chloro-2-methyl-1-phenylsulfonyl-2-butene (7) with tetramethoxytoluene 6. Surprisingly, the reaction itself was pretty complicated with the formation of many side products in much lowered E:Z selectivity of ca. 3:1 to 7:1 than reported. Use of zinc chloride and bromide resulted in 3.3:1 and 7.4:1 E:Z ratio, respectively (Table 1, entries 1-2) with yields usually less than 20%. Best ratio of 7.6:1 was obtained by the use of boron trifluoride etherate, however with low conversion of ca. 20%. Aluminum chloride also showed marginal conversion of 21% with 3.4:1 E:Z selectivity. Details of the HPLC profiles of these reaction mixtures were provided in Figure 2. Other Lewis acid such as MgBr2, AlEt3, TiCl4, FeCl3 led to complete decomposition of starting materials and no sign of the formation of the desired product was observed by HPLC analysis. To improve the yield and stereoselectivity of the reaction, we turned our attention to the use of montmollironite K-10 (MK-10), in which the reaction partners might be confined in the layered space, thus rendering the stereoselectivity increased. First we tested commercial MK-10 for the reaction of 6 with 7. Happily, first attempt led to much cleaner reaction profile with significantly improved E:Z selectivity (Table 2, entry 1). However, the reaction did not complete in prolonged reaction time – the maximum conversion was accomplished in 18 h (74%). We speculated that the low conversion might be caused by residual water in clay (6.9% for commercial MK-10). In this respect, we tested a series of