An improved route to coenzyme Q 10 (1) starting from commercially available coenzyme Q 1 is described. The key steps in this synthesis are the SeO 2 -mediated oxidation of the protected isoprenylhydroquinone 3 into the (E)-allyl alcohol 5 without the formation of undesired stereoisomer and the one-pot reductive elimination of the phenylsulfonyl and dibenzyl groups in 7 by using naphthalenyllithium.Introduction. -There is considerable interest in the chemical synthesis of coenzyme Q 10 (1) due to its unusual structure and biological significance [2] and commercial importance. Although various synthetic routes [3] to 1 have been developed to date, the synthetic strategy implying a terminally functionalized protected isoprenylhydroquinone as a key intermediate seems to be one of the most attractive approach. Our recently proposed route based on (2E)-4-[3,4-dimethoxy-2,5-bis(methoxymethoxy)-6methyl]-2-methylbut-2-en-1-ol allowed us to develop a convenient and highly stereoselective process for 1 starting from coenzyme Q 0 (= 2,3-dimethoxy-5-methyl-1,4-benzoquinone) [1]. However, in this process, one important issue became apparent: an incomplete H/Li exchange of the protected hydroquinone was observed 2 ). The overall yield of the thus prepared terminally functionalized protected isoprenylhydroquinone was compromised by this inefficient metalation procedure. The metalation step turned out to be a bottleneck, hampering the scale-up trials. Thus, the development of an efficient and highly regio-and stereoselective preparation of 1 is still in demand.The SeO 2 -mediated terminal oxidation of an isopropylidene moiety to an allyl alcohol is an important strategy towards the formation of an (E)-allyl alcohol under regio-1 ) For part 1, see [1]. 2 ) Only 65% of 1,4-bis(methoxymethoxy)-2,3-dimethoxy-5-methylbenzene was lithiated by BuLi in the presence of N,N,N',N'-tetramethylethane-1,2-diamine (TMEDA).