Abstract: Nonproteinogenic a,a-disubstituted-a-amino acids have attracted considerable attention because of their utility as conformational modifiers of biologically active peptides and as enzyme inhibitors. Typical methods for asymmetric synthesis of them involve chiral auxiliary-based enolate chemistry (ref. 1). However, the most straightforward strategy for the synthesis would involve direct asymmetric a-alkylation of parent a-amino acids in the absence of additional chiral sources such as chiral auxiliaries, chiral ligands, or even chiral catalysts. Since both L- and D-a-amino acids are readily commercially available, the synthetic route shown in Scheme 1 seems most attractive for the purpose. However, this process usually gives racemic a-alkylated products from either L- or D-a-amino acid because the enolate formation eliminates the chiral information of C(2) and an achiral enolate common to both L- and D-series is formed (A = B). If enolate intermediates enable to memorize the chiral information at C(2) of the starting materials, L- and D-a-amino acids would give optically active L- and D- (or D- and L-) a,a-disubsituted a-amino acids via intrinsically chiral enolate intermediates A and B, respectively. In this paper, we describe asymmetric a-alkylation of a-amino acid derivatives via the synthetic route