The nucleotide sequence of LEU4, a yeast gene encoding alpha-isopropylmalate synthase, has been determined. An open reading frame of 1857 nucleotides specifies a protein of 619 residues whose calculated molecular weight (68,416) and amino acid composition agree well with earlier estimates based on protein data. The 5' flanking region contains three blocks of sequence potentially involved in the general control of amino acid biosynthesis. It also has six blocks of homology in common with the 5'

more »

... lanking regions of two other LEU structural genes (LEU1 and LEU2). One of these blocks coincides with a palindromic element that has previously been demonstrated to be important for the specific leucine control of LEU2 (Martinez-Arias, A., Yost, H. J., and Casadaban, M. J. (1984) Nature 307, 740-742). Determination of the 5' ends of the LEU4 transcript indicates the existence of four major and several minor potential transcription start sites. Two of the major sites are located downstream from the ATG at the beginning of the long open reading frame. Utilization of these sites would lead to mRNA that could be translated from an in-frame AUG located 90 nucleotides downstream from the first one. The protein thus generated would be 30 amino acid residues shorter than the larger one. This situation might account for the occurrence of two alpha-isopropylmalate synthase-related proteins observed both in cell-free extracts and in in vitro translation mixtures (Hampsey, D. M., Lewin, A. S., and Kohlhaw, G. B. (1983) Proc. Natl. Acad. Sci. U.S.A. 80, 1270-1274). The larger of these proteins was incorporated into the mitochondria while the smaller one was not. We conclude that selection of appropriate transcription and translation start sites might control the subcellular localization of the LEU4 gene product. This conclusion is discussed with respect to other examples in yeast of genes that encode two forms of the same protein.