Nucleotide sequence of the human liver glutathione S-transferase subunit 1 cDNA.

Multiple human liver glutathione S-transferases (GST, EC 2.5.1.18) with overlapping substrate specificities may be essential to their multiple roles in xenobiotic metabolism, drug biotransformation and protection against peroxidative damage. Multiple GSTs have been purified from human liver cytosol and 13 isoenzymes were resolved by chromatofocusing (Vander Jagt et al., 1985). They are composed of at least two classes of subunits, Ha ( M , 26000) and H, ( M , 27 500), according to their relative electrophoretic mobility to rat GSTs (Vander Jagt et a[., 1985; Tu et al., 1986). We have demonstrated earlier a definitive immunological crossreactivity between the human liver and various rat GSTs and nucleotide sequence homology between their respective cDNAs (Tu et al., 1986). Human liver GSTs lack a mobility class equivalent to the rat liver GST Y, subunits ( M , 28 000) yet Y, cDNA (pGTR262) hybridized to the H, subunit cDNA (pGTH1) as strongly as did the Y, cDNA (pGTR261) (Lai et al., 1984; Tu et al., 1984; 1986). Furthermore, the H, subunit 1 cDNA, pGTH1, selected rat Y, and Y, subunit mRNAs with almost equal efficiency in hybridselected translation in vitro (Tu et al., 1986). In this communication, we provide a molecular basis to these observations by determining the nucleotide sequence of the H, subunit 1 cDNA, pGTH 1. The cDNA sequence of pGTHl was determined by a combination of the chemical method (Maxam & Gilbert, 1980) and the chain-termination method after subcloning into the EcoRI site of the M13 mp18 vector (Sanger et al., 1977; Messing, 1983). The cDNA is 810 nucleotides long, containing 66 nucleotides in the 5' non-coding region, a 222 amino acid open reading frame ( M , 25629) and a complete (78bp) 3' noncoding sequence including the poly(A) addition signal AATAAA (Fig. 1). The coding region nucleotide sequence is approx. 80% identical base-for-base to the rat liver Y, (pGTR261) subunit sequence (79.6%) (Lai et al., 1984) and to the rat liver Y, (pgTB42) subunit sequence (79.9%) (Telakowski-Hopkins et al., 1985). The predicted amino acid sequence of this human H, subunit is as different from the rat Y, subunit as is from the Y, subunit of rat liver GSTs, containing 55 substitutions among the 222 residues (75.2%). The 3' non-coding region of pGTHl cDNA (78bp) is shorter than the corresponding region (1Olbp) of the Y, cDNA pGTR261 (Lai et al., 1984) and much shorter than that in the Y, cDNA pGTR262 (Tu et al., 1984) (255bp). They share short, dispersed sequence homology, however. For example, nucleotides 756-797 of pGTHl (Fig. 1) are approx. 63% identical to the pGTR261 sequence (nucleotides 734-775) base-for-base without any deletion or insertion and nucelotides 761-794 of pGTHl are approx. 56% identical to the pGTB42 sequence (Telakowski-Hopkins et al., 1985) (nucleotides 807-840) base-for-base without any deletion or insertion. We have made a comparison of the pGTHl cDNA sequences with all the GST cDNA sequences available in the literature from rat and corn origins. In addition to the homology with Y, and Y, subunit cDNA of rat liver GSTs the H, subunit cDNA showed a region of considerable homology with the anionic (pGTR187) and a basic Y, subunit (pGTR200) near the N-terminal region with 62% and 69% homology, respectively (Fig. 2). This is consistent