SEQUENCE OF THE CDNA AND 5'-FLANKING REGION FOR HUMAN ACID ALPHA-GLUCOSIDASE, DETECTION OF AN INTRON IN THE 5' UNTRANSLATED LEADER SEQUENCE, DEFINITION OF 18-BP POLYMORPHISMS, AND DIFFERENCES WITH PREVIOUS CDNA AND AMINO-ACID-SEQUENCES

Acid maltase or acid α-glucosidase (GAA) is a lysosomal enzyme that hydrolyzes glycogen to glucose and is deficient in glycogen storage disease type II. Previously, we isolated a partial cDNA (1.9 kb) for human GAA; we have now used this cDNA to isolate and determine sequence in longer cDNAs from four additional independent cDNA libraries. Primer extension studies indicated that the mRNA extended approximately 200 bp 5′ of the cDNA sequence obtained. Therefore, we isolated a genomic fragment containing 5′ cDNA sequences that overlapped the previous cDNA sequence and extended an additional 24 bp to an initiation codon within a Kozak consensus sequence. The sequence of the genomic clone revealed an intron–exon junction 32 bp 5′ to the ATG, indicating that the 5′ leader sequence was interrupted by an intron. The remaining 186 bp of 5′ untranslated sequence was identified approximately 3 kb upstream. The promoter region upstream from the start site of transcription was GC rich and contained areas of homology to Sp1 binding sites but no identifiable CAAT or TATA box. The combined data gave a nucleotide sequence of 2,856 bp for the coding region from the ATG to a stop codon, predicting a protein of 952 amino acids. The 3′ untranslated region contained 555 bp with a polyadenylation signal at 3,385 bp followed by 16 bp prior to a poly(A) tail. This sequence of the GAA coding region differs from that reported by Hoefsloot et al. (1988) in three areas that change a total of 42 amino acids. Direct determination of the amino acid sequence in one of these areas confirmed the nucleotide sequence reported here but also disagreed with the directly determined amino acid sequence reported by Hoefsloot et al. (1988). At two other areas, changes in base pairs predicted new restriction sites that were identified in cDNAs from several independent libraries. The amino acid changes in all three ares increased the homology to rabbit–human isomaltase. Therefore, we believe that our nucleotide sequence for GAA is more precise. We have also identified single base-pair polymorphisms at 18 sites for human GAA, some of which are not silent. © 1990, Mary Ann Liebert, Inc. All rights reserved.