A human adenocarcinoma-associated antigen (KSA) defined by the monoclonal antibody KS1/4 has become the focus of several site-directed strategies for tumor therapy. KSA, a 40,000 Da cell surface glycoprotein antigen, is found at a high density in all adenocarcinomas examined to date and in corresponding normal epithelial tissues. Here we describe the cloning and sequencing of overlapping complementary DNA clones which encode the entire KSA as expressed in UCLA-P3, a human lung adenocarcinoma cell line. We have deduced the 314-amino acid sequence and have compared it to the N-terminal amino acid sequence data of the affinity-purified antigen. The KSA is synthesized as a 314-residue-long preproprotein that is then processed to a 232-residue-long antigen. KSA appears to have a single transmembrane domain of 23 residues that separates the highly charged 26-residue cytoplasmic domain from the extracellular domain. The N-terminal region of the propeptide is rich in cysteines and contains three potential N-glycosylation sites. Computer-assisted analyses at both the DNA and protein levels have found no significant similarities of this protein to known sequences, but a GC-rich 5' terminus is evident. Northern blot analysis shows that transcription of KSA can be detected in RNA isolated from normal colon but not in RNA isolated from normal lung, prostate, or liver.