Molecular genetic analyses of a 376-kDa Golgi complex (GC) membrane protein (giantin) are described. The immunoglobulin G fraction of a human serum containing antibodies against GC antigens as revealed by indirect immunofluorescence microscopy with Hep-2 cells was used to screen a HeLa cDNA expression library, yielding four overlapping cross-hybridizing clones. Additional cDNA clones were retrieved from a lambda gt11 human thyroid cDNA library or generated by reverse transcriptase-mediated PCR from HeLa cell mRNA. Alignment of the clones resulted in a consensus cDNA of 10,300 bp encoding a protein of 376 kDa. The corresponding mRNA with a size of about 10 kb was detected by Northern (RNA) blotting of HeLa, Hep-G2, and Jurkat cell RNA. Sequence analyses of the protein revealed an extraordinarily high content of heptad repeats with the probability of forming coiled coils similar to the proteins of the myosin family. Five overlapping recombinant proteins covering the entire sequence were synthesized and used for antibody production in rabbits and for affinity purification of human and rabbit antibodies. Indirect immunofluorescence experiments also done with brefeldin A-treated Hep-2 and Pt K1 cells revealed an identical GC staining of both the affinity-purified human and rabbit antibodies. Double labeling experiments with antibodies against the GC marker mannosidase II as well as immunoelectron microscopic studies confirmed the localization of the protein within the GC. A corresponding endogenous large-molecular-mass protein of about 390 kDa was found in [35S]methionine-labeled Hep-2 cell lysates as well as in GC-enriched subcellular fractions from rat liver. The protein as well as the recently described proteins golgin-95 and golgin-160 (M. J. Fritzler, J. C. Hamel, R. L. Ochs, and E. K. L. Chan, J. Exp. Med. 178:49-62, 1993) may belong to a new group of Golgi proteins with a high content of heptad repeats which may exert functions in scaffold formation or vesicle transport. As far as can be concluded from immunological and personally communicated partial cDNA sequence data, the protein seems to be identical with a 400-kDa Golgi protein (giantin) recently described (A. D. Linstedt and H. P. Hauri, Mol. Biol. Cell 4:679-693, 1993). Therefore, we agreed to adopt the name giantin.