Transcription mapping studies and DNA sequence analysis of the vaccinia virus HindIII D fragment predict that gene D8 encodes a protein 304 amino acids in length, with a molecular mass of 35,426 daltons, that is expressed at late times in infection. In order to determine whether the native D8 protein is required for virus propagation, we constructed a frameshift mutation in the D8 coding sequence. Virus containing this mutation were isolated and shown to replicate in a single-step growth experiment with wild type virus growth kinetics, demonstrating that the normal-length D8 protein is not essential for virus propagation in tissue culture. In order to investigate the synthesis of the wild-type and the mutant D8 proteins in virus-infected cells, we raised polyclonal antisera to a fusion protein consisting of a portion of the D8 coding sequence linked to the Escherichia coli trpE gene. Western blot (immunoblot) analysis of the time course of D8 protein synthesis in cells infected with either wild-type or mutant virus demonstrated that D8 protein was synthesized late in infection in each case and accumulated throughout the experiment. To determine whether the D8 protein was incorporated into the mutant or wild-type virus, purified virions were fractionated into Nonidet P-40-soluble, deoxycholate-soluble, and detergent-insoluble fractions. In both the wild-type and the mutant viruses, the D8 protein was an integral viral protein. The wild-type protein partitioned into the Nonidet P-40-soluble fraction, suggesting that it was a viral membrane protein. The mutant protein fractionated into the detergent-insoluble component, demonstrating that although the altered protein was incorporated into the virus, it was found in a abnormal location. In order to determine whether the D8 protein was present on the virion surface, the susceptibility of the D8 protein to proteolysis was tested by analyzing the products of incubation of the wild-type and mutant viruses with either chymotrypsin or trypsin. These studies demonstrated that the wild-type D8 protein was a transmembrane protein with a major extraviral domain that was released largely intact from the virus by trypsin. The mutant D8 protein was relatively refractory to proteolysis, confirming the hypothesis that although it is associated with the virus, it is in a conformation different from that of the wild-type protein. Tryptic digestion of the wild-type virus increased plaque formation severalfold, concomitant with the removal of the extraviral domain of the D8 protein.(ABSTRACT TRUNCATED AT 400 WORDS)