Glycoprotein D (gD) is a structural component of the herpes simplex virus envelope which is essential for virus penetration. The function of this protein is highly dependent on its structure, and its structure is dependent on maintenance of three intact disulfide bonds. gD contains six cysteines in its ectodomain whose spacing is conserved among all its homologs in other alphaherpesviruses as well as Marek's disease virus. For other proteins, conservation of cysteine spacing correlates with conservation of disulfide bond structure. We have now solved the disulfide bond structure of gD-1 and gD-2 of herpes simplex virus types 1 and 2, respectively. Two approaches were used. First, we constructed 15 double-Cys mutants of gD-1, representing all possible disulfide pairs. In each case, codons for cysteines were changed to serine. We reasoned that if two cysteines normally form a disulfide bond, double mutations which eliminate one proper bond should be less harmful to gD structure than double mutations which eliminate two disulfide bonds. The mutated genes were cloned into a eucaryotic expression vector, and the proteins were expressed in transiently transfected cells. Three double mutations, Cys-1,5, Cys-2,6, and Cys-3,4 permitted gD-1 folding, processing, transport to the cell surface, and function in virus infection, whereas 12 other double mutations each produced a malfolded and nonfunctional protein. Thus, the three functional double-Cys mutants may represent the actual partners in disulfide bond linkages. The second approach was to define the actual disulfide bond structure of gD by biochemical means. Purified native gD-2 was cleaved by CNBr and proteases, and the peptides were separated by high-performance liquid chromatography. Disulfide-linked peptides were subjected to N-terminal amino acid sequencing. The results show that cysteine 1 (amino acid [aa] 66) is bonded to cysteine 5 (aa 189), cysteine 2 (aa 106) is bonded to cysteine 6 (aa 202), and cysteine 3 (aa 118) is bonded to cysteine 4 (aa 127). Thus, the biochemical analysis of gD-2 agrees with the genetic analysis of gD-1. A similar disulfide bond arrangement is postulated to exist in other gD homologs.