The mechanism of DNA transmission between distinct organisms has remained a subject of long-standing interest. Agrobacterium tumefaciens mediates the transfer of plant oncogenes in the form of a 25-kb T-DNA sector of a resident Ti plasmid. A growing body of evidence leading to the elucidation of the mechanism involved in T-DNA transfer comes from studies on the vir genes contained in six major operons that are required for the T-DNA transfer process. Recent comparative amino acid sequence studies of the products of these vir genes have revealed interesting similarities between Tra proteins of Escherichia coli F factor, which are involved in the biosynthesis and assembly of a conjugative pilus, and VirB proteins encoded by genes of the virB operon of A. tumefaciens pTiC58. We have previously identified VirB2 as a pilin-like protein with processing features similar to those of TraA of the F plasmid and have shown that VirB2 is required for the biosynthesis of pilin on a flagella-free Agrobacterium strain. In the present work, VirB2 is found to be processed and localized primarily to the cytoplasmic membrane in E. coli. Cleavage of VirB2 was predicted previously to occur between alanine and glutamine in the sequence -Pro-Ala-Ala-Ala-Glu-Ser-. This peptidase cleavage sequence was mutated by an amino acid substitution for one of the alanine residues (D for A at position 45 [A45D]), by deletion of the three adjacent alanines, and by a frameshift mutation 22 bp upstream of the predicted Ala-Glu cleavage site. With the exception of the frameshift mutation, the alanine mutations do not prevent VirB2 processing in E. coli, while in A. tumefaciens they result in VirB2 instability, since no holo- or processed protein is detectable. All of the above mutations abolish virulence. The frameshift mutation abolishes processing in both organisms. These results indicate that VirB2 is processed into a 7.2-kDa structural protein. The cleavage site in E. coli appears to differ from that predicted in A. tumefaciens. Yet, the cleavage sites are relatively close to each other since the final cleavage products are similar in size and are produced irrespective of the length of the amino-terminal portion of the holoprotein. As we observed previously, the similarity between the processing of VirB2 in A. tumefaciens and the processing of the propilin TraA of the F plasmid now extends to E. coli.