The twin-arginine translocase (Tat) system is used by many bacteria to translocate folded proteins across the cytoplasmic membrane. The TatA subunit is the predicted pore-forming subunit and has been shown to form a homo-oligomeric complex. Through accessibility experiments using the thiol-reactive reagents 4-acetamido-4'-maleimidylstilbene-2,2'-disulfonic acid and Nalpha-(3-maleimidylproprionyl)biocytin toward site-specific cysteine mutants in TatA, we show that the N-terminus of TatA is located in the cytoplasm rather than the previously assumed periplasm. We also confirm previous observations that the C-terminus has a dual topology. By treatment with the membrane uncoupler carbonyl cyanide-m-chlorophenyl hydrazone, we show that the topological state of the C-terminus is dependent on the membrane potential. These results suggest two architectures of TatA in the membrane: one with a single transmembrane helix and the other with two transmembrane helices. Molecular models of both topologies were used to develop and cartoon a homo-oligomeric complex as a channel with a diameter of approximately 50 A and suggest that the double transmembrane helix topology might be the building block for the translocation channel. Additionally, in vivo cross-linking experiments of Gly2Cys and Thr22Cys mutants showed that Gly2, at the beginning of transmembrane helix-1, is in close proximity with Gly2 of a neighboring TatA, as Cys2 cross-linked immediately upon the addition of copper phenanthroline. On the other hand, Cys22, at the other end of the transmembrane helix, took at least 10 min to cross-link, suggesting that a possible movement or reorientation is required to bring this residue into proximity with a neighboring TatA subunit.