Consistency between the proposed elevator-like mechanism and previous experimental findings. (A) In the proposed elevator mechanism based on the repeat-swapped model (cartoon helices), the transport domain (teal and orange) is predicted to move relative to the dimerization domain (dark blue and dark red). Dimerization by cross-linking residues K551 (spheres) and K562 (not in known structure or model) from the TM5–TM6 loop (dark green) does not inhibit anion transport, consistent with their predicted location. Residue G796 in TM12 is in the domain interface (yellow sphere). (B) Biotinylation demonstrates accessibility of I684C (magenta) and, weakly, of Q683C (green) in TM8. These residues are only visible in the predicted cytoplasmic pathway of the IF model of AE1 (right) and not in the OF structure (left). (C) F792 (yellow sticks) in the dimerization domain forms more hydrophobic contacts with TM1 and TM3 of the transport domain in the model (right) than in the structure (left). Residues with C atoms <4.3 Å from an atom of F792 are shown as sticks. Helices are shown as cartoon helices (except TM3, for clarity) and labeled in bold. (D) E508 (magenta) at the cytoplasmic end of TM4 form contacts with G381 (white) and H703 in the OF structure (left) that may be lost in the IF conformation (right) as the distance between TM4 and TM5 increases. (E) Several ionizable residues on the extracellular surface (sticks) are predicted to form salt bridges only in the IF (right) or OF conformation (left). Three of these putative interactions line the extracellular pathway, bridging the transport and dimerization domains. The papain cleavage site (Q630) and its interaction partner in TM5 (D546) are also shown (green labels); the cleavage sites Q550 and Q564 are in the unresolved loop between TM5 and TM6. The 40-residue-long loop connecting TM7 and TM8 is colored black.