POTASSIUM-DEPENDENT MOLECULES 225
closer to the pore than S1 or S3; (4) intersubunit disulfi de bonds — formed after
residues in the S3 – S4 linker, within N - terminal S4 and near the C - terminal
end of the S5 helix, underwent site - directed mutagenesis to cysteine
residues — indicate that S4 lies in close proximity to the pore region; and
(5) patterns of accessibility in S1 – S3 and S4 helices indicate that their ends
all lie close to the extracellular surface under pore open conditions. Finally,
the reference 28 authors argue that more information is needed about the
resting state of voltage - dependent K + channels before defi nitive mechanisms
for reaching the activated state from the resting state can be elucidated.
Molecular modeling of voltage sensor – pore interactions were undertaken
by the reference 29 authors using many of the experimental constraints dis-
cussed in the preceding paragraphs. The results placed the S4 segment in the
groove formed at the interface between neighboring S5 helices from different
tetrameric domains. The researchers found that this was the only S4 position
that simultaneously satisfi ed all the starting constraints. Another theoretical
model for voltage gating has been developed in the laboratory of H. Robert
Guy.^36 These researchers described the general consensus concerning the
voltage - sensing mechanism for K + channels before the KvAP X - ray crystallo-
graphic structure (1ORQ, 1ORS) became available. First, it was believed that
the voltage - sensor (helices S1 – S4) spanned the membrane in all conforma-
tions (open, closed, or in transition). Second, accessibility studies suggested
that positively charged residues of S4 would be found in water - fi lled crevices
and that voltage dependency was due to these charges being moved a short
distance across some barrier. The distance that S4 would move was in
dispute.
The publication of the KvAP crystal structures of the full channel (1ORQ)
and the isolated voltage sensor domain (1ORS) and the model arising from
the publications (see discussion of references 18 and 25) did not agree with
the Guy and co - workers model. Because of similarities in sequence and con-
servation of key amino acid residues, the Guy research group did not believe
that the KvAP voltage sensor domain should have a very different structure
from that of the Shaker protein. Therefore, they believed that either (a) the
previous models were incorrect, allowing the crystal structure to yield a correct
interpretation, or (b) the crystal structure, especially of the full channel,
did not represent a native conformation of the protein. In the view of these
researchers, one of the main problems with the new interpretation based on
the KvAP crystal structure is the evidence that the voltage sensor paddle
(helices S3b and S4) moves through the lipid during the transition from the
resting state to the activated state. This movement appears to be inconsistent
with previously developed models and could possibly be an artifact arising
from a protein structure distorted from its native conformation. Possible dis-
tortions postulated for the full - length KvAP channel include: (1) In native
KvAP channels, antibody fragments (Fab) bind to the S3 – S4 loop only from
the extracellular side, whereas the S3 – S4 loop is in the cytoplasm (internal to
the cell) in 1ORQ; (2) in Shaker channels, the C - terminal ends of helices S1