42 M. H. Akabas
tracellular end of M3. Surprisingly, in the presence of GABA, cysteines at multiple
positions along the extracellular two thirds of M3 were accessible for reaction.
This implied that GABA activation altered the protein conformation and allowed
water to penetrate into what was ostensibly the hydrophobic interior of the protein
(Williams and Akabas 1999 ). This revealed that gating changes the conformation
of the outer ring of helices. A subsequent study that examined the ability to form
disulfide bonds between pairs of cysteines inserted into the M2 and M3 segments
within a single subunit. This showed that different pairs of residues were able to
form disulfide bonds in the absence and presence of GABA (Jansen and Akabas
2006 ). Thus, channel activation involved conformational rearrangements within
individual subunits.
Interestingly, allosteric modulators including benzodiazepines and the general
anesthetic propofol also increased the accessibility of M3 segment cysteines, even
though they did not open the channel (Williams and Akabas 2000 , 2001 , 2002 ;
Rosen et al. 2007 ). It is notable that different subsets of M3 residues were accessible
in the presence of GABA, benzodiazepines and propofol. This demonstrates that
the allosteric modulators stabilize distinct conformational states of the receptor and
that these distinct states can be distinguished by cysteine accessibility experiments.
Changes in substituted cysteine reactivity also demonstrate that the environ-
ment surrounding the GABAA α1 M1 segment changes in the presence of GABA
(Bali and Akabas 2012 ). Strikingly, at some positions the reaction rate of the
negatively charged reagent, pCMBS, was several orders of magnitude higher than
at most of the other positions and these positions were different in the absence
and presence of GABA. Mutating an arginine (M2-19’) on the backside of the
M2 segment in the adjacent β2 subunit to glutamine dramatically reduced the pC-
MBS reaction rates at these α1 M1 positions. We inferred that the positive charge
was increasing the apparent reaction rate either by increasing the local pCMBS
concentration or by orienting the pCMBS. This confirmed the close proximity
between the positive charge of the M2 segment arginine and the M1 segment in
the adjacent subunit that was apparent in the crystal structures (Unwin 2005 ).
Interestingly, the M1 positions that showed this effect were different in the ab-
sence and presence of GABA. This suggested that the relative position of M2
and M1 in adjacent subunits changes during activation. Consistent with the M2
segment moving towards the M1 segment in the adjacent subunit, a disulfide
bond between β2R19’C and α1I228C only formed in the presence of GABA.
This suggests that during channel activation the extracellular end of the M2 seg-
ment moves away from the channel axis towards the M1 segment in the adja-
cent subunit (Bali and Akabas 2012 ). These results were consistent with previous
studies in the nACh and glycine receptors that fluorophores at the 19’ position,
attached either by cysteine modification or via unnatural amino acids, underwent
a change in environment during gating (Dahan et al. 2004 ; Pless et al. 2007 ; Pan-
toja et al. 2009 ). Thus, SCAM experiments combined with disulfide crosslinking
can provide experimental insight into the conformational changes that occur dur-
ing channel gating.