Engineered Ionizable Side Chains 21
receptor, for example, or even a phenomenological quantity such as an EC 50 value,
it is the concentration of the ligand in bulk solution—not the concentration of the
ligand in the vicinity of the ligand-binding atoms—that we use in the calculation.
Acidic Side Chains We also studied the effect of engineering single acidic side
chains, although we did this only in M2. We found that aspartates and glutamates
substituted at positions facing the lumen of the open-channel pore (as identified by
the effect of systematically engineered basic side chains; Figs. 5 and 8 ) have highly
up-shifted pKa values ( pKa, pore ≅ 8.5; pKa, bulk ≅ 4.0–4.4), in marked contrast to the
nearly bulk-like pKas of basic side chains at these positions; as an example, Fig. 3
shows the acid–base behavior of a glutamate substituted at position 13ʹ. In other
words, whereas the negative charge of a side-chain’s carboxylate is destabilized
inside the AChR’s aqueous pore with respect to bulk water, the positive charge of
basic side chains is nearly as stable inside the pore lumen as it is in bulk water. It
is tempting to speculate that the same properties of the AChR’s pore that render it
cation selective underlie this differential stabilization of positive versus negative
side-chain charges. As for the other, non-lumen-facing positions in M2, we found
that channels containing single acidic side chains substituted at these positions have
wild type-like conductances with no evidence of excursions of the open-channel
signal to a current superlevel, even at pH 9.0; this is consistent with these side
chains being permanently protonated, and thus, with their pKa values being even
more up-shifted than they are when engineered on the lumen-facing stripe of M2.
As a result, when engineered in the AChR, acidic side chains turned out to be much
less informative than basic side chains.
Practical Aspects One of the appealing facets of the method we describe here is
its great simplicity. As far as equipment and skills are concerned, all that is needed
is a single-channel patch-clamp recording setup, a cell-culture room, basic molec-
ular-biology instruments and lots of patience to obtain long and stable recordings
containing single-channel bursts of activity. As far as software is concerned, any
program that allows the user to digitize current recordings, idealize single-channel
data and estimate the occupancies of the different current levels would do a good
job. If, however, estimates of the protonation and deprotonation rates were desired,
then a program that allows the user to estimate the rates of a kinetic model from
single-channel traces would be necessary. In our experience, we find that QuB (Qin
2004 ; Qin et al. 1996 ), which implements an approximate solution to the missed-
event problem, covers all the software requirements of the method.
A Word of Caution Although protonation–deprotonation events of single side
chains may certainly manifest as discrete interconversions between different
open-channel current levels, there are other reasons why these fluctuations may
occur. Before ascribing them to an acid–base reaction, it should be remembered
that proton-transfer events are expected to depend on pH in such a way that pro-
tonation becomes slower and deprotonation becomes faster as the pH increases.
Moreover, both the binding and unbinding rates of hydrogen ions are expected