Biomedical Engineering Reference
In-Depth Information
less compelling evidence for its validity. In contrast, a conventional approach commonly
found in the literature decomposes the photosignal into as many exponential terms as pos-
sible. This is still curve fitting by trial and error, even though the arbitrariness becomes hid-
den in the “black box” of computer algorithm. The experimental data thus impose
no
constraints
on the individual parameters—the conventional approach enlists twice as many
free parameters as the exponential terms: one for the decay time constant and another for
the amplitude of each exponential term (e.g., see [50]). The conventional approach thus
guarantees an agreement between the model and the data under
any
circumstance. This pit-
fall was seldom appreciated in the literature of photoelectric effects. But the conventional
approach appeared to be problematic because it often generated inconsistencies in the same
membrane system under different conditions and in similar systems under similar condi-
tions in different laboratories (see discussion in [59]). The resulting discrepancy cannot be
explained without taking into account the subtle effect of the presence of a series capaci-
tance—chemical capacitance. Furthermore, since the time-integral of a single exponential
decay function is nonzero
0
exp(
, the conventional practice of exponential
decomposition is a direct violation of the zero time-integral condition (cf. Figure 15.10 in
Chapter 5 of [48]). Therefore, the individual exponential components cannot possibly rep-
resent distinct molecular relaxation processes.
While the conventional approach essentially decomposes the signal by curve fitting
with multiple exponential terms, we sought to decompose the signal by
physical means
. We
used three different reconstitution methods to dissect and decompose the experimentally
measured photosignal into several individual components. The decomposition is thus
independent of the curve fitting process and is therefore model-independent. The detailed
experimental evidence in support of our approach can be found in the literature [59-66].
If bR is incorporated by means of the BLM technique into a genuine lipid bilayer mem-
brane, all three AC components,
B
1
,
B
2
, and
B
2
t
/
)d
t
0
, as well as the DC component, can be
detected. If a method, which was originally developed by Trissl and Montal [10], is used
reconstitute bR, only the
B
1
and the
B
2
components can be detected. In the latter method,
oriented purple membrane sheets are deposited on a transparent Teflon film with the
extracellular surface attached right next to the Teflon film. The Teflon film prevents the
extracellular surface of the purple membrane sheets from a direct contact with water, and,
therefore, the
B
2
component, which reflects interfacial proton transfers at the extracellular
surface, is missing. The DC component of the photosignal is also missing because Teflon
is an excellent insulator. We shall refer to this method as
Trissl-Montal
(TM) method.
Okajima and Hong [59] modified the TM method by depositing multiple layers of ori-
ented purple membrane sheets on the Teflon film and by allowing the multilayered (ML)
purple membrane-Teflon film assembly to dry in air for about 4 days or longer. When such
a film assembly is mounted in a chamber with two compartments and rehydrated with
electrolyte solution, the photosignal exhibits a dramatically different waveform as com-
pared with the signal generated by means of the TM method. We shall refer to the latter
method of reconstitution as the ML thin film method.
Typical photoresponses from a TM film and from an ML film are shown in Figure 15.6. All
photoelectric signals have a biphasic waveform, exhibiting a positive peak and a negative
peak. The signals from a TM film are highly sensitive to both pH and temperature (Figure
15.6A and Figure15.6B). Lowering the pH (or the temperature) causes an increase of the pos-
itive peak but a concurrent decrease of negative peak. In contrast, the signals elicited from an
ML film exhibit no detectable pH sensitivity (Figure 15.6C) and are only slightly sensitive to
temperature (Figure 15.6D). When the photosignals elicited from an ML film are subject to
equivalent circuit analysis, the measured signals are in good agreement with the calculated
response based on the equivalent circuit, again using only experimentally measurable input
parameters for the computation. The agreement holds under various conditions including
