Biomedical Engineering Reference
In-Depth Information
[89,100,109] to monitor the concentration of volatile species. Most of the work, thus far, has
concentrated on anesthetics, but these and other studies indicate that many organic com-
pounds will directly alter BR's photophysical properties [91,99,109]. The complexity of
both the photochromic and photovoltaic effects, and the sensitivity of these signals to the
external environment, combine to make BR a strong candidate for the design of hybrid
semiconductor sensors.
14.3.2.2 Bacteriorhodopsin-Based Chemical Sensor Architecture
The ability to utilize BR as a chemical sensor depends on two assertions. The first is that
BR has proven to be sensitive to a remarkable range of chemical antigens, and that sensi-
tivity is manifested as measurable changes in the protein's photochromic and photoelec-
tric properties. Sensitivity to its chemical environment, combined with the researcher's
ability to integrate it into device architectures, its rugged stability, high cyclicity, and the
ability to interrogate the protein by multiple means, makes BR a strong candidate for sen-
sor architectures. However, it is doubtful that wild-type BR will be useful as a sensor ele-
ment on its own. Even though the wild-type protein can respond to a wide variety of
chemical antigens, it is unlikely that wild-type BR alone will be able to discriminate
between more than a few chemical species—it simply cannot elicit a wide-enough variety
of responses. Fortunately, this limitation can be easily overcome by turning to genetically
engineered variants for the detection scheme. The second assertion, therefore, is that the
utilization of BR mutants will introduce differential sensitivity and chemical antigen speci-
ficity to the sensor architecture. Directed evolution can be used to custom-design proteins
that will respond specifically and selectively to individual chemical agents. The ultimate
approach would be to employ several BR mutants, each highly optimized to detect the
presence of a unique chemical species. The problem with this approach is that it involves
several rounds of directed evolution, consisting of random mutagenesis and screening
thousands of mutant strains for desirable proteins. Utilization of advanced BR mutations
therefore represents a long-term approach to sensor development.
A far simpler approach than employing directed evolution is possible, however, by tak-
ing advantage of the differential sensitivities exhibited by a library of mutant BR proteins:
as long as the protein is functional, it will respond in some way to external chemical stim-
uli. Looking for an optimized response is somewhat secondary to the ability to character-
ize a preexisting response; in other words, as engineering a specific change in the BR
photocycle is no longer the goal, screening need be done only for functional proteins, not
for a specific quality. Chemical agents are certain to affect the photokinetic and photo-
voltaic responses of mutant BR proteins, and as long as their responses are uniquely dif-
ferent from the native protein, they will be successful in chemical agent detection schemes.
Although individually, neither the wild type nor any specific mutant protein is capable of
accurately determining the identity of an unknown chemical agent, a collection of such
mutants with varied responses to the same toxin will enable a far more accurate determi-
nation, thereby reducing the possibility of false negatives. The result is a numeric matrix
describing the responses of each mutant protein that acts as a fingerprint for each chemi-
cal agent. Furthermore, access to a library of BR mutants with characterized responses to
external chemical stimuli has definitive implications for highly specific detection schemes.
A sensor based on a collection of BR mutants has the potential of displaying both sensi-
tivity and selectivity. The approach to implementing such a sensor is to simultaneously
monitor the response of a large number of mutant proteins to a given chemical antigen. As
detailed above, each mutant will exhibit a specific response, different from each and every
other mutant. By examining the simultaneous responses of a large set of mutant proteins
to a given chemical antigen, we now have the possibility of generating a numerical
