Improved multianalyte detection of organophosphates and carbamates with disposable multielectrode biosensors using recombinant mutants of Drosophila acetylcholinesterase and artificial neural networks
Tt. Bachmann et al., Improved multianalyte detection of organophosphates and carbamates with disposable multielectrode biosensors using recombinant mutants of Drosophila acetylcholinesterase and artificial neural networks, BIOSENS BIO, 15(3-4), 2000, pp. 193-201
Engineered variants of Drosophila melanogaster acetylcholinesterase (AChE)
were used as biological receptors of AChE-multisensors for the simultaneous
detection and discrimination of binary mixtures of cholinesterase-inhibiti
ng insecticides. The system was based on a combination of amperometric mult
ielectrode biosensors with chemometric data analysis of sensor outputs usin
g artificial neural networks (ANN). The multisensors were fully manufacture
d by screen-printing, including enzyme immobilisation. Two types of multise
nsors were produced that consisted of four AChE variants each. The AChE mut
ants were selected in order to obtain high resolution, enhanced sensitivity
and minimal assay time. This task was successfully achieved using multisen
sor I equipped with wild-type Drosophila AChE and mutants Y408F, F368L, and
F368H. Each of the AChE variants was selected on the basis of displaying a
n individual sensitivity pattern towards the target analytes. For multisens
or II, the inclusion of F368W, which had an extremely diminished paraoxon s
ensitivity, increased the sensor's capacity even further. Multisensors I an
d II were both used for inhibition analysis of binary paraoxon and carbofur
an mixtures in a concentration range 0-5 mu g/l, followed by data analysis
using feed-forward ANN. The two analytes were determined with prediction er
rors of 0.4 mu g/l for paraoxon and 0.5 mu g/l for carbofuran. A complete b
iosensor assay and subsequent ANN evaluation was completed within 40 min. I
n addition, multisensor II was also investigated for analyte discrimination
in real water samples. Finally, the properties of the multisensors were co
nfirmed by simultaneous detection of binary organophosphate mixtures. Malao
xon and paraoxon in composite solutions of 0-5 mu g/l were discriminated wi
th predication errors of 0.9 and 1.6 mu g/l, respectively. (C) 2000 Elsevie
r Science S.A. All rights reserved.