86

౪

4 Smell and Taste

types of signals for different compounds. The signals were detected by measuring an

electroantennogram (the electrical activity of an antennae); a program was used to reli-

ably learn the different signals from each antennae and correlate them to the presence

of a specific chemical with the four different elecroantennograms for it. After training,

the sensor was able to distinguish between eight different odors reliably. Background

molecules sometimes interfered with the identification, but not the detection of a com-

pound. Unfortunately, antennae, when removed from the insect, only last 60–90 min.

After that, another sensor needs to be built and trained.

A different approach using insect antennae is demonstrated with Drosophila flies

that can detect cancer by its smell [13]. Drosophila’s odor receptors use cAMP as a sig-

naling molecule. These flies are a model organism for genetic engineering; in this case,

flies were developed that have cAMP fluorescently marked. The Drosphila antennae has

different receptors in different regions that react to cancer cell odors differently. When

imaging all regions of the antennae, a difference map can be developed, which is highly

sensitive to the odor changes of people with cancer [13].

A rat taste neuron was used both to study taste cell responses as well as to develop a

sensor for sour taste [14]. The set-up is similar to a patch-clamp system (see Section 1.6),

but instead of rupturing the membrane to measure the transmembrane potential, the

potential change due to the change in ion concentrations is measured on the outside. The

sour taste neuron was grown on a silicon chip, in this case a light-addressable potentio-

metric sensor (LAPS) chip [14] (Figure 4.4). This system was stable for up to 30 minutes.

Different types of signals were measured; the signal had to be analyzed either in the

time or the frequency domain for optimal results, depending on the type of signal. Sour

signals between pH 2 and 4 could be detected. When several cells with all types of taste

receptors were used as the sensor, time histograms and interspike histograms combined

could be used to identify specific stimuli (e. g., sweetness) [15].

A similar system was used for the detection of smell [16]. A mixture of different rat

olfactory neurons was grown on a LAPS chip until maturity (3 days), and then their sig-

nals were recorded. A significant difference between inhibitory and stimulatory signals

could be detected. To make the sensor specific, the neurons were genetically engineered

to express a specific odorant receptor, ODR-10 [17]. The result was a sensor that was spe-

cific for diacetyl, the natural ligand of ODR-10. The sensor also exhibited different fir-

ing frequencies for different concentrations of ligand. The detection range was 0.1 mM

to 100 mM, but with limited reproducibility. To make each different cell separately ad-

dressable, each cell was immobilized onto a microelectrode of a microelectrode array

(MEA) chip via a DNA strand [18]. With improved sensitivity and reproducibility, this

system could result in a nanosensor that can detect more complex smells for some time.

All cells react to chemical stimuli (food, hormones, signal transduction molecules),

and thus can be used as chemical sensors. One example is a sensor based on endothelial

cells detecting the signal transduction molecule nitric oxide (NO) [19] (Figure 4.5). The

cells were immobilized onto graphene using the RGD peptide (RGD stands for arginine,

glycine, and aspartic acid and is a common part in cellular recognition) and kept alive