FET-Based Biosensors for The Direct Detection of ... - Semantic Scholar
1896
Full Paper
FET-Based Biosensors for The Direct Detection of Organophosphate Neurotoxins A. L. Simonian,*a A. W. Flounders,b J. R. Wildc a
Auburn University, Materials Research and Education Center, Auburn, AL, USA *e-mail: [email protected] b University of California, Microfabrication Laboratory, Berkeley, CA, USA c Biochemistry and Biophysics Department, Texas A&M University, College Station, TX, USA Received: April 15, 2004 Final version: May 10, 2004 Abstract Recent world-wide terrorist events associated with the threat of hazardous chemical agent proliferation, and outbreaks of chemical contamination in the food supply has demonstrated an urgent need for sensors that can directly detect the presence of dangerous chemical toxins. Such sensors must enable real-time detection and accurate identification of different classes of pesticides (e.g., carbamates and organophosphates) but must especially discriminate between widely used organophosphate (OP) pesticides and G- and V-type organophosphate chemical warfare nerve agents. Present field analytic sensors are bulky with limited specificity, require specially-trained personnel, and, in some cases, depend upon lengthy analysis time and specialized facilities. Most bioanalytical based systems are biomimetic. These sensors utilize sensitive enzyme recognition elements that are the in-vivo target of the neurotoxic agents which the sensor is attempting to detect. The strategy is well founded; if you want to detect cholinesterase toxins use cholinesterase receptors. However, this approach has multiple limitations. Cholinesterase receptors are sensitive to a wide range of non-related compounds and require lengthy incubation time. Cholinesterase sensors are inherently inhibition mode and therefore require baseline testing followed by sample exposure, retest and comparison to baseline. Finally, due to the irreversible nature of enzyme-ligand interactions, inhibition-mode sensors cannot be reused without regeneration of enzyme activity, which in many cases is inefficient and time-consuming. In 1996, we pioneered a new “kinetic” approach for the direct detection of OP neurotoxins based on agent hydrolysis by the enzyme organophosphate hydrolase (OPH; EC 3.1.8.2; phosphotriesterase) and further identified a novel multienzyme strategy for discrimination between different classes of neurotoxins. The major advantage of this sensor strategy is it allows direct and continuous measurement of OP agents using a reversible biorecognition element. We also investigated incorporation of enzymes with variations in substrate specificity (e.g., native OPH, site-directed mutants of OPH, and OPAA (EC 3.1.8.1), based upon preferential hydrolysis of PO, PF and PS bonds to enable discrimination among chemically diverse OP compounds. Organophosphate hydrolase enzymes were integrated with several different transduction platforms including conventional pH electrodes, fluoride ion-sensitive electrodes, and pH-responsive fluorescent dyes. Detection limit for most systems was in the low ppm concentration range. This article reviews our integration of organophosphate hydrolase enzymes with pH sensitive field effect transistors (FETs) for OP detection. Keywords: Organophosphate hydrolase, Organophosphorus acid anhydrolase, pHFET, ENFET, Biosensor, Diisopropyl fluorophosphate, Paraoxon, Demeton-S, Detection, Chemical warfare agents
1. Introduction The detection and identification of newly-introduced, organophosphate (OP) neurotoxins in air, water and soil poses an extremely difficult challenge. Soil and water samples are very likely to contain organophosphate pesticide residues due to heavy rural and urban use of these compounds. Yet, it is critical to detect military and terrorist activities which may result in environmental contamination with different chemical warfare (CW) agents. Thus, it is essential that CW neurotoxins can be readily and unequivocally distinguished from chemically similar agricultural compounds; ubiquitous organophosphate pesticides must never be interpreted as false positive indication of CW agents. Figure 1 compares the chemical structure of five Electroanalysis 2004, 16, No. 22
organophosphate compounds; all are toxic cholinesterase inhibitors but they vary in their neurotoxicity from organism to organism [1]. Paraoxon, diisopropyl fluorophosphate (DFP), and demeton-S are commercially available organophosphate pesticides. Paraoxon is representative of pesticides with neurotoxic phosphotriester bonds (PO bond); demeton-S is representative of phosphonothioate pesticides with a PS bond, and diisopropyl fluorophosphate is a phosphonofluoridate containing a PF bond. The structural similarity between demeton-S and DFP, and the different types of CW agents, have rendered them to be appropriate analogues for the development of detection and destruction technologies for these chemical agents. One of the main approaches used in the development of neurotoxin biosensors involves the inhibition of a specific
H 2004 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
DOI: 10.1002/elan.200403078
1897
Detection of Organophosphate Neurotoxins
Fig. 1. Chemical structures of some organophosphate neurotoxins.
enzyme in the presence of target analytes. The operational principle of inhibition-mode OP biosensors is based on the fact that the OP agents irreversibly bind at the enzyme active site, resulting in a loss of enzyme activity and hence a decrease in sensor signal. Several biosensors based on acetylcholinesterase (AChE) or butyryl cholinesterase (BChE) inhibition have been developed and used for OP neurotoxin detection [2 – 5]. A number of other enzymes such as urease and glucose oxidase have also been used as the target receptor in biosensors for their unique inhibitors [6, 7]. Inhibition-mode based sensors suffer from a number of limitations, such as: (i) any environmental or handling factors that cause loss of enzyme activity will result in false positive signals [8 – 10]; (ii) they requires baseline testing prior to sample application and lengthy sample incubation times to allow enzyme-analyte interaction; (iii) due to the irreversible nature of enzyme inhibition, inhibition-mode sensors cannot be reused without regeneration of enzyme activity, which in many cases is very inefficient. Despite several attempts to improve the specificity of cholinesterase sensors by monitoring of inhibition with both butyryl and acetyl cholinesterase [11] or analyzing inhibition patterns obtained with several acetyl cholinesterase enzymes from diverse sources [12, 13, 14], the inhibition-based sensors still have multiple unresolved problems, especially the inability to discriminate between neurotoxic agents and other unrelated enzyme inhibiting sample components such as heavy metals and pH variation. One of the ways to overcome the disadvantages of inhibition-based biosensors is to replace the inhibition receptor enzyme with a catalytic receptor enzyme. In this case, rapid, first-order enzyme kinetics may be utilized for real time or near-real time analysis. In 1996, we pioneered the application of the catalytic receptor approach to the direct detection of OP neurotoxins by utilizing the enzyme organophosphate hydrolase [15] for OP detection. Later, we developed a novel multi-enzyme strategy (using both organophosphate hydrolase and acetyl cholinesterase) for discrimination among different classes of neurotoxins [9]. Our approach was based on organophosphorus hydrolase (OPH; E.C. 3.1.8.1), a well-characterized metalloenzyme Electroanalysis 2004, 16, No. 22
originally isolated from Pseudomonas diminuta [16]. OPH is capable of cleaving PO, PF, PS, and PCN bonds via an SN2-type mechanism [17 – 19], resulting in unique sets of hydrolysis products, which also change the solution pH according to the reaction: Direct neurotoxin detection is thus possible via measurement of the pH change associated with enzyme activity [15]. OPH exhibits the unique ability to hydrolyze a large variety of organophosphate pesticides and neurotoxins including coumaphos, parathion, acephate, sarin, and VX [17]. OPH has been used to develop new biosensors for the direct detection of organophosphates based on pH monitoring of OPH activity with conventional pH electrodes [15, 20], or pH sensitive fluorescent dyes [21]. This strategy can be extended to further discriminate between different OP neurotoxins by monitoring additional organophosphate hydrolysis products. For example, hydrolysis of phosphofluoridates yields changes in pF as well as pH that can be detected with a fluoride specific ion-selective electrode [22]. These approaches have been subsequently validated by other research groups using the same enzyme (OPH or phosphotriesterase) who have verified the sensitivity and selectivity with different enzyme formats, such as whole cells or crude cell extracts [23, 24]. A separate class of hydrolytic enzymes, organophosphorus acid anhydrolase, (OPAA; EC 3.1.8.2.), has also been utilized for the direct detection of organophosphate compounds. The first OPAA discovered was isolated from a halophilic bacterial isolate designated JD6.5 and demonstrated high levels of DFP hydrolysis but lacked PO and PS bond hydrolysis [25]. The bacterial isolate was identified as a species of the genus Altermonas and a similar DFPhydrolyzing activity has been observed in several bacterial species [26]. As with OPH, the native function of OPAA is not yet known. However, OPAA has been unambiguously identified as a prolidase (E. C. 3.4.13.9) capable of hydrolyzing dipeptides with a prolyl residue in the carboxylterminal position [27]. A significant biosensor performance criterion is its analyte specificity. Enzymes with broad substrate specificity are suitable for detection of multiple analytes belonging to the same chemical class, while enzymes with high specificity for a single substrate are best for discriminative detection of a single target analyte. Initial investigations of OPAA substrate selectivity [25, 26] indicated that the enzyme is capable of cleaving PF bonds, while PO or PS bond hydrolysis is minimal. Since organophosphate pesticides, such as paraoxon and demeton-S, typically possess PO and PS bond structures (Fig. 1), such unique substrate prefer-
Scheme 1. H 2004 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
1898 ence makes OPAA very attractive for discriminative detection of fluorine containing organophosphates. Extending this strategy, a combination of OPH and OPAA in the same system may enable development of a multi-enzyme senor array capable of discriminating among PO, PS, and PF containing organophosphate compounds. Highly accurate and sensitive measurement of the pH changes associated with analyte hydrolysis is required for the development of such an enzyme-based biosensor system. A pH sensitive field effect transistor (pHFET) is the ideal device for monitoring these pH changes since the enzyme can be directly immobilized to the pH sensitive surface optimizing detection of enzyme hydrolysis products. PHFETS also provide a platform which lends itself to development of a multi-enzyme sensor array since multiple pH sensors can be integrated on a single sensor platform and different enzymes immobilized to each pH sensor. There is a significant literature related to pH-sensitive FETs (e.g., [28, 29]). Early investigations were frustrated by electrically leaky silicon dioxide and silicon nitride and physically leaky device encapsulation. More recent reports and the commercial availability of FET-based pH sensors indicate that many of these material issues have been resolved. A comprehensive overview of pHFET theory and operation written by the inventor of the device is available [27] and an interesting historical review of the evolution of pH sensitive FETs has been presented [30]. Enzyme modification of pHFETS (ENFETS) was first reported in 1980 but was hampered by unreliable pHFETS. A better understanding of FET device response led to creative and enhanced dynamic signal measurement techniques [31 – 33]. With more stable and reliable devices available, efforts have focused on resolving long recognized enzyme-modified FET issues: influence of solution buffer capacity [34], reference electrode miniaturization [35], and enzyme loading and stability. Enzyme-modified FETs have been proposed and investigated for many diverse analytes. The acetylcholinesterase-modified FET for organophosphate detection was a direct extension of cholinesterasemodified electrodes [36] and was proposed [37] soon after the initial enzyme-modified FET reports [38]. The investigation of cholinesterase-modified FETs has continued for several years [39 – 42] and the same strategy (inhibition mode/potentiometric detection) has been pursued with a commercially available, light addressable potentiometric sensor [43]. We previously reported the first use of the organophosphate hydrolase modified field effect transistor for organophosphate direct detection [44, 45] This present paper summarizes our efforts in the development of multi-enzyme biosensors based on pH-sensitive FET platforms and several OP-hydrolase enzymes (e.g., native OPH, genetically engineered OPH (site directed mutants), and OPAA).
Electroanalysis 2004, 16, No. 22
A. L. Simonian et al.
2. Experimental 2.1. Reagents and Buffers Paraoxon (diethyl-p-nitrophenyl phosphate), diisopropyl fluorophosphate (DFP), glutaraldehyde (grade 1), and reagents for buffer (NH4Cl, NaCl, MnCl2) were from Sigma (USA); demeton-S (O,O-dimethyl-S-2-ethylthiolethyl phosphorothioate), was from Chem Service (West Chester, Pa, USA). All solutions were prepared using 18 MW · cm ultrapure water (Milli-Q Plus, Millipore, ST).
2.2. Enzyme and Immobilization 2.2.1. Enzymes Wild-type OPH (E.C. 3.1.8.1) was isolated from a recombinant Escherichia coli strain using published procedures [19]. Various modified OPH enzymes with enhanced PF and PS bond cleaving capabilities have been developed and similarly purified and stored as lypholyzed powder [17]. OPAA (EC 3.1.8.2.) from Alteromonas sp. strain JD6.5 was isolated and purified to homogeneity [25] at the U.S. Army Edgewood Chemical Biological Center and obtained as a lyophilized enzyme powder (0.31 mg protein/mg powder). Since OPAA activity is enhanced with ammonia based buffers with low concentrations of MnCl2 [26], OPAA solutions were prepared in 50 mM NH4Cl, pH 8.5, with 0.1 mM MnCl2, 100 mM NaCl. 2.2.2. Immobilization on pHFET Chips Enzymes were immobilized to packaged pHFETchips following the aqueous aminopropyl-triethoxysilane (APTS): glutaraldehyde covalent attachment strategy described previously [46]. The APTS:glutaraldehyde was the preferred immobilization chemistry since it yielded higher total enzyme activity in bead and glass slide control studies. Sensor chips were treated with 1 N HCl, 15 min followed by 30% H2O2, 30 min for surface cleaning and silanol activation rather than the sulfuric acid, hydrogen peroxide activation used previously [46] to limit encapsulant degradation. Organosilane and glutaraldehyde treatments were performed by immersing packaged chips in solution. Enzyme was applied to only one device gate by pipetting enzyme into only one epoxy defined well. After extensive rinsing, chips were stored at 4 8C in 10 mM phosphate buffered saline (pH 7.4). Some packaged chips were reused by stripping enzyme in an oxygen plasma reactor (March Instruments, PX1000) at 50 W, 300 mtorr, 100 sccm O2, 5 minutes). Devices were retested for pH response after oxygen plasma stripping and then recoated with enzyme by repeating all immobilization chemistry treatments.
H 2004 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
1899
Detection of Organophosphate Neurotoxins
2.2.3. Sol-Gel Modification and Enzyme Immobilization To increase the surface area available for enzyme immobilization and provide a more stable enzyme microenvironment, silica microspheres were formed from tetraethoxysilane (Gelest, Inc., Tullytown, PA) via a base catalyzed solgel process then applied to diced chips by dip coating as described previously [46]. Dip coating was performed with diced chips prior to packaging, wire bonding and epoxy encapsulation. Sol-gel sphere diameters were determined with a sub-micron particle analyzer (Coulter, Model N4 MD) and verified with electron microscopy. Spheres were deposited over the entire chip surface. Spheres were scraped with tweezers from metal bond pads prior to wire bonding; spheres not in the gate region were completely covered with epoxy encapsulation. Sol-gel and non-sol-gel coated chips underwent identical enzyme immobilization and paraoxon testing procedures. 2.2.4. Testing Procedure All micro-chips were soaked in deionized water for at least 24 hours prior to testing. Devices were tested either (a) in batch mode by mounting a sealing ring over the encapsulated chip package which created a solution well (Fig. 2a) of approximately 1.2 mL volume; or (b) in a flow system consisting of a peristaltic pump and a flow cell which mounted over the PGA package providing a head space of approximately 75 mL (Fig. 2, b, c). A silver:silver chloride standard reference electrode (Orion Sure-Flow, Beverly, MA Orion, USA) was immersed in the test solution (batch mode) or a silver:silver chloride micro flow-through reference electrode (Microelectrode Inc., Bedford, USA) was introduced in the waste outlet stream (flow-through mode). Measurements were taken only with flow stopped to avoid streaming potential interferences. Device pH response was tested by monitoring changes in the common source voltage (Vcs, Fig. 3), with pH 4, 7, and 10 standard buffers (Sigma, St Louis, MO) before and after enzyme immobilization, sol-gel coating, and oxygen plasma treatment. Though pH 4.0 and 10.0 are not extreme conditions for enzyme exposure, chips with immobilized enzyme that were used for pH testing were not used for paraoxon testing unless fresh enzyme was immobilized to the chips. Chips with immobilized enzyme were tested for organophosphate sensitivity in 1) glycine-NaOH buffer (0.5, 1, or 10 mM with 0, 50, 100 mM NaCl) at pH 9.0 for OPH, or 2) in 50 mM NH4 Cl, pH 8.5, with 0.1 mM MnCl2, 100 mM NaCl for OPAA. Varying volumes of paraoxon were added, diluted in glycine buffer at pH 9.0 or ammonia buffer at pH 8.5, and changes in differential output voltage (Vdiff, Fig. 3) were monitored. The rate of hydrolysis of paraoxon in the absence of OPH was measured spectrophotometrically. This non-enzymatic hydrolysis was non-zero, but it was determined to be noncritical relative to the time scale of all sensor measurements. The independent measurement of pH after addition of paraoxon to the sensor solution was performed by removing Electroanalysis 2004, 16, No. 22
Fig. 2. a) Packaged pHFET chip with elastomer ring for bachmode measurements; b) dual pHFET chip with epoxy encapsulant; and, c) electronic control board with flow-through assembly.
a sample from the sensor solution and testing with the same pH meter used for all buffer testing and adjustments (Orion 601, Beverly, MA).
2.3. Apparatus and Procedures Sensor chips were prototype pH-sensitive field effect transistors (SenDx Medical, Inc., Carlsbad, CA). Each chip contained two discrete depletion modes, n-channel transistors with a non-metallized gate insulator stack of thermal silicon dioxide and chemical vapor deposited (CVD) silicon nitride (Figure 2, a). A porous ceramic matrix was formed at the gate surface by dip coating the chip prior to packaging into a solution of 200nm silica microspheres formed via a sol gel process as described above [46]. The chips were diced, packaged, and installed in a differential control circuit; Figure 3 shows a schematic of the chip control circuit. The control circuit was configured as a classical differential pair amplifier and used to monitor the gate to source voltage of the enzyme-modified device relative to the nonenzyme-modified device. The main advantage of the differential pair is that common mode variations such as temperH 2004 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
1900
A. L. Simonian et al.
Fig. 3. Schematic of the differential sensor circuit. Enzyme is immobilized to the gate of one pHFET while second pHFET serves as a non-enzyme coated reference. Bulk pH changes are measured as changes in the common source voltage Vcs and enzyme specific responses are measured as changes in Vdiff.
ature and bulk pH changes are eliminated while local pH changes at the FET gate with immobilized enzyme are amplified. Each experimental point was the average of 3 – 5 measurements.
ence of 1 mM paraoxon are : Vmaxapp ¼ 2.54 0.30 mV/s, and K app m ¼ 0.55 0.23 mM and in the presence of 1 mM demeton-S are 2.63 0.30 mV/s, and K app m ¼ 0.57 0.23 mM.
3.2. Influence of Buffer Capacity on the Sensor Response
3. Results and Discussion 3.1. Enzyme Kinetic Characteristics Table 1 presents kinetic characteristics of WT OPH, mutant RL-OPH, and OPAA. The comparison of OPH and OPAA kinetic parameters emphasizes the unique recognition capability of OPH and OPAA. OPAA displays a strong preference for PF bond hydrolysis making it suitable for discriminative detection of fluorine containing organophosphates, while OPH displays significant activity for PO, PF, and PS bonds hydrolysis making it suitable for detection of multiple organophosphates. PS bond hydrolysis by genetically modified OPH (OPH-RL) increased significantly (40 ) relative to wild type OPH) making possible demethon-S detection. Apparent kinetic parameters for soluble OPAA with DFP as a substrate, obtained by pH electrode, are: Vmaxapp ¼ 2.70 0.28 millivolts per second (mV/s), and K app m ¼ 0.54 0.08 mM. The same parameters in the presElectroanalysis 2004, 16, No. 22
The influence of the buffer capacity of the solution on device response is a well documented limitation of EnFETs. It is
Table 1. Comparative kinetic constants for OPH and OPAA enzymes Substrate/hydrolyzed bond
Enzyme
kcat (s1)
KM (mM )
DFP [ PF] [a]
OPAA OPH OPHRL OPAA OPH RL OPAA OPH OPH-RL
770 3500 8.2 4.64 11500 1400 0.028 2.6 68.0
2.81 1.42 1.2 0.6 0.12 0.036 3.5 3.0 2.5
Paraoxon [ PO] [b] Demeton-S [ PS]
[a] detected by fluoride-specific electrode; [b] detected by pH sensitive electrode.
H 2004 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
Detection of Organophosphate Neurotoxins
necessary to minimize buffer capacitance to optimize pH signal available from enzymatic hydrolysis of the analyte. However, this demand must be balanced by the conflicting need to provide some solution pH control to minimize background pH drift caused by sample addition and to maintain the pH required for enzyme activity. Figure 4 presents changes in batch mode sensor response for 1 mM paraoxon in solutions with different buffer strength. High salt concentration helps stabilize buffer, however, in deionized water solution salts only are not helpful. A buffer strength of 1 mM was selected as a compromise between satisfactory signal and the ability of an open system (subject to CO2 absorption especially in batch mode) to maintain constant pH. It should be noted that the independent pH measurements performed on removed samples indicated no measurable change in bulk pH associated with analyte hydrolysis. All buffer strengths tested were sufficient to compensate the acidic product generated by the immobilized enzyme once it diffused into the bulk. It was suspected that the sol-gel coatings may either act as a membrane and diminish the effect of buffer strength [35] or act as an additional solid state buffer and further diminish device response. However, no difference was noted between devices with and without sol-gel coatings with respect to buffer strength.
3.3. Sensor Response on Paraoxon with Native OPH Enzyme Sensor response to paraoxon concentrations from 10 to 50 mM (batch) and from 0.5 to 50 mM (flow) is presented in Figure 5. The “batch data” represents the averaged re-
Fig. 4. Effect of buffer concentration on sensor response. Batch mode measurements. Glycine-NaOH buffer with 120 mM NaCL and 2.5 mM KCl, pH 9.0. Initial 0 point – deionized H2O with salts, no glycine-NaOH. Electroanalysis 2004, 16, No. 22
1901 sponse of several different chips and the “flow data” represents the averaged response of multiple tests with the same chip. With the batch system, reproducible measurements were possible for paraoxon concentrations as low as 5 mM, while the flow system enabled reproducible measurements to 0.5 – 1 mM paraoxon. The two systems were set up at different laboratories and used by different experimenters, and the agreement between the two systems was excellent. To demonstrate the time scale of response, the sensor signals during repeated injections of paraoxon to the flow cell are presented in Figure 6. A comparison of the signal from sensors with and without sol-gel coatings to 10 – 100 mM paraoxon is presented in Figure 7. The overall device response was enhanced with the 200 nm sol-gel coating. Since the sol-gel coatings did not affect pH sensitivity, the measured signal increase is most likely due to an increase in enzyme loading or immobilized enzyme specific activity. This result correlated well with our earlier studies demonstrating that the total activity of OPH is enhanced by immobilizing enzyme to porous rather than non-porous particles [46]. Sensor response with a 20 nm solgel modification was the same as that of uncoated sensors. Assuming very close packed spheres, the pore size provided by the 20 nm sol-gel coating is one third of particle diameter or approximately 67 A. Unit cell dimensions of single crystal OPH enzyme from X-ray diffraction are 80.3, 93.4, and 44.8 A, [47]. This may explain why the 20 nm coating had no effect on sensor response, as the enzyme might have been unable to penetrate a silica microsphere matrix with an intersphere pore size equivalent to enzyme dimensions. The much larger pores provided by the 200 nm sol-gel coating would not limit enzyme penetration, and an increase in immobilized enzyme loading is expected. The 200 nm solgel coated sensors still displayed widely varying chip-to-chip response (CV ¼ 25 – 30%). In summary, the sol-gel coatings may have increased sensor response by increasing immobilized enzyme total activity; however, on a single chip surface,
Fig. 5. Non sol-gel coated sensor response to paraoxon. All tests conducted in 1 mM glycine-NaOH buffer pH 9.0. Inset shows flow system response to concentrations below 10 mM paraoxon. Flow system, N ¼ 5; batch system, N ¼ 6. H 2004 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
1902
A. L. Simonian et al.
were stored in PBS at 4 8C between measurements. The 200 nm sol-gel modified devices maintained greater than 80% of initial activity for the entire 6-week test period while the uncoated devices gradually lost more than 70% activity during the same period. This result supported our prior work [44], which demonstrated that porous immobilization platforms were superior to non-porous immobilization platforms at maintaining immobilized enzyme activity. A similar increase in enzyme stability with a preference for specific silica microsphere diameter has been reported [48]. All devices lost all activity after day 75. This is attributed to failure to include any preservative in the storage buffer.
3.4. Sensor Responses on Demethon-S with Genetically Modified Enzymes (RL-OPH
Fig. 6. Non sol-gel coated sensor response to paraoxon. All tests conducted in 1 mM glycine-NaOH buffer pH 9.0. Different concentrations of paraoxon injected at each dashed arrow, glycine buffer injected at each solid arrow.
As shown in Table 1, both wild type OPH and OPAA have low catalytic activity toward demeton-S. WT OPH is able to hydrolyze PS bonds only slowly and requires high concentration of analyte. Whereas immobilization of WT-OPH on a pHFET gate provides enough catalytic activity to hydrolyze paraoxon or DFP and produce measurable pH changes. It was not possible to obtain measurable response for demeton-S at concentrations up to 15 mM. To overcome this problem, the wild type OPH was genetically modified to increase catalytic activity toward PS bonds. This enzyme (OPH-RL, H254R/H257L) (and several others) was active enough to produce measurable pH changes in response to demeton-S exposure. (Figure 9). This enzyme displays a dramatic enhancement of the rate of hydrolysis of demeton-S and a significant activity decrease of hydrolysis of DFP (Table 1). This enhancement of the rate of demeton-S hydrolysis (kcat increased from 2.6 to 68.0 turnovers per second) was adequate to register positive biosensor responses.
3.5. Sensor Responses to DFP with OPAA
Fig. 7. Comparison of sensors with and without sol-gel coatings. All tests conducted in 1 mM glycine-NaOH buffer pH 9.0. N ¼ 4.
the sol-gel coating does not increase the surface area to volume ratio to such an extent that planar variations in enzyme loading are removed. With such a small area for direct immobilization, highly controllable and reproducible enzyme loading and activity continues to be a concern and the most significant source of sensor variability. The sol-gel coatings did not increase enzyme density to the point that dramatic changes in bulk pH were recorded. Devices with and without sol-gel coatings were tested with paraoxon at 1 – 2 week intervals over a 6-week period (Fig. 8). Devices Electroanalysis 2004, 16, No. 22
When OPAA was used as a biorecognition element, the sensor response to varying substrates was completely different, and detection time depended on the concentration of DFP. Figure 10 presents the responses of the OPAAmodified pH-sensitive FETs as a function of substrate concentration for different OP compounds. As expected, based on OPAA kinetic characteristics (see Table 1), the differences between sensor response to DFP and other substrates were quite large. Paraoxon generated barely measurable responses, even at high concentrations and there was no detectable signal for demeton-S. Linear, firstorder kinetics were observed for DFP concentrations from 12.5 to 500 mM. Similar high substrate specificity for DFP was observed also in the mix solutions contained 0.5 mM of each neurotoxin, paraoxon and demeton-S (Fig. 10).
H 2004 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
1903
Detection of Organophosphate Neurotoxins
Fig. 8. Sensor signal versus time since enzyme immobilization. All tests conducted in 1 mM glycine-NaOH buffer pH 9.0. Sensor response to 10, 50 and 100 mM paraoxon normalized with respect to sensor response on first day after immobilization. N ¼ 3.
however, the sensor response was not as sensitive to changes in pH as the soluble enzyme. [26] Satisfactory signal output was obtained from pH 7.5 – 9.0.
3.7. Long-Term Stability
Fig. 9. Sensor response on demeton-S. RL OPH, 1 mM glycineNaOH buffer pH 9.0.
OPAA immobilized on the sol-gel modified pH-sensitive chips lost 60% of its activity in 22 days, and there was no response after 30 days. This was especially surprising since previous results demonstrated that the silica sphere modification of the pH sensitive chips enabled the maintenance of a stable sensor response for more than 2 months [40]. Storage stability of OPAA immobilized on suspended silica gel particles was investigated separately [45]. In these studies, immobilized OPAA retained about 60% of the starting activity after five months. Such dramatic differences in long-term stability may be due to the significant differences in enzyme capacity of the two sensor formats; the silica gel has a much greater surface area than the pH-sensitive chip surface.
3.6. pH Dependence
3.8. Conceptual Design for Integrated Multi-Enzyme and Multi-Sensor Array
The previously reported pH optimum for OPAA activity (k app cat Þ with DFP as a substrate was 8.5, while the highest app catalytic efficiency (k app cat /K m Þ was observed at pH 6.8 [26]. However, enzyme immobilization can influence enzyme operational capability and change the pH optimum, depending on which sites on the enzymes were involved in binding to a surface. In order to determine the pH optimum for the OPAA sensor, the response to a fixed concentration of DFP (1mM) at pH 6.2 – 9.0 was measured (Fig. 11). Maximum sensor response was obtained at pH 8.5,
Measurable variations in enzyme-substrate specificity can be achieved with site directed mutagenesis and even subtle activity changes can be reproducibly detected with enzyme in solution. However, large variations in enzyme loading and enzyme activity after immobilization mean that only significant changes in enzyme activity can be detected and exploited for development of a multi-enzyme sensor array. Incorporation of enzyme variants from diverse species (OPAA vs OPH) provides far more discriminative capability than just incorporation of genetically modified
Electroanalysis 2004, 16, No. 22
H 2004 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
1904
A. L. Simonian et al.
Fig. 10. Signal versus concentration for DFP (black squares) paraoxon (inverted triangles), and mixture of 0.5mM of paraoxon, 0.5 mM of demeton-S and DFP (triangles): a) stop-flow assay with pH-sensitive field effect transistor, full curves; b) the linear ranges of the calibration plots, f(x)DFP ¼ 43.5 þ 1469x, R ¼ 0.99976, f(x)mix ¼ 42.5 þ 1255x, R ¼ 0.98891, f(x)paraoxon ¼ 13.9 þ 77.1x, R ¼ 0.85367.
Fig. 12. Anticipated response of integrated multi-enzyme and multisensor array. Proposed digital signal summarized in Table below. OPH* ¼ OPH genetic variant H254R/H257L abbreviated as OPH-RL.
Fig. 11. pH profile for DFP biosensor response. Stop-flow mode with pH-sensitive field effect transistor. Measurement condition: 1 mM NH4Cl buffer, containing 0.1 mM MnCl2 and 100 mM NaCl. DFP concentration 1 mM.
enzymes from the same source (OPH vs RL-OPH). Utilization of both strategies – genetically modified enzymes and multiple sources of enzymes provides the broadest discrimination. The appropriate combination of recognition elements and sensor platforms enables a set of almost digital responses, which are immune to the level of variation in enzyme loading and activity that is to be expected during production of an immobilized enzyme biosensor (Figure 12). Electroanalysis 2004, 16, No. 22
4. Conclusions Our investigations demonstrate the ability of the enzymes OPH, RL-OPH-RL, and OPAA to serve as highly discriminative biorecognition elements in biosensors for specific and direct detection of organophosphate neurotoxins. The immobilization of these enzymes to the exposed gate of a pH-sensitive field effect transistor produced a group of organophosphate sensors with rapid response to micromolar concentrations of paraoxon and DFP. The use of a specifically designed, genetically modified enzyme, RLOPH-RL, extends sensor discrimination and allows detection of demeton-S but with a higher detection limit. Combinations of organophosphate hydrolyzing enzymes, such as OPH, RL-OPH, and OPAA enable the construction of a multienzyme biosensor array capable of discriminating H 2004 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
Detection of Organophosphate Neurotoxins
among several related organophosphate compounds. Samples containing mixtures of both a G-type neurotoxin, such as sarin, and soman and organophosphate pesticide residues could be readily discriminatedidentified by this biosensor if combined with a front end sample separation system (e.g., HPLC). Though the sensors does do not provide the nanomolar organophosphate detection limits exhibited by acetylcholinesterase-modified FETs, OPH/OPAA-modified FETs do not require a continuous supply of consumable enzyme substrate, have a much faster response time, can be reused, are far less likely to generate false positive signal, and can discriminated between various neurotoxic pesticides and chemical warfare agents. Many of the early material problems of pH FETs have been resolved. Enzyme-modified FETs can now be re-evaluated as a suitable sensor platform in those cases where the detection needs correspond to the limits imposed by enzyme catalytic parameters and system buffer capacity. Due to its inherent miniaturization, low power consumption, and ease of integration, the FET is still an extremely appealing transduction platform. The sol-gel gate modifications investigated appear to provide a gate surface more suitable for enzyme immobilization. This sol-gel surface modification is not a mechanism for enhancing pH sensitivity or diminishing the dependence of the enzyme modified FET on the buffer capacity of the surrounding electrolyte.
5. Acknowledgements Support from Sandia National Laboratories and DOE Office of Nonproliferation Research and Engineering, USA-TATRC DREAMS, DARPA/ARO-MURI, NSFMURI, and Auburn University Detection and Food Safety Center is gratefully acknowledged.
6. References [1] D. P. Dumas, H. D. Durst, W. G. Landis, F. M. Raushel, J. R. Wild, Arch. Biochem. Biophys. 1990, 277, 155. [2] G. A. Evtugyn, H. C. Budnikov E. B. Nikolskaya, Analyst 1996, 121, 1911. [3] N. Mionetto, J.-L. Marty, I. Karube, Biosens. Bioelectron. 1994, 9, 463. [4] G. Palleschi, M. Bernabei, C. Cremisini, M. Mascini. Sens. Actuators B 1992, 7, 513. [5] J. J. Kulys, E. J. DQCosta. Biosens. Bioelectron. 1991, 6, 109. [6] A. Herbert, L. Guilhermino, H. C. Da Silva De Assis, P.-D. Hasen, Z. Angew. Zoologie 1995/96, 3, 1. [7] M. Brufani, M. Marta, M. Pomponi, Eur. J. Biochem. 1986, 157, 115. [8] A. L. Simonian, E. I. Rainina, J. R Wild, Anal. Lett. 1997, 30, 2453. [9] A. Herbert, L. Guilhermino, H. C. Da Silva De Assis, P.-D. Hasen, Z. Angew. Zoologie, 1995/96, 3, 1. [10] J. F. Payne, A. Mathieu, W. Melvin, L. L. Fancey, Marine Pollution Bulletin, 81. [11] T. T. Bachmann, R. D. Schmid, Anal. Chim. Acta 1999, 401, 95.
Electroanalysis 2004, 16, No. 22
1905 [12] O. Schafer, L. Weil, R. Niessner, Methoden Vom Wasser 1994, 82, 233. [13] T. T. Bachmann, R. D. Schmid, Anal. Chim. Acta 1999, 401, 95. [14] O. Schafer, L. Weil, R. Niessner, Methoden Vom Wasser 1994, 82, 233. [15] E. Rainina, A. Simonian, A. Efremenko, S.Varfolomeyev, J. Wild, Biosens. Bioelectron. 1996, 11, 991. [16] C. S. McDaniel, L. L. Harper, J. R. Wild, J. Bacteriol. 1988, 170, 2306. [17] D. P. Dumas, J. R. Wild, F. M. Raushel, Biotech. Appl. Biochem. 1989, 11, 235. [18] D. P. Dumas, H. D. Durst, W. G. Landis, F. M. Raushel, J. R. Wild, Arch. Biochem. Biophys. 1990, 277, 155. [19] K. Lai, N. .J. Stolowich, J. R. Wild, Arch. Biochem. Biophys. 1995, 318, 59. [20] A. L. Simonian, E. I. Rainina, J. R. Wild, in Methods in Biotechnology: Enzyme Biosensors – Techniques and Protocols (Eds: A. Mulchandani, K. Rogers), Humana Press, N. J. 1998, 17, pp. 237 – 248. [21] R. Russell, M. Pishko, A. Simonian, J. Wild, Anal. Chem. 1999, 71, 4909. [22] A. L. Simonian, B. D. diSioudi, J. R. Wild, Anal. Chim. Acta 1999, 389, 189. [23] P. Mulchandani, W. Chen, A. Mulchandani, J. Wang, L. Chen, Biosens. Bioelectron. 2001, 16, 433. [24] K. Roger, Y. Wang, A. Mulchandani, P. Mulchandani, W. Chen, Biotech. Prog. 1999, 15, 517. [25] J. J. DeFrank, T.-C. Cheng, J. Bacteriology 1991, 173, 1938. [26] J. J. DeFranik, W. T. Beaudry, T.-C. Cheng, S. P. Harvey, A. N. Stroup, L. L. Szafraniec, Chem.-Biol Interactions 1993, 87, 141. [27] T.-C. Cheng, J. J. DeFrank, V. K. Rastogi, Chim.-Biol. Interactions 1999, 119 – 120, 462. [28] Solid State Chemical Sensors (Eds: J. Janata, R. J. Huber), Academic Press, Orlando, FL 1985. [29] P. Bergveld, A. Sibbald, Analytical and Biomedical Applications of Ion Sensitive Field Effect Transistors Elsevier, Amsterdam 1988. [30] J. Janata, Analyst 1994, 119, 2275. [31] B. H. Van Der Schoot, P. Bergveld, Anal. Chim. Acta 1987, 199,160. [32] B. H. Van Der Schoot, P. Bergveld, Anal. Chim. Acta 1990, 233, 49. [33] A. Poghossian, A. Baade, H. Emons, M. J. Schçning, Sens. Actuators B 2001, 76, 634. [34] A. A. ShulQga, V. I. Strikha, A. P. Soldatkin, A. V. ElQskaya, H. Maupas, C. Martelet, P. Clechet, Anal. Chim. Acta 1993, 278, 233. [35] F. Lisdat, W. Moritz, Sens. Act. B 1993, 15 – 16, 228. [36] P. Strop, G. G. Guilbault, Anal. Chim. Acta 1972, 62, 425. [37] J. Janata, R. J. Huber, R. Cohen, E. S. Kolesar Jr., Aviat. Space Environ. Med. 1981, 52, 666. [38] S. Caras, J. Janata, Anal. Chem.1980, 52, 1935. [39] C. Dumschat,H. Muller, K. Stein, G. Schwedt, Anal. Chim. Acta 1991, 252, 7. [40] Y. Vlasov, A. Bratov, S. Levichev, Y. Tarantov, Sens. Actuators B 1991, 4, 283. [41] A. M. Nyamsi- Hendji, N. Jaffrezic-Renault, C. Martelet, P. Clechet, A. A. ShulQga, V. I. Strikha, L. I. Netchiporuk, A. P. Soldatkin, W. B. Wlodarski, Anal. Chim. Acta 1993, 281, 3. [42] M. J. Schçning, A. Poghossian, Analyst 2002, 127, 1137. [43] J. C. Fernando, K. R. Rogers, N. A. Anis, J. J. Valdes, R. G. Thompson, A. T. Eldefrawi, M. E. Eldefrawi, J. Agric. Food Chem. 1993, 41, 511.
H 2004 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
1906 [44] A. Flounders, A. Singh, J. Volponi, S. Carichner, K. Wally, A. Simonian, J. Wild, J. Schoeniger, Biosens. Bioelectron. 1999, 14, 713. [45] A. L. Simonian, J. K. Grimsley, A. W. Flounders, J. S. Schoeniger, T.-C. Cheng, J. J. DeFrank, J. R. Wild, Anal. Chim. Acta 2001, 442, 15.
Electroanalysis 2004, 16, No. 22
A. L. Simonian et al. [46] A. K. Singh, A. W. Flounders, J. V. Volponi, C. S. Ashley, K. Wally, J. S. Schoeniger, Biosens. Bioelectron. 1999, 14, 703. [47] M. M. Benning, J. M. Kuo, F. M. Raushel, H. M. Holden, Biochemistry 1994, 33, 15001. [48] Q. Yang, P. Atanasov, E. Wilkens, R. C. Hughes, Anal. Lett. 1995, 28, 2439.
H 2004 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
Full Paper
FET-Based Biosensors for The Direct Detection of Organophosphate Neurotoxins A. L. Simonian,*a A. W. Flounders,b J. R. Wildc a
Auburn University, Materials Research and Education Center, Auburn, AL, USA *e-mail: [email protected] b University of California, Microfabrication Laboratory, Berkeley, CA, USA c Biochemistry and Biophysics Department, Texas A&M University, College Station, TX, USA Received: April 15, 2004 Final version: May 10, 2004 Abstract Recent world-wide terrorist events associated with the threat of hazardous chemical agent proliferation, and outbreaks of chemical contamination in the food supply has demonstrated an urgent need for sensors that can directly detect the presence of dangerous chemical toxins. Such sensors must enable real-time detection and accurate identification of different classes of pesticides (e.g., carbamates and organophosphates) but must especially discriminate between widely used organophosphate (OP) pesticides and G- and V-type organophosphate chemical warfare nerve agents. Present field analytic sensors are bulky with limited specificity, require specially-trained personnel, and, in some cases, depend upon lengthy analysis time and specialized facilities. Most bioanalytical based systems are biomimetic. These sensors utilize sensitive enzyme recognition elements that are the in-vivo target of the neurotoxic agents which the sensor is attempting to detect. The strategy is well founded; if you want to detect cholinesterase toxins use cholinesterase receptors. However, this approach has multiple limitations. Cholinesterase receptors are sensitive to a wide range of non-related compounds and require lengthy incubation time. Cholinesterase sensors are inherently inhibition mode and therefore require baseline testing followed by sample exposure, retest and comparison to baseline. Finally, due to the irreversible nature of enzyme-ligand interactions, inhibition-mode sensors cannot be reused without regeneration of enzyme activity, which in many cases is inefficient and time-consuming. In 1996, we pioneered a new “kinetic” approach for the direct detection of OP neurotoxins based on agent hydrolysis by the enzyme organophosphate hydrolase (OPH; EC 3.1.8.2; phosphotriesterase) and further identified a novel multienzyme strategy for discrimination between different classes of neurotoxins. The major advantage of this sensor strategy is it allows direct and continuous measurement of OP agents using a reversible biorecognition element. We also investigated incorporation of enzymes with variations in substrate specificity (e.g., native OPH, site-directed mutants of OPH, and OPAA (EC 3.1.8.1), based upon preferential hydrolysis of PO, PF and PS bonds to enable discrimination among chemically diverse OP compounds. Organophosphate hydrolase enzymes were integrated with several different transduction platforms including conventional pH electrodes, fluoride ion-sensitive electrodes, and pH-responsive fluorescent dyes. Detection limit for most systems was in the low ppm concentration range. This article reviews our integration of organophosphate hydrolase enzymes with pH sensitive field effect transistors (FETs) for OP detection. Keywords: Organophosphate hydrolase, Organophosphorus acid anhydrolase, pHFET, ENFET, Biosensor, Diisopropyl fluorophosphate, Paraoxon, Demeton-S, Detection, Chemical warfare agents
1. Introduction The detection and identification of newly-introduced, organophosphate (OP) neurotoxins in air, water and soil poses an extremely difficult challenge. Soil and water samples are very likely to contain organophosphate pesticide residues due to heavy rural and urban use of these compounds. Yet, it is critical to detect military and terrorist activities which may result in environmental contamination with different chemical warfare (CW) agents. Thus, it is essential that CW neurotoxins can be readily and unequivocally distinguished from chemically similar agricultural compounds; ubiquitous organophosphate pesticides must never be interpreted as false positive indication of CW agents. Figure 1 compares the chemical structure of five Electroanalysis 2004, 16, No. 22
organophosphate compounds; all are toxic cholinesterase inhibitors but they vary in their neurotoxicity from organism to organism [1]. Paraoxon, diisopropyl fluorophosphate (DFP), and demeton-S are commercially available organophosphate pesticides. Paraoxon is representative of pesticides with neurotoxic phosphotriester bonds (PO bond); demeton-S is representative of phosphonothioate pesticides with a PS bond, and diisopropyl fluorophosphate is a phosphonofluoridate containing a PF bond. The structural similarity between demeton-S and DFP, and the different types of CW agents, have rendered them to be appropriate analogues for the development of detection and destruction technologies for these chemical agents. One of the main approaches used in the development of neurotoxin biosensors involves the inhibition of a specific
H 2004 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
DOI: 10.1002/elan.200403078
1897
Detection of Organophosphate Neurotoxins
Fig. 1. Chemical structures of some organophosphate neurotoxins.
enzyme in the presence of target analytes. The operational principle of inhibition-mode OP biosensors is based on the fact that the OP agents irreversibly bind at the enzyme active site, resulting in a loss of enzyme activity and hence a decrease in sensor signal. Several biosensors based on acetylcholinesterase (AChE) or butyryl cholinesterase (BChE) inhibition have been developed and used for OP neurotoxin detection [2 – 5]. A number of other enzymes such as urease and glucose oxidase have also been used as the target receptor in biosensors for their unique inhibitors [6, 7]. Inhibition-mode based sensors suffer from a number of limitations, such as: (i) any environmental or handling factors that cause loss of enzyme activity will result in false positive signals [8 – 10]; (ii) they requires baseline testing prior to sample application and lengthy sample incubation times to allow enzyme-analyte interaction; (iii) due to the irreversible nature of enzyme inhibition, inhibition-mode sensors cannot be reused without regeneration of enzyme activity, which in many cases is very inefficient. Despite several attempts to improve the specificity of cholinesterase sensors by monitoring of inhibition with both butyryl and acetyl cholinesterase [11] or analyzing inhibition patterns obtained with several acetyl cholinesterase enzymes from diverse sources [12, 13, 14], the inhibition-based sensors still have multiple unresolved problems, especially the inability to discriminate between neurotoxic agents and other unrelated enzyme inhibiting sample components such as heavy metals and pH variation. One of the ways to overcome the disadvantages of inhibition-based biosensors is to replace the inhibition receptor enzyme with a catalytic receptor enzyme. In this case, rapid, first-order enzyme kinetics may be utilized for real time or near-real time analysis. In 1996, we pioneered the application of the catalytic receptor approach to the direct detection of OP neurotoxins by utilizing the enzyme organophosphate hydrolase [15] for OP detection. Later, we developed a novel multi-enzyme strategy (using both organophosphate hydrolase and acetyl cholinesterase) for discrimination among different classes of neurotoxins [9]. Our approach was based on organophosphorus hydrolase (OPH; E.C. 3.1.8.1), a well-characterized metalloenzyme Electroanalysis 2004, 16, No. 22
originally isolated from Pseudomonas diminuta [16]. OPH is capable of cleaving PO, PF, PS, and PCN bonds via an SN2-type mechanism [17 – 19], resulting in unique sets of hydrolysis products, which also change the solution pH according to the reaction: Direct neurotoxin detection is thus possible via measurement of the pH change associated with enzyme activity [15]. OPH exhibits the unique ability to hydrolyze a large variety of organophosphate pesticides and neurotoxins including coumaphos, parathion, acephate, sarin, and VX [17]. OPH has been used to develop new biosensors for the direct detection of organophosphates based on pH monitoring of OPH activity with conventional pH electrodes [15, 20], or pH sensitive fluorescent dyes [21]. This strategy can be extended to further discriminate between different OP neurotoxins by monitoring additional organophosphate hydrolysis products. For example, hydrolysis of phosphofluoridates yields changes in pF as well as pH that can be detected with a fluoride specific ion-selective electrode [22]. These approaches have been subsequently validated by other research groups using the same enzyme (OPH or phosphotriesterase) who have verified the sensitivity and selectivity with different enzyme formats, such as whole cells or crude cell extracts [23, 24]. A separate class of hydrolytic enzymes, organophosphorus acid anhydrolase, (OPAA; EC 3.1.8.2.), has also been utilized for the direct detection of organophosphate compounds. The first OPAA discovered was isolated from a halophilic bacterial isolate designated JD6.5 and demonstrated high levels of DFP hydrolysis but lacked PO and PS bond hydrolysis [25]. The bacterial isolate was identified as a species of the genus Altermonas and a similar DFPhydrolyzing activity has been observed in several bacterial species [26]. As with OPH, the native function of OPAA is not yet known. However, OPAA has been unambiguously identified as a prolidase (E. C. 3.4.13.9) capable of hydrolyzing dipeptides with a prolyl residue in the carboxylterminal position [27]. A significant biosensor performance criterion is its analyte specificity. Enzymes with broad substrate specificity are suitable for detection of multiple analytes belonging to the same chemical class, while enzymes with high specificity for a single substrate are best for discriminative detection of a single target analyte. Initial investigations of OPAA substrate selectivity [25, 26] indicated that the enzyme is capable of cleaving PF bonds, while PO or PS bond hydrolysis is minimal. Since organophosphate pesticides, such as paraoxon and demeton-S, typically possess PO and PS bond structures (Fig. 1), such unique substrate prefer-
Scheme 1. H 2004 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
1898 ence makes OPAA very attractive for discriminative detection of fluorine containing organophosphates. Extending this strategy, a combination of OPH and OPAA in the same system may enable development of a multi-enzyme senor array capable of discriminating among PO, PS, and PF containing organophosphate compounds. Highly accurate and sensitive measurement of the pH changes associated with analyte hydrolysis is required for the development of such an enzyme-based biosensor system. A pH sensitive field effect transistor (pHFET) is the ideal device for monitoring these pH changes since the enzyme can be directly immobilized to the pH sensitive surface optimizing detection of enzyme hydrolysis products. PHFETS also provide a platform which lends itself to development of a multi-enzyme sensor array since multiple pH sensors can be integrated on a single sensor platform and different enzymes immobilized to each pH sensor. There is a significant literature related to pH-sensitive FETs (e.g., [28, 29]). Early investigations were frustrated by electrically leaky silicon dioxide and silicon nitride and physically leaky device encapsulation. More recent reports and the commercial availability of FET-based pH sensors indicate that many of these material issues have been resolved. A comprehensive overview of pHFET theory and operation written by the inventor of the device is available [27] and an interesting historical review of the evolution of pH sensitive FETs has been presented [30]. Enzyme modification of pHFETS (ENFETS) was first reported in 1980 but was hampered by unreliable pHFETS. A better understanding of FET device response led to creative and enhanced dynamic signal measurement techniques [31 – 33]. With more stable and reliable devices available, efforts have focused on resolving long recognized enzyme-modified FET issues: influence of solution buffer capacity [34], reference electrode miniaturization [35], and enzyme loading and stability. Enzyme-modified FETs have been proposed and investigated for many diverse analytes. The acetylcholinesterase-modified FET for organophosphate detection was a direct extension of cholinesterasemodified electrodes [36] and was proposed [37] soon after the initial enzyme-modified FET reports [38]. The investigation of cholinesterase-modified FETs has continued for several years [39 – 42] and the same strategy (inhibition mode/potentiometric detection) has been pursued with a commercially available, light addressable potentiometric sensor [43]. We previously reported the first use of the organophosphate hydrolase modified field effect transistor for organophosphate direct detection [44, 45] This present paper summarizes our efforts in the development of multi-enzyme biosensors based on pH-sensitive FET platforms and several OP-hydrolase enzymes (e.g., native OPH, genetically engineered OPH (site directed mutants), and OPAA).
Electroanalysis 2004, 16, No. 22
A. L. Simonian et al.
2. Experimental 2.1. Reagents and Buffers Paraoxon (diethyl-p-nitrophenyl phosphate), diisopropyl fluorophosphate (DFP), glutaraldehyde (grade 1), and reagents for buffer (NH4Cl, NaCl, MnCl2) were from Sigma (USA); demeton-S (O,O-dimethyl-S-2-ethylthiolethyl phosphorothioate), was from Chem Service (West Chester, Pa, USA). All solutions were prepared using 18 MW · cm ultrapure water (Milli-Q Plus, Millipore, ST).
2.2. Enzyme and Immobilization 2.2.1. Enzymes Wild-type OPH (E.C. 3.1.8.1) was isolated from a recombinant Escherichia coli strain using published procedures [19]. Various modified OPH enzymes with enhanced PF and PS bond cleaving capabilities have been developed and similarly purified and stored as lypholyzed powder [17]. OPAA (EC 3.1.8.2.) from Alteromonas sp. strain JD6.5 was isolated and purified to homogeneity [25] at the U.S. Army Edgewood Chemical Biological Center and obtained as a lyophilized enzyme powder (0.31 mg protein/mg powder). Since OPAA activity is enhanced with ammonia based buffers with low concentrations of MnCl2 [26], OPAA solutions were prepared in 50 mM NH4Cl, pH 8.5, with 0.1 mM MnCl2, 100 mM NaCl. 2.2.2. Immobilization on pHFET Chips Enzymes were immobilized to packaged pHFETchips following the aqueous aminopropyl-triethoxysilane (APTS): glutaraldehyde covalent attachment strategy described previously [46]. The APTS:glutaraldehyde was the preferred immobilization chemistry since it yielded higher total enzyme activity in bead and glass slide control studies. Sensor chips were treated with 1 N HCl, 15 min followed by 30% H2O2, 30 min for surface cleaning and silanol activation rather than the sulfuric acid, hydrogen peroxide activation used previously [46] to limit encapsulant degradation. Organosilane and glutaraldehyde treatments were performed by immersing packaged chips in solution. Enzyme was applied to only one device gate by pipetting enzyme into only one epoxy defined well. After extensive rinsing, chips were stored at 4 8C in 10 mM phosphate buffered saline (pH 7.4). Some packaged chips were reused by stripping enzyme in an oxygen plasma reactor (March Instruments, PX1000) at 50 W, 300 mtorr, 100 sccm O2, 5 minutes). Devices were retested for pH response after oxygen plasma stripping and then recoated with enzyme by repeating all immobilization chemistry treatments.
H 2004 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
1899
Detection of Organophosphate Neurotoxins
2.2.3. Sol-Gel Modification and Enzyme Immobilization To increase the surface area available for enzyme immobilization and provide a more stable enzyme microenvironment, silica microspheres were formed from tetraethoxysilane (Gelest, Inc., Tullytown, PA) via a base catalyzed solgel process then applied to diced chips by dip coating as described previously [46]. Dip coating was performed with diced chips prior to packaging, wire bonding and epoxy encapsulation. Sol-gel sphere diameters were determined with a sub-micron particle analyzer (Coulter, Model N4 MD) and verified with electron microscopy. Spheres were deposited over the entire chip surface. Spheres were scraped with tweezers from metal bond pads prior to wire bonding; spheres not in the gate region were completely covered with epoxy encapsulation. Sol-gel and non-sol-gel coated chips underwent identical enzyme immobilization and paraoxon testing procedures. 2.2.4. Testing Procedure All micro-chips were soaked in deionized water for at least 24 hours prior to testing. Devices were tested either (a) in batch mode by mounting a sealing ring over the encapsulated chip package which created a solution well (Fig. 2a) of approximately 1.2 mL volume; or (b) in a flow system consisting of a peristaltic pump and a flow cell which mounted over the PGA package providing a head space of approximately 75 mL (Fig. 2, b, c). A silver:silver chloride standard reference electrode (Orion Sure-Flow, Beverly, MA Orion, USA) was immersed in the test solution (batch mode) or a silver:silver chloride micro flow-through reference electrode (Microelectrode Inc., Bedford, USA) was introduced in the waste outlet stream (flow-through mode). Measurements were taken only with flow stopped to avoid streaming potential interferences. Device pH response was tested by monitoring changes in the common source voltage (Vcs, Fig. 3), with pH 4, 7, and 10 standard buffers (Sigma, St Louis, MO) before and after enzyme immobilization, sol-gel coating, and oxygen plasma treatment. Though pH 4.0 and 10.0 are not extreme conditions for enzyme exposure, chips with immobilized enzyme that were used for pH testing were not used for paraoxon testing unless fresh enzyme was immobilized to the chips. Chips with immobilized enzyme were tested for organophosphate sensitivity in 1) glycine-NaOH buffer (0.5, 1, or 10 mM with 0, 50, 100 mM NaCl) at pH 9.0 for OPH, or 2) in 50 mM NH4 Cl, pH 8.5, with 0.1 mM MnCl2, 100 mM NaCl for OPAA. Varying volumes of paraoxon were added, diluted in glycine buffer at pH 9.0 or ammonia buffer at pH 8.5, and changes in differential output voltage (Vdiff, Fig. 3) were monitored. The rate of hydrolysis of paraoxon in the absence of OPH was measured spectrophotometrically. This non-enzymatic hydrolysis was non-zero, but it was determined to be noncritical relative to the time scale of all sensor measurements. The independent measurement of pH after addition of paraoxon to the sensor solution was performed by removing Electroanalysis 2004, 16, No. 22
Fig. 2. a) Packaged pHFET chip with elastomer ring for bachmode measurements; b) dual pHFET chip with epoxy encapsulant; and, c) electronic control board with flow-through assembly.
a sample from the sensor solution and testing with the same pH meter used for all buffer testing and adjustments (Orion 601, Beverly, MA).
2.3. Apparatus and Procedures Sensor chips were prototype pH-sensitive field effect transistors (SenDx Medical, Inc., Carlsbad, CA). Each chip contained two discrete depletion modes, n-channel transistors with a non-metallized gate insulator stack of thermal silicon dioxide and chemical vapor deposited (CVD) silicon nitride (Figure 2, a). A porous ceramic matrix was formed at the gate surface by dip coating the chip prior to packaging into a solution of 200nm silica microspheres formed via a sol gel process as described above [46]. The chips were diced, packaged, and installed in a differential control circuit; Figure 3 shows a schematic of the chip control circuit. The control circuit was configured as a classical differential pair amplifier and used to monitor the gate to source voltage of the enzyme-modified device relative to the nonenzyme-modified device. The main advantage of the differential pair is that common mode variations such as temperH 2004 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
1900
A. L. Simonian et al.
Fig. 3. Schematic of the differential sensor circuit. Enzyme is immobilized to the gate of one pHFET while second pHFET serves as a non-enzyme coated reference. Bulk pH changes are measured as changes in the common source voltage Vcs and enzyme specific responses are measured as changes in Vdiff.
ature and bulk pH changes are eliminated while local pH changes at the FET gate with immobilized enzyme are amplified. Each experimental point was the average of 3 – 5 measurements.
ence of 1 mM paraoxon are : Vmaxapp ¼ 2.54 0.30 mV/s, and K app m ¼ 0.55 0.23 mM and in the presence of 1 mM demeton-S are 2.63 0.30 mV/s, and K app m ¼ 0.57 0.23 mM.
3.2. Influence of Buffer Capacity on the Sensor Response
3. Results and Discussion 3.1. Enzyme Kinetic Characteristics Table 1 presents kinetic characteristics of WT OPH, mutant RL-OPH, and OPAA. The comparison of OPH and OPAA kinetic parameters emphasizes the unique recognition capability of OPH and OPAA. OPAA displays a strong preference for PF bond hydrolysis making it suitable for discriminative detection of fluorine containing organophosphates, while OPH displays significant activity for PO, PF, and PS bonds hydrolysis making it suitable for detection of multiple organophosphates. PS bond hydrolysis by genetically modified OPH (OPH-RL) increased significantly (40 ) relative to wild type OPH) making possible demethon-S detection. Apparent kinetic parameters for soluble OPAA with DFP as a substrate, obtained by pH electrode, are: Vmaxapp ¼ 2.70 0.28 millivolts per second (mV/s), and K app m ¼ 0.54 0.08 mM. The same parameters in the presElectroanalysis 2004, 16, No. 22
The influence of the buffer capacity of the solution on device response is a well documented limitation of EnFETs. It is
Table 1. Comparative kinetic constants for OPH and OPAA enzymes Substrate/hydrolyzed bond
Enzyme
kcat (s1)
KM (mM )
DFP [ PF] [a]
OPAA OPH OPHRL OPAA OPH RL OPAA OPH OPH-RL
770 3500 8.2 4.64 11500 1400 0.028 2.6 68.0
2.81 1.42 1.2 0.6 0.12 0.036 3.5 3.0 2.5
Paraoxon [ PO] [b] Demeton-S [ PS]
[a] detected by fluoride-specific electrode; [b] detected by pH sensitive electrode.
H 2004 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
Detection of Organophosphate Neurotoxins
necessary to minimize buffer capacitance to optimize pH signal available from enzymatic hydrolysis of the analyte. However, this demand must be balanced by the conflicting need to provide some solution pH control to minimize background pH drift caused by sample addition and to maintain the pH required for enzyme activity. Figure 4 presents changes in batch mode sensor response for 1 mM paraoxon in solutions with different buffer strength. High salt concentration helps stabilize buffer, however, in deionized water solution salts only are not helpful. A buffer strength of 1 mM was selected as a compromise between satisfactory signal and the ability of an open system (subject to CO2 absorption especially in batch mode) to maintain constant pH. It should be noted that the independent pH measurements performed on removed samples indicated no measurable change in bulk pH associated with analyte hydrolysis. All buffer strengths tested were sufficient to compensate the acidic product generated by the immobilized enzyme once it diffused into the bulk. It was suspected that the sol-gel coatings may either act as a membrane and diminish the effect of buffer strength [35] or act as an additional solid state buffer and further diminish device response. However, no difference was noted between devices with and without sol-gel coatings with respect to buffer strength.
3.3. Sensor Response on Paraoxon with Native OPH Enzyme Sensor response to paraoxon concentrations from 10 to 50 mM (batch) and from 0.5 to 50 mM (flow) is presented in Figure 5. The “batch data” represents the averaged re-
Fig. 4. Effect of buffer concentration on sensor response. Batch mode measurements. Glycine-NaOH buffer with 120 mM NaCL and 2.5 mM KCl, pH 9.0. Initial 0 point – deionized H2O with salts, no glycine-NaOH. Electroanalysis 2004, 16, No. 22
1901 sponse of several different chips and the “flow data” represents the averaged response of multiple tests with the same chip. With the batch system, reproducible measurements were possible for paraoxon concentrations as low as 5 mM, while the flow system enabled reproducible measurements to 0.5 – 1 mM paraoxon. The two systems were set up at different laboratories and used by different experimenters, and the agreement between the two systems was excellent. To demonstrate the time scale of response, the sensor signals during repeated injections of paraoxon to the flow cell are presented in Figure 6. A comparison of the signal from sensors with and without sol-gel coatings to 10 – 100 mM paraoxon is presented in Figure 7. The overall device response was enhanced with the 200 nm sol-gel coating. Since the sol-gel coatings did not affect pH sensitivity, the measured signal increase is most likely due to an increase in enzyme loading or immobilized enzyme specific activity. This result correlated well with our earlier studies demonstrating that the total activity of OPH is enhanced by immobilizing enzyme to porous rather than non-porous particles [46]. Sensor response with a 20 nm solgel modification was the same as that of uncoated sensors. Assuming very close packed spheres, the pore size provided by the 20 nm sol-gel coating is one third of particle diameter or approximately 67 A. Unit cell dimensions of single crystal OPH enzyme from X-ray diffraction are 80.3, 93.4, and 44.8 A, [47]. This may explain why the 20 nm coating had no effect on sensor response, as the enzyme might have been unable to penetrate a silica microsphere matrix with an intersphere pore size equivalent to enzyme dimensions. The much larger pores provided by the 200 nm sol-gel coating would not limit enzyme penetration, and an increase in immobilized enzyme loading is expected. The 200 nm solgel coated sensors still displayed widely varying chip-to-chip response (CV ¼ 25 – 30%). In summary, the sol-gel coatings may have increased sensor response by increasing immobilized enzyme total activity; however, on a single chip surface,
Fig. 5. Non sol-gel coated sensor response to paraoxon. All tests conducted in 1 mM glycine-NaOH buffer pH 9.0. Inset shows flow system response to concentrations below 10 mM paraoxon. Flow system, N ¼ 5; batch system, N ¼ 6. H 2004 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
1902
A. L. Simonian et al.
were stored in PBS at 4 8C between measurements. The 200 nm sol-gel modified devices maintained greater than 80% of initial activity for the entire 6-week test period while the uncoated devices gradually lost more than 70% activity during the same period. This result supported our prior work [44], which demonstrated that porous immobilization platforms were superior to non-porous immobilization platforms at maintaining immobilized enzyme activity. A similar increase in enzyme stability with a preference for specific silica microsphere diameter has been reported [48]. All devices lost all activity after day 75. This is attributed to failure to include any preservative in the storage buffer.
3.4. Sensor Responses on Demethon-S with Genetically Modified Enzymes (RL-OPH
Fig. 6. Non sol-gel coated sensor response to paraoxon. All tests conducted in 1 mM glycine-NaOH buffer pH 9.0. Different concentrations of paraoxon injected at each dashed arrow, glycine buffer injected at each solid arrow.
As shown in Table 1, both wild type OPH and OPAA have low catalytic activity toward demeton-S. WT OPH is able to hydrolyze PS bonds only slowly and requires high concentration of analyte. Whereas immobilization of WT-OPH on a pHFET gate provides enough catalytic activity to hydrolyze paraoxon or DFP and produce measurable pH changes. It was not possible to obtain measurable response for demeton-S at concentrations up to 15 mM. To overcome this problem, the wild type OPH was genetically modified to increase catalytic activity toward PS bonds. This enzyme (OPH-RL, H254R/H257L) (and several others) was active enough to produce measurable pH changes in response to demeton-S exposure. (Figure 9). This enzyme displays a dramatic enhancement of the rate of hydrolysis of demeton-S and a significant activity decrease of hydrolysis of DFP (Table 1). This enhancement of the rate of demeton-S hydrolysis (kcat increased from 2.6 to 68.0 turnovers per second) was adequate to register positive biosensor responses.
3.5. Sensor Responses to DFP with OPAA
Fig. 7. Comparison of sensors with and without sol-gel coatings. All tests conducted in 1 mM glycine-NaOH buffer pH 9.0. N ¼ 4.
the sol-gel coating does not increase the surface area to volume ratio to such an extent that planar variations in enzyme loading are removed. With such a small area for direct immobilization, highly controllable and reproducible enzyme loading and activity continues to be a concern and the most significant source of sensor variability. The sol-gel coatings did not increase enzyme density to the point that dramatic changes in bulk pH were recorded. Devices with and without sol-gel coatings were tested with paraoxon at 1 – 2 week intervals over a 6-week period (Fig. 8). Devices Electroanalysis 2004, 16, No. 22
When OPAA was used as a biorecognition element, the sensor response to varying substrates was completely different, and detection time depended on the concentration of DFP. Figure 10 presents the responses of the OPAAmodified pH-sensitive FETs as a function of substrate concentration for different OP compounds. As expected, based on OPAA kinetic characteristics (see Table 1), the differences between sensor response to DFP and other substrates were quite large. Paraoxon generated barely measurable responses, even at high concentrations and there was no detectable signal for demeton-S. Linear, firstorder kinetics were observed for DFP concentrations from 12.5 to 500 mM. Similar high substrate specificity for DFP was observed also in the mix solutions contained 0.5 mM of each neurotoxin, paraoxon and demeton-S (Fig. 10).
H 2004 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
1903
Detection of Organophosphate Neurotoxins
Fig. 8. Sensor signal versus time since enzyme immobilization. All tests conducted in 1 mM glycine-NaOH buffer pH 9.0. Sensor response to 10, 50 and 100 mM paraoxon normalized with respect to sensor response on first day after immobilization. N ¼ 3.
however, the sensor response was not as sensitive to changes in pH as the soluble enzyme. [26] Satisfactory signal output was obtained from pH 7.5 – 9.0.
3.7. Long-Term Stability
Fig. 9. Sensor response on demeton-S. RL OPH, 1 mM glycineNaOH buffer pH 9.0.
OPAA immobilized on the sol-gel modified pH-sensitive chips lost 60% of its activity in 22 days, and there was no response after 30 days. This was especially surprising since previous results demonstrated that the silica sphere modification of the pH sensitive chips enabled the maintenance of a stable sensor response for more than 2 months [40]. Storage stability of OPAA immobilized on suspended silica gel particles was investigated separately [45]. In these studies, immobilized OPAA retained about 60% of the starting activity after five months. Such dramatic differences in long-term stability may be due to the significant differences in enzyme capacity of the two sensor formats; the silica gel has a much greater surface area than the pH-sensitive chip surface.
3.6. pH Dependence
3.8. Conceptual Design for Integrated Multi-Enzyme and Multi-Sensor Array
The previously reported pH optimum for OPAA activity (k app cat Þ with DFP as a substrate was 8.5, while the highest app catalytic efficiency (k app cat /K m Þ was observed at pH 6.8 [26]. However, enzyme immobilization can influence enzyme operational capability and change the pH optimum, depending on which sites on the enzymes were involved in binding to a surface. In order to determine the pH optimum for the OPAA sensor, the response to a fixed concentration of DFP (1mM) at pH 6.2 – 9.0 was measured (Fig. 11). Maximum sensor response was obtained at pH 8.5,
Measurable variations in enzyme-substrate specificity can be achieved with site directed mutagenesis and even subtle activity changes can be reproducibly detected with enzyme in solution. However, large variations in enzyme loading and enzyme activity after immobilization mean that only significant changes in enzyme activity can be detected and exploited for development of a multi-enzyme sensor array. Incorporation of enzyme variants from diverse species (OPAA vs OPH) provides far more discriminative capability than just incorporation of genetically modified
Electroanalysis 2004, 16, No. 22
H 2004 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
1904
A. L. Simonian et al.
Fig. 10. Signal versus concentration for DFP (black squares) paraoxon (inverted triangles), and mixture of 0.5mM of paraoxon, 0.5 mM of demeton-S and DFP (triangles): a) stop-flow assay with pH-sensitive field effect transistor, full curves; b) the linear ranges of the calibration plots, f(x)DFP ¼ 43.5 þ 1469x, R ¼ 0.99976, f(x)mix ¼ 42.5 þ 1255x, R ¼ 0.98891, f(x)paraoxon ¼ 13.9 þ 77.1x, R ¼ 0.85367.
Fig. 12. Anticipated response of integrated multi-enzyme and multisensor array. Proposed digital signal summarized in Table below. OPH* ¼ OPH genetic variant H254R/H257L abbreviated as OPH-RL.
Fig. 11. pH profile for DFP biosensor response. Stop-flow mode with pH-sensitive field effect transistor. Measurement condition: 1 mM NH4Cl buffer, containing 0.1 mM MnCl2 and 100 mM NaCl. DFP concentration 1 mM.
enzymes from the same source (OPH vs RL-OPH). Utilization of both strategies – genetically modified enzymes and multiple sources of enzymes provides the broadest discrimination. The appropriate combination of recognition elements and sensor platforms enables a set of almost digital responses, which are immune to the level of variation in enzyme loading and activity that is to be expected during production of an immobilized enzyme biosensor (Figure 12). Electroanalysis 2004, 16, No. 22
4. Conclusions Our investigations demonstrate the ability of the enzymes OPH, RL-OPH-RL, and OPAA to serve as highly discriminative biorecognition elements in biosensors for specific and direct detection of organophosphate neurotoxins. The immobilization of these enzymes to the exposed gate of a pH-sensitive field effect transistor produced a group of organophosphate sensors with rapid response to micromolar concentrations of paraoxon and DFP. The use of a specifically designed, genetically modified enzyme, RLOPH-RL, extends sensor discrimination and allows detection of demeton-S but with a higher detection limit. Combinations of organophosphate hydrolyzing enzymes, such as OPH, RL-OPH, and OPAA enable the construction of a multienzyme biosensor array capable of discriminating H 2004 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
Detection of Organophosphate Neurotoxins
among several related organophosphate compounds. Samples containing mixtures of both a G-type neurotoxin, such as sarin, and soman and organophosphate pesticide residues could be readily discriminatedidentified by this biosensor if combined with a front end sample separation system (e.g., HPLC). Though the sensors does do not provide the nanomolar organophosphate detection limits exhibited by acetylcholinesterase-modified FETs, OPH/OPAA-modified FETs do not require a continuous supply of consumable enzyme substrate, have a much faster response time, can be reused, are far less likely to generate false positive signal, and can discriminated between various neurotoxic pesticides and chemical warfare agents. Many of the early material problems of pH FETs have been resolved. Enzyme-modified FETs can now be re-evaluated as a suitable sensor platform in those cases where the detection needs correspond to the limits imposed by enzyme catalytic parameters and system buffer capacity. Due to its inherent miniaturization, low power consumption, and ease of integration, the FET is still an extremely appealing transduction platform. The sol-gel gate modifications investigated appear to provide a gate surface more suitable for enzyme immobilization. This sol-gel surface modification is not a mechanism for enhancing pH sensitivity or diminishing the dependence of the enzyme modified FET on the buffer capacity of the surrounding electrolyte.
5. Acknowledgements Support from Sandia National Laboratories and DOE Office of Nonproliferation Research and Engineering, USA-TATRC DREAMS, DARPA/ARO-MURI, NSFMURI, and Auburn University Detection and Food Safety Center is gratefully acknowledged.
6. References [1] D. P. Dumas, H. D. Durst, W. G. Landis, F. M. Raushel, J. R. Wild, Arch. Biochem. Biophys. 1990, 277, 155. [2] G. A. Evtugyn, H. C. Budnikov E. B. Nikolskaya, Analyst 1996, 121, 1911. [3] N. Mionetto, J.-L. Marty, I. Karube, Biosens. Bioelectron. 1994, 9, 463. [4] G. Palleschi, M. Bernabei, C. Cremisini, M. Mascini. Sens. Actuators B 1992, 7, 513. [5] J. J. Kulys, E. J. DQCosta. Biosens. Bioelectron. 1991, 6, 109. [6] A. Herbert, L. Guilhermino, H. C. Da Silva De Assis, P.-D. Hasen, Z. Angew. Zoologie 1995/96, 3, 1. [7] M. Brufani, M. Marta, M. Pomponi, Eur. J. Biochem. 1986, 157, 115. [8] A. L. Simonian, E. I. Rainina, J. R Wild, Anal. Lett. 1997, 30, 2453. [9] A. Herbert, L. Guilhermino, H. C. Da Silva De Assis, P.-D. Hasen, Z. Angew. Zoologie, 1995/96, 3, 1. [10] J. F. Payne, A. Mathieu, W. Melvin, L. L. Fancey, Marine Pollution Bulletin, 81. [11] T. T. Bachmann, R. D. Schmid, Anal. Chim. Acta 1999, 401, 95.
Electroanalysis 2004, 16, No. 22
1905 [12] O. Schafer, L. Weil, R. Niessner, Methoden Vom Wasser 1994, 82, 233. [13] T. T. Bachmann, R. D. Schmid, Anal. Chim. Acta 1999, 401, 95. [14] O. Schafer, L. Weil, R. Niessner, Methoden Vom Wasser 1994, 82, 233. [15] E. Rainina, A. Simonian, A. Efremenko, S.Varfolomeyev, J. Wild, Biosens. Bioelectron. 1996, 11, 991. [16] C. S. McDaniel, L. L. Harper, J. R. Wild, J. Bacteriol. 1988, 170, 2306. [17] D. P. Dumas, J. R. Wild, F. M. Raushel, Biotech. Appl. Biochem. 1989, 11, 235. [18] D. P. Dumas, H. D. Durst, W. G. Landis, F. M. Raushel, J. R. Wild, Arch. Biochem. Biophys. 1990, 277, 155. [19] K. Lai, N. .J. Stolowich, J. R. Wild, Arch. Biochem. Biophys. 1995, 318, 59. [20] A. L. Simonian, E. I. Rainina, J. R. Wild, in Methods in Biotechnology: Enzyme Biosensors – Techniques and Protocols (Eds: A. Mulchandani, K. Rogers), Humana Press, N. J. 1998, 17, pp. 237 – 248. [21] R. Russell, M. Pishko, A. Simonian, J. Wild, Anal. Chem. 1999, 71, 4909. [22] A. L. Simonian, B. D. diSioudi, J. R. Wild, Anal. Chim. Acta 1999, 389, 189. [23] P. Mulchandani, W. Chen, A. Mulchandani, J. Wang, L. Chen, Biosens. Bioelectron. 2001, 16, 433. [24] K. Roger, Y. Wang, A. Mulchandani, P. Mulchandani, W. Chen, Biotech. Prog. 1999, 15, 517. [25] J. J. DeFrank, T.-C. Cheng, J. Bacteriology 1991, 173, 1938. [26] J. J. DeFranik, W. T. Beaudry, T.-C. Cheng, S. P. Harvey, A. N. Stroup, L. L. Szafraniec, Chem.-Biol Interactions 1993, 87, 141. [27] T.-C. Cheng, J. J. DeFrank, V. K. Rastogi, Chim.-Biol. Interactions 1999, 119 – 120, 462. [28] Solid State Chemical Sensors (Eds: J. Janata, R. J. Huber), Academic Press, Orlando, FL 1985. [29] P. Bergveld, A. Sibbald, Analytical and Biomedical Applications of Ion Sensitive Field Effect Transistors Elsevier, Amsterdam 1988. [30] J. Janata, Analyst 1994, 119, 2275. [31] B. H. Van Der Schoot, P. Bergveld, Anal. Chim. Acta 1987, 199,160. [32] B. H. Van Der Schoot, P. Bergveld, Anal. Chim. Acta 1990, 233, 49. [33] A. Poghossian, A. Baade, H. Emons, M. J. Schçning, Sens. Actuators B 2001, 76, 634. [34] A. A. ShulQga, V. I. Strikha, A. P. Soldatkin, A. V. ElQskaya, H. Maupas, C. Martelet, P. Clechet, Anal. Chim. Acta 1993, 278, 233. [35] F. Lisdat, W. Moritz, Sens. Act. B 1993, 15 – 16, 228. [36] P. Strop, G. G. Guilbault, Anal. Chim. Acta 1972, 62, 425. [37] J. Janata, R. J. Huber, R. Cohen, E. S. Kolesar Jr., Aviat. Space Environ. Med. 1981, 52, 666. [38] S. Caras, J. Janata, Anal. Chem.1980, 52, 1935. [39] C. Dumschat,H. Muller, K. Stein, G. Schwedt, Anal. Chim. Acta 1991, 252, 7. [40] Y. Vlasov, A. Bratov, S. Levichev, Y. Tarantov, Sens. Actuators B 1991, 4, 283. [41] A. M. Nyamsi- Hendji, N. Jaffrezic-Renault, C. Martelet, P. Clechet, A. A. ShulQga, V. I. Strikha, L. I. Netchiporuk, A. P. Soldatkin, W. B. Wlodarski, Anal. Chim. Acta 1993, 281, 3. [42] M. J. Schçning, A. Poghossian, Analyst 2002, 127, 1137. [43] J. C. Fernando, K. R. Rogers, N. A. Anis, J. J. Valdes, R. G. Thompson, A. T. Eldefrawi, M. E. Eldefrawi, J. Agric. Food Chem. 1993, 41, 511.
H 2004 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
1906 [44] A. Flounders, A. Singh, J. Volponi, S. Carichner, K. Wally, A. Simonian, J. Wild, J. Schoeniger, Biosens. Bioelectron. 1999, 14, 713. [45] A. L. Simonian, J. K. Grimsley, A. W. Flounders, J. S. Schoeniger, T.-C. Cheng, J. J. DeFrank, J. R. Wild, Anal. Chim. Acta 2001, 442, 15.
Electroanalysis 2004, 16, No. 22
A. L. Simonian et al. [46] A. K. Singh, A. W. Flounders, J. V. Volponi, C. S. Ashley, K. Wally, J. S. Schoeniger, Biosens. Bioelectron. 1999, 14, 703. [47] M. M. Benning, J. M. Kuo, F. M. Raushel, H. M. Holden, Biochemistry 1994, 33, 15001. [48] Q. Yang, P. Atanasov, E. Wilkens, R. C. Hughes, Anal. Lett. 1995, 28, 2439.
H 2004 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
