Ion Channel Sensors Based on
Artificial Receptors
Yoshio Umezawa
A number of ion channel sensors
Hiroshi Aoki University of Tokyo (Japan)
have been developed based on various artificial receptors for inorganic and organic ions and bioactive substances.
I
on channel sensing is particularly suited to detecting large, hydrophilic ionic species that have high charge numbers. These analytes include bioactive substances such as peptides, proteins, polysaccharides, and oligonucleotides as well as multivalent inorganic ions. The working principle of this biomimetic sensor is similar to the transmembrane signaling used by ligand-gated ion channels in biological membranes (Figure 1). In 1987, the Langmuir–Blodgett technique was used to make a biomimetic Ca2+ sensor by coating the surface of a glassy carbon electrode with a thin layer of a synthetic lipid with a phosphate headgroup (1). The binding of a stimulus Ca2+ to the negatively charged lipid regulates the permeation of electroactive marker ions added to the sample solution—and the first ion channel sensor (ICS) was created (Figure 1; 1–3). A considerable number of ICSs have been developed based on various artificial receptors for inorganic and organic ions and bioactive substances (3–5).
Principles Several approaches have been suggested to mimic the efficiency, selectivity, and inherent signal amplification of ligand-gated bioreceptor membranes by use of synthetic receptors. In the 1970s, deprotonation of acidic functional groups in self-assembled monolayers (SAMs) was shown to affect redox currents at the underlying electrodes (6 ), although the first demonstration of this phenomenon was not reported until the late 1980s (1). Early examples of synthetic receptors based on the Langmuir– Blodgett technique used in this biomimetic approach to ICSs included anionic phospholipids, macrocyclic polyamines, and a cyclodextrin with ammonium groups (1, 7 ). More recently, thiolated receptors were self-assembled on gold electrodes by formation of covalent sulfur–gold bonds (8, 9), offering improved sensor stability for real-life applications. © 2004 AMERICAN CHEMICAL SOCIETY
S E P T E M B E R 1 , 2 0 0 4 / A N A LY T I C A L C H E M I S T R Y
321 A
Ion (e.g., Na+) Ligand Bilayer
(a) Closed
Open Intramolecular channel
FIGURE 1. A ligand-gated ion channel protein. Complexation of the channel protein with ligands allows the transport of transmembrane ions (e.g., Na+) across the lipid bilayer.
Physical blocking
Marker permeation
Two different types of ICSs can be distinguished according to their response mechanism. In one case, analyte molecules form inclusion-type complexes with the receptors and physically block the intramolecular channels, preventing redox markers from accessing the electrode surface (Figure 2a; 3, 10). This approach closely mimics ion channel proteins and is the basis for detecting electrically neutral, redox-inactive analytes. The other mechanism uses electrostatic attraction or repulsion. When ionic analytes bind to the synthetic receptor layer, they alter the net charge of the layer. This, in turn, regulates the oxidation or reduction of easily charged electroactive ions (often referred to as redox markers) at a receptor monolayer that coats the electrode. The background currents often observed at these electrodes in the absence of electrostatic repulsive interactions between the marker ions and surface receptor monolayers result from permeation of the markers through intermolecular voids, for example, uncharged receptor SAMs (Figure 2b –d). Many sensors based on these principles have been reported (4, 5). ICSs based on synthetic receptors were developed for hydrogen and metal ions, simple organic analytes such as phthalate and glucose, and more complex analytes such as antibodies and concanavalin derivatives. Phosphate esters, antibiotics, oligopeptides, DNA dendrimers, cyclodextrins, calixarenes, and antigenic groups are typical examples from the wide range of synthetic receptors that have been used in ICSs. Sensors for estrogen (11), inorganic cations (12–16 ), pH (17 ), avidin (18), cyclic AMP (cAMP; 19), DNA-binding substrates such as spermine or acridine orange (20), phosphate (21), protonated amines (22), protamine (23), heparin (24), and oligonucleotides (25–27 ) have recently been reported. FIGURE 2. Two different types of ICSs based on synthetic receptors. (a) Formation of inclusion-type complexes of the receptors with the analytes physically blocking the markers from traversing the intramolecular channels of the receptors. (b) Electrostatic repulsion between the charged receptors (+) and the markers with the same charge (+) prevents the markers from reaching the electrode surface in the absence of analytes with the opposite charge (–). Binding of the analytes to the receptors neutralizes the receptor charges, and intermolecular voids are formed between the receptor molecules. The markers can then permeate through the voids and react at the electrode surface. (c) The charged receptors (+) allow markers with the opposite charge (–) to access the electrode surface in the absence of analytes with the opposite charge (–). Complexation of the analytes to the receptors cancels the receptor charges and gives the electrode surface an excess of the opposite charge, hindering the redox reaction of the markers. (d) In the case of uncharged receptors, by using mixed SAMs of the receptors and thiol-anchored charged molecules (+), binding of the analytes with the opposite charge (–) controls the redox reaction of the markers with the same charge as (+, left) or opposite to (–, right) those of the thiol-anchored charged molecules in the same way as in (b) or (c), respectively.
322 A
A N A LY T I C A L C H E M I S T R Y / S E P T E M B E R 1 , 2 0 0 4
Redox marker
Electrode
Receptor
Analyte
(b)
Electrostatic repulsion
Marker permeation
Marker permeation
Electrostatic repulsion
Electrostatic Marker repulsion permeation
Electrostatic Marker repulsion permeation
(c)
(d)
Redox marker
Electrode
Charged receptor
Analyte
Uncharged receptor Thiol-anchored charged molecule
(a)
Advantages ICSs’ characteristics offer the inherent capability of signal amplification. A sensor for any charged and electroinactive analytes can, in principle, be constructed by immobilizing the proper receptors on the electrode surface. The receptors are not necessarily limited to synthetic compounds; biological receptors, such as antibodies and protein receptors, can also be applied. The mode of signal transduction is unique: gating of the immobilized receptor membrane’s permeability to redox marker ions in response to charged analytes. Therefore, the method of signal transduction does not have to be changed or tailored for each analyte–receptor combination, in contrast with many conventional chemical or biosensors. Just like membrane potential changes for ion-selective electrodes, permeability changes for redox markers are the only means for signal transduction for ICSs.
Current (A)
4 2 0 –2 –4 0
[Protamine] / g/mL 4.44 2.88 2.83 2.78 0 –0.2 –0.4
[Ru(NH3)6]3+
Potential (V) vs Ag/AgCl
Construction Gold electrode
Protamine
(b)
4
Current (A)
The synthesis of ICSs is similar to that of chemically modified electrodes (3). In principle, any approach to immobilizing functional receptors or molecular recognition reagents on electrode surfaces will work. Approaches include covalent attachment of receptor molecules, thiol-anchored SAMs on gold, surface electropolymerization of receptor sites, electrostatic (charge–charge) adsorption, and Langmuir–Blodgett films. The main point in designing the sensor membranes is that they need to provide receptor functions for selective recognition of analytes and chargedanalyte-gated permeability changes for redox marker ions across receptor SAMs. The approach for the intramolecular mechanism (Figure 2a) requires particular receptor compounds that have intramolecular channels. Therefore, only a few ICS applications have been reported, including cyclodextrin monolayers for hydrophobic organic analytes (4, 5, 28) and -hemolysin, a monomeric polypeptide with a heptameric pore for pharmaceutical analytes (29). The approach for the intermolecular channel mechanism (Figure 2b – d) is generally applied to charged analytes (4, 5). The receptor is responsible for sensor selectivity, and the charged-analytegated permeability change is for signal transduction. Uncharged receptor SAMs are generally not void-free, therefore marker ions can permeate to a degree. But in the case of charged receptors, the marker ion with the same charge is repelled from the receptor surface because of long-range electrostatic repulsion, thereby allowing minimum permeability (Figure 2b). For this reason, a redox marker with a negative or positive charge (often [Fe(CN)6]3–/4– or [Ru(NH3)6 ] 3+/2+) is selected according to the membrane charge. Ideally, the analyte-driven permeability change should be detected as a switch from completely closed to open. If that is not possible, a small background is preferable so that the subsequent increase in permeability upon charged-analyte binding can be detected. If the situation is the other way around— such as large background currents in the absence of analytes and background currents that decrease when analytes are added (Figure 2c)—it is as if absorption spectroscopy were used instead of fluorescence spectroscopy. But it still works in many cases (16, 18, 23, 24).
0 [Heparin] / g/mL 4.04 2.53 2.02 1.52 0
–4
–8 0.6
0.4
[Mo(CN)8 ] 4– or [Fe(CN)6] 3–
0.2
Potential (V) vs Ag/AgCl
Heparin Gold electrode
FIGURE 3. (a) Cyclic voltammogram of the influence of the protamine concentration (0, 2.78, 2.83, 2.88, and 4.44 µg/mL in the direction of the arrow) 3+ on 1 mM [Ru(NH3)6] in 0.1 M KCl and 10 mM Tris HCl buffer at pH 7.4 on a gold electrode modified with SAMs of thioctic acid. A negative charge of a deprotonated carboxy group of thioctic acid induces protamine adsorption onto the electrode surface, decreasing the redox current of 3+ the positively charged marker [Ru(NH3)6] . (b) Cyclic voltammogram of the influence of the heparin concentration (0, 1.52, 2.02, 2.53, and 4.04 4– µg/mL in the direction of the arrow) on 1 mM [Mo(CN)8] in 0.1 M KCl and 10 mM Tris HCl buffer at pH 7.4 on a gold electrode modified with SAMs of thioctic acid treated with protamine. Binding of heparin to the SAM provides the electrode surface with negative charges, decreas4– ing the redox current of the negatively charged marker [Mo(CN)8] . S E P T E M B E R 1 , 2 0 0 4 / A N A LY T I C A L C H E M I S T R Y
323 A
to 10–10 M for oligonucleotides (27 ), and 10–10 to 10–3 M for biotin (18). The large complexation formation constants for receptors and analytes and the permeability changes of redox markers are required for attaining wider dynamic ranges and superior detection limits.
Recent trends Gold electrode Negative
Gold electrode Positive
FIGURE 4. ICS for the detection of trivalent metal cations. Illustration of gold electrodes modified with SAMs of thiol-derivatized 4-aryl5-pyrazolone receptors to trivalent metal ions (M3+) for the cationic marker [Ru(NH3)6]3+.
Neat uncharged receptor membranes tend to exhibit a large background marker transport, thereby interfering with sensitivity. In this case, mixed monolayers with uncharged receptors of thiol-anchored alkyl amines (HS–(CH2)n–NH+3) or sulfates (HS–(CH2)n – SO3–) as a charged filler molecule should help control sharp charged-analyte gating of the marker ions (Figure 2d). Thiol-anchored SAMs are recommended for receptor immobilization on gold electrodes. In the case of charged-analyte gating of the receptor’s permeability to marker ions, receptor SAMs should be controlled and optimized so that the maximum change in permeability is obtained. This is achieved by changing the ratio of the receptor to filler molecules. This ratio depends on whether the receptor is charged or uncharged. The permeability change is typically evaluated by the change in current intensities from redox markers by differential pulse or cyclic voltammetry, which measure analyte concentrations in sample solutions. In most clinical applications of ICSs, measurements are taken in buffer solutions that do not contain macromolecules such as proteins. However, measurements in artificial and horse sera are problematic because the response is weak (23) or deviated (24), or the response time is increased (30) compared with the response for buffer solutions without sera; in one case, analyte detection in serum was impossible (21). The difficulties are probably caused by adsorption of macromolecules from the sera to the electrode surface, which prevents analytes from binding to the recognition sites of the receptor monolayers. In many cases, ICSs can be used repeatedly because of the strong sulfur–gold bonds. A protamine ICS based on a SAM of thioctic acid reported by Gadzekpo et al. was very stable without any observable change in performance over four months, during which the electrodes were used to measure >450 cyclic voltammograms (23). Takaya et al. also reported that the ICS for trivalent cations based on a phosphate ester SAM showed no change in selectivity during several hundred measurements after being stored for 20 months (12). Dynamic ranges of analyte concentrations are 10–6 to 10–2 M (16) or 10–9 to 10–2 M for trivalent metal cations, 10–8 to 10– 4 M for estrogen (11), 10–6 to 10–3 M for cAMP (19), 10–15 324 A
A N A LY T I C A L C H E M I S T R Y / S E P T E M B E R 1 , 2 0 0 4
Protamine and heparin. Polyions were expected to bind more strongly to receptor monolayers than analytes with a small charge number, which would significantly affect the access of marker ions to the electrode surface. Also, polyions are attractive analytes because they are widely applied in many fields. For example, polysaccharides, dermatan sulfate, and -carrageenan are used in the food industry; polyphosphates are fertilizers; and oligonucleotides are central to biochemistry. Heparin is used extensively as an anticoagulant, and protamine is used to neutralize the anticoagulant activity of heparin. Whereas protamine is a polycation with an average charge of ~+20, heparin is a polyanion with a typical average charge of ~–70. To detect protamine, we modified gold electrodes with SAMs of thioctic acid, which has two sulfur atoms that bind the receptor to the electrode (23). Upon deprotonation, the carboxy group provides a negative charge that induces protamine adsorption onto the electrode surface (Figure 3a). These sensors could detect protamine concentrations as low as 0.5 µg/mL (1.1 10 –7 M) when [Ru(NH3)6]3+ was used as a marker. Their response to polybrene, another polycation that neutralizes the anticoagulant activity of heparin, was ~1000 smaller than their protamine response. This high protamine selectivity can be explained by considering the structural differences between the two polycations. The positive charge is distributed over the outer surface of protamine, while the charge in polybrene is buried, to some degree. Because washing with 0.1 M KCl at pH 5.1 removes protamine bound to the receptor layer, the sensors can be used repeatedly. Preliminary observations showed no observable changes in the performance of this electrode over four months during which this sensor was used for the measurement of 450 cyclic voltammograms. At physiological concentration, the blood electrolytes sodium, potassium, calcium, and magnesium did not interfere. Moreover, the sensors responded to protamine in diluted horse serum and were used to detect the end point in heparin– protamine titrations. Subsequent efforts aimed at developing an ICS to measure heparin itself (24). The oxidation of [Mo(CN)8] 4 – or reduction of [Fe(CN)6]3– in heparin solutions was investigated with SAMs of thioctic acid that had been treated with protamine (Figure 3b). Heparin, with its multiple negative charges, neutralizes the positive charges on the protamine receptor and, at high concentrations, provides the electrode surface with an excess of negative charges, thereby repulsing the marker ions [Mo(CN)8] 4– 3– and [Fe(CN)6] from the electrode surface. As a result, redox currents for these negatively charged markers decrease. In a solution containing background inorganic ions at concentrations typical in blood, a linear concentration range of 0.05–1.5 µg/mL (3.3 10–9–1.0 10–7 M) was determined. Repeated measurements of 1.25 µg/mL heparin in a physiological ion back-
(a)
(b)
[Ru(NH3)6]3+
Complementary oligonucleotide Probe PNA
3
3
3
3
Gold electrode Positive
3
3
Gold electrode Negative
(c) 15
Current (A)
[Oligonucleotide] / M 1.0 10 –8 1.0 10 –9 1.0 10 –10 1.0 10 –11 1.0 10 –12 1.0 10 –13 1.0 10 –14 1.0 10 –15 0
10
5
0 0
–0.1
–0.2
–0.3
–0.4
Potential (V) vs Ag/AgCl
FIGURE 5. Detecting oligonucleotides. ICS based on SAMs of a PNA probe and 8-amino-1-octanethiol, whose amine group is protonated in a pH 7.0 phosphate buffer; the electrode surface is therefore positively charged. (a) The redox reaction of marker [Ru(NH3)6]3+ is hindered by the electrostatic repulsion in the absence of the complementary oligonucleotide. (b) Hybridization between the PNA probe and the negatively charged complementary oligonucleotide cancels the positive charge at the surface and provides excess negative charge at the surface, thereby facilitating the redox reaction of the marker. (c) Osteryoung square-wave voltammogram of the influence of the complementary oligonucleotide concentration on 1 mM [Ru(NH3)6]3+.
ground and in horse serum gave average concentrations of 1.30 and 1.56 µg/mL, respectively. Using a fresh electrode for each horse serum sample, however, yielded an average concentration of 1.21 µg/mL with a standard deviation of 0.026 µg/mL. Trivalent metal cations. An ICS was made of gold electrodes modified with SAMs of thiol-derivatized 4-acyl-5-pyrazolones, which are a -diketone and a good extracting reagent with two oxygen atoms as coordinating centers for hard ions
such as trivalent lanthanoid ions (16). ICSs based on chelatormodified electrodes allowed the selective detection of trivalent metal cations on the basis of the high affinity of the chelator. The pH dependence of the cyclic voltammograms of [Fe(CN)6]3– and [Ru(NH3)6]3+ as electroactive markers suggested the protonation of not only the -diketone but also the nitrogen moiety in the pyrazolone. With regard to the voltammetric responses to metal cations, an increase and a decrease in the redox current of [Fe(CN)6]3–/4 – and [Ru(NH3)6]3+/2+, respectively, were observed with increasing concentrations of the trivalent cations La3+, Gd3+, Yb3+, and Al3+ from 10– 6–10– 4 M up to 10–2 M at pH 5.5 (Figure 4), at which the chelating group is present in its deprotonated form. In contrast, such responses were negligible in the presence of up to 10–2 M of the divalent cations Mg 2+, Ca2+, Sr2+, and Ba2+; and Li+ and Na+. The order of the magnitudes of responses was Al3+ > Yb3+, Gd3+ > La3+ >> divalent cations, which is quite similar to the stability of 1:1 and 1:2 complexes between a -diketonate-type chelate and the metal ions. The highly selective responses to trivalent cations seem to reflect the selectivity of the chelating group as well as the large change in the surface charge induced by the complexation. Estrogen and cAMP. The basic features of a bioelectronic device are the immobilization of a biomaterial, such as an antibody or an enzyme, onto a conductive or semiconductive support and the electronic transduction of the biological functions associated with the biological matrixes. However, it may be difficult to adopt this technique for constructing an estrogen biosensor using a nuclear receptor-modified electrode, because affinity reactions between the nuclear receptors and their ligands are not directly linked to redox processes. A bioaffinity sensor was developed with the goal of detecting estrogen on the basis of specific binding of estrogen to its receptor immobilized on a gold disk electrode (11). The recombinant DNA encoding a human estrogen receptor ligand-binding domain was expressed in bacteria by using the histidine-tag fusion system. The protein was immobilized on a gold electrode with Ni(II)-mediated chemisorption using a histidine tag and thiol-modified iminodiacetic acid. Cyclic voltammetric measurements showed that the presence of estrogen suppressed the reversible electrochemical reaction of a [Fe(CN)6] 3–/4– redox couple in a concentration-dependent manner. It seems reasonable to suppose that the complexation with estrogen altered the electrostatic property of the protein layer on the electrode surface. These data suggest that this biosensor can be used to evaluate binding activity toward the human estrogen receptor. A similar approach was used to construct a sensor for cAMP (19). A novel 17-mer peptide ligand for cAMP was designed using the amino acid sequences of essential subsites in various cAMP-dependent protein kinase A families. The gold disk electrode modified with the designed 17-mer oligopeptide responded to cAMP but not to any other cyclic nucleotides. However, a scrambled peptide, which had the same amino acid composition but an amino acid sequence different from the 17mer oligopeptide, did not respond to any nucleotides. Oligonucleotides. Gold electrodes modified with SAMs of a 13-mer peptide nucleic acid (PNA) probe and 8-amino-1S E P T E M B E R 1 , 2 0 0 4 / A N A LY T I C A L C H E M I S T R Y
325 A
octanethiol were used to detect a complementary oligonucleotide (27 ). No response was observed to a single-base mismatched oligonucleotide. The electrode surface was positively charged in a pH 7.0 buffer solution because of the protonation of an amine group of the thiol, where the electron transfer between the positively charged marker [Ru(NH3)6]3+ and the surface was hindered because of the charge–charge repulsion between them (Figure 5a). Binding of the negatively charged complementary oligonucleotide to the probe cancels the positive charge at the surface and provides an excess negative charge at the surface, thereby facilitating the access of the marker to the electrode surface and its redox reaction (Figure 5b). The use of uncharged single-stranded PNA instead of single-stranded DNA tremendously affected the sensor’s ability to obtain a high charge-gated switching of the [Ru(NH3)6]3+ redox reaction. Using a 13-mer PNA probe for this sensing mode allowed detection of the oligonucleotide at the femtomolar level. As the number (length) of oligonucleotides increased, the detection limit improved (26, 27 ). Avidin–biotin binding assay. Biotin, known as vitamin H, is an important substance for mammalian bodies because it takes part in the metabolism of fatty acids and sugars. Biotin combines specifically with avidin, a basic tetrameric glycoprotein in egg white. The avidin–biotin interaction has very high affinity (dissociation constant of 10–15 M) and is one of the strongest noncovalent biological interactions. Numerous functionally labeled compounds such as radiolabels, fluorescent ligands, and enzyme conjugates have been used to monitor avidin–biotin binding. However, a procedure is usually required before the measurement, to separate the free, labeled biotin from that bound with avidin. The separation process in protein–ligand assays not only complicates the operation but may also cause the binding equilibrium between the protein and its ligand to deviate. The ICS detection of avidin and biotin is based on monitoring the changes in the electrode response of the redox marker (18). Avidin binds specifically with biotin in its SAM on the electrode surface. At pH lower than the isoelectric point of avidin, the electrode response of ferrocyanide ions, used as an anionic redox marker, increased at the avidin-bound electrode because of the electrostatic interaction between the avidin on the electrode surface and the ferrocyanide ions. Avidin’s detection limit of 2.9 10–10 M is a result of the increase in the electrode response of the anionic marker ions.
To understand the general nature of ICSs, scientists should compare them with other sensors, such as surface plasmon resonance and quartz crystal microbalance sensors, for which binding events of receptor molecules on solids are transduced to a respective common signal such as a change in dielectric constant or mass. In the case of ICSs, the signal transduction is a charged-analyte-gated permeability change of electroactive marker ions across the receptor SAMs. Yoshio Umezawa is a professor at the University of Tokyo (Japan). Umezawa’s research interests include optical probes for intracellular signaling in living cells, molecular tips for chemically selective scanning tunneling microscopy, and ICSs. Hiroshi Aoki is currently a researcher at the National Institute of Advanced Industrial Science and Technology (Japan). His research interests include voltammetric sensors for bioactive molecules based on molecular recognition chemistries at membrane–solution interfaces. Address correspondence about this article to Umezawa at
[email protected].
References (1) (2)
(3) (4) (5) (6) (7) (8) (9) (10) (11) (12) (13) (14) (15) (16) (17) (18) (19) (20)
Outlook Intramolecular channel mechanisms have not been used widely because of the difficulty in designing and preparing the relevant receptor compounds (4, 5), although it was initially believed they showed promise for detecting uncharged molecules. On the contrary, the intermolecular channel mechanism seems to present no particular limitation for constructing chemical and biosensors for any charged analytes that are redox-inactive by themselves. Because we rely on receptors to improve analyte selectivity, signal transduction should not be specific but generally applicable to any analyte. 326 A
A N A LY T I C A L C H E M I S T R Y / S E P T E M B E R 1 , 2 0 0 4
(21) (22) (23) (24) (25) (26) (27) (28) (29) (30)
Sugawara, M.; et al. Anal. Chem. 1987, 59, 2842–2846. Odashima, K.; Sugawara, M.; Umezawa, Y. In Interfacial Design and Chemical Sensing; Mallouk, T. E., Harrison, D. J., Eds.; American Chemical Society: Washington, DC, 1994; ACS Symposium Series Vol. 561, pp 123–134. Kutner, W.; et al. Pure Appl. Chem. 1998, 70, 1301–1318. Bühlmann, P.; et al. Electroanal. 1998, 10, 1149–1158. Sugawara, M.; et al. Bull. Chem. Soc. Jpn. 2002, 75, 187–201. Moses, P. R.; Wier, L.; Murray, R. W. Anal. Chem. 1975, 47, 1882–1886. Nagase, S.; et al. Anal. Chem. 1990, 62, 1252–1259. Ulman, A. In An Introduction to Ultrathin Organic Films: From Langmuir– Blodgett to Self-Assembly ; Academic Press: Boston, MA, 1991. Finklea, H. O. In Electroanalytical Chemistry ; Bard, A. J., Rubinstein, I., Eds.; Marcel Dekker: New York, 1996; Vol. 19, pp 109–335. Odashima, K.; et al. Anal. Chem. 1993, 65, 927–936. Murata, M.; et al. Anal. Sci. 2001, 17, 387–390. Takaya, M.; Bühlmann, P.; Umezawa, Y. Mikrochim. Acta 1999, 132, 55. Miura, Y.; et al. Biopolymers 2000, 55, 391–398. Bandyopadhyay, K.; et al. J. Chem. Soc., Chem. Commun. 2000, 141–142. Flink, S.; et al. J. Chem. Soc., Perkin Trans. 2 2000, 2, 2141–2146. Ito, T. J. Electroanal. Chem. 2001, 495, 87–97. Luo, L.; Yang, X.; Wang, E. Anal. Lett. 1999, 32, 1271–1286. Kuramitz, H.; Sugawara, K.; Tanaka, S. Electroanal. 2000, 12, 1299–1303. Katayama, Y.; et al. Anal. Chem. 2000, 72, 4671–4674. Nakano, K.; et al. In Polymers in Sensors; Mallouk, T. E., Harrison, D. J., Eds.; American Chemical Society: Washington, DC, 1998; ACS Symposium Series Vol. 690, pp 34–45. Aoki, H.; et al. Biosens. Bioelectron. 2003, 18, 261–267. Aoki, H.; Bühlmann, P.; Umezawa, Y. J. Electroanal. Chem. 1999, 473, 105–112. Gadzekpo, V. P. Y.; et al. Anal. Chem. 1999, 71, 5109–5115. Gadzekpo, V. P. Y.; et al. Anal. Chim. Acta 2000, 411, 163–173. Aoki, H.; Bühlmann, P.; Umezawa, Y. Electroanal. 2000, 12, 1272–1276. Aoki, H.; Umezawa, Y. Electroanal. 2002, 14, 1405–1410. Aoki, H.; Umezawa, Y. Analyst 2003, 128, 681–685. Chamberlain, R. V., II; et al. Langmuir 2000, 16, 1388–1396. Gu, L.-Q.; et al. Nature 1999, 398, 686–690. Xiao, K. P.; Kim, B. Y.; Bruening, M. L. Electroanal. 2001, 13, 1447–1453.
