- Email: [email protected]
Development of dual receptor biosensors: an analysis of FRET pairs
Biosensors & Bioelectronics 16 (2001) 231– 237 www.elsevier.com/locate/bios
Development of dual receptor biosensors: an analysis of FRET pairs Sheila A. Grant *, Juntao Xu, Edward J. Bergeron, Jennifer Mroz Center for Biomedical Engineering, Michigan Technological Uni6ersity, 1400 Townsend Dri6e, Houghton, MI 49931, USA Received 27 June 2000; received in revised form 23 November 2000; accepted 13 December 2000
Abstract The development of a dual receptor detection method for enhanced biosensor monitoring was investigated by analyzing potential fluorescent resonance energy transfer (FRET) pairs. The dual receptor scheme requires the integration of a chemical transducer system with two unique protein receptors that bind to a single biological agent. The two receptors are tagged with special molecular groups (donors and acceptors fluorophores) while the chemical transduction system relies on the well-known mechanisms of FRET. During the binding event, the two FRET labeled receptors dock at the binding sites on the surface of the biological agent. The resulting close proximity of the two fluorophores upon binding will initiate the energy transfer resulting in fluorescence. The paper focuses on the analysis and optimization of the chemical transduction system. A variety of FRET fluorophore pairs were tested in a spectrofluorimeter and promising FRET pairs were then tagged to the protein, avidin and its ligand, biotin. Due to their affinities, the FRET-tagged biomolecules combine in solution, resulting in a stable, fluorescent signal from the acceptor FRET dye with a simultaneous decrease in fluorescent signal from the donor FRET dye. The results indicate that the selected FRET pairs can be utilized in the development of dual receptor sensors. © 2001 Elsevier Science B.V. All rights reserved. Keywords: FRET; Avidin; Biotin; Biosensor; Dual receptors; Conjugation
1. Introduction There is an important need in the public health and defense arenas for ultrasensitive biosensors capable of rapidly detecting dangerous biological agents (Mulchandani et al., 1998, 1999; Anderson et al., 1998; Seo et al., 1999). For example, these sensors might be capable of quickly reporting the presence of infectious airborne, waterborne or blood-borne microorganisms or microorganism-produced toxins. A major limitation of current optical biosensors is the background ‘noise’ due to non-specific binding which may lead to false positive readings (Ferguson et al., 1991; Perez-Luna et al., 1999). Researchers have focused on improving immobilization techniques in order to help eliminate or reduce non specific adsorption (Diederick and Losche, 1997; Kroger et al., 1999). For example, surface func* Corresponding author. Tel.: + 1-906-4871729; fax: +1-9064871717. E-mail address: [email protected] (S.A. Grant).
tionalization of self-assembled monolayers continue to be studied in order to reduce the effects of non-specific adsorption (Blankenburg et al., 1989; Spinke et al., 1993). Prime investigated the protein adsorption effects in self assembled monolayers of omega-functionalized long chain alkanethiolates in the hopes of better understanding the interactions between proteins with organic surfaces, while Perez-Luna investigated binding reaction between streptavidin and biotin and their role on surface properties, in order to understand nonspecific adsorption events (Prime and Whitesides, 1991; PerezLuna et al., 1999). To overcome the problem of nonspecific binding, researchers have adsorbed blocking agents, such as bovine serum albumin, gelatin and Triton X-100 onto the substrate (Bhatia et al., 1989; Diederick and Losche, 1997). However, there was still a measurable level of nonspecific binding. To eliminate, or at least dramatically lower the adverse background, we present a technique that involves incorporating multiple receptors. When the multiple receptors bind to the target analyte, they will elicit some form of unique
0956-5663/01/$ - see front matter © 2001 Elsevier Science B.V. All rights reserved. PII: S 0 9 5 6 - 5 6 6 3 ( 0 0 ) 0 0 1 4 7 - 0
232
S.A. Grant et al. / Biosensors & Bioelectronics 16 (2001) 231–237
biological activity (i.e. conformational change) that will allow detection and help eliminate or reduce false positives. Cell receptors are a feasible line of approach for novel biosensor development and have been developed into viable sensor systems (Canziani et al., 1999). Naturally occurring protein receptors are typically found in cell membranes and they contain structural and functional domains that penetrate the intracellular and extracellular spaces. Within the body, these receptors act as intercellular communication links by reversibly binding to specific neurotransmitters and hormones; but they may also act as binding sites for many drugs and toxins. As a result, cell receptors have been utilized in the development of a variety of biosensor systems, from nitric oxide sensors to understanding renal cellular interactions (Nakai et al., 1998; Barker et al., 1999; Newman et al., 1999). Most of these researchers exploited the affinity between the agent and a single receptor. However, in addition to single receptor systems, there are many biological agents that utilize multiple receptors that offers a unique approach for sensing when utilized with the chemical transduction system, FRET. In order to function properly, the dual receptor method requires the integration of a chemical transducer system, FRET, with two unique protein receptors, each capable of binding to a single biological agent. Researchers have utilized FRET in structural biology and biochemistry as a method to measure protein structures (Selvin, 1995). In addition, the principles of FRET have been used in optical sensor development (Chang and Sipior, 1995; Ballerstadt and Schultz, 1997; Pearce, 1998). For example, the FRET pair of sulforhodamine 101 and Bromocresol green, was used to develop NH3 sensors (Chang and Sipior, 1995), while the FITC (fluorescein-5-isothiocyanate) and TRITC (tetramethylrhodamine isothiocyanate) FRET pair was adopted in competitive– binding assays (Ballerstadt and Schultz, 1997). Pearce utilized the FRET pair, 7-diethylaminocoumarin-3-carboxylic acid and lissamine rhodamine B sulfonyl chloride, to detect Ni(II) (Pearce, 1998). Although FRET sensors do exist, there has not been significant research in the development of FRET based dual receptor sensors.
1.1. Principles of fluorescence resonance energy transfer (FRET) The principles of FRET are well known and have been widely used in biochemical and biomedical analysis for many years (Wu and Brand, 1994). FRET occurs when radiation is absorbed by a fluorophore (termed the donor) on a typical 10 − 15 s time scale. Following the donor’s energy absorption, one of two possible events may occur: (1) the energy may be re-emitted in
the form of fluorescence on a 10 − 9 s time scale; or (2) the excited molecule (donor) can transfer its energy to another fluorophore (termed the acceptor) which subsequently fluoresces in the same order of time. This latter phenomenon is fluorescence resonance energy transfer or FRET (Chantal and Lebrun, 1998). FRET has been explained by Fo¨ rster in terms of the relative dipole orientation of the donor and the acceptor and the distance between these two fluorophores (Fo¨ rster, 1948). The efficiency of energy transfer is greatly affected by these parameters. The energy transfer yield is given by Eq. (1): E= R 60/(R 6 + R 60)
(1)
where R is the distance between the donor and acceptor in a biological environment. R0 is a constant for each donor–acceptor pair and is defined as the distance that energy transfer, E is 50% efficient. R0 is called the Fo¨ rster critical distance. The transferred energy is depended on both the Fo¨ rster critical distance, which is characterized by the FRET pairs themselves and the distance between the donor and acceptor fluorophores; hence, a distance dependent method of detection. An additional method that is utilized to calculate transfer efficiency, E, is from the fluorescence intensities (quantum yields) using Eq. (2): E= 1− (Fda/Fd)
(2)
where Fda is the donor fluorescence intensity determined at a given wavelength in the presence of the acceptor and Fd is the corresponding quantity determined in the absence of the acceptor (Lakowicz, 1983). This method was used to roughly calculate the transfer efficiency of the FRET fluorophore pairs. The design of the dual receptor biosensor is depended upon achieving high transfer efficiencies; hence, a change in fluorescent intensity, induced by donor and acceptor molecular distances brought about by the agent, can be detected. This paper focuses on investigating and optimizing the chemical transduction system, FRET, for the development of dual receptor biosensors. A number of fluorophores were tested and identified as possible pairs. The pairs that demonstrated high spectral overlap and energy transfer were then tagged to the biomolecules, avidin and biotin. Avidin and biotin were utilized as a test bed to prove the feasibility of the proposed method. The specificity of biotin binding to avidin provides the basis for developing FRET dual receptor sensor systems.
2. Materials and methods
2.1. Materials All reagents were reagent grade unless otherwise
S.A. Grant et al. / Biosensors & Bioelectronics 16 (2001) 231–237
indicated. The dyes, fluorescein-5-isothiocyanate (FITC) and 7-dimethylaminocoumarin-4-acetic acid (DMACA) were obtained from Molecular Probes (Eugene, OR). The dye, tetramethylrhodamine isothiocyanate (TRITC) and the biomolecules, streptavidin, avidin and biotin, as well as fluorescein-tagged biotin, were obtained from Sigma Chemical (St. Louis, MO). An additional fluorophore, 7-amino-4-methyl-3-coumarinylacetic acid-N-hydroxysuccinimide, (AMCANHS) was obtained from Fluka (Milwaukee, WI). Additional reagents, phosphate buffered saline (PBS), dimethyl sulfoxide (DMSO) and 1-ethyl-3-(3-dimethylaminopropyl)carbodiimide hydrochloride (EDC), were also obtained from Sigma. The dyes were mixed with methanol (MeOH; Mallinckrodt, Paris, KY) or PBS to make solutions that ranged from 10 − 4 to 10 − 6 M.
233
that creates a thiourea linkage with avidin. The avidin solution was prepared in 0.1 M sodium carbonate, at a pH= 9.0 and a concentration of at least 2 mg/ml. The FITC was then dissolved in dry DMSO at a concentration of 1 mg/ml in a darkened room. To protect from extraneous light, the solution was wrapped in aluminum foil. A 100 ml FITC solution was slowly added to each milliliter of protein solution while gently mixing. The solution reacted for 8–12 h at 4°C in a darkened environment. After the reaction, the derivative was purified in a column filled with Sephadex G-25 (Sigma) in PBS buffer. In order to obtain complete separation, the column size was : 15–20 times the size of the applied sample (: 30 ml). The smaller, unbounded fluorescent molecules nonspecifically adhere to the gel filtration support allowing for the separation of free dye from the conjugated dye.
2.2. Methods 2.2.1. Optimization of FRET pairs The FRET pairs must possess high-energy transfer from the donor fluorophore to the acceptor fluorophore in order to achieve elevated sensitivities. That is, the emission peak of the donor fluorophore must have a high degree of overlap with the absorption peak of acceptor fluorophore. A number of fluorescent dyes were tested. Approximately 2 ml of each dye solution was placed in individual polystyrene cuvettes. An ISA SPEX spectrofluorimeter (ISA Instruments SA, Inc., Edison, NJ) was utilized to acquire the absorption and emission spectra for each dye. After examining the scans, possible FRET dye pairs were then selected for further testing. Again, using the ISA, tests were performed involving time-based scans. While acquiring data, the donor (or acceptor) fluorophore was added and mixed with the acceptor fluorophore (or donor). The resulting change in fluorescence was recorded. Similar concentrations were maintained at all times in order to avoid dilution effects. 2.2.2. Labeling FRET pairs to a6idin and biotin Because the FRET pairs, AMCA/FITC, DMACA/ FITC and FITC/TRITC, demonstrated high spectral overlap, they were tagged to avidin and biotin. Biotin has very high affinity for avidin with a 10 − 15 disassociation constant, while each avidin molecule contains a maximum of four biotin-binding sites. Due to their high affinity, the biomolecules are able to bind to each other in solution. Details on the tagging protocol can be found elsewhere (Hermanson, 1996). Briefly, the procedures are summarized below. 2.2.3. Tagging FITC to a6idin The dye, FITC, contains an isothiocyanate group
2.2.4. Tagging AMCA-NHS to strepta6idin AMCA-NHS is an amine reactive derivative of AMCA containing an NHS ester on its carboxylate group. The result is direct reactivity toward amine-containing molecules, forming amide linkages with the AMCA fluorophore. In this case, streptavidin was prepared for conjugation. Streptavidin was dissolve in 50 mM sodium borate, pH 8.5 at a 10 mg/ml concentration. AMCA-NHS was dissolved in DMSO at a 2.6mg/ml concentration and was protected from light. AMCA-NHS was slowly added to the streptavidin solution in 100 ml increments to each milliliter of streptavidin solution. The reaction was concluded in 1 h. After the reaction, the solution was purified in a column filled with Sephadex G-25 (Sigma) in PBS buffer. 2.2.5. Tagging DMACA to a6idin Since DMACA can not be directly tagged to avidin, it was modified by introducing a chemical linker, 1ethyl-3-(3-dimethylaminopropyl)carbodiimide hydrochloride (EDC). EDC activated the carboxylate group on the DMACA to a highly reactive O-acylisourea intermediate. Derivation of DMACA from its carboxylate group caused no major effects on its fluorescent properties. The same procedure was followed as described above. 2.2.6. Tagging TRITC to biotin TRITC contains an isothiocyanate group that creates a thiourea linkage with biotin. The biotin solution was prepared in 0.1 M sodium carbonate, at a pH= 9.0 and a concentration of at least 2 mg/ml. The TRITC was then dissolved in dry DMSO at a concentration of 1 mg/ml in a darkened room. Next, the same procedure as tagging FITC to avidin was followed.
234
S.A. Grant et al. / Biosensors & Bioelectronics 16 (2001) 231–237
2.2.7. Tagging FITC to biotin In an effort to simplify procedures, FITC-tagged biotin was purchased from Molecular Probes. Information on the tagging procedure can be obtained from the Molecular Probes Handbook (Haugland, 1996). 2.3. Testing the FRET biosensor technique After labeling the FRET pair dyes with avidin (or streptavidin) and biotin, they were tested in an ISA SPEX spectrofluorimeter. Maintaining the same concentrations, 1 ml of the fluorophore-conjugated avidin was mixed with 1 ml of the fluorophore-conjugated biotin. Scans of the individual and mixed solutions were acquired.
3. Results and discussion
3.1. Determination of FRET pairs for biosensor applications A number of FRET dye pairs were examined which resulted in three FRET pairs having sufficiently high spectral overlaps. They were, DMACA/FITC, AMCA/ FITC and FITC/TRITC. Fig. 1 shows the excitation and emission scans of the fluorophore pair, FITC/ TRITC. FITC acted as the donor while TRITC was the acceptor fluorophore. FITC was excited at 495 nm and had an emission at 523 nm while TRITC had an excitation peak at 554 nm and an emission peak at 580 nm. The donor’s emission peak overlapped the excitation peak of the TRITC; hence energy transfer between the two fluorophores should occur when mixed together. In addition, Fig. 1 demonstrates one of the concerns of utilizing FRET as a detection method; the possibility
Fig. 1. Excitation and emission scans of the FRET pair, FITC and TRITC: excitation scan of FITC (a), emission scan of FITC (b), excitation scan of TRITC (c) and emission scan of TRITC (d).
Fig. 2. Emission spectrum of the individual fluorophores, DMACA (a), FITC (b) and a mixture of DMACA +FITC (c). The excitation wavelength was set at 373 nm.
of the tail of the donor’s excitation wavelength overlapping the absorption wavelength of the acceptor dye, resulting in an unwanted emission peak from the acceptor dye. This spectral overlap can be minimized through time discrimination of the signals and/or complex signal conditioning. Another approach involves tailoring the excitation wavelength of the donor dye, so that the overlap is minimized while still maintaining sufficient energy transfer. Fig. 2 shows the results of the FRET fluorophore pair, DMACA and FITC, in free solution of PBS. The DMACA acted as the donor dye with a 373 nm absorption peak and an emission peak at 457 nm, while the FITC was the acceptor dye having an absorption peak at 480 nm and an emission peak at 535 nm. Utilizing an excitation wavelength of 373 nm, Fig. 2 shows separate emission scans for free DMACA, free FITC and a mixture of FITC and DMACA. With a 373-nm emission, FITC demonstrated a small peak at 535 nm. This peak was small however, compared to the ‘mixture’ peak. When the FRET pair was mixed, a large emission peak occurred at : 535 nm, while the emission peak of the DMACA at : 460 nm was reduced. Subsequently, when the FRET dyes were mixed, a radiation-less transfer of excitation energy from the donor to the acceptor fluorophore was demonstrated. Roughly, the transfer efficiency can be determined using Eq. (2). However, Eq. (2) was derived under the assumption that the donor–acceptor pair was separated by a fixed distance. For a mixture of acceptors and donors in solution, more complex expressions are needed. These involved expressions are generally derived by averaging the transfer rate over the assumed spatial distribution of the donor–acceptor pair. For simplicity, the donor– acceptors in solution were assumed randomly stationary and the transfer efficiency was roughly predicted using Eq. (2). E= 1− (0.01/0.95)
S.A. Grant et al. / Biosensors & Bioelectronics 16 (2001) 231–237
The transfer efficiency was calculated on the donor fluorophore to be 99%. Time scans of DMACA and FITC fluorophores are shown in Fig. 3. The concentrations of the dyes were kept constant at 10 − 4 M. Fig. 3 shows two scans plotted on the same graph. The first scan consisted of setting the excitation wavelength at 373 nm and then monitoring the fluorescence at the 457 nm emission peak for the DMACA dye as the FITC dye was added. This data is shown on the left axis. The second scan consisted of, again, setting the excitation wavelength at 373 nm and then monitoring the fluorescence at the 535 nm emission peak for the FITC dye as DMACA was added to the cuvette. The right axis displays the intensity. Data from the first scan displays the donor’s fluorescence (DMACA) decreasing as the acceptor dye (FITC) was added. This is the ‘quenching’ effect that DMACA exhibits as the dye’s energy is transferred to the acceptor dye, FITC. On the right side of the graph, the plot shows that the acceptor’s (FITC) fluorescence was increased as the DMACA is added to the solution. This corresponds to the energy transfer taking place from the donor dye to the acceptor dye. In addition, the amount of energy transferred is not equal. Conservation of energy requires that the energy of the electronic state for the acceptor molecule must be either the same or less than that for the donor molecule. The excess of electronic energy upon transfer may be dissipated into vibrational energy. The transfer process, in effect, competes with other modes of de-excitation, including direct emission of photons with energy appropriate to the electronic state of the donor (Lakowicz, 1991). Fig. 4 displays the individual emission scans of AMCA-NHS and FITC and the mixture scan of these two FRET dyes. The AMCA-NHS acted as the donor dye with an absorption peak at 347 nm and emission peak at 450 nm, while the FITC was the acceptor dye with an emission peak at 520 nm. From the individual
Fig. 3. Time based scans of DMACA monitored at 457 nm with additions of FITC (— ) and FITC monitored at 535 nm with additions of DMACA (- - -).
235
Fig. 4. Emission spectrum of the individual fluorophores, AMCANHS (a) and FITC (b) and a mixture of AMCA-NHS +FITC (c). The excitation wavelength was set at 343 nm.
scan of the FITC fluorophore, there was a small emission peak resulting from the donor’s excitation wavelength. This indicates that only a small portion of the excitation wavelength of the AMCA-NHS dye will overlap the excitation wavelength of the FITC dye. When the dyes were mixed, a decrease in the emission peak of the AMCA-NHS dye at :450 nm resulted, while the fluorescent peak of the FITC dye at 520 nm increased. Therefore, it was demonstrated that a radiation-less transfer of excitation energy from the AMCANHS to the FITC dye occurred. Again, for simplicity, the donor –acceptor pairs in solution were assumed stationary and the transfer efficiency was roughly predicted using Eq. (2). The transfer efficiency was calculated to be 63%.
3.2. Conjugation of FRET pairs to biomolecules AMCA-NHS was tagged to streptavidin, while FITC was tagged to biotin. Fig. 5 shows the individual emis-
Fig. 5. Emission spectrum of the conjugated fluorophores, AMCANHS/streptavidin (a) and FITC/biotin (b) and a mixture of AMCANHS/streptavidin and FITC/biotin (c). The excitation wavelength was set at 343 nm.
236
S.A. Grant et al. / Biosensors & Bioelectronics 16 (2001) 231–237
4. Conclusion
Fig. 6. Emission spectrum of conjugated molecules FITC-tagged avidin (a), TRITC-tagged biotin (b) and a mixture (c), excited at 440 nm.
sion scans of the FRET labeled pairs and also the scan of the pair after they were mixed. There was a quenching of the donor peak, while the acceptor peak increased. When comparing Fig. 5 to Fig. 4, the AMCA-NHS peak was quenched to a greater degree for the streptavidin and biotin conjugates in comparison to the free dye conditions. Indeed, the energy transfer was calculated to be 88%, thus demonstrating a significant increase in transfer efficiency of :29%. In the case of the unconjugated dyes, the FRET pairs are randomly assembled, but with the streptavidin and biotin conjugates, the FRET pairs are assembled as a result of the high affinity between avidin and biotin. The energy transfer between the FRET pair is higher and more stable. Fig. 6 shows individual emission scans of TRITCtagged avidin and FITC-tagged biotin being excited at 440 nm. When the two were mixed together, there was a significant decrease in the donor’s emission. However, it was noted that there was no significant increase in the acceptor’s fluorescence. This could be due to several possible factors. One is that the energy was transferred into the form of vibrational energy to the surrounding protein structures, not to the acceptor dye. A more plausible reason is due to the general avidin and biotin binding structure. The acceptor dye was conjugated to the avidin molecule, while the donor dye was conjugated to the biotin molecule. There are four biotinbinding sites for every one avidin protein. Therefore, there is a 4:1 ratio between the donor dye and the acceptor dye. The concentration of the acceptor dye upon binding may have been too low to cause a significant increase in its fluorescence. To compensate for this problem, future experiments are planned whereby the acceptor dyes will be conjugated to biotin while the donor dyes will be conjugated to avidin.
This paper detailed the investigation and optimization of the FRET chemical transduction system. A number of fluorophores were tested as potential FRET pairs. The promising FRET pairs were further tested and conjugated to the biomolecules, streptavidin, avidin and biotin. These pairs were DMACA/FITC, AMCA/ FITC and FITC/TRITC. This identification and fluorescent labeling was the first step in developing enhanced biosensors that utilize the FRET dual receptor method. Because of their affinities, the FRETtagged biomolecules combine in solution resulting in a more stable, elevated fluorescent signal. The results indicate that the FRET pairs can be utilized in the development of a dual receptor sensor. Tests are currently being performed to further demonstrate the dual receptor method.
Acknowledgements The project was funded through a State of Michigan Research Excellence Fund (REF) and by the Whitaker Foundation Special Opportunity Grant.
References Anderson, G.P., King, K.D., Cao, L.K., Jacoby, M., Ligler, F.S., Ezzell, J., 1998. Clin. Diagn. Lab. Immunol. 5, 609. Ballerstadt, R., Schultz, J.S., 1997. Anal. Chem. Acta 345, 203. Barker, S.L., Zhao, Y., Marletta, M.A., Kopelman, R., 1999. Anal. Chem. 71, 2071. Bhatia, S.K., Shriver-Lake, L.C., Prior, K.J., Georger, J.H., Calvert, J.M., Bredehorst, R., Ligler, F.S., 1989. Anal. Biochem. 178, 408. Blankenburg, R., Meller, P., Ringsdorf, H., Salesse, C., 1989. Biochemistry 28, 8214. Canziani, G., Zhang, W., Cines, D., Rux, A., Willis, S., Cohen, G., Eisenberg, R., Chaiken, I., 1999. Methods 19, 253. Chang, Q., Sipior, J., 1995. Anal. Biochem. 232, 92. Chantal, M., Lebrun, C., 1998. Biochem. Edu. 26, 320 – 323. Diederick, A., Losche, M., 1997. Adv. Biophys. 34, 205. Ferguson, G.S., Chaudury, M.K., Sigal, G.B., Whitesides, G.M., 1991. Science 253, 776. Fo¨ rster, T., 1948. Ann. Phys. 2, 55. Haugland, R.P., 1996. Handbook on Fluorescent Probes and Research Chemicals. Molecular Probes, Eugene, OR. Herman, G.T., 1996. Bioconjugate Techniques. Academic Press, San Diego, CA. Kroger, D., Liley, M., Schiweck, W., Skerra, A., Vogel, H., 1999. Biosens. Bioelec. 14, 155. Lakowicz, J.R., 1983. Principles of Fluorescence Spectroscopy. Plenum Press, New York. Lakowicz, J.R., 1991. Topics in Fluorescence Spectroscopy, vol. 2. Plenum Press, New York. Mulchandani, A., Kaneva, I., Chen, W., 1998. Anal. Chem. 70, 5042. Mulchandani, A., Pan, S., Chen, W., 1999. Biotechnol. Prog. 15, 130. Nakai, J., Sekiguchi, N., Rando, T.A., Allen, P.D., Beam, K.G., 1998. J. Biol. Chem. 273, 13403.
S.A. Grant et al. / Biosensors & Bioelectronics 16 (2001) 231–237 Newman, D.J., Thakkar, H., Lam-Po-Tang, M.K., Kwan, J.T., 1999. Ren. Fail. 21, 349. Pearce, D.A., 1998. Biorgan. Med. Chem. Lett. 8, 1963. Perez-Luna, V.H., O’Brien, M.J., Opperman, K.A., Hampton, P.D., Lopez, G.P., Klumb, L.A., Stayton, P.S., 1999. J. Am. Chem. Soc. 121, 6469.
.
237
Prime, K.L., Whitesides, G.M., 1991. Science 252, 1164. Selvin, P.R., 1995. Methods Enzymol. 246, 300. Seo, K.H., Brackett, R.E., Hartman, N.F., Campbell, D.P., 1999. J. Food Prot. 62, 431. Spinke, J., Liley, M., Schmitt, F.J., 1993. J. Chem. Phys. 99, 7012. Wu, P., Brand, L., 1994. Anal. Biochem. 218, 1.
Development of dual receptor biosensors: an analysis of FRET pairs Sheila A. Grant *, Juntao Xu, Edward J. Bergeron, Jennifer Mroz Center for Biomedical Engineering, Michigan Technological Uni6ersity, 1400 Townsend Dri6e, Houghton, MI 49931, USA Received 27 June 2000; received in revised form 23 November 2000; accepted 13 December 2000
Abstract The development of a dual receptor detection method for enhanced biosensor monitoring was investigated by analyzing potential fluorescent resonance energy transfer (FRET) pairs. The dual receptor scheme requires the integration of a chemical transducer system with two unique protein receptors that bind to a single biological agent. The two receptors are tagged with special molecular groups (donors and acceptors fluorophores) while the chemical transduction system relies on the well-known mechanisms of FRET. During the binding event, the two FRET labeled receptors dock at the binding sites on the surface of the biological agent. The resulting close proximity of the two fluorophores upon binding will initiate the energy transfer resulting in fluorescence. The paper focuses on the analysis and optimization of the chemical transduction system. A variety of FRET fluorophore pairs were tested in a spectrofluorimeter and promising FRET pairs were then tagged to the protein, avidin and its ligand, biotin. Due to their affinities, the FRET-tagged biomolecules combine in solution, resulting in a stable, fluorescent signal from the acceptor FRET dye with a simultaneous decrease in fluorescent signal from the donor FRET dye. The results indicate that the selected FRET pairs can be utilized in the development of dual receptor sensors. © 2001 Elsevier Science B.V. All rights reserved. Keywords: FRET; Avidin; Biotin; Biosensor; Dual receptors; Conjugation
1. Introduction There is an important need in the public health and defense arenas for ultrasensitive biosensors capable of rapidly detecting dangerous biological agents (Mulchandani et al., 1998, 1999; Anderson et al., 1998; Seo et al., 1999). For example, these sensors might be capable of quickly reporting the presence of infectious airborne, waterborne or blood-borne microorganisms or microorganism-produced toxins. A major limitation of current optical biosensors is the background ‘noise’ due to non-specific binding which may lead to false positive readings (Ferguson et al., 1991; Perez-Luna et al., 1999). Researchers have focused on improving immobilization techniques in order to help eliminate or reduce non specific adsorption (Diederick and Losche, 1997; Kroger et al., 1999). For example, surface func* Corresponding author. Tel.: + 1-906-4871729; fax: +1-9064871717. E-mail address: [email protected] (S.A. Grant).
tionalization of self-assembled monolayers continue to be studied in order to reduce the effects of non-specific adsorption (Blankenburg et al., 1989; Spinke et al., 1993). Prime investigated the protein adsorption effects in self assembled monolayers of omega-functionalized long chain alkanethiolates in the hopes of better understanding the interactions between proteins with organic surfaces, while Perez-Luna investigated binding reaction between streptavidin and biotin and their role on surface properties, in order to understand nonspecific adsorption events (Prime and Whitesides, 1991; PerezLuna et al., 1999). To overcome the problem of nonspecific binding, researchers have adsorbed blocking agents, such as bovine serum albumin, gelatin and Triton X-100 onto the substrate (Bhatia et al., 1989; Diederick and Losche, 1997). However, there was still a measurable level of nonspecific binding. To eliminate, or at least dramatically lower the adverse background, we present a technique that involves incorporating multiple receptors. When the multiple receptors bind to the target analyte, they will elicit some form of unique
0956-5663/01/$ - see front matter © 2001 Elsevier Science B.V. All rights reserved. PII: S 0 9 5 6 - 5 6 6 3 ( 0 0 ) 0 0 1 4 7 - 0
232
S.A. Grant et al. / Biosensors & Bioelectronics 16 (2001) 231–237
biological activity (i.e. conformational change) that will allow detection and help eliminate or reduce false positives. Cell receptors are a feasible line of approach for novel biosensor development and have been developed into viable sensor systems (Canziani et al., 1999). Naturally occurring protein receptors are typically found in cell membranes and they contain structural and functional domains that penetrate the intracellular and extracellular spaces. Within the body, these receptors act as intercellular communication links by reversibly binding to specific neurotransmitters and hormones; but they may also act as binding sites for many drugs and toxins. As a result, cell receptors have been utilized in the development of a variety of biosensor systems, from nitric oxide sensors to understanding renal cellular interactions (Nakai et al., 1998; Barker et al., 1999; Newman et al., 1999). Most of these researchers exploited the affinity between the agent and a single receptor. However, in addition to single receptor systems, there are many biological agents that utilize multiple receptors that offers a unique approach for sensing when utilized with the chemical transduction system, FRET. In order to function properly, the dual receptor method requires the integration of a chemical transducer system, FRET, with two unique protein receptors, each capable of binding to a single biological agent. Researchers have utilized FRET in structural biology and biochemistry as a method to measure protein structures (Selvin, 1995). In addition, the principles of FRET have been used in optical sensor development (Chang and Sipior, 1995; Ballerstadt and Schultz, 1997; Pearce, 1998). For example, the FRET pair of sulforhodamine 101 and Bromocresol green, was used to develop NH3 sensors (Chang and Sipior, 1995), while the FITC (fluorescein-5-isothiocyanate) and TRITC (tetramethylrhodamine isothiocyanate) FRET pair was adopted in competitive– binding assays (Ballerstadt and Schultz, 1997). Pearce utilized the FRET pair, 7-diethylaminocoumarin-3-carboxylic acid and lissamine rhodamine B sulfonyl chloride, to detect Ni(II) (Pearce, 1998). Although FRET sensors do exist, there has not been significant research in the development of FRET based dual receptor sensors.
1.1. Principles of fluorescence resonance energy transfer (FRET) The principles of FRET are well known and have been widely used in biochemical and biomedical analysis for many years (Wu and Brand, 1994). FRET occurs when radiation is absorbed by a fluorophore (termed the donor) on a typical 10 − 15 s time scale. Following the donor’s energy absorption, one of two possible events may occur: (1) the energy may be re-emitted in
the form of fluorescence on a 10 − 9 s time scale; or (2) the excited molecule (donor) can transfer its energy to another fluorophore (termed the acceptor) which subsequently fluoresces in the same order of time. This latter phenomenon is fluorescence resonance energy transfer or FRET (Chantal and Lebrun, 1998). FRET has been explained by Fo¨ rster in terms of the relative dipole orientation of the donor and the acceptor and the distance between these two fluorophores (Fo¨ rster, 1948). The efficiency of energy transfer is greatly affected by these parameters. The energy transfer yield is given by Eq. (1): E= R 60/(R 6 + R 60)
(1)
where R is the distance between the donor and acceptor in a biological environment. R0 is a constant for each donor–acceptor pair and is defined as the distance that energy transfer, E is 50% efficient. R0 is called the Fo¨ rster critical distance. The transferred energy is depended on both the Fo¨ rster critical distance, which is characterized by the FRET pairs themselves and the distance between the donor and acceptor fluorophores; hence, a distance dependent method of detection. An additional method that is utilized to calculate transfer efficiency, E, is from the fluorescence intensities (quantum yields) using Eq. (2): E= 1− (Fda/Fd)
(2)
where Fda is the donor fluorescence intensity determined at a given wavelength in the presence of the acceptor and Fd is the corresponding quantity determined in the absence of the acceptor (Lakowicz, 1983). This method was used to roughly calculate the transfer efficiency of the FRET fluorophore pairs. The design of the dual receptor biosensor is depended upon achieving high transfer efficiencies; hence, a change in fluorescent intensity, induced by donor and acceptor molecular distances brought about by the agent, can be detected. This paper focuses on investigating and optimizing the chemical transduction system, FRET, for the development of dual receptor biosensors. A number of fluorophores were tested and identified as possible pairs. The pairs that demonstrated high spectral overlap and energy transfer were then tagged to the biomolecules, avidin and biotin. Avidin and biotin were utilized as a test bed to prove the feasibility of the proposed method. The specificity of biotin binding to avidin provides the basis for developing FRET dual receptor sensor systems.
2. Materials and methods
2.1. Materials All reagents were reagent grade unless otherwise
S.A. Grant et al. / Biosensors & Bioelectronics 16 (2001) 231–237
indicated. The dyes, fluorescein-5-isothiocyanate (FITC) and 7-dimethylaminocoumarin-4-acetic acid (DMACA) were obtained from Molecular Probes (Eugene, OR). The dye, tetramethylrhodamine isothiocyanate (TRITC) and the biomolecules, streptavidin, avidin and biotin, as well as fluorescein-tagged biotin, were obtained from Sigma Chemical (St. Louis, MO). An additional fluorophore, 7-amino-4-methyl-3-coumarinylacetic acid-N-hydroxysuccinimide, (AMCANHS) was obtained from Fluka (Milwaukee, WI). Additional reagents, phosphate buffered saline (PBS), dimethyl sulfoxide (DMSO) and 1-ethyl-3-(3-dimethylaminopropyl)carbodiimide hydrochloride (EDC), were also obtained from Sigma. The dyes were mixed with methanol (MeOH; Mallinckrodt, Paris, KY) or PBS to make solutions that ranged from 10 − 4 to 10 − 6 M.
233
that creates a thiourea linkage with avidin. The avidin solution was prepared in 0.1 M sodium carbonate, at a pH= 9.0 and a concentration of at least 2 mg/ml. The FITC was then dissolved in dry DMSO at a concentration of 1 mg/ml in a darkened room. To protect from extraneous light, the solution was wrapped in aluminum foil. A 100 ml FITC solution was slowly added to each milliliter of protein solution while gently mixing. The solution reacted for 8–12 h at 4°C in a darkened environment. After the reaction, the derivative was purified in a column filled with Sephadex G-25 (Sigma) in PBS buffer. In order to obtain complete separation, the column size was : 15–20 times the size of the applied sample (: 30 ml). The smaller, unbounded fluorescent molecules nonspecifically adhere to the gel filtration support allowing for the separation of free dye from the conjugated dye.
2.2. Methods 2.2.1. Optimization of FRET pairs The FRET pairs must possess high-energy transfer from the donor fluorophore to the acceptor fluorophore in order to achieve elevated sensitivities. That is, the emission peak of the donor fluorophore must have a high degree of overlap with the absorption peak of acceptor fluorophore. A number of fluorescent dyes were tested. Approximately 2 ml of each dye solution was placed in individual polystyrene cuvettes. An ISA SPEX spectrofluorimeter (ISA Instruments SA, Inc., Edison, NJ) was utilized to acquire the absorption and emission spectra for each dye. After examining the scans, possible FRET dye pairs were then selected for further testing. Again, using the ISA, tests were performed involving time-based scans. While acquiring data, the donor (or acceptor) fluorophore was added and mixed with the acceptor fluorophore (or donor). The resulting change in fluorescence was recorded. Similar concentrations were maintained at all times in order to avoid dilution effects. 2.2.2. Labeling FRET pairs to a6idin and biotin Because the FRET pairs, AMCA/FITC, DMACA/ FITC and FITC/TRITC, demonstrated high spectral overlap, they were tagged to avidin and biotin. Biotin has very high affinity for avidin with a 10 − 15 disassociation constant, while each avidin molecule contains a maximum of four biotin-binding sites. Due to their high affinity, the biomolecules are able to bind to each other in solution. Details on the tagging protocol can be found elsewhere (Hermanson, 1996). Briefly, the procedures are summarized below. 2.2.3. Tagging FITC to a6idin The dye, FITC, contains an isothiocyanate group
2.2.4. Tagging AMCA-NHS to strepta6idin AMCA-NHS is an amine reactive derivative of AMCA containing an NHS ester on its carboxylate group. The result is direct reactivity toward amine-containing molecules, forming amide linkages with the AMCA fluorophore. In this case, streptavidin was prepared for conjugation. Streptavidin was dissolve in 50 mM sodium borate, pH 8.5 at a 10 mg/ml concentration. AMCA-NHS was dissolved in DMSO at a 2.6mg/ml concentration and was protected from light. AMCA-NHS was slowly added to the streptavidin solution in 100 ml increments to each milliliter of streptavidin solution. The reaction was concluded in 1 h. After the reaction, the solution was purified in a column filled with Sephadex G-25 (Sigma) in PBS buffer. 2.2.5. Tagging DMACA to a6idin Since DMACA can not be directly tagged to avidin, it was modified by introducing a chemical linker, 1ethyl-3-(3-dimethylaminopropyl)carbodiimide hydrochloride (EDC). EDC activated the carboxylate group on the DMACA to a highly reactive O-acylisourea intermediate. Derivation of DMACA from its carboxylate group caused no major effects on its fluorescent properties. The same procedure was followed as described above. 2.2.6. Tagging TRITC to biotin TRITC contains an isothiocyanate group that creates a thiourea linkage with biotin. The biotin solution was prepared in 0.1 M sodium carbonate, at a pH= 9.0 and a concentration of at least 2 mg/ml. The TRITC was then dissolved in dry DMSO at a concentration of 1 mg/ml in a darkened room. Next, the same procedure as tagging FITC to avidin was followed.
234
S.A. Grant et al. / Biosensors & Bioelectronics 16 (2001) 231–237
2.2.7. Tagging FITC to biotin In an effort to simplify procedures, FITC-tagged biotin was purchased from Molecular Probes. Information on the tagging procedure can be obtained from the Molecular Probes Handbook (Haugland, 1996). 2.3. Testing the FRET biosensor technique After labeling the FRET pair dyes with avidin (or streptavidin) and biotin, they were tested in an ISA SPEX spectrofluorimeter. Maintaining the same concentrations, 1 ml of the fluorophore-conjugated avidin was mixed with 1 ml of the fluorophore-conjugated biotin. Scans of the individual and mixed solutions were acquired.
3. Results and discussion
3.1. Determination of FRET pairs for biosensor applications A number of FRET dye pairs were examined which resulted in three FRET pairs having sufficiently high spectral overlaps. They were, DMACA/FITC, AMCA/ FITC and FITC/TRITC. Fig. 1 shows the excitation and emission scans of the fluorophore pair, FITC/ TRITC. FITC acted as the donor while TRITC was the acceptor fluorophore. FITC was excited at 495 nm and had an emission at 523 nm while TRITC had an excitation peak at 554 nm and an emission peak at 580 nm. The donor’s emission peak overlapped the excitation peak of the TRITC; hence energy transfer between the two fluorophores should occur when mixed together. In addition, Fig. 1 demonstrates one of the concerns of utilizing FRET as a detection method; the possibility
Fig. 1. Excitation and emission scans of the FRET pair, FITC and TRITC: excitation scan of FITC (a), emission scan of FITC (b), excitation scan of TRITC (c) and emission scan of TRITC (d).
Fig. 2. Emission spectrum of the individual fluorophores, DMACA (a), FITC (b) and a mixture of DMACA +FITC (c). The excitation wavelength was set at 373 nm.
of the tail of the donor’s excitation wavelength overlapping the absorption wavelength of the acceptor dye, resulting in an unwanted emission peak from the acceptor dye. This spectral overlap can be minimized through time discrimination of the signals and/or complex signal conditioning. Another approach involves tailoring the excitation wavelength of the donor dye, so that the overlap is minimized while still maintaining sufficient energy transfer. Fig. 2 shows the results of the FRET fluorophore pair, DMACA and FITC, in free solution of PBS. The DMACA acted as the donor dye with a 373 nm absorption peak and an emission peak at 457 nm, while the FITC was the acceptor dye having an absorption peak at 480 nm and an emission peak at 535 nm. Utilizing an excitation wavelength of 373 nm, Fig. 2 shows separate emission scans for free DMACA, free FITC and a mixture of FITC and DMACA. With a 373-nm emission, FITC demonstrated a small peak at 535 nm. This peak was small however, compared to the ‘mixture’ peak. When the FRET pair was mixed, a large emission peak occurred at : 535 nm, while the emission peak of the DMACA at : 460 nm was reduced. Subsequently, when the FRET dyes were mixed, a radiation-less transfer of excitation energy from the donor to the acceptor fluorophore was demonstrated. Roughly, the transfer efficiency can be determined using Eq. (2). However, Eq. (2) was derived under the assumption that the donor–acceptor pair was separated by a fixed distance. For a mixture of acceptors and donors in solution, more complex expressions are needed. These involved expressions are generally derived by averaging the transfer rate over the assumed spatial distribution of the donor–acceptor pair. For simplicity, the donor– acceptors in solution were assumed randomly stationary and the transfer efficiency was roughly predicted using Eq. (2). E= 1− (0.01/0.95)
S.A. Grant et al. / Biosensors & Bioelectronics 16 (2001) 231–237
The transfer efficiency was calculated on the donor fluorophore to be 99%. Time scans of DMACA and FITC fluorophores are shown in Fig. 3. The concentrations of the dyes were kept constant at 10 − 4 M. Fig. 3 shows two scans plotted on the same graph. The first scan consisted of setting the excitation wavelength at 373 nm and then monitoring the fluorescence at the 457 nm emission peak for the DMACA dye as the FITC dye was added. This data is shown on the left axis. The second scan consisted of, again, setting the excitation wavelength at 373 nm and then monitoring the fluorescence at the 535 nm emission peak for the FITC dye as DMACA was added to the cuvette. The right axis displays the intensity. Data from the first scan displays the donor’s fluorescence (DMACA) decreasing as the acceptor dye (FITC) was added. This is the ‘quenching’ effect that DMACA exhibits as the dye’s energy is transferred to the acceptor dye, FITC. On the right side of the graph, the plot shows that the acceptor’s (FITC) fluorescence was increased as the DMACA is added to the solution. This corresponds to the energy transfer taking place from the donor dye to the acceptor dye. In addition, the amount of energy transferred is not equal. Conservation of energy requires that the energy of the electronic state for the acceptor molecule must be either the same or less than that for the donor molecule. The excess of electronic energy upon transfer may be dissipated into vibrational energy. The transfer process, in effect, competes with other modes of de-excitation, including direct emission of photons with energy appropriate to the electronic state of the donor (Lakowicz, 1991). Fig. 4 displays the individual emission scans of AMCA-NHS and FITC and the mixture scan of these two FRET dyes. The AMCA-NHS acted as the donor dye with an absorption peak at 347 nm and emission peak at 450 nm, while the FITC was the acceptor dye with an emission peak at 520 nm. From the individual
Fig. 3. Time based scans of DMACA monitored at 457 nm with additions of FITC (— ) and FITC monitored at 535 nm with additions of DMACA (- - -).
235
Fig. 4. Emission spectrum of the individual fluorophores, AMCANHS (a) and FITC (b) and a mixture of AMCA-NHS +FITC (c). The excitation wavelength was set at 343 nm.
scan of the FITC fluorophore, there was a small emission peak resulting from the donor’s excitation wavelength. This indicates that only a small portion of the excitation wavelength of the AMCA-NHS dye will overlap the excitation wavelength of the FITC dye. When the dyes were mixed, a decrease in the emission peak of the AMCA-NHS dye at :450 nm resulted, while the fluorescent peak of the FITC dye at 520 nm increased. Therefore, it was demonstrated that a radiation-less transfer of excitation energy from the AMCANHS to the FITC dye occurred. Again, for simplicity, the donor –acceptor pairs in solution were assumed stationary and the transfer efficiency was roughly predicted using Eq. (2). The transfer efficiency was calculated to be 63%.
3.2. Conjugation of FRET pairs to biomolecules AMCA-NHS was tagged to streptavidin, while FITC was tagged to biotin. Fig. 5 shows the individual emis-
Fig. 5. Emission spectrum of the conjugated fluorophores, AMCANHS/streptavidin (a) and FITC/biotin (b) and a mixture of AMCANHS/streptavidin and FITC/biotin (c). The excitation wavelength was set at 343 nm.
236
S.A. Grant et al. / Biosensors & Bioelectronics 16 (2001) 231–237
4. Conclusion
Fig. 6. Emission spectrum of conjugated molecules FITC-tagged avidin (a), TRITC-tagged biotin (b) and a mixture (c), excited at 440 nm.
sion scans of the FRET labeled pairs and also the scan of the pair after they were mixed. There was a quenching of the donor peak, while the acceptor peak increased. When comparing Fig. 5 to Fig. 4, the AMCA-NHS peak was quenched to a greater degree for the streptavidin and biotin conjugates in comparison to the free dye conditions. Indeed, the energy transfer was calculated to be 88%, thus demonstrating a significant increase in transfer efficiency of :29%. In the case of the unconjugated dyes, the FRET pairs are randomly assembled, but with the streptavidin and biotin conjugates, the FRET pairs are assembled as a result of the high affinity between avidin and biotin. The energy transfer between the FRET pair is higher and more stable. Fig. 6 shows individual emission scans of TRITCtagged avidin and FITC-tagged biotin being excited at 440 nm. When the two were mixed together, there was a significant decrease in the donor’s emission. However, it was noted that there was no significant increase in the acceptor’s fluorescence. This could be due to several possible factors. One is that the energy was transferred into the form of vibrational energy to the surrounding protein structures, not to the acceptor dye. A more plausible reason is due to the general avidin and biotin binding structure. The acceptor dye was conjugated to the avidin molecule, while the donor dye was conjugated to the biotin molecule. There are four biotinbinding sites for every one avidin protein. Therefore, there is a 4:1 ratio between the donor dye and the acceptor dye. The concentration of the acceptor dye upon binding may have been too low to cause a significant increase in its fluorescence. To compensate for this problem, future experiments are planned whereby the acceptor dyes will be conjugated to biotin while the donor dyes will be conjugated to avidin.
This paper detailed the investigation and optimization of the FRET chemical transduction system. A number of fluorophores were tested as potential FRET pairs. The promising FRET pairs were further tested and conjugated to the biomolecules, streptavidin, avidin and biotin. These pairs were DMACA/FITC, AMCA/ FITC and FITC/TRITC. This identification and fluorescent labeling was the first step in developing enhanced biosensors that utilize the FRET dual receptor method. Because of their affinities, the FRETtagged biomolecules combine in solution resulting in a more stable, elevated fluorescent signal. The results indicate that the FRET pairs can be utilized in the development of a dual receptor sensor. Tests are currently being performed to further demonstrate the dual receptor method.
Acknowledgements The project was funded through a State of Michigan Research Excellence Fund (REF) and by the Whitaker Foundation Special Opportunity Grant.
References Anderson, G.P., King, K.D., Cao, L.K., Jacoby, M., Ligler, F.S., Ezzell, J., 1998. Clin. Diagn. Lab. Immunol. 5, 609. Ballerstadt, R., Schultz, J.S., 1997. Anal. Chem. Acta 345, 203. Barker, S.L., Zhao, Y., Marletta, M.A., Kopelman, R., 1999. Anal. Chem. 71, 2071. Bhatia, S.K., Shriver-Lake, L.C., Prior, K.J., Georger, J.H., Calvert, J.M., Bredehorst, R., Ligler, F.S., 1989. Anal. Biochem. 178, 408. Blankenburg, R., Meller, P., Ringsdorf, H., Salesse, C., 1989. Biochemistry 28, 8214. Canziani, G., Zhang, W., Cines, D., Rux, A., Willis, S., Cohen, G., Eisenberg, R., Chaiken, I., 1999. Methods 19, 253. Chang, Q., Sipior, J., 1995. Anal. Biochem. 232, 92. Chantal, M., Lebrun, C., 1998. Biochem. Edu. 26, 320 – 323. Diederick, A., Losche, M., 1997. Adv. Biophys. 34, 205. Ferguson, G.S., Chaudury, M.K., Sigal, G.B., Whitesides, G.M., 1991. Science 253, 776. Fo¨ rster, T., 1948. Ann. Phys. 2, 55. Haugland, R.P., 1996. Handbook on Fluorescent Probes and Research Chemicals. Molecular Probes, Eugene, OR. Herman, G.T., 1996. Bioconjugate Techniques. Academic Press, San Diego, CA. Kroger, D., Liley, M., Schiweck, W., Skerra, A., Vogel, H., 1999. Biosens. Bioelec. 14, 155. Lakowicz, J.R., 1983. Principles of Fluorescence Spectroscopy. Plenum Press, New York. Lakowicz, J.R., 1991. Topics in Fluorescence Spectroscopy, vol. 2. Plenum Press, New York. Mulchandani, A., Kaneva, I., Chen, W., 1998. Anal. Chem. 70, 5042. Mulchandani, A., Pan, S., Chen, W., 1999. Biotechnol. Prog. 15, 130. Nakai, J., Sekiguchi, N., Rando, T.A., Allen, P.D., Beam, K.G., 1998. J. Biol. Chem. 273, 13403.
S.A. Grant et al. / Biosensors & Bioelectronics 16 (2001) 231–237 Newman, D.J., Thakkar, H., Lam-Po-Tang, M.K., Kwan, J.T., 1999. Ren. Fail. 21, 349. Pearce, D.A., 1998. Biorgan. Med. Chem. Lett. 8, 1963. Perez-Luna, V.H., O’Brien, M.J., Opperman, K.A., Hampton, P.D., Lopez, G.P., Klumb, L.A., Stayton, P.S., 1999. J. Am. Chem. Soc. 121, 6469.
.
237
Prime, K.L., Whitesides, G.M., 1991. Science 252, 1164. Selvin, P.R., 1995. Methods Enzymol. 246, 300. Seo, K.H., Brackett, R.E., Hartman, N.F., Campbell, D.P., 1999. J. Food Prot. 62, 431. Spinke, J., Liley, M., Schmitt, F.J., 1993. J. Chem. Phys. 99, 7012. Wu, P., Brand, L., 1994. Anal. Biochem. 218, 1.