- Email: [email protected]
Covalent DNA Immobilization on Polymer-Shielded Silver-Coated Quartz Crystal Microbalance Using Photobiotin-Based UV Irradiation
Biochemical and Biophysical Research Communications 290, 962–966 (2002) doi:10.1006/bbrc.2001.6297, available online at http://www.idealibrary.com on
Covalent DNA Immobilization on Polymer-Shielded Silver-Coated Quartz Crystal Microbalance Using Photobiotin-Based UV Irradiation Xiaodi Su 1 Micro- & Nano-System Lab, Institute of Materials Research & Engineering, 3 Research Link, Singapore 117602
Received December 14, 2001
The use of a commercial, silver-coated quartz crystal microbalance (QCM) as a disposable, low-cost, and reliable DNA sensor is presented. This is an incorporation of polymer-based silver electrode shielding and photochemistry-based surface modification for covalent DNA immobilization. To prevent undesired oxidation, the silver electrodes are coated with thin polystyrene films. The polymer surfaces are then modified by a photoreactive biotin derivative (photobiotin) under UV irradiation. The resulting biotin residues on the polymer-shielded surface react with a tetrameric avidin. Consequently a biotin-labeled DNA probe can be immobilized through a biotin–avidin– biotin bridge. A 14-mer single-stranded biotin–DNA probe and a 70mer single-stranded DNA fragment containing complementary or noncomplementary sequences are used as a model system for DNA hybridization assay on the proposed sensors. The shielding ability of the polystyrene coatings after photo irradiation is investigated. The DNA probe binding capacity, hybridization efficiency, and kinetics are also investigated. © 2002 Elsevier Science (USA)
Key Words: quartz crystal microbalance; polystyrene; DNA sensor; photochemistry; UV irradiation; photobiotin.
Quartz crystal microbalance (QCM) devices have been widely used in biological analysis. From an examination of the published QCM biosensors, gold electrode-coated crystals have been most frequently used (1, 2). This kind of quartz crystal exhibits long-term frequency stability and permits the regeneration of the used electrode because of the good stability of gold (3, 4). However, the high unit price of gold-coated crystals has weakened the advantages of QCM assay to some extent and has become a barrier to the commercialization of this sensor technique (5). Thus, there is an 1
Address correspondence and reprint requests to author. Fax: 65-8720785. E-mail: [email protected]. 0006-291X/02 $35.00 © 2002 Elsevier Science (USA) All rights reserved.
increased demand for low cost transducers for fabrication of disposable sensors and sensor arrays. Commercially available silver-coated quartz crystals, for example the standard component form of HC 49/U, are candidates in this regard. In a previous study, it was demonstrated that the use of this kind of low cost crystals as immunosensors is profitable with respect to molecular immobilization ability, acceptable mass sensitivity variability, frequency measurement stability, reliability, etc. However, the major problem encountered with the silver crystals is the aging of the metal electrode. Biologically activated sliver electrodes cannot retain long-term stability and reactivity. The continuous oxidation of silver makes it difficult to reach a stable initial frequency value in aqueous mediate. To overcome this problem, Sakti et al. (6) and Su et al. (7) have successfully demonstrated the use of polystyrene as shielding polymer to protect the electrodes from undesired oxidation. The resulting polymer coatings are suitable for protein adsorption through hydrophobic interaction. To realize covalent immobilization of protein and DNA on a polymer-protected-electrode, the polymer films need to be modified with additional functional groups. Among many of the conventional polymer surface modification techniques (e.g., plasma treatment, ion implantation, and graft polymerization) photochemical immobilization of molecules having defined characteristics has been considered as more durable, generic, and cost-effective method to alter the surface chemistry (8). Normally, a heterobifunctional photochemical coupling reagent consists of a photoactivable group, a spacer group to hold a functional group. Upon UV irradiation, the photoactivable group undergoes photolysis to generate intermediates that can insert readily into the chemical bonds of polymer substrates. As a result, the functional group is exposed at the interface. Photoactivable nitrene-generating biotin is a derivative of biotin, with an aryl azide group as the photo-
962
Vol. 290, No. 3, 2002
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
reactive group and a biotin as the functional group. After UV irradiation, the aryl azide generates a nitrene that can insert readily into COH, NOH, or OOH bonds. This reagent has been widely applied for protein and DNA labeling (9) and substrate modification for protein patterning (10 –12). In the latter applications, biotin/avidin/biotin chemistry is used for covalent protein and DNA immobilization. The tetrameric structure of avidin makes it possible to bind both to biotin groups on surfaces and biotinylated molecules simultaneously. The sandwiched or bridged structure leads to an irreversible covalent binding with a high affinity of K a ⫽ 1 ⫻ 10 15 M ⫺1 (13). In this study, we use polystyrene as shielding material to protect the silver electrodes of the QCM, photobiotin to activate the polymer coatings under UV irradiation, and biotin/avidin/biotin chemistry to realize covalent DNA immobilization. A 14-mer single-stranded biotin–DNA probe and a 70-mer single-stranded DNA fragment containing complementary or noncomplementary sequences are used as a model system for DNA hybridization assay. The shielding ability of the polystyrene coatings before and after photo irradiation is investigated. The probe binding capacity, hybridization efficiency, and kinetics are also investigated. This study leads to a cost-effective DNA hybridization assay by using silver-coated QCM as reaction carriers. MATERIALS AND METHODS Reagents. Polystyrene (PS), photobiotin [N-(4-azido-2-nitrophenyl)N-(N-d-biotinyl-3-aminopropyl)-N-methyl-1,3-propanediamine] (PHB), avidin (from hen egg white), and avidin fluorescence conjugate (avidin– FITC) were purchased from Sigma (St. Louis, MO). A biotinylated single-strand DNA probe with biotin attached to the 5⬘-phosphate end (biotin–ssDNA) was synthesized by GENSET Singapore Biotech Pte Ltd. The sequence was 5⬘-biotin-(CH2) 6-GCCTTACCTGGATT-3⬘. The single strand DNA targets (ssDNA) were 70-mer DNA fragments containing either complementary or noncomplementary sequences to the probe. The molecular weight of the probe and the targets were 4556 and 21,615, respectively. Hybridization buffer and buffer solutions used for avidin and probe immobilization were phosphate-buffered saline (PBS) containing 0.01 M phosphate and 0.15 M NaCl, pH 7.4. The washing buffer was PBS containing 0.05% Tween 20. QCM device. The reaction carriers used were AT-cut, 10-MHz quartz crystals with silver electrodes purchased from Hong Kong X’tals Limited (HK, China). They were in the standard component form of HC 49/U with a quartz wafer of 8.2 mm in diameter and a silver electrode of 4.5 mm. These crystals provided mass sensitivity of 1 Hz corresponding to a mass increase of 0.700 ⫾ 0.01 ng. A laboratory crystal oscillator provided by International Crystal Manufacture Co. Inc. (Oklahoma City, OK) was used to operate the crystals in both gas phase and liquid phase. The output frequency was continually monitored by a TF830 universal counter (ThurlbyThandar, UK) and finally transferred to a computer using the RS 232C interface. To realize liquid phase measurement, the crystals were clamped between two Plexiglas blocks with Neoprene O-ring seals. The upper surface of the crystal was exposed to 50 l of carrier buffer. Whenever stabilized, 5 l of sample solution was injected and frequency response was followed.
PS film preparation. PS film was prepared using top-down dropping method as described previously (7). It was accomplished by dropping 1 l of a PS toluene solution (10 mg/ml) on the silver electrodes. To ensure the even spread of the solution and a uniform PS thin film, the crystals were positioned such that the quartz plate remained horizontal. Upon solvent evaporation a thin layer of PS film remained. Photobiotinylation of the PS film. Ten microliters of PHB solution (2 mg/ml in distilled water containing 20% ethanol) was applied on the PS-coated surface (the presence of ethanol improved spread ability and rendered a homogeneous layer of biotin). After dried at room temperature in dark for 1 h, the PHB films were exposed for 30 min to the filtered light from a 150-W high intense UV lamp (U.S.A.). After the irradiation, the surfaces were rinsed with distilled water and sonicated for 5 min with 1% Tween 20 to remove unreacted photobiotin. Ten microliters of avidin (0.5 mg/ml in PBS) or avidin– FITC (1:5 diluted) was then applied to the biotinylated PS surfaces to incubate for 1 h. The avidin–FITC-coated surface was then imaged using a fluorescent microscope equipped with a 100-mW Hg arc lamp for illumination. To monitor the interface reactions, resonant frequency at each reaction step was recorded in dry state. Biotin–ssDNA immobilization and hybridization assay. To perform DNA immobilization and hybridization in liquid phase, a PHB/ PS-coated sensor was exposed to a 50-l PBS buffer. When the frequency was stabilized, avidin was injected (final concentration, 0.5 mg/ml) followed by biotin–ssDNA (42 g/ml) and DNA target (with varied concentrations). The resonant frequency was continuously monitored as the function of time.
RESULTS AND DISCUSSION Figure 1 illustrates the procedures of the PHB modification of a PS shielded electrode, sandwiched biotin/ avidin/biotin interaction for biotin–ssDNA immobilization and DNA hybridization. In a previous study (7) it was proved that the top down coated PS films from a 5 mg/ml toluene solution could protect the electrodes from undesired oxidation. In the current study, PS coatings were prepared from a 10 mg/ml solution, giving a thicker PS film of ⬃0.8 m (calculated using mass measurement sensitivity of 1 Hz ⫽ 0.7 ng). After a long-term exposure of the PS-coated sensors to ambient laboratory conditions, no frequency decrease was detectable. For an unprotected QCM, however, the oxidation of the silver electrode gave a frequency decrease at a rate of ⬃15 Hz day ⫺1 because of the formation of silver oxides (6, 7). In addition, the PS-coated electrodes have improved surface condition (i.e., smoother surface), making the sensors more suitable for liquid phase and kinetic study. To obtain a high degree of photobiotin attachment on the PS-coated surface, photobiotinylation was conducted on a dried PHB film (14) as described in experimental section. A X-ray photoelectron spectroscopic (XPS) study was conducted to confirm the success of the biotinylation and subsequent avidin binding. Table 1 shows the XPS data (elemental composition) for the surfaces with PS coating (I), PBH/PS coatings (II), and avidin/PHB/PS coatings (III). The values in the table are the average of the measurement results from three samples. The C (1s) and O (1s) peaks on surface I
963
Vol. 290, No. 3, 2002
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
FIG. 1. Schematic illustration of the procedures of photobiotinylation of polymer shielded silver QCM, biotin–ssDNA immobilization through biotin–avidin– biotin bridge, and consequent DNA hybridization.
ascribe to the polymer and the adsorbed oxygen, respectively. The exhibition of the N (1s) peak and the increased O (1s) peak on surface II are the evidences of the deposition of PHB. The subsequent avidin binding on surface III results in further increase of the N (1s) peak. To study the protective ability, PHB/PS-coated sensors were exposed to laboratory air. The continuous frequency recorded in 30 days indicated that the UV irradiated PS films remain protective ability to the silver electrodes. In the subsequent avidin binding experiments, avidin solutions were applied to a biotinylated PS surface and two control PS surfaces either treated by PHB but no irradiation or irradiated in the absence of PHB. To study the strength of avidin binding on these surfaces, the avidin coatings were rinsed sequentially with distilled water, PBS washing solution (containing detergent), and 0.1 N NaOH. Figure 2 shows the frequency changes caused by avidin binding upon rinsing with different solutions. Although avidin can adsorb on the control SP surfaces via physical adsorption or hydrophobic interaction (15), the resulting protein layers are
less stable. After rinsed with PBS washing buffer, the avidin coatings lose significantly. Under 0.1 N NaOH, the avidin coatings are stripped away entirely. In contrast, however, the avidin on biotinylated PS surface can withstand harsher washing buffer and is virtually unaffected by extreme pH shift. This is attributed to the strong covalent interaction between biotin and avidin. To verify the effect of the UV irradiation further, avidin-FITC was used to incubate with a PHB treated PS surface. A part of this surface was masked with an aluminum film when it was exposed to the UV light. After washing with distilled water and drying, the surface was imaged with a fluorescent microscope (Fig. 3). Although fluorescence is observed in both the irradiated (A) and masked (B) areas, the intensity and
TABLE 1
XPS Data for PS-Coated Sensor before and after PHB Treatment and Avidin Binding Elemental composition (%) Surface
C
O
N
PS (I) PHB/PS (II) Avidin/PHB/PS (III)
93.4 89.1 88.0
6.6 8.5 8.5
2.4 3.5
Note. Average results from three samples.
FIG. 2. Frequency changes caused by avidin adsorption on PScoated QCM. The PS coatings were either treated with photobiotin with/without UV irradiation or not treated by photobiotin. Washing solutions 1, 2, and 3 refer to distilled water, ELISA washing buffer with detergent, and 0.1 N NaOH, respectively.
964
Vol. 290, No. 3, 2002
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
FIG. 3. Fluorescence micrograph of avidin–FITC-immobilized on a PHB/PS-coated QCM with (area A) or without (area B) UV irradiation.
distribution of the fluorescence is entirely different. In the un-irradiated area, avidin–FITC is only observed occasionally owing to the nonspecific adsorption. Upon rinsing with PBS washing buffer, the nonspecifically adsorbed avidin-FITC in area B was removed and the fluorescence in the irradiated area remained unchanged (image not shown). Both the QCM result (Fig. 2) and fluorescent image (Fig. 3) indicate that nonspecific avidin binding on PS-coated sensor is minimal with no photobiotin or no irradiation. It should be acknowledge that the surface tension of the photobiotin on PS surface upon drying may be responsible for the uneven coverage of biotin residues and the consequent avidin-FITC deposition (Fig. 3, area A). Further improvement of the surface coverage of the photobiotin film may be able to achieve by spin coating method or optimizing the solution contents for example adding appropriate amount of surfactant. As illustrated in Fig. 1, the strong binding of avidin results in the formation of surfaces to which biotin– ssDNA probes can then be bound through the biotin– avidin– biotin bridge. To perform DNA immobilization, the avidin layers were not dried to avoid denaturation. Figure 4 shows a continuous frequency response of a PHB/PS-coated sensor upon avidin binding, biotin– ssDNA probe immobilization and target DNA hybridization. It can be seen that the avidin binding completes within 20 min and biotin–ssDNA immobilization within 10 min. Using the in air frequency change caused by avidin binding (⬃250 Hz) and the in liquid frequency change of biotin–ssDNA binding (36 Hz), the binding capacity of these two reagents are 2.6 ⫻ 10 ⫺3 and 5.5 ⫻ 10 ⫺3 nmol, respectively (assuming MW ⫽ 68,000 for avidin and 4556 for the biotin–ssDNA). A binding ratio of ⬃1:2 is obtained for the immobilized avidin to the biotin–ssDNA, implying that after reacting with biotin residues at the PS surface, two binding sites in tetravalent avidin remain available for the
FIG. 4. Frequency response of a PHB/PS-coated QCM to avidin (2 mg/ml) application (start at arrow 1), 14-mer biotin–ssDNA (42 g/ml) immobilization (at arrow 2), and a complementary 70-mer ssDNA (5 g/ml) hybridization (at arrow 3).
biotin–ssDNA interaction. In another control experiment, a biotinylated sensor was exposed to biotin– ssDNA directly without application of avidin. The undetectable frequency response (data not shown) confirms that the tetravalent avidin is attributable to the biotin–ssDNA immobilization. Figure 5 shows the concentration dependent hybridization rate. When the concentration of the target DNA is 0.5, 1, 2, and 4 g/ml, as can be seen, the hybridization reactions (70-mer target) complete at ⬃44, 29, 20, and 10 min, respectively. Referring to the frequency shift caused by the saturated hybridization (⬃40 Hz), the hybridization efficiency is estimated as 23.7%, which compares favorable with that reported elsewhere (16, 17). Curve 1 in Fig. 5 was obtained from a control experiment, in which the biotin–ssDNA modified sensor was exposed to a non-complementary ssDNA. The nondetectable frequency response indicated that the nonspecific background is minimal.
FIG. 5. Frequency responses of biotin–ssDNA modified sensors to the hybridization of a 70-mer ssDNA target at concentrations of 0.5, 1, 2, and 4 g/ml (curves 2 to 5). Curve 1 is obtained from a control experiment, in which a noncomplementary ssDNA sample is applied.
965
Vol. 290, No. 3, 2002
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
CONCLUSION In this study, we use silver-coated quartz crystals as reaction carriers to create a cost-effective QCM DNA sensor. Polystyrene is used to shield the silver electrode from undesirable oxidation. Photobiotin is used to treat the polymer surface through UV irradiation, so as to enable covalent DNA immobilization using biotin–avidin– biotin chemistry. This photochemistry method is suitable for activation of almost all kind of polymer materials for example, PE, PVC, polyimide. The incorporated electrode shielding and covalent immobilization strategy is suitable for any solid-phase bioassay where the electrode is an unstable material, for example surface acoustic wave (SAW) device with aluminum electrode. Since the binding affinity between biotin and avidin is extremely high, being able to withstand 0.1 N NaOH treatment, the hybridization reversibility of the sensor, i.e., separating the double-stranded DNA by pH shift and resetting the sensor with biotin–ssDNA probe for new cycles of hybridization is expected. ACKNOWLEDGMENT This work has been supported by Singapore NSTB Grant 93/11/4-2.
REFERENCES 1. Cavic, B. A., Hayward, G. L., and Thompson, M. (1999) Acoustic waves and the study of biochemical macromolecules and cells at the sensor-liquid interface. Analyst 32, 2353. 2. Su, X. D., Chew, F. T., and Li, S. F. Y. (2000) Design and application of piezoelectric quartz crystal-based immunoassay. Anal. Sci. 16, 107. 3. Steegborn, C., and Skladal, P. (1997) Construction and characterization of the direct piezoelectric immunosensor for atrazine operating in solution. Biosens. Bioelectron. 12, 19. 4. Su, X. D., Chew, F. T., and Li, S. F. Y. (2001) Piezoelectric quartz crystal based label-free analysis for allergy disease. Biosens. Bioelectron. 15, 629.
5. Minunni, M., Mascini, M., Guilbault, G. G., and Hock, B., 1995, The quartz crystal microbalance as biosensor. A status report on its future. Anal. Lett. 28, 749. 6. Sakti, S. P., Ro¯sler, S., Lucklum, R., Hauptmann P., Bu¨hling, F., and Ansorge, S. (1999) Thick polystyrene-coated quartz crystal microbalance as a basis of a cost effective immunosensor. Sens. Actuators A 76, 98. 7. Su, X. D., Ng, H. T., Dai, C. C., O’Shea, S. J., and Li, S. F. Y. (2000) Disposable, low cost, silver-coated, piezoelectric quartz crystal biosensor and electrode protection. Analyst 125, 2268. 8. Amos, R. A., Anderson, A. B., Clapper, D. L., Duquette, P. H., Duran, L. W., Hohle, S. G., Sogard, D. J., Swanson, M. J., and Guire, P. E. (1995) Biomaterial surface modification using photochemical coupling technology. In Encyclopedic Handbook of Biomaterials and Bioengineering, Part A: Materials (Wise, D. L., Trantolo, D. J., Altobelli, D. E., Yaszemski, M. J., Gresser, J. D., and Schwartz, E. R., Eds.), pp. 895–926. 9. Forster, A. C., Mcinnes, J. L., Skingle, D. C., and Symons, R. H. (1985) Nucleic Acids Res. 13, 745. 10. Brooks, S. A., Dontha, N., Davis, C. B., Stuart, J. K., O’Neill, G., and Kuhr, W. G. (2000) Segregation of micrometer-dimension biosensor elements on a variety of substrate surfaces. Anal. Chem. 72, 3253. 11. Brooks, S. A., Ambrose, W. P., and Kuhr, W. G. (1999) Micrometer dimension derivatization of biosensor surfaces using confocal dynamic patterning. Anal. Chem. 71, 2558. 12. Hengsakul, M., and Cass, A. E. G. (1996) Protein patterning with a photoactivatable derivative of biotin. Bioconjug. Chem. 7, 249. 13. Wilchek, M., and Bayer, E. A. (1988) The avidin– biotin complex in bioanalytical applications. Anal. Biochem. 171, 1. 14. Schwarz, A., Rossier, J. S., Roulet, E., Mermod, N., Roberts, M. A., and Girault, H. H. (1998) Micropatterning of biomolecules on polymer substrates. Langmuir 14, 5526. 15. Wybourne, M. N., Mingdi, Y., Keana, F. W., and Wu, J. C. (1996) Creation of biomolecule arrays by electrostatic immobilization on electron-beam-irradiated polystyrene thin films. Nanotechnology 7, 302. 16. Caruso, F., Rodda, E., Furlong, D. N., Niikura, K., and Okahata, Y. (1997) Quartz crystal microbalance study of DNA immobilization and hybridization for nucleic acid sensor development. Anal. Chem. 69, 2043. 17. Caruso, F., Rodda, E., Furlong, D. N., and Haring V. (1997) DNA binding and hybridization on gold and derivatized surfaces. Sens. Actuators B 41, 189.
966
Covalent DNA Immobilization on Polymer-Shielded Silver-Coated Quartz Crystal Microbalance Using Photobiotin-Based UV Irradiation Xiaodi Su 1 Micro- & Nano-System Lab, Institute of Materials Research & Engineering, 3 Research Link, Singapore 117602
Received December 14, 2001
The use of a commercial, silver-coated quartz crystal microbalance (QCM) as a disposable, low-cost, and reliable DNA sensor is presented. This is an incorporation of polymer-based silver electrode shielding and photochemistry-based surface modification for covalent DNA immobilization. To prevent undesired oxidation, the silver electrodes are coated with thin polystyrene films. The polymer surfaces are then modified by a photoreactive biotin derivative (photobiotin) under UV irradiation. The resulting biotin residues on the polymer-shielded surface react with a tetrameric avidin. Consequently a biotin-labeled DNA probe can be immobilized through a biotin–avidin– biotin bridge. A 14-mer single-stranded biotin–DNA probe and a 70mer single-stranded DNA fragment containing complementary or noncomplementary sequences are used as a model system for DNA hybridization assay on the proposed sensors. The shielding ability of the polystyrene coatings after photo irradiation is investigated. The DNA probe binding capacity, hybridization efficiency, and kinetics are also investigated. © 2002 Elsevier Science (USA)
Key Words: quartz crystal microbalance; polystyrene; DNA sensor; photochemistry; UV irradiation; photobiotin.
Quartz crystal microbalance (QCM) devices have been widely used in biological analysis. From an examination of the published QCM biosensors, gold electrode-coated crystals have been most frequently used (1, 2). This kind of quartz crystal exhibits long-term frequency stability and permits the regeneration of the used electrode because of the good stability of gold (3, 4). However, the high unit price of gold-coated crystals has weakened the advantages of QCM assay to some extent and has become a barrier to the commercialization of this sensor technique (5). Thus, there is an 1
Address correspondence and reprint requests to author. Fax: 65-8720785. E-mail: [email protected]. 0006-291X/02 $35.00 © 2002 Elsevier Science (USA) All rights reserved.
increased demand for low cost transducers for fabrication of disposable sensors and sensor arrays. Commercially available silver-coated quartz crystals, for example the standard component form of HC 49/U, are candidates in this regard. In a previous study, it was demonstrated that the use of this kind of low cost crystals as immunosensors is profitable with respect to molecular immobilization ability, acceptable mass sensitivity variability, frequency measurement stability, reliability, etc. However, the major problem encountered with the silver crystals is the aging of the metal electrode. Biologically activated sliver electrodes cannot retain long-term stability and reactivity. The continuous oxidation of silver makes it difficult to reach a stable initial frequency value in aqueous mediate. To overcome this problem, Sakti et al. (6) and Su et al. (7) have successfully demonstrated the use of polystyrene as shielding polymer to protect the electrodes from undesired oxidation. The resulting polymer coatings are suitable for protein adsorption through hydrophobic interaction. To realize covalent immobilization of protein and DNA on a polymer-protected-electrode, the polymer films need to be modified with additional functional groups. Among many of the conventional polymer surface modification techniques (e.g., plasma treatment, ion implantation, and graft polymerization) photochemical immobilization of molecules having defined characteristics has been considered as more durable, generic, and cost-effective method to alter the surface chemistry (8). Normally, a heterobifunctional photochemical coupling reagent consists of a photoactivable group, a spacer group to hold a functional group. Upon UV irradiation, the photoactivable group undergoes photolysis to generate intermediates that can insert readily into the chemical bonds of polymer substrates. As a result, the functional group is exposed at the interface. Photoactivable nitrene-generating biotin is a derivative of biotin, with an aryl azide group as the photo-
962
Vol. 290, No. 3, 2002
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
reactive group and a biotin as the functional group. After UV irradiation, the aryl azide generates a nitrene that can insert readily into COH, NOH, or OOH bonds. This reagent has been widely applied for protein and DNA labeling (9) and substrate modification for protein patterning (10 –12). In the latter applications, biotin/avidin/biotin chemistry is used for covalent protein and DNA immobilization. The tetrameric structure of avidin makes it possible to bind both to biotin groups on surfaces and biotinylated molecules simultaneously. The sandwiched or bridged structure leads to an irreversible covalent binding with a high affinity of K a ⫽ 1 ⫻ 10 15 M ⫺1 (13). In this study, we use polystyrene as shielding material to protect the silver electrodes of the QCM, photobiotin to activate the polymer coatings under UV irradiation, and biotin/avidin/biotin chemistry to realize covalent DNA immobilization. A 14-mer single-stranded biotin–DNA probe and a 70-mer single-stranded DNA fragment containing complementary or noncomplementary sequences are used as a model system for DNA hybridization assay. The shielding ability of the polystyrene coatings before and after photo irradiation is investigated. The probe binding capacity, hybridization efficiency, and kinetics are also investigated. This study leads to a cost-effective DNA hybridization assay by using silver-coated QCM as reaction carriers. MATERIALS AND METHODS Reagents. Polystyrene (PS), photobiotin [N-(4-azido-2-nitrophenyl)N-(N-d-biotinyl-3-aminopropyl)-N-methyl-1,3-propanediamine] (PHB), avidin (from hen egg white), and avidin fluorescence conjugate (avidin– FITC) were purchased from Sigma (St. Louis, MO). A biotinylated single-strand DNA probe with biotin attached to the 5⬘-phosphate end (biotin–ssDNA) was synthesized by GENSET Singapore Biotech Pte Ltd. The sequence was 5⬘-biotin-(CH2) 6-GCCTTACCTGGATT-3⬘. The single strand DNA targets (ssDNA) were 70-mer DNA fragments containing either complementary or noncomplementary sequences to the probe. The molecular weight of the probe and the targets were 4556 and 21,615, respectively. Hybridization buffer and buffer solutions used for avidin and probe immobilization were phosphate-buffered saline (PBS) containing 0.01 M phosphate and 0.15 M NaCl, pH 7.4. The washing buffer was PBS containing 0.05% Tween 20. QCM device. The reaction carriers used were AT-cut, 10-MHz quartz crystals with silver electrodes purchased from Hong Kong X’tals Limited (HK, China). They were in the standard component form of HC 49/U with a quartz wafer of 8.2 mm in diameter and a silver electrode of 4.5 mm. These crystals provided mass sensitivity of 1 Hz corresponding to a mass increase of 0.700 ⫾ 0.01 ng. A laboratory crystal oscillator provided by International Crystal Manufacture Co. Inc. (Oklahoma City, OK) was used to operate the crystals in both gas phase and liquid phase. The output frequency was continually monitored by a TF830 universal counter (ThurlbyThandar, UK) and finally transferred to a computer using the RS 232C interface. To realize liquid phase measurement, the crystals were clamped between two Plexiglas blocks with Neoprene O-ring seals. The upper surface of the crystal was exposed to 50 l of carrier buffer. Whenever stabilized, 5 l of sample solution was injected and frequency response was followed.
PS film preparation. PS film was prepared using top-down dropping method as described previously (7). It was accomplished by dropping 1 l of a PS toluene solution (10 mg/ml) on the silver electrodes. To ensure the even spread of the solution and a uniform PS thin film, the crystals were positioned such that the quartz plate remained horizontal. Upon solvent evaporation a thin layer of PS film remained. Photobiotinylation of the PS film. Ten microliters of PHB solution (2 mg/ml in distilled water containing 20% ethanol) was applied on the PS-coated surface (the presence of ethanol improved spread ability and rendered a homogeneous layer of biotin). After dried at room temperature in dark for 1 h, the PHB films were exposed for 30 min to the filtered light from a 150-W high intense UV lamp (U.S.A.). After the irradiation, the surfaces were rinsed with distilled water and sonicated for 5 min with 1% Tween 20 to remove unreacted photobiotin. Ten microliters of avidin (0.5 mg/ml in PBS) or avidin– FITC (1:5 diluted) was then applied to the biotinylated PS surfaces to incubate for 1 h. The avidin–FITC-coated surface was then imaged using a fluorescent microscope equipped with a 100-mW Hg arc lamp for illumination. To monitor the interface reactions, resonant frequency at each reaction step was recorded in dry state. Biotin–ssDNA immobilization and hybridization assay. To perform DNA immobilization and hybridization in liquid phase, a PHB/ PS-coated sensor was exposed to a 50-l PBS buffer. When the frequency was stabilized, avidin was injected (final concentration, 0.5 mg/ml) followed by biotin–ssDNA (42 g/ml) and DNA target (with varied concentrations). The resonant frequency was continuously monitored as the function of time.
RESULTS AND DISCUSSION Figure 1 illustrates the procedures of the PHB modification of a PS shielded electrode, sandwiched biotin/ avidin/biotin interaction for biotin–ssDNA immobilization and DNA hybridization. In a previous study (7) it was proved that the top down coated PS films from a 5 mg/ml toluene solution could protect the electrodes from undesired oxidation. In the current study, PS coatings were prepared from a 10 mg/ml solution, giving a thicker PS film of ⬃0.8 m (calculated using mass measurement sensitivity of 1 Hz ⫽ 0.7 ng). After a long-term exposure of the PS-coated sensors to ambient laboratory conditions, no frequency decrease was detectable. For an unprotected QCM, however, the oxidation of the silver electrode gave a frequency decrease at a rate of ⬃15 Hz day ⫺1 because of the formation of silver oxides (6, 7). In addition, the PS-coated electrodes have improved surface condition (i.e., smoother surface), making the sensors more suitable for liquid phase and kinetic study. To obtain a high degree of photobiotin attachment on the PS-coated surface, photobiotinylation was conducted on a dried PHB film (14) as described in experimental section. A X-ray photoelectron spectroscopic (XPS) study was conducted to confirm the success of the biotinylation and subsequent avidin binding. Table 1 shows the XPS data (elemental composition) for the surfaces with PS coating (I), PBH/PS coatings (II), and avidin/PHB/PS coatings (III). The values in the table are the average of the measurement results from three samples. The C (1s) and O (1s) peaks on surface I
963
Vol. 290, No. 3, 2002
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
FIG. 1. Schematic illustration of the procedures of photobiotinylation of polymer shielded silver QCM, biotin–ssDNA immobilization through biotin–avidin– biotin bridge, and consequent DNA hybridization.
ascribe to the polymer and the adsorbed oxygen, respectively. The exhibition of the N (1s) peak and the increased O (1s) peak on surface II are the evidences of the deposition of PHB. The subsequent avidin binding on surface III results in further increase of the N (1s) peak. To study the protective ability, PHB/PS-coated sensors were exposed to laboratory air. The continuous frequency recorded in 30 days indicated that the UV irradiated PS films remain protective ability to the silver electrodes. In the subsequent avidin binding experiments, avidin solutions were applied to a biotinylated PS surface and two control PS surfaces either treated by PHB but no irradiation or irradiated in the absence of PHB. To study the strength of avidin binding on these surfaces, the avidin coatings were rinsed sequentially with distilled water, PBS washing solution (containing detergent), and 0.1 N NaOH. Figure 2 shows the frequency changes caused by avidin binding upon rinsing with different solutions. Although avidin can adsorb on the control SP surfaces via physical adsorption or hydrophobic interaction (15), the resulting protein layers are
less stable. After rinsed with PBS washing buffer, the avidin coatings lose significantly. Under 0.1 N NaOH, the avidin coatings are stripped away entirely. In contrast, however, the avidin on biotinylated PS surface can withstand harsher washing buffer and is virtually unaffected by extreme pH shift. This is attributed to the strong covalent interaction between biotin and avidin. To verify the effect of the UV irradiation further, avidin-FITC was used to incubate with a PHB treated PS surface. A part of this surface was masked with an aluminum film when it was exposed to the UV light. After washing with distilled water and drying, the surface was imaged with a fluorescent microscope (Fig. 3). Although fluorescence is observed in both the irradiated (A) and masked (B) areas, the intensity and
TABLE 1
XPS Data for PS-Coated Sensor before and after PHB Treatment and Avidin Binding Elemental composition (%) Surface
C
O
N
PS (I) PHB/PS (II) Avidin/PHB/PS (III)
93.4 89.1 88.0
6.6 8.5 8.5
2.4 3.5
Note. Average results from three samples.
FIG. 2. Frequency changes caused by avidin adsorption on PScoated QCM. The PS coatings were either treated with photobiotin with/without UV irradiation or not treated by photobiotin. Washing solutions 1, 2, and 3 refer to distilled water, ELISA washing buffer with detergent, and 0.1 N NaOH, respectively.
964
Vol. 290, No. 3, 2002
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
FIG. 3. Fluorescence micrograph of avidin–FITC-immobilized on a PHB/PS-coated QCM with (area A) or without (area B) UV irradiation.
distribution of the fluorescence is entirely different. In the un-irradiated area, avidin–FITC is only observed occasionally owing to the nonspecific adsorption. Upon rinsing with PBS washing buffer, the nonspecifically adsorbed avidin-FITC in area B was removed and the fluorescence in the irradiated area remained unchanged (image not shown). Both the QCM result (Fig. 2) and fluorescent image (Fig. 3) indicate that nonspecific avidin binding on PS-coated sensor is minimal with no photobiotin or no irradiation. It should be acknowledge that the surface tension of the photobiotin on PS surface upon drying may be responsible for the uneven coverage of biotin residues and the consequent avidin-FITC deposition (Fig. 3, area A). Further improvement of the surface coverage of the photobiotin film may be able to achieve by spin coating method or optimizing the solution contents for example adding appropriate amount of surfactant. As illustrated in Fig. 1, the strong binding of avidin results in the formation of surfaces to which biotin– ssDNA probes can then be bound through the biotin– avidin– biotin bridge. To perform DNA immobilization, the avidin layers were not dried to avoid denaturation. Figure 4 shows a continuous frequency response of a PHB/PS-coated sensor upon avidin binding, biotin– ssDNA probe immobilization and target DNA hybridization. It can be seen that the avidin binding completes within 20 min and biotin–ssDNA immobilization within 10 min. Using the in air frequency change caused by avidin binding (⬃250 Hz) and the in liquid frequency change of biotin–ssDNA binding (36 Hz), the binding capacity of these two reagents are 2.6 ⫻ 10 ⫺3 and 5.5 ⫻ 10 ⫺3 nmol, respectively (assuming MW ⫽ 68,000 for avidin and 4556 for the biotin–ssDNA). A binding ratio of ⬃1:2 is obtained for the immobilized avidin to the biotin–ssDNA, implying that after reacting with biotin residues at the PS surface, two binding sites in tetravalent avidin remain available for the
FIG. 4. Frequency response of a PHB/PS-coated QCM to avidin (2 mg/ml) application (start at arrow 1), 14-mer biotin–ssDNA (42 g/ml) immobilization (at arrow 2), and a complementary 70-mer ssDNA (5 g/ml) hybridization (at arrow 3).
biotin–ssDNA interaction. In another control experiment, a biotinylated sensor was exposed to biotin– ssDNA directly without application of avidin. The undetectable frequency response (data not shown) confirms that the tetravalent avidin is attributable to the biotin–ssDNA immobilization. Figure 5 shows the concentration dependent hybridization rate. When the concentration of the target DNA is 0.5, 1, 2, and 4 g/ml, as can be seen, the hybridization reactions (70-mer target) complete at ⬃44, 29, 20, and 10 min, respectively. Referring to the frequency shift caused by the saturated hybridization (⬃40 Hz), the hybridization efficiency is estimated as 23.7%, which compares favorable with that reported elsewhere (16, 17). Curve 1 in Fig. 5 was obtained from a control experiment, in which the biotin–ssDNA modified sensor was exposed to a non-complementary ssDNA. The nondetectable frequency response indicated that the nonspecific background is minimal.
FIG. 5. Frequency responses of biotin–ssDNA modified sensors to the hybridization of a 70-mer ssDNA target at concentrations of 0.5, 1, 2, and 4 g/ml (curves 2 to 5). Curve 1 is obtained from a control experiment, in which a noncomplementary ssDNA sample is applied.
965
Vol. 290, No. 3, 2002
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
CONCLUSION In this study, we use silver-coated quartz crystals as reaction carriers to create a cost-effective QCM DNA sensor. Polystyrene is used to shield the silver electrode from undesirable oxidation. Photobiotin is used to treat the polymer surface through UV irradiation, so as to enable covalent DNA immobilization using biotin–avidin– biotin chemistry. This photochemistry method is suitable for activation of almost all kind of polymer materials for example, PE, PVC, polyimide. The incorporated electrode shielding and covalent immobilization strategy is suitable for any solid-phase bioassay where the electrode is an unstable material, for example surface acoustic wave (SAW) device with aluminum electrode. Since the binding affinity between biotin and avidin is extremely high, being able to withstand 0.1 N NaOH treatment, the hybridization reversibility of the sensor, i.e., separating the double-stranded DNA by pH shift and resetting the sensor with biotin–ssDNA probe for new cycles of hybridization is expected. ACKNOWLEDGMENT This work has been supported by Singapore NSTB Grant 93/11/4-2.
REFERENCES 1. Cavic, B. A., Hayward, G. L., and Thompson, M. (1999) Acoustic waves and the study of biochemical macromolecules and cells at the sensor-liquid interface. Analyst 32, 2353. 2. Su, X. D., Chew, F. T., and Li, S. F. Y. (2000) Design and application of piezoelectric quartz crystal-based immunoassay. Anal. Sci. 16, 107. 3. Steegborn, C., and Skladal, P. (1997) Construction and characterization of the direct piezoelectric immunosensor for atrazine operating in solution. Biosens. Bioelectron. 12, 19. 4. Su, X. D., Chew, F. T., and Li, S. F. Y. (2001) Piezoelectric quartz crystal based label-free analysis for allergy disease. Biosens. Bioelectron. 15, 629.
5. Minunni, M., Mascini, M., Guilbault, G. G., and Hock, B., 1995, The quartz crystal microbalance as biosensor. A status report on its future. Anal. Lett. 28, 749. 6. Sakti, S. P., Ro¯sler, S., Lucklum, R., Hauptmann P., Bu¨hling, F., and Ansorge, S. (1999) Thick polystyrene-coated quartz crystal microbalance as a basis of a cost effective immunosensor. Sens. Actuators A 76, 98. 7. Su, X. D., Ng, H. T., Dai, C. C., O’Shea, S. J., and Li, S. F. Y. (2000) Disposable, low cost, silver-coated, piezoelectric quartz crystal biosensor and electrode protection. Analyst 125, 2268. 8. Amos, R. A., Anderson, A. B., Clapper, D. L., Duquette, P. H., Duran, L. W., Hohle, S. G., Sogard, D. J., Swanson, M. J., and Guire, P. E. (1995) Biomaterial surface modification using photochemical coupling technology. In Encyclopedic Handbook of Biomaterials and Bioengineering, Part A: Materials (Wise, D. L., Trantolo, D. J., Altobelli, D. E., Yaszemski, M. J., Gresser, J. D., and Schwartz, E. R., Eds.), pp. 895–926. 9. Forster, A. C., Mcinnes, J. L., Skingle, D. C., and Symons, R. H. (1985) Nucleic Acids Res. 13, 745. 10. Brooks, S. A., Dontha, N., Davis, C. B., Stuart, J. K., O’Neill, G., and Kuhr, W. G. (2000) Segregation of micrometer-dimension biosensor elements on a variety of substrate surfaces. Anal. Chem. 72, 3253. 11. Brooks, S. A., Ambrose, W. P., and Kuhr, W. G. (1999) Micrometer dimension derivatization of biosensor surfaces using confocal dynamic patterning. Anal. Chem. 71, 2558. 12. Hengsakul, M., and Cass, A. E. G. (1996) Protein patterning with a photoactivatable derivative of biotin. Bioconjug. Chem. 7, 249. 13. Wilchek, M., and Bayer, E. A. (1988) The avidin– biotin complex in bioanalytical applications. Anal. Biochem. 171, 1. 14. Schwarz, A., Rossier, J. S., Roulet, E., Mermod, N., Roberts, M. A., and Girault, H. H. (1998) Micropatterning of biomolecules on polymer substrates. Langmuir 14, 5526. 15. Wybourne, M. N., Mingdi, Y., Keana, F. W., and Wu, J. C. (1996) Creation of biomolecule arrays by electrostatic immobilization on electron-beam-irradiated polystyrene thin films. Nanotechnology 7, 302. 16. Caruso, F., Rodda, E., Furlong, D. N., Niikura, K., and Okahata, Y. (1997) Quartz crystal microbalance study of DNA immobilization and hybridization for nucleic acid sensor development. Anal. Chem. 69, 2043. 17. Caruso, F., Rodda, E., Furlong, D. N., and Haring V. (1997) DNA binding and hybridization on gold and derivatized surfaces. Sens. Actuators B 41, 189.
966