Quartz Crystal Microbalance Study of DNA ... - ACS Publications

Immobilization and Hybridization for Nucleic Acid. Sensor Development. Frank Caruso, Elke Rodda, and D. Neil Furlong*. CSIRO, Division of Chemicals an...
0 downloads 0 Views 186KB Size
Download PDF
Anal. Chem. 1997, 69, 2043-2049

Quartz Crystal Microbalance Study of DNA Immobilization and Hybridization for Nucleic Acid Sensor Development Frank Caruso, Elke Rodda, and D. Neil Furlong*

CSIRO, Division of Chemicals and Polymers, Private Bag 10, Clayton South MDC, Victoria 3169, Australia Kenichi Niikura and Yoshio Okahata

Department of Biomolecular Engineering, Tokyo Institute of Technology, Nagatsuda, Midori-ku, Yokohama 227, Japan

The immobilization of two 30-mer oligonucleotides, one biotinylated (biotin-DNA) and the other having a mercaptohexyl group at the 5′-phosphate end (BS1-SH), onto modified gold surfaces has been examined using a quartz crystal microbalance (QCM). Both single-layer and multilayer DNA films were prepared. The single-layer films of biotin-DNA were constructed by binding to a precursor layer of avidin, which had been attached to the QCM either covalently using a water-soluble carbodiimide hydrochloride (EDC) and N-hydroxysuccinimide (NHS) or via electrostatic interaction with poly(allylamine hydrochloride) (PAH). Single-layer films of BS1-SH were also formed on PAH via the electrostatic attraction between the amine groups on PAH and the negatively charged phosphate backbone of DNA. Multilayer films of DNA were fabricated by the successive deposition of avidin and poly(styrenesulfonate) (PSS), up to a total of nine avidin/ PSS layers, followed by DNA adsorption. DNA immobilization and hybridization of the immobilized DNAs was monitored in situ from QCM frequency changes. Hybridization was induced by exposure of the DNAcontaining films to complementary DNA in solution. Equal frequency changes were observed for the DNA immobilization and hybridization steps for the single-layer films, indicating a DNA probe-to-hybridized DNA target ratio of 1:1. The multilayer DNA films also exhibited DNA hybridization, with a greater quantity of DNA hybridized compared with the single-layer films. The multilayer films provide a novel means for the fabrication of DNA-based thin films with increased capacity for nucleic acid detection. Recently there has been increasing interest in the detection of specific DNA sequences using methods which do not require the use of labels such as radioisotopes, enzymes, and fluorophores.1 Biosensor systems based on nucleic acids not only eliminate the need for such labels but also offer the potential advantage of rapid, real-time solution monitoring of DNA hybridization, as well as high sensitivity and specificity. A biosensor with a DNA probe can have a number of applications, including (1) Downs, M. E. A. Biochem. Soc. Trans. 1991, 19, 39. S0003-2700(96)01220-6 CCC: $14.00

© 1997 American Chemical Society

real-time detection of genome DNA (for example, in clinical diagnostics), forensic identification, and a host of research applications.1,2 The basis of operation for a DNA biosensor is the complementary coupling between the specific DNA sequences within target analytes and the specific nucleotide sequence immobilized onto the solid support (i.e., transducer). Methods used for the direct detection of DNA binding through base pairing (without using specific labels) include electrochemical,3,4 optical,5-8 and microgravimetric8-13 techniques. The quartz crystal microbalance (QCM) is a promising candidate for biosensor applications, and its potential for the detection of DNA hybridization has recently been demonstrated.8-13 The QCM is an extremely sensitive mass sensor, capable of measuring subnanogram levels.14,15 The crystal resonance frequency decreases with an increase in mass on the QCM.16 Although the QCM has a high inherent sensitivity, methods for improving the detection limit of this device are being sought to enable wide application of the technique for DNA hybridization detection and, for example, in other areas such as immunochemical sensing. A detection limit of around 10-18 M, which corresponds to approximately 10-12 g (depending on the number of DNA base pairs) of target DNA, is required for many applications.1,5 The sensitivity can, in some cases, be improved by using QCM crystals of higher frequencies17 or by amplification of the nucleic acids by polymerase chain (2) Landegren, U.; Kaiser, R.; Caskey, C. T.; Hood, L. Science 1988, 242, 229. (3) Krznaric, D.; Cosovic, B. Anal. Biochem. 1986, 156, 454. (4) Brabec, V. Bioelectrochem. Bioenerg. 1983, 11, 245. (5) Pollard-Knight, D.; Hawkins, E.; Yeung, D.; Pashby, D. P.; Simpson, M.; McDougall, A.; Buckle, P.; Charles, S. A. Ann. Biol. Clin. 1990, 48, 642. (6) Jost, J. P.; Munch, O.; Anderson, T. Nucleic Acids Res. 1991, 19, 2788. (7) Piscevic, D.; Lawall, R.; Veith, M.; Liley, M.; Okahata, Y.; Knoll, W. Appl. Surf. Sci. 1995, 90, 425. (8) Caruso, F.; Rodda, E.; Furlong, D. N. Sens. Actuators B, in press. (9) Fawcett, N. C.; Evans, J. C.; Chien, L. C.; Drozda, K. A.; Flowers, N. Sens. Technol. 1988, 4, 5. (10) Okahata, Y.; Matsunobu, Y.; Ijiro, K.; Mukae, M.; Murakami, A.; Makino, K. J. Am. Chem. Soc. 1992, 114, 8299. (11) Yamaguchi, S.; Shimomura, T. Anal. Chem. 1993, 65, 1925. (12) Su, H.; Kallury, K. M. R.; Thompson, M. Anal. Chem. 1994, 66, 769. (13) Ito, K.; Hashimoto, K.; Ishimori, Y. Anal. Chim. Acta 1996, 327, 29. (14) Guilbault, G. G. In Applications of Piezoelectric Quartz Crystal Microbalances; Lu, C., Czanderna, A. W., Eds; Elsevier: Amsterdam/New York, 1984; p 251. (15) Lucklum, R.; Henning, B.; Hauptmann, P.; Schierbaum, K. D.; Vaihinger, S.; Gopel, W. Sens. Actuators 1991, A25, 705. (16) Sauerbrey, G. Z. Phys. 1959, 155, 206.

Analytical Chemistry, Vol. 69, No. 11, June 1, 1997 2043

reaction (PCR).18 These two methods, however, have practical limitations, particularly in the development of automated biosensor systems for real-time monitoring: QCM devices of higher frequencies (>∼10 MHz) are often difficult to operate in liquids because of frequency stability problems, while use of PCR can be laborious and time consuming and requires a number of manipulations. Alternatively, the sensitivity of a sensor system (with respect to the concentration in solution) may be improved by the creation of multilayer DNA (receptor)-containing films. The construction of such films and their hybridization capacity are investigated in the present work. We employ QCM to investigate DNA hybridization in singlelayer and multilayer DNA films and explore the use of multilayer DNA films for increasing DNA hybridization capacity. The interaction between biotin and avidin is exploited to immobilize biotinylated DNA (biotin-DNA), and the electrostatic attraction between the polyelectrolyte poly(allylamine hydrochloride) (PAH) and the negative phosphate backbone of DNA is used to immobilize a 30-mer oligonucleotide (BS1-SH). (BS1-SH was used since it was available from a previous study.8) Sensitivity (with respect to DNA solution concentration) is increased by constructing multilayers of DNA by the successive deposition of avidin and poly(styrenesulfonate) (PSS) (up to a total of nine layers) on a precursor polyelectrolyte thin film, followed by biotin-DNA adsorption. The effect of nonspecific binding is also investigated using a noncomplementary 30-mer DNA. This work forms part of our on-going studies in examining and characterizing DNA binding and hybridization onto solid surfaces for biosensor applications. In a previous investigation,8 we employed QCM and surface plasmon resonance (SPR) to assess the degree of DNA binding to gold surfaces and hybridization with the attached DNA probes. The results obtained in the current study will be compared to those reported earlier.8

EXPERIMENTAL SECTION Materials. 3,3′-Dithiodipropionic acid, 2-aminoethanol (ethanolamine), poly(allylamine hydrochloride) (PAH, Mr 50 00065 000), and poly(sodium 4-styrenesulfonate) (PSS, Mr 70 000) were obtained from Aldrich Chemical Co. N-Ethyl-N′-(3-(dimethyl)aminopropyl)carbodiimide hydrochloride (EDC, AR grade), N-hydroxysuccinimide (NHS, AR grade), and 3-mercaptopropionic acid (MPA) were obtained from Sigma. The 30-mer biotinylated DNA, with biotin attached to the 5′-phosphate end (biotin-DNA), and the DNA 30-mer oligonucleotide, with a mercaptohexyl group at the 5′-phosphate end (BS1-SH), were purchased from Japan Bioservices Co., Ltd. (Saitama, Japan). These modified DNAs have the same sequence (5′dGTC ACG CTG CGC GTA ACC ACC ACA CCC GCC3′). The unmodified 30-mer DNAs, BS1c (complementary) and BS1nc (noncomplementary), were synthesized on an Applied Biosystems oligonucleotide synthesizer. The DNA 30mer oligonucleotide probes (biotin-DNA and BS1-SH) have a sequence that is complementary to that of BS1c, but they have the same sequence as that of BS1nc. Avidin (from egg white, extrapure reagent) was obtained from Nacalai Tesque, Inc. (Kyoto, Japan). All chemicals were used as received. (17) Bizet, K.; Gabrielli, C.; Perrot, H.; Therasse, J. Proceedings of The Fourth World Congress on Biosensors; Bangkok, Thailand, 29-31 May 1996, Elsevier Advanced Technology: Oxford, UK, 1996. (18) Saiki, R. K.; Gelfand, D. H.; Stoffel, S.; Scharf, S. J.; Higuchi, R. J.; Horn, G. T.; Mullis, K. B.; Erlich, H. A. Science 1988, 239, 487.

2044

Analytical Chemistry, Vol. 69, No. 11, June 1, 1997

HEPES buffer (4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid) was purchased from Aldrich. HEPES solutions (0.05 M) of pH 7.5 and ionic strength 0.1 M were prepared containing 0.024 M NaOH and 0.076 M NaCl. Sulfuric acid and nitric acid were AR grade and supplied by Rhoˆne-Poulenc. Sodium chloride (AR grade) and hydrogen peroxide (AR grade) were purchased from BDH. The silicon paste used to protect the electrical contacts and seal one side of the QCMs from solution was obtained from the Tokyo Toray Dow Corning Silicone Corp. (Tokyo, Japan). The water used in all experiments was obtained from a three-stage Milli-Q purification system with a conductivity of less than 1 µS cm-1. QCM Apparatus. AT-cut quartz crystals with a fundamental resonance frequency (Fo) of 9 MHz were supplied by Kyushu Dentsu Co. (Omura-City, Nagasaki, Japan). These crystals (4.5 mm electrode diameter) were supplied with gold-coated electrodes, formed by thermal evaporation of gold to an average thickness of 100 nm. The QCM is connected into a series resonant two-gate TTL circuit, which causes the QCM to oscillate at its Fo near 9 MHz. The TTL circuit is powered from a 5 V dc supply. The crystal frequencies were monitored by an Iwatsu frequency counter (SC 7201), which was connected to an IBMcompatible PC. All experiments were conducted at 22 ( 1 °C. Further details can be found in previous publications.8,19-21 Preparation of QCM Surfaces. Prior to use, the connecting wires to the QCM electrodes were covered with the protective silastic film to prevent degradation of the electrical contacts when submersed in solution. The gold QCM surfaces were then cleaned by exposure to piranha solution (one part of 30% H2O2 in three parts of H2SO4)22 for 2 min, followed by rinsing with pure water and drying with nitrogen. This process was repeated twice. Caution: Piranha solution should be handled with extreme care, and only small volumes should be prepared at any time. For in situ QCM experiments, one side of the quartz crystal was sealed with a rubber casing, maintaining it in an air environment, while the other was exposed to aqueous solution. The casing is essential for QCM frequency stability in liquids.8,19,20 For experiments where in air measurements only were performed, both sides of the crystal were exposed to the solution of interest. In all cases, the QCM crystals were used immediately after preparation. DNA Immobilization on QCM. (a) Biotin-DNA Immobilization via Interaction with Avidin (Scheme 1). The gold QCM was exposed to an ethanolic 5 mM solution of 3,3′dithiodipropionic acid for 20 min, followed by water rinsing. Five microliters of 100 mg mL-1 EDC solution was then placed on the surface, followed immediately by 5 µL of 100 mg mL-1 NHS solution. These solutions were allowed to interact with the 3,3′dithiodipropionic acid for 20 min in a 100% humidity environment to prevent solution evaporation. The surface was then rinsed with water and immersed in an aqueous 0.2 mg mL-1 avidin solution for at least 60 min, after which the surface was rinsed again. (In some cases, the avidin layer on the QCM was dried using nitrogen after the avidin adsorption step and the frequency change measured. These crystals were not, however, used for subsequent biotin-DNA immobilization, as avidin was found to denature on (19) Caruso, F.; Serizawa, T.; Furlong, D. N.; Okahata, Y. Langmuir 1995, 11, 1546. (20) Caruso, F.; Rodda, E.; Furlong, D. N. J. Colloid Interface Sci. 1996, 178, 104. (21) Caruso, F.; Rinia, H.; Furlong, D. N. Langmuir 1996, 12, 2145. (22) Bain, C. D.; Evall, J.; Whitesides, G. M. J. Am. Chem. Soc. 1989, 111, 7155.

Scheme 1. Immobilization of Biotin-DNA on Avidin-Modified QCM Electrodes

drying.) The QCM was then exposed to a 1 mM 2-aminoethanol solution (pH 8.0, adjusted using HCl) for 60 min, rinsed, and placed into HEPES buffer for DNA immobilization. When the QCM frequency had stabilized, biotin-DNA was injected (final concentration, 1 µg mL-1), and the QCM frequency was monitored as a function of time until immobilization was complete, as indicated by a constant (and maximum) frequency shift. The QCM was then removed from solution, rinsed, and used in the hybridization experiments. Multilayer biotin-DNA films were fabricated by the successive solution deposition of avidin and PSS (up to a total of nine layers) on a (PAH/PSS)2 precursor film of thickness 8 nm23 and subsequent exposure of this thin film to a solution of biotin-DNA. Fabrication of the precursor film of (PAH/PSS)2 on MPA-modified QCM electrodes has been described previously.23 This precursor layer was used to provide an outer layer of negatively charged polyelectrolyte (PSS) for the subsequent immobilization of avidin (isoelectric point, 10.0-10.5).24 Avidin (positively charged) was adsorbed onto the outer PSS layer from an aqueous 0.2 mg mL-1 avidin solution (pH ∼6) for 2 h, after which the surface was rinsed with water. The avidin-coated QCM was then exposed to ∼5 mL of a 3 mg mL-1 PSS solution for 1 min. (The PSS solution contained 0.01 M HCl and 10 mM MnCl2.) This surface was again rinsed with pure water. This procedure was repeated until a total of nine layers (five avidin and four PSS) were deposited. The multilayer film, which was kept wet to avoid avidin denaturation, was placed into a HEPES solution for DNA immobilization experiments. Upon stabilization of the QCM frequency, biotinDNA was injected (final concentration, 1 µg mL-1), and the QCM frequency was monitored as a function of time until adsorption was complete. The QCM was removed from solution, rinsed, and used in the hybridization experiments. In some experiments, the QCM was dried after the avidin and PSS adsorption steps, and the frequency changes were measured in order to monitor the quantity of adsorbed avidin and PSS. These crystals were not used for subsequent biotin-DNA immobilization. (b) Electrostatic Immobilization of BS1-SH. BS1-SH DNA was immobilized via electrostatic attraction between the positively (23) Caruso, F.; Niikura, K.; Furlong, D. N.; Okahata, Y. Langmuir, submitted for publication (part 1). (24) Lehninger, A. L. Biochemistry, 2nd ed.; Worth Publishers, Inc.: New York, 1975.

charged outer layer of PAH and the negatively charged phosphate backbone of DNA. A precursor trilayer of PAH/PSS/PAH on MPA-modified gold QCM electrodes was first prepared as previously described.23 The QCM was then immersed into a HEPES solution, and BS1-SH was injected (final concentration, 1 µg mL-1) once the QCM frequency had stabilized. The binding profile for BS1-SH was monitored by following the in solution frequency change as a function of time. The crystal was then removed from solution, rinsed, and dried. The quantity of immobilized BS1-SH was calculated from the in air frequency shift. Hybridization. Hybridization experiments were performed by exposing the QCMs with the immobilized DNA probes (biotinDNA or BS1-SH) to the hybridization buffer HEPES/NaCl (0.05 M HEPES containing 100 mM NaCl, pH 7.5) and subsequently injecting the complementary 30-mer DNA (BS1c; final concentration, 0.5 µg mL-1). Hybridization was evidenced by in situ frequency changes. In some cases, following completion of the hybridization reaction in situ, the QCMs were removed from solution, rinsed, and dried and the in air frequencies recorded. Noncomplementarity (i.e., nonspecific) binding was assessed using the noncomplementary 30-mer DNA (BS1nc; final concentration, 0.5 µg mL-1). In this work, hybridization reversibility (that is, melting of the double strands) was not investigated, since it has previously been found that the QCM crystal resonant frequency alters from its initial value (by tens of hertz) upon heating and does not revert back to its original (prior to heating) frequency when cooled.8 Due to this, hybridization reversibility on our QCM electrodes cannot be assessed via heating measurements. Further, pure water rinsing of BS1c hybridized with BS1-SH immobilized on a gold QCM does not cause removal of BS1c.8 RESULTS AND DISCUSSION Biotin-DNA Immobilization and Hybridization. The adsorption of 3,3′-dithiodipropionic acid on the QCM (Scheme 1, step 1) and its subsequent reaction with EDC and NHS were not followed by QCM because the observed QCM frequency changes for short-chain thiols and disulfides are often not reliable due to dissolution of gold from the QCM electrodes (which conceals any frequency decrease due to adsorption)25,26 and since the EDC/ NHS treatment step produces no measurable frequency shift by Analytical Chemistry, Vol. 69, No. 11, June 1, 1997

2045

Figure 1. QCM frequency change as a function of time for the adsorption of avidin from a 0.2 mg mL-1 aqueous avidin solution onto gold QCM electrodes modified as shown in Scheme 1. The arrow indicates the time at which avidin was injected into the water solution.

QCM.26 Avidin adsorption (from a 0.2 mg mL-1 aqueous avidin solution) was monitored in situ by QCM, as shown in Figure 1. The in solution frequency change (∆Fsol) is -100 Hz, and adsorption is complete within 30 min. The in air frequency change (∆Fair) for the avidin step (i.e., after drying) is -60 ( 9 Hz (average ( standard deviation for duplicate experiments). The factor of ∼2 difference between the ∆Fsol and ∆Fair values can be ascribed to water entrapped within the protein layers.20,27-29 Converting the ∆Fair value to a mass using 0.87 ng Hz-1 for our QCM operating in air8,19,20,23 yields 52 ( 8 ng for avidin on the QCM. [∆Fsol values cannot be directly transposed to mass changes; in the liquid phase, the QCM does not necessarily behave as predicted by the Sauerbrey equation due to viscoelastic effects arising from the solvent and the adsorbed layer(s).30-35 In addition, water entrapped within the layers also contributes to the observed in solution frequency changes (see above).] This experimental mass change of 52 ( 8 ng is very close to the calculated value of 54 ng for a monomolecular layer of closepacked avidin in a flat orientation (i.e., occupying an area of 33 nm2; avidin dimensions are 6.0 nm × 5.5 nm × 4.0 nm36). Subsequent exposure of dried avidin layers to biotin-DNA resulted in no biotin-DNA being immobilizedsboth ∆Fsol and ∆Fair remained unchanged, within experimental error. This suggests that avidin denatures with drying. Hence, the avidin layers were not dried for the biotin-DNA immobilization experiments. Figure 2 shows the QCM frequency response with time for the subsequent immobilization of biotin-DNA via interaction with avidin. Immobilization is confirmed by the decrease in frequency, ∆Fsol ) -18 Hz, and is complete in ∼5 min. The average frequency change ((standard deviation) for biotin-DNA im(25) Sondag-Huethorst, J. A. M.; Scho¨nenberger, C.; Fokkink, L. G. J. J. Phys. Chem. 1994, 98, 6826. (26) Geddes, N. J.; Paschinger, E. M.; Furlong, D. N.; Caruso, F.; Hoffmann, C. L.; Rabolt, J. F. Thin Solid Films 1995, 260, 192. (27) Ebara, Y.; Okahata, Y. Langmuir 1993, 9, 574. (28) Grabbe, E. S.; Buck, R. P.; Melroy, O. R. J. Electroanal. Chem. 1987, 223, 67. (29) Caruso, F.; Furlong, D. N.; Kingshott, P. J. Colloid Interface Sci., in press. (30) Nomura, T.; Okuhura, M. Anal. Chim. Acta 1982, 142, 281. (31) Bruckenstein, S.; Shay, M. Electrochim. Acta 1985, 30, 1295. (32) Kanazawa, K. K.; Gordon, J. Anal. Chim. Acta 1985, 175, 99. (33) Yang, M.; Thompson, M.; Duncan-Hewitt, W. C. Langmuir 1993, 9, 802. (34) Hinsberg, W.; Wilson, C.; Kanazawa, K. K. J. Electrochem. Soc. 1986, 133, 1448. (35) Karpovich, D. S.; Blanchard, G. J. Langmuir 1994, 10, 3315. (36) Green, N. M.; Joynson, M. A. Biochem. J. 1970, 118, 71.

2046 Analytical Chemistry, Vol. 69, No. 11, June 1, 1997

Figure 2. QCM frequency change versus time for the immobilization of biotin-DNA from HEPES buffer onto an avidin-modified QCM electrode (Scheme 1). The arrow indicates the time at which biotinDNA was injected into the HEPES solution.

Figure 3. Frequency response with time for the biotin-DNA coated QCM (Scheme 1) in HEPES/NaCl solution as a result of the addition of (a) BS1c or (b) BS1nc. The arrow indicates the time at which the BS1c or BS1nc was injected into the HEPES/NaCl solution.

mobilization for triplicate experiments is -18 ( 2 Hz, indicating that biotin-DNA is reproducibly immobilized. ∆Fair values for biotin-DNA could not be measured, as drying the avidin layer, which is required for the “before” QCM frequency measurement prior to biotin-DNA immobilization, resulted in no subsequent binding of biotin-DNA. However, assuming that the biotin-DNA frequency change in air is the same as that in solution,8 and using the in air frequency change for avidin adsorption (60 ( 9 Hz, see above), a binding ratio of 2:1 is calculated for biotin-DNA to immobilized avidin. Immersion of the avidin-coated QCM crystal in HEPES buffer, followed by injection of a 0.5 µg mL-1 aqueous biotin solution and then biotin-DNA (in fresh HEPES), showed no frequency change for the biotin-DNA step, indicating that biotin (from solution) blocked the avidin binding sites. (Biotin binding was not measured, as biotin is too small a molecule to give measurable frequency changes.) This result shows that biotinDNA is immobilized onto the avidin-coated QCM only via interaction with avidin. Exposure of the QCM with immobilized biotin-DNA to the hybridization solution HEPES/NaCl, followed by injection of a solution of complementary 30-mer DNA (BS1c), also showed a frequency decrease of 18 Hz (Figure 3, curve a). The interaction between immobilized biotin-DNA and BS1c is rapid, being complete within ∼5 min. A duplicate experiment yielded ∆Fsol ) -19 Hz, showing that the values obtained are reproducible. The frequency shifts obtained suggest hybridization of the immobilized

Figure 4. QCM frequency change versus time for the adsorption of avidin from water onto a precursor layer of (PAH/PSS)2 on MPAmodified QCM electrodes. Avidin is electrostatically adsorbed on the outer, negatively charged PSS layer. The arrow indicates the time at which avidin was injected into the water solution.

Figure 5. QCM frequency changes for the construction of alternating avidin/PSS multilayer films formed on (PAH/PSS)2 precursor films on MPA-modified gold QCM electrodes. The odd layer numbers correspond to avidin deposition and the even layer numbers to PSS. Experimental data points for two separate experiments are shown.

strands, with an immobilized probe-to-target ratio of 1:1. Exposing the nonhybridized biotin-DNA-coated QCM to a solution of noncomplementary 30-mer DNA (BS1nc) (under the same conditions that BS1c was interacted) showed a negligible frequency change (Figure 3, curve b). This confirms that the BS1c is hybridized by avidin-immobilized biotin-DNA on a QCM. (Exposure of the DNA-coated QCM crystal first to BS1nc and then to BS1c produced the same frequency changes for BS1c as those observed when the QCM was first exposed to a BS1c solution.) Avidin layers were also formed on precursor layers of (PAH/ PSS)2 on MPA-modified QCM electrodes. In these films, avidin is electrostatically immobilized on an outer, negatively charged PSS layer. Figure 4 shows the in solution frequency changes for avidin adsorption onto negatively charged PSS. A frequency change of -397 Hz is obtained at 50 min, by which time adsorption is complete. Drying this layer resulted in ∆Fair ) -212 Hz, which corresponds to a mass increase of 184 ng on the QCM surface. As before, a difference of ∼2 between the ∆Fsol and ∆Fair values is observed. The quantity of avidin on the surface suggests that multilayers/aggregates of avidin are formed and/or that penetration of the precursor film occurs (54 ng is expected for a monomolecular layer of close packed avidinssee earlier). For the biotin-DNA immobilization and hybridization experiments, drying of avidin immobilized on (PAH/PSS)2 was avoided because of avidin denaturation. The immobilization of biotin-DNA from HEPES buffer onto avidin electrostatically adsorbed onto a precursor (PAH/PSS)2 layer on MPA-modified gold electrodes yielded ∆Fsol of -15 ( 2 Hz and -11 ( 2 Hz for separate experiments (data not shown). Adsorption is complete within 5 min in each case. The frequency changes and the adsorption time are the same, within experimental error, as those obtained for biotin-DNA immobilization onto avidin attached to the QCM via Scheme 1. This suggests that, despite the large amount of avidin adsorbed on the precursor (PAH/PSS)2 layer, the equivalent of only a monolayer of avidin is actually interacting with biotin-DNA for its immobilization. Interaction of the QCM containing electrostatically bound avidin and immobilized biotin-DNA with BS1c in HEPES/NaCl buffer yielded a maximum frequency change of -13 ( 2 Hz in ∼4 min (data not shown), again implying an immobilized probe-to-target ratio of 1:1. Exposure of a similarly prepared QCM to BS1nc in HEPES/ NaCl buffer produced a frequency change of only -3 ( 2 Hz.

The above results show that complementary 30-mer DNA species can be detected in situ with DNA probes immobilized on a QCM via avidin-biotin interaction. Multilayer films of biotin-DNA were constructed in order to increase the DNA hybridization capacity of the film. These films were fabricated by the alternate deposition of avidin and PSS (up to nine layers) on a precursor 8 nm thick (PAH/PSS)2 layer on MPA-modified QCM electrodes and by then exposing them to a solution of biotin-DNA (see Experimental Section). The principle of the multilayer buildup is based on the electrostatic attraction between oppositely charged species37-42 (avidin and PSS in this work). This method has been extensively used to fabricate multilayer films of polyelectrolytes of alternating charge.23,37-42 To confirm formation of the avidin/PSS multilayer films, avidin and PSS were alternately adsorbed on a precursor (PAH/PSS)2 layer, with pure water rinsing and drying and frequency measurement between intermediate steps. This adsorption cycle was repeated until five avidin layers were deposited. Figure 5 shows the ∆Fair for avidin and PSS as a function of the number of layers and confirms the buildup of an avidin/PSS multilayer. The layer buildup is essentially linear after the first avidin/PSS deposition cycle. The average ∆Fair for the avidin and PSS layer pair, after the first avidin/PSS deposition, is -678 ( 76 Hz (-569 Hz avidin, 108 Hz PSS). The large frequency changes for avidin most likely represent aggregation and multilayers of avidin on the surface. These films were fabricated only to confirm the formation of avidin/PSS multilayers but were not used for immobilization of biotin-DNA because the avidin layers were dried; films fabricated in the same way but without intermediate drying were used instead for biotin-DNA binding and hybridization. It has recently been shown that multilayer films of anti-IgG could also be fabricated using this principle when alternately deposited with PSS.43 In that work, the anti-IgG multilayers were immunologically active, interacting with IgG, which penetrated the multilayer film. Since IgG, which is a globular protein of dimensions 10 nm × 14 (37) Decher, G.; Hong, J. D. Makromol. Chem., Macromol. Symp. 1991, 46, 321. (38) Decher, G.; Hong, J. D. Ber. Bunsenges. Phys. Chem. 1991, 95, 1430. (39) Decher, G.; Hong, J. D.; Schmitt, J. Thin Solid Films 1992, 210/211, 831. (40) Decher, G.; Schmitt, J. Prog. Colloid Polym. Sci. 1992, 89, 160. (41) Lvov, Y.; Decher, G.; Mo ¨hwald, H. Langmuir 1993, 9, 481. (42) Decher, G.; Lvov, Y.; Schmitt, J. Thin Solid Films 1994, 244, 772. (43) Caruso, F.; Niikura, K.; Furlong, D. N.; Okahata, Y. Langmuir, submitted for publication (part 2).

Analytical Chemistry, Vol. 69, No. 11, June 1, 1997

2047

Figure 6. QCM frequency change with time due to the interaction of biotin-DNA with avidin/PSS multilayer films constructed by alternate deposition of avidin and PSS on MPA/(PAH/PSS)2-coated gold QCM electrodes. The arrow indicates the time at which biotin-DNA was injected into the HEPES solution.

nm × 5 nm, was found to penetrate anti-IgG/PSS multilayers,43 it was expected that biotin-DNA (∼10 nm × 1.84 nm8) could also penetrate an avidin/PSS multilayer, producing multilayer DNAcontaining films. Furthermore, the complementary BS1-SH may similarly penetrate the film and hybridize. The QCMs with multilayers of avidin and PSS (total of five avidin and four PSS layers) were immersed in HEPES buffer for biotin immobilization experiments. Biotin-DNA was then injected and the frequency monitored with time. The QCM frequency was found to increase (Figure 6), rather than decrease. (A decrease in frequency is expected with an increase in mass on the QCM surface via immobilization.) The maximum increase in frequency of 55 Hz occurred after ∼30 min, compared to only 5 min for the immobilization of biotin-DNA onto avidin electrostatically adsorbed onto a precursor (PAH/PSS)2 layer (see earlier). This may be indicative of biotin-DNA diffusion into the multilayer film. The frequency increase may be attributed to avidin removal from the multilayer film with biotin-DNA immobilizationsweakly adsorbed/ immobilized avidin may be solubilized by biotin-DNA. Avidin removal was not observed when the QCM was placed into the hybridization buffer but only upon injection of biotin-DNA. Repeat experiments yielded the same trend and similar maximum frequency changes. Measurements to test for the removal of avidin from the QCM surface via in air frequency measurements could not be performed, because drying of avidin before injection of biotin-DNA (required for the “before” QCM frequency measurement) resulted in no subsequent binding of biotinDNA. Hence, the presence of biotin-DNA on the surface and in the multilayer film could not be confirmed by QCM frequency measurements. However, hybridization experiments were performed in order to examine the presence of biotin-DNA in the film. Figure 7 shows the QCM frequency response of a multilayer film of avidin and PSS, previously exposed to biotin-DNA, to the subsequent injection of BS1c in HEPES/NaCl solution. A frequency change of -52 Hz is obtained, suggesting binding of BS1c through hybridization. A repeat experiment yielded a frequency change of -54 Hz. To check that BS1c was not nonspecifically bound to avidin or PSS, the same experiment was repeated with BS1nc and yielded a frequency change of only -2 Hz. The experiments confirm the presence of immobilized biotin-DNA in 2048 Analytical Chemistry, Vol. 69, No. 11, June 1, 1997

Figure 7. QCM frequency change with time for the hybridization of BS1c with multilayer biotin-DNA films. These films were constructed by alternate deposition of avidin and PSS (total of nine layers) on MPA/(PAH/PSS)2-coated gold QCM electrodes, followed by exposure to a solution of biotin-DNA. The arrow indicates the time at which BS1c was injected into the HEPES/NaCl solution.

the multilayer film and hybridization between BS1c and immobilized biotin-DNA. The kinetics of hybridization for duplicate experiments were identical, with hybridization being complete in ∼20 min. Hybridization is much slower than that for the corresponding single-layer system (see earlier), possibly indicating that BS1c penetration into the film occurs. In addition, the frequency change is ∼4-5 times that observed for the singlelayer system. This shows that the sensitivity can be enhanced for a given concentration of BS1c in solution by constructing DNAcontaining multilayers. BS1-SH Immobilization and Hybridization. The electrostatic attraction between the positively charged outer layer of PAH and the negatively charged phosphate backbone of DNA was used to immobilize BS1-SH DNA on a precursor layer of PAH/PSS/ PAH on MPA-modified gold QCM electrodes. A maximum frequency change of -22 Hz was observed within 5 min (data not shown). Both the frequency change and the adsorption time are similar to those obtained for biotin-DNA binding onto singlelayer avidin films (see Figure 2). The frequency change for the BS1-SH layer in air was the same as that in solution (within experimental error) and similar to that observed for BS1-SH binding onto a bare gold QCM electrode (-24 ( 4 Hz).8 This implies a similar packing density of BS1-SH on PAH and on bare gold. (A horizontally packed monomolecular layer of BS1-SH is expected to produce an in air frequency change of ∼-14 Hz.) Hybridization of BS1-SH with BS1c occurs within 4 min and results in a frequency decrease of 13 Hz in solution (data not shown). The kinetics and magnitude of frequency shift for this hybridization reaction are also similar to those observed for biotinDNA binding onto single-layer avidin films (see earlier). Pure water rinsing did not remove the BS1c layer. Interaction of PAHimmobilized BS1-SH with BS1nc showed a negligible frequency change, i.e., within the experimental noise of the QCM ((2 Hz).

CONCLUSIONS We have demonstrated that the QCM with avidin-immobilized biotin-DNA and PAH-immobilized BS1-SH can be successfully used for the in situ detection of hybridization of a complementary 30-mer DNA oligonucleotide (BS1c). The frequency shifts indicate

an immobilized DNA probe-to-hybridized DNA target ratio of 1:1. Hybridization of BS1c with multilayer biotin-DNA films was also observed. These films displayed enhanced sensitivity with respect to the concentration of BS1c in solution, by hybridizing a greater amount at a given concentration. The data presented provide a promising basis for the development of nucleic acid sensors. In particular, the combination of a real-time monitoring device such

as the QCM and a DNA-containing multilayer makes an attractive biosensor for rapid detection of various DNA species. Received for review December 3, 1996. February 20, 1997.X

Accepted

AC961220R X

Abstract published in Advance ACS Abstracts, April 1, 1997.

Analytical Chemistry, Vol. 69, No. 11, June 1, 1997

2049