Kinetic Measurements of DNA Hybridization on an Oligonucleotide

Publication Date (Web): February 28, 1998 ... Detection of Sub-fM DNA with Target Recycling and Self-Assembly Amplification on ... Protein Binding Kin...
0 downloads 0 Views 267KB Size
Download PDF
Anal. Chem. 1998, 70, 1288-1296

Kinetic Measurements of DNA Hybridization on an Oligonucleotide-Immobilized 27-MHz Quartz Crystal Microbalance Yoshio Okahata,* Masanori Kawase, Kenichi Niikura, Fuyuka Ohtake, Hiroyuki Furusawa, and Yasuhito Ebara

Department of Biomolecular Engineering, Tokyo Institute of Technology, Nagatsuda, Midori-ku, Yokohama 226-8507, Japan

A highly sensitive 27-MHz quartz-crystal microbalance, on which a 10-30-mer oligonucleotide was immobilized as a probe molecule, was employed to detect hybridization of complementary oligonucleotides in aqueous solution. From frequency decreases (mass increases due to the hybridization) with passage of time, kinetic parameters such as association constants (Ka) and binding and dissociation rate constants (k1 and k-1) could be obtained, as well as binding (hybridization) amount at the nanogram level (∆m). Kinetic studies were carried out by changing various parameters: (i) the immobilization method of a probe oligonucleotide on Au electrode, (ii) number of mismatching bases in sequences of target oligonucleotides, (iii) length of both probe and target oligonucleotides, (iv) hybridization temperature, and (v) ionic strength in solution. The obtained results were compared with those obtained by a surface plasmon resonance method using a BIAcore system. Reversible hybridization of complementary DNAs and/or RNAs is fundamental to the biological processes of replication, transcription, and translation. Physical kinetic studies of nucleic acid hybridization are necessary for understanding these important biological processes on a molecular level. DNA hybridization studies in a homogeneous solution require the ability to separately monitor paired and unpaired nucleic acids. DNA hybridization in solution may be monitored by several methods such as hypochromisity (UV spectra),2 circular dichroism spectra,2 calorimetry,3 fluorescent probes,4 and NMR spectra.5 All of these techniques, however, require relatively large amounts (several micrograms per milliliter) of nucleic acids and separation processes of paired and unpaired nucleic acids. Nucleic acid hybridization on a solid surface has as an advantage the use of a small amount of DNA sample, in which an oligonucleotide probe is immobilized on a nitrocellulose or a nylon (1) Preliminary report: Okahata, Y.; Matsunobu, Y.; Ijiro, K.; Mukai, M.; Murakami, A.; Makino, K. J. Am. Chem. Soc. 1992, 114, 8299. (2) Bush, C. A. In Basic Principles in Nucleic Acid Chemistry; Ts’o, P. O. P., Ed.; Academic Press: New York, 1976; pp 91-169. (3) Breslaurer, K. J. In Thermodynamic Data for Biochemistry and Biotechnology; Hinz, H.-J., Ed.; Springer-Verlag: New York, 1986; pp 402-427. (4) (a) Yguerabide, J.; Ceballos, A. Anal. Biochem. 1995, 228, 208-220. (b) Tyagi, S.; Kramer, F. R. Nature Biotechnol. 1996, 14, 303-309. (c) Morrison L. E.; Stols, L. M. Biochemistry 1993, 32, 3095-3104. (5) Patel, D. J.; Pardi, A.; Itakura, K. Science 1982, 216, 581-590.

1288 Analytical Chemistry, Vol. 70, No. 7, April 1, 1998

filter membrane and target DNAs are added as droplets.6 These solid-phase methods, however, are necessary to separate paired and unpaired nucleic acids by washing or filtration, and target DNAs are usually labeled either directly or indirectly with fluorescent or radioactive molecules to detect hybridization amounts on a solid surface. It also takes a relatively long time to analyze the results. It is difficult to detect quantitatively the absolute amount of hybridization with fluorescent-labeling and radiolabeling methods, where in reality, these are used routinely as highly accurate methods of quantifying hybridization. In our preliminary report, we proposed a new methodology to detect DNA hybridization between an oligonucleotide immobilized on a 9-MHz quartz crystal microbalance (QCM) and a large target M13 phage single-strand DNA in aqueous solutions from frequency changes of the QCM.1 QCMs are known to provide very sensitive mass-measuring devices in gas phase7,8 and in aqueous (6) (a) Dunn, A. R.; Hassel, J. A. Cell 1977, 12, 23. (b) Meinkoth, J.; Wahl, G. Anal. Biochem. 1984, 138, 267. (c) Syvanen, A.-C.; Laaksonen, M.; Soderlund, H. Nucleic Acids Res. 1986, 14, 5037. (d) Keller, G. H.; Cumming, C. U.; Huang, D.-P.; Manak, M. M.; Ting, R. Anal. Biochem. 1988, 170, 441. (e) Chan, V.; Graves, D. J.; McKenzie, S. E. Biochem. J. 1995, 69, 2243-2255. (7) (a) Guilbault, G. G. Anal. Chem. 1983, 55, 1682. (b) Schierbaum, K. D.; Weiss, T.; Thoden von Velzen, E. U.; Engbersen, J. F. J.; Reinhoudt, D. N.; Go ¨pel, W. Science 1994, 265, 1413. (c) Yang, H. C.; Dermody, D. L.; Xu, C.; Ricco, A. J.; Crooks, R. M. Langmuir 1996, 12, 726. (d) Wells, M.; Dermody, D. L.; Yang, H. C.; Kim, Y. T.; Crooks, R. M. Langmuir 1996, 12, 1989. (e) Grate, J. W.; Patrash, S. J.; Abraham, M. H.; Du, C. M. Anal. Chem. 1996, 68, 913. (8) (a) Okahata, Y.; Ebato, H.; Taguchi, K. J. Chem. Soc., Chem. Commun. 1987, 1363-1365. (b) Okahata, Y.; Kimura, K.; Ariga, K. J. Am. Chem. Soc. 1989, 111, 9190-9194. (c) Okahata, Y.; Ebato, H. J. Chem. Soc., Perkin Trans. 2 1991, 457-479. (d) Okahata, Y.; Ebato, H. Anal. Chem. 1991, 63, 203207. (e) Okahata, Y.; Ebato, H. Trends Anal. Chem. 1992, 11, 344-354. (f) Okahata, Y. In Olfaction and Taste IV; Kurihara, X., Suzuki, N., Ogawa, H., Eds,; Springer-Verlag: Tokyo, 1994; pp 703-707. (g) Matsuura, K.; Ebara, Y.; Okahata. Y. Thin Solid Films 1996, 273, 61-65. (h) Okahata, Y.; Matsuura, K.; Ito, K.; Ebara, Y. Langmuir 1996, 12, 1023-1026. (i) Okahata, Y.; Matsuura, K.; Ebara, Y. Supramol. Sci. 1996, 3, 165-169. (j) Matsuura, K.; Okahata, Y. Chem. Lett. 1996, 119-120. (k) Matsuura, K.; Ebara, Y.; Okahata, Y. Langmuir 1997, 13, 814-820. (9) (a) Thompson, M.; Arthur, C. L.; Dhaliwal, G. K. Anal. Chem. 1986, 58, 1206. (b) Muramatsu, H.; Dicks, J. M.; Tamiya, E.; Karube, I. Anal. Chem. 1987, 59, 2760. (c) Ebersole, R. C.; Ward, M. D. J. Am. Chem. Soc. 1988, 110, 8623. (d) Tamiya, E.; Suzuki, M.; Karube, I. Anal. Chim. Acta 1989, 217, 321. (e) Ebersole, R. C.; Miller, J. A.; Moran, J. R.; Ward, M. D. J. Am. Chem. Soc. 1990, 112, 3239. (f) Yamaguchi, S.; Shimomura, T.; Tatsuma, T.; Oyama, N. Anal. Chem. 1993, 65, 1925. (g) Ebato, H.; Gentry, C. A.; Herron, J. N.; Mueller, W.; Okahata, Y.; Ringsdorf, H.; Suci, P. A. Anal. Chem. 1994, 66, 1683. (h) Wakamatsu, K.; Hosoda, K.; Mitomo, H.; Ooya, M.; Okahata, Y.; Yasunaga, K. Anal. Chem. 1995, 67, 3336. S0003-2700(97)00584-2 CCC: $15.00

© 1998 American Chemical Society Published on Web 02/28/1998

solution.9,10 Their resonance frequency was probed to decrease linearly upon the increase of mass on the QCM electrode at the nanogram level in the air phase.11 We had prepared a 10-mer nucleotide having a mercaptopropyl group at the 5′-phosphate end, whose sequence is complementary with the EcoRI binding site of M13 phage ssDNA, and immobilized it on an Au electrode of a QCM plate by using an Au-S interaction. We could observe the selective binding of a large single-strand M13 phage DNA (7249mer) to the 10-mer nucleotide immobilized on a QCM plate.1 Several researchers have reported nucleotide hybridizations by using microgravimetric techniques including a QCM system, in which the hybridization phenomena have been described qualitatively but the kinetic analyses have hardly been performed.12 Another commercially available and in situ biomolecular interaction analysis (BIA) is the BIAcore system (Pharmacia Biotech Co.).13 The basis of the BIAcore system is the optical phenomenon of surface plasmon resonance (SPR), in which the angle of incident beam required for SPR changes depends on the binding amount of guest molecules due to the change of refractive index and other factors near the Au surface. Although BIAcore/ SPR techniques have derived quantitative information from refractive index changes, these methods have been largely used qualitatively in the past because optical phenomena are affected by various factors near the interface of the surface and solution. In this paper, we report quantitative kinetic studies of DNADNA hybridization from time dependencies of frequency decrease (mass increase) of the 10-30-mer nucleotide-immobilized highly sensitive 27-MHz QCM system, responding to the addition of oligonucleotides in buffer solutions (see Figure 1). Kinetic studies were carried out by changing (i) the immobilization method of the oligonucleotide on Au electrode, (ii) the number of mismatching bases in sequences of target oligonucleotides, (iii) the chain length of both probe and target oligonucleotides, (iv) the hybridization temperature, and (v) the ionic strength in solution. The 27-MHz QCM used in this study is ∼10 times more sensitive than the our conventional 9-MHz QCM,8,10 and has a sensitivity of 0.6 ng cm-2 of mass change per 1 Hz of frequency decrease.14 This sensitivity is enough to detect a mass change of 50-100 ng cm-2 (10) (a) Okahata, Y.; Ebato, H.; Ye, X. J. Chem. Soc., Chem. Commun. 1988, 1037-1038. (b) Ebara, Y.; Okahata. Y. Langmuir 1993, 9, 574-576. (c) Okahata, Y.; Ijiro, K.; Matsuzaki, Y. Langmuir 1993, 9, 19-21. (d) Okahata, Y.; Matsuzaki, Y.; Ijiro, K. Sens. Actuators B 1993, 13, 380-383. (e) Ebara, Y.; Okahata, Y. J. Am. Chem. Soc. 1994, 116, 11209-11212. (f) Sato, T.; Serizawa, T.; Okahata, Y. Biochem. Biophys. Res. Commun. 1994, 204, 551. (g) Okahata, Y.; Yasunaga, K.; Ogura, K. J. Chem. Soc., Chem. Commun. 1994, 469. (h) Okahata, Y.; Ebara. Y. In Current Topics in Biophysics, Frangopol, T., Sanduloviciu, M., Eds.; Iasi University Press: Romania, 1995; Vol. 3, pp 152-171. (i) Sato, T.; Serizawa, T.; Okahata, Y. Biochim. Biophys. Acta 1996, 1285, 14. (j) Niikura, K.; Nagata, K.; Okahata, Y. Chem. Lett. 1996, 863-864. (11) Sauerbrey, G. Z. Phys. 1959, 155, 206. (12) (a) Fawcett, N. C.; Evans, J. C.; Chien, L. C.; Drozda, K. A.; Flowers, N. Sens. Technol. 1988, 4, 5. (b) Ward, M. R.; Buttry, M. Science 1990, 249, 1000. (c) Su, H.; Kallury, M. R. Thompson, M.; Roach, A. Anal. Chem. 1994, 66, 769-777. (d) Yamaguchi, S.; Shimomura, T. Anal. Chem. 1993, 65, 1925. (e) Su, H.; Thompson, M. Biosens. Bioelectron. 1995, 10, 329-340. (f) Ito, K.; Hashimoto, K.; Ishimori, Y. Anal. Chim. Acta 1996, 327, 29. (13) (a) Ikuta, S.; Takagi, K.; Wallace, B. R.; Itakura, K. Nucleic Acids Res. 1987, 15, 797-811. (b) Jost, J.-P.; Munch, O.; Anderson, T.; Nucleic Acids Res. 1991, 15, 731-739. (c) Wood, S. J. Microchem. J. 1993, 47, 330-337. (d) Gotoh, M.; Hasegawa, Y.; Shinohara, Y.; Shimizu, M.; Tosu, M. DNA Res. 1995, 2, 285-293. (e) Bates, P. J.; Dosanjh, H. S.; Kumar, S.; Jenkins, T. C.; Laughton, C. A.; Neide, S. Nucleic Acids Res. 1995, 23, 3627-3632. (f) Nilson, P.; Persson, B.; Uhler, M.; Nygren, P. A. Anal. Biochem. 1995, 224, 400-408.

Figure 1. Schematic illustration of DNA hybridization on an oligonucleotide-immobilized highly sensitive 27-MHz QCM.

obtained by binding of small oligonucleotides. We also compared our kinetic parameters obtained by the QCM method with those obtained by a SPR method by using a BIAcore system (Pharmacia Biotech Co.). EXPERIMENTAL SECTION Materials. Oligonucleotides were automatically synthesized using a DNA synthesizer (Pharmacia Biotech AB) and commercially available amidites. Oligonucleotides having a disulfide group were prepared by introducing the (β-hydroxyethyl)dithioethoxy group to the 5′-phosphate end with water-soluble carbodiimide. 5′-Biotinylated oligonucleotides were prepared using biotinylating amidite (Biodite, Pharmacia Biotech AB). All the resultant oligonucleotides were purified by anion-exchange chromatography, and their concentrations were determined by an optical density measurement taken at 260 nm. Avidin was obtained from Calzyme and other chemicals were purchased from Tokyo Kasei, Co. or Sigma Co. and used without further purification. Calibration of a 27-MHz QCM. A 27-MHz, AT-cut QCM is commercially available from Showa Crystals Co. (Chiba, Japan). The diameter of its quartz plate is 8 mm, and Au electrodes are (14) Ebara, Y.; Itakura, K.; Okahata, Y. Langmuir 1996, 12, 5165-5170.

Analytical Chemistry, Vol. 70, No. 7, April 1, 1998

1289

deposited on both sides (diameter, 2.5 mm; area) 4.9-mm2 ).14 The one side of the quartz crystal was sealed with a rubber casing, maintaining it in an air environment to avoid contact with the ionic aqueous solution, while the other is exposed to aqueous buffer solution (see Figure 1).8,10 A cased 27-MHz QCM was connected to an oscillation circuit designed to drive the quartz in aqueous solution. The frequency changes were followed by a universal frequency counter (Hewlett-Packard Co., Ltd., Tokyo, model 53131A) attached to a microcomputer system (Macintosh Power Book 170, Apple).8,10 The following Sauerbrey equation was obtained for the ATcut shear mode QCM in the air phase,11

∆F ) -

2F02

∆m AxFqµq

(1)

where ∆F is the measured frequency change (Hz), Fo the fundamental frequency of the QCM (27 × 106 Hz), ∆m the mass change (g), A the electrode area (4.9-mm2 ), Fq the density of quartz (2.65 g cm-3), and µq the shear modus of quartz (2.95 × 1011 dyn cm-2). When the QCM is employed in an aqueous solution, eq 1 cannot be simply applied due to the effects of interfacial liquid properties (i.e., density, viscosity, conductivity, and dielectric constant),15 thin-film viscoelasticity,16 electrode morphology,16,17 and mechanism of acoustic coupling18 on the QCM oscillation behavior. Therefore, we calibrated our cased 27-MHz QCM in the aqueous solution (pH 7.9, 10 mM Tris-HCl, 0.2 M NaCl, at 20 °C), analogously to the calibration of our conventional 9-MHz QCM.8,10 A respective amount of polymer solution was cast (0.1-15 µg cm-2), an LB film of lipid monolayers was deposited (0.04-4 µg cm-2), or 10-30-mer oligonucleotides having SS groups were directly immobilized (10-60 ng cm-2) on the bare Au electrode side of the cased QCM plate. One linear relationship was observed between the deposited amount of mass and the frequency decrease of the QCM both in air phase and in the aqueous solution, independent of deposition methods and chemical compounds. The slope of this curve showed that a frequency decrease of 1 Hz corresponded to a mass increase of 0.61 ( 0.1 ng cm-2 on the QCM electrode.14 This means that the Sauerbrey equation can be applied in the aqueous solution limited by these experimental conditions. When a long-chain ssDNA, such as 100-300 mer, was immobilized on the QCM, however, the linear correlation could not be obtained in the aqueous solution due to the effect of the flexible conformation of the long DNA chains on the electrode. The sensitivity for the mass change of a 27-MHz QCM was increased by ∼10 times, in comparison to our conventional 9-MHz (15) Nomura, T.; Okuhara, M. Anal. Chim. Acta 1982, 142, 281. (b) Bruckenstein, S.; Shay, M. Electrochim. Acta 1985, 30, 1295. (c) Kanazawa, K. K.; Gordon, J. Anal. Chim. Acta 1985, 175, 99. (16) (a) Yang, M.; Thompson, M.; Duncan-Hewitt, W. C. Langmuir 1993, 9, 802. (b) Hinsberg, W.; Wilson, C.; Kanazawa, K. K. J. Electrochem. Soc. 1986, 133, 1448. (c) Okahata, Y.; Ebato, H. Anal. Chem. 1989, 61, 2185. (d) Okahata, Y.; Kimura, K.; Ariga, K. J. Am. Chem. Soc. 1989, 111, 9190. (17) Yang, M.; Thompson, M. Langmuir 1993, 9, 1990. (b) Urbakh, M.; Daikhin, L. Langmuir 1994, 10, 2836. (18) (a) Muramatsu, H.; Tamiya, E.; Karube, I. Anal. Chem. 1988, 60, 2142. (b) Duncan-Hewitt, W. C.; Thompson, M. Anal. Chem. 1992, 64, 94.

1290 Analytical Chemistry, Vol. 70, No. 7, April 1, 1998

QCM (∆m ) 6.5 ng cm-2).8,10 The noise level of the 27-MHz QCM was (5 Hz in buffer solution at 20 °C, and the standard deviation of the frequency was 10 Hz for 12 h in the same condition. These values were the same level as the conventional 9-MHz QCM.8,10 Immobilization of Nucleotide Probes on a QCM. The bare Au electrode side of the QCM plate was cleaned with a solution of 25% H2O2/75% H2SO4 (piranha solution) to remove organic adsorbate impurities from the gold surfaces and was rinsed with Milli-Q water several times. A direct immobilization of a probe oligonucleotide on the Au electrode was simply carried out by soaking the QCM in a 0.3 M NaCl aqueous solution of 10-30-mer nucleotides having S-S groups for several hours at 25 °C (see Figure 1B). The immobilization amount was followed by a frequency decrease (mass increase) in the solution. When the frequency was decreased ∼100 Hz (mass increase of ∼60 ng cm-2) after 2-3 h, the QCM was picked up in the air phase to stop the immobilization and washed with buffer solution several times. The immobilized amount of 60 ng cm-2 on the Au electrode (4.9-mm2 ) is calculated to be 30% coverage of single-strand nucleotide (area per molecule, ∼2.2 nm2) on the electrode. In the direct immobilization method, some of nonspecific binding may be included. The immobilized oligonucleotides, however, hardly released from the QCM during soaking in the buffer solution for 1 day (∆m < 5 ng cm-2). Immobilization of biotinylated oligonucleotide (an avidin-biotin method) was carried out as shown in Figure 2. The cleaned Au electrode of the QCM was soaked in an aqueous solution (3 mL) of 3,3′-dithiopropionic acid (1 mM) at room temperature and the frequency decrease was saturated at about 100-150 Hz (mass increase of 60-90 ng cm-2) after 20 min (step 1). This means that 3,3′-dithiopropionic acid (area per molecule, 0.4 nm2) covered roughly as a monolayer on the electrode (4.9-mm2 ). Before drying, the carboxylic acid on the QCM was reacted with N-hydroxysuccinimide in the presence of water-soluble carbodiimide [EDC, 1-ethyl-3-[3-(dimethylamino)propyl]carbodiimide] in the aqueous solution (step 2). The frequency attained equilibrium at a decrease of 50-100 Hz after 30 min. The QCM having the activated carboxyl groups was immersed in 1 mL of the aqueous buffer solution (pH 7.9, 10 mM Tris-HCl, 0.2 M NaCl) of avidin (10 µg, MW 68 000) in step 3. The frequency decrease reached equilibrium at 800-1000 Hz (480-600 ng cm-2) in 1 h. This means that avidin (section area, ∼80 nm2) was estimated to bind as a monolayer on the electrode (25-30 ng on 4.9-mm2 ). Avidin remained on the electrode after rinsing with aqueous solution several times (∆F < -100 Hz). The QCM was immersed in the aqueous solution (1 mL) of ethanol amine (1 M) for 30 min to deactivate the carboxyl group as β-hydroxyethylamide (step 4). The avidin-immobilized QCM was immersed in 1 mL of the aqueous buffer solution (pH 7.9, 10 mM Tris-HCl, 0.2 M NaCl) of biotinylated oligonucleotides (10-30-mer, 1 µM) at 25 °C (step 5). After 30 min, the frequency decreased ∼100 ( 10 Hz (mass increase of 60 ( 5 ng cm-2) in the case of a 10-mer nucleotide. This means that one or two biotinylated oligonucleotides can bind to four binding sites of an avidin molecule. Hybridization on a 27-MHz QCM. A probe oligonucleotideimmobilized QCM was soaked into 5 mL of aqueous solution (pH 7.9, 10 mM Tris-HCl, 0.2 M NaCl) at the respective temperature (usually 20 °C), and the resonance frequency of the QCM was

Figure 2. Immobilization method of oligonucleotides on Au electrode by using an avidin-biotin interaction.

defined as zero position after the equilibrium. The stability and the drift of the 27-MHz QCM frequency in the solution were (5 Hz for 12 h at 20 °C. The frequency change of the QCM responding to the addition of 10-100 µL of aqueous solution of target oligonucleotide was recorded with time. The solution was stirred to avoid any effect of diffusion of guest molecules, and the stirring did not affect the stability and the amount of frequency changes. Measurements by Using a BIAcore System. The BIAcore system is composed of an optical detector system, a removable sensor tip coated with Au, and a microfluidic cartridge that controlled the automatic flow of sample solution onto the sensor tip surface. The instrument is controlled by a personal computer with a data evaluational program. A commercially available Au sensor tip is covered by a dextran layer having carboxymethyl groups, usually on which probe molecules are immobilized covalently through amide linkages. To compare hybridization experiments under the same conditions, a dextran layer on Au electrode was completely removed by washing with piranha solution several times. On the bare Au sensor tip, a probe oligonucleotide was immobilized in the same manner as the QCM plate according to the scheme in Figure 2, except the immobilization was carried out in the flow cell. A biotinylated oligonucleotide was also immobilized on avidin that was covalently bonded in a dextran layer on a commercially available sensor tip, as comparison. In BIAcore system, 1000 RU has been reported to correspond to ∼100 ng cm-2 by the radiolabeled calibration method.19 Thus, the immobilized amount of oligonucleotide corresponded to 70 ng cm-2, which was consistent with that of a 10-mer nucleotide on a QCM plate. Hybridization measurements were carried out in an aqueous solution (pH 7.9, 10 mM Tris-HCl, 0.2 M NaCl) of target oligonucleotide at 20 °C at a flow rate of 2 µL/min for 30 min. The SPR signal (∆RU) resultant to the changes of the refractive index near the sensor surface was followed with time. RESULTS AND DISCUSSION Figure 3 shows a typical frequency decrease with passage of (19) Stenberg, E.; Persson, B.; Roos, H.; Urbaniczky, C. J. Colloid Interface Sci. 1991, 143, 513-526.

Figure 3. Typical time dependencies of frequency decreases of the 5′GGGAATTCGT3′-immobilized QCM responding to the addition of (a) a fully complementary 3′CCCTTAAGCA5′ and (b) 3′CCCTGAAGCA5′ having one mismatching base (underlined) in the middle of the sequence. Conditions: 10 mM Tris-HCl, pH 7.9, 0.2 M NaCl, 20 °C, [immobilized probe] ) 60 ng cm-2 (1 pmol on 4.9-mm2 electrode through an avidin-biotin interaction), and [target nucleotide] ) 0.19 µM (3.0 µg in 5 mL).

time (∆F) of the 27-MHz QCM, on which a 10-mer nucleotide (5′GGGAATTCGT3′) is immobilized through an avidin-biotin interaction, responding to the addition of a complementary 10mer 3′CCCTTAAGCA5′ or a 10-mer nucleotide having one mismatching base in the middle of the sequence (3′CCCTGAAGCA5′, the mismatching base is underlined) in the aqueous buffer solution (10 mM Tris-HCl, pH 7.9, and 0.2 M NaCl) at 20 °C. The probe nucleotide was immobilized with a long spacer (∼5 nm) of a large avidin and a spacer alkyl chain to the Au plate by Au-S interaction as described in the Experimental Section (see Figures 1A and 2). The immobilized amount of a probe nucleotide was 60 ng cm-2 (3 ng, 1 pmol on a 4.9-mm2 Au electrode). When a large excess of complementary 10-mer nucleotide was injected (0.19 µM, 3.0 µg in 5 mL), the frequency gradually decreased in time and saturated at -∆F ) 80 ( 5 Hz (∆m ) 48 ( 2 ng cm-2) within 20 min. After the QCM plate was exposed to the air phase Analytical Chemistry, Vol. 70, No. 7, April 1, 1998

1291

the intercept of the linear correlation of Figure 4b, respectively. Ka and ∆mmax values were obtained from a saturation method to be 1.2 × 106 M-1 and 48 ( 5 ng cm-2, respectively (see Table 1). The ∆mmax value was roughly consistent with the immobilized amount of the probe nucleotide (60 ( 5 ng cm-2), which means a 1:1 binding between a target and probe nucleotides on the QCM plate. Binding kinetics can be also calculated from the time dependencies of frequency decreases (mass increases) in Figure 3. The binding between a probe and a target nucleotide is described by eq 3 k1

}[hybridization] [target] + [probe] {\ k -1

(3)

The hybridization amount formed at time t after injection is given by eqs 4-6.8g-8k,14

Figure 4. (A) Saturation hybridization behavior on a probe of 5′GGGAATTCGT3′-immobilized QCM depending on target 3′CCCTTAAGCA5′ concentration. (B) Linear reciprocal plot of [target]/∆m against [target].

and dried carefully it in air, the binding amount of the target 10mer nucleotide was roughly ∆m ) 50 ( 10 ng cm-2, obtained from the frequency decrease in the air phase before and after hybridization. This value was consistent with the in situ binding amount in the aqueous solution. This means that the effect of hydration or not-evaporated water molecules near nucleotides and proteins is small. When an excess amount (0.4 µM) of the 10-mer nucleotide having the same sequence of a probe nucleotide (5′GGGAATTCGT3′) was present in the solution in advance, binding of the target oligonucleotide was inhibited. This is due to the doublestrand formation in the solution, and a target oligonucleotide cannot bind to a probe oligonucleotide on a QCM. Binding of a 10-mer nucleotide having one mismatching base was minimal (∆m < 10 ng cm-2, curve b in Figure 3). When only avidin was immobilized on the QCM without a biotinylated oligonucleotide, nucleotide molecules were hardly bound. These results indicate that the complementary hybridization is observed in curve a of Figure 3. When the injected concentration of the target 10-mer DNA was increased in the solution, the binding amount (∆m) vs the target concentration in the solution showed a simple saturation curve containing one component, as shown in Figure 4a.

[target]0 [target]0 1 ) + ∆m ∆mmax ∆mmaxKa

(2)

Reciprocal plots between [target]0/∆m and [target]0 gave a simple straight line according to eq 2, where [target]0 indicates the initial concentration of target nucleotide. Association constants (Ka) and maximum binding amount (∆mmax) of the target nucleotide to the probe nucleotide was calculated from the slope and 1292 Analytical Chemistry, Vol. 70, No. 7, April 1, 1998

[hybridization]t ) [hybridization]∞(1 - e-(1/τ)t)

(4)

∆mt ) ∆mmax(1 - e-(1/τ)t)

(5)

τ-1 ) k1[target]0 + k-1

(6)

Figure 5 shows a linear correlation of the reciprocal of the relaxation time (τ-1) of the binding against various concentrations of target nucleotides (eq 6). k1 and k-1 could be obtained from the slope and intercept of Figure 5 to be 2.4 × 104 M-1 s-1 and 0.02 s-1, respectively. The association constant Ka was calculated to be 1.2 × 106 M-1 from k1/k-1. This Ka value obtained from the curve-fitting method according to eqs 4-6 was fairly consistent with that obtained from a saturation method (Ka ) 1.2 × 106 M-1, see eq 2 and Figure 4). Data are summarized in Table 1 (run 1). Comparison of Kinetic Parameters with BIAcore and Other Methods. Comparative hybridization behavior using a BIAcore system is shown in Figure 6. Shown are typical time dependencies of ∆RU changes of the sensor immobilized with the biotinylated 10-mer nucleotide (5′GGGAATTCGT3′, ∼60 ng cm-2) on an avidin monolayer, responding to binding with the complementary 10-mer nucleotide in the flow cell. At arrow A, flow of the target nucleotide solution was started, and at arrow B, buffer solution flow was reinstituted. With the complementary 10-mer nucleotide clear binding and dissociation behavior were observed. In contrast, with the 10-mer nucleotide having one mismatching base in the middle of the sequence, binding behavior was minimal. Kinetic parameters could be obtained from the curve-fitting method as well as from the QCM measurements using eqs 3-6. The obtained association constants (Ka) and binding and dissociation rate constants (k1 and k-1) are summarized in Table 1 (run 3). The obtained kinetic parameters were fairly consistent with those obtained from the QCM method (run 1). The immobilization method and the nucleotide sequences used are the same, but only the measurement concept is different in both the QCM (mass change) and SPR (refractive index change) methods. The commercially available sensor tip of BIAcore was coated with a dextran layer having carboxymethyl groups as side chains to immobilize biomolecules.13 To compare the results obtained

Table 1. Binding and Dissociation Rate Constants (k1 and k1 ) and Association Constants (Ka) of a Target 3′CCCTTAAGCA5′ to a Probe Biotinylated-5′GGGAATTCGT3′ on a QCM at 20 °Ca run

instrument

immobilization method

k1b (103 M-1s-1)

k-1b (10-3 s-1)

Kac (106 M-1)

∆mmaxd (ng cm-2)

1 2 3 4

27-MHz QCM 27-MHz QCM BIAcore BIAcore

avidin-biotin direct avidin-biotin on a bare Aue avidin-biotin in dextran matrixf

24 4.7 62 55

20 0.91 7.4 15

1.2 (1.2)d 5.2 8.4 3.8

48 43

a 10 mM Tris, pH 7.8, 0.2 M NaCl, 1 mM EDTA, 20 °C. Each run was examined at least three times and the experimental errors were within (10%. b Obtained from eq 4 (a curve-fitting method). c Obtained from k1/k-1. d Obtained from eq 2 (a saturation method). e Biotinylated nucleotide probes were immobilized on an avidin monolayer on a bare Au surface of a sensor tip. f Biotinylated nucleotide probes were immobilized on an avidin that is covalently bonded to carboxymethyl groups in a swelled dextran layer.

Figure 5. Linear reciprocal plot of relaxation time (τ) against [target]/ ∆m according to eq 6.

when avidin was directly immobilized on the bare Au surface, we also immobilized avidin in the carboxylic acid groups in the swelled dextran layer and the biotinylated 10-mer nucleotide was introduced in the same manner. Kinetic parameters are also summarized in Table 1 (run 4). Similar binding parameters were obtained when the nucleotide was immobilized both on the surface of the bare Au plate and in the swelled dextran layer on the sensor tip. We also compared the results obtained on the solid surface of the QCM or SPR with those obtained in the homogeneous solution. Morrison and Stols4c obtained kinetic parameters by a fluorescent-quenching method of hybridization of fluorescent probe-introduced 10-mer nucleotides between 5′AACCATCAGG3′ and 3′TTGGTAGTCC5′ in the aqueous solution (pH 8.0, 0.1 M NaCl, at 20 °C): k1 ) 8 × 104 M-1 s-1, k-1 ) 7 × 10-2 s-1, and Ka ) 1.1 × 106 M-1. These values are close to the kinetic parameters in Table 1 (runs 1, 3, and 4) obtained in the hybridization through an avidin-biotin immobilization on the solid surface. These results indicate that effects of the solid surface and steric hindrance are relatively small when a target nucleotide is small and a probe nucleotide is immobilized through a long spacer of avidin-biotin interaction. Effect of Spacer Length of Probe Nucleotides. A probe 10-mer 5′GGGAATTCGT3′ was immobilized different two ways (see Figure 1A and B). For the avidin-biotin method, the probe nucleotide was immobilized with a long spacer group (∼5 nm) of a large avidin on the Au plate. For the direct immobilization method, a probe nucleotide (∼3-nm length) was bound directly at the SS group linked to the 5′-end phosphate group of nucleotide

Figure 6. Typical ∆RU changes with passage of time of a BIAcore sensor tip immobilized with biotinylated-5′GGGAATTCGT3′ responding to the flow of (a) fully complementary 3′CCCTTAAGCA5′ and (b) 3′CCCTGAAGCA5′ having one mismatching base (underlined) in the middle of the sequence. The probe nucleotide was immobilized with an avidin-biotin interaction on a bare Au sensor tip. The target oligonucleotide flow started at the arrow A and the flow was changed to buffer solution at arrow B. Conditions: 10 mM Tris-HCl, pH 7.9, 0.2 M NaCl, 1 mM EDTA, at 20 °C, [immobilized probe] ) ∼60 ng cm-2 on the bare sensor tip, and [target nucleotide] ) 0.5 µM.

(spacer length ∼0.4 nm). Figure 7 shows typical time dependencies of hybridization between a probe nucleotide that was immobilized directly on the QCM and complementary or noncomplementary target nucleotide. The sequences of target and probe nucleotides and the immobilization amount of the target nucleotide on the QCM were the same as those for the avidin-biotin method (60 ( 5 ng cm-2). The very similar time dependencies of frequency changes were observed as compared with Figure 3, except for the time scale difference for the hybridization equilibrium. Thus, in the direct immobilization method, hybridization became very slow, taking ∼4 h to reach the binding equilibrium, compared with ∼20 min when the probe was immobilized with a long spacer of the avidinbiotin linkage in Figure 3. The maximum binding amount (∆mmax), association constant (Ka), and binding and dissociation rate constants (k1 and k-1) were obtained from both the saturation methods (eq 3) and the curvefitting method (eq 6) by changing concentrations of the target DNA, and the results are summarized in Table 1 (run 2). The Ka and ∆m values obtained from the direct immobilization method were relatively close to those obtained from the avidin-biotin Analytical Chemistry, Vol. 70, No. 7, April 1, 1998

1293

Figure 7. Time dependencies of frequency decreases of the QCM directly immobilized with 5′GGGAATTCGT3′ responding to the addition of (a) a fully complementary 3′CCCTTAAGCA5′ and (b) 3′CCCTGAAGCA5′ having one mismatching base (underlined) in the middle of the sequence. Conditions: 10 mM Tris-HCl, pH 7.9, 0.2 M NaCl, 20 °C, [immobilized probe] ) 60 ng cm-2 (1 pmol on 4.9-mm2 electrode), and [target nucleotide] ) 0.19 µM (3.0 µg in 5 mL).

immobilization method. However, both k1 and k-1 values were largely decreased in comparison with those obtained in the avidin-biotin immobilization method (run 1). Thus, in the direct immobilization method, the binding rate was decreased probably due to the steric hindrance near the substrate surface and the dissociation rate was also decreased probably due to the same interaction with the substrate surface. Since both ∆mmax and Ka values were similar to those obtained from the avidin-biotin immobilization method, the hybridization ability of the direct immobilized nucleotide was similar to that of the avidin-biotin immobilized nucleotide. The probe oligonucleotides were immobilized on a QCM by using the avidin-biotin method in the following experiments. Effect of Ionic Strength. Hybridizations between 10-mer probe 5′GGGAATTCGT3′ immobilized by an avidin-biotin method and the target 3′CCCTTAAGCA5′ were studied in different NaCl concentrations (0-0.5 M). The fundamental frequency of the 27MHz QCM was not affected by the NaCl concentration. Figure 8 shows the time course of hybridization processes in 0, 0.1, and 0.5 M NaCl as typical examples. Hybridization could be hardly observed without NaCl. The apparent binding amount was increased gradually with increasing ionic strength. The kinetic parameters k1, k-1, and Ka were obtained from the curve-fitting method (eq 6) in different target nucleotide concentrations, and the results are summarized in Table 2. The hybridization percentage increased gradually with increasing NaCl concentrations, because of the decrease in electric repulsion between nucleotides. When the NaCl concentration was increased from 0.1 to 0.5 M, Ka values increased gradually ∼50 times and this is due to the increase of binding rate constants (k1) and the decrease of dissociation rate constants (k-1). Thus, with the high ionic strength condition, the hybridization ability was increased due to the fast binding and the slow dissociation process. This is simply explained by the decrease of electrostatic repulsion between anionic chains of target and probe nucleotides in high concentrations of NaCl. 1294 Analytical Chemistry, Vol. 70, No. 7, April 1, 1998

Figure 8. Effect of ionic strength on hybridization behavior between a 5′GGGAATTCGT3′-immobilized QCM and a target 3′CCCTGAAGCA5′. Conditions: 10 mM Tris-HCl, pH 7.9, 20 °C, [immobilized probe] ) 60 ng cm-2 (1 pmol on 4.9-mm2 electrode), and [target nucleotide] ) 0.19 µM (3.0 µg in 5 mL). Table 2. Effect of Ionic Strength on Kinetic Parameters of Hybridization between a Probe 5′GGGAATTCGT3′-Immobilized QCM and a Target 3′CCCTTAAGCA5′ at 20 °Ca [NaCl] (M)

k1 (103 M-1 s-1)

k-1 (10-3 s-1)

Ka (106 M-1)

hybridizationb (%)

0.1 0.2 0.3 0.4 0.5

16 24 49 45 96

39 20 5.2 3.3 4.0

0.42 1.2 9.5 14 24

51 56 65 86 100

a 10 mM Tris-HCl, pH 7.9, [immobilized probe] ) 60 ng cm-2 (1 pmol on 4.9-mm2 electrode) on a QCM, and [target nucleotide] ) 0.19 µM (3.0 µg in 5 mL). b 1:1 hybridization was observed at 0.5 M NaCl.

Figure 9. Temperature effect on hybridization between a 5′GGGAATTCGT3′-immobilized QCM and a target 3′CCCTTAAGCA5′. Conditions: 10 mM Tris-HCl, pH 7.9, 0.2 M NaCl, [immobilized probe] ) 60 ng cm-2 (1 pmol on 4.9-mm2 electrode), and [target nucleotide] ) 0.19 µM (3.0 µg in 5 mL).

Effect of Temperature. Hybridizations of 10-mer nucleotide were carried out at different temperatures: 10, 15, 20, 25, and 30 °C. Figure 9 shows typical time dependencies of hybridization at 15 and 30 °C. At 30 °C, the hybridization attained equilibrium within 1 min but the hybridization amount was very low. In contrarst, at 15 °C, the hybridization took more than 50 min but finally hybridized with a ratio of 1:1 after 1 day. Kinetic parameters are summarized in Table 3. When the temperature increased from

Table 3. Effect of Temperature on Kinetic Parameters of Hybridization between a Prove 5′GGGAATTCGT3′-immobilized QCM and a Target 3′CCCTTAAGCA5′a temp °C

k1 (103 M-1 s-1)

k-1 (10-3 s-1)

Ka (106 M-1)

hybridizationb (%)

10 15 20 25 30

36 53 24 29 68

1.4 1.7 20 34 40

26 31 1.2 0.85 1.7

100 79 56 25 10

a 10 mM Tris-HCl, pH 7.9, 0.2 M NaCl, [immobilized probe] ) 60 ng cm-2 (1 pmol on 4.9-mm2 electrode) on a QCM, and [target nucleotide] ) 0.19 µM (3.0 µg in 5 mL). b 1:1 hybridization was observed at 10 °C after 1 day.

Figure 10. Frequency changes of the QCM when temperatures of the solution rose gradually with time (1 °C min-1). (a) The QCM on which avidin and single-strand probe 5′GGGAATTCGT3′ were immobilized and (b) the QCM on which avidin and both complementary target and probe 10-mer nucleotides were immobilized. Conditions: 10 mM Tris, pH 7.9, 0.2 M NaCl, [immobilized probe] ) 60 ng cm-2 (1 pmol on a 4.9-mm2 electrode).

10 to 30 °C, the hybridization percentage decreased from 100 to 10%. Ka values decreased with increasing temperatures mainly due to the increase of dissociation rate constants (k-1); binding rate constants (k1) were not greatly temperature dependent. Tm Measurements on the QCM. To confirm whether the hybridization occurs on the QCM plate or not, we measured a melting temperature (Tm) of hybridized 10-mer nucleotides on the QCM. The QCM on which the complementary target nucleotide bound to the probe nucleotide was soaked in a new buffer solution (pH 7.9, 10 mM Tris-HCl, 0.2 M NaCl, 10 °C), and solution temperatures rose gradually with time (∼1 °C min-1). Figure 10

shows the frequency changes of the QCM with increasing temperature. In the case of the QCM on which avidin and the single-stranded nucleotide were immobilized, the frequency simply increased with temperature (curve a) mainly because of viscosity and density decreases of the water medium.16c,d When the temperature of the solution decreased gradually with time, the frequency reverted along the same line. This means that the immobilized molecules did not release from the QCM plate during temperature experiments. A similar frequency increase was observed for the naked QCM in the solution. In contrast, the frequency of the QCM on which the target nucleotide was bound to the probe nucleotide abruptly increased near 25 °C (curve b), which was consistent with Tm ) 23 ( 2 °C for the same sequence of 10-bp nucleotides obtained from hypochromisity of UV absorption in the same aqueous solution.2 When the temperature decreased quickly (10 °C min-1), the frequency did not revert to the original value. Thus, the frequency increase near 25 °C indicates the mass decrease due to the release (melting) of the target nucleotide from the QCM plate. The bound target nucleotide was probed to form double strands (hybridize) on the QCM plate. Effect of DNA Chain Length of Nucleotides. Hybridizations among 10-, 20-, and 30-mer nucleotides were also studied. Hybridization kinetics obtained from curve-fitting methods are summarized in Table 4. When the complementary chain length was increased from 10-, 20-, and 30-mer, Ka values were clearly increased with increasing chain length due to the increase of k1 values and the decrease of k-1 values. Thus, in the longer nucleotide chains, hybridization becomes strong with the higher binding rate and with the slower dissociation rate. Effect of Mismatching Bases. Figure 11 shows frequency changes due to the hybridization between the 20-mer probe 5′TGCCAAGCTTGGGAATTCGT3′ on the QCM and 20-mer target nucleotides having 0, 1, 2, and 3 mismatching bases in the chain. When the number of mismatching bases was increased, the binding amount was clearly decreased. In the case of the hybridization between the 10-mer nucleotides, hybridization was hardly observed when only one mismatching base was introduced in the middle of chain (see Figure 3). In the case of the 20-mer hybridizations, the presence of one mismatching base decreased the binding to ∼50% and two mismatching bases caused no hybridization. This means that the QCM technique will find the number of mismatching base in sequences depending on the length of complementary base pairs. Hybridization kinetic parameters were obtained from the curvefitting method (eq 6). Ka values clearly decreased with increasing number of mismatching bases and this is due to the decrease of k1 values and the increase of k-1 values. Thus, when the target

Table 4. Effect of Chain Length of Nucleotides on Hybridization Kinetic Parametersa probe and target nucleotides

chain length (mer)

k1 (103 M-1 s-1)

biotin-5′GGGAATTCGT3′ + 3′CCCTTAAGC5′ biotin-5′TGCCAAGCTTGGGAATTCGT3′ + 3′ACGGTTCGAACCCTTAAGCA5′ biotin-5′GTCCTAGGCCGCTTAAGCGCTTTCGAAGCG3′ + 3′CAGGATCCGGCGAATTCGCGAAAGCTTCGC5′

10

24

20

120

0.21

570

30

390

0.01

39000

a

k-1 (10-3 s-1)

Ka (106 M-1)

20

1.2

10 mM Tris-HCl, pH 7.9, 20 °C, [immobilized probe] ) 1 pmol on 4.9-mm2 electrode on a QCM.

Analytical Chemistry, Vol. 70, No. 7, April 1, 1998

1295

Figure 11. Effect of mismatching bases in the target sequences on hybridization on a 20-mer probe of 5′TGCCAAGCTTGGGAATTCGT3′ on a QCM. The underlines indicate a mismatching base of the target nucleotides. Conditions: 10 mM Tris, pH 7.9, 0.2 M NaCl, 20 °C, [immobilized probe] ) 120 ng cm-2 (1 pmol on a 4.9-mm2 electrode) on a QCM, and [target nucleotide] ) 0.01 µM (0.3 µg in 5 mL).

molecules had a mismatching base, binding rate was decreased and dissociation rate was increased and consequently both the binding amount and association constants were decreased with increasing number of mismatching bases. CONCLUSION The 27-MHz QCM is highly sensitive to detect hybridization between oligonucleotides without using any DNA labeling in ionic aqueous solutions. From time courses of frequency changes, kinetic parameters such as binding amount (∆m), association constants (Ka), and binding and dissociation rate constants (k1 and k-1) could be obtained. The obtained results were consistent with those obtained by surface plasmon resonance using the BIAcore system. The QCM method has the following advantages for the conventional fluorescent-labeling or radiolabeling method: (i) pre- and posttreatments are not required to modify DNA probes, (ii) the absolute binding amount and its time course can be obtained, (iii) relatively speedy measurements, (iv) the QCM plate can be reusable by removing the target nucleotide by

1296 Analytical Chemistry, Vol. 70, No. 7, April 1, 1998

heating or alkali treatment, (v) relatively inexpensive instrument and QCM plate, (vi) since the QCM plate and its oscillation circuit system is small (10 × 10 × 10 mm2), a multiple sensor system can be easily prepared on which each different probe nucleotide is immobilized on each QCM plate. We believe that the highly sensitive 27-MHz QCM technique will provide a new tool for kinetic studies of DNA hybridization and for the new genetic diagnosis sensor system. ACKNOWLEDGMENT We deeply thank Drs. Masanori Gotoh and Yukio Hasegawa (Pharmacia Biotech K. K., Tokyo) for helpful support in the use of the BIAcore system.

Received for review June 5, 1997. Accepted January 13, 1998. AC970584W