Anal. Chem. 1997, 69, 5200-5202
Mismatch-Sensitive Hybridization Detection by Peptide Nucleic Acids Immobilized on a Quartz Crystal Microbalance Joseph Wang,*,† Peter E. Nielsen,‡ Mian Jiang,† Xiaohua Cai,† Joao Roberto Fernandes,† Douglas H. Grant,† Mehmet Ozsoz,† Asher Beglieter,§ and Michael Mowat§
Department of Chemistry and Biochemistry, New Mexico State University, Las Cruces, New Mexico 88003, Center for Biomolecular Recognition, IMBG, Department of Biochemistry B, The Panum Institute, Blegdamsvej 3c, DK 2200, Copenhagen, Denmark, and Manitoba Institute of Cell Biology, Winnipeg, Manitoba R5E OV9, Canada
A quartz crystal microbalance DNA hybridization biosensor, based on thiol-derivatized peptide nucleic acid (PNA) probes, offers unusual in situ differentiation of single-base mismatches. A large excess of a single-base mismatch oligonucleotide has no effect on the frequency response of the target. Such remarkable distinction between perfect matches and mismatches is illustrated by the detection of a common mutation in the p53 gene. The greater specificity of the new mass-sensitive indicatorless hybridization device over those of analogous PNA-based carbon electrodes is attributed to the formation of a PNA monolayer and the use of a hydrophilic ethylene glycol linker. The improved specificity is coupled to very fast (3-5 min) hybridization in a low-ionic-strength medium. Nucleic acid hybridization forms the basis for the diagnosis of genetic and infectious diseases.1 Miniaturized biosensors, coupling the inherent specificity of DNA recognition reactions with the high sensitivity of physical transducers, thus hold great promise for sequence-specific detection.2-6 Yet, most of these devices are not capable of selectively detecting a point mutation, as desired in numerous clinical situations. This note reports on the remarkable specificity accruing from the immobilization of thiol-derivatized peptide nucleic acids (PNA) on quartz crystal microbalance (QCM) transducers. QCMs are very sensitive mass measuring devices because of changes in their resonant frequency upon a weight increase on their surface. The use of QCM transducers offers an in situ sensitive detection of hybridization events, without the need for optical or redox indicators.2,7,8 However, such use of an oligonucleotide-im†
New Mexico State University. The Panum Institute. § Manitoba Institute of Cell Biology. (1) Skogerboe, K. J. Anal. Chem. 1995, 67, 449R-454R. (2) Fawcett, N.; Evans, J.; Chien, L. ; Flowers, N. Anal. Lett. 1988, 21, 10991114. (3) Piunno, P. A.; Krull, U. J.; Hudson, R.; Damha, M.; Cohen, H. Anal. Chim. Acta 1994, 288, 205. (4) Mikkelsen, S. R. Electroanalysis 1996, 8, 15-19 (5) Xu, X. H.; Bard, A. J. J. Am. Chem. Soc. 1995, 117, 2627-2631. (6) Johnston, D. H.; Glasgow, K.; Thorp, H. H. J. Am. Chem. Soc. 1995, 117, 8933-8938. (7) Okahata, Y.; Matsunobo, Y.; Ijiro, K.; Mukae, M.; Murakami, A.; Makino, K. J. Am. Chem. Soc. 1992, 114, 8299-8300. (8) Caruso, F.; Rodda, E.; Furlong, D.; Nikura,K.; Okahata,Y. Anal. Chem. 1997, 69, 2043-2049. ‡
5200 Analytical Chemistry, Vol. 69, No. 24, December 15, 1997
mobilized QCM to measure target DNA strands suffers from significant errors (30-90%) in the presence of similar concentrations of imperfect matches.7 PNA is a DNA mimic that has a neutral peptide-like backbone with nucleobases that allows the molecule to hybridize to complementary DNA strands with high affinity and specificity.9-11 Immobilized PNA probes have been shown recently to improve the distinction between closely related sequences when using electroactive hybridization indicators.12 Even in that case, the resulting PNA-modified electrodes displayed a substantial error (of ∼20%) upon exposure to equal concentrations of the singlebase mismatch oligonucleotide. In the following sections, we demonstrate that the mass-sensitive response of the new PNA QCM biosensor is not affected by a large excess of single-base mismatch oligonucleotides and exploit this unusual mismatch discrimination for detecting a single-base alteration relevant to the p53 gene. EXPERIMENTAL SECTION A 2-mL glass cell, containing a stirred phosphate buffer solution (0.02 M Na2HPO4 adjusted to pH 7.0 with phosphoric acid), was employed. The quartz wafer was placed at the bottom of the cell by clamping it between two glass joints, each covered with a 2-mmthick rubber ring (10-mm i.d.). A 20-mm-i.d. glass tube, fused to the upper joint, formed the cell. The upper surface of the wafer was exposed to the solution. The frequency changes were monitored with a Maxtek plating monitor (Model PM-740, Maxtek Inc., Torrance, CA), interfaced with a microcomputer. Stirring was provided with a glass propeller. All experiments were carried out at room temperature. AT-cut quartz crystals, with a fundamental frequency of 5 MHz, were received from International Crystal Manufacturing Co. (Oklahoma City, OK). These crystal wafers were coated with gold electrodes on both sides (area, 41 mm2 × 2; average gold thickness of 100 nm). Prior to the probe immobilization, the gold QCM surfaces were ultrasonically cleaned by a 20-min exposure to a Piranha solution (H2SO4:H2O2 ) 3:1). (Caution: such solution should be handled with extreme care.) The cleaned QCM crystal (9) Nielsen, P. E.; Egholm, M.; Buchardt, O. Bioconjugate Chem. 1994, 5, 3-7. (10) Hyrup, B.; Nielsen, P. E. Bioorg. Med. Chem. 1996, 4, 5-23. (11) Jensen, K. K.; Orum, H.; Nielsen, P. E.; Norden, B. Biochemistry 1997, 36, 5072-5077. (12) Wang, J.; Palecek, E.; Nielsen, P. E.; Rivas, G.; Cai, X.; Shiraishi, H.; Dontha, N.; Luo, D.; Farias, P. A. J. Am. Chem. Soc. 1996, 118, 7667-7670. S0003-2700(97)00607-0 CCC: $14.00
© 1997 American Chemical Society
Chart 1
was modified at 4 °C by maintaining one of its surfaces in contact with 40 µL of a 200 µg/mL solution of PNA in sterilized water in a sealed vial for 2 h, followed by evaporation of the solvent from the open vial. The wafer was subsequently rinsed with the phosphate buffer solution and pure water. The amount of the immobilized probe was calculated to be 1360 ng, which corresponds to 310 pmol, i.e., a surface coverage of 7.4 pmol/mm2. Cyclic voltammetric experiments using ferrocyanide indicated 87% coverage of the gold surface. The 15-mer thio-PNA probe was synthesized in Nielsen’s laboratory. Its sequence is Cys-egl-GGC AGT GCC TCA CAANH2, where Cys denotes a cysteine group and egl denotes ethylene glycol (Chart 1). The 15-mer thiolated DNA oligomer, with the same base sequence, but with a mercaptohexyl group at the 5′-phosphate end, was purchased from Integrated DNA Technologies (Coraville, IA). Life Technologies Inc. (Grand Island, NY) provided the 15-mer DNA target (T ) 5′-TTG TGA GGC ACT GCC-3′) and its one-base mismatch oligomer (M ) 5′-TTG TGA GAC ACT GCC-3′). Other 15- and 27-mer noncomplementary DNA oligomers had the following sequence: 5′TGT ACG TCA CAA CTA-3′ and 5′-GTC GTC AGA CCC AAA ACC CCG AGA GGG-3′. All aqueous media used were in sterile distilled water.
Figure 1. Frequency-time response of the PNA QCM biosensor for multiple additions of 10 µg/mL of the 15-mer DNA target (designated as T) and of 50 µg/mL of the 15-mer single-base mismatch oligomer (designated as M). Sequence T, 5′-TTG TGA GGC ACT GCC-3′; sequence M, 5′-TTG TGA GAC ACT GCC-3′.
RESULTS AND DISCUSSION The remarkable mismatch discrimination associated with the use of PNA QCM devices was illustrated in connection with the detection of a single-base alteration in a short DNA segment related to a frequent point mutation in the p53 gene, a mutation found in many types of cancer cells.13 The 15-mer thiol-derivatized PNA (Chart 1) was selected to target this common single-base alteration. Figure 1 shows the response to two additions of 10 µg/mL of the DNA target (designated as T), followed by two additions of 50 µg/mL of the single-base mismatch oligonucleotide (designated as M), followed by two more additions of 10 µg/mL of the DNA target and one 50 µg/mL addition of the mismatch. As indicated from the decrease in the resonant frequency, the sensor responds rapidly to these changes in the target concentration. Steady-state responses are produced within ∼3 min. In contrast, no response is observed upon adding the 5-fold excess of the mismatch. In comparison, DNA QCM biosensors exhibited a 26-31% error when the mismatch concentration was equal to the target level and required more than an hour to reach a steadystate response.7 Similar selectivity improvements were obtained upon exposing the sensor first to excess of the mismatch or when challenging the PNA QCM biosensor with various levels of other noncomplementary oligomers (not shown; see Experimental Section for exact sequences). The initial presence of such noncomplementary oligomers had no effect on the frequency
response of the target. The data of Figure 1 indicate that the unique binding specificity of PNA probes is maintained upon immobilization on the gold-plated QCM. Such use of a PNAderivatized thiolate anchor group to form a self-assembled monolayer accounts for the improved specificity over those of electrochemical sensors based on PNA-modified carbon paste surfaces.12,14 The packed PNA layer, coupled with the hydrophilic ethylene glycol linker, eliminates nonspecific adsorption effects and results in a readily accessible probe. Figure 1 indicates also that QCM transducers can be used to follow in real time the kinetics of hybridization of PNA with DNA, in a manner analogous to the recent use of a surface plasmon technique.11 The concentration dependence was assessed from 12 successive 10 µg/mL increments in the target DNA concentration. The response increased linearly with the target concentration at first up to 40 µg/mL, with a slight curvature up to 100 µg/mL, leveling off thereafter (not shown). The latter reflects saturation of the PNA hybridization sites. A detection limit of around 1 µg/mL was estimated on the basis of the signal-to-noise characteristics of these data (S/N ) 3). Accordingly, applications to real DNA samples would require coupling with PCR amplification. While the use of PNA allows the use of short probes (e.g., the 15-mer one used here), statistical considerations indicate that such applications may also require a slightly longer (>17-mer) probe to selectively recognize unique segments in samples of the size of the human genome.15 The reproducibility of the sensor was assessed from six measurements of 10 µg/mL of the DNA target using different crystals. This experiment yielded an average frequency change of 18.9 Hz, a range of 15.8-21.0 Hz, and a relative standard deviation of 18%. Such precision reflects the reproducibility of the probe immobilization. This frequency change corresponds to 136 ng (i.e., 0.70 pmol/mm2) of the target DNA. This, along with the PNA coverage of 7.4 pmol/mm2, explains the leveled off calibration plot above the 100 µg/mL target. In summary, the results given here demonstrate that a remarkable mismatch discrimination can be achieved by coupling
(13) (a) Levine, A. Biol. Chem. 1993, 374, 227-231. (b) Hollstein, M.; Sidransky, D.; Vogeelstein, B.; Harris, C. Science 1991, 253, 49. (c) The replacement of a guanine by adenine at base 524 results in an altered p53 protein with an arginine being replaced by a histidine amino acid.
(14) Wang, J.; Rivas, G.; Cai, X.; Chicharro, M.; Parrado, C.; Dontha, N.; Begleiter, A.; Mowat, M.; Palecek, E.; Nielsen, P. E. Anal. Chim. Acta 1997, 344, 111. (15) Millan, K. M.; Mikkelsen, S. M. Anal. Chem. 1993, 65, 2317-2323.
Analytical Chemistry, Vol. 69, No. 24, December 15, 1997
5201
PNA probes with QCM transducers. Such ability of PNA/QCM biosensors to distinguish between perfect matches and mismatches is of great importance in genetic screening and therapy. The unusual specificity is coupled with the advantages of a very fast hybridization, absence of external indicators, and use of lowionic-strength solutions or short probes. Coupling the QCM operation with a direct electric field control16 may lead to an even greater mismatch discrimination. (16) Sosnowski, R. G.; Tu, E.; Butler, W. F.; O’Connel, J.; Heller, M. J. Proc. Natl. Acad. Sci. U.S.A. 1997, 94, 1119-1123.
5202
Analytical Chemistry, Vol. 69, No. 24, December 15, 1997
ACKNOWLEDGMENT J.W. acknowledges a grant from the NM Water Resources Research Institute. J.R.F., D.H.G., and M.O. acknowledge fellowships from FAPESP (Brazil), Mount Allison University (Canada), and the Research Council (Turkey), respectively. Received for review June 11, 1997. Accepted September 12, 1997.X AC9706077 X
Abstract published in Advance ACS Abstracts, October 15, 1997.
