Electronic Sequencing by Recognition
One way that single base resolution might be achieved in a nanopore reader is to use electron tunneling to sense the bases. Electron tunneling is a localized effect. Vacuum tunneling rates decrease by about a factor of ten for each Angstrom that a gap between metal electrodes is increased. Thus, a system based on electron tunneling should only sense the base in the immediate vicinity of the electrode. Motivated by this fact, Zwolak and DiVentra1 have proposed that two electrodes, placed very close to a nanopore, could sense the chemical nature of a base via the magnitude of the tunnel current through each base as it passed between the electrodes. Simulations suggest that, with a carefully chosen bias, such discrimination may indeed be possible. The scheme is difficult to implement. The electrodes must be placed within an atomic diameter or so of the target sites on the DNA, the DNA must be oriented, and fluctuations suppressed. While it might be possible to constrain and align the DNA using a large electric field, other factors, such as contamination on the electrodes,2 and the random presence of water molecules and ions all pose problems. These are the very problems that have plagued molecular electronics for years. For instance, DNA has been described as an insulator,3 a semiconductor,4 a metal5 and even a superconductor.6 The most likely source of this extreme range of results lies in experimental methods, problems that now appear to be largely removed.7 The key to making reliable electronic measurements on single molecules lies in using chemistry to form predictable bonds between the metal electrodes and the molecule.8, 9
The bonds used to contact the molecules to the metal do not have to be strong and covalent. Ohshiro and Umezawa showed that tunnel current between bases could be significantly enhanced by hydrogen bonds.10
Can this scheme generate single molecule base-specific information? A simple model of the base-base interaction can be built by attaching one base to the probe of a scanning tunneling microscope (STM) and another (in the form of a thiolated nucleoside) to a conducting substrate. The movement of DNA past the "reading head" can be simulated by pulling the STM probe away from the surface and recording the decay of tunnel current. The results are quite striking11 and some of them are summarized in figure X(B - D). Figure X(B) shows data for a bare metal surface with various functional groups on the probe, X(C) data for a surface funtionalized with thymidine and X(D) data for a surface functionalized with deoxycytidine. The green and orange curves are for a bare probe and a thiophenol-fucntionalized probe. The black curves are for a guanine functionalized probe. Notably, signals obtained with the control group are insensitive to the chemical nature of the group attached to the tip. This illustrates the role played by the "weak link" (the non-bonded interaction at the bottom surface). A 100 MW resistor in series with a 100 kW resistor is still basically a 100 MW resistor. The signals obtained over a thymidine coated surface (figure X(C)) are notably different. The G-T mismatch is stabilized by two hydrogen bonds and the signals obtained with a guanine on the probe are clearly larger than the controls. The really striking difference occurs for the Watson-Crick G-C basepair (figure X(D)). The signal is considerably larger. Analysis of these signals suggests that good discrimination is possible, and that a small number of repeated reads could lead to highly accurate sequence data.11 A similar study of the entire "sandwich" (electrode-guanidinium-DNA-base-electrode) suggests that the readout will work on intact and unmodified single stranded DNA.12
Should it prove possible to implement a similar scheme in conjunction with a nanopore, this readout scheme could offer some advantages. The junction is self-assembling, driven by the hydrogen bonding. Thus manufacturing and assembly tolerances are reduced relative to the direct tunneling scheme. The readout is chemical with a "binary" detection. A junction either forms or it does not. In practice, mismatches also give signals, but these appear to be quite distinct from the matched reads. It is even possible that the transient hydrogen bonding could provide the much needed "clock" to control the translocation rate through the pore. A disadvantage of the scheme is that, in the embodiment shown here, one reader is required for each base. Thus alignment would get complicated unless a homogeneous set of DNA fragments was presented to the reader.
- Zwolak, M. and M. Di Ventra, Electronic Signature of DNA Nucleotides via Transverse Transport. Nano Lett., 2005. 5: p. 421-424.
- Smith, T., The hydrophillic nature of a clean gold surface. J. Colloid Interface Sci., 1980. 75: p. 51-53.
- Dunlap, D.D., et al., Masking generates contiguous segments of metal coated and bare DNA for STM imaging. Proc. Natl. Acad. Sci. (USA), 1993. 90: p. 7652-7655.
- Porath, D., et al., Direct measurement of electrical transport through DNA molecules. Nature, 2000. 403: p. 635-638.
- Fink, H.-W. and C. Schoenberger, Electrical conduction through DNA molecules. Nature, 1999. 398: p. 407-410.
- Kasumov, A.Y., et al., Proximity-Induced Superconductivity in DNA. Science, 2001. 291: p. 280-282.
- Lindsay, S.M. and M.A. Ratner, Molecular Transport Junctions: Clearing Mists. Advanced Materials, 2007. 19: p. 23-31.
- Cui, X.D., et al., Reproducible measurement of single-molecule conductivity. Science, 2001. 294: p. 571-574.
- Xu, B. and N.J. Tao, Measurement of Single-Molecule Resistance by Repeated Formation of Molecular Junctions. Science, 2003. 301: p. 1221-1223.
- Ohshiro, T. and Y. Umezawa, Complementary base-pair-facilitated electron tunneling for electrically pinpointing complementary nucleobases. Proc. Nat. Acad. Sci., 2006. 103: p. 10-14.
- He, J., et al., Identification of DNA base-pairing via tunnel-current decay. Nano Letters, 2007. 7: p. 3854-3858.
- He, J., et al., Hydrogen-bond electrical circuit for recognizing bases in single stranded DNA. submitted., 2008.