Portal de Periódicos da CAPES

Sequencing and the Single Channel

1999; Elsevier BV; Volume: 77; Issue: 6 Linguagem: Inglês

10.1016/s0006-3495(99)77122-5

1542-0086

Olaf S. Andersen,

Membrane-based Ion Separation Techniques

Single-channel electrophysiology provides for high-resolution measurements of the dynamics of single macromolecules, the channel proteins (Neher and Sakmann, 1976Neher E. Sakmann B. Single-channel currents recorded from membrane of denervated frog muscle fibers.Nature. 1976; 260: 779-802Crossref Scopus (1529) Google Scholar). Among the measurements that have proven fruitful have been ability to monitor, in almost real time, the interactions between the channel and a blocker, or modifier, of ion permeation (Neher and Steinbach, 1978Neher E. Steinbach J.H. Local anesthetics transiently block currents through single acetylcholine-receptor channels.J. Physiol. 1978; 277: 153-176PubMed Google Scholar, Prod'hom et al., 1987Prod'hom B. Pietrobon D. Hess P. Direct measurement of proton transfer rates to a group controlling the dihydropyridine-sensitive Ca2+ channel.Nature. 1987; 329: 243-246Crossref PubMed Scopus (154) Google Scholar, Hemsley and Busath, 1991Hemsley G. Busath D. Small iminium ions block gramicidin channels in lipid bilayers.Biophys. J. 1991; 59: 901-907Abstract Full Text PDF PubMed Scopus (15) Google Scholar). The emphasis usually has been on the channel, but different blockers/modifiers may have different characteristic "fingerprints" (Hemsley and Busath, 1991Hemsley G. Busath D. Small iminium ions block gramicidin channels in lipid bilayers.Biophys. J. 1991; 59: 901-907Abstract Full Text PDF PubMed Scopus (15) Google Scholar), which provides for possibilities in terms of using the characteristics of the blocking events to determine the chemical identity of the blocking species. In this issue of Biophysical Journal, Akeson et al. explore the feasibility of using single ion channels, in this case the Staphylococcus aureus α-hemolysin channels, to determine the chemical identity of polynucleotides, with the aim of establishing a method for DNA or RNA sequencing. Earlier work (Kasianowicz et al., 1996Kasianowicz, J. J., E. Brandin, D. Branton, and D. W. Deamer. 1996. Characterization of individual polynucleotide molecules using a membrane channel. Proc. Natl. Acad. Sci. USA. 93.Google Scholar) showed that single-stranded DNA or RNA molecules can cause brief, incomplete block of the current through α-hemolysin channels when the polarity of the applied potential is such as to drive the polynucleotides into the pore. Kinetic analysis provided evidence for three different kinds of incomplete blocking events. When examining the relation between polynucleotide length and block duration, the average durations of two of the three types of blocking events varied as linear functions of the polynucleotide length, suggesting they could be related to polynucleotide passage through α-hemolysin channels. If so, the average block duration should decrease as the magnitude of the applied potential increased, which was observed. To validate the suggestion, Kasianowicz et al., 1996Kasianowicz, J. J., E. Brandin, D. Branton, and D. W. Deamer. 1996. Characterization of individual polynucleotide molecules using a membrane channel. Proc. Natl. Acad. Sci. USA. 93.Google Scholar used quantitative polymerase chain reaction to demonstrate that single-stranded, but not double-stranded, DNA could pass through α-hemolysin-doped bilayers. The flux of single-stranded DNA through the bilayer, as evaluated by the polymerase chain reaction, could be related quantitatively to the frequency of long-lived blocking events. Moreover, polynucleotides that were constructed such that they would have both single- and double-stranded stretches caused long-lived current blocks, which could be "cleared" by reversing the polarity of the applied potential. Finally, the frequency of blocking events caused by RNA varied predictably when ribonuclease was added to the polynucleotide-containing solution. Do individual polynucleotides have distinct signatures that reflect their chemical composition, and are these signatures sufficiently different that one might be able to use single-channel methods for sequencing purposes? As shown by Akeson et al., 1990Akeson M. Branton D. Kasianowicz J.J. Brandin E. Deamer D.W. Microsecond time-scale discrimination among polycytidylic acid, polyadenylic acid, and polyuridylic acid as homopolymers or as segments within single RNA molecules.Biophys. J. 1999; 77: 3227-3233Abstract Full Text Full Text PDF PubMed Scopus (823) Google Scholar the answers are "yes" and "most likely." When examining blocking events caused by homopolynucleotides, one can distinguish the characteristic current reduction (∼85%) caused by the passage of polyadenylic acid (poly A) from the current reduction (90–95%) caused by the passage of polycytidylic acid (poly C). This difference in current levels allows Akeson et al. to detect the passage of tandem polynucleotides, such as a construct with 30 A and 70 C, which is very encouraging for future developments. But then matters get complicated. Polyuridylic acid (poly U), for example, causes a current reduction similar to that caused by poly A. The kinetics of polynucleotide transfer through the α-hemolysin channel are different, however, as the passage of poly U (∼1.4 μs/nucleotide at 120 mV) is about 10-fold faster than that of poly A (∼20 μs/nucleotide) for polynucleotides of comparable length. Because the distribution of blocking durations (polynucleotide passage times) is nonexponential and appears similar to a normal distribution (Kasianowicz et al., 1996Kasianowicz, J. J., E. Brandin, D. Branton, and D. W. Deamer. 1996. Characterization of individual polynucleotide molecules using a membrane channel. Proc. Natl. Acad. Sci. USA. 93.Google Scholar), one might be able to distinguish between poly A and poly U based simply on their characteristically different passage times, assuming polynucleotides of equal length. Complications also arise because the secondary structure of the polynucleotides influences both the magnitude and time distribution of the current reductions. The pyrimidine cytosine, for example, is less bulky than the purine adenosine. Even accepting the lack of a well-defined relation between lumen geometry and conductance, e.g., (Finkelstein, 1985Finkelstein A. The ubiquitous presence of channels with wide lumens and their gating by voltage.Ann. N. Y. Acad. Sci. 1985; 456: 26-32Crossref PubMed Scopus (17) Google Scholar), one therefore would expect a larger current reduction when poly A traverses the channel than when poly C does so, which is contrary to what is observed. Under "normal" conditions, however, poly C occurs primarily as single-stranded helices, with a diameter (∼1.3 nm) that is close to the luminal diameter of the α-hemolysin pore (∼1.4 nm; Song et al., 1996Song L. Hobaugh M.R. Shustak C. Cheley S. Bayley H. Gouax J.E. Structure of staphylococcal α-hemolysin, a heptameric transmembrane pore.Science. 1996; 274: 1859-1866Crossref PubMed Scopus (1937) Google Scholar). One therefore might surmise that the more organized single-stranded structure could cause a lower conductance and a longer passage time per nucleotide, ∼6 μs, than that of the more disorganized poly U. That appears to be the case, as polydeoxycytidylic acid (poly dC), which has less secondary structure, causes an only 85–90% current reduction and traverses the channel at a rate of ∼1 μs per nucleotide. The concern about secondary structure becomes even more acute in the case of poly A, which again forms single-stranded helix, with a diameter of ∼2 nm. In this case, the poly A structure would need to unfold for the polynucleotide to traverse the channel, which might account for the much longer passage times observed for poly A. In fact, poly dA traverses the channel at a rate of ∼3 μs per nucleotide, about 5-fold faster than poly A, consistent with the notion that the rate of polynucleotide passage (and maybe even the magnitude of the current reductions) may be biased by the secondary structure of the polynucleotide. However, if the polynucleotides have significant secondary structure, how does the channel catalyze the unwinding of the secondary structure at a rate sufficient to account for the observed kinetics of the blocking events? That remains an open question. The rate constant for the polynucleotide-induced block of the α-hemolysin channels has yet to be determined; however, Kasianowicz et al., 1996Kasianowicz, J. J., E. Brandin, D. Branton, and D. W. Deamer. 1996. Characterization of individual polynucleotide molecules using a membrane channel. Proc. Natl. Acad. Sci. USA. 93.Google Scholar reported that the blocking rate for equimolar polynucleotide concentrations was faster for shorter (∼200 nucleotides) than for longer (∼1000 nucleotides) polynucleotides, suggesting that polynucleotide unfolding and orientation is important. Given the results in Table 1 of Kasianowicz et al., 1996Kasianowicz, J. J., E. Brandin, D. Branton, and D. W. Deamer. 1996. Characterization of individual polynucleotide molecules using a membrane channel. Proc. Natl. Acad. Sci. USA. 93.Google Scholar the bimolecular rate constant for passage of polydeoxynucleotides with ∼150 nucleotides is ∼104 M−1·s−1, which should be comfortably below the diffusion-controlled limit and thus allow for significant "processing" of the polynucleotide as it associates with the pore entrance. The time course of any such processing should depend of the orientation of the partially structured polynucleotide in relation to the pore entrance, which might provide insights into why the distribution of passage times through the pore is much broader than predicted from "simple" polymer theory (Lubensky and Nelson, 1999Lubensky D.K. Nelson D.R. Driven polymer translocation through a narrow pore.Biophys. J. 1999; 77: 1824-1838Abstract Full Text Full Text PDF PubMed Scopus (373) Google Scholar). The results of Akeson et al. are very promising, but the question remains of how close we are to polynucleotide sequencing by single-channel methods. Two quite different limitations need to be overcome before one can construct a prototype. First, how do we overcome the limitations imposed by the secondary structure dependence? It may be possible to do the experiments at temperatures sufficiently high to disrupt the secondary structure, which will become particularly important for "real" polynucleotides that may form double helices, which would block polynucleotide transfer through the pores (Kasianowicz et al., 1996Kasianowicz, J. J., E. Brandin, D. Branton, and D. W. Deamer. 1996. Characterization of individual polynucleotide molecules using a membrane channel. Proc. Natl. Acad. Sci. USA. 93.Google Scholar). Gawrisch et al., 1992Gawrisch K. Ruston D. Zimmerberg J. Parsegian V.A. Rand R.P. Fuller N. Membrane dipole potentials, hydration forces, and the ordering of water at membrane surfaces.Biophys. J. 1992; 61: 1213-1223Abstract Full Text PDF PubMed Scopus (404) Google Scholar, in fact, could form bilayers from dipalmitoylphosphatidylcholine at ∼50°C, which might be sufficient to eliminate the secondary structure problem. Second, how do we deconvolute the current signal to give information about the underlying nucleotide sequence? Given appropriate reference data, similar to those provided by Akeson et al., 1990Akeson M. Branton D. Kasianowicz J.J. Brandin E. Deamer D.W. Microsecond time-scale discrimination among polycytidylic acid, polyadenylic acid, and polyuridylic acid as homopolymers or as segments within single RNA molecules.Biophys. J. 1999; 77: 3227-3233Abstract Full Text Full Text PDF PubMed Scopus (823) Google Scholar it should be possible to use computational methods similar to those that have been developed for extracting idealized single-channel current traces from noisy current records (Venkataramanan et al., 1998Venkataramanan L. Walsh J.L. Kuc R. Sigworth F.J. Identification of hidden Markov models for ion channel current. I: Colored background noise.IEEE Trans. Signal Proc. 1998; 46: 1901-1915Crossref Scopus (55) Google Scholar). Even then, however, one probably would need to slow down the polynucleotide transfer through the pore in order to get a satisfactory signal, which creates demands that seem to be orthogonal to the need for disrupting the secondary structure, unless one could use the slowing-down caused by unwinding the secondary structure productively. A saving grace may be that the atomic resolution structure of the α-hemolysin channel is known, which means that one may be able to introduce bulky residues that would slow down the passage through the pore, and maybe even catalyze the unraveling of any secondary structures. Eventually the bilayer technology is likely to become limiting, as it will be difficult to design a bilayer-based apparatus that would be sufficiently stable for routine sequencing purposes. Polymerizable lipids may offer a solution, but technical problems remain to resolved (Rhodes et al., 1994Rhodes D.G. Hui S.W. Xu Y.H. Byun H.S. Singh M. Bittman R. Structure of polymerizable lipid bilayers. VII: Lateral organization of diacetylenic phosphatidylcholines with short proximal acyl chains.Biochim. Biophys. Acta. 1994; 1215: 237-244Crossref PubMed Scopus (8) Google Scholar). Alternatively, one might be able to use nano-scale pores etched in irradiated micro-fabricated materials, as this allows for pore diameters close to what would be needed (Bean et al., 1970Bean C.P. Doyle M.V. Entine G. Etching of submicron pores in irradiated mica.J. Appl. Phys. 1970; 41: 1454-1459Crossref Scopus (75) Google Scholar) and has proven useful for the sizing of viruses (DeBlois and Wesley, 1977DeBlois R.W. Wesley R.K.A. Sizes and concentrations of several type C oncornaviruses and bacteriophage T2 by the resistive-pulse technique.J. Virol. 1977; 23: 227-233PubMed Google Scholar). Finally, how are other single-molecule methods for polynucleotide sequencing progressing? Currently, no fully implemented method exists, but the future is almost here. Eigen and Rigler, 1994Eigen M. Rigler R. Sorting single molecules: application to diagnostics and evolutionary biotechnology.Proc. Natl. Acad. Sci. USA. 1994; 91: 5740-5747Crossref PubMed Scopus (906) Google Scholar proposed a method for DNA sequencing based on single-molecule detection using fluorescence correlation microscopy using exonuclease digestion of DNA that was synthesized using fluorescence-labeled nucleotides, and Kinjo et al., 1998Kinjo M. Nishimura G. Koyama T. Mets Ü. Single-molecule analysis of restriction DNA fragments using fluorescence correlation spectroscopy.Anal. Biochem. 1998; 260: 166-172Crossref PubMed Scopus (41) Google Scholar showed how one can monitor the restriction enzyme cleavage of such fluorescent DNA fragments. Recently, at the 13th International Biophysics Congress in New Delhi, India, M. Gösch (Gösch et al., 1999Gösch M. Blom H. Thyberg P. Földes-Papp Z. Björk G. Holm J. Heino T. Rigler R. Detection and selection of single biomolecules with hydrodynamic flow in microstructures.J. Biosci. 1999; 24: 191Google Scholar) reported on the controlled digestion of labeled DNA, which was fixed by laser tweezers in a micro-structure with fast hydrodynamic flow, and detection of the released nucleotides by their fluorescent signature. Work needs to be done, but a practical implementation is likely to appear in the near term. These different methods are likely to be complementary, however, and the electrophysiological approach proposed by Kasianowicz et al., 1996Kasianowicz, J. J., E. Brandin, D. Branton, and D. W. Deamer. 1996. Characterization of individual polynucleotide molecules using a membrane channel. Proc. Natl. Acad. Sci. USA. 93.Google Scholar and Akeson et al. allows, in principle, sequencing of unmodified polynucleotide fragments.