Resistive-Pulse DNA Detection with a Conical Nanopore Sensor

Sep 8, 2006 - C. Chad Harrell,‡ Youngseon Choi, Lloyd P. Horne, Lane A. Baker,§ Zuzanna S. Siwy,| and. Charles R. Martin*. Department of Chemistry ...
0 downloads 0 Views 159KB Size
Download PDF
Langmuir 2006, 22, 10837-10843

10837

Resistive-Pulse DNA Detection with a Conical Nanopore Sensor† C. Chad Harrell,‡ Youngseon Choi, Lloyd P. Horne, Lane A. Baker,§ Zuzanna S. Siwy,| and Charles R. Martin* Department of Chemistry and Center for Research at the Bio/Nano Interface, UniVersity of Florida, GainesVille, Florida 32611-7200 ReceiVed May 3, 2006. In Final Form: July 28, 2006 In this paper, we describe resistive-pulse sensing of two large DNAs, a single-stranded phage DNA (7250 bases) and a double-stranded plasmid DNA (6600 base pairs), using a conically shaped nanopore in a track-etched polycarbonate membrane as the sensing element. The conically shaped nanopore had a small-diameter (tip) opening of 40 nm and a large-diameter (base) opening of 1.5 µm. The DNAs were detected using the resistive-pulse, sometimes called stochastic sensing, method. This entails applying a transmembrane potential difference and monitoring the resulting ion current flowing through the nanopore. The phage DNA was driven electrophoretically through the nanopore (from tip to base), and these translocation events were observed as transient blocks in the ion current. We found that the frequency of these current-block events scales linearly with the concentration of the DNA and with the magnitude of the applied transmembrane potential. Increasing the applied transmembrane potential also led to a decrease in the duration of the current-block events. We also analyzed current-block events for the double-stranded plasmid DNA. However, because this DNA is too large to enter the tip opening of the nanopore, it could not translocate the pore. As a result, much shorter duration current-block events were observed, which we postulate are associated with bumping of the double-stranded DNA against the tip opening.

Introduction There is increasing interest in the concept of using nanopores in synthetic or biological membranes as resistive-pulse sensors for both molecular and macromolecules analytes.1-32 The †

Part of the Electrochemistry special issue. * To whom correspondence should be addressed. E-mail: crmartin@ chem.ufl.edu. ‡ Current address: Vistakon, Inc. 7500 Centurion Parkway, Suite 100W-3B, Jacksonville, FL 32256. § Current address: Department of Chemistry, Indiana University, Bloomington, IN 47405. | Current address: Department of Physics, University of California, Irvine, Irvine, CA 92697. (1) Bayley, H.; Martin, C. R. Chem. ReV. 2000, 100, 2575-2594. (2) Braha, O.; Walker, B.; Cheley, S.; Kasianowicz, J. J.; Song, L.; Gouaux, J. E.; Bayley, H. Chem. Biol. 1997, 4, 497-505. (3) Henriquez, R. R.; Ito, T.; Sun, L. Crooks, R. M. Analyst 2004, 129, 478482. (4) Gu, L. Q.; Braha, O.; Conlan, S.; Cheley, S.; Bayley, H. Nature 1999, 398, 686-690. (5) Gu, L. Q.; Bayley, H. Biophys. J. 2000, 79, 1967-1975. (6) Gu, L. Q.; Cheley, S.; Bayley, H. Proc. Natl. Acad. Sci. U.S.A. 2003, 100, 1598-15503. (7) Gu, L. Q.; Cheley, S.; Bayley, H. Science 2001, 291, 636-640. (8) Gu, L. Q.; Braha, O.; Conlan, S.; Cheley, S.; Bayley, H. Nature 1999, 398, 686-690. (9) Howorka, S.; Nam, J.; Bayley, H.; Kahne, D. Angew. Chem., Int. Ed. 2004, 43, 842-846. (10) Movileanu, L.; Cheley, S.; Bayley, H. Biophys. J. 2003, 85, 897-910. (11) Ashkenasy, N.; Sa´nchez-Quesada, J.; Bayley, H.; Ghadiri, M. R. Angew. Chem., Int. Ed. 2005, 44, 1401-1404. (12) Kasianowicz, J. J.; Burden, D. L.; Han, L. C.; Cheley, S.; Bayley, H. Biophys. J. 1999, 76, 837-845. (13) Astier, Y.; Braha, O.; Bayley, H. J. Am. Chem. Soc. 2006, 128, 17051710. (14) Kasianowicz, J. J.; Henrickson, S. E.; Weetall, H. H.; Robertson, B. Anal. Chem. 2001, 73, 2268-2272. (15) Henrickson, S. E.; Misakian, M.; Robertson, B.; Kasianowicz, J. J. Phys. ReV. Lett. 2000, 85, 3057-3060. (16) Kasianowicz, J. J.; Brandin, E.; Branton, D.; Deamer, D. W. Proc. Natl. Acad. Sci. U.S.A. 1996, 93, 13770-13773. (17) Mathe´, J.; Askimentiev, A.; Nelson, D. R.; Schulten, K.; Meller, A. Proc. Natl. Acad. Sci. U.S.A. 2005, 102, 12377-12382. (18) Deamer, D. W.; Branton, D. Acc. Chem. Res. 2002, 35, 817-825. (19) Meller, A.; Branton, D. Electrophoresis 2002, 23, 2583-2591. (20) Meller, A.; Nivon, L.; Brandin, E.; Golovchenko, J.; Branton, D. Proc. Natl. Acad. Sci. U.S.A. 2000, 97, 1079-1084.

resistive-pulse method,1 which when applied to such molecular analytes is sometimes called stochastic sensing,1-13 entails mounting the membrane containing the nanopore between two electrolyte solutions, applying a transmembrane potential difference, and measuring the resulting ion current flowing through the electrolyte-filled nanopore. In simplest terms, when the analyte enters and translocates the nanopore, it transiently blocks the ion current, resulting in a downward current pulse. The frequency of such translocation-induced current-block events is proportional to the concentration of the analyte, and the identity of the analyte is encoded in the magnitude and duration of the current-block. The majority of such molecular resistive-pulse sensing data has been obtained using a biological nanopore, R-hemolysin (R-HL), embedded in a supported lipid-bilayer membrane.4,5-21 Numerous analytes including metal ions,8,12 DNA,11,13,15-21 proteins,9,14 and various small molecules4-7 have been detected with the R-HL nanopore, and these data remain the benchmark by which alternative nanopores are evaluated. However, it seems unlikely that practical sensing devices will result from a biological nanopore because of the fragility of the lipid-bilayer membrane that houses the nanopore. As a result, there is tremendous current research interest in developing synthetic analogues (i.e., a (21) Meller, A.; Nivon, L.; Branton, D. Phys. ReV. Lett. 2001, 86, 3435-3438. (22) Storm, A. J.; Chen, J. H.; Ling, X. S.; Zandbergen, H. W.; Dekker: C. Nat. Mater. 2003, 2, 537-540. (23) Chen, P.; Mitsui, T.; Farmer, D. B.; Golovchenko, J.; Gordon, R. G.; Branton, D. Nano Lett. 2004, 4, 1333-1337. (24) Saleh, O. A.; Sohn, L. L. Nano Lett. 2003, 3, 37-38. (25) Ito, T.; Sun, L.; Henriquez, R. R.; Crooks, R. M. Acc. Chem. Res. 2004, 37, 937-945. (26) Sun, L.; Crooks, R. M. J. Am. Chem. Soc. 2000, 122, 12340-12345. (27) Siwy, Z. S. AdV. Funct. Mater. 2006, 16, 735-746. (28) Siwy, Z. S.; Trofin, L.; Kohli, P.; Baker, L. A.; Trautmann, C.; Martin, C. R. J. Am. Chem. Soc. 2005, 127, 5000-5001. (29) Heins, E. A.; Siwy, Z. S.; Baker, L. A.; Martin, C. R. Nano Lett. 2005, 5, 1824-1829. (30) Heins, E. A.; Baker, L. A.; Siwy, Z. S.; Mota, M. O.; Martin, C. R. J. Phys. Chem. B 2005, 109, 18400-18407. (31) Chen, P.; Gu, J.; Kim, Y.; Wang, Q.; Branton, D. Nano Lett. 2004, 4, 2293-2298. (32) Storm, A. J.; Storm, C.; Chen, J.; Zandbergen, H.; Joanny, J.; Dekker, C. Nano Lett. 2005, 5, 1193-1197.

10.1021/la061234k CCC: $33.50 © 2006 American Chemical Society Published on Web 09/08/2006

10838 Langmuir, Vol. 22, No. 25, 2006

nonbiological nanopore embedded in a mechanically and chemically robust synthetic membrane).1,3,22-30 For example, nanopores fabricated in silicon and silicon nitride films by ion or electron beam irradiation methods have been used as resistivepulse sensors, using primarily DNA as the analyte.22,23,31,32 The track-etch method,33 a nanopore fabrication technology that has been practiced commercially for decades, can be used to make nanopores in synthetic polymer membranes. An important advance in this technology has allowed for the preparation of polymer membranes (e.g., polycarbonate, poly(ethylene terephthalate), polyimide, 5-10 µm thick) that contain only a single nanopore.34 Nanopores with the conventional cylindrical crosssection can be prepared, as well nanopores that are conically shaped.27-30,34-42,45 The conical nanopores have been used as sensors to detect small molecules,29,30 DNA,36,39 proteins,28 and nanoparticles.42 In this paper, we describe resistive-pulse sensing of two large DNAs, a single-stranded phage DNA (7250 bases) and a doublestranded plasmid DNA (6600 base pairs), using a conically shaped nanopore in a track-etched polycarbonate membrane as the sensing element. A conically shaped nanopore has two openings, the small-diameter opening at one face of the membrane and the large-diameter opening at the opposite face. We will call the small-diameter opening the “tip” of the nanopore and the largediameter opening the “base” of the nanopore. The nanopores used here had a tip opening of 40 nm and a base opening of 1.5 µm. The phage DNA was driven electrophoretically through the nanopore (from tip to base), and these translocation events were observed as transient blocks in the ion current. We found that the frequency of these current-block events scales linearly with the concentration of the DNA and with the magnitude of the applied transmembrane potential. Increasing the transmembrane potential also resulted in a decrease in the duration of the currentblock events. We also analyzed current-block events for the double-stranded plasmid DNA. However, because this plasmid DNA is too large to enter the tip opening of the nanopore, it could not translocate the pore. As a result, much shorter duration current-block events were observed, which we postulate are associated with bumping of the plasmid-DNA against the tip opening.36 Experimental Section Materials. Polycarbonate films (3 cm diameter, 12 µm thick) that had been irradiated with a single swift heavy ion to create a single damage track through the film were obtained from GSI, Darmstadt, Germany. All other chemicals were of reagent grade or better and used as received. Single-stranded M13mp18 phage DNA (7250 bases) and Col E1 plasmid DNA from E. coli strain C600 (6600 base pairs) were obtained from Sigma and used as received. (33) Fleischer, R. L.; Price, P. B.; Walker, R. M. Nuclear Tracks in Solids: Principles and Applications; University of California Press: Berkeley, CA, 1975. (34) Apel, P. Y.; Korchev, Y. E.; Siwy, Z.; Spohr, R.; Yoshida, M. Nucl. Instr. Methods Phys. Res. B 2001, 184, 337-346. (35) Harrell, C. C.; Siwy, Z. S.; Martin, C. R. Small 2006, 2, 194-198. (36) Schiedt, B.; Healy, K.; Morrison, A. P.; Neumann, R.; Siwy, Z. S. Nucl. Instr. Methods Phys. Res. B 2005, 236, 109-116. (37) Harrell, C. C.; Kohli, P.; Siwy, Z. S.; Martin, C. R. J. Am. Chem. Soc. 2004, 126, 15646-15647. (38) Siwy, Z. S.; Powell, M. R.; Kalman, E.; Astumain, R. D.; Eisenberg, R. S. Nano Lett. 2006, 6, 473-477. (39) Mara, A.; Siwy, Z. S.; Trautmann, C.; Wan, J.; Kamme, F. Nano Lett. 2004, 4, 497-501. (40) Siwy, Z. S.; Gu, Y.; Spohr, H. A.; Baur, D.; Wolf-Reber, A.; Spohr, R.; Apel, P.; Korchev, Y. E. Europhys. Lett. 2002, 60, 349-355. (41) Li, N.; Yu, S.; Harrell, C. C.; Martin, C. R. Anal. Chem. 2004, 76, 20252030. (42) Lee, S.; Zhang, Y.; White, H. S.; Harrell, C. C.; Martin, C. R. Anal. Chem. 2004, 76, 6108-6115.

Harrell et al. Preparation of the Conical Nanopores. Conically shaped nanopores were etched into the single-track polycarbonate membrane by anisotropic chemical etching of the heavy-ion irradiated polycarbonate films. As per the procedure developed by Apel et al.,34 the irradiated film was placed between the two halves of a U-tube cell, and an etch solution (9 M KOH) was added to one half-cell and a stop solution (1 M formic acid) to the other half-cell. The damage track is preferentially etched from the face in contact with the etch solution, so that the base opening is etched into this face of the membrane. Etching was continued until the etch solution broke through to the stop solution on the other side of the membrane. To detect breakthrough, each half-cell contained a Pt wire, and a transmembrane potential difference of 15 V was applied during etching using a Keithley 6487 picoammeter/voltage source (Keithley Instruments, Cleveland, OH). The electrodes were configured such that the anode was on the side of the cell containing the etch solution and the cathode was on the side containing the stop solution. The current was initially zero, and breakthrough was signaled by a sudden increase in current, typically after a period of 1.5 to 2 h. In addition to creating the conical nanopore, the etch process thinned the membrane from 12 to 10 µm. After etching, the membrane was removed from the U-tube cell and placed in water overnight. The base diameter, dbase, was measured using field-emission scanning electron microscopy (FE-SEM). The tip diameter, dtip, was measured using both FE-SEM and electrochemically as described previously.46 FE-SEM. A Hitachi S-4000 FE-SEM was used. In addition to measuring dbase and dtip, FE-SEM was used to explore the geometry of the conical nanopores. This was accomplished by plating gold into the nanopore using an electroless plating procedure described previously.47 This creates a gold nanocone that is a replica of the pore. In addition, both faces of the polycarbonate membrane become coated with thin films of gold. To image these nanocones, the Au surface film on the face of the membrane containing the tip opening was removed by mechanically polishing with a cotton swab wetted with ethanol. The polymer template was then removed using a reactive ion-etch process described previously.35 Electrochemical Measurements. The single conical nanopore membrane was mounted in a U-tube cell, and both half-cells were filled with an electrolyte solution that was 1 M KCl, 10 mM tris buffer (pH ) 8), and 1 mM EDTA. A Ag/AgCl electrode was inserted into each half-cell solution, and a constant transmembrane potential was applied. The resulting ion-current flowing through the electrolytefilled nanopore was measured using an Axopatch 200B current amplifier (Molecular Devices Corporation, Union City, CA) in the voltage-clamp mode with a low-pass Bessel filter at 2 kHz bandwidth. The signal was digitized using a Digidata 1233A analog-to-digital converter (Molecular Devices Corporation). Data were recorded and analyzed using pClamp 9.0 software (Molecular Devices Corporation). The sign convention for these experiments is as follows: At positive applied transmembrane potentials, the anode is on the side of the membrane containing the base opening of the nanopore, and the cathode is on the side of membrane facing the tip opening. The DNA was added to the electrolyte solution facing the nanopore tip.

Results and Discussion Nanopore Characterization. Figure 1 shows FE-SEM images of the tip and base openings of a typical conically shaped nanopore in a polycarbonate membrane. The tip diameter, dtip, obtained from such images was ∼40 nm, and the base diameter, dbase, was ∼1.5 µm. Knowing the base diameter allows dtip to be confirmed (43) Tinland, B.; Pluen, A.; Sturm, J.; Weill, G. Macromolecules 1997, 30, 5763-5765. (44) Muthukumar, M.; Kong, C. Y. Proc. Natl. Acad. Sci. U.S.A. 2006, 103, 5273-5278. (45) Siwy, Z. S.; Heins, E.; Harrell, C. C.; Kohli, P.; Martin, C. R. J. Am. Chem. Soc. 2004, 126, 10850-10851. (46) Harrell, C. C.; Lee, S. B.; Martin, C. R. Anal. Chem. 2003, 75, 68616867. (47) Martin, C. R. Science 1994, 266, 1961-1966.

ResistiVe-Pulse DNA Detection

Langmuir, Vol. 22, No. 25, 2006 10839

Figure 2. Typical current-time transients for the single conicalnanopore membrane sensor at an applied transmembrane potential of 900 mV. (a) Electrolyte only, no ss-DNA. (b) Electrolyte that was 10 nM in ss-DNA.

Figure 1. FE-SEM images of (a) surface of the single-nanopore membrane that was exposed to the formic acid (stop) solution during etching. The tip opening of the nanopore is seen as the dark spot near the center of the image. (b) Surface of the membrane exposed to the etch solution showing the base opening. (c) Templatesynthesized Au nanocone prepared in such a conically shaped nanopore.

via an electrochemical measurement of the ionic resistance of the electrolyte-filled nanopore. If the cone is truly conical, the ionic resistance of the pore (R) is related to dbase, dtip, the specific resistance of the electrolyte (F), and the length of the pore (l, membrane thickness) via the following equation:41

R)

4Fl πdbasedtip

(1)

The value of R was determined by simply obtaining the (linear) current-voltage curve for the electrolyte-filled nanopore. With this value of R, all of the other parameters in eq 1 are known, and dtip can be calculated. A value of 44 nm was obtained for the nanopores used here, in good agreement with the value obtained via FE-SEM. To further investigate the shape of the nanopore, an electroless plating method was used to deposit gold within the pore.35,47 This provides a nanocone that is a replica of the pore. Images of such nanocones show that the pore has a nearly ideal conical shape (Figure 1C). From the value of dbase and the membrane thickness (10 µm), an opening angle of ∼4.3° was calculated. Starting at the tip opening, this is the angle that the pore wall makes with a line drawn perpendicular to the membrane surface and down the center of the pore. Sensing the Single-Stranded DNA. An important feature of the conical nanopore sensor is that the voltage drop caused by the ion current flowing through the nanopore is focused at the nanopore tip.42 Indeed, calculations done by Lee et al. indicate that the field strength in the solution just inside the nanopore tip can be greater than 106 V per m, when the total voltage drop across the nanopore membrane is only 1 V.42 A consequence of this focusing effect is that the ion current is extremely sensitive to analyte species present in or near the nanopore tip. That is, there is an analyte “sensing zone” just inside the tip.28,29,42 This focusing effect makes conically shaped nanopores better suited for resistive-pulse sensing than cylindrical nanopores. In the absence of DNA, applying a transmembrane potential of 900 mV resulted in a steady-state ion current of 4.8 nA that was free of transient current-block events (Figure 2A). Upon addition of the single-stranded phage DNA (ss-DNA) to the solution facing the tip opening, numerous transient current-block events were observed (Figure 2B). Similar results have been obtained for resistive-pulse sensors based on both biological18

Figure 3. Current-block-event frequency vs concentration of the ss-DNA. Applied transmembrane potential ) 900 mV. Error bars here and other figures represent standard deviations obtained by averaging the number of counts in three 30-s windows of the currentblock data.

and synthetic22,23,31,32,36,39 nanopores and have been interpreted as resulting from the transient blockage of the ion current as the DNA translocates the nanopore. We have observed similar current-block events when a porphyrin molecule was driven through a similar conical nanopore sensor.29 In this case, however, the tip opening was 4 nm because of the smaller size of the porphyrin relative to the DNA sensed here. Figure 3 shows that there is a linear relationship between the frequency of these current-block events (fb) and the concentration of the ss-DNA. This is expected if the events result from electrophoretic transport of the ss-DNA through the nanopore. This is because the electrophoretic flux (J, mol s-1 cm-2) of ss-DNA through the nanopore tip is linearly related to the concentration of ss-DNA (C) via eq 248

J ) -zFDtCE/RT

(2)

where z is the charge of the ss-DNA, Dt is the diffusion coefficient associated with transport of the ss-DNA through the tip opening, and E is the electric field strength in the tip. The relationship between current-pulse frequency, fb, and concentration can be seen more clearly by multiplying both sides of eq 2 by the cross section area of the tip opening (πrtip2) and then by Avogadro’s number (A). This converts the left-hand side of eq 2 to molecules of ss-DNA translocating the tip per second, which is the currentblock frequency

molecule/s ) fb ) -zFDtCE(πrtip2)A/RT

(3)

The charge, z, on the ss-DNA is equivalent to the number of bases, since there is one phosphate per base, and we have an independent measure of rtip. Hence, according to eq 3, the slope of the line in Figure 3 should provide the product DtE. If we assume that E is equivalent to the value calculated by Lee et al.

10840 Langmuir, Vol. 22, No. 25, 2006

Harrell et al.

(104 V cm-2)42 for a similar nanopore with a similar applied transmembrane potential (1 V vs the 0.9 V used in Figure 3), a diffusion coefficient for the ss-DNA through the nanopore tip on the order of Dt ∼ 3 × 10-12 cm2 s-1 is obtained from the slope of the line in Figure 3. This is, however, an order-of-magnitude calculation because a close examination of the electric field profiles generated by Lee et al.42 shows that the field strength right at the tip opening varies significantly across the tip opening. Hence the electric field operating on the ss-DNA will be greater near the wall of the tip and less in the electrolyte solution in the middle of the tip opening. Nevertheless, this order-of-magnitude calculation provides useful information about the transport of the ss-DNA through the tip opening. For example, it is interesting to compare this nanopore-tip diffusion coefficient with the diffusion coefficient, Ds, for the ss-DNA in free solution calculated from the Stokes-Einstein equation

Ds ) kT/6πηrg

(4)

η in eq 4 is the viscosity (we assume η of pure water at 25 °C, 9 × 10-4 kg m-1 s-1), and rg is the radius of gyration of the ss-DNA “particle” in solution. The radius of gyration can be calculated via

rg ) (Lp/3)1/2

(5)

where L is the contour length and p is the persistence length of the ss-DNA. The contour length is calculated assuming an average spacing between bases of 0.43 nm for ss-DNA;43 hence, with a total of 7250 bases, L ) 3.1 µm for this DNA. The persistence length of ss-DNA in 1 M salt solution is ∼4 nm.43 Using these values in eq 5 gives rg ∼ 64 nm, and plugging this into eq 4 gives a Ds in free solution for the ss-DNA of ∼2 × 10-7 cm2 s-1. We learn that Dt associated with transport of the ss-DNA through the nanopore tip is roughly 5 orders of magnitude smaller than Ds for transport in free-solution. The reason for this can be seen from the value of the radius of gyration for the ss-DNA “particle” in solution. From rg, we see that the diameter of the ss-DNA particle in solution (128 nm) is over three-times larger than the diameter of the tip opening of the nanopore (40 nm). Hence, to translocate, the ss-DNA must adopt an extended conformation and reptate through the tip opening. (Mathematical simulations of this reptation process for ss-DNA transport through the R-HL nanopore have been reported.44) Put another way, the ss-DNA must conform itself to the size of the tip opening and be dragged (by the electrophoretic force) through the tip opening. It is not at all surprising that the diffusion coefficient, Dt, for this process would be much smaller than Ds associated with transport of the ss-DNA in free solution, where the only barrier to transport is the intervening water molecules. Equation 3 also indicates that for a fixed concentration of ss-DNA, the frequency of the current-block events should increase linearly with E. Figure 4 illustrates that this is, indeed, the case. Similar results have been obtained with other synthetic nanopore DNA sensors.31 Note, however, that eq 3 would predict that at zero volts of applied transmembrane potential the current-block frequency would likewise be zero, and this is not observed in Figure 4. This is undoubtedly because, although at zero volts there is no electrophoretic component to transport, the ss-DNA can still diffuse into the tip opening, and eq 3 does not take into account this diffusional contribution to the transport. Furthermore, this linear relationship does not extend to lower applied transmembrane potentials because no events are observed for potentials of 500 mV or less. A threshold potential below which

Figure 4. Current-block-event frequency vs applied transmembrane potential. Concentration of ss-DNA ) 20 nM.

translocation was not observed was also encountered in investigations of transport of ss-DNA through the R-HL nanopore.15 Figures 3 and 4 illustrate an important advantage of a synthetic nanopore resistive-pulse sensor relative to a biological-nanopore sensor of this type. The linear dependence of event frequency on concentration means that the detection limit for this general method must be defined in terms of how long the analyst is willing to wait to detect a current-block event due to the analyte. Eventually, at sufficiently low concentrations, the frequency becomes prohibitively low, or put another way, the average time interval between events becomes prohibitively long. The linear dependence of event frequency on field strength shows, however, that this time interval between events can be shortened by simply applying higher transmembrane potentials. This suggests a general strategy of applying higher transmembrane potentials for more dilute solutions so as to shorten the time interval the analyst must wait to see an analyte current-block event. This strategy would work well with our synthetic-nanopore membranes because we have applied up to 20 V across such membranes without membrane rupture.49 In contrast, the supported lipid-bilayer membranes that house the biological nanopore can withstand only about 200 mV before rupture.50 We did not, however, apply potentials greater than 1 V in the studies reported here because this is the maximum potential provided by the Axopatch potentiostat used. Figure 5 shows that the duration of the current-block event decreases with increasing applied transmembrane potential. This is also consistent with electrophoretic transport theory. The electrophoretic velocity (V) of an ion is related to the electric field strength (E) via eq 651

V ) |z|eE/6πηr

(6)

where r is the radius of the ion, e is the electronic charge, and η is the viscosity of the solution. Taking the inverse of both sides of eq 6 yields

1/V ) 6πηr/|z|eE

(7)

Multiplying both sides of eq 7 by the length of the detection zone, ld, makes the left-hand side the translocation time through the detection zone, τ, which is equivalent to the current-block duration as depicted in Figure 5

τ ) 6πηrld/|z|eE

(8)

(48) Bard, A. J.; Faulkner, L. R. Electrochemical Methods, 2nd ed.; John Wiley & Sons: New York, 2001; pp 29. (49) Yu, S.; Lee, S. B.; Martin, C. R. Anal. Chem. 2003, 75, 1239-1244. (50) Nakane, J. J.; Akeson, M.; Marziali, A. J. Phys.: Condens. Matter 2003, 15, R1365.

ResistiVe-Pulse DNA Detection

Figure 5. Effect of applied transmembrane potential on the duration of the current-block events. The values of the background currents were 700 mV, 3.4 nA; 800 mV, 4.0 nA; 900 mV, 4.8 nA; and 1000 mV, 5.5 nA.

Figure 6. Scatter plot of magnitude of a current block (∆i) vs the duration of that block. [ss-DNA] ) 10 nM. Applied transmembrane potential ) 900 mV.

We see from eq 8 that τ is inversely related to the electric field strength. Although this agrees with the general trend shown in Figure 5, we do not treat these data quantitatively because we do not yet have measured or calculated field strength values for most of the applied transmembrane potentials used in these studies, nor do we yet have good estimates of the corresponding ld values. Figure 6 shows a scatter plot of the magnitude of a ss-DNA current-block event (∆i) vs the duration of that event, τ. We see that the range in ∆i values is about a factor of 3 (∼400 to ∼1200 pA), whereas the range in τ is greater than a factor of 8. As noted above, to enter the nanopore tip, the ss-DNA must adopt an extended conformation. The large range in event duration (Figure 6) reflects the differences in time required for each individual ss-DNA molecule to adopt this preferred conformation. It is important to note, again, that simulations of ss-DNA transport through the R-HL nanopore clearly show this effect of conformational change on translocation time.44 The range in ∆i values is not as large as the range in τ (Figure 6) because once the preferred conformation is achieved, it is this conformation that determines the magnitude of ∆i. Sensing the Double-Stranded DNA. Figure 7A shows typical current-block events for a 10 nM solution of the double-stranded plasmid DNA (ds-pDNA), and Figure 7B shows the corresponding scatter plot. The ds-pDNA events are much shorter in duration and have much smaller ∆i values than the ss-DNA events. Analogous short duration and small ∆i events were observed by Schiedt et al.36 in their resistive-pulse investigations

Langmuir, Vol. 22, No. 25, 2006 10841

Figure 7. (a) Typical current-time transient for the single conicalnanopore membrane sensor in the presence of 10 nM ds-DNA. (b) Scatter plot of magnitude of a current block (∆i) vs the duration of that pulse for [ds-DNA] ) 10 nM. Applied transmembrane potential ) 900 mV.

of a smaller ds-pDNA (5386 base pairs) using a conical nanopore with smaller (20 nm) tip opening than used here. They interpreted these short-duration events as being due to collisions or “bumping” of the ds-pDNA against the tip opening of the nanopore without translocation through the nanopore. To ascertain if this is a reasonable interpretation for the data obtained here (Figure 7), it is important to know the size of the ds-pDNA we studied. Fortunately, light scattering data have been obtained for this ds-pDNA, which yielded a radius of gyration of 104 nm or a diameter of 208 nm.52 This is 1.6 times bigger than the ss-DNA discussed above and over four times larger than the diameter of the tip opening of our nanopore. Hence, if the ds-pDNA is to translocate, it must adopt an extended conformation, but two points are worth making in this regard. First, double stranded DNA is less flexible than single stranded DNA, as indicated by the longer persistence length of ds-DNA: 50 nm for ds-DNA vs 4 nm for ss-DNA in 1 M KCl.43 Second, the ds-pDNA is a unique form of ds-DNA in that it is a supercolied circular DNA.52 As a result of the supercoiling, it is even less flexible than conventional ds-DNA. Hence, it would be more difficult for the ds-pDNA to adopt a conformation needed to enter the 40 nm tip opening. These explanations combined with the fact that it is four times larger than the opening support the conclusion that the short duration events observed here (Figure 7) are due to bumping rather than translocation. Equation 8 can also shed light on this issue. Let us assume for the moment that both the ss-DNA and the ds-pDNA do translocate. We can then define the ratio τp/τs, which is average translocation time for the ds-pDNA divided by the average translocation time for the ss-pDNA. Equation 5 shows that this ratio is given by

τp/τs ) rg,pzs/rg,szp

(9)

where rg is the radius of gyration and z is the charge, and the subscripts p and s denote the ds-pDNA and the ss-DNA, respectively. The radii values were given above, and the z values are simply determined by the number of phosphate. Since the (51) Bard, A. J.; Faulkner, L. R. Electrochemical Methods, 2nd ed.; John Wiley & Sons: New York, 2001; p 66-67. (52) Voordouw, G.; Kam, Z.; Borochov, N.; Eisenberg, H. Biophys. Chem. 1978, 8, 171-189.

10842 Langmuir, Vol. 22, No. 25, 2006

double stranded DNA contains two phosphates per base pair, zp ) 2 × 6600 ) 13 200. Putting these radii and charge values into eq 9 yields τp/τs ) 0.9. We learn that the predicted translocation durations for these two DNAs are about the same because the ds-pDNA is bigger (slows down translocation) but has higher charge (speeds up translocation) than the ss-DNA. In contrast, the experimental data show that τp , τs, which strongly argues that both of these species cannot be translocating. However, this analysis based on eq 9 does not take into account the dramatically hindered diffusion of these DNAs as they reptate through the tip opening, as evidenced by the ∼10-12 cm2 s-1 Dt value calculated above for the ss-DNA. Because the ds-pDNA is larger, its effective diffusion coefficient through the tip opening would be even smaller than that of the ss-DNA. Hence, if it did translocate, the translocation time (current-block duration, τ) for the ds-pDNA would be larger than that for the ss-DNA, and larger than that predicted by eq 9, yet experimentally the opposite is observed. The other factor to consider is the magnitude of the currentblock event, ∆i. Because the ds-pDNA is larger and, if translocating, moving more slowly than the ss-DNA, one would expect that the ds-pDNA would produce a larger ∆i than the ss-DNA. Again, the opposite is observed experimentally. For all of these reasons, we conclude, in agreement with Schiedt et al.,36 that the short-duration events observed for the ds-pDNA are due to bumping of the ds-pDNA against the tip opening and not due to translocation. There is additional precedent for this idea that a nanoparticle can be driven by electrophoresis to the tip of a nanopore but be too big to translocate. Lee et al. observed this result with 150 nm nanometer charged polystyrene nanoparticles and a nanopore with tip opening of 60 nm.42 The experiment in this case was somewhat different from the experiment used here; for example, electrophoresis was driven by passing a constant current through the pore, as opposed to the constant applied transmembrane potential method used here. Furthermore, Lee et al. found that the particle became pinned at the tip opening for a long time as the electrophoretic current was applied. That is, the nanoparticle did not bump against the pore opening and “bounce” off as we and Schiedt et al. are proposing. This difference in the behavior of the Lee et al. polystyrene nanoparticle vs our ds-pDNA nanoparticle is easy to rationalize. First, the polystyrene nanoparticle had on average 42 000 negative charges on its surface, whereas our ds-pDNA nanoparticle has 13 200 negative charges. Hence, the electrophoretic force acting on the polystyrene nanoparticle is greater. Second, the electrophoretic force depends on the field strength at the tip opening, and because we and Lee et al. use a different experiment (constant current vs constant potential), it would be highly unlikely that the field strength is the same. Sensing the ss-DNA and ds-DNA in Solution Together. The very different magnitudes and durations of the currentblock events for the ds-DNA vs the ss-DNA would suggest that it would be easy to distinguish between these two analytes if they were in solution together. Figure 8A shows a current-time trace for a solution that was 10 nM in both DNAs. We see both the longer duration, and greater ∆i, events associated with the ss-DNA and the shorter duration, and smaller ∆i, events of the ds-DNA. This can be seen more clearly in the scatter plot in Figure 8B.

Harrell et al.

Figure 8. (a) Current-block events for a mixture of ds-DNA and ss-DNA (10 nM each). (b) Scatter plot of magnitude of a current block (∆i) vs the duration of that block for [ss-DNA] ) [ds-DNA] ) 10 nM. Applied transmembrane potential ) 900 mV.

Conclusion

conical nanopore sensor and that these translocation events result in corresponding transient current-block events. The frequency of these current-block events scales linearly with the concentration of the ss-DNA and with the magnitude of the applied transmembrane potential. Furthermore, in agreement with electrophoretic transport theory, the duration of these current-block events scales with the inverse of the applied transmembrane potential. We also analyzed current-block events for a large plasmid ds-DNA. Because of its large size and its restricted flexibility, the ds-pDNA can neither enter the tip opening nor translocate the nanopore. As a result, much shorter duration, and smaller ∆i, current-block events were observed, which we postulate are associated with bumping of the ds-DNA against the tip opening. Because of the dramatic differences in event duration and ∆i, the ss-DNA events can be distinguished from the ds-DNA events when the two DNAs are in solution together. The synthetic nanopores described here offer significant advantages for resistive-pulse sensing relative to a sensor based on a biological nanopore housed in a lipid-bilayer membrane. These advantages include the chemical and mechanical stability of the synthetic polymer membrane, the ability to tune the size of the tip opening to match the size of the analyte,28-30,35,37,42,45 and the ability to apply much larger transmembrane potentials than is possible with the bilayer-membrane-based sensor. However, it is clear that greater selectivity will be required than the simple size-based selectivity demonstrated here. We have shown that highly selective sensors can be obtained by attaching to the nanopore a molecular-recognition agent that selectively binds the target analyte.28 We are currently expanding on this paradigm that has proved so successful with the R-HL based resistive-pulse sensors.1 Finally, resistive-pulse sensors are fundamentally electroanalytical devices. They were not, however, pioneered by an electrochemist: neither the earlier Coulter counter53 nor the more recent “stochastic sensors”.1 This is perhaps because they are unusual electroanalytical devices in that they do not emphasize the Faradaic current, as do amperometric sensors, nor do they entail measurement of a Nernst potential, as per potentiometric

In this paper, we have demonstrated that single-stranded phage DNA molecules can be driven electrophoretically through a

(53) Lines, R. W. In Particle Size Analysis; Stanley-Wood, N. G., Lines, R. W., Eds.; Royal Society of Chemistry: London, 1992.

ResistiVe-Pulse DNA Detection

sensors. There is no doubt, however, that electrochemists can make important contributions to the future development of this technology, particularly as it applies to the sensing of molecular and macromolecular analytes. To paraphrase Prof. Feynman:54 (54) http://nanoparticles.org/pdf/Feynman.pdf.

Langmuir, Vol. 22, No. 25, 2006 10843

There’s “plenty of room” for electroanalytical chemists in these nanopore sensing devices. Acknowledgment. This work was supported by the Air Force Office of Scientific Research and the National Institutes of Health. LA061234K