Nucleotide Identification and Orientation Discrimination of DNA

Aug 13, 2008 - E-mail: [email protected]. ... Yun Ding , Aaron M. Fleming , Henry S. White , and Cynthia J. ... Robert F. Purnell and Jacob J. Sch...
0 downloads 0 Views 683KB Size
NANO LETTERS

Nucleotide Identification and Orientation Discrimination of DNA Homopolymers Immobilized in a Protein Nanopore

2008 Vol. 8, No. 9 3029-3034

Robert F. Purnell, Kunal K. Mehta, and Jacob J. Schmidt* Department of Bioengineering, UniVersity of California, Los Angeles, Los Angeles, California 90095 Received July 30, 2008

ABSTRACT Nanopores have been used as extremely sensitive resistive pulse sensors to detect analytes at the molecular level. There has been interest in using such a scheme to rapidly and inexpensively sequence single molecules of DNA. To establish reference current levels for adenine, cytosine, and thymine nucleotides, we measured the blockage currents following immobilization of single-stranded DNA polyadenine, polycytosine, and polythymine within a protein nanopore in chemical orientations in which either the 3′ or the 5′ end enters the pore. Immobilization resulted in low-noise measurements, yielding sharply defined current distributions for each base that enabled clear discrimination of the nucleotides in both orientations. In addition, we find that not only is the blockage current for each polyhomonucleotide orientation dependent, but also the changes in orientation affect the blockage currents for each base differently. This dependence can affect the ability to resolve polyadenine and polythymine; with the 5′ end entering the pore, the separation between polyadenine and polythymine is double that observed for the 3′ orientation. This suggests that, for better resolution, DNA should be threaded through the 5′ end first in nanopore DNA sequencing experiments.

Resistive pulse sensing can detect objects passing through an electrolyte-filled channel by measurement of changes in the channel’s electrical resistance. Reduction of the channel size to the nanometer scale enables detection of small molecules.1-3 Recently, researchers have begun to investigate the application of nanopore sensing to sequence single molecules of DNA.4 To enable single-molecule DNA sequencing with nanopores, the ionic currents flowing through the nanopore during the translocation of adenine (A), cytosine (C), guanine (G), and thymine (T) nucleotides must be differentiable and identifiable. Initial work showed that the magnitude and duration of ionic current blockages occurring during the electrophoretically driven translocation of different single-stranded DNA (ssDNA) or RNA polyhomonucleotides through the pore protein R-hemolysin (RHL) could be used to differentiate them.5-7 However, the translocation time of these polynucleotide strands was extremely short: on average, for RNA, it was approximately 5 µs/base for polyC and 22 µs/base for polyA,6 and for DNA, it was 1 µs/base for polyC and 3 µs/base for polyA.7 These short translocation times necessitate high measurement bandwidths, resulting in large measurement noise. Unfortunately, reducing the translocation speed by controlling solution viscosity, temperature, and driving voltage has been shown to reduce the measured signal as well.8 * Corresponding author. E-mail: [email protected]. 10.1021/nl802312f CCC: $40.75 Published on Web 08/13/2008

 2008 American Chemical Society

Immobilizing the ssDNA in the pore enables indefinite extension of the measurement time, significantly reducing the measurement bandwidth and noise. Since the narrowest portion of the RHL pore (1.4 nm in diameter9) can accommodate only single-stranded DNA, complete translocation of ssDNA strands through the pore can be interrupted by attaching a hybridized DNA strand, which cannot fit through the pore, as a steric “stopper”.10 Ashkenasy et al. exploited this property to capture and immobilize ssDNA in RHL utilizing a terminal hairpin.11 Another study examined the effect of orientation on blockage current using immobilized polyA strands terminated with hairpins at either the 3′ or the 5′ ends. 12 We have continued in this line of inquiry, measuring the blockage currents resultant when polyA, polyC, and polyT homopolymers were immobilized in RHL with both 3′ and 5′ ends entering the pore. Our initial work utilized a terminal hairpin; however, we observed that multiple dynamic current levels were possible when the hairpin was in contact with the pore, consistent with previous work.13 To eliminate the presence of these levels, we chose a different strategy of immobilization, which was to biotinylate the end of the ssDNA and bind streptavidin to it. Streptavidin, which is unable to fit in either the channel or the vestibule of the pore, has been used to immobilize strands of ssDNA in RHL to determine the minimum voltage required to attract a ssDNA

Figure 1. (a) Representative current trace (blue curve) from our experiment. Initially upon application of 120 mV, the pore is clear and the open-pore current (io) is measured. Subsequently, one ssDNA enters the pore, and the channel current decreases to a lower value, the blockage current (ib). After several seconds, the sign of the applied voltage (red curve) is reversed, ejecting the DNA strand from the pore. (b) Histogram of the current data in (a) between t ) 0 and t ) 8 s. The means of each distribution are determined by performing a Gaussian fit.

strand into the pore,14 to probe the bond strength of hybridized DNA,15 and to analyze primer-extension with a DNA polymerase enzyme.16 We observed that ssDNA containing a streptavidin termination gave a single stable blockage current level when immobilized in the pore (Figure 1). We created freestanding lipid bilayers by applying solutions of 1% (w/v) diphytanoylphosphatidylcholine (DPhPC, Avanti Polar Lipids, Alabaster, AL) in n-decane to 100 µm orifices in 10 µm thick Teflon partitions (Eastern Scientific) integrated into lipid bilayer measurement chambers filled with 1 M KCl, 1 mM EDTA, and 10 mM Tris-HCl, as described previously.6,17 The resultant self-assembly of the lipid bilayers was monitored electrically through resistance and capacitance measurements. All voltages were applied to the “trans” side, with the other, “cis”, side held at 0 V. The solution conductivity and temperature were measured before the start of each experiment using a conductivity probe (Oakton Instruments, Williston, VT) and a thermistor. When a high quality lipid bilayer was formed (identified by a capacitance of ∼25 pF, with a resistance typically greater than 200 GΩ), 1 µL of 1.7 µg/mL RHL in 100 mM TrisHCl, 200 mM NaCl, pH 8.2 was added to the cis side of the chamber, and 120 mV was applied. Protein incorporation 3030

was observed as an abrupt increase in transmembrane current to ∼115 pA. At this point, the cis solution was exchanged for a fresh solution containing no protein. Single-stranded DNA homopolymers of polyA, polyC, and polyT (each 40 nucleotides long) with a biotin group on the 3′ or 5′ end were obtained from Integrated DNA Technologies (Coralville, IA) and Midland Reagants (Midland, TX), respectively. Because of extensive self-hybridization,18,19 synthesis of a 40-base polyG strand could not be obtained (personal communication, Midland Certified Reagents, Midland, TX) and therefore was not included in our studies. Solutions of ssDNA dissolved in TE buffer (10 mM Tris, 1 mM EDTA, pH 7.5) at a concentration of 79-198 µM were mixed with 250 µL of KCl buffer solution containing 16.7 µM streptavidin (Sigma). The volume of ssDNA solution (20-50 µL) was chosen such that all ssDNA was bound to streptavidin. The solution of streptavidin-linked ssDNA (270-300 µL) was added to the cis side of the chamber following the single channel incorporation and solution exchange and mixed by gentle stirring with the pipet tip. To maintain the same fluidic volume, an identical amount of the mixed fluid was then removed. To capture, measure, and eject individual ssDNA strands for analysis, a LabVIEW (National Instruments, Austin, TX) or pClamp (Molecular Devices, Sunnyvale, CA) program was used to apply voltages of alternating polarity throughout each experiment. In each cycle, 120 mV was applied for 8 s, followed by -100 mV for 2 s (Figure 1a) while the resultant current was amplified with a transimpedance amplifier20 or a bilayer amplifier and headstage (BC-535, Warner Instruments, Hamden, CT) and acquired at 5 kHz. For the simultaneous measurement of different polyhomonucleotides, experimental protocol proceeded as described above, but a second solution of different ssDNA was added later. In addition, if polyA (or polyT) was already in the solution being examined and polyT (or polyA) was to be added, at least two equivalents of the second nucleotide were added (270-300 µL at a time) to account for hybridization between polyA and polyT, discussed further below. In the initial phase of the voltage protocol (Figure 1a, red trace), 120 mV was applied to the trans side of the membrane, and the negatively charged ssDNA was driven into the pore. This could be observed as an instantaneous change in the transchannel current from a high value (equal to the value of the open pore current, io, measured in the absence of ssDNA in the surrounding solution) to a lower value, the blockage current, ib. Reversal of the voltage to -100 mV ejected the DNA from the pore and restored the conductance of the pore to its open value. In other experiments (data not shown) where the immobilizing voltage was applied for much longer periods of time, the blocked state persisted until the voltage was reversed, indicating that at 120 mV the strand was immobilized indefinitely. A histogram of the open and blocked currents from the trace in Figure 1a is shown in Figure 1b. The long immobilization time results in extremely precise measurements of the average values of io and ib; fit to a Gaussian distribution, the standard errors of the mean of the values in the figure were 3 fA and Nano Lett., Vol. 8, No. 9, 2008

Figure 2. (a) Plot of blockage current versus time for a typical experiment. Initially, the solution presented to the pore contained polyC only. After approximately 15 min (left vertical dashed line), a solution containing polyA is added, and a second blockage current level appears. At 150 min (right vertical dashed line), two equivalents of polyT are added, resulting in the disappearance of the polyA signal and the appearance of a third signal. The text in each region of the graph states what nucleotides are present in solution at that time. (b) Histogram of the blockage currents in (a). By performing a three-term Gaussian fit to this data, the mean blockage current values for each nucleotide were determined.

9 fA for io and ib, respectively. The mean values of io and ib are referred to henceforth as io and ib. Repetition of the measurement protocol allowed a large number of strands to be immobilized and characterized. Over the course of an individual experiment, we found the blockage currents to be highly stable. To directly compare the blockage currents of polyC, polyA, and polyT ssDNA, we measured them sequentially and simultaneously in the same pore. Strands with the 5′ and 3′ ends free to enter were examined in separate experiments. Figure 2 shows a plot of the measured ib in an experiment in which 300 µL of polyC with streptavidin on the 3′ end was initially added and measured for 10 min; then the same volume of streptavidinterminated polyA was added. Following this addition, we saw effectively a superposition of two sets of data: one which was a continuation of the currents measured for polyC and another new set of ib values which we attributed to polyA. Finally, at the 140 min mark, we added two 300 µL Nano Lett., Vol. 8, No. 9, 2008

equivalents of streptavidin-terminated polyT (as described above). We saw that the values of ib that appeared following the addition of polyA disappeared, and a new ib level appeared with a blockage current value less than that of polyC. Hybridization of polyA to polyT in the solution would result in a strand of double stranded DNA with streptavidin present on both ends since the streptavidin cap was attached at the 3′ end for both strands. These strands would be unable to enter the pore. Since polyT was in excess over polyA and the polyA signals were no longer present, we infer that all polyA strands were hybridized. The remaining unhybridized polyT strands were free to enter the pore and therefore gave rise to the new blockage current level. The polyC signal persisting throughout the experiment served as a baseline for comparison for polyA and polyT. As can be seen in Figure 2a, the current signals produced by polyA, polyC, and polyT are readily distinguishable. Figure 2b shows a histogram of the data of Figure 2a. Each of the peaks in this histogram were fit to a Gaussian distribution; the mean ( standard deviation of the blockage currents of polyA, polyC, and polyT were 21.68 ( 0.11 pA, 20.23 ( 0.12 pA, and 18.45 ( 0.15 pA, respectively. In additional experiments with different membranes and RHL proteins, we repeated the measurement protocol described above. Again, we observed similar behavior: blockage currents increasing in the order of polyT, polyC, and polyA. The means and standard deviations of Gaussian fits of the histograms of the blockage current data are shown, along with the data from Figure 2, in Table 1. The membrane resistance was measured at the start of each experiment if possible (after the membrane is formed but before the RHL pore has inserted into the membrane); the temperature and conductivity were measured throughout the experiment, and average values are given in the table. A similar procedure was repeated with homopolymers having streptavidin on the 5′ end (Figure 3a). 300 µL of polyA was added first, followed by 270 µL of polyC; again, this produced two readily distinguishable blockage current levels. Subsequent addition of 846 µL (282 µL at a time) of polyT was coincident with the disappearance of the polyA blockage current, continuation of the polyC signal, and the appearance of a new blockage current level that we assigned to polyT. Again, we see that all three blockage current levels are distinct and distinguishable. Each of the three peaks in a histogram of this data (Figure 3b) was fit to a Gaussian distribution as before, giving values of 18.29 ( 0.11 pA, 21.69 ( 0.07 pA, and 19.11 ( 0.07 pA for ib of polyA, polyC, and polyT, respectively. As with the other orientation, we repeated the experiments several times; the values of these peaks and their standard deviations, along with membrane resistance, conductivity, and temperature for each experiment, are shown in Table 2. From this data, we see that, within a small range, the blockage current values for each nucleotide are repeatable and distinct, as with the 5′ leading strands. Several interesting experimental differences between the 3′ and the 5′ leading strands were observed. First, for strands that enter 5′ end first, ib for polyA was higher than those of polyC by 1-1.5 pA, and ib values for polyC were higher 3031

Table 1. Comparison of Blockage Currents from 5′ Leading Strands of PolyA, PolyC, and PolyTa. expt

A: ib (pA)

C: ib (pA)

T: ib (pA)

io (pA)

Rmem (GΩ)

T (°C)

σ (mS·cm-1)

1 21.94 (0.12 21.08 ( 0.21 19.08 ( 0.16 127.3 ( 2.2 720 ( 540 (4) 19.70 ( 0.02 119 ( 6 2 21.68 ( 0.11 20.23 ( 0.12 18.45 ( 0.15 117.04 ( 0.69 1128 ( 142 (15) 20.40 ( 0.07 113 ( 2 3 21.61 ( 0.11 19.82 ( 0.21 18.04 ( 0.19 115.12 ( 1.1 2352 ( 615 (3) 19.40 ( 0.12 116 ( 4 4 22.58 ( 0.13 20.96 ( 0.12 18.98 ( 0.07 119.62 ( 0.83 N/A 21.20 ( 0.06 118 ( 5 a Each column shows the mean and standard deviation of blockage and open pore currents from Gaussian fits to histograms for four experiments. For each experiment, the membrane resistance was measured before the pore inserted into the membrane, if possible; the average of the resistance measurements is given, with the standard deviation and the number of measurements. The solution temperature and conductivity were also measured over the span of the experiment; average values with standard deviations are given.

observed in the above experiment, 3 equivalents of polyT were added (282 µL at a time as described earlier) because 2 equivalents removed the polyA signal but did not produce enough polyT events for statistical analysis (data not shown). The probabilities of observing a polyA, polyC, or polyT event and the resulting concentrations of each homopolymer through the experiment with 3′ and 5′ leading strands is given in Supporting Information.

Figure 3. (a) Plot of blockage current values versus time for strands inserted 3′ end first. Initially, polyA is present in solution. After approximately 40 min, polyC is added to the solution, resulting in a second current signal. At 100 min, three equivalents of polyT are added, resulting in the disappearance of the polyA signal due to hybridization (see text) and the appearance of a new signal. (b) A histogram of the blockage currents in (a) is fit to a three-term Gaussian distribution.

than those of polyT by 1.8-2 pA. In contrast, for 3′ leading strands, ib of polyC was 2-3 pA larger than polyT, and ib of polyT was ∼.5-1 pA greater than polyA. Interestingly, the polyC ib was greater than polyT for both orientations, while ib for polyA was largest in the 5′ leading orientation and smallest in the 3′ direction. Second, for the 3′ leading strands, we noticed that the number of measurements in the range assigned to polyC (see Supporting Information) was much greater than the number for polyA. For this reason, we added polyA to the chamber first in the above experiment to establish the polyA level. In a separate experiment, in which polyT was added followed by polyC, we observed a similar effect. To ensure that the polyT level would be 3032

While our experiments show that strands of polyA, polyC, and polyT can be readily distinguished by blockage current and the magnitudes of these blockage currents change with the chemical orientation of the ssDNA, we note that absolute blockage current values for polyA, polyC, and polyT could not be determined. As can be seen in Tables 1 and 2, the average blockage currents for a polynucleotide (e.g., polyA) between experiments vary by more than several standard deviations, implying that the measured currents are probably different between experiments. One possible explanation for this is variation in membrane resistance between experiments, which varied from 69 GΩ to over 2000 GΩ (Table 1). This was typical; in separate experiments, we measured the membrane resistance for 50 membranes, obtaining values ranging from 50 GΩ to over 2000 GΩ, with an average of 1.3 GΩ (data not shown). At 120 mV, these membrane resistances would result in additional measured currents of 0.06-2.4 pA; these currents could account for a significant portion of the spread of the blockage and open pore currents that we observe. Previous work has also found that variation in solution temperature21 and conductivity22 can produce variations in measured currents of magnitudes consistent with the spread in our data. However, even with these variations, the ib values for each polyhomonucleotide across all experiments were completely distinguishable and did not overlap with each other. Examination of previous work shows that direct comparisons between themselves as well as with our work are problematic because of differing experimental conditions such as temperature, ionic strength, applied voltage, and immobilization mechanism. Two previously published experiments with similar conditions to ours (Meller et al., 1 M KCl, 20 °C, freely translocating polyA and polyC7 and Butler et al., 1 M KCl 21 °C freely translocating polyA, polyC, and polyT23) show bimodal distributions for both polyA and polyC with general quantitative agreement for the most probable blockage currents: 0.115 and 0.152 for ib/io of polyA and 0.125 and 0.157 for ib/io of polyC. In ref 23, the polyT distribution appears to be sufficiently broad (ib/io peak ≈ 0.125) that two peaks cannot be distinguished. Nano Lett., Vol. 8, No. 9, 2008

Table 2. Blockage Currents from 3′ Leading Strands of PolyA, PolyC, and PolyTa expt

A: ib(pA)

C: ib(pA)

T: ib(pA)

io (pA)

Rmem (GΩ)

T (°C)

σ (mS·cm-1)

1 17.78 ( 0.29 21.18 ( 0.16 18.59 ( 0.07 113.95 ( 0.45 N/A 20.15 ( 0.05 118 ( 3 2 18.29 ( 0.11 21.69 ( 0.08 19.11 ( 0.07 117.33 ( 0.46 N/A 20.26 ( 0.09 116 ( 3 3 18.35 ( 0.09 21.49 ( 0.08 18.77 ( 0.08 113.80 ( 0.46 N/A 19.97 ( 0.12 115 ( 2 4 19.79 ( 0.07 23.13 ( 0.25 20.51 ( 0.17 118.68 ( 2.8 69.4 ( 13.0 (11) 20.60 ( 0.30 114 ( 2 a Each column shows the average and standard deviation of blockage and open pore currents from Gaussian fits to histograms for four experiments. For each experiment the membrane resistance was measured before the pore inserted into the membrane, if possible; the average of the resistance measurements is given, with the standard deviation and the number of measurements. The solution temperature and conductivity were also measured over the span of the experiment; average values with standard deviations are given.

We highlight these previous results because our experimental conditions, with the exception of strand immobilization, are the same. To directly compare our results with these, we divided our average blockage current measured for each base by the average open pore current for each experiment and took the average of these values for all experiments, obtaining 0.160, 0.189, and 0.166 for polyA, polyC, and polyT, respectively, for strands with the 3′ end entering and obtaining 0.184, 0.171, and 0.156 for polyA, polyC, and polyT, respectively, for strands with the 5′ end entering. We consistently obtain blockage current ratios larger than the previous work with similar solution conditions. It is possible that the streptavidin immobilization is responsible for the difference in these results. Even with these differences, a qualitative agreement is found between our results and this previous work. First, we note that, in later work with ssDNA immobilized in aHL using terminal hairpins, the two peaks in the polyA data were shown to be due to the different strand orientations;12 for the data cited above, the peak at 0.115 may be attributed to the 3′ entering strand, and the peak at 0.152 may be attributed to the 5′ entering strand. Our results also show that ib for polyA is greater in the 5′ direction. Second, although two peaks in the polyC blockage current distribution were shown in the previously published work (at 0.125 and 0.157), no immobilization experiments were performed to attribute them to a particular insertion direction. We find that the blockage current for the 3′ insertion direction of polyC is greater than that for the 5′ insertion direction, in contrast with polyA. Therefore, we attribute the 0.125 peak of the previously published work to the 5′ insertion direction and the 0.157 peak to the 3′ insertion direction. Third, a bimodal distribution for the polyT distribution in Butler et al. could not be resolved. We find in our data that the blockage currents for polyT in the 3′ and 5′ insertion directions are different but very close: 0.166 and 0.156, respectively. As Butler’s data was for free translocation, the noise resulting from the high measurement bandwidth may have broadened the blockage current distribution sufficiently to obscure two closely spaced peaks. Fourth, we find in our data that ib/io of polyC in the 3′ insertion direction (0.189) is very similar to that of polyA with the 5′ insertion direction (0.184). From our assignment of the 0.157 polyC ib/io value of Meller to the 3′ direction, we find that this is also very similar to the 5′ direction polyA ib/io value of 0.152. We notice a similar relationship with ib/io for 3′ polyA (0.160) and 5′ polyC (0.171). The results presented show that each of the polyhomonucleotides produces a distinct blockage current which is dependent on the orientation of the strand in the pore. Nano Lett., Vol. 8, No. 9, 2008

Figure 4. Histograms of blockage currents from separate experiments with polyA, polyC, and polyT in which the 5′ and 3′ orientations were examined by a single pore simultaneously. In each case, the data fit well with a two-term Gaussian.

Interestingly, the blockage current of each homopolymer appears to be affected differently by orientation; for example, ib for polyA is greater in the 5′ direction than that in the 3′ direction, whereas ib of polyT and polyC is less in the 5′ direction. Furthermore, the difference in ib between the 3′ and the 5′ direction is also different for each base; a cursory examination of the data in Tables 1 and 2 indicates that the difference in ib between the 3′ and the 5′ directions is greatest for polyA, followed by polyC and polyT. To directly measure these differences, we examined both orientations simultaneously for each of the polyhomonucleotides in three separate experiments. Following the same procedure as before, 280-300 µL solutions of each homopolymer conjugated to streptavidin on the 3′ or 5′ end were added to the cis side of the chamber. Histograms of the measured ib are shown in Figure 4 and are consistent with the results of Table 1 and Table 2. The average ib values and experimental parameters are listed in Table 3. As expected, each experiment yielded a two-peaked distribution: for polyA, ssDNA with 3′ and 5′ leading ends produced ib of 18.58 ( 0.08 pA and 22.04 ( 0.17 pA, respectively, for a separation of 3.46 pA. For polyC, ib for 3′ and 5′ leading strands were 21.16 ( 0.07 pA and 19.54 ( 0.09 pA, respectively, for a separation of 1.61 pA. Blockage currents 3033

Table 3. Simultaneously Measured 3′ and 5′ Leading Blockage Currents for PolyA, PolyC, and PolyTa base

3′: ib (pA)

5′: ib (pA)

io (pA)

T (°C)

σ (mS·cm-1)

A 18.58 ( 0.08 22.04 ( 0.17 116.3 ( 0.8 20.09 ( 0.06 111.6 ( 0.1 C 21.16 ( 0.07 19.54 ( 0.09 115.0 ( 1.3 20.38 ( 0.09 113.1 ( 0.1 T 18.90 ( 0.06 18.19 ( 0.06 116.2 ( 0.2 20.36 ( 0.08 111.3 ( 0.1 a The average and standard deviation of blockage and open pore currents measured from Gaussian fits to histograms for experiments containing both orientations of a single polyhomonucleotide. The average and standard deviations of open pore currents and temperature for each experiment are reported. Membrane resistances could not be measured for these experiments because protein insertion occurred before they could be obtained. The solution conductivities were measured once per experiment, and the uncertainty listed is the instrumental error.

measured for the 3′ and 5′ leading strands of polyT were 18.90 ( 0.06 pA and 18.19 ( 0.06 pA, respectively, for a separation of 0.71 pA. As can be seen from Figure 4, there are preferential insertion orientations for each nucleotide; the probabilities of polyA, polyC, and polyT events and their concentrations are given in Supporting Information. One possible qualitative explanation for these results is the angle the bases in the homopolymer make with the phosphate backbone. Previously, Mathe et al. published the results of a molecular dynamics simulation that showed purine bases (A and G) of ssDNA homopolymers in free solution exhibited a base angle of 88°. (Base angles greater than 90° represent a tilt toward the 5′ end, and angles less than 90° tilt toward the 3′ end.) When confined in a cylinder of the same diameter as the narrow constriction of RHL, polyA made an angle of ∼115°, tilted toward the 5′ end.12 This asymmetry in angle was believed to be the cause of the larger ib for polyA in the 5′ leading direction. For polyA, we find ib to be larger for 5′ leading strands, in agreement with this work; however, for polyC and polyT, ib is larger in the 3′ leading direction. Mathe et al. also performed simulations of unconfined polypyrimidines (C and T) and found the base angle to be 108° (slightly tilted toward the 5′ direction). This angle should increase as the ssDNA is confined, giving a base tilt asymmetry in the same direction as that found for polyA. Therefore, this seems to contradict the hypothesis of base tilting toward the 5′ end of the homopolymer as the cause of larger ib for strands inserted 5′ end first. For resistive pulse-based nanopore DNA sequencing to be possible, each base must produce a unique blockage current. We have shown here that this is satisfied for polyA, polyC, and polyT ssDNA in 1 M KCl at 120 mV. Moreover, we find the separation in ib values for polyA and polyT in the 5′ direction (∼2 pA) allows for better resolution of polyA and polyT in the 3′ leading orientation (