Nanopore-Assisted, Sequence-Specific Detection, and Single

Jun 4, 2019 - We report here on the ability of the α-hemolysin (α-HL) nanopore to achieve label-free, selective, ... We found that with the applicat...
0 downloads 0 Views 2MB Size
Article Cite This: Anal. Chem. 2019, 91, 8630−8637

pubs.acs.org/ac

Nanopore-Assisted, Sequence-Specific Detection, and SingleMolecule Hybridization Analysis of Short, Single-Stranded DNAs Loredana Mereuta,†,§ Alina Asandei,‡,§ Irina Schiopu,‡ Yoonkyung Park,*,∥ and Tudor Luchian*,† Department of Physics and ‡Interdisciplinary Research Institute, Sciences Department, ‘Alexandru I. Cuza’ University, Iasi 700506, Romania ∥ Department of Department of Biomedical Science and Research Center for Proteinaceous Materials (RCPM), Chosun University, Gwangju 61452, Republic of Korea Downloaded via UNIV OF SOUTHERN INDIANA on July 19, 2019 at 08:13:38 (UTC). See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles.



S Supporting Information *

ABSTRACT: We report here on the ability of the α-hemolysin (α-HL) nanopore to achieve label-free, selective, and real-time detection of 15 nt long ssDNA fragments in solution, by exploiting their hybridization with freely added, polycationic peptides-functionalized PNAs. At the core of our work lies the paradigm that when PNAs and ssDNA are mixed together, the bulk concentration of free PNA decreases, depending upon the (mis)match degree between complementary strands and their relative concentrations. We demonstrate that the ssDNA sensing principle and throughput of the method are determined by the rate at which nonhybridized, polycationic peptidesfunctionalized PNA molecules arrive at the α-HL’s vestibule entrance and thread into the nanopore. We found that with the application of a 30-fold salt gradient across the nanopore, the method enhances single-molecule detection sensitivity in the nanomolar range of ssDNA concentrations. This study demonstrates that the transmembrane potential-dependent unzip of single PNA−DNA duplexes at the α-HL’s β-barrel entry permits discrimination between sequences that differ by one base pair.

W

nucleic acids,46,47 while others have demonstrated the ability of polyelectrolyte-modified capacitive EIS sensors for detection of DNA hybridization in aqueous buffers.48 Single-molecule, fluorescence resonance energy transfer (smFRET) and fluorescence correlation spectroscopy methods were also used to detect duplex formation within or between individual DNA strands.43,49 Nanopores presented an attractive approach to probing DNA strands by selective hybridization with DNA or PNA targets.50−61 Due to the fact that short nucleic acids fragments (9−12 bases) bind transiently with their complementary strands at room temperature,62 electrical force applied on the nucleic acid dimer inside the nanopore easily unzips the molecule. This imparts the nanopore approach with the ability to also quantify in real time the DNA unzipping kinetics from single-molecule kinetics analysis dwell times of the molecule inside the nanopore,50,58,63−65 which is crucial for nanoporebased single-nucleotide polymorphism detection. This approach was highly successful for the task of miRNA detection66 or selective detection of miRNAs with single-nucleotide variants.67−70 Here, we combine the use of wild-type α-HL homooligomer nanopores with polycationic peptide-functionalized PNA monomers, allowing duplex formation with complementary sequences, to detect short ssDNA fragments at the single-

ith the presentation of conductance studies through individual nanopores more than two decades ago, the prospect for a cheaper, versatile, easier-to-implement and run method to reveal properties of matter at the single-molecule level arose.1−10 The cornerstone of the approach is represented by pioneering studies by Kasianowicz and Bezrukov,11,12 revealing that the α-hemolysin (α-HL) protein nanopore of Staphylococcus aureus permits investigation of reaction dynamics. Soon after the idea of using nanopores to sequence DNA arose.13−27 Despite considerable advances, major challenges plaguing the accurate sequencing of DNA with nanopores lie in (i) the spatial resolution provided by the nanopore and (ii) translocation velocity (1−3 μs/nt)28 determined mainly by electrophoresis, hydrodynamic friction, and surface adsorption effects29−31 which, unimpeded, imposes severe constrains in detecting single-nucleotides passage across the nanopore. There are, however, many applications which do not require sequencing per se but instead need to detect a specific DNA target.32−38 For such tasks hybridization assays, including microarrays39 or qPCR,40 are usually employed to achieve sequence detection. These methods do present limitations, however, caused by nonspecific amplification or hybridization, insufficient base selectivity and sensitivity, costly infrastructure, and complex work.39−43 Alternatively, nanomaterials were employed for fluorescence assays of nucleic acids.44,45 Other approaches for detecting hybridization of short DNAs based particularly on nanoparticles were proven useful for multiplexed detection of © 2019 American Chemical Society

Received: May 2, 2019 Accepted: June 4, 2019 Published: June 4, 2019 8630

DOI: 10.1021/acs.analchem.9b02080 Anal. Chem. 2019, 91, 8630−8637

Article

Analytical Chemistry

and digitized with a NI PCI 6221, 16-bit acquisition board (National Instruments, USA) at a sampling frequency of 50 kHz. Statistical analyses on the relative blockage amplitudes on the electric current through a single α-HL protein as well as the frequency and duration of the PNA- or DNA−PNA-induced current blockades were analyzed within the statistics of exponentially distributed events, as previously described.73 To quantify the blockade probabilities of PNA−nanopore interactions in the absence or presence of distinct ssDNAs, allpoints histograms were constructed based on the ionic current A block recordings, leading to values of block probability =

molecule level. Statistical quantification of the DNA−PNA hybridization and of DNA−PNA duplex rupture at the singlemolecule level led to measurable properties including detection propensity of specific ssDNA sequences by the target PNA, hybridization energy, and base pair mismatches. Efficiencies of mismatch discrimination using size-varied captured ssDNA probes were examined.



EXPERIMENTAL SECTION Materials. The polypeptide-functionalized (PNA) and ssDNA sequences were synthesized and purified by Panagene Inc., Republic of Korea (PNA), and Sigma-Aldrich, Germany (ssDNAs), according to the sequences designed by us. The primary sequences and molecular weights of the PNA and ssDNAs used herein are shown in Table 1.

A block + Aopen

(Ablock and Aopen refer to the histogram areas assigned to the “block” and “open” states, respectively, on the corresponding histogram). Graphic representations and numerical analysis of the recorded data were done in Origin 6 (OriginLab, Northampton, MA, USA) and pClamp 6.03 (Molecular Devices, USA) software.

Table 1. Primary Sequences of the PolypeptideFunctionalized PNA and ssDNA Strandsa



RESULTS AND DISCUSSION Principle of the Experiment. Unlike previous studies,21,50,55,63 a conceptually different approach was taken here to detect ssDNAs with the α-HL nanopore. Instead of directly capturing ssDNAs, we monitored the decrease of the aqueous concentration of a complementary cationic polypeptidefunctionalized PNA sequence following selective DNA−PNA hybridization (Figure 1). With potential applicability to discoveries presented herein, in early work authors have shown that by engineering a specific analyte recognition site on selected polymers the nanopore-based single-molecule detection method has the potential of simultaneous detection and quantitation of multiple analytes.74 As the electrophoretic force needed to capture the functionalized PNA molecules inside the α-HL vestibule drives the anionic ssDNA fragments away from the nanopore,65,66 the approach seems ideally suited in applications seeking selective, individual ssDNA species detection within a mixture of other ssDNAs or anionic molecules with a high signal-to-noise ratio. As reported initially,75 to optimize such capture events leading to indirect ssDNA detection, we added all molecules on the cis side of the membrane. The slightly larger vestibule opening of the α-HL on the cis side (diameter of ∼2.6 nm) than its βbarrel on the trans side (diameter of ∼2 nm)75,76 provides a reduced entropy barrier for molecules partitioning within the α-HL. A similar strategy was reported in subsequent work.77 Electrical Detection of ssDNAs Hybridization by Polypeptide-Functionalized PNAs. In Figure 1 we illustrate that cis-side-added, cationic polypeptide-functionalized PNAs get captured by a single α-HL clamped at trans negative (Figure 1b), seen as reversible current blockades through the nanopore. The voltage-dependent kinetic behavior of PNA−αHL interactions is described in Figure S1. Trans-positive potentials, on the other hand, preclude PNA−α-HL nanopore interactions, indicated by the absence of reversible blockade events (Figure 1g) The hybridization-driven, duplex formation between cisadded ssDNAs and target PNAs leads to depletion of bulk PNA concentration. Given the net negative charge on the DNA−PNA duplex (bare charge of −5|e−|) and ssDNA fragments (bare charge of −15|e−|), at trans-negative potentials only the free, nonhybridized cationic polypeptide-functionalized PNAs will be captured electrophoretically by the nanopore (Figure 1b−e). As for a second-order reaction, the

a

Red-marked region within the nDNA1_5, nDNA3_5, nDNA3_3, and nDNA7_5 represents the mismatches introduced to the complementary domain (cDNA) relative to the PNA sequence. All sequences were aligned as to readily visualize the base pairs from the opposite strands. As previously,65 and in accord to the accepted convention, the N-terminal of the PNA is referred to as the 5′-end of the PNA. The buffer solutions and sample preparation are detailed in the Supporting Information.

Single-Molecule Electrophysiology. The electrophysiology experiments were performed in a recording chamber separated in two compartments (denoted by cis grounded and trans) by a 25 μm thick Teflon film (Goodfellow, Malvern, MA, USA) that contained a punctured aperture of about 120 μm in diameter for planar lipid membrane formation. Using the Montal--Muller technique65,71 the bilayers were formed across the aperture from the dissolved lipids in pentane, which were added on the surface of the electrolyte solution in both compartments. About 0.5−2 μL of the α-HL was added to the grounded, cis compartment, from a stock solution made in 0.5 M KCl. After insertion of a single heptameric α-HL nanopore into the membrane and depending on the particular experiment, ssDNAs and PNA were added from the stock solution to either the cis or the trans side of the membrane. Because of the slight anion selectivity of the α-HL at neutral pH (PK+/PCl− ≈ 0.7), when the electrolyte concentration gradient was applied across the membrane, a small reversal potential was observed (ΔVreversal = 8.2 mV).72 To mitigate the effect of the reversal potential upon the net transmembrane voltage sensed by the nanopore, prior to applying the transmembrane voltage, the reversal potential was nullified from the amplifier’s voltage offset knob. The ionic currents were recorded with an Axopatch 200B (Molecular Devices, USA) or EPC-8 (HEKA, Germany) amplifier, low-pass filtered at 10 kHz, 8631

DOI: 10.1021/acs.analchem.9b02080 Anal. Chem. 2019, 91, 8630−8637

Article

Analytical Chemistry

+180 mV when the cis chamber contained solely cDNA (Figure 1f) or PNA:cDNA (Figure 1h), PNA:nDNA3_5 (Figure 1i) or PNA:nDNA7_5 (Figure 1j) mixed initially at a molar ratio of 1:0.05. The blockade activity recorded under the above conditions (Figure 1f and 1h−j) demonstrates that apparent cessation of binding events to the nanopore at ΔV = −180 mV following cDNA−PNA hybridization (Figure 1c) does not constitute an experimental artifact or irreversible collapse of the nanopore. At present we lack a clear explanation for the phenomenon embodied by Figure 1c and aim for an indepth follow up of this investigation in upcoming work. The presented model of PNA−nanopore interactions following hybridization with various ssDNAs, was further tested. As shown in Figure 2, incremental, cis-side addition of

Figure 1. Nanopore-based detection of bulk PNAs depletion through hybridization with ssDNA fragments allows accurate and real-time detection of probe ssDNA. (a) At ΔV = −180 mV, anionic ssDNA fragments (cDNA) added on the cis side (0.05 μM) of the lipid membrane containing a single α-HL are driven away from the nanopore. (b) Cis-added polycationic peptide-functionalized PNAs (5 μM) are electrophoretically captured at the α-HL’s vestibule at ΔV = −180 mV, leading to reversible conductance changes of the nanopore (upwardly oriented events) (block probability PNA = 0.66). (c) When the cis-added PNAs are mixed with perfectly matched ssDNAs (cDNA) (PNA:cDNA; molar ratio of 1:0.05), a reduction in the PNA-induced blockade events ensues. Addition of ssDNAs containing three (d, nDNA3_5) and seven mismatched bases toward the PNA (e, nDNA7_5) leave the frequency of PNA-induced blockade events only slightly modified. At ΔV = +180 mV, cis-added cDNAs are reversibly captured at the α-HL’s vestibule entry (downwardly oriented spikes, f), whereas polycationic peptide-functionalized PNAs are precluded from capture (g). Further, at ΔV = +180 mV, cis-side mixing of PNA and cDNA at a molar ratio of 1:0.05 (h) results in distinct blockade events reflecting the nanopore capture of anionic PNA−cDNA duplexes (level “#”) or single-stranded cDNAs (level “*”). In contrast, PNAs mixing with the three- (i, nDNA3_5) or seven-mismatched ssDNAs (j, nDNA7_5) resulted in blockade events suggestive of sole ssDNAs interaction with the nanopore (level “*”).

Figure 2. Concentration-dependent effect of distinct ssDNAs on the blockade events reflecting PNA−nanopore interactions. Typical traces recorded at ΔV = −180 mV and corresponding all-points histograms, showing the effect of distinct relative concentrations of cDNA (I, a− c), nDNA3_5 (II, a−c), and nDNA7_5 (III, a and b) relative to PNA (5 μM, on PNA-induced ion current blockades through a single α-HL nanopore. All nucleic acid sequences were added on the cis side of the membrane. With the addition of nDNA7_5 fragments (III), a visible effect of PNA depletion was absent even at the largest molar ratio tested (PNA:nDNA7_5)/(1:0.05).

PNA−nanopore association rate depends on the PNA’s bulk concentration, which in our experiments lowers due to hybridization. We anticipated that subsequent to ssDNA− PNA hybridization in the cis chamber a less frequent appearance of free PNA-induced blockade events will ensue, as compared to that measured before hybridization. Thus, we found that addition of the complementary cDNA (0.25 μM) on the cis chamber containing PNA (5 μM) led to almost cessation in the occurrence of blockade events at ΔV = −180 mV (Figure 1c; block probability PNA:cDNA = 0.003) as compared to the case when cDNA was absent (Figure 1b; block probability PNA = 0.66). Furthermore, as shown in Figure 1d and 1e, addition of nDNA3_5 and nDNA7_5, respectively, barely affects the likelihood of PNA-induced blockade events on the nanopore as compared to the control case, and the phenomenon is correlated with the degree of base mismatch between the ssDNAs and the PNA strands (PNA alone; Figure 1b) (block probability PNA:nDNA3_5 = 0.62, block probability PNA:nDNA7_5 = 0.51). We noted that following PNA−cDNA hybridization, initially mixed at a molar ratio of 1:0.05, very few to no PNA-induced blockade events ensued at ΔV = −180 mV (Figure 1c). This is unexpected as nonhybridized PNAs would still be available for capture at the α-HL’s vestibule, as evidenced during experiments involving the nDNA3_5 and nDNA7_5 constructs (Figure 1d and 1e). To investigate this further, we assessed the analyte−nanopore interactions at positive potentials of ΔV =

various concentrations from the cDNA (I, Figure 2a−c), nDNA3_5 (II, Figure 2a−c), and nDNA7_5 (III, Figure 2a and 2b) led to a corresponding decrease in the blockade probability of events reflecting PNA−nanopore interactions at ΔV = −180 mV, and the prevalence of the effect is sensitive to the degree of DNA−PNA base mismatch. The molecular events related to PNA−nanopore interactions before ssDNA-induced hybridization (Figure 2I−III), were quantified through the block probability PNA = 0.57. Subsequent cDNA addition at distinct molar ratios relative to PNA’s concentration led to values of block probability PNA:cDNA/1:0.01 = 0.46 (Figure 2Ib) and block probability PNA:cDNA/1:0.05 = 0.004 (Figure 2Ic), and this phenomenon is dependent on the degree of PNA:ssDNA base mismatch (block probability PNA:nDNA3_5/1:0.01 = 0.38, Figure 2IIb; block probability PNA:nDNA3_5/1:0.05 = 0.18; Figure 2IIc; block probability PNA:nDNA7_5/1:0.05 = 0.47; Figure 2IIIb). Thus far, the findings showcase a relevant strength of the approach, namely, to achieve fast detection of a known ssDNA sequence from other fragments in aqueous solution, via 8632

DOI: 10.1021/acs.analchem.9b02080 Anal. Chem. 2019, 91, 8630−8637

Article

Analytical Chemistry progressive PNA depletion following DNA−PNA hybridization without the need of additional separation or purification steps. DNA−PNA hybridization was found sensitive to the degree of base complementarity and on the position of the mismatch along the duplex, which is in good agreement with previous reports.58 We illustrate this in Figure S2, showing that three bases mismatch engineered at the opposite ends of the ssDNA fragment (compare nDNA3_3 vs nDNA3_5) visibly influence the ssDNA propensity to deplete bulk PNA concentration via hybridization. Salt Gradient Enhances the Nanopore Sensitivity of the ssDNA Detection. To develop this approach into a viable ssDNA fingerprinting technology, for future real-world analytical applications the detection threshold should match the low concentration of ssDNAs in biological fluids; for instance, the concentration of free circulating DNA in plasma and serum is in the range of 1.8−35 ng mL−1.78 In the present approach, the (indirect) ssDNA sensing ability is limited by the lowest concentration of PNA, which can be optimally detected with the nanopore, and this creates a challenge. If the objective is to increase the detection limit of the probe ssDNA, from the simple bimolecular model of ssDNA−PNA hybridization it is optimal to start with a correspondingly low concentration of the PNA, namely, in a ssDNA−PNA mixture, once the ssDNA−PNA hybridization achieves equilibrium resulting in duplexes with the concentration [PNA:DNA]eq, the value of [PNA:DNA]eq relative to the initial concentration of PNA [PNA0], expressed as percentage values (

[PNA : DNA]eq [PNA 0]

Figure 3. Salt gradient across the nanopore facilitates detection of PNA depletion following hybridization with nanomolar concentrations from the complementary ssDNA. (I) With a symmetrical KCl buffer present across the nanopore and an applied transmembrane potential ΔV = −160 mV, as low as 3 μM from the cis-added polycationic peptide-functionalized PNAs, enables rare but yet visible capture events (a, b) at the nanopore’s vestibule entry. With a 3 M KCl (trans):0.1 M KCl (cis) gradient maintained across the nanopore and a similar applied ΔV as above, cis-added polycationic peptidefunctionalized PNAs can be detected at bulk concentrations as low as 0.5 μM (c, d). (II) Cis-side, incremental application from the complementary cDNA in the nanomolar range depletes the PNA’s bulk concentration via hybridization (a), and this phenomenon is seen as a progressive increase in the association time (τon) of the PNA (c− f). τon vs [cDNA] analysis (b) reveals that cis-side addition of cDNA at a concentration as low as 5 nM can be detected via changes entailed in the association time (τon) of the PNA. In b, dotted lines represent the 95% confidence intervals for the average τon values (upward triangles), and the corresponding digits reflect the events number entering the statistics.

× 100), is

augmented when PNA0 has correspondingly lower values (see Quantitative description of the DNA−PNA hybridization, within the Langmuir binding model with depletion in the Supporting Information). In other words, under conditions of low concentration of PNA0, the frequency of nonhybridized PNA-induced blockade events which is proportional to (PNA0[PNA:DNA]eq), would differ to a larger extent from the PNAinduced blockade events measured in the absence of hybridization (proportional to PNA0) and thus increase the likelihood of accurate ssDNA detection present at low concentrations. Thus far in our study we found that the optimal PNA concentration enabling ssDNA detection in symmetrical salts present in cis and trans chambers was 5 μM, resulting in a reliable sensitivity limit of the cis-added ssDNA of 0.05 μM (Figure 1). A way to optimally detect PNA-induced blockades with PNA present initially at lower concentrations (