Covalent Modification of Silicon Nitride Nanopore by Amphoteric Polylysine for Short DNA Detection.

In this work, we demonstrate a chemical modification approach, by means of covalent-bonding amphoteric poly-l-lysine (PLL) on the interior nanopore surface, which could intensively protect the pore from etching when exposed in the electrolyte under various pH conditions (from pH 4 to 12). Nanopore was generated via simple current dielectric breakdown methodology, covalent modification was performed in three steps, and the functional nanopore was fully characterized in terms of chemical structure, hydrophilicity, and surface morphology. I-V curves were recorded under a broad range of pH stimuli to evaluate the stability of the chemical bonding layer; the plotted curves demonstrated that nanopore with a covalent bonding layer has good pH tolerance and showed apparent reversibility. In addition, we have also measured the conductance of modified nanopore with varied KCl concentration (from 0.1 mM to 1 M) at different pH conditions (pHs 5, 7, 9, and 11). The results suggested that the surface charge density does not fluctuate with variation in salt concentration, which inferred that the SiN x nanopore was fully covered by PLL. Moreover, the PLL functionalized nanopore has realized the detection of single-stranded DNA homopolymer translocation under bias voltage of 500 mV, and the 20 nt homopolymers could be evidently differentiated in terms of the current amplitude and dwell time at pHs 5, 8, and 11.

Introduction

Nanopore sensing is an ultra-sensitive, flexible, and label-free approach for single-molecule measurement and DNA sequencing.[1−11] In the past decade, the single-nanometer-scale pores demonstrated great capability for the detection, identification, and characterization of a variety of analytes, such as nanoparticles,[12−17] and biomolecules like bacteria,[18] protein,[19−26] antibody,[27,28] nucleic acid,[29−32] DNA,[33−35] RNA,[36,37] and so on. In this approach, a membrane with a nanopore was sandwiched in a flow cell containing an electrolyte solution, and the analytes (usually some charged molecules or particles) were driven to translocate through the nanopore by applying an electric field across the monolayer membrane, during which a temporary resistive pulse was produced for the characterization of molecular information of analytes.

Solid-state nanopores exhibited remarkable chemical, thermal, and mechanical stability and extraordinary versatility in terms of the size, shape, and surface properties, as well as structural robustness compared with their biological counterparts.[38,39] A large number of researches have been reported on DNA sequencing with solid-state nanopores based on different substrates, such as silicon oxide,[40−42] silicon nitride (SiN),[43−46] aluminum oxide,[47,48] molybdenum sulfide,[49−52] boron nitride,[53−56] graphene,[57,58] and polymer,[59,60] whereas other work was also displayed by means of nanotube[61−66] and nanochannel.[67−71] Nanopore fabrication technologies along with a wide range of supporting membrane materials have been intensively developed, including electric beam lithography, ion beam sculpting, track-etch technique, focused-ion-beam milling, and dielectric current breakdown.[72−81] Our group has successfully generated nanopore with tunable diameter on silicon nitride and graphene based on dielectric breakdown methodology, and eventually realized the detection of short single-stranded DNA (ssDNA) homopolymers and tetracycline; a theoretical simulation of the ion transport properties in the pore channel has also been investigated in the light of Poisson and Nernst–Planck equation.[82−85]

To further exploit the application of abiotic nanopore toward tailoring single nucleotide resolution and specific target sensing, numerous modification strategies (containing covalent bonding[86−93] and physical absorption[94−98]) have been employed to help the understanding of fundamental chemical interactions at the nanoscale/single-molecule level and modulate the surface properties as well. In the present work, we displayed the design, preparation, characterization, and application of a synthetic nanopore based on SiN membrane by covalent attachment of hydrophilic, positively charged poly-l-lysine (PLL) on the wall of nanopore. The high biocompatibility and pH sensitivity of amphoteric PLL provide an insight for further investigation of the interaction between DNA and protein with the nanopore fabricated by the dielectric breakdown.[99,100] The chemical modification procedure was performed in three steps according to the reported procedure,[101] complete characterizations have been implemented for the modified nanopore. We have looked into the influence of conductance by varying the concentration and pH value of the electrolyte in situ and unveiled the correlation of conductance and surface properties of modified nanopores with salt concentration and pH value. The PLL-coated nanopore was applied ultimately for short ssDNA recognition under 500 mV as the input voltage.

Results and Discussion

Fabrication, Modification, and Characterization of SiN Nanopore

This work was performed based on the SiN membrane on which nanopore was generated via a current-stimulus dielectric breakdown. The freshly prepared nanopore underwent three steps of chemical reactions, the buffer or water rinse was applied for each step to remove the excess reactants and any physical absorption on the interior surface.

In terms of the characterization of the modified functional groups on nanopore, IR was firstly applied to check the characteristic groups of the chemical skeleton. Unfortunately, the expected absorption for −NH– and C=O stretch was not observed, this might be due to the very limited amount of functional molecules on the chip surface, which exhibits an extremely low absorption, and is thus invisible on the IR spectrum. Also, the hydrophilicity evidently increased after the modification based on the contact angle, as presented in Figure S1. Other characterizations such as energy-dispersive spectrometry (EDS) and X-ray photoelectron spectroscopy (XPS) have been applied for the element analysis of the modified and unmodified surface (Tables S1 and S2). The comparative results demonstrated that elements C and N were increased on the modified surface than on the bare one because of the introduction of abundant aliphatic groups on both the surface and inside the nanopore; elements O and Si were decreased, as these two elements were not the main components of the modified structure and thus occupied less percentage of the total elements on the modified nanochip than on the bare one. Although change in the value is different, the variation tendency is consistent between the two techniques. Optical property was examined by Raman spectroscopy, the result displayed no obvious distinction for the modified and unmodified chips. Both atomic force microscopy (AFM) and transmission electron microscopy (TEM) technologies have been employed to analyze the surface topography and morphology of modified nanochip. The little white particles with an average height of 1.4 nm in Figure B revealed the agglomeration of PLL chain on a dry nanopore surface, which differs from the relatively clearer bare chip surface in Figure A. Also, the TEM images demonstrated a characteristic size difference before (16 nm, Figure C) and after the covalent modification of the nanopore (11 nm, Figure D); a white ring observed at the periphery of nanopore indicated that the functional layer was immobilized around the pore.

Figure 1

AFM pictures for bare (A) and PLL-modified (B) SiN nanochips with a mean height of the particles of 1.4 nm. TEM pictures for bare SiN chip with nanopore 16 nm in diameter (C) and PLL-modified (D) SiN nanopore 11 nm in diameter.

Investigation of the Stability of Modified Nanopore

I–V characterization was conducted on a patch clamp amplifier for the freshly modified nanopore mounted on the fluidic setup with a buffer solution of pH 8. The pore diameter was estimated based on an empirical formula:[102]

In the formula, σ is the electrical conductivity (normally 11.93 S/m for 1 M KCl, pH 8), L is the thickness of SiN membrane (20 nm in this work), G stands for the pore conductance, and d represents the pore diameter. The diameter of a bare SiN nanopore (all of the diameters presented in the text were calculated according to the aforementioned formula) reduced from 3.5 nm to approximately 1 nm after covalently bonding PLL on the interior surface; the I–V curve was collected discontinuously for this pore during more than a month to monitor the variation in pore size with time under the same condition. As correlation between the slope of I–V curve and the nanopore size is reported, the greater the slope, the larger the pore size; so, we could intuitively deduce the size change from the I–V curve. Figure displays the I–V characterization acquired on the unmodified and PLL-modified SiN nanopores. The slope of the curves illustrated that nanopore significantly dwindled after PLL modification; it turned to approximately 1 nm and sustained for more than a month. Figure S2 demonstrates the correlation between modified nanopore size and time. The diameter maintained at around 1 nm for a month, but slightly enlarged to above 1 nm afterward, which might be on account of the damage during frequent electrolyte/water exchange for detection and storage. This work was scaled up on tens of PLL-modified nanopores and proved to be a reliable modification procedure for further application on other nanopore substrates.

Figure 2

I–V curves of unmodified and PLL-modified SiN nanopore after a month (black: bare SiN nanopore; red: PLL-modified nanopore; blue and violet: PLL-modified nanopore after 10 and 30 days).

pH Tolerance of Modified SiN Nanopore

Poly-l-lysine (PLL) is a type of amphoteric amino acid with a pKa value of 9. Its introduction on the wall of SiN nanopore could potentially realize the rectification of ion current and surface charge of the interior pore at different pH conditions. To testify the stability of chemical bonding layer, the I–V curves were recorded with a broad range of pH stimuli (from 4 to 12) with 1 M KCl. The results in Figure S3 indicate that the slope of I–V curves of modified nanopore does not change significantly at the pH from 4 to 8, but gradually became larger under very basic condition (pH from 9 to 12), presumably owing to the polarity switch of the modified nanopore from acidic to basic condition. Furthermore, the I–V curve returns to the initial status when switching the pH value to 8. It could be elucidated that the covalent layer prevents the nanopore from etching by alteration of pH value. Table S3 summarizes the details of pore size at different pHs with 1 M KCl as the electrolyte. It is worth mentioning that no current rectification was observed on the fundamental characteristics of the current response between bare and PLL-bonded nanopore, which might depend on the surface charge and nanopore geometry.

Ionic Conductance Behavior and Surface Charge Dependence of PLL-Modified Nanopore

To disclose the correlation of amphoteric PLL and surface charge property of functional pore with the salt concentration and pH condition, we have recorded the I–V curves for the modified pore with varying KCl concentration (0.1 mM, 1 mM, 10 mM, 100 mM, and 1 M) at each pH condition (pHs 5, 7, 9, and 11). Figure depicts the conductance versus KCl concentration for modified pore (pore size was calculated to be around 2.5 nm according to formula ) at four pH conditions. The conductance of the device at all of the four pH conditions matches that of the model described as fixed charge density behavior.[102−105] The experimentally extracted conductance of the modified pore does not fluctuate at low salt concentration but gradually increases with the augmentation of salt concentration at all of the four pH conditions; presumably, the surface charge is dominated by PLL at low salt concentrations.

Figure 3

Conductance of PLL-modified SiN nanopore with a pore size of around 2.5 nm, which measured at varied KCl concentrations under four pH conditions. The two solid lines are derived from the models called as a fixed charge density (red) and bulk behavior (green); the dash lines stand for the status that is close to the fixed charge density model.

Short ssDNA Homopolymers Translocation through PLL-Modified SiN Nanopore

The SiN nanopore fabricated by the dielectric breakdown approach was subsequently modified by the covalent bonding PLL and applied afterward for the detection of 20 nt short homopolymers (with chain length of around 7 nm). Figure displays the baseline and current blockades determined for 10 pM of single-stranded poly(dG)20, poly(dT)20 from a PLL-modified SiN nanopore with 500 mV as the applied voltage in the buffer solution of 1 M KCl, pH 8. The original translocation during 20 s is provided in Figure S4; the observed open-pore current is around 700 pA. The presented stochastic set of downward spikes in Figure B inferred the translocation of single-molecule DNA homopolymer through the nanopore (Figure A is the baseline ionic current without the addition of analytes), and the amplified blockade shown alongside (Figure C) indicated that the current amplitude for poly(dT)20 is higher than that for poly(dG)20 at pH 8. This was further exhibited by the histograms in Figure , which were extracted from hundreds of events of the two homopolymers at pHs 5, 8, and 11, the mean current amplitudes, and the corresponding dwell time (Figure , with the fitting method of exponential decay) for poly(dG)20 and poly(dT)20 at the three different pH conditions listed in Table . It is not difficult to discover that the current amplitude for poly(dG)20 is higher than that for poly(dT)20, which is consistent with their physical dimension, except in the case pH 8, when the current amplitude for poly(dG)20 is 208.9 ± 2.4 pA, lower than that for poly(dT)20 (254.1 ± 4.7 pA). At pH 8, which is near the isoelectric point of PLL, the modified surface is close to neutral. The lower surface charge together with the conformation of G may be responsible for the smaller current blockade; in addition, base current was also significantly influenced by the distinction of surface charges under diverse pH. Furthermore, the current amplitude decreases with increasing pH value. This could be explained by the fact that under acidic condition, the interaction of the positively charged PLL with the negatively charged DNA is stronger. The DNA probably showed a more complicated conformation rather than a straight chain, thus exhibiting a higher current amplitude during translocation.

Figure 4

Baseline (A) and events (B) of 10 pM single-stranded poly(dT)20, poly(dG)20 translocation through PLL-modified SiN nanopore with 500 mV voltage in the buffer solution of 1 M KCl, pH 8. (C) Zoomed-in event from the trace in (B).

Figure 5

Gaussian distributions of normalized histograms of current amplitudes (a1–b3) for poly(dG)20 and poly(dT)20 at pHs 5, 8, and 11.

Figure 6

Normalized histograms of dwell time (a4–b6) for poly(dG)20 and poly(dT)20 at pHs 5, 8, and 11, with fitting method of exponential decay.

Table 1

Details of Current Amplitude and Dwell Time for poly(dG)20 and poly(dT)20 at pHs 5, 8, and 11

	current amplitude (pA)		dwell time (ms)
pH variation	poly(dG)20	poly(dT)20	poly(dG)20	poly(dT)20
pH 5	421.7 ± 3.0	320.2 ± 2.6	0.54 ± 0.03	1.02 ± 0.10
pH 8	208.9 ± 2.4	254.1 ± 4.7	0.66 ± 0.03	0.94 ± 0.02
pH 11	229.2 ± 0.9	181.6 ± 1.0	0.83 ± 0.08	0.82 ± 0.10

Moreover, the corresponding dwell time has a similar tendency as the current blockade for poly(dT)20, whereas it is inverse for poly(dG)20, probably due to the conformation diversity of G-quadruplexes.[106] It has been reported that ions and pH have the largest effects on the formation of the G-quadruplex.[107] In addition, the translocation rate in our system is around 50 μs per base according to the values presented in Table , which signifies the existence of a significant interaction of DNA with the covalent bonding PLL on the wall of nanopore; here, the translocation rate has been increased more than an order of magnitude compared with our previously reported results.[82,84] We anticipate that our further work will be the exploration of this modification technique on other nanopore substrates.

Experimental Section

Materials

The nanochips (SiN membrane growing on both sides of the silica substrate and free standing on top of window) were purchased from Nanopore Solution, Portugal, with the thickness of 20 nm and the window size of 10 μm × 10 μm. 3-(Triethoxysilyl) propylamine, glutaraldehyde (25% solution in H2O), poly-l-lysine (PLL, Mw 1000–5000), methanol anhydrous were ordered from Aldrich, Shanghai. Phosphate buffer was prepared by using monopotassium phosphate and dipotassium phosphate obtained from Admas, Shanghai. Sulfuric acid, hydrogen peroxide, and ethanol absolute were obtained from Admas, Shanghai. All of the chemicals and solvents were used as received without any further treatment. HPLC-grade water was from Molecular 1850D.

Method of Covalent Modification of PLL

The experiment was carried out on the nanochip with a free-standing SiN membrane. First, the nanopore was fabricated on SiN membrane via the electric pulse breakdown approach by embedding the chip in flow cell, which was filled with electrolyte of 10 mM Tris–HCl, 1 M KCl, and 1 mM ethylenediaminetetraacetic acid (EDTA, pH 8) and by using Ag/AgCl as a reference electrode. Second, the freshly prepared nanopore was treated with piranha solution for 30 min followed by rinsing with deionized water to clean and simultaneously introduce hydroxyl group on the interior nanopore surface. The schematic diagram of the chemical modification of SiN nanopore by PLL is displayed in Scheme .

Scheme 1

Schematic Diagram of the Chemical Modification of SiN Nanopore by Polylysine

Afterwards, the chip with nanopore was soaked in 5% of (3-aminopropyl)triethoxysilane anhydrous methanol solution for 30 min, followed by washing with ethanol absolute before treating with 5% glutaraldehyde in phosphate-buffered saline (PBS, 10 mM, pH 7.4) for 1 h and then washing with PBS. The terminated aldehyde group on the wall of the nanopore was finally cross-linked with the lateral amine group on polylysine (Mw 1000–5000, 1 mg/mL) in PBS for 2 h, followed by rinsing with PBS at 50 °C to remove the physically absorbed PLL on the interior pore surface, and finally washed with deionized water for the subsequent use.

Characterizations and Measurement

The modified nanopore was characterized in terms of chemical structure (IR, Agilent Cary 630), element analysis (EDS, JSM-7800F, and XPS, Escalab 250Xi), optical properties (Raman, inVia Reflex), surface morphology (AFM, Brucker Dimension EDGE, and TEM, FEI tecnai F20), and hydrophilicity (Contact Angle, DSA100). Electric pulse breakdown was performed on a Keithley 2450 equipment, which was controlled by a Labview program to measure the conductivity of the SiN membrane through current–voltage (I–V) curves. I–V characteristics and ion current blockade measurements were recorded on a patch clamp amplifier (Axopatch 200B) with the supersensitive electronics housed inside a Faraday cage.

Stability of Modified Nanopore

The polylysine-modified nanopore was investigated in terms of the stability of the covalent coating layer under varied pH conditions. Nanochip with functional nanopore generated by electric pulse breakdown was assembled inside a PET or Teflon flow cell. Electrolyte containing 10 mM Tris–HCl, 1 M KCl, and 1 mM EDTA (pH 8) was filled in both reservoirs (cis and trans) before the insertion of Ag/AgCl electrode, which was connected to the resistive feedback amplifier. The I–V curves were recorded over 30 days without changing the pH value and electrolyte concentration and increasing the voltage from −200 mV to +200 mV, at a scanning rate of 50 mV/min. Moreover, the pH gradient (from 4 to 12) has been employed during I–V characterization. Conductance was measured with electrolyte concentration varying from 0.1 mM to 1 M. Finally, 10 pM of two single-stranded homopolymer nucleotides, poly(dG)20 and poly(dT)20, were respectively introduced into the cis reservoir for translocation detection under bias voltage of 500 mV at pHs 5, 8, and 11.

Conclusions

To sum up, we have presented the integration of positively charged amino acid PLL into SiN nanopore via covalent bonding and have performed overall characterization for functional nanopore. The stability of chemical coating layer of nanopore has been evaluated under variation of pH value by measuring the I–V curves on the patch clamp. An obvious tendency was observed that the size of PLL bonding nanopore remains unchanged under acidic conditions but enlarged gradually at pH above 9, which depends on the surface charge and polymer extension. This process was reproducible and it has been demonstrated that the nanopore size is reversible when the pH value decreases down to the acidic range. The functional nanopore was extremely stable and could remain unchanged evidently for a month. We have investigated the salt dependence of modified nanopore conductance and discovered that our device followed a fixed charge density model in a low-salt regime. The modified pore was ultimately employed for the detection of ssDNA homopolymer translocation under bias voltage of 500 mV, thereby realizing the distinction between the 20 nt homopolymers by ionic current amplitude and dwell time. Furthermore, the amphoteric PLL on the wall of nanopore showed virtually no current rectification in this system, whereas it could provide an insight into further investigation of the interaction between DNA and protein and suggest future applications in biosensor technology.