Nanopore Detection Using Supercharged Polypeptide Molecular Carriers

Abstract

The analysis at the single-molecule level of proteins and their interactions can provide critical information for understanding biological processes and diseases, particularly for proteins present in biological samples with low copy numbers. Nanopore sensing is an analytical technique that allows label-free detection of single proteins in solution and is ideally suited to applications, such as studying protein-protein interactions, biomarker screening, drug discovery, and even protein sequencing. However, given the current spatiotemporal limitations in protein nanopore sensing, challenges remain in controlling protein translocation through a nanopore and relating protein structures and functions with nanopore readouts. Here, we demonstrate that supercharged unstructured polypeptides (SUPs) can be genetically fused with proteins of interest and used as molecular carriers to facilitate nanopore detection of proteins. We show that cationic SUPs can substantially slow down the translocation of target proteins due to their electrostatic interactions with the nanopore surface. This approach enables the differentiation of individual proteins with different sizes and shapes via characteristic subpeaks in the nanopore current, thus facilitating a viable route to use polypeptide molecular carriers to control molecular transport and as a potential system to study protein-protein interactions at the single-molecule level.

Figure 1

Single-molecule nanopore sensing of proteins carried by supercharged unstructured polypeptides (SUPs). (a) Schematic of the experimental setup of single-protein detection using nanopipettes. An SEM image of a typical 13 nm nanopore formed at the tip of a quartz nanopipette (scale bar: 20 nm). An Ag/AgCl working electrode was inserted into the nanopipette (trans chamber), and the other Ag/AgCl reference electrode was fixed in the external bath, where proteins were placed (cis chamber). A positive voltage to the trans chamber was applied to capture anionic proteins from the cis chamber, whereas a reversed voltage was applied for cationic proteins, as shown in the schematic. The SUP is an elastin-like polypeptide with a repetitive sequence of VPGXG, where X is a variable amino acid for modular charge or hydrophobicity. X can be glutamic acid (E) or lysine (K) for the design of anionic or cationic SUPs. (b) Structural illustration of enhanced green fluorescent protein (eGFP, 28.6 kDa, pI of 5.58), anionic eGFP-E36 (46.6 kDa, pI of 4.57), and cationic eGFP-K36 (46.6 kDa, pI of 9.83). Representative ionic current traces and typical translocation events (filtered at 5 kHz for visualization) in 1 M KCl, 10 mM Tris-EDTA, pH 8.0 buffer for eGFP (5 nM, +500 mV), eGFP-E36 (5 nM, +500 mV), and eGFP-K36 (5 nM, −500 mV) were shown, respectively. Data were recorded at 1 MHz and further processed with a 10 kHz low-pass filter for statistics.

Figure 2

Charge- and length-dependence of the nanopore translocation for eGFP-SUPs. (a) Normalized capture rate (event frequency/concentration) for eGFP, eGFP-E36, and eGFP-K36 at the same voltage magnitude (500 mV). T scores were used to test the statistical significance between eGFP and eGFP-SUPs; ***P < 0.001 and ****P < 0.0001. (b) Scatter plots of peak amplitude versus dwell time for eGFP (n = 845), eGFP-E36 (n = 1074), and eGFP-K36 (n = 871) at 500 mV. Box and whisker plots of (c) dwell time and (d) peak amplitude for eGFP, eGFP-E36, and eGFP-K36, along with associated statistics. (e) Schematic of the cationic eGFP-SUPs translocated through a nanopore. Lysine is the key compound in the repeated sequence (VPGKG)_n, where n represents the net charge of the SUP. Application of a negative bias to the trans chamber captures cationic eGFP-SUPs by both electrophoretic and electro-osmotic flows. A series of eGFP-SUPs with different charges and chain lengths (K18, K36, and K72) were shown. Ionic current was recorded in 1 M KCl, 10 mM Tris-EDTA pH 8.0 buffer. Data were sampled at 1 MHz and low-pass filtered at 10 kHz. (f) Concentration dependence of capture rate for eGFP, eGFP-K18, eGFP-K36, and eGFP-72 at −500 mV. Voltage dependence of (g) normalized capture rate, (h) dwell time, and (i) peak amplitude for eGFP, eGFP-K18, eGFP-K36, and eGFP-72. The error bars represent one standard deviation of at least three independent experimental repeats.

Figure 3

Size-dependent single-protein identification using cationic SUPs and subpeak analysis. (a) Schematic of protein detection carried by K72 carriers, where the protein of interest was attached at one end of K72. (b) Schematic illustration of a typical event of one protein–SUP molecule during nanopore translocation. The relationship between the excluded volume (Λ) and the blockade current (ΔI_b) can be estimated as Λ = ΔI_bH_eff²/(σψ) (1), indicating that current blockade is proportional to the excluded volume of translocated molecules. A folded protein, therefore, has a greater excluded volume per unit length than that of an unfolded, linear SUP. Individual nanopore readouts are shown with the protein level as a higher current blockade and the SUP level as a lower current blockade but longer time residence. (c) Structures of a series of proteins with different sizes, Sn (13.0 kDa), SfCherry (20.3 kDa), eGFP (28.6 kDa), and mIFP (35.2 kDa), fused in K72 carriers. (d) Representative ionic current traces (scale bar: 100 pA; 5 s) and typical individual translocation events (scale bar: 50 pA; 80 ms) for the corresponding structures were filtered to 5 kHz for better visualization. Protein translocation was performed in 1 M KCl, 10 mM Tris-EDTA pH 8.0 buffer at −500 mV, and processed with a low-pass filter of 10 kHz. (e) Distributions of the subpeak amplitude extracted from individual translocation events for Sn-K72 (n = 578), SfCherry-K72 (n = 389), eGFP-K72 (n = 765), and mIFP-K72 (n = 341). The mean amplitude of subpeaks shows an increasing trend as the size of proteins increases. Distribution is fitted by the Gumbel function. (f) Distributions of fractional subpeak position (i.e., relative location) suggest subpeaks located at either start or end of individual events, consistent with the protein–SUP structure. Protein Data Bank (PDB) codes: Sn, 3BOI; SfCherry, 4KF4; eGFP, 2Y0G; mIFP, 5VIQ.

Figure 4

Subpeak analysis of flexible design of protein–SUP structure. (a) Schematic of the structure of K36-protein-K36, in which the net charge is equivalent to K72 carriers, but the target protein is bound at the center of the carrier. (b) An example translocation event for K36-protein-K36 shows that the subpeak is shifted to the center of the signal where the target protein is. (c) Structures of K36-SfCherry-K36 and K36-SfGFP-K36 used for single-molecule protein identification. (d) Representative translocation events (scale bar: 50 pA; 20 ms; filtered to 5 kHz) indicated that nanopore signals were consistent with the protein structures with a folded protein at the center. Nanopore experiments were performed in 1 M KCl, 10 mM Tris-EDTA pH 8.0 buffer at −500 mV with a low-pass filter of 10 kHz. (e) Distributions of the subpeak amplitude extracted from individual events for K36-SfCherry-K36 (n = 286) and K36-SfGFP-K36 (n = 307) were fitted with the Gaussian function. (f) Distributions of fractional subpeak position suggest that subpeaks are located at the center of individual events as designed in the structure.

Figure 5

Nanopore sensing of single protein–protein interactions using cationic SUPs. (a) Proof-of-concept demonstrated here is using eGFP-K72 as the antigen to detect anti-GFP antibody. Schematic representation of antibody–eGFP complexes and antibody-eGFP-K72 complexes. Representative ionic current traces for (i) antibody (10 nM)-eGFP (1 nM) interactions and (ii) antibody (10 nM)-eGFP-K72 (1 nM) interactions recorded in PBS + 1 M KCl, 10 mM Tris-EDTA pH 8.0 buffer (1: 9, v/v) at −500 mV. Typical individual events show an enhanced subpeak for antibody-eGFP-K72 complexes compared with eGFP-K72. Data were sampled at 1 MHz and re-filtered to 5 kHz for visualization. (b) Schematic illustration of antibody-eGFP binding and density scatter plots of peak amplitude versus dwell time for separate translocation experiments of (i) 10 nM antibody (n = 1220), (ii) 750 nM eGFP (n = 845), and (iii) 10 nM antibody + 1 nM eGFP (n = 979). (c) Same antigen-antibody binding, but the antigen was replaced with eGFP-K72. Density scatter plots for separate experiments of 10 nM antibody (n = 1220), 1 nM eGFP-K72 (n = 1865), and 10 nM antibody + 1 nM eGFP-K72 (n = 1378). (d) Subpeak analysis of eGFP-K72 (1 nM) interacting with antibody (10 nM). Distribution of subpeak amplitude and fractional position (n = 425) showed the antibody bound with eGFP at the ends of SUP carriers resulting in a boost in subpeak amplitude. Threshold: subpeak amplitude >50 pA; dwell time >1 ms was selected to distinguish antibody-eGFP-K72 complexes from unbound states. (e) Binding assays of 1 nM eGFP-K72 in the presence of anti-GEP antibody ranging from 1 pM to 50 nM. The binding curve was fitted using the Hill equation, and the dissociation constant (K_d value) was determined to be ∼7.85 nM under this condition. Error bars represent one standard deviation of at least three independent experimental repeats.