Nanoelectrochemical Study of Molecular Transport through the Nuclear Pore Complex

Abstract

The nuclear pore complex (NPC) is the proteinaceous nanopore that solely mediates the transport of both small molecules and macromolecules between the nucleus and cytoplasm of a eukaryotic cell to regulate gene expression. In this personal account, we introduce recent progress in our nanoelectrochemical study of molecular transport through the NPC. Our work represents the importance of chemistry in understanding and controlling of NPC-mediated molecular transport to enable the efficient and safe delivery of genetic therapeutics into the nucleus, thereby fundamentally contributing to human health. Specifically, we employ nanoscale scanning electrochemical microscopy to test our hypothesis that the nanopore of the NPC is divided by transport barriers concentrically into central and peripheral routes to efficiently mediate the bimodal traffic of protein transport and RNA export, respectively, through cooperative hydrophobic and electrostatic interactions.

Graphical Abstract

This personal account introduces recent progress in our nanoelectrochemical study of molecular transport through the nuclear pore complex (NPC). We employ nanoscale scanning electrochemical microscopy (SECM) to test our hypothesis that the nanpore of the NPC is spatially organized to efficiently mediate the bimodal traffic of protein transport and RNA export through peripheral and central routes, respectively, through cooperative hydrophobic and elecrostatic interactions.

1. Introduction

Greater chemical understanding and controlling of molecular transport through the nuclear pore complex (NPC) are urgently required to fundamentally advance various research fields including biology and biomedicine.^[1] The NPC solely transports both small molecules and macromolecules between the nucleus and cytoplasm of a eukaryotic cell to play imperative physiological roles.^[2] The NPC is crucial to the regulation of gene expression^[3] and is linked to many human diseases^[4] including cancers^[5] and neuronal diseases.^[6] Importantly, the NPC is relevant to aging and the resultant neurological disorders, because the NPC has a slow turnover rate to accumulate oxidative and other damages in long-lived neurons.^[4b] Moreover, nucleocytoplasmic transport can be genetically disrupted to cause amyotrophic lateral sclerosis and frontotemporal dementia also known as Lou Gehrig’s and Pick’s disease, respectively.^[6a] A cause of these neurodegenerative diseases is the expansion of a hexanucleotide repeat in gene chromosome 9,^[7] which produces proteins with repeats of proline–arginine dipeptides, PR_n, to demonstrate high neurotoxicity^[8] by blocking the NPC.^[9]

Biomedically, insights into chemical interactions between transported molecules and transport barriers of NPC are required toward the rational design of genetic therapeutics for their efficient and safe nuclear import through the NPC. The NPC has been targeted as the sole gate to enable gene therapy and nanomedicine of many human diseases and genetic disorders.^[10] For instance, CRISPR-based gene editing^[11] is highly promising therapeutically,^[12] but is limited by the nuclear import of RNA-guided Cas nuclease, which is prokaryotic and too large (~160 kDa) to passively diffuse through the NPC.^[13] Originally, Cas was directly modified with nuclear localization signal (NLS) peptides,^[14] which are recognized by naturally occurring nuclear transport receptors, e.g., importins, to chaperon the passively impermeable macromolecule through the NPC into the nucleus.^[15] More efficient and safe approaches, however, have been demanded for the in-vivo nuclear import of various macromolecules and nanomaterials^[16] beyond Cas nuclease.

Herein, we introduce recent progress in our nanoelectrochemical study of molecular transport through the NPC as driven by our original hypothesis that sophisticated molecular transport through the NPC is spatially regulated at the nanometer scale. Specifically, we hypothesize that the interior of the NPC nanopore is concentrically divided into central and peripheral routes by transport barriers. Moreover, we hypothesize that the transport barriers possess various populations of hydrophobic and charged amino acids to sort out proteins and RNAs into peripheral and central routes, respectively, not exclusively by hydrophobic interactions as a long-standing consensus, but cooperatively with electrostatic interactions. We test these hypotheses by employing nanoscale scanning electrochemical microscopy (SECM) based on the enabling technologies and protocols developed in our laboratory.

2. Molecular Transport through the NPC

2.1. Central and Peripheral Routes

Our work is built on the hypothesis that the transport barriers of NPC are heterogeneously distributed in the nanopore to efficiently transport different macromolecules, i.e., proteins and RNAs, through different pathways.^[17] The NPC represents one of the largest known protein complexes (in total ~120 MDa) and comprises multiple copies of 30 distinct proteins called nucleoporins (nups) to perforate the double-membraned nuclear envelope (NE).^[18] There has been a consensus that the passive transport of large macromolecules (typically >40 kDa^[13]) through the ~50 nm-diameter pore of NPC is prevented by hydrophobic repeats of phenylalanine and glycine (FG) that are rich in barrier-forming nups.^[15] By contrast, the NPC transports passively impermeable macromolecules with NLS peptides as even larger complexes with transport receptors, e.g., importins (~97 kDa), which have been considered to bind FG domains through hydrophobic interactions.^{[15, 19]}

Specifically, we proposed an advanced transport model where the NPC pore is concentrically divided into central and peripheral routes by FG nups (Figure 1) to mediate the bimodal traffic of RNA export and protein transport, respectively.^[17] In our model, the central pathway is constituted by FG domains assembled around Nup98.^[20] The diameter of the central pathway is determined by the static diameter of the midplane ring formed under Nup98 by oligomers of Nup54 and Nup58. This flexible ring dynamically contracts^[21] to increase the width of the peripheral pathway for the transport of large importin–cargo complexes. This model also accounts for the blockage of the peripheral pathway by a lectin, wheat germ agglutinin (WGA), which binds the N-acetylglucosamine groups of Nup62 projected from the periphery of the midplane ring. Thus, WGA can prevent the interaction of flexible ring with importins, which is needed to locally and dynamically open the peripheral pathway. Originally, the importin-driven conformational change of flexible ring was proposed as the gating mechanism of the central pathway.^[21] In our model, the central route is blocked by Nup98 against passively impermeable macromolecules as well as importin–cargo complexes.^[22] Accordingly, Nup98 mediates RNA export^[23] through the central route efficiently while importins chaperon NLS-tagged proteins separately through the peripheral route. In addition, Nup153^[24] and Nup 214^[25] also serve as major FG nups for protein transport and RNA export, respectively.

Figure 1.

Schematic side (left) and top (right) views of NPC. C and N represent cytoplasmic and nucleoplasmic sides, respectively. Wavy and dashed lines are nuclear filaments and basket, respectively. Adapted with permission from Kim et al., J. Am. Chem. Soc. 2013, Vol. 135, p. 2321. Copyright 2013 American Chemical Society.

2.2. Chemical Properties of Transport Barriers

In our hypothetical transport model, the permeability of central and peripheral routes is controlled by different types of transport barriers based on FG repeats, i.e., mesh and brush barriers, respectively. Specifically, the central mesh barrier is expected to be more hydrophobic and tighter than the peripheral polymer-brush barrier, which is more highly charged and aqueous. The higher hydrophobicity of the central barrier is expected from the higher contents of FG units in Nup98 with ~10 FG dipeptides per 100 amino acids (Table 1).^[22b] Subsequently, cross-linking among cohesive FG units results in the formation of a tight mesh barrier to sterically hinder protein transport.^[26] These tight mesh barriers are also expected for Nup98 because the pore center is more crowded^[27] and because more cohesive GLFG (L, leucine) repeats^[28] are incorporated only in Nup98. By contrast, flexible polymer-brush barriers are formed by the FG repeats of Nup62 and POM121 to block the peripheral route. These FG-nups contain only 1.7 and 3.2 FG dipeptides per 100 amino acids, respectively.^[22b] Thus, the intrinsically disordered FG-rich peptides may remain as flexible polymer-brush barriers to entropically hinder passive macromolecular transport.^[29]

Table 1.

Amino Acid Compositions of FG Domains

	Nup54	Nup58	Nup62	Nup153	Nup214	Pom121	Nup98
FGa	7.3	3.2	3.2	5.0	4.0	1.7	9.6
+a	1.4	2.3	1.4	4.8	3.4	4.8	3.5
–a	0.3	0.6	0.3	4.7	2.1	2.1	1.4
copiesb	32	16	48	32	16	16	48

^[a]The percentages of FG dipeptides, or amino acids with cationic (arginine and lysine) or anionic (asparate and glutamate) residues.

^[b]The number of FG nup in the single NPC. Adapted with permission from Pathirathna et al., Chem. Sci. 2019, Vol. 10, p. 7929. Copyright 2019 Royal Society of Chemistry.

Moreover, our model challenges the long-standing consensus of hydrophobic FG-based transport barriers to hypothesize that the transport barriers use also electrostatic interactions through charged amino acids intermingled between FG repeats (Table 1^{[22b, 30]}) to regulate molecular transport pathway-selectively. For instance, strong electrostatic interactions may be exerted on the molecule transported through the peripheral route owing to a higher content of charged amino acids in POM121 (~8%) than the central FG-rich nups (1.5–4.4%).^[22b] Importantly, POM121 contains ~3 times more positively charged amino acids than negative ones^[22b] to potentially facilitate the transport of negatively charged importin and its cargo complexes through the peripheral route.^[31]

3. Nanoscale SECM of NPCs

We investigated the permeability of NPCs to small probe molecules by employing nanoscale SECM to test and advance our hypothetical transport model. Nanoscale SECM is a powerful electrochemical method that can image and characterize single molecules and single nanostructures, quantitatively address the kinetics of extremely fast heterogeneous and homogeneous reactions, and dynamically resolve a complicated reaction into multiple steps.^[32] The passive transport of small probe molecules through the NPC is unmeasurably fast without nanoscale SECM, which accomplishes extremely high mass-transport conditions across a nanometer-wide gap between the tip and the nucleus to yield unprecedented information about chemical interactions between probe molecules and transport barriers. Moreover, nanoscale SECM is highly sensitive to a change in the permeability of NPCs to small molecules, for instance, by the co-existence of macromolecules in the nanopore and the structural change of transport barriers.

3.1. Enabling Technologies

Experimentally, nanoscale SECM allows us to probe molecular transport through the NPCs at the intact or solid-supported NE. A large intact nucleus was extracted from the oocyte of Xenopus laevis oocyte and studied by SECM (Figure 2A) to establish the intrinsic permeability of NPCs to several small probe molecules. Small nuclear proteins, however, were leached from the nucleoplasm of the intact nucleus to eventually foul the SECM tip, thereby limiting usable probe molecules. This problem was solved by employing the NE that was detached from the nucleoplasm and supported by a microporous solid-state membrane (Figure 2B). The nucleoplasm-free NE not only yielded the consistent permeability of NPCs to probe molecules as studied at the intact NE but also enabled the SECM study of highly charged probe molecules to reveal the electrostatic effect of transport barriers (see below).

Figure 2.

(A) The SECM cell with a swollen nucleus isolated from the oocyte of Xenopus laevis frog. Reprinted with permission from Kim et al., J. Am. Chem. Soc. 2013, Vol. 135, p. 2321. Copyright 2013 American Chemical Society. Photographs of (B) the swollen nucleus with the NE detached from the nucleoplasm and (C) the nucleoplasmic side of the NE spread on the microporous region of a Si₃N₄ membrane. Reprinted with permission from Pathirathna et al., Anal. Chem. 2019, Vol. 91, p. 5446. Copyright 2019 American Chemical Society.

We developed redox-sensitive and ion-sensitive SECM tips to monitor the fast transport of redox-active and redox-inactive probes through the NPCs, respectively. Both types of tips were prepared reproducibly by using a pipet puller, where a glass capillary with or without a Pt wire was heated by a CO₂ laser and pulled mechanically under the conditions controlled by a computer. Specifically, a Pt wire was sealed in a glass capillary, heat-pulled, and milled by using focused ion beam (FIB) technology to obtain a flat tip end (Figure 3A). By contrast, ion-selective tips were obtained simply by pulling a glass capillary and FIB milling the tip end (Figure 3B). Importantly, the glass sheath of the Pt tip was thinned to position the tip end at nanometer distances from the NE without the tip−NE contact (Figure 3C), where the Pt tip was externally biased to electrolyze a target probe, O, thereby yielding a current response. The glass pipet was filled with a water-immiscible electrolyte solution (yellow in Figure 3C) and immersed in the aqueous electrolyte solution to form a liquid/liquid interface at the tip. A Pt electrode was inserted in the electrolyte-filled pipet to apply an external bias to the liquid/liquid interface, where a target ion with a charge of z, I^z, was transferred to yield a current response.

Figure 3.

SEM images of (A) a redox-active Pt SECM tip and (B) a pulled pipet tip. Scale bars are 1 μm. Reprinted with permission from Nioradze et al., Anal. Chem. 2013, Vol. 85, p. 6198, and Kim et al., Anal. Chem. 2014, Vol. 86, p. 2090, respectively. Copyright 2013 and 2014 American Chemical Society. (C) scheme of a Pt tip (left) or an electrolyte-filled pipet (right) positioned near the NE.

3.2. SECM-Induced Mode

We employed the induced mode of SECM to measure the otherwise unmeasurably high permeability of NPCs to small probe molecules. With this operation mode (Figure 3C), an SECM tip locally disturbs the equilibrium partitioning of a target probe between the top and bottom solutions of the NE to induce and detect the local flux of the probe molecules. Accordingly, the tip current is enhanced in comparison with an impermeable membrane, which only blocks the access of the probe to the tip to yield a lower tip current, i.e., the negative feedback effect.

Experimentally, the tip approaches a substrate to obtain a plot of the tip current versus the tip–substrate distance, i.e., approach curve, which is quantitatively analyzed to determine the NE permeability. When the tip is located in the bulk solution containing a target probe, O or I^z, electron transfer or ion transfer at the tip yields a diffusion-limited current response, i_T,∞, given by

$$i_{T,\infty }=4xnF{D}_{w}ca$$ (1)

where x is a function of RG (=r_g/a; a and r_g are the inner and outer radii of a tip), n is the charge transferred at the tip (i.e., the number of electrons or the charge of a probe ion, z), F is the Faraday constant, D_w and c are the diffusion coefficient and concentration of the probe molecule in the aqueous solution. When the tip is positioned near the NE, the tip reaction depleted the target probe between the tip and the NE to induce the transport of the probe molecule from the bottom solution through the NPCs (Figure 3C). The transported probe molecules are also detected at the tip so that the current response of the tip near the NE serves as the measure of the permeability of NPCs to the probe molecule. Advantageously, the concentration gradient of a probe molecule is local and steep under the small tip to accomplish high mass-transport conditions, thereby enabling the determination of the extremely high permeability of NPCs to the probe molecule. Conveniently, the SECM induced mode does not require an external bias across the NE or different concentrations of the probe molecule between the top and bottom solutions to drive the transport of the probe molecule through the NPC, thereby simplifying the experimental setup and quantitative data analysis based on diffusional mass transport without an effect of migration or convection.

4. Central and Peripheral Routes

We employed nanoscale SECM not only to measure the high permeability of the NPCs to small probe molecules but also to demonstrate the presence of central and peripheral routes through the NPCs.^{[17a, 17b]} Previously, we found that the permeability of NPCs to small redox probe molecules was unmeasurably high for microscale SECM and was limited by the diffusion of the probe molecules across a micrometer-wide gap between the SECM tip and the NE.^[33] The extremely high permeability of the NE is attributed to the large and short nanopores of the NPCs as well as their high density. By contrast, nanoscale SECM accomplishes higher mass-transport conditions across the nanoscale tip–NE gap to accurately determine the permeability of the NPC controlled by the diffusion of the probe molecules through the nanopore. The corresponding permeability of the NE, k, as perforated by multiple NPCs is given by the effective medium theory^[34] as

$$k=\frac{2N{D}_{w}r}{2l/\pi r+1/f\left(\sigma \right)}$$ (2)

with

$$f\left(\sigma \right)=\frac{1+3.8{\sigma }^{5/4}}{1-\sigma }$$ (3)

where a planar membrane with a thickness of l is randomly perforated by cylindrical nanopores with a radius of r, a density of N, and a porosity of σ (=πr²N) to mediate the passive transport of a small probe molecule with a diffusion coefficient of D_NPC in the pore to yield γ = D_NPC/D_w. Eq 2 predicts a linear relationship between the permeability and the diffusion coefficient of probe molecules for water-filled nanopores (γ = 1) as ensured in SECM studies of small redox and ionic probes through ultrathin solid-state nanoporous membranes as the model of the NE.^[35]

4.1. Dimensions of Central and Peripheral Routes

We were able to determine the dimensions of central and peripheral routes from the permeability of multiple NPCs to small probe molecules by selectively blocking peripheral routes with WGA. This lectin possesses a radius of ~2.5 nm^[36] to nearly completely block importin-facilitated transport^[37] by plugging the peripheral route^[38] through binding to the N-acetylglucosamine groups of Nup62.^[39] Figure 4a shows the experimental and theoretical SECM approach curves at the intact nucleus (Figure 2A) with and without WGA. The current response based on the diffusion-limited oxidation of (ferrocenylmethyl)trimethylammonium (FcTMA⁺) at a 0.42 μm-radius Pt tip was lowered as the tip approached vertically to the NE because the diffusion of the redox probe to the tip was hindered by the impermeable part of the NE. The tip current, however, was higher than expected with an impermeable membrane (i.e., negative feedback) because the NE was perforated by the NPCs to transport the redox probe to the tip, i.e., the SECM-induced mode (Figure 3C). The higher tip current in the absence of WGA is attributed to the transport of the probe molecule through the entire pore. By contract, the tip current was lower in the presence of WGA, which blocks the peripheral route. The experimental approach curves fitted well with simulated curves, where the concentration of the probe molecule under the tip was calculated by the finite element method to obtain the tip current (Figure 4B). The good fits yielded the passive permeability of 3.1 × 10^–2 cm/s and 5.9 × 10^–2 cm/s with and without WGA, respectively. The numerical analysis also showed that the high permeabilities were measured by bringing the tip at nanoscale separations of ~130 nm from the NE, where the tip eventually contacted the NE to deviate the experimental curves from the simulated curves (Figure 4A).

Figure 4.

(A) Experimental and simulated approach curves at the NE with and without 1.0 g/L WGA. The respective simulation curves employed a = 0.44 μm and 0.42 μm with RG = 2.5. The theoretical negative approach curve was calculated for RG = 2.5.^[40] (B) Cross-section of the concentration profile of FcTMA⁺ around the tip−NE nanogap as simulated by the finite element method with d/a = 0.3. Reprinted with permission from Kim et al., J. Am. Chem. Soc. 2013, Vol. 135, p. 2321. Copyright 2013 American Chemical Society.

We analyzed the permeability of NPCs with and without WGA by the effective medium theory to determine the dimensions of peripheral and central routes. The theory indicates that the passive permeability of 5.9 × 10^–2 cm/s is consistent with eq 1 with N = 40 NPCs/μm², r = 24 nm, l = 35 nm, and γ = 1. These r and l values agree with those measured through the cryo-electron tomography of the nucleus isolated from a mature Xenopus oocyte.^[41] Moreover, the NPC density is consistent with a typical density of ~46 NPCs/μm² for a Xenopus oocyte nucleus as determined by AFM.^[42] The high density corresponds not only to the close packing of NPCs with an external diameter of ~140 nm but also to the presence of ~22 NPCs under the ~0.42 μm-radius Pt tip with an area of 0.55 μm². Moreover, γ = 1 indicates that the probe ions freely diffuse through the entire nanopore without significant interactions with transport barriers. By contrast, peripheral routes were blocked by WGA to yield two-fold lower permeability, which corresponds to 17 nm-radius central routes in eq 1 with otherwise same parameters. A difference of ~7 nm between pore radii with and without WGA is similar to the diameter of WGA (~5 nm)^[36] and corresponds to the static thickness of the peripheral route, which must expand transiently and locally to transport large importin–cargo complexes (see below).

4.2. Ion-Induced Permeabilization of Central and Peripheral Routes

We investigated the passive permeability of the NPCs to highly hydrophobic ions by SECM to find unexpectedly that central and peripheral routes of the NPC can be “opened” to naturally impermeable macromolecules.^[17b] This finding is significant therapeutically because further studies demonstrated that the peripheral route is more readily permeabilized to serve as an attractive route for the nuclear delivery of macromolecular and nanomaterial therapeutics. Specifically, we employed fluorescence microscopy to study how the permeabilization of the NPC by the hydrophobic ions depends on the pathway and the mode of macromolecular transport. We found that extremely hydrophobic ions, i.e., tetraphenylarsonium (TPhAs⁺) and perfluorobutyl sulfonate (PFBS^–), permeabilize the peripheral route of the NPC to the passive transport of bovine serum albumin (BSA), which is naturally impermeable owing to the large molecular weight of ~67 kDa. The anomalous BSA transport was blocked by WGA and reversibly prevented by incubating the permeabilized NPCs in the solution free from the hydrophobic ions. This result indicates that the hydrophobic ions interact with peripheral transport barriers and alter their permeability reversibly. By contrast, the NPC is permeabilized irreversibly by hydrophobic alcohols,^[43] which remove Nup98 from the NPC.^[39] Moreover, such detergents as CHAPS^[44] and Triton X^[45] dissolve lipids from the double membrane to irreversibly permeabilize the NE. Moreover, we found that the hydrophobic ions permeabilize the central route not to BSA but to the importin–BSA complex, which is no longer blocked by WGA.

We attributed the pathway-dependent permeabilization of the NPC to the different structures of central and peripheral transport barriers.^[17b] We proposed that the peripheral route is blocked by flexible polymer-brush barriers^[29] and, subsequently, is more readily permeabilized by the hydrophobic ions than the central route filled with tight mesh barriers.^[26] We speculated that multiple hydrophobic ions cooperatively interact with the hydrophobic FG units of the peripheral barrier to compress the polymer brushes to a lower height as supported by recent theoretical studies,^[46] thereby mediating the passive transport of otherwise impermeable BSA (Figure 5A). Similarly, the tighter central mesh barrier is weakened by the hydrophobic ions to transport importin–BSA complexes, where importin interacts with FG repeats to locally break the weakened mesh-like network among FG repeats (Figure 5B). Without the hydrophobic ions, the interactions between importins and FG repeats are strong enough to permeabilize not the mesh-like central barrier but the brush-like peripheral barrier. Subsequently, the ~7 nm-thick peripheral route is expanded transiently and locally to accommodate a large importin–BSA complex while maintaining the integrity of the central route.

Figure 5.

Mechanism of ion-induced permeabilization of (A) peripheral polymer-brush barriers and (B) central mesh barriers to the passive transport of BSA and the importin-facilitated transport of NLS-tagged BSA, respectively. Reprinted with permission from Kim et al., Anal. Chem. 2014, Vol. 86, p. 2090. Copyright and 2014 American Chemical Society.

5. Electrostatics of Transport Barriers

We also proposed and examined a hypothesis that charged amino acids of FG nups play crucial roles to electrostatically maintain the structure and selectivity of transport barriers in cooperation with hydrophobic FG residues. The synergy between electrostatic and hydrophobic interactions can be important in the rational design of genetically therapeutic substances for efficient nuclear import through the NPC.^[10a] The crucial role of electrostatic interactions, however, had been supported only by using synthetic FG-containing gels as an NPC mimetic.^[47] By contrast, we experimentally proved the theoretical prediction^[31b] that molecular transport through the authentic NPC can be electrostatically gated by cationic amino acids that are more abundant in FG domains than anionic amino acids and even FG dipeptides (Table 1).^[22b] In addition to the functional role of cationic residues, we demonstrated the structural role of anionic residues to maintain the spatial distribution of FG domains, which was disrupted by a high concentration of Ca^2+[48] to lower NPC permeability.^[49] Interestingly, nucleocytoplasmic protein transport is inhibited also in vivo at high intracellular Ca²⁺ concentrations,^[50] thereby exemplifying the importance of nuclear calcium signaling.^[51]

5.1. Ca²⁺ Effect on Barrier Structure and Permeability

Recently, we revealed synchronized Ca²⁺ effects on the permeability and conformation of NPCs by nanoscale SECM and atomic force microscopy (AFM), respectively. This study was enabled by supporting the NE of Xenopus oocyte nucleus on a microporous membrane.^[49] The NE was detached from the nucleoplasm and spread over solid support, i.e., a microporous Si₃N₄ membrane as originally employed for fluorescence transport studies, where the NPCs of the supported NE mediated importin-facilitated macromolecular transport^[52] as expected physiologically.^[1] The self-standing patch of micropore-supported NE was robust enough for AFM imaging (Figure 6A) to reveal the Ca²⁺-dependent conformations of individual NPCs.^[49] Specifically, plugged, entangled, and recessed states were observed at the central region of NPC by AFM when we imaged the cytoplasmic side of micropore-supported NE prepared and fixed in mock intracellular buffer (MIB) containing a physiological sub-micromolar concentration of free Ca²⁺ (Figure 6B), which mimics the concentration of free Ca²⁺ in oocytes. Remarkable differences in topography were demonstrated by cross-sections of AFM images of NPCs at the corresponding states (Figure 6C). A large “plug-like” feature (red line) can be the intrinsic transporter of the NPC^[53] or a macromolecule captured during the translocation to block the nanopore.^[54] A plug was as high as 10 nm in comparison with the NPC that was unplugged and recessed from the surrounding framework (blue line). In addition, a small protrusion with a height of ~1.7 nm was seen at the center of entangled NPC (purple line) and was attributed to the transient intermingling of FG domains as ensured by high-speed AFM.^[55] Overall, we found that ~50% of NPCs were entangled in MIB, where the population of plugged NPCs (~25%) was similar to that of recessed NPCs. By contrast, AFM images of NEs incubated in a nuclear isolation media (NIM) containing a high concentration of 1.5 mM Ca²⁺ (Figure 6D) demonstrated that NPCs were dominantly recessed (~80%) as reported previously^{[39, 56]} to yield ~0% and ~20% of plugged and entangled NPCs, respectively.

Figure 6.

AFM images of (A) nucleoplasmic and (B) cytoplasmic sides of the NE supported by a micropore in MIB. In part B, red, purple, and blue arrows indicate examples of plugged, entangled, and recessed NPCs. (C) Cross-sections of the respective NPCs in MIB. (D) AFM image of the cytoplasmic side of a micropore-supported NE prepared in NIM, where only recessed and entangled NPCs were observed as represented by blue and purple arrows, respectively. Reprinted with permission from Pathirathna et al., Anal. Chem. 2019, Vol. 91, p. 5446. Copyright 2019 American Chemical Society.

We employed nanoscale SECM to demonstrate that the permeability of the NPCs to four redox probes are systematically lower in NIM containing a high concentration of 1.5 mM Ca²⁺ than in MIB containing a physiological sub-micromolar concentration of Ca²⁺. Experimentally, the patch of micropore-supported NE was imaged by SECM (Figure 7A) to measure an approach curve at the center of the NE patch (Figure 7B). The experimental approach curves fitted well with simulated curves to quantitatively ensure the higher permeability of the NPCs in MIB than in NIM. The permeability was proportional to the diffusion coefficient of the redox probes in the aqueous solution to yield a higher slope in MIB than in NIM. The higher slope in MIB was consistent with that measured by using the intact nucleus in MIB to yield γ = 1 in eq 1 for the free diffusion of the redox probes through the entire pore in addition to the expected radius, length, and density as discussed above. By contrast, the lower slope in NIM yielded γ = 0.45 in eq 1 with the otherwise same parameters. The lower γ value corresponds to the higher viscosity of transport barriers in the nanopore by a factor of 2.2 (=1/γ) based on the Stokes–Einstein equation. Overall, we proposed that the high concentration of Ca²⁺ interacted with anionic residues of transport barriers to alter their structure, which viscosified the transport barriers to lower their permeability.

Figure 7.

(A) An SECM image of the micropore-supported NE in MIB as obtained by using FcTMA⁺. Colored scale bars indicate the normalized tip current. (B) SECM approach curves of FcTMA⁺ at micropore-supported NEs prepared in MIB and NIM. Simulated curves at the NE used (a, k) = (0.59 μm, 0.061 cm/s) and (0.45 μm, 0.040 cm/s) for MIB and NIM, respectively, with RG = 2. Reprinted with permission from Pathirathna et al., Anal. Chem. 2019, Vol. 91, p. 5446. Copyright 2019 American Chemical Society.

5.2. Electrostatic Gating

We applied SECM to find that molecular transport through the NPC is electrostatically gated by the cationic residues of the amino acids intermingled between FG repeats.^[17c] Specifically, we found that the transport of polycationic peptide, protamine (~4.5 kDa and ~20 positive charges), through the peripheral pathway of the NPC is electrostatically blocked in the low salt buffer (LSB) containing 1 mM KCl. In this work, protamine transport was driven and monitored by using a micropipet-supported liquid/liquid interface^[57] as an ionophore-based ion-selective tip.^[58] An approach curve with protamine in LSB (red line in Figure 8A) yielded low tip currents at the short tip–NE distances not only because the NE partially hindered the diffusional access of protamine to the tip but also because protamine transport was electrostatically hindered by positively charged transport barriers of NPCs. The tip current became more positive with higher concentrations of 10 and 100 mM KCl in LSB (orange and magenta lines, respectively) to represent higher NPC permeability. As the KCl concentration changed from 1 mM in LSB to 10 and 100 mM, the corresponding Debye length decreased from 3.5 nm to 2.4 and 0.93 nm.^[59] Protamine transport was electrostatically blocked with 1 and 10 mM KCl, where the Debye length was similar to the size of water-filled spaces (5.2 nm) among FG repeats.^[60] The Debye length with 100 mM KCl was much shorter than water-filled spaces of the network to mediate the free diffusion of protamine with a hydrodynamic diameter of 4.0 nm.^[35b]

Figure 8.

(A) Experimental (lines) and simulated (circles) approach curves of protamine at micropore-supported NEs and SiO₂/Si wafers in LSB. Tip diameters are a = 1.5 μm and RG = 1.4. (B) NE permeability to probe ions against their diffusion coefficients in LSB and MIB (red and blue symbols, respectively). Blue and red lines represent the best fits of eq 1 for monovalent ions in LSB and MIB and protamine in LBS, respectively. TBA⁺ represents tetrabutylammonium. Reprinted with permission from Pathirathna et al., Chem. Sci. 2019, Vol. 10, p. 7929. Copyright 2019 Royal Society of Chemistry.

We also found that protamine transport in LSB is freely mediated through the central pathway and electrostatically blocked at the peripheral pathway. Approach curves of protamine in LSB were not affected by WGA (blue line in Figure 8A). The binding of WGA to the peripheral pathway^{[38–39, 61]} did not affect protamine transport through the peripheral pathway, which was already blocked electrostatically. By contrast, protamine transport was blocked by WGA in MIB containing 90 mM KCl (data not shown), where a high ionic strength eliminated an electrostatic effect. Importantly, WAG does bind peripheral FG nups in LSB to partially block the passive transport of a polyanionic oligosaccharide, Arixtra (~1.7 kDa and ~10 negative charges). Moreover, we obtained a linear relationship between NE permeability and diffusion coefficient for small monovalent ions in LSB and MIB (Figure 8B). This result confirms that transport barriers are unaffected by buffers, thereby excluding that slower protamine transport in LSB than in MIB is due to different structures of transport barriers.

We analyzed lower protamine permeability in LSB quantitatively to determine the sizes of peripheral and central regions that are electrostatically affected and unaffected, respectively. We employed the same pore length and density for protamine as used in LSB and MIB to account for similar permeability to monovalent ions (blue line in Figure 8B). Accordingly, the smaller slope of 2.7 × 10³ cm^–1 (red line) for protamine in LSB corresponds to a smaller diameter of 20.8 nm in eq 1 for the protamine-permeable central region. This result indicates that the 13.6 nm-thick peripheral region of 48 nm-diameter pore was electrostatically blocked against protamine transport in LSB. We attribute the electrostatic effect to peripheral POM121 (Figure 1), which has a high population of cationic residues in comparison with anionic residues and even FG dipeptides (Table 1). In addition, each of the other peripheral FG-rich nups, i.e., Nup54, Nup58, and Nup62, can spatially distribute excess positive charges to extend the electrostatically blocked peripheral region beyond the Debye length of ~3.5 nm in LSB. Moreover, we speculate that the edge of Nup98 with a high percentage of cationic residues also contributes to electrostatically block the periphery of NPC, thereby thickening the protamine-impermeable region further.

6. Summary and Outlook: Toward Single NPC Imaging

Our study manifested the future need and feasibility of nanoscale SECM^[32] to investigate the mechanism of molecular transport at the single NPC level. Specifically, nanoscale SECM of single NPCs is required to resolve the permeability of individual NPCs at different conformations as revealed by AFM (Figure 6). Moreover, nanoscale SECM of single NPCs will be able to determine whether the central plug is the permeable transporter of the NPC or the impermeable macromolecule trapped in the nanopore. These two origins of the central plug have been debated for decades^{[53–54, 62]} because they cannot be distinguished by the structural imaging of the central plug with cryo-electron tomography^[62] and atomic force microscopy (AFM).^{[55, 63]} Ultimately, nanoscale SECM is attractive as a non-contact and non-destructive approach^[64] to stochastically monitor the NPC-mediated translocation of single macromolecules without chemical modification in contrast to electrophysiological^[65] and fluorescence^[66] approaches.

We have dramatically improved the spatial resolution of SECM by employing a ~30 nm-diameter pipet tip, which was quantitatively demonstrated by imaging ion transport through the single nanopores of a thin solid silicon membrane as the model of the NE.^[67] Experimentally, a nanopipet tip (Figure 9A) was filled with the water-immiscible electrolyte solution and brought to a few nanometers from a nanoporous solid-state membrane (Figure 9B) in water to obtain a permeability image (Figure 9C). Higher current responses were observed over nanopores, where a probe ion was transported to the nanopipet tip. The expected periodicity of the nanopore array was seen in the SECM image. Nanopores in the image looked larger than their actual diameter by the tip diameter,^[67a] because the tip current was enhanced by the flux of TBA⁺ through the nanopore even when the edge of the tip was positioned over the edge of a nanopore. Moreover, the SECM image of individual nanopores agreed with images simulated by the finite element method.

Figure 9.

TEM images of (A) a quartz nanopipet at ×200,000 and (B) a nanoporous Si₃N₄ membrane at ×60,000. (C) A 300 nm × 600 nm SECM image of single Si₃N₄ nanopores as obtained with a nanopipet tip in 1 M KCl containing 5 mM TBACl. Probe scan rate, 100 nm/s in the x-direction. Reprinted with permission from Chen et al., Anal. Chem. 2017, Vol. 89, p. 9946. Copyright 2017 American Chemical Society.

Biography

Professor Shigeru Amemiya was born in Osaka and raised in Yokohama. He received BS, MS, and PhD degrees in chemistry from the University of Tokyo, where he worked in the group of Professor Yoshio Umezawa. He received a postdoctoral fellowship from the Japan Society for the Promotion of Science to work in the group of Professor Allen J. Bard at the University of Texas at Austin. He joined the Department of Chemistry at the University of Pittsburgh as an assistant professor and is currently a full professor. His research interest resides in fundamental electrochemistry including the development of nanoelectrodes and nanoscale scanning electrochemical microscopy and the mechanistic studies of heterogeneous charge transfer, membrane transport, and electrocatalysis.