Karyopherin enrichment and compensation fortifies the nuclear pore complex against nucleocytoplasmic leakage

Abstract

Nuclear pore complexes (NPCs) discriminate nonspecific macromolecules from importin and exportin receptors, collectively termed "karyopherins" (Kaps), that mediate nucleocytoplasmic transport. This selective barrier function is attributed to the behavior of intrinsically disordered phenylalanine-glycine nucleoporins (FG Nups) that guard the NPC channel. However, NPCs in vivo are typically enriched with different Kaps, and how they impact the NPC barrier remains unknown. Here, we show that two major Kaps, importinβ1/karyopherinβ1 (Kapβ1) and exportin 1/chromosomal maintenance 1 (CRM1), are required to fortify NPC barrier function in vivo. Their enrichment at the NPC is sustained by promiscuous binding interactions with the FG Nups, which enable CRM1 to compensate for the loss of Kapβ1 as a means to maintain NPC barrier function. However, such a compensatory mechanism is constrained by the cellular abundances and different binding kinetics for each respective Kap, as evidenced for importin-5. Consequently, we find that NPC malfunction and nucleocytoplasmic leakage result from poor Kap enrichment.

Figure 1.

Kap enrichment in vivo and removal by Ran mix. (A) Transient transfections of MDCK cells with Kapβ1-EGFP, EGFP-CRM1, and Imp5-mCherry constructs reveal the subcellular localization of Kaps in vivo. Kapβ1 and CRM1 show visible nuclear rim stains indicating their enrichment at the NPCs, whereas Imp5 does not. (B) Fluorescence profiles obtained along the dashed lines shown in A. Kapβ1-EGFP and EGFP-CRM1 show fluorescence spikes (black) that coincide with the edges of the nuclear DAPI staining (blue), whereas similar features are lacking in the Imp5-mCherry signal. Line plots were created using Fiji after smoothing the images with a median filter (2-pixel radius) to minimize noise. (C) Retention of Kapβ1-EGFP, EGFP-CRM1, and Imp5-mCherry at NPCs following digitonin permeabilization and Ran mix treatments. Ran mix–treated cells are shown with original and brightness-adjusted settings for improved visualization. Each series of images was collected using the same imaging conditions. (D) Digitonin and Ran mix treatments significantly reduce the enriched pool of Kapβ1-EGFP (n = 10) and EGFP-CRM1 (n = 10) at the NE. Data points were normalized to the predigitonin NE fluorescence values of each cell. Note: Retention of Imp5-mCherry in digitonin-permeabilized HeLa cells lies below the detection limit, as shown with the brightness adjusted in C. The brightfield image confirms that cells were not removed from the field of view. Further quantification of Imp5-mCherry has been omitted. Error bars denote minimum and maximum measured values. Scale bars, 10 µm.

Figure 2.

Kapβ1, CRM1, and Imp5 bind to NPCs in a concentration-dependent manner. (A) Experimental sequence. (B–D) Representative images of permeabilized HeLa cells incubated in increasing concentrations of (B) exoKapβ1, (C) exoCRM1, and (D) exoImp5. The concentration-dependent accumulation of each Kap is measured from their respective nuclear rim stainings. Cells in the first row are shown with the same dynamic range settings. The brightness is adjusted in the second row to improve visualization of the nuclear rim. Percentages indicate the laser power used to image the cells. Fluorescent beads were used for signal normalization to facilitate comparisons between images (see Materials and methods). Representative images were chosen from the same dataset. (E) Quantification of exoKapβ1 (green), exoCRM1 (blue), and exoImp5 (magenta) at the NPCs and normalized by the maximum fluorescence measured for each Kap at 10 μM. The apparent binding affinity of each Kap to the NPCs was obtained by fitting a single-component Langmuir isotherm to each respective dataset. Data points, error bars, and K_D,NPC values were obtained by propagating means and errors across all replicates (n ≥ 3). Scale bars, 20 µm.

Figure S1.

Effect of 10% Kapβ1 occupancy on Langmuir isotherm analysis simulated using SPR data. This simulation uses SPR data of Kapβ1 binding to a mixed FG Nup layer comprising cNup62, cNup98, cNup153, and cNup214. The Langmuir fit to the original data (black) gives K_D1,SPR = 0.416 μM and K_D2,SPR = 393 μM. These data are then offset by 10% occupancy to simulate the effect of 10% preloading in permeabilized cells (Fig. 2 E). The Langmuir fit to the offset data (red) gives K_D = 1.11 μM. The original data (black) are taken from Kapinos et al. (2017).

Figure 3.

Pairwise binding reveals the relative occupancies of different Kaps. (A) exoCRM1 titration in the presence of 10 μM exoKapβ1. (B) Normalized fluorescence signals of exoCRM1 and exoKapβ1 plotted as a function of exoCRM1 concentration. The maximal observed change in the relative occupancy of exoCRM1 is obtained by subtracting its titration value (blue) from its standalone value (gray) at the highest concentration (i.e., 10 μM exoCRM1; blue arrow). The relative occupancy of exoKapβ1 obtained in the presence of 10 μM exoCRM1 is also shown (green arrow). A single-component Langmuir isotherm fit provides the K_D,NPC of exoCRM1 in the presence of 10 μM exoKapβ1. (C) exoImp5 titration in the presence of 10 μM exoKapβ1. (D) Normalized fluorescence signals of exoImp5 and exoKapβ1 plotted as a function of exoImp5 concentration. The maximal observed change in the relative occupancy of exoImp5 is obtained by subtracting its titration value (magenta) from its standalone value (gray) at 10 μM exoImp5 (magenta arrow). The relative occupancy of exoKapβ1 obtained in the presence of 10 μM exoImp5 is also shown (green arrow). (E) Titration of exoImp5 in the presence of 10 μM exoCRM1. (F) Normalized fluorescence signals of exoImp5 and exoCRM1 plotted as a function of exoImp5 concentration. The maximal observed change in the relative occupancy of exoImp5 is obtained by subtracting its titration value (magenta) from its standalone value (gray) at 10 μM exoImp5 (magenta arrow). The relative occupancy of exoCRM1 obtained in the presence of 10 μM exoImp5 is also shown (blue arrow). Cells in the first row are visualized within the dynamic range shown. The brightness has been adjusted in each second row to better visualize the nuclear rim. Percentages above the panels indicate the laser power used to image the cells. Data points, error bars, and K_D,NPC values were obtained by propagating means and errors across all replicates (n ≥ 3). Scale bars, 20 µm.

Figure S2.

Binding of CRM1 and Imp5 to FG Nup layers in vitro. (A and B) SPR response curves obtained for (A) CRM1 and (B) Imp5 binding to cNup153 (green), cNup62 (black), cNup98 (blue), and cNup214 (red). Vertical jumps in the signal correspond to triple BSA injections used to measure FG Nup layer height. RU, resonance units. (C) Equilibrium binding analysis of Kapβ1 (yellow), CRM1 (blue), and Imp5 (magenta) to cNup62, cNup98, cNup153, and cNup214. Lines represent single-component (solid) or two-component (dashed) Langmuir isotherm fits to the average SPR equilibrium response (R_eq). The mean apparent dissociation constant calculated from n ≥ 4 replicates was used for the fitting. For each replicate, data points were normalized to the maximum response value (R_max) or their sum (R_max1 and R_max2) obtained from the equilibrium fit. Note: The Kapβ1 data were reproduced from Kapinos et al. (2017).

Figure 4.

Equilibrium and kinetic analysis of Kap-FG Nup binding interactions. (A) Equilibrium dissociation constants obtained for Kapβ1 (yellow), CRM1 (blue), and Imp5 (magenta) binding to cNup62, cNup98, cNup153, and cNup214. Boxplots denote the median and the first and third quartiles. K_D,SPRs correspond to the mean values from n ≥ 4 measurements at each condition. Error bars denote SD. (B) Kinetic maps of Kapβ1 (yellow), CRM1 (blue), and Imp5 (magenta) binding to cNup62, cNup98, cNup153, and cNup214. Each map was constructed by averaging over at least four sensograms for every Kap-FG Nup pair. The color intensity indicates the fractional abundance of different kinetic states. All Kaps exhibit multivalent binding, and their kinetic behavior is characterized by different kinetic phases: high affinity (*), low affinity fast (▴), and low affinity slow (○). Arrowheads point to the mean fitted k_off value for each Kap. The data for Kapβ1 have been reproduced from Kapinos et al. (2017).

Figure 5.

Promiscuous binding is balanced by Kap size, binding affinity, and abundance. (A) Theoretically predicted shift in the occupancy of CRM1 from its standalone value at 10 µM CRM1 compared with the presence of 10 µM Kapβ1 background (blue arrow in Fig. 3 B) as a function of CRM1 and Kapβ1 K_D values. (B) Theoretically predicted shift in the pore occupancy of Imp5 from its standalone value at 10 µM Imp5 to when a background of 10 µM Kapβ1 is present (magenta arrow in Fig. 3 D) as a function of Imp5 and Kapβ1 K_D values. (C) Theoretically predicted shift in the relative occupancy of Imp5 from its standalone value at 10 µM Imp5 to when a background of 10 µM CRM1 is present (magenta arrow in Fig. 3 F) as a function of Imp5 and CRM1 K_D values. The bounded regions (black) indicate the K_D values, which are consistent with SPR measurements and are within 1 SD of experimentally measured occupancy shifts. Dashed contour lines indicate the K_D values that result in the average experimentally measured shift (white), and the K_D values that result in 1 SD from these relative occupancy values (gray). Note: The color scale of each heatmap is different.

Figure S3.

Reduction of Kapβ1 occupancy in response to CRM1 and Imp5 binding. (A) Theoretically predicted shift in the occupancy of Kapβ1 calculated between its standalone value at 10 µM Kapβ1 and in the presence of 10 µM CRM1 background (Fig. 3 B, green trace) as a function of Kapβ1 and CRM1 K_D values. (B) Same as A, except a background of 10 µM Imp5 is assumed (Fig. 3 D, green trace). See Fig. 5 in the main text. The marked region (black) indicates the K_D values, which are consistent with SPR measurements and are within 1 SD of experimentally measured occupancy shifts. Dashed contour lines indicate the pairs of K_D values that result in the average experimentally measured shift (white) and the K_D pairs that correspond to the 1 SD from the values of the average relative occupancy (gray). Note: The color scale in each heat map is different.

Figure S4.

Quantitative analysis of MDCK cells after siRNA treatment. (A) Quantification of Kapβ1 silencing in MDCK cells after treatment with Kapβ1-specific siRNA1 or siRNA2 at the amounts shown. The dashed line indicates the removal of unrelated sample lanes that were probed on the same membrane. (B) Proteomic analysis of Kapβ1, CRM1, and Imp5 cellular abundance before and after Kapβ1 silencing. Only Kapβ1 was significantly reduced in MDCK cells, whereas CRM1 and Imp5 levels were not affected. All data points were normalized to the mean value of Kapβ1 abundance in control siRNA cells (n = 4). (C) Analysis of CRM1 silencing efficiency in MDCK cells after treatment with CRM1-specific siRNA. See Materials and methods for details. Note: In all cases, the chemiluminescent signal was recorded using different exposure times to optimally visualize GAPDH or a given Kap. CRM1 and Kapβ1 signals were normalized to the corresponding GAPDH signal from the same lane. MW, molecular weight; M, marker. Source data are available for this figure: SourceData FS4.

Figure 6.

Evidence of Kap compensation at the NPC. (A) A significant fraction of Kapβ1 is depleted from the NE following Kapβ1 silencing. This correlates with (1) an increased enrichment of CRM1 at the NE and (2) increases in both the NE/C and NE/N ratios of Kapβ1 and CRM1, respectively. This suggests that the cytoplasmic pool of Kapβ1, together with the nuclear pool of CRM1, has been recruited to compensate for the depleted Kapβ1 molecules at the NPCs. (B) A small fraction of CRM1 is reduced at the NE following CRM1 silencing. This correlates with (1) an enriched pool of Kapβ1 at the NE that is relatively unchanged, (2) no change to the NE/N ratio of CRM1, and (3) a slight increase in the NE/C ratio of Kapβ1. This suggests that only a small fraction of Kapβ1 is being recruited from the cytoplasm to compensate for depleted CRM1 molecules at the NPCs. For explanation, see the main text. Statistical analysis was performed using the Kruskal-Wallis test. P adjusted values were calculated using the Benjamini-Hochberg procedure (****, P = 0.0001; **, P = 0.0021; * P = 0.0332; ns = 0.1). Scale bars, 10 µm.

Figure 7.

Kap enrichment fortifies the NPC permeability barrier in vivo. (A) Silencing Kapβ1 shifts the steady-state distribution (N/C ratio) of 2xEGFP-NES into the nucleus as a result of increased NPC permeability (i.e., leak). (B) An increase of NPC permeability due to Kapβ1 silencing also results in a shift of 3xEGFP-NES into the nucleus. Impairing 3xEGFP-NES export via CRM1 silencing results in a qualitatively similar but larger shift in the N/C ratio. (C) Silencing CRM1 does not show any detectable change to the N/C ratio of 2xEGFP-NLS. Statistical analysis was performed using the Kruskal-Wallis test. P adjusted values were calculated using the Benjamini-Hochberg procedure (****, P = 0.0001; ***, P = 0.0002; **, P = 0.0021; *, P = 0.0332; ns = 0.1). Scale bars, 10 µm.

Figure 8.

Kapβ1 depletion softens the NPC permeability barrier to nonspecific cargoes in vivo. (A) Representative image sequence showing the recovery of 2xEGFP in the nucleus obtained during a FRAP experiment in control siRNA-treated cells. Lightning indicates the nuclear photobleaching event at t = 0. Scale bar, 10 µm. (B) Fluorescence recovery curves (symbols) and their fits (black lines) as obtained in individual cells. In all cases, an increase in nuclear fluorescence (normalized fluorescence <1) correlates to a concomitant decrease in cytoplasmic fluorescence (normalized fluorescence >1). Both nuclear recovery and cytoplasmic loss of fluorescence are characterized by similar time constants because Kaps do not play a role in the passive diffusion of 2xEGFP. For clarity, only every 10th data point is shown. (C) Kapβ1 silencing expedites the passive exchange of 2xEGFP cargoes across NPCs. (D) Kapβ1 silencing leads to an increase in NPC permeability for 2xEGFP cargoes. Statistical analysis was performed using ordinary one-way ANOVA. P adjusted values were calculated using the Benjamini-Hochberg procedure (***, P = 0.0002; **, P = 0.0021; *, P = 0.0332; ns = 0.1). See main text for details.

Figure S5.

Fluorescence recovery of 2xEGFP within the nucleus in cells overexpressing Kapβ1-iRFP. Top: The time elapsed per frame is 1 s. Kapβ1-iRFP fluorescence is visualized in the last panel. Lightning indicates the nuclear photobleaching event at t = 0. Scale bar, 10 µm. Bottom: τ (left) and permeability (right) of 2xEGFP do not change significantly following Kapβ1-iRFP overexpression. Statistical analysis was performed using a nonparametric (Mann-Whitney) two-tailed test. Error bars denote minimum and maximum measured values.

Figure 9.

Summary of Kap enrichment and compensation at the NPC permeability barrier. (A) Enrichment of Kapβ1 and CRM1 at the NPC under WT conditions based on their respective cellular abundances (C_Kap), apparent binding affinities (K_D,Kap) to the FG Nups, and molecular volumes. (B) Depleting Kapβ1 significantly reduces its occupancy at the NPC, thereby allowing more CRM1 molecules to bind to the FG Nups. However, CRM1 compensation is constrained by its cellular concentration. (C) Depleting CRM1 does not elicit any detectable change to the permeability barrier due to (1) its low WT occupancy and (2) dominance of Kapβ1. Note that the size of CRM1 is larger than Kapβ1.