LSM12-EPAC1 defines a neuroprotective pathway that sustains the nucleocytoplasmic RAN gradient

Abstract

Nucleocytoplasmic transport (NCT) defects have been implicated in neurodegenerative diseases such as C9ORF72-associated amyotrophic lateral sclerosis and frontotemporal dementia (C9-ALS/FTD). Here, we identify a neuroprotective pathway of like-Sm protein 12 (LSM12) and exchange protein directly activated by cyclic AMP 1 (EPAC1) that sustains the nucleocytoplasmic RAN gradient and thereby suppresses NCT dysfunction by the C9ORF72-derived poly(glycine-arginine) protein. LSM12 depletion in human neuroblastoma cells aggravated poly(GR)-induced impairment of NCT and nuclear integrity while promoting the nuclear accumulation of poly(GR) granules. In fact, LSM12 posttranscriptionally up-regulated EPAC1 expression, whereas EPAC1 overexpression rescued the RAN gradient and NCT defects in LSM12-deleted cells. C9-ALS patient-derived neurons differentiated from induced pluripotent stem cells (C9-ALS iPSNs) displayed low expression of LSM12 and EPAC1. Lentiviral overexpression of LSM12 or EPAC1 indeed restored the RAN gradient, mitigated the pathogenic mislocalization of TDP-43, and suppressed caspase-3 activation for apoptosis in C9-ALS iPSNs. EPAC1 depletion biochemically dissociated RAN-importin β1 from the cytoplasmic nuclear pore complex, thereby dissipating the nucleocytoplasmic RAN gradient essential for NCT. These findings define the LSM12-EPAC1 pathway as an important suppressor of the NCT-related pathologies in C9-ALS/FTD.

Fig 1. LSM12 depletion attenuates SG formation upon arsenite-induced oxidative stress.

(A) LSM12 and ATXN2 promote arsenite-induced SG assembly, likely via the same genetic pathway. SH-SY5Y cells stably expressing each shRNA were incubated with 50-μM sodium arsenite (NaAsO₂) for the indicated time and then co-stained with anti-G3BP1 antibody (red), anti-PABPC1 antibody (green), and Hoechst 33258 (blue) to visualize SGs and the nucleus, respectively. (B) The percentage of SG-positive cells, the number of SGs per SG-positive cell, and the size of SGs were quantified using ImageJ software and averaged (n = 15–19 confocal images of random fields of interest obtained from 3 independent experiments; n = 329–1,045 cells). Error bars indicate SEM. **P < 0.01, ***P < 0.001 to control^shRNA cells at a given time point, as determined by 2-way ANOVA with Tukey post hoc test. (C) Immunoblotting of total cell extracts from individual shRNA cell lines with anti-ATXN2, anti-LSM12, and anti-tubulin (loading control) antibodies. The abundance of each protein was quantified using ImageJ and normalized to that of tubulin. Relative protein levels were then calculated by normalizing to those in control^shRNA cells. Data represent means ± SEM (n = 3). n.s., not significant; *P < 0.05, **P < 0.01, ***P < 0.001, as determined by 2-way ANOVA with Tukey post hoc test. (D, E) Overexpression of LSM12, but not ATXN2, restores arsenite-induced SG assembly in LSM12-depleted cells. Control^shRNA and LSM12^shRNA cells were transfected with an expression vector for FLAG, FLAG-tagged LSM12, or FLAG-tagged ATXN2. Arsenite-induced SG assembly was quantified 48 hours after transfection. Data represent means ± SEM (n = 15–16 confocal images obtained from 3 independent experiments; n = 200–820 cells). n.s., not significant; *P < 0.05, **P < 0.01, ***P < 0.001 to control^shRNA cells expressing FLAG at a given time point, as determined by 2-way ANOVA with Tukey post hoc test. All underlying numerical values are available in S1 Data. ANOVA, analysis of variance; ATXN2, ataxin-2; LSM12, like-Sm protein 12; SEM, standard error of the mean; SG, stress granule; shRNA, short hairpin.

Fig 2. LSM12 depletion disrupts the RAN gradient and impairs NCT upon oxidative stress.

(A) LSM12 depletion impairs NCT under oxidative stress conditions in a manner that is independent of ATXN2 or SG assembly. Individual shRNA cell lines were transfected with an expression vector for the S-GFP reporter. SG assembly was then quantified 48 hours after transfection. Transfected cells were co-stained with anti-G3BP1 antibody (red) and Hoechst 33258 (blue). Where indicated, cells were incubated with 50-μM NaAsO₂ or PBS (vehicle control) for 2 hours to induce oxidative stress. Phospho-EIF2α–dependent SG assembly was blocked by treating with 2-μM ISRIB for 3 hours before NaAsO₂ incubation. DMSO was used as vehicle control for ISRIB. (B) NCT of S-GFP reporter proteins was quantified by calculating the ratio of N/C fluorescence in individual cells. Two-way ANOVA detected significant interaction effects of arsenite and ISRIB treatments on NCT only in control^shRNA cells (P = 0.0004). Data represent means ± SEM (n = 100–137 cells from 3 independent experiments). n.s., not significant; ***P < 0.001, as determined by Tukey post hoc test. (C) LSM12 depletion disrupts the nucleocytoplasmic RAN gradient in a manner that is independent of ATXN2 or SG assembly. Cells were incubated with 2-μM ISRIB, 50-μM NaAsO₂, or vehicle controls as described above and then co-stained with anti-RAN antibody (red), anti-G3BP1 antibody (green), and Hoechst 33258 (blue). (D) The relative distribution of endogenous RAN proteins was quantified by calculating the ratio of N/C fluorescence. Two-way ANOVA detected significant interaction effects of arsenite and ISRIB treatments on the RAN gradient only in control^shRNA cells (P < 0.0001). Data represent means ± SEM (n = 95–104 cells from 3 independent experiments). n.s., not significant; **P < 0.01, ***P < 0.001, as determined by Tukey post hoc test. All underlying numerical values are available in S1 Data. ANOVA, analysis of variance; ATXN2, ataxin-2; GFP, green fluorescent protein; ISRIB, integrated stress response inhibitor; LSM12, like-Sm protein 12; N/C, nuclear to cytoplasmic; NCT, nucleocytoplasmic transport; SEM, standard error of the mean; SG, stress granule; shRNA, short hairpin.

Fig 3. LSM12 depletion facilitates the nuclear accumulation of C9ORF72-derived poly(GR) proteins and exacerbates their pathogenic effects.

(A) LSM12 depletion suppresses the maturation of poly(GR)-induced SGs but promotes the nuclear accumulation of poly(GR) granules. Control^shRNA and LSM12^shRNA cells were transfected with a GFP-GR₁₀₀ expression vector and then co-stained with anti-G3BP1 antibody (red) and Hoechst 33258 (blue) at the indicated time after transfection. (B) The assemblies of poly(GR)-induced SGs and nuclear poly(GR) granules were quantified as in Fig 1. Data represent means ± SEM (n = 23–25 confocal images obtained from 3 independent experiments; n = 424–513 GFP-GR₁₀₀–positive cells). n.s., not significant; *P < 0.05, ***P < 0.001, as determined by Student t test. (C) ISRIB treatment suppresses the poly(GR)-induced disruption of NCT in control cells, but not in LSM12-depleted cells. Control^shRNA and LSM12^shRNA cells were co-transfected with expression vectors for S-GFP and FLAG-GR₁₀₀. Where indicated, cells were incubated with 2-μM ISRIB or DMSO (vehicle control) for 5 hours and then co-stained with anti-FLAG antibody (red) and Hoechst 33258 (blue) 48 hours after transfection. (D) NCT of S-GFP reporter proteins was quantified as in Fig 2B. Data represent means ± SEM (n = 142–166 cells from 3 independent experiments). n.s., not significant; ***P < 0.001, as determined by 2-way ANOVA with Tukey post hoc test. (E) LSM12 depletion exacerbates poly(GR)-induced disruption of the nuclear lamina. Control ^shRNA and LSM12^shRNA cells were transfected with an expression vector for GFP or GFP-GR₁₀₀ and then co-stained with anti-lamin B1 antibody (red) and Hoechst 33258 (blue) to visualize nuclear envelope morphology 48 hours after transfection. Yellow arrows indicate GFP-GR₁₀₀–positive cells with severe disruption of the nuclear lamina. (F) Control cells expressing GFP-GR₁₀₀ were scored for nuclear poly(GR) granules and abnormal morphology of the nuclear lamina. The relative percentages of cells with severe nuclear laminar disruption were averaged from confocal images of 6 random fields of interest per condition (n = 55–69 GFP-GR₁₀₀–positive cells from 3 independent experiments). Error bars indicate SEM. ***P<0.001, as determined by Student t test. (G) The relative percentages of control^shRNA and LSM12^shRNA cells with severe nuclear lamina disruption were quantified as described above. Data represent means ± SEM (n = 10 confocal images obtained from 3 independent experiments; n = 183–297 GFP–or GFP-GR₁₀₀–positive cells). n.s., not significant; ***P < 0.001, as determined by 2-way ANOVA with Tukey post hoc test. All underlying numerical values are available in S1 Data. ANOVA, analysis of variance; GFP, green fluorescent protein; LSM12, like-Sm protein 12; N/C, nuclear to cytoplasmic; NCT, nucleocytoplasmic transport; SEM, standard error of the mean; SG, stress granule; shRNA, short hairpin.

Fig 4. An LSM12^V135I mutant allele exhibits dominant-negative effects on the RAN gradient.

(A) Overexpression of LSM12^V135I mutant protein impairs NCT. Control and LSM12^KO cells were co-transfected with expression vectors for S-GFP and FLAG-tagged LSM12 (wild-type or LSM12^V135I mutant). Transfected cells were treated with 50-μM NaAsO₂ or PBS (vehicle control) for 2 hours and then co-stained with anti-FLAG antibody (red) and Hoechst 33258 (blue) 48 hours after transfection. (B) NCT of S-GFP reporter proteins was quantified as in Fig 2B. Two-way ANOVA detected significant interaction effects of arsenite treatment and LSM12 deletion on NCT only in FLAG-expressing cells (P = 0.0026). Data represent means ± SEM (n = 119–138 cells from 3 independent experiments). n.s., not significant; ***P < 0.001, as determined by Tukey post hoc test. (C) Overexpression of LSM12^V135I mutant protein promotes the nuclear accumulation of poly(GR) granules. SH-SY5Y cells were co-transfected with expression vectors for GFP-GR₁₀₀ and FLAG-tagged LSM12 (wild-type or LSM12^V135I mutant) and then co-stained with anti-FLAG antibody (red), anti-G3BP1 antibody (magenta), and Hoechst 33258 (blue) 48 hours after transfection. (D) The assemblies of poly(GR)-induced SGs and nuclear poly(GR) granules were quantified as in Fig 1. Data represent means ± SEM (n = 19 confocal images obtained from 3 independent experiments; n = 279–327 GFP-GR₁₀₀–positive cells). n.s., not significant; *P < 0.05, ***P < 0.001, as determined by 1-way ANOVA with Dunnett post hoc test. (E) Overexpression of LSM12^V135I mutant protein disrupts the nucleocytoplasmic RAN gradient. Control and LSM12^KO cells were transfected with an expression vector for FLAG-tagged LSM12 (wild-type or LSM12^V135I mutant) and then co-stained with anti-RAN antibody (red), anti-FLAG antibody (green), and Hoechst 33258 (blue) 48 hours after transfection. (F) The nucleocytoplasmic RAN gradient was quantified as in Fig 2D. Data represent means ± SEM (n = 119–120 cells from 3 independent experiments). n.s., not significant; ***P < 0.001, as determined by 2-way ANOVA with Tukey post hoc test. All underlying numerical values are available in S1 Data. ANOVA, analysis of variance; GFP, green fluorescent protein; LSM12, like-Sm protein 12; SEM, standard error of the mean; N/C, nuclear to cytoplasmic; NCT, nucleocytoplasmic transport.

Fig 5. LSM12 posttranscriptionally up-regulates EPAC1 expression to sustain the RAN gradient for NCT.

(A) Fold changes in total transcript levels (x-axis) versus translating ribosome-associated transcript levels (y-axis) in LSM12-depleted cells were assessed by RNA sequencing (n = 2 biological replicates; n = 10,856 transcripts) and depicted as a scatter plot. (B) LSM12-depleted cells express low levels of EPAC1 transcript. Total RNA was prepared from control^shRNA and LSM12^shRNA cells. The abundance of each transcript was quantified by real-time RT-PCR and normalized to that of GAPDH. Relative mRNA levels in LSM12^shRNA cells were then calculated by normalizing to those in control^shRNA cells. Data represent means ± SEM (n = 3). *P < 0.05, ***P < 0.001, as determined by Student t test. (C) LSM12-depleted cells express low levels of EPAC1 protein. The abundance of each protein was quantified as in Fig 1C. Data represent means ± SEM (n = 3). **P < 0.01, as determined by Student t test. (D) LSM12 depletion posttranscriptionally decreases EPAC1 expression via the 5′ UTR. EPAC1 reporter plasmids encoding NLUC were generated as depicted in S7B Fig. Control^shRNA and LSM12^shRNA cells were co-transfected with each EPAC1 reporter and FLUC expression vector (normalizing control). Luciferase reporter assays were performed 48 hours after transfection. NLUC activity was first normalized to FLUC activity per condition. Relative expression of each EPAC1 reporter in LSM12^shRNA cells was then calculated by normalizing to the NLUC/FLUC value in control^shRNA cells. Data represent means ± SEM (n = 4). n.s., not significant; *P < 0.05, as determined by Student t test. (E) EPAC1 overexpression restores the nucleocytoplasmic RAN gradient in LSM12-deleted cells. Control and LSM12^KO cells were transfected with an expression vector for FLAG or FLAG-tagged EPAC1 and then co-stained with anti-RAN antibody (red), anti-FLAG antibody (green), and Hoechst 33258 (blue) 48 hours after transfection. (F) The nucleocytoplasmic RAN gradient was quantified as in Fig 2D. Data represent means ± SEM (n = 103–107 cells from 3 independent experiments). n.s., not significant; ***P < 0.001, as determined by 2-way ANOVA with Tukey post hoc test. (G) EPAC1 overexpression suppresses LSM12-deletion effects on the poly(GR)-induced disruption of NCT. Control and LSM12^KO cells were co-transfected with different combinations of expression vectors for S-tdT, GFP-GR₁₀₀, and FLAG-tagged EPAC1 and then co-stained with anti-FLAG antibody (magenta) and Hoechst 33258 (blue) 48 hours after transfection. (H) NCT of S-tdT reporter proteins was quantified as in Fig 2B. Two-way ANOVA detected significant interaction effects of GFP-GR₁₀₀ and LSM12 deletion on NCT only in FLAG-expressing cells (P = 0.0018). Data represent means ± SEM (n = 123–153 cells from 3 independent experiments). n.s., not significant; ***P < 0.001, as determined by Tukey post hoc test. All underlying numerical values are available in S1 Data. ANOVA, analysis of variance; EPAC1, exchange protein directly activated by cyclic AMP 1; FLUC, firefly luciferase; GFP, green fluorescent protein; LSM12, like-Sm protein 12; N/C, nuclear to cytoplasmic; NCT, nucleocytoplasmic transport; NLUC, Nano-luciferase; RNA-seq, RNA sequencing; RT-PCR, reverse transcription PCR; S-tdT, S-tdTomato; SEM, standard error of the mean; shRNA, short hairpin; TRAP-seq, translating ribosome affinity purification sequencing.

Fig 6. EPAC1 suppresses poly(GR) toxicity relevant to NCT and nuclear integrity.

(A) SH-SY5Y cells were transfected with control^siRNA or EPAC1^siRNA 24 hours before transfecting with an expression vector for GFP or GFP-GR₁₀₀. Transfected cells were co-stained with anti-RAN antibody (red), anti-G3BP1 antibody (magenta), and Hoechst 33258 (blue) 72 hours after siRNA transfection. Where indicated, cells were treated with 2-μM ISRIB or DMSO (vehicle control) for 5 hours before antibody staining. (B) The nucleocytoplasmic RAN gradient was quantified as in Fig 2D. Two-way ANOVA detected significant interaction effects of GFP-GR₁₀₀ and ISRIB treatment on the RAN gradient in control^siRNA cells (P = 0.0008), but not in EPAC1^siRNA cells (P = 0.4674 for EPAC1^siRNA #1; P = 0.7870 for EPAC1^siRNA #2); significant interaction effects of GFP-GR₁₀₀ and EPAC1 depletion on the RAN gradient regardless of ISRIB treatment (P < 0.0001 for both EPAC1^siRNA in DMSO; P = 0.0327 for EPAC1^siRNA #1 in ISRIB; P = 0.0013 for EPAC1^siRNA #2 in ISRIB). Data represent means ± SEM (n = 124–145 GFP–or GFP-GR₁₀₀–positive cells from 3 independent experiments). n.s., not significant; ***P < 0.001, as determined by Tukey post hoc test. (C) EPAC1 depletion facilitates the nuclear accumulation of poly(GR) protein. SH-SY5Y cells were co-transfected with each siRNA and a GFP-GR₁₀₀ expression vector as above. Transfected cells were co-stained with anti-G3BP1 antibody (red) and Hoechst 33258 (blue) at the indicated time points after plasmid DNA transfection. (D) The assembly of nuclear poly(GR) granules was quantified as in Fig 1. Data represent means ± SEM (n = 18–19 confocal images obtained from 3 independent experiments; n = 366–413 GFP-GR₁₀₀–positive cells). **P < 0.01, ***P < 0.001, as determined by Student t test. (E) EPAC1 depletion exacerbates the poly(GR)-induced disruption of the nuclear lamina. SH-SY5Y cells were co-transfected with each siRNA and a GFP-GR₁₀₀ expression vector, treated with 2-μM ISRIB or DMSO (vehicle control) and then co-stained with anti-lamin B1 antibody (red), anti-G3BP1 antibody (magenta), and Hoechst 33258 (blue) as described above. Yellow arrows indicate GFP-GR₁₀₀–positive cells with severe nuclear lamina disruption. (F) The abnormal nuclear laminar morphology was quantified as in Fig 3E. Two-way ANOVA detected significant interaction effects of GFP-GR₁₀₀ and EPAC1 depletion on the nuclear integrity regardless of ISRIB treatment (P < 0.0001 for both DMSO and ISRIB). Data represent means ± SEM (n = 15 confocal images obtained from 3 independent experiments; n = 313–545 GFP–or GFP-GR₁₀₀–positive cells). n.s., not significant; ***P < 0.001, as determined by Tukey post hoc test. (G, H) Overexpression of LSM12 or EPAC1 suppresses the poly(GR)-induced disruption of the RAN gradient. SH-SY5Y cells were co-transfected with different combinations of expression vectors for FLAG-tagged LSM12, FLAG-tagged EPAC1, and GFP-GR₁₀₀. Transfected cells were co-stained with anti-RAN antibody (red), anti-FLAG antibody (magenta), and Hoechst 33258 (blue). The nucleocytoplasmic RAN gradient was quantified as in Fig 2D. Two-way ANOVA detected significant interaction effects of GFP-GR₁₀₀ with LSM12 or EPAC1 overexpression on the RAN gradient (P < 0.0001 for both). Data represent means ± SEM (n = 101–109 GFP–or GFP-GR₁₀₀–positive cells from 3 independent experiments). n.s., not significant; ***P < 0.001, as determined by Tukey post hoc test. (I, J) Overexpression of LSM12 or EPAC1 suppresses the poly(GR)-induced disruption of the nuclear lamina. Yellow arrows indicate GFP-GR₁₀₀–positive cells with severe nuclear lamina disruption. Magenta arrows indicate cells overexpressing LSM12-FLAG or EPAC1-FLAG protein. The abnormal morphology of the nuclear lamina was quantified as in Fig 3E. Data represent means ± SEM (n = 11–19 confocal images obtained from 3 independent experiments; n = 235–508 GFP–or GFP-GR₁₀₀–positive cells). n.s., not significant; *P < 0.05, ***P < 0.001, as determined by 2-way ANOVA with Tukey post hoc test. All underlying numerical values are available in S1 Data. ANOVA, analysis of variance; EPAC1, exchange protein directly activated by cyclic AMP 1; GFP, green fluorescent protein; ISRIB, integrated stress response inhibitor; LSM12, like-Sm protein 12; N/C, nuclear to cytoplasmic; NCT, nucleocytoplasmic transport; SEM, standard error of the mean; siRNA, small interfering RNA.

Fig 7. Overexpression of LSM12 or EPAC1 rescues NCT-relevant pathologies in C9-ALS patient-derived neurons.

(A) C9-ALS iPSNs express low levels of LSM12 and EPAC1 proteins. C9-ALS iPSNs (CS28, CS29, and CS52) and control iPSNs (CS0, CS29-ISO, CS4) were harvested 21 days after neuronal differentiation from NPCs. Total cell extracts from 3-week-old iPSNs were resolved by SDS-PAGE and immunoblotted with anti-LSM12, anti-EPAC1, and anti-tubulin (loading control) antibodies. The abundance of each protein was quantified as in Fig 1C. Error bars indicate SEM (n = 3 independent differentiation experiments). *P < 0.05, **P < 0.01, as determined by 1-way ANOVA with Dunnett post hoc test. (B) C9-ALS iPSNs express low levels of LSM12 and EPAC1 transcripts. Total RNA was prepared from 3-week-old iPSNs, and the abundance of each transcript was quantified as in Fig 5B. Data represent means ± SEM (n = 3 independent differentiation experiments). n.s., not significant; *P < 0.05, as determined by 1-way ANOVA with Dunnett post hoc test. (C, D) Overexpression of LSM12 or EPAC1 rescues the RAN gradient in C9-ALS iPSNs. NPCs from C9-ALS iPSCs (CS29) and their isogenic control cells (CS29-ISO) were transduced with individual recombinant lentiviruses that express the indicated FLAG-tagged proteins along with a GFP reporter. iPSNs were fixed 21 days after neuronal differentiation from NPCs and co-stained with anti-RAN antibody (red), anti-MAP2 antibody (magenta), and Hoechst 33258 (blue). The nucleocytoplasmic RAN gradient was quantified as in Fig 2D. Two-way ANOVA detected significant interaction effects of C9-ALS and lentiviral overexpression on the RAN gradient (P = 0.0051 for LSM12; P = 0.0066 for LSM12^V135I; P = 0.0041 for EPAC1). Data represent means ± SEM (n = 100–105 GFP–positive cells from 4 independent differentiation experiments). n.s., not significant; *P < 0.05, **P < 0.01, ***P < 0.001, as determined by Tukey post hoc test. (E, F) Overexpression of LSM12 or EPAC1 suppresses the pathogenic mislocalization of TDP-43 in the cytoplasm of C9-ALS iPSNs. Three-week-old iPSNs (CS29-ISO and C9-ALS CS29) were co-stained with anti-TDP-43 antibody (red), anti-MAP2 antibody (magenta), and Hoechst 33258 (blue). The relative distribution of endogenous TDP-43 proteins was quantified similarly as above. Two-way ANOVA detected significant interaction effects of C9-ALS and lentiviral overexpression on the RAN gradient (P = 0.0154 for LSM12; P = 0.0120 for EPAC1). Data represent means ± SEM (n = 102–104 GFP–positive cells from 4 independent differentiation experiments). n.s., not significant; *P < 0.05, **P < 0.01, ***P < 0.001, as determined by 2-way ANOVA with Tukey post hoc test. (G, H) Overexpression of LSM12 or EPAC1 suppresses caspase-3 activation in C9-ALS iPSNs. Three-week-old iPSNs (CS29-ISO and C9-ALS CS29) were co-stained with anti-cleaved caspase-3 antibody (red), anti-MAP2 antibody, and Hoechst 33258 (blue). The relative percentages of iPSNs expressing cleaved caspase-3 were averaged from 15 confocal images of random fields of interest per condition (n = 771–1,129 GFP-positive cells from 3 independent differentiation experiments). Data represent means ± SEM. n.s., not significant; ***P < 0.001, as determined by 2-way ANOVA with Tukey post hoc test. All underlying numerical values are available in S1 Data. ANOVA, analysis of variance; ATXN2, ataxin-2; C9-ALS, C9ORF72-associated amyotrophic lateral sclerosis; EPAC1, exchange protein directly activated by cyclic AMP 1; GFP, green fluorescent protein; LSM12, like-Sm protein 12; N/C, nuclear to cytoplasmic; NCT, nucleocytoplasmic transport; NPC, neural progenitor cell; SEM, standard error of the mean.

Fig 8. EPAC1 depletion limits RAN-GTP availability for assembly of the RAN-associating nuclear pore complex and suppression of poly(GR) toxicity.

(A) LSM12 deletion dissociates RAN and importin β1 from the RANBP2-RANGAP1 complex. Soluble extracts from control and LSM12^KO cells were immunoprecipitated with control IgG or anti-RANBP2 antibody. Purified IP complexes were resolved by SDS-PAGE and immunoblotted with specific antibodies (left). Asterisks indicate SUMOylated RANGAP1. Input, 4.5% of soluble extracts used in each IP. (B) EPAC1 overexpression restores the assembly of RAN-associating nuclear pore complex in LSM12-deleted cells. Control and LSM12^KO cells were transfected with FLAG or EPAC1-FLAG expression vector. Soluble extracts were prepared 48 hours after transfection and then immunoprecipitated with anti-FLAG antibody. (C) The hydrolysis-resistant GTP analog, GTPγS, blocks the dissociation of RAN and importin β1 from the RANBP2-RANGAP1 complex in EPAC1-depleted cells. Where indicated, soluble cell extracts were preincubated with 0.1 mM GTPγS or DMSO (vehicle control) at 25°C for 30 minutes before IP. (D) RAN overexpression suppresses the nuclear assembly of poly(GR) granules. SH-SY5Y cells were co-transfected with siRNA, GFP-GR₁₀₀, and RAN-FLAG expression vectors, as in Fig 6A. Transfected cells were co-stained with anti-lamin B1 antibody (red), anti-FLAG antibody (magenta), and Hoechst 33258 (blue) 48 hours after plasmid DNA transfection. The assembly of nuclear poly(GR) granules was quantified similarly as in Fig 1. Data represent means ± SEM (n = 15–17 confocal images obtained from 3 independent experiments; n = 280–471 GFP-GR₁₀₀–positive cells). n.s., not significant; *P < 0.05, **P < 0.01, ***P < 0.001, as determined by 2-way ANOVA with Tukey post hoc test. (E, F) RAN overexpression suppresses the poly(GR)-induced disruption of the nuclear lamina. Yellow arrows indicate GFP-GR₁₀₀–positive cells with severe nuclear lamina disruption. Magenta arrows indicate cells overexpressing RAN-FLAG protein. The abnormal morphology of the nuclear lamina was quantified as in Fig 3E. Data represent means ± SEM (n = 16–17 confocal images obtained from 3 independent experiments; n = 325–547 GFP–or GFP-GR₁₀₀–positive cells). n.s., not significant; **P < 0.01, ***P < 0.001, as determined by 2-way ANOVA with Tukey post hoc test. All underlying numerical values are available in S1 Data. ANOVA, analysis of variance; EPAC1, exchange protein directly activated by cyclic AMP 1; GFP, green fluorescent protein; IP, immunoprecipitation; LSM12, like-Sm protein 12; SEM, standard error of the mean; siRNA, small interfering RNA.