Fig 2: Steric binding of KPNA3 interferes with the self-association and condensation of NPAT. (A) Coimmunoprecipitation of FLAG-tagged C-region and GFP-M-region in the presence or absence of GFP-KPNA3. (B) Coimmunoprecipitation of FLAG-tagged C-region or its ΔNLS mutant and GFP-M-region. (C) DIC images of FLAG-NPAT (12.5 μM) condensation in the presence of GFP, KPNA3, or KPNA3 ARM 3–5 mutants at a concentration of 12.5 μM. Scale bar, 10 µm. The size of droplets was quantified (NPAT+GFP: n = 62; NPAT+KPNA3: n = 56; NPAT+KPNA3 ARM3–5: n = 85). ****: P < 0.0001, ns: not significant. (D) Schematic diagram of KPNA3, ARM 3–5 and GFP-ARM 3–5. (E) DIC images of FLAG-NPAT (12.5 μM) condensation in the presence of GFP, FLAG-KPNA3 ARM 3–5, or GFP-KPNA3 ARM 3–5 mutant at a concentration of 12.5 μM. Scale bar, 10 µm. The size of droplets was quantified (NPAT+GFP: n = 78; NPAT+myc-KPNA3 ARM3–5: n = 74; NPAT+GFP-KPNA3 ARM3–5: n = 53). **: P < 0.01. (F) Coimmunoprecipitation of FLAG-tagged C-region and GFP-M-region in the presence of Myc-KPNA3, Myc-AMR 3–5, or GFP-ARM 3–5. (G) Schematic diagram of KPNA3, ARM 1–5, and ARM 3–10. (H) DIC images of FLAG-NPAT (12.5 μM) condensation in the presence of GFP, FLAG-KPNA3, FLAG-ARM 1–5, or FLAG-ARM 3–10 mutant at a concentration of 12.5 μM. Scale bar, 10 µm. The size of droplets was quantified (NPAT+GFP: n = 74; NPAT+KPNA3: n = 80; NPAT+KPNA3 ARM1-5: n = 97; NPAT+KPNA3 ARM3-10: n = 110). ****: P < 0.0001. (I) Coimmunoprecipitation of FLAG-tagged C-region and GFP-M-region in the presence of GFP, Myc-KPNA3, Myc-ARM 1–5, or Myc-ARM 3–10 mutant. Source data are available for this figure: SourceData F7.

Fig 3: The binding to KPNA3 suppresses the condensation of NPAT. (A) Immunofluorescence staining of endogenous NPAT in MCF-7 cells transfected with FLAG vector, FLAG-KPNA3, or FLAG-KPNA5. Scale bar, 10 µm. The insets show the magnification of the boxed region (scale bar, 5 μm). The numbers of endogenous NPAT foci per cell were quantified (FLAG vector: n = 109; FLAG-KPNA3: n = 102; FLAG-KPNA5: n = 107). ****: P < 0.0001, ns: not significant. (B) Schematic diagram of KPNA3, KPNA5, and KPNA5-3-hybrid mutant. (C) Coimmunoprecipitation of GFP-NPAT C-region and FLAG-tagged KPNA3 or KPNA5-3-hybrid mutant. (D) Immunofluorescence staining of endogenous NPAT in MCF-7 cells transfected with FLAG vector, FLAG-KPNA3, FLAG-KPNA5, or FLAG-KPNA5-3-hybrid. Scale bar, 10 µm. The insets show the magnification of the boxed region (scale bar, 5 μm). The numbers of endogenous NPAT foci per cell were quantified (FLAG vector: n = 104; FLAG-KPNA3: n = 92; FLAG-KPNA5: n = 95; FLAG-KPNA5-3-hybrid: n = 87). ****: P < 0.0001, ns: not significant. (E) Schematic diagram of NPAT, KPNA3, and NPAT-KPNA3 fusion mutant. (F) Western blot analysis of GFP-NPAT and GFP-NPAT-KPNA3 fusion mutant expression in MCF-7 cells. (G) Confocal images of MCF-7 cells transfected with GFP-NPAT or GFP-NPAT-KPNA3 fusion. Bar: 10 μm. The ratio of cells with nuclear NPAT foci was quantified for three independent experiments. ***: P < 0.001. Source data are available for this figure: SourceData F5.

Fig 4: KPNA3 interacts with the NLS of NPAT. (A) MCF-7 cells treated with or without ivermectin (25 μM) were stained with phalloidin for actin and NPAT antibody for HLB. Bar: 10 μm. The insets show the magnification of the boxed region (scale bar, 5 μm). The numbers of HLB foci per cell were quantified as violin plots (DMSO: n = 126; ivermectin: n = 132). ****: P < 0.0001. (B) MCF-7 cells treated with or without ivermectin (25 μM) were subjected to qPCR analysis for transcription levels of the indicated RD histone genes and two control genes. Data in the bar graphs represent the means ± SEM values of the relative transcription levels for three independent experiments. *: P < 0.05, **: P < 0.01, ***: P < 0.001, ****: P < 0.0001, ns: not significant. (C) Confocal images of 293T-GFP-NPAT cells treated with or without ivermectin (25 μM). Bar: 10 μm. The insets show the magnification of the boxed region (scale bar, 5 μm). The GFP-NPAT foci per cell were quantified as violin plots (DMSO: n = 85; ivermectin: n = 73). ****: P < 0.0001. (D) Coimmunoprecipitation of FLAG-NPAT and HA-tagged KPNA family members. (E) Schematic diagram of NPAT and its truncation mutants. (F) Coimmunoprecipitation of FLAG-tagged NPAT truncation mutants and GFP-KPNA3. (G) Schematic diagram of the NPAT C-region and its ΔNLS deletion mutant (Δ1,368–1,405). (H) Confocal images of representative MCF-7 cells expressing the GFP-NPAT C-region and its NLS deletion mutant. Scale bar: 5 μm. Data in the bar graphs represent the means ± SD values of the ratio of nuclear GFP signal of WT or NLS deletion mutant for three independent experiments. ****: P < 0.0001. (I) Coimmunoprecipitation of the FLAG-tagged NPAT C-region or its ΔNLS deletion mutant and HA-KPNA3. Source data are available for this figure: SourceData F1.

Fig 5: Schematic model for KPNA3-mediated NPAT decondensation and HLB remodeling. We found that KPNA3 determines the nuclear import of NPAT by interacting with the C-terminus NLS region. A C-SIF region proximal to NLS is essential for mediating the self-association of NPAT to drive the condensation of NPAT in nuclei. In the cytoplasm, NLS-mediated KPNA3 association sterically blocks the C-SIF motif-dependent NPAT self-association, which ensures the nuclear import of NPAT and prevents aberrant cytoplasmic HLB formation.