Palmitoylated importin α regulates mitotic spindle orientation through interaction with NuMA

Abstract

Regulation of cell division orientation is a fundamental process critical to differentiation and tissue homeostasis. Microtubules emanating from the mitotic spindle pole bind a conserved complex of proteins at the cell cortex which orients the spindle and ultimately the cell division plane. Control of spindle orientation is of particular importance in developing tissues, such as the developing brain. Misorientation of the mitotic spindle and thus subsequent division plane misalignment can contribute to improper segregation of cell fate determinants in developing neuroblasts, leading to a rare neurological disorder known as microcephaly. We demonstrate that the nuclear transport protein importin α, when palmitoylated, plays a critical role in mitotic spindle orientation through localizing factors, such as NuMA, to the cell cortex. We also observe craniofacial developmental defects in Xenopus laevis when importin α palmitoylation is abrogated, including smaller head and brains, a hallmark of spindle misorientation and microcephaly. These findings characterize not only a role for importin α in spindle orientation, but also a broader role for importin α palmitoylation which has significance for many cellular processes.

Keywords: Importin α; Microcephaly; NuMA; Palmitoylation; Spindle Orientation.

Figure 1. Palmitoylation mediated cortical localization of importin α and importin α cargo binding are required for proper mitotic spindle orientation.

(A) Immunofluorescence images of importin α localization in metaphase-arrested (refer to Methods) HCT116 cells incubated for 1 h with DMSO, drugs that inhibit importin palmitoylation (10 μM Wnt-C59) drugs that enhance importin palmitoylation (50 μM palmostatin), and drugs that inhibit importin cargo binding (25 μM ivermectin) or cargo release (40 μM importazole). Yellow arrows indicate cortical poles and white arrows indicate lateral cortex. Scale bar = 5 μm. Cell boundaries determined by brightfield. (B–D) Quantification of importin α localization in drug treated cells. Measurements of importin α signal intensity were made at three cellular locations: polar cortex, lateral cortex and cytoplasm. Polar cortex measurements were made for each cell at the pole with the higher measure of intensity, a similar method was used for lateral cortex measurements and cytoplasm measurements were made at the midline of the cell. These measurements were normalized to each other on a cell-by-cell basis by determining the ratio of cortical vs lateral importin α, cortical vs cytosolic importin α, and lateral vs cytosolic importin α. Mean +/− SEM n = 60 from 2 biological replicates. (C) **p = 0.0093, (D) **p = 0.0029 determined by Student’s t-test. (E) Western blot of HCT116 cell fractions following subcellular fractionation to separate PM, cytoplasmic, organelle, and nuclear fractions. PM and cytoplasmic fractions shown. Western blots were immunostained for PMCA1 (PM marker), β-tubulin (cytoplasmic marker), and importin α (KPNA2). (F) Cartoon representation of a metaphase cell with misoriented spindles mounted on a coverslip indicating the angle, α, which was measured as the arctangent of the horizontal distance, z, over the vertical distance, x, between the two centrosomes to determine the angle of spindle structures. (G) Immunofluorescence images of metaphase-arrested HCT116 cells in presence of DMSO or 10 μM Wnt-C59 stained for NuMA. DMSO treated cell represents a properly oriented cell with a spindle angle of 0 degrees relative to the parallel of the coverslip the cells were mounted on. Wnt-C59 treated cell represents a severely misoriented cell with a spindle angle of 90 degrees relative to the parallel of the coverslip the cells were mounted on. Scale bar = 5 μm. (H) Quantification of mitotic spindle angles for metaphase-arrested HCT116 cells incubated in DMSO, 10 μM Wnt-C59, 50 μM palmostatin, 40 μM importazole, 25 μM ivermectin, or 100 μM 2-bromopalmitate for 1 h prior to analysis. All drug treatments except palmostatin significantly increased the mean spindle angle of metaphase cells relative to a DMSO control. Mean +/− SEM, n = 60 mitotic cells from 2 biological replicates. *p = 0.0103, **p = 0.0012, ***p = 0.0009, ****p < 0.0001 determined by Student’s t-test. Refer to Methods for method of determining spindle angle. (I) Immunofluorescence images of metaphase-arrested HCT116 cells transfected via nucleofection with importin α-mCherry-HA-CaaX and incubated in either DMSO or 10 μM Wnt-C59 for 1 h prior to analysis stained for NuMA. Scale bar = 5 μm. (J) Quantification of mitotic spindle angles for metaphase-arrested HCT116 cells incubated with DMSO or 10 μM Wnt-C59 expressing importin α-mCherry-HA or importin α-mCherry-HA-CaaX. Cells expressing importin α-mCherry-HA-CaaX showed no spindle misorientation when treated with Wnt-C59. Mean +/− SEM, n = 60 mitotic cells from 2 biological replicates *p = 0.0134, ****p < 0.0001 determined by Student’s t-test. (K) Immunofluorescence images of metaphase-arrested RPE-1 cells incubated with DMSO, 10 μM Wnt-C59, or 50 μM palmostatin for 1 h prior to analysis stained for γ-tubulin. DMSO treated cells are representative of cells with properly oriented spindles. Wnt-C59 treated cells were significantly misoriented and palmostatin treated cells were properly oriented when compared to DMSO control. Scale bar = 5 µm. (L) Quantification of mitotic spindle angles for metaphase-arrested RPE-1 cells incubated with DMSO, 10 μM Wnt-C59 or 50 μM palmostatin for 1 h prior to analysis. Wnt-C59 treatment significantly increased the mean spindle angle of metaphase cells relative to a DMSO control while palmostatin treatment did not significantly increase the mean spindle angle relative to a DMSO control. Mean +/− SEM, n = 60 mitotic cells from 2 biological replicates. *p = 0.0175 determined by Student’s t-test. Source data are available online for this figure.

Figure 2. Importin α interacts with NuMA, but not Dlg at the metaphase cell cortex in a palmitoylation-dependent manner.

(A) Western blot of NuMA immunoprecipitation from HCT116 cells treated with DMSO, 10 μM Wnt-C59 or 50 μM palmostatin for 1 h prior to analysis. Immunoprecipitation of NuMA followed by importin α and NuMA western blot. (B) Schematic of PLA quantification. ROIs of quantification represented by dashed lines. (C) Immunofluorescence images of DuoLink proximity ligation assay (PLA) probing interaction of NuMA with importin α (KPNA2) in interphase and metaphase-arrested HCT116 cells in the presence of DMSO, 10 μM Wnt-C59 or 50 μM palmostatin for 1 h prior to analysis. White dashed lines indicate cell borders as determined by brightfield images. Scale bar = 5 μm. (D) Quantification of the percentage of importin α (KPNA2)-NuMA PLA foci at the polar cortex, lateral cortex and cytoplasm in DMSO, Wnt-C59 and palmostatin treated cells. Foci were enriched at the polar cortex in DMSO treated cells, the cytoplasm in Wnt-C59 treated cells, and the lateral cortex in palmostatin treated cells. Mean +/− SEM, n > 136 foci, *p = 0.0132, ****p < 0.0001 determined by Student’s t-test. (E) Immunofluorescence images of DuoLink PLA probing interaction of Dlg with importin α (KPNA2) in interphase and metaphase-arrested HCT116 cells in the presence of DMSO, 10 μM Wnt-C59 or 50 μM palmostatin. White dashed lines indicate cell borders. Scale bar = 5 μm. (F) Quantification of the percentage of importin α (KPNA2)-Dlg PLA foci at the polar cortex, lateral cortex and cytoplasm in DMSO, Wnt-C59, and palmostatin treated cells. Localization of PLA foci did not change across three drug treatments. Mean +/− SEM, n > 297 foci, all data points are non-significant as determined by Student’s t-test. (G) Immunofluorescence images of DuoLink proximity ligation assay probing interaction of NuMA with nucleofected importin α constructs (KPNA2-HA-mCherry and KPNA2-HA-mCherry-CaaX) in interphase and metaphase-arrested HCT116 cells in the presence of DMSO or 10 μM Wnt-C59 for 1 h prior to analysis. White dashed lines indicate cell borders as determined by brightfield images. Scale bar = 5 μm. (H) Quantification of the percentage of nucleofected importin α constructs (KPNA2)-NuMA PLA foci at the polar cortex, lateral cortex and cytoplasm in Wnt-C59 treated cells. Cells nucleofected with importin α-HA and treated with Wnt-C59 did not exhibit enrichment of foci at the PM and instead showed foci throughout the cell. Cells nucleofected with importin α-HA-CaaX and treated with Wnt-C59 exhibited foci enrichment at the PM indicating that expression of importin α-HA-CaaX rescued the effects of Wnt-C59 treatment. Mean +/− SEM, n > 195 foci. Polar *p = 0.0279, lateral *p = 0.0393 determined by Student’s t-test. (I) Western blot of importin α, NuMA and β-tubulin in metaphase-arrested HCT116 cells in the presence of DMSO, 10 μM Wnt-C59, 50 μM Palmostatin, 40 μM importazole, 25 μM ivermectin, or 100 μM 2-bromopalmitate. (J, K) Quantification of NuMA and importin α (KPNA2) expression levels, respectively, relative to β-tubulin protein levels for each condition. Mean +/− SEM, n = 3. Source data are available online for this figure.

Figure 3. Palmitoylated importin α regulates cortical localization of NuMA and dynein/dynactin in metaphase.

(A) Confocal images of NuMA localization in metaphase-arrested HCT116 cells in the presence of DMSO, 10 μM Wnt-C59, 50 μM palmostatin, 40 μM importazole or 25 μM ivermectin. Yellow arrows indicate cortical poles and white arrows indicate lateral cortex. Scale bar = 5 μm. (B–D) Quantification of NuMA localization in drug treated cells. Measurements of NuMA signal intensity were made at three cellular locations: polar cortex, lateral cortex and cytoplasm. Polar cortex measurements were made for each cell at the pole with the higher measure of intensity, a similar method was used for lateral cortex measurements and cytoplasm measurements were made at the midline of the cell. These measurements were normalized on a cell-by-cell basis by determining the ratio of polar vs lateral NuMA, polar vs cytosolic NuMA, and lateral vs cytosolic NuMA. Mean +/− SEM n > 40. (B) **p = 0.0064, ***p = 0.0005, ****p < 0.0001, (C) *p = 0.0142, **p = 0.0047, (D) **p = 0.0047, ****p < 0.0001 determined by Student’s t-test. (E) Immunofluorescence images of p150^Glued localization in metaphase-arrested HCT116 cells treated with either DMSO, 10 μM Wnt-C59, or 50 μM palmostatin. Yellow arrows indicate cortical poles and white arrows indicate lateral cortex. Scale bar = 5 μm. (F–H) Quantification of p150^Glued localization in drug-treated cells. Measurements of p150^Glued signal intensity were made at three cellular locations: polar cortex, lateral cortex and cytoplasm. Polar cortex measurements were made for each cell at the pole with the higher measure of intensity, the same method being used for lateral cortex measurements and cytoplasm measurements at the midline of the cell. These measurements were normalized on a cell-by-cell basis by determining the ratio of polar vs lateral p150^Glued, polar vs cytosolic p150^Glued, and lateral vs cytosolic p150^Glued. Mean +/− SEM n = 60. (F) *p = 0.0336, (G) *p = 0.0112 determined by Student’s t-test. Source data are available online for this figure.

Figure 4. Regulation of palmitoylation is required for proper brain development in Xenopus laevis.

(A) Brightfield images of NF stage 42 X. laevis grown in the presence of DMSO, 100 μM Wnt-C59 or 1 mM palmostatin. Scale bar = 500 μm. (B) Measurements of drug treated stage 42 X. laevis head shape by 3 metrics: distance between eyes, snout length, and overall head area. All 3 metrics of head shape were significantly altered from DMSO control in both Wnt-C59 and palmostatin treatments. Mean +/− SEM n > 55. Distance between eyes ****p < 0.0001, snout length ***p = 0.0003 DMSO-Wnt-C59 p = 0.0002 DMSO-Palmostatin, head area *p = 0.0389, ****p < 0.0001 determined by Student’s t-test. (C) Immunofluorescence images of DMSO treated NF stage 46 X. laevis immunostained for the neuroprogenitor marker nestin and stained with Hoechst to visualize DNA. Scale bar = 250 μm. (D) Quantification of total cell count in forebrains and percentage of forebrain cells positive for nestin signal in NF stage 46 X. laevis grown in the presence of DMSO, 100 μM Wnt-C59 or 1 mM palmostatin. Quantifications were performed on maximum projection images from z-stack images of X. laevis brains with a parent-child analysis to determine the number of total cells as determined by Hoechst signal that also were positive for nestin signal. All Wnt-C59 treated X. laevis embryos died before reaching NF stage 46 while all palmostatin treated X. laevis display a significantly reduced neuroprogenitor population by nestin positive cell count. Mean +/− SEM n = 12, ***p = 0.0001 determined by Student’s t-test. Source data are available online for this figure.

Figure 5. Overexpression of CaaX modified importin α in the developing X. laevis brain partially rescues developmental defects due to PORCN inhibition.

(A) Confocal images of NF stage 42 X. laevis brains from X. laevis grown in the presence of DMSO or 100 μM Wnt-C59 immunostained for phospho-histone H3, a marker of actively dividing cells. Scale bar = 250 μm. (B) Confocal images of NF stage 42 X. laevis brains from X. laevis expressing importin α modified with a C-terminal CaaX domain which forces cortical localization via farnesylation and grown in the presence of 100 μM Wnt-C59 immunostained for phospho-histone H3 and the modified importin α-CaaX construct. (C) Confocal images of NF stage 42 X. laevis brains from X. laevis expressing an mCherry construct modified with a C-terminal CaaX domain and grown in the presence of 100 μM Wnt-C59 immunostained for phospho-histone H3 and the modified CaaX construct. (D) Quantification of the number of phospho-histone H3 positive cells in stage 42 X. laevis brains of X. laevis grown in the presence of DMSO or 100 μM Wnt-C59 and expressing importin α-CaaX or mCherry-CaaX. Wnt-C59 treated X. laevis embryos showed a significantly reduced number of phospho-histone H3 positive cells in the brain compared to DMSO treated X. laevis. X. laevis embryos expressing importin α-CaaX in the brain display a partial rescue of the reduced phospho-histone H3 levels which was not recapitulated in X. laevis expressing mCherry-CaaX. Mean +/− SEM n > 10, *p < 0.0168, ****p < 0.0001 determined by Student’s t-test. Source data are available online for this figure.

Figure 6. Importin α regulates mitotic spindle orientation through mediating NuMA localization to the metaphase cortex and maintenance at the cell cortex through anaphase in a palmitoylation-dependent manner.

Proposed model of palmitoylated importin α’s role in astral microtubule anchoring as a transporter of NuMA and a scaffold at the cell cortex for astral microtubule anchoring proteins to maintain cortical localization throughout metaphase and anaphase.

Figure EV1. Proteome screens confirm palmitoylation of human importin α-1 and enrichment of NLS containing proteins by cellular localization and function.

(A) Prediction of palmitoylated cysteine residues in Human Importin α-1 (KPNA2) by GPS-Palm. Prediction score is on a 0-1 scale with 0 representing low confidence of palmitoylation and 1 representing high confidence of palmitoylation. Three highest confidence residues are highlighted in green. (B) Cellular localization of NLS containing proteins not predicted to be in the nucleus. (C) Cellular functions of NLS containing proteins found at the PM sorted by gene ontology terms.

Figure EV2. KPNA2 cellular localization boundary determination.

(A) Representative immunofluorescence and bright-field images for cells used in Fig. 1A to determine KPNA2 localization and cell boundary determination for each drug treatment. Scale bar = 5 μm. (B) Quantification of the circularity of metaphase-arrested HCT116 cells treated with DMSO, 10 μM Wnt-C59, 50 μM palmostatin, 40 μM importazole or 25 μM ivermectin for 1 h. Mean +/− SEM n = 60, **p = 0.0031 determined by Student’s t-test.

Figure EV3. CaaX modified importin α localizes to the plasma membrane independent of palmitoylation.

Immunofluorescent images of HCT116 cells transfected with importin α-mCherry, mCherry-CaaX or importin α-mCherry-CaaX treated with DMSO, Wnt-C59 or palmostatin. Importin α-mCherry-CaaX localizes to the plasma membrane in all drug treatments. Scale bar = 5 μm.

Figure EV4. Wnt-C59 treated X. laevis tadpoles exhibit shortened tail length.

(A) Brightfield images of NF Stage 42 X. laevis tadpole tails treated with DMSO from 24 to 72 hpf. DMSO treated tadpoles exhibit normal development and typical length tails. Scale bar = 500 μm. White dashed line indicates examined tadpole tail for each image. (B) Brightfield images of NF Stage 42 X. laevis tadpole tails treated with Wnt-C59 from 24 to 72 hpf. Wnt-C59 treated tadpoles exhibit abnormal development, shortened tails, and abnormally shaped tails. Scale bar = 500 μm. White dashed line indicates examined tadpole tail for each image.

Figure EV5. Importin α overexpression produces severe developmental defects in X. laevis embryos.

Immunofluorescent images of X. laevis embryos co-injected with importin α-mCherry-CaaX pcDNA4TO and pcDNA6TR at 24, 48, and 72 h post fertilization. Embryos were injected at either the 1 cell, 2 cell (injected into 1 of 2 cells), or 16 cell stage (injected into the D11 blastomere). All 1 and 2 cell injected embryos exhibit high mortality and severe developmental defects. D11 injected embryos had better survivability though some embryos still exhibited defects. Scale bar = 500 μm.