Nde1-mediated inhibition of ciliogenesis affects cell cycle re-entry

Abstract

The primary cilium is an antenna-like organelle that is dynamically regulated during the cell cycle. Ciliogenesis is initiated as cells enter quiescence, whereas resorption of the cilium precedes mitosis. The mechanisms coordinating ciliogenesis with the cell cycle are unknown. Here we identify the centrosomal protein Nde1 (nuclear distribution gene E homologue 1) as a negative regulator of ciliary length. Nde1 is expressed at high levels in mitosis, low levels in quiescence and localizes at the mother centriole, which nucleates the primary cilium. Cells depleted of Nde1 have longer cilia and a delay in cell cycle re-entry that correlates with ciliary length. Knockdown of Nde1 in zebrafish embryos results in increased ciliary length, suppression of cell division, reduction of the number of cells forming the Kupffer's vesicle and left-right patterning defects. These data suggest that Nde1 is an integral component of a network coordinating ciliary length with cell cycle progression and have implications for understanding the transition from a quiescent to a proliferative state.

Figure 1

Depletion of Nde1 induces longer cilia. (a) Immunofluorescence staining of centrin2 (green) or Nde1 (red) in NIH-3T3^WT cells. (b) Expression of endogenous Nde1 in asynchronous cultures of NIH-3T3^WT cells (lane 1), NIH-3T3^Nde1-KD1 (lane 2), NIH-3T3^Nde1-KD2 (lane 3), or NIH-3T3^KD-con. cells (lane 4) (upper panel) or α-tubulin (loading control, lower panel). (c) Time course of cilia formation induced by serum starvation in NIH-3T3^WT and NIH-3T3^Nde1-KD2 cells. Cilia or basal bodies were visualized by an antibody against acetylated α-tubulin (red) or γ-tubulin (green). Scale bar, 2.5 μm. (d) Average ciliary length of NIH-3T3^WT cells at 0 (n=17), 6 (n=20), 12 (n=55), or 24h of serum starvation (n=61) and NIH-3T3^Nde1-KD2 at 0 (n=21), 6 (n=42), 12 (n=67), and 24h (n=49) of serum starvation. Ciliary length was measured from ciliated cells in each group per time-point from a representative experiment. “*”, P<0.05, Student’s t test. (e) DAPI-, doublecortin (DCX)-, and adenylyl cyclase 3 (ACIII)-labeled freshly dissociated, mouse E18.5 embryonic cortical neurons transiently transfected with GFP or GFP plus an Nde1-specific shRNA. Scale bar: 5 μm. (f) Average ciliary length of GFP⁺/DCX⁺ embryonic neurons transiently transfected with GFP alone (n=84) or GFP plus Nde1-specific shRNA (n=76). “*”, P <0.05, Student’s t test. (g) Expression of Nde1 (upper panel) or α-tubulin (lower panel) in untransfected RPE1-hTERT cells (lane 1), RPE1-hTERT cells transiently transfected with a control siRNA for 48h (lane 2), human Nde1-specific siRNA (hNde1 siRNA) for 24h (lane 3), or hNde1 siRNA for 48h (lane 4). (h) RPE1-hTERT cells transiently transfected with siRNA control (Scm.^control) or siRNA targeting human Nde1 (hNde1^KD) were double-stained with antibodies against γ-tubulin (red) and acetylated α-tubulin (green) following 24h serum starvation. Scale bar, 10 μm. (i) Average ciliary length of Scm.^control at 0 (n=20), 12 (n=46), or 24h (n=54) of serum starvation and hNde1^KD RPE1-hTERT cells at 0 (n=80), 12 (n=80), and 24h (n=116) of serum starvation. Quantification was obtained from ciliated cells. “*”,P <0.05, Student’s t test. (j) Percentage of Ki-67 positive Scm.^control or hNde1^KD RPE1-hTERT cells synchronized in mitosis by nocodazole treatment (M sync.), followed by a 6h recovery in complete media (0h of serum starvation), followed by serum starvation for 12h or 24h (n=3 independent experiments).

Figure 2

Expression of flag-tagged human Nde1 (f-hNde1) rescues abnormally long cilia in NIH-3T3^Nde1-KD2 cells. (a) Expression levels of f-hNde1 in NIH-3T3^Nde1-KD2 cells. Lysates of NIH-3T3^WT cells stably expressing GFP (NIH-3T3^WT Mock), NIH-3T3^Nde1-KD2 cells stably expressing GFP (NIH-3T3^Nde1-KD2 Mock) or NIH-3T3^Nde1-KD2 cells stably expressing f-hNde1 (NIH-3T3^Nde1-KD2 hNde1) were immunoblotted for Nde1 and α-tubulin as loading control (upper panel) or α-flag (lower panel). (b) Average ciliary length of NIH-3T3^WT cells stably expressing GFP (NIH-3T3^WT Mock; n=57), NIH-3T3^Nde1-KD2 stably expressing GFP (NIH-3T3^Nde1-KD2 Mock; n=78), or NIH-3T3^Nde1-KD2 stably expressing f-hNde1 (NIH-3T3^Nde1-KD2 hNde1; n=100) following serum starvation for 24h. (c) Cilia staining in mock-infected NIH-3T3^WT (a) and NIH-3T3^Nde1-KD2 cells (b), or f-hNde1-infected NIH-3T3^Nde1-KD2 cells (c-l). Antibody against acetylated α-tubulin was used to visualize cilia (a-l), antibody against flag to detect f-hNde1 (c-g), and antibody to γ-tubulin to visualize the basal body (a, b, and h-l). Note that ciliary morphology and length changed according to the amount of f-hNde1 expressed (c-g). (Inset) Schematic representation of the dosage-dependent effect of f-hNde1 on ciliary length and morphology. Low levels of f-hNde1 converted abnormally long cilia back to cilia of normal size (compare a with c), while moderate or high levels of f-hNde1 resulted in bulged (d-f) or stumpy cilia (g), respectively. Scale bar, 2.5 μm. (d) RPE1-hTERT cells were transiently transfected with f-hNde1 and recovered for 48h followed by an additional 24h of serum starvation. Cells were double-stained with antibodies raised against the flag epitope (red) or acetylated α-tubulin (green). (e) Schematic representation of full length Nde1. Coiled-coil domains are shown as grey boxes (18–85 and 90–188). (f) Myc-tagged Nde1^{(L135P,F138P)} (Nde1^{(L135P, F138P)}-myc) was transiently transfected into NIH-3T3^WT cells. Cells were serum starved for 24h and double-stained with antibodies raised against acetylated α-tubulin (green) or the myc epitope (red). Scale bar, 7 μm. (g) Summary of structure-function analysis.

Figure 3

Nde1 suppresses ciliogenesis through LC8. (a) Overexpression of LC8 suppresses the effect of transfected Nde1 on ciliogenesis. NIH-3T3^WT cells were transiently co-transfected with wild type Nde1-myc and flag tagged LC8 (f-LC8) at plasmid ratios of 9:1 and 1:9. Co-transfection of Nde1-myc and flag tagged bacterial alkaline phosphatase (f-BAP) in a plasmid ratio of 1:9 was used as control. Scale bar, 7 μm. (b) Expression levels of LC8 (upper panel) or α-tubulin (loading control, lower panel) in NIH-3T3^WT cells transiently transfected with f-LC8 (f-LC8^OE, lane 1), a control siRNA (Scm.^control, lane 2), or a mouse LC8-specific siRNA (LC8^KD, lane 3). (c) Depletion of LC8 suppresses ciliogenesis. NIH-3T3^WT and NIH-3T3^Nde1-KD2 cells were transiently transfected with LC8-specific siRNA (LC8^KD) and double-stained with antibodies against γ-tubulin (red) and acetylated α-tubulin (green), following 24h of serum starvation. Scale bar, 10 μm. (d) Overexpression of LC8 promotes cilium formation. NIH-3T3^WT cells were transiently transfected with f-LC8 (f-LC8^OE) and double stained with an antibody raised against the flag epitope (f-LC8, red) or acetylated α-tubulin (green). Scale bar, 10 μm. (e and f) Quantification of ciliary length (e) or ciliation (f) of untransfected NIH-3T3^WT cells (n=35), transiently transfected with f-LC8 (f-LC8^OE, n=76), or LC8-specific siRNA (hNde1^KD, n=44). Ciliary length was measured from ciliated cells and percentage of ciliated cells was obtained from three independent experiments (n=3). “*”, P <0.05, Student’s t test. (g and h) Artificial tethering of LC8 at the basal body suppresses cilia formation. GFP-PACT (green, g) or GFP-PACT-LC8/BS (green, h) was transiently expressed in NIH-3T3^WT and NIH-3T3^Nde1-KD2 cells, followed by immunofluorescence staining with γ-tubulin (yellow) or acetylated α-tubulin (red). While both constructs were targeted specifically to the basal body, only GFP-PACT-LC8/BS caused the formation of bulged or stumpy cilia. Scale bar, 10 μm. (i) Schematic diagram summarizing the functional role of Nde1-LC8 interaction in ciliogenesis. Sequestration of LC8 at the basal body suppresses cilia formation, whereas increase of unbound LC8 at the basal body promotes cilia formation.

Figure 4

Nde1 expression inversely correlates with ciliogenesis. (a and c) Centriolar expression of Nde1 decreases upon ciliation. RPE1-hTERT (a) or NIH-3T3^WT (c) cells were serum-starved for the indicated time points. Nde1 (red) or acetylated α-tubulin (green) was visualized by indirect immunofluorescence. Arrows indicate Nde1 localization. Scale bar, 10 μm (a). Scale bar, 2.5 μm (c). (b and d) Quantification of fluorescence intensity ratio of Nde1/γ-tubulin (red/green) signals at the centrosome at 0 (n=128), 12 (n=140) or 24h (n=201) of serum starvation. Fluorescence intensity of coinciding green or red pixels within the boxed area (inset) was measured in a projection of z series collected in 0.5 μm intervals. The range of fluorescence intensity per pixel in box was from 0–255 (n is indicated on graph). (e) Cell cycle dependent regulation of Nde1 expression. Asynchronously proliferating NIH-3T3^WT cells (lane 1, Asyn.) were serum starved for 12h (lane 2, 12h serum −), and 24h (lane 3, 24h serum −), followed by serum re-stimulation for 6h (lane 4, 6h serum +), and 12h (lane 5, 12h serum +). Phosphorylation levels of RB (pRB^S807/811) and Cdc2 (pCdc2^Y15) or levels of cyclin A, cyclin E, and Nde1 were determined by immunoblotting. α-tubulin was used as a loading control. (f) Nde1 levels decrease as cells exit mitosis. Cell cycle analysis of NIH-3T3^WT cells synchronized in mitosis by nocodazole treatment (600ng/ml) for 12h (0h), followed by wash and release into complete media (10% calf serum) for 0.5, 1, 2, or 4h (inset). Lysates from cells arrested in mitosis (0h, lane 1), cells released from mitosis for 0.5 (lane 2), 1 (lane 3), 2 (lane 4), or 4 h (lane 5) were immunoblotted with antibodies against Nde1 (upper panel) or β-actin (lower panel). (g) Schematic diagram of Nde1 expression during the cell cycle and ciliogenesis.

Figure 5

Nde1 depletion causes a delay in cell cycle re-entry. (a) NIH-3T3^WT and NIH-3T3^Nde1-KD2 cells were arrested in G0 by 24h serum starvation (0h serum +) and induced to re-enter the cell cycle by serum re-stimulation for 12h (12h serum +). Cells were pulse-labeled with EdU and immunostained with antibodies against γ-tubulin (green), acetylated α-tubulin (red), and EdU (green, inset). For illustration purposes, EdU labeling is shown as a 25% reduction of the projected image. Arrows indicate EdU-labeled cells. Scale bar, 2.5 μm. (b) Percentage of EdU-positive NIH-3T3^WT (black) or NIH-3T3^Nde1-KD2 (gray) cells following serum re-stimulation (n=3 independent experiments). “*”, P <0.05. Student’s t test. (c) Ciliary length of NIH-3T3^WT (black) at 0 (n=43) or 12h (n=67) of serum re-stimulation and NIH-3T3^Nde1-KD2 (gray) at 0 (n=57) or 12 h (n=78) of serum re-stimulation. “*”, P <0.05. Student’s t test. (d) Percentage of ciliated NIH-3T3^WT (black bar) or NIH-3T3^Nde1-KD2 (gray bar) cells following serum re-stimulation (n=3 independent experiments). “*”, P <0.05. Student’s t test. (e) Ciliary length of EdU-positive NIH-3T3^WT (black, n=21) or NIH-3T3^Nde1-KD2 (gray, n=32) cells following serum re-stimulation. “*”, P <0.05. Student’s t test. (f) Scm.^control or hNde1^KD RPE1-hTERT cells were immunostained with antibodies against γ-tubulin (green), acetylated α-tubulin (red), and EdU (green, inset). Arrows indicate EdU-labeled cells. Scale bar, 10 μm. (g) Percentage of Scm.^control (gray) or hNde1^KD (white) RPE1-hTERT cells labeled with EdU at 24h following serum starvation (0h serum +), 12h (12h serum +), and 24h following serum re-stimulation (24h serum +) (n=3 independent experiments). “*”, P <0.05. Student’s t test. (h) Phospho-RB (pRB^S807/811) and cyclin A levels in Scm.^control (lanes 1, 3, 5, and 7) and hNde1^KD RPE1-hTERT cells (lanes, 2, 4, 6, and 8) at 6h, 12h, 18h, and 24h following serum re-stimulation (serum +). α-tubulin was used as loading control. (i) Ciliary length of RPE1-hTERT Scm.^control (gray) at 0 (n=109), 12 (n=109), or 24h (n=117) of serum re-stimulation and hNde1^KD (white) at 0 (n=110), 12 (112), or 24h (n=125) of serum re-stimulation. “*”, P <0.05. Student’s t test. (j) Percentage of ciliated Scm.^control or hNde1^KD RPE1-hTERT cells following serum re-stimulation (n=3 independent experiments). “*”, P<0.05. Student’s t test.

Figure 6

Knockdown of Nde1 causes a cilium-dependent delay in cell cycle re-entry. (a) Expression levels of Nde1 and IFT20 (upper panel) or α-tubulin (loading control, lower panel) in untransfected RPE1-hTERT cells (lane 1), RPE1-hTERT cells stably expressing an shRNAi construct targeting IFT20 (IFT20^KD) transiently transfected with an siRNA against hNde1 (lane 2), or IFT20^KD cells (lane 3). (b) Expression levels of IFT88/polaris (upper panel), Nde1 (middle panel), or α-tubulin (loading control, lower panel) in untransfected RPE1-hTERT cells (lane 1), transiently transfected with a siRNA against hNde1 (lane 2), siRNA against iFT88/polaris (lane 3), or both siRNAs for hNde1 and IFT88/polaris (lane 4). (c) RPE1-hTERT IFT20^KD (IFT20^KD), IFT20/hNde1^KD, IFT88^KD, or IFT88/hNde1^KD cells were double stained with antibodies against γ-tubulin (green) or acetylated α-tubulin (red). (d) Cilia length distribution of RPE1-hTERT Scm.^control, hNde1^KD, IFT20^KD, IFT20/hNde1^KD, IFT88^KD, and IFT88/hNde1^KD cells. Overall percentile of ciliation is shown next to legend. (e) RPE1-hTERT cells transfected with a control siRNA (Scm.^control) or hNde1 siRNA (hNde1^KD) were treated with Cytochalasin D (Scm.^control CD or hNde1^KD CD) and immunostained with antibodies against γ-tubulin (green) or acetylated α-tubulin (red). (f) Cilia length distribution of Scm.^control, hNde1^KD, Scm.^control CD, or hNde1^KD CD cells. (g) Time course of cell cycle re-entry of Scm.^control, hNde1^KD, IFT88^KD, IFT88/hNde1^KD, IFT20^KD, IFT20^KD/hNde1^KD, Scm.^control CD, and hNde1^KD CD RPE1-hTERT cells in response to serum re-stimulation. (n=3 independent experiments). (h) Time course of cell cycle re-entry of NIH-3T3^WT cells transfected with a control siRNA (NIH-3T3^WT Scm.^control) or IFT88/polaris-specific siRNA (NIH-3T3^WT IFT88^KD) and NIH-3T3^Nde1-KD2 cells transiently transfected with control scrambled siRNA (NIH-3T3^Nde1-KD2 Scm.^control) or IFT88/polaris-specific siRNA (NIH-3T3^Nde1-KD2 IFT88^KD). (n=3 independent experiments). (i) Rab8a^Q67L-GFP induces long cilia in NIH-3T3^WT cells. GFP- or Rab8a^Q67L-GFP – transfected NIH-3T3^WT cells were stained for γ-tubulin (yellow or green) or acetylated α-tubulin (red). Scale bars, 10 μm. (j) Time course of cell cycle re-entry of NIH-3T3^WT GFP^control or NIH-3T3^WT Rab8a^Q67L-GFP) cells in response to serum re-stimulation. (n=3 independent experiments).

Figure 7

Depletion of nde1 in zebrafish leads to longer cilia and smaller Kupffer’s vesicle. (a) Average ciliary length at 10 somite stages (ss) of wild-type zebrafish embryos (control; n=75), embryos injected with nde1 morpholino (nde1 MO; n=100), or co-injected with nde1 morpholino and human NDE1 cap mRNA (rescue; n=87). “*”,P <0.001. Student’s t test. (b) Whole-mount immunofluorescence staining of cilia at the Kupffer’s vesicle at 6 or 10ss embryos using acetylated α-tubulin (green) and atypical PKC (aPKC, red) in wild type (WT) embryos (left panels) or nde1 MO (right panels). Scale bars, 10 μm. (c) Average ciliary length in Kupffer’s vesicle of WT at 6 (n=232) or 10ss (n=241) embryos and nde1 MO at 6 (n=171) or 10ss (n=182) embryos. n represents ciliated cells from 6–10 embryos per group. “*”, P <0.05. Student’s t test. (d) Number of cells in the Kupffer’s vesicle in WT (black bar) at 6 (n=23) or 10ss (n=29) and nde1 MO (gray bar) at 6 (n=24) or 10ss (n=27) embryos. “*”, P <0.05. Student’s t test. (e) Percentage of ciliated cells in the Kupffer’s vesicle of WT (black bar) or nde1 MO (gray bar) at 6 and 10ss. (n=3 independent experiments). (f) Whole-mount immunofluorescence staining of the Kupffer’s vesicle of 10ss WT and nde1 MO with antibodies against phosphorylated histone H3 (pH3, green) or atypical PKC (aPKC, red). Scale bars, 10 μm. (g) Percentage of pH3-positive cells in the Kupffer’s vesicle of 10ss WT (black bar) and nde1 MO (gray bar). (n=3 independent experiments). “*”, P <0.001. Student’s t test. (h) Percentage of embryos in wild-type (control), nde1 MO, or human NDE1 mRNA plus nde1 MO co-injected embryos (rescue) with no expression of southpaw (Absent), expression at the right side (Right), left side (Left), or expression on both sides (Bilateral) of the lateral plate mesoderm at 14 hpf. Dorsal view of southpaw mRNA at 14 hpf. (i) Percentage of embryos in each group with no looping (No loop), leftward (Left), or rightward looping (Right) of the heart tube at 48 hpf determined by the expression pattern of myl-7 mRNA. Laterality defects are manifested as non-looping (No Loop) or leftward looping (Left), whereas wild-type embryos show rightward looping of the heart (Right).