Importin α regulates ciliogenesis and cilia length with implications for Xenopus nephrogenesis

Abstract

Cilia are microtubule-based organelles essential for a wide range of biological processes ranging from facilitating fluid flow to transducing developmental and growth signals. Defects in cilia structure or function can lead to ciliopathies. Ciliogenesis and cilia length regulation depend on protein transport to the ciliary base, but the underlying molecular mechanisms remain unclear. Here we identify that the nuclear adapter protein, importin α, has conserved localization in human epithelial primary cilia and Xenopus laevis (X. laevis) epidermal multiciliated cells. Importin α regulates both ciliogenesis and cilia length maintenance, dependent on its localization to the membrane via palmitoylation or in the cytoplasm when not palmitoylated. In addition, we identify key ciliary proteins, CEP164 and ARL13B, as candidate binding partners of importin α through their NLS sequence and the requirement of this binding interaction for proper ciliogenesis and cilia length. Disruption of importin α palmitoylation in X. laevis causes defects in nephrogenesis which is rescued by forced membrane localization of importin α. These findings reveal a previously unrecognized role for importin α in cilia biology and advances understanding of congenital kidney diseases.

Summary

Importin α possesses a novel regulatory role in ciliogenesis and cilia length in human primary cilia and Xenopus laevis multiciliated cells and is involved in Xenopus nephrogenesis.

Introduction

Cilia are specialized organelles that protrude from the plasma membrane and are composed of a microtubule-based axoneme encased by a ciliary membrane. Primary cilia play critical roles in development and tissue homeostasis, whereas multiciliated epithelial cells contain hundreds of motile cilia that are essential for fluid flow in various organs (Spassky & Meunier, 2017; Ishikawa & Marshall, 2011). Defects in cilia structure or function can lead to a class of human disorders known as ciliopathies that affect nearly all organ systems (Reiter & Leroux, 2017; Waters & Beales, 2011). Many ciliopathies are associated with defects in the structural components required for ciliogenesis, the complex process which describes the assembly of cilia. While critical features of ciliary biology have been identified such as proper basal body docking for ciliogenesis, and formation of the transition zone for selective protein entry for cilium length maintenance (Zhao et al, 2022), the molecular mechanisms of these processes remain poorly understood.

One interesting pattern that has emerged from previous work investigating these ciliary processes is the discovery that several proteins involved in nuclear import localize within the cilia and are essential for cilia formation (Johnson & Malicki, 2019). The karyopherin transport protein family, known as importins, comprise the primary nuclear import proteins. This includes importin α, with seven human isoforms (KPNA1-7) and importin β, which has 2 isoforms (KPNB1-2) (Oka & Yoneda, 2018; Lee et al, 2006). While importin β₂ can bind cargo directly through conserved PY motifs, importin β₁ requires heterodimerization with importin α to bind cargo. Importin α functions as an adapter protein, binding cargo in the cytoplasm via nuclear localization signal (NLS) sequences. When importin α binds cargo, it can subsequently traverse the nuclear pore by passing through the nuclear pore protein (NUP) selective filter (Lu et al, 2021; Lee et al, 2006). Bound cargo is then released via RanGTP binding to importin β. Several NUPs have been identified as localizing to the primary cilia and are required for ciliogenesis (Chen et al, 2023; Dishinger et al, 2010; Hurd et al, 2011; Fan et al, 2011; del Viso et al, 2016). Of interest, recent work has unveiled novel functions for importin α isoform 1 (KPNA2) when localized to the plasma membrane with four palmitate lipids, a post-translational lipid modification known as palmitoylation, regulating nuclear and spindle size scaling in Xenopus laevis (X. laevis). The X. laevis palmitoylation sites of importin α were confirmed to be conserved in humans and are required for mitotic spindle orientation regulation (Sutton et al, 2025; Brownlee & Heald, 2019). In addition, Sutton et al 2025 performed a genome-wide screen in human and X. laevis to identify proteins with conserved NLS motifs and sorted the proteins by cellular localization. Second only to the nucleus, the plasma membrane was found to be highly enriched for proteins with conserved NLS motifs, suggesting a possible role for palmitoylated importin α in their localization. When sorting these plasma membrane localized NLS-containing proteins by gene ontology terms, this study identified several ciliary proteins with conserved NLS motifs, implicating palmitoylated importin α in cilia biology through localization of these proteins.

Building upon this previous work, we propose that palmitoylated importin α mediates an underlying molecular mechanism of protein trafficking for ciliogenesis and cilia length maintenance through the binding of NLS-containing ciliary cargo proteins. We demonstrate through confocal and ultra-expansion microscopy (U-ExM) the conserved characterization of importin α localization in cilia and at the cilium base in both in vitro human epithelial primary cilia and in vivo X. laevis epidermal multiciliated cell (MCC) models. We show that the palmitoylation status of importin α and its ability to bind and release cargo proteins are required for both primary and motile ciliogenesis and demonstrate that membrane-bound importin α can rescue decreased ciliogenesis phenotypes by forcing importin α to the membrane via farnesylation. Furthermore, we highlight the critical requirement for a balance between palmitoylated membrane-bound and non-palmitoylated cytoplasmic importin α to mediate protein shuttling for cilia length maintenance in both primary and motile cilia. Proteins previously implicated in ciliogenesis and ciliary length, Centrosomal protein 164 (CEP164) and ADP-ribosylation factor-like protein 13b (ARL13B), respectively, contain NLS sequences and in this study, we suggest they may be binding partners of importin α. Furthermore, we found that importin α binding mediated transport of CEP164 is required for proper ciliogenesis, while ARL13B ciliary localization is attributed to importin α palmitoylation levels and binding/releasing activities. Lastly, we demonstrate how disruption of importin α palmitoylation leads to decreased renal primary cilia as well as a kidney morphological defects, reminiscent of congenital anomalies of the kidney and urinary tract (CAKUT) in developing X. laevis embryos. Remarkably, these kidney defects were rescued by forcing importin α back to the plasma membrane. Overall, this work offers a previously unexplored role for palmitoylated importin α in ciliogenesis and length maintenance in both motile and primary cilia and advances our understanding of the importance of renal primary cilia during nephrogenesis.

Results

Importin α localizes to ciliary structures in human epithelial primary cilia and X. laevis epidermal MCCs

Importin β₂ has previously been identified as having a role in trafficking proteins to primary cilia and NUPs have been shown to localize to the cilia base (Hurd et al, 2011; Dishinger et al, 2010; Chen et al, 2023; Obado & Rout, 2016; del Viso et al, 2016). This suggests that importin-mediated transport may be playing a role in ciliary functions. Consequently, we aimed to investigate the role of importin α in ciliogenesis by characterizing its localization to ciliary structures using the well-establish model system for ciliogenesis studies, hTERT-RPE-1 (RPE-1) cells (May-Simera et al, 2018; Zhao et al, 2022). Importin α localization was observed throughout RPE-1 primary cilia, with enrichment at the cilium base (Fig. 1A). To further elucidate its localization at the base, RPE-1 cells were labeled with a basal body associated protein, Chibby (CBY), which we found to colocalize with importin α at the cilium base (Fig. 1B). Given the diminutive size of basal bodies (centriolar width and length are approximately 250nm and 400nm, respectively (Winey & O’Toole, 2014; LeGuennec et al, 2021; Greenan et al, 2018), we employed U-ExM for increased resolution of importin α localization at the cilium base (Gambarotto et al, 2021, 2018; Laporte et al, 2024). To demonstrate the feasibility, we performed U-ExM on ciliated RPE-1 cells and successfully expanded the sample, allowing for visualization of more intricate staining for the primary cilium and basal body (Fig. 1C). Notably, U-ExM imaging clarified the localization of importin α throughout the basal body of RPE-1 cells, corroborating our previous observations of importin α at the cilium base (Fig. 1D). Next, we utilized the X. laevis model system to determine whether this localization is conserved within motile cilia. The Xenopus system is an ideal model for studying in vivo ciliogenesis, as the developing Xenopus embryo contains MCCs along the entire outer body axis, known as epidermal MCCs (Walentek & Quigley, 2017). The external localization of these epidermal MCCs along with the rapid development of X. laevis facilitate easy manipulation and visualization of motile cilia (Werner & Mitchell, 2012, 2013). Importin α localization was detected throughout motile cilia and at the cilia base, consistent with our findings in RPE-1 cell primary cilia (Fig. 1E). Another unique aspect of the Xenopus system is the ability to physically shear and collect motile cilia from the epidermal MCCs. This process, termed reversible deciliation, involves a calcium-shock to the embryo, shedding the cilia while the basal bodies remain at the apical surface followed by subsequent cilia regrowth (Werner & Mitchell, 2013; Seidl et al, 2023) (Fig. 1F). Immunoblot analysis with this technique revealed the presence of importin α in the isolated motile cilia, validating our microscopy results (Fig. 1G). Collectively, these findings demonstrate through a multifaceted approach, the conserved localization of importin α at ciliary structures.

Fig. 1. Importin α localizes to primary cilia in human epithelial RPE-1 cells and X. laevis epidermal multiciliated cells (MCCs).

To induce ciliogenesis, RPE-1 cells were serum starved for 48 hrs and immunostained to identify importin α localization. A) RPE-1 cell immunostained for KPNA2 (importin α) (green), acetylated-α-tubulin (magenta), and DNA (blue). Scale bars = 5μm (full image) and 1μm (inset image). B) RPE-1 cell immunostained for KPNA2 (importin α) (green), acetylated-α-tubulin (magenta), Chibby (CBY) (cyan) and DNA (blue). Scale bars = 5μm (full image) and 1μm (inset image). C) Ultra-Expansion confocal microscopy (U-ExM) of a ciliated RPE-1 cell immunostained for acetylated-α-tubulin (magenta) and γ-tubulin (yellow). Scale bar =2μm. D) U-ExM of ciliated RPE-1 cells immunostained for KPNA2 (importin α) (green) localization with acetylated-α-tubulin (magenta). Scale bar = 1μm. E) X. laevis NF stage 30 epidermal MCCs immunostained to characterize Kpna2 (importin α) (green) localization at cilia, acetylated-α-tubulin (magenta), the plasma membrane (white arrow) and at the cilia base (yellow arrow). Scale bar = 5μm. F) Schematic of deciliation protocol on X. laevis NF stage 30 embryos. G) Immunoblot of isolated cilia and X. laevis embryo for Kpna2 (importin α), acetylated-α-tubulin, Gapdh, and β-tubulin. Schematic created with BioRender.com.

Palmitoylated importin α regulates ciliogenesis in RPE-1 cells

Previous research has demonstrated that importin α can be targeted to the plasma membrane upon palmitoylation, thereby impacting nuclear and spindle size scaling, and mitotic spindle orientation (Brownlee & Heald, 2019; Sutton et al, 2025). Given our characterization of importin α at ciliary structures, we hypothesized that palmitoylated importin α may play a regulatory role in ciliogenesis. To investigate this, we explored the role of palmitoylated importin α in regulating ciliogenesis in RPE-1 cells that form primary cilia via the intracellular ciliogenesis pathway which is characterized by its hallmark ciliary pocket. This process begins with the mother centriole acquiring Golgi apparatus-derived vesicles containing proteins which begin the assembly of the ciliary sheath internally, before migrating apically to the plasma membrane, creating the ciliary pocket that serves as a platform for vesicular trafficking (Bernabé-Rubio & Alonso, 2017; Benmerah, 2013). To ascertain whether palmitoylated importin α regulates ciliogenesis, pharmacological inhibitors were used to disrupt palmitoylation status. S- and O-palmitoylation are the processes of attaching palmitate lipids to cysteine or serine residues, respectively, through either the action of palmitoyl acyl transferase or membrane-bound O-acyltransferase enzymes. In this study, we used the inhibitor Wnt-C59, a specific small molecule inhibitor of porcupine O-acyltransferase (PORCN), which is responsible for serine palmitoylation of only two known targets: Wnt, whose palmitoylation is required for its secretion (Proffitt et al, 2013), and importin α, whose palmitoylation is essential for its membrane localization (Brownlee & Heald, 2019). Previous work from our group has utilized and characterized Wnt-C59 treatment as resulting in a decrease of palmitoylated membrane-bound importin α and an increase in cytosolic importin α in both human cell culture and X. laevis embryos (Brownlee & Heald, 2019; Sutton et al, 2025) (Fig. S1A). Depalmitoylation, however, involves the removal of palmitate lipids, regulated by acyl protein thioesterases (APTs). To target depalmitoylation we used the inhibitor palmostatin which inhibits the activity of APT 1/2, broadly inhibiting the depalmitoylation of palmitoylated proteins, leading to an increase in palmitoylated membrane-bound importin α and a decrease in cytosolic importin α (Brownlee & Heald, 2019; Sutton et al, 2025) (Fig. S1A). Additionally, importazole was used to inhibit the ability of importin α to release NLS-containing cargo (Soderholm et al, 2011).

To elucidate the role of importin α in ciliogenesis, RPE-1 cells were serum starved for 48 hours to induce ciliogenesis; after 24 hours of serum starvation, RPE-1 cells were treated with the aforementioned drugs while remaining in serum free media for an additional 24 hours, for cilia formation (Fig. S1B) (Cooper, 2000; Pitaval et al, 2010; Ishikawa & Marshall, 2011). This experimental window was used to determine if palmitoylated importin α is required for cilia formation by quantification of the percentage of ciliated versus unciliated RPE-1 cells. Palmostatin-treated RPE-1 cells exhibited increased membrane-bound palmitoylated importin α, but no apparent effect on cilia formation (Fig. 2A,B). Wnt-C59-treated RPE-1 cells resulted in a reduction of membrane-bound palmitoylated importin α, translocating it to the cytoplasm and concomitantly decreasing the percentage of ciliated cells (Fig. 2A,B). Inhibition of PORCN activity leads to a decrease in Wnt signaling by abrogating Wnt palmitoylation and subsequent secretion. Therefore, in order to address the off-target effects of Wnt-C59 treatment, RPE-1 cells were treated with Wnt-C59 in exogenous Wnt-conditioned medium, which contains active WNT3A ligand, activating Wnt signaling independent of PORCN inhibition (Roelink & Nusse, 1991; Willert et al, 2003). In these conditions, we would predict that any ciliary defects observed would thus be attributable solely to the lack of palmitoylated importin α at the membrane. Notably, addition of exogenous Wnt did not rescue the percentage of ciliated cells, suggesting that the cilia defects observed upon PORCN inhibition are not due to a decrease in Wnt signaling. Additionally, siRNA-mediated knockdown of PORCN in RPE-1 cells phenocopied the decreased ciliation results observed with Wnt-C59 treatment, validating our previous pharmacological drug treatments with a genetic approach (Fig. S1C-E). Knockdown of PORCN in RPE-1 cells resulted in a ~15% ciliation reduction compared to control siRNA, and, intriguingly, ~40% of the ciliated cells exhibited a stubby cilia phenotype, suggesting a potential effect on regulation of proper cilia length (Fig. S1D,E). Lastly, RPE-1 cells treated with importazole also exhibited a decrease in ciliogenesis, suggesting that importin α releasing bound NLS-containing ciliary cargos is essential for ciliogenesis to occur (Fig. 2A,B). To determine if the observed ciliogenesis defects after Wnt-C59 treatment are attributable to the absence of membrane-bound importin α specifically, RPE-1 cells were transfected with an importin α-mCherry-HA-CAAX (importin α-CAAX) plasmid, which facilitates membrane localization via farnesylation through the CAAX motif, irrespective of palmitoylation status (Fig. 2C,D) (Choy et al, 1999; Ahearn et al, 2011; Sutton et al, 2025). We first validated the importin α-CAAX plasmid as significantly increasing the amount of importin α at the membrane vs cytoplasm compared to transfection with the importin α-WT plasmid in RPE-1 cells (Fig. S2). Next, as outlined in the experimental timeline in Supp. Figure 1A, 24 hours prior to ciliogenesis, RPE-1 cells were transfected with importin α-CAAX or a wild-type (WT) importin α plasmid, followed by drug treatment with DMSO or Wnt-C59 to elucidate the requirement of membrane-bound importin α for proper ciliogenesis. Under transfection of the importin α constructs, there was no observed alterations in importin α localization to the cilia or basal body, but strikingly, we observed a rescue of ciliogenesis with importin α-CAAX transfection in comparison to importin α WT + Wnt-C59 and Wnt-C59 treatment alone (Fig. 2C-E). It is worth noting that these results strongly suggest that the observed ciliogenesis defects are not due to pharmacological off target effects or from Wnt signaling disruption by Wnt-C59 treatment, but rather attributable to changes solely in importin α membrane localization, as forcing importin α to the membrane is sufficient to rescue the observed defects.

Fig. 2. Palmitoylated importin α regulates primary ciliogenesis of human epithelial RPE-1 cells.

A) Non-ciliated RPE-1 cells were serum starved for 24 hours and treated with listed the drugs for an additional 24 hours in serum free conditions: DMSO, 10μM Wnt-C59, 50μM palmostatin, or 40μM importazole. Drug treated RPE-1 cells immunostained for KPNA2 (importin α) (green), acetylated-α-tubulin (magenta), and DNA (blue). Scale bar = 5μm. B) Quantification of %of ciliated vs. % of unciliated RPE-1 cells after drug array treatment for 24hrs.* P<0.05; ** P<0.01; ns, not significant, one-way ANOVA with Dunnett’s multiple comparisons test, n>200 cells per treatment, 3 replicates. Mean ± SEM. C) Schematic of RPE-1 cells transfected via nucleofection, with 1μg of either importin α-mCH-HA or D) importin α-mCH-HA-CAAX plasmid. RPE-1 cells transfected with importin α constructs, serum starved and treated with DMSO or 10μM Wnt-C59 for 24 hours and immunostained for HA tag (green), acetylated-α-tubulin (magenta), and DNA (blue). Scale bars = 5μm (full image) and 1μm (inset image). E) Quantification of % of ciliated vs. % of unciliated RPE-1 cells transfected with either construct, serum starved and treated with DMSO or 10μM Wnt-C59 for 24hrs. * P<0.05, Student’s t-test, n=100 cells per treatment, 3 replicates. Mean ± SEM. Schematics created with BioRender.com.

Lastly, as RPE-1 cells form cilia specifically via the intracellular pathway we explored if there is a conserved role for palmitoylated importin α in cells which utilize a different ciliogenesis program termed the alternative pathway. The alternative ciliogenesis pathway involves the basal body docking directly to the plasma membrane, after which there is a recruitment of ciliary proteins involved in building the cilium (Bernabé-Rubio & Alonso, 2017; Sorokin, 1968; Ganga et al, 2021). To test the importance for palmitoylated importin α in regulating the alternative pathway, Madin-Darby Canine Kidney (MDCK-II) cells were utilized. These cells were treated prior to ciliogenesis with the same array of drugs and monitored for changes in ciliation. MDCK-II cell ciliation results corroborated those observed in Fig. 2B, demonstrating the requirement for membrane localization of importin α and its proper function to release bound cargo in the alternative ciliogenesis pathway (Fig. S1F,G). Taken together, our data demonstrate a novel role for palmitoylated importin α regulating ciliogenesis.

Palmitoylated importin α regulates ciliogenesis in X. laevis epidermal MCCs

Multiciliogenesis is a process involving centriole biogenesis that generates hundreds of basal bodies, which subsequently dock at the membrane and facilitate vesicle transport to deliver proteins required for cilia formation (Werner & Mitchell, 2012). To investigate whether palmitoylated importin α regulates the formation of multiciliated cells in the X. laevis epithelium (Fig. 3A), X. laevis embryos at NF stage 18 were treated with palmostatin or Wnt-C59 to inhibit importin α depalmitoylation and palmitoylation, respectively, until NF stage 37 to ensure maximum drug penetration into MCCs (Fig. 3B). Only Wnt-C59 treatment resulted in a decrease of the amount of MCCs per unit area in X. laevis embryos, suggesting a global effect on the amount of ciliated MCCs (Fig. 3C). Additionally, we observed that Wnt-C59 treated X. laevis embryos have a shortened body axis phenotype previously ascribed to defects in Wnt signaling related to development of the anterior-posterior axis (Hikasa & Sokol, 2013; Petersen & Reddien, 2009) (Fig. 3B). Previous studies have demonstrated that various mucociliary cells on the epidermis rely on the Wnt/β-catenin signaling pathway for proper cell differentiation and patterning of mucociliary epithelium, although the timing of this signaling during development is unclear (Walentek & Quigley, 2017). Therefore, we sought to elucidate whether the reduction in the quantity of epidermal MCCs, as evidenced by the absence of cells containing acetylated-α-tubulin, is attributable to a lack of cellular differentiation of MCCs or a defect in ciliogenesis. We also wanted to test whether the effects on ciliogenesis observed upon PORCN inhibition were due to defects in the Wnt pathway, or if these defects are due to disruption of importin α membrane localization specifically. To accomplish both of these objectives, we microinjected X. laevis embryos at the specific blastomere identified to give rise to the epidermis with an importin α-WT or importin α-CAAX plasmid, followed by drug treatment to determine if importin α membrane localization can rescue ciliogenesis defects. Ciliogenesis was rescued in importin α-CAAX microinjected X. laevis embryos as quantified by the integrated density of the motile cilia in each MCC (Fig. 3D,E), demonstrating that the previously observed multiciliogenesis phenotypes were due to importin α mislocalization specifically and not Wnt signaling defects. In support of this, we also disrupted Porcn expression by microinjecting a morpholino (MO) targeting Porcn into one cell at the two-cell stage of the developing X. laevis embryo, with the uninjected cell serving as an internal control. Porcn MO injected X. laevis embryos exhibited a loss of cilia formation in the epidermal MCCs and a loss of epithelial pigment, as an indicator of targeted knockdown expression (Yoon et al, 2022), compared to the uninjected side, further demonstrating the role for palmitoylation of importin α in regulating ciliogenesis (Fig. S3). Collectively, these results highlight the critical and conserved role of palmitoylated importin α in the formation of both primary cilia and multicilia, across in vivo and in vitro models.

Fig. 3. Palmitoylated importin α has a conserved role in regulating multiciliogenesis in X. laevis epidermal MCCs.

A) Schematic of hypothesized role for palmitoylated importin α regulating the formation of motile cilia in the X. laevis epidermal MCCs. B) X. laevis NF stage 37 embryos treated for 3 days with DMSO, 100μM Wnt-C59 or 1mM palmostatin followed by immunostaining of epidermal MCCs by acetylated-α-tubulin (gray). Image in bottom right highlight area of embryo for MCC counts quantified in 3C. Scale bars = 200μm (full image) and 50μm (inset image). C) Quantification of ciliated MCCs/area per drug treatment. ***P<0.001; ns, not significant, one-way ANOVA with Dunnett’s multiple comparisons test, n=30 embryos. D) X. laevis embryos microinjected at the 4-cell stage with 30pg of importin α-mCH-HA or importin α-mCH-HA-CAAX targeted to the epidermis. After microinjection embryos were treated with either DMSO or 100 μM Wnt-C59 until NF stage 38 and immunostained for HA tag (green) and acetylated-α-tubulin (magenta). Cilia masks were created by ImageJ mean thresholding of the acetylated-α-tubulin channel for quantification in Figure 3E. Scale bar = 5μm. E) Quantification of the integrated density in arbitrary units for each MCC cilia mask. ****P<0.0001; ns, not significant, Student’s t-test, n=30 MCCs, 3 embryos, per condition. Schematic created with BioRender.com.

Importin α regulates cilia length in a palmitoylation and cytoplasmic dependent manner

Following ciliogenesis, the cilium maintains its structure through protein trafficking, ensuring the maintenance of its length and facilitating extracellular to intracellular signal transduction. Cilia length regulation is mediated by proteins trafficking into the cilium via the transition zone, enabling selective entry and exit of proteins in a bi-directional manner (Wang et al, 2021). Previous studies have demonstrated the presence of a RanGTP gradient, in conjunction with the localization of nuclear pore proteins within cilia (Dishinger et al, 2010; Fan et al, 2011; Chen et al, 2023). These findings imply that cilia may share fundamental characteristics with nuclear import, where protein trafficking is governed by the entry of importins through the selective nuclear pore protein complex and subsequent release upon interaction with RanGTP. Thus far our data has demonstrated that palmitoylated importin α regulates ciliogenesis and we hypothesize that importin α contains a dual role in cilia biology as a regulator of cilia length maintenance by regulating the steady-state dynamics of cilia length (Fig. 4A). To assess for changes in cilia length, RPE-1 cells were treated with the drug array to affect either palmitoylation status or importin function, once the cilia had already formed, and analyzed for cilia length as measured by the CiliaQ ImageJ plug-in (Hansen et al, 2021; Schneider et al, 2012) (Fig. S4A). Ciliated RPE-1 cells treated with Wnt-C59, inhibiting palmitoylation of importin α and thus increasing levels of cytoplasmic importin α, exhibited a decrease in cilia length (Fig. 4B,C). Surprisingly, ciliated RPE-1 cells treated with palmostatin, which inhibits the removal of membrane localized importin α and therefore decreases the pool of cytoplasmic importin α, also resulted in shortened cilia (Fig. 4B,C). Previous work from our group has characterized the shift of importin α from the membrane to cytoplasm upon Wnt-C59 treatment and the inverse upon palmostatin treatment (Sutton et al, 2025; Brownlee & Heald, 2019) and these results suggest that both enrichment of importin α at the membrane and depletion of importin α at the membrane can result in misregulation of cilia length. Thus, a balance between the levels of membrane-bound and cytoplasmic importin α must be maintained for cilia length control. These two populations of importin α may play distinct but equally critical roles in ciliary trafficking, both of which are regulated by importin α palmitoylation.

Fig. 4. Cilia length is regulated by the balance of palmitoylated and unpalmitoylated (cytoplasmic) importin α.

A) Schematic of hypothesized role for palmitoylated importin α and cytoplasmic importin α regulating cilia length. B) Ciliated RPE-1 cells treated for 2 hrs with drugs: DMSO, 10μM Wnt-C59, 50μM palmostatin, 40μM importazole, 25μM Ivermectin, or 50μM Cycloheximide, immunostained for acetylated-α-tubulin (magenta), ARL13B (gray) and DNA (blue). Scale bar = 5μm. C) Quantification of cilia length (μm) of ciliated RPE-1 cells treated for 2 hrs with drugs. ****P<0.0001; ns, not significant, one-way ANOVA with Dunnett’s multiple comparisons test, n=100 cells per treatment, 3 replicates. D) Quantification of relative ARL13B intensity within cilia of drug treated ciliated RPE-1 cells. *P<0.05; **P< 0.01; ns, not significant, one-way ANOVA with Dunnett’s multiple comparisons test, n= 50 cilia per treatment, 3 replicates. Mean ± SEM. E) X. laevis NF stage 32 epidermal MCCs were deciliated and allowed to reciliate over 4 hrs. X. laevis MCCs pretreatment, deciliated MCCs (t= 0 hours post deciliation, hpd) and regrown MCCs 4hpd for MCCs recovered in DMSO, 100μM Wnt-C59 or 1mM palmostatin, immunostained for acetylated-α-tubulin (gray). Scale bars = 5μm. F) Quantification of cilia length (μm) before deciliation and 4hpd. *P<0.05; **P< 0.01; ****P<0.0001; ns, not significant, one-way ANOVA with Dunnett’s multiple comparisons test, n= 90 cilia, n>2 replicates per condition. Schematic created with BioRender.com.

Next, we explored the significance of importin function in releasing and binding cargo proteins during steady state dynamics of cilia length maintenance. To achieve this, ciliated RPE-1 cells were treated with importazole, inhibiting release of cargo bound to importin α, which resulted in a reduction in cilia length (Fig. 4B,C ). To understand the requirement for importin α binding to cargo in regulating cilia length, ciliated RPE-1 cells were treated with the competitive inhibitor of importin α cargo binding, ivermectin (Yang et al, 2020), which also decreased cilia length, suggesting a role for importin α binding of cargo in regulating cilia length (Fig. 4B,C ). A potential confounding effect of inhibiting importin α-NLS binding/release is the abrogation of normal transcription and subsequent protein expression as many transcription factors rely on importin α-mediated nuclear transport for their activity (Oka & Yoneda, 2018). The disruption of protein expression during the experimental window may impact cilia length maintenance, therefore we utilized cycloheximide, an inhibitor of eukaryotic translation elongation (Schneider-Poetsch et al, 2010). By inhibiting protein expression by cycloheximide treatment, we can determine if our ciliary length defects are due to inhibiting importin α-mediated transport to the cilium or a side-effect of disrupting the expression of new ciliary proteins. Cycloheximide-treated ciliated RPE-1 cells did not significantly alter cilia length, indicating that cilia can maintain their normal length without new protein synthesis during the experimental period (Fig. 4B,C). Overall these data suggest that there is a prerequisite transport system of importin α in both palmitoylated and cytoplasmic forms to bind and release cargo proteins, facilitating their trafficking into the cilium to maintain proper cilia length.

Our data has demonstrated that functioning importin α in binding and releasing cargo is required for both cilia formation and cilia length maintenance. However, it remains unknown if importin α-mediated ciliary transport is required to shuttle essential ciliary proteins for their function in these processes. Our research group has previously performed a genome-wide screen for NLS-containing proteins and identified a subset involved in cilia biology (Sutton et al, 2025). Among this screen, we identified ARL13B as a potential NLS-containing ciliary cargo known to be involved in cilia length regulation. ARL13B is a small regulatory GTPase ciliary protein that localizes to the ciliary membrane and has been characterized to regulate cilia length (Larkins et al, 2011). When analyzed with the NLS prediction tool, NLStradamus, ARL13B received a high NLS probability score of 0.921 (out of 1), suggesting ARL13B as a potential candidate importin α binding cargo for regulation of cilia length (Nguyen Ba et al, 2009). Therefore we sought to investigate the localization of ARL13B within cilia in ciliated RPE-1 cells drug treated for the disruption of importin α function (binding/release of cargo) or palmitoylation status. Interestingly, the relative ciliary ARL13B intensity in palmostatin treated ciliated RPE-1 cells increased, whereas Wnt-C59 treatment resulted in no change in ARL13B intensity (Fig. 4B,D). This result suggests that inhibition of depalmitoylation of importin α, which has been shown to cause an increase in cytoplasmic importin α, leads to increased transport of ARL13B into the cilium. Next, ciliated RPE-1 cells were treated with importazole, which inhibits the release of NLS-bound cargo from importin α, which resulted in an increase in the amount of ciliary ARL13B, whereas inhibiting binding of NLS-cargo to importin α via ivermectin treatment led to a decreased amount of ciliary ARL13B (Fig. 4B,D). Importantly, cycloheximide-treated ciliated RPE-1 cells did not result in changes in cilia length nor levels of ARL13B in cilia in the experimental window, ruling out alterations of transcription and/or translation as the cause of decreased in cilia length (Fig. 4C,D). Taken together, these results suggest that importin α plays an important role in ciliary transport of ARL13B into the cilia. After establishing the role of importin α in regulating primary cilia length in RPE-1 cells, we tested its role in MDCK-II cells and observed changes in cilia length, however these trends were not similar to our RPE-1 data, suggesting importin α may have different functions of cilia length regulation dependent on cell type (Fig. S4B,C).

Importin α regulates cilia length in X. laevis epidermal MCCs

Next, we investigated whether importin α also regulates cilia length in X. laevis epidermal MCCs. Deciliation of X. laevis epidermal MCCs is followed by cilia regrowth over a 4-hour period, thus providing a window to assess the impact of importin α on cilia regrowth kinetics (Werner & Mitchell, 2013). Following deciliation, X. laevis embryos recovered in either palmostatin or Wnt-C59 to determine the effect of inhibiting importin α membrane localization during the cilia regeneration period (Fig. 4E). While cilia length varied over time between both treatments, at the final time point the cilia regrown in the Wnt-C59 recovery embryos were shorter compared to control and palmostatin recovery embryos (Fig. 4F). Additionally, while the cilia were shorter in the Wnt-C59 recovery embryos, they also contained fewer cilia within the MCC (Fig. 4E). Altogether, our data shows that importin α plays a role in cilia length regulation in both steady state dynamics and cilia regrowth.

Importin α binds to the NLS sequence of CEP164 to mediate ciliogenesis

The previously mentioned bioinformatics screen employed to identify putative NLS-containing ciliary proteins by Sutton et al 2025, also identified the centrosomal protein, CEP164 as likely to contain a bona-fide NLS sequence with a probability score of 0.999 (out of 1) (Nguyen Ba et al, 2009). Previous work on CEP164 has demonstrated its localization to the distal appendages of the mother centriole and its requirement for basal body docking, thereby facilitating cilia formation (Graser et al, 2007). However the exact mechanism by which CEP164 localizes to the basal body remains unknown. Given that CEP164 is predicted to contain a NLS sequence, we sought to determine if importin α could potentially bind CEP164 through its NLS sequence to mediate its localization for proper ciliogenesis. We observed co-localization of importin α and CEP164 in ciliated RPE-1 cells and at the base of X. laevis epidermal MCC (Fig. 5A,C,D). U-ExM of importin α co-localization with CEP164 had a similar pattern to that of CEP164 and the distal appendages (Fig. 5B). In support of this, the DuoLink proximity ligation assay (PLA) was employed to determine whether this co-localization occurs within a 40nm distance, indicative of a direct interaction. The PLA was performed on ciliated RPE-1 cells and revealed the binding of importin α and CEP164 at the cilium base (Fig. 5E,F). To confirm that the binding of importin α and CEP164 is mediated by its NLS sequence, we generated a mutant version of CEP164 by alanine substitution of the positively charged lysine residues found within the predicted NLS sequence (CEP164 ΔNLS GFP) (Fig. 5G).(Graser et al, 2007) To test whether importin α binding to CEP164 via its NLS sequence is required for ciliogenesis, RPE-1 cells were transfected with CEP164 WT GFP and CEP164 ΔNLS GFP plasmids, along with ARL13B-mCardinal for cilia visualization (Aljiboury ID et al, 2023; Graser et al, 2007). Overexpression of CEP164 ΔNLS GFP resulted in a decrease in ciliation, indicating a dominant-negative effect on ciliogenesis (Fig. 5H,I). In summary, our data elucidates a potential mechanistic role of importin α-mediated transport of ciliary cargo proteins, such as CEP164 and ARL13B for proper regulation of ciliogenesis and cilia length (Fig. 4,5).

Fig. 5. CEP164 ΔNLS cannot bind importin α and acts as a dominant negative mutant, inhibiting ciliogenesis.

A) Ciliated RPE-1 cell immunostained for acetylated-α-tubulin (magenta), KPNA2 (importin α) (green), CEP164 (cyan) and DNA (blue). Scale bars = 5μm (full image) and 1μm (inset images). B) U-ExM top-down view of a ciliated RPE-1 cell immunostained for KPNA2 (importin α) (green) and CEP164 (cyan). Scale bar = 1μm. C) Schematic of a deciliated X. laevis MCC, for a top-down view of importin α localization at base. D) X. laevis NF stage 32 deciliated epidermal MCC immunostained for Kpna2 (importin α) (green) and Cep164 (cyan). Scale bars = 5μm (full image) and 1μm (inset image). E) Proximity Ligation Assay (PLA) performed on ciliated RPE-1 cells, negative control with KPNA2 (importin α) and GAPDH (gray). Zoomed in image of lack of PLA spot at base of cilia (white arrow), acetylated-α-tubulin (magenta) and DNA (blue). F) PLA performed on RPE-1 cells for KPNA2 (importin α) and CEP164 (gray). Zoomed in image of PLA spot at cilium base (white arrow), acetylated-α-tubulin (magenta) and DNA (blue). Scale bars = 2μm (full image) and 1μm (inset image). G) Predicted nuclear localization signal (NLS) sequence within CEP164 spanning amino acid (a.a) residues 110-128. Mutated version of the CEP164 NLS (CEP164 ΔNLS GFP) by replacement the lysines (K) with alanines (A). H) RPE-1 cells transfected with 750ng of ARL13B-mCardinal and 250ng of either CEP164 WT GFP or CEP164 ΔNLS GFP immunostained for plasmid with GFP (green), ARL13B (magenta), and DNA (blue). After transfection, RPE-1 cells were serum starved to induce ciliogenesis. White arrow indicates CEP164 localization but lack of a primary cilium. Scale bar = 2μm. I) Quantification of % of ARL13B+ cilia and % of unciliated cells in transfected RPE-1 cells from Figure 5H. * P<0.05, Student’s t-test, n=100 cells per treatment, 4 replicates. Mean ± SEM. Schematic created with BioRender.com.

Disruption of palmitoylated importin α contributes to X. laevis kidney morphology defects

Overall, this study has demonstrated a role for palmitoylated importin α in regulating ciliogenesis and cilia length through its interaction with different NLS-containing cilia cargo proteins and membrane association. Building upon this, we investigated whether disrupting palmitoylated importin α in these ciliary processes could contribute to kidney disease in vivo. Ciliopathies, disorders arising from defects in cilia structure or function, can impact various organs and for this study we focused on the kidney, as primary cilia located in the renal tubules of the kidney are essential for proper kidney development and morphology (McConnachie et al, 2021). Defects in the formation or function of renal primary cilia can lead to chronic kidney diseases such as congenital anomalies of the kidney and urinary tract (CAKUT) (Shao et al, 2020; Scarinci et al, 2022; Kalot et al, 2023; Barakat & Butler, 2024). As we’ve demonstrated altered cilia length in MDCK-II cells (kidney epithelial cells) by disrupting importin α palmitoylation (Fig. S4), we sought to leverage the X. laevis model to explore the role of palmitoylated importin α in kidney development and its potential contributions to kidney disease.

X. laevis are an ideal model to study kidney development (nephrogenesis) because its embryos contain one nephron, the functional unit of the kidney on each side of its body, while human kidneys normally contain 200,000 to over 2.5 million nephrons (Bertram et al, 2011). The genetic networks that regulate mammalian nephrogenesis are analogous to the X. laevis pronephros, also known as the embryonic kidney (Blackburn & Miller, 2019; Vize et al, 1995) (Fig. 6A). The X. laevis embryonic kidney is comprised of tubular sections, such as the distal and proximal tubules which filter waste products and reabsorb nutrients from the blood (Vize et al, 1995). To elucidate potential contributions of palmitoylated importin α to X. laevis nephrogenesis, embryos were treated with Wnt-C59, inhibiting importin α palmitoylation. Given the potential confounding effects of Wnt-C59 treatment on the Wnt pathway, embryos were also treated with okadaic acid (OA), a PP1 and PP2A phosphatase inhibitor. Previous studies have demonstrated that OA can activate the Wnt pathway downstream and independent of Wnt secretion, thereby rescuing the effects of PORCN inhibition on Wnt signaling (Seidl et al, 2023; Zhang et al, 2023). Consequently, we hypothesized that treatment with Wnt-C59 and OA would inhibit importin α localization to the membrane via palmitoylation while preserving active Wnt signaling within the embryo. Upon treatment of X. laevis embryos with Wnt-C59, we observed significant kidney morphology defects, such as loss of tubular structures and development of enlarged regions, reminiscent of those observed in kidney development abnormalities such as CAKUT (Kalot et al, 2023; Barakat & Butler, 2024) (Fig. 6B,C). Treatment of embryos with OA alone did not produce these morphological defects, but treatment with Wnt-C59 + OA resulted in the same kidney morphology defects as Wnt-C59 treatment alone (Fig. 6B,C). These results demonstrate that palmitoylation of importin α specifically, and not the activation of the Wnt pathway is required for proper X. laevis nephrogenesis.

Fig. 6. Disruption of importin α palmitoylation causes X. laevis kidney morphology defects.

A) X. laevis NF stage 42 embryo immunostained for the proximal tubules (3G8) and distal tubules (4A6) (gray), comprising of the embryonic kidney, outlined in the white box. Scale bar = 200μm. Image below is a X. laevis NF stage 42 embryonic kidney (gray) with the proximal and distal tubule outlined in yellow and white dashed lines respectively. Scale bar = 100μm. B) X. laevis embryonic kidneys (gray) drug treated with DMSO, 100μM Wnt-C59, 10nM Okadaic Acid (OA) or 100μM Wnt-C59 + 10nM OA until embryos reached NF stage 42. Scale bar = 100μm. C) Quantification of % X. laevis embryonic kidneys with normal, mild or severe kidney morphology for all drug treatments. ****P<0.0001; ns, not significant, Fisher’s exact test, n>14 embryos per condition. D) X. laevis embryos microinjected at the 4-cell or 8-cell stage with 60pg of importin α-HA-mCH-CAAX targeted to the kidney. After microinjection, embryos were treated with either DMSO or 100 μM Wnt-C59 until NF stage 42, immunostained for embryonic kidney (gray) and HA tag (green). White arrow and dashed line point out HA localization in embryonic kidney. Scale bar = 20μm. E) Quantification of % X. laevis embryonic kidneys with normal, mild or severe kidney morphology for all drug treatments. *P<0.05; ns, not significant, Fisher’s exact test, n>6 embryos per condition. F) Embryonic kidney (gray) renal primary cilia immunostained for acetylated-α-tubulin (magenta). Scale bar = 10μm. White dashed lines outline of kidney tubule G) Quantification of # of cilia for 100μm of proximal tubule for all drug treatments. ****P<0.0001; ns, not significant, one-way ANOVA with Dunnett’s multiple comparisons test (DMSO comparisons), Student’s t-test (Wnt-C59 vs. Wnt-C59 +OA), n>14 embryos per condition. H) Quantification of # of cilia for 100μm of distal tubule for all drug treatments. **P<0.01; ***P<0.001; **** P<0.0001; ns, not significant, one-way ANOVA with Dunnett’s multiple comparisons test (DMSO comparisons), Student’s t-test (Wnt-C59 vs. Wnt-C59 +OA), n>15 embryos per condition.

To investigate if forced membrane localization of importin α via the CAAX motif is sufficient to rescue the kidney morphology defects observed with Wnt-C59 treatment, we microinjected embryos with importin α-WT or importin α-CAAX plasmids targeted to the blastomere fate-mapped to give rise to the embryonic kidney and treated injected embryos with DMSO or Wnt-C59 until NF stage 42. While neither importin α-WT nor importin α–CAAX expression alone altered kidney morphology (Fig. S5A-C), remarkably, we observed a rescue in kidney morphology in our importin α–CAAX + Wnt-C59 treatment embryo, supporting our hypothesis that membrane-bound importin α is required for proper nephrogenesis (Fig. 6D,E). Moreover, this result also demonstrates that forced membrane localization of importin α is sufficient to rescue kidney morphology independent from Wnt signaling. We next sought to investigate how renal primary cilia in the distal and proximal tubule segments of the X. laevis kidney are affected (Fig. 6F). In the proximal tubule region, we observed a lack of primary cilia rescue in our Wnt-C59 + OA treatment, suggesting that under these conditions, importin α palmitoylation is essential for proper primary cilia formation (Fig. 6G). Conversely, we did not observe this effect in the distal tubule region (Fig. 6H). Our data suggest while palmitoylated importin α may not be necessary in both segments of the X. laevis kidney, there is still a requirement for palmitoylated importin α in regulating renal primary ciliogenesis for proper nephrogenesis.

Discussion

Our study offers a previously unrecognized function for palmitoylated and non-palmitoylated importin α in regulating ciliogenesis and cilia length in both human and X. laevis renal primary cilia and epidermal multicilia. Previous research has established the localization of RanGTP, nuclear pore proteins, and importin β at the cilia and their essential role in ciliogenesis (Fan et al, 2011; Dishinger et al, 2010; Chen et al, 2023; Hurd et al, 2011). We have identified a pathway that integrates these findings by characterizing the localization of palmitoylated and non-palmitoylated importin α at the cilia base and lumen, respectively, as well as multiple NLS-containing cargos involved in ciliogenesis and cilia length maintenance (Fig. 7). Additionally, our results suggest that palmitoylated importin α is crucial during nephrogenesis and required for renal primary ciliogenesis in the tubule segments of the X. laevis embryonic kidney (Fig. 7).

Fig. 7. Importin α regulates ciliogenesis and cilia length maintenance in a palmitoylation dependent manner and is essential for proper kidney development.

Importin α possesses a novel function in the cell apart from its classic role in nuclear transport. Importin α regulates ciliogenesis and cilia length in both primary and multiciliated cells. In addition, palmitoylated importin α is required for X. laevis nephrogenesis. Schematic created with BioRender.com.

We have characterized the localization of importin α at ciliary structures (Fig. 1), however this localization raises the question of how and where palmitoylated importin α associates with the primary cilia membrane. Current research in the field has identified the basal body distal appendage protein, CEP164, as localizing to the centrioles in the initial steps of basal body docking prior to ciliogenesis (Zhao et al, 2022; Schmidt et al, 2012; Chou et al, 2025). Additionally, it has been shown that CEP164 can associate indirectly with the ciliary pocket, a hub for vesicular trafficking, by a CBY/ciBAR complex which creates the membrane curvature of the pocket (Hoque et al, 2024). Notably, cholesterol and lipids normally found in lipid raft microdomains have been demonstrated to be enriched at the ciliary pocket (Benmerah, 2013; Rohatgi & Snell, 2010). Given that palmitoylated proteins are known to associate with cholesterol-rich lipids rafts (Emmer et al, 2009, 2010), it is an intriguing possibility that palmitoylated importin α might localize to the ciliary pocket, where it could interact with CEP164 to facilitate basal body docking (Fig. 5). This hypothesis is further supported by our findings that CEP164 associates with importin α at the base of the primary cilium and that CEP164 and importin α binding may be mediated via CEP164’s NLS sequence (Fig. 5). This may also explain the reduction in cilia observed when CEP164 ΔNLS GFP is overexpressed in RPE-1 cells, by functioning as a dominant-negative mutant (Fig. 5). In this case, the intact C-terminus of CEP164 would still enable distal appendage binding, while the mutated NLS would prevent CEP164 binding of importin α near the ciliary pocket, hindering basal body docking and in turn ciliogenesis (Schmidt et al, 2012).

We have demonstrated that palmitoylated importin α is required for proper ciliogenesis by disrupting its palmitoylation status with pharmacological inhibitors and siRNA knockdown of PORCN. However, it remains unclear how importin α is initially associated with the nascent cilia (Fig. 2). During the beginning of the intracellular ciliogenesis pathway, Golgi-derived vesicles dock and fuse on the distal appendages of the mother centriole mediated through CEP164 and then dock at the membrane to promote cilium growth. Previous studies exploring this process have demonstrated that these Golgi-derived vesicles contain ciliary proteins that range from Rab GTPases to various centrosomal proteins are required to initiate ciliogenesis (Schmidt et al, 2012; Ganga et al, 2021; Zhao et al, 2022). In addition, previous work on the early stages of the alternative ciliogenesis pathway have linked shared ciliary proteins from the intracellular ciliogenesis pathway such as Rab GTPases and intraflagellar transport proteins known to congregate at the plasma membrane for cilia formation(Bernabé-Rubio et al, 2016; Labat-de-Hoz et al, 2021). Of particular interest, it has been shown that importin α can also localize to the Golgi to undergo its palmitoylation modification for subsequent membrane localization (Rohatgi & Snell, 2010; Wei et al, 2015; Chang et al, 2019). Our data raises the intriguing possibility that palmitoylated importin α may be required to facilitate localization of CEP164 to the basal body in the initial stages of the intracellular ciliogenesis pathway, in addition to its role in the alternative pathway involving basal body docking to the plasma membrane.

We also identified the cilia length regulating protein ARL13B as containing a NLS sequence and potentially a binding cargo of importin α. Additionally, we explored the necessity for a balance of membrane-bound importin α and cytoplasmic importin α, with the ability to bind and release NLS-bound cargo such as ARL13B, as a potential regulator of cilia length maintenance (Fig. 4). Interestingly, previous work has found that ARL13B not only regulates cilia length but is also an essential factor for protein trafficking and the execution of Sonic hedgehog (Shh) signaling (Larkins et al, 2011). Shh signaling in cilia has been studied extensively and is critical for the development and maintenance of various organs (Larkins et al, 2011; Bangs & Anderson, 2017). Further investigation promises to elucidate if ARL13B trafficking by importin α has the potential to impact Shh signaling and will reveal new connections between membrane-bound importin α in regulating downstream cilia signaling events. Given these diverse roles of palmitoylated importin α during ciliogenesis and cilia length control, there are several opportunities for future studies.

These results introduce a new pathway which highlights an unexplored area of research in the current field, linking palmitoylation of importins to ciliogenesis and cilia length maintenance. These findings have profound implications for our understanding of the importance of primary cilia during development.

Materials and Methods

Cell culture, drug treatments, and RNAi

hTERT-RPE-1 (RPE-1) and canine MDCK-II cells were obtained from ATCC (CRL-4000 and CRL-2936 respectively). RPE-1 cells were grown in DMEM/F12 media supplemented with 10% v/v fetal bovine serum (FBS; VWR #F31016-500) and 1% v/v antibiotic-antimycotic (Anti-Anti; ThermoFisher Scientific, #15240096). MDCK-II cel,ls were grown in EMEM media supplemented with 10% v/v FBS and 1% v/v Anti-Anti. Cell lines were grown at 37°C and 5% CO₂ in a humidified chamber. To induce ciliogenesis RPE-1 cells were serum starved for 48 hrs and MDCK-II cells were grown in complete media for 120 hrs. For ciliogenesis analysis, MDCK-II cells were cultured on transwell 0.4μm pore size inserts (ThermoFisher Scientific, #07-200-165). Cells were routinely checked for mycoplasma using the Mycoplasma Detection kit (Southern Biotechnology, #13100-01). Pharmacological drug treatments for palmitoylation and importin function: DMSO (Spectrum Chemical, #40470004), Wnt-C59 (Selleck Chem, #S7037), palmostatin-B (Sigma Aldrich, #178501-5MG), importazole (Selleck Chem, #S8446), ivermectin (ThermoFisher Scientific, #J6277.03) and cycloheximide (Sigma Aldrich, #C7698-1G) were added at concentrations listed in figure legends in DMEM F12 or EMEM serum free media 24 hrs prior to ciliogenesis for RPE-1 and MDCK-II cells respectively. RNAi experiments on RPE-1 cells were performed using the TriFECTa DsiRNA kit https://www.idtdna.com/pages/products/functional-genomics/dsirnas-and-trifecta-rnai-kits#product-details validated for >70% knockdown of target mRNA. The oligonucleotides for PORCN and a negative control (IDT hs.RI.PORCN.13.2 and IDT DS NCS1 respectively) were transfected at an amount of 30 pmol per 6 well plate using the Lipofectamine RNAiMAX transfection reagent 3000 (ThermoFisher Scientific, #13778030) in accordance with the supplier’s recommendations. RPE-1 cells were knockdown with siRNA PORCN for 24 hrs prior to serum starvation of 48 hrs to induce ciliogenesis. Samples were either fixed for immunostaining of prepared for cell lysis for immunoblotting as described below.

Production of Wnt-3A conditioned medium

Wnt-3A conditioned medium was collected from ATCC L Wnt-3A cells (CRL-2647) according to manufactures’ protocol. Briefly, L Wnt-3A cells were grown in DMEM until confluency (~4 days), after which the medium was collected with a 0.4μm pore sterile filter. Fresh DMEM was added to the cells to allow for another 3 days of growth, after which this medium was collected and sterile filtered. The first and second batch of collected medium was mixed 1:1 and stored at 4°C until use.

Ultra-expansion microscopy (U-ExM)

RPE-1 cells were grown on 12mm fibronectin coated coverslips using fibronectin bovine plasma (ThermoFisher Scientific, #33010018) at 5μg/mL. RPE-1 cells were serum starved for 48 hrs to induce ciliogenesis. Once ciliated, RPE-1 cells were fixed with 4% PFA in PBS for 2 hrs at room temperature and then proceeded to the U-ExM protocol as defined by Laporte et al., 2024. Primary antibodies used for U-ExM; mouse anti-acetylated-α-tubulin (1:250,Sigma-Aldrich, # T7451; RRID: AB_609894), rabbit anti-gamma-tubulin (1:250, Sigma-Aldrich, # T5192; RRID: AB_261690), mouse anti-KPNA2 (1:100, Proteintech, # 66870-1-Ig; RRID: AB_2882207), and rabbit anti-CEP164 (1:200, Proteintech, # 22227-1-AP; RRID:AB_2651175), Hoechst (1:1000, Cell Signaling Technology Cat# 4082, RRID:AB_10626776). To image expanded gels, poly-d-lysine (0.1mg/mL) coated slides were used for mounting and imaged on the EVOS M7000 epifluorescent microscope.

Animal models

All animals were maintained in accordance with standards established by the Division of Laboratory Animal Resources at Stony Brook University.

Xenopus laevis (X. laevis) in-vitro fertilization

X. laevis adult females were induced to lay eggs by a priming injection of 100 U pregnant mare serum gonadotropin (PMSG) at least 48 hrs before use and a boosting injection of 500 U of human chorionic gonadotropin (hCG) 16 hrs before use. Following the hCG injection, adult female X. laevis were placed in a 2L water bath of 1X MMR (100 mM NaCl, 2 mM KCl, 1 mM MgCl2, 2 mM CaCl2, 0.1 mM EDTA, 5 mM HEPES pH 7.8) overnight at 17°C. Approximately 16 hrs following hCG injection, fresh eggs were collected by squeezing eggs from ovulating frogs into a 10cm plastic petri dish. To fertilize eggs, a sperm solution made from ¼ of a male frog testis was placed in 1mL 1X MR (100 mM NaCl, 1.8 mM KCl, 2.0 mM CaCl2, 1.0 MgCl2, 5.0 mM HEPES-NaOH, pH 7.6) and homogenized using scissors and a pestle. 1mL of sperm solution was added dropwise to the freshly squeezed eggs and the dish was swirled to form a monolayer of eggs and incubated for 3 min. Dishes were flooded with milli-Q water and incubated for an additional 10 min. Eggs were dejellied with a 2% cysteine solution for 6 min with occasional swirling and washed 5 times with 1/3 MR. Fertilized eggs were incubated at 23°C until the appropriate developmental stage as described by Nieuwkoop and Faber (NF) 1994. X. laevis embryos treated with DMSO, 100μM Wnt-C59, 1mM Palmostatin or 10nM Okadaic Acid in fresh 1/3 MR and drug treated until the developmental stage indicated throughout the text and figures when applicable.

Immunofluorescence staining

Cell culture:

RPE-1 cells were grown on 12mm fibronectin coated coverslips. Polarized MDCK-II cells were grown on 12mm 0.4μm pore size transwell filters and once ciliated the membrane was cut out and fixed in 4% PFA for 10 min at RT. Following fixation, fixed cells were permeabilized with PBST (0.2% Triton X-100 in PBS, Sigma-Aldrich, #9036-19-5) 3 x 5 min and blocked for 1hr at RT in Bovine Serum Albumin (BSA) -PBST (3% BSA in PBST). Primary antibodies used were: mouse anti-acetylated-α-tubulin (1:1000, Sigma-Aldrich, #T7451; RRID: AB_609894), rabbit anti-ARL13B (1:250, Proteintech, # 17711-1-AP; RRID:AB_2060867), mouse anti-KPNA2 (1:250, Proteintech, #66870-1-Ig; RRID: AB_2882207), rabbit anti-CBY (1:500, Gift from Dr. Ken-Ichi Takemaru), mouse anti-HA tag (1:250, Sigma-Aldrich, #SAB2702196; RRID:AB_2893412), rabbit anti-GFP tag (1:250, Proteintech, #50430-2-AP; RRID:AB_11042881), and rabbit anti-CEP164 (1:250, Proteintech, #22227-1-AP; RRID:AB_2651175). Following primary antibody incubation for 1 hr at RT, coverslips were washed 3 x 5min with PBST and then incubated for 30 min with secondary antibodies. Secondary antibodies used were: Donkey anti-mouse and anti-rabbit IgG-AF488 (Southern Biotech, #6411-30 and #6440-30; RRID: AB_2796314; RRID: AB_2796368, respectively), Donkey anti-mouse and anti-rabbit IgG-AF568 (ThermoFisher Scientific, #A10037, #A10042; RRID: AB_11180865; RRID: AB_2534017, respectively) and Donkey anti-mouse IgG-AF633 (ThermoFisher Scientific, #A-21050; RRID:AB_2535718) and all used at 1:1000 dilution. For Fig. 1A,1B, 5A, following incubation with secondary antibodies, samples were incubated with CoraLite^® Plus 488-conjugated Acetyl-Tubulin (Lys40) Monoclonal antibody (1:200, Proteintech, #CL488-66200; RRID:AB_2883282) for 2 hrs. Finally, coverslips were washed 3 x 5 min with PBST and mounted with ProLong Diamond Antifade Mountant (ThermoFisher Scientific, #P36961) and imaged on EVOS M7000 epifluorescent microscope.

X. laevis:

Whole X. laevis embryos at the appropriate stages indicated in figure legends were fixed in 4% PFA in PBS for 2 hrs at RT or overnight at 4°C, followed by 3 x 5 min with PBST (0.2% Triton X-100 in PBS). After permeabilization embryos were blocked in BSA-PBST (3% BSA in 0.2% Triton X-100 in PBS) for 1 hr at RT followed by overnight primary antibody incubation at 4°C with rotation. Primary antibodies used were: rabbit anti-acetylated-α-tubulin (1:500, Cell Signaling, #3971; RRID:AB_2210204), mouse anti-KPNA2 (1:250, Proteintech, #66870-1-Ig; RRID: AB_2882207), mouse anti-HA tag (1:250, Sigma-Aldrich, #SAB2702196; RRID:AB_2893412), rabbit anti-CEP164 (1:250, Proteintech, # 22227-1-AP; RRID:AB_2651175), Phalloidin (1:100, ThermoFisher Scientific, #A30107), mouse proximal kidney tubules 3G8 (1:30; EXRC, # 3G8.2X11; RRID: AB_10013600) and distal kidney tubules 4A6 (1:5, EXRC, #4A6.2C10; RRID: AB_10013600). Primary antibodies were washed out 3 x 1 hr with PBST and embryos were incubated overnight at 4°C with secondary antibodies in PBST. Secondary antibodies used were: Donkey anti-mouse and anti-rabbit IgG-AF488 (Southern Biotech, #6411-30 and #6440-30; RRID: AB_2796314; RRID: AB_2796368, respectively), Donkey anti-mouse and anti-rabbit IgG-AF568 (ThermoFisher Scientific, # A10037 and #A10042; RRID: AB_11180865; RRID: AB_2534017, respectively) and Donkey anti-mouse IgG-AF633 (ThermoFisher Scientific, # A-21050; RRID:AB_2535718) and all used at 1:500. Secondary antibodies were washes out 3 x 1 hr and mounted with ProLong Diamond antifade mountant (ThermoFisher Scientific, # P36961) and imaged on EVOS M7000 or Zeiss 980 confocal microscope.

Ciliation quantification

For quantification of changes in ciliogenesis, or percentage of ciliated cells was performed by counting the number of ciliated cells vs. unciliated cells using the point tool in ImageJ (v2.0.0). Quantification was plotted with GraphPad Prism 10 and number of samples are indicated in figure legends. For siRNA PORCN treated RPE-1 cells, the stubby cilia phenotype was determined by a ciliated cell possessing a cilium that appeared punctate.

X. laevis deciliation and motile cilia isolation

Approximately 70-100 X. laevis NF stage 30 embryos were subjected to the deciliation procedure as described previously by Werner and Mitchell, 2013. Briefly X.laevis embryos were deciliated with 75mM CaCl₂ and 0.02% NP-40 in 1/3 MR for 2 min or until cilia movement stopped. Deciliated embryos were then fixed immediately in 4% PFA and followed immunostaining protocol above, or deciliated embryos recovered in a drug bath of either DMSO, 100μM Wnt-C59 or 1mM palmostatin in 1/3 MR, with subsequent fixation and immunostaining after each hour for a total of 4 hrs post deciliation for cilia regrowth. Motile cilia were isolated and prepared for immunoblotting as previously described by Seidl et al., 2023.

Quantification of reversible deciliation:

Quantification of cilia regrowth after deciliation was performed using the Celleste Image Analysis Software (ThermoFisher Scientific, # AMEP4942). MCCs in each image were first selected as region of interest (ROI) using the Celleste ‘select’ tool. Next within each MCC/ROI, 3 clear individual cilia (from cilia base to tip) were traced using the direct measurement polyline tool. Cilia length was measured in microns (μm) and plotted with GraphPad Prism 10. Number of samples are indicated in figure legends.

X. laevis immunoblotting

Isolated cilia and whole X.laevis embryo were collected as described above and mixed 1:1 with 2X Laemmli buffer, boiled for 10 min at 70°C. and 5 min at 95°C. respectively. Samples were loaded for SDS-PAGE in a 7.5% Tris-glycine gel and then transferred on to a nitrocellulose membrane. The membrane was then washed 3 x 5 min with PBST (0.1% Tween-20 and 0.01% SDS in 1x PBS) and incubated with 5% milk blocking buffer for 1 hr at RT, followed by overnight primary antibody incubation at 4°C. Primary antibodies used were: rabbit anti-acetylated-α-tubulin (Cell Signaling, #3971; RRID:AB_2210204), mouse and rabbit anti-KPNA2 (Proteintech, #66870-1-Ig and #10819-1-AP; RRID: AB_2882207; RRID: AB_2265526, respectively), mouse and rabbit anti-GAPDH (Proteintech, # 60004-1-Ig and #10494-1-A; RRID:AB_2107436; RRID:AB_2263076, respectively), and mouse anti-β-tubulin (DSHB, # E7; RRID:AB_528499) in blocking buffer all diluted 1:1000. Next the membrane was washed 3 x 5 min with PBST (0.1% Tween-20 and 0.01% SDS in 1x PBS) and incubated with fluorescent secondary antibodies used at 1:10,000: IR-800 (Advansta, #R-05061-250; RRID:AB_3676152) and IgG-AF700 (ThermoFisher Scientific, #A-21038; RRID: AB_2535709) for 1 hr at RT, washed 3 x 5 min with PBST and then scanned with an Odyssey Infrared Imaging System (LI-COR Biosciences).

X. laevis plasmid and morpholino microinjections

For epidermal MCC experiments, 30pg of importin α-mCherry(mCH)-HA or importin α-mCH-HA-CAAX, along with 5ng/nL of dextran as a tracer were injected targeted for epidermis at the 4-cell stage. For embryonic kidney experiments, 60pg of importin α-mCH-HA or importin α-mCH-HA-CAAX, along with 5ng/nL of dextran as a tracer at either the 4- or 8-cell stage targeted for the kidney. For morpholino experiments, PORCN morpholino was purchased from Gene Tools and validated in Chen et al, 2015 (NM_001087954; GeneTools). X. laevis embryos were injected with 40ng of PORCN morpholino into one blastomere at the 2-cell stage, with the uninjected blastomere serving as an internal control. Morpholino sequence is: PORCN, (5′ CTCCCATCATCTCTCGGAATTC 3′). For PORCN morpholino injections, the injected half of the embryo was determined by observations in changes in pigment on the epithelium as the morpholino was target to the epidermis. All microinjected embryos were placed in 2.5% ficoll in 1/3 MR with gentamycin for 4 hrs, after which were then moved into a new dish of 1/3 MR and incubated at 23°C. Drug treated and microinjected X. laevis embryos were treated with DMSO, 100μM Wnt-C59 (Selleck Chem, #S7037), or 10nM Okadaic Acid (Cell Signaling, #5934S) at NF stage 18-19 and grown until the appropriate developmental stage indicated in figure legends for fixation and staining.

X. laevis epidermal MCCs quantification

MCCs/Area:

Quantification of the number of epidermal MCCs present on drug treated (DMSO, Wnt-C59 and palmostatin) X. laevis embryos was determined by an image taken with a 20x objective with an area of approximately 500 μm x 380 μm. Within this field the number of MCCs were counted with the point tool in Image J (v2.0.0). Number of samples are indicated in figure legends.

MCC cilia mask:

For quantification of the integrated fluorescence density of the cilia within a MCC, the acetylated-α-tubulin channel was used for threshold mean segmentation in ImageJ (v2.0.0). The segmented cilia were selected for measuring the integrated density in arbitrary units and plotted with GraphPad Prism 10. Number of samples are indicated in figure legends.

X. laevis kidney morphology and cilia quantification

Kidney morphology:

For quantification of X. laevis kidney morphology, the number of turns in the embryonic kidney, proximal and distal tubules, were counted with the point tool in ImageJ (v2.0.0). To categorize the number of turns into normal, mild or severe morphology, we established a threshold limit in which a kidney with 8 or more tubule turns are classified as normal, kidneys with less than 8 but greater than 4 turns were classified as mild and kidneys with less than 4 tubule turns were classified as severe. Morphology quantification was plotted with GraphPad Prism 10 and number of samples are indicated in figure legends.

Proximal and distal tubule cilia:

For quantification of the number of cilia in the proximal and distal tubule regions, 100μm region of interest (ROIs) were made using ImageJ (v2.0.0). Within each ROI, the cilia were counted using the ImageJ point tool. Cilia quantification in the proximal and distal tubule were plotted with GraphPad Prism 10 and number of samples are indicated in figure legends.

Plasmid construct development

A mutated version of the CEP164 NLS, spanning amino acid resides 110-128, was generated by VectorBuilder. Mutation of the NLS was performed by replacing the lysine residues in the NLS to alanines, to preserve protein structure folding dynamics. This construct has the vector identifier #VB240314-1449abt.

Plasmid Transfection

RPE-1 cells were transfected via nucleofection LONZA P3 cell line 4D-Nucleofector kit (#V4XP-3032) following recommended protocol. Amount of DNA for transfections are listed in figure legends when applicable. Following nucleofection, RPE-1 cells were incubated in serum free media to induce ciliogenesis for a total of 48 hrs. During this experimental time, 24 hrs prior to ciliogenesis, RPE-1 cells were treated with either DMSO or 10μM Wnt-C59 (Selleck Chem, #S7037). After which, RPE-1 cells were fixed and immunostained as described above. pCS2-Arl13b-mCardinal plasmid was acquired from Addgene (#194542) and pRcCMV-Cep164-GFP plasmid was acquired from Addgene (#41148). Plasmids, pUBC-hKPNA2-HA-mCherry and pUBC-hKPNA2-HA-mCherry-CAAX were used from Sutton et al, 2025.

Membrane/Cytoplasm ratio quantification

Quantification of the membrane/cytoplasm ratio of transfected RPE-1 cells was performed in ImageJ (v2.0.0) by using a fixed line of 50 pixels in length and 10 pixels in width, to measure the mean fluorescence intensity of each channel (GFP/HA tag and KPNA2) with one line at the membrane and cytoplasm for each cell measured. Mean fluorescence intensity measurements were plotted with GraphPad Prism 10 and number of samples are indicated in figure legends.

Cilia length quantification

RPE-1 cilia length was measured using the CiliaQ ImageJ plugins, CiliaQ Preparator (v0.1.2), CiliaQ Editor (v0.03) and CiliaQ (v0.1.7). The CiliaQ plugin was used as previously described by Hansen et al., 2021. X. laevis re-ciliation lengths were quantified in ImageJ (v2.0.0) by drawing a line tool tracing 3 distinct and clear cilia (from base to tip) for 30 MCCs for each sample.

ARL13B relative intensity quantification

Relative intensity of ARL13B in cilia was measured in ciliated RPE-1 cells treated with the previously described drug array for 2 hours post cilia formation and immunostained for ARL13B and acetylated-α-tubulin (to label cilia). ARL13B intensity was determined in ImageJ (v2.0.0) by measuring the integrated density of ARL13B signal in a 5 pixel width line along the length of each cilium and dividing this measure by the integrated density of acetylated-α-tubulin for the same cilia. This series of measurements was done for 50 ciliated cells per drug treatment in triplicate, totaling 150 ciliated cells per condition.

Proximity Ligation Assay (PLA)

PLA was performed using Duolink PLA fluorescence protocol (Sigma-Aldrich, #DUO92101). RPE-1 cells were serum starved to induce ciliogenesis and fixed with 4% PFA in 1x PBS for 10 min. After fixation, samples were permeabilized with PBST for 15 min followed by blocking with the Duolink blocking solution for 1 hr at 37°C. Next, samples were incubated with primary antibodies diluted in the Duolink Antibody Diluent for 1 hr at 37°C and washed twice with Wash buffer A for 5 min each. Samples were then incubated with the Duolink Antibody Diluent containing mouse PLA^minus and rabbit PLA^plus probes for 1 hr at 37°C followed by the ligation reaction for 30 min at 37°C. Next samples were incubated with the amplification reaction for 100 min at 37°C with the Duolink Fluorescent Detection Reagent Red. For cilia labeling, samples were then incubated for 2 hrs at 37°C with CoraLite^® 488-conjugated Acetyl-tubulin (Proteintech listed above) diluted in Duolink Antibody Diluent. Lastly, samples were washed with Wash buffer B and then mounted with the Duolink In Situ Mounting Media with DAPI (Sigma-Aldrich, #DUO92101) for imaging on EVOS M7000 epifluorescent microscope with an Olympus Super Apo 100x, 1.4 NA oil objective.

Microscopy

Images were acquired on an EVOS M7000 epifluorescent microscope or a Zeiss LSM 980 confocal microscope, as written in the figure legends. Images acquired on the EVOS M7000 were processed using the Celleste Imaging System 2D/3D Deconvolution at default settings. Images acquired on the Zeiss LSM 980 were acquired in either confocal mode or with Airyscan processing on Zen.

Cell culture:

Importin α ciliary localization (Fig. 1) and cilia length regulation (Fig. 4) in RPE-1 cells were imaged with an Olympus Semi-Apo 60x 0.7 numerical aperture (NA) oil objective. U-ExM RPE-1 cell (Fig. 1) was imaged on a Zeiss 980 confocal microscope with a Plan Apo 63x, 1.4 NA oil objective. Drug treated and/or transfected RPE-1 cells (Fig. 2,3) were imaged with an Olympus Semi-Apo 60x, 0.7 NA oil objective. Localization of importin α and CEP164 in RPE-1 cells (Fig. 5) and transfection of RPE-1 cells with CEP164 WT GFP or CEP164 ΔNLS GFP (Fig. 5) were imaged on EVOS M7000 epifluorescent microscope with an Olympus Super Apo 100x, 1.4 NA oil objective.

X. laevis:

Importin α ciliary localization in epidermal MCCs (Fig. 1) were imaged on EVOS M7000 epifluorescent microscope with an Olympus Super Apo 100x, 1.4 NA oil objective. Drug treated X. laevis NF stage 37 embryos (Fig. 3), X. laevis embryonic kidney images (Fig. 6A,B and Fig. S5), and PORCN MO X. laevis embryos (Fig. S2) were imaged on EVOS M7000 epifluorescent microscope with an Olympus Semi-Apo 4x, 0.13 NA objective. Microinjected epidermal MCCS (Fig. 3) and PORCN MO epidermal MCCs (Fig. S3) were imaged on EVOS M7000 epifluorescent microscope with an Olympus Semi-Apo 60x, 0.7 NA oil objective. Reversible deciliation epidermal MCCs (Fig. 4) were imaged on EVOS M7000 epifluorescent microscope with an Olympus UPLFLN 40x, 0.51 NA air objective. Importin α and CEP164 localization at the epidermal MCC base was imaged on a Zeiss 980 confocal microscope with a Plan Apo 100x (Fig. 5), 1.4 NA oil objective. X. laevis embryonic kidney, proximal and distal tubule were imaged on EVOS M7000 epifluorescent microscope with an Olympus Semi-Apo 60x, 0.7 NA oil objective (Fig. 6).

QUANTIFICATION AND STATISTICAL ANALYSIS

All statistical analyses were performed in GraphPad Prism 10.0. Comparisons between datasets were determined by one-way ANOVA with Dunnett’s post-hoc multiple comparisons test unless otherwise stated. Graphs represent the mean value ± standard error of the mean (SEM) unless otherwise stated. *P<0.05, **P <0.01, ***P<0.001, and ****P<0.0001 unless otherwise stated. Blinding was not performed unless otherwise stated.