Mutations in the β-tubulin TUBB impair ciliogenesis and are associated with ciliopathy-like phenotypes

Mollica, Antonio; Omer, Safia; Forguson, Georgiana; Steiman, Sydney; Evagelou, Sonia L.; Naumenko, Serhiy; Walker, Susan; Li, Lu Yi; Teeling, Aideen; Lindsay, Kyle; Erwood, Steven; Visuvanathan, Shagana; Vig, Anjali; Vernon, Robert M.; Akman, Benjamin; Smith-Hicks, Constance; Forman-Kay, Julie D.; Shroff, Manohar; Pai, Vivek; Harrison, Rene E.; Cohn, Ronald D.; Ivakine, Evgueni A.

doi:10.1038/s41467-025-65634-x

Download PDF

Article
Open access
Published: 27 November 2025

Mutations in the β-tubulin TUBB impair ciliogenesis and are associated with ciliopathy-like phenotypes

Nature Communications volume 16, Article number: 10637 (2025) Cite this article

6287 Accesses
Metrics details

Subjects

Abstract

Tubulinopathies and neurodevelopmental ciliopathies are two groups of genetic disorders that cause structural brain malformations. Tubulinopathies result from mutations in tubulins, the building blocks of microtubules, most of which are dominant. Neurodevelopmental ciliopathies are mostly recessive disorders caused by defects in the primary cilium, an organelle that regulates key signaling pathways during brain development. Although more than 40 genes have been associated with neurodevelopmental ciliopathies, many patients still lack a known genetic cause. Here, we present a de novo heterozygous missense variant (p.G308S) in Tubulin Beta Class I (TUBB) identified in a patient showing features of both ciliopathy and tubulinopathy. While microtubules are essential for primary cilia formation and function, an association between tubulin mutations and neurodevelopmental ciliopathies has not been established. Using patient-derived cells and gene-edited isogenic cell lines, we show that the identified variant impairs the early stages of cilia formation by altering microtubule dynamics and structure. Consistent with this, mice carrying the variant display decreased ciliation in the cerebellum and kidney. Furthermore, we demonstrate that the disease mechanism is not haploinsufficiency and that other patient mutations in TUBB affect cilia formation, putting forward defective ciliogenesis as a contributing pathogenic factor in a subset of tubulinopathy patients.

Bridging the gap: an emerging link between tubulinopathies and ciliopathies

Article Open access 06 May 2026

Expanding genetic and clinical spectra of β-tubulinopathies: A Korean study

Article 29 October 2025

Primary cilia as dynamic and diverse signalling hubs in development and disease

Article 18 April 2023

Introduction

Tubulinopathies are a group of disorders caused by mutations in genes encoding for tubulins^1,2,3,4. Tubulins are the building blocks of microtubules (MTs), one of the major components of the eukaryotic cytoskeleton. MTs are polymers of α- and β-tubulin heterodimers arranged head-to-tail to form protofilaments, which interact laterally to create the hollow cylindrical structure of MTs⁵. In humans, there are nine α- and ten β-tubulin isotypes, each encoded by a separate gene. Despite their high homology, there is increasing experimental and clinical evidence that the different tubulin isotypes are functionally distinct and not fully interchangeable^1,6,7. The nine α- and ten β-tubulins are differentially expressed in different tissues at different times, generating a diverse repertoire of αβ dimers that confer distinct properties to the resulting MTs⁸. Due to the importance of MTs in cellular functions, life-compatible tubulin mutations found in patients typically have subtle effects on MT properties and functions and do not severely compromise the overall functionality of MTs². To date, the vast majority of known pathogenic variants in tubulin genes are heterozygous missense mutations, often arising de novo in the affected individuals. MTs are known to play a role in key neurodevelopmental processes such as neurogenesis, neuronal migration, axon formation, and synaptogenesis^1,3. Of the over twenty tubulin genes encoded by the human genome, six have been associated with neurodevelopmental defects: one encoding α-tubulin (TUBA1A), four encoding β-tubulins (TUBB, TUBB2A, TUBB2B, TUBB3), and the γ-tubulin gene TUBG1². Additionally, variants in specific tubulin isotypes have also been shown to cause phenotypes unrelated to brain development, such as neurodegenerative diseases (TUBA4A and TUBB4A)^9,10,11 or diseases outside the nervous system (TUBB1 and TUBB8)^12,13. Neurodevelopmental tubulinopathies are characterized by an array of complex malformations visible on brain imaging, typically affecting the cortex and other brain structures such as the basal ganglia, cerebellum, and brainstem^2,3. Neurodevelopmental tubulinopathies caused by mutations in different tubulin genes generally present with characteristic features, even though there is extensive phenotypic overlap^1,2,3. However, the molecular mechanisms underlying each structural malformation have not been fully elucidated.

Neurodevelopmental ciliopathies comprise Joubert syndrome (JS) and Meckel syndrome (MKS)^14,15. Traditionally considered as two separate disorders due to their distinct phenotypes, JS and MKS significantly overlap in their causative genes and have recently been proposed to represent the mild (JS) and severe (MKS) ends of the same pathological spectrum¹⁶. JS and MKS belong to a group of diseases called ciliopathies because the pathogenic mutations affect the structure or the function of the primary cilium, a single-copy organelle present in most mammalian cells, including neurons and neuronal progenitors^16,17. Emerging as a hair-like protrusion of the plasma membrane, the primary cilium is a sensory organelle enriched in receptors and signaling factors¹⁸. Despite being continuous with the plasma membrane, it has a distinct protein and lipid composition that is maintained by a diffusion barrier at the ciliary base and by active transport mechanisms in and out of the cilium¹⁹. More than just an accumulation site for signaling molecules, the primary cilium plays an active role in modulating signal transduction by concentrating or excluding specific factors in a regulated fashion. In particular, the primary cilium regulates numerous signaling pathways that are important for embryogenesis, brain development, and adult tissue homeostasis, such as Hedgehog and Wnt^18,20. Consistent with the broad role of the primary cilium, most ciliopathies can present in some patients with symptoms in various organ systems, including the brain^14,15. However, in the case of JS and MKS, brain malformations are defining and diagnostic features. JS originates from defective embryogenesis of midbrain and hindbrain structures, and affects approximately 1 in 100000 live births²¹. It is almost exclusively recessive, but an autosomal dominant mild form of JS has recently been described^16,22,23. The distinctive diagnostic feature of JS is a cerebellar and brainstem malformation visible on brain imaging, referred to as the molar tooth sign (MTS). The MTS consists of cerebellar vermis hypoplasia or dysplasia, thick horizontal and elongated superior cerebellar peduncles, and a deep interpeduncular fossa^16,21. At the cellular level, this characteristic set of brain malformations has been linked to defects in neurogenesis, neuronal migration, and axon guidance^16,20. Clinically, JS typically presents with hypotonia evolving to ataxia, developmental delay, oculomotor apraxia, and breathing abnormalities. In addition to the hallmark neurological symptoms, patients with JS may suffer from an array of symptoms in other organ systems shared with other ciliopathies such as cystic kidney, liver fibrosis, and retinal defects^16,21. MKS is a recessive condition that affects approximately 1 in 135000 live births²⁴. It is the most severe ciliopathy, causing pre- or perinatal lethality due to developmental defects in multiple organs. MKS is associated with a highly variable phenotype, but the most common and distinctive features are occipital encephalocele, enlarged cystic kidneys, and liver fibrosis. Neurodevelopmental defects such as occipital encephalocele or other less common brain malformations are obligate features for MKS, which may also present with additional defects such as postaxial polydactyly, coloboma, cleft lip and palate, congenital heart defects, ambiguous genitalia, and skeletal abnormalities^16,24. Despite having greater than 40 JS- and 15 MKS-associated genes identified thus far, the exact genetic basis of disease for many patients suffering from neurodevelopmental ciliopathies remains undetermined^16,25.

In cell biology, the link between MTs and primary cilia is well established. Nine circularly arranged bundles of two MTs (doublets) make up the axoneme of primary cilia, a specialized structure that supports the cilium and allows for selective transportation of signaling factors in and out of the cilium¹⁹. Moreover, MTs are essential for the biogenesis of the primary cilium, which is assembled via the ciliogenesis pathway in the G1 phase of the cell cycle or in G0 following cell cycle exit²⁶. In particular, neurons and neuronal progenitors rely on the intracellular pathway of ciliogenesis^27,28,29,30. In intracellular ciliogenesis, endosomal and Golgi pre-ciliary vesicles (PCVs) are recruited to the mother centriole via both dynein- and myosin-mediated transport on MTs and actin filaments, respectively^31,32,33. PCVs then fuse to form the ciliary membrane, initiating the growth of the ciliary axoneme while the centrosome/forming cilium complex moves apically towards the plasma membrane^33,34,35. This movement is primarily mediated by pushing forces from dense MT bundles assembling against the cell cortex³⁵. Despite this tight connection, an association between mutations in tubulin genes and neurodevelopmental ciliopathies is lacking, and tubulinopathies and neurodevelopmental ciliopathies are considered as two independent groups of disorders. Additionally, the molecular mechanisms underlying tubulinopathies are poorly understood, and many patients with neurodevelopmental ciliopathies lack a genetic etiology.

In this work we investigate the potential role of primary cilia defects in the pathogenesis of tubulinopathies. Starting from a patient manifesting clinical features of both JS and tubulinopathy, we show that a subset of dominant brain development mutations in the β-tubulin gene TUBB on chromosome 6p21.33 affects primary cilia formation and is associated with ciliopathy-like phenotypes, potentially revealing an emerging disease mechanism for TUBB-related tubulinopathies and implicating TUBB as a candidate ciliopathy gene.

Results

Patient with a neurodevelopmental disorder shows atypical, ciliopathy- and tubulinopathy-like brain MRI features

A male patient, the sole in his pedigree, presented from an early age with neurodevelopmental delay and various neurological symptoms. The patient showed mild intellectual disability, expressive language delay, axial hypotonia and ataxia, and oculomotor apraxia. By early adulthood, his cognitive abilities were assessed to be within the average range, and his expressive language improved. The patient displayed only neurological symptoms without involvement of other organs or organ systems.

Magnetic resonance imaging (MRI) of the patient’s brain (Fig. 1a, compared to MRI from a healthy control in Fig. S1a) revealed features compatible with JS, i.e., superior vermis hypoplasia, mildly long, thick, and horizontal superior cerebellar peduncles, and elevated roof of the 4^th ventricle, which show up as the classic MTS on axial images (Fig. 1a–ii). In addition, the MRI showed other abnormalities such as dysplastic somewhat asymmetric appearance of the superior cerebellar hemispheres (Fig. 1a–iii) and asymmetric, somewhat bulbous and abnormal caudate nucleus on the right (Fig. 1a–iv and 1a–v), which are soft features of tubulinopathy³. Other findings were also present, such as hypoplastic olfactory bulbs (Fig. 1a–vi) and incomplete rotation of the left hippocampus (Fig. 1a–vii).

Fig. 1: A de novo heterozygous mutation in TUBB was identified in a patient with atypical brain imaging features. — **Fig. 1: A de novo heterozygous mutation in *TUBB* was identified in a patient with atypical brain imaging features.**

Overall, the patient described here represents an atypical case of neurodevelopmental disorder showing both ciliopathy- and tubulinopathy-like brain MRI features.

Whole genome sequencing analysis identified a presumed pathogenic variant in Tubulin Beta Class I

To identify the genetic cause of disease in the patient, we used whole genome sequencing (WGS) to analyze coding and non-coding germline variants segregating in the 5 members of the patient’s family: the patient, the patient’s parents, and the patient’s two unaffected siblings. Rare variants were assessed for pathogenicity both against the “rare multisystem ciliopathy disorders” panel (https://panelapp.genomicsengland.co.uk/panels/150) and in a gene panel-independent fashion, considering both dominant and recessive inheritance models for segregation analysis. Overall, our analysis predicted a de novo heterozygous missense variant (NM_178014.3: c.922 G > A/p.G308S) in Tubulin Beta Class I (TUBB) as the primary pathogenic candidate, since it was the only variant meeting defined pathogenicity criteria and segregating with the disease phenotype in the family (described in Methods). The TUBB gene is associated with two disorders with an autosomal dominant mode of inheritance: a brain development condition characterized by microcephaly and structural brain abnormalities (cortical dysplasia, complex, with other brain malformations 6 - OMIM #615771)³⁶, and a phenotypically complex disorder neurologically characterized by intellectual disability without gross brain malformations (circumferential skin creases Kunze type, CSC-KT - OMIM #156610)³⁷. Moreover, an association with congenital thrombocytopenia has recently been reported³⁸. As with disease-causing variants in the other tubulin genes, pathogenic mutations in TUBB are often de novo^3,36,37. Of note, TUBB is reported to be ubiquitously expressed in adult tissues and one of the most highly expressed β-tubulin genes in the developing brain^36,39.

Sequence alignment analysis showed that the G308 residue is conserved in all ten human β-tubulin isotypes and in all TUBB orthologs from yeast to primates (Fig. 1b). Mapping G308 on the 3D structure of the human αβ tubulin dimer and superimposing the dimer onto the structure of a fully assembled MT from Sus scrofa showed that G308 is exposed on the dimer surface and is not an interface residue (Fig. 1c). Positionally conserved in β-tubulin structures in different species, G308 lies in a sharp turn preceding β-sheet 8⁴⁰ and is followed by an arginine residue (R309) that is similarly conserved in all human β-tubulins except for TUBB8B and in all TUBB orthologs from yeast to primates (Fig. 1b, d). The positive side chain of R309 occupies a structurally important position and is involved in charge interactions with the negative dipole of two α-helices (10 and 12)⁴⁰ and the side chain of a glutamic acid residue (E343) (Fig. 1d). Considering the constrained position occupied by G308 and the unique flexibility of glycine residues, it is conceivable that G308 is crucial to hold R309 in the optimal position for its interactions.

The TUBB G308S variant affects MT assembly and alters MT structure

To determine the functional impact of the patient’s TUBB G308S mutation on αβ tubulin dimer formation and incorporation into MTs, we performed a MT incorporation assay where hTERT-RPE1 cells were transiently transfected with plasmids encoding either TUBB-V5 WT or TUBB-V5 G308S, and immunostained for exogenous TUBB and endogenous α-tubulin (Fig. 2a). In most transfected cells TUBB-V5 WT colocalized with α-tubulin and had a filamentous appearance, indicating stable incorporation into MTs (Fig. 2a–i, b). Conversely, TUBB-V5 G308S had a diffuse appearance in ~70% of transfected cells, while full or partial incorporation into MTs was observed in only ~30% of cells (Fig. 2a–ii, a–iii, and b). Quantification of the colocalization of TUBB-V5 WT and TUBB-V5 G308S with α-tubulin confirmed these results (Fig. 2c). These data suggest that TUBB G308S impairs αβ tubulin dimer formation or incorporation into MT protofilaments.

**Fig. 2: The G308S mutant TUBB protein shows reduced incorporation into microtubules.**

The observation that TUBB-V5 G308S showed residual MT incorporation in a small proportion of transfected cells prompted us to investigate the impact of this variant on MT function. To this end, we used prime editing⁴¹ to generate a TUBB G308S heterozygous cell line (TUBB^G308S) in human diploid p53-null hTERT-RPE1 cells (TUBB^WT). The genotype of the generated TUBB^G308S cell model was confirmed by Sanger sequencing (Fig. 3a) and no editing was detected at the four most likely predicted protein-coding off-target sites (Fig. S2). Compared to plasmid-based studies, where the mutant protein is overexpressed and present in variable amounts across different cells, TUBB^G308S cells represent a more physiologically relevant system in which the allele dosage of the heterozygous patient is preserved. Semiquantitative RT-PCR for TUBB mRNA showed comparable expression levels in TUBB^WT and TUBB^G308S cells (Fig. 3b–i), where both the wild-type and the mutant alleles are transcribed at a 1:1 ratio (Fig. 3b–ii). Mutations in β-tubulin can result in post-translational degradation of the mutant protein⁴². However, immunoblot analysis revealed that β-tubulin levels in total cell lysates were unchanged in TUBB^G308S compared to TUBB^WT cells (Fig. S3a), suggesting that the G308S mutation does not alter TUBB expression or protein stability.

Fig. 3: Heterozygous TUBBG308S hTERT-RPE1 cells nucleate fewer microtubules that show excessive curvature. — **Fig. 3: Heterozygous TUBB^G308S hTERT-RPE1 cells nucleate fewer microtubules that show excessive curvature.**

To understand how the TUBB G308S mutation impacts MT organization and dynamics, we first examined the entire MT array in TUBB^WT and TUBB^G308S cells by immunofluorescence. TUBB^G308S cells showed overall dimmer α-tubulin intensity than TUBB^WT cells (Fig. 3c). Indeed, the mean α-tubulin fluorescence intensity around the ninein-marked centrosome was significantly reduced in TUBB^G308S cells compared to TUBB^WT cells (Fig. 3d). A similar reduction in fluorescence intensity was observed when staining for newly formed MTs marked by tyrosinated α-tubulin (Fig. 3d). This finding suggests that MT nucleation or early polymerization is impaired in TUBB^G308S cells. To better visualize MT assembly in TUBB^G308S cells, we performed imaging of live cells transiently transfected with the MT plus-end binding protein EB1-GFP to monitor MT nucleation and growth (Fig. 3e). We quantified the number of EB1-GFP comets in cells expressing weak and medium EB1-GFP levels. Intriguingly, we found a significant decrease in the number of EB1-GFP comets emanating from the centrosome in TUBB^G308S cells compared to TUBB^WT cells (Fig. 3e, f). To complement these observations, we assessed MT regrowth after nocodazole treatment and washout in TUBB^G308S and TUBB^WT cells. Quantifying levels of tyrosinated α-tubulin after 2 or 4 min of recovery showed lower fluorescence intensity in TUBB^G308S cells compared to TUBB^WT cells (Fig. 3g), suggesting that fewer MTs are nucleated and assembled in mutant tubulin-expressing cells. To determine if this was from a defect in the MT organizing centre (MTOC), we examined the centrosomal recruitment of γ−tubulin and ninein, two proteins required for MT polymerization from the centrosome⁴³. The centrosomal levels of both γ−tubulin and ninein were similar in TUBB^G308S cells compared to control cells (Fig. 3h), suggesting that TUBB^G308S cells have intact MTOC structures but with lower MT nucleating capacity.

We also noticed, during imaging of the MT array, that MT architecture was different in TUBB^G308S cells, where wavy and buckled MTs were frequently observed (Fig. 3i). To analyze this further, we acquired thin confocal sections of individual MT filaments and measured their curvature. Analysis of MT curvature (detailed in Methods) showed that the distribution of the average MT curvature in TUBB^G308S cells deviated significantly from that of TUBB^WT cells by being wider and shifted to the right (Fig. 3j). We also examined MT plus-end trajectories in live cells by tracking EB1-GFP in TUBB^G308S cells compared to TUBB^WT cells. The curvature of MT plus-end trajectories was measured after tracing maximum intensity confocal projections of EB1-GFP in time-lapse series (Fig. S3b). We found that the EB1-GFP curvature distribution in mutant TUBB^G308S cells was significantly different from that of TUBB^WT cells (Fig. S3c), suggesting structural alterations, or bending, caused by the mutation. Next, we examined EB1 localization on MT plus-ends and found a significant decrease in the mean fluorescence of EB1-GFP comets in TUBB^G308S cells compared to TUBB^WT cells (Fig. S3d). Collectively, these results show that the TUBB G308S mutation affects MT assembly and overall MT organization.

The TUBB G308S variant causes a primary cilia defect in cultured cells

The ciliopathy-like features in the patient’s brain MRI prompted us to investigate the potential effect of the newly identified TUBB G308S variant on primary cilia. To this end, we first sought to characterize cilia-related phenotypes in primary skin fibroblasts from the patient and an age- and sex-matched healthy control (Fig. 4a). To validate primary skin fibroblasts as a suitable cell type for this purpose, we verified that TUBB is expressed. Semiquantitative transcript analysis by RT-PCR revealed that TUBB is highly expressed in skin fibroblasts, and at similar levels in control and patient cells (Fig. 4b–i). Moreover, both the wild-type and the mutant alleles are transcribed in patient cells, at a 1:1 ratio (Fig. 4b–ii). To study primary cilia in cell culture, ciliation is commonly induced by promoting cell cycle exit through serum starvation. Immunostaining of primary cilia in serum-starved quiescent fibroblasts (Fig. 4c) revealed that patient cells exhibited a significant decrease (~30%) in ciliation frequency compared to control cells (Fig. 4c, d), while showing normal cilium length (Fig. 4e). This finding supports a role for ciliation defects in the patient’s clinical presentation and provides a cellular phenotype for subsequent in vitro studies.

**Fig. 4: Patient fibroblasts show reduced ciliation and normal cilium length.**

To associate the ciliation phenotype observed in patient cells with the TUBB G308S variant, the effect of the mutation must be studied in a context that is different from the patient’s genetic background, and isogenic between control and mutant cells. Human hTERT-RPE1 cells are a common model to study primary cilia in vitro due to efficient ciliation and the functional properties of their cilium⁴⁴. Therefore, we assessed ciliation in the TUBB^G308S cell model compared to the isogenic TUBB^WT cell line. Immunostaining of cilia in serum-starved cells (Fig. 5a) revealed that the TUBB^G308S cell model exhibited a significant decrease (~35%) in ciliation frequency compared to control cells (Fig. 5a, b). Of note, similar results were also observed in a second TUBB G308S heterozygous cell line (TUBB^G308S-1) (Fig. S1b, S2) and the magnitude of the reduction was comparable to that observed in patient fibroblasts. The cell model accurately recapitulates the ciliation phenotype observed in patient cells, showing that the TUBB G308S mutation impairs ciliation and reinforces the hypothesis that it is causative of the patient’s clinical phenotype.

Fig. 5: Heterozygous TUBBG308S hTERT-RPE1 cells show reduced ciliation. — **Fig. 5: Heterozygous TUBB^G308S hTERT-RPE1 cells show reduced ciliation.**

Mice carrying the TUBB5 G308S variant show reduced ciliation in the cerebellum and kidney

To investigate the effect of the patient’s TUBB G308S variant on primary cilia in vivo, we used prime editing⁴¹ with the aim of generating a mouse model carrying the G308S mutation in TUBB5, the murine ortholog of TUBB that shares an identical protein sequence with its human counterpart. The F0 chimeric mice — all males — were either sterile with no viable or motile sperm, or failed to transmit the mutation to their offspring. Chimeric mice were sacrificed at postnatal day 17 (P17, juvenile) or at 20 weeks of age (adult), and brains and kidneys, which are relevant organs in neurodevelopmental ciliopathies, were collected for DNA analysis and histological evaluation. Mice with a degree of chimerism higher than 25% in the brain and kidney, as determined by Sanger sequencing, were selected for further assessment.

The brain region primarily affected in JS is the cerebellum. Histological analysis showed no overt morphological abnormalities in the cerebellum of TUBB5 G308S chimeric mice, which displayed normal foliation and preserved overall structure (Fig. S4a). Consistent with this, quantification of molecular layer (ML) thickness across different lobes revealed no significant differences compared to control mice (Fig. S4b). Based on the results obtained in cultured cells, we next examined the level of ciliation in the cerebellum of chimeric mutants. To this end, we immunostained brain sections from adult TUBB5 G308S mice and age- and sex-matched control mice for the primary cilia marker ARL13B. To distinguish ciliation rates across cerebellar cell types, we co-stained with cell type-specific markers: calbindin-D28 for Purkinje cells, GFAP for astrocytes, and IBA1 for microglia. In addition, ciliation was assessed in DAPI-positive cells within the granule cell layer (GCL) and the (ML) (Fig. 6a, S5a, c). TUBB5 G308S chimeric mice exhibited a significant decrease in ciliation within the GCL, while no significant differences were observed in the other cerebellar regions or cell types (Fig. 6b, S5b, d). Next, we investigated additional phenotypes associated with reduced brain ciliation. Reduced ciliation in adult mice has been linked to a decrease in VGLUT2-positive synapses on Purkinje cell dendrites⁴⁵. To investigate this, we immunostained for VGLUT2 (Fig. 6c) and quantified the number of synapses (VGLUT2-positive puncta) in the ML. We did not detect a statistically significant difference in the number of VGLUT2-positive puncta between chimeric and control mice (Fig. 6c, d).

**Fig. 6: Chimeric TUBB5 G308S mice show reduced ciliation in the cerebellar granule cell layer.**

Consistent with the established association between primary cilia dysfunction and obesity in both mice and humans⁴⁶, adult TUBB5 G308S mice exhibited a significant increase in total body weight and visceral fat compared to wild-type controls (Fig. S4c).

To investigate the potential effects of the TUBB5 G308S mutation on motile cilia function, we examined ARL13B-positive motile cilia in ependymal cells lining the third ventricle in chimeric mice. We observed no overt changes in cilia density or length (Fig. S6a). Consistent with these findings, sagittal histological sections showed no evidence of distended ventricles or hydrocephalus, hallmark features of ependymal cilia dysfunction⁴⁷ (Fig. S6b), suggesting that the function of ependymal motile cilia is intact in the chimeric mice.

Primary cilia also extend into the tubular lumen of the kidney, and kidneys are commonly affected in primary ciliopathies, with cyst formation being the hallmark histopathologic feature in both patients and mouse models^48,49. Therefore, we assessed ciliation in the kidneys of chimeric mice. Immunostaining for ARL13B (Fig. 7a) revealed a significant reduction in primary cilia in the kidneys of adult TUBB5 G308S chimeric mice compared to control mice (Fig. 7a, b). A similar trend towards reduced ciliation was also observed in juvenile TUBB5 G308S chimeric mice, although the difference was not statistically significant (Fig. S7a, b). Consistent with these findings, histopathological assessment of kidney sections revealed the presence, in all adult TUBB5 G308S mice, of glomerular cysts characterized by cystic dilatation of the Bowman’s spaces and the initial proximal convoluted tubule, as well as collapse of the glomerular tufts (Fig. 7c and d). Specifically, across the five chimeric mice analyzed, 76% of measured renal corpuscles (38 of 50) exhibited Bowman’s space areas exceeding twice the average size observed in control animals (Source Data, Fig. 7 tab). In addition, cystic dilatation of tubules was also observed in three out of five chimeric mice tested (Fig. 7c), though the cystic area was minimal (Fig. S8). In contrast, no cysts were observed in P17 mice (Fig. S7c).

**Fig. 7: Chimeric TUBB5 G308S mice show a reduction in kidney ciliation and develop glomerular and tubular cysts.**

Taken together, these results demonstrate that the TUBB5 G308S mutation impairs ciliation in vivo and results in histological ciliopathy phenotypes in the kidney, further supporting the connection between the patient’s variant and primary ciliopathies.

The TUBB G308S variant impairs primary cilia formation and centrosome migration

We next sought to understand the molecular mechanism by which the TUBB G308S mutation impairs ciliation. Since the primary cilium is assembled in the G0/G1 phase of the cell cycle²⁶ and MT function and organization are also closely interconnected with cell cycle progression^50,51, we first wondered if the TUBB G308S mutation affects ciliation by altering cell cycle regulation. To test this, we investigated whether TUBB^G308S cells are able to exit the cell cycle upon serum starvation. Cell cycle analysis by flow cytometry revealed that, despite cycling slower than TUBB^WT cells, TUBB^G308S cells responded to serum starvation (Fig. S9a, b). In particular, TUBB^G308S cells exited the cell cycle as effectively as TUBB^WT cells and no difference in the proportion of quiescent cells was observed following serum starvation in the conditions where the ciliation defect was apparent (Fig. S9c). This finding rules out the hypothesis that the observed ciliation phenotype is an indirect consequence of a primary cell cycle defect.

We next investigated whether the ciliogenesis pathway is compromised in mutant cells. hTERT-RPE1 cells rely on the intracellular pathway of ciliogenesis^28,29, which is the same pathway used by neurons and neuronal progenitors^28,29,30, the relevant cell types in neurodevelopmental disorders. In intracellular ciliogenesis, dynamic MTs play a critical role in trafficking PCVs to the centrosome and in the subsequent migration of the forming cilium to the plasma membrane^31,33,35. In hTERT-RPE1 cells, PCVs trafficking occurs within the first few hours of serum withdrawal, when axoneme elongation begins^31,33. To test whether the TUBB G308S mutation affects early stages of ciliogenesis, we quantified the presence of detectable cilia in TUBB^WT and TUBB^G308S cells after 2, 4, or 6 h of serum withdrawal. We found that TUBB^G308S cells showed lower ciliation frequencies than control cells, with a difference apparent at 4 h after serum withdrawal that reached statistical significance at 6 h (Fig. 8a), supporting the impact of the TUBB G308S mutation on the early steps of cilia formation. In the early stages of ciliogenesis, PCVs are recruited to the mother centriole via dynein-mediated MT transport^31,33. To investigate whether the TUBB G308S mutation affects dynein activity, we examined the transport of transferrin, a well-established model to study dynein-mediated vesicular trafficking⁵². TUBB^WT and TUBB^G308S cells were serum-starved for 24 h, then incubated with fluorescently-labeled transferrin and chased for 45 min to 1 h. Analysis of intracellular transferrin distribution (Fig. 8b) revealed that, in control cells, transferrin-positive vesicles showed strong accumulation in the perinuclear region. In contrast, in TUBB^G308S cells transferrin-positive vesicles were more dispersed throughout the cytoplasm (Figs. 8b, c), suggesting a significant impairment in vesicular trafficking along MTs containing the mutant TUBB G308S protein.

Fig. 8: Heterozygous TUBBG308S hTERT-RPE1 cells show defective primary cilia formation and migration. — **Fig. 8: Heterozygous TUBB^G308S hTERT-RPE1 cells show defective primary cilia formation and migration.**

MT pushing forces, generated by growing MTs, are proposed to apically move the centrosome/forming cilium complex towards the plasma membrane³⁵. Therefore, we wondered whether the MT nucleation defects and MT deformations observed in TUBB^G308S cells (Fig. 3) affect the MT subpopulations contributing to the migration of the primary cilium to the plasma membrane. To test this hypothesis, we induced primary cilia formation by serum starving cells for 24 h before immunostaining for primary cilia and the centrosomal protein CEP135 (Fig. 8d). Confocal images encompassing the entire thickness of cells were acquired (Fig. 8d) and converted to XZ projections to measure the relative distance between the ciliated centrosome and the nuclear apex. We found that ciliated centrosomes in TUBB^G308S cells traveled shorter distances compared to their counterpart in control cells (Fig. 8d). Similar results were obtained when measuring the spatial proximity of ciliated centrosomes to the nearest nucleus margin (Fig. S10a, b).

To investigate the underlying molecular mechanism of the reduced primary cilia formation and migration, we analyzed the assembly and structure of MTs in serum-starved cells. Immunofluorescence analysis of total MTs in TUBB^G308S cells showed a reduced MT density compared to control cells also in serum-starved conditions (Fig. 8e). We also analyzed MT nucleation through EB1-GFP expression in control and mutant cells under serum-starvation. TUBB^G308S cells showed significantly less MT nucleation events compared to TUBB^WT cells (Fig. 8f and Supplementary Movie 1), as well as reduced EB1-GFP accumulation at MT plus-ends (Fig. S10c). Since MT pushing forces are involved in primary cilia migration, we also examined the speed of EB1-GFP comets and found MTs in TUBB^G308S cells had reduced growth speed compared to MTs in control cells (Fig. 8g). Next, we examined acetylated microtubules (Ac-MTs), a stable subset of the MT network, in serum-starved cells by immunostaining for acetylated α-tubulin. In control cells, Ac-MTs appeared as filaments localized to the perinuclear region (Fig. 8h). In TUBB^G308S cells, we observed a significant increase in the percentage of cells with enhanced perinuclear Ac-MTs compared to TUBB^WT cells (Figs. 8h, i). Furthermore, fluorescence intensity measurements of the Ac-MT network revealed higher intensity levels in TUBB^G308S cells compared to TUBB^WT cells (Fig. 8j). Enhanced acetylation has been correlated with mechanical stress at sites of lattice defects, such as breaks and bends^53,54,55, further supporting potential structural irregularities within the MT network in TUBB^G308S mutant cells.

Collectively, our results suggest that the TUBB G308S mutation impairs primary cilia formation and migration, due to the critical roles of MT growth, dynamics and organization in ciliogenesis.

TUBB gene haploinsufficiency does not affect ciliation

Next, we intended to investigate the association between mutations in the TUBB gene and defective ciliation beyond the patient’s G308S variant. While most known pathogenic mutations in tubulin genes are dominant², both gain-of-function and haploinsufficiency pathogenic mechanisms have been proposed^36,56,57,58.

To evaluate the contribution of TUBB haploinsufficiency to the ciliation phenotype, a TUBB haploid cell model was generated in hTERT-RPE1 cells using CRISPR/Cas9⁵⁹. Haploidization of the TUBB locus was achieved by delivering a pair of guide RNAs (gRNAs) that span the TUBB gene from intron 1 to the downstream intergenic region (Fig. 9a), with the intronic gRNA being specific to a naturally occurring heterozygous SNP in hTERT-RPE1 cells (Fig. 9a–i). This genome editing strategy resulted in TUBB truncation after exon 1 (19 amino acids) on the targeted allele, leaving the TUBB gene intact on the non-targeted allele (Figs. 9a–ii and 9a–iii). Copy number analysis by digital droplet PCR (ddPCR) confirmed the successful haploidization of the TUBB locus (Fig. 9b). Gene expression analysis by quantitative RT-PCR (qPCR) showed a ~ 30% decrease in TUBB mRNA levels in the haploidized cell line compared to wild-type cells, suggesting partial compensation for the deleted allele by TUBB haploid cells (Fig. 9c). Immunostaining of cilia in serum-starved cells (Fig. 9d) revealed no difference in ciliation frequency between TUBB haploid and wild-type cells (Fig. 9d, e). This finding indicates that TUBB haploinsufficiency does not affect ciliation in hTERT-RPE1 cells.

Fig. 9: TUBB haploidized hTERT-RPE1 cells show normal ciliation. — **Fig. 9: *TUBB* haploidized hTERT-RPE1 cells show normal ciliation.**

Other clinically relevant dominant mutations in TUBB affect ciliation

We next sought to investigate the effect of other clinically relevant mutations in TUBB on primary cilia. In particular, we considered all the sixteen non-synonymous single nucleotide polymorphism (SNP) variants present in the ClinVar database in April 2020.

Since mutations in the TUBB gene do not affect ciliation by haploinsufficiency (Fig. 9e), we reasoned it should be possible to elicit their gain-of-function phenotype and test their effect in an overexpression system. To this end, hTERT-RPE1 cells were transiently transfected with either TUBB-V5 WT or TUBB-V5 bearing a specific variant, serum starved the following day for 24 h to induce ciliogenesis, and the percentage of ciliated cells was assessed in transfected (V5 positive) and non-transfected (V5 negative) cells (Fig. 10a, b). Overexpression of the patient’s G308S variant caused a significant decrease (~90%) in ciliation frequency compared to cells overexpressing the wild-type protein (Fig. 10b). This finding confirms that the TUBB G308S mutation impairs ciliation and validates our overexpression strategy for variant testing. Of note, the magnitude of the ciliation defect was greater upon overexpression than it was in patient fibroblasts or in the TUBB^G308S cell models (Figs. 4d, 5b and S1b), suggesting a dose-dependent gain-of-function pathogenic mechanism for the G308S variant. Consistent with this, overexpression of wild-type TUBB-V5 in TUBB^G308S cells led to an increase in ciliation (Fig. S11a).

**Fig. 10: A subset of brain development dominant mutations in TUBB affects ciliation in hTERT-RPE1 cells and is associated with ciliopathy-like brain features.**

Previously reported pathogenic mutations in TUBB are associated with CSC-KT (OMIM #156610)³⁷, which do not cause gross brain malformations, and with a specific tubulinopathy characterized by structural brain abnormalities (OMIM #615771)³⁶. Overexpression of the CSC-KT causing variants (Q15K and Y222F) did not cause any ciliation defect. However, overexpression of the tubulinopathy-associated variant E401K — but not of T221I, M299V or V353I — caused a significant decrease (~80%) in ciliation frequency compared to cells overexpressing the wild-type protein (Fig. 10b). With regard to the variants with undetermined clinical significance in ClinVar, all ten were preliminarily tested in single-replicate experiments. No effect on ciliation was found for the variants classified as “likely benign” or “of uncertain significance”, while the “likely pathogenic” variant E108K showed reduced ciliation and was prioritized for replicate studies (Fig. S11b). Overexpression of the E108K variant was confirmed to affect ciliation, although in a milder way compared to G308S or E401K (Fig. 10b). Brain MRI of the individual carrying the E108K variant (Fig. 10c) revealed structural brain abnormalities, including features compatible with JS such as superior vermian dysgenesis and a pattern reminiscent of a non-classic MTS on axial images (Fig. 10c–i and 10c–ii).

In the same experiment, the ability of each TUBB variant to incorporate into MTs was assessed based on the filamentous, diffuse or partially diffuse appearance of the V5-tagged exogenous TUBB protein (Fig. S11c, d). The results were concordant with the existing literature for all the variants on which incorporation studies had been previously performed (Q15K, Y222F, M299V, V353I, E401K)^36,37. Of note, a correlation was found between the ability of each TUBB variant to incorporate into MTs and its effect on ciliation. Incorporation defects were observed only for the three variants that also affected ciliation: G308S, E401K, and E108K. As with ciliation, G308S and E401K showed a strong incorporation defect while E108K displayed a milder phenotype (Fig. S11d).

These findings show that some brain development dominant mutations in TUBB affect ciliation in vitro, putting forward ciliation defects as a contributing pathogenic factor in a subset of tubulinopathy patients.

Discussion

Despite MTs being fundamental to primary cilia formation and function, an association between mutations in tubulin genes and neurodevelopmental ciliopathies has not been established, and tubulinopathies and neurodevelopmental ciliopathies are considered as two independent groups of disorders. Here, we present evidence suggesting that defective ciliogenesis is a contributing pathogenic factor in at least a subset of tubulinopathy patients. We show that three dominant mutations (G308S, E108K, and E401K) in a β-tubulin gene (TUBB) from patients with structural brain malformations affect ciliation, and that the patients whose MRI is available to us (G308S and E108K) exhibit varying degrees of ciliopathy-like brain features. The TUBB G308S variant was identified in a patient with brain abnormalities reminiscent of both JS and tubulinopathies. We validated it as the bona fide disease-causing mutation by showing that it affects MT dynamics and structure, impairs the ciliogenesis pathway, and results in reduced ciliation when modeled in mice. We also showed that TUBB haploinsufficiency does not impair ciliation and identified E108K and E401K in a screen for gain-of-function dominant ClinVar mutations in TUBB affecting brain development and ciliation.

Consistent with our finding that heterozygous loss of TUBB does not affect ciliation, the three variants all act in a gain-of-function fashion and impair ciliation upon overexpression in wild-type cells (Fig. 10b). However, the molecular mechanism by which each variant affects ciliation could be different. The G308S variant identified in this work shows reduced incorporation into cellular MTs in overexpression studies, impairs MT nucleation, affects MT stability, and alters MT architecture. From a structural point of view, G308 seems key in positioning R309 for charge interactions that are important for TUBB tertiary structure. The same pattern is also present in all human α-tubulins, where G308 is conserved and R309 is replaced by a lysine residue that provides similar electrostatic properties (TUBA1A is shown as an example in Fig. 1b). Considering the structural role of G308 and the unique flexibility of glycine residues, we hypothesized that the G308S mutation could have a subtle effect on the TUBB fold, where the slightly bulkier serine does not prevent TUBB from folding but makes the fold less stable. Consistent with this idea, in silico protein stability analysis predicted G308S to be among the most destabilizing glycine to serine substitutions in β-tubulins across different species (Fig. S12). In a plausible scenario, the mutant protein forms unstably folded αβ tubulin dimers whose incorporation into MTs is therefore unstable. Since TUBB is one of the most abundant β-tubulin isotypes, this would result in a net nucleation defect because MT assembly is a cooperative process where higher stabilization energy is required for the initial nucleation stage compared to the later elongation phase⁶⁰. For this reason, stochastic incorporation of mutant αβ dimers could instead be tolerated during elongation, resulting in the observed curved MTs. Moreover, the tendency of mutant αβ dimers to form curved MTs likely affects nucleation itself, as it was shown that curved tubulin oligomers reduce nucleation efficiency by hindering lateral interactions between growing protofilaments and destabilizing early nucleation intermediates^61,62. Similar abnormalities in MT organization and dynamics persist upon induction of ciliogenesis in serum-starved TUBB^G308S cells. Impaired ciliation is apparent at early time points after serum withdrawal, suggestive of defective primary cilia formation rather than reduced stability of the formed cilia. In the early stages of ciliogenesis, dynein-mediated MT transport is crucial to recruit PCVs to the mother centriole^31,33 and we observed that dynein-mediated vesicular trafficking is disrupted in serum-starved TUBB^G308S cells (Fig. 8c). In mammalian cells, dynein is mainly targeted to the plus-end of MTs via a cascade of protein interactions ultimately dependent on EB1 recruitment onto MTs⁶³. In serum-starved TUBB^G308S cells, we observed a significant reduction of EB1 accumulation on MTs. Reduced recruitment of EB1 has been associated with different types of structural alterations of the MT plus-end, including those induced by a dominant missense mutation in the β-tubulin gene TUBB2B causing brain malformations^64,65. Furthermore, variations in the overall MT structure such as bending and buckling have been shown to contribute to motor-mediated MT transport and influence vesicle motion⁶⁶. Overall, we hypothesize that the nucleation of fewer MTs and the structural abnormalities of the existing ones affect dynein-mediated transport of PCVs to the mother centriole, resulting in reduced ciliation. Consistent with impaired nucleation, altered structure, and decreased growth speed of MTs during ciliogenesis, we also observed reduced migration of ciliated centrosomes to the plasma membrane, a movement driven by pushing forces from MTs nucleation, bundling, and polymerization³⁵. The defect in cilia migration may also be influenced by the observed increase in MT acetylation in serum-starved TUBB^G308S cells. In vitro studies have shown that MT acetylation reduces nucleation frequency and accelerates the rate of shrinkage⁶⁷. Acetylation occurs at MT regions with structural defects, such as breaks and bends, which allow α-tubulin acetyltransferase 1 (αTAT1), an enzyme that acetylates α-tubulin at lysine 40, access to the MT lumen^53,54,55. In cells lacking deacetylase HDAC6 activity, MT hyperacetylation reduces both polymerization and depolymerization rates⁶⁸. This imbalance between growing and stabilized MT populations in mutant cells may interfere with the pushing forces required for cilia movement. However, the impact of impaired cilia migration on ciliation and disease pathology is hard to evaluate. In principle, it could have no effect at all if the slowly migrating cilia eventually reach the cell surface during ciliogenesis in vivo. Conversely, it is also possible that cilia are selectively disassembled if they remain for too long in the cytosol, where they are not fully functional. In this case, defective migration would also contribute to the observed reduction in ciliation. For clarity, the proposed model for the molecular mechanism by which TUBB G08S affects ciliogenesis is summarized in Fig. S13.

While a similar molecular mechanism is conceivable for the E108K variant, which also shows partial incorporation into cellular MTs, the E401K variant possibly acts through a different mechanism. TUBB E401K was previously shown to impair the chaperone-mediated assembly of αβ tubulin dimers and to be unable to incorporate into MTs³⁶. These features suggested a loss-of-function mode of action for E401K, which was partially confirmed by phenotypic similarities in the brain of E401K mutant mice and heterozygous knock-out mice, both showing excessive apoptosis of cortical upper-layer neurons leading to microcephaly⁵⁸. However, some gain-of-function phenotypes of the E401K variant were also observed in the same studies. Overexpression of the mutant protein in developing mouse brain resulted in aberrant neuronal migration and positioning³⁶, while the presence of ectopic progenitors and defects in mitotic spindle orientation were observed in the embryonic cortex of E401K mutant mice but not of heterozygous knock-out mice⁵⁸. These findings led to the hypothesis that the E401K variant acts through a complex dual mechanism where loss-of-function effects are associated to microcephaly while gain-of-function effects are linked to defective neuronal positioning and caused by undetected residual incorporation of the mutant protein into MTs or by a deleterious effect on the tubulin folding pathway^36,58. We propose that reduced ciliation is due to the gain-of-function mode of action of TUBB E401K and show that minimal (~2% of cells) incorporation of TUBB E401K into cellular MTs can occur, suggesting that the mutant protein is not intrinsically incapable of incorporating into MTs.

Furthermore, we observed a counterintuitive direct correlation between the diffuse appearance of each TUBB variant upon overexpression and its gain-of-function effect on ciliation. This finding suggests a possible alternative interpretation of the MT incorporation assay, especially for variants that show residual, albeit minimal, incorporation. In such cases, the assay may reflect how stably the mutant tubulin is incorporated into MTs rather than its intrinsic ability to incorporate. In this scenario, which is not mutually exclusive with poorly incorporated gain-of-function variants impairing the tubulin folding pathway, the proportion of cells showing diffuse tubulin would indicate the severity of the unstable incorporation phenotype and, in turn, of the destabilizing effect on MTs.

Although the three variants may act through different molecular mechanisms, it is likely that they all ultimately affect the early stages of ciliogenesis as our data argue against the presence of TUBB in the axoneme of human primary cilia, at least in the cell types used in this study. Indeed, primary skin fibroblasts from the patient showed normal cilia length and morphology, and V5-tagged wild-type TUBB did not localize to the ciliary axoneme when overexpressed in hTERT-RPE1 cells. Even though an axonemal role for TUBB in the developing brain cannot in principle be excluded, our work suggests that TUBB is present in the MT subpopulation involved in PCV trafficking and centrosome migration, shedding some light on the largely unexplored field of the roles of distinct tubulin isotypes in ciliogenesis.

In our overexpression screen, some pathogenic TUBB variants did not show any ciliation defect. These include all the CSC-KT causing mutations (Q15K and Y222F), which do not cause brain malformations, but also three variants associated with brain abnormalities in ClinVar (T221I, M299V, and V353I). No functional evidence of pathogenicity or description of the patient’s brain is available for T221I, while M299V and V353I have been validated to cause microcephaly with structural brain abnormalities³⁶. Although gain-of-function phenotypes have been reported for M299V and V353I, which caused defects in neuronal migration (both variants) and positioning (M299V) upon overexpression in developing mouse brain, their molecular mechanism of action is significantly different from that of the ciliation-impairing variants G308S, E401K and E108K as they have no (V353I) or moderate (M299V) impact on the tubulin folding pathway and show complete, stable incorporation into cellular MTs³⁶. Considering the role for defects in MT structure, nucleation and growth rate in impairing ciliogenesis and the proposed model linking those defects to unstable incorporation of the mutant tubulins, it is conceivable that these stably incorporating variants cause subtler effects on MT properties and dynamics that do not affect ciliation, or that affect it too mildly for the sensitivity of our assay. Furthermore, our screen is limited by the assumption of a gain-of-function mode of action for the tested variants, as loss-of-function mutations would not show any phenotype upon overexpression in wild-type cells. The ability of TUBB haploid cells to ciliate normally indeed suggests that heterozygous loss-of-function mutations in TUBB cannot affect ciliation. However, we observed that TUBB haploid cells partially compensate for the deleted allele at the transcriptional level. The mechanism used by cells to sense tubulin levels is currently unknown, nor is it known if this partial compensation is crucial to restore normal ciliation in TUBB haploid cells. Therefore, the possibility of a heterozygous loss-of-function missense mutation that escapes the sensing mechanism and impairs ciliation, although unlikely, cannot in principle be excluded.

The TUBB5 G308S chimeric mouse model exhibited impaired ciliation in both the kidneys and cerebellum, along with an increase in visceral fat which is consistent with the established link between primary cilia dysfunction and obesity⁴⁶. Primary cilia-related phenotypes were most prominent in the kidneys, where reduced ciliation was accompanied by cystic dilation of Bowman’s spaces affecting the majority of renal corpuscles analyzed, as well as tubular cyst formation, hallmark features of primary ciliopathies⁴⁸. In the cerebellum, the observable phenotype was limited to reduced ciliation in the granule cell layer (GCL), and the chimeric mouse model did not recapitulate the structural cerebellar phenotypes observed in JS.

Although cerebellar malformations are a hallmark of JS in human patients, cerebellar phenotypes in JS mouse models are known to be variable¹⁶. Some models exhibit only mild cerebellar abnormalities⁶⁹, while others show no overt cerebellar defects^70,71. This variability has been linked to factors such as the mouse genetic background^70,72, as well as allele penetrance even among mutant littermates^72,73 or between humans and mice⁷¹. In contrast, renal manifestations are typically more consistent across JS mouse models, which aligns with our observations.

The chimeric nature of the mice in our study may have limited the sensitivity of our analysis due to genetic heterogeneity, with varying proportions of mutant cells across animals and tissues. Additionally, the need to exclude mice with insufficient chimerism reduced the sample size, potentially limiting our ability to fully capture subtle phenotypes. To address these limitations, conditional mutant mouse lines could provide a more refined model to investigate the effects of TUBB5 mutations and may uncover additional phenotypes, for example, by facilitating assessments at different developmental stages. In-depth studies using these higher-penetrance models, potentially in different genetic backgrounds, will be needed to definitively establish the causal link between the TUBB G308S mutation and ciliopathy-related brain malformations, building upon the converging lines of evidence from our patient findings, cellular models, and the cilia reduction observed in our chimeric mouse model.

We propose defective ciliogenesis to contribute to disease pathogenesis in at least a subset of TUBB-related tubulinopathy patients, alongside other cellular phenotypes associated with MT dysfunction, such as the mitotic spindle defects described for the E401K variant⁵⁸. The heterogeneity in clinical manifestations and molecular phenotypes of the pathogenic TUBB variants studied so far suggests that each mutation has a specific mode of action, where the relative contribution of different pathogenic mechanisms may vary^36,37,58. In the two patients analyzed, the severity of the ciliopathy-like brain phenotype (Figs. 1a–ii and c–i, c–ii) correlates with the magnitude of the ciliation defect observed in vitro for their respective mutations (G308S and E108K) (Fig. 10b), with G308S being the more severe mutation. Interindividual variability in disease manifestation is consistent with the complexity of these conditions, as tubulinopathies and neurodevelopmental ciliopathies are two genetically heterogeneous groups of disorders characterized by a complex genotype-phenotype correlation. For both, different mutations in the same gene can result in specific phenotypes, and different phenotypes or disease severity can occur in individuals carrying the same variant, suggesting an important role for genetic modifier or environmental factors^2,16.

Recently, Dodd et al. identified de novo variants in the β-tubulin gene TUBB4B as causative for primary ciliary dyskinesia (PCD) and other non-neurodevelopmental ciliopathy features⁷⁴. PCD is a disorder affecting motile cilia, which are organelles homologous to the primary cilium. They demonstrated that TUBB4B is enriched in the axoneme of motile cilia and is essential for their assembly and function in specific tissues, where it appears to be restricted to centriolar and ciliary MTs⁷⁴. The present study complements these findings by suggesting a link between mutations in TUBB, primary cilia and neurodevelopmental ciliopathies, and by showing that defects in primary cilia may contribute to the classic neurodevelopmental features of tubulinopathies, which appear to be absent in TUBB4B patients⁷⁴.

The clinical features of TUBB patients, the absence of hallmark symptoms of motile ciliopathies such as hydrocephalus in TUBB5 G308S chimeric mice, and the lack of apparent ependymal cilia defects in these mice argue against a role for TUBB in motile cilia function. Furthermore, the cryo-EM studies that identified TUBB4B excluded TUBB as a component of the axoneme of human respiratory cilia⁷⁴. However, TUBB5 G308S chimeric mice — all males — exhibited infertility, a trait compatible with motile ciliopathies. In addition, ependymal cilia were also grossly normal in TUBB4B mutant mice as well^74,75, indicating that tissue specificity exists for motile cilia-related β-tubulin isoforms. Therefore, at least in mice, the involvement of TUBB5 in motile cilia function within specific tissues cannot in principle be ruled out, presenting an interesting avenue for future research using conditional knockout and knock-in mouse models. Nevertheless, primary cilia are also known to function in the mammalian testis, and defects in primary cilia have been associated with impaired fertility, likely due to disrupted testicular development^76,77. Moreover, functional MTs are crucial across all stages of spermatogenesis⁷⁸. Therefore, MT and primary cilia dysfunctions resulting from the TUBB5 G308S mutation may explain the infertility phenotype observed in chimeric mice.

Our data link tubulinopathies and neurodevelopmental ciliopathies, two groups of disorders affecting brain development. We showed that some gain-of-function patient mutations in the TUBB gene affect ciliogenesis and are associated with ciliopathy-like phenotypes, potentially revealing an emerging disease mechanism for TUBB-related tubulinopathies and implicating TUBB as a candidate ciliopathy gene. Moreover, these findings can prompt future investigations on the link between tubulinopathies and primary cilia beyond the TUBB gene. Indeed, six tubulin genes have been associated to neurodevelopmental defects to date (TUBA1A, TUBB, TUBB2A, TUBB2B, TUBB3, TUBG1)², but the contribution of distinct tubulin isotypes to cilia formation and homeostasis is still unknown. Furthermore, while tubulinopathies are considered incurable due to their developmental nature⁷⁹, primary cilia are also present in differentiated neurons of the adult brain, where they are crucial to maintain neuronal viability and connectivity^17,45. Therefore, the involvement of primary cilia defects in the pathogenesis of tubulinopathies reveals a potential point of intervention to improve disease outcome after birth.

Methods

Subject recruitment and clinical evaluation

This study was approved by the Research Ethics Board of The Hospital for Sick Children, Toronto, Ontario, Canada. Informed written consent to participate in research and to publish potentially identifying clinical information was obtained from the patient, the patient’s family members, and control subjects. A separate study allowing the sharing of deidentified patient information with patient consent was approved by the Review Board of the Johns Hopkins University School of Medicine, Baltimore, Maryland, USA. Clinical data were obtained by direct examination of participants and retrospective chart review.

Animal studies

The C57BL/6NCrl control mice and the CD-1 (ICR) mice used for the generation of chimeric mice were purchased from Charles River Laboratories or obtained from an in-house colony. All mice used in this study were housed in the specific-pathogen-free facility at The Centre for Phenogenomics (TCP; Toronto, Ontario, Canada). Mice were maintained on a 12-h light/dark cycle in a temperature-controlled (21–22 °C) and humidity-controlled (30%-55%) environment and provided with food and water ad libitum in individually ventilated units (Tecniplast Sealsafe Plus) at a maximum density of five animals per cage. Only male mice at postnatal day 17 (juvenile) or between 19 and 21 weeks of age (adult) were used as controls to ensure age and sex matching with the generated chimeric mice.

All procedures involving animals were performed in compliance with the Animals for Research Act of Ontario and the Guidelines of the Canadian Council on Animal Care. The TCP Animal Care Committee reviewed and approved all procedures.

Brain magnetic resonance imaging

Multiplanar multisequence magnetic resonance imaging (MRI) of the brain was performed at two distinct institutions. MRI shown in Fig. 1, S1 entailed the use of the following sequences, as per institutional protocol: sagittal 3DT1, axial DWI, axial FLAIR, axial/coronal T2, coronal T2 fat-saturated orbits, and FIESTA. MRI shown in Fig. 10, performed at a different institution, included the following sequences: sagittal T1 SPGR, T2 FLAIR CUBE, axial T2 FRF FSE, T1 MT, 3D SWAN, DTI, coronal T2 FRFSE, and high-resolution coronal T2 FRFSE through the hippocampus. Detailed acquisition parameters are provided in the Reporting Summary.

Whole genome sequencing

Genomic library preparation and whole genome sequencing (WGS) of the participants’ DNA were conducted at The Centre for Applied Genomics (TCAG) at The Hospital for Sick Children, Toronto, Ontario, Canada. 100 ng of DNA was used as input material for library preparation using the Illumina TruSeq PCR-free Library Prep Kit following the manufacturer’s recommended protocol. Libraries were validated on a Bioanalyzer DNA High Sensitivity chip to check for size and absence of primer dimers, and quantified by qPCR using Kapa Library Quantification Illumina/ABI Prism Kit protocol (KAPA Biosystems). Each sample was sequenced on one lane of an Illumina HiSeq X platform following Illumina’s recommended protocol to generate paired-end reads of 150 bases in length. Each lane generated 112.5 Gb of data for ~30X coverage of the human genome.

Whole genome sequencing analysis and variant identification

WGS reads were aligned to the UCSC hg19 human reference genome assembly. Bioinformatic analysis of genetic variants was performed using a pipeline established in the Care4Rare project⁸⁰. Namely, the germline variant calling pipeline used is implemented in the bcbio-nextgen framework (https://bcbio-nextgen.readthedocs.io/en/latest/index.html) and it is validated based on the Genome Analysis Toolkit’s (GATK) best practices⁸¹ and the standard NA12878 validation sample. GATK 4.0.7.0 was used, and variants in all 5 family members’ samples were called simultaneously. Variants were annotated using Ensembl Variant Effect Predictor (VEP)⁸², GEMINI⁸³ and Vcfanno⁸⁴. A detailed description of variants annotation can be found at https://docs.google.com/document/d/1zL4QoINtkUd15a0AK4WzxXoTWp2MRcuQ9l_P9-xSlS4/edit?usp=sharing.

To prioritize variants, we first restricted our analysis to gene coding regions and variants rare in the general population (gnomAD⁸⁵ popmax <1%). Among the resulting 1328 variants, we selected those that segregate with the disease phenotype in the family following both dominant and recessive modes of inheritance. No biallelic variant was found to segregate with the disease phenotype in the family under a recessive model. The dominant model with a de novo hypothesis, where a heterozygous variant is present in the proband and absent in parents and siblings, resulted in 12 variants. Of those, a missense variant in TUBB (NM_178014.3: c.922 G > A/p.G308S) became our top candidate since only the TUBB gene had an OMIM description (#615771: cortical dysplasia, complex, with other brain malformations 6, autosomal dominant - #156610: symmetric circumferential skin creases, congenital 1, autosomal dominant).

We also filtered all coding and rare (gnomAD⁸⁵ popmax <1%) non-coding variants in the 201 genes of the “Rare multisystem ciliopathy disorders” panel downloaded from https://panelapp.genomicsengland.co.uk/panels/150. This resulted in 21 coding variants, none of which segregated in the family according to the phenotype, and 348169 non-coding variants. Non-coding variants were filtered under a recessive model of inheritance (Hom in proband, Het in parents, non-Hom in siblings; OMIM gene description is present; OMIM inheritance mode is autosomal recessive). This filter resulted in 169 variants. We reviewed the variants manually, removing low coverage calling artifacts, and, based on the variant annotations, we reached the conclusion that none of the variants was contributing to the disease phenotype.

Additionally, we generated a more stringent variant report containing ultra-rare (gnomAD⁸⁵ popmax <0.5%, frequency in the internal Care4Rare database <5), high quality (variant quality score ≥ 1000, alt depth ≥ 20 in one of the samples) variants not restricted by a gene panel. This filter resulted in 40911 coding and non-coding variants, of which 3479 were in 1240 genes present in OMIM. We reviewed these variants using both dominant and recessive models and have not identified additional candidates that segregate in the family according to the phenotype.

Antibodies

For immunostaining experiments, the following primary antibodies were used: mouse monoclonal anti-α-tubulin (DM1A, Sigma-Aldrich T6199, 1:500 dilution. Fig. 2), mouse monoclonal anti-α-tubulin (DM1A, Sigma-Aldrich T9026, 1:10000 dilution. Figs. 3, 8), chicken polyclonal anti-V5 tag (Abcam ab9113, 1:500 dilution), mouse monoclonal anti-tyrosinated tubulin (TUB-1A2, Sigma-Aldrich T9028, 1:500 dilution), mouse monoclonal anti-γ-tubulin (GTU-88, Sigma-Aldrich T6557, 1:500 dilution for Fig. 3, 1:2000 dilution for Figs. 4, 5, 9), rabbit polyclonal anti-ninein (Sigma-Aldrich ABN1720, 1:300 dilution), mouse monoclonal anti-ARL13B (NeuroMab 75-287, 1:300 dilution. Fig. 6, S5, S6), rabbit polyclonal anti-ARL13B (Proteintech 17711-1-AP, 1:300 dilution), mouse monoclonal anti-acetylated tubulin (6-11B-1, Sigma-Aldrich T7451, 1:500 dilution), rabbit polyclonal anti-CEP135 (Sigma-Aldrich SAB4503685, 1:500 dilution), chicken polyclonal anti-calbindin-D28k (ThermoFisher PA5-143561, 1:1000 dilution), rabbit monoclonal anti-IBA1 (EPR16588, Abcam ab178846, 1:1000 dilution), mouse monoclonal anti-VGLUT2 (N29/29, ThermoFisher MA5-27613, 1:1000 dilution), chicken polyclonal anti-GFAP (Abcam ab4674, 1:1000 dilution).

For immunoblotting, the following primary antibodies were used: mouse monoclonal anti-β-tubulin (AA2, Sigma-Aldrich T8328, 1:500 dilution), mouse monoclonal anti-β-actin (AC-15, Sigma-Aldrich A5441, 1:1000 dilution).

The following secondary antibodies were used: goat anti-rabbit IgG (H + L) Alexa Fluor™ 488 (ThermoFisher A-11008), goat anti-rabbit IgG (H + L) Alexa Fluor™ 555 (ThermoFisher A-21429), goat anti-mouse IgG (H + L) Alexa Fluor™ 488 (ThermoFisher A-11001), goat anti-mouse IgG (H + L) Alexa Fluor™ 555 (ThermoFisher A-21422), goat anti-chicken IgY (H + L) Alexa Fluor™ 488 (ThermoFisher A-11039), donkey anti-mouse IgG (H + L) Alexa Fluor™ Plus 647 (ThermoFisher A-32787), goat anti-mouse IgG (H + L) HRP-conjugated (Jackson ImmunoResearch 115-035-146).

Plasmids

For gene editing experiments, the following plasmids were used: pSpCas9(BB)-2A-Puro (PX459) V2.0, which was a gift from Feng Zhang (Addgene plasmid # 62988); BPK1520_PuroR, the assembly of which was described previously⁸⁶; BPK1520_blastR, the assembly of which was described previously⁵⁹; pCMV-PE2, which was a gift from David Liu (Addgene plasmid # 132775); pU6-pegRNA-GG-acceptor, which was a gift from David Liu (Addgene plasmid # 132777). Plasmids for gRNA expression were generated by ligating annealed oligonucleotides into the linearized PX459, BPK1520_PuroR or BPK1520_blastR vectors. Plasmids for prime editing guide RNA (pegRNA) expression were generated by Golden Gate assembly of annealed oligonucleotides into the pU6-pegRNA-GG-acceptor vector, as previously described⁴¹.

For overexpression experiments, the following plasmids were used: pEGFP-N1 expressing human EB1-GFP (JB131), which was a gift from Tim Mitchison and Jennifer Tirnauer (Addgene plasmid # 39299); pCMV3-TUBB:V5, which was generated for this study.

The pCMV3-TUBB:V5 plasmid was generated using two rounds of site-directed mutagenesis to remove the N-terminal GFPSpark tag from the commercially available pCMV3-GFPSpark:TUBB plasmid (Sino Biological HG11626-ANG) and insert a C-terminal V5 tag (GKPIPNPLLGLDST) with a two amino acids linker (GS). Single-nucleotide polymorphism (SNP) TUBB variants were introduced into pCMV3-TUBB:V5 via site-directed mutagenesis. To preserve the sequence of the pCMV3 vector from unwanted mutations during PCR amplification, the whole TUBB:V5 insert was moved to the pUC19NotI cloning vector at the KpnI/NotI restriction sites and moved back to pCMV3 after mutagenesis.

All site site-directed mutagenesis experiments were performed using the Q5® Site-Directed Mutagenesis Kit (New England Biolabs E0554) following the manufacturer’s recommended protocol. The mutagenic primers used were designed using the online NEBaseChanger tool (https://nebasechanger.neb.com) and are listed in Supplementary Data 1. In all experiments, the complete open reading frame (ORF) was sequenced after mutagenesis to ensure that only the desired mutation was present.

All plasmids generated in this study are available from the corresponding author upon request.

Chimeric mice generation

The generation of TUBB5 G308S chimeric mice was performed at The Centre for Phenogenomics (TCP; Toronto, Ontario, Canada). Genome editing of mouse Embryonic Stem (ES) cells to introduce the desired variant (NM_011655.5: c.922 G > A/p.G308S) was performed at the TCP Transgenic Core by prime editing 3b (PE3b)⁴¹. Briefly, 1 ×106 C57BL/6NTac-C2 ES cells were cotransfected with 4 μg of the pegRNA- expression plasmid (pU6-pegRNA-GG-acceptor), 1.3 μg of the nicking gRNA (ngRNA)- and puromycin resistance- expression plasmid (BPK1520_PuroR), and 2.7 μg of the prime editor- expression plasmid (pCMV-PE2) using the Neon™ Transfection System (Thermo Fisher Scientific MPK5000; pulse voltage 1400 V, pulse width 10 ms, pulse number 3), and then plated onto a 100 mm tissue culture-treated dish covered with 5 × 105 mouse embryonic fibroblasts (MEFs) in DMEM medium supplemented as previously described⁸⁷. Selection with 1 μg/ml of puromycin was initiated 24 h post-transfection and continued for 3 days. At 1 week post-transfection, colonies were picked and replicated for cryopreservation and genomic DNA extraction. Genotyping was performed by PCR amplifying the target region and sequencing the targeted locus. The mutant ES cell clones — all heterozygous — were expanded, and genomic DNA was extracted for additional confirmation prior to aggregation. Cells were thawed in KnockOut™ Serum Replacement (KOSR)+2i medium, the formulation of which was previously described⁸⁷, and passaged 2 or 3 times before aggregation.

The details of morula aggregation have been previously described⁸⁸. Briefly, embryos from superovulated CD-1 (ICR) females were collected at 2.5 days post-coitum (dpc). Zonae pellucidae of embryos were removed by treatment with EmbryoMax^® Acidic Tyrode′s Solution (Sigma-Aldrich MR-004). ES cell colonies were treated with ESGRO Complete Accutase (Sigma-Aldrich SF006) to form loosely connected clumps of cells. Each zona-free embryo was aggregated with a clump of 6-10 ES cells inside a depression well made in the dish with an aggregation needle (BLS Ltd). Aggregates were cultured overnight in microdrops of Global^® medium (CooperSurgical^® LGGG-050) covered by embryo-tested LifeGlobal^® paraffin oil (CooperSurgical^® LGPO-500) at 37 °C and 6% CO₂. The following morning, morulae and blastocysts were transferred into the uteri of 2.5-dpc pseudopregnant CD-1 (ICR) females previously mated with vasectomized males. Chimeric mice — all males — were identified at birth by the presence of black eyes and later by coat pigmentation. Chimeric mice with over 50% ES cell contribution to the coat color were bred with different albino B6N (B6N-Tyr^c-Brd/BrdCrCrl) or CD-1 (ICR) females until they reached 20 weeks of age. Chimeric mice were either sterile and did not produce offspring, or did not achieve ES cell germline transmission, as determined by coat color. At 20 weeks of age, the chimeric mice were euthanized, and their brains and kidneys were collected as described below. Prior to organ collection, the caudae epididymis of apparently infertile chimeric mice were dissected to cryopreserve sperm and potentially rescue the line through in vitro fertilization. However, no viable or motile sperm were detected upon sperm analysis using CASA (Computer Assisted Sperm Analysis) image analysis system (Hamilton Thorne).

To perform the experiments required for manuscript revision, three more rounds of chimeric mouse generation were conducted. Chimeric mice — all males — were then euthanized for analysis at either postnatal day 17 or at 20 weeks of age.

The sequences of the pegRNAs and ngRNAs used in this study are listed in Supplementary Data 2.

Mouse body weight and fat composition analysis

Total body weight and fat mass were assessed in TUBB5 G308S chimeric mice and C57BL/6NCrl control mice using dual-energy X-ray absorptiometry (DEXA) at The Centre for Phenogenomics (TCP; Toronto, Ontario, Canada), following the facility’s standard operating protocol. Body fat composition was calculated as the ratio of fat mass to total body weight.

Mouse organ collection and processing

TUBB5 G308S chimeric mice and C57BL/6NCrl control mice were euthanized by cervical dislocation (for H&E kidney experiments and sperm analysis in adult mice) or by transcardiac perfusion with cold PBS under isoflurane anesthesia (for all other experiments), and their brains and kidneys were collected. The brains were hemisected along the midsagittal plane immediately after dissection. One kidney and one brain hemisphere were snap-frozen in liquid nitrogen and stored at -20 °C for genomic DNA isolation. The other kidney and brain hemisphere were immersion-fixed in 10% neutral-buffered formalin (Sigma-Aldrich HT501128) overnight at 4 °C and paraffin-embedded for histological staining and immunofluorescence at The Centre for Phenogenomics (TCP; Toronto, Ontario, Canada), according to their standard protocols. Kidneys were bisected longitudinally prior to embedding, and all samples were embedded cut side down.

Cell culture and transfection

hTERT-RPE1 cells (ATCC CRL-4000) and p53-null hTERT-RPE1 cells (a gift from Daniel Durocher) were cultured in DMEM/F12 1:1 Mix (Wisent Bioproducts 319-085-CL) supplemented with 10% heat-inactivated FBS (Wisent Bioproducts 080-450) and 1% penicillin-streptomycin solution (Wisent Bioproducts 450-201-EL), and maintained at 37 °C in a humidified atmosphere (5% CO₂). Primary skin fibroblasts were cultured in Alpha MEM (Wisent Bioproducts 310-010-CL) supplemented with 10% heat-inactivated FBS (Wisent Bioproducts 080-450) and 1% penicillin-streptomycin solution (Wisent Bioproducts 450-201-EL), maintained at 37 °C in a humidified atmosphere (5% CO₂) and used in experiments below passage 10. Where indicated, serum starvation was performed by washing cells twice with pre-warmed PBS (Wisent Bioproducts 311-010-CL) and culturing them in serum-free Opti-MEM™ (Thermo Fisher Scientific 31985-062) for the indicated time.

Transfections of hTERT-RPE1 cells and p53-null hTERT-RPE1 cells were carried out using Lipofectamine™ 3000 Transfection Reagent (Thermo Fisher Scientific L3000008) following the manufacturer’s recommended protocol. Unless otherwise indicated, cells were transfected at ~80% confluence 24 h after plating, and the medium was replaced 6h after transfection. Only for the generation of the TUBB haploid cell model, hTERT-RPE1 cells were transfected using the Neon™ Transfection System (Thermo Fisher Scientific MPK5000) with a pulse voltage of 1050 V, a pulse width of 30 ms, and a pulse number of 2.

Cell line generation

Primary fibroblast cell lines were established from a skin biopsy of the patient and an age- and sex-matched control subject at The Centre for Applied Genomics (TCAG), Tissue Culture Facility - Biobanking Services at The Hospital for Sick Children, Toronto, Ontario, Canada according to their standard protocols. The patient-derived cell line was authenticated by Sanger sequencing for the TUBB NM_178014.3: c.922 G > A/p.G308S variant (Fig. 4a).

The TUBB G308S heterozygous hTERT-RPE1 cell lines (TUBB^G308S and TUBB^G308S-1) were generated by prime editing 3b (PE3b)⁴¹. Consistent with previous findings that functional p53 lowers the efficiency of Cas9-mediated gene editing even in the absence of double-strand breaks^89,90,91, p53-null hTERT-RPE1 cells (a gift from Daniel Durocher) were needed to achieve an editing efficiency detectable via Sanger sequencing in the bulk cellular population after antibiotic selection. Cells were plated at 2 × 105 cells/well on 12-well plates (Corning 3513) and transfected the following day as described above. 845 ng of the pegRNA- expression plasmid (pU6-pegRNA-GG-acceptor), 280 ng of the nicking gRNA (ngRNA)- and blasticidin resistance- expression plasmid (BPK1520_blastR), and 375 ng of the prime editor- expression plasmid (pCMV-PE2) were cotransfected. At 24 h post-transfection, blasticidin (Thermo Fisher Scientific A1113903) was added to the cells at 30 μg/ml in complete medium and selection medium was replaced daily for the following 6 days. Next, cells were resuspended in FACS buffer (2% FBS, 2.5 mM EDTA in PBS) at a concentration of 1 × 106 cells/well, filtered through a 40 μm cell strainer (VWR CA21008-949), and single-cell sorted using a MoFlo XDP cell sorter (Beckman Colter) into a 96-well plate (Corning 3599) containing complete medium. Clonal lines were expanded, and genomic DNA was extracted upon passaging for genotyping. Genotyping was performed by PCR amplifying the target region, and sequencing the targeted locus. Attempts to generate this cell line by homology-directed repair (HDR) and base editing (BE) were unsuccessful.

The TUBB haploid cell line was generated in wild-type hTERT-RPE1 cells (ATCC CRL-4000). WGS data from hTERT-RPE1 cells were accessed from the NCBI Sequence Read Archive (experiment ID: SRX858057). The target heterozygous polymorphism in TUBB intron 1 (NM_178014.4: c.57+235 G > C) was manually identified using the Integrative Genomics Viewer⁹² and its presence was confirmed via Sanger sequencing of genomic DNA (Fig. 9a–i). The target polymorphism was chosen as it generates or destroys an SpCas9 PAM sequence, allowing for allele-specific editing. The final selection of both gRNAs used in this experiment (the allele-specific one targeting TUBB intron 1, and the one targeting the intergenic region downstream of TUBB) was based on minimizing off-target sites, as determined by the CHOPCHOP v.3 software⁹³. hTERT-RPE1 cells are endogenously resistant to puromycin because of the method used for their immortalization, and the TUBB haploid cell line was established before the gRNA- and blasticidin resistance- expression plasmid (BPK1520_blastR) was assembled in our laboratory⁵⁹. Therefore, EGFP cotransfection was used to allow FACS-based enrichment of transfected cells. Cells were transfected using the Neon™ Transfection System (Thermo Fisher Scientific MPK5000) as described above, and plated at 1 × 105 cells/well on 24-well plates (Corning 3524). The two gRNAs were cloned into the gRNA- and SpCas9- expression vector PX459 and transfected in equal proportions (1000 ng each) into cells, together with a 7.5 kb CMV-EGFP expression plasmid assembled in our laboratory (200 ng). At 48 h post-transfection, EGFP positive cells were sorted using a MoFlo XDP cell sorter (Beckman Colter), replated and allowed to recover for 5 more days. At 1 week post-transfection, cells were single-cell sorted into a 96-well plate containing complete medium, as described above. Clonal lines were expanded, and genomic DNA was extracted upon passaging for genotyping. Genotyping was performed by amplifying and sequencing the predicted deletion junction formed by the joining of the two cut-sites (Fig. 9a–ii and a–iii). To confirm the successful haploidization of the target area and ensure that the deleted fragment was not retained elsewhere in the genome, the copy number of the targeted region was quantified by digital droplet PCR as described below (Fig. 9b).

The sequences of the pegRNAs, ngRNAs and gRNAs used in this study are listed in Supplementary Data 2.

Genomic DNA isolation and PCR amplification

Genomic DNA isolation from cultured cells and mouse tissue was performed using the DNeasy Blood & Tissue Kit (Qiagen 69506) according to the manufacturer’s recommended protocol. PCR amplification was performed using CloneAmp™ HiFi PCR Premix (Takara Bio 639298) for genotyping the human TUBB locus and using DreamTaq Green PCR Master Mix 2X (Thermo Fisher Scientific K1082) in all other instances, following the manufacturer’s recommended protocols.

The sequences of the primers used for PCR amplification are listed in Supplementary Data 3. Uncropped and unprocessed scans of the gels are provided in the Source Data file.

RNA isolation, semiquantitative and quantitative RT-PCR

Total RNA isolation was performed using the RNeasy Mini Kit (Qiagen 74106) according to the manufacturer’s recommended protocol. 1 μg of total RNA was reverse transcribed to cDNA using SuperScript™ III Reverse Transcriptase (Thermo Fisher Scientific 18080044) following the manufacturer’s recommended protocol (20 μl reaction volume).

For semiquantitative RT-PCR, PCR amplification was performed on a 1:5 dilution of the template cDNA for 25 cycles using DreamTaq Green PCR Master Mix 2X (Thermo Fisher Scientific K1082) following the manufacturer’s recommended protocol.

Quantitative RT-PCR (qPCR) was performed on a 1:100 dilution of the template cDNA using PowerUp™ SYBR™ Green Master Mix (Thermo Fisher Scientific A25776) in a QuantStudio™ 3 Real-Time PCR System (Thermo Fisher Scientific, Applied Biosystems A28136). The relative expression levels of TUBB were compared using the ΔΔCt method, with GAPDH used as housekeeping control gene for normalization. The primers for TUBB are located within the deleted region in the TUBB haploid cell line, were previously used in qPCR⁹⁴, and their amplification efficiency was verified by generating a standard curve with four 1:10 serial dilutions of the template cDNA.

The sequences of the primers used for semiquantitative and quantitative RT-PCR experiments are listed in Supplementary Data 3. To prevent amplification of genomic DNA, at least one primer of each pair was designed to span an exon-exon boundary in the target mRNA. Uncropped and unprocessed scans of the gels are provided in the Source Data file.

Protein isolation and immunoblotting

To immunoblot for total β-tubulin and β-actin, TUBB^WT and TUBB^G308S cells were plated on 6-well plates (Corning 3506) and grown overnight in complete medium. Cells were resuspended in ice-cold 1X RIPA Lysis Buffer IV with Triton-X-100 pH 7.4 (Bio Basic RB4478) supplemented with Protease Inhibitor Cocktail (Sigma-Aldrich P8340). Protein concentration was determined using the Pierce™ BCA Protein Assay Kit (Thermo Fisher Scientific 23225) following the manufacturer’s recommended protocol. Next, an equal volume of 2X Laemmli Sample Buffer (Bio-Rad 1610737) was added to the samples, and lysates were boiled for 10 min. Equal amounts of protein lysates were separated on 4-20% Mini-PROTEAN® TGX™ Precast Protein Gels (Bio-Rad 4561096) in 1X Tris/Glycine/SDS buffer (Bio-Rad 1610732) and electro-blotted to nitrocellulose membranes in 1X Tris/Glycine buffer (Bio-Rad 1610734) supplemented with 20% methanol (Fisher Scientific A412-4). Membranes were blocked in 1X TBST [Tris-Buffered Saline (50 mM Tris pH 7.5, 150 mM NaCl) with 0.1% Tween® 20 detergent (Fisher Scientific BP337-500)] containing 5% non-fat powdered milk (Bio Basic NB0669) for 1 h at room temperature. Membranes were probed at 4 °C overnight with the primary antibodies indicated above, diluted in blocking buffer. Next, membranes were washed in 1X TBST and incubated for 1 h at room temperature with the HRP-conjugated anti-mouse secondary antibody indicated above (1:10000 dilution in blocking buffer). Chemiluminescence signals were detected using SuperSignal™ West Pico PLUS Chemiluminescent Substrate (Thermo Fisher Scientific 34580) and acquired using a ChemiDoc Imaging System (Bio-Rad). Immunoblots were exposed for durations ranging from 0.3 to 2 seconds without saturating the camera’s pixels.

Uncropped and unprocessed scans of the blots are provided in the Source Data file.

Copy number analysis by digital droplet PCR

Genomic copy number estimation of TUBB was conducted at The Centre for Applied Genomics (TCAG) at The Hospital for Sick Children, Toronto, Ontario, Canada. Copy number analysis was performed on the QX200 Droplet Digital PCR System (Bio-Rad) using a Custom TaqMan™ Copy Number Assay (Thermo Fisher Scientific 4400294). The primer sequences are shown in Supplementary Data 4. 50 ng of genomic DNA was digested with 5U of DraI (New England Biolabs R0129) in a 3 μl reaction with a 1-h incubation at 37 °C and no enzyme denaturation. The 20 μl copy number reaction mix consisted of 10 μl of 2X ddPCR™ Supermix for Probes (Bio-Rad 1863026), 1 μl of the copy number target assay (labeled with FAM), 1 μl of the VIC-labeled copy number reference assay (RNAseP, Thermo Fisher Scientific 4403326 or NPC1, Supplementary Data 4), 5 μl of water, and 3 μl of 16.7 ng/μl digested genomic DNA. Cycling conditions for the reaction were 95 °C for 10 min, followed by 45 cycles of 94 °C for 30 seconds and 60 °C for 1 min, 98 °C for 10 min, and finally a 4 °C hold on a Life Technologies Veriti thermal cycler. Data were analyzed using QuantaSoft v1.4 (Bio-Rad).

Copy number estimates were calculated as the ratio of the TUBB gene concentration to the reference gene concentration, multiplied by the number of reference gene copies per genome (i.e., 2).

Immunofluorescence

For indirect immunofluorescence in cultured cells, cells were plated on glass coverslips (Electron Microscopy Sciences 72230-01) in 12-well plates (Corning 3513) and treated according to each experiment. Cells were then washed once with PBS (Wisent Bioproducts 311-010-CL) and fixed in methanol (Fisher Scientific A412-4) at -20 °C for the time indicated for each experiment. Next, coverslips were blocked with 5% FBS (2, 4, 5, 9, 10, S1 and S11) or 3% BSA, 1% FBS (Figs. 3, 8 and S10) in PBST [PBS with 0.05% Tween® 20 detergent (Fisher Scientific BP337-500)]. Coverslips were subsequently incubated with the primary antibodies indicated above diluted in blocking buffer, in a humidified chamber overnight at 4 °C. Coverslips were then washed three times in PBST and incubated in a humidified chamber with the appropriate Alexa Fluor™-conjugated secondary antibodies (1:1000 dilution in blocking buffer) for 1 h at room temperature, and counterstained for 5 min with 1 μg/ml DAPI (Thermo Fisher Scientific 62248) or Hoechst 34580 (Thermo Fisher Scientific H21486) in PBS. Following three washes in PBST, coverslips were mounted on microscope glass slides (VWR 48312-401) using ProLong™ Gold Antifade Mountant (Thermo Fisher Scientific P36930).

For indirect immunofluorescence in mouse tissues, paraffin-embedded organs were sectioned on a Leica RM2235 microtome (Leica Biosystems) at a thickness of 7 μm for the brain and 4 μm for the kidney. The sections were mounted onto Premium Superfrost® Plus microscope slides (VWR 48311-703), dewaxed in xylene, and rehydrated in a series of graded ethanol to water. Antigen retrieval was performed in 10 mM trisodium citrate (pH 6.0) for 30 min at 96 °C. Slides were then washed three times with PBS (Wisent Bioproducts 311-010-CL) and the sections were circled using an Elite PAP Pen (Diagnostic BioSystems Inc. K039) to create a hydrophobic barrier that allows for the use of small solution volumes. Next, sections were permeabilized with 0.1% Triton-X-100 (Sigma-Aldrich T8787) in PBS for 10 min and blocked with 2.5% BSA in PBS with 0.1% Triton-X-100. Sections were subsequently incubated with the primary antibodies indicated above diluted in blocking buffer, in a humidified chamber overnight at 4 °C. Sections were then washed three times in PBS and incubated in a humidified chamber with the appropriate Alexa Fluor™-conjugated secondary antibody (1:1000 dilution in blocking buffer) for 2 h at room temperature. Sections were next counterstained with 1 μg/ml DAPI (Thermo Fisher Scientific 62248) in PBS for 20 min. Following three washes in PBS, slides were coverslipped with ProLong™ Gold Antifade Mountant (Thermo Fisher Scientific P10144).

Fluorescent imaging

Whole slide scans were acquired on a Pannoramic 250 Flash III slide scanner (3DHistech) using a 40×0.95 NA air objective (Zeiss). The instrument was operated in extended focus mode (seven focal planes spanning 5 μm axial distance) to capture the entire cell volume and all primary cilia that may lie on different focal planes.

Epifluorescence images were acquired on a Zeiss Axio Observer Z1 inverted microscope using a Axiocam 506 mono camera (Zeiss) with a 40×1.3 NA oil immersion objective (Zeiss), configured using Zeiss Zen 3.1 image acquisition software.

Confocal images in Figs. 2, 4, 5, 6c, 9, 10, S5, S6, S7 and S11 were acquired on a Leica SP8 Lightning confocal microscope (Leica Microsystems) using HyD detectors in combination with a 40×1.1 NA water immersion objective (Fig. 6c, S5, S6 and S7) or 63×1.3 NA oil immersion objective (other figures) controlled by Leica Application Suite X (LAS X) image acquisition software. Confocal images in Figs. 6a and 7 were acquired on a Leica Stellaris 5 confocal microscope (Leica Microsystems) using HyD detectors in combination with a 63×1.4 NA oil immersion objective controlled by Leica Application Suite X (LAS X) image acquisition software. Confocal images used for the MT incorporation assay in Fig. 2b were acquired on a Leica DMi8 spinning disk confocal microscope (Leica Microsystems) using a cooled electron-multiplying charged-coupled device (EM-CCD) camera (Hamamatsu) with a 40×1.3 NA oil immersion objective, configured using Volocity® image acquisition software (Quorum Technologies). Confocal images used for the colocalization analysis in Fig. 2c were acquired on a Leica Stellaris 5 confocal microscope (Leica Microsystems) using HyD detectors in combination with a 63×1.4 NA oil immersion objective controlled by Leica Application Suite X (LAS X) image acquisition software. Confocal images used for the ciliation assay in Figs. 6b and 7b were acquired on a Leica SP8 Lightning confocal microscope (Leica Microsystems) using HyD detectors in combination with a 40×1.1 NA water immersion objective controlled by Leica Application Suite X (LAS X) image acquisition software. Confocal images in Figs. 3, 8 and S10 were acquired on a Quorum WaveFX-X1 spinning disk confocal microscope (Quorum Technologies) using a cooled electron-multiplying charged-coupled device (EM-CCD) camera (Hamamatsu) with a 40×1.3 NA or a 63×1.4 NA oil immersion objective, configured using MetaMorph® image acquisition software (Molecular Devices). Unless otherwise indicated, consecutive Z-stacks were acquired at Nyquist intervals to capture the entire cell or tissue volume and all primary cilia that may lie on different focal planes. Unless otherwise indicated, maximum intensity projections of all the Z-planes are shown in the figures.

All live-cell imaging was acquired on a Quorum WaveFX-X1 spinning disk confocal microscope with a 63×1.4 NA oil immersion objective. Cells were kept at 37 °C with 5% CO₂ throughout imaging.

The Pannoramic 250 Flash III slide scanner, the Leica SP8 Lightning microscope, the Leica DMi8 microscope, and the Leica Stellaris 5 microscope are housed in The Imaging Facility at The Hospital for Sick Children, Toronto, Ontario, Canada. The Quorum WaveFX-X1 microscope is housed in the Centre for the Neurobiology of Stress, University of Toronto Scarborough, Ontario, Canada.

Histological staining and imaging

Hematoxylin and eosin (H&E) staining of mouse kidneys, including sectioning, was performed at The Centre for Phenogenomics (TCP; Toronto, Ontario, Canada) according to their standard protocols. Brightfield images were acquired on an Olympus BX51 microscope housed at TCP using a 40X objective and an Olympus DP71 camera.

H&E staining of mouse brains was performed according to a publicly available protocol from the University of Rochester Medical Center (www.urmc.rochester.edu/musculoskeletal-research/core-services/histology/protocols), using Mayer’s Hematoxylin (Sigma-Aldrich 51275) and Eosin (ELITechGroup SS-171C2). After staining, slides were coverslipped with Cytoseal™ XYL mounting medium (Thermo Fisher Scientific 8312-4). Brightfield whole-slide scans were acquired on a Pannoramic 250 Flash III slide scanner (3DHistech) equipped with a 40×0.95 NA air objective (Zeiss), housed in The Imaging Facility at The Hospital for Sick Children, Toronto, Ontario, Canada.

Histological image analysis

H&E-stained kidney slides were scanned at 20X magnification using a Axioscan 7 slide scanner (Zeiss) at The Centre for Phenogenomics Pathology Core (Toronto, Ontario, Canada). HALO image analysis software (Indica Labs) was used to measure the Bowman’s capsular space area (Fig. 7d) and the percentage of tubular cystic area (Fig. S8b). For the capsular space area, ten to twenty renal corpuscles were randomly selected per sample. For each, the inner surface of the Bowman’s capsule and the outer surface of the glomerulus were annotated using the Pen Annotation tool. The capsular space area was then calculated for each renal corpuscle by subtracting the glomerular area from the capsular area. To measure the percentage of tubular cystic area, the inner surface of each cyst and the outer boundary of the kidney section were annotated using the Flood Fill tool. The percentage of cystic area relative to the total kidney section area was then calculated.

Cerebellar molecular layer (ML) thickness was measured in Fiji using the Measure tool on images exported from whole-slide scans at a constant zoom level. For each mouse, one field of view from each of the cerebellar lobules III, IV/V, VI, and VII/IX was assessed, and ML thickness was measured at two distinct points.

Fluorescence intensity measurements

Measurements of fluorescence intensity of MTs, EB1-GFP comets, ninein, and γ−tubulin were generated by using the selection tools in Fiji⁹⁵ to encompass intensities from single frames or maximum intensity projections, followed by background intensity subtraction from each measurement. To quantify cells with an enhanced Ac-MT network, a unified intensity threshold was applied to the confocal images in Fiji and cells exceeding this threshold were identified.

MT incorporation assay

MT incorporation assay was performed in wild-type hTERT-RPE1 cells (ATCC CRL-4000). Cells were plated at 2 ×105 cells/well on glass coverslips in 12-well plates and transfected the following day with either wild-type or TUBB G308S pCMV3-TUBB:V5 plasmid, as described above. At 24 h post-transfection cells were fixed in methanol at -20 °C for 5 min and processed for immunofluorescence as described above, without staining cell nuclei.

MT incorporation was assessed on individual confocal images, acquired at 40X magnification. To avoid sampling biases, fields of view were randomly selected across the whole coverslip using only the α-tubulin channel, and all the low and medium-TUBB-V5-expressing cells in each field of view were scored. Analysis was performed blind to sample identity. MT incorporation was visually quantified by assigning each transfected cell to the “Incorporated”, “Partially incorporated”, or “Diffuse” category based on the filamentous, partially diffuse, or diffuse appearance of the fluorescent signal in the V5 channel.

Representative images shown in Fig. 2 are maximum intensity projections of confocal Z-stacks encompassing the entire cell volume.

Colocalization analysis

Colocalization analysis was performed on maximum intensity projections of confocal Z-stacks using the Coloc 2 plugin in Fiji⁹⁵. Mander’s coefficients for TUBB-V5 and α-tubulin were calculated from cells with medium and low TUBB-V5 expression in two-color images. For each cell, a fixed-size region of interest (ROI) was selected for analysis. To avoid sampling biases, the ROI was randomly selected using only the α-tubulin channel. Fluorescence intensity thresholds were applied prior to analysis to exclude background pixels.

MT dynamics assay

For EB1-GFP live imaging in complete medium, 3 ×105 cells were plated into 35-mm glass-bottom single-well dishes (MatTek P35G-1.5-14-C) and transiently transfected the following day with the EB1-GFP expressing plasmid, as described above. For EB1-GFP live imaging in serum-starved conditions, 4 ×105 cells were plated into 35mm glass-bottom single-well dishes and serum-starved the following day in Opti-MEM™ as described above, upon transient transfection with the EB1-GFP expressing plasmid. In both experiments, cells were imaged the day after plasmid transfection, around and no later than 24 h post-transfection.

The speed of EB1 comets was quantified from the maximum intensity projection of confocal time-lapse images of low and medium-EB1-GFP expressing cells using the TrackMate plugin⁹⁶ in Fiji⁹⁵. Images were subjected to bleach correction using the simple ratio method prior to track analysis utilizing the LoG Detector and the Simple LAP Tracker. The integrity of EB1-GFP tracks was verified individually. The number of EB1-GFP comets was quantified using the TrackMate⁹⁶ plugin in Fiji⁹⁵ from the maximum intensity projection of confocal Z-stack images of medium and low-EB1-GFP-expressing cells. To reduce biases from variations in cell shape, a fixed-sized area at the centrosomes was used to count the EB1-GFP comets. EB1-GFP comets were estimated as 1 µm-diameter spots that were further analyzed using LoG Detector.

MT regrowth assay

MT regrowth assay was performed by depolymerizing the MT network by treating cells with 10 μM nocodazole (Sigma-Aldrich SML1665) in complete medium for 20 min at 37 °C with 5% CO₂. After depolymerization, cells were washed extensively with PBS to remove the nocodazole and incubated with pre-warmed complete culture medium to induce MT nucleation and regrowth at 37 °C with 5% CO₂. After 2 or 4 min of MT regrowth, cells were fixed in methanol at -20 °C for 10 min.

MT curvature analysis

MT curvature analysis was generated using the Kappa plugin⁹⁷ in Fiji⁹⁵. Semiautomated tracking of individual MTs was done on confocal slices using open B-spline configurations. The initialization curve is fitted using point distance minimization algorithm.

Ciliation assays

For primary skin fibroblasts, 2 ×105 cells/well were plated on glass coverslips in 12-well plates. The following day, cells (~95% confluence) were serum starved in Opti-MEM™ as described above for 72 h to induce ciliation, before fixation in methanol at -20 °C for 20 min and processing for immunofluorescence.

For hTERT-RPE1 cells, 3 ×105 cells/well were plated on glass coverslips in 12-well plates. The following day, cells (~90% confluence) were serum starved in Opti-MEM™ as described above for 2, 4, 6 (Fig. 8a), 24 (Fig. 8d and S10) or 48 (Figs. 5, 9 and S1) h to induce ciliation, before fixation in methanol at -20 °C for 20 min and processing for immunofluorescence.

For the overexpression experiments, wild-type hTERT-RPE1 cells (ATCC CRL-4000) (Fig. 10b and S11b) or TUBB^G308S cells (Fig. S11a) were plated at 2 ×105 cells/well on glass coverslips in 12-well plates. The following day, cells were transfected as described above with either wild-type or mutant pCMV3-TUBB:V5 plasmid. At 24 h post-transfection (~90% confluence) ciliation was induced by serum starving cells in Opti-MEM™ as described above for 24 h, before fixation in methanol at -20 °C for 20 min and processing for immunofluorescence.

The percentage of ciliated cells, the length of primary cilia and MT incorporation in the overexpression screen were assessed on whole slide scans visualized using CaseViewer 2.4 or SlideViewer 2.5 (3DHistech). For each coverslip, three different fields (ROI: regions of interest) were selected for analysis. To avoid sampling biases, the three ROIs were randomly selected across the whole coverslip using only the DAPI or Hoechst channel, and all the cells in each ROI were scored. All analyses were performed at the same zoom level (90X) and blind to sample identity. The percentage of ciliated cells was calculated as the number of primary cilia divided by the total number of DAPI- or Hoechst-labeled cell nuclei, as scored manually using the Marker Counter tool in CaseViewer. The length of primary cilia was measured using the Measurement Annotation tool in CaseViewer, approximating primary cilia as linear structures. In the overexpression experiments, only low and medium-TUBB-V5-expressing cells were scored for the analysis of the transfected cell population. MT incorporation was visually quantified by assigning each transfected cell to the “Incorporated”, “Partially incorporated” or “Diffuse” category based on the filamentous, partially diffuse or diffuse appearance of the fluorescent signal in the V5 channel. The percentage of ciliated cells in Fig. S1b was manually scored on an Agilent BioTek Lionheart LX Automated Microscope with a 40×0.6 NA air objective housed at The Hospital for Sick Children, Toronto, Ontario, Canada.

Ciliation in mouse cerebellar sections was assessed on maximum intensity projections of confocal Z-stacks encompassing the entire thickness of the tissue, acquired at 40X magnification. For each sample, three (Fig. S5b) or four (Fig. 6b and S5d) different fields of view from cerebellar lobules I/II, IV/V and IX were acquired for analysis. To avoid sampling biases, the fields of view were randomly selected using only the DAPI channel. All analyses were performed blind to sample identity. Ciliation levels were determined by manually scoring the number of primary cilia either per region (GCL, ML, and white matter) or per cell type (Purkinje cells, microglia). For region-based quantification, the GCL, ML, or white matter area was manually delimited as a region of interest (ROI), cilia were counted within that ROI, and normalized by area.

Ciliation in mouse kidney sections was assessed on maximum intensity projections of confocal Z-stacks encompassing the entire thickness of the tissue, acquired at 40X magnification. For each sample, four different fields of view were acquired for analysis. To avoid sampling biases, the four fields of view were randomly selected across the whole coverslip using only the DAPI channel. All analyses were performed blind to sample identity. Ciliation levels were quantified by manually scoring the number of primary cilia per the area of the field of view (291² μm² for Fig. 7b; 166² μm² for Fig. S7b).

Length of motile ependymal cilia lining the third ventricle of the mouse brain was measured in Fiji⁹⁵ using the Measure function on maximum intensity projections of confocal Z-stacks encompassing the full thickness of the tissue. For each sample, three different fields of view were acquired for analysis, and cilia length was measured at three points per field of view.

Representative images shown in the figures are maximum intensity projections of confocal Z-stacks encompassing the entire cell or tissue volume.

Quantification of VGLUT2-positive puncta

VGLUT2-positive puncta were quantified in the ML of the mouse cerebellum from maximum projection images using Fiji⁹⁵. A consistent intensity threshold was applied to remove background signal, and puncta were detected using the Analyze Particles function with a size filter of 2-6 µm². Four fields of view per mouse were acquired from cerebellar lobules I/II, IV/V, and IX, with the ML manually delimited as a region of interest (ROI), as shown in Fig. 6c.

Transferrin uptake assay

Transferrin uptake assay was performed in cells that had been serum-starved for 24 h. Cells were incubated with 50 µg/mL Alexa Fluor™ 488-conjugated human transferrin (Jackson ImmunoResearch 009-540-050) on ice for 30 min. After incubation, the transferrin-containing medium was removed, and cells were further incubated in complete culture medium for 45 min to 1 h. Cells were washed with acid wash buffer (0.5 M NaCl in 0.2 M acetic acid) to remove surface-bound transferrin and fixed in 4% paraformaldehyde (PFA) for 10 min. Following fixation, cells were washed with PBS and counterstained with 1 µg/mL DAPI (Thermo Fisher Scientific, 62248) for 5 min.

Centrosome migration measurement

The migration distance of ciliated centrosomes was measured from confocal Z-stack slices (0.2 µm apart) encompassing the entire thickness of the cell. Stacks were displayed as XY and XZ views using XYZ projection tool in Fiji⁹⁵. Straight line function was used on XZ projection to measure the distance between CEP135 and the apex of DAPI-stained nucleus.

Protein structural analyses

All 3D protein structures were rendered using the PyMOL Molecular Graphics System version 2.0 (Schrödinger, LLC). Mutant protein stability was calculated as change in Gibbs folding free energy (ΔΔG) using FoldX software⁹⁸ with default settings. ΔΔG was calculated for every possible glycine to serine substitution over the four indicated protein structures (PDB IDs: 4i4t, 5nqu, 5yl2, 5ca1).

Cell cycle analysis

Cell cycle analysis of cycling and serum-starved TUBB^WT and TUBB^G308S hTERT-RPE1 cells was performed by flow cytometry using propidium iodide (PI) to measure DNA content. For cycling samples, 1.5 × 105 cells/well were plated in complete medium in 12-well plates. For serum-starved samples, the same conditions as the ciliation assay were used: 3 × 105 cells/well were plated in complete medium in 12-well plates, and serum-starved in Opti-MEM™ for 48 h the following day, as described above. In preparation for flow cytometry ~2 × 106 cells/sample were collected, resuspended in 50 μl eBioscience™ Flow Cytometry Staining Buffer (Thermo Fisher Scientific 00-4222-26) and fixed with 1 ml of ice-cold 80% ethanol overnight at 4 °C. Fixed cells were then collected, incubated with 500 μl of 2 mg/ml RNase A (Qiagen 19101) in HBSS with calcium and magnesium (Thermo Fisher Scientific 14025-092) for 5 min at room temperature, and stained for 30 min at room temperature with 500 μl of 0.1 mg/ml PI (Thermo Fisher Scientific P3566) in HBSS 0.6% NP-40 (Sigma-Aldrich I8896). Cells were next filtered through a 40 μm cell strainer (VWR CA21008-949) and flow cytometry was performed using a BD LSR II analyzer (BD Biosciences). Analysis of PI was performed using 561 nm (yellow-green) laser excitation and a 575/26 nm bandpass filter. Flow cytometry data were gated for single nuclei, analyzed, and plotted using FlowJo™ software version 10.8 (BD Biosciences).

Statistics and reproducibility

Data normality was assessed using the Anderson-Darling or D’Agostino-Pearson omnibus tests. For normally distributed data, statistical significance was determined using a two-tailed Student’s t-test or an ordinary one-way ANOVA followed by Tukey’s or Šídák’s multiple comparisons test. For non-normally distributed data, a non-parametric two-tailed Mann-Whitney U test was used. Statistical analyses were conducted using GraphPad Prism version 9.4.1 (GraphPad Software) or BioRender Graph (using R version 4.2.2) with p-value (p) <0.05 considered statistically significant.

Sample sizes were determined based on literature precedents in the field and not predetermined by statistical methods. No data were excluded from the analyses. Experiments were not randomized, as experimental groups were defined by genotype. Investigators were blinded to sample identity for cell culture and mouse experiments, as detailed in each relevant methods section. Further details on study design, randomization and blinding are provided in the Reporting Summary.

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.

Data availability

All experimental data that support the findings of this study are available in the main text or in the supplementary information files. MRI data generated in this study are not publicly available due to data sharing restrictions. Relevant genotype-phenotype data are provided within the article. Whole genome sequencing data cannot be publicly shared, as the study protocol permits publication of findings and results but does not allow deposition of raw data in public repositories. Access to whole genome sequencing data may be granted to qualified researchers upon request to the corresponding author for non-commercial research purposes only, subject to approval by the relevant ethics board. We aim to respond to requests within 20 business days. The following Protein Data Bank (PDB) entries were used in this study: 6i2i, 6o2s, 4i4t, 5nqu, 5yl2, 5ca1. The following NCBI Sequence Read Archive (SRA) experiment ID was accessed for this study: SRX858057. Source data are provided with this paper.

References

Breuss, M. W., Leca, I., Gstrein, T., Hansen, A. H. & Keays, D. A. Tubulins and brain development - The origins of functional specification. Mol. Cell Neurosci. 84, 58–67 (2017).
Article CAS PubMed Google Scholar
Maillard, C. et al. Tubulin mutations in human neurodevelopmental disorders. Semin Cell Dev. Biol. https://doi.org/10.1016/j.semcdb.2022.07.009 (2022).
Romaniello, R. et al. Tubulin genes and malformations of cortical development. Eur. J. Med Genet 61, 744–754 (2018).
Article PubMed Google Scholar
Romero, D. M., Bahi-Buisson, N. & Francis, F. Genetics and mechanisms leading to human cortical malformations. Semin Cell Dev. Biol. 76, 33–75 (2018).
Article CAS PubMed Google Scholar
Goodson, H. V. & Jonasson, E. M. Microtubules and microtubule-associated proteins. Cold Spring Harb Perspect Biol. 10, 10.1101/cshperspect.a022608 (2018).
Hoyle, H. D. & Raff, E. C. Two Drosophila beta tubulin isoforms are not functionally equivalent. J. Cell Biol. 111, 1009–1026 (1990).
Article CAS PubMed Google Scholar
Saillour, Y. et al. Beta tubulin isoforms are not interchangeable for rescuing impaired radial migration due to Tubb3 knockdown. Hum. Mol. Genet 23, 1516–1526 (2014).
Article CAS PubMed Google Scholar
Vemu, A., Atherton, J., Spector, J. O., Moores, C. A. & Roll-Mecak, A. Tubulin isoform composition tunes microtubule dynamics. Mol. Biol. Cell 28, 3564–3572 (2017).
Article CAS PubMed PubMed Central Google Scholar
Miyatake, S. et al. Expanding the phenotypic spectrum of TUBB4A-associated hypomyelinating leukoencephalopathies. Neurology 82, 2230–2237 (2014).
Article CAS PubMed Google Scholar
Simons, C. et al. A de novo mutation in the beta-tubulin gene TUBB4A results in the leukoencephalopathy hypomyelination with atrophy of the basal ganglia and cerebellum. Am. J. Hum. Genet 92, 767–773 (2013).
Article CAS PubMed PubMed Central Google Scholar
Smith, B. N. et al. Exome-wide rare variant analysis identifies TUBA4A mutations associated with familial ALS. Neuron 84, 324–331 (2014).
Article CAS PubMed PubMed Central Google Scholar
Chen, B. et al. The comprehensive mutational and phenotypic spectrum of TUBB8 in female infertility. Eur. J. Hum. Genet 27, 300–307 (2019).
Article CAS PubMed Google Scholar
Stoupa, A. et al. TUBB1 mutations cause thyroid dysgenesis associated with abnormal platelet physiology. EMBO Mol. Med. 10, 10.15252/emmm.201809569 (2018).
Oud, M. M., Lamers, I. J. & Arts, H. H. Ciliopathies: Genetics in Pediatric Medicine. J. Pediatr. Genet 6, 18–29 (2017).
Article CAS PubMed Google Scholar
Wheway, G. & Mitchison, H. M. Genomics england research, c. opportunities and challenges for molecular understanding of ciliopathies-The 100,000 Genomes Project. Front Genet 10, 127 (2019).
Article CAS PubMed PubMed Central Google Scholar
Van De Weghe, J. C., Gomez, A. & Doherty, D. The joubert-meckel-nephronophthisis spectrum of ciliopathies. Annu Rev. Genomics Hum. Genet 23, 301–329 (2022).
Article PubMed Google Scholar
Guemez-Gamboa, A., Coufal, N. G. & Gleeson, J. G. Primary cilia in the developing and mature brain. Neuron 82, 511–521 (2014).
Article CAS PubMed PubMed Central Google Scholar
Wheway, G., Nazlamova, L. & Hancock, J. T. Signaling through the Primary Cilium. Front Cell Dev. Biol. 6, 8 (2018).
Article PubMed PubMed Central Google Scholar
Mirvis, M., Stearns, T. & James Nelson, W. Cilium structure, assembly, and disassembly regulated by the cytoskeleton. Biochem J. 475, 2329–2353 (2018).
Article CAS PubMed Google Scholar
Park, S. M., Jang, H. J. & Lee, J. H. Roles of primary cilia in the developing brain. Front Cell Neurosci. 13, 218 (2019).
Article CAS PubMed PubMed Central Google Scholar
Romani, M., Micalizzi, A. & Valente, E. M. Joubert syndrome: congenital cerebellar ataxia with the molar tooth. Lancet Neurol. 12, 894–905 (2013).
Article PubMed Google Scholar
Schroder, S. et al. Heterozygous truncating variants in SUFU cause congenital ocular motor apraxia. Genet Med 23, 341–351 (2021).
Article PubMed Google Scholar
Serpieri, V. et al. SUFU haploinsufficiency causes a recognisable neurodevelopmental phenotype at the mild end of the Joubert syndrome spectrum. J. Med Genet 59, 888–894 (2022).
Article CAS PubMed Google Scholar
Hartill, V., Szymanska, K., Sharif, S. M., Wheway, G. & Johnson, C. A. Meckel-Gruber Syndrome: An Update on Diagnosis, Clinical Management, and Research Advances. Front Pediatr. 5, 244 (2017).
Article PubMed PubMed Central Google Scholar
Vilboux, T. et al. Molecular genetic findings and clinical correlations in 100 patients with Joubert syndrome and related disorders prospectively evaluated at a single center. Genet Med 19, 875–882 (2017).
Article PubMed PubMed Central Google Scholar
Kasahara, K. & Inagaki, M. Primary ciliary signaling: links with the cell cycle. Trends Cell Biol. 31, 954–964 (2021).
Article CAS PubMed Google Scholar
Sorokin, S. Centrioles and the formation of rudimentary cilia by fibroblasts and smooth muscle cells. J. Cell Biol. 15, 363–377 (1962).
Article CAS PubMed PubMed Central Google Scholar
Molla-Herman, A. et al. The ciliary pocket: an endocytic membrane domain at the base of primary and motile cilia. J. Cell Sci. 123, 1785–1795 (2010).
Article CAS PubMed Google Scholar
Ghossoub, R., Molla-Herman, A., Bastin, P. & Benmerah, A. The ciliary pocket: a once-forgotten membrane domain at the base of cilia. Biol. Cell 103, 131–144 (2011).
Article PubMed Google Scholar
Baudoin, J. P. et al. Tangentially migrating neurons assemble a primary cilium that promotes their reorientation to the cortical plate. Neuron 76, 1108–1122 (2012).
Article CAS PubMed Google Scholar
Westlake, C. J. et al. Primary cilia membrane assembly is initiated by Rab11 and transport protein particle II (TRAPPII) complex-dependent trafficking of Rabin8 to the centrosome. Proc. Natl Acad. Sci. USA 108, 2759–2764 (2011).
Article ADS CAS PubMed PubMed Central Google Scholar
Firestone, A. J. et al. Small-molecule inhibitors of the AAA+ ATPase motor cytoplasmic dynein. Nature 484, 125–129 (2012).
Article ADS CAS PubMed PubMed Central Google Scholar
Wu, C. T., Chen, H. Y. & Tang, T. K. Myosin-Va is required for preciliary vesicle transportation to the mother centriole during ciliogenesis. Nat. Cell Biol. 20, 175–185 (2018).
Article CAS PubMed Google Scholar
Lu, Q. et al. Early steps in primary cilium assembly require EHD1/EHD3-dependent ciliary vesicle formation. Nat. Cell Biol. 17, 228–240 (2015).
Article CAS PubMed PubMed Central Google Scholar
Pitaval, A. et al. Microtubule stabilization drives 3D centrosome migration to initiate primary ciliogenesis. J. Cell Biol. 216, 3713–3728 (2017).
Article CAS PubMed PubMed Central Google Scholar
Breuss, M. et al. Mutations in the beta-tubulin gene TUBB5 cause microcephaly with structural brain abnormalities. Cell Rep. 2, 1554–1562 (2012).
Article CAS PubMed PubMed Central Google Scholar
Isrie, M. et al. Mutations in either TUBB or MAPRE2 cause circumferential skin creases Kunze type. Am. J. Hum. Genet 97, 790–800 (2015).
Article CAS PubMed PubMed Central Google Scholar
Shah, Y. B. et al. Inherited bone marrow failure with macrothrombocytopenia due to germline tubulin beta class I (TUBB) variant. Br. J. Haematol. 200, 222–228 (2023).
Article CAS PubMed Google Scholar
Leandro-Garcia, L. J. et al. Tumoral and tissue-specific expression of the major human beta-tubulin isotypes. Cytoskeleton (Hoboken) 67, 214–223 (2010).
Article CAS PubMed Google Scholar
Lowe, J., Li, H., Downing, K. H. & Nogales, E. Refined structure of alpha beta-tubulin at 3.5 A resolution. J. Mol. Biol. 313, 1045–1057 (2001).
Article CAS PubMed Google Scholar
Anzalone, A. V. et al. Search-and-replace genome editing without double-strand breaks or donor DNA. Nature 576, 149–157 (2019).
Article ADS CAS PubMed PubMed Central Google Scholar
Wang, Y., Tian, G., Cowan, N. J. & Cabral, F. Mutations affecting beta-tubulin folding and degradation. J. Biol. Chem. 281, 13628–13635 (2006).
Article CAS PubMed Google Scholar
Luders, J. & Stearns, T. Microtubule-organizing centres: a re-evaluation. Nat. Rev. Mol. Cell Biol. 8, 161–167 (2007).
Article PubMed Google Scholar
Gomez, A. E., Christman, A. K., Van De Weghe, J. C., Finn, M. & Doherty, D. Systematic analysis of cilia characteristics and Hedgehog signaling in five immortal cell lines. PLoS One 17, e0266433 (2022).
Article PubMed PubMed Central Google Scholar
Bowie, E. & Goetz, S. C. TTBK2 and primary cilia are essential for the connectivity and survival of cerebellar Purkinje neurons. Elife 9, https://doi.org/10.7554/eLife.51166 (2020).
Vaisse, C., Reiter, J. F. & Berbari, N. F. Cilia and Obesity. Cold Spring Harb Perspect Biol 9, https://doi.org/10.1101/cshperspect.a028217 (2017).
Ji, W. et al. Ependymal cilia: Physiology and role in hydrocephalus. Front Mol. Neurosci. 15, 927479 (2022).
Article CAS PubMed PubMed Central Google Scholar
McConnachie, D. J., Stow, J. L. & Mallett, A. J. Ciliopathies and the kidney: A review. Am. J. Kidney Dis. 77, 410–419 (2021).
Article CAS PubMed Google Scholar
Ko, J. Y. & Park, J. H. Mouse models of polycystic kidney disease induced by defects of ciliary proteins. BMB Rep. 46, 73–79 (2013).
Article CAS PubMed PubMed Central Google Scholar
Rusan, N. M., Fagerstrom, C. J., Yvon, A. M. & Wadsworth, P. Cell cycle-dependent changes in microtubule dynamics in living cells expressing green fluorescent protein-alpha tubulin. Mol. Biol. Cell 12, 971–980 (2001).
Article CAS PubMed PubMed Central Google Scholar
Forth, S. & Kapoor, T. M. The mechanics of microtubule networks in cell division. J. Cell Biol. 216, 1525–1531 (2017).
Article CAS PubMed PubMed Central Google Scholar
Driskell, O. J., Mironov, A., Allan, V. J. & Woodman, P. G. Dynein is required for receptor sorting and the morphogenesis of early endosomes. Nat. Cell Biol. 9, 113–120 (2007).
Article CAS PubMed Google Scholar
Coombes, C. et al. Mechanism of microtubule lumen entry for the alpha-tubulin acetyltransferase enzyme alphaTAT1. Proc. Natl Acad. Sci. USA 113, E7176–E7184 (2016).
Article CAS PubMed PubMed Central Google Scholar
Janke, C. & Montagnac, G. Causes and consequences of microtubule acetylation. Curr. Biol. 27, R1287–R1292 (2017).
Article CAS PubMed Google Scholar
Xu, Z. et al. Microtubules acquire resistance from mechanical breakage through intralumenal acetylation. Science 356, 328–332 (2017).
Article ADS CAS PubMed PubMed Central Google Scholar
Tischfield, M. A., Cederquist, G. Y., Gupta, M. L. Jr. & Engle, E. C. Phenotypic spectrum of the tubulin-related disorders and functional implications of disease-causing mutations. Curr. Opin. Genet Dev. 21, 286–294 (2011).
Article CAS PubMed PubMed Central Google Scholar
Aiken, J., Buscaglia, G., Aiken, A. S., Moore, J. K. & Bates, E. A. Tubulin mutations in brain development disorders: Why haploinsufficiency does not explain TUBA1A tubulinopathies. Cytoskeleton (Hoboken) 77, 40–54 (2020).
Article CAS PubMed Google Scholar
Breuss, M. et al. Mutations in the murine homologue of TUBB5 cause microcephaly by perturbing cell cycle progression and inducing p53-associated apoptosis. Development 143, 1126–1133 (2016).
CAS PubMed Google Scholar
Erwood, S. et al. Saturation variant interpretation using CRISPR prime editing. Nat. Biotechnol. 40, 885–895 (2022).
Article CAS PubMed Google Scholar
Roostalu, J. & Surrey, T. Microtubule nucleation: beyond the template. Nat. Rev. Mol. Cell Biol. 18, 702–710 (2017).
Article CAS PubMed Google Scholar
Ayukawa, R. et al. GTP-dependent formation of straight tubulin oligomers leads to microtubule nucleation. J. Cell Biol. 220, 10.1083/jcb.202007033 (2021).
Gardner, M. K. Straightening up is required to nucleate new microtubules. J. Cell Biol. 220, https://doi.org/10.1083/jcb.202102123 (2021).
Duellberg, C. et al. Reconstitution of a hierarchical +TIP interaction network controlling microtubule end tracking of dynein. Nat. Cell Biol. 16, 804–811 (2014).
Article CAS PubMed Google Scholar
Reid, T. A. et al. Structural state recognition facilitates tip tracking of EB1 at growing microtubule ends. Elife 8, 10.7554/eLife.48117 (2019).
Denarier, E., Brousse, C., Sissoko, A., Andrieux, A. & Boscheron, C. A neurodevelopmental TUBB2B beta-tubulin mutation impairs Bim1 (yeast EB1)-dependent spindle positioning. Biol Open 8, 10.1242/bio.038620 (2019).
Kulic, I. M. et al. The role of microtubule movement in bidirectional organelle transport. Proc. Natl Acad. Sci. USA 105, 10011–10016 (2008).
Article ADS CAS PubMed PubMed Central Google Scholar
Portran, D., Schaedel, L., Xu, Z., Thery, M. & Nachury, M. V. Tubulin acetylation protects long-lived microtubules against mechanical ageing. Nat. Cell Biol. 19, 391–398 (2017).
Article CAS PubMed PubMed Central Google Scholar
Tran, A. D. et al. HDAC6 deacetylation of tubulin modulates dynamics of cellular adhesions. J. Cell Sci. 120, 1469–1479 (2007).
Article CAS PubMed Google Scholar
Lancaster, M. A. et al. Defective Wnt-dependent cerebellar midline fusion in a mouse model of Joubert syndrome. Nat. Med 17, 726–731 (2011).
Article CAS PubMed PubMed Central Google Scholar
Hynes, A. M. et al. Murine joubert syndrome reveals hedgehog signaling defects as a potential therapeutic target for nephronophthisis. Proc. Natl Acad. Sci. USA 111, 9893–9898 (2014).
Article ADS CAS PubMed PubMed Central Google Scholar
Suciu, S. K., Long, A. B. & Caspary, T. Smoothened and ARL13B are critical in mouse for superior cerebellar peduncle targeting. Genetics 218, 10.1093/genetics/iyab084 (2021).
Ramsbottom, S. A. et al. Mouse genetics reveals barttin as a genetic modifier of Joubert syndrome. Proc. Natl Acad. Sci. USA 117, 1113–1118 (2020).
Article ADS CAS PubMed Google Scholar
Lee, J. E. et al. CEP41 is mutated in Joubert syndrome and is required for tubulin glutamylation at the cilium. Nat. Genet 44, 193–199 (2012).
Article CAS PubMed PubMed Central Google Scholar
Dodd, D. O. et al. Ciliopathy patient variants reveal organelle-specific functions for TUBB4B in axonemal microtubules. Science 384, eadf5489 (2024).
Article CAS PubMed Google Scholar
Sewell, M. T., Legue, E. & Liem, K. F., Jr. Tubb4b is required for multi-ciliogenesis in the mouse. Development 151, 10.1242/dev.201819 (2024).
Nygaard, M. B., Almstrup, K., Lindbaek, L., Christensen, S. T. & Svingen, T. Cell context-specific expression of primary cilia in the human testis and ciliary coordination of Hedgehog signalling in mouse Leydig cells. Sci. Rep. 5, 10364 (2015).
Article ADS PubMed PubMed Central Google Scholar
Girardet, L., Augiere, C., Asselin, M. P. & Belleannee, C. Primary cilia: biosensors of the male reproductive tract. Andrology 7, 588–602 (2019).
Article PubMed Google Scholar
Gadadhar, S., Hirschmugl, T. & Janke, C. The tubulin code in mammalian sperm development and function. Semin Cell Dev. Biol. 137, 26–37 (2023).
Article CAS PubMed Google Scholar
Bahi-Buisson, N. & Maillard, C. Tubulinopathies Overview. in GeneReviews((R)) (eds. Adam, M. P. et al.) (Seattle (WA), 1993).
Kernohan, K. D. et al. Diagnostic clarity of exome sequencing following negative comprehensive panel testing in the neonatal intensive care unit. Am. J. Med Genet A 176, 1688–1691 (2018).
Article PubMed Google Scholar
DePristo, M. A. et al. A framework for variation discovery and genotyping using next-generation DNA sequencing data. Nat. Genet 43, 491–498 (2011).
Article CAS PubMed PubMed Central Google Scholar
Cunningham, F. et al. Ensembl 2022. Nucleic Acids Res 50, D988–D995 (2022).
Article CAS PubMed Google Scholar
Paila, U., Chapman, B. A., Kirchner, R. & Quinlan, A. R. GEMINI: integrative exploration of genetic variation and genome annotations. PLoS Comput Biol. 9, e1003153 (2013).
Article ADS CAS PubMed PubMed Central Google Scholar
Pedersen, B. S., Layer, R. M. & Quinlan, A. R. Vcfanno: fast, flexible annotation of genetic variants. Genome Biol. 17, 118 (2016).
Article PubMed PubMed Central Google Scholar
Karczewski, K. J. et al. The mutational constraint spectrum quantified from variation in 141,456 humans. Nature 581, 434–443 (2020).
Article ADS CAS PubMed PubMed Central Google Scholar
Erwood, S. et al. Modeling Niemann-Pick disease type C in a human haploid cell line allows for patient variant characterization and clinical interpretation. Genome Res 29, 2010–2019 (2019).
Article CAS PubMed PubMed Central Google Scholar
Gertsenstein, M. et al. Efficient generation of germ line transmitting chimeras from C57BL/6N ES cells by aggregation with outbred host embryos. PLoS One 5, e11260 (2010).
Article ADS PubMed PubMed Central Google Scholar
Behringer, R., Gertsenstein, M., Nagy, K. V. & Nagy, A. Manipulating the mouse embryo: a laboratory manual, xxii, 814 pages (Cold Spring Harbor Laboratory Press, Cold Spring Harbor, New York, 2014).
Haapaniemi, E., Botla, S., Persson, J., Schmierer, B. & Taipale, J. CRISPR-Cas9 genome editing induces a p53-mediated DNA damage response. Nat. Med 24, 927–930 (2018).
Article CAS PubMed Google Scholar
Bowden, A. R. et al. Parallel CRISPR-Cas9 screens clarify impacts of p53 on screen performance. Elife 9, https://doi.org/10.7554/eLife.55325 (2020).
Geisinger, J. M. & Stearns, T. CRISPR/Cas9 treatment causes extended TP53-dependent cell cycle arrest in human cells. Nucleic Acids Res 48, 9067–9081 (2020).
Article CAS PubMed PubMed Central Google Scholar
Robinson, J. T. et al. Integrative genomics viewer. Nat. Biotechnol. 29, 24–26 (2011).
Article CAS PubMed PubMed Central Google Scholar
Labun, K. et al. CHOPCHOP v3: expanding the CRISPR web toolbox beyond genome editing. Nucleic Acids Res 47, W171–W174 (2019).
Article CAS PubMed PubMed Central Google Scholar
Gasic, I., Boswell, S. A. & Mitchison, T. J. Tubulin mRNA stability is sensitive to change in microtubule dynamics caused by multiple physiological and toxic cues. PLoS Biol. 17, e3000225 (2019).
Article CAS PubMed PubMed Central Google Scholar
Schindelin, J. et al. Fiji: an open-source platform for biological-image analysis. Nat. Methods 9, 676–682 (2012).
Article CAS PubMed Google Scholar
Tinevez, J. Y. et al. TrackMate: An open and extensible platform for single-particle tracking. Methods 115, 80–90 (2017).
Article CAS PubMed Google Scholar
Mary, H. & Brouhard, G. J. Kappa (κ): Analysis of curvature in biological image data using B-splines. Preprint at bioRxiv, 852772, https://doi.org/10.1101/852772 (2019).
Schymkowitz, J. et al. The FoldX web server: an online force field. Nucleic Acids Res 33, W382–W388 (2005).
Article CAS PubMed PubMed Central Google Scholar

Download references

Acknowledgements

We thank Paul Paroutis (The Imaging Facility, The Hospital for Sick Children, Toronto) for his assistance with whole slide scanning. We thank Emily Reddy (Flow Cytometry Facility, The Hospital for Sick Children, Toronto) for her help with flow cytometry data acquisition and analysis. We thank Daniel Durocher (Department of Molecular Genetics, University of Toronto) for sharing the p53-null hTERT-RPE1 cells. We thank Marina Gertsenstein and Monica Pereira (Model Production Core, The Centre for Phenogenomics, Toronto) for their technical support with chimeric mice generation. We thank Mohammad Eskandarian and Vivian Bradaschia (Pathology Core, The Centre for Phenogenomics, Toronto) for their assistance with histopathological analysis and image analysis of mouse kidney samples. We thank Elise Héon (Genetics and Genome Biology Program, The Hospital for Sick Children, Toronto) for sharing the control primary skin fibroblasts and for helpful discussions. We thank Dan Doherty (Center for Integrative Brain Research, Seattle Children’s Research Institute) and Julie Van De Weghe (Department of Pediatrics, University of Washington) for extensive and helpful discussions throughout this project. We are grateful to all the current and past members of the Cohn laboratory and the Ivakine laboratory for helpful discussions and input. In particular, we thank Hong Anh Truong for her dedicated care and assistance with mouse colony maintenance. This study was supported by the Nicol Family Foundation and the SickKids Foundation, with funding awarded to E.A.I. and R.D.C. and by a grant from the Rare Disease Foundation and the BC Children’s Hospital Foundation (grant #2808 to A.M.). A.M. is a recipient of an Ontario Trillium Scholarship (OTS). S.O. is a recipient of a University of Toronto Provost’s Postdoctoral Award and is supported by the Canadian Institutes of Health Research (CIHR)-REDI award (grant #ED6-190720). Research in R.E.H.’s laboratory is supported by a grant from the Canadian Institutes of Health Research (CIHR, grant #PJT-166084). The Centre for the Neurobiology of Stress at University of Toronto Scarborough is supported by a grant from the Canada Foundation for Innovation (CFI, grant #493864).

Author information

Authors and Affiliations

Genetics and Genome Biology Program, The Hospital for Sick Children, Toronto, ON, Canada
Antonio Mollica, Georgiana Forguson, Sydney Steiman, Sonia L. Evagelou, Lu Yi Li, Aideen Teeling, Kyle Lindsay, Steven Erwood, Shagana Visuvanathan, Anjali Vig, Ronald D. Cohn & Evgueni A. Ivakine
Department of Biochemistry, University of Toronto, Toronto, ON, Canada
Antonio Mollica & Julie D. Forman-Kay
Department of Biological Sciences, University of Toronto Scarborough, Toronto, ON, Canada
Safia Omer & Rene E. Harrison
Department of Physiology, University of Toronto, Toronto, ON, Canada
Sydney Steiman & Evgueni A. Ivakine
Department of Biostatistics, Harvard Chan School of Public Health, Boston, MA, USA
Serhiy Naumenko
The Centre for Applied Genomics, The Hospital for Sick Children, Toronto, ON, Canada
Susan Walker
Department of Molecular Genetics, University of Toronto, Toronto, ON, Canada
Steven Erwood, Ronald D. Cohn & Evgueni A. Ivakine
Molecular Medicine Program, The Hospital for Sick Children, Toronto, ON, Canada
Robert M. Vernon & Julie D. Forman-Kay
Department of Neurology and Developmental Medicine, Kennedy Krieger Institute, Baltimore, MD, USA
Benjamin Akman & Constance Smith-Hicks
Department of Neurology, Johns Hopkins University School of Medicine, Baltimore, MD, USA
Constance Smith-Hicks
Division of Neuroradiology, Department of Diagnostic and Interventional Radiology, The Hospital for Sick Children, Toronto, ON, Canada
Manohar Shroff & Vivek Pai
Department of Medical Imaging, University of Toronto, Toronto, ON, Canada
Manohar Shroff & Vivek Pai
Department of Pediatrics, The Hospital for Sick Children, Toronto, ON, Canada
Ronald D. Cohn

Authors

Antonio Mollica
View author publications
Search author on:PubMed Google Scholar
Safia Omer
View author publications
Search author on:PubMed Google Scholar
Georgiana Forguson
View author publications
Search author on:PubMed Google Scholar
Sydney Steiman
View author publications
Search author on:PubMed Google Scholar
Sonia L. Evagelou
View author publications
Search author on:PubMed Google Scholar
Serhiy Naumenko
View author publications
Search author on:PubMed Google Scholar
Susan Walker
View author publications
Search author on:PubMed Google Scholar
Lu Yi Li
View author publications
Search author on:PubMed Google Scholar
Aideen Teeling
View author publications
Search author on:PubMed Google Scholar
Kyle Lindsay
View author publications
Search author on:PubMed Google Scholar
Steven Erwood
View author publications
Search author on:PubMed Google Scholar
Shagana Visuvanathan
View author publications
Search author on:PubMed Google Scholar
Anjali Vig
View author publications
Search author on:PubMed Google Scholar
Robert M. Vernon
View author publications
Search author on:PubMed Google Scholar
Benjamin Akman
View author publications
Search author on:PubMed Google Scholar
Constance Smith-Hicks
View author publications
Search author on:PubMed Google Scholar
Julie D. Forman-Kay
View author publications
Search author on:PubMed Google Scholar
Manohar Shroff
View author publications
Search author on:PubMed Google Scholar
Vivek Pai
View author publications
Search author on:PubMed Google Scholar
Rene E. Harrison
View author publications
Search author on:PubMed Google Scholar
Ronald D. Cohn
View author publications
Search author on:PubMed Google Scholar
Evgueni A. Ivakine
View author publications
Search author on:PubMed Google Scholar

Contributions

Conceptualization: A.M., E.A.I., R.D.C., S.O., R.E.H. Methodology: A.M., S.O., S.L.E., G.F., S.S., E.A.I., S.N., S.W., S.E., R.M.V. Investigation: A.M., S.O., S.L.E., G.F., S.S., S.N., S.W., L.Y.L., A.T., K.L., S.E., S.V., A.V., R.M.V., M.S., V.P. Resources: B.A., C.S.-H, R.D.C. Visualization: A.M., S.O., G.F., S.S., M.S., V.P. Supervision: E.A.I., R.D.C., R.E.H., J.D.F.-K. Writing-original draft: A.M., M.S., S.O., G.F. Writing-review & editing: A.M., G.F., E.A.I., R.D.C., S.O., R.E.H.

Corresponding author

Correspondence to Evgueni A. Ivakine.

Ethics declarations

Competing interests

The authors declare no competing interests.

Peer review

Peer review information

Nature Communications thanks the anonymous reviewers for their contribution to the peer review of this work. A peer review file is available.

Additional information

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary information

Supplementary Information (download PDF )

Description of Additional Supplementary File (download PDF )

Supplementary Data 1 (download XLSX )

Supplementary Data 2 (download XLSX )

Supplementary Data 3 (download XLSX )

Supplementary Data 4 (download XLSX )

Supplementary Movie 1 (download AVI )

Reporting Summary (download PDF )

Transparent Peer Review file (download PDF )

Source data

Source Data (download XLSX )

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/.

Reprints and permissions

About this article

Cite this article

Mollica, A., Omer, S., Forguson, G. et al. Mutations in the β-tubulin TUBB impair ciliogenesis and are associated with ciliopathy-like phenotypes. Nat Commun 16, 10637 (2025). https://doi.org/10.1038/s41467-025-65634-x

Download citation

Received: 16 June 2023
Accepted: 20 October 2025
Published: 27 November 2025
Version of record: 27 November 2025
DOI: https://doi.org/10.1038/s41467-025-65634-x