A proteomic survey of microtubule-associated proteins in a R402H TUBA1A mutant mouse

Microtubules play a critical role in multiple aspects of neurodevelopment, including the generation, migration and differentiation of neurons. A recurrent mutation (R402H) in the α-tubulin gene TUBA1A is known to cause lissencephaly with cerebellar and striatal phenotypes. Previous work has shown that this mutation does not perturb the chaperone-mediated folding of tubulin heterodimers, which are able to assemble and incorporate into the microtubule lattice. To explore the molecular mechanisms that cause the disease state we generated a new conditional mouse line that recapitulates the R402H variant. We show that heterozygous mutants present with laminar phenotypes in the cortex and hippocampus, as well as a reduction in striatal size and cerebellar abnormalities. We demonstrate that homozygous expression of the R402H allele causes neuronal death and exacerbates a cell intrinsic defect in cortical neuronal migration. Microtubule sedimentation assays coupled with quantitative mass spectrometry demonstrated that the binding and/or levels of multiple microtubule associated proteins (MAPs) are perturbed by the R402H mutation including VAPB, REEP1, EZRIN, PRNP and DYNC1l1/2. Consistent with these data we show that the R402H mutation impairs dynein-mediated transport which is associated with a decoupling of the nucleus to the microtubule organising center. Our data support a model whereby the R402H variant is able to fold and incorporate into microtubules, but acts as a gain of function by perturbing the binding of MAPs.

R402H mutation in mice, and the generation of transgenic yeast harbouring an analogous variant (R403H) [25]. Here, we explore this idea further by exploiting a new R402H conditional TUBA1A mouse mutant. We show that this mouse line recapitulates multiple aspects of the human disease including neuronal migration defects in the cortex and hippocampus as well as a reduction in striatal and cerebellar size. Coupling microtubule sedimentation assays with quantitative mass spectrometry we identify a number of MAPs that are dysregulated in this mouse mutant including; VAPB, REEP1, EZRIN, PRNP and DYNC1l1/2.

Generation of the R402H Tuba1a mouse line
To generate a conditional R402H Tuba1a mouse mutant we designed a targeting construct that carries two copies of the terminal exon 4 (Fig 1A, S1A and S1B Fig). The first copy encodes wild-type Tuba1a and is flanked by LoxP sites. Downstream we cloned a R402H mutant version of exon 4, that is expressed in the presence of a Cre recombinase. Following the targeting of ES cells, southern blot analysis, and blastocyst injection, chimeric offspring were backcrossed to C57/BL6. We confirmed the functionality of our genetic approach by crossing these animals with a Emx1-Cre driver line that expresses Cre in the dorsal forebrain [26]. cDNA sequencing of R402H/+; Emx1-cre animals showed that the transcript encoding the R402H mutation is expressed in the cortex of adult mice, but not in the cerebellum ( Fig  1B). We confirmed that our manipulation of the Tuba1a genomic locus by itself does not lead to any overt phenotypes by assessing the neuroanatomical features of heterozygous and homozygous adult animals that do not express Cre recombinase. Both R402H/+ and R402H/R402H are indistinguishable from +/+ controls (S1C-S1E and S1F-S1I Fig).

Defects in neuronal migration in the cortex and hippocampus in R402H/+; Emx1-cre mice
To directly assess neuronal migration in R402H/+; Emx1-cre mice, we injected pregnant dams with BrdU at E14.5 and harvested brains at P0 and P10. We divided the cortex into 10 bins In the presence of a Cre recombinase, the wild-type variant of exon 4 is excised, and the R402H mutant is expressed. (B) Sequencing traces generated from mRNA extracted from cortex and cerebellum of heterozygous R402H/+ animals that express Emx1-Cre in the dorsal telencephalon, but not in the cerebellum. The R402H mutation is only detected in the cortex. (C-E) Nissl stained sections of adult +/+; Emx1-cre, R402H/+, and R402H/+; Emx1-cre animals. A fractured pyramidal cell layer in CA1 region (indicated by arrowheads) of the hippocampus is present in R402H/+; Emx1-cre animals (n = 5). (F-H) Representative coronal sections of the adult cortex stained with the pan-neuronal marker NEUN from +/+; Emx1-cre, R402H/+, and R402H/+; Emx1-cre animals. (I) Quantification of cortical thickness reveals no significant difference between genotypes (n = 5, P>0.05). (J-R) Cortical sections stained with layerspecific markers: CUX1 for layers II to IV, ER81 for layer V and FOXP2 for layer VI. (S-U) Relative distribution of CUX1 (S), ER81 (T) and FOXP2 (U) and quantitated the distribution of BrdU-positive cells in R402H/+; Emx1-cre mutants and littermate controls (Fig 2A-2D). At P0 we observed a significant reduction in the relative number of BrdU positive cells in bin 9 in R402H/+; Emx1-cre mutants with a concomitant increase in bins 1-2 ( Fig 2C) (n = 5, R402H/+ vs. R402H/+; Emx1-cre bin 9: P<0.0001, bin 1: P<0.01 and bin 2: P<0.01). At P10 there was a significant reduction of cells in bin 7, with an accumulation of BrdU-positive cells in bins 1 and 2 in R402H/+; Emx1-cre mutants, when compared to their littermate controls ( Fig 2D) (n = 5, R402H/+ vs. R402H/+ Emx1-cre, bin 7: P<0.0001, bin 1: P<0.001 and bin 2: P<0.001). To assess neuronal migration in the hippocampus we focused on our P10 cohort and quantitated BrdU-positive cells in the oriens layer, the pyramidal cell layer, the radiatum layer, the lacunosum moleculare, the molecular layer of the dentate gyrus, the dentate gyrus and in the hilus (Fig 2E and 2F). This showed a significant increase in the relative number of BrdU-positive cells in the oriens layer in R402H/+; Emx1-cre mutants in comparison to littermate controls, with a decrease in the pyramidal cell layer (n = 5, R402H/ + vs. R402H/+; Emx1-cre oriens layer: P<0.0001, pyramidal cell layer: P<0.0001). Quantitation of the total number of BrdU-labelled neurons in both the cortex and hippocampus revealed no differences between genotypes suggesting that neurogenesis is normal in R402H/ +; Emx1-cre animals (S2 Fig) (n = 5, R402H/+ vs. R402H/+; Emx1-cre, P>0.05). These data show that the R402H mutation in TUBA1A causes defects in neuronal migration in the cortex and hippocampus of mice.

Reduction in striatal size and cerebellar abnormalities in R402H/+; Nestincre mice
A recurrent feature described in patients with TUBA1A mutations is a dysmorphic striatum with a hypoplastic and/or dysplastic cerebellum [28,29]. To assess whether the R402H/+ mouse line recapitulates these features we exploited a Nestin driver line that expresses Cre recombinase throughout the developing nervous system [30]. To determine whether the striatum was affected in the mutants, we stained coronal brain sections of P6 mice with Nissl ( Fig  3A-3C). There was a significant reduction in the striatal area in R402H/+; Nestin-cre animals in comparison to littermate controls ( Fig 3D) (n = 5, R402H/+ vs. R402H/+; Nestin-cre P<0.05), accompanied by an enlargement of the lateral ventricles ( Fig 3E) (n = 5, +/+; Nestincre vs. R402H/+; Nestin-cre P<0.01). This phenotype was mirrored in the cerebellum where we again observed a significant reduction in the total area of sagittal sections in P6 R402H/+; Nestin-cre animals (Fig 3F-3I) (n = 5, R402H/+ vs. R402H/+; Nestin-cre P<0.0001). To assess cerebellar lamination, we focused on the 3 rd folia staining sagittal sections with calbindin (which labels Purkinje cells) and NeuN (which labels all neurons) (Fig 3J-3O). We quantitated the number of NeuN-positive cells within the internal granule cell layer (IGL), the Purkinje cell layer (PCL), the molecular layer (ML) and in the external granule cell layer (EGL). We observed a significant accumulation of NeuN-positive cells in both the ML and PCL in R402H/ +; Nestin-cre mutants, concomitant with a reduction in the number of NeuN-positive cells in the IGL (Fig 3P) (n = 5, R402H/+ vs. R402H/+ Nestin-cre ML: P<0.0001, PCL: P<0.01, and IGL: P<0.0001). These results show that the R402H mutation causes a reduction in the size of the striatum and cerebellum, and defects in cerebellar lamination. positive cells in control mice (+/+; Emx1-cre in light grey and R402H/+ in dark grey) and heterozygous mutants (R402H/+; Emx1-cre in blue). R402H/+; Emx1-cre mutants presented with significant lamination defects for all three layer-markers (n = 5, see S1 Table)

Homozygous expression of the R402H mutation exacerbates the migration phenotype and leads to neuronal cell death
The conditional approach we employed allowed us to generate homozygous, R402H/R402H; Emx1-cre animals, however, these mice die shortly after birth. We therefore harvested R402H/ R402H; Emx1-cre embryos at E18.5 when Tuba1a is expressed at high levels and neuronal Neuronal migration defects in the cortex and hippocampus of R402H/+; Emx1-cre mice. BrdU was injected into pregnant dams at E14.5, and brains harvested at either P0 (n = 5) or P10 (n = 5). (A) Image of a P10 brain stained with antibodies that target BrdU. (B) Enlargement showing segmentation of the P10 cortex into 10 bins. (C) Graph showing the relative distribution of BrdU+ cells in the cortex of P0 control animals (+/+; Emx1-cre in light grey and R402H/+ in dark grey) and heterozygous mutants (R402H/+; Emx1-cre in blue). In R402H/+; Emx1-cre animals we noted a significant accumulation of BrdU+ cells in the deeper layers of the cortex (bins 1-4) concomitant with a decrease of BrdU+ cells in the upper layers (bins 8-9). (D) Relative distribution of BrdU+ cells across the cortex at P10. We found a significant increase in the proportion of BrdU+ cells in the deeper layers (bins 1-2) and a reduction in proportion of BrdU+ cells in bin 7. (E) Image showing the segmentation of a P10 hippocampus stained with antibodies that target BrdU. We quantitated BrdU+ cells in the oriens layer (OL), pyramidal cell layer (PCL), radiatum layer (RL), lacunosum moleculare (LM), molecular dentate gyrus (MDG), dentate gyrus (DG) and the hilus (H). (F) Graph showing the distribution of BrdU+ cells in the different layers of the hippocampus of P10 mice (n = 5). R402H/+; Emx1-cre mutants show a significant accumulation of cells in the OL, accompanied by a decrease in BrdU+ cells in the PCL in comparison to controls. Error bars show mean ± s.e.m. � P<0.05; �� P<0.01; ��� P< 0.001; ���� P< 0.0001 (See also S1 Table). Scale bars show 1000 μm in A, 100 μm in B and 200 μm in E.

The R402H mutation alters the microtubule proteome
To gain insight into the molecular mechanisms that underlie the neuronal migration phenotype in R402H/R402H; Emx1-cre animals, we isolated microtubules from the developing forebrain of mutants and littermate controls at E18.5 by taxol-mediated polymerisation coupled with sucrose gradient centrifugation (n = 4) ( Fig 5A). We confirmed that expression of R402H TUBA1A did not affect α-tubulin protein levels in either brain lysates or microtubule pellets Exploiting isobaric labelling and mass spectrometry, we undertook a quantitative comparison of the proteins associated with microtubules in R402H/R402H; Emx1-cre mutants and R402H/R402H controls. In total this experiment resulted in the identification of 5283 proteins with 2 or more unique peptides, including DCX, Lis1, multiple kinesin motors, as well as EB1, EB2 and EB3 (S2 Table). Amongst these 1330 were significantly altered in R402H/R402H; Emx1-cre mutants (P<0.01) (S3 Table). This number was narrowed to 286 following the application of a false discovery rate of 5% (Fig 5B and 5C, S4 Table). To assess the reliability of our data, we compared the results of four biological replicates and observed similar fold change ratios for those proteins altered in R402H/R402H; Emx1-cre mutants (S4D Fig). Next, we performed an unbiased gene ontology analysis on the 286 proteins altered in R402H/R402H; Emx1-cre mutants which revealed an enrichment of proteins involved in cellular component organization, neuron and neuron projection development ( Fig 5D). We further refined our list of candidates to 7  Table). (T-V) R402H/R402H animals were electroporated with either a pCAGEN control vector or pCAGEN-Cre at E14.5 and the embryos collected at E17.5. (V) Relative distribution of proteins known to directly bind to microtubules; VAPA, VAPB, REEP1, EZRIN, PRNP, KIF5C and DYNC1l1. To confirm that these proteins are dysregulated in R402H/R402H; Emx1-cre mutants, we again isolated microtubules from the forebrain of E18.5 animals and undertook western blot analyses (Fig 6A-6G and S5 Fig). For each of our candidates we quantitated the total amount of protein in forebrain lysates, as well as the amount that sedimented with microtubules. Unexpectedly we found that the total amount of VAPB, REEP1, EZRIN and PRNP was significantly lower in the R402H/R402H; Emx1-cre brain lysates with a GFP+ cells across the cortex at E17.5. Acute expression of the R402H allele results in an accumulation of neurons in the SVZ and iIZ, which is accompanied by a significantly lower proportion of neurons that reach the cortical plate (n = 6, see S1 Table). In all panels control animals (+/+; Emx1-cre) are shown in grey, heterozygotes (R402H/+; Emx1-cre) in blue and homozygotes (R402H/R402H; Emx1-cre) in orange. VZ denotes ventricular zone; SVZ, subventricular zone; IZ, intermediate zone; CP, cortical plate; MZ, marginal zone; iIZ, mIZ, oIZ, and iCP, mCP, oCP correspond to inner, medial and outer IZ and CP. Scale bars show 400 μm in B, 500 μm in C, 100 μm in F, 100 μm in J, N, R and U, and 10 μm in J'. In G, O, S and V error bars show mean ± s.e.m., and in K, data is represented as median ± 95% CI. � P<0.05; �� P<0.01, ��� P<0.001, ���� P< 0.0001.
https://doi.org/10.1371/journal.pgen.1009104.g004 These data suggest that the decreased protein observed in lysates is due to post-translation dysfunction, and not a transcriptional phenotype. Only in the case of dynein intermediate chain (DYNC1l1/2) were there comparable amounts of protein in control brain lysates but a significant reduction in protein levels that sedimented with microtubules in R402H/R402H; Emx1-cre mutants (Fig 6A)

The R402H mutation perturbs dynein-mediated lysosomal transport and coupling of the nucleus to the centrosome
Given the results of our proteomic analysis we asked whether dynein-mediated lysosomal transport is perturbed in R402H/R402H; Emx1-cre animals. To assess this, we cultured primary

PLOS GENETICS
travelled by lysosomes towards the soma, which was significantly reduced in R402H/R402H; Emx1-cre mutants in comparison to littermate controls ( Fig 7F) (n = 39 lysosomes for R402H/ R402H and n = 35 lysosomes for R402H/R402H; Emx1-cre P<0.05). The dynein complex is known to play a critical role in neuronal migration as it generates the force necessary to move the nucleus towards the microtubule organising center [5,33]. Previous studies have shown that dynein perturbation disrupts nucleus-centrosome coupling causing an impairment in nuclear translocation [5,34]. In light of these data we analysed nucleus-centrosome coupling in our R402H/R402H mice following in utero electroporation of a pCAGEN-Cre vector at E14.5, and harvesting at E17.5. We labelled the centrosome with gamma-tubulin and measured the distance from the nucleus to the centrosome in neurons entering the cortical plate (Fig 8A-8C). This revealed a statistically significant increase in nucleus-centrosomal distance when the R402H allele is expressed in comparison to controls (n = 6, R402H/R402H; pCAGEN vs. R402H/R402H; pCAGEN-Cre, P<0.0001). Taken together these results show that the R402H mutation perturbs dynein-mediated transport and coupling of the nucleus to the microtubule organising center.

Discussion
In this study we have investigated the pathology and molecular mechanisms that result from an R402H mutation in TUBA1A, a variant that is known to cause lissencephaly in the human population. To do so we generated a R402H conditional mouse model and have shown that it recapitulates multiple aspects of the human disease phenotype. Heterozygous animals present with laminar defects in the cortex and hippocampus, as well as cerebellar and striatal hypoplasia. We attribute the laminar abnormalities observed in the forebrain to defects in neuronal migration, which is exacerbated in homozygous animals. In homozygotes the developing cortical plate is inverted and there is a significant increase in neuronal cell death which is associated with postnatal lethality. Exploiting in utero electroporation, we have shown that acute expression of the R402H allele perturbs neuronal migration in a cell-autonomous manner, decoupling the nucleus from the centrosome. While there are a number of existing mouse mutants

PLOS GENETICS
for TUBA1A, this animal is the first to directly recapitulate a known disease-causing variant. Consistent with the data presented in this paper, the Jenna (S140G), Tuba1a ND (N102D), and the Tuba1a M101736 (D47G) mouse mutants all present with neuroanatomical phenotypes indicative of impairments in neuronal migration and/or cell death [20,35,36]. Interestingly, the ablation of TUBA1A by CRISPR-mediated genome editing (Tuba1a d4304/d4304 , Tuba1a d4262/d4262 ) results in viable heterozygous animals, that are not reported to have laminar abnormalities in the forebrain [37]. In contrast the homozygous ablation of Tuba1a results in lethality, with severe brain malformations with enlarged ventricles [37]. These data, coupled with the results presented in this manuscript indicate that TUBA1A is necessary for cortical development in vertebrates and that the R402H variant acts as a gain of function. This mode of action is supported by the work of Aiken and colleagues who have convincingly shown that ectopic overexpression of the R402H mutation in the developing murine cortex disrupts neuronal migration [25].
To explore the molecular mechanisms that are perturbed by the R402H mutation we sedimented microtubules from the forebrain of R402H/R402H animals and performed quantitative mass spectrometry. This resulted in the identification of 286 proteins that were significantly dysregulated in R402H/R402H mutants, surprisingly only 7 of these were reported to be MAPs (VAPA, VAPB, REEP1, EZRIN, PRNP, KIF5C and DYNC1l1). While a proportion of the remaining 279 proteins may be uncharacterised MAPs, it is probable that some form complexes with MAPs. For instance, Hap1 is known to interact with dynein participating in vesicular trafficking [38] and β-catenin immunoprecipitates with ezrin [39]. An alternative possibility is that the pellet analysed included precipitated proteins that have simply co-sedimented with the microtubules during sample preparation [40]. Previous work by Yu and colleagues developed a system to identify tubulin associated proteins (TAPs), by expressing a TUBB3-P2A-bio-Sumo-GFP-TUBA1A construct in HEK293T cells. Tubulin heterodimers were then purified using a biotinylated tag and mass spectroscopy performed, resulting in the identification of 3091 putative TAPs. In our study we identified a total of 5283 putative MAPs of which 933 overlap with the Yu dataset. The larger number of proteins identified in our study may be attributed to the fact that we assessed microtubule bound proteins (as opposed to TAPs), our analysis was performed in an in vivo setting (as opposed to HEK293T cells), and because we relied on sedimentation (as opposed to tagged proteins).
Western blot analysis confirmed that 5 out of 7 MAPs (VAPB, REEP1, EZRIN, PRNP and DYNC1l1/2) identified by mass spectrometry were dysregulated in R402H/R402H, Emx1-cre animals. We also observed a reduction in the levels of VAPA in mutant animals in lysates (P = 0.15), but this reduction was not statistically significant. Our inability to confirm a reduction of microtubule associated KIF5C in R402H/R402H, Emx1-cre animals may be due to the reduced sensitivity of Western blot analysis in comparison to spectroscopic methods or it may be a false positive. VAPA and VAPB are vesicle-associated membrane proteins that are associated with axonal transport of mitochondria and in the case of VAPB implicated in the pathology of amyotrophic lateral sclerosis (ALS8) [41,42]. REEP1 plays an important role in the organisation of the microtubule endoplasmic reticulum network, is expressed in the developing cortex and is known to cause spastic paraplegia 31 [43,44]. EZRIN is a cytoskeletal linker that is associated with the plasma membrane and has been shown to regulate RhoA/Rho kinase; it affects neuritogenesis and is highly expressed in migrating neurons within the intermediate zone [45,46]. PRNP encodes the prion protein PrP C , responsible for the neurodegenerative process in Creutzfeldt-Jakob disease and has been shown to influence the stability and assembly of microtubules [47,48]. Interestingly, it has been shown that retrograde transport of PrP C within vesicles in hippocampal neurons is mediated by the dynein motor [49]. In the case of VAPB, REEP1, EZRIN and PRNP, we unexpectedly found that the R402H mutation influences the total level of proteins in forebrain lysates. This reduction could be a consequence of a negative feedback loop whereby a reduction in microtubule binding due to the R402H mutation influences the levels of these proteins, but not their transcript.
Our finding that dynein binding and functionality is perturbed by the R402H mutation mirrors the conclusions reached by Aiken and colleagues, who mutated the equivalent tubulin residue (R403) in yeast. They showed that GFP-tub1-R403H still incorporates into microtubules and that dynein is correctly recruited to the plus-end of microtubules, but reported that dynein mediated microtubule sliding events during mitosis were impaired [25]. By deleting the minor tubulin Tub3, they further demonstrated that this phenotype scales with the ratio of the R403H allele present in cells. This result is consistent with our observation that a more severe neuronal migration phenotype occurs in homozygous R402H/R402H animals compared to heterozygous R402H/+ animals. We have further demonstrated that in vivo expression of the R402H mutation decouples the nucleus from the centrosome, consistent with a defect in dynein-mediated nucleokinesis. Interestingly, patients with dynein mutations (DYNC1H1) present with a range of developmental phenotypes including polymicrogyria, dysgenesis of the corpus callosum and dysmorphic basal ganglia [7,50]. Dynein mutations are not, however, limited to patients with developmental phenotypes. They have also been reported in individuals with degenerative disorders such as Charcot-Marie-Tooth disease (CMT2) and spinal muscular atrophy [51,52]. It is therefore conceivable that dynein dysfunction contributes to both the migratory and cell death phenotype that we observe in R402H/R402H animals.
In summation, our data are consistent with a model whereby the R402H variant is able to fold and incorporate into microtubules, but acts as gain of function by perturbing the binding and/or protein levels of multiple MAPs. We have shown that this is not limited to dynein alone, but includes other MAPs including VAPB, REEP1, EZRIN and PRNP. Future studies will explore what role these proteins play in neuronal migration and/or survival, providing a foundation to design appropriate therapeutic interventions.

Ethics statement
All experiments were carried out according to legal requirements and covered by an approved license (M58/006093/2011/14) from the City of Vienna.

Animal generation, genotyping and maintenance
Mice carrying the Tuba1a R402H allele were generated by Ozgene (Perth, Australia) exploiting their go Germline platform. Briefly, targeting constructs were electroporated into C57/BL6 ES cells and positive clones selected by southern blot analysis following digestion with Nde I. Positive clones were injected into blastocysts to generate chimeras, which were backcrossed to C57/BL6 animals. Animals were genotyped by PCR amplification of Tuba1a locus using primers: ATGCTGCCAATAACTATGCT and TGGGAGAAAAGGAAATGTAA. Transgenic mice expressing Cre recombinase were genotyped using the primers: GCTCGACCAGTTTAGTTAC CC and TCGCGATTATCTTCTATATCTTCAG. We sequenced both the Tuba1a genomic locus and cDNA which confirmed that R402H is the only coding variant present. Animals were housed at the animal research laboratories of the Institute of Molecular Pathology on a 14:10 hour light:dark cycle. Food and water were provided ad libitum.

Nissl staining
Slides were washed three times in PBS before incubation with Nissl stain (0.1% cresyl violet acetate; Sigma, C5042), and dehydrated in increasing concentrations of ethanol (30%, 70%, 95% and 100%). Slides were kept in xylol prior to coversliping with DPX mounting media. The slides were then digitized using a Pannoramic 250 Flash II slide scanner (3DHISTECH), with a 20x/0.8 Plan-Apochromat objective. To quantify the cortical thickness four (E18.5) sections per animal were measured with ImageJ. Quantifications were performed blind to genotype.

BrdU injection and staining
Pregnant mice were injected intraperitoneally with BrdU (5 μL at 20 mg/mL per gram of mouse body weight) at E14.5. Pups were collected at birth (P0) or at postnatal day 10 (P10) and samples drop-fixed in 4% paraformaldehyde overnight, before dehydration in 30% sucrose and embedding in Neg-50 Frozen Section Medium. Littermate controls were sectioned on a cryostat (12 μm) and stained against BrdU using a permanent staining protocol. Briefly, the sections were pre-treated for 5 min with 3% H 2 O 2 and 0.1% NaN 3 in water to block endogenous peroxidases. Heat-mediated antigen retrieval was performed using a tris-based buffer (Vector Laboratories, H-3301), followed by an HCl treatment (2N in water) for 70 min at 37˚C. Sections were incubated with the primary antibody (1:300, Bio-Rad, OBT0030) prepared in 2% milk in PBS + 0.3% Triton at 4˚C, overnight. The next day, sections were washed and incubated with a species-specific secondary antibody, conjugated to horseradish peroxidase (ImmPRESS HRP anti-rat detection kit, Vector Laboratories, MP-7404) for 1 h at room temperature. 3,3'-diaminobenzidine (DAB, Vector Laboratories SK-4100) was used as substrate for HRP. Finally, sections were dehydrated in increasing concentrations of ethanol before mounting with DPX mounting medium (Sigma-Aldrich, 06522). Matching sections were scanned and four images per animal were quantitated using ImageJ. For P0 and P10 brains, BrdU-positive cells were manually counted in a 370 μm-wide box, placed across the cortex and divided in ten equally spaced bins. The different layers of the hippocampus were manually segmented and the BrdU-positive cells, quantified using ImageJ. All quantifications were done blind to genotype.

In utero electroporation
In utero electroporations were performed as previously described [53]. For these experiments, the following plasmids were used: pCAGEN (control) or pCAGEN-Cre (used to induce the expression of the R402H mutant) which were injected at 1 μg/μL, together with pCAGEN-GFP at 0.2 μg/μL to label the migrating neurons. Electroporations were performed on E14.5 pregnant R402H/R402H females and the embryos collected at E17.5. Following fixation and dehydration, the samples were sliced (12 μm thick) and stained with a GFP antibody. Image acquisition was done on a laser scanning confocal microscope (LSM710 Zeiss). Five images per animal were analysed and the number of GFP positive cells was counted manually using ImageJ, blind to the condition. To assess the distribution of the labelled cells, we divided the cortex in different zones: ventricular zone, the subventricular zone, the inner, medial, and outer intermediate zone (iIZ, mIZ, oIZ), and the inner, medial, and outer cortical plate (iCP, mCP, oCP) [53]. These were defined based on cell density, using the DAPI channel as a reference. Inner/medial/outer sub-zones were defined by sub-diving each cortical zone in three equally spaced layers. To quantify the distance between the nucleus and centrosome, the sections were stained with antibodies against GFP (abcam, ab13970, 1:1000) and gamma-tubulin (Sigma, T6557, 1:100), and counterstained with DAPI. Ten to twelve images were acquired on a laser scanning confocal microscope (LSM800 Zeiss) equipped with a 20x/0.8 Pan-Apochromat DIC objective. Z-stacks were collected at 1 μm steps and maximum projection images generated before quantification. The distance between the nucleus and centrosome was measured manually in ImageJ, blinded to the genotype, in GFP positive migrating neurons in the intermediate zone, with a clear bipolar morphology and identifiable centrosome within the leading process.

Microtubule sedimentation assay
Microtubules pellets were prepared as previously described [7]. Briefly, forebrain samples were isolated from E18.5 embryos by dissection in cold PBS and subsequent snap-frozen in liquid nitrogen. Protein lysates of wild-type and mutant embryos were prepared simultaneously in tubulin buffer: 100 mM PIPES pH 6.9 (Sigma, P6757), 1 mM MgCl 2 and 1 mM EGTA (Sigma, E3889) supplemented with protease inhibitor (Roche, 04693159001). The tissue was homogenized by pipetting the cell suspension, kept on ice. After incubation for 15 min on ice, the lysate was centrifuged at maximum speed at 4˚C in a tabletop centrifuge, followed by two ultracentrifugations at 60 000 rpm at 4˚C for 15 min. Total protein concentration was quantified using Bradford reagent (Bio-Rad, 5000006) and 70% of the cleared lysate was used in the polymerization reaction (approximately 60 μl). For microtubule polymerisation taxol (Sigma, T7191), GTP (Sigma, G8877) and AMP-PNP (Sigma, 10102547001) were added to the lysate to a final concentration of 20 μM, 1 mM and 2 mM, respectively. Microtubule polymerization was performed at 37˚C for 30 min. To pellet taxol-stabilized microtubules and microtubuleassociated proteins, the reaction was loaded onto a sucrose cushion prepared with 1 M sucrose, 10 μM taxol, 0.5 mM GTP and 1 mM AMPPNP. The samples were then centrifuged at 60 000 rpm at 37˚C for 20 min. All ultracentrifugations were performed in a Beckman TLA 100.3 rotor in Optima MAX-XP ultracentrifuge. After removal of the supernatant, the pellet was resuspended directly in loading dye, boiled at 95˚C for 3 min and kept at -20˚C. For mass spectrometry analysis the pellets were snap-frozen in liquid nitrogen until further processing.

Mass spectrometry analysis
Protein pellets were processed using previously described protocols with minor modifications [54]. Briefly, pellets were dissolved in 30 μL of SDT buffer (4% SDS, 0.1 M DTT in 0.1 M TEAB; pH = 7.5) and incubated 10 min at 95˚C. Protein extracts were mixed with 200 μL of UA solution (8 M urea in 0.1 M TEAB; pH = 8.5), placed into a microcon unit (MWCO = 30 kDa) and centrifuged for 20 min at 14000 x g. All centrifugations were carried out at room temperature unless stated otherwise. After the washing step, proteins were alkylated by adding 100 μL of 0.1 M iodoacetamide in UA with following 20 min incubation in dark and at room temperature. Following centrifugation proteins were digested directly in the ultrafiltration unit by adding 50 μL of TEAB buffer which contains a specific protease trypsin (Mass spectrometry grade; Promega) in 1:40 ratio (protease:protein; w/w). After overnight digestion at 37˚C peptides were eluted into a fresh collection tube by centrifugation at 14000 x g. The ultrafiltration unit was then washed twice by adding 50 μL of 50 mM TEAB buffer, centrifuged again, and the eluate collected. Collected tryptic peptides were subsequently labelled with tandem mass tags (TMT). TMT10-plex labelling reagents were equilibrated to room temperature and 41 μL of anhydrous acetonitrile were added. Reagents and the peptide samples (80 μg for each sample) were mixed together and incubated for 1h. Different amounts of the channels were added to correct the median value to 1:1:1:1:1:1:1:1:1:1. Following labelling, peptides were desalted using C18 cartridges (Sep-Pak Vac 1cc-50 mg, Waters). These were equilibrated, then the peptide sample was loaded on the columns without applying pressure. The cartridge was washed 3 times and to elute the peptides, the cartridge was flushed 3 times with 150 μL of 70% ACN and 0.1% FA. The sample was diluted to a volume of 200 μL with 0.1% FA, snap frozen with liquid nitrogen and lyophilized overnight. The lyophilized peptide sample was dissolved in SCX buffer A (5 mM phosphate buffer pH 2.7, 15% ACN). The SCX fractionation was performed on an Ultimate system (Thermo Fisher Scientific) using a TSKgel SP-25W (ToSOH) column (5 μm particles, 1 mm i.d. x 300 mm) at a flow rate of 40 μL/min. For the separation, a ternary gradient was used, starting with 100% buffer A for 10 min, followed by a linear increase to 10% buffer B (5 mM phosphate buffer pH 2.7, 1 M NaCl, 15% ACN) and 50% buffer C (5 mM phosphate buffer pH 6.0, 15% ACN) in 10 min, to 25% buffer B and 50% buffer C in next 10 min, to 50% buffer B and 50% buffer C in next 5 min and an isocratic elution for further 15 min. The flow-through was collected as a single fraction, along the gradient fractions that were collected every minute and stored before analysis. The system used was an UltiMate 3000 RSLC nano system coupled to a Q Exactive HF-X mass spectrometer, equipped with an EASY-spray ion source (Thermo Fisher Scientific) and JailBreak 1.0 adaptor insert for a spray emitter (Phoenix S&T, Inc., USA). Peptides were loaded onto a trap column (Thermo Fisher Scientific, PepMap C18, 5 mm × 300 μm ID, 5 μm particles, 100 Å pore size) at a flow rate of 25 μL/min using 0.1% TFA as mobile phase. After 10 min, the trap column was switched in line with the analytical column (Thermo Fisher Scientific, PepMap C18, 500 mm × 75 μm ID, 2 μm, 100 Å). Peptides were eluted using a flow rate of 230 nL/min, and a binary 2h gradient, respectively 165 min. The gradient starts with the mobile phases: 98% A (water/formic acid, 99.9/0.1, v/v) and 2% B (water/acetonitrile/formic acid, 19.92/80/0.08, v/v/ v), increases to 35% B over the next 130 min, followed by a gradient in 5 min to 90% B, stays there for 5 min and decreases in 2 min back to the gradient 98% A and 2% B for equilibration at 30˚C. The Q Exactive HF-X mass spectrometer was operated in data-dependent mode, using a full scan (m/z range 380-1650, resolution of 120 000, target value 3x10 6 ) followed by MS/MS scans (resolution of 45000, target value 1x10 5 , maximum injection time 250ms) of the 10 most abundant ions. MS/MS spectra were acquired using normalized collision energy 35% and an isolation width of 0.7 m/z. For the detection of the TMT reporter ions a fixed first mass of 110 m/z was set for the MS/MS scans. Precursor ions selected for fragmentation were put on a dynamic exclusion list for 60 s. The intensity threshold was set to 4x10 4 . Raw files were processed with Proteome Discoverer (version 2.3.0.523, Thermo Fisher Scientific, Germany). Database searches were performed using MS Amanda (version 2.0.0.9849) against the Swis-sProt mouse database (version 2019.01.26). The raw files were loaded as fractions into the processing workflow. Trypsin was defined as the proteolytic enzyme, cleaving after lysine or arginine, except when followed by proline, and up to two missed cleavages were allowed. Precursor and fragment ion tolerance were set to 5 ppm and 15 ppm respectively. Identified spectra were rescored using Percolator and filtered to 1% FDR at the peptide spectrum match level. Protein grouping was performed in Proteome Discoverer applying a strict parsimony principle. Proteins were subsequently filtered to a false discovery rate of 1% at protein level. TMT Reporter ion S/N values were extracted from the most confident centroid mass within an integration tolerance of 20 ppm. Protein quantification was determined based on unique and razor peptides. Statistical significance of differentially abundant peptide/proteins between different conditions was determined using a paired LIMMA test [55]. The functional enrichment analysis was performed using by g:Profiler (version e98_eg45_p14_bca6d38) with g:SCS multiple testing correction method applying significance threshold of 0.05 as previously described [56].

Primary neuron culture and live-cell imaging of lysosomes
Cortical primary neurons were prepared from E18.5 embryos as previously described [57]. The meninges were carefully removed before adding trypsin solution (0.05% Trypsin-EDTA, buffered with 1 M HEPES) and incubating the tissue at 37˚C for 15 min. The cortices were then washed three times with HBSS (Gibco, 14025050) before being triturated using P1000 and P200 pipette tips. The resulting cell suspension was filtered through a 40 μm filter (Falcon, 352340) before cell counting. For the lysotracker experiments, 2x10 5 neurons were seeded in poly-L-lysine coated MatTek dishes (MatTek Corporation, P35G-0-14-C). Following 3 days in vitro (DIV), the neuronal cultures were incubated with fresh neurobasal media containing lysotracker diluted at 1:2000 for 25-30 min at 37˚C (Thermo Fisher Scientific, L12492). The cultures were washed prior to addition of imaging media (neurobasal medium without phenol red, Thermo Scientific 12348017). Control and homozygous embryos were imaged together and blinded to genotype in four independent experiments. Imaging was performed on a spinning disc microscope equipped with a Hamamatsu EMCCD 9100-13 camera from PerkinElmer UltraView VoX, using a 63x/1.15 LD C-Apochromat water objective. Trafficking was recorded in the axon for 3 min every 0.5 sec. Files were exported from Volocity (version 6.3.0, PerkinElmer) and analysed in ImageJ. Lysosomal trafficking was measured in the axon, considered as the longest process that also showed a thicker protuberance on the cell body, resembling an axon hillock.

Kymograph analysis and lysosome trafficking
To determine speed and run length, kymographs were generated with the KymoToolBox [58] and the motion of individual lysosomes traced using the Simple Neurite Tracer plug-in for ImageJ. For each kymograph paths were created as a user-defined selection, converted to a segmented line and exported as a ROI, before analysis with a Python-based script. Movement that lasted for at least 3 frames (1.5 sec), was considered a run and was followed until the next change of state. Each track was then classified as retrograde, anterograde, bidirectional or paused, and only tracks traced for at least 60 sec in the retrograde direction were considered. All analysis was done blinded to genotype.

Statistical methods
All statistical analysis was performed in GraphPad Prism software package (v8.0.2). We employed a variety of statistical tests including unpaired t-tests, one-and two-way repeated measures ANOVAs, with Tukey's or Sidak's multiple comparisons test (See S1 Table). Datasets were checked for normality by the Shapiro-Wilk test and whenever the data did not follow a normal distribution, nonparametric tests were applied. These included: Mann-Whitney test and Kruskal-Wallis test with Dunn's multiple comparisons test. When necessary, a Bonferroni correction for multiple comparisons was applied. Samples and animals were not subject to randomization but were assigned to experimental groups based on their genotype.
Supporting information S1 Fig. Generation and characterisation of R402H Tuba1a mice. (A) Schematic representation of the Tuba1a targeting construct designed to conditionally express the R402H patient mutation. After removal of the neomycin cassette, the conditional knock-in (cKI) allele consists of two copies of exon 4, one flanked by LoxP sites and the other carrying the R402H mutation. Upon Cre recombination, the wild-type exon is excised and the mutant exon 4 is expressed. (B) Southern blot analysis of four different clones. The expected size of the wildtype allele is 9.1 kb, and 7 kb for the conditional knock-in. A009F and A044F are both heterozygous for the cKI insertion and A004M and A040F are wild-type mice. (C-H) Representative Nissl stained hippocampal (C-E) and cortical sections (F-H) from adult animals. There is no difference between wild-type (+/+), heterozygous (R402H/+) and homozygous (R402H/ R402H) animals in the absence of a Cre recombinase, demonstrating that introduction of the LoxP sites into the Tuba1a locus has no effect. There is no significant difference in the total number of BrdU-labelled neurons when comparing +/+; Emx1-cre controls, R402H/+ controls and R402H/+; Emx1-cre mutants in the P10 cortex, P0 cortex or the P10 hippocampus (n = 5, P>0.05). Error bars show mean ± s.e.m. Twoway repeated-measures ANOVA with Tukey's test for multiple comparisons; n.s.-not significant. (TIF) S3 Fig. Total number of TBR1+ and BRN2+ cells. (A-B) Quantification of the total number of (A) TBR1+ and (B) BRN2+ neurons in E18.5 sections. Labeled cells were manually counted in a 300 μm-wide box placed across the developing somatosensory cortex. For both layer markers there was no significant difference between +/+; Emx1-cre controls and R402H/+; Emx1-cre heterozygotes (n = 5, P>0.05), however, there was a significant reduction in TBR1 + and BRN2+ cells in R402H/R402H; Emx1-cre homozygotes (n = 5, TBR1: +/+; Emx1-cre vs. Kymographs were generated using a plug-in for ImageJ and each lysosome traced individually. The entire analysis was done blinded to genotype, using a Python-based custom-made script allowing for run length and average speed calculation. (TIF) S1  Table. Mass spectrometry results of 286 proteins significantly dysregulated in R402H/ R402H; Emx1-cre mutants, following the application of an FDR of 5%. (XLSX) S1 Video. Live cell imaging of R402H/R402H primary cortical neurons incubated with Lysotracker. Lysosome trafficking was recorded for 3 min every 0.5 sec. The playback speed of the video is 20 fps. This relates to the kymograph presented in Fig 7A. (AVI) S2 Video. Live imaging of R402H/R402H; Emx1-cre primary cortical neurons incubated with Lysotracker. Wild-type and mutant animals were processed simultaneously, and imaging was done blinded to genotype. Lysosome trafficking was recorded for 3 min every 0.5 sec. The playback speed of the video is 20 fps. This relates to the kymograph presented in Fig 7A. (AVI)