Human TUBB3 mutations perturb microtubule dynamics, kinesin interactions, and axon guidance

Abstract

We report that eight heterozygous missense mutations in TUBB3, encoding the neuron-specific beta-tubulin isotype III, result in a spectrum of human nervous system disorders that we now call the TUBB3 syndromes. Each mutation causes the ocular motility disorder CFEOM3, whereas some also result in intellectual and behavioral impairments, facial paralysis, and/or later-onset axonal sensorimotor polyneuropathy. Neuroimaging reveals a spectrum of abnormalities including hypoplasia of oculomotor nerves and dysgenesis of the corpus callosum, anterior commissure, and corticospinal tracts. A knock-in disease mouse model reveals axon guidance defects without evidence of cortical cell migration abnormalities. We show that the disease-associated mutations can impair tubulin heterodimer formation in vitro, although folded mutant heterodimers can still polymerize into microtubules. Modeling each mutation in yeast tubulin demonstrates that all alter dynamic instability whereas a subset disrupts the interaction of microtubules with kinesin motors. These findings demonstrate that normal TUBB3 is required for axon guidance and maintenance in mammals.

Figures

Figure 1. Clinical spectrum and orbital imaging of the TUBB3 syndromes

(A–I) Study participant photographs. R262C can cause bilateral ptosis and severe CFEOM3 with the resting position of both eyes infraducted and abducted (A), moderate CFEOM3 that can be unilateral (B), and mild CFEOM3 (not shown). A similar spectrum is seen with D417N; severe CFEOM3 is shown in (E). A302T (C) and R380C (D) cause moderate to severe CFEOM3. Participants in (A–E) have full facial movements. The axonal neuropathy in the participant with D417N (E) results in atrophy of the intrinsic foot muscles and a high arch (F). E410K (G) and R262H (H) result in severe CFEOM3 and facial weakness, and R262H also results in congenital ulnar deviation of the hand with joint contractures of the thumbs and fingers (I). (J–L) MRI of the brainstem at the level of the oculomotor nerve (J) and orbital contents posterior to the globe (K) in a participant with predominantly left-sided CFEOM3 and a D417N substitution. Note unilateral hypoplasia of the left oculomotor nerve (J, arrow) and the atrophy of the levator palpebrae superioris (LPS), superior rectus (SR), and medial rectus (MR) muscles in (K). The inferior rectus (IR), lateral rectus (LR), and superior oblique (SO) muscles appear normal. (L) Control orbital MRI for comparison. (ON) denotes optic nerve.

Figure 2. Spectrum of human brain malformations correlate with specific TUBB3 mutations

(A–G) Midline sagittal MRI showing the spectrum of corpus callosum (CC) dysgenesis; corresponding amino acid substitutions are noted to the left. R62Q (A) and most R262C (B) participants have normal CC development, whereas D417N subjects have hypoplasia of the posterior body (C, arrow). Subjects with A302T, E410K, and R262H have diffuse CC hypoplasia (D–F). (G) Both R380C siblings have CC agenesis, and brainstem (arrow) and mild vermian hypoplasia (asterisk). (H–N) Axial MRI from the same patient scans showing the spectrum of anterior commissure (AC) dysgenesis and overall loss of white matter compared to the normal R62Q scan (H, arrow indicates AC). (I–L) Subjects have hypoplastic AC. R262H (M) and R380C (N) patient scans show AC agenesis and dysmorphic basal ganglia. The anterior limb of the left internal capsule is hypoplastic in R262H (M, arrow), while there is lack of clear separation between the caudate and putamen and bilateral hypoplasia of the anterior limbs of the internal capsule with R380C (N, arrows). (A′-H′) Anterior (A′-D′) and posterior (E′-H′) coronal sections showing the spectrum of basal ganglia dysmorphisms present in individuals with R262C, R262H, and R380C substitutions (Figure 3A′-H′). (A′-D′) Compared to a TUBB3+/+ scan (A′), R262C reveals asymmetric basal ganglia with enlargement of the left caudate head and putamen (B′ arrow). (C′) The twin of the R262H patient in (F, M) has dysgenesis of the left and right anterior limbs of the internal capsule (C′ arrows), apparent fusion of an enlarged left caudate head with the putamen, and dilatation of the left and right anterior horn of the lateral ventricle. (D′) The older sibling of the patient scanned in (G, N) harboring an R380C substitution has hypoplasia of the anterior limb of the internal capsule (D′ arrow) and fusion of the left caudate head and underlying putamen. (E′-H′) Coronal MRI at the level of the caudate body and lateral ventricles is normal in (E′). R262C and R262H subjects have hypoplasia of the left caudate body (F′, G′, arrows) and tail, and the R262H patient has dilatation of the left lateral ventricle. (H′) The R380C patient has bilateral hypoplasia of the caudate body and tail, with Probst bundles of callosal axons that line the bodies of the lateral ventricles (arrow heads).

Figure 3. TUBB3R262C/R262C mice have normal cortical layering but show defects in axon guidance

(A–D, n=5) WT and TUBB3R262C/R262C (E–H, n=5) E18.5 coronal sections immunostained with markers specific for cortical layers show that the cortex has developed properly. Mild midline changes result from large Probst bundles (pb), comprised of stalled commissural axons adjacent to the midline. (I and K) Coronal sections from E18.5 WT (I, n=4) and TUBB3R262C/R262C (K, n=5) embryos show that the anterior commissure (red arrow) appears broken and fails to cross the midline, while the CC has crossed but is abnormally thick (red arrowhead) in the mutant. (J and L) E18.5 coronal sections from embryos immunostained with the axonal marker L1 show Probst bundles in a TUBB3R262C/R262C (L, arrowheads) mutant compared to WT (J). (M–P) Whole-mount neurofilament staining of E11.5-12 WT (M, N; n=13) and TUBB3R262C/R262C (O, P; n=6) embryos. Mutant oculomotor (III) and trochlear (IV) nerves, as well the maxillary (Vm) and ophthalmic (Vo) divisions of the trigeminal nerve are stalled at E11.5 (O) compared to WT (M). At E12, the mutant oculomotor nerve follows an aberrant course adjacent to the trochlear nerve (P) compared to WT (N, ). CC = corpus callosum; e = eye; asterisk (*)= distal tip of oculomotor nerve.

Figure 4. TUBB3R262C/R262C mice have low TUBB3 protein levels, altered microtubule stability, and decreased Kif21a interactions

(A) TUBB3 protein levels are reduced in TUBB3+/R262C (WT/KI) and TUBB3R262C/R262C (KI/KI) vs. WT mice. (B) TUBB3 R262C heterodimers incorporate into microtubules throughout cell bodies, neurites, and growth cones of dissociated cortical neurons as seen by co-localization with tyrosinated -tubulin and actin. Variable reductions in TUBB3 staining intensity are noted between WT and mutant neurons. (C) Levels of de-tyrosinated -tubulin are increased in brain lysates from mutant versus WT mice. (D) Brain lysates from TUBB3R262C/R262C mice show increased microtubule polymerization at steady-state levels despite lower levels of β-tubulin. Mutant TUBB3 is detected in the pellets (p). (E) Levels of Kif21a are reduced on TUBB3R262C/R262C mutant microtubules polymerized in vitro from brain lysates and incubated with ATP, whereas levels of Kif3a remain constant. (F) In vitro transcription and translation of WT and TUBB3 mutant heterodimers in rabbit reticulocyte lysate. Products analyzed by SDS (top) and non-denaturing (native, bottom) gel electrophoresis and stained with an anti-V5 antibody against the C-terminal tag demonstrate that although transcription and translation are not affected by the mutations, there can be significant and variable decreases in the yield of native heterodimers. (G) Synthesized WT, R262H, E410K, and D417H/N heterodimers cycle with native bovine tubulin at equivalent efficiency; vertical lines denote removed empty lanes. *P<0.05, **P<0.001.

Figure 5. TUBB3 amino acid substitutions and phenotype-genotype-function correlations

(A) Schematic of the TUBB3 protein; arrows indicate the location of each TUBB3 amino acid substitution. (B) Structure of the αβ-tubulin heterodimer from a rotated side view (pdb: 1JFF). α-tubulin is dark-shaded on right, and the arrows on the left point to each mutant residue, depicted in black, on β-tubulin. (C) Outside view of β-tubulin depicting each mutant residue in black with the exception of R62Q, which cannot be seen. The area proposed for motor protein interactions is depicted as a shaded grey oval. Residues R380, E410 and D417 are found in helices H11 and H12, respectively, on the external surface of β-tubulin. R262, found in the loop between helix H8 and strand 7, and A302, found in the loop following helix H9, are positioned laterally and below to helices H12 and H11, respectively. (D) Side interior view of β-tubulin showing the location of each mutant residue in black, with the exception of R380C, which is occluded by D417. R62 is positioned in the H1-S2 loop (N-loop) that forms lateral contacts with the M-loop of adjacent protofilaments. (E, top) Magnified image of β-tubulin depicting the putative hydrogen bond between the arginine side chain of R262 and the carbonyl oxygen of D417, both of which are shown in a stick representation. (E, middle and bottom) Both the R262C and R262H substitutions are predicted to break the hydrogen bond. (F) Phenotype-genotype and summary of functional data. AA = amino acid substitution; # ped = number of pedigrees; # aff = number of affected individuals; AC = anterior commissure hypoplasia; CC = corpus callosum hypoplasia; BG = basal ganglia dysgenesis; DD = developmental delay; FW = facial weakness; PN = progressive axonal sensorimotor polyneuropathy; CJC = congenital joint contractures; HD = heterodimer formation; HeLa = HeLa cell incorporation; MT dyn = microtubule dynamics (HS = highly stable, S = stable); Kip3p/2p = Kip3p and Kip2p microtubule plus-end accumulation ( ↓ decreased, ↑ increased). + denotes present; - denotes absent; −/+ denotes that only a subset of subjects have the feature and/or the findings are mild; variable (−) denotes rare participants with the substitution without CFEOM3.

Figure 6. TUBB3 disease amino acid substitutions results in changes to microtubule dynamic instability

(A, B) 4-D life-time plots depicting the lengths of microtubules (Y axis) over time (X axis) in G1 cells from (A) haploid WT and TUB2 mutants and (B) heterozygous diploid WT and TUB2 mutants demonstrate that Tub2p substitutions perturb microtubule dynamic instability. For WT and each mutation, one microtubule representing data from the collective analysis has been selected and plotted. (C) Summary table of individual dynamic instability parameters. Number of events is listed in parentheses. *P<0.05, **P<0.001, ***P<0.0001.

Figure 7. Kip3p and Kip2p levels are reduced at the plus-ends of mutant microtubules

Merged z-stack images showing the levels of Kip3p-3YFP (A–E, red) and Kip2p-3YFP (K–O, red) on WT and mutant microtubules labeled with CFP-Tub1p (α-tubulin, green) in budding yeast. Corresponding Kip3p (F–J) and Kip2p (P–T) YFP channels with signal intensities adjusted equally for WT and each TUB2 mutant are provided for comparison. In WT cells, Kip3p-3YFP (A, F) and Kip2p-3YFP (K, P) are speckled along the length and accumulate at the plus-ends of growing microtubules labeled by white arrows. In contrast, mutant R262C (B), R262H (C), E410K (D), and D417N (E) cells all have a significant reduction of Kip3p-3YFP along the length and plus-ends of microtubules (white-arrows), and Kip2p-3YFP is speckled along the length but reduced or absent on microtubule plus-ends (white-arrows) in R262C (L), R262H (M), E410K (N), and D417H (O) cells. Graphs depicting the overall mean levels of Kip3p-YFP (U) and Kip2p-3YFP (V) on microtubule plus-ends in WT and mutant cells. *P<0.05, ** P<0.001, *** P<0.0001.