|Home | About | Journals | Submit | Contact Us | Français|
A spectrum of neurological disorders characterized by abnormal neuronal migration, differentiation, and axon guidance and maintenance have recently been attributed to missense mutations in the genes that encode α– and β-tubulin isotypes TUBA1A, TUBA8, TUBB2B, and TUBB3, all of which putatively co-assemble into neuronal microtubules. The resulting nervous system malformations can include different types of cortical malformations, defects in commissural fiber tracts, and degeneration of motor and sensory axons. Many clinical phenotypes and brain malformations are shared among the various mutations regardless of structural location and/or isotype, while others segregate with distinct amino acids or functional domains within tubulin. Collectively, these disorders provide novel paradigms for understanding the biological functions of microtubules and their core components in normal health and disease.
Twenty-three heterozygous missense mutations in TUBA1A have been reported [1–7]. All are sporadic and likely de novo, and four have been found in more than one unrelated individual. Most children harboring these mutations have microcephaly, severe motor and intellectual disabilities, and seizures. A subset also has simple strabismus and/or facial weakness, but none has paralytic strabismus or congenital fibrosis of the extraocular muscles (CFEOM). Fetopsy and imaging studies reveal that most TUBA1A mutations cause various grades of classic lissencephaly, ranging from the complete loss of gyri and sulci (agyria) to brains with simplified, abnormally thick convolutions (pachygyria). Cortical lamination is significantly disturbed and the six-layer cortex is reduced to four thicker layers, and in some individuals, only two thin layers. The hippocampus and brainstem are usually malformed and hypoplastic. There is typically partial or complete absence of the corpus callosum and hypoplasia of the internal capsule and corticospinal tract associated with dysmorphic basal ganglia. Cerebellar vermian hypoplasia is also prominent, and TUBA1A mutations are estimated to account for approximately 30% of lissencephaly associated with cerebellar malformations  (Table).
Five heterozygous missense mutations in TUBB2B that are all sporadic and de novo have been reported . Similar to the TUBA1A phenotype, children harboring TUBB2B mutations are typically microcephalic and have severe motor and intellectual disabilities often accompanied by seizures. Fetopsy and imaging studies reveal perturbed cortical cell migration and abnormal development of radial glia. Instead of classic lissencephaly, however, these children have polymicrogyria, a brain malformation characterized by excessive small gyri that are separated by shallow sulci, giving the brain a cobblestone-like appearance. The polymicrogyria in affected individuals is bilateral, asymmetric, and typically more predominant in the frontal and temporal lobes. Similar to TUBA1A mutations, partial or complete agenesis of the corpus callosum, dysmorphisms of the basal ganglia and cerebellum, and brainstem hypoplasia are common radiological findings (Table).
TUBB3 mutations cause a more diffuse spectrum of brain malformations and neurological disabilities, and certain phenotypes often segregate with particular amino acid substitutions [9,10]. Fourteen heterozygous missense mutations have been reported, which are a mixture of familial and de novo mutations, and six have been identified in more than one unrelated individual. The first eight TUBB3 missense mutations to be reported cause the paralytic eye movement disorder CFEOM3, which results from hypoplasia of the oculomotor nerve(s) and secondary atrophy of extraocular muscles. Depending upon the specific amino acid substitution, patients can develop sensorimotor polyneuropathy due to the progressive degeneration of motor and sensory axons in the limbs, and are sometimes born with wrist and finger contractures, facial paralysis, and mild to moderate intellectual and behavioral disabilities. Brain malformations include agenesis or hypoplasia of commissural axon tracts, hypoplasia of the corticospinal tract, and dysmorphic basal ganglia with fusion of the caudate and putamen. Overall, this initial combination of clinical and radiological findings point to a generalized defect in axon guidance and maintenance  (Table).
More recent findings have expanded the TUBB3 phenotypic spectrum. Six additional TUBB3 missense mutations have been identified in individuals with predominant frontal polymicrogyria or simplified and disorganized gyral patterning . Five of the mutations alter different residues from the initial set of TUBB3 mutations, while one results in a different substitution at the same residue. Remarkably, none of these patients have CFEOM3, facial paralysis, or signs of sensorimotor polyneuropathy, although most have intermittent or permanent non-paralytic strabismus. Similar to TUBA1A and TUBB2B mutations, brainstem and cerebellar vermian hypoplasia are also common; however, these are not prominent findings in the initial series of TUBB3 mutations . Both sets of TUBB3 mutations commonly cause corpus callosum dysgenesis and corticospinal tract hypoplasia, and fetopsy and diffusion tensor imaging (DTI) from the second set of TUBB3 mutations further support a generalized defect in axon guidance . Basal ganglia dysmorphisms are also present, and thus appears to be a defining phenotype of the dominant tubulin related disorders (Table).
Finally, two consanguineous pedigrees segregating severe developmental delay, seizures, and optic nerve hypoplasia as a recessive trait have been reported to harbor the same homozygous 14 base pair intronic deletion upstream of exon 2 of the TUBA8 gene, altering transcript splicing and resulting in greatly reduced TUBA8 protein levels in patient derived lymphoblastoid cells. Brain imaging of affected family members revealed extensive bilateral polymicrogyria, dysplastic or absent corpus callosa, and brainstem dysmorphisms . Although these phenotypes converge with those described above, additional mutations are needed to definitively support TUBA8 as a polymicrogyria associated gene.
Microtubules are dynamic polymers comprised of tandem repeats of αβ tubulin heterodimers, which assemble in a head to tail fashion at the growing ends of microtubules to form a sheet of longitudinal protofilaments. Lateral interactions between neighboring protofilaments cause the sheet to close, thereby forming the hollow, cylindrical microtubule body [12,13]. The structural conformation of longitudinal protofilaments is tightly regulated and, therefore, the structures of α– and β–tubulin are highly conserved throughout eukaryotes.
Tubulin is comprised of three separate structural domains that are formed by β-sheets that alternate with α-helices; these are the N-terminal, intermediate, and C-terminal domains and, in β–tubulin, correspond to residues 1–229, 230–371, and 372–450 respectively . These three structural domains serve at least five functions (Figure 1). The GTP binding pocket is formed by residues in the N-terminal structural domain (Figure 1A); GTP binding is important for protein folding, the structure and stability of tubulin heterodimers, and the conformation of longitudinal protofilaments necessary for lateral interactions and microtubule nucleation. Interactions between the N-terminal domain and the adjacent intermediate domain participate in structural rearrangements resulting from the hydrolysis of GTP bound by β–tubulin (Figure 1B). GTP hydrolysis causes straight protofilaments to curl outwards, resulting in microtubule depolymerization [14, 15]. Other residues within this domain mediate longitudinal (Figure 1C) and lateral (Figure 1D) interactions that are necessary for both heterodimer and microtubule stability. Finally, motor protein and microtubule-associated protein (MAP) interactions occur through residues at the C-terminus that form alpha helices on the external surface of tubulin (Figure 1E). These residues mediate interactions with kinesin and dynein motors that facilitate intracellular transport in the anterograde and retrograde directions, respectively, as well as other MAPs that extrinsically regulate the dynamic properties of microtubules. Certain residues found in this domain are also important for the stability of longitudinal protofilaments [14,16–18].
Disease-causing amino acid substitutions in TUBA1A, TUBB2B, and TUBB3 are widely distributed among the three domains of tubulin and are therefore predicted to perturb different microtubule functions according to their structural locations (Figure 2). Several mutations alter residues that interact directly with the GTP nucleotide (TUBB2B S172P and TUBB3 T178M) or those located directly adjacent (TUBB2B L228P and TUBB3 E205K). Other mutations are located at contact surfaces between the intra- and inter-heterodimer and thus are predicted to alter longitudinal protofilament interactions (Figure 2, row 1). Both of these regions of tubulin are highly complementary and contain stretches of absolutely conserved residues; thus, a single amino acid substitution affecting one of these residues could be expected to significantly impede the formation of heterodimers, the overall ability of microtubules to polymerize, or the conformational changes essential to the dynamic properties of microtubules . Another subset of mutations affects residues that lie in close proximity to the interface between the N-terminal and intermediate domains (Figure 2, row 1) . Likewise, these amino acid substitutions may also affect heterodimer stability and/or the dynamic properties of microtubules by altering the structure of tubulin during nucleotide exchange and hydrolysis. Overall, most of these types of mutations cause a range of gyral malformations, suggesting that microtubule stability may be diminished during key processes of cell migration, such as the coupling of the nucleus with the centrosome [19,20].
Several mutated residues in TUBA1A and TUBB3 fall within or immediately adjacent to the regions that participate in lateral interactions. Lateral interactions occur primarily between N-terminal loops of one heterodimer with the large loops in the intermediate domains of an adjacent heterodimer in the flanking longitudinal protofilament (Figure 1D, Figure 2, row 1) . These interactions permit the assembly of microtubules, regulate their growth and shortening properties, and also provide the stabilizing force that allows them to curve and bend without breaking . In neurons, this is particularly important because microtubules are arranged in dense networks, and in order to maintain their structural integrity, they must be resistant to forces that cause bending or buckling as is commonly seen in vivo . At least one of these residues, R62Q in TUBB3 associated with isolated CFEOM3, was found to diminish the growth and shortening properties of microtubules and made them more resistant to depolymerization when introduced into yeast β–tubulin . The remaining mutations cause gyral and other brain malformations, but it is not known how they affect the stability or dynamic properties of microtubules.
Finally, many substitutions mutate residues on the surface of microtubules that mediate protein interactions with kinesin, dynein, and other MAPs. For example, DCX (doublecortin) is a MAP that causes X-linked lissencephaly, and it has recently been shown to interact with α-tubulin R264, a recurrent mutation found in TUBA1A (Figure 2B, row 2) [23,24]. Intriguingly, several of these types of mutations are associated with distinct neurological impairments and/or brain malformations that suggest specific perturbations to developmental pathways involved in cell migration or axon guidance and maintenance. For example, recurrent R402C and R402H substitutions found on the external surface of TUBA1A (Figure 2B, row 2) cause a phenotype indistinguishable from the classical form of lissencephaly resulting from LIS1 mutations or LIS1 and YWHAE co-deletions (Miller-Dieker syndrome), respectively [7,25]. LIS1 forms a large protein complex with dynein that is essential for proper neuronal migration and leading neurite extension, suggesting R402 is particularly important for some aspect of LIS1/dynein function [7,26,27]. TUBB3 residues associated with CFEOM3 predominately affect amino acids that are important for kinesin-microtubule interactions (Figure 2A, row 2) [9,17]. Many of these mutations are associated with sensorimotor polyneuropathy, underscoring the association between neurodegeneration and motor trafficking defects, as well as CFEOM3 . CFEOM1, a nearly indistinguishable eye movement disorder, results from heterozygous mutations in the motor and stalk regions of the kinesin KIF21A, suggesting its involvement in the disease [9,29,30]. Considering that phenotypic distinctions can occur between those mutations that affect kinesin and/or MAP interactions, this provides an opportunity to study how specific residues on the surface of microtubules differentially regulate the vast number of protein interactions .
All mutations reported thus far in TUBA1A, TUBB2B, and TUBB3 have been heterozygous missense mutations. Missense mutations in the absence of nonsense, frameshift, or genomic deletions support altered protein function rather than haploinsufficiency as a primary genetic etiology of these tubulin related disorders, and this is reinforced by phenotype-genotype correlations associated with recurrent mutations [9,31]. By contrast, homozygous splice-site mutations in TUBA8 delete amino acids that are necessary for the structural integrity of tubulin and likely result in loss of protein function; however, heterozygous carriers do not have reported phenotypes, suggesting nervous system development is presumably more or less normal in the context of TUBA8 haploinsufficiency . Thus, with the exception of TUBA8, several lines of evidence support dominant-negative effects rather than tubulin haploinsufficiency as sufficient to cause the various types of neurological and structural brain impairments.
Other findings, however, suggest that some mutations cause a reduction in the overall amount of tubulin heterodimers containing the mutant isotype, adding an important caveat to this argument. In vitro assays have revealed that approximately half of the amino acid substitutions in TUBA1A, TUBB2B, and TUBB3 that have been tested result in a significant decrease in the production of functional tubulin heterodimers due to the failure to properly interact with one or more chaperone proteins that capture and fold nascent tubulin polypeptides [8–10,32–34] (Figure 3). Upon overexpression in mammalian cells, some of these epitope tagged tubulin mutants demonstrate a moderate to severe decrease in microtubule incorporation as would be expected from compromised folding. These results are confounded by the ability of many of these poorly folded mutant heterodimers to still incorporate into interphase microtubules at levels equivalent to wild type, although this could be an artifact from overexpression. Nonetheless, knockdown of Tubb2b in the embryonic rat cortex by in utero RNAi has reduced protein levels by 60%, causing arrested cell migration in a manner reminiscent of the human TUBB2B phenotype .
There are not always clear phenotypic differences between those tubulin mutations that significantly diminish heterodimer synthesis and microtubule incorporation versus those that permit levels equivalent to wild type. However, in some instances, there appears to be a trend among recurrent mutations such that increased levels of microtubule incorporation correlates with more widespread and often severe nervous system malformations. This is best demonstrated by different recurrent mutations altering the same residue. For example, the most common TUBB3 substitution, R262C, predominately results in isolated eye movement restrictions, whereas R262H causes severe eye movement restrictions in addition to other neurological impairments and brain malformations. Introducing both of these mutations into yeast β-tubulin similarly altered the growth and shortening properties of microtubules and diminished kinesin-microtubule interactions; however, in vitro folding assays and overexpression in mammalian cells demonstrated that the R262H substitution resulted in the generation of more tubulin heterodimers with higher levels of microtubule incorporation than R262C . Perhaps in a similar manner, recurrent TUBA1A substitutions R402H and R402C are distinguishable because the former usually cause a more severe lissencephaly phenotype with complete agyria versus R402C, and also permit higher levels of heterodimer formation and microtubule incorporation [7,33]. Thus, in some circumstances, the manifestation and/or severity of nervous system impairments may depend on the relative abundance of mutant heterodimers compared to wild-type, combined with their ability to incorporate into the microtubule cytoskeleton such that dynamics, motor protein, and/or MAP interactions are affected in different dominant-negative fashions.
Many questions still remain with regard to how the different mutations in α– and β–tubulin cause the range of reported neurological impairments and structural brain malformations. At present, much of the evidence supports that the dynamic properties and functions of microtubules are altered in several fashions. These include diminishing the overall abundance of functional tubulin heterodimers, altering GTP binding, altering longitudinal and lateral protofilament interactions, and impairing microtubule interactions with kinesin, dynein, and other MAPs (Figure 3). It is conceivable that each mutation disturbs one or more of these processes to different extents, and therefore, it is often difficult to predict phenotypic outcomes based solely upon structural locations and putative functional effects. Nonetheless, certain phenotypes including classic lissencephaly, CFEOM3, and polyneuropathy are unique or more common to mutations in one particular isotype or in different functional domains within tubulin. This is intriguing because expression levels are known to vary across isotypes according to cell type and developmental stage, and there are reported differences in their subcellular localizations and ability to influence microtubule dynamics [35–37]. Therefore, all of these factors should be considered when designing experiments to examine the function of different tubulin isotypes and their disease-related amino acid substitutions in the developing nervous system.
This work was funded by National Institutes of Health NEI R01EY12498. Dr. Engle is an Investigator of the Howard Hughes Medical Institute.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.