The α-Tubulin gene TUBA1A in Brain Development: A Key Ingredient in the Neuronal Isotype Blend

Microtubules are dynamic cytoskeletal polymers that mediate numerous, essential functions such as axon and dendrite growth and neuron migration throughout brain development. In recent years, sequencing has revealed dominant mutations that disrupt the tubulin protein building blocks of microtubules. These tubulin mutations lead to a spectrum of devastating brain malformations, complex neurological and physical phenotypes, and even fatality. The most common tubulin gene mutated is the α-tubulin gene TUBA1A, which is the most prevalent α-tubulin gene expressed in post-mitotic neurons. The normal role of TUBA1A during neuronal maturation, and how mutations alter its function to produce the phenotypes observed in patients, remains unclear. This review synthesizes current knowledge of TUBA1A function and expression during brain development, and the brain malformations caused by mutations in TUBA1A.


Introduction
Brain development is a highly complex process that requires careful coordination of cell proliferation, differentiation, migration, axon and dendrite growth and guidance, and synapse formation. Each of these cellular tasks rely on the proper regulation of microtubules, which are dynamic cytoskeletal polymers that help provide structure and generate force in all eukaryotic cell types. Neurons require elaborate networks of microtubules to support their complex morphologies and to perform a variety of critical functions. In developing neurons, dynamic microtubules are necessary to support essential functions such as morphological changes that underlie polarization and migration [1][2][3]. In particular, microtubules are key players in the initiation and extension of neurites, axon specification, and neuronal migration. In adult neurons, microtubules provide a stable backbone for established axons and dendrites, acting as essential trafficking routes for microtubule-based motors to carry various cargos from one compartment of the neuron to another. Microtubules also play an important role in synaptic plasticity, where they facilitate the morphological changes to dendritic spines that are thought to underlie synaptic enhancement and depression [4][5][6]. How microtubules are regulated to support distinct roles in different compartments of the neuron and at different stages of development to adulthood is widely unknown.
Exploring how TUBA1A contributes to neuronal function will provide important insight into how cells "tune" their microtubule networks using the programmed expression of a specific tubulin isotype. A major outstanding question in the microtubule field is how this tuning is achieved at the molecular level. Distinct tubulin isotypes could lead to outright functional differences due to intrinsic changes to the tubulin protein structure, which could alter microtubule dynamic instability or extrinsic binding of important microtubule-associated proteins (MAPs). Alternatively, carefully timed expression of a specific tubulin isotype could increase the concentration of tubulin present in the cell, supplying a necessary burst of new microtubule assembly during morphological changes such as axon outgrowth. In this review, we will discuss the current understanding of the regulation of microtubule function during neuronal development. We will particularly focus on TUBA1A, discussing its regulated expression, mutations associated with brain malformations, and how this α-tubulin might uniquely contribute to normal brain development.

Microtubule Basics
Tubulin, the fundamental protein subunit of the microtubule, is a heterodimer of αand β-tubulin polypeptides. αand β-tubulins are conserved across eukaryotic evolution and are related to filament-forming proteins in prokaryotes and bacteriophage [22,23]. These proteins share the common characteristic of polymerizing into dynamic filaments, and this characteristic depends on their ability to bind and hydrolyze GTP. The tubulin heterodimer binds to two molecules of GTP at two separate sites. The so-called "non-exchangeable" or "N-site" is located at the intradimer interface, between αand β-tubulin [24] ( Figure 1A). GTP bound to the N-site is not hydrolyzed and exchanges at a slow rate [25]. In contrast, the "exchangeable", or "E-site", is located at the interdimer interface formed by the β-subunit of one heterodimer and the α-subunit of a neighboring heterodimer [24] ( Figure 1A). GTP at the E-site can be hydrolyzed to GDP and subsequently exchanged for a new GTP nucleotide [26]. Importantly, the nucleotide status at the E-site plays a determining role in the conformation of the heterodimer and its interactions with other heterodimers (see description of α/β-Heterodimers polymerize into a sheet-like conformation, known as the microtubule lattice ( Figure 1B). The lattice consists of longitudinal chains of heterodimers arranged head-to-tail, called protofilaments. Longitudinal interactions involve an extensive hydrophobic interface between αtubulin of one heterodimer and β-tubulin of the adjacent heterodimer, which completes the E-site [32]. Protofilaments bind along their lateral sides to other protofilaments. Lateral interactions involve α-tubulin of one protofilament binding to the α-tubulins of neighboring protofilaments, and βtubulin binding to neighboring β-tubulins. In contrast to longitudinal interactions, the interfaces of lateral interactions consist of flexible loop regions of α-and β-tubulin and feature prominent electrostatic contributions [28,32]. The binding energy of lateral interactions is predicted to be much weaker than longitudinal interactions, based on computational modeling [33,34]. In most cells, the α/β-Heterodimers polymerize into a sheet-like conformation, known as the microtubule lattice ( Figure 1B). The lattice consists of longitudinal chains of heterodimers arranged head-to-tail, called protofilaments. Longitudinal interactions involve an extensive hydrophobic interface between α-tubulin of one heterodimer and β-tubulin of the adjacent heterodimer, which completes the E-site [32]. that of kinesin. The dynein mechanism involves a coordinated progression of conformational changes to mechanically link the nucleotide cycle of its AAA + ATPase domains to a microtubule-binding domain that is located~15 nm away [44]. Furthermore, dyneins predominantly move toward the minus ends of microtubules, while most kinesins move toward the plus ends. Dynein motility is regulated by an assortment of dynein-binding proteins that regulate its speed, level of force production, and cargoes [45][46][47][48][49][50][51][52][53][54]. This regulation may explain how a single cytoplasmic dynein gene could be necessary for diverse roles in cells, while kinesins have undergone evolutionary diversification giving rise to 45 genes with specified functions. Because of the differences between dyneins and kinesins, it is likely that these motors use different mechanisms to bind and move along the microtubule surface. However, the contribution of the microtubule surface to motor activity and regulation is poorly understood.
Although much research in the microtubule field focuses on MAPs and motors that regulate microtubule networks by binding and moving along them, it is becoming clear that tubulin heterodimers themselves are intrinsic key regulators. Tubulins exhibit molecular differences that can be genetically encoded through different αand β-tubulin genes, or conferred by posttranslational modifications. These molecular differences provide cells with a toolkit for changing the properties of microtubule networks.

Tubulin Isotypes
Nearly all eukaryotes express multiple, distinct genes for αand β-tubulin, known as tubulin isotypes. Recent analysis of the human genome identified nine α-tubulin isotypes and ten β-tubulin isotypes, along with dozens of pseudogenes [7,8]. In addition to their different chromosomal locations, isotypes can be distinguished by three features: (1) the amino acid sequences they encode; (2) nucleotide sequences of the 3 untranslated region (UTR); and (3) expression levels in different tissues or developmental stages. We will focus on key differences amongst the α-tubulin isotypes.
The amino acid sequence of α-tubulin is strongly conserved across eukaryotic evolution. For example, the human α-tubulin TUBA1A exhibits 86% sequence identity to α-tubulin in the unicellular eukaryote Giardia lambia and 75% identity to the α-tubulin in the budding yeast, Saccharomyces cerevisiae. Within the human α-tubulin isotypes (Table 2), there are a small number of amino acid sequence differences. Importantly, these differences are conserved in isotype homologues across species, suggesting that selective pressure may maintain isotype-specific sequence differences ( Table 2). The majority of these differences are found in the 15-27 amino acids at the very carboxy-terminus; a region known as the Carboxy-Terminal Tail (CTT). CTTs decorate the outer surface of the microtubule, contain an abundance of amino acids with negatively-charged side chains, and are major sites of post-translational modifications (PTMs), which will be discussed in the next section. Current models propose that the molecular diversity generated by genetically-encoded differences and PTMs in the CTTs act as a "tubulin code" that regulates the activities of MAPs and motors at the microtubule surface [55,56]. Consistent with this model, several studies show that altering the amino acid sequence within the CTT causes changes in microtubule function in vivo [57][58][59][60][61]. Besides the CTTs, human α-tubulin isotypes also exhibit amino acid differences at other positions. For example, the human TUBA1A and TUBA1B isotypes have identical CTTs, but have different amino acids at position 232 and 340. Position 232 is buried deep within α-tubulin, while position 340 lies on the microtubule surface, near the interdimer interface [62]. Whether these amino acid differences lead to functional differences between TUBA1A and TUBA1B has not been investigated. This exemplifies a general deficit in our understanding of α-tubulin isotypes-although we have known of amino acid differences between α-tubulin isotypes for over 30 years, the functional consequences of these differences remain largely unexplored.  [63]. Gray denotes residues that differ from TUBA1A.
In contrast to the highly similar coding sequences of α-tubulin isotypes, the 3 -UTR regions are highly divergent. Vertebrate tubulin isotypes were originally identified from cDNA clones, and it was noted at that time that each isotype contained a distinct 3 UTR sequence [10,64,65]. What makes the differences in the 3 -UTR region especially intriguing is that they are conserved across species, with the 3 -UTR regions of human TUBA1A and TUBA1B sharing interspecies homology with rat Tuba1a and Tuba1b, respectively [64]. This conservation of the noncoding region implies selective pressure, and raises the question of how 3 -UTR regions contribute to function. Presumably, the 3 -UTR regions could differentially regulate mRNA stability and/or localization within a cell. A beautiful example of this regulation comes from studies in zebrafish, where the 3 -UTR of the β-tubulin isotype tubb5 targets the mRNA to axons and distal growth cones during development [66]. This mRNA targeting could provide an appealing mechanism for increasing the supply of tubulin heterodimers at a region of the cytoplasm that is far from the nucleus. Whether the 3 -UTRs of α-tubulin isotypes provide similar regulation awaits discovery.
The third distinguishing feature of tubulin isotypes is their pattern of expression across different tissues and developmental stages. The β-tubulin isotypes have been extensively mapped to different tissues, cell types, and, in some cases, sub-cellular localization [67][68][69][70][71]. We have a comparatively poor understanding of the distributions of α-tubulin isotypes, with the exception of TUBA1A. TUBA1A is strongly and specifically expressed in the developing nervous system, and provides over 95% of the α-tubulin in the embryonic brain [10,72]. We will extensively discuss TUBA1A expression and its regulation in subsequent sections. Spatial and temporal expression data on α-tubulin isotypes remain sparse and may need to be readdressed. Tracking the expression of tubulin isotypes at the protein level is particularly challenging due to the high degree of homology, and will benefit from new approaches to selectively label isotypes without impairing their functions.
Why different cell types express specific tubulin isotypes is a long-standing question. On one hand, the coding and non-coding differences between isotypes could impart functional differences within microtubule networks. The strongest evidence for specific functional roles for α-tubulin isotypes comes from studies of in Drosophila and C. elegans, which demonstrate isotype-specific requirements for generating proper axonemal structures within cilia and flagella [73][74][75]. These findings underscore the possibility that isotypes may play specific roles in building complex microtubule architectures and raise the question of whether specific isotypes could be required to build other complex microtubule structures, such as those in neurons. An alternative, but not mutually exclusive explanation for differential isotype expression is that it provides a convenient mechanism to regulate the levels of tubulin protein in a cell. This is a particularly important challenge considering that cells must balance the levels of αand β-tubulin to form heterodimers, and excess monomer, particularly β-tubulin, can be toxic [76][77][78]. Studies investigating tubulin isotypes have provided some answers to these questions, but our overall understanding of isotype biology is still in its infancy.

Tubulin Post-Translational Modifications
In addition to different isotypes, the tubulin subunits can be regulated by diverse PTMs. Various PTMs have been identified on neuronal microtubules, including detyrosination/tyrosination, polyglutamylation, acetylation, and polyamination. In some cases, these PTMs can further amplify genetically-encoded differences between isotypes, since the modified amino acid residues on αor β-tubulin are only found in a subset of isotypes. In this section, we will briefly summarize the current evidence of major classes of PTMs and their functional impacts on neuronal microtubules. For comprehensive reviews of tubulin PTMs, the reader is referred to several recent reviews [55,79,80].
Detyrosination/tyrosination refers to the enzymatic removal of tyrosine from the CTT of α-tubulin (detyrosination), and subsequent re-ligation (tyrosination). This tyrosine is genetically encoded by six α-tubulin isotypes in humans, including TUBA1A (Table 2). Although the enzyme that catalyzes the detyrosination reaction in vivo has not been identified, it is well-established that tyrosination is catalyzed by Tubulin Tyrosine Ligase (TTL), which exclusively acts on free heterodimers to ligate tyrosine onto a detyrosinated α-tubulin [81,82]. Antibodies that selectively bind to either tyrosinated or detyrosinated α-tubulin show that these PTMs can be differentially enriched on specific microtubules within a network, or specific regions of an individual microtubule, and appear to correlate with the age of the microtubule lattice [83][84][85]. Accordingly, neurons exhibit an enrichment of tyrosinated α-tubulin at regions containing more newly-assembled microtubules (e.g., neurites and the distal ends of axons) while detyrosinated α-tubulin is primarily enriched in regions with older and highly stable microtubules (the axon shaft) [86][87][88]. Neurons lacking the TTL enzyme undergo aberrant neuronal development, indicating an important role for the cycle of detyrosination/tyrosination [89]. Currently, there is little to no evidence that tyrosination status influences the intrinsic stability of a microtubule. In contrast, tyrosination status does impact the binding of various proteins to the microtubule surface. Cytoskeleton-Associated Protein Glycine-rich (CAP-Gly) domains selectively bind to EEY/F motifs, such as those found in the CTT of α-tubulins, and this binding strongly depends on the aromatic side chain of the terminal Y/F residue [58,[90][91][92]. Detyrosination therefore inhibits the microtubule-binding activity of CAP-Gly-domains. CAP-Gly domains are found in several microtubule-associated proteins, including CLIP170, the tubulin binding co-factors TBCB and TBCE, and the p150 glued subunit of dynactin. Recent studies show that the binding of p150 glued and CLIP170 to tyrosinated tubulin promotes the initiation of retrograde dynein-dependent transport in vitro and at the distal ends of axons [93,94]. Thus, detyrosination/tyrosination provides a system for local control of microtubule-MAP interactions and transport within a microtubule network.
Polyglutamylation is the most abundant tubulin PTM within the brain and involves the addition of variable-length chains of glutamate residues to genetically-encoded glutamates in the CTT regions of αor β-tubulin. Polyglutamylation was originally found to extend from α-tubulin peptide sequences that are only present in the TUBA1A and TUBA1B isotypes, with the glutamate chains extending from the γ-carboxyl group of the genetically-encoded glutamate at position 445 [63]. This may represent the primary polyglutamylation site on α-tubulin; however, alternative glutamates in the α-tubulin CTT may also be targeted for modification. The enzymes that catalyze polyglutamylation belong to the Tubulin Tyrosine Ligase Like (TTLL) family [95]. To date, twelve mouse TTLLs have been identified based on sequence homology, although only six of these exhibit polyglutamylase activity in vitro [96]. Among these, TTLLs 5, 6, 11, and 13 selectively modify α-tubulin [95][96][97]. The removal of glutamate residues is catalyzed by Cytosolic Carboxypeptidase (CCP) enzymes, which can either shorten polyglutamate chains or remove the genetically encoded, penultimate glutamate from detyrosinated α-tubulin, creating a truncated species known as ∆2-tubulin [98]. Similar to detyrosination/tyrosination, polyglutamylation is thought to alter the interactions of MAPs and motors at the microtubule surface. The clearest example is the regulation of microtubule severing by spastin. Here, the length of glutamate chains provides a tunable signal for directing spastin's severing activity [99,100]. Although the mechanistic role of polyglutamylation in neurons is still unclear, it appears to be developmentally regulated and correlate with degeneration. Both in vivo and in vitro experiments show that levels of polyglutamylation vary in different brain regions and increase over development, reaching their highest level in mature neurons [101]. Loss of function mutations affecting the CCP1 enzyme disrupt the normal program of tubulin PTMs in mouse neurons, increasing the amount of polyglutamylated tubulin and decreasing the ∆2-tubulin species [98]. CCP1 disruption causes Purkinje cell degeneration in mice, in a manner that depends on increased polyglutamylation [102][103][104]. This suggests an important yet poorly understand role for regulated polyglutamylation in neurons.
The presence or absence of specific PTMs can directly affect the stability of the microtubule. For example, the adult brain contains a much higher cold stable population of microtubules than the developing brain, and this increase in the proportion of cold stable microtubules in the brain has been attributed to an accumulation of polyamination on neuronal microtubules [105,106]. Polyamination describes the addition of polyamine to multiple, genetically-encoded glutamine residues in αand β-tubulin, by the transglutaminase enzyme TG2 [106]. Polyamination is sufficient to stabilize microtubules in vitro, and TG2 protein and activity levels increase postnatally, suggesting a role in neuronal maturation [106]. Acetylation of α-tubulin also correlates with microtubule stability, and staining with antibodies to acetylated tubulin is commonly used as a measure of microtubule stability in cells. However, while long-lived microtubules tend to be acetylated, the direct impact of acetylation on microtubule stability is not well established. Two recent studies indicate that acetylation of α-tubulin may promote microtubule stability through an unexpected mechanism-softening the microtubule lattice and allowing it to withstand bending without breaking and depolymerizing [107,108]. Studies in C. elegans demonstrate that microtubule acetylation is necessary to form specialized microtubules with 15-protofilament lattices in touch receptor neurons [109,110]. In addition to direct effects on the microtubule lattice, acetylation has been reported to promote the activities of kinesin-1 motors in vivo [111] and dynein motors in vitro [112]. This is intriguing, since the canonical acetylation site at lysine residue 40 of α-tubulin is located on the luminal side of the microtubule cylinder. How acetylation impacts lattice stability, interactions on the microtubule surface, and the larger implications of microtubule acetylation in vivo remain active areas of research.

Roles of Microtubules during Neuronal Development and Adulthood
Microtubules play numerous important roles during brain development, particularly in neurons, where the microtubule cytoskeleton has been intensely studied. Once neuronal progenitors exit the cell cycle to become neurons, diverse microtubule-based maturation stages must occur for the neuron to correctly extend dynamic neurites, migrate to the proper position, form and guide long axons, set-up synapses, and sustain the diverse regions of the neuron. Microtubules are essential for these phases of neuronal maturation and are differentially regulated by numerous MAPs to perform diverse neuronal microtubule-based functions ( Figure 2).
As shown in Figure 2, microtubules interact with numerous MAPs during each stage of neuronal maturation to properly perform various cellular tasks. During neurite initiation, dynamic actin forms lamellipodia that become stabilized by invading microtubules (Figure 2A) [113]. Interestingly, the dynamic properties of microtubules at this stage may inform which neurite becomes the future axon. Locally stabilizing microtubules in a particular neurite with stabilizing drugs results in that neurite becoming the axon [114]. Later, as neurites mature, microtubules become stabilized by MAPs such as tau and MAP2 ( Figure 2B) [39,115]. The presence of tau helps identify the neurite as the axon, while MAP2 identifies a dendritic fate. At this stage, motors such as kinesin-1 and dynein are important in sorting and pushing microtubules to the end of the neurite [116].
As development proceeds, neuronal microtubules begin to exhibit a distinct polarity and decoration by MAPs depending on which type of process they inhabit. Microtubules within the axon become uniformly oriented with their plus ends directed away from the soma ( Figure 2E), while dendritic microtubules retain a mixed polarity ( Figure 2H) [117,118]. Dynein has been implicated in establishing and maintaining this plus-end-out (away from the soma) orientation in axons [116].
Plus-end-out polarity is critical for axonal transport, discussed below. The quantity and diversity of MAPs in the neuronal environment, and the interplay between them makes understanding each MAPs' role difficult. For example, XMAP215 directly regulates microtubule plus-end dynamics in vitro, but in the growth cone it regulates the linkage of translocating microtubules to the F-actin network, thereby constraining microtubule growth velocity ( Figure 2G) [119]. Future studies are needed to determine how microtubules are regulated throughout each stage of neuronal maturation, with attention given to how microtubule dynamics are regulated by the plethora of MAPs present during each developmental stage. neurons, where the microtubule cytoskeleton has been intensely studied. Once neuronal progenitors exit the cell cycle to become neurons, diverse microtubule-based maturation stages must occur for the neuron to correctly extend dynamic neurites, migrate to the proper position, form and guide long axons, set-up synapses, and sustain the diverse regions of the neuron. Microtubules are essential for these phases of neuronal maturation and are differentially regulated by numerous MAPs to perform diverse neuronal microtubule-based functions ( Figure 2).
As shown in Figure 2, microtubules interact with numerous MAPs during each stage of neuronal maturation to properly perform various cellular tasks. During neurite initiation, dynamic actin forms lamellipodia that become stabilized by invading microtubules (Figure 2A) [113]. Interestingly, the dynamic properties of microtubules at this stage may inform which neurite becomes the future axon. Locally stabilizing microtubules in a particular neurite with stabilizing drugs results in that neurite becoming the axon [114]. Later, as neurites mature, microtubules become stabilized by MAPs such as tau and MAP2 ( Figure 2B) [39,115]. The presence of tau helps identify the neurite as the axon, while MAP2 identifies a dendritic fate. At this stage, motors such as kinesin-1 and dynein are important in sorting and pushing microtubules to the end of the neurite [116].  One of the most important functions of microtubules in adult neurons is to facilitate the efficient trafficking of organelles, proteins, mRNAs, and other cargoes across long distances in the axon. To appreciate the importance of efficient transport, consider that the distal regions of an axon can be as far as one meter away from critical protein and mRNA synthesis occurring in the soma. The unique microtubule polarity of the axon that is established early in development organizes transport in the anterograde and retrograde directions, with the help of motor proteins such as kinesin and dynein that move towards the plus or minus ends, respectively ( Figure 2F). Kinesins, particularly kinesin-1 and 3 family members, carry cargoes towards the axon terminal by walking towards the plus ends of axonal microtubules [120]. Conversely, dynein carries cargoes back toward the soma by walking towards the minus ends [121]. The combination of uniform microtubule orientation and processive motility by directional molecular motors allows neurons to effectively transport cargoes across large distances.
In addition to establishing the architecture and transport networks within individual neurons, microtubule function is critical for neuronal migration, which is required to form the layers of the cortex and many other structures in the brain. During neuronal migration, microtubules extend into the leading process where they help to steer the protrusive growth cone at the end of the axon ( Figure 2D) [122]. Microtubules also generate force to move the nucleus, in a process known as nucleokinesis. Here, microtubules extend back from the centrosome to form a cage-like network around the nucleus [123]. Dynein and its regulator LIS1 generate pulling forces to draw the nucleus toward the centrosome and move the centrosome toward the leading process [124]. Consistent with the important roles of the microtubule network in neuronal migration, mutations in LIS1 and the microtubule regulator doublecortin/DCX are associated with migration disorders that give rise to brain malformations [125].

Tubulinopathies Reveal Essential Role of TUBA1A in Brain Formation and Function
Recent studies investigating the genetic cause of brain malformation disorders have revealed that heterozygous, missense mutations to TUBA1A and other neuronal tubulin isotypes play an important role in brain development. First shown by Keays et al. in 2007, sequencing of tubulin genes has proven fruitful in uncovering the genetic source of numerous patients exhibiting complex neurological and physical phenotypes with cortical malformations. From these genotype-phenotype analyses, TUBA1A is recognized as vital for neurodevelopment based on the devastating effects observed in patients containing heterozygous, de novo TUBA1A missense mutations [11][12][13][14][15][16][17][18][19][20]. Brain malformation disorders caused by mutations to TUBA1A and other neuronally expressed tubulin isotypes are collectively termed "tubulinopathies", and lead to severe cortical abnormalities, mental retardation, and commonly epilepsy and paralysis [11][12][13][14][15][16][17][18][19][20]. Patients containing TUBA1A mutations exhibit a wide variety of cerebral cortex malformation phenotypes including lissencephaly, pachygyria, microlissencephaly, and polymicrogyria. While known genetic causes of these phenotypes hint at which developmental processes may be disrupted, little is known about how the TUBA1A mutations disrupt microtubule functions, or even how these disruptions could cause the larger-scale cellular and tissue problems seen in patients. The wide variety of brain malformations observed in patients leads to the prediction that different missense mutations in TUBA1A may disrupt different neuronal maturation phases. Disease-causing mutations to TUBA1A therefore provide a valuable opportunity to investigate the numerous neurodevelopmental stages that require TUBA1A, and how microtubules must be regulated for each stage to occur appropriately.
The identification of neurodevelopment disorder-causing TUBA1A mutations supports the hypothesis that this particular α-tubulin isotype is essential for neuronal maturation; however how can subtle changes to one α-tubulin protein lead to drastic changes in the formation of the brain? Expression studies of TUBA1A mRNA make it clear that TUBA1A is by far the most prevalent α-tubulin isotype in the embryonic nervous system, accounting for more than 95% of α-tubulin mRNA [72]. Thus, mutations to TUBA1A could greatly affect the available pool of neuronal tubulin. Tubulin mutations can act to alter numerous tubulin/microtubule characteristics and functions, but most of the possible changes fit into three general categories: (A) disrupting tubulin folding and heterodimer formation, thereby depleting the pool of α-tubulin that is competent to assemble microtubules; (B) disrupting polymerization activity and/or microtubule dynamics regulation, which could either deplete the pool of assembly-competent α-tubulin or alter dynamics once mutant heterodimers have formed the microtubule lattice; or (C) assembling appropriately into microtubules but altering microtubule function by disrupting interactions with MAPs and motors ( Figure 3). Identifying which category TUBA1A mutations fit into will greatly increase our understanding of tubulinopathy disease progression. This task is not easy, however, as structure-function predictions for tubulin residues are rarely straightforward. Tubulin is a complex protein that interacts with numerous binding partners and undergoes long range conformational changes as part of its principle biochemical activity. Therefore, knowing the location of a change in the tubulin sequence does not necessarily give insight into its functional consequences. This underscores the need for functional studies that can test and refine structural predictions. However, despite the growing list of tubulinopathy mutations, few studies provide insight into how individual mutations impact the tubulin protein and its function in vivo. Future studies must be performed to fill in the missing molecular, mechanistic steps between the known TUBA1A mutations and the final brain phenotype. Therefore, knowing the location of a change in the tubulin sequence does not necessarily give insight into its functional consequences. This underscores the need for functional studies that can test and refine structural predictions. However, despite the growing list of tubulinopathy mutations, few studies provide insight into how individual mutations impact the tubulin protein and its function in vivo. Future studies must be performed to fill in the missing molecular, mechanistic steps between the known TUBA1A mutations and the final brain phenotype. Either of these defects may produce haploinsufficiency/loss of function consequences, resulting in fewer polymerization competent tubulin heterodimers available to form microtubules, or changes in microtubule dynamics. This also may lead to changes in the ratio of α-tubulin isotypes present in the microtubule lattice; (C) TUBA1A mutations may lead to mutant tubulin heterodimers that appropriately polymerize and cause toxic, gain of function consequences from within the microtubule lattice. Once in the lattice, mutant dimers may intrinsically change microtubule behavior or extrinsically alter MAP binding. Either of these defects may produce haploinsufficiency/loss of function consequences, resulting in fewer polymerization competent tubulin heterodimers available to form microtubules, or changes in microtubule dynamics. This also may lead to changes in the ratio of α-tubulin isotypes present in the microtubule lattice; (C) TUBA1A mutations may lead to mutant tubulin heterodimers that appropriately polymerize and cause toxic, gain of function consequences from within the microtubule lattice. Once in the lattice, mutant dimers may intrinsically change microtubule behavior or extrinsically alter MAP binding.
While few studies have specifically addressed how mutations that disrupt TUBA1A cause cortical malformations, understanding the known mechanisms can provide clues into how attributes of TUBA1A contribute to normal brain development. However, there is much more work to be done to elucidate the mechanistic role of TUBA1A. Several different aspects of TUBA1A function could go awry to disrupt cellular processes and ultimately lead to abnormal brain development. For example, interference with neuronal migration can lead to lissencephaly, but not all the aspects of TUBA1A that are important for neuronal migration are known. For example, a mutation to TUBA1A could cause haploinsufficiency through many mechanisms, including disruptions to TUBA1A folding, heterodimer formation, or microtubule assembly. This could disrupt neuronal migration due to a pool of incompetent tubulin dimers forming less stable microtubules. Alternatively, specific TUBA1A mutants could appropriately assemble into microtubules but then cause dominant changes that "poison" the microtubule network. The dominant disruption could be achieved through changes to microtubule dynamics or disruption of binding of MAPs involved in migration. The following sections attempt to marry the current understanding of cortical malformation progression with how identified TUBA1A mutations could potentially act in the established pathway.

TUBA1A Mutations Linked to Lissencephaly
Lissencephaly describes a set of cortical malformations where at least part of the brain surface appears smooth, lacking the cortical folds that are a hallmark of a healthy human brain [126]. Severe lissencephaly manifests as a complete lack of cortical folds (agyria), while milder forms present as fewer broad folds (pachygyria) or as bands of heterotopic gray matter embedded below the white matter of the cortex (subcortical band heterotopia (SBH)). All of these manifestations of lissencephaly are attributed to neuronal migration errors in the developing cortex. In agyria and pachygyria, neurons fail to reach their appropriate positions leading to a disordered four-layered cortical structure lacking gyri or sulci folds, instead of the normal, six-layered cortex containing folds. In SBH, cortical neurons inappropriately migrate to an area deep to the cortex, forming band-like patterns of grey matter beneath the cortex [127].
In principle, any perturbation to cortical migration could lead to lissencephaly phenotypes, but most identified cases are attributed to mutations to LIS1 (also known as PAFAH1B1), doublecortin/DCX, and more recently TUBA1A [125]. These genes are encompassed by the cortical migration pathway, and provide the molecular basis for the malformation. LIS1 and DCX play important roles in regulating microtubule-based tasks. LIS1 is an adaptor for the microtubule-motor dynein, acting as a "clutch" that allows dynein to remain attached to microtubules for longer periods of time [128]. This modulation of dynein activity is important during neuronal migration, when dynein is responsible for pulling the microtubule-caged nucleus [129]. While LIS1 modulates dynein behavior, DCX binds directly to microtubules to stabilize and promote polymerization [130]. During migration, DCX facilitates the formation/maintenance of the microtubule cage around the nucleus, as well as stabilizing microtubules in the leading process of the migrating neuron [131]. These proteins are vital to migration, as disruption leads to defective migration, and overexpression of either LIS1 or DCX is sufficient to increase migration rates [131].
As lissencephaly-associated proteins are known to modulate the microtubule network, proper microtubule function must be requisite for correct cortical migration to occur. In fact, lissencephaly is the most prominent malformation attributed to TUBA1A mutations, with over 90% of patients exhibiting some form of lissencephaly malformation. This suggests that the consequence of many TUBA1A mutants is to disrupt migration in some way. The molecular basis of this disruption, or whether many different changes to tubulin function could lead to migrational defects, remains largely unknown. Table 3 provides the known TUBA1A mutations that cause lissencephaly.

TUBA1A Mutations Linked to Polymicrogyria
Polymicrogyria describes the cortical malformation characterized by excessive gyration (i.e., multiple, small folds) of the cerebral cortex [143]. The classification of this malformation is complex as cases of polymicrogyria are heterogeneous, with variable pathological results, clinical features, and etiologies. Even key characteristics used to define the malformation are controversial, as some sources point to abnormal cortical lamination as a key feature [144], while others affirm that cortical layering remains normal, and there are simply fewer neuronal populations inhabiting the layers [143]. A number of seemingly unrelated environmental and genetic causes have been implicated as the molecular basis of polymicrogyria, but the details of this developmental disorder remain mysterious.
The major non-genetic causes of polymicrogyria relate to ischemic insults in utero, such as hypoxia, hypoperfusion, congenital infections, and/or inflammation of the microvasculature [144]. Metabolic and mitochondrial diseases have also been implicated, with a mouse model of Zellweger syndrome revealing that polymicrogyria may be caused by defects in glutamate receptor-mediated calcium mobilization during neuronal migration [145]. Additionally, the transcription factor PAX6, which is important for neuronal migration and axon guidance, has been implicated in polymicrogyria. Contrary to these few hints implicating neuronal migration in polymicrogyria, the disorder is generally considered a "post-migrational" malformation, with issues occurring after neurons have completed their migratory pathway to form the cortical layers [143,146]. Polymicrogyria seems to be caused by numerous, seemingly unrelated etiologies that all result in an excessive gyration phenotype. This is supported by the identification of many genetic roots of the disorder, including signaling molecules, cytoskeletal elements, and others.
Contrary to lissencephaly, where cortical migration provides a clear culprit for the malformation, the complex etiology of polymicrogyria makes it difficult to predict how TUBA1A mutations contribute to the molecular basis of the disease. Polymicrogyria-causing TUBA1A mutations are significantly less common than lissencephaly-causing mutations (~13%), suggesting that perhaps only very specific disruptions of tubulin function can lead to polymicrogyria. Understanding how these TUBA1A mutations alter intrinsic properties of microtubules, interactions with MAPs, and consequently cellular functions will provide a window into the cellular progression of the disease, and will help shed light on the "post-migration" vs. "migration" debate. The TUBA1A mutations that lead to polymicrogyria are described in Table 4.

TUBA1A Mutations Linked to Microcephaly
Microcephaly describes a brain that is significantly smaller than average, typically measured by occipitofrontal circumference (OFC). Different definitions exist as to where the cutoff from small head to microcephaly occurs, with some defining it more stringently as less than 4 SD (i.e., below the 1st centile) below the average [148], while others include a broader range with less than 2 SD (i.e., below the 3rd centile) [149]. However, arguments have been made about the relevance of using such a broad label, as many infants within the −2 SD to −3 SD population will be "normal" [150]. Microcephaly can be categorized into two main divisions: primary microcephaly and secondary microcephaly. Primary microcephaly describes a non-progressive, significantly small head detected prior to 36 gestational weeks (GW), and generally results from reductions in neurogenesis or loss of neural stem cells [150]. The major causes of primary microcephaly include non-genetic, damaging events prior to birth or mutations to genes regulating mitosis in neuronal progenitors [148,150]. In contrast, secondary microcephaly describes when microcephaly progresses postnatally, and is considered a neurodegenerative condition [148]. Causes of secondary microcephaly are numerous and varied, encompassing anything that disrupts orderly development and function of the brain.
As TUBA1A is not expressed in neuronal progenitors, microcephaly associated with TUBA1A likely fits into the secondary microcephaly classification. Neuronal maturation relies on appropriate regulation of microtubules (Figure 2), so it follows that disrupting microtubule function could alter neuronal development and cause progressive microcephaly. This idea is supported by TUBA1A patient cases where head circumference measurements have been taken more than once, such as the patient with p.E27Q whose OFC decreased from within the normal range at birth (−1 SD) to microcephalic (−3.3 SD) by two months of age [134]. Table 5 provides TUBA1A mutations linked to microcephaly. In this review, we use the broadest definition of microcephaly and include cases where the OFC is 2 SD below the appropriate mean (i.e., less than the 3rd percentile). Using this classification,~74% of TUBA1A mutations lead to "microcephaly", or smaller than normal head size. It is important to note that microcephaly is never the primary cortical malformation associated with TUBA1A mutants, but is a common accompaniment to both lissencephaly and polymicrogyria cortical phenotypes.    Bold indicates first report of mutation; * Asterisk indicates mutation was identified through whole exome sequencing; Abbreviations: OFC, occipitofrontal circumference; M, male; F, female; GW, gestational weeks; SD, standard deviation; SBH, subcortical band heterotopia.

TUBA1A Mutations Linked to Cerebellar Dysplasia
In addition to the cortical malformations described above, TUBA1A mutations also commonly cause dysplasia of a variety of other brain regions, most notably the cerebellum. In fact, "lissencephaly with cerebellar hypoplasia" is a common class of TUBA1A mutation-induced phenotype [12,16]. In addition, some identified cases hint that tubulin mutations may cause specific cerebellar phenotypes with only subtly disrupted, or normal, cortical folds [137]. Further sequencing of TUBA1A in patients without the characteristic cortical lissencephaly phenotype may prove fruitful in uncovering additional TUBA1A mutations.

Summary of TUBA1A Mutations Associated with Brain Malformations
In this section, we have merely provided broad categories of brain malformations linked to mutations in TUBA1A. Within each category are numerous mutations that lead to the identified cortical malformation, but also other detrimental neurological and physical phenotypes. The increasing number of TUBA1A mutations discovered and the growing list of phenotypic consequences call for a greater understanding of the mechanism(s) of tubulinopathy disease progression.

Cellular Impact of TUBA1A Mutations
While numerous studies have identified missense mutations in TUBA1A as the genetic basis of brain malformations, few studies have been conducted to determine how these mutations alter tubulin molecularly, or how the mutations alter neuronal maturation/function. Below we describe studies investigating the molecular and cellular impact of TUBA1A mutations.
Studies investigating the stability of the mutant tubulin heterodimer and its ability to incorporate into microtubules provide hints at the underlying mechanism of the mutant and point to whether they lead to haploinsufficiency ( Figure 3A,B), or dominant disruption of microtubule function from within a polymerized microtubule ( Figure 3C), or potentially both. To date, Tian et al. published the most comprehensive and informative study on the consequences of TUBA1A mutations on heterodimer formation and stability [151]. TUBA1A mutant proteins were expressed in vitro to perform transcription/translation analysis in rabbit reticulocyte to test tubulin heterodimer yield. They discovered that p.L397P, p.V303G, and p.R402C lead to significantly reduced amount of unstable tubulin heterodimer, and that p.I188L, p.I238V, p.P263T, p.L286F, p.R402H, and p.S419L caused slight reduction to the amount of tubulin heterodimer [151]. These data suggest that mutations such as p.P263T may act more dominantly, while p.L397P, p.V303G, and p.R402C may act more as haploinsufficient mutants. This hypothesis is supported by the discovery that when ectopically expressed, p.P263T can incorporate into microtubules, dampen microtubule dynamics in COS-7 cells, and decrease microtubule growth in neurites of E15.5 primary cortical neurons 1 DIV [16,151]. These data all point to p.P263T acting dominantly, as described by Figure 3C. Alternatively, when p.V303G was ectopically expressed, it caused no change in microtubule dynamics in COS-7 or primary cortical neurons [151], as might be expected from a mutant that leads to few competent tubulin heterodimers entering the microtubule ( Figure 3A).
Few TUBA1A mutations that lead to lissencephaly have been satisfactorily linked to the migratory pathway. However, inferences can be made about the functions of TUBA1A that may be important for migration based on how MAPs interact with microtubules and on the limited molecular data available on some of the lissencephaly patient TUBA1A mutations. For instance, p.R264C mutation occurs frequently in the TUBA1A patient population (12.5%), causes the less severe pachygyria phenotype, and is one of the most investigated mutations. Molecular and cellular data interrogating the consequence of p.R264C reveal that the mutation reduces the frequency of heterodimer formation due to compromised folding efficiency [152]. This reduced folding efficiency does not completely cause tubulin heterodimers containing the p.R264C mutant to drop out of the polymerization-competent tubulin pool, as they have been seen to form microtubules in HeLa, COS-7, and P19 cells [16,151]. p.R264C has also been modeled using iPSCs derived from patient cells, revealing that neurospheres generated from the patient iPSCs are capable of normal neurite extension [138]. Taken together, these data reveal that the p.R264C TUBA1A mutation probably follows the mutant scenario outlined by Figure 3A, and more importantly that partial haploinsufficiency to TUBA1A is sufficient to disrupt cortical migration enough to cause pachygyria. Similarly, cellular studies with p.I188L, which causes laminar heterotopia, show that FLAG-tagged mutant Tuba1a incorporates into microtubules in P19 cells [16], but that the mutation also causes slightly reduced amounts of tubulin heterodimer formation [151].
Three other lissencephaly-causing mutations have been investigated for their effects on cellular function, p.N329S, p.C25F, and p.R64W. Mutation p.N329S causes lissencephaly with cerebellar hypoplasia [16,138], and when neurospheres were generated from patient iPSCs, the mutation caused abnormal neurite extension [138]. While these cellular data suggest that defective neurite extension leads to lissencephaly with cerebellar hypoplasia, how the microtubules are altered by the p.N329S mutation to cause this diminished neurite growth is unknown. Mutation p.C25F leads to lissencephaly/thin cortex, causes decreased microtubule density in mutant transfected COS7 cells, and also caused patient fibroblast microtubules to depolymerize faster in cold [133]. These results are hard to interpret without further molecular evidence, but suggest that p.C25F may cause a shift in the stability in microtubules and that this decrease in stable microtubules is also sufficient to derail cortical migration enough to cause lissencephaly. Mutation p.R64W leads to severe brain malformations, exhibiting extremely thin cerebral parenchyma, agenesis of the cerebellum, and additional devastating defects to other brain regions. Similar to p.C25F, this mutation also causes decreased microtubule density and unstable microtubules [133]. However, because the phenotypes of p.C25F and p.R64W are divergent, with R64W exhibiting much more severe brain malformations, additional data must be collected to distinguish the molecular consequences of these mutations.
Whereas several lissencephaly-causing mutations lead to a loss of microtubules in vivo, there are no examples of mutants acting dominantly to assemble into microtubules and "poison" their function. These gain-of-function effects could be exerted by mutations that alter amino acid residues on the microtubule surface, such as mutations of R402, which cause lissencephaly. No studies have yet investigated how the patient mutations p.R402C and p.R402H impact microtubule function in vivo; therefore, these are high priority for future investigation.
While a few mutations leading to lissencephaly have been examined in terms of molecular and cellular consequence, to our knowledge none of the polymicrogyria-causing TUBA1A mutants have been investigated. Understanding how these mutations alter microtubule function will be an important step in understanding the molecular basis of the polymicrogyria malformation.
In addition to the cellular data referenced above, two mouse models that contain Tuba1a point mutations have been developed. Mice with N-ethyl N-nitrosourea (ENU) induced mutations in Tuba1a have provide insights into how Tuba1a contributes to neuronal development. For example, an S140G substitution in Tuba1a interferes with efficient GTP binding and heterodimer formation, but heterodimers that include S140G mutant Tuba1a can incorporate into the lattice [11]. One copy of the S140G Tuba1a allele is sufficient to cause defective neuronal migration in the hippocampus and the auditory, visual, and somatosensory cortices [11]. This mutation increases neuronal branching and alters the direction of migration in the rostral migratory stream that populates the olfactory bulb [153]. The neuronal migration phenotypes caused by the S140G mutation can be rescued with increased expression of wild-type Tuba1a, indicating that this mutation reduces the function of the protein [11]. Intriguingly, in vitro data suggest that the S140G mutation cannot be rescued by expressing another α-tubulin isotype, Tuba8 [153]. Additionally, presence of the Tuba8 isotype significantly altered microtubule polymerization compared to control or S140G mutant cells. This study provides the first evidence of functionally distinct α-tubulin isotypes in neurons [153]. However, it is important to note that Tuba8 expression has not been detected in neurons [154].
An N102D substitution in Tuba1a provides different insight into the role of Tuba1a in neural development. Homozygous Tuba1a N102D mice have significantly less acetylated tubulin in growth cones, suggesting that microtubules containing Tuba1a N102D are less stable in extending growth cones [155]. On a cellular level, homozygous Tuba1a N102D mice have defects in motor axon extension and synapse function [155]. On a tissue level, homozygous Tuba1a N102D mice have gross defects in brain development including disorganized cortical layers [155]. The asparagine residue at position 102 is highly conserved across eukaryotes, which allows for the impact of N102D to be isolated and tested in the simple model organism budding yeast. The N to D substitution in yeast α-tubulin interferes with heterodimer stability, impairs its ability to incorporate into the lattice, and increases the frequency of microtubule catastrophes/depolymerization events [155]. Interestingly, yeast cells expressing the mutant allele in one α-tubulin isotype appear to compensate by increasing the cellular levels of an alternative, wild-type α-tubulin isotype. This suggests a novel consequence of TUBA1A mutations-cells may respond by shifting the blend of α-tubulin isotypes, which could in turn alter the normal regulation of microtubule networks.
Together, these mouse models demonstrate that Tuba1a is important for axon extension, efficient synapse function, and migration of neurons during development. However, we do not understand how mutations that have been identified in tubulinopathy patients impact microtubule properties or cellular function to disrupt development. Genetic models of patient derived Tuba1a mutations would lend much needed insights into the attributes of Tuba1a that are important for its role in neurodevelopment.

TUBA1A Expression: A Burst of Tubulin to Fuel Morphogenesis?
To understand the critical role of TUBA1A in brain development, it is important to consider which cells express TUBA1A, and how expression is regulated.

TUBA1A Expression Pattern
TUBA1A was originally identified in a cDNA library from human fetal brain tissue [64]. Subsequent northern blot studies in cell lines showed that TUBA1A is expressed in neural-derived cells and is particularly abundant in cell lines with characteristic long cytoplasmic processes (such as adherent neuroblastomas), but not in cell lines that exhibit small, round morphologies [9]. Since then, numerous groups have tracked the expression of TUBA1A and its vertebrate homologues to neuronal populations after terminal mitosis, as they begin to extend long processes.
A series of classic studies elucidated the expression program of mouse Tuba1a mRNA across tissues and developmental stages. The earliest study investigating mouse Tuba1a expression via northern blot with probes recognizing the distinct Tuba1a 3 -UTR provides the most extensive inquiry into which organs express Tuba1a postnatally [10]. This study identified Tuba1a mRNA in the adult brain, lung, and at low levels in the testes; however, only alternative α-tubulin mRNAs were detected in the heart, kidney, liver, muscles, spleen, stomach, and thymus. A time course analysis of brain-derived samples revealed that Tuba1a is abundant through early postnatal stages, begins to decline by P10, but still is expressed beyond P32 through adulthood [10]. Subsequent studies sought to elucidate the program of Tuba1a expression in the brain; to our knowledge there are no other published studies confirming the presence of Tuba1a in the adult lung or testes.
Much of our understanding of Tuba1a expression in the developing brain comes from studies conducted by Miller and colleagues, who inserted lacZ behind the native Tuba1a promoter and used this Tuba1a:nlacZ transgenic mouse to monitor Tuba1a expression in vivo [156][157][158]. The expression profiles of mouse Tuba1a generated by these approaches, and later a transgenic reporter line using Tuba1a promoter-driven EYFP expression [159], are outlined in Tables 6 and 7. These studies show that Tuba1a is expressed highly throughout the nervous system over the course of development, starting around E9.5 and continuing through early postnatal stages [10,156,158,159]. This trend holds true through the developing central and peripheral nervous system, with Tuba1a induction onset correlating to neurogenesis of numerous neuronal structures including the forebrain, midbrain, hindbrain, spinal cord, retina, and cranial nerves, among others [158]. Primary cultures generated from the transgenic reporter mice revealed that Tuba1a is expressed in cells positive for neuronal markers neuron-specific enolase (NSE) and βIII tubulin, but not in cells positive for stem cell or glial markers [156,[158][159][160]. More recent RNA-seq studies from the Barres lab show that Tuba1a remains the most abundant α-tubulin isotype mRNA in neurons late in brain development (Postnatal Day 7) [154] (Figure 4). Based on these studies, the current model is that Tuba1a expression is induced coincident with or shortly after the terminal cell division that gives rise to neurons, and provides >95% of the α-tubulin mRNA in these cells [72].  [154] (Figure 4). Based on these studies, the current model is that Tuba1a expression is induced coincident with or shortly after the terminal cell division that gives rise to neurons, and provides >95% of the α-tubulin mRNA in these cells [72].   Distribution of α-tubulin isotype mRNA expression in various cell types generated from P7-17 mouse cerebral cortex. OPC population contains 5% microglial contamination. Adapted from [141].
While much of the RNA probing data suggest that Tuba1a expression is limited primarily to neurons, the recent RNA-seq dataset contends that Tuba1a may be expressed in more cell types than the earlier studies indicate [154]. As stated previously, numerous studies claim that Tuba1a mRNA is absent in stem cell, glial, and neuron progenitor populations, limited only to early born neurons [156,[158][159][160]. However, the RNA-Seq data suggest that Tuba1a may in fact be expressed in numerous cell types found in the brain, with Tuba1a expression equally high in newly-formed oligodendrocytes as it is in neurons at Postnatal Day 7 ( Figure 4). As immunohistochemistry double labeling studies suggest that oligodendrocytes in embryonic and adult mouse do not express Tuba1a [156,159,160], the contradictory expression of Tuba1a in glial populations calls for additional studies to conclusively uncover what cell types contain TUBA1A. After brain development, Tuba1a mRNA levels persist, albeit at lower levels, in the adult mouse brain. Which adult cell populations express Tuba1a is somewhat controversial, due to conflicting data. Bamji and Miller found that Tuba1a mRNA was panneuronal in the adult mouse brain, with the highest levels found in neuronal populations that have the potential for morphologic growth, such as neurons of the piriform cortex [157] (Table 7). In contrast, Coksaygan and colleagues only detected Tuba1a mRNA in some regions of the adult brain (Table 7). How experimental differences might contribute to these discrepancies is not immediately clear.
While much of the Tuba1a expression data have been conducted in mouse, studies in rat models yield similar findings. Rat Tuba1a mRNA is highly enriched in the embryonic nervous system and is less abundant in adult brain [72]. Additionally, the rat data also support the idea that Tuba1a is induced in neurons undergoing neurite extension, such as in vivo at the cortical plate as well as in vitro in PC12 cells induced to differentiate with NGF treatment [72].
Interestingly, studies of the zebrafish tuba1a yield results that overlap with those from mammalian models, but with several important differences. Whereas zebrafish tuba1a is dramatically induced in the developing nervous system and declines coincident with the maturation, it is maintained at high levels in progenitor cells in the retinal periphery, lining the brain ventricles, and around the central canal of the spinal cord [161]. When cultured in vitro, these tuba1a-expressing cells divide and give rise to new neurons, suggesting that zebrafish may express tuba1a in neuronal progenitors as well as newly-born neurons [161]. Consistent with this notion, morpholino-based knock down of zebrafish tuba1a suppresses CNS development and leads to fewer differentiating neurons [162]. Intriguingly, Ramachandran et al. generated transgenic fish to conditionally and permanently label tuba1a-expressing cells, and found transient tuba1a expression in neural progenitors, skeletal muscle, heart, and intestine progenitors [163]. This experiment suggests that tuba1a expression is not neuronally limited in zebrafish.
These expression studies strongly suggest that TUBA1A is highly induced in neurons as they undergo cellular morphology changes to generate cytoplasmic processes. This suggests a simple model for why neurons, but not other cell types, use the conserved expression program for TUBA1A-neurons need significantly more α-tubulin at the time when they are maturing to form the long processes necessary to wire the nervous system.

Mechanisms Regulating TUBA1A Expression
While questions remain regarding which cell populations express TUBA1A, it is clear that expression is carefully regulated during development and adulthood. What pathways control this expression program for TUBA1A? A variety of factors and conditions that mediate TUBA1A mRNA production have been identified, and growth factors involved in neuronal differentiation play a prominent role.
Several studies have shown that growth factors are sufficient to stimulate Tuba1a expression. For example, PC12 cells from rats can be induced to differentiate into neurons with the addition of Nerve growth factor (NGF), and consequently Tuba1a mRNA levels increase as the differentiated PC12 cells begin to actively extend neurites [72]. Similarly, NGF treatment of neonatal rats causes a 5-10-fold increase of Tuba1a mRNA during a developmental period when Tuba1a levels normally decrease [164]. In addition to NGF, Fibroblast Growth Factor (FGF) and Platelet-Derived Growth Factor (PDGF) also stimulate Tuba1a expression. FGF was first shown to stimulate Tuba1a expression in rat sensory neurons, where treatment with FGF-1 increases Tuba1a mRNA levels by nearly ten-fold, and causes a concomitant increase in neurite outgrowth [165]. These findings suggest that exogenous stimuli such as growth factors may be sufficient to induce TUBA1A expression, at least in certain cell types.
The most detailed evidence of the regulatory pathways that act downstream of growth factors to increase Tuba1a transcription comes from a study using mouse primary cultures of E12-E13 cortical progenitors [166]. This study revealed a phosphorylation cascade initialized by FGF and PDGF activation that ends in drastic upregulation of TUBA1A, and ultimately neurogenesis ( Figure 5). Inhibiting factors in this pathway, MEK or C/EBP, leads to a significant decrease in Tuba1a expression and stops neurogenesis. Conversely, introducing an activated C/EBP phosphomimetic enhances Tuba1a expression and promotes neurogenesis. Furthermore, this study identified three C/EBP binding sites in the conserved TUBA1A promoter ( Figure 5B), and revealed that direct C/EBP binding to these promoter regions greatly enhances TUBA1A transcription. Interestingly, inhibiting the Tuba1a neurogenesis pathway causes the would-be neurons to take on an astrocyte fate. This reveals that the same progenitors can be induced to become either neurons or astrocytes, and that Tuba1a activation is an important step in promoting the neuronal fate.
While the studies described above implicate common neurogenesis growth factors in Tuba1a expression, the specific requirements for Tuba1a activation remain largely unexplored. However, these data provide an additional link between the presence of Tuba1a and neurogenesis, further promoting the idea that Tuba1a may be an important regulator of neuronal fate.

TUBA1A in Regeneration
The unique temporal expression of TUBA1A during brain development indicates a critical, albeit poorly defined role for TUBA1A as a growth-associated neuronal α-tubulin. Interestingly, there is evidence that TUBA1A may also play important roles in remodeling the cytoskeleton during nerve regeneration and repair. Although Tuba1a mRNA expression declines after development, it is rapidly induced following nerve injury [72,[167][168][169][170][171]. Mathew and Miller proposed that NGF could regulate Tuba1a expression following injury like in development [164,172]. The p75 NGF receptor is upregulated following axotomy, providing some support for this model [164,172]. The induction of a developmental α-tubulin, such as TUBA1A, during regeneration supports the idea that developing and regenerating neurons may utilize common mechanisms and that TUBA1A is of particular importance to growing neurons, regardless of age.
How might nerve injury trigger the re-emergence of TUBA1A expression? Several studies suggest that Tuba1a is normally repressed in adult neurons, but if normal processes are disrupted, Tuba1a expression is induced. For example, blocking fast axonal transport has been shown to induce Tuba1a expression [173,174]. These findings specifically implicate normal axonal transport as part of the mechanism that represses TUBA1A. For example, the neuron could be constantly monitoring its axonal integrity using intrinsic cues, such as backup of protein cargoes that would occur within the

TUBA1A in Regeneration
The unique temporal expression of TUBA1A during brain development indicates a critical, albeit poorly defined role for TUBA1A as a growth-associated neuronal α-tubulin. Interestingly, there is evidence that TUBA1A may also play important roles in remodeling the cytoskeleton during nerve regeneration and repair. Although Tuba1a mRNA expression declines after development, it is rapidly induced following nerve injury [72,[167][168][169][170][171]. Mathew and Miller proposed that NGF could regulate Tuba1a expression following injury like in development [164,172]. The p75 NGF receptor is upregulated following axotomy, providing some support for this model [164,172]. The induction of a developmental α-tubulin, such as TUBA1A, during regeneration supports the idea that developing and regenerating neurons may utilize common mechanisms and that TUBA1A is of particular importance to growing neurons, regardless of age.
How might nerve injury trigger the re-emergence of TUBA1A expression? Several studies suggest that Tuba1a is normally repressed in adult neurons, but if normal processes are disrupted, Tuba1a expression is induced. For example, blocking fast axonal transport has been shown to induce Tuba1a expression [173,174]. These findings specifically implicate normal axonal transport as part of the mechanism that represses TUBA1A. For example, the neuron could be constantly monitoring its axonal integrity using intrinsic cues, such as backup of protein cargoes that would occur within the axon upon injury, or extrinsic cues such as constant uptake of growth factors like CNTF from nearby Schwann cells [167,173]. Alternatively, it has been proposed that repression could be maintained through physical contacts between neurons and neighboring cells [174]. However, current evidence does not support this model, as studies have found that loss of contact with either the basal lamina or the neuron's target structure alone is insufficient to induce Tuba1a expression [167,172]. Although the exact mechanisms by which its expression is induced after nerve injury remain unclear, the model that TUBA1A expression may be controlled through mechanisms that monitor normal, homeostatic processes in neurons is intuitively appealing. This could explain how TUBA1A expression is repressed late in development, when these processes become active, and de-repressed when these processes are disrupted.
The expression of TUBA1A during regeneration after nerve injury provides an interesting opportunity to study regenerative capacity in neurons. While peripheral nerves are able to regenerate after injury, most neurons of the central nervous system do not have this capacity [175]. What allows peripheral axons to regenerate, while central axons cannot? Many of the initial studies identifying Tuba1a expression in response to injury focused on the peripheral nervous system; however, increases in Tuba1a expression have been noted following injury to central nerves as well [170]. This observation provides an opportunity to probe the differences in regenerative capacity between the central and peripheral nervous systems. Interestingly, one study found that Tuba1a gene expression after injury was broadly similar between peripheral nerves that regenerate and proximal central nerves that do not [170]. Importantly, in peripheral neurons, Tuba1a expression remained elevated until regeneration was complete. In central rubrospinal neurons, which did not regenerate, Tuba1a expression remained elevated for as long as two weeks post injury, even after levels of other injury-related cytoskeletal proteins had returned to normal [170]. However, if rubrospinal neurons are transplanted into a peripheral environment, they can regenerate [176]. Other studies have demonstrated that Tuba1a expression correlates with the timing of regeneration, and that in general Tuba1a will remain elevated until the axon is able to regenerate [172,177]. Thus, these data show that specific central neurons have regenerative capacity, but likely an unfavorable environment prevents complete regeneration. Conversely, study of a different population of central neurons found no evidence of Tuba1a upregulation after injury [169]. This may indicate some heterogeneity in regenerative capacity between different populations of central neurons; some neurons may lack the capacity to regenerate, regardless of whether the environment is conducive to regeneration [169]. Interestingly, Tuba1a can become upregulated after injury, even in uninjured regions of the nervous system. For example, elevated Tuba1a expression was not limited solely to injured peripheral axons, as neighboring uninjured axons can also display increased Tuba1a mRNA [172,177]. Additionally, elevated Tuba1a expression was found in the red nucleus, a brainstem nucleus that sends axons to the spinal cord, following a spinal cord compression injury [171]. The outlined studies show a selective upregulation of Tuba1a in regenerating neurons. The properties of TUBA1A α-tubulin that make it particularly suited to these functions are yet unknown. However, it is interesting to speculate that combinations of different tubulin isotypes could produce microtubules that are preferentially suited for specific functions.

Concluding Remarks
In recent years, the α-tubulin gene TUBA1A has emerged as a vital component in the development of the nervous system. With the identification of tubulinopathy-causing TUBA1A mutations, understanding how TUBA1A must function during neural development becomes an urgent inquiry. Microtubules perform many complex roles throughout development, but it remains unclear how TUBA1A, and the tubulin heterodimers and microtubules it produces, can be regulated to perform such diverse tasks in different neuronal compartments and through developmental and adult stages. The neuronal tubulin gene profile, as well as the poorly understood, yet remarkably complex, regulation of tubulin PTMs and MAPs may provide a starting point for determining how microtubules are tuned throughout neuron development. Brain malformation-causing TUBA1A mutations provide an irreplaceable opportunity to explore the role of microtubules at the cellular level of neuronal maturation, and at the tissue level of brain formation. While previous studies have performed the important task of sequencing tubulin genes to uncover underlying genetic mutations, future tubulinopathy studies must focus on the molecular and cellular consequences of these mutations to fully understand how altered TUBA1A function contributes to brain malformation progression. Discovering the molecular mechanisms behind TUBA1A mutations will be an important undertaking in cellular neuroscience, and could bring powerful new insight into the essential roles of microtubules in neural development.