Elongator is a histone acetyl-transferase complex, first isolated from yeast and consisting of six subunits named ELP1–ELP6 that are highly conserved in eukaryotic organisms including yeast, humans and Arabidopsis. Experiments with yeast indicate that the Elongator complex is a functional entity (1). Similarly, knockouts of the Arabidopsis homologs of ELP1, ELP3 and ELP4 subunits show identical phenotypes, suggesting that each is required for the complex's function in plants (2). A human homolog of ELP1, IκB kinase complex-associated protein (IKAP), is the protein responsible for familial dysautonomia (FD). ELP3 is a conserved acetylase and is crucial for the function of the complex. Although the well-established target of this acetylase activity is nuclear histones, localization data suggest that there is likely to be another, cytoplasmic, target (3–5).
Familial dysautonomia (FD), a disease of the autonomic and sensory nervous systems, involves mutations in the protein IκB kinase complex-associated protein, which is a component of the human Elongator acetylase complex. We suggest a hypothesis in which defects in tubulin acetylation and impairment of microtubule-based protein trafficking may be an underlying cause of FD. In addition, an Arabidopsis homolog of the Elongator subunit ELP3 has been found to bind to the αβ-tubulin heterodimer, suggesting that α-tubulin may be a cytoplasmic target of Elongator acetylase activity. Studies of synergistic double mutants in yeast indicate a novel role for Elongator in cytoskeletal dynamics, although this is probably because of an effect on actin rather than microtubules. Finally, we suggest that tubulin deacetylase inhibitors may prove useful in the treatment of FD.
Role of Elongator and tubulin acetylation in FD
What could be the cytoplasmic target of the Elongator acetylase? α-Tubulin, a subunit of αβ-tubulin heterodimers that assemble to form the microtubule cytoskeleton, is acetylated on a conserved lysine residue (6,7), although the tubulin acetylase has not been identified. There are notable similarities between enzymes with histone and tubulin deacetylase activities. In fact, a member of the histone deacetylase family, histone deacetylase 6, also acts as a tubulin deacetylase (8). Similarly , SIRT(sirtuin)2, a mammalian ortholog of the histone deacetylase Sir2p (silent mating type information regulation 2), deacetylates α-tubulin and colocalizes with microtubules (9). In addition, the histone deacetylase inhibitor trichostatin also promotes tubulin acetylation (8,10). If the activities of histone and tubulin deacetylases are so closely related, it is reasonable to postulate that histone and tubulin acetylase activities may also be related. Thus, the Elongator complex, which is known to possess histone acetylase activity in vitro and in vivo (3,11), could also acetylate tubulin in vivo, especially because acetylases are generally known to act on diverse substrates (12,13).
Some evidence from IKAP-depleted cells suggests a role for Elongator in cytoskeletal dynamics. The cytoskeleton is known to be important for cell motility, and IKAP-depleted HeLa cells showed reduced motility at a wound edge (14). In addition, certain cytoskeletal genes show altered expression in IKAP-depleted cells (14). It is possible that this is a result of transcriptional regulation. For example, the actin filament-severing protein gelsolin was downregulated probably because of a defect in transcriptional elongation (14).
A defect in tubulin acetylation and hence impairment in microtubule-based transport could be a contributing factor to FD, considering that there is a well-established link between protein trafficking along microtubules or actin filaments and neurodegeneration. Trafficking defects have been implicated in numerous neurodegenerative diseases, including Alzheimer's disease, lissencephaly, amyotrophic lateral sclerosis and others (15,16 and references therein). Neurons are likely to be highly sensitive to defects in motor proteins or their tracks (microtubules and actin filaments) because of their long extensions and rapid turnover of trafficking components such as synaptic vesicles.
Indeed, a recent study has found that α-tubulin acetylation influences the binding and motility of the microtubule motor kinesin-1 in vitro (17). Treatment of neurons with inhibitors of tubulin deacetylase increased the transport of the kinesin-1 cargo protein JIP1 [c-Jun N-terminal kinase (JNK)-interacting protein 1] to neurite tips, thus showing a key role for tubulin acetylation in microtubule-based transport.
Proposed mechanism of interactions between Elongator complex, JIP/JNK module and microtubules
It appears that three players, namely the Elongator complex, the JIP/JNK module of proteins and microtubule-regulating proteins, may interact mutually in a mechanism that promotes tubulin acetylation, microtubule stabilization and microtubule-based transport (Figure 1).
- 1The ELP1 subunit of the Elongator complex activates JNK (5), which is probably bound to the scaffold protein JIP, in concert with mitogen-activated protein kinase kinase and mitogen-activated protein kinase kinase kinase. ELP1 itself may in turn be phosphorylated, and it could then activate ELP3, which might subsequently acetylate a microtubule. In this way, the Elongator-dependent acetylation of microtubules would promote kinesin-based transport of the JIP/JNK module, together with its resident JNK-activating kinases, toward the microtubule plus-end.
- 2The scaffold protein JIP1 is important in the activation of JNK by upstream kinases (18). JIP1 can either enhance or inhibit JNK signaling depending on its concentration (19). While mice deficient in JIP1 do not show brain abnormalities (20), those deficient in another JNK-scaffolding protein, JIP3, are defective in axon guidance (21), although JIP1 can partially compensate for the loss of JIP3 (22). Caenorhabditis elegans JIP3 homolog UNC-16 binds to, and is transported by, kinesin (23), suggesting that JIP3 and JIP1 may be transported along microtubules in a similar manner. Therefore, tubulin acetylation is likely to affect the transport and localization of all JIP proteins in neurons and may be crucial for correct brain development.
- 3Activated JNK in the JIP/JNK module also phosphorylates other growth-cone resident proteins, including the microtubule-depolymerizing protein SCG10 and the microtubule-stabilizing protein doublecortin, mutations which are also responsible for lissencephaly (24,25). The phosphorylation causes the SCG10 to release two sequestered tubulin heterodimers, which polymerize at the plus-end of the microtubule. Phosphorylation of SCG10 on S62 and S73 by activated JNK regulates neurite elongation during brain development (26), and this phosphorylation is sufficient to inactivate the protein's microtubule-depolymerizing activity (27).
It is possible that the interaction between ELP1 and JNK may regulate correct tubulin acetylation because the region of ELP1 that is affected in FD contains a site of phosphorylation by JNK (5), and evidence from yeast suggests that phosphorylation of ELP1 is crucial to Elongator function (28). In the neurodegenerative disease FD, reduced activation of JNK might lead to enhanced microtubule turnover and hence microtubule destabilization and shorter neurites (28; Figure 1). It is important to note here, however, that neurite elongation requires a balance between microtubule stabilization and destabilization (29). Too much of either leads to shorter neurites. Indeed, gene silencing of SCG10 leads to a suppression of neurite extension (30).
Interestingly, a recent study shows that SIRT2 decelerates oligodendrocyte differentiation through tubulin deacetylation (31). Our model, or a similar model, could equally apply to the differentiation of such glial cells, as similar processes are probably at work during their differentiation (32). In this case, a defect in tubulin acetylation might lead to abnormality of oligodendroglia or other glial cells, thus causing neuronal degeneration.
Retrograde transport of neurotrophic factors along microtubules and FD
One of the most striking symptoms seen in people suffering FD is the absence of papillae on the tongue (33). A substantial body of work has been devoted to the study of the development of taste buds and papillae on the palate and tongue (34 and references therein). Correct development of taste buds and papillae requires innervation, and survival of the necessary gustatory neurons requires both neurotrophins secreted by epithelial cells and neurotrophin receptors on the neurons. The secreted neurotrophins and their receptors are endocytosed by the neurons and trafficked to the cell body (35) along microtubules (36,37). Thus, it appears that there may be a problem with retrograde signaling to the cell body in FD. Indeed, huntingtin, the protein involved in the neurodegenerative disorder Huntington's disease, controls survival of neurons by enhancing transport of brain-derived neurotrophic factor along microtubules (37).
Retrograde transport along microtubules requires a different motor protein to anterograde transport – dynein as opposed to kinesin. Activated and internalized neurotrophic receptors as well as neurotrophic factors themselves are transported to the nucleus by dynein (38). Indeed, this helps explain the neurodegeneration seen in individuals with impaired dynein function (39). It appears that tubulin acetylation is likely to be important for the association of dynein with microtubules, similar to kinesin, because binding of dynein to microtubules is enhanced if the microtubules are acetylated (40). We suggest that the ELP3 acetylase of the Elongator complex may acetylate microtubules. The acetylation may be deficient in sensory neurons of those with FD, thus inhibiting the retrograde transport of neurotrophic factors required for sensory neuron survival. This in turn may lead to the absence of papillae on the tongue and other symptoms associated with FD.
Elongator and the cytoskeleton in yeast
A large-scale survey of synthetic lethal mutations in yeast identified genetic interactions of ELP3 with proteins involved in cell polarity, including Myo2p, Bim1p, Bni1p, Cdc42p and Bbc1p (41,42). Knockouts of these genes are synergistic when combined with a knockout of ELP3, demonstrating that ELP3 plays a novel role in the maintenance of cell polarity, in addition to its documented role in histone acetylation and transcriptional elongation. Similar genetic interactions were found for other components of the Elongator complex. There is also a genetic interaction of ELP6 with the tubulin chaperone proteins Yke2p, Pac10p, Gim3p, Gim4p and Gim5p, again linking Elongator function to microtubules and cell polarity.
The Saccharomyces cerevisiae Elongator complex consisting of the six ELP1–ELP6 proteins has been proposed to participate in three distinct cellular processes: transcriptional elongation (43), polarized exocytosis (44) and formation of modified wobble uridines in transfer RNA (tRNA) (45). Studies indicate that the physiologically relevant function of the evolutionarily conserved Elongator complex is in formation of modified nucleosides in tRNAs (46). One study raised the possibility that regulation of polarized exocytosis by ELP1 is an evolutionarily conserved function of the entire Elongator complex and that FD results from a dysregulation of neuronal exocytosis (44).
The subunits of microtubules, α- and β-tubulin, are highly conserved proteins in all eukaryotes. Microtubules in different eukaryotic lineages also play a somewhat conserved role in maintaining cellular polarity. Some of the proteins that interact with microtubules and help regulate their dynamics are also homologous. For example, the family of proteins represented by Bim1p in yeast is also present in animals and plants, as are the microtubule motor proteins kinesins. The function of these microtubule-associated proteins can also be conserved in disparate taxa. Interestingly, a study on Arabidopsis identified a plant homolog of ELP3, Elongata3, as a tubulin-binding protein (47). As ELP3 is the subunit with acetyl-transferase activity, this supports our hypothesis.
There is some controversy about whether α-tubulin is acetylated in yeast. One study found that Schizosaccharomyces pombe lacked acetylated tubulin using the specific antibody 6-11B-1 (48). Moreover, the conserved lysine residue at amino acid position 40 is lacking in both S. pombe and S. cerevisiae α-tubulins. However, a more recent study found that acetylated tubulin was present in S. cerevisiae also using the 6-11B-1 antibody (49). This study does not explain how S. cerevisiae α-tubulin can be acetylated even though it lacks the conserved lysine. However, there is a lysine residue at position 42 that could possibly be acetylated. Nonetheless, trafficking in yeast and metazoans is very different. Yeast uses actin filaments and myosin motors for trafficking and not microtubules and their motors, so the effects of Elongator in yeast cells are more likely on actin-based processes rather than microtubule-dependent processes.
Conclusion and possible treatments for FD
We have suggested a hypothesis that disruption of tubulin acetylation followed by impaired trafficking along microtubules may be an underlying cause of FD, possibly also leading to aberrant microtubule dynamics in axons. This opens the possibility that drugs affecting either the microtubular cytoskeleton directly or its acetylation status may prove useful in treating the disease. In particular, drugs that inhibit tubulin deacetylase activity and thus promote tubulin acetylation might be beneficial. The compound tubacin (10) or novel SIRT2 inhibitors (50), for example, would be first-line candidate drugs to treat FD.
One further study is of particular interest in relation to a possible treatment for FD. The study (51) found that the Huntington's disease protein, huntingtin, interacts with IKAP in a yeast two-hybrid assay. Although this result is preliminary, it provides, for the first time, a possible direct link between FD and Huntington's disease and even suggests that treatments effective against Huntington's disease may be applicable, though perhaps in a modified form, to FD. Indeed, tubulin acetylation is dramatically reduced in brains of Huntington's disease sufferers (40). Thus, even if the Elongator complex does not directly acetylate tubulin it may prove that inhibitors of tubulin deacetylases will be useful in treating FD through increasing vesicular flux and the transport of neurotrophic factors along microtubules (40).
J. G., J. M. and R. O. would like to thank the Australian Research Council for financial support and the manuscript reviewers for their helpful comments.