betaTubulin56D (ß1 tubulin) Seven Drosophila dynein heavy chain genes have been characterized. Sequence analysis of partial clones reveals thateach encodes a highly conserved portion of the putative hydrolytic ATP-binding site thatincludes a consensus phosphate-binding (P-loop) motif. One of the clones is derived from Dhc64C, a Drosophilacytoplasmic dynein heavy chain gene that shows extensive amino acid identity to cytoplasmicdynein isoforms found in other organisms. Two other Drosophila dynein clones are 85 and 90% identical at theamino acid level to the corresponding region of the beta heavy chain of sea urchin axonemal dynein. Probesfor all seven of the dynein-related sequences hybridize to transcripts that are the appropriate size(approximately 14 kilobases) to encode the characteristic high molecular weight dynein heavy chainpolypeptides. The Dhc64C transcript is readily detected in RNA from ovaries, embryos, and testes.Transcripts from five of the six remaining genes are also detected in tissues otherthan testes, but in much lesser amounts. All but one of the dynein transcripts are expressed at comparable levels in testes, suggestingtheir participation in flagellar axoneme assembly and motility (Rassmuson, 1994). Dhc-Yh3, another of the Drosophila dynein genes, is located in Y chromosome region h3. This region is contained within kl-5,a locus required for male fertility. The PCR clone derived from Dhc-Yh3 is 85% identical to the correspondingregion of the beta heavy chain of sea urchin flagellar dynein but only 53% identical to a cytoplasmic dyneinheavy chain from Drosophila. In situ hybridization to Drosophila testes shows Dhc-Yh3 is expressed inwild-type males but not in males missing the kl-5 region. These results are consistent with the hypothesisthat the Y chromosome is needed for male fertility because it contains conventional genes that functionduring spermiogenesis (Gepner, 1993). Yet another dynein gene, Dhc64C, encodes a cytoplasmic dynein heavy chain polypeptide. The primary structure of theDrosophila cytoplasmic dynein heavy chain polypeptide has been determined by the isolation and sequenceanalysis of overlapping cDNA clones. Drosophila cytoplasmic dynein is highly similar in sequence andstructure to cytoplasmic dynein isoforms reported for other organisms. The Dhc64C dynein transcript isdifferentially expressed during development; the highest levels are detected in the ovaries of adultfemales. Within the developing egg chambers of the ovary, the dynein gene is predominantly transcribed inthe nurse cell complex. In contrast, the encoded dynein motor protein displays a striking accumulation in thesingle cell that will develop as the oocyte. The temporal and spatial pattern of dynein accumulation in theoocyte is remarkably similar to that of several maternal effect gene products that are essential for oocytedifferentiation and axis specification. This distribution and its disruption by specific maternal effectmutations lends support to recent models suggesting that microtubule motors participate in the transport ofthese morphogens from the nurse cell cytoplasm to the oocyte (Li, 1994). The cytoplasmic dynein light-chain gene, ddlc1, encodes the first identified cytoplasmic dynein light chain. Its geneticanalysis represents the first in vivo characterization of cytoplasmic dynein function in higher eucaryotes.The ddlc1 gene maps to 4E1-2 and encodes an 89-amino-acid polypeptide with a high similarity to theaxonemal 8-kDa outer-arm dynein light chain from Chlamydomonas flagella. The ddlc1 gene is expressedubiquitously. Anti-DDLC1 antibody analyses show that the DDLC1 protein is localized in the cytoplasm.P-element-induced partial-loss-of-function mutations cause pleiotropic morphogenetic defects in bristle andwing development, as well as in oogenesis, resulting in female sterility. The morphologicalabnormalities found in the ovaries are always associated with a loss of cellular shape and structure, asvisualized by a disorganization of the actin cytoskeleton. Total-loss-of-function mutations are lethal. Alarge proportion of mutant animals degenerate during embryogenesis, and the dying cells showmorphological changes characteristic of apoptosis: cell and nuclear condensation andfragmentation, as well as DNA degradation. Cloning of the human homolog of the ddlc1 gene, hdlc1,demonstrates that the dynein light-chain 1 is highly conserved in both flies and humans. Northern blot analysisand epitope tagging show that the hdlc1 gene is ubiquitously expressed and that the human dynein lightchain 1 is localized in the cytoplasm (Dick, 1996).
Seven Drosophila dynein heavy chain genes have been characterized. Sequence analysis of partial clones reveals thateach encodes a highly conserved portion of the putative hydrolytic ATP-binding site thatincludes a consensus phosphate-binding (P-loop) motif. One of the clones is derived from Dhc64C, a Drosophilacytoplasmic dynein heavy chain gene that shows extensive amino acid identity to cytoplasmicdynein isoforms found in other organisms. Two other Drosophila dynein clones are 85 and 90% identical at theamino acid level to the corresponding region of the beta heavy chain of sea urchin axonemal dynein. Probesfor all seven of the dynein-related sequences hybridize to transcripts that are the appropriate size(approximately 14 kilobases) to encode the characteristic high molecular weight dynein heavy chainpolypeptides. The Dhc64C transcript is readily detected in RNA from ovaries, embryos, and testes.Transcripts from five of the six remaining genes are also detected in tissues otherthan testes, but in much lesser amounts. All but one of the dynein transcripts are expressed at comparable levels in testes, suggestingtheir participation in flagellar axoneme assembly and motility (Rassmuson, 1994).
Dhc-Yh3, another of the Drosophila dynein genes, is located in Y chromosome region h3. This region is contained within kl-5,a locus required for male fertility. The PCR clone derived from Dhc-Yh3 is 85% identical to the correspondingregion of the beta heavy chain of sea urchin flagellar dynein but only 53% identical to a cytoplasmic dyneinheavy chain from Drosophila. In situ hybridization to Drosophila testes shows Dhc-Yh3 is expressed inwild-type males but not in males missing the kl-5 region. These results are consistent with the hypothesisthat the Y chromosome is needed for male fertility because it contains conventional genes that functionduring spermiogenesis (Gepner, 1993).
Yet another dynein gene, Dhc64C, encodes a cytoplasmic dynein heavy chain polypeptide. The primary structure of theDrosophila cytoplasmic dynein heavy chain polypeptide has been determined by the isolation and sequenceanalysis of overlapping cDNA clones. Drosophila cytoplasmic dynein is highly similar in sequence andstructure to cytoplasmic dynein isoforms reported for other organisms. The Dhc64C dynein transcript isdifferentially expressed during development; the highest levels are detected in the ovaries of adultfemales. Within the developing egg chambers of the ovary, the dynein gene is predominantly transcribed inthe nurse cell complex. In contrast, the encoded dynein motor protein displays a striking accumulation in thesingle cell that will develop as the oocyte. The temporal and spatial pattern of dynein accumulation in theoocyte is remarkably similar to that of several maternal effect gene products that are essential for oocytedifferentiation and axis specification. This distribution and its disruption by specific maternal effectmutations lends support to recent models suggesting that microtubule motors participate in the transport ofthese morphogens from the nurse cell cytoplasm to the oocyte (Li, 1994).
The cytoplasmic dynein light-chain gene, ddlc1, encodes the first identified cytoplasmic dynein light chain. Its geneticanalysis represents the first in vivo characterization of cytoplasmic dynein function in higher eucaryotes.The ddlc1 gene maps to 4E1-2 and encodes an 89-amino-acid polypeptide with a high similarity to theaxonemal 8-kDa outer-arm dynein light chain from Chlamydomonas flagella. The ddlc1 gene is expressedubiquitously. Anti-DDLC1 antibody analyses show that the DDLC1 protein is localized in the cytoplasm.P-element-induced partial-loss-of-function mutations cause pleiotropic morphogenetic defects in bristle andwing development, as well as in oogenesis, resulting in female sterility. The morphologicalabnormalities found in the ovaries are always associated with a loss of cellular shape and structure, asvisualized by a disorganization of the actin cytoskeleton. Total-loss-of-function mutations are lethal. Alarge proportion of mutant animals degenerate during embryogenesis, and the dying cells showmorphological changes characteristic of apoptosis: cell and nuclear condensation andfragmentation, as well as DNA degradation. Cloning of the human homolog of the ddlc1 gene, hdlc1,demonstrates that the dynein light-chain 1 is highly conserved in both flies and humans. Northern blot analysisand epitope tagging show that the hdlc1 gene is ubiquitously expressed and that the human dynein lightchain 1 is localized in the cytoplasm (Dick, 1996).
Polo is regulated by phosphorylation and has preferred microtubule-associated substrates in Drosophila embryo extracts. Wild type Polo protein migrates as a tight doublet of 67 kDA. By phosphatase treatment, which also inactivates the kinase, this is converted to a single band. Putative Polo substrates include proteins of sizes 220 kDa, 85 kDa and 54 KDa. Monoclonal antibody to ß-tubulin precipitates the phosphorylated 54 kDa protein, together with an associated 85 kDa protein also phosphorylated by Polo. Moreover, Polo binds to an 85 kDa protein that is enriched in microtubule preparations (Tavares, 1996).
Costa (also known as Costal2 or Cos2) iscytoplasmic and binds microtubules in a taxol-dependent, ATP-insensitve manner, while kinesin heavy chain binds microtubules in a toxol-dependent, ATP-insensitive manner. Cubitus interruptus associates with Cos2 in a largeprotein complex, suggesting that Cos2 directly controls the activity of Ci. This association does not involve microtubules. Elevated cytoplasmic Ci staining is seen in cos2 clones in the anterior compartment. The level of Ci staining is independent of the clone's distance from the A/P border. Nuclear Ci is not evident in the clones (Sisson, 1997).
Actin and microtubules (MTs) are tightly coordinated during neuronal growth cone navigation and are dynamically regulated in response to guidance cues; however, little is known about the underlying molecular mechanisms. Drosophila Pod-1 can crosslink both actin and MTs. In cultured S2 cells, Pod-1 colocalizes with lamellar actin and MTs, and overexpression remodels the cytoskeleton to promote dynamic neurite-like actin-dependent projections. Consistent with these observations, Pod-1 localizes to the tips of growing axons, regions where actin and MTs interact, and is especially abundant at navigational choice points. In either the absence or overabundance of Pod-1, growth cone targeting but not outgrowth is disrupted. Taken together, these results reveal novel activities for pod-1 and show that proper levels of Pod-1, an actin/MT crosslinker, must be maintained in the growth cone for correct axon guidance (Rothenberg, 2003).
To test whether Pod-1 can crosslink actin and MTs, soluble Pod-1 was purified from a stable S2 cell line engineered to inducibly express full-length His-tagged Pod-1. When purified Dpod-1HIS (80-100 nM) is added to fluorescently labeled phalloidin-stabilized actin filaments (30-60 nM), it rapidly (within minutes) induces the formation of long (20-50 μm), mostly unbranched actin bundles that frequently have bends or curls. Their appearance does not change over time, suggesting that a steady state is reached. When buffer alone or BSA (even at a 25-fold higher concentration) is added to the actin filaments, no bundling activity is observed, demonstrating that the activity is specific. Similarly, when taxol-stabilized fluorescent MTs are combined with purified Pod-1HIS (80-100 nM Pod-1HIS; 500 nM MTs), rapid crosslinking of MTs is observed, whereas no crosslinking occurswith buffer alone or with BSA -- even at a 25-fold higher concentration. Thus, Pod-1HIS possesses both actin and MT crosslinking activities (Rothenberg, 2003).
To test whether Pod-1 can remodel the cytoskeleton in cells, Pod-1 was studied in S2 cells, a system that can be used to study cytoskeletal dynamics. In S2 cells plated and spreading on concanavalinA-coated (conA) cover slips, endogenous Pod-1 localizes to sites of new actin polymerization, particularly at the lamellar edge. Costaining fixed cells with Pod-1 and Alexa488-phalloidin shows that Pod-1 is enriched at the edge of ruffling lamellae in an uneven, punctate pattern. Notably, Pod-1 colocalizes with a subset of actin, particularly the newly polymerized actin assembled at the lamellar edge that supports retrograde flow. Prominent staining is also seen on intralamellar actin filaments and filopodia-like structures (Rothenberg, 2003).
Costaining spreading cells for Pod-1 and Tubulin shows Pod-1 colocalizing with a subset of MTs -- especially those whose polymerizing ends are meeting the lamellar edge. In the rare cells that project filopodia, Pod-1 accumulates to high levels in those projections (comprised of actin bundles) and colocalizes with invading MTs. This subcellular localization is consistent with a molecule playing a role in actin/MT interactions (Rothenberg, 2003).
To investigate the dependence of Pod-1's subcellular localization on actin and MTs, cells were treated with latrunculin, a drug that leads to the depolymerization of actin microfilaments, or nocodazole, a drug that causes depolymerization of MTs. Nocodazole does not change the subcellular localization of Pod-1 even when MTs are completely depolymerized; thus, Pod-1 localization is independent of MTs. In contrast, latrunculin disrupts the actin cytoskeleton and causes Pod-1 to lose its characteristic localization at the lamellar edge. Costaining for Pod-1 and Tubulin shows that in latrunculin-treated cells, Pod-1 relocalizes from actin filaments to MTs, suggesting that Pod-1 had a high affinity for actin and a lower affinity for MTs or that its association with MTs may beregulated. This finding, together with the endogenous Pod-1 localization and the observed biochemical activities, is consistent with a role for Pod-1 in the interaction of actin and MTs in dynamic cellular structures (Rothenberg, 2003).
When Pod-1HIS (80-100 nM) is added simultaneously to phalloidin-stabilized fluorescent actin filaments (30-60 nM) and taxol-stabilized fluorescent MTs (500 nM), a dramatic crosslinking activity is observed in which actin bundles colocalize with MT bundles. Significantly, when the same experiment is performed with purified α-actinin, a well-characterized actin bundling protein, actin aggregates are seen, but there is no bundling of MTs. Thus, in vitro, Pod-1 could crosslink actin filaments and MTs, activities that are likely to be significant for the remodeling and coordination of actin and MT networks in dynamic cells (Rothenberg, 2003).
Slep, K. C. and Vale, R. D. (2007). Structural basis of microtubule plus end tracking by XMAP215, CLIP-170, and EB1. Mol. Cell 27(6): 976-91. Medline abstract: 17889670
Microtubule plus end binding proteins (+TIPS) localize to the dynamic plus ends of microtubules where they stimulate microtubule growth and recruit signaling molecules. Three main +TIP classes have been identified (XMAP215, EB1 and CLIP-170), but whether they act upon microtubule plus ends through a similar mechanism has not been resolved. This study reports crystal structures of the tubulin binding domains of XMAP215 (yeast Stu2p and Drosophila Msps), EB1 (yeast Bim1p and human EB1), and CLIP-170 (human), which reveal diverse tubulin binding interfaces. Functional studies, however, reveal a common property that native or artificial dimerization of tubulin binding domains (including chemically-induced heterodimers of EB1 and CLIP-170) induces tubulin nucleation/assembly in vitro and, in most cases, plus end tracking in living cells. It is proposed that +TIPs, although diverse in structure, share a common property of multimerizing tubulin, thus acting as polymerization chaperones that aid in subunit addition to the microtubule plus end (Slep, 2007).
This work analyzes the biochemical mechanisms of three plus end tracking protein families (EB1, CLIP-170 and XMAP215) using x-ray crystallography, in vitro tubulin binding/assembly assays, and in vivo plus end tracking assays. Five crystal structures are presented of +TIP tubulin binding domains across multiple species and families illustrating structural conservation within families and functional convergence across families. In vitro and in vivo analysis reveal an underlying general theme to these unrelated domains. While tubulin binding can be conferred by a single +TIP domain, the promotion of microtubule nucleation and polymerization in vitro and plus end tracking in vivo require multiple domains acting in concert. In the case of EB1 and CLIP-170 members, two domains are required for robust plus end localization and these two structurally unrelated domains can even cooperate if artificially joined together to achieve this activity. XMAP215 members are more complicated, perhaps due to the relative weaker association between TOG domains and tubulin in higher eukaryotic members. However, multimerization of TOG domains clearly facilitates tubulin nucleation in vitro and overall microtubule association in living cells. From these data, a model is presented in which unrelated tubulin binding domains have convergently evolved a similar mechanism for enhancing tubulin assembly on growing microtubule ends by acting as multivalent tubulin polymerization chaperones (Slep, 2007).
The tubulin binding domains of these three +TIP proteins are quite diverse in their architecture and show no evidence of a common evolutionary origin. However, the structural conservation within a +TIP class is very high, as evidenced by the near superimposition of yeast Bim1p with human EB1 and yeast Stu2p with Drosophila Mini spindles. The different architectures also suggest varying interactions modes with tubulin and potential synergy between the +TIPs. XMAP215 has a very flat binding interface created by a series of rigid, highly conserved loops on one face of the TOG domain; the length of the TOG domain is fairly closely matched to a tubulin monomer (α or β) and likely makes extensive contacts through a combination of hydrophobic and electrostatic interactions. In contrast to the elongated TOG domain, the CH domain of EB1 is spherical and approximately half the size. Mutagenesis suggests that several residues around one hemisphere contribute to tubulin binding. This data is consistent with recent work suggesting that EB1 might nestle in the groove between tubulin protofilaments. CLIP-170 also has a globular structure and the data suggests that it might bind tubulin’s disordered C-terminal tail as part of its binding interface (Slep, 2007).
Comparison of the +TIPS and protein engineering studies suggest a considerable structural variation in how +TIP proteins can be interconnected to achieve their activities in plus end tracking and promoting tubulin assembly. In some cases, tubulin binding domains are arrayed in tandem along a polypeptide chain with presumably unstructured linkers in between (e.g., Mini spindles, Ch-TOG, CLIP-170). The cis linking domains within a polypeptide appears to be critical, since it is shown that TOG domains added in trans do not reconstitute function. In other cases, single α/β tubulin binding domains are found in a polypeptide and multivalent tubulin binding is achieved by polypeptide dimerization [e.g., EB1 family and Bik1 (the CLIP-170 homologue from yeast)]. Hybrid strategies are also employed (e.g. CLIP-170, which has tandem Cap-Gly domains and is dimerized via a coiled coil and Stu2p which has tandem TOG domains, a dimerization domain and an additional C-terminal microtubule binding domain). It was also found that native dimerization sequences are not essential, as a variety of artificial dimerization strategies (e.g. GCN4 leucine zippers, gluthione S-transferase, and FKBP-rapamycin-FRB) are capable of reconstituting +TIP protein function (Slep, 2007).
Collectively, these studies reveal structural variation in how multiple +TIP tubulin binding domains can be combined to promote tubulin oligomerization for microtubule assembly. However, cis versus trans arrangements of +TIP domains may produce certain unique outcomes. For example, the linear arrays of TOG domains in the extended XMAP215 structure may generate a pseudo protofilament-like arrangement that is particularly effective for microtubule nucleation. Indeed, XMAP215 has a potent microtubule nucleation ability compared with CLIP-170, potentially related to an in vivo nucleation activity given XMAP215’s TACC-dependent localization to the centrosome. Analogous mechanisms exist for the actin cytoskeleton where the Spire protein utilizes several arrayed actin binding domains to template the nucleation of an actin filament (Slep, 2007).
The data show that single +TIP tubulin binding domains do not promote microtubule nucleation or growth in vitro (and in some cases are inhibitory), nor do they localize to microtubule plus ends in vivo. In contrast, multimerized +TIP tubulin binding domains are potent microtubule nucleator in vitro, promote microtubule growth in vitro and are requisite in vivo for microtubule plus end localization. For EB1 and CLIP-170, it was dynamically shown that plus end tracking requires an ability to bind more than one tubulin subunit. These results suggest a model in which +TIPs bind multiple tubulin dimers in solution and then deliver these larger tubulin oligomers to the ends of microtubules. This general idea was first introduced two decades ago by (Gard, 1987) to explain the high rates of tubulin assembly induced by XMAP215. Multimerization overcomes the inherent polymerization barrier tubulin heterodimers face due to single longitudinal and lateral tubulin:tubulin affinities estimated to be on the order of mM and M respectively. +TIP-induced multimerization of tubulin would increase the effective affinity for the microtubule lattice through cooperative binding, thereby decreasing the critical concentration for polymerization. By stabilizing tubulin-tubulin interactions prior to their full incorporation into a mature, cylindrical lattice, the +TIPs would act as polymerization chaperones. Physiologically, such chaperones would become particularly important for enhancing the growth of microtubules when free tubulin dimers are below the critical concentration of microtubule assembly or when microtubule destabilizing proteins are active (e.g in mitosis). This mechanism also enhances the spatial and temporal regulation of microtubule assembly in the cell, in part by regulating the localization and/or activity of +TIPs without modifying tubulin, the basic unit of polymerization (Slep, 2007).
Support for this general model comes from a number of other laboratories. Tandem Cap-Gly domains in CLIP-170 have been shown to plus end track whereas single Cap-Gly domains of the homolog CLIP-115 show greatly reduced microtubule association. Optical trapping studies analyzing microtubule growth showed step-wise growth that increased in size in the presence of full length XMAP215, a result they interpreted as XMAP215-facilitated incorporation of tubulin oligomers onto the microtubule end (Slep, 2007).
An apparent exception to the multiple tubulin binding rule for +TIPs is the CLASP family, which contains only one TOG domain at its N-terminus. However, informed by TOG structures and secondary structure predictions several TOG-like (TOGL) domains can be identified in the CLASP family. Two additional dodeca-helical domains are predicted with alternating loops that exhibit high homology to TOG domain intra-HEAT loops. If folded into a HEAT-like structure, the conservation profile of alternating loops would be localized to one face of the domain, suggesting that TOGL domains may bind tubulin by a mechanism similar to what is described in this study for the TOG domain. Thus, although poorly conserved at the primary structure level, XMAP215 and CLASP appear to be ancient +TIP relatives that may employ a similar general strategy for plus end tracking. Support for CLASP’s possible polymerization chaperone role comes from studies that observed CLASP-dependent microtubule subunit incorporation into fluxing kinetochore fibres (Slep, 2007).
The model for +TIPs as multimeric tubulin chaperones undoubtedly oversimplifies the complexity of interactions that are occurring at the microtubule plus end in vivo. +TIP families show divergent effects on microtubule dynamics, some promoting growth while others act as anti-pause, destabilization, rescue and perhaps even nucleation factors. Some of these differences in activity may reside in variations in the affinity constants of +TIP domains with tubulin monomers, oligomers and microtubules. As an example, non-EB1 family microtubule-binding CH domain proteins, including CLAMP and HEC1 appear to have higher affinity for the microtubule lattice than EB1 and localize along the entire length of the microtubule rather than just at plus ends. The spatial arrangement of +TIP domains, arrayed versus dimerized, may lead to distinct effects on microtubule dynamics by preferentially stabilizing lateral versus longitudinal tubulin-tubulin interactions. Finally, the intricate web of +TIP-+TIP interactions may generate unique outcomes on tubulin assembly. Multi-component in vitro assays, high resolution EM analysis of microtubules, and x-ray structures of +TIP-tubulin complexes will be required to further understand how +TIPs regulate the microtubule lattice (Slep, 2007).
Hereditary spastic paraplegias (HSPs), a group of neurodegenerative disorders characterized by lower-extremity spasticity and weakness, are most commonly caused by mutations in the spastin gene, which encodes a AAA+ ATPase related to the microtubule-severing protein katanin. A Drosophila homolog of spastin (D-spastin) has been identified, and D-spastin RNAi-treated or genetic null flies were shown to exhibit neurological defects; protein overexpression decreases the density of cellular microtubules. Elucidating spastins function and disease mechanism will require a more detailed understanding of its structure and biochemical mechanism. This study has investigated the effects of D-spastin, individual D-spastin domains, and D-spastin proteins bearing disease mutations on microtubules in cellular and in vitro assays. D-spastin, like katanin, displays ATPase activity and uses energy from ATP hydrolysis to sever and disassemble microtubules; disease mutations abolish or partially interfere with these activities (Roll-Mecak, 2005).
D-spastin from a Drosophila Schneider (S2) cell cDNA library, GFP was fused to its C terminus, and it was expressed in S2 cells in order to establish its subcellular distribution and in vivo activity. Previous localization studies of mammalian spastin have yielded diverse and often conflicting results, including nuclear localization, a mixture of nuclear and cytoplasmic distributions, and endosomal localization. At low expression levels, this study found that D-spastin-GFP was localized to discrete punctate structures that were distributed throughout the cytoplasm, suggestive of membrane vesicle association. Strikingly, cells expressing even relatively low levels of D-spastin-GFP showed diminished microtubule staining and many small microtubule fragments. Cells expressing higher levels of D-spastin-GFP displayed a nearly complete absence of cytoplasmic microtubules. These results are consistent with previous studies showing that spastin overexpression results in microtubule disassembly and with a recent report showing that spastin is targeted to membrane vesicles (Roll-Mecak, 2005).
Spastins are composed of three domains: an N-terminal region containing a putative transmembrane-spanning sequence (TM), a microtubule interacting and trafficking (MIT) domain that is well conserved in the spastin family, and an ATP binding AAA domain (AAA). Because the functions of these domains remain poorly understood, various truncated D-spastin-GFP constructs were expressed in S2 cells. A construct lacking the transmembrane-containing N terminus (ΔTM D-spastin-GFP) was diffusely distributed in the cytoplasm, indicating that the TM domain is required for the vesicular staining pattern observed for the full-length construct. In contrast, a construct that lacked the AAA domain (TM + MIT) still localized to punctate structures but did not affect the microtubule cytoskeleton. The MIT or AAA domains alone were diffusely localized and did not disassemble microtubules. Immunoblot analysis confirmed that these GFP fusion constructs expressed intact proteins. These results indicate that the TM/MIT region is involved in membrane targeting and the MIT/AAA region has activity on the microtubule cytoskeleton (Roll-Mecak, 2005).
The cellular expression do not prove that spastin is a microtubule disassembly agent because such effects might be indirect or require additional cellular cofactors. Therefore whether purified D-spastin is capable of microtubule disassembly in vitro was tested. D-Spastin lacking the TM region was expressed in E. coli as a GST fusion protein and was purified to higher than 95% homogeneity. The purified D-spastin was tested for ATPase activity and it was found to hydrolyze 1.8 ATP/subunit/s, similar to the maximal microtubule-stimulated ATPase rate reported for katanin. Thus far, no stimulation of the ATPase rate by taxol-stabilized microtubules has been observed. When purified D-spastin and ATP were applied to taxol-stabilized, rhodamine-labeled microtubules bound to a glass surface, the microtubules were completely disassembled after 2 min. At earlier time points, microtubules developed discrete breaks along their length. Severing of non-taxol-stabilized microtubules was also observed when D-spastin was added to interphase Xenopus extracts containing self-assembled microtubules. Thus, like katanin, D-spastin is capable of severing microtubules. The microtubule-severing reaction requires ATP hydrolysis; microtubules remained intact if ATP was omitted from the reaction or if the nonhydrolyzable ATP analog ATPγS was added instead of ATP. Microtubules became resistant to spastin-mediated severing when the negatively charged C-terminal peptide of tubulin was cleaved by subtilisin digestion. Subtilisin-treated microtubules are also resistant to severing by katanin (Roll-Mecak, 2005).
These experiments demonstrate that D-spastin alone can couple ATP hydrolysis to microtubule disassembly, and they confirm the cellular results that the MIT-AAA region is sufficient for this activity. It has been reported that the N-terminal TM region in human spastin is crucial for microtubule binding because truncated spastin lacking this region did not cosediment with microtubules. Although it remains possible that the TM region contributes to MT binding affinity, the current data show that the N terminus of D-spastin is not required for microtubule binding and severing (Roll-Mecak, 2005).
Attempts were made to express full-length spastin and the AAA domain in bacteria but they were found to be unstable and formed large aggregates. However, the MIT domain could be expressed and purified. This domain did not sever microtubules, and in cosedimentation binding experiments, it bound to microtubules much more weakly than the ΔTM construct in the absence of nucleotide. The lack of high-affinity binding is consistent with the finding that the MIT domain expressed in S2 cells did not colocalize with microtubules. Thus, high-affinity microtubule binding appears to require the combination of the MIT and the AAA domain or at least additional sequence beyond the minimal conserved-MIT domain investigated in this study (Roll-Mecak, 2005).
More than 140 spastin gene mutations have been isolated from HSP patients. Nonsense, frameshift, or splice site mutations are mostly scattered throughout the gene, whereas missense mutations are located almost exclusively in the AAA domain, underscoring the importance of this ATP binding module for spastin’s function. The 28 known missense mutations provide an array of tools with which to probe spastin’s function. Three such mutations were examined: a critical Walker A residue (K488R) that is predicted to impair ATP binding and mutations situated at the N- (S462C) and C-terminal (D655N) ends of the AAA domain. On the basis of structure-based sequence alignments, D655 resides in the C-terminal helical subdomain of the AAA module that contributes to nucleotide-dependent conformational changes. In addition to these known disease mutations, a mutation was also investigated in a key Walker B residue (E542A) known to be involved in ATP hydrolysis in other AAA+ ATPases (Roll-Mecak, 2005).
When full-length (FL) D-spastin constructs harboring these mutations were expressed in S2 cells, it was found that they localized to distinct punctate structures, as seen with wild-type D-spastin. However, for the Walker A (K488R) and Walker B (E542A) mutants, microtubule disassembly was not observed; instead, at moderate to high expression levels, microtubule bundling was observed around the perinuclear region. Surprisingly, the S462C and D655N mutants still severed microtubules when transfected into cells, although their activity was not as robust as wild-type D-spastin because higher levels of expression appear to be needed to disrupt the microtubule network. This result differs from a study showing that expression of the equivalent serine-to-cysteine human disease mutant did not cause microtubule destabilization (Roll-Mecak, 2005).
In the ΔTM constructs, the Walker A and B site mutants showed very different subcellular localizations. At low expression levels, the ATP-binding-deficient mutant K488R was diffusely distributed, and at high levels it coated and bundled microtubules. In contrast, the ATP-hydrolysis-deficient mutant E542A colocalized with and bundled only a subset of microtubules, perhaps suggestive of cooperative binding or binding to a specific subpopulation of microtubules. As noted for the full-length constructs, the expressed ΔTM S462C and D655N mutant proteins also caused partial microtubule disassembly (Roll-Mecak, 2005).
The biochemical activities of purified D-spastin mutant proteins were examined. The Walker A and B mutations, as expected, displayed neither detectable ATPase activity nor in vitro microtubule-severing activity. In contrast, the S462C and D655N mutant proteins both showed ATPase activity but were 40% and 80% decreased in maximal activity when compared with the wild-type protein. Surprisingly, the S462C and D655N mutants displayed a greater impairment in severing taxol-stabilized microtubules than might be expected on the basis of their ATPase activities. Whereas wild-type spastin almost completely disassembled taxol-stabilized microtubules on a surface within 2 min, the S462C mutant protein generated a comparable degree of disassembly only after >10 min, and the D655N mutant generated only a few microtubule breaks after a 20 min incubation. These data suggest that the S462C and D655N mutations impair ATPase activity but also produce a defect in the coupling of ATPase activity to microtubule destabilization (Roll-Mecak, 2005).
In summary, this study shows that D-spastin severs and disassembles microtubules both in cells and in vitro. These results provide the first direct biochemical evidence of spastin's ATPase and microtubule-destabilizing activities. It is interesting to compare spastin to katanin, another AAA+ ATPase that severs microtubules. Spastin's AAA domain is highly homologous to katanin, yet these proteins share no sequence similarity in their N-terminal regions. Thus, spastin and katanin appear to have evolved distinct microtubule binding domains that communicate with a similar motor module to bring about the severing reaction. Unlike katanin, which appears to be largely soluble or centrosome associated, spastin appears to associate with vesicular membranes through its N-terminal domain containing a predicted transmembrane-spanning sequence. This raises the possibility that spastin may be involved in remodeling the microtubule cytoskeleton near membrane surfaces, which may be important for spastin's function in synaptic architecture and transmission. Further investigations linking spastin's enzymatic activity to its cellular function will be required to better understand spastin's role in the nervous system as well as the mechanism by which spastin mutations give rise to the pathology of hereditary spastic paraplegias (Roll-Mecak, 2005).
Spastin, the most common locus for mutations in hereditary spastic paraplegias, and katanin are related microtubule-severing AAA ATPases involved in constructing neuronal and non-centrosomal microtubule arrays and in segregating chromosomes. The mechanism by which spastin and katanin break and destabilize microtubules is unknown, in part owing to the lack of structural information on these enzymes. This study reports the X-ray crystal structure of the Drosophila spastin AAA domain and provides a model for the active spastin hexamer generated using small-angle X-ray scattering combined with atomic docking. The spastin hexamer forms a ring with a prominent central pore and six radiating arms that may dock onto the microtubule. Helices unique to the microtubule-severing AAA ATPases surround the entrances to the pore on either side of the ring, and three highly conserved loops line the pore lumen. Mutagenesis reveals essential roles for these structural elements in the severing reaction. Peptide and antibody inhibition experiments further show that spastin may dismantle microtubules by recognizing specific features in the carboxy-terminal tail of tubulin. Collectively, these data support a model in which spastin pulls the C terminus of tubulin through its central pore, generating a mechanical force that destabilizes tubulin-tubulin interactions within the microtubule lattice. This work also provides insights into the structural defects in spastin that arise from mutations identified in hereditary spastic paraplegia patients (Roll-Mecak, 2008).
The combination of X-ray crystallography, SAXS ab initio reconstructions and structure-guided mutagenesis provides the first structural information on microtubule-severing proteins and allows the proposal of a molecular model for spastin-mediated severing. Owing to their similar domain organization and high sequence similarity, this model probably pertains to katanin as well. It is proposed that face A of the spastin AAA ring docks onto the microtubule, placing the positively charged N-terminal pore entrance in contact with the negatively charged C terminus of tubulin. The translocation from face A to face B would correspond to the direction of substrate translocation proposed for the distantly related AAA ATPases ClpX, ClpA and ClpB. The linker and MIT domains extending from the ring would make additional contacts with the microtubule, thus increasing microtubule avidity and potentially stabilizing the hexamer on the microtubule. On the basis of affinity measurements, only a subset of the six arms is likely to make strong binding interactions with the microtubule (Roll-Mecak, 2008).
It is proposed that the tubulin polypeptide is threaded through the pore, perhaps driven by nucleotide-driven conformational changes of the pore loops. However, spastin may not need to completely translocate the tubulin polypeptide substrate, but instead just grip the C-terminal tubulin tail and exert mechanical 'tugs' that might partially unfold tubulin or locally destabilize protomer-protomer interactions, leading to catastrophic breakdown of the microtubule lattice. It also remains possible that the MIT domains could participate in this nucleotide-driven process by binding and 'feeding' the C-terminal tails to the pore. Further biophysical characterization will be needed to decipher the structural details of substrate recognition and mechanical force production. The data also suggest that spastin may selectively recognize post-translationally modified tubulins ('Glu' tubulins) that are part of stable microtubules. Consistent with this idea, loss of spastin in Drosophila results in the accumulation of stabilized polyglutamylated tubulin in neurons and spastin knockout mice show axonal swellings enriched in detyrosinated, stable microtubules. The structure also provides the first glimpse into how spastin disease mutations contribute to spastin dysfunction and disease, most of which are likely to be involved in destabilizing protomer-protomer interactions, microtubule- or ATP-binding; in such cases, spastin-linked HSP is probably caused by haploinsufficiency and not a dominant negative effect. Further elucidation of the mechanistic details of how spastin interacts with particular tubulin isoforms and post-translational modifications and leads to microtubule destabilization may provide insight into the origin of spastin paraplegias and potential treatments for this disease (Roll-Mecak, 2008).
Return to beta Tubulin 56D Protein interactions part 1/2
Home page: The Interactive Fly © 1995, 1996 Thomas B. Brody, Ph.D.
The Interactive Fly resides on the
Society for Developmental Biology's Web server.