Structure of the γ-tubulin ring complex-capped microtubule.

Microtubules are composed of α-tubulin and ß-tubulin dimers positioned head-to-tail to form protofilaments that associate laterally in varying numbers. It is not known how cellular microtubules assemble with the canonical 13-protofilament architecture, resulting in micrometer-scale α/ß-tubulin tracks for intracellular transport that align with, rather than spiral along, the long axis of the filament. We report that the human ~2.3 MDa γ-tubulin ring complex (γ-TuRC), an essential regulator of microtubule formation that contains 14 γ-tubulins, selectively nucleates 13-protofilament microtubules. Cryogenic electron microscopy reconstructions of γ-TuRC-capped microtubule minus ends reveal the extensive intra-domain and inter-domain motions of γ-TuRC subunits that accommodate luminal bridge components and establish lateral and longitudinal interactions between γ-tubulins and α-tubulins. Our structures suggest that γ-TuRC, an inefficient nucleation template owing to its splayed conformation, can transform into a compacted cap at the microtubule minus end and set the lattice architecture of cellular microtubules.

Spastin regulates anaphase chromosome separation distance and microtubule-containing nuclear tunnels.

Nuclear envelope reassembly during the final stages of each mitosis depends on disassembling spindle microtubules without disrupting chromosome separation. This process involves the transient recruitment of the ESCRT-III complex and spastin, a microtubule-severing AAA (ATPases associated with diverse cellular activities) mechanoenzyme, to late-anaphase chromosomes. However, dissecting mechanisms underlying these rapid processes, which can be completed within minutes, has been difficult. Here, we combine fast-acting chemical inhibitors with live-cell imaging and find that spindle microtubules, along with spastin activity, regulate the number and lifetimes of spastin foci at anaphase chromosomes. Unexpectedly, spastin inhibition impedes chromosome separation, but does not alter the anaphase localization dynamics of CHMP4B, an ESCRT-III protein, or increase γ-H2AX foci, a DNA damage marker. We show spastin inhibition increases the frequency of lamin-lined nuclear microtunnels that can include microtubules penetrating the nucleus. Our findings suggest failure to sever spindle microtubules impedes chromosome separation, yet reforming nuclear envelopes can topologically accommodate persistent microtubules ensuring nuclear DNA is not damaged or exposed to cytoplasm.

Structure of the γ-tubulin ring complex-capped microtubule.

Microtubules are composed of α/ß-tubulin dimers positioned head-to-tail to form protofilaments that associate laterally in varying numbers. It is not known how cellular microtubules assemble with the canonical 13-protofilament architecture, resulting in micrometer-scale α/ß-tubulin tracks for intracellular transport that align with, rather than spiral along, the filament's long-axis. We report that the human ∼2.3MDa γ-tubulin ring complex (γ-TuRC), an essential regulator of microtubule formation that contains 14 γ-tubulins, selectively nucleates 13-protofilament microtubules. Cryo-EM reconstructions of γ-TuRC-capped microtubule minus-ends reveal the extensive intra- and inter-domain motions of γ-TuRC subunits that accommodate its actin-containing luminal bridge and establish lateral and longitudinal interactions between γ- and α-tubulins. Our structures reveal how free γ-TuRC, an inefficient nucleation template due to its splayed conformation, transforms into a stable cap that blocks addition or loss of α/ß-tubulins from minus-ends and sets the lattice architecture of cellular microtubules. One Sentence Summary: Structural insights into how the γ-tubulin ring complex nucleates and caps a 13-protofilament microtubule.

A nucleotide binding-independent role for γ-tubulin in microtubule capping and cell division.

The γ-tubulin ring complex (γ-TuRC) has essential roles in centrosomal and non-centrosomal microtubule organization during vertebrate mitosis. While there have been important advances in understanding γ-TuRC-dependent microtubule nucleation, γ-TuRC capping of microtubule minus-ends remains poorly characterized. Here, we utilized biochemical reconstitutions and cellular assays to characterize the human γ-TuRC's capping activity. Single filament assays showed that the γ-TuRC remained associated with a nucleated microtubule for tens of minutes. In contrast, caps at dynamic microtubule minus-ends displayed lifetimes of ∼1 min. Reconstituted γ-TuRCs with nucleotide-binding deficient γ-tubulin (γ-tubulinΔGTP) formed ring-shaped complexes that did not nucleate microtubules but capped microtubule minus-ends with lifetimes similar to those measured for wild-type complexes. In dividing cells, microtubule regrowth assays revealed that while knockdown of γ-tubulin suppressed non-centrosomal microtubule formation, add-back of γ-tubulinΔGTP could substantially restore this process. Our results suggest that γ-TuRC capping is a nucleotide-binding-independent activity that plays a role in non-centrosomal microtubule organization during cell division.

Biochemical reconstitutions reveal principles of human γ-TuRC assembly and function.

The formation of cellular microtubule networks is regulated by the γ-tubulin ring complex (γ-TuRC). This ∼2.3 MD assembly of >31 proteins includes γ-tubulin and GCP2-6, as well as MZT1 and an actin-like protein in a "lumenal bridge" (LB). The challenge of reconstituting the γ-TuRC has limited dissections of its assembly and function. Here, we report a biochemical reconstitution of the human γ-TuRC (γ-TuRC-GFP) as a ∼35 S complex that nucleates microtubules in vitro. In addition, we generate a subcomplex, γ-TuRCΔLB-GFP, which lacks MZT1 and actin. We show that γ-TuRCΔLB-GFP nucleates microtubules in a guanine nucleotide-dependent manner and with similar efficiency as the holocomplex. Electron microscopy reveals that γ-TuRC-GFP resembles the native γ-TuRC architecture, while γ-TuRCΔLB-GFP adopts a partial cone shape presenting only 8-10 γ-tubulin subunits and lacks a well-ordered lumenal bridge. Our results show that the γ-TuRC can be reconstituted using a limited set of proteins and suggest that the LB facilitates the self-assembly of regulatory interfaces around a microtubule-nucleating "core" in the holocomplex.

Force-dependent stimulation of RNA unwinding by SARS-CoV-2 nsp13 helicase.

The superfamily 1 helicase nonstructural protein 13 (nsp13) is required for SARS-CoV-2 replication. The mechanism and regulation of nsp13 has not been explored at the single-molecule level. Specifically, force-dependent unwinding experiments have yet to be performed for any coronavirus helicase. Here, using optical tweezers, we find that nsp13 unwinding frequency, processivity, and velocity increase substantially when a destabilizing force is applied to the RNA substrate. These results, along with bulk assays, depict nsp13 as an intrinsically weak helicase that can be activated >50-fold by piconewton forces. Such force-dependent behavior contrasts the known behavior of other viral monomeric helicases, such as hepatitis C virus NS3, and instead draws stronger parallels to ring-shaped helicases. Our findings suggest that mechanoregulation, which may be provided by a directly bound RNA-dependent RNA polymerase, enables on-demand helicase activity on the relevant polynucleotide substrate during viral replication.

Force-dependent stimulation of RNA unwinding by SARS-CoV-2 nsp13 helicase.

The superfamily-1 helicase non-structural protein 13 (nsp13) is required for SARS-CoV-2 replication, making it an important antiviral therapeutic target. The mechanism and regulation of nsp13 has not been explored at the single-molecule level. Specifically, force-dependent unwinding experiments have yet to be performed for any coronavirus helicase. Here, using optical tweezers, we find that nsp13 unwinding frequency, processivity, and velocity increase substantially when a destabilizing force is applied to the dsRNA, suggesting a passive unwinding mechanism. These results, along with bulk assays, depict nsp13 as an intrinsically weak helicase that can be potently activated by picoNewton forces. Such force-dependent behavior contrasts the known behavior of other viral monomeric helicases, drawing stronger parallels to ring-shaped helicases. Our findings suggest that mechanoregulation, which may be provided by a directly bound RNA-dependent RNA polymerase, enables on-demand helicase activity on the relevant polynucleotide substrate during viral replication.

MKLP2 Is a Motile Kinesin that Transports the Chromosomal Passenger Complex during Anaphase.

During cytokinesis, signals from the anaphase spindle direct the formation and position of a contractile ring at the cell cortex [1]. The chromosomal passenger complex (CPC) participates in cytokinesis initiation by signaling from the spindle midzone and equatorial cortex [2], but the mechanisms underlying the anaphase-specific CPC localization are currently unresolved. Accumulation of the CPC at these sites requires the presence of microtubules and the mitotic kinesin-like protein 2, MKLP2 (KIF20A), a member of the kinesin-6 family [2-7], and this has led to the hypothesis that the CPC is transported along microtubules by MKLP2 [3-5, 7]. However, the structure of the MKLP2 motor domain with its extended neck-linker region suggests that this kinesin might not be able to drive processive transport [8, 9]. Furthermore, experiments in Xenopus egg extracts indicated that the CPC might be transported by kinesin-4, KIF4A [10]. Finally, CPC-MKLP2 complexes might be directly recruited to the equatorial cortex via association with actin and myosin II, independent of kinesin activity [4, 8]. Using microscopy-based assays with purified proteins, we demonstrate that MKLP2 is a processive plus-end directed motor that can transport the CPC along microtubules in vitro. In cells, strong suppression of MKLP2-dependent CPC motility by expression of an MKLP2 P-loop mutant perturbs CPC accumulation at both the spindle midzone and equatorial cortex, whereas a weaker inhibition of MKLP2 motor using Paprotrain mainly affects CPC localization to the equatorial cortex. Our data indicate that control of cytokinesis initiation by the CPC requires its directional MKLP2-dependent transport.

CLASP Mediates Microtubule Repair by Restricting Lattice Damage and Regulating Tubulin Incorporation.

Microtubules play a key role in cell division, motility, and intracellular trafficking. Microtubule lattices are generally regarded as stable structures that undergo turnover through dynamic instability of their ends [1]. However, recent evidence suggests that microtubules also exchange tubulin dimers at the sites of lattice defects, which can be induced by mechanical stress, severing enzymes, or occur spontaneously during polymerization [2-6]. Tubulin incorporation can restore microtubule integrity; moreover, "islands" of freshly incorporated GTP-tubulin can inhibit microtubule disassembly and promote rescues [3, 4, 6-8]. Microtubule repair occurs in vitro in the presence of tubulin alone [2-6, 9]. However, in cells, it is likely to be regulated by specific factors, the nature of which is currently unknown. CLASPs are interesting candidates for microtubule repair because they induce microtubule nucleation, stimulate rescue, and suppress catastrophes by stabilizing incomplete growing plus ends with lagging protofilaments and promoting their conversion into complete ones [10-17]. Here, we used in vitro reconstitution assays combined with laser microsurgery and microfluidics to show that CLASP2α indeed stimulates microtubule lattice repair. CLASP2α promoted tubulin incorporation into damaged lattice sites, thereby restoring microtubule integrity. Furthermore, it induced the formation of complete tubes from partial protofilament assemblies and inhibited microtubule softening caused by hydrodynamic-flow-induced bending. The catastrophe-suppressing domain of CLASP2α, TOG2, combined with a microtubule-tethering region, was sufficient to stimulate microtubule repair, suggesting that catastrophe suppression and lattice repair are mechanistically similar. Our results suggest that the cellular machinery controlling microtubule nucleation and growth can also help to maintain microtubule integrity.

Mechanisms of Motor-Independent Membrane Remodeling Driven by Dynamic Microtubules.

Microtubule-dependent organization of membranous organelles occurs through motor-based pulling and by coupling microtubule dynamics to membrane remodeling. For example, tubules of endoplasmic reticulum (ER) can be extended by kinesin- and dynein-mediated transport and through the association with the tips of dynamic microtubules. The binding between ER and growing microtubule plus ends requires End Binding (EB) proteins and the transmembrane protein STIM1, which form a tip-attachment complex (TAC), but it is unknown whether these proteins are sufficient for membrane remodeling. Furthermore, EBs and their partners undergo rapid turnover at microtubule ends, and it is unclear how highly transient protein-protein interactions can induce load-bearing processive motion. Here, we reconstituted membrane tubulation in a minimal system with giant unilamellar vesicles, dynamic microtubules, an EB protein, and a membrane-bound protein that can interact with EBs and microtubules. We showed that these components are sufficient to drive membrane remodeling by three mechanisms: membrane tubulation induced by growing microtubule ends, motor-independent membrane sliding along microtubule shafts, and membrane pulling by shrinking microtubules. Experiments and modeling demonstrated that the first two mechanisms can be explained by adhesion-driven biased membrane spreading on microtubules. Optical trapping revealed that growing and shrinking microtubule ends can exert forces of ∼0.5 and ∼5 pN, respectively, through attached proteins. Rapidly exchanging molecules that connect membranes to dynamic microtubules can thus bear a sufficient load to induce membrane deformation and motility. Furthermore, combining TAC components and a membrane-attached kinesin in the same in vitro assays demonstrated that they can cooperate in promoting membrane tubule extension.

Feedback-Driven Assembly of the Axon Initial Segment.

The axon initial segment (AIS) is a unique neuronal compartment that plays a crucial role in the generation of action potential and neuronal polarity. The assembly of the AIS requires membrane, scaffolding, and cytoskeletal proteins, including Ankyrin-G and TRIM46. How these components cooperate in AIS formation is currently poorly understood. Here, we show that Ankyrin-G acts as a scaffold interacting with End-Binding (EB) proteins and membrane proteins such as Neurofascin-186 to recruit TRIM46-positive microtubules to the plasma membrane. Using in vitro reconstitution and cellular assays, we demonstrate that TRIM46 forms parallel microtubule bundles and stabilizes them by acting as a rescue factor. TRIM46-labeled microtubules drive retrograde transport of Neurofascin-186 to the proximal axon, where Ankyrin-G prevents its endocytosis, resulting in stable accumulation of Neurofascin-186 at the AIS. Neurofascin-186 enrichment in turn reinforces membrane anchoring of Ankyrin-G and subsequent recruitment of TRIM46-decorated microtubules. Our study reveals feedback-based mechanisms driving AIS assembly.

Exocyst Subcomplex Functions in Autophagosome Biogenesis by Regulating Atg9 Trafficking.

During autophagy, double-membrane vesicles called autophagosomes capture and degrade the intracellular cargo. The de novo formation of autophagosomes requires several vesicle transport and membrane fusion events which are not completely understood. We studied the involvement of exocyst, an octameric tethering complex, which has a primary function in tethering post-Golgi secretory vesicles to plasma membrane, in autophagy. Our findings indicate that not all subunits of exocyst are involved in selective and general autophagy. We show that in the absence of autophagy specific subunits, autophagy arrest is accompanied by accumulation of incomplete autophagosome-like structures. In these mutants, impaired Atg9 trafficking leads to decreased delivery of membrane to the site of autophagosome biogenesis thereby impeding the elongation and completion of the autophagosomes. The subunits of exocyst, which are dispensable for autophagic function, do not associate with the autophagy specific subcomplex of exocyst.

CLASP Suppresses Microtubule Catastrophes through a Single TOG Domain.

The dynamic instability of microtubules plays a key role in controlling their organization and function, but the cellular mechanisms regulating this process are poorly understood. Here, we show that cytoplasmic linker-associated proteins (CLASPs) suppress transitions from microtubule growth to shortening, termed catastrophes, including those induced by microtubule-destabilizing agents and physical barriers. Mammalian CLASPs encompass three TOG-like domains, TOG1, TOG2, and TOG3, none of which bind to free tubulin. TOG2 is essential for catastrophe suppression, whereas TOG3 mildly enhances rescues but cannot suppress catastrophes. These functions are inhibited by the C-terminal domain of CLASP2, while the TOG1 domain can release this auto-inhibition. TOG2 fused to a positively charged microtubule-binding peptide autonomously accumulates at growing but not shrinking ends, suppresses catastrophes, and stimulates rescues. CLASPs suppress catastrophes by stabilizing growing microtubule ends, including incomplete ones, preventing their depolymerization and promoting their recovery into complete tubes. TOG2 domain is the key determinant of these activities.

Tipping microtubule dynamics, one protofilament at a time.

Microtubules are polymeric tubes that switch between phases of growth and shortening, and this property is essential to drive key cellular processes. Microtubules are composed of protofilaments formed by longitudinally arranged tubulin dimers. Microtubule dynamics can be affected by structural perturbations at the plus end, such as end tapering, and targeting only a small subset of protofilaments can alter the dynamics of the whole microtubule. Microtubule lattice plasticity, including compaction along the longitudinal axis upon GTP hydrolysis and tubulin dimer loss and reinsertion along microtubule shafts can also affect microtubule dynamics or mechanics. Microtubule behaviour can be fine-tuned by post-translational modifications and tubulin isotypes, which together support the diversity of microtubule functions within and across various cell types or cell cycle and developmental stages.

Septins are involved at the early stages of macroautophagy in S. cerevisiae.

Autophagy is a conserved cellular degradation pathway wherein double-membrane vesicles called autophagosomes capture long-lived proteins, and damaged or superfluous organelles, and deliver them to the lysosome for degradation. Septins are conserved GTP-binding proteins involved in many cellular processes, including phagocytosis and the autophagy of intracellular bacteria, but no role in general autophagy was known. In budding yeast, septins polymerize into ring-shaped arrays of filaments required for cytokinesis. In an unbiased genetic screen and in subsequent targeted analysis, we found autophagy defects in septin mutants. Upon autophagy induction, pre-assembled septin complexes relocalized to the pre-autophagosomal structure (PAS) where they formed non-canonical septin rings at PAS. Septins also colocalized with autophagosomes, where they physically interacted with the autophagy proteins Atg8 and Atg9. When autophagosome degradation was blocked in septin-mutant cells, fewer autophagic structures accumulated, and an autophagy mutant defective in early stages of autophagosome biogenesis (atg1Δ), displayed decreased septin localization to the PAS. Our findings support a role for septins in the early stages of budding yeast autophagy, during autophagosome formation.This article has an associated First Person interview with the first author of the paper.

Deconvolution of Buparlisib's mechanism of action defines specific PI3K and tubulin inhibitors for therapeutic intervention.

BKM120 (Buparlisib) is one of the most advanced phosphoinositide 3-kinase (PI3K) inhibitors for the treatment of cancer, but it interferes as an off-target effect with microtubule polymerization. Here, we developed two chemical derivatives that differ from BKM120 by only one atom. We show that these minute changes separate the dual activity of BKM120 into discrete PI3K and tubulin inhibitors. Analysis of the compounds cellular growth arrest phenotypes and microtubule dynamics suggest that the antiproliferative activity of BKM120 is mainly due to microtubule-dependent cytotoxicity rather than through inhibition of PI3K. Crystal structures of BKM120 and derivatives in complex with tubulin and PI3K provide insights into the selective mode of action of this class of drugs. Our results raise concerns over BKM120's generally accepted mode of action, and provide a unique mechanistic basis for next-generation PI3K inhibitors with improved safety profiles and flexibility for use in combination therapies.

The C-terminal tails of heterotrimeric kinesin-2 motor subunits directly bind to α-tubulin1: Possible implications for cilia-specific tubulin entry.

The assembly of microtubule-based cytoskeleton propels the cilia and flagella growth. Previous studies have indicated that the kinesin-2 family motors transport tubulin into the cilia through intraflagellar transport. Here, we report a direct interaction between the C-terminal tail fragments of heterotrimeric kinesin-2 and α-tubulin1 isoforms in vitro. Blot overlay screen, affinity purification from tissue extracts, cosedimentation with subtilisin-treated microtubule and LC-ESI-MS/MS characterization of the tail-fragment-associated tubulin identified an association between the tail domains and α-tubulin1A/D isotype. The interaction was confirmed by Forster's resonance energy transfer assay in tissue-cultured cells. The overexpression of the recombinant tails in NIH3T3 cells affected the primary cilia growth, which was rescued by coexpression of a α-tubulin1 transgene. Furthermore, fluorescent recovery after photobleach analysis in the olfactory cilia of Drosophila indicated that tubulin is transported in a non-particulate form requiring kinesin-2. These results provide additional new insight into the mechanisms underlying selective tubulin isoform enrichment in the cilia.

Centriolar CPAP/SAS-4 Imparts Slow Processive Microtubule Growth.

Centrioles are fundamental and evolutionarily conserved microtubule-based organelles whose assembly is characterized by microtubule growth rates that are orders of magnitude slower than those of cytoplasmic microtubules. Several centriolar proteins can interact with tubulin or microtubules, but how they ensure the exceptionally slow growth of centriolar microtubules has remained mysterious. Here, we bring together crystallographic, biophysical, and reconstitution assays to demonstrate that the human centriolar protein CPAP (SAS-4 in worms and flies) binds and "caps" microtubule plus ends by associating with a site of ß-tubulin engaged in longitudinal tubulin-tubulin interactions. Strikingly, we uncover that CPAP activity dampens microtubule growth and stabilizes microtubules by inhibiting catastrophes and promoting rescues. We further establish that the capping function of CPAP is important to limit growth of centriolar microtubules in cells. Our results suggest that CPAP acts as a molecular lid that ensures slow assembly of centriolar microtubules and, thereby, contributes to organelle length control.