The oncogenic tyrosine phosphatase PTP4A3 is an attractive cancer therapeutic target because it is overexpressed in many types of cancer and promotes tumor metastasis, contributing to poor patient prognosis. We recently synthesized JMS‐053, which is a potent, selective, reversible, iminothienopyridinedione inhibitor of PTP4A3. We modeled possible interactions between JMS‐053 and PTP4A3 and the goal of our current work was to test the model and further define the crucial amino acids involved in the PTP4A3 inhibition by JMS‐053. We generated three PTP4A3 single amino acid mutants: C104S, K144I and C49S. The enzymatic activities of these phosphatases were determined in vitro in the absence and presence of JMS‐053, using the artificial substrate 6,8‐difluoro‐4‐methylumbelliferyl phosphate. The C104S mutant showed no catalytic activity, consistent with C104 being the catalytic cysteine. The K144I mutant retained full enzymatic activity. In contrast to our molecular modeling predictions, the K144I mutant was inhibited by JMS‐053 to the same extent as the wild type enzyme. We concluded K144 was not primarily involved in JMS‐053 inhibition. Tyrosine phosphatases are known to be inhibited by intramolecular disulfide bond formation and C49S has been reported to be the only cysteine involved with disulfide bond formation with the catalytic C104. The C49S mutant had reduced catalytic activity compared to the wild type PTP4A3 but JMS‐053 retained partial inhibition. Thus, C49‐C104 disulfide bond formation does not appear to be the primary mode of JMS‐053 inhibition, although we cannot completely exclude involvement of C49. We are now examining the oxidation state of C104 after JMS‐053 treatment and new analogs of JMS‐053. The information obtained from mapping the critical amino acids involved in PTP4A3 inhibition should assist in the design of the next generation phosphatase inhibitors. Support or Funding Information The Department of Defense (BC170507), the Fiske Drug Discovery Fund and the Ivy Foundation. This abstract is from the Experimental Biology 2019 Meeting. There is no full text article associated with this abstract published in The FASEB Journal .
ABSTRACT Dynein harnesses ATP hydrolysis to move cargo on microtubules in multiple biological contexts. Dynein meets a unique challenge in meiosis by moving chromosomes tethered to the nuclear envelope to facilitate homolog pairing essential for gametogenesis. Though processive dynein motility requires binding to an activating adaptor, the identity of the activating adaptor required for dynein to move meiotic chromosomes is unknown. We show that the meiosis-specific nuclear-envelope protein KASH5 is a dynein activating adaptor: KASH5 directly binds dynein using a mechanism conserved among activating adaptors and converts dynein into a processive motor. We map the dynein-binding surface of KASH5, identifying mutations that abrogate dynein binding in vitro and disrupt recruitment of the dynein machinery to the nuclear envelope in cultured cells and mouse spermatocytes in vivo . Our study identifies KASH5 as the first transmembrane dynein activating adaptor and provides molecular insights into how it activates dynein during meiosis.
Oncogenic protein tyrosine phosphatases (PTPs) are overexpressed in numerous human cancers but they have been challenging pharmacological targets. The emblematic oncogenic PTP4A tyrosine phosphatase family regulates many fundamental malignant processes. 7-Imino-2-phenylthieno[3,2-c]pyridine-4,6(5H,7H)-dione (JMS-053) is a novel, potent, and selective PTP4A inhibitor but its mechanism of action has not been fully elucidated, nor has the chemotype been fully investigated. Because tyrosine phosphatases are notoriously susceptible to oxidation, we interrogated JMS-053 and three newly synthesized analogs with specific attention on the role of oxidation. JMS-053 and its three analogs were potent in vitro PTP4A3 inhibitors, but 7-imino-5-methyl-2-phenylthieno[3,2-c]pyridine-4,6(5H,7H)-dione (NRT-870-59) appeared unique among the thienopyridinediones with respect to its inhibitory specificity for PTP4A3 versus both a PTP4A3 A111S mutant and an oncogenic dual specificity tyrosine phosphatase, CDC25B. Like JMS-053, NRT-870-59 was a reversible PTP4A3 inhibitor. All of the thienopyridinediones retained cytotoxicity against human ovarian and breast cancer cells grown as pathologically relevant three-dimensional spheroids. Inhibition of cancer cell colony formation by NRT-870-59, like JMS-053, required PTP4A3 expression. JMS-053 failed to generate significant detectable reactive oxygen species in vitro or in cancer cells. Mass spectrometry results indicated no disulfide bond formation or oxidation of the catalytic Cys104 after in vitro incubation of PTP4A3 with JMS-053 or NRT-870-59. Gene expression profiling of cancer cells exposed to JMS-053 phenocopied many of the changes seen with the loss of PTP4A3 and did not indicate oxidative stress. These data demonstrate that PTP4A phosphatases can be selectively targeted with small molecules that lack prominent reactive oxygen species generation and encourage further studies of this chemotype.
SIGNIFICANCE STATEMENT
Protein tyrosine phosphatases are emerging as important contributors to human cancers. We report on a new class of reversible protein phosphatase small molecule inhibitors that are cytotoxic to human ovarian and breast cancer cells, do not generate significant reactive oxygen species in vitro and in cells, and could be valuable lead molecules for future studies of PTP4A phosphatases.
Article Figures and data Abstract Editor's evaluation Introduction Results Discussion Materials and methods Data availability References Decision letter Author response Article and author information Metrics Abstract Dynein harnesses ATP hydrolysis to move cargo on microtubules in multiple biological contexts. Dynein meets a unique challenge in meiosis by moving chromosomes tethered to the nuclear envelope to facilitate homolog pairing essential for gametogenesis. Though processive dynein motility requires binding to an activating adaptor, the identity of the activating adaptor required for dynein to move meiotic chromosomes is unknown. We show that the meiosis-specific nuclear-envelope protein KASH5 is a dynein activating adaptor: KASH5 directly binds dynein using a mechanism conserved among activating adaptors and converts dynein into a processive motor. We map the dynein-binding surface of KASH5, identifying mutations that abrogate dynein binding in vitro and disrupt recruitment of the dynein machinery to the nuclear envelope in cultured cells and mouse spermatocytes in vivo. Our study identifies KASH5 as the first transmembrane dynein activating adaptor and provides molecular insights into how it activates dynein during meiosis. Editor's evaluation This manuscript identifies a meiosis-specific protein that recruits and activates the motility of the dynein-1 transport machinery at the nuclear envelope. In prophase I of meiosis, dynein moves chromosomes tethered to the nuclear envelope to expedite the search and pairing between homologous chromosomes. Previous studies have shown that dynein tethers to chromosomes via the LINC complex, which consists of a SUN protein and transmembrane KASH protein. KASH5 comprises all the known features of bona fide cargo adaptors of dynein, and this manuscript demonstrated that KASH5 directly binds dynein and activates its processive motility. https://doi.org/10.7554/eLife.78201.sa0 Decision letter Reviews on Sciety eLife's review process Introduction Cytoplasmic dynein-1 (dynein) is the primary retrograde, microtubule-associated molecular motor in most eukaryotes (Cianfrocco et al., 2015; Schroer et al., 1989). Dynein performs cellular work by coupling the energy derived from ATP hydrolysis to the movement of cellular cargo along the microtubule filament (Roberts et al., 2013; Carter, 2013; Carter et al., 2016). Dynein is a multi-subunit protein complex that comprises two copies of six different subunits, including the heavy chain (HC) that contains the ATPase motor, the intermediate chain, the light-intermediate chain (LIC), and three different light chains (Canty et al., 2021; Figure 1—figure supplement 1A). Dynein is not a processive motor on its own. Dynein achieves processive and directional motility by assembling into the activated dynein complex that, in addition to dynein, includes the 23-subunit protein complex dynactin and one of a class of proteins called activating adaptors (McKenney et al., 2014; Schlager et al., 2014; Urnavicius et al., 2018; Grotjahn et al., 2018; Urnavicius et al., 2015; Figure 1—figure supplement 1A). There are ~12 confirmed activating adaptors, defined by their ability to bind dynein and induce fast and processive dynein motility on immobilized microtubules in vitro (Reck-Peterson et al., 2018; Olenick and Holzbaur, 2019). Each activating adaptor contains an N-terminal domain that binds to the dynein LIC, followed by a coiled-coil domain (CC) that binds to both dynein HC and dynactin (Reck-Peterson et al., 2018; Olenick and Holzbaur, 2019; Figure 1A). In addition to binding dynein-dynactin, each activating adaptor uses its C-termthe coinal domain to bind cellular cargo and link it to the dynein motor (Reck-Peterson et al., 2018; Olenick and Holzbaur, 2019; Redwine et al., 2017). The currently known activating adaptors fall into three families based on the structure of their LIC-binding N-terminal regions (Lee et al., 2018; Lee et al., 2020; Schroeder et al., 2014). LIC-binding domains are either ‘CC-boxes’ (as in BICDR1 and BicD2) (McKenney et al., 2014; Schlager et al., 2014; Urnavicius et al., 2018; Elshenawy et al., 2019), Hook-domains (as in Hook1 and Hook3) (McKenney et al., 2014; Urnavicius et al., 2018; Olenick et al., 2016; Schroeder and Vale, 2016), or EF-hand-pairs (as in ninein, ninein-like, and CRACR2a) (Redwine et al., 2017; Lee et al., 2020; Wang et al., 2019). An EF-hand in its canonical form is a Ca2+-binding structural motif, although several EF-hand proteins adopt this structure without the aid of Ca2+ Kawasaki et al., 1998. Consistent with this, some dynein activating adaptors like CRACR2A bind Ca2+ and require it for dynein activation, while others like Rab45 and ninein-like do not (Lee et al., 2020; Wang et al., 2019). Figure 1 with 1 supplement see all Download asset Open asset KASH5 binds dynein via a direct NCC-light-intermediate chain (LIC) interaction. (A) Domain arrangement of a typical dynein activating adaptor showing an N-terminal LIC-interacting domain, a central coiled-coil (CC), and a C-terminal cargo-binding domain. Note that this description is simplified for BicD2 as it is primarily CC and thus uses a CC region toward the N-terminus to bind LIC. (B) Schematic for how the telomere and dynein tether to each other at the nuclear envelope with the help of the SUN1-KASH5 complex to move chromosomes and facilitate homolog pairing during meiosis. (C) Domain diagram of human KASH5 FL and domain deletion constructs used in this study. (D) Pull down of purified proteins on glutathione (GSH) beads. Indicated glutathione S-trasferase (GST)-tagged KASH5 constructs were incubated with dynein LIC and pulled down on GSH-beads followed by visualization on an SDS-PAGE using Coomassie-blue staining. Number of replicates, n=2. (E and F) UV280 absorbance profile (top) and Coomassie-blue staining analysis (bottom) of size-exclusion chromatography (SEC) of KASH5-NCC alone, LIC alone, and a mixture of KASH5-NCC and LIC using either wild type LIC (E) or the LIC F447A/F448A double mutant (F). n represents number of replicates for each SEC run. n=3, 3, 2, 2, and 2 for KASH5-NCC alone, LIC alone, LIC F447A/F448A alone, KASH5-NCC with LIC, and KASH5-NCC with LIC F447A/F448A, respectively. Figure 1—source data 1 Unedited SDS-PAGE gels relating to Figure 1D–F. https://cdn.elifesciences.org/articles/78201/elife-78201-fig1-data1-v2.pdf Download elife-78201-fig1-data1-v2.pdf As the primary motor used to traffic all cellular cargo toward the microtubule minus-end (typically clustered near the nucleus), dynein must traffic hundreds of types of cargo, including, but not limited to, membrane-bound vesicles, organelles, RNA-protein complexes, lipid droplets, and viruses (Reck-Peterson et al., 2018). How cargo specificity is defined in different biological contexts and for myriad cargo is not well understood, but dynein activating adaptors likely play an important role in this regard. Perhaps some of the most unique cargos trafficked by dynein are chromosomes in prophase I of meiosis. Meiosis involves one round of DNA replication followed by two rounds of cell division to generate haploid gametes (Zickler and Kleckner, 2015). Crossover, which is the product of homologous recombination between homologous chromosomes in prophase I, ensures proper segregation of chromosomes and promotes genetic diversity. Dynein drives chromosomal movements to facilitate homolog pairing and meiotic recombination (Wynne et al., 2012; Lee et al., 2015). These dynein-driven chromosomal movements increase the search space that homologous chromosomes explore in the nucleus to untangle mispaired nonhomologous chromosomes and facilitate proper pairing between homologous chromosomes (Figure 1B; Shibuya and Watanabe, 2014c). The nuclear envelope (NE) is intact during prophase I, implying that dynein is separated from its cargo by two lipid bilayers (the inner nuclear membrane [INM] and outer nuclear membrane [ONM]) and the perinuclear space between them (Figure 1B). In this context, dynein tethers to chromosomes via the highly conserved linker of nucleoskeleton and cytoskeleton (LINC) complex at the NE (Meier, 2016). The LINC complex consists of a SUN protein and a KASH protein that span the INM and ONM, respectively and bind each other in the perinuclear space (Hieda, 2017; Figure 1B). Mammals encode multiple SUN and KASH proteins that bind to their LINC partner using highly conserved SUN and KASH domains, respectively. Cytosolic dynein immunoprecipitates with KASH5, a KASH protein that is expressed exclusively in prophase I of meiosis and essential for mouse fertility (Morimoto et al., 2012; Horn et al., 2013). A short ~20 amino acid KASH peptide at the very C-terminus of KASH5 interacts with the SUN protein, SUN1, in the perinuclear space (Figure 1B and domain diagram of KASH5-FL in Figure 1C). Inside the nucleus, SUN1 binds telomeres at the ends of chromosomes, completing the attachment between dynein and the chromosomal cargo (Ding et al., 2007). Despite being a process essential to meiosis progression and fertility, how dynein is activated to drive these movements across the NE remains unknown. KASH5 protein consists of an N-terminal EF-hand pair followed by a CC domain, reminiscent of the domain architecture of EF-hand pair-containing dynein activating adaptors (Figure 1C and Figure 1—figure supplement 1B). Here, we tested the hypothesis that KASH5 is a dynein activating adaptor that activates dynein specifically during prophase I of meiosis. We show that KASH5 binds dynein LIC in a manner comparable to other known dynein activating adaptors. Using human cell lysates and purified recombinant protein, we demonstrate that KASH5 is a bona fide activating adaptor that converts dynein and dynactin into a processive motile complex. We identify specific residues in the KASH5 EF-hand pair that mediate the interaction with dynein LIC and demonstrate that disruption of the KASH5-LIC interaction impairs dynein recruitment to the NE of both KASH5-transfected HeLa cells and mouse spermatocytes undergoing meiosis. Our findings establish KASH5 as a novel EF-hand pair-type activating adaptor, the first identified transmembrane (TM) dynein activating adaptor, and the dynein activator responsible for moving chromosomes in meiotic prophase I. Results KASH5 binds dynein’s LIC using a mechanism conserved in other activating adaptors To determine the regions of KASH5 that promote binding to dynein and dynactin, we generated 3×-FLAG-tagged KASH5-FL (wild type [WT]) and domain deletion constructs lacking one or more of its domains, namely the TM domain, EF-hand pair-containing N-terminal domain (N), CC domain, and the unstructured region between its CC and TM (Figure 1C and Methods). Deletion of TM greatly increased the level of soluble KASH5 compared to KASH5 FL (compare lanes five and six in Input, Figure 1—figure supplement 1C), allowing us to define a stable, TM-less construct of KASH5 for subsequent biochemical analysis in buffers lacking harsh detergents. Indeed, KASH5-ΔTM co-immunoprecipitates with dynein’s HC and LIC and dynactin’s p150 subunit (Figure 1—figure supplement 1C), consistent with the association of the dynein transport machinery and KASH5 reported previously (Horn et al., 2013). KASH5-NCC, and to a lesser extent, KASH5-N, retained association with dynein and dynactin, suggesting that KASH5-NCC recapitulates the entire dynein-interaction interface of KASH5 as suggested previously (Horn et al., 2013). Co-immunoprecipitation (Co-IP) data strongly suggest a direct interaction between KASH5 and dynein, but this has not been demonstrated with purified components. We hypothesized that the KASH5 N-terminus, like in other dynein activating adaptors, directly binds dynein LIC (Figure 1A and C). We performed glutathione (GSH) bead pull-down of purified GST-fusions of human KASH5 N, CC, and NCC with full-length, untagged human dynein LIC1 (hereafter referred to as LIC). Consistent with co-IP analysis, KASH5 N and NCC, but not CC, bind directly to LIC (Figure 1D). Together, our results show that KASH5 directly binds the dynein transport machinery using its EF-hand pair-containing N domain and CC domain. To determine if KASH5 binds to dynein like other EF-hand pair activating adaptors (Figure 1—figure supplement 1B), we incubated purified, untagged KASH5-NCC and LIC (Figure 1—figure supplement 1D) and performed size-exclusion chromatography (SEC). Indeed, KASH5-NCC co-eluted with LIC in SEC as a peak distinguishable from KASH5-NCC or LIC alone (Figure 1E). Dynein activating adaptors tested to date bind dynein’s LIC via a well-conserved pair of consecutive phenylalanine residues in LIC’s helix 1 (F447 and F448 in the human LIC1 sequence) (Figure 1—figure supplement 1E; Lee et al., 2018; Lee et al., 2020; Celestino et al., 2019). LICF447A/F448A, which has both key phenylalanine residues mutated to alanine, disrupts binding to BicD2, Hook3, and CRACR2a (Lee et al., 2018; Lee et al., 2020). KASH5-NCC and LICF447A/F448A failed to co-elute (Figure 1F), consistent with KASH5 using a structural mechanism shared with other dynein activating adaptors to bind dynein LIC. KASH5 binds dynein LIC with an affinity comparable to other activating adaptors but with unique stoichiometry To determine the stoichiometry of the KASH5-LIC interaction, we used SEC coupled to multi-angle light scattering (MALS) to determine the molecular weight of the NCC:LIC complex. NCC:LIC forms a homogenous complex with an experimentally determined molecular weight of 139 kDa, consistent with a 2:1 NCC:LIC stoichiometry (theoretical molecular weight (MW) of 133.5 kDa) (Figure 2A). The 2:1 stoichiometry was also observed at threefold higher protein concentrations (~10-fold greater than the Kd determined with isothermal titration calorimetry [ITC]; see below) and despite adding a twofold molar excess of LIC over KASH5-NCC, suggesting that it is not a result of incomplete binding or dissociation of the complex (Figure 2—figure supplement 1A). SEC-MALS analysis of NCC (alone) is consistent with a homodimer, defining a quaternary structure for KASH5 that is conserved among other dynein activating adaptors. Figure 2 with 1 supplement see all Download asset Open asset KASH5 directly binds dynein light-intermediate chain (LIC) with a 2:1 stoichiometry. (A) Size-exclusion chromatography (SEC)-multi-angle light scattering (MALS) analysis of KASH5-NCC alone, LIC alone, and the KASH5-NCC-dynein-LIC complex (SEC profile data same as in Figure 1E) showing that KASH5-NCC is homodimeric while the NCC-LIC complex adopts a 2:1 stoichiometry. n represents number of replicates for each SEC-MALS run. n=2 each for KASH5-NCC alone, LIC alone, and KASH5-NCC with LIC. (B and C) SEC-MALS analysis of KASH5-N alone, LIC alone, and a mixture of KASH5-N and LIC using either untagged KASH5-N (B) or GST-tagged KASH5-N (C). Coomassie-blue stained SDS-PAGE analysis for the indicated KASH5-LIC mixtures is shown below the SEC-MALS profile. n represents number of replicates for each SEC-MALS run. n=1, 1, 3, and 1 for KASH5-N, GST-KASH5-N, KASH5-N with LIC, and GST-KASH5 with LIC, respectively. (D) Isothermal titration calorimetry (ITC) analysis of KASH5-NCC with the dynein LIC433–458 peptide containing F447 and F448. Mean and SE of the mean of the dissociation constant (Kd) are indicated for a biological duplicate. Figure 2—source data 1 Unedited SDS-PAGE gels relating to Figure 2B and C. https://cdn.elifesciences.org/articles/78201/elife-78201-fig2-data1-v2.pdf Download elife-78201-fig2-data1-v2.pdf The 2:1 stoichiometry suggests that CC dimerization allows two KASH5’s EF-hand pairs to bind a single copy of LIC. To further explore the role that KASH5 dimerization plays in LIC binding, we performed SEC-MALS with two different constructs of KASH5-N: GST-KASH5-N and untagged KASH5-N. Both constructs contain KASH5’s EF-hand pair, but only GST-KASH5-N would dimerize through its GST tag. SEC-MALS confirmed that GST-KASH5-N is a dimer and untagged KASH5-N is a monomer (Figure 2B and C). Next, we incubated each construct with purified LIC. Both GST-KASH5-N and KASH5-N were capable of binding LIC (Figure 2B and C). The experimentally determined mass of the complex formed between KASH5-N and LIC was ~10 kDa lower than the theoretical mass for a 1:1 complex (Figure 2B). This lower apparent molecular weight could indicate either partial dissociation of the complex during SEC analysis or partial overlap between the peak of the 1:1 complex and excess LIC in the mixture. To ensure that the observed stoichiometry was not limited by the concentration of the proteins, we repeated the SEC-MALS analysis of the LIC-KASH5-N mixture at threefold higher concentration of each protein (Figure 2—figure supplement 1B). Increased protein concentration did not alter the ~1:1 stoichiometry. In contrast to KASH5-N and consistent with what we observed with KASH5-NCC, GST-KASH5-N bound with 2:1 stoichiometry, confirming that dimerization of an EF-hand is not accompanied by association of a second LIC polypeptide (Figure 2B and C). Together, these data suggest that dimeric KASH5 EF-hands bind a single copy of LIC. We employed ITC to determine the affinity between KASH5 and LIC. We titrated a LIC peptide that corresponds to the sequence in helix 1 (amino acids (aa) 433–458) into purified KASH5-NCC (Lee et al., 2018; Lee et al., 2020; Celestino et al., 2019). We observed that KASH5-NCC bound the LIC peptide with a dissociation constant Kd = 4.3 μM, which is comparable to the LIC-binding affinities of other activating adaptors (Lee et al., 2018; Lee et al., 2020; Figure 2D). Some EF-hand pair activating adaptors bind dynein in a Ca2+-dependent fashion (Lee et al., 2020; Wang et al., 2019). To test if Ca2+ regulates KASH5 binding to dynein’s LIC, we performed ITC in the presence of CaCl2 or EGTA to chelate any Ca2+ that co-purified with KASH5-NCC. Dissociation constants were very similar for each condition, suggesting that Ca2+ does not regulate the NCC-LIC interaction (Figure 2—figure supplement 1C, D). Finally, we used ITC to test if the EF-hand pair of KASH5 binds Ca2+. Titration of Ca2+ into KASH5-NCC did not result in a change of enthalpy, suggesting that the KASH5 EF-hand pair does not bind Ca2+ (Figure 2—figure supplement 1E). KASH5 activates dynein motility To determine if KASH5 is a dynein activating adaptor, we appended a C-terminal FLAG and an N-terminal green fluorescent protein (GFP) tag to KASH5-ΔTM in a pcDNA3 backbone for mammalian cell expression (Redwine et al., 2017). We expressed these constructs in HEK 293T cells, immunoprecipitated the cellular lysates on anti-FLAG resin, and determined if the GFP-tagged KASH5 constructs displayed processive motility on microtubules using total internal reflection fluorescence microscopy (TIRF) (Figure 3A). Observed processive motility would indicate that KASH5 co-precipitated the activated dynein complex. Robust motility was observed with the positive control (GFP-BicD2-FLAG), while very little motility was observed with the negative control (GFP-FLAG) (Figure 3B and C and Figure 3—figure supplement 1A-C). GFP-KASH5-ΔTM-FLAG and GFP-KASH5-NCC-FLAG displayed processive movements on microtubules (Figure 3B and C). We quantified the velocity, run length, percentage of processive events, and the landing rate of dynein on microtubules in the presence of BicD2 and the KASH5 constructs. Dynein associated with GFP-KASH5-ΔTM-FLAG, GFP-KASH5-NCC-FLAG, and GFP-BicD2-FLAG all moved at velocities ~0.5 µm/s (Figure 3C). Interestingly, BicD2 immunoprecipitates displayed slightly enhanced run lengths, processivity, and landing rate compared to either KASH5 construct (Figure 3—figure supplement 1A-C). Although the significance of the differences between BicD2- and KASH5-dependent dynein motility is unclear, it may reflect the different biological contexts the two proteins function in. Additionally, KASH5-ΔTM performed slightly better than KASH5-NCC in the immunoprecipitation (IP)-TIRF experiments, suggesting that regions downstream of the KASH5 CC may contribute to dynein motility (Figure 3C and Figure 3—figure supplement 1A-C). Indeed, C-terminal to the CC in KASH5 is a putative Spindly motif, which in other activating adaptors binds to the pointed end of dynactin and facilitates dynein motility (Gama et al., 2017; Figure 3—figure supplement 2A-C). Altogether, the IP-TIRF data support the hypothesis that KASH5 is a newly identified dynein activating adaptor. Figure 3 with 2 supplements see all Download asset Open asset KASH5 activates dynein. (A) Schematic for the IP-total internal reflection fluorescence (TIRF) assay to measure dynein motility of anti-FLAG immunoprecipitates of GFP-FLAG-tagged KASH5-ΔTM and BicD2. (B) Kymographs showing motility of immunoprecipitated complexes containing indicated GFP-FLAG-tagged constructs monitored by GFP fluorescence using TIRF microscopy. (C) Velocity of processive events from a total of two movies from two biological replicates (four movies analyzed in total). Each data point represents an individual processive event; n=694, 615, and 484 for BicD2, KASH5-ΔTM, and KASH5-NCC, respectively. Median and interquartile range shown. Significance determined from a Kruskal-Wallis test with Dunn’s multiple comparisons test. ns, not significant; **p≤0.01. (D) Schematic for the TIRF assay with purified proteins. (E and F) Kymographs of KASH5-dynein-dynactin complexes monitored by differentially fluorophore-labeled dynein and KASH5 in the absence (E) or presence (F) of Lis1. (G) Velocity of processive events from a total of two biological replicates (six movies analyzed in total). Each data point represents an individual processive event; n=398, 414, 174, and 146 for dynein and dynactin with KASH5, KASH5 +Lis1, Buffer, and Lis1, respectively. Median and interquartile range shown. Significance determined from a Kruskal-Wallis test with Dunn’s multiple comparisons test. ns, not significant; **p≤0.01; ****p≤0.0001. (H) Landing rate for the observed motile events with purified proteins and complexes from a total of two biological replicates (six movies analyzed in total). n values are derived from the average processive events/micron from all microtubules analyzed in a movie; n=6. Mean and SE of the mean shown. Significance determined from a Brown-Forsythe and Welch ANOVA test with Dunnett’s T3 multiple comparison test. *p≤0.05; **p≤0.01; ***p≤0.001. Figure 3—source data 1 Numerical source data relating to Figure 3C, G and H. https://cdn.elifesciences.org/articles/78201/elife-78201-fig3-data1-v2.xlsx Download elife-78201-fig3-data1-v2.xlsx Activating adaptors are also able to promote dynein motility in a fully reconstituted system (McKenney et al., 2014; Schlager et al., 2014). To test if KASH5 can activate dynein with only purified components, we purified SNAP-tagged recombinant human dynein from an insect cell expression system and human dynactin from HEK 293T cells stably expressing FLAG-p62, as we have described previously (Htet et al., 2020), and recombinant Halo-KASH5-ΔTM from Escherichia coli (Figure 3—figure supplement 1D and see Methods). We selected KASH5-ΔTM over Halo-tagged KASH5-NCC in this analysis because the former encompasses all soluble domains of KASH5, including the putative Spindly motif (Figure 1C and Figure 3—figure supplement 2A-C) and showed improved dynein motility in the IP-TIRF analysis (Figure 3C and Figure 3—figure supplement 1A-C). The purified proteins were assembled in the presence of ATP and an oxygen scavenger system and imaged as they associated with microtubules using TIRF (Figure 3D). Dynein was labeled with either tetramethylrhodamine (TMR) or Alexa-647 via the SNAP tag in all experiments. In experiments where dynein was labeled with Alexa-647, we labeled KASH5 with TMR via the Halo tag. Processive dynein motility was observed in the presence of KASH5-ΔTM (Figure 3E, G and H and Figure 3—figure supplement 1E, F) but not in its absence (Figure 3G and H, and Figure 3—figure supplement 1E-G). Moving Alexa-647 dynein was almost always colocalized with TMR-labeled KASH5-ΔTM (Figure 3E). Together, the TIRF motility data qualify KASH5 as a newly identified dynein activating adaptor, making it one of about a dozen characterized members of this family of dynein regulators. The velocity of the processive events recorded with purified dynein, dynactin, and KASH5 was significantly slower than in the IP-TIRF motility experiments (median velocity of 0.320 µm/s with purified KASH5-ΔTM versus 0.538 µm/s with dynein immunoprecipitated by KASH5-ΔTM) (Figure 3C and G). We reasoned that the reconstitutions with purified proteins were missing a dynein regulatory factor that promotes activity. Lis1 increases the velocity of dynein-dynactin-activating adaptor motility by promoting dynein conformations that drive association with dynactin and an activating adaptor (Htet et al., 2020; Elshenawy et al., 2020; Gutierrez et al., 2017; Markus et al., 2020; Baumbach et al., 2017; Qiu et al., 2019). Inclusion of purified Lis1 in the in vitro reconstituted TIRF experiments resulted in a significant increase in the velocity of motile events (Figure 3F). In fact, the velocity of dynein-dynactin-KASH5 processive events in the presence of Lis1 is comparable to that observed in the IP-TIRF motility experiments (median velocity of 0.577 µm/s). Consistent with observations made with other activating adaptors, Lis1 also increased the landing rate, run length, and percent processivity of dynein-dynactin-KASH5 complexes (Htet et al., 2020; Elshenawy et al., 2020; Gutierrez et al., 2017; Baumbach et al., 2017; Figure 3H and Figure 3—figure supplement 1E, F). These data suggest that Lis1 likely drives the association of dynein-dynactin-KASH5 complexes during meiosis, as has been demonstrated for other activators in the cell (Htet et al., 2020; Qiu et al., 2019; Splinter et al., 2012). Specific residues of the KASH5 EF-hand that mediate the interaction with LIC and dynein activation To identify residues in KASH5 that mediate the interaction with dynein, we generated a homology model of KASH5’s EF-hand pair bound to LIC’s helix 1 peptide harboring the FF motif based on the structure of CRACR2a EF-hand pair bound to this peptide (Lee et al., 2020) (PDB: 6PSD) (Figure 4A–C; see Methods). Based on this model, we identified nine conserved amino acids in the KASH5 EF-hand pair that surround the hydrophobic surface of LIC harboring F447 and F448 (Figure 4C): I36, T40, Y60, V64, R73, L77, F97, L98, and M101 (Figure 4A and C). We substituted each of these residues with aspartate residues (D) in the GFP-KASH5-ΔTM-FLAG background to disrupt the putative hydrophobic interface, expressed each construct in HEK 293T cells, and performed both co-IP and IP-TIRF motility experiments as described in Figure 1—figure supplement 1C and Figure 3A–C, respectively. All the KASH5 EF-hand mutants tested showed impaired binding to dynein, albeit to varying extents (Figure 4D and E). All the mutants co-precipitated at least some motile dynein, as visualized by GFP-KASH5-ΔTM-FLAG in TIRF (Figure 4F). Most of the mutants displayed motile events with comparable velocity, processivity, landing rate, and run length as WT KASH5-ΔTM (Figure 4F–H and Figure 4—figure supplement 1A, B). However, three of the mutants, KASH5-ΔTML77D, KASH5-ΔTMF97D, and KASH5-ΔTMM101D displayed severely impaired dynein association and very little processive motility as visualized via TIRF (Figure 4D–G and Figure 4—figure supplement 1A). Figure 4 with 1 supplement see all Download asset Open asset Mutations in the EF-hand of KASH5 abrogate association with active dynein complexes in the cell. (A) Alignment of the EF-hand pairs of KASH5 and previously characterized dynein activator CRACR2A showing the putative Ca2+-binding loop. Conserved and similar residues are shown in red and orange, respectively. Residues that were mutated are indicated with green asterisks. (B and C) Homology model of the KASH5 EF-hand pair bound to the dynein light-intermediate chain (LIC) helix 1 with LIC F447/448 shown in orange and the putative KASH5 EF-hand residues mutated in this study that form a binding pocket around the helix shown in green. (D) Anti-FLAG co-immunoprecipitation (co-IP) analysis of HEK 293T cell lysates containing transiently transfected FLAG-tagged KASH5-ΔTM and indicated mutants; the three mutants with the most drastic binding defect highlighted in red. (E) The immunoprecipitation signal for dynein heavy chain (HC) was quantified for each of the samples represented in panel D by dividing the western blot band intensity of HC by that of the FLAG-KASH5-ΔTM band in that lane of the IP fraction. Mean and SE of the mean from a triplicate set of experiments are shown. (F) Representative kymographs from motility experiment for each mutant. (G) Percent processive events for each mutant. Mean and SE of the mean shown. Percent of processive events from a total of two movies from two biological replicates. n values are derived from the average percent processive events from all microtubules analyzed in a movie; n=4. Significance determined from a Brown-Forsythe and Welch ANOVA test with Dunnett’s T3 multiple comparison test. **p≤0.01; ***p≤0.001. Only pairwise comparisons with a p-value p≤0.05 are shown. (H) Velocities of processive events for each mutant. Median and interquartile shown. The distribution of velocities for each mutant was compared to KASH5-ΔTM with a Kruskal-Wallis test with Dunn’s multiple comparisons test. No pairwise comparison was significantly different. For most mutants, data was obtain
Dynein is the primary minus-end-directed microtubule motor protein. To achieve activation, dynein binds to the dynactin complex and an adaptor to form the "activated dynein complex." The protein Lis1 aids activation by binding to dynein and promoting its association with dynactin and the adaptor. Ndel1 and its paralog Nde1 are dynein- and Lis1-binding proteins that help control dynein localization within the cell. Cell-based assays suggest that Ndel1-Nde1 also work with Lis1 to promote dynein activation, although the underlying mechanism is unclear. Using purified proteins and quantitative binding assays, here we found that the C-terminal region of Ndel1 contributes to dynein binding and negatively regulates binding to Lis1. Using single-molecule imaging and protein biochemistry, we observed that Ndel1 inhibits dynein activation in two distinct ways. First, Ndel1 disfavors the formation of the activated dynein complex. We found that phosphomimetic mutations in the C-terminal domain of Ndel1 increase its ability to inhibit dynein-dynactin-adaptor complex formation. Second, we observed that Ndel1 interacts with dynein and Lis1 simultaneously and sequesters Lis1 away from its dynein-binding site. In doing this, Ndel1 prevents Lis1-mediated dynein activation. Together, our work suggests that in vitro, Ndel1 is a negative regulator of dynein activation, which contrasts with cellular studies where Ndel1 promotes dynein activity. To reconcile our findings with previous work, we posit that Ndel1 functions to scaffold dynein and Lis1 together while keeping dynein in an inhibited state. We speculate that Ndel1 release can be triggered in cellular settings to allow for timed dynein activation.
Dynein harnesses ATP hydrolysis to move cargo on microtubules in multiple biological contexts. Dynein meets a unique challenge in meiosis by moving chromosomes tethered to the nuclear envelope to facilitate homolog pairing essential for gametogenesis. Though processive dynein motility requires binding to an activating adaptor, the identity of the activating adaptor required for dynein to move meiotic chromosomes is unknown. We show that the meiosis-specific nuclear-envelope protein KASH5 is a dynein activating adaptor: KASH5 directly binds dynein using a mechanism conserved among activating adaptors and converts dynein into a processive motor. We map the dynein-binding surface of KASH5, identifying mutations that abrogate dynein binding in vitro and disrupt recruitment of the dynein machinery to the nuclear envelope in cultured cells and mouse spermatocytes in vivo. Our study identifies KASH5 as the first transmembrane dynein activating adaptor and provides molecular insights into how it activates dynein during meiosis.
Cytoplasmic dynein-1 (dynein) is the primary microtubule minus-end directed molecular motor in most eukaryotes. As such, dynein has a broad array of functions that range from driving retrograde-directed cargo trafficking to forming and focusing the mitotic spindle. Dynein does not function in isolation. Instead, a network of regulatory proteins mediate dynein’s interaction with cargo and modulate dynein’s ability to engage with and move on the microtubule track. A flurry of research over the past decade has revealed the function and mechanism of many of dynein’s regulators, including Lis1, dynactin, and a family of proteins called activating adaptors. However, the mechanistic details of two of dynein’s important binding partners, the paralogs Nde1 and Ndel1, have remained elusive. While genetic studies have firmly established Nde1/Ndel1 as players in the dynein transport pathway, the nature of how they regulate dynein activity is unknown. In this review, we will compare Ndel1 and Nde1 with a focus on discerning if the proteins are functionally redundant, outline the data that places Nde1/Ndel1 in the dynein transport pathway, and explore the literature supporting and opposing the predominant hypothesis about Nde1/Ndel1’s molecular effect on dynein activity.
Dynein is the primary minus-end-directed microtubule motor [1]. To achieve activation, dynein binds to the dynactin complex and an adaptor to form the "activated dynein complex" [2, 3]. The protein Lis1 aids activation by binding to dynein and promoting its association with dynactin and adaptor [4, 5]. Ndel1 and its orthologue Nde1 are dynein and Lis1 binding proteins that help control where dynein localizes within the cell [6]. Cell-based assays suggest that Ndel1/Nde1 also work with Lis1 to promote dynein activation, although the underlying mechanism is unclear [6]. Using purified proteins and quantitative binding assays, we found that Ndel1's C-terminal region contributes to binding to dynein and negatively regulates binding to Lis1. Using single-molecule imaging and protein biochemistry, we observed that Ndel1 inhibits dynein activation in two distinct ways. First, Ndel1 disfavors the formation of the activated dynein complex. We found that phosphomimetic mutations in Ndel1's C-terminal domain increase its ability to inhibit dynein-dynactin-adaptor complex formation. Second, we observed that Ndel1 interacts with dynein and Lis1 simultaneously and sequesters Lis1 away from its dynein binding site. In doing this, Ndel1 prevents Lis1-mediated dynein activation. Our work suggests that
