The molecular motor dynein is essential for mitotic spindle orientation, which defines the axis of cell division. The light intermediate chain subunits, LIC1 and LIC2, define biochemically and functionally distinct vertebrate dynein complexes, with LIC2-dynein playing a crucial role in ensuring spindle orientation. We reveal a novel, mitosis-specific interaction of LIC2-dynein with the cortical actin-bundling protein transgelin-2. Transgelin-2 is required for maintaining proper spindle length, equatorial metaphase chromosome alignment, spindle orientation and timely anaphase onset. We show that transgelin-2 stabilizes the cortical recruitment of LGN-NuMA, which together with dynein is required for spindle orientation. The opposing actions of transgelin-2 and LIC2-dynein maintain optimal cortical levels of LGN-NuMA. In addition, we show that the highly conserved serine 194 phosphorylation of LIC2 is required for proper spindle orientation, by maintaining mitotic centrosome integrity to ensure optimal astral microtubule nucleation. The work reveals two specific mechanisms through which LIC2-dynein regulates mitotic spindle orientation; namely, through a new interactor transgelin-2, which is required for engagement of LGN-NuMA with the actin cortex, and through mitotic phosphoregulation of LIC2 to control microtubule nucleation from the poles.

This article has an associated First Person interview with the first author of the paper.

The mitotic spindle consists of the two centrosomes (spindle poles) that nucleate kinetochore-directed microtubules, interpolar microtubules and the astral microtubules that engage with the cortex to anchor the spindle, thus positioning and orienting it properly within the cell (Glotzer, 2009; Kulukian and Fuchs, 2013). Proper positioning and orientation of the mitotic spindle decides the plane of cell division within a tissue and, hence, governs several fundamental physiological functions impacting embryonic and tissue development, body axis elongation and asymmetric division of stem cells leading to self-renewal or differentiation (Morin and Bellaiche, 2011; Noatynska et al., 2012; Juschke et al., 2014). The mechanisms of spindle orientation are important to understand from both a fundamental standpoint and a therapeutic angle, as spindle misorientation leads to several disease conditions (Quyn et al., 2010; Pease and Tirnauer, 2011).

Cytoplasmic dynein (‘dynein’) is a conserved microtubule-based, multisubunit motor protein complex that transports diverse cargoes primarily towards its minus ends, and is therefore required for crucial functions during interphase and mitosis (Niclas et al., 1996; Ma et al., 1999). One essential mitotic function of dynein is to help position and orient the mitotic spindle correctly (Gönczy et al., 1999). Dynein transports microtubule nucleating proteins such as γ-tubulin (Young et al., 2000) to both spindle poles to enable optimal astral microtubule nucleation (Doxsey et al., 1994; Young et al., 2000; Chen et al., 2014). In addition, cortically localized dynein captures astral microtubules by engaging with a conserved cortical protein complex consisting of Gαi-LGN-NuMA and simultaneously generating a pulling force on the spindle towards the cortex, thus anchoring the spindle to the cell boundary (Peyre et al., 2011; Kiyomitsu and Cheeseman, 2012; Kotak and Gönczy, 2013; Zheng et al., 2013). Overall, engagement of the plasma membrane with this force generator complex consisting of Gαi-LGN-NuMA-dynein is a conserved and vital event required for orienting the spindle in multiple systems (Kiyomitsu and Cheeseman, 2012; di Pietro et al., 2016).

The cortical actin meshwork just under the plasma membrane is another key cellular component thought to be required for spindle orientation (Theesfeld et al., 1999; Zheng et al., 2013; Zhu et al., 2013; Machicoane et al., 2014; Carminati et al., 2016; Kschonsak and Hoffmann, 2018). This hypothesis has gained further traction with recent reports describing the specific roles of cortical actin-binding proteins such as the ezrin-radixin-moesin (ERM) proteins, afadin, MISP and CAP-Z as key contributors to spindle orientation (Zhu et al., 2013; Machicoane et al., 2014; di Pietro et al., 2016, 2017; Kschonsak and Hoffmann, 2018), thus establishing the indispensable role of the actin cortex and its interacting proteins in orienting the mitotic spindle. It is therefore plausible that other actin-cortex binding proteins also play important roles in maintaining mitotic spindle orientation.

Cytoplasmic dynein exists in two mutually exclusive complexes in the cell, containing either one or the other of the light intermediate chain homologs, LIC1 or LIC2 (Tynan et al., 2000). LIC2-dynein plays a crucial role in maintaining spindle orientation (Mahale et al., 2016a). In this study, we sought to understand the molecular mechanisms that underlie the function of LIC2-dynein in mitotic spindle orientation. We identified the actin bundling protein transgelin-2 (Tagln2) as a novel interactor of mitotic LIC2-dynein and a novel contributor to spindle orientation. Tagln2 is a small (22 kDa) actin-binding protein that is associated with modulation of actin dynamics and membrane ruffling through regulation of actin cytoskeletal stability and organization, and is known to have functions in the immune system (Shapland et al., 1993; Na et al., 2015, 2016; Kim et al., 2017). However, the function of Tagln2 in mitosis has hardly been explored, making this the first report to our knowledge to implicate it in chromosome congression, maintenance of spindle length and spindle orientation. We show that Tagln2 stabilizes the cortical recruitment of the orientation controlling factors LGN and NuMA, conserved proteins that collaborate with dynein to capture astral microtubules and help anchor the spindle to the cortex. We further show that the conserved phosphorylation of LIC2 at the serine 194 (S194) residue is required for maintaining spindle pole integrity and optimal microtubule nucleation, thus helping to maintain spindle orientation. This study delineates two key mechanisms through which the LIC2 subunit of dynein ensures proper spindle orientation by ensuring proper centrosome integrity and astral microtubule nucleation: one at the cell cortex through Tagln2 and the other at the spindle pole through phosphoregulation.

Transgelin-2 interacts with dynein and is required for mitotic progression

We had shown earlier that LIC1-dynein and LIC2-dynein unevenly distribute their mitotic functions, with LIC2-dynein being the major contributor to spindle orientation (Mahale et al., 2016a). Given that LIC1 and LIC2 largely occupy mutually exclusive dynein complexes (Tynan et al., 2000), and also that the LICs are major determinants of the cargo-binding diversity of dynein, we predicted that the differential functions of both LICs would be dictated by binding to distinct interaction partners. In an effort to identify other proteins that impart specificity to LIC2-dynein in mediating spindle orientation, we determined the mitotic interactome of LIC2 using a U2OS (human osteosarcoma) cell line stably expressing a multifunctional tandem affinity purification (MTAP)-tagged LIC2 (Mahale et al., 2016a,b). Among the proteins reproducibly coming up as LIC2 interactors (Tables S1-S7), we biochemically validated the actin-bundling protein transgelin 2 (Tagln2) as a bona fide novel interactor of mitotic LIC2. Interestingly, Tagln2 interacted only with LIC2-dynein, but not with LIC1-dynein, exclusively in mitotic but not interphase lysates (Fig. 1A). We observed that the LIC2-Tagln2 interaction was abrogated upon dynein heavy chain (DHC) depletion (Fig. 1B), confirming that the interaction of Tagln2 is with the LIC2-dynein complex and not with free LIC2. This result and the presence of DHC in the affinity precipitation experiments (Fig. 1A) confirmed that Tagln2 is a novel mitosis-specific interactor of the dynein motor.

Fig. 1.

Transgelin 2 is a novel dynein interactor required for mitotic progression. (A) Immunoprecipitation of empty vector eMTAP (control), human LIC1-MTAP and human LIC2-MTAP at the indicated cell cycle stages followed by immunoblotting (IB) with the respective antibodies, as indicated. (B) Transgelin-2 does not interact with dynein-LIC2 upon depletion of dynein heavy chain (DHC). Top: siRNA-mediated depletion of DHC. Bottom: immunoprecipitation of stably expressing human LIC2-MTAP from mitotic lysates probed with the indicated antibodies. (C) Representative fluorescence images of asynchronous HeLa cells treated with GFP (control) or Tagln2 siRNA SMARTpool. Metaphase (white arrows) and cytokinetic (yellow arrows) cells are indicated. (D,E) Fraction of cells delayed in metaphase (D) and cytokinesis (E) upon GFP- (control) and Tagln2-specific siRNA treatment (n=3 experiments, with at least 500 cells counted per experiment). (F) Stills from time-lapse videos of control or Tagln2 siRNA SMARTpool-treated HeLa cells showing a delay in the metaphase-to-anaphase transition as well as in cytokinesis (green, GFP-tubulin; red, mcherry-Histone 2B). (G,H) Scatter plots for the data shown in F, for NEB to anaphase (G) and anaphase to abscission (H). (I,J) Fraction of cells delayed in metaphase for more than 80 min following NEB (I) and in cytokinesis for more than 120 min following anaphase onset (J). n=3 experiments, with 30 metaphase (I) or cytokinetic (J) cells counted per experiment. Error bars represent mean±s.d. **P<0.01, ***P<0.001, ****P<0.0001. Scale bars: 100 µm (C).

Fig. 1.

Transgelin 2 is a novel dynein interactor required for mitotic progression. (A) Immunoprecipitation of empty vector eMTAP (control), human LIC1-MTAP and human LIC2-MTAP at the indicated cell cycle stages followed by immunoblotting (IB) with the respective antibodies, as indicated. (B) Transgelin-2 does not interact with dynein-LIC2 upon depletion of dynein heavy chain (DHC). Top: siRNA-mediated depletion of DHC. Bottom: immunoprecipitation of stably expressing human LIC2-MTAP from mitotic lysates probed with the indicated antibodies. (C) Representative fluorescence images of asynchronous HeLa cells treated with GFP (control) or Tagln2 siRNA SMARTpool. Metaphase (white arrows) and cytokinetic (yellow arrows) cells are indicated. (D,E) Fraction of cells delayed in metaphase (D) and cytokinesis (E) upon GFP- (control) and Tagln2-specific siRNA treatment (n=3 experiments, with at least 500 cells counted per experiment). (F) Stills from time-lapse videos of control or Tagln2 siRNA SMARTpool-treated HeLa cells showing a delay in the metaphase-to-anaphase transition as well as in cytokinesis (green, GFP-tubulin; red, mcherry-Histone 2B). (G,H) Scatter plots for the data shown in F, for NEB to anaphase (G) and anaphase to abscission (H). (I,J) Fraction of cells delayed in metaphase for more than 80 min following NEB (I) and in cytokinesis for more than 120 min following anaphase onset (J). n=3 experiments, with 30 metaphase (I) or cytokinetic (J) cells counted per experiment. Error bars represent mean±s.d. **P<0.01, ***P<0.001, ****P<0.0001. Scale bars: 100 µm (C).

To assess function, we depleted Tagln2 from cells using sequence-specific siRNA treatment [SMARTPool of four siRNAs, or two published individual Tagln2 siRNAs, T2A and T2B (Han et al., 2017)], as indicated in Figs S1–S3. Each of these siRNAs robustly depleted Tagln2 protein levels (Fig. S3A,B). Tagln2 depletion, followed by confocal immunofluorescence imaging of HeLa cells, revealed robust increases in the metaphase and cytokinetic indices (Fig. 1C-E; Fig. S2D,E), suggesting specific Tagln2-induced delays in both these stages of mitotic progression. Time-lapse, live cell confocal imaging in a HeLa cell line labeled with GFP-tagged α-tubulin (green, microtubules) and mCherry-tagged histone 2B (red, chromosomes) confirmed that Tagln2 depletion led to a time delay of over 2.5-fold in both metaphase to anaphase progression and in completion of cytokinesis (Fig. 1F-J; Movies 1, 2). Immunostaining and confocal imaging of both interphase and mitotic Hela cells showed localization of Tagln2 at regions of the actin cortex, which was lost upon Tagln2 depletion (Fig. S1A,D). Prominent cortical enrichment of Tagln2 was observed using anti-GFP immunofluorescence in a HeLa cell line stably expressing Tagln2-GFP (see next section). To ascertain whether Tagln2 depletion impacted cortical F-actin integrity, we compared it with latrunculin A (Lat A; an F-actin depolymerizing agent) treatment. Treatment of cells with LatA at 200 nM or higher drastically disrupted cortical F-actin integrity in mitosis (Fig. S1A), as also reported earlier (Kotak et al., 2014). However, the F-actin cortex looked normal after either 100 nM LatA treatment or robust Tagln2 depletion using siRNA (Fig. S1A). Tagln2 depletion did not appear to disturb significantly the levels (Fig. S1B) or gross organization (Fig. S1C) of the cortical actin meshwork, as measured using region-of-interest (ROI) quantification (for Fig. S1B) or linescan analysis of the actin cortex thickness (for Fig. S1C) from confocal micrographs. We also did not observe major differences in cell morphology after Tagln2 depletion in interphase (Fig. S1D,E). Taken together, the results revealed a novel, mitosis-specific interaction of Tagln2 with LIC2-dynein and a strong function in mitotic progression.

Transgelin-2 is required for chromosome congression, maintaining spindle length and spindle orientation

The pronounced mitotic delay caused by Tagln2 depletion prompted us to analyze the potential underlying mechanisms. We had observed imperfect metaphase plate formation upon Tagln2 depletion during the live cell imaging, which suggested defects in chromosome congression. Quantitative analysis of the live imaging data revealed that of approximately 40% mitotic cells that became delayed in mitosis (Fig. 1), about half showed chromosome miscongression (Fig. 2A,B). This phenotype was also observed in immunostained cells (Fig. 2C). Analysis of live movies (Movies 1, 2, 11) suggested that this defect was a genuine chromosome congression defect during metaphase plate formation, with no significant cohesion fatigue observed (Movie 11) (Stevens et al., 2011). Another effect that was visible was the presence of shortened mitotic spindles [calculated from three-dimensional (3D) confocal reconstructions of mitotic cells] upon Tagln2 depletion, with the average spindle length (interpolar distance) being significantly reduced (Fig. 2D,E). In light of these mitotic defects, we examined the mitotic localization of Tagln2.

Fig. 2.

Transgelin 2 is required for multiple mitotic functions including spindle orientation. (A) Stills from time-lapse videos of control or Tagln2 siRNA SMARTpool-treated HeLa cells showing chromosome congression defects (green, GFP-tubulin; red, mcherry-Histone 2B). White arrows show unaligned chromosomes. (B) Quantification of the fraction of mitotic cells showing chromosome congression defects upon control or Tagln2 siRNA SMARTpool-mediated depletion; n=3 independent experiments, with a minimum of 30 mitotic cells per experiment. (C,D) Representative images of mitotic Hela cells treated with control or Tagln2 siRNA SMARTpool and immunostained with the dyes and antibodies indicated to show chromosome congression defects (yellow arrows indicate unaligned chromosomes) and spindle length shortening (red line indicates the spindle axis used to measure spindle length). (E) Quantification of spindle lengths from metaphase cells. Each dot represents the spindle length from one mitotic cell. (F) Representative images showing the spindle angle (angle between the spindle axis and the substratum, enclosed by white lines) in 3D reconstructed images of confocal z-stacks of HeLa cells treated with GFP (control) or Tagln2 siRNA SMARTpool for 48 h and immunostained for γ tubulin (spindle poles, red), α tubulin (microtubules, green) and chromosomes (DAPI, blue). (G) Average spindle angle made by metaphase cells upon control and Tagln2 siRNA treatment. Individual dots represent values from individual cells. For D-G, n=3 independent experiments, with a minimum of 35 mitotic cells analysed per experiment. Error bars represent mean±s.d. (B) or mean±s.e.m. (E,G). ***P<0.001, ****P<0.0001. Scale bars: 8 µm (D), 5 µm (F).

Fig. 2.

Transgelin 2 is required for multiple mitotic functions including spindle orientation. (A) Stills from time-lapse videos of control or Tagln2 siRNA SMARTpool-treated HeLa cells showing chromosome congression defects (green, GFP-tubulin; red, mcherry-Histone 2B). White arrows show unaligned chromosomes. (B) Quantification of the fraction of mitotic cells showing chromosome congression defects upon control or Tagln2 siRNA SMARTpool-mediated depletion; n=3 independent experiments, with a minimum of 30 mitotic cells per experiment. (C,D) Representative images of mitotic Hela cells treated with control or Tagln2 siRNA SMARTpool and immunostained with the dyes and antibodies indicated to show chromosome congression defects (yellow arrows indicate unaligned chromosomes) and spindle length shortening (red line indicates the spindle axis used to measure spindle length). (E) Quantification of spindle lengths from metaphase cells. Each dot represents the spindle length from one mitotic cell. (F) Representative images showing the spindle angle (angle between the spindle axis and the substratum, enclosed by white lines) in 3D reconstructed images of confocal z-stacks of HeLa cells treated with GFP (control) or Tagln2 siRNA SMARTpool for 48 h and immunostained for γ tubulin (spindle poles, red), α tubulin (microtubules, green) and chromosomes (DAPI, blue). (G) Average spindle angle made by metaphase cells upon control and Tagln2 siRNA treatment. Individual dots represent values from individual cells. For D-G, n=3 independent experiments, with a minimum of 35 mitotic cells analysed per experiment. Error bars represent mean±s.d. (B) or mean±s.e.m. (E,G). ***P<0.001, ****P<0.0001. Scale bars: 8 µm (D), 5 µm (F).

Immunostaining revealed that Tagln2 was present both at the cortex and in the cytoplasm of mitotic cells; however, we did not observe any polar, spindle or chromosomal localization with the antibody used (Fig. S1A), perhaps due to low signals at these sites and high background staining of cytosolic Tagln2. We re-assessed the localization of Tagln2-GFP stably expressed in Hela cells, both in the presence and absence of anti-Tagln2 siRNA treatment to deplete endogenous Tagln2 levels. We only observed faint but clear localization of Tagln2-GFP (through GFP immunostaining) at metaphase kinetochores after Tagln2 siRNA treatment (Fig. S1E). This could be because of better incorporation of exogenous tagged Tagln2 at the kinetochore due to reduced competition from the endogenous protein (following siRNA-mediated depletion). Additionally, the absence of background Tagln2 immunostaining might have helped highlight this localization. Analogous methods have been shown to help visualize protein localization in cells when antibody staining was not successful (Rivera-Molina and Toomre, 2013; Ahmed et al., 2018; Sana et al., 2018). This method showed clear cortical and kinetochore localization of Tagln2 in mitotic cells, although we could still not observe it at spindle poles or on the spindle microtubules (Fig. S1E).

The cortical localization of Tagln2 prompted us to explore whether Tagln2 is required for spindle orientation. Indeed, 3D analysis of confocal z-stack images revealed that Tagln2-depleted cells showed strong spindle orientation defects compared with control siRNA (anti-GFP)-treated cells, with the average spindle-axis angle with respect to the substratum increasing approximately twofold (Fig. 2F,G; Fig. S2A,B). The cortical enrichment of Tagln2 in mitotic cells suggested a role at the cortex during spindle orientation, as we did not observe any localization of Tagln2 at spindle poles (Fig. S1E), which also regulate spindle orientation. We therefore assessed whether Tagln2 depletion impacted the localization of LGN and NuMA, two key cytosolic components of the cortical orientation machinery (Bowman et al., 2006; Woodard et al., 2010; Zheng et al., 2010). We observed significant reduction in the cortical accumulation of both LGN and NuMA upon Tagln2 depletion (Fig. 3A-D; Fig. S2F,G). Interestingly, the levels of both LGN and NuMA appeared to increase at the polar cortex upon depletion of LIC2, but not of LIC1. To obtain a quantitative estimate of these observations, we used two methods to quantify the fluorescence intensity at the cortex.

Fig. 3.

Transgelin 2 is required for proper recruitment of LGN at the cortex in mitosis. (A) Representative confocal immunofluorescence images following siRNA-mediated depletion (control or Tagln2 siRNA SMARTpool) in HeLa cells and immunostained with the respective antibodies. (B) Representative confocal fluorescence images of a HeLa cell line stably expressing GFP-LGN. (C) Scatter plots showing the quantification of data in A for LGN using linescans. n=3 experiments, with at least 15 metaphase cells counted per experiment for each condition. (D) Quantification of data in B for GFP-LGN using area-based quantification from small boxes (such as represented in yellow) drawn encompassing the cortex (Sana et al., 2018); 20 cells were counted for each condition. In C and D, each dot represents the intensity ratio for one cell. (E) Immunoprecipitation of empty GFP (control) or GFP-LGN-expressing stable HeLa cell mitotic lysates using GFP-Trap beads followed by immunoblotting (IB) using the respective antibodies. (F) Anti-FLAG immunoprecipitation of empty vector eMTAP (control) and human LIC2-MTAP mitotic lysates obtained from the respective stably expressing U2OS cell lines, followed by immunoblotting with the indicated antibodies to probe whether actin binds to mitotic LIC2-dynein. (G,H) Images showing the co-localization of GFP-Tagln-2 with LGN (G) and NuMA (H). White boxes indicate representative ROIs used for quantifying the colocalization coefficient. (I) Pearson's coefficient of co-localization of GFP-Tagln2 with LGN and NuMA at the cortex in mitotic cells. Error bars represent mean±s.e.m. (C,D) or mean±s.d. (I). *P<0.05, ***P<0.001; ns, not significant. Scale bars: 8 µm.

Fig. 3.

Transgelin 2 is required for proper recruitment of LGN at the cortex in mitosis. (A) Representative confocal immunofluorescence images following siRNA-mediated depletion (control or Tagln2 siRNA SMARTpool) in HeLa cells and immunostained with the respective antibodies. (B) Representative confocal fluorescence images of a HeLa cell line stably expressing GFP-LGN. (C) Scatter plots showing the quantification of data in A for LGN using linescans. n=3 experiments, with at least 15 metaphase cells counted per experiment for each condition. (D) Quantification of data in B for GFP-LGN using area-based quantification from small boxes (such as represented in yellow) drawn encompassing the cortex (Sana et al., 2018); 20 cells were counted for each condition. In C and D, each dot represents the intensity ratio for one cell. (E) Immunoprecipitation of empty GFP (control) or GFP-LGN-expressing stable HeLa cell mitotic lysates using GFP-Trap beads followed by immunoblotting (IB) using the respective antibodies. (F) Anti-FLAG immunoprecipitation of empty vector eMTAP (control) and human LIC2-MTAP mitotic lysates obtained from the respective stably expressing U2OS cell lines, followed by immunoblotting with the indicated antibodies to probe whether actin binds to mitotic LIC2-dynein. (G,H) Images showing the co-localization of GFP-Tagln-2 with LGN (G) and NuMA (H). White boxes indicate representative ROIs used for quantifying the colocalization coefficient. (I) Pearson's coefficient of co-localization of GFP-Tagln2 with LGN and NuMA at the cortex in mitotic cells. Error bars represent mean±s.e.m. (C,D) or mean±s.d. (I). *P<0.05, ***P<0.001; ns, not significant. Scale bars: 8 µm.

We used linescan analysis (Kiyomitsu and Cheeseman, 2012; Seldin et al., 2013; Carminati et al., 2016) in the immunostained cells to determine the relative enrichment of these proteins at the cortex with respect to the adjacent cytoplasm (explained in Materials and Methods). We observed that the cortical levels of both LGN (Fig. 3A,C) and NuMA (Fig. S2F,G) were significantly reduced upon Tagln2 depletion, suggesting that Tagln2 is required to recruit LGN to the cortex. Interestingly, depletion of dynein LIC2 showed the opposite effect, an accumulation of both LGN (Fig. 3A,C) and NuMA (Fig. S2F,G) at the cortex. Our observations collectively indicate that Tagln2 is required to stabilize LGN-NuMA recruitment to the cortex. Indeed, co-depletion of both Tagln2 and LIC2 led to levels of LGN (Fig. 3A,C) and NuMA (Fig. S2F,G) at the cortex being restored to control levels, validating this hypothesis. Notably, depletion of LIC1 did not lead to accumulation of cortical LGN-NuMA, affirming the crucial role of LIC2-dynein in controlling spindle orientation at the cortex. In addition to this, we also recorded the effect of LIC1/LIC2/Tagln2/LIC2+Tagln2 depletion on the cortical localization of LGN in mitosis using a stably expressing GFP-LGN Hela cell line (Kiyomitsu and Cheeseman, 2012) (a kind gift from Iain Cheeseman, Department of Biology, Whitehead Institute of Biomedical Research, Massachusetts Institute of Technology, USA). We quantified the data from these experiments using a standard method employing small rectangular boxes drawn around the cortex (Kotak et al., 2013, 2014; Sana et al., 2018). These results (Fig. 3B,D) mirrored our fixed cell quantification data for LGN (Fig. 3A,C) for all conditions tested, thus confirming the earlier results. Together, all of the above observations suggest that Tagln2 depletion affects LGN-NuMA localization in metaphase without significantly dismantling the actin cortex.

We also checked for the cortical levels of both proteins in anaphase, wherein NuMA enriches at the cortex and engages directly with the plasma membrane phospholipids independently of LGN (Kotak et al., 2014). Tagln2 depletion led to a visible reduction in cortical LGN (Fig. S2I), but did not appear to affect the cortical enrichment of NuMA in anaphase (Fig. S2H), suggesting that the major effect of Tagln2 in metaphase and anaphase is to stabilize the cortical recruitment of LGN, but not NuMA, directly. We tested whether Tagln2 interacts with LGN using mitotic lysates from a HeLa cell line stably expressing GFP-LGN (Kiyomitsu and Cheeseman, 2012) by immunoprecipitation of the GFP tag. We observed a specific interaction of Tagln2 with GFP-LGN but not with the GFP tag alone (Fig. 3E). Given that Tagln2 interacted both with LIC2-dynein (Fig. 1A) and with LGN (Fig. 3E) in mitosis, and also that cortical LGN levels were oppositely influenced by these two proteins (Fig. 3A-D), we checked to see whether the dynein-Tagln2 interaction could be mediated through the actin cortex. However, a negligible amount of actin, if any, pulled down with human LIC2 in affinity precipitation experiments performed on the MTAP-tagged human LIC2 stable cell line (Fig. 3F), suggesting that the actin cortex itself is not a direct mediator. Consistent with the observed biochemical interaction, we also observed significant colocalization of both LGN and NuMA with GFP-Tagln2 at the polar cortex (Fig. 3G-I). Overall, these results established Tagln2 as a novel cortical protein required for spindle orientation by virtue of its stabilization of LGN at the cortex through a physical interaction.

Phosphorylation of LIC2 at a conserved residue is required for spindle orientation

We next investigated whether intrinsic post-translational modification of LIC2 is required for spindle orientation. The LIC subunits of dynein undergo conserved hyperphosphorylation by the master mitotic kinase cdk1 concomitant with entry into mitosis (Niclas et al., 1996; Dell et al., 2000; Nousiainen et al., 2006; Dephoure et al., 2008; Olsen et al., 2010). However, the functional significance of most of these phosphorylation events is not understood, especially for the LIC2 subunit. We systematically investigated the importance of the three cdk1 phosphorylation sites (S194, S383 and S391), all of which are reported to be phosphorylated in mitosis (Nousiainen et al., 2006; Dephoure et al., 2008; Olsen et al., 2010). Accordingly, we generated single-site phosphodeficient mutant constructs (serine to alanine substitutions S194A, S383A, S391A) of rat LIC2 at these residues, as the wild-type rat LIC2 orthologue is competent to functionally rescue the depletion of human LIC2 during mitosis (Fig. 4A) (Mahale et al., 2016a,b). We exogenously expressed each of these FLAG-tagged phosphodeficient constructs of rat LIC2 in a HeLa cell line stably expressing GFP-α-tubulin and histone 2B-mCherry (used in Fig. 1F) and ascertained robust expression by immunoblotting against the FLAG tag (Fig. S3C). We surmised that expression of any phosphodeficient constructs defective in important functions would lead to mitotic defects in a dominant negative manner. Confocal time-lapse imaging following 48 h of plasmid transfection revealed a clear pattern of mitotic phenotypes, wherein the cells displayed significant delays in either early (S194A) or late (S383A/S391A) stages of mitosis (Fig. 4A-E; Movies 3-6). These results highlighted the relative importance of individual mitotic phosphorylation events of LIC2 for its mitotic functions.

Fig. 4.

Site-specific phosphorylation of the LIC2 subunit of dynein is required for metaphase to anaphase and cytokinetic progression. (A) Left: Diagram representing wild-type and single site phosphomutant rat LIC2 constructs. Red and green wedges represent the phosphodeficient alanine substitution and the phosphomimetic glutamic acid substitution of the native serine, respectively. Right: Stills from time-lapse videos of HeLa cells transfected with rat LIC2 wild-type or phosphomutant constructs shown on the left (green, GFP-tubulin; red, mcherry-Histone 2B). (B) Scatter plots for the data shown in A, for NEB to anaphase (B) and anaphase to abscission (C). (D,E) Quantification showing the fraction of mitotic cells delayed for longer than 80 min in metaphase (D) and for longer than 120 min in cytokinesis (E) upon expression of the indicated rat LIC2 constructs. n=3 experiments each, with at least 35 mitotic cells counted per experiment. **P<0.01, ***P<0.001, ****P<0.0001; ns, not significant. Scale bars: 25 µm.

Fig. 4.

Site-specific phosphorylation of the LIC2 subunit of dynein is required for metaphase to anaphase and cytokinetic progression. (A) Left: Diagram representing wild-type and single site phosphomutant rat LIC2 constructs. Red and green wedges represent the phosphodeficient alanine substitution and the phosphomimetic glutamic acid substitution of the native serine, respectively. Right: Stills from time-lapse videos of HeLa cells transfected with rat LIC2 wild-type or phosphomutant constructs shown on the left (green, GFP-tubulin; red, mcherry-Histone 2B). (B) Scatter plots for the data shown in A, for NEB to anaphase (B) and anaphase to abscission (C). (D,E) Quantification showing the fraction of mitotic cells delayed for longer than 80 min in metaphase (D) and for longer than 120 min in cytokinesis (E) upon expression of the indicated rat LIC2 constructs. n=3 experiments each, with at least 35 mitotic cells counted per experiment. **P<0.01, ***P<0.001, ****P<0.0001; ns, not significant. Scale bars: 25 µm.

We also generated stably expressing U2OS cells containing MTAP-tagged human LIC2 (wild-type and S194A mutants; see input lanes of Fig. S3F,H). We depleted endogenous LIC2 using siRNAs and performed anti-GFP confocal immunofluorescence microscopy in order to visualize the subcellular localization of the transgenically expressed proteins. Both LIC2 variants localized to mitotic kinetochores, spindles, spindle poles and the cortex (Fig. S5D,E). Under these conditions, which essentially mimicked a complementation of endogenous human LIC2 depletion with wild-type or S194A rat LIC2 phosphomutants, the S194A mutant failed to rescue the early mitotic arrest but rescued the cytokinetic arrest, equivalent to wild-type rat LIC2 (Fig. S5F,G). Transient expression of the corresponding rat LIC2 constructs in a DHC-GFP stable Hela cell line (Kiyomitsu and Cheeseman, 2012) depleted of endogenous LIC2 (by siRNA treatment) did not alter the normal localization of the dynein complex, as DHC-GFP could be observed at the poles, cortex, kinetochores and mitotic spindles (Fig. S5C). In addition, the IC74 subunit localized to prometaphase kinetochores (enriched by treatment with nocodazole) in an indistinguishable manner following rescue of human LIC2 depletion with either wild-type or S194A rat LIC2 (Fig. S5H-J). Together, these assays confirmed that the subcellular localization of these phosphomutants was largely normal during mitosis.

We investigated whether expression of the S194A mutant caused spindle misorientation by performing functional complementation assays with exogenously expressed wild-type or phosphodeficient rat LIC2 constructs (resistant to siRNA targeting), in the background of endogenous human LIC2 depletion using sequence-specific siRNAs (Fig. S3D). Three-dimensional reconstruction of confocal z-stacks indeed revealed a strong spindle orientation defect with the S194A mutation but not with the other phosphodeficient constructs (Fig. 5A,B), with a significant fraction of mitotic cells (close to 40%) exhibiting a large tilt in the spindle angle (>20°). Analysis of the live cell imaging data (Fig. 4A) suggested that spindle misorientation either precedes or broadly coincides with metaphase plate formation, and is not a consequence of prolonged metaphase arrest (Fig. S6D,E; Movies 14, 15). Expression of the S194A rat LIC2 mutant did not lead to any significant displacement of prophase centrosomes from the nuclear membrane (Fig. S6F,G), a function attributed to dynein. Exogenous expression of the S194E (phosphomimetic) rat LIC2 mutant did not lead to any delays in NEB-to-anaphase progression (Fig. 4). Expression of the S194E phosphomimetic mutant (Fig. S3E) also rescued the spindle misorientation defect observed upon LIC2 depletion (Fig. 5A,C), attesting to the specific role of S194 phosphorylation in mitotic progression and spindle orientation. To probe the integration of the various LIC2 phosphomutants into the dynein complex, we generated stably expressing MTAP-LIC2 wild-type (Mahale et al., 2016a,b) and phosphomutant (S194A/S383A/S391A) U2OS cell lines containing each of the three individual human LIC2 constructs as stable integrants, both in the presence and depletion of endogenous LIC2 (Fig. S3F-H). The phosphorylation sites on both human and rat LIC2 bear the same residue numbers as a result of the high sequence conservation (Fig. S4). Using FLAG-based affinity precipitations (by exploiting the FLAG tag present in the human LIC2-MTAP constructs), we observed efficient interactions of all LIC2 phosphomutants with the core dynein subunits DHC and IC74, demonstrating proper integration of these mutants into the dynein complex (Fig. S3F-H).

Fig. 5.

Phosphorylation of LIC2 at S194 regulates spindle orientation. (A) Representative images showing the spindle angle (enclosed by white lines) in 3D reconstructed images of confocal z-stacks of human LIC2-depleted HeLa cells complemented with expression of wild-type or phosphomutant rat LIC2 constructs as indicated and immunostained for γ tubulin (spindle poles, red), α tubulin (microtubules, green) and DAPI (chromosomes, blue). (B,C) Scatter plots depicting the average spindle angle made by metaphase cells, as represented in A. Individual dots represent values from individual cells. For B, n=3 experiments each, with at least 30 cells were counted per experiment. For C, 30 metaphase cells were counted over three experiments. (D) FLAG mitotic immunoprecipitates of cell lines stably expressing empty MTAP vector, human LIC2-MTAP and MTAP-tagged S194A-LIC2 (in the background of siRNA-mediated human LIC2 depletion, as indicated), probed by immunoblotting for the binding of transgelin 2, γ-tubulin and LGN with the respective antibodies as indicated. (E) Representative images depicting the status of pole focusing (γ tubulin, red) upon complementation of human LIC2 depletion with the indicated wild-type or phosphodeficient rat LIC2 constructs. (F) Quantification of integrated fluorescence intensity of γ tubulin at the spindle poles from the cells imaged in E. n=3 experiments, with at least 25 cells were counted per experiment. (G) Representative images from time-lapse videos of HeLa cells showing microtubule nucleation upon complementation with wild-type or S194A rat LIC2 in the background of human LIC2 depletion. Control represents untreated cells. (H) Quantification of microtubule nucleation using Imaris software upon complementation of human LIC2 depletion with the indicated rat LIC2 constructs. n=3 experiments for each condition, with at least 15 cells counted per experiment. Error bars represent mean±s.e.m. (B,C,F) or mean±s.d. (H). *P<0.05, ***P<0.001, ****P<0.0001; ns, not significant. Scale bars: 5 µm (A,G), 8 µm (E).

Fig. 5.

Phosphorylation of LIC2 at S194 regulates spindle orientation. (A) Representative images showing the spindle angle (enclosed by white lines) in 3D reconstructed images of confocal z-stacks of human LIC2-depleted HeLa cells complemented with expression of wild-type or phosphomutant rat LIC2 constructs as indicated and immunostained for γ tubulin (spindle poles, red), α tubulin (microtubules, green) and DAPI (chromosomes, blue). (B,C) Scatter plots depicting the average spindle angle made by metaphase cells, as represented in A. Individual dots represent values from individual cells. For B, n=3 experiments each, with at least 30 cells were counted per experiment. For C, 30 metaphase cells were counted over three experiments. (D) FLAG mitotic immunoprecipitates of cell lines stably expressing empty MTAP vector, human LIC2-MTAP and MTAP-tagged S194A-LIC2 (in the background of siRNA-mediated human LIC2 depletion, as indicated), probed by immunoblotting for the binding of transgelin 2, γ-tubulin and LGN with the respective antibodies as indicated. (E) Representative images depicting the status of pole focusing (γ tubulin, red) upon complementation of human LIC2 depletion with the indicated wild-type or phosphodeficient rat LIC2 constructs. (F) Quantification of integrated fluorescence intensity of γ tubulin at the spindle poles from the cells imaged in E. n=3 experiments, with at least 25 cells were counted per experiment. (G) Representative images from time-lapse videos of HeLa cells showing microtubule nucleation upon complementation with wild-type or S194A rat LIC2 in the background of human LIC2 depletion. Control represents untreated cells. (H) Quantification of microtubule nucleation using Imaris software upon complementation of human LIC2 depletion with the indicated rat LIC2 constructs. n=3 experiments for each condition, with at least 15 cells counted per experiment. Error bars represent mean±s.e.m. (B,C,F) or mean±s.d. (H). *P<0.05, ***P<0.001, ****P<0.0001; ns, not significant. Scale bars: 5 µm (A,G), 8 µm (E).

Our observations that Tagln2 was a specific mitotic interactor of LIC2-dynein (Fig. 1A), which regulates spindle orientation (Fig. 2F,G), prompted us to test whether phosphorylation of LIC2 at S194 was required for the interaction with Tagln2. Interestingly, we observed that Tagln2 could interact efficiently with both wild-type LIC2 and S194A-LIC2 proteins in FLAG affinity precipitates from mitotic lysates of the respective stable cell lines upon depletion of endogenous LIC2 (Fig. 5D). This result demonstrated that the interaction of Tagln2 with LIC2-dynein was largely independent of S194 phosphorylation of LIC2. We therefore probed whether the integrity of the spindle poles, the other cellular hotspot (apart from the cortex) crucial in maintaining proper spindle orientation, was affected by the loss of phosphorylation at S194. We performed functional complementation experiments (as described above) followed by confocal imaging of HeLa cells immunostained for γ-tubulin, a bona fide marker for centrosomes (Moudjou et al., 1996; Leguy et al., 2000; Lecland and Lüders, 2014). It was apparent from these images that the mild pole-splitting defect observed upon human LIC2 depletion (Mahale et al., 2016a) could not be rescued by expression of the S194A-LIC2 mutant (Fig. 5E). However, expression of either of the other two LIC2 C-terminal phosphodeficient mutants (S383A, S391A; Fig. 5E), or the S194E phosphomimetic mutant (Fig. S6A), restored tightly focused poles. We quantified the levels of γ-tubulin at the spindle poles using 3 µm diameter ROIs marked around the mitotic centrosomes (which encompassed most control centrosomes) (Lecland and Lüders, 2014). Interestingly, we observed the accumulation of higher amounts of γ-tubulin around the centrosomes upon rescue with S194A-LIC2, but no change upon rescue with the other two LIC2 C-terminal phosphodeficient mutants (Fig. 5F) or with the S194E phosphomimetic mutant (Fig. S6A-C). These results suggested that the fragmented γ-tubulin at the centrosomes correlates with the spindle misorientation phenotype observed with the S194A mutation.

We also investigated whether the S194A-dependent pole fragmentation affected microtubule nucleation. γ-Tubulin is the central component of the γ-tubulin ring complex (γ TURC), the major microtubule-nucleating factor at the centrosomes (Joshi et al., 1992; Stearns and Kirschner, 1994; Leguy et al., 2000; Piehl et al., 2004; Wiese and Zheng, 2006; Kollman et al., 2011), and is known to be transported to the poles by dynein (Smith et al., 2000; Young et al., 2000). Defects in astral microtubule nucleation are well documented in the literature to be a major cause of spindle misorientation (Raynaud-Messina and Merdes, 2007; Kollman et al., 2011). We therefore predicted that the minor pole fragmentation observed with the S194A mutant would lead to defects in microtubule nucleation capability. To test this hypothesis, we performed functional complementation experiments (as shown in Fig. 5E) on mitotic HeLa cells stably expressing EB1-GFP (to observe growing microtubule tips) and histone 2B-mCherry (to observe the chromosomes). We performed live cell, time-lapse confocal microscopy on these cells (Movies 7-10, 13, 16; representative images shown in Fig. 5G) and quantified the number of EB1 comets observed per frame (∼3.6 s between successive frames). Depletion of LIC2 led to more than 30% reduction in the number of EB1-GFP spots observed, which could be rescued by the expression of wild-type and S194E rat LIC2 mutants but not by the S194A mutant (Fig. 5G,H), indicating a severe microtubule nucleation defect as a result of the S194A phosphomutation. In consonance with this data, affinity purification experiments on mitotic lysates of the respective stable cell lines (as used in Fig. S3F-H) showed that the interaction of human LIC2 with γ-tubulin was significantly reduced by S194A phosphomutation (Fig. 6D). This result was replicated, even after siRNA-mediated depletion of endogenous LIC2 (Fig. 5D). These observations imply that LIC2-S194 phosphorylation is required for proper focusing of γ-tubulin at the centrosome to enable optimal microtubule nucleation. Overall, this set of experiments demonstrated the hitherto-unknown importance of the conserved S194 phosphorylation (Fig. S4) in ensuring proper microtubule nucleation, which presents a plausible explanation for its essential role in spindle orientation.

Fig. 6.

S194 phosphorylation of LIC2 is required for preventing premature centriole disengagement. (A) Representative confocal immunofluorescence micrographs of Hela cells immunostained for the indicated antibodies following LIC2 siRNA treatment and complementation with transiently expressed rat LIC2 (wild type or S194A). Insets in the merged images show zoomed areas around the respective centrosomes; arrows indicate individual centrioles. Scale bar: 5 µm; optical zoom factor 5, 63× oil lens. (B) Quantification of the fraction of mitotic poles (visualized by γ-tubulin immunostaining, red) that contain the number of centrioles shown on the x-axis (visualized by centrin-1 immunostaining, green). n=3 independent experiments, with at least 15 metaphase cells analysed per experiment. (C) SBP affinity purification (AP SBP) of mitotic lysates obtained from stable U2OS cell lines expressing empty vector (eMTAP) or the indicated MTAP-human LIC1 or LIC2 proteins, followed by immunoblotting (IB) with the respective antibodies as indicated. IN, input; FT, flow through; W, wash; AP, affinity purified eluate. Error bars indicate mean±s.d. ***P<0.001; ns, not significant.

Fig. 6.

S194 phosphorylation of LIC2 is required for preventing premature centriole disengagement. (A) Representative confocal immunofluorescence micrographs of Hela cells immunostained for the indicated antibodies following LIC2 siRNA treatment and complementation with transiently expressed rat LIC2 (wild type or S194A). Insets in the merged images show zoomed areas around the respective centrosomes; arrows indicate individual centrioles. Scale bar: 5 µm; optical zoom factor 5, 63× oil lens. (B) Quantification of the fraction of mitotic poles (visualized by γ-tubulin immunostaining, red) that contain the number of centrioles shown on the x-axis (visualized by centrin-1 immunostaining, green). n=3 independent experiments, with at least 15 metaphase cells analysed per experiment. (C) SBP affinity purification (AP SBP) of mitotic lysates obtained from stable U2OS cell lines expressing empty vector (eMTAP) or the indicated MTAP-human LIC1 or LIC2 proteins, followed by immunoblotting (IB) with the respective antibodies as indicated. IN, input; FT, flow through; W, wash; AP, affinity purified eluate. Error bars indicate mean±s.d. ***P<0.001; ns, not significant.

S194 phosphorylation of LIC2 prevents premature centriole disengagement in mitosis

We further characterized the pole fragmentation defect observed upon LIC2 depletion or S194A mutation. We performed rescue experiments with either wild-type or S194A rat LIC2 (through transient transfections) in the background of human LIC2 depletion, and assessed the number of centrioles (with centrin-1 immunostaining) per pole (γ-tubulin immunostaining). The major population of control cells showed two centrioles per γ-tubulin-positive pole, as expected (Fig. 6A,B). However, quantitative analysis revealed an increase in monocentriolar poles (single centrin-1 dot inside a γ-tubulin puncta) as the major defect upon LIC2 depletion (Fig. 6A,B), as also reported earlier (Jones et al., 2014). Interestingly, this defect could be rescued by expression of wild-type rat LIC2 but not by the S194A mutant (Fig. 6A,B). We did not observe any significant increase in the fraction of bi- or multinucleate cells with S194A-LIC2 mutation (Fig. S5G), suggesting that this phosphorylation does not impact cytokinetic cleavage, consistent with the data shown in Fig. 4 regarding cytokinetic timing. Our results suggest that prevention of S194 phosphorylation underlies the premature centriole disengagement observed upon LIC2 depletion in mitosis.

The overall results from this study reveal two novel molecular mechanisms, centering around the dynein motor, that regulate mitotic spindle orientation. First, we uncover a new, mitosis-specific interaction of dynein with the cortically localized actin-bundling protein Tagln2, whose depletion also led to mitotic delay (Fig. 1). This is the first report, to our knowledge, of mitotic functions for Tagln2. The Mitocheck study had earlier suggested a role for Tagln2 in mitosis (https://www.mitocheck.org/gene.shtml?gene=ENSG00000158710); however, there are no studies published on a mitotic function for Tagln2. The robust protein depletion and the phenocopy of mitotic defects observed with multiple sequence-specific siRNAs (Fig. 1, Fig. S2B-E) suggests that the mitotic phenotypes are specific and unlikely to be off-target effects. Tagln2 depletion did not significantly perturb the integrity of the F-actin cortex (Fig. S1). Moreover, when we depleted both LIC2 and Tagln2, LGN-NuMA was reinstated at the cell cortex in a manner indistinguishable from untreated cells (Fig. 3; Fig. S5). These results suggest that LGN retains the ability to recruit to the cortex independently of Tagln2, which may serve as an additional stabilizing factor for LGN at the cortex. Such a mechanism has been proposed for Band 4.1, which is a stabilizing factor for NuMA retention at the cortex in mitosis (Seldin et al., 2013).

Chromosome miscongression (Fig. 2A-C) upon Tagln2 depletion could conceivably be a result of defective capture of spindle microtubules at the kinetochores, although this hypothesis remains to be experimentally demonstrated. The shortening of the mitotic spindle upon Tagln2 depletion (Fig. 2D,E) could be perceived as counterintuitive with respect to the known effect of dynein (a Tagln2 interactor), whose depletion normally lengthens the spindle (Gatlin and Bloom, 2010). However, this observation suggests a few possibilities worthy of future exploration. First, the displacement of LGN-NuMA from the cortex upon Tagln2 depletion (Fig. 3; Fig. S2F,G) would impede dynein-mediated pulling forces on the poles along astral microtubules. This force is opposite to that generated by spindle-localized dynein, which pulls the two poles together (Gatlin and Bloom, 2010), thus leading to spindle shortening upon Tagln2 depletion. Second, there are reports that the F-actin cytoskeleton regulates mitotic spindle length. Specifically, cortical F-actin has been shown to lengthen the mitotic spindle in Xenopus laevis oocytes (Woolner et al., 2008). Tagln2, which localizes at the cortex in mitosis (Fig. 3G,H; Fig. S1), could aid in this function, therefore leading to spindle shortening upon Tagln2 depletion (Fig. 2). Chromosome misalignment and spindle length dysregulation are expected to cause mitotic delays due to prolonged activation of the spindle assembly checkpoint.

Tagln2 was enriched at the cortex and at the kinetochores in mitotic cells, but not visibly on the mitotic spindles or on the spindle poles (Fig. 3G,H; Fig. S1). This observation led us to explore the role of Tagln2 in spindle orientation, for which the cortex is a major hub of essential biochemical activity. At metaphase, Tagln2 was required for the optimal stabilization of LGN and NuMA at the cortex (Fig. 3; Fig. S2), key cogs in the cortical machinery responsible for anchoring the spindle to the cell periphery (Schaefer et al., 2000; Yu et al., 2000; Woodard et al., 2010; Peyre et al., 2011). LGN is known to be recruited prior to NuMA in mitosis to the cortex up to metaphase (Merdes et al., 2000; Kiyomitsu and Cheeseman, 2012; Kotak et al., 2014; Kotak and Gönczy, 2014; Seldin et al., 2016). At anaphase, NuMA directly engages with membrane phospholipids and enriches at the cortex without a requirement for LGN (Kotak et al., 2014). Tagln2 depletion led to the visible cortical loss of LGN but not of NuMA in anaphase (Fig. S2H,I), suggesting that the primary effect of Tagln2 is on the stabilization of LGN at the cortex. However, co-depletion of LIC2-dynein and Tagln2 restored the cortical localization of LGN, which is known to be recruited by the membrane-anchored Gαi protein (Woodard et al., 2010). These observations suggested that although LGN recruits to the polar cortex independently through cortical Gαi, its cortical localization is further stabilized by its interaction with Tagln2. The cortical stabilization of LGN-NuMA by Tagln2 and the LGN-Tagln2 interaction are consistent with a model in which opposing forces imparted on LGN-NuMA at the cortex (namely cortical anchoring through Tagln2 versus poleward pulling forces generated primarily by LIC2-dynein) help maintain optimal levels of cortical LGN-NuMA to capture astral microtubules and thus orient the spindle. This model is consistent with the known role of dynein in transporting LGN-NuMA towards spindle poles along astral microtubules (Zheng et al., 2013; Mahale et al., 2016a). Alternatively, Tagln2 could serve as an additional anchor for the dynein complex at the cortex through its interaction with LIC2, and could help stabilize the dynein-NuMA-LGN complex at the cortex. In this scenario, Tagln2 depletion weakens or releases the anchoring of dynein at the cortex, allowing it to transport LGN-NuMA poleward along astral microtubules. More studies are needed to elucidate the exact mechanism of Tagln2 function in spindle orientation. Regardless, the potent role of Tagln2 in spindle orientation emphasizes the emerging role of the actin cortex in this important mitotic function through the ERM proteins afadin, CAPZ and MISP (Wee et al., 2011; Zhu et al., 2013; Machicoane et al., 2014; Carminati et al., 2016; di Pietro et al., 2016, 2017; Kschonsak and Hoffmann, 2018). Tagln2 is a novel player added to the list of specific cortical actin-binding proteins that oppose the poleward force exerted by dynein along astral microtubules, a concept that has been earlier proposed (Palmer et al., 1992; Zheng et al., 2013). It remains to be seen whether Tagln2 acts in collaboration with or independently of these other cortical actin-binding proteins.

The second major insight from this work is the importance of mitotic phosphorylation of the LIC2 subunit of dynein in spindle orientation. Our study uncovers distinct contributions of N-terminal phosphorylation (S194, in metaphase functions such as chromosome congression, pole-mediated microtubule nucleation, spindle orientation and timely anaphase onset) and clustered C-terminal phosphorylations (S383 and S391, during cytokinesis) (Fig. 4). These observations reveal finely tuned regulation of the mitotic functions of LIC2, depletion of which has been shown to impede the progression of both metaphase and cytokinesis (Palmer et al., 2009; Raaijmakers et al., 2013; Mahale et al., 2016a,b). Phosphorylation at S194 is required for orienting the metaphase spindle by enabling optimal astral microtubule nucleation (Fig. 5), which could probably be impaired because of the requirement for S194 phosphorylation in maintaining mitotic centrosome integrity (Fig. 6; Fig. S6A,B). The premature centriole disengagement induced by LIC2 depletion (Fig. 6) (Jones et al., 2014) could be rescued by wild-type LIC2 but not by the S194A mutant of LIC2 (Fig. 6), suggesting that S194 mitotic phosphorylation is necessary for maintaining cohesion of the centrioles in mitosis and enabling proper microtubule nucleation. Failure of the S194A-LIC2 mutant to bind to γ-tubulin (Fig. 6C) is consistent with a hypothesis whereby centrosomal LIC2-dynein helps properly anchor γ-tubulin transported to the vicinity of the pole to confer optimal microtubule nucleation capability to the centrosome. Despite its evident importance in multiple mitotic dynein functions, prevention of S194 phosphorylation is unlikely to impair all mitotic dynein functions. This conclusion is corroborated by the efficient incorporation of S194A-LIC2 into the dynein complex (Fig. S3F,H) and the proper localization of S194A-LIC2 on the spindle poles, cortex and kinetochores of mitotic cells (Fig. S5E-J). This localization of S194A-LIC2 is largely mirrored by the localization of DHC upon rescue of endogenous LIC2 depletion with S194A-LIC2 (Fig. S5C). Moreover, S194A-LIC2 is able to bind to Tagln2 as efficiently as wild-type LIC2 (Fig. 5D). Functionally, S194 phosphorylation does not appear to be essential for maintaining the adherence of the late G2/prophase centrosome to the nuclear membrane (Fig. S6F,G) or for cytokinetic progression (Fig. S5G). Together, these observations suggest that not all dynein functions in mitosis require S194 phosphorylation. Our data suggest that the prominent contributions of S194 mitotic phosphorylation are in microtubule nucleation, spindle orientation and chromosome congression, which together probably contribute to the metaphase-to-anaphase delay. Further exploration is also needed to delineate the deeper mechanisms governing the role of S194 phosphorylation in ensuring centriole cohesion. Interestingly, S194 phosphorylation did not affect the binding of LIC2 to the cortical proteins LGN or Tagln2 (Fig. 5D), which appear to bind robustly to both wild-type LIC2 and S194A-LIC2 in mitosis, suggesting that S194 phosphorylation impacts the pole-localized pool of LIC2-dynein, but not the cortically localized pool.

The high sequence conservation of human LIC2 S194 (Fig. S4) suggests that the mitotic functions of this phosphorylation are likely to be conserved in other vertebrates. Notably, our results reinforce the idea that dynein LIC1 and LIC2 are capable of performing distinct cellular functions through their interaction with separate subsets of cellular proteins, a notion with precedent in the literature (Tynan et al., 2000; Mahale et al., 2016a; Gonçalves et al., 2019). Dynein LICs play central roles in engaging with the dynactin complex through various adaptor proteins, thus imparting both processivity and cargo-binding specificity to the motor (Reck-Peterson et al., 2018). Our study reveals aspects of the finely orchestrated regulation of the LICs that govern dynein function. Further investigation is likely to illuminate the contributions of the two LICs in different stages of the cell cycle.

Cell culture and cell synchronization

U2OS (human osteosarcoma cell line) and HeLa (human epithelial cervical cancer cell line) cells were obtained from the ATCC and authenticated. HeLa and U2OS cell lines were grown in Dulbecco's modified Eagle's medium (DMEM, high glucose). HeLa cells expressing GFP-tagged α-tubulin and mCherry-tagged histone 2B were grown in DMEM medium supplemented with puromycin and hygromycin B antibiotics, respectively. The LGN-GFP HeLa stable lines and EB1-GFP and histone 2B-mCherry stable cell lines were grown in DMEM medium. For synchronization in prometaphase, nocodazole (Sigma) was used at 150 nM and 400 nM for HeLa and U2OS cells, respectively, for 12 h. Prometaphase cells were washed with PBS and incubated for 1 h in prewarmed DMEM medium to release them into metaphase and then harvested by gentle shake off.

Plasmid constructs, siRNA and transfection

Full-length rat LIC2, GFP Tagln2 and empty GFP plasmid constructs were used for transfection. These constructs were transfected using Lipofectamine 2000 (Invitrogen) according to the manufacturer's instructions. Single siRNAs against different target genes were procured from Dharmacon. The sequences for the different siRNAs (5′ to 3′) were as follows: GFP (negative control), CAUGAAGCAGCACGACUUC (100 nM); LIC1, GAAAGUUUGUACAUGAGAA (100 nM); LIC2a, ACCUCGACUUGUUGUAUAA (100 nM); LIC2b, GCCGGAAGAUGCAUAUGAA (100 nM); T2A, GCAAGAACGUGAUCGGGUU (100 nM); and T2B, UAUGUGAGCUCAUUAAUGC (100 nm).

All of these siRNAs have been previously used and published by different groups (Palmer et al., 2009; Sivaram et al., 2009; Mahale et al., 2016a,b). A SMARTpool of four siRNAs was used against Tagln2 (Dharmacon). All siRNAs were transfected using Dharmafect 1 reagent (Dharmacon) as described in the manual. For rescue experiments in HeLa cells, plasmids were transfected on day 1 using Lipofectamine 2000 and siRNAs transfected after 24 h. The metaphase index was calculated using DAPI-stained HeLa cells to visualize a clear metaphase plate. The cytokinetic index was calculated by visualizing dense α-tubulin staining in the intercellular bridge. In live movies, the visible severing of the dense α-tubulin bundle in the cytokinetic bridge was used as a marker for the onset of abscission.

Immunoblotting

Cells were scraped in 2× Laemmli buffer containing loading dye and heated at 95°C for 5 min. Samples were resolved using SDS-PAGE, followed by transfer of the proteins onto a polyvinylidene difluoride (PVDF) membrane (Millipore). Protein was transferred onto PVDF membranes, which were then blocked in 5% skimmed milk followed by incubation with primary antibody overnight at 4°C or at room temperature for 1 h. Subsequently, the blots were washed using PBST (phosphate buffered saline with 0.1% Tween 20) followed by incubation with secondary horse radish peroxidase (HRP)-conjugated antibodies for 1 h at room temperature and washed using PBST. Chemiluminescence signals were developed using the Luminata Forte reagent (Millipore) and signals detected using Image Quant 4000 (GE). The various antibodies used for immunoblotting were as follows: LIC2 (Abcam ab178702; 1:1000), LIC1 (Thermo Fisher Scientific PA531644; 1:1000), γ-tubulin (Bethyl A302-631A; 1:2000), FLAG M2 (Sigma F-1804; 1:10,000), β-actin (Sigma A3853; 1:2000), GAPDH (Invitrogen 437000; 1:500), Tagln2 (Abcam ab121146, 1:200; Cloud-Clone PAP-790Hu01, 1:250), IC74 (Abcam ab23905; 1:1000), DHC (Protein Tech, 12345-1-AP; 1:500), NuMA (Thermo Fisher Scientific MA5-17293; 1:1000) and LGN (Abcam ab84571; 1:250); secondary antibodies used were anti-mouse HRP (Jackson Laboratories 715-035-150; 1:10,000) and anti-rabbit HRP (Jackson Laboratories 711-035-152; 1:10,000).

Affinity purification

Human LIC2 was cloned into a multi-affinity tagged vector (pCDNA4-TO-hygromycin-mVenus-MAP vector; henceforth referred to as ‘MTAP’), which imparts a tandem 6His-FLAG-streptavidin binding protein (SBP) motif embedded in a fluorescent yellow fluorescent protein (YFP) tag at the C-terminus of the cloned gene. Human LIC2 was amplified from the pLVX-mGFP LIC2 vector using a gene-specific set of primers that contained EcoRV and HindIII restriction sites. The amplified human LIC2 was cloned into the MTAP vector using EcoRV HF and HindIII restriction sites. The identities of clones were confirmed by double restriction digestion followed by sequencing of the entire insert to confirm its identity (Eurofins). We prepared a stable cell line using this multi-affinity tagged vector in a U2OS (human osteosarcoma) cell line. U2OS cells transiently transfected with the human LIC2-MTAP transgene were cultured as above and treated with 300 µg/ml of hygromycin antibiotic for selection of stable integrants. The fluorescence of this stable cell line was observed under a Nikon fluorescence microscope and confocal microscopes (Leica TCS SP5 and Leica TCS SP8). The U2OS cells stably expressing human LIC2-MTAP were further sorted using flow cytometry based on the YFP fluorescence tag (BD Influx).

For further experiments, we chose cells expressing human LIC2 at levels equivalent to endogenous LIC2. These cells were cultured in DMEM medium containing hygromycin B, treated with 400 nM nocodazole for 10-12 h to enrich them in prometaphase, washed with PBS and cultured for 1 h in fresh medium without nocodazole to release the cells from prometaphase and allow them to progress to metaphase. Metaphase cells were shaken off from the plate and immediately frozen in liquid nitrogen. These cells were then ground in cryogenic conditions (under liquid nitrogen using a mortar and pestle) to generate a ‘cryo-grindate’ that was later reconstituted in 1× lysis buffer (50 mM Tris-HCl pH 7.5, 125 mM NaCl, 1 mM EDTA, 0.2% NP-40 and 5% glycerol) along with phosphatase inhibitors and protease inhibitors. The reconstituted cryo-grindate was loaded on to a StrepTrap HP 1 ml column (GE Healthcare) at a rate of 0.5 ml/min. This tagged transgene binds to Streptactin beads through its SBP tag. The column was washed with ten column volumes of wash buffer (50 mM Tris-HCl pH 7.5, 250 mM NaCl, 0.05% NP-40). Finally, elution of the bound protein was achieved by passing 5-10 column volumes of elution buffer (25 mM Tris-HCl pH 7.5, 125 mM NaCl and 2.5 mM desthiobiotin).

Immunoprecipitation

Cells were lysed using lysis buffer (20 mM Tris-Cl pH 7.5, 125 mM NaCl, 2 mM EGTA, 0.5% Triton X-100) containing 1× HALT protease inhibitor (Thermo Fisher Scientific). Pre-conjugated FLAG-agarose beads (Sigma) were washed three times with lysis buffer. Cell lysate was incubated for 4 h or overnight with the washed FLAG-agarose beads. The beads were washed using lysis buffer, and beads containing the bound proteins were heated at 95°C in Laemmli buffer. Immunoprecipitated lysate was separated using SDS-PAGE and subjected to immunoblotting as described. GFP immunoprecipitation was carried out using GFP-Trap beads (Chromotek) in a similar fashion as per manufacturer's instructions. In all figures, IN, input; FT, flowthrough, and IP/AP, immunoprecipitate.

Immunostaining

We performed immunostaining in HeLa cells using antibodies against microtubules (α tubulin, 1:1000; Sigma T6199), γ-tubulin (spindle poles, Bethyl A302-631A, 1:500 and Abcam ab11316, 1:800), β-actin (Sigma A3853; 1:500), Tagln2 (Abcam ab121146; 1:200), LGN (Abcam ab84571; 1:500), NuMA (Thermo Fisher Scientific MA5-17293; 1:500), centrin-1 (Abcam ab101332; 1:500), CREST (Antibodies Incorporated; 1:50), GFP (Abcam ab6556; 1:1000) and chromatin (DAPI, Sigma). Anti-mouse-IgG and anti-rabbit-IgG AlexaFluor-488-conjugated secondary antibodies (715-545-150, 711-545-152), anti-mouse- and anti-rabbit-IgG AlexaFluor-594-conjugated secondary antibodies (715-585-150, 711-585-152) and anti-human Cy5-conjugated secondary antibody (109-175-008) for immunofluorescence were purchased from Jackson Immunoresearch. For immunostaining, cells grown on coverslips were fixed using chilled methanol for at least 20 min at −20°C, washed with PBS containing 1% bovine serum albumin and 0.5% Triton X-100 (PBSAT) and incubated with primary antibody for 45 min. After incubation, the cells were washed with PBSAT and incubated with secondary antibodies conjugated to suitable fluorophores for 45 min. Finally, the coverslips were mounted on glass slides using Prolong Gold or Prolong Diamond anti-fade reagent (Invitrogen) to minimize fluorescence signal bleaching. Imaging was performed after drying and setting of the anti-fade mounting medium overnight in the dark at room temperature.

Immunofluorescence imaging of cells

High resolution immunofluorescence confocal images were acquired using a Leica TCS SP8 laser scanning confocal microscope using either a HCX PL APO CS 63× oil immersion objective or a HCX PL APO CS 40× oil immersion objective. All acquisition settings for imaging were kept identical for both control and test samples.

Image analysis and quantification

Confocal image analysis was performed using the Imaris software suite (Bitplane) for quantification of fluorescence intensity. For intensity quantification of γ tubulin at the poles, metaphase cells were suitably immunostained. Three-dimensional images were reconstructed using z-stacks in the Imaris software suite. Integrated fluorescence intensities were calculated at poles for γ tubulin in mitotic cells by drawing a sphere of 3 µm diameter (average size of the pole is 2-3 μm in HeLa cells) (Colello et al., 2012; Lecland and Lüders, 2014) around each pole and measuring the intensity of γ tubulin, with subtraction of the local background. The total integrated intensity graph was plotted using GraphPad Prism software. For spindle orientation measurements, immunofluorescence images were visualized in 3D using the Imaris software suite. These images were exported into the tiff file format and opened in ImageJ software. The spindle angle of mitotic cells made by the spindle axis with the substratum (bottom of the cover slip) was calculated using ImageJ software and methods described earlier (Delaval et al., 2011; Hehnly and Doxsey, 2014; Juschke et al., 2014; Mahale et al., 2016a). For measuring the rates of microtubule nucleation, time-lapse imaging was performed using an EB1-GFP and histone 2B-mCherry HeLa stable cell line. The EB1-GFP stable cell line was used to calculate the rate of nucleation in control versus treated cells. The time between two successive frames was approximately 3.6 s. The rate of nucleation of the microtubules was calculated using the Imaris software, which was instructed to pick automatically any spot of 0.25 nm above the designated threshold fluorescence intensity and to track each spot over time to yield the number of spots recorded per frame. The parameters for this analysis were set according to previous reported analyses using Imaris (Stout et al., 2011; Yamashita et al., 2015). Spindle length was calculated as the distance between the two poles measured from 3D reconstructions using the Imaris 3D software.

Quantification of LGN and NuMA intensities at the cortex

We quantified LGN and NuMA intensities at the cortex using the LASX offline software from Leica. We drew a line bisecting the two poles and running across the mitotic cell cortex and quantified the intensity at the cortex and 10 pixels inside the cell from the cortex to record cytosolic intensity of the molecule for comparison with the cortical intensity, as described earlier (Carminati et al., 2016). We determined the cortex/cytoplasm ratios for LGN and NuMA independently for the purpose of quantification. Additionally, we quantified LGN, NuMA and actin intensities at the cortex using ImageJ software. To calculate the cortex/cytosol mean intensity ratios, we drew rectangular ROIs measuring 1.8 µm×4 µm placed on the mitotic cell cortex, in the cytosol or outside the cytosol (to calculate the background for subtraction from each of the cortical and cytosolic values) (Sana et al., 2018). We determined the cortex/cytoplasm ratio for LGN, NuMA and actin independently for the purpose of quantification. To calculate IC74 kinetochore intensity, we made a 0.45 µm diameter sphere around the kinetochore and measured the intensity of dynein IC74 within this sphere, normalized to the intensity of the kinetochore (CREST) from the same sphere, and plotted the ratio using GraphPad Prism software. We measured actin cortex thickness in cells using LASX (Leica) offline software. We drew three lines bisecting the cell and running across the mitotic cell cortex at both ends and quantified the cortical peak widths from the linescan intensity profiles. The mean peak width per mitotic cell was calculated as the average of six cortical peaks (from three linescan profiles per cell). The average actin cortex thickness (average of mean peak widths) over 12 mitotic cells was calculated per condition.

Site-directed mutagenesis

Primers were designed for site-directed mutagenesis with a primer length of 35-40 bases for human and rat LIC2 inserts, as shown in Table S8. The site of mutations was kept closer to the centre of the primers. PCR reactions were set up using Phusion polymerase (Thermo Fisher Scientific). After amplification, the PCR product was digested overnight at 37°C using Dpn1 enzyme, which cleaves only the parental methylated strand. The newly formed strand contained the desired mutations. This PCR product was transformed either in DH5α or XLBlue1 ultracompetent Escherichiacoli cells. Transformed competent cells were plated onto LB agar plates with the desired antibiotic selection and incubated in the bacterial incubator at 37°C overnight. The bacterial colonies were picked and the plasmid isolated. The desired mutations in our gene of interest were confirmed by DNA sequencing (Scigenom/Agrigenom). The primers used to obtain the desired mutants in human or rat LIC2 genes are shown in Table S8.

Statistical analysis

The total numbers of cells counted per experiment for statistical analyses are stated in the corresponding figure legends. The graphs in all figures were made using GraphPad Prism software. Error bars were calculated from at least three experiments and represent standard deviation (s.d.) or standard error of the mean (s.e.m.). Statistical significance of experiments was measured by the one-way ANOVA or Student's t-test with Tukey's comparison method or using the Kruskal–Wallis test.

We thank Prof. Stephen J. Doxsey for providing U2OS cells, Prof. Dannel McCollum for the pMTAP-mVenus expression vector, Prof. Daniel W. Gerlich for the EGFP-α tubulin - mCherry-H2B stable HeLa cell line, Dr Iain Cheeseman for the GFP-LGN cell line, Dr Fumio Matsuzaki for the LGN antibody, Dr Mahak Sharma for the EB1-GFP and histone 2B-mCherry stable cell line, and Dr Sandeep Saxena for the centrin-1 and γ-tubulin antibodies. We are grateful to the Regional Centre for Biotechnology (RCB) for providing the requisite infrastructure and to all members of the Laboratory of Cellular Dynamics, RCB for critical comments and suggestions during the study.

Author contributions

Conceptualization: S.V.S.M.; Methodology: A.S., S.D., S.V.S.M.; Validation: A.S., S.D., S.V.M.; Formal analysis: A.S., S.D., S.V.S.M.; Investigation: A.S., S.D., S.V.S.M.; Resources: S.V.S.M.; Data curation: A.S., S.D., S.V.S.M.; Writing - original draft: A.S., S.D., S.V.S.M.; Writing - review & editing: A.S., S.D., S.V.S.M.; Visualization: A.S., S.V.S.M.; Supervision: S.V.S.M.; Project administration: S.V.S.M.; Funding acquisition: S.V.S.M.

Funding

A.S. was funded by the Indian Council of Medical Research, India and S.D. by the Regional Centre for Biotechnology (RCB) for their respective PhD fellowships. This research was supported by extramural funds by the Department of Biotechnology, Ministry of Science and Technology, India and by intramural funds from Regional Centre for Biotechnology provided to S.V.S.M.

Ahmed
,
S. M.
,
Nishida-Fukuda
,
H.
,
Li
,
Y.
,
McDonald
,
W. H.
,
Gradinaru
,
C. C.
and
Macara
,
I. G.
(
2018
).
Exocyst dynamics during vesicle tethering and fusion
.
Nat. Commun.
9
,
5140
.
Bowman
,
S. K.
,
Neumüller
,
R. A.
,
Novatchkova
,
M.
,
Du
,
Q.
and
Knoblich
,
J. A.
(
2006
).
The Drosophila NuMA Homolog Mud regulates spindle orientation in asymmetric cell division
.
Dev. Cell
10
,
731
-
742
.
Carminati
,
M.
,
Gallini
,
S.
,
Pirovano
,
L.
,
Alfieri
,
A.
,
Bisi
,
S.
and
Mapelli
,
M.
(
2016
).
Concomitant binding of Afadin to LGN and F-actin directs planar spindle orientation
.
Nat. Struct. Mol. Biol.
23
,
155
-
163
.
Chen
,
C.-T.
,
Hehnly
,
H.
,
Yu
,
Q.
,
Farkas
,
D.
,
Zheng
,
G.
,
Redick
,
S. D.
,
Hung
,
H.-F.
,
Samtani
,
R.
,
Jurczyk
,
A.
,
Akbarian
,
S.
, et al. 
(
2014
).
A unique set of centrosome proteins requires pericentrin for spindle-pole localization and spindle orientation
.
Curr. Biol.
24
,
2327
-
2334
.
Colello
,
D.
,
Mathew
,
S.
,
Ward
,
R.
,
Pumiglia
,
K.
and
LaFlamme
,
S. E.
(
2012
).
Integrins regulate microtubule nucleating activity of centrosome through mitogen-activated protein kinase/extracellular signal-regulated kinase kinase/extracellular signal-regulated kinase (MEK/ERK) signaling
.
J. Biol. Chem.
287
,
2520
-
2530
.
Delaval
,
B.
,
Bright
,
A.
,
Lawson
,
N. D.
and
Doxsey
,
S.
(
2011
).
The cilia protein IFT88 is required for spindle orientation in mitosis
.
Nat. Cell Biol.
13
,
461
-
468
.
Dell
,
K. R.
,
Turck
,
C. W.
and
Vale
,
R. D.
(
2000
).
Mitotic phosphorylation of the dynein light intermediate chain is mediated by cdc2 kinase
.
Traffic
1
,
38
-
44
.
Dephoure
,
N.
,
Zhou
,
C.
,
Villen
,
J.
,
Beausoleil
,
S. A.
,
Bakalarski
,
C. E.
,
Elledge
,
S. J.
and
Gygi
,
S. P.
(
2008
).
A quantitative atlas of mitotic phosphorylation
.
Proc. Natl. Acad. Sci. USA
105
,
10762
-
10767
.
di Pietro
,
F.
,
Echard
,
A.
and
Morin
,
X.
(
2016
).
Regulation of mitotic spindle orientation: an integrated view
.
EMBO Rep.
17
,
1106
-
1130
.
di Pietro
,
F.
,
Valon
,
L.
,
Li
,
Y.
,
Goïame
,
R.
,
Genovesio
,
A.
and
Morin
,
X.
(
2017
).
An RNAi screen in a novel model of oriented divisions identifies the actin-capping protein Zβ as an essential regulator of spindle orientation
.
Curr. Biol.
27
,
2452
-
2464.e8
.
Doxsey
,
S. J.
,
Stein
,
P.
,
Evans
,
L.
,
Calarco
,
P. D.
and
Kirschner
,
M.
(
1994
).
Pericentrin, a highly conserved centrosome protein involved in microtubule organization
.
Cell
76
,
639
-
650
.
Gatlin
,
J. C.
and
Bloom
,
K.
(
2010
).
Microtubule motors in eukaryotic spindle assembly and maintenance
.
Semin. Cell Dev. Biol.
21
,
248
-
254
.
Glotzer
,
M.
(
2009
).
The 3Ms of central spindle assembly: microtubules, motors and MAPs
.
Nat. Rev. Mol. Cell Biol.
10
,
9
-
20
.
Gonçalves
,
J. C.
,
Dantas
,
T. J.
and
Vallee
,
R. B.
(
2019
).
Distinct roles for dynein light intermediate chains in neurogenesis, migration, and terminal somal translocation
.
J. Cell Biol.
218
,
808
-
819
.
Gönczy
,
P.
,
Pichler
,
S.
,
Kirkham
,
M.
and
Hyman
,
A. A.
(
1999
).
Cytoplasmic dynein is required for distinct aspects of MTOC positioning, including centrosome separation, in the one cell stage Caenorhabditis elegans embryo
.
J. Cell Biol.
147
,
135
-
150
.
Han
,
M.-Z.
,
Xu
,
R.
,
Xu
,
Y.-Y.
,
Zhang
,
X.
,
Ni
,
S.-L.
,
Huang
,
B.
,
Chen
,
A.-J.
,
Wei
,
Y.-Z.
,
Wang
,
S.
,
Li
,
W.-J.
, et al. 
(
2017
).
TAGLN2 is a candidate prognostic biomarker promoting tumorigenesis in human gliomas
.
J. Exp. Clin. Cancer Res.
36
,
155
.
Hehnly
,
H.
and
Doxsey
,
S.
(
2014
).
Rab11 endosomes contribute to mitotic spindle organization and orientation
.
Dev. Cell
28
,
497
-
507
.
Jones
,
L. A.
,
Villemant
,
C.
,
Starborg
,
T.
,
Salter
,
A.
,
Goddard
,
G.
,
Ruane
,
P.
,
Woodman
,
P. G.
,
Papalopulu
,
N.
,
Woolner
,
S.
and
Allan
,
V. J.
(
2014
).
Dynein light intermediate chains maintain spindle bipolarity by functioning in centriole cohesion
.
J. Cell Biol.
207
,
499
-
516
.
Joshi
,
H. C.
,
Palacios
,
M. J.
,
McNamara
,
L.
and
Cleveland
,
D. W.
(
1992
).
γ-tubulin is a centrosomal protein required for cell cycle-dependent microtubule nucleation
.
Nature
356
,
80
-
83
.
Juschke
,
C.
,
Xie
,
Y.
,
Postiglione
,
M. P.
and
Knoblich
,
J. A.
(
2014
).
Analysis and modeling of mitotic spindle orientations in three dimensions
.
Proc. Natl. Acad. Sci. USA
111
,
1014
-
1019
.
Kim
,
H.-R.
,
Lee
,
H.-S.
,
Lee
,
K.-S.
,
Jung
,
I. D.
,
Kwon
,
M.-S.
,
Kim
,
C.-H.
,
Kim
,
S.-M.
,
Yoon
,
M.-H.
,
Park
,
Y.-M.
,
Lee
,
S.-M.
, et al. 
(
2017
).
An essential role for TAGLN2 in phagocytosis of lipopolysaccharide-activated macrophages
.
Sci. Rep.
7
,
8731
.
Kiyomitsu
,
T.
and
Cheeseman
,
I. M.
(
2012
).
Chromosome- and spindle-pole-derived signals generate an intrinsic code for spindle position and orientation
.
Nat. Cell Biol.
14
,
311
-
317
.
Kollman
,
J. M.
,
Merdes
,
A.
,
Mourey
,
L.
and
Agard
,
D. A.
(
2011
).
Microtubule nucleation by γ-tubulin complexes
.
Nat. Rev. Mol. Cell Biol.
12
,
709
-
721
.
Kotak
,
S.
and
Gönczy
,
P.
(
2013
).
Mechanisms of spindle positioning: cortical force generators in the limelight
.
Curr. Opin. Cell Biol.
25
,
741
-
748
.
Kotak
,
S.
and
Gönczy
,
P.
(
2014
).
NuMA phosphorylation dictates dynein-dependent spindle positioning
.
Cell Cycle
13
,
177
-
178
.
Kotak
,
S.
,
Busso
,
C.
and
Gönczy
,
P.
(
2013
).
NuMA phosphorylation by CDK1 couples mitotic progression with cortical dynein function
.
EMBO J.
32
,
2517
-
2529
.
Kotak
,
S.
,
Busso
,
C.
and
Gönczy
,
P.
(
2014
).
NuMA interacts with phosphoinositides and links the mitotic spindle with the plasma membrane
.
EMBO J.
33
,
1815
-
1830
.
Kschonsak
,
Y. T.
and
Hoffmann
,
I.
(
2018
).
Activated ezrin controls MISP levels to ensure correct NuMA polarization and spindle orientation
.
J. Cell Sci.
131
,
jcs214544
.
Kulukian
,
A.
and
Fuchs
,
E.
(
2013
).
Spindle orientation and epidermal morphogenesis
.
Philos. Trans. R. Soc. Lond. B Biol. Sci.
368
,
20130016
.
Lecland
,
N.
and
Lüders
,
J.
(
2014
).
The dynamics of microtubule minus ends in the human mitotic spindle
.
Nat. Cell Biol.
16
,
770
-
778
.
Leguy
,
R.
,
Melki
,
R.
,
Pantaloni
,
D.
and
Carlier
,
M.-F.
(
2000
).
Monomeric γ-tubulin nucleates microtubules
.
J. Biol. Chem.
275
,
21975
-
21980
.
Ma
,
S.
,
Triviños-Lagos
,
L.
,
Gräf
,
R.
and
Chisholm
,
R. L.
(
1999
).
Dynein intermediate chain mediated dynein-dynactin interaction is required for interphase microtubule organization and centrosome replication and separation in Dictyostelium
.
J. Cell Biol.
147
,
1261
-
1274
.
Machicoane
,
M.
,
de Frutos
,
C. A.
,
Fink
,
J.
,
Rocancourt
,
M.
,
Lombardi
,
Y.
,
Garel
,
S.
,
Piel
,
M.
and
Echard
,
A.
(
2014
).
SLK-dependent activation of ERMs controls LGN-NuMA localization and spindle orientation
.
J. Cell Biol.
205
,
791
-
799
.
Mahale
,
S.
,
Kumar
,
M.
,
Sharma
,
A.
,
Babu
,
A.
,
Ranjan
,
S.
,
Sachidanandan
,
C.
and
Mylavarapu
,
S. V. S.
(
2016a
).
The light intermediate chain 2 subpopulation of dynein regulates mitotic spindle orientation
.
Sci. Rep.
6
,
22
.
Mahale
,
S. P.
,
Sharma
,
A.
and
Mylavarapu
,
S. V. S.
(
2016b
).
Dynein light intermediate chain 2 facilitates the metaphase to anaphase transition by inactivating the spindle assembly checkpoint
.
PLoS ONE
11
,
e0159646
.
Merdes
,
A.
,
Heald
,
R.
,
Samejima
,
K.
,
Earnshaw
,
W. C.
and
Cleveland
,
D. W.
(
2000
).
Formation of spindle poles by dynein/dynactin-dependent transport of NuMA
.
J. Cell Biol.
149
,
851
-
862
.
Morin
,
X.
and
Bellaiche
,
Y.
(
2011
).
Mitotic spindle orientation in asymmetric and symmetric cell divisions during animal development
.
Dev. Cell
21
,
102
-
119
.
Moudjou
,
M.
,
Bordes
,
N.
,
Paintrand
,
M.
and
Bornens
,
M.
(
1996
).
gamma-Tubulin in mammalian cells: the centrosomal and the cytosolic forms
.
J. Cell Sci.
109
,
875
-
887
.
Na
,
B.-R.
,
Kim
,
H.-R.
,
Piragyte
,
I.
,
Oh
,
H.-M.
,
Kwon
,
M.-S.
,
Akber
,
U.
,
Lee
,
H.-S.
,
Park
,
D.-S.
,
Song
,
W. K.
,
Park
,
Z.-Y.
, et al. 
(
2015
).
TAGLN2 regulates T cell activation by stabilizing the actin cytoskeleton at the immunological synapse
.
J. Cell Biol.
209
,
143
-
162
.
Na
,
B.-R.
,
Kwon
,
M.-S.
,
Chae
,
M.-W.
,
Kim
,
H.-R.
,
Kim
,
C.-H.
,
Jun
,
C.-D.
and
Park
,
Z.-Y.
(
2016
).
Transgelin-2 in B-cells controls T-cell activation by stabilizing T cell - B cell conjugates
.
PLoS ONE
11
,
e0156429
.
Niclas
,
J.
,
Allan
,
V. J.
and
Vale
,
R. D.
(
1996
).
Cell cycle regulation of dynein association with membranes modulates microtubule-based organelle transport
.
J. Cell Biol.
133
,
585
-
593
.
Noatynska
,
A.
,
Gotta
,
M.
and
Meraldi
,
P.
(
2012
).
Mitotic spindle (DIS)orientation and DISease: cause or consequence?
J. Cell Biol.
199
,
1025
-
1035
.
Nousiainen
,
M.
,
Sillje
,
H. H. W.
,
Sauer
,
G.
,
Nigg
,
E. A.
and
Korner
,
R.
(
2006
).
Phosphoproteome analysis of the human mitotic spindle
.
Proc. Natl. Acad. Sci. USA
103
,
5391
-
5396
.
Olsen
,
J. V.
,
Vermeulen
,
M.
,
Santamaria
,
A.
,
Kumar
,
C.
,
Miller
,
M. L.
,
Jensen
,
L. J.
,
Gnad
,
F.
,
Cox
,
J.
,
Jensen
,
T. S.
,
Nigg
,
E. A.
, et al. 
(
2010
).
Quantitative phosphoproteomics reveals widespread full phosphorylation site occupancy during mitosis
.
Sci. Signal.
3
,
ra3
.
Palmer
,
R. E.
,
Sullivan
,
D. S.
,
Huffaker
,
T.
and
Koshland
,
D.
(
1992
).
Role of astral microtubules and actin in spindle orientation and migration in the budding yeast, Saccharomyces cerevisiae
.
J. Cell Biol.
119
,
583
-
593
.
Palmer
,
K. J.
,
Hughes
,
H.
and
Stephens
,
D. J.
(
2009
).
Specificity of cytoplasmic dynein subunits in discrete membrane-trafficking steps
.
Mol. Biol. Cell
20
,
2885
-
2899
.
Pease
,
J. C.
and
Tirnauer
,
J. S.
(
2011
).
Mitotic spindle misorientation in cancer--out of alignment and into the fire
.
J. Cell Sci.
124
,
1007
-
1016
.
Peyre
,
E.
,
Jaouen
,
F.
,
Saadaoui
,
M.
,
Haren
,
L.
,
Merdes
,
A.
,
Durbec
,
P.
and
Morin
,
X.
(
2011
).
A lateral belt of cortical LGN and NuMA guides mitotic spindle movements and planar division in neuroepithelial cells
.
J. Cell Biol.
193
,
141
-
154
.
Piehl
,
M.
,
Tulu
,
U. S.
,
Wadsworth
,
P.
and
Cassimeris
,
L.
(
2004
).
Centrosome maturation: measurement of microtubule nucleation throughout the cell cycle by using GFP-tagged EB1
.
Proc. Natl. Acad. Sci. USA
101
,
1584
-
1588
.
Quyn
,
A. J.
,
Appleton
,
P. L.
,
Carey
,
F. A.
,
Steele
,
R. J.
,
Barker
,
N.
,
Clevers
,
H.
,
Ridgway
,
R. A.
,
Sansom
,
O. J.
and
Näthke
,
I. S.
(
2010
).
Spindle orientation bias in gut epithelial stem cell compartments is lost in precancerous tissue
.
Cell Stem Cell
6
,
175
-
181
.
Raaijmakers
,
J. A.
,
Tanenbaum
,
M. E.
and
Medema
,
R. H.
(
2013
).
Systematic dissection of dynein regulators in mitosis
.
J. Cell Biol.
201
,
201
-
215
.
Raynaud-Messina
,
B.
and
Merdes
,
A.
(
2007
).
γ-tubulin complexes and microtubule organization
.
Curr. Opin. Cell Biol.
19
,
24
-
30
.
Reck-Peterson
,
S. L.
,
Redwine
,
W. B.
,
Vale
,
R. D.
and
Carter
,
A. P.
(
2018
).
The cytoplasmic dynein transport machinery and its many cargoes
.
Nat. Rev. Mol. Cell Biol.
19
,
382
-
398
.
Rivera-Molina
,
F.
and
Toomre
,
D.
(
2013
).
Live-cell imaging of exocyst links its spatiotemporal dynamics to various stages of vesicle fusion
.
J. Cell Biol.
201
,
673
-
680
.
Sana
,
S.
,
Keshri
,
R.
,
Rajeevan
,
A.
,
Kapoor
,
S.
and
Kotak
,
S.
(
2018
).
Plk1 regulates spindle orientation by phosphorylating NuMA in human cells
.
Life Sci Alliance
1
,
e201800223
.
Schaefer
,
M.
,
Shevchenko
,
A.
,
Shevchenko
,
A.
and
Knoblich
,
J. A.
(
2000
).
A protein complex containing Inscuteable and the Gα-binding protein Pins orients asymmetric cell divisions in Drosophila
.
Curr. Biol.
10
,
353
-
362
.
Seldin
,
L.
,
Poulson
,
N. D.
,
Foote
,
H. P.
and
Lechler
,
T.
(
2013
).
NuMA localization, stability, and function in spindle orientation involve 4.1 and Cdk1 interactions
.
Mol. Biol. Cell
24
,
3651
-
3662
.
Seldin
,
L.
,
Muroyama
,
A.
and
Lechler
,
T.
(
2016
).
NuMA-microtubule interactions are critical for spindle orientation and the morphogenesis of diverse epidermal structures
.
eLife
5
,
e12504
.
Shapland
,
C.
,
Hsuan
,
J. J.
,
Totty
,
N. F.
and
Lawson
,
D.
(
1993
).
Purification and properties of transgelin: a transformation and shape change sensitive actin-gelling protein
.
J. Cell Biol.
121
,
1065
-
1073
.
Sivaram
,
M. V. S.
,
Wadzinski
,
T. L.
,
Redick
,
S. D.
,
Manna
,
T.
and
Doxsey
,
S. J.
(
2009
).
Dynein light intermediate chain 1 is required for progress through the spindle assembly checkpoint
.
EMBO J.
28
,
902
-
914
.
Smith
,
D. S.
,
Niethammer
,
M.
,
Ayala
,
R.
,
Zhou
,
Y.
,
Gambello
,
M. J.
,
Wynshaw-Boris
,
A.
and
Tsai
,
L.-H.
(
2000
).
Regulation of cytoplasmic dynein behaviour and microtubule organization by mammalian Lis1
.
Nat. Cell Biol.
2
,
767
-
775
.
Stearns
,
T.
and
Kirschner
,
M.
(
1994
).
In vitro reconstitution of centrosome assembly and function: the central role of γ-tubulin
.
Cell
76
,
623
-
637
.
Stevens
,
D.
,
Gassmann
,
R.
,
Oegema
,
K.
and
Desai
,
A.
(
2011
).
Uncoordinated loss of chromatid cohesion is a common outcome of extended metaphase arrest
.
PLoS ONE
6
,
e22969
.
Stout
,
J. R.
,
Yount
,
A. L.
,
Powers
,
J. A.
,
Leblanc
,
C.
,
Ems-McClung
,
S. C.
and
Walczak
,
C. E.
(
2011
).
Kif18B interacts with EB1 and controls astral microtubule length during mitosis
.
Mol. Biol. Cell
22
,
3070
-
3080
.
Theesfeld
,
C. L.
,
Irazoqui
,
J. E.
,
Bloom
,
K.
and
Lew
,
D. J.
(
1999
).
The role of actin in spindle orientation changes during the Saccharomyces cerevisiae cell cycle
.
J. Cell Biol.
146
,
1019
-
1032
.
Tynan
,
S. H.
,
Purohit
,
A.
,
Doxsey
,
S. J.
and
Vallee
,
R. B.
(
2000
).
Light intermediate chain 1 defines a functional subfraction of cytoplasmic dynein which binds to pericentrin
.
J. Biol. Chem.
275
,
32763
-
32768
.
Wee
,
B.
,
Johnston
,
C. A.
,
Prehoda
,
K. E.
and
Doe
,
C. Q.
(
2011
).
Canoe binds RanGTP to promote PinsTPR/Mud-mediated spindle orientation
.
J. Cell Biol.
195
,
369
-
376
.
Wiese
,
C.
and
Zheng
,
Y.
(
2006
).
Microtubule nucleation: γ-tubulin and beyond
.
J. Cell Sci.
119
,
4143
-
4153
.
Woodard
,
G. E.
,
Huang
,
N.-N.
,
Cho
,
H.
,
Miki
,
T.
,
Tall
,
G. G.
and
Kehrl
,
J. H.
(
2010
).
Ric-8A and Giα recruit LGN, NuMA, and dynein to the cell cortex to help orient the mitotic spindle
.
Mol. Cell. Biol.
30
,
3519
-
3530
.
Woolner
,
S.
,
O'Brien
,
L. L.
,
Wiese
,
C.
and
Bement
,
W. M.
(
2008
).
Myosin-10 and actin filaments are essential for mitotic spindle function
.
J. Cell Biol.
182
,
77
-
88
.
Yamashita
,
N.
,
Morita
,
M.
,
Legant
,
W. R.
,
Chen
,
B.-C.
,
Betzig
,
E.
,
Yokota
,
H.
and
Mimori-Kiyosue
,
Y.
(
2015
).
Three-dimensional tracking of plus-tips by lattice light-sheet microscopy permits the quantification of microtubule growth trajectories within the mitotic apparatus
.
J. Biomed. Opt.
20
,
101206
.
Young
,
A.
,
Dictenberg
,
J. B.
,
Purohit
,
A.
,
Tuft
,
R.
and
Doxsey
,
S. J.
(
2000
).
Cytoplasmic dynein-mediated assembly of pericentrin and γ tubulin onto centrosomes
.
Mol. Biol. Cell
11
,
2047
-
2056
.
Yu
,
F.
,
Morin
,
X.
,
Cai
,
Y.
,
Yang
,
X.
and
Chia
,
W.
(
2000
).
Analysis of partner of inscuteable, a novel player of Drosophila asymmetric divisions, reveals two distinct steps in inscuteable apical localization
.
Cell
100
,
399
-
409
.
Zheng
,
Z.
,
Zhu
,
H.
,
Wan
,
Q.
,
Liu
,
J.
,
Xiao
,
Z.
,
Siderovski
,
D. P.
and
Du
,
Q.
(
2010
).
LGN regulates mitotic spindle orientation during epithelial morphogenesis
.
J. Cell Biol.
189
,
275
-
288
.
Zheng
,
Z.
,
Wan
,
Q.
,
Liu
,
J.
,
Zhu
,
H.
,
Chu
,
X.
and
Du
,
Q.
(
2013
).
Evidence for dynein and astral microtubule-mediated cortical release and transport of Gαi/LGN/NuMA complex in mitotic cells
.
Mol. Biol. Cell
24
,
901
-
913
.
Zhu
,
M.
,
Settele
,
F.
,
Kotak
,
S.
,
Sanchez-Pulido
,
L.
,
Ehret
,
L.
,
Ponting
,
C. P.
,
Gönczy
,
P.
and
Hoffmann
,
I.
(
2013
).
MISP is a novel Plk1 substrate required for proper spindle orientation and mitotic progression
.
J. Cell Biol.
200
,
773
-
787
.

Competing interests

The authors declare no competing or financial interests.

Supplementary information