Phase separation of TPX2 enhances and spatially coordinates microtubule nucleation - PubMed (original) (raw)
Phase separation of TPX2 enhances and spatially coordinates microtubule nucleation
Matthew R King et al. Nat Commun. 2020.
Abstract
Phase separation of substrates and effectors is proposed to enhance biological reaction rates and efficiency. Targeting protein for Xklp2 (TPX2) is an effector of branching microtubule nucleation in spindles and functions with the substrate tubulin by an unknown mechanism. Here we show that TPX2 phase separates into a co-condensate with tubulin, which mediates microtubule nucleation in vitro and in isolated cytosol. TPX2-tubulin co-condensation preferentially occurs on pre-existing microtubules, the site of branching microtubule nucleation, at the endogenous and physiologically relevant concentration of TPX2. Truncation and chimera versions of TPX2 suggest that TPX2-tubulin co-condensation enhances the efficiency of TPX2-mediated branching microtubule nucleation. Finally, the known inhibitor of TPX2, the importin-α/β heterodimer, regulates TPX2 condensation in vitro and, consequently, branching microtubule nucleation activity in isolated cytosol. Our study demonstrates how regulated phase separation can simultaneously enhance reaction efficiency and spatially coordinate microtubule nucleation, which may facilitate rapid and accurate spindle formation.
Conflict of interest statement
The authors declare no competing interests.
Figures
Fig. 1. TPX2 forms a co-condensate with tubulin in vitro and in the cytosol.
All scale bars are 3 μm. a Secondary structure and intrinsic disorder predictions in TPX2. b Epifluorescent image of GFP-TPX2 (green) condensates (see Supplementary Movie 1) (left) and DIC image of untagged TPX2 condensates (right), both at a final concentration of 1 µM. Representative of six experimental replicates. c Schematic for assaying phase separation—TPX2 (with or without other proteins) is purified and maintained in a high-salt (0.5 M) buffer and this is transferred at 1:4 volume:volume into a no salt buffer to achieve physiological salt levels (0.1 M). d Epifluorescent image of GFP-TPX2 (green) condensates prepared with Cy5-labeled tubulin (magenta) (both at 4 µM) prepared as shown in c and imaged in a flow chamber (see Supplementary Fig. 1d for control). Representative of six experimental replicates. e TIRF image of TPX2-Tubulin co-condensates (green and magenta, 1 and 10 µM, respectively) prepared in MT polymerization buffer in a flow chamber, 18 min after reaction started. Representative of three experimental replicates. Partition coefficients for d and e are mean values (points) with ± 1 SD as error bars from 225 and 170 condensates, respectively. f Experimental setup for g—pre-formed TPX2 condensates are overlaid with Xenopus egg cytosol containing fluorescent tubulin. g Oblique-TIRF microscopy of GFP-TPX2 (green) and tubulin (Alexa568-labeled—red) taken 5 min after reaction started (minutes:seconds). h In the same experiment as shown in Fig. 1g, the tubulin channel imaged over time (minutes:seconds) and depicted. Data representative of three experimental replicates. i Quantification of integrated tubulin signal from indicated areas corresponding to initial condensates (gray) and MT fan structures (blue). Mean values shown as circles with ± 1 SD shown as error bars from n = 5 technical replicates. Source data are provided as a Source Data file.
Fig. 2. TPX2 preferentially co-condenses with tubulin on MTs.
a Experimental setup for b and c: non-phase-separated GFP-TPX2 (50 nM) and Alexa568-labeled tubulin mixed with Xenopus meiotic egg cytosol and imaged via oblique-TIRF microscopy. b GFP-TPX2 (green) localization to growing MT network imaged over time (minutes:seconds) in the cytosol (see Supplementary Movie 2), representative of three experimental replicates. Microtubules are labeled red and growing plus-tips are blue. All scale bars are 3 µm, unless indicated. c GFP-TPX2 (green) - and Alexa568-labeled tubulin (red) colocalization in the cytosol treated with Nocodazole to prevent MT polymerization, representative of five experimental replicates. Images taken 10 min into reaction (see Supplementary Movie 3). Scale bar is 3 µm. TPX2 was 50 nM and tubulin was 2 µM. d TIRF images (contrast-optimized), representative of four experimental replicates, and e quantification of GFP-TPX2 (green) and Cy5-labeled tubulin (magenta) (at indicated concentrations, both proteins equimolar) localized to pre-formed microtubules (Alexa568-labeled, GMPCPP stabilized—red) and spun down onto coverslips, fixed, and imaged; scale bar, 3 µm. f Oblique-TIRF images of only Cy5-labeled tubulin (magenta) condensed with GFP-TPX2 (not shown) either in solution (top panel) or on a pre-formed MT (lower panel—MT not shown) at concentrations shown. Images displayed at matched brightness and contrast. Enhanced contrast of 10 nM and 25 nM images shown is shown in the box. Scale bar, 2 µm. g Quantification of relative tubulin signal from TPX2-tubulin co-condensates in solution (blue curve) and on microtubules (purple curve) at concentrations shown in e. Mean (points) and SEM (error bars) of three replicate experiments (error bars) shown, n of microtubules per condition are: 145, 216, 114, 217, 243, 335, 309, and 309 for 1, 2.5, 5, 10, 25, 50, 100, and 250 nM [TPX2/TB], respectively; n of condensates per condition are: 280, 225, 266, and 282 for 10, 25, 50, 100, and 250 nM [TPX2/TB], respectively. Endogenous concentration range of TPX2 (q = 25–100 nM) indicated. h Schematic of experimental setup—live observation of TPX2 (green) and soluble tubulin (magenta) localization to a stable microtubule (red) attached to a coverslip-bottomed well. i TIRF images of a time series of GFP-TPX2 (green) and soluble Cy5-labeled tubulin (magenta) localization to a GMPCPP-stabilized Alexa568-labeled microtubule seed (red) (see Supplementary Movie 4). Images are contrast enhanced to show early non-association events and are representative of three experimental replicates. Arrowheads indicate fusion of two condensates. Time in minutes:seconds, 00:00 corresponds to addition of TPX2 into the well. Scale bar, 2 µm. Source data are provided as a Source Data file.
Fig. 3. TPX2-tubulin co-condense at the physiologically relevant concentration of TPX2.
a TIRF images and b quantification of TPX2-mediated branching MT nucleation in Xenopus meiotic cytosol at indicated concentrations of TPX2. Shown images were taken at 1000 s (indicated). Cy5-labeled tubulin (red) and mCherry-EB1 (green) highlight microtubules and growing microtubule plus ends, respectively. Scale bar, 10 µm. See also Supplementary Movie 5. c Rate of MT nucleation as a function of TPX2 concentration for four independent replicates of data shown in a and b. Source data are provided as a Source Data file. See also Table 1. Rates normalized to maximum rate within an experiment. Line of best fit shown and approximate physiologically relevant range (25–100 nM) highlighted. Endogenous concentration range of TPX2 (30–100 nM) indicated (also in Fig. 2e).
Fig. 4. TPX2 N-terminus enhances condensation and MT nucleation efficiency.
a Schematic of full-length TPX2 and various N-terminal and C-terminal truncation constructs used. b Epifluorescent images of GFP-TPX2 (green) and Cy5-labeled tubulin (magenta) co-condensates at 1 µM and 10 µM (equimolar concentration) for indicated TPX2 truncation construct. Representative of three experimental replicates. c Quantification of relative tubulin signal in co-condensate (partition coefficient) as a function of TPX2 concentration. Mean values (points) with ± 1 SD as error bars from a representative experiment are plotted and a line was fit; n of condensates per condition are: NT_1-319—406, 572, 278, 217, 292, 276, and 268 for 250, 500 nM, 1, 2.5, 5, 10, and 25 μM [TPX2/TB], respectively; NT_1-480—254, 49, 256, 286, 283, 273, 271, and 198 for 100, 250, 500 nM, 1, 2.5, 5, 10, and 25 μM [TPX2/TB], respectively; CT_319-716—275, 323, 296, and 297 for 2.5, 5, 10, and 25 μM [TPX2/TB], respectively; CT_480-716—327 and 253 for 10 and 25 μM [TPX2/TB], respectively. See also Supplementary Fig. 5a–d for partition coefficient measures of GFP-TPX2 signal. d Rate of MT nucleation as a function of TPX2 concentration for indicated constructs. Full-length data previously shown (Fig. 3c). Rates for each concentration of a given construct are normalized to the maximum rate of that construct (absolute maximum rates between all constructs are within a twofold range). Lines of best fit shown. See also Supplementary Fig. 5e, f for individual rate curves and Supplementary Movie 6. e Different efficiencies of MT nucleation for the full-length and CT_480-716 TPX2 construct. Efficiency values are the inverse of the TPX2 concentration at which half the maximum rate of MT nucleation is achieved ([TPX2]−1 at Rate1/2Max); efficiency values were normalized to full length (efficiency of 1). The difference (fold change) in efficiencies is shown (∆). Source data are provided as a Source Data file.
Fig. 5. TPX2-tubulin co-condensation underlies efficient branching MT nucleation.
a Schematic of chimera design: the endogenous N-terminal 1–480aa of TPX2 are replaced with exogenous domains containing distinct features shown in table. b Quantification of tubulin partition coefficient as a function of TPX2 concentration. For partition coefficient graphs, mean values (points) with ±1 SD as error bars from a representative experiment are plotted and a line was fit; n of condensates per condition are: IDR_NoTB—111, 80, 142, 99, 198, and 198 for 50, 100, 250, 500 nM, 1, and 2.5 μM [TPX2/TB], respectively; NoIDR_TB—198, 136, and 86 for 5, 10, and 25 μM [TPX2/TB], respectively; IDR_TB—87, 122, 153, 198, and 198 for 100, 250, 500 nM, 1, and 2.5 μM [TPX2/TB], respectively; Syn_Pos—326, 302, 100, 261, 270, and 323 for 50, 100, 250, 500 nM, 1 and 2.5 μM [TPX2/TB], respectively. See also Supplementary Figs. 6a, c, f and 7e for partition coefficients of TPX2. c Rate of MT nucleation as a function of TPX2 concentration for indicated constructs. Rates for each concentration of a given construct are normalized to the maximum rate of that construct (absolute maximum rates between all constructs are in a ±2× range). Lines of best fit shown. Data for full length (Figs. 2e and 3d) and CT_TPX2 (Fig. 4d) were previously shown. See also Supplementary Fig. 6b, d, e, g and 7f for individual rate curves and Supplementary Movies S7 and S8. Source data are provided as a Source Data file. d Summary table of results.MT nucleation efficiencies are relative to full-length TPX2.
Fig. 6. Importins α/β inhibit TPX2 condensation and activity.
a Epifluorescent images of TPX2-tubulin co-condensates (green and magenta, respectively) in vitro prepared with importins-α/β at indicated excess (0 × = no importins-α/β) TPX2 and tubulin both at 500 nM. Scale bar, 1 µm. b TIRF Images of TPX2-mediated MT nucleation in Xenopus meiotic cytosol with TPX2 and importins-α/β added at 100 nM TPX2 and indicated excess of importins-α/β. Cy5-labeled tubulin (red) and mCherry-EB1 (green) highlight microtubules and growing microtubule plus ends, respectively. Images taken at 1800 s. Scale bar, 10 µm. See Supplementary Movie 9. c Quantification of data in b. d Quantification of relative tubulin signal (partition coefficient) as a function of excess importins-α/β. Mean values (points) with ± 1 SD as error bars are shown, n = 522 and 420 condensates for 0× and 1× importins-α/β, respectively. e Rate of MT nucleation as a function of excess importins-α/β, normalized to 0× importins-α/β. Data pooled from three experimental replicates of (b, c). Line of best fit shown. Source data are provided as a Source Data file.
Fig. 7. Model.
Left-side cartoon: at low concentrations (
<endogenous), tpx2="" localizes="" to="" mts="" but="" does="" not="" recruit="" soluble="" tubulin,="" likely="" due="" electrostatic="" repulsion="" (denoted="" by="" “+”="" and="" “−”).="" right-side="" cartoon:="" at="" high="" concentrations="" of="" (≥endogenous),="" colocalizes="" with="" tubulin="" on="" microtubules.="" center:="" graphical="" abstraction="" data="" demonstrating="" that="" promotes="" branching="" mt="" nucleation="" (in="" the="" cytosol)="" forms="" a="" co-condensate="" vitro)="" in="" switch-like="" manner="" or="" above="" its="" endogenous="" concentration.="" top="" gradient:="" relative="" importin-α="" β="" levels="" existing="" as="" gradient="" around="" chromosomes="" also="" affect="" tpx2-tubulin="" co-condensation="" tpx2-mediated="" nucleation.<="" div=""> </endogenous),>
References
- Su Xiaolei, Ditlev Jonathon A., Hui Enfu, Xing Wenmin, Banjade Sudeep, Okrut Julia, King David S., Taunton Jack, Rosen Michael K., Vale Ronald D. Phase separation of signaling molecules promotes T cell receptor signal transduction. Science. 2016;352(6285):595–599. doi: 10.1126/science.aad9964. - DOI - PMC - PubMed
Publication types
MeSH terms
Substances
LinkOut - more resources
Full Text Sources
Miscellaneous