Katanin contributes to interspecies spindle length scaling in Xenopus

Abstract

Bipolar spindles must separate chromosomes by the appropriate distance during cell division, but mechanisms determining spindle length are poorly understood. Based on a 2D model of meiotic spindle assembly, we predicted that higher localized microtubule (MT) depolymerization rates could generate the shorter spindles observed in egg extracts of X. tropicalis compared to X. laevis. We found that katanin-dependent MT severing was increased in X. tropicalis, which, unlike X. laevis, lacks an inhibitory phosphorylation site in the katanin p60 catalytic subunit. Katanin inhibition lengthened spindles in both species. In X. tropicalis, k-fiber MT bundles that connect to chromosomes at their kinetochores extended through spindle poles, disrupting them. In both X. tropicalis extracts and the spindle simulation, a balance between k-fiber number and MT depolymerization is required to maintain spindle morphology. Thus, mechanisms have evolved in different species to scale spindle size and coordinate regulation of multiple MT populations in order to generate a robust steady-state structure.

Figure 1. Katanin-Dependent MT Severing is Elevated in X. Tropicalis Compared toX. Laevis

(A) Time course showing individual fields from a flow cell MT depolymerization assay in which taxol-stabilized, rhodamine-labeled MTs were incubated with either X. laevis (left panels) or X. tropicalis (right panels) egg extract. Movie S1 shows that MTs disappear much more rapidly in X. tropicalis extract. Scale bar 10 μm. (B) Quantification of MT disappearance measured in % remaining total MT length in X. laevis, X. tropicalis and mixed extracts. L:T indicates the ratio of X. laevis to X. tropicalis extracts. Example of representative data from 3 independent experiments. (C) Quantification of MT disappearance measured in % remaining total MT length in X. tropicalis extract in the presence of control and inhibitory antibodies to kinesin-13s MCAK and Kif2A, and severing factor katanin. Movie S2 shows that MTs persist only in the presence of katanin antibodies. (D) Severing assay in control IgG-depleted compared to p60 katanin-depleted X. tropicalis egg extract. Scale bar 10 μm.

Figure 2. Comparison of X. Laevis and X. Tropicalis Katanin

(A) Western blot of X. laevis and X. tropicalis extracts at different dilutions, probed for p60 katanin, shows that protein levels are similar in the two species. (B) Mixing experiment in which extract from one species was combined with extract from the other that had been depleted with control or katanin antibodies. Quantification of MT disappearance measured in % remaining total MT length indicates that the difference in MT severing activity measured in the two extracts is due to intrinsic differences in katanin itself. Representative data from 3 independent experiments. (C) Immunofluorescence analysis of p60 katanin in X. laevis and X. tropicalis extracts reveals a similar localization, but greater enrichment of katanin at X. tropicalis spindle poles. Upper panels are merged images of MTs (red), DNA (blue) and katanin p60 (green), which is shown individually in the lower panels. (D) Representative quantification of the average ratio of katanin to tubulin fluorescence intensity in line scans from X. laevis and X. tropicalis spindles stained on the same coverslip (N = 18, 25 spindles). (E) Comparison of X. laevis and X. tropicalis amino acid sequences reveals minor differences only in the region of the protein thought responsible for MT binding. The % identity and number of amino acid differences is shown for katanin domains: the p80 binding domain (yellow), the MT binding domain (blue), the AAA ATPase domain (red) and the oligomerization domain (green). See Figure S1A for full alignment. (F) Representative Western blot of MT co-pelleting assay in X. laevis and X. tropicalis egg extracts probed for katanin and tubulin. Taxol-stabilized MTs were incubated with crude egg extracts in the presence of ATPγS to inhibit severing activity. Quantification of the ratio of katanin to tubulin intensity from the pellet from three independent experiments is shown underneath. See Figure S1B for complete data.

Figure 3. Phosphorylation of X. laevis Katanin Inhibits its Activity

(A) Time course showing individual fields from a flow cell MT destabilization assay in which taxol-stabilized, Alexa 488-labeled MTs were incubated with either X. laevis (left panels) or X. tropicalis (right panels) recombinant purified p60. Movie S3 shows that MTs disappear at approximately the same rate. Scale bar 10 μm. See Figure S2A-B for pure MT binding data and Kd estimation. (B) Quantification of MT disappearance measured in % remaining total MT length in the presence of X. laevis (red line) and X. tropicalis (blue line) recombinant p60 at indicated concentrations. See Figure S2C for data showing ATP-dependence of activity. (C) X. laevis or X. tropicalis p60 was added to either X. laevis or X. tropicalis egg extracts in the presence of ATPγS, retrieved using the MBP tag. Incorporated thio-phosphate was modified with PNBM and analyzed by SDS-PAGE and Western blot for total retrieved protein and incorporated nitrobenzyl thio-phosphate ester. X. tropicalis p60 incorporated less labeled phosphate than X. laevis p60, which was further reduced by addition of Aurora kinase inhibitor VX680 to the extracts or mutation of serine 131 to alanine (S131A). See Figure S2D for multiple species sequence alignment of this putative Aurora phosphorylation site. (D) Quantification of thio-phosphate incorporation into X. laevis and X. tropicalis p60. Mean± S.D., N ≥ 2 extracts. (E) In vitro phosphorylation of purified recombinant X. laevis and X. tropicalis p60 and N-terminal MT binding domains by Aurora B. Phosphorylation was detected using Pro-Q Diamond gel stain (see Methods). (F) Quantification of MT disappearance when recombinant control protein (MBP, green line), X. laevis p60 (red line), X. tropicalis p60 (blue line), X. laevis p60 S131E (yellow line), or X. laevis p60 S131A (orange line) was added to X. laevis egg extract in the MT severing assay. Disappearance is measured in % remaining total MT length. Movie S4 shows that both X. tropicalis p60 and X. laevis p60 S131A possess high severing activity. See Figure S2E for bar graph comparing MT half-lives.

Figure 4. MT Severing Scales Spindle Length in a Computational Simulation of Spindle Assembly

(A) Fundamental activities present in of the simulation. Dynamic MTs (red and green) are nucleated throughout the spindle and slide away from the midzone through the action of kinesin-5 (orange). Dynein (cyan) transports small oligomers of NuMA (yellow) to MT minus-ends. NuMA recruits kinesin-13 MT depolymerization (blue gradient) and MT severing activity to the vicinity of MT minus-ends. (B) Shortening of spindles resulting from increasing severing rates. (C) Spindle length (mean ± SD) as a function of severing rate. Color indicates MT structure organization, defined as “bipolarity” in (Loughlin et al., 2010), as strongly bipolar (green) or bipolar (yellow).

Figure 5. Katanin Inhibition Lengthens Spindles and Reveals K-Fibers

(A) Spindle morphologies following addition of control (IgG) or katanin antibodies to either X. laevis or X. tropicalis spindles formed around chromatin-coated beads with quantification of spindle length increases (mean ± SE) upon katanin inhibition in the different spindle types (N = 94, 83 spindles from two extracts for X. laevis and 305, 401 spindles from two extracts for X. tropicalis). Proportional spindle length scaling is observed. (B) Spindle morphologies following addition of control (IgG) or katanin antibodies to either X. laevis or X. tropicalis spindles formed around X. laevis sperm nuclei containing 18 chromosomes with quantification of spindle length increases (mean ± SE) upon katanin inhibition in the different spindle types (N = 116, 144, 97 spindles from three extracts for X. laevis and 263, 65, 152 spindles from three extracts for X. tropicalis). Proportional spindle length scaling is observed in X. laevis reactions but not in X. tropicalis reactions, in which long, stable, wispy MT extensions protrude through spindle poles, shown in Movie S5. Fluorescent speckle microscopy shows MT disappearance at spindle poles in the control, but persistence of stable MTs in the wispy extensions (Movie S6). (C) Inhibition of k-fiber formation prevents the wispy MT phenotype upon katanin inhibition. Image panels are time-lapse movie frames from reactions containing katanin antibodies or a combination of katanin and kinetochore (Nuf2) antibodies. (D) Quantification of pole phenotypes upon addition of Nuf2, katanin, or both antibodies. Inhibiting k-fiber formation with Nuf2 antibodies does not affect spindle poles. Wispy MT extensions are observed upon katanin inhibition, but largely rescued by co-inhibition of Nuf2 (N = 64, 50, 51, 172, 53, 57, 104, 73, 59 spindles in three extracts). (E) Schematic of spindle phenotypes observed upon katanin inhibition.

Figure 6. K-Fiber Number and MT Depolymerization Rate Determine Pole Stability

(A) Schematic showing k-fiber addition to the spindle simulation. K-fiber plus-ends are tethered to the midplane of the spindle and grow at the speed of flux and never catastrophe, while K-fiber minus-ends are indistinguishable from other MT minus-ends. (B) Representative image of simulated spindle with 5 k-fibers (yellow). Scale bar (blue) 10 μm. (C) Quantification of spindle length with decreasing MT depolymerization rates in the simulation reveals a threshold for failure to reach steady state (red x's) in the presence of 5 k-fibers. (D) Quantification of spindle lengths upon addition of k-fibers. Simulations containing greater than 10 k-fibers failed to reach a steady state length within 80 minutes (red x's), indicating a threshold number of k-fibers that overwhelms polar MT depolymerization. (E) Effects on spindle morphology of adding increasing numbers of k-fibers (yellow) to the simulation. Lower numbers of k-fibers (<10) were tolerated and increased spindle length proportionally. Higher numbers caused spindle pole disruption and never reached a steady state length as k-fibers continued to elongate (Movie S7). Scale bar (blue) 10 μm. (F) Quantification of spindle lengths with 5 k-fibers and changing MT severing rate. Spindles maintain pole morphology as severing and depolymerization activity in the simulation can efficiently destabilize the 5 k-fibers. (G) Quantification of spindle lengths with 15 k-fibers and changing MT severing rate. Simulations containing 15 k-fibers failed to reach a steady state length with low amounts of severing (red x's) indicating a threshold number of k-fibers that overwhelms polar MT destabilization. (H) Movie frames from live analysis comparing X. tropicalis spindle morphology in the presence of 10 or 20 k-fibers upon katanin inhibition. Spindles containing two sperm nuclei and 20 chromosomes were infrequent but clearly contain increased DNA content as measured by DAPI fluorescence intensity (data not show). As in the simulation, a smaller number of k-fibers could be accommodated without disrupting spindle structure, whereas with 20 k-fibers pole structure was disrupted and the spindle elongated by 10.2±1.2 μm (mean ± SE; N = 88 normal and 30 pole disrupted spindles in 2 extracts).

Figure 7. Aurora B Phosphorylation at X. Laevis p60 Serine 131 Regulates Spindle Length

(A) Quantification of X. tropicalis spindle lengths in depletion/add-back experiments. P60 depletion leads to longer spindle lengths, which can be partially rescued by adding back X. tropicalis p60 or X. laevis p60S131A, but not MBP, X. laevis p60 or X. laevis p60S131E. Pairwise t-test between MBP and other rescue conditions pvalue <0.0001 except X. laevis p60 S131E (pvalue = 0.48). (mean± SE N = 240-500 spindles from 3 extracts per condition). (B) Quantification of X. tropicalis spindle pole morphology in depletion/add-back experiments. P60 depletion leads to a high percentage of spindles with wispy MT extensions protruding through the spindle pole (see Figure 5B and C), which can be rescued by adding back X. tropicalis p60 or X. laevis p60S131A, but not MBP, X. laevis p60 or X. laevis p60S131E. (mean± SD; N = 3 extracts, at least 150 spindles evaluated per condition). (C and D) Aurora B inhibition decreases spindle length in X. laevis and is rescued by inhibiting p60 (D) but does not decrease spindle length in X. tropicalis (E). Quantification of spindle length in the presence of 5 μm ZM447439 with and without p60 katanin inhibition in the two extracts. (mean ± SD; N > 70 spindles from 2 extracts). Scale bar 10 μm. (E) Model of spindle length scaling through phosphorylation of katanin p60 by Aurora B. P60 (red) is shown in its hexameric ring form.