Multiple domains of human CLASP contribute to microtubule dynamics and organization in vitro and in Xenopus egg extracts

Abstract

Cytoplasmic linker associated proteins (CLASPs) comprise a class of microtubule (MT) plus end-binding proteins (+TIPs) that contribute to the dynamics and organization of MTs during many cellular processes, among them mitosis. Human CLASP proteins contain multiple MT-binding domains, including tumor over-expressed gene (TOG) domains, and a Ser-x-Ile-Pro (SxIP) motif known to target some +TIPs though interaction with end-binding protein 1 (EB1). However, how individual domains contribute to CLASP function is poorly understood. We generated full-length recombinant human CLASP1 and a series of truncation mutants and found that two N-terminal TOG domains make the strongest contribution to MT polymerization and bundling, but also identified a third TOG domain that further contributes to CLASP activity. Plus end tracking by CLASP requires the SxIP motif and interaction with EB1. The C-terminal coiled-coil domain mediates dimerization and association with many other factors, including the kinetochore motor centromere protein E (CENP-E), and the chromokinesin Xkid. Only the full-length protein was able to rescue spindle assembly in Xenopus egg extracts depleted of endogenous CLASP. Deletion of the C-terminal domain caused aberrant MT polymerization and dramatic spindle phenotypes, even with small amounts of added protein, indicating that proper localization of CLASP activity is essential to control MT polymerization during mitosis.

Fig. 1

CLASP is a thin, elongated protein that can homodimerize. (A) Schematic of human CLASP1 domain architecture and the constructs generated for this study. (B) Representative sypro-ruby stained gel of purified full-length CLASP and smaller constructs, which were all expressed as StrepII-CLASP-GFP-10XHis fusions. (C) Electron micrograph of negatively stained, full-length CLASP showing elongated particles (see right panels for individual examples). (D) Elution profiles of two CLASP proteins after gel filtration. In the profile for full-length CLASP (left) there is a peak at fractions 12–18 corresponding to a MW of 180 kD that likely represents monomeric CLASP, and a second peak (arrow) for fractions 2–5 that corresponds to a protein MW of ∼350 kD and presumably represents dimeric CLASP. The elution profile for the CLASP1-1171 fragment (right) lacks this second peak, indicating that the C-terminal coiled-coil region is required for CLASP dimerization.

Fig. 2

CLASP domains contribute differentially to MT binding, bundling, and rates of tubulin polymerization. (A) Representative gels of MT pelleting assays for CLASP1-662 and CLASP 1-1171 constructs. CLASP (0.5 μM) construct was added to 0–4.0 μM of taxol-stabilized MTs. Binding reaction mixtures were centrifuged, and both supernatant (S) and pellet (P) recovered and analyzed by SDS-PAGE (10% Bis-Tris) and visualized with sypro-ruby stain. (B) Affinity curves for the five different CLASP constructs derived from the pelleting assay gels. Band intensities were quantified using ImageJ (NIH) and plotted to construct an affinity curve from which the kd was estimated. Standard errors (not shown) were large and therefore only relative affinities should be evaluated. (C) 90° light scattering curves at 350 nm absorbance showing tubulin copolymerization with various CLASP constructs. Data was collected at 30 s intervals. Inset shows an EM micrograph of a MT bundle formed in the presence of the CLASP1-1171 construct, which also contributes to absorbance in this assay.

Fig. 3

CLASP SxIP motifs mediate MT plus tip tracking though EB1. CLASP-GFP fusion proteins were added to human centrosomes in metaphase-arrested Xenopus egg extract. (A) Time-lapse images of full-length CLASP, CLASP1-662, and CLASP662-1463 in the assay. EB1 (4 μM) was added to enhance visualization of MT plus end association. Arrows indicate individual MTs or bundles (Supporting Information Movies S1–S3). (B) Centrosome tracking assay for construct CLASP662-1463 in EB1-depleted egg extract, and depleted extract to which recombinant EB1 was added back at the endogenous concentration of approximately 1 μM (Supporting Information Movies S4 and S5). (C) Summary of results for all tracking experiments with different CLASP constructs. All time-lapse images were collected at 2.5 s intervals. Scale bars = 10 μm.

Fig. 4

CLASP interacts with multiple binding partners through different domains. (A) Identification of CLASP interacting partners and binding domains. Full-length and shorter constructs of CLASP were used to pull down interacting proteins from Xenopus egg extract. For each CLASP construct, potential binding partners were separated by SDS-PAGE and probed by Western blot using available antibodies. Blots reveal the chromokinesin XKID as a novel partner that interacts with CLASP most likely through the C-terminal domain, similarly to XCENP-E and CLIP-170. PRC1 interacts with CLASP constructs that contain the S/R region and the putative crTOG3 domain. EB1 binds to all fragments containing the conserved SxIP motif. (B) Pelleting assays of full-length CLASP binding to MTs in the presence of increasing NaCl concentrations analyzed by SDS-PAGE (10% Bis-Tris) and stained with Sypro Ruby. (C) Pelleting assays of full-length CLASP binding to subtilisin treated (+) and untreated MTs (−) analyzed by SDS-PAGE (12% Bis-Tris) and stained with Sypro Ruby.

Fig. 5

Rescue of Xorbit depletion with full-length CLASP. (A) Fluorescence images of spindles in egg extracts that were depleted with control antibodies (Mock depletion) or Xorbit antibodies followed by addition of buffer control (Xorbit depletion) or 1.0 μM full-length CLASP protein. (B) Western blot analysis of depletion and add-back reactions indicating that the endogenous CLASP concentration is approximately 1.5 μM.

Fig. 6

Exogenous CLASP protein lacking the C-terminal domain causes spindle MT polymerization and organization defects in Xenopus egg extract. (A) Representative images of major spindle phenotypes with addition of 0.2 μM CLASP1-1171 upon entry into mitosis. Scale bar = 10 μM. (B) Quantification of phenotype distribution comparing buffer control. Greater than 50 spindles were counted for each condition in three separate experiments.

Fig. 7

Exogenous CLASP protein lacking the N-terminal TOG domains causes defects in chromosome congression. (A) Representative images showing CCDs with addition of 0.2 μM CLASP662-1463 upon entry into mitosis. Scale bar = 10 μM. (B) Quantification of defects comparing buffer control. Greater than 50 spindles were counted for each condition in three separate experiments.