An unconventional TOG domain is required for CLASP localization

Abstract

Cytoplasmic linker-associated proteins (CLASPs) form a conserved family of microtubule-associated proteins (MAPs) that maintain microtubules in a growing state by promoting rescue while suppressing catastrophe.¹ CLASP function involves an ordered array of tumor overexpressed gene (TOG) domains and binding to multiple protein partners via a conserved C-terminal domain (CTD).^2,3 In migrating cells, CLASPs concentrate at the cortex near focal adhesions as part of cortical microtubule stabilization complexes (CMSCs), via binding of their CTD to the focal adhesion protein PHLDB2/LL5β.^4,5 Cortical CLASPs also stabilize a subset of microtubules, which stimulate focal adhesion turnover and generate a polarized microtubule network toward the leading edge of migrating cells. CLASPs are also recruited to the trans-Golgi network (TGN) via an interaction between their CTD and the Golgin protein GCC185.⁶ This allows microtubule growth toward the leading edge of migrating cells, which is required for Golgi organization, polarized intracellular transport, and cell motility.⁷ In dividing cells, CLASPs are essential at kinetochores for efficient chromosome segregation and anaphase spindle integrity.^8,9 Both CENP-E and ASTRIN bind and target CLASPs to kinetochores,^10,11 although the CLASP domain required for this interaction is not known. Despite its high evolutionary conservation, the CTD remains structurally uncharacterized. Here, we find that the CTD can be structurally modeled as a TOG domain. We identify a surface-exposed and conserved arginine residue essential for CLASP CTD interaction with partner proteins. Together, our results provide a structural mechanism by which the CLASP CTD directs diverse sub-cellular localizations throughout the cell cycle.

Keywords: CLASP; Golgi; TOG domain; cell division; focal adhesion; microtubules.

Figure 1

The CTD of CLASPs folds like a TOG domain (A) Schematic of the HsCLASP1 protein domain delimitations. (B) Alignment and secondary structure comparison of the CTDs from HsCLASP1, HsCLASP2, and CeCLS-2, and of the TOG2 of HsCLASP1. The secondary structure of HsCLASP1 TOG2 is assigned from the crystal structure and that of HsCLASP1 CTD from the AlphaFold2 model. (C and D) AlphaFold2 3D structure prediction of HsCLASP1 CTD, color-coded by either (C) HEAT repeats (HRs) A–F or (D) amino acid conservation. Right: zoom on the C-terminal-conserved patch highlighting the R1481 residue. Amino acid conservation is color-coded (red, high conservation, yellow, low conservation). (E) Still images of spindle in C. elegans zygotes expressing CeCLS-2-GFP (orange) or CeCLS-2^ΔCTD::HsCTD::GFP (blue) upon depletion of endogenous CLS-2. Scale bars, 5 μm. (F) Quantification of mean WT CeCLS-2-GFP (orange) or CeCLS-2^ΔCTD::HsCTD::GFP (blue) along the mitotic spindle upon depletion of endogenous CLS-2. Error bars, SEM. Sample sizes (n metaphase plate) are indicated on the graph. (G) Quantification of embryonic viability in the indicated conditions. Sample sizes (n embryos) are on the left of each bar plot. See also Figure S1.

Figure 2

A conserved arginine in the CTD of CLASPs is essential for their proper sub-cellular localizations (A–F) Immunofluorescence images of DLD-1 cells transiently expressing wild-type (WT) or R1481A EGFP-HsCLASP1, and stained for (A) the trans-Golgi marker TGN46, (C) the focal adhesion marker paxillin, or (E) the centromeric marker ACA (bottom right: zoom on one kinetochore pair; scale bars, 1 μm). Scale bars, 5 μm. White arrowheads indicate centrosomes (localized with an anti-pericentrin antibody, not shown). Quantification of mean intensity of WT or R1481A EGFP-HsCLASP1 at (B) the Golgi, (D) the cell cortex (CMSCs), and (F) kinetochores. Error bars: SD for (B) and (D), SEM for (F). Sample sizes (n areas, N kinetochores) are indicated on each graph. Unpaired t tests, p < 0.0001. (G) Still images of C. elegans zygotes expressing WT or R970A CeCLS-2-GFP and the chromosomal marker mCherry-H2B. Scale bars, 5 μm. (H) Quantification of WT or R970A CeCLS-2-GFP along the mitotic spindle. Error bars, SD. Sample sizes (n metaphase plate) are indicated on each graph. See also Figure S2.

Figure 3

Identification of HsCLASP1 CTD interacting proteins (A) Schematic of the strategy used to compare WT or R1481A HsCLASP1 CTD protein partners using EGFP immunoprecipitation followed by mass-spectrometric identification. (B) Result of LC-MS/MS analysis from four independent experiments. Volcano plot of −log₁₀ p values against log₂ fold change (R1481A versus WT). Proteins significantly enriched in the WT but not the R1481A mutant (log₂ enrichment > 1, p < 0.05) are indicated in pink, with selected proteins highlighted. (C) Result from two independent yeast-two-hybrid screens from a human fibroblast or a human placental library, respectively. The predicted biological score (PBS) of each interaction is ranked from A (very high confidence) to D (moderate confidence). Proteins are highlighted as significantly enriched (pink) or not enriched (gray) in the WT compared with the R1481A mutant (identified in the proteomic analysis). Proteins not found in the proteomic analysis are displayed in black. (D) Venn diagram displaying the overlaps between hits from the fibroblast and placenta Y2H screens, and the results from the mass spectrometry analysis (only known CLASP interactors, or proteins also identified in at least one of the Y2H screen, are included). See also Figure S3 and Data S1.