Talin rod domain–containing protein 1 (TLNRD1) is a novel actin-bundling protein which promotes filopodia formation

Figures

• 

  Figure 1.. TLNRD1 is homologous to talin R7R8. (A) The domain structure of talin. Talin has an N-terminal head domain and a large rod domain comprised of 13 helical bundles, R1–R13 (Goult et al., 2013). The R7 (blue) and R8 (orange) domains are highlighted. The 11 VBSs are shown in pink. (B) Schematic of talin R7R8 (left) and TLNRD1 (right); the four-helix bundle (orange) is inserted into the loop between α3 and α8 helices of the five-helix bundle (blue). The VBS consensus sequence (Gingras et al., 2010) in talin R7 and R8 (pink) is absent from TLNRD1. (C) Crystal structure of TLNRD1-FL reveals an antiparallel, symmetric dimer. The domains of each monomer are labeled as 4h and 5h, and 4h’ and 5h’. (D) Top-down view of C with the helices labeled. The two four-helix bundles dimerize via the extensive interface between helices α6 and α7. (E) Schematic representation of the TLNRD1 dimer. (F) Overlay of the TLNRD1-4H structure (orange) on top of TLNRD1-FL (partially transparent) showing the dimerization interface is identical in both. (G) Overlay of one monomer of TLNRD1 (orange and blue) with talin R7R8 (gray; Protein Data Bank accession no. 2X0C). The individual domains are structurally…

• 

  Figure S1.. Conservation of TLNRD1 and its alignment with talin R7R8 domains. (A) Amino acid sequence alignment of human TLNRD1 (UniProt accession no. Q9H1K6) and talin1 residues 1357–1653 encompassing R7R8 (UniProt accession no. Q9Y490). Domain boundaries are highlighted with orange corresponding to the four-helix domain and blue corresponding to the five-helix domain. (B) Tree diagram representing TLNRD1 presence and absence over evolution. TLNRD1 first appears in choanoflagellates and sponges and is conserved throughout vertebrate evolution. TLNRD1 has been lost from nematodes, arthropods, and Cnidaria. (C) TLNRD1-FL sequence conservation mapped onto the full-length TLNRD1 structure using ConSurf (Ashkenazy et al., 2016). Structure colored according to extent of conservation with blue indicating variability and purple indicating highly conserved residues. Residues important for dimerization (Q213, R246, F250, A260, and E267) are highlighted.

• 

  Figure 2.. Characterization of the TLNRD1 dimer. (A and B) The symmetrical antiparallel TLNRD1 dimer interface mediated by F250. One monomer is shown as a white surface representation, F250 from the opposing monomer (orange) inserts into the pocket on the surface of the other. A salt bridge between R246 and E267’ is shown. (C and D) SEC-MALS analysis of 106 µM TLNRD1-FL WT (black) and F250D (red). (D) Analysis of the molar mass of the major TLNRD1-FL species “a” yields a molecular weight of ∼74 kD. The TLNRD1-F250D yields a molecular weight of ∼37.8 kD, peak “c.” (E) TLNRD1-FL dimerization measured using MST at 25°C with 40% Red LED laser excitation. Unlabeled TLNRD1-FL was titrated into a fixed concentration (50 nM) of fluorescently labeled TLNRD1-FL. Data were analyzed in MO.Affinity Analysis software (v2.1.3) using the law of mass action to generate an apparent equilibrium dimer Kd, Kd,dimer of 80 ± 0.6 nM. F250D Kd,dimer not determined. FNorm, normalized fluorescence; ND, not defined.

• 

  Figure S2.. Biochemical analysis of TLNRD1. (A) CD spectra of TLNRD1-FL (black) and TLNRD1-F250D (red) showing no change in overall secondary structure as a result of the mutation. (B) nNMR analysis of the F250D mutant shows it closely resembles the WT. 1H,15N heteronuclear single quantum coherence (HSQC) spectra of 150 µM 15N-labeled TLNRD1–4H WT (black) or TLNRD1-4H-F250D (red). (C) CD thermostability analysis of TLNRD1-FL and TLNRD1-F250D showing a reduction in stability with loss of dimerization. (D) Binding of fluorescein-labeled RIAM (residues 4–30) peptide with TLNRD1-FL WT (black) and TLNRD1-FL F250D (red) measured using a fluorescence polarization assay. (E) Low-speed actin cosedimentation bundling assay with TLNRD1-4H serial dilution against 15 µM F-actin. The two gels show the (left) pellet and (right) supernatant fractions. At high TLNRD1-4H concentrations, unbound TLNRD1-4H is present in the supernatant, suggesting the binding becomes saturated. At low TLNRD1-4H concentrations, the actin is predominantly in the supernatant. (F) Actin filaments alone visualized by negative stain EM. (G and H) Actin binding of the F250D mutant. (G) High-speed actin binding assay of TLNRD1-FL and…

• 

  Figure 3.. TLNRD1 and talin R7R8 both bind LD motif proteins and actin.(A) Schematics of TLNRD1-FL, TLNRD1-4H, and talin R7R8 domain structures. (B–G) TLNRD1 binds to LD motif–containing ligands. (B–D) Binding of fluorescein-labeled RIAM(4–30) peptide with (B) TLNRD1-FL, (C) TLNRD1-4H, and (D) talin R7R8 measured using a fluorescence polarization assay. The binding of lamellipodin (20–46; blue) is also shown. (E–G) BODIPY-labeled KANK1(30–60); wild-type (black) and 4A mutant (blue) peptides (E) binding to TLNRD1-FL, (F) not binding to TLNRD1-4H, and (G) binding to talin R7R8. ND, not defined. Experiments performed in triplicate. (H) Structural model of TLNRD1-FL bound to RIAM (cyan) and KANK (pink) LD motifs. (I) High-speed actin cosedimentation assay showing that both TLNRD1-FL and TLNRD1-4H interact with F-actin. TLNRD1-4H alone does not pellet. S/N, supernatant; P, pellet.

• 

  Figure 4.. TLNRD1 is an actin-bundling protein.(A) Schematic of talin ABS2 binding a single actin filament (left). TLNRD1-FL (middle) and TLNRD1-4H (right) bind two actin filaments. (B and C) Actin bundling assays using serial dilutions of TLNRD1-FL (B) and TLNRD1-4H (C) against 15 µM F-actin. (D–F) EM images of negative-stained F-actin bundles with (D) TLNRD1-FL, (E) TLNRD1-4H, and (F) TLNRD1-F250D. Red box (left) enlarged on right. (G) U2OS cells expressing GFP-TLNRD1 were plated on fibronectin, incubated for 2 h with SiR-actin to label the actin cytoskeleton, and imaged live using an Airyscan confocal microscope (1 picture every 16 s; Video 1). Scale bars: (main) 25 µm, (inset) 5 µm. The yellow arrows highlight GFP-TLNRD1 localizing to actin stress fibers.

• 

  Figure 5.. TLNRD1 localizes to, and modulates, filopodia.(A and B) U2OS cells transiently expressing MYO10-mScarlet and GFP, GFP-TLNRD1, or GFP-TLNRD1-F250D were plated on fibronectin for 2 h, fixed, and imaged using a spinning disk confocal microscope. Representative images are displayed (A). Yellow square indicates region of interest, which is magnified on the right. Scale bars: (main) 25 µm, (inset) 5 µm. (B) The number of MYO10-positive filopodia per cell was then quantified (n > 125 cells, three biological repeats; ***, P = 0.002). (C and D) Efficiency of siRNA-mediated (oligos nos. 6 and 7) silencing of TLNRD1 in U2OS cells as detected using quantitative PCR (C) or Western blotting (D). To detect TLNRD1 protein levels by Western blot, TLNRD1 was first immuno-precipitated from cell lysate as indicated. (E) TLNRD1-silenced (oligos nos. 6 and 7) U2OS cells transiently expressing MYO10-GFP were plated on fibronectin for 2 h and fixed, and the number of MYO10-positive filopodia per cell was quantified (n > 150 cells, three biological repeats; ***, P < 0.001). (F and G) U2OS cells transiently expressing GFP, GFP-TLNRD1, or GFP-TLNRD1-F250D were plated on fibronectin for 20 min, fixed, and…

• 

  Figure S3.. TLNRD1 localizes to the tip of MYO10 filopodia. U2OS cells expressing MYO10-mScarlet and GFP-TLNRD1 or GFP-TLNRD1-F250D were plated on fibronectin for 2 h, stained for F-actin, and imaged using SIM. (A) Representative maximum intensity projections are displayed. The yellow squares highlight regions of interest, which are magnified; scale bars: (main) 10 µm; (inset) 5 µm. (B) Heatmap highlighting the subcellular localization of F-actin, MYO10, TLNRD1, and TLNRD1-F250D within filopodia based on >360 intensity profiles (see Materials and methods for details). (C) Random 2D migration assay of U2OS cells plated on fibronectin and nontransfected or transiently expressing GFP or GFP-TLNRD1. GFP-TLNRD1 overexpression increases migration velocity in 2D (n = 120 cells from three independent repeats; ***, P < 0.001). Cell trajectories for nontransfected and GFP-TLNRD1–expressing cells are shown. P values were determined using a randomization test (see Materials and methods for details).