Structure and tropomyosin binding properties of the N-terminal capping domain of tropomodulin 1

Abstract

Two families of actin regulatory proteins are the tropomodulins and tropomyosins. Tropomodulin binds to tropomyosin (TM) and to the pointed end of actin filaments and "caps" the pointed end (i.e., inhibits its polymerization and depolymerization). Tropomodulin 1 has two distinct actin-capping regions: a folded C-terminal domain (residues 160-359), which does not bind to TM, and a conserved, N-terminal region, within residues 1-92 that binds TM and requires TM for capping activity. NMR and circular dichroism were used to determine the structure of a peptide containing residues 1-92 of tropomodulin (Tmod1(1-92)) and to define its TM binding site. Tmod1(1-92) is mainly disordered with only one helical region, residues 24-35. This helix forms part of the TM binding domain, residues 1-35, which become more ordered upon binding a peptide containing the N-terminus of an alpha-TM. Mutation of L27 to E or G in the Tmod helix reduces TM affinity. Residues 49-92 are required for capping but do not bind TM. Of these, residues 67-75 have the sequence of an amphipathic helix, but are not helical. Residues 55-62 and 76-92 display negative 1H-15N heteronuclear Overhauser enhancements showing they are flexible. The conformational dynamics of these residues may be important for actin capping activity.

FIGURE 1

Alignment of the sequences to chicken Tmod1_1-92 of vertebrate tropomodulins (Tmods) encoded by four different genes, invertebrate Tmod analogs encoded by three genes, and vertebrate leiomodins (Lmods) encoded by three different genes. The residues that are identical or homologous (same charge and hydrophobicity) in all of the Tmods and Lmods are highlighted in cyan. The residues that are homologous in both vertebrate and invertebrate Tmods but not Lmods are highlighted in yellow, whereas those that are homologous only in vertebrate Tmods are in bold typeface. (*) Residue 3 of Chicken Tmod1 aligns with residue 33 of tmd-1 of C. elegans. (#) Residue 3 of Chicken Tmod1 aligns with residue 61 of Sanpodo CG1539-PC. (+) Residue 3 of Chicken Tmod1 aligns with residue 20 of Ensangp00000025288. References are as follows: Chicken Tmod1 (Babcock and Fowler, 1994), Human Tmod1 (Strausberg et al., 2002); Human Tmod2 (Cox and Zoghbi, 2000), Human Tmod3 (Cox and Zoghbi, 2000), Xenopus Tmod3 (Klein et al., 2002), Chicken Tmod4 (Almenar-Queralt et al., 1999), Human Tmod4 (Cox and Zoghbi, 2000), Zebra Fish Tmod4 (Almenar-Queralt et al., 1999), C. elegans Tmd-1 (Wilson, 1998), D. melanogaster Sanpodo, (Adams et al., 2000), A. gambaiae EnsangP00000025288 NCBI XP_309162, Mouse Lmod1 (Okazaki et al., 2002), Human Lmod2 (NCBI XP_351745), Human Lmod3 (Strausberg et al., 2002).

FIGURE 2

Circular dichroism spectra of fragments of tropomodulin 1. (▪) Tmod1_1-130, 10 μM, (•) Tmod1_{1-92, 1} 10 μM, (∇) Tmod1_1-48, 10 μM, and (⋄)Tmod1_38-92, 20 μM. All peptides were in 100 mM NaCl, 10 mM sodium phosphate, pH 6.5, 0°C.

FIGURE 3

The ¹⁵N-¹H HSQC spectrum of the backbone resonances of ¹⁵N- Tmod1_1-92, 0.8 mM (cyan) overlaid with the spectrum of the complex of ¹³C¹⁵N-Tmod_1-92, 0.9 mM with an excess (1.3 mM) of TM1a_1-14Zip (black) at 10°C. Several of the residues following prolines showed two sets of resonances in both spectra. The set with lower intensities, following prolines in the cis conformation are labeled in magenta. The crosspeaks arising from the Gln and Asn side chains, and the folded crosspeaks from the Arg side chains, are not shown to simplify the figure. The samples were dissolved in 100 mM NaCl, 10 mM sodium phosphate 5% deuterium oxide, pH 6.5.

FIGURE 4

The displacements of the ¹H^α and ¹³C^α resonances and the sequential NOEs of Tmod1_1-92 show that residues 24–33 are helical but there is little other regular structure. (A) The chemical shift displacements of the resonances of the ¹H^α, ¹H^N, ¹³C^α, and ¹³C^β atoms of Tmod1_1-92 compared to the shifts of residues in random conformations (Wang and Jardetzky, 2002). The minor sets of resonances, following prolines in the cis conformations are not shown. (B) Summary of the principal NOE crosspeaks used to complete the assignments and determine the secondary structure of Tmod1_1-92.

FIGURE 5

The structure of Tmod1_1-92. Residues 22–36 are aligned. (A) The 10 backbone structures of Tmod1_1-92, with lowest constraint violations, calculated from the NOE data using the program AutoStructure. (B) A detail of the heavy atoms of residues 19–37.

FIGURE 6

The ¹⁵N-¹H HSQC spectrum of the ¹⁵N¹³C- Tmod1_1-92/¹⁴N¹²C-GlyTM1a_1-14Zip complex at 20°C overlaid with a heteronuclear ¹⁵N-¹H HNOE spectrum. The concentrations of Tmod and TM peptides were 0.9 and 1.3 mM, respectively, in 100 mM NaCl, 10 mM sodium phosphate, 10% D₂O, pH 6.5, at 20°C. The HSQC spectrum is shown in blue, the positive crosspeaks in the HNOE spectrum are magenta, and the negative crosspeaks are yellow. Resonances originating from residues in the binding site become more dispersed and have positive HNOEs showing that they become structured upon binding. The minor crosspeaks following proline in the cis conformation are labeled in magenta. The crosspeaks arising from Arg, Gln, and Asp side chains are not shown.

FIGURE 7

The binding of wild-type and mutated fragments of tropomodulin 1 to peptides containing the N-terminus of rat αα-tropomyosin encoded by exon 1b. Panels A, B, D, and E show the effect of complex formation on the folding of the AcTM1b_1-19Zip peptide, and panel C the effect on TM1b_1-19Zip, monitored by CD spectroscopy. Panel F shows native gel electrophoresis studies. (∇) Sum of the folding curves of the TM1b_1-19Zip peptides and Tmod fragment alone; (•) the folding of the mixture of the TM and Tmod peptides. The data are normalized to a scale of 0–1. Mixing AcTM1b_1-19Zip, with Tmod1_1-92 (A), or Tmod_1-48 (B) increases the T_M and cooperativity of folding showing complex formation. In contrast Tmod_38-92 (C) and Tmod1_1-92 containing the mutations L27E (D) and L27G (E) have little effect on the folding. All folding studies were performed in 100 mM NaCl, 10 mM phosphate, pH 6.5. The concentrations of the TM and Tmod peptides were 10 μM except in the case of TMod_39-92, where both peptides were 20 μM. (F) Native gel electrophoresis of Tmod peptides alone (lanes 1, 3, and 5) and Tmod peptides mixed with AcTM1b_1-19Zip (lanes 2, 4, and 6). Lane 1 and 2 contain wild-type Tmod1_1-92, lanes 3 and 4 have Tmod1_1-92 L27 G, and lanes 5 and 6 contain Tmod1_1-92 L27G. Only the wild-type peptide shows the formation of a complex with the TM peptide. AcTM1b_1-19Zip alone does not enter the gel.

FIGURE 8

Circular dichroism at 222 nm of Tmod1_1-48, with peptides containing the N-termini of tropomyosins encoded by exon 1a, AcTM1a_1-14Zip (A) and exon 1b, AcTM1b_1-19Zip (B) and their mixtures as a function of temperature. (∇) AcTM1a_1-14Zip or AcTM1b_1-19Zip, (•) Tmod1_1-48, (⋄) addition of the curves of the unbound TM and Tmod peptides, (▪) mixture of the TM and Tmod peptides. The concentration of Tmod1_1-48 was 10 μM and the concentration of the TM peptides were each 5 μM. All peptides were in 100 mM NaCl, 10 mM sodium phosphate, pH 6.5.

FIGURE 9

The intensities of the C^α-H^N crosspeaks in the HN(CO)CA spectrum of Tmod1_1-92 relative to the average intensities. The crosspeaks from the residues preceding P39 have ∼25% of the intensity of the residues following P39 suggesting that the C-terminal region is more flexible than the N-terminal region. Crosspeaks to proline residues, which do not have a H^N residue, are missing from the graph, as are crosspeaks that were not resolved in the spectrum.