Extensive and modular intrinsically disordered segments in C. elegans TTN-1 and implications in filament binding, elasticity and oblique striation

Abstract

TTN-1, a titin like protein in Caenorhabditis elegans, is encoded by a single gene and consists of multiple Ig and fibronectin 3 domains, a protein kinase domain and several regions containing tandem short repeat sequences. We have characterized TTN-1's sarcomere distribution, protein interaction with key myofibrillar proteins as well as the conformation malleability of representative motifs of five classes of short repeats. We report that two antibodies developed to portions of TTN-1 detect an approximately 2-MDa polypeptide on Western blots. In addition, by immunofluorescence staining, both of these antibodies localize to the I-band and may extend into the outer edge of the A-band in the obliquely striated muscle of the nematode. Six different 300-residue segments of TTN-1 were shown to variously interact with actin and/or myosin in vitro. Conformations of synthetic peptides of representative copies of each of the five classes of repeats--39-mer PEVT, 51-mer CEEEI, 42-mer AAPLE, 32-mer BLUE and 30-mer DispRep--were investigated by circular dichroism at different temperatures, ionic strengths and solvent polarities. The PEVT, CEEEI, DispRep and AAPLE peptides display a combination of a polyproline II helix and an unordered structure in aqueous solution and convert in trifluoroethanol to alpha-helix (PEVT, CEEEI, DispRep) and beta-turn (AAPLE) structures, respectively. The octads in BLUE motifs form unstable alpha-helix-like structures coils in aqueous solution and negligible heptad-based, alpha-helical coiled-coils. The alpha-helical structure, as modeled by threading and molecular dynamics simulations, tends to form helical bundles and crosses based on its 8-4-2-2 hydrophobic helical patterns and charge arrays on its surface. Our finding indicates that APPLE, PEVT, CEEEI and DispRep regions are all intrinsically disordered and highly reminiscent of the conformational malleability and elasticity of vertebrate titin PEVK segments. The proposed presence of long, modular and unstable alpha-helical oligomerization domains in the BLUE region of TTN-1 could bundle TTN-1 and stabilize oblique striation of the sarcomere.

Fig. 1. The domain organization, predicted disorder regions and sites for protein expression, antibody production and peptide synthesis in TTN-1

Domain organization: Purple boxes represent Ig domains, green boxes represent Fn3 domains, yellow indicates the 2400 residue PVET/K region, the blue box represents the 1500 residue BLUE region predicted to form a coiled-coil, gray denotes the 250 residue region composed of AAPLE repeats, orange represents additional predicted coiled-coil sequences, and the thin blue boxes with projecting lines denote the 30 residue dispersed repeat (DispRep). Also, note the serine/threonine protein kinase domain near the C-terminus. Above the schematic is a bar showing the possible location of the TTN-1 domains in the scaromere. Expressed proteins and peptides: Black boxes denote the position and extent of ~300 residue regions that were expressed in E. coli (5/6, 1/2, 3/4, 9/10, 11/12 and 7/8), and peptides that were synthesized (AAPLE42, CEEEI, PEVT39, BLUE32, DispRep). Four polyclonal antibodies (EU145, EU102, 9/10 and EU143) were generated to the indicated regions. Immediately below the schematic are indicated whether or not (+ or −) a given region was able to bind to F-actin or myosin in vitro. The graph is the PONDR VLXT disorder probability profile for TTN-1. Disorder probability above 0.5 implies disorder.

Fig. 2. Identification of a 2.2 MDa TTN-1 by immunoblot and site-specific antibodies

(A) TTN-1 specific antibodies: Total proteins from a mixed-stage population of C. elegans (Ce), or rabbit myofibril proteins (Rb) were separated on a 2–8% gradient gel and transferred to PVDF membrane. Individual strips were incubated with either antibodies to vertebrate titin (3.2 and 2.2. MDa) or nebulin (700 kDa) or a mixture of antibodies to nematode myosin heavy chain (MHC, 200 kDa), paramyosin (90 kDa) or actin (43 kDa) to serve as size markers. Of the four antibodies produced to different regions of TTN-1(EU145, EU102, 9/10 and EU143), only 9/10 and EU143 detect a polypeptide from worm extracts running at approx. 2 MDa, the size predicted for TTN-1 from previous sequence analysis. (B) Kettin-specific antibody: Total proteins from either wild type (wt) or a strain carrying an intragenic deletion in the kettin gene (ok1641) were separated on a 5% gel, transferred to membrane and reacted with either EU102 or MH44, an monoclonal shown previously to recognize kettin. As shown, both antibodies recognize a protein of approx. 500 kDa, the size of kettin, from wild type and detect a slightly smaller protein in ok1641. Thus, although EU102 was generated to TTN-1 sequence, it clearly detects a different protein, kettin and displayed no reactivity with TTN-1 on immunoblot.

Fig. 3. Immunofluorescence localization of TTN-1 to the I-band region of the sarcomere

Each panel of three (A–C, D–F etc.) confocal images shows results of co-staining adult body wall muscle with the indicated pairs of antibodies. (A–C) Anti-TTN-1 antibodies EU143 localize both between dense bodies and outside of dense bodies in the I-band. Both myosin heavy chain B (MHC B) and twitchin had been shown previously to localize to the A-band except for the middle of the A-band. Note that co-staining of antibodies generated to the C-terminus of TTN-1 (EU143) with either MHC B (panels A–C) or with twitchin (G–I) shows some overlap at the outer edge of the A-band (yellow in panels C and I). (J–L) Anti-TTN-1 antibodies 9/10 localize in the same region as the EU143 antibody. (M) Drawing of a portion of body wall muscle indicating the organization of the myofilament lattice in this obliquely striated muscle. The drawing is oriented such that the I-bands and A-bands have the same orientation as in the immunofluorescence images. Dense bodies are analogous to Z-disks in mammalian cross striated muscle. (N) Proposed model of TTN-1 orientation, maximal span and variable environment in the sarcomere. The TTN-1 molecules span from the dense bodies toward the A-band in an N to C orientation, with some reaching as far as the edge of the thick filaments. Such molecules are mis-registrated across the sarcomere (bottom three filaments), due to the shape of the dense body. The molecules are also offset significantly between adjacent sarcomeres due to the lateral offset of the sarcomeres in this obliquely striated muscle (top two filaments). Scale bar, 5 μm.

Fig. 4. TTN-1 interacts with F-actin and myosin

(A) Data demonstrating actin binding to some TTN-1 fragments. (Upper panel) Cosedimentation of TTN-1 fragment 7/8 with F-actin. Fragment 7/8 at 0, 0.25, 0.5, 1.0, 2.0 and 4.0 μM were incubated with 10 μM F-actin, and then pelleted by ultracentrifugation. Pellets (p) and supernatants (s) were resolved by SDS–PAGE (10% Tris-Glycine) and visualized by Coomassie Blue staining. (Lower panel) Plots of F-actin binding derived from cosedimentation experiments with fragments 7/8, 9/10, 5/6 and 1/2. (B) Data demonstrating myosin binding to some TTN-1 fragments. Total conventional myosin from nematodes was purified and adsorbed to the wells of a microtiter plate. The indicated 6His tagged TTN-1 fragments were added at 0 to 500 nM and an ELISA was conducted using anti-6His antibodies for detection.

Fig. 5. Secondary structure interconversion, disorderness and coiled-coil of BLUE32

(A–F). CD spectra of BLUE32 peptide (50 μM) in 10 mM K-phosphate, pH 7.0, 150 mM KCl: (A) at 2, 25, 50 and 75 °C; (B) in 0%, 20%, 40% and 80% (v/v) TFE. (C, D) Changes in the relative content (mole fractions) of conformational states of BLUE32 at different temperature (C) and in different concentrations of TFE (D) (●, ▲, ■ for component I, II and III, respectively). (E) Three conformational states I, II and III resulting from CCA deconvolution of all spectra at 2 (with and without 150 mM KCl), 25 (with and without 150 mM KCl), 50, 75 °C and in 0%, 20%, 40%, 80% (v/v) TFE. (F) Thermal titration of Blue 32 was monitored by continuous measurement of CD value at 222 nm from 2 to 80 °C; (G) in the presence of 0, 2, 5 M urea; and (H) the addition of 4 M NaCl.

Fig. 6. Helical nature of BLUE-32 and comparison with tropomyosin

(A) Helical net diagram of hydrophobic residues in BLUE peptide sequence, residues 7036–7088. Large yellow circles are hydrophobic residues with the 8-4-2-2 pattern. The split green/blue circles are the locations of the As or Lys residues between the two leucines separated by 8 residues in the 8-4-2-2 pattern. Small colored circles are charged (red and blue) and polar (green) residues. (B) Space filling model of the BLUE-16 repeat. Structure of BLUE16 was generated by ROSETTA and then subjected to 4 ns of MD in a water box at low ionic strength. Structure orientated to show the residues QAKD that are in the white strip on the left side of the inset helical net between the two structures. (C) Same BLUE-16 structure rotated 180 degrees to show the hydrophobic strip as in the helical net, residues LALNL. Sequence of peptide with 8-4-2-2 repeat pattern shown at top between structures. (D) The 8-4-2-2 pattern is shown with only the location of the hydrophobic residues highlighted for clarity. (E) Helical net typically found in globular proteins. (F) Helical net found in coiled-coil proteins such as tropomyosin. (G) Helical net for residues 61–113 of porcine tropomyosin that the BLUE-32 peptide threaded to.

Fig. 7. BLUE helices in parallel configuration

(A) Model of the first 250 residues of BLUE threaded onto tropomyosin before MD. The BLUE32 peptide is highlighted to show residue type. (B) The structure after 15 ns of MD. Asterisks indicate regions of the α-helices that transition from the curved form as found in coiled-coil structures to the straighter conformation preferred by BLUE. (D) Close-up of the BLUE-32 region after 15 ns of MD. Residues that were interacting after 15 ns of MD simulation are shown in space filling form. L7029 and L7053 are interacting in the threaded model and remain in contact throughout the simulation. A network of salt-bridges has formed between the two helices. (D) The peptide sequences for the two helices are shown with the octad repeat below the sequence and the 8-4-2-2 hydrophobic residues indicated. (E) Helical net as in Fig. 6A. The residues that are interacting between the two chains have × over them. The line indicates the interface between the two chains.

Fig. 8. BLUE helices in an anti-parallel configuration

(A) Model of the first 280 residues of BLUE threaded onto α-actinin. Structure shown is before minimization and MD simulation. Residues that are close to anti-parallel and then cross are highlighted in color of the residue type. (B) Structure after minimization and 21 ns of MD simulation. (C) Structure of residues 7158–7183 KQEADAKLQKENDDKLKQEADAKLKK and 7217–7239 NDDKLKQEADAKLQKENDDKLKQ) from α-actinin threading after minimization and 21 ns of MD. MD greatly disrupted the anti-parallel bundle conformation of the actinin based structure and the two helices crossed as shown. The interacting hydrophobic residues are the leucines at opposite ends of two LQKENDDKL repeats with the asparagine residue between the leucines the center of the interaction (asterisked residues in sequences in D). The residues of the LNL interaction and the salt bridges are labeled on the structure and the the sequence in D. (D) Sequences of the peptides with the location of the 8-4-2-2 hydrophobic residues and octad repeat. (E) Helical net showing the interacting residues. Residues on the peptide from 7158 to 7183 are marked with a ×. Residues on the peptide from 7217 to 7239 are marked with a ‘+’. Helical net does not represent 7158 to 7239 only a schematic of the interacting residues on these two peptides.

Fig. 9. Conformations of APPLE42, PEVT39, BLUE32 and DispRep peptides from distinct modules of TTN-1

Conformations as determined from circular dichroism spectroscopy, including single α-helix (H, dark green), bundled α-helix (CC, pink), polyproline II helix (PPII, green), β-turn (βT, grey) and unordered structure (U, line) at various temperatures, ionic strengths and solvent polarity were characterized by circular dichroism. Peptides from PEVT, DispRep and AAPLE consist of a combination of PPII helix and unordered structure in aqueous solution. In trifluoroethanol (TFE) they are converted to single α-helix (for PEVT, DispRep) or β-turn (for AAPLE), respectively. Urea increases PPII contents of AAPLE, PEVT and DispRep peptides. NaCl decreases PPII contents. The BLUE32 peptide from the long 1500 residue region forms higher ordered helical structure. Its single α-helix content decreases with higher temperatures (converts to unordered) and increases with higher concentration of TFE, converted from an unordered structure and the unraveling of a coiled-coil helix. Urea denatures and NaCl destabilizes α-helical structure.