Diagrams of domains and motifs found in myosin-X and DCC and their physical properties. (A) Myosin-X consists of an N-terminal motor domain, three IQ motifs, a coiled-coil (CC) region, three PH domains and MyTH4 and FERM domains. The tandem MyTH4 and FERM domain cassette has a central role in cargo recognition. Three PH domains contain one split PH domain (PHn and PHc) and sequentially appear as PHn–PH1–PHc–PH2. (B) The DCC cytoplasmic tail possesses three conserved regions, P1, P2 and P3 (orange ovals), which might facilitate protein–protein interactions. The transmembrane helix (TH) is represented by a grey rectangle. (C) ITC profiles and thermodynamic data for binding experiments at 20°C. Injections of 6 μl of 733 μM DCC peptide into 50 μM MyTH4–FERM cassette solution. The protein exhibited a K_D value of 0.53±0.06 μM, ΔH of −4.98±0.04 kcal/mol and ΔS of 11.7 cal/mol. Raw data for 30 sequential injections (the upper panel) and the plot of the heat evolved (kcal) per mole of DCC peptide added, corrected for the heat of the peptide dilution, against the molar ratio of the peptide to the cassette. The data (filled squares) were fitted using the software ‘one set of sites', and the solid line represents the best fit. (D) Hydrodynamic property of the complex determined by AUC measurements. The distribution of apparent molecular mass obtained from sedimentation velocity analysis of the complex shows a mono-modal peak with apparent molecular mass of 62.7±8.6 kDa.

Overall structure of the complex between the myosin-X MyTH4–FERM cassette and the DCC P3 peptide. (A) Ribbon representation of the myosin-X MyTH4–FERM–DCC P3 complex. The MyTH4–FERM cassette contains the MyTH4 extension (yellow), and MyTH4 (green) and FERM (cyan) domains. The FERM domain consists of subdomains A, B and C. The DCC P3 peptide (orange) binds subdomain C. (B) Electrostatic surface potentials of the myosin-X MyTH4–FERM cassette. The electrostatic surface potentials are shaded blue in positively and red in negatively charged regions, respectively. The bound DCC P3 peptide is shown as a ribbon model (orange). The yellow oval indicates the positively charged patch, which was identified as the microtubule-binding site in this study. (C) Comparison of MyTH4 and VHS domains. The myosin-X MyTH4 domain (green) superimposed on the STAM2 VHS domain (grey) bound to ubiquitin (Ub, pink) (3LDZ). Six helices (α1–α4, α6–α7) out of the eight conserved helices of the VHS domain are overlapped with corresponding helices. The MyTH4 domain lacks the α5 helix of the VHS domain. A large flipping over of the α8M helix of the MyTH4 domain relative to the corresponding α8 helix of the VHS domain is indicated with a blue arrow. The acidic-cluster dileucine sorting signal peptide (ACLL, pink) of the cation-dependent mannose-6-phosphate receptor bound to the GGA3 VHS domain (1JUQ) is modelled on the STAM VHS domain structure to show the peptide-binding site. (D) The positively charged patch on the MyTH4 domain. A cluster of basic residues (yellow) surrounds nonpolar residues. The basic residues include the C-terminal part of the α2M helix (His1597) and the N-terminal part of α3M helix (Arg1600 and Arg1603), the N-terminal part of the α6M helix (Arg1643, Lys1647 and Lys1650), the C-terminal part of the α7M helix and the flanking loop (Lys1677, Lys1678 and Arg1682). (E) The interface between the MyTH4 and FERM domains. The short linker between the α8M helix and β1A strand links the MyTH4 (green) and FERM (cyan) domains. The interface involves the α7M–α8M loop and α8M helix from the MyTH4 domain and the β2A–α1A loop and α2A helix from subdomain A.

DCC–FERM interactions. (A) The DCC P3 peptide docked into the groove of FERM subdomain C. The MyTH4–FERM domain is shown as electrostatic potential surfaces as in Figure 2B. The salt bridge between Glu1756 and Lys2034 bridges the cleft (A–C cleft) between subdomain A (Sub A) and subdomain C (Sub C). (B) Sequence alignment of DCC P3 peptide with related receptor C-terminal peptides. The netrin receptors are DCC, Neogenin (Ngn) and Frazzled_A (FzlA). Conserved residues are highlighted in yellow (conserved in all members) and grey (conserved in DCC and neogenin). Heptad repeat residues are highlighted in green. The P3 region (Kolodziej et al, 1996) is indicated at the top. Residues that make direct contact with myosin-X are summarized (Contacts). In our current model of the human DCC P3 peptide, 24 N-terminal residues (1390–1413) and 2 C-terminal residues (1446–1447) are missing (indicated with dotted lines) due to poor electron density. (C) Direct interactions between the myosin-X MyTH4–FERM cassette and DCC in cells. Co-immunoprecipitation of myosin-X and DCC from HEK293T cells in a manner dependent on subdomain C of the FERM domain. Myc–Myo-X (residues 1486–2058) contains the myc-tagged full-length MyTH4–FERM cassette and myc–Myo-X(ΔC) (residues 1486–1954) lacks subdomain C of the FERM domain. Cell lysates were incubated with anti-myc antibodies to immunoprecipitate the myc–Myo-X–DCC complex, resolved on SDS–PAGE and then immunoblotted with antibodies against myc or DCC. One percent of the input was included in the blot to indicate the amount of relevant proteins. (D) Effect of mutations of the DCC P3 peptide on myosin-X binding. Pull-down binding assays of mutated DCC P3 peptides with the myosin-X MyTH4–FERM cassette (Myo-X) were performed with mutations of the αP3′ helix (Tyr1420 and Glu1421) and αP3 helix (Leu1432, Glu1433, Leu1435, Met1436 and Leu1439). Eluted samples were analysed by SDS–PAGE.

Close-up views of the DCC–myosin-X FERM domain interactions. (A) At the middle of the αP3 helix (orange), DCC Gln1438 forms a hydrogen bond with the main chain of myosin-X Phe2002 from the β5C strand of subdomain C, which may contribute to lifting of the strand and result in enlargement of the groove. The DCC-bound form (cyan) is superimposed on the free form (magenta). (B) At the C-terminal end of the αP3 helix, three DCC residues (Asn1440, Thr1443 and Ser1445) participate in the hydrogen bonding network formed by multiple hydrogen bonds (solid lines) and the salt bridge between Glu1756 from subdomain A and Lys2034 from subdomain C.

Myosin-X MyTH4–FERM cassette interactions with microtubule, DCC and integrin. (A) Co-sedimentation of microtubules and the myosin-X MyTH4–FERM cassette (Myo-X). The MyTH4–FERM cassette co-sedimented with microtubules in pellets (P) with little remaining in the soluble fraction (S). (B) In vitro pull-down binding assays with GST-fused C-terminal tail peptides of α- and β-tubulins. Purified MyTH4–FERM cassette and GST-fused α1A or β2B-tubulin peptide were mixed in 20 mM Tris–HCl (pH 8.0), 50 mM NaCl and 1 mM DTT and incubated at 4°C for 30 min, followed by incubation with glutathione-Sepharose 4B at 4°C for 10 min. After washing five times, proteins were eluted with 20 mM glutathione in 50 mM Tris–HCl buffer (pH 8.0) containing 150 mM NaCl and then subjected to SDS–PAGE. (C) Tubulin tail binding to the MyTH4 domain and the DCC P3 peptide binding to the FERM domain. Pull-down assay with the wild-type myosin-X MyTH4–FERM cassette (Myo-X (wt)) shows binding to the GST-fused C-terminal acidic tail peptide of β2B-tubulin (GST-tubulin tail) but the K1647D/K1650D mutation abolished binding (Myo-X (K1647D/K1650D)). The F2002K mutation had no effect on tubulin tail binding (Myo-X (F2002K)). On the other hand, the charge-reversal mutation only weakly affected binding to the GST-DCC P3 peptide and the nonpolar to polar mutation of subdomain C abolished the binding. (D) Microtubules bind the MyTH4 but not the FERM domain. Microtubule co-sedimentation experiments with wild-type and mutated myosin-X MyTH4–FERM cassettes used in (C). The charge-reversal mutation of the MyTH4 domain abolished or reduced co-sedimentation and the nonpolar to polar mutation of FERM subdomain C had no effect on sedimentation. (E) Competition assays show interference between tubulin tail and DCC bindings. Pull-down assay of the myosin-X MyTH4–FERM cassette (Myo-X, 30 μM) with the GST-fused β2B-tubulin peptide was performed in the presence of the DCC P3 peptide (0, 30, 60, 120 and 300 μM). The relative amount of myosin-X pulled down with GST-tubulin in the absence of DCC is shown at the bottom. In this assay, the 30 μM DCC P3 (the 1:1 molar ratio of DCC to MyTH4–FERM) completely abolished tubulin binding. (F) Competition assays show interference between microtubule and DCC bindings. Microtubule co-sedimentation experiments with the myosin-X MyTH4–FERM cassette (Myo-X, 4 μM) were performed in the presence of the DCC P3 peptide (0, 4, 20 and 40 μM). The relative ratio of myosin-X to co-precipitated myosin-X with microtubules in the absence of DCC is shown at the bottom. Error bars are standard deviations of each mean value from three independent measurements. (G) Competition assays show interference between integrin and DCC bindings. The myosin-X MyTH4–FERM cassette (Myo-X, 10 μM) was pulled down by GST-fused integrin β5 cytoplasmic tail (743–799) in the presence of the DCC P3 peptide (0, 10, 50 and 100 μM). The relative amount of myosin-X pulled down with GST-integrin in the absence of DCC is shown at the bottom with error bars from three independent measurements. (H) Competition assays show interference between microtubule and integrin bindings. Microtubule co-sedimentation experiments with the myosin-X MyTH4–FERM cassette (Myo-X, 4 μM) were performed in the presence of the integrin β5 peptide (0, 4, 20 and 40 μM). The relative ratio of myosin-X to co-precipitated myosin-X with microtubules in the absence of DCC is shown at the bottom with error bars from three independent measurements.

Peptide recognition by FERM domains. Comparison of complex structures of the myosin-X and radixin FERM domains bound to their binding partners. (A) The FERM domain of the MyTH4–FERM cassette bound to the DCC P3 peptide of the current structure. (B) The radixin–ICAM-2 complex (1J19) exhibits the prototypic binding mode of adhesion molecule recognition by β–β association. A similar binding mode (Motif-1α recognition) is found in both the radixin–PSGL-1 (2EMT) and radixin–CD43 (2EMS) complex structures. The radixin–CD44 (2ZPY) and radixin–NEP (2YVC) complexes also display this binding mode with the modified recognition motif Motif-1β. (C) The radixin–NHERF complex (2D10) displays the second class of binding mode with the NHERF peptide docked into the β-sandwich loops. (D) The radixin–IP3 complex (1GC6) reveals the PtdIns(4,5)P₂-binding site between subdomains A and C. (E) Superimposition of DCC P3 (orange) and ICAM-2 (green) peptides bound to the myosin-X (cyan) and radixin (grey) FERM domains, respectively. The ICAM-2 peptide forms a short β strand followed by a 1-turn 3₁₀ helix. (F) Superimposition of the myosin-X (cyan) and radixin (grey) FERM domains bound to DCC P3 and ICAM-2 peptides, respectively. The bound peptides are omitted for clarity. Nonpolar residues of the radixin FERM domain are highlighted as stick models (yellow).