Structural Basis and Mechanism of TrkA Activation by NGF through Ligand-Induced Rotation of Transmembrane Domain Dimers

Trk receptors are essential for the nervous system development. The molecular mechanism of TrkA activation by its ligand NGF is still unsolved. Recent data indicates that at endogenous levels most of TrkA is in an equilibrium monomer-dimer and the binding of NGF induces an increase of the dimer and oligomer forms of the receptor. An unsolved issue is the role of the transmembrane domain (TMD) in the dimerization of TrkA and the structural details of the TMD in the active dimer receptor. We found that TrkA-TMD can form dimers, identified the structural determinants of the dimer interface in the active receptor and validated this interface using site-directed mutagenesis together with functional and cell differentiation studies. As around 20% of TrkA is in an inactive dimeric form in the absence of ligand how NGF binding is able to active this pre-formed dimer is unknown. Using in vivo crosslinking we identified a reordering of the extracellular juxtamembrane (JTM) region after ligand binding. This conformational change could be mimicked by replacement of some residues in the JTM with cysteine that form ligand-independent active dimers and reveal a preferred dimer interface. In addition to that, insertion of leucine residues into the TMD helix induces a ligand-independent TrkA activation suggesting that a rotation of the TMD dimers could be behind TrkA activation by NGF. Altogether our data indicate that the transmembrane and juxtamembrane regions of the receptor play a key role in the dimerization and conformational activation of TrkA by NGF.


23
Abstract 24 Trk receptors are essential for the nervous system development. The molecular mechanism 25 of TrkA activation by its ligand NGF is still unsolved. Recent data indicates that at 26 endogenous levels most of TrkA is in an equilibrium monomer-dimer and the binding of 27 NGF induces an increase of the dimer and oligomer forms of the receptor. An unsolved 28 issue is the role of the transmembrane domain (TMD) in the dimerization of TrkA and the 29 structural details of the TMD in the active dimer receptor. We found that TrkA-TMD can 30 form dimers, identified the structural determinants of the dimer interface in the active 31 receptor and validated this interface using site-directed mutagenesis together with 32 functional and cell differentiation studies. As around 20% of TrkA is in an inactive dimeric 33 form in the absence of ligand how NGF binding is able to active this pre-formed dimer is 34 unknown. Using in vivo crosslinking we identified a reordering of the extracellular 35 juxtamembrane (JTM) region after ligand binding. This conformational change could be 36 mimicked by replacement of some residues in the JTM with cysteine that form ligand-37 independent active dimers and reveal a preferred dimer interface. In addition to that, 38 insertion of leucine residues into the TMD helix induces a ligand-independent TrkA 39 activation suggesting that a rotation of the TMD dimers could be behind TrkA activation by 40 NGF. Altogether our data indicate that the transmembrane and juxtamembrane regions of 41

Introduction 50
Nerve growth factor (NGF) is a member of the mammalian neurotrophin (NT) protein 51 family implicated in the maintenance and survival of the peripheral and central nervous 52 systems (1-3). NGF is a dimer that interact with two distinct receptors, TrkA a cognate 53 member of the Trk receptor tyrosine kinase family and the p75 neurotrophin receptor, 54 which belongs to the tumor necrosis factor receptor (TNFR) superfamily of death receptors 55 (4-6). TrkA signalling is essential for sensory and sympathetic neuron survival during 56 development (7). Genetic mutations in the NTRK1 gene cause Congenital Insensitivity to 57 Pain with Anhidrosis (CIPA), a rare disease (8), and somatic mutations and chromosomal 58 rearrangements generate aberrant protein fusions with constitutive kinase activation causing 59 several types of cancer (9, 10). 60 Despite all these important roles, the molecular mechanisms of TrkA activation have been 61 poorly studied in comparison with those of other receptor tyrosine kinase (RTK) family 62 members (11,12). The first three extracellular domains of TrkA consist of a leucine-rich 63 region, LRR (Trk-d1) that is flanked by two cysteine-rich domains . 64 The fourth and fifth domains (Trk-d4 and Trk-d5) are immunoglobulin (Ig)-like domains, 65 and these are followed by a 30-residue-long linker that connects the extracellular portion of 66 the receptor to the single transmembrane domain and a juxtamembrane intracellular region 67 that is connected to the kinase domain. TrkA is activated by nerve growth factor (NGF) a 68 member of the neurotrophin family (3). The NGF binding domain is located in the Trk-69 d5(Ig2) domain (13) although other domains also participate in activation by NTs through 70 an unknown mechanism (14,15). 71 Two models for TrkA activation are postulated; a ligand-induced dimerization of TrkA 72 monomers and a ligand-induced conformational activation of pre-formed inactive dimers. 73 The first model, which is based on the crystal structure of NGF with the ligand-binding 74 domain of TrkA (13,16), assumed that the dimerization of TrkA is solely ligand-mediated 75 and that receptor-receptor interactions are not present in the absence of its ligand. In the 76 second model TrkA exists as a pre-formed inactive dimer suggesting receptor-receptor 77 contacts in the absence of NGF (17,18). The most recent data using single-particle tracking 78 (19) and FRET studies (20) suggests that TrkA is, at endogenous levels, predominantly 79 monomer (80%) and NGF binding induces an increase and a stabilization of the TrkA 80 dimers and the formation of oligomers, together with a conformational change leading to 81 kinase activation. This mechanism of activation has been called the "transition model" (20) 82 and postulates a dynamic transition from a monomer to an inactive dimer to a ligand-bound 83 active dimer, suggesting that Trk receptors are activated through a combination of the two 84 mentioned models. 85 Whatever the model, it is clear that dimerization of TrkA is required for its activation. 86 Deletion constructs suggested that dimerization of TrkA in the absence of NGF is mediated 87 by the transmembrane (TM) and by the intracellular domains (ICD) (20). In the case of the 88 ICD this is supported by the crystal structure of the kinase domain of TrkA that showed the 89 presence of dimers in the crystallographic unit (21,22). However the structural determinants 90 of the TMD dimerization are not know and in this regard, it is important to understand the 91 conformation of the TrkA-TMD dimer and identify the active dimer interface that may 92 represent the functional state of the full-length receptor. In addition biochemical data 93 supporting a conformational activation of TrkA pre-formed dimers are lacking. 94 In the present work we investigated the role of TrkA-TMDs in its activation, using several 95 complementary structural and molecular biological approaches. Here we present the 96 structure of the dimeric TMD in the receptor active state and show that the coupling of 97 ligand-binding to rotation of the TMD may underlie the TrkA activation by NGF. 98 Results 100

Structural basis of TrkA transmembrane domain dimerization. 101
It has been shown that the isolated TMDs of all human RTKs form dimers in bacterial 102 membranes (23). In addition functional studies indicate that TMDs play an important role 103 as a modulator of RTK homodimerization and kinase activation (reviewed in (23-25)). 104 Switching between two dimerization modes of the transmembrane helix has recently been 105 described as part of the activation mechanism of the EGF, . However, to date the role of TrkA-TMD dimerization in TrkA receptor activation has 107 not been studied in detail. Here we used two different approaches to characterize the 108 dimerization of the TrkA-TMD. First, to study TrkA-TMD dimerization in a biological 109 membrane context we used ToxRED, a genetic bacterial system, to assess TrkA 110 transmembrane dimerization using ToxCAT as a reporter system (30, 31). The TrkA-TMD 111 self-associated in biological membranes in this system ( Figure 1A). To determine the TrkA 112 amino acid residues implicated in this dimerization we assayed the effect in this assay of 113 individual mutation of each of the small residues (Ala, Gly or Ser) in the TMD sequence to 114 a bulky residue such as isoleucine. This approach to disruption of dimerization was similar 115 to that used in classical studies of glycophorin-A, in which the mutation G83I impaired its 116 TMD dimerization (31). In our study, less TMD dimerization was observed in the S419I 117 and G423I mutants compared to wt ( Figure 1B), implicating the sequence S 419 xxxG 423 as 118 part of the TrkA-TMD dimerization motif in bacterial membranes. 119 To obtain a structural insight into TrkA-TMD dimerization we solved the structure of 120 human TrkA-TMD dimers in lipid micelles using NMR. For this study, the human TrkA-121 TMD was produced in a cell-free system (see Experimental Procedures) as previously 122 described (32). When the peptide is solubilized in dodecylphosphocholine (DPC) micelles 6 at a lipid-to-protein molar ratio (LPR) of 50:1 the TrkA-TMD is in equilibrium between 124 monomeric, dimeric and other oligomeric states. The ratio of these states varies as the LPR 125 value is altered ( Figure S1A). We then titrated TrkA-TMD in DPC micelles using a 126 standard technique in our laboratory (33) to measured the standard free energy of 127 dimerization (ΔG 0 ) ( Figure S1B). The ΔG 0 value obtained (-1.9±0.2 kcal/mol) suggested 128 that the TrkA-TMD dimer is quite stable compared to the TMD dimers of other RTKs. 129 Thus, although its dimerization energy is weaker than that of the VEGFR2 dimer (ΔG 0 = -130 2.5 kcal/mol in DPC) (34), it is stronger than that of FGFR3 (ΔG 0 = -1.4 kcal/mol in 131 DPC/SDS 9:1 mixture) and ErbB4 (ΔG 0 = -1.4 kcal/mol in DMPC/DHPC 1:4 bicelles) 132 (35). The 15 N-HSQC spectrum of 15 N labeled-TrkA-TMD ( Figure S2) contained the 133 expected number of cross-peaks, and the good quality of the spectra allowed solving of the 134 structure of the dimer in DPC micelles (Figures 1C, Figure S3, Figure S4, Figure S5 and 135 Table S1). The α-helical region of the TrkA-TM dimer starts at G417, ends at N440, and is 136 ~38 Å in length ( Figure S3). The crossing angle of the TrkA-TM helices is 40º, and the 137 minimal distance between two monomers is 8.8 Å ( Figure 1C and Table S1). The 138 hydrophobicity plot and contact surface area of the dimer is shown in Figure S5. The 139 dimerization interface lies along the sequence motif L 424 xxF 427 A 428 xxF 431 ( Figure 1C and 140 Figure S5) that is conserved in the TrkA-TMD of several species and also in TrkC but not 141 in TrkB ( Figure 1D). 142 These analyses of TrkA-TMD dimerization suggested the existence of two possible 143 dimerization motifs; S 419 xxxG 423 and L 424 xxF 427 A 428 xxF 431 , which are located on opposite 144 faces of the TMD helix ( Figure 1C). The latter motif was found in analysis of the 145 dimerization of TMDs in micelles using a direct approach and in the context of the isolated 146 domain, while the former motif was observed in analysis of the dimerization of TMDs in 147 real membranes, using an indirect approach and in the context of a chimeric protein. Thus, 148 the biological relevance of each motif can be questioned, as the presence of large 149 extracellular and intracellular globular domains or the lipid environment of the plasma 150 membrane may favour or hinder a specific interaction interface (36). We therefore used 151 different functional assays to verify the found dimer interfaces in the context of the full-152 length receptor. 153 154 Functional identification of the dimer interface upon NGF stimulation. 155 The state of full-length TrkA was followed by assay of three different aspects of its 156 activity: dimerization of the receptor, phosphorylation of intracellular tyrosine residues and 157 neurite differentiation of PC12nnr5 cells. To investigate the dimerization of TrkA, we 158 individually mutated most of the N-terminal residues of the rat TrkA-TMD to cysteine 159 ( Figure 2A), expressed these constructs in Hela cells, and then measured the amount of 160 cross-linked species. To facilitate cross-linking via these cysteine residues in the 161 transmembrane domains we used oxidation with molecular iodine (I 2 ) as previously 162 described (37). Such oxidation allows the formation of a disulfide bond between two close 163 cysteine residues inside the lipid bilayer. Plasma membrane fractions from cells expressing 164 different single-cysteine mutants were incubated in the absence or presence of NGF, 165 together with molecular I 2 and were then analyzed by non-reducing SDS-PAGE and 166 western immunoblotting. As shown in Figures  To further study the significance of the found interfaces, we mutated the small residues Ala, 173 Gly and Ser within this region to the bulky Ile residue and we then assayed TrkA activation 174 upon NGF stimulation (Figures 2F and 2G). The rationale behind this approach was that the 175 mutation of a small residue to a bulky one on the relevant interface would prevent the 176 formation of the active dimeric state by inducing steric clashes, and would, therefore, 177 reduce TrkA activation. To perform this assay, we transfected Hela cells that do not express 178 endogenous TrkA with these mutants, stimulated these cells with non-saturating 179 concentrations (10 ng/mL) of NGF, and then assayed TrkA activation by analysis of TrkA 180 autophosphorylation using western blotting. Upon transfection, two TrkA electrophoretic 181 bands are present in the TrkA immunoblots of Hela cells; a lower band (approx.110 kDa) of 182 intracellular immature TrkA that has not completed Golgi-mediated processing of high-183 mannose N-glycans (38) and an upper band (approx.140 kDa) with mature sugars that is 184 expressed in the plasma membrane. Exposure to NGF substantially increased the 185 phosphorylation of the upper TrkA band as assessed by blotting with a phospho-specific 186 antibody against the phospho-tyrosine residues of the activation loop, Tyr674 and Tyr675 187 (P-Y674/5). This autophosphorylation was quantified to follow TrkA activation. 188 Constitutive (t=0, no NGF added) and ligand-dependent phosphorylation of plasma 189 membrane-localized TrkA after 5 and 15 minutes were measured. Since overexpression of 190 TrkA induces ligand-independent autophosphorylation, we first transfected the Hela cells 191 with increasing concentrations of TrkA to determine a TrkA level that could still be 192 detected but that displayed no autophosphorylation in the upper band in the absence of 193 NGF ( Figure S6). It is noteworthy that all mutants are expressed at the plasma membrane as 194 evidenced by both immunofluorescence localization in the absence of Triton X-100 using 195 an antibody against an epitope in the TrkA N-terminus ( Figure S7) and by flow cytometry 196 ( Figure 2H). Out of the seven single-point mutants tested only the A428I substitution 9 demonstrated a pronounced inhibitory effect on receptor autophosphorylation. A428 is the 198 only small-chain residue that is found deep in the dimerization interface of the TrkA TMD 199 conformation determined using NMR, which further supports the relevance of the obtained 200 NMR structure. The inhibitory effect of A428 substitution on receptor activity was further 201 enhanced when all three Ala residues that are at least somehow involved in the TMD 202 dimerization in the NMR-based structure, A421, A425 and A428, were simultaneously 203 substituted (TrkA-3A/3I). 204 Lastly, we studied the effect of the same mutations on the NGF-induced differentiation of 205 transfected PC12nnr5 cells ( Figure 2H). Again, the A428I mutant displayed substantial 206 inhibition of this TrkA activity. Unexpectedly, although the mutation S419I had no effect 207 on TrkA activation by NGF in these two assays, the mutation G423I did have an impact on 208 cell differentiation ( Figure 2H). This result suggested that G423 is located at a helix 209 interface that is important in the downstream activation and signalling of TrkA leading to 210 cell differentiation (See Discussion). 211 The combined results of the functional assays agree regarding the importance of the NMR-212 derived TMD structure for TrkA activation. The critical role of the residue A428 in TrkA 213 activation by NGF together with its location deep in the dimer interface indicates that the 214 NMR-based structure corresponds to the mode of TMD interaction in the NGF-induced 215 active state of the receptor 216 217 A specific conformation of extracellular juxtamembrane residues is necessary for 218

TrkA activation. 219
Stimulation of Hela cells transfected with TrkA-wt with NGF induces the formation of 220 TrkA dimers that are crosslinked with BS3 ( Figure 3). While it is known that TrkA dimers 221 are formed in the absence of NGF, no cross-linking was observed without the ligand, which 222 suggests that ligand binding is accompanied by changes in the conformation of the 223 extracellular part of the protein. As BS3 reacts only with free amines (the side-chains of 224 Lys residues or a free N-terminus) we searched for possible sites in TrkA that might have 225 caused the observed crosslinking. According to the crystal structure of the TrkA/NGF 226 complex (15) ( Figure 3A), there are no lysine residues in the TrkA-ECD that are located in 227 a position where crosslinking of the side chains of Lys residues of two monomers could 228 occur. Since BS3 does not cross the plasma membrane, and since we used the full-length 229 TrkA receptor in our assays, we wondered if the observed BS3 cross-linking was mediated 230 via crosslinking of K410 and K411 in the extracellular juxtamembrane region (eJTM) of 231 TrkA ( Figure 3A) since this region is not observed in the crystal structure (16). To verify 232 this hypothesis, we mutated both K410 and K411 to Arg and repeated the initial experiment 233 using HEK293 cells transfected with this TrkA-KK/RR construct ( Figure 3B). No BS3-234 induced TrkA crosslinking was observed in the TrkA-KK/RR-transfected cells suggesting 235 that NGF binding brings this region of the eJTM into close proximity. 236 We considered that, if NGF indeed induces contacts between these eJTM regions, then we 237 should be able to mimic this activity of NGF by forcing the dimerization of eJTMs in the 238 absence of NGF. For this purpose, we individually mutated most of the residues in the 239 eJTM of TrkA to cysteine and subsequently analyzed the dimerization of these transfected 240 single point mutants ( Figure 4A). After transfection of Hela cells, disulfide dimers were 241 spontaneously formed in all constructs but the amount of dimers differed between the 242 various mutants ( Figure 4B). The amount of dimer is significantly higher in the positions 243 D412C and K411C. As a functional assay we then transfected these mutants into HeLa 244 cells, which do not express endogenous TrkA, and quantified the phosphorylation of the 245 tyrosines from the kinase activation loop (Y674/675) in the absence and presence of NGF 246 ( Figure 4C). This analysis showed the presence of active dimers (D406C, K410C and K411C) which are activated constitutively in the absence of NGF, and dimers that are not 248 active in the absence of NGF (V408C, D412C, E413C and T414C). In general there is a 249 poor correlation ( Figure 4D) between the amount of dimer formation and constitutive 250 activation, suggesting that dimerization by itself is not enough for TrkA activation. 251 However the mutants with higher constitutive activation (D406C, K410C and K411C) 252 showed a good correlation between dimer formation and activation ( Figure 4D, green dots, 253 r 2 =0.92). We then transfected some of the active mutants in PC12nnr5 cells. In the absence 254 of NGF the R405C, K410C and K411C mutants induced the formation of neurites in 255 PC12nnr5 cells (Figures 4E and 4F) supporting the constitutive activation of these mutants 256 and suggested that disulfide bond formation through this interface mimics the binding of 257 NGF. If we assume that the TMD α-helix continues in the juxtamembrane region the 258 residues R405, K410 and K411 are in one face of the helix (in green in the Figure 4G). By 259 contrast the residues whose mutation to cysteine induces inactive dimers are located in 260 another face (in red in the Figure 4G). All the mutants are correctly expressed at the plasma 261 membrane as found by flow cytometry ( Figure 4H). 262 In the presence of NGF these mutants showed no further activation by NGF ( Figure 4C) 263 suggesting they are fully active and the dimer interface adopted by the cysteine dimers is 264 similar or identical to the one obtained with NGF binding. By contrast the mutants that 265 form inactive dimers respond to NGF in a similar manner or even higher than the wt 266 ( Figure 4C), indicating that covalent dimer formation do not preclude the activation by the 267 ligand. How can NGF activate a covalent dimer if the cysteine are located downstream of 268 the ligand binding domain? The most plausible explanation is that NGF can induce the 269 dimerization of two pre-formed inactive dimers as shown in the Figure 4H. Thus, Leu insertion allows evaluation of whether a change in the rotation angle of the intracellular domain plays any role in TrkA activation ( Figure 5C). We transfected the 296 constructs TrkA-ins1L, TrkA-ins2L, TrkA-ins3L and TrkA-ins4L that included 1 to 4 297 inserted leucines respectively. The insertion of one Leu, TrkA-ins1Leu, significantly 298 increased both the constitutive activation of TrkA in transfected Hela cells compared to that 299 of transfected TrkA-wt ( Figure 5D and 5E) and the differentiation of PC12nnr5 cells 300 compared to wt-transfected cells ( Figure 5F) in a ligand-independent manner. Equal levels 301 of all constructs were expressed at the plasma membrane as determined using flow 302 cytometry ( Figure S8). These data indicate that a rotation of the TMD is likely to underlie 303 TrkA activation. 304 305 306

Discussion 307
TrkA belongs to a subfamily of RTKs that includes the other family members TrkB and 308 TrkC. These RTKs are essential for the formation of the nervous system and mediate a 309 variety of cellular responses in normal biological processes and in pathological states (39). 310 An understanding of their mechanism of action is necessary to facilitate the design of new 311 pharmacological agents targeted to the processes in which they play a role. Additionally, we found another possible TMD dimerization interface using a genetic assay, 347 and functional tests indicated that this interface was also relevant for TrkA activation. 348 Although destruction of this interface by replacement of a small amino acid residue with a 349 bulky residue inhibited the dimerization in the bacterial membranes, it does not affect the 350 activation of TrkA by NGF, suggesting that this dimer interface is not populated in the full-351 length TrkA receptor at least in the ligand-bound state. The structural data indicates that 352 this interface is an opposite side as the active dimer interface. Interestingly in an ideal α-353 helix connected to the TMD the residues from the eJTM that form inactive dimers when 354 mutated to cysteine are located in the same helix side than the SxxxG motif. This suggests 355 that rotation of the eJTM will also rotate the TMD. Based on this it is possible that the 356 SxxxG interface forms part of an inactive pre-dimer state, although we cannot detect the 357 formation of dimers when the SxxxG motif is mutated to cysteine. One possibility is that, in 358 the inactive state, the distance between the SxxxG motifs are longer than the required for 359 the disulfide bond formation. This could be due to the presence of steric clashes from the 360 TrkA ICD or ECD, that are absent in the ToxRed constructs. However the finding that the 361 mutation of the G423I reduces significantly the differentiation of PC12 cells with NGF but 362 not the activation of the kinase domain (as shown by the phosphorylation of the Tyr-363 674/675) is intriguing. PC12 cells differentiation needs of a sustained MAPK activation to 364 differentiate and it is plausible that a higher activation of TrkA is reached by the lateral 365 association of liganded dimers; i.e oligomerization of TrkA dimers. Actually the formation 366 of TrkA oligomers by NGF has been already described by single-particle tracking in 367 neuroblastoma cells (19). And oligomerization of ligand-bound RTKs seems a general 368 mechanism to regulate differential intracellular signalling (44). The important role of G423 in cell differentiation and its location in an opposite side of the α-helix to the active dimer 370 interface, suggests that G423 may participate in a protein-protein interface corresponding to 371 a supra-dimer (oligomer) form of the TrkA receptor that is populated upon NGF binding 372 (19). In this context further investigation on the TrkA oligomer formation is warranted. 373 To answer the second question posed above, we investigated the role of the eJTM regions 374 of the TrkA receptor upon NGF binding in receptor activation and dimerization. We 375 showed that full-length TrkA receptors can be activated by specific single-point mutations 376 in the eJTM, in which the position of the mutation relative to the TMD was more important 377 for activation than the dimerization propensity of the mutant. 378 Bearing all of these findings in mind, we propose a mechanism of receptor activation that is 379 outlined in Figure 6. In the model of the Figure 6, NGF binding to the Ig2 will induce a 380 rotation of the eJTM to facilitate the dimerization of the TMD via a preferred interface. This 381 model is supported by our data that showed that cysteine mutants in the eJTM in some 382 specific positions can activate TrkA without ligand and NGF binding brings the N-termini of 383 the α-helices into close proximity. The insertion of Leu residues downstream of the TMD 384 dimerization motif activated TrkA in the absence of ligand. This suggests that rotation of the 385 downstream domain may be behind the activation of the kinase domain. The rotational 386 mechanism of RTK activation has been proposed by other authors (45,46). In this model the 387 ligand would induce a rotation of the TMD dimer interface that will re-orient the kinase 388 domains in order to facilitate the trans-phosphorylation. Although our results support this 389 model of activation other alternative possibilities may exist. For instance the insertion of 390 extra-residues increases the TMD length and could induce a piston-like mechanism of 391 activation by a pulling-down force. However an increase in the length by two, three or four 392 residues should also activate the kinase and this was not the case, as only the Ins1L mutant 393 showed activation. Also as the insertion of the residues are located into the TMD it may alter the dimer interface of the TMD dimer leading to the formation of another dimer interface 395 compatible with a higher activation of the kinase domains. Although we cannot discard this 396 possibility, in the constructs we made the Leu residues are inserted upstream of the active 397 dimer interface in order to not alter the dimer interface found by NMR studies. In any case 398 the results obtained from the insertion mutagenesis suggest a conformational activation of 399  were lysed with TNE buffer (Tris-HCl pH 7.5, 150 mM NaCl, 1 mM EDTA) supplemented 576 with 1% triton X-100 (Sigma), protease inhibitors (Roche), 1 mM PMSF (Sigma), 1 mM 577 sodium orthovanadate (Sigma), and 1 mM sodium fluoride (Sigma). In the experiments 578 involving the TrkA cysteine mutants, 10 mM iodoacetamide (Sigma) was added to the lysis 579 buffer. Lysates were kept on ice for 10 minutes and centrifuged at 12,000 g for 15 minutes 580 in a tabletop centrifuge. The protein level of the lysates was quantified using a Bradford kit 581 (Pierce) and lysates were analyzed by SDS-PAGE 582 583 Western blot analysis 584 Cellular debris was removed by centrifugation at 12,000 g for 15 minutes and the protein 585 level of cell lysates was quantified using the Bradford assay (Pierce). Proteins were resolved 586 in SDS-PAGE gels and transferred to nitrocellulose membranes that were incubated 587 overnight at 4 ºC with one of the following antibodies: mouse monoclonal anti-HA (1:2000, 588 Sigma); rabbit polyclonal MBP-probe (1:1000, Santa Cruz); rabbit anti-phosphoTyr674/5 589 (1:1000, Cell Signaling); rabbit anti-TrkA (1:1000, Millipore). Following incubation with the 590 appropriate secondary antibody, membranes were imaged and bands quantified using 591 enhanced chemiluminescence and autoradiography. 592 593

ToxRED assays 594
Bacterial colonies were inoculated into LB medium (with 50 µg/ml ampicillin). Fresh LB 595 cultures (with 50 µg/ml ampicillin) were inoculated from fresh plates and were grown at 37 596 °C until they reached an O.D. of approximately 600≈0.6. They were then harvested by 597 centrifugation. Cells were subsequently lysed in lysis buffer (TNE, see recipe above) under 598 gentle conditions to avoid mCherry denaturation. After transfer to 1.5-mL Eppendorf tubes, 599 the mixture was incubated at room temperature with gentle agitation for 30 min. Samples 600 were then centrifuged for 10 min at 12000 × g to remove cell debris and to clarify the 601 supernatants for analysis. Aliquots (150 µl) of clarified supernatant were transferred to black, 602 opticlear 96-well plates and mCherry emission spectra were collected using a plate reader 603 (Tecan, Maennedorf, CH) with an excitation wavelength of 587 nm and emission 604 wavelengths of 610-650 nm. Subsequently, aliquots were transferred from black to clear 96-605 well plates and the absorbance was measured from 450 to 750 nm. Measurements of 606 mCherry and of construct expression were performed from at least 10 different colonies and 607 were normalized for the relative expression level of each construct using western blotting 608 with the anti-MBP antibody. For western blots, samples were mixed with equal volumes of 2× SDS-PAGE sample buffer, heated to 95 °C for 10 min, separated on 10% (w/v) 610 polyacrylamide mini-gels, and were blotted onto nitrocellulose membranes that were probed 611 with anti-MBP antibody. To analyze disulfide bond formation bacterial colonies were lysed 612 using TNE plus 10 mM iodoacetamide and were analyzed using non-reducing SDS-PAGE. 613 The cell-free reaction mixture was diluted three-times with buffer A (50 mM Tris pH 8.0 and 646 200 mM NaCl). After 10 minutes of incubation the mixture was centrifuged for 10 min at 647 18000 g at room temperature. The precipitate was washed consecutively with buffer A 648 containing 30 µg/ml RNAse A (Fermentas) and buffer B (50 mM Tris pH 8.0 and 100 mM 649 NaCl). The target protein was solubilized with 200 µl buffer B containing 1% lauryl 650 sarcosine. After each step the protein was centrifuged for 10 min at 18.000 g at room 651 temperature and aliquots of the supernatant were analyzed using 12,5% Tricine SDS-PAGE 652 (53). The clarified protein solution was applied onto a 10/300 Tricorn column prepacked 653 with Superdex 200 (GE Healthcare) and pre-equilibrated with buffer B containing 0.2% 654 lauryl sarcosine. Protein-containing fractions were combined and precipitated using the 655 TCA/acetone procedure (50). 656 657

Preparation of NMR samples in a membrane mimetic media 658
The so-called "isotopic-heterodimer" (1:1 mixture of unlabeled and 15N/ 13 С-labeled peptides) 659 samples were prepared corresponding to the TrkA-TMD construct in order to solve its 660 structure. The powder containing the peptides of both samples was first dissolved in a 1:1 661 trifluoroethanol-H2O mixture with the addition of deuterated DPC (d38, 98%, CIL) and 662 phosphate buffer, and was then kept for several minutes in an ultrasound bath and 663 lyophilized. Subsequently, the dried samples were dissolved in 350 µl of a 9:1 H2O:D2O 664 mixture. To attain a uniform micelle size and uniform distribution of the peptide throughout 665 the micelles, the samples were sonicated in an ultrasound bath for several minutes until the 666 solution was completely transparent. The TrkA-TMD concentration in the isotopic-667 heterodimer sample was 1.9 mM, and other conditions were: LPR 50:1, pH 5.9, and 20 mM 668 phosphate buffer. Samples were placed in Shigemi NMR tubes with a glass plunger. 669 Selective-residue labeling was implemented to avoid peaks overlapping while processing the 670 NMR spectra. 671 672 NMR spectroscopy and spatial structure calculation 673 NMR spectra were acquired at 45 °C using 600 and 800 MHz AVANCE III spectrometers 674 (Bruker BioSpin, Germany) equipped with pulsed-field gradient triple-resonance cryoprobes. 675 1H, 13C, and 15N resonances of TrkA-TMD were assigned with CARA software (51)  As cysteine residues are buried deep in the phospholipid bilayer, membranes were isolated 705 and I2 was used as the oxidation agent following as in (Schwem & Fillingame, 2006). 706 Membrane fractions were diluted in the homogenization buffer (250 mM sucrose, 1mM 707 using a Bardford kit (Pierce) and equal amounts of isolated membranes fractions were 709 incubated for 10 min with or without NGF (10 ng/mL). A solution of 2.5 mM I2 in absolute 710 ethanol was freshly prepared immediately before the crosslinking of cysteine residues and 711 was added to the incubated membrane fractions with NGF ( TrkA exists in an equilibrium between a monomer and an inactive pre-dimer, which is 871 stabilized by interactions between Ig2 domains and between the inactive kinase domains. 872 Upon NGF binding a rotation of the Ig2 domain induces the formation of a crossed TMD 873 dimer that is stabilized by the protein interface determined in the NMR and functional 874 studies. The rotation of the TMD is transmitted to the intracellular region in order to 875 activate TrkA kinase autophosphorylation.     TrkA   wt  S404C  R405C  D406C  V408C  E409C  K410C  K411C  D412C  E413C  T414C   EV  TrkA   wt  S404C  R405C  D406C  V408C  E409C  K410C  K411C  D412C  E413C  T414C