Search

The three-dimensional solution structure of the ligand binding D2 domain of the fibroblast growth factor receptor (FGFR) is determined using multidimensional NMR techniques. The atomic root-mean-square distribution for the backbone atoms in the structured region is 0.64 A. Secondary structural elements in the D2 domain include 11 beta-strands arranged antiparallely into two layers of beta-sheets. The structure of the D2 domain is characterized by the presence of a short flexible helix that protrudes out of the layers of beta-sheets. Results of size exclusion chromatography and sedimentation velocity experiments show that the D2 domain exists in a monomeric state both in the presence and in the absence of bound sucrose octasulfate (SOS), a structural analogue of heparin. Comparison of the solution structure of the D2 domain with the crystal structure of the protein (D2 domain) in the FGF signaling complex reveals significant differences, suggesting that ligand (FGF) binding may induce significant conformational changes in the receptor. SOS binding sites in the D2 domain have been mapped on the basis of the 1H-15N chemical shift perturbation data. SOS binds to the positively charged residues located in beta-strand III and the flexible helix. Isothermal titration calorimetry data indicate that the ligand (hFGF-1) binds strongly (Kd approximately 10(-9) M) to the D2 domain even in the absence of SOS. Binding of SOS to either the D2 domain or hFGF-1 does not seem to be the driving force for the formation of the D2-hFGF-1 binary complex. The function of SOS binding appears to stabilize the preformed D2-FGF binary complex.

Lateral dimerization of membrane proteins has evolved as a means of signal transduction across the plasma membrane for all receptor tyrosine kinases (RTKs). The transmembrane (TM) domains of RTKs are proposed to play an important role in the dimerization process. We have investigated whether the TM domains of one RTK, fibroblast growth factor receptor 3 (FGFR3), dimerize in lipid vesicles in the absence of the extracellular domains and ligands. We have performed sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) with peptides produced via solid-phase peptide synthesis that correspond to the TM domain of FGFR3. We have carried out Forster resonance energy transfer (FRET) measurements using two donor-acceptor pairs, fluorescein/rhodamine and Cy3/Cy5, as a function of peptide concentration and donor-to-acceptor mole ratios. Our results suggest that FGFR3 TM domains form sequence-specific dimers in lipid bilayers. However, the dimerization propensity of FGFR3 TM domain is much weaker than the dimerization propensity of glycophorin A (GpA), the well-characterized "membrane dimer standard". We discuss our findings in the context of cell signaling across the plasma membrane and diseases or disorders that occur due to single amino acid mutations in the TM domain of FGFR3.

The current working model for fibroblast growth factor receptor (FGFR) dimerization and activation requires the assembly of a ternary complex of fibroblast growth factor (FGF), FGFR, and heparin or heparan sulfate proteoglycan (HSPG) on the plasma membrane. The recent FGF2-FGFR1-heparin crystal structure provides a detailed but static view of the FGF-FGFR-heparin complex. However, the kinetics of ternary complex assembly has yet to be investigated. Here, we characterize FGF2, FGFR1, and heparin interactions using surface plasmon resonance (SPR). Binding constants for binary FGF2/FGFR1 (KD = 62 nM), FGF2/heparin (KD = 39 nM), and FGFR1/heparin (KD = 3.2 microM) interactions correlate to the magnitude of binding interface observed in the FGF2-FGFR1-heparin crystal structure. Interestingly, comparison of sensorgrams of sequential injections of FGF2 and FGFR1 and equimolar FGF2-FGFR1 injections onto a heparin neoproteoglycan surface demonstrates that FGF2 dramatically enhances the association of FGFR1 with heparin and leads us to propose a model for the stepwise assembly of a ternary FGF-FGFR-HSPG complex. The weak binding affinity of the FGFR1-heparin interaction suggests that in this model, FGFR and HSPG are unbound in the absence of FGF ligand. The availability of FGF results in formation of initial FGF-HSPG complexes, which promotes the rapid binding of FGFR and creates a ternary complex capable of undergoing dimerization and subsequent FGFR activation. In contrast, alternative models for the kinetic assembly of a ternary complex in which binary FGF-FGFR or FGFR-HSPG complexes are intermediates do not conform well with the experimental data.

The 22 members of the FGF family have been implicated in cell proliferation, differentiation, survival, and migration. They are required for both development and maintenance of vertebrates, demonstrating an exquisite pattern of affinities for both protein and proteoglycan receptors. FGF19, one of the most divergent human FGFs, is unique in binding solely to one receptor, FGFR4. We have used molecular replacement to solve the crystal structure of FGF19 at 1.3 A resolution using five superimposed FGF structures as the search model. The structure shows that two novel disulfide bonds found in FGF19, one of which appears to be conserved among several of the other FGFs, stabilize extended loops. The key heparin-binding loops of FGF19 have radically different conformations and charge patterns, compared to other FGFs, correlating with the unusually low affinity of FGF19 for heparin. A model for the complex of FGF19 with FGFR4 demonstrates that unique sequences in both FGF19 and FGFR4 are key to the formation of the complex. The structure therefore offers a clear explanation for the unusual affinity of FGF19 for FGFR4 alone.

The conformational stability of the human acidic fibroblast growth factor (hFGF-1) is investigated using amide proton exchange and temperature-dependent chemical shifts, monitored by two-dimensional NMR spectroscopy. The change in free energy of unfolding (DeltaG(u)) of hFGF-1 is estimated to be 5.00 +/- 0.09 kcal.mol(-)(1). Amide proton-exchange rates of 74 residues (in hFGF-1) have been unambiguously measured, and the exchange process occurs predominately according to the conditions of the EX2 limit. The exchange rates of the fast-exchanging amide protons exposed to the solvent have been measured using the clean SEA-HSQC technique. The amide proton protection factor and temperature coefficient estimates show reasonably good correlation. Residues in beta-strands II and VI appear to constitute the stability core of the protein. Among the 12 beta-strands constituting the beta-barrel architecture of hFGF-1, beta-strand XI, located in the heparin binding domain, exhibits the lowest average protection factor value. Amide protons involved in the putative folding nucleation site in hFGF-1, identified by quench-flow NMR studies, do not represent the slow-exchanging core. Residues in portions of hFGF-1 experiencing high conformational flexibility mostly correspond to those involved in receptor recognition and binding.

The assembly and activation of oligomeric complexes of FGF, the transmembrane receptor kinase (FGFR), and heparan sulfate transmit intracellular signals regulating growth and function of cells. An understanding of the structural relationships between the three subunits and their redundancy and specificity is essential for understanding the ubiquitous FGF signaling system in health and disease. Previously, we reported that a primary heparin or heparan sulfate binding site resides in a distinct sequence in immunoglobulin (Ig)-like module II of the three modules of FGFR. Here we report that in the absence of flanking sequences, isolated Ig module II of FGFR1 supports the binding of FGF-1, FGF-2, and FGF-7 in respective order of affinity. None of the three FGFs detectably bind Ig module I or the IIIb and IIIc splice variants of Ig module III in the absence of flanking sequences. Ig module I and the C-terminus of Ig module III are dispensable for high-affinity binding of FGF-1, FGF-2, and FGF-7. Alterations in highly conserved Ig module II in the heparin binding domain and substitution of individual sequence domains spanning the entire sequence of Ig module II with those from Ig module I obliterated FGF binding. Addition of a specific number of FGFR sequences to the C-terminus of Ig module II resulted in a gain in affinity for FGF-7. Several site-specific alterations in the C-terminus of full-length FGFR1IIIc, an isoform that otherwise absolutely rejects FGF-7, resulted in gain of FGF-7 binding. These results suggest that a complex of Ig module II and heparan sulfate is the base common active core of the FGFR ectodomain and that flanking structural domains modify FGF affinity and determine specificity.

Acidic fibroblast growth factor (FGF-1), keratinocyte growth factor (FGF-7), and FGF-10 are homologues with distinct specificity. In the presence of heparin, FGF-1 binds and activates in vitro all FGFR subtypes, while FGF-7 exhibits absolute specificity for the IIIb splice variant of FGFR2. FGF-10 exhibits a similar specificity but also binds the FGFR1IIIb isoform. Neither FGF-7 nor FGF-10 will bind to IIIc isoforms of FGFR. Molecular models of FGF, heparin, and the FGFR ectodomain suggested that sequences between beta-strands 10 and 12 of FGF may be important for the interaction of FGF with the heparin-FGFR ectodomain duplex. Site-directed mutants of FGF-7 and FGF-10 were prepared to test whether this domain might underlie failure of FGF-7 and FGF-10 to bind to the FGFRIIIc isoforms. Constructions with substitution of FGF-1 sequences spanning the entire C-terminus encoded in exon 3 or only C-terminal sequences spanning beta-strands 10 through 12 conferred ability on FGF-7 to bind to and activate FGFRIIIc without a significant loss in binding to or activation of FGFR2IIIb. A series of twelve different substitutions of shorter segments of FGF-1 sequences into the C-terminal portion of FGF-7 or FGF-10 revealed that substitution of GSCKRG for GIPVRG or the tri-peptide sequence KKN for NQK just N-terminal to it conferred dual activities on both the FGF-7 and FGF-10 backbones. The results suggest that the combined sequence domain, which we call the FGF glycine box (G-box), is a major determinant for the specificity of the binding of FGF to heparan sulfate-FGFR duplexes.

The fibroblast growth factor receptor 2 (FGFR2) and the keratinocyte growth factor receptor (KGFR) have different ligand binding specificities despite differing only in the second half of their immunoglobulin-like (Ig-like) domain III. Three-dimensional model structures were generated for domain III on the basis of variable (V) Ig domains. The region that differs between the two receptors is predicted to include two loops: one connects beta-strands F-G and is analogous to the complementarity determining region 3 (CDR3) of immunoglobulins; the other connects beta-strands D-E. These regions were targeted for mutagenesis. Single mutations in the F-G loop were found to only slightly alter ligand binding, whereas a double mutant, KGFR Y345-->S,Q348-->I, acquired significant affinity for bFGF. Notably, the affinity of this double mutant KGFR for KGF and aFGF was essentially unaltered. A mutant FGFR2, in which the D-E beta-hairpin (T319TDKEI) is replaced with the KGFR D-E beta-hairpin (S319SNA), has 9-fold reduced affinity for bFGF. These results demonstrate that the F-G or CDR3 analogous loop in FGFRs plays a key role in determining ligand binding and specificity. In addition, however, the protein loop connecting beta-strands D and E may also be involved in ligand binding. Several point mutations in FGFR2, shown recently to give rise to multiple inherited skeletal defects, are localized according to our models to the F-G or D-E loops of domain III. Our results strongly suggest that these naturally occurring mutations specifically alter ligand binding by FGFR2.

The binding interactions for the three primary reactants of the fibroblast growth factor (FGF) system, basic FGF (bFGF), an FGF receptor, FGFR1, and the cofactor heparin/heparan sulfate (HS), were explored by isothermal titrating calorimetry, ultracentrifugation, and molecular modeling. The binding reactions were first dissected into three binary reactions: (1) FGFR1 + bFGF<==>FGFR1/bFGF, K1 = 41 (+/- 12) nM; (2) FGFR1 + HS<==>FGFR1/HS, K2 = 104 (+/- 17) microM; and (3) bFGF + HS<==>bFGF/HS, K3 = 470 (+/- 20) nM, where HS = low MW heparin, approximately 3 kDa. The first, binding of bFGF to FGFR1 in the absence of HS, was found to be a simple binary binding reaction that is enthalpy dominated and characterized by a single equilibrium constant, K1. The conditional reactions of bFGF and FGFR1 in the presence of heparin were then examined under conditions that saturate only the bFGF heparin site (1.5 equiv of HS/bFGF) or saturate the HS binding sites of both bFGF and FGFR1 (1.0 mM HS). Both 3-and 5-kDa low MW heparins increased the affinity for FGFR1 binding to bFGF by approximately 10-fold (Kd = 4.9 +/- 2.0 nM), relative to the reaction with no HS. In addition, HS, at a minimum of 1.5 equiv/bFGF, induced a second FGFR1 molecule to bind to another lower affinity secondary site on bFGF (K4 = 1.9 +/- 0.7 microM) in an entropy-dominated reaction to yield a quaternary complex containing two FGFR1, one bFGF, and at least one HS. Molecular weight estimates by analytical ultracentrifugation of such fully bound complexes were consistent with this proposed composition. To understand these binding reactions in terms of structural components of FGFR1, a three-dimensional model of FGFR1 was constructed using segment match modeling. Electrostatic potential calculations confirmed that an elongated cluster, approximately 15 x 35 A, of nine cationic residues focused positive potential (+2kBT) to the solvent-exposed beta-sheet A, B, E, C' surface of the D(II) domain model, strongly implicating this locus as the HS binding region of FGFR1. Structural models for HS binding to FGFR1, and HS binding to bFGF, were built individually and then assembled to juxtapose adjacent binding sites for receptor and HS on bFGF, against matching proposed growth factor and HS binding sites on FGFR1. The calorimetric binding results and the molecular modeling exercises suggest that bFGF and HS participate in a concerted bridge mechanism for the dimerization of FGFR1 in vitro and presumably for mitogenic signal transduction in vivo.(ABSTRACT TRUNCATED AT 400 WORDS)