Identification of tyrosines 154 and 307 in the extracellular domain and 653 and 766 in the intracellular domain as phosphorylation sites in the heparin-binding fibroblast growth factor receptor tyrosine kinase (flg).

Four tyrosine residues have been identified as phosphorylation sites in the tyrosine kinase isoform of the heparin-binding fibroblast growth factor receptor flg (FGF-R1). Baculoviral-insect cell-derived recombinant FGF-R1 was phosphorylated and fragmented with trypsin while immobilized on heparin-agarose beads. Phosphotyrosine peptides were purified by chromatography on immobilized anti-phosphotyrosine antibody and analyzed by Edman degradation and electrospray tandem mass spectrometry. Tyrosine residue 653, which is in a homologous spatial position to major autophosphorylation sites in the catalytic domain of the src and insulin receptor kinases, is the major intracellular FGF-R1 phosphorylation site. Residue 766 in the COOH-terminus outside the kinase domain is a secondary site. Tyrosine residues 154 and 307, which are in the extracellular domain of transmembrane receptor isoforms and are in an unusual sequence context for tyrosine phosphorylation, were also phosphorylated.     

Expression and immunochemical analysis of rat and human fibroblast growth factor receptor (flg) isoforms.

Potentially 96 splice variants among four genes that code for the human heparin-binding fibroblast growth factor receptor family complicate study of structure, metabolism, and function of single isoforms in mammalian cells. As an alternative, we expressed structural subdomains and isoforms of the flg receptor gene in bacteria and baculoviral-infected insect cells. We developed and characterized a panel of 16 isoform and domain-specific polyclonal and monoclonal antibodies. The panel of antibodies was used to distinguish mature glycosylated ligand-binding and kinase-active and -inactive recombinant isoforms in baculoviral insect cells and transfected mammalian cells and natural isoforms in rat prostate and human liver cells. The results revealed a cell type-specific expression of the flg gene and isoforms that result from combinations of splice variations. Reactive epitopes of monoclonal antibodies against both the three (alpha) and two (beta) immunoglobulin-like disulfide loop extracellular domain isoforms were mapped by cross-reactivity with synthetic polypeptide sequences and deletion mutants expressed in bacteria. The native alpha and beta receptor isoforms differed in display of shared epitopes and suggested that the NH2-terminal Loop I and COOH-terminal Loops II and III of the alpha isoform are interactive. Although the common Loops II and III appear qualitatively sufficient for ligand binding, the results suggest that tertiary relationships among loops in the three and two loop isoforms are distinct and, therefore, the two isoforms may have distinct activities. Spatial models for arrangement of immunoglobulin-like loops in the extracellular domain of the two isoforms are presented.     

The expression of two isoforms of the human fibroblast growth factor receptor (flg) is directed by alternative splicing.

Structural analysis of the cDNA for the human fibroblast growth factor receptor (flg) revealed the existence of a larger and a shorter isoform of the receptor. The larger form has three extracellular immunoglobulin-like domains. On the other hand, the shorter form deletes the first (the most external) immunoglobulin-like domain region. Two consecutive amino acids (Arg Met) between the first and second immunoglobulin-like domains are sometimes deleted from the shorter form. In this paper, we isolated and analyzed the gene for the human fibroblast growth factor receptor. Organization of the gene revealed that the isoforms are produced by two different types of alternative splicing (the cassette and internal donor types) from the common gene. In human placenta, the shorter form is expressed as the major isoform.     

Control of fibroblast growth factor receptor kinase signal transduction by heterodimerization of combinatorial splice variants.

A differentiated liver cell (HepG2), which exhibits a dose-dependent growth-stimulatory and growth-inhibitory response to heparin-binding fibroblast growth factor type 1 (FGF-1), displays high- and low-affinity receptor phenotypes and expresses specific combinatorial splice variants alpha 1, beta 1, and alpha 2 of the FGF receptor (FGF-R) gene (flg). The extracellular domains of the alpha and beta variants consist of three and two immunoglobulin loops, respectively, while the intracellular variants consist of a tyrosine kinase (type 1) isoform and a kinase-defective (type 2) isoform. The type 2 isoform is also devoid of the two major intracellular tyrosine autophosphorylation sites (Tyr-653 and Tyr-766) in the type 1 kinase. An analysis of ligand affinity, dimerization, autophosphorylation, and interaction with src homology region 2 (SH2) substrates of the recombinant alpha 1, beta 1, and alpha 2 isoforms was carried out to determine whether dimerization of the combinatorial splice variants might explain the dose-dependent opposite mitogenic effects of FGF. Scatchard analysis indicated that the alpha and beta isoforms exhibit low and high affinity for ligand, respectively. The three combinatorial splice variants dimerized in all combinations. FGF enhanced dimerization and kinase activity, as assessed by receptor autophosphorylation. Phosphopeptide analysis revealed that phosphorylation of Tyr-653 was reduced relative to phosphorylation of Tyr-766 in the type 1 kinase component of heterodimers of the type 1 and type 2 isoforms. The SH2 domain substrate, phospholipase C gamma 1 (PLC gamma 1), associated with the phosphorylated type 1-type 2 heterodimers but was phosphorylated only in preparations containing the type 1 kinase homodimer. The results suggest that phosphorylation of Tyr-653 within the kinase catalytic domain, but not Tyr-766 in the COOH-terminal domain, may be stringently dependent on a trans intermolecular mechanism within FGF-R kinase homodimers. Although phosphotyrosine 766 is sufficient for interaction of PLC gamma 1 and other SH2 substrates with the FGF-R kinase, phosphorylation and presumably activation of substrates require the kinase homodimer and phosphorylation of Tyr-653. We propose that complexes of phosphotyrosine 766 kinase monomers and SH2 domain signal transducers may constitute unactivated presignal complexes whose active or inactive fate depends on homodimerization with a kinase or heterodimerization with a kinase-defective monomer, respectively. The results suggest a mechanism for control of signal transduction by different concentrations of ligand through heterodimerization of combinatorial splice variants from the same receptor gene.     

Mutations uncouple human fibroblast growth factor (FGF)-7 biological activity and receptor binding and support broad specificity in the secondary receptor binding site of FGFs

The fibroblast growth factor (FGF) family plays a key role in a multitude of physiological and pathological processes. The activities of FGFs are mediated by a family of tyrosine kinase receptors, designated FGFRs. The mechanism by which FGFs induce receptor activation is controversial. Despite their structural similarity, FGFs display distinct receptor binding characteristics and cell type specificity. Previous studies with FGF-2 identified a low affinity receptor binding site that is located within a loop connecting its 9th and 10th beta-strands. The corresponding residues in the other family members are highly variable, and it was proposed that the variability might confer on FGFs unique receptor binding characteristics. We studied the role of this loop in FGF-7 by both site-directed mutagenesis and loop replacement. Unlike the other members of the FGF family, FGF-7 recognizes only one FGFR isoform and is, therefore, ideal for studies of how the specificity in the FGF-FGFR interaction is conferred at the structural level. Point mutations in the loop of FGF-7 did not change receptor binding affinity but resulted in reduced mitogenic potency and reduced ability to induce receptor-mediated phosphorylation events. These results suggest that the loop of FGF-7 fulfills the role of low affinity binding site required for receptor activation. The observation that it is possible to uncouple FGF-7 receptor binding and biological activity favors a bivalent model for FGFR dimerization, and it may be clinically relevant to the design of FGF-7 antagonists. Reciprocal loop replacement between FGF-7 and FGF-2 had no effect on their known receptor binding affinities nor did it alter their known specificity in eliciting a mitogenic response. In conclusion, these results suggest that, despite the diversity in the loop structure of FGF-2 and FGF-7, the loop has a similar function in both growth factors.     

Involvement of the conserved acidic amino acid domain of FGF receptor 1 in ligand-receptor interaction

The fibroblast growth factor receptor 1 (flg) contains eight acidic amino acids between the first and second immunoglobulin domain. This report examines the role of the acidic domain in the interaction of the flg receptor with its ligands. We observed a marked inhibition of binding of bFGF to the receptor when the acidic domain was completely deleted, but mutants with two and four amino acids deleted (flgΔ₂ and flgΔ₄, respectively) still bound the ligand. After addition of a bifunctional cross-linking reagent, cross-linked complexes (between bFGF and receptor) with the expected size were observed in cells expressing mutants lacking two or four acidic residues, but not in cells expressing mutants lacking six or eight acidic residues. Immunoprecipitation with anti-flg antibody followed by electrophoresis produced a band of 90 Kd in tunicamycin-treated cells expressing the mutant as well as the wild-type receptors, indicating that the inhibition of binding was not due to defective expression of the protein. The ability of flgΔ₈ to mediate a mitogenic response to FGFs was also greatly reduced when this mutated receptor was expressed in receptor-negative cells. The effect of replacing the acidic amino acids with lysine residues was also studied. Binding of bFGF to cells transfected with a plasmid encoding a mutated protein with four amino acid substitutions was totally inhibited, but an eight amino acid substitution did not alter ligand binding to the receptor. In this case the mutation with four amino acids substitution caused a drastic impairment of protein expression. Thus the acidic domain of the FGFR-1 plays an essential role in receptor function, either because it is important for a stable protein configuration or for ligand-receptor interaction. © 1993 Wiley-Liss, Inc.     

Decidualization and implantation: embryo-uterine bioinformatics at work

The implantation of the blastocyst into a nurturing endometrium involves two overlapping steps: 1. The blastocyst-endometrial luminal epithelial attachment. 2. The decidualization of the endometrial stroma. An intriguing question is how does the blastocyst identify the uterine implantation site. Current research is focused on hypothetical soluble signaling molecules released by the blastocyst for conditioning a discrete uterine luminal epithelial domain for implantation. A still unresolved issue is the functional significance of receptor autophosphorylation following binding of uterine epithelial cell-derived heparin-binding epidermal growth factor-like growth factor to the epidermal growth factor receptor on trophoectodermic cell surfaces. With recent results hinting at the role of signaling proteins associated with the bone morphogenetic protein, fibroblast growth factor, WNT and hedgehog families to enable embryo implantation, the dynamics of uterine-embryo interaction becomes linked to fundamental cellular pathways of growth, differentiation and apoptosis.     

Substitution of putative half-cystine residues in heparin-binding fibroblast growth factor receptors. Loss of binding activity in both two and three loop isoforms.

Alternate use of an exon coding for an 89-residue NH2 terminal immunoglobulin-like disulfide loop results in isoforms of the heparin-binding fibroblast growth factor receptor (FGF-R) with three (FGF-R alpha) and two (FGF-R beta) Ig-like loops in the extracellular domain. Both FGF-R alpha and FGF-R beta isoforms exhibit qualitatively similar ligand-binding activities. In this report, we show by site-directed mutagenesis and analysis of ligand-binding activity in transfected cells that substitution of a cysteine that potentially forms an intra-loop disulfide in either juxtamembrane Loop II or III disrupted maturation and formation of the ligand-binding site in both FGF-R alpha and FGF-R beta isoforms. Neither three loop FGF-R alpha constructions coding for intact Loops I and II adjacent to defective Loop III nor intact Loops I and III separated by defective Loop II exhibited ligand-binding activity. In addition, a two-loop molecule of tandem Loops I and III was inactive. The results suggest that single Loops I, II, or III of FGF-R are insufficient to form a ligand-binding site. Loop I does not form an independent ligand-binding site with either Loop II or III, but interacts with a common ligand-binding site formed by Loops II and III (Xu, J., Nakahara, M., Crabb, J. W., Shi, E., Matuo, Y., Fraser, M., Kan, M., Hou, J., and McKeehan, W. L. (1992) J. Biol. Chem. 267, 17792-17803, 1992).     

A tyrosine-phosphorylated carboxy-terminal peptide of the fibroblast growth factor receptor (Flg) is a binding site for the SH2 domain of phospholipase C-gamma 1.

Phospholipase C-gamma (PLC-gamma) is a substrate of the fibroblast growth factor receptor (FGFR; encoded by the flg gene) and other receptors with tyrosine kinase activity. It has been demonstrated that the src homology region 2 (SH2 domain) of PLC-gamma and of other signalling molecules such as GTPase-activating protein and phosphatidylinositol 3-kinase-associated p85 direct their binding toward tyrosine-autophosphorylated regions of the epidermal growth factor or platelet-derived growth factor receptor. In this report, we describe the identification of Tyr-766 as an autophosphorylation site of flg-encoded FGFR by direct sequencing of a tyrosine-phosphorylated tryptic peptide isolated from the cytoplasmic domain of FGFR expressed in Escherichia coli. The same phosphopeptide was found in wild-type FGFR phosphorylated either in vitro or in living cells. Like other growth factor receptors, tyrosine-phosphorylated wild-type FGFR or its cytoplasmic domain becomes associated with intact PLC-gamma or with a fusion protein containing the SH2 domain of PLC-gamma. To delineate the site of association, we have examined the capacity of a 28-amino-acid tryptic peptide containing phosphorylated Tyr-766 to bind to various constructs containing SH2 and other domains of PLC-gamma. It is demonstrated that the tyrosine-phosphorylated peptide binds specifically to the SH2 domain but not to the SH3 domain or other regions of PLC-gamma. Hence, Tyr-766 and its flanking sequences represent a major binding site in FGFR for PLC-gamma. Alignment of the amino acid sequences surrounding Tyr-766 with corresponding regions of other FGFRs revealed conserved tyrosine residues in all known members of the FGFR family. We propose that homologous tyrosine-phosphorylated regions in other FGFRs also function as binding sites for PLC-gamma and therefore are involved in coupling to phosphatidylinositol breakdown.     

Identification of residues important both for primary receptor binding and specificity in fibroblast growth factor-7

Fibroblast growth factors (FGFs) mediate a multitude of physiological and pathological processes by activating a family of tyrosine kinase receptors (FGFRs). Each FGFR binds to a unique subset of FGFs and ligand binding specificity is essential in regulating FGF activity. FGF-7 recognizes one FGFR isoform known as the FGFR2 IIIb isoform or keratinocyte growth factor receptor (KGFR), whereas FGF-2 binds well to FGFR1, FGFR2, and FGFR4 but interacts poorly with KGFR. Previously, mutations in FGF-2 identified a set of residues that are important for high affinity receptor binding, known as the primary receptor-binding site. FGF-7 contains this primary site as well as a region that restricts interaction with FGFR1. The sequences that confer on FGF-7 its specific binding to KGFR have not been identified. By utilizing domain swapping and site-directed mutagenesis we have found that the loop connecting the beta4-beta5 strands of FGF-7 contributes to high affinity receptor binding and is critical for KGFR recognition. Replacement of this loop with the homologous loop from FGF-2 dramatically reduced both the affinity of FGF-7 for KGFR and its biological potency but did not result in the ability to bind FGFR1. Point mutations in residues comprising this loop of FGF-7 reduced both binding affinity and biological potency. The reciprocal loop replacement mutant (FGF2-L4/7) retained FGF-2 like affinity for FGFR1 and for KGFR. Our results show that topologically similar regions in these two FGFs have different roles in regulating receptor binding specificity and suggest that specificity may require the concerted action of distinct regions of an FGF.     

Interactions of putative heparin-binding domains of basic fibroblast growth factor and its receptor, FGFR-1, with heparin using synthetic peptides

We have examined structure-function relationships that have been proposed to account for the heparin-binding properties of basic fibroblast growth factor and its receptor, FGFR-1, using synthetic peptides, DNA synthesis assays and binding assays in a resonant mirror biosensor. The results suggest that the interaction of FGFR-1 with heparin may not be physiologically relevant and that the site of interaction of the polysaccharide on bFGF is more complex than has been anticipated. © 1998 Rapid Science Ltd     

FGFR-4, a novel acidic fibroblast growth factor receptor with a distinct expression pattern.

We have previously identified two novel members of the fibroblast growth factor receptor (FGFR) gene family expressed in K562 erythroleukemia cells. Here we report cDNA cloning and analysis of one of these genes, named FGFR-4. The deduced amino acid sequence of FGFR-4 is 55% identical with both previously characterized FGFRs, flg and bek, and has the structural characteristics of a FGFR family member including three immunoglobulin-like domains in its extracellular part. Antibodies raised against the carboxy terminus of FGFR-4 detected 95 and 110 kd glycoproteins with a protein backbone of 88 kd in COS cells transfected with a FGFR-4 cDNA expression vector. The FGFR-4 protein expressed in COS cells could also be affinity-labeled with radioiodinated acidic FGF. Furthermore, ligand binding experiments demonstrated that FGFR-4 binds acidic FGF with high affinity but does not bind basic FGF. FGFR-4 is expressed as a 3.0 kb mRNA in the adrenal, lung, kidney, liver, pancreas, intestine, striated muscle and spleen tissues of human fetuses. The expression pattern of FGFR-4 is distinct from that of flg and bek and the yet additional member of the same gene family, FGFR-3, which we have also cloned from the K562 leukemia cells. Our results suggest that FGFR-4 along with other fibroblast growth factor receptors performs cell lineage and tissue-specific functions.     

Association of the signaling adaptor FRS2 with fibroblast growth factor receptor 1 (Fgfr1) is mediated by alternative splicing of the juxtamembrane domain.

Fibroblast growth factor receptors (FGFRs) are a family of transmembrane tyrosine kinases involved in signaling via interactions with the family of fibroblast growth factors. Alternative splicing of the juxtamembrane region of FGFR1-3 leads to the inclusion or exclusion of two amino acids, valine and threonine, the VT site. The presence or absence of VT (VT+ or VT-, respectively) affects the signaling potential of the receptor. The VT+ receptor isoform is required for Erk2 phosphorylation, a component of the mitogen-activated protein kinase signaling pathway. FRS2 is an adaptor protein that links FGFRs to the mitogen-activated protein kinase signaling pathway. FRS2 interacts with a region of the juxtamembrane domain of FGFR1 that includes the alternatively spliced VT site. We investigated the interaction of FRS2 with murine Fgfr1 juxtamembrane domain. We showed the alternatively spliced VT motif, at the juxtamembrane domain of Fgfr1 is required for FRS2 interaction with Fgfr1. Activation of signaling pathways from FRS2 is likely to be regulated by controlling the Fgfr1/FRS2 interaction through alternative splicing of the VT motif of Fgfr1.     

Structure-based mutational analyses in FGF7 identify new residues involved in specific interaction with FGFR2IIIb

Receptor binding specificity is an essential element in regulating the diverse activities of fibroblast growth factors (FGFs). FGF7 is ideal to study how this specificity is conferred at the structural level, as it interacts exclusively with one isoform of the FGF-receptor (FGFR) family, known as FGFR2IIIb. Previous mutational analysis suggested the importance of the beta4/beta5 loop of FGF7 in specific receptor recognition. Here a theoretical model of FGFR2IIIb/FGF7 complex showed that this loop interacts with the FGFR2IIIb unique exon. In addition, the model revealed new residues that either directly interact with the FGFR2IIIb unique exon (Asp63, Leu142) or facilitate this interaction (Arg65). Mutations in these residues reduced both receptor binding affinity and biological activity of FGF7. Altogether, these results provide the basis for understanding how receptor-binding specificity of FGF7 is conferred at the structural level.     

The cell surface receptor FGFRL1 forms constitutive dimers that promote cell adhesion

FGFRL1 is a novel member of the fibroblast growth factor (FGF) receptor family. Utilizing the FRET (fluorescence resonance energy transfer) technique, we demonstrate that FGFRL1 forms constitutive homodimers at cell surfaces. The formation of homodimers was verified by co-precipitation of differentially tagged FGFRL1 polypeptides from solution. If overexpressed in cultivated cells, FGFRL1 was found to be enriched at cell-cell contact sites. The extracellular domain of recombinant FGFRL1 promoted cell adhesion, but not cell spreading, when coated on plastic surfaces. Adhesion was mediated by heparan sulfate glycosaminoglycans located at the cell surface. It could specifically be blocked by addition of soluble heparin but not by addition of other glycosaminoglycans. When the amino acid sequence of the putative heparin-binding site was modified by in vitro mutagenesis, the resulting protein exhibited decreased affinity for heparin and reduced activity in the cell-binding assay. Moreover, a synthetic peptide corresponding to the heparin-binding site was able to neutralize the effect of heparin. With its dimeric structure and its adhesion promoting properties, FGFRL1 resembles the nectins, a family of cell adhesion molecules found at cell-cell junctions.     

Interactions of multiple heparin binding growth factors with neuropilin-1 and potentiation of the activity of fibroblast growth factor-2

The hypothesis that neuropilin-1 (Npn-1) may interact with heparin-binding proteins other than vascular endothelial growth factor has been tested using an optical biosensor-based binding assay. The results show that fibroblast growth factor (FGF) 1, 2, 4, and 7, FGF receptor 1, hepatocyte growth factor/scatter factor (HGF/SF), FGF-binding protein, normal protease sensitive form of prion protein, antithrombin III, and Npn-1 itself are all able to interact with Npn-1 immobilized on the sensor surface. FGF-2, FGF-4, and HGF/SF are also shown to interact with Npn-1 in a solution assay. Moreover, these protein-protein interactions are dependent on the ionic strength of the medium and are inhibited by heparin, and the kinetics of binding of FGF-2, FGF-4 and HGF/SF to Npn-1 are characterized by fast association rate constants (270,000-1,600,000 m(-1) s(-1)). These results suggest that Npn-1 possesses a "heparin" mimetic site that is able to interact at least in part through ionic bonding with the heparin binding site on many of the proteins studied. Npn-1 was also found to potentiate the growth stimulatory activity of FGF-2 on human umbilical vein endothelial cells, indicating that Npn-1 may not just bind but also regulate the activity of heparin-binding proteins.