The Acidic Domain and First Immunoglobulin-Like Loop of Fibroblast Growth Factor Receptor 2 Modulate Downstream Signaling through Glycosaminoglycan Modification

Abstract

Fibroblast growth factor receptors (FGFRs) are membrane-spanning tyrosine kinases that have been implicated in a variety of biological processes including mitogenesis, cell migration, development, and differentiation. We identified a unique isoform of FGFR2 expressed as a diffuse band with an unusually large molecular mass. This receptor is modified by glycosaminoglycan at a Ser residue located immediately N terminal to the acidic box, a stretch of acidic amino acids. The acidic box and the glycosaminoglycan modification site are encoded by an alternative exon of the FGFR2 gene. The acidic box appears to play an important role in glycosaminoglycan modification, and the presence of this domain is required for modification by heparan sulfate glycosaminoglycan. Moreover, the presence of the first immunoglobulin-like domain encoded by another alternative exon abrogated the modification. The high-affinity receptor with heparan sulfate modification enhanced receptor autophosphorylation, substrate phosphorylation, and ternary complex factor-independent gene expression. It also sustained mitogen-activated protein kinase activity and increased eventual DNA synthesis, a long-term response to fibroblast growth factor stimulation, at physiological ligand concentrations. We propose a novel regulation mechanism of FGFR2 signal transduction through glycosaminoglycan modification.Various types of growth factors bind to heparin and heparan sulfate. It has been suggested that fibroblast growth factors (FGFs) are stored and protected from proteolytic degradation in the extracellular matrix and on the cell surface by interaction with heparan sulfate proteoglycans, which thereby serve as a reservoir of the growth factors (9,26,30,33,44,45). Heparan sulfate proteoglycans not only play such a modulatory role but are involved in the binding of FGF to its high-affinity receptors as an essential component. Thornton et al. (42) demonstrated that heparin potentiates the mitogenic activity of crude preparations of FGF-1. Further studies of the interaction of FGF-1 with heparin suggested that the potentiation of mitogenic activity is caused by the ability of heparin to increase the affinity of FGF-1 for its receptor (35). It has recently been shown that free heparin or heparan sulfate glycosaminoglycan (HSGAG) is required for high-affinity binding of FGF-2 to FGF receptor (FGFR) 1 expressed in heparan sulfate-deficient CHO cells (50). Based on these findings, a dual-receptor model has been proposed, in which a complex composed of a classical proteintype receptor and a low-affinity glycosaminoglycan-type receptor mediates the cellular responses to FGF (15). FGFR1, FGFR2, FGFR3, and FGFR4, have been identified. A variety of isoforms are generated by alternative splicing of these FGFR genes, which contributes to the functional diversity of this receptor family (13). While the extracellular domain of FGFRs determines ligand-binding specificity (24), the C-terminal domain regulates the basal activity of the receptor (20). The extracellular doma...

Proliferation of Neointimal Smooth Muscle Cells after Arterial Injury

“…3A, lane 4) or the DN-FGFR-3␣IIIcAB adenovirus (Fig. 3A, lane 6) resulted in the predominant expression of the predicted 44-kDa protein with very little post-translational modification evident, which is consistent with findings that the presence of Ig-like domain I in these FGFR-3 isoforms largely abrogates any glycosaminoglycan modification occurring near the acidic box (19). To assess the ability of the DN-FGFR adenoviruses to inhibit neointimal SMC proliferation, carotid arteries injured 7 days earlier were exposed to each adenovirus (100 l of a 5 ϫ 10 9 -plaque-forming unit solution), and 72 h later the effects on DNA synthesis were assessed by immunohistochemical detection and quantification of BrdUrd incorporation (percentage of BrdUrd-positive neointimal SMCs).…”

“…Two days after infection with the DN-FGFR-2␤IIIcAB adenovirus (Fig. 3A, lane 2), SMCs expressed a minor peptide band of 37 kDa, which is the predicted size of the receptor protein in the absence of any significant glycosaminoglycan modification, and an intense broad band of ϳ55-80 kDa, which is consistent with this receptor isoform undergoing extensive post-translational modification by glycosaminoglycans (19). In contrast, infection of the SMCs with either the DN-FGFR-3␣IIIbAB adenovirus (Fig.…”

“…For example, FGF-1 binds to all FGFR isoforms with high affinity, FGF-2 exhibits a high affinity for binding to FGFR-1IIIb, FGFR-1IIIc, and FGFR-3IIIc, and FGF-9 exhibits a high affinity for binding to FGFR-2IIIc, FGFR-3IIIb, and FGFR-3IIIc (17); conversely, FGF-7 and FGF-10 are restricted in binding, with high affinity only to FGFR-2IIIb (14,17,18). Finally, it has become increasingly apparent that the cellular potency of FGFR signaling is also influenced by post-translational receptor modification, in particular the enhanced mitogenic capacity of FGFR-2␤ isoforms arising from glycosaminoglycan modification at the acidic box motif (19).…”

Paracrine and autocrine effects of fibroblast growth factor-4 in endothelial cells

“…Our data demonstrate that human recombinant FGF2 and FGF4 elicit similar biological responses in MAE cells by triggering receptor downstream signaling, exempli®ed by ERK 1/2 phosphorylation (Besser et al, 1995), and stimulating DNA synthesis, although the potency of the mitogenic activity of FGF4 appears to be lower The e ect appears to be speci®c since 2-Odesulfated heparin, that does not interact with FGF4 (Guimond et al, 1993), is ine ective and unmodi®ed heparin does not a ect the receptor-binding capacity of FGF2. Recent data have shown that the acidic box of the FGFR2/IIIc/Ig-2 loops isoform can be covalently modi®ed by HS chains in its Ser-Ser-Gly-motif (Sakaguchi et al, 1999) and that both the receptor core protein and the glycosaminoglycan modi®cation are required for binding. By RT ± PCR analysis of the cellular transcripts, we demonstrate that MAE cells express both the FGFR2/IIIc/Ig-3 loops and the FGFR2/IIIc/Ig-2 loops isoforms and that the latter one contains the Ser-Ser-Gly-motif.…”

“…Nucleotide sequence (not shown) indicates that this fragment actually encodes for the N-terminus of the FGFR2/IIIc/Ig-2 loops variant (Twigg et al, 1998) and contains the HSmodi®ed Ser-Ser-Gly motif in its acidic domain (Sakaguchi et al, 1999) (Figure 2c). Cross-linking of MAE cell surface receptors with 125 I-FGF2, followed by SDS ± PAGE and 1 week exposure of the gel to a FLA2000 PhosphoImager screen (Fuji), showed the presence of a faint broad band with M r 140 000 ± 210 000 (data not shown), as already observed for NIH3T3 cells overexpressing the HS-modi®ed FGFR2/ IIIc/Ig-2 loops isoform (Sakaguchi et al, 1999), suggesting that both the IIIc/Ig-3 loops and the IIIc/ Ig-2 loops isoforms of FGFR2 are exposed on the cell surface. However, the limited number of FGFRs and the coexpression of two isoforms made unfeasible the assessment of the contribution of each individual FGFR2 isoform to FGF2 and FGF4 binding.…”