1. Nomenclature
It has been nearly 75 years since an extract of bovine brain was shown to promote proliferation of fibroblasts in cell culture and was called fibroblast growth factor or FGF. About 25 years later, the mitogenic activity was attributed to a purified basic (based on isoelectric point) 15 kDa protein. At around the same time, an acidic form of the protein was identified and the two proteins became known as aFGF and bFGF. Upon the discovery of the binding activity of these proteins to heparan sulfate, heparin affinity chromatography led to the discovery of further proteins of this family. The original two were renamed with numbers, FGF1 and FGF2, and the numbering system was applied to the members of the family discovered through protein purification, biological activity, and genetic analysis over the next few decades.
2. FGF Subfamilies: Structure and Receptor Binding
The structural features and receptor binding activity of the FGFs are summarized in Figure 1. The size of the proteins varies between 155 and 257 amino acids and they share a conserved core sequence of about 120 amino acids (Figure 1A). These are arranged in three sets of four antiparallel β-sheets which make up the globular portion of the molecules. The N- and C-terminal portions form loops which distinguish the shapes of the proteins and account for their differences in receptor binding and activity. The FGFs are divided into numbered subfamilies, as shown by the grouping by color in 1B based on phylogeny and sequence homology.
Fig1. Fibroblast growth factor (FGF) family. A. Structural schematic of FGF structure. The 22-member family of human fibroblast growth factors, each encoded by a separate gene, range in length from 155 to 267 amino acids. All share a common conserved globular core composed of 12 antiparallel β-sheets (blue). The degree of conservation of this core as well as some functional considerations have led to the grouping of the 22 FGFs into seven subfamilies as listed in parts A and B of the figure. These in turn can be placed into three groups based on differences in the N- and C-termini of the proteins which lead to distinct functional differences. Four of the five paracrine subfamilies that act on nearby cells have a signal sequence (SS) which allows them to be secreted from the cell and a heparin-binding sequence (HB) which ensures their accumulation and binding to receptors on nearby cells. FGFs 1 and 2 lack the N-terminal signal sequence and are released by damaged cells or by a non-Golgi exocytotic mechanism. FGFs 11-14, sometimes referred to as FGF homologous factors (FHF), comprise the intracrine family because, having no signal sequence to bring about their secretion, they work in the cells in which they are produced. The members of endocrine subfamily, FGFs-19, -21, and -23, lack a functional heparin binding site and thus are not concentrated in the vicinity of the cell but are released into the circulation to reach distant target cells. B. Binding of FGFs to the FGF receptor. The binding of each FGF in a mitogenic bioassay to the seven forms of the FGF receptor (see legend to Figure 2) is shown by a closed circle. All measurements are relative to FGF1 and, for the purposes of this table, 15% relative binding was the arbitrary cutoff for binding activity. Data is from Zhang et al. (2006). J. Biol. Chem 281:15694–15700 and Ornitz et al. (1996) J. Biol. Chem. 271: 15292–15297.
Fig2. FGF receptor (FGFR) structure. A. FGFR monomer. This diagram represents the common structural features of each of the four separately encoded FGF receptors, FGFR1–FGFR4. The extracellular domain of the FGF monomer consists of three immunoglobulin (Ig)-like loops, D1, D2, and D3. Loops D2 and D3 comprise the ligand binding domain and are separated from loop D1 by a stretch of acidic amino acids referred to as the acid box. Alternative splicing of mRNA in loop D3 (purple) generates the –b and –c forms of FGFRS 1, 2, and 3 listed in Figure 1B (there is an –a form which is a truncated secreted version of the receptor, not shown). In order to form a stable dimer upon ligand binding (panel B) interaction between heparan sulfate and its binding site on the FGFR (red) is required. The FGF receptor has a single pass transmembrane domain (dark green). The intracellular portion consists of a juxtamembrane region followed by a split tyrosine kinase (TK) in which the TK catalytic domain is interrupted by a nonfunctional amino acid sequence. B. FGF-HS-mediated dimerization of FGFR (paracrine FGF subfamilies 1,4,7,8,9). In this diagram, The 2:2:2 complex of FGF, FGFR, and heparan sulfate (HS) is shown. HS is attached to membrane proteoglycan (PG), ensuring its abundance in the extracellular membrane environment, and is required for high-affinity binding of FGF to its receptor and stabilization of the complex shown. C. FGF-klotho-mediated activation of FGFR (endocrine) subfamily 19. FGFs that do not have a heparan sulfate (HS) binding site are released into the circulation and react with distant target cells that have both an FGFR and a membrane bound protein called klotho. The details (for example, dimerization) of klotho-FGFR interactions are not as well understood as those with heparan sulfate, but at least one molecule each of the ligand (FGF, gold), receptor, FGFR (green), and the co-receptor, klotho (blue), are required for activation of the receptor.
Based on their mechanism of action, the FGF sub families are grouped as paracrine, endocrine, or intracrine. The largest of these, the paracrine FGFs (subfamilies FGF-1,- 4, -7, -8, and -9), contain both an N-terminal signal sequence (except for subfamily 1; see legend to Figure 1A) and a heparan sulfate binding region in the C-terminus. The latter is necessary for high-affinity FGF receptor binding. Thus this group of subfamilies are secreted from the cells that make them to bind nearby target cells, in a 2:2:2 complex of FGF:HS:FGFR (see Figure 2). The members of the endocrine group (subfamily 19) lack the heparan sulfate binding site and therefore are not retained in the region in which they are secreted. They are, instead, released into the general circulation for action at dis tant target cells, acting as classical endocrine hormones. As will be discussed following, the binding of these FGFs to the FGF receptor in their target cells requires the presence in the membrane of a co-receptor protein, klotho. Finally, since the FGFs of subfamily 11 lack the signal sequence, they are not secreted from the cells in which they are made and their biological activity is confined to these cells. Figure 1B shows that the six secreted subfamilies of FGFs have distinct receptor binding affinities and therefore different biological activities in their target cells. The receptor subtypes are discussed in section III.B.
FGFs 11-14, which are also known as FGF homologous factors, FHF, are not always included as bona fide members of the FGF family since their mechanism of action differs so substantially from those of the other FGFs. Rather than interacting with the extra cellular domain of the FGF receptor (see Figure 1B) these proteins bind to and regulate the activity of volt age gated sodium channels and are largely active in neurons and cardiomyocytes. The remainder of our discussion of FGFs in this chapter will focus on their interaction with their cognate FGF receptors, consequent intracellular signaling, and biological activities in various cells and tissues and will not, therefore, include the intracrine FGF11 subfamily.
