The C terminal regions in both FGF and FGF define
The C-terminal regions in both FGF19 and FGF21 define their KLB interaction , , , . However, less than 40% identity in comparative alignment of C-terminal sequences for these two proteins (Table 1) suggests that their interaction with KLB may require higher-order association with other regions within these molecules. Our results reveal that relatively short C-terminal peptides of twenty-five amino acids from both of these endocrine hormones are fully sufficient to support interaction with KLB. Moreover, the limited sequence identity in the C-terminal region of FGF19 and FGF21 defines to a substantial degree the common functional elements of utmost importance to KLB binding. This was clearly demonstrated by the comparative Ala-scan peptides with eleven of the thirteen positional sites that display sizable change in activity being common between the two proteins (Table 1, Group B). The correlative manner in which Ala-substituted peptides block signaling by either hormone (Figure 1C, Table 1) also suggests that FGF19 and FGF21 utilize common KLB binding interfaces. Finally, tight association between activities assessed in signaling and cell-free binding assays (Figure 1D, Table S2) is supportive of the mechanism in peptide action occurring via direct KLB association. These results fit well with the recent report employing a biophysical approach in assessment of structural determinants that define FGF21 interaction with KLB . Additionally, sequence alignment of the C-terminal sequences of FGF19 and FGF21 from numerous species (Fig. S5) reveals a high degree of conservation among the critical residues highlighted by the Ala-scan data, which fortifies the evolutionary basis of retaining important biological functions. Alanine scanning is customarily used to identify locations where decreased activity implies structural importance. It is highly uncommon to observe a logarithmic increase in potency upon substitution with alanine . Therefore, the selective increase in peptide antagonism through a single C-terminal amino need that moment change in 19C26,A26 was unexpected. Through study of additional substitutions, it is nevertheless clear that alanine is not unique in its potency-enhancing properties. Within the additional set of representative changes, only the inversely charged, anionic glutamic acid produced a subtle reduction in potency when compared to 19C26,A26 (Figure 2B). Furthermore, the introduction of a terminal alanine to the FGF21-based C-terminal peptide did not replicate the potency of 19C26,A26 (Table 1), yet replacement of the natural serine in the FGF21 peptide with lysine (21C25,K25) similarly reduced its antagonistic activity (Figure 2A). These results collectively indicate that the presence of a C-terminal lysine decreases potency and infers that other differences in native peptide sequences constitute the basis for higher potency antagonism. Lastly, we investigated the 19C26,A26 super-antagonist for its ability to block FGF signaling in vivo and found it efficiently inhibits the transcriptional activity of FGF19 and FGF21 in pancreas and adipose, two highly relevant target tissues (Figure 5A,B). Thus, this peptide constitutes the smallest functional antagonist to be reported, and it should become a useful reagent to study in vivo physiology of endocrine FGFs. The wide range of peptide-based antagonistic activities identified through Ala-scan (Table 1) led us to investigate whether comparable alanine mutations in full-length proteins would similarly impact FGF agonism. A single site mutation that selectively destroyed peptide-based antagonism demonstrated an analogous effect to eliminate agonism of full-length proteins. Conversely, a benign substitution for peptide-based antagonism was without impact when introduced into the wild type hormone (Figure 3A). Most importantly, the enhanced KLB binding potency of 19C26,A26 translated to a full-length agonist when integrated as a C-terminal substitution in the native FGF21 sequence (FGF21-19A) and yielded a FGF21 analog with increased in vitro activity (Figure 3C,D) that was also pharmacologically superior when studied in DIO mice (Figure 5C,D).