Structural basis for agonism and antagonism of hepatocyte growth factor - PubMed (original) (raw)

Structural basis for agonism and antagonism of hepatocyte growth factor

W David Tolbert et al. Proc Natl Acad Sci U S A. 2010.

Abstract

Hepatocyte growth factor (HGF) is an activating ligand of the Met receptor tyrosine kinase, whose activity is essential for normal tissue development and organ regeneration but abnormal activation of Met has been implicated in growth, invasion, and metastasis of many types of solid tumors. HGF has two natural splice variants, NK1 and NK2, which contain the N-terminal domain (N) and the first kringle (K1) or the first two kringle domains of HGF. NK1, which is a Met agonist, forms a head-to-tail dimer complex in crystal structures and mutations in the NK1 dimer interface convert NK1 to a Met antagonist. In contrast, NK2 is a Met antagonist, capable of inhibiting HGF's activity in cell proliferation without clear mechanism. Here we report the crystal structure of NK2, which forms a "closed" monomeric conformation through interdomain interactions between the N- domain and the second kringle domain (K2). Mutations that were designed to open up the NK2 closed conformation by disrupting the N/K2 interface convert NK2 from a Met antagonist to an agonist. Remarkably, this mutated NK2 agonist can be converted back to an antagonist by a mutation that disrupts the NK1/NK1 dimer interface. These results reveal the molecular determinants that regulate the agonist/antagonist properties of HGF NK2 and provide critical insights into the dimerization mechanism that regulates the Met receptor activation by HGF.

PubMed Disclaimer

Conflict of interest statement

The authors declare no conflict of interest.

Figures

Fig. 1.

Fig. 1.

Heparin independent binding of NK2 to Met (A and B). Binding of NK1 and NK2 to Met, as determined by competition of the binding of biotinylated NK1 to Met in AlphaScreen Assays, in the absence (A) and presence (B) of heparin. Both NK1 and NK2 contain the K132E and R134E mutations that were designated as 1K1 (18). (C and D). Heparin independent binding of NK1 mutants to Met, as determined by AlphaScreen competition curves, for the Met567-NK1 interaction in the absence (A) and presence (B) of heparin. The NK1 mutants tested are: NK1(R134G) (black), NK1(A)—NK1(R134G,R73E) (red), NK1(B)—NK1(R134G,K58A,K60A,T61A,K62A) (blue), NK1(C)—NK1(R134G,K58E,K60E,T61A,K62E) (green), and NK1(D)—NK1(R134G,K58E,K60E,T61A,K62E,R73E) (orange). IC50 values derives from these curves are summarized in Table 1.

Fig. 2.

Fig. 2.

NK2 is a Met antagonist (A) uPA induction of MDCK cells by HGF (60 ng/mL), the NK1(Y124A) antagonist, and NK2. (B) MDCK cell scatter assays for HGF (60 ng/mL), NK1(Y124A) (1 μM), NK2 (0.0125 μM), or HGF + NK1 (Y124A), and HGF + NK2, in the presence or absence of heparin. (C) Inhibition of HGF-mediated uPA activation by NK2 and the NK1(Y124A) antagonist. Experiments are identical to (A) except that 60 ng/mL HGF was added to wells that were pretreated with NK1(Y124A) or NK2. (D) NK2 fails to induce Met dimerization regardless of the presence of heparin. NK1 and NK1 (Y124A) are used as positive and negative controls for their ability to induce Met-dimerization, which is assayed based on a modified AlphaScreen reported previously (8). Values are normalized to the dimerization signal of NK1 at 0.50 μM in the presence of heparin.

Fig. 3.

Fig. 3.

Crystal structure of NK2 (A) NK2 is a monomer. The overall structure of NK2 is represented by a ribbon diagram. Individual domains are designated by N for the N-terminal domain, K1 for the first kringle domain, and K2 for the second kringle domain. (B) The N-domain has a rigid structure. An overlay of the mouse (purple) and human (blue) N-domain crystal structures with the N-domain from the NK2 structure (green) shows the rigid fold of the N-domain. K1 and K2 domains of NK2 are omitted for clarity. (C) Kringle domain has a conserved rigid structure as shown by an overlay of K1 (purple) and K2 (green) domains of NK2. (D) The dimeric structure of NK1 (PDB ID 2QJ2) (blue) is presented for direct comparison of the monomeric NK2 structure (green) represented in (A). (E) The N/K2 interface. The interactions between intramolecular N and K2 domains are are shown as dashed yellow lines and disulfide bonds are shown as solid yellow lines. Nitrogen and oxygen atoms in side chain or main chain atoms are colored blue or red respectively. Only the alpha carbon residues are shown for the residues not involved in either interface. The first kringle domain of the N/K2 interface is omitted for clarity. All structure figures were made with Bobscript (–35) and Raster3D (36).

Fig. 4.

Fig. 4.

Conversion of NK2 to a Met agonist by disrupting the N/K2 interface (A) Mutations that disrupt the N/K2 interface promote the ability of NK2 to induce Met dimerization, which is assayed based on a modified AlphaScreen reported previously (8). Three different mutated NK2 are shown in the presence and absence of heparin. (B) Induction of uPA by HGF (60 ng/mL) and the three NK2 mutated variants (0.0125 μM) in the presence and absence of heparin in MDCK cells. The D257A/N258A NK2 has nearly the same ability to induce uPA activation as HGF. (C) Scattering of MDCK cells by the D257A/N258A NK2 mutant (0.0125 μM) and HGF (60 ng/mL). Top row (-heparin), bottom row (+heparin).

Fig. 5.

Fig. 5.

Conversion of the D257A/N258A NK2 from a Met agonist to a Met antagonist by mutating Y124A in the NK1 interface. (A) uPA induction assays in MDCK cells for HGF (60 ng/mL), the D257A/N258A NK2, and the D257A/N258A NK2 containing the Y124A mutation in the NK1 interface show the Y124A mutation abolishes the ability of the D257A/N258A NK2 to activate uPA. (B) An antagonist assay shows the ability of Y124A mutated NK1 and NK2 (D257A/N258A) to inhibit HGF-mediated activation of uPA in MDCK cells. (C) An antagonist assay shows the ability of Y124A mutated NK1 and NK2 (D257A/N258A) to inhibit HGF-mediated scattering of MDCK cells. The Y124A NK1 and NK2 were used as antagonist controls. (D) The Y124A mutation abolishes the ability of the D257A/N258A NK2 to promote Met-dimerization as compared to Fig. 4_A_.

Similar articles

Cited by

References

    1. Birchmeier C, Birchmeier W, Gherardi E, Vande Woude GF. MET, metastasis, motility, and more. Nature Reviews Molecular Cell Biology. 2003;4:915–925. - PubMed
    1. Buchstein N, et al. Alternative proteolytic processing of hepatocyte growth factor during wound repair. Am J Pathol. 2009;174:2116–2128. - PMC - PubMed
    1. Eder JP, Vande Woude GF, Boerner SA, LoRusso PM. Novel therapeutic inhibitors of the c-Met Signaling pathway in cancer. Clin Cancer Res. 2009;15:2207–2213. - PubMed
    1. Kirchhofer D, et al. Structural and functional basis of the serine protease-like hepatocyte growth factor β-chain in Met binding and signaling. J Biol Chem. 2004;279:39915–39924. - PubMed
    1. Gherardi E, et al. Functional map and domain structure of MET, the product of the c-met protooncogene and receptor for hepatocyte growth factor/scatter factor. Proc Nat’l Acad Sci USA. 2003;100:12039–12044. - PMC - PubMed

Publication types

MeSH terms

Substances

Grants and funding

LinkOut - more resources