Structure and functional analysis of the IGF-II/IGF2R interaction
2007; Springer Nature; Volume: 27; Issue: 1 Linguagem: Inglês
10.1038/sj.emboj.7601938
ISSN1460-2075
AutoresJames Brown, Carlie Delaine, Oliver Zaccheo, Christian Siebold, Robert J. C. Gilbert, Gijs van Boxel, Adam Denley, John C. Wallace, A. Bassim Hassan, Briony E. Forbes, E. Yvonne Jones,
Tópico(s)Glycosylation and Glycoproteins Research
ResumoArticle29 November 2007free access Structure and functional analysis of the IGF-II/IGF2R interaction James Brown James Brown Cancer Research UK Receptor Structure Research Group, Division of Structural Biology, Wellcome Trust Centre for Human Genetics, University of Oxford, Headington, Oxford, UK Search for more papers by this author Carlie Delaine Carlie Delaine School of Molecular and Biomedical Science, The University of Adelaide, Adelaide, Australia Search for more papers by this author Oliver J Zaccheo Oliver J Zaccheo Tumour Growth Control Group, Weatherall Institute of Molecular Medicine, John Radcliffe Hospital, University of Oxford, Headington, Oxford, UK Search for more papers by this author Christian Siebold Christian Siebold Cancer Research UK Receptor Structure Research Group, Division of Structural Biology, Wellcome Trust Centre for Human Genetics, University of Oxford, Headington, Oxford, UK Search for more papers by this author Robert J Gilbert Robert J Gilbert Cancer Research UK Receptor Structure Research Group, Division of Structural Biology, Wellcome Trust Centre for Human Genetics, University of Oxford, Headington, Oxford, UK Search for more papers by this author Gijs van Boxel Gijs van Boxel Cancer Research UK Receptor Structure Research Group, Division of Structural Biology, Wellcome Trust Centre for Human Genetics, University of Oxford, Headington, Oxford, UK Search for more papers by this author Adam Denley Adam Denley School of Molecular and Biomedical Science, The University of Adelaide, Adelaide, Australia Search for more papers by this author John C Wallace John C Wallace School of Molecular and Biomedical Science, The University of Adelaide, Adelaide, Australia Search for more papers by this author A Bassim Hassan A Bassim Hassan Tumour Growth Control Group, Weatherall Institute of Molecular Medicine, John Radcliffe Hospital, University of Oxford, Headington, Oxford, UK Search for more papers by this author Briony E Forbes Briony E Forbes School of Molecular and Biomedical Science, The University of Adelaide, Adelaide, Australia Search for more papers by this author E Yvonne Jones Corresponding Author E Yvonne Jones Cancer Research UK Receptor Structure Research Group, Division of Structural Biology, Wellcome Trust Centre for Human Genetics, University of Oxford, Headington, Oxford, UK Search for more papers by this author James Brown James Brown Cancer Research UK Receptor Structure Research Group, Division of Structural Biology, Wellcome Trust Centre for Human Genetics, University of Oxford, Headington, Oxford, UK Search for more papers by this author Carlie Delaine Carlie Delaine School of Molecular and Biomedical Science, The University of Adelaide, Adelaide, Australia Search for more papers by this author Oliver J Zaccheo Oliver J Zaccheo Tumour Growth Control Group, Weatherall Institute of Molecular Medicine, John Radcliffe Hospital, University of Oxford, Headington, Oxford, UK Search for more papers by this author Christian Siebold Christian Siebold Cancer Research UK Receptor Structure Research Group, Division of Structural Biology, Wellcome Trust Centre for Human Genetics, University of Oxford, Headington, Oxford, UK Search for more papers by this author Robert J Gilbert Robert J Gilbert Cancer Research UK Receptor Structure Research Group, Division of Structural Biology, Wellcome Trust Centre for Human Genetics, University of Oxford, Headington, Oxford, UK Search for more papers by this author Gijs van Boxel Gijs van Boxel Cancer Research UK Receptor Structure Research Group, Division of Structural Biology, Wellcome Trust Centre for Human Genetics, University of Oxford, Headington, Oxford, UK Search for more papers by this author Adam Denley Adam Denley School of Molecular and Biomedical Science, The University of Adelaide, Adelaide, Australia Search for more papers by this author John C Wallace John C Wallace School of Molecular and Biomedical Science, The University of Adelaide, Adelaide, Australia Search for more papers by this author A Bassim Hassan A Bassim Hassan Tumour Growth Control Group, Weatherall Institute of Molecular Medicine, John Radcliffe Hospital, University of Oxford, Headington, Oxford, UK Search for more papers by this author Briony E Forbes Briony E Forbes School of Molecular and Biomedical Science, The University of Adelaide, Adelaide, Australia Search for more papers by this author E Yvonne Jones Corresponding Author E Yvonne Jones Cancer Research UK Receptor Structure Research Group, Division of Structural Biology, Wellcome Trust Centre for Human Genetics, University of Oxford, Headington, Oxford, UK Search for more papers by this author Author Information James Brown1, Carlie Delaine2, Oliver J Zaccheo3, Christian Siebold1, Robert J Gilbert1, Gijs van Boxel1, Adam Denley2, John C Wallace2, A Bassim Hassan3, Briony E Forbes2 and E Yvonne Jones 1 1Cancer Research UK Receptor Structure Research Group, Division of Structural Biology, Wellcome Trust Centre for Human Genetics, University of Oxford, Headington, Oxford, UK 2School of Molecular and Biomedical Science, The University of Adelaide, Adelaide, Australia 3Tumour Growth Control Group, Weatherall Institute of Molecular Medicine, John Radcliffe Hospital, University of Oxford, Headington, Oxford, UK *Corresponding author. Cancer Research UK Receptor Structure Research Group, Division of Structural Biology, Wellcome Trust Centre for Human Genetics, University of Oxford, Roosevelt Drive, Headington, Oxford, OX3 7BN, UK. Tel.: +44 1865 287546; Fax: +44 1865 287547; E-mail: [email protected] The EMBO Journal (2008)27:265-276https://doi.org/10.1038/sj.emboj.7601938 PDFDownload PDF of article text and main figures. ToolsAdd to favoritesDownload CitationsTrack CitationsPermissions ShareFacebookTwitterLinked InMendeleyWechatReddit Figures & Info Embryonic development and normal growth require exquisite control of insulin-like growth factors (IGFs). In mammals the extracellular region of the cation-independent mannose-6-phosphate receptor has gained an IGF-II-binding function and is termed type II IGF receptor (IGF2R). IGF2R sequesters IGF-II; imbalances occur in cancers and IGF2R is implicated in tumour suppression. We report crystal structures of IGF2R domains 11–12, 11–12–13–14 and domains 11–12–13/IGF-II complex. A distinctive juxtaposition of these domains provides the IGF-II-binding unit, with domain 11 directly interacting with IGF-II and domain 13 modulating binding site flexibility. Our complex shows that Phe19 and Leu53 of IGF-II lock into a hydrophobic pocket unique to domain 11 of mammalian IGF2Rs. Mutagenesis analyses confirm this IGF-II 'binding-hotspot', revealing that IGF-binding proteins and IGF2R have converged on the same high-affinity site. Introduction The insulin-like growth factor (IGF)-system includes insulin-like growth factors I and II (IGF-I and IGF-II) along with the type I (IGF-1R) and type II (IGF2R) cell-surface receptors, the insulin receptor (IR) and circulating IGF-binding proteins (IGFBPs) (Denley et al, 2005). The biological actions of the IGFs are mediated by IGF-1R and IR, leading to cell growth, differentiation and survival. Their distribution and activity is controlled via high-affinity association with IGFBPs, and the binding sites on IGFs for IGFBPs have been delineated in detail by structural studies (Headey et al, 2004; Carrick et al, 2005; Sitar et al, 2006). In mammals, the activity of IGF-II (but not IGF-I) is further moderated by IGF2R, which sequesters IGF-II for internalization and degradation. IGF2R is classed as a growth inhibitor, with loss of function causing increased growth (Foulstone et al, 2005). In line with this, Igf2r is a putative tumour suppressor gene and mutations have been found in several cancers (reviewed in Falls et al, 1999). Unusually high levels of circulating IGF-II and simultaneous downregulation of IGF2R are correlated with the growth of human and murine tumours (Toretsky and Helman, 1996; Hassan and Howell, 2000). Hence, disruption of IGF-II action is a potential method of tumour control, with the natural scavenging action of IGF2R an obvious tool for such intervention. IGF2R, also known as the cation-independent mannose-6-phosphate receptor, is found ubiquitously in human tissues with a truncated soluble form of the receptor present in the circulation (Lobel et al, 1988; Oshima et al, 1988). Full-length, 300-kDa, IGF2R comprises a large N-terminal extracellular region of 15 homologous domains, a single membrane-spanning region and a small cytoplasmic tail. In addition to IGF-II binding, major IGF2R functions include sorting newly synthesized lysosomal enzymes and endocytosis of extracellular lysosomal enzymes (reviewed in Ghosh et al, 2003). To perform these disparate functions, the extracellular region contains binding sites for IGF-II and phosphomannosyl residues (Oshima et al, 1988; Schmidt et al, 1995). Domain 11, the first in the IGF2R extracellular region to be characterized by X-ray crystallography, contains the putative IGF-II-binding site (Schmidt et al, 1995; Brown et al, 2002). Mannosylated proteins are bound by domains 3, 5 and 9 (Reddy et al, 2004). Sequence alignments suggest that all 15 domains share the same β-barrel architecture and the crystal structure of a domain 1–3 fragment supports this hypothesis (Olson et al, 2004). The only large deviation from the canonical repeat occurs in domain 13, where sequence analysis predicts that an insertion forms an independent module with a fibronectin type II fold (FNII). Domain 13 enhances IGF-II binding, an effect believed in someway to be mediated by FNII (Lobel et al, 1988), resulting in a reduction in the rate of IGF-II release (Devi et al, 1998; Linnell et al, 2001). An Ile1572Thr mutation in IGF2R domain 11 essentially abrogates IGF-II binding (Garmroudi et al, 1996; Linnell et al, 2001). The crystal structure of domain 11 mapped Ile1572 to a hydrophobic pocket, which was therefore proposed as the putative IGF-II-binding site (Brown et al, 2002); subsequently a cluster of residues lining this pocket have been implicated in IGF-II binding by mutational analyses (Zaccheo et al, 2006). In contrast, a series of mutational analyses of IGF-II have highlighted effects on IGF2R binding for a disparate set of residues (Burgisser et al, 1991; Sakano et al, 1991; Delaine et al, 2007). Sequence comparisons across species potentially harbour additional functional insights. IGF2R gained IGF-II-binding activity coincident with the evolution of placental development; the cation-independent mannose-6-phosphate receptors of non-mammalian species, including monotremes such as platypus, as well as the evolutionarily more distant chicken and Xenopus, are unable to bind IGF-II, but those of marsupials and mammals can (Killian et al, 2000). Thus, there is a relative wealth of functional data to relate to the IGF-II/IGF2R interaction; however, the lack of a definitive structural context has hindered development of a robust understanding of this system. Major questions regarding the IGF-II/IGF2R interaction remain unanswered, including precise definition of the residues forming the interface and understanding the role of domain 13 in raising the affinity for IGF-II. We still do not understand how the structurally homologous domains are arranged in the full-length receptor and whether any rules drive this tertiary assembly. To address these questions, we report the crystal structure of a complex between IGF-II and domains 11–13 of IGF2R, complemented by crystal structures of IGF2R domains 11–12 and 11–12–13–14 in isolation. We have validated our structure-based hypotheses with additional functional analyses of the IGF-II/IGF2R interaction using surface plasmon resonance and mutation of both IGF-II and IGF2R. Our conclusions allow us to analyse the IGF-II/IGF2R interaction as an example of evolutionary gain-of-function, and to uncover features common to IGF-II/IGFBP and IGF-II/IGF2R binding. The latter insights provide potentially important caveats to consider in the development of novel anti-cancer therapies. Results The structures of IGF2R fragments and the IGF2R/IGF-II complex Three multi-domain IGF2R fragments were generated: IGF2R-Dom11–12 (domains 11 and 12), IGF2R-Dom11–13 (domains 11, 12 and 13) and IGF2R-Dom11–14 (domains 11, 12, 13 and 14). Crystal structure determinations were achieved for IGF2R-Dom11–12 (3.2 Å resolution), IGF2R-Dom11–14 (2.9 Å resolution) and for an IGF2R-Dom11–13/IGF-II complex (4.1 Å resolution) (summarized in Materials and methods; Table I and Supplementary Table SI). As expected, the structures of domains 12, 13 and 14 of IGF2R closely resemble that of IGF2R-Dom11 (Figure 1; Supplementary Figure S1A). The root-mean-squared deviation between the domains for which structures are now available (1, 2, 3, 11, 12, 13 and 14) ranges between 1.8 and 2.5 Å for an average of 125 Cα equivalences (Supplementary Table SII). The core domain structure is a flattened β-barrel consisting of nine β-strands, termed βA–βI forming two crossed β-sheets, the first formed by βA–βD and the second by βE–βI. Each of domains 11–14 is labelled with the standard-strand nomenclature in Figure 1B. The N-terminal region preceding βA contains two additional β-strands (βNA′ and βNA″), which form a β-hairpin capping off the β-barrel. Each domain contains four conserved disulphide bonds, with domain 13 having an extra two disulphide bonds in the FNII insert. FNII may be defined as a 48-residue insert spanning Glu1896 to Arg1943, which adopts the typical FNII fold (Figure 1B and C; Supplementary Figure S1B); four β-strands form two sheets of two antiparallel β-strands stabilized by two disulphide bonds. In addition, the first two FNII residues form a third β-strand in the lower β-sheet and a short α-helix packs against the top β-sheet. Figure 1.Cartoon representations of IGF2R fragment structures. (A) IGF2R-Dom11–12. (B) IGF2R-Dom11–14. (C) The complex between IGF2R-Dom11–13 and IGF-II. Glycans are shown as spheres and strand labelling is given in panel B. Download figure Download PowerPoint Table 1. Statistics for crystallographic data collection and refinement Multiple anomalous dispersion data Peak Inflection Remote Resolution range (Å) 30.0–3.2 30.0–3.2 30.0–3.2 P6122 cell dimensions (Å) 64.7, 64.7, 269.5 64.8, 64.8, 269.9 64.8, 64.8, 269.7 Wavelength (Å) 0.9789 0.9791 0.8856 Unique reflectionsa 6204 (592)b 6183 (594)b 6147 (581)b Completeness (%)a 99.9 (100) 99.1 (100) 100 (100) I/σIa 27.0 (3.8) 40.0 (5.0) 44.0 (5.2) Redundancya 14.9 (15.5) 15.7 (16.7) 19.7 (20.6) Rmerge (%)a 9.2 (69.5) 7.9 (81.3) 8.4 (79.0) Native data IGF2R-Dom11–12 IGF2R-Dom11–14 Complex Data collection Resolution range (Å)a 20–3.2 (3.3–3.2) 30–2.9 (3.0–2.9) 50–4.1 (4.2–4.1) Space group P6122 C2 C2 Cell dimensions (Å) 64.8, 64.8, 269.6 138.8, 69.3, 97.5; β=103.5° 165.9, 116.9, 116.6; β=123.4° Wavelength (Å) 0.9789 0.931 0.933 Unique reflections 6123 19 919 14 662 Completeness (%)a 99.5 (96.6) 98.6 (92.8) 99.0 (99.8) I/σIa 25.5 (4.9) 13.9 (3.0) 5.4 (2.5) Redundancya 19.6 (19.5) 3.7 (3.6) 3.7 (3.7) Rmerge (%)a 9.4 (74.4) 7.7 (49.9) 28.8 (54.5) Refinement No. of reflections 5495 18 893 13 919 Rfactor/Rfree (%) 25.3/31.7 25.7/30.4 29.1/32.7 Non-hydrogen protein atoms 2017 4642 7880 Non-hydrogen sugar atoms 28 42 109 R.m.s.d bonds (Å) 0.007 0.007 0.006 R.m.s.d angles (deg) 1.132 1.164 0.906 Overall B-factor (Å2) 97.6 65.8 90.4 Ramachandran plot (%)c 80.6/17.6/1.8/0 79.7/18.3/1.9/0 78.9/19.0/2.1/0 a Values in parentheses correspond to the highest-resolution data shell. b Friedel pairs are treated as different reflections. c Calculated using PROCHECK (core/allowed/generously allowed/disallowed regions). A unique juxtaposition of domains for IGF-II binding Domains 11–13 pack into a compact arrangement, which concurs with this section of IGF2R forming the high-affinity IGF-II-binding site (Figure 1; Supplementary Figure S2). The most distinctive feature, the FNII insert, projects out from the FG-loop of domain 13 to nestle beneath the AB-loop of domain 11, making multiple contacts with both domains 11 and 12. Despite its importance in high-affinity IGF-II binding, FNII does not directly contact IGF-II in the complex. Approximately 4100 Å2 of surface area is buried at interfaces in the IGF2R-Dom11–14 structure. The largest interface buries ∼1330 Å2 between domains 11 and 12, where the domain 12 BC-loop packs against the domain 11 βE–βI sheet. Many residues previously identified as forming a hydrophobic patch on the isolated IGF2R-Dom11 surface are found in this interface (Brown et al, 2002). A further ∼730 Å2 are buried where the domain 11 AB and GH-loops interface with FNII. The domain 12–13 interface is slightly smaller than that between 11 and 12, burying ∼1160 Å2 where FNII packs against the domain 12 NA″A and BC-loops. The smaller interface of ∼890 Å2 between domains 13 and 14 involves mainly the domain 13 NA′–NA″ hairpin loop and the DE-loop and βE of domain 14. Structural superposition shows the relative orientations of domains 11 and 12 in the three- and four-domain fragments to be conserved from the IGF2R-Dom11–12 structure, implying that this is a comparatively rigid two-domain unit. Furthermore, an overlay of IGF2R-Dom11–13 from the complex with IGF2R-Dom11–14 gives a close match, indicating that this rigidity extends to the tri-domain 11–13 unit. The relative orientations of domains 11–14 as multi-domain assemblies do not resemble the tri-domain motif reported for domains 1–3 (Olson et al, 2004), and there is little commonality of surface use in assembly, despite conservation of the core fold (Figure 2). Figure 2.Interface regions in IGF2R. 'Front' and 'back' surface views are shown of each IGF2R domain so far crystallized as part of a multi-domain fragment. Cartoon representations are included for orientation purposes. Coloured patches highlight regions involved in interfaces with other domains, with corresponding patches coloured the same. Residues involved in interfaces are defined as those undergoing a change in ASA. Red patches represent residues with ASA buried in more than one interface. Download figure Download PowerPoint Molecular details of the IGF-II/IGF2R interaction The complex structure described in this paper provides a good 'coarse-grained' model for analysing the interaction between IGF-II and IGF2R; however, the resolution of 4.1 Å is insufficient for detailed interpretation of side-chain placement and explicit bonding information should not be inferred. IGF-II bound to IGF2R buries ∼750 Å2 of solvent-accessible surface area (ASA) on IGF-II and ∼710 Å2 on IGF2R (Figure 3; Supplementary Figure S3). Analysis of the interface indicates that IGF2R residues burying ASA at the interface are almost all contained in the domain 11 AB-, CD-, FG- and HI-loops (Figure 3A; Supplementary Figure S3). At the core, a hydrophobic cluster (Figure 3B) comprising Tyr1542, Phe1567 and Leu1629 of IGF2R surrounds Phe19 of IGF-II. Only domain 11 directly contacts IGF-II; minor ASA changes for Trp1939 and Phe1941 in FNII reflect their proximity to IGF-II side chains. At the binding site, IGF-II has an overall negatively charged surface and IGF2R an overall positively charged surface (Figure 3C). The involvement of Tyr1542, Glu1544, Phe1567, Thr1570 and Ile1572 is consistent with mutational studies, which implicate these side chains in IGF-II binding (Zaccheo et al, 2006; Figure 3D). Figure 3.Views of the IGF2R-Dom11–13/IGF-II complex and summaries of prior mutagenesis studies. (A) An open-book style view in cartoon representation. Side chains undergoing ASA change upon complexation are shown as sticks and key side chains are labelled. (B) Hydrophobic patches (green) on the surface, defined using the programme GRID (Goodford, 1985) and shown here as volumes of pseudo-energies contoured at −2.3 kcal/mole. (C) The electrostatic potential surface, produced using the APBS Tools plug-in for Pymol (Baker et al, 2001) and contoured ±10 kT (blue denotes positive and red negative potential). (D) IGF2R side chains in the proposed IGF-II-binding region, which have been mutagenized (Zaccheo et al, 2006). IGF2R is shown in cartoon representation and IGF-II as surface representation. (E) IGF-II side chains mutagenized before this study (note Ala54 is not visible). IGF-II is shown in cartoon representation and IGF2R as surface representation. Download figure Download PowerPoint For IGF-II, Phe19 undergoes the largest change in ASA (Supplementary Figure S3) and is the anchor residue of the interaction. This concurs with previous mutagenesis studies of IGF-II, which demonstrated that Thr16, Phe19, Asp52 and Leu53 are critical for IGF2R binding (Delaine et al, 2007). Of these, Thr16 has been highlighted as central to the markedly different binding affinities of the IGFs for IGF2R. Our structure shows that Thr16 is buried in the binding interface where mutation to Ala as in IGF-I most likely causes significant loss of interface interactions. Of the other interface side chains differing between IGF-II and IGF-I for which mutation data are available, Leu55Arg causes a 3.3-fold drop in affinity while Ala54Arg causes a fourfold drop (Forbes et al, 2001). The complex structure shows that Ala54 does indeed lie within the interface, whereas Leu55 does not appear to be involved. Both Ala54 and Leu55 are close to the crucial Leu53 side chain (Delaine et al, 2007) and so mutation introducing large charged side chains could induce local conformational changes responsible for the observed reduction in IGF2R binding affinity. Our crystal structure indicates that prior attempts to predict this complex in silico did not provide an accurate model (Roche et al, 2006). The direct crystallographic insights into the interface architecture detailed above provided fresh impetus for us to evaluate the IGF-II/IGF2R interaction by using mutagenesis to probe further IGF-II side-chain involvement and to analyse the contribution of FNII. IGF2R binding of novel IGF-II mutations Our complex structure showed that the residues Phe48, Arg49 and Ser50 are not close to the binding interface (Figure 3E). This observation contradicts previous reports suggesting that these residues are critical for IGF2R binding (Burgisser et al, 1991; Sakano et al, 1991). We therefore decided to make single alanine mutations of these residues and to remake the Phe48Thr, Arg49Ser, Ser50Ile IGF-II reported in the previous studies, and performed surface plasmon resonance using our IGF2R multi-domain fragments (IGF2R-Dom10–13, IGF2R-Dom11–13 and IGF2R-Dom11–14). All mutants retained the native IGF-II fold as measured by circular dichroism and were active in IGF-1R binding assays (data not shown), except Tyr27Leu IGF-II, which has previously been characterised (Sakano et al, 1991). Sensorgrams and affinities are presented in Figure 4 and Table II, respectively. Figure 4.Novel IGF-II mutations. (A) A view of the IGF-II/IGF2R-Dom11–13 complex with side chains investigated by mutagenesis shown as sticks and coloured according to the legend. (B) BIAcore analyses of IGF-II and analogues (representative curves each at 50 nM) binding to multi-domain IGF2R fragments. Curves are coloured as in the legend to panel A. (C) Summary of all IGF-II mutagenesis studies, which used IGF2R fragments. IGF-II side chains are coloured according to their effect on IGF2R binding (see legend). Download figure Download PowerPoint Table 2. BIAcore kinetic analysis of ligand binding by IGF2R fragments ka1 (1/Ms) ( × 105) kd1 (1/s) ( × 10−2) ka2 (1/Ms) ( × 10−3) kd2 (1/s) ( × 10−3) KA (1/M) ( × 108) Relative KA χ2 (A) IGF-II analogues IGF2R-Dom10–13 IGF-II 9.02 1.28 3.43 1.44 2.49 1.00 2.1 Phe48Ala 6.54 3.22 2.19 0.66 0.93 0.37 0.9 Arg49Ala 9.55 2.08 1.62 1.76 0.91 0.37 1.2 Ser50Ala 11.25 1.71 3.10 1.77 1.84 0.74 3.4 Phe48Thr, Arg49Ser, Ser50Ile 4.66 6.02 3.61 0.52 0.58 0.23 0.6 Glu57Ala 9.38 1.56 3.31 1.76 1.90 0.76 2.7 Thr58Met 11.13 2.29 2.35 1.52 1.27 0.51 2.7 Leu8Ala 2.87 6.25 6.05 0.63 0.56 0.22 0.5 Glu6Arg 9.16 1.20 4.47 0.76 5.35 2.15 2.5 Tyr27Leu 5.81 1.98 5.13 0.53 3.97 1.59 1.7 IGF2R-Dom11–13 IGF-II 6.85 1.28 3.85 1.48 2.01 1.00 8.1 Phe48Ala 5.50 2.58 1.90 0.92 0.72 0.36 2.1 Arg49Ala 6.48 1.88 2.08 2.34 0.66 0.33 6.1 Ser50Ala 9.43 1.67 3.61 1.82 1.77 0.88 11 Phe48Thr, Arg49Ser, Ser50Ile 3.84 5.30 3.20 0.58 0.53 0.26 1.2 Glu57Ala 7.42 1.55 3.84 1.80 1.82 0.90 1 Thr58Met 8.90 1.93 2.57 1.83 1.18 0.59 4.2 Leu8Ala 2.36 7.01 5.75 0.69 0.31 0.15 9.7 Glu6Arg 7.55 1.09 4.56 0.92 4.28 2.13 13 Tyr27Leu 5.24 1.66 4.71 0.67 2.92 1.45 2.6 IGF2R-Dom11–14 IGF-II 6.92 1.40 3.81 1.45 1.82 1.00 1.7 Phe48Ala 3.51 2.82 2.22 0.87 0.51 0.28 1.1 Arg49Ala 6.33 2.06 2.03 1.95 0.64 0.35 1.1 Ser50Ala 8.28 1.57 3.19 1.55 1.61 0.89 2 Phe48Thr, Arg49Ser, Ser50Ile 1.14 4.88 3.19 0.74 0.12 0.07 0.7 Glu57Ala 7.72 1.63 3.71 1.73 1.65 0.90 1.8 Thr58Met 9.66 2.28 2.71 1.53 1.18 0.65 1.5 Leu8Ala 0.79 4.30 6.49 0.66 0.22 0.12 6.3 Glu6Arg 6.98 1.26 5.04 0.75 4.30 2.36 4.4 Tyr27Leu 4.90 1.75 5.48 0.57 3.48 1.91 1.3 (B) IGF2R-Dom10–13 mutants Wild type 19.0 0.04 0.05 0.02 142 1.00 2.1 ΔFNII 15.3 4.48 6.99 0.17 14.3 0.10 0.5 Glu1544Lys 16.2 0.31 2.46 1.36 14.9 0.10 1.2 Glu1544Lys+ΔFNII 23.4 0.63 2.34 0.17 56.1 0.40 0.5 FNII, fibronectin type II insert; IGF-II, insulin-like growth factor II; IGF2R, type II IGF receptor. Association rate constants (ka), dissociation rate constants (kd) and association equilibrium constants (KA=ka/kd) are given. Relative association equilibrium constants form the basis of discussion in the text and are highlighted in bold (either KA IGF-II analogue/KA IGF-II or KA IGF2R mutant/KA IGF2R wild type). χ2 values are an average of values for all experiments with each analogue/mutant. (A) IGF-II analogues binding to multi-domain IGF2R fragments and (B) IGF2R mutants binding to IGF-II. BIAcore curves were fitted using a two-state conformational change model as outlined in the Materials and methods. Our Phe48Ala and Arg49Ala mutants display a 2.7- to 3.6-fold reduction in affinity for the multi-domain IGF2R fragments compared with wild-type IGF-II, and the Ser50Ala mutant no more than a 1.4-fold reduction in binding affinity. Our triple mutant where the equivalent insulin residues are substituted into IGF-II had a more significant effect, with a 4.3-fold (IGF2R-Dom10–13), 3.8-fold (IGF2R-Dom11–13) and 14.3-fold (IGF2R-Dom11–14) reduction in binding affinity. However, this is a much smaller effect than previously reported for the Phe48Thr, Arg49Ser, Ser50Ile IGF-II mutant (Sakano et al, 1991; Bach et al, 1993), and our new mutagenesis and structural data demonstrate that these side chains are not crucial determinants of IGF2R binding affinity, but perhaps indirectly influence IGF2R binding. Further mutants predicted to lie within the binding interface were also tested for IGF2R binding. Mutation of Leu8 to Ala has a significant effect, resulting in up to an 8.3-fold reduction in IGF2R binding affinity. Our complex structure reveals that Leu8 is buried in the IGF-II:IGF2R interface in a predominantly hydrophobic environment. Mutation to Ala would lead to a smaller hydrophobic contact area, thus providing an explanation for the observed reduction in binding affinity. Neither of Glu57Ala IGF-II, Thr58Met IGF-II or Tyr27Leu IGF-II mutants differed greatly from IGF-II in their IGF2R binding affinities. Interestingly, Glu57 is partially buried in the IGF-II/IGF2R complex and there is a significant difference in ASA upon binding (Figure 3; Supplementary Figure S3), but evidently substitution with Ala can be tolerated. Thr58 and Tyr27 are not part of the interface as seen in the complex structure (Figures 3 and 4), and this observation correlates nicely with the mutagenesis data. Tyr27 is crucial for IGF-1R binding and is not situated near the recently described IGF2R-binding surface (Delaine et al, 2007). Glu6Arg IGF-II shows an increase in IGF2R binding affinity of over twofold compared with IGF-II. The aliphatic portion of the Glu6 side chain is partially buried in the complex close to the FG-loop of domain 11 and mutation to Arg could conserve this interaction, while introducing new hydrogen bonds via its side-chain nitrogen atoms. Taken together with previous IGF-II mutation studies using the same IGF2R fragments (Delaine et al, 2007), these results give an overall view of the impact of IGF-II mutations on IGF2R binding (Figure 4C). While mutations at the interface drastically affect IGF2R binding, mutating more remote side chains can still influence IGF2R binding affinity, although more subtly than previously reported. The IGFs are relatively small polypeptide hormones and their core structure is perhaps more susceptible to changes in surface-exposed side-chain arrangements than larger proteins. Such changes could be transmitted to other regions of the protein. Although discussion of indirect mechanisms is speculative, it is obvious that mutations of surface-exposed IGF-II residues not directly involved in IGF-II binding can influence binding affinity. Role of FNII in high-affinity IGF-II binding Our structure shows that FNII does not directly contact
Referência(s)