Structure-function studies
The group has been in the forefront of efforts to devise new analytical techniques for studying the molecular structure of HS [1, 2] and GAG interactions with proteins [3, 4]. The picture that has emerged for HS is of a semi-ordered polymeric structure in which the sulphated sugar residues are arranged in clusters, called composite sulphated regions CSR), which are fairly regularly spaced along the polymer chain (Figure 1).
Figure 1: Schematic model of HS proteoglycan and the composite structure of the HS chain
A transmembrane HS proteoglycan (eg a syndecan) is represented with three HS chains.
The latter are comprised of a series of distinctive CSRs embedded within N-acetylated sequences ([GlcNAc – GlcUA]n).
A single expanded CSR is shown with its core S-domain ([GlcNS(±6S) – IdoUA(2S)]n) flanked by lesser-sulphated transition zones in which [GlcNAc(±6S) – GlcUA – GlcNS(±6S) – IdoUA] sequences occur.
Each CSR contains a highly sulphated domain (S-domain), flanked by lower sulphated sequences (transition zones). CSRs are the principal sites for growth factor interactions, and for promoting the assembly of signaling complexes together with high-affinity receptors. Identifying the structures of these active sites is one of the group's principal aims. It is clear that there may be several different mechanisms involved in HS-mediated cell signaling. In some cases the S-domain may induce a conformational change in the growth factor, whereas in others, for example FGF1, S-domains of appropriate length and sulphation bind the growth factor in a cooperative trans-dimeric orientation that efficiently recruits two FGF-receptors into a bioactive complex [4] (Figure 2A).
Figure 2: Models of HS-growth factor interactions(A) Cooperative mechanism of FGF1 dimerisation by HS
HS is an absolute requirement for fibroblast growth factor (FGF) signaling as it forms an active ternary complex with two FGFs and two tyrosine kinase receptors (FGFRs). We identified a potential cooperative mechanism of HS-induced FGF1 dimerisation that might facilitate the subsequent dimerisation and activation of the FGFRs. This requires specific sequences within HS, as well as a putative conformational change in the saccharide induced by initial FGF1 monomer binding (4).
(B) VEGF165 binding to HS
The most abundant and bioactive isoform of vascular endothelial growth factor (VEGF) is VEGF165. This is secreted as covalently-linked, anti-parallel homodimers and thus contains two identical C-terminal HS-binding domains. The dimer-binding domain within HS is thus proposed to comprise two separate monomer-binding S-domains, containing critical 6-O-sulphate groups, which need to be sufficiently spaced within a CSR (5).
(C) HGF/SF binding to HS
Hepatocyte growth factor / scatter factor (HGF/SF) is a modular protein comprising an N-terminal domain (N), a sequence of four Kringles (K1-K4) and a C-terminal serine protease homology domain (SP). The primary GAG-binding site, which accommodates a HS or DS tetra-/hexa-saccharide sequence, is located in the N-domain (3), though a weaker co-operative interaction with K1 may also occur.
For the naturally dimeric VEGF165, where HS interactions also enhance receptor binding/activation, the two constituent monomers bind in a cis configuration within an extended CSR [5] (Figure 2B). In the case of HGF/SF, another HS-dependent factor [6], there is further complexity in that dermatan sulphate (DS), another GAG that is structurally distinct from HS, is also able to bind and act as a co-receptor, indicating an interesting specificity of molecular recognition [3] (Figure 2C).
Elucidating the size and sequence specificities for oligosaccharide bioactivities in these different protein systems will hopefully provide a platform from which suitable drugs can be designed with potential activities as growth factor antagonists in cancer therapy.