Glycosaminoglycan-Binding Proteins Lecture 25, Chapter 29 May 11, 2004 Jeff Esko

Types of Glycan-Binding Proteins Glycosyltransferases and modifying enzymes Antibodies induced by carbohydrate antigens Animal Lectins: P,C,S,R,L and I-type (Lectures 19-23) Plant Lectins: Con A, PHA, WGA, Ricin, and many others (Lecture 24) Glycosaminoglycan-binding proteins (Lecture 25) Bacterial adhesins and Viral hemagglutinins (Lecture 26) Lectin - term usually restricted to proteins that share primary sequence homologies

Glycosaminoglycans (GAG) GlcNAc GlcA b3 b4 Hyaluronan GalNAc GlcA b3 b4 4S 2S IdoA Chondroitin/Dermatan Sulfate a4 b4 6S NS NS3S 2S GlcNAc GlcA IdoA Heparin/Heparan Sulfate

GAG Binding Proteins Hyaluronan-binding proteins have a binding motif called the Link module Chondroitin sulfate binds to many proteins, but with low affinity, no apparent fold Hundreds of heparin binding proteins exist and do not generally sort into families of genes related through a common fold Dermatan sulfate binds to many of the same proteins as heparin

Hyaluronan (HA) b 4 3 b4 b3 n≥1000 GlcNAc GlcA Synthesized at plasma membrane, extruded from cell Abundant in skeletal tissues, synovial fluid, skin, elevated in expanding tissues (morphogenesis, invasion) Interesting biophysical properties (hydration, viscous solutions, resiliency) Present as capsule in some bacteria

Hyaluronan (HA) Bent, helical, relatively stiff structures Relatively stiff structure can assume unusual shapes. Not clear how this would affect protein interactions Bent, helical structures Fragments are potent signaling molecules Day & Sheehan (2001) COSB 1:1617

Aggrecan and CD44: Hyaluronan Binding Proteins Aggrecan forms link-protein stabilized complexes with HA, load bearing function

Hyaluronan-Binding Proteins (HABPs) Aggrecan Versican Neurocan Link Protein Brevican TSG-6 CD44 LYVE-1 = Link Module Link Module shared in common with Hyaladherens Aggrecan members (cartilage) involved in tissue integrity Versican in connective tissues (Fibroblasts) Neurocan and Brevican in brain matrix TSG-6 Tumor necrosis factor induced gene 6, inflammatory protein of unknown function CD44 Cell attachment, turnover of HA, migration of leukocytes, multiple isoforms based on splicing and glycosylation Layilin, a recently characterized transmembrane protein with homology to C-type lectins, is a membrane-binding site for talin in peripheral ruffles of spreading cells. Similar to CD44 in that regard LYVE-1 Lymph vessel endothelium, HA receptor for degradation SPACR, SPACRCAN, CDC37, CD38, Layilin, Bikunin RHAMM (intracellular) May be intracellular O-GlcNAc binding proteins Aggrecan Family Tissue architecture, stability Link protein Stabilizes aggrecan-HA aggregates CD44 Cell adhesion TSG-6 Inflammation LYVE-1 Clearance

Members deduced by sequence homologies SS refers to structural motifs, a-helices and b-sheets SS: b = b-sheet a = a-helix Members deduced by sequence homologies Note position of four conserved Cys residues, plus other amino acids in consensus sequence

The two a-helices and two triple-stranded anti-parallel b-sheets make up the Link Module 3-D structure of TSG-1 determined by NMR Consensus fold Only identified in vertebrates. Consistent with the fact that HA is relatively recent invention, Drosophila does not make it. Similar orientation to C-type lectins, but no Ca2+ b6 a2 b1 Day and Prestwich (2001) JBC 277:4585

Notice positively charged residues and aromatics Binding residues determined by mutagenesis Binding site is actually generated by folding of different segments of the chain, bringing key residues into proximity Notice positively charged residues and aromatics www.glycoforum.gr.jp/science/hyaluronan/HA16/HA16E.html

Heparin Binding Proteins NS3S 2S a4 b4 NS 2S a4 6S NS 2S b4 a3 4S 2S

Conformational Considerations GAG chains assume helical configurations, which causes charged residues to alternate across the helix a4 6S NS 2S = GlcNAc = GlcA NS 2S NS and 2S groups are on the same side COO- locations depend on whether its GlcA or IdoA CO2- CO2- 6S Zipper effect Also means that proteins could bind to both sides of chain NS 2S

Sugar Conformation Most sugars prefer the 4C1 conformation IdoA which is formed by epimerization of GlcA has the 1C4 or 1S0 conformation The greater conformational flexibility means that the sulfate and carboxylates can shift position more readily Greater binding possibilities and induced fit

Do Consensus Sequences Exist? Generally, GAG binding proteins contain clustered Lysine and Arginine residues In 1989, Cardin and Weintraub proposed a consensus sequence for heparin binding proteins, B = basic residue -XBBXBX- -XBBBXXBX- Spacing would place basic residues on the same face of an a-helix (3.4 residues/turn) or a b-strand (alternating faces) The consensus is predictive, not diagnostic. As many proteins don’t have this motif as do It turns out that most binding proteins do not fit this pattern and binding site is composed of positive residues contributed by different segments of the protein

Antithrombin Antithrombin, a serpin (serine protease inhibitor) Inactivates proteases involved in coagulation (Factors IIa and Xa) Blocks coagulation Antithrombin deficiency results in thrombosis (clot formation) Heparin binds to antithrombin, alters its conformation, and enhances rate of inactivation of Xa and IIa by a factor of 104 Only need a heparin pentasaccharide to activate O S O - - C O O - ±OSO3- O S O 3 3 O O O O O C O O - O O H O H O O S O - O O H 3 O H O O O NAc O H N H S O - O S O - N H S O - 3 3 3

Antithrombin-Heparin Binding occurs by a two step mechanism, initial weak interaction followed by a conformational change, then tight binding. KD ~ 2.5 x 10-10 M DG ~ 13.3 kcal/mol

Heparin-Antithrombin D A Binding site for heparin is in a cleft formed by two helices Binding is oriented, with pentasaccharide in cleft and flanking chain to the non-reducing side extending up and over the protein An 18-mer is actually needed to inactivate thrombin, so it acts like a template to approximate antithrombin-thrombin Interaction with thrombin does not require specific oligosaccharide sequence (low affinity)

D 3.6 kcal D E 2.1 kcal E 1.8 kcal F G 6.9 kcal H P Jin et al. (1997) PNAS 94:14683 D 3.6 kcal D E 2.1 kcal E 1.8 kcal F Explain helix lettering and origin of P helix\ Explain numbering of sugar residues Stars represent human mutations that affect affinity and thrombosis Lys 114 effect is mostly due to ionic effects since adjusting salt concentration has a large effect (1000x). G 6.9 kcal H P

Contribution of Individual Groups to Affinity Blue numbers refer to kcal binding deduced by altering the glycan groups Red numbers refer to kcal binding deduced by mutating amino acids 3.7 0.4 5.1 1.7 2.8 1.8 Pentasaccharide analog Notice the GlcNS at D does not contribute directly GlcA is adjacent, which is unusual, especially when bracketed by two N-sulfated units Carboxyl group not important, but it creates target site for 3OST1 IdoA is critical. Conformational effect on chain DEF are thought to bind first causing the conformational change Residues to the N-terminal side play a role in binding larger oligosaccharides. 3.6 2.1 6.9 Atha et al. (1985) Biochemistry 24:6723

a4 b3 a4 b4 6S NS NS3S 2S a4 b4 NS 2S a4 6S NS 2S

Heparin versus Heparan Sulfate The difference between heparin and heparan sulfate is quantitative not qualitative

Heparan Sulfate Proteoglycans: Co-receptors and Signaling Molecules Wnts TGF-/BMPs HGF HB-EGF Hedgehog FGF VEGF Angiopoietin Heparan sulfate FGF FGF Signaling Event Mitogenesis

FGF-Heparin Hexasaccharide Crystal structure shows surface binding 119KRTGQYKLGSKTGPGQK135

FGF/FGF Receptor Co-crystals Plotnikov et al. Cell 98:641 (1999)

FGF2/FGFR1 Symmetric structure Heparin Symmetric structure Heparin interacts with both ligands and receptors Two heparin oligosaccharides present in crystal FGF FGF Mulloy & Linhardt (2001) COSB 11:623

Potential Docking Site for Heparin Top View Top View with basic residues shaded blue Therefore GAG chain acts as a template or scaffold on which ligand and receptor anneal Catalytic Side View

FGF-2 Activation Sequence If symmetric dimer structure is correct: a4 b4 NS 2S 6S a4 FGF-2 Binding Domain Receptor Binding Domain FGF-2 Binding Domain = GlcA = IdoA = GlcNAc

Expression of “Active” Heparan Sulfates - FGF + FR1 Heparinase FGF-2 plus AP-tagged receptor FGF alone Locate all HS by antibody staining K= keratinocytes, BM = basement membrane, V = blood vessel, FR1-AP = alkaline phosphatase fusion to FGF receptor-1, 3G10 = monoclonal antibody to heparinase treated HS Chang et al. FASEB J. 14:137 (2000)

FGF2/FGFR1 FGF1/FGFR2 Its never simple! FGF FGF FGF FGF Two chains found in first model One chain binds asymmetrically in second model FGF FGF Its never simple! Mulloy & Linhardt (2001) COSB 11:623

FMDV Depression that defines binding site for heparin is made up of segments from all three major capsid proteins Foot and Mouth disease virus Fry et al. (1999) Embo J 18:543

a4 b3 a4 b4 6S NS NS3S 2S a4 b4 NS 2S a4 6S NS 2S b4 a3 4S 2S