Computational Study of Glycan Conformation in Glycoprotein

Glycosylation is an important post-translational modification of proteins. Considerable efforts have been made to understand how glycosylation affects the structure, dynamics and function of proteins, but it remains an enigma due to the diversity and variability in the glycosylation. In this talk, I’ll discuss about glycan conformations in RCSB structure database. Using algorithms to extract glycan conformations in the structure database, I’ll demonstrate how proteins might affect the glycan conformations when glycans are covalently attached to a protein via N-glycosylation.

Transcript

1. Schrödinger Sunhwan Jo 4/14/2016 Computational Study of Glycan Conformation in Glycoprotein
2. My Ph.D. In One Slide: CHARMM-GUI: A Web-Based Graphical User Interface for CHARMM • Jo, Kim, Iyer, and Im. (2008) CHARMM-GUI: a web-based graphical user interface for CHARMM., J. Comput. Chem. • Jo, Kim, and Im. (2007) Automated builder and database of protein/membrane complexes for molecular dynamics simulations., PLoS ONE • Jo, Vargyas, Vasko-Szedlar, Roux, and Im. (2008) PBEQ-Solver for online visualization of electrostatic potential of biomolecules., Nucleic Acids Res. • Jo, Lim, Klauda, and Im. (2009) CHARMM-GUI Membrane Builder for mixed bilayers and its application to yeast membranes., Biophys. J. • Lee, Jo, Rui, Egwolf, Roux, Pastor, and Im. (2012) Web interface for brownian dynamics simulation of ion transport and its applications to beta-barrel pores., J. Comput. Chem. • Jo, Jiang, Lee, Roux, and Im. (2013) CHARMM-GUI Ligand Binder for Absolute Binding Free Energy Calculations and Its Application., J. Chem. Inf. Model. • Cheng, Jo, Lee, Klauda, and Im. (2013) CHARMM-GUI Micelle Builder for Pure/Mixed Micelle and Protein/Micelle Complex Systems., J. Chem. Inf. Model. SSNMR Ensemble Dynamics • Jo and Im. (2011) Transmembrane helix orientation and dynamics: insights from ensemble dynamics with solid-state NMR observables., Biophys. J. • Kim, Jo, and Im (2011) Solid-State NMR Ensemble Dynamics as a Mediator between Experiment and Simulation., Biophys. J. • Im, Jo, and Kim. (2012) An ensemble dynamics approach to decipher solid-state NMR observables of membrane proteins., Biochim. Biophys. Acta Glycan Structure Modeling • Jo, Song, Desaire, Mackerell, and Im. (2011) Glycan Reader: automated sugar identification and simulation preparation for carbohydrates and glycoproteins., J. Comput. Chem. • Jo and Im. (2012) Glycan fragment database: a database of PDB-based glycan 3D structures., Nucleic Acids Res. • Jo, Lee, Skolnick, and Im. (2013) Restricted N-glycan Conformational Space in the PDB and Its Implication in Glycan Structure Modeling. PLoS Comp. Biol. • Wu, Engström, Jo, Stuhlsatz, Yeom, Klauda, Widmalm, and Im. (2013) Molecular Dynamics and NMR Spectroscopy Studies of E. coli Lipopolysaccharide Structure and Dynamics. Biophys. J. Type III Secrection System • Zhong, Lefebre, Kaur, McDowell, Gdowski, Jo, Wang, Benedict, Lea, Galan, and De Guzman. (2012) The Salmonella Type III Secretion System Inner Rod Protein PrgJ Is Partially Folded., J. Biol. Chem. Cholesterol Orientation in Lipid Bilayer • Jo, Rui, Lim, Klauda, and Im. (2010) Cholesterol flip-flop: insights from free energy simulation studies., J. Phys. Chem. B
3. Life after Ph.D. Protein-Protein Binding Free Energy and Entropy • Jo, Chipot, and Roux (2015) Efficient determination of relative entropy using combined temperature and Hamiltonian replica-exchange molecular dynamics., J. Chem. Theo. Compute. • Jo and Jiang (2015) A generic implementation of REST2 algorithm in NAMD for complex biophysical simulations, Compute. Phys. Comm. • Jo, Suh, He, Chipot, and Roux (2016) Leveraging the Information from Markov State Models to Improve the Convergence of Umbrella Sampling Simulations (in preparation) • Jo, Jiang, Chipot, and Roux (2016???) Quantifying Protein-Protein Binding Energy and Entropy using Molecular Dynamics Simulations (on going) ANMPathway • Das, Gur, Cheng, Jo, Bahar, and Roux. (2014)., Exploring the Conformational Transitions of Biomolecular Systems Using a Simple Two-State Anisotropic Network Model., PloS Comp. Biol.
4. Glycan Refers Carbohydrates O OH HO HO OH OH O OH HO HO NHAc OH O HO HO HO HO OH O OH HO HO OH OH O OH HO HO NHAc OH O HO HO OH OH O HO HO OH OH OH O O OH HO OH OH O AcHN OH OH O OH HO OH OH ᴅ-Glucose (Glc) N-acetyl-ᴅ- Glucosamine (GlcNAc) ᴅ-Mannose (Man) ᴅ-Galactose (Glc) N-acetyl-ᴅ- Galactosamine (GalNAc) ᴅ-Xylose (Xyl) ᴅ-Glucuronic acid (GlcA) ʟ-Fucose (Fuc) N-Acetylneuraminic acid (NeuAc) Common sugars (= monosaccharide) in vertebrates
5. Glycan Refers Carbohydrates O OH HO HO OH OH O OH HO HO NHAc OH O HO HO HO HO OH O OH HO HO OH OH O OH HO HO NHAc OH O HO HO OH OH O HO HO OH OH OH O O OH HO OH OH O AcHN OH OH O OH HO OH OH ᴅ-Glucose (Glc) N-acetyl-ᴅ- Glucosamine (GlcNAc) ᴅ-Mannose (Man) ᴅ-Galactose (Glc) N-acetyl-ᴅ- Galactosamine (GalNAc) ᴅ-Xylose (Xyl) ᴅ-Glucuronic acid (GlcA) ʟ-Fucose (Fuc) N-Acetylneuraminic acid (NeuAc) Common sugars (= monosaccharide) in vertebrates Disaccharide O OH OH HO O OH O OH HO HO OH 1 2 3 4 6 Glc (1➞4) Glc
6. Glycan Refers Carbohydrates O OH HO HO OH OH O OH HO HO NHAc OH O HO HO HO HO OH O OH HO HO OH OH O OH HO HO NHAc OH O HO HO OH OH O HO HO OH OH OH O O OH HO OH OH O AcHN OH OH O OH HO OH OH ᴅ-Glucose (Glc) N-acetyl-ᴅ- Glucosamine (GlcNAc) ᴅ-Mannose (Man) ᴅ-Galactose (Glc) N-acetyl-ᴅ- Galactosamine (GalNAc) ᴅ-Xylose (Xyl) ᴅ-Glucuronic acid (GlcA) ʟ-Fucose (Fuc) N-Acetylneuraminic acid (NeuAc) Common sugars (= monosaccharide) in vertebrates Disaccharide O OH OH HO O OH O OH HO HO OH Glc (1➞4) Glc β1→4 α1→6 β1→4 α1→3 O O HO O NH OH O O HO O HO O CH3 OH O HO NH OH O CH3 O HO HO HO OH O HO HO HO OH A B GlcNac GlcNac Mannose Mannose Mannose Oligosaccharide (1➞4) (1➞4) (1➞3) (1➞6) 1 2 3 4 6
7. Glycan Refers Carbohydrates O OH HO HO OH OH O OH HO HO NHAc OH O HO HO HO HO OH O OH HO HO OH OH O OH HO HO NHAc OH O HO HO OH OH O HO HO OH OH OH O O OH HO OH OH O AcHN OH OH O OH HO OH OH ᴅ-Glucose (Glc) N-acetyl-ᴅ- Glucosamine (GlcNAc) ᴅ-Mannose (Man) ᴅ-Galactose (Glc) N-acetyl-ᴅ- Galactosamine (GalNAc) ᴅ-Xylose (Xyl) ᴅ-Glucuronic acid (GlcA) ʟ-Fucose (Fuc) N-Acetylneuraminic acid (NeuAc) Common sugars (= monosaccharide) in vertebrates Disaccharide Glc (1➞4) Glc Oligosaccharide (1➞4) β1→4 α1→6 β1→4 α1→3 O O HO O NH OH O HO O HO O OH O HO NH OH O HO HO HO OH O HO HO HO OH A B
8. N-glycan Biosynthesis and N-glycosylation for Glycoprotein ER Lumen P Dolichol- phosphate P P P P P P P P P P P P P P P "Flip" P P P P P P P P P P P P P P P N-Acetylglucosamine Glucose Mannose β1→4 α1→6 α1→3 β1→4 α1→6 α1→3 α1→2 α1→2 α1→2 α1→2 α1→3 α1→3 Glc 3 Man 9 GlcNAc 2 α1→2 Cytosolic side • Different glycosyltransferases act on specific glycan moieties and monosaccharides are added onto a dolichol-phosphate lipid molecule. • Dolichol-phosphate initially face cytosolic side of the ER membrane, later flipped to face inside of ER lumen. N-glycan precursor
9. N-glycan Biosynthesis and N-glycosylation for Glycoprotein ER P P Ribosome Nascent peptide N OST Folded? N Protease degradation Medial-Golgi Trans-Golgi Cis-Golgi N Matured Complex Glycan N N N N Mammals Plants Invertebrates N-Acetylglucosamine Glucose Mannose Galactose Fucose N-Acetylgalactosamine Xylose Sialic acid Species specific core modification • Glycosylation is important for “quality control” of the proteins synthesized in ER and increase stability of the protein. • In the maturation process, species specific core modification is performed. • In the maturation process, complex glycan sequence is acquired by various deglycosylase and glycosyltransferase in Golgi. N Matured Complex Glycan N
10. N Matured Complex Glycan N
11. Why Do We Study Glycan? 1. Glycans can modulate protein structure, dynamics, and function. 2. Glycans are related with vaccine development. 3. Alteration of glycosylation pattern is related to aging and disease. Fucose GlcNac Man Gal Sialic acid Anti-inflamatory Pro-inflamatory Fc receptor binding site Different glycan sequence affects the binding affinity of IgG1 to Fc receptor Barb and Prestgard. Nat Chem Biol (2011) vol. 7 (3) pp.147-53
12. Why Do We Study Glycan? 1. Glycans can modulate protein structure, dynamics, and function. 2. Glycans are related with vaccine development. 3. Alteration of glycosylation pattern is related to aging and disease. HIV gp120 envelop proteins has lots of glycans (about 50% of the total mass) Trimeric spike Image adapted from: http://www.avert.org/aids-photo-gallery.php?photo_id=508&gallery_id=4 Burton et al. Proc Natl Acad Sci USA (2005) vol. 102 (42) pp. 14943-8
13. Why Do We Study Glycan? 1. Glycans can modulate protein structure, dynamics, and function. 2. Glycans are related with vaccine development. 3. Alteration of glycosylation pattern is related to aging and disease. Onset of disease or aging alters the glycosylation pattern Fan and Hendrickson. Nature (2005) vol. 433(7023) 269-277 Varki, Cummings, Esko, Cold Spring Harbor Press (2009) 2nd Ed. FSH Hormone has four glycosylation site. Different glycosylation pattern affects the activity of the hormone as well as generates different signal at the cellular level.
14. Motivation: Knowing primary sequence of a glycan using mass spectrometry and NMR is relatively easy Understanding its impact on the protein structure and dynamics is hard Glycans are involved in many biological processes, but poorly understood. • How different glycoforms affect the structure and dynamics of glycoproteins? • What glycoform can be targeted for vaccine development or assays?
15. Carbohydrate structures found in Cambridge Structural Database contains mostly monomeric carbohydrates and only small number of oligosaccharide structures. “No” According to recent NMR studies, oligosaccharides are likely have several well defined structure, and it is just hard to have electron diffraction when multiple conformations are favorable. “Yes” Moreover, we have many glycan structure present in the Protein Data Bank database, and there are cases where same glycan structure adopts similar structures. Glycan Structure Prediction/Modeling Does N-glycans Have Well-Defined Structure? Carbohydrate conformational freedom is independent of protein structure, thus it will be also very flexible on protein surface.
16. 3AVE Human Fc Fragment 3AY4 Fc Fragment 3C2S Human Fc Fragment 3D6G Fc Fragment 3DO3 Human Fc Fragment 2DTS Human Fc Fragment 3FJT Human Fc Fragment 1H3X Human Fc Fragment 1I1A Neonatal Fc Fragment 1I1C Rat Fc Fragment 1L6X Rituximab Fc Fragment 1OQO Fc Fragment 2QL1 Human Fc Fragment 2RGS Mouse Fc Fragment 3SGJ Fc Fragment 3SGK Fc Fragment 2VUO Rabbit Fc Fragment • Large number of PDB entries of IgG1 antibody has conserved glycosylation motif, and the structure of such glycan is very similar from one structure to another having average RMSD ~ 0.5 Å. Some N-glycans Does Have Well-Defined Structure RMSD ~ 0.5 Å
17. Detection of carbohydrate-like molecules using graph representation Automatic annotation of carbohydrates based on their 3D structures Automatic recognition of glycosidic linkages Glycan Reader: Automatic Recognition of Carbohydrates in PDB Jo, Song, Desaire, MacKerell, and Im (2011) J. Comput. Chem. 32:3135-3141 About 30% of PDB containing carbohydrate have at least one error.
18. Glycan Reader: Automatic Recognition of Carbohydrates in PDB
19. CHARMM PSF/CRD with complete glycosidic linkage patch PDB:1L6X; IgG1 constant domain PDB:1GZM; Bovine rhodopsin Molecular dynamics simulation setup of glycoprotein in lipid bilayer Electrostatic surface of glycosylated protein PDB:2A0Z; Toll-like receptor Glycan Reader: Automatic Recognition of Carbohydrates in PDB
20. Total Number of PDBs with Glycan 6% 94% PDBs with Glycan 77709 (Dec. 2011) 4952 Number of glycan chains Glycan Reader: Automatic Recognition of Carbohydrates in PDB Jo, Song, Desaire, MacKerell, and Im (2011) J. Comput. Chem. 32:3135-3141 56.5% N-glycosylated 4.3% O-glycosylated 39.2% Ligands Biologically relevant glycans have ~6-20 carbohydrate units. 5637 88,078 (Dec. 2012) PDBs with Glycan Number of PDBs with Glycan Length of Glycan Chain
21. Glycan Structure Prediction/Modeling Glycan Fragment Database Jo and Im (2011) Nucleic Acids Res. 2012 www.glycanstructure.org Disaccharide O OH OH HO O OH O OH HO HO OH Glc (1➞4) Glc
22. PDB Glycan Structure Survey Jo, Lee, Skolnick, and Im (2013) PLoS Comp Biol
23. PDB Glycan Structure Survey PDB:1TG7 RMSD <5Å RMSD Population PDB:1MWE RMSD <3Å PDB:2QC1 RMSD <1Å • What is the range of structural difference between glycans having same sequence?
24. PDB Glycan Structure Survey RMSD Population Homologous Non-homologous • Does protein structure affect N-glycan structure? If protein structure affects the glycan structure glycans on the homologous glycoprotein will have similar structures.
25. PDB Glycan Structure Survey p-value Cumulative Population Homologous Non-homologous RMSD Population Homologous Non-homologous Random Statistical test • Does protein structure affect N-glycan structure? If protein structure affects the glycan structure glycans on the homologous glycoprotein will have similar structures.
26. PDB Glycan Structure Survey • For each glycan sequence, 10,000,000 random conformers and 124,750 conformers were used to calculate the conformational diversity. Jo, Lee, Skolnick, and Im (2013) PLoS Comp Biol Homologous Glycan Non- homologous Glycan
27. Homologous Glycans Have Smaller Structural Variability Homologous Non-homologous Random • For each glycan sequence, 10,000,000 random conformers and 124,750 conformers were used to calculate the conformational diversity. Random Non-homologous Homologous Jo, Lee, Skolnick, and Im (2013) PLoS Comp Biol Homologous Glycan Non- homologous Glycan
28. Homologous Glycans Have Conserved Structure Homologous Non-homologous On average, ~75% of glycans on homologous proteins have similar structures. Jo, Lee, Skolnick, and Im (2013) PLoS Comp Biol p=0.05
29. Homologous Glycans Have Conserved Structure Homologous Non-homologous Jo, Lee, Skolnick, and Im (2013) PLoS Comp Biol p=0.05
30. Homologous Glycans Have Conserved Structure Homologous Non-homologous Jo, Lee, Skolnick, and Im (2013) PLoS Comp Biol
31. Summary: PDB Glycan Structure Survey • N-glycan structures on the surface of homologous glycoproteins are significantly conserved, suggesting that the protein structures does affect the N-glycan structures. • N-glycan orientations are diverse even in homologous glycoproteins. However, if we have the first few carbohydrate residues in the crystal structure can be used to increase the modeling accuracy. • Internal substructures of glycan structures in the PDB are well conserved even for glycans on the non-homologous proteins. • There could be a sampling bias; this observations are based on crystal structures of N-glycan and the structure of N-glycans having strong interaction with protein may have higher chance to solved. • Glycoprotein’s sequence similarity can be used to identify “good” template structures and such template significantly enhances the performance of N-glycan modeling.
32. Can I predict reasonable glycan structure models using homology modeling? Aglycoprotein structure Glycosylation site + Glycan sequence + (Naïve) Glycan Modeling Protocol:
33. Aglycoprotein structure Glycosylation site + Glycan sequence + Glycoprotein structure (Naïve) Glycan Modeling Protocol: Can I predict reasonable glycan structure models using homology modeling?
34. Cluster template Search fragment Cluster fragment Search template structure - Glycan sequence - Protein structure - Glycosylation site Discard disallowed conformers (bad contact) Refine / Rank Sequence from user: Template structure Fragment #2 Fragment #1 Assemble Glycan Structure Prediction/Modeling (Naïve) Glycan Modeling Protocol: • Performance of modeling protocol depends on the ability to: • Generate large number of native-like structures • Discriminate native-like structure vs. others
35. Cluster template Search fragment Cluster fragment Search template structure - Glycan sequence - Protein structure - Glycosylation site Discard disallowed conformers (bad contact) Refine / Rank Sequence from user: Template structure Fragment #2 Fragment #1 Assemble Glycan Structure Prediction/Modeling (Naïve) Glycan Modeling Protocol: Homology Modeling Approach Select template structures that are more native-like • Performance of modeling protocol depends on the ability to: • Generate large number of native-like structures • Discriminate native-like structure vs. others
36. Glycan Structure Prediction/Modeling • PDB:1L6X was used to perform glycan structure modeling using the homology modeling approach and the performance is compared by the RMSD with respect to the crystal structure. • ~60 % of the generate structure models have RMSD < 3Å with respect to the crystal structure. Glycan-only Glycoprotein (Naïve) Glycan Modeling Protocol: Homology Modeling Approach
37. Glycan Structure Prediction/Modeling • Orientation of glycan structure seems to be more difficult suggesting more sampling is required after modeling of glycan structures. Glycan-only Glycoprotein (Naïve) Glycan Modeling Protocol: Homology Modeling Approach
38. Glycan Structure Prediction/Modeling What is the Impact of Glycosylation to Glycan Structures? Free Glycosylation
39. Glycan Structure Prediction/Modeling What is the Impact of Glycosylation to Glycan Structures? Free Glycosylation PDB MD Simulation
40. • The N-glycan core pentasaccharide sequence can be found in virtually every N-glycan sequence. • The pentasaccharide sequence is small enough for extensive characterization using simulation, and at the same time large enough to see diverse conformational preferences. β1→4 α1→6 β1→4 α1→3 O O HO O NH OH O O HO O HO O CH3 OH O HO NH OH O CH3 O HO HO HO OH O HO HO HO OH A B (1) (2) (3) (4) (1) (2) (3) (4) Conformational Preference of N-glycan Core Pentasaccharide: N Matured Complex Glycan N N N N Mammals Plants Invertebrates
41. Torsion Angle Clusters Glycosidic Linkage Cluster Population 1 2 3 4 (1→4) (1→4) (1→3) (1→6) #1 (-92, 94) (-84, 104) (78, -112) (74, 168, 60) 28% #2 (-80, 126) (-77, 113) (75, -107) (65, 117, -60) 20% #3 (-77, 117) (-80, 130) (71, -133) (64, 91, 74) 12% #4 (-93, 108) (-83, 129) (74, -145) (102, 83, -50) 8% #5 (-78, 114) (-83, 108) (141, -73) (128, 157, -23) 4% Initial Conformations for Simulation: Overlay of average structures from the clusters 1-5: Ring heavy atoms of the first three residues are used for alignment β1→4 α1→6 β1→4 α1→3 O O HO O NH OH O O HO O HO O CH3 OH O HO NH OH O CH3 O HO HO HO OH O HO HO HO OH A B (1) (2) (3) (4) 1-6 Branch 1-3 Branch 1-6 Branch 1-6 Branch
42. Simulation Detail: MD System Size Nwater Simulation Time 1 44 x 44 x 44 2611 700 ns 2 45 x 45 x 45 2872 700 ns 3 44 x 44 x 44 2611 700 ns 4 45 x 45 x 45 2870 700 ns 5 45 x 45 x 45 2872 700 ns RE 44 x 44 x 44 2611 100 ns 12.5 Å • Five independent simulation system was built using initial conformation as the representative conformations from 5 largest clusters from GFDB. • System size is determined to have at least 12.5 Å of water layer at the system edge, and no ions are included. • Simulation systems are minimized for 100 steps of SD and ABNR method, and CPT dynamics was performed at 300 K using NAMD. Switching function was used for nonbonded interaction (12/10; cutoff/switching). • 8 cpu x 105 days x 5 replicates = 360,000,000 CPU hours! • 152 cpu x 25 days = 300,000,000 CPU hours! 44-45 Å
43. Conformational Fluctuation of Pentasaccharide: • Conformational distribution from the pentasaccharide in solution is more restricted than the conformers from high-temperature simulation and random conformation pool. • Distinctive peaks in the conformational distribution suggests that there are few preferential conformations exists. At the higher temperature, the conformational variability moves towards random conformation. Random Conformation Pool REXMD (450K) PDB (N=89) MD Simulation (Aggregated)
44. Conformational Fluctuation of Pentasaccharide: • Conformational distribution from the pentasaccharide in solution is more restricted than the conformers from high-temperature simulation and random conformation pool. • Distinctive peaks in the conformational distribution suggests that there are few preferential conformations exists. At the higher temperature, the conformational variability moves towards random conformation. • Although the number of PDB entries are not large, the glycans in the PDB showed similar conformational variability as random conformation pool. REXMD (450K) PDB (N=89) Random Conformation Pool MD Simulation (Aggregated)
45. Conformational State Designation by Glycosidic Torsion Angle α1→3 O O HO O NH OH O O HO O HO O CH3 OH O HO NH OH O CH3 O HO HO HO OH O HO HO HO OH B A B C D 1) 1) A B C D 2) 2) 3) 3) A B 4) 5) 4) 5) A B C G g t
46. Conformational State Designation by Glycosidic Torsion Angle α1→3 O O HO O NH OH O O HO O HO O CH3 OH O HO NH OH O CH3 O HO HO HO OH O HO HO HO OH B A B C D 1) 1) A B C D 2) 2) 3) 3) A B 4) 5) 4) 5) A B C G g t ABAAG 288 possible conformational states 42 states visited
47. Conformational States of N-glycan Core Pentasaccharide in Solution • 10 largest conformational states cover more than 99 % of conformations in solution at 300 K. • A single largest conformational states occupies > 75 % of all trajectories! • The free energy difference between the largest state (AAAAG) and the second largest state (AAAAg) is ~1.3 kcal/mol.
48. 2.7 3.5 4.3 5.0 AAAAg AAAAG 6.1 4.6 AAABg 4.8 5.1 4.2 5.9 5.6 AAABG 3.9 2.9 3.8 3.3 ABAAG 4.3 • The state fold-back state (8%) appears to have more interaction than the extended state (76%). Why the extended state is more favorable than fold-back state? • Extended conformation is entropically more favorable in general. • Fold-back state has not enough number of H-bond to compensate the entropic penalty and these H-bonds are rather weak. Conformational States of N-glycan Core Pentasaccharide in Solution “extended” “fold-back”
49. • Although the number of sample is not enough, compared to free glycan in solution, glycans in close contact with protein have significant population shift. • The ones closer to protein have less population in the smaller basin (“B”) but the terminal residues seems more flexible. Conformational States of N-glycan Core Pentasaccharide in Glycoprotein
50. AAAAG PDB:3PPS AAABG PDB:3GLY AAAAg PDB:2DTS • Glycans in the crystal structure have extensive contact with protein, which may explain the change of conformational preferences from solution. Conformational States of N-glycan Core Pentasaccharide in Glycoprotein 1-6 branch 1-6 branch 1-6 branch Increased protein-glycan interaction allows glycan to adopt conformational state unfavorable in solution
51. AAACg PDB:1B5F 1-6 branch • Glycans in the crystal structure have extensive contact with protein, which may explain the change of conformational preferences from solution. • N-glycans are found on the surface of the protein and may have contact with neighboring crystal units. Conformational States of N-glycan Core Pentasaccharide in Glycoprotein
52. AAACg PDB:1B5F 1-6 branch Crystal neighbor • Glycans in the crystal structure have extensive contact with protein, which may explain the change of conformational preferences from solution. • N-glycans are found on the surface of the protein and may have contact with neighboring crystal units. Conformational States of N-glycan Core Pentasaccharide in Glycoprotein
53. Glycan Structure Prediction/Modeling Summary: Conformational Preference of N-glycan Core Pentasaccharide • N-glycan pentasaccharide conformation in solution is rather restricted compared to glycosylated N-glycans. • The balance between the entropy and enthalpy appears to play a big role in determining the conformation of glycan. • At least in the pentasaccharide, the hydrogen bond are weak. This may not be the case for longer glycan, which can have more interaction to its GlcNAc region when folded-back to itself. • Increased protein-glycan interaction allows glycan to adopt conformational states that are unfavorable in solution. This, in turn, suggests the dynamics of glycan in glycoprotein could be slower than in solution.
54. Acknowledgement Im Lab Wonpil Im Huisun Lee Soohyung Park Yifei Qi Kevin Song (Chicago) Collaboration Heather Desaire (KU) Alex Mackerell (UMD) Jeff Skolnick (GATECH) Wei Jiang (ANL) Roux Lab Benoît Roux Yillin Meng Hui Li Huan Rui