A Glycan-Receptor Complex: Structural and Binding Analyses
( \newcommand{\kernel}{\mathrm{null}\,}\)
Literature-Based Guided Assessment (LGA) | A Glycan-Receptor Complex: Binding and X-ray structure analyses |
Instructors: Email hjakubowski@csbsju.edu for answers
Material derived from:
Harry M. Williams, Jesper B. Moeller, Ian Burns, Trevor J. Greenhough, Uffe Holmskov, and Annette K. Shrive. Crystal structures of human immune protein FIBCD1 suggest an extended binding site compatible with the recognition of pathogen-associated carbohydrate motifs. J. Biol Chem, 300, January 2024, https://doi.org/10.1016/j.jbc.2023.105552. Creative Commons Attribution (CC BY 4.0)
Introduction
Our immune system recognizes molecular structures from pathogens. One example is the outer lipopolysaccharide (LPS) from the bacterial cell wall. Since all bacteria have a cell wall, it would make sense that all organisms mounting an immune response to bacteria would recognize a common motif, or partial structure, from the cell wall. Such motifs are called pathogen-associated molecular patterns, or PAMPs. We've explored PAMPs before in Chapter 5.4. One interesting "cell wall" structure is chitin, which is a glycan of N-acetylglucosamine (GlcNAc) in β(1,4) link so it is very analogous to cellulose, a similar polymer of just glucose.
Chitin is found in the cell walls of fungi and walls of cysts in amoebae, so we encounter chitin in fungal and amoebae diseases. It is also found in egg shells of nematodes and in exoskeletons of invertebrates like mosquitoes, ticks, and snails, which are also vectors of disease.
Here is iCn3D model of monomeric cyclic GlcNAc (PDB file from PubChem). The N-acetyl group is attached to carbon 2 of the pyranose ring.
Here is a iCn3D model of anhydrous β-Chitin derived from X-ray crystallography at a 1 nm resolution. (PDB file from: https://polysac3db.cermav.cnrs.fr/db....php?number=73)
The fungal cell wall is a complicated mixture of chitin, β-glucans (glucose polymers), and proteins with extensive mannose glycans, as shown in Figure 1 below.
Figure 1: Fungal cell wall components. Vega K, Kalkum M. Chitin, chitinase responses, and invasive fungal infections. Int J Microbiol. 2012;2012:920459. doi: 10.1155/2012/920459. Epub 2011 Dec 11. PMID: 22187561; PMCID: PMC3236456.Creative Commons Attribution License.
The wall contains a cell membrane with various membrane proteins, a protective layer of chitin (yellow), glucans (mostly beta), and mannoproteins on its surface. Different fungal cell walls contain different glucans. For example, the cell wall of A. fumigatus contains β-1,3- and β-1,4-glucan, and α-1,3-glucan, while C. albicans contains β-1,3- and β-1,6-glucan.
Figure 2 below shows fungal PAMPs, host receptors that recognize them, and signaling pathways that are activated on binding of PAMPs to their receptors. Note the question mark by the putative chitin receptor. More about that below.
Figure AGlycan−ReceptorComplex.2 Fungal cell wall pathogen-associated molecular patterns (PAMPs) and their host-pattern recognition receptors (PRRs). Vega et al., ibid.
Various fungal cell wall components are recognized by specific PRRs. Some PAMPs are recognized by multiple PRRs; for example, N-linked mannan is recognized by mannan receptor (MR), dectin-2, and FcγR. Phospholipomannan-(PLM-) coated β-glucans are recognized by both TLR-6 and TLR-2. Other receptors may involve the signaling pathway of another PRR. For example, galectin-3, which recognizes β-mannosides, signals through TLR-2 (represented by a curved arrow), dectin-1, when activated by β-glucans can signal to activate the nuclear factor kappa-light chain enhancer of activated B cells (NF-κB) on its own or with the help of TLR-2. Fc gamma receptor (FcγR) may signal through dectin-2 when activated by N-linked mannan. Recognition of these fungal cell wall components mediates fungal recognition and defense by the host. Recognition by host PRRs usually involves signaling through NF-κB and activation of proinflammatory cytokines, such as TNF-α, or in some instances, and anti-inflammatory cytokines such as interleukin (IL)-10. The possibility of an alternative chitin receptor exists, activation of which leads to the recruitment of IL-4 producing cells. However, chitin has been shown to function as a T helper (Th)1 immune modulator, which stands in contrast an IL-4 associated Th2 response
Questions
A pre-modeled structure of beta chitin is shown below. H bonds are shown as dotted yellow lines. 3 H bonds are labeled with distances (1.9, 2.0, and 2.2 angstroms). The labeled H bonds are on the front of the structure
a. How many layers of chitin strands are shown?
b. Complete the following table with the properties of the H bonds. For donors indicate X-H and acceptors Y and include the number of X and Y in the structure (use ring number systems)
H bond | Intra or Inter-strand | H bond donor | H bond acceptor |
1.9 | Intrastrand | C3OH | ring (acetal) O |
2.0 | Interstrand | C2NH | C2-NHC(=O)CH3 |
2.2 | Interstrand | C6H2OH | C2-NHC(=O)CH3 |
- Answer
-
sdsds
Knowing a bit about protein structure and binding, to what might a host protein that recognizes a pathogen through its chitin glycan bind? Hint: compare it to the enzyme lysozyme which cleaves between NAG-NAM in cell walls.
a. A single GlcNAc within chitin or its degradation products
b. A (GlcNAc)2 within chitin or its degradation products
c. A (GlcNAc)3 within chitin or its degradation products
d. A larger section of the chitin polymer
- Answer
-
Protein Structural Analyses
The structure of the chitin receptor and its properties have been elusive. A leading candidate is the Fibrinogen C domain-containing protein 1 (FIBCD1), which is a single-pass (biotopic, as it crosses both leaflets of the membrane) Type II membrane protein (N terminal in the cytoplasm and close to the transmembrane domain in sequence number).
Figure 3 shows an interactive iCn3D model of the predicted AlphaFold structure of the human Fibrinogen C domain-containing protein 1 (FIBCD1) membrane protein (Q8N539). The alpha fold structure was aligned with the model for the Membranome database.
Figure 3: AlphaFold structure of the human Fibrinogen C domain-containing protein 1 membrane protein (Q8N539). (Copyright; author via source). Click the image for a popup or use this external link: https://structure.ncbi.nlm.nih.gov/i...cVEchayWfdjXT8
Uniprot shows that amino acids 1-33 (the N-terminal, gold coil backbone in the iCn3D model) are in the cytoplasm, not in the membrane as shown in the model above. Amino acids 34-54, colored cyan, comprise the single-pass transmembrane domain, while amino acids 55-461 (shown in cartoon form) represent the extracellular domain which interacts with chitin derivatives. The extracellular domain is abbreviated as FReD (fibrinogen-like recognition domain). Key amino acids involved in the binding of chitin derivatives are shown as CPK-colored sticks and labeled. On the iCn3D display, you can toggle the "a" key on your keyboard to move between the Membranome model and the iCn3D model.
You will note that the amino acids in the extracellular domain involved in binding chitin derivatives seem to be projecting into space so the question remains as to how this binds to chitin derivatives. The crystal structure of the extracellular domain has been determined. Figure 4 shows an interactive iCn3D model of the tetrameric fibrinogen-like recognition domain of FIBCD (6ZR4). The protein complex is a homotetramer or homo 4-mer.
Figure 4: Tetrameric fibrinogen-like recognition domain of FIBCD (6ZR4). (Copyright; author via source). Click the image for a popup or use this external link: https://structure.ncbi.nlm.nih.gov/i...DNVea1SMurWrj9
The amino acids involved in binding are shown in spacefill and labeled. Again, they are projecting to the outside of the tetramer.
Using the iCn3D model above, what kind of symmetry is displayed by the tetramer?
- Answer
-
Remember that the structures available from the Protein Data Bank are just visualized data based on electron density maps which are converted to 3D structures of connected atoms using computer processing and modeling to give atoms with XYZ coordinates. Often even the default PDB structures need more analyses to optimize our understanding of the structure and function of the molecules. Such is the case with the FIBCD PDB ID 6ZR4 structure. As mentioned above the key amino acids interacting with chitin derivatives are not shown interacting with anything. We can further process the file to show these interactions.
The PDB page for 6ZR4 actually contains three different structures. Two are called the Biological Assemblies (or the biologically functional unit). Both biological assemblies are homotetramers. One consists of just the A chain, each with an N-linked (GlcNAc)2Man trisaccharide. The other biological assembly contains just B chains without the trisaccharide. These can be seen by clicking Figure 5 below which will take you to the PDB 6ZR4 page.
Figure 5: Biological assemblies and unit cell/asymmetric unit for FIBCD (6ZR4)
Toggle the > circled in red in the above figure to move between biological assembly 1 for the A chain with trisaccharide, 2 for the B chain tetramer.
The last structural file for 6ZR4 is the asymmetric unit, which is the smallest part of a crystal lattice that repeats to create the entire crystal lattice to fill out space. In this case, the asymmetric unit consist of one A chain and one B chain, each in monomeric form, that are separated in space from each other. The asymmetric unit can be seen by toggling the > again.
Why might the A chain have an N-linked (GlcNAc)2Man attached to it?
- Answer
-
In the interactive iCn3D model below, you can toggle between an even larger biological assembly that contains multiple unit cells (shown in the static Figure 6 below) and the asymmetric unit (a separate A and B chain) of the tetrameric fibrinogen-like recognition domain of FIBCD (6ZR4). (FReD)
Figure 6: A Biological assembly and the asymmetric unit of the tetrameric fibrinogen-like recognition domain of FIBCD (6ZR4). (Copyright; author via source). Click the image for a popup or use this external link: https://structure.ncbi.nlm.nih.gov/i...SSKsqaE9JZjax8.
The A chains are shown in magenta and the B chains in dark blue. With your mouse toggle between the biological assembly to the unit cell asymmetric unit (one A and one B monomer, which comprise the unit cell) as follows: choose Analysis, Assembly, Asymmetric Unit. For this iCn3D model, the PDB (not the mmdb) file was loaded into iCn3D.
The magenta subunits are the 4 monomers of the A chain tetramer, each of which has an N-linked (GlcNAc)2Man trisaccharide at Asn 340. The top right trisaccharide is shown in SNFG cartoons and labeled. Even in this larger assembly, we don't see the interactions of the trisaccharide shown with SNFG coloring with the proteins.
How many asymmetric subunits are shown in the image above?
- Answer
-
Trisaccharide Interactions with the extracellular domain
To reiterate, our goals is to explore how the extracellular FReD domain of this membrane protein recognizes chitin. To explore this, a series of separate crystal structures were made containing bound GlcNAc, N-Acetyl-Gly, or(GlcNAc)2 to probe the binding site of chitin derivatives. Let's use iCn3D to model the interactions of a single bound GlcNAc that co-crystallized with the tetramer (6ZQR).
Let's create an iCn3D model of a single GlcNAc bound to the tetramer using iCn3D and these instructions
- File, input 6ZQR into the box, and select Load Biological Unit (Assembly)
- Analysis, Seq. and Annotations, and then in right hand window choose the details tab
- Orient the tetramer so the gold monomer is to the upper right
- Click Protein 6ZQR_A2, which is the gold chain
- Select, Save Selection, and name it AChain
- Hover over the GlcNAc to the right of the gold chain and verify it is labeled as NAG 3584, then alt-click it to select it
- Select, Save Selection, and name it NAG-A2
- In the selected sets, Define sets window, Ctrl-click AChain so it and NAG-A2 are highlighted
- View, View Selection to limit the view to these two molecules
- To view interaction: Analysis, Interactions, and then unclick Halogens and Contacts/Interactions
- keep the rest of the defaults
- For select the first set, choose NAG-A2; For select the second set, choose AChain
- Click 4. 3D Display Interactions
- Select, Save Selection, name it Interact
- Close H bond interaction window
- In the Defined Sets window, choose AChain; Color, Unicolor, Grey, Light Gray
- In the Defined Sets window, choose Interact; Color, Atom
- With interact still highlighted: Analysis, Label, Per residue and label
- Analysis, Label Scale, 2
- In Defined Sets window, choose NAG-A2; Style, Chemical, Sphere
- Style, Background, Transparent
- File, Save File, iCn3D PNG image, Original Size.
- File, Share Link, Lifetime Short Link and paste it below.
You can use either of these File Save methods to reopen your final rendered structure.
- Answer
-
Chitin and fragments of it would logically bind across a more extended binding site in the extracellular domain of the protein compared to a single GlcNAc, which would bind in a localized binding site. Fortuitously (a word used in the paper), the N-linked glycan trisaccharide from subunit A is bound by subunit B in an extended S1-binding site on the FIBCD1-FReD surface in the crystal structure of the protein obtained without added ligand (6ZR4)! Hence this fortuitously-bound state would give clues as to the interacting residues required for the binding of longer chitin and chitin derivatives! We will model that in iCn3D now.
Specifically, let's look at the complex of the larger glycan, the N-linked (GlcNAc)2Man trisaccharide bonded to Asn 340, and the B chain to get hints about the probable extended binding site.
We have to process the PDF yet again to see the interaction of the trisaccharide on an A monomer with the B monomer which doesn't have the glycan. The above 6ZR4 PDB file does NOT show an interaction of the N-linked trisaccharide on the A chain with the B chain in either of the Biological Assemblies or in the asymmetric unit. Some transformations of the PDB file must be made to produce the coordinates to show that. These transformations are explored in detail in the iCn3D Tutorial: Asymmetric Unit, Unit cell, and Molecular Assemblies found in Volume V. An iCn3D model of the new file showing A:B chain interactions and the binding of the N340-linked trisaccharide to the B chain is shown below.
You will use the output file of this PDB transformation to model the interaction of the N-linked trisaccharide on the A monomer with the non-glycosylated B monomer below. First, let's look at the interactions described in the paper. We will then try to duplicate them with iCn3D.
Figure 7 below shows an image of the trisaccharide interacting with the monomer chain. The first GlcNAc residue in the glycosylation chain interacts with the subunit B ligand-binding site S1(1) as it did in the structure containing one GlcNAc (see above); the second GlcNAc crosses the Cys414–His415 and bindings at the S2 site. The mannose, a proxy for a third GlcNAc in chitin derivatives, presumably binds in the S3 site.
Figure 7: The S1 binding site on FIBCD1.
Panel (A) shows the native glycan ((GlcNAc)2Man) bound in the FIBCD1 subunit B S1 site. The Asn340-linked glycan is shown as a stick model (green) binding to the subunit B S1 ligand-binding site (blue).
Panel (B) Sequence alignment in the S1 ligand-binding region in FIBCD1 and homologous proteins TL5A, Ficolin-1, Ficolin-2, and Ficolin-3. Sequence numbers for each protein domain are indicated. The primary N-acetyl S1(1) binding pocket residues are highlighted in orange and calcium coordinating residues in red. Additional residues involved in the neighboring pocket (S1(2)) in FIBCD1 are highlighted in green and residues His396 and Arg412, involved in forming S1(3) on the FIBCD1-FReD, are highlighted in cyan. Ficolin-1 and the homologous tachylectin 5A (TL5A) also bind acetylated ligands via the S1 site with each coordinating ligands in a similar manner to FIBCD1. In ficolin-2, changes to a number of residues at the S1 site render it inactive and, instead, ligands are coordinated at alternative sites (designated S2, S3, and S4), which in ficolin-2 form an extended binding surface upon which a range of carbohydrate and noncarbohydrate ligands are bound
Now let's model the same interactions using iCn3D.
Continue with with these step. Use must download and use a local copy of this processed PDB file for the A:B chain trisaccharide interactions. The file was kindly made by Jiyao Wang.
- File, Open File, PDB and choose the concatenated PDB file (if not already opened)
- Analysis, Seq. and Annotations, and then in right hand window choose the details tab
- Click Chem Struct_C, which is the trisaccharide
- Select, Save Selection, and name it trisacch
- Analysis, Interactions and then unclick Halogens and Contacts/Interactions
- Uncheck Contacts/Interactions, set H bond to 4.2
- Keep default sets: (set 1 select - ie. the trisacch; set 2 unselected - ie. everything else)
- Click 4. 3D Display Interactions
- Select, Save Selection, name it Interact
- Close H bond interaction window
- In the Defined Sets window, choose Interact and trisacch
- View Selection
- Color Atom
- Analysis, Label, Per residue and label
- Style, Background, Transparent
- File, Save File, iCn3D PNG image, Original Size
Crop their image with your favorite program and paste it below.
- Answer
-
A colored-coded surface view of the B chain bound to the trisaccharide is shown in Figure 8 below. It shows the three subsites (1), (2) and (3) within the more extend S1 binding site.
Fig 8: (C) FIBCD1 surface view showing the native glycan (green) bound in subunit B with the N-linked GlcNAc in the primary S1(1) conserved pocket, the second GlcNAc in the neighboring pocket, designated S1(2), and mannose in the designated S1(3) pocket.
Overlaid (least-squares fit of main chain residues of each structure) in the S1(1) pocket are GlcNAc from the GlcNAc-bound structure (grey) and from the (GlcNAc)2 ligand-bound structure (cyan) and acetate (pink, present in the subunit A S1(1) of native and subunit B of the GlcNAc, GlcNAc2, Neu5Ac and GalNAc4S structures). The sulfate ion in S1(2) of subunit A of the Native structure, and in S1(3) of subunit B of all the ligand-bound structures are also shown along with the acetate ion in the subunit B S1(2) pocket of the N-acetylalanine structure.
Let's try to replicate that image in iCn3D.
Take a snip of the surface of the B chain in interaction with the trisaccharide following the instructions below.
- Choose File, Open, PDB Appendable and the file you downloaded for the previous model
- Analysis, Seq. and Annotations, and then in right hand window choose the details tab
- Click Protein Struct_b (B chain)
- Select, Save Selection, and name it BChain
- Click Struct_c (trisaccharide)
- Select, Save Selection, and name it trisacch
- In the selected Sets Window, ctrl-click just trisacch and Chain B
- View, View Selected
In the sequences and annotations window for B Chain use your mouse to select then save
- H396 (for mannose interaction), then Select, Save Selection, name it H396-Mann
- R412, N413 for Mann, adjacent NAG, the save with name R412N413
- H415, Y431 for terminal Mann, then save with name H415Y431
- Click B chain in Defined sets,
- Color, Unicolor, White
- Click the 3 sets of interacting amino acids sets in Defined Sets,
- Protein, Sidechains, spheres
- Select H396Mann, Color, Unicolor Red; Select R412N413, Color, Unicolor Cyan; Select H415Y431, Color, Unicolor, Orange
- Choose Chain B, Style, Surface Type, MS (molecular surface) with Context
- In Defined sets, ctrl-click trisach and 3 sets of amino acid interactions
- Analysis, Label, Residue & number
- Style, Background, Transparent
- File, Save File, iCn3D PNG image, Original Size
Crop their image with your favorite program and paste it below.
- Answer
-
Let's summarize the binding interaction results.
Using the above figures and iCn3D models, describe the binding interactions between the trisaccharide and the S1(1), S1(2), and S1(3) subsites of the S1 binding site
- Answer
-
The first GlcNAc residue in the glycosylation chain interacts with the subunit B ligand-binding site S1(1) in essentially the same manner as the other bound ligands; the second GlcNAc crosses the Cys414–His415 backbone allowing the N-acetyl group to fit into a pocket delineated by the residues His382 and Arg412–His415, establishing a second acetyl-binding pocket S1(2). Mannose, the third carbohydrate of the CHO-derived glycan in the native structure, further extends the glycan-binding site to a third pocket S1(3) by interacting with the His396 and Arg412 side chains.
Binding Analyses
To complement the structural analyses, binding analyses were conducted using GlcNAc and larger acetylchitooligosaccharides up to (GlcNAc)11. They used these GlcNAc derivatives in an ELISA assay to alter the binding of the extracellular domain of the Fibrinogen C domain-containing protein 1 (FIBCD1) to acetylated bovine serum albumin (AcBSA) that was irreversibly adsorbed to microtiter plate (a previously validated method). The acetylated-BSA mimics the acetylated Glc in chitin. The extracellular (soluble) domain is abbreviated as FReD (fibrinogen-like recognition domain). See Chapter 5.7 for a description of ELISAs).
How would the addition of GlcNAc-containing glycans affect the binding of FReD in the ELISA?
- Answer
-
The results of one set of ELISAs are shown in Figure 9 below.
Figure 9: Inhibition of FIBCD1-FReD binding to acetylated bovine serum albumin by N-acetylchitooligosaccharides.
Panel A, representative ELISA-based setup demonstrating acetyl group–specific inhibition of binding between FIBCD1-FReD and AcBSA by the tested N-acetylchitooligosaccharides. B
Panel B, statistical comparison of the tested N-acetylchitooligosaccharides and their ability to inhibit binding between FIBCD1-FReD and AcBSA at half maximal inhibitory concentration (IC50). Results shown in A and B are combined data from three to four independently performed experiments. Statistics: Data are presented as mean ± SD. ∗p < 0.05 by one-way analysis of variance (ANOVA) with post hoc Tukey test.
Interpret the results
- Answer
-
Another method, thermophoresis, was used to measure the actual dissociation constant, KD, for GlcNAc derivatives for FReD. This technique does not require immobilization of the ligand. Instead, it measures how the movement of a molecular species through a region containing a temperature gradient is affected by the binding of another species. Movement is typically detected by laser-induced of the sample. The movement in the temperature gradient occurs through diffusion which would be affected by the size, mass, hydration, etc. of the molecules of interest. The data can be translated into binding curves, which are shown in Figure 10 below.
Figure 10: Analysis of FIBCD1-FReD interactions with acetylated molecules by microscale thermophoresis.
Representative microscale thermophoresis binding curves of recombinant FIBCD1-FReD with (A) N-acetylated glucosamine (GlcNAc), (B) N-acetylchitooligosaccharide (GlcNAc)5, (C) N-acetylated mannosamine (ManNAc), and (D) acetylated BSA (AcBSA). The calculated KD values of FIBCD1-FreD to AcBSA (53 ± 11 nM), GlcNAc (27 ± 9 mM), (GlcNAc)5 (1.2 ± 0.4 mM), and ManNAc (31 ± 8 mM)
Interpret the result.
- Answer
-
Finally, site-directed mutagenesis was used to convert the wildtype enzyme to the H396A enzyme. The results are shown in Figure 11 below.
Figure 11: Repeated measurements of GlcNAc, (GlcNAc)2, and (GlcNAc)3 inhibition of FIBCD1-FReD variants binding to acetylated bovine serum albumin. Repeated measurements demonstrating acetyl group–specific inhibition of binding between
Panel (A) wildtype (WT) FIBCD1-FReD and (B) mutant variant H396A FIBCD1-FReD to immobilized acetylated bovine serum albumin by the tested N-acetylchitooligosaccharides. Results displayed are combined data from eight and five independently performed experiments, respectively. Statistics: Data are presented as mean ± SD. ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001 by one-way analysis of variance (ANOVA) with post hoc Tukey test.
Why did they make the H396A mutation? What effect did it have?
- Answer
-