Cyanovirin-N Binding to N-Acetyl-d-glucosamine Requires Carbohydrate-Binding Sites on Two Different Protomers

Abstract

Cyanovirin-N (CV-N) binds high-mannose oligosaccharides on enveloped viruses with two carbohydrate-binding sites, one bearing high affinity and one low affinity to Manα(1-2)Man moieties. A tandem repeat of two CV-N molecules (CVN2) was tested for antiviral activity against human immunodeficiency virus type I (HIV-1) by using a domain-swapped dimer. CV-N was shown to bind N-acetylmannosamine (ManNAc) and N-acetyl-d-glucosamine (GlcNAc) when the carbohydrate-binding sites in CV-N were free to interact with these monosaccharides independently. CVN2 recognized ManNAc at a K_d of 1.4 μM and bound this sugar in solution, regardless of the lectin making amino acid side chain contacts on the targeted viral glycoproteins. An interdomain cross-contacting residue Glu41, which has been shown to be hydrogen bonding with dimannose, was substituted in the monomeric CV-N. The amide derivative of glucose, GlcNAc, achieved similar high affinity to the new variant CVN-E41T as high-mannose N-glycans, but binding to CVN2 in the nanomolar range with four binding sites involved or binding to the monomeric CVN-E41A. A stable dimer was engineered and expressed from the alanine-to-threonine-substituted monomer to confirm binding to GlcNAc. In summary, low-affinity binding was achieved by CVN2 to dimannosylated peptide or GlcNAc with two carbohydrate-binding sites of differing affinities, mimicking biological interactions with the respective N-linked glycans of interest and cross-linking of carbohydrates on human T cells for lymphocyte activation.

Figure 1

Binding affinity of CVN2 to ManNAc. (A) Schematic representation of monomeric CV-N vs domain-swapped CVN2 binding to carbohydrates, i.e., ManNAc. Green halves represent low-affinity binding domain A, and blue halves represent high-affinity binding domain B. (B) ITC data showing the binding of CVN2 to ManNAc. Thermograms are shown, and binding curves are fitted to the one-set of sites binding model. T = 298 K.

Figure 2

ITC data showing the binding of CVN2, mutant E41T dimer, and CVN-E41A to GlcNAc. (A) Schematic representation of monomeric CVN-E41A and dimeric CVN2 binding to GlcNAc. (B) Binding affinity of CVN2 to GlcNAc. Thermogram is shown, and the binding curve fitted to the one-set of sites model. (C) Mutant E41T dimer binds to GlcNAc. Thermogram is shown, and binding analyzed based on one-set of sites binding model. T = 298 K. Ligand in cell. (D) 3D X-ray structure from CV-N (PDB ID:3S3Y) with mutated Ala41 shown as a stick. High-affinity carbohydrate-binding site (AAs 40–89) in blue; low-affinity carbohydrate-binding site (AAs 1–39, 90–101) in green. (E) Binding of CVN-E41A to GlcNAc. Thermogram is shown, and binding analyzed based on one-set of sites binding model. T = 298 K. The experiments were performed in triplicate.

Figure 3

(A–D) Interaction between CVN, CVN-E41T, CVN2, mutant dimer, and GlcNAc. Bottom trace: ¹H NMR of GlcNAc in deuterated phosphate buffer pH = 7.4 (T = 298 K). Upper trace: STD-NMR experiments in the presence of (A) CVN, (B) CVN-E41T, (C) CVN2, and (D) mutant dimer. (E–H) Interaction between CVN, CVN-E41T, CVN2, mutant dimer, and ManNAc. Bottom trace: ¹H NMR of ManNAc in deuterated phosphate buffer pH = 7.4 (T = 298 K). Upper trace: STD-NMR experiments in the presence of (E) CVN, (F) CVN-E41T, (G) CVN2, and (H) mutant dimer. Orange, yellow, and blue bars highlighting the peaks of the amide group: orange (−NH), yellow (−COCH₃), and blue (−CH₃).