The amino terminus of varicella-zoster virus (VZV) glycoprotein E is required for binding to insulin-degrading enzyme, a VZV receptor

Abstract

Varicella-zoster virus (VZV) glycoprotein E (gE) is required for VZV infection. Although gE is well conserved among alphaherpesviruses, the amino terminus of VZV gE is unique. Previously, we showed that gE interacts with insulin-degrading enzyme (IDE) and facilitates VZV infection and cell-to-cell spread of the virus. Here we define the region of VZV gE required to bind IDE. Deletion of amino acids 32 to 71 of gE, located immediately after the predicted signal peptide, resulted in loss of the ability of gE to bind IDE. A synthetic peptide corresponding to amino acids 24 to 50 of gE blocked its interaction with IDE in a concentration-dependent manner. However, a chimeric gE in which amino acids 1 to 71 of VZV gE were fused to amino acids 30 to 545 of herpes simplex virus type 2 gE did not show an increased level of binding to IDE compared with that of full-length HSV gE. Thus, amino acids 24 to 71 of gE are required for IDE binding, and the secondary structure of gE is critical for the interaction. VZV gE also forms a heterodimer with glycoprotein gI. Deletion of amino acids 163 to 208 of gE severely reduced its ability to form a complex with gI. The amino portion of IDE, as well an IDE mutant in the catalytic domain of the protein, bound to gE. Therefore, distinct motifs of VZV gE are important for binding to IDE or to gI.

FIG. 1.

Expression of VZV gEt and gEt truncation mutants. (A) Map of the extracellular domain of VZV gE fused to immunoglobulin Fc (gEt) and gEt truncation mutants. (B) Immunoblot of VZV gEt and truncation mutants, using anti-human Fc antibody. The numbers to the right of panel B are molecular weights in thousands.

FIG. 2.

Deletion of gE amino acids 32 to 71 abolishes binding to IDE. (A) gEt or gEt truncation mutants were immobilized onto protein A-Sepharose beads, human melanoma cell lysates were added, and after washing and boiling, the proteins were immunoblotted with anti-IDE antibody. (B) The pull-down assay described for panel A was performed in the presence or absence of peptides at the indicated concentrations, followed by immunoblotting with anti-IDE antibody. (C) Pull-down assays were performed with cells cotransfected with gEt and gIt and immunoblotted with anti-IDE antibody. The numbers to the right of the panels are molecular weights in thousands.

FIG. 3.

VZV-HSV chimeric gE protein does not bind to IDE. (A) Structure of VZV-HSV chimeric gE. (B) Expression of VZV-HSV chimeric gE (lane 2) and HSV-2 gE (lane 4) was detected by immunoblotting with antibody to HA. Control plasmid 1 contains the chimeric gE sequence but does not express the protein; control plasmid 2 expresses GFP. (C) Immunoblot of cell lysate with anti-IDE antibody shows IDE in the lysate, but immunoprecipitation of HA-tagged chimeric gE or HSV gE using antibody to HA does not coimmunoprecipitate IDE. Control plasmids 1 and 2 are as described for panel B and do not coimmunoprecipitate IDE. The numbers to the right of panel B and the left of panel C are molecular weights in thousands.

FIG. 4.

VZV gE amino acids 163 to 208 are important for interacting with VZV gI. CV1/EBNA cells were transfected with plasmids encoding gEt or gEt mutant proteins in the presence or absence of plasmid gIt, which encodes the extracellular domain of gI without an Fc tag. gEt-gIt complexes were then precipitated with protein A-Sepharose and resolved by sodium dodecyl sulfate-polyacrylamide gel electrophoresis. gIt was detected, using antibody to gI. A plasmid expressing GFP was used as a control.

FIG. 5.

gE interacts primarily with the amino portion of IDE. (A) Expression of human IDE-w.t., the IDE-N fragment, the IDE-C fragment, and the IDE-E111Q mutant in E. coli, detected by immunoblotting with antibody to IDE. The additional bands in IDE-E111Q are presumably due to protein degradation. (B) Binding of IDE-w.t. (lane 1), IDE-E111Q (lane 2), or melanoma cell lysate containing IDE (lane 5) to gEt. gEt was immobilized onto protein A-Sepharose beads, IDE-w.t. or IDE-E11Q was added, and after washing and boiling, proteins were immunoblotted with anti-IDE antibody. Control 1 encodes the vaccinia 7.5 protein fused to Fc, and control 2 encodes soluble Jam fused to Fc. (C) IDE-E111Q, IDE-N, and IDE-C bind to gEt. Pull-down assays were performed as described for panel B, using 10 μg of each purified IDE protein. The controls in lanes 2, 4, and 6 encode vaccinia P7.5 fused to Fc, while the control in lane 8 encodes EBV BZLF2 fused to Fc. Lanes 7 and 8 are from a separate experiment in which a lower-stringency buffer (phosphate-buffered saline) was used for the binding assay than for the experiment shown in lanes 1 to 6. The numbers to the right of the panels are molecular weights in thousands.

FIG. 6.

Map of VZV gE motifs that are important for binding to IDE and gI. N, putative N-linked glycosylation sites; C, cysteine residues; SP, signal peptide; TM, transmembrane domain; CT, cytoplasmic tail.