Characterization of EBV gB indicates properties of both class I and class II viral fusion proteins

Abstract

To gain insight into Epstein-Barr virus (EBV) glycoprotein B (gB), recombinant, secreted variants were generated. The role of putative transmembrane regions, the proteolytic processing and the oligomerization state of the gB variants were investigated. Constructs containing 2 of 3 C-terminal hydrophobic regions were secreted, indicating that these do not act as transmembrane anchors. The efficiency of cleavage of the gB furin site was found to depend on the nature of C-terminus. All of the gB constructs formed rosette structures reminiscent of the postfusion aggregates formed by other viral fusion proteins. However, substitution of putative fusion loop residues, WY(112-113) and WLIY(193-196), with less hydrophobic amino acids from HSV-1 gB, produced trimeric protein and abrogated the ability of the EBV gB ectodomains to form rosettes. These data demonstrate biochemical features of EBV gB that are characteristic of other class I and class II viral fusion proteins, but not of HSV-1 gB.

Figure 1

Schematic representation of A) Full-length gB: ectodomain (residues 23 to 683) is shown as filled with dots, the C-terminal intracellular domain (CTD) is presented by the clear box. Mature protein is produced after cleavage of the 22-residues long signal peptide at the N-terminus (horizontal lines). The three membrane-proximal regions residues 686-705, 710-730 and 733-753 are between the ectodomain and the CTD, B) Short ectodomain variant (EctoS) comprises residues 23-683 of gB and lacks all three membrane-proximal regions, C) Long ectodomain variant (EctoL), residues 23-733 of gB, includes first two membrane-proximal regions, and a GCN4 trimerization domain (black box), D) EctoL/EK is the same as the variant as described in C), but contains an enterokinase recognition site DDDDK instead of the wild-type furin cleavage site RRRRR (residues 428-432).

Figure 2

Western blot of fractions collected during purification of EctoS gB variant on CL55 antibody column. Primary antibody used for blotting recognizes histidine tag. The samples were run under non-reducing conditions. Elution was done with 0.2 M glycine pH 2.5 as described in Materials and Methods section.

Figure 3

Coomassie brilliant blue stained SDS-PAGE gels: A) EctoS and EctoL variant migrate as single polypeptide chains in the absence of reducing agent (left). Upon addition of DTT a significant fraction of the EctoL, but not of EctoS variant, stays uncleaved and is labeled with a star (right). The presence of ~70 kDa and ~40 kDa EctoS fragments under reducing conditions indicating cleavage of EctoS variant (right). B) EctoL/EK variant migrates as a single chain regardless of the presence of reducing agent indicating the lack of cleavage (left and middle lanes). Extraneously added enterokinase (EK) cleaves the protein into two fragments.

Figure 4

A) Chromatogram obtained for immunoaffinity purified EctoS variant run on S200 column. gB elutes in the void volume of the column, indicating presence of oligomers of molecular mass larger than 1300 kDa. B) Material from the void volume fractions was visualized by electron microscopy. Irregular rosettes, reminiscent of rosettes formed by class I and class II fusion proteins, were detected. Each arm is 15-20 nm long and represents most likely a gB trimer.

Figure 5

Sucrose gradient centrifugation of: A) supernatant containing EctoS gB variant, and B) EctoS gB purified using immuno-affinity chromatography. The sucrose gradient distribution of the recombinant gB from supernatant does not differ from the distribution profile observed for the purified variant. These data suggest strongly that oligomerization state of the recombinant gB did not change during immuno-affinity purification.

Figure 6

A) Chromatogram obtained for immunoaffinity purified EctoL variant run on S200 column in the presence of deoxycholate in the sample and running buffer. gB elutes in the 8.67 ml peak, and is most likely in trimeric form. B) gB rods were visualized by EM in the EctoL gB sample obtained from fractions corresponding to the gB trimer peak. Rods are 15-20 nm long, although some longer ~30 nm rods are present (marked with asterisk sign) and may correspond to two gB trimer rods held together.

Figure 7

Putative fusion loops and their role in rosette formation. X-ray structure of HSV-1 gB ectodomain: A) for clarity only HSV-1 gB monomer is shown, B) gB HSV-1 trimer is shown. Residues forming putative fusion loops in HSV-1 gB HR^177-178 and RVEA^258-261 are represented, respectively, as blue and red sticks. C) Hypothetical model of rosettes observed for EBV gB created by docking HSV-1 gB trimers to interact thought the putative fusion loops (which in EBV gB contain hydrophobic residues WY^112-113 and WLIW^193-196). EBV gB rosettes shown in Figure 4 are irregular indicating that the geometry constrains allow association of a variable number EBV gB trimers. However, the model shown here is descriptive and was not created to model rosettes in any particular manner. Images were generated using PyMOL (8). PDB accession number for HSV-1 gB is 2gum (23).

Figure 8

Elution profile of EctoS/FL gB variant run on S200 column. A) gB elutes in the 8.86 ml peak which corresponds to a gB trimer, and gB is detected by SDS-PAGE and Western blotting using anti-gB antibody. B) Numerous individual rods, 15 nm long, representing gB trimers, were observed by electron microscopy. Rosette structures were not observed in EctoS/FL protein preparations.