CrmE, a novel soluble tumor necrosis factor receptor encoded by poxviruses

Abstract

Cytokines and chemokines play a critical role in both the innate and acquired immune responses and constitute prime targets for pathogen sabotage. Molecular mimicry of cytokines and cytokine receptors is a mechanism encoded by large DNA viruses to modulate the host immune response. Three tumor necrosis factor receptors (TNFRs) have been identified in the poxvirus cowpox virus. Here we report the identification and characterization of a fourth distinct soluble TNFR, named cytokine response modifier E (CrmE), encoded by cowpox virus. The crmE gene has been sequenced in strains of the orthopoxviruses cowpox virus, ectromelia virus, and camelpox virus, and was found to be active in cowpox virus. crmE is expressed as a secreted 18-kDa protein with TNF binding activity. CrmE was produced in the baculovirus and vaccinia virus expression systems and was shown to bind human, mouse, and rat TNF, but not human lymphotoxin alpha, conjugates of lymphotoxins alpha and beta, or seven other ligands of the TNF superfamily. However, CrmE protects cells only from the cytolytic activity of human TNF. CrmE is a new member of the TNFR superfamily which is expressed as a soluble molecule that blocks the binding of TNF to high-affinity TNFRs on the cell surface. The remarkable finding of a fourth poxvirus-encoded TNFR suggests that modulation of TNF activity is complex and represents a novel viral immune evasion mechanism.

FIG. 1

Sequences of vTNFRs. Shown are pairwise alignments of the predicted amino acid sequences of CrmE in different orthopoxviruses (A) and of vTNFRs identified in CPV-BR (CrmB, -C, and -D) and CPV-EP (CrmE) (B). Solid backgrounds represent differences, and shaded backgrounds represent regions of high similarity. Dots and stars indicate deletions and stop codons, respectively. Solid circles and triangles show predicted N-glycosylation and signal peptide cleavage sites, respectively. The positions of CRDs are indicated. The accession numbers of the sequences are as follows: Y15035 (CPV-GRI90 crmE), AJ272008 (CPV-EP crmE), AJ272005 (EV Hampstead crmE), AJ272006 (EV Moscow crmE), AJ272009 (camelpox virus crmE), Q85308 (CPV-BR crmB), U87234 (CPV-BR crmD), and U55052 (CPV-BR crmC).

FIG. 2

Expression of vTNFRs in the baculovirus and VV expression systems. (A) Sf cells infected with AcNPV or the indicated recombinant viruses were pulse-labeled with [³⁵S]cysteine and [³⁵S]methionine from 26 to 29 h p.i. (B) BSC-I cells were infected with VV-WR or vCrmE and pulse-labeled from 4 to 8 h p.i. In both panels A and B, proteins present in cells and media were analyzed by SDS-PAGE and visualized by fluorography. The positions of the expressed proteins in supernatants and cell extracts are indicated. Molecular masses (in kilodaltons) are shown.

FIG. 3

TNF binding activity of vTNFRs expressed in the baculovirus (A) or VV (B) system. Supernatants from Sf (100 μl, corresponding to 2 × 10⁵ cells) or BSC-I (50 μl, corresponding to 2.5 × 10⁵ cells) cell cultures infected with the indicated recombinant baculovirus or VV, respectively, were incubated with 200 pM human ¹²⁵I-TNF. Levels of bound ¹²⁵I-TNF were determined by precipitation with PEG and filtration. The background radioactivity precipitated with PEG in the presence of binding medium has been subtracted. The specific ¹²⁵I-TNF binding of duplicate samples (means ± standard deviations) is shown.

FIG. 4

TNF binding specificity of CrmE. Supernatants (2 μl, corresponding to 3.8 × 10³ cells) from BSC-I cells infected with recombinant VV expressing CrmE were incubated with 150 pM human ¹²⁵I-TNF in the absence (no competitor) or in the presence of the indicated fold excess of unlabeled human (Hu), mouse (Mo), or rat TNF (A), a 500-fold excess of cold human LTα or the conjugate LTα1/β2 or LTα2/β1 (B), or a 500-fold excess of human GITR, CD40L, BAFF, TWEAK, TRAIL, 4-1BBL, RANK, or MIP-1α (C). Levels of bound ¹²⁵I-TNF were determined by precipitation with PEG and filtration. The percent specific ¹²⁵I-TNF binding of duplicate samples (means ± standard deviations) is calculated relative to binding in the absence of competitor, which was 4,828 cpm (A and B) or 3,606 cpm (C).

FIG. 5

Competition of TNF binding to U937 cells. Different amounts of medium (25 or 50 μl, corresponding to 5 × 10⁴ or 1 × 10⁵ cells, respectively) from cultures of Sf cells infected with the indicated recombinant baculoviruses were incubated with 200 pM human ¹²⁵I-TNF for 1 h at 4°C. U937 cells were added and incubated for 2 h at 4°C, and the amount of radioactivity bound to cells was determined by phthalate oil centrifugation. The specific ¹²⁵I-TNF binding of duplicate samples (mean ± standard deviation) is shown.

FIG. 6

Biological activity of CrmE. Shown is the effect of CrmE on the cytolysis of mouse L929 cells induced by human (A), mouse (B), or rat (C) TNF or by human LTα (D). Crystal violet staining was used to determine, in triplicate samples, the cell viability after 12 h of treatment with TNF or LTα alone or in the presence of binding medium (Medium) or supernatants from BSC-I cells infected with VV-WR or recombinant VV expressing CrmE (vCrmE). The amount of supernatant was 10 μl, corresponding to 2.5 × 10³ cells. Percent cytotoxicity was calculated as described in Materials and Methods.