Variola virus F1L is a Bcl-2-like protein that unlike its vaccinia virus counterpart inhibits apoptosis independent of Bim

Abstract

Subversion of host cell apoptosis is an important survival strategy for viruses to ensure their own proliferation and survival. Certain viruses express proteins homologous in sequence, structure and function to mammalian pro-survival B-cell lymphoma 2 (Bcl-2) proteins, which prevent rapid clearance of infected host cells. In vaccinia virus (VV), the virulence factor F1L was shown to be a potent inhibitor of apoptosis that functions primarily be engaging pro-apoptotic Bim. Variola virus (VAR), the causative agent of smallpox, harbors a homolog of F1L of unknown function. We show that VAR F1L is a potent inhibitor of apoptosis, and unlike all other characterized anti-apoptotic Bcl-2 family members lacks affinity for the Bim Bcl-2 homology 3 (BH3) domain. Instead, VAR F1L engages Bid BH3 as well as Bak and Bax BH3 domains. Unlike its VV homolog, variola F1L only protects against Bax-mediated apoptosis in cellular assays. Crystal structures of variola F1L bound to Bid and Bak BH3 domains reveal that variola F1L forms a domain-swapped Bcl-2 fold, which accommodates Bid and Bak BH3 in the canonical Bcl-2-binding groove, in a manner similar to VV F1L. Despite the observed conservation of structure and sequence, variola F1L inhibits apoptosis using a startlingly different mechanism compared with its VV counterpart. Our results suggest that unlike during VV infection, Bim neutralization may not be required during VAR infection. As molecular determinants for the human-specific tropism of VAR remain essentially unknown, identification of a different mechanism of action and utilization of host factors used by a VAR virulence factor compared with its VV homolog suggest that studying VAR directly may be essential to understand its unique tropism.

Figure 1

(a) Sequence alignment of vaccinia virus (VV) with variola virus (VAR) F1L. Red denotes sequence differences in the Bcl-2 domain, H indicates helices. (b) Sequence variations between VV and VAR F1L are mapped in red on a molecular surface of VV F1L (gray). (c) BH3-binding profile of VAR_F1L. K_D values were determined by isothermal calorimetry

Figure 2

Structures of VAR F1L:Bak and Bid BH3 complexes. (a) Cartoon diagram of F1L:Bak BH3 complex. F1L chains are shown as a cartoon (lime and salmon), with helices α1-7 labeled. Two Bak BH3 chains are show in cyan and light blue. (b) Cartoon diagram of F1L:Bak BH3 complex. This view looks into the hydrophobic BH3-binding grove, which is formed by helices α2–5. F1L helices α2–7 from monomer 1 (lime) are labeled, as is helix α1′ from monomer 2 (salmon). Bak BH3 is shown in cyan. (c) Cartoon diagram of F1L:Bid BH3 complex. F1L (lime and salmon) in complex with Bid BH3 (wheat and lilac). The view is as in a. (d) Cartoon diagram of F1L:Bid BH3 complex. This view is as in b. F1L helices α2–7 from monomer 1 (lime) are labeled, as is helix α1′ from monomer 2 (salmon). Bid BH3 is shown in wheat. (e) Stereo diagram of the F1L (lime):Bak (cyan) complex interface. The F1L surface is shown in gray, except for magenta shading indicating the floor of the peptide-binding groove. F1L residues are labeled in lime, Bim BH3 residues are labeled in cyan. (f) Stereo diagram of the F1L (lime):Bim (wheat) complex interface. View and labeling is as in c, except for labeling of Bid residues (in black)

Figure 3

VAR F1L inhibits Bax- but not Bak-mediated apoptosis. (a) Viability of wild-type, Bax^−/−, Bak^−/− and Bax^−/−/Bak^−/− DKO cells. MEF cells stably overexpressing VAR F1L or vector, treated with 1.5 μM thapsigargin and cultured for 24 h. (b) Viability of 293T cells transiently overexpressing either Bim_EL or Bid as well as F1L, Bcl-2 or vector control. Error bars are ±S.E.M. with n=3