|Home | About | Journals | Submit | Contact Us | Français|
The four serotypes of dengue virus present a formidable challenge for the development of efficacious human vaccines. Cockburn and colleagues, in this issue of Structure, describe the structural basis of a cross-reactive neutralizing antibody, providing greater insight into immune protection and pathogenesis.
Dengue virus (DENV) is a mosquito-borne member of the flavivirus genus responsible for roughly 50 million human infections each year. Four antigenically-related serotypes of DENV circulate in tropical and subtropical regions of the globe. Infection by any of the four DENV serotypes may cause a self-limiting febrile illness that is rarely fatal, and is thought to confer immunity to reinfection by the same serotype of DENV. However, the secondary infection of a DENV-seropositive individual with a heterologous DENV serotype may lead to more severe clinical manifestations of disease, including hemorrhage and shock. The increased hyperendemicity of DENV in many regions of the globe has resulted in a dramatic increase in the number of severe DENV cases each year. The immunological mechanisms responsible for the more aggressive disease following the secondary infection of DENV-immune individuals is incompletely understood and may involve both arms of the adaptive immune response (Rothman, 2011). DENV-reactive antibodies are thought to contribute to not only protection from infection but also to exacerbated disease following secondary infections. Antibody-dependent enhancement of infection describes the more efficient infection of Fcγ-receptor-expressing cells in the presence of DENV-reactive antibodies and may contribute to severe disease (Kliks et al., 1989). While vaccines that protect against DENV are urgently needed, their development is complicated by a requirement to simultaneously protect against all four serotypes of DENV (Whitehead et al., 2007). Thus, understanding the molecular basis for the recognition of DENV by homologous and cross-reactive antibodies is critical for understanding factors that contribute to pathogenesis.
The envelope (E) protein is an elongated three domain structure incorporated into virus particles that orchestrates the attachment and entry of virions into cells (Mukhopadhyay et al., 2005). The majority of anti-flavivirus neutralizing antibodies recognize epitopes contained within or between the three domains of the E protein (Pierson et al., 2008). In this issue of Structure, Cockburn and colleagues (2012) present exciting new structural insights into the recognition of DENV by a murine monoclonal antibody (mAb) capable of neutralizing all four serotypes of DENV. mAb 4E11 is a well-characterized cross-reactive murine mAb raised against DENV1 that binds domain III (E-DIII) of the E protein. Biochemical analysis of the binding of this mAb to recombinant variants of E-DIII of DENV1 identified nine residues that significantly impact antibody recognition (Lisova et al., 2007). These critical residues map to the A- and G- βstrands that form the edge of E-DIII β-sandwich and have been shown to be involved the recognition of many group-reactive anti-flavivirus antibodies (Sukupolvi-Petty et al., 2007). The structure of 4E11 in complex with E-DIII from all four serotypes of DENV now described extends these studies by defining how variation in the 4E11 epitope among DENV strains modulates antibody binding and neutralization. 4E11 binds each DENV serotype in a similar manner via interactions with residues on the Aand G-strands surrounding a hydrophobic core region of E-DIII. These studies identify the conserved residues required for cross-reactivity among DENV strains as well as the contribution of more variable side chains toward the marked difference in binding affinity among DENV serotypes. Of significant interest, the relative position of 4E11 in complex with E-DIII from each DENV serotype reveals that the antibody shifts position on the A-strand epitope to accommodate the nonconserved residues among DENV strains. This raises the intriguing possibility that subtle differences in the angle of antibody engagement of an epitope may translate into substantial differences in the interaction of intact antibody molecules with the virus particle and in what numbers.
On mature DENV virions, 180 E proteins are arranged as rafts of antiparallel dimers organized in an unusual herringbone pattern (Mukhopadhyay et al., 2005). This dense arrangement imposes steric constraints for antibody recognition on the intact virion (Pierson et al., 2008). The A-strand epitope recognized by mAb 4E11 is not predicted to be accessible on the mature virion. The distal end of E domain II (E-DII) encodes a highly conserved hydrophobic fusion loop that is required for viral membrane fusion. On mature virions, the fusion loop packs into the hydrophobic pocket adjacent to the A-strand of E-DIII on the opposing E protein of a dimer. The E-DIII residue that coordinates this interaction is a critical contact for 4E11. Thus, the binding of 4E11 requires displacement of the fusion loop away from the dimer interface.
So how does 4E11 bind the virion? Proteins incorporated into flaviviruses are in constant motion as they sample related conformations at equilibrium. Our understanding of the ensemble of conformations sampled by the virion is incomplete but has been greatly informed by studies with antibodies in complex with virions. 1A1D-2 is a DENV group-reactive mAb that shares 85% amino acid identity with 4E11, binds the same A-strand epitope, and is capable of neutralizing DENV 1, 2, and 3 (Lok et al., 2008). Binding of 1A1D-2 to virus particles is temperature-dependent. Cryo-electron microscopic reconstructions of 1A1D-2 in complex with DENV suggested this antibody binds the virion by trapping a conformation not predicted by the existing structural models of the mature DENV structure (Figure 1) (Mukhopadhyay et al., 2005). Presumably, 1A1D-2 and 4E11 bind and stabilize the A-strand epitope as it becomes accessible on the surface of a “ breathing” virion. In agreement, a recent study suggested that time- and temperature-dependent effects on antibody- mediated neutralization may be a common feature of anti-flavivirus antibodies (Dowd et al., 2011). The contribution of structural dynamics on the serotype- dependent interactions between 4E11 and DENV remains unexplored.
Virion breathing may have significant consequences not only for antibody recognition, but also for receptor binding and virion disassembly. As Cockburn and colleagues (2012) speculate, if the four DENV serotypes bind to the same receptor, they have already identified a potential common binding site on the E protein for such a receptor interaction. Under physiological conditions, the site would be transiently accessible and could be captured by a cellular receptor. The proximity of this site on the A-strand of DIII to the fusion loop suggests that this hydrophobic sequence may become exposed at the cell surface, analogous to what has been proposed for alphaviruses (Meyer and Johnston, 1993). This capture event by a cellular receptor might irreversibly bind the virus to the cell and promote the subsequent fusion and disassembly of the virion. Taken together, these and previous studies suggest that understanding the molecular mechanisms of antibody-virus binding may provide much broader biological insights than simply antibody-mediated neutralization.