As the sole target of HIV-1 neutralizing antibodies (NAbs), the Env protein covers much of its surface with carbohydrate to limit the host response to the regions of Env that are incapable of controlling viral replication (e.g. hyper-variable loops). While the glycan shield is a well-adapted immune-escaping strategy among viruses, the HIV-1 Env, in particular, belongs to the most heavily glycosylated proteins known. Indeed, only 3% of Env surface is immunoglobulin-accessible.
To account for the extensive sequence diversity of Env, NAbs need to target conserved protein surface. However, about two-thirds of N-linked glycosylation sites are encoded in the conserved regions of Env. We observed that the glycosylation sites in the outer domains of Env are highly variable; in some cases, 40% of these sites may be missing due to natural sequence variation.
We were able to group the missing glycosylation sites (glycan holes)and underlying protein structure into Carbohydrate-Occluded Neutralization Epitopes (CONEs). Rabbit immunization with native-like Env trimer (BG505 SOSIP) induced NAbs with immunogenic epitopes at glycan holes (position 241 and 289). Hence, we envisioned the development of individual CONE-immunogens to generate Env-specific antibodies and direct immune responses to conserved Env-surface.
To this end, we developed a versatile computational protein design platform and grafted the CONE epitopes on heterologous protein scaffolds with preserved structural properties. We demonstrated high accuracy structural modeling with our method by comparing the crystal structures with the designed models. We designed a panel of immunogens engrafting four different epitopes (CONEs 1, 2, 4, and 5) with subtype C sequences and performed immunogenic evaluation in animal models. The in vivo efficacy of antigen presentation was evaluated using two methods (adjuvanted subcutaneous immunization and self-assembling nanoparticles). Antibodies raised by rabbit immunization exhibit specificity and affinity toward SOSIP Env protein and neutralization effects toward HIV-1 subtype C, which accounts for 50% of new infections worldwide.
Our immunization strategy was designed so that we could compare immunogens of different sequence conservation and different structural stabilities, providing valuable insights for vaccine design: (a) Immunization with nanoparticles may elicit antisera with higher specificity; (b) CONE immunogen of metastable protein scaffolds may induce antibodies with elevated non-specificity (CONE 5 vs. CONE 1/4); and (c) Sequence conservation is an important criterion of immunogenicity of CONEs (CONE 2 vs. CONE 1/4).
The transmitted HIV-1 is a heterogeneous population of viruses because different copies of Env are missing glycans at different CONEs. Hence, our strategy has the potential when used combinatorically (i.e., a cocktail of antibodies) to have a broad neutralization effect. CONE immunogens, when immobilized on a column, can also be applied to screen polyclonal human sera for reactivity to CONEs. The existence of such anti-CONE antibodies would provide valuable insights to the relationship between the presence of a glycosylation site and its impact on viral fitness. We expect the combination of two fields (rational protein design and vaccinology) will provide tailor-made protein tools to halt the persistent global threat imposed by HIV-1 virus.
Authors: Cheng Zhu and Nikolay V. Dokholyan