Germline VRC01 antibody recognition of a modified clade C HIV-1 envelope trimer and a glycosylated HIV-1 gp120 core

“…2c-f). Specifically, as previously described in structures involving gp120 monomers 22, 23, 29 , germline-encoded R71 HC in the BG24 iGL -GT1 and BG24mat-BG505 structures formed a salt bridge with the conserved D368 gp120 side chain, an Ab interaction that mimics the interaction of host receptor residue R59 CD4 with D368 gp120 (Extended Data Fig. 2c).…”

“…The lineage then affinity matures into two main branches, where one remains dependent on the N276 glycan for binding and the other gaining affinity for the CD4bs epitope in the absence of the N276 glycan. Together with the recent finding that VRC01 inferred GL Ab interacted with the N276-glycan of 426c gp120 core lacking the V1-V2-V3 loops (Borst et al., 2018), these data suggest an alternative vaccine design strategy where the N276-glycan is present on the germline-targeting immunogens at the initial priming stages to select for early intermediates with glycan-binding properties, which could be subsequently boosted with immunogens to affinity mature intermediate precursors to relinquish dependency on the glycan and acquire higher affinity for the CD4bs epitope. A similar strategy was suggested for V2-apex directed Abs upon observation that varying affinities for particular glycoforms were associated with the elicitation of the PCT64 lineage (Landais et al., 2017, Rantalainen et al., 2018).…”

“…A 3.6-Å resolution crystal structure of antibody P-p3b3 bound to the 426c Core and a 3.2-Å resolution crystal structure of antibody P-p1f1 bound to eOD-GT8 were solved (Figure 2; Table S1). These antibodies bind 426c Core and eOD-GT8 with the same angles of approach that human glVRC01 binds 426c WT Core (all atoms root-mean-square deviation [RMSD] = 0.3 Å for gp120-Fv) (Borst et al, 2018) or eOD-GT6 (all atoms RMSD = 0.8 Å for gp120-Fv) (Jardine et al, 2013) (Figure 2A). Critical contacts, both in the HC (Figure 2B) and in the LC (Figure 2C), were maintained in these interactions: Trp50 HC , Asn58 HC , Arg71 HC , and Trp100B HC residues were not mutated in the mouse VRC01-like antibodies, adopted the same orientations, and participated in the same hydrogen bonding as observed with glVRC01 (Figure 2B).…”

“…The mouse VRC01-like antibodies have longer CDRL1 than most human VRC01-class antibodies (17 versus 11–12 aa long based on the Kabat nomenclature); however, several human VRC01-class antibodies, including VRC01c-HuGL2, have been isolated by sorting naive B cells with eOD-GT8 that contain 17-aa-long CDRL1 whose sequence is similar to that of the VRC01-like antibodies identified here (Figure 2C) (Havenar-Daughton et al, 2018; Jardine et al, 2016). Whereas the shorter CDRL1 domain of the human inferred glVRC01 Ab is well ordered in the complex of this antibody with eOD-GT6 (Jardine et al, 2013) or with 426c WT Core (Borst et al, 2018), the CDRL1 of the mouse VRC01-like antibodies bound to 426c Core or eOD-GT8 and that of VRC01c-HuGL2 bound to eOD-GT8 (Jardine et al, 2016) were disordered, indicating extensive CDRL1 flexibility. Such flexibility is likely necessary to accommodate glycans present on N276 in the presence of longer CDRL1.…”