Fig. 3: Reduction of V3 dominance increases selection of trimer-reactive clones.

a Modifications in DANA 1^mod, 3^mod, 5^mod focus on (I) recombinant HIV-1 Env trimer panning targets and (II) type and number of ribosome display selection rounds. V3-depl: depletion of open, V3-exposing trimers from panning trimer preparation through V3 mAb binding); HS: use of highly stabilized DS-SOSIP trimer; PP: pre-panning of DARPin library with V3 peptide to remove V3-reactive DARPins before DANA; WO: DANA without off-rate round that would favor selection of high-affinity binders. b–e Comparison of binding, neutralizing and sequence properties of DARPins with valid ORF (n = 159–163) among 190 randomly picked clones from each DANA. Data are derived from single experiments. b Binding properties of DARPins based on ELISA using Env trimer probes and the V3-crown mimetic peptide V3-IF (BG505) (see Table S4). Trimer binding is categorized as binding to at least one of several trimers but not the V3-mimetic. V3 specificity includes solely V3-reactive and V3- and trimer dual-reactive clones. Low-level binders categorize DARPins with no trimer- and V3-reactivity >3-fold over background. c Neutralization score against a multi-clade 5-pseudovirus panel reflecting breadth and potency. The max. score is 75, a score 5–14 indicates low, a score of >15 high neutralizing activity. The left panel depicts the neutralization score, the right panel the number of clones with low and high neutralizing activity, respectively. d DARPin frequency without mutations (i.e., typical DARPins), with insertions/deletions and with >5% framework mutations. e Distribution of DARPin types.