Figure 1. Crystal structure of gp41FP-TM in complex with 2H10.

(A) Schematic drawing of gp41 and expression constructs of gp41 chains N and C. Sequence numbering is based on the HIV-1-HBX2 envelope gp160 sequence. Color coding is as follows: FP, fusion peptide, red; FPPR, fusion peptide proximal region, orange; HR1, heptad repeat region 1, yellow; HR2, heptad repeat region 2, blue; MPER, membrane proximal external region, violet; TM, transmembrane region, beige; CC, cys loop region, light blue and cyt, cytoplasmic domain brown. Expression tags used are TrxA, thioredoxin fusion protein, His, His-tag, TEV, TEV protease cleavage sequence, Tag, chain N contains a Flag-tag (DYKDDDDK sequence) and chain C an N-terminal enterokinase cleavage site (DDDDK). (B) Ribbon presentation of gp41TM-FP in complex with 2H10. Color-coding of the different segments is as indicated in the gp41 scheme (A), the 2H10 nanobody is colored in green. (C) Ribbon presentation of gp41TM-FP including the core six-helical bundle trimer axis (black line) revealing the different orientations of FP and TM. (D) Close-up of the interaction of gp41FP-TM with 2H10. Residues in close enough contact to make polar interactions are shown as sticks. (E, F, G) Ribbon diagram of the individual protomers named chain A, B, and C. Residues within the FPPR and MPER hinge regions are indicated.

Figure 1—figure supplement 1. Characterization of gp41 containing FP and TM.

(A) Size exclusion chromatography of the gp41FP-TM complex composed of chains N and C and SDS-PAGE showing the two bands corresponding to gp41 chains N and C. (B) SEC of gp41FP-TM in complex with the llama nanobody 2H10 and corresponding SDS PAGE showing the three bands corresponding to gp41 chains N and C and 2H10.

Figure 1—figure supplement 2. Biophysical characterization of gp41FP-TM and MPER Ab interaction.

(A) Circular dichroism of gp41FP-TM shows that FP and TM increase the melting temperature of gp41. Temperature-dependent unfolding of gp41FP-TM monitored by circular dichroism spectroscopy recorded at 222 nm in a buffer containing 1% β-OG. Gp41FP-TM has an estimated Tm of ~93°C. (B) Gp41FP-TM complex formation with 2H10. ITC data were recorded on successive injections of 2H10 at a concentration of 267 µM into the cell containing gp41FP-TM at a concentration of 19,5 µM. Three experiments were performed, with an average stoichiometry N = 1.1 +/- 0.2, which suggests that on average only one 2H10 binds to trimeric gp41FP-TM under these conditions. The calculated KD is 2.1 µM ± 0.9. (C) Gp41FP-TM complex formation with 2H10. Bio-layer interferometry (BLI) binding of gp41FP-TM to 2H10. GP41FP-TM concentrations analyzed are dilutions between 156 and 2500 nM. The estimated KD based on the steady state binding model is 170 ± 17 nM. Note that the calculated Kds of the ITC and BLI experiments are only estimates since 2H10 needs to optimally stabilize the gp41FP-TM-binding conformation. Therefore only a fraction of gp41FP-TM may adopt the required conformation during the injection time used to record binding.

Figure 1—figure supplement 3. Close-ups of the model and its corresponding electron density.

2Fo-Fc composite omit maps contoured at 1 σ of a central 6HB core region (A and B) of the kinked MPER conformation of protomer C.

Figure 1—figure supplement 4. Comparison of the gp41FP-TM structure with gp41 core structures.

Ribbon presentation of gp41-MPER (pdb 3k9a), gp41FPPR-MPER (pdb 2x7r) and gp41FP-TM (numbering is shown for chain B). Cα super positioning of all three structures onto chains N-B (residues 546–574) and C-B (residues 628–662) of gp41FP-TM, revealing an r.m.s.d of 0.55 Å between pdb 3k9a and gp41FP-TM and an r.m.s.d. of 0.29 Å between pdb 2x7r and gp41FP-TM for the straight helices of chain B.

Figure 1—figure supplement 5. Positioning of gp41FP-TM-2H10 in a bilayer by MD simulation.

(A) Model of gp41FP-TM-2H10 before simulation and (B) after 1 µs simulation, which repositions the 2H10 CDR3 in the membrane and reveals movement of FP of chain C. The orange spheres represent the phosphate atoms of the phospholipids and mark the membrane boundaries. Residues 512–517 and 701–707 have been modeled in a helical conformation to provide complete models of FP and TM. (C) Close-up of the proposed membrane interaction of the 2H10 interface. CDR3 W100 and S100d mutated to F could as well insert into the membrane. Furthermore, basic residues at positions S30K, S27R, and S74R (shown as sticks) are positioned to make polar interactions with lipid head groups.

Figure 1—figure supplement 6. Crystal lattice packing.

Crystal packing of protein 2-D layers arranged in the c direction of the crystal unit cell do not show defined crystal contacts. The inset shows the distances between the protein layers indicating that the defined C-terminus of chain C is close (14 Å) to the N-terminus of an HR1 helix (yellow). Thus crystal lattice stabilization is likely unregular and poorly defined at the resolution of the crystal diffraction data. We hypothesize that weak crystal contacts are formed by the C-terminal extensions of TM, which may be able to adopt different orientations and are therefore not present in the structure.

Figure 1—figure supplement 7. Membrane interaction of 2H10.

Membrane interaction of nanobodies 2H10, 2H10-RKRF, and bnAb 10E8 was tested using liposomes containing the lipid composition of the HIV-1 envelope. Nanobodies 2H10 and 2H10-RKRF as well as bnAb10E8 were incubated alone and with liposomes and subsequently separated in a sucrose gradient flotation assay. Samples of each fraction of the gradient were analyzed by SDS-PAGE, which demonstrates that 2H10 and 2H10-RKRF do not float with liposomes in this assay indicating no or very low non-specific membrane interaction as reported previously for 2H10 (Lutje Hulsik et al., 2013) and some weak interaction (*) for 10E8 as reported previously (Chen et al., 2014). The figure supplement is also related to data presented in Table 2.