Solid-state nuclear magnetic resonance measurements of HIV fusion peptide 13CO to lipid 31P proximities support similar partially inserted membrane locations of the α helical and β sheet peptide structures.

Fusion of the human immunodeficiency virus (HIV) membrane and the host cell membrane is an initial step of infection of the host cell. Fusion is catalyzed by gp41, which is an integral membrane protein of HIV. The fusion peptide (FP) is the ∼25 N-terminal residues of gp41 and is a domain of gp41 that plays a key role in fusion catalysis likely through interaction with the host cell membrane. Much of our understanding of the FP domain has been accomplished with studies of "HFP", i.e., a ∼25-residue peptide composed of the FP sequence but lacking the rest of gp41. HFP catalyzes fusion between membrane vesicles and serves as a model system to understand fusion catalysis. HFP binds to membranes and the membrane location of HFP is likely a significant determinant of fusion catalysis perhaps because the consequent membrane perturbation reduces the fusion activation energy. In the present study, many HFPs were synthesized and differed in the residue position that was (13)CO backbone labeled. Samples were then prepared that each contained a singly (13)CO labeled HFP incorporated into membranes that lacked cholesterol. HFP had distinct molecular populations with either α helical or oligomeric β sheet structure. Proximity between the HFP (13)CO nuclei and (31)P nuclei in the membrane headgroups was probed by solid-state NMR (SSNMR) rotational-echo double-resonance (REDOR) measurements. For many samples, there were distinct (13)CO shifts for the α helical and β sheet structures so that the proximities to (31)P nuclei could be determined for each structure. Data from several differently labeled HFPs were then incorporated into a membrane location model for the particular structure. In addition to the (13)CO labeled residue position, the HFPs also differed in sequence and/or chemical structure. "HFPmn" was a linear peptide that contained the 23 N-terminal residues of gp41. "HFPmn_V2E" contained the V2E mutation that for HIV leads to greatly reduced extent of fusion and infection. The present study shows that HFPmn_V2E induces much less vesicle fusion than HFPmn. "HFPtr" contained three strands with HFPmn sequence that were chemically cross-linked near their C-termini. HFPtr mimics the trimeric topology of gp41 and induces much more rapid and extensive vesicle fusion than HFPmn. For HFPmn and HFPtr, well-resolved α and β peaks were observed for A6-, L9-, and L12-labeled samples. For each of these samples, there were similar HFP (13)CO to lipid (31)P proximities in the α and β structures, which evidenced comparable membrane locations of the HFP in either structure including insertion into a single membrane leaflet. The data were also consistent with deeper insertion of HFPtr relative to HFPmn in both the α and β structures. The results supported a strong correlation between the membrane insertion depth of the HFP and its fusogenicity. More generally, the results supported membrane location of the HFP as an important determinant of its fusogenicity. The deep insertion of HFPtr in both the α and β structures provides the most relevant membrane location of the FP for HIV gp41-catalyzed membrane fusion because HIV gp41 is natively trimeric. Well-resolved α and β signals were observed in the HFPmn_V2E samples with L9- and L12- but not A6-labeling. The α signals were much more dominant for L9- and L12-labeled HFPmn_V2E than the corresponding HFPmn or HFPtr. The structural model for the less fusogenic HFPmn_V2E includes a shorter helix and less membrane insertion than either HFPmn or HFPtr. This greater helical population and different helical structure and membrane location could result in less membrane perturbation and lower fusogenicity of HFPmn_V2E and suggest that the β sheet fusion peptide is the most functionally relevant structure of HFPmn, HFPtr, and gp41.