Kinetically coupled folding of a single HIV-1 glycoprotein 41 complex in viral membrane fusion and inhibition

HIV-1 glycoprotein 41 (gp41) mediates viral entry into host cells by coupling its folding energy to membrane fusion. Gp41 folding is blocked by fusion inhibitors, including the commercial drug T20, to treat HIV/AIDS. However, gp41 folding intermediates, energy, and kinetics are poorly understood. Here, we identified the folding intermediates of a single gp41 trimer-of-hairpins and measured their associated energy and kinetics using high-resolution optical tweezers. We found that folding of gp41 hairpins was energetically independent but kinetically coupled: Each hairpin contributed a folding energy of ∼−23 k_BT, but folding of one hairpin successively accelerated the folding rate of the next one by ∼20-fold. Membrane-mimicking micelles slowed down gp41 folding and reduced the stability of the six-helix bundle. However, the stability was restored by cooperative folding of the membrane-proximal external region. Surprisingly, T20 strongly inhibited gp41 folding by actively displacing the C-terminal hairpin strand in a force-dependent manner. The inhibition was abolished by a T20-resistant gp41 mutation. The energetics and kinetics of gp41 folding established by us provides a basis to understand viral membrane fusion, infection, and therapeutic intervention.The HIV-1 glycoprotein 41 (gp41) homotrimer constitutes the transmembrane stem of the envelope glycoprotein complex (Env) and plays key roles in viral entry, the first step of viral infection (1–3) (Fig. 1A). Comprising three gp120/gp41 heterodimers, the Env complex recognizes receptors on target cells and primes gp41 for membrane fusion (4, 5). Gp41 is initially held in a largely unfolded state and shielded by gp120 in the metastable Env complex like a loaded spring (1). During membrane fusion, gp41 inserts its fusion peptide into the host cell membrane with the help of gp120, forming an extended gp41 prehairpin conformation (6, 7) (Fig. 1B). Then the extended gp41 folds back, drawing the two membranes into proximity. In this process, gp41 uses its folding energy to lower the energy barrier of membrane fusion and thus increases the rate of fusion. After fusion, gp41 forms an extraordinarily stable trimer-of-hairpins in a six-helix bundle (6HB) conformation, in which three N-terminal heptad repeats (NHRs) form a central coiled coil and three C-terminal heptad repeats (CHRs) bind in the grooves of the NHR coiled coil in an antiparallel fashion (8, 9) (Fig. 1 B and C). The 6HB further extends to the membrane-proximal external region (MPER) and its complementary fusion peptide-proximal region (FPPR) (10). Although significant progress has been made by defining gp41 structures in the initial and final stages of membrane fusion, how gp41 transits between the two structures remains unclear. Addressing this question requires better determination of the intermediates, energy, and kinetics associated with gp41 folding.Open in a separate windowFig. 1.Sequences, conformations, and constructs of gp41 and the experimental setup. (A) Domain structure of HIV-1 gp41. The three truncated gp41 sequences used in this study, the sequence of T20, and T20-resistant mutation (N554D) are indicated. CHR, C-terminal heptad repeat; CT, cytoplasmic tail; FP, fusion peptide; FPPR, fusion peptide-proximal region; MPER, membrane-proximal external region; NHR, N-terminal heptad repeat; TMD, transmembrane domain. (B) Diagram showing gp41 folding and viral membrane fusion. In the prehairpin state, three gp41 monomers form a stable three-helix bundle in their NHRs with largely unfolded CHRs. We refer to the conformational transition from this prehairpin state to the six-bundle (6HB) state as gp41 folding, in contrast to de novo formation of the 6HB from free gp41 monomers as measured in previous ensemble experiments (18). (C) Experimental setup to pull a single gp41 complex using dual-trap optical tweezers. (D) Topology of the construct used to pull the third hairpin in the gp41 complex. Two constructs with different loop sizes were made, gp41-ll with a 19-aa loop and gp41-sl with a 5-aa loop. (E) Topology of the construct to specifically pull the second gp41 hairpin. The construct contained a cleavage site of the TEV protease (indicated by x). After TEV cleavage and purification, the C-terminal pulling site of gp41-ll-hp2 was generated.Knowledge of gp41 folding is crucial to understanding the mechanisms of viral membrane fusion and infection as well as pharmaceutical intervention in AIDS (11, 12). Membrane fusion needs to overcome an energy barrier over ∼50 k_BT (13). A commensurate energy is demanded from folding of one or multiple gp41 complexes to catalyze viral membrane fusion. However, a large gp41 folding energy may also destabilize the metastable Env complex (6, 14). Mutations that decrease or increase gp41 stability in turn stabilize or destabilize the Env complex (15). Thus, the folding energy of gp41 seems to be optimized to balance the rate of membrane fusion and the stability of the Env complex, thereby achieving highest infectivity. However, the folding energy of a single gp41 trimer has not been accurately measured owing to irreversibility and protein aggregation associated with gp41 folding in traditional ensemble-based experiments (16).Once attached to the host cell, an HIV virion requires at least 15 min to fuse (6, 17). Several gp120-dependent gp41 folding intermediates have been implicated in the fusion process (6, 7). After gp120 is shed, folding intermediates of the gp41 complex alone have not been reported (18). However, intermediates may appear due to independent or cooperative folding of three gp41 monomers. All these intermediates, regardless of their dependence on gp120, are important, because they are essential for membrane fusion and primary targets of anti-HIV drugs, including various fusion inhibitors and broadly neutralizing antibodies (11, 12). In contrast, the conserved gp41 and gp120 regions are protected by gp120 glycan shield on the surface of the Env complex, and thus are not accessible for most drugs and antibodies (1, 2). T20 (enfuvirtide) is the first commercial antiretroviral fusion inhibitor (19, 20). It is a 36-aa synthetic polypeptide that shares its sequence with regions of CHR and MPER in gp41 (Fig. 1A). Thus, T20 competes with these regions to bind their cognate sequences on gp41 to inhibit gp41 folding (Fig. 1B). T20 tightly associates with the five-helix bundle (5HB) missing one CHR with a dissociation constant around 30 nM and efficiently inhibits HIV-1 viral infection (21). However, the efficacy of T20 is compromised by certain gp41 mutations. Gp41 mutation N554D (or N43D based on gp41 amino acid numbering) is a common T20-resistant mutation identified from AIDS patients treated with T20 (22). How T20 efficiently inhibits folding of the wild-type, but not the drug-resistant, gp41 is not well understood.Gp41 folding is profoundly affected by membranes. Several studies showed that gp41 fails to fold into the 6HB conformation in the presence of membranes or detergent micelles but remains in largely unfolded monomer form (23, 24). However, other studies suggested that gp41 does form the 6HB structure that can even be strengthened by SDS (25). These different reported affinities between gp41 monomers and membranes or membrane mimics inversely correlate with the stability of the 6HB and lead to different models of membrane fusion. In principle, proteins that mediate membrane fusion, including viral fusion proteins and SNARE (soluble N-ethylmaleimide-sensitive factor attachment receptor) proteins, may promote membrane fusion either mechanically or chemically. In the mechanical model, fusion proteins, such as SNARE proteins, transduce their folding energy to mechanical force to draw two membranes into proximity for fusion (26). A single neuronal SNARE complex folds/assembles in multiple stages and outputs a total energy of −65 k_BT (27), or −41 kcal/mol. In the chemical model, fusion proteins or chemical agents such as polyethylene glycol alter membrane properties (for example, by dehydration of membranes) to enhance membrane merging without large energy output from folding of fusion proteins, much like strewing salt on snow. However, these two extreme models may not be exclusive. It will be interesting to see whether gp41 fuses membranes similarly to SNARE proteins or by a different mechanism.Finally, HIV-1 gp41 represents a family of structurally and mechanistically conserved class-I viral fusion proteins, including the glycoprotein GP2 of Ebola virus and hemagglutinin of the influenza virus (1, 4). To our knowledge, none of these proteins has been well studied in terms of their folding energy and kinetics. Therefore, detailed measurements of gp41 folding energy and kinetics will shed light on the working mechanism of these viral fusion proteins and help to develop antiretroviral fusion inhibitors against diseases caused by these viruses (11, 20).We investigated folding of a single gp41 complex using high-resolution optical tweezers (28). Our single-molecule manipulation approach used mechanical force to unfold the gp41 complex under a physiological solution condition, to probe its folding energy and kinetics, and to mimic the membrane repulsive force during functional gp41 folding. In addition, our method overcame ensemble averaging and synchronization required by traditional protein folding studies that often obscure folding intermediates (27, 29). We found that each gp41 hairpin in the complex folded in a two-state manner and kinetically coupled with other hairpins. The total folding energy of a single gp41 complex was estimated to be −71 k_BT. We discovered that T20 inhibited gp41 folding by a novel force-dependent strand-displacement mechanism. Although detergent micelles destabilize the gp41 complex, they do not abolish the complex formation.