Friday, 25 May 2012

Lecture material: Structure of the dengue virus envelope protein after membrane fusion


  • Membrane fusion is the central molecular event during the entry of enveloped viruses into cells. 
  • Viral surface proteins facilitate bilayer fusion
  • They are triggered by viral interaction with target cell.
  • The best-studied example is the influenza virus haemagglutinin (HA),
  • HA is synthesized as a single-chain precursor
  • It is then cleaved into two chains, known as HA1and HA2, during transport of the trimeric glycoprotein to the cell surface. 
  • The binding of HA1 to a cell-surface receptor leads to endocytic uptake
  • acidification of the endosome triggers dramatic conformational rearrangement of HA2
  • The latter is a two-stage process.
  •  Exposure of the aminoterminal ‘fusion peptide’ of HA2
  •  allows it to insert into the endosomal membrane
  • Whole HA2 pp chain subsequently folds over
  • This brings together its N and C termini  forces the target-cell membrane (held by the fusion peptide) and theviral membrane (held by the C-terminal transmembrane anchor ofHA) against each other.
  • HA is the prototype of a large class of viral fusion proteins—for example, those of other myxo- and paramyxoviruses such as measles virus, retroviruses such as HIV, and filoviruses such as Ebola virus
  • All of these ‘class I’ viral fusion proteins are two-chain products of a cleaved, single-chain precursor, 
  • All bear a hydrophobic fusion peptide at or near the N terminus created by the cleavage 
  • in all class I fusion proteins, a three-chain, a-helical, coiled-coil assembles during the conformational change
  • this drives the fusion peptide towards the target-cell membrane
  • This creates central structural element of the fusion machinery.
  • An architecturally and evolutionarily distinct class of fusion proteins is found on flaviviruses, such as yellow fever, West Nile,and dengue viruses, and on alphaviruses, such as Semliki Forest and Sindbis viruses.
  • These proteins associate with a second, ‘protector’ protein, 
  • the cleavage of protector protein primes the fusion protein to respond to acidic pH.
  • Structures have been determined for the ectodomains of three class II proteins in their prefusion state
  • Those of two flaviviruses, tick-borne encephalitis (TBE) and dengue viruses, are dimeric, both in solution and on the viral membrane surface
  • They have three domains, with folds based largely on b-sheets. 
  • One of these (domain II), an elongated, finger-likestructure, bears a loop at its tip with a hydrophobic sequence conserved among all flaviviruses. 
  • Experiments with TBE virus show that this ‘cd loop’ (residues 98–109 in dengue type 2) is responsible for attachment of soluble E ectodomains to target membranes 
  • The hydrophobic residues are essential for its activity
  • cd loop of class II fusion proteins has a function analogous to that of the N-terminal fusion peptide in class I fusion proteins: 
  1. insertion into the host-cell membrane
  2. provision of an attachment point for drawing host-cell and viral membranes together
  • We refer to the cd loop as the ‘fusion loop’, reserving ‘fusion peptide’ for the N-terminal segment of class I fusion proteins
  • In dengue virus type 2 protein (sE) in its trimeric, postfusion conformation,
  • The fusion loops of the three subunits come together to form a membrane-insertable,‘aromatic anchor’ at the tip of the trimer. 
  • The fusion loop retains its prefusion conformation. 
  • Neighbouring hydrophilic groups restrictinsertion to the proximal part of the outer lipid-bilayer leaflet.
  • The entire ectodomain of the protein folds back on itself,
  • This directs the C-terminal, viral membrane anchor towards the fusion loop
Membrane insertion and trimer formation
  • Dimer formed by dengue sE dissociates reversibly.
  • At acidic pH, dissociation is essentially complete at protein concentrations of 1 mg ml
  • at neutral pH, the dissociation constant is one to two orders of magnitude smaller. 
  • The fusion loop at the tip of domain II would be exposed in the monomer
  • but exposure does not cause nonspecific aggregation of the protein
  • experiments show that the fusion loop of monomeric TBE sE allows association with lipid membranes and that this membrane association catalyses irreversible formation of sE trimers at low pH 
  • on acidification, sE dimers dissociate, bind liposomes and trimerize 
  • The trimers are tapered rods, about 70–80 A˚long and 30–50 A˚in diameter, with the long axis perpendicular to the membrane and their wide end distal to it. 
  • They tend to cluster on the liposome surface, often forming a continuous layer. 
  • These heavily decorated areas appear to have a greater than average membrane curvature, resulting in smaller vesicles
  • This observation suggests that E trimers can induce curvature, a property that may help promote fusion
  • The dengue sE trimers can be solubilized with the detergent n-octyl-b-D-glucoside (b-OG); they remain trimeric at all pH values between 5 and 9, as determined by gel filtration chromatography 
The fusion loop
  •  the fusion loop is stable when fully exposed. 
  • It thus appears that the fusion loop retains essentially the same conformation, whether buried against another subunit, inserted into a lipid membrane, or exposed to aqueous solvent.
  • In the trimer, the three hydrophobic residues in the fusion loop conserved among all flaviviruses—Trp 101, Leu 107 and Phe 108—are fully exposed on the molecular surface, near the three-fold axis.
  • They form a bowl-like concavity at the trimer tip, with a hydrophobic rim 
  • Tryptophans tend to appear in membrane proteins at the interface between the hydrocarbon and head-group layers of the lipid
  • if the indole amine participates in a hydrogen bond, as is the case for Trp 101, the side chain may be completely buried in the hydrocarbon layer.
  • We therefore propose that the E trimers penetrate about 6 A˚ into the hydrocarbon layer of the target membrane. 
  • They cannot penetrate further, because of exposed carbonyls and charged residues on the outside rim of the fusion-loop bowl
  • the fusion loop is held in the membrane mainly by an ‘aromatic anchor’ formed by Trp 101 and Phe 108. 
  • The bowl is lined by the hydrophobic side chains of Leu 107 and Phe 108, so that it cannot accommodate lipid headgroups
  • fatty-acid chains from the inner leaflet of the membrane may extend across to contact the base of the fusion-loop bowl, or that fatty-acid chains from the outer leaflet may bend over to fill it. 
  • In either case, insertion will produce a distortion in the bilayer
A postfusion conformation
  • Domain III folds back
  • B-strands rearrange at trimer interface
  • This projects C terminus if sE towards fusion loop
  • Position fusion loop at entrance of a channel which extends towards fusion loops
  • Stem connecting end of sE fragment with viral TM anchor could span length of channel
  • By binding channel, stem would contribute more trimer contacts with domain II of another subunit. 
  • Proposed stem conformation puts viral TM domain in vicinity of fusion loop
  • Like postfusion conformation of class I viral fusion proteins.
Propsed mechanism for fusion mediated by class II viral fusion proteins. Full-length E is represented as in Fig. 1c, with the stem and viral transmembrane anchor in cyan. a, E binds to a receptor on the cell surface and the virion is internalized to an endosome. b, Reduced pH in the endosome causes domain II to hinge outward from the virion surface, exposing the fusion loop, and allowing E monomers to rearrange laterally in the plane of the viral membrane. c, The fusion loop inserts into the hydrocarbon layer of the host-cell membrane, promoting trimer formation. d, Formation of trimer contacts spreads from the fusion loop at the tip of the trimer, to the base of the trimer. Domain III shifts and rotates to create trimer contacts, causing the C-terminal portion of E to fold back towards the fusion loop. Energy release by this refolding bends the apposed membranes. e, Creation of additional trimer contacts between the stem-anchor and domain II leads first to hemifusion and then (f) to formation of a lipidic fusion pore.

Mechanism of membrane fusion
  • The structure described here, combined with previous knowledge, allows us to propose the following mechanism for how conformational changes in the flavivirus E protein promote membrane fusion.
  1. E associates with a CS R, through domain III probably.there is evidence for glycan-mediated interactions as well. Receptor binding leads to endosomal uptake.
  2. Reduced pH in endosome causes E dimers on virion surface to dissociate, exposing their fusion loops and letting domains I and II to flex relative to one another. Mutations alter pH threshold of fusion. Orientation for pre and postfusion structures are different. This indicates a pH-dependent hinge at domain I-domain II interface. Release of constrains by dimer contacts may let stem extend away from mebrane. Domain II turns outwards, away from virion surface, to insert its fusion loop into target cell membrane.
  3. Domain II outwardly projects. This destroys tight packing interactions on virion outer surface, allowing lateral rearrangement of E monomers. Target membranes probably catalyse trimerisation, leading to a prefusion intermediate. in which the trimer bridges host-cell and viral membranes. Its fusion loops bind  the former and TM tail anchors in the latter. 
  4. Trimer contacts form from fusion loops at trimer tip to domain I at the base. Domain II shifts and rotates, folding C terminus of sE back towards fusion loop.  Free energy released by this refolding can drive the two membranes to bend towards each other. Fusion loop insertion induces positive bilayer curvature. This may stabilise lateral surfaces of protrusions. 
  5. A hemifusion stalk forms. Its proximal leaflets are fused and distal leaflets are unfused. It is thought to be an essential intermediate.  Hemifusion stalks can ‘flicker’ open into narrow fusion pores.  Migration of the transmembrane segments along a transient pore will prevent its closing.  If TM segments of adj stem regions snap into place round tips of domains II, formation of symmettrical final structure could drive transition from stalk to pore. 
  6. Trimer reaches conformation, with stems docked along surface of domains II. Fusion loops and TM anchors are next to each other in the fused membrane.
Comparison with class I fusion
  • Class I and II have some common features.
  • During fusion transitions, the protein folds back so its 2 membrane attachment points come together in postfusion structure.
  • Class I proteins fold back by zippering up an outer layer around a central trimeric coiled coil.
  • For trimeric dengue sE, class II proteins fold back by nucleating trimer formation around an elongated fingerlike fusion domain, rearranging 2 other domains, and probably zippering an extended C-terminal stem along trimer surface.
  • Class II viral fusion proteins form trimers from monomers.
  • Class I proteins are trimeric in prefusion state.
  • In influenza HA and trimerisation of dengue E, the important trimer interactions in final state form during transition
  •  postulated prefusion intermediate is, both for class I fusion proteins and now for class II, a structure in which these central trimer contacts have formed but the zippering-up of the outer layer has not yet begun

No comments:

Post a Comment