Virus Spikes - How Viruses Access Cells

Adenovirus

Above: many viruses, both non-enveloped (naked) like the adenovirus shown above, and enveloped viruses (possessing a lipid bilayer membrane or envelope) like influenza, coronavirus and HIV possess spikes, made of protein (or glycoprotein). What is the function of these virus spikes and why are they medically important?

In this article we look at how viruses gain access to the cells they infect and how antibodies, and hence vaccines, can sometimes prevent this.

Adenovirus

Adenovirus causes sore throats and eye infections (conjunctivitis) in humans and is one of the agents of the common cold, bronchitis and peneumonia and has been implicated in obesity. Additionally, some adenoviruses cause some bladder infections and gastroenteritis and other infections. There are a large number of serotypes in humans, with different serotypes causing a different spectrum of disease. There are a number of adenovirus species with some infecting mammals, some birds and some amphibians.

The T = 25 capsid consists of hexon protein subunits (arranged in complexes of 3) making up the 20 equilateral triangular faces and penton protein subunits (in complexes of 5) making up the 12 vertices. Each of the 12 pentons has a fiber (spike) attached non-covalently to it and each fiber is a trimer of three fiber protein subunits. Each fiber has a globular head. Each facet of the capsid is made up of 12 trimers of the hexon protein subunit (36 hexon protein subunits per face or 720 in total; with 3 in each hexon we have 240 hexons in total).

Viruses are 'pirates of the cell' as they take control of an infected cell and use its machinery and resources to manufacture more of themselves. However, they can only infect certain specific cell types as the virus needs a specific 'key' to gain access. This key is a specific adhesin (adhesive molecule) or ligand that binds to a specific target molecule or receptor on the surface of the target cell. Often adhesins work in pairs: the first adhesin is efficient at grappling the target cell, to prevent the virus from being swept past. The spikes of viruses are well positioned for this task since they protrude from the surface and can engage target receptors more readily. The globular head of the fibers in adenovirus is the primary adhesin and binds to the CAR receptor on the surface of target cells. (CAR = Coxsackievirus and Adenovirus Receptor).

To gain access to the target cell the adenovirus must then use its 'key' which is the second adhesin. This adhesin resides in the 5 protein subunits of each penton base which have a surface domain that binds integrin coreceptors on the target cell to initiate uptake of the virus by the unsuspecting host. (A coreceptor is a secondary receptor that binds the secondary adhesin). This domain is the integrin-binding domain or RGD loop that contains a group of three amino acid residues, RGD (arginine-glycine-aspartate or Arg-Gly-Asp) which bind αvβ3 or αvβ5 integrins.

A predicted structure of a penton subunit is illustrated, as a ribbon diagram, below:

Adenovirus penton subunit

This is the penton subunit of human adenovirus 2 (Modeled in Phyre2 using the FASTA sequence provided by Borcherding, F. 2001, from NCBI, GenBank: CAC67483.1). The subunit consists of 571 amino acid residues.


Recall that 5 such pentons form the penton base at each vertex to which the fiber trimer is attached. The integrin-binding domain is uppermost, with its RGD sequence. (A domain is part of a protein which performs a specific function). There is also a hypervariable loop of unknown function.

The base of the penton subunit contains a barrel-like domain made up 8 strands of beta-pleated sheet. A β-pleated sheet is a region of a protein formed of strands of the polypeptide chain aligned parallel or antiparallel to one-another to form a corrugated sheet, which may be further folded. In this case the sheet is folded around to form a barrel. the strands align since they hydrogen-bond to one-another, forming a  strong structure. This is a type of β-barrel. Beta-barrels are often used in membrane-spanning protein pores, in this case it is a structural feature, making up part of the main capsid shell. In this case the strands are antiparallel and the barrel is formed from a type of beta-pleated sheet called a jellyroll (jelly roll):

Jellyroll motif

The jellyroll consists of 8 antiparallel beta-strands, as can be seen, connected by loops. This motif consists of two Greek key motifs. (A motif is a region of a protein with a particular structure). A polypeptide chain has chemically distinguishable ends: one end called the N-terminus and the other called the C-terminus and by convention the sequence begins at the N-terminus, with the first amino acid residue, and proceeds to the C-terminus. In these ribbon diagrams the arrows indicate beta-strands and point from N to C.

A hypervariable region (HVR) of a viral protein is a region where the sequence of amino acids varies considerably from one type to another. These regions are highly variable between different strains or serotypes of the same virus species. In contrast highly conserved regions are parts of the sequence where the combination of amino acids is more critical to function and so have less room to mutate without losing function and so only evolve slowly.

Sequence variability sometimes means a region has no specific function and so its sequence is free to mutate and evolve; such sections of proteins are often structural and perform their function as long as the amino acid changes do not significantly disrupt their shape. In pathogens, however, HVRs are often important for pathogenesis, even if they are not critical to pathogen replication. For example, they may surround conserved regions and a large antibody binding to them may still block the function of the conserved region (e.g. by 'steric hindrance' that is physically getting in the way so an adhesin, for example, can not bind to its receptor when an antibody is attached to a surrounding region).

The specific adhesin, the RGD sequence we would expect to be conserved since a change in this sequence could prevent binding to integrins (though may open up binding to a different receptor and possibly a different cell type). Viruses are of course targets for host defenses and hypervariable regions frequently provide protection by reducing binding of host proteins. For example, an antibody that recognizes and binds to one specific sequence of amino acids may cease binding if that sequence changes sufficiently. In the penton protein the hypervariable loop could be a target for antibodies whose binding may prevent the RGD loop engaging with integrins, blocking infection. Thus this loop may serve to protect the RGD sequence from attack.

A predicted structure of the human adenovirus 2 fiber (spike protein) is shown below:

Adenovirus fiber

This model (582 amino acids) was produced from the human Adenovirus serotype 2 (Ad2) virus, from the sequence on NCBI (AP_000190.1) provided by: Davison, A.J., Benko, M. and Harrach, B. 2003. Genetic content and evolution of adenoviruses. J Gen Virol 84 (PT 11): 2895-2908. The model was produced in Phyre2. The head domain was modeled with 100% confidence but the fiber shaft is a flexible structure and so its exact morphology is uncertain but very similar to the crystal structure (see review by: Nicklin et al. 2005. The Influence of Adenovirus Fiber Structure and Functionon Vector Development for Gene Therapy. Molecular Therapy 12(3) 384-393. doi:10.1016/j.ymthe.2005.05.008).

Note that the fiber shaft contains pairs of beta-strands arranged in a β-helix whilst the terminal knob or head consists in large part of a β-barrel and surface loops.

Remember that three such monomers make up a complete spike as shown below:

Adenovirus fiber - ribbon model

Above: ribbon-model of human Adenovirus 2 spike (fiber) trimer; below surface model. The trimer arrangement was predicted using SymmDoc.The three feet of the spike (fiber) bind to the penton at the capsid vertex. The penton consists of a pentagonal array of 5 penton protein subunits and each fiber tail can bind to the space between neighbouring penton subunits. This means we have a symmetry mismatch as only 3 of the 5 available binding sites are used by the fiber.

Adenovirus fiber - surface

Notice that the spike shown here is considerably shorter than in our model at the top of the page. Spike length is very variable in adenovirus and Ad2 is a short-spike variant. Longer spikes facilitate binding to CAR, but some adenoviruses have other 'keys' to access the cell by other mechanisms. As well as causing pharyngitis (and pneumonia in children) Ad2 can infect liver cells (along with serotypes 1 and 5). Adenovirus is thought to gain access to liver cells by the hexon proteins acting as ligands for blood-clotting factor X, triggering their uptake by liver cells, so the length of the spikes may become less important. Each hexon is made of 3 subunits, where each subunit is a bit like a 'double' protein with two 8-stranded jellyroll beta-barrels. These 3 subunits form a pseudohexagonal hexon, in which the arrangement is rather like a 6-petalled flower with an hexagonal contour.

The fiber shaft also varies in flexibility between serotypes. The beta-strands of the β-helix give some rigidity, but there are some flexible hinges between them, one of them towards the base of the spike below the third beta-twist (the third pair of beta-strands in the fiber shaft). Flexibility has been shown to make it easier for the head domain to reach and bind the CAR receptor. Once the penton base binds integrin, the fibers are shed.

Below,the fiber head as seen from above, as a ribbon model and as a surface-filling model.

Adenovirus fiber head - ribbon view

Adenovirus fiber head - surface

Antibodies can block adhesion

Antibodies are host proteins that recognize specific foreign molecules, usually a short amino acid sequence in a foreign protein, and bind tightly to it. Antibodies are large proteins and their binding will interfere with the normal functioning of the target protein. In adenovirus, antibodies that can successfully neutralize virus function can be produced to the penton base, the fiber, especially the globular head, and the hexon. Clearly, binding of antibodies to the fiber head may prevent the head binding to CAR; similarly antibodies bound to the penton base may prevent it binding integrins. The virus 'keys' will no longer fit into their locks and the virus particle will be less able to infect a potential host cell.

How does influenza infect cells?

A well-studied example of antibody binding to virus spikes is influenza. The structure of the influenza virion is illustrated below:


Influenza

The haemagluttinin (HA or H) spikes are trimers of three subunits, each of which has a binding site (receptor site) that recognizes the receptor sialic acid, a carbohydrate found on the surface of epithelial cells in the respiratory tract. The structure of HA is illustrated below:

Influenza HA

Influenza HA

HA is antigenic (a target for antibodies) and subject to mutation (antigenic drift) and that variant H1 has undergone several mutations since 1968, especially at regions bordering the receptor site. This is likely an adaptive evolutionary response: when a host develops an effective antibody that binds tightly in a region that blocks the receptor site (preventing the virus docking with and gaining access to target cells) there is selective pressure on the virus. (Antibodies may block the receptor site by physically getting in the way or perhaps by triggering a conformational change in the antigen). Those mutations to which the effective antibody no longer binds well will be selected for.

Enveloped Virions - Membrane Fusion

The HA spike of influenza serves two functions: docking to a target cell and gaining entry to the cell. Influenza is an enveloped virus: it has a double phospholipid membrane enclosing its capsid. This envelope is acquired from membrane lipids of the host cell in which the virion was made and is capable of fusing with the membrane of the target cell. It is by fusion between the viral envelope and host cell membrane that the virion gains access to the host cell. The general process of membrane fusion is illustrated below:

Membrane fusion step 1

Above: the viral envelope (top) approaches the target cell membrane (bottom).

Membrane fusion step 2

Above: the inner phospholipid leaflets begin to fuse.

Membrane fusion step 3

Above: the outer phospholipid leaflets meat.

The membranes are in an unstable intermediate state.

Membrane fusion step 4

Above: an hour-glass shaped fusion pore opens between the two membranes.

This pore enlarges as the entire viral envelope is incorporated into the host cell membrane and the virus enclosed by the fusing envelope is released into the target cell's cytoplasm through the pore, once it is large enough. Note that the fusion pore exhibits negative curvature: in a sphere where the two great circles meet they curve in the same direction - towards the opposite side, a sphere has positive curvature. In a saddle, one axis curves upwards from front-to-back, the other downwards from side-to-side, the saddle has negative curvature. Wrap or rotate the saddle around its long axis and you have an hour-glass shape, which also has negative curvature like the fusion pore.

This fusion process does not happen spontaneously as it has to go from two positive curvature states through the high-energy intermediate steps before forming the pore with its negative curvature and to do this the membranes have to approach very closely. Membranes are also generally negatively charged and so repel one-another at close range: energy must be supplied to overcome this repulsion barrier. This is one of the function of the HA spikes. The HA spikes store elastic potential energy which is released when the HA adheres to its receptor (sialic acid). Part of the HA spike acts as a fusion protein. Other enveloped viruses frequently employ a similar mechanism, e.g. HIV. The mechanism in influenza is well illustrated on ViralZone. In general the mechanism is shown below (omitting the initial adhesion step and showing only the fusion protein):

Viral membrane fusion step 1

Above: the viral membrane is shown (bottom, violet) bearing fusion proteins and the target cell membrane is above it (orange). Binding of the adhesin to its receptor (not shown) brings the fusion machinery within range of the target cell and causes the cleavage of part of the spike to expose the fusion peptide. The stored potential energy is released and the fusion peptide (translucent orange) springs outwards and is driven into the target cell membrane.

Viral membrane fusion step 2

Above:the rest of the fusion protein acts as a hinge that springs shut as it changes to a more stable conformation (forming hairpin loops) causing it to fold over, driving the two membranes together, overcoming the energy barrier to membrane fusion which then proceeds as above to form a fusion pore through which the virus enters the host cell.

Below: the structure of the fusion protein in HIV 1. The fusion peptide (FP) is shown in orange, near the N-terminal, TM is a transmembrane domain that anchors the protein in the HIV envelope near the C-terminus. HR refers to heptad repeats: repeated sequences of seven amino acids near the N-terminus (NHR) in blue and near the C-terminus (CHR) in red which make up the hinge mechanism along with the loop-forming strand connecting them. This fusion protein also occurs as a trimer of three such subunits (as illustrated in the fusion diagrams above).

HIV fusion protein (gp41)

Part of the sequence (from residue 660 to residue 683) is shown (Cardoso et al., 2005). The regions in red are sites to which known neutralizing antibodies bind. The binding of antibodies to these regions prevents the hinge from operating and although the virus may still dock with a target cell it will be unable to initiate membrane fusion and so unable to enter. Without treatment most people succumb to HIV within about 5 to 10 years, despite the production of anti-HIV antibodies. Thus, although these antibodies can keep the infection in check for a time, they are not enough to eliminate the infection. (In those with natural immunity this probably involves additional immune factors). It takes time for the immune system to manufacture suitable antibodies, by which time many of the T cells, which HIV destroys, may already be infected.


Additional References

Cardoso, R.M.F., M.B. Zwick. R.L. Stanfield, R. Kunert, J.M. Binley, H. Katinger, D.R. Burton and I.A. Wilson. 2005.  Broadly Neutralizing Anti-HIV Antibody 4E10 Recognizes a Helical Conformation of a Highly Conserved Fusion-Associated Motif in gp41. Immunity, 22, 163–173.

Kelley LA et al. 2015. The Phyre2 web portal for protein modeling, prediction and analysis. Nature Protocols 10, 845-858.

Schneidman-Duhovny D, Inbar Y, Nussinov R, Wolfson HJ. PatchDock and SymmDock: servers for rigid and symmetric docking. NAR, 33: W363-W367, 2005.

Schneidman-Duhovny D, Inbar Y, Nussinov R, Wolfson HJ. Geometry based flexible and symmetric protein docking. Proteins, 60: 224-231, 2005.


Article created:8 Dec 2020.