Virion Structure and Composition

Why the virion is built the way it is

The virion is the virus in its extracellular phase: the particle that has to survive the journey from one host cell to the next. It has two jobs, and the whole of virion architecture follows from them. First, it must protect the genome from heat, drying and chemical attack. Second, it must deliver that genome into a new cell, which means binding a receptor, crossing a membrane, and releasing its cargo at the right moment.

The constraint that shapes everything is genetic economy. A virus genome is small, and a container large enough to hold it cannot be encoded as one huge unique protein. The solution, used by almost every virus, is to build the shell from a large number (usually hundreds) of identical subunits that lock together. A handful of genes can then specify a structure orders of magnitude larger than themselves. This single idea explains why virions are so often symmetrical, and why their proteins are repeated rather than diverse.

How virion structure is studied

Viruses are smaller than the resolution limit of the light microscope (about 300 nm), so seeing them at all required new tools, and virology grew up alongside them.

Electron microscopy came first. Thin-sectioning of infected cells showed virions in context, while negative staining (Brenner and Horne, 1959), in which an electron-dense salt fills the interstices around a particle and casts it in relief, revealed surface detail and remains a fast diagnostic technique.

X-ray crystallography then took resolution to the atomic scale, revealing, for example, the wedge-shaped β-barrel subunits of picornaviruses and the structure of the influenza haemagglutinin that underpins vaccine strain selection.

Cryo-electron microscopy (cryo-EM) images flash-frozen, unstained particles in a hydrated state, close to how they exist in life. Once a runner-up to crystallography, it now routinely reaches near-atomic resolution and is the dominant method for virus particles and their protein machines, with X-ray crystallography reserved for the cases it still does best (small viral enzymes, and complexes of glycoprotein domains with antibody fragments). The flood of cryo-EM structures of the SARS-CoV-2 spike bound to antibodies is a measure of how far the technique has come. Its main limit is particle size: structures below roughly 100 kDa remain hard, but most viral spikes and polymerases are comfortably larger.

A few landmarks fix the scale. Poxviruses, at about 300 nm, are the largest viruses of vertebrates and were just visible by light microscopy; circoviruses, at 17–22 nm, are the smallest. Between them sit familiar landmarks: picornaviruses around 30 nm, flaviviruses around 50 nm, and influenza viruses and retroviruses around 100 nm. The giant amoebal viruses discovered in the last two decades (mimivirus, pandoravirus, pithovirus) reach 0.5–1.5 µm, the size of small bacteria, forcing a re-examination of where the limits of “virus” lie. None is a human pathogen.

The anatomy of a virion

Strip a virion to its essentials and there is a genome (one or a few molecules of DNA or RNA) wrapped in a protein coat, the capsid. The capsid plus its enclosed nucleic acid is the nucleocapsid. A few enzymes may be packaged inside.

The deepest structural division among viruses is whether they stop there or add a membrane. Non-enveloped (naked) viruses are just the nucleocapsid: a robust protein shell that is both shield and the agent of attachment.

Enveloped viruses wrap the nucleocapsid in a lipid bilayer envelope, acquired by budding through a cellular membrane (the plasma membrane, or an internal one such as the endoplasmic reticulum). The lipids are therefore host-derived and vary with the budding site, while the proteins embedded in them are virus-coded.

These surface proteins, the peplomers or spikes, are usually glycoproteins assembled as dimers or trimers, with the receptor-binding site at the tip. Many enveloped viruses also lay a non-glycosylated matrix protein against the inner face of the envelope, giving the particle rigidity and linking the envelope to the nucleocapsid.

Because the envelope is host membrane, it can carry the occasional host protein (such as major histocompatibility complex, MHC, molecules) outward with the virus, which can have unexpected effects on pathogenesis and on diagnostic assays.

The genome inside

The genome is the cargo, and its physical form varies more widely in viruses than anywhere else in biology. It may be DNA or RNA, single- or double-stranded, linear or circular, and carried as one molecule (monopartite) or several segments (multipartite). The way each genome is read and copied is its Baltimore class.

DNA genomes of vertebrate viruses are monopartite and usually double-stranded (the parvoviruses, circoviruses and anelloviruses are single-stranded). Some are linear, some circular; the papilloma- and polyomaviruses carry a circular, supercoiled genome, and the hepadnaviruses a distinctive partially double-stranded circle that is completed to covalently closed circular DNA (cccDNA) in the cell.

Several DNA viruses carry special ends: poxviruses have covalently closed hairpin termini, adenoviruses and parvoviruses have inverted terminal repeats, and adenoviruses, hepadnaviruses and parvoviruses (and some RNA viruses such as the picornaviruses) carry a protein covalently linked to the genome end that primes its replication.

RNA genomes are single-stranded except in three families (the reoviruses, birnaviruses and picobirnaviruses, all double-stranded and segmented).

Single-stranded RNA genomes are defined by their polarity. A positive-sense genome reads directly as messenger RNA (picornaviruses, flaviviruses, coronaviruses, togaviruses, retroviruses); a negative-sense genome is its complement and must be transcribed first (paramyxoviruses, rhabdoviruses, filoviruses, orthomyxoviruses); a few are ambisense, part each (the arenaviruses, one bunyavirus genus). Segment numbers are characteristic: two for arenaviruses, three for bunyaviruses, six to eight for the orthomyxoviruses, and ten to twelve for the reoviruses.

Every RNA virus encodes its own RNA-dependent RNA polymerase, because host cells have none; the exceptions are the retroviruses and hepadnaviruses, which carry a reverse transcriptase instead. Hepatitis delta is the one vertebrate virus with a circular single-stranded RNA, a structure reminiscent of plant viroids.

Two further features matter. Viral genomes are haploid (one copy of each gene), with a single exception: the retroviruses are diploid, carrying two copies. And because coding space is precious, viruses stretch it by reading the same sequence in overlapping frames, transcribing both DNA strands, splicing one transcript several ways, or cleaving a single polyprotein into many.

Symmetry: how the shell is built

Because a capsid is assembled from many identical subunits making the same contacts over and over, the result is symmetrical, and only two true symmetries occur in viruses: icosahedral and helical. Anything that is neither is called complex.

Icosahedral symmetry encloses the most volume with the least protein. Crick and Watson reasoned that a closed shell of identical subunits in identical environments could hold at most 60 copies. Most “spherical” viruses have many more, so Caspar and Klug proposed that the subunits sit in quasi-equivalent positions: nearly, but not exactly, the same, with slight flexing of the contacts.

The number of subunits is then a multiple of 60 set by the triangulation number, T (T = 1 gives 60 subunits, T = 3 gives 180, and larger values build larger shells). An icosahedron has twelve vertices, thirty edges and twenty triangular faces, with axes of five-, three- and twofold rotational symmetry.

Its surface units are capsomeres: subunits surrounded by six neighbours are hexons, and the twelve at the vertices, surrounded by five, are pentons. Adenovirus is the classic example, its faces tiled with hexons and a penton at each vertex.

Icosahedral capsids are everywhere in medical virology, in both naked and enveloped viruses. The naked ones are capsid and genome and nothing more: the adenoviruses, human papillomavirus, the picornaviruses (poliovirus and rhinovirus), and norovirus. In the enveloped ones the icosahedral capsid sits inside a membrane, as in the herpesviruses (herpes simplex virus and varicella-zoster virus), hepatitis B virus, and the flaviviruses (dengue and Zika).

Helical symmetry winds the subunits into a cylinder with the nucleic acid spiralling inside, as in the rigid rod of tobacco mosaic virus. In the vertebrate viruses that use it the helical nucleocapsid is always coiled up and wrapped in an envelope rather than left bare: influenza viruses (Orthomyxoviridae), measles and mumps (Paramyxoviridae), rabies virus (Rhabdoviridae), Ebola (Filoviridae) and the coronaviruses are all built this way.

Complex covers everything that is neither. The poxviruses are the largest and most familiar: brick-shaped, multilayered virions with no single overall symmetry, among them monkeypox (mpox) virus and molluscum contagiosum virus. The tailed bacteriophages are complex in their own way, a polyhedral head joined to a helical tail.

The mature HIV-1 capsid is another, and a particularly elegant one: it is a fullerene cone, a lattice of capsid-protein hexamers closed off by exactly twelve pentamers (the same rule that closes a football), curved into a cone rather than a sphere. It is neither icosahedral nor helical.

Symmetry describes the capsid, not necessarily the whole particle. Many enveloped viruses, influenza and HIV among them, are pleomorphic, varying in size and shape from one particle to the next, because the lipid envelope is not a rigid symmetric shell. An ordered nucleocapsid sits inside a far less ordered coat.

What capsids and envelopes do

A virion is not an inert box. Its surface proteins are machines for the most dangerous step in the cycle, getting in. They carry out at least four jobs: binding the receptor, fusing or breaching membranes, uncoating, and sometimes modifying the receptor.

The receptor is the molecule on the host cell that the virus binds; the ligand is the molecule on the virus that does the binding. The influenza haemagglutinin is a ligand for its receptor, sialic acid; the SARS-CoV-2 spike is a ligand for angiotensin-converting enzyme 2 (ACE2), and its receptor-binding domain flips between “down” (hidden) and “up” (exposed) states to engage it, a much-imaged example of how surface proteins gate attachment.

Fusion proteins illustrate a deeper principle. The influenza haemagglutinin is synthesised as an inactive precursor (HA0) and cleaved by a host protease into two linked chains. That cleavage does not switch the protein on; it leaves it metastable, a spring held under tension behind a high energy barrier (which is why a primed, infectious particle can sit stable in the cold for weeks).

The actual trigger, the low pH of the endosome for influenza, or receptor engagement for others, simply lowers the barrier and lets the protein snap into its fusion-active shape. This “primed but waiting” design recurs across enveloped viruses, the coronavirus spike included.

Fusion, physically, is a membrane merger: the outer leaflets of the viral and cellular membranes merge into a hemifusion stalk, and that stalk then opens into a fusion pore through which the genome passes into the cytoplasm. Non-enveloped viruses, having no membrane to fuse, instead breach the endosomal membrane to reach the cytosol.

Stability and inactivation

How long a virus stays infectious outside a host is not a fixed number but a rate: particles in a population lose infectivity progressively, so the time to sterilise a sample depends on both the conditions and the starting titre. There is no single figure for how long a virus survives outside the body.

Temperature is the dominant factor. As a rough approximation, the half-life of most viruses runs to seconds at 60 °C, minutes at 37 °C, hours at 20 °C, days at 4 °C, and years at −70 °C or below; long-term stocks are kept at −70 °C or in liquid nitrogen, or freeze-dried. Enveloped viruses are the more heat-labile, their fragile membrane proteins denaturing readily, which is why respiratory specimens are best delivered to the laboratory quickly and chilled rather than frozen.

Most viruses prefer isotonic, physiological pH and ionic conditions, but there are telling exceptions. Enveloped viruses are generally inactivated at pH 5–6, whereas the rotaviruses and many picornaviruses survive the stomach’s acid intact, consistent with their faecal-oral spread.

Lipid solvents and detergents are the sharpest divide. Ether, chloroform and detergents such as sodium deoxycholate dissolve the lipid envelope and destroy the infectivity of enveloped viruses, while leaving naked viruses untouched. This is the basis of the classic ether-sensitivity test, a quick way to tell whether an unknown virus is enveloped. (Virologists turn the same trick to their advantage, using mild detergents to strip envelopes and liberate proteins for vaccines and serological reagents.)

By way of contrast, prions, which are not viruses but fall within the virologist’s domain, are extraordinarily resistant, surviving boiling, irradiation and ordinary disinfection. They are a standing reminder that the rules above are rules for viruses, not for every transmissible agent.

References and recommended reading

• Burrell CJ, Howard CR, Murphy FA. Virion Structure and Composition. In: Fenner and White’s Medical Virology, 5th edition. Academic Press / Elsevier; 2017. Chapter 3, pp. 27–37. The readable backbone for this material: composition, genome architecture, symmetry, and the stability rules.
• Harrison SC. Principles of Virus Structure. In: Fields Virology, 7th edition, Volume 4: Fundamentals. Wolters Kluwer; 2023. Chapter 3. The current, deeper treatment; the source for the cryo-EM resolution shift, the metastable cleaved fusion protein, the hemifusion-to-pore pathway, viral pleomorphy, and the HIV-1 fullerene-cone capsid.