Influenza Virus: Structure, Entry, Replication, and Evasion

Influenza is one of the most studied viruses on Earth, yet it remains remarkably effective at infecting people year after year. That persistence comes down to how the virus is built, how it hijacks your cells, and how it outmaneuvers your immune system. The influenza A virus carries its genetic material on eight separate RNA segments, a design quirk that gives it an unusual ability to evolve and reshuffle its genes. Understanding how it moves from the surface of a cell all the way to the nucleus, copies itself, and slips past immune defenses explains not only why flu keeps coming back but also why it is so difficult to develop a universal vaccine.

Anatomy of the Virus Particle

An influenza virion is roughly spherical (though filamentous forms also exist) and is wrapped in a lipid envelope stolen from the host cell it last budded out of. Studding that envelope are two main surface proteins: hemagglutinin (HA) and neuraminidase (NA). HA is the protein the virus uses to latch onto cells, and NA is the enzyme that frees newly made viruses so they can spread. The two have to work in careful balance; if HA binds too tightly and NA cuts too weakly, new virions get stuck on the cell surface, and if the reverse is true, the virus cannot attach in the first place.¹ A smaller but critically important protein called M2 sits in the envelope and acts as a proton channel, letting hydrogen ions flow into the virus interior at the right moment during infection.

Beneath the envelope lies a shell of matrix protein (M1), which gives the particle structural rigidity. Inside that shell are the eight RNA segments, each wrapped by many copies of a protein called nucleoprotein (NP) and capped at one end by a three-part polymerase complex made of the subunits PA, PB1, and PB2. Each segment plus its proteins forms a ribonucleoprotein complex (RNP), and these RNPs are the functional units the virus needs for copying its genome once it reaches the host cell nucleus.² Having eight separate segments instead of one continuous genome is unusual among viruses and carries enormous evolutionary consequences discussed later in this article.

Latching On and Getting Inside

Entry begins when HA on the virus surface recognizes sialic acid, a sugar molecule that decorates many proteins and lipids on the surface of your cells. The specific type of linkage between sialic acid and its neighboring sugar matters enormously. Human influenza strains prefer a particular linkage arrangement abundant in the human upper airway, while avian influenza strains prefer a different arrangement found mainly in bird intestinal cells.³ This receptor preference is one of the main reasons bird flu viruses do not easily infect humans. But the picture is more complex than just two linkage types: the length of the sugar chain, chemical modifications along it, and even sulfur-containing additions all influence how well a given flu strain can grab hold.⁴

Once HA binds sialic acid, the cell engulfs the virus in a bubble of membrane called an endosome. The endosome’s interior gradually becomes more acidic as the cell pumps protons into it. When the pH drops to around 5.0–5.5, HA undergoes a dramatic shape change: a hidden stretch of amino acids called the fusion peptide shoots outward and inserts into the endosomal membrane.⁵ Cryo-electron microscopy has captured intermediate structures during this process, including a long triple-helical coil that may act as a bridge between the viral and endosomal membranes, pulling them together until they fuse.⁶ At a finer level, lowering the pH from about 7.8 to 5.2 first releases fusion peptides from their resting pockets and then triggers an unwinding of the central helices of HA into an extended form, a rearrangement that provides the mechanical force to merge the two membranes.⁷

Uncoating and Freeing the Genome

Membrane fusion alone does not finish the job. The viral RNPs are still bundled together and wrapped in M1 matrix protein. They need to be released into the cell’s cytoplasm as individual segments so they can travel to the nucleus. This is where the M2 proton channel earns its keep. While the virus sits in the early endosome, M2 lets protons seep through the viral envelope and acidify the virus interior.⁸ The internal M1 protein shell dissolves at a pH of roughly 5.5 to 6.0, loosening its grip on the RNPs.⁹

This is a two-step process, and both steps are essential. If you skip the early-endosomal acidification stage or block M2 with a drug, the genome does not release properly and infection stalls.¹⁰ The older antiviral drugs amantadine and rimantadine worked exactly this way, plugging the M2 channel. Unfortunately, widespread resistance has made them largely useless against circulating strains today.

Getting Into the Nucleus

Unlike many RNA viruses that replicate in the cytoplasm, influenza needs to get its RNPs into the cell nucleus. This is unusual and creates extra hurdles for the virus, but it also gives influenza access to the host’s own gene-copying machinery in a way that turns out to be extremely useful during transcription.

After M1 dissolves and the RNP bundles are freed, a transport protein called transportin 1 recognizes a specific signal on residual M1 still clinging to the RNPs. Transportin 1 strips off the remaining M1, which allows the RNP bundles to separate into individual segments.¹¹ Once separated, each RNP is recognized by a different set of import proteins (importin-α and importin-β), which ferry them through nuclear pores. The nucleoprotein coating each RNP carries two distinct nuclear import signals, and both contribute to getting the job done, though one appears to be the primary driver.¹² The virus essentially uses two different members of the import machinery family for two distinct steps: one for stripping the coat and one for nuclear entry.

Cap-Snatching and Genome Copying

Once inside the nucleus, the viral polymerase does something no other RNA virus polymerase is known to do in quite the same way. It physically associates with the host cell’s own RNA polymerase II (the enzyme that copies the cell’s genes) and steals the beginnings of the cell’s freshly made messenger RNA transcripts. Specifically, the PB2 subunit of the viral polymerase grabs the protective “cap” structure on a host transcript, while the PA subunit acts as an endonuclease, cutting the host RNA roughly 10 to 13 nucleotides downstream.¹³ The resulting short capped fragment then serves as a primer, giving the polymerase a running start to transcribe viral messenger RNA from the viral RNA template.¹⁴ This trick, called cap-snatching, is how the virus disguises its mRNAs as the cell’s own, letting them be recognized and translated by the cell’s protein-making machinery.

Cap-snatching handles transcription, the making of messenger RNA for protein production. But the virus also needs to copy its full-length genome for packaging into new virions, and that process (replication) is different: it produces complete RNA copies without a stolen cap. The balance between these two activities shifts over the course of infection. Early on, when host capped mRNAs are plentiful, the polymerase favors transcription. As infection progresses and polymerase concentrations rise, the system tilts toward genome replication.¹⁵ Small viral RNA fragments generated during replication may also help trigger this switch, accumulating as a signal that it is time to start making more genomes rather than more protein instructions.²

Exiting the Nucleus and Budding Out

Newly copied RNPs must leave the nucleus and reach the cell surface for assembly into new virions. This export depends on a complex involving the viral nuclear export protein (NEP), M1, and the host cell’s own export machinery. NEP recruits M1 to the RNPs inside the nucleus, forming an export-competent complex.¹⁶ M1 itself carries a nuclear export signal; when that signal is mutated, M1 gets trapped in the nucleus, and with it, the RNPs — virus production drops sharply.¹⁷

At the cell’s surface, newly made HA, NA, and M2 are already embedded in the plasma membrane, clustered in cholesterol-rich patches called lipid rafts. Spherical virions bud from these rafts in a process that does not require the cell’s actin skeleton, while filamentous virions — long, thread-like particles that some strains produce — do depend on actin filaments for their formation.¹⁸ The final step is release. Freshly budded virions are still tethered to the cell by sialic acid residues on the cell surface that HA would happily re-bind. NA cuts those sialic acid tethers, freeing the new virions to spread to neighboring cells and clearing a path through the sialic-acid-rich mucus lining the airway.¹⁹ Oseltamivir (Tamiflu) and zanamivir (Relenza) work by blocking NA, leaving virions stuck and unable to spread.

Shutting Down the Alarm

Your cells are not defenseless. When an infection is detected, cells normally produce type I interferons — signaling proteins that warn neighboring cells to prepare antiviral defenses. Influenza counters this with a protein called NS1, encoded on the smallest of its eight genome segments. NS1 is a potent interferon antagonist: it blocks the production of type I interferon in both airway cells and immune cells called dendritic cells, buying the virus critical time to replicate before the immune system ramps up.²⁰

NS1 does this partly by targeting a cellular sensor called RIG-I, which normally detects viral RNA and triggers the interferon alarm. Specifically, NS1 interferes with two host enzymes (TRIM25 and Riplet) that activate RIG-I, preventing the sensor from sounding the alarm. Interestingly, this interference is species-specific: the NS1 proteins from different flu strains vary in how well they suppress RIG-I in human versus avian cells, which has implications for cross-species transmission.²¹

Drift, Shift, and Why Flu Keeps Coming Back

Influenza evolves through two broad mechanisms. The first, antigenic drift, is the slow accumulation of mutations as the virus replicates. The viral polymerase lacks a proofreading function, so copying errors are frequent. Some of those errors land in the antigenic sites on HA — the very spots your antibodies recognize. Over time, enough mutations accumulate that antibodies from a past infection or vaccination no longer bind well, and you become susceptible again. Some of these HA mutations alter receptor binding at the same time they change the virus’s appearance to the immune system, meaning that antigenic diversification can happen even without direct antibody pressure, as compensatory mutations in other parts of HA restore receptor binding.²²

The second mechanism, antigenic shift, is far more abrupt and is the reason pandemics happen. Because the genome is split across eight segments, when two different influenza A viruses infect the same cell, their segments can mix and match during replication — a process called reassortment.²³ If a pig cell, for example, is simultaneously infected by a human strain and an avian strain, the new viruses emerging from that cell could carry HA from the bird virus on a backbone adapted to humans. The result is a virus that the human population has never encountered, and immune protection built up over years of seasonal flu exposure may offer little help. This reassortment-driven shift has been the engine behind every flu pandemic of the last century.²⁴

Why Bird Flu Rarely Crosses Into Humans

Beyond receptor preference, there is a deeper molecular barrier to cross-species infection. Avian influenza polymerases work poorly in human cells, and a host protein called ANP32A is the reason. The avian version of ANP32A has an extra 33 amino acids that the human version lacks, and avian flu polymerases depend on that extra stretch to replicate efficiently. In human cells, the shorter ANP32A does not support avian polymerase activity well.²⁵ For avian flu to replicate in people, the polymerase typically needs to acquire adaptive mutations — the best known is a single amino acid change in PB2 — that compensate for the shorter human ANP32A.

Pigs are a concern in pandemic planning partly because their version of ANP32A supports a broader range of influenza polymerases than the human version does, binding the polymerase more tightly and allowing both avian and mammalian viruses to replicate.²⁶ This makes pigs an ideal mixing vessel where avian and human strains can co-infect and reassort, potentially generating viruses with pandemic potential.

Original Antigenic Sin and Immune Memory

Your immune system’s memory of past flu strains is generally a good thing, but it comes with a quirk that complicates vaccination. When you encounter a new flu strain, your immune system preferentially boosts antibodies against the strain it first met in childhood, even if those antibodies are a poor match for the current threat. This phenomenon, called original antigenic sin, means your antibody response to a novel strain is partly hijacked by old memories.²⁷ The practical consequence is that people exposed to a new flu variant can produce antibodies skewed toward older strains at the expense of making antibodies against the novel parts of the new virus, potentially worsening the current infection.²⁸

This is one reason universal flu vaccine design is so challenging. A vaccine targeting the conserved stem region of HA — the part that changes less from year to year — is one leading approach. Antibodies that bind this stem can neutralize a broad range of flu strains and subtypes.²⁹ Nanoparticle-based vaccines designed to present the stem in a way that activates the right precursor immune cells are in development for both major groups of influenza A.³⁰ However, the virus can still acquire mutations in the stem and in NA that reduce the effectiveness of even broadly neutralizing stem antibodies, so the arms race continues.³¹

Defective Particles That Fight the Virus

Not every virion the cell produces is functional. During replication, the polymerase sometimes produces shortened, defective copies of genome segments. Virions carrying these truncated genomes are called defective interfering particles (DIPs). They cannot replicate on their own, but they compete with normal viruses for the cell’s replication machinery, effectively slowing down production of fully functional virus.³² Cells infected with high concentrations of DIPs show stronger interferon responses and higher rates of programmed cell death, and a significant fraction of those cells produce no functional viral proteins at all.³³

DIPs are more than a curiosity. Researchers are exploring them as a potential therapeutic strategy: administering synthetic DIPs to an infected airway could, in theory, outcompete the real virus and dampen infection. They also matter in vaccine manufacturing, where accumulation of DIPs during egg- or cell-based flu virus production can reduce the yield of usable vaccine virus.

How Researchers Manipulate Flu in the Lab

Much of what we know about each step in the flu lifecycle comes from a technique called reverse genetics. Instead of isolating a virus from nature and studying it as-is, researchers build it from scratch using DNA copies of all eight genome segments cloned into plasmids. Transfecting cells with these plasmids produces infectious virus, and by introducing deliberate mutations into any segment, scientists can figure out exactly which parts of each protein matter for attachment, replication, immune evasion, and everything else.³⁴

The same technology is central to vaccine production. Seasonal flu vaccines need to be updated constantly, and reverse genetics allows manufacturers to quickly swap in the HA and NA genes from newly circulating strains onto a well-characterized laboratory backbone that grows efficiently in eggs or cell culture. The approach works for both influenza A and influenza B: an eight-plasmid system for influenza B, for example, eliminated the need for older, slower methods that involved co-infecting cells with multiple viruses and selecting the right reassortant.³⁵ More recent refinements allow researchers to assemble all eight segments in a single test-tube reaction before transfecting them into cells, streamlining the process further.³⁶ Without reverse genetics, the detailed mechanistic picture of influenza described throughout this article — from cap-snatching to M2 channel function to NS1-mediated immune evasion — would be far less complete.