The Coronavirus Replication Cycle Explained

Coronaviruses reproduce by commandeering the machinery inside your own cells, running through a tightly choreographed sequence that begins the moment the virus latches onto a cell surface receptor and ends when freshly assembled particles escape to infect new cells. The full cycle takes roughly ten hours for SARS-CoV-2, and every step along the way represents a potential vulnerability. Understanding it has driven the development of antivirals like remdesivir and Paxlovid, and it explains why coronaviruses are both remarkably successful and surprisingly targetable.

Docking and Entering the Cell

The first contact between a coronavirus and a host cell happens through the spike protein, the large, club-shaped projection that studs the viral surface. The spike is really two functional halves fused together. Its upper portion, called S1, contains a receptor-binding domain that grabs onto a protein on the outside of the host cell. For SARS-CoV and SARS-CoV-2, that target is ACE2, a protein found on cells lining the lungs, gut, blood vessels, and other tissues. Crystal structures have shown that the receptor-binding domain presents a gently concave surface that cradles a specific part of ACE2, fitting it snugly like a hand cupping a ball.¹ Lab experiments confirmed that SARS-CoV-2’s binding domain grips human ACE2 more tightly than the original SARS virus does, which partly explains its greater transmissibility.²

Binding alone is not enough. For the virus to actually merge with the host cell membrane and inject its genome, the spike must be cut at specific sites by host enzymes. Two cleavage events matter. One occurs at the boundary between S1 and S2, and another occurs at a site within S2 called S2′. That second cut is the trigger that exposes a stretch of amino acids called the fusion peptide, which punches into the host membrane and pulls the two membranes together.³ A host enzyme called TMPRSS2, sitting on the surface of lung cells, can perform this cleavage and dramatically accelerate membrane fusion.⁴ When TMPRSS2 is unavailable, the virus can still get in through an alternative route: it gets swallowed into an internal compartment called an endosome, where other enzymes (cathepsins) do the cutting instead.⁵ This backup plan is one reason the virus is so hard to block at the entry stage alone.

Translating the Viral Genome

Once the viral RNA spills into the cytoplasm, the cell’s own ribosomes treat it much like a normal messenger RNA and begin reading it immediately. No DNA copy is ever made; the entire replication cycle of a coronavirus unfolds in the cytoplasm, never touching the nucleus. The first two-thirds of the genome encodes two enormous chains of linked proteins called polyproteins. These are translated directly from the incoming RNA.

One of the cleverest tricks in the cycle happens right here. The shorter polyprotein, pp1a, is produced by straightforward translation. But to make the longer polyprotein, pp1ab, the ribosome has to stumble at a very precise point in the RNA where a tightly knotted structure called a pseudoknot jams against the ribosome’s entry channel. This causes the ribosome to slip backward by one nucleotide and continue reading in a shifted frame, giving access to a whole additional stretch of genetic code that includes the viral copying enzyme.⁶ Without this frameshift, the virus cannot produce its RNA-copying machinery, so replication dies on the spot.⁷ Drug designers have explored targeting this pseudoknot, since small molecules that lock it in place or disrupt it could shut down the entire downstream process.⁸

Cutting the Polyproteins Into Working Parts

The huge polyprotein chains are useless in their fused form. They must be chopped into individual nonstructural proteins, each with a specific job. Two viral proteases handle this task. The first few cuts near the front of the chain are made by one or two papain-like proteases, which separate the initial nonstructural proteins.⁹ The bulk of the remaining cuts, at roughly eleven different sites, are carried out by the main protease, often abbreviated Mpro. It chops the polyprotein in a defined order, and the sequence of cuts matters because it coordinates the timing of the whole operation. Structural studies have shown that Mpro’s active site is flexible enough to accommodate many different recognition sequences, yet the polyprotein’s own shape and dynamics partly govern which sites get cut first.¹⁰ This protease is the target of nirmatrelvir, the active ingredient in Paxlovid, because blocking it freezes the polyprotein in its unprocessed state and stops replication cold.

Building a Hidden Replication Factory

Coronaviruses do not scatter their copying machinery loose in the cytoplasm. Instead, they hijack the cell’s internal membranes to build a sheltered workspace. Several of the newly freed nonstructural proteins remodel the endoplasmic reticulum, the cell’s membrane-rich manufacturing network, into clusters of double-membrane vesicles. These structures form a so-called reticulovesicular network that is visible under electron microscopy.¹¹ Studies of SARS-CoV and SARS-CoV-2 in multiple cell types confirm that these double-membrane vesicles originate from ER membranes.¹²

These vesicles serve at least two purposes. First, they concentrate the replication enzymes and RNA templates in a small space, improving the efficiency of copying. Second, the double membrane physically hides the double-stranded RNA intermediates that would otherwise be detected by the cell’s innate immune sensors. It is essentially a cloaking device: the virus builds itself a private room inside the cell so that the alarm does not sound while it copies itself.

The RNA Copying Engine

The centerpiece of the replication factory is the RNA-dependent RNA polymerase, the enzyme that copies the viral genome. In SARS-CoV-2, this is the protein called nsp12, and it does not work alone. It forms a complex with two smaller cofactors, nsp7 and nsp8, that stabilize its structure and help it grip the RNA template. Structural studies resolved this complex at near-atomic resolution and revealed that nsp12 adopts a right-hand-like shape with finger, palm, and thumb regions enclosing the active site where new RNA is synthesized.¹³ Two copies of nsp8 extend long helical arms along the exiting RNA, forming positively charged “sliding poles” that keep the polymerase attached to the template over long stretches. This processivity is critical because coronavirus genomes are enormous for RNA viruses, around 30,000 nucleotides, and the polymerase must copy the full length without falling off.¹⁴

Notably, the polymerase itself is error-prone. Left unchecked, it would introduce mutations at a rate that could destroy the genome’s integrity. Coronaviruses solve this problem with a built-in proofreading enzyme, an exonuclease embedded in nsp14 that reads newly copied RNA and snips out mismatched nucleotides.¹⁵ This proofreading step is extremely unusual among RNA viruses and is one reason coronaviruses can maintain such large genomes without collapsing under the weight of their own errors. It also creates a headache for drug developers, because the proofreader can remove certain antiviral drugs after they get mistakenly incorporated into the RNA chain. Experiments showed that nsp14’s exonuclease can rescue an RNA primer poisoned with the hepatitis C drug sofosbuvir, allowing the polymerase to resume copying even in the presence of the drug.¹⁵ Not all nucleoside drugs are equally vulnerable to this editing, though, and that distinction has shaped which antivirals actually work against coronaviruses.

Making Messenger RNAs for Structural Proteins

The replication machinery must produce two kinds of RNA output. First, it copies the full-length genome for packaging into new virus particles. Second, it generates a set of shorter messenger RNAs, called subgenomic RNAs, that encode the structural and accessory proteins the virus needs. Structural proteins like the spike, envelope, membrane, and nucleocapsid are all translated from these shorter messages.

The way subgenomic RNAs are made is distinctive. Rather than simply reading portions of the genome, the polymerase uses a “discontinuous transcription” mechanism. As it copies RNA from the body of the genome, it encounters signal sequences called transcription-regulating sequences that sit upstream of each gene. When the polymerase hits one of these signals, it pauses and jumps to the far end of the genome, attaching a common leader sequence to the front of the new RNA.¹⁶ The result is a set of messenger RNAs that all share the same beginning but encode different proteins from the latter part of the genome.¹⁷ This template-switching mechanism is a hallmark of coronaviruses and is relevant to diagnostics: PCR tests often detect subgenomic RNAs as a sign of active viral replication rather than leftover genomic fragments.

Capping the RNA to Avoid Detection

Newly made viral messenger RNAs need a chemical cap at their front end. Without it, the cell’s own defenses would recognize them as foreign, chew them up with enzymes, and trigger immune alarm signals. Coronaviruses replicate entirely in the cytoplasm, meaning they cannot borrow the capping machinery in the nucleus that the cell uses for its own messages. Instead, they carry their own multi-step capping assembly line.

The process begins with the polymerase complex itself. The nsp12 protein contains a domain called NiRAN that attaches the initial cap structure to the RNA.¹⁸ Next, nsp14 adds a methyl group to create what is called a cap-0 structure. Finally, the enzyme nsp16, which requires nsp10 as an activator, adds a second methyl group to complete a cap-1 structure that closely mimics the caps found on the cell’s own messenger RNAs.¹⁹ The full capping pathway has been reconstituted in the lab using only viral proteins, confirming the virus is self-sufficient for this task.²⁰ This self-reliance ensures that the virus’s mRNAs are translated efficiently while flying under the cell’s immune radar.²¹

Assembly and Budding

Once the structural proteins are translated, they converge at a membrane compartment between the ER and the Golgi apparatus, known as the ERGIC. The spike protein, left to its own devices, would travel to the cell surface and cause infected cells to fuse with their neighbors, forming giant multinucleated cells called syncytia. The envelope (E) and membrane (M) proteins prevent this by retaining the spike inside the cell at the ERGIC. The M protein uses a retrieval signal on the spike’s cytoplasmic tail to hold it in place, while the E protein modulates the secretory pathway to keep spike from escaping.²²

Meanwhile, the nucleocapsid (N) protein wraps around newly copied genomic RNA, bundling it into a compact ribonucleoprotein package. N protein achieves this partly through a biophysical process called phase separation: it forms concentrated liquid droplets with RNA that bring the two components together efficiently.²³ N protein also interacts with the M protein at the ERGIC membrane, but these two interactions are mutually exclusive. In other words, N protein first condenses with RNA to package the genome, then hands off the package to M protein at the budding site.²⁴

The M protein is the true driver of particle assembly. When the ribonucleoprotein complex associates with E, S, and M at the ERGIC membrane, M undergoes a conformational switch from a short form to a long form, which bends the membrane and wraps it around the viral cargo to form a new virus particle.²⁵

Escaping the Cell

Most enveloped viruses leave the cell through the standard secretory pathway, the same route cells use to ship proteins to their surface. Coronaviruses take a different exit. Research has shown that beta-coronaviruses, including SARS-CoV-2, hijack lysosomes for their escape. Lysosomes are compartments normally used to break down cellular waste, but the virus repurposes them as transport vehicles.²⁶

A viral protein called ORF3a plays a central role in this hijacking. It promotes the movement of lysosomes toward the cell surface by facilitating the recruitment of transport machinery, and it triggers those lysosomes to fuse with the outer membrane and dump their contents, including new virus particles, outside the cell.²⁷ The virus also raises the pH inside lysosomes, making them less acidic. This deacidification protects the virus particles from being degraded by the harsh enzymes that normally fill these compartments, and it further boosts viral release.²⁸ The standard secretory pathway still serves as a backup, but lysosomal exocytosis appears to be the primary and more efficient route.

Shutting Down the Host’s Defenses

While the replication cycle churns along, the virus is simultaneously sabotaging the cell’s ability to fight back. One of the very first proteins produced during infection, nsp1, binds directly to the host ribosome and plugs the channel where messenger RNA enters. This blocks translation of the cell’s own messages, including those encoding antiviral proteins, while the virus’s own RNA uses a special leader structure to bypass the blockade.²⁹ The effect is a broad shutdown of host protein production that cripples the cell’s ability to mount an immune response.³⁰

Beyond the translational blockade, multiple viral proteins team up to suppress interferon signaling, the cell’s primary alarm system for viral infection. At least three proteins attack different steps in this pathway. Nsp6 binds to a key signaling protein called TBK1 and prevents it from activating the transcription factor IRF3, which normally turns on interferon genes. Nsp13 blocks TBK1 through a different mechanism. And the accessory protein ORF6 intercepts IRF3 after it has been activated, preventing it from entering the nucleus to switch on interferon production.³¹ Nsp6 also interferes with the downstream response to interferons by blocking the movement of a signaling molecule called STAT2 into the nucleus, suppressing the cell’s ability to respond even if interferons do get produced.³² Altogether, the SARS-CoV-2 genome encodes numerous proteins that collectively suppress both the production and signaling of type I and type III interferons.³³

Where Antiviral Drugs Intercept the Cycle

Knowing the replication cycle in detail has given drug developers clear targets. The two most prominent antivirals approved against SARS-CoV-2 each hit a different step.

Remdesivir mimics a natural building block of RNA. Once the polymerase incorporates it into the growing RNA chain, the chain stalls a few positions later because the drug creates a structural clash that the polymerase cannot push past.³⁴ What makes remdesivir effective where some other nucleoside drugs fail is that it partially evades the nsp14 proofreading mechanism. The delay between incorporation and chain termination means the drug gets buried several nucleotides deep, making it harder for the exonuclease to reach and remove it.

Nirmatrelvir, the antiviral component of Paxlovid, takes a completely different approach. It inhibits the main protease Mpro, preventing the polyproteins from being cut into functional parts.³⁵ Without this processing step, the replication machinery never assembles, and the virus cannot copy its genome or produce structural proteins. Because Mpro has no close human equivalent, the drug can target it with relatively few side effects. Nirmatrelvir is co-administered with ritonavir, which slows its breakdown in the liver to keep drug levels high enough to work.³⁶

Other nucleoside analogs work by a different strategy entirely: rather than stalling the chain, some are incorporated so frequently that the genome accumulates a lethal number of copying errors, a process sometimes called lethal mutagenesis.³⁷ Molnupiravir works roughly along these lines. Each antiviral strategy exploits a specific vulnerability in the replication cycle, which is why the cycle’s details matter beyond the lab.

Cellular Stress and the ER Under Siege

The massive replication activity does not go unnoticed by the cell. The flood of viral proteins pouring through the endoplasmic reticulum, combined with the physical remodeling of ER membranes into double-membrane vesicles, triggers a stress response called the unfolded protein response. This is a protective system the cell normally uses when too many misfolded proteins accumulate. It activates three sensor molecules in the ER membrane that collectively slow down protein production, ramp up the cell’s folding capacity, and dispose of proteins that cannot be rescued.³⁸

The relationship between this stress response and the virus is complicated. In some contexts, the unfolded protein response helps control infection by shutting down protein synthesis and triggering cell death before the virus can finish its cycle. But coronaviruses appear to selectively manipulate parts of this response, keeping the cell alive long enough to finish producing new virus particles while co-opting autophagy components to help form the double-membrane replication vesicles.³⁹ The virus, in effect, walks a tightrope: it stresses the cell enough to remodel its membranes but manages the response carefully enough that the cell doesn’t self-destruct before assembly is complete.

How Spike Mutations Reshape Host Range

The replication cycle’s fidelity mechanisms keep the mutation rate lower than that of most RNA viruses, but mutations still accumulate, especially in the spike protein’s receptor-binding domain, where even single amino acid changes can alter which species the virus infects. The substitution N501Y, present in several major SARS-CoV-2 variants of concern, significantly boosted the virus’s ability to infect cells expressing mouse ACE2, indicating a broader potential host range.⁴⁰ Another substitution, T478I, even enabled pseudoviruses to use chicken ACE2 in laboratory assays, hinting at a potential ability to jump into avian species.

Evolutionary analyses have also uncovered how the virus adapted to humans in the first place. Position 519 of the spike protein holds a histidine in all human SARS-CoV-2 sequences but an asparagine in closely related bat and pangolin coronaviruses. Viruses carrying the ancestral asparagine at this position show reduced replication in human lung cells and lower binding to human ACE2, suggesting the switch to histidine was a key step in the virus’s adaptation to humans.⁴¹ Occasionally, mutations in the proofreading exonuclease itself can accelerate the overall mutation rate, giving a cluster of viruses an evolutionary fast lane. One such proofreading mutation was linked to a burst of rapid evolution in a lineage of SARS-CoV-2.⁴² These findings underscore that the replication cycle is not just a static process. Its own machinery can evolve in ways that reshape the virus’s future trajectory.