Herpes Simplex Virus: Structure, Entry, Interaction, and Resistance

Herpes simplex virus is one of the most structurally complex human pathogens known, built from layers of protein and lipid that each serve distinct roles in infection. Its ability to enter a wide range of cell types, travel long distances inside neurons, hide from the immune system for a lifetime, and develop resistance to frontline drugs all trace back to specific molecular features that researchers have been mapping for decades. Understanding this machinery reveals not just how the virus works, but why it has been so difficult to cure.

Anatomy of the Virus Particle

An HSV particle, or virion, is organized into four concentric layers. At the center sits the viral DNA genome, packaged inside an icosahedral capsid roughly 125 nanometers across. That capsid is enclosed by an amorphous protein layer called the tegument, which in turn is wrapped by a lipid envelope studded with glycoproteins. Each layer has a specific job during infection, and the proteins within them interact in surprisingly complex networks.

The capsid itself is an intricate shell assembled from several proteins. High-resolution imaging has revealed the structures of the major capsid protein VP5, the triplex proteins VP19c and VP23, and the small capsid protein VP26 at near-atomic detail, showing that HSV capsid proteins display an extraordinary range of shapes and structural flexibility compared to simpler viruses.¹ At a single vertex of each capsid sits a ring-shaped portal complex, built from twelve copies of a protein called UL6. This portal functions as the channel through which viral DNA is threaded into the capsid during assembly and later ejected into the host cell nucleus during infection.²

The tegument is arguably the most functionally diverse layer. It contains at least 21 known proteins, some of which are released into the cell immediately upon entry to begin manipulating the host, while others remain stuck to the capsid and help ferry it toward the nucleus.³ One of the most important tegument proteins, UL36, adopts a remarkably elongated stalk shape, measured at about 34 nanometers long but only about 3 nanometers wide.⁴ UL36 acts as something of a central hub, forming direct interactions with UL37, the major tegument protein UL48, and the capsid itself, linking the inner and outer layers of the virion together. A yeast two-hybrid screen of 13 tegument proteins identified nine distinct protein-protein interactions, including self-associations of several tegument components, underscoring how densely cross-linked this layer really is.⁵

How the Virus Attaches to Cells

HSV entry is a multi-step process that begins with the virus sticking to sugar chains on the cell surface. The initial receptor for both HSV-1 and HSV-2 is heparan sulfate, a type of sugar chain found on proteoglycans decorating most human cells.⁶ Two viral envelope glycoproteins, gB and gC, mediate this attachment. Viruses lacking both gC and gB are severely impaired in their ability to bind heparan sulfate, making this step essential for the initial grab.⁷ The specificity of this binding is notable: heparin and highly sulfated heparan sulfate can compete with the cell surface receptor and block attachment, whereas related sugars like chondroitin sulfate and dermatan sulfate cannot.⁸

Attachment to heparan sulfate, however, is just the opening handshake. The virus cannot get inside without engaging a second, more specific receptor. This is where glycoprotein D takes over. gD binds to one of several host cell receptors, with nectin-1 being the most broadly used entry receptor in human tissues. Structural studies have shown that nectin-1 binds to gD at a site distinct from the binding site used by another receptor called HVEM. A single amino acid at the tip of a loop on nectin-1, phenylalanine 129, protrudes into a groove on gD that is otherwise occupied by gD’s own tail, and mutating that residue prevents both gD binding and viral entry.⁹ Both HSV-1 and HSV-2 use a conserved mode of nectin-1 recognition, meaning the same basic receptor-binding strategy is shared between the two serotypes.¹⁰

Membrane Fusion and Getting Inside

Once gD binds its receptor, the virus still needs to merge its envelope with a host membrane. This fusion step requires a cascade of signals among glycoproteins. After gD engages nectin-1 (or HVEM), its C-terminal region is displaced, which sends a signal to a heterodimer of gH and gL. The gH/gL complex in turn activates gB, the actual fusion protein that refolds and drives the two membranes together.¹¹ The details of how gH/gL activates gB remain an open question, but the general relay is well established: gD senses the receptor, gH/gL transduces the signal, and gB executes fusion.

HSV can fuse directly at the cell’s outer membrane or can be swallowed into an endosome first, fusing with the endosome wall instead. In classic laboratory cell lines like Vero cells, the virus fuses at the plasma membrane without needing an acidic environment. But in HeLa and CHO cells, entry depends on endocytosis and exposure to low pH.¹² This means the virus can switch between at least two distinct entry pathways depending on the cell type it encounters, a flexibility that likely helps it infect the wide range of tissues it is known to target.¹³

Human keratinocytes, the primary target during skin infection, reveal yet another variation. HSV-1 enters these cells through nectin-1-dependent plasma membrane fusion with striking speed, reaching maximum penetration within about five minutes. Even more unusually, entry into keratinocytes still worked at temperatures as low as 7°C, conditions that completely shut down endocytosis. Electron microscopy confirmed naked capsids sitting just beneath the plasma membrane with no sign of virions inside vesicles.¹⁴ This rapid, cold-tolerant pathway appears to be unique to keratinocytes and has not been observed in other cell types tested.

Traveling to the Nucleus

After the capsid is dumped into the cytoplasm, it must reach the cell nucleus to deliver its DNA. In most cells, the capsid rides the cell’s microtubule network like a highway, using motor proteins to pull it toward the nuclear envelope. This process is especially dramatic in neurons, where the virus enters at nerve endings in the skin and must travel long distances back along axons to reach the neuronal cell body in a ganglion.¹⁵

Once the capsid arrives at the nucleus, it docks at a nuclear pore complex. Two proteins are critical for this step: the capsid protein pUL25 and the tegument protein pUL36, the same elongated stalky protein that bridges layers inside the virion.¹⁶ Docking alone is not enough. The virus must then eject its tightly wound DNA through the pore and into the nucleus. Studies using temperature-sensitive mutants of both pUL25 and pUL36 found that capsids with defective versions of these proteins could still dock at nuclear pores but failed to release their DNA, sometimes remaining stuck for at least four hours.¹⁷ The high internal pressure inside the fully loaded capsid is thought to help drive the DNA out once the pore opens, somewhat like releasing a compressed spring.¹⁸

Transport within neurons during a later stage of the virus’s life cycle, when newly assembled particles move outward from the cell body toward the nerve terminals, involves a mix of strategies. Some newly made capsids travel separately from the envelope-bearing vesicles (a “separate” model), while others travel as fully assembled, envelope-and-all particles (a “married” model). Live-cell imaging of fluorescently tagged HSV in human neurons found that the majority of particles moving through axons were of the separate type, though married particles were also present.¹⁹

Evading the Immune System

HSV has evolved multiple strategies to dodge immune detection, targeting both the innate and adaptive arms of the host defense. One of its most important tricks involves blocking a cellular alarm system called the cGAS/STING pathway, which normally detects foreign DNA in the cytoplasm and triggers interferon production to alert neighboring cells. HSV-1 disables this pathway through at least two tegument proteins. UL41, a protein with RNA-degrading activity, reduces levels of the cGAS sensor itself, preventing the cell from recognizing viral DNA.²⁰ A second tegument protein, VP22, tackles the same pathway differently: it physically binds to cGAS and inhibits its enzymatic activity, blocking the production of the signaling molecule that would activate STING.²¹

On the adaptive immunity side, HSV targets the process by which infected cells display viral fragments on their surface to alert killer T cells. The viral protein ICP47 blocks a transporter called TAP, which normally ferries peptide fragments from the cytoplasm into the compartment where they are loaded onto MHC class I molecules. Without these loaded molecules reaching the cell surface, cytotoxic T cells cannot recognize the infected cell, giving the virus a window to replicate before the adaptive immune response catches up.²²

Latency and the Epigenetic Switch

HSV’s most remarkable survival strategy is latency. After the initial infection, the virus retreats to sensory nerve ganglia, where its DNA persists in neuronal nuclei as a circular molecule, largely silent and invisible to the immune system. The only viral transcript produced abundantly during this dormant period is the latency-associated transcript, or LAT.²³ LAT is not a messenger RNA that makes a protein in the usual sense. It is a stable intron, and part of its sequence runs antisense to a key viral regulatory gene called ICP0, which normally helps kick-start the virus’s active replication program. LAT can inhibit ICP0’s ability to turn on other genes, which may help keep the virus quiet.²⁴

LAT also plays a role in how efficiently latency is established in the first place. In mouse experiments, a wild-type strain established latent infections in about 30% of trigeminal ganglion neurons, and roughly a third of infected mice subsequently reactivated after heat stress. Mutant viruses with deletions in the LAT promoter or the 5′ end of LAT established latency in only about 10% of neurons, and those mice were impaired in reactivation as well.²⁵ So LAT appears to boost the number of neurons harboring latent virus rather than directly flipping the switch on reactivation.

The deeper mechanism behind latency involves chemical modifications to the histone proteins that package the viral DNA inside the nucleus. During latency, the LAT region of the genome carries histone marks associated with active transcription, while the rest of the viral genome, including the genes needed for active replication, is decorated with marks characteristic of silenced, tightly packed chromatin.²⁶ This epigenetic landscape is thought to be central to maintaining latency: the virus avoids chromatin silencing during active infection but allows itself to be silenced during dormancy, with reactivation involving a partial disruption of that silencing.²⁷ DNA methylation, by contrast, does not appear to play a significant role in regulating latent HSV gene expression.

Acyclovir Resistance and Its Molecular Roots

Acyclovir has been the mainstay of HSV treatment since the 1980s. It works by exploiting a viral enzyme, thymidine kinase, which converts acyclovir into its active monophosphate form. Cellular enzymes then add two more phosphate groups, and the resulting triphosphate jams the viral DNA polymerase, halting genome replication.²⁸ Because the drug’s activation depends on a virus-encoded enzyme, it selectively targets infected cells, which is why acyclovir has relatively few side effects in people with healthy immune systems.

Resistance arises most commonly through mutations in the thymidine kinase gene. Mutations at specific positions, including codons 51, 83, and 175 of HSV-1 thymidine kinase, can abolish the enzyme’s activity entirely, preventing acyclovir from being phosphorylated and thus rendering it useless.²⁹ Resistance-conferring mutations have been found in both conserved and nonconserved regions of the thymidine kinase gene, making it difficult to predict exactly where the next resistance mutation might appear.³⁰

Less commonly, resistance can also emerge from mutations in the viral DNA polymerase itself. Mutations within several conserved regions of the polymerase gene have been shown to cause cross-resistance not just to acyclovir but also to foscarnet and cidofovir, two backup drugs used when acyclovir fails. Some of these polymerase mutations even reduced sensitivity to cidofovir, narrowing the available treatment options considerably.³¹ Cross-resistance of this kind is a particular concern in immunocompromised patients, who are more likely to harbor resistant strains because their weakened immune systems allow prolonged viral replication under drug pressure.

Helicase-Primase Inhibitors as an Alternative

Because acyclovir resistance is driven by thymidine kinase or polymerase mutations, drugs targeting a completely different viral enzyme would sidestep those resistance pathways. This is the rationale behind helicase-primase inhibitors like pritelivir and amenamevir, which block the enzyme complex the virus uses to unwind and copy its DNA at the replication fork. Recent structural work using cryo-electron microscopy has revealed exactly how these drugs bind to the HSV-1 helicase-primase complex, showing that both pritelivir and amenamevir block helicase activity by fitting into the same general pocket but making distinct molecular contacts.³²

Resistance to helicase-primase inhibitors can emerge, but the picture is more complicated than with acyclovir. Certain mutations at equivalent positions in the UL5 helicase subunit across HSV-1, HSV-2, and varicella-zoster virus can reduce sensitivity to pritelivir or amenamevir by hundreds to thousands of fold. However, the two drugs do not always share the same resistance profile. Structural modeling predicts, and lab data confirm, that some mutations that severely reduce pritelivir’s effectiveness still leave amenamevir fully active, and vice versa.³³ There is also evidence that some helicase-primase resistance mutations come at a cost to the virus: one resistance mutation in UL5 helicase produced a slow-growing virus that was clearly less virulent in a mouse infection model compared to wild type, though a different mutation at the same amino acid position did not reduce virulence at all.³⁴ Whether a given resistance mutation impairs the virus or not depends on the specific substitution, which makes predicting real-world resistance trajectories tricky.

Gene Editing and Host-Targeted Approaches

All existing antiviral drugs share a fundamental limitation: they suppress active viral replication but leave the latent reservoir in ganglionic neurons untouched. The virus can always reactivate once drug pressure is removed. Gene editing offers a conceptually different approach, aiming to destroy the latent DNA itself. Using engineered meganucleases delivered by adeno-associated virus vectors, researchers have eliminated 90% or more of latent HSV-1 DNA in mouse models of orofacial infection, and up to 97% in mouse models of genital infection, with corresponding reductions in viral shedding.³⁵ CRISPR-Cas9-based editing has also shown promise in cell culture and three-dimensional tissue models, effectively reducing both latent virus and reactivation.³⁶ These approaches are still in preclinical stages, and delivering gene-editing tools efficiently to every latently infected neuron in a human ganglion remains a major challenge.

A separate strategy sidesteps viral targets altogether. Some researchers have turned to host-targeted antivirals, reasoning that if you block a host protein the virus depends on, resistance becomes much harder to evolve since the virus cannot simply mutate a host gene. One such target is RACK1, a scaffold protein the virus co-opts to translate its messenger RNAs. Small-molecule RACK1 inhibitors have shown effectiveness at blocking HSV-1 replication in cell culture.³⁷ The appeal of this approach is broad-spectrum potential: because multiple viruses rely on RACK1, a single drug could theoretically work against several unrelated pathogens. The tradeoff, inevitably, is that interfering with a host protein raises more safety concerns than targeting a viral one, and these compounds have not yet been tested in animals or people.