# Chapter 118. Virus-Specific RNA Condensation, Assembly, and Packaging

## Scope Note

This chapter owns virus-specific RNA condensation during replication and assembly, viral factories and replication organelles, context-specific viral RNP material states, capsid-genome recognition and packaging, virus-like particle (VLP) cargo selection, assembly-targeting interventions, and measurement of viral material states. It connects viral RNA structure ([Chapter 117](chapter1111.md)) to particle biology through tobacco mosaic virus, bacteriophages, picornaviruses, nodaviruses, HIV-1, influenza viruses, coronaviruses, and hepatitis B virus. General condensate physics and phase-separation evidence standards belong to [Chapter 58](chapter1053.md), nuclear RNA bodies to [Chapter 95](chapter1090.md), and the life cycles and viral manipulation of host stress granules, P-bodies, germ granules, and neuronal granules to [Chapter 105](chapter1100.md).

## Executive Summary

Viral RNA is not merely an information carrier; it is a physical object whose length, charge density, secondary structure, and segmental flexibility determine how it interacts with viral proteins, how it is condensed into particles, and how it is selectively packaged among the many host RNAs present in an infected cell. A typical single-stranded RNA viral genome carries one negative charge per nucleotide along a backbone of roughly 3,000 to 30,000 nucleotides. To package this polyelectrolyte into a capsid with an internal diameter of 20 to 80 nanometers requires overcoming charge repulsion, conformational entropy, and steric constraints. Viral proteins solve this problem through complementary electrostatic surfaces, nucleic acid chaperone activity, ATP-driven packaging motors, and assembly pathways that couple genome recognition to capsid closure.

The chapter organizes viral RNA condensation along two physical axes. The first axis concerns selectivity: some viruses recognize specific RNA sequences or structures called packaging signals that ensure only full-length viral genomes enter capsids, while others use charge-neutralization packaging in which any RNA of approximately the right length and charge can be encapsidated if assembly conditions permit. The second axis concerns the material state of the viral ribonucleoprotein: viral RNPs can exist as ordered helical filaments, quasi-spherical condensates, liquid-like replication factories, or disordered protein-RNA coacervates, and transitions among these states are governed by protein concentration, RNA length, ionic strength, temperature, and post-translational modifications.

Understanding viral RNA condensation as a materials problem has practical consequences. Virus-like particles -- capsids assembled without an infectious genome -- form the basis of several licensed vaccines, including those against hepatitis B virus and human papillomavirus. Cargo-selection rules determine what host, vector, or designed nucleic acids enter VLPs and therefore affect consistency, immunogenicity, and delivery applications. Assembly inhibitors that block or misdirect capsid formation or genome packaging are validated antiviral strategies: the HIV-1 capsid inhibitor lenacapavir and hepatitis B virus core protein allosteric modulators show that the viral assembly pathway is druggable. Synthetic mRNA formulation and lipid nanoparticle physics are treated in [Chapter 153](chapter1137.md) and [Chapter 156](chapter1139.md).

The chapter synthesizes biophysical mechanism, structural biology, and virological function. It begins with the thermodynamics and kinetics of RNA condensation by viral proteins, moves to the phase behavior of replication factories and RNPs, explains how packaging signals and charge-neutralization achieve selective genome packaging, describes the design and application of virus-like particles, and closes with antiviral and vaccine strategies that target or exploit viral RNA condensation. Throughout, the chapter distinguishes established biophysical mechanisms from hypotheses under active investigation and highlights measurement methods whose artifacts and assumptions must be understood to interpret the literature.

## Concept Inventory

- **RNA condensation in virology:** the compaction of a viral RNA genome into a volume orders of magnitude smaller than its free-solution radius of gyration, mediated by interactions with viral proteins, multivalent cations, and/or crowding agents. Condensation is distinguished from nonspecific aggregation by its reversibility, its structural order, and its functional coupling to infectivity. Boundary case: some viral genomic RNAs are not fully condensed inside the virion but exist as ribonucleoprotein filaments with segmental flexibility, as in influenza virus RNPs.
- **Viral ribonucleoprotein (RNP):** a complex of viral RNA with viral nucleocapsid or coat protein. Viral RNPs span a continuum from ordered helical nucleocapsids (tobacco mosaic virus, rabies virus) to disordered but compact cores (HIV-1, some retroviruses) to loosely associated protein-RNA assemblies. The term does not presuppose a specific structural arrangement.
- **Capsid assembly:** the process by which viral structural proteins self-associate or assemble in an energy-dependent manner to form a protein shell that encloses the viral genome. Assembly can be templated by nucleic acid (co-condensation assembly, as in many positive-strand RNA viruses and helical plant viruses) or can occur first as an empty procapsid into which the genome is subsequently packaged by an ATP-driven motor (as in many double-stranded DNA bacteriophages and herpesviruses, and the double-stranded RNA bacteriophage phi6).
- **Packaging signal (Psi element):** a structured RNA element, or collection of elements distributed across the genome, that is recognized by viral structural or nucleocapsid proteins and that confers selective advantage in genome packaging relative to non-viral or subgenomic RNAs. Packaging signals have been most definitively characterized in retroviruses, bacteriophages, and some positive-strand RNA plant viruses. Not all viruses use discrete packaging signals; many small positive-strand RNA viruses rely on electrostatic packaging coupled to genome replication.
- **Selective packaging:** that full-length viral genomic RNA is preferentially incorporated into virions when host RNAs, subgenomic RNAs, and defective viral RNAs are present at comparable or higher concentrations. Selectivity can arise from packaging signal recognition, from coupling between genome replication and assembly, from temporal and spatial compartmentalization, or from RNA-length discrimination by capsid geometry.
- **Liquid-liquid phase separation (LLPS) in viral factories:** a proposed or supported mechanism for forming some membraneless viral protein-RNA compartments used in replication, transcription, or assembly. Evidence is strongest for selected negative-strand RNA virus inclusion bodies and remains context-dependent. Many positive-strand RNA virus replication organelles are membrane-bound spherules or double-membrane vesicles, and morphology or component exchange alone does not establish LLPS or its necessity for replication.
- **Virus-like particle (VLP):** a macromolecular assembly of viral structural proteins that resembles the authentic virion in morphology and antigenicity but lacks the infectious viral genome. VLPs can be produced by recombinant expression of capsid or envelope proteins in heterologous systems, and they may encapsidate host RNA, no RNA, or specifically loaded cargo RNA depending on the expression system and purification conditions.
- **In vitro assembly:** reconstitution of viral capsids, VLPs, or nucleocapsid-like particles from purified components in a cell-free system. In vitro assembly is the gold standard for demonstrating that viral proteins and RNA are sufficient for particle formation and for measuring the thermodynamic and kinetic parameters of assembly.
- **Charge neutralization packaging:** a mode of genome encapsidation in which electrostatic attraction between a positively charged protein surface and the negatively charged RNA backbone drives co-condensation, with limited or no sequence specificity. The capsid interior of many positive-strand RNA viruses carries a high density of basic residues (arginine and lysine), and RNA of sufficient length and negative charge can drive assembly even if it is heterologous or non-viral.
- **Nucleic acid chaperone:** a protein that facilitates RNA conformational rearrangements by destabilizing misfolded structures and promoting the formation of thermodynamically stable folds, without ATP hydrolysis. Retroviral nucleocapsid proteins, coronavirus nucleocapsid proteins, and some plant virus coat proteins exhibit nucleic acid chaperone activity. Chaperone activity is important for genome dimerization, packaging signal exposure, and the structural transitions that accompany assembly.
- **Single-stranded RNA secondary structure in capsids:** the ensemble of base-paired helices, loops, and junctions that the viral genome adopts when condensed inside the virion. The in-capsid structure can differ substantially from the in-solution structure because protein contacts, molecular crowding, and charge neutralization alter the free-energy landscape of RNA folding. The in-capsid structure is constrained by capsid geometry and by specific protein-RNA contacts.
- **Genome-length determination:** the mechanism by which a virus ensures that capsids package genomes of approximately the correct length. In packaging-signal-based systems, length is determined by the number and distribution of signals along the genome. In charge-neutralization systems, length is determined by the electrostatic capacity of the capsid interior: RNA that is too short cannot neutralize enough positive charge to drive assembly to completion, and RNA that is too long cannot fit or creates excessive strain.
- **Cargo selection in VLPs:** the set of nucleic acids -- host RNA, heterologous RNA, DNA, or no nucleic acid -- that are encapsidated when viral structural proteins assemble in a given expression system. Cargo selection is controlled by assembly conditions (pH, ionic strength, protein concentration), the presence or absence of packaging signals on the cargo, the length and secondary structure of the cargo, and whether assembly occurs in vivo or in vitro.

## What to Know Before Reading This Chapter

The reader needs background in three areas. First, biophysical chemistry: the concepts of electrostatic interaction between charged polymers, excluded volume, molecular crowding, the radius of gyration of a flexible polymer, and the thermodynamic driving forces for macromolecular assembly (enthalpy-driven charge neutralization, entropy-driven counterion release, hydrophobic burial). The reader does not need a full course in polymer physics, but they should be comfortable with the idea that RNA is a polyanion and that proteins can present polycationic surfaces.

Second, the reader should know the general structure of viral capsids as covered in introductory virology: icosahedral and helical symmetry, the distinction between capsid protein and nucleocapsid protein, the concepts of T-number (triangulation number) and quasi-equivalence, and the presence or absence of a lipid envelope. This chapter assumes those terms and concepts are available from [Chapter 115](chapter1109.md) or from general virology background.

Third, the reader should distinguish between three uses of the word "assembly" in virology. Assembly can mean capsid protein oligomerization into a closed shell (the capsid assembly problem). It can mean genome packaging into a preformed capsid (the packaging motor problem). It can mean co-condensation of proteins and RNA into an RNP particle (the co-assembly problem). These are mechanistically distinct, and the chapter specifies which meaning is intended at each point.

The chapter uses examples from plant viruses, bacteriophages, and animal viruses. Plant viruses such as tobacco mosaic virus, cowpea chlorotic mottle virus, and brome mosaic virus have historically provided the most tractable in vitro assembly systems and remain the best-understood examples of RNA condensation biophysics. Bacteriophages provide the best-understood packaging motor systems. Animal viruses provide the medical relevance and the most complex RNP architectures. The reader should not dismiss plant and phage examples as esoteric; the principles they reveal apply broadly.

Several topics are intentionally separated from this chapter. General phase-separation theory and criteria are covered in [Chapter 58](chapter1053.md); nuclear RNA bodies in [Chapter 95](chapter1090.md); cellular granule life cycles and viral manipulation of host granules in [Chapter 105](chapter1100.md); membrane fusion, receptor binding, and entry in [Chapter 115](chapter1109.md); polymerase enzymology in [Chapter 116](chapter1110.md); innate immune sensing in [Chapter 108](chapter1103.md); mRNA vaccine and lipid nanoparticle design in [Chapter 153](chapter1137.md) and [Chapter 156](chapter1139.md); drug pharmacology and resistance in [Chapter 160](chapter1143.md); and retroviral Gag processing and maturation in [Chapter 120](chapter1114.md).

## 118.1. Virus-specific RNA condensation during replication and assembly

Viral RNA condensation is the process by which a flexible, highly charged polynucleotide is compacted by viral proteins into a structure that fits inside a capsid, protects the RNA from nucleases, and presents the genome in a form competent for subsequent uncoating and translation or transcription. The physical problem is formidable. A 10,000-nucleotide single-stranded RNA in free solution under physiological ionic strength has a radius of gyration on the order of 30 to 50 nanometers, whereas the internal cavity of a typical icosahedral capsid is 20 to 40 nanometers in diameter. The RNA must therefore be compacted by a factor of roughly 10 to 100 in volume. The compaction must be reversible: the RNA must be released upon entry into a new host cell. And the compaction must not destroy the information content of the RNA, meaning that intra-strand base pairing that is functional for translation, replication, or packaging must either be preserved or refoldable after release.

The driving forces for condensation are electrostatic and entropic. Viral capsid proteins and nucleocapsid proteins contain basic residue-rich domains -- often arginine- and lysine-rich N-terminal or C-terminal arms, or internal RNA-binding clefts -- that present a polycationic surface to the RNA polyanion. When protein and RNA associate, the counterions that previously neutralized the charged groups on both polymers are released into solution. Counterion release carries a large entropic gain because many small ions, each with high translational entropy in the bound state, are liberated into bulk solvent. This entropic driving force is the dominant thermodynamic contribution to RNA-protein co-condensation under many conditions, analogous to the entropic driving force for DNA condensation by polyamines and multivalent cations.

![Figure 118.1. Tobacco mosaic virus assembly as the paradigm of viral RNA condensation](../assets/figures/chapter1112_figure1.png)

**Figure 118.1. Tobacco mosaic virus assembly as the paradigm of viral RNA condensation.** Tobacco mosaic virus coat-protein disks recognize an RNA assembly origin, convert to a helical form, and nucleate bidirectional elongation that threads and condenses the genome; the nucleation barrier separates specific initiation from downhill growth.

Tobacco mosaic virus (TMV) provides the historical and pedagogical paradigm of RNA condensation in viral assembly. TMV is a helical virus: its coat protein assembles into a rod 300 nanometers long and 18 nanometers in diameter, with the single-stranded genomic RNA of approximately 6,400 nucleotides embedded between turns of the protein helix. The assembly process was reconstituted in vitro by Fraenkel-Conrat and Williams in 1955, establishing that purified coat protein and genomic RNA are sufficient for infectious particle formation. The assembly mechanism was subsequently dissected by Butler, Klug, and others in the 1970s: assembly initiates when a coat protein disk (a two-layer ring of 34 subunits) recognizes a specific RNA hairpin, the assembly origin, located within the coat protein gene. The RNA threads through the central hole of the disk, the disk converts from a cylindrical to a helical conformation, and additional disks or monomers add bidirectionally, drawing the RNA through the elongating helix and condensing it into the final rod. The TMV system established four principles that recur throughout viral RNA condensation biology: (1) an assembly origin or nucleation site that provides specificity, (2) a protein conformational switch between assembly-inactive and assembly-active states, (3) bidirectional or unidirectional growth of the protein shell coupled to RNA incorporation, and (4) the requirement that the RNA be long enough to span the full particle length.

The physics of RNA condensation by basic peptides and protein domains has been studied extensively through simplified model systems. Peptides corresponding to the arginine-rich RNA-binding domains of HIV-1 Tat, Rev, and Gag, of hepatitis B virus core protein, and of various plant virus coat proteins condense RNA in vitro in a charge-ratio-dependent manner. At charge ratios near unity (one positive charge from peptide per negative charge from RNA phosphate), these peptides form compact, roughly spherical RNA-protein particles with diameters of 15 to 50 nanometers. Below charge neutrality, condensation is partial; above it, charge inversion can occur, producing positively charged particles that are stable due to excess peptide. These model systems demonstrate that the electrostatic driving force alone, without specific RNA sequence recognition, can produce compact RNP particles with morphologies resembling authentic nucleocapsids.

Real viral systems add layers of specificity and regulation atop this electrostatic baseline. Many viral coat proteins undergo a pH-dependent or proteolytic conformational switch that activates their RNA-binding and assembly-competent state. In flock house virus (Nodaviridae), the capsid protein alpha undergoes an autocatalytic cleavage to produce beta and gamma polypeptides, and this maturation cleavage is required for particle stability and infectivity. In HIV-1, the Gag polyprotein contains a matrix domain that targets the assembling particle to the plasma membrane, a capsid domain that drives protein-protein association, a nucleocapsid domain that binds and chaperones the genomic RNA, and a p6 domain that recruits the ESCRT machinery for budding. The proteolytic cleavage of Gag by the viral protease during maturation reorganizes the particle from an immature, roughly spherical assembly into a mature, conical capsid core. This proteolytic cascade couples RNA condensation to particle architecture, membrane acquisition, and maturation -- all in a single polyprotein.

The kinetics of assembly matter as much as the thermodynamics. Capsid protein assembly in the presence of RNA is typically a nucleated polymerization. The rate-limiting step is the formation of a critical nucleus -- a small oligomer of coat protein bound to RNA -- after which elongation proceeds rapidly. The nucleation barrier ensures that assembly does not initiate on every RNA molecule simultaneously and that assembly is coupled to the availability of full-length genomic RNA. If nucleation were too facile, coat protein would be consumed in abortive assembly on short RNAs, fragments, and host transcripts. In vitro assembly kinetics show that the rate of particle formation depends on protein concentration with a sigmoidal profile characteristic of nucleated polymerization, and that RNA acts as a catalyst for assembly by providing a scaffold on which the critical nucleus can form.

Host factors influence condensation efficiency. Cellular RNA-binding proteins, helicases, and molecular chaperones can compete with viral proteins for RNA binding, can remodel RNA structures to expose or occlude packaging signals, or can be co-opted by the virus as assembly cofactors. The RNA helicase DDX6 and related DEAD-box proteins have been implicated in the assembly of hepatitis C virus and HIV-1 by remodeling viral RNA structures. Cellular polyanions such as polyphosphates and glycosaminoglycans can compete with RNA for binding to basic protein domains, potentially titrating assembly if their concentration is high. These host-virus interactions are poorly understood relative to the core assembly mechanism and represent an active area of investigation.

"RNA condensation in viruses is equivalent to DNA condensation in bacteriophages and eukaryotic chromatin." While both involve compaction of a polynucleotide by proteins or polycations, viral RNA condensation differs in several respects: single-stranded RNA has a different persistence length and folding landscape than double-stranded DNA; many RNA viruses co-condense protein and RNA simultaneously rather than packaging pre-condensed DNA into a preformed shell; and the electrostatic forces in RNA condensation often involve basic protein domains rather than small polyamines or histones. The principles are related but the molecular mechanisms and biological contexts are distinct.

## 118.2. Viral factories, replication organelles, and context-specific RNP material states

Viral replication does not occur in a homogeneous cytoplasmic solution. Infected cells reorganize their internal membranes and compartments to create specialized environments -- viral factories or replication organelles -- where genome replication, transcription, translation, and assembly are coordinated. The material state of these compartments spans a continuum from membrane-bound spherules and double-membrane vesicles to membraneless inclusion bodies with liquid-liquid phase separation properties to solid-like or gel-like RNP condensates. Understanding these material states as physical chemistry problems reveals how viruses concentrate components, exclude antiviral factors, and couple replication to assembly.

Positive-strand RNA viruses predominantly use membrane-bound replication organelles. As described in [Chapter 116](chapter1110.md), picornaviruses, flaviviruses, and coronaviruses remodel endoplasmic reticulum, Golgi, or mitochondrial membranes into spherules (invaginations with a narrow neck connecting the interior to the cytoplasm), double-membrane vesicles (concentric membrane pairs), or membranous webs (interconnected convoluted membranes). These membrane structures concentrate the replication-transcription complex, shield double-stranded RNA replication intermediates from cytoplasmic innate immune sensors, and provide a two-dimensional surface that increases the effective concentration of replication components. The membrane neck or pore is thought to allow NTP import and product single-stranded RNA export while excluding ribosomes and large cytoplasmic proteins. The material state of these organelles is that of a membrane-enclosed aqueous compartment with protein and RNA concentrations substantially higher than bulk cytoplasm, but the compartment itself is bounded by a lipid bilayer and is not a phase-separated condensate in the conventional sense.

![Figure 118.2. Context-specific material states of viral factories and replication organelles](../assets/figures/chapter1112_figure2.png)

**Figure 118.2. Context-specific material states of viral factories and replication organelles.** Flavivirus spherules, coronavirus double-membrane vesicles, and rabies Negri bodies concentrate viral RNA through distinct membrane-bound or condensate-like architectures; morphology, component exchange, and chemical sensitivity do not make their material states interchangeable.

Negative-strand RNA viruses, by contrast, form cytoplasmic inclusion bodies that often behave as liquid-liquid phase-separated condensates. Rabies virus (Rhabdoviridae) forms roughly spherical Negri bodies in infected neurons; these bodies contain the viral nucleoprotein, phosphoprotein, polymerase, and genomic RNA. Respiratory syncytial virus (Pneumoviridae) and Ebola virus (Filoviridae) form similar inclusion bodies. The evidence for LLPS behavior includes: roughly spherical shape under light microscopy, fusion of adjacent inclusion bodies over minutes to hours, dissolution upon treatment with 1,6-hexanediol (an aliphatic alcohol that disrupts weak hydrophobic interactions characteristic of condensates), and fluorescence recovery after photobleaching experiments showing that components exchange between the inclusion body and the cytoplasm. The condensate model proposes that multivalent weak interactions among the nucleoprotein (which coats the RNA), the phosphoprotein (which links nucleoprotein to the polymerase), and the RNA drive phase separation, and that the resulting condensate concentrates replication components while excluding cytoplasmic proteins that do not interact with condensate components.

The material state of a viral factory has functional consequences. Liquid-like condensates allow rapid diffusion of components within the factory, facilitating the assembly of replication complexes and the exchange of templates and products. More solid-like or gel-like states may provide greater protection against host defenses but may slow down replication or make product release less efficient. Ebola virus inclusion bodies provide an interesting case: they contain a population of dynamic, liquid-like condensates but also can mature into more static, gel-like structures over the course of infection, and the balance between these states may determine whether replication is productive or stalled. The coronavirus nucleocapsid protein, which coats the genomic RNA, also undergoes LLPS with RNA in vitro, and this property has been proposed to contribute to the formation of replication-transcription complexes or assembly sites, although the in vivo relevance remains under investigation.

Viral RNP condensation must be evaluated in the context of infection. High viral-protein concentration, multimeric nucleoprotein-RNA filaments, long viral RNAs, phosphorylation, membrane surfaces, and host factors can all change assembly. [Chapter 58](chapter1053.md) defines the general physical models, saturation behavior, alternative mechanisms, and evidence standards. The virus-specific question here is whether perturbing a proposed material-state determinant changes factory organization and viral replication without merely disrupting protein folding, RNA binding, membrane remodeling, or expression.

The relationship between membrane-bound replication organelles and membraneless condensates is not an either/or dichotomy. Some positive-strand RNA virus replication complexes may contain condensate-like protein-RNA clusters on the cytoplasmic face of the membrane. The assembly of flavivirus capsid protein on lipid droplets has properties of a phase transition. Negative-strand RNA virus inclusion bodies can associate with membranes, and the interplay of membrane surfaces with phase separation is an active research frontier. The term "replication organelle" is used in this chapter as an inclusive descriptor for any virus-induced compartment dedicated to replication, regardless of whether it is membrane-enclosed, phase-separated, or a hybrid.

The coupling of replication to assembly is a key functional requirement. In many viruses, newly synthesized genomic RNA must be captured by structural proteins before it is degraded, translated, or used as a template for further replication. The spatial organization of the replication factory is thought to facilitate this handoff: replication occurs at one site within the factory, newly synthesized RNA is translocated to a distinct subregion, and coat or nucleocapsid proteins co-condense with the RNA to initiate assembly. In influenza virus, the viral RNPs are exported from the nucleus after replication and assembly, with the matrix protein M1 and the nuclear export protein NS2 mediating the transition from nuclear RNP to cytoplasmic assembly site at the plasma membrane. The physical chemistry of RNP transport -- how an RNP is handed from one protein complex to another while maintaining its folded state -- is poorly characterized and likely involves nucleic acid chaperone activity.

"All viral replication organelles are liquid-liquid phase-separated condensates." The evidence for LLPS is strong for some negative-strand RNA virus inclusion bodies but largely absent for the membrane-bound spherules and double-membrane vesicles of positive-strand RNA viruses. Even for inclusion bodies that show LLPS properties, the demonstration that phase separation is required for replication typically requires mutation of specific interaction domains and correlation of condensation with replication competence, which has been performed for only a subset of systems.

"Phase separation of viral proteins in vitro proves that viral factories are condensates in vivo." In vitro phase separation demonstrates capacity under the tested conditions. A virus-specific cellular claim requires orthogonal structural and dynamic evidence, separation-of-function perturbations, and a replication phenotype while controlling for expression, RNA binding, membrane integrity, and toxicity; [Chapter 58](chapter1053.md) treats the complete evidence hierarchy.

Include cellular phase separation principles from [Chapter 58](chapter1053.md) for cross-reference.

## 118.3. ssRNA packaging physics and capsid-genome interactions

How does a viral capsid select its own genome from the sea of host RNAs, subgenomic RNAs, defective RNAs, and replication intermediates present in an infected cell? This is the packaging selectivity problem. The mechanisms fall into two broad categories: packaging-signal-dependent packaging and charge-neutralization packaging. Most viruses use a combination of both, with one mechanism predominant.

Packaging-signal-dependent packaging uses specific RNA sequences or structures, called packaging signals or Psi elements, that are recognized by viral structural proteins. The best-characterized examples come from retroviruses, bacteriophages, and some positive-strand RNA plant viruses. In HIV-1, the Psi element is a structured region in the 5′ untranslated region of the genomic RNA that contains several stem-loops, including SL1 (the dimerization initiation site), SL2, SL3, and SL4. The Gag nucleocapsid domain, which contains two CCHC zinc finger motifs, recognizes the Psi element through a combination of structure-specific and sequence-specific contacts. Mutations that disrupt Psi reduce genome packaging by 10- to 100-fold, and deletion of the zinc fingers drastically reduces selective packaging while still allowing nonspecific RNA incorporation. The Psi element also functions as a dimerization signal: two copies of the genomic RNA are brought together through a kissing-loop interaction between SL1 sequences before or during packaging, ensuring that the virion contains two copies of the genome, a defining feature of retroviruses.

The bacteriophage MS2 (Leviviridae) packaging system is the best-understood example of a discrete, high-affinity packaging signal. The MS2 coat protein binds as a dimer to a specific 19-nucleotide RNA stem-loop with sub-nanomolar affinity. The genomic RNA contains a single copy of this operator hairpin, and binding of a coat protein dimer initiates assembly of the T=3 icosahedral capsid around the RNA. The operator-coat protein interaction has been characterized by X-ray crystallography, NMR, and extensive mutagenesis: specific hydrogen bonds between the coat protein and the RNA bases in the loop region provide sequence specificity, while electrostatic interactions with the backbone provide affinity. The MS2 system is a canonical example of a packaging-signal-driven assembly in which a single high-affinity site nucleates capsid assembly on the cognate RNA.

Bacteriophage phi6 (Cystoviridae), a double-stranded RNA virus, packages its three genomic segments (S, M, L) into a preformed procapsid using an ATP-driven packaging motor. The procapsid contains a hexameric P4 packaging ATPase at each of its twelve vertices. The P4 motor recognizes specific packaging signals near the 5′ ends of the single-stranded plus-sense precursor RNAs and translocates them into the procapsid in a defined order -- S segment first, then M, then L -- consuming ATP. After all three segments are packaged, minus-strand synthesis occurs inside the capsid, converting the single-stranded precursors into double-stranded genomic RNA. This system illustrates packaging signal recognition coupled to an ATP-driven translocation motor, a strategy shared with many double-stranded DNA bacteriophages and herpesviruses.

Charge-neutralization packaging operates through a fundamentally different mechanism. In many small positive-strand RNA viruses, the interior surface of the capsid is lined with basic residues, and assembly is driven by electrostatic interaction between this polycationic interior and the RNA polyanion. The capsid interior has a finite electrostatic capacity: RNA that is shorter than the genome does not provide enough negative charge to stabilize the closed capsid, and RNA that is much longer cannot be accommodated without strain or cannot be fully neutralized. Under these conditions, the virus does not need a specific packaging signal because the genome-length RNA is the optimal ligand for assembly. Cowpea chlorotic mottle virus (Bromoviridae) is the classic example: its capsid protein can assemble in vitro around a wide variety of RNAs, including heterologous viral RNAs and synthetic polyanions, as long as the RNA is within a certain length range and the assembly conditions (pH, ionic strength, protein-to-RNA ratio) are appropriate. At pH 7 and moderate ionic strength, assembly requires RNA; at low pH and high ionic strength, the coat protein can form empty capsids. This pH dependence is thought to mimic the in vivo regulation of assembly: the coat protein is maintained in an assembly-incompetent state at neutral pH in the cytoplasm, and the conditions inside the viral factory -- possibly lower pH or higher coat protein concentration -- trigger assembly selectively on newly replicated genomic RNA.

![Figure 118.3. Two strategies for selective ssRNA genome packaging](../assets/figures/chapter1112_figure3.png)

**Figure 118.3. Two strategies for selective ssRNA genome packaging.** Single-stranded RNA genomes can be selected by discrete packaging signals, distributed electrostatic neutralization, or hybrid strategies; HIV-1, bacteriophage MS2, and cowpea chlorotic mottle virus illustrate sequence-specific nucleation and charge-driven packaging.

The distinction between packaging-signal and charge-neutralization mechanisms is not absolute. Many viruses use electrostatic interactions as the driving force for assembly while also using packaging signals to bias assembly toward genomic RNA. Brome mosaic virus, often cited as a charge-neutralization packager, does preferentially package its own RNA over heterologous RNAs in plant infections, suggesting that additional selectivity mechanisms operate in vivo even though they are not evident in simplified in vitro systems. In hepatitis B virus, a reverse-transcribing DNA virus whose pregenomic RNA is packaged, the viral polymerase binds a specific stem-loop (epsilon) on the pregenomic RNA, and this interaction nucleates both packaging and reverse transcription. The electrostatic contributions of the core protein C-terminal arginine-rich domain are also required for RNA packaging, illustrating a hybrid mechanism.

The RNA secondary structure inside the capsid is a distinct structural state. In solution, viral genomic RNA adopts an ensemble of conformations determined by the base-pairing free energies, the ionic conditions, and the presence of bound proteins. Inside the capsid, the high protein concentration (often 200-400 mg/mL in the capsid interior), the molecular crowding, the neutralization of backbone charge, and the specific protein-RNA contacts all alter the RNA folding landscape. The in-capsid RNA structure has been probed by chemical modification inside virions, by crosslinking, and by cryo-EM reconstruction that resolves RNA density. In bacteriophage MS2, the RNA adopts a defined conformation in which the operator hairpin is bound at each of the 90 coat protein dimers, and the remaining RNA forms an ordered network of contacts with the capsid interior. In flock house virus, cryo-EM reveals a dodecahedral cage of RNA duplexes lining the capsid interior, suggesting that the RNA forms a semi-ordered shell beneath the protein capsid. In HIV-1, the RNA structure inside the mature capsid core is less ordered, consistent with the nucleocapsid protein's function as a nucleic acid chaperone that prevents the RNA from being trapped in nonfunctional folds.

How do viruses package a genome of exactly the right length? For packaging-signal systems, the genome contains a defined number and distribution of signals, and the capsid geometry and assembly stoichiometry ensure that only RNA molecules that present the correct number of signals can complete assembly. For charge-neutralization systems, length selectivity is an emergent property of the electrostatic interactions between the RNA and the capsid interior, combined with the capsid's finite volume. In vitro assembly experiments with cowpea chlorotic mottle virus coat protein show that RNA length strongly influences assembly efficiency: short RNAs form small, non-icosahedral particles or fail to assemble, RNAs near the native genome length form regular T=3 particles with high efficiency, and very long RNAs form multiparticle assemblies or large aberrant structures. The length selectivity of charge-neutralization systems is statistical rather than absolute, which may explain why these viruses sometimes package subgenomic or defective RNAs at detectable frequencies.

The packaging of segmented genomes adds another layer of complexity. Influenza A virus packages eight distinct RNPs, and the virus must incorporate at least one copy of each segment to produce infectious particles. Electron tomography of influenza virions shows that the eight RNPs are arranged in a characteristic "7+1" pattern, with seven RNPs surrounding a central one. How do the eight segments find each other? The current model proposes that each segment carries a distinct set of packaging signals -- short, conserved RNA sequences at the 3′ and 5′ termini, within the coding regions, or both -- and that intersegment RNA-RNA interactions, mediated by these signals, assemble the complete set of eight RNPs prior to or during budding. Mutations in the packaging signals of one segment can reduce incorporation of that segment and of specific other segments, supporting the intersegment interaction model. The influenza packaging system is therefore a case of packaging-signal-mediated selectivity applied to a segmented genome, where the signals must encode both segment identity and intersegment compatibility.

"All viruses package only one copy of their genome." Retroviruses package two copies (diploid). Paramyxoviruses and filoviruses package a single negative-sense RNA, but the nucleocapsid is helical and the number of genome copies per particle can deviate from one. Influenza packages eight distinct segments, each presumably in one copy in the archetypal virion, but virions with fewer or more than eight segments are produced. The phenomenon of polypoidy -- packaging more than one genome copy -- occurs naturally in some viruses and can be induced in others by altering assembly conditions.

The misconception "charge-neutralization packaging means the virus packages every RNA nonselectively" overlooks length selectivity and additional biological filters. Charge-neutralization mechanisms can be coupled to temporal and spatial compartmentalization, the availability of newly replicated genomic RNA in replication factories, and interactions with replication proteins that discriminate genomic from subgenomic and host RNAs. The degree of selectivity in vivo is often higher than in vitro charge-neutralization experiments would suggest.

## 118.4. Virus-like particles and RNA cargo selection

Virus-like particles are the technological and therapeutic face of viral RNA condensation biology. A VLP is a macromolecular assembly of viral structural proteins that resembles the authentic virion but lacks the infectious genome. VLPs can be produced by expressing one or more viral structural proteins in a heterologous system -- bacteria, yeast, insect cells, plant cells, or mammalian cells -- where they self-assemble into particles. Because VLPs present viral surface antigens in their native conformation and at high density, they are often more immunogenic than soluble protein subunits and have become successful vaccine platforms.

The licensed VLP vaccines illustrate the range of systems. The hepatitis B virus vaccine, first licensed in 1986, is produced by expressing the hepatitis B virus surface antigen (HBsAg) in Saccharomyces cerevisiae; the HBsAg protein self-assembles into 22-nanometer spherical particles that contain host-cell lipids and no nucleic acid. The human papillomavirus vaccines (Gardasil, Cervarix) are produced by expressing the L1 major capsid protein in yeast or insect cells; L1 assembles into T=7 icosahedral VLPs that are morphologically identical to authentic HPV virions. The hepatitis E virus vaccine (Hecolin) uses a truncated capsid protein expressed in E. coli that assembles into VLPs. These vaccines demonstrate that viral structural proteins carry the information for self-assembly, that the assembled VLP structure can be authenticated by cryo-EM and biophysical methods, and that VLPs elicit protective antibody responses without requiring an adjuvant in some cases.

![Figure 118.4. Virus-like particle production platforms and RNA cargo composition](../assets/figures/chapter1112_figure4.png)

**Figure 118.4. Virus-like particle production platforms and RNA cargo composition.** The virus-like-particle expression system determines assembly compartment and incidental RNA cargo, whereas intentional loading by packaging signals, coexpression, or disassembly-reassembly must be quantified by cargo identity, copies per particle, and heterogeneity.

The choice of expression system determines cargo selection. When viral structural proteins are expressed in the cytoplasm of eukaryotic cells, the assembling VLPs can encapsidate host RNA. The amount and identity of packaged RNA depend on the virus family, the expressed proteins, the assembly compartment, and the presence or absence of a packaging signal. Hepatitis B virus core protein expressed in E. coli assembles into capsids that contain bacterial RNA predominantly, because the C-terminal arginine-rich domain binds RNA nonspecifically and the high concentration of RNA in the bacterial cytoplasm drives co-assembly. If the arginine-rich domain is deleted or if assembly is performed in vitro with purified protein and controlled RNA concentrations, the RNA cargo composition can be manipulated. HPV L1 VLPs produced in yeast or insect cells contain little or no nucleic acid because L1 lacks a strong RNA-binding domain and assembly proceeds through a nucleic-acid-independent pathway. Understanding these differences is important for VLP vaccine characterization: regulatory agencies require demonstration that the VLP product is consistent in composition, including its nucleic acid content, batch after batch.

Intentional cargo loading extends VLPs beyond vaccines to gene therapy and targeted delivery. The basic principle is to incorporate a packaging signal into a therapeutic RNA, co-express or co-assemble the RNA with the viral structural proteins, and purify the resulting loaded VLPs. Adeno-associated virus (AAV), a parvovirus used extensively for gene therapy, packages its single-stranded DNA genome into a preformed capsid; AAV vectors are produced by providing the therapeutic DNA flanked by inverted terminal repeats along with the AAV Rep and Cap proteins and helper virus functions. While AAV is a DNA virus, its packaging principles are conceptually similar: a specific cis-acting signal (the inverted terminal repeat) is recognized by the packaging machinery (Rep protein), and the genome is inserted into the capsid. RNA-loaded VLPs have been explored using bacteriophage MS2 coat protein VLPs (with the operator hairpin fused to the cargo RNA), hepatitis B virus core protein VLPs (with the epsilon packaging signal on the cargo), and retroviral Gag VLPs (with the Psi element). Challenges include serum stability, innate immunogenicity, reproducible cargo stoichiometry, and payload limits imposed by capsid volume.

The RNA cargo inside a VLP can affect the VLP's physical properties. RNA acts as a polyanion that contributes to particle stability, and empty capsids are often less stable than RNA-filled capsids to thermal or chemical denaturation. The RNA also contributes to the particle's buoyant density in cesium chloride or iodixanol gradients, allowing separation of empty and full particles -- an important quality-control step in VLP manufacturing. The RNA sequence can affect assembly efficiency: in bacteriophage MS2, the presence of the operator hairpin dramatically increases the rate and yield of capsid assembly compared with non-cognate RNA. These effects mean that cargo RNA is not an inert passenger but an active participant in the assembly thermodynamics and kinetics.

The measurement of RNA cargo identity and quantity in VLPs requires specialized methods. Sedimentation velocity analytical ultracentrifugation can resolve empty capsids from RNA-containing capsids because their sedimentation coefficients differ. Native mass spectrometry can determine the mass of intact VLPs and, by subtraction, the mass of encapsulated RNA. RNase protection assays, followed by extraction and sequencing of the protected RNA, identify which RNA sequences are inside the particle. Quantitative PCR or RT-qPCR measures the amount of specific RNA species. For regulatory characterization, these methods are combined to provide a comprehensive picture of VLP composition, including the size distribution (dynamic light scattering, nanoparticle tracking analysis), the structural integrity (cryo-EM), the antigenic authenticity (monoclonal antibody binding panels), and the nucleic acid cargo.

"VLPs are empty capsids." Many VLPs contain host-cell RNA, sometimes in amounts comparable to the genomic RNA content of authentic virions. The phrase "VLP" describes the absence of an infectious viral genome, not the absence of all nucleic acid. Regulatory documents specify "empty" versus "filled" VLPs, where "empty" means lacking nucleic acid within the detection limit of the characterization method.

"Any RNA can be loaded into any VLP by simply mixing the RNA with the capsid protein." Loading efficiency depends on the presence of a packaging signal, the RNA length, the assembly conditions (pH, ionic strength, protein-to-RNA ratio, temperature), and whether assembly is performed in vitro or in vivo. Many attempts to load arbitrary therapeutic RNAs into VLPs have failed because the RNA lacks a packaging signal, is too short or too long, or assembles into non-native structures.

## 118.5. Antiviral intervention, vaccine design, and measurement of viral material states

The physical chemistry of viral RNA condensation is not only a basic science problem; it is a rational basis for intervention. Assembly inhibitors, packaging inhibitors, and VLP-based vaccines all exploit the thermodynamic and kinetic principles described in the preceding sections. This section connects the biophysics to therapeutic design and to the methods that validate both mechanism and intervention.

Assembly inhibitors are small molecules or peptides that bind to viral capsid or nucleocapsid proteins and block their assembly into functional particles. The mechanism can be thermodynamic (stabilizing an assembly-incompetent conformation, reducing the affinity of protein-protein or protein-RNA interactions) or kinetic (slowing nucleation, capping growing intermediates, inducing misassembly into nonfunctional aggregates). The two most clinically advanced examples are HIV-1 capsid inhibitors and hepatitis B virus core protein allosteric modulators.

Lenacapavir (formerly GS-6207) is a first-in-class HIV-1 capsid inhibitor that binds at the interface between capsid hexamers and pentamers, stabilizing the assembled capsid and preventing both early uncoating and late assembly. It binds with picomolar affinity to the capsid protein and has a long pharmacokinetic half-life allowing dosing every six months. Lenacapavir is mechanistically distinct from most antivirals, which target enzymes (protease, reverse transcriptase, integrase), and demonstrates that targeting the structural biology of capsid assembly is a viable drug strategy. Its success has reinvigorated interest in assembly-targeting antivirals across virus families.

Hepatitis B virus core protein allosteric modulators represent a different assembly intervention strategy. Class I modulators bind the core protein dimer-dimer interface and induce the formation of aberrant, non-capsid polymers or empty capsids that lack pregenomic RNA. Class II modulators (sulfamoylbenzamides) bind a different pocket and accelerate capsid assembly into morphologically normal but empty capsids. Both classes reduce infectious virus production, and several have entered clinical trials. The hepatitis B virus example illustrates the principle that assembly can be inhibited by misdirecting it -- driving the capsid protein into nonproductive assembly pathways rather than simply blocking assembly.

Packaging inhibitors target the interaction between the viral structural protein and the genomic RNA, preventing selective genome incorporation without necessarily blocking capsid assembly. The HIV-1 nucleocapsid zinc fingers are the best-studied packaging-inhibitor target: compounds that eject zinc from the CCHC motifs, such as disulfide benzamides and certain electrophilic agents, abolish selective packaging of the HIV-1 genome and produce noninfectious particles that still bud from cells but contain predominantly host RNA. Although nucleocapsid inhibitors have not yet produced a licensed drug, the concept is mechanistically validated. The challenge is that the nucleocapsid zinc fingers are small, solvent-exposed, and part of a flexible domain that is difficult to target with high specificity. Ribonuclease-targeting chimeras and proteolysis-targeting chimeras that degrade the nucleocapsid protein or the capsid protein have been proposed as alternative strategies to block assembly by eliminating the structural protein rather than inhibiting its function.

![Figure 118.5. Mechanisms of assembly-targeting antiviral interventions](../assets/figures/chapter1112_figure5.png)

**Figure 118.5. Mechanisms of assembly-targeting antiviral interventions.** Assembly-targeting antivirals disrupt different steps: lenacapavir alters HIV-1 capsid dynamics, hepatitis B core modulators misdirect capsid assembly, and zinc-finger disruption prevents selective genome packaging even when particles still bud.

VLP vaccine design exploits the self-assembly properties of viral structural proteins. The key design decisions include: which structural protein or proteins to express; whether to include the envelope glycoproteins (for enveloped-virus VLPs); which expression system to use; how to purify the VLPs from host-cell contaminants including host nucleic acids; how to characterize particle integrity, antigenicity, and consistency; and whether to include an adjuvant. For enveloped VLPs such as those based on influenza HA and NA or SARS-CoV-2 spike, the glycoproteins are displayed on a lipid bilayer that may bud from the expressing cell or may be assembled in vitro. The structure and stability of VLP-displayed antigens can be assessed by cryo-EM, by monoclonal antibody binding panels that probe conformational epitopes, and by hydrogen-deuterium exchange mass spectrometry that measures protein dynamics.

The convergence of VLP technology with structure-based immunogen design offers new vaccine strategies. The principle of "germline targeting," developed for HIV-1 broadly neutralizing antibody induction, involves designing a series of immunogens -- starting with a VLP or nanoparticle that binds unmutated B cell receptors, progressing through intermediate immunogens, and ending with a native-like trimer -- to guide the antibody response through somatic hypermutation toward broadly neutralizing antibodies. The structural biology of these immunogens, including their RNA cargo and its effect on immunogen conformation and stability, must be characterized to ensure that each immunogen in the series maintains the correct epitope presentation. The role of RNA cargo in VLP immunogenicity is an underexplored but potentially important variable: RNA inside a VLP can activate Toll-like receptors 7 and 8 if the VLP is taken up by endosomes and partially disassembled, providing a built-in adjuvant effect.

Measurement methods for viral RNA condensation, packaging, and VLP characterization form a toolkit that spans structural biology, biophysics, biochemistry, and virology. Sedimentation velocity analytical ultracentrifugation resolves capsid assembly states, empty from RNA-filled particles, and soluble protein from assembled particles based on differences in sedimentation coefficient. Cryo-electron microscopy and cryo-electron tomography provide near-atomic resolution structures of VLPs and viral cores, and cryo-ET of intact virions can resolve the in situ arrangement of RNPs. Hydrogen-deuterium exchange mass spectrometry measures the dynamics and solvent accessibility of capsid proteins in different assembly states and in the presence or absence of RNA. Native mass spectrometry measures the mass of intact capsids and VLPs, distinguishing different oligomeric states and revealing the stoichiometry of bound RNA. Small-angle X-ray scattering and small-angle neutron scattering provide low-resolution structural information on the overall shape and internal density distribution of capsids and RNPs in solution. Fluorescence correlation spectroscopy and single-particle tracking measure the size, concentration, and dynamics of assembling particles in real time. Isothermal titration calorimetry measures the thermodynamics of protein-RNA interactions that drive assembly. Optical tweezers and magnetic tweezers measure the forces generated by packaging motors at the single-molecule level.

**Table 118.1. Biophysical methods for characterizing viral particles, assembly, and packaging.** - **Rows (methods to include):** 1. Sedimentation velocity analytical ultracentrifugation (SV-AUC): sedimentation coefficient distribution, 100-500 uL at 0.1-1 mg/mL, resolves empty/full capsids and assembly intermediates, Johnston-Ogston effect and nonideality at high concentration, assembly state and cargo analysis. 2. Cryo-electron microscopy (cryo-EM) single-particle analysis: 3D structure at near-atomic resolution, ~3 uL at 0.1-5 mg/mL, atomic model of capsid and ordered RNA density, heterogeneity and flexible RNA limit resolution, capsid and RNP structure. 3. Cryo-electron tomography (cryo-ET): 3D structure of pleomorphic or unique particles in situ, ~3 uL at physiological concentration in thin layer, native arrangement of RNPs inside virions, missing wedge and low signal-to-noise ratio, influenza RNP organization, HIV-1 core structure. 4. Native mass spectrometry: mass of intact capsids and VLPs, ~5 uL at 0.1-1 mg/mL in volatile buffer, stoichiometry of protein and RNA components, desolvation can dissociate weak interactions, cargo stoichiometry. 5. Hydrogen-deuterium exchange mass spectrometry (HDX-MS): protein backbone dynamics and solvent accessibility, ~50 uL at 0.1-1 mg/mL per time point, maps assembly-induced changes in protein dynamics, back-exchange correction and peptide coverage limitations, conformational changes during assembly. 6. Small-angle X-ray scattering (SAXS) / Small-angle neutron scattering (SANS): overall shape, radius of gyration, internal density distribution, ~30 uL at 1-10 mg/mL, low-resolution envelope and internal RNA density, radiation damage (SAXS), requires contrast matching (SANS), solution structure of capsids and RNPs. 7. Isothermal titration calorimetry (ITC): binding affinity, stoichiometry, and enthalpy of protein-RNA interaction, ~300 uL protein + 100 uL RNA at 10-100 uM, thermodynamics of the binding event, requires purified components, binding cannot be too tight or too weak, protein-RNA interaction thermodynamics. 8. Optical tweezers / Magnetic tweezers: force and step size of packaging motors at single-molecule level, single DNA/RNA molecule tethered between beads, force-velocity relationship, step size, and stall force, surface attachment artifacts, limited throughput, packaging motor mechanism. 9. Dynamic light scattering (DLS) / Nanoparticle tracking analysis (NTA): particle size distribution and concentration, ~50 uL at 0.01-1 mg/mL, hydrodynamic radius and polydispersity, intensity-weighted (DLS), low resolution for mixtures (DLS), VLP size and aggregation quality control. 10. Fluorescence correlation spectroscopy (FCS): diffusion coefficient and concentration of fluorescently labeled particles, ~20 uL at nanomolar concentration, hydrodynamic radius and particle brightness, photobleaching, labeling can perturb assembly, assembly kinetics in real time.

| Method | Measured property | Approximate sample requirements | Information content | Key artifacts and limitations | Typical application in viral assembly research |
| --- | --- | --- | --- | --- | --- |
| **Sedimentation velocity analytical ultracentrifugation (SV-AUC)** | Sedimentation coefficient distribution | 100-500 uL at 0.1-1 mg/mL | Resolves empty and full capsids and assembly intermediates | Johnston-Ogston effect and nonideality at high concentration | Assembly state and cargo analysis |
| **Cryo-electron microscopy (cryo-EM) single-particle analysis** | 3D structure near atomic resolution | About 3 uL at 0.1-5 mg/mL | Atomic capsid model and ordered RNA density | Heterogeneity and flexible RNA limit resolution | Capsid and ribonucleoprotein (RNP) structure |
| **Cryo-electron tomography (cryo-ET)** | 3D structure of pleomorphic or unique particles in situ | About 3 uL at physiological concentration in a thin layer | Native RNP arrangement inside virions | Missing wedge and low signal-to-noise | Influenza RNP organization and HIV-1 core architecture |
| **Native mass spectrometry** | Mass of intact capsids and virus-like particles (VLPs) | About 5 uL at 0.1-1 mg/mL in volatile buffer | Protein/RNA stoichiometry | Desolvation may dissociate weak interactions | Cargo stoichiometry |
| **Hydrogen-deuterium exchange mass spectrometry (HDX-MS)** | Protein backbone dynamics and solvent accessibility | About 50 uL at 0.1-1 mg/mL per time point | Maps assembly-induced dynamics | Back-exchange and peptide coverage limit spatial detail | Conformational changes during assembly |
| **Small-angle X-ray scattering (SAXS) / small-angle neutron scattering (SANS)** | Shape, radius of gyration, and internal density | About 30 uL at 1-10 mg/mL | Envelope and RNA-density information | Radiation damage and contrast-matching assumptions | Solution structure |
| **Isothermal titration calorimetry (ITC)** | Affinity, stoichiometry, and enthalpy | About 300 uL protein plus 100 uL RNA at 10-100 uM | Thermodynamics of RNA-protein binding | Purified components and affinity-window limitations | Capsid-RNA binding energetics |
| **Optical or magnetic tweezers** | Force and step size at single-molecule level | Single DNA/RNA tethered between beads | Force-velocity curves, step size, and stall force | Surface artifacts and low throughput | Packaging motor mechanism |
| **Dynamic light scattering / nanoparticle tracking analysis (DLS/NTA)** | Particle size and concentration | About 50 uL at 0.01-1 mg/mL | Hydrodynamic radius and polydispersity | Intensity weighting and poor resolution of complex mixtures | Virus-like-particle quality control |
| **Fluorescence correlation spectroscopy (FCS)** | Diffusion coefficient and concentration | About 20 uL nanomolar fluorescent particles | Hydrodynamic radius and molecular brightness | Photobleaching and label perturbation | Real-time assembly kinetics |

**Table 118.1. Biophysical Methods for Characterizing Viral Particles, Assembly, and Packaging**

| Method | Measured property | Sample requirements | Information content | Key artifacts and limitations | Typical application |
|---|---|---|---|---|---|
| Sedimentation velocity AUC | Sedimentation coefficient distribution | 100–500 µL at 0.1–1 mg/mL | Resolves empty/full capsids and assembly intermediates | Johnston-Ogston effect; nonideality at high concentration | Assembly state and cargo analysis |
| Cryo-EM single-particle analysis | 3D structure at near-atomic resolution | ~3 µL at 0.1–5 mg/mL | Atomic model of capsid and ordered RNA density | Heterogeneity and flexible RNA limit resolution | Capsid and RNP structure |
| Cryo-electron tomography (cryo-ET) | 3D structure of pleomorphic particles in situ | ~3 µL at physiological concentration | Native arrangement of RNPs inside virions | Missing wedge; low signal-to-noise ratio | Influenza RNP organization; HIV-1 core structure |
| Native mass spectrometry | Mass of intact capsids and VLPs | ~5 µL at 0.1–1 mg/mL in volatile buffer | Stoichiometry of protein and RNA components | Desolvation can dissociate weak interactions | Cargo stoichiometry |
| HDX-MS | Protein backbone dynamics and solvent accessibility | ~50 µL at 0.1–1 mg/mL per time point | Maps assembly-induced changes in protein dynamics | Back-exchange correction; peptide coverage limitations | Conformational changes during assembly |
| SAXS / SANS | Overall shape, radius of gyration, internal density | ~30 µL at 1–10 mg/mL | Low-resolution envelope and internal RNA density | Radiation damage (SAXS); requires contrast matching (SANS) | Solution structure of capsids and RNPs |
| Isothermal titration calorimetry (ITC) | Binding affinity, stoichiometry, enthalpy | ~300 µL protein + 100 µL RNA at 10–100 µM | Thermodynamics of the binding event | Requires purified components; binding cannot be too tight or weak | Protein-RNA interaction thermodynamics |
| Optical tweezers / Magnetic tweezers | Force and step size of packaging motors | Single DNA/RNA molecule tethered between beads | Force-velocity relationship, step size, stall force | Surface attachment artifacts; limited throughput | Packaging motor mechanism |
| Dynamic light scattering (DLS) / NTA | Particle size distribution and concentration | ~50 µL at 0.01–1 mg/mL | Hydrodynamic radius and polydispersity | Intensity-weighted (DLS); low resolution for mixtures | VLP size and aggregation quality control |
| Fluorescence correlation spectroscopy (FCS) | Diffusion coefficient and concentration | ~20 µL at nanomolar concentration | Hydrodynamic radius and particle brightness | Photobleaching; labeling can perturb assembly | Assembly kinetics in real time |

Each measurement method has assumptions and artifacts that must be understood to avoid misinterpretation. Analytical ultracentrifugation at high protein concentrations can be perturbed by nonideality and by the Johnston-Ogston effect, in which slower-sedimenting species are artificially elevated because they are pushed forward by faster-sedimenting species. Cryo-EM structure determination requires that the particle population be structurally homogeneous; heterogeneous assembly intermediates or flexible RNA can blur density and prevent high-resolution reconstruction, so the structures obtained may represent only the most ordered subpopulation. Native mass spectrometry requires volatile buffers, typically ammonium acetate at neutral pH, and the desolvation process can dissociate weakly bound components; the mass spectrum reflects the species that survive the electrospray process, which may differ from the species in solution. Hydrogen-deuterium exchange reports on backbone amide solvent accessibility, but the exchange kinetics must be corrected for the intrinsic exchange rate of each amide, which depends on the neighboring amino acid sequence, pH, and temperature.

The FDA and EMA regulatory framework for VLP-based products requires specific characterization assays beyond those used in academic structural biology. Identity testing (SDS-PAGE, western blot, mass spectrometry of digested protein), purity testing (host-cell protein ELISA, host-cell DNA qPCR, residual plasmid DNA), potency testing (antigen content by ELISA, in vitro or in vivo immunogenicity), and consistency testing (particle size distribution, empty-to-full ratio, thermal stability by differential scanning calorimetry or differential scanning fluorimetry) are all required. Understanding these requirements is essential for translating VLP discovery into licensed vaccines.

The misconception "assembly inhibitors work by preventing the capsid protein from binding RNA" is too narrow. Assembly inhibitors can act at protein-protein interfaces, at protein-RNA interfaces, or by stabilizing nonproductive conformations. Some hepatitis B virus core protein allosteric modulators accelerate rather than inhibit assembly, driving the protein into empty capsids that are assembly products but not functional virions. The discovery of lenacapavir, which stabilizes rather than prevents the assembled capsid, further illustrates that "assembly inhibitor" is a functional category, not a mechanistic one.

## Experimental Foundations and Evidence Standards

The evidence for viral RNA condensation mechanisms comes from four integrated approaches. In vitro reconstitution with purified components establishes sufficiency: if coat protein and RNA assemble into infectious particles in a test tube, as demonstrated for TMV in 1955, then no other factors are strictly required. Mutagenesis of protein basic domains, RNA packaging signals, or protein-protein interfaces shows which interactions are necessary for assembly in infected cells or in in vitro systems. Structural biology -- cryo-EM, X-ray crystallography, and NMR -- reveals the atomic contacts between protein and RNA and the overall architecture of the assembled particle. Biophysical measurements -- analytical ultracentrifugation, calorimetry, and single-molecule force spectroscopy -- quantify the thermodynamics, kinetics, and forces of assembly and packaging.

Evidence for phase behavior in viral factories is more recent and less complete. In vitro phase separation demonstrates capacity, while live-cell morphology, fusion, exchange, or chemical sensitivity provides suggestive but individually non-diagnostic evidence. Stronger causal tests use orthogonal measurements and separation-of-function perturbations that change factory organization and replication while controlling for protein abundance, folding, RNA binding, membrane integrity, and toxicity. [Chapter 58](chapter1053.md) supplies the general evidence framework.

The evidence for cargo selection in VLPs is primarily compositional. Sequencing of RNA extracted from VLPs identifies the cargo. Manipulating the RNA -- adding or removing a packaging signal, changing RNA length, altering the assembly conditions -- shows how cargo composition responds. Regulatory characterization requires demonstrating that the cargo is consistent across batches and that the identity and amount of nucleic acid do not affect the product's safety or efficacy profile. These are manufacturing rather than mechanistic requirements, but they rely on the same analytical methods.

Interpretive hazards recur in this literature. One hazard is conflating in vitro assembly conditions with in vivo assembly mechanisms. Many viral coat proteins assemble under a wider range of conditions in vitro than they encounter in cells, and the fact that a protein can assemble without RNA at low pH does not prove that such empty capsids form in vivo. A second hazard is attributing packaging selectivity solely to a characterized packaging signal when other mechanisms (replication-assembly coupling, compartmentalization, RNA length discrimination) also contribute. A third hazard is overinterpreting cryo-EM density attributed to RNA. RNA inside a capsid is often flexible and does not resolve to high resolution; density features attributed to RNA organization can be influenced by the averaging and symmetry imposition inherent in single-particle cryo-EM reconstruction.

## Biological Contexts and Cross-Chapter Boundaries

This chapter connects viral RNA structure ([Chapter 117](chapter1111.md)) to viral particle biology and antiviral intervention. The packaging signals described in [Chapter 117](chapter1111.md) as RNA structural elements are treated here as physical determinants of selective condensation. The replication organelles described in [Chapter 116](chapter1110.md) as membrane-bound compartments are treated here as material states with phase behavior. The viroids covered in [Chapter 119](chapter1113.md) present an interesting boundary case: viroids are unencapsidated, infectious single-stranded circular RNAs that do not encode proteins. Their replication does not involve capsid assembly or packaging, but their pathogenesis may involve RNA-protein condensation in the host, and the evolutionary relationship between viroids and viral capsids is an open question.

Host stress granules and P-bodies can interact with viral factories, but their assembly, remodeling, and manipulation by viruses belong to [Chapter 105](chapter1100.md). This chapter follows viral proteins or RNAs only after they enter a virus-specific replication, assembly, or packaging state. Nuclear RNA bodies remain in [Chapter 95](chapter1090.md), and general condensate physics remains in [Chapter 58](chapter1053.md).

VLP findings can inform comparisons with synthetic delivery particles, but mRNA design and lipid nanoparticle composition, assembly, trafficking, and delivery belong to [Chapter 153](chapter1137.md) and [Chapter 156](chapter1139.md). Here VLPs are treated as virus-derived assemblies whose cargo selection and analytical control follow capsid-specific rules.

Retroviruses ([Chapter 120](chapter1114.md)) are treated here primarily for their packaging signal and Gag-mediated assembly; the reverse transcriptase enzymology and integration biology are outside this chapter's scope. The retroviral nucleocapsid protein's nucleic acid chaperone activity is central to selective packaging and genome dimerization and is a model for nucleic acid chaperone function in other virus families.

## Recent Consensus

Several points are established consensus: (1) viral RNA condensation is driven primarily by electrostatic interactions between basic protein domains and the RNA phosphate backbone, with counterion release providing the dominant entropic driving force; (2) some viruses, notably retroviruses and bacteriophages, use specific packaging signals for selective genome incorporation, while others, notably many small icosahedral plant viruses, use charge-neutralization packaging with limited sequence specificity; (3) the material state of negative-strand RNA virus inclusion bodies includes liquid-liquid phase separation properties, while the replication organelles of positive-strand RNA viruses are predominantly membrane-bound spherules and double-membrane vesicles; (4) VLPs are established vaccine platforms with strong safety and immunogenicity profiles demonstrated by licensed products; (5) targeting capsid assembly with small molecules is a validated antiviral strategy, as shown by lenacapavir for HIV-1 and core protein allosteric modulators for hepatitis B virus.

Recent advances have expanded these consensuses. Single-molecule force spectroscopy with optical tweezers has directly measured the forces generated by bacteriophage packaging motors, including phi29 and phi6, providing quantitative models for ATP-driven genome translocation. Cryo-ET of intact virions has revealed the in situ organization of influenza RNPs and HIV-1 cores, connecting in vitro structural models to native particle architecture. Hydrogen-deuterium exchange mass spectrometry has mapped the dynamic changes in capsid protein conformation during assembly, identifying regions that become ordered or disordered upon RNA binding and capsid closure.

## Open Questions, Controversies, Deprecated Models, and Common Misconceptions

Open questions:

- Is the liquid-liquid phase separation of viral proteins in vitro and in cells functionally required for replication or is a nonfunctional consequence of high protein concentration and multivalent interactions? For some systems, notably rabies virus Negri bodies, the correlation between phase separation propensity and replication competence has been tested by mutagenesis. For most claimed viral condensates, functional validation is incomplete.
- What is the mechanism of RNA length determination in charge-neutralization packaging systems? The electrostatic capacity model successfully predicts that assembly efficiency is RNA-length-dependent, but it does not fully explain the sharp length optimum observed for some viruses. Contributions from RNA secondary structure, from the kinetics of nucleation and elongation, and from the mechanical properties of the partially assembled capsid may all contribute but are not yet quantitatively integrated.
- What is the role of the packaged RNA cargo in VLP immunogenicity? RNA inside VLPs can act as a TLR7/8 agonist if the VLP is endocytosed and partially disassembled, but whether this contributes to the superior immunogenicity of VLPs over soluble protein subunits in humans is not definitively established. The answer matters for vaccine design because it could inform whether to include, exclude, or modify the RNA cargo.

Controversies:

- A fourth area of active controversy is whether assembly-targeting antivirals will be broadly useful or will remain niche therapies. Lenacapavir's success for HIV-1 is clear, and hepatitis B virus core protein allosteric modulators are advancing, but assembly inhibitors have not yet been approved for any acute viral infection. The concern is that assembly inhibitors, which act late in the viral life cycle, may be less effective than polymerase inhibitors at rapidly reducing viral load, and that the genetic barrier to resistance may be lower if the targeted protein-protein or protein-RNA interface can accommodate mutations without losing function.

Common misconceptions:

- "Viral RNA condensation is the same as DNA condensation." Viral RNA condensation depends on RNA length, structure, protein interactions, replication timing, and particle architecture.
- "All viral replication organelles are phase-separated condensates." Some viral factories show condensate-like behavior, but others are membrane-bound, scaffolded, or actively maintained assemblies.
- "VLPs are empty capsids that contain no RNA." VLPs can package host, vector, or designed RNA depending on expression system and assembly mechanism.
- "Any RNA can be loaded into any VLP." Packaging depends on length, structure, charge, packaging signals, assembly pathway, and production context.
- "Assembly inhibitor is a single mechanism." Assembly-targeting antivirals can affect protein-protein interfaces, protein-RNA interfaces, conformational switching, or capsid stability.