The team, led by Morgan Williamson, Caleb Boren, Rahul Kumar and Lauren Gollahon, fused the two structural proteins of HPV16 to the two halves of a yellow fluorescent protein called Venus. The technique, known as bimolecular fluorescence complementation or BiFC, splits a fluorescent protein into an N-terminal and a C-terminal fragment, neither of which emits light on its own. When the two fragments are brought together by proteins that interact with each other, the fluorophore reconstitutes and fluoresces. By attaching the C-terminal half of Venus to L1 and the N-terminal half to L2, the researchers ensured that only particles containing both structural proteins would produce a detectable YFP signal, providing a built-in reporter of productive capsid assembly.

A key design challenge was to attach the fluorophore halves without straining the capsid architecture. L1 assembles into pentamers that interlock through conserved disulfide bonds between the residues Cys175 and Cys428, forming an icosahedral shell roughly 50 to 60 nanometers in diameter. The native capsid contains approximately 360 copies of L1 along with 12 to 36 copies of the minor protein L2, which sits between pentamers and slightly tightens the shell. The researchers designed a specialized linker with the sequence HPPAPKED to join the fluorophore fragments to the capsid proteins, and structural predictions generated with AlphaFold2 and visualized in ChimeraX suggested the linker allowed the fluorophore halves to approach closely without obstructing the tertiary and quaternary contacts needed for assembly.

The constructs were built using overlap extension PCR, with the L1 and L2 coding sequences copied from the established p16sheLL expression vector and the Venus fragments taken from standard BiFC plasmids. Notably, the new coding regions were placed under a general CMV promoter rather than the SV40 large T-dependent system used in most published HPV16 VLP protocols, which achieves high plasmid copy numbers only in the specialized 293TT producer cell line. By using ordinary 293T cells instead, the authors argue, the platform becomes accessible to far more laboratories. The change also sidesteps recent concerns about producer cells expressing large T antigen, including the possibility of spontaneous packaging of large T coding sequences and the delivery of oncogenic DNA into healthy cells, a risk the team directly tested and did not observe.

Expression of the fusion proteins matched theoretical predictions. Immunoblotting of transfected 293T cells detected L1 at around 60 kilodaltons and L2 at around 70 kilodaltons, consistent with the calculated molecular weights of the L1-VC155 and L2-VN173 fusion proteins. Flow cytometry tracked the BiFC signal over time and found peak intensity at approximately 72 hours after transfection, when roughly 17.6 percent of the total cell population produced detectable YFP, indicating successful co-uptake of both plasmids and intracellular capsid assembly. Live-cell imaging revealed the same cell swelling that earlier researchers famously described as producer cells filling up like beach balls, and the fluorescent signal was visible across multiple microscope platforms and filter sets, from dedicated YFP filters down to standard FITC channels.

Particles were harvested by alkaline lysis with ammonium sulfate, which allows capsid maturation as disulfide bonds form between L1 subunits over 24 to 36 hours, and then purified on a discontinuous iodixanol density gradient by ultracentrifugation. The BiFC particles banded at the expected position, matching the density of the positive control p16sheLL particles, and transmission electron microscopy confirmed distinctive pentamer formation and icosahedral morphology. Dynamic light scattering placed the particles near 50 nanometers, in agreement with the electron micrographs, and showed less size variance than the conventional construct, which displayed multiple peaks including one near 4,000 nanometers corresponding to large aggregates.

That reduction in aggregation is one of the study’s most practical findings. Conventional L1-only VLPs are notorious for clumping into mega-particle assemblies that complicate purification, measurement, and formulation, and the largest aggregate observed in the control samples spanned a full ten microns, roughly a thousand times the diameter of a single particle. In contrast, no mega-aggregates appeared in the BiFC preparations under either aqueous or dry conditions. The authors suggest that the linker and fluorophore slightly alter the electrostatic and hydrophobic surface interactions that drive mesoscale particles together, yielding more uniform suspensions and potentially a higher fraction of usable particles recovered after purification.

The fluorescent tag also opens a route to sorting. Because particles carrying both L1 and L2 differ from L1-only shells in fluorescence, overall charge, and estimated mass, the authors lay out several strategies for separating them. At physiological pH, BiFC-modified L1 carries a net positive charge of roughly plus 27 compared with minus 3 for unmodified L1, a difference that could be exploited with field-flow fractionation using a charged membrane and crossflow. The added fluorophore mass shifts total particle weight enough that the extreme populations, L1-only shells and fully loaded L1/L2 capsids, become distinguishable by resonant mass measurement, which gauges buoyant mass from the frequency shift of a vibrating resonator. The authors caution that the incremental weight difference per additional L2 molecule remains too small for conventional size-based separation, but the combination of charge, fluorescence, and mass could support high-throughput workflows in future flow virometry systems.

Estimated yields were workable for laboratory use. Counting particles per unit area on electron microscopy grids, the team calculated roughly 33.59 times ten to the seventh particles per milliliter fraction, sufficient to achieve a multiplicity of infection of about 6.7 when dosing cell cultures, within the commonly recommended range of 0.1 to 10. Quality control analyses, including PCR screening of extracted capsid contents with primers for L1, L2, and SV40 large T sequences, found no detectable encapsidated DNA, confirming that the particles are empty shells. This emptiness is an important safety feature for a platform intended to serve as a foundation for packaging therapeutic genes or chemotherapeutic payloads for targeted delivery.

The work remains at an early stage, and several questions are unresolved, including whether L1 and L2 fusion proteins made in separate cells can assemble during cell lysis, and whether the higher relative L2 expression in the new system saturates each particle with fluorophore and boosts per-particle brightness. Immunogenicity and in vivo functionality have yet to be tested. Still, the study establishes that a fluorescently self-reporting HPV16 capsid can be produced correctly, imaged on routine equipment, kept free of contaminating nucleic acid, and maintained in a less aggregated state than the current standard. If further studies confirm its immunological performance, the platform could support both next-generation vaccines aimed at broader cross-type protection through L2 epitopes and precision delivery vehicles that exploit the virus’s natural tropism for the heparan sulfate proteoglycans displayed on disturbed and cancerous cells, bringing together vaccine science, nanotechnology, and cancer therapy around a single glowing shell.

Subject of Research: Fluorescently tagged HPV16 virus-like particles engineered for real-time assembly monitoring, reduced aggregation, and potential particle sorting

Article Title: A novel approach to HPV16 VLP construction and production with improved accessibility and particle sorting potential

Article References: Williamson, M., Boren, C., Kumar, R., & Gollahon, L. (2026). A novel approach to HPV16 VLP construction and production with improved accessibility and particle sorting potential. Frontiers in Molecular Medicine, 6, Article 1941186. https://doi.org/10.3389/fmmed.2026.1941186

Image Credits: AI Generated

DOI: 10.3389/fmmed.2026.1941186

Keywords: HPV16, virus-like particles, BiFC, Venus fluorophore, L1 capsid protein, L2 capsid protein, nanoparticle characterization, aggregation, flow virometry, vaccine platform, drug delivery, 293T cells