The team, based in Hungary and collaborating internationally, engineered the K1F bacteriophage, a T7-family virus that specifically infects Escherichia coli strains shielded by a K1 polysaccharide capsule, a feature associated with invasive, clinically dangerous E. coli such as those causing neonatal meningitis. Rather than relying on the classical strategy of reporter phages, which carry a transgene that the infected bacterium must transcribe and translate before any signal appears, the researchers exploited the sequence of ejection proteins that T7-like phages deliver into the host cytoplasm at the very start of infection. These intracapsid proteins, known as ICPs, are packed inside the viral head and are propelled into the bacterium in a defined order, where several of them refold to build the ejectosome, a tubular channel through which the viral genome passes.

The reporters in this system exploit the peculiar biochemistry of nanoluciferase, a small and intensely bright enzyme derived from the deep-sea shrimp Oplophorus gracilirostris. Nanoluciferase can be split into two fragments, a larger C-terminal piece and a smaller N-terminal piece, neither of which glows on its own. Crucially, when the fragments meet in solution in the presence of the substrate Furimazine, they spontaneously reassemble into an active enzyme that emits blue light. The researchers fused the small N-terminal fragment, abbreviated NnLuc, to two of the phage’s ICPs, gp6.7 and gp14, ensuring that the reporter fragment would travel inside the capsid, inert and sequestered, until the virus encountered a host it could infect.

Constructing such phages required careful genetic engineering. The team built donor plasmids carrying extra copies of the g6.7 or g14 genes fused in-frame to the NnLuc sequence and inserted these constructs into the K1F genome at a position previously shown to tolerate foreign DNA without instability. Recombinant phages were then selected using a CRISPR-Cas9 counter-selection system that cleaves the genomes of wild-type phages, allowing only the engineered versions to propagate. Phenotypic testing confirmed that the modified viruses grew to titres similar to wild type, with only a slight reduction in plaque size for one construct, indicating that the fusion proteins imposed minimal fitness costs on the viruses.

Initial experiments verified that the fusion proteins were indeed packaged into the phage heads and remained functional. When the researchers heat-treated the engineered phages to crack open the capsids and released their contents into a solution containing the purified C-terminal nanoluciferase fragment, the mixtures glowed brightly, confirming that the packaged NnLuc fusions could reconstitute active luciferase once liberated. Controls using wild-type phage, plasmid-free bacterial extracts, or intact capsids produced little or no signal, with one exception: intact K1Fe6.7::NnLuc particles produced a modest background glow, apparently due to soluble fusion protein contaminating the phage lysate, a problem the team believes can be solved with ultrafiltration purification.

The decisive experiments followed in living cells. When E. coli EV36, a K1-capsulated strain carrying a plasmid that expressed the C-terminal fragment, was challenged with either engineered phage, luminescence rose significantly above controls within minutes. In contrast, E. coli Nissle 1917, which expresses the reporter fragment but wears a non-cognate K5 capsule that the K1F phage cannot dock onto, produced no signal at all, demonstrating that phage binding and injection are strict prerequisites for light emission. Kinetic measurements showed the signal became statistically significant roughly seven minutes and fifty seconds after phage addition, and a more sensitive protocol detected a robust difference as early as five minutes and forty-five seconds after infection.

Those early time points matter because, in the T7 phage family, transcription and translation of late viral genes, including the g6.7 and g14 genes whose products were tagged, do not begin until at least seven to eight minutes after infection. Any luminescence detected before that window therefore cannot originate from the host expressing the reporter gene; it must come from proteins physically injected from the viral capsid. To seal the argument, the researchers treated infected cells with rifampicin, a transcription inhibitor, or tetracycline, a translation inhibitor. At six and a half minutes post-infection, the luminescent signal from cells infected with K1Fe14::NnLuc was completely unaffected by either inhibitor, while K1Fe6.7::NnLuc retained the majority of its signal, providing the first direct evidence that protein injection alone can betray a successful phage infection.

Of course, a diagnostic that only detects bacteria pre-engineered to express half of a luciferase would be of little practical use. To address this, the team built a model diagnostic system in which wild-type, unmodified bacteria are infected by the reporter phages and then supplied externally with column-purified C-terminal nanoluciferase along with the NanoGlo reagent, which contains the Furimazine substrate and a lysis agent that opens the bacterial envelope and allows the two fragments to meet. In this configuration, the cognate K1-capsulated strain generated signals two orders of magnitude brighter than in earlier experiments, and positive versus negative samples were clearly distinguished just three minutes after phage addition, the strongest evidence yet that the emitted light arises purely from injected protein.

Sensitivity testing revealed an important trade-off between speed and detection limit. When cell lysis occurred after only eight minutes of infection, relying solely on the injected enzyme fragments, roughly 690,000 cells per well, corresponding to about 23 million cells per milliliter, were required for reliable detection. Extending the infection to fifteen minutes, which allowed some de novo reporter production, and lengthening the luminescence incubation lowered the limit to 12,000 cells per well, roughly 3.5 million cells per milliliter. The authors note that reported bacterial concentrations in cerebrospinal fluid of children with E. coli K1 meningitis range from about 20,000 to 40 million cells per milliliter, with a mean near 200,000, meaning that even this unoptimized prototype already operates within a clinically relevant range.

The researchers envision the technology as the foundation of cheap, lateral-flow-style tests in which all reactive components, the dried C-terminal fragment, the substrate, and the engineered phage, are pre-loaded in a device, and the user supplies only the sample and moisture. Background noise could be further reduced by purifying phage lysates more stringently, deleting the wild-type copies of the ICP genes to load more reporter per capsid, or deploying the chemistry in microfluidic lab-on-a-chip devices that concentrate reactions into tiny volumes. In the longer term, combinatorial multichannel chips could classify bacteria by genus, species, and strain in sequence. Because phage binding directly reports viral sensitivity, the same platform could help physicians assemble personalized phage-therapy cocktails on the spot, potentially shrinking the turnaround of precision antibacterial treatment from hours to the few minutes that can decide outcomes in severe infection.

Subject of Research: Expression-free bacteriophage-based detection of target bacteria via conditional release of encapsidated reporter proteins

Article Title: Rapid, expression‐free bacteriophage‐based specific detection of target bacteria by conditional release of encapsidated reporter molecules

Article References: Avramucz, Á., Wheatley, J. P., Liyanagedera, S. B. W., Fehér, T., & Amaee, R. (2026). Rapid, expression‐free bacteriophage‐based specific detection of target bacteria by conditional release of encapsidated reporter molecules. Bioengineering & Translational Medicine, Article e70171. https://doi.org/10.1002/btm2.70171

Image Credits: AI Generated

DOI: 10.1002/btm2.70171

Keywords: bacteriophage, bacterial detection, nanoluciferase, diagnostics, K1F phage, E. coli, ejectosome, intracapsid proteins, biosensor, phage therapy, antimicrobial resistance, rapid testing