Nature Microbiology Unveils HIDEN-SEQ to Decode the Genetic “Dark Matter” of Bacteriophages
Bacteriophages contain enormous numbers of genes whose functions remain unknown, even in viruses that have been studied for decades. That genomic “dark matter” is becoming increasingly important as phages move from model systems in molecular biology toward potential tools against multidrug-resistant bacterial infections. A study published in Nature Microbiology now introduces a high-throughput strategy designed to systematically determine which phage genes matter, when they matter and which bacterial defences they help viruses overcome.
The method, called HIDEN-SEQ, or hidden Acr-enabled transposon-insertion sequencing, was developed by Dorentina Humolli, Damien Piel, Jessica Ransome, Kathrin Bausch, Sarah Tschudin-Sutter, Monica Ortelli, Christoph Dehio, Jan-Willem Veening and Alexander Harms. The work brings together researchers from ETH Zürich, University Hospital Basel, the University of Basel and the University of Lausanne. The authors describe HIDEN-SEQ as a way to connect viral genes with experimentally selectable phenotypes across entire bacteriophage genomes, with potential applications ranging from fundamental phage biology to biotechnology and phage therapy.
The problem HIDEN-SEQ addresses is substantial. Most tailed phages carry genomes containing dozens or hundreds of genes. A relatively conserved core is involved in processes such as replication, gene expression, DNA metabolism and virion assembly, but a large proportion of the remaining genes encode hypothetical proteins with no experimentally established function. These accessory genes are particularly interesting because many appear to determine how a phage adapts to specific hosts, environmental conditions and bacterial antiviral defence systems.
Understanding these genes is also becoming a practical problem for phage therapy. A phage may look highly effective against a laboratory bacterial strain but encounter a completely different molecular environment in a clinical isolate carrying toxin–antitoxin systems, restriction enzymes, cyclic-nucleotide defences or other antiviral mechanisms. Without knowing which viral genes allow a phage to overcome those barriers, rational selection and engineering of therapeutic phages remains difficult.
HIDEN-SEQ adapts the principle of transposon-insertion sequencing, or TnSeq, to bacteriophages. In bacterial genetics, TnSeq works by generating large populations of mutants carrying transposon insertions throughout the genome and then measuring which mutants disappear under a particular experimental condition. If disruption of a gene causes a mutant to vanish from the population, that gene is likely important for fitness under that condition.
Applying the same logic to phages has historically been technically difficult because viruses provide fewer convenient genetic selection markers and because many phages possess modified or protected DNA that can frustrate genome-targeting systems.
The ETH-led team devised an elegant solution. They used the Himar1 mariner transposon, which inserts at TA dinucleotide sites, and engineered the transposon to carry the anti-CRISPR gene acrVIA1. Meanwhile, the bacterial host expresses CRISPR–Cas13a programmed to attack the transcript of an essential phage gene. A normal phage is therefore suppressed by Cas13a, while a phage that has successfully acquired the transposon also gains AcrVIA1 and can escape the CRISPR pressure.
This creates a selectable pool containing enormous numbers of phage transposon mutants. Deep sequencing then determines where each transposon has inserted. Researchers can compare the abundance of each mutant before and after infection under a particular condition. Genes that tolerate many insertions are relatively dispensable, whereas genes whose mutants disappear are required either absolutely or specifically under the condition being tested.
The workflow is illustrated particularly clearly in Figure 1 of the paper. A transposon carrying the anti-CRISPR gene is introduced randomly into phage genomes, CRISPR–Cas13a eliminates phages that failed to acquire the protective element, and the surviving transposon library is subsequently exposed to different bacterial hosts or environmental conditions. Sequencing before and after those challenges converts changes in mutant abundance into a genome-wide functional map.
The researchers first benchmarked HIDEN-SEQ using bacteriophage T4, one of the most extensively studied viruses in molecular biology. T4 carries a genome of approximately 168 kilobases encoding nearly 300 genes, yet around 130 of those genes remain uncharacterized despite decades of research. The initial HIDEN-SEQ library achieved transposon insertions at 97.2% of available TA sites across the viral genome.
After allowing the mutant population to replicate on laboratory Escherichia coli K-12, the researchers determined that 190 T4 genes were not essential under the standard conditions tested. The resulting essentiality map closely reproduced knowledge accumulated through decades of conventional T4 genetics, providing an important validation that the massively parallel approach was identifying biologically meaningful effects rather than simply generating insertion noise.
HIDEN-SEQ also produced more information than a simple essential versus non-essential classification. Some disruptions caused modest reductions in viral fitness rather than complete loss of replication, allowing genes to be separated into categories ranging from weak fitness defects to essential functions. In some cases the density of insertions was sufficient to resolve functional differences within individual genes. T4 RNA ligase A, for example, tolerated insertions in one part of the protein but not another, reproducing known domain-level functional differences.
The most powerful aspect of the method emerged when the same mutant library was exposed to different bacterial defence systems.
A gene can be dispensable during infection of one host but indispensable in another. This is particularly true for anti-defence genes. If a bacterial strain does not possess the defence being targeted, deleting the corresponding viral countermeasure may have almost no consequence. Put the same mutant phage against a bacterium carrying that defence, however, and the virus may suddenly become unable to replicate.
The researchers demonstrated this principle using several well-characterized T4 systems. The rIIA and rIIB genes became specifically important when T4 encountered the RexAB antiviral system. The dmd gene was required against the RnlAB toxin–antitoxin system. The known 57B/Acb1 protein became important against CBASS, while the anti-DarT factor AdfN was specifically required against DarTG1.
Once HIDEN-SEQ had shown that it could recover known phage counter-defence mechanisms, the team asked whether it could uncover ones that had been missed.
When T4 was challenged with the DarTG2 defence system, a single gene, previously called tk.4, emerged as strongly conditionally essential. Removing the gene made T4 sensitive to DarTG2. Structural analysis indicated similarity to the macrodomain of the Mycobacterium tuberculosis DarG antitoxin, leading the researchers to rename the protein AdfM, for anti-DarT factor macrodomain. The authors propose that AdfM counteracts DarTG2 by removing ADP-ribose modifications from viral DNA.
The method was not restricted to bacterial immune systems. When T4 mutants were compared between nutrient-rich LB and glucose-supplemented minimal medium, disruption of the viral vs gene produced one of the strongest condition-dependent defects. The gene modifies valyl-tRNA synthetase, although the biological purpose of that modification remains unclear. Deleting vs specifically impaired T4 plaque formation under nutrient limitation, showing that HIDEN-SEQ can identify viral genes needed for different physiological states of the bacterial host as well as for escaping immunity.
The researchers then moved beyond T4 to determine whether HIDEN-SEQ could work on less intensively studied viruses.
They first applied it to Bas37, a relative of T4. The overall essentiality profiles of the two phages were strikingly similar, but HIDEN-SEQ also exposed gene-specific differences. For example, the Bas37 version of RNA ligase A showed a different domain-essentiality pattern from its T4 counterpart, and an orthologue of the T4 translational regulator regA was essential in Bas37 but dispensable in T4 under the same laboratory conditions.
A more challenging test came from Bas54, a much less characterized member of the Vequintavirinae with a 136-kb genome containing 227 predicted genes. Its HIDEN-SEQ library reached 91.6% saturation of available TA sites. After replication under standard conditions, 145 genes could be disrupted without complete loss of viability.
One of the most revealing observations concerned the supposedly essential portion of this poorly understood phage. Of 44 Bas54 genes classified as essential, 15 lacked a functional annotation. In other words, even among the genes without which the virus apparently cannot complete productive infection, roughly one-third encoded proteins whose roles were unknown. Targeted knockout experiments confirmed severe growth defects for several of these genes.
The Bas54 experiments also revealed an intriguing feature relevant to host range. The phage carries numerous predicted tail-fibre genes, yet several of them could be disrupted without preventing infection of laboratory E. coli K-12. Later experiments across different bacterial isolates suggested that alternative tail fibres become important on different hosts, potentially allowing the virus to switch between receptor-recognition strategies depending on the bacterial strain.
This led to one of the study's most clinically relevant experiments. Instead of continuing to use only laboratory E. coli, the researchers challenged the phage mutant libraries with uropathogenic E. coli isolates obtained from patients evaluated for urinary tract infection or asymptomatic bacteriuria at University Hospital Basel.
When T4, Bas37 and Bas54 infected these clinical bacterial isolates, different sets of viral genes became conditionally essential. Because naturally occurring E. coli strains carry highly variable antiviral defence repertoires, the same bacteriophage can face a fundamentally different molecular battlefield from one patient isolate to another.
HIDEN-SEQ allowed the researchers to move from those genome-wide signals toward specific phage–defence relationships.
T4 Alt and a Bas37 β-α glucosyltransferase were implicated in resistance to Mokosh type I defence. Alt was already known as an ADP-ribosyltransferase capable of modifying host RNA polymerase and interfering with other cellular systems, but the experiments uncovered a previously unrecognized role in helping T4 overcome Mokosh. The researchers also found evidence that viral DNA glucosylation contributes to escaping the same defence.
Two previously uncharacterized genes provided even more striking examples. T4 nrdC.1 proved necessary for overcoming the Septu defence system, while Bas37 bas37_0203 protected against Druantia type III. Neither protein contained known domains that immediately explained its activity, meaning the functional screen provided biological information that sequence annotation alone had failed to reveal.
Another Bas37 gene, bas37_0260, emerged as a previously unknown inhibitor of the bacterial McrBC type IV restriction system. The gene occupies the same general genomic region as T4 proteins involved in type IV restriction evasion but shares essentially no sequence similarity with them beyond an N-terminal signal associated with packaging into viral particles. The knockout became specifically sensitive to McrBC, providing direct functional evidence for its role.
The researchers also identified two Bas54 genes required to overcome a predicted bacterial defence called DS-28. One encodes a MoxR ATPase, whereas the second, bas54_0140, had no functional annotation. Across the study, HIDEN-SEQ uncovered new bona fide phage factors acting against six different bacterial defence systems: DarTG2, Mokosh type I, Septu, Druantia type III, McrBC type IV restriction and DS-28. Five of the implicated viral genes had no previously established function.
That result captures why functional genomics could be particularly important for phage research. Genome sequencing can identify genes and computational approaches can predict structures, domains and evolutionary relationships, but many phage proteins remain too divergent to annotate confidently. HIDEN-SEQ asks a different question: not “what does this sequence resemble?” but “under which biological condition does the virus fail when this sequence is disrupted?”
For therapeutic phage development, this could eventually provide a much deeper understanding of why an otherwise promising phage succeeds against one bacterial isolate and fails against another. A phage's host range is not controlled solely by adsorption to a receptor. Even after successful attachment and genome entry, dozens of bacterial defence systems can terminate infection. Mapping which viral genes neutralize those systems could therefore help identify phages better equipped to infect clinically relevant strains.
The same information could also support phage engineering.
HIDEN-SEQ directly identifies genomic regions that can tolerate disruption. For synthetic biology, these dispensable regions could indicate where genes might be removed, replaced or potentially used as insertion sites for engineered functions. Conversely, genes that become essential only against particular clinical isolates highlight components that should not be removed if the goal is to preserve a broad therapeutic host range. The authors explicitly identify mapping of dispensable regions as one potential route toward improved clinical and biotechnological phage engineering.
HIDEN-SEQ also has several important limitations.
Because Himar1 inserts specifically at TA dinucleotides, genes containing very few TA sites cannot be interrogated at the same resolution; in the T4 analysis, genes with five or fewer sites were not assigned a standard essentiality classification. Transposon insertions can also alter expression of nearby genes, producing so-called polar effects. The researchers attempted to minimize this by using a transposon without a transcriptional terminator, but they acknowledge that such effects may explain some disagreements with previous T4 genetics.
There is another conceptual limitation: a gene that is absolutely essential cannot easily be studied for secondary conditional effects using a disruptive transposon because mutants lacking it cannot propagate in the first place. Knockdown approaches such as CRISPR interference therefore remain complementary rather than obsolete. The authors also note that individual phages may require technical adaptations, including alternative promoters or transposon systems, particularly when unusual DNA modifications interfere with mutagenesis.
The field is evolving rapidly. During revision of the manuscript, the authors note that two independent studies reported related transposon-sequencing approaches in other phage systems. Rather than diminishing the significance of HIDEN-SEQ, these parallel developments suggest that genome-scale transposon mutagenesis may become a broader platform for experimental phage genetics.
What distinguishes HIDEN-SEQ is the combination of dense genome-wide mutagenesis, CRISPR-based enrichment and the ability to reuse the same mutant library against different hosts, immune systems and environmental conditions. A single phage genome can therefore be converted into thousands of competing genetic experiments performed simultaneously.
That could alter how researchers approach the immense unexplored genetic diversity of bacteriophages.
Instead of studying hypothetical genes individually, scientists can perturb large fractions of a viral genome in parallel and allow the biological condition itself to reveal which genes matter. A defence system identifies its corresponding viral countermeasure because mutants lacking that countermeasure disappear. A clinical bacterial isolate exposes the viral machinery required specifically for infection of that strain. Nutrient limitation reveals genes involved in adapting to altered host physiology.
For phage therapy, the immediate importance is not that HIDEN-SEQ has produced a new therapeutic phage. It has not. The advance is a tool for understanding the viruses that researchers may eventually want to use therapeutically.
As phage therapy moves toward increasingly rational selection and engineering, simply knowing that a phage forms plaques may no longer be enough. Researchers will want to understand which parts of its genome are essential, which genes determine its performance against different bacterial strains, which proteins neutralize bacterial immunity and which genomic regions can safely be modified.
HIDEN-SEQ provides a way to obtain much of that information in a single systematic framework.
The study therefore addresses one of the central contradictions in modern phage biology: bacteriophages are among the most genetically diverse biological entities on Earth and potentially powerful sources of antimicrobial and biotechnological tools, yet much of their genetic machinery remains unidentified. By experimentally linking thousands of viral mutations to precise biological conditions, HIDEN-SEQ offers a route into that unexplored genetic space — and potentially a more rational foundation for engineering the phages that will be used in the future.
Sources :
Dorentina Humolli, Damien Piel, Jessica Ransome, Kathrin Bausch, Sarah Tschudin-Sutter, Monica Ortelli, Christoph Dehio, Jan-Willem Veening & Alexander Harms. Systematic mapping of bacteriophage gene essentiality with HIDEN-SEQ. Nature Microbiology (2026).
DOI: https://doi.org/10.1038/s41564-026-02455-8
.png){kind=logo}
Comments
Post a Comment