A Virus That Kills Bacteria In 17 Minutes Is Being Used To Hack Evolution

Lab-Engineered Virus Drives Bacterial Evolution 160,000 Times Faster Than Normal

A virus that kills bacteria in 17 minutes is now being used to speed up evolution itself. Researchers at the National University of Singapore have turned a bacteria-infecting virus into a turbocharged mutation engine, compressing weeks of laboratory work into roughly a day-long cycle. Called lytic selection and evolution (LySE), the method could change how scientists engineer bacteria to study drug resistance, improve metabolic pathways, and one day support applications from plastic-linked recycling processes to medicine production.

At its core, the system exploits bacteriophage T7, a tiny virus that naturally hunts E. coli bacteria, copies itself inside the cell, and bursts out in about 17 minutes, releasing roughly 180 new viral particles in the process. Most viruses try to copy DNA as accurately as possible. LySE does the opposite.

Researchers rewired T7’s DNA-copying machinery to make deliberate, targeted errors at extraordinarily high rates, flooding specific genes with copying mistakes while leaving the rest of the bacterial DNA untouched. That precision lets the team evolve chosen genetic traits at speeds far beyond what standard lab methods can achieve.

How LySE Outpaces Standard Evolution Methods

For decades, scientists have used a process called directed evolution to engineer proteins and biological pathways. In the classic approach, researchers manually introduce random mutations into a gene, test which versions perform best, and repeat the cycle. It works, but it is slow and labor-intensive. Newer, automated “continuous” methods run without human intervention, but they lose control: mutations accumulate in the wrong places, and cells sometimes find shortcuts, called cheater mutations, that game the system without actually improving the target gene.

LySE splits the difference. During each cycle, the system alternates between two phases. In the lytic phase, T7 viruses infect bacteria at a high ratio, triggering an error-prone copying process that churns out mutated versions of the target genes and packages them into new viral particles. In the cellular phase, those particles deliver the mutated genes to fresh, uninfected bacteria at a low ratio, where the bacteria grow and compete. Only cells carrying improved gene variants survive whatever pressure the researchers have applied, whether that is an antibiotic or a difficult food source. Critically, the old bacterial hosts are completely destroyed during the lytic phase, wiping out any stray mutations that may have built up elsewhere in the bacterial DNA. Each new cycle starts with a clean slate.

Inside the Broken Copying Machine

Engineering the mutation engine at the heart of LySE took multiple rounds of modification. Researchers started with a previously known error-prone version of T7’s DNA-copying enzyme and stacked additional changes on top of one another to make it progressively sloppier.

Early variants reached mutation rates roughly 70 times higher than the normal enzyme. From there, the team disabled the enzyme’s error-catching function entirely and fused it to a protein that chemically alters DNA letters as they are being copied. One intermediate version, called v8, reached a mutation rate approximately 160,000 times higher than the baseline rate seen in the host E. coli genome. A refined final version, v9, had a somewhat lower but still very high mutation rate while maintaining better viral packaging efficiency, making it the better practical choice for running LySE campaigns.

Both versions favored one broad class of DNA changes, called transitions (where one DNA letter swaps for a chemically similar one), rather than producing every possible single-letter change equally. Within that constraint, mutations were spread evenly across the full target sequence, with no obvious hot spots or dead zones.

An elegant side effect of making the copying enzyme so error-prone was an automatic safety feature the researchers call multiplicity tuning. Because the error-prone enzyme sabotages its own replication, widespread bacterial killing only occurs when researchers deliberately add a large number of viral particles relative to bacteria. Dialing that ratio up or down gives precise control over when each phase of the cycle begins, with no genetic switches or chemical triggers required.

From Drug Resistance Genes to Plastic-Linked Pathways

To test the system, the researchers ran two experiments. First, they evolved a gene called tetA, which confers resistance to the antibiotic tetracycline, to instead resist tigecycline, a more powerful antibiotic that normally defeats tetA-based defenses. After five LySE cycles, the evolved bacteria tolerated 25 times more tigecycline than bacteria with the original gene. A standard comparison method, adaptive laboratory evolution, reached only a fraction of that improvement over the same number of rounds, and those gains largely disappeared when the evolved gene was moved to fresh bacteria.

In a second test, the team evolved a five-gene pathway that allows E. coli to use ethylene glycol, a chemical building block released when PET plastic is broken down, as its sole food source. After five LySE cycles, the best-performing bacterial strains showed a 50.9% increase in final cell density compared to the starting version, according to the study published in Nature Microbiology. Bacteria evolved by the comparison method showed no mutations in the target genes at all; their gains came entirely from changes elsewhere in the genome.

LySE requires nothing beyond standard lab tools: pipettes, culture tubes, and an incubator. It may also handle genetic packages far larger than most competing methods allow, opening the door to evolving multi-gene pathways that have long been out of reach for rapid directed evolution.

---