Somewhere inside every human cell, a molecular alarm system watches for DNA that does not belong. When it spots a viral intruder, it fires off a chemical distress signal that rallies the body’s antiviral defenses. That system, anchored by a protein called cGAS, has been a focus of immunology research for over a decade. But a growing body of work, highlighted in recent Quanta Magazine reporting, reveals something striking: bacteria have been running a nearly identical program for more than a billion years, using it to fight off their own viral enemies. The discovery is reshaping how scientists understand the origins of human immunity and where its future therapies might come from.
A shared molecular alarm
The protein at the center of this story is cyclic GMP-AMP synthase, or cGAS. In mammalian cells, cGAS acts as a cytosolic DNA sensor, first identified by Zhijian “James” Chen’s lab at the University of Texas Southwestern Medical Center in 2013. When cGAS encounters foreign DNA drifting in the cell’s interior, whether from an invading virus or a damaged cell, it produces a small signaling molecule called cyclic GMP-AMP (cGAMP). That molecule, carrying a distinctive 2′-5′ chemical bond, directly activates the adaptor protein STING, which switches on type I interferon signaling and launches a frontline antiviral response.
What surprised researchers was finding that bacteria run a remarkably parallel program. In 2019, a team led by Rotem Sorek at the Weizmann Institute of Science published work in Nature demonstrating that bacteria deploy a four-gene system called CBASS, short for cyclic-oligonucleotide-based anti-phage signaling system. When a bacteriophage, a virus that attacks microbes, infects a bacterium carrying CBASS, the system produces its own version of cGAMP. But instead of rallying neighboring defenders, the infected bacterium essentially destroys itself before the phage can finish replicating, sacrificing one cell to protect the surrounding colony. Sorek’s team showed this abortive infection response works against diverse phages, not just a single strain, making it a broad-spectrum defense.
The enzymatic family behind these signals turns out to be vast. A broader group of bacterial enzymes called CD-NTases (cGAS/DncV-like nucleotidyltransferases) generates a chemically diverse set of cyclic nucleotide messengers. Biochemical screens, including work by Philip Kranzusch’s group at Harvard Medical School, confirmed that CD-NTases are widespread across microbial genomes and produce signals with varied chemical structures. Bacteria, in other words, do not carry a single signaling weapon. They carry an entire arsenal of related molecules, each capable of activating different downstream effectors.
Tracing the evolutionary thread
If bacteria and human cells use such similar molecular tools, the obvious question is: where did the overlap come from?
Evolutionary analysis published in Molecular Cell by Kranzusch and colleagues established that the cGAS-STING pathway has ancient evolutionary roots. That work revealed deep conservation at the protein-structure level, even when raw DNA sequences have diverged dramatically. The three-dimensional shapes and biochemical activities of these proteins have been preserved across vast evolutionary distances, while the underlying genetic code has drifted beyond easy recognition. The study also showed that 2′,3′-cGAMP can activate STING proteins from very different organisms, suggesting the signaling molecule functions as something close to a universal chemical password across species.
Further evidence comes from the broader search for microbial “defense islands,” regions of bacterial and archaeal genomes packed with genes that help fend off viral invaders. A 2018 study in Science by Shany Doron, Rotem Sorek, and colleagues at the Weizmann Institute used a systematic genome-neighborhood strategy to discover and experimentally validate multiple previously unknown anti-phage systems distributed widely across microbial genomes. Many of these defense islands contain CD-NTases and their associated effectors, reinforcing the idea that cyclic nucleotide signaling is a central organizing principle in microbial antiviral defense, not an isolated curiosity.
What remains unsettled
For all the structural parallels, the exact route by which cGAS-STING moved from bacteria into animal cells has not been definitively established. Structural conservation strongly suggests a shared ancestor, but whether the pathway was inherited vertically through an unbroken chain of descent, acquired through horizontal gene transfer at some point in deep history, or arose independently through convergent evolution acting on similar biochemical building blocks is still debated. The emphasis on structural conservation over sequence similarity complicates standard phylogenetic methods, which rely on DNA comparisons, and makes it harder to reconstruct a precise evolutionary timeline.
It is also unclear how much of the bacterial signaling repertoire has direct analogs in animals. While the core architecture of cyclic nucleotide messengers paired with sensor-effector modules appears conserved, many bacterial CD-NTases synthesize unusual cyclic molecules with no obvious counterparts in human cells. Whether these exotic messengers represent evolutionary dead ends or inspired now-unrecognizable descendants in animal immunity remains an open question that will require both computational comparison and hands-on biochemical work with poorly understood animal proteins.
The full catalog of microbial defense mechanisms is almost certainly incomplete. The Doron et al. study validated several new systems, but the researchers themselves noted that many predicted defense gene clusters await experimental testing. Future discoveries could reveal additional links between microbial and animal antiviral strategies, or they could show that many bacterial defenses have no direct analog in higher organisms. Either outcome would sharpen the picture considerably.
The path toward medicine
The shared logic of cGAS-STING signaling between bacteria and humans has already attracted pharmaceutical interest. Synthetic STING agonists, molecules designed to mimic cGAMP and activate the pathway on demand, are being tested in clinical trials as cancer immunotherapies and vaccine adjuvants. Companies including Merck, Bristol Myers Squibb, and several biotech firms have advanced STING-targeting compounds into early-phase human studies, aiming to boost immune responses against tumors that have learned to evade detection.
But the same pathway that fights viruses and tumors can, when overactivated, drive damaging inflammation. Gain-of-function mutations in STING are linked to a group of rare autoinflammatory conditions called STING-associated vasculopathy with onset in infancy (SAVI), and dysregulated cGAS-STING signaling has been implicated in lupus-like disease and other interferonopathies. How to exploit the system therapeutically without triggering harmful inflammation is a problem that laboratory findings alone cannot resolve, particularly given that patients vary widely in their baseline immune reactivity.
No primary trial data yet links bacterial-derived cGAMP variants specifically to enhanced human immunity in a way that clearly outperforms existing approaches. The therapeutic promise is real but early, and the gap between understanding a billion-year-old signaling language and safely manipulating it in a clinic remains substantial.
Reading the evidence clearly
The strongest claims in this field rest on direct experimental data. The identification of cGAS as a DNA sensor and the biochemical characterization of cGAMP come from controlled laboratory experiments with defined cell lines and purified proteins. The CBASS findings derive from phage challenge assays in bacteria, where researchers measured survival rates and tracked cGAMP production in real time. These are reproducible, mechanistic results, not statistical correlations.
Evolutionary claims carry a different weight. The ancient origin of cGAS-STING is inferred from structural comparisons across species, supported by the finding that proteins from distantly related organisms fold into similar shapes and catalyze the same reactions. That is strong evidence of shared ancestry, but it is not the same as watching the pathway transfer from one organism to another. The distinction matters: “these proteins do the same thing in bacteria and humans” is well established; “humans inherited this pathway directly from bacteria” is the model that best fits current data, not a directly observed historical event.
As of May 2026, the evidence supports a compelling but still-developing narrative. Human cells and bacteria share a molecular language for detecting and responding to invading genetic material, built around cyclic nucleotide messengers and conserved sensor proteins. The details of how that language evolved, and how it can be safely turned into medicine, are still being worked out. Each new microbial defense system that comes to light, each structural study that bridges the gap between distant species, adds resolution to a picture that stretches back more than a billion years and forward into the next generation of immunotherapies.
More from Morning Overview
*This article was researched with the help of AI, with human editors creating the final content.
