Bacteria Turn Phage Immune Evasion Into a Deadly Self-Destruct Signal

Bacteria and the viruses that infect them, bacteriophages, have been locked in an evolutionary arms race for billions of years. Phages outnumber bacteria roughly ten to one, and this relentless pressure has driven bacteria to evolve a diverse arsenal of immune systems that detect and repel viral infection. Phages, in turn, have developed equally elaborate countermeasures. A new review published in PLOS Pathogens examines one particularly striking twist in this conflict: the recently discovered Panoptes defense system, which turns a phage’s own evasion tactic into a trigger for the infected cell’s self-destruction.

At the heart of this molecular conflict lies nucleotide signaling. Bacteria use small, diffusible nucleotide molecules as second messengers to report infection and coordinate their immune responses. Several major defense systems rely on this principle, including the cyclic oligonucleotide-based antiphage signaling system (CBASS), Thoeris, Pycsar, and type III CRISPR-Cas. In each case, sensor proteins detect phage infection, synthesize nucleotide signals, and activate downstream effector proteins that execute defense responses such as cell death or growth arrest, depriving the virus of a replicating host.

CBASS systems, found in roughly ten percent of sequenced bacterial genomes, are built around cGAS/DncV-like nucleotidyltransferases, enzymes that generate diverse cyclic di- and trinucleotide signals upon recognizing phage infection. These signals activate effector proteins that carry out growth arrest or cell death through membrane disruption, DNA degradation, or metabolite depletion. For example, the effector CapV, the most abundant CBASS effector, is a phospholipase activated by the signal 3′,3′-cGAMP that degrades membrane phospholipids. Another effector, Cap15, recognizes cyclic dinucleotides containing at least one uridine base and oligomerizes to disrupt the inner membrane, while NucC is an endonuclease activated by the cyclic trinucleotide cA3 that nonspecifically cleaves bacterial DNA.

Thoeris systems operate on a related principle using NAD+-derived signals. The ThsB sensor protein, through its Toll/interleukin-1 receptor domain, synthesizes signals such as 1”–3′ gcADPR in type I Thoeris or His-ADPR in type II Thoeris. These signals bind the ThsA effector, which cleaves NAD+, leading to NAD+ depletion and subsequent cell death or growth arrest. A distinct type IV Thoeris variant employs a TIR-domain sensor that produces N7-cADPR upon infection, activating a caspase-like effector that nonspecifically degrades cellular proteins. Pycsar follows similar logic: its cyclase PycC synthesizes the cyclic mononucleotides cUMP and cCMP, which activate effectors that either cleave NAD+ or disrupt the inner membrane.

Phages have answered these defenses with an array of counterstrategies. Some anti-defense proteins directly destroy immune components: Acb1 and Apyc1 degrade cyclic nucleotide signals, while Tad3 physically blocks effector activation. Particularly effective, however, are sponge-like proteins, small phage-encoded polypeptides that sequester bacterial nucleotide signals without degrading them. By soaking up these signaling molecules, sponges prevent effector activation and neutralize the host immune response, enabling phage replication. Seven distinct families of phage-encoded sponge proteins have been experimentally characterized to date: Acb2, Acb4, Tad1, Tad2, Sequestin, Lockin, and TadIV-1.

Despite their diverse amino acid sequences and oligomeric structures, these sponge proteins share a common functional principle: they assemble into multimeric complexes presenting deep, positively charged binding pockets that accommodate specific nucleotide ligands with varying degrees of promiscuity. Acb2, for instance, forms a hexamer with two distinct binding sites that simultaneously capture cyclic dinucleotides such as 3′,3′-cGAMP and cyclic trinucleotides such as cA3, antagonizing multiple CBASS subtypes at once. Acb4 assembles into a tetramer that sequesters diverse ligands. Tad1 and Tad2 are cross-system sponges that soak up signals from both CBASS and Thoeris, with Tad1 displaying a remarkable two-plus-six binding architecture. Sequestin and Lockin are dedicated Thoeris inhibitors, and TadIV-1 targets type IV Thoeris by binding N7-cADPR through a unique flexible loop. Recent work has shown that homologs of these sponges exhibit even broader targeting than previously appreciated, with some Acb2 homologs sequestering the Thoeris signal 3′cADPR and MoTad2 binding the Pycsar signals cUMP and cCMP.

Bacteria, however, have countered with the Panoptes system, which inverts the logic of this arms race. Panoptes belongs to a growing list of antiphage systems that detect phage evasion rather than infection directly, joining PARIS, which monitors viral anti-defense proteins; PrrC, which tracks restriction system status; and Hailong, which uses a decoy single-stranded DNA to detect infection. What makes Panoptes distinctive is that it uses nucleotide second messengers as its sensing cue. In type I Panoptes, encoded by a two-gene operon of optS and optE, the OptS enzyme is a constitutively active minimal CRISPR polymerase that continuously produces cyclic dinucleotides, including 2′,3′-c-di-AMP and 3′,2′-cGAMP. These molecules bind directly to OptE, a transmembrane effector, locking it in an inactive conformation and preventing its toxicity under steady-state conditions.

The novelty emerges upon phage infection. When a phage deploys its sponge proteins to sequester nucleotide signals, an essential step in its immune evasion strategy, those sponges bind not only their intended CBASS and Thoeris targets but also the decoy cyclic dinucleotides produced by OptS. In stripping away these inhibitory ligands from OptE, the sponges liberate the effector from repression. Freed OptE undergoes conformational changes and oligomerizes in the absence of its nucleotide, triggering disruption of the bacterial inner membrane and restriction of phage replication, ultimately leading to cell death or growth arrest. Type II Panoptes, which accounts for approximately forty percent of all Panoptes instances, works analogously: its mCpol cyclase generates the cyclic trinucleotide cA3, which holds a CARF-TM effector in an inactive dimeric state until sponges remove it, at which point the effector drives oligomerization, outer membrane permeabilization, and growth arrest. This arrangement is itself highly unusual, since most immune systems are activated by infection signals rather than constitutively primed with repressive ligands that must be removed for activation.

Genetic and biochemical experiments have confirmed that in type I Panoptes, Acb2 directly sequesters 2′,3′-c-di-AMP and that this sequestration activity is sufficient to trigger activation. Acb2 thus exhibits a dual function: it helps phages evade CBASS and, in some cases, Thoeris, yet in Panoptes-expressing bacteria the very same activity becomes a lethal trigger. For type II Panoptes, heterologous expression of Acb2 has similarly been shown to sequester cA3 and activate the CARF-TM effector, causing cell toxicity, although whether phage-encoded Acb2 activates type II Panoptes during natural infection remains to be demonstrated. Activation is not limited to Acb2: co-expression of Tad1 or Tad2 homologs with the OptSE operon also reduces bacterial colony formation, suggesting Panoptes responds to multiple antagonists that deplete its cognate signals. Consistent with a model of sensing signal loss rather than recognizing sponge proteins themselves, the degrading enzyme Acb1 activates Panoptes, while Apyc1 and AcrIII-1, which target cyclic mono- and tetranucleotides, do not. However, direct biochemical evidence for how Tad proteins trigger type I Panoptes is still lacking.

The discovery carries broader implications. The decoy-based logic of Panoptes fits the guard/decoy model well established in plant NLR immunity, in which a sensor monitors a host target or its mimic and triggers defense when a pathogen effector modifies either. It suggests that additional trap systems monitoring cellular metabolites and signaling molecules may await discovery in bacterial defense islands. The evolutionary connection between CBASS and the eukaryotic cGAS-STING pathway has also prompted interest in whether higher organisms employ analogous strategies, particularly since TIR-domain proteins in mammals produce Thoeris-like signaling molecules, though no eukaryotic equivalent of phage sponge proteins has yet been identified. From a translational standpoint, Panoptes offers conceptual inspiration for antiviral strategies that exploit a pathogen’s own evasion mechanisms rather than directly targeting viral functions, an approach that may face fewer challenges with resistance development. Key open questions remain, including how OptE oligomerization forms membrane pores, the full repertoire of antagonists capable of triggering Panoptes, and whether the system exists in bacterial lineages beyond those currently identified.

Subject of Research: Bacterial antiphage immunity and phage-encoded sponge proteins

Article Title: From CBASS and Thoeris to Panoptes: Phage sponges and the inversion of bacterial immunity

Article References: Zheng, R., & Zhang, L. (2026). From CBASS and Thoeris to Panoptes: Phage sponges and the inversion of bacterial immunity. PLOS Pathogens, 22(10), e1014717. https://doi.org/10.1371/journal.ppat.1014717

Image Credits: AI Generated

DOI: 10.1371/journal.ppat.1014717

Keywords: bacteriophages, bacterial immunity, CBASS, Thoeris, Panoptes, nucleotide signaling, sponge proteins, phage evasion, innate immunity, host-virus arms race, second messengers, microbiology