Phage DNA Research Articles

Structural and biochemical insights into the mechanism of the Gabija bacterial immunity system

The Gabija system is a newly discovered bacterial immune system that consists of GajA and GajB. Here we report the cryo-EM structure of the Gabija complex from Bacillus cereus VD045 at 3.6 Å, which provides the direct evidence of interactions between GajA and GajB. The Gabija complex is an octameric ring structure with four GajA and four GajB. GajA is an OLD nucleases family protein, while GajB belongs to the SF1 helicases. The Gabija complex has sequence-specific DNA nuclease activity and prefers circular rather than linear DNA as substrate, its activity is more sensitive to concentrations change of nucleotides compared to GajA alone. Our data suggest a mechanism of Gabija immunity: the nuclease activity of Gabija complex is inhibited under physiological conditions, while it is activated by depletion of NTP and dNTP upon the replication and transcription of invading phages and cleave the circular DNA to prevent phage DNA replication.

Bacteriophage P22 SieA-mediated superinfection exclusion.

Many temperate phages encode prophage-expressed functions that interfere with superinfection of the host bacterium by external phages. Salmonella phage P22 has four such systems that are expressed from the prophage in a lysogen that are encoded by the c2 (repressor), gtrABC, sieA, and sieB genes. Here we report that the P22-encoded SieA protein is necessary and sufficient for exclusion by the SieA system and that it is an inner membrane protein that blocks DNA injection by P22 and its relatives, but has no effect on infection by other tailed phage types. The P22 virion injects its DNA through the host cell membranes and periplasm via a conduit assembled from three "ejection proteins" after their release from the virion. Phage P22 mutants that overcome the SieA block were isolated, and they have amino acid changes in the C-terminal regions of the gene 16 and 20 encoded ejection proteins. Three different single-amino acid changes in these proteins are required to obtain nearly full resistance to SieA. Hybrid P22 phages that have phage HK620 ejection protein genes are also partially resistant to SieA. There are three sequence types of extant phage-encoded SieA proteins that are less than 30% identical to one another, yet comparison of two of these types found no differences in phage target specificity. Our data strongly suggest a model in which the inner membrane protein SieA interferes with the assembly or function of the periplasmic gp20 and membrane-bound gp16 DNA delivery conduit.IMPORTANCEThe ongoing evolutionary battle between bacteria and the viruses that infect them is a critical feature of bacterial ecology on Earth. Viruses can kill bacteria by infecting them. However, when their chromosomes are integrated into a bacterial genome as a prophage, viruses can also protect the host bacterium by expressing genes whose products defend against infection by other viruses. This defense property is called "superinfection exclusion." A significant fraction of bacteria harbor prophages that encode such protective systems, and there are many different molecular strategies by which superinfection exclusion is mediated. This report is the first to describe the mechanism by which bacteriophage P22 SieA superinfection exclusion protein protects its host bacterium from infection by other P22-like phages. The P22 prophage-encoded inner membrane SieA protein prevents infection by blocking transport of superinfecting phage DNA across the inner membrane during injection.

Phage treatment of Pseudomonas aeruginosa yields a phage-resistant population with different susceptibility to innate immune responses and mild effects on metabolic profiles

In this study, we have investigated innate immune activation capacity and metabolic features of a population of P. aeruginosa PAO1 phage-resistant mutants with diverse genetic modification (large genomic deletions and point mutations) arising after exposure to phages targetting lipopolysaccharide (LPS) or Type-4 pili (T4P). Deletions led to the loss of genes involved in LPS synthesis, cell envelope permeability, efflux systems, biofilm production, oxidative stress tolerance, and DNA repair. Loss of LPS O antigen resulted in bacterial sensitivity to serum complement and stimulation of inflammatory cascades but did not cause increased phagocytosis, while T4P phage-resistant mutants were more effectively phagocytized than LPS-defective mutants. Changes in the utilization of different carbon, nitrogen, sulphur, and phosphorus sources were identified, especially in mutants where the two phage DNA persisted in the bacterial population (pseudolysogeny). However, the metabolic changes did not directly correlate with single-gene mutations or the large gene deletions, suggesting they reflect adaptive changes to the gene modifications that arise during the selection of resistant mutants. In contrast, phage-resistant mutants were susceptible to humoral innate immune responses, suggesting that phage resistance may be a beneficial outcome of phage therapy.

Recruitment of complete crAss-like phage genomes reveals their presence in chicken viromes, few human-specific phages, and lack of universal detection.

The order Crassvirales, which includes the prototypical crAssphage (p-crAssphage), is predominantly associated with humans, rendering it the most abundant and widely distributed group of DNA phages in the human gut. The reported human specificity and wide global distribution of p-crAssphage makes it a promising human fecal marker. However, the specificity for the human gut as well as the geographical distribution around the globe of other members of the order Crassvirales remains unknown. To determine this, a recruitment analysis using 91 complete, non-redundant genomes of crAss-like phages in human and animal viromes revealed that only 13 crAss-like phages among the 91 phages analyzed were highly specific to humans, and p-crAssphage was not in this group. Investigations to elucidate whether any characteristic of the phages was responsible for their prevalence in humans showed that the 13 human crAss-like phages do not share a core genome. Phylogenomic analysis placed them in three independent families, indicating that within the Crassvirales group, human specificity is likely not a feature of a common ancestor but rather was introduced on separate/independent occasions in their evolutionary history. The 13 human crAss-like phages showed variable geographical distribution across human metagenomes worldwide, with some being more prevalent in certain countries than in others, but none being universally identified. The varied geographical distribution and the absence of a phylogenetic relationship among the human crAss-like phages are attributed to the emergence and dissemination of their bacterial host, the symbiotic human strains of Bacteroides, across various human populations occupying diverse ecological niches worldwide.

Short-range translocation by a restriction enzyme motor triggers diffusion along DNA

Cleavage of bacteriophage DNA by the Type III restriction-modification enzymes requires long-range interaction between DNA sites. This is facilitated by one-dimensional diffusion (‘DNA sliding’) initiated by ATP hydrolysis catalyzed by a superfamily 2 helicase-like ATPase. Here we combined ultrafast twist measurements based on plasmonic DNA origami nano-rotors with stopped-flow fluorescence and gel-based assays to examine the role(s) of ATP hydrolysis. Our data show that the helicase-like domain has multiple roles. First, this domain stabilizes initial DNA interactions alongside the methyltransferase subunits. Second, it causes environmental changes in the flipped adenine base following hydrolysis of the first ATP. Finally, it remodels nucleoprotein interactions via constrained translocation of a ∼ 5 to 22-bp double stranded DNA loop. Initiation of DNA sliding requires 8–15 bp of DNA downstream of the motor, corresponding to the site of nuclease domain binding. Our data unify previous contradictory communication models for Type III enzymes.

Screening of Diabetes-Associated Autoantigens and Serum Antibody Profiles Using a Phage Display System.

Aims/Introduction: Phage display method is a crucial tool to find novel clinically valuable diabetes-associated autoantigens and identify known autoantigen epitopes that are associated with diabetes and could provide scientific support and guidance for the artificial construction and synthesis of Type I diabetes mellitus (T1DM) novel biomarkers. Materials and Methods: The phage display system was used for the "biopanning" of T1DM serum. Following the sequencing of the phage DNAs, the homologous sequences of the above fusion heptapeptide were further investigated by BLAST to track the origin of the polypeptide sequences. The antibody spectrum revealed new T1DM-associated epitopes and antibodies. Results: A total of 1200 phage DNA were sequenced and 9 conserved polypeptide sequences were collected. It was confirmed that the zinc transporter and islet amyloid protease were among them. The conserved polypeptide sequence 8 and another three distinctive polypeptide sequences derived from Proteus were discovered. Furthermore, we expressed recombinant proteins with homologous polypeptide sequences for the human islet amyloid polypeptide (IAPP) and polypeptide precursor human zinc transporter 8 (ZNT8). Through clinical sample detection for the serum from T1DM (n = 100) and T2DM (n = 200) patients, results demonstrate the importance and relevance of these polypeptides in the recognition and classification of various forms of diabetes. Conclusion: Human pancreatic and concurrent bacterial-derived protein antigens and their epitopes were identified in this research by the phage display system, which is crucial for distinguishing different types of diabetes.

Development of PCR primers enabling the design of flexible sticky ends for efficient concatenation of long DNA fragments.

We developed chemically modified PCR primers that allow the design of flexible sticky ends by introducing a photo-cleavable group at the phosphate moiety. Nucleic acid derivatives containing o-nitrobenzyl photo-cleavable groups with a tert-butyl group at the benzyl position were stable during strong base treatment for oligonucleotide synthesis and thermal cycling in PCR reactions. PCR using primers incorporating these nucleic acid derivatives confirmed that chain extension reactions completely stopped at position 1 before and after the site of the photo-cleavable group was introduced. DNA fragments of 2 and 3 kbp, with sticky ends of 50 bases, were successfully concatenated with a high yield of 77%. A plasmid was constructed using this method. Finally, we applied this approach to construct a 48.5 kbp lambda phage DNA, which is difficult to achieve using restriction enzyme-based methods. After 7 days, we were able to confirm the generation of DNA of the desired length. Although the efficiency is yet to be improved, the chemically modified PCR primer offers potential to complement enzymatic methods and serve as a DNA concatenation technique.

Multiplex Functional Characterization of Protein Variant Libraries in Mammalian Cells with Single-Copy Genomic Integration and High-Throughput DNA Sequencing.

Sequencing-based, massively parallel genetic assays have enabled simultaneous characterization of the genotype-phenotype relationships for libraries encoding thousands of unique protein variants. Since plasmid transfection and lentiviral transduction have characteristics that limit multiplexing with pooled libraries, we developed a mammalian synthetic biology platform that harnesses the Bxb1 bacteriophage DNA recombinase to insert single promoterless plasmids encoding a transgene of interest into a pre-engineered "landing pad" site within the cell genome. The transgene is expressed behind a genomically integrated promoter, ensuring only one transgene is expressed per cell, preserving a strict genotype-phenotype link. Upon selecting cells based on a desired phenotype, the transgene can be sequenced to ascribe each variant a phenotypic score. We describe how to create and utilize landing pad cells for large-scale, library-based genetic experiments. Using the provided examples, the experimental template can be adapted to explore protein variants in diverse biological problems within mammalian cells.

Use of Specific Borrelia Phages as a New Strategy for Improved Diagnostic Tests.

The high failure rate of tick-borne infection (TBI)-related testing underscores the need for novel approaches that do not rely on serology and two-tier testing. Delayed diagnosis of TBIs, especially Borrelia infections, results in high healthcare costs and great suffering. There is a significant need for a reliable blood test that can aid in the diagnosis of Lyme disease, particularly when the current FDA-approved serological test is not sensitive enough to detect early Lyme patients who have not yet produced antibodies against Borrelia. Bacteriophages are viruses that specifically associate with their bacterial hosts, particularly prophages, bacteriophages residing in bacteria, and have proven to be tightly correlated with their bacterial hosts. They are poised to have wider applications as markers to detect bacteria, particularly in infectious disease. The gene of choice depends on the prevalence of phages within a particular group of bacteria. Phage genes that have been used as molecular markers to examine phage diversity include structural genes encoding the major capsid protein, the portal protein, the DNA polymerase, and the terminase. Borrelia species carry specific phage sequences that can be used as a proxy to identify the bacteria. Using phages as a proxy for bacteria is beneficial, as phages can be detected more easily than bacteria and can be used to bypass the cryptic and tissue-bound feature that typifies human Borrelia infections.We explored a completely new way of detecting Borrelia using Borrelia-specific bacteriophages as a diagnostic tool. Our detection method, patented by Phelix R&D and Leicester University (WO2018083491A1), could potentially transform infectious disease diagnostics through the innovative use of real-time PCR to target circulating bacteriophage DNA in blood from patients with Lyme disease. Firstly, this bacteriophage-based approach offers increased sensitivity since bacteriophages are typically present in five- to tenfold excess over bacterial cells, making it more accurate and sensitive than conventional bacteria-targeting PCR tests. One of the reasons bacteria-based PCR tests are frequently negative is due to the low bacterial concentration in the blood. Bacteriophage-based PCR surpasses this barrier and offers a direct test, as phages are part of bacteria's own genetic material, in contrast to all existing indirect tests (ELISA, Western BLOT, LTT/ELISPOT test). Secondly, a phage-based test can differentiate between different Lyme disease-causing and relapsing fever-causing Borrelia subtypes (B. burgdorferi s.l., B. miyamotoi, etc.), given that bacteriophages are indicators of bacterial identity. Finally, this test can detect Lyme disease in both early and late stages.

Viral Genomic DNA Packaging Machinery.

Tailed double-stranded DNA bacteriophage employs a protein terminase motor to package their genome into a preformed protein shell-a system shared with eukaryotic dsDNA viruses such as herpesviruses. DNA packaging motor proteins represent excellent targets for antiviral therapy, with Letermovir, which binds Cytomegalovirus terminase, already licensed as an effective prophylaxis. In the realm of bacterial viruses, these DNA packaging motors comprise three protein constituents: the portal protein, small terminase and large terminase. The portal protein guards the passage of DNA into the preformed protein shell and acts as a protein interaction hub throughout viral assembly. Small terminase recognises the viral DNA and recruits large terminase, which in turn pumps DNA in an ATP-dependent manner. Large terminase also cleaves DNA at the termination of packaging. Multiple high-resolution structures of each component have been resolved for different phages, but it is only more recently that the field has moved towards cryo-EM reconstructions of protein complexes. In conjunction with highly informative single-particle studies of packaging kinetics, these structures have begun to inspire models for the packaging process and its place among other DNA machines.

Genomic characterization of SNW-1, a novel prophage of the deep-sea vent chemolithoautotroph Sulfurimonas indica NW79.

The globally widespread genus Sulfurimonas are playing important roles in different habitats, including the deep-sea hydrothermal vents. However, phages infecting Sulfurimonas have never been isolated and characterized to date. In the present study, a novel prophage SNW-1 was identified from Sulfurimonas indica NW79. Whole genome sequencing resulted in a circular, double-stranded DNA molecule of 37,096 bp with a mol% G+C content of 37. The genome includes 64 putative open reading frames, 33 of which code for proteins with predicted functions. Presence of hallmark genes associated with Caudoviricetes and genes involved in lysis and lysogeny indicated that SNW-1 should be a temperate, tailed phage. Phylogenetic and comparative proteomic analyses suggested that Sulfurimonas phage SNW-1 was distinct from other double stranded DNA phages and might represent a new viral genus.

Infant gut DNA bacteriophage strain persistence during the first 3 years of life

Bacteriophages are key components of gut microbiomes, yet the phage colonization process in the infant gut remains uncertain. Here, we establish a large phage sequence database and use strain-resolved analyses to investigate DNA phage succession in infants throughout the first 3 years of life. Analysis of 819 fecal metagenomes collected from 28 full-term and 24 preterm infants and their mothers revealed that early-life phageome richness increases over time and reaches adult-like complexity by age 3. Approximately 9% of early phage colonizers, which are mostly maternally transmitted and infect Bacteroides, persist for 3 years and are more prevalent in full-term than in preterm infants. Although rare, phages with stop codon reassignment are more likely to persist than non-recoded phages and generally display an increase in in-frame reassigned stop codons over 3 years. Overall, maternal seeding, stop codon reassignment, host CRISPR-Cas locus prevalence, and diverse phage populations contribute to stable viral colonization.

Emergence of cefiderocol resistance during ceftazidime/avibactam treatment caused by a large genomic deletion, including ampD and piuCD genes, in Pseudomonas aeruginosa.

We report the emergence of cefiderocol resistance during the treatment of a ST312 Pseudomonas aeruginosa respiratory infection with ceftazidime/avibactam. whole genome sequencing (WGS) revealed that resistance was caused by a large genomic deletion, including PiuDC (iron transport system) and AmpD (ampC negative regulator), driven by the integration of phage DNA. Thus, our findings alert that this type of deletion could be an efficient (two mechanisms in one step) specific cefiderocol resistance mechanism that might occur nonspecifically upon treatment with β-lactams that select for AmpC overexpression.

Enrichment, Sequencing, and Identification of DNA Bacteriophages from Fecal Samples.

Research on individual viruses and phages, as well as viral populations (viromes), is greatly expanding. Phages and viromes are increasingly suspected to have numerous impacts on the ecosystem in which they reside by interacting directly or indirectly with the other organisms present in their environment. In particular, phage communities of the gut microbiota have been associated with a wide range of diseases. However, properly investigating intestinal viromes is still very challenging, both experimentally and analytically. This chapter proposes a simple and reproducible protocol to separate and enrich DNA phage particles from fecal samples, to sequence them, and finally obtain a basic but robust bioinformatic characterization and classification of the global bacteriophage community.

Comparative genomics and proteomics analysis of phages infecting multi-drug resistant Escherichia coli O177 isolated from cattle faeces

The increasing prevalence of antimicrobial-resistant (AMR) pathogens has become a major global health concern. To address this challenge, innovative strategies such as bacteriophage therapy must be optimised. Genomic characterisation is a crucial step in identifying suitable phage candidates for combating AMR pathogens. The aim of this study was to characterise seven phages that infect the Escherichia coli O177 strain using a whole genome sequencing. The analysis of genome sequences revealed that these phages had linear dsDNA, with genome sizes spanning from 136, 483 to 166,791 bp and GC content varying from 35.39 to 43.63%. Taxonomically, the phages were classified under three different subfamilies (Stephanstirmvirinae, Tevenvirinae, and Vequintavirinae) and three genera (Phapecoctavirus, Tequatrovirus, and Vequintavirus) within the class Caudoviricetes. In silico PhageAI analysis predicted that all the phages were virulent, with confidence levels between 96.07 and 97.26%. The phage genomes contained between 66 and 82 ORFs, which encode hypothetical and putative functional proteins. In addition, the phage genomes contained core genes associated with molecular processes such as DNA replication, transcription modulation, nucleotide metabolism, phage structure (capsid and tail), and lysis. None of the genomes carried genes associated with undesirable traits such as integrase, antimicrobial resistance, virulence, and toxins. The study revealed high genome and proteome homology among E. coli O177 phages and other known Escherichia phages. The results suggest that the seven phages are new members of the genera Phapecoctavirus, Tequatrovirus, and Vequintavirus under the subfamilies Stephanstirmvirinae, Tevenvirinae, and Vequintavirinae, respectively.

Novel WYL domain-containing transcriptional activator acts in response to genotoxic stress in rapidly growing mycobacteria

The WYL domain is a nucleotide-sensing module that controls the activity of transcription factors involved in the regulation of DNA damage response and phage defense mechanisms in bacteria. In this study, we investigated a WYL domain-containing transcription factor in Mycobacterium smegmatis that we termed stress-involved WYL domain-containing regulator (SiwR). We found that SiwR controls adjacent genes that belong to the DinB/YfiT-like putative metalloenzymes superfamily by upregulating their expression in response to various genotoxic stress conditions, including upon exposure to H2O2 or the natural antibiotic zeocin. We show that SiwR binds different forms of single-stranded DNA (ssDNA) with high affinity, primarily through its characteristic WYL domain. In combination with complementation studies of a M. smegmatis siwR deletion strain, our findings support a role of the WYL domains as signal-sensing activity switches of WYL domain-containing transcription factors (WYL TFs). Our study provides evidence that WYL TFs are involved in the adaptation of bacteria to changing environments and encountered stress conditions.

RNA and Single-Stranded DNA Phages: Unveiling the Promise from the Underexplored World of Viruses

RNA and single-stranded DNA (ssDNA) phages make up an understudied subset of bacteriophages that have been rapidly expanding in the last decade thanks to advancements in metaviromics. Since their discovery, applications of genetic engineering to ssDNA and RNA phages have revealed their immense potential for diverse applications in healthcare and biotechnology. In this review, we explore the past and present applications of this underexplored group of phages, particularly their current usage as therapeutic agents against multidrug-resistant bacteria. We also discuss engineering techniques such as recombinant expression, CRISPR/Cas-based genome editing, and synthetic rebooting of phage-like particles for their role in tailoring phages for disease treatment, imaging, biomaterial development, and delivery systems. Recent breakthroughs in RNA phage engineering techniques are especially highlighted. We conclude with a perspective on challenges and future prospects, emphasizing the untapped diversity of ssDNA and RNA phages and their potential to revolutionize biotechnology and medicine.

The mobilome of Lactobacillus crispatus M247 includes two novel genetic elements: Tn7088 coding for a putative bacteriocin and the siphovirus prophage ΦM247.

Lactobacillus crispatus is a member of the vaginal and gastrointestinal human microbiota. Here we determined the complete genome sequence of the probiotic strain M247 combining Nanopore and Illumina technologies. The M247 genome is organized in one circular chromosome of 2 336 109 bp, with a GC content of 37.04 % and 2303 ORFs, of which 1962 could be annotated. Analysis of the M247 mobilome, which accounts for 14 % of the whole genome, revealed the presence of: (i) Tn7088, a novel 14 105 bp long integrative and mobilizable element (IME) containing 16 ORFs; (ii) ΦM247, a novel 42 510 bp long siphovirus prophage containing 52 ORFs; (iii) three clustered regularly interspaced short palindromic repeats (CRISPRs); and (iv) 226 insertion sequences (ISs) belonging to 14 different families. Tn7088 has a modular organization including a mobilization module encoding FtsK homologous proteins and a relaxase, an integration/excision module coding for an integrase and an excisionase, and an adaptation module coding for a class I bacteriocin and homologous to the listeriolysin S (lls) locus of Listeria monocytogenes. Genome-wide homology search analysis showed the presence of Tn7088-like elements in 12 out of 23 L. crispatus complete public genomes. Mobilization and integration/excision modules are essentially conserved, while the adaptation module is variable since it is the target site for the integration of different ISs. Prophage ΦM247 contains genes for phage structural proteins, DNA replication and packaging, lysogenic and lytic cycles. ΦM247-like prophages are present in seven L. crispatus complete genomes, with sequence variability mainly due to the integration of ISs. PCR and sequencing showed that the Tn7088 IME excises from the M247 chromosome producing a circular form at a concentration of 4.32×10-5 copies per chromosome, and reconstitution of the Tn7088 chromosomal target site occurred at 6.65×10-4 copies per chromosome. The ΦM247 prophage produces an excised form and a reconstituted target site at a level of 3.90×10-5 and 2.48×10-5 copies per chromosome, respectively. This study identified two novel genetic elements in L. crispatus. Tn7088 represents the first example of an IME carrying a biosynthetic gene cluster for a class I bacteriocin in L. crispatus.

Characterizations of novel broad-spectrum lytic bacteriophages Sfin-2 and Sfin-6 infecting MDR Shigella spp. with their application on raw chicken to reduce the Shigella load

The evidence and prevalence of multidrug-resistant (MDR) Shigella spp. poses a serious global threat to public health and the economy. Food- or water-borne MDR Shigella spp. demands an alternate strategy to counteract this threat. In this regard, phage therapy has garnered great interest from medical practitioners and researchers as a potential way to combat MDR pathogens. In this observation, we isolated Shigella phages from environmental water samples and tested against various clinically isolated MDR Shigella spp. In this study, we have defined the isolation and detailed physical and genomic characterizations of two phages Sfin-2 and Sfin-6 from environmental water samples. The phages exhibited potent lytic activity against Shigella flexneri, Shigella dysenteriae, and Shigella sonnei. They showed absorption within 5–10 min, a burst size ranging from ~74 to 265 PFU/cell, and a latent period of 5–20 min. The phages were stable at a broad pH range and survived an hour at 50°C. The purified phages Sfin-2 and Sfin-6 belong to the Siphoviridae family with an isometric head (64.90 ± 2.04 nm and 62.42 ± 4.04 nm, respectively) and a non-contractile tail (145 ± 8.5 nm and 148.47 ± 14.5 nm, respectively). The in silico analysis concluded that the size of the genomic DNA of the Sfin-2 phage is 50,390 bp with a GC content of 44.90%, while the genome size of the Sfin-6 phage is 50,523 bp with a GC content of 48.30%. A total of 85 and 83 putative open reading frames (ORFs) were predicted in the Sfin-2 and Sfin-6 phages, respectively. Furthermore, a comparative genomic and phylogenetic analysis revealed that both phages represented different isolates and novel members of the T1-like phages. Sfin-2 and Sfin-6 phages, either individually or in a cocktail form, showed a significant reduction in the viable Shigella count on raw chicken samples after 72 h of incubation. Therefore, these results indicate that these phages might have a potential role in therapeutic approaches designed for shigellosis patients as well as in the biological control of MDR Shigella spp. in the poultry or food industry during the course of meat storage.

Structural basis of Gabija anti-phage defence and viral immune evasion

Bacteria encode hundreds of diverse defence systems that protect them from viral infection and inhibit phage propagation1–5. Gabija is one of the most prevalent anti-phage defence systems, occurring in more than 15% of all sequenced bacterial and archaeal genomes1,6,7, but the molecular basis of how Gabija defends cells from viral infection remains poorly understood. Here we use X-ray crystallography and cryo-electron microscopy (cryo-EM) to define how Gabija proteins assemble into a supramolecular complex of around 500 kDa that degrades phage DNA. Gabija protein A (GajA) is a DNA endonuclease that tetramerizes to form the core of the anti-phage defence complex. Two sets of Gabija protein B (GajB) dimers dock at opposite sides of the complex and create a 4:4 GajA–GajB assembly (hereafter, GajAB) that is essential for phage resistance in vivo. We show that a phage-encoded protein, Gabija anti-defence 1 (Gad1), directly binds to the Gabija GajAB complex and inactivates defence. A cryo-EM structure of the virally inhibited state shows that Gad1 forms an octameric web that encases the GajAB complex and inhibits DNA recognition and cleavage. Our results reveal the structural basis of assembly of the Gabija anti-phage defence complex and define a unique mechanism of viral immune evasion.