Systematic discovery of antiphage defense systems in the microbial pangenome

Abstract

The arms race between bacteria and phages led to the development of sophisticated antiphage defense systems, including CRISPR-Cas and restriction-modification systems. Evidence suggests that known and unknown defense systems are located in "defense islands" in microbial genomes. Here, we comprehensively characterized the bacterial defensive arsenal by examining gene families that are clustered next to known defense genes in prokaryotic genomes. Candidate defense systems were systematically engineered and validated in model bacteria for their antiphage activities. We report nine previously unknown antiphage systems and one antiplasmid system that are widespread in microbes and strongly protect against foreign invaders. These include systems that adopted components of the bacterial flagella and condensin complexes. Our data also suggest a common, ancient ancestry of innate immunity components shared between animals, plants, and bacteria.

Figure 1. Discovery of new anti-phage defense systems in defense islands.

(A) Illustration of the computational analysis employed for each pfam found to be enriched in defense islands. Pfams that are enriched in the vicinity of known defense genes are identified, and their neighboring genes are clustered based on sequence homology to identify conserved cassettes that represent putative defense systems. (B) Tendency of protein families to occur next to defense genes. The genomic neighborhood for each member gene in each pfam is examined, and the fraction of member genes occurring in the vicinity (10 genes on each side) of one or more known defense genes is recorded. Pink, a set of 123 pfams known to participate in anti-phage defense (“positive set”); blue, the remaining 13,960 pfams analyzed in this study. (C) Neighborhood variability score for the analyzed pfams. Score represents the fraction of pfam members occurring in different defense neighborhoods out of total occurrences of pfam members (see Methods). Pink, the 123 positive pfams; blue, a set of 576 pfams that passed the 65% threshold for fraction of members occurring with defense genes in proximity.

Figure 2. Experimentally verified defense systems.

(A) Flowchart of the experimental verification strategy. (B) Active defense systems cloned into B. subtilis. (C) Active defense systems cloned into E. coli. For B-C, fold protection was measured using serial dilution plaque assays, comparing the system-containing strain to a control strain that lacks the system and has an empty vector instead. Data represent average of 3 replicates, see Figures S2 and S3. Numbers below phage names represent phage genome size. On the right, gene organization of the defense systems, with identified domains indicated (DUF, domain of unknown function). Gene sizes are drawn to scale; scale bar represents 400 amino acids.

Figure 3. The Zorya system.

(A) Representative instances of the Type I Zorya system and their defense island context. Genes known to be involved in defense are orange. Mobilome genes are in dark grey. RM, restriction-modification; TA, toxin-antitoxin; Abi, abortive infection; Wadjet and Druantia are systems identified as defensive in this study (see below). (B) Representative instances of the Type II Zorya system. (C) Domain organization of the two types of Zorya. (D) Model of the flagellum base. The position of the MotAB complex is indicated. (E) Efficiency of plating (EOP) of phage SECphi27 infecting WT Type I Zorya, deletion strains, and strains with point mutations. Data represent PFU/ml, average of 3 replicates with error bars representing STD. ZorA:T147A/S184A and ZorB:D26N are predicted to abolish proton flux; ZorC:E400A/H443A are mutations in two conserved residues in pfam15611 (“EH domain”) whose function is unknown (23); ZorD:D730A/E731A are mutations in the Walker B motif, predicted to abolish ATP hydrolysis.

Figure 4. The Thoeris system.

(A) Representative instances of the Thoeris system and their defense island context. Thoeris genes thsA (containing NAD+ binding domain) and thsB (TIR domain) are marked dark and light green, respectively. Genes known to be involved in defense are orange. Mobilome genes are in dark grey. RM, restriction-modification; TA, toxin-antitoxin; Abi, abortive infection. (B) The two Thoeris systems shown in this study to protect against myophages. Locus tag accessions are indicated for the individual genes. (C) EOP of phage SBSphiJ infection with WT and mutated versions of the B. amyloliquefaciens Y2 Thoeris (top set) or B. cereus MSX-D12 Thoeris (bottom set) cloned into B. subtilis BEST7003. Average of 3 replicates, error bars represent STD.

Figure 5. The Wadjet system provides protection against plasmid transformation in B. subtilis.

(A) Representative instances of the Wadjet system and their defense island context. Genes known to be involved in defense are orange. RM, restriction-modification; TA, toxin-antitoxin; Abi, abortive infection. (B) Domain organization of the three types of Wadjet. Pfam and COG domains were assigned according to the information in the IMG database (48). (C) Wadjet reduces plasmid transformation efficiency in B. subtilis. Instances of Wadjet systems were taken from Bacillus cereus Q1 (Type I), Bacillus vireti LMG 21834 (Type II) and Bacillus thuringiensis serovar finitimus YBT-020 (Type III) (Table S4) and cloned into B. subtilis BEST7003. Gene deletions and point mutations are of the B. cereus Q1 Type I Wadjet. Transformation efficiency of plasmid pHCMC05 into Wadjet-containing strains is presented as a percentage of the transformation efficiency to B. subtilis BEST7003 carrying an empty vector instead of the Wadjet system. Average of 3 replicates; error bars represent STD.