Structure and rational engineering of the PgIX methyltransferase and specificity factor for BREX phage defence

Bacteriophages (phages) pose a significant threat to bacteria, driving the evolution of diverse bacterial defense systems. These systems, often clustered in defense islands, include well-known mechanisms like restriction-modification (RM), abortive infection, and CRISPR-Cas. Recent research has uncovered numerous additional defense systems, such as BREX, CBASS, BstA, retrons, viperins, pycsar, and PARIS, often identified through guilt-by-association approaches or functional selection. Many of these systems are regulated by conserved systems, such as the BrxR family, suggesting coordinated responses to phage infection. BREX systems are found in approximately 10% of bacterial and archaeal genomes and are related to the Pgl system. They are frequently found alongside GmrSD-family type IV restriction enzymes. Type I BREX systems consist of six genes: *brxA*, *brxB*, *brxC*, *pglX*, *pglZ*, and *brxL*. BrxA is a DNA-binding protein and BrxL is a DNA-stimulated AAA+ ATPase. PgIX, a key component, shows homology to methyltransferases and is hypothesized to methylate non-palindromic 6 bp sequences (BREX motifs) at the N6 adenine in the fifth position, distinguishing self from non-self DNA. Ocr, a phage T7 protein mimicking dsDNA, inhibits BREX by binding to PgIX. The *Salmonella enterica* serovar Typhimurium StySA locus (SenLT2III) exhibits BREX activity and is found in the clinically relevant ST313 strain D23580, encoding a BREX locus closely related to the LT2 BREX locus, forming a defense island with PARIS. The relative simplicity and clinical importance of the *Salmonella* BREX system, combined with the availability of a Durham phage collection, made it a suitable model for investigating BREX function. This study aimed to investigate the activity of the *Salmonella* D23580 BREX system, elucidate the role of individual BREX genes, characterize the structure of PgIX, and explore the potential for rational engineering of the system.

Previous research has established the existence and prevalence of various bacterial defense systems against phage infection. The discovery and characterization of BREX as a novel phage resistance system, its relationship to the Pgl system, and its frequent co-localization with Type IV restriction enzymes have been documented. Studies have also explored the individual roles of BREX genes, particularly *brxA*, *brxB*, *brxC*, *pglX*, *pglZ*, and *brxL*. The functional roles of BrxA (DNA binding) and BrxL (DNA-stimulated AAA+ ATPase) have been partially characterized. The mechanism of BREX activity, involving methylation of specific DNA motifs, has been explored. Previous work highlighted the role of Ocr, a phage-encoded protein mimicking dsDNA, in inhibiting BREX activity. The *Salmonella* StySA locus has been investigated in the context of BREX activity. Genomic analyses have revealed the prevalence of BREX and its association with other defense systems. The study builds upon this foundation, providing structural and functional insights into the *Salmonella* BREX system.

The study employed a multi-faceted approach combining microbiology, genetics, structural biology, and bioinformatics. First, the activity of the *Salmonella* D23580 BREX defense island (BREX_sty) was confirmed using efficiency of plating (EOP) assays against environmentally isolated *Salmonella* phages after generating a BREX-deficient strain (D23850ΔBREX). The role of each gene in BREX_sty was investigated through systematic gene deletions in *E. coli*, using TB34 and T7 phages to assess the impact on phage defense. The X-ray crystal structures of PgIX, both alone (bound to SAM) and in complex with Ocr, were determined. The crystal structures were solved using molecular replacement, using AlphaFold predicted model for PgIX. Methylation patterns were analyzed using MinION and PacBio sequencing to identify the methylation motifs. A SAM-dependent methyltransferase assay was performed with purified PgIX to assess its in vitro activity. The effect of Ocr and its *Salmonella* homologue Gp5 on BREX activity was tested using EOP assays. Analytical size-exclusion chromatography (SEC) was used to determine the solution state of PgIX and the PgIX:Ocr complex. Rational engineering of PgIX was performed to modify the BREX motif recognized by altering specific amino acid residues within the protein, based on sequence alignments and structural comparisons with Mmel, a closely related methyltransferase. The functional effects of these mutations were assessed using EOP assays and PacBio sequencing to confirm changes in methylation patterns. Mass photometry was employed to assess the interaction of PgIX variants (wild-type, N-terminal domain only, and a C-terminal mutant) with both DNA and Ocr. Bioinformatics tools, including DELTA-BLAST, HMMER, Biopython, PADLOC v2.0.0, and Defence-finder v1.2.0, were used to analyze genomic data and predict defense systems. Standard molecular biology techniques like cloning (LIC and Gibson Assembly), PCR, and gene knockouts were used. Protein expression and purification involved standard protocols employing affinity chromatography (Ni-NTA, Heparin) and SEC.

The *Salmonella* D23580 BREX_sty island provides protection against environmentally isolated *Salmonella* phages. Genetic analysis revealed that *brxA*, *brxB*, *brxC*, *pglX*, and *pglZ* are essential for both phage defense and methylation. Surprisingly, *brxL* was not essential for phage defense in *Salmonella*, unlike in *E. coli* and *Acinetobacter*. The crystal structure of PgIX revealed two distinct domains: an N-terminal methyltransferase domain (binding SAM) and a C-terminal domain involved in motif recognition. The PgIX-Ocr complex structure showed that Ocr inhibits BREX by forming a heterotetrameric complex with PgIX, preventing DNA binding. Two potential DNA-binding modes were proposed based on structural comparisons with EcoKI and Mmel. Mass photometry confirmed that the C-terminal domain of PgIX is essential for both DNA and Ocr binding. Rational engineering of PgIX successfully altered the BREX motif recognized and expanded the range of phages targeted by the system. Specifically, the T802A and S838N mutations in PgIX broadened recognition to include both GATCAG and GATAAG motifs, resulting in the new GATMAG motif.

This study provides comprehensive insights into the structure and function of the *Salmonella* BREX system. The finding that *brxL* is dispensable for phage defense in *Salmonella* contrasts with previous findings in other species, suggesting that the role of BrxL might vary depending on the bacterial host. The structures of PgIX, both alone and in complex with Ocr, reveal the molecular mechanisms underlying BREX activity and its inhibition. The identification of two potential DNA-binding modes highlights the complexity of the system. The successful rational engineering of PgIX demonstrates the feasibility of manipulating BREX specificity for targeted phage defense and methylation pattern generation. The observation that PgIX alone lacks in vitro methylation activity suggests that higher-order complexes are required for optimal function. The work suggests that the C-terminal domain of PgIX plays a crucial role in DNA recognition and binding. The surprising lack of PARIS activity in the current study requires further investigation, with the possibility that a different susceptible phage remains to be identified.

This research provides the first detailed structural and functional characterization of the *Salmonella* BREX phage defense system. The study definitively identifies PgIX as the key specificity factor, critical for both host methylation and phage targeting. The successful engineering of PgIX to alter both phage susceptibility and methylation patterns highlights the potential for developing tailored BREX systems for various applications. Future research could focus on elucidating the role of other BREX components in complex formation and function, clarifying the role of BrxL, and identifying phages susceptible to the PARIS system.

The study primarily focused on the *Salmonella* D23580 BREX system and its interaction with a limited set of phages. In vitro methylation assays with PgIX alone were unsuccessful, possibly due to the need for additional BREX components for full activity. The exact mechanism of DNA binding by PgIX requires further investigation, as the crystal structure with DNA could not be obtained. The unexpected lack of PARIS system activity remains to be resolved. The functional assays in *E. coli* may not entirely reflect the behavior of the system in its natural *Salmonella* host.