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

The study addresses how BREX systems recognise their cognate non-palindromic 6-bp DNA motifs to distinguish self from non-self and prevent phage replication. Although BREX is widespread and related to Pgl systems, the mechanism of motif recognition has remained unclear. The authors focus on the Salmonella Typhimurium ST313 D23580 BREX island (BREXsty), which is clinically relevant and comparatively simple, to determine genetic requirements for defence and methylation, define the structure and function of the conserved methyltransferase PgIX, and test whether PgIX dictates recognition specificity. The work aims to clarify PgIX’s roles in methylation and phage defence, delineate DNA-binding modes, assess inhibition by phage DNA mimics (Ocr), and enable rational retargeting of BREX specificity to broaden anti-phage activity and host methylation motifs.

Defense islands in bacteria encode diverse anti-phage systems including RM, abortive infection, CRISPR-Cas, CBASS, retrons, viperins, Pycsar, BstA and PARIS. BREX is common (~10% of bacterial/archaeal genomes) and often co-localizes with Type IV restriction enzymes (GmrSD) or PARIS. Prior work established that type I BREX uses PglX to N6-methylate adenine at position 5 of a 6-bp non-palindromic motif for self/non-self discrimination. Ocr, a phage T7 DNA mimic, inhibits E. coli BREX by binding PgIX. The Salmonella LT2 StySA system targets GATCAG. Some components (brxB, brxC, pglX, pglZ) are essential for restriction and methylation; brxL’s role varies across species. Structural homologs (e.g., Type IIL RM enzyme MmeI) provided insights into methyltransferase fold and target recognition domains, but BREX motif recognition mechanisms and PgIX structural details were unresolved.

- Strains and phages: Constructed Salmonella D23580 Δprophage and ΔBREX strains; cloned the D23580 BREX-PARIS island into E. coli (pBrxXLsty) and generated individual gene deletions (brxA, brxB, brxC, pglX, pglZ, brxL) and PARIS deletions (ariA, ariB, double). Isolated Salmonella phages from sewage/river water; used Durham coliphage collection (12 phages) for E. coli assays. - Phage defence assays: Monitored growth/infection curves and performed efficiency of plating (EOP) assays comparing test strains to controls. Assessed inhibition of BREX via induced expression of phage Ocr (T7) or Gp5 (Sp6) from pBAD30 vectors. - Methylation detection: Extracted genomic DNA and assessed adenine methylation using Oxford Nanopore MinION (Megalodon) and PacBio SMRT sequencing (SMRTLink Base Modification Analysis). Included WGA control to remove modifications. Determined target motifs and percent methylation genome-wide. - Bioinformatics: Surveyed co-localization of BREX with other defence systems using DELTA-BLAST to identify pglZ/pglX homologs, HMMER to filter genomes encoding PgIZ and PgIX, and PADLOC/Defense Finder to annotate defence systems between pglZ and pglX. - Protein work: Expressed and purified PgIX, attempted CTD/NTD constructs, and purified Ocr. Performed analytical SEC to assess oligomeric states and complex formation with Ocr. Conducted mass photometry to measure complex masses and test DNA/Ocr binding with PgIX variants (WT, NTD-only, and a CTD multi-site mutant targeting charged residues involved in Ocr binding). Attempted in vitro methyltransferase assays (MTase-Glo) using genomic DNA substrate with/without additional BREX proteins (PgIZ, BrxB) and SAM. - Crystallography: Solved X-ray structures of Salmonella PgIX bound to SAM (PDB 8C45) and PgIX-SAM bound to Ocr (PDB 8Q56) at 3.5 Å, via molecular replacement using AlphaFold PgIX model and Ocr structure (1S7Z). Analyzed domain architecture, SAM-binding motifs (GxG, NPPY), and interfaces with Ocr (salt bridges, H-bonds). Compared to MmeI (5HR4) and RM Type I specificity subunits to infer target recognition domain (TRD) and potential DNA-binding modes. - Rational engineering: Identified candidate PgIX residues for motif specificity based on alignment with MmeI specificity determinants and REBASE motif associations. Designed 23 pglX mutants targeting five non-modified positions in the 6-bp motif, screened using a complementation system (pBrxXLsty-ΔpglX + pBAD30-pglX mutant) in E. coli with phages TB34 and Trib (which lacks native Salmonella BREX motifs but encodes predicted re-engineered motifs). Validated active mutants by PacBio methylome analysis.

- BREX activity in Salmonella: D23580 BREXsty confers phage defence against environmental Salmonella phages; e.g., SCW1 EOP ~3.33×10^-4 ± 1.43×10^-4, SCW3 EOP ~0.18 ± 0.09, while others were largely unaffected. - Genetic requirements in E. coli: In pBrxXLsty background, brxA, brxB, brxC, pglX and pglZ are required for defence against TB34 and for host methylation; brxL is dispensable and its deletion can enhance defence (up to ~10,000-fold improvement in EOP reduction vs WT for TB34). PARIS genes (ariA/ariB) were not required for observed defence or methylation and did not show activity under tested conditions. - Target motif and methylation: PacBio identified near-complete methylation of GATCAG sites in WT pBrxXLsty (≈97.8–100%), confirming the StySA-like motif with N6mA at the 5th adenine. WGA controls showed 0% modified by PacBio. MinION trends agreed but with less quantitative dynamic range. - Structural biology of PgIX: PgIX comprises an N-terminal methyltransferase domain (harbouring GxG motif for SAM binding and NPPY motif for adenine interaction; γ-class amino-methyltransferase) and a C-terminal domain with a TRD-like region and extended helical spacer reminiscent of Type I RM specificity subunits. SAM was observed in the conserved pocket. Lack of nuclease motifs supports a methyltransferase-only role. - Ocr inhibition and complex formation: Ocr (and Salmonella phage Sp6 Gp5) fully inhibited BREX defence in E. coli carrying pBrxXLsty. Analytical SEC and crystallography showed PgIX binds an Ocr dimer to form a heterotetramer (2 PgIX monomers + Ocr dimer). Interfaces include multiple salt bridges (e.g., Ocr D35/D42/D62/D76/E109/R79 with PgIX K1201/K1097/K1070/K1110/D1213/K516). Ocr binding does not induce large PgIX domain rearrangements, consistent with DNA mimic sequestration. - DNA-binding modes and CTD role: Structural superpositions with EcoKI Ocr/DNA complexes and the DNA-bound MmeI suggest alternative DNA-binding orientations; a positively charged hinge/CTD groove likely accommodates DNA with a flipped adenine near the SAM pocket. Mass photometry showed WT PgIX forms specific complexes with 120-bp DNAs, while an NTD-only construct and a CTD multi-site mutant (mutating charged residues implicated in Ocr binding) failed to bind DNA or Ocr, indicating essential roles of CTD charged residues for binding. - Rational engineering of specificity: Among 23 designed PgIX variants, mutant 3 (T802A, S838N; residues aligned to MmeI specificity determinants) conveyed defence against Trib and TB34 in complementation and in-locus contexts. PacBio methylomes revealed broadened motif recognition from GATCAG to GATMAG (recognizing both GATCAG and GATAAG) with ~99.6–99.8% methylation, demonstrating PgIX as the sole specificity factor for both host methylation and phage targeting and showing successful engineering to retarget BREX specificity.

The findings resolve a central question in BREX biology: PgIX is the methyltransferase that also provides target specificity for BREX, linking host methylation and anti-phage discrimination. Structural elucidation of PgIX with SAM identifies the catalytic architecture and demonstrates a TRD-like CTD that mediates DNA recognition. The PgIX:Ocr heterotetramer explains potent Ocr-mediated inhibition via DNA mimicry and sequestration of the recognition interface. Genetic and biochemical data show that BREX defence and methylation depend on brxA, brxB, brxC, pglX and pglZ, but not brxL in Salmonella; notably, brxL deletion can enhance defence for some phages, suggesting BrxL may modulate or regulate BREX activity rather than act as an essential effector in this system. The inability to observe PgIX-only methylation in vitro and the structural models imply that larger BREX assemblies are needed for productive DNA engagement and catalysis, potentially involving coordinated recognition of non-palindromic motifs on dsDNA by multiple PgIX units. Rational engineering of PgIX broadened recognition (GATMAG) and shifted phage susceptibility profiles (enabling defence against Trib), indicating that specificity is tunable by targeted substitutions in the TRD. These insights refine the mechanistic model of BREX specificity and provide a foundation for programmable phage defence and epigenomic editing via engineered BREX.

This work provides the first X-ray structures of the BREX methyltransferase PgIX (with SAM and in complex with the DNA mimic Ocr), demonstrates that PgIX is the specificity factor for motif recognition in both phage defence and host methylation, and maps key CTD residues required for DNA and Ocr binding. Genetic dissection shows that brxA, brxB, brxC, pglX and pglZ are essential for defence and methylation in the Salmonella BREX system, while brxL is dispensable and can modulate defence strength. The target methylation motif in Salmonella D23580 is GATCAG with near-saturated N6mA modification in vivo. Engineering PgIX specificity (T802A, S838N) broadened motif recognition to GATMAG and retargeted BREX to new phages, validating the feasibility of programmable BREX. Future work should: (i) resolve structures of PgIX (and higher-order BREX complexes) bound to cognate DNA to define the catalytic trajectory and domain dynamics; (ii) determine the composition and stoichiometry of functional BREX assemblies required for methylation and restriction; (iii) clarify BrxL’s regulatory or effector roles across diverse BREX systems; and (iv) expand rational design strategies to reprogram BREX specificity for phage control and epigenetic applications.

- PgIX did not exhibit detectable in vitro methyltransferase activity alone or with limited partner proteins, suggesting missing components, stoichiometries, or conditions required for activity. - Crystallization with DNA was unsuccessful; DNA-binding modes are inferred from superpositions and electrostatics rather than direct DNA-bound structures. - PARIS activity was not observed under tested conditions; its contribution to defence in Salmonella remains unresolved. - Some defence and methylation assessments used a heterologous E. coli host, which may not fully capture native regulation or interactions. - Nanopore-based methylation quantification showed reduced dynamic range compared to PacBio; quantitative conclusions relied on PacBio data. - Only one engineered PgIX mutant demonstrated robust functional effects; the broader design space and rules for specificity modulation require further exploration.