Journal: bioRxiv

Article Title: A methylome-derived m 6 -dAMP trigger assembles a PUA-Cal-HAD immune filament that depletes dNTPs to abort phage infection

doi: 10.64898/2026.01.15.699771

Figure Lengend Snippet: PUA senses m 6 -dAMP to activate nucleotide depletion by Cal and HAD. (A) Quantification of nucleotide metabolites by LC-MS, shown as relative intensity levels in log10, in cells expressing YFP or PUA-Cal-HAD upon infection with phage T2. (B) Infectivity of T2 phage escape mutants on YFP- and PUA-Cal-HAD-expressing cells. Escape mutants have mutations in gene denA . (C) Infectivity of Bas38 and T2 phages on YFP (control) or PUA-Cal-HAD cells containing both methylases Dcm and Dam (BW25113), or versions with deleted methylases. (D) LC-MS measurement of PUA-Cal-HAD in vitro activity on dNTPs in the presence of m 6 -AMP or m 6 -dAMP. (E) Reaction scheme for PUA-Cal-HAD activity. Step 1: HAD hydrolyses the β-γ phosphoanhydride bond of dATP to generate dADP and inorganic phosphate (Pi). Calcineurin-CE then hydrolyses the α-β phosphoanhydride bond of dADP to yield dAMP and Pi.

Article Snippet: The total run time was 8 min with a gradient between 1 to 5 min. m 6 -dAMP was prepared by digesting 10 mM m 6 -dATP with Apyrase (New England Biolabs) for 1h at 37°C, following manufacturer instructions, followed by filtration with an Amicon Ultra-0.5 Centrifugal Filter Unit 3LJkDa.

Techniques: Liquid Chromatography with Mass Spectroscopy, Expressing, Infection, Control, In Vitro, Activity Assay

Journal: bioRxiv

Article Title: A methylome-derived m 6 -dAMP trigger assembles a PUA-Cal-HAD immune filament that depletes dNTPs to abort phage infection

doi: 10.64898/2026.01.15.699771

Figure Lengend Snippet: Sensing of m 6 -dAMP by PUA drives PUA-Cal hexamer reorganisation and filament formation. (A) Molecular mass distributions measured by mass photometry for purified PUA-Cal in the absence and presence of HAD and m 6 -dAMP. Data are shown as the percentage of particles per molecular mass (kDa). (B) Representative cryo-EM micrograph of PUA-Cal mixed with HAD and m 6 -dAMP, showing filamentous assemblies (white arrows). (C) Cryo-EM reconstruction of the m 6 -dAMP bound PUA-Cal filament. Top, side view showing stacked PUA-Cal hexamers forming an extended fibre with an axial repeat of ∼181 Å. Bottom, top view illustrating PUA-like domains on the outer surface and the Cal domains forming the inner core. (D) Structure of a single m 6 -dAMP bound PUA-Cal hexamer extracted from the filament, shown in two orientations. m 6 -dAMP is bound in each PUA-like domain (six ligands per hexamer), while NMN occupies the Cal active sites. (E) Superposition of m 6 -dAMP and m 6 -AMP bound PUA-Cal hexamers, highlighting ligand-dependent conformational differences. Close up views of Cal and PUA-like domains are shown in (F) and (G), respectively. (F) Close-up view of the Cal dimer comparing m 6 -dAMP and m 6 -AMP bound states. In the m 6 -dAMP bound state, the two Cal subunits within each dimer move closer together (inward displacements of helices α1, α2, and α4 of up to ∼5.2 Å). The loop spanning residues 295-316 shifts toward the dimer interface, reducing the distance between opposing loops from ∼27 Å to ∼13 Å. This arrangement is accompanied by a reorientation of bound NMN driven by movement of loop 426-442, while coordination of the catalytic metal centre and the overall active-site geometry remain conserved. (G) Superposition of PUA-like domains from m 6 -dAMP and m 6 -AMP bound hexamers. In the m 6 -dAMP-bound state, PUA-like domains undergo pronounced rotations/translation shifts, bringing the two PUA protomers within each dimer into closer apposition and accommodating paired ligand binding. (H) m 6 -dAMP-dependent conformational rearrangements at the PUA-Cal interface. Superpositions highlight ligand-dependent interdomain changes: (1) ligand-induced rearrangement of W38 in the PUA nucleotide-binding pocket; (2) m 6 -dAMP specific remodelling of the PUA N-terminal loop (residues 10-16), including reorientation of K14; (3) Ligand-dependent remodelling of the PUA-Cal junction loop (residues 173-193), consistent with an interdomain relay that stabilises an assembly-competent configuration. (I) Structural basis of m 6 -dAMP recognition by the PUA-like domain. Top: overall structure of a PUA-Cal dimer within the filament, highlighting m 6 -dAMP bound within each PUA-like domain. Middle: close-up view showing two m 6 -dAMP molecules bound symmetrically within the PUA dimer and coordinated by a Mg 2+ ion, with phosphate groups stabilised by basic residues (including R10 and K40). Bottom: detailed view of the PUA-like binding pocket, in which the adenine bases are stabilised by π–π stacking with conserved aromatic residues (F9, H23, W35, W38, Y100, and Y101). (J) Overall architecture of staked PUA-Cal hexamers forming a continuous filament, shown in two orientations. Insets highlight filament-stabilising interfaces: (1) Cal-Cal contacts between adjacent hexamers formed by hydrogen-bonding and salt bridge networks; (2) PUA-PUA interactions between neighbouring hexamers, showing polar contacts between PUA protomers that are engaged in the filament state but not observed in basal hexameric assemblies.

Article Snippet: The total run time was 8 min with a gradient between 1 to 5 min. m 6 -dAMP was prepared by digesting 10 mM m 6 -dATP with Apyrase (New England Biolabs) for 1h at 37°C, following manufacturer instructions, followed by filtration with an Amicon Ultra-0.5 Centrifugal Filter Unit 3LJkDa.

Techniques: Purification, Cryo-EM Sample Prep, Ligand Binding Assay, Binding Assay

Journal: bioRxiv

Article Title: A methylome-derived m 6 -dAMP trigger assembles a PUA-Cal-HAD immune filament that depletes dNTPs to abort phage infection

doi: 10.64898/2026.01.15.699771

Figure Lengend Snippet: DNA mimics inhibit the anti-phage activity of PUA-Cal-HAD. (A) Effect of heterologous expression of DNA mimics on PUA-Cal-HAD (P-C-H) activity against phage T5. Bars represent the average of triplicates, with individual data points overlaid. YFP, yellow fluorescent protein (control); Asterisks indicate statistically significant differences to P-C-H (p < 0.05). (B) Effect of expressing Gam under Anderson promoter at variable expression levels (0.1, 1) on PUA-Cal-HAD activity against phage T5. Bars represent the average of triplicates, with individual data points overlaid. Asterisks indicate statistically significant differences to P-C-H (p < 0.05). (C) Alignment of T4-Arn and T2-Arn protein sequences, highlighting conserved regions. The Genbank accession numbers of the proteins are shown between brackets. (D) AlphaFold 3 (AF3) model of the PUA-Cal-T4-Arn. Left, model showing T4-Arn bound to the PUA-like domain. Right, close-up view of the T4-Arn-PUA interface, highlighting contacts between a β-sheet region of T4-Arn and residues on the PUA surface. (E) Effect of T4-Arn, T2-Arn, and T4-Arn with mutation N70S on PUA-Cal-HAD activity against phage T5. Bars represent the average of triplicates, with individual data points overlaid. Asterisks indicate statistically significant differences to P-C-H (p < 0.05). (F) Co-immunoprecipitation assays showing interaction between PUA-Cal and T4-Arn. His-tagged PUA-Cal efficiently co-precipitates with Flag-tagged T4-Arn, but not T2-Arn. (G) Structural superposition of the PUA-Cal-T4-Arn AF3 model with the cryo-EM PUA-Cal dimer. RMSD, root mean square deviation. (H) Superposition of the PUA-T4-Arn AF3 model onto the m 6 -dAMP-bound PUA-Cal filament obtained by cryo-EM.

Article Snippet: The total run time was 8 min with a gradient between 1 to 5 min. m 6 -dAMP was prepared by digesting 10 mM m 6 -dATP with Apyrase (New England Biolabs) for 1h at 37°C, following manufacturer instructions, followed by filtration with an Amicon Ultra-0.5 Centrifugal Filter Unit 3LJkDa.

Techniques: Activity Assay, Expressing, Control, Mutagenesis, Immunoprecipitation, Cryo-EM Sample Prep