Anti-viral defence by an mRNA ADP-ribosyltransferase that blocks translation

Statistics and reproducibility

Unless otherwise noted, representative images depict one of three biological replicates.

Bacterial and phage growth and culture conditions

E. coli was routinely grown in LB medium at 37 °C unless otherwise stated. Phages were propagated and handled as described previously⁴.

Plasmid and strain construction

Primers, strains and plasmids are listed in Supplementary Tables 2, 3 and 4, respectively. In all cases, when plasmids were used as PCR templates, PCR samples were treated with DpnI at 37 °C for 1 h to eliminate the plasmid template before transformation. Finished DNA constructs were transformed into DH5α and verified with Sanger and/or long-read sequencing (Primordium) before transformation into the wild-type MG1655 background. For phage assays, the cmdTAC operon was present in low-copy pCD1 (Chloramphenicol-resistant (Cm^r), pSC101 origin of replication) and expressed from its native promoter (P_native). To construct the variant pCV49 (cmdT*AC) by site-directed mutagenesis, complementary primers containing the Y to A mutation (CV109 and CV110) were used to amplify pCD2 for 15 cycles with KAPA DNA polymerase. pCV39 was constructed using the same primer set with pCD4 as the plasmid template.

Low-copy, P_native–cmdTA_NT-3×HAC (pCV43) was engineered using primers CV127 and CV128 to amplify pCD2 such that amplicon ends were located downstream of the CmdA start codon. A synthetic DNA fragment with three YPYDVPDYA codons plus GGGSGGG linker codons (3×HA tag, CV115) with ends complementary to the PCR-amplified vector was then ligated to this vector amplicon by Gibson assembly. P_native–cmdTAC_NT-Flag (pCV44) (NT Flag indicates an N-terminal Flag tag) was constructed by amplification of pCD2 with outward facing and 5′-phosphorylated primers CV120 and CV122 which included DYKDDDDK codons followed by intramolecular blunt-end ligation with T4 DNA ligase. To construct P_native–cmdT_CT-FlagAC, primers CD41 and CD42 were used to amplify pCD2 and primers CD43 and CD44 were used to amplify C-terminally Flag-tagged (CT-Flag) CmdT sequence from pCD10 (see below for pCD10 construction) which was then ligated using Gibson assembly. To construct P_native–cmdT_CT-FlagA_NT-3×HAC_NT-His6 (pCV42), first, a His₆ epitope tag encoding fragment was added to cmdC on pCD2 by the intramolecular blunt-end ligation strategy using primers CV120 and CV122. Next, primers CD41 and CD42 were used to amplify this plasmid and primers CD43 and CD44 were used to amplify CT-Flag cmdT from pCD10 which was then assembled using Gibson assembly. Finally, to insert three tandem HA tags onto the N terminus of CmdA, this intermediate construct was amplified by PCR with CV123 and CV124 such that amplicon ends were located downstream of the CmdA ATG start codon. The synthetic DNA fragment CV115 (HA tag) was used as a PCR template with primers CV125 and CV126 which was then ligated to the vector amplicon by Gibson assembly to produce pCV42.

To construct pCV45 used for the deletion of alt.-3 from T4, complimentary oligonucleotides CV118 and CV119 with pCAS9 compatible overhanging sites were annealed by slow cooling from 98 °C in the presence of 50 mM NaCl to form a duplex spacer insert. pCAS9 and the annealed oligonucleotide were incubated with T4 DNA ligase and BsaI-v2 (NEB) in a one-pot reaction.

For pBAD30 constructs (kanamycin-resistant (Km^r), medium-copy p15a origin), primers CD5 and CD6 were used to amplify and linearize pBAD30. Insert fragments were amplified with the relevant primers (CD7-10, CD13-15, CD20-21, CD30-33, CD38-40) using T4 genomic DNA, plasmid DNA, or MG1655 genomic DNA as a template. pBAD-cmdT_NT-HISA (pCD19) was created using PCR site-directed mutagenesis of pCD4 using primers CD45 and CD46. pBAD-cmdT_NT-HIS (∆cmdA, pCD9) was created by using primers CV1 and CV2 (which exclude the open reading frame of cmdA) to amplify pCD19. This PCR amplicon was intramolecularly ligated with T4 DNA ligase. In some experiments, pAJM677 (Km^r, p15A origin), a variant of pBAD, was used to express CmdTA due to its higher expression after induction with arabinose and tighter repression (pCV41). To engineer pCV41, pAJ677 was amplified and linearized with primers CV113 and CV114. The insert containing cmdTA was amplified from pCD4 with primers CV116 and CV117. Plasmid and insert fragments were ligated by Gibson Assembly. To add the 3×HA tag to cmdA in this context, pCV41 was used as a PCR template with primers CV127 and CV128, and this amplicon was ligated to CV115 fragment by Gibson assembly. Anhydrotetracycline inducible (P_tet) pIF (carbenicillin-resistant (Cb^r), low-copy pSC101 origin) and pKVS45 (Cb^r, p15A origin) constructs were similarly constructed by PCR amplification of the vectors using primers CD24 and CD25 and inserts (CD26-29, CD36-37), followed by Gibson assembly.

To construct the gp23 expression plasmid pCD16, primers CD16 and CD17 or CD18 and CD19 were used to amplify the high-copy origin from pUC19 (pMB1* origin) and pBAD30 without its origin, respectively. These fragments were assembled using Gibson assembly to create a high-copy inducible vector. Subsequently, the backbone was amplified by PCR using primers CD5 and CD6 and gp23 was amplified by PCR from T4 genomic DNA using primers CD34 and CD35. The two fragments were ligated using Gibson assembly.

Plaque and phage assays

Overnight cultures of indicated strains were mixed 1:80 with melted LB with 0.5% agar and then overlaid on plates containing LB with 1.2% agar. For plaque assays done with induction of an arabinose-inducible promoter, base layer plates contained 0.2% w/w arabinose. A tenfold dilution series of the indicated phage was spotted onto plates and the plates grown at 30 °C overnight and plaque-forming units (PFU) were enumerated. log₁₀(protection) (Fig. 1c) was measured as −log₁₀ EOP, where EOP is the ratio PFU_experimental/PFU_control, where the subscript indicates the conditions. Unless otherwise noted, experiments were performed in biological triplicate and representative images are shown.

To measure survival of strains infected with T4, overnight cultures were diluted to OD₆₀₀ 0.1. Cultures were grown to OD₆₀₀ 0.3 and then adjusted to ~3 × 10⁷ colony-forming units (CFUs) in a 1 ml volume in a 1.7 ml Eppendorf tube. Cells were infected with T4 at an MOI of 10 and incubated at 37 °C with rotation. At 0 and 18 min post-infection, cells were pelleted and washed twice with PBS to remove excess phages. One-hundred microlitres of tenfold dilutions were spread onto LB agar plates with chloramphenicol or kanamycin and CFUs were quantified. Survival was measured as CFU ml^−1 at 18 min post-infection divided by CFU ml^−1 at 0 min post-infection. To combat progeny phages in the empty vector strain inhibiting CFU formation, all samples were plated with ~10⁷  chloramphenicol-treated, chloramphenicol-sensitive companion plating cells (MG1655).

ECOI assays were conducted by diluting overnight cultures to OD₆₀₀ 0.1 in 20 ml LB. Cultures were grown until they reached OD₆₀₀ 0.3–0.4 at which point they were infected with T4 at an MOI of 0.1. After 20 min, 1 ml was pelleted and washed twice with PBS. One-hundred microlitres of tenfold dilutions were mixed with 50 µl of indicator strain and 3 ml LB 0.5% agar and overlayed onto LB plates. To control for unadsorbed phages, a ∆ompC strain (OmpC is the receptor for T4) was assayed in parallel. CFUs were enumerated and ECOI was calculated as PFU ml^−1 of the cmdTAC-containing strain divided by PFU ml^−1 of empty vector, after subtracting PFU ml^−1 of the ∆ompC control experiment from each value.

To determine burst size, cell cultures of empty vector and CmdTAC-containing strains were grown in LB + 20 µg ml^−1 chloramphenicol in a water bath at 37 °C until OD₆₀₀ measured 0.5. l-Tryptophan was then added to 20 µg ml^−1 to each culture to assist adsorption of T4. One hundred microlitres of a 10⁷ PFU ml^−1 T4 stock were added to 9.9 ml of each culture and incubated without shaking for 2 min to allow adsorption. Next, for each culture, 100 µl T4-infected culture from this adsorption flask was added to 9.9 ml LB + 20 µg ml^−1 chloramphenicol (flask A). Flask A was again diluted 1:10 into flask B, and again 1:10 into flask C. Five hundred microlitres from flask A was added to 200 µl ice-cold chloroform and vortexed for 10 s. Viable PFUs from this chloroform-treated sample represent unadsorbed phage (adsorption control). Next, 100 µl from each flask A (time 0 sample) or the adsorption control was mixed with 3.5 ml LB 0.5% agar maintained at 50 °C to which was added 50 µl of an overnight culture of indicator strain. This mixture was vortexed briefly and overlayed onto LB + 20 µg ml^−1 chloramphenicol + 1.2% agar plates. All flasks were then left to incubate in a shaking water bath at 30 °C. After 60 min, 100 µl from flask C of the empty vector strain and flask A of the +cmdTAC strain were overlayed with indicator strain on agar plates. After overnight incubation at 37 °C, plaques were enumerated, and normalized to the adsorption control. Burst size was recorded as the number of plaques from each plate multiplied by their dilution factor, and then divided by the number of plaques at time 0.

Growth curves

For measuring growth during T4 infection, overnight cultures of +cmdTAC and empty vector cells were back-diluted 1:200 in 96-well plates and infected with T4 at the indicated MOIs. Cultures were grown at 37 °C with orbital shaking on a plate reader (Biotek) for 6 h. For ectopic expression of Gp23 and Gp31 with CmdTAC, overnight cultures were back-diluted to OD₆₀₀ of 0.05 in M9L + 0.2% w/w glucose + 100 ng ml^−1 anhydrotetracycline (aTc) and grown for 3 h at 37 °C to pre-induce Gp31. Cultures were then pelleted and resuspended at an OD₆₀₀ of 0.05 in fresh M9L + 0.2% w/w glucose + 100 ng ml^−1 aTc or M9L + 0.2% w/w glucose + 100 ng ml^−1 aTc. Cultures were grown at 37 °C with orbital shaking on a plate reader for 12 h.

RNA extraction following phage infection

Overnight cultures of +cmdTAC and empty vector cells were back-diluted and grown at 37 °C to OD₆₀₀ between 0.2 and 0.3 before being infected with T4 at a MOI of 10. RNA was extracted from cells at multiple timepoints post-infection as previously described⁴⁴. In brief, 1 ml of cells was mixed with 1 ml of boiling lysis buffer (SDS 2%, 4 mM EDTA pH 8) and incubated at 100 °C for 5 min before flash freezing in liquid nitrogen. Two millilitres of acid-buffered phenol solution (pH 4.3, Sigma) heated to 67 °C was added to thawed samples, vortexed, and then incubated at 67 °C for 2 min. Samples were spun down at 20,000g for 10 min and hot phenol extraction repeated on the collected aqueous layer. A third extraction was then done using 2 ml of acid-buffered phenol-chloroform solution (Ambion). RNA from the final extraction was then precipitated at −20 °C for at least 1 h or at −80 °C overnight with 1× volume isopropanol, 1/10× volume 3 M sodium acetate (pH 5.5, Thermo Fisher), and 1/100× volume GlycoBlue. RNA was pelleted by centrifugation at 4 °C and 20,000g for 30 min. Pellets were washed twice with 800 ml of ice-cold 70% ethanol, air-dried, and resuspended in 90 μl RNAse-free H₂O (Thermo Fisher).

To remove DNA, 10 μl of 10× Turbo DNase buffer (Ambion) and 2 μl of Turbo DNase I (Ambion) was added to each sample and incubated at 37 °C for 20 min. An additional 2 μl of Turbo DNase I was then added, and samples again incubated at 37 °C for 20 min. RNA was extracted from this digest by precipitation with 3× volume ethanol, 1/10× volume 3 M sodium acetate (pH 5.5), and 1/100× volume GlycoBlue. Pelleting and washing were performed the same as described above. RNA yield was verified using a NanoDrop spectrophotometer.

RNA extraction from non-infected cells

Cells were grown until desired conditions and then 900 μl of culture was mixed with 100 μl of stop solution (5% acid phenol, 95% ethanol) and inverted to mix. Samples were then spun down at 13,000g for 30 s, the supernatant removed, and pellets flash frozen in liquid nitrogen. To each pellet, 400 μl of TRIzol Reagent (Invitrogen) heated to 65 °C was added and mixed using a thermomixer for 10 min at 65 °C and 2,000 rpm before freezing at −80 °C for at least 10 min. Samples were thawed and then centrifuged at 20,000g for 5 min at 4 °C to pellet any debris and the TRIzol solution moved to a new tube. RNA was purified using the Direct-zol RNA Miniprep kit (Zymo Research) following manufacture’s protocol including optional on-column DNAse treatment. RNA yield was verified using a NanoDrop spectrophotometer.

Immuno-northern blotting

Novex 6% TBE-urea gels in 1× TBE buffer (Invitrogen) were pre-run at 180 V for at least 50 min prior to sample loading. Each RNA sample was mixed with equal volume of Novex 2× TBE-urea sample buffer (Invitrogen), heated at 90 °C for 10 min, and then placed on ice for 2–3 min just before loading. Gels were run at 180 V for 30–50 min depending on expected product length. Gels were removed from casing and incubated in 40 ml 1× TBE with added 4 μl of SYBR Gold stain (Thermo Fisher) for 10 min. Gels were imaged on a ChemiDoc MP imaging system (Bio-Rad) set for SYBR Gold imaging. RNA was transferred from the gel to a Hybond-N⁺ nylon membrane (Cytiva) via semi-dry transfer at 0.38 A for 90 min. After transfer, RNA was bound to the membrane by exposure to 120,000 μJ of UV radiation in a Stratalinker UV Crosslinker. Membranes were then incubated with shaking in 0.2% iBlock (Invitrogen) in 1× PBST for 10 min at room temperature or overnight at 4 °C. Primary antibody treatment was done with Poly/Mono-ADP Ribose rabbit antibody (Cell Signaling Technologies) diluted 1:1,000 in 0.2% iBlock + 1× PBST either for 2 h at room temperature or overnight at 4 °C with shaking. Following primary antibody treatment, membranes were washed 3 times for 10 min each with 1× PBST. For secondary antibody treatment, membranes were incubated for 1 h with shaking at room temperature with goat anti-rabbit IgG (H + L) secondary antibody, HRP (Invitrogen) diluted 1:1,000 in 0.2% iBlock + 1× PBST. Membranes were then again washed 3 times for 10 min each in 1× PBST. Signal was developed using SuperSignal West Femto maximum sensitivity substrate (Thermo Fisher) and imaged on a ChemiDoc MP imaging system set for chemiluminescence detection. Dot blots were conducted identically except 250 ng DNA or 1 µg RNA were spotted on membranes.

For agarose immuno-northern blots, 0.8 g of agarose was melted in 66.7 ml of H₂O and allowed to cool to 65 °C. 8 ml of 0.2 M (10×) 3-(N-morpholino)propanesulfonic acid (MOPS) buffer, 5.4 ml of formaldehyde and 5 µl of 10 mg ml^−1 ethidium bromide were added to the agarose, and a 14×12 cm gel was cast and allowed to cool. 4 µg RNA were added to 17 µl sample buffer (2 µl 10× MOPS, 4 µl formaldehyde, 10 µl de-ionized formamide, and 1 µl ethidium bromide) and samples were denatured at 80 °C for 10 min then cooled on ice for 5 min. Prior to sample loading, the empty gel was run at 115 V for 5 min. 2 µl of loading dye (50% glycerol, bromophenol blue and xylene cyanol) were added to each RNA sample. Samples were then electrophoresed at 100 V for 80 min in 1× MOPS buffer. The gel was visualized before soaking in H₂O for 10 min followed by a 20-minute equilibration in transfer buffer (3 M NaCl, 0.01 N NaOH). RNA was transferred onto Hybond-N+ nylon membrane by upward capillary transfer at room temperature for 75 min in transfer buffer. Immunoblotting was performed as described above. All immunoblotting experiments were performed in at least biological duplicate.

RNA immunoprecipitation and sequencing

Cells were collected and RNA collected as described above for infected cells. rRNA was removed using a previously described ribosomal RNA subtraction method⁴⁶. rRNA-depleted RNA was then fragmented using sonication. For each sample to be sonicated, 4 μg of RNA was added to 100 μl 1× TE buffer (Sigma) in a 1.5 ml TPX microtube (Diagenode) and incubated on ice for 15 min. Tubes were then placed in a Bioruptor 300 sonicator water bath chilled to 4 °C for 60 cycles of 30 s on, 30 s off at high power setting. Every ten cycles tubes were briefly spun down in a microcentrifuge to ensure all liquid stayed below the water line in the sonicator. Each sample was then brought to a total volume of 200 μl with RNAse-free water and then precipitated at −20 °C for at least 1 h or at −80 °C overnight with 600 μl 100% ethanol, 20 μl of 3 M sodium acetate (pH 5.5), and 2 μl GlycoBlue. RNA was pelleted by centrifugation at 4 °C and 21,000g for 30 min. Pellets were washed twice with 800 ml of ice-cold 70% ethanol, air-dried, and resuspended in 90 μl RNAse-free H₂O.

ADP ribose RNA immunoprecipitation was based on a methylated RNA immunoprecipitation sequencing (MeRIP-seq) protocol for low-input samples⁴⁷. One-hundred microlitres of Dynabeads Protein G beads were washed 3 times in IP buffer (150 mM NaCl, 10 mM pH 7.5 Tris-HCl, 0.1% NP-40 substitute). Ten microlitres of Poly/Mono-ADP Ribose rabbit antibody (Cell Signaling Technologies) was added to washed beads resuspended in 500 μl IP buffer and then incubated overnight at 4 °C with end-to-end rotation. Following incubation, antibody conjugated beads were washed twice with IP buffer and then resuspended in 500 μl IP buffer with 20 μg fragmented, rRNA-depleted RNA and 5 μl Superase-In RNAse inhibitor and incubated overnight at 4 °C with end-to-end rotation. Samples were then washed twice with 1 ml IP buffer, twice with 1 ml low-salt wash (50 mM NaCl, 10 mM pH 7.5 Tris-HCl, 0.1% NP-40 substitute), and twice with 1 ml high-salt wash (500 mM NaCl, 10 mM pH 7.5 Tris-HCl, 0.1% NP-40 substitute). For each wash, beads were incubated in the wash solution for 10 min at 4 °C with end-to-end rotation. After the final wash, beads were incubated in 200 μl RLT buffer from the Qiagen RNeasy kit for 2 min at room temperature with end-to-end rotation. Supernatant was separated from the beads using a magnetic rack, transferred to a new tube, and mixed with 200 μl of 100% ethanol. This mixture was passed through a RNeasy MiniElute spin column by centrifugation at 20,000g at 4 °C for 1 min. Spin columns were then washed once with 500 μl RNeasy RPE buffer and once with 500 μl 80% ethanol with each spin done at 20,000g for 1 min at 4 °C. Columns were then spun at 20,000g for 5 min to remove residual ethanol. RNA was eluted from the column in 15 μl RNAse-free H₂O with a spin at 20,000g for 5 min at 4 °C. RNA yield and integrity was verified using a NanoDrop spectrophotometer and a Novex 6% TBE-urea gel (Invitrogen), respectively.

Pre- and post-immunoprecipitation RNA (50–100 ng) was then used to make RNA-seq libraries using the NEBNext Ultra II RNA Library Prep Kit for Illumina following the manufacturer’s protocol for use with rRNA-depleted formalin-fixed, paraffin-embedded RNA. Paired-end sequencing of the libraries was performed on a Singular G4 machine at the MIT BioMicroCenter. FASTQ files were then mapped to the MG1655 genome (NC_00913.2), the T4 genome (NC_000866), and the plasmid pKVS45-CmdTAC as previously described^44,48.

Library preparation for RNA-seq

Cells were collected and RNA collected as described above for infected cells. rRNA was removed using a previously described ribosomal RNA subtraction method⁴⁶. One-hundred nanograms of each rRNA-depleted RNA sample was then used to make RNA-seq libraries using the NEBNext Ultra II RNA Library Prep Kit for Illumina following the manufacturer’s protocol for use with purified mRNA or rRNA-depleted RNA. Paired-end sequencing of the libraries was performed on an Illumina NextSeq 5000 machine at the MIT BioMicroCenter. FASTQ files were then mapped to the MG1655 genome (NC_00913.2), the T4 genome (NC_000866), and the plasmid pKVS45-CmdTAC, as previously described^44,48.

Co-immunoprecipitation and LC–MS/MS

Overnight cultures of +cmdTAC and +cmdTA/Flag-C or +cmdT-Flag/AC cells were back-diluted in 250 ml LB and grown at 37 °C to an OD₆₀₀ of 0.3 and then for +cmdTAC and +cmdTA/Flag-C samples infected with T4 at an MOI of 10. At 0 min for all samples and 15 min post-infection for +cmdTAC and +cmdTA/Flag-C cultures 64 ml of sample was pelleted by centrifugation at 7,500g for 5 min. Pellets were decanted and resuspended in 1 ml of lysis buffer (25 mM Tris-HCL, 150 mM NaCl, 1 mM EDTA, 5% glycerol, 1% Triton X-100) supplemented with 1 μl ml^−1 Ready-Lyse Lysozyme (Fischer Scientific), 1 μl ml^−1 benzonase (Sigma), and cOmplete Protease Inhibitor Cocktail (Roche) and then flash frozen in liquid nitrogen. Samples were kept in liquid nitrogen until all timepoints were collected. Samples were subjected to two freeze-thaw cycles in liquid nitrogen to ensure complete lysis of cells. Additional lysis buffer was added to samples as needed to normalize sample concentrations by OD₆₀₀. Samples were spun at 20,000g for 10 min at 4 °C to pellet any debris. For each sample, 50 μl of Pierce Anti-DYKDDDDK magnetic agarose beads was mixed with 450 μl lysis buffer and then collected to the side of the tube using a magnetic rack. Beads were then washed twice with 500 μl lysis buffer. After the final wash, beads were mixed with 1 ml of sample and incubated for 20 min at room temperature on an end-to-end rotor. After incubation, beads were washed in wash buffer (1× PBS, 150 mM NaCl) twice and then once with MilliQ H₂O.

On-bead reduction, trypsin digest, and LC–MS/MS were done by the MIT Biopolymers and Proteomics Core as previously described⁴². In brief, proteins were reduced for 1 h at 56 °C with 10 mM dithiothreitol (Sigma) and then alkylated for 1 h at 25 °C in the dark with 20 mM iodoacetamide (Sigma). Proteins were digested with modified trypsin (Promega) overnight in 100 mM, pH 8 ammonium bicarbonate at a 1:50 enzyme:substrate ratio. Formic acid (99.9%, Sigma) was added to stop trypsin digest. Digested peptides were desalted using Pierce Peptide Desalting Spin Columns (Thermo) then lyophilized. Peptides were separated on a PepMap RSLC C18 column (Thermo) over 90 min by reverse phase HPLC (Thermo Ultimate 3000) before nano-electrospray with an Orbitrap Exploris 480 mass spectrometer (Thermo). Mass spectrometer run was done in data-dependent mode. Full scan parameters were resolution of 120,000 across 375–1600 m/z and maximum IT 25 ms. This was followed by MS/MS for as many precursor ions in a two second cycle with a resolution of 30,000, dynamic exclusion of 20 s, and a NCE of 28. Detected peptides were mapped to MG1655, plasmid, and T4 protein sequences and the abundance of proteins were estimated by number of spectrum counts/molecular mass to normalize for protein sizes. The ratio of spectral counts between the Flag pulldowns and untagged pulldowns at each timepoint were used to generate the data in the figures with a pseudocount added to each count.

In vitro transcription and translation

In vitro transcription–translation assays were conducted using the PURExpress kit (NEB) according to the manufacturer’s protocol with a 2-h incubation at 37 °C. Each reaction was supplemented with 1 U µl^−1 Riboguard RNase inhibitor (LGC Biosearch Technologies), with or without 1 mM NAD⁺ and protein eluants as indicated. When supplying mRNA as a translation template, primers were used to amplify the DHFR gene using PCR from the PURExpress control DHFR plasmid. The PCR amplicon was purified using the DNA Clean & Concentrator Kit (Zymogen). Then, mRNA was synthesized from the PCR template by incubating 300 ng DNA with 200 U T7 RNA polymerase, 0.5 mM NTPs, and 5 mM DTT in a final reaction volume of 40 µl at 37 °C for 4.5 h. The resulting RNA was purified from the reaction using the RNA Clean and Concentrator Kit (Zymogen) with on-column DNase I treatment. Pure mRNA was then treated with CmdT or control mock purified protein in 1× ADPr buffer (20 mM Tris-HCl pH 8.0 and 150 mM NaCl) with 1 mM NAD⁺ and 1 U µl^−1 Riboguard at 37 °C for 2 h. RNA was again purified as before, and 1 µg was supplied in the PURExpress reaction for 4 h at 37 °C. From this reaction, 2.5 µl was then denatured in Laemmli buffer and run on a 8–16% polyacrylamide gel by SDS–PAGE and stained with either Brilliant Blue R250 or Coomassie Fluor Orange (Molecular Probes) and visualized on a Bio-Rad ChemiDoc MP imager.

Protein immunoblotting

Cell cultures were grown overnight and diluted 1:200 in fresh LB containing the appropriate antibiotics. Cultures were grown at 37 °C to mid-exponential phase and then treated with T4 at an MOI of 10, or the appropriate inducers as dictated by the experiment. At various timepoints, cells were pelleted, flash frozen and subsequently resuspended in Laemmli buffer with 2.5% 2-mercaptoethanol in a volume normalized to culture turbidity (100 µl OD600^−1 ml^−1). Samples were run by standard SDS–PAGE on 12% polyacrylamide gels. Transfer onto 0.2 µm PVDF membranes was done at 90 V for 40 min for CmdA–HA, and otherwise was done at 100 V for 60 min. Membranes were blocked in Tris-buffered saline with 0.05% Tween-20 (TBST) and 5% non-fat milk for 60 min at room temperature and incubated with primary monoclonal antibody (1:1,000 rabbit anti-Flag or anti-HA, Cell Signaling Technologies) overnight at 4 °C. Membranes were washed with TBST and incubated with HRP-conjugated goat anti-rabbit IgG (Invitrogen) in blocking buffer for 60 min at room temperature. Membranes were again washed and incubated with SuperSignal West Femto Maximum Sensitivity Substrate (Thermo Scientific) before exposure on a Bio-Rad ChemiDoc MP imager. Membranes were stained with Brilliant blue R250 as a loading control. To quantify band intensity, we used the gel analysis tool in ImageJ. Pixel intensity from the antibody signal was normalized to pixel intensity of total protein stain.

For immunoblots, membrane chemiluminescence was imaged directly followed by imaging of pre-stained molecular weight markers. Images were aligned, as shown in Supplementary Fig. 1, to relate chemiluminescent bands to molecular weight markers, as shown in main figures.

Non-denaturing blots were performed by lysing cells with a buffer composed of 50% BPER-II (Thermo Scientific), 0.1 mg ml^−1 lysozyme, cOmplete protease inhibitor (Sigma-Aldrich), 6 U DNase I (NEB) and 3 µl RNase A (NEB) in volumes normalized to OD₆₀₀ of culture samples. Samples were incubated until clear at room temperature and then spun at max speed in a table-top centrifuge for 5 min to pellet insoluble material. Native loading dye (6×; 600 mM Tris-HCl, 50% glycerol, 0.02% bromophenol blue) was added to samples and loaded onto a 12% polyacrylamide Mini-Protean TGX pre-cast gel (Bio-Rad, does not contain SDS). Samples were electrophoresed in 25 mM Tris, 192 mM glycine running buffer at 75 V for 90 min. Transfer was conducted onto 0.2 µm PVDF membranes as described in this section at 100 V for 1 h at 4 °C. Blots were processed as described above. Blots shown are representative images of at least two biological replicates.

T4 genome engineering and evolution

The evolution of T4 on cmdTAC-containing cells was performed as described previously⁴⁹ for four rounds, resulting in the alt.-3 C-terminal extensions in all five replicates. To generate alt.-3 mutants for further evolution experiments, T4 stock was overlayed onto strains containing Cas9 and spacers directed toward alt.-3 or a control plasmid with no spacer (ML4233 and ML4234). The number of plaques formed on the spacer-containing strain was compared to the control to determine whether there was any selection imposed by the spacer. Despite attempting eight potential spacers, no selection was observed. To mitigate this, we repeated the experiment with T4 ∆agt ∆bgt (DNA contains non-glucosylated, 5-hydroxymethyl cytosine) on an E. coli ∆mcrA ∆mcrBC background required for viability of T4 ∆agt ∆bgt. This T4 formed fewer plaques in the presence of the alt.-3 spacer, suggesting that selection for alt.-3 mutants was imposed in this condition. The alt.-3 region was amplified by PCR and Sanger sequenced from plaques that were able to form on the spacer-containing strain. Of those plaques, we isolated a strain that harboured a mutation encompassing nearly the entire open reading frame of alt.-3. The T4 ∆alt.-3 strain was propagated in the presence of the spacer and stored as a stock at 4 °C. Evolution of this T4 strain on cmdTAC-containing cells was conducted the same as before, for 17 rounds, without observing mutations that increased plaquing ability.

Radiolabel incorporation assays

Overnight cultures of +cmdTAC and +cmdT*AC cells were back-diluted in LB + 20 µg ml^−1 chloramphenicol and grown at 37 °C to an OD₆₀₀ between 0.2 and 0.3. An aliquot of each culture was collected before T4 infection at an MOI of 10 and again at each indicated timepoint (t = 10, 20, 30, 40 min post-infection). Each collected sample was incubated with EasyTag EXPRESS-³⁵S protein labelling mix (Perkin Elmer) at 23 µCi ml^−1 for 2 min at 37 °C. Labelling was chased with an unlabelled cysteine/methionine mixture at 5 mM and then samples precipitated in 13% w/v ice-cold TCA. Samples were pelleted by centrifugation at 16,000g for 10 min at 4 °C, washed twice with 500 µl ice-cold acetone, and then resuspended in resuspension buffer (100 mM Tris pH 11.0, 3% w/v SDS). Samples were run on 4–20% SDS–PAGE gels (Bio-Rad), the gels incubated in Gel-Dry Drying Solution (Invitrogen) for 10 min, and then dried on a vacuum gel dryer for 2 h at 80 °C. Dried gels were exposed to a phosphorimaging screen overnight before imaging on an Amersham Typhoon imager.

Protein purification

Five millilitres of cultures of ML4207 and ML4232 were grown overnight at 37 °C in LB + 0.2% glucose. The following day, 5 ml of each culture was washed of glucose twice and used to inoculate 495 ml of LB + 25 µg ml^−1 kanamycin. After 1 h of additional growth, arabinose was added to a final concentration of 0.2%. Cultures were grown an additional 95 min, pelleted, washed with H₂O, again pelleted, and flash frozen in liquid N₂. The following day, cell pellets were resuspended in 4 ml lysis buffer (50 mM Tris pH 7.5, 500 mM NaCl, 0.05% Tween-20, EDTA-free protease inhibitor, 0.5 mM PMSF, 0.5 mg ml^−1 lysozyme, 5 mM imidazole, and 5% glycerol) on ice. Cells were then lysed by sonication in a Bioruptor 300 for two rounds of 10 cycles each, high setting, 30 s on/30 s off. One millilitre of Ni-NTA agarose resin (Qiagen, 0.5 ml bed volume) was equilibrated in lysis buffer. Cell lysate was clarified by centrifugation then incubated with the Ni resin for 1 h at 4 °C with gentle rocking. The following steps were conducted at 4 °C. The resin was then passed through a 10 ml chromatography column and then washed 5× with 2.5 ml of wash buffer (same as lysis buffer but without lysozyme, and imidazole at 20 mM). Protein was then eluted 5× with 2.5 ml elution buffer (wash buffer with imidazole concentration at 300 mM). Eluted proteins were buffer-exchanged into Tris pH 7.4 using Micro Bio-Spin chromatography columns (Bio-Rad) and concentrated using Amicon Ultra 0.5 ml centrifugal filters with a 3-kDa cutoff.

In vitro ADP-ribosylation by CmdT

A typical reaction was assembled on ice as follows. To a buffer composed of 20 mM Tris-HCl pH 8.0 and 150 mM NaCl, we added 1 U µl^−1 Riboguard RNase inhibitor, 1 mM NAD⁺ (NEB), 4 µg of DNA or RNA oligonucleotide, and protein at the concentrations indicated. The reactions were then incubated in a thermocycler at 37 °C. To stop the reaction, an equal volume of 2× 6 M urea sample buffer (Novex) was added. RNA was denatured at 95 °C for 10 min and then immediately placed on ice. One microgram of RNA samples were then subject to electrophoresis in 15% polyacrylamide TBE-urea gels at 180 V for 75 min. Gels were stained both with SYBR Gold and with a concentrated solution of 0.2% methylene blue in 0.1× TBE buffer for 15 min, de-stained with several changes of H₂O and imaged.

In vitro ADP ribose RNA pulldown

Twenty micrograms of total RNA were ADP-ribosylated with CmdT as described above with 0.25 mM 6-Biotin-17-NAD⁺ (Cayman Chemical) for 4 h at 37 °C and then continued at 4 °C overnight. Two control reactions were set up identically except with mock purified protein, or with standard NAD⁺ in place of 6-Biotin-17-NAD⁺. Ten micrograms of each reaction were kept at −80 °C as the pre-pulldown sample. The remaining 10 µg were incubated with streptavidin conjugated superparamagnetic beads (Dynabeads MyOne Streptavidin C1) following the manufacturer’s protocol. RNA was stripped from the beads by addition of 0.5 ml Trizol and incubation at 25 °C for 15 min on a thermomixer at 1,000 rpm. Beads were then precipitated with a magnet and 100 µl of chloroform were added. The phases were separated by centrifugation at 14,000g for 15 min. Finally, the aqueous phase was purified using the RNA Clean and Concentrate Kit (Zymogen). Pre- and post-pulldown samples were electrophoresed on a 6% TBE-urea gel for 45 min, stained with SYBR Gold and imaged. The samples from pre- and post-pulldown reactions containing 6-Biotin-17-NAD⁺ and CmdT were subject to RNA-seq as described in this section, but without rRNA depletion.

HPLC analysis of ribonucleosides

Ten micrograms each of no-U and no-C RNA oligonucleotides (Fig. 4d) were subjected to ADP-ribosylation as described above. Controls were included in which purified CmdT was replaced by a mock purification, or in which NAD⁺ was omitted. Next, samples were split and treated with either 100 U Nuclease P1 (NEB) and 10 U antarctic phosphatase, or, the same with the addition of 1 U Phosphodiesterase I from Crotalus adamanteus venom (SVPD, Millipore Sigma). Reactions were incubated in digest buffer (25 mM Tris-HCl pH 7.6, 50 mM NaCl₂, 1 mM ZnCl₂, and 10 mM MgCl₂) at 37 °C for 3.5 h in a total volume of 110 µl. One-hundred microlitres of digested and dephosphorylated nucleosides (10 µg) were injected onto a Vydac C18 4.6 ×250 mm reverse phase silica column (218TP54) equilibrated with 90% buffer A (0.1 M triethylammonium acetate (TEAA), pH 7.0)/10% buffer B (0.1 M TEAA, 20% acetonitrile, pH 7.0). HPLC was run with a mobile phase gradient composed of buffer A and B, 10–60% B from 1–21 min and 60–97% B from 21–26 min at a flow rate of 1 ml min^−1. Analytes were measured at A₂₅₄. On a replicate run, samples without SVPD treatment were collected as fractions and relevant fractions were lyophilized. The samples were then resuspended in digest buffer, and again incubated for 3.5 h with 10 U antarctic phosphatase and 1 U SVPD and analysed by HPLC as described above.

Mass spectrometry of modified nucleosides

Fractions collected from HPLC analysis were dried in a speed-vac and resuspended in 200 µl of 50% acetonitrile in 0.1% formic acid. The fractions, or a buffer blank, were directly infused by syringe pump into a Thermo Q Exactive with an API source and electrospray ionization probe at a flow rate of 5 µl min^−1. The instrument was operated in positive ion mode. MS/MS was conducted at collision energies of both 25 and 40 CE. Instrument parameters were as follows: spray voltage, 3.8 kV; capillary tube temperature, 280 °C; sheath gas, 20; auxiliary gas, 5; sweep gas, 5.

Bioinformatic analyses

The CmdT sequence logo was generated from the CmdT hmm file from DefenseFinder⁵⁰ using skylign.org.

Hidden Markov model profiles of CmdT and CmdC were downloaded from DefenseFinder⁵⁰ and searched against the RefSeq non-redundant protein database using hmmscan and default parameters. Protein hits were then identified in all available RefSeq bacterial genomes and CmdTAC was called if both CmdT and CmdC were present within two proteins of each other in the genome. CmdA was not included in calling as it both has higher sequence variability and is often unannotated in clearly homologous systems. All datasets were downloaded in July 2023. The complete taxonomic lineage of RefSeq genomes was created and filtered to include bacteria of current interest (genera with greater than 1,000 sequenced genomes). A taxonomic relationship of these genera was produced with NCBI Common Tree, and presence/absence was recorded from the taxonomic profiles of the CmdT/C hmmscan.

Structural predictions

Protein homology was assessed using HHpred⁵¹. Predictions of the structures of individual components and CmdTAC as a complex were done using AlphaFold2⁵² with the multimer module and default parameters on the reduced database with 1 prediction generated per model. Structural homology searches based on the AlphaFold2-predicted structures of CmdT and CmdC were done using Foldseek⁵³ and the top hit for each search used for subsequent analyses. Electrostatics modelling was done using the coulombic function in ChimeraX. All predicted structure visualization was done using ChimeraX.

RNA-seq and RIP-seq read mapping

FASTQ files for each sample were trimmed using cutadapt⁵⁴ (version 1.15) and then mapped to the E. coli MG1655 genome (NC_00913.2), the T4 genome (NC_000866), and the plasmid pKVS45-CmdTAC using bowtie2⁵⁵ (version 2.3.4.1) with the following arguments: -D 20, -I 40, -X 300, -R 3, -N 0, -L 20, -i S,1,0.50. Sam files generated from bowtie2 mapping were then converted to bam files using samtools⁵⁶ (version 1.7) and then further converted to numpy arrays using the genomearray3 python library⁵⁷ for use in downstream analyses. For in vivo RIP–seq analyses only highly expressed transcripts as determined by transcripts with an RNA TPM for both replicates greater than or equal to the minimum mean TPM of any T4 transcript were used. For statistical comparisons of TPM ratios between RNA types a Welch’s t-test was used.

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.