Bacterial strains and growth conditions
Escherichia coli strains that were used in this study are listed in Supplementary Table 4. All bacterial cultures were grown in 3.5 ml of media in 14-ml culture tubes shaking at 220 rpm at 37 °C, unless otherwise indicated. ‘Overnight’ cultures are defined as those that were grown for 16–20 h following inoculation from a single colony or glycerol stock. Where applicable, culture media was supplemented with carbenicillin (100 µg ml−1) and/or chloramphenicol (20 µg ml−1) for plasmid maintenance or strain selection. Escherichia coli OmniPir49 was used for strain construction and storage of plasmids and E. coli MG1655 (CGSC6300) was used for all phage, colony formation and nucleotide extraction experiments.
All E. coli cultures used for cloning, strain construction, protein expression and indicated colony formation assays were grown in LB medium (1% tryptone, 0.5% yeast extract and 0.5% NaCl). All strains were frozen for long-term storage in LB plus 30% glycerol (v/v) at −70 °C. Strains used to perform phage amplification, phage infection assays, 2′,3′-c-diAMP extraction and measurement, and indicated colony formation assays were cultivated in ‘MMCG’ minimal medium (47.8 mM of Na2HPO4, 22 mM of KH2PO4, 18.7 mM of NH4Cl, 8.6 mM of NaCl, 22.2 mM of glucose, 2 mM of MgSO4, 100 mM of CaCl2, 3 mM of thiamine and trace metals at 0.1× (Trace Metals Mixture T1001, Teknova; final concentration: 8.3 μM of FeCl3, 2.7 μM of CaCl2, 1.4 μM of MnCl2, 1.8 μM of ZnSO4, 370 nM of CoCl2, 250 nM of CuCl2, 350 nM of NiCl2, 240 nM of Na2MoO4, 200 nM of Na2SeO4 and 200 nM of H3BO3)). When a strain with two plasmids was cultivated in MMCG medium, bacteria were grown in carbenicillin (20 μg ml−1) and chloramphenicol (4 μg ml−1). When growing strains that required induction, 500 μM of IPTG or 0.2% arabinose was used to induce, as appropriate. MMCG and LB agar plates contain 1.6% agar and media components described above.
Plasmid construction
The plasmids used in this study are listed in Supplementary Table 4. Cloning and plasmid construction were performed as previously described7. In brief, genes of interest were amplified from phage genomic DNA or previously constructed plasmids using Q5 Hot Start High-Fidelity Master Mix (New England Biolabs (NEB)), or were synthesized as FragmentGENEs (Genewiz). Gene inserts were flanked by at least 18 base pairs of homology to the vector backbone outside the restriction digest sites. Ligation of genes into the digested, linearized backbone vector was done using modified Gibson Assembly50 with HiFi DNA Assembly Master Mix (NEB). Gibson assemblies were transformed by electroporation into competent OmniPir and plated onto LB (1.6% agar) plates with appropriate antibiotics to select for successful transformations. Phage genes and optS with point mutations were generated by amplifying the gene of interest in two parts from a plasmid template, with the desired mutation occurring in the overlapping region between the two amplicons. Unless otherwise indicated, all enzymes were purchased from NEB.
For the optSE operon in a pLOCO3 backbone, complete vectors with the indicated operons were generated as ValueGENEs (Genewiz). The pLOCO3 vector (including superfolder green fluorescent protein (sfGFP)) was initially constructed using Gibson assembly to join and circularize two FragmentGENEs, one with the pLOCO3 backbone and one with the sfGFP gene (Genewiz; gene fragment sequences are listed in Supplementary Table 4).
For all vectors using the pTACxc backbone, pAW1608 was amplified and purified from OmniPir. Purified plasmid was then linearized using BamHI-HF and NotI-HF, or EcoRV-HF and PstI-HF. Gibson ligation was used to circularize the plasmid with the new insert.
For all vectors using the pBAD30 backbone, pAW1640 was amplified and purified from OmniPir. Purified plasmid was then linearized using EcoRI-HF and NotI-HF. Gibson ligation was used to circularize the plasmid with the new insert.
For all vectors using the pET16SUMO2 backbone, pAW1123 was amplified and purified from Sure1. Purified plasmid was then linearized using BamHI-HF and NotI-HF. Gibson ligation was used to circularize the plasmid with the new insert.
Plasmid sequences were verified with Sanger sequencing (Quintara Biosciences or Azenta) and/or Oxford Nanopore Sequencing (Plasmidsaurus). Reads were mapped to the predicted plasmid sequence using the Map to Reference feature of Geneious Prime (default settings).
Phage amplification and storage
The phages used in this study are listed in Supplementary Table 4. Information on the Bacteriophage Selection for your Laboratory (BASEL) collection used in this study is reported in ref. 51. Phage lysates were generated through plate amplification using a modified double agar overlay52. For plate amplification, 400 ml of mid-log MG1655 were mixed with 3.5 ml of MMCG soft agar mix (MMCG with 0.35% agar and 10 mM of MgCl2, 10 mM of CaCl2 and 100 mM of MnCl2) and 100–1,000 plaque-forming units (PFU). Plates were then incubated overnight at 37 °C. Phages were collected by adding 5 ml of SM buffer (100 mM of NaCl, 8 mM of MgSO4, 50 mM of Tris-HCl (pH 7.5) and 0.01% gelatin) to the plate and incubating for 1 h at room temperature. To increase phage titre, the top agar overlay was scraped and collected along with the SM buffer. The SM buffer and top agar mixture was centrifuged at 4,000g for 10 min and the supernatant was transferred to a new tube. The resulting liquid was passed through a 0.2-mm filter or treated with two or three drops of chloroform, followed by vortexing, to remove viable bacteria. All amplified phages were stored at 4 °C in SM buffer.
Phage infection assays
Phage infection assays and phage titre quantifications were performed using a modified double agar overlay technique52. Strains containing the indicated plasmids were cultivated overnight in MMCG or LB medium (including appropriate antibiotics) and were diluted 1:10 in fresh medium the following day. The bacteria were grown until they reached mid-logarithmic phase (OD600 0.1–0.8). A total of 400 µl of mid-log bacteria were mixed with 3.5 ml of 0.35% MMCG agar (plus 5 mM of MgCl2 and 0.1 mM of MnCl2) or LB agar (plus 10 mM of MgCl2, 10 mM of CaCl2 and 0.1 mM of MnCl2) and poured on top of an MMCG (1.6% agar) or LB (1.6% agar) plate, respectively. The plate was allowed to cool for 15 min. Once cooled, 2 μl of a phage tenfold serial dilution series was spotted onto the soft agar overlay and allowed to dry, after which the plates were incubated at 37 °C overnight. Plates were imaged roughly 18–24 h after infection.
The resulting phage titre was quantified in PFU per millilitre for each phage lysate tested. PFU were enumerated on the basis of the lowest phage dilution spot with individual, quantifiable PFU. The dilution at that spot was used to calculate the PFU per millilitre appropriately. When there was a hazy zone of clearance, rather than identifiable plaques, the lowest phage concentration at which this was seen was counted as ten plaques. When no clearance was observed, the least dilute spot was counted as 0.9 plaques, and this was used as the limit of detection for the assay. Phage infection data were reported as PFU per millilitre ± standard error of the mean (s.e.m.) of n = 3 biological replicates.
Phage infection time course in liquid culture
Bacterial strains containing the indicated plasmid were grown overnight in 25-ml MMCG or LB media plus appropriate antibiotics. Cultures were diluted to an OD600 of 0.1 in fresh media without antibiotics (total volume, 30 ml) and were grown for 2 h at 37 °C with shaking at 220 rpm. After 2 h, the cultures were infected with phage at the indicated multiplicity of infection. The OD600 of the cultures was measured at the indicated time points after infection. To enumerate PFU at each time point, 250 µl of culture was collected and centrifuged at 20,000g for 5 min at 4 °C. The supernatant was transferred to a new tube and three to five drops of chloroform were added, followed by vortexing, to kill any bacteria that remained. The resulting lysates were titred using the phage infection assay protocol explained above.
Escaper phage generation and amplification
T4 escaper phages were generated from three unrelated, clonal T4 lysates (‘parents’) that were separately plate amplified on wild-type E. coli MG1655. To make the T4 escaper phages (‘daughters’), 400 µl of mid-log bacteria expressing the optSE operon (in MMCG plus 100 µg ml−1 carbenicillin) was mixed with 100 µl of parent T4 lysate (about 4 × 105 PFU) and 3.5 ml of MMCG top agar and poured onto an MMCG agar plate. The plate was allowed to dry and was incubated overnight at 37 °C. The next day, five single escaper plaques were individually isolated from each parent T4 plate using a Pasteur pipette, soaked in 500 µl of SM buffer in an Eppendorf tube and filter sterilized using a Nanosep 0.2-µm spin filter (Pall Labs). A dilution series of each T4 escaper phage was spot-plated onto E. coli MG1655 expressing Panoptes to confirm replication in the presence of OptSE. From this plate, single plaques from each escaper were individually purified and plate amplified (as described in the phage amplification and storage protocol above) before storage.
Phage genome sequencing and escaper analysis
The genomes of the parent and escaper T4 phages were purified as previously described53. To do this, 450 ml of phage lysate (more than 108 PFU ml−1) was mixed with 50 µl of 10× DNAse I buffer (100 mM of Tris-HCl (pH 7.6), 25 mM of MgCl2 and 5 mM of CaCl2) and treated with DNAse I (final concentration 2 × 10−3 U µl−1) and RNAse A (final concentration 2 × 10−2 mg ml−1). This mixture was incubated for 1.5 h at 37 °C to remove extracellular nucleic acids. After, EDTA was added to a final concentration 20 mM to stop the reaction. Each parent and escaper phage genome was then isolated and purified using the Qiagen DNeasy Cleanup Kit, starting at the proteinase K digestion step53.
The purified phage genomes were prepared for Illumina sequencing using a modification of the Nextera kit protocol as previously described54. Illumina sequencing was performed using a MiSeq V2 Micro 300-cycle kit (SeqCenter). The resulting reads were mapped to Genome accession AF158101.6 using Geneious software’s ‘Map to Reference’ feature. Each read was trimmed to remove the Nextera adapter sequences before mapping (sequence trimmed: AGATGTGTATAAGAGACAG) using the ‘Trim sequences’ option; otherwise, Geneious default settings were used. The trimmed sequences were mapped to the phage genome using default settings with the ‘Map to Reference(s)’ feature. The Geneious feature ‘Find Variations/SNPs’ was used to identify variants in daughter phage genomes. Called variants were identified as escaper mutations if they were present in 75% or more of reads and were not present in parent phage genomes.
Construction of phage gene deletions
T4 knockout phages were generated as previously described55. In brief, 5 ml of E. coli MG1655 expressing a pET vector encoding the template for homologous repair were grown in LB to mid-log phase. Bacteria were then infected with about 4 × 108 PFU of T4 and grown for 2–3 h before collecting lysate. Phage lysates amplified in this way were then mixed with 400 μl of mid-log bacteria expressing eLbuCas13a constructs with spacers targeting the gene of interest and poured onto an MMCG agar plate as in the method described in solid plate amplification, detailed above. Individual plaques were isolated and spot-plated on E. coli MG1655 expressing the same spacer to confirm mutation and to plaque-purify each clone. Target gene deletion was validated using PCR.
Homologous repair templates encoded 250 bp of homology on either side of the target gene. In-frame deletions retained the first and last six amino acids of the target gene, but deleted the intervening sequence. Two 31-nt spacers were selected to target the beginning of each gene and induced as needed using anhydrotetracycline at 5 nM in the top agar of the soft agar overlay.
The acb1 knockout was constructed with repair template pEK0220 and spacers pEK0223 and pEK0224. The knockout was PCR verified using the primers oEK0490 and oEK0491.
The acb2 knockout was constructed with repair template pEK0221 and spacers pEK0225 and pEK0226. The knockout was PCR verified using the primers oEK0492 and oEK0493.
The double knockout was created by doing the ∆acb2 knockout steps on a confirmed ∆acb1 knockout phage.
Colony formation assays for bacterial growth inhibition analysis
Bacterial growth inhibition was tested using colony formation assays. Bacterial strains with indicated plasmids were grown overnight in MMCG media plus appropriate antibiotics. The cultures were then tenfold serially diluted in fresh MMCG media (without antibiotics) and 5 µl of each dilution was spotted onto an MMCG agar plate containing the appropriate antibiotics, as well as IPTG (500 µM; induced condition) as indicated. Data in Fig. 4e were collected using LB media with appropriate antibiotics, with or without glucose (0.2% w/v; uninduced condition), IPTG (500 µM; induced condition) or arabinose (0.2% w/v; induced condition), as indicated. After the spotted bacteria were allowed to dry, plates were incubated at 37 °C for roughly 16–18 h for LB plates or around 24 h for MMCG plates. Growth inhibition was quantified the next day by counting the number of colony-forming units (CFU) of the lowest dilution that had individual colonies. When no individual colonies could be counted, the lowest bacterial concentration at which growth was observed was counted as 10 CFU. In instances where no growth was visible, the least dilute spot was counted as 0.9 CFU and used as the limit of detection. Colony formation data were reported as CFU per millilitre ± s.e.m. of n = 3 biological replicates.
Acb2 protein expression and isothermal titration calorimetry
The vector expressing WT 6xHis-hSUMO2-Acb2 was transformed into Rosetta2 expressing the pRARE2 plasmid and plated onto 1.6% MMCG agar plates, plus 100 μg ml−1 of carbenicillin and 20 μg ml−1 of chloramphenicol. An individual colony was picked the following day and inoculated into 100 ml of M9ZB media (47.8 mM of Na2HPO4, 22 mM of KH2PO4, 18.7 mM of NH4Cl, 85.6 mM of NaCl, 1% Casamino acids (VWR), 0.5% v/v glycerol, 2 mM of MgSO4, trace metals at 0.5× (Trace Metals Mixture) plus 100 μg ml−1 of carbenicillin and 20 μg ml−1 of chloramphenicol. The culture was then grown overnight, shaking at 37 °C and 220 rpm. The following day, the culture was used to inoculate 2 l of the same, fresh media to an OD600 of 0.05, then grown to an OD600 of about 1.5. Cultures were crash-cooled on ice for 30 min before IPTG was added to 500 μM to induce protein expression. The culture was then moved to a 16-°C shaking incubator and allowed to grow overnight.
Cultures were collected by centrifugation for 30 min at 4,600g and 4 °C in an Avanti JXN-26 Floor Centrifuge using the JXN 12.500 rotor (Beckman Coulter). The resulting pellets were resuspended in 40 ml of lysis buffer (20 mM of HEPES (pH 7.5), 400 mM of NaCl, 10% v/v glycerol, 20 mM of imidazole, 0.1 mM of dithiothreitol (DTT)). After resuspension, cells were lysed by sonication at 80% amplitude, with 15-s-on, 45-s-off pulses for a total processing time of 10 min using a Q500 sonicator (Qsonica). Cellular debris was removed from sonicated lysates by centrifugation for 45 min at 4 °C and 16,000g in an Avanti JXN-26 Floor Centrifuge using the JA 25.50 rotor (Beckman Coulter). The soluble lysate was then decanted and protein was purified using immobilized metal affinity chromatography. Briefly, the soluble lysate was run over 2 ml of HisPure cobalt slurry (Fisher Scientific) equilibrated in lysis buffer. The resin was then washed with 2 × 25 ml of wash buffer (20 mM of HEPES (pH 7.5), 1 M of NaCl, 10% v/v glycerol, 20 mM of imidazole, 0.1 mM of DTT) and protein was eluted in 10 ml of elution buffer (20 mM of HEPES (pH 7.5), 400 mM of NaCl, 10% v/v glycerol, 300 mM of imidazole, 0.1 mM of DTT). Proteins were then dialysed against 2 × 1 l of dialysis buffer (20 mM of HEPES (pH 7.5), 250 mM of KCl, 0.1 mM of DTT) overnight at 4 °C using 3.5-kDa molecular weight cut-off (MWCO) SnakeSkin Dialysis Tubing (VWR). The 6×His-SUMO-tag was cleaved using 6×His-hSENP2 (produced in-house; final concentration of 1:100 hSENP2:protein w/w) during the overnight dialysis step. After dialysis, proteins were run over 2 ml of HisPure cobalt slurry equilibrated in dialysis buffer to remove any 6×His-SUMO tagged proteins.
After dialysis, the protein was concentrated as needed using 3-kDa MWCO Nanosep spin concentration columns (Pall Labs) and stored as 200–500-μl aliquots in dialysis buffer at −70 °C. Protein concentrations were measured using A280 on a Nanodrop OneC (Thermo Fisher Scientific) and protein purity was visualized using SDS-PAGE followed by Coomassie staining.
ITC assays were adapted from the protocol described in ref. 13. In brief, the Kd, ΔH and ΔS for the binding of WT Acb2 with 2′,3′-c-di-AMP were determined using a MicroCal ITC200 calorimeter. Purified Acb2 and 2′,3′-c-di-AMP were dialysed into the ITC buffer (20 mM of HEPES (pH 7.5) and 200 mM of NaCl) at 4 °C overnight. The titration was carried out with 27 successive 1.5-μl injections of 150 µM of 2′,3′-c-di-AMP into the sample cell containing 50 µM of WT Acb2. Each injection was spaced by 180 s, and the cell was kept at 25 °C and stirred at 750 rpm. Origin software was used for integration and global curve fitting with a ‘one set of sites’ binding model. We selected the ‘one set of sites’ model, which is appropriate to use for binding between a macromolecule and a ligand that includes any number of binding sites in which all sites have the same association constant (Ka) and ∆H.
Laser scanning confocal microscopy
Strains were grown overnight in LB plus appropriate antibiotics and were diluted 1:10 into fresh medium the following day. Strains were grown until they reached mid-logarithmic phase (OD600 0.1–0.8) and were then normalized to an OD600 of 0.55 before induction with 500 µM of IPTG. At each time point, culture samples were removed from the liquid culture, OD normalized to 0.55 in a volume of 100 μl with fresh MMCG medium and stained with 25 µg ml−1 of DAPI, 5 µg ml−1 of FM1-43 and 5 µM of propidium iodide by incubating samples with dyes for 2 min at room temperature. A total of 5 µl of each sample was pipetted onto an MMCG imaging pad containing 500 µM of IPTG and allowed to dry for about 5 min before imaging. Laser scanning confocal microscopy was performed on a Nikon A1R microscope. For FM1-43, a 488-nm laser and a Chroma ET525/50m emission filter were used. For propidium iodide, a 561-nm laser and a Chroma ET600/50m emission filter were used. All images were acquired using the same laser power and gain settings. For each strain condition and time point, a 3 × 3 image scan was collected; representative areas in these scans are shown in Fig. 4f.
AbCap5 protein expression and purification
The vector for expressing 6×His-hSUMO2-AbCap5 was transformed into Rosetta2 expressing the pRARE2 plasmid and plated onto 1.6% LB agar plates (plus 100 µg ml−1 of carbenicillin and 20 µg ml−1 of chloramphenicol). The plate was allowed to dry and incubated at 37 °C overnight. The next day, an individual colony was picked to inoculate 25 ml of liquid LB media (plus 100 µg ml−1 of carbenicillin and 20 µg ml−1 of chloramphenicol), which was then grown overnight shaking at 220 rpm and 37 °C. The following day, the overnight culture was diluted 1:100 into fresh media plus antibiotics, then grown to an OD600 of around 0.6 before IPTG addition to a final concentration of 500 µM to induce protein expression. Each culture flask was then moved to a 16-°C shaking incubator (220 rpm) and allowed to grow overnight.
The next day, cultures were collected by centrifugation for 30 min at 4,600g and 4 °C in an Avanti JXN-26 Centrifuge using the JXN 12.500 rotor (Beckman Coulter). Each bacterial pellet was resuspended to a total volume of 40 ml of lysis buffer (20 mM of HEPES (pH 7.5), 400 mM of NaCl, 10% v/v glycerol, 30 mM of imidazole and 1 mM of DTT) plus 1 µl of Pierce Universal Nuclease (Thermo Fisher Scientific). The resuspended bacterial pellets were kept chilled and were lysed by sonication at 80% amplitude, with 30 s on, followed by 30 s off, for a total processing time of 30 min using a Sonicator 4000 (Misonix). The lysed bacteria were centrifuged at 4 °C for 1 h at 14,000g in a 5910 R tabletop centrifuge (Eppendorf) to pellet cellular debris left over from sonication.
The resulting soluble lysate was transferred to a new conical tube and kept on ice. To purify protein, the entire soluble lysate was run over 1 ml of lysis buffer-equilibrated Ni-NTA resin (Thermo Fisher Scientific). The flow-through was collected and reapplied to the resin. The resin was then washed with 5 × 25 ml of wash buffer (20 mM of HEPES (pH 7.5), 1 M of NaCl, 10% v/v glycerol, 30 mM of imidazole and 1 mM of DTT). The protein was eluted in 10 ml of elution buffer (20 mM of HEPES (pH 7.5), 400 mM of NaCl, 10% v/v glycerol, 300 mM of imidazole and 1 mM of DTT). The eluted protein was added to 10-kDa MWCO tubing (VWR) and then dialysed in 1 l of dialysis buffer (20 mM of HEPES (pH 7.5), 250 mM of KCl and 1 mM of DTT) for 1 h at 4 °C. Afterwards, the dialysis tubing and protein were placed into 1 l of fresh dialysis buffer and allowed to dialyse overnight at 4 °C. The 6×His-SUMO-tag was cleaved from the N terminus of AbCap5 using 6×His-hSENP2 (purified in-house; final concentration of 1:100 hSENP2:protein w/w), which was added to the purified protein immediately before it was placed in the dialysis tubing. After dialysis and 6×His-SUMO cleavage, the purified proteins were applied to 1 ml of dialysis buffer-equilibrated Ni-NTA and the flow-through was collected and reapplied to the resin to remove any uncleaved 6×His-SUMO tagged proteins.
The purified AbCap5 was concentrated as needed using 30-kDa MWCO Macrosep spin concentration columns (Pall Labs) and stored in 1-ml aliquots in 50% glycerol v/v at −20 °C. A Nanodrop OneC (Thermo Fisher Scientific) was used to measure A280 to determine the protein concentration. Protein purity was determined by using SDS-PAGE followed by Coomassie staining. The AbCap5 protein purified in this way was used in nuclease assays for the measurement of intracellular 2′,3′-c-di-AMP.
Determination of AbCap5 nucleotide specificity
For each 30-µl reaction, AbCap5 (final concentration 325 nM) was incubated with 500 ng of linear PCR-amplified DNA and 3 µl of the indicated nucleotide (final concentration 100 nM) in reaction buffer (final concentrations: 10 mM of Tris-HCl (pH 7.5), 25 mM of KCl and 10 mM of MgCl2). The mixture was incubated for 1 h at 37 °C, then boiled for 5 min at 95 °C. Reactions were mixed with 6 µl of 6× Gel Loading Dye plus SDS (NEB) and 20 µl was loaded and visualized on a 2% (w/v) agarose gel stained with SYBR Safe (Thermo Fisher Scientific).
Bacterial lysate preparation and nucleotide extraction
Bacterial strains expressing either the optSE operon or an empty vector were grown overnight in 25 ml of MMCG plus carbenicillin (100 µg ml−1). The following day, the overnight culture was diluted 1:10 in 100 ml (total volume) of fresh media (without antibiotics) and was allowed to grow to an OD600 of about 0.6–0.7 (noting the specific OD600 at collection). Once the appropriate OD600 was reached, the cultures were centrifuged at 4,000g for 10 min at 4 °C in a 5910 R tabletop centrifuge (Eppendorf) to pellet the bacteria. The resulting supernatant was discarded, and the bacterial pellet was resuspended in 500 µl of lysis buffer (10 mM of Tris-HCl (pH 7.5) and 25 mM of KCl) and transferred to a new Eppendorf tube. Hen-lysozyme (0.2 mg ml−1 final concentration; VWR) and roughly 200 µl of zirconium beads were added before boiling at 95 °C for 2.5 min and then vortexing at max speed for 2 min. The samples were boiled and vortexed again and then treated with proteinase K (0.03 mg ml−1 final concentration; Qiagen) and 1-µl Pierce Universal Nuclease (Thermo Fisher Scientific) for 30 min at 37 °C. After treatment, the samples were boiled and vortexed once again and Triton-X was added to a final concentration of 0.5% v/v. The samples underwent a final boil and vortex, and were centrifuged at 17,500g for 10 min at 4 °C. The resulting lysate/supernatant was transferred to a new Eppendorf tube and kept on ice until ready for nucleotide extraction.
Nucleotides were extracted from bacterial lysates using a modified version of a previously described phenol-chloroform/chloroform extraction protocol13. In brief, 450 µl of phenol-chloroform (Thermo Fisher Scientific) was added to 450 µl of bacterial lysate (prepared above), vortexed for 30 s and then centrifuged at 17,500g for 45 min at 4 °C. A total of 400 µl of the top aqueous layer was carefully removed and added to 400 µl of fresh phenol-chloroform in a new tube. This mixture was vortexed for 30 s and centrifuged at 17,500g for 10 min at 4 °C. A total of 350 µl of the resulting top aqueous layer was carefully removed and added to 350 µl of chloroform (VWR) in a new tube. The mixture was vortexed again for 30 s and centrifuged one final time at 17,500g for 10 min at 4 °C. The resulting top aqueous layer was removed and placed in a new tube for storage at −20 °C. Nucleotide extraction from indicated strains was carried out for a total of n = 3 biological replicates.
AbCap5-based intracellular 2′,3′-c-di-AMP measurements
Commercially available 2′,3′-c-di-AMP (Enzo) was used to make a standard curve by twofold serially diluting the nucleotide in nuclease-free water. A dilution series of nucleotide extracts from the tested bacterial strains were made by twofold serially diluting the extracts in nuclease-free water. For each 25-µl reaction, AbCap5 (final concentration 50 nM) was incubated with 500 ng of linear PCR-amplified DNA and 5 µl of one dilution of nucleotide extract, 2′,3′-c-di-AMP or water in reaction buffer (final concentrations: 10 mM of Tris-HCl (pH 7.5), 25 mM of KCl and 10 mM of MgCl2). The mixture was incubated for 2 h at 37 °C, and then boiled at 95 °C for 5 min. Reactions were mixed with 5 µl of 6× Gel Loading Dye plus SDS (NEB) and 20 µl were loaded and visualized on a 2% (w/v) agarose gel stained with SYBR Safe (Thermo Fisher Scientific).
The concentration of intracellular 2′,3′-c-di-AMP ([n.t.]) was calculated using the following equation:
$$[{\rm{n}}.\,{\rm{t}}.]=\frac{({\rm{S}}{\rm{C}})\times ({\rm{l}}{\rm{y}}{\rm{s}}{\rm{a}}{\rm{t}}{\rm{e}}\,{\rm{v}}{\rm{o}}{\rm{l}}{\rm{u}}{\rm{m}}{\rm{e}})}{({\rm{C}}{\rm{V}})\times ({{\rm{O}}{\rm{D}}}_{600})\times ({\rm{c}}{\rm{u}}{\rm{l}}{\rm{t}}{\rm{u}}{\rm{r}}{\rm{e}}\,{\rm{v}}{\rm{o}}{\rm{l}}{\rm{u}}{\rm{m}}{\rm{e}})}$$
SC is the upper or lower 2′,3′-c-di-AMP standard curve concentration (in [M]) that corresponds to a similar DNA degradation amount as the last visible nucleotide extract dilution reaction (from OptSE-expressing cells) that contains non-degraded or partially degraded DNA PCR product, respectively. Lysate volume is the volume of lysis buffer (in litres) that was used to resuspend the bacterial pellet before the lysis steps. CV is the OD600-specific cell volume (in litres per OD per millilitre) of the bacteria based on the E. coli strain and growth conditions56. OD600 is the optical density of the bacterial culture at the time of collection (before pelleting). Culture volume is the volume of bacterial culture (in millilitres) at the time of collection. Intracellular [2′,3′-c-di-AMP] is reported as an average for the upper and lower range of concentrations (in nanomolars) for n = 3 biological replicates of the nucleotide extraction.
DNase alert-based intracellular 2′,3′-c-di-AMP measurements
Single-use DNase alert (Integrated DNA Technologies) was resuspended in 43 μl of 10× nuclease buffer per tube (1 M of Tris-HCl (pH 7.5), 2.5 M of KCl, 1 M of MgCl2). The following was mixed in each well of a black 96-well, clear-bottom plate: 10 μl of 10× DNase alert, 35 μl of salmon sperm DNA (20 ng μl−1), 5 μl of 2′3′-c-di-AMP nucleotide standard or bacterial lysate and 50 μl of AbCap5 (50 nM). The reaction was monitored at an excitation/emission of 536/556 nm in a TECAN Spark plate reader at 37 °C every 2 min for 30 min.
GraphPad Prism was used to plot the data and determine the velocity of the reaction (that is, slope measured in relative fluorescence units per minute) in the linear range. A chemical standard of 2′,3′-c-di-AMP was used to determine a standard curve by means of linear regression ([agonist] versus response (three parameters); Y = bottom + X × (top − bottom)/(EC50 + X)), which allowed for interpolation of nucleotide concentration in the well from bacterial lysate reaction velocities.
The concentration of intracellular 2′,3′-c-di-AMP ([n.t.]) was calculated using the following equation:
$$[{\rm{n}}.\,{\rm{t}}.]=\frac{([{\rm{lysate}}])\times ({\rm{lysate}}\;{\rm{volume}})}{({\rm{CV}})\times ({{\rm{OD}}}_{600})\times ({\rm{culture}}\;{\rm{volume}})}$$
Recombinant protein expression and purification
The genes for full-length optS from K. pneumoniae strain KP67, optS from V. navarrensis, optE from V. navarrensis (soluble β-barrel only, residues S73-END) and bacteriophage T4 acb2 were codon optimized and synthesized as double-stranded DNA fragments (gBlocks; Integrated DNA Technologies). The genes were cloned using Gibson assembly into a linearized (restriction enzyme digested with BamHI and NotI) in-house pET16 expression vector modified to allow in-frame cloning of an N-terminal hexa-histidine tag with or without a hSUMO2 tag with a Gly-Ser linker. After Sanger sequencing confirmation, the resulting plasmids were subsequently transformed into BL21-RIL E. coli cells (BL21 derivative, DE3; Invitrogen, Life Technologies) for protein expression.
Transformant colonies were grown overnight at 37 °C on MDG media (0.5% glucose, 25 mM of Na2HPO4, 25 mM of KH2PO4, 50 mM of NH4Cl, 5 mM of Na2SO4, 2 mM of MgSO4, 0.25% aspartic acid and trace metals) plates supplemented with 100 μg ml−1 of ampicillin and 34 μg ml−1 of chloramphenicol for selection. Single colonies were used to inoculate liquid MDG media cultures (same components without agar), which were grown overnight for about 16–20 h at 37 °C and 230-rpm shaking. Overnight cultures were used to inoculate 1 l of liquid M9ZB expression media (0.5% glycerol, 1% Cas-amino acids, 47.8 mM of Na2HPO4, 22 mM of KH2PO4, 18.7 mM of NH4Cl, 85.6 mM of NaCl, 2 mM of MgSO4 and trace metals) contained in 2.5-l flasks. After cultures reached an OD600 of greater than 2.5, flasks were placed on ice for 15 min; following this, 0.5 mM of IPTG (final) was added and the mixture was incubated at 16 °C with 230-rpm shaking overnight (about 16–20 h). Cells were collected by centrifugation, washed with PBS buffer and flash-frozen with liquid N2 and stored at −80 °C until needed.
Conventional nickel-affinity chromatography was carried out using gravity flow at 4 °C. Briefly, E. coli cell pellets were resuspended with lysis buffer (20 mM of HEPES-KOH (pH 7.5), 400 mM of NaCl, 10% glycerol, 30 mM of imidazole (pH 7.5), 1 mM of DTT) and subjected to sonication to release cellular contents (10 s on, 20 s off, 70% amplitude, 5 min total on time). The lysate was clarified with centrifugation and the supernatant loaded onto 4–6 ml of packed Ni-NTA resin pre-equilibrated with lysis buffer. The column was then sequentially washed with 20 ml of lysis buffer, 70 ml of wash buffer (20 mM of HEPES-KOH (pH 7.5), 1 M of NaCl, 10% glycerol, 30 mM of imidazole (pH 7.5), 1 mM of DTT) and a final 35 ml of lysis buffer to remove high concentrations of salt. Protein was eluted with 20 ml of elution buffer (20 mM of HEPES-KOH (pH 7.5), 400 mM of NaCl, 10% glycerol, 300 mM of imidazole (pH 7.5), 1 mM of DTT). The eluent was dialysed overnight in 20 mM of HEPES-KOH (pH 7.5), 250 mM of KCl and 1 mM of DTT at 4 °C with gentle stirring. In the case of Acb2, hSENP2 protease (D364–L589, M497A) was added to the dialysis tubing to cleave the N6xHis-SUMO2 tag. The dialysed eluent was concentrated and further purified through size-exclusion chromatography using a HiLoad 16/600 Superdex 200 pg column (Cytiva) equilibrated with 20 mM of HEPES (pH 7.5), 250 mM of KCl and 1 mM of TCEP-KOH. Size-exclusion chromatography fractions were analysed with SDS-PAGE; fractions that contained recombinant proteins of interest were pooled and concentrated to more than 10 mg ml−1, were flash-frozen in liquid N2 and then stored at −80 °C until needed.
Protein crystallization and structure determination
Apoprotein crystals for OptS from K. pneumoniae strain KP67 were obtained at room temperature by mixing 1 µl of reservoir solution consisting of 8% (w/v) PEG 4000 with an equal volume of 7 mg ml−1 of the protein solution using the hanging-drop vapour diffusion method. The crystals were soaked in mother liquor with 15% ethylene glycol as cryoprotectant and subsequently plunged into liquid N2 and shipped for remote data collection.
Crystals for KpOptS bound to ATP analogue ApCpp were grown at room temperature using the hanging-drop vapour diffusion method in 16% w/v PEG 4000 (reservoir solution). KpOptS was diluted to 7 mg ml−1 with buffer containing 100 mM of KCl, 20 mM of HEPES-KOH (pH 7.5), 1 mM of TCEP with the addition of 1 mM of ApCpp, 10 mM of MgCl2 and 1 mM of MnCl2 at their final concentration. The crystals were grown in 15-well trays (NeXtal) containing 350 μl of reservoir solution and 2-μl drops. Drops were mixed 1:1 with the purified protein-ligand mixture and reservoir solution. Crystals were cryoprotected with reservoir solution supplemented with 30% glycerol with the addition of 1 mM of ApCpp, 10 mM of MgCl2 and 1 mM of MnCl2 and collected by flash-freezing in liquid nitrogen.
X-ray diffraction data for apo KpOptS were collected at experimental station 12-2 at the Stanford Synchrotron Radiation Lightsource (SSRL) using a Dectris Pilatus 16 M PAD detector. Beamline 12-2 was set at 0.97946 Å and the crystal data were collected at 100 K. The dataset was processed using the X-ray Detector Software package57 that is incorporated into the automated ICEflow (autoPROC) processing pipeline used at SSRL. POINTLESS and AIMLESS were used for scaling and space group assignment, and TRUNCATE for conversion to structure factors. X-ray diffraction data for ApCpp-bound KpOptS were collected at beamline 5.0.1 at the Advanced Light Source using a Dectris Pilatus 32 M detector. The beamline was set at 0.97741 Å wavelength and the crystal data were collected at 100 K. The structures were determined using molecular replacement conducted by the Phaser-MR program in the PHENIX suite (v.1.21-5207)58 using a predicted structural model of KpOptS generated by ColabFold v.1.5.5, which uses a homology search by MMseqs2 with AlphaFold259. The structures of KpOptS were iteratively refined using the phenix.refine program and the residue positions were manually adjusted in Coot (v.1.1.17)60. The ApCpp ligand was a predefined ligand searchable in the PDB under ligand identifier APC, whereas the pCpp fragment ligand restraints were generated using the Elbow tool in Phenix based on the SMILES code: O = P(O)(O)CP(=O)(O)OP(=O)(O)O. The final refined structures had the following Ramachandran statistics—apoprotein: 98.09% favoured, 1.91% allowed, 0.00% outlier; ApCpp-bound: 96.98% favoured, 3.02% allowed, 0% outlier. The full data collection and refinement statistics are summarized in Extended Data Table 1. The models were visualized and analysed using PyMOL (The PyMOL Molecular Graphics System, v.3.1.6.1, Schrödinger). The atomic coordinates for apo KpOptS and KpOptS bound to ApCpp have been deposited to the PDB using IDs 9MNR and 9PD0, respectively. Buried surface area analysis was conducted on the apo KpOptS structure by uploading the coordinate file to the PDBePISA web-based program provided by the PDB in Europe61.
Mass photometry analysis of KpOptS
Coverslips (24 mm × 50 mm, Thorlabs) and silicon gaskets (Grace Bio-Labs) were cleaned with several rinses of ultrapure water and HPLC-grade isopropanol and then dried with a clean nitrogen stream. A clean gasket was adhered to a coverslip, and the assembly was placed on the stage of a Refyn TwoMP mass photometer (Refeyn) and centred on a single well. A total of 15 μl of KpOptS diluted to 50 nM in freshly prepared sample buffer (20 mM of HEPES (pH 7.5), 250 mM of KCl) was added to the well and a 60-s video was recorded at room temperature after auto-focusing using the Refeyn AcquireMP (v.2.3.0) software. Each measurement was done in triplicate on two different days. A total of 50 nM of β-amylase in sample buffer was used as a calibration standard (56 kDa, 112 kDa and 224 kDa). Data were processed using DiscoverMP (v.2.3.0; Refeyn).
HPLC analysis of enzyme activity
Enzymatic reactions were performed in total volumes of 100–600 µl with, typically, 250 µM in total of equimolar ribonucleotide triphosphates (ATP, GTP, CTP, UTP), 1 mM of MnCl2, 10 mM of MgSO4, 20 mM of HEPES-KOH (pH 7.4) and 100 mM of NaCl, with 100 µM of OptS protein added last. Reactions were transferred to a 10-kDa cut-off filter and subjected to centrifugation for 10 min at 13,500g. When appropriate, purified Acb2 was added at 250 µM and incubated for a further 1 h at 37 °C. A total of 1 µl of about 20 mg ml−1 of proteinase K (NEB, catalogue no. P8107S) was added to degrade Acb2 and release cyclic dinucleotide before filtering. Likewise, samples were at times treated with P1 nuclease and/or CIP as indicated. All samples were injected at 10 µl onto an Agilent 1200 Series HPLC equipped with a 4.6 × 150 mm and 5 µm particle-size Zorbax Bonus-RP C18 column using an isocratic elution method including 97% 50 mM NaH2PO4 (pH 6.8) and 3% acetonitrile buffer system held at 40 °C. Nucleotide separation was monitored at 254 nm using a multiwavelength detector. For mass spectrometry analysis, enzymatic reactions were prepared using 20 mM of ammonium acetate (pH 8.0) instead of HEPES-KOH and NaCl, but otherwise treated the same with or without an liquid chromatography fractionation step to isolate specific peaks. For fractionation on HPLC, the temperature was kept at 23 °C with a gradient elution method as follows: solvent A—20 mM of ammonium acetate (pH 8.0) and solvent B—methanol; 0–2 min, 100% A; 2–8 min, 0–100% B; 8–10 min, 100% B; 10–11 min, 0–100% A; 11–17 min, 100% A. Roughly 1-ml fractions were collected and concentrated through SpeedVac (1–3 h, 30 °C) before injection on a Waters Acquity ultra-performance liquid chromatography (UPLC)-QDa instrument equipped with a C18 50-mm UPLC column and single quadrupole mass spectrometry detector (QDa; m/z 50–1250) and photodiode array detector (200–500 nm). Standard gradient elution was used from 0–100% 0.1% formic acid:acetonitrile over 5 min. In ESI+ mode, analyte m/z ions (M + H)+ or (M + NH3)+ were validated to less than ±5 ppm relative to the nearest sodiated polyethylene glycol (CAS: 25322-68-3, av. Mwt 400) or sodiated methoxypolyethyleneglycol (CAS: 990-74-4, av. Mwt 350) calibrant peak lockmass.
Thermal shift analysis of effector nucleotide binding
The soluble β-barrel domain of the S-2TMβ VnOptE effector (residues S73-END) was briefly incubated at a final concentration of 20 µM in reaction buffer containing 3× SYPRO Orange Dye, 100 mM of NaCl, 20 mM of HEPES-KOH (pH 7.5) and 500 µM of the respective cyclic nucleotide as indicated. Sample volumes of 25 µl each were heated from 20 °C to 95 °C over the course of about 1.2 h using a Bio-Rad CFX Duet Real-Time PCR System. The Förster resonance energy transfer channel with excitation at 450–490 nm and detection at 560–580 nm was used to monitor SYPRO fluorescence with every 0.5-°C increment. The first derivative of each fluorescence curve was calculated, and the melting temperature was identified as the peak of each derivative curve. Data are presented as an average of nine technical replicates and representative of at least three biological replicates.
Computational analysis of mCpol sequences and structures
PSI-BLAST (Research Resource Identification (RRID):SCR_001010)62 and jackhmmer (RRID:SCR_005305)63 were used to carry out iterative sequence profile searches against the non-redundant protein database (nr) from the National Center for Biotechnology Information (NCBI) clustered down to 90% identity (nr90) to eliminate redundancy. These searches were used to collect the complete complement of sequences belonging to the mCpol family. False-positive hits to Cas10 were eliminated using reverse profile searches with the RPSBLAST program64. Clustering on the basis of percentage or bit-score similarity was performed using MMseqs (RRID: SCR_008184)65, with parameters changed according to the clustering objective. Domains of mCpol-containing proteins and the proteins translated from surrounding genes on the genome with known homology were annotated using a database of domain sequence profiles including Pfam A models (RRID: SCR_004726)66. Profile–profile comparisons were performed with the HHpred program67. Structural models were constructed using the AlphaFold3 program68. Structure similarity searches were performed using the DALI69 and FoldSeek programs70.
Phylogenetic distribution and significance analysis
Taxonomic lineages were obtained from the NCBI Taxonomy Database (RRID: SCR_003256). In total, 13,710 completely sequenced prokaryotic genomes encompassing the organisms from these lineages were downloaded from NCBI GenBank and checked for the presence or absence of representative conflict system components using the PSI-BLAST iterative search program. This was followed by clustering for homologue groups and reverse searches as described above. NCBI genome record (GCA) numbers were used as markers for unique organisms. The significance of pairwise system overlap within the sampled genome space was calculated using Fisher’s exact test implemented in the R command fisher.test(). Data processing (knitr and dplyr libraries), statistical analysis (stats and Rmpfr libraries) and visualization (VennDiagram library) were performed using the R language.
Synthetic nucleotide ligands
Synthetic cyclic dinucleotide ligands used for HPLC, thermofluor and mass spectrometry analysis were purchased from Biolog Life Science Institute: 3′,3′-c-di-AMP (catalogue no. C 088), 2′,3′-c-di-AMP (catalogue no. C 187), 2′,2′-c-di-AMP (catalogue no. C 188), 3′,3′-c-di-GMP (catalogue no. C 057), 2′,3′-c-di-GMP (catalogue no. C 182), 2′,2′-c-di-GMP (catalogue no. C 162), 3′,3′-cGAMP (catalogue no. C 117), 2′,3′-cGAMP (catalogue no. C 161), 3′,2′-cGAMP (catalogue no. C 238), 2′,2′-cGAMP (catalogue no. C 210), pppA(2′,5′)pA (catalogue no. T 073), 3′,3′-cUAMP (catalogue no. C 357), 3′,2′-cUAMP (catalogue no. C 398), 2′,3′-cUAMP (catalogue no. C 399), c-tri-AMP (catalogue no. C 362), c-tetra-AMP (catalogue no. C 335), c-penta-AMP (catalogue no. C 394) and c-hexa-AMP (catalogue no. 332). ApCpp was purchased from Jena Bioscience (catalogue no. NU-421).
Accession numbers
The crystal structure data for apo KpOptS and KpOptS bound to ApCpp have been deposited in the PDB (9MNR and 9PD0, respectively). All other relevant accession numbers can be found in Supplementary Table 4.
Statistics and reproducibility
Each experiment presented was performed with independent biological replicates using bacterial cultures grown on separate days. The n, error bars and bar graph data are defined in the figure legends. Error bars were selected according to published recommendations71,72,73. Data were plotted using GraphPad Prism 9. Statistical analysis of specified comparisons was performed, and all results can be found in the corresponding source data. In all cases, a two-sided, unpaired, parametric t-test was used. To compare multi-log differences, data were log-transformed before statistical analysis. When four or fewer comparisons are made in a single graph, statistical comparisons are shown in the figure. In all other cases, statistics are presented in the source data to aid clarity. *P < 0.05, **P < 0.001. Illumina sequencing results were analysed using Geneious Prime Software (v.2024.0.1).
Reporting summary
Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.
Source link
