Genetic identification of a candidate novel anti-phage system

Previous studies have demonstrated that bacterial defense systems cluster together in chromosomal hotspots, defined as ‘defense islands’ and that their mobilisation is dependent on mobile genetic elements [17,32,33]. Genetic neighborhood analysis of known anti-phage systems has allowed the systematic discovery of many novel defense systems [2,4,13,14,16,17,32]. Following the same principle, we aimed to identify bacterial operons, situated in the context of defense islands, whose role has not yet been associated with defense against invasion of foreign DNA.

Pseudomonas aeruginosa is among the species that have been reported to encode many distinct anti-phage systems [34]. To identify new candidate anti-phage systems in this organism and across the genus we downloaded all available genome sequences of Pseudomonas species from the Refseq database to the ‘Scaffold’ assembly level (ftp://ftp.ncbi.nlm.nih.gov/genomes/ASSEMBLY_REPORTS/assembly_summary_refseq.txt, as of 11th June 2022/~3500 genomes), including entries that represent only portion of genomes (Scaffolds) that have been reconstructed from the assembly of whole-genome shotgun (WGS) sequencing data. Subsequently we used the Defense Finder tool to identify known anti-phage systems [34] (S1 Table). We then manually screened for flanking genes or operons that associated with the defense systems identified in S1 Table and whose function was either unknown (genes annotated to encode hypothetical proteins), associated with interaction/degradation of nucleic acids, or that could mediate cell death or stasis (i.e. peptidoglycan hydrolase, lipases, pore-formation, NAD⁺ depletion, etc.). Our search consistently revealed a gene that was encoded next to several known anti-phage systems (Fig 1A). We performed function predictions using the protein encoded by P. aeruginosa NCTC 11442/ATCC 33350 (assembly ID: GCF_001420205.1, locus tag: AN400_RS26690, protein ID: WP_023115263). This protein represents one of the most frequently occurring homologues of the new candidate defence gene we identified (S2 Table). Predictions with hmmscan against a PFAM database (26) revealed that the C-terminal region of AN400_RS26690 contains a PF02646 domain, typical of RmuC proteins (Fig 1B). Additionally, AN400_RS26690 harbors a predicted trans-membrane domain (TMH) at its N-terminus (Fig 1B).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. Identification of a novel anti-phage system.

(a) Schematic of the genomic neighbourhood of Shield systems in representative genomes. Known defense genes were predicted using PFAM and Defense-finder [34]. ShdA is in blue and other Shield partners are represented with coloured outline, as indicated. The full set of Shield subtypes is shown in S2 Fig and known anti-phage systems annotations are reported in S6 Table. (b) Schematic representation of the predicted domain organisation of the candidate defence protein, ShdA. (c) Phylogenetic tree based on ShdA homologues. Coloured blocks indicate ShdA homologues belonging to each Shield subtype.

https://doi.org/10.1371/journal.pgen.1010784.g001

Subsequently, to extend the search and detect a wider diversity of homologues, we built a Hidden Markov Model (HMM) model, including RmuC-like proteins from several species. We then queried a local Refseq of all bacterial genomes to the ‘contig’ assembly level (ftp://ftp.ncbi.nlm.nih.gov/genomes/ASSEMBLY_REPORTS/assembly_summary_refseq.txt, as of 11th June 2022/~150000 genomes). With the inclusion of contigs, which cover a smaller percentage of the full genome than Scaffolds, we aimed to expand the breadth of bacterial species considered in our search, to find more instances of RmuC-like domains associated with defense. The dataset was manually curated to include only those homologues found within defense islands (thus excluding housekeeping RmuC) (Figs 1A and S1). This resulted in ~70 unique protein identifiers and we renamed these protein homologues as ShdA. We subsequently used flanking gene (FlaGs) analysis [35] to define the genomic neighborhood of the shdA genes (Figs 1A and S1 and S3 and S4 Tables). From this we noted that although in some instances shdA could be found as an orphan gene, in other instances it co-occurred with subsets of additional genes which were always encoded downstream (Figs 1A and S1 and S3 Table). We renamed shdA-containing operons ‘Shield’ and classified them in 4 subtypes according to their gene composition. (Fig 1A and Table 1).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Summary of predicted protein components for each Shield subtype.

https://doi.org/10.1371/journal.pgen.1010784.t001

Despite the relaxed parameters used in search of RmuC homologues, Shield subtypes were, to date, only found in Pseudomonas spp and predominantly in P. aeruginosa (S2 Table). In particular, of ~14000 Pseudomonas sequences available to contig level in the Refseq database analysed, Shield subtypes were found in 6.8% of Pseudomonas genomes.

Where ShdA was encoded alone, we designated this Shield I. Of the other three candidate systems, Shield subtype II is the most widely distributed, representing 72% of the total Shield systems found and we renamed its partner gene shdB. Shield I, III and IV instead represented 15.1%, 6.8% and 5.5% of the total Shield systems found, respectively (Figs 1 and S1 and S2 Table). We named the other partner genes shdC—shdE according to the order they appeared in the FlaGs schematic representation of Shield subsystems (Fig 1A, Table 1). A BLAST search with strict query coverage and sequence similarity parameters (80–100% and 70–100%, respectively) revealed that ShdC, ShdD and ShdE proteins are predominantly found in association with Shield subsystems.

The 15 proteins encoded directly upstream and downstream of shdA were annotated using Defense-finder and PFAM predictions (Figs 1A and S1 and S4–S6 Tables), confirming that Shield subtypes localize adjacent to known anti-phage systems (Figs 1A and S1, and S4–S6 Tables). They also often associate with WYL-domain containing proteins, transcriptional regulators that are enriched in phage defense islands [3,36,37]. Furthermore, the genome neighborhood also encodes several viral proteins and integrases, in agreement with recent reports that mobile genetic elements (MGE) represent primary carriers of defense islands [17,32,33]. These findings further support the involvement of the Shield systems in bacterial immunity.

Alignment of ShdA homologues reveals a high degree of shared sequence identity, particularly towards the C-terminus (S2 and S3 Figs). A phylogenetic tree (based on the alignment in S3 Fig), reveals that although Shield subtype III is more similar to Shield subtype II in terms of gene composition (both having a ShdA and ShdB homologue), ShdA III homologues cluster closer to homologues from subtype I and IV (Fig 1C), whilst some ShdA I proteins are found in-between ShdA II branches (Fig 1C).

Within our ShdA homologues hits, we also found ~70 homologues embedded within DISARM-like operons (S4A and S5 Figs and S7 Table). DISARM-associated ShdA proteins exhibit a longer sequence length, with an additional sequence stretch at the N-terminus (S6 Fig). Phylogenetic analysis shows that DISARM-related ShdA homologues cluster separately from those of Shield, suggesting ShdA homologues associated to Shield and DISARM have diverged early (S4B Fig). Interestingly, the taxonomic distribution of DISARM-associated ShdA proteins is also exclusively limited to Pseudomonas spp. (S8 Table).

ShdA is the core anti-phage defense module in Shield subtypes

Since ShdA is the sole component of Shield I, we reasoned that this protein might represent the core ‘defense’ module of Shield systems. To investigate this, we cloned representative examples of shdAI—shdAIV under control of the arabinose-inducible promoter in plasmid pBAD18. We subsequently transformed Escherichia coli MG1655 with these constructs and tested their ability to protect E. coli growing in liquid culture from lysis by the unrelated lytic phages ϕSipho and ϕAlma (Fig 2).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2. Shield I, Shield II and Shield IV prevent phage infection.

Growth curves of E. coli MG1655 carrying empty pBAD18 (VC) or the same plasmid encoding (a) ShdA I (the sole component of the Shield I system), (b) the Shield II system or ShdA II only, (c) Shield III or ShdA III only, and (d) Shield IV or ShdA IV only. Strains were grown in LB medium supplemented with 0.2% L-arabinose and phages ϕSipho or ϕAlma were added at the start of the growth curve at a MOI = 1. Points show mean +/− SEM (n  =  3 biological replicates).

https://doi.org/10.1371/journal.pgen.1010784.g002

We observed that Shield I, which only comprises the ShdA I module, confers protection against the phage ϕSipho (Fig 2A), but was ineffective against ϕAlma. Similarly, ShdA II and ShdA IV also provided defense against phage-mediated lysis (Fig 2B and 2D), with ShdA II protecting against the action of both ϕSipho and ϕAlma, and ShdA IV against ϕAlma only. We also tested the ability of the full Shield II and Shield IV systems to provide protection, and noted that in each case the cognate ShdA elicited at least similar levels of protection as the full system (Fig 2B and 2D). In addition, we also observed that for ϕSipho, ShdA II alone provided better protection compared to Shield II (Fig 2B). Conversely, ShdA III and Shield III were ineffective against both phages (Fig 2C). We conclude that at least three of the Shield subtypes are involved in bacterial immunity, and that the defense phenotype is dependent on the conserved component ShdA. We also conclude that the different subtypes we tested show phage-specific patterns of protection.

Shield II and ShdA II mode of action involves reduction of phage burden

To better investigate the mode of action of Shield in phage defense in vivo, we focused on Shield II from P. aeruginosa NCTC 11442/ATCC 33350 (assembly ID: GCF_001420205.1), as this represents the most common Shield subtype (S2 Table).

In this strain, the Shield II locus is preceded by a WYL-domain containing protein. Sequence analysis by BLAST and structural predictions indicated this is an homologue of BrxR, a regulator frequently found in defence islands [3,36,37](Fig 1). Shield II alone or Shield II and the upstream BrxR were cloned with their native promoter in a transposon mini-Tn7 vector (S11 Table, see Material and Methods), for site-specific insertion into the chromosome of a P. aeruginosa PA14 strain with a non-functional CRISPR-Cas system (csy3::LacZ) [38,39]. We determined the EOP of Shield II-harboring strains against several P. aeruginosa phages and confirmed that Shield II is a bona fide anti-phage system in Pseudomonas (Fig 3A). We further observed that the upstream BrxR homologue appeared to repress Shield II-mediated protection (Fig 3A).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 3. Characterisation of the anti-phage activity of Shield subtype II.

(a) Efficiency of plating (EOP) measurement for P. aeruginosa PA14 csy3::LacZ (strain with a disrupted CRISPR-Cas system) carrying either a chromosomally-integrated Shield II with its native promoter, Shield II and the upstream WYL-domain containing protein or the empty transposon mini-Tn7 vector used for chromosomal integration. Strains were challenged with several phages as shown in panel a, Points show mean +/− SEM (n  =  3 biological replicates). (b) Evaluation of ShdA II fold protection against a suite of phages. Fold protection was obtained by dividing the value of the efficiency of plating (EOP) calculated for a strain expressing ShdA II by the EOP value of a strain carrying the empty vector, when infected with phages, as shown in panel b. (c) Efficiency of plating (EOP) measurement for E. coli MG1655 carrying empty vector (VC, pBAD18) or the same plasmid encoding Shield subtype II when infected with phages that showed susceptibility to ShdA II in Fig 2B. Points show mean +/− SEM (n  =  3 biological replicates). (d) Efficiency of plating (EOP) measurement for E. coli MG1655 carrying empty vector (VC, pBAD18) or the same plasmid encoding the Shield II system, ShdA II only or ShdB II only when challenged with phages ϕSipho, ϕTB34, ϕAlma, ϕBaz and ϕNR1. Points show mean +/− SEM (n  =  3 biological replicates) except for phages ϕSipho and ϕTB34 where n  =  4 biological replicates. E. coli MG1655 was used as a reference strain for all EOP measurements. (e) Average burst size assessment for the same strain and plasmid combinations as (c) following infection with ϕSipho at MOI 0.1. Points show mean +/− SEM (n  =  3 biological replicates). Statistical analysis for panel a-c was performed with GraphPad applying unpaired student t test. No significance was detected, unless indicated (*p ≤ 0.05). For panels d-e statistical relevance was measured using one-way ANOVA with Dunnett’s multiple comparison test. No significance was detected, unless indicated (*p ≤ 0.05).

https://doi.org/10.1371/journal.pgen.1010784.g003

Having confirmed activity in the original host species, we returned to heterologous expression of Shield systems in E. coli for downstream studies. We first tested the ability of ShdA II and Shield II to provide protection against a suite of E. coli phages. For ShdA II we calculated a fold protection value (ratio between efficiency of plating (EOP) of ShdA II and the EOP of cells carrying empty vector) when challenged with a panel of phages. This ratio provided a quick read-out of ShdA II ability to confer protection against the group of phages tested (Fig 3B). We found that ShdA II conferred resistance to several different phages, which were also susceptible to full-length Shield II (Fig 3C).

Subsequently, we directly compared the EOP on strains producing either the full Shield II system or ShdA II or ShdB II alone when infected with phages ϕSipho, ϕTB34, ϕAlma, ϕNR1 or ϕBaz. We observed that both Shield II and ShdA II conferred a decrease in EOP, albeit with a more modest effect against ϕAlma and ϕBaz (Fig 3C and 3D). Furthermore, Shield II and ShdA II decreased the burst size of ϕSipho, with a more marked effect for ShdA II-harbouring cells (Figs 3E and S7A).

Next, we investigated Shield II and ShdA II-mediated defense in liquid culture against ϕSipho or ϕTB34. Cells harboring Shield II or ShdA II showed better survival than cells carrying empty vector (VC) or ShdB II only, especially at lower MOI values. Additionally, ShdA II- expressing cells showed a better growth rate than Shield II-harboring cells, during phage infection (S7B Fig).

To better determine how Shield and ShdA II mediates phage defense, we assessed the progression of ϕSipho infection over the course of 24 hrs, measuring PFU/mL, CFU/mL and OD_600nm at several timepoints (Fig 4A–4C). At 3 and 6 hrs post-infection, cells containing vector only or producing ShdB II alone exhibited a complete culture collapse (Fig 4A–4C). At later timepoints, these strains formed small colonies that increased over time, an indication that escape mutants had arisen (Fig 4C). Conversely, cells harboring Shield II or ShdA II displayed an increased growth rate and CFU/mL counts compared to empty vector (Fig 4B). Nevertheless, at later time points the cell counts of Shield II and ShdA II cells were reduced compared to the same strains at t = 0 hrs, whereas their growth in liquid remained unaffected (Fig 4B and 4C).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 4. Shield II and ShdA II provide population immunity.

E. coli MG1655 harboring pBAD18 (pBAD18) or the same plasmid encoding Shield II, ShdA II only or ShdB II only were cultured in LB medium containing 0.2% L-arabinose and infected with ϕSipho at MOI 0.1. Following infection, the a) titre (PFU/mL), b) cell counts (CFU/mL) and c) the growth rate (OD_600nm) of each culture was measured at several timepoints, as shown in panels a-c, over the course of 24 hrs post infection. Points show mean +/− SEM (n  =  3 biological replicates). Statistical analysis for panel a was performed with GraphPad using one-way ANOVA with Dunnett’s multiple comparison test. No significance was detected, unless indicated (*p ≤ 0.05).

https://doi.org/10.1371/journal.pgen.1010784.g004

Quantification of phage produced at each timepoint showed that in the presence of Shield II or ShdA II, ϕSipho replication was hindered despite the reduced cell survival (Fig 4A and 4B). Notably, at 24 hrs post infection the titre of ϕSipho on ShdA II-harboring cells did not show any increase compared to 0 hr (Fig 4A). Conversely, ϕSipho retained partial ability to replicate in cells carrying the full Shield II system (Fig 4A). Efficiency of Centre Of Infection (ECOI) measurements further confirmed that in presence of ShdA II a fewer number of infective centres were formed compared to the full Shield II (S7C Fig).

Taken together, our results show that Shield II and ShdA II promote population-wide immunity by decreasing the cell survival of infected cells (Fig 4B and 4C). Whilst the reduced survival of Shield II- and ShdA II-harboring cells could be part of the immunity mechanism, the phenotypes observed could be partially due to incomplete clearance of the tested phage infection, especially in the case of Shield II (Fig 4A–4C).

From the experiments evaluating the cell survival in liquid cultures, phage replication, burst size, and ECOI measurements of strains carrying Shield II and ShdA II, we consistently noted that presence of ShdA II alone reduces the amount of released phage particles to a greater extent than that observed for cells harboring Shield II. These results suggest that ShdB II may negatively affect the activity of ShdA II (Figs 2, 4 and S7B and S7C).

ShdA is a membrane protein that has nuclease activity in vitro

Since ShdA proteins are sufficient to elicit protection against phage infection (Fig 2), our next aim was to characterise ShdA activity. Most of ShdA homologues we identified are predicted to have a transmembrane domain at their N-terminus (S9 Table). To confirm that this prediction is correct, we fractionated E. coli cells producing C-terminally His-tagged ShdA II. Western blot analysis revealed that the protein was only detected in the membrane fraction (Fig 5A). We then asked the question whether membrane-localization of ShdA is essential for its anti-phage activity. To this end we attempted to clone truncated shdA genes lacking the 5′ regions encoding the membrane anchor. However, all our efforts were unsuccessful, even when we used tightly repressible vectors, suggesting that the encoded proteins were toxic to E. coli. Thus, it seems likely that anchoring of ShdA to the membrane modulates its toxicity.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 5. ShdA homologues exhibit nuclease activity in vitro.

(a) Cells carrying plasmid pBAD18 encoding ShdA II-His₆ were grown for 5 hrs in the presence of 0.2% L-arabinose. Cells were fractionated to produce soluble and membrane samples and analysed by immunoblot with antibodies to the His6 tag, GroEL (cytoplasmic control) and TatA (membrane control). (b-e) In vitro DNAse activity assays using (b) ShdA II_138-524 (c) ShdA I_140-526 (d) ShdA III_135-523 and (e) ShdA IV_135-521. ShdA proteins and DHFR were synthesised using the cell-free PURExpress kit (NEB). DNAse activity was tested against 10 ng of input DNA. DNA types tested were phage DNA, E. coli MG1655 chromosomal DNA and plasmid (pSG483) DNA. For ShdA I_140-526, ShdA III_135-523 and ShdA II_138-524 phage DNA was from ϕSipho. For ShdA IV_135-521 phage DNA was from ϕAlma. (f) Evaluation of transformation efficiency for E. coli MG1655 carrying empty vector (VC, pBAD18) or the same plasmid encoding Shield subtype II with plasmid DNA. For all panels, 0.2% L-arabinose was added at time zero to induce expression of the encoded genes in pBAD18. Points show mean +/− SEM (n  =  3 biological replicates). Statistical analysis was performed using one-way ANOVA with Dunnett’s multiple comparison test. No significance was detected, unless indicated (*p ≤ 0.05).

https://doi.org/10.1371/journal.pgen.1010784.g005

It was previously shown that E. coli RmuC displays a restriction endonuclease-like fold and is predicted to regulate genome inversion events through DNA cleavage [40]. We therefore next sought to test whether the RmuC domain in ShdA has nuclease activity. As we were unable to clone the truncated shdA alleles, we used cell-free synthesis to generate the proteins. Following cell-free synthesis using a template for ShdA II_138-524, the products caused degradation of phage ϕSipho DNA, E. coli chromosomal DNA and plasmid DNA (Fig 5B). No degradation was observed when using products generated with the kit-provided template for the control protein dihydrofolate reductase (DHFR) (Fig 5B).

To confirm whether this was a common feature of ShdA proteins, we also tested the ability of in vitro synthesised ShdA I_140-526, ShdA III_135-523 and ShdA IV_135-521 to degrade DNA. Whilst the yield of each was low, the products all showed DNase activity (Fig 5C–5E), confirming that DNA degradation is a common feature of ShdA proteins. Interestingly, although we were not able to identify any phages in our collection that were sensitive to Shield III, the ShdA III protein was capable of partially degrading DNA from phage ϕSipho. This suggests that ShdA III is also likely to be active in phage defense, but that our phage collection does not contain representative phage that are sensitive to its activity.

As ShdA homologues exhibited the conserved ability to degrade plasmid DNA, we tested whether Shield II/ShdA II could affect the efficiency of plasmid DNA acquisition during transformation. Indeed, both Shield II and ShdA II decreased transformation efficiency (Fig 5F).

ShdA induces bacterial nucleoid re-arrangement that is modulated by ShdB

The in vitro nuclease assays (Fig 5B–5E) indicated that the RmuC domain of ShdA has DNase activity that can be active against host chromosomal DNA. To determine whether cells producing ShdA II were anucleate or exhibited other types of aberrant cellular DNA organisation such as changes in DNA condensation, we examined them using the DNA-specific stain 4′,6-diamidino-2-phenylindole (DAPI) and fluorescence microscopy. We observed that over a time-course of ShdA II induction, E. coli chromosomal DNA was not detectably depleted/degraded, as indicated by the absence of anucleate cells. However, the bacterial nucleoid underwent drastic spatial rearrangement within the cell (Fig 6A). After 2 hr of induction, the nucleoid of ShdA II-expressing cells adopted a distinct morphology and cellular localisation, with an increasingly clustered (S8A Fig) and peripheral DNA distribution pattern (Fig 6A and 6B). This data suggests that ShdA II can disturb the native organisation of the nucleoid and recruit chromosomal DNA to the cell periphery, most likely through direct protein-DNA interactions. Cells producing Shield II showed a less extensive redistribution of DAPI fluorescence, again consistent with a role for ShdB II in negatively impacting the function of ShdA II. As evidenced by unaltered or increased overall DAPI staining levels (S8B Fig), the ShdA-mediated disturbances in host chromosome organisation are not associated with extensive degradation of chromosomal DNA, although DNA fragmentation cannot be ruled out. Measurements of the cell-to-cell correlation of fluorescence variance and integrated intensity allowed to confirm that the increased variance in ShdA- and Shield-expressing cells (S8A Fig) is not due to increased DAPI staining levels (S8C–S8F Fig).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 6. ShdA alters bacterial nucleoid morphology.

(a). E. coli MG1655 harboring pBAD18 (pBAD18) or the same plasmid encoding Shield II, ShdA II only or ShdB II only were cultured in LB medium containing 0.2% L-arabinose for 2 hours. At t = 0 hours, t = 1 hour and t = 2 hours an aliquot of each culture was removed and stained with DAPI (for DNA visualisation) and FM 5–95 dye (to stain cell periphery/ outer membranes). Cells were subsequently imaged using fluorescence microscopy (b) The ratio between DAPI fluorescence intensity at the cell periphery and the whole cell was quantified to assess ShdA II-mediated recruitment of DNA at the cell periphery for strains in (a). Points show mean +/− SEM (n  =  100 cells). Statistical analysis was performed using one-way ANOVA with Dunnett’s multiple comparison test. No significance was detected, unless indicated (*p ≤ 0.05).

https://doi.org/10.1371/journal.pgen.1010784.g006

Accessory Shield components may modulate ShdA-mediated defense or provide additional defence components

To investigate the role of ShdB in the Shield II system we generated a high confidence structural model using AlphaFold (S9A–S9C Fig) and used it to search the Dali server [41,42]. This predicted structural similarity to the metallopeptidase M15 family and to several carboxypeptidases (S9A–S9C Fig and S10 Table). While M15 family proteins are normally involved in peptidoglycan biosynthesis and turnover, ShdB harbors no detectable signal peptide that would direct it to the periplasm. Furthermore, when we provided ShdB II with the signal peptide from E. coli OmpA protein, to mediate its secretion to the periplasm through the general secretory (Sec) pathway, we observed no effect on the growth of E. coli, suggesting that the protein is unlikely to cleave peptidoglycan (S10A Fig).

Interestingly, some toxin-antitoxin systems have antitoxins with predicted metallopeptidase activity that exert their effect through degradation of the toxin partner [16,43,44]. We therefore reasoned that ShdB may similarly be a protease whose specific substrate is ShdA. To explore this further we first tested the ability for ShdA II and ShdB II to interact using the bacterial two hybrid assay (BTH) [45]. For this purpose, ShdA II and ShdB II were fused to the two fragments of the catalytic domain of Bordetella pertussis adenylate cyclase, carried by the pUT18 and pT25 plasmids [45]. For a detailed description of the bacterial two hybrid principle, see Materials and Methods. We observed an interaction between ShdA II and ShdB II when the T25 fragment of the adenylate cyclase was fused to ShdA II and the UT18 fragment to ShdB II (S10B Fig). The T25 domain is fused to the N-terminus of candidate proteins whereas the UT18 domain is at the C-terminus. Hence, we concluded that the N-terminus of ShdA II and the C-terminus of ShdB are involved in the interaction. Additionally, a self-interaction between ShdB proteins was also observed (S10B Fig). Furthermore, when we modified shdAII in the constructs that were used to assess anti-phage activity to produce a protein with a C-terminal hexahistidine epitope, a decrease in ShdA-His₆ protein levels was observed when co-produced with ShdB II (S10C Fig). Taken together our results are consistent with the hypothesis that ShdB II negatively affects the activity of ShdA II through direct interaction and possibly, degradation of ShdA.

Finally, we sought to determine whether a similar regulatory mechanism is found for Shield IV, a subtype that contains the ShdD and ShdE components. Neither of these proteins have any identifiable domains or predicted functions, and share no homology with ShdB. Using phage ϕAlma, we found that ShdD alone was, like ShdA IV, capable of providing protection (Fig 7A), indicating that two phage defense proteins are part of Shield IV. However, while the full Shield IV system was competent to mediate resistance to ϕAlma infection, none of the pairwise combinations conferred any significant protection, suggesting that unlike Shield II, the regulation and fine-tuning of Shield IV anti-phage activity is more complex (Fig 7B).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 7. ShdA-associated components in Shield subtypes regulate ShdA activity and/or provide additional defence modules.

(a). Growth in liquid LB supplemented with 0.2% L-arabinose of E. coli MG1655 carrying (VC, pBAD18) or plasmids directing the expression of the Shield IV system (Shield IV), ShdAIV only (ShdAIV), ShdD only (ShdD), ShdE only (ShdE). Strains were infected with MOI = 1 of ϕAlma as described in Methods. Point show mean +/− SEM (n = 3 biological replicates). (b) E. coli MG1655 carrying (VC, pBAD18) or plasmids directing the expression of the Shield IV system (Shield IV), the ShdA IV and ShdD pair (ShdAD), ShdA IV and ShdE pair (ShdAE), or the ShdA and ShdE pair (ShdDE) were grown in LB medium supplemented with 0.2% L-arabinose and infected with MOI = 1 of ϕAlma as indicated in Methods. Point show mean +/− SEM (n = 3 biological replicates).

https://doi.org/10.1371/journal.pgen.1010784.g007

Bacterial strains, plasmids and culture conditions

The strain P. aeruginosa NCTC 11442 (equivalent to ATCC 33350) was grown overnight at 37°C for chromosomal DNA extraction, whereas of P. aeruginosa PA14 csy3::lacZ was grown at 30°C. E. coli BL21 (DE3) and MG1655 were grown at 37°C on either solid media or liquid culture, shaking at 200 rpm. For liquid growth, Luria broth (LB) was used as the standard medium. For growth on solid media, LB was supplemented with 1.5% (w/v) or with 0.35% (w/v) agar for solid or soft agar, respectively. When required, LB was supplemented with ampicillin (Amp, 50 μg/mL), chloramphenicol (Chl, 25 μg/mL), gentamycin (Gm, 30 μg/mL), isopropyl-β-D-thiogalactopyranoside (IPTG, 0.5 mM), L-arabinose (0.2% w/v) or D-glucose (0.2% w/v). Strains and plasmids used in this study are listed in S11 Table.

Construction of P. aeruginosa mutants

Shield II or Shield II with the upstream WYL-containing domain were amplified from P. aeruginosa NCTC 11442 and cloned into pUC18-mini-Tn7T-Gm [38] using NEBuilder HiFi assembly (S11 and S12 Tables). For each construct the regions upstream and downstream were included to encompass the promoter and terminator of each locus. Four mL of P. aeruginosa PA14 csy3::lacZ (S11 and S12 Tables) were washed twice with 300 mM sucrose and resuspended in a final volume of 200 μl of 300 mM sucrose. Cells were electroporated following addition of 50 ng of each pUC18-mini-Tn7T construct and 50 ng of pTNS2 helper plasmid, as previously described [38].

Selection of mutants with a chromosomally integrated construct was performed on LB agar supplemented with 30 μg/ mL of gentamycin.

Measurement of bacterial growth

Overnight cultures were diluted 1:200 in fresh medium and grown until an OD_600nm of 0.3. Cultures were infected with phage at an MOI of 1, 0.1 and 0.01 in a final volume of 200 μL and aliquoted into a 96-well plate. The plate was incubated with continuous shaking in a TECAN infinite nano M+ and absorbance at 600 nm was measured every 20 min.

DNA manipulation and transformation

The chromosomal DNA of P. aeruginosa NCTC 11442 was extracted using the GenElute Bacterial Genomic DNA kit (Merck). Plasmid backbones and inserts for cloning were amplified from the purified chromosomal DNA using Q5 High-Fidelity DNA Polymerase (NEB). PCR products and plasmids were purified with Monarch DNA kits (NEB). Overlapping primers for amplification were designed using the NEBuilder assembly tool (https://nebuilderv1.neb.com/) (S12 Table). Plasmids and inserts were assembled using NEBuilder HiFi DNA Assembly (NEB), followed by incubation at 50°C for 20 min. Point mutations or small deletions in pGM34 and pGM71 (S11 and S12 Tables) were performed using the KLD enzyme mix (NEB)

Phage propagation and lysate preparation

Coliphage lysates were diluted in phage buffer (10 mM Tris–HCl pH 7.4, 10 mM MgSO4, 0.1% gelatin) and propagated in E. coli DH5α. Pseudomonas phages were propagated in P. aeruginosa PA01 in King’s B media.

For this purpose, neat lysates or their dilutions were added to 200 μL of E. coli DH5α or in P. aeruginosa PA01. The mixture was added to 5 mL of soft agar and poured onto LB agar plates. Plates were incubated at 37°C overnight. Following incubation, the top agar containing confluent phage plaques was scraped off and added to 3 mL of phage buffer and 500 μL of chloroform. Samples were vortexed for 2 minutes and then incubated at 4°C for 30 min. Samples were subsequently centrifuged at 4000 x g for 20 min and the supernatant collected and added to 100 μL of chloroform for storage.

Efficiency of plating measurement and fold protection calculation

To measure the efficiency of plating (EOP), 10 μL of phage lysate was added to 200 μL of an overnight culture of E. coli MG1655 carrying empty pBAD18 or pBAD18 encoding the Shield II system, ShdAII or ShdB II only. Five mL of soft agar was added to each culture and poured onto LB agar plates. As a control strain, plasmid-free MG1655 was used. EOP was measured as the number of PFU mL^-1 of a test strain divided by the number of PFU/mL of the control strain.

For P. aeruginosa, 10 μL of phage lysate was added to 100 μL of an overnight culture of P. aeruginosa csy3::lacZ or the same strain carrying chromosomal integrations of either Shield II or WYL-Shield II. Five mL of soft agar was added to each culture and poured onto LB agar plates and incubated at 30°C. P. aeruginosa csy3::lacZ was used as a control and EOP was measured as above.

For the calculation of fold protection, the EOP values calculated for ShdA II and empty vector when challenged with the phage panel indicated in Fig 3A were used. The ratio between ShdA II and empty vector EOP values was used to calculate fold protection.

Efficiency of centre of infection

E. coli MG1655 harboring pBAD18 or the same vector encoding Shield II, ShdA II or ShdB II was grown in LB with ampicillin and 0.2% arabinose to an OD_600nm of ~ 0.6 and infected with ϕSipho at MOI of 0.1 for ECOI measurement and MOI 2 for cell survival assessment. Infected cells were incubated for 15 min at 37°C shaking at 200 rpm to favour adsorption. Following incubation, cells were washed with ice-cold phage buffer twice and then serially diluted in phage buffer. For ECOI, 10 μL of the desired dilution was added to 200 μL of E. coli DH5α in 5 mL of soft agar. Infective centres were measured as the number of PFUs on each plate. ECOI measurement are reflective of the proportion of productive infections and generally represent a more absolute measure of phage viability compared to EOP, which is always calculated in relation to a reference strain.

Measurement of PFU/mL, CFU/mL and growth rate during phage infection

E. coli MG1655 harboring pBAD18 or the same vector encoding Shield II, ShdA II or ShdB II was grown in LB with ampicillin and 0.2% arabinose to an OD_600nm of ~ 0.6 and infected with ϕSipho at MOI of 0.1. An aliquot of each culture was collected at t = 0 hr, t = 3 hr, t = 6 hr, t = 9 hr, t = 12 hr and t = 24 hr and the OD_600nm was measured. For each timepoint, the cultures’ aliquots were also serially diluted and plated on LB agar plates to measure CFU/mL or plated onto E. coli DH5α top lawns to measure phage titre (PFU/mL).

One step growth curves

To calculate the burst size, E. coli MG1655 carrying pBAD18 or the same vector encoding Shield II, ShdA II or ShdB II were grown to an OD_600nm of ~0.6. ϕSipho was added to an MOI of 0.1. Cultures were incubated at 37°C at 200 rpm and duplicate samples were collected at t  =  5 min, t  =  10 min, t  =  15 min, t  =  20 min, t  =  30 min, t  =  45 min, t  =  60 min, t  =  75 min, t  =  90 min, t  =  105 min and t  =  120 min post infection. For each timepoint, one sample was immediately serially diluted and plated onto E. coli DH5α top lawns, accounting for free phages and phage-infected cells. For t  =  5 min, a second duplicate sample was collected and treated with chloroform before being plated as above, to represent the adsorption control. The PFU/mL from the adsorption control was subtracted from the average PFU/mL of the bottom plateau shown in S7A Fig (Beginning of infection cycle). Burst size was calculated by dividing this value into the average PFU/mL of the top plateau (S7A Fig) reached at the end of the infection cycle.

Cell free protein synthesis and Nuclease assay

In vitro synthesis of the RmuC domains of ShdA I, ShdA II, ShdA III and ShdA IV was performed using the PURExpress cell-free transcription/translation kit (NEB). Protein synthesis was performed with either 250 ng and 500 ng of DNA template. Synthesis was performed for 4 hr at 37°C according to the manufacturer’s recommendation. Following incubation, 10 mM MgCl₂ was added to the reaction and the final volume was adjusted to 10 μL. Ribosomes were removed through centrifugation for 60 min at 15,000 rpm at 4°C, through an Amicon Ultracel 0.5mL spin concentrator with a 100 KDa filter (Merck). The flowthrough was collected and the His-tagged PURExpress kit components were removed from the reaction following incubation with Ni-NTA agarose beads (Thermo) for 45 min at 4°C. Agarose beads were removed through centrifugation with Biorad micro Bio-spin columns at 15,000 g for 10 min at 4°C. As a control, the same reactions were performed in parallel with the dihydrofolate reductase (DHFR) control provided by the PURExpress kit.

To test for nuclease activity the in vitro synthesised ShdA proteins or DHFR control were incubated with 20 ng of phage ϕSipho, E. coli MG1655 chromosome or plasmid pSG483 DNA, followed by agarose gel electrophoresis and staining with GelRed (Cambridge Bioscience).

Subcellular fractionation

Overnight cultures of MG1655 carrying ShdA II with a C-terminal His₆ tag were diluted in 25 mL LB containing 0.2% L-arabinose and grown for 5 hrs. Following growth, cells were centrifuged at 4000 x g for 10 min at 4°C. Cells were resuspended in 1 mL of Tris HCl pH 8 and lysed by sonication. Cellular debris were removed by centrifugation at 14,000 x g for 10 min at 4°C. The supernatant was collected and subjected to ultracentrifugation (200,000 x g, 30 min, 4°C). The supernatant was collected, representing the cytoplasmic fraction and the pellet resuspended in 1 mL of Tris HCl pH 8, representing the total membrane fraction. His₆ tagged proteins and GroEL (used as cytoplasmic control) were detected as above. TatA was used as membrane control protein and detected with anti-TatA serum as previously described [55].

Bacterial two hybrid assay

ShdA II and ShdB II were fused to the two fragments of adenylate cyclase encoded on plasmids pT25 and pUT18 (SI Appendix, S12 Table) [45]. Plasmids were introduced into E. coli BTH101 and selected on MacConkey medium (Difco) supplemented with maltose (1%), Amp, 50 μg/mL and Chl 25 μg/mL. Plates were incubated for 48 hrs at 30°C and positive interactions were identified as dark red colonies. When an interaction occurs between two proteins, the two adenylate cyclase fragments are brought in proximity, producing the cAMP signal. This in turns activates the operon involved in maltose catabolism and produces red colonies on MacConkey medium.

Cell free protein synthesis and Nuclease assay

In vitro synthesis of the RmuC domains of ShdA I, ShdA II, ShdA III and ShdA IV was performed using the PURExpress cell-free transcription/translation kit (NEB). Protein synthesis was performed with either 250 ng and 500 ng of DNA template. Synthesis was performed for 4 hr at 37°C according to the manufacturer’s recommendation. Following incubation, 10 mM MgCl₂ was added to the reaction and the final volume was adjusted to 10 μL. Ribosomes were removed through centrifugation for 60 min at 15,000 rpm at 4°C, through an Amicon Ultracel 0.5mL spin concentrator with a 100 KDa filter (Merck). The flowthrough was collected and the His-tagged PURExpress kit components were removed from the reaction following incubation with Ni-NTA agarose beads (Thermo) for 45 min at 4°C. Agarose beads were removed through centrifugation with Biorad micro Bio-spin columns at 15,000 g for 10 min at 4°C. As a control, the same reactions were performed in parallel with the dihydrofolate reductase (DHFR) control provided by the PURExpress kit.

To test for nuclease activity the in vitro synthesised ShdA proteins or DHFR control were incubated with 20 ng of phage ϕSipho, E. coli MG1655 chromosome or plasmid pSG483 DNA, followed by agarose gel electrophoresis and staining with GelRed (Cambridge Bioscience).

Fluorescence microscopy and quantification of DAPI fluorescence

Overnight cultures (5 ml) were diluted into 25 mL LB containing 0.2% arabinose and 50 μg/mL ampicillin and grown for 2 hr. 200 μL of each culture were collected at timepoints t = 0 hr, t = 1 hr and t = 2 hr and stained with 4′,6-diamidino-2-phenylindole (DAPI) at a final concentration of 5 μg /mL and F M5-95(Thermo) at 200 μg /mL. Staining was carried out at 37°C for 15 min and then 2 μL of each culture was transferred on a microscope slide with a pad of 1% UltraPure agarose (Invitrogen) in H₂O. Imaging was carried out on Nikon Eclipse Ti equipped with CoolLED pE-300^white light source, Nikon Plan Apo 100×/1.40 NA Oil Ph3 objective, and Photometrics Prime sCMOS, and Chroma 49008 (Ex 560/40, Dm 585, Em 630/75) filter set for FM 5–95 and Chroma 49000 (Ex 350/50, DM 400, EM 460/50) filter set for DAPI. The images were captured using Metamorph 7.7 (Molecular Devices) and analysed using Fiji [56].

Quantification of DAPI fluorescence was performed using Fiji. In brief, individual cells were identified from thresholded phase contrast images and converted to regions of interest (ROI). The ROI areas associated with the cell periphery were defined as a 3-pixel wide band extending towards the cell interior from the cell boundary which, in turn, was determined by a steep change in phase contrast. These peripheral cell areas represent approx. 200 nm wide sections along the shorter cell axis, flanking the approx. 800 nm wide cell interior region. These whole cell and cell periphery -ROIs, and background-subtracted fluorescence images were then used to quantify the integrated density of DAPI fluorescence signals for the whole cell, and for the respective cell periphery. At last, the ratio of DAPI fluorescence signals between the cell periphery and the whole cell, the integrated density (sum of pixel values) of the whole cell, and the standard deviation thereof were measured, followed by calculation of the variance of cell DAPI fluorescence. The ratio of cell periphery and whole cell fluorescence, the whole cell integrated density itself, and variance thereof were represented as swarm plots, and further analysed by one-way ANOVA with Dunnett’s multiple comparison tests. The potential correlation between the integrated density and variance values for each strain was analysed visually by means of a scatter plot, together with a linear regression between the data points and calculation of the coefficient of determination (R²). The linear regression analysis was carried out with GraphPAd Prism 9.

ShdA II detection

ShdA II with a C-terminal His₆ tag was expressed from the arabinose inducible plasmid, pBAD18 alone or in presence of ShdB II. Induction was performed with 0.2% L-arabinose for 5 hrs. Following induction, cells were harvested by centrifugation at 14,000 x g for 10 min at 4°C and resuspended in Laemmli buffer. His₆ tagged proteins were detected with anti-His₆ primary antibody (Pierce; 1:6,000) and anti-mouse secondary antibody (Biorad; 1:10,000). GroEL was detected with anti-GroEL primary antibody (Pierce; 1:10,000) and anti-rabbit secondary antibody (Biorad; 1:20,000).

Homology searches and gene neighbourhood analysis

ShdA alignments were generated using MUSCLE (v3.8.1551) [57]. Alignments were used to generate Hidden Markov Models with the HMMER suite (v 3.3.2) [58]. The obtained models were used to query a local Refseq database of Pseudomonas spp or all bacterial genomes.The cutoff value was set to a bit score of 30 over the overall sequence/profile comparison. Efetch from the entrez utilities [59] was employed to retrieve an identical protein group report (IPG) for each hit protein obtained from the HMMER searches. Neighboring genes to Shield system were identified using FlaGs version 1.2.7 [35] with a non-redundant ShdA protein set as query. Neighboring genes were analysed using the Defense-finder tool to determine their association to known anti-phage systems.

Annotation of anti-phage systems

Flanking genes of Shield systems were retrieved using the FlaGs tool [35]. Clustered flanking genes retrieved by FlaGs were subjected to searches against a local PFAM database [60]. Additionally, assembly IDs of genomes encoding Shield homologues, retrieved using Efetch, were used to download genome proteomes. Proteomes were then used as an input to predict anti-phage systems with Defense-finder [34]. Genome neighbourhoods were scanned and anti-phage system predictions were manually curated by comparing Defense-finder results with PFAM predictions [60]. For multi-gene loci where PFAM and Defense-finder predictions were discordant, the prediction that best matched their operon organisation was chosen.

Protein function prediction

Where possible, protein function predictions were performed using a local PFAM database [60]. Protein structures were predicted using Alphafold and the Dali server (50, 51). Presence of signal peptides and transmembrane domains were predicted using DeepTMHMM and SignalP 6.0 [61,62]. [60]

Phylogenetic analysis

The alignment of ShdA proteins was used to build a maximum likelihood phylogenetic tree with IQTREE (v 2.1.4) [63] with 1000 ultrafast bootstraps. Trees were plotted and annotated in iTOL [64].