Discovery of coordinately regulated pathways that provide innate protection against interbacterial antagonism.

Bacterial survival is fraught with antagonism, including that deriving from viruses and competing bacterial cells. It is now appreciated that bacteria mount complex antiviral responses; however, whether a coordinated defense against bacterial threats is undertaken is not well understood. Previously, we showed that Pseudomonas aeruginosa possess a danger-sensing pathway that is a critical fitness determinant during competition against other bacteria. Here, we conducted genome-wide screens in P. aeruginosa that reveal three conserved and widespread interbacterial antagonism resistance clusters (arc1-3). We find that although arc1-3 are coordinately activated by the Gac/Rsm danger-sensing system, they function independently and provide idiosyncratic defense capabilities, distinguishing them from general stress response pathways. Our findings demonstrate that Arc3 family proteins provide specific protection against phospholipase toxins by preventing the accumulation of lysophospholipids in a manner distinct from previously characterized membrane repair systems. These findings liken the response of P. aeruginosa to bacterial threats to that of eukaryotic innate immunity, wherein threat detection leads to the activation of specialized defense systems.

Introduction

Antagonism from other organisms is a threat faced nearly universally by bacteria living in mixed populations, yet we are only beginning to understand the mechanisms employed in defense against these assaults (Hersch et al., 2020a; Peterson et al., 2020; Robitaille et al., 2021). One defense mechanism that is common in the gut microbiome and potentially other habitats is the production of immunity proteins that grant protection against specific toxins delivered by the type VI secretion system (T6SS) (Ross et al., 2019). These ‘orphan’ immunity proteins share homology with and are likely evolved from cognate immunity proteins that protect bacteria from undergoing self-intoxication, which is inherent to the indiscriminate nature of the T6SS delivery mechanism. Core cellular structures have also been implicated in promoting survival during interbacterial antagonism. Extracellular polysaccharide capsules protect Vibrio cholerae and Escherichia coli from T6SS-based attack of other species (Hersch et al., 2020b; Toska et al., 2018), and in Acinetobacter baumannii, lysine modification at peptidoglycan crosslinks was suggested to render the cell wall resistant to degradation by amidase toxins (Le et al., 2020). Finally, characterized stress response pathways also appear to contribute to antagonism defense. Indeed, it has been suggested that bacterial stress responses evolved in part as a means of detecting and responding to antagonistic competitors (Cornforth and Foster, 2013; Lories et al., 2020). A candidate-based approach applied to E. coli and P. aeruginosa found that stress response genes involved in envelope integrity and acid stress, among others, can contribute to resistance against a V. cholerae phospholipase toxin (Kamal et al., 2020). While the P. aeruginosa genes identified impacted survival when the toxin was produced heterologously within the organism, they did not influence the fitness of the bacterium during interbacterial competition with V. cholerae.

We previously discovered that P. aeruginosa mounts an effective but largely undefined defense upon exposure to an antagonistic competitor, which we named the response to antagonism (PARA) (LeRoux et al., 2015a). The response is activated when a subpopulation of P. aeruginosa cells are lysed, releasing intracellular contents that trigger the response in neighboring survivors (LeRoux et al., 2015b). PARA is coordinated by the Gac/Rsm global regulatory pathway, which post-transcriptionally controls the expression of nearly 400 genes (Goodman et al., 2004; Lapouge et al., 2008). This regulon includes the Hcp secretion island I-encoded type VI secretion system (H1-T6SS), the activity of which enables P. aeruginosa to kill or disable competing bacteria through the delivery of a cocktail of toxic effector proteins (Hood et al., 2010). While an important component of PARA, data suggest that the H1-T6SS is just one feature of the response; P. aeruginosa strains lacking functionality of the two-component system required to initiate Gac/Rsm signaling are severely crippled in interbacterial antagonism defense relative to those lacking only the H1-T6SS (LeRoux et al., 2015a). The majority of the genes under Gac/Rsm control encode proteins of unknown function, and we hypothesized that these could represent uncharacterized, novel mechanisms by which P. aeruginosa defends against interbacterial antagonism.

Results

Discovery of coordinately regulated P. aeruginosa interbacterial toxin defense pathways

Prior work by our laboratory has shown that Burkholderia thailandensis (B. thai) fiercely antagonizes P. aeruginosa using its antibacterial T6SS (LeRoux et al., 2015a). Thus, to identify its interbacterial defense factors, we subjected a transposon library of P. aeruginosa to antagonism by B. thai or a derivative lacking antibacterial T6SS activity (ΔT6S) and used high-throughput sequencing to measure gene-level fitness changes. Duplicate screening of this library revealed 34 genes critical for the survival of P. aeruginosa specifically while undergoing antagonism by B. thai (Figure 1A, Figure 1—figure supplement 1A and B , Figure 1—source data 1). Strikingly, 25 of these belong to the Gac/Rsm regulon and three additional hits included genes directly involved in the Gac/Rsm signaling system (Goodman et al., 2004; Lapouge et al., 2008). Indeed, disruption of genes encoding GacS and GacA, the central sensor kinase and response regulator required for activation of the pathway, respectively, elicited the strongest interbacterial defense defect among all genes in the P. aeruginosa genome (Figure 1A and B). On the contrary, inactivating insertions in retS, encoding a hybrid sensor kinase that negatively regulates GacS, increased interbacterial competitiveness.

Figure 1.

Multiple pathways under Gac/Rsm control contribute to P. aeruginosa defense against antagonism.

(A) Transposon library sequencing-based comparison of the fitness contribution of individual P. aeruginosa genes during growth competition with wild-type B. thai versus B. thai ∆T6S. Genes under Gac/Rsm control (blue) and those encoding core Gac/Rsm regulatory factors (purple, labeled) are indicated. (B) Overview of the Gac/Rsm pathway (SD, Shine–Dalgarno). The rsmY and rsmZ genes encode small RNA molecules that sequester the translational regulator RsmA. Hybrid sensor kinase PA1611 is here renamed RetS-regulating sensor kinase A (RskA). (C) Left: P. aeruginosa gene clusters hit (greater than threefold in replicate screens; Figure 1—source data 1) in this study. Numbers below genes indicate transposon insertion ratio (B. thai ∆T6S/wild-type) for screen in (A) followed by the replicate screen (Figure 1—figure supplement 1); light toned genes were hit in only one replicate. Asterisk indicates genes for which insertion was undetected in libraries obtained from B. thai wild-type competition. Right: zoom-in of boxed region of (A) with genes colored corresponding to clusters at left. (D, E) Recovery of P. aeruginosa cells with the indicated genotypes following growth competition against B. thai wild-type (D) or ∆T6S (E). For interbacterial competition assays, the mean ± SD of biological duplicates and associated technical duplicates is shown: *p<0.05 using unpaired two-tailed Student’s t-test.

(A) Growth of wild-type P. aeruginosa in co-culture with the indicated B. thai strains. (B) Transposon library sequencing-based comparison of the fitness contribution of individual P. aeruginosa genes during growth competition with wild-type B. thai versus B. thai ∆T6S. Genes under Gac/Rsm control (blue) and those encoding core Gac/Rsm regulatory factors (purple, labeled) are indicated. This experiment is a biological replicate of that in Figure 1A. (C) P. aeruginosa gene clusters hit (Figure 1—source data 1) in this study. (D) Zoom-in of boxed region of (B) with genes colored corresponding to clusters at left. The interbacterial competition assay represents two biological replicates with two associated technical replicates.

(A) Schematic depicting the arc1 gene cluster of P. aeruginosa. Predicted enzymatic activity of Arc1A, B, D, F is indicated (right) and placed onto a phospholipid biosynthesis pathway (bottom). (B) Depiction of arc2 genes from the indicated species. Genes encoding the predicted MoxR-like ATPase and von Willebrand factor (VWF) domain protein are indicated below the respective genes. Orthologs are colored to highlight synteny. (C) Depiction of arc3 genes from the indicated species. S. dynata exemplifies an instance of translational fusion of arc3A,B. Domains of unknown function constituting the two proteins are indicated below.

(A) Recovery of P. aeruginosa strains bearing the indicated single or double deletions within arc1-3 following growth competition against wild-type B. thai. (B) Recovery of P. aeruginosa cells with the indicated genotypes following growth competition against wild-type B. thai. The mean ± SD of biological duplicates and associated technical duplicates is shown: n.s. corresponds to p>0.05 by unpaired two-tailed Student’s t-test. The fold difference in growth yield between mutants and the parental strain is indicated above.

Figure 1—figure supplement 1.

Multiple pathways under Gac/Rsm control contribute to P. aeruginosa defense against antagonism.

(A) Growth of wild-type P. aeruginosa in co-culture with the indicated B. thai strains. (B) Transposon library sequencing-based comparison of the fitness contribution of individual P. aeruginosa genes during growth competition with wild-type B. thai versus B. thai ∆T6S. Genes under Gac/Rsm control (blue) and those encoding core Gac/Rsm regulatory factors (purple, labeled) are indicated. This experiment is a biological replicate of that in Figure 1A. (C) P. aeruginosa gene clusters hit (Figure 1—source data 1) in this study. (D) Zoom-in of boxed region of (B) with genes colored corresponding to clusters at left. The interbacterial competition assay represents two biological replicates with two associated technical replicates.

Our screens identified previously uncharacterized genes belonging to three operons under Gac/Rsm control (Goodman et al., 2004), several genes within two operons of the HSI-I-encoded T6SS (H1-T6SS) (Hood et al., 2010), and two genes within an operon not subject to Gac/Rsm regulation (Figure 1C, Figure 1—figure supplement 1C, Figure 1—source data 1). Given that false positives are less common among multiple gene hits within an operon, we focused on these genes for subsequent validation in pairwise competition assays. Strains bearing in-frame deletions of the genes most strongly hit in each operon showed substantially reduced fitness in competition with wild-type B. thai, and in each instance this could be partially restored by genetic complementation (Figure 1D). Further consistent with our screening results, each of the single-gene deletion mutants grew similarly to the wild-type in competition with B. thai unable to antagonize P. aeruginosa (B. thai ΔT6S) (Figure 1E).

Arc1-3 operate independently and defend in a manner distinct from general stress responses

The P. aeruginosa Gac/Rsm pathway has been implicated in a process we termed danger sensing (LeRoux et al., 2015a; LeRoux et al., 2015b). We previously showed that the lysis of neighboring kin cells triggers the pathway, leading to T6SS activation. However, the Gac/Rsm pathway post-transcriptionally activates hundreds of genes whose function has largely remained cryptic (Figure 1B). Bioinformatic analyses suggested that each of the uncharacterized Gac/Rsm-regulated operons hit in our screen likely constitute a discrete pathway. Five of the six genes within the PA2536-41 cluster are predicted to encode proteins related to phospholipid biosynthesis or metabolism, including four genes encoding products with significant homology to enzymes required for phosphatidylglycerol synthesis (Figure 1—figure supplement 2A). Orthologs of core bacterial phospholipid biosynthetic machinery are encoded elsewhere in the genome, indicating likely specialization of this putative pathway. The PA4317-23 genes co-occur across many species within the Xanthomonadaceae and Pseudomonadaceae (Figure 1—figure supplement 2B). While none of the proteins encoded by this gene cluster have been characterized, they include a predicted MoxR-like AAA+-ATPase and a protein with a von Willebrand factor (VWF) domain. Related protein pairs co-occur widely and are known to cooperate in a chaperone-like function that releases protein inhibition or stimulates metal cofactor insertion (Snider and Houry, 2006; Tsai et al., 2020). Finally, PA5113 and PA5114 also co-occur widely, and several organisms encode single polypeptides that represent a fusion of the two proteins (Figure 1—figure supplement 2C).

Figure 1—figure supplement 2.

Antagonism resistance clusters possess functionally related genes and are in diverse bacteria.

(A) Schematic depicting the arc1 gene cluster of P. aeruginosa. Predicted enzymatic activity of Arc1A, B, D, F is indicated (right) and placed onto a phospholipid biosynthesis pathway (bottom). (B) Depiction of arc2 genes from the indicated species. Genes encoding the predicted MoxR-like ATPase and von Willebrand factor (VWF) domain protein are indicated below the respective genes. Orthologs are colored to highlight synteny. (C) Depiction of arc3 genes from the indicated species. S. dynata exemplifies an instance of translational fusion of arc3A,B. Domains of unknown function constituting the two proteins are indicated below.

Next, we sought to obtain experimental support for our hypothesis that the Gac/Rsm-regulated operons hit in our screen constitute independent functional units. First, we compared the fitness of our single-deletion strains to those bearing in-frame deletions in the two genes most strongly hit within each operon. In each operon, the second deletion did not impact competitiveness with B. thai, suggesting that the pairs of genes encode components of the same pathway (Figure 1—figure supplement 3A). On the contrary, additive effects on interbacterial competitiveness were observed when genes from each of the three operons were inactivated in a single strain. Indeed, the fitness deficiency of this triple mutant strain approaches that of the ∆gacS strain, which lacks Gac/Rsm function entirely (Figure 1—figure supplement 3B). Together, these data reveal that the Gac/Rsm pathway and discrete functional units under its control are critical for P. aeruginosa survival when antagonized by the T6SS of B. thai. Based on these data, we named the genes with the three Gac/Rsm-regulated operons hit in our screen arc1A-F (antagonism resistance cluster 1, PA2541-36), arc2A-G (PA4317-23), and arc3A,B (PA5114-13) (Figure 1C).

Figure 1—figure supplement 3.

Interbacterial competition assays measure the contribution of arc1-3 to P. aeruginosa fitness during antagonism by B. thai.

(A) Recovery of P. aeruginosa strains bearing the indicated single or double deletions within arc1-3 following growth competition against wild-type B. thai. (B) Recovery of P. aeruginosa cells with the indicated genotypes following growth competition against wild-type B. thai. The mean ± SD of biological duplicates and associated technical duplicates is shown: n.s. corresponds to p>0.05 by unpaired two-tailed Student’s t-test. The fold difference in growth yield between mutants and the parental strain is indicated above.

As a first step toward understanding how Arc1-3 influence the competitiveness of P. aeruginosa, we asked whether they serve defensive or offensive roles. In our initial screen, B. thai dramatically outnumbered P. aeruginosa (~50:1); therefore, the diminished P. aeruginosa viability observed under these conditions suggested each pathway can function defensively. However, these conditions do not exclusively isolate the defensive contribution of the pathways. For example, we observed that killing competitor cells via the H1-T6SS indirectly contributes to P. aeruginosa viability under these conditions, likely by expanding the available growth niche. To probe the capacity of Arc1-3 to antagonize B. thai, we initiated co-culture growth competition assays with an excess of P. aeruginosa. We previously demonstrated that in co-cultures initially dominated by P. aeruginosa, antagonism by B. thai has a negligible impact on P. aeruginosa growth (Schwarz et al., 2010). Under these conditions, inactivating Arc2 and Arc3 had no impact on B. thai survival. Arc1 inactivation led to an increase in B. thai survival; however, this effect was modest relative to its impact on P. aeruginosa defense (Figure 2A). In contrast, inactivation of the H1-T6SS increased B. thai survival by approximately four orders of magnitude, eclipsing its impact on P. aeruginosa defense by ~100-fold. Together, these finding indicate that the contributions of Arc1-3 to P. aeruginosa competitiveness during antagonism arise predominately from defensive mechanisms.

Figure 2.

Arc pathways can provide toxin-specific antagonism defense that eclipses that of other cellular factors.

(A) Recovery of B. thai cells following growth competition against a relative abundance of P. aeruginosa containing the indicated gene deletions, colored according to Figure 1C. (B) Relative survival of P. aeruginosa Arc1-3-inactivated mutants following growth in competition with an excess of the indicated B. thai strains (mutant CFU/parental CFU × 100 for each B. thai competitor). The loss of ColA or Tle3 activity in B. thai increases the relative survival of P. aeruginosa lacking Arc2 or Arc3 activity, respectively. (C, D) Transposon library sequencing-based comparison of the fitness contribution of individual P. aeruginosa genes during growth competition with wild-type B. thai versus B. thai lacking Tle3 (C) or ColA (D) activity, colored according to Figure 1C. Values in parentheses correspond to the number of genes hit (greater than threefold change in transposon insertion frequency) within each cluster compared to the number within each cluster hit in the initial screens (B. thai wild-type versus B. thai ∆T6S) followed by the average transposon insertion frequency ratio (B. thai mutant/B. thai wild-type) of the genes in the depicted screen. (E) Fitness contribution of each P. aeruginosa Arc pathway and the H1-T6SS to defense against Tle3 versus ColA. For interbacterial competition assays, the mean ± SD of biological duplicates and two technical replicates is shown: *p<0.05 using unpaired two-tailed Student’s t-test.

Survival of the indicated P. aeruginosa strains during exposure to hydrogen peroxide (50 mM, 30 min) (A), high salinity (3 M NaCl, 20 hr) (B), high temperature (55°C, 30 min) (C), or detergent (0.5% [w/v] SDS) (D). The mean ± SD of a representative biological replicate with three technical replicates is shown.

Anti-VSV-G immunoblot analysis was used to probe the levels of the indicated VSV–G-tagged (–V) Arc proteins encoded at their respective native genomic loci in P. aeruginosa wild-type, ∆retS, or ∆gacS backgrounds. The asterisk denotes a nonspecific VSV-G antibody-reactive band. The arrow highlights the position of Arc2G–V. Blot images are representative of at least two biological replicates. See Figure 2—figure supplement 2—source data 1–6.

(A) Schematic depicting the tle1/tli1 (top) and tle3/tli3 (bottom) loci in the genome B. thai E264 strain. Predicted catalytic serine residues mutagenized in this study are indicated by red asterisks. (B) Recovery of B. thai strain sensitized to Tle3 intoxication (∆tle3∆tli3) following growth competition against B. thai wild-type or a strain in which both orthologs of Tle3 contain an inactivating point mutation. The mean ± SD of two biological replicates with two technical replicates is shown.

Our results suggested that the Gac/Rsm system functions analogously to innate immune sensors of eukaryotic organisms, wherein danger sensing activates the expression of independently functioning downstream effectors that defend the cell against specific threats (Paludan et al., 2021). By extension, this response should be distinct from that provoked by general stressors. To test this, we measured the contribution of Arc1-3 to P. aeruginosa survival in conditions known to induce general stress response systems, including heat shock, exposure to H2O2, and propagation in high-salinity media or media containing detergent (Jørgensen et al., 1999; Munguia et al., 2017). Unlike previously established stress-intolerant control strains (∆rpoS and ∆vacJ), the survival of strains lacking Arc1-3 function was equivalent to wild-type P. aeruginosa in these assays, suggesting that their function is distinct from those involved in general stress resistance (Figure 2—figure supplement 1). This is also in line with the regulatory profile of Arc1-3; immunoblotting showed that basal expression of each is exceedingly low, yet highly inducible through stimulation of the antagonism-responsive Gac/Rsm pathway (Figure 2—figure supplement 2).

Figure 2—figure supplement 1.

Arc1-3 do not contribute to the survival of P. aeruginosa exposed to common environmental stresses.

Survival of the indicated P. aeruginosa strains during exposure to hydrogen peroxide (50 mM, 30 min) (A), high salinity (3 M NaCl, 20 hr) (B), high temperature (55°C, 30 min) (C), or detergent (0.5% [w/v] SDS) (D). The mean ± SD of a representative biological replicate with three technical replicates is shown.

Figure 2—figure supplement 2.

Arc1-3 are subject to tight regulation by the Gac/Rsm signaling pathway.

Anti-VSV-G immunoblot analysis was used to probe the levels of the indicated VSV–G-tagged (–V) Arc proteins encoded at their respective native genomic loci in P. aeruginosa wild-type, ∆retS, or ∆gacS backgrounds. The asterisk denotes a nonspecific VSV-G antibody-reactive band. The arrow highlights the position of Arc2G–V. Blot images are representative of at least two biological replicates. See Figure 2—figure supplement 2—source data 1–6.

Arc1-3 provide defense against distinct threats

The specific involvement of Arc1-3 in defense against antagonism led us to hypothesize that the pathways could grant protection against mechanistically distinct threats. Notably, B. thai delivers a cocktail of toxins to target cells including two predicted phospholipases (Tle1 and Tle3) (Russell et al., 2013), a peptidoglycan-degrading amidase (Tae2) and a colicin-like toxin predicted to form inner membrane pores (ColA) (Russell et al., 2012; Salomon et al., 2014). To determine the role of Arc1-3 in defense against insults caused by toxins with biochemically diverse modes of action, we subjected strains with individual Arc pathways inactivated to pairwise competition against B. thai strains lacking single toxins in its arsenal. For Tle1 and Tle3, which are duplicated in the B. thai genome and required for basal T6SS function, strains bearing mutated codons corresponding to predicted catalytic residues of the proteins were utilized (Figure 2—figure supplement 3). Strikingly, we found that Tle3 inactivation abrogated the antagonism defense defect of P. aeruginosa lacking Arc3 function (Figure 2B), whereas its defense defect was maintained in competition experiments against other toxin-inactivated strains of B. thai. In contrast, the survival of P. aeruginosa lacking Arc2 function was significantly restored by deletion of colA in B. thai, while none of the single-effector mutations within B. thai impacted the defense defect of P. aeruginosa lacking Arc1 function. These results support the hypothesis that Arc pathways can afford protection in a manner specific to the damage arising from mechanistically distinct threats (Arc2, Arc3) or play a more general role in defense (Arc1). However, B. thai delivers toxins beyond those we inactivated in this study; therefore, we cannot rule out that Arc1 – like Arc2 and Arc3 – provides effector-specific defense.

Figure 2—figure supplement 3.

The loci encoding Tle1 and Tle3 are duplicated in B. thai, and Tle3 induces potent self-intoxication.

(A) Schematic depicting the tle1/tli1 (top) and tle3/tli3 (bottom) loci in the genome B. thai E264 strain. Predicted catalytic serine residues mutagenized in this study are indicated by red asterisks. (B) Recovery of B. thai strain sensitized to Tle3 intoxication (∆tle3∆tli3) following growth competition against B. thai wild-type or a strain in which both orthologs of Tle3 contain an inactivating point mutation. The mean ± SD of two biological replicates with two technical replicates is shown.

We found that Arc2 and Arc3 are major defense factors against insults caused by ColA and Tle3, respectively, suggesting that these toxins may damage cells by a mechanism not efficiently countered by other cellular pathways. To evaluate the relative defensive contribution of Arc2 and Arc3 genome-wide, we performed additional transposon-based screening in which a mutant library of P. aeruginosa was grown in competition with B. thai strains lacking ColA or Tle3 activity. Remarkably, this experiment defined arc3A as the gene most critical for P. aeruginosa survival during intoxication by Tle3 (Figure 2C, Figure 2—source data 1). Insertions in the second Arc3 gene, arc3B, also crippled P. aeruginosa defense against Tle3, though to a lesser degree. Arc2 genes were dispensable for Tle3 defense, and only one gene within Arc1 contributed to Tle3 defense beyond the threefold insertion frequency ratio cutoff we employed. Inactivation of ColA had a relative minor impact on the fitness of most P. aeruginosa transposon mutants (Figure 2D, Figure 2—source data 2). However, four of the six Arc2 genes hit in our initial screen provided defense against ColA that exceeded our threefold cutoff. Additionally, one Arc1 gene met these criteria, whereas Arc3 genes were wholly dispensable for defense against ColA. These results support the model that Arc3 provides specific defense against Tle3, Arc2 provides a degree of specific protection against ColA, and Arc1 serves a broader role in antagonism defense. A comparison of the fitness contributions of each Arc pathway and the H1-T6SS to P. aeruginosa undergoing intoxication by Tle3, ColA or Tle1 highlights the degree of their specificity (Figure 2E).

Arc3 is a broadly conserved pathway that prevents lysophospholipid accumulation

The highly specific defense afforded by Arc3 against the predicted phospholipase Tle3 prompted us to further interrogate its function. Arc3A is a large protein containing 35 predicted transmembrane domains and lacking residues strongly indicative of enzymatic activity (Figure 3A). With predicted outer-membrane-anchored N-terminal lipidation and a C-terminal inner-membrane transmembrane helix, Arc3A is expected to span the periplasm. Neither Arc3A nor Arc3B share significant homology with characterized proteins; however, apparent orthologs of each are encoded by neighboring open-reading frames across an exceptionally wide distribution of bacteria (Figure 3B). Arc3A-related proteins are composed of domain of unknown function 2339 (DUF2339), possess large, yet varying numbers of transmembrane segments, and are found in bacteria belonging to most Gram-negative and Gram-positive phyla. Consistent with the predicted N-terminal localization of Arc3B to the outer membrane, its related proteins are restricted to Gram-negative phyla. Co-immunoprecipitation studies using epitope-tagged variants of Arc3A and Arc3B encoded at their native chromosomal loci provided evidence that the two proteins stably associate (Appendix 1—table 1). The arc3A and arc3B genes abut a third gene encoded in the same orientation, estA, and there is evidence that transcripts containing all three genes are generated by P. aeruginosa (Gebhardt et al., 2020). Although estA encodes an esterase (Wilhelm et al., 1999), which could have relevance for phospholipase defense, estA was not detected in our co-immunoprecipitation analyses, it did not exhibit differential insertion frequency in our screens, and in-frame deletion of estA did not impact P. aeruginosa survival during intoxication by Tle3 (Figure 1—source data 1, Figure 2—source data 1 and 2, Figure 3—figure supplement 1A).

Figure 3.

Arc3B is a predicted massive polytopic membrane protein with functionally complementing family members that occur widely in Gram-negative and -positive bacterial phyla.

(A) Schematic depiction of Arc3A and Arc3B based on bioinformatic predictions. (B) Phylogeny of bacterial phyla with representative arc genes depicted and colored as in (A). Membrane-associated regions (grayed) and the number of predicted transmembrane segments are indicated. (C) Recovery of P. aeruginosa strains bearing the indicated mutations and containing a control chromosomal insertion (–) or insertion constructs expressing Arc3B proteins derived from assorted bacteria (P. aer, P. aeruginosa; P. pro, P. protegens; B. cen, B. cenocepacia; P. syr, P. syringae; C. rod, Citrobacter rodentium) following growth competition against B. thai wild-type. (D) Recovery of the indicated P. protegens strains following growth competition against B. thai wild-type or a strain lacking Tle3 activity. For interbacterial competition assays, the mean ± SD of biological duplicates and two or three associated technical replicates is shown: *p<0.05 using unpaired two-tailed Student’s t-test.

Recovery of the indicated P. aeruginosa strains following growth in competition with B. thai wild-type or a strain lacking Tle3 activity (tle3). Means ± SD of biological duplicates with three technical replicates s shown.

Appendix 1—table 1.

Mass spectrometric analysis of P. aeruginosa Arc3A and Arc3B immunoprecipitation samples.

Locus ID[*]	Peptide counts of sample (%)[†]	Fold change (VSV-G/Ctrl ‡)
Arc3B-VSV-G sample
PA5113 (Arc3B)	5.86	> 200
PA4385	1.62	2.29
PA5016	1.45	0.82
PA0090	1.17	0.72
PA3729	1.12	1.30
PA4761	1.12	0.71
PA5114 (Arc3A)	1.12	N.D. [§]
PA5239	1.06	1.10
PA3861	1.00	1.10
PA4269	1.00	1.10
PA1092	0.95	0.78
PA0141	0.89	0.71
PA0963	0.84	1.27
PA2976	0.84	0.83
PA4260	0.84	0.69
PA4265	0.84	0.61
PA4270	0.84	1.04
PA1803	0.78	0.74
PA2151	0.78	0.77
PA3950	0.78	1.19
PA3001	0.73	1.10
PA3656	0.73	0.80
PA5173	0.73	0.76
PA5554	0.73	1.20
PA5556	0.73	2.05
Arc3A-VSV-G sample
PA5554	2.92	4.81
PA5113 (Arc3B)	2.22	> 200
PA5114 (Arc3A)	1.44	N.D.
PA4751	1.30	12.87
PA5556	1.26	3.55
PA3729	1.10	1.28
PA5016	1.10	0.62
PA4429	0.99	N.D.
PA4942	0.96	6.36
PA2493	0.92	2.60
PA3160	0.90	N.D.
PA4246	0.87	0.79
PA0090	0.85	0.53
PA1552	0.83	N.D.
PA0077	0.81	5.33
PA2494	0.81	ND
PA2976	0.72	0.71
PA0659	0.70	N.D.
PA4761	0.70	0.44
PA4941	0.65	2.57
PA4265	0.63	0.46
PA3656	0.61	0.67
PA3794	0.61	N.D.
PA2151	0.56	0.55
PA3821	0.56	N.D.

Proteins containing at least two peptides identified in a given immunoprecipitation sample are included. Arc3A and Arc3B are bolded.

Value corresponds to the abundance (peptide counts) of the protein within the total immunoprecipitation sample. Only the 25 most abundant proteins in each sample are shown.

Immunoprecipitation from a P. aeruginosa strain lacking a VSV-G-tagged protein served as the control sample (Ctrl).

The protein was not detected in the control sample.

Figure 3—figure supplement 1.

EstA and Arc3B inactivation does not significantly impact P. aeruginosa fitness in pairwise growth competition experiments with B. thai.

Recovery of the indicated P. aeruginosa strains following growth in competition with B. thai wild-type or a strain lacking Tle3 activity (tle3). Means ± SD of biological duplicates with three technical replicates s shown.

The observation that Arc3A proteins are found in bacteria lacking an Arc3B homolog, combined with our finding that Arc3B contributes relatively little to Tle3 defense, led us to speculate that Arc3A plays an intrinsic role in defense, independent of its interaction Arc3B. We tested this by examining the ability of Arc3A-related proteins from other species to complement the defense defect of P. aeruginosa Δarc3A. Despite high-sequence divergence that would likely preclude specific interactions with Arc3B or other P. aeruginosa factors, four of these proteins restored Tle3 defense to a statistically significant degree and two provided protection equivalent to that of P. aeruginosa arc3A expressed in the same manner (Figure 3C). We also examined one of these proteins in its native context. Inactivation of arc3A in Pseudomonas protegens conferred a strong interbacterial antagonism defense defect specific to B. thai Tle3 (Figure 3D). In total, these data strongly suggest that proteins in the Arc3A family possess the intrinsic capacity to protect bacteria against Tle3-like toxins, and potentially phospholipase toxins more broadly.

While the ability of heterologously expressed Arc3A homologs to contribute to P. aeruginosa defense demonstrates that these proteins can function independently, this finding does not rule out the possibility that in P. aeruginosa Arc3B enhances the defensive function of Arc3A. In two of our Tn-seq screens (Figure 1A and C, Figure 2C), we observed a significant difference in arc3B insertion frequency during competition with wild-type B. thai compared to the T6SS- or Tle3-inactivated mutants. However, in pairwise competition experiments with B. thai, we were unable to detect a significant fitness defect associated with arc3B inactivation in either the wild-type or the Δarc3A background (Figure 3—figure supplement 1). These results suggest that Arc3B may play an auxiliary role in defense that is only detectable under the more stringent conditions of the Tn-seq screens, where P. aeruginosa strains compete with each other as well as B. thai.

Enzymes with PLA activity cleave glycerophospholipids at the sn1 or sn2 position, generating fatty acid and detergent-like lysophospholipid products (Filkin et al., 2020). For toxins like Tle3, which are delivered to target cells in exceedingly low quantities (Fridman et al., 2020; Hernandez et al., 2020), it is likely toxic products, rather than phospholipid depletion per se, that most immediately contribute to cell death. Based on the intrinsic capacity of Arc3A to grant protection against Tle3, a predicted PLA, we hypothesized that it directly mitigates damage caused by the toxic products of Tle3. To test this, we measured phospholipid content within extracts derived from competing P. aeruginosa and B. thai strains at a time point immediately prior to detectable changes in P. aeruginosa viability. Strikingly, relative to mixtures containing the wild-type organisms, those containing P. aeruginosa ∆arc3A possessed highly elevated levels of lysophosphatidylethanolamine (LPE) and monolysocardiolipin (MLCL) in a manner dependent on the activity of Tle3 in B. thai (Figure 4A). Lysophosphatidylglycerol, free fatty acid, and corresponding parent phospholipid levels were not measurably altered by Tle3 (Figure 4—figure supplement 1).

Figure 4.

Arc3 prohibits Tle3-catalyzed lysophospholipid accumulation by a mechanism distinct from known pathways.

(A–D) Mass spectrometric (A) and radiographic thin layer chromatography (TLC) (B) analysis of major phospholipid and lysophospholipid species within lipid extracts derived from the indicated mixtures of P. aeruginosa and B. thai strains. P. aeruginosa were grown in 32PO42- prior to incubation with B. thai. Strains were allowed to interact for 1 hr before phospholipids were harvested. Radiolabeled molecules of interest within biological triplicate experiments resolved by TLC were quantified by densitometry (C, D) (E) Mass spectrometric analysis of major phospholipid and lysophospholipid species within lipid extracts derived from mixtures containing B. thai lacking Tle3-specific immunity factors competing with the indicated B. thai strains. (F) Recovery of the indicated P. aeruginosa strains following growth competition against B. thai wild-type or a strain lacking Tle3 activity. (G) Radiographic TLC analysis of products extracted after incubation of P. aeruginosa-derived spheroplasts with purified radiolabeled lysophosphatidylethanolamine (LPE). For interbacterial competition assays, the mean ± SD of biological replicates and associated technical replicates is shown: *p<0.05 using unpaired two-tailed Student’s t-test. TLC images shown are representative of at least two biological replicates. See also Figure 4—source data 1–3.

Mass spectrometric analysis of free fatty acid (A) and parent phospholipid species (B) within lipid extracts derived from the indicated mixtures. Data represent the mean ± SD of biological duplicates with two technical replicates.

Schematic depicting arc3B and aas loci in the indicated bacterial species.

Recovery of the indicated P. aeruginosa strains carrying empty pPSV39 (control) or pPSV39-aas grown in competition with an excess of B. thai with Isopropyl β-D-1-thiogalactopyranoside (IPTG) for induction of expression from the plasmid. Data represent mean and standard deviation from technical duplicates from one biological replicate representative of two conducted.

Figure 4—figure supplement 1.

Tle3 intoxication does not impact free fatty acids nor intact parent phospholipids in P. aeruginosa.

Mass spectrometric analysis of free fatty acid (A) and parent phospholipid species (B) within lipid extracts derived from the indicated mixtures. Data represent the mean ± SD of biological duplicates with two technical replicates.

The accumulated lysophospholipids we observed could derive from P. aeruginosa or they could result from retaliatory activity against B. thai. To distinguish these possibilities, we subjected P. aeruginosa strains with radiolabeled phospholipids to Tle3 intoxication by unlabeled B. thai strains. Radiographic TLC analysis of the phospholipids generated in these strain competition mixtures confirmed that P. aeruginosa lacking Arc3B accumulates LPE and MLCL during Tle3 intoxication (Figure 4B–D). Tle3 belongs to a family of effectors for which there remains no definitively characterized member. We were unable to measure the activity of the enzyme in vitro; thus, to determine whether the accumulation of LPE and MLCL were a direct result of Tle3 activity, we quantified phospholipids in extracts derived from B. thai strains undergoing Tle3-based self-intoxication (Figure 2—figure supplement 3). Samples derived from mixtures composed of a B. thai strain with Tle3 and a derivative strain sensitized to Tle3 intoxication through cognate immunity gene inactivation accumulated LPE and MLCL, mirroring our interspecies competition findings (Figure 4E).

In E. coli, lysophospholipids are transported across the inner membrane by the dedicated transporter LplT, after which they are reacylated by Aas (Zheng et al., 2017). Both LplT and Aas have been shown to play a critical role in defending E. coli cells from phospholipase attack (Lin et al., 2018). In P. aeruginosa, LplT and the acylation domain of Aas are found in a single predicted polypeptide, encoded by PA3267 (aas). Interestingly, the aas gene was illuminated as a potential P. aeruginosa antagonism defense factor in two of the transposon mutant screens we performed and aas genes neighbor arc3 genes in many bacteria (Figure 2C, Figure 1—figure supplement 1B, Figure 4—figure supplement 2, Figure 1—source data 1, Figure 2—source data 1). In pairwise interbacterial competition assays, we were unable to detect the impact of aas inactivation on P. aeruginosa fitness; however, we found that P. aeruginosa strains lacking both aas and arc3A exhibited a Tle3 defense defect substantially greater than that associated with Arc3A inactivation alone (Figure 4F). Aas overexpression complements the Δaas-dependent defect in this strain, but it does not compensate for the inactivation of Arc3A (Figure 4—figure supplement 3). Additionally, we found that Aas is required for PE regeneration from LPE, whereas Arc3A neither contributed to this process, nor could we detect changes in LPE stability upon Arc3A inactivation in the absence of antagonism (Figure 4G). These findings indicate that arc3 encodes a previously undescribed, widespread and independent pathway that prohibits lysophospholipid accumulation by an interbacterial phospholipase toxin.

Figure 4—figure supplement 2.

Arc3B genes are encoded adjacent to aas genes in diverse bacteria.

Schematic depicting arc3B and aas loci in the indicated bacterial species.

Figure 4—figure supplement 3.

Overexpression of aas fails to complement the competitive defect associated with Arc3A inactivation.

Recovery of the indicated P. aeruginosa strains carrying empty pPSV39 (control) or pPSV39-aas grown in competition with an excess of B. thai with Isopropyl β-D-1-thiogalactopyranoside (IPTG) for induction of expression from the plasmid. Data represent mean and standard deviation from technical duplicates from one biological replicate representative of two conducted.

Discussion

Evidence suggests that the existence of bacteria is characterized by an onslaught of challenges to their integrity (Granato et al., 2019; Peterson et al., 2020). This is borne out in part by recent work in the phage arena, which has highlighted a vast number of dedicated bacterial defense systems against these parasites (Bernheim and Sorek, 2020; Hampton et al., 2020). It is now appreciated that the amalgam of these systems constitutes a bacterial immune system, with both innate and adaptive arms that are functionally analogous, and even evolutionarily related to those of eukaryotic organisms (Morehouse et al., 2020). Our work shows that bacteria also possess a multitude of coordinately regulated pathways that provide defense against the threats posed by other bacteria. This further cements the analogy between the immune systems of bacterial and eukaryotic cells, the latter of which also rely on the coordinated production of factors specialized for countering threats of viral or bacterial origin.

We found that Arc3A is the single cellular factor of P. aeruginosa most critical for defense against a PLA toxin. Whether Arc3A, perhaps in concert with Arc3B, offers a wider breadth of defensive activity is not known. However, it is notable that phospholipase toxins are ancient and among the most prevalent bacterial toxins known. Even in the absence of direct antagonism by a PLA toxin, environmental lysophospholipids and those generated during normal cellular processes may represent ongoing threats to bacterial viability (Flores-Díaz et al., 2016). We observed low Arc3A expression in the absence of Gac/Rsm activation, arguing that its role in PARA serves an adaptive purpose.

The precise mechanism by which Arc3 proteins defend against intoxication by phospholipase toxins remains unknown. Although we found that Arc3 grants specific protection against Tle3, the relatively low levels of P. aeruginosa intoxication achieved by Tle1 do not permit us to rule out that it might also provide broader phospholipase defense. The number and density of transmembrane helices in Arc3A family proteins is unusual, and we postulate that these anomalous features of the protein are intimately connected to its mechanism of action. These segments may sequester lysophospholipids and facilitate their turnover or it is conceivable that they directly recruit and interfere with phospholipase enzymes.

We have found that several previously uncharacterized gene clusters under Gac/Rsm control contribute critically to PARA. Our bioinformatic analyses show that the arc clusters are widely conserved among Bacteria, yet the Gac/Rsm pathway is restricted to γ-Proteobacteria. If arc genes play similar roles in these diverse bacteria, which our data suggest, their expression is also likely subject to induction by antagonism. It will be of interest in future studies to probe the generality of coordinated interbacterial defenses by identifying additional signaling systems responsive to antagonism. Collectively, our findings provide new insight into the diversity of mechanisms bacteria have evolved to defend against interbacterial attack and suggest that many more cryptic specialized defense pathways await discovery.

Materials and methods

Bacterial strains and culture conditions

A detailed list of all strains and plasmids used in this study can be found in Appendix 2—key resources table. Bacterial strains under investigation in this study were derived from P. aeruginosa PAO1, B. thai E264, and P.protegens pf-5 (Paulsen et al., 2005; Stover et al., 2000; Yu et al., 2006). All strains were grown in Luria-Bertani (LB) broth and incubated either at 37°C (P. aeruginosa and B. thai) or 30°C (P. protegens). For Pseudomonas strains, the media was supplemented with 25 μg/ml irgasan, 15 μg/ml gentamicin, 5% (w/v) sucrose, and 0.2% (w/v) arabinose as needed. B. thai was cultured with 15 μg/ml gentamicin and 200 μg/ml trimethoprim as necessary, and counterselection for allelic exchange was performed on M9 minimal medium agar plate containing 0.4% (w/v) glucose and 0.2% (w/v) p-chlorophenylalanine. E. coli strains used in this study included DH5α for plasmid maintenance, SM10 for conjugal transfer of plasmids into P. aeruginosa, P. protegens, and B. thai, and UE54 for radiolabeled phosphatidylethanolamine (PE) synthesis. E. coli strains were grown in LB supplemented with 15 μg/ml gentamicin, 150 μg/ml carbenicillin, and 200 μg/ml trimethoprim as needed.

Plasmid construction

All primers used in plasmid construction and generation of mutant strains are listed in Appendix 3—table 1. In-frame chromosomal deletions and point mutations were created using the suicide vector pEXG2 for P. aeruginosa and P. protegens, and pJRC115 for B. thai (Chandler et al., 2009; Rietsch et al., 2005). For the generation of chromosomal mutation constructs, 750 bp regions flanking the mutation site were PCR amplified and inserted stitched together into the appropriate vector via Gibson assembly (Gibson et al., 2009). Site-specific chromosomal insertions in P. aeruginosa were generated using pUC18-Tn7t-pBAD-araE as previously described (Hoang et al., 2000). Our trimethoprim-resistant B. thai strain was generated using pUC18T-mini-Tn7-Tp-PS12-mCherry via the mini-Tn7 system (LeRoux et al., 2012).

Appendix 3—table 1.

Oligonucleotides used in this study.

Oligonucleotides 5′–3′	Source
pEXG2_∆PA2541_F1CAAGCTTCTGCAGGTCGACTCTAGAAGGTGGAACCGGACCTGAAG	Integrated DNA Technology
pEXG2_∆PA2541_R1CTCAGGCCTGGGAAATCATGTCAGCCAGTCC	Integrated DNA Technology
pEXG2_∆PA2541_F2CATGATTTCCCAGGCCTGAGAGGGAACG	Integrated DNA Technology
pEXG2_∆PA2541_R2TAAGGTACCGAATTCGAGCTCGTATGACCCAGGCGGTCG	Integrated DNA Technology
pEXG2_∆PA4323_F1AGCTCGAGCCCGGGGATCCTCTAGAGACGCAGAACCCGATCGAG	Integrated DNA Technology
pEXG2_∆PA4323_R1AAAGCACGCCCGATGGCTTCATACGCGG	Integrated DNA Technology
pEXG2_∆PA4323_F2GAAGCCATCGGGCGTGCTTTGACGGATC	Integrated DNA Technology
pEXG2_∆PA4323_R2GGAAGCATAAATGTAAAGCAAGCTTGAGCGATGCCGAGCTGGA	Integrated DNA Technology
pEXG2_∆PA5114_F1CAAGCTTCTGCAGGTCGACTCTAGAGCCGCATGCCCTTGACCT	Integrated DNA Technology
pEXG2_∆PA5114_R1CGACCTCGGCAATCCATTGCATGCGTCGAATC	Integrated DNA Technology
pEXG2_∆PA5114_F2GCAATGGATTGCCGAGGTCGCATCGGAG	Integrated DNA Technology
pEXG2_∆PA5114_R2AATTCGAGCTCACCAGATCATCGGCGCCG	Integrated DNA Technology
pEXG2_∆PA2536_F1CAAGCTTCTGCAGGTCGACTCTAGACTCGGCCTGGCCCGAGC	Integrated DNA Technology
pEXG2_∆PA2536_R1CAGGCGAACCTCAGGGCGTTTCGTTCATCTCAGGCCTCC	Integrated DNA Technology
pEXG2_∆PA2536_F2GGAGGCCTGAGATGAACGAAACGCCCTGAGGTTCGCCTG	Integrated DNA Technology
pEXG2_∆PA2536_R2TAAGGTACCGAATTCGAGCTCCACATCGGTCTCAGCGAAGC	Integrated DNA Technology
pEXG2_∆PA4320_F1AGCTCGAGCCCGGGGATCCTCTAGATGTTCGGCTTCTACATCATGAAC	Integrated DNA Technology
pEXG2_∆PA4320_R1TGCGCCAGCCCACGCTGGCGTCAGTCAG	Integrated DNA Technology
pEXG2_∆PA4320_F2CGCCAGCGTGGGCTGGCGCAGCCTTTTC	Integrated DNA Technology
pEXG2_∆PA4320_R2GGAAGCATAAATGTAAAGCAAGCTTGTCGGCGGTTTCCTCGCT	Integrated DNA Technology
pEXG2_∆PA5114/PA5113_F1CAAGCTTCTGCAGGTCGACTCTAGAGCCGCATGCCCTTGACCT	Integrated DNA Technology
pEXG2_∆PA5114/PA5113_R1TGTTCGGCGAAATCCATTGCATGCGTCGAATC	Integrated DNA Technology
pEXG2_∆PA5114/PA5113_F2GCAATGGATTTCGCCGAACAAGCCATGAG	Integrated DNA Technology
pEXG2_∆PA5114/PA5113_R2TAAGGTACCGAATTCGAGCTCAACAGCCAGACCACGATGTAGC	Integrated DNA Technology
pEXG2_∆PA5113_F1CAAGCTTCTGCAGGTCGACTCTAGATTGCTAGGGGTGCTGGCG	Integrated DNA Technology
pEXG2_∆PA5113_R1ATGGCTTGTTGGAGCGCGTCATGGCTGC	Integrated DNA Technology
pEXG2_∆PA5113_F2GACGCGCTCCAACAAGCCATGAGCCGGTTC	Integrated DNA Technology
pEXG2_∆PA5113_R2TAAGGTACCGAATTCGAGCTCAACAGCCAGACCACGATGTAG	Integrated DNA Technology
pEXG2_∆PA5112_F1CAAGCTTCTGCAGGTCGACTCTAGATACCGCTGGCAGTTGCCG	Integrated DNA Technology
pEXG2_∆PA5112_R1AGAAGTCCAGCGCCATTCTGATCATTCTCTTACTC	Integrated DNA Technology
pEXG2_∆PA5112_F2CAGAATGGCGCTGGACTTCTGAAACGGCGGC	Integrated DNA Technology
pEXG2_∆PA5112_R2TAAGGTACCGAATTCGAGCTCCGCGCAACCGCCGGTTGG	Integrated DNA Technology
pEXG2_∆PA3267_F1CAAGCTTCTGCAGGTCGACTCTAGATCTACATCGACTTCGACG	Integrated DNA Technology
pEXG2_∆PA3267_R1ATCAGCGAGCTTGGGTCATCGTCCTTGTTAC	Integrated DNA Technology
pEXG2_∆PA3267_F2GATGACCCAAGCTCGCTGATCGATCCGC	Integrated DNA Technology
pEXG2_∆PA3267_R2TAAGGTACCGAATTCGAGCTCTTCGCCGGCCTGTTCGAAG	Integrated DNA Technology
pEXG2_∆PFL_RS01995_F1CAAGCTTCTGCAGGTCGACTCTAGAAACTACAACGTCAGCCTG	Integrated DNA Technology
pEXG2_∆PFL_RS01995_R1TTCATTCAACGGTTTGCCTATCCATTGCATGTGTCG	Integrated DNA Technology
pEXG2_∆PFL_RS01995_F2CGACACATGCAATGGATAGGCAAACCGTTGAATGAA	Integrated DNA Technology
pEXG2_∆PFL_RS01995_R2TAAGGTACCGAATTCGAGCTCCTGCGACCACACCAGCG	Integrated DNA Technology
pEXG2_PA2541_VSVG_F1CAAGCTTCTGCAGGTCGACTCTAGAGCGCTGATACGGACCATGC	Integrated DNA Technology
pEXG2_PA2541_VSVG_R1TTTTCCTAATCTATTCATTTCAATATCTGTATAGGCCTGGCCGTCGCTGCCCCG	Integrated DNA Technology
pEXG2_PA2541_VSVG_F2TATACAGATATTGAAATGAATAGATTAGGAAAATGAGAGGGAACGGGCGAAC	Integrated DNA Technology
pEXG2_PA2541_VSVG_R2TAAGGTACCGAATTCGAGCTCCTCCTGCGGCGCACGATG	Integrated DNA Technology
pEXG2_PA4323_VSVG_F1CAAGCTTCTGCAGGTCGACTCTAGATGGAGTTCCACCAGTTGCGCG	Integrated DNA Technology
pEXG2_PA4323_VSVG_R1TGATCCGTCATTTTCCTAATCTATTCATTTCAATATCTGTATAAAGCACGCCGGCGCGCT	Integrated DNA Technology
pEXG2_PA4323_VSVG_F2TATACAGATATTGAAATGAATAGATTAGGAAAATGACGGATCAGACCGACTG	Integrated DNA Technology
pEXG2_PA4323_VSVG_R2TAAGGTACCGAATTCGAGCTCAAGGTACACTTCTCCGCC	Integrated DNA Technology
pEXG2_PA5114_VSVG_F1CAAGCTTCTGCAGGTCGACTCTAGACTGGAGCTGCGCTACCTGTTCG	Integrated DNA Technology
pEXG2_PA5114_VSVG_R1TAATCTATTCATTTCAATATCTGTATATGGCTGCTCGGCCTCCGA	Integrated DNA Technology
pEXG2_PA5114_VSVG_F2GATATTGAAATGAATAGATTAGGAAAATCGGAGGCCGAGCAGCCA	Integrated DNA Technology
pEXG2_PA5114_VSVG_R2TAAGGTACCGAATTCGAGCTCGCTCGGACCAGATCATCGGC	Integrated DNA Technology
pEXG2_PA5113_VSVG_F1CAAGCTTCTGCAGGTCGACTCTAGACTGCGTCTGGCCTGGCCG	Integrated DNA Technology
pEXG2_PA5113_VSVG_R1TTTTCCTAATCTATTCATTTCAATATCTGTATATGGCTTGTTCGGCGAGGAAC	Integrated DNA Technology
pEXG2_PA5113_VSVG_F2TATACAGATATTGAAATGAATAGATTAGGAAAATGAGCCGGTTCCGCGCTATG	Integrated DNA Technology
pEXG2_PA5113_VSVG_R1TAAGGTACCGAATTCGAGCTCGTCGGGCAACAGCCAGAC	Integrated DNA Technology
pUC18_PA2541_FAGC GAATTCGAGCTCGGTACCACGGGAGGAAAG ATGATTTCCGTCTATCAACTC	Integrated DNA Technology
pUC18_PA2541_RCTCATCCGCCAAAACAGCCAAGCTTTCAGGCCTGGCCGTCGC	Integrated DNA Technology
pUC18_PA4323_FAGCGAATTCGAGCTCGGTACCACGGGAGGAAAGATGAAGCCATCGCGCGCCCTGCTGG	Integrated DNA Technology
pUC18_PA4323_RCTCATCCGCCAAAACAGCCAAGCTTTCAAAGCACGCCGGCGCGCTTCCAG	Integrated DNA Technology
pUC18_PA5114_FAGCGAATTCGAGCTCGGTACCACGGGAGGAAAGATGCAATGGATTTTCATGCTGG	Integrated DNA Technology
pUC18_PA5114_RCTCATCCGCCAAAACAGCCAAGCTTTCATGGCTGCTCGGCCTC	Integrated DNA Technology
pUC18_pfl_rs01995_FAGCGAATTCGAGCTCGGTACCACGGGAGGAAAGATGCAATGGATATTCATGCTGC	Integrated DNA Technology
pUC18_pfl_rs01995_RCTCATCCGCCAAAACAGCCAAGCTTTCACGATGACACTCCTTCATTC	Integrated DNA Technology
pUC18_i35_rs27920_FAGCGAATTCGAGCTCGGTACCACGGGAGGAAAGATGAACTGGGCATTCGCCG	Integrated DNA Technology
pUC18_i35_rs27920_RCTCATCCGCCAAAACAGCCAAGCTTTCATGGCTGGCCGTCCTG	Integrated DNA Technology
pUC18_pspto_rs26695_FAGCGAATTCGAGCTCGGTACCACGGGAGGAAAGATGCTTTGGATTTGTCTGGTAG	Integrated DNA Technology
pUC18_pspto_rs26695_RCTCATCCGCCAAAACAGCCAAGCTTTCATGGCGTTTCAGGCGCTG	Integrated DNA Technology
pUC18_ta05_rs00690_FAGCGAATTCGAGCTCGGTACCACGGGAGGAAAGATGGACGACCTTTTAATCCTG	Integrated DNA Technology
pUC18_ta05_rs00690_RCTCATCCGCCAAAACAGCCAAGCTTTCATTTGTTTTCTCCAGCTTTG	Integrated DNA Technology
pJRC115_ BTH_II0090S264A_F1TAAAACGACGGCCAGTGCCAAGCTTCGACGGATCAGCGTTTCAAGCTGC	Integrated DNA Technology
pJRC115_ BTH_II0090S264A_R1ACCTTGGGCATGACCCATCACCGTGATCGTTTCGTG	Integrated DNA Technology
pJRC115_ BTH_II0090S264A_F2GGTCATGCCCAAGGTACGATCATCACGCTGCTCG	Integrated DNA Technology
pJRC115_ BTH_II0090S264A_R2GCTCGGTACCCGGGGATCCTCTAGACGGCTCGGCACGATGCGC	Integrated DNA Technology
pJRC115_ BTH_I3226S264A_F1TAAAACGACGGCCAGTGCCAAGCTTCGACGGATCAGCGTTTCAAGCTGC	Integrated DNA Technology
pJRC115_ BTH_I3226S264A_R1ACCTTGGGCATGACCCATCACCGTGATCGTTTCGTG	Integrated DNA Technology
pJRC115_ BTH_I3226S264A_F2GGTCATGCCCAAGGTACGATCATCACGCTGCTCG	Integrated DNA Technology
pJRC115_ BTH_I3226S264A_R2GCTCGGTACCCGGGGATCCTCTAGACGGCTCGGCACGATGCGC	Integrated DNA Technology
pJRC115_ BTH_I2698S267A_F1TCAATCAGTATCTAGAGGGACACCTTTCTCAAGCGAAATC	Integrated DNA Technology
pJRC115_ BTH_I2698S267A_R1GCCGCGCGCGAAACCAAACACATAAAGGCGAATGCG	Integrated DNA Technology
pJRC115_ BTH_I2698S267A_F2GGTTTCGCGCGCGGCGCAGCGGAAGCTCGCACGTTTTC	Integrated DNA Technology
pJRC115_ BTH_I2698S267A_R2TGTTAAGCTAGAATTCCATCAAACCCGCTGTCCCATGCTC	Integrated DNA Technology
pJRC115_ BTH_I2701S267A_F1TAAAACGACGGCCAGTGCCAAGCTTGAAAGACGAGCAGGACGCG	Integrated DNA Technology
pJRC115_ BTH_I2701S267A_R1ACCCCGCGCGAAACCGAATACGTATAGCCGAATGCG	Integrated DNA Technology
pJRC115_ BTH_I2701S267A_F2GGTTTCGCGCGGGGTGCAGCGGAGGCTCGCAC	Integrated DNA Technology
pJRC115_ BTH_I2701S267A_R2GCTCGGTACCCGGGGATCCTCTAGACTCGACCTCCAGCAGATCG	Integrated DNA Technology
pJRC115_ ∆BTH_I2691_F1TAAAACGACGGCCAGTGCCAAGCTTATTTCAAGCGCGGCCAGTC	Integrated DNA Technology
pJRC115_ ∆BTH_I2691_R1ATCTATGCGAGCTTCCCGCCATTTTTATTCC	Integrated DNA Technology
pJRC115_ ∆BTH_I2691_F2GGCGGGAAGCTCGCATAGATGAGTGATG	Integrated DNA Technology
pJRC115_ ∆BTH_I2691_R2GCTCGGTACCCGGGGATCCTCTAGATTCTGTCAATACTTAAAATACAATTTTC	Integrated DNA Technology
pJRC115_ ∆BTH_II0089-94_F1TAAAACGACGGCCAGTGCCAAGCTTGAATCAGTGCATCGCTGTAC	Integrated DNA Technology
pJRC115_ ∆BTH_II0089-94_R1ATGTTTTGCTTAGCGTATCGTTCAAATTGG	Integrated DNA Technology
pJRC115_ ∆BTH_II0089-94_F2CGATACGCTAAGCAAAACATGAGATTATTGAAGAG	Integrated DNA Technology
pJRC115_ ∆BTH_II0089-94_R2GCTCGGTACCCGGGGATCCTCTAGATATGCAACGCATTGCCGAAAC	Integrated DNA Technology
pJRC115_ ∆BTH_I3225-8_F1TAAAACGACGGCCAGTGCCAAGCTTCCGGTCAATATACCACCATC	Integrated DNA Technology
pJRC115_ ∆BTH_I3225-8_R1ATGTGCGCCCCGTATCGTTCAAATTGGTCAC	Integrated DNA Technology
pJRC115_ ∆BTH_I3225-8_F2GAACGATACGGGGCGCACATAATAAGACTTG	Integrated DNA Technology
pJRC115_ ∆BTH_I3225-8_R2GCTCGGTACCCGGGGATCCTCTAGAATTTGCTGTTTCTGCTGATG	Integrated DNA Technology
PCR_1AGTCTCGTGGGCTCGGAGATGTGTATAAGAGACAGGGGGGGGGGGGGGGGG	Integrated DNA Technology
PCR_1BTCATCGGCTCGTATAATGTGTGG	Integrated DNA Technology
PCR_2ACAAGCAGAAGACGGCATACGAGATTCGCCTTAGTCTCGTGGGCTCGG	Integrated DNA Technology
PCR_2BCAAGCAGAAGACGGCATACGAGATCTAGTACGGTCTCGTGGGCTCGG	Integrated DNA Technology
PCR_2CAATGATACGGCGACCACCGAGATCTACACCTAGAGACCGGGGACTTATCAGCCAACCTGTTA	Integrated DNA Technology
Seq_primerCTAGAGACCGGGGACTTATCAGCCAAC	Integrated DNA Technology

Generation of mutant strains

For P. aeruginosa and P. protegens strain generation, pEXG2 mutation constructs were transformed into E. coli SM10, and SM10 donors were subsequently mixed and incubated together with Pseudomonas recipients on a nitrocellulose membrane on top of an LB agar plate at a ratio of 10:1 donor–recipient. The cell mixtures were incubated for 6 hr at 37°C to allow for conjugation. These cell mixtures were then scraped up, resuspended into 200 μl LB, and plated on LB agar plates containing irgasan and gentamicin to select for cells containing the mutant construct inserted into the chromosome. Pseudomonas merodiploid strains were then grown overnight in nonselective LB media, followed by counterselection on LB no-salt agar plates supplemented with sucrose. Gentamicin-sensitive, sucrose-resistant colonies were screened for allelic replacement by colony PCR, and mutations were confirmed via Sanger sequencing of PCR products.

For B. thai strain generation, pJRC115 mutation constructs were transformed into E. coli SM10. Conjugation to enable plasmid transfer was performed as described above and plated on LB agar containing gentamicin and trimethoprim to select for B. thai with the chromosomally integrated plasmid. Merodiploids were grown overnight in LB media at 37°C, then plated on M9 minimal medium containing p-chlorophenylalanine for counterselection. Trimethoprim-sensitive, p-chlorophenylalanine-resistant colonies were screened for allelic replacement by colony PCR, and mutations were confirmed via Sanger sequencing of PCR products.

For expression of Arc1F, Arc2G, Arc3A, and Arc3A-homologs in P. aeruginosa, pUC18-Tn7t-pBAD-araE containing the gene of interest and the helper plasmid pTNS3 were co-transformed into P. aeruginosa strains by electroporation (Choi et al., 2008). After 6 hr of outgrowth in LB at 37°C, transformants were plated on an LB agar plate with gentamicin to select for cells with mini-Tn7 integration.

Bacterial competition assay

For each co-culture competition experiment, donor and recipient strains were first grown for 20 hr in LB medium. Cultures were spun for 1 min at 20,000 × g to pellet cells, the supernatant removed, and cells were washed once with fresh LB medium. Cell pellets were resuspended in LB and normalized to OD600 = 20. Mixtures of donor and recipient strains were then established at 20:1 (B. thai vs. P. aeruginosa, P. protegens, or B. thai) or 10:1 (P. aeruginosa vs. B. thai) v/v ratios. The initial ratios of donor and recipient strains in these mixtures were measured by performing 10-fold serial dilutions and plating on appropriate selective media to evaluate by colony-forming unit (CFU) analysis. The co-culture competitions were initiated by spotting 5 µl of each mixture onto nitrocellulose filters placed on 3% (w/v) agar LB no-salt plates supplemented with L-arabinose for induction of gene expression as necessary. Competitions were incubated for 6 hr at 37°C between B. thai and P. aeruginosa, or 6 hr at 30°C for competition between B. thai and P. protegens. For B. thai self-intoxication, competitions were incubated at 37°C for 3 hr. Cells were harvested by scraping individual spots from excised sections of the nitrocellulose filter into LB medium. Suspensions were serially diluted and plated on selective media for CFU quantification.

Immunoblotting analysis

To analyze the expression of Arc1F-VSV-G, Arc2G-VSV-G, and Arc3A-VSV-G, P. aeruginosa strains were grown in LB broth at 37°C to log phase. Cells were pelleted and resuspended in lysis buffer (20 mM Tris-HCl, pH 7.5, 300 mM NaCl, 10% [v/v] glycerol) and then mixed 1:1 with 2X SDS-PAGE sample loading buffer. Samples were then boiled at 100°C for 10 min and loaded at equal volumes to resolve using SDS-PAGE, then transferred to nitrocellulose membranes. Membranes were blocked in TBST (10 mM Tris-HCl pH 7.5, 150 mM NaCl, and 0.1% [w/v] Tween-20) with 5% (w/v) non-fat milk for 30 min at room temperature, followed by incubation with anti-VSV-G or with anti-ribosome polymerase β subunit primary antibodies diluted in TBST for 1 hr at room temperature on an orbital shaker. Blots were then washed with TBST, followed by incubation with secondary antibody (goat anti-rabbit or anti-mouse HRP conjugated) diluted in TBST for 30 min at room temperature. Finally, blots were washed with TBST again and developed using Radiance HRP substrate and visualized using iBright imager.

Lipidomic analysis

To analyze lipid content after intra- and inter-species intoxication, overnight cultures of donor and recipient cells were harvested and mixed at 1:1 (v/v) ratio in LB medium. These mixtures were spotted onto a 3% (w/v) agar LB no-salt plate and incubated for 1 hr at 37°C to allow for intoxication. Cells were then collected, and the lipids were extracted using the Bligh–Dyer method (Bligh and Dyer, 1959). Briefly, bacterial pellets were resuspended in 0.5 ml of solution containing 0.5 M NaCl in 0.5 N HCl, followed by adding 1.5 ml of chloroform/methanol mixture (1:2, v/v). Suspensions were vortexed at room temperature for 15 min, 1.5 ml of 0.5 M NaCl in 0.5 N HCl solution was added to each sample, and vortexed for an additional 5 min. Lipid and aqueous layers were separated by centrifugation at 2000 × g for 5 min. The lower lipid phase was collected, and these dried lipid samples were analyzed for PE, PG, CL, LPE, LPG, and MLCL content by the Kansas State Lipidomics Research Center.

Lipid samples were analyzed by electrospray ionization triple quadrupole mass spectrometry in direct infusion mode, and data were processed as described by with slight modifications (Shiva et al., 2013), which include use of a Waters Xevo TQS mass spectrometer (Milford, MA). Internal standards used to normalize data are shown in Appendix 1—table 2. For PE and LPE, SPLASH standards were used for normalization (Avanti Polar Lipids, Alabaster, AL). Mass spectrometry global parameters, scan modes, and data processing parameters are shown in Appendix 1—table 2.

Appendix 1—table 2.

Internal standards and parameters used in lipidomic analysis.

Internal standards used
Compound formula	Compound Name	Alternative name
C20H41O9P	LysoPG(14:0)
C24H49O9P	LysoPG(18:0)
C34H67O10P	PG(14:0/14:0)
C46H91O10P	PG(20:0/20:0), i.e. diphytanoyl PG
C19H40O7PN	LysoPE(14:0)
C23H48O7PN	LysoPE(18:0)
C29H58O8PN	PE(12:0/12:0)
C45H90O8PN	PE(20:0/20:0), i.e. diphytanoyl PE
C66H120O17P2	CL(57:4)	14:1(3)–15:1 CA
C70H134O17P2	CL(61:1)	15:0(3)–16:1 CA
C89H166O17P2	CL(80:4)	22:1(3)–14:1 CA
Mass spectrometry global parameters on Waters Xevo TQS mass spectrometer
Parameter	Value
Cycle time	Automatic
Source temperature (°C)	150
Desolvation temperature (°C)	250
Cone gas flow (L/h)	150
Desolvation gas fow (L/h)	650
Collision gas flow (mL/min)	0.14
Nebuliser gas (Bar)	7
LM 1 Resolution	2.8
HM 1 Resolution	14.8
Mass spectral acquisition and data processing parameters
Acquisition parameters
Compound	Function	m/z range	Start/end time (min)	Scan duration (s)	Ionization mode	Cone voltage (V)	Collision energy (V)	Total scan time (min)
PG	Neutral Loss of 189.04	680–880	1.6–5	2	ES+	30	8	3.4
LPE, PE	Neutral Loss of 141.02	422–922	1.6–5	5	ES+	40	20	3.4
LPG	Parents of 152.9	420–570	5.2–10	0.75	ES-	30	38	4.8
CL, MLCL	Parents of 153	865–1,572	1.6–30	4.7	ES-	40	75	28.4
Data processing parameters
Subtraction	Smooth mean	Centroid mid peak width at half height (all top 50%)
1, 40, 0.01	2 × 0.4	2
1, 40, 0.01	2 × 0.4	2
1, 40, 0.01	1 × 0.6	5
1, 40, 0.01	2 × 0.6	5

Proteomic analysis

P. aeruginosa strains containing VSV-G-tagged proteins were grown in 50 ml to mid-log phase, centrifuged at 2500 × g for 15 min, and the pellets harvested by resuspension in lysis buffer (20 mM Tris-HCl pH 7.5, 300 mM NaCl, 5% [w/v] glycerol, 0.5% [v/v] Triton X-100). Cell lysates were prepared by sonication, and tagged proteins were enriched by incubating cell lysates with 30 μl of anti-VSV-G agarose beads at 4°C for 4 hr with constant rotation. Agarose beads were then pelleted by centrifugation at 100 × g for 2 min and washed three times with 20 mM ammonium bicarbonate. VSV-G agarose beads and bound proteins were then treated with 10 μl of 10 ng/μl (100 ng total per sample) sequencing grade trypsin for 16 hr at 37°C. After digestion, 40 μl of 20 mM ammonium bicarbonate was added to the agarose beads and peptide mixture and lightly mixed. Beads were centrifuged at 300 × g for 3 min, and the supernatant collected as the peptide fraction. This peptide mixture was reduced with 5 mM Tris(2-carboxyethyl) phosphine hydrochloride for 1 hr, followed by alkylation using 14 mM iodoacetamide for 30 min in the dark at room temperature. Alkylation reactions were quenched using 5 mM 1,4-dithiothreitol. Samples were then diluted with 100% acetonitrile (ACN) and 10% (w/v) trifluoroacetic acid (TFA) to a final concentration of 5% ACN (v/v) and 0.5% TFA (w/v) and applied to MacroSpin C18 columns (30 μg capacity) that had been charged with 100% ACN and ddH2O. Bound peptides were then washed twice in 5% (v/v) ACN and 0.5% (w/v) TFA, before elution with 70% (v/v) ACN and 0.1% (v/v) formic acid (FA).

Peptides were analyzed by LC-MS/MS using a Dionex UltiMate 3000 Rapid Separation nanoLC and a Q Exactive HF Hybrid Quadrupole-Orbitrap Mass Spectrometer (Thermo Fisher Scientific Inc, San Jose, CA). Approximately 1 μg of peptide samples was loaded onto the trap column, which was 150 μm × 3 cm in-house packed with 3 μm C18 beads. The analytical column was a 75 μm × 10.5 cm PicoChip column packed with 3 μm C18 beads (New Objective, Inc, Woburn, MA). The flow rate was kept at 300 nl/min. Solvent A was 0.1% FA in water and Solvent B was 0.1% FA in ACN. The peptide was separated on a 120 min analytical gradient from 5% ACN/0.1% FA to 40% ACN/0.1% FA. The mass spectrometer was operated in data-dependent mode. The source voltage was 2.10 kV, and the capillary temperature was 320°C. MS 1 scans were acquired from 300 to 2000 m/z at 60,000 resolving power and automatic gain control (AGC) set to 3 × 106. The top 15 most abundant precursor ions in each MS 1 scan were selected for fragmentation. Precursors were selected with an isolation width of 2 Da and fragmented by higher-energy collisional dissociation (HCD) at 30% normalized collision energy in the HCD cell. Previously selected ions were dynamically excluded from re-selection for 20 s. The MS 2 AGC was set to 1 × 105.

Proteins were identified from the tandem mass spectra extracted by Xcalibur version 4.0. MS/MS spectra were searched against the Uniprot P. aeruginosa PAO1 strain database using Mascot search engine (Matrix Science, London, UK; version 2.5.1). The MS 1 precursor mass tolerance was set to 10 ppm, and the MS 2 tolerance was set to 0.05 Da. A 1% false discovery rate cutoff was applied at the peptide level. Only proteins with a minimum of two unique peptides above the cutoff were considered for further study. The search result was visualized by Scaffold (version 4.8.3, Proteome Software, Inc, Portland, OR).

Transposon mutant library construction

A P. aeruginosa PAO1 transposon mutant library of ~80,000 unique Himar1 insertions was prepared using established protocols (Lee et al., 2017). Briefly, P. aeruginosa were mutagenized by delivery of the transposon- and transposase-bearing suicide vector pBT20 from E. coli SM10 λpir (Kulasekara et al., 2005). Insertion mutants were selected on LB agar containing irgasan and gentamycin and pooled. The library was harvested and transposon insertion sites of the complete pool were defined by transposon insertion sequencing as described below.

Transposon mutant library screen in bacterial competition

The P. aeruginosa transposon mutant library was grown at 37°C to mid-log phase. Cell suspensions were then normalized to OD600 = 1 and incubated with 100 OD600 of B. thai donor strains (wild-type, ∆icmF, tle3S264A, or ∆colA). Competitions were performed on 3% (w/v) agar LB no-salt plates at 37°C for 7 hr. Cells were washed once with PBS, diluted, and plated on LB agar plates supplemented with irgasan to select for viable P. aeruginosa cells. After overnight incubation, cells were harvested, and genomic DNA was directly extracted from pellets using QIAGEN Blood and Tissue Midi gDNA prep kit. Transposon insertion sequencing libraries were generated from 3 µg gDNA per sample using the C-tailing method as described (Gallagher, 2019) with the transposon-specific primers listed in Appendix 3—table 1. For PCR round 1, PCR_1A and PCR_1B were used. For PCR round 2, PCR_2A, PCR_2B, and PCR_2C were used. For sequencing, Seq_primer was used. Libraries were pooled and sequenced in multiplex using an Illumina MiSeq with a 5% PhiX spike-in.

Tn-seq data analysis

Seqmagick was used to trim the first six bases from each read (https://github.com/fhcrc/seqmagick; Matsen Group, 2020). Reads were then aligned, and counts were enumerated using TRANSIT TPP (https://transit.readthedocs.io/en/latest/tpp.html). An annotation GFF was created from the original P. aeruginosa PAO1 GFF file (GCF_000006765.1) and was translated to TRANSIT portable format using the TRANSIT 'convert gff_to_prot_table' command (https://transit.readthedocs.io/en/latest/transit_running.html#prot-tables-annotations). Finally, the TRANSIT 'export combined_wig' command was used to combine and annotate the counts. Appendix 1—table 3 provides summary data of the sequencing runs and their processing by TRANSIT.

Appendix 1—table 3.

Transposon sequencing data summary.

Growth experiment	Mutant pool pre-competition	Pool with WT B. thai - rep 1	Pool with δt6s B. thai - rep 1
fastq file name	tn_mutant_pool.fastq	pool_v_wt_rep1.fastq	pool_v_deltaT6_rep1.fastq
Total reads	515,434	5,788,728	6,452,899
Trimmed (valid Tn prefix)	515,293	5,786,427	6,450,594
Mapped	350,923	4,666,597	5,283,645
Mapped to TA sites	349,915	4,634,942	5,241,672
TA_sites	94,404	94,404	94,404
TAs_hit	50,348	57,644	60,331
Max reads per TA site	324	9,519	5,886
Mean reads per hit TA site	6.9	80.4	86.9
Pool with WT B.thai - rep 2	Pool with ΔT6S B.thai - rep 2	Pool with tle3S264A B. thai	Pool with Δ colA B. thai
pool_v_wt_rep2.fastq	pool_v_deltaT6_rep2.fastq	pool_v_tle3.fastq	pool_v_colA.fastq
411,515	720,133	3,226,978	3,130,545
411,401	719,895	3,225,816	3,129,574
266,466	481,466	3,024,047	2,938,318
265,661	479,769	2,997,466	2,922,995
94,404	94,404	94,404	94,404
46,500	53,148	59,145	58,101
669	447	3,484	3,568
5.7	9.0	50.7	50.3

Bioinformatic analysis of Arc3 gene distribution

The phylogenetic profiler tool in IMG (Chen et al., 2021) was used to determine the distribution of Arc3A homologs (identified as proteins classified under the Pfam 10101) across bacterial phyla. The number and location of predicted transmembrane domains present in representative Arc3A homologs from each phylum were calculated using CCTOP and TMpred (https://embnet.vital-it.ch/software/TMPRED_form.html) (Dobson et al., 2015). Homologs of Arc3B encoded adjacent to Arc3A-like genes were identified on the basis of their classification in the Pfam 13,163.

Bacterial stress assays

P. aeruginosa cultures were assayed for the ability to survive oxidative stress (50 mM H2O2), heat stress (55°C), osmotic stress (3 M NaCl), and detergent stress (0.5% [w/v] SDS and 0.25 mM EDTA) in LB media as previously described with modifications (Jørgensen et al., 1999). After incubation with the oxidative, heat, and osmotic stress conditions, cultures were serially diluted 10-fold in LB medium and plated on LB agar for subsequent analysis of remaining viable cells by CFU determination. For oxidative stress, stationary-phase cultures were diluted to OD600 = 0.1 and incubated in LB media containing 50 mM H2O2 for 30 min at 37°C with continuous shaking. For heat stress, stationary-phase cultures were diluted in pre-warmed LB media to OD600 = 0.1 and incubated at 55°C for 30 min with continuous shaking. For osmotic stress, overnight cultures were diluted in LB media containing 3 M NaCl to OD600 = 0.1. Cultures were incubated at 37°C for 20 hr with shaking. For detergent stress, cultures were serially diluted in LB media and directly plated on LB agar plates containing 0.5% (w/v) SDS and 0.25 mM EDTA for CFU quantification.

Phospholipid analysis

To prepare [32P]-labeled bacteria, P. aeruginosa strains were grown in LB media supplemented with 5 μCi/ml [32P]-orthophosphoric acid. Tle3 intoxication was performed by mixing 200 OD600 of B. thai and [32P]-P. aeruginosa at a 10:1 ratio (v/v) and incubating this mixture at 37°C for 1 hr on 3% (w/v) agar LB no-salt plates containing L-arabinose as necessary. After 1 hr of intoxication, cells were collected, and the lipids were extracted using the Bligh–Dyer method as described above. Purified lipid samples were loaded onto a silica gel thin-layer plate and developed with chloroform/methanol/acetic acid/ddH2O (85:15:10:3.5, v/v/v/v) solvent system (Lopalco et al., 2017). The air-dried plate was exposed to a storage phosphor screen. Individual phospholipid was visualized and quantified using phosphorimaging to calculate phospholipid content expressed as mol% of the total phospholipid pool.

PE regeneration assay

To generate [32P]-labeled PE, E. coli UE54 strain was grown in LB media supplemented with 5 μCi/ml [32P]-orthophosphate overnight at 37°C (Harvat et al., 2005). Cells were collected and the lipids were extracted using the Bligh–Dyer method as described above. The extracted lipids were dried and resuspended in buffer containing 100 mM HEPES-NaOH, pH = 7.5, 100 mM KCl, 10 mM CaCl2, and 1% (w/v) n-Dodecyl-β-d-maltoside. 10 units of pancreas phospholipase A2 were added to digest PE to LPE at 37°C overnight with shaking. After incubation, lipids were extracted and loaded onto a silica gel thin-layer plate and developed with chloroform/methanol/acetic acid/ddH2O (85:15:10:3.5, v/v/v/v) solvent system. The dried plate was exposed to an x-ray film for 2 hr. The phospholipid bands were visualized by developing the film, and bands corresponding to [32P]-LPE on the TLC plate were scraped, extracted, and resuspended in 100% ethanol.

P. aeruginosa spheroplasts were generated by resuspending log-phase culture in 25 mM Tris-HCl, pH 8, 450 mM sucrose, and 1.4 mM EDTA. After addition of 20 μg/ml lysozyme, cells were incubated on ice for 30 min. Intact spheroplasts were collected by centrifugation (3000 × g for 5 min) at 4°C and gently resuspended in 25 mM Tris-HCl, pH 8, and 450 mM sucrose. To examine PE regeneration, [32P]-LPE was added into spheroplast solutions and incubated at 37°C. The reactions were terminated at the indicated time by adding chloroform-methanol mixture (1:2, v/v). The lipids were extracted, separated by TLC, and analyzed using phosphorimaging as described above (Lin et al., 2019).

Statistical analyses

GraphPad Prism 7 software (San Diego, CA) was used for statistical analysis. We used an unpaired two-tailed Student’s t-test for pairwise comparisons. In all cases, p<0.05 was considered statistically significant.

This study identifies novel (Arc) pathways that mediate resistance of bacteria to attacks by other bacteria. The identified genes appear to specifically protect host cells from the action of bacterial toxins. In particular, one system (Arc3) is shown to protect Pseudomonas aeruginosa from Type-6 secretion phospholipase effectors. Thus, while it is growing increasingly clear that bacteria have evolved numerous protection mechanisms against phages, the study expands this concept to bacteria.

Our editorial process produces two outputs: i) public reviews designed to be posted alongside the preprint for the benefit of readers; ii) feedback on the manuscript for the authors, including requests for revisions, shown below. We also include an acceptance summary that explains what the editors found interesting or important about the work.

Decision letter after peer review:

Thank you for submitting your article "Coordinately regulated interbacterial antagonism defense pathways constitute a bacterial innate immune system" for consideration by eLife. Your article has been reviewed by 2 peer reviewers, and the evaluation has been overseen by a Reviewing Editor and Wendy Garrett as the Senior Editor. The following individuals involved in review of your submission have agreed to reveal their identity: Lotte Sogaard-Andersen (Reviewer #2).

The reviewers have discussed their reviews with one another, and the Reviewing Editor with input from the senior editor. Consensus was that the manuscript is of significant import and interest but will require revision. We have drafted this document to help you prepare a revised submission.

Essential revisions:

1) The authors define the arc1-arc3 genes as part of a "bacterial innate immune system". This should be toned down, because it is not proven that they allow self- vs non-self-discrimination. In fact, the heterologous expression of Arc3B protects as well in Pa (presumably even Arc3B encoded in a B. thai) which argues against true self -vs non-self-discrimination.

2) The protective function of Arc3B could be further clarified: does it act directly on Tle3 or indirectly via a converging pathway that removes LPE or MLCL or acts on them? While it is understood that defining the exact molecular function of Arc3B is beyond the immediate scope of this manuscript, it should be possible to discriminate, testing for example if Arc3B can can protect against another phospholipase or whether LPE/MCL levels can be affected by Arc3B independently of Tle3. And last, does Arc3B require Arc3A for activity?

3) The manuscript is densely written and should be edited to make it an easier read for a broad audience.

Reviewer #1 (Recommendations for the authors):

1. line 126: Arc1 inactivation is described as resulting in a modest reduction in B. thai survival. My reading of the figure is that Arc1 inactivation results in a modest increase in B. thai survival. Please clarify.

2. line 126-127: I do not understand these comparisons. Maybe explain in a bit more detail.

3. Figure 2B: What is plotted on the Y-axis?

Reviewer #2 (Recommendations for the authors):

The authors define the arc1-arc3 genes as part of a "bacterial innate immune system". I found this choice questionable since it is not proven that they allow self- vs non-self-discrimination. In fact, since heterologous expression of Arc3B can protect as well in Pa (presumably even Arc3B encoded in a B. thai), it's not a true self -vs non-self-discrimination. Would referring to it as an antitoxin not be suitable instead?

The underpinnings protective (antitoxin) effect of Arc3B remain murky: does it fulfill a structural role, perhaps sequestering Tle3 or interfering with its activity, stability or translocation? Or does it act via a converging pathway that removes LPE or MLCL or acts on them, in parallel to Aas? In the latter case, Arc3B should be possible to protect against phospholipid cleavage caused in other ways than Tle3-mediated intoxication. Can this be tested with another phospholipase for example (expressed in Pa for self-intoxication)? The authors have not shown that LPE and MLCL levels can be modulated by Arc3B (with or without Arc3A) in the absence of Tle3, begging the question whether Arc3B acts specifically on Tle3 or phospholipase-mediated damage in general. Conversely can Aas overexpression compensate for loss of Arc3B and/or Arc3A in response to Tle3 intoxication?

Were the Arc3B orthologs that were chosen for complementation only from species that encode Arc3A or also from those that encode a solitary Arc3B? This could be discussed.

Arc3B interacts with Arc3A, but it does not appear to have enzymatic activity. It would be useful to know if Arc3B requires Arc3A for activity (assessed by epistasis experiments using Arc3B/A single and double mutants). One would predict that Arc3B can act independently of Arc3A since Arc3A is not always co-conserved with Arc3B, in which case the association between Arc3A and Arc3B should not be important to protect against Tle3.

Line 111 and panel A with B of Figure1—figure supplement 3: I don't see the cumulative effect claimed when compared the triple mutant Δarc123 (panel A) to single deletions (panel B).

Figure 1—figure supplement 3A: the names of the deleted strains seem incorrect: Δarc1DΔarc1G should be Δarc2DΔarc2G; as well Δarc1AΔarc1B is Δarc3AΔarc3B.

Line 218: Protection by Arc3A is valid for Tle3 and not against Tle1 (that is also a phospholipase). Change "against interbacterial phospholipase toxins" to "against Tle3 interbacterial phospholipase toxin".

Essential revisions:

1) The authors define the arc1-arc3 genes as part of a "bacterial innate immune system". This should be toned down, because it is not proven that they allow self- vs non-self-discrimination. In fact, the heterologous expression of Arc3B protects as well in Pa (presumably even Arc3B encoded in a B. thai) which argues against true self -vs non-self-discrimination.

We thank the reviewer(s) for raising this matter. One possibility here is that the reviewer(s) may misunderstand how producers of Tle3 (such as B. thailandensis) prevent self-intoxication. These organisms do not rely on Arc3A homologs to provide protection; indeed, B. thai lacks a discernable homolog of arc3A and we generally see no correlation between the presence of arc3A homologs and predicted phospholipase toxin genes across bacterial genomes. Rather, self-intoxication is prevented (with 100% efficacy) through the production of dedicated immunity proteins encoded adjacent to the effector genes.

Additionally, we note that in eukaryotic innate immune systems, discrimination between self and non-self typically involves a means of sensing the foreign invader coupled to induction of protective responses. In P. aeruginosa, production of Arc3A is controlled by the Gac/Rsm globally regulatory system, and we show that this is associated with low expression of Arc3A during pure culture growth (Figure 2—figure supplement 2). Previously, we demonstrated that the Gac/Rsm pathway is activated specifically by the presence of an antagonistic competitor, via the lysis of a subset of the population and release of an as yet unidentified cellular component that serves as the activating signal (LeRoux et al., 2015. eLife). Thus, paralleling eukaryotic innate immune system effectors, Arc3A becomes activated only under conditions where a foreign threat is perceived.

Finally, we would respectfully disagree with the assertion that self-vs. non-self discrimination is the sole function of an immune system. A key function of the eukaryotic innate immune system is the detection of and response to damaged cells (Kono and Rock, 2008. Nat. Rev. Microbiol.). This occurs through the detection of so-called danger associated molecular patterns (DAMPs), which can be generated either through the action of foreign invaders or self-derived sources of cellular damage (Shi et al., 2003 Nature; Martinon et al., 2006. Nature).

However, these considerations aside, at the reviewers’ request, we have toned down the language we use to describe the pathways we have discovered in both the discussion (p15, line 390) and the title of the paper (now modified to “Discovery of coordinately regulated pathways that provide innate protection against interbacterial antagonism”).

2) The protective function of Arc3B could be further clarified: does it act directly on Tle3 or indirectly via a converging pathway that removes LPE or MLCL or acts on them? While it is understood that defining the exact molecular function of Arc3B is beyond the immediate scope of this manuscript, it should be possible to discriminate, testing for example if Arc3B can can protect against another phospholipase or whether LPE/MCL levels can be affected by Arc3B independently of Tle3. And last, does Arc3B require Arc3A for activity?

We agree that it would be of interest to further clarify the protective function of Arc3A. We have attempted numerous experiments in an attempt to determine whether Arc3A can provide protection against other phospholipases, but have been met with a number of technical challenges. B. thailandensis produces a second phospholipase effector, Tle1. This protein has a less significant impact on P. aeruginosa growth than Tle3, but by reducing P. aeruginosa defenses through inactivation of the T6SS in this organism, we were able to detect a small but significant level of inhibition attributable to Tle1 (Author response image 1) . Inactivation of Ar3A under these conditions did not further sensitize P. aeruginosa to the activity of Tle1. However, a key caveat of this experiment is that the impact of Arc3A on Tle1-mediated intoxication in these assays may be masked by the stronger effect of Tle3. A “publishable” experiment would involve inactivating Tle3 to isolate the effect of Tle1-mediated intoxication. However, two homologous but non-identical copies of both tle3 and tle1 are encoded in B. thai, and generating the quadruple mutant necessary to perform this experiment in a controlled fashion has proved technically challenging. A second avenue we have explored is to evaluate whether inactivation of Arc3A results in sensitivity to other antagonists that encode known or predicted phospholipase toxins. Inactivation of Arc3A failed to sensitize P. aeruginosa to intoxication by six different phospholipase-encoding strains from five species. However, we are not confident in the interpretation of this result because each of the toxin-producing strains employed is significantly out-competed by P. aeruginosa, even when the experiment is initiated at a high donor:recipient ratio (, see strongly negative C.I. values), which is perhaps not surprising given the extensive offensive and defense arsenal this organism can employ. Thus, we have been unable to conclusively evaluate whether Arc3A confers protection against additional lysophospholipid-generating toxins.

Author response image 1.

Inactivation of Arc3A does not detectably impact P. aeruginosa growth in competition with assorted phospholipase producing species.

(A) Recovery of P. aeruginosa with the indicated genotypes following growth in competition with an excess of wild-type or Tle1-inactivated B. thai. (B and C) Recovery of P. aeruginosa (B) or competitive index obtained (C; final donor:recipicient ratio divided by initial donor:recipienct ratio) following growth in competition with the indicated organisms.

To evaluate whether Arc3A can affect LPE levels in the absence of Tle3, we have isolated spheroplasts of P. aeruginosa expressing Arc3A or a derivative in which arc3A is deleted and incubated these with radiolabeled LPE. We find that Arc3A does not significantly affect the rate of LPE degradation in this assay (Figure 4G). This result was included but not highlighted in the original manuscript. We have now added mention of this finding to the results (p. 14, lines 365-366).

Finally, while our Tn-seq results indicate that Arc3B contributes to P. aeruginosa competitiveness, in pairwise competition experiments we see no evidence that Arc3B is required for Arc3A activity. This finding is now incorporated into the revised manuscript (revised Figure 3—figure supplement 1 and p12, lines 267-276). Arc3B is encoded downstream of Arc3A, suggesting the difference in phenotype we observed between the two assays is not due merely to polar effects of transposon insertion. Of note, we also found that transposon insertions in aas were associated with reduced P. aeruginosa fitness under the conditions of our tn-seq screen, but similarly, deletion of this gene did not have an effect on competitiveness in pairwise competitions. These findings suggest the transposon screen represents a more sensitive assay for the detection of genes important for fitness during interbacterial competition, and thus that Arc3B plays an important but auxiliary role in P. aeruginosa defense against antagonism.

3) The manuscript is densely written and should be edited to make it an easier read for a broad audience.

We regret any difficulty that our writing style may have caused and we thank the reviewer for providing this comment. We have incorporated revisions throughout the manuscript to improve readability:

Reviewer #1 (Recommendations for the authors):

1. line 126: Arc1 inactivation is described as resulting in a modest reduction in B. thai survival. My reading of the figure is that Arc1 inactivation results in a modest increase in B. thai survival. Please clarify.

The reviewer is correct in that inactivation of Arc1 causes a modest increase in B. thai survival, i.e. a modest decrease in the ability of P. aeruginosa ability to kill B. thai. We regret the confusion caused by our original statement and have revised for clarity (p8, lines 143-145).

2. Line 126-127: I do not understand these comparisons. Maybe explain in a bit more detail.

We have revised the description of these experiments to provide more clarity (p. 7, line 131-p. 8, line 145)

3. Figure 2B: What is plotted on the Y-axis?

The specific metric presented in this figure was calculated by dividing the P. aeruginosa CFUs obtained for the indicated mutant by those obtained for the wild-type strain during competition with the indicated B. thai strain, and then multiplying by 100 to obtain a percentage. We have revised the figure legend to make this more clear.

Reviewer #2 (Recommendations for the authors):

The authors define the arc1-arc3 genes as part of a "bacterial innate immune system". I found this choice questionable since it is not proven that they allow self- vs non-self-discrimination. In fact, since heterologous expression of Arc3B can protect as well in Pa (presumably even Arc3B encoded in a B. thai), it's not a true self -vs non-self-discrimination. Would referring to it as an antitoxin not be suitable instead?

See the response to Essential Revision #1 above.

The underpinnings protective (antitoxin) effect of Arc3B remain murky: does it fulfill a structural role, perhaps sequestering Tle3 or interfering with its activity, stability or translocation? Or does it act via a converging pathway that removes LPE or MLCL or acts on them, in parallel to Aas? In the latter case, Arc3B should be possible to protect against phospholipid cleavage caused in other ways than Tle3-mediated intoxication. Can this be tested with another phospholipase for example (expressed in Pa for self-intoxication)? The authors have not shown that LPE and MLCL levels can be modulated by Arc3B (with or without Arc3A) in the absence of Tle3, begging the question whether Arc3B acts specifically on Tle3 or phospholipase-mediated damage in general. Conversely can Aas overexpression compensate for loss of Arc3B and/or Arc3A in response to Tle3 intoxication?

We agree with the reviewer that a mechanistic explanation for the protected effect of Arc3B remains unclear. From an evolutionary perspective, it seems unlikely that Arc3A acts specifically to provide defense against one phospholipase toxin produced by another species. However, for technical reasons discussed above (see Essential Revision #2), we have yet to identify other phospholipase-targeting toxins against which Arc3A conclusively provides protection. As requested by the reviewer, we have evaluated whether Aas over-expression can compensate for Arc3A inactivation. Our data show that Aas over-expression does not alter the compromised defense of ∆arc3A (new Figure 4—figure supplement 3 and p14, lines 362-364).

Were the Arc3B orthologs that were chosen for complementation only from species that encode Arc3A or also from those that encode a solitary Arc3B? This could be discussed.

The Arc3A homologs we evaluated for their ability to complement the arc3A deletion in P. aeruginosa were all from organisms that also encoded a homolog of Arc3B. The organisms encoding unaccompanied Arc3A homologs are all distantly related to P. aeruginosa (Gram-positives, cyanobacteria, etc., see Figure 3B) and thus we reasoned that poor expression of the genes would confound the experiment.

Arc3B interacts with Arc3A, but it does not appear to have enzymatic activity. It would be useful to know if Arc3B requires Arc3A for activity (assessed by epistasis experiments using Arc3B/A single and double mutants). One would predict that Arc3B can act independently of Arc3A since Arc3A is not always co-conserved with Arc3B, in which case the association between Arc3A and Arc3B should not be important to protect against Tle3.

The prediction of the reviewer is correct, in that Arc3B is not required for Ar3A activity. In the revised manuscript, we include the results of the suggested experiment, which demonstrates that inactivation of Arc3B alone or in combination with Arc3A does not affect P. aeruginosa competitiveness in pairwise competition assays with B. thailandensis (revised Figure 3—figure supplement 1 and p12, lines 267-276). However, we note that in our Tn-seq assays, arc3A was depleted more than 3-fold following competition with antagonistic B. thailandensis. Importantly, arc3A is encoded downstream of arc3B, thus eliminating the possibility that this phenotype results from polar effects of the transposon. Thus, it appears that Arc3A does contribute to competitiveness under some conditions, but is clearly not required for Arc3B function in P. aeruginosa.

Line 111 and panel A with B of Figure1—figure supplement 3: I don't see the cumulative effect claimed when compared the triple mutant Δarc123 (panel A) to single deletions (panel B).

The cumulative effect we are referring to here is apparent when comparing the parental final CFU counts to those of the individual mutants or the triple mutant for a given experiment. We observe a degree of variability in the growth yield of the parent strain competed against B. thai in different experiments. To highlight the difference we are referring to, we have now added the fold difference value for CFUs obtained for the mutants vs. the parent strain to each graph (revised Figure 1—figure supplement 3).

Figure 1—figure supplement 3A: the names of the deleted strains seem incorrect: Δarc1DΔarc1G should be Δarc2DΔarc2G; as well Δarc1AΔarc1B is Δarc3AΔarc3B.

The reviewer is correct and we have fixed this error in the revised manuscript.

Line 218: Protection by Arc3A is valid for Tle3 and not against Tle1 (that is also a phospholipase). Change "against interbacterial phospholipase toxins" to "against Tle3 interbacterial phospholipase toxin".

We appreciate the issue raised here. We have revised this sentence to indicate more clearly that we currently only have evidence for protection against one type of phospholipase toxin provided by Arc3A (p 12, line 266).

Appendix 2—key resources table

Reagent type (species) or resource	Designation	Source or reference	Identifiers	Additional information
Strain, strain background (Pseudomonas aeruginosa)	PAO1	Stover et al., 2000
Strain, strain background (P. aeruginosa)	PAO1 ∆pa2541	This study		See Materials and methods
Strain, strain background (P. aeruginosa)	PAO1 ∆pa2541∆pa2536	This study		See Materials and methods
Strain, strain background (P. aeruginosa)	PAO1 ∆pa4323	This study		See Materials and methods
Strain, strain background (P. aeruginosa)	PAO1 ∆pa4323∆pa4320	This study		See Materials and methods
Strain, strain background (P. aeruginosa)	PAO1 ∆pa5114	This study		See Materials and methods
Strain, strain background (P. aeruginosa)	PAO1 ∆pa5114∆pa5113	This study		See Materials and methods
Strain, strain background (P. aeruginosa)	PAO1 ∆pa2541, Tn7::AraE::pa2541	This study		See Materials and methods
Strain, strain background (P. aeruginosa)	PAO1 ∆pa4323, Tn7::AraE::pa4323	This study		See Materials and methods
Strain, strain background (P. aeruginosa)	PAO1 ∆pa5114, Tn7::AraE::pa5114	This study		See Materials and methods
Strain, strain background (P. aeruginosa)	PAO1 ∆pa2541∆pa4323∆pa5114	This study		See Materials and methods
Strain, strain background (P. aeruginosa)	PAO1 ∆gacS (∆pa0928)	LeRoux et al., 2015a
Strain, strain background (P. aeruginosa)	PAO1 ∆icmF1 (∆pa0077)	Silverman et al., 2011
Strain, strain background (P. aeruginosa)	PAO1 ∆pa5112	This study		See Materials and methods
Strain, strain background (P. aeruginosa)	PAO1 ∆pa5114,Tn7::AraE::empty	This study		See Materials and methods
Strain, strain background (P. aeruginosa)	PAO1 ∆pa5114,Tn7::AraE:: pfl_rs01995	This study		See Materials and methods
Strain, strain background (P. aeruginosa)	PAO1 ∆pa5114,Tn7::AraE:: i35_rs27920	This study		See Materials and methods
Strain, strain background (P. aeruginosa)	PAO1 ∆pa5114, Tn7::AraE::pspto_rs26695	This study		See Materials and methods
Strain, strain background (P. aeruginosa)	PAO1 ∆pa5114, Tn7::AraE::ta05_rs00690	This study		See Materials and methods
Strain, strain background (P. aeruginosa)	PAO1 ∆pa3267	This study		See Materials and methods
Strain, strain background (P. aeruginosa)	PAO1 ∆pa5114∆pa3267	This study		See Materials and methods
Strain, strain background (P. aeruginosa)	PAO1 ∆retS (∆pa4856)	Mougous et al., 2006
Strain, strain background (P. aeruginosa)	PAO1 ∆retS ∆pa3267	This study		See Materials and methods
Strain, strain background (P. aeruginosa)	PAO1 ∆retS ∆pa5114	This study		See Materials and methods
Strain, strain background (P. aeruginosa)	PAO1 ∆rpoS (∆pa3622)	Jørgensen et al., 1999
Strain, strain background (P. aeruginosa)	PAO1 ∆vacJ (∆pa2800)	Munguia et al., 2017
Strain, strain background (P. aeruginosa)	PAO1 pa2541_vsv-g	This study		See Materials and methods
Strain, strain background (P. aeruginosa)	PAO1 ∆retS, pa2541_vsv-g	This study		See Materials and methods
Strain, strain background (P. aeruginosa)	PAO1 ∆gacS, pa2541_vsv-g	This study		See Materials and methods
Strain, strain background (P. aeruginosa)	PAO1 pa4323_vsv-g	This study		See Materials and methods
Strain, strain background (P. aeruginosa)	PAO1 ∆retS, pa4323_vsv-g	This study		See Materials and methods
Strain, strain background (P. aeruginosa)	PAO1 ∆gacS, pa4323_vsv-g	This study		See Materials and methods
Strain, strain background (P. aeruginosa)	PAO1 pa5114_vsv-g	This study		See Materials and methods
Strain, strain background (P. aeruginosa)	PAO1 ∆retS, pa5114_vsv-g	This study		See Materials and methods
Strain, strain background (P. aeruginosa)	PAO1 ∆gacS, pa5114_vsv-g	This study		See Materials and methods
Strain, strain background (P. aeruginosa)	PAO1 ∆retS, pa5113_vsv-g	This study		See Materials and methods
Strain, strain background (Burkholderia thailandensis)	E264 (ATCC 700388)	Yu et al., 2006
Strain, strain background (B. thailandensis)	E264 ∆icmF1 (∆BTH_I2954)	LeRoux et al., 2015a
Strain, strain background (B. thailandensis)	E264 tle3S264A(BTH_II0090S264A + BTH_I3226S264A)	This study		See Materials and methods
Strain, strain background (B. thailandensis)	E264 tle1S267A(BTH_I2698S267A + BTH_I2701S267A)	This study		See Materials and methods
Strain, strain background (B. thailandensis)	E264 ∆colA (∆BTH_I2691)	This study		See Materials and methods
Strain, strain background (B. thailandensis)	E264 ∆tae2 (∆BTH_I0068)	Russell et al., 2012
Strain, strain background (B. thailandensis)	E264 Tn7::Tp-PS12-mCherry	LeRoux et al., 2012
Strain, strain background (B. thailandensis)	E264 tle3 sensitized strain (∆BTH_I3225-8, ∆BTH_II0089-94)	This study		See Materials and methods
Strain, strain background (B. thailandensis)	E264 Tn7::Tp-PS12-mCherry tle3 sensitized strain (∆BTH_I3225-8, ∆BTH_II0089-94)	This study		See Materials and methods
Strain, strain background (Pseudomonas protegens)	pf-5 (ATCC BAA-477)	Paulsen et al., 2005
Strain, strain background (P. protegens)	Pseudomonas protegens pf-5 ∆pfl_rs01995	This study		See Materials and methods
Strain, strain background (Escherichia coli)	DH5α	Thermo Fisher ScientificCat# 18258012
Strain, strain background (E. coli)	SM10	Biomedal LifescienceCat# BS-3303
Strain, strain background (E. coli)	UE54MG1655 lpp-2Δara714 rcsF::mini-Tn10 cam pgsA::FRT-kan-FRT	Harvat et al., 2005
Recombinant DNA reagent	pEXG2-∆pa2541 (plasmid)	This study		Deletion construct, see Materials and methods
Recombinant DNA reagent	pEXG2-∆pa2536 (plasmid)	This study		Deletion construct, see Materials and methods
Recombinant DNA reagent	pEXG2-∆pa4323 (plasmid)	This study		Deletion construct, see Materials and methods
Recombinant DNA reagent	pEXG2-∆pa4320 (plasmid)	This study		Deletion construct, see Materials and methods
Recombinant DNA reagent	pEXG2-∆pa5114 (plasmid)	This study		Deletion construct, see Materials and methods
Recombinant DNA reagent	pEXG2-∆pa5114/pa5113 (plasmid)	This study		Deletion construct, see Materials and methods
Recombinant DNA reagent	pEXG2-∆pa3267 (plasmid)	This study		Deletion construct, see Materials and methods
Recombinant DNA reagent	pEXG2-∆pa5113 (plasmid)	This study		Deletion construct, see Materials and methods
Recombinant DNA reagent	pEXG2-∆pa5112 (plasmid)	This study		Deletion construct, see Materials and methods
Recombinant DNA reagent	pEXG2-∆pfl_rs01995 (plasmid)	This study		Deletion construct, see Materials and methods
Recombinant DNA reagent	pEXG2-pa2541_vsv-g (plasmid)	This study		Construct to introduce VsvG tag, see Materials and methods
Recombinant DNA reagent	pEXG2-pa4323_vsv-g (plasmid)	This study		Construct to introduce VsvG tag, see Materials and methods
Recombinant DNA reagent	pEXG2-pa5114_vsv-g (plasmid)	This study		Construct to introduce VsvG tag, see Materials and methods
Recombinant DNA reagent	pEXG2-pa5113_vsv-g (plasmid)	This study		Construct to introduce VsvG tag, see Materials and methods
Recombinant DNA reagent	pUC18-Tn7t-pBAD-araE (plasmid)	Hoang et al., 2000		Arabinose-inducible expression system, see Materials and methods
Recombinant DNA reagent	pUC18-Tn7t-pBAD-araE-pa2541 (plasmid)	This study		Arabinose-inducible expression system, see Materials and methods
Recombinant DNA reagent	pUC18-Tn7t-pBAD-araE-pa4323 (plasmid)	This study		Arabinose-inducible expression system, see Materials and methods
Recombinant DNA reagent	pUC18-Tn7t-pBAD-araE-pa5114 (plasmid)	This study		Arabinose-inducible expression system, see Materials and methods
Recombinant DNA reagent	pUC18-Tn7t-pBAD-araE-pfl_rs01995 (plasmid)	This study		Arabinose-inducible expression system, see Materials and methods
Recombinant DNA reagent	pUC18-Tn7t-pBAD-araE-i35_rs27920 (plasmid)	This study		Arabinose-inducible expression system, see Materials and methods
Recombinant DNA reagent	pUC18-Tn7t-pBAD-araE-pspto_rs26695 (plasmid)	This study		Arabinose-inducible expression system, see Materials and methods
Recombinant DNA reagent	pUC18-Tn7t-pBAD-araE-ta05_rs00690 (plasmid)	This study		Arabinose-inducible expression system, see Materials and methods
Recombinant DNA reagent	pTNS3 (plasmid)	Choi et al., 2008
Recombinant DNA reagent	pJRC115-BTH_II0090S264A (plasmid)	This study		Point mutation construct, see Materials and methods
Recombinant DNA reagent	pJRC115-BTH_I3226S264A (plasmid)	This study		Point mutation construct, see Materials and methods
Recombinant DNA reagent	pJRC115-BTH_I2698S267A (plasmid)	This study		Point mutation construct, see Materials and methods
Recombinant DNA reagent	pJRC115-BTH_I2701S267A (plasmid)	This study		Point mutation construct, see Materials and methods
Recombinant DNA reagent	pJRC115-∆BTH_I2691 (plasmid)	This study		Deletion construct, see Materials and methods
Recombinant DNA reagent	pJRC115-∆BTH_I3225-8 (plasmid)	This study		Deletion construct, see Materials and methods
Recombinant DNA reagent	pJRC115-∆BTH_II0089-94 (plasmid)	This study		Deletion construct, see Materials and methods
Recombinant DNA reagent	pUC18T-miniTn7T-Tp-PS12-mCherry (plasmid)	LeRoux et al., 2012		Construct for mCherry expression
Software, algorithm	Seqmagick	Matsen Group, 2020, https://github.com/fhcrc/seqmagick
Software, algorithm	TRANSIT TPP	https://transit.readthedocs.io/en/latest/tpp.html
Antibody	Anti-VSV-G (rabbit polyclonal)	MilliporeSigma	Cat# V4888-200UG	Western blot (1:5000 dilution)
Antibody	Anti-rabbit IgG HRP conjugated (goat)	MilliporeSigma	Cat# A6154-1ML	Western blot (1:5000 dilution)
Antibody	Anti-ribosome polymerase β (mouse monoclonal)	BioLegend	Cat# 663903; RRID: AB_2564524	Western blot (1:1000 dilution)
Antibody	Anti-mouse IgG HRP conjugated (sheep)	MilliporeSigma	Cat# AC111P	Western blot (1:4000 dilution)