VgrG Spike Dictates PAAR Requirement for the Assembly of the Type VI Secretion System

ABSTRACT

Widely employed by Gram-negative pathogens for competition and pathogenesis, the type six protein secretion system (T6SS) can inject toxic effectors into neighboring cells through the penetration of a spear-like structure comprising a long Hcp tube and a VgrG-PAAR spike complex. The cone-shaped PAAR is believed to sharpen the T6SS spear for penetration but it remains unclear why PAAR is required for T6SS functions in some bacteria but dispensable in others. Here, we report the conditional requirement of PAAR for T6SS functions in Aeromonas dhakensis, an emerging human pathogen that may cause severe bacteremia. By deleting the two PAAR paralogs, we show that PAAR is not required for T6SS secretion, bacterial killing, or specific effector delivery in A. dhakensis. By constructing combinatorial PAAR and vgrG deletions, we demonstrate that deletion of individual PAAR moderately reduced T6SS functions but double or triple deletions of PAAR in the vgrG deletion mutants severely impaired T6SS functions. Notably, the auxiliary-cluster-encoded PAAR2 and VgrG3 are less critical than the main-cluster-encoded PAAR1 and VgrG1&2 proteins to T6SS functions. In addition, PAAR1 but not PAAR2 contributes to antieukaryotic virulence in amoeba. Our data suggest that, for a multi-PAAR T6SS, the variable role of PAAR paralogs correlates with the VgrG-spike composition that collectively dictates T6SS assembly.

IMPORTANCE Gram-negative bacteria often encode multiple paralogs of the cone-shaped PAAR that sits atop the VgrG-spike and is thought to sharpen the spear-like T6SS puncturing device. However, it is unclear why PAAR is required for the assembly of some but not all T6SSs and why there are multiple PAARs if they are not required. Our data delineate a VgrG-mediated conditional requirement for PAAR and suggest a core-auxiliary relationship among different PAAR-VgrG modules that may have been acquired sequentially by the T6SS during evolution.

INTRODUCTION

To survive in polymicrobial communities, bacteria have developed various weapon-like tools to outcompete neighboring species. One highly effective molecular weapon is the type VI secretion system (T6SS) that has been widely found in about 25% of Gram-negative bacteria (1–3). The T6SS-susceptible targets include Gram-positive and Gram-negative bacteria, yeast, amoeba, and mammalian cells (2, 4–8). These diverse functions of the T6SS largely depend on its secreted toxic effectors (9–11). Thousands of effectors have been predicted in Gram-negative bacteria, and known toxins possess broad functions, ranging from toxins targeting the cell wall, membrane, and nucleic acids to enzymes capable of cross-linking the cytoskeleton and hydrolyzing NAD(P)+, as well as metal scavenging (12–17). Concomitant with these diverse effector functions is the continuous pressure of T6SS-mediated competition on the evolution of various microbial defense mechanisms, including specific defenses through cognate immunity proteins and general defenses through the innate immunity-like stress responses (18–22).

T6SS effectors are secreted by a complex nanomachine comprising a transmembrane TssJ/L/M complex, a baseplate TssE/F/G/K, and a TssB/C-Hcp double tubular structure with a VgrG-PAAR spike complex (11, 23–25). Stacks of TssB/C hexamers form a contractile sheath that wraps around an inner tube made of Hcp hexamers (23, 26). Atop the Hcp tube sits a spike complex, made of a VgrG trimer and a cone-shaped PAAR protein (12, 24). Effector delivery can be achieved via multiple routes, including binding to the PAAR-VgrG spike complex and the Hcp inner tube (24, 27–29). Sheath contraction results in the ejection of the Hcp tube, the spike complex, as well as the associated effectors (30).

The trimeric VgrG complex is essential for T6SS assembly, even for the assembly of the highly stable noncontractile sheath variant (12, 30, 31). In contrast, PAAR, the other component of the spike, has been shown to be crucial for T6SS assembly in Acinetobacter baylyi, Serratia marcescens, and Enterobacter cloacae but dispensable in Vibrio cholerae and Aeromonas dhakensis (24, 32–34). What accounts for such difference remains elusive. A major challenge in addressing this question is that T6SS organisms often encode multiple VgrG and PAAR proteins, as well as multiple independent T6SS systems (24, 27, 30, 32). For example, the human pathogen Pseudomonas aeruginosa type strain PAO1 possesses 8 PAAR genes, 10 vgrG genes, and 3 independent T6SS clusters (35, 36). Another convoluting factor is that many PAAR and VgrG have acquired extended domains and can act as dual functional effectors as well as carriers for dedicated effectors and their chaperones (12, 13, 24, 32, 36, 37). In addition, effectors and chaperones have also been shown to be crucial for the T6SS assembly (9, 34, 38, 39).

To understand the relations of multiple PAAR and VgrG proteins, we herein employ a simple T6SS model of A. dhakensis, which is an emerging important human pathogen that can cause severe soft tissue infection and bacteremia (40, 41). The A. dhakensis type strain SSU displays a constitutively active T6SS with strong antibacterial activities by secreting three antibacterial effectors (13, 33, 42, 43). These effectors are a colicin-like TseC, a lysozyme TseP, and a nuclease TseI, whose secretion is specifically dependent on each cognate upstream-encoded VgrG protein, respectively (13, 33, 43). In addition, there are two PAAR proteins encoded in the SSU genome (33). By constructing a set of combinatorial PAAR and vgrG mutants, we systemically examined how they contribute to T6SS functions. We demonstrate that PAAR proteins do not dictate effector specificity and deletion of PAAR genes reduces but not abolishes T6SS functions. However, combinatorial deletions of PAAR and vgrG can severely impair T6SS functions, indicating that PAAR proteins are conditionally indispensable depending on the presence of VgrG proteins. Specifically, the T6SS maintains functionality in the absence of the auxiliary-cluster-encoded PAAR2 and VgrG3 but is severely impaired when lacking the main-cluster-encoded PAAR1 and either VgrG1 or VgrG2 proteins. We observed a similar conditional requirement of PAAR for the T6SS of V. cholerae as well. Using an amoeba infection model, we found that deletion of PAAR1 but not PAAR2 impairs A. dhakensis virulence to amoeba cells. Our findings highlight that PAAR and VgrG proteins are functionally interconnected in governing T6SS assembly and activities.

RESULTS

PAAR proteins are not essential for T6SS functions.

Of the two PAAR genes in A. dhakensis SSU, PAAR1 is located in the main T6SS cluster, upstream of vgrG2, while the PAAR2 gene resides in the auxiliary cluster, upstream of vgrG3 (Fig. 1A). Sequence alignment shows that their C-terminal sequences are identical while the N terminus is less conserved (Fig. 1B), and neither has an extended functional domain. Structural modeling analysis using the closest homolog VCA0105 (PDB: 4JIV) as the template shows that both PAAR proteins are predicted to form a cone-shaped structure with a flat triangular base that interacts with the top of the beta-helix of the VgrG heterotrimeric structure (Fig. 1B) (24). Notably, one of the three base fragments (D29-G34) resides in the variable N-terminal sequence, suggesting the two PAAR proteins may differ in interacting with the VgrG heterotrimer.

FIG 1.

PAAR proteins are not essential to T6SS functions. (A) T6SS gene clusters in A. dhakensis SSU. Genes that are directly involved in this study are indicated. (B) Sequence alignment and structural modeling of PAAR proteins. Sequence alignment was generated using Clustal and the structural model of PAAR1 using Phyre2 with the VCA0105 protein as a template (PDB: 4JIV). Predicted VgrG-interacting fragments are indicated in the alignment as well as in the bottom view of the PAAR1 structure. (C) Western blotting of Hcp secretion in PAAR deletion mutants. Exponential-phase growing cells were collected at OD₆₀₀ = 1 and Hcp levels were detected using a custom-made polyclonal antibody in whole cells (Cell) and secreted supernatants (Sec). The RNA polymerase beta subunit RpoB serves as a loading and cell lysis control. (D) Bacterial competition assay of the PAAR mutants against three effector-immunity defective mutants. Killer and prey strains, as indicated, are mixed at a 5:1 ratio and coincubated for 3 h on LB plates, followed by serial dilutions on selective medium to enumerate the survival of prey cells. Error bars indicate the mean +/– standard deviation and statistical significance was calculated using one-way ANOVA analysis. ****, P < 0.0001. (E) Mass spectrometry analysis of secreted proteins in wild type and T6SS mutants. Total levels of VgrG and Hcp homologs are shown as indicators for T6SS secretion activity and for comparison with the secreted levels of individual effectors.

We first examined whether delivery of effectors requires the two PAAR proteins in SSU. Western blotting of secreted Hcp levels, a hallmark of T6SS activities, shows that deletion of single or double PAAR genes attenuated but did not abolish Hcp secretion, indicating that PAAR proteins are not essential for T6SS (Fig. 1C). Hcp was secreted at a lower level in the ΔPAAR1 mutant than in the ΔPAAR2 mutant, suggesting PAAR1 is more critical (Fig. 1C). To test whether PAAR dictates the specificity of effector delivery, we performed competition assays using PAAR deletion mutants and the three previously constructed effector-specific immunity-deficient mutants (13, 33, 43). Results show that all three effector-immunity mutants were efficiently killed by the PAAR deletion mutants (Fig. 1D).

To confirm that effectors were secreted independent of PAAR, we used an LC-MS/MS secretome approach by comparing all secreted proteins of the wild type, individual ΔvgrG mutants, and the ΔPAAR1&2 double deletion mutant (Fig. 1E). Results show that effectors were detected in the wild type and the ΔPAAR1&2 mutant but not in their corresponding ΔvgrG mutants. There was background cell lysis as evidenced for the detection of Hcp in the ΔvasK (also known as ΔtssM) mutant and cytosolic RNA polymerase subunit RpoC in all samples. The known secreted flagellar hook-associated protein FlgK was detected at comparable levels across all samples, suggesting good quality of sample preparation. Collectively, these results indicate PAAR1 and PAAR2 are not essential for T6SS secretion or specificity of effector delivery.

PAAR proteins are critical for T6SS functions in vgrG mutants.

Because PAAR sits on top of the VgrG trimer (24), we next tested whether PAAR proteins work in concert with VgrG proteins for effector delivery. By introducing PAAR gene deletions to the previously constructed ΔvgrG mutants (33, 43), we made a panel of PAAR and vgrG single and combinatorial deletion mutants and tested their killing activities against an E. coli prey (Fig. 2A). Results show that single or double deletions of the PAAR genes and single deletion of the vgrG genes did not result in significant changes in the killing efficiency of the T6SS against the E. coli prey. However, when the PAAR1 deletion was introduced to any of the ΔvgrG mutants, the killing activities against E. coli were severely impaired in comparison with the single ΔPAAR or ΔvgrG mutants. When the PAAR2 deletion was introduced to those three ΔvgrG mutants, similar defects in killing E. coli were observed except for the ΔPAAR2 ΔvgrG3 mutant that still exhibited strong E. coli killing (Fig. 2A). In addition, when each of the individual vgrG deletions was introduced to the double PAAR deletion mutant, only the ΔPAAR1/2 ΔvgrG3 mutant showed intermediate T6SS killing, while the other two mutants showed abolished killing against the E. coli prey (Fig. 2B).

FIG 2.

Effect of combinatorial PAAR and vgrG deletions on T6SS functions. (A) Competition assay of wild type (WT), the T6SS-null ΔvasK mutant, and PAAR/vgrG deletion mutants against an E. coli prey. Survival of E. coli was enumerated after coincubation with the indicated SSU strains. (B) Competition assay of wild type (WT), the ΔvasK mutant, and the double PAAR/vgrG deletion mutants against E. coli. Survival of E. coli was enumerated after coincubation with the indicated SSU strains. (C) Western blotting of Hcp secretion in deletion mutants of PAAR/vgrG. (D) Western blotting of Hcp secretion in the double deletion mutants of PAAR/vgrG. (C and D) Exponential-phase growing cells were collected at OD₆₀₀ = 1 and Hcp levels were detected using a custom-made polyclonal antibody in whole cells (Cell) and secreted supernatant (Sec). The RNA polymerase beta subunit RpoB serves as a loading and cell lysis control. (A and B) Error bars indicate the mean +/– standard deviation and statistical significance was calculated using one-way ANOVA analysis. All assays were repeated at least once. ****, P < 0.0001; ns, not significant.

Because competition assays may be affected by both effector delivery and prey defense (18, 19, 44), we compared Hcp secretion of these PAAR and vgrG deletion mutants using Western blotting. Results of secretion assays are consistent with the competition results. Of the six PAAR vgrG double deletion mutants, only the ΔPAAR2 ΔvgrG3 mutant showed a strong wild-type level of Hcp secretion, while Hcp secretion was substantially reduced in the other mutants (Fig. 2C). In mutants lacking both PAAR genes and each of the three vgrG genes, we found that secretion of Hcp was reduced to an intermediate level in the ΔvgrG3 ΔPAAR1&2 mutant and was barely detectable in the ΔvgrG1 and the ΔvgrG2 backgrounds (Fig. 2D). As a control, all mutants showed comparable cytosolic Hcp levels and the cytosolic RpoB, the beta subunit of RNA polymerase, was not detectable in the secreted samples (Fig. 2C and D). These results collectively suggest that PAAR is conditionally required for T6SS assembly depending on the composition of the VgrG complex.

Ectopic expression of PAAR1 but not PAAR2 restores T6SS activities.

Because PAAR1 and PAAR2 reside upstream of vgrG2 and vgrG3, respectively, it is possible that the above-mentioned T6SS defects of the double deletion PAAR vgrG mutants result from polar effects of PAAR gene deletions. To test this, we expressed PAAR1 and PAAR2 using the arabinose-inducible pBAD plasmid in the different PAAR vgrG double deletion mutants and compared their T6SS activities in competition assays. Results show that expression of PAAR1 significantly increased the killing against prey E. coli cells in all three ΔPAAR1 vgrG mutants, relative to the corresponding mutants expressing the empty plasmid control (Fig. 3A). In contrast, expression of PAAR2 had little effect in the three ΔPAAR2 vgrG mutants (Fig. 3A). Next, we tested the effect of expressing individual VgrG proteins in the three ΔPAAR2 vgrG mutants. Competition assays against the E. coli prey show that expression of VgrG1 did not increase the killing activities of the ΔPAAR2 ΔvgrG1 and expression of VgrG2 only moderately increased the killing activities of the ΔPAAR2 ΔvgrG2 (Fig. 3B). The T6SS-active ΔPAAR2 ΔvgrG3 mutant was not affected by expression of either PAAR2 or VgrG3 (Fig. 3A and B). Collectively, these results show that ectopic expression of PAAR1 could complement the three PAAR1 vgrG double deletions. The ineffective complementation with either PAAR2 or VgrG1/2 suggests that the T6SS defect of PAAR2 vgrG1/2 mutants may not be simply caused by polar effects of PAAR2 deletion on its downstream vgrG3 but, at least partially, due to disturbed stoichiometry and composition of the PAAR-VgrG-effector complexes.

FIG 3.

Complementation by PAAR1 but not PAAR2 restores T6SS activities. (A) Competition assay of PAAR/vgrG deletion mutants carrying either an empty vector (pBAD) or an arabinose-inducible PAAR1 or PAAR2 plasmid. Survival of the E. coli prey was enumerated after coincubation with the indicated killer strains. (B) Competition assay of PAAR/vgrG deletion mutants carrying either an empty vector (pBAD) or an arabinose-inducible VgrG plasmid. Survival of E. coli was enumerated after coincubation with the indicated killer strains. The competition assays in both A and B were performed similarly as in Fig. 2 except that 0.1% arabinose was included in the LB agar during competition. Error bars indicate the mean +/– standard deviation and statistical significance was calculated using one-way ANOVA analysis. All assays were repeated at least once. ****, P < 0.0001; ns, not significant.

PAAR1 but not PAAR2 contributes to A. dhakensis virulence in amoeba.

Because the T6SS is essential for A. dhakensis SSU virulence in the phagocytosis Dictyostelium discoideum (Dicty) model (33), we tested whether PAAR proteins contribute to such virulence. As expected, the number of Dicty cells required to form a plaque on a bacterial lawn was substantially higher for wild-type SSU than the ΔvasK mutant. Interestingly, although single or double deletion of PAAR had little effect on antibacterial activities (Fig. 1D), deletion of PAAR1 or both PAAR1 and PAAR2 substantially reduced the virulence against D. discoideum while the ΔPAAR2 remained as virulent as the wild-type cells (Fig. 4). These results suggest that PAAR1 but not PAAR2 is critical for T6SS-mediated antieukaryotic functions.

FIG 4.

PAAR1 is critical to anti eukaryotic functions. A dilution series of Dicty cells were plated on bacterial lawns containing SSU and the derivative mutants, individually. The number of Dicty cells at each dilution was indicated. A clear plaque indicates that Dicty cells can survive and feed on the bacteria, and the minimal number of Dicty cells to form plaques directly correlate with the antieukaryotic virulence of bacteria.

PAAR is conditionally required for T6SS functions in V. cholerae.

We next asked why different PAAR proteins contribute to T6SS functions differently. We noticed that PAAR2 is located on a separate cluster distant from the PAAR1 and T6SS main cluster (Fig. 1A). Based on our observations, we postulate that the main-cluster-associated PAAR be considered a core PAAR that plays a more important role in dictating T6SS than the other PAARs encoded on auxiliary islands for multi-PAAR T6SSs. To test whether this is applicable to other T6SSs, we examined PAAR-VgrG functional dependence in the T6SS-active V. cholerae strain V52. The genome of V52 encodes two PAAR proteins and three VgrG proteins that reside in one main cluster and three auxiliary clusters, with the PAAR1 gene and the vgrG3 gene associated with the main cluster and the PAAR2 and the other two vgrG genes on separate distant clusters (2, 24) (Fig. 5A).

FIG 5.

Effect of combinatorial PAAR and vgrG deletions in V. cholerae. (A) Organization of T6SS genes in V. cholerae V52. Genes encoding PAAR, VgrG, and Hcp proteins are highlighted in color. (B) Competition assay of wild type (WT), the T6SS-null ΔvasK mutant, and PAAR/vgrG deletion mutants against an E. coli prey. Survival of E. coli was enumerated by serial dilutions after coincubation with the indicated V. cholerae strains. Error bars indicate the mean +/– standard deviation and statistical significance was calculated using one-way ANOVA analysis. ***, P < 0.001; ****, P < 0.0001; ns, not significant. No prey survival was detected in the WT and ΔPAAR1 samples and thus an arbitrary value of one colony at the highest concentration was used for statistical analysis. (C) Secretion of Hcp in different mutants. Secretion samples were prepared similarly from exponentially growing cells at OD₆₀₀ = 1. Hcp was detected with a custom-made V. cholerae Hcp antibody in whole cells and the supernatants.

Based on their gene localization, we hypothesized that PAAR1 would contribute more to T6SS functions than PAAR2. To test this, we introduced single and double deletion of PAAR genes to the ΔvgrG1 and the ΔvgrG3 mutants (VgrG2 is required for T6SS secretion in V. cholerae and thus not deleted) (12, 45). Using bacterial competition assays against an E. coli prey, we compared the effect of PAAR and vgrG deletions on T6SS-mediated killing. We confirmed that double PAAR deletions attenuated T6SS-mediated killing (24) to a level comparable to that of single deletion of vgrG1 or vgrG3 (Fig. 5B). Deletion of PAAR1 in the ΔvgrG1 mutant significantly reduced its antibacterial activity while deletion of PAAR2 did not. Similarly, the deletion of PAAR1 but not the PAAR2 in the ΔvgrG3 mutant also significantly reduced the ΔvgrG3 killing activity. In addition, double deletion of both PAAR genes resulted in severely impaired T6SS killing (Fig. 5B). The impaired secretion of Hcp in these double PAAR mutants lacking either vgrG1 or vgrG3 correlated well with the interbacterial competition data (Fig. 5C). These results support that PAAR1 is a core PAAR and agree well with the observations in SSU that the conditional essentiality of PAAR proteins for T6SS secretion depends on the types of VgrG proteins available for T6SS assembly.

DISCUSSION

Previous studies have revealed multiple functions of PAAR that include being a core structural component (24, 32), an effector itself (16, 32), an effector carrier (27), and a sorting control for specific VgrG secretion (32, 35). The large number of PAAR-VgrG pairs and the dual role of PAAR/VgrG-extended effectors have made deciphering the role of PAAR in T6SS assembly challenging. Unlike the T6SSs in most other species, all PAAR and VgrG proteins in A. dhakensis SSU possess just the canonical structural domains, making its T6SS one of the simplest models for examining PAAR-VgrG relations. Here, we demonstrate the conditional dependence of PAAR for T6SS functions in A. dhakensis SSU as well as in V. cholerae V52, depending on the composition of the VgrG complex. In both strains, PAAR is crucial for T6SS assembly in single vgrG deletion mutants except for the SSU ΔvgrG3 mutant, in which the VgrG complex composed of VgrG1 and VgrG2 is still capable of supporting T6SS assembly at a moderate efficiency (Fig. 6). Collectively, our findings suggest the indispensable role of PAAR in some T6SSs is dictated by the types of the VgrG spikes in those species.

FIG 6.

Schematic of conditional requirement of PAAR and VgrG for T6SS functions. Top panel summarizes key findings for the T6SS assembly in A. dhakensis. Different composition of the VgrG-PAAR spike results in differential T6SS activities. For the severely impaired group, PAAR and VgrG may not form a stable complex themselves or with the other essential structural components to promote assembly. Bottom panel depicts a hypothetical model for sequential acquisition and diversification of PAAR and VgrG. For multi-PAAR/VgrG T6SSs, there is likely a core PAAR/VgrG module associated with the main T6SS components. This original PAAR-VgrG is strengthened by auxiliary PAAR-VgrG proteins encoded by distant clusters that may have been acquired separately. Some PAAR-VgrG proteins may carry extended functional domains. Although these auxiliary PAAR-VgrG components work together with the core but cannot replace the core PAAR-VgrG functions, they promote diversification of the PAAR-VgrG spike and expansion of T6SS functions.

In addition to previous findings that different PAAR-VgrG effector groups can compete for secretion in several organisms (27, 32, 35), we find that different PAAR and VgrG pairs collectively contribute to total T6SS activities in A. dhakensis. This is supported by the observations that all PAAR and vgrG deletion pairs severely impaired T6SS functions except for the deletions of PAAR2 vgrG3 that moderately attenuated T6SS secretion. PAAR1 appears to play a more important role in T6SS secretion than PAAR2. In the amoeba infection model, PAAR1 is also distinctive from PAAR2 for its crucial role in virulence. Notably, both PAAR2 and vgrG3 genes are located in the same auxiliary cluster. In V. cholerae, the PAAR1 cluster is located immediately upstream of the main T6SS cluster encoding structural proteins, while the PAAR2 cluster is isolated from the rest of T6SS structural gene clusters (Fig. 5A). Deletion of PAAR1 further reduced killing abilities of the vgrG1 or the vgrG3 mutants while the PAAR2 deletion had little effect. It is possible that such a VgrG-determined PAAR requirement is generally applicable to more complex multi-PAAR/VgrG T6SSs as well.

Building on these results and our earlier findings that double vgrG deletion mutants cannot assemble T6SS (39), we propose a model to illustrate the differential effects of core and auxiliary PAAR and VgrG paralogs on T6SS assembly (Fig. 6). The VgrG trimeric spike is absolutely required for the T6SS assembly, while PAAR interacts with the VgrG spike. For any given multi-PAAR/VgrG T6SSs, there is at least a core PAAR-VgrG module that is sufficient to support the T6SS assembly and has likely coevolved with the main structural components. The incorporation of distantly encoded auxiliary PAAR/VgrG proteins provides complementary functions to the T6SS assembly as well as contributes to the diversification of T6SS functions, likely through evolved PAAR/VgrG with extended domains or interacting effectors. Future biochemical and structural studies are required to reveal the molecular details of PAAR and VgrG interactions. Nonetheless, from an evolutionary perspective, although the T6SS main or original cluster may have encoded all necessary components for T6SS assembly, acquisition and integration of the auxiliary yet nonessential clusters promote T6SS activities and competitive fitness.

MATERIALS AND METHODS

Bacterial strains and growth conditions.

All strains, plasmids, and primers are listed in Table S1 in the supplemental material. Strains were routinely grown in Lysogeny Broth (LB) media ([wt/vol] 1% tryptone, 0.5% yeast extract, 0.5% NaCl) aerobically at 37°C. Antibiotics were supplemented at the following concentrations whenever appropriate: streptomycin (100 μg/mL), ampicillin (100 μg/mL), kanamycin (50 μg/mL), gentamicin (20 μg/mL), chloramphenicol (25 μg/mL for E. coli, 2.5 μg/mL for SSU). Plasmids were constructed using standard molecular techniques, and sequence confirmed by Sanger sequencing. All plasmids and primers are available upon request.

Bioinformatics analysis and homology modeling.

All gene sequences of A. dhakensis SSU are retrieved from the draft genome assembly (GenBank: NZ_JH815591.1) and sequence corrected by Sanger sequencing for unresolved nucleotide sequences.

Both PAAR proteins were subject to structural modeling with the template PAAR protein VCA0105 (PDB: 4JIV) with 100% confidence and 53% identity for PAAR1 and 41% identity for PAAR2 using the Phyre2 analysis (46). Sequences were aligned using Clustal Omega (47) and the alignment was visualized using ESPript with default settings (https://espript.ibcp.fr) (48). Structural models were analyzed using Chimera (49).

Protein secretion assay.

Secretion assay was performed as previously described (33). Briefly, aerobically grown LB cultures were collected at an OD₆₀₀ of 1 by centrifugation at 2,500 × g. Pellets were resuspended in fresh LB and incubated at 30°C for 1 h. Cells were collected again by centrifugation at 2,500 × g for 8 min. Cell pellets were resuspended in SDS-loading dye and the supernatants were centrifuged again to remove residue cells and then precipitated in 20% [vol/vol] TCA (trichloroacetic acid) at −20°C for 20 min. Precipitated proteins were collected by centrifugation at 15,000 × g for 30 min at 4°C, washed twice with acetone, and air-dried, followed by the addition of SDS-loading dye. Samples with the SDS-loading dye were boiled for 10 min before SDS-PAGE analysis. For secretome analysis, LC-MS/MS analysis of excised bands was performed at the Southern Alberta Mass Spectrometry core facility.

Western blotting.

Proteins were subject to separation on an SDS-PAGE gel and transferred to a PVDF membrane (Bio-Rad). The membrane was blocked with 5% [wt/vol] nonfat milk in TBST (50 mM Tris, 150 mM NaCl, 0.05% [vol/vol] Tween 20, pH 7.6) for 1 h at room temperature, and incubated with primary and secondary antibodies, sequentially. Signals were detected using the Clarity ECL solution (Bio-Rad). Monoclonal antibodies were purchased from Biolegend (RpoB, product no. 663905). RpoB, the RNA polymerase beta subunit was used as a control for equal loading and cell lysis wherever appropriate. The polyclonal antibody to Hcp was custom-made by Shanghai Youlong Biotech (43). The secondary HRP-linked antibodies were purchased from Cell Signaling Technology (CST, product no. 7076S and 7074S, respectively).

Bacterial competition assay.

Killer and prey strains were grown aerobically in liquid LB to an OD₆₀₀ of 1 and 2, respectively. Cells were collected by centrifugation and gently resuspended in fresh LB. Killer and prey cells were mixed at a ratio of 5:1 and then spotted on LB-agar plates, followed by 3 h coincubation at 37°C. For competition assays involving arabinose-inducible constructs, 0.1% [wt/vol] l-arabinose was added to the LB-agar plates during the killer-prey coincubation. The mixed cells were recovered in fresh LB and vortexed vigorously to release cells from agar, and the mixtures were serially diluted by 10-fold and plated on appropriate antibiotic-containing plates selecting for killer and prey survival, respectively. All competition assays were performed in biological triplicates and repeated at least once.

Amoeba plaque assay.

Killing of amoeba was performed as previously described (45, 50). Bacterial cells were grown in LB overnight at 37°C, centrifuged at 10,000 × g for 1 min, and resuspended in phosphate buffered saline (PBS). Cells (100 μL of 10⁸ CFU/mL) were then plated on SM/5 plates at room temperature. D. discoideum (Dicty) cells were grown as axenic culture in HL5 medium and collected at the density of about 1.5 × 10⁶ cells/mL. Dicty cells were washed by centrifugation in PBS and resuspended to a final concentration of 1 × 10⁷ cells/mL. A series of 3-fold dilutions of D. discoideum cells were plated on top of the bacterial lawn on the SM/5 plates. Plates were incubated at 22°C for 3 to 7 days and plaques were monitored daily. The minimal number of D. discoideum cells required to form plaque on the bacterial lawn is indicated.

Data availability.

The data that support the findings of this study are available within the paper or available from the corresponding author upon reasonable request.