The Legionella autoinducer LAI-1 is delivered by outer membrane vesicles to promote interbacterial and interkingdom signaling

Legionella pneumophila is an environmental bacterium, which replicates in amoeba but also in macrophages, and causes a life-threatening pneumonia called Legionnaires’ disease. The opportunistic pathogen employs the α-hydroxy-ketone compound Legionella autoinducer-1 (LAI-1) for intraspecies and interkingdom signaling. LAI-1 is produced by the autoinducer synthase Legionella quorum sensing A (LqsA), but it is not known, how LAI-1 is released by the pathogen. Here, we use a Vibrio cholerae luminescence reporter strain and liquid chromatography-tandem mass spectrometry to detect bacteria-produced and synthetic LAI-1. Ectopic production of LqsA in Escherichia coli generated LAI-1, which partitions to outer membrane vesicles (OMVs) and increases OMV size. These E. coli OMVs trigger luminescence of the V. cholerae reporter strain and inhibit the migration of Dictyostelium discoideum amoeba. Overexpression of lqsA in L.pneumophila under the control of strong stationary phase promoters (PflaA or P6SRNA), but not under control of its endogenous promoter (PlqsA), produces LAI-1, which is detected in purified OMVs. These L. pneumophila OMVs trigger luminescence of the Vibrio reporter strain and inhibit D. discoideum migration. L. pneumophila OMVs are smaller upon overexpression of lqsA or upon addition of LAI-1 to growing bacteria, and therefore, LqsA affects OMV production. The overexpression of lqsA but not a catalytically inactive mutant promotes intracellular replication of L. pneumophila in macrophages, indicating that intracellularly produced LA1-1 modulates the interaction in favor of the pathogen. Taken together, we provide evidence that L. pneumophila LAI-1 is secreted through OMVs and promotes interbacterial communication and interactions with eukaryotic host cells.

The gram-negative bacterium Legionella pneumophila is an opportunistic pathogen, which upon inhalation of bacterialaden aerosols replicates in lung macrophages and causes a life-threatening pneumonia termed Legionnaires' disease (1-3).
Gram-negative bacteria form and release so-called outer membrane vesicles (OMVs), which are spherical vesicles of ca. 10 to 300 nm diameter confined by a membrane bilayer with lipopolysaccharide in the outer leaflet (38,39).OMVs are produced and released by planktonic, biofilm, and intracellular bacteria under different physiological conditions, and they transport proteins, hydrophobic small molecules, and nucleic acids between bacteria or between bacteria and eukaryotic cells (40,41).OMVs adopt a plethora of functions, including disposal of "waste" material (proteins, lipids, and peptidoglycan), secretion of virulence factors (toxins, proteases, and lipases), nutrient scavenging (carbon sources, and micronutrients: e.g., iron), inactivation of antibiotics, titration of phages, transport of signaling molecules, as well as delivery of DNA and regulatory RNA (42).While the production of OMVs comes with many advantages for the producing bacteria, it also comes with a price, as OMVs are complex and energetically costly to produce, and they may elicit adverse immune responses (43).
L. pneumophila produces OMVs, which are enriched in virulence-relevant proteins such as proteolytic and lipolytic enzymes, the macrophage infectivity potentiator (Mip) protein, Icm/Dot components and substrates, as well as the major flagellum component flagellin (44,45).These OMVs can fuse with eukaryotic membranes, but do not kill host cells and rather promote the growth of amoeba, activate mammalian cells, and modulate their cytokine response (44,(46)(47)(48).The host cell's innate immune response is also targeted by small RNAs delivered by L. pneumophila OMVs (49).Moreover, L. pneumophila OMVs bind to and are internalized by macrophages (46,50), where-independently of the Icm/Dot type IV secretion system-they inhibit phagosome-lysosome fusion (51) and promote intracellular bacterial replication at later time points of infection (47).
In this study, we tested the hypothesis that the L. pneumophila signaling compound LAI-1 partitions to OMVs and is secreted by these vesicles.We provide evidence that LAI-1 partitions to and is secreted by OMVs formed in either Escherichia coli or L. pneumophila overexpressing the autoinducer synthase gene lqsA.These OMVs mediate intrabacterial as well as interkingdom communication.Finally, the overexpression of lqsA, but not a catalytically inactive form, promotes intracellular replication of L. pneumophila in macrophages.

LAI-1 detection by a Vibrio reporter strain and LC-MS/MS
The L. pneumophila signaling compound, LAI-1, has been identified as 3-hydroxypentadecane-4-one by liquid chromatography-tandem mass spectrometry (LC-MS/MS) upon overexpression of lqsA controlled by the P tac promoter in E. coli and L. pneumophila (21).Under the conditions tested, L. pneumophila strains did not produce endogenous LAI-1 detectable by mass spectrometry (MS) or by the V. cholerae αhydroxyketone luminescence reporter strain MM920 (21) (Fig. 2A).Upon P tac -controlled overexpression of lqsA in E. coli, Legionella micdadei or L. pneumophila growing on charcoal Figure 1.The Legionella pneumophila Lqs system:production, detection, and release of LAI-1.The Lqs (Legionella quorum sensing) system produces, detects, and responds to the small signaling molecule LAI-1 (Legionella autoinducer-1, 3-hydroxypentadecane-4-one).The system comprises the autoinducer synthase LqsA, the cognate membrane-bound sensor kinases LqsS and LqsT, and the prototypic response regulator LqsR.LAI-1 produced by LqsA partitions into outer membrane vesicles (OMVs), through which the hydrophobic compound is released from the bacteria to promote interbacterial as well as interkingdom signaling.or cqsA (pTS-6) under control of the P tac promoter were grown on CYE agar in 24 well plates for 2 days, before the V. cholerae reporter strain MM920 was streaked out in parallel.After another day, growth was assessed (upper panels), and bioluminescence was determined with a luminometer (lower panels).V. cholerae MM920 was treated with the concentrations of synthetic LAI-1 indicated, DMSO was included as negative control, and luminescence intensity was yeast extract agar plates, the V. cholerae strain MM920 detected a signal produced by E. coli and L. micdadei, but not by several L. pneumophila strains (JR32, Corby, clinical isolate #883) (Fig. 2A).However, upon growing in N-(2-acetamido)-2-aminoethanesulfonic acid-buffered yeast extract (AYE) broth, L. pneumophila overexpressing lqsA produced a signal, which was detected by the V. cholerae reporter strain in culture supernatants, albeit only during a short period of time during growth (Fig. S1).Contrarily, the signal produced by L. pneumophila overexpressing cqsA was robustly detected by the V. cholerae reporter strain, regardless of whether L. pneumophila was grown on agar plates or in broth (Figs.2A and S1).

LAI-1 secretion by outer membrane vesicles
Given the apparently low amounts of LAI-1 produced by L. pneumophila, we next tested whether and with which sensitivity the V. cholerae reporter strain detects LAI-1.To this end, 1 to 50 μM synthetic LAI-1 was added to the reporter strain and luminescence was measured (Fig. 2, B and C). 1 to 20 μM synthetic LAI-1 triggered the luminescence response by the V. cholerae reporter strain in a dose-dependent manner, indicating that the reporter strain indeed recognizes the L. pneumophila signaling molecule.Fifty micromolar of LAI-1 interfered with the growth of V. cholerae, and accordingly, the luminescence emitted by the reporter strain was reduced.The detection limit for LAI-1 of the V. cholerae luminescence reporter assay was ca. 1 μM (Fig. 2C).
Synthetic LAI-1 was also detected by LC-MS/MS (Figs.  and S2).The detection limit of the LC-MS/ MS approach for LAI-1 was ca. 10 fmol corresponding to 1 μl of a 10 nM standard solution.

LAI-1 partitions to OMVs of E. coli ectopically expressing lqsA
To assess whether LAI-1 affects OMV formation and/or partitions into OMVs, we ectopically expressed lqsA in E. coli under the control of the P tac promoter.OMVs were isolated from bacterial culture supernatants and enriched by ultracentrifugation as outlined in the Experimental procedures section (Fig. 3A).While E. coli expressing lqsA produced an OMV population ranging from approximately 10 to 140 nm in diameter, the control strain produced a population of OMVs, the majority of which ranged from 10 to 60 nm in diameter (Fig. 3B).Hence, the ectopic production of LqsA in E. coli seems to increase the average OMV size and thus to possibly affect the formation of OMVs.
We then exposed the V. cholerae reporter strain MM920 to OMV preparations purified from E. coli expressing lqsA or not (Fig. 3C).OMV preparations (50-200 μg protein/ml) purified from E. coli expressing lqsA resulted in a dose-dependent luminescence signal, while OMV samples prepared from the E. coli control strain or untreated reporter strain did not produce any luminescence (Fig. 3D).LAI-1 was also detected by LC-MS/MS in OMV preparations from E. coli expressing lqsA (Fig. 3E) but not from the E. coli control strain (Fig. 3F, Table 1).
Previously, we have shown that synthetic LAI-1 at a concentration of 10 μM promotes the expression of P flaA -gfp and motility in L. pneumophila (33).Based on these findings, we sought to test whether OMVs purified from E. coli expressing lqsA also induced the expression of gfp under the control of P flaA in L. pneumophila.Using this readout, OMV samples prepared from the E. coli strain producing LqsA induced the expression of P flaA -gfp in L. pneumophila in a dose-dependent manner (Fig. 3G).In contrast, OMVs from the E. coli strain harboring the empty plasmid barely induced the expression of P flaA -gfp in L. pneumophila.The expression of P flaA -gfp above background levels by OMVs from the E. coli strain harboring the empty plasmid is likely due to unidentified components of the E. coli OMVs.Accordingly, not only the V. cholerae reporter strain, but also L. pneumophila detects and responds to OMV-associated LAI-1.Taken together, the ectopic production of LqsA in E. coli produces LAI-1, which partitions to OMVs, affects OMV formation of the donor strain, and promotes interbacterial communication.

E. coli OMVs containing LAI-1 inhibit the migration of D. discoideum
The migration of D. discoideum is reduced by pretreatment with 10 μM synthetic LAI-1 (37) (Fig. S3).Based on this finding, we asked the question, whether E. coli OMVs containing LAI-1 would exhibit a comparable effect and promote interkingdom signaling.To this end, we treated D. discoideum amoeba with OMVs purified from E. coli expressing lqsA or not (Fig. 4).Compared to the (2-(N-morpholino)ethanesulfonic acid)-buffered (MB) medium control (Fig. 4A), the migration of the amoeba was inhibited in a dose-dependent manner by OMVs purified from E. coli expressing lqsA (Fig. 4B), but not by OMVs from E. coli harboring an empty plasmid (Fig. 4C).At the highest OMV concentration used (500 μg/ml protein), the velocity of the amoeba was reduced ca.3-fold compared to the controls (Fig. 4D).Accordingly, OMV-associated LAI-1 not only promotes interbacterial communication but also interkingdom signaling.

L pneumophila OMV production is regulated by LqsA and LAI-1
Similar to E. coli (Fig. 3B), the production of OMVs by L. pneumophila might be affected by LAI-1.In order to maximize the potential effect of LqsA and LAI-1 production on OMV formation, we thought to correlate P lqsA activity with OMV formation.In L. pneumophila, lqsA (25) and P lqsA (Fig. S4) are expressed from early stationary throughout later stationary growth phase, and therefore, we routinely isolated OMVs from L. pneumophila in stationary growth phase.
To test whether LqsA affects OMV formation of L. pneumophila, we first compared the OMV populations produced by the parental L. pneumophila strain JR32 or the ΔlqsA mutant strain (Fig. 5A).The strains JR32 and ΔlqsA produced similar OMV populations, with individual OMVs LAI-1 secretion by outer membrane vesicles ranging in size from approximately 20 to 140 nm in diameter and a median diameter of 62 nm (Fig. 5B).Hence, under the conditions used, the absence of LqsA does not seem to affect the formation of OMVs in L. pneumophila.We then exposed the V. cholerae reporter strain MM920 to OMV preparations purified from L. pneumophila JR32 or ΔlqsA (Fig. 5C).Strikingly, none of the OMV preparations tested at a concentration of 50 to 200 μg/ml produced a luminescence signal above the background level of the untreated reporter strain (Fig. 5D).Similarly, LAI-1 was also not detected by LC-MS/MS in OMV preparations from strains JR32 (Fig. 5E) or ΔlqsA (Fig. 5F, Table 1).Taken together, under the conditions tested, L. pneumophila does not produce and purified OMVs do not contain detectable levels of LAI-1.
To increase the levels of LAI-1 produced by L. pneumophila, we sought to express lqsA from known strong promoters, which initially were tested as transcriptional fusions with gfp.Compared to the promoter P lqsA , the stationary phase promoters P flaA and P 6SRNA indeed led to much higher levels of GFP (Fig. 6A).Moreover, judged from the production of GFP, P flaA and P 6SRNA are strongly expressed in the early or later stationary growth phase, respectively, while P lqsA is barely expressed (Fig. 6B).
Given the features of these promoters, we expressed lqsA under the control of P flaA or P 6SRNA and compared the constructs to lqsA expressed under the control of its endogenous promoter P lqsA .The resulting L. pneumophila strains expressing lqsA (WT or a catalytically inactive mutant) under the control of P lqsA , P flaA , or P 6SRNA showed the same morphology and grew indistinguishably in AYE broth (Fig. S5).
L. pneumophila expressing lqsA under the control of its endogenous promoter P lqsA (Fig. 6C) produced an OMV population similar to the corresponding strain overproducing a catalytically inactive LqsA mutant.The OMVs ranged in size from approximately 10 to 150 nm diameter.L. pneumophila expressing lqsA under the control of P flaA (Fig. 6D) or P 6SRNA (Fig. 6E) produced OMV populations with individual OMV ranging in size from approximately 10 to 100 nm in diameter.The median OMV size was similar in strains expressing lqsA or lqsA K258A from the same promoter in the case of P lqsA and P flaA : 67 nm/62 nm (P lqsA ), and 37 nm/32 nm (P flaA ).However, the median size of OMVs produced by L. pneumophila expressing lqsA under the control of P 6SRNA was significantly smaller than the control (34 nm/45 nm).Overall, the OMVs produced by L. pneumophila expressing lqsA from its endogenous promoter P lqsA were bigger than the OMVs from the strains overproducing LqsA under the control of P flaA or P 6SRNA .This was observed independently of the catalytic activity of LqsA, and the reason is unknown.
Based on the finding that overexpression of lqsA under the control of P 6SRNA results in smaller OMVs, we tested whether synthetic LAI-1 affects OMV formation.To this end, we added 10 μM LAI-1 to L. pneumophila JR32 or ΔlqsA growing in mid-logarithmic phase.Under these conditions, the median size of OMVs from L. pneumophila treated with LAI-1 (Fig. 6F) was significantly smaller than OMVs from untreated bacteria (Fig. 6G): JR32 -47 nm/62 nm, ΔlqsA -42 nm/58 nm.Therefore, exogenously added synthetic LAI-1 reduced the median OMV size of L. pneumophila.Taken together, the L pneumophila OMV size distribution and possibly production is affected by the overexpression of lqsA or by exogenously added synthetic LAI-1.

LAI-1 partitions to OMVs of L. pneumophila overexpressing lqsA
To assess whether LAI-1 produced by L. pneumophila strains overexpressing lqsA from different promoters partitions into OMVs, we exposed the V. cholerae reporter strain MM920 to the corresponding OMV preparations (Fig. 7A).OMV preparations (50-250 μg protein/ml) purified from L. pneumophila strains expressing lqsA under control of P flaA or P 6SRNA yielded robust, dose-dependent luminescence signals with a detection limit of ca. 100 μg protein/ml (Fig. 7B).In contrast, L. pneumophila strains expressing lqsA under control of its endogenous promoter P lqsA did not elicit any luminescence from the reporter strain.Accordingly, the overexpression of lqsA under control of strong promoters produces LAI-1, which partitions to OMVs and activates quorum sensing (luminescence) in a Vibrio reporter strain.
Analogous results were obtained upon detection of LAI-1 by LC-MS/MS (Fig. 7C, Table 1).LAI-1 was robustly detected and identified in OMV preparations from L. pneumophila expressing lqsA under control of P flaA or P 6SRNA , but barely from an L. pneumophila strain expressing lqsA under control of P lqsA .Finally, the expression of a catalytically inactive lqsA mutant under control of any of these promoters did not yield LAI-1 in OMV preparations detectable by MS above background levels (Fig. S6, Table 1).In summary, the overproduction in L. pneumophila of WT but not catalytically inactive LqsA produces LAI-1, which partitions to OMVs.

L. pneumophila OMVs containing LAI-1 inhibit the migration of D. discoideum
To address the question, whether L. pneumophila OMVs containing LAI-1 affect amoeba migration and thus LAI-1 secretion by outer membrane vesicles promote interkingdom signaling, we treated D. discoideum with OMVs purified from L. pneumophila JR32 producing WT LqsA or the catalytically inactive mutant LqsA K258A under control of P 6SRNA (Fig. 8).Compared to the MB medium control (Fig. 8A), the migration of the amoeba was inhibited in a dose-dependent manner by OMVs purified from L. pneumophila expressing WT lqsA (Fig. 8B), but not by OMVs purified from L. pneumophila expressing catalytically inactive lqsA (Fig. 8C).At the highest OMV concentration used (1000 μg/ml protein), the velocity of the amoeba was reduced ca.1.7-fold compared to the controls (Fig. 8D).

LqsA overexpression promotes intracellular growth of L. pneumophila
To test the hypothesis that LAI-1 produced by L. pneumophila might affect intracellular pathogen-host interactions, we analyzed the replication in macrophages of L. pneumophila strains constitutively producing GFP and LqsA WT or catalytically inactive mutant protein under the control of P lqsA , P flaA, or P 6SRNA .These strains produced GFP at a comparable level (Fig. S7).Using these plasmids producing GFP and LqsA, LAI-1 was detected by the V. cholerae reporter strain MM920 in OMVs purified from strains expressing lqsA under the control of P flaA or P 6SRNA , but not under the control of P lqsA (Fig. 9A).Intriguingly, upon infection of RAW 264.7 macrophages, the L. pneumophila strains overexpressing lqsA under the control of the P flaA or P 6SRNA promoters replicated more efficiently in the host cells (Fig. 9B).Contrarily, the L. pneumophila strain expressing lqsA under control of its endogenous P lqsA promoter or strains expressing a catalytically inactive lqsA mutant did not show enhanced intracellular growth.Hence, the overexpression of lqsA but not a catalytically inactive mutant promotes intracellular replication of L. pneumophila in macrophages, indicating that intracellularly produced LA1-1 modulates the interaction in favor of the pathogen.

Discussion
In this study, we analyzed the mechanism of LAI-1 release as well as signaling effects of OMVs produced by E. coli and L. pneumophila strains expressing the autoinducer synthase gene lqsA.We found that the expression of lqsA in E. coli (Fig. 3) and L. pneumophila (Figs 5-7) produces LAI-1, which partitions to OMVs.The OMV-localizing LAI-1 is bioactive in interbacterial as well as interkingdom signaling.Accordingly, LAI-1 in OMVs produced by E. coli triggers luminescence of V. cholerae MM920 (Fig. 3, C and D), induces expression of P flaA -gfp in L. pneumophila (Fig. 3G), and inhibits migration of D. discoideum (Fig. 4).Analogously, LAI-1 in OMVs produced by L. pneumophila triggers luminescence of V. cholerae MM920 (Fig. 7) and inhibits migration of D. discoideum (Fig. 8).
Gene expression analysis by quantitative reverse transcription PCR indicated that lqsA is expressed at the onset of the stationary growth phase (25).We observed a similar expression profile for gfp under the control of P lqsA (Fig. S4).However, compared to the strong stationary promoters P flaA and P 6SRNA , the expression of P lqsA was very low (Fig. 6).Upon ectopic expression of lqsA under the control of different promoters, the strength of promoter activity correlated with the amount of LAI-1 detected by LC-MS/MS in OMVs: P 6SRNA > P flaA > P lqsA (Table 1).No LAI-1 was detected in OMVs from L. pneumophila producing catalytically inactive LqsA, or in OMVs from L. pneumophila JR32 (Table 1).The expression of lqsA under control of the P tac promoter yielded ca.30 times less LAI-1 in OMVs from L. pneumophila than E. coli (Table 1).Taken together, these findings correspond to the observation that L. pneumophila strains do not produce LAI-1 detectable by the α-hydroxyketone V. cholerae reporter strain MM920 or by LC-MS/MS, unless lqsA is overexpressed under control of P flaA or P 6SRNA (Figs. 2, 5, 7 and 9).Hence, under the conditions assessed, L. pneumophila tightly regulates the production of only low amounts of LAI-1.
L. pneumophila might produce higher concentrations of LAI-1 under different physiological conditions.In particular, conditions regulated by the pleiotropic transcription factor LvbR, such as biofilm formation or intracellular growth, could determine lqsA expression and LqsA activity.Notably, the production of LAI-1 might be regulated on a transcriptional level (e.g., positive or negative regulation of lqsA expression) and/or on a posttranslational level (e.g., covalent attachment of the cofactor pyridoxal-5 0 -phosphate to LqsA).In any case, the accumulation of LAI-1 in OMVs likely enhances its bioactivity over short distances in highly populated, densely packed biofilms or in the intracellular, organelle-rich milieu of eukaryotic host cells.

LAI-1 secretion by outer membrane vesicles
The results obtained here with OMV-localizing LAI-1 are similar to results obtained with synthetic LA1-1 for interbacterial signaling (33) and for interkingdom signaling (37).However, OMV-associated LAI-1 and synthetic LAI-1 might be solubilized in different microenvironments, that is, in phospholipid bilayer vesicles or in micelles, respectively.Accordingly, the mode of action (membrane fusion versus membrane insertion) and the bioavailability of OMVassociated LAI-1 and synthetic LAI-1 might differ substantially, and consequently, it is challenging to compare the efficacy of the two distinct LAI-1 delivery forms.
While we cannot exclude other transport mechanisms for LAI-1, overall, our findings provide ample evidence that LAI-1 is secreted by OMVs (Figs. 3, 7 and 9).Since LAI-1 is rather hydrophobic, the compound likely partitions into the OMV membrane, but might also be present in the OMV lumen to some extent.As a general concept, packaging in OMVs of hydrophobic, low molecular weight signaling molecules LAI-1 secretion by outer membrane vesicles solubilizes these compounds in aqueous environments and facilitates their distribution over rather large distances at biologically active concentrations (42).Hence, by being distributed through OMVs, LAI-1 and other hydrophobic bacterial signaling molecules experience an increased solubility and enhanced bioactivity.
With regard to interkingdom signaling, LAI-1 is bioactive regardless of whether the compound is delivered extracellularly through purified OMVs from E. coli (Fig. 4) or L. pneumophila (Fig. 8), or whether the compound is produced intracellularly in host cells infected with L. pneumophila overexpressing lqsA (Fig. 9).These results are in agreement with the notion that the (unknown) eukaryotic LAI-1 receptor faces the extracellular milieu and, upon uptake of L. pneumophila, the lumen of the LCV.This topology would allow the recognition of LAI-1 upon extracellular delivery as well as upon intracellular production.Alternatively, the hydrophobicity of LAI-1 might allow for a fairly unrestricted membrane diffusion and receptor targeting.
The overexpression of lqsA, but not a catalytically inactive mutant, promotes intracellular replication of L. pneumophila in macrophages (Fig. 9).This result indicates that intracellular -and thus likely continuous or repeated-production of LA1-1 promotes pathogen-host interactions of L. pneumophila.Similarly, the overexpression of lqsA by intracellular L. pneumophila inhibited cell migration of host cells (37).While the inhibition of eukaryotic cell migration could be faithfully recapitulated by the addition of 10 μM synthetic LAI-1, synthetic LAI-1 added prior to or concomitantly with infected macrophages, did not enhance the uptake (37) or intracellular replication of L. pneumophila.Perhaps, LAI-1 needs to be produced in a continuous or repeated manner to exert intracellular effects.OMVs shed by intracellular L. pneumophila have been reported to inhibit fusion of phagosomes with lysosomes (51) and to promote bacterial replication in macrophages (47).These processes are mechanistically not well understood, but might at least in part rely on LAI-1.Further studies will elucidate the pathways underlying the interkingdom detection of and response to LAI-1 by eukaryotic host cells.

Bacteria, cells, and reagents
The bacterial strains used in this study are listed in Table S1.E. coli TOP10 was cultured overnight at 37 C in LB broth supplemented with 30 μg/ml Cam.L. pneumophila strains were grown for 3 days on charcoal yeast extract agar plates (57), with or without chloramphenicol (Cam; 10 μg/ml) at 37 C. Bacterial colonies were used to inoculate liquid cultures (starting concentration A 600 of 0.1) in AYE medium (58) at 37 C on a wheel (80 rpm) and grown for approximately 18 h, with Cam (5 μg/ml) added to maintain plasmids if required.
V. cholerae MM920 (22) was cultured overnight at 30 C in LB broth supplemented with 5 μg/ml tetracycline (Tet) prior to an experiment.V. cholerae MM920 lacks the autoinducer synthase gene cqsA (ΔcqsA) and the sensor kinase gene luxQ (ΔluxQ), and therefore, does not respond to the quorum sensing signal AI-2, and does not produce but responds to the α-hydroxyketone compounds CAI-1 and LAI-1.Strain MM920 harbors plasmid pBB1, which contains the luxCDABE luciferase operon of V. harveyi and produces light upon detection of CAI-1 and LAI-1.
Murine macrophage-like RAW 264.7 cells (American Type Culture Collection TIB-71, laboratory collection) were cultivated in RPMI 1640 medium (Life Technologies) supplemented with 10% heat-inactivated fetal calf serum (FCS; Life Technologies) and 1% glutamine (Life Technologies) at 37 C with 5% CO 2 in a humidified atmosphere.The macrophages were grown in T75 flasks (Faust Laborbedarf AG) and split every second or third day.
D. discoideum strain Ax3 was cultivated in HL5 medium without glucose (Formedium) supplemented with D-maltose (Roth) at 23 C. The amoebae were grown in T75 flasks and split every third day.

Molecular biology and plasmid construction
The plasmids utilized in this study are listed in Table S1.Cloning procedures followed standard protocols, with plasmid isolation performed using commercially available kits from Qiagen.DNA fragments were amplified using Phusion High Fidelity DNA polymerase, using the primers listed in Table S2.
Gibson assembly was carried out using the NEBuilder HiFi DNA assembly kit (NEB).All constructed plasmids were validated through Sanger sequencing (Microsynth).To construct plasmid pAK14, P lqsA -lqsA was amplified by PCR using the LqsA-nat-fw and LqsA-mod-rev primers and genomic DNA as a template, and the fragment was cloned into plasmid pTS-28 cut with BamHI and MluI.
The plasmids pMF04 and pMF15 were created by replacing the gfp gene in pCM-9 (33) with either the lqsA gene or a pointmutated lqsA K258A gene, resulting in the transcriptional fusions P flaA -lqsA and P flaA -lqsA K258A , respectively.Amplification of the lqsA or lqsA K258A gene regions was performed using the oMF015 and oMF016 primers, with JR32 genomic DNA or pTS-24 as the respective template.The backbones for pMF04 and pMF15 were identical and obtained by amplifying pCM-9 using the oMF013 and oMF014 primers.Similarly, plasmids pMF03 and pMF16 were constructed as transcriptional P 6SRNA -lqsA and P 6SRNA -lqsA K258A fusions by replacing the gfp gene in pRH049 (36).The lqsA or lqsA K258A gene regions were amplified using the oMF011 and oMF012 primers, with JR32 genomic DNA or pTS-24 as the respective template.The PCR products were then cloned into pCM-9, amplified by the oMF009/oMF010 primer set.
To construct plasmid pMF17-pMF21, plasmid pNT31 was used as the backbone by digesting with restriction enzyme BamHI.For pMF17 and pMF18, the insert DNA fragments were amplified with the oMF064 and oMF065 primers, using pMF03 and pMF16 as template.The insert DNA fragments of pMF19 and pMF20 were amplified with the oMF066 and oMF065 primers, using pMF04 and pMF15 as templates.The insert DNA fragment of pMF21 was amplified with the oMF067 and oMF065 primers, using JR32 genomic DNA as template.To construct pMF22, plasmid pMF21 was used as the backbone by amplifying with the primers oMF070 and oMF071 primers.Its insert DNA fragment was amplified by oMF069 and oMF065 primers, using pMF16 as template.The amplified DNA insert fragments were cloned into their corresponding backbones using a NEBuilder HiFi DNA assembly kit.

Isolation of bacterial OMVs
To prepare the bacterial cultures for OMV isolation, L. pneumophila strains were grown to late stationary phase (A 600 5.0) in 420 ml AYE medium, supplemented with 5 μg/ ml Cam if required.The E. coli strains harboring pTS-2 or pTS-10 were cultured in 420 ml of LB medium supplemented with 30 μg/ml Cam.The cultures were maintained at 37 C with shaking at 120 rpm for 16 h.To induce the expression of lqsA in E. coli harboring pTS-2 or pTS-10, 1 mM IPTG was added to the bacterial culture, and the incubation was continued for another 5 h.
OMVs were isolated basically as described (54).A low-speed centrifugation step (1500g, 15 min, 4 C) and subsequent membrane filtration (0.45 μm pore size, Huberlab) were carried out to remove bacterial cells and debris from the culture supernatant.The resulting filtrate was then divided equally into six 70 ml polycarbonate bottles (Beckman Coulter) and subjected to ultracentrifugation (150,000g, 2 h, 4 C) using an Optima L-100XP centrifuge with a 45 Ti fixed angle rotor (Beckman Coulter).The resulting pellets were washed with PBS buffer and subjected to another round of ultracentrifugation under the same conditions.This process yielded a clear yellowish solution containing the secreted OMVs.The pellets from the six polycarbonate bottles were resuspended in a total of 1 ml of PBS buffer, and the resuspended samples were filtered once again using 0.45 μm pore size syringe filters.The OMV yield was determined by quantifying their protein content using the Pierce BCA Protein Assay Kit (Thermo Fisher Scientific) according to the manufacturer's instructions.The OMV extracts were either used immediately for experiments or stored at −80 C for future use, with caution taken to avoid multiple freeze and thaw cycles.

Gene expression measurements with microplate reader
L. pneumophila strains harboring transcriptional gfp fusions were grown to late stationary phase (A 600 5.0) in 3 ml AYE medium, supplemented with 5 μg/ml Cam if needed.Each bacterial culture was then diluted with fresh medium to A 600 0.2, and 200 μl of diluted bacterial suspension was pipetted in triplicates in a Falcon 96-well tissue culture-treated microplate (Corning).The growth curve (A 600 ) and GFP fluorescence intensity (excitation, 488 nm, emission, 528 nm; bottom read) was measured with a BioTek Cytation 5 microplate reader (Agilent Technologies) every 0.5 h for 48 h at 37 C with continuous orbital shaking.

LC-MS/MS analysis
For the preparation of MS samples, OMVs extracted from different L. pneumophila or E. coli strains were subjected to the following procedure.Initially, an OMV sample was combined with dichloromethane in a glass tube at a volume ratio of 2:3.The mixture was then gently agitated on a bench rotator for 1 h at room temperature (RT), with a rotation speed of 4 rpm.Subsequently, the glass tube was allowed to stand undisturbed for 1 h, promoting the separation of the two liquid phases.The organic phase (lower phase) containing the target compounds was carefully collected.To remove the solvent, the collected organic phase was evaporated to dryness using a Concentrator Plus (mode "vacuumhigh vapor", V-HV) for 45 min at RT.The residual solution was reconstituted by adding 150 μl of acetonitrile.The reconstituted sample was then stored at −80 C until used for analysis.
MS-product reaction monitoring analysis was carried out with a Thermo Q Exactive plus instrument (Thermo Fisher Scientific) in the positive Fourier transform mass spectrometry.Both experiments were performed with a mass resolution of 17,000 (m/z = 200).In case of product reaction monitoring, a ramped collision energy (20, 25 and 30 eV) was applied.LA1-1 precursor ion at m/z 438.2 (unit resolution) applying high energy C-trap collision dissociation.A heated electrospray ionization probe was used applying following source parameters: vaporizer temperature, 380 C; sheath gas, 50; aux gas, 20; sheath gas, 50; sweep gas, 0; RF level, 50.0; capillary temperature, 275 C.
For reproducible estimation of absolute concentrations, external standards with concentrations of 10, 100, 1000, and 8000 nM were measured with each batch.Standard curves were determined from summed peak areas of fragment ions at m/z 180.1748, 222.0338, and 420.2323 whereby fragment ion m/z 180.1748 was used as identification fragment.MS-level 1 analysis was used to confirm exact precursor ion mass.

Vibrio reporter assay
The V. cholerae strain MM920 was inoculated in LB liquid medium supplemented with 5 μg/ml Tet and incubated for 18 h at 37 C.The overnight culture was diluted with fresh medium to an A 600 of 0.25.The medium was supplemented with either synthetic LAI-1 (1-50 μM) or OMV samples (50-250 μg/ml protein concentration).

LAI-1 secretion by outer membrane vesicles
The mixtures were then transferred to a 96-well plate (Chemie Brunschwig AG) and bioluminescence (luminescence; bottom read) intensity was measured using a Biotek Cytation 5 microplate reader every 0.5 h for 8 to 10 h at 30 C with continuous orbital shaking.Images were captured after 4 to 5 h incubation (when bioluminescence intensity usually reached maximum levels) using the FluorChem SP imaging system (Alpha InnoTec) with an exposure time of 15 min.All experiments were performed in biological triplicates.
Single amoeba tracking D. discoideum Ax3 amoeba were seeded at a density of 2 × 10 4 cells per well into an 8-well Ibidi chamber (Ibidi) and incubated for 3 to 4 h in HL5 medium to allow attachment to the bottom.The medium in each well was then replaced by a diluted OMV sample and incubated at 23 C for 1 h before microscopy imaging.E. coli OMVs were diluted using MB medium (20 mM (2-(N-morpholino)ethanesulfonic acid), pH 6.9, 0.7% yeast extract, 1.4% BBL thiotone E peptone) to a final protein concentration of 100 μg/ml, 250 μg/ml, or 500 μg/ml.L. pneumophila OMVs were diluted using MB medium to a final protein concentration of 250 μg/ml, 500 μg/ml, or 1000 μg/ml.

Intracellular replication assay
For infection assays, RAW 264.7 macrophages were seeded at a density of 1 × 10 5 cells per well onto 24-well tissue culture plates and incubated overnight in RPMI 1640/10% FCS medium.L. pneumophila was inoculated at an A 600 0.2 in AYE medium and grown on a wheel at 37 C to stationary phase (A 600 5.0, 2 × 10 9 bacteria/ml).The bacterial cultures were then diluted to the desired density in pre-warmed RPMI 1640/ 10% FCS, and the infection was synchronized by centrifugation (1050g, 10 min, RT).After 1 h, the infected cells were washed 4 times with prewarmed RPMI 1640/10% FCS, and further incubated for the time indicated.Depending on the experimental set-up, the infected cells were imaged live, lysed with 0.1% Triton X-100 and/or fixed with 4% paraformaldehyde.

Transmission electron microscopy
Transmission electron microscopy (TEM) was used for the morphological characterization and size distribution of OMVs extracted from different L. pneumophila or E. coli strains.TEM was carried out by using a FEI Tecnai G2 Spirit (FEI Company), operating at 120 kV and equipped with a Gatan Orius 1000 (4k × 2.6k pixels) camera (Gatan Inc).FEI was used as the control software, and Gatan DigitalMicrograph was applied for image acquisition.
Prior to TEM, negative staining was performed on OMV samples as follows: A droplet of 5 μl OMV sample (1000 μg/ml diluted with PBS buffer) was added onto a copper a grid (300 square mesh, Formvar Carbon support film; micro to nano) and incubated for 60 s to allow attachment.The liquid was gently blotted away using a Whatman filter paper.A drop of ddH 2 O was then added to wash the grid once.Then a drop of 1% uranyl acetate solution was added onto the grid and incubated for 45 s.After blotting off the liquid, the grid was placed on a filter paper to let dry completely and transferred for TEM.
The size of the OMVs was manually measured from the images acquired by TEM using ImageJ.In each sample, a minimum of 1000 OMVs were individually measured for diameter, and the data were then used to plot the diameter distribution.The average number or concentration of OMVs could not be determined due to technical reasons, since (i) OMVs were nonhomogeneously distributed on the electron microscopy grids, and (ii) the size detection limit of OMVs was <50 nm (smaller OMVs could not be quantified).However, the size distribution and the protein concentration ( ±5%) of OMVs purified from the same bacterial species was very similar, indicating that the overall number of OMVs was also similar.

Confocal laser scanning microscopy
Fluorescence-based imaging was conducted with a Leica SP8 confocal laser scanning microscope.The following imaging acquisition was used for all imaging experiments: white light laser at 488 nm (2% intensity), fluorescent signal was detected by a power HyD detector (emission range 500-520 nm), transmissive light was detected by a photomultiplier tube detector with a gain of 380.For single amoeba tracking, three fields of interest were randomly selected for each sample and recorded continuously for 2 h with 2 min time interval.Image analyses were performed using ImageJ (https://imagej.net/ij/)and Chemotaxis and Migration Tool version 2.0 (Ibidi).

Quantification and statistical analysis
Each experiment was independently replicated at least three times, and representative images are shown.All statistical analyses were performed using GraphPad Prism version 7.01 for Windows, GraphPad Software (www.graphpad.com).Twotailed student's t tests were used to compare the means of three technical replicates between experimental samples and controls.Significance levels were represented in the figures as follows: *, **, ***, or **** to indicate probability values of less than 0.05, 0.01, 0.001, and 0.0001, respectively.

Figure 2 .
Figure 2. LAI-1 detection by Vibrio cholerae MM920 and LC-MS/MS.A-C, luminescence of the V. cholerae α-hydroxyketone reporter strain MM920.A, Legionella pneumophila strains (JR32, Corby, clinical isolate 883), Legionella micdadei, or Escherichia coli harboring an empty vector (pTS-10) or plasmids expressing lqsA (pTS-2)or cqsA (pTS-6) under control of the P tac promoter were grown on CYE agar in 24 well plates for 2 days, before the V. cholerae reporter strain MM920 was streaked out in parallel.After another day, growth was assessed (upper panels), and bioluminescence was determined with a luminometer (lower panels).V. cholerae MM920 was treated with the concentrations of synthetic LAI-1 indicated, DMSO was included as negative control, and luminescence intensity was (B) visualized by a gel documentation system (15 min exposure time), or (C) measured by a plate reader (30 C, 8 h).D-F, MS fragment spectra of LAI-1.D, the fragment ions at m/z 180.174 and 222.221 cover the complete structure of LAI-1 and exhibit high specificity.For detection of LAI-1, the focus was placed on fragment ions at m/z 180.174 (light blue arrowhead), 222.221 (orange arrowhead), and 420.232 (green arrowhead).Positive detection and identification results were obtained from the extracted ion chromatogram (EIC) of (E) the parent ion at m/z 438.24, along with (F) compound-specific fragment ions at m/z 180.174 (light blue line), 222.221 (orange line), and 420.232 (green line).LC-MS/MS analysis was performed with 1 μM synthetic LAI-1 as a standard.The data shown are means and standard deviations of technical triplicates (C) and representative of three independent experiments.CYE, charcoal yeast extract; DMSO, dimethyl sulfoxide; LAI-1, Legionella autoinducer-1; LC-MS/ MS, liquid chromatography-tandem mass spectrometry; Lqs, Legionella quorum sensing; RLU, relative light units.
Figure 2. LAI-1 detection by Vibrio cholerae MM920 and LC-MS/MS.A-C, luminescence of the V. cholerae α-hydroxyketone reporter strain MM920.A, Legionella pneumophila strains (JR32, Corby, clinical isolate 883), Legionella micdadei, or Escherichia coli harboring an empty vector (pTS-10) or plasmids expressing lqsA (pTS-2)or cqsA (pTS-6) under control of the P tac promoter were grown on CYE agar in 24 well plates for 2 days, before the V. cholerae reporter strain MM920 was streaked out in parallel.After another day, growth was assessed (upper panels), and bioluminescence was determined with a luminometer (lower panels).V. cholerae MM920 was treated with the concentrations of synthetic LAI-1 indicated, DMSO was included as negative control, and luminescence intensity was (B) visualized by a gel documentation system (15 min exposure time), or (C) measured by a plate reader (30 C, 8 h).D-F, MS fragment spectra of LAI-1.D, the fragment ions at m/z 180.174 and 222.221 cover the complete structure of LAI-1 and exhibit high specificity.For detection of LAI-1, the focus was placed on fragment ions at m/z 180.174 (light blue arrowhead), 222.221 (orange arrowhead), and 420.232 (green arrowhead).Positive detection and identification results were obtained from the extracted ion chromatogram (EIC) of (E) the parent ion at m/z 438.24, along with (F) compound-specific fragment ions at m/z 180.174 (light blue line), 222.221 (orange line), and 420.232 (green line).LC-MS/MS analysis was performed with 1 μM synthetic LAI-1 as a standard.The data shown are means and standard deviations of technical triplicates (C) and representative of three independent experiments.CYE, charcoal yeast extract; DMSO, dimethyl sulfoxide; LAI-1, Legionella autoinducer-1; LC-MS/ MS, liquid chromatography-tandem mass spectrometry; Lqs, Legionella quorum sensing; RLU, relative light units.

Figure 5 .
Figure 5. Legionella pneumophila strain JR32 and ΔlqsA produce similar OMV populations.A, negative staining TEM images of OMVs derived from L. pneumophila JR32 (left) or ΔlqsA mutant strain (right).B, size distribution analysis of OMVs from L. pneumophila JR32 or ΔlqsA.C and D, luminescence of Vibrio cholerae MM920 mixed with OMVs from L. pneumophila JR32 or ΔlqsA of the protein concentrations indicated (control: DMSO).Intensity was (C) visualized by a gel documentation system (15 min exposure time), or (D) measured by a plate reader (30 C, 7 h).E and F, LC-MS/MS analysis of OMVs of L. pneumophila JR32 or ΔlqsA (1 mg protein).Extracted ion chromatograms (EICs) of fragment ions at m/z 180.174 (light blue), 222.221 (orange), and 420.232 (green) did not indicate the presence of LAI-1 in these OMVs.The data shown are means and standard deviations of technical triplicates (D) and representative of three independent experiments.DMSO, dimethyl sulfoxide; LAI-1, Legionella autoinducer-1; LC-MS/MS, liquid chromatography-tandem mass spectrometry; Lqs, Legionella quorum sensing; OMVs, outer membrane vesicles; RLU, relative light units; TEM, transmission electron microscopy.

Table 1
Detection of LAI-1 in OMV samples by LC-MS/MS For each sample, 1 mg of OMVs (protein concentration) was processed and analyzed by LC-MS/MS as outlined in the Experimental procedures section.The values are representative for >2 measurements. a