Extracellular and intracellular destruction of Pseudomonas aeruginosa by Dictyostelium discoideum phagocytes mobilize different antibacterial mechanisms

Ingestion and killing of bacteria by phagocytic cells are critical processes to protect the human body from bacterial infections. In addition, some immune cells (neutrophils, NK cells) can release microbicidal molecules in the extracellular medium to eliminate non‐ingested microorganism. Molecular mechanisms involved in the resulting intracellular and extracellular killing are still poorly understood. In this study, we used the amoeba Dictyostelium discoideum as a model phagocyte to investigate the mechanisms allowing intracellular and extracellular killing of Pseudomonas aeruginosa. When a D. discoideum cell establishes a close contact with a P. aeruginosa bacterium, it can either ingest it and kill it in phagosomes, or kill it extracellularly, allowing a direct side‐by‐side comparison of these two killing modalities. Efficient intracellular destruction of P. aeruginosa requires the presence of the Kil2 pump in the phagosomal membrane. On the contrary, extracellular lysis is independent on Kil2 but requires the expression of the superoxide‐producing protein NoxA, and the extracellular release of the AplA bacteriolytic protein. These results shed new light on the molecular mechanisms allowing elimination of P. aeruginosa bacteria by phagocytic cells.

In addition to being delivered to maturing phagosomes, microbicidal molecules can also be released in the extracellular medium, a process referred to as degranulation (Faurschou & Borregaard, 2003;Mok et al., 2021;Othman et al., 2021).This allows phagocytic cells to kill uningested microorganisms in their close vicinity.Highlighting the importance of this killing mechanism, some bacterial pathogens have developed strategies to interfere with this process in order to increase their ability to mount harmful infections (Eichelberger & Goldman, 2020).For example, during pulmonary infections, Yersinia pestis, etiologic agent of plague, injects YopE and YopH proteins in neutrophils, inhibiting granule fusion with the plasma membrane and promoting bacterial survival (Eichelberger et al., 2019).The ionic and biochemical composition of the extracellular medium clearly differs from that of the acidic phagosomal lumen, and it seems likely that in these two very different environments, different cellular effectors are used to kill bacteria, but this has not been studied extensively.
In this study, we used D. discoideum, a well-characterized phagocytic cell, to study the cellular mechanisms involved in intracellular and extracellular killing of P. aeruginosa bacteria.D. discoideum is a soil amoeba feeding on microorganisms.Its small haploid genome allows relatively easy genetic manipulation and has made it a convenient model system to study many biological processes, in particular the complex interactions between phagocytic cells and bacteria (Cosson & Soldati, 2008), and the largely conserved mechanisms ensuring killing of bacteria (Dunn et al., 2017).P. aeruginosa is an opportunistic pathogen, ubiquitous in the soil and water (Hardalo & Edberg, 1997).Pathogenic strains of P. aeruginosa can produce and secrete virulence factors that allow them to kill amoebae (Cosson et al., 2002;Pukatzki et al., 2002).When bacterial virulence is reduced (e.g., by specific mutations), amoebae can ingest bacteria, kill them (Jauslin et al., 2021), and feed upon them (Cosson et al., 2002).
The balance between bacterial virulence factors and D. discoideum killing mechanisms determines the outcome of the encounter between these two organisms.
In this study, we show that D. discoideum kills ingested P. aeruginosa bacteria in phagosomes, as well as uningested bacteria with which it establishes extracellular contacts.Extracellular and intracellular killing make use of distinct sets of cellular bactericidal mechanisms.

| D. discoideum frequently establishes transient contacts with P. aeruginosa without ingesting them
This study was initiated by the serendipitous observation that when a D. discoideum cell establishes a contact with a P. aeruginosa bacterium, this contact results in ingestion of the bacterium only in a minority of cases.To visualize the interactions between amoebae and bacteria, non-pathogenic GFP-expressing P. aeruginosa bacteria were deposited on a glass slide together with phagocytic amoebae and imaged every 30 s for two hours.D. discoideum occasionally established a contact with a P. aeruginosa bacterium and ingested it (Figure 1a).In this situation, as previously described, the ingested bacterium remained fluorescent for a few minutes, until its destruction in phagosomes (Jauslin et al., 2021) (Figure 1a).In other cases, D. discoideum established a contact with P. aeruginosa, but did not ingest it (Figure 1b).After a brief contact (mean duration, 2.2 min; n = 150 contacts), the D. discoideum cell detached and moved away from the bacterium (Figure 1b).We assessed this trait quantitatively by counting how often direct contact between a bacterium and an amoeba resulted in engulfment of the bacterium, as has been previously described (Delince et al., 2016).Quantification revealed that D. discoideum ingested P. aeruginosa in 12% of the cases after it established a contact with them (Figure 1c).The behavior of D. discoideum was different when it encountered other gram-negative (Klebsiella pneumoniae or E. coli) or gram-positive (S. aureus) bacteria: contact with these three bacteria resulted in a very efficient phagocytosis, with at least 80% of contacts resulting in ingestion of bacteria (Figure 1c).When a D. discoideum cell establishes a transient contact with a bacterium, the bacterium may in principle remain extracellular, or it may be transiently ingested and excreted a few minutes later.Light microscopy does not allow to distinguish unambiguously between these two possibilities.In order to determine if bacteria remained extracellular, we labeled the surface of P. aeruginosa with antibodies coupled with fluorescein (FITC) and Alexa 546.When bacteria were ingested, the FITC fluorescence was rapidly quenched in acidifying phagosomes while the fluorescence of Alexa 546 did not diminish (Figure 2a,c).On the contrary, when bacteria were only transiently in contact with D. discoideum cells, no quenching of the FITC was observed (Figure 2b,d), indicating that either these bacteria were not ingested, or they were regurgitated without having reached acidifying phagosomes.
Together these observations indicate that two scenarios can be distinguished when D. discoideum cells establish contacts with P. aeruginosa: either the bacterium is ingested, or it remains extracellular.This situation allowed us to characterize in parallel the destruction of ingested and non-ingested bacteria by D. discoideum.

| Intracellular destruction of P. aeruginosa requires Kil2 activity
As previously described, when bacteria were ingested by amoebae, their GFP fluorescence disappeared a few minutes later, when the bacteria were destroyed in maturing phagosomes (Jauslin et al., 2021) (Figure 1a).Recording the time between phagocytosis and extinction of the GFP fluorescence reveals the kinetics of intracellular destruction of ingested bacteria (Figure 3a).As previously reported (Jauslin et al., 2021), ingested P. aeruginosa were rapidly destroyed in phagosomes, and destruction was slower in kil2 KO cells.
Kil2 is a phagosomal P-type ATPase that has been proposed to transport magnesium ions from the cytosol to the phagosomal lumen (Lelong et al., 2011).Measuring the level of fluorescence of bacteria revealed that full extinction of GFP fluorescence was preceded by a gradual decline during approximately 2-3 min (Figure 3b).The average time before GFP fluorescence declined was longer in kil2 KO cells than in WT cells (Figure 3a), but once initiated, the kinetics of GFP extinction were similar in WT and mutant cells (Figure 3b).We tested systematically a collection of mutants to determine the importance of various gene products in intracellular destruction of P. aeruginosa (Figure 3c).Our results are in good agreement with results previously reported (Jauslin et al., 2021).Initial observations suggested that intracellular destruction of P. aeruginosa may be slower in noxA KO cells and bpiC KO cells than in wild-type (WT) cells (Jauslin et al., 2021).However, the differences observed were relatively small and their statistical significance was not firmly established.NoxA is the main superoxide-producing NADPH oxidase in D. discoideum phagosomes (Lardy et al., 2005) and BpiC is a bactericidal permeability-increasing protein which binds lipopolysaccharides (LPS) in the cell wall of gram-negative bacteria (Jauslin et al., 2021).
We reanalyzed our original experiments and performed new experiments, and the resulting set of results indicates that noxA KO cells and bpiC KO cells destroy ingested P. aeruginosa as fast as WT cells (Figure 3c).We also established that aplA KO, aplB KO, aplH KO, aplN KO, and WT D. discoideum cells destroyed ingested bacteria with indistinguishable kinetics (Figure 3c).

| P. aeruginosa is lysed following transient contact with D. discoideum
As described in Figure 1, many P. aeruginosa bacteria are not ingested by D. discoideum.However, after 2 h of continuous observation the number of live bacteria drastically decreased in the presence of D. discoideum.To quantify this observation, we counted in five independent experiments the number of live bacteria present in the field of observation at time 0 and after 2 h, as well as the number of ingested bacteria.Overall, 1094 bacteria were present at time 0, 371 remained after 2 h, but only 12 were ingested over this time period.This observation strongly suggests that in the presence of D. discoideum, a large fraction (≈65%) of P. aeruginosa bacteria were destroyed extracellularly (Figure S1a, WT).We then assessed more precisely the fate of individual bacteria following contact with D. No contact red line; ≈ 20% lysis).In the absence of D. discoideum, only a very small percentage of P. aeruginosa lysed spontaneously (Figure 4e; Buffer green line; <10% lysis).We ascertained that these differences were significant by quantifying six independent experiments (Figure 4f).
In order to determine the sequence of events leading to bacterial lysis, we recorded movies with shorter time intervals (1 image every 6 s).Among 80 bacteria touched by a D. discoideum cell, 3 presented an intermediate state where loss of GFP fluorescence preceded bacterial lysis by one frame, that is, by 6 s (Figure S2).These observations indicate that permeabilization of the bacterial membrane and loss of intracellular GFP preceded by few seconds extracellular lysis of bacteria.
While some bacteria were not lysed even several hours after being touched by D. discoideum cells, their viability (i.e., their ability to grow and divide) may be affected by the contact.In order to evaluate this possibility, we mixed bacteria and amoeba cells and observed them for 25 min.We then replaced the buffer with LB medium containing a small amount of paraformaldehyde and increased the temperature to 37°C (Figure 5a).These conditions kill D. discoideum cells, but allow bacterial growth (Crespo-Yanez et al., 2022).
As a control, we also measured growth of bacteria incubated in the ) were not ingested when touched by D. discoideum, and lysis of these bacteria was never observed (Figure S3a,B).

| Extracellular lysis of P. aeruginosa requires D. discoideum NoxA and AplA
In order to identify cellular mechanisms allowing D. discoideum to lyse bacteria extracellularly, we followed the fate of individual extracellular bacteria exposed to a panel of D. discoideum mutants.
kil2 KO D. discoideum cells lysed extracellular bacteria as efficiently as WT cells (Figure 6a), indicating that while Kil2 was essential for efficient intracellular destruction (Figure 3), it was not required for extracellular destruction.Two mutant cells exhibited strongly defective extracellular killing of bacteria: noxA KO and aplA KO cells (Figure 6a).Global analysis also revealed that the number of extracellular bacteria decreased less in the presence of noxA KO and aplA KO cells than in the presence of WT cells (Figure S1a,b).Smaller defects in the extracellular lysis of bacteria were observed in kil1 KO and alyL KO cells (Figure 6b).No significant defects were observed when other genes were mutated, notably in members of the apl family of genes, aplB, aplH, or aplN (Figure 6b).While a role for NoxA and AlyL are coherent with the literature (see Discussion), the putative role of AplA in bacterial killing was more unexpected, and we present below a detailed analysis of the putative role of AplA in extracellular killing.
The AplA protein is composed of 5 saposin B-type domains termed SAPLIP domains (Figure 7a).Saposin-like proteins typically interact with lipids and this property conveys an antibacterial activity to some of them (e.g., granulysin (Pena & Krensky, 1997)).
We overexpressed in aplA KO cells an AplA protein tagged at its C-terminus with an ALFA epitope (Gotzke et al., 2019).The AplA-ALFA protein was detected by western blot as a doublet with an approximate molecular weight of 70 kDa, slightly higher than expected for the full-length AplA protein (59 kDa), presumably due to

| AplA is secreted in the extracellular medium
The role of AplA in extracellular killing suggests that it is secreted extracellularly.To test this hypothesis, we incubated cells expressing AplA-ALFA in HL5 for 2.5 or 5 h and measured by western blot the presence of AplA-AFLA in cells and in the extracellular medium (SN: supernatant).Cells contained a mixture of high-(≈72 kDa) and low (≈68 kDa) molecular weight AplA (Figure 8, Cell).In HL5 conditions, only a small portion of the 72 kDa AplA was found in the medium after 2.5 and 5 h (Figure 8, SN) showing that AplA was mostly retained in the cells.When cells were incubated in nutrient-depleted phosphate buffer (PB*), AplA was efficiently secreted in the medium at 2.5 and 5 h (Figure 8, SN).As observed in HL5, only the 72 kDa AplA was secreted, which presumably represents the mature form of AplA (Figure 8).
We next assessed whether the secretion of AplA is stimulated in the presence of P. aeruginosa or K. pneumoniae bacteria.The amount of AplA detected in the SN of cells exposed to P. aeruginosa or K.
pneumoniae was the same as in the absence of bacteria (Figure S4a).
In summary, the secretion of AplA resembles the secretion of lysosomal enzymes which were previously shown mostly retained in unstarved cells, and secreted in the extracellular medium by starved cells (Dimond et al., 1981).To verify that the secretion of AplA mirrors the secretion of lysosomal enzymes in our system, we measured the activity of two well-characterized lysosomal enzymes in the cell pellet and in the extracellular medium.As expected, N-acetylglucosaminidase and α-mannosidase are secreted by starved cells (PB* buffer) but not by unstarved cells (HL5) (Figure S4b,c).The secretion of these enzymes was not increased in the presence of P. aeruginosa or K. pneumoniae (Figure S4b,c).Together these results indicate that, like lysosomal enzymes, AplA is secreted most efficiently in the extracellular medium by starved cells, and that its secretion is not increased in the presence of bacteria.
We then used immunofluorescence to detect in which compartment the AplA-ALFA was present within cells.For this we used a panel of recombinant antibodies labeling specific D. discoideum subcellular compartments (Figure 9).AplA-ALFA was mostly detected in a compartment at the center of the cell, which was distinct from p25-positive recycling endosomes (Charette et al., 2006) (Figure 9a), from p80-positive endosomes (Ravanel et al., 2001) (Figure 9b), from VatA-positive endosomes (Figure 9c), from sctA-positive pycnosomes (Sabra et al., 2016) (Figure 9d), from lysosomes enriched in sulfated oligosaccharides (Knecht et al., 1984) (Figure 9e), from the Rhesus-positive contractile vacuole (Benghezal et al., 2001) (Figure 9f), and from the PDI-positive endoplasmic reticulum (Marchetti, 2021) (Figure 9g).AplA-positive compartments are in close proximity with the Golgi apparatus, but the two compartments are clearly distinct (Figure 9h).Overall, these results indicate that AplA is present in a cellular compartment distinct from all cellular compartments tested here, presumably of lysosomal nature, the content of which is secreted upon starvation.

| Efficient intracellular and extracellular bacterial destruction require different sets of effectors
This study is based on the observation that when D. discoideum cells encounter P. aeruginosa bacteria, this can generate two different out-

| Role of NoxA, AlyL, and AplA in extracellular lysis of P. aeruginosa
NoxA is the only ortholog of human Nox proteins expressed in vegetative D. discoideum cells (Lardy et al., 2005).Like its Nox2 human counterpart (Vermot et al., 2021), it produces superoxide-free radicals both in the extracellular medium and within phagosomes.In neutrophils and in D. discoideum, the relative importance of superoxide production at the cell surface and in phagosomes has not been directly assessed so far.Our results indicate that in D. discoideum, superoxide production is necessary for efficient lysis of extracellular P. aeruginosa, but dispensable for P. aeruginosa destruction within phagosomes.
AlyL is one of the very diverse lysozymes encoded by the D. discoideum genome (Lamrabet et al., 2020), and our previous results showed that it plays an important role in the killing of K. pneumoniae bacteria in D. discoideum phagosomes (Jauslin et al., 2021).In humans, bactericidal lysozymes similar to AlyL are released in the extracellular environment by a variety of cells (Ragland & Criss, 2017).
While secreted lysozyme is thought to participate in human innate immunity at least in part by destroying extracellular bacteria, its relative role in intracellular versus extracellular destruction of bacteria is not known, and its exact mode of action is still under investigation.
Our observations indicate that AlyL plays a role in the extracellular destruction of P. aeruginosa, while its activity is dispensable for destruction of P. aeruginosa in phagosomes.
AplA is a saposin-like protein.The antibacterial role of proteins containing a saposin-like lipid-binding (SAPLIP) domain was first characterized in Entamoeba histolytica.E. histolytica amoebapore A is composed of one saposin-like domain.Granules containing amoebapore A can fuse with the cell surface as well as with phagosomal membranes.Secreted amoebapore A can lyse extracellular bacteria by interacting with their membranous lipids but its role in phagosomes remains to be established (Leippe & Herbst, 2004).Similar saposin-like antibacterial proteins were membranes (Dhakshinamoorthy et al., 2018).Like amoebapore A and AplD, human granulysin is composed of a single saposin-like domain.It is produced in non-phagocytic cytotoxic T cells and natural killer cells (Krensky & Clayberger, 2009).These cells can release granulysin in the extracellular medium at sites where they engage into close contact with infected cells or extracellular pathogens, and this release allows the extracellular killing of pathogenic bacteria such as S. typhimurium, L. monocytogenes, E. coli, and S. aureus (Lu et al., 2014;Stenger et al., 1998;Walch et al., 2014).Surfactant protein B contains 3 SAPLIP domains.It is secreted by type II epithelial cells into the pulmonary alveolar space (Weaver & Whitsett, 1989), binds to K. pneumoniae and S.
aureus and enhances their phagocytosis and killing in macrophages (Yang et al., 2010).

| Predator-prey relationship between P. aeruginosa and D. discoideum
In its natural environment, D. discoideum frequently encounters pathogenic bacteria which have evolved virulence mechanisms to survive this encounter.Especially problematic are P. aeruginosa bacteria since they secrete factors that can kill D. discoideum cells before D. discoideum can ingest these bacteria (Cosson et al., 2002;Pukatzki et al., 2002).In particular, secreted bacterial rhamnolipids can lyse D. discoideum or mammalian cells (Cosson et al., 2002).The type III secretion system also allows P. aeruginosa to kill amoebae or mammalian cells with which they establish a physical contact (Alibaud et al., 2008;Pukatzki et al., 2002).In addition, several P. aeruginosa factors were shown to prevent its ingestion by mammalian phagocytes, such as LPS (Engels, 1985 #66), alginate (Oliver, 1985 #67;Leid, 2005 #68), and the ExoS and ExoT toxins delivered into phagocytic cells by the type III secretion system (Rangel, 2014 #69;Garrity-Ryan, 2000 #70).Our results show that P. aeruginosa also escapes phagocytosis by D.  -10 (Caterina et al., 1994;Cornillon et al., 2000) cells, referred to as WT were grown at 21°C in HL5 medium (Froquet et al., 2009) containing 15.6 μg/mL of tetracycline in 10 mL Petri dishes.All mutants used in this study are derived from WT DH1-10 and were used previously (Jauslin et al., 2021), with the exception of aplB and aplH mutants.A detailed description of the strategy used to create all mutants is provided (Figure S5a), as well as, the sequence of the primers used (Figure S5b) and a picture of the relevant agarose gels (Figure S5c).

| Intracellular destruction and extracellular lysis of bacteria
The intracellular or extracellular destruction of fluorescent bacteria in the presence of D. discoideum cells was visualized and measured as described previously (Bodinier et al., 2020).Briefly, fluorescent bacteria and 7 × 10 5 D. discoideum cells were washed in phosphate buffer (PB: 2 mM Na 2 HPO 4 + 14.7 mM KH 2 PO 4 pH 6.3) supplemented with 100 mM sorbitol (PB-Sorbitol).Fluorescent bacteria were deposited in a glass-bottom well (μ-slide 8-well; IBIDI) and slides were centrifuged after addition of the D. discoideum cells to ensure efficient sedimentation of bacteria and cells (183 × g, 10 min).When K. pneumoniae and S. aureus were used, no centrifugation was needed to sediment bacteria before the addition of D. discoideum cells.Movies were recorded for 2 h at 21°C by taking a picture every 30 s.
Intracellular bacteria were identified based on the fact that at the time of ingestion, they detached from the glass substrate and moved inside the cell in the central area.Extracellular bacteria remained immobile and attached to the glass substrate.
Alternatively extracellular bacteria detached from the substrate, remained attached to the amoeba surface, but visibly on the outside of the cell.
Detaching and floating bacteria were easily identified (Figure S6) and when analyzing the fate of individual bacteria, they were excluded from analysis.

| Measuring bacterial viability
Viable bacteria were defined here as bacteria which can grow in LB and this was measured as described previously (Crespo-Yanez et al., 2022).Briefly, cells were mixed with bacteria as described above, and observed for 25 min instead of 2 h (movie 1).Then the medium was aspirated gently and replaced with LB medium containing 0.0016% of paraformaldehyde and 100 μg/mL of Carbenicillin and cells were incubated at 37°C for 10 h.These conditions kill D. discoideum but allow growth of viable P. aeruginosa bacteria.Growth of viable bacteria was visualized by taking a picture every 5 min during this second incubation (movie 2).

| Expression of AplA recombinant proteins
The sequences coding for various Apl proteins (AplA, AplB, AplH, AplA domain 1, or 5) fused at the C-terminal end to an SG spacer and the ALFA tag sequence (SGSRLEEELRRRLTE) (Gotzke et al., 2019) were synthesized by Thermo Fisher Scientific and cloned into the prepSC3 vector (G418-resistant) as described previously (Froquet et al., 2012).The expression vectors were transfected into aplA KO cells, and clones were selected using G418 at a concentration of 15 mg/L.The domain 1 of AplA, extended from Met 1 to Glu130.To express the domain 5, the signal sequence of AplA (Met 1 to Ala 22) was followed by the sequence of domain 5 (Glu 400 to Phe 522).

| Secretion of AplA-ALFA
aplA KO cells overexpressing AplA-ALFA (4 × 10 6 ) were pelleted and resuspended in 200 μL of HL5 or PB-Sorbitol supplemented with 0.5% HL5 (PB*) to increase cell viability and incubated at 21°C at 100 rpm for 5 h.An aliquot of 100 μL was taken after 2 h 30 min of secretion in each condition.The supernatant and cells were separated by centrifugation (2 min, 1503 × g).For each condition, 2 × 10 5 cells were loaded onto SDS-PAGE, separated by electrophoresis, and analyzed by western blot as described above.

| Immunofluorescence
D. discoideum cells overexpressing AplA-ALFA were deposited on a sterile glass coverslip overnight in HL5 at 21°C to reach 1 × 10 6 cell/mL the next day.Then, HL5 was replaced with PB* for 10 min prior to fixation and permeabilization, to favor efficient attachment of cells to the coverslip.Cells were fixed in PB* + 4% paraformaldehyde for 30 min at room temperature.Then cells were washed 5 min in PBS + 40 mM NH 4 Cl and permeabilized with methanol at −20°C for 2 min.After 1 wash with PBS for 5 min, coverslips were blocked with PBS-BSA (PBS + 0.2% BSA) for 15 min prior to incubation with primary antibodies (AL626-R or AL626-M, AJ513-M, AJ154-M, AJ520-M, AK422-M, AK426-M, AJ514-M, and Rhesus-R) at 1 μg/mL in PBS-BSA for 30 min.
After 3 washes in PBS-BSA, coverslips were incubated with secondary antibodies (goat anti-mouse coupled to Alexa 488 (Invitrogen, A11029) and goat anti-rabbit coupled to Alexa 647 (Invitrogen, A21245)) diluted 1:400 in PBS-BSA for 30 min.Then, coverslips were washed 3 times in PBS-BSA and once in PBS and mounted on microscope slides with 10 μL of Moewiol Dabco.Picture were taken with Confocal Laser Scanning Microscopy (Zeiss LSM700).

F
Dictyostelium discoideum cells often release Pseudomonas aeruginosa bacteria after establishing a transient contact with them.(a) The top row of consecutive images shows superimposed phase contrast and fluorescence images, allowing to visualize the ingestion and the intracellular destruction of P. aeruginosa by a WT D. discoideum cell.The bottom row (GFP) shows the GFP channel with the outline of the D. discoideum cell drawn in white.White arrowheads indicate an ingested bacterium and empty arrowheads indicate the time where the bacterium is destroyed.Ingested bacterium detached from the glass substrate and moved inside the amoeba cell.Scale bar 10 μm.(b) One D. discoideum cell establishes a close contact with a P. aeruginosa bacterium, without ingesting it.Non-ingested bacterium did not move relative to the glass substrate.Scale bar 10 μm.(c) For different bacteria, the percentage of contacts with D. discoideum resulting in ingestion was determined.The indicated values are means ± SEM of 5 independent experiments (50 events per experiment).Pa: Pseudomonas aeruginosa, Kp: Klebsiella pneumoniae, Ec: E. coli, Sa: S. aureus.and observed that many P. aeruginosa bacteria lost their fluorescence while still extracellular.Extracellular lysis was observed while D. discoideum was still attached to the bacterium (Figure 4a) or after the D. discoideum cell had detached from the bacterium (Figure 4b).Some bacteria were not lysed following contact with D. discoideum (Figure 4c) even after prolonged observation for up to 120 min.Loss of fluorescence of extracellular bacteria following contact with D. discoideum occurred abruptly within less than 30 s (the time separating two successive frames) (Figure 4d).Extracellular bacteria were clearly visible by phase contrast microscopy when they were not in contact with D. discoideum, and in this situation, loss of bacterial fluorescence was concomitant with bacterial lysis visualized by phase contrast microscopy (Figure 4b, insets).To assess more quantitatively the extracellular bacterial lysis, lysis of bacteria was recorded as a function of time following the initiation of a contact with D. discoideum.Approximately 60% of bacteria were F I G U R E 2 Pseudomonas aeruginosa bacteria are not exposed to an acidic pH during a transient contact with Dictyostelium discoideum.P. aeruginosa bacteria were labeled with a primary antibody recognizing surface O5-antigens and secondary antibodies coupled with Alexa 546 and FITC.(a) An antibody-coated bacterium is ingested by a D. discoideum cell.FITC is rapidly quenched in the acidic phagosomal compartment.(b) An antibody-coated bacterium is touched and then released by a D. discoideum cell.The FITC is not quenched, indicating that the bacterium is not exposed to an acidic pH.(c, d) The fluorescence intensities of FITC and Alexa 546 signals was quantified for bacteria ingested by D. discoideum (mean ± SEM, n = 19 events) (c), and for bacteria establishing a transient contact with D. discoideum (mean ± SEM of 2 independent experiments; n = 19 events) (d).Time 0 indicates the time when each bacterium was initially ingested (c) or touched (d) by D. discoideum.120 min following a contact with a D. discoideum cell (Figure 4e; Contact blue line).Bacteria that were not touched by amoebae also lysed occasionally, but less frequently (Figure 4e; absence of D. discoideum cells (Figure 5a,d).In this experiment, we can follow unambiguously individual bacteria and D. discoideum cells over both movies.When bacteria were in contact with D. discoideum, F I G U R E 3 Ingested Pseudomonas aeruginosa are destroyed in Dictyostelium discoideum phagosomes.(a) Survival of intracellular P. aeruginosa following ingestion by D. discoideum cells.For each bacterium analyzed, time 0 indicates the time when it is ingested by D. discoideum cell.Extinction of GFP is synonymous with bacterial destruction.The cumulative survival curve of bacteria is shown.P. aeruginosa destruction is slower in kil2 KO cells compared to aplA KO and WT cells.(WT: N = 16 independent experiments, n = 381 bacteria ingested, kil2 KO: N = 9, n = 245 and aplA KO: N = 7, n = 197).(b)Fluorescence level of P. aeruginosa ingested by D. discoideum cells in the minutes preceding bacterial GFP extinction.Data are normalized using 100% as the fluorescence when the bacterium was ingested by the amoeba and 0% as the fluorescence when GFP is extinguished.In this plot, time 0 corresponds to the time when GFP fluorescence became undetectable.Extinction of GFP fluorescence was gradual, extending over a period of 2-3 min (WT: N = 8, n = 60, kil2 KO: N = 4, n = 60 and aplA KO: N = 4, n = 60).(c) The area under the survival curve (AUC) was determined in independent experiments, normalized to WT as described in FigureS7a,b (mean ± SEM).Some of the data presented in this figure was obtained from movies generated in a previous study(Jauslin et al., 2021) and these experiments are indicated with gray dots.Except for kil2 KO cells, all mutant cells destroyed ingested P. aeruginosa as efficiently as WT cells.A.U. arbitrary units.Kruskal-Wallis test (p = 0.0004) followed by Dunn's multiple comparison test.***p = 0.0005.Statistical analysis was performed combining both datasets.

F
Dictyostelium discoideum can destroy Pseudomonas aeruginosa bacteria without ingesting them.Examples of the different fates of extracellular P. aeruginosa following a contact with D. discoideum cells are shown.The insets show separate phase contrast and fluorescence images of bacteria.White arrowheads indicate live bacteria and empty arrowheads indicate lysed bacteria.Scale bar 10 μm.(a) An uningested bacterium is lysed during its contact with a D. discoideum cell.(b) A bacterium is touched (time 0 min), then released (time 1 min) by a D. discoideum cell.4 min later (time 5 min), the bacterial fluorescence disappears, and the phase contrast image indicates that the bacterium is lysed.(c) A bacterium is touched (time 0 min), then released (time 3 min) by a D. discoideum cell, but no bacterial lysis is seen over the duration of the movie (90 min).(d) The bacterial fluorescence level was quantified in the minutes preceding extracellular bacterial lysis.Data are normalized using 100% as the fluorescence when the bacterium is first touched by the amoeba and 0% as the background fluorescence when GFP is extinguished.In this graph, time 0 corresponds to the time when GFP fluorescence disappeared.Extinction of GFP fluorescence occurred abruptly within 30 s, the time separating consecutive pictures.(N = 2 independent experiments, n = 30).(e) Survival of uningested P. aeruginosa following contact with D. discoideum cells (blue line).As a comparison, bacteria that did not come into contact with D. discoideum were also analyzed (red line), as well as bacteria incubated in buffer in the absence of D. discoideum cells (green line).(N = 6 independent experiments, Contact: n = 180, No contact: n = 172, Buffer: n = 180 bacteria).(f) Quantification of the area over the survival curve (AOC) in each experiment as described in Figure S7c,d (mean ± SEM).AU, arbitrary units.Mann-Whitney test.*p ≤ 0.05; **p ≤ 0.01.I G U R E 5 Following a contact with Dictyostelium discoideum, bacteria not lysed within a few hours retained the ability to divide.Pseudomonas aeruginosa and D. discoideum cells were mixed and incubated for 25 min at 21°C (movie 1), then the medium was changed to LB and the temperature increased to 37°C to allow growth of bacteria (movie 2).209 bacteria were touched by D. discoideum during the first 25 min, and 56 of them lysed before the end of movie 1 (a).Of the 153 remaining bacteria, 121 grew during the second incubation (a and b, arrowheads), and 27 lysed during the second incubation (a and c, asterisks).Only 5 bacteria that did not lyse failed to grow during the second incubation.In control conditions (without amoeba cells), most bacteria (190/197) were able to grow during the second incubation (a and d, arrows).These observations indicate that following contact with D. discoideum, bacteria either lyse, or remain fully viable (i.e., capable of growth and division).Scale bar 10 μm.viable, that is, they grew in LB (121/209 = 58%) (Figure 5a,b).Other bacteria (56 + 27/209 = 40%) were lysed during the experiment (Figure 5a,c).Only a very small fraction of bacteria (5/209 = 2%) were not lysed yet failed to grow in LB.As expected, in the absence of D. discoideum cells, most bacteria (190/200; 95%) remained viable (Figure 5a,d).A few bacteria were lysed spontaneously (3 + 4/200 = 3.5%).A very small number of bacteria (3/200 = 1.5%) remained intact but failed to grow when bacteria were not exposed to D. discoideum cells.Together these experiments indicate that when D. discoideum amoebae established a contact with P. aeruginosa bacteria, approximately 50% of the bacteria were heavily damaged, leading to their lysis in the following hours.The remaining 50% appeared unharmed and fully viable.Extracellular destruction was not observed with other bacteria: a small number of K. pneumoniae, E. coli and S. aureus glycosylation (Figure 7b, star).As detailed below, the higher band in the AplA doublet presumably corresponds to the mature form of the protein.Expression of AplA-ALFA in aplA KO cells restored efficient extracellular lysis (Figure 7c,d).To determine if a single SAPLIP domain of AplA is sufficient to restore an efficient extracellular lysis of P. aeruginosa, we produced the domain 1 or the domain 5 of AplA in aplA KO amoeba cells.These proteins were detected by western blot at the expected molecular weight (≈16 kDa for domain 1 and ≈17 kDa for domain 5) (Figure 7b).A band around 80 kDa in cells overexpressing AplA domain 5 (Figure 7b, double stars) may F I G U R E 6 Dictyostelium discoideum NoxA and AplA are required for efficient extracellular lysis of Pseudomonas aeruginosa.(a) Survival curves of extracellular P. aeruginosa following contact with D. discoideum (N = 10 independent experiments, WT: n = 482, noxA KO: n = 266, aplA KO: n = 330 and kil2 KO: n = 210 bacteria).noxA KO and aplA KO amoeba cells lyse extracellular bacteria less efficiently than WT.No defect in bacterial lysis was observed when using kil2 KO cells.(b) Quantification of the defect in extracellular lysis in independent experiments for various mutants compared to WT cells (data are AUCs normalized to WT as described in FigureS7a,b, mean ± SEM).Among all the tested mutants, only noxA KO, aplA KO, alyL KO, and kil1 KO lysed P. aeruginosa significantly less efficiently than WT amoeba.A.U., arbitrary units.Kruskal-Wallis test (p < 0.0001) followed by Dunn's multiple comparison test.***p ≤ 0.001, ****p ≤ 0.0001.
cases, the bacterium is ingested and destroyed in phagosomes or (ii) in ≈ 90% of cases the bacterium remains extracellular.Extracellular contact is sufficient to induce extracellular destruction of ≈ 50% of uningested bacteria.This situation allowed us to assess in parallel the molecular mechanisms involved in intracellular and extracellular destruction of P. aeruginosa.Our observations indicate that efficient intracellular destruction requires the presence of Kil2, a putative Mg 2+ pump present in the phagosomal membrane (Lelong et al., 2011).No other gene product tested in this study was required for efficient intra-phagosomal destruction.On the contrary, extracellular lysis did not require the expression of Kil2, but was decreased by genetic inactivation of noxA and aplA, and to a lesser extent kil1 and alyL.The fact that AplA, NoxA, and AlyL are essential for efficient extracellular but not intracellular destruction demonstrates that intracellular and extracellular destruction of bacteria mobilize largely different molecular mechanisms.AplA, NoxA, and AlyL may either not participate in intracellular destruction of P. aeruginosa, or the existence of redundant destruction mechanisms in phagosomes may make these proteins non-essential in the phagosomal context.

F
I G U R E 8 Dictyostelium discoideum secretes AplA.We expressed AplA-ALFA in aplA KO cells and detected it in cells and in the extracellular medium (SN) by Western blot using a recombinant antibody against the ALFA tag.Cells were incubated in culture medium (HL5) or in starvation buffer (PB*) for 2.5 or 5 h.aplA KO cells cultivated in HL5 were used as a negative control.Cells contained a mixture of high (≈72 kDa) and low (≈68 kDa) molecular weight AplA (black and white arrowheads, respectively).The 72 kDa AplA was secreted most efficiently in starvation medium.many species ranging from other amoebae (17amoebapore-like Apl proteins encoded in the D. discoideum genome(Dhakshinamoorthy, 2018 #33)) to human (Granulysin and Surfactant protein B).In D. discoideum, AplD is expressed during multicellular development and participates in the defense of multicellular slugs against invading bacteria by permeabilizing bacterial Our results indicate that in D. discoideum, expression of either a single saposin domain of AplA or of other Apl proteins (AplB and AplH) in aplA KO cells restored efficient extracellular lysis of P. aeruginosa.This seems to indicate a rather low degree of specificity between different Apl proteins, and between the various saposin domains of each Apl protein.However, the use of efficient expression plasmids presumably leads to overexpression of these various proteins, and a more refined analysis may reveal more subtle differences between different Apl proteins.Within cells, AplA was found in unidentified intracellular granules.The nature and fate of AplA-containing granules remains to be firmly established, although our results clearly suggest that they can fuse with the cell surface since AplA can be released in the extracellular medium (Figure 8).Our results do not necessarily imply a dichotomy between granules fusing with the cell surface and with phagosomes.Indeed, AlyL was previously shown to participate in the destruction of K. pneumoniae in phagosomes (Jauslin et al., 2021), while this study demonstrates that it plays a role in extracellular killing of P. aeruginosa.It seems that, like in other phagocytic cells, cytosolic granules containing bacteriolytic effectors can fuse both with phagosomes and with the plasma membrane of D. discoideum cells.The different properties of the phagosomal and extracellular environment, as well as the characteristics of the bacteria, ultimately determine which bacteriolytic effector is most efficient.In summary, D. discoideum makes use of a vast array of mechanisms to ensure intracellular and extracellular destruction of bacteria.A similar set of mechanisms presumably allows the F I G U R E 9 AplA localizes in unidentified granules in the vicinity of the Golgi apparatus.We expressed AplA-ALFA in aplA KO cells and detected it by immunofluorescence using an anti-ALFA antibody together with a panel of known markers: (a) p25 (plasma membrane and recycling endosomes), (b) p80 (endosomal membranes), (c) VatA (contractile vacuole and endosomal compartments), (d) SctA (pycnosomes), (e) CA1 (lysosomes), (f) Rhesus (contractile vacuole), (g) PDI (endoplasmic reticulum), and (h) Golgi.AplA is detected in a cellular compartment distinct from all compartments detected here.Bar scale, 5 μm.
to destroy bacteria.However, in humans, these different mechanisms have been distributed in a number of different cell types implicated in different facets of innate immunity.
discoideum.Further studies will be necessary to determine if the same mechanisms allow P. aeruginosa to escape phagocytosis by D. discoideum and mammalian cells.D. discoideum amoebae would be at a severe disadvantage if they faced pathogens that can kill them at a distance or during a simple contact, while the amoebae could only kill pathogens after ingesting them.Our results indicate that D. discoideum is capable of lysing P. aeruginosa bacteria without ingesting them.Our observations further indicate that bacteria simply placed in the vicinity of D. discoideum amoebae were lysed to some extent, indicating that at least some D. discoideum bacteriolytic factors can act even without a direct contact between phagocytic cells and bacteria.Extracellular killing of P. aeruginosa by D. discoideum is a logical solution to deal with the pathogenic mechanisms developed by P. aeruginosa.This alternative antibacterial strategy apparently targets a specific weak point of P. aeruginosa, that is, its high sensitivity to extracellular lysis.Indeed, the other bacteria tested in this study (K.pneumoniae, S. aureus and E. coli) are not subject to extracellular lysis.It remains to be seen whether the sensitivity of P. aeruginosa to extracellular lysis is specific to the P. aeruginosa strain used in this study, or is a general property of the whole P. aeruginosa species.It is likely that a thorough search would identify other microorganisms sensitive to extracellular lysis by D. discoideum.The existence of anti-amoebal traits in P. aeruginosa and of additional anti-bacterial traits in D. discoideum presumably reflect the ecological fight between predatory amoebae and their P. aeruginosa prey.4 | MATERIAL S AND ME THODS 4.1 | Cells and reagents D. discoideum DH1