We were interested in investigating to what extend variations in membrane lipid composition impact Lp infection and reasoned that one way to approach this question experimentally is though metabolic reprogramming of the GPL repertoire via exogenous lipid supplementation because eukaryotic cells preferentially utilize fatty acids internalized from interstitial fluids (exogenous dietary fatty acids) over de novo synthesized lipids for GPL assembly. To this end, we infected primary bone-marrow derived murine macrophages (BMMs) with Lp and measured bacterial growth over several infection cycles under conditions where different fatty acids from the two major membrane classes (saturated and monounsaturated) ( Table 1 ) were supplemented in the culture medium for the duration of the infection ( Figure 1A ). These infections were carried out under serum-free conditions to ensure preferential utilization by macrophages of the specific fatty acid tested over other serum-derived lipids. For these assays, the fatty acids were complexed with albumin, which is the main transport carrier of non-esterified fatty acids in interstitial fluids (Spector, 1975; van der Vusse, 2009). A bioluminescent Lp strain encoding the luxR operon under a constitutive promoter was used which allowed us to monitor bacterial growth continuously in high throughput manner (Ondari et al., 2022).

When supplied exogenously myristic (C14:0), stearic (C18:0) and palmitoleic acid (C16:1) significantly altered Lp intracellular growth in primary BMMs in a dose-dependent manner ( Figure 1B ). Overall, all SFA tested reduced Lp growth in primary BMMs infections, although only the differences for myristic and stearic acid treatments were statistically significant ( Figure 1B ). Conversely, MUFAs either significantly enhanced Lp growth during infection (palmitoleic acid) or did not produce a growth phenotype (oleic acid) ( Figure 1B ). An alternative assay measuring the colony forming units (CFUs) recovered at distinct times after infection also produced similar bacterial growth enhancement by palmitoleic acid ( Figure 1C ). Comparable Lp growth patterns were seen in infections of immortalized BMMs (iBMMs) where palmitoleic acid enhanced and all of the SFA except lauric acids (C12:0) reduced bacterial growth ( Supplementary Figures 1A, B ). The mean amount of circulating palmitoleic acid in the serum in a large cohort of 827 young healthy adults from diverse ethnic backgrounds was measured at 133.0 ± 67.2µM (min 27.7µM and max 555.9µM) (Abdelmagid et al., 2015). Therefore, we conclude that palmitoleic acid supplied at physiologically relevant concentrations can promote Lp growth in infections of primary and immortalized BMMs.

The growth-altering phenotypes produced by dietary fatty acids in Lp infections could be a result of a nutrient effect due to direct metabolic utilization by the bacterium, which should be evident when bacterial growth is measured in the absence of host cells. Therefore, we measured Lp growth in AYE liquid cultures in the presence of different fatty acids or vehicle control condition. The initiation of Lp replication was delayed in axenic cultures in the presence of 300µM lauric, myristic, palmitic or oleic acids to various degrees ( Figure 2A ), indicating that these fatty acids can induce measurable direct growth defect in the absence of host cells. Under the same growth conditions Escherichia coli growth in axenic medium was enhanced ( Figure 2B ) consistent with E. coli capacity for metabolic utilization of exogenous fatty acids (Jaswal et al., 2021). Palmitoleic acid at concentrations that enhanced Lp intracellular replication did not affect Lp growth in axenic cultures ( Figure 2A ). Thus, palmitoleic acid in the absence of host cells is not a growth-promoting metabolite for Legionella. We also conclude that Lp tolerance to fatty acids-induced lipotoxicity is reduced compared to E.coli.

For Legionella, the number of bacteria egressing from each infected macrophage is determined by several variables – (i) internalization kinetics; (ii) trafficking and remodeling of the LCV into an ER-like compartment; (iii) maturation of the LCV into an organelle that supports bacterial replication; (iv) synchronization of LCV expansion and bacterial replication and (v) LCV membrane disruption during egress. Even moderate variations in these key parameters can be amplified over successive infection cycles to produce a growth phenotype. We sought to determine which variable(s) is/are directly or indirectly regulated by the presence of exogenous palmitoleic acid.

First, we compared the uptake of BSA-coupled fatty acids by BMMs as different rates of internalization could potentially explain the distinct Lp growth phenotypes we observed during infection. Fatty acids uptake occurs through direct membrane diffusion of the lipid or facilitated transport by membrane transporters (Su and Abumrad, 2009), where the transport rates depend on size and saturation states of the fatty acid (Kampf et al., 2006; LeBarron and London, 2016). One fate of internalized fatty acids is the rapid re-esterification into newly synthesized triacylglycerols for storage into lipid droplets (Walther and Jr., 2012). Thus, we quantified the number of lipid droplets per cell as an indirect readout for fatty acids uptake by iBMMs ( Figure 3 ). After a 4hrs treatment all fatty acids increased the lipid droplets’ size ( Figure 3A ) and significantly increased the number of lipid droplets per cell ( Figures 3A, B ) consistent with increased internalization and incorporation of dietary fatty acids in triacylglycerols. Approximately, a two-fold increase in the number of lipid droplets was observed in MUFAs-treated iBMMs compared to SFA-treated cells potentially as a result of higher transport rates ( Figure 3B ). Nevertheless, an increased lipid uptake is unlikely to account for the Lp growth phenotype caused by dietary palmitoleic acid because all three MUFAs induced lipid droplet formation comparably ( Figure 3B ), yet Lp growth enhancement was specific for palmitoleic acid ( Figure 1B and Supplementary Figure 1A ).

Next, we investigated whether Lp phagocytic uptake was enhanced by dietary palmitoleic acid, which may account for the increased bacterial growth seen in the BMM infection assays. To this end, we pre-treated iBMMs with palmitoleic acid or vehicle control for 16hrs after which Lp internalization was measured at 2hpi (hours-post-infection) using inside/out microscopy approach ( Figure 4A, B ). Palmitoleic acid treatment did not alter the number of internalized bacteria by iBMMs, unlike the disruption of actin polymerization by cytochalasin D which interfered with phagocytosis ( Figure 4B ). Therefore, we conclude that an increase in bacteria uptake is not the cause for the higher bacterial growth supported by palmitoleic acid-laden BMMs.

Because Lp uptake was not enhanced by dietary palmitoleic acid, we next investigated the effect on LCV biogenesis and expansion. Lp replication is initiated in mature LCVs at ~ 4 hours post internalization, after trafficking and remodeling of the LCV into an ER-like organelle is completed (Ivanov and Roy, 2009). While majority of Lp phagocytosed by BMMs successfully evade endosome maturation (> 90%) (Roy et al., 1998) and the organelle remodeling is successfully completed, only ~50% of mature LCVs support robust bacterial replication (Nogueira et al., 2009). Therefore, we investigated whether dietary palmitoleic acid increases the number of LCVs supporting bacterial replication which could account for the growth-enhancement phenotype. To this end, we used a well-established microscopy-based assay to calculate the percentage of LCVs supporting bacterial replication (Nogueira et al., 2009). In this method the percentage of LCV supporting bacterial replication is derived as an index of the percentage of infected cells harboring LCVs that support bacterial replication at 10hpi over the percentage of infected cells at 2hpi ( Figure 5 ). In general, the LCVs in palmitoleic acid-laden iBMMs at 10hpi harbored more bacteria ( Figures 5A, B ) with fewer vacuoles containing less than 3 bacteria as compared to vehicle control (7% vs 23% in vehicle control) ( Figure 5B ). The RV (replicative vacuole) index also reflected the capacity of palmitoleic acid to increase significantly the percentage of LCVs supporting bacterial replication from 58.3% to 81.5% ( Figure 5C ). Moreover, the percentage of large LCVs (bacteria n >13) in palmitoleic acid treated iBMMs doubled from 35% to 71% indicating an overall increased LCV capacity to support bacterial replication. Based on these data we conclude that dietary palmitoleic acid (i) enhances the capacity of LCVs to support Lp replication and (ii) increases the number of LCVs that support bacterial replication.

To gain insight into which aspects of LCV homeostasis are impacted by different fatty acids we performed live-cell microscopy of BMMs infected with GFP-expressing Lp, where cells were imaged every 4 hours over a 2-day period ( Figure 6 ). Using this approach, three distinctly identifiable phases were observed during the LCV lifespan – expansion, stasis and collapse. During the expansion phase the size of the LCV increases, whereas during the stasis phase little measurable change in the LCV size was detected ( Figure 6A ). The collapse phase of the LCV is initiated by an abrupt drop in LCV size accompanied by sudden decrease in intensity of the GFP signal consistent with membrane rupture and is terminated when the GFP signal disappears entirely indicating egress completion ( Figure 6A ). Bulk analysis at the population level revealed that in palmitoleic acid-treated cells the LCV size on average increased significantly faster as compared to vehicle or palmitic acid-treated cells even though comparable number of LCVs were detected across all conditions for the first 20 hours of the infection cycle ( Figure 6B ), which suggest that LCV expansion might be affected. However, from averaged population data of an asynchronous infection we could not determine whether this phenotype is due to earlier initiation of bacterial replication, faster bacterial replication or prolonged LCV lifespan. Therefore, we tracked the size of individual LCVs over their lifespan from each experimental condition to determine the maximum LCV size ( Figure 6D ), the duration of the LCV expansion phase ( Figure 6E ), the entire LCV lifespan ( Figure 6F ) and the LCV rate of expansion ( Figure 6G ). From single particle tracking analysis, we observed that in palmitoleic acid-laden macrophages the duration of LCV expansion on average was prolonged (20hrs vs 16hrs in vehicle control) ( Figure 6E ), the LCV reached a larger peak size (295.4µm² vs 205µm² in vehicle control) ( Figure 6D ), and the LCVs doubled in size quicker (in 5.1 hours vs 6.1 hours for vehicle control) ( Figure 6G ). The individual size traces of 10 randomly selected LCVs showed that bacterial replication did not initiate earlier in palmitoleic acid-laden macrophages compared to vehicle control treated cells ( Supplementary Figure 2 ). In addition, palmitoleic acid treatment shortened the LCV collapse phase resulting in higher percentage of LCVs that broke down within 4hrs (19% vs 6% in vehicle control) ( Supplementary Figure 3 ). Thus, we conclude that palmitoleic acid treatment remodels macrophages in a manner that makes them more permissive for Lp growth by prolonging the LCV expansion phase and accelerating bacterial replication rather than triggering early bacterial replication. In addition, we found evidence for shortened LCV collapse phase, which likely accelerated egress and the initiation of the subsequent infection cycle. The LCV membrane is one element common to all these phenotypes, suggesting that remodeling of host membrane lipidome by palmitoleic acid may be at the center.

To gain insight into the remodeling of the macrophage lipidome during Legionella infection mediated by dietary palmitoleic acid we performed lipidomics analysis. To this end, primary BMMs were uninfected or infected for 8hrs in the presence/absence of 200µM palmitoleic acid followed by bulk lipids extraction and liquid chromatography tandem mass spectrometry analysis ( Figure 7 ). The lipid abundance for 555 distinct lipids across all experimental conditions was determined. Both C16:0 and C16:1 acylated glycerophospholipids increased upon Lp infection, however palmitoleic acid supplementation further increased significantly the amount of C16:1 acylated GPLs but not C16:0 acylated GPLs demonstrating the internalization and specific incorporated of dietary palmitoleic acid into membrane glycerophospholipids ( Figure 7A ). Preferential incorporation of C16:1 acyl chains was detected in phosphatidylcholines (PC) as well as phosphatidylglycerols (PG) and to a lesser extend in phosphoethanolamines (PE) and phosphatidylserines (PS) ( Figure 7A ). GPLs containing C16:1 acyl chain at either sn-1 or sn-2 positions were enriched following palmitoleic acid supplementation indicating certain membrane GPLs are also remodeled via Land’s cycle through direct acyl chain substitution involving PLA₁ and PLA₂ phospholipases (O’Donnell, 2022) ( Figure 7A ). Acyl chain analysis on the top 100 lipids significantly changed in response to infection revealed that palmitoleic acid treatment caused also notable decrease in C18:0 acylated GPLs specifically ( Figure 7B ), which indicates potential replacement of C18:0 acyl chains by exogenous C16:1. The consequence of exchanging saturated C18:0 with monounsaturated C16:1 on membrane GPLs is an overall increase in the degree of unsaturation resulting in enhanced membrane disorder, fluidity and permeability. Increased membrane fluidity at the ER increases ER tolerance to proteostatic stress and membrane rigidification-induced lipotoxicity (Erbay et al., 2009; Jacquemyn et al., 2017).

Dietary palmitoleic acid also altered the infection-induced enrichment of certain membrane lipid classes ( Figure 7C ). For example, PC and ceramides were further enriched by palmitoleic acid treatment, whereas the enrichment of PE was reversed ( Figure 7C ). Predictably the levels of di- and triacylglycerols, which can store fatty acids, also increased likely to accommodate the influx of exogenous palmitoleic acid ( Figure 7C ). Taken together, the data indicates that exogenous palmitoleic acid in Lp-infected macrophages remodels multiple aspects of the host lipidome one of which – replacement of C18:0 with C16:1 in membrane GPLs through preferential incorporation in PC and PG – would increase membrane disorder and fluidity. Potentially, membrane disorder could be an important parameter for sustained optimal LCV membrane expansion.

The lipidomics data prompted us to investigate whether alteration in membrane disorder can account for the growth-enhancing capacity of palmitoleic acid. The Lp intracellular growth defects induced by supplementation of saturated fatty acids, which rigidify membranes upon GPLs incorporation, also are consistent with that idea ( Figure 1B and Supplementary Figure 1A ). In general, the negative impact of exogenous SFAs on LCV homeostasis is evident from the smaller LCV size ( Figures 6C, D ), the decrease in the rate of LCV expansion ( Figure 6G ) and the prolonged egress ( Supplementary Figure 3 ). However, at higher concentrations SFA exhibited some lipotoxicity to both Lp ( Figure 2A ) and BMMs ( Supplementary Figure 4A ); thus, SFA treatment produces multiple phenotypes in both macrophages and Lp making it difficult to attribute the growth restriction phenotype solely to LCV expansion. Hence, we took a different approach by taking advantage of cis-trans geometric and positional isomerism of MUFAs to explore the role of membrane disorder in Lp intracellular growth. The trans-MUFA isomers have linear structures similar to SFAs ( Figures 1A and 8A ) and like SFAs increase membrane rigidification when incorporated in GPLs (Deguil et al., 2011; Tyler et al., 2019). Palmitelaidic acid (trans-C16:1, ω-7) – the geometric isomer of palmitoleic acid (cis-C16:1, ω-7) - reduced Lp intracellular growth in iBMM infections in a dose-dependent manner when supplied exogenously ( Figure 8B ). Importantly, palmitelaidic acid did not impact Lp growth under axenic conditions ( Figure 8C ) demonstrating that this lipid lacked microbiocidal activity and the growth phenotype is likely a result of lipidome reprogramming of the host cell. Similarly, elaidic acid (trans-C18:1, ω-9) - the linear geometric isomer of oleic acid (cis-C18:1, ω-9) – also produced a significant Lp growth defect phenotype in macrophage infections in a dose-dependent manner whereas oleic acid treatment had a neutral effect on Lp growth in biolux assays ( Figure 8B ) and a small but measurable increase in LCV size in live-cell imaging assays ( Figure 6D ). Macrophages exposed to elaidic acid harbored smaller LCVs as compared to oleic acid-treated macrophages ( Figures 6C, D ), which was a result of decreased LCV rate of expansion ( Figure 6G ) and did not affect the overall LCV lifespan ( Figure 6F ). Treatment with elaidic acid also accelerated bacterial egress ( Supplementary Figure 3 ). Elaidic acid did not reduce macrophage viability ( Supplementary Figure 4B ) and had only a minor impact on Lp growth in axenic cultures at high concentrations ( Figure 8C ). Taken together the data from treatments with matched positional isomers demonstrates that membrane rigidifying trans-MUFAs interfere with LCV expansion and Lp growth unlike their geometric non-linear cis-isomers which increase membrane disorder. Thus, we conclude that the physiochemical properties especially pertaining to membrane disorder dictate the optimal rate of niche expansion and housing capacity of the LCV. The distinct Lp growth phenotypes elicited by palmitoleic vs oleic acid indicate that the carbon chain length of the fatty acid is also a determinant.

Robust microbiocidal activity towards Lp in axenic cultures ( Figure 8C ) and in macrophage infections ( Figure 8B ) was observed for sapienic acid - a non-linear positional isomer of palmitoleic acid in which the double bond is at the ω−10 carbon ( Figure 8A ). Sapienic acid is produced by Δ-6-desaturation of palmitic acid by FADS6 and is a major component of human sebum with antimicrobial properties (Ge et al., 2003; Drake et al., 2008). Thus, the position of the double bond on cis-C16:1 fatty acids can have remarkably different biological effects on Legionella viability and intracellular replication.

The specificity of the Lp growth phenotype for palmitoleic acid over other fatty acids raised the possibility that palmitoleic acid might directly or indirectly interfere with cell intrinsic host defenses because palmitoleic acid can also function as a lipokine (Cao et al., 2008). As an adipose-tissue derived lipokine, palmitoleic acid regulates cellular responses through the activation of G protein-coupled receptors independently of membrane GPL remodeling. To investigate whether dietary palmitoleic acid broadly interferes with macrophage capacity to limit bacterial replication we decided to infect cells with other Legionella species that traffic and establish ER-like intracellular niches similar to L. pneumophila (Maruta et al., 1998; Asare and Kwaik, 2007; Ivanov and Roy, 2009; Wood et al., 2015). To this end, we engineered L. jordanis, L. longbeachae and L. dumoffii strains that chromosomally encode the luxR operon under an IPTG-inducible promoter ( Supplementary Figure 5 ) and infected human U937 macrophages in the presence of palmitic or palmitoleic acid ( Figure 9 ). Similar to the data from murine BMMs ( Figures 1B, C and Supplementary Figures 1A, B ), palmitoleic acid but not palmitic acid significantly increased Lp growth in human macrophages demonstrating that the growth-promoting capacity of C16:1 is conserved ( Figure 9 ). For the other Legionella species, palmitoleic acid did not enhance bacterial growth indicating that macrophages are not being affected in a manner that renders them broadly permissive for bacteria replication. Therefore, we conclude that the palmitoleic acid-induced growth advantage is conserved in human macrophages and is specific for L. pneumophila.

Alveolar macrophages are a major cellular niche supporting Lp growth in human pulmonary infections (Nash et al., 1984), however they also perform a central role in the turnover of lung surfactant - a complex mixture of lipids, proteins and carbohydrates - that acts to decrease surface tension at the air–liquid interface of the alveoli in mammalian lungs (Bernhard, 2016). Type II alveolar epithelial (AE2) cells synthesize and secrete pulmonary surfactant, whereas both AE2 and alveolar macrophages carry out surfactant internalization and turnover (Stern et al., 1986; Lopez-Rodriguez et al., 2017). Phospholipids constitute ~70% of pulmonary surfactant with dipalmitoyl-PC (PC16:0/16:0), palmitoyl-myristoyl-PC (PC16:0/14:0) and palmitoyl-palmitoleoyl-PC (PC16:0/16:1) being the major components (Bernhard, 2016). Thus, alveolar macrophages must be highly adapted to resolve the lipotoxic stress brought by continuous uptake and processing of GPLs containing palmitic acid. Eukaryotic cells cope with palmitate induced membrane rigidification by either (i) increasing the abundance of SCD – the enzyme that resolves membrane stress by desaturating palmitate to palmitoleic acid (Erbay et al., 2009; Piccolis et al., 2018; Oshima et al., 2020) or (ii) by transporting palmitate via the lipid chaperone Fatty Acid Binding Protein 4 (FABP4) to mitochondria for β-oxidation (Erbay et al., 2009). Gene expression profiles from single-cell RNAseq analysis of human lungs sourced from the Human Protein Atlas data repository (Karlsson et al., 2021) show very high levels of SCD expression in pulmonary macrophages and in AE2 cells, which would be needed to counteract membrane rigidification during surfactant turnover through SFA-to-MUFA conversion ( Figure 10 ). The expression of SCD5, which encodes the other Δ9-desaturase enzyme allele in humans (Igal and Sinner, 2021) was low in both AE2 cells and macrophages ( Figure 10 ). In addition, FABP4 transcripts were abundant in pulmonary macrophages but not AE2 cells consistent with the need for fatty acid transport to mitochondria for β-oxidation as part of surfactant turnover ( Figure 10 ). Based on our data, palmitoleic acid sourced from surfactant-derived palmitoyl-palmitoleoyl-PC (PC16:0/16:1) is likely to increase the permissiveness of alveolar macrophages for Lp replication. Similarly, the high Δ9-desaturase activity in pulmonary macrophages would result in desaturation of surfactant-derived palmitic acid into palmitoleic acid and thus increase macrophage capacity to support Lp growth. Therefore, it is possible that the adaptive responses guarding against lipotoxic stress during the continuous turnover of pulmonary surfactant may inadvertently increase alveolar macrophages permissiveness for Lp replication.

All L. pneumophila strains used in this study were derived from the L. pneumophila serogroup 1, strain Lp01 (Berger and Isberg, 1993) and have a clean deletion of the flaA gene to avoid NLRC4-mediated pyroptotic cell death response triggered by flagellin when BMMs from C57BL/6J mice are infected with flagellated Legionella. The following Lp strains were used in this study: (1) Lp01 ΔflaA; (2) Lp01 ΔflaA Ptac-GFP, which expresses GFP under isopropyl-beta-D-thiogalactoside (IPTG)-inducible promoter; (3) Lp01 ΔflaA LuxR strain in which the luxR operon (luxCDABE) from Photorhabdus luminescens was inserted via homologous recombination on the bacterial chromosome downstream of the icmR promoter (Ondari et al., 2022). The following bioluminescent Legionella strains that encode the LuxR operon under the IPTG-inducible Ptac promoter on the chromosome were generated as described below: L. dumoffii strain NY 23, L. jordanis strain BL-540 and L. longbeachae strain Long Beach4. A list of the bacterial strains used in this study is provided in Table 2 .

Legionella strains were grown for 2 days at 37°C on charcoal yeast extract (CYE) plates [1% yeast extract, 1%N-(2-acetamido)-2-aminoethanesulphonic acid (ACES; pH6.9), 3.3mM l-cysteine, 0.33mM Fe(NO₃)₃, 1.5% bacto-agar, 0.2% activated charcoal] (Feeley et al., 1979). For all infections, each Legionella strain was harvested from CYE plates and grown in liquid cultures. For liquid cultures, bacteria from day 2 heavy patches grown on CYE plates were suspended to optical densities of 0.1 or 0.3 in 2.5ml of complete ACES buffered yeast extract (AYE) broth (10mg/ml ACES: pH6.9, 10mg/ml yeast extract, 400mg/l L-cysteine, 135mg/l) supplemented with 100µg/ml streptomycin and grown aerobically at 37°C to early stationary phase (between 18-22 hours, to OD_{600 =} 3.0–4.0). Liquid cultures of the GFP-expressing strains were supplemented with 10μg/ml chloramphenicol.

For the generation of inducible bioluminescent Legionella strains, genomic loci that contain an intergenic MfeI restriction site were chosen as follows: for L. dumoffii (Ld3X locus [315,471 → 317,917 nt]); for L. jordanis (Lj5X locus [2,542,670 → 2,545,129 nt]); for L. longbeachae (Ll7X locus [854,416 → 856,719 nt]). The synthetic alleles that encode the luxR operon under the control of the tac promoter (Ptac) were built in stepwise fashion. Step one - the target genomic loci were PCR amplified from genomic DNA from each respective species with the following primers: Ld3X (FC_BglII_MfeI003X and RC_SacI_MfeI003X), Lj5X (FC_BglII_MfeI005X and RC_SacI_MfeI005X) and Ll7X (FC_BglII_MfeI007X and RC_SacI_MfeI007X). Each PCR amplicon product was double digested with BglII/SacI and cloned into the suicide plasmid pSR47s that was digested with BamHI/SacI. Step two - the Ptac regulon containing lacI (1,572 bp fragment) was PCR amplified using the pMR33 plasmid (Ramsey et al., 2012) as template with the primers FC_EcoRI_Ptac regulon and RC_MfeI_Ptac regulon and was cloned in the MfeI-digested pSR47s derivatives (from step 1) that contained the Ld3X, Lj5X and Ll7X loci. Correct directionality of the Ptac regulon within each locus was confirmed by sequencing. Step three - the pSR47s intermediates from step 2 were digested with BamHI and NotI and the luxR operon (5,854 bp fragment), which was similarly digested, was introduced to complete the construction of the pSR47s-Ld3X-Ptac-LuxR, pSR47s-Lj5X-Ptac-LuxR, and pSR47s-Ll7X-Ptac-LuxR plasmids.

The replacement of the target loci on the Legionella chromosome was carried out by allelic exchange via double homologous recombination and was confirmed via bioluminescence emission; kanamycin sensitivity and sucrose resistance (Merriam et al., 1997). For allelic exchange, plasmids were introduced in the respective Legionella species by either tri-parental mating (pSR47s-Ld3X-Ptac-LuxR and pSR47s-Lj5X-Ptac-LuxR) or via electroporation (pSR47s-Ll7X-Ptac-LuxR) with the following parameters voltage-1800, capacitance- 25µF and resistance- 200ohms. All primer sequences are listed in Table 3 and plasmid information is included in Table 4 .

The following reagents were purchased from Cayman Chemicals - lauric acid (cat# 10006626), myristic acid (cat# 13351), palmitic acid (cat# 10006627), palmitoleic acid (cat# 10009871), stearic acid (cat# 1011298), oleic acid (cat# 9004089), elaidic acid (cat# 90250), sapienic acid (cat #9001845), palmitelaidic acid (cat# 9001798), and Nile Red (cat#30787). Fatty acid-free Bovine Serum Albumin (BSA) was purchased from VWR (MP Biological, cat# 152401). Hoechst 33342 was purchased from ThermoScientific (Life Technologies, cat # H3570). The purified α-Lp IgY chicken antibody was custom generated by Cocalico Biologicals against formalin-killed bacteria (Abshire et al., 2016). The α-chicken-IgY-TRITC (cat#A16059) was purchased from ThermoFisher Scientific.

C57BL/6J mice were purchased from Jackson Laboratories and housed at LSUHSC-Shreveport animal facility. Ethical approval for animal procedures for experiments in this study was granted by the Institutional Animal Care and Use Committee at LSUHSC-Shreveport (protocol# P-15-026).

Primary BMMs derived from bone marrow progenitors isolated from C57BL/6J mice were cultured on 10 cm petri dishes in RPMI 1640 with L-glutamine (BI Biologics, cat #01-100-1A) supplemented with 10% v/v FBS (Atlas Biologics, cat #FS-0500-AD), conditioned medium from L929 fibroblast cells (ATCC, CCL-1) (20% final volume) and penicillin/streptomycin (ThermoScientific, cat# 15140163) at 37°C with 5% CO₂. Additional medium was added at days 3, 5 and 7 from the start of differentiation. Macrophages were collected and seeded for infections at day 9. Immortalized BMMs from C57BL/6J mice were a kind gift from Dr. Jonathan Kagan (Harvard Medical School) (Evavold et al., 2021) and were propagated in High-Glucose DMEM supplemented with L-glutamine (ThermoScientific, cat# BP379-100), 10% v/v FBS (Atlas Biologics, cat# F-0500-DR), 1mM sodium pyruvate (Quality Biologicals, cat#116-079-721) and penicillin/streptomycin. U937 monocytes (ATCC, CRL1593.2) were cultured in RPMI-1640 with L-glutamine (BI Biologics, cat #01-100-1A), 10% v/v FBS and penicillin/streptomycin at a temperature of 37°C in the presence of 5% CO₂.

Fatty acid stocks were prepared in ethanol at the maximum possible concentration. Tissue-culture grade fatty acid-free BSA (MP Biomedicals, cat#152401) was dissolved in 150mM NaCl at a final concentration of 1.6mM and was filter sterilized through a 0.22µm PES filter. Fatty acid stock solution was mixed with the BSA stock solution in a 6:1 molar ratio to a final FA concentration of 10mM in 1.5ml Eppendorf tubes, which were subsequently incubated at 37°C for 60min while being vigorously vortexed every 10min for at least 30sec. For the preparation of the BSA-only vehicle control stocks equivalent ethanol amount without the fatty acid was complexed with BSA.

BMMs were seeded in 6 well tissue culture plates (2x10⁶ cells/well) in re-plating medium (RPMI 1640 with L-glutamine, 10%FBS, 10% M-CSF-conditioned medium) for 2hrs followed by 14hr incubation with serum-free RPMI (SF-RPMI). Cells were infected with liquid culture-grown Lp01 ΔflaA at MOI=5 (Multiplicity of infection denotes the ratio of the number of bacteria in the inoculum/number of host cells) for 8hrs in SF-RPMI. Plates were centrifuged for 5 min at 800rpm to increase bacteria attachment to host cells immediately after the inoculum was added. Palmitoleic acid complexed with BSA (200µM final concentration) was added at 2hpi in SF-RPMI. Each infection condition in every experiment was carried out in two technical replicates, which were pooled together after sample preparation. Samples from eight biological repeats were prepared for LCMS analysis as detailed below.

For all LCMS methods LCMS grade solvents were used. Media was removed and discarded from cell culture samples in 6 well plates and washed briefly with 1ml of 0.9% sodium chloride. Wash was gently removed, and 0.4ml of ice-cold methanol was added to each well. Plates were placed on ice for 5 min. To each sample 0.4ml of water was added and the well was scraped to suspend cell-associated solids. Samples were transferred to microtubes and 0.4ml of chloroform was added to each. Samples were agitated for 30 minutes at 4°C and subsequently centrifuged at 16,000g for 20 min to induce layering. An 0.4ml aliquot of the organic layer (bottom) was dried in a Savant SpeedVac SPD130 (Thermo Scientific) and resuspended in 1ml of 5µg/ml butylated hydroxytoluene in 6:1 isopropanol:methanol.

LCMS grade water, methanol, isopropanol and acetic acid were purchased through Fisher Scientific. Bulk lipids were analyzed as previously described (Schwarz et al., 2020; Foliaki et al., 2023). A Shimadzu Nexera LC-20ADXR was used for chromatography across a Waters XBridge® Amide column (3.5µm, 3mm X 100mm) with a 12-minute binary gradient from 100% 5mM ammonium acetate, 5% water in acetonitrile apparent pH 8.4 to 95% 5mM ammonium acetate, 50% water in acetonitrile apparent pH 8.0 was used to separate organic fraction samples. Lipids were detected using a Sciex 6500+ QTRAP® mass spectrometer with polarity flipping. Lipids were detected using scheduled MRMs based on acyl chain product ions in negative mode or a combination of invariant or neutral loss combinations in positive mode. All signals were integrated using MultiQuant® Software 3.0.3. Signals with greater than 50% missing values were discarded and remaining missing values were replaced with the lowest registered signal value. All signals with a QC coefficient of variance greater than 40% were discarded. Filtered datasets were total sum normalized prior to analysis. Single and multi-variate analysis was performed in MarkerView® Software 1.3.1. Datasets were z-scaled for display across the displayed groups. Z-score sums were calculated by summing all lipid signals with a given lipid characteristic in the z-scaled dataset that passed a statistical criterion. The entire lipidomics data set is available at Figshare via https://doi.org/10.6084/m9.figshare.24309139.v1.

Cells were seeded on coverslips in re-plating medium (DMEM with L-glutamine, 10% FBS, sodium pyruvate) overnight at 3x10⁵ cells/well seeding density in 24-well plates. Cells were cultured in serum-free RPMI (SF-DMEM) for 6hrs and then were treated with either 48µM BSA alone or 150µM BSA-complexed fatty acid in SF-DMEM for 4hrs. Cells were gently washed with warm PBS (3X) and fixed with 2%-paraformaldehyde for 30min at ambient temperature. For quantitative analysis of lipid droplets, cells were stained with 200nM NileRed in PBS for 16hrs at 4°C and with Hoechst for 60min at ambient temperature. Coverslips were subsequently washed with PBS (5X) and mounted with ProLong Gold antifade reagent (ThermoFisher) onto glass slides. Images were captured with inverted wide-field Keyence BZX-800 microscope. Image acquisition parameters - Hoechst (ExW 395nm/EmW 455nm) and NileRed (ExW 555nm/EmW 605nm). The z-axis acquisition was set based on the out-of-focus boundaries and the distance between individual z-slices was kept at 0.3µm. Image analysis was performed with Keyence imaging software and the number of lipid droplets per cell was enumerated.

BMMs were seeded on coverslips as described in the previous section. Following attachment, cells were cultured in SF-RPMI containing either 300µM Palmitoleic acid complexed with BSA or BSA alone for 16hrs. Next, BMMs were infected with liquid culture grown Lp01 ΔflaA pTac::GFP at MOI=10 for 2hrs. One set of the BSA-treated cells was treated with 5µM cytochalasin D at 30min prior to the infection to block phagocytosis. After cells were washed with warm PBS (3X), the infection was stopped by the addition of 2% paraformaldehyde for 60min at ambient temperature. The surface-associated bacteria were immunolabeled without plasma membrane permeabilization with chicken α-Legionella IgY antibody for 90 min in PBS containing goat serum (2% vol/vol). Next, coverslips were washed with PBS (3X) and were stained with tetramethyl rhodamine conjugated goat α-chicken IgY (ThermoFisher, cat# A16059) at 1:500 dilution and Hoechst at 1:2000 dilution for 60min in PBS containing goat serum (2% vol/vol). Coverslips were mounted with ProLong glass anti-fade mountant (ThermoFisher, cat# P36984) onto slides.

Microscopy analyses of infected cells. Images were captured with inverted wide-field Keyence BZX-800 microscope. Image acquisition parameters - Hoechst (ExW 395nm/EmW 455nm); GFP (ExW 470nm/EmW 525nm) and TRITC (ExW 555nm/EmW 605nm). The z-axis acquisition was set based on the out-of-focus boundaries and the distance between individual Z-slices was kept at 0.3µm. Only linear image corrections in brightness or contrast were completed. For each condition, over 100 bacteria were imaged and scored as either intracellular (single positive – green only) or not-internalized (double positive – green/red).

Liquid cultures of Lp01 ΔflaA LuxR in complete AYE were set-up at starting OD_{600 =} 0.4 from plate grown bacteria (day-2 heavy patches) and were distributed in white-wall clear-bottom 96-well plates (Corning, cat# 3610). BSA-complexed fatty acids were added at the beginning of the assay at the indicated concentrations. All conditions in all assays were performed in technical triplicates. Plate was incubated in a luminometer (Tecan Spark) at 37°C for 24hrs. Optical density (OD = 600nm) data were automatically collected every 10 mins after the cultures were agitated for 180sec (double orbital rotation, 108rpm).

Primary BMMs were seeded at 1x10⁵ cells per well in white-wall, clear bottom 96-well plates (Corning cat# 3610) in re-plating medium (RPMI 1640 with L-glutamine, 10% FBS, and 10% M-CSF-conditioned medium) for 2hrs. Immortalized BMMs were seeded in re-plating medium of High Glucose-DMEM with L-glutamine (5.5ml of 100X, ThermoScientific (cat# BP379-100), FBS [10% final volume, Atlas Biologics, (cat# F-0500-DR)], sodium pyruvate (5.5ml of 100mM, Quality Biologicals (cat# 116-079-721), and penicillin/streptomycin [5.5ml of 100X, ThermoScientific (cat# 15140163)]. For human U937 macrophages infections, U937 monocytes were seeded at a density of 1×10⁵ cells per well in white-wall clear-bottom 96-well plates and differentiated into macrophages after 24hr treatment with 10ng/ml Phorbol 12-myristate 13-acetate (Adipogen) after which cells were cultured without PMA and antibiotics for additional 48hrs prior to infection. Cells were cultured with serum-free medium for 14hrs (BMMs and U937 macrophages) or for 6hrs (iBMMs) prior to infection. Infections were carried out under serum free conditions with liquid culture grown Lp01 ΔflaA icmRp-LuxR at MOI=5. BSA-complexed fatty acids were added at 2hpi at the indicated final concentration. Plates were kept in a tissue culture incubator at 37°C with 5% CO₂ and periodically, the bioluminescence output from each well was acquired (integration time 2 or 5sec). The data is presented as fold change (Δ) in bioluminescence from the T₀ reading. All conditions in all assays were performed in technical triplicates.

Macrophages were seeded, infected, and treated as listed in “Bioluminescence assay for Legionella intracellular replication”. All infection conditioned were synchronized at 2hpi by washing each well three times with warm PBS. Plates were kept in a tissue culture incubator at 37°C and 5% CO₂ and periodically removed to collect bacteria. For bacterial recovery, cell culture medium from the well was moved into a 1.5ml Eppendorf tube and replaced with 100µL of sterile water for lysis of the infected macrophages via hypotonic shock and the plate was returned to the incubator for 10mins. Next, the contents of the well were pipetted at least 20 times and were collected in the respective Eppendorf tube which already contained the cell culture medium. Eppendorf tubes were vortexed for 30sec and contents (~300µL) were serially diluted five times for a total of six dilutions. 25µL from each dilution was plated on a CYE agar plate and incubated until colonies were visible. Colonies from a single dilution were counted and CFUs were calculated for each condition. All infection conditions for every experiment were performed in three technical replicates. The data is presented as fold change (Δ) in recovered CFU over the CFUs recovered after synchronization.

iBMMs were seeded on coverslips at 3x10⁵ cells per coverslip in re-plating medium (DMEM with L-glutamine, 10% FBS, sodium pyruvate). After overnight incubation at 37°C and 5% CO₂, cells were cultured with serum-free medium and infected with Lp01 ΔflaA GFP+. At 2hpi, infections were gently washed with warm PBS (3X) to remove extracellular bacteria. One infection condition was stopped at 2hpi, the remaining infections were allowed to proceed until 10hpi in the presence of serum-free media containing either BSA only, 300µM palmitoleic acid or 300µM palmitic acid. At 10hpi, cells were washed with warm PBS (3X) and the infection was stopped by the addition of 2% paraformaldehyde for 60min at ambient temperature. Next, coverslips were stained with Hoechst (1:2000 dilution) for 60min and were mounted with ProLong glass antifade mountant (ThermoFisher, cat# P36984) onto slides. The number of infected cells and the number of bacteria per LCV were enumerated for every condition. The Replicative Vacuole (RV) Index is calculated as the ratio of the percentage of infected macrophages at 10hpi that harbor LCV with >6 bacteria over the percentage of infected macrophages at 2hpi.

For infections, BMMs were seeded in 96-well black-wall clear-bottom plates (Corning cat# 3904) at 8.0×10⁴ cells per well in re-plating medium (RPMI-1640 with L-glutamine, 10% FBS, and 10% M-CSF-conditioned medium) for 2hrs and then were serum starved for 14hrs in serum-free phenol red-free DMEM (Gen Clone 25-501C). All infections were carried out in serum-free medium supplemented with 1mM IPTG. BSA-complexed fatty acids were added at 2hpi. Final media volume was 150µl per well. In each experiment, all conditions were performed in technical triplicates. Plates were centrifuged (1,000rpm, 5min) and were loaded into the IncuCyte™ S3 housing module. The IncuCyte™ S3 HD live-cell imaging platform (Sartorius) is a wide-filed microscope mounted inside a tissue culture incubator and is run by the IncuCyte control software. For each well, four single plane images in bright field and green (ExW 440-480nm/EmW 504-544nm) channels were automatically acquired with S Plan Fluor 20X/0.45 objective every four hours. Images were analyzed with the IncuCyte Analysis software. For LCV analysis, bacterial fluorescence was used to generate a binary mask to define individual LCV objects and to measure the object’s area size (GCU x µm²) and the number of objects per image. For single object LCV tracking, the LCV area was recorded for ~100 LCVs per condition for every timeframe from appearance of the LCV to its disappearance. The IncuCyte™ S3 imaging analysis was performed in the Innovative North Louisiana Experimental Therapeutics program (INLET) core facility at LSU Health-Shreveport. GCU – Green Calibrated Units.

BMMs were seeded at 1x10⁵ cells per well in white-walled, clear bottom 96-well plates (Corning cat# 3610) in re-plating medium (RPMI 1640 with L-glutamine, 10% FBS, and 10% M-CSF-conditioned medium) until attached. Macrophages were serum-starved overnight prior to treatment with different fatty acids. Cell viability was assessed continuously based on ATP production with the RealTime-Glo™ MT Cell Viability Assay kit (Promega, cat# G9711) through bioluminescence output using the TECAN Spark plate reader.

Single cell RNAseq data from normal human lung samples for SCD, SCD5 and FABP4 was sourced from The Human Protein Atlas resource platform (https://www.proteinatlas.org/)(Uhlén et al., 2015; Karlsson et al., 2021). For each transcript the original sourced data is presented in normalized protein-coding transcripts per million (nTPM). For each cell group the transcript expression data is pooled from all representative cell types. The original gene expression data can be found through these links: SCD (www.proteinatlas.org/ENSG00000099194-SCD/single+cell+type/lung); SCD5 (www.proteinatlas.org/ENSG00000145284-SCD5/single+cell+type/lung); FABP4 (www.proteinatlas.org/ENSG00000170323-FABP4/single+cell+type/lung).

Calculations for statistical differences were completed with Prism v10.0.3 (GraphPad Software). The statistical tests applied for the different data sets are indicated in the figure legends and the resultant p-values are shown in the figures.