The short chain fatty acid butyrate prevents intracellular replication of Legionella by regulating cysteine levels in macrophages

28 Macrophages can prevent infections from intracellular pathogens by restricting access to 29 essential nutrients, termed nutritional immunity. With the exception of tryptophan depletion, 30 it is unclear if other amino acids are similarly regulated in infected macrophages. Here, we 31 show that the expression of nutrient transporters in Legionella -infected macrophages is 32 modulated by the short chain fatty acid butyrate. Butyrate prevented the upregulation of the 33 cystine/glutamate exchanger, Slc7a11, in macrophages infected with L. pneumophila , which 34 decreased cellular cysteine levels. Butyrate and the Slc7a11 inhibitor erastin impaired 35 intracellular Legionella replication in macrophages in vitro , with these being restored by 36 exogenous supplementation with cysteine. Butyrate caused increased histone acetylation in 37 infected macrophages, and pan- and class II HDAC inhibitors also restricted intracellular 38 Legionella growth in a cysteine-dependent manner. Intranasal administration of butyrate 39 reduced L. pneumophila lung burdens in mice. Our data suggest that butyrate alters the 40 metabolism of macrophages to promote nutritional immunity by decreasing cysteine levels 41 and that this can be harnessed to treat bacterial lung infections. 42 immunoblotting. Tubulin is shown as loading control. Representative of two independent experiments. d, CFUs of ∆flaA infected BMDMs treated with butyrate, QVP or ABT-737 at 968 6 and 48 hours post infection. Mean and SEM from at least three independent experiments. e, 969 Uninfected and ∆flaA infected BMDMs treated with or without butyrate were analysed for 970 cell death using live-cell imaging. Mean and SEM three independent experiments.


Introduction 44
Macrophages are front line immune cells capable to sensing and eliminating invading 45 microbes. The delivery of microbes to hydrolytic lysosomes promotes cell-intrinsic immunity 46 (1). The induction of additional anti-microbial mechanisms are aimed at pathogens that are 47 able to survive in macrophages (2). This includes nutritional immunity whereby macrophages 48 limit the availability of essential nutrients to prevent intracellular microbial replication. 49 Nutritional immunity is best understood in the case of iron and zinc which can be selectively 50 depleted in pathogen-containing vacuoles and thus prevent microbial survival (3)(4)(5). 51 Similarly, sensing of microbes induces the expression of indoleamine 2,3-dioxygenase 1 52 (IDO1) in macrophages, which depletes cellular tryptophan levels and thus restricts 53 During infections, macrophages rewire their metabolic pathways to induce anti-microbial and 80 inflammatory responses, referred to as immunometabolism (19). There is limited information 81 on how the changes in host metabolism affect intracellular pathogens that replicate 82 preferentially in macrophages, such as Legionella (20,21). Furthermore, the nutritional 83 environment within infection sites affect macrophage responses, as exemplified by the 84 microbial short chain fatty acids (SCFAs) acetate, propionate and butyrate, which reach 85 millimolar concentrations in the gut due to bacterial fermentation of fibre or consumption of 86 SCFA-rich foods (22). Butyrate dampens inflammation in the gut by reducing the expression 87 of inflammatory cytokines in macrophages likely by inhibiting histone deacetylase activity 88 (23). As a fatty acid, butyrate is utilized by gut macrophages and other cells as a carbon 89 source, augmenting metabolic and immune pathways that dampen inflammation (24). 90 Furthermore, sensing of butyrate by cell surface receptors, such as G protein-coupled 91 receptors (GPRs), stimulates the differentiation of immune cells to control gut inflammation 92 (25). Intake of butyrate or fibre increases serum butyrate levels and thus affects macrophage 93 and immune responses in other organs, including in the lung (24,(26)(27)(28). Because of the anti-94 inflammatory effects of butyrate, the SCFA is currently pursued in preclinical and clinical 95 trials to improve outcomes in inflammatory diseases, such as Crohn's disease, type-1 96 diabetes, allergy, and asthma (29,30). The physiological consequence of SCFAs in 97 infections, however, has resulted in contrasting outcomes, as butyrate is beneficial in some 98 bacterial and viral infections (27,(31)(32)(33)(34), but detrimental in other infectious diseases (35). 99 100 Here, we reveal that butyrate affects the metabolism of macrophages and thus prevents 101 intracellular replication of Legionella. We show that butyrate controls the expression of cell 102 surface solute carriers, such as Slc7a11 (also known as cystine/glutamate exchanger, xCT) 103 thereby controlling amino acid levels in infected macrophages. Intranasal butyrate 104 administration reduced bacterial lung burdens, suggesting that SCFAs increase the ability of 105 macrophages to control infections by restricting amino acid availability as a potential new 106 therapeutic option. 107 108

Results 109
Butyrate decreases intracellular L. pneumophila burdens in macrophages. 110 To test whether short chain fatty acids affect intracellular Legionella survival, we infected 111 bone marrow-derived macrophages (BMDMs) with L. pneumophila deficient in flagellin 112 expression (∆flaA). Unlike wild type L. pneumophila, which induces rapid macrophage cell 113 death due to the activation of the NLRC4/caspase-1 inflammasome (36), ∆flaA was able to 114 replicate in BMDMs as determined by colony forming units 48 hours post infection (Fig 1a). 115 L. pneumophila growth under these conditions is restricted to BMDMs, as GFP-expressing 116 bacteria were primarily detected inside macrophages but not in the culture media (Fig. 1b). 117 The supplementation of the macrophage culture media with the short chain fatty acids, 118 acetate, propionate, and butyrate, during macrophage infections markedly reduced L. 119 pneumophila burdens in BMDMs at 48 hours post infection (Fig 1a). Intracellular bacterial 120 burdens remained unaffected at 6 hours post infection (Fig 1a), suggesting that SCFAs do not 121 markedly affect bacterial uptake and elimination, but rather prevent subsequent bacterial 122 growth. To identify which short chain fatty acid affects intracellular Legionella, BMDMs 123 were exposed to acetate, propionate and butyrate separately. While acetate had no effect on 124 ∆flaA burdens, butyrate and propionate prevented increased Legionella numbers at 48 hours 125 post infection (Fig 1a). In the following experiments, we focused on butyrate to probe 126 specific mechanisms. Live-cell imaging of GFP-expressing L. pneumophila confirmed that 127 butyrate had little effect on initial BMDM infections, but prevented subsequent increased 128 fluorescent signals, consistent with a lack of intracellular growth (Fig 1b). 129 130 Propionate and acetate can prevent axenic Legionella growth in rich culture media by acting 131 as signalling molecules (37). Similarly, 4 mM and higher concentrations of butyrate 132 markedly prevented axenic ∆flaA replication, which was not observed with Salmonella or E. 133 coli (Fig S1a, b and c). Butyrate concentrations of 1mM and lower only marginally reduced 134 axenic L. pneumophila growth rates, but prevented intracellular growth in BMDMs,135 suggesting that the SCFA may not directly affect intracellular bacteria (Fig S1a and Fig S1d). 136 In contrast to BMDMs, 1mM butyrate failed to prevent L. pneumophila growth in the human 137 macrophage cell line THP1 (Fig S1e) but also in immortalized mouse macrophages 138 (iBMDMs) (Fig S1f), suggesting altered responses in cell lines rather than species 139 differences. Recent studies indicated that butyrate can affect macrophage differentiation from 140 bone marrow progenitor cells or blood monocytes (27,33). The presence of butyrate during 141 BMDM differentiation resulted in reduced intracellular L. pneumophila numbers, similar to 142 when butyrate was administered at the time of BMDM infections ( Fig S1g). Butyrate still 143 caused reduced intracellular L. pneumophila burdens when the SCFA was used to pre-treat 144 BMDMs but was removed during infections (Fig 1c). Butyrate also affected intracellular L. 145 pneumophila numbers when applied after infections were established (post-infection) (Fig  146   1c). Butyrate also reduced intracellular L. pneumophila numbers in isolated mouse alveolar 147 macrophages (Fig 1d). As reported previously, butyrate can stimulate autophagy as evidenced 148 by decreased p62 levels and increased lipidation of the autophagosome LC3 member 149 GABARAP in uninfected BMDMs (Fig S1h). L. pneumophila, however, inhibits autophagy 150 to ensure intracellular survival (12). Consistent with this notion, butyrate failed to markedly 151 reduce p62 levels in infected BMDMs (Fig S1h). Inhibiting autophagy with Bafilomycin A 152 did not increase bacterial burdens in butyrate treated BMDMs (Fig S1i). Taken together, out 153 data shows that butyrate treatment of BMDMs and alveolar macrophages causes reduced 154 intracellular L. pneumophila numbers independent of autophagy. 155 156

Butyrate controls the expression of nutrient transporters during infections 157
To understand how butyrate affects macrophage responses in Legionella infections, we next 158 performed RNAseq analysis of butyrate-treated BMDMs. Consistent with previous studies 159 using purified LPS, butyrate prevented the upregulation of cytokines and mediators of 160 inflammation during L. pneumophila infection, although some cytokines such as CSF3 were 161 induced (Fig 2a and b). Genes that were up-or downregulated in butyrate treated and infected 162 BMDMs were associated with innate immune responses, LPS and inflammation based on 163 GO-term analysis (Fig 2a). In addition, we noticed that genes associated with metabolism 164 were similarly regulated by butyrate. These included several solute carriers that were 165 upregulated in L. pneumophila-infected BMDMs compared to uninfected macrophages (Fig 166 S2a), but markedly reduced in butyrate-treated macrophages, such as the cysteine/glutamate 167 exchanger (Slc7a11), importer of organic anions (Slco3a1), lysosomal copper (Slc31a2) and 168 glucose (Slc2a6) and mitochondrial amino acids (Slc25a44) (Fig 2b). Butyrate treatment 169 upregulated additional solute carriers that were otherwise largely unaffected by L. 170 pneumophila infections, such as Slc45a4 an alternative sugar transporter (Fig 2b). Besides 171 transporters, butyrate also regulated the expression of metabolic enzymes, such as IRG1 172 (Acod1) which was highly upregulated in infected BMDMs and generates the metabolite 173 itaconate known to inhibit L. pneumophila growth (38). While Glut1 (Slc2a1) was 174 upregulated in infected BMDMs, independent of butyrate, other glycolytic enzymes remained 175 relatively unchanged during infections, including glycolytic (Hk1, Tpi1, Gpi1, Pdha1, Pgm2, 176 Pfkfb3) and TCA enzymes (Mdh1, Aco1) (Fig 2b). Butyrate stimulated the expression of 177 hexokinase 1 (Hk1) as well as TCA cycle enzymes malate dehydrogenase 1 and aconitase 1, 178 regardless of infection (Fig 2b). This suggests that butyrate affects multiple cellular metabolic 179 pathways in macrophages. Butyrate also modulated gene expression in uninfected BMDMs, 180 but the effects were not as pronounced as in infected BMDMs, further suggesting that 181 butyrate may primarily affect pathways associated with infections in macrophages ( Fig S2b). 182

183
To gain insights into how metabolic pathways are regulated by butyrate and whether this 184 affects intracellular L. pneumophila replication, we focused on Slc7a11 as this transporter 185 was the most highly upregulated gene in infected BMDMs and its expression was abrogated 186 by butyrate treatment (Fig 2b). We confirmed by qRT-PCR and immunoblot analysis that 187 Slc7a11 is induced in L. pneumophila infection but not in the presence of butyrate (Fig 2c  188 and d). The Slc7a11 inhibitor, erastin, reduced intracellular L. pneumophila burdens in 189 BMDMs (Fig 2e), suggesting that the effects of butyrate may be connected to Slc7a11 190 downregulation. Inhibition of Slc7a11 via erastin can trigger the programmed cell death 191 pathway of ferroptosis (39). Treatment with the ferroptosis inhibitor, ferrostatin, however, 192 failed to restore L. pneumophila burdens during erastin treatment and ferrostatin by itself did 193 not affect L. pneumophila infections (Fig 2e). Butyrate treatment can also induce apoptosis 194 (40), which does abrogate Legionella replication (41). Consistent with this notion, we noticed 195 that butyrate reduced the expression of the anti-apoptotic BCL-2 member, MCL-1, and 196 induces caspase-3 expression (Fig 2a). Immunoblot analysis confirmed that butyrate, but not 197 propionate, reduced MCL-1 levels in uninfected BMDMs, but not infected BMDMs (Fig 198 S2c). However, butyrate did not cause caspase-3 cleavage, including in L. pneumophila 199 infected BMDMs (Fig S2c). Inhibition of caspase activity, using the pan-caspase inhibitor 200 QVD, failed to restore intracellular L. pneumophila burdens after butyrate treatment, 201 although QVD restored L. pneumophila replication in control treatment with ABT-737 which 202 induced apoptosis ( Fig S2d). Finally, butyrate failed to trigger increased cell death in infected 203 macrophages, suggesting that the decreased CFUs are not due to macrophage cell death ( pneumophila-infected BMDMs treated with butyrate (Fig 3a). Similarly, there was a trend 219 towards lower levels of the cysteine-related metabolites glutathione and g-glutamylcysteine in 220 butyrate-treated macrophages, including those infected with L. pneumophila (Fig S3a and b). 221 Besides Slc7a11-mediated cystine uptake, macrophages can utilise cysteine directly via 222 different transporters. We thus tested whether the addition of cysteine in the culture media 223 would affect L. pneumophila intracellular survival, particularly in the presence of butyrate. 224 Cysteine supplementation restored L. pneumophila burdens in butyrate-treated BMDMs (Fig  225   3b). Imaging of GFP-expressing L. pneumophila demonstrated that the green fluorescent 226 signal increased within macrophages over time, whereas extracellular bacterial growth was 227 not observed (Fig 3c). The addition of cysteine 24 hours after butyrate treatment of infected 228 BMDMs similarly restored intracellular L. pneumophila numbers (Fig 3d). Conversely, 229 cysteine supplementation of butyrate pre-treated BMDMs enabled L. pneumophila to survive 230 in the absence of the SCFA (Fig 3d). Furthermore, intracellular L. pneumophila loads were 231 decreased in BMDMs cultured in cysteine/cystine-free media, with bacterial replication being 232 restored by cysteine supplementation (Fig 3e). Macrophage viability was not affected in 233 cysteine/cystine-free media (data not shown). In addition, supplementation of cystine-234 containing culture media with the reducing agent 2-mercaptoethanol increased intracellular L. 235 pneumophila burdens in butyrate-treated BMDMs (Fig 3f). Collectively, these findings 236 support a model in which butyrate limits L. pneumophila replication in macrophages by 237 restricting cysteine availability. 238 239 Besides Slc7a11, butyrate affected additional nutrient transporters and metabolic enzymes in 240 Legionella infected macrophages (Fig 2b). Thus, we also measured intracellular metabolites 241 using unbiased metabolite profiling which identified 28 metabolites with high confidence that 242 were significantly altered between L. pneumophila infected BMDMs treated with or without 243 butyrate (Fig 3g). Butyrate was enriched in butyrate-treated macrophages, validating the 244 approach (Fig 3g). Similarly, glutamate levels were increased in butyrate-treated 245 macrophages, consistent with the reduced expression of Slc7a11 which exports glutamate in 246 exchange for cystine. Conversely, glutamate was the most highly decreased amino acid in 247 infected BMDMs compared to uninfected cells ( Fig S3d). Glutamine was similarly 248 decreased, suggesting assimilation to glutamate ( Fig S3d). Intriguingly, cysteine-dependent 249 metabolites such as hypotaurine and taurine showed reduced levels in infected BMDMs, 250 which may further increase free cysteine levels ( Fig S3d). Conversely, butyrate treatment 251 increased hypotaurine and taurine levels in infected macrophages, thus further depleting 252 available cysteine (Fig 3g). Similarly, serine levels were increased in infected BMDMs and 253 reduced in the presence of butyrate, although this did not reach statistical significance ( longbeachae failed to increase bacterial burdens, unless cysteine was supplemented (Fig 4i). 280 These data support the notion that intracellular growth of Legionella strains and species 281 depends on sufficient access to amino acids, particularly cysteine. 282 283

Butyrate acts as a HDAC inhibitor to control intracellular Legionella replication 284
The effect of butyrate on mammalian cells can be mediated either by signalling via G protein-285 coupled receptors (GPCRs) or by inhibiting histone deacetylases (HDACs) (23, 40). We 286 initially tested whether GPCRs, which have previously been identified to sense butyrate and 287 other SCFAs (43), affect intracellular L. pneumophila numbers. For this, we generated 288 BMDMs from mice deficient in either GPR35, 41, 43, 65 or 109 and determined intracellular 289 L. pneumophila numbers in the presence of butyrate. We observed reduced bacterial burdens 290 in all GPCR-deficient BMDMs after butyrate treatment (Fig S4a), suggesting redundancy 291 between GPCR signalling or that other mechanisms are involved in reducing bacterial 292 numbers. 293 To assess whether butyrate inhibits host HDAC activity and thus prevents intracellular 294 bacterial replication, we next treated BMDMs with the pan-HDAC inhibitor trichostatin A 295 (TSA). Similar to butyrate treatment, TSA affected the expression of inflammatory mediators 296 in L. pneumophila infected BMDMs, but also nutrient transporters, including Slc7a11 (Fig  297   5a). TSA prevented intracellular L. pneumophila replication to a similar extent as butyrate 298 treatment (Fig 5b). Consistent with HDAC inhibition by butyrate in this experimental system, 299 lysine acetylation of histone H3 was markedly increased after butyrate treatment in 300 uninfected and L. pneumophila infected BMDMs, even more so than was apparent with TSA 301 treatment (Fig 5c). We next wanted to identify the specific HDAC that enables intracellular 302 L. pneumophila replication in macrophages. For this we initially used RGF966, which 303 inhibits HDAC3, as this class I HDAC has been shown to prevent inflammatory and 304 antimicrobial responses in macrophages (33,44). Consistent with the notion that Legionella 305 can evade antimicrobial responses, HDAC3 inhibition via RGF966 did not reduce 306 intracellular bacterial burdens (Fig 5d). While loss of HDAC3 reduced the expression of 307 several solute transporters in LPS-treated BMDMs, it did not prevent Slc7a11 upregulation 308 ( Fig S4b) (44). In contrast, the HDAC inhibitor tubastatin A, which targets class IIb HDAC6 309 and 10 (45), mimicked butyrate treatment in preventing intracellular L. pneumophila 310 replication (Fig 5e) and cysteine supplementation restored intracellular burdens (Fig 5f). 311 Consistent with these phenotypes, tubastatin A prevented the upregulation of Slc7a11 in L. 312 pneumophila infected BMDMs (Fig 5g). L. pneumophila was able to replicate in HDAC6-313 deficient BMDMs to the same degree as in WT macrophages (Fig 5h), suggesting a role of 314 HDAC10 or additional HDACs. LMK235, which inhibits multiple HDACs but has greatest 315 potency against class IIa HDACs, also markedly reduced intracellular L. pneumophila 316 numbers (Fig 5i). Taken together, this suggests that butyrate likely acts by inhibiting one or 317 more HDACs to attenuate Slc7a11-dependent cystine import, thus preventing intracellular L. 318 pneumophila replication in macrophages. 319 320

Butyrate controls L. pneumophila lung infections 321
We have previously shown that high-fiber diets increase serum butyrate levels, which then 322 affect immune responses in several tissues (28). Mice fed on high fiber diets, however, did 323 not display any changes in their ability to control L. pneumophila lung burdens compared to 324 no-fiber fed control animals (Fig 6a). In contrast, intranasal administration of the mixture of 325 SCFAs, acetate, propionate, and butyrate, reduced L. pneumophila lung burdens (Fig 6b). 326 Finally, butyrate alone was sufficient to reduce L. pneumophila lung burdens after intranasal 327 administration (Fig 6c) (23)  351 and this is also true in Legionella infections as shown here. This would suggest that increased 352 levels of butyrate may predispose to certain infections, an important consideration for current 353 efforts to treat unrelated diseases with SCFA supplementation. Butyrate also affects the 354 metabolism of host cells such as macrophages by promoting oxidative phosphorylation (24) 355 and reducing glycolysis (33), thus controlling additional immune responses. However, we did 356 not detect marked changes in core mitochondrial and glycolytic enzymes in butyrate treated 357 macrophages after Legionella infections. This may relate to our observation that Legionella 358 infections only marginally affected the expression of core metabolic enzymes (Fig 2b), in 359 stark contrast to other infections which upregulate the expression of metabolic genes (49, 50). 360 Legionella infections, nevertheless, triggered metabolic changes that are consistent with the 361 concept of immunometabolism, including increased levels of glycolytic metabolites, such as 362 PEP, 2P glycerate and DHAP, and mitochondrial intermediates, such as succinate and 363 itaconate. This suggests that intracellular Legionella interfere with host metabolic pathways 364 to promote their own survival. Studies on cultured L. pneumophila have highlighted unique 365 metabolic programs that drive bacteria growth whereby amino acids such as serine and 366 cysteine are used as major carbon sources, even in the presence of glucose (10). Despite this 367 notion, intracellular L. pneumophila inhibit mitochondrial function and promote glycolysis of 368 macrophages (51). The latter is dependent on the degradation of host glycogen via a L. 369 pneumophila effector that contains amylase activity, which evolved to prevent amoeba cyst 370 formation rather than to generate glucose as carbon source (8) decreased such as leucine and isoleucine which are essential for Legionella growth. Whether 382 this is due to bacterial activity of host factors remains unknown. 383 384 Slc7a11 is highly upregulated in Legionella infected macrophages likely as part of the 385 oxidative burst that occurs during phagocytosis. This is because cystine import via Slc7a11 386 enables glutathione synthesis, which is important to prevent cellular damage during oxidative 387 stress. Consistent with previous reports, we find that glutathione itself plays only a minor role 388 in Legionella infections (54,55). Treatments that prevent the expression of Slc7a11 or its 389 activity, such as butyrate and erastin, reduce intracellular Legionella burdens in macrophages, 390 unless the culture media is supplemented with cysteine. Cysteine supplementation also 391 enabled intracellular Legionella growth after 24 hours of butyrate treatment, suggesting that 392 the decreased cysteine levels in macrophages primarily affect bacterial replication not so 393 much their survival. Commonly used culture media contain cystine as it is more stable that 394 cysteine. Media that contain reducing agents to preserve cysteine and other supplements alter 395 the effects of butyrate on intracellular Legionella, highlighting that subtle changes to culture 396 conditions affect experimental outcomes. In addition, Slc7a11 expression is tightly regulated. 397 In contrast to antimicrobial responses, Slc7a11 expression is independent of HDAC3 activity 398 (44)

Quantification of Legionella burdens in vitro 506
To determine bacterial burdens, macrophages were seeded at a density of 5× 10 5 cells/mL into 507 24-well tissue culture plates and infected with Legionella strains at a MOI of 10 with or without 508 treatments. After 2 h, medium was replaced, cells washed with PBS and then further incubated 509 in media with or without treatment. At the indicated time points, cells were lysed in 0.05 % 510 digitonin for 5 min at room temperature and serial dilutions of the cell lysates and the 511 corresponding culture media were plated on BCYE agar plates, and bacterial colonies counted 512 after 48 h at 37 °C. 513 514

Live-cell imaging 515
Cells were seeded at a density of 5 × 10 5 cells/mL in 96-well tissue culture plates and cell death 516 was determined essentially as described previously (41). Cells were stained with 1 µM Cell 517 Tracker Green (CTG) (Invitrogen) for 20 min in serum-free RPMI 1640. Medium was then 518 replaced with RPMI 1640 supplemented with 15 % FBS and 15 % L-cell-conditioned medium. 519 Cells were stained with 0.6 µM Draq7 (Abcam). Cells were infected at a MOI between 10 (or 520 as indicated) in triplicate biological repeats. Alternatively, unstained BMDMs were infected 521 with GFP-expressing L. pneumophila ∆flaA (MOI = 10). Cells were imaged every 60 minutes 522 on Leica AF6000 LX or DMi8 epi-fluorescence microscope containing a heated 5% carbon 523 dioxide incubator and motorized stage with a 10× objective (NA: 0.8). Images were processed 524 and analysed using MetaMorph, Excel and GraphPad Prism. Technologies) antibodies conjugated to HRP (1:10,000 dilution in T-BST + 5 % skim milk or 538 3%BSA). Membranes were developed with the luminol-based enhanced chemiluminescence 539 (ECL) and exposed to film (Kodak). Scanned images were processed in Photoshop Adobe. 540 541 Quantitative RT-PCR 542 Total RNA was isolated using the TRIZOL Reagent (Ambion). Equal amounts of cDNA 543 were synthesized using Superscript III (Invitrogen). Quantitative PCR (qPCR) was performed 544 on Agilent AriaMx Real-TIme PCR system using the FastStart Universal SYBR Green 545 Master Rox (Roche) master mix using the following conditions: 95 o C, 10min; 95 o C, 20sec; 546 60 o C, 20sec (45x); 72 o C, 20sec; 95 o C, 15sec; 60 o C, 15sec; slow increase of temperature to 547 95 o C within 20min; 95 o C, 15sec; 12 o C. The specificity of the reaction was verified by melt 548 curve analysis. The threshold crossing value was noted for each transcript and normalized to 549 the internal control. The relative quantitation of each mRNA was performed using the 550 comparative Ct method. Experimental data processing was performed using AriaMx 551 (Agilent System). The following qPCR primers were used for this study. 552

RNA Seq 554
For RNA-seq experiment, 2 x 10 6 BMDMs/ml were uninfected or infected with 555 L. pneumophila ∆flaA at a MOI of 10 for 3 hrs and treated with or without butyrate (1mM) or 556 Trichostatin A (TSA, 10nM). Cells were washed in PBS and treated with Trizol reagent to 557 isolate total RNA. The quality of total RNA was verified with the Bioanalyzer. One 558 microgram of total RNA was used for library preparation by the PAT-seq method (60). In 559 brief, a biotinylated PAT-anchor primer cagacgtgtgctcttccgatctttttttttttttttttt compatible with 560 Illumina adaptor sequences was used to generate a 3′ tag on RNA ensuring that reverse 561 transcription is only possible from true 3′ends. A limited RNA digest was performed using 10 562 units of RNase T1 for 2 minutes, followed by chloroform/phenol extraction. Streptavidin 563 magnetic beads were used to collect the 3′ fragments and 5′ ends were phosphorylated with 564 T4 Polynucleotide Kinase. To ligate a 5′ splinted linker, the PAT-seq Splint A 5'-565 ccctacacgacgctcttccg(rA)(rC)(rT)-3' and PAT-seq Splint B 3'-566 gggatgtgctgcgagaaggctagannnn-5' were pre-annealed and then ligated to the 5′ end of the 3′ 567 fragments with T4 RNA ligase 2. Excess splint was removed by washing the magnetic beads. 568 Reverse transcription was performed with SuperScript III utilizing the PAT-seq end-extend 569 primer on the magnetic matrix. The cDNA was eluted from the beads in 2x formamide gel 570 loading buffer and size selected (~150 -400bp) on a 6% urea-PAGE. After excision from 571 the gel, the cDNA was eluted by the "crush and soak" method followed by ethanol 572 precipitation. One-third of the purified cDNA was used as input for 16 cycles of 573 amplification with PAT-seq Universal forward sequencing primer 5'-574 aatgatacggcgaccaccgagatctacactctttccctacacgacgctcttccg-3' and ScriptSeq Index PCR reverse 575 primer and AmpliTaq Gold 360 Master Mix. PAT-seq libraries were sequenced with the 576 Illumina Hiseq1500 platform using 150 base rapid chemistry according to the manufacturer's 577 instructions. To assign reads to a specific genome, reads were first clipped of poly(A) tail and 578 low-quality sequence using Tail Tools (60). Clipped reads shorter than 20 bases were 579 discarded. The analysis of the gene expressions was done using Degust software and the data 580 is available online 581 (https://degust.erc.monash.edu/degust/version/4.1/compare.html?code=01485e0b3f4264ada6f 582 c6090ecba83a4#/). 583 584

Metabolomics 585
The BMDMs were plated at cell density of 4x10 6 cells/10 cm dish overnight. Next day, 586 BMDMs were infect the L. pneumophila ∆flaA at MOI of 10 for 6hr with or without butyrate 587 (1mM). Media was removed, cells washed in ice cold (4°C) PBS and dish placed on ice for 588 the remaining of the extraction. Metabolites were extracted with 750uL of extraction solvent 589 (with or without 25 mM NEM in 80% methanol 20% 10mM ammonium formate, pH 7) at 590 4°C. Cells extracts were removed from the dish with a plastic cell scraper and thoroughly 591 vortexed for 60 min at 4°C in Eppendorf tubes. After centrifugation at 20,000 x g for 10 min 592 4°C, supernatant was transferred to new tubes and evaporated under nitrogen stream at 20-593 25 o C. Extract was stored at -80°C and prior to analysis solubilized in 120 µL of 80% 594 methanol. After centrifugation at 20,000 x g for 10 minutes at 4°C, supernatant was 595 transferred into liquid-chromatography and mass spectrometry (LC-MS) vials. Samples were 596 analysed by LC-MS as and analysed with IDEOM software as described previously (61)        Supplementary Files This is a list of supplementary les associated with this preprint. Click to download. Table1.xlsx