Salmonella manipulates the host to drive pathogenicity via induction of interleukin 1β

Acute gastrointestinal infection with intracellular pathogens like Salmonella Typhimurium triggers the release of the proinflammatory cytokine interleukin 1β (IL-1β). However, the role of IL-1β in intestinal defense against Salmonella remains unclear. Here, we show that IL-1β production is detrimental during Salmonella infection. Mice lacking IL-1β (IL-1β -/-) failed to recruit neutrophils to the gut during infection, which reduced tissue damage and prevented depletion of short-chain fatty acid (SCFA)-producing commensals. Changes in epithelial cell metabolism that typically support pathogen expansion, such as switching energy production from fatty acid oxidation to fermentation, were absent in infected IL-1β -/- mice which inhibited Salmonella expansion. Additionally, we found that IL-1β induces expression of complement anaphylatoxins and suppresses the complement-inactivator carboxypeptidase N (CPN1). Disrupting this process via IL-1β loss prevented mortality in Salmonella-infected IL-1β -/- mice. Finally, we found that IL-1β expression correlates with expression of the complement receptor in patients suffering from sepsis, but not uninfected patients and healthy individuals. Thus, Salmonella exploits IL-1β signaling to outcompete commensal microbes and establish gut colonization. Moreover, our findings identify the intersection of IL-1β signaling and the complement system as key host factors involved in controlling mortality during invasive Salmonellosis.


Introduction
The cytokine interleukin 1β (IL-1β) is a proinflammatory alarmin that is released mainly from activated myeloid cells during acute and chronic inflammation.To facilitate its rapid secretion, IL-1β is stored in the cell as a propeptide which is cleaved to its mature form by Caspase 1 following activation of the inflammasome by invasive bacteria.While inflammasome-mediated secretion of IL-1β is well characterized, IL-1β is also known to be secreted via inflammasomeindependent mechanisms [1].Once secreted, IL-1β affects multiple cell types via binding to the IL-1 receptor [2].Notably, IL-1β secretion affects endothelial cell permeability to allow massive infiltration of neutrophils from the periphery into the infected tissue [3].While these activated neutrophils contribute to bacterial killing, they also have a damaging effect on host tissues.
The foodborne pathogen Salmonella enterica serovar Typhimurium (hereafter Salmonella) is a common cause of acute gastrointestinal inflammation, caused by consumption of contaminated food.While this infection is usually self-limiting in healthy humans, it can lead to lifethreatening bacteremia in immune-compromised individuals.In mice carrying a mutated Nramp1 gene, such as all mice on a C57BL/6 background, oral infection leads to acute colitis and systemic dissemination of Salmonella, ultimately resulting in death within several days [4].The Nramp1 gene encodes an ion channel expressed in macrophages and it is thought that loss of this gene renders phagocytic cells unable to control growth of intracellular bacteria [5].Yet, the mechanism leading to mortality caused by Salmonella infection in mice is not clear.
When Salmonella is ingested via contaminated food it reaches the host intestine, where it is confronted by the gut microbiota.These microbes are well adapted to the unique environment in the gut and are thus well suited to outcompete Salmonella for the limited resources available in the gut [6].Commensal microbes also maintain a symbiotic relationship with the host, where dietary fibers from the host diet are fermented to produce short-chain fatty acids (SCFAs) by the microbiota, mainly by bacteria from the Clostridia class [7].These SCFA are used as an energy source by host colonocytes via fatty-acid oxidation, a process which utilizes oxygen in the tissue, thus leaving the gut lumen hypoxic [7].To gain a foothold in the gut, Salmonella influences the host to drive remodeling of the gut niche through inflammation [8], yet exactly how this process occurs is not clear.
Studies in mice have revealed that Salmonella can exploit the host's innate immune responses to allow its own expansion.Mice lacking toll-like receptors (TLRs) 2, 4, and 9 are more resistant to Salmonella infection than mice lacking either TLRs 2 and 4 or 4 and 9 alone, indicating that bacterial sensing by the host can be detrimental [9].In another example, mice lacking the cytokine IL-22 were found to limit Salmonella luminal growth better than wildtype mice (WT), as secretion of lipocalin-2 and calprotectin under the control of IL-22 suppressed commensal microbes which directly compete with Salmonella [10].Thus, Salmonella has evolved to exploit the host's defense mechanisms for its own advantage [11].Here, we set out to determine whether host production of IL-1β is protective during Salmonella infection and discovered that the pathogen exploits this cytokine to drive pathogenicity.

Loss of IL-1β in mice reduces Salmonella burden and prevents mortality during oral infection
Previous studies have shown that IL-1β has a protective role during chemical-induced colitis, by driving epithelial repair following injury [12].To determine whether IL-1β is required for host defense during infection-induced colitis, we challenged streptomycin-treated WT and IL-1β -/-mice (on a C57BL/6 background; Nramp1 s S1 Fig) with a single oral dose of Salmonella.Pretreatment with streptomycin is crucial to drive robust gastrointestinal inflammation [13].We found that Salmonella levels were reduced 100-fold in the cecal lumen of IL-1β -/-mice 4 days post-infection (d.p.i.) compared to WT mice (Fig 1A).IL-1β -/-mice also harbored 10-fold less Salmonella in gut-draining mesenteric lymph nodes (M.L.N.), spleen, and liver compared to WT mice (Fig 1B -1D).This reduction in bacterial loads in IL-1β -/-mice was accompanied by a marked reduction in histological colonic tissue damage (Fig 1E -1G).The composition of the gut microbiota before infection can affect the ability of Salmonella to thrive in the intestine [14].To determine whether preinfection differences in microbiota composition between WT and IL-1β -/-mice affect Salmonella colonization, we performed co-housing experiments (to normalize the microbiota composition [15]) and found that co-housing did not lead to different results than separately housed mice (S2A-S2D Fig).Thus, loss of IL-1β limits Salmonella expansion in the gut lumen and systemic sites.We then tested whether loss of IL-1β affects infection-induced mortality.Remarkably, we found that IL-1β -/-mice were completely resistant to infection-induced mortality, while WT mice all succumbed to the infection (Fig 1H).This resistance of IL-1β -/-mice was also present in co-housed WT and IL-1β -/-mice (S2E Fig) , indicating that microbiota differences before infection did not affect mortality in these experiments.To determine whether this resistance to infection is dependent on mode of infection, we infected mice intravenously with Salmonella.We found that IL-1β -/-mice infected intravenously with a low dose of Salmonella die from the infection in a similar manner as WT mice (Fig 1H).These results indicate that IL-1β plays a role in mortality during oral Salmonella infection.

IL-1β -/-mice fail to recruit neutrophil to the gut during Salmonella infection
We next wanted to determine how IL-1β production drives Salmonella expansion in the gut lumen.Mice carrying the mutant Nramp1 allele, such as C57BL/6 mice, are thought to fail in controlling Salmonella expansion because of defective macrophage function [4].Thus, we hypothesized that IL-1β impairs intracellular killing of Salmonella by macrophages.To test this, we extracted peritoneal macrophages from WT and IL-1β -/-mice and tested their ability to kill intracellular Salmonella in vitro.Contrary to our hypothesis, we found that macrophages from IL-1β -/-mice were less efficient at killing intracellular Salmonella (S3A Fig) .This result indicates that loss of IL-1β impairs Salmonella growth in a manner which is not dependent on intracellular killing by macrophages.
To understand how IL-1β deficiency affects the colonic tissue during Salmonella infection, we performed bulk RNA sequencing followed by pathway analysis of differentially expressed genes.We found that genes which were down-regulated in IL-1β -/-mice are involved in neutrophil function and recruitment to the colon (Fig 2A).Indeed, key chemokines that attract neutrophil and monocytes to the colon, along with neutrophil-specific transcripts such as lipocalin-2 and calprotectin, were expressed at lower levels in infected IL-1β -/-mice (Fig 2B).These differences were absent in uninfected WT and IL-1β -/-mice (S4 Fig) .We validated these findings by counting myeloperoxidase (MPO)-and neutrophil elastase (NE)-positive cells in colonic tissue sections from infected mice.Accordingly, IL-1β -/-mice had on average 4-fold fewer neutrophils in their colons (Fig 2C -2E).We then tested whether loss of IL-1β leads to reduced neutrophils levels during Salmonella infection because of neutrophil development defects, or failure to recruit neutrophils to the colonic tissue.By analyzing circulating blood from naïve and infected mice, we found that loss of IL-1β does not affect circulating neutrophil levels during steady state, but rather prevents neutrophil recruitment to circulation during infection (Fig 2F and 2G).Thus, the resistance of IL-1β -/-mice to Salmonella is not due to an elevated antimicrobial response as they fail to draw neutrophils to the site of infection.

IL-1β leads to collapse of gut short-chain fatty acid-producing Clostridia during Salmonella infection
Commensal gut microbes provide colonization resistance by directly inhibiting pathogen growth and by affecting the host [16].We next tested whether changes in the gut microbiota of IL-1β -/-mice can explain their resistance to Salmonella infection.We reasoned that lack of neutrophil recruitment to the gut would dampen the harmful damage of infection-induced inflammation on the microbiota [16].16S rRNA microbiota analysis of feces from infected WT and IL-1β -/-mice revealed that the latter contained a more diverse microbiota after Salmonella infection (Fig 3A).Jaccard analysis confirmed that the gut microbial composition of infected IL-1β -/-mice was distinct from that of WT mice (Fig 3B).Analysis of composition of microbiomes (ANCOM) showed that bacteria from the Clostridia class were enriched in the gut of infected IL-1β -/-mice while these microbes were virtually absent from the gut of infected WT mice (Figs 3C and S5A).Previous reports have shown that Salmonella infection leads to depletion of SCFA-producing Clostridia microbes [7].We found that in naïve WT and IL-1β -/-mice, the levels of Clostridia bacteria were similar.However, Salmonella infection dramatically reduced the levels of Clostridia bacteria in infected WT mice while not affecting these microbes in IL-1β -/-mice (Fig 3D).This post-infection reduction in levels of Clostridia were not dependent on preinfection conditions as co-housing and streptomycin treatment of WT and IL-1β -/-mice did not result in differences in Clostridia levels (S5B Fig).Thus, Salmonella infection does not lead to depletion of SCFA-producing Clostridia bacteria in IL-1β -/-mice.

Preservation of SCFA-producing bacteria in IL-1β -/-mice inhibits Salmonella expansion by maintaining beta oxidation in colonocytes and hypoxic condition in the gut
Next, we wanted to determine how this different microbiota in infected IL-1β -/-mice affects the colonic tissue.We found that genes which were up-regulated in the colon of infected IL-1β -/-mice were enriched in metabolism pathways (Fig 4A).Specifically, we found that genes encoding proteins that participate in energy production via fatty acid beta oxidation were upregulated in infected IL-1β -/-mice as compared to infected WT mice (Fig 4B).These differences were not present in uninfected WT and IL-1β -/-mice (S6A and S6B Fig).
SCFAs produced by commensal Clostridia bacteria serve as the preferred energy source for colonocytes.These epithelial cells produce ATP via beta oxidation of SCFA, which renders the colonic lumen hypoxic as this process utilizes oxygen.Salmonella infection has been shown to deplete SCFA-producing bacteria from the colon, thus forcing colonocytes to produce energy via glycolysis [7].As this process does not require oxygen, Salmonella infection effectively leads to elevated oxygen levels in the colonic lumen.Oxygen in the lumen is toxic to the obligatory anaerobic commensal microbiota, which is beneficial for Salmonella as it disrupts the colonization resistance provided by these commensals [16].Indeed, we found that all the genes in the glycolysis pathway were elevated in infected WT compared to IL-1β -/-mice (Fig 4C).These differences were not present in uninfected WT and IL-1β -/-mice (S6C-S6E Fig) .Accordingly, we found that the colonocytes in IL-1β -/-mice remained hypoxic during Salmonella infection while colons of WT mice became rich in oxygen (Fig 4D and 4E).
These observations led us to hypothesize that failure to recruit neutrophils in IL-1β -/-mice preserves hypoxic conditions in the colon by sparing SCFA-producing Clostridia bacteria.This in turn can inhibit Salmonella growth by providing direct competition and by reducing oxygen levels that are important for Salmonella growth [16].To test our hypothesis, we depleted Clostridia bacteria from the colon of WT and IL-1β -/-mice by vancomycin treatment (in addition to the standard streptomycin treatment) and then infected these mice with Salmonella.We validated that vancomycin treatment indeed depleted the major SCFA-producing members of the Clostridia class [17] via 16S sequencing (S7 Fig).Indeed, we found that Salmonella growth in the intestine of IL-1β -/-mice was indistinguishable from WT mice after vancomycin treatment (Fig 4F).Thus, preservation of SCFA-producing bacteria in IL-1β -/-mice maintains energy production via beta oxidation and hypoxic condition in the gut which in turn inhibit Salmonella growth.

IL-1β drives mortality in Salmonella-infected mice by suppressing the anaphylatoxin-inactivator carboxypeptidase N
Next, we wanted to determine whether preserving SCFA-producing Clostridia is the mechanism that prevents mortality in Salmonella-infected IL-1β -/-mice.However, we found that vancomycin-treated IL-1β -/-mice still survived after oral Salmonella infection (S8A Fig) .We then hypothesized that the reason for resistance to Salmonella infection in IL-1β -/-mice is that they can clear the infection.Yet, we found that even 17 d.p.i.these IL-1β -/-mice were still colonized by Salmonella (S8B Fig), and that the pathogen was still fully virulent as transmission to WT hosts was still lethal (S8C Fig).
To understand why IL-1β -/-mice do not die from Salmonella infection, we performed RNA sequencing on mice 6 d.p.i., as this is the time point when WT mice are moribund.We reasoned that genes that are differently expressed in WT mice 6 d.p.i.compared to WT mice 2 d.p.i., and to IL-1β -/-mice 6 d.p.i., are related to the process that leads to mortality.We excluded genes that are differently expressed in IL-1β -/-mice 6 d.p.i.versus 2 d.p.i. as these mice do not die from the infection, thus these genes are not related with mortality (Fig 5A).Pathway analysis revealed that WT mice express genes that are related to vasculature morphology 6 d.p.i.(Fig 5A).We found this interesting, as septic shock which occurs during systemic infection by pathogens is characterized by vasculature permeability, dilation, and a severe hypotensive response [18,19].These harmful changes to the vasculature are mediates by complement proteins, also known as anaphylatoxins, which can affect vasculature permeability at sub-nanomole concentrations [18].We found that transcription levels of the complement C3 gene and the complement receptor C3ar1 gene were highly induced in WT mice at 6 d.p.i.(Fig 5B and 5C).While these genes were also induced in IL-1β -/-mice, their levels at 2 and 6 d.p.i. were still significantly lower than in WT mice (Fig 5B and 5C).Thus, IL-1β drives expression of anaphylatoxin proteins and receptors during Salmonella infection.
Under normal conditions, the damaging effects of anaphylatoxins are balanced by carboxypeptidase N (CPN1), which inactivates anaphylatoxins by peptide cleavage within seconds at physiological concentrations [20,21].We found that IL-1β -/-mice express 4-fold higher transcription levels of Cpn1 after Salmonella infection, compared with WT mice (Fig 5D).Indeed, a previous report has shown that IL-1 receptor signaling inhibits Cpn1 expression [22].These results imply that low expression and inactivation of the complement system in IL-1β -/-mice preserves their viability during infection.To test this, we treated mice with a CPN1 inhibitor.We found that inhibiting CPN1 in uninfected mice does not affect their viability (S9A Fig) .In infected WT mice, administration of CPN1 inhibitor did not affect survival (Fig 5E).However, infected IL-1β -/-mice treated with CPN1 inhibitor suffered from over 60% mortality, while vehicle-treated IL-1β -/-mice were completely resistant to infection-induced mortality (Fig 5E).We also verified that the inhibitor does not affect Salmonella growth (S9B Fig).Thus, loss of IL-1β protects mice from Salmonella-induced mortality by removing suppression from CPN1 expression.

Expression of complement 3 receptor correlates with IL-1β expression in sepsis patients
Finally, we wanted to determine whether the link between the complement system, IL-1β and life-threatening infection, which we show here in mice, is also relevant for human disease.We first analyzed a publicly available whole-blood transcriptional dataset from healthy individuals that underwent experimental endotoxemia via intravenous injection of lipopolysaccharide (LPS) [23].We found that LPS administration led to an increase in expression levels of IL-1β, C3AR1, and the IL-1β receptor IL1R1 transcripts (Fig 6A -6C).Additionally, we found that expression of IL-1β positively correlated with expression of C3AR1 in individuals treated with LPS, but not untreated controls (Fig 6D and 6E).Thus, endotoxemia leads to positive correlation between expression of IL-1β and complement receptor gene C3AR1 in humans.
Next, we analyzed peripheral blood RNA sequencing data from a cohort of sepsis patients [24].We found higher levels of IL-1β, C3AR1, and IL1R1 transcripts in patients with sepsis (Fig 6F -6H) and a positive correlation between expression levels of IL-1β and C3AR1 in sepsis patients but not healthy controls (Fig 6I and 6J).Finally, we analyzed whole-blood transcriptional data from a cohort that included patients admitted to the intensive care unit (ICU) with sepsis, patients admitted to the ICU with non-infectious conditions, and healthy controls [25].We found positive correlation between expression of IL-1β and C3AR1 only in patients suffering from sepsis (Fig 6K -6M).Together, these results indicate that expression of IL-1β, the IL-1β receptor, and the complement 3 receptor transcripts are induced in humans suffering from sepsis and that positive correlation between expression of IL-β and C3AR1 occurs only in sepsis patients.

Discussion
Salmonella has evolved to be a master manipulator of host response.By driving acute inflammation in the gut, Salmonella facilitates the killing of its microbial competitors via the host.Indeed, recent research has shown that, under experimental settings, an immune-compromised host (through loss of TLR signaling, IL-22 production, or STAT2 signaling) is better equipped to contain Salmonella infection [9,10,26].Here, we show that IL-1β is a central component of host manipulation by Salmonella.Using IL-1β signaling to facilitate its colonization is an effective mechanism deployed by Salmonella, as multiple microbial detection apparatuses converge on processing and release of IL-1β [1,27].Indeed, secretion of IL-1β can be performed even in the absence of inflammasome activation [1] and without resulting in pyroptosis [28].This capability of different pathways to drive IL-1β secretion can also explain why mice deficient in individual components of the inflammasome have not shown the same resistance to Salmonella infection as we show here in IL-1β -/-mice [29].Along these lines, a recent report has shown that only combined ablation of all the pathways which drive NLRC4 inflammasome-dependent cell death (deletion of Asc, Caspase 1 and Caspase 11, simultaneously) can protect mice from fatal Salmonella infection [30].These studies, along with the data presented here, indicate that deletion or inhibition of specific inflammasomes or Caspases does not result in complete loss of IL-1β production.Thus, it is possible that Salmonella has evolved to depend on IL-1β for host colonization, as infection consistently results in IL-1β secretion, even if not all pathogen-detection mechanisms are triggered.
Multiple studies have focused on the role of inflammasomes in host response to Salmonella infection [29].Most of these studies have focused on the ability of inflammasome activation to limit bacterial growth in inner tissues, such as liver and spleen.Thus, we chose to focus our study on the role that IL-1β plays in remodeling of the intestinal niche, rather than the mode of its production and release.Our results highlight the exploitation of IL-1β by Salmonella to establish a foothold in the luminal niche.We show that IL-1β production in response to infection activates a cascade of events that results in damage to the structure of the commensal microbiota and release of oxygen into the colonic lumen (Fig 7).This ultimately allows Salmonella to outcompete resident gut microbes by using the host immune response, rather than direct competition for resources in the gut.
Many studies in the fields of immunology and pathogenesis rely on mouse mortality as a measurable readout for virulence.In the current study, we found that IL-1β -/-mice do not succumb to Salmonella infection.To help us identify the mechanism leading to this resistance, we conducted a literature search to understand why this infection is lethal in C57BL/6 WT mice.Yet, we were surprised that our literature search on the mechanisms that drive mortality during non-typhoidal Salmonellosis did not result in definitive explanations.Death from nontyphoidal Salmonellosis in mice carrying the susceptibility allele in Nramp1 is attributed to an inability to control intracellular bacteria by macrophages [5].Yet, we could not find definitive evidence to support this notion.Another study has found that mortality from intravenous infection is dependent on bacterial lipid A and is characterized by robust IL-1β secretion [31].Yet, the downstream target of IL-1β that leads to mortality was not known.Our discovery that IL-1β dramatically up-regulates the expression of complement proteins and down-regulates the expression complement-inactivator in vivo provides a possible mechanism for Salmonelladriven mortality (Fig 7).This mode of regulation can thus maximize the antimicrobial activity of the complement system as it drives complement expression and inhibits its degradation simultaneously.
In humans, a global systematic review and meta-analysis found that the most common reason for non-typhoidal Salmonella disease-associated complications and fatalities was septicaemia [32].The definition of septicaemia is "an infection that occurs when bacteria enter the bloodstream and spread," according to the Cleveland Clinic [33].As in mice, this explanation does not provide a testable mechanism.Studies in nonhuman primates have shown that inhibition of complement proteins can prevent mortality from intravenous bacterial infection [34].Complement proteins are known to elevate vascular permeability, cause smooth muscle contraction and drive cardiomyopathy, all of which lead to fatal outcome [35].Thus, control of complement expression needs to be linked to acute immune activation, while also being tightly regulated to prevent excessive damage.Our finding that IL-1β drives expression of complement anaphylatoxins, and suppresses expression of complement-inactivating CPN1, provides an example where immune regulation of the complement system can spiral out of control, leading to death ( Fig 7).This also provides a possible mechanism for non-typhoidal Salmonella-induced mortality.While we show a correlation between expression of IL-1β and the complement receptor C3AR1 in humans suffering from sepsis, it will need to be tested whether this is the mechanism which results in fatalities from septicaemia in humans.
Finally, our study presents results in Salmonella-infected IL-1β -/-mice which differ from a previous report by Raupach and colleagues which also used Salmonella-infected IL-1β -/-mice [36].The work by Raupach and colleagues did not examine luminal bacterial loads in the intestine but found higher bacterial loads in inner tissues of IL-1β -/-mice as compared to WT mice.Also, Raupach and colleagues did not find that IL-1β -/-mice exhibit reduced Salmonella-induced mortality.One reason for this discrepancy might be that here we used the streptomycin-depletion model [13], which reduces the influence of the preinfection microbiota (which can vary between animal facilities), while Raupach and colleagues did not pretreat the mice with antibiotics.Indeed, variations in the preinfection microbiota can widely influence infection outcome and mouse survival, especially when antibiotic pretreatment is not performed [14].Another possible reason for this discrepancy is that Raupach and colleagues used a 10-fold higher inoculation dose compared to the dose we used here.At such high numbers (10 8 C.F.U.), the host might be overwhelmed to a point where genetic differences are less influential.Indeed, we also show here that loss of IL-1β does not provide protection from mortality during systemic infection.Finally, the genetic background of the mice might also explain this discrepancy, as is the case in Caspase 1-deficient mice [36][37][38].

Ethics statement
All experiments and procedures involving animals were approved by the institutional animal care and use committee (IACUC) of the Bar-Ilan University (study ID#14-02-2020).This study was performed according to national laws and regulations, and with compliance to EU directive regarding the protection of animals used for experimental and other scientific purposes.The Bar-Ilan IACUC has approved the animal experiments only after ethical balancing of the benefits of the project against the suffering caused to the animals.The approval is firmly based on the principle of the 3 Rs, to replace, reduce, and refine the use of animals used for scientific purposes.

Mice
C57BL/6 WT mice and IL-1β -/-mice (backcrossed to C57BL/6 for over 10 generations [12]) were separately bred and maintained in the conventional barrier at the Azrieli Faculty of Medicine, Bar-Ilan University, Israel; 8-to 14-week-old mice were used for all experiments.For cohousing experiments, mice were co-housed for 2 weeks prior to infection, which is a timeframe shown to be sufficient to eliminate microbiota differences [15].

Salmonella infection
Mice were treated with 20 mg streptomycin via gavage 24 h before oral infection with 10 7 C.F. U. Salmonella enterica serovar typhimurium (SL1334) or intravenous infection with 10 5 C.F. U. Salmonella.For vancomycin treatment model, mice were treated with 300 mg of streptomycin and 150 mg of vancomycin in drinking water in a volume of 300 ml throughout the course of the infection.At the designated time points, mice were euthanized and cecal content, liver, spleen, and mesenteric lymph nodes were removed, weighed, homogenized in PBS, and plated on LB agar plates containing streptomycin following serial dilutions.For survival assays, mice that were unresponsive to touch, cold or could not reach their food were euthanized.

Microbiota profiling and analysis
Bacterial DNA was extracted from feces, using the Mobio PowerSoil DNA extraction kit (MoBio) following a 2-min bead-beating step (Biospec).The V4 of the 16S rRNA gene was amplified using PCR with barcoded primers.DNA was then purified using AMPURE XP magnetic beads (Beckman Coulter) and quantified using Quant-iT PicoGreen dsDNA Assay (Thermo Fisher) and equal amounts of DNA were then pooled and sequenced.After sequencing on an Illumina MiSeq platform at the Faculty of Medicine Genomic Center (Bar Ilan University, Safed, Israel), single-end sequences reads were import and demultiplexed using QIIME 2 (version 2023.2) [42].Sequencing errors were corrected by DADA2 [43] and taxonomic classification was done using Greengenes reference database with confidence threshold of 99% [44].Principal coordinate analysis (PCoA) was performed using Jaccard distances, which calculate the difference between the presence or absence of features [45].ANCOM was used to identify differentially abundant taxa [46].Raw data is deposited in GSE252071.

Detection of hypoxia in vivo
Mice were treated intraperitoneally with a 1.2 mg solution of pimonidazole HCl (Hypoxyprobe Kit) 1 h before euthanasia.Distal colon tissues were fixed in 4% paraformaldehyde, paraffin embedded, sectioned, and stained with of 11.23.22.rRat Mab for 1 h.Sections were visualized with a Zeiss Axio imager M2 according to manufacturer's instructions.Scoring was performed in a blinded fashion.

Carboxypeptidase-inhibitor treatment
Mice were infected orally with Salmonella as above.After infection, mice were treated with 1.25 mg of DL-Benzylsuccinic acid (in 10% ethanol) via intraperitoneal injection every 8 h.Control mice were treated with 10% ethanol in PBS.Mice were physically examined before and after injections to exclude damage due to the procedure.Mice hurt during the injection were excluded from the experiment.

Statistical analysis
Results were analyzed using Graph-Pad Prism 10 software (GraphPad, La Jolla, California, United States of America).

Fig 2 .
Fig 2. Loss of IL-1β impairs neutrophil recruitment to the gut during Salmonella infection.(A) Pathway analysis of transcripts that are down-regulated in colonic tissue of Salmonella-infected IL-1β -/-according to GO biological function.Bars represent -log (P value) and dots represent number of genes in pathway.(B) Heatmap depicting differentially expressed innate immune genes with a P < 0.05.Each column represents a mouse and each row a gene.(C, D) Numbers of NE-positive (C) and myeloperoxidase-positive (D) cells in colonic section of Salmonella-infected mice 4 d.p.i.Each dot represents a mouse.(E) Representative immunohistochemistry images of colonic section from Salmonella-infected mice stained with anti-neutrophil elastase antibody.Scale bar, 50 μm.(F) % of neutrophils out of total circulating WBC and (G) concentration of neutrophils in serum of WT and IL-1β -/-mice infected as indicated 4 d.p.i.*P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001.(C and D) Student's t test.(F and G) One-way ANOVA.d. p.i., days post-infection; S. Tm, Salmonella typhimurium.These data are representative of 2 independent experiments.Numerical values are in S1 Data.The underlying data for this figure can be found at GSE252071.NE, neutrophil elastase; WT, wild-type.https://doi.org/10.1371/journal.pbio.3002486.g002

Fig 3 .
Fig 3. Salmonella infection does not deplete SCFA-producing Clostridia from the gut of IL-1β -/-mice.16 S rRNA sequencing was performed to characterize gut microbiota composition.(A) Shannon index representing microbial diversity in gut microbiota of Salmonella-infected mice.(B) PcoA of fecal microbiota based on Jaccard similarity coefficient.Each dot represents a mouse.(C) Relative taxonomic composition at the class levels.Each column represents a mouse.(D) qPCR analysis of levels of the class Clostridia in naïve or infected mice as indicated.Each dot represents a mouse.*P < 0.05; **P < 0.01; (A) Student's t test, (D) one-way ANOVA.d.p.i., days post-infection; S. Tm, Salmonella typhimurium.These data are representative of at least 2 independent experiments.Numerical values are in S1 Data.The underlying data for this figure can be found at GSE252071.PCoA, principal coordinate analysis; SCFA, short-chain fatty acid.https://doi.org/10.1371/journal.pbio.3002486.g003

Fig 4 .
Fig 4. Preservation of fatty acid beta-oxidation in IL-1β -/-mice inhibits Salmonella growth in vivo.(A) Pathway analysis of transcripts that are up-regulated in colonic tissue of Salmonella-infected IL-1β -/-according to GO biological function.Bars represent -log (P value) and dots represent number of genes in pathway.(B) Heatmap depicting differentially expressed genes involved in fatty acid oxidation with a P < 0.05.Each column represents a mouse and each row a gene.(C) Heatmap depicting differentially expressed genes in the glycolysis pathway with a P < 0.05.The enzymatic activity of each gene in the glycolysis pathway is presented on the left.Each column represents a mouse and each row a gene.(D) Immunofluorescence microscopy of colonic section from mice.Red staining shows pimonidazole that indicates hypoxia levels.Nuclei were stained with DAPI.Scale bar, 50 μm.(E) Quantification of red signal in (D).Each dot represents a mouse.(F) Salmonella C.F.U. in cecal contents of infected mice 4 d.p.i.treated as indicated.Each dot represents a mouse.**P < 0.01, ****P < 0.0001; (E) one-way ANOVA; (F) Mann-Whitney test.d. p.i., days post-infection; S. Tm, Salmonella typhimurium.These data are representative of 2 independent experiments.Numerical values are in S1 Data.The underlying data for this figure can be found at GSE252071.https://doi.org/10.1371/journal.pbio.3002486.g004

Fig 5 .
Fig 5. IL-1β promotes expression of anaphylatoxin and inhibits expression of carboxypeptidase N that drives mortality in Salmonella-infected mice.(A) Pathway analysis of transcripts that are differently expressed as indicated in Venn diagram in colonic tissue of Salmonella-infected mice according to GO biological function.Bars represent fold enrichment and dots represent number of genes in pathway.(B-D) Normalized reads of the indicated genes from colons of mice based on RNA sequencing.Each dot represents a mouse.(E) Survival of Salmonella-infected mice treated with vehicle or CPN1 inhibitor.*P < 0.05, ****P < 0.0001.(B-D) Student's t test.(E) Mantel-Cox test.d.p.i., days post-infection; CPN, carboxypeptidase N.These data are representative of 2 independent experiments.Numerical values are in S1 Data.The underlying data for this figure can be found at GSE252071.https://doi.org/10.1371/journal.pbio.3002486.g005

Fig 7 .
Fig 7. IL-1β is a key axis in host manipulation by Salmonella.(I) In naïve mice, commensal microbes produce SCFA.(II) These SCFA are used to produce energy via β-oxidation by colonocytes.As this process uses oxygen, the gut lumen remains hypoxic.(1) Salmonella infection in WT mice leads to IL-1β secretion and (2) production of chemokines that attract neutrophil to the gut lumen.(3) These neutrophil produce ROS and NOS and secrete metal chelators which (4) kills commensal microbes, thus reducing SCFA levels.This forces colonocytes to produce energy via fermentation which does not uses oxygen (5), thus allowing oxygen to seep into the gut lumen, further damaging the microbiota.(6) Release of IL-1β following infection leads to elevated levels of anaphylatoxins and suppressed levels of CPN1, leading to mortality.(A) In mice lacking IL-1β, Salmonella infection (B) does not lead to production of chemokines and (C) does not attract neutrophils to the gut lumen.(D) These preserve SCFA-producing commensals and (E) allow colonocytes to produce energy via β-oxidation, which preserves luminal hypoxia which impedes Salmonella expansion.(F) Lack of IL-1β reduces anaphylatoxin levels and maintains high levels of CPN1 that prevents mortality.CPN1, carboxypeptidase N; SCFA, short-chain fatty acid.https://doi.org/10.1371/journal.pbio.3002486.g007