Disease tolerance P. aeruginosa strains couple EPS synthesis with impaired succinate bioenergetics

To understand the basis of tolerance to P. aeruginosa lung infection, we examined the immunometabolic features of strains associated with this state. When transitioning to a tolerant program, hosts favor P. aeruginosa strains that are metabolically primed to manufacture the EPS alginate, driven by algD expression (Supplementary Fig. 1a). These alginate-producing communities help maintain respiratory integrity by suppressing IL-1β, while sparing cytokines like IL-6 or tumor necrosis factor α (TNFα), as shown by the laboratory strain of P. aeruginosa PAO1 (WT PAO1), its ΔalgD PAO1 mutant (Fig. 1a, Supplementary Fig. 1b, c), and alginate-programmed isolates from tolerant individuals (Supplementary Fig. 1d). These alginate-specialized strains do not produce type III secretion system toxins like pcrV, exoS and exoT, which trigger inflammasome activation and alveolar destruction^24,25,26 (Supplementary Fig. 1e). This milieu limits alveolar permeabilization (Fig. 1b, Supplementary Fig. 1f) and host mortality (Fig. 1c, Supplementary Fig. 1g), hallmarks of disease tolerance^11,12,20. The tolerogenic program enabled by alginate-producing pathogens does not rely on other major immunomodulatory mechanisms, such as recruitment of regulatory T cells (Treg)²⁷ or enrichment of the anti-inflammatory cytokine IL-10²⁸ (Supplementary Fig. 1h, i).

Mice were exposed to either PBS, WT PAO1, or ΔalgD PAO1 (n = 3, total of 9–10 mice per group). Measures outcomes included: a BAL cytokines; b BAL albumin; c host survival. d Bacterial energy production (Biolog Technology) (n = 3). e, f Bacterial oxygen consumption rate (Seahorse) (e) and total oxygen consumption (AUC: arbitrary unit count) (f) (n = 5). g Caloric expenditure (μW/OD₆₀₀). Right graph: total heat along growth (n = 4). h Growth curves (OD₆₀₀) (n = 3). i Biofilm quantification (n = 12). j BAL succinate (n = 3, total of 6–10 mice per group); k bacterial burden (n = 3, total of 12 mice per group); l BAL itaconate (n = 3, total of 6–9 mice per group). WT and Irg1^−/− mice were exposed to either PBS or WT PAO1 (n = 3, total of 8–10 mice per group). The following were measured: m BAL cytokines; n bacterial burden. Data are shown as average +/− SEM. d, f–g, i: t-Student test. a, b, j–n One-Way ANOVA (Tukey multiple comparison test). e, g, h Two-Ways ANOVA; C: Kaplan-Meier test. All statistical tests are two-sided. Source data are provided as a Source Data file.

Alginate-specialized P. aeruginosa strains also exhibit distinct bioenergetic profiles. Producing alginate through algD constrains TCA cycle activity, limiting succinate and citrate bioenergetics (Fig. 1d). This correlates with lower aerobic metabolic rates (Fig. 1e, f). Alginate synthesis demands substantial calories at the initial phase of proliferation, as illustrated by integrating isothermal calorimetric curves with growth assays²⁹ (μM/OD₆₀₀ – energy utilization per biomass unit) (Fig. 1g, gray and red zones). This energy expenditure limits growth, as shown by lower WT PAO1 biomass versus its ΔalgD PAO1 mutant in stationary phase (Fig. 1h, blue zone), similar to alginate-programmed isolates from tolerant hosts (Supplementary Fig. 1j–l). To balance this energy trade-off, P. aeruginosa prioritizes biofilm formation (Fig. 1I), a hallmark of limited bioenergetic resources³⁰. In the respiratory tract, these alginate-rich biofilms reduce succinate levels (Fig. 1j) and pathogen burden (Fig. 1k, Supplementary Fig. 1m, n), leading to decreased IL-1β signaling (Fig. 1a, Supplementary Fig. 1d). However, pre-nourishing these strains with succinate increases biomass and IL-1β but not IL-6 or TNFα (Supplementary Fig. 2). Overall, these findings illustrate how alginate-rich P. aeruginosa strains adapt bioenergetically to support lung health through disease tolerance by prioritizing alginate manufacture over succinate metabolism.

Alginate enhances host itaconate signaling during P. aeruginosa pneumonia

Alginate-rich P. aeruginosa strains promote airway itaconate enrichment (Fig. 1l). In vitro assays using bone-marrow derived macrophages (BMDM) showed that alginate specifically induced itaconate signaling (Supplementary Fig. 3a, b). Soluble alginate dampens BMDMs inflammatory responses to LPS, particularly reducing the release of IL-1β and TNFα (Supplementary Fig. 3c, d), but not IL-6 (Supplementary Fig. 3e). Using Irg1^−/− animals, which cannot produce itaconate¹⁴ (Supplementary Fig. 3f), we demonstrated that the immunometabolite during P. aeruginosa pneumonia restricted IL-1β signaling without affecting IL-6 and TNFα (Fig. 1m). This IL-1β inhibition linked with decreased NLPR3 inflammasome activity, particularly expression of both pro-Caspase 1 and its cleaved p20 form¹⁵ (Supplementary Fig. 3g). Itaconate promoted downstream cytoprotective routes that abrogate IL-1β production and favor mucosal integrity¹⁷, like Hemoxygenase 1 (HO-1) (Supplementary Fig. 3h). This environment reduced P. aeruginosa burden in the respiratory tract (Fig. 1n). These results show that P. aeruginosa alginate manufacture aligns with increased pulmonary itaconate levels, creating an immunometabolic state that promotes disease tolerance.

Itaconate blocks P. aeruginosa succinate bioenergetics and drives alginate-mediated biofilms

Itaconate promoted P. aeruginosa alginate-based biofilms, especially when combined with carbohydrates like glucose (Fig. 2a). Adding succinate, the preferred P. aeruginosa nutrient^31,32,33, dispersed these biofilms (Fig. 2a), showing that itaconate enables this lifestyle by altering succinate bioenergetics in the TCA cycle.

a Biofilm specialization by WT and ΔalgD PAO1 (OD₅₄₀/OD₆₀₀) (n = 3, 3-5 replicates per assay). WT PAO1 was exposed or not to itaconate in nutrient-rich media (LB). The following were measured: b intracellular TCA cycle metabolite abundance (n = 3); c Global chemoproteomic profiling of S-itaconation of the WT PAO1 proteome (n = 3). d Succinate oxidation (generation of anion superoxide (O₂^*−)) in WT PAO1 and Δict PAO1, which cannot degrade itaconate (n = 4). e oxygen consumption rates (OCR) by Seahorse technology (n = 4); f total oxygen consumed along time (n = 4); g growth (OD₆₀₀) (n = 3). Data are shown as average +/− SEM. a: One-Way ANOVA (Tukey multiple comparison test). g Two-Way ANOVA. b–d, f t-Student test. All statistical tests are two-sided. Source data are provided as a Source Data file.

In PAO1, itaconate provoked a ~ 15,000-fold accumulation of the succinate precursor succinyl-CoA (Fig. 2b). This succinyl-CoA enrichment was accompanied by accrual of all its predecessors, such as αketoglutarate (αKG), aconitate, and citrate (Fig. 2b), but not the succinyl-CoA byproducts fumarate, malate, and oxaloacetate (Fig. 2b). This indicates a bottleneck in the TCA cycle, impairing succinyl-CoA conversion to succinate.

In P. aeruginosa, the TCA cycle enzyme that converts succinate into succinyl-coA is succinyl-CoA ligase (sucCD). Chemoproteomic profiling using the bioorthogonal probe of itaconate, C3A, which retained the α,β-unsaturated carboxylic acid groups of the immunometabolite, revealed that itaconate modifies key cysteine residues of succinyl-CoA ligase through “S-itaconation”^18,34, inactivating its function. Among all proteins exhibiting C3A-modified cysteines, we found sucC and sucD, subunits β and α of succinyl-CoA ligase, respectively, clustering among the most significantly S-itaconated proteins (Fig. 2c). Modifications in sucC mapped to Cys101 (Log₂FC ~ 1.08) and Cys211 (Log₂FC ~ 3.58), and sucD PTM to Cys13 (Log₂FC ~ 2.45) and Cys124 (Log₂FC ~ 3.79) (Fig. 2c). Of note, in Gram-negative organisms, sucD Cys124 is highly conserved (Supplementary Fig. 4a), as it coordinates the removal of the CoA group from succinyl-CoA to form succinate³⁵. Transcriptomic data confirmed sucD inhibition by itaconate (Supplementary Fig. 4b).

Itaconate also modified sdhB at Cys209 and Cys213 (Log₂FC ~ 1.56), a succinate dehydrogenase subunit, the enzyme that oxidizes succinate in the P. aeruginosa TCA cycle-ETC interface (Fig. 2c). Functional assays using a P. aeruginosa strain unable to breakdown itaconate (Δict PAO1) indicated that this metabolite limited the ability of the organism to oxidize succinate and produce anion superoxide (O₂^*−) at the ETC (Fig. 2d). This bioenergetic impairment provoked compensatory increases in oxygen consumption to maintain ATP homeostasis (Fig. 2e, f). Consistently, the growth of Δict PAO1 in itaconate was rescued when succinate was added in increasing concentrations (Fig. 2g). Glucose failed to restore Δict PAO1 growth, showing itaconate’s effects are succinate-specific (Supplementary Fig. 4c–e). Thus, itaconate blocks P. aeruginosa succinate bioenergetics.

Itaconate promotes P. aeruginosa EPS reprogramming

We found itaconate supports alginate-based P. aeruginosa biofilms via bacterial metabolic remodeling. The specialization of P. aeruginosa into less immunostimulatory communities – alginate-rich biofilms – requires the concerted participation of many anabolic platforms, such as gluconeogenesis, the pentose phosphate pathway (PPP), and the de novo pathway of pyrimidine synthesis (DNPPS)^36,37,38,39 (Fig. 3a). These circuits generate different EPS precursors, like carbohydrates, cycled sugars, and nucleotides^36,37,38,39, which further stimulate algD function for alginate synthesis. We found that itaconate stimulated PAO1 enrichment of many gluconeogenic intermediates that facilitate alginate generation, like phosphoenolpyruvate (PEP), fructose-6-phosphate, glucose-6-phosphate and glucose (Fig. 3b). Furthermore, itaconate provoked PAO1 concentration of diverse sugars that emerge from gluconeogenesis, such as glucuronate, glucosamine (Glc), and N-Acetyl-Glc-1-phosphate (GlcNAc-1-P) (Fig. 3c). Itaconate also augmented in P. aeruginosa different cycled carbohydrates formed in the PPP, including D-Ribose, D-Sedoheptulose-7-phosphate, and D-Erythrose-4-phosphate (Fig. 3d). This metabolic reconfiguration was consistent with the accumulation of both orotate, the rate-limiting step of the DNPPS, and uracil, which is in direct equilibrium with orotate via uridine monophosphate (UMP) (Fig. 3e). The priming by itaconate of all P. aeruginosa platforms involved in EPS was confirmed by the increase of both UDP-glucuronate and GDP-glucose, nucleotide-activated carbohydrates that act as building blocks for many of these biopolymers (Fig. 3f). Of note, other EPS precursors like ADP-glucose and UDP-glucose were depleted by itaconate (Fig. 3f), indicating their rapid redirection to the extracellular milieu to strengthen the biofilm shield.

a P. aeruginosa central metabolism; b–f Intracellular metabolite abundance in WT PAO1 exposed or not to itaconate (n = 3); g–k Isotope carbon tracing in WT PAO1 exposed or not to ¹³C-itaconate. Different isotopologues per metabolite are coded with numbers-colors (n = 3). Data are shown as average +/− SEM. b–f: t-Student test. All statistical tests are two-sided. Source data are provided as a Source Data file.

Itaconate fuels P. aeruginosa EPS remodeling

We wondered how itaconate promoted P. aeruginosa EPS reprograming. We hypothesized that this process occurs as part of the pathogen’s strategy to withstand the immunometabolite toxicity⁴⁰. Itaconate is considered a major bactericidal agent⁴⁰. P. aeruginosa, unlike other opportunists, can endure itaconate burden by breaking it down into acetyl-CoA and pyruvate, two sources of gluconeogenesis⁴⁰ (Supplementary Fig. 5a and b). We postulated that itaconate drives P. aeruginosa anabolic reprograming by incorporating carbon flux into gluconeogenesis.

We tracked the integration of ¹³C-itaconate into the P. aeruginosa PAO1 central metabolism through isotope carbon labeling (Fig. 3g). These assays showed that itaconate highly assimilated into the backbone of both acetyl-CoA and pyruvate, as ~70% and ~60% of all isotopologues of these molecules, respectively, showed ¹³C signature (Fig. 3h). This incorporation enabled the avid circulation of ¹³C-itaconate carbon atoms into gluconeogenesis; while ~60% of all isotopologues from both PEP and glyceraldehyde-3-phosphate (G3P) presented the ¹³C signature, ~80–90% of isotopologues from fructose-1,6-biphosphate, fructose-6-phosphate, and glucose-6-phosphate, end products of gluconeogenesis, manifested this labeling (Fig. 3h).

Carbon atoms of ¹³C-itaconate reached upstream networks that employ byproducts of gluconeogenesis to facilitate EPS synthesis, like the PPP and DNPPS. This was illustrated by a prominent ¹³C signature in substantial isotopologue fractions of D-Ribose-5-P ( ~ 80%), orotate ( ~ 40%), uracil ( ~ 80%), uridine ( ~ 90%) and nucleotides like UMP ( ~ 90%), UDP ( ~ 90%) and UTP ( ~ 90%) (Fig. 3i–k). Ultimately, carbon atoms of itaconate dominated the architecture of many EPS precursors like UDP-glucoronate ( ~ 95%), UDP-GlcNAc ( ~ 95%), and UDP-glucose ( ~ 95%) (Supplementary Fig. 5c), confirming how this mitochondrial factor effectively assumes control over the P. aeruginosa production of pro-biofilm determinants.

Itaconate suppresses harmful inflammation by facilitating pulmonary glutaminolysis

Given that disease tolerance is coordinated by bilateral host-pathogen adaptations, we anticipated that itaconate, in addition to reprogram P. aeruginosa metabolism, would also impact pulmonary cells. Itaconate’s inhibition of host mitochondrial complex II¹⁴ would initiate an array of compensatory mechanisms to replenish the TCA cycle and restore cell bioenergetics, such as glutamine catabolism via glutaminolysis⁴¹. Indeed, in neurons, complex II blockade by itaconate prompts glutaminolysis and accumulation of the glutamine byproduct αKG⁴². Of interest, in macrophages, αKG represses IL-1β signaling via epigenetic changes, particularly through the H3K27 demethylase Kdm6b⁴¹. We anticipated that during P. aeruginosa pneumonia itaconate would suppress IL-1β in AMs via glutaminolysis and Kdm6b stimulation.

To promote glutaminolysis, AMs assimilate glutamine through transporters like Slc38a2, Slc38a1, Slc1a5, and Slc3a2⁴¹. Then, glutamine is converted into glutamate via Gls (rate-limiting step) and subsequently to αKG by Glud1 (Fig. 4a, blue arrows)⁴¹. Transcriptomic analyzes of AMs through single-cell RNA sequencing (scRNA-Seq) in lungs of infected Irg1^+/+ and Irg1^−/− mice confirmed that itaconate upregulates glutaminolysis during P. aeruginosa pneumonia, particularly by promoting expression of Slc38a1, Slc38a2, Slc3a2, Gls and Kdm6b (Fig. 4b–d). This setting was supported by reduced levels of Glul, which antagonizes Gls (Fig. 4c), and lower mRNA abundance of the inflammasome constituents nlpr3 and Il1b (Fig. 4c). Consistently, itaconate provoked glutamine depletion from the infected airway (Fig. 4e), its enrichment within the pulmonary tissue (Fig. 4f, left panel), and its conversion into glutamate (Fig. 4f, right panel). Inhibition of glutaminolysis with the Gls inhibitor BPTES in infected mice exacerbated IL-1β levels and bacterial burden (Fig. 4g, h), affirming its role in controlling inflammation without affecting IL-6 or TNFα (Fig. 4g). Itaconate did not affect other routes involved in AMs glutamine assimilation and stimulation of IL-1β signaling, such as the GABA shunt (Gad1, Abat, Aldh5a1, hif1a)⁶ (Fig. 4a, red arrows; Supplementary Fig. 6a).

a Glutamine metabolism in AMs. Lungs from Irg1^+/+ and Irg1^−/− mice exposed to either PBS or WT PAO1 were studied by scRNA-Seq: b cell subsets; c glutamine metabolism genes in AMs; d Glutaminolysis score in AMs (n = 1, pool of two mice per group, total of 12-48 cells); e BAL metabolite enrichment (n = 3, total of 6-9 mice per group); f lung tissue metabolite levels (DESI-2D). Mice were exposed or not to WT PAO1 and administered or not with BPTES (n = 3, total of 3–5 mice per group). The following were measured: g BAL cytokines; h bacterial burden. i Bacterial energy production with glutamine (Biolog Technology) (n = 3). j Bacterial growth in glutamine (n = 3). k BAL metabolite enrichment (n = 3, total of 6–9 mice per group). l Bacterial energy production with glutamate (Biolog Technology) (n = 3). m Bacteria growth in glutamate (n = 3). Data are shown as average +/− SEM. e, i, k, l t-Student test. d, g, h One-Way ANOVA (Tukey multiple comparison test); j, m Two-Way ANOVA. All statistical tests are two-sided. Source data are provided as a Source Data file.

Together, these findings indicate that, during P. aeruginosa infection, respiratory cells employ glutamine to mitigate pathological inflammation through glutaminolysis, a process facilitated by itaconate.

Itaconate stimulates AMs glutaminolysis by interfering with ETC

Since glutaminolysis is typically activated to compensate for mitochondrial bioenergetic insufficiency, particularly by fueling the TCA cycle⁴³, we hypothesized that itaconate promotes this response during P. aeruginosa pneumonia by disrupting ETC integrity. ETC activity, driven by complexes I, II, III and IV in the inner membrane of the mitochondrion, is coordinated by a mega molecular structure known as the “respirasome” (I₂II₂III₂IV₂)⁴⁴. The respirasome supports ATP synthesis by complex V, which energizes the inflammasome activity to produce IL-1β ¹⁶. ScRNA-Seq studies and pathway enrichment analyzes of lungs from infected Irg1^+/+ and Irg1^-/- mice revealed that itaconate alters the expression of key ETC subunits involved in respirasome integrity in AMs (Supplementary Data 1–2). The immunometabolite increased mRNA levels of mt-Nd1234, mt-Cytb, mt-Co123, and mt-Atp6, which, respectively, form the core membrane embedded subunits of complexes I, III, IV and V⁴⁴ (Supplementary Fig. 6b, c). However, itaconate reduced expression of essential supernumerary subunits required for the proper assembly of these complexes, such as Ndufa4-Ndufa11–Ndufb2-Ndufb4 (complex I), Uqcrb-Uqcrh (complex III), and Cox7a2-Cox4i1-Cox5b-Cox7c (complex IV)⁴⁵ (Supplementary Fig. 6b, c). While reducing Cox7a2, itaconate conserved its isoform Cox7a2l (P > 0.05) (Supplementary Fig. 6b, c, red quadrant), which forms less efficient respirasomes⁴⁵. These effects are specific to AMs, with no significant changes in other pulmonary cell subsets (P > 0.05 for all genes) (Supplementary Fig. 7). Together, these findings confirm that itaconate facilitates glutaminolysis and tolerance to P. aeruginosa infection through bioenergetic impairment of specific respiratory cells.

Alginate-producing P. aeruginosa strains facilitate airway glutaminolysis precursors

Unlike the ΔalgD PAO1 strain, its WT PAO1 counterpart, which produces alginate, relies less on environmental glutamine for energy generation and growth (Fig. 4i, j). This metabolic configuration enabled more airway glutamine availability during infection (Fig. 4k). Additionally, alginate did not affect P. aeruginosa’s ability to employ glutamate as energy source to proliferate (Fig. 4l, m), facilitating its depletion from the airway by the host (Fig. 4k). Collectively, these findings confirm how P. aeruginosa tailored for disease tolerance facilitates immunoregulatory pathways that rely on glutamine accessibility.

P. aeruginosa reprograms its bioenergetics in response to itaconate

The bioenergetic adjustment imposed by itaconate on the host prompted us to evaluate whether this might also occur in P. aeruginosa, leading to pathogen adaptation to survive. Integrating RNA-Seq with unbiased pathway analysis showed that P. aeruginosa PAO1 responded to itaconate by activating catabolic networks producing the ketone body acetoacetate (AcAc), including metabolism of the nutrient lipoic acid, degradation of the branched amino acids valine, leucine, and isoleucine, catabolism of terpenes like pinene, camphor, and geraniol, oxidation of propanoate, and metabolism of 2-oxocarboxylic acids (Supplementary Fig. 8a, Supplementary Data 3). Adding exogenous AcAc restored growth in strains unable to cope with itaconate stress, like Δict PAO1 (Supplementary Fig. 8b). P. aeruginosa also downregulated energy-intensive processes that incur substantial ATP expenditure, such as repressors of biofilm formation, phenazine biosynthesis, quorum sensing, the type VI secretion system (T6SS), sulfur assimilation routes and ABC transporters (Supplementary Fig. 8a). These changes conserved the pathogen ATP-GTP homeostasis (Supplementary Fig. 8c, d). However, the preserved energy was redirected to early alginate production, limiting biomass growth (Supplementary Fig. 8e–g). This indicates that P. aeruginosa adapts to itaconate but remains constrained by the metabolic demands of alginate synthesis.

P. aeruginosa inactivation of the anti-sigma factor mucA is a hallmark of adaptation to disease tolerance

The metabolic constrain found in the tolerant lung leads P. aeruginosa to evolve compensatory mutations^46,47,48. The adaptation of P. aeruginosa to this milieu, particularly the interaction between itaconate and alginate synthesis, attenuates the anti-sigma factor mucA (Fig. 5a). MucA inactivation is common in the lung of long-term infected hosts, such as people with cystic fibrosis (CF), chronic obstructive pulmonary disease (COPD), and bronchiectasis (BC), which associates with tolerance disruption and severe pulmonary deterioration^49,50. Among 51 types of inactivating mucA mutations found in 1014 genomes of P. aeruginosa isolates from these tolerant subjects (Supplementary Data 4), mucA22 emerged as the most frequent adaptation (Fig. 5b, c, Supplementary Fig. 9a). This mutation is not observed in strains from individuals at intensive care unit (ICU), which have not transitioned into disease tolerance yet (Fig. 5b, c, Supplementary Fig. 9a).

a mucA mRNA levels (n = 3, 8-12 technical replicates). Number of P. aeruginosa genomes studied: ICU: 182; CF: 532; COPD: 90; COPD: 210. In these genomes, the following was evaluated: b Frequency of mucA mutations in P. aeruginosa isolates genomes. c Frequency of different mucA mutations in isolates; d, e Unbiased pathway enrichment analyzes between WT and mucA22 PAO1 (n = 3). Top 5 significantly changed pathways are shown. f, g Bacterial oxygen consumption rate (Seahorse) (f) and total oxygen consumption (area under the curve) (g) (AUC: arbitrary unit count) (n = 4). h Bacterial energy production (Biolog Technology) (n = 3). i growth curves (OD₆₀₀) (n = 3). Data are shown as average +/− SEM. d, e t-Student test; a, g, h One-Way ANOVA (Tukey multiple comparison test); f, i Two-Ways ANOVA. All statistical tests are two-sided. Source data are provided as a Source Data file.

MucA22 is a stop-codon mutation that fully abrogates mucA function^50,51,52. Despite bolstering P. aeruginosa alginate production⁵³ (Supplementary Fig. 1b, c), mucA22 causes severe pulmonary exacerbations linked with secretion of cytokines that are not suppressed by itaconate, such as TNFα and TNFα-dependent chemoattractants like IL-8, MIP1α and MIP2^54,55,56, and massive infiltration of phagocytes that hinder the airway⁵⁷. MucA22 also boosts P. aeruginosa succinyl-coA ligase (sucCD) activity⁴⁸, potentially mitigating the TCA cycle impairment induced by itaconate (Fig. 2). We anticipated that, by offsetting the bacterial bioenergetic stress, this mutation would disrupts the host-pathogen synchrony that maintains disease tolerance.

MucA22 enhances P. aeruginosa TCA cycle bioenergetics

Analysis of publicly available transcriptomic datasets for WT PAO1 and its isogenic mucA22 PAO1 mutant^58,59 revealed that inactivation of the anti-sigma factor boosts P. aeruginosa TCA cycle function, making the pathogen resilient to the bioenergetic stress imposed by itaconate. By integrating unbiased pathway enrichment analyzes with the Kyoto Encyclopedia of Genes and Genomes (KEGG) and the Gene Ontology for Biological Processes (GOTERM BP) libraries, we identified the top five P. aeruginosa networks with the highest frequency of significantly altered (P < 0.05) genes by mucA22 (Fig. 5d–e, Supplementary Data 5). Within the “metabolic pathways” category, notable increased genes included aconitate hydratase (acnA), which converts citrate to aconitate, a αKG’s precursor that stimulates the TCA cycle⁶⁰ (Fig. 5d, Supplementary Data 5), as well as adenylosuccinate lyase (PA3517), fumarate hydratase (fumC2), and cytochrome C (PA2266), which collectively facilitate production of the succinate byproduct fumarate⁶¹, enable its oxidation to malate⁶², and transport electrons from this reaction across the ETC to synthesize ATP⁶³ (Fig. 5d, Supplementary Data 5). Consistently, mucA22 PAO1 demonstrated enhanced aerobic metabolism with succinate compared to WT PAO1, a process independent of alginate, as shown by an isogenic mucA22ΔalgD PAO1 mutant (Fig. 5f–g, Supplementary Fig. 1b, c). This heightened oxygen consumption linked to a 2-fold increase in energy generation with succinate, rising to 4-fold when alginate synthesis occurred (Fig. 5h). Such bioenergetic reprogramming likely provides P. aeruginosa with an energy surplus under itaconate stress. Indeed, despite reduced proliferation during EPS production, alginate-generating mucA22 PAO1 exhibited resilience to itaconate-induced growth inhibition, both in vitro (Fig. 5i) and in vivo (Supplementary Fig. 9b). Furthermore, this setting dispersed surface-attached P. aeruginosa biofilms by itaconate (Supplementary Fig. 9c–d). These findings demonstrate how mucA22 equips P. aeruginosa to overcome metabolic stress imposed by the itaconate-alginate axis, enhancing its bioenergetic capacity and adaptive fitness.

MucA22 P. aeruginosa conserves alginate-mediated tolerance but exacerbates TNFα signaling

Transcriptomic analyzes revealed that mucA22 enhances P. aeruginosa’s resilience to environmental threats while activating inflammatory pathways independent of itaconate, such as TNFα signaling. MucA22 elicited loci involved in the breakdown of complex carbohydrates, resistance to osmotic pressure, and tolerance to desiccation stress^49,64,65, such as trehalase (treA), glycogen synthase (PA2165), glycogen phosphorylase (glgP), UDP-glucose 6-dehydrogenase (PA2022), ABC transporter permeases PA3888-PA3891 and glutamine amidotransferase (PA3459) (Supplementary Data 5). This stress-resistance program aligned with activation of the “alginate biosynthetic process” (Fig. 5d), utilizing sugar byproducts to sustain EPS production through algD and auxiliary algABEFGJKLU8 genes^49,64,65.

Interestingly, mucA22 downregulated pyoverdine synthesis genes but paradoxically boosted pyoverdine release, a siderophore that exacerbates airway destruction (Fig. 5e, Supplementary Fig. 9e, Supplementary Data 5). This equilibrium might be part of a negative feedback loop triggered by the siderophore itself. Additionally, mucA22 induced pro-inflammatory adjustments linked to TNFα signaling, such as pilA and pilB, components of type IV pili, and the algZ (amrZ)-algR system, which disperses biofilms into planktonic forms by repressing PelA and Psl EPS genes (Fig. 5d–e, Supplementary Data 5). These adaptations escalate airway immunopathology, undermining disease tolerance through alginate and itaconate-driven mechanisms.

MucA22 promotes airway disruption through TNFα signaling

In mice infected with mucA22 P. aeruginosa, TNFα and related chemoattractants – e.g., MIP1α (Ccl3), MIP2 (Cxcl2), and the murine IL-8 homolog KC (Cxcl1) – increased, leading to higher pathogen loads and airway inflammation (Fig. 6a, b). As indicated by the mucA22ΔalgD PAO1 mutant, alginate partially attenuated bacterial burden (Fig. 6b), but not TNFα signaling (Fig. 6a). Although alginate production by mucA22 PAO1 favored many parameters of disease tolerance, such as airway itaconate enrichment (Supplementary Fig. 10a), less generation of T3SS necrotizing toxins (Supplementary Fig. 10B), decreased IL-1β and IL-6 (Fig. 6c), reduced alveolar permeabilization (Fig. 6d), and limited host mortality (Fig. 6e), these pathogens exacerbated the pulmonary recruitment of myeloid cells, such as monocytes (Supplementary Figs. 10c–e; recruited monocytes 2 by scRNA-Seq, Fig. 4b) and neutrophils (Supplementary Figs. 10c–e; recruited neutrophils 1 by scRNA-Seq, Fig. 4b).

a Mice were exposed to either PBS, WT PAO1, mucA22 PAO1 or mucA22ΔalgD PAO1 (n = 3, total of 8-11 mice per group). The following were analyzed: a, c BAL cytokines; b bacterial burden; d BAL albumin; e host survival. Mice were exposed to either PBS, WT PAO1, or mucA22 PAO1 (n = 2, total of 4–9 mice per group). The following were analyzed: f, g numbers and viability of type 1 pneumocyte in BAL; h, i, k lung cell numbers; j BAL VEGF. Data are shown as average +/− SEM. a–d, f–k: One-Way ANOVA (Tukey multiple comparison test). e: Kaplan-Meier test. All statistical tests are two-sided. Source data are provided as a Source Data file.

As indicated by scRNA-Seq and unbiased pathway enrichment analyzes, the shared signature of recruited monocytes 2 and recruited neutrophils 1 is expression of Tnf, its receptor Tnfrsf1b, and the regulatory protein Tnfaip3, as well as different inflammatory circuits elicited by TNFα, such as toll-like receptors 2 and 4 (Tlr2, Tlr4), cytoplasmic kinases (Map3k8, Ikbke), NFkΒ signaling (Nfkb1, Nfkb2, Nfkbia), and the chemoattracting factors Ccl3 (MIP1α), Cxcl2 (MIP2) and Cxcl1 (KC) (Supplementary Figs. 10f, g, Supplementary Data 6). Separately, recruited monocytes 2 downregulate pathways critical for pathogen clearance, including the respirasome (mt-Nd1234, Sdha, mt-Cytb, mt-Co123, mt-Atp6, Ndufa11–Ndufb247, Uqcrb, and Coxi1-Cox7c-Cox7a2-Cox7a2l)^44,45 and phagocytosis promoters (Pi3kr1, Sirpb1bc, Sirpa, Mapk14, c-Fos, Calm2, Calm3)^66,67,68 (Supplementary Fig. 10f). Analogously, recruited neutrophils 1 abrogate expression of essential genes for effector function (Csf1r, Csf3r)^69,70, phagocytosis (c-Fos)⁶⁶, oxidative burst (Rock2, Ncf2)^71,72, bacterial degradation (Lyz2)⁷³, and priming of naïve T cells against foreign antigens (H2-Ab1, H2-Q10, Cd74, Itga4, Itgal, Ptgs2, Il6ra, Cxcr4, Sell, Cst3, and Hsp90ab1)⁷⁴ (Supplementary Fig. 10g). Together, these findings reveal how mucA22 hinders airway integrity by facilitating the accretion of TNFα-producing phagocytes incapable of eliminating P. aeruginosa.

TNFα-driven inflammation damages type 1 pneumocytes, vital for alveolar function^75,76,77, leading to their detachment into the airway space (Fig. 6f), subsequent death (Fig. 6g), and compensatory type 2 pneumocyte hyperplasia to replenish them in the lung tissue^78,79 (Fig. 6h, i). This repair response increases VEGF levels⁸⁰ (Fig. 6j) and ciliated cell proliferation (Fig. 6k), attempting to clear airway debris but ultimately disrupting respiratory-vascular homeostasis. Thus, mucA22 exacerbates airway pathology by driving ineffective, necrotic immune responses that destabilize respiratory equilibrium.

Evidence of P. aeruginosa adaptation to alginate-driven disease tolerance in the human lung

To gain clinical insights in the mechanistic described, we studied FRD1, a widely used, multidrug resistant (MDR) isolate of P. aeruginosa from the lung of a chronically infected CF patient^81,82. Although this strain produces copious alginate, its mucA22 mutation drives severe immunopathology^83,84. Our analyzes confirmed FRD1 harbors vestiges of progressive adaptation to disease tolerance; 1) by preserving many traits linked with itaconate immunoregulation, such increased alginate synthesis and less production of inflammatory toxins; and 2) by heightening succinate bioenergetics via mucA22.

Compared with WT PAO1, FRD1 harbors 9,603 NSM, including 82 that significantly impact gene function (frameshift, stop gained, start lost, etc), 5,635 with moderate effects on gene activity (missense variant, conservative in frame deletion, conservative in frame insertion, etc), and 3,886 affecting intergenic regions (Supplementary Data 7). While few NSM affected alginate-related genes, including algC, algI, algK, alg8, and alg44 (Fig. 7a), and the integrity of the algD locus was fully preserved (Fig. 7a), clusters for alternative EPS like Pel and Psl accumulated numerous inactivating mutations (Fig. 7a). This mutation pattern correlated with reduced mRNA levels (Fig. 7a). In vivo, FRD1’s alginate-centered metabolism reduces bacterial burden (Fig. 7b) and dampens immunopathology, lowering markers like pro-inflammatory cytokines (IL-1β, IL-6, TNFα, MIP1α, MIP2, and KC) (Fig. 7c), airway permeability (Fig. 7d), and loss of body temperature regulation (Fig. 7e), as shown by an isogenic ΔalgD FRD1 strain. However, alginate has no effect on FRD1 lethality (Fig. 7f), suggesting complementary adaptations that mitigate bacterial pathology in the absence of the EPS.

a Number of non-synonymous mutations (# NSM) and gene expression level (logFC) for loci involved in EPS synthesis in FRD; control: WT PAO1. Mice were exposed to either PBS, WT PAO1, FRD1 or ΔalgD FRD1 (n = 2, total of 11-12 mice per group). The following were analyzed: b pathogen burden; c BAL cytokines; d BAL albumin; e body temperature; f Animal survival. Data are shown as average +/− SEM. B-E: One-Way ANOVA (Tukey multiple comparison test); f Kaplan-Meier. All statistical tests are two-sided. Source data are provided as a Source Data file.

FRD1 accrued substantial NSM in genes related to immunopathology, such as the involved in LPS O-antigen modification (e.g., WbpMZYW, wzt, PA5455-PA5459), LPS lipid A synthesis and modification (e.g., wzz2, waaGPL, eptC, lpxBD), and LPS surface exposure (e.g., lptD, lptG) (Supplementary Fig. 11a). LptD inactivation in FRD1 is sufficient to preserve host survival, as confirmed by a PAO1 strain lacking this locus (ΔlptD PAO1) (Supplementary Fig. 11b). Thus, FRD1 exhibits vestiges of adaptation to disease tolerance, balancing EPS and LPS integrity to mitigate disease.

P. aeruginosa FRD1 enhances succinate bioenergetics

As result of its mucA mutation mucA22, FRD1 reprograms its TCA cycle to prioritize succinate metabolism. Compared with WT PAO1, key genes for succinate generation and oxidation (sucCD, sdhABCD) are preserved and upregulated, whereas earlier TCA cycle steps (icd, idh, sucB, lpdG, lpdV, lpd3) are less conserved (Fig. 8a). FRD1 compensates for a mutation in fumC1, which utilizes the succinate byproduct fumarate, by maintaining the sequence integrity of fumC2, mqoA, mqoB, and gltA, as well as promoting their expression (Fig. 8a). This setting ensures efficient succinate bioenergetics and increases aerobic metabolism, independent of alginate (Fig. 8b, c). Like mucA22 PAO1 (Fig. 5h), FRD1 couples alginate synthesis with enhanced energy generation from succinate and fumarate, overcoming the metabolic burden of EPS production (Fig. 8d). Indeed, despite synthesizing alginate, FRD1 still generated higher biomass than WT PAO1 (Fig. 8e). Thus, by optimizing succinate-fumarate bioenergetics, FRD1 outcompetes the bioenergetic encumbrance of alginate production, enabling increased bacterial loads.

a Number of non-synonymous mutations (# NSM) and gene expression level (logFC) in FRD1 for TCA cycle clusters; control: WT PAO1. b, c oxygen consumption rates (OCR) by Seahorse technology (b) and total oxygen consumed along time (c) (AUC: arbitrary unit count) (n = 4). d bacterial energy production (Biolog Technology) (n = 3). e growth curves (OD₆₀₀) (n = 3). Data are shown as average +/− SEM. c, d: One-Way ANOVA (Tukey multiple comparison test); e Two-Ways ANOVA. All statistical tests are two-sided. Source data are provided as a Source Data file.