A respiro-fermentative strategy to survive nanoxia in Acidobacterium capsulatum Authors: Daniela Trojan, Emilio García-Robledo, Bela Hausmann, Niels Peter Revsbech, Dagmar Woebken, Stephanie A Eichorst Published: 2024 ABSTRACT AbstractMicrobial soil habitats are characterized by rapid shifts in substrate and nutrient availabilities, as well as chemical and physical parameters. One such parameter that can vary in soil is oxygen; thus, microbial survival is dependent on adaptation to this substrate. To better understand the metabolic abilities and adaptive strategies to oxygen-deprived environments, we combined genomics with transcriptomics of a model organism, Acidobacterium capsulatum, to explore the effect of decreasing, environmentally relevant oxygen concentrations. The decrease from 10 to 0.1 µM oxygen (3.6 to 0.036 pO2% present atmospheric level, respectively) caused the upregulation of the transcription of genes involved in signal transduction mechanisms, energy production and conversion and secondary metabolites biosynthesis, transport, and catabolism based on clusters of orthologous group categories. Contrary to established observations for aerobic metabolism, key genes in oxidative stress response were significantly upregulated at lower oxygen concentrations, presumably due to an NADH/NAD+ redox ratio imbalance as the cells transitioned into nanoxia. Furthermore, A. capsulatum adapted to nanoxia by inducing a respiro-fermentative metabolism and rerouting fluxes of its central carbon and energy pathways to adapt to high NADH/NAD+ redox ratios. Our results reveal physiological features and metabolic capabilities that allowed A. capsulatum to adapt to oxygen-limited conditions, which could expand into other environmentally relevant soil strains.Keywords: Acidobacteriota, oxygen limitation, microaerobic respiration, fermentation, NADH imbalances, transcriptome Introduction Microorganisms face a multitude of fluctuating and often limiting conditions across various environments, such as soils, human gut, and aquatic environments. Carbon, electron acceptors (such as oxygen (O2)), and/or nutrients can vary over space and time. As such, microorganisms need to compensate and employ strategies to survive during these potentially growth-restricting conditions. One such strategy is respiratory flexibility. The utilization of both high- and low-affinity terminal oxidases enables exploitation of the full range of O2 concentrations for oxidative phosphorylation and energy conservation, providing a great benefit in the ever-changing O2 concentrations across environments. This can be attained by inducing branched respiratory chains that terminate in multiple oxidases with different affinities for O2 (Bueno et al. 2012) recently shown in members of ubiquitous soil bacteria, the Acidobacteriota (Eichorst et al. 2018, Trojan et al. 2021). Other strategies by which cells respond to limitations are, e.g. modifying enzyme synthesis to take up growth-limiting nutrients or by modulating uptake rates for nutrients available in excess (Roszak and Colwell 1987). Alternatively, they can reroute metabolic fluxes, which enables them to shift to alternative sources of energy and building blocks while avoiding possible blockages due to specific nutrient limitations (Roszak and Colwell 1987, Bergkessel et al. 2016). Catabolism and ATP production are often incongruent during these periods of limitation (Stouthamer 1979). As a result of this incongruency, a trade-off can occur between catabolic rate and ATP yield, whereby bacteria utilize pathways for the most efficient production of molar ATP yield (YATP: mole of ATP/mole of oxidized substrate). For example, when catabolic rates are high but O2 limiting, fermentative pathways (when available) are employed, together with respiratory pathways, commonly referred to as respiro-fermentation physiology (Pfeiffer et al. 2001, Vemuri et al. 2007) allowing bacteria to maximize ATP production during electron acceptor limitation. This respiro-fermentative physiology has been observed in Escherichia coli, Bacillus subtilis, and Saccharomyces cerevisiae, yet the evolution and regulation of this metabolism is still under debate (Molenaar et al. 2009). Presumably, bacteria have evolved to harbor greater metabolic flexibility for ATP production, rather than pathways yielding optimal growth yield (Stouthamer 1979). Members of the phylum Acidobacteriota are ubiquitous across numerous soils (Fierer 2017, Delgado-Baquerizo et al. 2018) with a central role in carbon mineralization and plant decomposition (Fierer 2017, Crowther et al. 2019). Still very little is known about factors controlling their abundance in the environment or their effects on biogeochemical cycles under changing environmental conditions. In this study, we investigated the adaptive capability to O2-limited conditions with a model member of the phylum Acidobacteriota, Acidobacterium capsulatum 161. It is a member of the family Acidobacteriaceae that is commonly found across many environments, such as soils. Acidobacterium capsulatum 161 has originally been documented to be capable of microaerophilic growth and only later of weak fermentative growth as well (Pankratov et al. 2012, Myers and King 2016). Recently, its capacity for respiratory flexibility due to the presence and functionality of high- and low-affinity terminal oxidases was demonstrated (Trojan et al. 2021). Here, we expanded our investigation of this strain to ascertain if it has additional abilities to alter its metabolism, such as the rerouting of metabolic fluxes, by profiling the whole transcriptomic response of A. capsulatum 161 to decreasing O2 concentrations in the micro- and nanomolar range, the latter referred to as nanoxic (<1 µmol O2 l−1) (Berg et al. 2022). To date, no reports have closely documented the catabolic routes of carbon and energy metabolism of Acidobacteriota or have evaluated its global transcriptomic response to O2 deprivations. By combining genomics and hypoxic culture incubations using highly sensitive optical O2 sensors (Lehner et al. 2015), we were able to investigate transcription patterns at the oxic–anoxic interface and could observe a transition from respiratory to respiro-fermentative metabolism in A. capsulatum 161. Material and methods Growth conditions and experimental setup As previously described (Trojan et al. 2021), A. capsulatum 161 (ATCC 51196, DSM 11244) was grown up in biological quadruplicates in a vitamins and salts base medium (Eichorst et al. 2007, 2011) amended with 10 mM glucose as the sole carbon source with a pH of 5 under fully aerated conditions at room temperature. The setup of the 225-min microoxic incubations using two LUMOS systems (Lehner et al. 2015) for O2 concentration monitoring and sample collection at discrete, declining O2 concentrations (10 µmol O2 l−1, 0.1 µmol O2 l−1, and 0.001 µmol O2 l−1) down to anoxia (0 µmol O2 l−1 is < 0.0005 µmol O2 l−1) was previously described in Trojan et al. (2021). Briefly, at select time points (Fig. S1), 30 ml of culture were collected for RNA extraction and transcriptome sequencing by a syringe. For an immediate inactivation, the syringes were prefilled with an acidic phenol-stop solution (Kits et al. 2014) and precooled at 4°C. After centrifugation, cell pellets were snap frozen in liquid nitrogen and then stored at –80°C. RNA extraction and purification RNA was extracted from frozen cell pellets using an acidic phenol/chloroform/isoamyl alcohol protocol (Griffiths et al. 2000) with mechanical disruption (30 s, 4 ms−1, FastPrep-24 bead beater, MP Biomedicals, Heidelberg, Germany) (Trojan et al. 2021). Purification of RNA and verification of complete DNA removal were described previously (Trojan et al. 2021). Transcriptome sequencing Triplicate RNA samples from selected O2 concentrations and time points were sent to the Vienna BioCenter Core Facilities for sequencing. rRNA was depleted using the NEB Ribo-Zero rRNA removal kit for bacteria. Sequencing was performed on an Illumina NextSeq 550 system resulting in a total of 8.2–18.2 million 75-nucleotide reads per sample; more details can be found in Trojan et al. (2021). Data processing and statistical analyses Raw reads were trimmed of sequencing adapters and low-quality 3′ ends using BBduk (BBtools v37.61, https://jgi.doe.gov/data-and-tools/bbtools/) with default parameters and error-corrected using Bayes–Hammer module of SPAdes assembler version 3.13.0 (Nikolenko et al. 2013). Any reads mapping to either the SILVA SSU or LSU releases 132 (Quast et al. 2013) or the 5S rRNA database (Szymanski et al. 2016) with a sequence identity >70% (performed with BBmap, BBtools, https://jgi.doe.gov/data-and-tools/bbtools/) were removed from the dataset (Table S1). The remaining reads were mapped to the publicly available genome of the Acidobacterium capsulatum 161 (Eichorst et al. 2020). The RNA reads per gene were summarized using the featureCounts tool from the Subread package v1.6.2 (Liao et al. 2014). Based on the generated read count tables, transcripts per million (TPMs) were calculated in R v3.6.0. Differential expression analysis, such as the calculation of log2-fold changes of relative transcript abundance and the significance of these changes were calculated in DESeq2 v1.26.0 using default parameters and a P-value cutoff of .05 (Love et al. 2014). Results Global transcriptomic response under decreasing O2 concentrations Our analyses revealed that the decrease from 10 to 0.1 µmol O2 l−1 had the greatest impact on gene expression with the highest number of significantly differentially expressed genes observed (Fig. 1A). Subsequent transitions to 0.001 and further to 0 µmol O2 l−1 invoked only a few to no significant expression changes (Fig. 1A). At 0.001 µmol O2 l−1, oxygen was still being supplied at 10.1 µmol O2 l−1 but could no longer be accurately determined, therefore is defined as “apparent anoxia.” Ninety-seven percent of all annotated genes (n = 3321 genes) were transcribed at least at one time point across the O2 concentrations. A significant difference in transcript numbers between 10 and 0.1 µmol O2 l−1 was detected for 2677 (∼81%) of transcribed genes. Of these genes, 41% were upregulated and 40% were downregulated at 0.1 µmol O2 l−1 (Fig. 1B). Genes encoding hypothetical proteins with no further functional annotation accounted for ∼25% of the differentially transcribed genes, whereas ∼26% and ∼30% of significantly upregulated and downregulated genes, respectively, were assigned to protein-coding genes with annotated function (Fig. 1B). Impact of decreasing O2 concentrations on the transcriptome of A. capsulatum 161. (A) Heatmap depicts the proportion of genes that were differentially expressed (P < .05) between two O2 concentrations (µmol O2 l−1). The darker the color, the higher the proportion of genes whose expression has significantly changed between two O2 concentrations. (B) Breakdown of differentially transcribed genes (P ≤ .05) upon the decrease from 10 to 0.1 µmol O2 l−1. Downregulation is depicted in blue, upregulation in red. Proportion of genes encoding hypothetical proteins with no further functional annotation are shaded in light blue and light red, respectively. (C) The sum of TPM for all protein-coding genes transcribed in given COG categories in the transcriptomes at 10, 0.1, and 0.001 µmol O2 l−1. The number of transcribed genes per category is given in parentheses. Comparisons were done after 60 min at each respective O2 concentration with two exceptions: * depicts differential expression after 10 min, and ° depicts differential expression after 50 min. (D) Proportion of significantly differentially expressed genes upon the decrease from 10 to 0.1 µmol O2 l−1 for individual COG categories. Blue and red bars represent the percentages of genes showing lower and higher expression, respectively. Number of genes showing significantly differential higher or lower expression for each category is represented next to the bars. 0.001 µmol O2 l−1 represents an “apparent anoxia” as this concentration is close to the detection limit (0.0005 µmol O2 l−1) and could not be measured anymore, although 10.1 µmol O2 min−1 was still provided. Differential response of function-based categories to decreasing O2 concentrations Genes that exhibited significant differential expression in response to O2 decrease were classified in clusters of orthologous groups (COGs) (Fig. 1C and D). The COG categories of energy production and conversion (C), amino acid transport and metabolism (E), nucleotide transport and metabolism (F), and lipid transport and metabolism (I) were the categories that were most altered by the decrease from 10 to 0.1 µmol O2 l−1 as more than 80% of the total genes assigned to these categories exhibited significant differential gene expression (Fig. 1D). More specifically, the transition from 10 to 0.1 µmol O2 l−1 reduced the number of TPMs in COGs related to translation, ribosomal structure, and biogenesis (J); transcription (K); carbohydrate transport and metabolism (G); cell wall, membrane, and envelope biogenesis (M); and intracellular trafficking, secretion, and vesicular transport (U) (Fig. 1C), and many of these categories had a higher proportion of significantly downregulated genes due to reduced O2 (Fig. 1D). Interestingly, the number of transcripts per million (TPMs) in COG categories pertaining to energy production and conversion (C), secondary metabolites biosynthesis, transport, and catabolism (Q), and signal transduction mechanisms (T) increased with decreasing O2 concentrations (Fig. 1C), and some of these categories also had a higher proportion of significantly upregulated genes due to reduced O2 (Fig. 1D). Universal and reactive oxygen stress response to decreasing O2 The transcription of genes encoding proteins involved in general stress response (COG category: T-signal transduction mechanisms) was significantly stimulated upon the decrease of O2 (Fig. 1C and D). In particular, a two-component sensor histidine kinase exhibited an 84-fold increase upon the shift to 0.1 µmol O2 l−1 (P ≤ .001) (Fig. 2A, Table S2). Furthermore, several homologs of the universal stress response gene uspA were significantly upregulated upon the decrease in oxygenation from 10 to 0.1 µmol O2 l−1. Four of those uspA genes had already been highly expressed at 10 µmol O2 l−1, whereas the expression of two of them, those with the highest fold increase (7- and 56-fold, respectively) seemed to be specifically stimulated by the drop to 0.1 µmol O2 l−1 (Fig. 2A, Table S2). The transcription levels of genes coding for various chaperone proteins differed between 10 and 0.1 µmol O2 l−1, but with no discernible trend of regulation (Fig. 2A, Table S2). Transcription patterns of proteins involved in (A) stress response and ROS defense and (B) enzyme complexes in the ETC of A. capsulatum 161 exposed to decreasing O2 concentrations. Time-resolved gene expression at 10, 0.1, and 0.001 µmol O2 l−1 after 60 min at each respective O2 concentration with one exception: # depicts expression after 10 min at 0.001 µmol O2 l−1. Heatmaps depict average TPM values of biological replicates (n = 3). The last column depicts statistically significant differential (P ≤ .05) log2-fold changes (log2FC) of transcripts between 10 and 0.1 µmol O2 l−1 after 60 min at the respective O2 concentration. Downregulation is depicted in blue, upregulation in red. Asterisks depict non-significant differential expression (P > .05). 0.001 µmol O2 l−1 is defined as apparent anoxia: O2 was still supplied (10.1 µmol O2 min−1) but could no longer be accurately determined. Data for all replicates, gene locus tags, and further details are listed in Tables S2 and S3. The transcriptome data at 10, 0.1, and 0.001 µmol O2 l−1 of the continuously decreasing O2 incubations exhibited the downregulation of key genes (ahpCD, trxR, katG) involved in oxidative stress defense (Fig. 2A, Table S2). Surprisingly, several other key genes coding for oxidative stress defense enzymes, catalase C (katE), non-heme manganese-containing catalase (kat), manganese superoxide dismutase (soda), heme-dependent chlorite dismutase (cld), rubrerythrin (rbr), thioredoxin (trx), ferroxidase (bfr), organic hydroperoxide resistance protein ohrB, and various homologs of alkyl hydroperoxide reductase subunits C and D (ahpCD) did not follow this trend and were transcribed at significantly higher levels at diminishing O2 concentrations (Fig. 2A, Table S2). Most of all, rbr, oda, trx, bfr, and one ahpC homolog (ahpC-2) were transcribed at high levels with an up to 13-fold upregulation upon the shift from 10 to 0.1 µmol O2 l−1 (Table S2). Expression of electron transport chain and oxidative phosphorylation We detected gene expression for all complexes I–IV of the electron transport chain (ETC) and the ATP synthase (complex V) (Table S3). Acidobacterium capsulatum 161 harbors several complexes IV of the respiratory chain, and the transcriptional responses to the decrease in O2 from 10 to 0.1 µmol O2 l−1 (Fig. 2B, Table S3) were published and discussed recently (Trojan et al. 2021). The genes encoding the ATP synthase (complex V) were expressed during all O2-limiting conditions but showed a continuous decrease in transcription level with diminishing O2 availability (Fig. 2B, Table S3). Acidobacterium capsulatum 161 expressed two proton-translocating NADH dehydrogenases (NDH-I, complex I). Out of the whole NDH-I nuoA-N operon, the nuoAC genes were continuously expressed at high transcription levels, whereas the transcription levels of other subunits (nuoE-N) decreased significantly with O2 (P ≤ .001; Fig. 2B, Table S3). The other NADH dehydrogenase was an unusual complex I (Chadwick et al. 2018), as it had a duplicated nuo-M gene. It had similar expression patterns as the NDH-I; the nuoABC transcripts were detected in significantly higher numbers at diminishing O2 concentrations (Fig. 2B, Table S3). Two homologs of the type II NADH dehydrogenase (NDH-II), which do not translocate protons across the cell membrane (Blaza et al. 2017), were transcribed. One homolog was detected at a very low level with a maximum TPM value of 10 at 10 µmol O2 l−1, whereas the second homolog was significantly higher expressed (P ≤ .001) upon the shift from 10 to 0.1 µmol O2 l−1 (21-fold), and this high expression was maintained at all subsequent O2 conditions with an average TPM value of 740 (Fig. 2B, Table S3). The operon (sdhABC) encoding the succinate dehydrogenase (SDH), complex II of the ETC, did not show any discernible pattern of regulation; e.g. the membrane subunit was upregulated, whereas the catalytic subunit (sdhB)was downregulated at 0.1 µmol O2 l−1 (Fig. 2B, Table S3). Decreasing O2 concentrations significantly altered the central metabolism With glucose in excess, decreasing O2 concentrations resulted in significant changes in the gene expression between 10 and 0.1 µmol O2 l−1 for the Embden–Meyerhof–Parnas (EMP) pathway, the Entner–Doudoroff (ED) pathway, the pentose phosphate (PP) pathway, genes of the pyruvate metabolism and the tricarboxylic acid (TCA) cycle as well as the acetate, glycogen, and gluconate metabolism (Fig. 3, Fig. S2, Table S4). More details are discussed below. Differential gene expression due to low-nanomolar O2 concentrations across the central carbon and energy pathways of A. capsulatum 161 as inferred from analysis of the annotated genome and transcriptomes. Differential expression of genes involved in the EMP pathway, the ED pathway, the PP pathway, the TCA cycle as well as the gluconate-, acetate-, and glycogen metabolism. The genes represented in red were upregulated, while those represented in blue were downregulated upon the decrease from 10 to 0.1 µmol O2 l−1. Genes in white were non-significantly upregulated nor downregulated (P > .05). “?” depicts missing gene in the annotated genome. Expression of every gene was observed in all replicates (n = 3). The metabolite abbreviations are as follows: Glc, glucose; GDL, glucono-1,5-lactone; G6P, glucose-6-phosphate; 6-PGL, 6-phosphogluconolactone; 6PG, 6-phosphogluconate; KDPG, 2-keto-3-deoxy-6-phosphogluconate; KDG, 2-keto-3-deoxygluconate; Ru5P, ribulose 5-phosphate; R5P, ribose 5-phosphate; PRPP, phosphoribosyl diphosphate; Xu5P, xylulose 5-phosphate; S7P, sedoheptulose 7-phosphate; E4P, erythrose 4-phosphate; F6P, fructose 6-phosphate; FBP, fructose 1,6-bisphosphate; DHAP, dihydroxyacetone-3-phosphate; GAP, glyceraldehyde 3-phosphate; 1,3-bisPG, 1,3-biphosphoglycerate; 3-PG, 3-phosphoglycerate; 2-PG, 2-phosphoglycerate; PEP, phosphoenolpyruvate; Acetyl-P, acetyl phosphate; HAc, protonated form of acetate; CIT, citrate; ICT, isocitrate; 2-OG, 2-oxalglutarate, SUC-CoA, succinyl-coenzyme A; SUC, succinate; FUM, fumarate; MAL, malate; OAA, oxaloacetate; L-Asp, L-aspartate; L-Glu, L-glutamate; L-Gln, L-glutamine; CDP-Glc, cytidine diphosphate glucose; G1P, glucose-1-phosphate; UDP-Glc, uridine-5-diphosphate glucose; ADP-Glc, adenosine diphosphate glucose; Tre-6P, trehalose-6-phosphate; Tre, trehalose; M1P, maltose-1-phosphate; Q, quinone; QH2, quinol. ETC, electron transport chain. Data and abbreviations for all differentially expressed genes, gene locus tags, and further details are listed in Tables S2–S4. Glucose transport and pyruvate production were downregulated at low-nanomolar O2 concentrations The overall expression of genes in the glycolytic pathways decreased from 10 to 0.1 µmol O2 l−1 (Fig. 3, Fig. S2, Table S4). The transcription level of the symporter gene responsible for transporting glucose from the periplasm to the cytoplasm, galP, was significantly decreased (3.5-fold, P ≤ .001) upon the shift of oxygenation from 10 to 0.1 µmol O2 l−1 (Fig. 4, Fig. S1, Table S4). Conceptual figure depicting the response of A. capsulatum 161 to NADH/NAD+ ratio when transitioning from high to low nanomolar O2 concentrations and the associated physiological response. (A) At high O2 concentrations in glucose-unlimited medium, the ETC sufficiently consumes the NADH and electrons generated by glycolysis and TCA to keep the NADH/NAD+ redox ratio low. (B) Under O2-limited conditions, the ETC is limited by the lack of the terminal electron acceptor and cannot metabolize all the accumulated NADH, leading to NADH imbalances, ROS stress, and a potential toxicity. (C) By the onset of mixed acid fermentation, the production of acetate, and ethanol as well as shifting to the use of NDH-II rather than NDH-I, A. capsulatum 161 counteracts NADH imbalances and produces fewer NADH per ATP. It also oxidizes extracellular glucose to gluconate (glucono-1,5-lactone). Key for the abbreviations is depicted on the top right, except for the quinol/quinone cycle (QH2/Q) and cytochrome c (cytc). Figure adapted from Szenk et al. (2017). The production of pyruvate was downregulated across various paths. Pyruvate stemming from 2-keto-3-deoxy-6-phosphogluconate (KDPG) in the ED pathway via KDPG aldolase (eda gene) (Bennett et al. 2009, Flamholz et al. 2013) was expressed at significantly lower levels (Fig. 3, Fig. S2, Table S4). And the pfkA gene, which is unique to the EMP pathway in the catabolic direction, was significantly downregulated at 0.1 µmol O2 l−1 (1.3-fold, Fig. S1, Table S4). The conversion of PEP to pyruvate, which is coupled to the synthesis of ATP, is catalyzed by pyruvate kinase (pyk gene) that exhibited a lower expression at diminishing O2 concentrations (Fig. S2, Table S4). Apart from three genes, all genes involved in the PP pathway were significantly downregulated at 0.1 µmol O2 l−1 (Fig. 3, Fig. S2, Table S4). Pyruvate metabolism was highly transcribed at low-nanomolar O2 concentrations Pyruvate occupies a key position in central carbon metabolism and is an important branch point between catabolic and biosynthetic pathways (Fig. 3). Genes encoding proteins involved in the pyruvate metabolism (such as poxB, aceEF, lpdA, por, and maeA) were transcribed at higher levels at lower O2 concentrations (Fig. 3, Fig. S2, Table S4). Upregulation of glycogen and downregulation of gluconate metabolism at lower O2 concentrations Extracellular glucose can be oxidized to gluconate in the periplasm instead to G6P (Fig. 3). The membrane-bound glucose 1-dehydrogenase (gdh), which oxidizes glucose to glucono-1,5-lactone (Fig. 3), was one of the top upregulated genes (45-fold; log2-fold change of 5.5) and was transcribed at significantly higher levels at 0.1 µmol O2 l−1 (Fig. S1, Table S4). Yet gluconolaconase (gnl) had very low transcript levels at low O2 concentrations (Fig. 3, Table S4). Another alternative route to the glycolytic pathways is glycogen metabolism. The genes that metabolize G6P to glycogen were upregulated in A. capsulatum 161. These glucose polymers can be stored and mobilized upon future demand. We found genes for both, biosynthesis as well as degradation of glycogen, in the genome of A. capsulatum 161 and genes encoding proteins involved in the glycogen metabolism were all upregulated upon the shift of O2 from 10 to 0.1 µmol O2 l−1 (Fig. 3, Fig. S2, Table S4). Differential transcriptional response of TCA cycle genes to O2 concentrations Genes of the TCA cycle responded differently to the decrease of O2 (Fig. 3, Fig. S2, Table S4). The oxidative decarboxylation of acetyl-CoA is one of the main catalytic functions of the TCA cycle and provides reducing equivalents to the respiratory complexes. The first of four oxidative steps is a key rate-limiting step of the TCA cycle and is catalyzed by icd. The multiple homologs of the genes icd exhibited a high transcription level (high TPM levels) but were significantly downregulated at low-nanomolar O2 concentrations (Fig. S2, Table S4). In addition, the TCA cycle flux can be constricted by the availability of the oxaloacetate (OAA). The enzyme, phosphoenolpyruvate carboxykinase (pckA), was transcribed at significantly higher levels (7.5-fold) at 0.1 µmol O2 l−1, making OAA less available for the TCA cycle (Fig. 3). However, the TCA cycle also functions in a biosynthetic capacity, primarily in the synthesis of amino acids, heme, and glucose. Glutamate and aspartate are synthesized from 2-oxoglutarate and OAA, respectively, via transamination, and both genes gdhA as well as aspB, exhibited an increased expression at 0.1 µmol O2 l−1. We detected neither the genes aceA and aceB, encoding the enzymes isocitrate lyase and malate synthase, nor the gene glcB, encoding the malate synthase G, neither in the genome nor in the transcriptome of A. capsulatum 161, which are involved in the glyoxylate shunt of the TCA cycle. In addition to the oxidative conversion of pyruvate into acetyl-CoA, acetyl-CoA can also be synthesized through the catabolism of branched-chain amino acids by the branched-chain alpha-ketoacid dehydrogenase (BCKDC) complex (Fig. 3). The decrease in oxygenation to 0.1 µmol O2 l−1 increased the expression of this BCKDC complex up to 3.2-fold (Fig. 3, Fig. S2, Table S4). Significant transcriptional response of genes involved in production of acetate and ethanol due to decreasing O2 concentrations Genes involved in various pathways for the production of acetate exhibited significantly higher expression levels at lower nanomolar O2 concentrations, such as phosphotransacetylase (pta) (Fig. 3, Fig. S2, Table S4). The conversion of acetate to acetyl-CoA via acetyl-CoA synthetase (acs) was downregulated at lower O2 concentrations (Fig. 3, Fig. S2, Table S4). The peripheral membrane protein pyruvate oxidase PoxB (also sometimes referred to as pyruvate dehydrogenase [ubiquinone]), which oxidatively decarboxylates pyruvate to form acetate and directly coupled to the respiratory chain (Gennis and Hager 1976, Koland et al. 1984, Abdel-Hamid et al. 2001) (Fig. 3), was one of the topmost upregulated genes of the central carbon and energy metabolism with a 45-fold increase in transcription at 0.1 µmol O2 l−1 (Fig. 3, Fig. S2, Table S4). Apart from acetate production, ethanol production was significantly altered by decreasing O2 concentrations, and the homologs adhP, encoding for alcohol dehydrogenases, exhibited higher expression levels, up to 14.9-fold higher, at 0.1 µmol O2 l−1 (Fig. 3, Fig. S2, Table S4). Discussion Adaptations and fast responses to changes in environmental conditions often occur at the metabolic level and in this work, we gained new insights into the transcription response of A. capsulatum 161 to diminishing O2 concentrations at low micro- and nanomolar levels. Our data indicate that diminishing O2 played a pivotal role in regulating the expression of genes involved in central metabolism under C excess conditions (Figs 3 and 4). To counter the toxic accumulation of respiration byproducts building up from the lack of O2, it shifted its metabolism and rerouted fluxes from an energy favorable respiratory state (Fig. 4B) to a respiro-fermentative condition, in which acetate together with ethanol seemed to be major end-products (Figs 3 and 4C). Glucose transport, PP pathways, and pyruvate production were downregulated at low-nanomolar O2 concentrations (Fig. 3)—presumably to reduce the NADH/NAD+ redox ratio, which is a critical regulator of cell metabolism ultimately controlling the onset of respiro-fermentative metabolism (Shen and Atkinson 1970, Szenk et al. 2017). As A. capsulatum 161 transitioned from oxic to nanoxic conditions under C excess, the transcripts of many NADH-generating enzymes related to oxidative respiration were reduced (Fig. 4B). Glucose import (galP) exhibited reduced expression from 10 to 0.1 µmol O2 l−1, presumably as a means to limit the amount of available glucose. Yet, glucose 1-dehydrogenase (gdh) was overexpressed, potentially modulated a great part of the carbon flow through the gluconate bypass, thus reducing glucose concentration in the cell (Fig. 3). However, it appears that the cell did not use gluconate, as gluconolaconase had very low transcript levels at low O2 concentrations (<10 µmol O2 l−1). This could suggest that the enzyme requires a certain oxygen concentration to function. Pyruvate oxidase (PoxB) was upregulated in A. capsulatum 161, which catalyzes the decarboxylation of pyruvate to acetate and CO2 (Figs 3 and 4C), suggesting that pyruvate catabolism is the major switch point between the respiratory and fermentative responses. The glycolytic flux was redirected toward the production of fermentation products, acetate (upregulation of pta and acyP) and ethanol (upregulation of adhP) (Figs 3 and 4C, Fig. S2, Table S4), to prevent carbon intermediates to enter the TCA cycle (El-Mansi and Holms 1989). Cells can then convert acetyl-CoA through the Pta-AckA pathway, producing and excreting acetate while generating ATP (El-Mansi and Holms 1989). Since the flux from acetyl-CoA to acetate does not generate any NADH (while the flux from acetyl-CoA through the TCA cycle generates 8 NAD(P)H and 2 FADH2), carbon flow diversion to acetate could be viewed as a means of A. capsulatum 161 to reduce or prevent further NADH accumulation (El-Mansi and Holms 1989, Holms 2001). This is in congruence with previous work on Staphylococcus aureus, where acetate production was enhanced under low O2 and glucose excess conditions (Ferreira et al. 2013). In addition, the conversion of acetaldehyde to ethanol via adhP was upregulated, consuming NADH and hence a way to counteract the NADH/NAD+ imbalance (Figs 3 and 4C). Taken together, we hypothesize that the concomitant rise in NADH levels from glucose excess and low O2 conditions drove the onset of fermentative metabolism (acetate and ethanol production) to avoid toxic levels of NADH in the cell. Acetate and ethanol production stemming from pyruvate bypasses any energy-conserving steps associated with NADH, allowing a fast oxidation of pyruvate and efficient shuttling of protons/electrons to the ETC. Various studies have shown that in concentrated glucose environments, E. coli and other organisms switch to and obtain some of their energy anaerobically by acetate fermentation, even when O2 is plentiful, if the rate of glucose consumption is greater than the capacity to reoxidize the reduced equivalents generated (Farmer and Jones 1976, Hollywood and Doelle 1976, Andersen and Meyenburg 1980, Meyer et al. 1984, Farmer and Liao 1997, Kayser et al. 2005, Vemuri et al. 2006, Vazquez et al. 2008, Molenaar et al. 2009, Nahku et al. 2010, Valgepea et al. 2011, 2013, Zhuang et al. 2011, Basan et al. 2015, Peebo et al. 2015, Schütze et al. 2020). Although a major role of NADH is to supply electrons to the ETC thereby fueling the production of ATP, the strategy of A. capsulatum 161 was to reduce the NADH production stemming from respiratory pathways to avoid NADH imbalance while generating ATP (Szenk et al. 2017). The use of alternative pathways for NAD+ regeneration concomitant was also reported in other facultative anaerobes such as E. coli (Vemuri et al. 2006, Farhana et al. 2010, Martínez-Gómez et al. 2012, Szenk et al. 2017) and members of the genera Salmonella and Shigella (Gray et al. 1966, Wolfe 2005). The metabolic flexibility would allow these bacteria to cope with varying concentrations of carbon and O2 in such environments like soils. This respiro-fermentative strategy might extend into the Acidobacteriota, as many genomes harbor this potential as evidence by the presence of acetate kinase and alcohol dehydrogenase (Eichorst et al. 2018). Our experimental conditions also invoked a significant upregulation of the glycogen metabolism suggesting cells transform excess glucose to the storage compound glycogen (Fig. 3, Fig. S2, Table S4). The accumulation of glycogen provides a metabolic reserve for A. capsulatum 161 under potential carbon-limited conditions in the future, which could be an important strategy in environments such as soils allowing cells to cope with transient limiting conditions. Stress response to O2 and reactive oxygen species (ROS) is crucial for the ability to exist in habitats that are characterized by fluctuating O2 concentrations. Whether microbes can occupy such a habitat or a microniche within partly depends upon whether they are able to withstand local concentrations of high or low O2. Several universal stress proteins in A. capsulatum 161 were significantly upregulated by decreasing O2 concentrations approaching anoxia; especially two of these usp genes were affected by the drop of O2 to 0.1 µmol O2 l−1 (Fig. 2A, Table S2). Furthermore, the drop of O2 invoked a significant increase of transcription of a sensor histidine kinase (Fig. 2A, Table S2), which presumably allowed A. capsulatum 161 to sense environmental stimuli and manage various environmental changes by coupling environmental cues to gene expression (Stock et al. 2000, Mascher et al. 2006, Kaczmarczyk et al. 2014). Cellular stress can further lead to protein denaturation (Hightower 1991), and proteolytic removal of non‐functional proteins is crucial for optimal metabolic activities (Porankiewicz et al. 1999). We detected an upregulation of the ATP-dependent Clp proteases in the transcriptomic response of A. capsulatum 161 to diminishing O2 concentrations (Fig. 2A, Table S2), suggesting that they are important in removing irreversibly damaged polypeptides that may interfere with metabolic pathways under O2-limited stress conditions. In A. capsulatum 161, a clear differential upregulation of genes involved in counteracting oxidative stress was observed upon the decrease of oxygenation from 10 to 0.1 µmol O2 l−1 (Fig. 2A, Table S2), indicating that it is capable of adapting to different redox states. Oxidative stress defense genes such as manganese superoxide dismutase, thioredoxins, and glutaredoxins were highly expressed under stimulated at low O2 (Fig. 2A, Table S2), as seen previously in Nitrosomonas europaea (Sedlacek et al. 2020). The increased demand for proteins involved in ROS defense could be caused by NADH/NAD+ redox ratio imbalances, as NADH accumulates and becomes toxic. Under O2-limiting conditions, an increased level of NADH builds up, as it is less efficiently reoxidized to NAD+ as a result of reduced aerobic respiration. The high-affinity bd-type oxidase (cydAB) and NADH dehydrogenase (ndh-II) were upregulated upon the drop of oxygenation to 0.1 µmol O2 l−1 (Fig. 2B, Table S3). We previously hypothesized that the upregulation of the bd-type oxidase could suggest a contribution to respiratory activity at trace O2 conditions or favor the more faster electron flux than cbb3-type oxidases to permit more rapid reducing potential from carbon surplus (Trojan et al. 2021). The uncoupled NADH dehydrogenase NDH-II was highly upregulated at 0.1 µmol O2 l−1 (Fig. 2B, Table S3), presumably to compensate for the slow regeneration of NAD+ due to the low O2 availability. NDH-II only catalyzes the oxidation of NADH and reduction of quinones without the ability to pump protons, which, based on our data, seemed to be beneficial under micro- and nanoxic conditions. Alternative electron‐transfer routes seem to allow A. capsulatum 161 adjusting its energy transduction efficiency to its needs and substrate availability. In A. capsulatum 161, electrons can flow from the NADH dehydrogenase and SDH (complex II of the ETC) to the quinone/quinol pool, from where the electrons may either bypass the cytochrome bc1 complex (complex III) and directly flow to the bd-type quinol terminal oxidase or flow via a cytochrome c either to the low-affinity caa3‐type or the high-affinity cbb3‐type cytochrome c terminal oxidase. Under aerobic- or carbon-limiting conditions, the proton-coupled NDH-I might be the main driver of NADH oxidation and maintain the proton motive force required for ATP synthesis (Fig. 4A). Under the reduced environment in our incubations, when the reduction state of the quinone/quinol pool increases, the pathway to the NDH-II and bd-type quinol oxidase seemed to increase, with NDH-II being then the dominant dehydrogenase to oxidize excess NADH and support redox balance (Figs 2B and 4C). This strategy has been observed for E. coli, which may use NDH-II to counteract increasing NADH/NAD+ triggered by faster metabolism due to increased glucose uptake in order to support fast growth (Vemuri et al. 2006, Liu et al. 2019). Conclusions In this study, we examined the transcriptional response of A. capsulatum 161 to diminishing O2 oncentrations in the low nanomolar range. Overall, O2-limiting conditions invoked a significant stress response in A. capsulatum 161. Our data indicate that A. capsulatum 161 has the genomic potential for multiple routes for the early steps of glucose catabolism. Under O2-limited but glucose-unlimited conditions, A. capsulatum 161 reroutes fluxes through its central metabolism from glycolysis to fermentative end products to counteract NADH/NAD+ imbalances building up due to loss of respiratory capacities under electron acceptor-limiting conditions. Understanding these capacities advances the knowledge on the metabolic responses A. capsulatum 161 is capable of in order to successfully thrive and persist under fluctuating substrate availabilities in terrestrial environments. The investigated oxygen range (10 to 0.1 µmol O2 l−1 or 3.6–0.036 pO2% present atmospheric level) is environmentally relevant (Sexstone et al. 1985), presumably seen in various soil niches. Coping with dynamic O2 tensions is therefore vital for aerobic bacteria dwelling in (temporarily) O2-deprived habitats. During “spring snowmelt” or in the rhizosphere, catabolism and ATP yields can be uncoupled due to O2 limitation and carbon availability. To survive reductive stress during O2 deprivation, soil bacteria depend on metabolic strategies to maintain a proton motive force and redox balance. But modifications in metabolic routes at trace O2 levels extend beyond soils; these findings have implications in other environments, such as oxygen minimum zones (OMZs) in the Earth’s oceans. OMZs are large water masses with low oxygen concentrations, thus favoring anaerobic metabolism (Kalvelage et al. 2015). Interestingly, aerobic metabolism was previously detected in regions of apparent anoxic conditions (“anoxic” OMZs) (Garcia-Robledo et al. 2017), along with the presence of terminal oxidases (Kalvelage et al. 2015, Tsementzi et al. 2016) and production of O2 at trace levels (Canfield and Kraft 2022). These regions could provide a niche where bacteria transition from a respiratory to a respiro-fermentative metabolism to maximize energy yield, prior to using less favorable electron acceptors, such as nitrate. Taken together, the transition from aerobic respiration to a respiro-fermentative metabolism could provide bacteria the flexibility to generate energy during periods of limiting O2 in fluctuating environments for maintenance and survival of their populations. Supplementary Material Acknowledgments We thank the Division of Computational Systems Biology for providing and maintaining the Life Science Compute Cluster (LiSC) at the University of Vienna. We are grateful to Lars B. Pedersen at Aarhus University for his technical assistance, Sergey Borisov at Technical University of Graz for the production of optical oxygen sensors, and Dimitri V. Meier for bioinformatics support. Contributor Information Daniela Trojan, Department of Microbiology and Ecosystem Science, Centre for Microbiology and Environmental Systems Science, University of Vienna, 1030 Vienna, Austria. Emilio García-Robledo, Department of Biology, Faculty of Marine and Environmental Sciences, University of Cádiz, 11002 Cádiz, Spain. Bela Hausmann, Department of Microbiology and Ecosystem Science, Centre for Microbiology and Environmental Systems Science, University of Vienna, 1030 Vienna, Austria; Joint Microbiome Facility of the Medical University of Vienna and the University of Vienna, 1030 Vienna, Austria; Department of Laboratory Medicine, Medical University of Vienna, 1030 Vienna, Austria. Niels Peter Revsbech, WATEC, Department of Biology, Aarhus University, 8000 Aarhus, Denmark. Dagmar Woebken, Department of Microbiology and Ecosystem Science, Centre for Microbiology and Environmental Systems Science, University of Vienna, 1030 Vienna, Austria. Stephanie A Eichorst, Department of Microbiology and Ecosystem Science, Centre for Microbiology and Environmental Systems Science, University of Vienna, 1030 Vienna, Austria. Author contributions Daniela Trojan (Conceptualization, Data curation, Formal analysis, Investigation, Methodology, Visualization, Writing – original draft, Writing – review & editing), Emilio García-Robledo (Conceptualization, Data curation, Formal analysis, Investigation, Methodology, Resources, Visualization, Writing – review & editing), Bela Hausmann (Formal analysis, Resources, Software, Writing – review & editing), Niels Peter Revsbech (Methodology, Resources, Writing – review & editing), Dagmar Woebken (Conceptualization, Funding acquisition, Project administration, Resources, Supervision, Writing – review & editing), and Stephanie A. Eichorst (Conceptualization, Data curation, Funding acquisition, Project administration, Resources, Supervision, Writing – review & editing) Conflict of interest The authors declare no competing interests. Funding This research was funded in whole or in part by the Austrian Science Fund FWF (Grant-DOI: 10.55776/P26392; co-principal investigators, D.W. and S.E.), the European Research Council (ERC) Starting Grant 636928 (to D.W.) under the European Union’s Horizon 2020 research and innovation program, the Dr. Anton Oelzelt-Newin'sche Stiftung of the Austrian Academy of Sciences (ÖAW) (to D.W.), and the Spanish Research Agency, project PID2020-117340RA-I00, RYC2019-027675-I (to E.G.R.). For open access purposes, the author has applied a CC BY public copyright license to any author accepted manuscript version arising from this submission. Data availability The genome of A. capsulatum 161 is publicly available in the NCBI database with the following assembly accession number GCA_000022565.1. The raw transcriptomic reads are publicly available under BioProject accession number PRJNA635786. References Associated Data Supplementary Materials Data Availability Statement The genome of A. capsulatum 161 is publicly available in the NCBI database with the following assembly accession number GCA_000022565.1. The raw transcriptomic reads are publicly available under BioProject accession number PRJNA635786.