Complex regulation in a Comamonas platform for diverse aromatic carbon metabolism

Wilkes, Rebecca A.; Waldbauer, Jacob; Carroll, Austin; Nieto-Domínguez, Manuel; Parker, Darren J.; Zhang, Lichun; Guss, Adam M.; Aristilde, Ludmilla

doi:10.1038/s41589-022-01237-7

Download PDF

Article
Open access
Published: 06 February 2023

Complex regulation in a Comamonas platform for diverse aromatic carbon metabolism

Nature Chemical Biology volume 19, pages 651–662 (2023)Cite this article

10k Accesses
10 Citations
345 Altmetric
Metrics details

Subjects

An Author Correction to this article was published on 05 February 2024

This article has been updated

Abstract

Critical to a sustainable energy future are microbial platforms that can process aromatic carbons from the largely untapped reservoir of lignin and plastic feedstocks. Comamonas species present promising bacterial candidates for such platforms because they can use a range of natural and xenobiotic aromatic compounds and often possess innate genetic constraints that avoid competition with sugars. However, the metabolic reactions of these species are underexplored, and the regulatory mechanisms are unknown. Here we identify multilevel regulation in the conversion of lignin-related natural aromatic compounds, 4-hydroxybenzoate and vanillate, and the plastics-related xenobiotic aromatic compound, terephthalate, in Comamonas testosteroni KF-1. Transcription-level regulation controls initial catabolism and cleavage, but metabolite-level thermodynamic regulation governs fluxes in central carbon metabolism. Quantitative ¹³C mapping of tricarboxylic acid cycle and cataplerotic reactions elucidates key carbon routing not evident from enzyme abundance changes. This scheme of transcriptional activation coupled with metabolic fine-tuning challenges outcome predictions during metabolic manipulations.

Heat flows enrich prebiotic building blocks and enhance their reactivity

Article Open access 03 April 2024

Thomas Matreux, Paula Aikkila, … Christof B. Mast

Elucidation of genes enhancing natural product biosynthesis through co-evolution analysis

Article 12 April 2024

Xinran Wang, Ningxin Chen, … Xiaozhou Luo

Electrochemically coupled CH4 and CO2 consumption driven by microbial processes

Article Open access 10 April 2024

Yue Zheng, Huan Wang, … Feng Zhao

Main

Comamonas species are promising cellular chassis for the bioprocessing of aromatic-containing waste streams^1,2,3,4,5,6. Strains of C. testosteroni lack the genes required for carbohydrate utilization and have an innate preference for gluconeogenic substrates such as aromatic compounds^3,7. In particular, C. testosteroni KF-1 was isolated from sewage sludge for its capacity to degrade aromatic synthetic laundry surfactants³. Metabolic regulation in C. testosteroni strains during assimilation of aromatic compounds remains to be elucidated, thus severely limiting the potential exploitation of these wastewater isolates to metabolize xenobiotic aromatic compounds. In this work, we investigate native mechanisms controlling carbon routing in C. testosteroni KF-1 grown on two lignin-associated monomers (4-hydroxybenzoate (4HB) and vanillate (VAN)) and a synthetic polymer-associated monomer (terephthalate (TER)), which are all funneled through protocatechuate (PCA) toward the central carbon metabolism (CCM; Fig. 1a)³.

**Fig. 1: Peripheral metabolic pathways for aromatic compound catabolism.**

PCA is a key intermediate in the catabolism of hydroxybenzoates, methoxybenzoates and phthalates^8,9. There are three known catabolic pathways to initiate ring opening of PCA at the ortho position, the 2,3-meta position or the 4,5-meta position (Fig. 1a)⁹. Enzymes in each pathway were characterized in Pseudomonas putida KT2440 (ortho pathway)¹⁰, Paenibacillus sp. JJ-1B (2,3-meta pathway)¹¹, Sphingobium sp. SYK-6 (4,5-meta pathway)^12,13, and C. testosteroni CNB-1 (4,5-meta pathway)⁵, among others (Fig. 1a)⁵. Catabolism of aromatic compounds can both generate and consume cofactors before carbon influx into the tricarboxylic acid (TCA) cycle. For instance, 4HB hydroxylation into PCA requires reducing power in the form of NADPH or NADH^5,14, VAN demethylation through VanAB (a Rieske domain ring-hydroxylating oxygenase) requires NAD(P)H and produces formaldehyde^3,15, and TER decarboxylation to PCA requires NADPH during oxidation and produces NADPH in a dehydrogenase reaction (Fig. 1a)^16,17. To generate NADPH in CCM, C. testosteroni strains rely on isocitrate dehydrogenase (IDH) in the TCA cycle and malic enzyme (ME) in the cataplerotic direction because these strains lack the genes for a complete oxidative pentose phosphate (PP) pathway^3,18. Transhydrogenase reactions can also interconvert between NADH and NADPH pools to supply cofactors for catabolic reactions or biosynthetic demand¹⁸.

The presence of the complete pmd operon in the genome of C. testosteroni KF-1 implicates the 4,5-meta pathway for catabolism³, but this has not yet been verified experimentally. The 4,5-meta pathway consists of six enzymatic steps to convert PCA into pyruvate and oxaloacetate (OAA) without any carbon loss to CO₂ (Fig. 1a)⁵. In this pathway, one NADPH is generated during the production of 2-pyrone-4,6-dicarboxylate (PDC; Fig. 1a)^19,20. Potential genes for the ortho and 2,3-meta pathways are dispersed throughout the C. testosteroni KF-1 genome, but homologs of pcaG in the ortho pathway and praA and praH in the 2,3-meta pathway were not annotated (Fig. 1a)³. Due to possible multifunctional enzymes compensating for absent genome annotations²¹, the functionality of the ortho and 2,3-meta cleavage pathways remains to be ascertained.

Turning pathways ‘on’ or ‘off’ by regulating gene expression is a recognized strategy in bacteria to promote growth by conserving costs associated with producing unused enzymes^22,23. Importantly, flux directionality through metabolic reactions was determined to be related to negative Gibbs free energy (∆G), which is log proportional to both the concentration ratio of product to reactant and the flux ratio of forward to backward flux^24,25. Enzyme abundance was reported to be a poor predictor of carbon fluxes in P. putida, Bacillus subtilis and Saccharomyces cerevisiae, and, instead, additional regulation through metabolite abundances, post-translational modifications or allosteric regulation was proposed^26,27,28. Metabolite concentrations in Escherichia coli, Clostridium acetobutylicum and C. ljungdahlii were found to drive the thermodynamic potential controlling metabolic flux directionality without altering enzyme abundance^24,29,30. Leveraging orthogonal methods to elucidate the metabolic controls underlying aromatic carbon fluxes in C. testosteroni KF-1, we combined transcriptomics and proteomics measurements with targeted metabolomics data, ¹³C-kinetic profiling, ¹³C-fluxomics analysis, and ¹³C-isotopomer fragmentation analysis during feeding on 4HB, VAN and TER. Additionally, we established a toolkit for genetic manipulation of C. testosteroni KF-1 to validate hypotheses about flux through aromatic carbon cleavage pathways. New insights were obtained on the regulatory mechanisms during aromatic carbon utilization that involve complex relationships between transcript expression, protein abundance, metabolite levels, and metabolic fluxes.

Results

Transcript and enzyme levels regulate aromatic ring cleavage

The C. testosteroni cells grew well on all three aromatic substrates and succinate (a TCA cycle intermediate used as a reference substrate), with less than 30% difference in the growth rates (Fig. 1b). However, growth on 4HB exhibited a 50% greater carbon equivalent uptake rate than growth on the other compounds (P < 0.001; Fig. 1b,c and Extended Data Fig. 1a). The occurrence of metabolite secretions only during growth on 4HB, representing 28% of the 4HB-derived carbon uptake, explained the discrepancy between uptake rate and growth rate (Fig. 1d). This discrepancy implied a surplus of carbon uptake beyond the enzymatic capacity to direct 4HB-derived carbons to biomass. In fact, 60% of the carbon loss from 4HB uptake stemmed from secretions of metabolites (PCA and PDC) from the aromatic catabolism pathways before CCM; the remaining 40% was from pyruvate (a product of the 4,5-meta pathway), α-ketoglutarate (αKG) and malate (both TCA cycle metabolites) and glutamate (an amino acid derived from αKG; Fig. 1d and Extended Data Fig. 1b).

Consistent with the 4,5-meta pathway as the primary catabolic route from PCA, growth on the aromatic substrates triggered a greater than 18-fold upregulation of 4,5-meta pathway transcripts, transcriptional regulators, and proteins relative to growth on succinate; there was either no differential transcript expression or a lack of detectable proteins for the putative ortho and 2,3-meta pathways (Fig. 1e, Extended Data Fig. 2 and Supplementary Tables 4–8). As a confirmation, C. testosteroni KF-1 mutant cells without the 4,5-meta pathway (C. testosteroni ∆pmd, called strain AG9493) were incapable of growing on 4HB, VAN or TER (Extended Data Fig. 3). Gene products involved in both the transport and oxidation of VAN or TER were only detected or were differentially expressed (by up to 1,230-fold) during growth of C. testosteroni KF-1 on the respective substrate (Fig. 1e). However, gene products of CtesDRAFT_PD1893, a PmdK homolog with 98% identity to PmdK in Comamonas sp. E6 and annotated as both a 4HB and PCA transporter⁶, were detected or elevated by 34- to 72-fold during growth on all three aromatic substrates relative to growth on succinate (Fig. 1e and Supplementary Table 9). During single-substrate growth on 4HB or VAN, there was an increase (from 27- to 200-fold) in both the transcript and protein for CtesDRAFT_PD2627 (PobA homolog; responsible for the oxidation of 4HB to PCA; Fig. 1e). In sum, relative to succinate-fed cells, growth on all three aromatic substrates triggered the upregulation of gene products relevant to transport and initial catabolism of each substrate toward the 4,5-meta pathway as the assimilation route into CCM (Fig. 1e).

The transcriptomics data further implied the capacity for 4HB uptake in the presence of other aromatic substrates (Fig. 1e). In subsequent growth experiments of C. testosteroni KF-1 on a mixture of 4HB and VAN, 4HB contributed about 90% to the intracellular PCA pool until late exponential growth (Fig. 1f). During growth on a mixture of all three aromatic compounds, VAN and TER were only co-utilized at a detectable rate after 4HB was depleted (Fig. 1g, time point 3). However, even with their minimal depletion extracellularly during early exponential growth, VAN and TER each contributed about 13% each to the intracellular PCA pool (Fig. 1g). Our data seemed to indicate possible carbon catabolite repression to support preferential uptake of 4HB before VAN or TER, but the lack of a complete shutdown of the initial catabolic pathways for both VAN and TER pointed instead to a disproportionate rate utilization of the different substrates. Uptake regulation of substrate mixtures would be an important research avenue to address moving forward with C. testosteroni as a performance strain. Here, toward building the requisite foundation for future metabolic manipulations of C. testosteroni, we focus on gaining fundamental mechanistic insight on aromatic carbon assimilation into CCM.

Aromatic carbon flux into the TCA cycle is regulated at the OAA node

After switching cells from ¹³C-succinate to unlabeled aromatic substrates (or unlabeled succinate as a reference), we tracked the kinetic incorporation of the non-labeled fraction (M + 0) over 120 s (Fig. 2a). A rapid increase in non-labeled pyruvate and malate (up to 70–77% within 10 s) indicated minimal delay of aromatic carbon incorporation despite initial feeding on only succinate (Fig. 2a). Slow incorporation (only 2% after 10 s) into succinate confirmed an inactive ortho cleavage pathway (Fig. 2a). Aromatic carbon influx into OAA was retained initially in the TCA cycle due to impeded gluconeogenic flux, as revealed by a low fraction of non-labeled aspartate (an amino acid directly synthesized from OAA; only 10%) and phosphoenolpyruvate (PEP; less than 1%) after 120 s (Fig. 2a). Despite the lack of gluconeogenic flux into the Embden-Meyerhof-Parnas (EMP) and PP pathway metabolites immediately after the substrate switch, there were less than twofold changes in the abundance of gene products in these pathways in cells fed solely on the aromatic substrates relative to succinate-fed cells (Fig. 2b). Enzymes in the Entner–Doudoroff (ED) pathway were not detected here during aromatic substrate catabolism, consistent with a previous report of an active ED pathway during glycolytic growth but not during gluconeogenic growth¹⁸. Except for pyruvate (a product of the 4,5-meta pathway), pools of metabolites in the EMP and PP pathways remained relatively unchanged between growth conditions (Fig. 2b). While fluxes in the TCA cycle for aromatic-derived carbons were not substantially altered relative to growth on succinate, there were significant changes in transcripts, proteins and metabolites. During growth on 4HB compared to succinate, there was 2.5-fold higher protein abundance of both AcnA and FumC, whereas there was 2.5-fold lower protein abundance of the cataplerotic enzymes PckG and ME-2, all of which would facilitate carbon retention in the TCA cycle (Fig. 2b). Additionally, 4HB-grown cells had higher metabolite levels than succinate-grown cells, with 17.5-fold higher αKG, 11-fold higher pyruvate and nearly 2.5-fold higher citrate levels (Fig. 2b). By contrast, growth on VAN exhibited depletion in both proteins and metabolites in the TCA cycle relative to growth on succinate, including undetectable proteins (IDH2, ME-2, Ppc, SdhA and SdhD), 3.3-fold lower protein abundance (SdhB, SucC and SucD), a 2.7-fold lower αKG pool, and an undetectable pyruvate pool (Fig. 2b). However, during growth on TER, the general trend was a lack of appreciable differences in transcripts, proteins or metabolites relative to growth on succinate (Fig. 2b). Despite these substrate-specific differences, the similar kinetic isotopic flux profiling of the TCA cycle implied an underlying regulation that maintained consistent fluxes in a robust metabolic network for C. testosteroni KF-1 (Fig. 2a).

**Fig. 2: Isotope kinetics, transcriptomics, proteomics, and metabolomics analysis of CCM during assimilation of aromatic compounds.**

Across growth on the three different aromatic substrates relative to growth on succinate, there were differential changes in the expression of gene products that control the pool of OAA, which can act as an enzymatic inhibitor (Fig. 2b)^31,32,33. Lack of detectable levels of Pyc protein may mean a limited carboxylation reaction flux from pyruvate to OAA (Fig. 2b). Also, a specific OAA decarboxylase was not found annotated in the C. testosteroni KF-1 genome, but PmdF (the last step in the 4,5-meta pathway) has previously been shown to decarboxylate OAA into pyruvate in vitro⁵. At the node between OAA and malate, Mqo was downregulated 50-fold at the transcript level and was not detected at the protein level during growth on the aromatic substrates relative to succinate; but, there was an increase (albeit less than twofold) in malate dehydrogenase (Mdh) across both gene products (Fig. 2b). This different regulation of Mqo and Mdh in C. testosteroni would support reductive flux to modulate the OAA pool based on previous works with E. coli and Corynebacterium glutamicum^33,34 that reported Mqo operating in the direction of malate oxidation to OAA and Mdh favoring reduction of OAA to malate (Fig. 2b). Collectively, during feeding on the aromatic substrates, we found a lack of transcriptional or translational regulation for TCA cycle fluxes but gene product regulation of the OAA pool, which would in turn modulate the flux of the aromatic substrate-derived carbons into CCM (Fig. 2a,b). Using 4HB-grown cells versus succinate-grown cells, we further examined the relationships between transcripts, proteins, metabolites, and fluxes.

Central carbon fluxes are driven by metabolite pools

Due to the replenishment of the OAA pool by the 4,5-meta pathway in aromatic carbon catabolism, contribution of malate carbons to OAA as expected in a canonical TCA cycle was not required (Fig. 3a). Instead, ¹³C mapping revealed that about 40% of fumarate was derived from OAA and 60% was derived from succinate, indicating flux in the reductive direction (Fig. 3b and Supplementary Fig. 2). Near-zero negative values of ∆G for the reactions between succinate and OAA demonstrated that these reactions were close to equilibrium (Fig. 3c). For cataplerotic reactions, positional ¹³C mapping showed that pyruvate was the primary source (approximately 82%) of PEP; OAA contributed only 18% (Fig. 3d). Flux to pyruvate was also exclusively from malate and not OAA (Figs. 3e and 4a), which further indicated that PmdF was not involved in cataplerosis. Thus, the main route for cataplerosis was through ME, which was further supported by the nearly tenfold greater carbon flux from malate to pyruvate during 4HB metabolism than during succinate metabolism (Fig. 4a), despite only a modest driving force from malate to pyruvate (∆G = −3.95 ± 1.44 kJ mol^–1) in 4HB-grown cells (Fig. 3c).

**Fig. 3: Carbon mapping and free energy elucidate metabolic routing.**

**Fig. 4: Metabolic flux distributions depict energy and reducing power yields.**

In the clockwise canonical TCA cycle, flux from OAA to succinate increased 4-fold, but flux from succinate to malate decreased by 1.6-fold, and flux from malate to OAA decreased by 12.5-fold (Fig. 4a). This substantial decrease in the flux between malate and OAA agreed with the absence of detectable levels of Mqo during growth on 4HB (Figs. 2b and 4a). Consistent with undetected or less than twofold changes in the transcripts and proteins of AceA and GlcB, only a small flux (approximately 3% of 4HB uptake) was through the glyoxylate shunt (Fig. 4a). Likewise, there were less than 50% changes in the relative flux in the EMP and PP pathways between 4HB-grown cells and succinate-grown cells, in agreement with the lack of changes in the amounts of transcripts, enzymes and metabolites in these pathways (Figs. 2b and 4a). However, the observed differences in the TCA cycle and cataplerotic reactions implied substrate-specific mechanisms for converting carbon flux into energy generation (Figs. 2b and 4a,b).

Indeed, we determined different pathways involved in maintaining cofactor balance in response to feeding on 4HB versus succinate (Fig. 4b,c). Comparison of the protein sequences to enzymes with known cofactor specificity^14,35 suggested that NADPH was the likely cofactor for PobA and PmdC in C. testosteroni KF-1 (Supplementary Tables 8 and 16). Subsequent enzymatic assays of purified proteins from C. testosteroni KF-1 demonstrated that NADPH was the cofactor for PobA and for ME-1, IDH1 and IDH2, but NADH was the preferred cofactor for ME-2 and Mdh (Fig. 4b). Thus, a net-zero balance in NADPH was calculated from the uptake and channeling of 4HB-derived carbons through the 4,5-meta pathway, assuming that PmdC is exclusively NADPH dependent, as confirmed experimentally for PobA (Fig. 4c). The high carbon flux through IDH and ME-1 (ME-2 was not detected during growth on 4HB) downstream of the 4,5-meta pathway yielded NADPH in excess of the NADPH requirement for anabolism during growth on 4HB (Fig. 4a–c). Therefore, in contrast to succinate-grown cells that required transhydrogenase-dependent electron transfer to produce NADPH from the NADH pool, 4HB-grown cells instead may use electron transfer from excess NADPH to generate NADH in support of ATP production via oxidative phosphorylation (Fig. 4b), albeit the only transhydrogenase (PntAB) annotated in the C. testosteroni KF-1 genome was not differentially expressed (Supplementary Table 10). Despite a similar ATP requirement for anabolism, the flux analysis for 4HB-grown cells determined a 20% greater ATP surplus potentially available for cellular maintenance, which was corroborated by a 50% greater pool of measured intracellular ATP, presumably due to the relatively higher rate of oxidative phosphorylation and substrate-level phosphorylation than observed during growth on succinate (Fig. 4c,d).

Metabolic flux regulation was further examined through correlation relationships between transcriptomics, proteomics, metabolomics, and fluxomics (Fig. 5a). There was a strong positive linear relationship between the relative levels of RNA and protein (Pearson correlation, r = 0.815, n = 31, P < 0.001; Fig. 5a). The two outliers from this relationship were FumC and AcnA, which were 2.5-fold higher at the protein level during 4HB feeding than during succinate feeding despite the lack of differential expression of the genes, implying potential additional regulation of translation or protein turnover (Figs. 2b and 5a). Notably, there was no correlation between the relative metabolic fluxes and either relative transcript levels (Pearson correlation, r = 0.276, n = 29, P = 0.147) or relative protein abundance (Pearson correlation, r = 0.187, n = 27, P = 0.351), primarily due to nodes in the TCA cycle and cataplerosis (Fig. 5a). Excluding the reaction through ME, there was a strong agreement (r = 0.960, n = 5, P = 0.009) between the central carbon fluxes and the substrate-to-product ratio of metabolite pools (Fig. 5a). These correlation analyses thus revealed that metabolite pools were strongly correlated to TCA cycle fluxes, which ultimately controlled the overall cellular redox balance and energy metabolism in C. testosteroni.

**Fig. 5: Correlation of omics levels indicates multilevel regulation.**

Allosteric regulation influences key metabolic reactions

High flux from malate to pyruvate during growth on 4HB could not be explained at the transcript, protein or metabolite levels. In E. coli, favorable operation of ME in the forward decarboxylation direction was found to be regulated positively by succinate, fumarate, aspartate, and glutamate³⁶. Here, both ME-1 and ME-2 from C. testosteroni KF-1 exhibited strong activation (up to 200% and 1,100% increases in activity, respectively) in the presence of fumarate and aspartate and moderate activation in the presence of succinate and glutamate (up to a 160% increase in activity); only high concentrations of αKG had a significant inhibitory impact on ME-2 (Fig. 5b). Flux through PEP synthetase (PEPs) appeared to be stalled during growth on the aromatic substrates despite the relatively high pool of pyruvate and ATP compared to growth on succinate (Figs. 2a,b and 4d). In the presence of αKG, the utilization of pyruvate by PEPs was decreased by up to 35% (Fig. 5c). Thus, these findings implied that the high flux through ME could be facilitated by fumarate and aspartate levels (and, moderately, by succinate and glutamate), whereas the flux through PEPs could be impeded by the αKG pool.

Resolution of identified bottlenecks during 4HB catabolism

Due to the 50% greater carbon uptake rate during growth on 4HB than during growth on VAN or TER (Fig. 1c), we investigated the associated in vivo cellular kinetic response using ¹³C-labeled 4HB (Fig. 6a). A slowed ¹³C incorporation into PCA (25% lower at 15 s than 4HB labeling kinetics) was consistent with PCA buildup, which led to PCA secretion at 0.49 ± 0.06 mmol per cell dry weight in grams per hour (mmol g_CDW⁻¹ h⁻¹); PDC was secreted at a similar rate (0.46 ± 0.16 mmol g_CDW⁻¹ h⁻¹; Fig. 1d). These data thus reflected backlogs at both PmdAB and PmdD during initial catabolism of 4HB (Fig. 1a). Transcripts and proteins in the 4,5-meta pathway were 1.2- to 4-fold lower during growth on 4HB than during growth on VAN or TER (Fig. 1e and Supplementary Fig. 6), suggesting that catabolism of 4HB-derived carbons was limited by gene expression.

**Fig. 6: Kinetics of 4HB assimilation highlight bottlenecks in metabolism and targets for metabolic engineering.**

To resolve these bottlenecks, we engineered three strains that overexpressed PmdAB (strain AG9402), PmdD (strain AG9404), or the entire 4,5-meta operon (strain AG9480; Fig. 6b). During growth on 4HB, AG9480 grew comparably to the wild type, whereas both AG9402 and AG9404 exhibited slower growth by about 30% (Fig. 6c and Supplementary Fig. 7). Regarding the bottleneck at PCA, there was 83% and 53% lower secretion of PCA with AG9402 and AG9480, respectively, indicating that increased expression of PmdAB facilitated higher flux through this node (Fig. 6d). Interestingly, the resolution of the bottleneck at PmdAB for strain AG9402 was accompanied by both a 60% faster consumption of 4HB and a near 20-fold increase in PDC secretion rate relative to the wild type, highlighting that targeting PmdAB overexpression increased 4HB uptake flux through the 4,5-meta pathway but amplified the bottleneck at PmdD (Fig. 6d). Targeting the bottleneck at PmdD by overexpression did not appreciably alter PDC secretion (Fig. 6c,d). However, overexpression of the entire operon alleviated the bottleneck at PCA while maintaining a comparable PDC secretion to the wild type (Fig. 6d). Therefore, the bottleneck at PmdAB can be overcome at the transcriptional level, whereas additional regulation likely modulates flux through PmdD.

Despite a pyruvate secretion rate of 0.62 ± 0.13 mmol g_CDW⁻¹ h⁻¹, the pyruvate pool in 4HB-grown cells was the highest of the quantified intracellular metabolites (close to fourfold greater than αKG, the second highest metabolite pool; Fig. 6a), suggesting limited metabolic usage of pyruvate (Fig. 6a). This bottleneck at pyruvate was further exacerbated in AG9402, AG9404 and AG9480, which had over 500%, 200% and 50% increases in pyruvate secretion, respectively (Fig. 6d). These data thus revealed CCM metabolite buildup as a consequence of addressing the bottlenecks in the the 4,5-meta pathway. Despite similar ¹³C-labeling kinetics for pyruvate and PEP (reaching a plateau at about 75%), an 11-fold higher pyruvate pool than PEP pool further indicated a bottleneck in the production of PEP from pyruvate (Fig. 6a), as previously discussed (Fig. 2a). A twofold higher fraction of doubly ¹³C-labeled PEP than doubly ¹³C-labeled pyruvate confirmed the contribution of OAA decarboxylation to the PEP pool, as illustrated in Figs. 3d and 4a. The pyruvate pool was 148-fold greater than acetyl-coenzyme A (acetyl-CoA), which would support a strong thermodynamically favorable drive from pyruvate to acetyl-CoA, as demonstrated by the rapid ¹³C labeling of acetyl-CoA in the 4HB-fed cells (Fig. 5). However, comparing 4HB-grown cells to succinate-grown cells, a 12-fold greater pyruvate-to-acetyl-CoA ratio coupled with only a minimal (13%) increase in the PDH enzyme implied that pyruvate conversion to acetyl-CoA was potentially limited by enzyme abundance or the pool of free CoA (Fig. 2b and Supplementary Table 19).

The metabolite OAA represents an important partitioning node between aromatic carbon-derived influx from the 4,5-meta pathway, flux to the oxidative side of the TCA cycle, reductive flux to malate and cataplerotic reactions from the TCA cycle (Fig. 3a). Consistent with faster OAA reduction to malate than OAA decarboxylation to pyruvate, there was a higher rate of triply ¹³C-labeled malate (30% after 15 s) than doubly ¹³C-labeled pyruvate (less than 1% after 15 s; Fig. 6a). Furthermore, ¹³C metabolic flux analysis (MFA) indicated minimal flux from OAA to PEP (5.6%; Fig. 4a). Therefore, the OAA pool was primarily partitioned between the oxidative TCA cycle and the reductive flux to succinate. Despite comparable citrate and malate pools in 4HB-grown cells, ¹³C incorporation of OAA-derived carbons was nearly 30% greater into malate (following the M + 3 fraction) than into citrate (following the M + 5 fraction) after 15 s, indicating a higher initial flux to malate than to the oxidative TCA cycle (Fig. 6a). Consistent with this impaired carbon flux in the oxidative direction from citrate to succinate, 50% of citrate was fully ¹³C labeled, but the fully ¹³C-labeled fraction of αKG and succinate was less than 20% after 1 min (Fig. 6a). Moreover, the αKG pool was 3.2-fold greater than the succinate pool during growth on 4HB, leading to secretion of αKG (Figs. 1d and 6a). Collectively, kinetic isotopic flux profiling coupled with metabolite quantitation during aromatic carbon metabolism provided evidence of an incomplete bifurcation of the TCA cycle that culminates at succinate.

Discussion

Biological upcycling or remediation of recalcitrant aromatic compounds from waste streams warrants systematic investigations of metabolism across model and non-model organisms^{20,37,38,39,40,41}. Here, we showed that the wastewater isolate C. testosteroni KF-1 grew well on three aromatic compounds (4HB, VAN and TER) related to lignin and plastic derivatives, funneled substrate carbons through PCA into the 4,5-meta pathway, produced NADPH in excess of anabolic demand, and generated an ATP surplus. High 4HB uptake was accompanied by secretions of bioproduction value, including PDC, αKG, glutamate, malate, and pyruvate^20,42,43,44. These inherent characteristics highlight the potential of C. testosteroni as an emerging cellular chassis, but it has remained a relatively untapped organism compared to model species such as E. coli or P. putida¹. As a requisite to exploiting C. testosteroni and other non-model species, comprehensive multiomics studies of their cellular physiology are necessary³⁷.

Discrepancies between changes in gene products and changes in metabolic fluxes were reported in S. cerevisiae, Pseudomonas species, and Bacillus species^{26,27,28,45,46}. Here, relative flux changes in the TCA cycle could not be explained by changes in either transcript or protein abundance in C. testosteroni, in contrast to transcriptional regulation of TCA cycle flux reported for E. coli⁴⁷ or B. subtilis fed on non-aromatic substrates²⁸. The only exception was the downregulation of Mqo versus upregulation of Mdh in aromatic substrate-fed versus succinate-fed C. testosteroni to favor reductive flux from OAA, in accordance with previously reported differential roles of Mdh versus Mqo in E. coli³³. However, the remaining fluxes in the TCA cycle were correlated positively with the substrate-to-product ratio of metabolite pools, highlighting that directionality of reversible reactions relied on mass action principles, as previously demonstrated in S. cerevisiae²⁷. Lastly, we concluded that the constrained flux through PEPs could be facilitated by our confirmed allosteric inhibition by αKG^31,32.

In contrast to the TCA cycle fluxes, cataplerotic fluxes that partition carbon between the TCA cycle and gluconeogenesis were not controlled by the substrate-to-product ratios. Our quantitative ¹³C flux analysis and ¹³C mapping determined low flux from OAA to PEP, no flux between OAA and pyruvate, and high flux from malate to pyruvate. The downregulation of PckG during growth on 4HB indicated that transcript-level regulation moderated flux from OAA to PEP. Despite reported in vitro decarboxylation of OAA to pyruvate by PmdF from C. testosteroni CNB-1 (ref. ⁵), the in vivo absence of this activity in C. testosteroni KF-1 suggested either a lack of decarboxylase activity for CtesDRAFT_PD1895 or a different metabolic control. Comparing aromatic substrate-fed cells to succinate-fed cells, the relative ratio of the malate to pyruvate pools was incongruent with the high flux through ME. Further, there was a discrepancy between flux changes and changes in gene product abundance for ME. In agreement with previous work with E. coli³⁶, we found that fumarate, succinate, aspartate, and glutamate increased the activity of ME-2 and ME-1 from C. testosteroni. Consistent with our metabolome data and metabolic flux modeling, we concluded that the flux through ME in C. testosteroni may be facilitated by the reductive flux from OAA to fumarate.

Rational strain engineering requires a fundamental understanding of the regulatory mechanisms responsible for substrate catabolism, carbon fluxes in CCM, and energy metabolism³⁷. A multiomics analysis of several E. coli strains identified those best suited for producing amino acids and non-native compounds based on strain-specific regulation of metabolic pathways⁴⁸. Similarly, in B. subtilis, identification of altered fluxes due to genome reduction or glutamate presence revealed that high PP pathway flux promoted the production of cellulase⁴⁹. Relative to the other aromatic catabolic pathways, expression of an exogenous 4,5-meta pathway in P. putida resulted in the highest production of pyruvate⁵⁰. Here, overexpression of genes in the 4,5-meta pathway of C. testosteroni alleviated the intracellular buildup and secretion of PCA by mitigating the bottleneck at PmdAB. However, the bottleneck at PmdD was not relieved by overexpression, suggesting that post-translational modification or allosteric regulation may influence this enzyme. Alternatively, gene knockouts could be paired with targeted overexpression of PmdAB to promote secretion of value-added products such as PDC, but metabolic manipulation may result in the imbalance of cellular cofactor production. Thus, rational engineering will need to consider how to sustain the cofactor supply, for instance by maintaining or increasing fluxes through IDH and ME, both of which were identified here as critical to sustaining cofactor production.

In sum, we deciphered the following complex regulatory mechanisms in C. testosteroni KF-1 grown on aromatic compounds relevant to lignin and plastic feedstocks: transcriptional regulation for aromatic carbon catabolism before CCM, metabolite-level regulation through primarily mass action principles or coupled with protein abundances for the TCA cycle fluxes, and protein-level or allosteric regulation for the cataplerotic reactions partitioning TCA cycle fluxes to gluconeogenesis. These new multiomics perspectives combined with targeted ¹³C mapping of specific metabolic nodes present a framework of guiding principles to exploit metabolic reactions for aromatic carbon catabolism in C. testosteroni and other bacterial specialists of biotechnological relevance.

Methods

Culturing conditions

Cells of C. testosteroni KF-1 (DSMZ 14576) were obtained from Deutsche Sammlung für Mikroorganismen und Zellkulturen. Stocks of cells were made in nutrient-rich lysogeny broth (LB) and 25% glycerol for storage at −80 °C. Growth experiments of C. testosteroni KF-1 were performed in an incubator (model I24, New Brunswick Scientific) set at 30 °C with shaking at 220 r.p.m. All growth experiments were conducted in three replicates from the same stock culture. The minimal-nutrient medium contained the major nutrient salts (5.0 mM NaH₂PO₄, 20 mM K₂HPO₄, 37 mM NH₄Cl, 17 mM NaCl, 0.81 mM MgSO₄·7H₂O and 34 μM CaCl₂·2H₂O) and essential trace metal nutrients (13 μmol liter^–1 CuSO₄·5H₂O, 0.49 μmol liter^–1 H₃BO₃, 35 μmol liter^–1 ZnSO₄·5H₂O, 2.9 μmol liter^–1 MnSO₄·5H₂O, 0.11 μmol liter^–1 NiCl₂·5H₂O, 0.6 μmol liter^–1 Na₂MoO₄·5H₂O and 30 μmol liter^–1 FeSO₄·7H₂O), as previously reported¹⁸. The carbon source in the filter-sterilized (Waters Corporation; 0.22-μm nylon filters) and pH-adjusted (7.0) minimal-nutrient medium was 100 mM C 4HB (14.3 mM), TER (12.5 mM), VAN (12.5 mM) or succinate (25 mM). For growth on mixtures, the total substrate concentration was kept at 100 mM C in the 1:1 4HB:VAN mixture and 1:1:1 in the 4HB:VAN:TER mixture. Chemicals for standards and growth medium were purchased from Millipore-Sigma, Fisher Scientific or Cayman Chemical. All experiments were conducted with two transfers into experimental growth conditions to acclimate cells. The final transfer for experimental growth was into 125-ml or 250-ml baffled flasks at one-fifth volume to promote oxygen exchange.

Physiological characterizations

The optical density at 600 nm (OD₆₀₀) on an Agilent Cary UV-visible spectrophotometer was used to monitor cell growth until late stationary phase. Cell suspensions were diluted when the OD₆₀₀ reading was above 0.4 to maintain accurate measurements. To determine substrate consumption rates, metabolite secretion rates and g_CDW, cell suspension samples of three biological replicates were collected periodically throughout growth and pelleted with 5 min of centrifugation at 9,800 g and 4 °C. The supernatant was removed, filtered (Costar Spin-X, 0.22-μm-pore-size filter) and stored at −20 °C until liquid chromatography–mass spectrometry (LC–MS) or ultra-high-performance liquid chromatography (UHPLC) analysis to determine substrate depletion and metabolite secretion profiles. The pelleted cells were lyophilized for g_CDW analysis as previously described¹⁸. Substrate consumption rates (mmol g_CDW⁻¹ h⁻¹) and secretion rates (mmol g_CDW⁻¹ h⁻¹) were determined from the rates of disappearance and appearance of metabolites, respectively, in the medium using linear fit regression analysis and accounting for growth rate and the g_CDW. Metabolites released with increasing trends during exponential growth were considered secretions. We did not attempt to distinguish between active and passive release of metabolites but assumed that cell damage or lysis contributed minimally to these secretions, as determined previously⁵¹. Conversion factors between OD₆₀₀ and g_CDW were calculated for C. testosteroni KF-1 grown on 4HB (0.26 g_CDW liter⁻¹ OD₆₀₀⁻¹), VAN (0.42 g_CDW liter⁻¹ OD₆₀₀⁻¹), TER (0.47 g_CDW liter⁻¹ OD₆₀₀⁻¹), the 1:1 mixture of 4HB and VAN (0.25 g_CDW liter⁻¹ OD₆₀₀⁻¹) and the 1:1:1 mixture of all three aromatic compounds (0.30 g_CDW liter⁻¹ OD₆₀₀⁻¹). For the overexpression strains grown on 4HB, the conversion factors were 0.29 g_CDW liter⁻¹ OD₆₀₀⁻¹, 0.26 g_CDW liter⁻¹ OD₆₀₀⁻¹ and 0.25 g_CDW liter⁻¹ OD₆₀₀⁻¹ for AG9402, AG9404 and AG9480, respectively.

Transcriptomics analysis

Cells were grown in three replicates on minimal-nutrient medium with 100 mM C succinate, 4HB, VAN or TER as sole carbon sources. At an OD₆₀₀ of about 1.0 (mid-exponential growth), an aliquot of cell culture (500 μl) was added to a double volume (1 ml) of RNAprotect bacteria reagent (Qiagen). Samples were pelleted by centrifugation at 5,000 g for 10 min and stored overnight at −80 °C before total RNA extraction. Enzymatic lysis of the bacteria cells was conducted using lysozyme prepared in molecular-grade TE buffer, and extractions were conducted using an RNeasy mini kit (Qiagen). Samples were sent to the NUSeq Core Facility (Northwestern University) for TruSeq total RNA-sequencing library prep and Illumina HiSeq sequencing (50-base pair (bp), single-end reads). Reads were mapped to the C. testosteroni KF-1 genome (ASM16885v1) obtained from NCBI³ using Kallisto version 0.46.0 (ref. ⁵²). Transcript abundance and differential gene expression were calculated with Voom and Limma using Degust⁵³.

Proteomic analysis

Cells of C. testosteroni KF-1 (15 ml) were collected in quadruplicate at an OD₆₀₀ of approximately 1.0 during growth on minimal-nutrient medium with 100 mM C 4HB, VAN or TER as sole carbon sources. Cell pellets were resuspended in a reducing and denaturing buffer (SDS (1%)/Tris (200 mM, pH 8.0)/DTT (10 mM)) and extracted by vortexing and heating (95 °C, 20 min), and cysteine thiols were alkylated with 40 mM iodoacetamide. A modified enhanced filter-aided sample preparation protocol⁵⁴ with a Sartorius Vivacon 500 concentrator (30-kDa nominal cutoff) were used to purify proteins before overnight digestion with MS-grade trypsin at 37 °C. Peptides eluted from the concentrator were dried by vacuum centrifugation and isotopically labeled at both N and C termini using the diDO-IPTL methodology for quantitative analysis, as described previously⁵⁵. In brief, N termini were labeled with either undeuterated or dideuterated formaldehyde, whereas C termini were labeled with either ¹⁶O or ¹⁸O (ref. ⁵⁵). An internal standard for quantification was made from pooled peptide aliquots of each sample labeled with formaldehyde/¹⁸O. For LC–MS analysis, dideuterated formaldehyde/¹⁶O-labeled sample peptides and formaldehyde/¹⁸O-labeled internal standard were mixed 1:1 (vol/vol). Quantification of peptides was conducted on a Dionex UltiMate 3000 LC system with nanoelectrospray ionization (Proxeon Nanospray Flex source) coupled to an Orbitrap Elite mass spectrometer (Thermo Fisher Scientific) operating in a data-dependent acquisition mode, as described previously⁵⁵. In brief, chromatographic separation occurred on a monolithic capillary C18 column (GL Sciences Monocap Ultra, 100-µm inner diameter × 200-cm length) using a water–acetonitrile and 0.1% formic acid gradient, and mass spectra were collected with high-resolution (120,000 m/∆m) MS1 parent ion full scan with fragment ion scans of selected precursors. Analysis of proteomic mass spectra was conducted using MorpheusFromAnotherPlace⁵⁵. For MorpheusFromAnotherPlace analysis, the mass tolerance of precursor and product ions was set at 20 ppm and 0.6 Da, respectively, and modifications included static cysteine carbamidomethylation and variable methionine oxidation, N-terminal (d4)-dimethylation and C-terminal ¹⁸O₂. Target–decoy searching was used to control the false discovery rate for peptide spectrum matches to <0.5%. Protein-level relative abundances and standard errors were calculated in R using the Arm postprocessing scripts for diDO-IPTL data⁵⁵ (github.com/waldbauerlab).

Intracellular metabolite extraction and isotope experiments

Intracellular metabolites were extracted from three biological replicates during exponential growth on unlabeled or ¹³C-labeled growth medium by filtering and lysing cells as described previously¹⁸. In brief, cells were collected via filtration and rapidly quenched with 2:2:1 cold (−20 °C) methanol:acetonitrile:water. After centrifugation, the liquid suspensions of metabolites were dried under N₂ gas and stored at −20 °C until LC–MS analysis. Metabolites from unlabeled intracellular extracts were quantified using commercial standards on the software Thermo Scientific Xcalibur 3.0 Quan Browser with an R² coefficient of 0.990 or higher for the calibration curve. For long-term isotopic enrichment experiments, cells were washed with minimal-nutrient medium between transfers into ¹³C-labeled medium to remove unlabeled extracellular compounds. The ¹³C-labeled substrates used in this study were [U-¹³C₄]-succinate, [1-¹³C]-(carboxyl)-4HB, [¹³C₆]-(phenyl)-VAN and [1,4-¹³⁻C₂]-aspartic acid acquired from Cambridge Isotopes and [¹³C₆]-(phenyl)-4HB obtained from Millipore-Sigma. To determine substrate incorporation into PCA, cells were grown on [1-¹³C]-(carboxyl)-4HB and unlabeled VAN for the mixture of two aromatic compounds and [1-¹³C]-(carboxyl)-4HB, [¹³C₆]-(phenyl)-VAN and unlabeled TER for the mixture of all three aromatic compounds. After intracellular metabolites were identified and examined by LC–MS, isotopolog fractions were analyzed using Metabolomic Analysis and Visualization Engine software version 2011.6.17 (ref. ⁵⁶). All isotopolog data were corrected for the natural abundance of ¹³C using IsoCor v2 (ref. ⁵⁷).

Isotope and carbon switch kinetics

Cells were grown on two transfers of [U-¹³C]-succinate to obtain ¹³C-enriched intracellular metabolites. At mid-exponential growth (OD₆₀₀ of about 0.8–1.0) of the second transfer on [U-¹³C]-succinate, 1.5 ml of ¹³C-labeled cells was filtered on to a 0.22-μm filter, rinsed with an equal volume of sterile minimal-nutrient medium and transferred to an agar plate containing unlabeled succinate, 4HB, VAN or TER. Each condition was conducted in triplicate on separate days with different initial stocks of C. testosteroni KF-1 cells (independent biological replicates). Cells adhered to filters and exposed to unlabeled plates were rapidly quenched in 1 ml of cold 2:2:1 methanol:acetonitrile:water after 10 s, 30 s, 1 min, 2 min, 5 min, 15 min and 30 min. Time 0 was quenched immediately after rinsing and before transfer to a plate. After quenching, lysed cells were scraped off of filters, and the solution was centrifuged to remove cell debris and dried under N₂ as described previously¹⁸. Incorporation of unlabeled substrate was tracked with LC–MS and analyzed as described above. No adaptive time was required after the substrate switch to the unlabeled substrate to detect incorporation of non-labeled fractions in central carbon metabolites.

Single carbon source kinetics

Cells were grown in minimal-nutrient medium containing unlabeled 4HB for two transfers to acclimate the cells. During early exponential growth (OD₆₀₀ of approximately 0.4–6.0) of the second transfer, an aliquot of the unlabeled cells (3 ml) was filtered, rinsed and transferred onto minimal-nutrient agar plates containing unlabeled 4HB. After cells doubled on the plates, the filter with adhered cells was transferred to agar plates containing 100% [¹³C₆]-(phenyl)-4HB. At specific time points on the ¹³C-labeled plates (15 s, 1 min, 5 min, 15 min and 30 min), the cells were rapidly quenched using the aforementioned solvent solution, and the intracellular metabolites were extracted and analyzed as described above.

Instrumental analysis of metabolites

LC–MS analysis was conducted on an UHPLC (Thermo Fisher Scientific Dionex UltiMate 3000) coupled to a high-resolution/accurate mass MS (Thermo Fisher Scientific Q Exactive quadrupole-Oribitrap hybrid MS) with electrospray ionization. Metabolomics analysis was conducted on the intracellular and extracellular extracts as described previously¹⁸. In brief, metabolites were analyzed by reversed-phase ion-pairing LC using an Acquity UPLC Waters 1.7-µm particle size column (2.1 × 100 mm) at a constant column temperature of 25 °C and an injection volume of 10 µl. The mass spectrometer was operated in full-scan negative mode. Metabolite identification was based on accurate mass and matches with standard retention time. For positional ¹³C-labeling analysis, the same LC–MS parameters were used with the addition of data-dependent tandem MS to obtain product ions and map ¹³C-carbons using previously outlined fragmentation pathways⁵⁸. The following data-dependent tandem MS parameters were used: resolution, 17,500; isolation window, 1.0 m/z; normalized collision energy, 15, 30 and 45; automatic gain control target, 1 × 10⁵; maximum inject time, 50 ms.

Quantification of VAN and PDC from extracellular samples was performed using UHPLC with UV detection at 275 nm (Thermo Scientific Vanquish Flex with diode array detector). The PDC standard was prepared from cultures of P. putida CJ251 (refs. ^20,59) and purified as previously reported⁵⁹.To quantify PDC, we dissolved the standard first in DMSO and then diluted 1:100 to the standard range in high-purity water (Thermo Scientific Barnstead GenPure). A matrix mix spike of PDC into minimal-nutrient medium was used to evaluate the impact of the sample matrix. All other standards were prepared in minimal-nutrient medium. Chromatographic separation was performed using a reversed-phase C18 column (ZORBAX Eclipse Plus, 4.6 × 100 mm, 5 μm; Agilent) with a guard column (4.6 × 12.5 mm, 5 μm; Agilent). Column temperature was maintained at 25 °C. Mobile phases consisted of 1% formic acid in LC–MS-grade water (solvent A) and 80:10:10 acetonitrile:methanol:LC–MS water (solvent B), which was reported to be appropriate for phenolic acid separation⁶⁰. The method described previously⁶⁰ was modified to a total run time of 10.5 min at a flow rate of 0.9 ml min^–1 and injection volume of 10 µl. The following multistep gradient was used for solvent B: 0 min, 6%; 1 min, 6%; 8.5 min, 25%; 9 min, 25%; 9.5 min, 6%; 10.5 min, 6%.

Flux ratio analysis

To determine the fractional contribution of OAA versus pyruvate to the PEP pool, we conducted metabolic flux ratio analysis during exponential growth of cells grown on [1-¹³C]-(carboxyl)-4HB. We used the positional ¹³C labeling of citrate determined from fragmentation analysis to estimate the percentage of OAA that was ¹³C labeled at the first or fourth position. We obtained optimized flux ratios by optimizing across three replicate datasets to obtain the minimum error difference between the experimental and simulated labeling data. The following equations were used:

$${{{\mathrm{PEP}}}}_0 = {{{\mathrm{f}}}}_1\left( {{{{\mathrm{pyruvate}}}}_0} \right) + {{{\mathrm{f}}}}_2\left( {{{{\mathrm{OAA}}}}_0 + {{{\mathrm{OAA}}}}_{{{{1}}},4{{{\mathrm{th}}}}}} \right)$$

(1)

$${{{\mathrm{PEP}}}}_1 = {{{\mathrm{f}}}}_1\left( {{{{\mathrm{pyruvate}}}}_1} \right) + {{{\mathrm{f}}}}_2\left( {{{{\mathrm{OAA}}}}_{1,1{{{\mathrm{st}}}}}} \right)$$

(2)

Here, f₁ and f₂ refer to the fractions of PEP originating from pyruvate and OAA, respectively. The subscript 0 indicates the non-labeled fraction, and the subscript 1 indicates the singly labeled fraction, while 1st and 4th refer to the position of the ¹³C-carbon in OAA.

Quantitative flux analysis

A metabolic network model consisting of 57 reactions was constructed for ¹³C MFA of C. testosteroni KF-1, as described previously¹⁸. The spontaneous reaction in the 4,5-meta pathway was assumed to be at equilibrium between the open and cyclic hemiacetal forms (Fig. 1a)³⁵. The biomass efflux rates of metabolite precursors were calculated from experimentally determined growth rates and the stoichiometric ratio of polymer components in biomass composition to metabolites in the model¹⁸. The stoichiometric molar ratio was next converted to a rate by accounting for the biomass and substrate-specific growth rate. The anabolic demand for NADPH, NADH, and ATP was determined from the growth rates, metabolic network and biomass stoichiometry, as reported previously¹⁸.

Quantification of metabolic fluxes in CCM during growth on 4HB was conducted on the software OpenFlux2 using an extension of the elementary metabolic unit algorithm to simulate isotopomer balances across parallel-labeling experiments⁶¹. The ¹³C MFA model was constrained by measured metabolite secretions and biomass effluxes. The TCA cycle oxidative reactions from OAA to succinate were considered irreversible in the model. However, the TCA cycle reactions between succinate and OAA and the three potential cataplerotic reactions (OAA to pyruvate, OAA to PEP and malate to pyruvate) were all considered reversible in the model and were left unconstrained. Optimization of the fluxes was conducted across six isotopomer datasets (two labeling schemes [1-¹³C]-(carboxyl)-4HB and [¹³C₆]-(phenyl)-4HB and three replicates per labeling scheme) and using the isotopomer distributions of metabolites in the TCA cycle (malate, fumarate, citrate and αKG), EMP pathway (pyruvate, PEP, 3-phosphoglycerate, fructose-6-phosphate and glucose-6-phosphate) and PP pathway (sedoheptulose-7-phosphate). The labeling of CO₂ was determined experimentally by comparing the ¹³C labeling of N-carbamoyl-l-aspartate to aspartate, as demonstrated previously¹⁸. Addition of ¹³C-labeled carbons from aspartate to N-carbamoyl-aspartate was taken as the incorporation of ¹³C-labeled dissolved CO₂. A global solution was obtained after 100 iterations starting from random initial values for all fluxes and evaluated by assessing the variance-weighted sum of squared residuals between the experimentally determined and model-estimated isotopomer distributions and efflux rates. When the minimized sum of squared residuals values were below the χ² statistical test^. cutoff at a 95% confidence level, the model was considered to have a statistically acceptable fit. A non-linear search algorithm was used to calculate 95% confidence intervals for the optimized metabolic fluxes⁶¹. The standard deviation (or individual flux precision) was calculated from the 95% confidence intervals as previously reported¹⁸.

Protein production and purification

Mdh (CtesDRAFT_PD1281), ME (CtesDRAFT_PD0934 and CtesDRAFT_PD5092), IDH (CtesDRAFT_PD1820 and CtesDRAFT_PD1824), 4HB 3-monooxygenase (CtesDRAFT_PD2627) and PEP synthase (CtesDRAFT_PD3828) from C. testosteroni KF-1 were recombinantly produced in E. coli BL21 (DE3) and subsequently purified (Supplementary Fig. 9). The gene fragments encoding the different enzymes were synthesized by Twist Bioscience. The native nucleotide coding sequences from C. testosteroni KF-1 were used for all the enzymes except for ME-1 (CtesDRAFT_PD0934), for which the codon usage was optimized for expression in E. coli K12 to decrease its GC content. The coding sequences CtesDRAFT_PD0934, CtesDRAFT_PD5092, CtesDRAFT_PD1820 and CtesDRAFT_PD3828 were divided into two gene fragments to ease the synthetic process. The synthesized fragments were amplified by PCR and inserted into a modified pET28a(+) vector by USER cloning⁶², and the corresponding USER primers were designed using the AMUSER tool2. The selected pET28a(+) vector carries a TEV cleavage site instead of the canonical thrombin target sequence between the N-terminal His tag and the multiple cloning site. Except for CtesDRAFT_PD1281, the enzymes were cloned as N-terminal His-tag fusion proteins, and their native start codon was removed. For CtesDRAFT_PD1281, its native start and stop codons were deleted, and the coding sequence was fused with the vector C-terminal His tag. E. coli DH5α λpir competent cells were used for plasmid construction. Plasmids were sequence verified and cloned into the expression host E. coli BL21 (DE3). The final strains were preserved in 25% glycerol at −80 °C. Supplementary Tables 22 and 23 display the sequences corresponding to the synthetic gene fragments and the USER primers, respectively.

For protein production, transformed E. coli BL21 (DE3) cells were grown in 2xYT medium supplemented with 50 μg ml^–1 kanamycin. Single colonies were used to inoculate precultures, which were grown overnight at 37 °C and 250 r.p.m. The cultures were prepared in 2-liter baffled flasks and incubated at 37 °C and 200 r.p.m. until the OD₆₀₀ reached 0.6–0.8. At this point, cultures were cooled down at 4 °C, and isopropyl-d-1-thiogalactopyranoside was added at a final concentration of 0.4 mM to induce protein expression. After induction, cultures were incubated at 20 °C and 200 r.p.m. for 18 h and then collected by centrifugation (4,000 g, 20 min, 4 °C). Cell pellets were resuspended in buffer A (20 mM sodium phosphate (pH 7.5), 300 mM NaCl and 20 mM imidazole), and cell lysis was performed by three passes through an Avestin Emulsiflex C5 French press (ATA Scientific Instruments). Nucleic acids were hydrolyzed by adding 2.5 U ml^–1 Pierce Universal Nuclease for Cell Lysis (Thermo Fisher Scientific) to the cell extracts, followed by incubation for 30 min at room temperature with gentle shaking. Before purification, the lysates were centrifuged (10,000 g, 20 min, 4 °C), and the supernatants were filtered (0.2-μm-pore size).

Purification was performed by immobilized metal affinity chromatography using 1 ml of Ni-NTA resin (HisPur Ni-NTA resin, Thermo Fisher) loaded in 10-ml Pierce disposable columns. The resin was equilibrated with 10 ml of buffer A. The corresponding lysate was passed through the column followed by a washing step with 20 ml of buffer A to remove the non-bound proteins. Elution was performed with 4 ml of buffer B (20 mM sodium phosphate (pH 7.5), 300 mM NaCl and 500 mM imidazole). The buffer of the collected fractions was exchanged using PD-10 desalting columns (Cytiva) according to the instructions of the manufacturer. The new buffer was 20 mM sodium phosphate (pH 7.5) and 300 mM NaCl for all the purified enzymes with the exception of PEPs (CtesDRAFT_PD3828). The proteins were stored at 4 °C with the addition of 0.5 mM DTT until use. For PEPs, the storage buffer was 20 mM HEPES (pH 7.5) and 50 mM NaCl supplemented with 1 mM EDTA and 10 mM DTT to ensure stability⁶³; the protein was stored at 4 °C until use.

The concentration of the purified enzymes was determined in a NanoDrop 2000 spectrophotometer using the theoretical extinction coefficient of the respective proteins. The enzyme purity was confirmed by SDS–PAGE, loading 1–2 μg of protein per sample (Supplementary Fig. 10).

Enzyme assays

Unless otherwise stated, the enzymatic activities were determined following the reduction or oxidation of the NAD(P)⁺/NAD(P)H pair. Samples were prepared in triplicate at a final volume of 200 μl. Reactions were incubated at 30 °C for 8 h with no shaking using an ELx808 plate reader (BioTek Instruments). Absorbance measurements at 340 nm were taken in intervals of 20 s. Absolute consumption or formation of NAD(P)H was determined using the respective standard curve from 0.016 mM to 1 mM NAD(P)H.

Cofactor specificity

The cofactor specificity of Mdh (CtesDRAFT_PD1281), ME (CtesDRAFT_PD0934 and CtesDRAFT_PD5092), IDH (CtesDRAFT_PD1820 and CtesDRAFT_PD1824) and 4HB 3-monooxygenase (CtesDRAFT_PD2627) was determined by comparing the activity of the enzymes with both cofactors. For Mdh, the reaction mixture contained 50 mM HEPES (pH 7.5), 10 mM sodium OAA (preparing the stock solution within the same day), 5 mM MgCl₂ and 0.01% bovine serum albumin (BSA). NADH and NADPH were added at 0.25 mM and 0.5 mM, respectively. The reactions were prepared using 0.0020 μM CtesDRAFT_PD1281. The two MEs (ME-1 and ME-2) were assayed in a reaction mixture containing 50 mM HEPES (pH 7.5), 20 mM sodium l-malate, 5 mM MnCl₂ and 0.01% BSA. NAD⁺ and NADP⁺ were added at 0.5 mM and 1 mM, respectively. The reactions were performed using 0.040 μM CtesDRAFT_PD0934 or 0.75 μM CtesDRAFT_PD5092. The two IDHs (IDH1 and IDH2) were assayed in a reaction mixture containing 50 mM HEPES (pH 7.5), 10 mM sodium d,l-isocitrate, 5 mM MnCl₂ and 0.01% BSA. NAD⁺ and NADP⁺ were added at 0.5 mM and 1 mM, respectively. The reactions were prepared using 0.0025 μM CtesDRAFT_PD1820 or 0.030 μM CtesDRAFT_PD1824. For 4HB 3-monooxygenase, the reaction mixture contained 50 mM HEPES (pH 7.5), 10 mM sodium 4HB, 5 mM MgCl₂ and 0.01% BSA. NADH and NADPH were added at 0.25 mM and 0.5 mM. The reactions were prepared using 0.20 μM CtesDRAFT_PD2627.

Allosteric regulation

The effect of the potential allosteric agents on enzyme activity was analyzed relative to a control reaction without the selected compound. For the two MEs, the reaction mixture contained 50 mM HEPES (pH 7.5), 1 mM sodium l-malate, 5 mM MnCl₂, and 0.01% BSA. The cofactors NADP⁺ and NAD⁺ were used at 0.5 mM for CtesDRAFT_PD0934 and CtesDRAFT_PD5092, respectively. The potential allosteric reagents (l-glutamate, l-aspartate, fumarate, succinate and αKG) were assayed at 0.5 mM and 2 mM. The reactions were prepared using 0.30 μM CtesDRAFT_D0934 or 3.2 μM CtesDRAFT_PD5092. For PEPs (CtesDRAFT_PD3828), the potential inhibitory role of αKG was analyzed. The reaction mixture contained 50 mM HEPES (pH 7.5), 1 mM sodium pyruvate, 1.5 mM ATP, 5 mM MgCl₂ and 0.01% BSA. The reactions were prepared using 0.20 μM CtesDRAFT_PD3828. The effect of αKG was assayed at 2 mM and 5 mM. Samples were incubated at 30 °C for 30 min with no shaking, and the reaction was stopped by the addition of 5% formic acid. The activity of CtesDRAFT_PD3828 was determined based on the consumption of pyruvate measured on a Dionex UltiMate 3000 HPLC system equipped with an RI‐101 refractive index detector (Shodex). Samples were loaded onto an Aminex HPX‐87X ion exclusion (300 × 7.8 mm) column (Bio-Rad) kept at 30 °C. The mobile phase was 5 mM sulfuric acid at 0.6 ml min^–1, and an isocratic elution was applied for 11 min. Data processing was performed using the Chromeleon Chromatography Data System software 7.2.9 (Thermo Fisher Scientific).

Plasmid construction

Plasmids were synthesized by Genscript USA. Plasmids were designed based on the published sequences for C. testosteroni KF-1 (NCBI Reference Sequence NZ_AAUJ02000001.1) and either a pBBR1 backbone^64,65 for autonomous replication or a pk18mobsacB backbone⁶⁶ for chromosomal modification via homologous recombination. For pALC785, the P_tac promoter and pmdAB genes were synthesized and cloned into pBBR1-MCS1. For pALC786, the P_tac promoter and pmdD gene were synthesized and cloned into pBBR1-MCS1. For each of these plasmids, the pmd genes were codon optimized for expression in P. putida to prevent recombination with the native pmd locus⁶⁷. For deletion of the chromosomal pmd operon, a DNA region 1,058 bp in length upstream of the operon and 1,055 bp downstream of the operon were synthesized together into pK18mobsacB to serve as loci for homologous recombination, resulting in plasmid pALC783. For insertion of a strong promoter in front of the pmd operon, DNA was synthesized containing 1,058 bp of DNA homologous to the region upstream of the pmd operon followed by the P_tac promoter, lacO operator, and a ribosome binding site previously characterized in P. putida KT2440 (ref. ⁶⁸), followed by the first 1,008 bp of the coding region. This sequence was cloned into pK18mobsacB, resulting in plasmid pALC784. Synthesized DNAs, plasmid sequences and primers are listed in Supplementary Table 24 and Supplementary Data.

Transformation and strain construction in C. testosteroni

To enable genetic modifications in C. testosteroni KF-1, a transformation protocol was first developed. C. testosteroni was grown overnight in 50 ml of LB liquid medium at 30 °C and 250 r.p.m. The following day, stationary-phase cells were collected by centrifugation at 5,000 g. The supernatant was removed, and cells were resuspended in 25 ml of 10% glycerol. This wash was repeated two additional times before a final resuspension in 1 ml of 10% glycerol. This suspension was divided into 50-µl aliquots, which were ready for electroporation or frozen at −80 °C for later use. All cell collection and wash steps were performed at room temperature.

For electroporation, 100 ng of plasmid was added to 50 µl of competent cells at room temperature. Cells were electroporated at 1,600 V, 25 µF and 200 Ω in a 1-mm cuvette using a Gene Pulser Xcell (Bio-Rad). Observed time constants were ~4 ms. Cells were immediately recovered in 950 µl of SOC medium and incubated at 30 °C and 250 r.p.m. for 2 h. Cells were plated on LB solid medium (25 g liter^–1 LB granules (Fisher Scientific) and 15 g liter^–1 agar) with 50 µg liter^–1 kanamycin sulfate (Fisher Scientific) to select for transformants. It was observed that some untransformed cells were able to escape selection at concentrations of 50 µg liter^–1 kanamycin sulfate, so an additional selection was performed at a concentration of 200 µg liter^–1 kanamycin sulfate by streaking on agar plates to isolate individual colonies that were resistant to this higher level of kanamycin. Successful transformation with the pBBR1-based replicating vectors pALC785 and pALC786 was confirmed by colony PCR (Supplementary Fig. 11), resulting in strains AG9402 and AG9404, respectively. For homologous recombination plasmids, sacB was used as a counterselectable marker. After kanamycin selection, cells were plated on YT 20% sucrose plates (10 g liter⁻¹ tryptone, 5 g liter^–1 yeast extract and 200 g liter^–1 sucrose). Colonies appeared after several days; however, no discernable differences in colony size were observed. Additionally, these colonies remained kanamycin insensitive, even after an additional propagation on YT + 20% sucrose plates, suggesting that no counterselection occurred. Cells were then plated on LB + 10% sucrose plates, and after 24 h, differences in colony size were observed. To create strain AG9493, C. testosteroni KF-1 was transformed with pALC783. Deletion of the pmd operon was confirmed by colony PCR (Supplementary Fig. 12), and colonies were confirmed to be sensitive to kanamycin concentrations of 200 µg liter^–1 in both liquid and solid media. For transformation with pALC784, all colonies screened after counterselection showed bands consistent with the wild-type genotype. As an alternative, initial transformed colonies were screened using colony PCR for a single crossover event (after kanamycin selection and before sucrose counterselection) into the downstream region of homology, which would introduce the intended promoter in front of the entire pmd operon (Supplementary Fig. 13) as a semistable merodiploid (strain AG9480). Kanamycin selection (200 µg liter^–1) was maintained during experiments with this strain to ensure that an additional recombination event would not revert the genotype to the wild type. All primers are listed in Supplementary Table 24.

For future genetic modification of C. testosteroni KF-1, use of 200 µg liter^–1 kanamycin rather than 50 µg liter^–1 in liquid and solid media is recommended to ensure elimination of cells without kanR. Electroporation of pALC783-786 was repeated using 200 µg liter^–1 kanamycin for selection and resulted in transformation efficiencies of 16 and 120 colony-forming units per µg of DNA for the homologous recombination plasmids pALC783 and pALC784, respectively, and 154,000 and 176,000 colony-forming units per µg of DNA for the autonomously replicating plasmids pALC785 and pALC786, respectively. For counterselection using sacB, LB + 10% sucrose plates are recommended, as they produced clear phenotypic differences in transformants after overnight incubation.

Statistical analysis

Physiology experiments were analyzed for statically significant differences (P ≤ 0.05) between growth conditions using one-way analysis of variance (ANOVA) combined with Tukey honestly significant difference (HSD) post hoc tests. The Pearson correlation coefficient (r) was calculated to determine the strength of the relationship between transcriptomics and proteomics and between fluxomics and transcriptomics, proteomics or substrate-to-product ratios. A two-sided t-test was used to assess the significance of the relationships. For proteomics analysis, statistically significant differential protein abundance between aromatic-grown cells and succinate-grown cells was determined by calculating a z score for protein abundance differences as described previously¹⁸. In brief, assuming a standard normal distribution, z-score values were translated to P values and further corrected using the q value method (controlled to 0.05) to correct for multiple-testing familywise error rate.

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.

Data availability

Proteomic mass spectral data are available via ProteomeXchange under accession number PXD029813 and the MassIVE repository (massive.ucsd.edu) under accession number MSV000088418. Stable isotope-assisted metabolomics LC–HRMS data are deposited in the MetaboLights repository (www.ebi.ac.uk/metabolights/) under accession number MTBLS3947. The RNA-sequencing datasets generated during this study are available at the Gene Expression Omnibus with accession number GSE192852. The NCBI Reference Sequence for C. testosteroni KF-1 was NZ_AAUJ02000001.1.

Change history

05 February 2024
A Correction to this paper has been published: https://doi.org/10.1038/s41589-024-01561-0

References

Tang, Q., Lu, T. & Liu, S. J. Developing a synthetic biology toolkit for Comamonas testosteroni, an emerging cellular chassis for bioremediation. ACS Synth. Biol. 7, 1753–1762 (2018).
Article CAS PubMed Google Scholar
Wang, Y. H., Huang, Z. & Liu, S. J. Chemotaxis towards aromatic compounds: insights from Comamonas testosteroni. Int. J. Mol. Sci. 20, 2701 (2019).
Article CAS PubMed PubMed Central Google Scholar
Weiss, M. et al. Permanent draft genome sequence of Comamonas testosteroni KF-1. Stand. Genom. Sci. 8, 239–254 (2013).
Article Google Scholar
Zhu, D. et al. Genomics and biochemistry investigation on the metabolic pathway of milled wood and alkali lignin-derived aromatic metabolites of Comamonas serinivorans SP-35. Biotechnol. Biofuels 11, 338 (2018).
Article CAS PubMed PubMed Central Google Scholar
Ni, B. et al. Assimilation of aromatic compounds by Comamonas testosteroni: characterization and spreadability of protocatechuate 4,5-cleavage pathway in bacteria. Appl. Microbiol. Biotechnol. 97, 6031–6041 (2013).
Article CAS PubMed Google Scholar
Hosaka, M. et al. Novel tripartite aromatic acid transporter essential for terephthalate uptake in Comamonas sp. strain E6. Appl. Environ. Microbiol. 7, 6148–6155 (2013).
Article ADS Google Scholar
Liu, L. et al. High correlation between genotypes and phenotypes of environmental bacteria Comamonas testosteroni strains. BMC Genomics 16, 110 (2015).
Article PubMed PubMed Central Google Scholar
Ladino-Orjuela, G., Gomes, E., da Silva, R., Salt, C. & Parsons, J. R. in Reviews of Environmental Contamination and Toxicology, Vol. 237 (ed. de Voogt, W.P.) 105–121 (Springer, 2016).
Fuchs, G., Boll, M. & Heider, J. Microbial degradation of aromatic compounds—from one strategy to four. Nat. Rev. Microbiol. 9, 803–816 (2011).
Article CAS PubMed Google Scholar
Jiménez, J. I., Miñambres, B., Luis García, J. & Díaz, E. Genomic analysis of the aromatic catabolic pathways from Pseudomonas putida KT2440. Environ. Microbiol. 4, 824–841 (2002).
Article PubMed Google Scholar
Kasai, D. et al. Uncovering the protocatechuate 2,3-cleavage pathway genes. J. Bacteriol. 191, 6758–6768 (2009).
Article CAS PubMed PubMed Central Google Scholar
Hara, H., Masai, E., Miyauchi, K., Katayama, Y. & Fukuda, M. Characterization of the 4-carboxy-4-hydroxy-2-oxoadipate aldolase gene and operon structure of the protocatechuate 4,5-cleavage pathway genes in Sphingomonas paucimobilis SYK-6. J. Bacteriol. 185, 41–50 (2003).
Article CAS PubMed PubMed Central Google Scholar
Varman, A. M. et al. Decoding how a soil bacterium extracts building blocks and metabolic energy from ligninolysis provides road map for lignin valorization. Proc. Natl Acad. Sci. USA 40, E5802–E5811 (2016).
Google Scholar
Westphal, A. H. et al. Pyridine nucleotide coenzyme specificity of p-hydroxybenzoate hydroxylase and related flavoprotein monooxygenases. Front. Microbiol. 9, 3050 (2018).
Article PubMed PubMed Central Google Scholar
Providenti, M. A., O’Brien, J. M., Ruff, J., Cook, A. M. & Lambert, I. B. Metabolism of isovanillate, vanillate, and veratrate by Comamonas testosteroni strain BR6020. J. Bacteriol. 188, 3862–3869 (2006).
Sasoh, M. et al. Characterization of the terephthalate degradation genes of Comamonas sp. strain E6. Appl. Environ. Microbiol. 72, 1825–1832 (2006).
Article ADS CAS PubMed PubMed Central Google Scholar
Schlafli, H. R., Weiss, M. A., Leisinger, T. & Cook, A. M. Terephthalate 1,2-dioxygenase system from Comamonas testosteroni T-2: purification and some properties of the oxygenase component. J. Bacteriol. 176, 6644–6652 (1994).
Article CAS PubMed PubMed Central Google Scholar
Wilkes, R. A., Waldbauer, J. & Aristilde, L. Analogous metabolic decoupling in Pseudomonas putida and Comamonas testosteroni implies energetic bypass to facilitate gluconeogenic growth. mBio 12, e0325921 (2021).
Article PubMed Google Scholar
Otsuka, Y. et al. Efficient production of 2-pyrone 4,6-dicarboxylic acid as a novel polymer-based material from protocatechuate by microbial function. Appl. Microbiol. Biotechnol. 71, 608–614 (2006).
Article CAS PubMed Google Scholar
Johnson, C. W. et al. Innovative chemicals and materials from bacterial aromatic catabolic pathways. Joule 3, 1523–1537 (2019).
Article CAS Google Scholar
Rosenberg, J. & Commichau, F. M. Harnessing underground metabolism for pathway development. Trends Biotechnol. 37, 29–37 (2019).
Article CAS PubMed Google Scholar
Lanzavecchia, A. et al. Interdependence of cell growth and gene expression: origins and consequences. Science 330, 1099–1102 (2010).
Article Google Scholar
E. J. O’Brien, J. & Utrilla, B. O. Palsson, quantification and classification of E. coli proteome utilization and unused protein costs across environments. PLoS Comput. Biol. 12, e1004998 (2016).
Article Google Scholar
Bennett, B. D. et al. Absolute metabolite concentrations and implied enzyme active site occupancy in Escherichia coli. Nat. Chem. Biol. 5, 593–599 (2009).
Article CAS PubMed PubMed Central Google Scholar
Park, J. O. et al. Metabolite concentrations, fluxes and free energies imply efficient enzyme usage. Nat. Chem. Biol. 12, 482–489 (2016).
Article CAS PubMed PubMed Central Google Scholar
Kukurugya, M. A. M. A. et al. Multi-omics analysis unravels a segregated metabolic flux network that tunes co-utilization of sugar and aromatic carbons in Pseudomonas putida. J. Biol. Chem. 294, 8464–8479 (2019).
Article CAS PubMed PubMed Central Google Scholar
Hackett, S. R. et al. Systems-level analysis of mechanisms regulating yeast metabolic flux. Science 354, aaf2786-1 (2016).
Article Google Scholar
Chubukov, V. et al. Transcriptional regulation is insufficient to explain substrate-induced flux changes in Bacillus subtilis. Mol. Syst. Biol. 9, 709 (2013).
Article CAS PubMed PubMed Central Google Scholar
Tepper, N. et al. Steady-state metabolite concentrations reflect a balance between maximizing enzyme efficiency and minimizing total metabolite load. PLoS ONE 8, e75370 (2013).
Article ADS CAS PubMed PubMed Central Google Scholar
Richter, H. et al. Ethanol production in syngas-fermenting Clostridium ljungdahlii is controlled by thermodynamics rather than by enzyme expression. Energy Environ. Sci. 9, 2392–2399 (2016).
Article CAS Google Scholar
Chulavatnatol, M. & Atkinson, D. E. Phosphoenolpyruvate synthetase from Escherichia coli. J. Biol. Chem. 248, 2712–2715 (1973).
Article CAS PubMed Google Scholar
Cheng, V. W. T. et al. Redox state of flavin adenine dinucleotide drives substrate binding and product release in Escherichia coli succinate dehydrogenase. Biochemistry 54, 1043–1052 (2015).
van der Rest, M. E., Frank, C. & Molenaar, D. Functions of the membrane-associated and cytoplasmic malate dehydrogenases in the citric acid cycle of Escherichia coli. J. Bacteriol. 182, 6892–6899 (2000).
Article PubMed PubMed Central Google Scholar
Molenaar, D., van der Rest, M. E., Petrović, S. P. & Zentrallabor, B. Biochemical and genetic characterization of the membrane-associated malate dehydrogenase (acceptor) from Corynebacterium glutamicum. Eur. J. Biochem. 254, 395–403 (1998).
Masai, E. et al. Genetic and biochemical characterization of 4-carboxy-2-hydroxymuconate-6-semialdehyde dehydrogenase and its role in the protocatechuate 4,5-cleavage pathway in Sphingomonas paucimobilis SYK-6. J. Bacteriol. 182, 6651–6658 (2000).
Article CAS PubMed PubMed Central Google Scholar
Bologna, F. P., Andreo, C. S. & Drincovich, M. F. Escherichia coli malic enzymes: two isoforms with substantial differences in kinetic properties, metabolic regulation, and structure. J. Bacteriol. 189, 5937–5946 (2007).
Article CAS PubMed PubMed Central Google Scholar
Calero, P. & Nikel, P. I. Chasing bacterial chassis for metabolic engineering: a perspective review from classical to non-traditional microorganisms. Microb. Biotechnol. 12, 98–124 (2019).
Article CAS PubMed Google Scholar
Blank, L. M., Narancic, T., Mampel, J., Tiso, T. & O’Connor, K. Biotechnological upcycling of plastic waste and other non-conventional feedstocks in a circular economy. Curr. Opin. Biotechnol. 62, 212–219 (2020).
Article CAS PubMed Google Scholar
Tiso, T. et al. Towards bio-upcycling of polyethylene terephthalate. Metab. Eng. 66, 167–178 (2021).
Article CAS PubMed Google Scholar
Becker, J. & Wittmann, C. A field of dreams: lignin valorization into chemicals, materials, fuels, and health-care products. Biotechnol. Adv. 37, 107360 (2019).
Article CAS PubMed Google Scholar
Kamimura, N. et al. Bacterial catabolism of lignin-derived aromatics: new findings in a recent decade: update on bacterial lignin catabolism. Environ. Microbiol. Rep. 9, 679–705 (2017).
Article CAS PubMed Google Scholar
Panda, S. K., Sahu, L., Behera, S. K. & Ray, R. C. in Bioprocessing for Biomolecules Production (eds Molina, G. et al.) 195–209 (Wiley, 2019).
Akita, H., Nakashima, N. & Hoshino, T. Pyruvate production using engineered Escherichia coli. AMB Express 6, 94 (2016).
Article PubMed PubMed Central Google Scholar
Sano, C. History of glutamate production. Am. J. Clin. Nutr. 90, 728S–732S (2009).
Article CAS PubMed Google Scholar
Lassek, C., Berger, A., Zühlke, D., Wittmann, C. & Riedel, K. Proteome and carbon flux analysis of Pseudomonas aeruginosa clinical isolates from different infection sites. Proteomics 16, 1381–1385 (2016).
Article CAS PubMed Google Scholar
Schilling, O. et al. Transcriptional and metabolic responses of Bacillus subtilis to the availability of organic acids: transcription regulation is important but not sufficient to account for metabolic adaptation. Appl. Environ. Microbiol. 73, 499–507 (2007).
Article ADS CAS PubMed Google Scholar
Haverkorn Van Rijsewijk, B. R. B., Nanchen, A., Nallet, S., Kleijn, R. J. & Sauer, U. Large-scale ¹³C-flux analysis reveals distinct transcriptional control of respiratory and fermentative metabolism in Escherichia coli. Mol. Syst. Biol. 7, 477 (2011).
Article PubMed PubMed Central Google Scholar
Monk, J. M. et al. Multi-omics quantification of species variation of Escherichia coli links molecular features with strain phenotypes. Cell Syst. 3, 238–251 (2016).
Article CAS PubMed PubMed Central Google Scholar
Toya, Y. et al. ¹³C-metabolic flux analysis in heterologous cellulase production by Bacillus subtilis genome-reduced strain. J. Biotechnol. 179, 42–49 (2014).
Article CAS PubMed Google Scholar
Johnson, C. W. & Beckham, G. T. Aromatic catabolic pathway selection for optimal production of pyruvate and lactate from lignin. Metab. Eng. 28, 240–247 (2015).
Article CAS PubMed Google Scholar
Paczia, N. et al. Extensive exometabolome analysis reveals extended overflow metabolism in various microorganisms. Microb. Cell Fact. 11, 122 (2012).
Article CAS PubMed PubMed Central Google Scholar
Bray, N. L., Pimentel, H., Páll, M. & Pachter, L. Near-optimal probabilistic RNA-seq quantification. Nat. Biotechnol. 34, 525–527 (2016).
Article CAS PubMed Google Scholar
Powell, D. R. Degust: interactive RNA-seq analysis. Zenodo https://doi.org/10.5281/zenodo.3258932 (2019).
Erde, J., Loo, R. R. O. & Loo, J. A. Enhanced FASP (eFASP) to increase proteome coverage and sample recovery for quantitative proteomic experiments. J. Proteome Res. 13, 1885–1895 (2014).
Article CAS PubMed PubMed Central Google Scholar
Waldbauer, J., Zhang, L., Rizzo, A. & Muratore, D. diDO-IPTL: a peptide-labeling strategy for precision quantitative proteomics. Anal. Chem. 89, 11498–11504 (2017).
Article CAS PubMed Google Scholar
Clasquin, M. F., Melamud, E. & Rabinowitz, J. D. LC–MS data processing with MAVEN: a metabolomic analysis and visualization engine. Curr. Protoc. Bioinformatics 37, 14.11.1–14.11.23 (2012).
Millard, P. et al. IsoCor: isotope correction for high-resolution MS labeling experiments. Bioinformatics 35, 4484–4487 (2019).
Article CAS PubMed Google Scholar
Kappelmann, J., Klein, B., Geilenkirchen, P. & Noack, S. Comprehensive and accurate tracking of carbon origin of LC–tandem mass spectrometry collisional fragments for ¹³C-MFA. Anal. Bioanal. Chem. 409, 2309–2326 (2017).
Article CAS PubMed PubMed Central Google Scholar
Notonier, S. et al. Metabolism of syringyl lignin-derived compounds in Pseudomonas putida enables convergent production of 2-pyrone-4,6-dicarboxylic acid. Metab. Eng. 65, 111–122 (2021).
Article CAS PubMed Google Scholar
Hefni, M. E., Amann, L. S. & Witthöft, C. M. A HPLC-UV method for the quantification of phenolic acids in cereals. Food Anal. Methods 12, 2802–2812 (2019).
Article Google Scholar
Shupletsov, M. S. et al. OpenFLUX2: ¹³C-MFA modeling software package adjusted for the comprehensive analysis of single and parallel labeling experiments. Microb. Cell Fact. 13, 152 (2014).
PubMed PubMed Central Google Scholar
Cavaleiro, A. M., Kim, S. H., Seppälä, S., Nielsen, M. T. & Nørholm, M. H. H. Accurate DNA assembly and genome engineering with optimized uracil excision cloning. ACS Synth. Biol. 4, 1042–1046 (2015).
Article CAS PubMed Google Scholar
Burnell, J. N. Cloning and characterization of Escherichia coli DUF299: a bifunctional ADP-dependent kinase–π-dependent pyrophosphorylase from bacteria. BMC Biochem. 11, 1 (2010).
Article PubMed PubMed Central Google Scholar
Kovach, M. E. et al. Four new derivatives of the broad-host-range cloning vector pBBR1MCS, carrying different antibiotic-resistance cassettes. Gene 166, 175–176 (1995).
Article CAS PubMed Google Scholar
Lee, T. S. et al. BglBrick vectors and datasheets: a synthetic biology platform for gene expression. J. Biol. Eng. 5, 12 (2011).
Article CAS PubMed PubMed Central Google Scholar
Marx, C. J. Development of a broad-host-range sacB-based vector for unmarked allelic exchange. BMC Res. Notes 1, 1 (2008).
Article PubMed PubMed Central Google Scholar
Halper, S. M., Hossain, A. & Salis, H. M. Synthesis success calculator: predicting the rapid synthesis of DNA fragments with machine learning. ACS Synth. Biol. 9, 1563–1571 (2020).
Article CAS PubMed Google Scholar
Elmore, J. R., Furches, A., Wolff, G. N., Gorday, K. & Guss, A. M. Development of a high efficiency integration system and promoter library for rapid modification of Pseudomonas putida KT2440. Metab. Eng. Commun. 5, 1–8 (2017).
Article PubMed PubMed Central Google Scholar

Download references

Acknowledgements

The US National Science Foundation (NSF) Graduate Research Fellowship Program (DGE-1650441) provided support for R.A.W. Metabolomics studies were funded by a grant awarded to L.A. from the US NSF (CBET-2022854). This work was authored, in part, by Oak Ridge National Laboratory, which is managed by UT-Battelle, LLC, for the US Department of Energy (DOE) under contract DE-AC05-00OR22725. Funding was provided, in part, by the US DOE, Office of Energy Efficiency and Renewable Energy, Advanced Manufacturing Office and Bioenergy Technologies Office as part of the BOTTLE Consortium. We thank D. Salvachúa and C. Hoyt of the National Renewable Energy Laboratory (NREL) for synthesizing and supplying the 2-pyrone-4,6-dicarboxylic acid standard for quantification by HPLC. We thank G. Beckham of NREL for his feedback to early drafts of this manuscript. We acknowledge E. Hartmann of Northwestern University for granting access to their laboratory for RNA extractions.

Author information

Authors and Affiliations

Department of Biological and Environmental Engineering, College of Agriculture and Life Sciences, Cornell University, Ithaca, NY, USA
Rebecca A. Wilkes & Ludmilla Aristilde
Department of Civil and Environmental Engineering, McCormick School of Engineering and Applied Science, Northwestern University, Evanston, IL, USA
Rebecca A. Wilkes & Ludmilla Aristilde
Department of the Geophysical Sciences, University of Chicago, Chicago, IL, USA
Jacob Waldbauer & Lichun Zhang
Biosciences Division, Oak Ridge National Laboratory, Oak Ridge, TN, USA
Austin Carroll, Darren J. Parker & Adam M. Guss
The Novo Nordisk Foundation Center for Biosustainability, Technical University of Denmark, Kongens Lyngby, Denmark
Manuel Nieto-Domínguez
Northwestern Center for Synthetic Biology, Evanston, IL, USA
Ludmilla Aristilde

Authors

Rebecca A. Wilkes
View author publications
You can also search for this author in PubMed Google Scholar
Jacob Waldbauer
View author publications
You can also search for this author in PubMed Google Scholar
Austin Carroll
View author publications
You can also search for this author in PubMed Google Scholar
Manuel Nieto-Domínguez
View author publications
You can also search for this author in PubMed Google Scholar
Darren J. Parker
View author publications
You can also search for this author in PubMed Google Scholar
Lichun Zhang
View author publications
You can also search for this author in PubMed Google Scholar
Adam M. Guss
View author publications
You can also search for this author in PubMed Google Scholar
Ludmilla Aristilde
View author publications
You can also search for this author in PubMed Google Scholar

Contributions

R.A.W. and L.A. designed research. R.A.W. performed growth experiments, metabolomics experiments and metabolic flux modeling. R.A.W. analyzed metabolomics data. J.W. and L.Z. performed proteomics measurements. R.A.W. and J.W. analyzed proteomics data. A.C. prepared and characterized mutant strains. M.N.D. conducted enzyme assays. D.J.P. processed RNA-sequencing reads. R.A.W., D.J.P. and A.M.G. analyzed transcriptomics data. R.A.W. and L.A. wrote the manuscript. L.A. provided funding, resources and supervision. All authors provided feedback on the findings and final edits to the manuscript.

Corresponding author

Correspondence to Ludmilla Aristilde.

Ethics declarations

Competing interests

The authors declare no competing interests.

Peer review

Peer review information

Nature Chemical Biology thanks Yinjie Tang and the other, anonymous, reviewer(s) for their contribution to the peer review of this work.

Additional information

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Extended data

Extended Data Fig. 1 Growth profiles, substrate depletion, and metabolite secretions.

a) Kinetics of 4HB depletion (orange), VAN depletion (blue), and TER depletion (red) throughout cellular growth (grey) in C. testosteroni KF-1. Biological replicates (n = 3) are shown separately with the average represented by the line through the points. b) Kinetics and rates of metabolite secretion during growth on 4HB. Secretion data were from biological replicates (n = 3) shown as individual data points. Metabolite abbreviations: protocatechuate, PCA; 2-pyrone-4,6-dicarboxylate, PDC; pyruvate, Pyr; α-ketoglutarate, αKG.

Extended Data Fig. 2 Transcriptional regulators upregulated during growth on aromatic substrates.

Transcript abundance represents triplicate measurements of log2 fold changes for cells grown on the aromatic compounds relative to succinate.

Extended Data Fig. 3 Growth profile of 4,5-meta knockout mutant.

Strain AG9493 was grown on 100 mM C of succinate (green), 4-hydroxybenzoate (orange), vanillate (blue), or terephthalate (red). No growth was observed on the aromatic compounds after 42 h.

Supplementary information

Supplementary Information

Supplementary Figs. 1–13 and Tables 1–24.

Reporting Summary

Supplementary Data 1

DNA synthesis and plasmid sequences for plasmid construction and statistical information for Fig. 1b,c.

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/.

Reprints and permissions

About this article

Cite this article

Wilkes, R.A., Waldbauer, J., Carroll, A. et al. Complex regulation in a Comamonas platform for diverse aromatic carbon metabolism. Nat Chem Biol 19, 651–662 (2023). https://doi.org/10.1038/s41589-022-01237-7

Download citation

Received: 21 January 2022
Accepted: 29 November 2022
Published: 06 February 2023
Issue Date: May 2023
DOI: https://doi.org/10.1038/s41589-022-01237-7

This article is cited by

Engineering Comamonas testosteroni for the production of 2-pyrone-4,6-dicarboxylic acid as a promising building block
- Tom Delmulle
- Stijn Bovijn
- Wim K. Soetaert
Microbial Cell Factories (2023)
Mining microbial metabolism
- Yinjie J. Tang
- Tae Seok Moon
Nature Chemical Biology (2023)