ABSTRACT
The metabolism of one- and two-carbon compounds by the methylotrophic bacterium Methylobacterium extorquens AM1 involves high carbon flux through the ethylmalonyl coenzyme A (ethylmalonyl-CoA) pathway (EMC pathway). During growth on ethylamine, the EMC pathway operates as a linear pathway carrying the full assimilatory flux to produce glyoxylate, malate, and succinate. Assimilatory carbon enters the ethylmalonyl-CoA pathway directly as acetyl-CoA, bypassing pathways for formaldehyde oxidation/assimilation and the regulatory mechanisms controlling them, making ethylamine growth a useful condition to study the regulation of the EMC pathway. Wild-type M. extorquens cells were grown at steady state on a limiting concentration of succinate, and the growth substrate was then switched to ethylamine, a condition where the cell must make a sudden switch from utilizing the tricarboxylic acid (TCA) cycle to using the ethylmalonyl-CoA pathway for assimilation, which has been an effective strategy for identifying metabolic control points. A 9-h lag in growth was observed, during which butyryl-CoA, a degradation product of ethylmalonyl-CoA, accumulated, suggesting a metabolic imbalance. Ethylmalonyl-CoA mutase activity increased to a level sufficient for the observed growth rate at 9 h, which correlated with an upregulation of RNA transcripts for ecm and a decrease in the levels of ethylmalonyl-CoA. When the wild-type strain overexpressing ecm was tested with the same substrate switchover experiment, ethylmalonyl-CoA did not accumulate, growth resumed earlier, and, after a transient period of slow growth, the culture grew at a higher rate than that of the control. These findings demonstrate that ethylmalonyl-CoA mutase is a metabolic control point in the EMC pathway, expanding our understanding of its regulation.
INTRODUCTION
The function of central carbon metabolism is to generate essential metabolic intermediates and the energy and reducing power necessary to convert them to biomass. Metabolism must be fine-tuned to maintain cell homeostasis, balancing the production and consumption of metabolic intermediates and preventing a lethal accumulation of toxic compounds. The regulation of enzymes that control this balance is vital to the cell, particularly when challenged with an environmental perturbation such as a sudden change in the growth substrate. Identifying metabolic control points leads to a better understanding of fundamental physiology and can suggest physiological significance, such as preventing a toxic intermediate from accumulating or balancing the growth rate and growth yield. In addition, characterization of the metabolic strategies employed to effectively navigate an environmental perturbation provides valuable insight for strain manipulation and applied goals.
Methylobacterium extorquens AM1 is a facultative methylotroph capable of growth on single-carbon (C1) compounds such as methanol and methylamine, multicarbon compounds such as succinate (C4), and two-carbon compounds such as ethylamine and acetate (C2) (1–3). M. extorquens is widespread in nature and is associated with the leaf surfaces of plants, a heterogeneous and dynamic growth environment (4, 5). In this niche, M. extorquens must acclimate to sudden changes in growth substrate availability while preventing the lethal accumulation of the toxic methanol oxidation product formaldehyde and the downstream toxic metabolites glyoxylate and glycine (6, 7). Recent studies have demonstrated that M. extorquens utilizes specific regulatory strategies when challenged with a sudden change in the growth substrate from succinate to methanol (8). Under these conditions, M. extorquens restricts carbon flux to the assimilatory pathways until the necessary enzymes are upregulated sufficiently to accommodate the full assimilatory flux. These strategies allow M. extorquens to acclimate to the change in the growth substrate more efficiently and prevent a metabolic imbalance. While regulatory checkpoints have been identified in the methanol oxidation pathways (9), it was not possible to assess the control of the assimilatory pathways, since carbon flux into assimilation was restricted, limiting the magnitude of the observable change.
When M. extorquens grows on multicarbon substrates such as succinate, it utilizes a typical aerobic, heterotrophic metabolic network including the oxidation of acetyl coenzyme A (acetyl-CoA) via the tricarboxylic acid (TCA) cycle, NADH dehydrogenase, and an electron transport chain (1, 10). During growth on C2 compounds such as ethylamine and acetate, the primary growth substrate is oxidized to acetyl-CoA, which has two discrete inputs to central metabolism. The oxidation of acetyl-CoA to CO2 occurs primarily through the TCA cycle, while the assimilation of acetyl-CoA occurs through the ethylmalonyl-CoA pathway (EMC pathway) (11). The EMC pathway was recently described as an alternative pathway for the regeneration of glyoxylate in many isocitrate lyase-negative microbes (12). When M. extorquens is growing on a C2 substrate, the EMC pathway converts two molecules of acetyl-CoA plus two molecules of CO2 into one molecule of glyoxylate and one molecule of succinate (13, 14). Succinate is converted to malate via succinate dehydrogenase. Glyoxylate can be condensed with one molecule of acetyl-CoA to generate one molecule of malate or can be aminated to glycine (15). Glycine is converted to 2-phosphoglycerate for biomass production (11) (Fig. 1).
Metabolic pathway for two-carbon metabolism. Three known fates of beta-methylmalyl-CoA are shown. Genes encoding pathway enzymes are shown in gray. mauFBEDACJGLMN, methylamine dehydrogenases; aldB, aldehyde dehydrogenase; alo, aldehyde oxidase; acsA, acetyl-CoA synthase; phaA, β-ketothiolase; phaB, acetoacetyl-CoA reductase; croR, crotonase; ccr, crotonyl-CoA reductase/carboxylase; epi, ethylmalonyl-CoA/methylmalonyl-CoA epimerase; ecm, ethylmalonyl-CoA mutase; msd, methylsuccinyl-CoA dehydrogenase; mcd, mesaconyl-CoA hydratase; mclA1 and mclA2, β-methylmalyl-CoA lyases; mcl2, malyl-CoA thioesterase; pccAB, propionyl-CoA carboxylase; mcm, methylmalonyl-CoA mutase; sct, succinyl-CoA transferase; sga, serine-glyoxylate aminotransferase; gcv, glycine cleavage system; glyA, serine hydroxymethyltransferase. The asterisk indicates that genes for the conversion of serine to 2-phosphoglycerate are not shown. Genes in parentheses denote a predicted function not confirmed by the mutant phenotype. amcy, amicyanin; me-H4F, methylene tetrahydrofolate; PHB, poly-β-hydroxybutyrate cycle; FAD, flavin adenine dinucleotide; FADH, reduced flavin adenine dinucleotide.
Because acetyl-CoA enters directly into the EMC pathway, growth on a C2 compound provides a useful model for studying the EMC pathway independent of other C1 pathways and the mechanisms that regulate them. Furthermore, M. extorquens must balance the production and consumption of glyoxylate, a toxic intermediate that inhibits growth at millimolar concentrations (16). We hypothesized that the EMC pathway might harbor a metabolic control point to restrict the production of glyoxylate. Relatively little is known about the control of the EMC pathway enzymes, with only a single transcriptional regulator (CcrR, regulating ccr) having been reported (17). Most of the genes encoding the EMC pathway enzymes are not colocated in the M. extorquens genome, complicating the study of their regulation (18). Identifying a metabolic control point and characterizing its control would provide insight into how M. extorquens regulates carbon flux through the EMC pathway. This would also provide evidence concerning how cells acclimate to an environmental perturbation, such as a switch in the growth substrate, and how cells achieve metabolic balance to restore growth.
In this study, we report that ethylmalonyl-CoA mutase (Ecm), a reversible enzyme, is a regulatory control point in the EMC pathway that allows M. extorquens to efficiently restore metabolic balance when challenged with a sudden change in the growth substrate. Our multilayered analysis indicates that Ecm levels could be controlled at the transcriptional level.
Since its full description in 2007, the EMC pathway has been of interest for biotechnological applications due the variety of CoA esters produced as metabolic intermediates in the pathway (19–21). Identifying a regulatory control point in the EMC pathway will assist future engineering efforts to develop M. extorquens as a platform for the production of value-added chemicals.
MATERIALS AND METHODS
Materials.Chemicals and chemical standards were obtained from Sigma-Aldrich (St. Louis, MO) unless noted otherwise. Analytical-reagent-grade potassium dihydrogen phosphate and sodium dihydrogen phosphate were purchased from Fisher Scientific (Hampton, NH). High-performance liquid chromatography (HPLC)-grade methanol and absolute ethanol (Sigma, St. Louis, MO) were used for extraction procedures and buffer preparation for HPLC analysis. Nanopure water (Barnstead Life Sciences model; Thermo Fisher Scientific, Waltham, MA) was used for the preparation of all media, buffers, standards, and sample solutions.
Bacterial strains and growth conditions.M. extorquens strain AM1 and plasmids used in this work are described in Table 1. M. extorquens AM1 cells were grown in shake flasks, as reported previously, with succinate or ethylamine as the growth substrate (8, 9, 15).
M. extorquens strains and plasmids used in this study
Chemostat cultivation.Cultures were grown as reported previously (8, 22), with minor changes, as described here in brief. For the initial time point (time [T] = 0 min) of the substrate transition experiment, M. extorquens AM1 cells were grown to steady state in a chemostat culture with a limiting succinate concentration. Cultivation was performed by using minimal medium (8) with 3.7 mM succinate as a growth-limiting nutrient in a 2.2-liter bench-top BioFlo/Celligen 115 Modular BioReactor-Fermentor (New Brunswick Scientific) with a working volume of 2.0 liters, resulting in an optical density at 600 nm (OD600) of ∼0.63. The dilution rate was sustained at 0.163 h−1. Once the culture reached steady state, defined as maintaining a constant OD600 and a balanced flow rate for a minimum of three regenerations of the chemostat volume, the flows were stopped, and ethylamine was added to the succinate-limited steady-state culture to a final concentration of 20 mM. Cells were harvested at several time points for enzyme activity measurements, transcriptome sequencing (RNA-seq), and metabolomics analyses, as described below. For the substrate switchover experiment with overexpression of ecm via pCM80, 1.5 μg/ml tetracycline was added to retain the plasmid, and 13 μM cobalt chloride was added to the culture medium to negate the inhibitory effects of pCM80. The same concentration of cobalt chloride was added to the wild-type (WT) culture (23).
LC-MS and GC-MS metabolomics.Samples for gas chromatography (GC)-mass spectrometry (MS) and liquid chromatography (LC)-MS were collected by fast filtration using 20 ml of cell culture. Cells in the filter were frozen immediately with liquid nitrogen in a precooled 50-ml plastic tube. Samples were lyophilized for 2 h. Hot extraction of the metabolites was achieved by adding 20 ml of boiling water and boiling the samples for 10 min. Samples were then cooled on ice for 30 min and centrifuged for 20 min at 4,800 × g at 4°C. The supernatant was transferred into a new 50-ml tube and lyophilized for 2 days. The lyophilized product was resuspended in 1 ml of water and lyophilized for 1 day. Samples were resuspended in 60 μl of water and divided in two. One half was used for metabolomics analysis using two-dimensional gas chromatography. Samples were derivatized by using tert-butyldimethylsilyl trifluoromethanesulfonate (TBDMS). GC-MS measurements were used to analyze amino acid and organic acid content as described previously (8). The second half of the sample was analyzed by LC-MS to determine changes in the levels of CoA derivatives using a Waters Xevo LC-MS system, consisting of an Acquity ultraperformance liquid chromatography (UPLC) system and a Xevo triple-quadrupole mass spectrometer (Waters, Milford, MA). A CSH-C18 column (130 Å, 1.7 μm, and 2.1 mm by 100 mm) was used for liquid chromatography, with mobile phase A consisting of 25 mM ammonium acetate, 2% (vol/vol) acetic acid, and 1% (vol/vol) formic acid in water and mobile phase B consisting of 2% (vol/vol) acetic acid and 1% (vol/vol) formic acid in acetonitrile. The gradient consisted of 100% mobile phase A at the start, 80% mobile phase A at 2.5 min, 55% mobile phase A at 5 min, 5% mobile phase A at 6 min, 100% mobile phase A at 7 min, and 100% mobile phase A at 8 min, at a flow rate of 0.3 ml/min. All data were analyzed by using Masslynx Quanlynx Applications Manager software.
RNA-seq.Samples (50 ml) of cells were collected, and RNA was immediately extracted and digested with DNase I (Ambion), as described previously (16). RNA purity was analyzed by reverse transcription-PCR (RT-PCR) using iScript reverse transcriptase (Bio-Rad). RNA integrity was verified by using a 2100 BioAnalyzer (Agilent). rRNA depletion, library preparation, and Illumina Hi-Seq sequencing were carried out at the University of Washington (UW) High Throughput Genomics Center (HTGC). For each sample condition, two biological replicates were used. rRNA was depleted by using Ribo-Zero (Epicentre, Madison, WI). Each replicate was sequenced to a target depth of 20 million single-end 36-bp reads by using a multiplex strategy with an Illumina Hi-Seq 2000 sequencer (Illumina, San Diego, CA). Quality filtering of the raw read counts was performed by using CASAVA software (Illumina). The filtered data were processed as follows. The raw reads were aligned to the M. extorquens AM1 genome (MAGE database, 2 March 2012) by using BWA version 0.7.4-r385 with default parameters (24). The alignments were postprocessed into sorted BAM files with SAMTools version 0.1.19-44428cd (25). Reads were attributed to open reading frames (ORFs) by using the htseq-count tool from the HTSeq version 0.5.4p5 framework in the “intersection-nonempty” mode (26). The final abundance was measured in reads per kilobase of transcript per million reads mapped (RPKM) (27). Student t tests were used to assign significance with post hoc adjustment, according to the method of Storey and Tibshirani (28).
Enzyme assays.Cells for enzyme activities (50 ml) were harvested and processed as reported previously (8). A minimum of 2 biological replicates were assayed for each enzyme. Crotonyl-CoA reductase (Ccr) activity was measured, as reported previously (29), with 2 mM NADPH (8). Ecm and Epi activities were measured as reported previously (12), with minor changes. Briefly, the assay mixtures were quenched by boiling at 100°C for 5 min. The methylsuccinate concentration was quantified by GC-MS as described above. Forty-five micrograms Ecm and 3 to 10 μg of Epi were added to the reaction mixture as indicated in Fig. 4B. Ccr activity was measured as reported previously (8).
Calculation of hypothetical minimum enzyme activity.The standard approved equation correlating specific substrate (S) consumption (dS) per unit of time (dt) to growth rate u, dS/dt = (μ/Y) · X, was used as reported previously (30), with minor changes. For growth yield, Y, a value of 48% of bacterial cell dry mass as carbon was used (31). X represents 1 g (dry mass) of cells, and X, a value of 55% of cell dry mass as protein was used. Activity was calculated by accounting for 5 carbon atoms, since 2 molecules of acetyl-CoA are condensed and then carboxylated to generate ethylmalonyl-CoA.
Cloning and heterologous expression of ccr, ecm, and epi from M. extorquens AM1 in E. coli.The gene encoding crotonyl-CoA reductase (ccr) was amplified from M. extorquens AM1 chromosomal DNA by using the forward primer 5′-ATCATATGGCTGCAAGCGCAGCAC-3′, with the insertion of an NdeI restriction site, and the reverse primer 5′-ATGGATCCTCACATCGCCTTGAGCGGGAT-3′, with the insertion of a BamHI restriction site (restriction sites are italicized). The PCR product was generated by using Herculase II Fusion DNA polymerase (Agilent Technologies, Santa Clara, CA) according to the manufacturer's protocol, with 30 cycles including a denaturation step for 20 s, a primer annealing step at 65°C for 20 s, and an extension step at 72°C for 1.5 min. The PCR product was isolated and cloned into pET16B (EMD Millipore, Darmstadt, Germany), producing pNG221 for the expression of ccr and the production of N-terminal decahistidine-tagged fusion proteins. Competent Escherichia coli (BL21-AI) cells were transformed with pNG221 and grown at 37°C in LB medium containing 50 μg · ml−1 ampicillin. When the culture reached an OD600 of 0.6 to 0.7, expression was induced by the addition of 1.0 mM isopropyl thiogalactopyranoside (IPTG) and 0.2% (wt/vol) l-arabinose (Sigma). The cultures were grown for an additional 3.5 h at 30°C.
The gene encoding ethylmalonyl-CoA mutase was amplified by using the forward primer 5′-ACATATGAGCGCGCAAGCGAG-3′, with the insertion of an NdeI restriction site, and the reverse primer 5′-ATGGATCCGAAGACCTGCGCCC-3′, with the insertion of a BamHI restriction site (restriction sites are italicized). PCR was performed for 30 cycles, with 20 s of denaturation, 20 s of annealing at 65°C, and an extension step at 72°C for 2 min. The PCR product was isolated and cloned into pET16b to obtain pNG235 for the expression of ecm and the production of an N-terminal decahistidine-tagged fusion protein.
The gene encoding ethylmalonyl-CoA/methylmalonyl-CoA epimerase was PCR amplified by using the forward primer 5′-TCACAGCAGCGGCCATATCGAAGGTCGTCAATGATCGGACGGCTCAAT-3′ and the reverse primer 5′-CAGCTTCCTTTCGGGCTTTGTTAGCAGCCGTCAGACCTGCTCCAGCTC-3′, generating an insert with ends overlapping the pET16b insertion site. The pET16b backbone fragment was amplified by using the forward primer 5′-CGGCTGCTAACAAAGCCCGAAAGGAAGCTGA-3′ and the reverse primer 5′-TGACGACCTTCGATATGGCCGCTGCTGTG-3′. The backbone and insert fragments were assembled by using Gibson Assembly (NEB) to generate pNG252. This plasmid was transformed into BL21-AI cells for protein expression.
Cultures for the expression of ecm and epi were grown at 37°C in LB medium containing 100 μg · ml−1 ampicillin. For the expression of ecm, 1 ml of 20 mg/liter vitamin B12 was added per liter of LB medium (12, 32). Cultures were induced with 0.5 mM IPTG and 0.2% (wt/vol) l-arabinose (Sigma) when the OD600 was 0.6 to 0.9. The cultures were grown for another 3 h at 30°C. All cells were harvested by ultracentrifugation using a Sorvall RC-5B centrifuge (Thermo Fisher Scientific) at 7,500 rpm at 4°C and stored at −80°C until needed.
Purification of Ccr, Ecm, and Epi.All enzymes were purified as reported previously (9). Buffer A contained 50 mM Tris-HCl (pH 7.8) and 5 mM imidazole. Buffer B contained 50 mM Tris-HCl (pH 7.8), 200 mM NaCl, and 5 mM imidazole. Buffer C contained 50 mM Tris-HCl (pH 7.8) and 500 mM imidazole. Glycerol (15%, vol/vol) was added to the purified enzyme, which was stored at −80°C until needed.
Generation of expression mutant constructs.ecm was PCR amplified by using the forward primer 5′-ATAAGCTTCGGACAGGGAGTCGAGTG-3′, with the insertion of a HindIII restriction site, and the reverse primer 5′-ATTCTAGAGGCGGCCCTCAGAAGACC-3′, with the insertion of an XbaI restriction site (restriction sites are italicized). The PCR product was inserted into pCM80, generating pNG109, which was mated into M. extorquens AM1 as reported previously (33).
Microarray data accession number.RNA-seq information has been deposited in the National Center for Biotechnology Information Gene Expression Omnibus (GEO) database (http://www.ncbi.nlm.nih.gov/geo) under accession no. GSE63115.
RESULTS AND DISCUSSION
A metabolic perturbation strategy was used to assess the response of the assimilatory pathways. Chemostat cultures of wild-type M. extorquens AM1 were grown with a limiting succinate concentration at a growth rate of 0.163 h−1. After the culture reached steady state, the in- and outflows were stopped, and 20 mM ethylamine was added. This method achieves a rapid substrate switch, since a succinate-limited culture has a very low succinate concentration (8), forcing the metabolic network to acclimate to metabolize the new, available substrate. Samples were taken immediately before the substrate switch (time zero) and at several time points thereafter (2, 4.5, 9, 18, and 22.5 h). Culture densities (OD600), metabolite concentrations, RNA transcript levels, and specific enzyme activities were measured at these time points. These data were compared in order to generate a hypothesis for a metabolic control point in the EMC pathway.
Growth response to the switch from succinate to ethylamine.Immediately after the switchover of the growth substrate, the culture reproducibly entered a transient lag period, during which the culture density decreased from an OD600 of 0.065 ± 0.01 to an OD600 of 0.060 ± 0.02. After 9 h, the culture resumed growth at a rate of 0.051 h−1, which was similar to the growth rate of the batch culture in shake flasks, demonstrating the transient nature of the lag.
Gene expression response of ethylamine oxidation and central metabolic pathways.RNA-seq was used to assess the response of M. extorquens to the change in the growth substrate at the transcriptional level. The oxidation of ethylamine to acetaldehyde occurs via methylamine dehydrogenase (Mau), the conversion of acetaldehyde to acetate has been proposed to be carried out by an aldehyde dehydrogenase and/or aldehyde oxidase activity, and the conversion of acetate to acetyl-CoA has been proposed to be carried out by acetyl-CoA synthetase activity (Fig. 1) (15). The transcription levels of the genes encoding these enzymes began to increase by 2 h and peaked at 9 h, with these transcripts remaining highly upregulated for the duration of the experiment (data not shown). Peak increases ranged from 8- to 59-fold compared to the time zero transcript levels. At later time points, the transcriptomic profile for ethylamine oxidation was consistent with transcripts reported previously for batch cultures (15).
The TCA cycle is the primary assimilatory pathway utilized by M. extorquens growing on succinate, and carbon flux supports a growth rate of 0.163 h−1 (22, 34). For growth on ethylamine, however, M. extorquens splits acetyl-CoA flux between the TCA cycle (oxidation) and the EMC pathway (assimilation) (11), and the growth rate is much lower (0.051 h−1). As expected due to the downshift in the growth rate, TCA cycle genes were immediately downregulated (data not shown) between 2- and 8-fold 2 h after the substrate switch. Nine hours after the switch of the substrate, several TCA cycle genes were upregulated again but did not return to the same levels of expression as those during growth on succinate. On average, TCA cycle genes were downregulated ∼4-fold.
The EMC pathway genes, on the other hand, did not respond as a uniform module and showed multiple regulation patterns. The genes encoding the first two steps, β-ketothiolase (phaA) and acetoacetyl-CoA reductase (phaB), were initially upregulated 2- to 3-fold (2 h), but by 9 h, transcript levels of both genes decreased to near-initial levels (see Fig. S2 in the supplemental material). The genes encoding crotonase (croR), the third enzyme of the EMC pathway, and crotonyl-CoA reductase/carboxylase (ccr), the fourth enzyme, showed a pattern of initial downregulation with subsequent upregulation above the levels for the time zero point (Fig. 2; see also Fig. S2 in the supplemental material). The gene encoding the next enzyme, ethylmalonyl-CoA/methylmalonyl-CoA epimerase (epi), was the most differentially regulated of the EMC pathway genes. It was downregulated 6-fold at 2 h and 9-fold by 4.5 h. At 9 h, transcript levels increased 2-fold, and by 18 h, transcript levels had returned to the levels in succinate-grown cells, but the response was never greater than that at time zero (Fig. 2). A similar pattern was observed for the gene encoding the eighth enzyme, mesaconyl-CoA dehydratase (mcd) (Fig. 2). However, the two steps between epi and mcd showed very different responses. Transcript levels of ethylmalonyl-CoA mutase (ecm) were upregulated initially, peaked at 9 h, and then decreased 2-fold relative to the initial levels (Fig. 2). The expression of methylsuccinyl-CoA dehydrogenase (msd) did not change significantly over the course of the switchover. Likewise, the genes encoding the l-malyl-CoA/β-methylmalyl-CoA lyases (mclA1 and mclA2) displayed disparate expression patterns. Transcript levels of mclA1 initially increased nearly 4-fold (2 h) and remained upregulated at 4.5 h (2-fold), but by 9 h, they had returned to the levels observed for growth on succinate. Transcript levels of mclA2, on the other hand, decreased slightly initially and then increased >2-fold at 9 h and remained upregulated for the duration of the experiment. The expression of the gene encoding l-malyl-CoA thioesterase (mcl2) was initially downregulated 2-fold and remained at this level until 18 h after the substrate switch, when levels returned to those observed for growth on succinate. The genes encoding the two subunits of propionyl-CoA carboxylate (pccAB) both displayed the same expression pattern, being downregulated 4-fold at 2 h and 5-fold at 4.5 h and being upregulated 2-fold at 9 h. By 18 h, transcript levels had returned to the levels observed for growth on succinate. Transcripts of the methylmalonyl-CoA mutase gene mcmA responded in a manner similar to that of the transcripts of ecm, while transcripts of mcmB showed a pattern similar to that of mcl2 transcripts. The expression of the genes encoding succinyl-CoA transferase (sct) and succinate dehydrogenase (sdh) was downregulated until 18 h after the substrate switch, and expression levels returned to levels near those observed for growth on succinate by the end of the time course. This transcriptomic analysis suggests that M. extorquens undergoes substantial remodeling of the metabolic network to switch from growth on succinate to growth on ethylamine and that the genes of the EMC pathway are not uniformly regulated.
Transcriptional profile of EMC pathway genes during the transition from growth on succinate to growth on ethylamine, as measured by RNA-seq. Transcript levels for genes encoding enzymes for the conversion of crotonyl-CoA to mesaconyl-CoA are shown. Time (in hours) after the addition of ethylamine is indicated in grayscale. The RPKM was normalized to the read count at time zero (0 h). Error bars represent the standard deviations of data from two biological replicates.
Response of EMC pathway metabolites and glyoxylate.Metabolic intermediates of the EMC pathway in samples from the switchover time course were measured to determine whether any intermediates accumulated during the transient growth lag. Accumulation of a metabolite during the growth lag would indicate an imbalance in the production and consumption of the compound. The metabolites of the EMC pathway were measured either directly via LC-MS as CoA esters or, in the case of methylsuccinate, by hydrolyzing the ester and measuring the organic acid content via GC-MS. In the latter case, the CoA derivative was poorly detectable, but the acid form was detected in a more robust manner. By this method, we were able to determine relative concentrations of most EMC pathway intermediates as well as several other compounds of interest, such as glyoxylate. Relative concentrations were normalized to concentrations at time zero and compared across substrate switchover time points.
Ethylamine is oxidized to acetaldehyde, which is converted to acetyl-CoA, which is the entry molecule for the EMC pathway (Fig. 1). Acetyl-CoA levels over the course of the switchover experiment did not change significantly from the initial levels during growth on succinate (Fig. 3). Relative levels of β-hydroxybutyryl-CoA did not change within the first 2 h after the switch from succinate to ethylamine but increased >3-fold by 4.5 h, suggesting an imbalance of carbon flux at the branch point between the poly-β-hydroxybutyrate (PHB) cycle and the EMC pathway (Fig. 1). However, by 9 h, β-hydroxybutyryl-CoA levels had returned to their initial relative concentration, suggesting that the imbalance had been alleviated, even though the culture had not yet resumed growth. Relative levels of crotonyl-CoA decreased slightly after the switch from succinate to ethylamine.
Metabolite profile of the mid-EMC pathway during the transition from growth on succinate to growth on ethylamine. Concentrations of CoA intermediates were measured by LC-MS. Organic acid concentrations were measured by GC-MS. Time (in hours) after the addition of ethylamine is indicated in grayscale. 3-OHB-CoA, 3-hydroxybutyryl-CoA; ethylmalonyl-CoA, ethylmalonyl-CoA-derived butyryl-CoA (see the text). Error bars show the standard deviations of data from two biological replicates. Each sample represents a minimum of two injections.
The relative concentration of butyryl-CoA increased the most during the time course. Butyryl-CoA is produced from crotonyl-CoA by crotonyl-CoA reductase/carboxylase (Ccr) at a very low rate and only when bicarbonate/CO2 is absent (29). When bicarbonate/CO2 is present, ethylmalonyl-CoA is formed exclusively. Since air was constantly flowing into the bioreactors during the substrate switchover experiment and the cells produce CO2 during metabolism, it is unlikely that CO2 was limiting. Instead, it seemed possible that butyryl-CoA was formed via a spontaneous decarboxylation of ethylmalonyl-CoA during metabolite extraction. No signal above the background for ethylmalonyl-CoA was reliably detected during the substrate switchover, also suggesting that it was unstable during sample processing. A sample of pure ethylmalonyl-CoA was treated with the same hot water extraction process (see Materials and Methods), and both compounds were measured by LC-MS. After hot water treatment, the ethylmalonyl-CoA concentration decreased 4-fold, whereas the butyryl-CoA concentration increased 4-fold, demonstrating that ethylmalonyl-CoA was unstable during the extraction process and that the degradation generated butyryl-CoA (see Fig. S1 in the supplemental material). Based on these results, we interpreted the butyryl-CoA signal as being indicative of ethylmalonyl-CoA.
By 4.5 h after the growth substrate change, ethylmalonyl-CoA-derived butyryl-CoA began to accumulate (>2-fold), and by 9 h, its concentration increased >3.5-fold, suggesting a metabolic imbalance in the EMC pathway after the ethylmalonyl-CoA production step within the first 9 h. Relative levels of ethylmalonyl-CoA decreased at 18 and 22.5 h, concurrent with the culture resuming growth, but still remained higher than the levels during growth on succinate (time zero).
Two EMC pathway intermediates that follow ethylmalonyl-CoA are methylsuccinyl-CoA and propionyl-CoA (Fig. 1). As noted above, methylsuccinyl-CoA was hydrolyzed and measured as methylsuccinate. The concentrations of both intermediates decreased relative to their concentrations at time zero and returned to initial levels only at later time points. Another intermediate, mesaconyl-CoA, was not detected by our method, suggesting that the relative concentration of this compound did not increase greatly above background levels at any time point of the experiment. The relative level of glyoxylate decreased ∼2-fold at 2 h and did not increase significantly during the time course. This result corroborates the concept that the metabolic network is set in a way such that this toxic intermediate does not accumulate during perturbations (8).
The accumulation of ethylmalonyl-CoA and the decreased levels of the downstream metabolites methylsuccinate and propionyl-CoA suggest that carbon flux was restricted at ethylmalonyl-CoA during the growth lag.
Response of key C2 enzymes.Since ethylmalonyl-CoA accumulated during the transitory period between the substrate switchover and resumption of growth, we measured the activities of the two enzymes that utilize ethylmalonyl-CoA as a substrate. As shown in Fig. 1, the epimerase converts (2S)-ethylmalonyl-CoA to (2R)-ethylmalonyl-CoA. The ethylmalonyl-CoA mutase then converts (2R)-ethylmalonyl-CoA to (2S)-methylsuccinyl-CoA in an adenosylcobalamin (AdoCbl)-dependent manner. By measuring these activities, we could identify which enzyme was limiting and compare activity to transcripts, which would suggest whether the enzyme levels were regulated transcriptionally or posttranscriptionally. Since the substrates of these enzymes are not available commercially, S-ethylmalonyl-CoA was produced by using commercially purchased crotonyl-CoA and purified recombinant Ccr (see Materials and Methods). We first measured the rate of conversion of S-ethylmalonyl-CoA to methylsuccinyl-CoA, a reaction encompassing both the epimerase and mutase activities, in cell extracts from the substrate switchover experiment. The product was monitored by a GC-MS assay (see Materials and Methods). The conversion activity measured should reflect the rate-limiting enzyme at the time when the sample was taken. The activity profile over time (Fig. 4A) is similar to the RNA expression profile of ecm, suggesting that Ecm is the rate-limiting enzyme. To verify this result, purified recombinant Epi was added in excess to the 4.5-h cell extract, and activity did not increase (Fig. 4B). When purified recombinant Ecm was added in excess to the 4.5-h cell extract, activity increased nearly 2-fold, from 9 nmol · min−1 · mg−1 to 17 nmol · min−1 · mg−1. When both recombinant Epi and Ecm were added to the extract, activity increased >2-fold, to 14 nmol · min−1 · mg−1, which is not significantly different from the result when Ecm was added alone (P = 0.1, determined by Student's t test). Because β-hydroxybutyryl-CoA accumulated during the early time points, when activity was low, we also tested 5 to 400 μM concentrations of β-hydroxybutyryl-CoA for inhibition of the reaction, but no effect on activity was observed (data not shown). This result does not, however, eliminate the possibility of another small molecule regulating the reaction, and more studies are be needed to identify such an effector.
Ethylmalonyl-CoA-to-methylsuccinyl-CoA conversion. Specific activity was calculated by using the amount of methylsuccinate (hydrolyzed methylsuccinyl-CoA) produced from (2S)-ethylmalonyl-CoA. (A) Activity measured for each time point of the switchover from succinate to ethylamine as the substrate. (B) Activity measured in a pure-component system with only purified Epi and Ecm (+Epi+Ecm) and no extract; extract alone from the 4.5-h time point (ex); and extract with Epi (ex+Epi), Ecm (ex+Ecm), or both (ex+Epi+Ecm) added. GC-MS measurements are from a minimum of two injections per sample. Error bars represent the standard deviations of data from at least two biological replicates.
Since the conversions of (2S)-ethylmalonyl-CoA to (2R)-ethylmalonyl-CoA and (2R)-ethylmalonyl-CoA to (2S)-methylsuccinyl-CoA are catalyzed by reversible enzymes, ethylmalonyl-CoA and methylsuccinyl-CoA are assumed to be at equilibrium during steady state (35, 36). However, the switch from succinate to ethylamine creates a transient metabolic imbalance, as evidenced by the long growth lag and the change in the metabolite pools. A similar response was seen during the substrate switchover from succinate to methanol (8). Under this condition, when carbon begins to flow suddenly into the middle portion of the EMC pathway, the metabolite pools are no longer at equilibrium. Our data suggest that Ccr generates (2S)-ethylmalonyl-CoA much more rapidly than the combined activity of Epi/Ecm can consume it. In support of this hypothesis, the Ccr activity measured in the switchover samples was high (150 to 240 nmol · min−1 · mg−1) (data not shown) compared to the Epi/Ecm activity measured in extracts from the 4.5-h time point (9 nmol · min−1 · mg−1). The data shown in Fig. 4B suggest that the in vitro Epi activity is higher than that of Ecm, since added Epi does not increase activity. Msd activity is expected to be on the order of five times higher than that of Ecm (30), which would draw down the (2S)-methylsuccinyl-CoA pool and drive the Ecm reaction toward the product. Over time, the substrate and product pools should begin to reach a new equilibrium sufficient for normal growth on ethylamine. Since Ecm activity is lower than those of Epi and Msd, it should be the limiting step in the reaction sequence. As the cells reach a new metabolic set point while upregulating necessary enzymes to drive metabolism, the metabolite pools would return to equilibrium. Consistent with this model, ethylmalonyl-CoA pools return to baseline levels as the cells resume growth, beginning at 9 h after the switchover. The correlation of a growth lag with accumulation of ethylmalonyl-CoA and resumed growth with decreased ethylmalonyl-CoA concentrations coupled to the enzyme assay results suggested that Ecm is a metabolic control point.
The minimal Ecm activity needed for the growth rate observed was calculated to be 12 nmol · min−1 · mg−1 (see Materials and Methods). Our enzyme activity results show that at the time when this level of activity was reached (9-h time point), culture growth resumed. Since the Epi activity measured was higher than this level, the results support the model that Ecm activity, not Epi activity, was restricting the consumption of ethylmalonyl-CoA and preventing growth during the growth lag.
Overexpression of ecm removes control in the EMC pathway.The results described above suggested that Ecm was functioning as a control point during the transition from growth on succinate to growth on ethylamine, restricting the consumption of ethylmalonyl-CoA and the production of downstream intermediates during the lag period of the transition. To test this model, we overrode the control point by overexpressing ecm from a strong promoter (the mxaF promoter) that is highly expressed during growth on both succinate and ethylamine (16, 37). Ecm activity increased 4-fold (data not shown). We then repeated the switchover from succinate to ethylamine with the overexpression strain. The concentration of cobalt chloride in the medium was increased to 13 μM, which is known to reduce a vector-induced growth defect (23). Under these higher-cobalt growth conditions, the wild-type strain responded to the substrate switchover as we had observed previously with lower cobalt concentrations in the medium (data not shown). The growth response of the strain harboring the empty plasmid was similar to the response of the wild-type strain, with a lag in growth following the substrate switchover for 9 h and resumed growth corresponding to a 12- to 13-h doubling time (Fig. 5A). However, the strain overexpressing ecm reproducibly began growing after the 4.5-h time point (Fig. 5A). When growth resumed, the initial rate was 0.024 h−1. Between 8 and 9 h after the switchover, the growth rate of the culture increased to 0.071 h−1, 21% higher than that of the wild-type and control strains. In the control strain, ethylmalonyl-CoA accumulated over time, with concentrations peaking at ∼5-fold at 9 h and then decreasing when the culture resumed growth (Fig. 5B), a pattern similar to that of the wild-type strain (Fig. 3). In contrast, ethylmalonyl-CoA did not accumulate significantly in the strain overexpressing ecm.
Switchover from growth on succinate to growth on ethylamine in cells overexpressing ecm. (A) Growth response of the ecm-overexpressing strain. Squares, wild-type M. extorquens carrying pCM80_ecm; triangles, wild-type M. extorquens carrying pCM80. Time zero indicates when chemostat flows were stopped and 20 mM ethylamine was added to the bioreactor. The optical density (OD600) was measured at the plotted time points. Growth curves are representative of data from a minimum of two biological replicates. (B) Relative concentrations of ethylmalonyl-CoA during the substrate switchover. The ethylmalonyl-CoA concentration was measured by LC-MS. Measurements are normalized to values at time zero (0 h). Time is shown in hours. Black bars, wild-type M. extorquens carrying pCM80; white bars, wild-type M. extorquens carrying pCM80_ecm. Measurements were calculated from a minimum of two injections each. Error bars represent the standard deviations of data from two biological replicates.
Under these conditions, the growth lag decreased, and an earlier onset of growth in the overexpression strain correlated with a lack of accumulation of ethylmalonyl-CoA compared to the control. In addition, the pattern of growth for the switchover was different, with initial growth rates of the ecm overexpression strain being lower than those of both the wild type and the strain containing the empty vector. Later, growth of the ecm overexpression strain increased to a higher rate than that of the controls. This shift in growth might indicate a transient accumulation of a toxic intermediate, but the later, faster growth suggests that Ecm activity limits flux through the EMC pathway after growth occurs. This finding for C2 metabolism in a switchover experiment is consistent with a previous suggestion that the EMC pathway could contain a metabolic control point (8, 11).
One explanation is that if the required cofactor AdoCbl was present at subsaturating levels, Ecm would catalyze its reaction at a lower rate. However, the overexpression of ecm results in increased Ecm activity, suggesting that AdoCbl is not limiting.
We hypothesized that a metabolic control point in the EMC pathway could be in place to restrict the production of the toxic intermediate glyoxylate. However, metabolite analysis of the substrate switchover with the ecm-overexpressing strain did not show an accumulation of glyoxylate (data not shown). Another possibility is that Ecm functions to restrict carbon in the EMC pathway until downstream consumption enzymes are induced sufficiently to handle the full carbon flux. In support of this hypothesis, the genes encoding phosphoenolpyruvate carboxykinase (pckA), enolase (eno), pyruvate dehydrogenase (pdhABCD), the glycine cleavage system (gvcP), the phosphoserine pathway (serCAB), and malyl-CoA lyase (mclA2), all of which encode enzymes utilized in C2 metabolism, are not upregulated until 9 h after the substrate switchover. By limiting Ecm activity, the cell undergoes a longer growth lag than when this mechanism is bypassed but grows more quickly after the lag is finished. It is possible that in the dynamic growth environment of leaf surfaces and/or soil (normal niches for this bacterium), the metabolic trade-off may provide an overall growth advantage.
Summary.Determining how metabolic balance is restored after a metabolic perturbation in the ethylmalonyl-CoA pathway is fundamental to an understanding of the physiology of EMC pathway-containing bacteria such as the model methylotroph M. extorquens. Furthermore, insight into regulatory mechanisms of the EMC pathway could be vital to the engineering of such bacteria as a platform for producing value-added chemicals such as methylmalonyl-CoA, ethylmalonyl-CoA, mesaconic acid, methylsuccinic acid, and butanol (19–21). It is difficult to study the control of the EMC pathway during C1 metabolism due to the regulation of the upstream steps (8, 9, 38). However, the experimental approach used here, involving C2 metabolism, successfully bypassed these reported regulatory mechanisms for C1 oxidation and assimilation and allowed us to focus specifically on the EMC pathway. The results presented here indicate that during a transition from growth on succinate to growth on ethylamine, ethylmalonyl-CoA mutase transiently restricts carbon flux through the EMC pathway, and the regulatory mechanism is likely transcriptional.
ACKNOWLEDGMENTS
This work was funded by a grant from the DOE (DE-SC0006871).
We thank Mila Chistoserdova and Frances Chu for critical reading of the manuscript, Lisa Yuki-Bloch for her assistance with experiments, and Martin Sadilek for assistance with mass spectrometry.
FOOTNOTES
- Received 6 November 2014.
- Accepted 26 November 2014.
- Accepted manuscript posted online 1 December 2014.
- Address correspondence to Nathan M. Good, nategood{at}uw.edu.
Citation Good NM, Martinez-Gomez NC, Beck DAC, Lidstrom ME. 2015. Ethylmalonyl coenzyme A mutase operates as a metabolic control point in Methylobacterium extorquens AM1. J Bacteriol 197:727–735. doi:10.1128/JB.02478-14.
Supplemental material for this article may be found at http://dx.doi.org/10.1128/JB.02478-14.
REFERENCES
- Copyright © 2015, American Society for Microbiology. All Rights Reserved.
