Subpopulation-Specific Metabolic Pathway Usage in Mixed Cultures as Revealed by Reporter Protein-Based ¹³C Analysis

Abstract

Most large-scale biological processes, like global element cycling or decomposition of organic matter, are mediated by microbial consortia. Commonly, the different species in such consortia exhibit mutual metabolic dependencies that include the exchange of nutrients. Despite the global importance, surprisingly little is known about the metabolic interplay between species in particular subpopulations. To gain insight into the intracellular fluxes of subpopulations and their interplay within such mixed cultures, we developed here a ¹³C flux analysis approach based on affinity purification of the recombinant fusion glutathione S-transferase (GST) and green fluorescent protein (GFP) as a reporter protein. Instead of detecting the ¹³C labeling patterns in the typically used amino acids from the total cellular protein, our method detects these ¹³C patterns in amino acids from the reporter protein that has been expressed in only one species of the consortium. As a proof of principle, we validated our approach by mixed-culture experiments of an Escherichia coli wild type with two metabolic mutants. The reporter method quantitatively resolved the expected mutant-specific metabolic phenotypes down to subpopulation fractions of about 1%.

Consortia of microorganisms are responsible for diverse natural processes that range from the decomposition of organic matter or biodegradation of anthropogenic xenobiotics in ecosystems (36) to beneficial effects of commensal microbiotas in the mammalian gut (31). Directly or indirectly, these processes involve interactions of metabolic activities between species, such as syntrophy in anaerobic global carbon cycles for decomposition (20, 33) or complex interactions in the competitive environment of the gut microbiome (6, 32). In sharp contrast to this environmental, nutritional, and disease relevance of microbial communities, our methodological repertoire is geared mainly to the analysis of single-species culture. To gain insights into the metabolic status and operation of single species or subpopulations within such consortia, currently used approaches include (i) differential gene expression (11, 18), (ii) proteome analyses of metabolic enzymes either as a shotgun technique or with fluorescence-assisted sorting prior to analysis (4, 45), (iii) metabolite investigations without species separation (9, 19, 22, 44), and (iv) isotope probing combined with metagenomics (12, 37, 42), metabolite profiling (40), or imaging techniques (21, 41).

Although these approaches provide valuable metabolic information, they monitor only the inventory of the metabolic network and not its functional operation (25). This functional operation of the metabolic network is described by the intracellular fluxes (i.e., in vivo reaction rates) that integrate the response of all catalytic protein-metabolite and regulatory interactions at the genetic, posttranslational, allosteric, and kinetic levels (26). Most important for the study of metabolic status and operation, the energy, redox factor, and precursor generating intracellular fluxes in central carbon metabolism are required. These intracellular carbon fluxes are per se not measurable and must be estimated by model-based interpretation of measured data, for which several approaches exist.

Based on mass balances of measured substrate uptake rates with formation rates of biomass, by-products, and CO₂ within stoichiometric models, flux balance analysis typically provides a whole range of possible flux distributions, unless strong assumptions are made (5, 39). To avoid such assumptions, intracellular information that can be provided by ¹³C isotope labeling experiments is required. Here, the identification and quantification of specific metabolic pathways in vivo are based on unique ¹³C labeling patterns in metabolites that result from the metabolic distribution and rearrangement of the introduced ¹³C tracer (26, 35, 43, 46). The two complementary mathematical approaches, iterative isotopologue balancing and local flux ratio analysis, are used for such ¹³C-based metabolic flux analysis (47). Akin to flux balance analysis, iterative isotopologue balancing requires extracellular physiological rates and can thus be applied only to mixed populations when their individual rates are known. Flux ratio analysis, in contrast, relies solely on the intracellular ¹³C labeling patterns to calculate ratios of fluxes from converging pathways (14, 27, 34) and thus is potentially applicable to mixed populations for which extracellular rates are typically not available.

Conventional ¹³C flux ratio analysis is based on ¹³C labeling patterns in protein-bound amino acids in cellular biomass (47) and hence is not readily applicable to mixed cultures. To specifically resolve ¹³C labeling patterns of a given population within a mixed culture, either the populations have to be separated (13) or a unique protein representative for one population with selective purification is needed (28). For population separation, commonly either centrifugation or fluorescence-assisted cell sorting is used. The problem with the former approach is that centrifugation can separate species only with sufficient density differences and is not accurate enough for complex mixtures, while fluorescence-assisted sorting takes too long for a reliable ¹³C labeling pattern analysis to be done. Here, we develop a method for ¹³C flux analysis of subpopulations within mixed cultures using a plasmid-based reporter protein with glutathione S-transferase (GST) as the high-affinity purification tag and green fluorescent protein (GFP). For a proof of principle, we used cocultivation experiments of an Escherichia coli wild type and metabolic deletion mutants to define the methodological requirements for quantitative ¹³C flux analysis of subpopulations and to characterize the resolution power of the method.

MATERIALS AND METHODS

Bacterial strains, growth conditions, and media.

This study used E. coli K-12 MG1655 (German Collection of Microorganisms and Cell Cultures, Braunschweig, Germany) and its isogenic phosphoglucose isomerase (pgi protein) and malate dehydrogenase (mdh protein) deletion mutants (2). For the reporter protein plasmid, the published vector pGEX-KG, which carries already glutathione S-transferase (GST), was used, and green fluorescent protein (GFP) inserted after the polylinker (16). For all experiments, frozen glycerol stocks were used to inoculate 5 ml of Luria-Bertani (LB) medium with ampicillin and/or kanamycin at final concentrations of 50 mg·liter⁻¹ for metabolic mutants and strains containing a plasmid. After 5 h of incubation at 37°C and 300 rpm on a gyratory shaker, 5-ml volumes of M9 minimal media were inoculated at 500- to 2,000-fold dilutions as precultures. The mid-exponential-growth-phase M9 preculture at an optical density at 600 nm (OD₆₀₀) of 0.8 to 1.2 was then used to inoculate a 70-ml M9 batch monoculture in a 1-liter baffled shake flask to a maximal OD₆₀₀ of 0.03. For mixed cultures of E. coli strains, the LB and M9 precultures were prepared as monocultures. For the M9 batch coculture of the wild type and mdh mutant, two 70-ml volumes were inoculated with both strains at an OD₆₀₀ of 0.02. For the M9 batch coculture of the wild type and the pgi mutant, six 70-ml volumes were inoculated with the wild type at an OD₆₀₀ of 0.005 and with the pgi mutant at an OD₆₀₀ of 0.08, due to the growth rate of the pgi mutant being four times lower than that of to the wild type.

The filter-sterilized M9 minimal medium consisted of the following per liter of deionized water: 7.5 g of Na₂HPO₄·2H₂O, 3.0 g of KH₂PO₄, 1.5 g of (NH₄)₂SO₄, and 0.5 g of NaCl. The following components were filter sterilized separately and then added (per liter of final medium): 1 ml of 1 M MgSO₄, 1 ml of 0.1 M CaCl₂, 0.6 ml of 0.1 M FeCl₃, 20 ml of 25% (wt/vol) glucose, 2 ml of 1.5 mM thiamine, and 10 ml of a trace element solution containing (per liter) 180 mg of ZnSO₄·7H₂O, 180 mg of CoCl₂·6H₂O, 120 mg of MnSO₄·H₂O, and 120 mg of CuCl₂·2H₂O. For plasmid-containing strains, ampicillin was added to the M9 medium at a final concentration of 50 mg·liter⁻¹. For ¹³C flux analysis in batch cultures, either a mixture of 20% (wt/wt) [U-¹³C]glucose and 80% (wt/wt) naturally labeled glucose or 100% [1-¹³C]glucose was used (both at >99% isotopic purity; Cambridge Isotope Laboratories, Andover, MA). For M9 precultures of cocultivation experiments with E. coli strains, the same ¹³C-labeled glucose composition was used as for the M9 batch culture to avoid fractions with more than 3% unlabeled biomass at the time of harvest that interfere with ¹³C flux ratio analysis.

Determination of physiological parameters.

Supernatant samples were prepared by centrifugation of 1 ml culture broth for 3 min at 4°C and 14,000 × g to determine glucose and organic acids by the signals of a refractive index and diode array detector on a high-pressure liquid chromatographer (HPLC) (Agilent 1100), using a Aminex HPX-87H column at 60°C with 5 mM H₂SO₄ as an eluent. Cell growth was determined spectrophotometrically at 600 nm. The specific growth rate was calculated by linear regression versus time in monocultures. Growth curves of individual E. coli strains in mixed cultures were estimated based on start inocula and individually determined growth rates using the standard exponential growth equation. The fluorescence signal from the GFP fusion reporter protein in the culture broth was measured (Tecan Infinite M200).

Reporter protein production and purification.

Reporter protein production was induced with a 10 mM IPTG (isopropyl-β-d-thiogalactoside) stock solution at an OD₆₀₀ of 0.1 to 0.2 during exponential growth (29) at a final concentration of 0.05 mM. Batch-grown cultures were harvested during the mid-exponential growth phase at an OD₆₀₀ of about 2 by centrifugation for 1 min at room temperature and 15,500 × g. The biomass pellets were directly frozen in liquid nitrogen and kept at −80°C until reporter protein purification. For reporter protein purification, biomass pellets resulting from 70 ml culture broth were resuspended in 4 ml lysis buffer (50 mM Tris [pH 7.6], 100 mM NaCl, 1 mM MgCl₂, 2 mM dithiothreitol, and 4 mM phenylmethylsulfonyl fluoride) and disrupted by two passages through a French press cell at 4°C. Cell-free lysates were obtained by centrifugation at 23,000 × g for 10 min at 4°C and directly applied to 1 ml equilibrated glutathione Sepharose beads according to the manufacturer manual (GE Healthcare Europe, Switzerland). After incubation for 1 h on ice under gentle shaking, the beads were recovered by centrifugation at 800 × g at 4°C for 5 min and washed seven times with 5 ml lysis buffer without phenylmethylsulfonyl fluoride. To elute the reporter protein, the beads were incubated four times with 1 ml 15 mM fresh glutathione solution (reduced form) in 100 mM Tris [pH 7.6] at room temperature for 30 min.

To remove residual Tris buffer and glutathione, which can interfere with the derivatization for GC-MS analysis, 16 ml −20°C acetone was added to the pooled eluates for protein precipitation. After at least 1 h of incubation at −20°C, the solution was centrifuged at 15,000 × g for 10 min at −10°C, and the precipitated reporter protein resuspended in 400 μl deionized water and precipitated with 1.6 ml −20°C cold acetone. After three wash steps, the precipitated reporter was stored at −20°C or directly hydrolyzed in 1.5 ml 6 M HCl for 24 h at 105°C in sealed microtubes for gas chromatography mass spectrometry (GC-MS) analysis described below.

¹³C labeling pattern analysis of protein-bound amino acids by GC-MS and ¹³C-constrained metabolic flux ratio analysis.

Samples of protein-bound amino acids from whole-cell protein for GC-MS were prepared as described previously (47). Briefly, cell pellets were obtained by centrifugation of 1 ml culture broth for 3 min at 4°C and 14,000 × g, washed once with 1 ml 0.9% NaCl, and hydrolyzed in 1.5 ml 6 M HCl for 24 h at 105°C in sealed microtubes. After drying at 95°C, the samples were derivatized in 30 μl dimethylformamide (Fluka, Switzerland) and 30 μl N-(tert-butyldimethylsilyl)-N-methyl-trifluoroacetamide (Sigma, Switzerland) at 85°C for 1 h. For analysis of derivatized amino acids by GC-MS, an Agilent 5973 mass spectrometer coupled to an Agilent 6890 gas chromatograph and an Rtx-5Sil MS fused-silica column (10 m by 0.18 mm inside diameter, 0.18-μm film thickness, Restek U.S., Bellefonte, PA) was used. The oven gradient raised from 160°C with 20°C/min to a final temperature of 280°C.

The ¹³C labeling patterns were obtained by the software FiatFlux (48), with implemented peak integration, removal of faulty ¹³C labeling patterns, and correction for naturally occurring isotopes (47). Metabolic flux ratio analysis was based on previously established methods implemented in the used software FiatFlux (14, 47, 48). Briefly, corrected mass distributions of amino acids were mapped to the precursor metabolites in central carbon metabolism by a least-squares fit and locally interpreted by algebraic equations to calculate ratios of fluxes in convergent pathways. Standard deviations for flux ratios were determined from redundant mass distributions.

RESULTS

The reporter protein provides equivalent ¹³C labeling information to whole-cell protein.

To resolve ¹³C labeling patterns in subpopulations, we used a plasmid-based reporter protein, consisting of GST for highly selective purification by affinity chromatography (16) and GFP for monitoring of expression. For initial method validation, a tac promoter induced by the lactose analogue IPTG (1) was used to induce the reporter protein. First, we determined the optimal IPTG concentration that allows for sufficient reporter protein expression and unperturbed exponential growth to avoid physiological effects (29). For reporter induction in exponentially growing E. coli wild-type cultures, 0.05 mM IPTG was optimal for strong induction without significant growth perturbation (Fig. 1 A and B).

FIG. 1.

Batch growth of E. coli wild type containing the GST-GFP reporter plasmid upon increasing IPTG supplementation at an OD₆₀₀ of 0.1 (dashed line) (A) and the fluorescence signal resulting from the GFP-fusion reporter protein (B). (C) SDS-PAGE of the different protein purification steps. M, marker; 1 and 2, different sample volumes of cell extract; 3 and 4, different amounts of cell debris after disruption; 5, cell extract after loading of the affinity matrix; 6 to 12, wash steps; 13 to 15, glutathione elution steps; 16, empty.

Next, we determined whether the applied purification procedure yielded exclusively the reporter protein. For this purpose, we induced the reporter protein in 70-ml batch cultures of the E. coli wild type and compared aliquots from different protein purification steps by SDS gel electrophoresis (Fig. 1C). During purification, the affinity matrix turned green and yielded a green eluate that consisted of a single protein band in the SDS-PAGE gel with the expected reporter protein size of 51 kDa. Thus, we concluded that a pure reporter protein was obtained with a yield of about 1 mg dry reporter protein from 70 ml E. coli batch culture harvested at an OD₆₀₀ of 2, which corresponds to a total biomass dry weight of about 50 mg.

Having now a potentially suitable reporter protein system, we next verified that the ¹³C label information of the reporter protein was equivalent to the label information of the whole-cell protein. Hence, we purified the reporter protein from two 70-ml batch cultures of the E. coli wild type grown on either 20% (wt/wt) [U-¹³C]glucose and 80% naturally labeled glucose or 100% [1-¹³C]glucose. After purification, the GC-MS-derived ¹³C labeling patterns of hydrolyzed reporter protein from 70-ml batch cultures were compared to those of the whole-cell protein hydrolyzate from a 1-ml culture aliquot of the same cultures. For a comprehensive comparison of ¹³C labeling patterns, we determined intracellular ratios of converging fluxes to yield information about key pathways in central carbon metabolism (14, 47) (Fig. 2). The obtained flux ratios were highly similar for both ¹³C labeling pattern sources and matched previously published results for E. coli wild-type batch cultures with glycolysis as the main route for glucose breakdown (cf. serine derived through glycolysis), absent gluconeogenesis (cf. phosphoenolpyruvate [PEP] originating from oxaloacetate), and almost equal contributions of tricarboxylic acid (TCA) cycle and anaplerosis to oxaloacetate (cf. oxaloacetate originating from PEP) (14, 24). Hence, we concluded that the reporter protein is a reliable source of ¹³C labeling patterns and equivalent to whole-cell protein for determination of metabolic pathway activity as was shown previously (28).

FIG. 2.

(A) ¹³C-determined ratios of converging fluxes calculated from ¹³C labeling patterns in amino acids of whole-cell protein or purified reporter protein of E. coli wild-type batch grown on either 20% [U-¹³C]glucose and 80% (wt/wt) naturally labeled glucose or 100% [1-¹³C]glucose (indicated by the asterisk). Standard deviations were calculated by linearized error propagation (14, 48). (B) The directly calculated flux ratio values (black underlined numbers) at the resolved nodes and their inferred complementary values (gray numbers). Abbreviations: PP pathway, pentose phosphate pathway; ED pathway, Entner-Doudoroff pathway; TCA cycle, tricarboxylic acid cycle; ub, upper bound; lb, lower bound.

The reporter protein allows for identification of metabolic mutant-specific ¹³C patterns in cocultivation experiments.

To examine the general applicability of the reporter protein to mixed cultures, we cocultivated the E. coli wild type with the pgi or mdh deletion mutant containing the reporter plasmid. To verify the mutant-specific metabolism, we determined flux ratios in separate batch cultures on a mixture of 20% (wt/wt) [U-¹³C]glucose and 80% naturally labeled glucose (Fig. 3 A). In contrast to the primary glucose catabolism via glycolysis in the wild type, the pgi mutant showed the expected rerouting through the pentose phosphate (PP) pathway (14). Also for the mdh mutant with an impaired TCA cycle, the local flux rerouting with an increased relative anaplerotic flux compared to the wild type was confirmed (Fig. 3A). Since the mdh deletion can be partially replaced by the isoenzyme malate:quinone-oxidoreductase (38), an anaplerotic flux significantly lower than the theoretically expected value of 100% was observed.

FIG. 3.

(A) ¹³C-determined ratios of converging fluxes calculated from ¹³C labeling patterns in amino acids of the purified reporter protein obtained from separate E. coli wild-type and pgi and mdh deletion mutant batch cultures grown on a mixture of 20% [U-¹³C]glucose and 80% naturally labeled glucose. The affected enzymatic reactions of the pgi and mdh deletion mutants in the metabolic network of E. coli are highlighted by black boxes. Cocultivation experiments of E. coli wild type with the pgi deletion mutant (B) or mdh deletion mutant (C) containing the reporter plasmid and grown on a mixture of 20% [U-¹³C]glucose and 80% naturally labeled glucose. Reporter protein production was induced upon addition of 0.05 mM IPTG at an OD₆₀₀ of 0.1. Dashed lines represent growth curves of each strain estimated from inocula and individual growth rates from monocultures. The black lines are the sums of those estimated growth curves. Black circles mark the time of harvest of the complete culture with final population fractions of 17% for the pgi mutant and 49% for the mdh mutant. (D) ¹³C-determined ratios of converging fluxes calculated from ¹³C labeling patterns in amino acids of the purified reporter protein from the cocultivation experiments described for panels B and C. Abbreviations: G6P, glucose 6-phosphate; GAP, glyceraldehyde 3-phosphate; E4P, erythrose-4-phosphate; F6P, fructose 6-phosphate; FBP, fructose 1,6-bisphosphate; PGA, phosphoglycerate; R5P, ribose 5-phosphate; Ru5P, ribulose 5-phosphate; S7P, sedoheptulose 7-phosphate; X5P, xylose 5-phosphate.

Next, we tested whether these mutant-specific flux ratios could be resolved by a reporter-based ¹³C flux analysis in cocultivation experiments. Each mutant containing the reporter plasmid was grown in a mixed batch culture with the wild type on a mixture of 20% (wt/wt) [U-¹³C]glucose and 80% naturally labeled glucose (Fig. 3B and C). The determined overall growth curves could be well described as a superposition of two independently, exponentially growing populations calculated from their inoculum and previously determined growth rates in monocultures (i.e., pgi mutant, 0.21 h⁻¹; mdh mutant, 0.50 h⁻¹; and wild type, 0.71 h⁻¹). Thus, a pseudo-steady state could be assumed for both cocultures, and the ¹³C labeling patterns were derived from the individual purified reporter proteins in the two mutants. Since the thereby-derived flux ratios matched favorably with the ones obtained from separate mutant cultures (Fig. 3A and D), we concluded that the reporter protein can be used to investigate the metabolic pathway usage of a specific population in cocultivated, mixed cultures.

Subpopulation resolvability with the reporter protein depends mainly on the purified protein amount.

To elucidate the resolvability of metabolic pathway activity in mixed cultures with various subpopulation fractions, we prepared fractions of 0%, 1%, 5%, 10%, and 100% of induced E. coli pgi mutant batch cultures containing the reporter plasmid together with the E. coli wild type in a total volume of 70 ml. Both cultures were grown separately to the same OD₆₀₀ value and mixed in the various proportions before cell disintegration. To focus on one key pathway, we investigated the relative PP pathway flux by using a 100% [1-¹³C]glucose experiment, and ¹³C labeling patterns were derived from the purified, hydrolyzed reporter protein and whole-cell protein, representing the culture average (Fig. 4).

FIG. 4.

Dependency of the ¹³C-determined split ratio between glycolysis and PP pathway on the relative proportion of the E. coli pgi deletion mutant and wild type in artificially mixed cultures determined from ¹³C labeling patterns in amino acids of the purified reporter protein (black line) or whole-cell protein (gray line) as the culture average. Separate batch cultures were grown on 100% [1-¹³C]glucose and mixed to the indicated biomass fractions in a total volume of 70 ml. We used, as controls for the relative PP pathway activity, purified reporter proteins from 70-ml monocultures of the E. coli wild type and pgi deletion mutant. Errors for flux ratios were within 2% based on linearized error propagation (14, 48).

As expected, the culture average displayed an increase in the relative PP pathway activity in proportion to the fraction of the pgi mutant population in the overall culture. The reporter protein clearly resolved the increased relative PP pathway activity in the pgi mutant down to a population fraction of 1%, which corresponds to a volume as small as 0.7 ml for the pgi mutant in 70 ml total culture volume. Presumably, even smaller pgi mutant fractions could be clearly distinguished from the culture average based on the hyperbolic curve for the reporter protein-derived flux ratios. With this hyperbolic curve, the increased relative PP pathway activity should be traceable in theory to a population fraction of 0.25% for the pgi mutant, for which a PP pathway ratio of about 38% could be clearly distinguished from the culture average.

Nevertheless, the hyperbolic curve also indicated an underestimation of the relative PP pathway activity in the pgi mutant, which should be at a value of 100% for the PP pathway ratio (Fig. 4). Such an underestimation was not observed in the case of the cocultivation experiment of the E. coli wild type and pgi mutant, in which a 17% pgi mutant culture fraction was unambiguously identified with the same PP pathway ratio as that of the pgi mutant monoculture (Fig. 3A and D). The major difference between the subpopulation resolvability experiment and the cocultivation experiment were the used culture volumes. While in the subpopulation resolvability experiment, 0.7 ml pgi mutant culture broth was used in a 70-ml total culture volume, the cocultivation experiment used larger volumes, with a ca. 70-ml pgi mutant culture volume in a 410-ml total culture volume, corresponding to a 17% pgi mutant culture fraction. Based on these results, we concluded that the underestimated PP pathway ratio values in the subpopulation resolvability experiment were caused by unspecific binding of wild-type proteins to the affinity matrix and can be significantly reduced by greater reporter protein amounts for purification. Thus, for quantitative ¹³C flux analysis, 1 mg of purified reporter protein obtained from a 70-ml subpopulation volume, corresponding to a ca. 50-mg subpopulation biomass dry weight, is optimal with the used purification protocol.

DISCUSSION

We present here a reporter protein-based ¹³C method to quantitatively resolve subpopulation-specific intracellular fluxes. For a proof of principle, we unambiguously identified the E. coli pgi and mdh mutant-specific metabolic phenotypes in cocultivations with the wild type. The ability to resolve populations down to fractions of 1% renders the method applicable to microbial consortia with few species, such as (i) certain natural phototropic consortia (23), (ii) enriched microbial communities for degradation/removal of anthropogenic substances (7, 17, 30, 45), or (iii) mixed starter cultures in food biotechnology (8, 10). Provided that ¹³C-labeled substrates are directly taken up by the species of interest, this method allows us to quantify species-specific intracellular flux patterns within the consortium.

The current limit for this type of quantitative ¹³C flux ratio analysis is the requirement of about 50 mg of subpopulation dry weight to obtain about 1 mg of purified reporter protein. Should that be difficult to achieve for particular applications, a further reduction of subpopulation biomass would be possible by (i) reducing the affinity matrix volume to decrease unspecific binding, (ii) including a size exclusion chromatography step after affinity purification to obtain only the fluorescent fraction for subsequent GC-MS analysis, and (iii) optimizing GC-MS analysis with large-volume injection techniques (3) to obtain reliable ¹³C labeling patterns with only 0.1 mg of reporter protein. The size exclusion chromatography step would not only significantly reduce unspecific binding to the affinity matrix but also reduce glutathione binding proteins from other populations based on the size difference of the reporter protein and its GFP fluorescence signal that can be used for a simple assay.

While the resolution of intracellular fluxes would be desirable, one key metabolic question in microbial consortia is typically much more trivial (15): which nutrients does a particular species or subpopulation actually consume? This question is hard to address directly because mixed cultures thrive in or generate themselves often complex nutritional conditions with several possible carbon substrates. Without ab initio knowledge of the consumed substrates, ¹³C flux analysis cannot resolve intracellular fluxes in subpopulations. Here, our reporter protein method can be used as a discovery tool to identify the consumed or exchanged carbon substrate for subpopulation growth. This could be achieved, for example, by adding occurring substrates as the ¹³C-labeled tracers and detecting ¹³C-labeled enrichment in the amino acids of the reporter protein. The IPTG induction used here could potentially cause physiological changes in consortia, e.g., changes in the population distribution. Hence, the experimenter should choose an appropriately inducible promoter or use instead a constitutive, species-specific promoter for the reporter protein. By choosing a naturally induced, species-specific promoter, one can directly study the metabolic response that activates this promoter, e.g., different nutrients or stress conditions.