Engineering in Life Sciences RESEARCH ARTICLE Stability of a Mutualistic Escherichia coli Co-Culture During Violacein Production Depends on the Kind of Carbon Source Simon Schick1 Tobias Müller2 Ralf Takors2 Georg A. Sprenger1 1Institute of Microbiology, University of Stuttgart, Stuttgart, Germany 2Institute of Biochemical Engineering, University of Stuttgart, Stuttgart, Germany Correspondence: Georg A. Sprenger (georg.sprenger@imb.uni-stuttgart.de) Received: 4 March 2024 Accepted: 21 August 2024 Funding: This study was funded by the German Research Foundation (DFG) under research project (Project number: 427751089, SP503/11-1, and TA 241/12-1, respectively) within the priority program InterZell (SPP2170). Keywords: co-culture | D-xylose | E. coli | mutualism | violacein ABSTRACT The L-tryptophan–derived purple pigment violacein (VIO) is produced in recombinant bacteria and studied for its versatile applications. Microbial synthetic co-cultures are gaining more importance as efficient factories for synthesizing high-value compounds. In this work, a mutualistic and cross-feeding Escherichia coli co-culture is metabolically engineered to produce VIO. The strains are genetically modified by auxotrophies in the tryptophan (TRP) pathway to enable a metabolic division of labor. Therein, one strain produces anthranilate (ANT) and the other transforms it into TRP and further to VIO. Population dynamics and stability depend on the choice of carbon source, impacting the presence and thus exchange of metabolites as well as overall VIO productivity. Four carbon sources (D-glucose, glycerol, D-galactose, andD-xylose) were compared. D-Xylose led to co-cultures which showed stable growth andVIO production, ANT-TRP exchange, and enhancedVIO production. Best titers were∼126mgL–1 in shake flasks. The study demonstrates the importance and advantages of a mutualistic approach in VIO synthesis and highlights the carbon source’s role in co-culture stability and productivity. Transferring this knowledge into an up-scaled bioreactor system has great potential in improving the overall VIO production. 1 Introduction Microbial cells such as Escherichia coli offer an environmentally and economically friendly alternative in the synthesis of high- value chemicals and pharmaceuticals [1, 2]. Several carbon sources from sustainable origins can be utilized by E. coli. In addition to glucose, other sugars include the hemicellulose constituents D-xylose or L-arabinose, D-galactose (a moiety of lactose frommilk whey), or glycerol (a by-product from biodiesel production). Among others, E. coli is well known for the indus- trial production of aromatic amino acids [3] or derivatives of aromatic amino acids [4]. For the latter, the introduction of heterologous genes is often necessary. Violacein (VIO) is a low-molecular-weight purple pigment and a natural bisindole product derived from L-tryptophan biosynthesis [5–7]. It displays interesting biological properties as it acts as an antimicrobial compound versus many bacte- ria, protists, or viruses [7–9]. There is a plethora of further applications of VIO in cosmetics or textile production [10, 11]. VIO (and side products like deoxyviolacein [DVIO]) is naturally produced by Gram-negative bacteria such as Chro- mobacter violaceum [12, 13] or Janthinobacterium sp. [14, 15] which are psychrophilic microorganisms [16, 17]. Both, how- ever, are considered potential pathogens and therefore are not amenable to industrial production of the compounds [7]. Abbreviations: ANT, anthranilate; DVIO, deoxyviolacein; TRP, L-tryptophan; VIO, violacein. This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited. © 2024 The Author(s). Engineering in Life Sciences published by Wiley-VCH GmbH. Engineering in Life Sciences, 2024; 24:e202400025 https://doi.org/10.1002/elsc.202400025 1 of 10 https://doi.org/10.1002/elsc.202400025 https://orcid.org/0000-0002-9696-0254 https://orcid.org/0000-0002-7879-8978 mailto:georg.sprenger@imb.uni-stuttgart.de http://creativecommons.org/licenses/by/4.0/ https://doi.org/10.1002/elsc.202400025 Summary ∙ Co-cultures of twomutualisticE. colimutant strains can be successfully used to produce violacein fromvarious carbon sources. ∙ Best carbon source is D-xylose can be derived from sustainable plant biomass. ∙ Care must be taken to use the optimal starting ratios of both strains to achieve stable bacterial populations and productivity. Heterologous expression of the vioABCDE genes encoding enzymes for the oxygen-dependent transformation of two tryp- tophan (TRP) molecules into VIO (or DVIO) has already allowed production with recombinant microbial cells such as yeast [18] or Corynebacterium glutamicum [19]. The main recombinant producer, however, is E. coli [20–30]. The predominant approach forVIOproductionwith recombinant E. coli cells is the biosynthesis in one producer strain with improved TRP production to provide this as a precursor for VIO [21, 24, 27, 28, 31]. In recent years, the idea of microbial co- cultures with the metabolic division of labor in the production of chemicals has gained considerable interest [32–40]. Our group has recently shown that a stable and mutualistic co-culture of two E. coli strains could be obtained [41]. Here, one strain had a genetic block in the trpD gene and formed anthranilate (ANT) from glucose. ANT was excreted and served as an auxotrophic compound for the second strain which is unable to synthesize ANT on its own (trpE mutation) but can convert ANT to TRP. Excreted TRP then served to complement the growth defect of the first strain. Thus, these strains mutually benefit from each other in an obligatory interdependence [41]. We nowwent on and studied whether this system may be exploited for the production of the TRP-based VIO. Almost all recombinant E. coli VIO producers had been grown in monocultures and at temperatures below E. coli’s optimal growth temperature of 37◦C. This is due to the specific demands of VIO production genes which stem from psychrophilicmicroor- ganisms [13, 42]. Recently, by the engineering of the ribosome binding sites, VIO production with E. coli from LB media was found to be improved at higher temperatures (30◦C and 37◦C) [42]; these authors did not report on minimal media with single carbon (C) sources. While the present study was conducted, another group [22] reported on a modular co-culture for VIO synthesis. As C sources, several sugars have been reported such as glucose [24], L-arabinose [21], glycerol [23], and recently D-galactose (with TRP feeding) [30]. We show below that D-xylose— which is from sustainable origins—may be preferable over other C sources for a co-culture production of VIO with E. coli strains in a one-pot shake flask regime. Care should be taken to choose the right ratio of the two partner strains in a co-culture to avoid an unwanted collapse of the co-culture, however. TABLE 1 List of strains and plasmids. Name Relevant genotype Source Strains E. coli K-12 LJ110 W3110 fnr+ [46] ANT-3 LJ110 ΔtrpD ΔtnaA::FRT ΔtrpR::FRT-KmR-FRT [41] ANT-3a LJ110 ΔtrpD ΔtnaA::FRT ΔtrpR::FRT This study ANT-4 ANT-3a ΔlacZ This study ANT-5 ANT-4 ΔaraBAD::FRT This study TRP-3 LJ110 ΔtrpE ΔtrpR::FRT ΔtnaA::FRT [41] TRP-4 TRP-3 Δfuc::Ptac-vioD This study TRP-5 TRP-4 ΔaraBAD This study Plasmids pCas repA101(ts) km Pcas-cas9 ParaB-Red lacIq [43] pTarget-vioD cat sgRNA-Δfuc::Ptac-vioD This study pTarget-araBAD cat sgRNA-ΔaraBAD This study pKD46 PBADgam-bet-exo [44] pCP20 flp [45] pCO1 KmR [47] pVioABCE-km PBAD-vioABCE KmR [21] pVioempty-km KmR This study 2 Materials andMethods 2.1 Strain Construction Strains and plasmids of this study are listed in Table 1. Respective primers are listed in the supporting information (Table S1). The basis for the VIO-producing co-culture is themutualistic E. coli— E. coli co-culture (ANT-3 and TRP-3) as described previously [41]. Therein, ANT-3 produces ANT and is auxotrophic for TRP, while strain TRP-3 is also auxotrophic for TRP but can be supplemented aswellwithANT to satisfy its need for TRP. Furthermodifications were implemented via a plasmid-based CRISPR/Cas9 method as described earlier [43]. The deletion of chromosomal genes for L-arabinose degradation (ΔaraBAD::FRT) was created by the lambda-red-based recombineering method [44]. Flippase recog- nition target (FRT)-flanked kanamycin cassetteswere removed by the transformation of pCP20 [45]. ANT-5 was created by deleting the lacZ and araBAD genes. The partner strain TRP-5 received a chromosomally integrated vioD gene (from C. violaceum ATCC 12472, codon-optimized for E. coli K-12) replacing parts of the fuc locus, in addition to the araBAD deletion. For VIO production, the TRP-5 strain was transformed with pVioABCE-km [21], comprising the vioABCE genes from C. violaceum, while ANT-5 was transformedwith pVioempty-km to allow the use of kanamycin in the medium. Genetic changes were verified by PCR (gene deletions) and DNA sequencing (gene insertions). All strains were stored in glycerol stocks at −70◦C. 2 of 10 Engineering in Life Sciences, 2024 2.2 Cultivation andMedia Composition 2.2.1 Culture Medium All strains were cultivated in minimal medium (MM) [48] with 20 mg L−1 thiamin-hydrochloride and 5 g L−1 of carbon source (either glucose, glycerol, galactose, or xylose), adjusted to pH 7.0 with 1 mol L−1 HCl. Kanamycin was added to a final concentration of 0.05 g L−1 to prevent plasmid loss. For VIO production, 20 mmol L−1 L-arabinose and 0.1 mmol L−1 IPTG were supplemented as inducers of the vio genes at the beginning of the main cultivation. 2.2.2 Shake Flask Cultivation Pre-culture flasks (20 mLMM in a 250 mL shake flask with a cot- ton plug) were inoculated from a single colony of either ANT-5 + pVioempty-km or TRP-5 + pVioABCE-km from a MM agar plate. As these strains are auxotrophic, 0.1 g L−1 ANT (for TRP-5) or TRP (for ANT-5) were supplemented to the medium. Incubation took place at 30◦C and 200 rpm for 18–42 h to reach an optical density of∼4 at 600 nm (OD600). For the main co-culture in flasks (50 mL MM in a 500 mL shake flask with a cotton plug), cell suspensions of both pre-cultures were washed with 500 µL MM to remove residual ANT or TRP and inoculated with a combined OD600 of 0.2. The initial inoculation ratio (ANT-5/TRP-5) was determined as shown: 50/50 (OD600 of 0.1/0.1), 70/30 (OD600 of 0.14/0.06), and 30/70 (OD600 of 0.06/0.14). Incubation took place at 30◦C or 25◦C (for optimal VIO production) and 200 rpm for up to 120 h in a shaking incubator until the carbon source was depleted. If not indicated otherwise, all co-cultures were performed in triplicates. 2.3 Strain Ratio Identification To monitor the strain ratio of ANT-5 and TRP-5 within the cultivation flask, cell suspensions were plated out onMacConkey agar plates with 10 g L−1 lactose and incubated at 37◦C for 24 h. The lacZ deletion in ANT-5 leads to pale colonies as the strain cannot degrade the lactose, while TRP-5 forms deep red colonies (see Figure S1). The ratio was determined by counting 100– 200 pale and red colonies on each plate (minimum: 50 colonies counted). 2.4 Violacein Extraction For the extraction of VIO, method adapted from Rodrigues et al. [23], the cell pellet was mixed with 1 mL of pure ethanol and disrupted with glass beads under eight-shaped shaking move- ments (Silamat S6, Ivoclar: 4500 rpm, 30 s). The glass beads and cell debris were removed by centrifugation (14,000 rpm, 1 min) and the supernatant (VIO extract) was collected for analytics. This step was repeated until no violet color was visible in the supernatant. The supernatant of the cell culture (ANT, TRP, and carbon sources) or the supernatant of extracts (VIO) were used for all HPLC analysis. 2.5 Analytical Methods 2.5.1 Determination ofOpticalDensity (OD600) andCell Dry Weight The growth of the strains was monitored by photometric mea- surement (Cary 60 UV-Vis, Agilent) at OD600 during cultivation. The cell dryweight (CDW)was determined by a correlation factor between OD600 and CDW, as calculated in previous work [41]. 2.5.2 Violacein Determination by HPLC VIO analytics were performed by an HPLC method described earlier [20]. Therefore, an UHPLC (Dionex UltiMate 3000, Thermo Fisher Scientific) with a C-18 column (Luna C18(2), 5 µm, 250 × 4.6 mm, Phenomenex) was used. The temperature of the column oven was at 30◦C and the samples were cooled to 5◦C in the sampler unit. The injection volume of the VIO sample was 10 µL. An isocratic flow of 50% (v/v) ethanol (mobile phase) with a flow rate of 0.5 mL min−1 over 20 min was used. A diode array detector (DAD-3000) was used at a wavelength of 258 nm, one of the absorption maxima of VIO [5]. As external standards commercial VIO (CaymanChemical Company) andDVIO (Santa Cruz Biotechnology) were used, dissolved in pure ethanol. Note that the quantification of VIO in this work is displayed by the combined concentrations of VIO and DVIO. 2.5.3 Anthranilate and Tryptophan HPLC The quantification of ANT and TRP concentrations was per- formed similarly by an HPLC method (gradient see Figure S2) used in a previous publication [41]. 2.5.4 Sugar HPLC The consumption of carbon sources during the cultivation was measured via sugar HPLC method adapted from Nargesi et al. [49]. The HPLC (Infinity 1260, Agilent Technologies) with an organic acid column (Organic Acid-Resin, 300 × 8 mm, Chromatographie-Service GmbH) was used. The temperature of the column oven was at 40◦C, the sampler unit was cooled to 10◦C. The injection volume of the sugar samples was 5 µL. An isocratic flow of 5 mmol L−1 sulfuric acid (mobile phase) with a flow rate of 0.8mLmin−1 over 25minwas used. A refractive index detector (RID 1260) was used for the sugar detection. External commercial standards were used for calibration. 2.6 Calculation of Rates and Yields The mean biomass-specific production rates of compound i 𝑞i [mg (g h)−1] or [mg (OD600 h)−1] was calculated based on the concentration difference of the produced compound Δ𝑐i[mg L−1] between two measurements, related to the mean biomass in that interval (optical density 𝑐X [OD600] or concentration 𝑐X [g L−1]; correlation factor as described recently [41]) and the time 3 of 10 difference Δ𝑡 [h]. 𝑞i = Δ𝑐i 𝑐X ⋅ Δ𝑡 For the calculation of the specific VIO production rate, HPLC- quantified VIO concentrations were used. The VIO production was related to the total co-culture population (ANT-5 and TRP- 5). The L-tryptophan supply rate of strain TRP-5 for ANT-5 was calculated based on the estimated requirement of this amino acid for ANT-5 to realize the growth that was experimentally determined (assumption: 72% protein CDW−1 [g g−1] [50] and an L-tryptophan protein fraction of 1.1% [51]). The estimated L-tryptophan production was related to the mean biomass con- centration of the producer strain TRP-5. The VIO yield from substrate YVIO∕S [mg g−1] in a considered time period was calculated based on the concentration difference of produced VIO Δ𝑐VIO[mg L−1] related to the concentration difference of the carbon source used Δ𝑐S[g L−1] in that interval. YVIO∕S = Δ𝑐VIO Δ𝑐S 3 Results 3.1 Considerations for Strain Construction and Implementation We adopted our successful approach of using two E. coli strains, one with a genetic block (a partial deletion in the trpD gene) that leads to excretion of ANT, and the other which is ANT- auxotrophic but can produce and excrete TRP (deletion of trpE gene). The strains were genetically further developed by intro- ducing deletions of the trpR (for the TRP repressor) and tnaA (for tryptophanase) genes to avoid undesired effects of repression by TRP or degradation of TRP (see Table 1) [41]. For VIO production, the TRP producer strain was chromosomally equipped with vioD gene (under Ptac control) and with a plasmid carrying vioABCE genes fromC. violaceum [21]. To allow a co-culture in the presence of the antibiotic kanamycin (selectionmarker for the pVioABCE- km plasmid), the ANT producer was transformed with the empty vector (pVioempty-km) (see Figure 1). To distinguish the two strains in a mixed culture and to be able to determine the ratio of the two strains in a co-culture, the ANT producer was made lactose-negative (ΔlacZ). The relative abundance of the two strains in co-culture could thus be assessed by taking samples from the different growth phases and plating them out (after appropriate dilution) on MacConkey agar plates with lactose (1%). The ANT producers developed pale colonies, whereas the Lac+ TRP producers formed red colonies (see Figure S1). As the vioABCE genes are under the control of a ParaBAD promoter, L-arabinose was added to the culture media for induction; to avoid the further catabolism of this pentose, the strains ANT-5 and TRP-5 carry deletions of the araBAD genes (see Table 1). Further, IPTG (0.1 mM final concentration) was used to induce the chromosomally encoded vioD gene. As expected, the cell suspension of strain TRP-5 when equipped with the pVioABCE-km plasmid developed the typical deep-violet color indicative for the successful expression of the vio genes (see Figure S3). This, however, was only the case when TRP-5 grew (in the presence of L-arabinose as inducer) on LB agar plates, or on minimal agar plates with a single C source and supplementation of an inducer and either ANT or TRP. 3.2 Co-Cultures of Strains Allow Production of VIO From Various Carbon Sources We then testedwhether a stable co-culture of the two strainsANT- 5 and TRP-5 with their respective plasmids could be established. Therefore, both strains were first grown separately (as overnight cultures) and were then mixed in a 1:1 (50:50) ratio (based upon the OD600 values determined from the precultures). Incubation took place in shake flasks at 25◦C in MM with 5 g L−1 glucose. Without addition of inducers, the co-culture behaved as stable over the time course of 24 h (see Figure S4). This verified that these two strains, indeed, fed each other, when not induced for VIO production, similar to what we have reported before [41]. A recent study showed that synthetic E. coli consortia were able to synthesize up to roughly 50 mg L−1 of VIO in shake flask cultivation with 10 g L−1 D-galactose at 30◦C after optimizing the process (inoculation ratios, surfactant addition, and induction time point) [22]. To get a first understanding of the general VIO production ability and co-culture dynamics of our mutualistic co-cultures, the consortium was grown on four different C sources (D-glucose, glycerol, D-galactose, and D-xylose; 5 g L−1 intended starting concentration each), induced throughout with L-arabinose and IPTG. After cultivation, the final VIO concentra- tionswere∼18mgL−1 forD-glucose and∼24mgL−1 with glycerol, the best values were achieved with D-galactose and D-xylose,∼33 and 39 mg L−1, respectively (see Figure 2). We next compared the growth, VIO production, and population structure over a period of up to 5 days (120 h) in shake flasks to further exploit the shown potential of these mutualistic co-cultures. As VIO producers are known to be temperature- dependent, we chose 25◦C for further experiments. The four different C sources were compared, again. As expected, growth on D-glucose was fastest and the cell cultures reached an OD600 of ∼4.2 after already 24 h of incubation (Figure 3A). Growth on the three other C sources lagged clearly behind, glycerol as C source allowing the slowest growth. Final OD600 values ranged between ∼2.9 (D-galactose), ∼3.8 (D-glucose, D-xylose), and ∼4.4 (glycerol). Carbon sources were depleted in quite different fashions. Glucose was consumed after 24 h, galactose after 48 h, xylose after 72 h, and glycerol at the end of incubation (120 h) (see Figure 3C). VIO concentrations were low in the D-glucose shake flasks (lower than 8 mg L−1 after 120 h) (Figure 3B). D-Galactose yielded more VIO with a maximum of 34 mg L1 reached already after 48 h. Glycerol showed VIO production which was correlated to cell growth and up to 70 mg L−1 after 120 h. Finally, D-xylose gave the best results. VIO accumulated after a lag of ∼24 h but then developed continuously, reaching a maximal value of ∼126 mg L−1 after 120 h. 3.3 Population Structures Vary Enormously Depending on C Source As described above, we determined the percentage of the two strains ANT-5 and TRP-5 in the total population in the shake 4 of 10 Engineering in Life Sciences, 2024 FIGURE 1 Basic principle of the violacein-producing, synthetic mutualistic co-culture. (A) Simplified principle of the interaction of the strains ANT-5 (ANT-supplier) and TRP-5 (TRP-supplier, VIO-producer), based on the previously described co-culture (ANT-3/TRP-3) [41]. (B) Empty plasmid backbone of pVioABCE-km, providing kanamycin resistance for ANT-5. (C) Integration of the vioD gene (from Chromobacterium violaceum) into the bacterial genome (fuc locus) of TRP-5, plus transformation of pVioABCE-km [21] for complete VIO production. Both strains were additionally modified by deletion of their araBAD genes to not use L-arabinose (inducer of PBAD) as a carbon source. FIGURE 2 Final violacein concentration of shake flask experiments in minimal medium with novel co-culture (ANT-5/TRP-5; initial ratio: 50/50) at 30◦C with four different carbon sources (glucose, glycerol, galactose, and xylose). The data presented are from three biological replicates (two replicates for glucose). flask cultivations by plating out samples. To discern between strains, MacConkey agar plates with added lactose were used. We intended to establish 50:50 ratios at the starting time by mixing equivalent cell numbers as determined from the overnight cell densities. It should be noted that slight deviations from the intended 50:50 start-ratios were observed (see Figure 3C, 0 h data points). During growth on glucose, strain ANT-5 reached a share of almost 90% after 24 h. At that point of time no residual glucose was detectable and the co-culture had reached its maximum OD600 value. Thereafter, ANT-5’s share went down to about 50% and stayed thus till the end of incubation after 120 h. Growth on glycerol started with an actual ratio of roughly 58% (ANT-5):42% (TRP-5). ANT-5 stayed at this share after 24 h and eventually reached about 74% at the end of incubation. TRP-5 went from 42% in the beginning to 26% at the end of incubation. D-Galactose as C-source displayed a reverse population structure development. TRP-5 started at about 43% and went up to 90% after 48 h. At this point of time, galactose was consumed and the maximum OD value was noted. Afterward, the percentage of TRP-5 still went higher (up to >99%) at the end of incubation. D-Xylose retained a more stable ratio. Both strains had roughly a 50:50 share in the beginning. Thereafter, ANT-5 went up to about 61% after 24 h and then dropped to values below 30%at 72 h,which was the time point when no more D-xylose was present and the OD value had reached its maximum. TRP-5 instead went from 50% via 39% to 74% at 72 h. Eventually, both strains had an almost equal share at the end of incubation. It should be noted here that three biological replicates were run which all gave very similar trajectories. 3.4 Variations in the Initial Ratios Result in Quite Different Trajectories in Population Structure AsD-xylose had turned out asmost favorable C source for a stable co-culture, we wanted to know what the influence of variable start ratios was. We therefore ran shake flask incubation with changed initial (calculated) ratios of 70% (ANT-5):30% (TRP-5), or to the reverse (30% [ANT-5]:70% [TRP-5]). As can be seen, the two mixing ratios resulted in very similar growth trajectories (Figure 4A) with final OD values of about 4; the C source was consumed after 72 h. 5 of 10 FIGURE 3 Shake flask experiment in minimal medium with novel co-culture (ANT-5/TRP-5; initial ratio: 50/50) for violacein production at 25◦C with four different carbon sources (glucose, glycerol, galactose, and xylose). (A) Optical density at 600 nm, (B) violacein concentration over time, and (C) strain ratio percentages over cultivation time (determined on MacConkey agar plates). The vertical dashed line represents the end of the carbon source. The data presented are from three biological replicates. The VIO product yields, however, were different with the 30/70 ratio as indicated by higher VIO values throughout the incuba- tions over 120 h. The absoluteVIO values, however,were less (∼64 and∼46mg L−1, respectively) than in the 50/50 ratio experiments. Themeasured strain ratios showed quite different behavior. In the 70/30 experiments, ANT-5 increased slightly at 24 h to 82%, then dropped to values of 56% and 70% after 48 and 72 h. Finally, the share of ANT-5 was less than 3% at the end of incubation. TRP- 5 instead went from 26% to <20%, then increased to 44% and to 30% at 72 h. Then, curiously, strain TRP-5 completely overtook the population to have a share of >97% at the end of incubation. The 30/70 experiment on the other side showed an unbalanced ratio after C source consumption (after 48 h), but after 120 h a similar ratio than the other approach was detected (see Figure 4C). We analyzed the concentrations of ANT and TRP in the super- natants of the three xylose co-cultures. Free TRP could not be detected at any time (detection limit: 0.1 mg L−1). ANT concentrations peaked at 24 h with less 12 mg L−1 with the 50:50 and 30:70 ratio co-cultures (Figure 5A). The 70:30 co-culture 6 of 10 Engineering in Life Sciences, 2024 FIGURE 4 Shake flask experiment in minimal medium with xylose with novel co-culture (ANT-5/TRP-5; initial ratio: 70/30 and 30/70) for violacein production at 25◦C. (A) Optical density at 600 nm, (B) violacein concentration over time, and (C) strain ratio percentages over cultivation time (determined on MacConkey agar plates). The vertical dashed line represents the end of the carbon source. FIGURE 5 Analytics of the xylose at different initial ratios (50/50, 70/30, and 30/70). (A) Anthranilate concentrations during cultivation time. (B) Biomass-specific violacein production rate (qVIO) calculated from mean values. showed ANT concentrations above 30 mg L−1 at 24, 48, 72, and 120 h. We also determined the biomass-specific VIO production rate (qVIO). As can be seen from Figure 5B, the 50:50 initial ratio gave highest qVIO of about 3.0 and 2.4mg g−1 h−1 for the 30:70 ratio. The 70:30 ratio gave slightly more than 1.2 mg g−1 h−1. 4 Discussion In a pioneering work, two recombinant E. coli strains carrying different plasmids (pVioABX,wherebyX contained the vioE gene, and pVioCD, respectively) were co-cultured in 2xYTmedium and 7 of 10 FIGURE 6 Estimated specific L-tryptophan providing rate to strain TRP-5 for the growth of the auxotrophic strain ANT-5 (qTRP; TRP-5 for ANT-5) in a common shaking flask environment using different carbon sources. The rates are calculated based on the results presented in Figure 3 (for details, see Section 2). The dashed lines represent the point after carbon source depletion, which was therefore not considered. Error bars represent the standard deviation of biological triplicates. shown to produce a mixture of compounds consisting of VIO, DVIO, and proviolacein [52]. This approach, however, was not a classical co-culture but rather served to enlighten the biosynthetic pathway of VIO. A group recently used a commensalistic co-culture to produce VIO [22]. They identified several optimizing parameters regard- ing the carbon source used, the inoculation strain ratio, the time of induction and the addition of surfactants. A final VIO concentration of ∼50 mg L−1 was achieved and a final overall OD600 of about 9 (already after 24 h, presumably due to depletion of the initial 10 g L−1 supplied C source). In comparison, using the same carbon source (galactose, set starting concentration 5 g L−1) and temperature, our mutualistic co-culture achieved substantially higher VIO production performance parameters both in terms of substrate-to-product yield and biomass-specific VIO productivity. Since the respective reference values are not given for the commensalistic approach [22], we calculated the values based on the data isolated from the corresponding figure (Figure 4D in [22]).Without pursuing any optimization strategies, our syntheticmutualistic co-culture achieved a∼48%higher yield value YVIO/GAL (mg g−1) and a mean specific production rate qVIO (mg [OD600 h]−1) of∼2.2 × the commensalistic reference, suggest- ing that our approach is promising. It is worth mentioning that comparing the performance parameters after the 24 h mark (end of growth in [22]) would lead to an even more evident benefit. As a VIO production improvement strategy, we examined the well-known approach of temperature reduction. When lowering the temperature to 25◦C, the glycerol and D-xylose batch cul- tivations even reached final values of ∼70 and ∼126 mg L−1 of VIO. In the mentioned E. coli co-culture, D-galactose was the most favorable C source [22], while our mutualistic consortium achieved a 3.9-fold higher final VIO concentrationwhen supplied with D-xylose compared to D-galactose. Our findings showed that VIO production is clearly determined by the cultivation temperature (for glycerol and D-xylose), as shown by several studies before [12, 42, 53]. The stable growth of themutualistic co- culture depends on the cross-exchange of metabolites. We found that the different carbon sources permit this process to occur to different degrees, resulting in different population dynamics. The ANTproducerANT-5 consistently provides an oversupply ofANT in almost all cases (except D-xylose 48 h time point), which can be deduced from the different levels of ANT accumulation in the medium (data not shown). For the TRP auxotrophic strain ANT-5, no such TRP oversupply was detected. However, in the case of GLY as carbon source, a sufficient supply of TRP seems to be ensured, as indicated by relatively stable population ratios. In contrast, the decrease of ANT-5 in the total population of the GAL and XYL approaches (from 24 h) points toward an insufficient TRP supply (see Figure 6). Comparing the assumed TRP supply rates of TRP-5 to ANT-5 with those of the stable GLY approach (see Figure 6), it was found that the TRP supply rates are permanently or temporarily below the value of qTRP; TRP-5 for ANT-5 of ∼0.42, which is considered as necessary for stable population dynamics (mean value GLY approach). 8 of 10 Engineering in Life Sciences, 2024 The strain ratio analysis revealed that a stable co-culture resulted in a higher production of VIO, which was concurrently dependent on the kind of C source. During cultivation especially the use of D-xylose maintained stability that led to higher VIO concentrations. This underlines the complexity of different co- cultures, but nonetheless shows the potential of our mutualistic approach. While other groups surpassed our VIO production with values up to 4.45 g L−1 from monocultures grown on several C sources like D-glucose and glycerol in bioreactors [23, 26], we could expand feasible operating strategies by moving from shake flasks to spatially separated, interconnected two-compartment bioreactor systems with strain-specific temperature and process conditions [41, 53]. Inoculation ratios of co-cultures also have an impact on the productivity and stability of co-cultures [22, 41]. Shifting the initial ratio of D-xylose grown co-culture strains ANT-5 and TRP-5 to 70/30 and 30/70 gave new insights. The biomass-specific VIO production rate (qVIO) was generally lower than the 50/50 co- culture, especially the 70/30 approach. However, stability of the co-culture was not the reason for the lower production rate, as it had a nearly 50/50 ratio at the complete consumption of the C source. An ANT accumulation was detectable in the extracellular metabolite concentration, which was not fully consumed by the TRP-5 VIO producer strain, compared to the others. The C source here was mainly used for ANT production than for VIO production. Our results indicate that the kind of carbon source and co- culture stability are closely related and are crucial parameters for VIO productivity, and for the state of metabolic cross-feeding balance. The impact of metabolic cross-feeding at the core of the mutualistic co-cultures was already highlighted in our previous work [41]. This mutualistic and stable VIO-producing E. coli co- culture on D-xylose was able to produce a final concentration of ∼126 mg L−1 of VIO in a shake flask batch approach. In the future, the co-culture with its promising characteristics could be applied to a scaled-up continuous bioreactor fermentation for VIO production. Additionally, the co-culture principle is not limited to VIO production only but has the potential to also synthesize other valuable molecules. Acknowledgments The authors acknowledge Dr. Natalia Trachtmann for her helpful dis- cussions on TRP strain engineering and Dr. Jung-Won Youn for his HPLC expertise. In addition, the authors thank the German Research Foundation (DFG) for funding the research project (Project number: 427751089, SP503/11-1, and TA 241/12-1, respectively) within the priority program InterZell (SPP2170). Open-access publication was arranged within the context of the project DEAL. Conflicts of Interest The authors declare no conflicts of interest. Data Availability Statement The data that support the findings of this study are available from the corresponding author upon reasonable request. References 1. Y.-S. Ko, J. W. Kim, J. A. Lee, et al., “Tools and Strategies of Systems Metabolic Engineering for the Development of Microbial Cell Factories for Chemical Production,” Chemical Society Reviews 49 (2020): 4615– 4636. 2. J. Nielsen and J. D. Keasling, “Engineering Cellular Metabolism,” Cell 164 (2016): 1185–1197. 3. A. Rodriguez, J. A. Martnez, N. Flores, et al., “Engineering Escherichia coli to Overproduce Aromatic Amino Acids and Derived Compounds,” Microbial Cell Factories 13 (2014): 126. 4. J.-H. Lee and V. F. Wendisch, “Biotechnological Production of Aro- matic Compounds of the Extended Shikimate Pathway From Renewable Biomass,” Journal of Biotechnology 257 (2017): 211–221. 5. D. Rettori and N. Durán, “Production, Extraction and Purification of Violacein: An Antibiotic Pigment Produced by Chromobacterium Violaceum,” World Journal of Microbiology and Biotechnology 14 (1998): 685–688. 6. N. Durán, G. Z. Justo, C. V. Ferreira, et al., “Violacein: Properties and Biological Activities,” Biotechnology and Applied Biochemistry 48 (2007): 127–133. 7. I. Rivero Berti, M. E. Gantner, S. Rodriguez, et al., “Potential Biocide Roles of Violacein,” Frontiers in Nanotechnology 5 (2023): 1186386. 8. H. Park, S. Park, Y.-H. Yang, and K.-Y. Choi, “Microbial Synthesis of Violacein Pigment and Its Potential Applications,” Critical Reviews in Biotechnology 41 (2021): 879–901. 9. S. Y. Choi, S. Lim, K. Yoon, et al., “Biotechnological Activities and Applications of Bacterial Pigments Violacein and Prodigiosin,” Journal of Biological Engineering 15 (2021): 10. 10. R. K. Suryawanshi, C. D. Patil, H. P. Borase, et al., “Towards an Understanding of Bacterial Metabolites Prodigiosin and Violacein and Their Potential for Use in Commercial Sunscreens,” International Journal of Cosmetic Science 37 (2015): 98–107. 11. M. Kanelli, M. Mandic, M. Kalakona, et al., “Microbial Production of Violacein and Process Optimization for Dyeing Polyamide Fabrics With Acquired Antimicrobial Properties,” Frontiers in Microbiology 9 (2018): 1495. 12. R. Riveros, M. Haun, and N. Durán, “Effect of Growth Conditions on Production of Violacein by Chromobacterium Violaceum (BB-78 Strain),” Brazilian Journal of Medical and Biological Research 22 (1989): 569– 577. 13. J. M. Pemberton, K. M. Vincent, and R. J. Penfold, “Cloning and Heterologous Expression of the Violacein Biosynthesis Gene Cluster From Chromobacterium Violaceum,” Current Microbiology 22 (1991): 355–358. 14. F. Pantanella, F. Berlutti, C. Passariello, et al., “Violacein and Biofilm Production in Janthinobacterium lividum,” Journal of Applied Microbiology 102 (2007): 992–999. 15. X. Xu, X. Chu, B. Du, et al., “Functional Characterization of a Novel Violacein Biosynthesis Operon From Janthinobacterium sp. B9-8,” Applied Microbiology and Biotechnology 106 (2022): 2903–2916. 16. S. Yada, Y. Wang, Y. Zou, et al., “Isolation and Characterization of Two Groups of Novel Marine Bacteria Producing Violacein,” Marine Biotechnology 10 (2008): 128–132. 17. Y. Lu, L.Wang, Y. Xue, et al., “Production of Violet Pigment by aNewly Isolated Psychrotrophic Bacterium From a Glacier in Xinjiang, China,” Biochemical Engineering Journal 43 (2009): 135–141. 18. Y. Tong, J. Zhou, L. Zhang, and P. Xu, “A Golden-Gate Based Cloning Toolkit to Build Violacein Pathway Libraries in Yarrowia lipolytica,” ACS Synthetic Biology 10 (2021): 115–124. 19. H. Sun, D. Zhao, B. Xiong, et al., “Engineering Corynebacterium glutamicum for Violacein Hyper Production,” Microbial Cell Factories 15 (2016): 148. 9 of 10 20. A. L. Rodrigues, Y. Göcke, C. Bolten, et al., “Microbial Production of the Drugs Violacein and Deoxyviolacein: Analytical Development and Strain Comparison,” Biotechnology Letters 34 (2012): 717–720. 21. A. L. Rodrigues, N. Trachtmann, J. Becker, et al., “Systems Metabolic Engineering of Escherichia coli for Production of the Antitumor Drugs Violacein and Deoxyviolacein,”Metabolic Engineering 20 (2013): 29–41. 22. D. Gwon, E. Seo, and J. W. Lee, “Construction of Synthetic Microbial Consortium for Violacein Production,” Biotechnology and Bioprocess Engineering 28 (2023): 1005–1014. 23. A. L. Rodrigues, J. Becker, A. O. de Souza Lima, et al., “Systems Metabolic Engineering of Escherichia coli for Gram Scale Production of the Antitumor Drug Deoxyviolacein From Glycerol,” Biotechnology and Bioengineering 111 (2014): 2280–2289. 24. M.-Y. Fang, C. Zhang, S. Yang, et al., “High Crude Violacein Pro- duction From Glucose by Escherichia coli Engineered With Interactive Control of Tryptophan Pathway and Violacein Biosynthetic Pathway,” Microbial Cell Factories 14 (2015): 8. 25. M. Fang, T. Wang, C. Zhang, et al., “Intermediate-Sensor Assisted Push–Pull Strategy and Its Application in Heterologous Deoxyviolacein Production in Escherichia coli,”Metabolic Engineering 33 (2016): 41–51. 26. Y. Zhou, M.-Y. Fang, G. Li, et al., “Enhanced Production of Crude Violacein From Glucose in Escherichia coli by Overexpression of Rate- Limiting Key Enzyme(S) Involved in Violacein Biosynthesis,” Applied Biochemistry and Biotechnology 186 (2018): 909–916. 27. A. Ahmed, A. Ahmad, R. Li, et al., “Recent Advances in Synthetic, Industrial and Biological Applications of Violacein and Its Heterologous Production,” Journal of Microbiology and Biotechnology 31 (2021): 1465– 1480. 28. N. Sarwar, S. Sarwar, S. Ejaz, et al., “Metabolic Engineering ofMicroor- ganisms to Increase Production of Violacein,” International Journal of Agriculture Environment and Biotechnology 6 (2021): 295–306. 29. D. Yang, S. Y. Park, and S. Y. Lee, “Production of RainbowColorants by Metabolically Engineered Escherichia coli,” Advanced Science (Weinheim, Baden-Wurttemberg, Germany) 8 (2021): e2100743. 30. D. Gwon, J. Y. Seok, G. Y. Jung, and J. W. Lee, “Biosensor-Assisted Adaptive Laboratory Evolution for Violacein Production,” International Journal of Molecular Sciences 22 (2021): 6594. 31. R. D. Shukla, Á. Zvara, Á. Avramucz, et al., “inPOSE: A Flexi- ble Toolbox for Chromosomal Cloning and Amplification of Bacterial Transgenes,”Microorganisms 10 (2022): 236. 32. H. Zhang, B. Pereira, Z. Li, and G. Stephanopoulos, “Engineering Escherichia coli Coculture Systems for the Production of Biochemical Products,” Proceedings of the National Academy of Sciences 112 (2015): 8266–8271. 33. H. Zhang and X. Wang, “Modular Co-Culture Engineering, a New Approach for Metabolic Engineering,” Metabolic Engineering 37 (2016): 114–121. 34. N. I. Johns, T. Blazejewski, A. L. Gomes, and H. H. Wang, “Principles for Designing Synthetic Microbial Communities,” Current Opinion in Microbiology 31 (2016): 146–153. 35. N. S. McCarty and R. Ledesma-Amaro, “Synthetic Biology Tools to Engineer Microbial Communities for Biotechnology,” Trends in Biotech- nology 37 (2019): 181–197. 36. E. Sgobba and V. F. Wendisch, “Synthetic Microbial Consortia for Small Molecule Production,” Current Opinion in Biotechnology 62 (2020): 72–79. 37. T. Chen, X. Wang, L. Zhuang, et al., “Development and Optimization of a Microbial Co-Culture System for Heterologous Indigo Biosynthesis,” Microbial Cell Factories 20 (2021): 154. 38. M. Diender, I. Parera Olm, and D. Z. Sousa, “Synthetic Co-Cultures: Novel Avenues for Bio-Based Processes,” Current Opinion in Biotechnol- ogy 67 (2021): 72–79. 39. X. Chen, M. Wang, Y. Xing, et al., “Assembly of Microbial Commu- nities Engaging in Metabolic Division of Labor in a Diffusion-Limited Environment is Governed by Metabolic Flux,” ACS Synthetic Biology 12 (2023): 1972–1980. 40. M. Sauer and H. Marx, “Co-Cultures and Synthetic Microbial Com- munities for Green Chemical Production,” Current Opinion in Green and Sustainable Chemistry 42 (2023): 100842. 41. T.Müller, S. Schick, J. Beck, et al., “SyntheticMutualism inEngineered E. coli Mutant Strains as Functional Basis for Microbial Production Consortia,” Engineering in Life Sciences 23 (2023): e2100158. 42. Y. Zhang, H. Chen, Y. Zhang, et al., “Direct RBS Engineering of the Biosynthetic Gene Cluster for Efficient Productivity of Violaceins in E. coli,”Microbial Cell Factories 20 (2021): 38. 43. Y. Jiang, B. Chen, C.Duan, et al., “MultigeneEditing in theEscherichia coli Genome Via the CRISPR-Cas9 System,” Applied and Environmental Microbiology 81 (2015): 2506–2514. 44. K. A. Datsenko and B. L. Wanner, “One-Step Inactivation of Chromo- somal Genes in Escherichia coliK-12 Using PCR Products,” Proceedings of the National Academy of Sciences 97 (2000): 6640–6645. 45. P. P. Cherepanov and W. Wackernagel, “Gene Disruption in Escherichia coli: TcR and KmR Cassettes With the Option of Flp- Catalyzed Excision of the Antibiotic-Resistance Determinant,” Gene 158 (1995): 9–14. 46. T. Zeppenfeld, C. Larisch, J. W. Lengeler, and K. Jahreis, “Glucose Transporter Mutants of Escherichia coli K-12 With Changes in Substrate Recognition of IICB Glc and Induction Behavior of the ptsG Gene,” Journal of Bacteriology 182 (2000): 4443–4452. 47. N. Trachtmann, K. F. Alvarez Fong, E. Guitart Font, and G. A. Sprenger, “Construction of Chromosomally Encoded lacZ and gfp Reporter Strains of Escherichia coli for the Study of Global Regulation of Metabolism,” Engineering in Life Sciences 16 (2016): 675–681. 48. S. Tanaka, S. A. Lerner, and E. C. Lin, “Replacement of a Phosphoenolpyruvate-Dependent Phosphotransferase by a Nicotinamide Adenine Dinucleotide-Linked Dehydrogenase for the Utilization of Mannitol,” Journal of Bacteriology 93 (1967): 642–648. 49. B. Mohammadi Nargesi, G. A. Sprenger, and J.-W. Youn, “Metabolic Engineering of Escherichia coli for Para-Amino-PhenyleThanol and Para- Amino-Phenylacetic Acid Biosynthesis,” Frontiers in Bioengineering and Biotechnology 6 (2019): 201. 50. J. L. Ingraham, O.Maaløe, and F. C. Neidhardt,Growth of the Bacterial Cell (Sunderland, MA, USA: Sinauer Associates, 1983). 51. J. Pramanik and J. D. Keasling, “Effect of Escherichia coli Biomass Composition on Central Metabolic Fluxes Predicted by a Stoichiometric Model,” Biotechnology and Bioengineering 60 (1998): 230–238. 52. C. Sánchez, A. F. Braña, C. Méndez, and J. A. Salas, “Reevalu- ation of the Violacein Biosynthetic Pathway and Its Relationship to Indolocarbazole Biosynthesis,” Chembiochem 7 (2006): 1231–1240. 53. T. Müller, S. Schick, J.-S. Klemp, et al., “Synthetic Co-Culture in an Interconnected Two-Compartment Bioreactor System: Violacein Pro- duction With Recombinant E. coli Strains,” Bioprocess and Biosystems Engineering 47 (2024): 713–724. Supporting Information Additional supporting information can be found online in the Supporting Information section. 10 of 10 Engineering in Life Sciences, 2024 Stability of a Mutualistic Escherichia coli Co-Culture During Violacein Production Depends on the Kind of Carbon Source 1 | Introduction 2 | Materials and Methods 2.1 | Strain Construction 2.2 | Cultivation and Media Composition 2.2.1 | Culture Medium 2.2.2 | Shake Flask Cultivation 2.3 | Strain Ratio Identification 2.4 | Violacein Extraction 2.5 | Analytical Methods 2.5.1 | Determination of Optical Density (OD600) and Cell Dry Weight 2.5.2 | Violacein Determination by HPLC 2.5.3 | Anthranilate and Tryptophan HPLC 2.5.4 | Sugar HPLC 2.6 | Calculation of Rates and Yields 3 | Results 3.1 | Considerations for Strain Construction and Implementation 3.2 | Co-Cultures of Strains Allow Production of VIO From Various Carbon Sources 3.3 | Population Structures Vary Enormously Depending on C Source 3.4 | Variations in the Initial Ratios Result in Quite Different Trajectories in Population Structure 4 | Discussion Acknowledgments Conflicts of Interest Data Availability Statement References Supporting Information