Metabolic engineering of Corynebacterium glutamicum for anthocyanin production
- 856 Downloads
Anthocyanins such as cyanidin 3-O-glucoside (C3G) have wide applications in industry as food colorants. Their current production heavily relies on extraction from plant tissues. Development of a sustainable method to produce anthocyanins is of considerable interest for industrial use. Previously, E. coli-based microbial production of anthocyanins has been investigated extensively. However, safety concerns on E. coli call for the adoption of a safe production host. In the present study, a GRAS bacterium, Corynebacterium glutamicum, was introduced as the host strain to synthesize C3G. We adopted stepwise metabolic engineering strategies to improve the production titer of C3G.
Anthocyanidin synthase (ANS) from Petunia hybrida and 3-O-glucosyltransferase (3GT) from Arabidopsis thaliana were coexpressed in C. glutamicum ATCC 13032 to drive the conversion from catechin to C3G. Optimized expression of ANS and 3GT improved the C3G titer by 1- to 15-fold. Further process optimization and improvement of UDP-glucose availability led to ~ 40 mg/L C3G production, representing a > 100-fold titer increase compared to production in the un-engineered, un-optimized starting strain.
For the first time, we successfully achieved the production of the specialty anthocyanin C3G from the comparatively inexpensive flavonoid precursor catechin in C. glutamicum. This study opens up more possibility of C. glutamicum as a host microbe for the biosynthesis of useful and value-added natural compounds.
KeywordsAnthocyanin Corynebacterium glutamicum Catechin Flavonoid UDP-glucose
At present, anthocyanins used in industry are mainly obtained by extraction from plant tissues, which is subject to seasonal supply and quality control concerns inherent in agriculture [7, 8]. An alternative way of production is biosynthesis in metabolically engineered microorganisms, attributed to some advantages of microbes over plants, including ease of cultivation and fast growth, availability of sophisticated genetic tools, and well-defined metabolic networks and models. The most commonly used chassis microbe in metabolic engineering is E. coli, which has been extensively engineered for the biosynthesis of several natural flavonoids such as naringenin, kaempferol, and quercetin [9, 10, 11]. Saccharomyces cerevisiae and Streptomyces venezuelae have also been deployed for natural flavonoid production [12, 13, 14].
The biosynthesis of anthocyanins has been conducted in microorganisms for over a decade. In 2005, Yan et al. cloned and expressed in E. coli the genes of flavanone 3-hydroxylase (F3H) and ANS from Malus domestica, DFR from Anthurium andraeanum, and flavonoid 3-O-glucosyltransferase (F3GT) from Petunia hybrida . The recombinant strain produced 6.0 μg/L of C3G and 5.6 μg/L pelargonidin 3-O-glucoside using naringenin and eriodictyol as the respective precursors. Subsequent selection of plant-derived gene orthologs, optimization of UDP-glucose pool, regulation of precursor uptake and optimization of the production process dramatically enhanced production of pelargonidin 3-O-glucoside and C3G, with their titers reaching 113 mg/L and 350 mg/L, using afzelechin and catechin precursors, respectively [16, 17, 18]. Recently, de novo production of ~ 10 mg/L pelargonidin 3-O-glucoside from glucose has been achieved via an E. coli consortium. In this study, the first node strain was a highly efficient tyrosine producer and the entire pathway from tyrosine to pelargonidin 3-O-glucoside was split into four strains . However, all the reported recombinant hosts producing anthocyanins are currently limited to E. coli derivatives.
Corynebacterium glutamicum, having been widely used in industrial production of amino acids such as L-glutamate and l-lysine [20, 21], is advantageous over other bacteria in several aspects: (1) it does not produce endotoxins like E. coli and is generally regarded as safe for the production of pharmaceuticals, food and cosmetics; (2) it has been broadly applied in industry, and current facilities can be retrofitted to produce chemicals of interest; (3) its metabolism can be easily rewired for target compounds through the readily available genetic tools and metabolic models [22, 23, 24, 25]. Recently, this strain has been successfully engineered to produce flavanones or stilbenes by expressing CHS and CHI or stilbene synthase, respectively . Subsequently, the heterologous pathways introduced into C. glutamicum have been extended to flavonols (such as kaempferol and quercetin) and pterostilbene .
In this study, we constructed recombinant C. glutamicum strains that could produce the anthocyanin C3G from catechin. Through optimization of gene parts, expression levels, fermentation process parameters, and supply of the cosubstrate UDP-glucose, the engineered strain was able to produce ~ 40 mg/L C3G from 500 mg/L of catechin. To the best of our knowledge, this is the first report of biosynthesis of any anthocyanin in C. glutamicum, and this study further potentiates C. glutamicum for its application in flavonoid bioproduction.
Optimization of 3GT expression for C3G production in C. glutamicum
Enhanced expression of ANS improves C3G production
Comparison of different promoters for C3G pathway gene expression
Production of C3G by different engineered strains with combinations of constitutive promoters
C3G titer (mg/L)
1.02 ± 0.09
1.20 ± 0.11
3.13 ± 0.02
2.68 ± 0.01
Regulation of UDP-glucose supply for improved C3G production
C3G production using optimized fermentation conditions
When the concentration of the inducer IPTG was studied for its impact on C3G generation, a trend similar to that of inoculum size was observed, with 0.5 mM IPTG induction increasing C3G yield by ~ 10% compared to induction by 1 mM IPTG (Fig. 7b). The selection of carbon and nitrogen sources also played a significant role in C3G bioconversion. Supplementation of 2 g/L casamino acids  did not considerably alter C3G yield, whereas 4 g/L casamino acids slightly reduced production, and yeast extract and peptone largely inhibited C3G generation. For all the tested nitrogen sources, a concomitant addition of a carbon source had the same pattern of impact, with glucose > sucrose > fructose in supporting C3G formation, except when peptone was used as the nitrogen source (Fig. 7c). Based on these observations, the optimal bioprocess for C3G production was established as 2.5% inoculum size with 500 μM IPTG induction at 6 h after sub-culture, in modified AMM medium supplemented with 20 g/L glucose and 2 g/L casamino acids, with the maximal C3G titer reaching 41.7 mg/L. To evaluate the fermentation performance in shake flasks, the same strain was tested using the optimized fermentation conditions, and a titer of ~ 33 mg/L was observed in flasks at 48 h post induction. Additionally, we carried out C3G production using whole cells as a biocatalyst and obtained 43.7 mg/L of C3G (Additional file 1: Table S3), which was slightly higher than that obtained from the growing cells. In the whole cell biotransformation, C3G was only produced when cells were resuspended in AMM (pH7.0), suggesting that the C3G pathway might be sensitive to pH, salts, and/or certain nutrients.
Microbial production of anthocyanins may be used as a feasible way of producing anthocyanins for research and industrial applications. Previously, efficient E. coli-based bioproduction of anthocyanins have been successfully achieved in our lab [16, 39]. However, the host strain E. coli contains some unfavorable intrinsic properties such as production of endotoxins and underlying pathogenicity. Thus, in the present study, we referred to a new host strain, C. glutamicum, for the production of C3G. Through a series of engineering and optimization, ~ 40 mg/L C3G was generated by the modified strain. Recently, employment of C. glutamicum as the host strain to produce some flavanones has been performed . These attempts demonstrate that C. glutamicum is a possible choice of microbial production of flavonoids.
In engineered microbes, the heterologous expression of plant-derived genes is generally challenging, and issues such as incorrect protein folding and formation of inclusion bodies lead to poor production of functional enzymes involved in the metabolic pathways. Codon optimization and fusion expression are commonly used strategies to partially solve these problems [40, 41, 42]. In this study, we found that codon optimization of ANS and 3GT had very limited positive effects on improving C3G production, indicating that the codon usage in C. glutamicum might fit well with that in plants for these two genes. In another study of flavanone pathway expression in C. glutamicum, however, codon-optimization is indispensable . Fusion expression with tags are known to improve soluble expression of alien proteins in common host bacteria. In the present study, MBP and SUMO fusion led to enhanced C3G biosynthesis, suggesting better expression of 3GT in its active form. Another study also benefited from this method, in which the enzyme cis-itaconate decarboxylase with an MBP fusion had > 2-fold higher activity, and the fusion enzyme led to one-fold increase of itaconate production in C. glutamicum .
Besides the expression level of each individual gene, the ratio of expression levels of 3GT and ANS was also found to be critical for anthocyanin production in C. glutamicum. ANS is pivotal in the biosynthesis of anthocyanins in plants, as a direct correlation between ANS expression and anthocyanin accumulation has been observed in fruits such as bilberries and apples [44, 45]. More transcripts of the ANS gene lead to more copies of the ANS enzyme, which can help to convert more catechin to cyanidin. In this study, the expression level of 3GT decreased in the monocistronic form due to shortened translation time compared to that in the operon organization , resulting in a lower ratio of 3GT and ANS expression levels. Thus, a balanced expression of 3GT and ANS was achieved in the monocistronic form of pathway architecture, which could lead to enhanced C3G production as extensively demonstrated in metabolic engineering [47, 48].
Although sufficient expression of ANS and 3GT was necessary for better generation of C3G in C. glutamicum, extremely strong expression did not translate to a higher yield. The same negative effect was again observed in our study of polyglutamic acid production in C. glutamicum, in which the sod-controlled pathway expression led to less efficient biocatalysis than the tac-controlled expression (unpublished data). This was unexpected because these strong promoters have been widely used in the production of amino acids and other chemicals [49, 50, 51]; and in naringenin production in S. cerevisiae, constitutive expression of pathway genes driven by strong promoters (such as TDH3) resulted in a much higher titer compared with gene expression driven by weak inducible promoters (GAL1 and GAL10) [52, 53]. A possible explanation is that eftu and sod promoters are not applicable in the production of secondary metabolites in C. glutamicum. Given that the transcript threshold of ANS and 3GT tolerated by the host cells may be much lower than that of the pathway genes in amino acid biosynthesis, particularly strong expression of ANS or 3GT driven by promoter sod or eftu could bring severe metabolic burden to cells, thus leading to imbalanced metabolic pathways and limited generation of cofactors and cosubstrates. In this sense, moderate expression of both genes is crucial for high-titer production of anthocyanins, as in the case of resveratrol production in E. coli, in which modest constitutive expression (gap promoter), instead of strong inducible expression (T7 promoter), of 4CL and the stilbene synthase gene led to a higher yield . It could be inferred that optimal expression of flavonoid biosynthesis genes depends on the host strains, and the suitable expression configuration varies among strains and systems.
In the present study, the maximal conversion yield based on consumed catechin was ~ 30%, equivalent to the yield in E. coli . In our preliminary test, the substrate catechin was shown to be stable in the growing culture. Thus, it could be postulated that the consumed catechin was converted to cyanidin. Given that cyanidin is very unstable at neutral pH, and that an obvious cyanidin peak was not detected in HPLC analysis, it can be inferred that the fast degradation of cyanidin is a possible limiting factor in C3G production. It should be noted that C3G is also unstable at neutral pH ; thus, C3G stabilization is important for its biosynthesis. This has been achieved in E. coli by conducting the biocatalysis at a low pH (e.g., pH 5.0) . However, such a strategy was not feasible for C. glutamicum due to its high sensitivity to low pHs (Additional file 1: Table S3). Adaptation of C. glutamicum for better tolerance to low pHs could be a possible solution to improve C3G production in an acidic environment.
Apart from the modification on the anthocyanin pathway and stability of cyanidin and C3G, the supply of UDP-glucose is one of the most important factors in derteming C3G production. The intracellualr UDP-glucose is relatively stable and strictly controlled, with limited flow towards the formation of glycosylated anthocyanidin (C3G) as has been extensively demonstrated in the production of anthocaynins and other glycosylated flavonoids in E. coli [17, 56]. In the present study, coexpression of pgm and galU1 increased the production of C3G, indicating that their expression could channel more glucose-6-phosphate to UDP-glucose. This strategy could be used in C. glutamicum-based biosynthesis of other UDP-glucose derived products, such as glycogen, glycosylated proteins, and sophorolipids. In addition, inhibition of UDP-glucose degradation pathways through gene knockout or CRISPR interference could be conducted to improve the accumulation of UDP-glucose and further elevate C3G production [57, 58].
We have demonstrated the successful production of C3G in C. glutamicum from the comparatively abundant and inexpensive catechin. Through controlled regulation of the expression of the plant-derived anthocyanin pathway genes (ANS and 3GT), fine-tuned supply of UDP-glucose, and optimized fermentation process, C3G titer was elevated from ~ 0.37 mg/L to ~ 40 mg/L, representing > 100-fold improvement. This is the first report of anthocyanin bioproduction in C. glutamicum, and opens up new possibilities of microbial production of flavonoids by the GRAS strain C. glutamicum beyond E. coli. The inter-correlation of the flavonoid pathway with aromatic amino acid production pathway, and the extensive application of C. glutamicum in industrial production of amino acids make this bacterium promising for high-titer flavonoid biosynthesis from inexpensive feedstocks. So far, the production of naringenin from extracellular tyrosine has been achieved in C. glutamicum , and high titer production of tyrosine (26 g/L) from glucose in C. glutamicum has been well established [59, 60]. Based on these advances, it could be anticipated that de novo production of C3G from cheap carbon sources such as glucose or sucrose by a single recombinant C. glutamicum or a mixed culture of C. glutamicum strains can be fulfilled in the near future [19, 61].
Bacterial strains and media
The strains used in the study are listed in Additional file 1: Table S1. E. coli DH5α was used for cloning and plasmid propagation, and was grown in Luria Broth (LB) medium (Sigma) supplemented with 50 mg/L kanamycin when necessary; agar (Sigma) was added to 15 g/L for the preparation of medium-agar plates. C. glutamicum ATCC 13032 was used as the host for flavonoid production in this study. C. glutamicum cells were generally grown in Brain Heart Infusion (BHI) medium (BD) and kept in BHI with glycerol (20%, v/v) at − 80 °C for long-term storage. Fermentation by C. glutamicum was conducted in AMM medium supplemented with 0.2 mg/L biotin . AMM medium contained (per liter): glucose, 20 g; KH2PO4, 3.5 g; K2HPO4, 5.0 g; (NH4)2HPO4, 3.5 g; casamino acids, 2 g; MgSO4, 0.12 g; CaCl2, 11 mg; thiamine HCl, 0.5 mg; MOPS, 8.37 g; Tricine, 0.72 g; FeSO4·7H2O, 2.8 mg; NaCl, 2.92 g; NH4Cl, 0.51 g; MgCl2 0.11 g; K2SO4 0.05 g; and micronutrient mix ((NH4)6Mo7O24, 0.4 μg; H3BO3, 2.5 μg; CuSO4, 0.24 μg, MnCl2, 1.6 μg; and ZnSO4, 0.28 μg).
The plasmids and primers used in the present study are listed in Additional file 1: Tables S1 and S2. The ANS gene from Petunia hybrida and 3GT from Arabidopsis thaliana were acquired through PCR amplification (ACCUZYME 2X mix, Bioline) using the plasmid pETM6-At3GT-m-PhANS in the Koffas lab . Similarly, maltose-binding protein (MBP) tag or small ubiquitin-like modifier (SUMO) tag was amplified using the plasmid pMAL-c2X-PhANS  or pET His6 SUMO TEV LIC cloning vector (Addgene plasmid 29711). The codon-optimized genes of ANS and 3GT were synthesized by Integrated DNA technologies (IDT, USA). The fusion of MBP or SUMO tag with wildtype or codon-optimized 3GT genes was achieved by overlap extension PCR.
To construct expression plasmids of operon configurations, different versions of 3GT genes were first inserted into pEC-XK99E using traditional restriction enzyme-based cloning, followed by insertion of the ANS gene. Expression plasmids with ANS and 3GT in a monocistronic form were constructed by insertion of a fragment, consisting of rrnB terminator and a tac promoter, into plasmids of operon configurations described above. To obtain the plasmids expressing the fused gene of 3GT and ANS (termed 3AO) in C. glutamicum, 3AO was amplified using the plasmid pCDF-3AO as the template and subsequently cloned into the expression plasmids pEC-XK99E and pZ8-1 by EcoRI and SalI, respectively.
Other expression plasmids were built on the basis of pZM1 (to be published separately), which was created from the plasmid pZ8-Ptac (Addgene plasmid 740694) along the principle in the construction of ePathBrick vector pETM6 . The genes in the UDP-glucose biosynthesis pathway in E. coli (cmk, ndk, galU, pgm and ycjU) or in C. glutamicum (galU1 and pgm) were amplified from the genomic DNA of BL21 Star (DE3) or C. glutamicum ATCC 13032, which was extracted by PureLink Genomic DNA Kit (Invitrogen). Each gene was then cloned into pZM1 and assembled in a monocistronic form using a previously published method .
Construction of recombinant C. glutamicum strains
A single colony of wildtype C. glutamicum ATCC 13032 was inoculated into 3 mL of BHI medium and grown at 30 °C and 225 rpm. After overnight growth, 2 mL culture was transferred to 50 mL fresh BHI medium and grown to OD600 of ~ 1.75. Cells were chilled on ice for 10 min and centrifuged for 5 min at 3500 rpm and 4 °C. The pellet was washed once with 50 mL of ice-cold 10% (v/v) glycerol containing 1 mM Tris (pH 7.5) in ultrapure water and once with 50 mL of ice-cold 10% (v/v) glycerol, and was then resuspended in 1 mL ice-cold 10% glycerol. Aliquots (100 µL) were stored at − 80 °C. For electroporation, cells were thawed on ice (10 min), mixed with ~ 100 ng plasmid, and transferred to an electroporation cuvette (2 mm gap). Electroporation was performed with an electroporator (Bio-Rad) at 25 μF, 200 W and 2.5 kV, yielding a pulse duration of ~ 5 ms. Immediately after electroporation, cells were mixed with 1 mL pre-warmed BHI in the cuvette, and were transferred to a 2-mL microcentrifuge tube. Cells were heat-shocked at 46 °C for 6 min in a water bath, transferred to a 14-mL culture tube (VWR), incubated for 2 h at 30 °C, and plated on LB-agar plates containing 25 mg/L kanamycin. Positive clones were validated by colony PCR, plasmid miniprep, and gene sequencing (Genewiz).
Glycerol stocks were streaked onto LB agar plates with 25 mg/L kanamycin. Single colonies were inoculated into 3 mL of BHI medium with 25 mg/L kanamycin in a 14-mL culture tube for overnight growth at 30 °C and 225 rpm. Fresh AMM (1 mL) with 25 μg/mL kanamycin in a single well of a polypropylene deep 48-well plate (5 mL, VWR) was inoculated with 25 μL of the overnight culture, or other volumes when noted. In the process of optimization of carbon and nitrogen sources in AMM, 20 g/L of glucose, fructose or sucrose as well as different nitrogen sources (2 g/L yeast extract or peptone, 2 or 4 g/L casamino acids, or 4 g/L casamino acids plus 14 g/L yeast extract ) was used to prepare AMM and to test their effect on C3G production. The culture was then incubated at 30 °C and 225 rpm for 6 h. IPTG and catechin (prepared as a 50 g/L stock solution in dimethylformamide: ethanol = 8:2, v/v) were added to final concentrations of 1 mM and 500 mg/L, respectively. Necessary supplements (2-oxoglutarate, 0.1 mM; sodium ascorbate, 2.5 mM; orotic acid, 0.1 mM) were also added from 50-fold concentrated stock solutions (for strains containing constitutive version of C3G module, catechin and supplements were fed at the beginning of the subculture). The culture was further grown for 24 h at 30 °C and 225 rpm, and then mixed with equal volume of acidified methanol (with 1% hydrochloric acid, v/v), followed by brief vertexing. Following centrifugation at 21,000×g for 10 min, the supernatant was used for subsequent HPLC analysis. Scaled-up fermentation was carried out similarly in a 125-mL PYREX Erlenmeyer Flask containing 15 mL fermentation medium. Three biological replicates were used in all experiments.
The supernatants of cell extracts were analyzed by a previously established method . Briefly, 25 µL of each sample was loaded into Agilent 1200 series HPLC consisting of a ZORBAXSB-18 column (5 μm, 150 mm × 4.6 mm) and a diode array detector, and was separated by solvent A (0.1% formic acid in water) and solvent B (0.1% formic acid in acetonitrile) with a linear gradient change (10–40% B at 0–10 min and 40–60% B at 10–15 min) at 1 mL/min flow rate. Absorbance at 280 nm and 520 nm was monitored. Peak areas were calculated for concentrations of the relevant compounds using standards of catechin (Sigma) and C3G (Alkemist Labs). Student’s t test was used for statistical analysis.
Agilent 1200 series HPLC equipped with an Eclipse XDB-C18 column (5 μm, 150 mm × 4.6 mm) and an LTQ-ORBITRAP XL mass spectrometer was used. HPLC analysis was performed with solvent A (0.1% formic acid in water) and solvent B (0.1% formic acid in acetonitrile) at a flow rate of 250 µL/min with a linear gradient (5% B at 0–5 min, 5–45% B at 5–40 min, 45–90% B at 40–45 min, 90% B at 45–49.9 min, 90–5% B at 49.9–50 min, and 5% B at 50–60 min). Mass spectrometer was operated in a positive ion mode with 2-ppm mass accuracy. Mass spectra were acquired at a resolution of 60,000 in a detection range of M/Z 100–700. Acquisition parameters were set as follows: spray voltage 4.5 kV, capillary voltage 44 V, tube lens voltage 150 V, capillary temperature 250 °C, sheath flow rate 25, and auxiliary gas flow rate 5.
JZ, MM, JP and MAGK conceived the design of this study. JZ, YZ, MG, XW and SC performed experiments and analyzed data. JZ and XW wrote the manuscript. MAGK revised the manuscript. All authors read and approved the final manuscript.
The authors acknowledge Dr. Andrew Jones and Dr. Brady Cress for kind discussions. It is stated that Jens Plassmeier participated in this work while still an employee at Conagen.
A patent related to this work is under preparation.
Availability of data and materials
The data supporting our findings can be found in the main paper and the additional file.
Consent for publication
Ethics approval and consent to participate
Funding for this study was provided by Phase I STTR Grant (Award Number 1549767) from the National Science Foundation.
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
- 27.Kallscheuer N, Vogt M, Bott M, Marienhagen J. Functional expression of plant-derived O-methyltransferase, flavanone 3-hydroxylase, and flavonol synthase in Corynebacterium glutamicum for production of pterostilbene, kaempferol, and quercetin. J Biotechnol. 2017;258:190–6.CrossRefPubMedGoogle Scholar
- 28.Venkiteswaran S. Using genetic and molecular techniques to study anthocyanin production and protein expression in recombinant Escherichia coli. Buffalo: State University of New York at Buffalo, Department of Chemical and Biological Engineering; 2011.Google Scholar
- 44.Jaakola L, Määttä K, Pirttilä AM, Törrönen R, Kärenlampi S, Hohtola A. Expression of genes involved in anthocyanin biosynthesis in relation to anthocyanin, proanthocyanidin, and flavonol levels during bilberry fruit development. Plant Physiol. 2002;130:729–39.CrossRefPubMedPubMedCentralGoogle Scholar
- 55.Nielsen ILF, Haren GR, Magnussen EL, Dragsted LO, Rasmussen SE. Quantification of anthocyanins in commercial black currant juices by simple high-performance liquid chromatography. Investigation of their pH stability and antioxidative potency. J Agric Food Chem. 2003;51:5861–6.CrossRefPubMedGoogle Scholar
Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.