Introduction

Metabolic engineering by remodeling metabolic system of the cultured organisms has become a critically important tool to synthesize bioactive secondary metabolites. There are indeed more substantial numbers of biofuels, drugs, and industrially essential chemicals being produced by metabolomics-modified organisms [2, 4, 25, 29, 41, 51]. Monoculture systems have generated most of these bioactive compounds under optimal conditions. Although monoculture system is an efficient way to produce platform chemicals, it suffers from several disadvantages including lack of an optimal environment for the functioning of all pathway-specific enzymes, the increment in metabolic burden due to reconstruction or heterologous expression of complex pathways, and the formation of undesired byproducts [56, 58].

Several secondary metabolites with high industrial and medicinal values are produced by microorganisms and plants [53, 54]. Among them, flavonoids are a class of ubiquitous metabolites found in different plant species with profound bioactivity. Most of them exist in their glycosylated form with more stability, bioactivity, and solubility. However, aglycones are also present in large quantities [21]. Among various flavonoid reported, apigenin (API) and its derivative have been proven to be a well-known anti-aging, anti-fungal, anti-tumor, and anti-inflammatory agents [27, 28, 44]. Moreover, its derivative apigenin-7-O-β-d-glucopyranoside, also known as apigetrin (APG), has shown additional anti-proliferative and anti-oxidant activity against reactive oxygen species (ROS) [36, 44]. Besides glycosylation, hydroxylation, malonylation, methylation, and sulfation are common modifications that diversify flavonoids. Therefore, there is an increasing demand for these molecules in clinics and food-processing practices. Since lower solubility in water, short retention time in the intestine, and lower absorption rate restrict the consumption of flavonoids, the development of water-soluble flavonoids has been implicated for the treatment of many medical problems [15]. Glycosylation is one of the ways to enhance water solubility, the stability, and pharmacokinetic properties of flavonoids [21, 22, 46]. The biological activities of most natural products are attributed to the sugar moieties attached to them [11]. While chemical synthesis is used to synthesize such derivatives, the use of site-specific glycosyltransferases (GTs) that catalyzes the formation of the glycosidic bond are often used for the in vitro and in vivo synthesis of glycosides [5, 6, 8, 14, 47, 49]. Due to the development of chemical and molecular biology techniques, APG has been synthesized by chemical, semi-, and biosynthetic approaches [26, 32, 38, 50]. Among them, most biological approaches for APG production were carried out using plant-originated genes including TAL, 4CL, CHS, CHI, FNSI (or FNSII), and GTs via reconstruction in recombinant plasmids and hosts such as E. coli, Bacillus, and Saccharomyces cerevisiae to construct a monoculture system [13, 19, 23, 35].

Recently, the co-culture system simultaneously culturing two distinct metabolically engineered strains for production of desired metabolites has become an exciting and alternative tool. Numerous studies have reported the use of E. coli, Bacillus, and S. cerevisiae as a co-culture system to produce plenty of known bioactive compounds such as flavonoids, alkaloids, terpenoids, etc. [7, 10, 18, 30, 55,56,57]. Therefore, in the present study, we described the use of a co-culture system by genetically engineered E. coli for the production of APG. To achieve this endeavor, the culture system was compartmentalized into two parts: upstream strain containing biosynthetic pathway for production of API and downstream strain for overproduction of UDP-glucose and heterologous expression of a glycosyltransferase (Fig. 1).

Fig. 1
figure 1

Scheme of E. coli co-culture for production of APG

Full size image

Materials and methods

Culture media and chemicals

LB (Luria–Bertani) and M9 minimal medium (per liter, 6 g Na2HPO4, 3 g K2HPO4, 0.5 g NaCl, 0.1% NH4CL, 1% glucose, 1 mM MgSO4·7H2O, 100 µM CaCl2, 1.46 mg biotin, 1 mg thiamine-HCl) containing 2% glucose were used for seed culture and substrate fermentation, respectively. Antibiotics including ampicillin, kanamycin, and chloramphenicol (Biobasics, Canada) were used in the final concentration of 100, 50 and 30 µg/mL, respectively. Ethyl acetate, methanol, acetonitrile, and dimethyl sulfoxide (DMSO) were purchased from Merck (Germany) or Sigma (USA). Nanodrop 2000 UV–Vis Spectrophotometer (Thermo, USA) and HPLC–PDA (Agilent, USA) were used for chromatographic analysis.

DNA manipulation, bacterial transformation, recombinant plasmid reconstruction

DNA manipulation such as DNA plasmid extraction, purification, digestion, and ligation followed standard protocols as described by [43]. pACYC-Deut-1 containing chalcone synthase (CHS) and chalcone flavanone isomerase(CHI) from Petunia hybrida was previously described [24, 31]. 4CL-2 [4-coumarate: CoA ligase (4CL2) from Nicotiana tabacum cv. Samsun] gene was digested from the pAC-4CL-STS plasmid by NcoI/NotI restriction enzyme and then transferred to the multicloning site 1 (MCS1) of pCDFDuet-1 bearing streptomycin resistance gene (Str) [3]. Similarly, flavone synthase I (FNSI) from parsley Petroselinum crispum cv was digested with NdeI/XhoI from pET15b [33] and cloned into MCS2 of this duet vector. Those vectors were then introduced into E. coli BL21(DE3) to construct the recombinant E. coli API to produce API. In another way, E. coli MA3 strain, previously constructed for enhancing production of UDP-glucose [40, 48], was transformed with the recombinant pQE-30 vector containing glycosyltransferase-encoded gene (PaGT3) [37, 39] to generate E. coli GLP. All strains and DNA plasmids were listed in Table 1.

Table 1 Genes and strains in this study
Full size table

Use of E. coli GLP strain for the synthesis of apigetrin

Escherichia coli GLP strain was cultured in 3 mL of LB medium, 220 rpm, and 37 °C for overnight. Then, 500 µL of this seed culture was transferred to 250 mL flask containing 50 mL minimal M9 medium plus 2% glucose and maintained at 37 °C, 220 rpm until its optical density (OD) at 600 nm achieved 0.6. Total 1 mM IPTG was then added and continuously cultured under the same condition for 3–4 h for induction of protein. Total 100 µM purified API was supplemented to the induced broth and continuously maintained until 60 h. Bioconversion of substrate was checked at 12 h of interval time.

Synthesis of apigetrin and apigenin from p-coumaric acid

Escherichia coli API was used to produce API by feeding various concentration of p-coumaric acid (PCA). Similarly, minimal M9 plus 2% glucose was used for culture with 37 °C and 220 rpm. To increase the bioconversion of substrate, PCA was supplemented gradually into the induced E. coli API culture broth to avoid its growth inhibition by excessive substrate, i.e., 20, 20, 30, and 30 µM of PCA. Moreover, the sample was taken out at 12-h intervals for production testing.

In another way, the E. coli MA2 strain [40, 48] was transformed with various recombinant vectors (pAC-4CL-FNSI, pAC-CHS_CHI, and pQE30-PaGT3) to generate E. coli APG strain and used for the synthesis of APG. PCA was also gradually supplemented to culture to study the dependence of substrate bioconversion upon a time. Mainly 20, 20, 30, and 30 µM PCA was gradually added at 12, 24, 36 and 48 h, respectively, and the broth culture was maintained until 60 h.

Experimental co-culture for production of apigetrin

Effect of initial inoculum ratio on the production of apigetrin

Escherichia coli API and GLP were separately cultured by 3 mL of LB broth medium containing appropriate antibiotics at 37 °C and 220 rpm for overnight. Five hundred µL (unless stated) of each strain was subsequently diluted into 50 mL of M9 minimal media supplementing necessary antibiotics in 250 mL flasks; then the flasks were incubated at 37 °C and 220 rpm till the OD600nm of 0.6. The culture broths were centrifuged, and the cell pellets were resuspended in 50 mL of M9 media plus 2% glucose. The two populations were mixed in different ratios by OD, i.e., various ratios of API/GLP = 1: 1; 2:1; 4:1; 6:1; 8:1 and 10:1 (v/v) to generate a co-culture system. The broth culture was added with 1 mM IPGT and continuously cultured for 3–4 h at 37 °C, and 220 rpm to induce protein expression. Subsequently, the system was supplemented by total 100 µM PCA, i.e., 20, 20, 30, 30 µM at 12, 24, 36, and 48 h. Formation of NRN, API, and APG was investigated at interval time of 12 h until 60 h of culture. Similarly, inverse various ratios (in which API was kept in constant, but GLP was varied) of API/GLP, i.e. 1:2, 1:4, 1:6, 1:8, and 1:10 were also used for evaluating the production of APG under the same conditions and substrate feeding as described above.

Effect of culture temperature

Various range of culture temperature from 25 to 37 °C, i.e., 25, 27, 30, 32, and 37 °C was used to investigate the effect on the production of API and its derivatives. The total PCA concentration of 100 µM was supplemented to the culture under 220 rpm. The converted naringenin, API, and APG concentration were checked after interval time of 12 h until 60 h of culture.

Analysis and quantification

HPLC–DAD analysis was performed by injecting 20 µL of the samples on an Agilent 1260 HPLC system equipped with a photodiode array detector (DAD), degasser, an autosampler. An Agilent Zorbax SB C18 column (250 mm × 4.6 mm i.d., 5 μm; Agilent, Santa Clara, CA, USA) was used. The mobile phase of 0.1% trifluoroacetic acid (TFA) aqueous solution (solvent A) and acetonitrile (solvent B) were used with 1 mL/min flow rate. The concentration of acetonitrile during the binary gradient condition was as: 0–10 min, 0–60%; 10–20 min, 75%; 20–35 min, 75–90%. Peak detection was carried out at UV absorbance at 295 nm whereas.

Naringenin, API, and APG were purified using an MPLC instrument equipped the silica gel RP-packed column (YMC gel ODS-A, AA12SA5, Japan). The corresponding peaks were collected and then lyophilized by freeze drier (FDU-1200, Eyela, Japan). All chemicals were dissolved in DMSO-d6 and subjected to NMR analysis. A series of concentrations ranging from 10 to 100 mg/L of the product were prepared to construct a calibration curve. The molecular mass of the compounds was determined in LC–ESI–MS using Phenomenex Synergi Polar-RP column (150 × 4.6 mm, 4 µm), positive-ion mode.

Statistical analysis

The student’s t test was performed on the biological replicates to determine the statistical significance of the difference between control and experiment samples at each time point. Differences with a P value < 0.05 were considered statistically significant.

Results

Production of apigenin from p-coumaric acid using E. coli API strain

Since glucose increases the intracellular acetyl-CoA concentration via glycolysis and acts as the primary carbon source for flavonoid production [45], instead of LB, M9 minimal media with 2% glucose was used to restrict the rapid degradation of PCA [31]. The recombinant E. coli harboring biosynthetic gene cluster for API production was constructed using 4-coumarate: CoA ligase (4CL2) gene from Nicotiana tabacum cv. Samsun [3], chalcone synthase (CHS) and chalcone flavanone isomerase(CHI) from Petunia hybrida and Medicago sativa, respectively, as described early, and flavone synthase I (FNSI) from parsley Petroselinum crispum cv. [33]. When cultured at 37 °C, 220 rpm, and fed with 20 µM of PCA, the formation of naringenin (NRN) and API were observed by HPLC analysis (Fig. 2b). Further analysis by LC–ESI–MS resulted in the mass – charge ratio (m/z) of [M + H]+ of 272.9 and 270.8 for NRN and API in positive mode, respectively (Fig. 3a, b). Also, the time and concentration-dependent study showed highest concentration of API (31.8 µM) at 60 h, when 100 µM PCA was fed entirely (Fig. 4). Moreover, the yield of conversion was not improved when a higher concentration of PCA was supplied into the broth culture.

Fig. 2
figure 2

HPLC trace analysis of the biotransformation products of recombinant E. coli recombinant strains. a authentic PCA and API standard compounds, b E. coli API supplemented with the substrate PCA for production of API, c synthesis of APG by E. coli APG strain utilizing PCA as initial substrate, d E. coli GLP used for bioconversion of API to APG, and e co-culture of API and GLP (ratio of 1:2) for production of APG. triangle, PCA peak; star, APG peak; diamond, API peak; and 4 pointed star shape, naringenin peak

Full size image
Fig. 3
figure 3figure 3

LC-ESI–MS profile of products in positive mode: a naringenin, mass-to-charge ratio (m/z) of [M + H]+ = 272.9; b API, m/z of [M + H]+ = 270.8; c APG, m/z of [M + H]+ = 433.0

Full size image
Fig. 4
figure 4

Production of API by E. coli API strain, during 60 h of culture, using PCA as the substrate

Full size image

Synthesis of apigetrin using E. coli APG strain using PCA as initial substrate

To avoid growth inhibition of E. coli by extra-cellular addition of PCA [51], 10 µM PCA was initially supplemented to the induce broth culture. Post substrate addition after 5 h, the production of NRN, API, and APG were analyzed. As evident from Fig. 3c, we observed the presence of APG peak with retention time of 20.50 min in HPLC chromatogram (Fig. 2c) with m/z of [APG + H] + = 433.0. Further analysis of bioconversion at 12 h interval showed 77.7% PCA to be consumed, while the maximum yield of APG, NRN, and API reached 15.5, 32.8, and 27.4 µM after 60-h culture, respectively (Fig. 5).

Fig. 5
figure 5

Production of APG in E. coli APG strain, during 60 h of culture, using PCA as the substrate

Full size image

Biotransformation of API to APG by E. coli GLP

To compare the yield of APG production, 10 µM API was initially supplemented to induced culture broth, which gradually increased up to 100 µM in total at 48 h maintaining till 60 h. As evident from HPLC chromatogram monitored at the 12-h interval, the maximum accumulation of APG reached to 45.8 µM (19.8 mg/L) after 60 h (Figs. 2d, 6), where the rate of biotransformation of API to APG was higher than a synthesis of APG from PCA.

Fig. 6
figure 6

Bioconversion of API to APG using E. coli GLP as the whole-cell biocatalyst

Full size image

De novo co-culture for production of APG

Optimization of inoculum ratio in co-culture

Co-culture for the production of natural products requires testing proper ratio of initial inoculum because each strain was designed to carry specific work. Mainly, E. coli API and GLP were stepwise reconstructed to synthesis API and APG.

By changing the ratio of API/GLP from 1:1 to 10:1, the production of APG concentration was achieved to maximum, yielding of 32.5 µM (14.1 mg/L) or 2.1-fold compared to the use of mono-culture E. coli APG (Figs. 2e, 7). Hence, E. coli GLP expressed the strong bioactivity of PaGT3 again as proved by biotransformation of API to APG too. Moreover, the higher the ratio of E. coli API the more production of APG. Concurrently, the more production of API the more consumption of initial substrate, PCA.

Fig. 7
figure 7

Effect of primary inoculum culture on the production of APG by various types of E. coli co-culture

Full size image

In contrast, the ratio of API/GLP was changed from 1/2 to 1/10 the maximum concentration of APG has achieved the highest yield of 23.4 µM at a ratio of 1/10 and gradually decreased to 15.6 µM at the ratio of 1/2. Therefore, it was observed that higher APG concentration in accompanies with a higher ratio of GLP/API. However, the ratio of E. coli API was constant then the maximum yield of APG just achieved 72% (23.4/32.5 × 100) compared to the previous case. Based on above result analysis the ratio of API/GLP was selected between 8 and 10 for further experiments.

Effect of temperature on co-culture for production of APG

Temperature is one of the most important factors that affect bacterial growth and heterologous expression of enzymes. It has been proven that 35–37 °C is a suitable range for E. coli culture. However, glycosyltransferases are optimally expressed at a lower temperature. Thereby, various range of temperature was tested for co-culture to obtain higher production of APG. As a result, API was strongly synthesized, PCA concentration was fast consumed resulting in the highest yield of APG of 38.5 µM (16.6 mg/L) (Fig. 8) when the range of temperature maintained between 25 and 32 °C. However, the production of API was not increased when the temperature reached 37 °C with a maximal yield of APG was 32.5 µM.

Fig. 8
figure 8

Effect of various types of culture temperature on production of APG by E. coli co-culture

Full size image

Structural confirmation of NRN, API, and APG

The structures of API, APG, and NRN were confirmed by 1D and 2D NMR. We observed two meta-coupled [δH 6.19 (H-6) and 6.48 (H-8), each 1H, d, J = 2.0 Hz] and four ortho-coupled [δH 7.92 (H-2′ and H-6′) and 6.92 (H-3′ and H-5′), each 2H, d, J = 9.0 Hz] aromatic protons of API in 1H NMR spectrum. Also, the signals of one olefinic [δH 6.76 (1H, s, H-2)] and one chelated hydroxyl [δH 12.95 (1H, s, 5-OH)] protons were observed. The 13C NMR spectrum of API exhibited signals of 15 carbons for a flavone skeleton (Table 2). Comparison of the 1H and 13C NMR data of API with the reported values [1, 49] as well as detailed analysis of its HMBC correlations (Fig. 9) elucidated its structure as 5,7,4′-trihydroxyflavone. Similarly, the other compounds were characterized as APG (apigenin-7-O-β-d-glucoside) [12, 49] and 2S-5,7,4′-trihydroxyflavanone (naringenin) [52], which is in good agreement with their 1H and 13C NMR chromatograms, earlier report, and combination with their HMBC evidence (Figs. S2, S3, and S4, supporting information).

Table 2 1H NMR (500 MHz, DMSO-d6) and 13C NMR (125 MHz, DMSO-d6) data of API, APG and NRN
Full size table
Fig. 9
figure 9

Structures and critical HMBC (H → C) correlations of NRN (a), API (b) and APG (c)

Full size image

Discussion

It is observed that flavonoids are stable in the form of glycosides (O-, C- or N-glycosidic linked derivatives) in vascular plants. Till today, synthesis of flavonoid glycosides was carried out by a chemical approach with several disadvantages such as protection of un-reacted moiety, generation of various derivatives under severe conditions [50]. However, glycosyltransferase acts as a biocatalyst to catalyze the attachment of glycosyl moiety to various flavonoid acceptor in vitro and in vivo. Microbial synthesis of glycosides by mono-culture requires the construction of recombinant plasmids with numbers of genes, inhibition of bacterial chemo-physiological process, and challenge to achieve optimal culture conditions. For example, biological synthesis of flavanones (naringenin, pinocembrin) by recombinant mono-culture E. coli has been carried out by [16, 19, 24, 35]. Also, flavones (API, chrysin) have been synthesized by FNSI [34] including its derivative, genkwanin [23]. Therefore, as an alternative approach, co-culture system was designed in this study to fix the bottlenecks of monoculture. For example, resveratrol, afzelechin, muconic acid, and 3-aminobenzoic acid have been successfully produced by co-culture approach [9, 18, 55, 57]. Here, we have experimentally biosynthesized API and APG from PCA by monoculture system, where the maximal yield of each metabolite was 31.8 (8.6 mg/L) and 15.5 µM (6.7 mg/L), respectively. Besides, API substrate was also used for direct biotransformation by E. coli GLP and resulted in 45.8 µM of APG (19.8 mg/L). Furthermore, co-culture based on the modular system including upstream (E. coli API) and downstream (E. coli GLP) was optimized by the ratio of inoculum and temperature condition. By changing the various ratio of inoculum API/GLP, it resulted in the maximal yield of APG of 32.5 µM. Furthermore, the best yield of APG was achieved (38.5 µM or 16.6 mg/L) under optimal temperature (32 °C). Thereby, it meant increment in 2.5-fold of APG compared to mono-culture E. coli APG (16.6 mg/L), however, still lower than direct biotransformation of API by E. coli GLP (19.8 mg/L). For reference, the yield of API was mentioned in different records such as 30 mg/L [23] or 13 mg/L [34] via mono-culture. In comparison to other literature, the titer of afzelechin and resveratrol was achieved 40.7 ± 0.1, and 22.6 mg/L by E. coli co-culture; our yield of APG was in comparative level.

To further increase the titer of APG, several works need to be continuously carried out including strain improvement of E. coli API such as over-expression of malonyl-CoA [25, 34] or biosynthesis pathway of naringenin from the direct amino acid by the introduction of tyrosine ammonium lyase (TAL) or phenylalanine ammonia-lyase (PAL) [34, 35]. Similarly, improvement of E. coli GLP can be achieved by strengthening of PaGT3 via site-directed mutagenesis [20, 42]. Finally, it is necessary to optimize the culture conditions for co-culture system [17, 18].

Conclusions

Bioactive compounds are becoming essential in clinical practices, production of cosmetics, adjuvant, antibiotics as well as functional foods. Therefore, biosynthesis of bioactive compounds has evolved from simple to sophisticated approaches. Combination of genetic engineering, recombinant protein, metabolic engineering, and fermentation technology creates a novel ways for biosynthesis of these valuables including co-culture system. Our observations showed co-culture as an applicable and reproductive approach, which can be used to solve the bottlenecks of monoculture, yielding higher production of APG under simple conditions.