Background

Isoprenylated quinones, in which the length of the isoprenoid side chain varies, are widely distributed in almost all living organisms [1]. One isoprenylated quinone of particular interest is vitamin K2, a well-known high-value product. Vitamin K2 (menaquinone) is a series of vitamers with multiple isoprene units at the 3′ position of the naphthoquinone ring structure. Menaquinones have variable side chain lengths of 4–13 isoprene units, referred to as MK-n, where M stands for menaquinone, K stands for vitamin K, and n denotes the number of isoprenoid residues [2, 3]. Vitamin K is necessary for the functional modification of proteins that are involved in hepatic blood anticoagulation, the maintenance of bone health and cardiovascular health, cancer prevention, inflammation suppression, prevention of oxidative damage to the brain, sphingolipid synthesis, osteoporosis prevention, and even treatment for mitochondrial pathologies, such as Parkinson’s disease and amyotrophic lateral sclerosis [4,5,6,7,8]. To date, some of the widely examined vitamin K2-producing microorganisms include Bacillus subtilis natto, B. subtilis 168, Elizabethkingia meningoseptica, and Escherichia coli. Among them, naphthoquinone-type menaquinone-4 (MK-4), MK-5, and MK-6 have been previously identified in E. meningoseptica [9, 10]. E. coli primarily produces a benzoquinone-type ubiquinone-8 (UQ-8) under aerobic conditions but mainly synthesises MK-8 under anaerobic conditions [11].

To date, isopentenyl pyrophosphate (IPP) and its isomer dimethylallyl pyrophosphate (DMAPP), which is the precursor with variable side chain lengths, can be produced mainly via two different pathways: the methylerythritol 4-phosphate (MEP) pathway and the mevalonate (MVA) pathway [12]. Although stoichiometric studies have revealed that the 1-deoxy-d-xylulose-5-phosphate (DXP) pathway is more efficient than the MVA pathway, the production of isoprenoids by optimising the DXP pathway cannot go beyond the levels accomplished by expressing a heterologous MVA pathway in prokaryotes [13,14,15].

We sequenced the E. meningoseptica sp. F2 genome, and the results of KEGG pathway analysis revealed that E. meningoseptica sp. F2 utilises the MVA pathway (Additional file 1: Fig. S1) [16]. Previous studies have shown that triggering the MVA pathway could increase the level of terpenoid synthesis [17, 18]. However, there are only a few reports on the molecular modification of E. meningoseptica, and the main problems are listed as follows: (1) E. meningoseptica is resistant to multiple antibiotics, such as extended-spectrum β-lactam agents and aminoglycosides [19, 20], (2) E. meningoseptica lacks available molecular tools for molecular manipulation. Compared with other expression systems, E. coli remains the preferred host for producing recombinant proteins because of several advantages, including rapid, inexpensive, and high-yield protein production, due to the well-characterised genetics and variety of available molecular tools [21]. Several studies have explored the benefits of developing microbial consortia based on the engineering of microbial chassis [22,23,24,25]. Two E. coli strains or E. coli strains co-cultured with other strains were constructed individually to accommodate different pathway modules, which helped reduce the overwhelming metabolic stress on each strain [26, 27]. Therefore, the microbial consortium (engineered E. meningoseptica and E. coli) may be a better strategy to solve the aforementioned problems.

In this study, vitamin K2 was produced by the E. meningoseptica and E. coli co-culture system by introducing the MVA pathway and prenyltransferase genes from E. meningoseptica sp. F2 into E. coli. Compared to a single strain, the co-culture system performed effectively in the production of vitamin K2. To the best of our knowledge, this is the first study to report the usefulness of a co-culture system for vitamin K2 production.

Results and discussion

Engineering E. meningoseptica and E. coli co-culture systems to produce MK-n

Based on our previous antibiotic susceptibility tests, E. meningoseptica sp. F2 was resistant to ampicillin (0–200 µg/mL), caramycin (0–200 µg/mL), tetracycline (0–150 µg/mL), chloramphenicol (0–150 µg/mL), gentamicin (0–200 µg/mL), and streptomycin (0–200 µg/mL). These results also revealed that E. meningoseptica sp. F2 was sensitive to chloramphenicol (approximately 100–150 µg/mL) and tetracycline (approximately 75–150 µg/mL) [16]. To the best of our knowledge, fewer molecular modification tools for E. meningoseptica have been reported; therefore, performing molecular modifications on E. meningoseptica sp. F2 directly may be difficult. Considering that E. coli has many advantages from quick growth to high cell densities, a wide range of genetic tools and many expression vectors have been commercialised. More importantly, several researchers successfully used an E. coli co-culture system to improve target products [26]. In addition, E. coli has a complete vitamin K2 synthesis pathway (KEGG pathway: eco00130). Thus, we selected the method of co-cultivation of E. meningoseptica and E. coli to overcome these problems and improve vitamin K2 biosynthesis.

The products of MK-n (n = 4, 5, 6, 7, 8) were detected by co-culture of E. meningoseptica and E. coli (Additional file 1: Fig. S2). Previously, we identified that E. meningoseptica can produce MK-4, MK-5, and MK-6 and that E. coli can produce MK-8 [28, 29]. As shown in Fig. 1, the MK-n (n = 4, 5, 6, 7, 8) production and cell concentration [dry cell weight (DCW)] were measured as 14.08 ± 0.81 mg/L and 11.37 ± 0.28 mg/L, respectively. Among them, the MK series compounds were identified as MK-6 (approximately 75%) and MK-7 (approximately 10%). The results of the KEGG pathway analysis showed that E. meningoseptica sp. F2 did not contain either hexaprenyl diphosphate synthase (C30) or heptaprenyl diphosphate synthase (C35). The results of the KEGG pathway analysis revealed that geranylgeranyl diphosphate synthase (EmGGPPS) [EC: 2.5.1.12.5.1.10.2.5.1.29] and octaprenyl pyrophosphate synthase (EmOPPS) [EC: 2.5.1.90] existed in E. meningoseptica sp. F2 (Additional file 1: Fig. S1). In a previous study, we identified that EmGGPPS can catalyse consecutive condensation reactions of substrate molecules to form a C20 short-chain product GGPP [30]. Nevertheless, how the synthesis of hexaprenyl diphosphate in E. meningoseptica occurs remains unclear. E. meningoseptica sp. F2 possesses MK-6 as the major menaquinone component, and the results of the KEGG pathway analysis did not show the presence of hexaprenyl diphosphate synthase (HexppS). Therefore, we speculate that the biosynthesis of MK-6 may be related to EmOPPS in E. meningoseptica sp. F2. We also analysed the enzyme activity of EmOPPS derived from E. meningoseptica sp. F2.

Fig. 1
figure 1

Production of vitamin K2 by the monoculture and co-culture systems of Escherichia coli and Elizabethkingia meningoseptica sp. F2. The error bars represent the standard error of at least three biological replicates. *Statistical significance (p < 0.05) compared to the monoculture system of E. coli and E. meningoseptica sp. F2

Full size image

Expression and purification of a recombinant version of EmOPPS

To further investigate the significance of EmOPPS in E. meningoseptica sp. F2, the amino acid sequences of several trans-prenyltransferases, including C40-OPPS from E. coli K-12 MG1655 (P0AD57_Ecoli), E. meningoseptica sp. F2 (A0A1V3U058_ELIME), and Haemophilus influenzae Rd KW20 (P44916_HAEIN) and C35-HexppS from B. subtilis natto BEST195 (D4FY42_BACNB), were aligned with Clustal X and showed using ESPript. The results are shown in Fig. 2A. Two DDXXD motifs in the amino acid sequence were found in trans-prenyltransferases. The first motif is responsible for binding with farnesyl diphosphate (FPP), and the second with isopentenyl diphosphate (IPP) [31,32,33]. The amino acid located in the fifth position before the first DDXXD is alanine in the E. meningoseptica OPPS, E. coli OPPS, H. influenzae OPPS, and B. subtilis Hexpps and is important in determining the product’s chain length [34].

Fig. 2
figure 2

Bioinformatic and SDS-PAGE analysis of EmOPPS. A Alignment of amino acid sequences of prenyltransferases. The black box on a colour background indicates similar amino acid residues. The two conserved DDXXD motifs are indicated. B SDS-PAGE analysis of recombinant protein levels in Escherichia coli BL21 (DE3) cells at 30 °C for 4 h. Lanes 1, 3, 5, 7: supernatant. Lanes 2, 4, 6, 8: precipitate

Full size image

We constructed the plasmid PETA-1 (pET-28a-GB1-EmOPPS) and expressed it in E. coli BL21 to produce the strain J01. PETD-1 (pETDuet-1-MBP-EmGGPPS) was successfully expressed. To supplement the precursor of the side chain, we also synthesised GB1-OPPS on the second multiple cloning site of pETDuet-1 to form PETD-2 and transformed it into E. coli BL21, yielding strain J02, which can produce GB1-EmOPPS and MBP-EmGGPPS. The cells were induced with a low concentration of isopropylβ-D-1- thiogalactopyranoside (IPTG, 50 µM) and incubated at 30 °C for 4 h. The molecular masses of GB1-OPPS and MBP-GGPPS were 45.9 kDa and 80.4 kDa (Fig. 2B), respectively. To further analyse GB1-EmOPPS, we purified the proteins (Additional file 1: Fig. S3A). GB1-OPPS was purified using Ni-nitriloacetic acid (Ni–NTA) resin using an internal 6 × histidine affinity tag. Purified GB1-EmOPPS proteins were confirmed by mass spectrometry (Additional file 1: Fig. S3B).

Enzymatic assay to detect EmOPPS activity and MK-n production in E. coli

To functionally characterise EmOPPS, affinity-purified His-tagged recombinant proteins were incubated with IPP and FPP as co-substrates. After the reaction, we hydrolysed the diphosphate product using phosphatases to convert it into the corresponding alcohol, which was then detected using liquid chromatography-mass spectrometry (LC–MS). As shown in Additional file 1: Fig. S4A and B, two products [C25H42O (m/z = 358) and C30H50O (m/z = 426)] were detected (Additional file 1: Table S1). This suggests that C25 and C30 are the products of EmOPPS with IPP and FPP as substrates. This is different from previously characterised OPPS, which catalyse consecutive condensation reactions of FPP with five molecules of IPP to generate C40 octaprenyl diphosphate [35, 36]. To further test the enzymatic activity of EmOPPS in vivo, we utilised the vitamin K2 biosynthesis system in E. coli. Previous research has shown that E. coli mainly synthesises one type of vitamin K2, menaquinone-8 (MK-8), under micro-anaerobic conditions [29]. MK-7 can also be produced by heterogeneous heptaprenyl pyrophosphate synthetase (HepPPS) [37]. In this study, LC–MS revealed the accumulation of MK-6, -7, and -8 by strain J01 (Additional file 1: Fig. S4C, D, Fig. S5). Combining with the results of the enzyme activity test, we concluded that the biosynthesis of MK-6 (m/z = 581) is related to the activity of EmOPPS. Therefore, this assay demonstrated that EmOPPS synthesised hexaprenyl diphosphate (C30) both in vivo and in vitro. In the coagulation experiments, MK-4, -5, -6 exhibited high biological activity; with the increase of the number of isoprene units, the biological activity decreased [38]. Currently, most studies focus on microbial biosynthesis of MK-7, due in part to longer biological half-life and higher yield [39, 40] and less studies on MK-6. Therefore, our findings are meaningful for the study of vitamin K2.

The synthesis of one molecule each of C30, C35, and C40 required several molecules of IPP and one molecule of FPP. For FPP synthesis, the reaction of two molecules of IPP and one molecule of DMAPP was catalysed by GGPPS. Thus, we cloned EmGGPPS in pETDuet-1 to form PETD-1, and then the amplified fragments of GB1-EmOPPS were also cloned into the PETD-1 vector to form PETD-2. Finally, this vector was transformed into E. coli to form the J02 strain. This strain could biosynthesise MK-n (n = 4, 5, 6, 7, 8) (Fig. 3A, B, C, Additional file 1: Fig. S5, Additional file 1: Table S1). In a previous study, we confirmed that GGPPS can catalyse IPP and DMAPP to produce FPP and synthesise MK-4 in E. coli [30]. Thus, the appearance of MK-4 in the LC–MS results was related to the expression of EmGGPPS. This is the first time that a series of vitamin K2 (MK-n, n = 4, 5, 6, 7, 8) was synthesised simultaneously in E. coli.

Fig. 3
figure 3figure 3

Menaquinone-n (n = 4, 5, 6, 7, 8) production in the engineered Escherichia coli (J02 strains). LC–MS analysis of the production of J02 strains (A–C). A HPLC chromatograms of different MK-n B, C mass spectra of MK-n, MK-4 (m/z = 445), MK-5 (m/z = 513), MK-6 (m/z = 581), MK-7 (m/z = 649), and MK-8 (m/z = 717)

Full size image

Co-cultures of E. meningoseptica sp. F2 and E. coli J02 (CO1 system) were investigated. This consortium, with a 3:1 inoculation ratio, was able to produce a titre of 16.74 ± 0.58 mg/L of vitamin K2 (Fig. 4A). In addition, a previous study has shown that the inoculation ratio of different strains has a greater influence on the co-cultivation system for the target product [41]. The variation in population ratio changes population dynamics, which leads to the optimal function of each pathway module for the efficient conversion of substrates to the products with little to no accumulation of intermediate metabolites [42]. Therefore, in this study, we optimised the inoculation ratio of E. meningoseptica sp. F2 and E. coli. The highest vitamin K2 production measured was 22.62 ± 0.75 mg/L and was achieved with the inoculation ratio of 3:3 (Fig. 4A). The vitamin K2 concentration was two-fold higher than that of the monoculture of E. meningoseptica sp. F2.

Fig. 4
figure 4

Co-culture compatibility determination and upstream strain optimisation. A Optimisation of the inoculation ratios for the CO-1 cultivation system. Elizabethkingia meningoseptica and Escherichia coli were inoculated into the co-culture system at a ratio of 3:1, 3:2, 3:3, and 3:4 v/v. Statistical significance (p < 0.05) was determined by comparing with the initial inoculation ratio (3:1). B Different types of co-cultivation systems. The co-culture strains were inoculated at a ratio of 3:3. The error bars represent the standard error of at least three biological replicates. Statistical significance (p < 0.01) compared to the CO-1 cultivation system

Full size image

Identification of the mevalonate pathway of E. meningoseptica and the enhancement of vitamin K2 production by mevalonate supply from E. coli

Endogenous IPP and DMAPP levels are relatively low in E. coli cells and are not sufficient for the high expression of heterogeneous terpenoids [24, 26]. Previous study have also shown that the production of isoprenoids by modulating the DXP pathway cannot surpass the levels achieved by expressing a heterologous MVA pathway in prokaryotes [27]. In contrast, the MVA pathway for isoprenoid biosynthesis has been frequently engineered to produce valuable compounds [28]. To increase IPP supply, we introduced the MVA pathway genes in E. meningoseptica sp. F2 into E. coli. Bioinformatic analysis revealed that E. meningoseptica sp. F2 uses the MVA pathway to produce isoprene precursors; no studies have previously identified and analysed this pathway. Therefore, we introduced the MVA pathway genes of E. meningoseptica sp. F2 into E. coli and identified their products. First, the four genes EmHMGR, EmHMGS, EmPVD, and EmIDI, derived from E. meningoseptica sp. F2, as well as SceMK and ScePMK derived from Saccharomyces cerevisiae, were constructed into two plasmids: PETA-2 and pACD2. They were then introduced into the J02 strain to form the strain H01. Subsequently, we identified the mevalonate product of H01 strains, and the mass spectrometry results showed that the engineered strain H01 could metabolise and produce mevalonate and secrete it outside the cell (Additional file 1: Fig. S6). Co-cultures of E. meningoseptica sp. F2 and E. coli H01 (CO2 system) produced 25.51 ± 1.25 mg/L of MK-n (Fig. 4B).

The structure of menaquinone consists of a naphthoquinone ring and an isoprene side chain. The naphthoquinone ring is biosynthesised by seven enzymes encoded by the menFDHCEBA genes [39]. MenA encoded 1,4-dihydroxy-2-naphthoate octaprenyl transferase, which catalysed 1,4-dihydroxy-2- naphthoate to demethylmenaquinone as a rate-limiting enzyme. MenG/UbiE (MK-n biosynthesis methyltransferase) converted demethylmenaquinone to menaquinone. A previous study has shown that the overexpression of MenA and UbiE/MenG could be effective in improving vitamin K2 production [43]. Here, the gene fragments from EcMenA derived from E. coli K12 and EmMenE derived from E. meningoseptica sp. F2 were constructed into the plasmid PETA-2 to synthesise PETA-3. PETA-3 with pACD-2 was introduced into strain J02 to produce strain H02. When MenA or MenE/G was overexpressed, co-cultures of E. meningoseptica sp. F2 and E. coli H02 (CO3 system) produced 32.62 ± 1.1 mg/L of MK-n (n = 4, 5, 6, 7, 8), which was three-fold higher than that in the monoculture system of E. meningoseptica sp. F2 (Fig. 4B). As shown in Additional file 1: Fig. S7, under the same culture conditions, the yield of the E. coli H02 monoculture was 4.5 ± 0.13 mg/L, and the yield of the E. meningoseptica sp. F2 monoculture was 11.37 ± 0.28 mg/L. We speculate that strain E. meningoseptica sp. F2 was the dominant vitamin K2 producer in this co-culture system. These results support the further increase in vitamin K2 production using the microbial consortia platform of E. meningoseptica and E. coli. Figure 5 shows the co-culture design of E. meningoseptica sp. F2 and E. coli H02 (CO3 system) for the biosynthesis of MK-n (n = 4, 5, 6, 7, 8). The intracellular production of mevalonate by E. coli can be detected in the fermentation medium [44]. MVA can also be transported inside the cell as a precursor for the biosynthesis of downstream isoprenoids [15]. In addition, a previous study also indicated that the MVA feeding strategy could be more extensively utilised in the biosynthesis of isoprenoids [45]. Vitamin K2 is also an important isoprenoid product [46]; its isopentenyl side chain requires the MVA pathway to provide a precursor. In addition, both E. coli H02 and E. meningoseptica sp. F2 independently produced MK-n (n = 4, 5, 6, 7, 8). Furthermore, our previous study also implied that adding 1% polyoxyethyleneoleyl (POE) can accelerate the growth of microorganisms and improve the target product yield during the E. meningoseptica fermentation process [47]. Therefore, we added 1% surfactant to the co-culture system to increase the production of vitamin K2. Unexpectedly, the yield of MK-n decreased when 1% POE was added to the medium (Fig. 4B).

Fig. 5
figure 5

Design of Elizabethkingia meningoseptica and Escherichia coli co-culture system for vitamin K2 biosynthesis. (I) Mevalonate pathway: atoB acetoacetyl-CoA acetyltransferase; HMGS HMG-CoA synthase; HMGR HMG-CoA reductase; MK mevalonate kinase; MVD1 mevalonate pyrophosphate decarboxylase; idi IPP isomerase. MenA (1,4-dihydroxy-2-naphthoate octaprenyltransferase), MenG, and UbiE (Q-8/MK-8 biosynthesis methyltransferase) encoded the head structure biosynthesis enzymes that are involved in the vitamin K2 biosynthetic pathway starting from isochorismate. Gene names in red represent the overexpressed genes. (II) HPLC analysis of vitamin K2 products in E. meningoseptica and E. coli co-culture system

Full size image

The characterisation of the microbial consortium morphology

To investigate the presence of E. coli and E. meningoseptica sp. F2, particularly the co-existence of the microbes, the four groups (E. coli, E. meningoseptica, co-culture of E. meningoseptica and E. coli, and co-culture of E. meningoseptica and E. coli (+ POE) were visualised using scanning electron microscopy and confocal laser scanning microscopy, as shown in Fig. 6 and Additional file 1: Fig. S8. The scanning electron microscopy (SEM) images of the samples showed that E. meningoseptica and E. coli cells co-existed in the cultivation medium and that the two were closely connected. The monocultures of E.coli or E. meningoseptica sp. F2 are shown in Fig. 6A, B as controls, respectively. The size of the adjacent bacteria was uniform, and E. coli was partially autolysed and ruptured (Fig. 6A). E. coli cells were approximately 2 μm long and appeared as single straight rod bacteria. In another group, the size of the adjacent cells was generally uniform, and the shape of the cells was smooth but partially depressed (Fig. 6B). The cell volumes were smaller than those of E. coli (Additional file 1: Fig. S8A, B). The cells were non-spore-forming rods of approximately 1 μm in length. The co-existence of E. meningoseptica and E. coli displayed closely connected associations of large and small bacteria in a microbial consortium (Fig. 6C, Additional file 1: Fig. S8C, D). E. coli are gram-negative bacteria, with short rods, 2.0–6.0 μm in length; E. meningoseptica sp. F2 are gram-negative bacteria, with non-motile, non-spore-forming rods (1.0–2.0 μm length) [48,49,50,51]. Therefore, it can be seen that the longer bacteria is E. coli, and the other closely connected bacteria is E. meningoseptica sp. F2 (Fig. 6C, Additional file 1: Fig. S8C, D). The members of a consortium communicate by exchanging metabolites or signals that allow them to coordinate their activities through the division of labour [52]. Here, we speculate that the closely connected combination of bacteria may be more conducive to the exchange of materials and energy between different bacteria. Microbial interactions usually can be exploited to improve cell viability and productivity. Mutualistic interactions are established between the used strains; therefore, a relationship can benefit the interacting strains, thus promoting the overall performance [22, 53]. Furthermore, the number of cells of strain E. meningoseptica sp. F2 was far more than that of strain E.coli (Additional file 1: Fig. S8C and D). Confocal laser scanning microscopic observation also confirmed that strain E. meningoseptica sp. F2 was the dominant strain in this co-culture system. The yield of monoculture of E. meningoseptica sp. F2 was higher than that of monoculture of E. coli; therefore, we speculate that E. meningoseptica sp. F2 is the main vitamin K2-producing strain in this system. However, remarkable morphological changes were observed after the cells were subjected to POE treatment (Fig. 6D). Only a small part of the combined bacterial form was different in size and connection. In addition, there were several autolysed and ruptured bacterial fragments between the bacteria.

Fig. 6
figure 6

Visualisation of different inoculates using scanning electron microscopy images. A Escherichia coli. Blue circles surround the connected E. coli cells in view. B Elizabethkingia meningoseptica sp. F2. Green circles surround the connected E. meningoseptica sp. F2 cells in view. C E. meningoseptica and E. coli co-culture. Red circles surround the connected E. meningoseptica and E. coli cells in view. D E. meningoseptica and E. coli co-culture treated with 1% POE. Yellow circles surround the connected E. meningoseptica and E. coli cells in view

Full size image

Esterase activity and membrane permeability

Flow cytometry (FCM) was used to determine the cell permeability and esterase activity of the E. meningoseptica and E. coli co-culture. Dual parameter dot plots of E. coli BL21, E. meningoseptica sp. F2, CO3, and CO3 + POE cells were stained with carboxyfluorescein diacetate (cFDA) and propidium iodide (PI), as shown in Fig. 7. The subpopulations were identified based on their differential staining characteristics with PI and carboxyfluorescein (cF): PI-stained cells (cF-negative, PI-positive; Q1 quadrant) exhibited inactivated esterase and damaged membranes; PI and cF double-stained cells (cF-positive, PI-positive; Q2 quadrant) exhibited sublethal injuries with residual esterase activity and compromised membranes; cF-stained cells (cF-positive, PI-negative; Q3 quadrant) exhibited high esterase activity and intact membranes; the unstained area (cF-negative, PI-negative; Q4 quadrant) most likely corresponded to debris or lysed cells attributable to bacterial autolysis [54, 55].

Fig. 7
figure 7

Flow cytometry dot plots of monoculture system of Escherichia coli (A), B and Elizabethkingia meningoseptica sp. F2 as a reference and C co-culture strains of E. meningoseptica and E. coli cells treated without or D with 1% POE

Full size image

Approximately, 52.8% and 89.2% of E. coli and E. meningoseptica sp. F2 cells in the control group, respectively, were located in the Q4 quadrant (Fig. 7A, B), indicating that debris or lysed cells were attributable to bacterial autolysis during this period. Moreover, we observed 59.2% double-stained cells (cF-positive, PI-positive; Q2 quadrant) in Fig. 7C, which indicated that they exhibited higher esterase activity and membrane permeability in the co-culture of E. meningoseptica and E. coli. When compared with other groups (Fig. 7A, B, D), a minor proportion was distinguished in the Q1 quadrant (2.73%) (cells presumably without metabolic activity) and Q4 quadrant (22.7%) (most likely corresponded to debris or lysed cells attributable to bacterial autolysis). Therefore, we speculate that the construction of a microbial consortium of E. meningoseptica and E. coli improves membrane permeability and may be beneficial for the exchange and metabolism of substances between bacteria. Such microbial consortia consist of several organisms that, together, are more stable to environmental challenges, display reduced metabolic burden due to a division of labour and exchange of resources, possess expanded metabolic capabilities relative to monocultures, and can communicate (chemically or physically) between strains [56,57,58]. However, different phenomena were observed with initial concentrations of 1% POE, with most cells of CO3 located in the Q4 quadrant (Fig. 7D). The results revealed that most of the cell membranes of the CO3 co-cultivation system cells were damaged by exposure to 1% POE. According to a previous study, surfactants have been added to the culture to improve production; low concentrations of surfactants have been reported to increase membrane permeability; therefore, extracellular productivity was promoted by restricting intracellular diffusion of target metabolites [59]. Combined with the electron microscopy results in Fig. 6, we found that several cells were autolysed and lysed. Previous studies have found that a single E. meningoseptica sp. F2 can tolerate a concentration of 1% POE and not die. We speculate that, in this mixed bacterial system, the death of E. coli may be accelerated by POE.

Conclusion

To the best of our knowledge, this is the first report on the use of a bacterial co-culture system for the synthesis of vitamin K2. The present study showed that co-culture of E. meningoseptica and E. coli could be used to improve MK-n (n = 4, 5, 6, 7, 8) production by introducing vitamin K2 synthesis pathway genes from E. meningoseptica sp. F2 in E. coli. To investigate the molecular mechanism of MK-n production in E. meningoseptica, an enzymatic assay was performed, which revealed that EmOPPS could catalyse consecutive condensation reactions of FPP with three molecules of IPP to generate C30 HexPP in vivo and in vitro, which constitutes the side chain of MK-6. In addition, MK-n (n = 4, 5, 6, 7, 8) production was initially achieved in engineered E. coli by overexpressing E. meningoseptica sp. F2-derived EmOPPS and EmGGPPS. Under the optimised inoculation ratio and overexpression of synthetic pathway genes, MK-n (n = 4, 5, 6, 7, 8) production significantly increased to 32.62 ± 1.1 mg/L, which was approximately three-fold higher than that in the E. meningoseptica sp. F2 monoculture. Further expansion of this co-culture, morphological analysis, and FCM illustrated the physiological state of bacteria under co-culture conditions. The twin-bacillus consortium of E. meningoseptica and E. coli that we developed might serve as a promising platform for improved vitamin K2 biosynthesis.

Materials and methods

Bacterial strains, media, and cultivation

E. coli DH5α and E. coli BL21 (DE3) were used as host strains for plasmid construction and protein expression, respectively. Luria–Bertani (LB) medium (10 g/L tryptone, 5 g/L yeast extract, and 10 g/L NaCl) was used to grow E. coli cells unless otherwise specified. The strain E. meningoseptica sp. F2 is commercially used for industrial vitamin K2 production (China Centre for Type Culture Collection, CCTCC, AB2011070). The fermentation medium for the cultivation of the monoculture or co-culture systems was prepared as previously described [47]. The E. meningoseptica sp. F2 strain was inoculated onto beef extract peptone agar slants (3 g/L beef extract, 10 g/L peptone, 5 g/L NaCl, and 20 g/L agar; pH 7) at 37 °C for 24 h [12 h for E. coli BL21 (DE3)] and then stored at 4 °C. Subsequently, the strain on the agar slants was inoculated into fresh seed medium (20 g/L glycerol, 33 g/L peptone, 3 g/L K2HPO4·7H2O, 4.5 g/L NaCl, 0.3 g/L MgSO4·7H2O, and 1.5 g/L yeast extract; pH 7.2) in 250 mL Erlenmeyer flasks with a working volume of 50 mL, incubated at 22 °C, and 300 rpm for 24 h [12 h for E. coli BL21 (DE3)]. Then, the pre-inoculum was transferred to a 500 mL shake flask containing 100 mL of fermentation medium (20 g/L glycerol, 33 g/L peptone, 3 g/L K2HPO4·7H2O, 4.5 g/L NaCl, 0.3 g/L MgSO4·7H2O, 1.5 g/L yeast extract, pH 7.2), incubated at 37 °C, and 250 rpm for 9 days. After 9 days of fermentation, biomass accumulation was measured and estimated as DCW. In brief, 25 mL of the fermentation broth was collected and centrifuged at 15,000 rpm for 15 min, and the supernatant was discarded. The wet cells were washed with sterile distilled water and freeze-dried to a constant weight.

Genetic manipulations

All strains and plasmids used are listed in Table 1. The primer sequences are provided in Additional file 1: Table S2. The gene encoding EmOPPS derived from E. meningoseptica sp. F2 was amplified by polymerase chain reaction (PCR) using genomic DNA as the template. The PCR conditions were as follows: initial denaturation at 94 °C for 5 min; denaturation at 94 °C for 30 s, annealing at 54 °C for 30 s, and extension at 72 °C for 2 min for a total of 30 cycles; and the final extension at 72 °C for 10 min. Pfu PCR MasterMix (TIANGEN, Beijing, China) was used for DNA amplification. The MBP-EmGGPPS from V29-MBP-EmGGPPS was amplified and cloned into the NcoI and BamHI sites of the pETDuet-1 vector to form PETD-1. The amplified fragments GB1-EmOPPS were also cloned into the NdeI and KpnI sites of a PETD-1 vector to form PETD-2 (Table 1). To construct PETA-1, GB1-EmOPPS was cloned into the NcoI and BamHI sites of pET28a. The four genes (EmHMGR, EmHMGS, EmPVD, and EmIDI) derived from E. meningoseptica sp. F2 and SceMK and ScePMK derived from S. cerevisiae were constructed into two plasmids PETA-2 and PACD-2. The amplified fragments EmHMGR and EmHMGS were also cloned into the NcoI and BamHI sites of a pACYCDuet-1 vector to form PACD-1. Then, SceMK and EmIDI were cloned into the NdeI and KpnI sites of PACD-1 to form PACD-2. The amplified fragments EmPVD and ScePMK were also cloned into the NcoI and BamHI sites of the pET28a vector to form the vector PETA-2. EcMenA genes derived from E. coli K12 and EmMenE derived from E. meningoseptica sp. F2 were constructed on the XhoI and SalI sites of plasmid PETA-2 to form PETA-3. The recombinant plasmids were constructed using Gibson assembly [60]. NEBuilder HiFi DNA Master Assembly Mix was purchased from New England Bio Labs (NEB, Beijing, China). Restriction enzyme digestion, transformation, and other standard molecular biology techniques were performed as described by Sambrook et al. [61]. Protein expression was induced by adding 50 μM IPTG at OD600 = 0.6, and the cells were incubated for 4 h at 30 °C. The plasmid mini-prep kit and DNA gel extraction kit were purchased from Sangon Biotech (Shanghai, China).

Table 1 Strains and plasmids that were used in this study
Full size table

The protein purification procedure has been previously described [62]. Briefly, the cells were disrupted and purified using Ni–NTA column chromatography. The bacterial cells (containing pET-28a-GB1-EmOPPS) were sonicated and centrifuged at 12,000×g for 15 min. The supernatant was then loaded onto an Ni–NTA-Sefinose™ column (Sangon Biotech). After the column was washed with a wash buffer (20 mM Tris–HCl, 500 mM NaCl, pH 8.0) and different concentrations of imidazole, the recombinant protein was eluted with an elution buffer (250 mM imidazole, 20 mM Tris–HCl, 500 mM NaCl, pH 8.0). The purified EmOPPS was analysed using mass spectrometry by Sangon Biotech.

EmOPPS enzymatic assay and product identification

In vitro enzyme assays were performed as previously described [63, 64]. Briefly, a 400 μL reaction system containing 400 μM IPP and 200 μM FPP (100 mM HEPES, 5 mM MgCl2, 10 mM KCl, pH 7.5) and 0.5 μmol/L of the purified recombinant fusion protein were incubated at 30 °C for 2 h. After inoculation, the solution was mixed with 200 μL of 0.2 mol/L Tris–HCl (pH 9.5) containing bovine intestinal alkaline phosphatase (20 mg/mL, more than 10 DEA units/mg, Sigma-Aldrich) and two units of shrimp alkaline phosphatase (1 unit/μL; TaKaRa). The reaction mixture was then incubated overnight at 30 °C to hydrolyse the diphosphate products into their corresponding alcohols. The mixture was extracted with hexane, and then three parallel samples of hydrolysate were concentrated to 100 µL in the experiment. Finally, the sample products were analysed using an LC–MS system equipped with an LTQ Orbitrap XL ETD analyser (Thermo Fisher Scientific, USA). The samples were separated using an Agilent 1200 Series HPLC system at a flow rate of 250 μL/min, and the mobile phase consisted of methanol (100%). The electrospray potential was 4.5 kV in positive electrospray ionisation (ESI) mode, and the source temperature was 275 °C. Three independent cultures were analysed for each set of experiments.

Extraction and analysis of vitamin K2 and mevalonate

The extraction of vitamin K2 and measurement of its concentration were adapted from Wei et al. [9]. Briefly, vitamin K2 analysis was performed using high-performance liquid chromatography (HPLC, Shimadzu, Kyoto, Japan) with an eclipse plus a C18 column (Shimadzu, 250 mm × 4.6 mm ID); aliquots of 20 μL were injected manually using a loop injection valve (Shimadzu). The column temperature was maintained at 35 °C. The mobile phase consisted of methanol and dichloromethane (4:1, v/v) at a flow rate of 1 mL/min. The UV–Visible detector was operated at 248 nm for the menaquinones. The menaquinone series of compounds was determined by their absorption and mass spectra using an LC–MS system equipped with a 6200 series TOF/6500 series analyser (Agilent Technologies, Santa Clara, CA, USA).

Mevalonate was determined using an Agilent System 7820A GC equipped with a flame ionisation detector and an HP-5 column (30 m × 0.25 mm, 0.25 μm film thickness) [65, 66]. Briefly, 3 mL of fermentation broth was centrifuged at 14,000×g for 10 min at room temperature. The supernatant was adjusted to pH 2.0–3.0 with HCl and incubated at 45 °C for 2 h to convert mevalonate to mevalonolactone. This solution was then extracted with ethyl acetate and saturated with anhydrous Na2SO4. The ethyl acetate phase containing mevalonolactone was analysed. The column temperature profile was 70 °C for 1 min, a ramp of 20 °C/min to 150 °C with a 5 min hold, and a ramp of 30 °C/min to 300 °C with a 3 min hold. The inlet temperature was 150 °C, with a split ratio of 20:1. The retention time of mevalonolactone was confirmed using the commercial standard DL-mevalonolactone.

Morphological analysis of microbial cells by SEM and confocal laser scanning microscopy

SEM: After 3 days of fermentation, the cells were centrifuged (6000 rpm, 5 min), washed twice with phosphate-buffered saline (PBS, pH 7.4), placed on coverslips, and dried naturally. The cells were then immersed in 2.5% glutaraldehyde for 2–4 h at 4 °C, washed three times with phosphate buffer, and subjected to gradient ethanol dehydration (30%, 50%, 70%, and 90% ethanol once and 100% ethanol twice, each for 15 min). After dehydration, the cells were metal-sprayed and observed under a scanning electron microscope (Schottky Field Emission Scanning Electron Microscope, GeminiSEM 500, USA).

Laser confocal scanning microscopy: The co-culture cells (CO4 system: co-culture of E. meningoseptica sp. F2 and E. coli H03) were diluted ten-fold with PBS (pH 7.4) and placed on a glass slide. The fluorescent images were observed using a confocal laser scanning microscope (LSM 980, ZEISS, Germany) with a 488 nm bandpass filter.

Staining procedure and flow cytometric measurement

Double staining with cFDA (Solarbio, Shanghai, China) and PI (Sigma-Aldrich, St Louis, MO, USA) was used for FC analysis [47, 67, 68]. Briefly, non-treated stained cells and cells treated at 85 °C for 15 min and subsequently stained with cFDA or PI, respectively, were used as controls to define negative and positive histogram regions. Moreover, a non-stained and non-treated control was used to determine the autofluorescence of cells. E. coli BL21, E. meningoseptica sp. F2, CO3, and CO3 + POE cells were initially incubated with 50 μM FDA at 37 °C for 15 min to allow intracellular enzymatic conversion of cFDA to cF. After labelling, the cells were centrifuged at 12,000×g for 5 min at 10 °C and resuspended in 1 mL PBS buffer (pH 7.0). Then, 30 μM PI was added, and the mixture was placed in an ice bath and incubated for 10 min in the dark to allow labelling of membrane-compromised cells. Following the incubation with PI, the samples were placed on ice in the dark until analysis (maximum of 1 h).

Analysis was performed using a Gallios flow cytometer (Beckman Coulter Inc., Miami, FL, USA). The forward scatter (FS), side scatters (SS), and green (FL1) and red (FL3) fluorescence of each cell were measured, amplified, and converted into digital signals for further analysis. The FL1 of the cells stained with cFDA was collected in the FL1 channel (525 ± 15 nm), whereas the FL3 of cells labelled with PI was collected in the FL2 channel (620 ± 15 nm). The flow rates and cell concentrations of the samples were adjusted to maintain the acquisition at 200 microbial cells per second. A total of 50,000 events were recorded. The trials were replicated at least three times with three samples for each short-wave UV dose. Flow cytometry data were analysed using FlowJo version vX.0.7 (TreeStar Inc., Ashland, OR, USA).

Bioinformatic and data analyses

Multiple sequence alignments of E. coli K-12 MG1655 (P0AD57_Ecoli), E. meningoseptica sp. F2 (A0A1V3U058_ELIME), H. influenzae RdKW 20 (P44916_HAEIN), and C35-HexppS from B. subtilis subsp. natto BEST195 (D4FY42_BACNB) were performed using Clustal X and then analysed using ESPript 3.0. The sequences of four prenyltransferases obtained from the National Center for Biotechnology Information database: P0AD57_Ecoli (NP_417654), A0A1V3U058_ELIME (AQX05701), P44916_HAEIN (AAC22540), and D4FY42_BACNB (BAI85773). Three independent cultures were analysed for each set of experiments. All data are presented as mean ± standard error and were compared with those the respective control groups. Statistical significance was assessed by one-way ANOVA via GraphPad Prism 9 software (GraphPad Software, Inc., USA), and differences with p < 0.05 were designated as significant in all the cases.