Establishing microbial co‐cultures for 3‐hydroxybenzoic acid biosynthesis on glycerol

Abstract

Converting renewable feedstocks to aromatic compounds using engineered microbes offers a robust approach for sustainable, environment‐friendly, and cost‐effective production of these value‐added products without the reliance on petroleum. In this study, rationally designed E. coli–E. coli co‐culture systems were established for converting glycerol to 3‐hydroxybenzoic acid (3HB). Specifically, the 3HB pathway was modularized and accommodated by two metabolically engineered E. coli strains. The co‐culture biosynthesis was optimized by using different cultivation temperatures, varying the inoculum ratio between the co‐culture strains, recruitment of a key pathway intermediate transporter, strengthening the critical pathway enzyme expression, and adjusting the timing for inducing pathway gene expression. Compared with the E. coli mono‐culture, the optimized co‐culture showed 5.3‐fold improvement for 3HB biosynthesis. This study demonstrated the applicability of modular co‐culture engineering for addressing the challenges of aromatic compound biosynthesis.

Abbreviations

3HB

3‐hydroxybenzoic acid

CHR

chorismic acid

DHS

3‐dehydroshikimic acid

IPTG

isopropyl β‐D‐1‐thiogalactopyranoside

1. INTRODUCTION

Production of aromatic compounds at present is mainly based on utilization of petroleum resources, which often brings about economic, environmental, and sustainability concerns. Microbial biosynthesis has been developed as a viable method for converting renewable feedstocks to a wealth of aromatic products, offering an alternative approach for effective production and thus reducing the reliance on the use of petroleum. To date, significant success has been achieved for microbial biosynthesis of various aromatic compounds 1, 2, 3, 4. On the other hand, novel engineering approaches, such as modular co‐culture engineering, have recently been developed for improving microbial biosynthesis performance 5. Modular co‐culture engineering employs microbial co‐cultures composed of multiple microbial strains of the same or different species to accommodate the modularized biosynthesis pathways of the target products. This strategy allows for the pathway modularization in the context of a microbial co‐culture, as each co‐culture strain is only assigned with one module of the entire pathway. As a result, it facilitates the reduction of the metabolic burden on the recruited microbial strains, which has been considered to a major challenge metabolic engineering 6. In addition, balancing of the biosynthesis strengths between individual pathway modules harbored by the corresponding co‐culture strains can be achieved by straightforward manipulation of the strain inoculum ratio, offering an effective way for addressing the need of pathway balancing 5, 7. To date, the modular co‐culture engineering approach has been utilized for production of a variety of biochemicals ranging from simple biofuel molecules to complex natural products 5, 8, 9.

In the present study, we designed and constructed E. coli–E. coli co‐cultures for producing 3‐hydroxybenzoic acid (3HB) from renewable and abundant carbon substrate glycerol 10, 11. 3HB is structurally similar to 2‐hydroxybenzoic acid, the key component of aspirin. It has been found to possess stress response desensitizing, anti‐radical, and antioxidant activities 12, 13, 14, 15. It also has potential beneficial effect for maintaining normal blood lipid level 16. In addition, 3HB can be used for synthesis of industrially important chemicals such as polymer, herbicide, and plasticizer 17, 18, 19, 20. 3HB biosynthesis using the mono‐cultures of engineered E. coli and C. glutamicum has been reported, both of which utilized glucose as the carbon substrate 21, 22. For this study, we chose to use glycerol as the pathway substrate for 3HB bioproduction for a few reasons. First, glycerol is an important byproduct of the biodiesel industry and there has been increasing interest in using glycerol for microbial biosynthesis applications. To this end, 3HB bioproduction on glycerol has not been reported before. Moreover, there are relatively few studies using glycerol as the carbon substrate for co‐culture engineering, which leaves extensive unexplored areas for further investigation.

The 3HB biosynthetic pathway is illustrated in Figure 1A. Carbon substrate, such as glycerol used in this study, is first converted to 3‐dehydroshikimic acid (DHS) via the central metabolism and the shikimate pathway 23. DHS is then utilized to produce chorismic acid (CHR), which is subsequently converted to 3HB by 3‐hydroxybenzoate synthase 24. For constructing the co‐cultures in this study, two E. coli strains were individually engineered to accommodate the upstream (Glycerol to DHS) and downstream (DHS to 3HB) pathway modules, respectively. The co‐cultivation of the two strains in one consolidated culture presented a complete biosynthetic system for production of 3HB from Glycerol. The biosynthesis performance by the co‐cultures was investigated and compared with the mono‐culture control under different conditions to demonstrate the effectiveness of the modular co‐culture engineering for 3HB biosynthesis.

Figure 1.

(A) The 3HB biosynthetic pathway and the E. coli–E. coli co‐culture design for Glycerol‐to‐3HB conversion. The ratio between the co‐culture strains determined the relative biosynthetic strengths of the upstream and downstream pathway modules. DHS: 3‐dehydroshikimic acid, SHK: shikimic acid, S3P: shikimate 3‐phosphate, EPSP: 5‐enolpyruvoyl‐shikimate 3‐phosphate, CHR: chorismic acid, 3HB: 3‐hydroxybenzoic acid. (B) Effect of the cultivation temperature on 3HB biosynthesis using the engineered mono‐culture PMH1 and co‐culture PUN1:BDZ1. Specific 3HB production is defined by 3HB concentration over the culture cell density

2. MATERIALS AND METHODS

2.1. Chemicals and medium

The chemicals and culture medium used in this study were purchased from Sigma‐Aldrich (St. Louis, MO, USA) or Thermo Fisher Scientific (Waltham, MA, USA). The DNA cloning agents, including Q5 High‐Fidelity DNA polymerase, restriction enzymes, T4 ligase, were obtained from New England Biolabs (Ipswich, MA, USA). PCR primers were obtained from Sigma‐Aldrich (St. Louis, MO, USA). LB medium was used for all DNA cloning manipulations. The 3HB biosynthesis was conducted using a glycerol medium. One liter glycerol medium contained 5 g glycerol, 2 g of NH₄Cl, 5 g (NH4)₂SO₄, 3 g KH₂PO₄, 7.3 g K₂HPO₄, 8.4 g 3‐(N‐morpholino) propanesulfonic acid (MOPS), 0.5 g NaCl, 0.24 g MgSO₄, 0.5 g yeast extract, 40 mg of tyrosine, 40 mg of tryptophan, 100 mg phenylalanine, and trace elements (0.4 mg Na₂EDTA, 0.03 mg H₃BO₃, 1 mg thiamine, 0.94 mg ZnCl₂, 0.5 mg CoCl₂, 0.38 mg CuCl₂, 1.6 mg MnCl₂, 3.77 mg CaCl₂, and 3.6 mg FeCl₂). When needed, 50 mg/L kanamycin, 34 mg/L chloramphenicol, and 0.1 mM isopropyl β‐D‐1‐thiogalactopyranoside (IPTG) were supplemented into the medium.

PRACTICAL APPLICATION

The establishment of the E. coli–E. coli co‐cultures in this study offers an innovative and effective approach for converting renewable glycerol to value‐added compound 3‐hydroxybenzoic acid. After several rounds of engineering efforts, the optimized co‐culture system biosynthesized 294 mg/L 3HB from 5 g/L glycerol. The mass yield of 0.059 g/g was achieved in the test tube scale, showing strong potentials for further bioproduction improvement using the bioprocess engineering approaches. In addition, the constructed co‐cultures can be recruited for utilization of other renewable feedstock molecules such as glucose and xylose, as E. coli has native capabilities for consuming these molecules as carbon substrates. More broadly, the co‐culture engineering methodology developed in this work can be adapted for biosynthesis of other aromatic compounds derived from the same or similar pathways.

2.2. Plasmid and strain construction

The plasmids and strains used in this study are listed in Table 1. All plasmids were constructed using E. coli NEB5‐alpha high‐efficiency competent cell (New England Biolabs, Ipswich, MA, USA). Commercial kits (Zymo research, Irvine, CA, USA) were used for DNA purification, gel recovery and miniprep. The hyg5 gene encoding 3‐hydroxybenzoate synthase was codon‐optimized and synthesized by Bio Basic Inc (Amherst, NY, USA). To construct plasmid pYZ1, the hyg5 gene was PCR amplified using primers pHyg5‐up (CGGTCGACTTAAGGAGGAACAGACATGTTGAACCCGTCTAGC) and pHyg5‐down (ATGCGGCCGCCTAGTTTACATAACAACACCTTCA), then cloned into a previously constructed plasmid pBA5 (pUC57‐Kan carrying E. coli aroE, aroL, aroA, aroC, and ubiC genes, unpublished data) to replace the ubiC gene using SalI/NotI restriction sites. Similarly, the PCR amplified hyg5 gene was inserted into plasmid pET21c using SalI/NotI restriction sites to generate plasmid pYZ2.

Table 1.

Plasmids and strains used in this study

Plasmids	Description	Source
pET28a	T7 promoter, KanR	Novagen
pACYCDuet‐1	two T7 promoters, CmR	Novagen
pET21c	T7 promoter, AmpR	Novagen
pYZ1	pUC57(Kan) carrying E. coli aroE, aroL, aroA, aroC and codon optimized S. hygroscopicus hyg5 genes under the control of the constitutive Zymomonas mobilis pyruvate decarboxylase promoter (Ppdc)	This study
pYZ2	pET21c carrying codon optimized S. hygroscopicus hyg5 gene	This study
pCS5	pACYCDuet‐1 carrying the E. coli aroE, aroL, aroA and aroC genes, as well as the shiA gene that is under the control of a constitutive lacuv5 promoter	25

Strains	Description	Source
BL21(DE3)	F− ompT hsdSB (rB −, mB −) gal dcm (DE3)	Invitrogen
P5	E. coli K12 ΔpheA ΔtyrR lacZ::PLtetO‐1‐tyrA fbr aroG fbr tyrR::PLtetO‐1‐tyrA fbr aroG fbr hisH(L82R) ΔaroE ΔydiB	25
BH2	E. coli BL21(DE3) ΔxylA ΔtyrA ΔpheA	25
PUN1	P5 harboring pET28a	This study
PUN2	P5 harboring pET28a and pACYCDuet‐1	This study
PUN3	P5 harboring pET21c and pACYCDuet‐1	This study
PMH1	P5 harboring pYZ1	This study
PMH2	P5 harboring pYZ1 and pCS5	This study
BDZ1	BH2 harboring pYZ1	This study
BDZ2	BH2 harboring pYZ1 and pCS5	This study
BDZ4	BH2 harboring pYZ2 and pCS5	This study

2.3. 3HB biosynthesis by engineered mono‐cultures and co‐cultures

For monoculture biosynthesis, E. coli PMH1 and PMH2 were initially grown overnight in the LB medium at 275 rpm under 37°C. The cell culture was then centrifuged at 3500 rpm for 5 min, and the cell pellets were re‐suspended in 5 mL glycerol medium to reach an initial OD of 0.5. The inoculated culture was grown at 250 rpm under 30°C for 72 h. When needed, 0.1 mM IPTG was added at the beginning of the cultivation.

For E. coli–E. coli co‐culture biosynthesis, E. coli PUN1, PUN2, PUN3, BDZ1, BDZ2, and BDZ4 were cultivated separately in the LB medium. The overnight cultures were centrifuged and re‐suspended in the fresh glycerol medium in desired combinations (PUN1 and BDZ1, PUN2 and BDZ2, PUN3 and BDZ4). The total initial OD₆₀₀ of the co‐cultures was set to be 0.5, although the inoculum ratio between the upstream and downstream strain was varied. Except the PUN1:BDZ1 co‐culture, 0.1 mM IPTG was added at the beginning of the cultivation. For the IPTG induction time point experiment, 0.1 mM IPTG was added to the PUN3:BDZ4 co‐culture at 0, 2, 4, 6 and 8 h after the co‐culture inoculation.

2.4. Analytical method for 3HB quantification

To quantify 3HB, 1 mL culture samples were centrifuged at 13 000 rpm for 5 min. The supernatant was then filtered through 0.45 μm PVDF syringe filter (Whatman Inc., NJ, USA). The filtered samples were analyzed by Agilent HPLC equipped with a Waters C18 column HyperSelect ODS 5 μm 150 × 4.6 mm (Milford, MA, United States) and a UV detector set to a wavelength of 293 nm. The following water‐acetonitrile gradient elution program was used at a flow rate of 0.5 mL/min; 0 min, 25% acetonitrile; 4 min, 50% acetonitrile; 6 min, 99% acetonitrile; 8 min, 25% acetonitrile; 10 min, 25% acetonitrile.

3. RESULTS AND DISCUSSION

In this study, we used glycerol, an important byproduct of the biodiesel industry and renewable raw material, as the carbon substrate for biosynthesis of valuable compound 3HB. In order to construct the first co‐culture system, two E. coli strains were utilized for harboring the upstream and downstream 3HB pathway modules, respectively, as shown in Figure 1A. The upstream strain PUN1 was derived from a previously constructed strain for over‐production of metabolite 3‐dehydroshikimic acid (DHS) 25, 26. Notably, the genes aroE and ydiB involved in the competing pathway for DHS were deleted to improve the DHS availability. The downstream strain was engineered for establishing the DHS‐to‐3HB conversion. Specifically, plasmid pYZ1 was constructed to over‐express E. coli’s aroE, aroL, aroA, aroC genes for enhancing the formation of CHR from DHS. In addition, this plasmid contained the codon‐optimized hyg5 gene encoding S. hygroscopicus 3‐hydroxybenzoate synthase for converting CHR to 3HB 22, 24. Strain BDZ1 containing plasmid pYZ1 was used as the downstream stream for the co‐culture.

The PUN1:BDZ1 co‐culture was then grown on 5 g/L glycerol for the 3HB production. Two strains were inoculated at a ratio of 1:1. Strain PMH1 over‐expressing the selected shikimate pathway genes and 3HB synthase gene hyg5 was employed as the mono‐culture control. The mono‐culture and co‐culture were cultivated at three different temperatures, including 25, 30, and 37°C, for 72 h before subjecting to HPLC analysis. As shown in Figure 1B, the change of the cultivation temperature caused significant biosynthesis performance variation. For all temperatures, the engineered co‐culture showed higher production than the mono‐culture. For both mono‐culture and co‐culture, the optimal 3HB production was achieved at 30°C. At this temperature, the co‐culture produced 74 mg/L 3HB and the mono‐culture produced 55 mg/L 3HB. It was therefore indicated that 30°C was the best temperature for meeting the need of cell growth and heterologous 3HB biosynthesis, and this temperature was thus employed for all the following experiments. To our knowledge, this is the first report utilizing glycerol as the carbon substrate for biosynthesis of value‐added 3HB using the engineered E. coli mono‐culture and co‐culture.

Next, the 3HB biosynthesis by the PUN1:BDZ1 co‐culture was optimized by changing the inoculum ratio between the upstream and downstream strains, which helped to balance the biosynthetic strengths of the corresponding pathway modules carried by these strains. As shown in Figure 2A, when the downstream BDZ1 strain was inoculated at low ratios (19:1, 9:1 4:1), the DHS‐to‐3HB conversion and 3HB production were relatively low. With the increase of the BDZ1 inoculum, the 3HB concentration gradually increased. The highest production of 99 mg/L was achieved at the PUN1:BDZ1 ratio of 1:19. These results indicated that the DHS conversion in the downstream strain was the limiting step for the overall 3HB biosynthesis. The specific 3HB production, defined by 3HB concentration over the culture cell density, showed similar trend with the inoculum ratio change. Notably, the 3HB concentration and specific production of the mono‐culture control PMH1 were merely 55 mg/L and 17.2 mg/L/OD, respectively. The 3HB biosynthesis improvement by the co‐culture could be largely due to (1) reduction of metabolic burden on each strain improved their fitness and biosynthetic capabilities and (2) flexible increase of the downstream strain's sub‐population strengthened the DHS‐to‐3HB conversion. Although the DHS‐to‐3HB conversion can also be enhanced in the context of mono‐culture, it requires laborious optimization of the gene expression promoter, ribosomal binding site, plasmid copy number, etc. in a trial‐and‐error manner. In contrast, the co‐culture engineering approach offers a new perspective to circumvent these technical barriers for pathway balancing, and thus holds considerable potentials for wider application in the broad field of microbial biosynthesis.

Figure 2.

Optimization of the 3HB biosynthesis by employing engineered microbial co‐cultures of (A) PUN1:BDZ1, (B) PUN2:BDZ2, (C) PUN3:BDZ4. E. coli PMH1 and PMH2 were used as the mono‐culture controls. (D) The effect of IPTG induction time point on 3HB biosynthesis using the PUN3:BDZ4 co‐culture inoculated at 1:1 ratio

To further improve 3HB biosynthesis, a new co‐culture PUN2:BDZ2 was constructed in which the E. coli membrane‐bound transporter ShiA of the pathway intermediate DHS was employed to strengthen the DHS assimilation by the downstream strain. In fact, pathway intermediate mass transfer between the co‐culture strains is a common challenge for modular co‐culture engineering. The recruited ShiA transporter has been found to facilitate the importation of extracellular DHS into the E. coli cell and has been used for microbial biosynthesis of different compounds 25, 26, 27. For this study, the gene shiA encoding the transporter was over‐expressed using plasmid pCS5 in the downstream strain. pCS5 also carried additional copies of aroE, aroL, aroA, aorC genes for attracting stronger metabolic flux towards CHR formation. Notably, due to the modular nature of the co‐culture system, such metabolic engineering modification in the downstream strain did not interfere with the upstream strain's biosynthesis capability, highlighting the advantage of the modular co‐culture engineering. As shown in Figure 2B, the 3HB concentration and specific production of the PUN2:BDZ2 co‐culture were increased dramatically when the inoculum ratio was changed from 19:1 to 1:4, suggesting that the enhancement of the downstream pathway module for DHS‐to‐3HB conversion was critical for the overall bioproduction performance. No significant production improvement was observed when inoculum ratio was further changed to 1:9 and 1:19. 137 mg/L 3HB was produced at the inoculum ratio of 1:9, which was 38% higher than the optimal production by the PUN1:BDZ1 co‐culture. Interestingly, although the ShiA over‐expression was also expected to take up the secreted DHS for biosynthesis enhancement for mono‐culture, the mono‐culture control strain PMH2 engineered using the same strategy did not show any biosynthesis improvement over PMH1 (55 vs 52 mg/L). Therefore, it was demonstrated that the same engineering strategy could result in different outcomes in mono‐cultures and co‐cultures.

Next, we further modified the co‐culture design by adapting a dedicated plasmid for expressing 3‐hydroxybenzoate synthase Hyg5 using the strong T7 promoter, rather than the mild constitutive promoter Ppdc. Also, other downstream pathway genes (aroE, aroL, aroA, aorC) were solely expressed by plasmid pCS5.The correspondingly constructed strains PUN3 and BDZ4 showed slightly compromised DHS production and improved DHS uptake capabilities, respectively (Supporting Information Figure S1). Also, these strains’ glycerol uptake rates did not show significant difference from the co‐culture strains constructed early on (Supporting Information Figure S2). The PUN3:BDZ4 co‐culture was then grown in the glycerol medium for 3HB biosynthesis. As shown in Figure 2C, the 3HB concentration and specific production both varied depending on the inoculum ratio between PUN3 and BDZ4. Importantly, the 3HB concentrations at all tested inoculum ratios were increased to different levels, compared with the production by the PUN2:BDZ2 co‐culture. The biosynthesis improvement hereby indicated that improving the expression strength of the key pathway enzyme Hyg5 played a critical role in strengthening the overall biosynthesis performance. In the PUN3:BDZ4 co‐culture, this strategy resulted in 229 mg/L 3HB production after the inoculum ratio optimization, which was significantly higher than the previous mono‐cultures and co‐cultures. Notably, compared with the previous co‐cultures, the optimal inoculum ratio of PUN3:BDZ4 was shifted to 1:1. This finding showed that the modification of the pathway genes expression strategy altered the relative biosynthetic capabilities of the corresponding co‐culture strains and thus changed the optimal strain‐to‐strain ratio for 3HB bioproduction. Compared with the inoculum ratio, the endpoint strain‐to‐strain ratios showed dramatic changes at the end of the co‐culture cultivation (Supporting Information Table S1). Interestingly, the observed endpoint ratios converged to the range between 1.2:1 and 0.2:1, which is much smaller than the inoculum ratio range. This finding suggests that the co‐culture strains’ growth competitions under different starting situations eventually stabilized at a similar level.

We then investigated the influence of IPTG induction timing on 3HB production using the PUN3:BDZ4 co‐culture inoculated at the ratio of 1:1. In fact, it was found that the use of IPTG to induce strong gene expression could generate a negative impact on cell growth (Supporting Information Figure S3). On the other hand, although adding the inducer, e.g. IPTG, at the beginning of cultivation is a convenient and straightforward way for initiating pathway gene expression, it may not be necessarily the best strategy for maximizing the bioproduction. As shown in Figure 2D, early induction at the beginning of the co‐culture cultivation was found to be a sub‐optimal strategy for 3HB biosynthesis. As the induction timing was delayed to 2 and 4 h, the 3HB concentration gradually increased. Further delay of the induction timing to 6 and 8 h resulted in reduction of the bioproduction performance. The highest production of 294 mg/L achieved at 4 h induction was 5.3‐fold higher than the original mono‐culture strain PMH1 constructed for 3HB biosynthesis. The specific 3HB production also peaked at induction time point of 4 h. In fact, early induction of pathway genes’ overexpression often results in overwhelming metabolic burden on the host cells at early growth phases and can lead to reduced growth profile and sub‐optimal bioproduction performance. On the other hand, overly late induction cannot give the host cells sufficient time to dedicate metabolic resources on target production biosynthesis. As such, varying time point of induction is an effective way for balancing the need of cell growth and biosynthesis. Overall, the achievement of our study showed significant 3HB biosynthesis improvement in a step‐wise manner and demonstrated the general applicability of modular co‐culture engineering for de novo biosynthesis of various biochemicals.

CONFLICT OF INTEREST

The authors have declared no conflict of interest.

Supporting information

Supporting Information

Zhou Y, Li Z, Wang X, Zhang H. Establishing microbial co‐cultures for 3‐hydroxybenzoic acid biosynthesis on glycerol. Eng Life Sci 2019;19:389–395. 10.1002/elsc.201800195