Metabolic Engineering of Carotenoid Biosynthesis in Escherichia coli by Ordered Gene Assembly in Bacillus subtilis

Bacterial strains and plasmids. B. subtilis strain RM125 (24) was used for the OGAB assembly. BUSY9166 [RM125 proB::(cI857 spc)], which was constructed by insertion of a cI857-spc cassette from pCISP310B (9) into the NotI site of the proB gene, was also used for the OGAB assembly of a p_R promoter-equipped vector to repress expression from the p_R promoter. E. coli Top10 (Invitrogen) [F′{lacI^q Tn10 (Tet^r)} mcrAΔ(mrr-hsd RMS-mcrBC) φ80lacZΔM15 ΔlacX74 deoR recA1 araD139Δ(ara-leu)7697 galU galK rpsL(Str^r) endA1 nupG], JM109 (Takara) [recA1 endA1 gyrA96 thi hsdR17(r_K⁻ m_K⁺) e14⁻ (mcrA) supE44 relA1 Δ(lac-proAB)/F′ traD36 proAB⁺lacI^qlacZ DM15], and GI724 (Invitrogen) (13, 16) [F⁻λ⁻lacI^qlacPL8 ampC::P_trcI mcrA mcrB INV(rnnD-rnnE)] were used as host strains for cloning of assembled genes into the pCR-XL-TOPO vector (Invitrogen), pCrtP series, and pPrP series (see below), respectively. Plasmids pCAR25 and pACCAR25ΔcrtX (20) were kindly provided by N. Misawa (Marine Biotechnology Institute, Japan).

OGAB.The protocol of the OGAB method was previously described (22). The oligonucleotide sequences used in this study are listed in Table 1.

The putative 240-bp promoter region (P_crtE) upstream of crtE from Pantoea ananatis (Erwinia uredovora) (DDBJ accession number D90087 ) was amplified by PCR from pCAR25 with an Ex Taq DNA polymerase (Takara), using primers CrtPr-Acc65I-F1 and CrtPr-DraIII-R2. Similarly, the 128-bp p_R promoter region of the λ phage was amplified by PCR from pBEST515C (8), using primers PrP-Acc65I-F1 and PrP-DraIII-R2. The PCR products were subcloned into pCR-XL-TOPO (Invitrogen) and confirmed by DNA sequencing with an ABI PRISM 3100 genetic analyzer (Applied Biosystems). Each plasmid was digested with Acc65I, and the resultant small fragment was inserted into the identical site of pGETS109SfiI (22), resulting in pGETS109SfiI-Pcrt and pGETS109SfiI-PrP, respectively (Fig. 2B).

• Open in new tab
• Download powerpoint

FIG. 2.

(A) Schematic representation of the assembly of five genes for carotenoid biosynthesis. (B) DraIII fragments of shuttle vectors pGETS109SfiI-Pcrt and pGETS109SfiI-PrP used in this study. (C) Five circularly permuted DNA constructs assembled in this study.

Three types of five Pantoea ananatis genes, crtE (fragments E1 to E3, 932 bp), crtB (B1 to B3, 954 bp), crtI (I1 to I3, 1,502 bp), crtY (Y1 to Y3, 1,173 bp), and crtZ (Z1 to Z3, 549 bp), with DraIII sites at both ends, were amplified by PCR from pCAR25, using primers (Table 1) CrtE-DraIII-F1 and CrtE-DraIII-R1 (E1), CrtE-DraIII-F1 and CrtE-DraIII-R2 (E2), CrtE-DraIII-F2 and CrtE-DraIII-R1 (E3), CrtB-DraIII-F1 and CrtB-DraIII-R1 (B1), CrtB-DraIII-F1 and CrtB-DraIII-R2 (B2), CrtB-DraIII-F2 and CrtB-DraIII-R1 (B3), CrtI-DraIII-F1 and CrtI-DraIII-R1 (I1), CrtI-DraIII-F1 and CrtI-DraIII-R2 (I2), CrtI-DraIII-F2 and CrtI-DraIII-R1 (I3), CrtY-DraIII-F1 and CrtY-DraIII-R1 (Y1), CrtY-DraIII-F1 and CrtY-DraIII-R2 (Y2), CrtY-DraIII-F2 and CrtY-DraIII-R1 (Y3), CrtZ-DraIII-F1 and CrtZ-DraIII-R1 (Z1), CrtZ-DraIII-F1 and CrtZ-DraIII-R2 (Z2), and CrtZ-DraIII-F2 and CrtZ-DraIII-R1 (Z3), respectively. These 15 fragments were also subcloned into pCR-XL-TOPO and confirmed by DNA sequencing.

Two vector plasmids, pGETS109SfiI-Pcrt and pGETS109SfiI-PrP, and the 15 genes described above were digested with DraIII and gel purified. Each vector and an appropriate combination of five insert genes (Fig. 2C) were mixed (1.25 fmol/μl each) in a ligation buffer (66 mM Tris-HCl [pH 7.6], 6.6 mM MgCl₂, 10 mM dithiothreitol, 100 μM ATP, 150 mM NaCl, 10% polyethylene glycol 6000 [Wako Pure Chemical]) and incubated with 4 U of T4 DNA ligase (Toyobo) at 37°C for 30 min. Each ligation solution was transformed into competent B. subtilis cells and incubated overnight on LB plates containing tetracycline at 37°C (for RM125) or 30°C (for BUSY9166). The resultant plasmids, the pCrtP series (pCrtP-EBIYZ, pCrtP-ZEBIY, and pCrtP-BIYZE) and the pPrP series (pPrP-EBIYZ, pPrP-ZEBIY, pPrP-YZEBI, pPrP-IYZEB, and pPrP-BIYZE), were confirmed by restriction mapping.

Carotenoid biosynthesis in E. coli.The pCrtP series and pPrP series plasmids were retransformed into E. coli JM109 and GI724 cells, respectively. The JM109 transformants were grown in 50 ml of LB medium containing 100 μg/ml ampicillin at 37°C for 6, 12, or 24 h, and the cells were harvested by centrifugation. The GI724 transformants were grown in 50 ml of M9 minimal medium (21) containing 0.5% glucose, 0.2% Casamino Acids, 1 mM MgCl₂, and 10 μg/ml tetracycline at 30°C. When the culture attained an optical density of 0.5 at 550 nm, l-tryptophan was added to a final concentration of 100 μg/ml to cancel the repression of the p_R promoter by the cI repressor. After an additional 4, 8, or 12 h incubation at 37°C, the cells were harvested by centrifugation.

Carotenoids were extracted from the cells with acetone as previously described (5) and analyzed by high-performance liquid chromatography on a COSMOSIL 5C₁₈-MS-II column (4.6 by 250 mm; Nacalai Tesque) with acetonitrile-methanol-isopropanol (90:6:4) at a flow rate of 1 ml/min. The column was calibrated with commercially available zeaxanthin, β-cryptoxanthin, β-carotene (DHI Water and Environment), and lycopene (Sigma).

Real-time RT-PCR.Real-time reverse transcription (RT)-PCR was performed with a QuantiTect SYBR green RT-PCR kit (Roche) and gene-specific primer sets (the amplicon size was 113 to 137 bp; see below) on a LightCycler (Roche). The targeting region of the RT-PCR was positioned about two-thirds downstream from the initiation codon of each gene. Template mRNA was purified from E. coli cells by using an RNeasy Protect Bacteria mini kit (QIAGEN). The standard template mRNA was prepared as follows. Each gene was amplified by PCR with Ex Taq DNA polymerase (Takara), using specific primers (Table 1) RT-PCRcrtE-F and RT-PCRcrtE-R for crtE (137 bp), RT-PCRcrtB-F and RT-PCRcrtB-R for crtB (113 bp), RT-PCRcrtI-F and RT-PCRcrtI-R for crtI (129 bp), RT-PCRcrtY-F and RT-PCRcrtY-R for crtY (114 bp), and RT-PCRcrtZ-F and RT-PCRcrtZ-R for crtZ (115 bp). These PCR products were cloned into pDrive (QIAGEN), resulting in pDrive-E, pDrive-B, pDrive-I, pDrive-Y, and pDrive-Z. These plasmids were transcribed by the RiboMAX large-scale RNA production system SP6 (Promega), purified with an RNeasy mini kit (QIAGEN), and served as the standard mRNA.

Assembly of five genes for carotenoid biosynthesis by the OGAB method.In this study, we designed new zeaxanthin operons consisting of five carotenoid biosynthetic genes, crtE, crtY, crtI, crtB, and crtZ, assembled as one transcriptional unit in a polycistronic manner by the OGAB method. Constructed plasmids were transferred into E. coli to assay the production of carotenoids (Fig. 2A). To this end, all genes and an assembly vector were prepared as DNA fragments with arbitrary 3′ 3-nucleotide protrusions that allowed the genes to be linked together in a specific order and orientation by the OGAB method. The protrusions were generated by type II restriction endonucleases that recognize two separate recognition sites, such as DraIII, which recognizes 5′-CACNNN/GTG-3′, where N is any nucleotide in arbitrary sequence; DraIII was chosen in this study since there is no DraIII site in the relevant genes and vectors.

As the first choice of a promoter of these operons, a natural carotenoid promoter in the upstream region of crtE (P_crtE) of Pantoea ananatis was inserted into the B. subtilis-E. coli shuttle vector pGETS109SfiI (22), together with two DraIII sites for assembly, resulting in pGETS109SfiI-Pcrt (Fig. 2B, top). Prior to assembly, pGETS109SfiI-Pcrt was converted to a fragment with 3′ 3-nucleotide protrusions at both ends by DraIII digestion.

Each gene used in this study was defined as a DNA fragment with one open reading frame from a ribosome-binding site to the termination codon of the protein-coding region. Five carotenoid genes were separately amplified by PCR, using primers with DraIII recognition sites at the 5′-end portions, and cloned into E. coli plasmid vector. After sequence confirmation, the relevant gene fragments were excised from these plasmids with DraIII to generate 3-nucleotide protrusions at the 3′ ends. For the five circularly permuted carotenoid operons, at least three gene fragments that have different protrusions are required for every five genes for defined assembly. Thus, we prepared in total 15 gene fragments, crtE (E1 to E3), crtB (B1 to B3), crtI (I1 to I3), crtY (Y1 to Y3), and crtZ (Z1 to Z3), as shown in Fig. 2C.

Taking the construction of pCrtP-EBIYZ as a representative example, we adjusted the concentrations of six of the DraIII fragments, crtE (E3), crtB (B1), crtI (I1), crtY (Y1), crtZ (Z2), and pGETS109SfiI-Pcrt, to be equimolar (Fig. 2A). After an equal volume of each was mixed, the mixture was ligated in a tandemly repeated form by T4 DNA ligase and transformed into B. subtilis RM125 cells, resulting in pCrtP-EBIYZ. As the colonies of B. subtilis transformant with the plasmid did not show any color, due to carotenoids, we concluded that production of carotenoids in B. subtilis cells is difficult and did not perform further investigation of B. subtilis. The plasmid could be used for carotenoid biosynthesis in the noncarotenogenic bacterium E. coli.

Similarly, for the construction of the other four circularly permuted plasmids, appropriate genes were selected, mixed, ligated, and transformed into B. subtilis RM125. However, only two out of four constructs (Fig. 2C), pCrtP-ZEBIY and pCrtP-BIYZE, were obtained by the OGAB method. The efficiency of the emergence of colonies with correct assemblies (the ratio of the number of positive B. subtilis colonies to the number of tetracycline-resistant colonies) was 58% for pCrtP-EBIYZ and pCrtP-ZEBIY and 17% for pCrtP-BIYZE. Despite two trials of OGAB assembly, the remaining two plasmids, pCrtP-YZEBI and pCrtP-IYZEB, could not be obtained in B. subtilis.

Carotenoid biosynthesis in E. coli JM109 with the pCrtP series plasmids.The three pCrtP plasmids were transformed into E. coli JM109, which was then cultivated for 6, 12, or 24 h. The zeaxanthin production in the E. coli transformants with each of three pCrtP plasmids was confirmed by high-performance liquid chromatography, using commercially available zeaxanthin as a standard. As shown in Fig. 3, the highest zeaxanthin content of 820 ± 39 μg/g (dry weight) was obtained in the transformant with pCrtP-EBIYZ, whose order of the five genes (P_crtE-crtE-crtB-crtI-crtY-crtZ) corresponds to that for the zeaxanthin biosynthetic pathway (Fig. 1B). With pCrtP-BIYZE and pCrtP-ZEBIY, the amounts of zeaxanthin were 257 ± 16 μg/g (dry weight) and 172 ± 32 μg/g (dry weight), respectively. We also measured the yields of β-cryptoxanthin and β-carotene: the amounts of these intermediates were below the detection limit. Synthesis of zeaxanthin with pGETS109SfiI-derived pCrtP-EBIYZ was 4.4-fold higher than that of zeaxanthin with the original pACCA25ΔcrtX (187 ± 23 μg/g [dry weight]), whose gene order is native (P_crtE-crtE-crtY-crtI-crtB and convergent directional P_crtZ-crtZ [19]). Since differences between plasmids are known to affect carotenoid synthesis (20), we examined the effect of a pGETS109SfiI vector by inserting an NheI-HindIII fragment (6.5 kbp) carrying the carotenoid biosynthesis genes. Synthesis of zeaxanthin with the resultant pGETS109CAR25ΔcrtX plasmid (609 ± 14 μg/g [dry weight]) was 3.3-fold higher than that of zeaxanthin with pACCAR25ΔcrtX (Fig. 3).

• Open in new tab
• Download powerpoint

FIG. 3.

Time course of the production of zeaxanthin (μg/g [dry weight {dw}]) in E. coli JM109 transformants with pCrtP series plasmids.

Improvement of the assembly efficiency of the OGAB method.A possible reason why pCrtP-YZEBI and pCrtP- IYZEB could not be obtained by the OGAB method is the toxicity of the gene products of these plasmids to B. subtilis. In order to rigorously repress the gene expression of these plasmids in B. subtilis, we used a p_R/cI repressor system (4). By replacement of the promoter and host cells with a p_R promoter from the λ phage (Fig. 2B, bottom) and B. subtilis BUSY9166 cells expressing the cI repressor, respectively, all five circularly permuted plasmids (Fig. 2C) were obtained by the OGAB method; these were named the pPrP series. The efficiencies of the emergence of colonies with correct assemblies were 100% for pPrP-EBIYZ, 75% for pPrP-ZEBIY, 42% for pPrP-YZEBI, 33% for pPrP-IYZEB, and 92% for pPrP-BIYZE. These values are superior to those for the pCrtP series.

Carotenoid biosynthesis in E. coli GI724 with the pPrP series plasmids.Five pPrP plasmids were separately introduced into E. coli GI724, in which the cI repressor is expressed and might repress the p_R promoter to allow the expression of the relevant operons. In GI724, only a limited amount of carotenoids (<20 μg/g [dry weight]) was produced with pPrP series plasmids (Fig. 4, 0 h). When tryptophan was added to the culture in order to cancel the cI repression on the p_R promoter, all the transformants produced carotenoids (Fig. 4). Analogously with the case of the pCrtP series, the highest carotenoid content was obtained in the transformant with pPrP-EBIYZ, which is the order in the zeaxanthin biosynthetic pathway (Fig. 1B), though not only zeaxanthin but also the precursors β-cryptoxanthin and β-carotene were included in the products. With the other four plasmids, little or no β-cryptoxanthin or β-carotene was observed. We also measured the concentrations of lycopene for all plasmids, but the values were below the detection limit. In most cases, the zeaxanthin content gradually increased, reaching a maximum after 12 h, whereas that of the transformant with pPrP-ZEBIY rapidly reached a maximum after 4 h (Fig. 4).

• Open in new tab
• Download powerpoint

FIG. 4.

Time course of the production of β-carotene, β-cryptoxanthin, and zeaxanthin in E. coli GI724 transformants with pPrP series plasmids. dw, dry weight.

Expression profiles of the carotenoid genes in E. coli.By use of real-time RT-PCR, the mRNA copy numbers of the five carotenoid genes in E. coli GI724 cells transformed with pPrP plasmids were determined. Before induction of gene expression, the copy numbers were low, but the expression level of the fifth gene positioned at the 3′ end of each operon seemed relatively high compared with those of genes at other positions (Fig. 5, 0 h). This was probably a result of measuring antisense mRNA transcribed from a promoter (P1) in the downstream flanking pBR322 sequence (6), since both mRNA strands can be amplified by the one-step RT-PCR used here. The mRNA expression levels of almost all carotenoid genes were increased at 4 to 8 h after their induction (Fig. 5), correlating with the production of total carotenoids (Fig. 4). Interestingly, the rank order of the gene expression levels in each plasmid corresponded to the sequential order of the five genes in the plasmid, irrespective of the gene content; for all transformants, the most abundant gene among the five genes was located at a downstream flanking position relative to the promoter, and the copy number of the transcript decreased monotonically with increasing separation from the promoter. For example, the five mRNA species in GI724/pPrP-EBIYZ at 8 h were present in the following order of decreasing abundance: crtE (100%), crtB (72%), crtI (22%), crtY (10%), and crtZ (8%), where the percentage is calculated with respect to the mRNA of crtE. The average profiles of mRNA distribution (means ± standard deviations) for the five operons at 8 h decreased monotonically as follows: 100% (first position from the promoter/first position from the promoter), 55% ± 20% (second/first), 27% ± 28% (third/first), 10% ± 6% (fourth/first), and 7% ± 4% (fifth/first). Thus, the mRNA abundance decreases by roughly half from one gene to the next. However, the average reduction rates (means ± standard deviations) between specific pairs of genes at 8 h were as follows: 109% ± 23% (crtB-crtE), 44% ± 29% (crtI-crtB), 55% ± 15% (crtY-crtI), 54% ± 22% (crtZ-crtY), and 33% ± 15% (crtE-crtZ).

• Open in new tab
• Download powerpoint

FIG. 5.

RNA copy numbers of the carotenoid genes (orange, crtE; yellow, crtB; dark green, crtI; light green, crtY; blue, crtZ) in 1 ng of total RNA extracted from the E. coli GI724 transformants with assembled plasmids pPrP-EBIYZ (A), pPrP-ZEBIY (B), pPrP-YZEBI (C), pPrP-IYZEB (D), and pPrP-BIYZE (E) for incubation times of 0, 4, 8, and 12 h after induction. The quantitative real-time PCR data were obtained from at least two independent experiments, and the errors were in the range of 10 to 30%.