Lambda Red recombinase-mediated integration of the high molecular weight DNA into the Escherichia coli chromosome

Background Escherichia coli K-12 is a frequently used host for a number of synthetic biology and biotechnology applications and chassis for the development of the minimal cell factories. Novel approaches for integrating high molecular weight DNA into the E. coli chromosome would therefore greatly facilitate engineering efforts in this bacterium. Results We developed a reliable and flexible lambda Red recombinase-based system, which utilizes overlapping DNA fragments for integration of the high molecular weight DNA into the E. coli chromosome. Our chromosomal integration strategy can be used to integrate high molecular weight DNA of variable length into any non-essential locus in the E. coli chromosome. Using this approach we integrated 15 kb DNA encoding sucrose catabolism and lactose metabolism and transport operons into the fliK locus of the flagellar region 3b in the E. coli K12 MG1655 chromosome. Furthermore, with this system we integrated 50 kb of Bacillus subtilis 168 DNA into two target sites in the E. coli K12 MG1655 chromosome. The chromosomal integrations into the fliK locus occurred with high efficiency, inhibited motility, and did not have a negative effect on the growth of E. coli. Conclusions In addition to the rational design of synthetic biology devices, our high molecular weight DNA chromosomal integration system will facilitate metabolic and genome-scale engineering of E. coli. Electronic supplementary material The online version of this article (doi:10.1186/s12934-016-0571-y) contains supplementary material, which is available to authorized users.


Background
Escherichia coli K-12 is a gram-negative model bacterium used as a host for a number of synthetic biology and biotechnology applications [1][2][3]. Furthermore, E. coli is considered to be among the most promising chassis for the development of the minimal cell factories [4][5][6][7]. Novel approaches for introducing synthetic DNA, particularly the high molecular weight DNA, into E. coli would therefore greatly facilitate engineering efforts in this bacterium.
The chromosomal integration approaches differ in their flexibility, speed and the size of the DNA fragments they can integrate into the chromosome. While there had been studies describing high molecular weight DNA integration into the chromosomes of some bacteria, such as Bacillus subtilis [17][18][19], this was not reported in E. coli yet. The currently used lambda Red recombination methods relying on a selection with a single antibiotic gene flanked by FLP recombinase target (FRT) sites are time-consuming due to the necessity to flip-out the antibiotic resistance gene in order to use it again. Furthermore, the traditional lambda Red recombineering employing antibiotic cassette flanked by FRT sites leaves a single FRT site in the chromosome after flipping-out of the antibiotics gene [20]. The undesired recombination between the two FRT sites (first on the introduced DNA fragment and second in the chromosome) makes the repeated use of the cassettes flanked with FRT sites unsuitable for integration of large DNA sequences into the same target locus. Plasmid pSB1K3(FRTK), clonetegration, KIKO-vectors, and yeast homing mitochondrial I-SceI endonuclease-based methods were shown to successfully integrate up to 10 kb DNA fragments into the E. coli chromosome [12][13][14][15].
Here, we developed a reliable and flexible lambda Red recombinase-based approach that utilizes overlapping DNA fragments for integration of the high molecular weight DNA of virtually any length into any non-essential target site in the E. coli chromosome. We demonstrate the utility of this method by integrating 15 kb DNA encoding sucrose and lactose catabolism pathways and 50 kb of B. subtilis 168 DNA into the E. coli K12 MG1655 chromosome.

Results and discussion
fliK of the flagellar region 3b as the chromosomal integration target site The potential integration target sites in the E. coli chromosome vary in their suitability for integration of synthetic DNA [16]. Genes involved in the biogenesis and regulation of the flagellum are considered to be particularly suitable target sites for integration of genetic circuits due to their high expression, good characterization and non-essentiality for E. coli. Flagellar genes are situated in the genomic regions frequently occupied by DNA-binding proteins (highly expressed extended protein occupancy domains, heEPODs) [21,22]. Furthermore, unlike the traditionally used phage attachment (att) sites, flagellar genes are present and well-conserved in all commonly used industrial E. coli strains, such as MG1655, DH10B, BL21-DE3, and W3110 [4,21,22,25].
Here, we chose to integrate high molecular weight DNA into the fliK locus of the E. coli K12 MG1655 flagellar region 3b. We selected the fliK locus as the integration target site because our previous analyses of the E. coli K12 MG1655 flagellar regions 1, 2, 3a, and 3b revealed that the flagellar region 3b supports the highest integration efficiency of all flagellar regions [16,23,24]. Furthermore, fliK was shown to support the highest integration efficiency from the E. coli K12 MG1655 flagellar region 3b.
Some commonly used industrial E. coli strains, such as BL21-DE3, are often causing problems by genomic engineering. This is usually due to the missing att sites used in many chromosomal integration approaches. Unlike att sites, the fliK locus in the strain BL21-DE3 has 100 percent homology with the strain K12 MG1655. To further verify the suitability of the fliK locus for integration of synthetic DNA, we compared the integration efficiency in strains BL21-DE3 and K12 MG1655. The efficiency of integration of the kanamycin resistance cassette (approximately 2 kb) into the fliK locus of the strain BL21-DE3 was high and comparable to that observed in the strain K12 MG1655 (Fig. 1a). This confirms the suitability of the fliK locus for integration of the synthetic DNA across variable E. coli strains. The exact position within the fliK open reading frame where the integrations occured is shown in Fig. 1b.

Strategy of the high molecular weight DNA integration into the E. coli chromosome
Our strategy for the integration of the high molecular weight DNA molecules into the E. coli chromosome is outlined in Fig. 2. This approach utilizes the Red recombinase system of the bacteriophage lambda for integration of DNA fragments into the E. coli chromosome via recombination between homologous sequences [16]. Phage-based systems are a useful tool for genome engineering as they increase the otherwise very low rate of homologous recombination in some bacteria, such as E. coli [13,15,20,25]. The lambda Red recombinase system employed in our approach is composed of three proteins Bet, Exo and Gam. Bet and Exo bind to the ends of the introduced DNA fragment and facilitate homologous recombination and integration into the chromosome. This is possible due to Gam-facilitated inhibition of the host RecBCD exonuclease V which, when active, degrades foreign DNA [20]. Lambda Red recombinase system in our approach is introduced into the host E. coli cells on plasmid pKM208. The expression of the lambda Red recombinase system from the lacZ promoter is tightly controlled by the presence or absence of Isopropyl β-d-1-thiogalactopyranoside (IPTG) in the growth medium [26].
DNA fragments to be integrated into the E. coli chromosome are first cloned into the plasmids p14kan and pJScav next to the kanamycin cassette (kan) and chloramphenicol-yellow fluorescent protein cassette (cat-yfp), respectively (Fig. 2a). Cloning of DNA fragments into p14kan and pJScav is facilitated by Gibson isothermal assembly approach [27,28]. Next, DNA fragments are PCR amplified together with the attached kan or cat-yfp cassettes and sequences homologous to the target region in the chromosome. Sequences homologous to the target region in the E. coli chromosome (at least 60 bp in length) are introduced into the PCR amplified DNA fragments on primers. The length of the homology regions can be further increased by designing primers which generate longer overlapping regions between the DNA fragments for increased integration efficiency.
DNA fragments tagged with kan or cat-yfp cassettes are transformed into the electro-competent E. coli cells harbouring plasmid pKM208 with lambda Red recombination system. The pKM208-encoded lambda Red recombination system facilitates integration of DNA fragments into the E. coli chromosome (Fig. 2a). If the first DNA fragment is tagged with kan cassette, successful recombinants are selected by growing on kanamycin (Fig. 2b). The integration of the following DNA fragment tagged with cat-yfp cassette results in replacement of the kan resistance marker with cat-yfp cassette and successful recombinants are easily detectable by growth on chloramphenicol and yellow fluorescent protein (mVenus) expression (Fig. 2b). Elongation of the integrated DNA sequence is facilitated by the alternative use of kan and cat-yfp cassettes (Fig. 2a).

Integration of the sucrose and lactose catabolism pathways into the E. coli chromosome
To demonstrate the utility of our system we first integrated sucrose and lactose catabolism pathways into the E. coli strain K12 MG1655 (EcΔlac) chromosome. E. coli K12 MG1655 can not normally utilize sucrose. We obtained DNA encoding the sucrose catabolism pathway from E. coli W, which is the only non-pathogenic E. coli strain that can utilize sucrose as a carbon source [15,29].
Four DNA fragments (each approximatelly 3 kb in length) were first cloned into plasmids p14kan and pJScav for tagging with kan or cat-yfp cassettes (Fig. 2a). The first DNA fragment tagged with kan cassette was PCR amplified using primers harbouring sequences homologous to the fliK locus in the EcΔlac chromosome (Additional file 1) and electroporated into the host E. coli with pKM208-encoded lambda Red recombination system. The successful transformants were selected on plates with kanamycin (Fig. 2b). The second DNA fragment tagged with cat-yfp cassette was PCR amplified using forward primer with sequence homologous to the fliK locus and reverse primer with homology to the sequence of the first DNA fragment already integrated into the chromosome (Additional file 1). Electroporation of the second fragment into the host E. coli with pKM208-encoded lambda Red recombination system led to its integration into the chromosome and replacement of the kan resistance marker with cat-yfp cassette. Successful recombinants grew on plates with chloramphenicol and were and sequencing. PCR amplifications and sequencing across junctions showed that the DNA fragments were integrated into the E. coli chromosome in the correct order ( Fig. 3a; Additional file 1). Furthermore, the chromosomal integration of the yellow fluorescent protein (mVenus), chloramphenicol resistance, sucrose and lactose catabolism pathways was confirmed phenotypically (Fig. 3b-e). The engineered strain EcΔlac(I1234) grew on chloramphenicol (Fig. 3b), emitted yellow fluorescent light (Fig. 3c) and fermented lactose (Fig. 3d) and sucrose (Fig. 3e). DNA fragments to be integrated into the E. coli chromosome are first tagged with the selectable markers by cloning into plasmids p14kan and pJScav. p14kan and pJScav encode the kanamycin cassette (kan) and chloramphenicolyellow fluorescent protein cassette (cat-yfp), respectively. DNA fragments attached to kan or cat-yfp cassettes are transformed into the electrocompetent E. coli cells harbouring plasmid pKM208. pKM208-encoded lambda Red recombination system integrates DNA fragments into the chromosome by homologous recombination. Successful transformants from the chromosomal integration of the first DNA fragment attached to kan cassette are selected on medium with kanamycin. Second fragment attached to cat-yfp cassette integrates into the chromosome next to the first fragment, thus replacing the kan resistance marker. Alternative use of kan and cat-yfp cassettes allows elongation of the integrated DNA sequence. b Growth and antibiotic selection of E. coli harbouring integrated high molecular weight DNA tagged with kan or cat-yfp cassette. Strains with kan and cat-yfp cassettes grow only on kanamycin and chloramphenicol plates, respectively. Furthermore, cells with cat-yfp cassette produce fluorescent light

Integration of the 50 kb B. subtilis DNA into the E. coli chromosome
We obtained high molecular weight DNA for cloning into the E. coli K12 MG1655 chromosome from B. subtilis 168. We chose B. subtilis due to high DNA sequence variability between B. subtilis and E. coli to avoid undesired recombination between the host and the donor DNA. Furthermore, the B. subtilis 168 DNA region chosen does not harbour any known operons whose expression could be detrimental to the host E. coli (Additional file 3).
Seven DNA fragments (each approximatelly 6-7 kb in length) were first cloned into p14kan and pJScav (Fig. 2a). The first PCR amplified DNA fragment tagged with kan cassette and harbouring sequences homologous to the fliK locus (Additional file 1) was electroporated into host E. coli with pKM208-encoded lambda shown on the right. b LB agar with chloramphenicol was used to examine chloramphenicol resistance in EcΔlac (left) and EcΔlac(I1234) (right). Only the engineered strain EcΔlac(I1234) grew on chloramphenicol. c EcΔlac (left) and EcΔlac(I1234) (right) grown on LB plates were investigated for the expression of the integrated yellow fluorescent protein (mVenus) using blue light. EcΔlac(I1234) emitted yellow fluorescent light. d MacConkey agar was used to analyze the ability to ferment lactose in EcΔlac (left) and EcΔlac(I1234) (right). The red color of EcΔlac(I1234) is caused by a pH change resulting from lactose fermentation. e M9 medium supplemented with sucrose (20 mg/ml) was used to investigate the ability to utilize sucrose as a carbon source in EcΔlac (left) and EcΔlac(I1234) (right). EcΔlac(I1234) grew on the minimal sucrose medium Red recombination system. The successful transformants were selected on plates with kanamycin (Fig. 2b). The integration of the next DNA fragment tagged with catyfp cassette was detected by yellow fluorescent protein (mVenus) expression and growth on chloramphenicol (Fig. 2b). The alternative use of kan and cat-yfp cassettes led to the integration of all seven DNA fragments (with total length of approximately 50 kb) into the E. coli K12MG1655 chromosome. The integrations into the E. coli K12 MG1655 chromosome were verified by diagnostic PCR with flanking primers (Fig. 4a; Additional file 1) and sequencing. PCR amplifications and sequencing across junctions showed that the DNA fragments were integrated into the E. coli K12 MG1655 chromosome in the correct order (Fig. 4a).
To assess the suitability of the fliK locus for integration of the high molecular weight DNA, we compared the integration efficiency of the fliK locus with another randomly selected integration target site in the E. coli chromosome, cheW (Fig. 4b). Integration efficiencies were determined from the number of colony forming units per microgram of DNA. This analysis showed that the fliK locus supports integration of the high molecular weight B. subtilis DNA with higher efficiency than the cheW locus (Fig. 4b).
Genes encoding flagellar functions are important for the cell's motility. To verify that integration of the high molecular weight DNA into the fliK locus of E. coli K12 MG1655 impacts host cell's motility, we analyzed the engineered strain Ec(I1234567Bs) on the motility agar plate (Fig. 4c). The motility the strain Ec(I1234567Bs) was completely eliminated (Fig. 4c). Integrations of the high molecular weight DNA into the E. coli chromosome should not have a negative effect on the host cell's growth. To test the effect of chromosomal integration of the high molecular weight DNA on the growth rate, we used the engineered strains Ec(I1Bs), Ec(I12Bs), Ec(I123Bs), Ec(I1234Bs), Ec(I12235Bs), Ec(I123456Bs), and Ec(I1234567Bs) harbouring integrated DNA of increasing length (up to 50 kb). The cell growth was quantified by measuring absorbance of the bacterial cultures with the microplate reader (Fluostar Omega). The absorbance of the E. coli K12 MG1655 wild type was compared with that of the engineered strains Ec(I1Bs), Ec(I12Bs), Ec(I123Bs), Ec(I1234Bs), Ec(I12235Bs), Ec(I123456Bs), and Ec(I1234567Bs) grown at 37 °C for 24 h (Fig. 4d; Additional file 4). Integration of B. subtilis DNA fragments of various lengths into the fliK locus of the E. coli chromosome produced viable transformants whose growth was not impaired when compared with the wild type E. coli K12 MG1655. The strains Ec(I1Bs), Ec(I12Bs), Ec(I123Bs), Ec(I1234Bs), Ec(I12235Bs), Ec(I123456Bs), and Ec(I1234567Bs) harbouring integrated DNA grew slightly better than the wild type E. coli K12 MG1655 ( Fig. 4d; Additional file 4). Better growth of the strains with integrations into the fliK locus could be due to the lower metabolic burden to the cell resulting from the loss of the flagellar function. The fliK knockout strain grew only slightly better than the wild type E. coli K12 MG1655 (Additional file 4). This corroborates our previous findings, where similar growth pattern was observed for other mutants in the flagellum-encoding genes [16,23,24]. Good growth is a desirable feature for metabolic engineering and synthetic biology applications to generate more efficient and productive microbial cell factories.
The comparison of the integration efficiencies with another chromosomal integration target site, unimpaired growth and eliminated motility of the host cells show that the fliK locus of E. coli K12 MG1655 is suitable for chromosomal integration of the high molecular weight DNA.

Conclusions
We set out to provide the synthetic biology community with the reliable system for the integration of the high molecular weight DNA into the E. coli chromosome. To this end we developed a lambda Red recombinase-based method which utilizes overlapping DNA sequences for chromosomal integration of the high molecular weight DNA. Our chromosomal integration strategy is more flexible and allows integration of larger DNA molecules into the E. coli chromosome than the previous methods. Furthermore, the alternative use of kan and cat-yfp cassettes allows faster chromosomal integration process than the standard lambda Red recombineering methods, which rely on curing-out the lambda Red recombinaseencoding vector from the cell, before introducing second, FLP recombinase-encoding plasmid [20]. The length of the DNA fragments integrated into the E. coli chromosome by previous lambda Red recombinase-based approaches was about 10 kb [12][13][14][15][16]. The chromosomal integration strategy presented here can be used to integrate high molecular weight DNA of virtually any length into the E. coli chromosome. Furthermore, it does not target only a few integration sites, but allows DNA integration into any non-essential locus. Using this approach we integrated 15 kb DNA encoding sucrose and lactose catabolism pathways into the fliK locus of the flagellar region 3b and 50 kb B. subtilis DNA into two target sites in the E. coli K12 MG1655 chromosome. We showed that integration of 50 kb of the foreign DNA into the fliK locus inhibits motility of E. coli, but does not have a negative effect on the growth. Our chromosomal integration system will facilitate rational design of synthetic biology devices in E. coli. By allowing integration of the high molecular weight DNA, including whole operons and metabolic pathways into the E. coli chromosome, it can be used for metabolic and genome-scale engineering of this bacterium. Furthermore, it can be easily adapted for use in other bacteria with the functional lambda Red recombination system.

Bacterial strains and growth conditions
Bacterial strains and plasmids used in this study are listed in Table 1. Luria-Bertani medium (LB) was used for the routine growth of bacteria. Bacteria cultivated in the liquid LB were grown on a rotatory shaker at 200 rpm and 37, 30 or 42 °C, depending on the requirements of the experiments. Bacteria cultivated on LB agar (10 mg/ ml) plates were grown at 37, 30 or 42 °C for 24 h. When necessary, LB medium was supplemented with kanamycin (50 µg/ml), ampicillin (100 µg/ml) or chloramphenicol (30 µg/ml) to grow E. coli and kanamycin (5 µg/ml) or chloramphenicol (5 µg/ml) to cultivate B. subtilis. Lactose fermentation experiments were performed using MacConkey agar (Sigma). The ability to utilize sucrose was investigated using M9 medium supplemented with thiamine (1 µg/ml) and sucrose (20 mg/ml) [15].

DNA integration into the E. coli chromosome
Integrations of DNA fragments into the E. coli chromosome were performed employing previously described method [16]. Briefly, DNA fragments were first PCR amplified using primers harboring flanking sequences homologous to the target region in the E. coli K12 MG1655 chromosome (at least 60 bp in length) and purified by gel extraction. As longer sequences greatly improve integration efficiency, the length of the homology regions can be further increased by designing primers generating longer overlapping regions between the DNA fragments. Plasmid pKM208 encoding lambda Red recombination system was introduced into E. coli K12 MG1655 by electroporation and successful transformants were selected on ampicillin at 30 °C. Overnight E. coli K12 MG1655 culture harboring pKM208 was inoculated into a fresh liquid LB medium (1:100 dilution) and grown to OD 600 of 0.2 at 30 °C. Then, 1 mM IPTG was added and bacteria were grown to the final OD 600 of 0.4-0.6. Bacteria were washed twice with glycerol (100 µl/ml) and resuspended in glycerol (100 µl/ ml) (100 µl per 100 ml culture). Bacterial cells were then transformed with the gel purified DNA fragments by electroporation. In our settting, the best transformation efficiency was achieved when using 100-150 µl of the PCR product (approximatelly 5 µg of DNA). This volume was electroporated into 100 µl of the freshly prepared electro-competent cells harbouring pKM208. Transformants harboring the chromosomaly integrated DNA fragments were selected on kanamycin or chloramphenicol plates at 37 °C. Successful chromosomal integrations were verified by diagnostic PCR with flanking primers and sequencing. In general, approximately 50 % of the transformants harbored correctly integrated DNA fragment. The cells were transformed newly with pKM208 after each round of integration to increase chromosomal integration efficiency. After the last round of integration the temperature sensitive plasmid pKM208 can be cured out by growing transformants at 42 °C. Chemically competent bacteria were prepared by the modified Hanahan method [30]. Electro-competent E. coli were prepared by the modified Miller and Nickoloff method [31]. The whole process for integration of a single DNA fragment into the E. coli chromosome using this system takes about 4 days. This is faster than the standard lambda Red recombineering methods which necessitate curing-out the lambda Red recombinase system-encoding vector from the cell, prior to introducing another, FLP recombinase-encoding, plasmid. On the first day, DNA fragment is cloned into pJScav or p14kan by Gibson Assembly and transformed into competent E. coli strain. On the second day, correctly assembled plasmids are identified by colony PCR, correctly assembled colonies are inoculated for plasmid preparation and the recipient E. coli harbouring pKM208-encoded lambda Red recombination system is inoculated into LB with ampicillin. On the third day, plasmids pJScav/p14kan with assembled DNA are isolated and sequenced, DNA fragments tagged with cat-yfp

Motility assessment
The motility of the bacterial strains was evaluated using the motility agar plates. The motility agar plates were prepared by pouring 100 ml of motility agar [tryptone (100 mg/ml), NaCl (50 mg/ml), Bacto-Agar (2.5 mg/ml) (Difco)] in the 13 cm plates and let to set overnight. 2 μl of the overnight bacterial cultures diluted to OD 600 of 1.0 were inoculated in the middle of the motility plates pre-warmed to 37 °C. Plates were incubated for 4-6 h at 37 °C prior to taking the pictures.