Application of the FLP/LoxP-FRT recombination system to switch the eGFP expression in a model prokaryote

Abstract In prokaryotes, few studies have applied the flippase (FLP)/P1-flippase recombination target (LoxP-FRT) recombination system to switch gene expression. This study developed a new method for switching gene expression by constructing an FLP/LoxP-FRT site-specific recombination system in Escherichia coli. To this end, we placed the Nos terminator flanked by a pair of LoxP-FRT in front of enhanced green fluorescent protein (eGFP). The Nos terminator was used to block the expression of the eGFP. When a plasmid expressing FLP was available, deletion of the Nos terminator would allow expression of eGFP. The regulatory effect was demonstrated by eGFP expression. The efficiency of the gene switch was calculated as high as 89.67%. The results showed that the FLP/LoxP-FRT recombinase system could be used as a gene switch to regulate gene expression in prokaryotes. This new method for switching gene expression could simplify the gene function analysis in E. coli and other prokaryotes, as well as eukaryotes.


Introduction
Escherichia coli is a Gram-negative bacterium. Most E. coli strains have a capsular structure and fimbriae. Due to its simple structure and well-understood genetic background, it is an important model organism and one of the most commonly used bacteria for gene transformations currently [1]. Terminator is a DNA sequence that functions to terminate DNA transcription and release RNA. Hence, the terminator plays an important role in regulating gene expression [2,3]. At present, some universal terminators are used in plant genetic transformation vectors and most of them are derived from viruses or other microorganisms. Among them, Nos terminator is one of the most widely used terminators in plant molecular breeding [4].
The flippase (FLP) recombinase system is derived from the 2 µm plasmid of Saccharomyces cerevisiae. Depending on the location and direction of the flippase recombination target (FRT) recognition sites, FLP recombinase can invert, recombine, and position the DNA sequence between the recognition sites [5]. When FRT recognition sites have different orientations and are located on the same chromosome, FLP recombinase inverts the sequences between the recognition sites; when FRT recognition sites have the same orientation and are located on the same chromosome, FLP recombinase removes the entire region between the recognition sites, leaving only one recognition site; and when FRT recognition sites are in the same orientation but are located on different chromosomes, the two chromosomes will exchange DNA around the recognition sites under the action of FLP recombinase [6]. As the FLP recombinase system has been shown to have high recombination efficiency and target in eukaryotes, it is widely used in higher eukaryotes such as rice, Arabidopsis and tobacco for deletion of exogenous and marker genes at present [7][8][9][10][11][12]. Studies have shown that in the FLP recombination system, gene recombination efficiency was higher when using the locus of crossing over in the P1-flippase recombination target (LoxP-FRT) fusion recognition site compared with the FRT single recognition site [13,14]. Introduction of a nuclear localization signal (NLS) can also improve gene recombination efficiency [10].
In addition to eukaryotes, the FLP/FRT recombination system has also been used in prokaryotes. A markerless mutant was generated in cyanobacteria by using FLP recombinase to delete the kanamycin (KAN) resistance gene between two FRT recognition sites [15]. The FLP recombinase was used to eliminate the resistance genes between two FRT sites in E. coli [16]. Apart from this, there have been few studies about the FLP/LoxP-FRT site-specific recombination system operating as a gene switch to regulate target gene expression.
The purpose of this study was to develop a new method for switching gene expression by constructing an FLP/LoxP-FRT site-specific recombination system in E. coli. The regulatory effect was demonstrated by enhanced green fluorescent protein (eGFP) expression. This new method of switching gene expression could simplify the analysis of the gene function in prokaryotes and eukaryotes. The effectiveness of this method provides a reference for the FLP/LoxP-FRT recombinase system as a gene switch and basis for the future gene switch of other recombinase systems.

Bacterial strains and transformations
E. coli strain BL21 was used for this study. The liquid medium used for E. coli growth was Luria Broth (LB; 10.0 g tryptone, 5.0 g yeast extract, and 10.0 g sodium chloride per 1 L water), whereas LB agar was used as the solid medium (LB with 15.0 g agar per 1 L). For selective media (used to isolate transformed bacteria), the final concentrations of KAN, ampicillin (AMP), and isopropyl-β-D-thiopyrgalactoside (IPTG) were 50 µg/mL, 100 µg/mL, and 0.5 mmol/L, respectively.
Vectors pET-21a (+) and pRSFDuet-1 were purchased from GenScript Biotechnology Co. Ltd. The Ep30TL vector was constructed using EcoRI and SacI enzymes (Thermo Fisher Scientific, Tokyo, Japan) to double digest the pET-21a (+) vector, and then the synthetic fragment seq30TL (containing Nos terminator between two LoxP/FRT sites and tagged with eGFP) was ligated to the pET-21a (+) vector by T4 DNA ligase (Promega, Wisconsin, USA) ( Figure 1). The p30TK vector was constructed by double digestion of the pRSFDuet-1 vector with EcoRI and SacI enzymes followed by the ligation of the synthetic fragment seq30TK (containing FLP between NLS and tagged with Nos terminator) to the pRSFDuet-1 vector by T4 DNA ligase ( Figure 2). Ep30TLK competent cells were prepared as described by Sambrook and Russell [17]. The vector p30TK was transformed into Ep30TL competent cells by electroporation and the transformed line (Ep30TLK) was grown at 37℃ with shaking at 180 rpm/min for 1 h. Approximately, 100-200 µL of the resulting culture was spread on LB agar containing KAN + AMP + IPTG.

Vector sequence verification
The following primer sequences were designed for verification to ensure the sequences of the Ep30TL, p30TK, and Ep30TLK vectors were correct. The primers F1 and R1 were used to verify the Loxp-FRT-Nos-Loxp-FRT-eGFP vector Ep30TL sequence, where LoxP-FRT sequences with the same orientation were added at both ends of the Nos terminator, and the eGFP gene was added at the 3′ end. The primers F2 and R2 were used to verify the FLP recombinase expression vector p30TK. The EcoRI restriction site was introduced at the 5′ end of the FLP gene and the SacI restriction site was introduced at the 3′ end. The Ep30TLK vector, the product of p30TK transformation into Ep30TL competent cells by electroporation, was verified by sequencing with F1, R1, F2, and R2 primers. The synthesis and sequencing of the primers were completed by Tsingke Biotechnology Co. Ltd.

Regulatory efficiency calculations
Ep30TLK was cultured on LB medium supplemented with KAN + AMP + IPTG at 37°C. Samples of 50 cultures of 100 µL each were mounted on glass slides. Ep30TL competent cells were negative control without p30TK. Prepared slides were then analyzed with a Confocal 880 laser, confocal scanning microscope and stimulated with a laser at 488 nm to obtain scanning images of E. coli at different points in time post inoculation.
The study has set three replicates to calculate the efficiency of gene switch. The calculations of total bacteria were made from bacterial populations that were more or less similar in number in the negative control and treated bacterial cells. The efficiency of the gene switch was calculated as the number of single green fluorescent colonies divided by the total number of colonies.
DNA extraction was performed according to the instructions of the plasmid DNA microextraction kit (Magen, Guangzhou, China). Gel purification was performed according to the instructions of the gel midi purification kit (Tiangen, Beijing, China). Restriction digests were incubated at 37°C for 30 min according to the manufacturer instructions (Thermo Fisher Scientific, Tokyo, Japan).

Statistical analysis
Preliminary calculations were performed on the data using excel tables, SPSS16.0 statistical analysis software was used to perform variance analysis on the data of deletion efficiency of FLP recombinase, and all the data were analyzed by the least significant difference method at 5% level.

Verification of transformants with Ep30TL and p30TK
As shown in Figure 3, the Ep30TL plasmid was identified by single and double enzyme digestion with restriction endonucleases, EcoRI and SacI. The single digests each resulted in DNA fragments of 6,613 bp while the double digest resulted in two fragments of 5,441 bp and 1,172 bp, respectively; one was the pET-21a (+) plasmid vector and the other was the seq30TL insertion fragment. The electrophoresis results showed that the expected fragments were obtained and the sequencing results were consistent with the enzyme digestion results, further indicating that the cells were successfully transformed with the Ep30TL plasmid.  The p30TK plasmid was verified using the same method. As shown in Figure 4, the EcoRI enzyme single digest resulted in 4,666 bp and 800 bp fragments, while the SacI enzyme single digest resulted in a fragment of 5,466 bp. Double digest with EcoRI and SacI resulted in fragments of 3,819 bp, 847 bp, and 800 bp, consistent with the expectations. The p30TK plasmid isolated from the transformed bacteria was sequenced and found to be 100% homologous to the reference sequence, thus demonstrating that the transformation with p30TK was successful.

Expression of p30TK in Ep30TL competent cells
Through electroporation method, p30TK was transformed into Ep30TL competent cells, and the transformed product (Ep30TLK) and negative control were added to LB medium separately. Bacterial culture was then sampled to observe the fluorescence with a laser confocal microscope. As shown in Figure 5, eGFP did not express in the negative control (Figure 5a), demonstrating that the gene eGFP was locked in Ep30TL. The expression of eGFP was observed in Ep30TLK in five different fields (Figure 5b-f), which proved that p30TK had been transformed into Ep30TL competent cells, and the Nos terminator was deleted by FLP recombinase in Ep30TLK. The FLP/LoxP-FRT recombination system to switch the eGFP expression was successful in E. coli by observing the expression of eGFP.

Regulatory efficiency of gene switch
In the negative control, expression of eGFP was not observed in bacterial cells (Figure 6a). In Ep30TLK, expression of eGFP was observed in single colonies under a fluorescence microscope, with three different types of colonies being identified: those with no eGFP expression ("single white colonies," Figure 6b); those with some eGFP expression ("single green and white colonies," Figure 6c); and those with full eGFP expression ("single green colonies," Figure 6d). Single white colonies resulted from Nos terminator not being deleted, preventing eGFP expression. Similarly, the appearance of single green and white colonies was due to the incomplete deletion of Nos terminator by FLP recombinase, while the occurrence of green fluorescent single colonies was because eGFP could be expressed after the complete deletion of Nos terminator. Thus, it could be concluded that the FLP/LoxP-FRT recombination system could switch the gene eGFP expression in a single colony.
Then, the regulatory efficiency of gene switch in E. coli could be calculated. Statistical results are shown in Table 1; the efficiency of the gene switch in Ep30TLK was calculated in this study as ∼87.10-89.67%.

Verification of Ep30TLK plasmid
The three different single colonies shown in Figure 6 were cultured in liquid medium, then plasmids were extracted from each. Extracted plasmids were named Ep30TLK1, Ep30TLK2, and Ep30TLK3, each corresponding to the white, green and white, and green colonies, respectively. As shown in Figure 7a, EcoRI and SacI double digestion of Ep30TLK1 and Ep30TLK2 results in 1,172 bp fragments; when the same digestion was performed with Ep30TLK3, an 808 bp fragment was seen but not a 1,172 bp fragment, thus indicating that Nos terminator in Ep30TLK3 had been deleted by FLP recombinase. Results of each plasmid from single digests with EcoRI and SacI were also in line with the expectations. Finally, the Ep30TLK1, Ep30TLK2, and Ep30TLK3 plasmids were sequenced to unambiguously verify their identity and were found to be consistent with the reference sequence.
The double digestion model of vector Ep30TLK is shown in Figure 7b. After p30TK was transformed into Ep30TL competent cells, if FLP recombinase fully deleted Nos terminator in Ep30TL, the plasmid Ep30TLK3 could be obtained after identification from eGFP expression. The deletion of the Nos terminator would lead to an 808 bp fragment resulting from double digestion with   EcoRI and SacI. The results shown in Figure 5 were consistent with the double digestion model.

Discussion
In this study, the FLP/LoxP-FRT recombination system was used as a gene switch, controlling the temporal and spatial specificity to precisely regulate target gene expression. The FLP recombination acted as the "gene key" while Nos terminator was the "gene lock" that hindered the expression of eGFP. When the FLP recombinase recognized the LoxP-FRT fusion site and deleted the Nos terminator, the "gene key" opened the "gene lock" and eGFP could be expressed. The efficiency of the gene switch was calculated as high as 89.67%. Exploring the efficiency of the FRT/LoxP-FRT recombination system in E. coli provides a reference for the research of prokaryotes with other site-specific recombination systems. For example, when studying the Cre/ LoxP system from bacteriophage P1, based on the prior construction of a label-free transgenic Chlamydomonas reinhardtii [18], the Cre recombinase can be combined with the LoxP-FRT fusion site and applied to other prokaryotes. Alternatively, the method presented in this study could be used with the recombinase/recombination site (R/RS) system from Zygosaccharomyces to verify the effectiveness and gene switch efficiency of the R/RS recombinant systems quickly and efficiently.
This principle has been shown previously using the Cre/loxP site-specific recombination system as a gene switch in hybrid rice; one cassette was the KEY, containing a nuclear-localized Cre recombinase driven by the green-tissue-specific promoter rbcS, while another cassette was the LOCK, containing a Nos terminator between two loxP sites. When the two cassettes were pyramided into hybrid rice, the Cre recombination from the KEY would excise loxP-NosT in the LOCK, the gene of interest could express in green tissues but not express in the endosperm [19]. For future research in eukaryotes, we can consider changing the "gene key" promoter to regulate gene expression specifically. In the FLP/LoxP-FRT gene switch, the expression of FLP recombinase is designed to be driven by the specific promoter. The design strategy is to first deactivate all gene expressions and then activate expression in only the desired locations.
In addition to the FLP/LoxP-FRT gene switch that we constructed in this study for switching eGFP gene expression in the E. coli, we have considered another method for switching gene expression. When the promoter flanks by a pair of FRT in the opposite direction (FRT-promoter-FRT [opposite]), it can drive the expression of the eGFP. When FLP is expressed, the expression of eGFP is hindered because the promoter has been reversed. Compared with the "FRT-promoter-FRT (opposite)" method, our FLP/LoxP-FRT gene switch is stable. In the "FRT-promoter-FRT (opposite)" method, the promoter sequences between the FRT sites (opposite) would be inverted constantly and that may not regulate gene expression precisely. Therefore, as shown in our FLP/LoxP-FRT gene switch, the Nos terminator was added between the LoxP-FRT sites (same orientation) to hinder the expression of eGFP, when a plasmid expressed FLP recombinase, the Nos terminator was deleted instead of inverted, eGFP was expressed, and a stable gene switch was constructed in E. coli.
In conclusion, the FLP/LoxP-FRT gene switch has good development potential in prokaryotes and eukaryotes. First, in the research on the synthesis of natural products by recombinant microorganisms, the FLP/LoxP-FRT gene switch can be used to dynamically regulate the metabolic flow, that is, when the microorganisms grow to a certain stage, the "gene key" opens the "gene lock" to specifically express a specific protein at a specific time, thereby effectively distributing the intracellular metabolic flux and ultimately increasing the output of natural products. Second, for future research in eukaryotes, such as the cultivation of transgenic plants, the FRT/LoxP-FRT gene switch has the advantage of the ability to regulate the expression of foreign genes and marker genes to produce marker-free transgenic plants, thereby relieving public concerns about transgenic plants.
Funding information: This work was supported by the Bureau of Science and Technology of Changsha Municipality (grant kq2004052), the Natural Science Foundation of Hainan province (321MS108).