1st gain & loss screening system
Two different BAC recombineering systems are widely used including those based on bacterial phage-encoded recombinases; one uses episomal plasmids to supply RecET of the Rac phage [4, 5] and the other utilizes a temperature-sensitive lamda repressor to control the expression of lamda-red recombinase [6, 7]. Lamda-red recombinase appears to be at least 50- to 100-fold more efficient than the RecET system [8]. Thus, methodology based on the lamda-red recombinase was selected in this study.
SW105 bacteria strain has been genetically modified to have the lamda-red recombinase system. It has a PL operon encoding lamda-red recombinase (exo, bet, and gam) that plays a crucial role in the recombineering process. The PL operon is under strict control of the temperature-sensitive lamda repressor (cI857). At low temperatures in the range of 30–34 °C, cI857 is active and binds to the operator site thereby preventing transcription of the recombinant genes. Thermal upshift to 42 °C reversibly inhibits the activity of cI857, thereby activating transcription of the recombinant genes.
For the recombineering, a 5′ homology region (HR) and 3′ HR was introduced into the BAC targeting vector (BTV, Addgene ID: 131589) containing GOI and a kanamycin resistant gene (KanR) (Fig. 1A). To purify BAC targeting cassette (BTC), the BTV with HRs was cut with appropriate XhoI and XmaI (Fig. 1A). As BTC contains KanR with two flanking FRT sites, the recombination of BTC with a BAC clone after brief heat shock at 42 °C will result in a positive clone, which has dual antibiotic resistant genes, KanR (from BTC) and chloramphenicol resistant gene (CamR; from the BAC).
Following the recombineering, a LB plate containing chloramphenicol (Cam) and kanamycin (Kan) was used to discriminate the candidate clone(s). The number of candidate colonies appeared after 48 h incubation at 32 °C was summarized in Fig. 1B. As BTV in supercoiled form cannot be efficiently cleaved with restriction enzymes, non-cleaved BTVs may be included in the purified BTC. BTV contamination will result in a non-recombinant clone, which has triple antibiotic resistant genes, an ampicillin resistant gene (AmpR; from BTV), KanR (from BTV) and CamR (from the BAC). To identify the clones derives from the desired recombination, the 1st gain & loss screening system was applied (Fig. 1C). The candidate clones on the plate was inoculated in the 1st gain & loss screening system and incubated for 24 h (Fig. 1C). This system will exhibit the three possible cases (Fig. 1C).
1) Case#1: positive in LB with Cam, positive with ampicillin (Amp), positive with Kan (no recombination)
2) Case#2: negative in LB with Cam, positive with Amp, positive with Kan (no recombination)
3) Case#3: positive in LB with Cam, negative with Amp, positive with Kan (recombination)
Case#3 only occurs if the recombineering is successful. The average efficiency of obtaining a case#3 was 31.94% (Fig. 1D). To confirm the correct recombination, colony PCR was conducted on all clones from case#3. The forward (Fwd) primer was located in the BAC vector and reverse (Rev) primer in BTC (Fig. 1E). The size of PCR product in 5′ recombineered and 3′ recombineered region was approximately 280 bp and 500 bp, respectively, indicating the successful recombineering (Fig. 1F). Following PCR, DNA sequencing was performed to verify the recombinant region. All positive clones were found to exhibit recombinant sequences indicating 100% accuracy of 1st gain & loss screening system (Fig. 1F). Taken together, these results imply that simple inoculation of cells into the screening system visually displayed the positive clones with 100% accuracy.
2nd gain & loss screening system
During the 1st recombineering step, successful recombination was carried out through the introduction of KanR. However, KanR should be deleted for subsequent experiments. The SW105 strain harbors an endogenous L-arabinose-inducible flippase (FLP) gene. Since KanR is flanked by two FRT sites, the induced FLP gene in the presence of L-arabinose will delete KanR (Fig. 2A).
Following FLP induction, a LB plate containing Cam was used to discriminate the candidate clone(s). The number of candidate colonies appeared after 48 h incubation at 32 °C was summarized in Fig. 2B. The 2nd gain & loss screening system was used to determine whether the flip-out reaction was successful (Fig. 2C). The positive clones on the plate was inoculated and incubated for 24 h (Fig. 2C). This system will exhibit the two possible cases (Fig. 2C).
1) Case#1: positive in LB with Cam, positive with Kan (no flip-out)
2) Case#2: positive in LB with Cam, negative with Kan (flip-out)
Case#2 only occurs if the flip-out is successful. The average efficiency of obtaining a case#2 was 100% (Fig. 2D). To confirm the correct flip-out, colony PCR was conducted on all the clone from case#2. Two primers were located outside of the two FRT sites (Fig. 2E). The size of PCR product in flip-out region was approximately 100 bp, indicating the successful flip-out (Fig. 2F). Following PCR, DNA sequencing was performed to verify the flip-out region. All positive clones were found to exhibit recombinant sequences indicating 100% accuracy of 2nd gain & loss screening system (Fig. 2F). Taken together, these results confirmed the accuracy of the 2nd gain & loss screening system to 100%.
3rd gain & loss screening system
To facilitate integration of the BAC construct into the genome, the Tol2 transposon system was used [9]. The Tol2 transposon system yields the highest rate of genomic integration in the germ lineage resulting in the increased production of desired proteins[9, 10]. As CamR is located on a backbone vector of BAC clones, CamR was targeted to introduce the Tol2 transposon system. For the recombineering, a 5′ & 3′ CamR HR was introduced into the Cam targeting vector (CTV, Addgene ID: 131590) containing inverted left & right Tol2 transposons, AmpR for gain & loss screening, and a neomycin resistance (NeoR) gene for a mammalian selection marker (Fig. 3A). To purify Cam targeting cassette (CTC), CTV was cut with BamHI and XhoI (Fig. 3A).
As the recombination of CTC with a flip-out positive clone will lose CamR but acquire AmpR (from CTC), a LB plate containing Amp was used to discriminate the candidate clone(s). The number of candidate colonies appeared after 48 h incubation at 32 °C was summarized in Fig. 3B. As CTV in supercoiled form cannot be efficiently cleaved with restriction enzymes, non-cleaved CTVs may be included in the purified CTC. CTV contamination will result in a non-recombinant clone, which has double antibiotic resistant genes, AmpR (from CTV) and CamR (from the BAC). To differentiate whether the clone derives from the desired recombination, the 3rd gain & loss selection system was used. The candidate clones on the plate was inoculated in the 3rd gain & loss selection system and incubated for 24 h (Fig. 3C). This system will exhibit the three possible cases (Fig. 3C).
1) Case#1: positive in LB with Cam, negative with Amp (no recombination)
2) Case#2: positive in LB with Cam, positive with Amp (no recombination)
3) Case#3: negative in LB with Cam, positive with Amp (recombination)
Case#3 only occurs if the recombineering is successful. The average efficiency of obtaining a case#3 was 75.69% (Fig. 3D). To confirm the correct recombination, colony PCR was conducted on all the clone from case#3. The Fwd primer was located in the BAC vector and the Rev primer in the CTC (Fig. 3E). If a recombineering occurred, the PCR product of 5′ recombineered and 3′ recombineered region should be approximately 400 bp and 390 bp, respectively (Fig. 3F). Following PCR, DNA sequencing was performed to verify the recombinant region. All positive clones were found to exhibit recombinant sequences indicating 100% accuracy of 3rd gain & loss screening system (Fig. 3F). Taken together, these results also confirmed the accuracy of the third screening system to 100%.
More strategies to increase recombination efficiency
The efficiency of the 2nd BAC recombineering (75.69%) was 2 times higher than that of the 1st BAC recombineering (31.94%). This discrepancy is presumed to be due to the difference in HR length, the decisive for the recombination efficiency [11, 12]. However, the length of 5′ HR (473 bp) and 3′ HR (486 bp) in BTV was longer than 5′ HR (200 bp) and 3′ HR (200 bp) in CTV, so other factors may be involved in this discrepancy. Incomplete enzymatic digestion of the targeting vector increases the number of false positive clones, thereby reducing recombination efficiency [13]. Strategies to induce complete enzymatic digestion are required, but incomplete digestion may occur due to several other factors including DNA methylation (23). Thus, we hypothesized that strategy to distinguish between linear DNA and uncut circular DNA would increase the recombination efficiency. As a low percentage of agarose gel can distinguish fast-moving linear DNA fragments from slow-moving circular DNA [14], we used 0.2% agarose gel and differentiated the linearized BTC (4.7 kb; red rectangle) from uncut BTV (7.6 kb; yellow rectangle) (Fig. 4A). Then, the 1st BAC recombineering was performed again. The efficiency of the 1st BAC recombineering increased from 31.94–59.38%, indicating that a strategy to purify the targeting cassette on a low percentage of agarose gel can be another decisive factor for the efficient recombineering (Fig. 4A).
Targeting vector is high copy number plasmid, while BAC is low copy number plasmid. A high copy number plasmid replicates autonomously from the bacterial chromosome and is generally present in more than one copy per cell, providing higher antibiotic resistance [15]. CTV has AmpR for selection, so bacteria with uncut CTV with non-recombinant BACs can grow better than bacteria with recombinant BACs. Thus, we hypothesized that the selection of small colony will increase the recombination efficiency. Following the 2nd recombineering, the candidate clones were distinguished using LB plates containing Amp. After 48 h incubation at 32 °C, colonies appeared as shown graphically in Fig. 4B. To test our hypothesis, large colonies (larger than 1 mm) were selected from one group and small colonies (smaller than 1 mm) from the other group. The recombination efficiency obtained when selecting small colonies was increased by about 1.6-fold compared to when selecting large colonies (47.92% vs. 29.17%, Fig. 4B).