Antibiotics blasticidin S, puromycin and G418 are frequently used for selection of stably transfected mammalian cell lines [1]. For this purpose plasmid expressing a gene of interest may be cotransfected with a plasmid containing a convenient antibiotic resistance gene [2, 3]. Alternatively the antibiotic resistance gene and the gene of interest can be combined in one plasmid [4].
Unfortunately, the choice of several antibiotic resistance markers is available only for few types of expression vectors (for example, pcDNA3.1 vector series, Invitrogen) [5]. That is why usually researchers have to introduce new antibiotic resistance genes into the original vector. In this case, the cloning strategy may be complicated by the absence of the unique and convenient restriction sites in the plasmids containing long inserts.
Here we propose to introduce the antibiotic resistance genes using recombination (Fig. 1). We have created several insertion vectors (pINS-Puro, pINS-Neo, pINS-Blast) containing the pac (puromycin-N-acetyl transferase) [6, 7], aph (aminoglycoside phosphotransferase) [8, 9] and bsd (blasticidin S deaminase) [10] genes that provide resistance to puromycin, G418 (G418 is an aminoglycoside, similar in structure to neomycin) and blasticidin S respectively (Fig 2). pINS vectors can be introduced via Cre-recombination [11] into several commercially available target vectors containing the LoxP sites, for example phrGFP vector (Stratagene). In addition we created several new target vectors: pT-FLAG, pT-BS and pT-TK (Fig. 2).
Construction of the insertion vectors pINS
We have used the backbone of the pUNI-10 plasmid [12, 13] (Fig 1, 2) for construction of the insertion vectors pINS-Puro, pINS-Neo and pINS-Blast. pUNI-10 contains the R6Kγ origin of replication [14, 15] and the LoxP site [11] recognized by Cre recombinase [16]. R6Kγ origin is active only in E. coli strains expressing the π-protein encoded by the pir gene. Cloning and production of the pINS plasmids was performed in the pir+ E. coli strain BW23474 expressing the mutant form of the π-protein (pir-116) that allows to maintain a plasmid with the R6Kγ origin at a high copy number [17, 13].
Thus the pINS vectors contain four principal elements:
- R6Kγ origin of replication;
- LoxP site required for Cre-mediated recombination with target vector;
- Genes coding for either chloramphenicol acetyl transferase [18] or aminophosphotransferase [19] providing the resistance to the antibiotics chloramphenicol (Cam) or kanamycin (Kan), respectively. These genes are required for the selection of the recombinant constructs in E. coli;
- Genes coding for either pac (puromycin-N-acetyl transferase), aph (aminoglycoside phosphotransferase) or bsd (blasticidin S deaminase) controlled by the SV40 promoter. These genes provide mammalian cells with the resistance to puromycin, G418, or blasticidin S.
Conventional E. coli strains (XL-1 Blue, DH5α, JM-109 etc.) are pir- and cannot maintain the pINS plasmid. In contrast, the products of in vitro recombination between the pINS plasmid and the target vector can successfully replicate in the pir- strains due to the presence of the active origin of replication provided by the target vector. The selection of the recombinant plasmids is achieved by the markers Kan or Cam provided by pINS plasmid. This selection procedure allows to achieve 100% yield of recombinant plasmids (Fig. 1).
Construction of the target vectors
Target vectors compatible with our pINS plasmids must contain only three necessary elements (Fig. 1):
- the LoxP site;
- An origin of replication active in the pir- E. coli strain, for example, pUC-origin [20];
- An appropriate antibiotic resistance gene, for example beta-lactamase (bla) [21] providing resistance to ampicillin (Amp).
We have modified several commercially available plasmids (phRL-TK (Promega) and pBluescriptII (Stratagene) by introduction of the LoxP sites resulting in the target vectors pT-TK and pT-BS respectively (Fig 2).
pT-TK vector contains the Renilla luciferase gene under control of the herpes simplex virus thymidine kinase promoter (TK) [22]. pT-TK vector can be used for the expression of a gene of interest at the levels that are 10-20 times lower than produced by the CMV promoter at least in some types of mammalian cells (HeLa, NIH 3T3 [23] and MEF [23]). For this purpose, the luciferase gene has to be cut out by NheI and XbaI and replaced by the gene of interest. Alternatively, any other vector can be used as a target vector in our system if upgraded by insertion of the LoxP sites as described [12].
pT-BS vector contains the convenient pBluescriptII polylinker [24] suitable for cloning of the expression modules containing a gene of interest under the control of appropriate promoter.
We have also used the commercially available target vector phrGFP (Stratagene) already containing the LoxP site. We have also created a pT-FLAG vector by replacing the GFP via FLAG-tag in the phrGFP vector (Fig 2).
pT-FLAG vector is coding for the FLAG-tag (DYKDDDDK) [25] and the cytomegalovirus promoter (CMV) [26]. It is suitable for cloning and expression of proteins with the N-terminal FLAG-tag.
All target vectors were cloned and produced in the XL-1 Blue strain (pir-).
Introduction of an antibiotic resistance gene in the target vectors by in vitro recombination
We have performed in vitro recombination between the pINS and the target vectors using Cre-recombinase. We have transformed the pir- and pir+ E. coli strains (XL-1 Blue and BW23474 respectively) with the reaction mixture in order to test the efficiency of the reaction and selected the transformants using either kanamycin, chloramphenicol or ampicillin.
Recombination mix contains the product of recombination (pINS × target vector) as well as the initial pINS and target vectors that did not take part in the reaction (Fig. 3). Recombination mix produced ampicillin-resistant colonies in cases of pir- and pir+ strains due to the presence of the initial target vector (Amp). The pir+ strain transformed by the recombination mix also produced kanamycin or chloramphenicol-resistant colonies due to the presence of the initial pINS vector (Can or Kan). In contrast, we have observed much fewer kanamycin- or chloramphenicol-resistant colonies in the pir- strain transformed by the recombination mix. These colonies only appear if the cells receive replication-competent product of recombination containing the kanamycin/chloramphenicol resistance gene (pINS × target vector). Alternatively, these colonies could appear from the contaminants of the initial plasmids.
In order to test the purity of our plasmid preparations, we have also transformed the pINS vectors and the target vectors into both pir- and pir+ strains and selected the transformants using either kanamycin/chloramphenicol or ampicillin (Fig. 3 and data not shown). As expected, the pINS vectors did not transform the pir- strain. In contrast, the pir+ strain transformed by the pINS vector can grow on either kanamycin or chloramphenicol, but not on ampicillin. The target vector transformed both XL1-Blue (pir-) and BW23474 (pir+) strains since the activity of the pUC origin of replication did not depend on the presence of the pir gene and produced the ampicillin-resistant, but neither kanamycin- nor chloramphenicol-resistant colonies. This confirmed the purity of the initial plasmids.
We calculated the yield of recombination (0.02%) by counting the kanamycin-resistant colonies of the pir- strain transformed by the recombination mix and taking into account the transformation efficiency (2.2 × 10^8 colonies/mkg DNA) (Fig. 3, and data not shown).
In order to test the integrity of the recombination product, we have picked either kanamycin- or chloramphenicol-resistant colonies, isolated plasmid DNA and digested it with an appropriate restriction enzyme. We used EcoRI in case of recombination between pINS-Puro and phrGFP. All colonies gave the restriction pattern expected for the product of recombination, thus efficiency of the resistance marker introduction is close to 100% (Fig 4 and data not shown). Moreover, due to the directional nature of the LoxP sites, integration occurs in only one orientation depending on the orientation of the LoxP sites. This feature makes the population of recombinant vectors highly homogenous (Fig. 4 and data not shown).
Next we have verified whether the function of Puro-, Blast- or G418-resistance genes from the pINS-plasmids and the gene of interest from the target vector is preserved in the product of recombination. For this purpose we have performed the recombination between each of the three insertion vectors (pINS-Neo, pINS-Puro and pINS-Blast) and the target vector phrGFP. Then we have transformed the recombination mix into the pir- strain and selected the cells containing the product of recombination by growing them on the kanamycin- or chloramphenicol-containing plates. Then we transfected the initial plasmids and the product of recombination (Neo × GFP, Puro × GFP and Blast × GFP) into HeLa cells and analyzed their resistance to either puromycin, blasticidin S or G418. As expected, only pINS vectors and the products of recombination provided the HeLa cells with the resistance against blasticidin S, puromycin and G418 (Fig 5).
Then HeLa cells resistant to the antibiotics were inspected under the microscope for the expression of GFP. Only cells transfected by the recombination products were GFP-positive. Moreover, the proportion of the GFP-positive cells was considerably higher than in the case of transient transfection by the phrGFP plasmid (Fig 6). We conclude that our recombination procedure can "safely" merge the antibiotic resistance gene and the gene of interest in one plasmid.
