Full Text (PDF)
We are interested in identifying and characterizing recombinases that can achieve efficient site-specific insertion or cassette exchange in mammalian cells. The combination of gene targeting techniques and site-specific recombination systems (Cre/loxP and Flp/FRT) have been widely exploited in genetic analysis and in engineering complex chromosomal alterations in higher eukaryotes (1). The Cre/loxP system has been the preferred tool for genome engineering in murine embryonic stem cells (ESCs) and in mice. Cre and Flp are best suited for generating site-specific deletions (knockouts), because the 34-bp loxP and FRT recombination target sites, respectively, are recreated during recombination, and excision is preferred over integration. Mutant loxP sites have been developed to stabilize the integration event, but the efficiency of integration is still low (2). Attempts to exploit integrases from ? phage and closely related HK022 phage have resulted in limited success because of cofactor requirement and poor efficiency of intermolecular integration and intrachromosomal deletion (3,4). Recent work on integrases from FC31, R4, and TP901-1 phages demonstrated that these enzymes catalyze site-specific recombination between attP (phage attachment) and attB (bacterial attachment) sites in mammalian cells (5,6,7,8). These integrases belong to the resolvase/invertase or serine recombinase family that utilizes an N-terminal catalytic serine to mediate recombination and are structurally different from Cre, Flp, and ? int of the ? integrase or tyrosine recombinase family that utilizes a C-terminal catalytic tyrosine (9).
Recently it was shown that phage Bxb1 integrates into Mycobacterium smegmatis groEL1 gene, and Bxb1 integrase can catalyze recombination between the attP and attB sites in vitro in the absence of supercoiled DNA, cofactors, and divalent cations (10,11,12,13). Here we report that Bxb1 integrase, a serine recombinase, is functional in mammalian cells and catalyzes highly efficient unidirectional recombination between short heterologous attP and attB target sites resulting in the integration or deletion of DNA depending on the orientation and location of attP and attB sites.
To determine if the Bxb1 integrase functions in mammalian cells, we created a recombination assay plasmid pCMV-attP/attB containing a transcription termination or stop sequence, flanked by 52-bp attP and 46-bp attB sites (10), placed between the cytomegalovirus (CMV) promoter and the luciferase reporter gene ((Figure 1)A). Recombination between the attP and attB sites catalyzed by Bxb1 integrase would result in deletion of the stop sequence and activation of luciferase gene. Since the integrase gene is from bacteriophage, we totally synthesized the gene using codons optimized for high-level human and mouse expression without changing the encoded amino acid sequence (GenBank® accession no. NP_075302) and cloned it into a CMV promoter-driven mammalian expression plasmid to obtain pCMV-Bxb1. To test the recombination in various cell types, we obtained and cultured HEK 293 (ATCC, Manassas, VA, USA), NIH 3T3 (ATCC), mouse 129/S6 ESCs (Primogenix, St. Louis, MO, USA), mouse C17.2 neural stem cells (NSCs; Evan Snyder, The Burnham Institute, La Jolla, CA, USA), and rat bone marrow stromal cells (BMSCs; Osiris Therapeutics, Baltimore, MD, USA) following the recommended protocols. One day before transfection, cells were plated in a 96-well (HEK 293, NIH 3T3) or 48-well plate (BMSC, ESC, NSC) at different densities depending on the cell type (HEK 293,20,000 cells; NIH 3T3, 5000 cells; ESC, 10,000 cells, NSC, 120,000 cells; and BMSC, 3200 cells). The cells were then transfected with 25 ng recombination assay plasmid alone or along with varying amounts of pCMV-Bxb1 DNA (0, 25, 50, or 100 ng) using Lipofectamine™ 2000 (Invitrogen, Carlsbad, CA, USA) according to the manufacturer's instructions. Constitutively expressed Renilla luciferase reporter plasmid (pRL-CMV; Promega, Madison, WI, USA) was cotransfected (2 ng/well) and used as an internal control to normalize the transfection efficiency. Twenty-four hours (HEK 293, NIH 3T3, ESC) or 48 h (BMSC, NSC) after transfection, the media was discarded, and the cells were lysed with 50 µL passive lysis buffer (Promega) and 25-µL extracts were then assayed using the Dual Luciferase® Assay kit (Promega) on a plate reader equipped with injectors (Dynex Technologies, Chantilly, VA, USA). Introduction of pCMV-attP/attB and pCMV-Bxb1 plasmid DNAs into human HEK 293 cells, mouse NIH 3T3, NSCs, and ESCs, and rat BMSCs resulted in 215- to 2886-fold induction of luciferase activity depending on the cell type ((Figure 1)B, only the values obtained for 100 ng pCMV-Bxb1 are shown). Thus, the results clearly show that the Bxb1 integrase can catalyze the site-specific recombination reaction in mammalian cells, and its activity in cells from three different species suggests that the integrase may function autonomously of cellular factors.