complete control system for inducible mammalian expression
Denise Wyborski Peter Vaillancourt
The complete control mammalian expression system allows tight control of gene expression in a wide range of mammalian cell types. The inducible promoter used in the system is naturally repressed in the absence of the ecdysone analog ponasterone A (ponA). In a double-stable cell line, a linear dose-response is achieved over a wide range of ponA concentrations (4 fold to >500 fold) and, in a time-course experiment, a linear increase was achieved from 1 hour post induction (5 fold) to 20 hours (1,030 fold). Coexpression of both receptors from a single dicistronic transcript from the vector pERV3 facilitates replacement of the CMV promoter with a cell type-specific promoter of interest. The inducible vector pEGSH is engineered so that expression of the gene of interest can be monitored by Western blot analysis or RNA detection.
DNA vector-based systems that allow precise control of gene expression in vivo are invaluable for studying gene function in a variety of organisms, particularly when studying developmental and other biological processes for which the timing or dosage of gene expression is critical to gene function. Such systems are successfully used to overexpress toxic or disease-causing genes, induce gene targeting, and express antisense RNA. Pharmaceutical companies currently use inducible systems to facilitate screening for inhibitors of clinically relevant biological pathways and to explore potential applications for gene therapy.1
Most inducible mammalian systems currently available employ either natural promoters that are induced by heavy-metal ions, heat shock, growth factors, and steroid hormones or employ synthetic promoters and inducible activators that often contain cis and trans elements derived from bacteria or yeasts.2 However, the majority of these systems have drawbacks: pleiotropic effects caused by the inducer or the transcriptional activator, prohibitively high background expression in the absence of inducer, or poor penetrance and/or clearance of the inducer in some tissues.
Stratagene has recently introduced a better system for controlled mammalian expression based on the finding that the insect molting hormone, ecdysone, stimulates transcriptional activation in mammalian cells harboring the ecdysone receptor protein from the fruit fly Drosophila melanogaster.3 The ecdysone analogs ponA and muristerone A (murA) efficiently penetrate all tissues, including the brain, due to their lipophilic nature and short in vivo half-life. This results in rapid and potent inductions and rapid clearance. Ecdysteroids are not known to affect mammalian physiology in any measurable way.
The ecdysone receptor (EcR) is a member of the RXR heterodimer family of nuclear receptors. In mammalian cells, the EcR heterodimerizes with the retinoid- X-receptor (RXR), the mammalian homologue of USP. The EcR-RXR heterodimer is capable of binding to and activating reporters that contain multiple copies of the ecdysone-responsive element (EcRE). However, because EcRE-containing reporters can be specifically trans-activated by some other lipophilic steroids, the EcR protein and EcRE recognition sequence were modified to create both a synthetic ecdysone-inducible receptor that would not bind to and trans-activate any endogenous host genes, as well as a synthetic recognition site that would not be recognized by host transcription factors.
Three amino acids in the EcR DNA-binding domain (DBD) were mutated to change its DNA-binding specificity to that for the glucocorticoid receptor (GR).3 The GR-EcR (GEcR) fusion protein retains the ability to dimerize with RXR and trans-activate in a ponA-dependent manner and is able to recognize and activate reporters containing the synthetic binding site AGTGCA N1 TGTTC (E/GRE). This binding site is extremely unlikely to be recognized by steroid family receptors, which require perfect inverted half-sites, or by natural RXR heterodimer family receptors, which require single nucleotide spacing between half-sites. Finally, the GEcR receptor was further modified by replacing the EcR AD with the more potent VP16 trans-activator to create the receptor protein VgEcR (Figure 1).
Two vectors are required for ponA-inducible expression of the gene of interest: the receptor expression vector pERV3, from which VgEcR and RXR are constitutively expressed, and the ecdysone-inducible vector pEGSH (Figure 2). The pERV3 vector is engineered such that both receptors are expressed from a single mRNA transcribed from the CMV promoter. We accomplished this by placing the internal ribosome entry site (IRES) from encephalomyocarditis virus (EMCV) upstream of the second (RXR) open reading frame (ORF), which allows high-level internal (cap-independent) initiation of translation of ORFs positioned downstream in an appropriate context.4 Expressing both receptors from a single transcript has many advantages. For example, transcription of this expression cassette can be achieved in a wide variety of cell types from a single plasmid, due to the versatility of the CMV promoter; and the CMV promoter in this construct can be readily replaced with other promoters to confer cell-type specificity to receptor expressionan advantage that is particularly attractive for the construction of transgenic animals. The plasmid also contains a neomycin-resistance gene, which is expressed in both E. coli (kanamycin-resistance) and mammalian cells (G418-resistance).
The ecdysone-inducible expression vector pEGSH contains the ponA-inducible expression cassette comprised of 5 E/GRE binding sites upstream of a minimal promoter consisting of three SP1 sites, followed by the D. melanogaster hsp27 minimal promoter (Figure 2B ). The vector contains the hygromycin-resistance gene to allow stable selection in cells transformed with the pERV3 plasmid. The MCS of pEGSH was engineered to contain an array of restriction sites positioned for directional cloning of inserts derived from Lambda ZAP-derived cDNA vectors*, lgt10, lgt11, HybriZAP vectors, and other two-hybrid libraries, in addition to most other popular cDNA cloning and expression vectors.
In addition to this versatile array of restriction sites, the pEGSH MCS contains the following features: the ability to monitor expression of the gene of interest, by either a-FLAG immunodetectionll ll or by RNA detection using T3 antisense RNA probes; and the ability to seamlessly fuse the insert to the FLAG epitope and/or the HSP leader by using the seamless cloning kit** for which 80% cloning efficiencies are routinely achieved.
To demonstrate the quality of both the receptor expression and ecdysone-inducible vectors, we performed transient expression assays using the reporter pEGSH-luc in which the coding sequence for the firefly luciferase gene is inserted into the MCS of pEGSH. In these experiments, reporter was cotransfected with various amounts of the pERV3 receptor vector, induced with murA, and assayed for luciferase activity. Figure 3 shows that, in CHO cells, optimal expression is achieved with a relatively small amount of receptor vector (8 ng) to give an induction ratio of over three orders of magnitude. We achieved comparable results for similar experiments carried out in NIH3T3, 293, and CV-1 cells (data not shown).
The best method to engineer double-stable cell lines is to first produce stable receptor-expressing cell lines using the plasmid pERV3 and then screen stable receptor-expressing lines by transient transfection of the inducible reporter pEGSH-luc to find one that mediates the highest level of ponA-dependent trans-activation. Once produced, this line can be used for constructing pEGSH-derived lines. To expedite this process, we have produced three engineered cell lines derived from CHO, NIH3T3, and 293 cells, respectively, in which optimal levels of the receptors are stably expressed. These three lines show high-level inducer-dependent activation in transient expression assays using the pEGSH-luc reporter (Figure 4).
The stable receptor line ER-CHO (Figure 4A ) was stably transfected with the plasmid pEGSH-luc, and hygromycin- resistant clones were screened for ponA-dependent induction of luciferase activity. Figure 5 shows the results for this double-stable cell line designated HSL-34. The results in Figure 5A show that both murA and ponA induce luciferase activity in HSL-34 cells to comparable levels over a 100-fold range of inducer concentration. To more accurately assess the sensitivity of the system, both ponA concentration and induction time were varied. As the results in Figure 5B indicate, a linear response was observed from 80 nM ponA (four fold: the lowest concentration tested) to 10 M.
To more directly assess the control of protein expression in this cell line, we performed a Western blot. Lysates from uninduced cells and cells induced with increasing amounts of ponA were fractionated by SDS-PAGE, blotted, and probed with a-luciferase polyclonal antisera (Figure 5C ). No significant detectable luciferase is expressed in the uninduced extract, whereas a linear increase in signal is observed from 300 nM to 5 M ponA. In a time-course experiment, a five-fold induction was achieved only 1 hour after adding 10 mM ponA, and a linear increase was observed for up to 20 hours, at which time an induction ratio of 1,030-fold was observed (Figure 5D ). We also observed these induction kinetics in transient assays (data not shown).
Taken together, these results indicate that, by varying the ponA concentration, precise control of gene dosage can be achieved with this system. Furthermore, comparatively brief induction periods can result in low to moderate levels of gene expression.
Current commercially available systems lack a convenient tool to monitor the induced expression of the protein of interest when specific antibody is unavailable. To this end, the pEGSH vector has been engineered to provide a convenient means to detect both RNA and protein expression. For RNA detection, the T3 promoter has been positioned downstream of the MCS in the antisense orientation relative to the inducible expression cassette (Figure 2B ); labeled antisense RNA can be readily produced in vitro to use as a probe in Northern or dot blots, in situ hybridization, or RNase protection (data not shown).
To directly detect protein expression without the need for protein-specific antibody, the eight amino acid FLAG tag is positioned for C-terminal fusion to the protein of interest. Therefore, the induction of FLAG fusion proteins can be monitored by Western blotting or other immunodetection methods using the a-FLAG M2 antibody (data not shown).
The Complete Control system has several advantages over other inducible systems. The inducer exhibits no pleiotropic effects on cellular physiology and the inducers lipophilic nature and short in vivo half-life ensure that it rapidly penetrates and clears all tissues, as well as exhibits dose-dependent control of gene expression. In addition, the ecdysone receptor and its DNA recognition element are genetically altered so there is no cross-talk between the system and endogenous pathways. The uninduced basal activity of the system is extremely low and can be induced to over three orders of magnitude. Finally, the systems pERV3 and pEGSH vectors achieve more consistent expression of both receptors in a wide range of cell types, provide convenient means to monitor gene expression , and allow high-efficiency directional cloning of any insert due to the versatile MCS.
We thank Drs. Ronald M. Evans, Enrique Saez, David No (Salk Institute), Tanya Hosfield, Quinn Lu, Grady Howe, Tim Sanchez, Cathy Chang, Mary Buchanan, and Joe Sorge for helpful discussions and providing reagents; John Bauer for his devotion to the Quest for the Perfect Inducer; and Alan Greener for his continuing support.
Schockett, P.E. and Schatz, D.G. (1996) Proc. Natl. Acad. Sci. 93: 5173-5176.
Deuschle, U., et. al. (1995) Molec. Cell Biol. 15(4): 1907-1914.
No, D., et. al. (1996) Proc. Natl. Acad. Sci. 93: 3346-3351.
Borman, A.M. (1997) Nuc. Acid Res. 25(5): 925-932.
Aslandis, C. and de Jong, P.J. (1990) Nuc. Acid Res. 18: 6069-6074.
* U.S. Patent Nos. 5,128,256, 5,286,636 and European Patent No.
* * Patent pending