Introduction

In most eukaryotes, RNA interference (RNAi) can mediate post-transcriptional gene silencing in a nucleotide sequence-specific manner. To initiate RNAi, double-stranded RNA is delivered to cells or tissues in one of two different forms: small interfering RNAs (siRNAs) or short hairpin RNAs (shRNAs) (Fire et al. 1998; Hammond et al. 2000; Paddison et al. 2002; Zou et al. 2003). Gene silencing by siRNA and shRNAs is faster and cheaper than a gene knockout mediated by homologous recombination. Chemically synthesized double-stranded siRNAs of about 21 bp are generally introduced into cells where they either induce the degradation of the target mRNA or inhibit its translation. To deliver shRNAs into cells, vectors containing RNA polymerase III promoters (e.g., U6 and H1 promoters) are used to express shRNAs that are processed into siRNAs inside the cell (Xia et al. 2003; Borchert et al. 2006; Pan et al. 2006; Yuan et al. 2006a, b).

Since vector systems that use RNA polymerase III promoters are not able to express shRNAs in a tissue-specific manner, RNA polymerase II promoter-based expression systems have been deployed to express shRNAs more specifically (Silva et al. 2005; Zhou et al. 2005; Yuan et al. 2006a, b; Giering et al. 2008). In such systems, shRNA is expressed in the form of a microRNA precursor in an intron, which is first processed into a mature shRNA and then into an siRNA inside the cell. Thus, shRNAs can be expressed in a tissue-specific manner, which is usually marked by the expression of a fluorescent protein. Several vectors, including retroviral and lentiviral vectors, have been developed to efficiently and robustly express intronic shRNA precursors that have different hairpin loops and promoters (Lin et al. 2006; Noguchi et al. 2007, 2011; Ying and Lin 2009). In most cases, constitutively active promoters are involved in these vector systems, although temporally inducible promoters such as the tetracycline response element (TRE) can be used to carry out more controlled experiments (Czauderna et al. 2003; Strebhardt et al. 2007; Rangasamy et al. 2008; Wiederschain et al. 2009; MATSUSHITA et al. 2013).

Lentiviral vectors have been used as the prime tool for the expression of shRNAs, particularly in cells and tissues that do not rapidly divide and in which plasmid transfection is not successful (Manjunath et al. 2009; Moore et al. 2010). Since lentiviral vectors integrate into the host genome after infection, long-term expression of shRNAs is another advantageous feature of lentiviral vectors used in gene delivery. Efforts to improve the efficiency and safety of gene delivery have resulted in several vector modifications, such as reducing the chance of generating replication-competent viral particles (Naldini et al. 1996; Kordower et al. 2000; Lois et al. 2002; Levine et al. 2006; Montini et al. 2006; Pluta and Kacprzak 2009; MD et al. 2014).

In this study, we developed a lentiviral vector system that contains both reverse tetracycline-controlled transactivator 3 (rtTA3) and the TRE promoter, which are necessary for the doxycycline-inducible expression of shRNAs. In this system, shRNAs are expressed as intronic miR-30a precursors. We found that cross-talk between the cytomegalovirus (CMV) and TRE promoters affects their activities. The relative orientation and distance between these promoters affects their strength. The maximal level of doxycycline-induced gene expression and the lowest level of transcription in the absence of doxycycline were achieved using the head-to-head orientation of both promoters and locating them far from one another. Based on these findings, we were able to construct a lentiviral vector that faithfully expresses intronic miR-30a shRNA precursors in a doxycycline-inducible manner, which could be used to temporally control shRNA expression and thus similarly regulate overall gene expression.

Materials and methods

Cell culture and transfection

HEK 293T and CHO-K1 cells were grown in Dulbecco’s modified Eagle’s medium (DMEM) (CORNING, Corning, USA) supplemented with 10% FBS (Lonza, Basel, Switzerland), penicillin (100 U/mL), and streptomycin (100 µg/mL) at 37 °C and 5% CO2. To generate a stable cell line expressing the murine EGFP-p53 fusion protein, CHO-K1 cells were transfected using the JetPRIME transfection reagent (PolyPlus, New York, USA) and selected in media containing 100 µg/mL hygromycin B. A homogeneous cell population expressing GFP was enriched by fluorescence-activated cell sorting (FACS). The CHO-K1 stable cell line expressing EGFP-p53 was maintained in DMEM containing 10% FBS, 1% penicillin/streptomycin, and 50 µg/mL hygromycin B.

For transient expression of shRNAs, 5 × 105 cells were seeded into six-well plates 24 h before transfection and transfected with plasmids expressing various forms of shRNA (2 µg) using PolyPlus following the manufacturer’s instructions. Doxycycline was added 6 h after transfection at a final concentration of 1 µg/mL, and fresh medium containing doxycycline (1 µg/mL) was added every 24 h.

Plasmids

Vectors containing the miR-30a shRNA precursor sequence (5′ end: 122 bp; shRNA-3′ end: 127 bp) and genes for different fluorescent proteins (TagBFP, EGFP, and RFP670) were prepared by the sequential ligation of PCR products and amplification of the ligation mixtures. Different combinations of PCR products were cloned between the cPPT and 3′ LTR of the pGreenPuro lentiviral vector (SBI, Palo alto, USA) as follows: pHAB267: cPPT-CMV promoter-TagBFP-T2A-rtTA3-WPRE-TRE promoter-RFP670-EGFP-3′ LTR; pHAB271: cPPT-TRE promoter-RFP670-EGFP-CMV promoter-TagBFP-T2A-rtTA3-WPRE; pHAB371: cPPT-CMV promoter-TagBFP-T2A-rtTA3-WPRE-EGFPreverse-RFP670reverse-TRE promoterreverse-3′ LTR; and pHAB372: cPPT-WPRE-rtTA3reverse-T2A-TagBFPreverse-CMV promoterreverse-TRE promoter-RFP670-EGFP-3′ LTR. In these constructs, EGFP was flanked by the 5′ and 3′ ends of the miR-30a precursor. To assemble the intronic miR-30a shRNA precursor sequences, shRNA sequences against target genes were cloned between the 5′ and 3′ ends of the precursor after removing EGFP by XhoI (NEB, Ipswich, USA) and EcoRI (NEB, Ipswich, USA) digestion.

Intronic miR30a shRNA precursor

The following sequence represents the intronic miR-30a shRNA precursor; the sense and antisense strands are represented by underlined, lowercase letters and the hairpin loop by italic, bold capital letters: TTGAATGAGGCTTCAGTACTTTACAGAATCGTTGCCTGCACATCTTGGAAACACTTGCTGGGATTACTTCTTCAGGTTAACCCAACAGAAGGCTCGAGAAGGTATATTGCTGTTGACAGTGAGCGAccctgtcatcttttgtcccttTAGTGAAGCCACAGATGTAaagggacaaaagatgacagggGTGCCTACTGCCTCGGAATTCAAGGGGCTACTTTAGGAGCAATTATCTTGTTTACTAAAACTGAATACCTTGCTATCTCTTTGATACATTTTTACAAAGCTGAATTAAAATGGTATAAATTAAATCACTTT. This shRNA precursor is located in the intron after the RFP670 ORF.

For experiments to reduce the expression of EGF-p53 using p53-specific shRNAs, the following shRNA sequences, which were shown to knock down the expression of p53, were replaced by the XhoI (CTCGAG)/EcoRI (GAATTC) fragment in the miR-30a shRNA precursor.

p53-393 shRNA

CTCGAGAAGGTATATTGCTGTTGACAGTGAGCGAccctgtcatcttttgtcccttTAGTGAAGCCACAGATGTAaagggacaaaagatgacagggGTGCCTACTGCCTCGGAATTC.

p53-814 shRNA

CTCGAGAAGGTATATTGCTGTTGACAGTGAGCGCccactacaagtacatgtgtaaTAGTGAAGCCACAGATGTAttacacatgtacttgtagtggATGCCTACTGCCTCGGAATTC.

Negative control shRNA

CTCGAGAAGGTATATTGCTGTTGACAGTGAGCGAggatgtttcaccaaggaaataTAGTGAAGCCACAGATGTAtatttccttggtgaaacatccGTGCCTACTGCCTCGGAATTC.

The above oligonucleotide templates were amplified by PCR using the primers (5′- GTGCAATGCTCGAGAAGGTATATTG-3′ and 5′- TAAGCGATGAATTCCGAGGCAGTAG-3′). Then XhoI/EcoRI digests of the PCR products were cloned into the same restriction enzyme (RE) sites in the vector containing the miR-30a shRNA precursor.

Flow cytometric analysis

To quantitate the expression of fluorescence proteins (TagBFP, EGFP, and RFP670) by flow cytometry, HEK 293T cells were transiently transfected with pHAB267, pHAB271, pHAB371, or pHAB372. After treatment with doxycycline, cells were treated with trypsin (0.25%), briefly washed with 1 × PBS, and resuspended in the FACS buffer (1 × PBS, 2% FBS) before harvest.

A Moflo XDP flow cytometer (Beckman Coulter, Brea, USA) was used for the analysis. The instrument settings were as follows: log forward scatter (FSC) and log side scatter (SSC): 440 V. TagBFP was excited at 405 nm. EGFP was excited at 488 nm, and RFP670 was excited at 633 nm. TagBFP, EGFP, and RFP670 fluorescence was collected using 450/15 nm, 529/28 nm, and 670/30 nm bandpass filters, respectively. The singlet cell population was gated on FSC and SSC. The typical sampling rate was 500 cells per second, and the typical sample size was 50,000 cells per measurement unless otherwise stated. The data were analyzed using Kaluza software (Beckman Coulter, Brea, USA).

Results and discussion

A lentiviral vector for the expression of shRNAs from a doxycycline-inducible TRE promoter

To understand how the regulatory elements in lentiviral vectors affect the expression of intronic shRNAs, we constructed a stable cell line in which the EGFP-p53 fusion protein was produced from the UBC promoter. Most cells expressed a uniform level of GFP, as measured by fluorescence microscopy and flow cytometry.

We constructed a lentiviral vector to express shRNAs from a doxycycline-inducible TRE promoter: TagBFP and rtTA3 were constitutively expressed from the CMV promoter, and RFP670 and intronic miR-30a shRNA were expressed from the doxycycline-inducible TRE promoter (Fig. 1). We also generated a vector with the EGFP ORF that contained the region between the 5′ and 3′ ends of the miR-30a precursor, flanked by XhoI and EcoRI RE sites (Fig. 1). We chose the intronic miR-30a precursor because it is known to efficiently proceed to a mature shRNA after transcription by the RNA polymerase II promoter (Lee et al. 2004). The vector also contains the rtTA3 transactivator that is necessary for doxycycline-inducible transcription from the TRE promoter (Strebhardt et al. 2007).

Fig. 1
figure 1

A lentiviral vector is used to express shRNAs from a doxycycline-inducible promoter. Both TagBFP and rtTA3 are expressed as a single RNA transcript from the CMV promoter and translated independently due to the presence of the T2A peptide between them. During translation, a stalled ribosome on the T2A sequence falls off and reinitiates translation from an ATG codon in rtTA3. TagBFP marks the cells infected with the lentiviral vector. The rtTA3 protein is the transactivator necessary for transcription from the TRE promoter. The shRNAs is expressed in the form of an intronic miR-30a precursor. For enzymatic cloning, an shRNA precursor sequence is cleaved into 5′ and 3′ segments with the EGFP ORF cloned between them. When necessary, the shRNA is cloned into the precursor region using EcoRI and XhoI digestion

Full size image

In the pHAB267vector, TagBFR-T2A-rtTA3 was transcribed as a single transcript but translated as independent polypeptides: TagBFP and rtTA3. Thus, the cells that contained this vector were expected to express TagBFP and support the doxycycline-inducible expression of RFP670 and the intronic miR-30a precursor as a single transcript. During splicing, the mature miR-30a shRNA was produced from its intronic precursor. As described in the materials and methods, different shRNA sequences can be cloned into the vector as precursor forms using EcoRI and XhoI RE sites (Fig. 1).

shRNAs against p53 block its expression regardless of doxycycline treatment

When two different shRNAs (mp53-393 and mp53-814) that specifically target p53 were transiently expressed in the CHO-K1 stable cell line expressing EGFP-p53, GFP expression was significantly reduced, as measured by flow cytometry (Fig. 2). A negative control shRNA that does not target p53 did not cause any reduction of the GFP signal (Fig. 2). Thus, it appears that intronic shRNA precursor RNAs are transcribed from the transiently transfected plasmid and then processed into mature shRNAs. Interestingly, the specific knockdown of GFP-p53 expression by p53-specific shRNAs was observed even without doxycycline treatment, suggesting the existence of other mechanism(s) that express intronic shRNA precursors other than the doxycycline-inducible TRE promoter (Fig. 2).

Fig. 2
figure 2

shRNAs against p53 knock down p53 expression independent of doxycycline. a Timeline of doxycycline treatment and FACS analysis. b Flow cytometric analysis of the CHO-K1 stable cell line expressing EGFP-p53; 1.0 × 104 cells were counted for each FACS experiment. c The EGFP intensity in each cell was divided by the number of cells analyzed to obtain the average EGFP signal and EGFP-p53 level

Full size image

To discern whether doxycycline-independent transcription is caused by the vector itself or by the CHO-K1 cell line expressing EGFP-p53, we transiently transfected HEK 293T cells with the pHAB267 vector, which has the EGFP ORF in place of the miR-30a shRNA. As shown in Fig. 3, TagBFP was constitutively expressed from the CMV promoter, with or without doxycycline. RFP670 expression was induced by doxycycline, as seen in the previous experiment, but RFP670 expression without doxycycline treatment was significantly higher than the control. The same amount of EGFP was produced regardless of doxycycline treatment (Fig. 3).

Fig. 3
figure 3

EGFP in the RFP670 intron is strongly expressed in the absence of doxycycline. HEK 293T cells were transfected with pHAB267 to measure the transcriptional activity of the CMV and TRE promoters. TagBFP is directly downstream of the CMV promoter, RFP670 is downstream of the TRE promoter, and EGFP is located in the RFP670 intron, flanked by the 5′ and 3′ ends of the miR-30a precursor; 1.0 × 104 cells were counted for each FACS experiment. Dox (−): without doxycycline, Dox (+): with doxycycline

Full size image

The doxycycline-independent expression of shRNAs or EGFP suggests the existence of a cryptic promoter in the 200-bp intronic region containing the 5′ end of the miR-30a precursor. Of note, the RFP670 ORF before the intron contains an in-frame stop codon, ruling out a possible RFP670-GFP fusion protein. Different expression patterns of RFP670 with GFP also preclude the existence of a fusion protein (Fig. 3).

To identify the putative cryptic promoter, we carried out deletion and site-directed mutagenesis studies of the 200-bp region between RFP670 and EGFP. As shown in Fig. 4, a deletion of the majority of the 200-bp region and site-directed mutagenesis of a putative TATA element did not change the expression patterns of RFP670 and EGFP, or the doxycycline independence of EGFP expression. Based on these results, we conclude that the aberrant expression of EGFP is not due to the presence of a cryptic promoter.

Fig. 4
figure 4

Mutagenesis of the intronic 5′ miR-30a precursor region to look for a cryptic promoter. a Deletion mutants in which 50-bp regions were deleted in the intronic 5′ miR-30a precursor region (50-bp deletion in pHAB286, 100-bp deletion in pHAB287, and 150-bp deletion in pHAB288) in the pHAB267 vector were analyzed for RFP670 and EGFP expression using flow cytometry. b HEK 293T cells were transfected with the pHAB267, pHAB286, pHAB287, or pHAB288 vector and treated with doxycycline, and the expression levels of RFP670 and EGFP were analyzed by FACS. c Site-directed mutagenesis of a putative TATA element in the 5′ region of the miR-30a precursor. ‘TATATTG’ in pHAB267 was modified to ‘TACCTTG’ (pHAB275), ‘CCCCTTG’ (pHAB276), and ‘TACCCCG’ (pHAB277). d HEK 293T cells were transfected with each site-directed mutant construct, and EGFP expression was analyzed by FACS

Full size image

The effects of promoter combinations on doxycycline-inducible TRE transcription

To determine whether different combinations of the CMV and TRE promoters cause different expression patterns of RFP670 and EGFP, we constructed different vectors in which the position and orientation of the promoters differed from each other. The pHAB267 vector contains a head-to-tail orientation of CMV-TRE, pHAB271 contains a head-to-tail orientation of TRE-CMV, pHAB371 contains a head-to-head orientation of CMV-TRE, and pHAB372 contains a tail-to-tail orientation of CMV-TRE (Fig. 5a). When each vector was transiently transfected into HEK 293T cells, doxycycline-inducible expression from the TRE promoter was only evident in the pHAB371 vector, as measured by the expression levels of EGFP and RFP670 using flow cytometry (Fig. 5b, c). When both promoters were closest to each other (tail-to-tail orientation in pHAS372), the highest expression of RFP670 and EGFP from the TRE promoter occurred in the absence of doxycycline. When both promoters were in the head-to-tail orientation, the expression levels of RFP670 and EGFP were the same with and without doxycycline, regardless of the order of the promoters (pHAB267 vs pHAB271). From these results, we conclude that both the orientation and distance between nearby promoters affect their activity and that promoter cross-talk is significantly reduced when promoters are in a head-to-head orientation (far from each other), which is important for the differential expression of a gene from a constitutively active promoter and a gene from an inducible promoter in a single vector.

Fig. 5
figure 5

Promoter cross-talk depends on the orientation and distance of promoters with respect to one another. a Vectors with different distances between and orientations of the CMV and TRE promoters. b HEK 293T cells were transfected with each vector and grown in media without doxycycline; the expression levels of RFP670 and EGFP were measured by flow cytometry. c HEK 293T cells were transfected with pHAB267 or pHAB371, and grown in the media with or without doxycycline; the expression levels of RFP670 and EGFP were measured by flow cytometry

Full size image

Doxycycline-inducible expression of intronic shRNAs from TRE promoter

To determine whether intronic shRNAs are expressed from the TRE promoter in the pHAB371 vector in the presence of doxycycline, a negative control and two different shRNAs (that specifically target murine p53) were first cloned in the pHAB371 vector and then transiently expressed in the CHO-K1 stable cell line expressing EGFP-p53. GFP-p53 expression was subsequently measured by flow cytometry (Fig. 6). Compared to the pHAB267 vector, which contains the CMV-TRE promoters in a head-to-tail orientation and thus expresses RFP670 and EGFP in a doxycycline-independent manner (Figs. 2, 6), cells transfected with pHAB371 expressed shRNAs only when doxycycline was added to the medium. Thus, shRNAs against p53 (mp53-393 and mp53-814), but not a negative control shRNA, knocked down GFP-p53 expression in the presence of doxycycline.

Fig. 6
figure 6

Inhibition of EGFP-p53 expression by the doxycycline-inducible expression of intronic shRNAs from the TRE promoter. a CHO-K1 cells stably expressing EGFP-p53 were transfected with pHAB267 or pHAB371 vectors containing negative control shRNA, p53-393 shRNA, or p53-814 shRNA. Cells were grown in media with or without doxycycline, and GFP levels were measured by flow cytometry. b, c The EGFP intensity in each cell was divided by the number of cells analyzed to obtain the average EGFP signal and EGFP-p53 level

Full size image

It is unclear how the EGFP ORF was expressed without its own promoter. A deletion of the intronic region and site-directed mutagenesis of a putative TATA element ruled out the presence of a cryptic promoter upstream of EGFP. Since the EGFP ORF is located in an intron, translation could be initiated from the 5′ end of the intron that is separated from the RFP670 exon by RNA splicing.

Conclusions

In this study, we showed that promoter cross-talk depends on the relative position and orientation of the promoters involved. The head-to-head orientation of promoters results in minimal effects on promoter activity, compared to other orientations (tail-to-tail and head-to-tail). Based on this information, we constructed a vector that could inducibly express intronic shRNAs from the TRE promoter, allowing the performance of controlled experiments of temporally regulated shRNA expression in the presence or absence of doxycycline.