Woodchuck Hepatitis Virus Posttranscriptional Regulatory Element Enhances Expression of Transgenes Delivered by Retroviral Vectors

Plasmids. (i) HIV-1 vector plasmids.

Plasmids pHR′CMV-GFP and pHRCMV-Luc have been described previously (29). A PCR-amplified WPRE cassette (nucleotides 1093 to 1684; GenBank accession no. J04514) was modified with ClaI or EcoRI ends and inserted into pHR′CMV-GFP either at the unique ClaI site upstream of the cytomegalovirus (CMV) promoter or at the unique EcoRI site downstream of the GFP stop codon, resulting in plasmids pHR′W-CMV-GFP and pHR′CMV-GFP-W, respectively. Subsequently, the BamHI-XhoI GFP coding sequence was replaced with a BamHI-XhoI luciferase coding sequence to generate pHR′CMV-Luc-W. Plasmids pMD.G and pCMVΔR8.91 have been described previously (29).

(ii) MLV vector plasmids.

WPRE was inserted as a ClaI cassette into the unique ClaI site of plasmid pCLNCX (21). Subsequently, a BamHI-HindIII thymidine kinase (TK)-luciferase cassette was substituted for the CMV promoter, resulting in plasmid pCLNTluc-W. The ClaI WPRE cassette was deleted to generate control plasmid pCLNTluc.

Tissue culture and transfection.

Dulbecco’s modified Eagle’s medium (Gibco) was supplemented with 10% fetal calf serum (Gibco), a combination of penicillin and streptomycin (Gibco), and glutamine (Gibco). 293T, HeLa, HeLa-tat, HOS, 208F, and NIH 3T3 cells were cultured in supplemented Dulbecco’s modified Eagle’s medium in a 10% CO₂ atmosphere. Gamma irradiation was delivered by a 3-min exposure to a ⁶⁰CO source. Vector stocks were prepared and cells were transduced as previously described (29).

To determine the titers of GFP-transducing vectors, five serial 1:2 dilutions of each filtered vector stock were used to transduce HeLa cells in six-well plates (2 × 10⁵ cells/well). The highest and lowest inocula corresponded to 100 ml and 6.25 ml of undiluted supernatant, respectively. Vector particles were added to 2 ml of culture medium in the absence of Polybrene and left on the cells for 48 to 60 h. At this time, the percentage of GFP-positive cells was determined with a fluorescence-activated cell sorter on a Beckton Dickinson FACScan. To calculate titers (transducing units per milliliter), 2 × 10⁵ cells/well was multiplied by the percentage of GFP-positive cells, and this product was divided by the number of microliters in the inoculum. Numerous titer determinations have shown that the percentage of transduced cells correlates linearly with the vector input when the percentage is lower than 12%. Therefore, all titers were based on at least two values lower than 12% and showing the expected linearity.

Southern analysis.

Genomic DNA was isolated from three 10-cm plates of each cell line. Cells were lysed, phenol extracted, and ethanol precipitated by standard methods. Ten micrograms of genomic DNA was digested overnight with BamHI, EcoRI, and XhoI restriction enzymes. Digested DNA was ethanol precipitated and electrophoresed on a 0.9% agarose gel. DNA was visualized with ethidium bromide staining before being transferred to a nylon membrane by standard methods.

RNA isolation and analysis.

Cells were washed with phosphate-buffered saline (PBS) and pelleted by centrifugation. For total RNA, the cell pellet was resuspended in 250 μl of PBS, and 750 μl of RNA Stat LS-50 (Tel-Test) was added. For nuclear and cytoplasmic fractionation, cells were resuspended in cytoplasmic lysis buffer (10 mM HEPES [pH 7.8], 10 mM KCl, 0.1 mM EDTA, 20% glycerol, 0.5% Nonidet P-40). The lysed cells were spun for 3 min at 8,000 × g, and the supernatant was recovered and spun for an additional 5 min at 14,000 × g. The supernatant was transferred to 1 ml of RNA Stat LS-50. The nuclear pellet from the first spin was resuspended in 1 ml of cytoplasmic lysis buffer and then spun at 8,000 × g for 3 min. The supernatant was discarded, and the pellet was resuspended in 800 μl of nuclear buffer (10 mM Tris [pH 8.4], 1.5 mM MgCl₂, 140 mM NaCl, 20% glycerol). The sample was spun at 8,000 × g, and the supernatant was discarded. The pellet was resuspended in 300 μl of nuclear buffer and lysed with 1 ml of RNA Stat LS-50. The RNA Stat LS-50 protocol was followed.

For RNA half-life analysis, cells were grown to 70% confluency and treated with 5 mg of actinomycin D per ml. For each time point, two plates of cells were harvested. The cells were pelleted and resuspended in 250 μl of PBS, and 750 μl of RNA Stat LS-50 was added. After RNA purification, the samples were treated with DNase for 15 min at 37°C. Five micrograms of nuclear RNA and 10 μg of cytoplasmic RNA were separated on a 1% agarose–formaldehyde gel, transferred to a nylon membrane, and hybridized with a GFP probe by use of Quickhyb (Qiagen) and the manufacturer’s protocol.

Nuclear run-on assays.

For each experiment, 5 × 10⁷ cells were harvested. The nuclei were prepared by lysing the cells in cell lysis buffer (10 mM Tris [pH 8.3], 10 mM NaCl, 5 mM MgCl₂). The nuclei were washed once in cell lysis buffer and frozen overnight. The run-on reactions were performed with 25 mM Tris (pH 8.0)–12.5 mM MgCl₂–750 mM KCl–1.25 mM each ATP, CTP, and GTP–30 μl of UTP (800 Ci/mmol). The reaction mixtures were incubated for 30 min at 30°C. The nuclei were homogenized in 750 μl of RNA Stat LS-50 (Tel-Test). The labeled RNA was purified in accordance with the manufacturer’s specifications. The RNA samples were treated with DNase, phenol extracted, and centrifuged through G-25 columns. The samples were ethanol precipitated and resuspended, and counts were determined with a scintillation counter. Equivalent counts were hybridized with nitrocellulose filters containing plasmid DNAs for histone H2B and glyceraldehyde-3-phosphate dehydrogenase (GAPDH) and PCR products derived from vector sequences by use of Quickhyb and the manufacturer’s protocol. Hybridized filters were treated with RNase A and analyzed.

WPRE enhances the expression of transgenes delivered by HIV-based vectors.

To test whether the WPRE enhances the expression of intronless reporter genes delivered by HIV-based vectors, a 600-nucleotide-long cassette containing the WPRE sequence was cloned into the previously described plasmids pHR′CMV-Luc and pHR′CMV-GFP (29). The WPRE was inserted in the 3′ untranslated region (UTR) of the reporter genes between the stop codon and the polypurine tract (Fig. 1A). To produce vesicular stomatitis virus G protein (VSV-G)-pseudotyped transducing particles, vector plasmids with or without the WPRE were cotransfected into 293T cells with the envelope plasmid pMD.G and the packaging plasmid pCMVΔR8.91 by a published protocol (20). The resulting vectors were used to transduce in parallel different cell lines. These experiments were done at a multiplicity of infection (MOI) of 0.1 to favor a single integration per cell. Dilution analysis of the vector stocks on 293T cells showed that the WPRE did not influence titers (data not shown). However, fluorescence-activated cell sorter quantification of GFP expression in cells transduced with HR′CMV-GFP or HR′CMV-GFP-W demonstrated that the presence of the WPRE in the transgene 3′ UTR (indicated by the suffix W in the plasmid name) increased the mean expression of the transduced population by at least fivefold in 293T cells (Fig. 2A) and threefold in HeLa cells (Fig. 2B). In contrast, the WPRE inserted in the opposite orientation (indicated by the suffix W(as) in the plasmid name) inhibited GFP expression in HeLa cells (Fig. 2C). Therefore, WPRE function is orientation dependent, as has been reported previously (8).

Further studies revealed that the WPRE had to be present within the transgene transcript to function. When the WPRE was inserted upstream of the CMV promoter governing the transcription of the GFP gene (Fig. 1A), it was unable to stimulate GFP expression in HeLa cells (Fig. 2D). This result demonstrates that the WPRE must be present within a transcript to stimulate expression, consistent with a posttranscriptional mode of action.

To exclude the possibility that the WPRE effect was gene specific, we tested the ability of the WPRE to stimulate luciferase expression. The luciferase activities induced in various target cells by either HR′CMV-Luc or HR′CMV-Luc-W were compared. Viral stocks were normalized for the amounts of HIV-1 p24 capsid antigen present in the inoculum (Fig. 2E). The WPRE in the 3′ UTR of the luciferase cDNA increased luciferase production by a factor of 5 to 6 (mean, 5.27) in all cell lines tested (293T, HeLa, HOS, 208F, and NIH 3T3). The WPRE is thus active in both human and rodent cell lines and in cells of epithelial, osteoblastic, or fibroblastic origin.

To explore if WPRE function is sensitive to the proliferation status of transduced cells, we analyzed the level of luciferase expression in dividing and gamma-irradiated 293T cells. Cells were transduced with HIV-based luciferase vectors carrying the WPRE sequence in either the sense or the antisense orientation (Table 1). When a vector without the WPRE was used as a reference, the WPRE in the sense orientation resulted in a eightfold increase in luciferase expression in both dividing and arrested 293T cells. The WPRE in the antisense orientation was nonfunctional. This experiment reveals that the proliferation status of the target cells has no impact on WPRE function, an important feature for in vivo expression in terminally differentiated tissues.

WPRE enhances the expression of transgenes carried by MLV-based vectors.

Since the WPRE is functional in settings as different as those of WHV and HIV-based vectors, it was considered likely that it would also be functional when incorporated into a vector derived from a simple retrovirus. To test this hypothesis, the WPRE was inserted in the 3′ UTR of the luciferase cDNA delivered by an MLV-based vector (21). Instead of the CMV promoter, the promoter from the human herpes simplex virus TK gene was chosen to demonstrate that WPRE action is not dependent on a particular promoter (Fig. 1C). Stocks of VSV-G-pseudotyped MLV-based vectors were generated by cotransfection of 293T cells with three plasmids, pMD.G, pCMVGagPol, and pCLNTluc or pCLNTluc-W, and used to transduce 293T cells. The induction of luciferase activity in these targets was measured and normalized for the amount of reverse transcriptase activity present in the inoculum. The results revealed that the presence of the WPRE increased the levels of expression of luciferase delivered by MLV-based vector-mediated transduction more than fourfold (relative light units with CLNTluc and CLNTluc-W [mean ± standard error of the mean for triplicate batches of each vector in two independent experiments], 139,518 ± 11,349 and 571,887 ± 7,319, respectively).

WPRE also acts on spliced mRNAs.

In its natural context, the WPRE is located within intronless mRNAs. To test whether this element could facilitate the expression of spliced mRNAs, the WPRE was inserted in HIV-based vectors expressing GFP or luciferase but devoid of the internal CMV promoter (Fig. 1B). In these vectors, the reporter-encoding RNAs are produced by the 5′ HIV long terminal repeat (LTR). Transgene expression measured in this system is from the spliced message. It is not expected that unspliced messages will substantially contribute to transgene expression because 10 ATG triplets present in the intron sequence may act as aberrant translational start sites. Since the HIV LTR is a weak promoter unless stimulated by the HIV Tat protein or the adenovirus early protein E1A, we used HeLa-tat and 293T cells as targets. As shown in Table 2, the WPRE enhanced luciferase expression in 293T cells eightfold. This level of stimulation by the WPRE is comparable to that observed for intronless mRNAs. In HeLa and HeLa-tat cells, the WPRE increased luciferase expression by factors of 4 and 5, respectively. The similar levels of enhancement observed in HeLa and HeLa-tat cells indicated that the WPRE is effective over a wide range of promoter activities, since the LTR is 30 times more active in HeLa-tat cells than in HeLa cells. Comparable results were obtained with the GFP gene (data not shown). As noted for intronless transcripts, the action of the WPRE on spliced mRNAs was orientation dependent.

Mechanism of WPRE action.

It was previously established that HPRE and WPRE act at a posttranscriptional level (7, 11, 13). To explore the mechanism of WPRE action in a lentivirus context, 293T cells were transduced with HR′CMV-GFP and HR′CMV-GFP-W at an MOI of 0.2. Fluorescent cells were sorted, expanded, and used for comparative DNA and RNA analyses (Fig. 3). Southern blot analysis demonstrated that the GFP transgene was present at comparable copy numbers in both populations (Fig. 3A). Nuclear run-on analysis (Fig. 3B) demonstrated that the WPRE did not influence the rate of transcription of vector-based messages. The observation that the WPRE does not function by stimulating transcription is consistent with previous reports indicating that WHV does not have an enhancer within the WPRE (6, 8, 10, 26).

To determine the effects of the WPRE on GFP transcripts, RNA from cytosolic and nuclear fractions of transduced cells was analyzed by Northern blotting with a GFP-specific probe (Fig. 4B). Three RNA species can be detected by this probe; the two larger ones correspond to unspliced and spliced transcripts initiated at the 5′ HIV LTR, and the smaller one corresponds to RNAs originating from the internal CMV promoter. All GFP transcripts extracted from HR′CMV-GFP-W-transduced cells were larger than their counterparts extracted from HR′CMV-GFP-transduced cells due to the presence of the WPRE. The levels of expression of all three classes of WPRE-containing transcripts were six times higher than those in the respective controls in both nuclear and cytosolic fractions.

To determine if the WPRE altered the RNA half-life, populations transduced with either HR′CMV-GFP or HR′CMV-GFP-W were incubated with actinomycin D. Total RNA was extracted at different times and analyzed by Northern blotting (Fig. 4C). From the earliest time, all the WPRE-containing transcripts were much more abundant than the respective controls. However, the half-life of the WPRE-containing GFP RNA was increased less than twofold compared to that of the GFP RNA, as determined by phosphorimager analysis (Fig. 4D). Taken together, these results suggest that the WPRE acts very early during the biogenesis of RNA transcripts, perhaps by directing their efficient processing as soon as they emerge from the transcriptional machinery.