Background
The tRNAs, U6 small nuclear (sn) RNA and adenovirus-virus-associated RNAs are normally expressed in cells at high levels by RNA polymerase III. These polymerase III promoters have been used in gene therapy applications to express a variety of inhibitory RNAs, including RNAi, aptamers, ribozymes, antisense RNAs, and decoy RNAs [
1‐
5]. Retroviral and lentiviral vectors have been the primary method of gene delivery to carry these inhibitor cassettes [
1,
4]. Importantly, high levels of expression and inhibitory activity have been demonstrated in several studies targeting HIV-1 sequences with RNA polymerase III-driven inhibitory RNAs [
4,
5].
While stable transduction of the transgene and vector sequences is necessary for gene expression, several studies reporting lack of gene function after transduction with retroviral vectors have identified various mechanisms limiting vector function. In several systems including our studies, the expression cassettes have been positioned in the vectors the in the reverse orientation to allow for the inclusion of important regulatory sequences (poly A signals, introns, self-splicing ribozymes, etc.) so that these signals will not affect vector genomic RNA production. However, deletions or aberrant processing have been identified in several of these studies due to the introduction of new cryptic regulatory sequences. For β-globin expression, which is normally developmentally tissue-specific regulated, generation of a retroviral vector with the β-globin gene in the reverse orientation lead to the identification of several RNA termination signals in the insert [
6]. Jonsson
et al also found fortuitous splice sites in the purine nucleoside phosphorylase (PNP) genomic construct when inserted in a retroviral vector in the reverse orientation [
7]. Additionally, multidrug resistance 1 (MDR1) expression in transduced bone marrow and spleen colonies resulted in both truncated and full-length mRNA, because the cDNA was found to contain cryptic splice donor and splice acceptor sites, even though it is expressed in the forward direction [
8]. Additionally, when a retroviral vector contains direct repeats, deletions can occur during reverse transcription with RNA template misalignment of the polymerase growing point and the first direct repeat [
7,
9]. It has been suggested that regulatory sequences may potentially emerge in retroviral vectors from sequences that are not naturally transcribed (i.e. antisense sequences or the U6 promoter) [
10].
Using retroviral vectors to introduce HIV-1 specific RNA aptamers into CD4+ cell lines, we found consensus splicing signals in the reverse orientation of the U6 promoter that lead to partial deletion of the U6 promoter after retroviral-mediated gene transfer. The splicing deletion lead to reduced promoter activity, and low expression of the aptamer in transduced cells was associated with a lack of HIV-1 inhibition. These studies could potentially be generalized to other vector systems that express small RNA inhibitors from this pol III promoter and should serve as a warning to other investigators in the design of their gene delivery vectors.
Discussion
Towards a stem cell gene therapy strategy for AIDS, we studied transduction of three different HIV-1-specific RNA aptamers and one control aptamer transcriptionally regulated by the U6 promoter in the antisense orientation within a retroviral vector. After shuttle packaging and transduction of CD4+ cell lines, these vectors were no longer able to inhibit HIV-1 replication. Instead, we found deletions within the U6 promoter in all 4 vectors after retroviral mediated gene transfer and consensus splicing signals in the U6 promoter in the reverse orientation.
The classical splicing consensus sequences and basal splicing machinery have been described for many years [
14‐
16]. The basal splicing machinery recognizes the classical splicing sequences and catalyzes the reaction removing the intron and joining the exons together. At the ends of the exon and introns are the well-described splice donor-splice acceptor sequences (see Figure
5) and within the intron is an internal branch point with a requisite adenosine that is necessary for the intramolecular lariat bond. Additionally, auxiliary elements within exon or introns are commonly required for efficient constitutive or alternative splicing. Finding similar classical splicing consensus sequences surrounding the deletion in the U6 promoter suggests that the splicing of this cryptic intron during transcription of the retroviral genomic RNA in the packaging cell lines caused the deletion.
Although the antisense U6 promoter closely matches the consensus splice donor/splice acceptor sites and has 5 potential internal branch sites, splicing of the antisense U6 promoter is inefficient. In the RT-PCR analysis of aptamer expression in cells transfected with the four different aptamer vectors, we observed only a small fraction of the shorter deleted product after only a single cycle of transcription and potential RNA splicing. In the genomic DNA from transduced CEMx174 cells after the retroviral vector was shuttle packaged, amplification of both the full-length and truncated products was observed, even after the vector underwent multiple rounds of viral replication (i.e., transcription, reverse transcription and integration). It appears that the mismatches in the splicing consensus sequences or that other auxiliary splicing signals may be necessary for efficient splicing of the antisense U6 promoter. Although splicing is detectable, the frequency is low, and this provides some optimism for vectors that have been produced with the U6 promoter in the antisense direction. However, our finding a similar deletion in four separate MMP-aptamer transduced populations reinforces that this splicing-induced deletion occurs reproducibly and that multiple cycles of retroviral packaging and transduction would only serve to increase the fraction of truncated product.
The molecular mechanism of transcriptional activation with U6 promoters has been thoroughly studied [
18]. RNA polymerase (pol) II promoters and pol III promoters for snRNA genes are very similar, even using common binding factors (reviewed in [
18]). The U6 pol III promoter contains 3 protein-binding domains: the proximal sequence element (PSE) and the TATA box are approximately 50 and 25 bps upstream of the transcriptional initiation site and the distal enhancer element is approximately 200 bps upstream of the transcriptional initiation site. For basal transcription, the PSE binds the snRNA activating protein complex (SNAPc) of five proteins (as they do in the pol II PSEs), while the TATA box binds to the initiation complex TFIIIB (reviewed in [
18]). TFIIIB, like the TFIIB for pol II promoters, facilitates interactions with the RNA polymerase and is required for all pol III promoters, even the separated tRNA-type promoters. For the human U6 promoter, TFIIIB consists of the TATA box binding protein (TBP), TFIIIB50 and TFIIIB90. Meanwhile, the distal enhancer element binds multiple factors (Oct-1, ZNF143 and ZNF76) that stimulate the formation of preinitiation complexes, with some proteins activating both pol II and pol III snRNA gene transcription. Correct nucleosome positioning between the enhancer and PSE during chromatin assembly also brings into close proximity the two binding domains and allows for cooperative interactions with the Oct-1 and SNAPc. A stable initiation complex recruits either RNA pol II or III and results in snRNA transcription.
The effects of the splicing deletion we observed in the aptamer vectors on cooperative binding and transcriptional activity may be complex because the 139 bps splicing deletion mutates potential transcription factor binding sites and changes the alignment of the remaining sites. In the proximal promoter, five of 19 bps from the PSE are deleted, which may limit SNAPc binding [
18]. In the distal U6 enhancer, two of the 4 total octamer consensus sequences (ATTTGCAT [
19]) are also deleted. We do not believe loss of these two non-functional [
19] domains contributes to the decreased promoter activity, since the deleted U6 promoter still contains the two functional octamer binding sequences: 1) the SPH motif, which binds to ZNF76 and ZNF143 [
20], and 2) the Oct-1 binding sites, which is involved in the cooperative binding of SNAPc to the U6 promoter that leads to increase recruitment of pol III and transcriptional initiation [
21]. Additionally, the splicing deletion may affect transcriptional activity because correct nucleosome positioning in the U6 promoter allows for the juxtaposed positioning of the distal enhancer with the proximal PSE necessary for their cooperative interactions [
22]. Thus, removing 134 bps from the U6 promoter will change the spatial alignment of the transcription factors with only 13 bps remain between the two DNA binding domains. In gel shift assays, cooperative binding of Oct-1 and SNAPc still occurs in recombinant promoters with only 18, 23, and 27 bps separation [
17]. Our Northern blots demonstrate that the deletion in the U6 promoter decreased, but did not completely eliminate, transcriptional activity. Some limited interaction must be possible between the SNAPc on the PSE and with Oct-1 in the enhancer.
Many factors in the configuration of retroviral vectors play a role in gene expression and activity. We initiated these studies when we observed a lack of inhibition of HIV-1 replication in cells that had been transduced with HIV-specific aptamer vectors generated using a shuttle packaging protocol. Interestingly, we found the splicing deletion in the U6 promoter that reduced transcriptional activity. However, not all transduced cells contained the splicing deletion, indicating that the transduced populations contained a mixture of wildtype and deleted promoters. In our earlier work [
11], we showed inhibition of HIV-1 with a gamma-retroviral vector expressing GFP and containing the U6-aptamer cassettes in the reverse orientation. These previous studies used a single round of packaging and pools of individual clones after limiting dilution plating for expression and challenge studies. In the present studies, we shuttle packaged the same vectors between amphotropic- and GaLV-pseudotyped Phoenix packaging cell lines to generate higher titer clones for subsequent
in vivo experiments. In the process, amplification of the vector during shuttle packaging also allowed for several cycles of RNA processing and the enrichment of aptamer cassettes with the splicing deletion in transduced cells. Additionally, the present studies used bulk populations of transduced cells (sorted for GFP expression) rather than pools of clones. The position-dependent effects of the integration site in the T cell clones may have resulted in higher levels of aptamer expression than we observed in bulk populations of transduced cells with heterogeneous integration sites. Consequentially, abundant aptamer expression was detected previously in the T cell clones by RNase protection assay [
11], while aptamer expression in the bulk transduced cells analyzed in this report was not detected by northern blot and RNase protection assay (data not shown). While the splicing deletion we observed in the U6 promoter reduced transcriptional activity, other factors in the vector design must also contribute to the overall lack of aptamer expression and viral inhibition that we observed in the transduced cells. We are currently examining other pol III promoters in the reverse and forward orientations to determine which vectors will permit stable gene transfer, the highest level of expression and the strongest inhibition. Interestingly, many gene transfer studies with the U6 promoter have used the complete U6 promoter, even though a minimal promoter with only the distal and proximal elements is active. Recently, the U6 promoter has also been used in lentiviral-based retroviral vectors to regulate siRNA inhibitors of either HIV-1 or of host cellular genes [
4,
23,
24]. These SIN-type lentiviral vectors, with a deletion in the 3' LTR to inactivate the HIV-1 promoter after reverse transcription and integration, have the U6 promoter-inhibitor cassette in the forward direction, and thus would not be subject to the deletion via the cryptic splice site described in this report.
Methods
Cell culture and transduction conditions
The CD4
+ cell line CEMx174 was cultured in RPMI-1640 (Sigma, St. Louis, MO) plus 20% fetal bovine serum (FBS, HyClone, Logan, UT), 10 mM HEPES (Cellgro/Mediatech, Fisher Scientific, Federal Way, WA), 50 U/ml penicillin and 50 μg/ml streptomycin (Cellgro/Mediatech), 2 mM L-glutamine (Cellgro/Mediatech), (R20 medium) at 37°C with 5% CO
2. CEMx174 cells were exposed to viral supernatants plus 8 μg/ml polybrene (Sigma) with a 20 minutes spin at 200 rpm and then overnight culture. After 24 hours, the cells were washed and expanded in R20 media. Phoenix (amphotropic) [
12], Phoenix (GaLV) packaging cell lines [
13] (both based on 293T cells), and U2OS cells were cultured in DMEM plus 10% FBS, 10 mM HEPES, 50 U/ml penicillin and 50 μg/ml streptomycin, 2 mM L-glutamine (D10 medium) at 37°C with 5% CO
2.
Generation and evaluation of high-titer producer cell lines
The MMP vectors were previously described [
11]. Plasmid DNAs of the four retroviral vectors was transfected into the Phoenix (GaLV) packaging cell line by calcium phosphate co-precipitation and washed after 24 hours. After 48 hours, culture supernatant was collected, passed through a 0.45-micron filter and used to transduce the Phoenix (amphotropic) packaging cell line. Aptamer-transduced Phoenix (amphotropic) cells were sorted for expression of GFP using a Becton Dickinson FACS Vantage. These heterogeneous populations of cells were used to generate supernatant for titering and for transducing Phoenix (GaLV) packaging cell lines. Aptamer-transduced Phoenix (GaLV) cells were sorted for expression of GFP. These heterogeneous populations of cells were used to generate supernatant for titering and for transducing CEMx174 cells. Dilutions of viral stocks were used to transduce U2OS cells and the percent GFP positive cells determined by flow cytometry. Stock titers ranged from 1.0 to 2.6 × 10
5 transduction units (TU)/ml (Table
1).
Viral inhibition assays
The HIV-1 strain NL4-3 was kindly provided by Ronald C. Desrosiers (NEPRC, HMS). Viral stocks were generated by infection of CEMx174 cells and harvesting cell-free supernatants on day 8–12 after infection. Virus stocks were analyzed for Gag production by ELISA (HIV p24, Coulter HIV-1 Core Antigen Assay; Coulter International Corp., Miami, FL) per the manufacturer's instructions and/or titered for TCID
50 values by limiting dilution assay as previously described [
25].
To challenge aptamer-transduced and control CEMx174 cells for viral replication, transduced cells were resuspended for 4 hours in HIV-1 at MOIs of 0.001 to 0.01 TCID50/cell, before being washed and cultured in 15 ml R20. Production of p24 Gag in the supernatant was assessed by ELISA (Beckman Coulter, Fullerton CA) in cell-free supernatants to quantify viral replication.
Molecular analysis
Genomic DNA was isolated from transduced cells using the QIAamp Blood Mini Kits (Qiagen, Valencia, CA). The inhibitor sequence from each transduced population was amplified using the forward 3'GFP (5'-GCT GGA GTT CGT GAC CGC-3') and the reverse MMP 3' UTR (5'-AGC TGG TGA TAT TGT TGA GTC AAA AC-3') primers as shown in Figure
1. Reaction mixtures containing 500 ng genomic DNA, 200 nM of each primer, 0.2 mM dNTPs in 50 μl of 1× PCR buffer (Invitrogen Corp., Carlsbad, California) were amplified with annealing temperature of 56°C, elongation temperature of 72°C, and denaturation temperature of 94°C for 35 cycles, using Platinum Taq DNA polymerase (Invitrogen). Samples were separated by agarose electrophoresis and stained with ethidium bromide. The image was captured using the Alpha Innoteck Corp. Imager program version 3.24 (Alpha Innoteck Corp., San Leandro, CA).
The PCR product for each band was gel purified (Qiagen) and cloned into pCR®II-blunt-TOPO® using Zero Blunt TOPO PCR Cloning Kit (Invitrogen). At least 18 individual clones for each PCR product were isolated and sequenced on a CEQ™8000 Genetic Analysis System (Beckman Coulter, Fullerton CA) using CEQ™DTCS-Quick Start Kit (Beckman Coulter) according to the manufacturer's instructions.
Sequence data and chromatogram data from the individual clones were aligned using ContigExpress in the Vector NTI Suite (Invitrogen). Consensus sequences were generated and compared to the predicted amplicon sequence using BEAUTY (BLAST Enhanced Alignment Utility) [
26], an enhanced version of the Basic Local Alignment Search Tool (BLAST) from the National Center for Biotechnology Information (NCBI), on the Baylor College of Medicine website [
29]. Multiple sequence alignments were generated using the Multiple Sequence Alignment Program [
27] at the Baylor College of Medicine website.
The Phoenix (amphotropic) packaging cell line was transfected with the different retroviral plasmids containing the HIV-1-specific RT aptamers using lipofectamine 2000 (Invitrogen) 48 hours prior to total RNA isolation with using Qiagen RNA easy (Qiagen). The inhibitor sequence from each transfected population was amplified using the same forward 3'GFP and the reverse MMP 3' UTR primers. The reaction mixtures contained 1 μg of total RNA and 200 nM of each primer, 0.2 mM dNTPs, 1.6 mM MgSO4, in 50 μl of 1× One-step Master mix (Invitrogen). The samples was reverse transcribed at 56°C for 30 mins with (RT+) or without (RT-) 25 units of the reverse transcriptase Superscript III (Invitrogen), and amplified with annealing temperature of 56°C, elongation temperature of 72°C, and denaturation temperature of 94°C for 40 cycles, using Platinum Taq DNA polymerase (Invitrogen). Samples were separated on 1.2% agarose gel and stained with ethidium bromide. The image was captured using the Alpha Innoteck Corp. Imager program version 3.24 (Alpha Innoteck Corp., San Leandro, CA).
Northern Blot
The pCRII TOPO plasmids corresponding to the wildtype (wt) and splicing mutant (mut) U6 promoter-70.15 aptamer expression cassettes were transiently transfected into 293T cells with lipofectamine 2000 (Invitrogen) per the manufacturer's instructions. As a control for the transfection efficiency, the constitutively active SEAP expression cassette was also transfected and soluble SEAP activity quantified. Two days after transfection, total RNA was isolated from the aptamer-expressing 293T cells by Trizol (Invitrogen). The amount of RNA in each sample was calculated from the optical density at 260 nm (OD260). RNA samples were subjected to electrophoresis in 8% denaturing polyacrylamide gel, transferred to a nylon membrane (Hybond N+, Amersham Pharmacia Biotech) and hybridized with a 32P-labeled RNA aptamer, tRNALys3 or U6 snRNA probe. After stringent washing of the membrane, images were captured and the bands were quantified by the phosphorimager (STORM 820; Molecular Dynamics).
Acknowledgements
These studies were supported by National Institute of Health grants AI 61797, CA 73473, RR00168, and by a developmental award from the Partners/Fenway/Shattuck Center for AIDS Research (CFAR), an NIH-funded program (AI 42851). We thank Drs. Richard C. Mulligan (Children's Hospital, HMS) for the MMP vector, Hans-Peter Kiem (Fred Hutchinson Cancer Research Center) for providing the Phoenix (GaLV) packaging cell line, Welkin Johnson (NEPRC, HMS) for the T2-SEAP cells, and Ronald Desrosiers (NEPRC, HMS) for the HIVNL4-3 viral stocks. We also thank Michelle Connole and the sequencing core facility for technical assistance, and Carolyn O'Toole and Noel Bane for assistance with manuscript preparation.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
PPJ participated in the design of and created the retroviral vectors. FEW carried out the viral production (packaging, titering and transductions) and viral challenge experiments (culture and immunoassays). GQ carried out the viral challenge experiments (culture and immunoassays), the molecular genetic studies and participated in the sequence alignments. XS performed the Northern blots and quantified expression. VRP participated in the design of the vectors and the subsequent study. RPJ participated in the design of the vectors and the subsequent study, and edited the manuscript. SEB participated in the design of the vectors and the subsequent study, coordinated the study, performed the sequence alignment, matched the consensus sequences and wrote the manuscript. All authors read and approved the final manuscript.