|Home | About | Journals | Submit | Contact Us | Français|
To make human immunodeficiency virus type 1 (HIV-1)-based vectors safer for use in the research and clinical setting, a significant modification to the HIV-1 genome has been the deletion of promoter and enhancer elements from the U3 region of the long terminal repeat (LTR). Vectors containing this deletion are thought to have no LTR-directed transcription and are called self-inactivating (SIN) lentivectors. Using four distinct approaches, we show that SIN lentivectors continue to have promoter activity near the 5′ LTR, which is responsible for the production of full-length vector transcripts. To verify that transcripts derived from the LTR in SIN lentivectors are competent for encapsidation and integration, we transduced a lentiviral packaging cell line with a SIN lentivector and then observed the production of viable vector particles containing full-length SIN lentivector genomes. We have also attempted to identify sequences in the SIN lentivector which are responsible for transcriptional activation at the 5′ LTR. Using different segments of the vector LTR and leader region in a promoter assay, we have determined that the residual promoter activity is contained entirely within the leader region and that, although this element is downstream of the transcription initiation site, it is capable of initiating transcription from the 5′ end of R in the LTR. Mutation of leader region binding sites for the transcriptional activators downstream binding factor 1 (DBF1) and SP1 reduces transcription from the SIN LTR by up to 80%. Knowledge of the potential for mobilization of HIV-1-derived SIN lentivectors will be important for the design of future gene therapy trials with such vectors.
The inability of simple retroviral vectors based on the genome of the Moloney murine leukemia virus (MMLV) to transduce nondividing cells has limited their usefulness for both research and clinical purposes, which require efficient gene transfer into quiescent cells types such as hematopoietic stem cells. The ability of lentiviral vectors to transduce such nondividing cells (3, 13, 34, 51) separates them from MMLV-based vectors, and this feature has fueled intensive development of vector systems based on human immunodeficiency virus type 1 (HIV-1) and other mammalian lentiviruses such as HIV-2 (8), simian immunodeficiency virus (37, 45), feline immunodeficiency virus (7, 38), and equine infectious anemia virus (36). A paramount concern in the development of these systems has been to create transfer vectors which maintain the ability to integrate into the genomes of nondividing cells but which have been modified in ways that maximize their biosafety.
Three primary biosafety concerns involving vectors based on the HIV-1 genome are (i) the risk of the vector packaging system giving rise to replication-competent lentiviruses through recombination between the transfer vector and packaging constructs (53), (ii) the risk of insertional oncogenesis if vectors with transcriptionally active long terminal repeats (LTRs) are used, and (iii) the risk of mobilization of HIV-1-based transfer vectors from transduced target cells by subsequent (or prior) infection with wild-type HIV-1 (1, 14). Perhaps the most important modification to HIV-1-based transfer vectors which addresses all of these biosafety issues has been the creation of self-inactivating (SIN) lentivectors by deleting many of the core promoter and enhancer elements from the U3 region of the vector (33). Due to the fact that the U3 deletion is copied to the 5′ and 3′ LTRs during reverse transcription, integrated SIN vectors contain only LTRs with U3 deleted.
Transcriptional activation at the HIV-1 LTR has been shown to depend on the presence of sequences both upstream and downstream of the traditional transcription initiation site at the U3-R junction (Fig. (Fig.1A).1A). Directly upstream of the transcription initiation site is a core promoter containing a classical TATA sequence, flanking E-boxes, and tandem binding sites for SP1. In the region 5′ of the core promoter, a large number of binding sites for T-lymphoid- and myeloid-specific transcription factors are present. Importantly, sequences overlapping and downstream of the transcription initiation site are also involved in activation of the HIV-1 promoter. An initiator (Inr) element overlapping the U3-R junction (sequence −6 to +30 where the first base of R is designated +1, a convention which will be used throughout this report) has been found to be required for transcription from the HIV-1 LTR (56). Additionally, two NF-κB sites in the R region are involved in activating the LTR in response to mitogenic stimuli (32). Further downstream in the leader region (LR), three binding sites for AP-1, as well as sites for NF-AT, DBF1, and SP1, have been shown to be located in DNase-hypersensitive regions in vivo, suggesting that some or all of these factors are bound to this region in actively transcribed integrated HIV-1 proviruses (10, 11). Elimination of these binding sites by deletion or mutation has been shown to decrease transcription from the HIV-1 LTR (11, 52). Additionally, a second Inr-like sequence consisting of a CT triplet located at position +227 has been found to be important for transcriptional competence of the HIV-1 LTR (29).
The standard SIN deletion removed 400 of the 455 bases comprising the U3 portion of the LTR. The deletion removes the TATA sequence and SP1 sites at the core promoter as well as many of the T-lymphoid- and myeloid-specific transcription factor binding sites present upstream of the core promoter (Fig. (Fig.1A).1A). When integrated into the genome of target cells, this SIN U3 proved to be severely disabled with respect to serving as a promoter, and because no transcripts derived from the LTR in SIN lentivectors were detected by Northern analysis (60), the LTR in these vectors is commonly thought to be completely transcriptionally inactive. We now demonstrate that the transcriptional inactivation of the LTR of SIN vectors is not complete and, moreover, that SIN lentivectors are capable of producing full-length genomic transcripts containing the Ψ sequence, making them competent for encapsidation and mobilization. We also show that mobilized SIN lentivectors are fully competent for a subsequent round of integration. Using several different methods, we were able to confirm the transcriptional activity remaining near the SIN LTR (henceforth referred to as sinLTR for clarity). In an attempt to attribute this transcriptional activity to specific cis-acting sequences in the LTR and LR of sinLTR vectors, we evaluated portions of the LTR and LR for transcriptional activity in a promoter screen. Additionally, we have analyzed transcription patterns from vectors containing specific mutations at sites previously identified to function as Inr and transcription factor binding sequences. Through these analyses, we found no independent transcription activating sequences in the sinLTR U3-U5 region and determined that residual promoter activity in SIN vectors is not dependent on the presence of the HIV-1 start site Inr nor the Inr-like sequence located at position +227 relative to the beginning of R. Rather, the elements responsible for promoting transcription from the sinLTR are located downstream of U5 in the LR and are largely dependent on the presence of DBF1 and SP1 binding sites in the +200 to +287 region.
HEK 293 cells (henceforth designated 293 cells) and 293T cells were grown in Dulbecco's modified Eagle's medium (DMEM) containing 10% (vol/vol) fetal bovine serum (FBS), penicillin (100 U/ml), streptomycin (100 μg/ml), and 2 mM l-glutamine. Jurkat T lymphoid cells, CEM T lymphoid cells, K562 erythromyeloid cells, and HL60 metamyeloid cells were all obtained from the American Type Culture Collection. Jurkat, CEM, and K562 cells were grown in RPMI 1640 medium containing 10% FBS, penicillin (100 U/ml), streptomycin (100 μg/ml), amphotericin B (250 ng/ml), and 2 mM l-glutamine. HL60 cells were grown in Iscove's modification of Dulbecco's medium with 20% FBS, penicillin (100 U/ml), streptomycin (100 μg/ml), amphotericin B (250 ng/ml), and 0.01 mM 2-mercaptoethanol. The lentiviral packaging cell line, WAN-1 (T. Kafri, unpublished data), was grown in DMEM with 10% tetracycline-free FBS, penicillin (100 U/ml), streptomycin (100 μg/ml), 2 mM l-glutamine, and 1 μg of doxycycline (Sigma Scientific, Inc., Brighton, Mass.)/ml until expression of HIV-1 packaging proteins and the vesicular stomatitis virus envelope glycoprotein (VSV-G) were induced by culture in medium free of doxycycline.
The HIV-1-derived vector plasmid pHR′-hCMV-eGFP (35), which contains wild-type LTRs and expresses the enhanced green fluorescent protein (eGFP), was the kind gift of Inder Verma (Salk Institute, San Diego, Calif.). The transfer vector expressed from pHR′-hCMV-eGFP is referred to as wtLTR-hCMV-eGFP in this report. The SIN HIV-1-derived vector plasmid pCCL-hCMV-eGFP (9) was the kind gift of Luigi Naldini (Cell Genesys, Foster City, Calif.). The transfer vector expressed from pCCL-hCMV-eGFP is referred to as sinLTR-hCMV-eGFP in this report. wtLTR and sinLTR vectors used in this work were derived from these two plasmids, respectively.
To create wtLTR and sinLTR vectors with a transgene directly 5′ of the internal promoter, the herpes simplex virus thymidine kinase (HSV-TK) gene was cloned into a unique ClaI site upstream of the human cytomegalovirus (hCMV) promoter in pHR′-hCMV-eGFP and pCCL-hCMV-eGFP. The HSV-TK gene was recovered from pG1NASVTK (31) by PCR with Pfu Turbo polymerase (Stratagene, La Jolla, Calif.) by using oligonucleotide primers which create ClaI sites at both ends of the gene: 5TKClaI (5′-AGAATCGATTATGGCTTCGTACCCCTG-3′) and 3TKClaI (5′-CCGAATCGATTCAGTTAGCCTCCCCCC-3′).
Lentiviral vectors with wtLTRs or sinLTRs containing no internal promoter were produced by removing the hCMV promoter from pHR′-hCMV-eGFP or pCCL-hCMV-eGFP, respectively, by digestion with BamHI and ClaI, followed by blunting with Pfu Turbo polymerase and ligation with T4 DNA ligase. These vectors are referred to as wtLTR-eGFP and sinLTR-eGFP. Vectors with sinLTRs containing mutations in the LTR and LR but lacking internal promoters were constructed in a similar fashion from the mutated vector constructs described below.
sinLTR lentivectors with additional regions of U3 deleted were prepared by PCR-mediated site-directed mutagenesis. The mutated sequences and their locations in the LTR and LR are depicted in Fig. 1B and C. All site-directed mutagenesis reactions were performed by using polyacrylamide gel electrophoresis-purified oligonucleotides with Pfu Turbo polymerase according to directions accompanying the QuikChange-XL kit (Stratagene) with the following thermocycler conditions: 17 cycles of 95°C for 1 min, 58°C for 1 min, and 68°C for 2 min per kb of plasmid length. The resulting PCR products were digested with DpnI (Invitrogen, Carlsbad, Calif.) to destroy the input template plasmid, and 2 μl of this mixture was used to transform competent DH5α. The vector containing the minimal 24-bp U3 sequence (the attL site) was generated by following the method of Iwakuma et al. (23) with oligonucleotide primers 5CCLattL (5′-GCTAATTCACTCCCAGGGTCTCTCTGGTTAG-3′) and 3CCLattL (5′-CTAACCAGAGAGACCCTGGGAGTGAATTAGC-3′), with pCCL-hCMV-eGFP as a template. Final constructs were verified by sequencing. To further compromise the sequence of the Inr at the transcription initiation site of the LTR in pCCL-hCMV-eGFP, three bases were changed from GGG to TTT by following the method of Rittner et al. (41) with oligonucleotides 5CCL+1TTT (5′-GTTTAGTGAACCGTTTTCTCTCTGGTTAG-3′) and 3CCL+1TTT (5′-CTAACCAGAGAGAAAACGGTTCACTAAAC-3′) (mutated bases are underlined).
Vectors containing mutated transcription factor binding sites in U5 and the LR were generated by PCR-mediated site-directed mutagenesis by following the method of Van Lint et al. (52). The SP1 mutant was generated with oligonucleotides 5SP1mut (5′-GCGCGCACGGCAAGATTCGAGGTGCGTCGACTGGTGAGTACG-3′) and 3SP1mut (5′-CGTACTCACCAGTCGACGCACCTCGAATCTTGCCGTGCGCGC-3′). Because the DBF1 mutation involves changing six A residues to G residues, we found that mutation primers making all six mutations at once could not be used successfully, perhaps due to the excessive length and C/G content of the primers. Consequently, the DBF1 mutant was generated in two steps by using oligonucleotides 5aDBF1mut (5′-GACTTGAAAGCGCCCGGGAAACCAGAG-3′) and 3aDBF1mut (5′-CTCTGGTTTCCCGGGCGCTTTCAAGTC-3′) for the first step, the product of which was used as the template for a second round of mutagenesis with the oligonucleotides 5bDBF1mut (5′-GAACAGGGACTTGCCCGCGCCCGGGAAC-3′) and 3bDBF1mut (5′-GTTTCCCGGGCGCGGGCAAGTCCCTGTTC-3′). The AP-1/AP-3L mutant was generated by using oligonucleotides 3AP3Lmut (5′-CCCTCAGACCCTTGTAGGCAGTGGTTTAAATCTCTAGCAGTG-3′) and 3AP3Lmut (5′-CACTGCTAGAGATTTAAACCACTGCCTACAAGGGTCTGAGGG-3′). The CTCTCTC sequence at positions +227 to +233, which is suspected to have Inr activity (29), was modified to GAGAGAG by using oligonucleotides 5CCL+227GA3 (5′-GAAACCAGAGGAGGAGAGAGGACGCAGGAC-3′) and 3CCL+227GA3 (5′-CTCCCTGCGTCCTCTCTCCTCCTCTGGTTTC-3′). The successful mutation of each of the above sites was verified by sequencing.
Vector supernatants were prepared by calcium phosphate-mediated cotransfection (calcium phosphate transfection kit; Invitrogen) of 293T cells as previously described (48) by using 10 μg of transfer vector plasmid, 10 μg of pRΔ8.9 packaging plasmid (59), and 2 μg of pMD.G(VSV-G) envelope plasmid (35). 293T cells were plated on poly-l-lysine-coated 10-cm-diameter plates in DMEM with 10% FBS and allowed to adhere for 6 h prior to transfection. Twelve hours after application of the DNA precipitate, the cells were rinsed three times with Dulbecco's phosphate-buffered saline (PBS) and then subjected to induction with 10 mM sodium butyrate (Sigma) in DMEM with 10% FBS. After 12 h, the cells were rinsed once with PBS and then refed with fresh medium. Vector supernatants were collected 24 or 48 h after application of fresh medium and then filtered through a 0.2-μm-pore-size syringe filter and stored at −80°C. Pseudo-sinLTR supernatants were produced in the same fashion; however, 10 μg of pHIT60, a plasmid expressing the MMLV packaging proteins Gag and Pol (48), was used in place of pRΔ8.9. To determine vector titers, 293 cells were plated at 2 × 105 cells per well in six-well plates and allowed to adhere for 6 h, at which time the media were removed and replaced with 1 ml of undiluted or diluted vector supernatant supplemented with 8 μg of Polybrene (Sigma)/ml. At the time of vector addition, the number of cells per well was recorded for use in calculating the titer. The vector supernatants remained on the cells overnight and were then replaced with fresh medium. Seventy-two hours after the addition of vector supernatant, the cells were harvested and assessed by flow cytometry for eGFP expression. Titers were calculated by using the following equation: [(% of eGFP-positive cells/100) × dilution factor × number of cells].
293 and WAN-1 cells were plated at 105 cells per well on six-well plates in DMEM with 10% FBS and permitted to adhere for 6 h prior to transduction. The medium was then aspirated and replaced with 1 ml of vector supernatant supplemented with 8 μg of Polybrene/ml. Transduction was verified 1 week after the application of the vector supernatant by assessment of eGFP expression from the vectors with a FACSCalibur flow cytometer and CellQuest software (Becton Dickinson, San Jose, Calif.). In vector mobilization experiments, transduced 293 and WAN-1 cells were passaged with Trypsin-EDTA every other day for 1 week to remove input vector particles prior to assessment of cell media over these cells for the presence of mobilized vector particles. After 1 week of passaging, the cells were replated at near confluence on poly-l-lysine-coated six-well plates in 1 ml of fresh medium. Once daily for five consecutive days, the 1 ml of medium over these cells was removed and replaced with fresh medium. The removed medium was filtered through a 0.2-μm-pore-size syringe filter and placed onto 293 cells which had been freshly plated. All five supernatant collections were supplemented with 8 μg of Polybrene/ml and added to the same batch of 293 cells. These cells were passaged in culture for 2 weeks and then assessed by flow cytometry for the presence of mobilized vector.
To verify that the eGFP-positive 293 cells which had been exposed to medium conditioned by WAN-1 cells transduced with the sinLTR-hCMV-eGFP vector had been transduced by full-length vector capable of genomic integration, cells were passaged for an additional 2 weeks (total of 1 month in culture posttransduction) and genomic DNA was then harvested from these cells with a DNeasy kit (QIAGEN, Valencia, Calif.). Genomic DNA was then analyzed by ALU-mediated PCR (4, 49) to verify the presence of integrated sinLTR vectors in the target cells. The first-round PCR was performed on 100 ng of genomic DNA with the following primers, the first of which anneals to the ALU sequence present at random locations in the human genome while the second anneals to the 3′ end of the U5 region of the LTR: ALU(s), 5′-TCCCAGCTACTGGGGAGGCTGAGG-3′; 3U5.1end, 5′-CTGCTAGAGATTTTCCACACTGAC-3′. The first-round PCR was performed on genomic DNA from mock-transduced 293 cells and genomic DNA from 293 cells exposed to media conditioned by WAN-1 cells transduced with the sinLTR-hCMV-eGFP vector (designated 293sin) by using the following thermocycler conditions: 30 cycles of 95°C for 50 s, 60°C for 50 s, and 72°C for 3 min. The second-round PCR was performed on 1/50th of the PCR product from the first round by using primers which amplify the region of the LTR between the 5′ end of U3 and the portion of U5 just 5′ to the 3U5.1end primer, 5U3 (5′-CTAATTCACTCCCAACGAAGACAAG-3′) and 3U5.2int (5′-CCACACTGACTAAAAGGGTCTGAG-3′). The second-round PCR was performed with the following thermocycler conditions: 30 cycles of 95°C for 50 s, 64°C for 50 s, and 72°C for 30 s. To control for the possibility that input genomic DNA in the first-round PCR was responsible for positive amplification in the second round, we included a control reaction in which the first-round PCR was performed without primers on 100 ng of 293sin genomic DNA. When 1/50th of this reaction mixture was used in the second-round PCR, the absence of a PCR product indicated that visible product in other lanes was amplified only from the product of the first-round PCR product.
For studies of vector transcript production, lymphocytic and myeloid cell lines were transduced as follows: 105 cells were plated in 100 μl of RPMI 1640 (no serum) in 48-well plates to which 100 μl of undiluted or diluted vector supernatant was added. Cells were incubated for 12 h, followed by the addition of 200 μl of media. Each day thereafter, the volume was doubled with fresh media and cells were successively passaged up to six-well plates. Transduced cells were cultured for a minimum of 1 week prior to any subsequent analysis. When transduction of cells with equivalent numbers of vector particles was desired, the titers of vector supernatants were determined by p24 immunoassay (Coulter Corp., Fullerton, Calif.).
Jurkat cells were transduced with the wtLTR-hCMV-eGFP and sinLTR-hCMV-eGFP vectors with or without the HSV-TK gene inserted upstream of the internal hCMV promoter. After passage in culture for 1 week or more, cells were assessed by flow cytometry for eGFP expression. Pools of cells containing each vector which were 10 to 30% eGFP positive were then plated in duplicate for treatment or nontreatment with ganciclovir (GCV; Roche Pharmaceuticals, Nutley, N.J.) in six-well plates at 5 × 105 cells per well. The treated cells were passaged in media containing 20 μg of GCV/ml for 2 weeks, with medium and GCV replacement every 3 days. Nontreated cells were passaged at the same frequency in medium lacking GCV. After 2 weeks, the treated cells were replated in fresh medium lacking GCV and all cells were passaged for an additional 2 weeks. All groups were then analyzed by flow cytometry for eGFP expression. In each group, dividing the percentage of cells remaining eGFP positive with GCV treatment by the percentage of eGFP-positive cells without GCV illustrates the GCV sensitivity of cells transduced with each vector.
RNA from stably transduced cells was harvested by using RNAqueous (Ambion, Inc., Austin, Tex.). The RNA was then treated with DNA-free (Ambion), with a slight modification of the manufacturer's directions in that the DNase I treatment was performed twice to completely eliminate residual genomic DNA. Five hundred nanograms of total DNase I-treated RNA served as the template for cDNA synthesis by using a reverse transcriptase PCR (RT-PCR) kit (Clontech, Palo Alto, Calif.) with a reverse transcription primer which anneals to the eGFP transgene present in each of the vectors, GFP.RTp (5′-CCGTCCTCCTTGAAGTCGATGCCC-3′). For every reaction, a control was set up which lacked the RT enzyme. Dilutions of cDNA were used as templates in PCRs with primers which detect a 314-bp fragment of eGFP 5′ to the site at which the GFP.RTp primer anneals, 5GFP (5′-ATGGTGAGCAAGGGCGAGGAGCTG-3′) and 3GFP (5′-GCCGTCGTCCTTGAAGAAGATGGTG-3′). One to eight microliters of the cDNA was used as the template for PCRs to identify the presence of the Ψ region by using primers which anneal to the R region and the 3′ end of the LR, 5R (5′-GAGCCTGGGAGCTCTCTGGCTAAC-3′) and 3u/sGag (5′-CTCTCTCCTTCTAGCCTCCGC-3′). Similar dilutions were also used to detect the presence of the HSV-TK gene in cDNA from cells transduced with HSV-TK containing vectors by using oligonucleotides 5TK-ID (5′-ACGATCTTGGTGGCGTGAAACTCC-3′) and 3TK-ID (5′-GTCATCGGCTCGGGTACGTAGACG-3′). For semiquantitative analysis of transcript expression, ethidium bromide-stained gels were visualized with an Eagle Eye charge-coupled device and densitometry was performed with EagleSight software (Stratagene).
293 cells were stably transduced with wtLTR-eGFP and sinLTR-eGFP vectors lacking an internal promoter. These cells were plated onto poly-l-lysine-coated six-well plates and transfected in duplicate with increasing amounts of the plasmid pSV2-tat72 (16) acquired from the AIDS Research and Reference Reagent Program, Division of AIDS, National Institute of Allergy and Infectious Diseases, National Institutes of Health (NIH; Rockville, Md.). Seventy-two hours after transfection, cells were harvested and the levels of eGFP expression within each group were analyzed by flow cytometry. At the same time, RNA was harvested from the transfected cells and subjected to DNase I digestion three times to remove input plasmid and expression from pSV2-tat72 was verified by RT-PCR with oligonucleotides 5tat72 (5′-ATGGAACCGGTCGACCCGCGTC-3′) and 3tat72 (5′-AGACAGAGAAACCTGGTGGGTC-3′). Ψ-Containing transcripts were also detected in this cDNA by using the previously described Ψ primers.
The promoterless pGL3-Basic plasmid containing the firefly luciferase gene was acquired from Promega (Madison, Wis.). Several fragments of the wtLTR and sinLTR regions, as well as portions of the SIN lentivector LR, were cloned into the KpnI-to-NheI sites in pGL3-Basic by using the listed primers (which contain the appropriate restriction enzyme sites, as designated in their names) to recover the fragments from pHR′hCMV-eGFP or pCCL-hCMV-eGFP where appropriate. All PCRs were performed with Pfu Turbo polymerase. The wtLTR and sinLTR regions (U3-U5) were recovered from the appropriate plasmid by using oligonucleotides 5U3KpnI (5′-AAAAGGTACCACTGGAAGGGCTAATTC-3′) and 3U5NheI (5′-ATGAGCTAGCACTGCTAGAGATTTTCC-3′). The LR (1 to 180) segment was recovered by using oligonucleotides 5LR1KpnI (5′-TGCTGGTACCGGGTCTCTCTGGTTAG-3′) and 3U5NheI (described above). The LR (181 to 335) segment was recovered by using 5LR181KpnI (5′-ATCTGGTACCGTGGCGCCCGAACAGG-3′) and 3LR335NheI (5′-TCGCGCTAGCCTCTCTCCTTCTAGCC-3′). The LR (200 to 287) segment was recovered by using 5LR200KpnI (5′-GAACGGTACCTTGAAAGCGAAAGGG-3′) and 3LR287NheI (5′-GCATGCTAGCGTCGCCGCCCCTCGC-3′). Segments containing mutations in either the DBF1 or SP1 site were recovered by using pCCL-hCMV-eGFP(ΔDBF1/ΔSP1) with those two mutations as a template. The LR (200 to 287) ΔDBF1/ΔSP1 segment was recovered by using 5LR200ΔDBF1KpnI (5′-GAACGGTACCTTGCCCGCG-3′) and 3LR287ΔSP1NheI (5′-GCATGCTAGCGTCGACGCACCTCG-3′). The LR (200 to 235) ΔDBF segment was recovered by using 5LR200ΔDBF1KpnI (described above) and 3LR235NheI (5′-GAGTGCTAGCTCGAGAGAGCTCTGG-3′). The LR (236 to 287) ΔSP1 segment was recovered by using 5LR236KpnI (5′-CTCTGGTACCGCAGGACTCGGCTTGC-3′) and 3LR287ΔSP1NheI (described above). The LR (253 to 287) ΔSP1 segment was recovered by using 5LR253KpnI (5′-CTCGGGTACCTGAAGCGCGCACGGC-3′) and 3LR287ΔSP1NheI (described above). The LR (288 to 335) segment was recovered by using 5LR288KpnI (5′-GGGCGGTACCTGGTGAGTACGCCAA-3′) and 3LR335NheI (described above).
293 cells were plated on poly-l-lysine-coated six-well plates at 2 × 105 cells per well. After 6 h, the medium was aspirated and replaced with fresh medium and the cells were then transfected in triplicate by calcium phosphate precipitation with 1,250 fmol of each of the pGL3 constructs into which various fragments of the LTR or LR had been cloned. To control for differing transfection levels, 14.7 fmol of the control plasmid pRL-CMV, which expresses Renilla luciferase from the hCMV promoter, was cotransfected with each of the pGL3 constructs. After 12 h of culture, the transfection mixture was aspirated and replaced with fresh media. The cells were cultured until 72 h after the initiation of transfection, at which time the cells were rinsed once with PBS and then lysed with passive lysis buffer (Promega) for assessment of firefly and Renilla luciferase expression by using the dual-luciferase assay kit (Promega) according to the manufacturer's instructions. Luciferase units in each sample were measured with a Lumat 9501 luminometer (Berthold, Wildbad, Germany).
Mutations were made to the sinLTR-eGFP vector plasmid as described above. Equivalent amounts of each vector, as determined by p24 immunoassay, were used to transduce Jurkat cells in triplicate. Cells were carried in culture for 1 week as described above and then analyzed for eGFP expression by flow cytometry. To assess transduction efficiency with each vector, genomic DNA was isolated from transduced cells and checked by PCR for the eGFP sequence by using the eGFP primers described above. Simultaneous amplification of glyceraldehyde-3-phosphate dehydrogenase (GAPDH) with primers included with the RT-PCR kit (Clontech) from each sample allowed us to verify that equal amounts of genomic DNA were added to each reaction mixture. PCR products were quantified on an ethidium bromide agarose gel by densitometry. The percentage of cells expressing eGFP from each vector was then divided by the GAPDH-corrected measurement of eGFP transduction into cells.
The prediction of the most thermodynamically favorable conformation of the Ψ loop was made with Mfold (61) by using a tool available online (http://www.bioinfo.rpi.edu/applications/mfold/old/rna/).
To determine whether SIN lentivectors produce full-length genomic transcripts capable of mobilization by wild-type HIV-1, we used the lentiviral packaging cell line WAN-1 (Kafri, unpublished) to represent an HIV-1-infected cell. WAN-1 cells are a clone of 293 cells which have been engineered to express the HIV-1 proteins Gag/Pol and Rev, each from separate expression cassettes under the control of Tet-off promoters (18). Additionally, this cell line expresses the VSV-G envelope and eGFP from a bidirectional Tet-off promoter. Expression of the HIV-1 packaging proteins, the VSV-G envelope, and eGFP is largely repressed when doxycycline is present in the media. Repression of these proteins is required for the expansion of WAN-1 cells because VSV-G and the protease component of Pol are cytotoxic when expressed constitutively (2, 15, 25, 26, 54). The coexpression of VSV-G and eGFP from a single expression cassette permits verification that the cells have been fully derepressed when cultured in doxycycline-free media by observation of eGFP expression by fluorescent microscopy or flow cytometry. After the WAN-1 cells had been successfully derepressed, they were transduced in parallel with 293 cells, using lentivector supernatants produced by transient transfection of 293T cells (Fig. (Fig.2A).2A). Vectors used included the wtLTR-hCMV-eGFP vector packaged with HIV-1 Gag/Pol and VSV-G, the sinLTR-hCMV-eGFP vector packaged with HIV-1 Gag/Pol and VSV-G, and a pseudo-sinLTR supernatant in which the sinLTR-hCMV-eGFP vector was packaged with MMLV Gag/Pol and VSV-G. Although MMLV Gag is capable of encapsidating unspliced HIV-1 vector genomes to a low degree (57), we have never produced infectious vector particles with this combination (unpublished data); thus, the pseudo-sinLTR supernatant serves as a control for the passive transfer of transfer vector plasmid or preformed eGFP within the vector supernatant. To reduce the amount of plasmid in the supernatant used to transduce the WAN-1 and 293 cells, we diluted the supernatants 1:10 in fresh D10 and used two hits of this dilution separated by 24 h to transduce the cells. Twelve hours after the second addition of vector supernatant, the cells were washed several times with PBS and refed with fresh media. During the week after transduction, the cells were passaged every other day with trypsin-EDTA to facilitate the removal of residual input vector particles. After 1 week, the transduced WAN-1 and 293 cells were analyzed by flow cytometry for eGFP expression (Fig. (Fig.2A).2A). As expected, even nontransduced WAN-1 cells were largely eGFP positive due to the aforementioned expression of eGFP in these cells when they are derepressed. This made it difficult to monitor the transduction of WAN-1 cells with eGFP-expressing vectors; however, the eGFP expression seen in 100% of the 293 cells which were transduced in parallel with the WAN-1 cells by the wtLTR and sinLTR vectors packaged with HIV-1 Gag/Pol demonstrate that these supernatants had high titers (Fig. (Fig.2A).2A). The lack of eGFP expression in 293 cells subjected to the pseudo-sinLTR supernatant verifies that no viable vector particles were present in this supernatant. It has been shown that lentivector supernatants prepared by transient transfection contain a large amount of residual transfer vector plasmid (44) which theoretically could transfect the WAN-1 and 293 cells exposed to the supernatants we used; however, we observed no spontaneous transfection of cells exposed to the pseudo-sinLTR supernatant (Fig. (Fig.2A2A).
The transduced WAN-1 and 293 cells were plated at near confluence, and for five consecutive days, medium conditioned by these cells was removed, filtered though a 0.2 μm-pore-size syringe filter, and placed onto a fresh batch of 293 cells. These 293 cells were passaged in culture for 2 weeks and then analyzed for eGFP expression. As seen in Fig. Fig.2A,2A, 293 cells exposed to media conditioned by mock-transduced WAN-1 cells showed no evidence of transfer of the eGFP gene from the WAN-1 cells. As expected, WAN-1 cells transduced with the wtLTR vector mobilized this vector, such that 9.38% of the target 293 cells were eGFP positive. The lack of eGFP expression in target 293 cells exposed to media conditioned by 293 cells transduced with the wtLTR vector demonstrates that there were no residual input vector particles present on these cells which could be carried to the final target 293 cells. We thus conclude that all eGFP-positive target 293 cells were transduced by vector genomes mobilized from the WAN-1 cells. In this experiment, which is representative of three similar experiments, we observed a relatively low amount of wtLTR vector mobilization. We believe this is partially attributable to the fact that WAN-1 cells do not express Tat, which is required for high-level expression from the wtLTR. Additionally, when WAN-1 cells are used to package high-titer vector supernatants, it is usually advantageous to grow them in the presence of the histone deacetylase inhibitor sodium butyrate (Kafri, unpublished). For these experiments, we did not add sodium butyrate to the media because we were concerned that this might artificially favor expression from the wtLTR and sinLTR.
When target 293 cells were exposed to media conditioned by WAN-1 cells which had been transduced by the sinLTR vector, we observed a significant level of eGFP positivity in the target 293 cells of 0.71%. Again, the absence of eGFP expression in target 293 cells exposed to media conditioned by sinLTR vector-transduced 293 cells shows that transfer of input vector particles is not responsible for the eGFP positivity of the target cells. Thus, we conclude that we have demonstrated mobilization of a sinLTR vector from a cell producing HIV-1 packaging proteins. To control for the possibility that residual transfer vector plasmid present in the supernatant transfected the WAN-1 cells and that expression from such spontaneously transfected plasmids could account for the apparent mobilization of the sinLTR vector from these cells, we also performed a control transduction of WAN-1 cells with the pseudo-sinLTR supernatant which did not contain HIV-1 vector particles. The lack of transfer of eGFP positivity to target 293 cells exposed to medium conditioned by these WAN-1 cells permits us to exclude that explanation for our results. Further characterization of our ability to mobilize sinLTR vectors from an HIV-1 packaging cell line is under way (Kafri, unpublished).
In Fig. Fig.2B,2B, we demonstrate, by ALU-mediated PCR analysis, that the target 293 cells exposed to medium conditioned by WAN-1 cells transduced by the sinLTR vector contained integrated sinLTR vector. The ALU element is a short interspersed nuclear element which is randomly located throughout the human genome and accounts for roughly 11% of its sequence (27). Although it has been demonstrated that SIN lentivectors favor integration into actively transcribed regions of the genome (46), a portion of these semirandomly integrated vectors will be located within a few kilobases of an ALU element. In the first round of our ALU-mediated PCR analysis, the sequence between ALU elements and the 5′ LTR of nearby integrated vectors was amplified. Then, a second-round PCR was performed on 1/50th of the first-round PCR product, using primers which amplified the region between the 5′ end of U3 and the 3′ end of U5 upstream of the annealing site for the U5 primer used in the first-round PCR. When this series of reactions was performed on mock-transduced 293 cells, the visible smear from the first-round PCR represents products primed solely by the ALU primer but no product was present in the second-round PCR. In contrast, when performed on the 293 cells into which the sinLTR vector had been mobilized (293sin), the second-round PCR yielded a specific product, indicating that the vector was integrated into the genome of these cells. Additionally, we note that the length of this product is specific to the sinLTR vector. Since the second-round primers anneal to the 5′ and 3′ ends of the LTR, the product would be 400 bp longer if amplified from a wtLTR vector. Consequently, we can confirm our conclusion that this experiment demonstrates mobilization of a sinLTR vector rather than an artifact produced by accidental introduction of the wtLTR vector at any stage of the experiment. To control for the possibility that genomic DNA put into the first-round PCR was amplified by the second-round primers, we performed a control in which the 293sin genomic DNA was subjected to the first-round PCR in the absence of primers and then to the second-round PCR with primers. Because this result was negative, we can conclude that the product seen in the ALU-mediated PCR performed on 293sin cell genomic DNA was amplified only from integrated vectors.
To demonstrate that a transgene placed downstream of the LTR in a sinLTR vector but upstream of the vector's internal promoter could be expressed, we constructed vectors containing the HSV-TK gene downstream of the LTR in both the wtLTR-hCMV-eGFP and sinLTR-hCMV-eGFP vectors (Fig. (Fig.3A).3A). These vectors were packaged with HIV-1 Gag/Pol and VSV-G and then used to transduce Jurkat T lymphoid cells.
Expression of transcripts from the LTR of both the wtLTR and sinLTR vectors was detected by using RT-PCR on DNase I-treated RNA isolated from the transduced Jurkat cells. As seen in Fig. Fig.3B,3B, we were able to specifically detect the HSV-TK sequence in RNA isolated from cells transduced with the wtLTR and sinLTR vectors bearing this gene upstream of the internal promoter. This assay was not performed quantitatively, and the brighter intensity of the HSV-TK product from the RNA isolated from cells transduced with the sinLTR-TK-hCMV-eGFP vector compared with RNA from cells transduced with the wtLTR-TK-hCMV-eGFP vector is accounted for by the fact that a larger percentage of cells were positive for the sinLTR vector (21.9%) than the wtLTR vector (12.8%). RT-PCR with primers detecting the eGFP transgene expressed from the internal promoter of each vector was used as a control to verify cDNA synthesis.
To provide further evidence for expression of HSV-TK from the LTR in both the wtLTR and sinLTR vectors, we determined the susceptibility of Jurkat cells transduced with wtLTR and sinLTR vectors which did or did not contain the HSV-TK gene upstream of the internal promoter to GCV-induced cell death. Pools of Jurkat cells which were 10 to 30% positive for each vector were plated in duplicate for exposure to media which did or did not contain 20 μg of GCV/ml. Cells were passaged for 2 weeks under these conditions, and then all cells were carried for an additional 2 weeks in media lacking GCV. When cells treated with GCV had recovered to normal growth patterns, they were analyzed for eGFP expression by flow cytometry. We calculated the ratio of eGFP positivity in populations which had been treated with GCV compared to those which had not been so treated (Fig. (Fig.3C).3C). Figure Figure3C3C represents the average of data from five separate experiments. As expected, Jurkat cells transduced with wtLTR and sinLTR vectors lacking the HSV-TK gene did not exhibit significant changes in the ratio of eGFP positivity with or without GCV. Cells transduced with wtLTR-TK-hCMV-eGFP demonstrated high-level sensitivity to GCV-induced cell death, and the ratio of eGFP positivity with or without GCV treatment was 0.25 ± 0.2 (mean ± standard deviation). The sensitivity of Jurkat cells transduced with wtLTR-TK-hCMV-eGFP compared to cells transduced with wtLTR-hCMV-eGFP was found by Student's t test to be highly significant (P < 0.0001). Jurkat cells transduced with sinLTR-TK-hCMV-eGFP exhibited a mild sensitivity to GCV with an eGFP positivity with or without GCV ratio of 0.76 ± 0.08, and this sensitivity was significant compared to Jurkat cells transduced with sinLTR-hCMV-eGFP (P = 0.004).
Another method we used to demonstrate expression from the LTR of sinLTR vectors was to construct wtLTR and sinLTR vectors completely lacking internal promoters. To accomplish this, the hCMV promoter was removed from the wtLTR-hCMV-eGFP and sinLTR-hCMV-eGFP vectors to create wtLTR-eGFP and sinLTR-eGFP (Fig. (Fig.4A).4A). These vectors were packaged with HIV-1 Gag/Pol and VSV-G and used to transduce Jurkat T lymphoid cells and K562 erythromyeloid cells. Due to the low-level expression from the sinLTR, we did not expect to be able to get a reliable titer for the sinLTR-eGFP vector by standard titer determining protocols, wherein limiting dilution transduction of 293 cells is performed, followed by analysis of eGFP expression by flow cytometry. Consequently, we determined the vector titers by p24 immunoassay. For unknown reasons, the titer of sinLTR-eGFP (132.5 ± 8.1 ng/ml) was significantly lower than the titer of wtLTR-eGFP (1401.7 ± 25.1 ng/ml). Nevertheless, wtLTR-eGFP was diluted with media to equal the p24 titer of sinLTR-eGFP, and serial dilution transductions with these two vectors were performed on Jurkat and K562 cells. Figure Figure4B4B demonstrates the patterns of eGFP expression in these cells 1 week after transduction with each vector at a concentration of 26.5 ng/ml. Significant levels of eGFP expression were observed in cells transduced with both vectors; however, the level of expression from sinLTR-eGFP was lower than that from wtLTR-eGFP, as indicated by the lower mean fluorescence intensities (MFIs) in cells transduced with sinLTR-eGFP (Fig. (Fig.4B).4B). These data indicate that the 5′ LTR in SIN lentivectors retains transcriptional activity, but it is significantly lower than the level of transcription from the wtLTR. A similar comparison of expression from the wtLTR and sinLTR, in which differences in transduction efficiency are accounted for, is described below (see Fig. Fig.8A8A).
We developed a semiquantitative RT-PCR assay with which we could determine the relative activities of the wtLTR and sinLTR in the context of vector constructs containing internal promoters. DNase I-treated RNA from stably transduced lymphoid (Jurkat and CEM) and myeloid (K562 and HL60) cells was used as a template for reverse transcription by using a primer which anneals to the eGFP transgene present in the vectors (Fig. (Fig.5A).5A). This produced a pool of strictly vector-derived cDNAs with a high likelihood that all would be complete to the 5′ end of the transcript. RT is prone to early termination; thus, minimizing the distance the enzyme must travel increases the likelihood it will reach the 5′ end of the RNA template. In this case, the 5′ ends of LTR-derived transcripts were within 2.6 kb of the RT primer. Transcripts from the LTR were detected by using primers which anneal to R (just downstream of the traditional transcriptional start site at the U3-R junction) and the LR, producing a product 330 bp in length (Fig. (Fig.5A).5A). We refer to these as Ψ-containing transcripts, and these can only be derived from the 5′ LTR. We also performed PCR with primers which detect a 314-bp fragment of eGFP. Transcripts containing this sequence can be derived from either the LTR or the internal promoter. Since the Ψ and eGFP reactions are of similar efficiency (Fig. (Fig.5B),5B), we semiquantitatively determined the percentage of transcripts originating from the LTR by densitometrically measuring the amount of Ψ and eGFP transcripts and dividing the former by the latter (Fig. (Fig.5B5B).
We found that in cells transduced with the wtLTR vector, transcripts from the LTR represented 6 to 37% of the vector-derived transcripts, with the highest percentage being seen in T cells. In cells transduced with the sinLTR vector, the percentage of transcripts derived from the LTR ranged from 1 to 5.5%, again with the highest LTR transcriptional activity in T cells. Comparison of these data allow us to conclude that the sinLTR retains roughly 15% of the transcriptional activity found in the wtLTR in the context of vectors containing internal hCMV promoters. We have confirmed that Ψ-containing transcripts can also be detected from sinLTR vectors containing other internal promoters (data not shown). Using this assay, transcripts originating from the LTR in a SIN lentivector were also identified in freshly transduced mouse bone marrow (data not shown), indicating that expression from the sinLTR is not restricted to cell lines or to human cells.
Having demonstrated by several different approaches that SIN vectors produced transcripts from the 5′ LTR, we sought to characterize the mechanism by which these full-length genomic transcripts are produced. As a first step toward identifying whether transcription from the sinLTR involves an RNA polymerase II (RNAP II) holoenzyme with the same constituents as that found in the holoenzyme which nucleates on the wtLTR during transcription, we tested the Tat-responsiveness of the sinLTR. To accomplish this, we transduced 293 cells with wtLTR-eGFP and sinLTR-eGFP and found that the sinLTR vector was capable of expressing in this cell type (data not shown). We then transfected the 293 cells transduced with the two vectors with increasing amounts of pSV2-tat72, a plasmid which expresses an active but truncated version of HIV-1 Tat (16). Seventy-two hours after transfections, the level of eGFP expression was determined by flow cytometry (Fig. (Fig.6A).6A). Increases in the MFI of eGFP expression demonstrated transcriptional activation at the LTR by Tat. We only saw such transactivation in cells transduced with the wtLTR vector. No change in eGFP intensity was detected in cells transduced with sinLTR-eGFP, even at the highest levels of transfected Tat (Fig. (Fig.6A).6A). To ensure that Tat was expressed in all cells, RT-PCR detections of Tat were performed on DNase I-treated RNA from all cells (Fig. (Fig.6B).6B). Thus, we conclude that transcription from the wtLTR, but not that from the sinLTR, is transactivated by Tat.
To facilitate identification of sequences in the sinLTR vector which are responsible for the residual promoter activity in or near the LTR, we adopted a promoter screening assay in which fragments of the LTR and LR were cloned into a plasmid, pGL3-Basic, which has a firefly luciferase (F-luc) gene situated downstream of the site into which the fragments were cloned. Equimolar amounts of each construct were cotransfected into 293 cells with pRL-CMV, a plasmid which serves as a transfection control by expressing Renilla luciferase (R-luc) from the hCMV promoter. Seventy-two hours after transfection, relative expressions of F-luc and R-luc were determined with a luminometer. The relative luciferase expression from each construct was normalized to the expression from pGL3-Basic, which was low, but not negligible (Fig. (Fig.7A).7A). When the wtLTR U3-U5 was assayed, it produced high levels of F-luc, greater even than a control plasmid from which F-luc was expressed by the simian virus 40 promoter (Fig. (Fig.7A).7A). However, when the sinLTR U3-U5 region was analyzed, it gave no detectable expression while a construct containing the 335-bp LR sequence between the 5′ end of R and the 3′ end of the Ψ region, designated LR (1 to 335), gave a significant level of expression (roughly 2% of the expression level from the wtLTR). Using random hexamer-primed RT-PCR followed by PCR with the Ψ primers identified in Fig. Fig.5,5, we verified that the LR (1 to 335) construct was producing transcripts originating from the 5′ end of this sequence or perhaps upstream of it (Fig. (Fig.7B).7B). The level of expression seen with the LR (1 to 335) construct was equaled by that from a construct containing the region +181 to +335 (Fig. (Fig.7A),7A), indicating that no LTR sequences in R or U5 are required for the level of transcription seen with the complete LR construct.
We then analyzed the effect of mutating transcription factor binding sites in this region which have been identified by other investigators to play a role in transcription from the HIV-1 LTR. Binding sites for DBF1 and SP1 are located within the LR (181 to 335) fragment, and these factors have been found in nuclease-sensitive sites in vivo, indicating that they may be bound in integrated HIV-1 proviruses and thus may play a role in transcription (10, 40, 52). Indeed, Van Lint et al. demonstrated that mutation of these sites significantly decreased transcription from an intact wtLTR in HIV-1 (52). When we analyzed the LR (181 to 335) fragment with either the DBF1 or the SP1 site mutated individually, the expression decreased slightly; however, combining the DBF1 and SP1 mutations led to a roughly 50% decrease in expression from the construct. This led us to conclude that DBF1 and SP1 work cooperatively to promote transcription at this site, but because expression was not completely eliminated, we believe there are other factors in this region continuing to promote transcription.
In an attempt to isolate these other sequences, we made further mutations and truncations within the LR (181 to 335) construct. Through these analyses, we localized the bulk of the transcriptional promotion to the 87-bp fragment LR (200 to 287). Other than the DBF1 and SP1 sites, the only other potential transcriptional activation site in this region which has been previously studied is a CT triplet, designated (CT)3, which has been identified as having Inr-like activity within HIV-1 proviruses (29). When this site was mutated to a GA triplet in HIV-1, transcription was reduced two- to threefold (29); however, when we added this mutation to the LR (200 to 287) construct, we did not see a significant change in the level of transcription (Fig. (Fig.7A).7A). Through the analysis of even shorter segments, we conclude that DBF1 is the only relevant site in the LR (200 to 235) segment, since the LR (200 to 235) ΔDBF1 construct did not express. Ultimately, we isolated a 34-bp fragment, which contained the SP1 site, and which expressed at a significant level even if the SP1 site was mutated, indicating that elimination of SP1 binding may be incomplete or there may be an unidentified binding site for another factor located in this segment (Fig. (Fig.7A7A).
We next analyzed the effect of DBF1 and SP1 binding site mutations in the context of a vector to determine whether the decrease in expression seen in the promoter screen was predictive of what would happen in the vector context. We constructed sinLTR vectors with the appropriate mutations in the LTR or LR by using PCR to mutagenize the specific sites (see Materials and Methods and Fig. 1B and C for mutant sequences). To simplify analysis of the effects of LR and LTR mutations on LTR-derived transcription, we used the assay described for Fig. Fig.4,4, in which vectors lacking internal promoters were used to demonstrate expression derived from the LTR. To accomplish this, we mutated sites within the LTR and LR in the vector sinLTR-eGFP, packaged the mutant vectors with HIV-1 Gag/Pol and VSV-G, and then transduced Jurkat cells with equal amounts of each vector as determined by p24 immunoassay. For both cell types, transduction with each vector was performed in triplicate. One week after transduction, cells were analyzed by flow cytometry to determine the percentage that were positive for eGFP expression. To correct our interpretation of the eGFP expression for possible differences in the transduction efficiency using each vector, we harvested genomic DNA from the transduced cells and the amount of transduced vector was determined by PCR amplification of the eGFP gene by using the eGFP primers described previously (Fig. (Fig.33 and and5).5). To control for the amount of genomic DNA put into the PCRs, we simultaneously amplified a GAPDH product and then quantitated the GFP and GAPDH products densitometrically (data not shown). Figure Figure8A8A shows the relative expression levels from each construct in Jurkat cells after correction for differences in transduction efficiency, all normalized to the expression level from the sinLTR-eGFP vector having no additional mutations.
In accord with our finding from the promoter assay that DBF1 and SP1 are primarily responsible for transcription from the sinLTR, the sinLTR-eGFP vector bearing mutations in the DBF1 and SP1 binding sites, designated sinLTR(ΔDS), exhibited a 79% decrease in expression compared to the sinLTR-eGFP vector without additional mutations. Because mutation of the DBF1 and SP1 sites did not completely eliminate expression in the promoter assay constructs (Fig. (Fig.7A),7A), we believe that either SP1 binding to the mutant site is not completely eliminated or that there are additional sequences near the SP1 site involved in transcriptional activation, and that this may be wholly or partially responsible for expression from the sinLTR(ΔDS) vector.
In an attempt to further decrease the expression from this mutant, we mutated additional sites which have been implicated in transcriptional activation in HIV-1. First, we mutated an AP-1/AP-3-like (henceforth AP1/3L) site located in the U5 region. Van Lint and colleagues have demonstrated this region to be a binding site for NF-AT which is involved to a low degree in transcriptional activation of the HIV-1 LTR (52). We did not see a significant change in the level of expression when the AP1/3L mutation was added to sinLTR(ΔDS) to create sinLTR(ΔADS) (Fig. (Fig.8A).8A). We further added the Δ(CT)3 mutation to this construct to produce sinLTR(ΔADCS) and found that this mutation did not alter the expression of eGFP (Fig. (Fig.8A),8A), confirming our finding in the promoter assay that the CT triplet does not play a role in transcriptional activation.
We also used this system to evaluate the possibility that the previously identified HIV-1 Inr ascribed to the sequence −6 to +30 relative to the beginning of R at +1 (56) may play a role in promoting transcription from the sinLTR. To alter the sequence at the beginning of this Inr, we deleted an additional 31 bp of U3, including the sequence −6 to −1 (see Materials and Methods). Iwakuma et al. demonstrated that the U3 region can tolerate deletion of all but the 5′ 24 bp which constitute the left attachment (attL) site, which is recognized by integrase (23). This construct, designated sinLTR(attL) (Fig. (Fig.1C),1C), expressed at nearly the same level as the standard sinLTR vector. Since other investigators have provided evidence that the sequence 5′ of the first base of R may not be important for Inr function in HIV-1 (41), we also constructed a vector in which the GGG sequence at the U3-R junction was mutated to TTT, which was shown to severely decrease transcription from the HIV-1 LTR, presumably by altering function of the Inr (41). This vector, designated sinLTR(T3), also expressed at nearly the same level as the standard sinLTR (Fig. (Fig.8A).8A). Finally we constructed a vector combining the attL with the T3 mutation which demonstrated only a mild defect in expression (Fig. (Fig.8A),8A), indicating that the start site Inr does not play a significant role in the residual promoter activity from the sinLTR.
Having demonstrated that mutation of the DBF1 and SP1 sites in the LR can significantly decrease transcription from the sinLTR, we next analyzed whether this mutation alters the titers of mutated vectors. We thus constructed versions of the sinLTR-hCMV-eGFP vector containing mutations of individual or multiple sites and then packaged them with HIV Gag/Pol and VSV-G and determined their titers in duplicate on 293 cells (see Materials and Methods). A vector containing the DBF1 and SP1 mutations had a titer which was only 20.7% of the titer of the nonmutated vector (Fig. (Fig.8B).8B). Analysis of vectors with the DBF1 and SP1 sites mutated individually permit us to attribute the decrease in titer solely to the SP1 mutation (Fig. (Fig.8B8B).
When creation of SIN lentiviral vectors was first reported (33, 60), the lack of transcripts from the LTR as assayed by Northern blot was provided as evidence of the complete abolition of LTR-mediated transcripts in these vectors. We have also observed that the level of transcription from the sinLTR is generally insufficient to be detectable by Northern analysis (data not shown). We were, however, concerned that this issue had not been investigated thoroughly enough, and that prompted us to develop more sensitive methods for detecting transcripts derived from the LTR of SIN vectors.
In this work, we have presented four independent lines of evidence that the sinLTR lentivectors retain promoter activity near the 5′ LTR which is responsible for the production of full-length vector genome transcripts competent for encapsidation and integration. First, we demonstrated that transduction of a lentiviral packaging cell line with a standard SIN lentivector results in the production of viable vector particles containing full-length SIN lentivector genomes (Fig. (Fig.2).2). We also demonstrated that an HSV-TK gene placed downstream of the sinLTR, but upstream of the vector's internal promoter, is produced in quantities high enough to render cells containing this vector partially susceptible to GCV-induced cell death (Fig. (Fig.3).3). Additionally, we demonstrated that SIN lentivectors containing no internal promoter express an eGFP transgene located directly downstream of the packaging signal and Rev-responsive element located at the 5′ end of the vector (Fig. (Fig.4).4). Finally, using RT-PCR, we demonstrated that transcripts initiated at the U3-R junction of the 5′ LTR in SIN lentivectors can be readily detected (Fig. (Fig.5),5), and using this assay, we conclude that the sinLTR/LR retains approximately 15% of the transcriptional activity of the wtLTR, depending on cell type.
After demonstrating the residual promoter activity in sinLTR vectors, we then sought to identify the elements remaining in the standard sinLTR vector constructs which are responsible for this activity. To provide some insight as to whether transcription initiation on the sinLTR was similar to that seen on the wtLTR, we rigorously analyzed the responsiveness of the sinLTR to the HIV-1 transactivator Tat. Tat transactivates the transcription of HIV-1 by binding to the TAR loop in nascent transcripts and directing the phosphorylation of the C-terminal domain of RNAP II by CDK9 which increases the processivity of the holoenzyme (17, 58). In 293 cells transduced with wtLTR-eGFP and sinLTR-eGFP vectors, we observed high-level and dose-dependent transactivation of the wtLTR by Tat (Fig. (Fig.6).6). In contrast, the level of transcription from the sinLTR was unchanged, even at the highest level of transfected Tat plasmid (Fig. (Fig.6).6). We thus conclude that the RNAP II holoenzyme that nucleates on the sinLTR is not Tat responsive.
The lack of Tat-responsiveness provides a clue that the process of transcriptional initiation on the sinLTR involves transcriptional activators which nucleate a RNAP II holoenzyme capable of only basal levels of transcription. Thus, we analyzed the potential for other elements in the LTR and LR to be involved in this basal promoter activity. An obvious candidate for involvement in promoting transcription from the sinLTR is the HIV-1 start site Inr, which has been identified to encompass the sequence −6 to +30 relative to the beginning of R at +1 (56). Inr elements are found near the transcriptional start sites of many TATA-containing and TATA-less promoters (reference 47 and references therein). In the latter, the Inr is usually thought to serve as a nucleation point for the RNAP II holoenzyme and thus dictates the site at which transcripts begin. It is notable that HIV-1 has an Inr at the traditional transcriptional initiation site which has been shown to be involved in initiating transcription from the LTR in vivo (56). There is evidence, however, that the HIV-1 start site Inr does not function efficiently in the absence of a TATA sequence at the normal −30 position relative to the transcriptional start site (56). Indeed, a very high level of conservation of the core promoter TATA sequence has been noted in proviral isolates from HIV-1-infected individuals, indicating its importance to HIV-1 viability (12). With the standard 400-bp U3 deletion used to create the sinLTR, the TATA sequence at this location is entirely removed (Fig. (Fig.1B).1B). The remaining U3 sequence in the sinLTR does not fortuitously contain a TATA sequence, nor does it contain a sequence which can be predicted to function as a noncanonical TATA sequence as determined by analysis with Matinspector (39) (Genomatix).
We have now presented substantial evidence that the HIV-1 Inr is not responsible for the residual promoter activity in the sinLTR. In our promoter screen, the sinLTR U3-R-U5 segment alone had no detectable promoter activity (Fig. (Fig.7A).7A). Additionally, lentivector constructs containing a deletion that removes the first six nucleotides of the HIV-1 Inr, a mutation that converts the first three nucleotides of R from GGG to TTT, and a mutant combining both of those modifications, were capable of expressing transgene at levels near that of the control sinLTR vector containing no mutations (Fig. (Fig.8A).8A). From this, we conclude that transcription from the LTR in SIN vectors does not require the HIV-1 Inr sequence at the U3-R junction to be intact.
Several studies have demonstrated that factors that bind downstream of the transcription initiation site play a key role in transcriptional activation from the LTR in integrated HIV-1 proviruses. This led us to question whether the LR binding sites may be involved in basal transcription from the sinLTR. We approached this issue by first demonstrating that the LR from the sinLTR vector is indeed capable of promoting transcription (Fig. (Fig.7A)7A) and that it does so in a manner which initiates transcription at or upstream of the first nucleotide of R (Fig. (Fig.7B).7B). After mutating the binding sites for DBF1 and SP1, we observed a significant decrease in the transcriptional activity from the LR, both as an isolated fragment in a promoter screen (Fig. (Fig.7A)7A) and as a mutation to the LR of a sinLTR vector (Fig. (Fig.8A).8A). The mutations we made were previously demonstrated by Van Lint et al. to significantly eliminate binding of the factors to the mutated sites (52). Although we could not eliminate all sources of transcriptional activation from the LR, we have demonstrated that mutation of DBF1 and SP1 sites in this region can reduce residual transcription from the sinLTR to roughly 20% of its original level. Unfortunately, mutation of the SP1 site is also accompanied by a significant (roughly 80%) drop in titer (Fig. (Fig.8B).8B). The SP1 site in the LR overlaps stem-loop 1 (SL1) of the Ψ region (Fig. (Fig.9).9). Although the SP1 site mutation we have used from Van Lint et al. (52) was selected to avoid disruption of SL1, it is possible that other thermodynamically favorable conformations of Ψ are disrupted by the mutation, leading to a decrease in transcript packaging and thus vector titer. Indeed, it has been demonstrated that maintenance of the SL1 stem-loop is crucial to the packaging of HIV-1 (5, 6, 19, 21, 55). Maintaining a high vector titer is of pivotal importance for the transduction of certain cell types, such as hematopoietic stem cells (20, 43, 50), so the decrease in titer we have seen with the sinLTR vector bearing the SP1 mutation will limit its utility. It remains to be determined whether other mutations to the SP1 site will be superior in terms of both eliminating factor binding and maintaining packageability and infectivity of the vector genomes.
In conclusion, our finding that HIV-1-derived vectors containing the SIN deletion in the U3 region of the LTR are capable of expressing full-length genomic transcripts raises the possibility that sinLTR vectors may be susceptible to mobilization by wild-type HIV-1, which will be a concern when these vectors enter into clinical usage. Due to the susceptibility of only certain cell types to HIV-1 infection, many potential users of HIV-1-derived vectors need never worry about mobilization (e.g., those using lentivectors to transduce nonhematopoietic tissues). Mobilization is, however, a possibility in the case of lentivector-transduced progenitor or mature hematopoietic cells susceptible to HIV-1 infection, and this should be borne in mind until a sinLTR vector with a fully transcriptionally silent LTR can be developed. The lack of Tat-responsiveness of the sinLTR is fortuitous, since even in their present configuration, sinLTR vectors will be at an extreme disadvantage for mobilization by HIV-1. If a cell transduced with a sinLTR vector were to become infected with HIV-1, the virus will transcribe its own genome at very high levels in the presence of Tat, whereas the sinLTR vector would only be capable of expression at a basal level. This will result in HIV-1 genomic transcripts far outnumbering vector genomic transcripts, making the vector transcripts unlikely to be packaged.
Another issue to be considered on the basis of our data is that the basal transcription from the sinLTR may have some significance for other lentivector applications such as the development of transgenic animals by using lentiviral vectors carrying expression cassettes with cell type-specific internal promoters (30). It is possible that low-level expression from the sinLTR may be adequate for the production of some transgene in a non-cell-type-specific manner. Such studies may benefit from a decrease in the basal transcription from the sinLTR achieved by mutating the LR DBF1 and SP1 sites.
This work was supported by two National Research Service awards (AI-07078-18 and CA-09569-13) to A.C.L. and by grants from the National Cancer Institute, NIH (1 P01 CA59318), and the Kenneth T. and Eileen L. Norris Foundation to D.B.K. . D.B.K. is the recipient of a Distinguished Clinical Scientist Award from the Doris Duke Charitable Foundation. T.K. was supported by a grant from the NIH (RO1 DK58702-01) and by the National Hemophilia Career Development Award.
We thank the AIDS Research and Reference Reagent Program for providing the pSV2-tat plasmid.