|Home | About | Journals | Submit | Contact Us | Français|
Conceived and designed the experiments: CP BY MK. Performed the experiments: CP BY. Analyzed the data: CP BY HK MK. Contributed reagents/materials/analysis tools: MJL HGK. Wrote the paper: CP BY MK.
RNA interference (RNAi) has been considered as an efficient therapeutic approach against the human immunodeficiency virus type 1 (HIV-1). However, to establish a durable inhibition of HIV-1, multiple effective short hairpin RNAs (shRNAs) need to be stably expressed to prevent the emergence of viral escape variants. In this study, we engineered a randomized lentiviral H1-promoter driven shRNA-library against the viral genome. Potent HIV-1 specific shRNAs were selected by ganciclovir treatment of cell lines stably expressing the cDNA of Herpes Simplex Virus thymidine kinase (HSV-TK) fused to HIV-1 nucleotide sequences. More than 50% of 200 selected shRNAs inhibited an HIV-1 based luciferase reporter assay by more than 70%. Stable expression of some of those shRNAs in an HIV-1 permissive HeLa cell line inhibited infection of wild-type HIV-1 by more than 90%. The combination of a randomized shRNA-library directed against HIV-1 with a live cell selection procedure yielded non-toxic and highly efficient HIV-1 specific inhibitory sequences that could serve as valuable candidates for gene therapy studies.
RNAi is a highly conserved process that protects the host from transposable elements and viruses , . In mammalian cells, RNAi can be initiated by the introduction of synthetic double stranded short interfering RNAs (siRNAs) of 19–21 basepairs, which are recognized by the cytoplasmic RNA induced silencing complex (RISC) and trigger the specific degradation of complementary mRNA, thereby circumventing an interferon response . Besides its role as an experimental technique for loss-of-function studies , , RNAi emerges as a powerful therapeutic approach towards human diseases as well as viral infections , . Different vectors have been developed both for transient and stable expression of the inhibitory RNA species , more commonly by RNA Polymerase III driven transcription of the inverted sense and antisense strands of shRNAs, connected by a linker-sequence (loop) , . In the case of HIV-1, siRNA- and shRNA-mediated post-transcriptional gene silencing offers a promising strategy to inhibit viral replication at different stages of viral infection by targeting viral as well as host genes , , , . However, the stable expression of antiviral shRNAs did not evoke the expected long-term inhibition of the virus due to the emergence of viral escape mutants , , . To circumvent the emergence of RNAi-resistant variants it seems to be mandatory to identify multiple potent inhibitory sequences targeting different or conserved regions of the viral genome , , , .
There are several publicly available design algorithms , , but reliable rules for efficient siRNA-sequences are still missing and designed siRNAs need to be assayed for their inhibitory potential and possible cytotoxic or other off-target effects . Furthermore, the costly and time-consuming synthesis and validation of designed siRNAs limit their use in large-scale studies. Consequently, several groups have developed techniques for the generation of randomly prepared shRNA libraries , , . All these approaches employ a similar principle for the construction of the library: cDNA is enzymatically converted to short inverted repeats separated by a linker-sequence and transcribed into shRNAs by appropriate expression systems. However, the feasibility and efficiency of these techniques has been disputed .
We have adopted and modified these technologies to generate the first randomized shRNA-library against the HIV-1 genome. A stringent cell-based selection procedure was introduced, which allowed for the identification of novel and potent HIV-1 specific shRNAs with little or no host cell toxicity. Moreover, a range of these virus-specific shRNAs were applied as synthetic siRNAs with a comparable efficacy. Additionally, the same sequences were stably expressed in permissive HeLa CD4 (HeLa P4) cell lines , where they effectively inhibited wild-type HIV-1 infections. Thus, our data demonstrate for the first time that a straightforward shRNA-library construction protocol in line with a stringent cellular selection procedure yield potent and readily convertible shRNAs against a viral genome.
The construction of the shRNA-library can be divided into 7 steps. Step 1: 5 µg viral DNA (pNL4.3; NIH AIDS Research and Reference Reagent Program) was partially digested with 1×10−3 U DNaseI in the presence of 1 mM MnCl2 (10′, 37°C). DNaseI was inactivated by vortexing (30″) and heat-inactivation (10′, 65°C). Cleaved 5′ and 3′ ends were repaired with 5 U T4 DNA Polymerase and 10 U Klenow Fragment in the presence of 200 µM dNTPs. DNA was separated on 20% polyacrylamid in TBE and fragments of the appropriate size (150–300 bp) were isolated by diffusion in 0.5 M NaCl; 1 mM EDTA (1h, 50°C, shaking) and purified by EtOH precipitation. Step 2: 4.5 µg of the DNA was ligated to 1.5 µg of the 3′loop. Subsequently the T4 ligase was heat-inactivated (10′, 65°C) and the DNA was digested with MmeI (1h, 37°C). The reaction was separated on 20% polyacrylamid in TBE and the product of 3′ loop and gene fragment was isolated and purified. The PAGE-purified 3′ loop (5′-GTTGAATCCCGGTTCAAGAGACCGGGATCCAAC) was purchased from QIAGEN (Hilden, Germany) and contained a MmeI restriction enzyme site. Step 3: 1 µg of the purified DNA fragment was ligated to 500 ng 5′loop. The reaction was separated on 20% polyacrylamid in TBE and the single-stranded circular DNA ring of 3′ loop, gene fragment and 5′ loop was isolated and purified. The PAGE-purified 5′loop (5′-GGAGAGACTCACTGGCCGTCGTTTTACCAGTGAAGATCTCCNN) was purchased from Qiagen. Step 4: The purified DNA ring was used for a RCA using 10 U Φ29 Polymerase and the 5′loop specific primers RCA1 (5′-ACTGGTAA) and RCA2 (5′-GCCGTCGT) in the presence of 0.01 U Pyrophosphatase and 200 µM dNTPs (o.n., 30°C). The RCA products were digested with BglII and MlyI (1h, 37°C) and separated on 20% polyacrylamid in TBE. The palindromic, inverted repeats were isolated and purified. Step 5: 10 ng of the BglII-MlyI – digested fragments were ligated to 100 ng of pENTR/shLib. The transformed bacterial colonies were pooled and the primary shRNA-library isolated. Step 6: The isolated plasmids were digested with BamHI and re-ligated (pENTR/shLib -BamHI). The resulting bacterial colonies were pooled as the secondary shRNA-library. Step 7: The shRNA-expression cassettes were subcloned into pLPac/shLib (LR Clonase™ II Enzyme Mix, Invitrogen, Karlsruhe, Germany).
pENTR/ siH1/mDD-stuffer  was linearized with BbsI and the resulting 5′ T-overhangs were blunt-ended using 5 U T4 DNA Polymerase in the presence of 200 µM dATP. After digestion with BglII, shRNA-sequences were ligated between the H1-Promoter and the polyT cassette.
2 inverted Xcm I restriction enzyme sites were created by annealing 36 nt single-stranded oligonucleotides with 5′BamHI and 3′HindIII overhangs (MWG, sequences on request) and ligated into pENTR/siH1/mDD-stuffer. Digestion of the pENTR/TA with XcmI allows for TA-cloning of the Taq-Polymerase amplified shRNA-expression cassettes.
The puromycin-gene was PCR-amplified from pPGK-puro (K. Rajewski, Harvard medical school; primer-sequences on request). The PCR-product was digested with SmaI and Bpu1102I and ligated into the GATEWAY® compatible lentiviral vector pL .
The HSV-TK gene was PCR-amplified from pRapid-flirt (A. Waisman, Medical Hospital Mainz, ) and ligated into pENTR/siH1/mDD-stuffer to yield pENTR/TK. Afterwards, three overlapping HIV-1 genomic segments were PCR-amplified and ligated between HSV-TK ORF and the polyA to yield the selection constructs pENTR/TK/HIV-Lib1, pENTR/TK/HIV-Lib2 and pENTR/TK/HIV-Lib3, respectively (primer-sequences on request). The selection cassettes were next subcloned into the lentiviral vector pL by means of LR Clonase™ II Enzyme Mix (Invitrogen, Karlsruhe, Germany).
The production of lentiviral particles was performed as described in the Manual ‘ViraPower™ Lentiviral Expression System’ (Invitrogen, Karlsruhe, Germany).
HeLa wt or HeLa P4 cells were transduced with viral particles at a (MOI: 0.1) in the presence of 6 µg/ml polybrene. 48h post transduction, cells were treated with 10 µg/ml blasticidin for 10 days. Surviving clones were expanded in 3 µg/ml blasticidin and analysed for stable integration of the transgene and expression of HSV-TK protein and the fusion-mRNAs.
The library selection cell lines HeLaTK/HIV-LIB1, HeLaTK/HIV-LIB2 and HeLaTK/HIV-LIB3 were transduced with lentiviral particles of the relevant shRNA-libraries (MOI: 0.1). 48h post transduction, selection cell lines were treated two days with 1µg/ml puromycin followed by ganciclovir treatment (2×10−6 M) for 5 days, replacing the selection media each day. Surviving cells were expanded clonally and the integrated shRNAs were recovered by nested PCR (primer sequences on request).
Equal amounts of proteins were probed either with HSV-TK antibody (W.C. Summers, Yale University), HIV-1 Integrase antibody (acris antibodies, Herford, Germany), Hsp70 antibody (BD Pharmingen, Heidelberg, Germany) or β-actin antibody (Sigma-Aldrich, Muenchen, Germany), and with horse-radish peroxidase-conjugated antibody to rabbit or mouse IgG (Sigma-Aldrich, Muenchen, Germany) as the primary and secondary antibodies in blocking buffer (10% milk in TBS), respectively.
Total RNA was isolated using TRIZOL® (Invitrogen, Karlsruhe, Germany). Enrichment of mRNA was achieved with the NucleoTrap mRNA kit (Macherey-Nagel, Düren, Germany). 1 µg mRNA was separated on denaturing 1.8% Agarose in MOPS buffer and transferred onto positively charged nylon membranes in 20×SSC. Membranes were pre-hybridized (4h, 65°C) and hybridized with 32P-labeled HSV-TK cDNA probe (overnight, 65°C). Washed membranes were exposed to Phosphoimager screens.
HEK 293 FT cells (Invitrogen, Karlsruhe, Germany) were co-transfected with either PCR-amplified shRNA-expression cassettes or shRNA-expressing vectors and pNL4.3Luc.R-E- (NIH AIDS Research and Reference Reagent Program). Luciferase expression was measured 48h post transfection using the Luciferase Assay Reporter Kit (Promega, Mannheim, Germany) and normalized to total protein.
The PCR amplified shRNA expression cassettes were cloned into pENTR/TA and sequenced via TouchDown-PCRs using the BigDye® Terminator v3.1 Cycle Sequencing Kit in combination with the dGTP BigDye® Terminator v3.1 Cycle Sequencing Kit (both applied biosystems) in a ratio of 31 and in the presence of 0.83 M Betaine.
IFN-β ELISAs (PBL Biomedical Laboratories, Piscataway, NJ) and p24 ELISAs (Aalto Bio Reagents Ltd., Dublin, Ireland) were performed as described by manufacturers instructions.
3×103 cells were seeded overnight per well of a 96 well plate. Adherent cells were quantified with crystal violet dye as previously described  for up to 5 days and normalized to day 1.
The construction of the randomized lentiviral shRNA-library against HIV-1 was based on a previously described technology with some modifications . The consecutive enzymatic steps to produce the shRNA-library based on the cDNA of HIV-1 are shown in Figure 1 and are described in detail in the Material and Method section. In order to circumvent the production of wild-type virus in selection cell lines, the HIV-1 genome (pNL4.3) was divided into three internal overlapping segments, comprising all genes of HIV-1. The three parts HIV-1.1 (2.5 kb), HIV-1.2 (2.5 kb) and HIV-1.3 (2.8 kb) comprised the NL4.3 nucleotides 855–3374, 3297–5796 and 6067–8447, respectively. In order to maximize the diversity of the shRNA-library, the HIV-1 segments were randomly fragmented with DNase I (Fig. 1a) which exhibits only minor if any sequence specificity , . Blunted fragments of about 150–300 bp were ligated to the 3′loop which connects the final sense and antisense strands and enables the restriction of the target sequence to 20 basepairs by the insertion of the Mme I recognition site at the start of its stem (Fig. 1b). The subsequent ligation of the 5′loop produces single-stranded DNA rings which were amplified by rolling circle amplification (RCA) to concatemers of inverted repeats that are flanked by Bgl II and Mly I restriction sites (Fig. 1c–d). The resulting palindromic, inverted repeats encoding shRNA molecules were inserted downstream of the RNA Polymerase III Promoter H1 derived from pSUPER  and upstream of the poly-T termination site of the GATEWAY® compatible linearized vector pENTR/shLib (Fig. 1e, primary library). Excess 3′ loops were removed by digestion with Bam HI. The following religation yielded expression-ready shRNA-vectors (Fig. 1e, secondary library). To generate the final HIV-1 shRNA-library, the shRNA-sequences were subcloned by LR recombination into the lentiviral vector pLP/EGFP/shLib (Fig. 1e, final library) which allows for the enrichment of transduced cell lines by puromycin selection and fluorescence monitoring by means of enhanced green fluorescence protein (EGFP).
The combined complexity of the primary libraries was about 1.4×104 (1750 colonies per kB of cDNA). Sequencing of 10 individual clones revealed 10 different HIV-1 specific shRNA-sequences consisting of sense and antisense strand as well as the loop sequence (not shown). Thus, our modifications allowed us to elevate the complexity of our shRNA-library as compared to the original technology, which utilized restriction enzymes for fragmentation of target DNA. Furthermore, we amplified the number of colonies by a factor of about 10 in order to maintain the complexity of the libraries throughout the subsequent cloning steps (Table 1). Hence, our data suggest that the randomization of a cDNA by DNaseI produces a broader diversity and complexity of a given library .
In order to maximize the stringency of the selection scheme, our method employed a fourfold combination of negative/positive selection measures, comprising ganciclovir (negative)  and blasticidin selection of stable selection cell lines (Fig. 1a) in line with selection of the shRNA library via puromycin, and EGFP fluorescence (positive), respectively. Several stably transduced HeLa selection clones (HeLaTK/HIV-LIB1, HeLaTK/HIV-LIB2 and HeLaTK/HIV-LIB3) were shown to express the HSV-TK protein (Fig. 2b) and to be functional by means of ganciclovir induced cell death, as visualised by crystal violet staining of adhering cells (Fig. 2c). Furthermore, we tested if our pooled lentiviral shRNA library is able to downregulate the HSV-TK expression in the corresponding selection cell lines. Upon transient transfection of the shRNA libraries into the selection cell line clones indicated by asterisks in Fig. 2c, only the complementary library is capable of downregulating the HSV-TK expression significantly as compared to the non-matching libraries (Fig. 2d).
Furthermore, the proper expression of the fusion-mRNAs was verified by Northern blot analysis of mRNA from these selection cell lines using a HSV-TK specific probe (Fig. 2e). Finally, we show that in the presence of puromycin and gancyclovir only the transduction of a relevant shRNA-library into the corresponding selection cell line enables the transduced cells to proliferate and form colonies, suggesting that our selection design represents a highly stringent procedure for the selection of shRNA-libraries (Fig. 2f).
The lentiviral shRNA-libraries were stably transduced into the corresponding selection cell lines and treated with ganciclovir in order to select for potent shRNAs. Transduced cells were then treated with puromycin and only surviving, EGFP positive colonies were picked for analysis. Next, shRNAs were recovered by nested PCR amplification of shRNA-expression cassettes  from the genomic DNA of surviving cell clones. PCR amplified shRNA cassettes were subjected to co-transfections into HEK 293 FT cells along with the HIV-1 specific luciferase reporter construct pNL4.3luc.R-E-. Of the 200 individual cassettes tested, more than 50% inhibited luciferase activity by over 70% as measured 48h post transfection and are thus considered as effective (Fig. 3a).
To enable a vector-based expression of those shRNA-sequences we subcloned a collection of highly potent PCR-amplified shRNA-expression cassettes into the GATEWAY® compatible TA-cloning vector pENTR/TA. 36 recovered shRNA-expression cassettes were sequenced and their inhibitory potential was reconfirmed in the HIV-1 specific luciferase assay described above (Fig. 3b). The relative nucleotide positions and HIV-1 target sequences of this selection are depicted in Figure 3b and Table 2. 24 individual shRNAs were directed against the gag (g) and gag-pol transcript (g/p), 2 shRNAs against vif (vif), one against rev (revE1) and one against env (env), respectively. Subsequent BLAST analysis against the HIV-sequence data base of the Los Alamos National Laboratory (http://www.hiv.lanl.gov) revealed that some shRNAs target well conserved HIV-1 sequences, whereas some target more variable regions. It is worth emphasizing that none of our shRNA-sequences has been described so far, pushing us to analyse the selected sequences using RNAi design algorithms . Interestingly, none of our sequences matched all parameters of those algorithms, suggesting that there are still unknown sequence characteristics for potent siRNAs and shRNAs.
These data suggest that our stringent fourfold selection protocol yielded a high proportion of potent inhibitory shRNAs that are unique in their sequence preferences as compared to those predicted by publicly available design algorithms.
In eukaryotic cells, shRNAs are transcribed in the nucleus and exported to the cytoplasm by exportin-5 where they are processed into functional siRNAs by Dicer . For that reason, we expected a similar inhibitory potential of exogenously administered siRNA derivatives compared to their equivalent vector based shRNAs. 11 effective shRNA sequences were randomly selected and synthesized as siRNAs to be co-transfected with the corresponding selection constructs into HEK 293 FT cells. Western blot analysis revealed that all of them were able to inhibit the TK expression from the HSV-TK/HIV fusion-mRNA constructs even at the lowest concentration used (Fig. 4a), demonstrating that our shRNA sequences could also be converted to efficient siRNAs.
It has been described that some short inhibitory sequences may exert a cytotoxic effect which was linked to the interferon (IFN) type I responses (IFN-α and IFN-β) , . In order to test our sequences in that regard, we performed IFN-β ELISAs on HeLa cells transfected with the siRNA as described previously . Even at the highest concentration used none of the above sequences induced detectable levels of IFN-β (Fig. 4b).
In addition, we investigated the inhibitory potential of the identified shRNA sequences with regard to wild-type HIV-1 infections. The corresponding 11 shRNAs were subcloned into the lentiviral vector pL and stably transduced into HIV-1 permissive HeLa P4 cells. Stable shRNA-expressing clones were infected with wt HIV-1 and the amount of the HIV-1 capsid protein p24 was determined in the culture supernatants by p24-antigen ELISA 42 h post infection. As shown in Fig. 4c, all cell clones expressing the verified shRNAs showed a significant reduction of p24 in the culture supernatant in comparison to cell clones expressing an HIV-1 irrelevant shRNA (scr). Finally, we investigated whether stable lentiviral expression of the shRNAs may have any detrimental cellular effects. As shown in Fig. 4d, none of the shRNAs impaired the division rate of the corresponding cell line compared to the controls as seen by quantitative crystal violet staining of adherent cells through four days of culture. Taken together, our data indicate that our selected shRNAs can be converted to potent siRNAs and exert a high inhibitory effect on wt HIV-1 infections, when stably expressed, both without any detectable cytotoxicity.
To develop an effective RNAi-based gene therapy against HIV-1, a combinatorial approach using several shRNAs expressed from different polymerase III promoters is supposed to be mandatory . To this end, we tested the human U6 promoter for expression of our shRNA sequences and compared its activity against H1 promoter driven shRNA expression. Western blot analysis of integrase gene expression after co-transfection of the shRNA vectors in combination with pNL4.3R-E-Luc revealed that there is no significant difference in the efficacy of the two promoters tested (Fig. 5a).
Additionally, we compared our newly identified shRNAs with previously reported anti-HIV shRNAs. For this purpose, we generated plasmids for U6 promoter driven shRNAs against gag/pol, nef, rev and vpu/enf. The shRNAs against gag/pol  as well as vpu/env  were directed at highly conserved regions of the HIV-1 genome, whereas shNef  has been designed according to the siRNA algorithm provided by Dharmacon and the shRev  has been chosen due to the accessibility of the target mRNA. As shown in Fig. 5b, shRNAs selected from our randomized shRNA-library show comparable inhibitory potential to the published shRNA sequences. To confirm this observation, we quantified the inhibitory potential of the shRNAs by performing luciferase assays on the lysates (Fig. 5c). Together, these results demonstrate that our highly potent HIV-1 specific shRNA sequences exert comparable efficacy as to the published sequences. They could also be expressed from different promoters, which suggest them as possible candidates for combinatorial gene-therapeutic approaches against HIV-1.
RNA interference by means of siRNAs and shRNAs has emerged as a promising therapeutic approach against HIV-1 . However, the design of highly efficient and specific siRNAs and shRNAs still remains critical. This is mainly due to the absence of a standard algorithm for the design of siRNAs , . We generated an HIV-directed shRNA library based on the proviral DNA of HIV-1 (NL4.3), in order to systematically identify efficient HIV-1 specific shRNAs. Our study is the first application of a randomly engineered shRNA-library technology on the human immunodeficiency virus (Fig. 1). The proviral DNA of HIV-1 was fragmented using DNase I (Fig. 1a) in order to yield diverse overlapping cDNA-fragments , . A stringent fourfold selection protocol based on positive selection via blasticidin, puromycin and EGFP fluorescence as well as negative selection by means of ganciclovir/HSV-TK in selection cell lines was applied and enabled us to enrich for cell lines expressing potent inhibitory shRNAs (Fig. 2). Besides their efficiency our sequences were devoid of any detectable cytotoxicity, which is probably also due to the stringent selection procedure (Fig. 4).
By comparing these sequences with published shRNAs, we could show that the selection of our randomized shRNA-library indeed identified efficient shRNAs with a comparable inhibitory potential (Fig. 5). Many published siRNAs or shRNAs were designed according to publicly available in silico algorithms. Surprisingly, none of our sequences satisfied all parameters of them. Certainly, more shRNA-sequences from randomized libraries have to be analysed to determine general sequence preferences. However, these functionally but not structurally selected shRNAs may contribute to a better understanding of the critical parameters for the design of potent siRNA and shRNA sequences. Moreover, this information could be combined with studies concerning the secondary structure and therefore the accessibility of the target-mRNA, which also influences the potential of RNAi. For this purpose Jakobsen et al. generated a sense and antisense oligonucleotide RNA-library to identify accessible target sites in the HIV-1 leader sequence, which is considerably structured, and identified five particular sites .
It has been shown that the activity of a lentiviral vector combining 2–3 anti-HIV shRNAs was markedly reduced with the increasing number of shRNA expression cassettes  based on recombination of repeat sequences in shRNA expression cassettes. We conclude, that this could be abrogated using different polymerase III promoters. Here, we could show that the shRNAs selected from our randomized library can be sufficiently expressed from different polymerase III promoters (Fig. 5). This suggests them as promising candidates for multiple shRNA approaches.
Although, retroviral shRNA gene therapy approaches showed only moderate success in the case of patients suffering from X-linked SCID , the durable potential of related vector systems is undeniable. Furthermore, siRNAs, e.g. against VEGF to combat AMD (age-related macula degeneration) , cytomegalovirus retinitis  and RSV infections (respiratory syncitical virus)  are currently being validated in clinical phase III trials, highlighting the significance of RNA-based therapeutics . Thus, we liked to examine if our selected shRNAs would be suitable for a gene therapeutic approach against HIV-1. Therefore, we subcloned a range of our selected shRNAs into the lentiviral vector pL and stably transduced them into HIV-1 permissive HeLa P4 cells at a low multiplicity of infection (MOI: 0.1) to obtain low copy integrants and reduce the risk of insertional mutagenesis. After HIV-1 infection, all HeLa P4 cell lines stably expressing a HIV-1 specific shRNA showed a considerable reduction in progeny virus release in comparison to scrambeled shRNA expressing or parental cells by inhibiting the production of progeny virus.
It has been shown that more than one shRNA against HIV-1 is necessary to circumvent the generation of viral escape mutants , . Others reported that an efficient T-cell specific delivery of three siRNAs directed against the co-receptor for HIV CCR5 and the viral vif and tat genes resulted in a significant suppression of HIV-1 infections in humanized mouse models . Moreover, as an optimal strategy to prevent the emergence of escape mutants, it has been recently proposed to combine a polymerase III driven shRNA targeting the shared exon of tat and rev in combination with a TAR decoy and a chimeric VA1-ribozyme targeting human CCR5 . Indeed, a pilot feasibility study using this gene therapeutic approach is already underway . We suggest that e.g. a combination of our shRNAs vif1 or vif2 with the gag/pol shRNAs could target all alternatively spliced viral mRNAs and thus might provide a strong defence against HIV escape mutants (Fig. 3, Table 2). Nevertheless, a putative application of shRNA approaches would greatly benefit from the development of efficient non-integrating but stable vectors. Additionally, the present expression systems could be optimized by the use of inducible or lineage-specific promoters that drive the shRNA expression , , , thereby reducing possible off-target or mutagenic effects.
Recently, three whole genome screens used siRNA libraries to identify host genes critical for HIV infections , . These screens have suggested hundreds of previously unrecognized host cell genes being involved in viral propagation and replication and hence should be considered in future approaches. Although a few HIV-specific shRNAs have been previously described, the increased numbers of potent and non-toxic shRNAs are probably needed to identify an optimal set of inhibitory sequences. This is especially true in light of the fact that shRNAs can mediate unintentional sequence-specific and –unspecific silencing of non-targeted genes. Our study clearly demonstrates that a stringently selected shRNA-library is well suited for the unbiased identification of novel potent and non-toxic shRNAs in addition to those predicted by commonly used public algorithms.
We thank K. Rajewski (Harvard medical school), A. Waisman (Medical Hospital Mainz) and W.C. Summers (Yale University) for pPGK-puro, pRapid-flirt and HSV-TK antibody, respectively.
Competing Interests: The authors have declared that no competing interests exist.
Funding: This work was supported by the Deutsche Forschungsgemeinschaft (DFG, www.dfg.de). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.