Mol Biotechnol (2010) 45:116–120 DOI 10.1007/s12033-010-9247-8
RESEARCH
shRNA Expression Constructs Designed Directly from siRNA Oligonucleotide Sequences Tuva Barøy • Kirsten Sørensen • Mona Mari Lindeberg Eirik Frengen
•
Published online: 30 January 2010 Ó Springer Science+Business Media, LLC 2010
Abstract Gene silencing by RNA interference (RNAi) is a widely used approach for target-specific knockdown of gene expression. Induction of RNAi in mammalian cells can be achieved by introduction of synthetic small interfering RNA (siRNA) resulting in transient knockdown, or alternatively by stable expression of short hairpin RNA (shRNA). Several algorithms for efficient siRNA design exist, but recent reports have suggested that these cannot be directly used to design efficient shRNAs. In this study, 25 siRNAs targeting independent sequences in five transcripts were used for the construction of shRNA cassettes. Both the siRNAs and shRNA constructs were transfected into HEK293T cells. Quantitative real-time PCR analysis revealed that 19 out of the 25 shRNA constructs reduced the average expression level to less than 30%. Our data support that sequences designed by siRNA algorithms efficiently reduce the expression of the target gene when converted into shRNA expression constructs.
Electronic supplementary material The online version of this article (doi:10.1007/s12033-010-9247-8) contains supplementary material, which is available to authorized users. T. Barøy K. Sørensen E. Frengen (&) Institute of Medical Genetics, Faculty of Medicine, University of Oslo, P.O. Box 1036, Blindern, 0315 Oslo, Norway e-mail:
[email protected] T. Barøy e-mail:
[email protected] K. Sørensen e-mail:
[email protected] T. Barøy M. M. Lindeberg Department of Medical Genetics, Ulleva˚l University Hospital, Oslo, Norway e-mail:
[email protected]
Keywords siRNA shRNA shRNA cassette Cloning Transfection
Introduction RNA interference (RNAi) is involved in post-transcriptional gene regulation in eukaryotic cells mediated either by degradation or translation arrest of the target mRNA (reviewed in [1]). Double-stranded RNA (dsRNA) molecules of more than 30 bp activate a general antiviral response mechanism when introduced into mammalian cells [2]. However, when small interfering RNA (siRNA) molecules of 20–30 bp are used, specific down-regulation of gene expression is obtained. The siRNAs trigger the RNAi response leading to degradation of mRNAs complementary to one of the siRNA strands (reviewed in [3]). To prevent off-target effects the siRNAs should ideally contain multiple mismatches to all nontarget mRNA sequences (reviewed in [4]). Functional experiments where wild-type function was restored in heterozygous tumor cells by specific RNAi-based down-regulation of mutant transcripts have demonstrated the specificity and efficiency provided by the siRNA approach [5]. The genomes of higher eukaryotes contain hundreds of microRNA (miRNA) genes whose transcripts are processed into shorter hairpin structures [6]. These hairpin RNAs are processed into mature miRNAs that inhibit translation of the target RNAs. When short dsRNAs are introduced into mammalian cells, the sequence identity to the target mRNA determines whether mRNA silencing is achieved via the siRNA or miRNA mechanism: Perfect sequence identity induces site-specific degradation of mRNA [7], whereas mismatches between the dsRNA and the target lead to translation arrest [8].
Mol Biotechnol (2010) 45:116–120
117
Synthetic siRNAs have been successfully applied to mammalian cell lines, primary cells, and embryonic stem cells, often resulting in more than 80% down-regulation of the target gene expression [9]. This approach gives a transient effect where mRNA levels in proliferating cells return to normal after 3–7 days. In order to obtain long-term suppression of target gene expression, alternative approaches have been developed utilizing expression vectors where the siRNAs usually are expressed as siRNA hairpins (shRNAs) [10]. Extensive research over the last years has revealed sequence specific characteristics important for siRNA functionality, such as 30–50% G/C content, high 50 -end stability of the sense strand, absence of internal repeats or palindromes [11, 12]. Even though many siRNA design algorithms have been described based on these observations [12–16], the efficacy of siRNAs can only be determined experimentally based on the inhibition of the target gene expression [17, 18]. Furthermore, recent reports have suggested that specific design algorithms are required for the design of efficient shRNAs [19, 20], conceivably because specific sequence characteristics are important for the processing of shRNAs to siRNAs in the cells. We have measured the knockdown of gene expression using 25 siRNAs targeting independent sequences in five transcripts, and converted all siRNAs into shRNA expression constructs. Nineteen out of 25 shRNA constructs gave more than 70% knockdown of gene expression supporting that a direct conversion of siRNAs to shRNA expression constructs is an efficient approach. In addition, our data indicate that the spacer sequence ATCAAGAGA can replace the commonly used spacer sequence TTCAAGAGA in order to avoid a premature termination signal when
converting siRNA target sequences ending with three Ts into shRNAs.
Fig. 1 Conversion of siRNA sequences to shRNAs. The siRNA to shRNA conversion, as shown here for the redesign of GGHsi1 into GGHsh1, was performed for all 25 siRNA sequences listed in the Supplementary Table 1. The forward oligo contains the 19-nt target sequence (bold), a 9-nt spacer sequence, a 19-nt sequence complementary to the target sequence (bold), and a termination signal (T5).
In addition, the forward oligo carries 50 - and 30 -end overhangs compatible with the end produced by the restriction enzymes BglII and BstXI, facilitating cloning into the vector pSiRPG, which was used to express the shRNAs as hairpin molecules. The reverse oligo is shown below the forward oligo
Materials and Methods Construction and Cloning of shRNA Inserts Four siRNA oligos targeting each of the following transcripts were purchased from Dharmacon (Dharmacon; CO, USA): NDRG1 [GenBank: NM_006096], GGH [GenBank: NM_003878], SQLE [GenBank: NM_003129], WWOX [GenBank: NM_016373], and ADIPOR1 [GenBank: NM_ 015999]. In addition, three siRNA oligos targeting the GGH transcript were purchased from Ambion (Applied Biosystems/Ambion, TX, USA). Two siRNAs targeting NDRG1 have previously been described [21]. All siRNA target sequences are listed in Supplemental Table 1. In order to allow expression of shRNAs containing sequences identical to the 25 siRNAs, the sequences were redesigned as shRNA cassettes for cloning into the vector pSiRPG. An example of the conversion of the siRNA GGHsi1 to the corresponding shRNA GGHsh1 is shown in Fig. 1. In short, the sense oligo contains the specific 19-nucleotide (nt) target sequence followed by a 9-nt spacer sequence (TTCAAGAGA), a 19-nt sequence complementary to the target and a stretch of five thymidines (T) for transcription termination. The sense oligos also contained the 50 -end and the 30 -end 4-bp overhangs required for cloning into the BglII and BstXI-sites in the vector [21]. Because the target sequences of siRNA GGHsi3, NDRG1si3, SQLEsi3, and WWOXsi1 ended with three Ts, these were converted to shRNAs using an alternative spacer ATCAAGGA in order to avoid a premature termination in the target-spacer
118
junction. The alternative spacer was also used in the construct ADIPOR1sh4. All shRNA oligos were cloned into the vector pSiRPG and plasmid DNA from colonies obtained in the cloning step was analyzed by sequencing using the primers pSiRPG-F (GGGGAACTTCCTGACT AGGG) and pSiRPG-R (GTCTCTCCCCCTTGAACCTC).
Mol Biotechnol (2010) 45:116–120
Hs00963625_m1). The quantification of cDNA was performed using the comparative CT method, where the amplification of the target and endogenous control (PMM1) was run in separate tubes. The amount of target, normalized to the PMM1 expression and relative to mock-transfected cells, was estimated.
Cell Culture and Transient Transfection Results and Discussion The human cell line HEK293T (Invitrogen, CA, USA) was maintained in RPMI 1640 (Invitrogen, CA, USA) containing 10% FBS (Sigma, MO, USA). Cells were plated 24 h prior to transfection in 6-well trays at a density of 4 9 105 cells/well. Transfection of pSiRPG-constructs was performed using a Lipofectamine 2000 (Invitrogen, CA, USA) to DNA ratio of 10:4 (ll to lg). In order to select for transfected cells, the cells were added puromycin (InvivoGen, CA, USA) at a final concentration of 2 lg/ml 2 days after transfection and incubated for 8 days. Transfection of siRNA oligos was performed according to the siRNA Transfection Protocol for HEK293T cells (Dharmacon; CO, USA). Briefly, the HEK293T cells were seeded out at a density of 2 9 105 cells/well in 12-well trays the day before transfection. The transfections were performed by adding to each well 2 ll of DharmaFECT 1 (Dharmacon; CO, USA) and siRNA to a final concentration of 100 nM. All transfection experiments were performed in duplicate. Analysis of Gene Expression by Quantitative Real-Time PCR The siRNA- or shRNA-transfected cells were harvested 2 and 10 days after the transfection, respectively, in 19 Nucleic Acid Purification Lysis Solution (Applied Biosystems, Foster City, CA, USA). Total RNA was isolated with the ABI PRISMÒ 6100 Nucleic Acid PrepStation (Applied Biosystems) using the Isolation of Total RNA from Cultured Cells kit (Applied Biosystems). Total RNA was reverse transcribed by random-primed cDNA synthesis using the High Capacity cDNA Archive Kit according to the manufacturer’s instructions (Applied Biosystems, CA, USA). Real-time PCR experiments were performed using an ABI Prism 7900 instrument (Applied Biosystems, CA, USA) at 50°C for 2 min, 95°C for 10 min followed by 40 cycles with 95°C for 15 s and 60°C for 1 min. All reagents were purchased from Applied Biosystems; TaqMan Universal PCR Master Mix and the TaqMan Gene Expression Assays: GGH (Assay ID Hs00608257_m1), SQLE (Assay ID Hs01123765_m1), NDRG1 (Assay ID Hs00608389_m1), ADIPOR1 (Assay ID Hs01114951_m1), WWOX (Assay ID WWOX-E78), and PMM1 (Assay ID
Comparing Knockdown Efficiencies of siRNAs and shRNAs with Identical Target Sequences In order to compare knockdown efficiencies of siRNAs and shRNAs with identical targets, sequences from 25 commercially available siRNAs (Supplemental Table 1) were converted into shRNAs (Fig. 1). All shRNA cassettes were cloned into the BglII/BstXI cloning sites downstream of the polymerase III H1 RNA promoter in pSiRPG [21], and the cloned inserts were verified by sequencing. The siRNAs and shRNAs were transfected into HEK293T cells, and the levels of target transcripts were measured by quantitative real-time PCR using mock-transfected cells as a reference. A summary of the knockdown efficiencies for all siRNAs and shRNAs is presented in Fig. 2. In total, the target gene expression level was reduced to less than 30% compared to the gene expression level in control cells for 19 out of 25 shRNAs. In 18 out of the 25 shRNA/siRNA-pairs, the siRNAs gave slightly better knockdown efficiencies than the shRNAs (Fig. 2). This can most probably be accounted for by reduced transfection efficiency alone, since transfection of shRNA constructs is usually less effective than transfection of siRNA oligos [22]. Only two shRNAs showed a better knockdown efficiency than the corresponding siRNA. The ADIPOR1 siRNA4 gave only marginally reduced transcript level compared to the control cells while the corresponding shRNA gave a significant reduction. We cannot exclude that the low efficiency of ADIPOR1 siRNA4 is caused by mismatch between the siRNA and the target transcript. In total, the results presented support that sequences selected by siRNA algorithms efficiently reduce the expression of the target gene when converted into shRNA expression constructs. Comparison of Efficiencies of shRNA Cassettes with Regular and Alternative Spacers The low knockdown efficiency obtained by SQLEsh1 (Fig. 2) could possibly be a result of premature termination at a stretch four T residues, which has been shown to be sufficient for termination in many human tRNA genes [23]. Because four Ts are present in the target-spacer junction in
Mol Biotechnol (2010) 45:116–120
119
Fig. 2 Efficient down-regulation of gene expression using siRNAs and shRNAs containing identical target sequences. HEK293T cells were transfected with siRNA oligos (dark grey) or pSiRPG-shRNA constructs (light grey) targeting the transcripts of NDRG1, GGH,
SQLE, WWOX, or ADIPOR1 as indicated. Normalized mean levels of the target transcript in two independent experiments each performed in duplicates were plotted with standard deviation. The transcript levels were normalized to the levels in mock-transfected cells
SQLEsh1, we made an additional construct, SQLEsh1-AT, using the alternative spacer ATCAAGGA. When transfected into HEK293T cells the efficiency of SQLEsh1-AT was markedly improved compared to the shRNA1 which contains the original spacer (Fig. 3), thus suggesting that the alternative spacer could be useful in order to improve the knockdown efficiency of some constructs that contain a stretch of four Ts. In parallel, we also made the constructs
NDRG1sh4-AT and NDRG1sh7-AT with the alternative spacer in order to compare the efficiency of the shRNA constructs with the two spacers. The resulting knockdown efficiencies were similar for the two constructs regardless of the spacers for both NDRG1 shRNAs (Fig. 3). In addition it should be noted that GGHsh3, SQLEsh3, and WWOXsh1, which were constructed using the alternative spacer, also showed efficient knockdown of the target transcript (Fig. 2). Commercially available siRNAs are usually designed by the most updated algorithms for siRNA selection. A fraction of these siRNAs contain three T residues at the 30 -end. Our results support that these sequences can directly be converted into efficient shRNA cassettes by using the alternative spacer ATCAAGAGA.
Conclusion
Fig. 3 Knockdown efficiencies using shRNA cassettes with alternative spacers. HEK293T cells were transfected with pSiRPG-shRNA constructs containing the commonly used TTCAAGAGA spacer (dark grey) or the alternative ATCAAGAGA spacer (light grey) targeting NDRG1 and SQLE as indicated. Normalized mean levels of the target transcript in two independent experiments each performed in duplicates were plotted with standard deviation. The transcript levels were normalized to the levels in mock-transfected cells
Our results support that a large fraction of sequences designed by modern siRNA algorithms efficiently reduces target gene expression when converted into shRNA expression constructs. When siRNAs containing three T residues at the 30 -end are converted into shRNA cassettes, the spacer ATCAAGAGA could be used to avoid a premature termination in the target-spacer junction. Our results suggest that a direct conversion of commercially available siRNAs into shRNA expression constructs represents an efficient approach to achieve stable RNAibased knockdown of target gene expression in human cells. Acknowledgments This study was supported by The Research Council of Norway through grants from the Functional Genomics Program (FUGE, Grant Number 151882), and from Helse Sør-Øst (Grant Number 2007060). TB, ML, and EF are supported by Ulleva˚l University Hospital Research Fund (VIRUUS).
120
References 1. Mello, C. C., & Conte, D., Jr. (2004). Revealing the world of RNA interference. Nature, 431, 338–342. 2. Elbashir, S. M., et al. (2001). Duplexes of 21-nucleotide RNAs mediate RNA interference in cultured mammalian cells. Nature, 411, 494–498. 3. Meister, G., & Tuschl, T. (2004). Mechanisms of gene silencing by double-stranded RNA. Nature, 431, 343–349. 4. Amarzguioui, M., & Rossi, J. J. (2008). Principles of Dicer substrate (D-siRNA) design and function. Methods in Molecular Biology, 442, 3–10. 5. Martinez, L. A., et al. (2002). Synthetic small inhibiting RNAs: Efficient tools to inactivate oncogenic mutations and restore p53 pathways. Proceedings of the National Academy of Sciences of the United States of America, 99, 14849–14854. 6. Lee, Y., et al. (2003). The nuclear RNase III Drosha initiates microRNA processing. Nature, 425, 415–419. 7. Caudy, A. A., et al. (2003). A micrococcal nuclease homologue in RNAi effector complexes. Nature, 425, 411–414. 8. Doench, J. G., Petersen, C. P., & Sharp, P. A. (2003). siRNAs can function as miRNAs. Genes and Development, 17, 438–442. 9. Krueger, U., et al. (2007). Insights into effective RNAi gained from large-scale siRNA validation screening. Oligonucleotides, 17, 237–250. 10. Bernards, R., Brummelkamp, T. R., & Beijersbergen, R. L. (2006). shRNA libraries and their use in cancer genetics. Nature Methods, 3, 701–706. 11. Khvorova, A., Reynolds, A., & Jayasena, S. D. (2003). Functional siRNAs and miRNAs exhibit strand bias. Cell, 115, 209–216. 12. Reynolds, A., et al. (2004). Rational siRNA design for RNA interference. Nature Biotechnology, 22, 326–330. 13. Amarzguioui, M., & Prydz, H. (2004). An algorithm for selection of functional siRNA sequences. Biochemical and Biophysical Research Communications, 316, 1050–1058.
Mol Biotechnol (2010) 45:116–120 14. Hsieh, A. C., et al. (2004). A library of siRNA duplexes targeting the phosphoinositide 3-kinase pathway: determinants of gene silencing for use in cell-based screens. Nucleic Acids Research, 32, 893–901. 15. Ui-Tei, K., et al. (2004). Guidelines for the selection of highly effective siRNA sequences for mammalian and chick RNA interference. Nucleic Acids Research, 32, 936–948. 16. Huesken, D., et al. (2005). Design of a genome-wide siRNA library using an artificial neural network. Nature Biotechnology, 23, 995–1001. 17. Hung, C. F., et al. (2006). A novel siRNA validation system for functional screening and identification of effective RNAi probes in mammalian cells. Biochemical and Biophysical Research Communications, 346, 707–720. 18. Birmingham, A., et al. (2007). A protocol for designing siRNAs with high functionality and specificity. Nature Protocols, 2, 2068–2078. 19. Li, L., Lin, X., Khvorova, A., Fesik, S. W., & Shen, Y. (2007). Defining the optimal parameters for hairpin-based knockdown constructs. RNA, 13, 1765–1774. 20. Taxman, D. J., et al. (2006). Criteria for effective design, construction, and gene knockdown by shRNA vectors. BMC Biotechnology, 6, 7. 21. Storvold, G. L., et al. (2007). A retroviral vector for siRNA expression in mammalian cells. Molecular Biotechnology, 35, 275–282. 22. Buchholz, F., Kittler, R., Slabicki, M., & Theis, M. (2006). Enzymatically prepared RNAi libraries. Nature Methods, 3, 696– 700. 23. Braglia, P., Percudani, R., & Dieci, G. (2005). Sequence context effects on oligo(dT) termination signal recognition by Saccharomyces cerevisiae RNA polymerase III. Journal of Biological Chemistry, 280, 19551–19562.
