Inhibition of fibrillin 1 expression using U1 snRNA as a vehicle for the presentation of antisense targeting sequence
Inhibition of fibrillin 1 expression using U1 snRNA as a vehicle for the presentation of antisense targeting sequenceRobert A.Montgomery and Harry C.Dietz1,*
Department of Surgery and1Departments of Pediatrics, Medicine and Molecular Biology & Genetics, Johns Hopkins University School of Medicine,Baltimore, MD 21205,USA
Received October 2, 1996;Revised and Accepted January 14, 1997
This study examines whether the mimicking of selected properties of naturally occurring antisense RNAs in prokaryotes allows efficient inhibition of gene expression byin situ-expressed recombinant molecules in mammalian cells. Prokaryotic regulatory transcripts are expressed at high levels and have hairpin structures at their termini, features reminiscent of small nuclear RNAs (snRNAs) which are abundant and stable in the nucleus of all mammalian cells. A sequence complementary to fibrillin-1 (FBN1) mRNA, interrupted in its center by a hammerhead ribozyme, was substituted for the Sm protein binding site between the stem-loop structures of U1 snRNA. Expression of the chimeric antisense RNA resulted in dramatic inhibition of expression of fibrillin-1 message and protein in stably transfected cultured cells. The inhibitory effect was localized to the nucleus. The biological properties of U1 snRNA may provide a widely applicable vehicle for thein vivo delivery of antisense targeting sequences.
Antisense technologies for the targeted inhibition of gene expression could provide an effective strategy for the management of inherited disorders with dominant-negative or gain-of-function pathogenetic mechanisms, for the suppression of oncogenes, or for the control of a variety of infectious agents.In situ expression of antisense complementary RNA (cRNA) has many theoretical and practical advantages over methods based upon the exogenous administration of synthetic oligodeoxynucleotides which have a propensity for producing non-sequence-specified biological effects (1 ,2 ). When considered in the context ofin vivo therapeutic applications, foremost among these is the potential for enduring activity (3 ).
Many naturally occurring cRNAs with well documented regulatory functions have been described in prokaryotes (4 ,5 ). All are characterized by stable stem-loops that encompass or flank the targeting sequence, with the 3' loop generally having a high GC content. These structures are believed to enhance stability of the targeting molecules by conferring resistance to the activity of exonucleases (6 ,7 ). Besides the addition of stabilizing structural elements, other factors that may increase the effectiveness of cRNAs derived from recombinant expression vectors include the use of a potent promoter, the targeting of the initiating AUG codon and directly adjacent sequences, the inclusion of autocatalytic ribozyme structures that cleave the target in a site-specified manner and enrichment for the cRNA within the nuclear compartment (1 ,8 ,9 ).
An antisense expression construct was fashioned in this study that incorporates most of these potentially enhancing features. In that the structure described for naturally occurring antisense RNAs is highly similar to that for small nuclear RNAs (snRNAs), essential components of the spliceosome complex that are abundant and stable in the nucleus of mammalian cells, the U1 snRNA gene was selected as the framework for vector construction (10 ). Other theoretical attributes of this choice include the potent and constitutively active nature of the U1 snRNA promoter, the ability of the unusual trimethylguanosine 5' cap and Sm protein interactions to signal transport of U1 snRNA back into the nucleus (11 ,12 ), and the lack of polyadenylation of mature snRNAs, a factor which may favorably influence transcript trafficking and localization (9 ). Moreover, unlike other spliceosome components, U1 snRNA is widely dispersed in the nucleoplasm (13 ).
Dominant-negative forms of fibrillin-1 cause Marfan syndrome, an autosomal dominant systemic disorder of connective tissue (14 ). The antisense targeting `core' used in our expression construct contains sequences exactly complementary to coding nucleotides 1-15 and 17-30 of fibrillin-1 mRNA, separated by the 22 bp hammerhead ribozyme loop (Fig.1 ). The regions of complementarity are predicted to align this autocatalytic structure with the consensus sequence for ribozyme cleavage (5'-GUC-3') within the target message (15 ,16 ). A chimeric construct (pU1/FIB) was produced by substituting this core targeting sequence for the short Sm protein binding site between the two hairpin loops of U1 snRNA (10 ). We show that potent and specific inhibition of an abundant mRNA species can be achieved by expression of this chimeric cRNA and propose a novel gene therapy strategy for Marfan syndrome and other diseases with a dominant-negative pathogenetic mechanism.
A human osteosarcoma cell line that expresses fibrillin-1 (MG63) was transfected with either pU1/FIB or a control construct with the identical vector backbone and selectable markers (pZeoSVLacZ). Both transfected cell lines were subsequently brought through the identical selection process to generate multiple polyclonal and monoclonal colonies of stable transfectants. All lines were assayed by immunohistochemistry after an identical number of passages and at a highly comparable cell density. The amount of immunoreactive extracellular fibrillin-1 was strikingly reduced in multiple lines harboring pU1/FIB, while cells transfected with pZeoSVLacZ showed a pattern of protein deposition that was indistinguishable from untransfected controls (Fig.2 ). No differences were observed between the lines harboring pU1/FIB, pZeoSVLacZ or untransfected cells upon immunohistochemical analysis with a monoclonal antibody to fibronectin, suggesting that the chimeric cRNA did not globally impair pre-mRNA splicing or protein expression and metabolism. All cell lines showed similar growth parameters, suggesting some degree of specificity for the targeting process.
Antisense RNAs expressed by pU1/FIB might inhibit fibrillin-1 expression by multiple mechanisms. Ribozyme cleavage would remove the 5' cap from targeted transcripts, an event predicted to effect their rapid degradation. Alternatively, in the absence of ribozyme cleavage, involvement of the initiating AUG codon in duplex formation could impair translation. Northern blot analysis of multiple clonal colonies harboring pU1/FIB revealed a dramatic reduction in the level of FBN1 message (Fig.3 ). FBN1 transcripts were easily detected upon Northern analysis of mRNA extracted from untransfected cells and those transfected with pZeoSVLacZ. Only trace amounts of product corresponding to the targeted transcripts were observed when cell lines harboring pU1/FIB were assayed by RT-PCR (Fig.4 ). No apparent difference in amplicon abundance was observed when amplification was performed across the predicted ribozyme cleavage site versus a distant location, suggesting that targeted transcripts are rapidly degraded. Untransfected and pZeoSVLacZ-transfected cells showed comparable accumulation of fibrillin-1 message (Fig.4 ). All cell lines accumulated similar amounts of mature [beta]-actin mRNA (Figs3 and4 ), again attesting to the specificity of targeting. These data suggest that most, if not all of the inhibitory effect was achieved at the level of target message abundance.
The effectiveness of antisense inhibition of gene expression is greatly enhanced by colocalization of the targeting sequence and the target transcripts in the same cellular compartment (17 ,18 ). Based upon the interpretation of previous studies, antisense-induced mRNA degradation takes place within the nucleus (19 ). We selectively isolated RNA from the nuclear and cytoplasmic fractions of cells expressing pU1/FIB to determine the subcellular localization of chimeric transcripts. As shown in Figure5 , despite our modification of native U1 snRNA sequence, transcripts derived from pU1/FIB are greatly enriched in the nucleus.
Nuclear targeting of cytoplasmic U1 snRNA is influenced by hypermethylation of the 5' cap structure and binding of at least one common U snRNP protein, both influenced by the Sm protein binding site which has been altered in pU1/FIB (10 ). As seen in Figure1 , a sequence at the 5' end of the inserted antisense `core' (AAUUGG) is highly similar to the consensus site [PuA(U)nGPu, Pu = A or G] for Sm protein binding (24 ). It has also been shown that Sm binding sites are highly tolerant of mutations including internal nucleotide substitutions and deletions (24 ). Alternatively, Boelens and co-workers have recently demonstrated that mutant U1 snRNAs can be retained in the nucleus due to interaction with undefined saturable nuclear binding sites (25 ). Our data clearly indicate that the chimeric transcripts expressed from pU1/FIB are greatly enriched in the nuclear compartment. Indeed, unlike native U1 snRNA, no chimeric transcripts were seen in the cytoplasmic fractions. It is therefore possible that Sm protein binding, 5' cap hypermethylation, and hence nuclear targeting are maintained and perhaps enhanced in transcripts derived from pU1/FIB. Alternatively, the amount of cytoplasmic pU1/FIB cRNA may be below the level of detection for the methods employed, or the chimeric transcripts may never exit the nuclear compartment.
Further experimentation will be necessary to determine the generalized efficiency of CIAO-mediated antisense inhibition of gene expression. If other chimeric cRNAs are as effective as those derived from pU1/FIB, this strategy could be immediately applied to the inhibition of selected oncogenes or infectious agents. Its use for inherited dominant-negative or gain-of-function disorders will be complicated by the indiscriminate targeting of innocent and essential wild-type transcripts derived from the unaffected allele. Attempts at allele-specific targeting will likely sacrifice efficiency. In consideration of the extremely low levels of mutant protein needed to achieve phenotypic expression for many disorders, including Marfan syndrome, this may not be tolerable (26 ). Instead a replacement phase could be added to reconstitute wild-type protein. Based upon the degenerate nature of the genetic code, multiple nucleotide substitutions could be made across the target region of the coding sequence within an exogenously supplied expression construct encoding the protein of interest. This would prohibit duplex formation between the cRNA and transcripts derived from the replacement vector while maintaining the fidelity of the amino acid sequence. The consensus sequence for ribozyme cleavage could also be abolished. An additional benefit of this approach, currently being explored for Marfan syndrome, is that a single strategy would be applicable to all patients, despite the presence of allelic heterogeneity.
The pU1/FIB vector was constructed on the backbone of the pZeoSV (Invitrogen) prokaryotic/eukaryotic expression vector. The SV40 promoter, polyadenylation site and polylinker were excised from pZeoSV at theBamHI sites. A U1 snRNA expression cassette cloned into pUC13 (a gift from K. Beemon) (27 ) was excised withBamHI digestion and ligated into theBamHI sites of the modified pZeoSV. Two rounds of site-directed mutagenesis (28 ) were then performed to change 4 nt flanking the Sm protein binding site of U1 snRNA, creating uniqueEcoRI andSpeI restriction sites (pZeoU1EcoSpe). Complementary oligonucleotides that encode the antisense `core' sequence shown in Figure1 , including the 24 highly conserved nucleotides of hammerhead ribozymes (16 ), were synthesized and annealed at 40oC such that the remaining 5' and 3' overhangs were exactly complementary to the overhangs left byEcoRI andSpeI digestion. The sequences of the oligonucleotides were as follows: 5'-AATTGGCGATCTCCAGCACTGATGAGTCCGTGAGGA-CGAAACGCCCTCGACGCAT-3' and 5'-CTAGATGCGTCGAGGGCGTTTCGTCCTCACGGACTCATCAGTGCTGGAG- ATCGCC-3' (sense and antisense, respectively). The resulting duplex was ligated into theEcoRI andSpeI sites of pZeoU1EcoSpe to create pU1/FIB. All ligation junctions were sequenced to verify the identity and orientation of the insert. The sequence of the resulting chimeric RNA (pU1/FIB) was analyzed using a program that predicts RNA structure (Mike Zuker's RNA page: http://ibc.wustl.edu/~zuker/rna/). The boundaries of the targeting sequence utilized in the construct were selected to maximize preservation of the U1 snRNA stem-loops, the ribozyme secondary structure and the accessibility of the sequence complementary to the target message.
The parent (MG63) human osteosarcoma cell line (29 ) was a gift from F. Ramirez. Cells were grown to 60% confluency and transfected with either linearized pU1/FIB or an unmodified reporter gene construct (pZeoSVLacZ, Invitrogen), used as a control. The transfections were performed using the DOTAP liposome formulation (Boehringer Mannheim, 1 mg/ml) according to manufacturer's instructions. Cells were grown in MEM media (Cellgro) with 10% fetal calf serum and 250 µg/ml of zeocin (Invitrogen). Cell death was evident after 48 h. To remove dead and dying cells, cultures were rinsed daily with Hank's balanced salt solution (GIBCO BRL) and overlaid with fresh zeomycin-containing medium. After 14 days, widely spaced clonal colonies of 10-100 cells were observed and harvested using 8 × 8 mm cloning cylinders (Specialty Media). After treatment with trypsin, cells were transferred to single wells of 24-well tissue culture plates and were clonally expanded. Alternatively, multiple clonal colonies were allowed to coalesce in the original plates used for transfection to establish polyclonal colonies. Zeomycin selection (250 µg/ml) was maintained throughout all phases of experimentation.
Immunohistochemistry was performed as previously described (30 ) using either anti-fibrillin-1 mAb 69 (a gift from L. Y. Sakai) or an anti-fibronection mAb (Sigma).
The magnetic porous glass direct mRNA purification technique was used to isolate poly(A) RNA according to the manufacturer's instructions (CPG, Inc.). Electrophoresis of 3.5 µg of mRNA was performed under denaturing conditions, as previously described (31 ). The gel was exposed to 60 mJ UV light to facilitate transfer of large mRNA species. RNA was transferred to nylon membrane using the turboblotting system according to manufacturer's instructions (Schleicher & Schuell). The membrane was washed briefly in 2* SSC, crosslinked with 125 mJ UV light and prehybridized in Expresshyb (Clontech) for 30 min at 68oC. Human cDNA probes encoding fibrillin-1 (nt 370-1183) and [beta]-actin (Clontech) were labeled by random priming (32 ). The membrane was hybridized with the fibrillin-1 probe for 1 h at 68oC, washed first in 2* SSC with 0.05% SDS at room temperature and then in 0.1* SSC with 0.1% SDS at 50oC, and exposed to X-ray film for autoradiography. This process was then repeated using the identical filter without stripping and the [beta]-actin probe.
MG63 cells were grown to confluency and washed with cold phosphate-buffered saline (PBS). The cells were scraped into a small amount of residual PBS and pelleted. The pellet was resuspended in NP-40 buffer (0.14 M NaCl, 1.5 mM MgCl2, 10 mM Tris-HCl pH 8.6, 0.5% NP-40, 1 mM dithiothreitol, 500 U/ml placental RNase inhibitor), vortexed and allowed to stand on ice for 5 min. The solution was pelleted by centrifugation at 12 000g. The aqueous phase (cytoplasmic fraction) and pellet (nuclear fraction) were separated. An aliquot of 1 ml RNAzol (Tel-Test, Inc.) was added to each fraction, followed by 0.2 ml chloroform. The lysate was centrifuged and the aqueous phase containing the RNA was isopropanol precipitated. RNase free, DNase (Ambion, 0.5 U/µl) along with 10 mM MgCl2 was mixed with each resuspended sample. The mixture was incubated at 15oC for 30 min, then 25oC for 30 min. The DNase was destroyed by heating to 100oC for 5 min. The RNA was phenol/chloroform extracted.
Synthesis of cDNA and RT-PCR were performed using 1.5 µg of total RNA and reagents supplied in a kit (Perkin Elmer) according to manufacturer's instructions. Conditions for PCR were as follows: 95oC, 30 s; 58oC, 30 s; 72oC, 30 s. The primers used were designed to amplify selectively either pU1/FIB-derived transcripts (primers A and C) or endogenous U1 snRNA (primers B and C). The identical amounts of the same RNA preparations and the identical methods were used to amplify fibrillin-1 (primers 5'UTR-S and FB1B-AS, across the predicted cleavage site; primers FB2-S and FB3-AS, 3' to the cleavage site) and [beta]-actin (primers actin-S and actin-AS) messages. The sequences of the primers were as follows:
primer A (sense), 5'-AGGACGAAACGCCCTCGAC-3';
primer B (sense), 5'-GCTTATCCATTGCACTCCGG-3';
primer C (antisense), 5'-GAAAGCGCGAACGCAGTCC-3';
actin-S (sense), 5'-GCACTCTTCCAGCCTTCC-3';
actin-AS (antisense), 5'-GCGCTCAGGAGGAGCAAT-3';
5'UTR-S (sense), 5'-GCAAGAGGCGGCGGGAG-3';
FB1B-AS (antisense), 5'-AAGGTTTTCCATCCAGGC-3';
FB2-S (sense), 5'-ATGAATGGAGGTAGCTGC-3';
FB3-AS (antisense), 5'-TTACCATAGGAACAGAGCAC-3'.
Resulting amplicons were run in 1% agarose, 2% NuSieve (FMC) gels, stained with ethidium bromide, and photodocumented using the Eagle Eye II system (Stratagene). Negative images are shown.
We thank F. Ramirez, L. Y. Sakai, and K. Beemon for the provision of reagents and helpful discussions. Supported by NIH grant 2RO1AR/HL41135, The Smilow Foundation and The Dana and Albert Broccoli Center for Aortic Disease. R.A.M. is a scholar of the American College of Surgeons. H.C.D. is a Richard Starr Ross Research Scholar.
1 Neckers, L. and Whitesell, L. (1993) Antisense technology: biological utility and practical considerations. Am. J. Physiol.265, L1-12.MEDLINE Abstract
3 Izant, J.G. and Weintraub, H. (1984) Inhibition of thymidine kinase gene expression by anti-sense RNA: a molecular approach to genetic analysis. Cell36, 1007-1015.MEDLINE Abstract
4 Wagner, E.G.H. and Simons, R.W. (1994) Antisense RNA controls in bacteria, phages, and plasmids. Annu. Rev. Microbiol.48, 713-742.
6 Case, C.C. et al. (1989)The unusual stability of the IS10 anti-sense RNA is critical for its function and is determined by the structure of its stem-domain. EMBO J. 8, 4297-4305.MEDLINE Abstract
7 Hjalt, T.A. and Wagner, E.G. (1995) Bulged-out nucleotides protect an antisense RNA from RNase III cleavage. Nucleic Acids Res.23, 571-579.MEDLINE Abstract
8 Hirashima, A., Sawaki, S., Inokuchi, Y. and Inouye, M. (1986) Engineering of the mRNA-interfering complementary RNA immune system against viral infection. Proc. Natl. Acad. Sci. USA 83, 7726-7730. MEDLINE Abstract
9 Zhong, L., Batt, D.B. and Carmichael, G.G. (1994) Targeted nuclear antisense RNA mimics natural antisense-induced degradation of polyoma virus early RNA. Proc. Natl. Acad. Sci. USA91,4258-4262.
10 Guthrie, C. and Patterson, B. (1988) Spliceosomal snRNAs. Annu. Rev. Genet.22, 387-419.MEDLINE Abstract
11 Hamm, J., Darzynkiewicz, E., Tahara, S.M. and Mattaj, W.I. (1990) The trimethylguanosine cap structure of U1 snRNA is a component of a bipartite nuclear signal. Cell62, 569-577.MEDLINE Abstract
12 Fischer, U. et al. (1991) Diversity in the signals required for nuclear accumulation of U snRNA and variety in the pathways of nuclear transport. J. Cell Biol. 113, 705-714.MEDLINE Abstract
13 Carmo-Fonseca, M. et al. (1991) Mammalian nuclei contain foci which are highly enriched in components of the pre-mRNA splicing machinery. EMBO J. 10, 195-206.MEDLINE Abstract
14 Dietz, H.C. et al. (1991) Marfan syndrome caused by a recurrent de novo missense in the fibrillin gene. Nature352, 337-339.MEDLINE Abstract
15 Cech, T.R. and Bass, B. L. (1986) Biological catalysis by RNA. Annu. Rev. Biochem. 55, 599-629.MEDLINE Abstract
16 Haseloff, J. and Gerlach, W.L. (1988) Simple RNA enzyme with new and highly specific endoribonuclease activities. Nature334, 585-591.MEDLINE Abstract
17 Sullenger, B.A and Cech, T.R. (1993) Tethering ribozymes to a retroviral packaging signal for destruction of viral RNA. Science262, 1566-1569.MEDLINE Abstract
18 Michienzi, A., Prislei, S. and Bozzoni, I. (1996) U1 small nuclear RNA chimeric ribozymes with substrate specificity for the Rev pre-mRNA of human immmunodeficiency virus. Proc. Natl. Acad. Sci. USA93, 7219-7224.MEDLINE Abstract
19 Cornelissen, M. (1989) Nuclear and cytoplasmic sites for anti-sense control. Nucleic Acids Res.17, 7203-7209.MEDLINE Abstract
20 Hendry, P., McCall, M.J., Santiago, F.S. and Jennings, P.A. (1992) A ribozyme with DNA in the hybridising arms displays enhanced cleavage ability. Nucleic Acids Res.20, 5737-5741.MEDLINE Abstract
21 Deshler, J.O., Li, H., Rossi, J.J. and Castanotto, D. (1995) Ribozymes expressed within the loop of a natural antisense RNA from functional transcription terminators. Gene155, 35-43.MEDLINE Abstract
22 Bertrand, E., Pictet, R. and Grange, T. (1994) Can hammerhead ribozymes be efficient tools to inactivate gene function. Nucleic Acids Res. 22, 293-300.MEDLINE Abstract
23 De Young, M.B. et al. (1994) Functional characterization of ribozymes expressed using U1 and T7 vectors for the intracellular cleavage of ANF mRNA. Biochemistry33, 12127-12138.MEDLINE Abstract
24 Jones, M.H. and Guthrie, C. (1990) Unexpected flexibility in an evolutionarily conserved protein-RNA interaction: genetic analysis of the Sm binding site. EMBO J. 9, 2555-2561.MEDLINE Abstract
25 Boelens, W.C., Palacios, I. and Mattaj, I.W. (1995) Nuclear retention of RNA as a mechanism for localization. RNA1, 273-283.MEDLINE Abstract
26 Dietz, H.C. and Pyeritz, R.E. (1995) Mutations in the human gene for fibrillin-1 (FBN1) in Marfan syndrome and related disorders. Hum. Mol. Genet. 4, 1799-1809.MEDLINE Abstract
27 Zhuang, Y. and Weiner, A.M. (1986) A compensatory base change in U1 snRNA suppresses a 5' splice site mutation Cell 46, 827-835.MEDLINE Abstract
28 Deng, W.P. and Nickoloff, J.A. (1992) Site-directed mutagenesis of virtually any plasmid by eliminating a unique site. Anal. Biochem.200, 81-88.MEDLINE Abstract
29 Sakai, L.Y., Keene, D.R. and Engvall, E. (1986) Fibrillin, a new 350-kD glycoprotein, is a component of extracellular microfibrils. J. Cell Biol.103, 2499-2509.MEDLINE Abstract
30 Eldadah, Z.A. et al. (1995) Expression of a mutant human fibrillin allele upon a normal human or murine genetic background recapitulates a marfan cellular phenotype. J. Clin. Invest. 95, 874-880.MEDLINE Abstract
31 Lehrach, H., Diamond, D.,Wozney, J.M. and Boedtker, H. (1977) RNA molecular weight determinations by gel electrophoresis under denaturing conditions, a critical reexamination. Biochemistry16, 4743-4751.MEDLINE Abstract
32 Feinberg, A.P. and Vogelstein, B. (1983) A technique for radiolabeling DNA restriction endonuclease fragments to high specific activity. Anal. Biochem. 132, 6-13.MEDLINE Abstract
*To whom correspondence should be addressed
This page is maintained by OUP admin. Last updated Thu Mar 13 17:29:21 GMT 1997. Part of the OUP Journals World Wide Web service.Copyright Oxford University Press, 1996
