The effect of structure in a long target RNA on ribozyme cleavage efficiency
The effect of structure in a long target RNA on ribozyme cleavage efficiencyThomas B. Campbell1,2,*, Cheryl K. McDonald1 and Moira Hagen3
1Department of Medicine and 2Department of Biochemistry and Molecular Genetics, University of Colorado Health Sciences Center, Denver, CO 80262, USA and 3Department of Chemistry and Biochemistry, University of Colorado, Boulder, CO 80309, USA
Received August 25, 1997;Revised and Accepted October 24, 1997
ABSTRACT
Inhibition of gene expression by catalytic RNA (ribozymes) requires that ribozymes efficiently cleave specific sites within large target RNAs. However, the cleavage of long target RNAs by ribozymes is much less efficient than cleavage of short oligonucleotide substrates because of higher order structure in the long target RNA. To further study the effects of long target RNA structure on ribozyme cleavage efficiency, we determined the accessibility of seven hammerhead ribozyme cleavage sites in a target RNA that contained human immunodeficiency virus type 1 (HIV-1) vif-vpr. The base pairing-availability of individual nucleotides at each cleavage site was then assessed by chemical modification mapping. The ability of hammerhead ribozymes to cleave the long target RNA was most strongly correlated with the availability of nucleotides near the cleavage site for base pairing with the ribozyme. Moreover, the accessibility of the seven hammerhead ribozyme cleavage sites in the long target RNA varied by up to 400-fold but was directly determined by the availability of cleavage sites for base pairing with the ribozyme. It is therefore unlikely that steric interference affected hammerhead ribozyme cleavage. Chemical modification mapping of cleavage site structure may therefore provide a means to identify efficient hammerhead ribozyme cleavage sites in long target RNAs.
The hammerhead ribozyme is a relatively small (~34 nt) RNA that has a conserved central core that is required for catalytic activity surrounded by stems I, II and III (1 -3 ). In the three-dimensional structure of the hammerhead ribozyme, stems I, II and III are A-form helices arranged in a Y-shape with stems I and II in close proximity (Fig. 1 ) (4 ,5 ). Substrate cleavage by the hammerhead ribozyme occurs most efficiently 3' to the nucleotide sequence NUH (where N is any nucleotide and H is any nucleotide but G), although not all NUH sites are cleaved equally well (6 ). To maintain high specificity when the hammerhead ribozyme is used as a gene inactivating agent, stems I and III are each formed by five or six specific base pair interactions between the ribozyme and the substrate (7 ,8 ). Thus, the substrate recognition sequence N5UHN6 allows efficient and specific cleavage by the hammerhead ribozyme.
The use of hammerhead ribozymes to inhibit gene expression requires that ribozymes efficiently cleave specific sites in target RNA molecules (9 ). However, secondary structure in short oligonucleotide substrates results in an increased Km and inhibition of assembly of the ribozyme substrate complex (10 ). Likewise, ribozyme cleavage of a long target RNA at sites with known secondary structure is much less efficient than cleavage of short oligonucleotide substrates (11 ) because of decreases in both the rate of the association step (increased Km) and the chemical step (decreased kcat; 12 -16 ). That long target RNA structure affects ribozyme cleavage is further supported by the finding that cleavage of target RNAs by protein ribonucleases predicts efficient hammerhead ribozyme cleavage (16 -18 ). Although the structure of long target RNAs predicted by energy minimization algorithms sometimes correlates with ribozyme cleavage efficiency (19 ,20 ), this method has not proven to be a reliable predictor of ribozyme cleavage efficiency (14 ,16 ,21 ,22 ) and has not provided insight on how structure at individual target RNA nucleotides affects efficient cleavage. Thus, previous studies have demonstrated that long target RNA structure affects hammerhead ribozyme cleavage but they have not provided a detailed understanding of the determinants of efficient long target RNA cleavage.
Higher order structure of target RNA molecules could interfere with ribozyme association by at least two mechanisms. First, intramolecular base pairing within the target RNA could preclude base pairing between the ribozyme and the cleavage site. Second, the tertiary structure of large target RNAs could inhibit ribozyme association by steric hindrance. To date, the effects of intramolecular base pairs in long target RNAs on hammerhead ribozyme cleavage have not been determined and the role of steric hinderance in the inhibition of ribozyme cleavage is unknown. If the ability of hammerhead ribozymes to inhibit gene expression is to be optimized, it is necessary to better understand the determinants of efficient cleavage of long target RNAs. Toward this goal, the present study was undertaken to determine the effects of structure in a long target RNA on hammerhead ribozyme cleavage. First, the accessibility of sites in a long target RNA to hammerhead ribozyme cleavage was determined. Second, the availability of individual nucleotides at each cleavage site in the long target RNA for base pairing with the hammerhead ribozymes was determined by quantitative chemical modification mapping. The results of these studies demonstrate that the ability of hammerhead ribozymes to efficiently cleave a long target RNA is directly determined by the availability of nucleotides near the cleavage site for base pairing with the ribozyme and that it is unlikely that steric hindrance significantly affects the accessibility of hammerhead ribozyme cleavage sites.
The 4903HH ribozyme and substrate RNA oligonucleotides were provided by Dharmacon Research, Inc. and were deprotected as previously described (23 ). All other oligonucleotide RNA ribozymes and substrates were synthesized on an Applied Biosystems 394 DNA/RNA synthesizer. Ribozyme and substrate oligonucleotides for 5218HH and 5320HH were deprotected in concentrated NH4OH/ethanol (3:1) overnight at 55°C. After drying in vacuum, silyl protecting groups were removed by resuspending the pellet in 40 equivalents of tetrabutylammonium fluoride (TBAF) per equivalent of silyl (400 µl of 1 M TBAF in THF), and incubated at room temperature overnight, in the dark. Following deprotection, the mixture was brought to 1.6 ml with 0.01 M Tris-HCl, pH 7.5, 0.001 M EDTA, and adjusted to 0.5 M NaCl. Ribozyme and substrate oligonucleotides for 4993HH, 5055HH, 5257HH and 5288HH were deprotected in 40% aqueous methylamine at 65°C for 10 min. After drying in vacuum, 2'-hydroxyl protecting groups were removed by incubation in triethylamine:triethylamine 3HF:n-methyl pyrollidinone (23:31:46, v:v) at 65°C for 90 min. Oligonucleotides were ethanol precipitated and dried. Oligonucleotide RNA hammerhead ribozymes were purified by electrophoresis on a denaturing 15% polyacrylamide gel, localized by UV shadowing. Excised gel slices were eluted with H2O, and eluents desalted by chromatography on C-18 Sep-Pak columns (Waters). Concentrations of RNA oligonucleotides were determined by absorption spectrophotometry.
Oligonucleotide RNA substrates were 5' end-labeled in a 10 µl reaction that contained: 25 pmol RNA, 150 µCi [[gamma]-32P]ATP (6000 Ci/mmol, New England Nuclear ), 10 U T4 polynucleotide kinase (New England Biolabs), 70 mM Tris-HCl, pH 7.6, 10 mM MgCl2, and 5 mM dithiothreitol. Reactions were incubated for 30 min at 37°C. End-labeled oligonucleotides were purified by electrophoresis on a denaturing 20% polyacrylamide gel, excised gel slices were eluted with H2O, and eluents desalted by chromatography on C-18 Sep-Pak columns. The concentration of purified end-labeled oligonucleotides was estimated by scintillation counting.
Long substrate RNA was transcribed from pT7vif which contains human immunodeficiency virus type 1 (HIV-1) LAI vif and the 5' region of vpr (nt 4585-5345 of HIV-1 LAI) in the SpeI-PstI site of pGEM5zf+. Transcription mixtures contained 25 µg of pT7vif linearized with SalI, 40 mM Tris-HCl, pH 7.5, 12 mM MgCl2, 1 mM each of CTP, UTP, ATP and GTP, 4 mM spermidine, 10 mM dithiothreitol, and 1000 U T7 RNA polymerase in a 5 ml final volume. The 818 nt RNA was localized by UV shadowing, excised and eluted in 5 ml of 0.01 M Tris-HCl, pH 7.5, 0.001 M EDTA and 0.25 M NaCl at 4°C overnight. RNA was precipitated with 2.5 vol of 100% ethanol, washed twice with 70% ethanol and resuspended in 200 µl H2O. Purified vif-vpr RNA was quantitated spectrophotometrically (assuming 1 OD260 = 40 µg/ml of RNA). The vif-vpr transcript was 3' end-labeled in a 50 µl reaction which contained 25 pmol RNA, 100 µCi [[alpha]-32P]cordycepin 5'-triphosphate (5000 Ci/mmol, New England Nuclear), 2500 U yeast polyA polymerase (United States Biochemical), 20 mM Tris-HCl, pH 7.0, 50 mM KCl, 7 mM MnCl2, 0.2 mM EDTA, 100 µg/ml acetylated BSA and 10% glycerol, and incubated at 30°C for 20 min (24 ). Following end-labeling, vif-vpr RNA was gel purified as described above and quantitated by scintillation counting. Immediately prior to use in kinetic or chemical modification assays, the vif-vpr transcript was heated to 50°C for 10 min and cooled to 37°C for 2 min.
The apparent second order rate constant for reaction of free substrate and ribozyme, (kcat/Km)S, was determined under single-turnover conditions. Ribozyme was renatured in 20 mM MgCl2 at 50°C for 10 min and cooled to 37°C for 2 min. Final reactions (50 µl) contained 2-100 nM ribozyme, ~0.5 nM end-labeled substrate, 50 mM Tris-HCl, pH 7.5 (at 37°C), 10 mM MgCl2, 10 mM NaCl and 140 mM KCl. Reactions were initiated by addition of substrate (pre-warmed to 37°C for 2 min), and 7 µl aliquots were removed at specified times and mixed with an equal volume of stop solution (82% formamide, 50 mM EDTA, 0.04% xylene cyanol and 0.04% bromophenol blue, and 2× TBE). Substrate and product were separated by denaturing polyacrylamide gel electrophoresis and quantitated with a Molecular Dynamics PhosphorImager. The data obtained for short substrates were corrected for unreactive substrate (range 0.04-0.2) by (fraction S - fraction Sunreactive)/(1 - fraction Sunreactive) and the pseudo-first order rate constant determined from plots of fraction substrate versus time, followed to <5% substrate remaining. For determination of (kcat/Km)long S the pseudo-first order rate constant was determined from initial rates of reaction obtained from plots of substrate remaining versus time, followed to no less than 10% substrate remaining.
Chemical modification mapping of the vif-vpr transcript was performed under conditions that were similar to the conditions used for ribozyme kinetic studies. The extent of base pairing at A and C was determined by reaction of vif-vpr RNA with dimethyl sulfate (DMS). DMS (Aldrich Chemical Co, Inc.) was diluted 1:1 (vol:vol) in ethanol immediately prior to use. The reaction mixture contained 14 pmol of purified vif-vpr RNA, 70 mM HEPES-KOH, pH 7.8, 10 mM MgCl2, 10 mM NaCl, 140 mM KCl, 1 mM DTT in a final volume of 200 µl. The reaction mixture was heated to 50°C for 10 min, then cooled to 37°C. The vif-vpr transcript was modified with DMS by addition of 1 µl of dilute DMS to the 200 µl reaction mixture and incubated at 37°C for 10 min. This incubation time was chosen because in time course experiments it was the minimum time required to give primer extension stops that were easily detected and quantitated (see below). The reaction was terminated by the addition of 50 µl DMS stop solution (1 M [beta]-mercaptoethanol, 1.5 M NaOAC, 1.0 M Tris pH 7.5, 0.1 mM EDTA) and incubated on ice. Modified RNA was precipitated by addition of 750 µl of ethanol and washed with 450 µl of 70% ethanol. Modified RNA was dried and resuspended in 150 µl of 300 mM Na acetate, extracted with phenol:chloroform:isoamyl alcohol, ethanol precipitated, washed with 70% ethanol, dried and resuspended in 30 µl of H2O. Controls were treated in an identical fashion and included RNA to which no DMS was added, and RNA to which the DMS stop solution was added prior to the addition of DMS.
The extent of base pairing at G and U was determined by reaction of vif-vpr RNA with 1-cyclohexyl-3-(2-morpholinoethyl)carbodiimide metho-p-toluenesulfonate (CMCT). Prior to use, CMCT (Sigma) was dissolved in CMCT buffer (70 mM HEPES-KOH pH 7.8, 10 mM MgCl2, 10 mM NaCl, 140 mM KCl) at a final concentration of 42 mg/ml. The reaction mixture contained 14 pmol vif-vpr RNA in 20 µl reaction buffer (see above). The reaction mixture was heated to 50°C for 10 min and cooled to 37°C for 10 min. After addition of 20 µl of CMCT, the reaction mixture was incubated at 37°C for 20 min. This incubation time was chosen because it was the time required to give primer extension stops that were comparable in intensity to DMS primer extension stops at adjacent nucleotides. The CMCT reaction was terminated by the addition of 100 µl 0.3 M NaOAc, followed by ethanol precipitation and phenol:chloroform:isoamyl alcohol extraction as described above for DMS modified RNA. Controls were treated in an identical fashion and included RNA to which no CMCT was added, and RNA to which the CMCT stop solution was added prior to the addition of CMCT.
Modifications in the vif-vpr transcript by DMS or CMCT were detected by extension of oligonucleotide primers by reverse transcriptase. Four primers were designed to anneal to sites located ~100 nt apart in the vif-vpr transcript: V5330A, 5'-GGGCTGCAGCAGTTGTTGCAG-3'; V5228A, 5'-ATCCTAGGAAAATGTCTAACAGC-3'; V5122A, 5'-CCATCTATCCTCTGTCAGTTTCG-3'; V5031A, 5'-GATCCTACCTTGTTA-TGTCCTGC-3'. Primers were 5' end-labeled with T4 polynucleotide kinase as described above for RNA oligonucleotides. Each primer was annealed to either modified or unmodified vif-vpr RNA in an annealing mix that contained 2 pmol RNA, 1 pmol 32P end-labeled primer, 0.05 M Tris-HCl pH 8.3, 0.06 M NaCl, 0.01 M DTT in a final volume of 12 µl, by incubation at 90°C for 3 min followed by cooling to room temperature. The primer extension reaction contained 2 µl of annealing mix and 0.4 mM dNTPs, 1 U AMV reverse transcriptase (Life Sciences, Inc.), 0.05 M Tris-HCl pH 8.3, 0.06 M NaCl, 0.01 M DTT and 0.03 M MgOAc in a final volume of 5 µl. Primer extension reactions were incubated at 50°C for 30 min. In addition, an RNA sequencing ladder was created by primer extension of unmodified vif-vpr RNA under the same conditions except that either ddGTP, ddCTP, ddATP or ddTTP were added to the reaction at a final concentration of 80 µM. Primer extension was terminated by the addition of stop solution (0.4× TBE, 0.04% bromophenol blue, 0.04% xylene cyanol, 94% formamide) and all reactions were resolved by electrophoresis on denaturing 8% polyacrylamide gels. Each primer extension reaction was analyzed twice on the same polyacrylamide gel: in a long run in which the xylene cyanol was run to the bottom of the gel and a short run in which the bromophenol blue was run to the gel bottom. Treatment of vif-vpr RNA with DMS or CMCT as described above resulted in 80-90% decreases in the amount of full-length primer extension product. Thus, it is possible that in this analysis a single molecule of vif-vpr transcript was modified more than once.
Individual bands in the lanes that corresponded to modified or unmodified RNA were quantitated by volume integration with a Molecular Dynamics PhosphorImager and ImageQuaNT software. The identity of each band was confirmed by comparison to the dideoxynucleotide sequencing lanes. To correct for differences in gel loading between the short and long runs, the integrated volumes of individual bands that were present in each lane in both the short and long runs were used to normalize the long run values to the short run values. The integrated volume of each band was then corrected for variation in gel loading and phosphorimager screen exposure by dividing each individual value by the total integrated volume above the primer in that lane in the short run (25 ). The normalized band volume for each nucleotide was then calculated as the difference between the corrected band volume in the modified lane and the corrected band volume in the unmodified lane. All statistical analyses were performed with StatView software (Abacus Concepts, Berkeley, CA).
Hammerhead ribozymes were targeted to 15 sites in the vif-vpr region of HIV-1. The vif-vpr region was chosen as a target for anti-HIV-1 ribozymes because this region is important for viral infectivity, and the nucleotide sequence of this region is highly conserved among different HIV-1 isolates (26 ). All hammerhead ribozymes contained the stem II and conserved catalytic core nucleotides shown in Figure 1 , and differed only in the nucleotide composition of stems I and III. The collective GC content of the hammerhead recognition sites (40%) was similar to the overall GC content of the 818 nt vif-vpr transcript (43%).
Summary of rate constants measured for hammerhead ribozymes targeted to specific sites in vif-vpr
Ribozyme
Short substrate
(kcat/Km)S (×107 M-1min-1)a
Cleavage site
short Sb
long Sc
accessibilityd (%)
4903HH
GGAGAUA <=> UAGCACe
3
0.005
0.2
4993HH
AGGCCUU <=> AUUAGG
1
0.003
0.3
5055HH
GGAUCUC <=> UACAAU
0.6
0.001
0.2
5218HH
GGAGCUU <=> AAGAAU
2
0.0002
0.01
5257HH
UUGGCUC <=> CAUGGC
0.9
0.0002
0.02
5288HH
GAAACUU <=> AUGGGG
5
0.1
2
5320HH
AGCCAUA <=> AUAAGA
2
0.07
4
a(kcat/Km)S is the apparent second order rate constant for the reaction of free ribozyme and free substrate.bSynthetic 13 nt substrate. Standard error from ±2% to ±13%.c818 nt vif-vpr substrate. Standard error from ±4% to ±35%.dAccessibility of cleavage sites in the vif-vpr substrate is [(kcat/Km)long S/(kcat/Km)short S] × 100.eSequence of the 13 nt synthetic substrates (Fig. 1). These are also the sequences of the hammerhead ribozyme reognition sites in vif-vpr. Arrows indicate site of cleavage.
The accessibility of cleavage sites in the long target RNA to hammerhead ribozymes was determined by analysis of cleavage reactions performed under conditions in which the association step was expected to be rate-limiting. First, the apparent second order rate constant for the reaction of free ribozyme and free short substrate, (kcat/Km)short S, was determined with 13 nt synthetic substrates that were not predicted to form stable secondary structures. Analysis of short substrate cleavage reactions by electrophoresis on denaturing polyacrylamide gels revealed only a single cleavage product and, on semi-logarithmic plots, short substrate cleavage was linear to no less than 5% of substrate remaining (Fig. 2 A). For all ribozymes, the rate of short substrate cleavage was proportional to ribozyme concentration (Fig. 3 A) and the apparent second order rate constant, (kcat/Km)short S, was determined from the slope of these lines (Table 1 ). For a hammerhead ribozyme, (kcat/Km)S = konS [approx] 2 × 107 M-1min-1 (27 ). Cleavage of short oligonucleotide substrates by ribozymes 4903HH, 4993HH, 5055HH, 5218HH, 5257HH, 5288HH and 5320HH gave values of (kcat/Km)short S that are similar to the value of konS for a hammerhead ribozyme. The value of (kcat/Km)short S for these ribozymes is also within the range of the second order rate constant for RNA bimolecular helix formation (107-108 M-1min-1 at 32.4°C; 28 ,29 ). Thus, it is likely that (kcat/Km)short S for these seven ribozymes represents the rate of association of free ribozyme and free substrate. Eight other hammerhead ribozymes targeted to sites in vif-vpr (4957, 4975, 5016, 5025, 5057, 5084, 5234 and 5276) were found to have values of (kcat/Km)short S that were significantly less than the range of the second order rate constant for RNA bimolecular helix formation; these ribozymes were not evaluated further.
In contrast to the hammerhead ribozyme, the ribozyme derived from the self-splicing intron of Tetrahymena thermophila is a much larger molecule (388 nt) that base pairs to a 6 nt substrate cleavage site (N5U). Previously, we determined the accessibility of sites in the vif-vpr transcript for five Tetrahymena ribozymes(Table 2 ). The conditions of temperature, pH, magnesium and monovalent salt concentration were identical to those employed in the present study of hammerhead ribozyme accessibility. In contrast to hammerhead ribozymes, the accessibility of Tetrahymena ribozyme cleavage sites in the long substrate was not dependent on the availability of any nucleotide in the N5U recognition site (Fig. 6 C). Likewise, Tetrahymena ribozyme cleavage site accessibility was not dependent on the mean availability of each N5U site (Table 2 ; r2 = 0.12).
It was not surprising to find in the present study that cleavage of the long target RNA by hammerhead ribozymes was inhibited 25-10 000-fold, compared to cleavage of short substrates. The present study is unique, however, because quantitative chemical modification mapping of the individual nucleotides in each ribozyme recognition site was used to determine the relationship between the availability of cleavage sites for base pairing with the hammerhead ribozyme and long target RNA cleavage efficiency. The simple three-dimensional structure of a hammerhead ribozyme suggests that its ability to associate with cleavage sites within a long target RNA is determined by the ability to form stems I and III by specific base pair interactions with the target. Intramolecular base pairs at the cleavage site, that result from either secondary or tertiary structure in the target RNA, would be expected to inhibit association. We observed a direct relationship between the accessibility of long substrate cleavage sites to hammerhead ribozymes and the availability of cleavage site nucleotides for chemical modification (Fig. 6 B). It is important to note that this relationship existed over a 400-fold range of cleavage site accessibility and, because we studied ribozymes that had a broad spectrum of long target RNA cleavage efficiencies, these results are not overly influenced by highly efficient or inefficient ribozymes. Therefore the relationship between the accessibility of long substrate cleavage sites to hammerhead ribozymes and the availability of cleavage site nucleotides for chemical modification explains not only why the efficient ribozymes worked well, but also explains why the inefficient ribozymes worked poorly.
We thank Tom Cech for helpful discussions and providing support for the ribozyme cleavage assays. We thank Anne Gooding for synthesis of RNA oligonucleotides. We also thank Dharmacon Research, Inc. for providing the 4903 hammerhead ribozyme and substrate. This work was supported by grants from the Colorado RNA Center of the Colorado Advanced Technology Institute and the National Institutes of Health (U01AI35226, F32AI9506, K11AI01159 and R01GM55503).
26 Myers,G., Korber,B., Berzofsky,J.A., Smith,R.F. and Pavlakis,G.N. (1992) Human Retroviruses and AIDS 1992. Theoretical Biology and Biophysics, Los Alamos, NM.
27 Hertel,K.J., Herschlag,D. and Uhlenbeck,O.C. (1994) Biochemistry, 33, 3374-3385.MEDLINE Abstract
28 Porschke,D. and Eigen,M. (1971) J. Mol. Biol., 62, 361-381.MEDLINE Abstract
29 Porschke,D., Uhlenbeck,O.C. and Martin,F.H. (1973) Biopolymers, 12, 1313-1335.
30 Peattie,D.A. and Gilbert,W. (1980) Proc. Natl. Acad. Sci. USA, 77, 4679-4682.MEDLINE Abstract
31 Moazed,D., Stern,S. and Noller,H.F. (1986) J. Mol. Biol., 187, 399-416.MEDLINE Abstract
T. Sugiyama, M. Gursel, F. Takeshita, C. Coban, J. Conover, T. Kaisho, S. Akira, D. M. Klinman, and K. J. Ishii CpG RNA: Identification of Novel Single-Stranded RNA That Stimulates Human CD14+CD11c+ Monocytes
J. Immunol.,
February 15, 2005;
174(4):
2273 - 2279.
[Abstract][Full Text][PDF]
S. Atsumi, Y. Ikawa, H. Shiraishi, and T. Inoue Selections for constituting new RNA-protein interactions in catalytic RNP
Nucleic Acids Res.,
January 15, 2003;
31(2):
661 - 669.
[Abstract][Full Text][PDF]
J. Kurreck, B. Bieber, R. Jahnel, and V. A. Erdmann Comparative Study of DNA Enzymes and Ribozymes against the Same Full-length Messenger RNA of the Vanilloid Receptor Subtype I
J. Biol. Chem.,
February 22, 2002;
277(9):
7099 - 7107.
[Abstract][Full Text][PDF]
B. Bramlage, E. Luzi, and F. Eckstein HIV-1 LTR as a target for synthetic ribozyme-mediated inhibition of gene expression: site selection and inhibition in cell culture
Nucleic Acids Res.,
November 1, 2000;
28(21):
4059 - 4067.
[Abstract][Full Text][PDF]
M. Amarzguioui, G. Brede, E. Babaie, M. Grotli, B. Sproat, and H. Prydz Secondary structure prediction and in vitro accessibility of mRNA as tools in the selection of target sites for ribozymes
Nucleic Acids Res.,
November 1, 2000;
28(21):
4113 - 4124.
[Abstract][Full Text][PDF]
B. O’Neill, S. Millington–Ward, M. O’Reilly, G. Tuohy, A.-S. Kiang, P. F. Kenna, P. Humphries, and G. J. Farrar Ribozyme-Based Therapeutic Approaches for Autosomal Dominant Retinitis Pigmentosa
Invest. Ophthalmol. Vis. Sci.,
September 1, 2000;
41(10):
2863 - 2869.
[Abstract][Full Text]
CiteULike 
Connotea 
Del.icio.us What's this?
