Strategems in vitro for gene therapies directed to dominant mutations
Strategems in vitro for gene therapies directed to dominant mutationsSophia Millington-Ward+, Brian O'Neill+, Gearoid Tuohy, Najma Al-Jandal, Anna-Sophia Kiang, Paul F. Kenna, Arpad Palfi1, Patrick Hayden, Fiona Mansergh, Avril Kennan, Peter Humphries and G. Jane Farrar*
The Wellcome Ocular Genetics Unit, Genetics Department, Trinity College Dublin, Dublin 2, Ireland and 1Department of Zoology and Cell Biology, Jozsep Attila University, Egyetem U-2, PO Box 659, Szeged H-6722, Hungary
Received February 6, 1997;Revised and Accepted June 3, 1997
A major difficulty associated with the design of gene therapies for autosomal dominant diseases is the immense intragenic heterogeneity often encountered in such conditions. In order to overcome such difficulties we have designed, and evaluated in vitro,three strategies which avoid a requirement to target individual mutations for genetic suppression. In the first, normal and mutant alleles are suppressed by targeting sequences in transcribed but untranslated regions of transcript (UTRs), enabling introduction of a replacement gene with the correct coding sequencing but altered UTRs to prevent suppression. A second approach involves suppression in coding sequence and concurrent introduction of a replacement gene by exploiting the degeneracy of the genetic code. A third strategy utilises intragenic polymorphism to suppress the disease allele specifically, the advantage being that a proportion of patients with different disease mutations have the same polymorphism. These approaches provide a wider choice of target sequence than those directed to single disease mutations and are appropriate for many mutations within a given gene. General methods for suppression may be directed towards the primary defect or a secondary effect associated with the disease process, such as apoptosis. Three general methods targeting the primary defect which circumvent problems of allelic genetic heterogeneity are explored in vitro using hammerhead ribozymes designed to target transcripts from the rhodopsin, peripherin and collagen 1A1 and 1A2 genes, extensive genetic heterogeneity being a feature of associated disease pathologies.
Gene therapies directed towards recessive disorders typically involve introduction and expression of a wild-type gene to ameliorate disease pathology. In contrast to recessive disorders, dominant disorders may be caused by a reduction in the level of wild-type protein, by mutant protein (gain of function mutation) or by a combination of both. Gene therapies for dominant diseases, where the pathology is due at least in part to a gain of function mutation rather than haplo-insufficiency alone, may require suppression of the disease allele while in many cases maintaining expression of the wild-type allele.
There are techniques by which to attempt gene suppression. These, in conjunction with an understanding of the molecular etiology of the disease, result in an increased number of disease targets for therapy. Complete suppression may be difficult to achieve. For some disorders it may be necessary to block expression of a disease allele completely, whereas others may tolerate low levels of mutant protein. Gene silencing has previously been achieved with varying efficiencies using, for example, antisense DNA and RNA, ribozymes, triple helix or peptide-nucleic acids (1 -14 ). Modifications, such as phosphorothioates, have been made to antisense oligonucleotides to increase resistance to nuclease degradation, binding affinity and cellular uptake (15 -17 ). Ribozymes, RNA enzymes which can elicit sequence-specific cleavage of target RNAs (18 -20 ), may also be modified; increased ribozyme stability has been achieved by replacing regions of a ribozyme with DNA or using chemically modified nucleotides (21 ,22 ). Hammerhead ribozyme stability and catalysis can be enhanced by proteins such as p7 nucleocapsid protein from HIV-1 and glyceraldehyde-3-phosphate dehydrogenase (23 -25 ). Ribozymes have also been proposed as a means of replacing defective RNAs with correct copies by targeted trans-splicing (26 ,27 ).
Despite the availability of methods for gene silencing, strategies to differentiate between normal and disease alleles and to target disease alleles specifically may be difficult; frequently they differ by only a single nucleotide. Furthermore, many dominant disorders are heterogeneous. Different mutations within a gene may give rise to a similar disease pathology. Development of `designer' therapies for each mutation may not be viable. Three strategies for suppression of dominant mutations which are independent of the disease mutation are explored in vitro using retinal and collagen genes implicated in retinitis pigmentosa (RP) and osteogenesis imperfecta (OI) respectively. Such approaches could potentially be applied to many dominant disorders involving gain of function mutations.
Studies of degenerative eye disorders, including RP and various macular dystrophies, have resulted in a substantial elucidation of the molecular pathogenesis of these debilitating human retinopathies (28 -37 ). RP involves loss of rod and cone photoreceptors, whereas macular degeneration involves largely loss of cone photoreceptors. Applying genetic linkage, X-linked RP genes, autosomal dominant RP (adRP) genes and autosomal recessive RP genes have been localised. In some cases the disease gene has been characterised and specific mutations identified. Mutations in the genes encoding two photoreceptor proteins, rhodopsin and peripherin, have been implicated in autosomal dominantly inherited retinopathies (38 -42 ). Approximately 100 rhodopsin mutations have been identified in patients with RP or congenital stationary night blindness and ~40 peripherin mutations in patients with adRP or macular dystrophies. Studies of transgenic mice with retinopathies have indicated that at least some of the disease pathology is due to dominant gain of function mutations (43 -44 ); however, accurate expression of the wild-type allele may also be required, since over- or under-expression of, for example, wild-type rhodopsin can lead to disease pathology. OI is an autosomal dominantly inherited brittle bone disorder affecting ~1 in 12 000 people. The pathology in some forms of OI is due to mutations in the collagen 1A1 and 1A2 genes (Col1A1 and Col1A2); >150 mutations have been identified to date (45 ,46 ). The genes encode two type I pro-collagen proteins, the most abundant proteins in man and major components of bone and fibrous tissues. These collagens form triple helical structures essential for normal function which contain two pro[alpha]1(I) chains and one pro[alpha]2(I) chain. Type I OI is typically the mildest, caused by null mutations and therefore reduced levels of wild-type protein. In contrast, types II, III and IV are more severe and are due to structural deformities in the protein.
With an increasing knowledge of the molecular etiology of inherited disorders it is timely to explore methods of therapeutic intervention. Towards this end many mouse models have been created (43 ,44 ,47 ,48 ). In parallel, efficient viral and non-viral vectors for gene delivery will be required (49 ,50 ). Additionally, dominant disorders will require strategies suppressing either the primary defect or a secondary effect. Three approaches based on the former are explored in vitro in this study.
Two approaches involving gene suppression and replacement have been investigated. Hammerhead ribozymes (19 ), which can be designed to elicit sequence-specific autocatalytic cleavage of target RNAs, have been used in both strategies. In the first, hammerhead ribozymes were designed to target non-coding regions of RNAs, i.e. to target rhodopsin and peripherin untranslated regions (UTRs). Targeting UTRs provides flexibility in choice of sequence for suppression. In contrast, strategies directed towards single disease mutations restrict the sequence that can be chosen to achieve gene silencing. The approach has the advantage that the same strategy, when directed to non-coding sequences, could be used, in principle, to suppress many different mutations in a given gene.
A second approach employs ribozymes directed towards transcript coding sequence and has been explored in vitro using rhodopsin and peripherin. The approach again provides a wide choice of target sequence which optimises the probability of efficient suppression being achieved. Wild-type transcript is provided using a replacement gene which is altered, exploiting the degeneracy of the genetic code, such that the replacement gene codes for wild-type protein but is protected from ribozyme cleavage; the replacement gene is altered at third base positions. In essence the strategies involve gene suppression and replacement such that the replacement gene cannot be silenced. The same suppression and replacement steps could in principle be used to ameliorate many different mutations in a given gene, circumventing the need for mutation-dependent therapies. This is particularly relevant when large numbers of mutations within a single gene cause the disease pathology, as is the case with many dominant disorders. Rhodopsin and peripherin were chosen in this study due to the identification of many dominantly inherited mutations within these genes in patients with retinopathies.
A third approach to achieve suppression in a mutation-independent manner has been addressed. As sequence databases expand it is becoming evident that levels of intragenic polymorphism are substantial. Such polymorphism has been used to direct ribozyme cleavage specifically to transcripts from one allele of a polymorphism while allowing expression of the other allele. The approach also has the advantage that it is independent of the disease mutation. The proportion of patients who potentially could be treated would depend on the allelic frequencies of polymorphisms in a population (2pq). Additionally, the approach allows continued expression of one allele and hence a replacement gene would only be required in dominant diseases where some pathology was due to reduced levels of wild-type protein. Two examples of allele-specific suppression have been tested in vitro using polymorphisms in the collagen 1A1 and 1A2 genes, mutations in both of which cause OI.
All three approaches were explored in vitro as follows. Rhodopsin, peripherin and collagen 1A1 and 1A2 cDNAs were expressed in vitro from the T7 promoter in pCDNA3; expression products were the correct size. Appropriate modifications to replacement genes were made using primer-directed, PCR-based mutagenesis. Expression products from replacement genes were again the correct size. Ribozyme cleavage of unadapted and adapted transcripts was tested in vitro at various MgCl2 concentrations and for varying times (Figs 1 -5 ). Cleavage products with all hammerhead ribozymes tested were of the predicted size.
Mouse rhodopsin. Ribozyme 3 (Rz3) was designed to target a GUC target site (at position 1393-1395) in a predicted open loop structure in mouse rhodopsin 5'-non-coding sequence (Table 1 ). Antisense arms were 7 and 8 bases respectively, to provide sequence specificity while supporting autocatalytic cleavage of target RNA. Mouse rhodopsin cDNA was altered in non-coding sequence; 5'-UTR sequence was replaced by human peripherin 5'-UTR sequence (mRhoH1). UTR sequence was chosen from a gene expressed in the same tissue as rhodopsin, i.e. in photoreceptor cells, to minimise effects on tissue-specific expression. A second hybrid with a single base change (U -> G) in the 5'-UTR, altering GUC to GGC and hence removing the ribozyme cleavage site, was generated (mRhoH2). Unadapted mouse rhodopsin transcripts were cleaved by Rz3 (Fig. 1 ). In most cases RNA cleavage was complete and cleavage products were observed rapidly after addition of divalent ions. Residual uncleaved transcript was present at 5 mM MgCl2 (Fig. 1 A, lane 4). Both mRhoH1 and mRhoH2 transcripts with altered UTR sequences remained intact (Fig. 1 B and C), indicating the sequence specificity of the Rz3 target and that a single base change protected transcripts from cleavage. In contrast, most unadapted rhodopsin RNA was cleaved by Rz3 even in the presence of the altered transcripts. Subtle sequence modifications such as that in mRhoH2 may be all that are required to prevent cleavage. However, transcripts from mRhoH1 would be protected from cleavage and binding by a range of ribozymes/antisense and therefore could be of particular use when multiple suppressors were required to achieve efficient silencing.Human peripherin. Ribozymes 8 and 9 (Rz8 and Rz9) were designed to target NUX and GUX sites (CUA and GUU at positions 234-236 and 190-192 respectively) at two predicted open loop structures in human peripherin 5'-UTR sequence (Fig. 2 ). A human peripherin cDNA with altered 5'-non-coding sequences was generated by replacing a portion of human peripherin 5'-UTR sequence (76-212) with mouse peripherin 5'-UTR sequence (bases 84-250), thereby altering sequences around the target sites for Rz8 and Rz9. Subsequent to expression in vitro the majority of the unadapted RNA was cleaved by Rz8 (Fig. 3 A and B); the intensity of the uncleaved transcript decreased over time. In contrast, the adapted RNA with altered non-coding sequence remained intact. Similar results were obtained with Rz9 (Fig. 3 C and D), which targets a different open loop structure than Rz8. However, although efficient cleavage was obtained with both ribozymes, cleavage profiles at various MgCl2 concentrations varied between Rz8 and Rz9; Rz9 was active over a broader range. Some ribozymes may be preferentially active under physiological conditions; combinations of ribozymes targeting different predicted two-dimensional RNA conformations or requiring different intracellular conditions may be required to achieve effective cleavage in vivo.
Human rhodopsin. Ribozyme 10 (Rz10) was designed to target a GUX site (at position 475-477) in the coding sequence of human rhodopsin at a large robust open loop structure as assessed by predictions of two-dimensional conformations (Fig. 4 ). Rz10 cleaves human rhodopsin RNA effectively in vitro at this site, no residual uncleaved transcript being observed after 3 h cleavage in 10 mM MgCl2 at 37oC. A replacement gene coding for wild-type amino acids but with a single base change at a wobble position (GUCVal -> GUGVal) was generated using PCR-based mutagenesis. Transcripts from this modified gene remained intact (Fig. 4 ) and were resistant to ribozyme attack. Unadapted human rhodopsin was expressed to the AcyI site after the stop codon and therefore includes the complete coding sequence (Fig. 4 ).
Human peripherin. Ribozyme 30 (Rz30) was designed to target a NUX site (CUA at position 255-257) at a predicted open loop structure in human peripherin coding sequence (Fig. 4 ). A human peripherin DNA fragment coding for wild-type amino acids but with a single nucleotide change (CTALeu -> CTGLeu) was generated using PCR-based mutagenesis. Subsequent to expression in vitro most unadapted RNA was cleaved by Rz30 (Fig. 4 ). In contrast, adapted RNA with modified coding sequence remained intact. Exploiting the wobble hypothesis enables suppression and replacement with a gene resistant to the suppressor(s). Many such sites are available in both human rhodopsin and peripherin and indeed in most genes.AB
Human collagen 1A1 and 1A2. A single base polymorphism has previously been reported in the 3'-UTR of the human collagen 1A1 gene at position 3210, one allele having a C nucleotide, the other a T nucleotide. Notably, this polymorphism occurs at a predicted open loop structure in the transcript and, moreover, allele T creates a ribozyme cleavage site (GUC), whereas allele C has a GAC sequence at this position. A hammerhead ribozyme (Rzcol1A1) was designed to target allele T of the polymorphism at the GUX site. Both alleles were cloned and expressed in vitro, however, only allele T was cleaved by Rzcol1A1, allele C remaining intact. Rzcol1A1 cleaved allele T transcripts rapidly upon addition of divalent ions (Fig. 5 ). Similarly, a polymorphism in human collagen 1A2 has been used to achieve allele-specific ribozyme cleavage. The polymorphism is found at position 907 in the coding sequence, one allele containing an A nucleotide while the other has a T nucleotide. A hammerhead ribozyme (Rzcol1A2) designed to target the ribozyme cleavage site on allele T, a GUC site, elicited cleavage of transcripts from this allele specifically while leaving transcripts from allele A intact (Fig. 5 ).
Translating developments in gene delivery to the treatment of dominant and polygenic disorders will require methods to suppress the effect(s) of mutant alleles. Frequently autosomal dominant diseases are heterogeneous in their etiologies. The intra-allelic heterogeneity associated with rhodopsin and peripherin-linked adRP is mirrored in many other dominant disorders. Osteogenesis imperfecta, epidermolysis bullosa and hypertropic cardiomyopathy, induced by various mutations in the collagen 1A1 and 1A2 genes, collagen VII and keratin 14 genes and the [beta]-myosin gene respectively, represent a few of the very large number of examples. Therapeutic approaches based on targeting specific disease mutations may be problematic, since individual mutations are often rare; for example a mutation may be present in a single family.
Distinguishing between wild-type and disease alleles also presents problems, as wild-type and mutant alleles often differ by a single nucleotide. While such specificity has been achieved, for example suppression of a mutation in codon 12 of the K-ras gene using a hammerhead ribozyme which cleaves mutant RNA specifically (6 ,51 ), in many situations it may be difficult to obtain. The presence of a ribozyme target site in the mutant would facilitate specific suppression of the disease allele; for example human rhodopsin Pro23Leu and Del255 mutations, which have previously been observed in human adRP patients, are cleaved specifically in vitro as in Figure 4 . However, the probability of an appropriate site is low given the specificity of the target site and the incidence of open loop structures. If antisense RNA flanking a ribozyme core were able to discriminate between normal and mutant alleles it would extend the number of mutations that could be targeted specifically. However, this approach was explored for one rhodopsin mutation, a Met207Arg mutation; both normal and mutant transcripts were cleaved by a ribozyme whose antisense arms were complementary to the mutant (data not shown). Likewise, the discriminating power of triple helices and peptide-nucleic acids may frequently not be sufficient to specifically silence the disease allele. Clearly there is a need for approaches which provide a general strategy for suppression, circumventing difficulties with genetic heterogeneity and with discriminating between disease and normal alleles. This study explores three approaches in vitro which are independent of the disease mutation.
Examples of suppression and replacement are provided using three trans-acting hammerhead ribozymes directed to open loop structures in mouse rhodopsin and human peripherin transcripts. Hammerhead ribozymes with a consensus sequence (19 ) have been used, however, various modifications can be made to ribozyme design to achieve optimal cleavage. Antisense RNA flanking the ribozyme catalytic core was used successfully to elicit binding and cleavage of target RNAs in a sequence-specific manner. Hammerhead ribozymes were designed to target 5'-UTR sequences; however, 5'-UTR sequences, 3'-UTR sequences, introns or any combination of non-coding sequence could be targeted. UTR sequences have the advantage that they are present in both immature and mature message and therefore may be more efficient targets for achieving gene silencing. Three retinal cDNAs with altered non-coding sequences were generated by mutagenesis. RNAs from altered cDNAs were protected entirely from ribozyme cleavage due to absence of the target site and in some cases sequence complementary to antisense RNA flanking the catalytic core. Notably, the presence of a mixed population of transcripts did not prevent cleavage of unadapted transcripts. Modifications involved replacement of UTR sequence with UTR sequence from a gene expressed in the same tissue or UTR sequence from the same gene but from a different mammalian species, the rationale being to limit possible subsequent effects on level and tissue-specific expression. There is considerable evidence suggesting that gene expression can be regulated at the level of translation and that 5'-UTRs may be involved in such control (61 -64). In this regard it is of interest that transcripts with a single base change at the ribozyme target site were protected from cleavage by Rz3 (Fig. 1 ). Mechanisms involved in regulation of gene expression have not been fully established for retinal genes such as rhodopsin and peripherin; however, non-coding sequences may be important in this. Therefore, subtle modifications to replacement genes to prevent ribozyme cleavage or nucleic acid binding would be preferable. This study provides preliminary data in vitro suggesting that this approach may be feasible.
The degeneracy of the genetic code has provided geneticists with high levels of intragenic polymorphism. This degeneracy has been exploited in two ways with respect to general methods for suppression. Firstly, degeneracy has been utilised to test in vitro an approach also involving suppression and replacement. The target site can be chosen in any part of the coding sequence, increasing the likelihood of obtaining efficient suppression. A hammerhead ribozyme was designed to target a cleavage site in the coding sequence of human rhodopsin such that a replacement gene could be created with a single base change at a wobble position which would eliminate the cleavage site but code for the same amino acid. Cleavage of rhodopsin transcripts was achieved while replacement transcripts generated by PCR-directed mutagenesis remained intact. A single base change protected against cleavage and enabled the introduction of a replacement gene coding for the wild-type protein. Similar results were obtained for human peripherin using a replacement gene with a single base alteration in a wobble position which in principle could provide wild-type protein but whose transcripts would be resistant to ribozyme cleavage (Fig. 4 ). Codon usage has been shown previously to influence gene expression (52 ). However, altering one or indeed a few codons within a gene, together with utilising codons which are used at other sites in the sequence, should minimise possible effects of such modifications and provides an opportunity to alter DNA and RNA without changing the encoded protein. Again, this approach provides a wider choice of target sequence for suppression, increasing the probability that efficient suppression could be achieved.
Polymorphism has been used extensively over the past decade to create genetic maps of the human genome. A further application of such polymorphism would be as a tool to direct therapies to specific alleles. Given expanding numbers of intragenic polymorphisms, this approach should be possible for many genes. Polymorphisms in the collagen 1A1 and 1A2 genes have been chosen, as multiple dominant mutations in these genes cause OI. Notably, one allele of each polymorphism contains a ribozyme cleavage site whereas the target site is absent in both cases in the alternative allele. Allele-specific ribozyme cleavage was obtained in vitro for one allele of both polymorphisms, while the other allele remained intact (Fig. 5 ). The scope of a polymorphism in terms of distinguishing between wild-type and mutant alleles will depend on its frequency in the population. Given hypothetical allele frequencies of, for example, p = 0.5 and q = 0.5, then 50% of individuals (2pq = 0.5) will be heterozygous. Multiple intragenic polymorphisms could be used to increase the proportion of heterozygotes. An advantage of the approach is that fewer suppressors will be required if directed to polymorphisms than if directed to specific disease mutations. In contrast to the suppression and replacement approaches above, this approach involves only suppression. However, replacement may be required if some disease pathology is due to haplo-insufficiency. Notably, the mechanism of action of mutant collagens in OI would suggest that a reduction in the ratio of mutant to wild-type collagen molecules alone would lead to an amelioration in the disease pathology (45 ,46 ). Mice heterozygous for null mutations in the rhodopsin and peripherin genes exhibit abnormal photoreceptors, hence, while suppression of the mutant may ameliorate symptoms, replacement may also be required. Harnessing polymorphism in gene therapies has previously been suggested for tumor therapy by suppressing one allele of a gene involved in cell viability. As tumor tissues are often haploid, this may lead to tumor cell death whereas surrounding normal tissues would potentially be unaffected (53 ). Undoubtedly the resource of intragenic polymorphisms will be utilised in the future development of therapeutic approaches for human disorders.
Transferring efficient RNA cleavage obtained in the study in vitro to in vivo situations may require significant modifications to ribozyme constructs (7 ,54 -61 ) to optimise ribozyme efficiency, stability and localisation in cells. Efficiency and accuracy of ribozyme cleavage has previously been shown to be enhanced using proteins such as nucleocapsid protein; endogenous proteins may confer similar effects in vivo (23 -25 ). Despite difficulties, accurate cleavage of target transcripts by hammerhead ribozymes has been demonstrated in vivo in transgenic animals. Additionally, ribozymes have been delivered and expressed in vivo using adenovirus-mediated transfer (61 ). Although these strategies have been tested in vitro in this study using trans-acting hammerhead ribozymes, other methods for gene silencing, such as antisense RNA or triple helices, could be used if ribozymes were inefficient in vivo.
In the case of the first approach, transcript cleavage in UTRs possibly may not affect the ability of cleavage products to generate protein. Moreover, although lowering RNA levels may lead to a parallel lowering of protein levels, this is not always the case. Cellular mechanisms may prevent a significant decrease in protein levels despite a substantial decrease in levels of RNA. However, in many instances suppression at the RNA level has been shown to be effective in reducing protein levels (60 ). Multiple ribozymes could be designed, for example connected or shotgun ribozymes, to optimise gene silencing if required.
The approaches outlined above are subject to problems of gene delivery which are associated with the development of many gene therapies, i.e. the requirement for safe and efficient vectors to achieve tissue- and level-specific expression. The level at which a suppression effector is present in a cell will be a crucial element in the efficiency of suppression. Moreover, the level of expression of a replacement gene for some disorders will be central to the effectivenss of the therapy. This is particularly relevant in situations where haplo-insufficiency alone has been shown to produce aberrant effects even in the absence of mutant protein, which indeed is the case for rhodopsin and peripherin, as discussed above (48 ). Hence incorporating appropriate transcription control elements into vectors will be vitally important in the process of translating therapeutic approaches into viable future therapies. Additionally, rapid flux in mRNA synthesis may be deleterious to certain cell types, hence, the feasibility of these approaches needs to be evaluated thoroughly in vivo.
General strategies for therapies could be targeted towards the primary defect, as above, or to a secondary effect, such as programmed cell death (apoptosis). Apoptotic cell death has been encountered in many diseases, including many mouse models of human retinopathies. Inhibiting apoptotic mechanisms may provide viable routes for future therapies. A recent study using wild-type mice reported that light-induced photoreceptor apoptosis could largely be eliminated when mice were bred onto a c-fos-/- background (62 ). c-fos has previously been shown to promote apoptosis, moreover, c-fos has also been shown to be elevated in apoptotic photoreceptors in some retinopathies. Down-regulating c-fos expression in photoreceptors may provide a general approach for therapies. Many potential targets for prevention of apoptosis exist (63 ), some of which may prove to be appropriate routes for gene therapy. However, targeting secondary effects may not alleviate or may only partially alleviate disease symptoms or indeed may have deleterious effects. In such situations targeting the primary defect may be necessary to ameliorate disease pathology. Alternatively, a combination of therapies directed to both primary defects and secondary effects may be required to treat some diseases. Using the strategies explored in this study in vitro the same methods of suppression and gene replacement may be a feasible therapeutic approach for many mutations within a given gene. Such systems should expedite the future development of therapeutic strategies for some dominant disorders where inherent genetic heterogeneity together with problems associated with distinguishing between disease and normal alleles have presented significant hurdles.
cDNAs, cDNA hybrids with altered sequences and ribozyme DNA fragments were cloned into commercial expression vectors (pCDNA3 and Bluescript) which enable expression from T7, T3, SP6 or CMV promoters. Inserts were placed in the multiple cloning site (MCS) at or near the terminal ends. Clones containing cDNAs, hybrid cDNAs with altered sequences and ribozymes were sequenced by ABI automated sequencers using standard protocols.
RNA was obtained from clones in vitro using Ribomax kits (Promega) and standard protocols. RNA purifications were undertaken using Bio-101 RNA purification kits or a solution of 0.3 M sodium acetate and 0.2% SDS. Cleavage reactions were performed using standard protocols with varying MgCl2 concentrations (0-15 mM) at 37oC typically for 3 h. Time points were performed at predetermined optimal MgCl2 concentrations for up to 5 h. Radioactively labeled RNA products were obtained by incorporating [[alpha]-32P]rUTP (Amersham) in expression reactions. Labeled products were run on polyacrylamide gels prior to cleavage reactions for the purpose of RNA purification and subsequent to cleavage reactions to establish if RNA cleavage had been achieved. Molar ratios of ribozyme:template were estimated at 100:1 for many of the ribozymes tested (Rz10, Rz20, Rz255, Rzcol1A1 and Rz30), however, higher ratios of ~1000:1 were used for Rz3, Rz8 and Rz9. Threshold ratios required for efficient cleavage are under evaluation for each ribozyme and template before cellular work will be undertaken.
Predicted secondary structures for human and mouse rhodopsin and peripherin mRNAs were obtained using the RNA PlotFold program. Ribozymes were designed to target open loop structures in RNAs that were likely to be accessible to ribozymes and were in non-coding sequence. Integrity of open loops was evaluated from the 15 most probable two-dimensional conformations. RNA structures for truncated RNA products were generated and the integrity of open loops between full-length and truncated RNAs compared.
Mouse rhodopsin cDNA. The full length mouse rhodopsin cDNA was cloned into the EcoRI site of pCDNA3 in an orientation enabling subsequent expression from T7 or CMV promoters. The clone was generated using an EcoRI site present at position 1120 5' of the start of transcription (accession no. M55171).
Hybrid cDNAs with altered 5'-non-coding sequence, mRhoH1. A mouse rhodopsin hybrid cDNA with altered 5'-UTR sequence was generated by PCR primer-directed mutagenesis using a HindIII (in pCDNA3)-Eco47III (in exon 2) cassette. Mouse rhodopsin 5'-UTR sequence was replaced by 5'-UTR sequence from human peripherin, i.e. with sequence from a gene expressed in the same tissue, photoreceptor cells. Mouse rhodopsin sequence begins at the mouse rhodopsin ATG start site.
Hybrid cDNAs with altered 5'-non-coding sequence, mRhoH2. A mouse rhodopsin hybrid cDNA designed to eliminate the GUC ribozyme binding site targeted by Rz3 by altering the U to G (GUC -> GGC) was generated by primer-driven PCR mutagenesis using a HindIII-Eco47III cassette.
Ribozyme construct. A hammerhead ribozyme (Rz3) targeting an open loop in the 5'-UTR was cloned (subsequent to primer synthesis and annealing) into the HindIII and XbaI sites of pCDNA3. The target site (GUC) was at position 1393-1395 (accession no. M55171). Antisense flanks are underlined. The clone includes a hammerhead ribozyme (19 ) consensus sequence.
Rz3:
CUUCGUACUGAUGAGUCCGUGAGGAC GAAACAGAGAC
Human rhodopsin cDNA. A human rhodopsin cDNA with full-length 5'-UTR was cloned into the HindIII and EcoRI sites of pCDNA3. The full-length 5'-UTR sequence was inserted into the clone using primer-driven PCR mutagenesis and a HindIII (in pCDNA3)-BstEII (in exon IV) DNA fragment.
Hybrid cDNA, hRhoH1. A single base change (C -> G) at a ribozyme cleavage site at positions 475-477 (GUC -> GUG) was introduced by primer-directed mutagenesis into human rhodopsin using a HindIII-BstEII cassette.
Mutant cDNAs, hRhoM1 and M2. A single base change (ProCCC23LeuCTC) and a 3 bp deletion (Del255ATC) were introduced by primer-directed mutagenesis into human rhodopsin using HindIII-BstEII and HindIII-AcyI cassettes respectively.
Ribozyme constructs. Hammerhead ribozymes (Rz10, Rz20 and Rz255) designed to target open loop structures in rhodopsin transcripts were cloned into the HindIII and XbaI sites of pCDNA3. Target sites were GUC, CUC and AUC at positions 475-477, 361-363 and 4161-4163 of the human rhodopsin sequence respectively (accession no. K02281). Antisense flanks are underlined.
Rz10:
GGACGGUCUGAUGAGUCCGUGAGGAC GAAACGUAGAG
Rz20:
UACUCGAACUGAUGAGUCCGUGAGGAC GAAAGGCTGC
Rz255:
AUGACCATCUGAUGAGUCCGUGAGGAC GAAAUGACCA
Human peripherin cDNA. Human peripherin cDNAs cloned into the EcoRI site of Bluescript or pCDNA3 were expressed from the T7 promoters in the vectors.
Hybrid cDNAs, hPerH1. A clone with altered 5'-UTR sequence was generated using primer-driven PCR mutagenesis of a BamHI (in the 5'-UTR)-BglII (in the coding sequence) DNA fragment. The clone contains human peripherin 5'-UTR sequences until the BamHI site at position 76, then mouse peripherin 5'-UTR sequence (at position 84) until the ATG start site, where it returns to human peripherin sequence.
Hybrid cDNAs, hPerH2. A DNA fragment containing the T7 promoter and a single base change at a wobble position (257 A -> G) was generated by PCR mutagenesis.
Ribozyme constructs. Hammerhead ribozymes (Rz8 and Rz9) designed to target open loops in the 5'-UTR were cloned into the HindIII and XbaI sites of pCDNA3. The target sites were CUA and GUU at positions 234-236 and 190-192 respectively of the human peripherin sequence (accession no. M62958). Rz30, also cloned into the HindIII and XbaI sites of pCDNA3, targets a cleavage site (CUA) in the coding sequence at positions 255-257. Antisense flanks are underlined.
Rz8:
CCAAGUGCUGAUGAGUCCGUGAGGAC GAAAGUCCGG
Rz9:
CAAACCUUCUGAUGAGUCCGUGAGGAC GAAACGAGCC
Rz30:
ACUUUCAGCUGAUGAGUCCGUGAGGAC GAAAGCGCCA
Human collagen cDNA
Collagen 1A1. Human collagen 1A1 cDNA clones containing the C and T alleles of the polymorphism at 3210 were cloned into the HindIII and XbaI sites of pCDNA3. Clones contain collagen 1A1 sequence from position 2977 to 3347.
Collagen 1A2. Clones with polymorphic variants (position 907 T/A) of collagen 1A2 were generated by PCR mutagenesis using a HindIII-XbaI PCR cassette and were cloned in pCDNA3 using these sites.
Ribozyme constructs. Hammerhead ribozymes (Rzcol1A1 and Rz907) designed to target open loops were cloned into the HindIII and XbaI sites of pCDNA3. The target sites were CUC and GUC at positions 3209-3211 and 906-908 respectively of the human collagen 1A1 and 1A2 sequences respectively (accession no. K01228). Antisense flanks are underlined.
We would like to thank Drs W.Baehr and G.Travis for providing rhodopsin and peripherin cDNA clones. The research unit is supported by grants from the Wellcome Trust, Retinitis Pigmentosa Ireland Fighting Blindness, the US Foundation Fighting Blindness, the Health Research Board (Ireland), EU Copernicus Co-operation, the Ulvescroft Foundation and the British Retinitis Pigmentosa Society.
1 Postel,E.H., Flint,S.J., Kessler,D.J. and Hogan,M.E. (1991) Evidence that a triplex-forming oligodeoxyribonucleotide binds to the c-myc promoter in HeLa cells, thereby reducing c-myc mRNA levels. Proc. Natl. Acad. Sci. USA, 88, 8227-8231.MEDLINE Abstract
2 Duval-Valentin,G., Thuong,N.T. and Helene,C. (1992) Specific inhibition of transcription by triple helix-forming oligonucleotides. Proc. Natl. Acad. Sci. USA, 89, 504-508.MEDLINE Abstract
3 Lange,W., Cantin,E.M., Finke,J. and Dolken,G. (1993) In vitro and in vivo effects of synthetic ribozymes targeted against BCR/ABL mRNA. Leukemia, 7, 1786-1794.MEDLINE Abstract
4 Valera,A. et al. (1994) Expression of GLUT-2 antisense RNA in [beta] cells of transgenic mice leads to diabetes. (1994) J. Biol. Chem., 269, 28543-28546.MEDLINE Abstract
5 Dosaka-Akita,H., Akie,K., Hiroumi,H., Kawakami,Y. and Murakami,A. (1995) Inhibition of proliferation by L-myc antisense DNA for the translational initiation site in human small cell lung cancer. Cancer Res., 55, 1559-1564.MEDLINE Abstract
6 Feng,M., Cabrera,G., Deshane,J., Scanlon,K. and Curiel,D.T. (1995) Neoplastic reversion accomplished by high efficiency adenoviral-mediated delivery of an anti-ras ribozyme. Cancer Res., 55, 2024-2028.MEDLINE Abstract
7 Chowrira,B.M., Pavco,P.A. and McSwiggen,J.A. (1994) In vitro and in vivo comparison of hammerhead, hairpin and hepatitis delta virus self-processing ribozyme cassettes. J. Biol. Chem., 269, 25856-25864.MEDLINE Abstract
8 Knudsen,H. and Nielsen,P.E. (1996) Antisense properties of duplex- and triplex-forming PNAs. Nucleic Acids Res., 24, 494-500.MEDLINE Abstract
9 Ohta,Y., Kijima,H., Ohkawa,T., Kashani-Sabet,M. and Scanlon,K.J. (1996) Tissue-specific expression of an anti-ras ribozyme inhibits proliferation of human malignant melanoma cells. Nucleic Acids Res., 24, 938-942.MEDLINE Abstract
10 Jankowsky,E. and Schwenzer,B. (1996) Oligonucleotide facilitators may inhibit or activate a hammerhead ribozyme. Nucleic Acids Res., 24, 423-429.MEDLINE Abstract
11 Hardenbol,P and Van Dyke,M.W. (1996) Sequence specificity of triplex DNA formation: analysis by a combinatorial approach, restriction endonuclease protection selection and amplification. Proc. Natl. Acad. Sci. USA, 93, 2811-2816.MEDLINE Abstract
12 Porumb,H. et al. (1996) Temporary ex vivo inhibition of the expression of the human oncogene HER2 (NEU) by a triple helix-forming oligonucleotide. Cancer Res., 56, 515-522.
13 Trauger,J.W., Baird,E.E. and Dervan,P.B. (1996) Recognition of DNA by designed ligands at subnanomolar concentrations. Nature, 382, 559-561.
14 Wei,Z. et al. (1996) Hybridisation properties of oligodeoxynucleotide pairs bridged by polyarginine peptides. Nucleic Acids Res., 24, 655-661.MEDLINE Abstract
15 Cazenave,C. et al. (1989) Comparative inhibition of rabbit globin mRNA translation by modified antisense oligodeoxynucleotides. Nucleic Acids Res., 17, 4255-4273.MEDLINE Abstract
17 McKay,R.A. et al. (1996) Enhanced activity of an antisense oligonucleotide targeing murine protein kinase C-[alpha] by the incorporation of 2'-O-propyl modifications. Nucleic Acids Res., 24, 411-417.
18 Cech,T. (1987) The chemistry of self-splicing RNA and RNA enzymes. Science, 236, 1532-1539.MEDLINE Abstract
19 Haseloff,J. and Gerlach,W.L. (1988) Simple RNA enzymes with new and highly specific endoribonuclease activities. Nature, 334, 585-591.MEDLINE Abstract
20 Pley,W.H., Flaherty,K.M. and McKay,D. (1994) Three-dimensional structure of a hammerhead ribozyme. Nature, 372, 68-74.
21 Paolella,G., Sproat,B.S. and Lamond,A.I. (1992) Nuclease resistant ribozymes with catalytic activity. EMBO J., 11, 1913-1919.
22 Taylor,N.R., Kaplan,B.E., Swiderski,P., Li,H. and Rossi,J.J. (1992) Chimeric DNA-RNA hammerhead ribozymes have enhanced in vitro catalytic efficiency and increased stability in vivo. Nucleic Acids Res., 20, 4559-4565.MEDLINE Abstract
23 Tsuchihashi,Z., Khosla,M. and Herschlag,D. (1993) Protein enhancement of hammerhead ribozyme catalysis. Science, 262, 99-102.MEDLINE Abstract
24 Muller,G. et al. (1994) Amino acid requirements of the nucleocapsid protein of HIV-1 for increasing catalytic activity of a Ki-ras ribozyme in vitro. J. Mol. Biol., 242, 4222-429.
25 Sioud,M. and Jespersen,L. (1996) Enhancement of hammerhead ribozyme catalysis by glyceraldehyde-3-phosphate dehydrogenase. J. Mol. Biol., 257, 775-789.MEDLINE Abstract
26 Sullenger,B.A. and Cech,T.R. (1994) Ribozyme mediated repair of defective mRNA by targeted trans-splicing. Nature 371, 619-622.MEDLINE Abstract
27 Jones,J.T., Lee,S.W. and Sullenger,BA. (1996) Tagging ribozyme reaction sites to follow trans-splicing in mammalian cells. Nature Med., 2, 643-648.
28 Humphries,P., Kenna,P.F. and Farrar,G.J. (1992) On the molecular genetics of retinitis pigmentosa. Science, 256, 804-808.MEDLINE Abstract
29 Farrar,G.J. et al. (1993) Extensive genetic heterogeneity in autosomal dominant retinitis pigmentosa. In Hollyfield,J.G. et al. (eds), Retinal Degenerations. Plenun Press, New York, NY.
30 Weber,B.H.F., Vogt,G., Pruett,R.C., Stohr,H. and Felbor,U. (1994) Mutations in the tissue inhibitor of metalloproteinases-3 (TIMP3) in patients with Sorsby's fundus dystrophy. Nature Genet., 8, 352-356.
31 Kajiwara,K., Berson,E.L. and Dryja,T.P. (1994) Digenic retinitis pigmentosa due to mutations at the unlinked peripherin/RDS and ROM1 loci. Science, 264, 1604-1608.MEDLINE Abstract
32 Dryja,T.P., Finn,J.T., Peng,Y.W., McGee,T.L., Berson,E.L. and Yau,K.W. (1995) Mutations in the gene encoding the [alpha] subunit of the rod cGMP-gated channel protein in auotsomal recessive retinitis pigmentosa. Proc. Natl. Acad. Sci. USA, 92, 10177-10181.MEDLINE Abstract
33 Ott,J. et al. (1990) Localising multiple X chromosome-linked retinitis pigmentosa loci using multilocus homogeneity tests. Proc. Natl. Acad. Sci. USA, 87, 701-704.MEDLINE Abstract
34 Jordan,S.A. et al. (1993) Localisation of an autosomal dominant retinitis pigmentosa gene to chromosome 7q. Nature Genet., 4, 54-58.MEDLINE Abstract
35 Inglehearn,C.F. et al. (1993) A new locus for autosomal dominant retinitis pigmentosa on chromosome 7p. Nature Genet., 4, 51-53.MEDLINE Abstract
36 Van Soest,S. et al. (1994) Assignment of a gene for autosomal recessive retinitis pigmentosa (RP12) to chromosome 1q31-q32.1 in an inbred and genetically heterogeneous disease population. Genomics, 22, 499-504.MEDLINE Abstract
37 McWilliam,P. et al. (1989) Autosomal dominant retinitis pigmentosa (adRP): localization of an adRP gene to the long arm of chromosome 3. Genomics, 5, 612-619.
38 Dryja,T.P. et al. (1990) A point mutation of the rhodopsin gene in one form of retinitis pigmentosa. Nature, 343, 364-366.MEDLINE Abstract
39 Farrar,G.J. et al. (1991) A three-base pair deletion in the peripherin-RDS gene in one form of retinitis pigmentosa. Nature, 354, 478-480.MEDLINE Abstract
40 Kajiwara,K. et al. (1991) Mutations in the human retinal degeneration slow gene in autosomal dominant retinitis pigmentosa. Nature, 354, 480-483.MEDLINE Abstract
41 Nichols,B.E. et al. (1993) Butterfly-shaped pigment dystrophy of the fovea caused by a point mutation in codon 167 of the RDS gene. Nature Genet., 3, 202-207.MEDLINE Abstract
42 Wells,V. et al. (1993) Mutations in the RDS gene can cause either RP or macular dystrophy. Nature Genet., 3, 213-218.
43 Prockop,D.J., Kuivaniemi,H. and Tromp,G. (1994) Molecular basis of osteogenesis imperfecta and related disorders of bone. Clin. Plastic Surg., 21, 407-413.
44 Prockop,D.J. and Kiviriko,K.I. (1995) Collagens: molecular biology, diseases and potential for therapy. Annu. Rev. Biochem., 64, 403-434.MEDLINE Abstract
45 Olsson,J.E. et al. (1992) Transgenic mice with a rhodopsin mutation (Pro23His): a mouse model of autosomal dominant retinitis pigmentosa. Neuron, 9, 815-830.MEDLINE Abstract
46 Nassh,M.I., Hollyfield,J.G., Al-Ubaidi,M.R. and Baehr,W. (1993) Simulation of human autosomal dominant retinitis pigmentosa in transgenic mice expressing a mutated murine opsin gene. Proc. Natl. Acad. Sci. USA, 90, 5499-5503.
47 Tsang,S.H. et al. (1996) Retinal degeneration in mice lacking the [gamma] subunit of the rod cGMP phosphodiesterase. Science, 272, 1026-1029.MEDLINE Abstract
48 Humphries,M. et al. (1997) Mouse retinal degeneration due to a targeted disruption of the rhodopsin gene. Nature Genet., 15, 216-219.MEDLINE Abstract
49 Li,T. et al. (1994) In vivo transfer of a reporter gene to the retina mediated by an adenoviral vector. Invest. Ophthalmol. Vision Sci., 35, 2543-2549.
50 Bennett,J. et al. (1996) Photoreceptor cell rescue in retinal degeneration (rd) mice by in vivo gene therapy. Nature Med., 2, 649-654.
51 Kashani-Sabet,M., Funato,T., Florenes,V.A., Fodstad,O. and Scanlon,K.J. (1994) Suppression of the neoplastic phenotype in vivo by an anti-ras ribozyme. Cancer Res., 54, 900-902.MEDLINE Abstract
52 Eyre-Walker,A. (1991) An analysis of codon usage in mammals: selection or mutation bias? J. Mol. Evol., 33, 442-449.MEDLINE Abstract
53 Housman,DE. (1995) Inhibitors of alternative alleles of genes as a basis for cancer therapeutic agents. World Intellectual Property Organisation, WO 95/03335.
54 Sonenberg,N. (1993) Remarks on the mechanism of ribosome binding to eukaryotic mRNAs. Gene Expression, 3, 317-323.MEDLINE Abstract
55 Curtis,D., Lehmann,R. and Zamore,P.D. (1995) Translational regulation in development. Cell, 81, 171-178.MEDLINE Abstract
56 Wood,M.W., VanDongen,H.M.A. and VanDongen,A.M.J. (1996) The 5'-untranslated region of the N-methyl-D-asparate receptor NR2 subunit controls efficiency of translation. J. Biol. Chem., 271, 8115-8120.MEDLINE Abstract
57 Cameron,F.H. and Jennings,P.A. (1989) Specific gene suppression by engineered ribozymes in monkey cells. Proc. Natl. Acad. Sci. USA, 86, 9139-9143.MEDLINE Abstract
58 Sullenger,B.A. and Cech,T.R. (1993) Tethering ribozymes to a retroviral packaging signal for destruction of viral RNA. Science, 262, 1566-1569.MEDLINE Abstract
59 Lieber,A. and Strauss,M. (1995) Selection of efficient cleavage sites in target RNAs by using a ribozyme expression library. Mol. Cell. Biol., 15, 540-551.MEDLINE Abstract
60 Efrat,S. et al. (1994) Ribozyme-mediated attenuation of pancreatic [beta]-cell glucokinase expression in transgenic mice results in impaired glucose-induced insulin secretion. Proc. Natl. Acad. Sci. USA, 91, 2051-2055.MEDLINE Abstract
61 Lieber,A. and Kay,M.A. (1996) Adenovirus-mediated expression of ribozymes in mice. J. Virol., 70, 3153-3158.MEDLINE Abstract
62 Hafezi,F. et al. (1997) Absence of c-fos prevents light-induced apoptotic cell death of photoreceptors in retinal degeneration in vivo. Nature Med., 3, 346-349.
63 Nicholson,D.W. (1996) ICE/CED3-like proteases as therapeutic targets for the control of inappropriate apoptosis. Nature Biotechnol., 14, 297-301.
*To whom correspondence should be addressed. Tel: +35 31 6082482; Fax: +35 31 6719394; Email: gjfarrar@vax1.tcd.ie
+These authors contributed equally to this work
-->
This page is maintained by OUP admin. Last updated Wed Aug 13 15:52:16 BST 1997. Part of the OUP Journals World Wide Web service.
Copyright
Oxford University Press, 1996
