Antisense Derivatives of U7 Small Nuclear RNA As Modulators of Pre-Mrna Splicing - Stamm

Comments by Stefan (yellow)

Glossary (green)

Referral to other chapters (pink)

List of vendors (gray)

Reagents added to the database (blue)

Antisense derivatives of U7 small nuclear RNA as modulators of pre-mRNA splicing

Overview:

PCR-based production of splicing modulation tools in U7 SmOPT (protocol 1)

Subcloning strategy into transfer vector, e.g. pWPTS (protocol 2)

Outcome:U7 snRNA tool to modulate splicing of a particular target gene by transfection or vector-based transduction

Questions answered:
How can a specific splicing pattern be modulated?
What is the therapeutic outcome?

Antisense derivatives of U7 small nuclear RNA as modulators of pre-mRNA splicing

Kathrin Meyer and Daniel Schümperli *

Institut für Zellbiologie, Universität Bern, Bern, Switzerland

*Address correspondence to: Daniel Schümperli, Institut für Zellbiologie, Universität Bern, Baltzerstrasse 4, CH-3012 Bern, Switzerland,
Phone: +41-31-6314675, Fax +41-31-6314616, e-mail:

1. Abstract

Although it is possible to modulate splicing with antisense oligonucleotides, in animals or patients, this approach is confronted with problems of efficient delivery to the tissue of interest and requires repeated administrations of large amounts of costly materials. Daniel, can you start with the positive here, and put the first sentence after the first: In vivo expressed short splicing-modulating RNAs represent an interesting alternative, be it for gene therapy, to understand the nature of splicing mutations be it for gene therapy or for basic studies on alternative splicing. The U7 system is therefore a good alternative to antisense oligos… that have problems. Modified derivatives of the U7 small nuclear RNA (snRNA) involved in histone RNA 3' end processing are particularly well suited for this kind of approach. Two important features are the nuclear accumulation and high stability of the RNA as part of a small nuclear ribonucleoprotein particle. In particular, U7 derivatives containing two tandem antisense sequences directed against targets upstream and downstream of an exon can induce the efficient and specific skipping of that exon. U7 snRNA derivatives can also be equipped with an additional functional moiety, e.g. an exonic splicing enhancer sequence capable of tethering a splicing activator protein of the SR family to the exon of interest, in order to promote the inclusion of this exon into the mRNA. This article will describe the constructions of such U7 derivatives starting from the original U7 Sm OPT plasmid. U7 expression cassettes have been successfully introduced into a great number of cell lines, primary cells or tissues with the help of lentiviral and adeno-associated viral vectors, and we will also describe a procedure used for subcloning into such viral delivery vectors.

Keywords: splicing modulation / in vivo expressed U7 snRNA derivatives / gene therapy / PCR mutagenesis

1. 
   Theoretical background

2.1 What makes U7 snRNA a suitable in vivo splicing modulation tool?

To be effective, splicing-modulating antisense RNAs must accumulate in the nucleoplasm where splicing occurs (see chapter 42 Aartsma Rus). This is why derivatives of U small nuclear RNAs (snRNAs), and in particular of U7 snRNA, have been widely used for this purpose [1]. Apart from the advantage that the antisense RNA accumulates as part of a stable small nuclear ribonucleoprotein (snRNP), U7 snRNA expression cassettes, with their small size, will fit into all types of gene therapy vectors so that they can be efficiently targeted to many different tissues and cell types. Toxic side effects have not been observed, and the desired antisense effect should only be exerted in those cells expressing the targeted pre-mRNA. Thus, in contrast to gene replacement therapies, the target gene keeps its natural regulatory network in terms of temporal and spatial expression.

The U7 snRNP is a ribonucleoprotein complex specialized in 3' end processing of histone pre-mRNA (Fig. 1a) [2] (see chapter 2 and 5, Meister and Luhrmann for discusson of snRNPs). The first 18-20 nucleotides of the ~60 nucleotide-long U7 snRNA are complementary to a conserved histone downstream element located 3' of the histone pre-mRNA cleavage site. The next 11 nucleotides constitute a binding site for Sm and Sm-like proteins, which form a heptameric Sm core structure during the cytoplasmic maturation phase of the snRNP. The RNA ends in a relatively (what is relatively?) stable hairpin. A trimethyl guanosine cap structure at the 5' end and the Sm core is an are important signals for nuclear accumulation and, together with the 3' hairpin, serves to stabilise the snRNP particle. Importantly, the Sm binding sequence of U7 snRNA deviates from the consensus 5’-(A/G)AUUU(G/U)UG(G/A)-3’ found in spliceosomal snRNAs. It associates with five Sm proteins also found in spliceosomal snRNPs and two U7-specific Sm-like proteins, termed Lsm10 and Lsm11 (Lsm for like sm), which are essential for histone RNA processing [3, 4]. However, if the U7-specific Sm binding sequence is converted to the consensus found in spliceosomal snRNAs, a new U7 Sm OPT construct is obtained which binds all seven Sm proteins found in spliceosomal snRNAs (Fig. 1b). This particle can no longer induce the cleavage of the histone pre-mRNAs, but can still bind to them by RNA:RNA base pairing, thus becoming a competitive inhibitor for wild-type U7 snRNPs [2]. Another consequence is that the U7 Sm OPT RNA accumulates as nuclear snRNP about three times more efficiently than its wild-type counterpart. (does OPT has any meaning? Optimal?)

After having discovered these features, we thought U7 Sm OPT might be an ideal tool for splicing modulation. As U7 Sm OPT could inhibit histone processing by an antisense mechanism, it seemed likely that, by exchanging the natural sequence binding to histone pre-mRNA with one directed against a new target, one could turn U7 Sm OPT into a sequence-specific competitor for components of the splicing machinery or make it bind any RNA molecule found in the nucleoplasm (Fig. 1b). Its efficient accumulation in the nucleoplasm seemed to be another bonus for its action in splicing modulation.

These expectations concerning splicing modulation have been fully met in studies on U7-mediated exon skipping in a variety of genes and disorders including thalassemic -globin mutations [5-7], Duchenne Muscular Dystrophy [8-10] and HIV-1/AIDS [11, 12]. U7 tools have also been used to promote the inclusion of a poorly used exon in the case of Spinal Muscular Atrophy [13-15] (see also chapter 17a, b, Singh) or to block intronic cryptic 5' splice sites (ss) activated by mutation of the natural 5' ss [16].

2.2 Strategic considerations

To induce exon skipping, several types of sequences in and around the exon of interest can be targeted. Truly comparative results have been mostly obtained in tissue culture models for three-thalassemic mutations in the second intron of the human -globin gene that create 5' splice sites (IVS2-654, -705 and -745) and activate a common cryptic 3' ss further upstream in the same intron, resulting in the inclusion of an aberrant exon in -globin mRNA and in the loss of -globin protein production (Fig. 2A). In this system, U7 Sm OPT derivatives targeting the cryptic 3' ss or the 5' ss created by the mutations were able to induce exon skipping, whereas a U7 construct targeting the branch point upstream of the aberrant exon had no effect (Fig. 2B) [5, 6]. Based on theoretical considerations, a length of 17-20 nucleotides for the antisense sequences tested should be sufficient (what are these considerations, TM calculation?) Although sequences in this length range often work, we have found the optimal length for an antisense sequence targeting the cryptic 3' ss to be 24 nucleotides. It also seems to be important that the antisense sequence directly abutts abuts the Sm binding sequence, as the introduction of spacer nucleotides abolished exon skipping in several examples [6]. Compared to such single target U7 snRNAs, more efficient exon skipping can be obtained with U7 snRNA derivatives carrying two tandem antisense sequences, targeting sequences upstream and downstream of the targeted exon or partly overlapping with the 3' and 5' ss [6]; L. Angehrn, J. Marquis and D.S., unpublished results; Fig. 2B). This double-target approach has proven to be very effective as a means to induce exon skipping in the context of other transcription units such as the human [9] or mouse [8, 10] dystrophin gene, the cyclophilin A gene [11] or the multiply spliced HIV-1 transcripts [12]. This is why we suggest to begin a new exon skipping project by selecting such a combination of two antisense sequences of 18-20 nucleotides each. If it is then necessary to further optimise the strategy, one has to use different sequence combinations or use one of the exon-internal targeting strategies described below. Do you suggest to target putative exon enhancer or silencer or does it not make any difference?

Efficient skipping of the aberrant -globin exon has also been obtained when the U7 snRNA derivative carried a sequence complementary to exon-internal sequences (Fig. 2B) [7]. In this case, the U7 snRNP may have acted by masking exonic splicing enhancer (ESE) sequences or by affecting the flexibility of the exon. In our hands, a U7 Sm OPT derivative could induce skipping of exon 7 in the human survival of motoneuron 1 (SMN1) and SMN2 genes, even if it did not target a characterised ESE [14]. The exon skipping activity of a U7 Sm OPT derivative targeting exon-internal sequences may be additionally stimulated by adding a 5' end tail to the U7 RNA that can bind the splicing silencing protein hnRNP A1 as recently shown for exon 51 of the human dystrophin gene [17].

Two different approaches have been proposed to promote the inclusion of a weak exon. One approach, originally developed by Hertel and collaborators, is based on the idea that the weak exon and the subsequent exon compete for forming an active spliceosome with the 5' ss of the preceding exon (Fig. 3A). In the case of the SMN2 gene, whose poor exon 7 inclusion is responsible for the disease Spinal Muscular Atrophy (SMA), can you reword this: ..where inclusion of exon 7 is a therapeutic approach for the disease… becauseSMA is caused by the loss of SMN1, U7 constructs targeting the 3' ss of exon 8 were indeed able to improve exon 7 inclusion [13, 14]. The other approach, which, in our hands, proved to be more efficient, makes use of U7 snRNAs containing an antisense sequence targeting exon 7 as well as an additional splicing enhancer sequence, so that a splicing activator of the SR protein family gets recruited to the weak exon (Fig. 3B) [14]. We could recently demonstrate that improving SMN2 exon 7 use by this approach indeed has a therapeutic benefit and can even entirely cure a severe SMA phenotype, when the corresponding U7 expression cassette was introduced into a severe mouse model for SMA by germline transgenesis [15].

The recent demonstration that defects caused by the weakening of a 5' ss accompanied by activation of a cryptic 5' ss in the downstream intron can also be corrected by U7 snRNAs targeting the cryptic site [16] demonstrates that the U7 approach is not limited to exon skipping/inclusion decisions. Additionally, we point out that the approach could also be useful beyond the realm of splicing defects caused by mutation. Since the majority of human genes undergo alternative splicing [18] (see chapter 4 and 8 Hertel and Smith), U7 snRNA-based splicing modulators could also be used to affect some of these alternative splicing decisions in medically relevant situations such as neoplasias or diseases of the immune system.

2.3 Gene transfer and regulated expression

It is also important to consider in which system a splicing correcting U7 gene will be used. Luckily, due to the small size of the U7 expression cassette (~600 bp), it can be delivered to cells or tissues with many types of vector system.

Even if the ultimate application will necessitate a viral vector, we prefer to modify the functionally important sequences (antisense sequences with or without SR or hnRNP protein binding sequence) in the context of the original pSP64-derived U7 Sm OPT plasmid (protocol 1). The cassette can then be amplified by PCR with mutagenic primers containing restriction sites to allow for insertion into the vector system of choice (protocol 2).

For any cellular assay system that is amenable to DNA transfection by means of lipid agents or calcium phosphate precipitation (e.g. splicing reporter genes in HeLa cells or in other easily transfectable cell types), we directly introduce the pSP64-derived U7 Sm OPT plasmid by this route. The splicing reporter can either be a stable component of the cell genome or it can be co-transfected along with the modified U7 plasmid.

For cells in culture that are difficult to transfect with DNA refractory to DNA transfection techniques or if a stable integration of the U7 cassette into the cell genome is desired, we routinely use lentiviral transfer vectors (see chapter 14, gabelini) . However, as lentiviral vector technology is not established in all laboratories that may want to use U7-based splicing correction, we point out that it is also possible to cotransfect the modified U7 plasmid with a suitable selection marker. After the selection has been applied, individual surviving cell colonies must then be screened for cointegration of the U7 cassette. An example of such a selection for hygromycin resistance has been described in the first paper on U7-based splicing correction [5]. Stable cell lines can also be rapidly made using the flp recombination system described in chapter 38, Kjems).

Lentiviral vectors can also be used to transfer U7 cassettes to various cell types or to produce transgenic animals. Lentiviral vectors offer the ability of stable integration and hence prolonged expression of the U7 cassette in many cell types, including cells of the hematopoietic system [19]. Moreover, they can be injected into pre-implantation embryos to generate transgenic mice [20]. We have used this approach to introduce a splicing correction U7 cassette into mice and then transfer it by conventional breeding into mice of a severe SMA model [15]. However, it should be made clear that animals transgenic for U7 cassettes can also be produced by more broadly available techniques such as pronuclear injection of, for example, a simple U7 Sm OPT-derived plasmid.

For gene transfer and long-term expression in non-dividing cell types such as myofibers, neuronal cells or hepatocytes, the use of non-integrating vectors based on adenovirus associated virus (AAV) seems to be the ideal choice [21]. Impressive and very long-lasting results have been obtained with such a vector system with a double-target U7 construct able to induce skipping of exon 21 of the mouse dystrophin gene in the mdx mouse, a common animal model for Duchenne Muscular Dystrophy [10].

Even though the packaging capacity of AAV vectors is limited, the small U7 cassettes can easily be accommodated, even together with a marker gene for the monitoring or selection of transduced cells. It is also possible to introduce several U7 cassettes in tandem, e.g. to exploit a synergism between different cassettes or to obtain a higher expression due to the increased copy number of a single one. However, one should not attempt to incorporate tandem U7 cassettes into lentiviral vectors because of the high tendency of the viral reverse transcriptase for template switching, which means that all but one copy of a tandem repeat will be deleted during reverse transcription.

Important concerns for in vivo studies and especially for any therapeutic applications in humans are a potential toxicity and the risk of insertional mutagenesis. In this context we note that we have on two occasions generated transgenic mice with different U7 cassettes. One transgenesis experiment was carried out by pronuclear injection, the other by lentiviral vector transgenesis. In each of these events, mice were bred from multiple founder animals and followed over 5-10 generations. In none of these mice have we observed any kind of toxicity or increased occurrence of neoplasias. While some concerns regarding insertional mutagenesis with lentiviral vectors are still justified [22], this risk is minimal for AAV vectors because of their non-insertional behaviour. Furthermore, U snRNA promoters are not known to act as enhancers on nearby genes (do you have a reference for this?), which may translate into a lower risk of insertional mutagenesis if the remainder of the transfer vector is chosen carefully.

For studies in animal models to determine the development of an inherited disease as well as the therapeutic time window, it may be desirable to induce and repress the expression of a U7 cassette at will or to express it specifically in certain tissues or cell types. In this respect it is important to note that the transcription of U snRNA genes is fundamentally different from that of mRNAs [23]. Except for the U6 and U6 atac RNAs which are transcribed by RNA polymerase III, all other spliceosomal U snRNAs as well as U7 snRNA are transcribed by a specialised RNA polymerase II complex. This complex interacts with U snRNA-specific promoter sequences, which differ from those of mRNA promoters. Transcription is constitutive and terminates in a weakly conserved 3' box located downstream of the U snRNA coding sequence. Replacement of the promoter by a regulatable or tissue-specific mRNA promoter is not an option, as it results in a failure to recognise the 3' box.

However, by adapting a system for the regulated expression of short hairpin RNAs for RNA interference studies [24], we have recently been able to develop a doxycycline-inducible version of U7 expression cassettes [25]. This system requires two lentiviral vectors and relies on a doxycycline-sensitive version of the KRAB/KAP1 trancriptional silencing protein. Concerning tissue-specific expression systems, it may be possible to combine U7 genes with tissue-specific enhancers [8] or to engineer "floxed" U7 genes that can be activated or inactivated by tissue-specific Cre recombination .

Finally, although this has not been attempted so far, it may well be possible to adapt the U7-based splicing modulation technology to non-mammalian systems, including invertebrates and plants.

1. Protocols

2.1 Protocol 1: PCR-based introduction of functional sequences into U7 SmOPT

Generation of PCR primers for a desired U7 construct

The original U7 SmOPT vector [26] contains the murine U7 gene as a 570 bp HaeIII fragment inserted into the unique SmaI site of the pSP64 polylinker. The wild-type U7 Sm binding site has been converted to the Sm OPT sequence and single StuI and HpaI sites have been inserted upstream and downstream of the U7 snRNA sequence, respectively, all by site directed mutagenesis. The orientation of the U7 gene is against the Sp6 promoter so that riboprobes for RNAse protection assays can be generated by run-off transcription of StuI-linearised plasmids with SP6 RNA polymerase [6]. Figs. 4a and 4b show a scheme and the sequence of the regions relevant for cloning, respectively.