WO2012114111A1

WO2012114111A1 - Gene silencing

Info

Publication number: WO2012114111A1
Application number: PCT/GB2012/050410
Authority: WO
Inventors: Nicholas Jarvis Proudfoot; Monika GULLEROVA
Original assignee: Oxford University Innovation Ltd
Current assignee: Oxford University Innovation Ltd
Priority date: 2011-02-24
Filing date: 2012-02-23
Publication date: 2012-08-30
Anticipated expiration: 2013-08-24
Also published as: GB201103167D0

Abstract

The present invention provides a method of reducing or preventing the expression of a gene of interest comprising providing a DNA molecule having a sequence which comprises in a 5' to 3' direction (i) a Pol II promoter element in sense orientation, (ii) the gene of interest or a portion thereof comprising 40 nucleotides or more, and (iii) a Pol II promoter element in antisense orientation, and expressing the gene of interest incorporated into the DNA molecule in a mammalian cell. It also provides a method of reducing or preventing the expression of a gene of interest in a mammalian cell comprising integrating a Pol II promoter element in antisense orientation downstream (3') of the gene of interest and inducing expression from the antisense promoter. It also provides an in vitro method of preparing a pool of siRNA molecules specific for a gene of interest comprising providing a DNA molecule having a sequence which comprises in a 5' to 3' direction (i) a Pol II promoter element in sense orientation, (ii) the gene of interest or a portion thereof comprising 100 nucleotides or more, and (iii) a Pol II promoter element in antisense orientation, and transcribing the gene of interest in the presence of Dicer, preferably recombinant Dicer, to produce siRNA molecules.

Description

GENE SILENCING

Field of the invention

The present invention relates to a method of reducing or silencing expression of genes in mammalian cells.

Background of the invention

RNA interference (RNAi) was first identified as a process by which double strand (ds) RNA is introduced into cells/organisms which can result in the inhibition of genes containing homologous sequence. It is now appreciated that most eukaryotes display some facet of RNAi. RNAi appears to operate at two levels of gene expression inhibition. On one level the gene may be switched off by the induction of repressive chromatin structures, a process called transcriptional gene silencing (TGS). Alternatively the messenger RNA made by the gene may be blocked from being translated either through its direct degradation or through blocking its access to protein synthesis, a process called post transcriptional gene silencing (PTGS).

TGS appears to be exclusively used in some fungi such as fission yeast (S. pombe) as well as in plants. Previous studies in fission yeast have demonstrated that natural transcriptional gene silencing is a cis specific process. Some organisms use both TGS and PTGS mechanisms such as C. elegans and Drosophila. Mammals appear to use only PTGS mechanisms. Such post transcriptional gene silencing may be achieved using either exogenous siRNAs or endogenous micro RNAs. This type of gene silencing is transient in nature.

The exclusive use of PTGS in mammals has meant that experimental induction of RNAi has routinely employed small dsRNAs either synthetic or alternatively gene constructions that produce small hairpin RNAs. Small dsRNAs or hairpin RNA so introduced into cells by transfection are processed in the cytoplasm into microRNAs that then target specific mRNA inactivation. RNAi using longer dsRNAs is not believed to function in mammals because dsRNA will activate a nonspecific viral defence mechanism, the interferon response, which leads to an arrest of protein synthesis and nonspecific mRNA degradation in the affected cells (Tuschl, 2001). Interferons are a group of signalling molecules which are induced and secreted when cells are infected by R A viruses or exposed to dsR A. This activates a dsR A sensitive protein kinase called PKR that is known to inhibit cellular translation with resulting cell death. The interferon response is not sequence specific and PKR is potentially activated by any cytoplasmic dsRNA greater than 30 bp in length (Clemens et al, 1997; Cole, 2007).

Some studies on gene silencing have focused on dsRNA derived from convergent transcription. Gullerova and Proudfoot (2008) showed that convergent genes (CGs) in S. pombe display a regulated transcription termination process during the cell cycle. In Gl phase, CGs generate read-through transcripts and so form dsRNA. This induces transient heterochromatin through a nuclear RNAi pathway. However following the replication S phase, CGs no longer generate read-through transcription as heterochromatin recruited cohesin acts to block G2 read-through transcription so that in G2, CGs form shorter mRNA transcripts that do not overlap to form dsRNA. At the end of G2, mitosis results in cohesin degradation allowing sister chromatid segregation to new daughter cells, and consequent CG read-through transcription in the Gl phase of the new cell cycle. Even though TGS is not considered a viable gene silencing approach for mammalian cells, some studies described in the literature do employ convergent transcription to induce RNAi in other organisms. In trypanosomes, studies have employed artificial convergent transcription units comprising convergent bacteriophage T7 promoters flanking a gene sequence to be silenced that was then trans fected into the parasite. These trypanosomes were also engineered to express the T7 phage RNA polymerase and show significant gene silencing effects assumed to be PTGS (Alibu et al, 2005; Shi et al, 2000; Wang et al, 2000).

Giordano et al (2002) describe a study in Drosophila, where TGS is known to occur, in which convergent transcription was engineered by use of heterologous Gal4 regulated RNA polymerase II promoters (from budding yeast) placed convergently flanking a test gene sequence. Following transfection into flies, also expressing Gal4 transcription factor, significant gene silencing was observed and assumed to be of a PTGS nature. In a study in mammalian cell culture, Tran et al (2003) used a gene construct containing two U6 R A polymerase III convergent promoters (DualU6) to generate very short dsRNA less than 30 bps. They indicate that it is essential to use such short dsRNA to avoid activation of the cytoplasmic dsRNA interferon response which was predicted to cause general inhibition of longer dsRNA expressing cells. Significant gene silencing of target genes was again observed.

WO 01/77350 proposes the use of convergent transcription to inhibit gene expression particularly in cereal crops.

These previous studies show the potential for convergent transcription as a way to induce gene silencing. However, this technology has not been widely applied. In particular the practicality of this methodology is restricted by the complexity of arranging such transcription systems and the fact that the silencing observed is thought to be elicited through PTGS mechanisms.

Summary of the invention

Accordingly, as to a first aspect the invention provides a method of reducing or preventing the expression of a gene of interest comprising providing a DNA molecule having a sequence which comprises in a 5' to 3' direction (i) an RNA polymerase II (Pol II) promoter element in sense orientation, (ii) the gene of interest or a portion thereof comprising 40 nucleotides or more in length, and (iii) a Pol II promoter element in antisense orientation, and expressing the gene of interest incorporated into the DNA molecule in a mammalian cell, for example in a mammal or in a mammalian expression system. Surprisingly this method does not induce an interferon response. Advantageously this method results in a long-lasting gene silencing effect, both of the gene in the expressed construct (cis) and also of any endogenous copy of the same gene present in the host cell genome (trans). Furthermore, convergent transcription constructs are cheap and easy to make and use, such that this method may have wide- ranging applicability. The Pol II promoter elements (i) and (iii) may be the same or different. Preferably the Pol II promoter element selected is either a highly active viral or composite promoter. Preferred promoters are the CMV promoter and the SV40 late promoter. It may also be useful in some applications to have a regulatable promoter or one that is only active in particular tissues. For example an inducible system may be used using one of the above-mentioned promoters in combination with one of tetracycline (tet) regulatable, Doxycycline or oestrogen.

Preferably the gene of interest or the portion thereof comprises 50 nucleotides or more, more preferably 60, 70, 80, 90, 100, 150, 200 or 500 nucleotides or more, and most preferably a substantial part or the full sequence of the gene. By a substantial part we mean at least about 50% of the gene sequence, more preferably at least about 60%, 70%), 80%o, 90%), 95%o or 99% of the gene sequence. Using a longer sequence will provide more specificity to the gene silencing effect, while still avoiding any interferon response.

Preferably the DNA molecule does not contain termination sequences associated with the gene of interest in either the sense or antisense orientations. The expression system may be selected from a range of tissue culture cells, for example any human primary cell line, or cells isolated from patients, cancer cell lines, and stem cells, for example Hela cells, 293T cells, CHO cells, HEP G2 cells, HEK293, cos7 cells. The mammal may be a human, cow, pig or sheep. Preferably expression of the gene of interest is reduced or prevented for at least about 72 hours, more preferably at least about 96 hours or at least about 120 hours. Such a long-lasting effect has not previously been demonstrated using RNAi techniques such as siRNA or shRNA. Preferably expression of the gene of interest is decreased at least 10-fold.

The gene of interest may be selected as one which it is desirable to silence for therapeutic purposes or may be one which it is desirable to silence in order to investigate its function. Preferably the gene of interest is one or more genes selected from viral genes, for example from HIV, herpes or CMV, such as env, gag, pol, tat, rev, nef or vif; genes required for the exosome to function efficiently, such as yeast Rrp6 or any homolog thereof, or yeast Rrp44 or any homolog thereof; genes required for the proteosome to function efficiently, such as ubiquitine activating enzyme, ubiquitine conjugating enzymes or ubiquitine ligase; a region of translocation associated with a cancer, for example all or part of the c-myc fusion protein on chromosome 8; 15-200 trinucleotide repeats specific for a particular trinucleotide repeat disorder, preferably 15-100 trinucleotide repeats, for example 15-100 repeats of the sequence -CAG-. According to a second aspect the invention provides a DNA molecule having a sequence which comprises in a 5' to 3 ' direction (i) a Pol II promoter element in sense orientation, (ii) a gene of interest or a portion thereof comprising 40 nucleotides or more, and (iii) a Pol II promoter element in antisense orientation, for use in reducing or preventing the expression of the gene of interest in a mammalian cell, for example in a mammal or a mammalian expression system. This aspect of the invention has the advantages set out above. Furthermore, preferred features set out above and herein are also preferable in this aspect of the invention.

According to a further aspect the invention provides the use of a DNA molecule having a sequence which comprises in a 5' to 3' direction (i) a Pol II promoter element in sense orientation, (ii) a gene of interest or a portion thereof comprising 40 nucleotides or more, and (iii) a Pol II promoter element in antisense orientation, for reducing or preventing the expression of the gene of interest in a mammalian cell, for example in a mammal or a mammalian expression system.

According to yet a further aspect the invention provides a DNA molecule having a sequence which comprises in a 5' to 3 ' direction (i) a Pol II promoter element in sense orientation, (ii) a gene of interest or a portion thereof comprising 40 nucleotides or more, and (iii) a Pol II promoter element in antisense orientation, for use in therapy.

According to a yet a further aspect the invention provides the use of a DNA molecule having a sequence which comprises in a 5' to 3' direction (i) a Pol II promoter element in sense orientation, (ii) a gene of interest or a portion thereof comprising 40 nucleotides or more, and (iii) a Pol II promoter element in antisense orientation, for the manufacture of a medicament for the treatment of cancer or a trinucleotide repeat disorder or a viral infection. Preferred features set out above and herein are also preferable in these aspects of the invention.

According to a further aspect the invention provides a method of reducing or preventing the expression of a gene of interest in a mammalian cell comprising integrating a Pol II promoter element in antisense orientation downstream (3') of the gene of interest and inducing expression from the antisense promoter. Preferably this method is an in vitro method. This aspect of the invention has the advantages set out above. Additionally, this approach may be simpler from the cloning point of view and one can integrate such antisense promoters behind any gene or genes in a cell. Furthermore, preferred features set out above and herein are also preferable in this aspect of the invention.

In a further aspect the invention provides an in vitro method of preparing a pool of siRNA molecules specific for a gene of interest comprising:

- providing a DNA molecule having a sequence which comprises in a 5 ' to 3 ' direction (i) a Pol II promoter element in sense orientation, (ii) the gene of interest or a portion thereof comprising 100 nucleotides or more, and (iii) a Pol II promoter element in antisense orientation,

- transcribing the gene of interest in the presence of Dicer, preferably recombinant Dicer, to produce siRNA molecules.

Preferably the transcription is achieved using nuclear extract or in presence of active Pol II. Advantageously, this provides a simple and cheap in vitro method to prepare a pool of siRNA specific to a gene of interest and which may be used to reduce or prevent expression of the gene of interest. Furthermore the invention provides a method of reducing or preventing the expression of a gene of interest in a mammalian cell comprising transfecting the cell with siR A molecules produced according to the above method. In another aspect the invention provides siRNA molecules produced according to the above method, for use in therapy, as well as the use of siRNA molecules produced according to the above method, for the manufacture of a medicament for the treatment of cancer or a trinucleotide repeat disorder or a viral infection. Preferred features set out above and herein are also preferable in these aspects of the invention.

Brief description of the drawings

Figure 1 shows data indicating that convergent genes in plasmids induce transcriptional cis silencing.

Figure 2 shows data indicating that plasmid derived convergent genes induce transcriptional trans silencing.

Figure 3 shows data indicating that convergent transcription induces trans silencing. Figure 4 shows data indicating that trans silencing of essential genes is induced by convergent transcription.

Figure 5 shows data indicating that convergent transcription induces trans silencing of mammalian genes.

Figure 6 shows data indicating that transcriptional silencing induced by the CT cassette leads to more efficient inhibition of test gene expression than siRNA or shRNA treatment.

Figure 7 shows a model of induced transcriptional silencing in trans.

Figure 8 shows data indicating that convergent transcription induces trans silencing of mammalian genes.

Figure 9 shows examples of other cell lines and other target genes.

Figure 10 shows in vitro experiments proving that Dicer interacts with pol II, dsRNA formation derived from CT plasmid and siRNA formation derived from CT plasmid. Figure 11 shows gene silencing effect of CT, sense, antisense and sense+antisense transcription. Only CT induces most efficient TGS. Full gene inserted in CT plasmid is more efficient to induce TGS than exon or intron only. CT without terminators (polyA signals) is more efficient in inducing TGS.

Figure 12 shows CT induced TGS on TDP43 gene.

Figure 13 shows limited spreading of CT induced heterochromatin and long lasting effect.

Figure 14 shows different CT gene constructs used. Detailed description of the invention

The invention is based on the approach of inducing nuclear TGS either by transfecting long convergent transcription units into mammalian cell nuclei or by integrating an antisense promoter downstream of an endogenous target gene in a mammalian cell. By this approach, since the dsRNAs produced as a result of convergent transcription remain in the nuclei, activation of the interferon dsR A response is avoided. The benefits of this procedure are that convergent transcription constructs are cheap and easy to make and use. Furthermore TGS so induced produces a longer term gene silencing effect unlike more transient PTGS. In particular, it is shown herein in both S. pombe and mammalian cells that convergent transcription can be utilized as a simple and effective way to mediate long term TGS. It is shown that such convergent transcription systems may have substantial advantage over currently employed gene silencing procedures.

These findings may have wide ranging implications for in vivo and in vitro protein production, which may be reduced or prevented by positioning a gene of interest between convergent promoter elements, or by positioning a Pol II promoter element in antisense orientation downstream of an endogenous gene of interest. This is in contrast to current processes for silencing genes of interest in mammalian cells, which use dsRNA molecules comprising only a small fragment of the gene of interest. The invention may be particularly valuable because any gene can be silenced in accordance with the invention and so this provides a novel and simple method to investigate the function of any gene in a cell or mammal. Furthermore, the invention may be used therapeutically to treat or prevent a disease by silencing one or more genes that are contributing to or causing the disease. According to one aspect the invention provides a method of reducing or preventing the expression of a gene of interest comprising providing a DNA molecule having a sequence which comprises in a 5' to 3 ' direction (i) a Pol II promoter element in sense orientation, (ii) the gene of interest or a portion thereof comprising 40 nucleotides or more, and (iii) a Pol II promoter element in antisense orientation, and expressing the gene of interest incorporated into the DNA molecule in a mammalian cell.

According to another aspect the invention provides a DNA molecule having a sequence which comprises in a 5' to 3' direction (i) a Pol II promoter element in sense orientation, (ii) a gene of interest or a portion thereof comprising 40 nucleotides or more, and (iii) a Pol II promoter element in antisense orientation, for use in reducing or preventing the expression of the gene of interest in a mammalian cell.

According to a further aspect the invention provides the use of a DNA molecule having a sequence which comprises in a 5' to 3' direction (i) a Pol II promoter element in sense orientation, (ii) a gene of interest or a portion thereof comprising 40 nucleotides or more, and (iii) a Pol II promoter element in antisense orientation, for reducing or preventing the expression of the gene of interest in a mammalian cell. According to yet a further aspect the invention provides a method of reducing or preventing the expression of a gene of interest in a mammalian cell comprising integrating a Pol II promoter element in antisense orientation downstream (3') of the gene of interest and inducing expression from the antisense promoter. Any RNA polymerase II (Pol II) promoter may be used in accordance with the invention. Preferably the Pol II promoter element selected is either a highly active viral or composite promoter. Preferred promoters are the CMV promoter and the SV40 late promoter. It may also be useful in some applications to have a regulatable promoter or one that is only active in particular tissues. For example an inducible system may be used using one of the above-mentioned promoters in combination with one of tetracycline (tet) regulatable, Doxycycline or oestrogen. The same Pol II promoter may be used in (i) and (iii) or alternatively different promoter combinations may also be used effectively. In one embodiment it is preferred that promoters (i) and (iii) are inducible as this will allow the control of target gene (ii) at anytime, for example by adding the appropriate drug. When a Pol II promoter is integrated directly downstream of an endogenous gene of interest then the Pol II promoter is preferably a regulatable promoter or one that is only active in particular tissues. For example an inducible system may be used using one of the above-mentioned promoters (e.g. a highly active viral or composite promoter such as the CMV promoter or SV40 late promoter) in combination with one of tetracycline (tet) regulatable, Doxycycline or oestrogen.

The gene of interest may be any gene which it is desired to reduce or prevent expression from. The gene of interest may be a single gene or may be 2 or more genes, for example 3, 4 or 5 genes, or portions thereof. For example, it may be desired to investigate the function of a particular protein in the cell and this may be done in accordance with the invention by reducing or preventing expression from the gene encoding that protein. This may be useful in discovering how different proteins interact, in learning about different signalling pathways and in learning which proteins are required for a particular end result. Alternatively, one or more different genes from the same pathway, for example 2, 3 or 4 genes, may be cloned in sequence between (i) and (iii) which may provide a maximal effect on the pathway. For example one or more HIV genes, such as gag, pol, env, tat, rev, nef and vif, may be cloned in sequence between (i) and (iii) in order to target HIV.

There is no requirement for the convergent transcription units described herein to contain any termination sequences associated with the gene of interest. Therefore, preferably the DNA molecules described herein have a sequence which consists essentially of in a 5' to 3' direction (i) a Pol II promoter element in sense orientation, (ii) a gene of interest or a portion thereof comprising 40 nucleotides or more, and (iii) a Pol II promoter element in antisense orientation. Termination sequences provide polyadenylation of the mRNA and/or transcriptional termination signal. If desired one or more termination sequences may be associated with the gene of interest in the sense orientation and/or one or more termination sequences may be associated with the gene of interest in the antisense orientation, but preferably no termination sequences are associated with the gene of interest in either orientation. It may also be desired to reduce or prevent the expression of a gene for a therapeutic purpose, for example by gene therapy of a patient. For example a particular gene may be known to be overexpressed or expressed anomalously in a particular disease condition. In such a case the gene may be silenced in accordance with the invention in order to provide a therapeutic effect. An example of this might be the expression of a particular gene in cancer cells. Silencing of this gene may be used as a way to treat or prevent the spread of the cancer (see for example Lotem et al. PNAS 2005 vol. 102 no. 51, 18556-18561). Preferably the gene of interest is one or more genes selected from viral genes, for example from HIV, herpes or CMV, such as env, gag, pol, tat, rev, nef or vif; genes required for the exosome to function efficiently, such as yeast Rrp6 or any homolog thereof, or yeast Rrp44 or any homolog thereof; genes required for the proteosome to function efficiently, such as ubiquitine activating enzyme, ubiquitine conjugating enzymes or ubiquitine ligase; a region of translocation associated with a cancer, for example all or part of the c-myc fusion protein on chromosome 8; 15-200 trinucleotide repeats specific for a particular trinucleotide repeat disorder, preferably 15-100 trinucleotide repeats, for example 15-100 repeats of the sequence -CAG-. These embodiments are described in more detail below.

As an example of this, chromosome translocations, i.e. abnormal rearrangement between non-homologous chromosomes, are common trigger in cancer. Such a translocation can lead to the fusion of normally separated genes, if the joining of chromosomes happens in coding regions. For example, translocation of c-myc on chromosome 8 causes a fusion protein, which gives lymphocyte proliferation ability. Such occurrence of fusion proteins can be fatal for cell and lead to various phenotypes, typical of cancer cells. For example it has been described that several leukemias are caused by acquired translocations. The present invention could be used to silence the regions of translocations and so kill the cancer cells, thereby allowing normal cells (which do not have regions of translocation) to proliferate. Thus in accordance with the invention the gene of interest may comprise a region of translocation associated with cancer, such regions being known to the skilled person. In particular the gene of interest may comprise the whole or part of the c-myc fusion protein on chromosome 8. Other translocations which may be targeted may be found for example in Igbokwe and Lopez-Terrada (Molecular Testing of Solid Tumors. Archives of Pathology & Laboratory Medicine: January 2011, Vol. 135, No. 1, pp. 67-82) and Sandberg and Avery (Cancer Genetics and Cytogenetics, 2003 (2010) 102-126).

In other embodiments it may be desirable that the gene of interest encodes a protein comprising a growth hormone, a clotting factor, a viral antigen, an antibody, an ion channel or an enzyme. In general, convergent transcription in accordance with the invention can be used with advantage in any case where the value of an organism or cell line to commerce, including agriculture, is determined by the level of expression of one of its natural genes or of an artificially introduced gene. For example, the invention could be used to treat, prevent or reduce viral infection by targeting one or more viral genes, for example from HIV-1, Herpes or CMV, for example env, gag, pol, tat, rev, nef or vif

The invention could also be used to increase the yield of transgene expression by inactivating the exosome (RNA level) or the proteasome (protein level) by targeting genes required for these complexes to function efficiently. Specific genes which may be targeted include any homolog of yeast Rrp6 or rrp44, in order to inactivate the exosome. In order to inactivate the proteasome, many genes may be targeted, for example ubiquitine activating enzyme, a ubiquitine conjugating enzyme or ubiquitine ligase. In accordance with the invention the gene of interest is cloned into a DNA molecule sandwiched between convergent Pol II promoters. Preferably the whole coding sequence of the gene is included in the DNA construct. Alternatively it is possible to use only a portion of the sequence of gene, preferably at least about 40 nucleotides or more, more preferably about 50 60, 70, 80, 90, 100, 125, 150, 200, 250, 500 or 1000 nucleotides. Typically the maximum length of gene insert is about 10 kb as beyond this size gene cloning and cell transfection are less efficient. In other embodiments the maximum length of gene insert may be about 5kb or about 2 kb. The DNA molecule may be made by any method known to the skilled person (such as restriction digestion/PCR/ligation). For example standard cloning techniques may be used as described in Molecular Cloning: A Laboratory Manual (Maniatis et al., published by Cold Spring Harbor Laboratory Press). There is no requirement for the gene to be in frame with either of the promoter sequences. The gene is preferably located adjacent to the promoter but there may be a portion of non-coding sequence between the promoter and the gene provided that the promoter is able to direct expression of the gene. For example the portion of non-coding sequence between the promoter and the gene may be up to 2kb.

Preferably the DNA molecule is isolated, in the sense that the invention is not intended to encompass naturally-occurring DNA molecules that include convergent Pol II promoters. The DNA molecule is usually in the form of an expression vector. An expression vector is any vector capable of expressing those DNA sequences contained therein which are operably linked to other sequences capable of effecting their expression. An example of an expression vector is a plasmid. Once the DNA molecule has been produced, it can be amplified and used in a suitable expression system, as is well known to the skilled person. The DNA construct may be introduced into a mammalian cell or a mammalian cell culture stably or transiently where the DNA sequence is expressed by transcription and translation. In the case of stable expression the DNA must be replicable in the host either as an episome or as an integral part of the chromosomal DNA. There are various methods of introducing foreign DNA into a eukaryotic cell: some rely on physical treatment (electroporation, nanoparticles, magnetofection), other on chemical materials or biological particles (viruses) that are used as carriers. Any of these methods, or others, may be used as known to the skilled person. For example the construct may be transfected into the cell using calcium phosphate, by electroporation, or using liposomes.

Transfection into the mammalian cell may be transient or stable. Transient transfection/ expression of the convergent gene transcription construct may be suitable in order to achieve short or mid term gene silencing applications (for example in the region of 3, 4, 5, 6 or 7 days). For long term gene silencing (for example longer than 7 days, more preferably longer than 14 or 28 days) stable transfection is preferred and this may be conveniently achieved by stable integration of the DNA construct into the host cell chromosomes. The addition of a selectable marker gene on the DNA construct (such as neomycin) will allow for the selection of stable transfected cells that will be permanently silenced at the target genes. Alternatively, use of inducible/repressible promoters would allow to switch integrated CT construct on or off, just by adding an appropriate drug, with no need of transfection.

Expression from the gene of interest can be induced in the usual way and this will depend on the type of promoter element(s) used. Expression from the gene of interest will result in the production of dsRNA molecules. Since transcription occurs from both the sense and antisense promoters, sense and antisense RNA molecules are produced which then hybridise together to form dsRNA molecules. These dsRNA molecules then act to suppress further transcription from the gene, both from the construct and any other copy of the gene found in the cell, by heterochromatin modification and by other RNAi pathways. Any mammalian cell or expression system may be used in accordance with the invention. For example the mammal may be a human, cow, pig, sheep or chicken. Preferred expression systems are human and mammalian tissues and cell lines, for example any human primary cell line, or cells isolated from patients, cancer cell lines, and stem cells, for example Hela cells, 293T cells, CHO cells, HEP G2 cells, HEK293, cos7 cells. The mammal may be a human, cow, pig, sheep or chicken.

The constructs described herein can be used in any situation where it is desired to reduce or prevent the expression of a gene. Yields of mRNA and protein can be quantified by numerous techniques known to the skilled person, including real-time PCR, northern blot, RNAse protection and SI nuclease analysis for mRNA yields, and Western blot for protein yields. In accordance with one aspect of the invention a Pol II promoter element is integrated in antisense orientation downstream (3') of a gene of interest. In this case it is preferred that the Pol II promoter element is an inducible element, such as TetO. The promoter element may be integrated downstream of a gene of interest using standard site specific genome integration using homology regions for recombination, as is well known to the skilled person. This approach may be used to reduce or prevent the expression of one or more genes of interest. The method may be used to create a stable cell line with an inducible knock down system. When the term "reducing or preventing gene expression" is used herein, this refers to a decrease in the mRNA and/or protein levels which are observed when a particular gene is expressed in a particular expression system from a DNA molecule in accordance with the invention, as compared to expression of the same gene in the same expression system which does not contain a DNA molecule in accordance with the invention. Thus, the DNA molecule induces a trans effect on any copy of the gene found in the cells of the expression system. The term 'silence' is also used herein and is intended to have the same meaning as 'reduce or prevent'. In the aspect of the invention in which a Pol II promoter element is integrated in antisense orientation downstream (3') of a gene of interest, then the term "reducing or preventing gene expression" is used to refer to a decrease in the mRNA and/or protein levels which are observed when transcription of the gene of interest is induced from the antisense promoter in accordance with the invention, as compared to when there is no transcription of the gene of interest from the antisense promoter. Although the level of silencing is likely to vary between genes, preferably, according to the invention expression of the gene of interest is reduced by from about 2-fold to about 10-fold or more. Most preferably the gene is silenced entirely in that no protein encoded by the gene is produced or is capable of detection. Advantageously, the amount of nuclear mRNA and cytoplasmic mRNA produced according to the invention is at least 2- to 3 -fold lower, preferably at least 10-fold lower, than the amount produced in the absence of providing such a DNA construct. For example, the amount of nuclear mRNA and cytoplasmic mRNA produced according to the invention may be from 2-fold to 20-fold lower, or from 4-fold to 12- fold lower, than previous methods. In contrast to siR A methods in which only the amount of cytoplasmic mRNA is reduced, the amounts of both nuclear mRNA and cytoplasmic mRNA are reduced in accordance with the invention. Advantageously, by down-regulating amount of nuclear mRNA, mRNA in the cytoplasm is also down- regulated.

Advantageously, the amount of protein produced according to the invention is at least 2- to 3-fold lower, preferably at least 10-fold lower, than the amount produced in the absence of providing such a DNA construct. For example, the amount of protein produced according to the invention may be from 2-fold to 20-fold lower, or from 4- fold to 12-fold lower, than previous methods. Most preferably the gene is silenced entirely in that no protein encoded by the gene is produced. This may be determined by quantitating mRNA levels, from situations plus the DNA construct according to the invention compared with situations minus the DNA construct according to the invention, using RT-PCR (or other such techniques known to the skilled person). Following this, a western blot may be performed to detect the target protein produced from the two situations (plus and minus the DNA molecule).

The gene silencing effect brought about in accordance with the invention may vary and it likely to last at least about 72 hours, more preferably at least about 96 hours or at least about 120 hours. Furthermore the gene silencing effect may be permanent. The present invention may be used to reduce or prevent the expression of genes integrated into chromosomal locations in cells by providing the same gene, or a portion thereof, contained in an expression vector, such as a plasmid, in vitro. Because reduced protein production can be achieved by simply inserting the gene of interest between convergent promoter sequences, this technique is an incredibly cheap and easy technology to implement and requires nothing more than cloning techniques. Further, no alterations in the coding portion of the gene are required. Furthermore the present invention may be used to reduce or prevent the expression of genes integrated into chromosomal locations in cells by integrating a Pol II promoter element in antisense orientation downstream of the selected genes. Because reduced protein production can be achieved by simply integrating a promoter element downstream the gene of interest, this technique is an incredibly cheap and easy technology to implement and requires nothing more than cloning techniques. Further, no alterations in the coding portion of the gene are required.

A further great advantage of the invention is for the production of stable cell lines, where promoters (i) and (iii) are inducible and are integrated into a chromosome thereby allowing the silencing of target gene (ii) at anytime, just by adding an appropriate drug. Alternatively the same effect may be achieved if the sense promoter is constitutive and the antisense promoter is inducible. Furthermore the invention provides therapeutic uses. Thus the invention also provides a DNA molecule having a sequence which comprises in a 5' to 3 ' direction (i) a Pol II promoter element in sense orientation, (ii) a gene of interest or a portion thereof comprising 40 (preferably 100) nucleotides or more, and (iii) the Pol II promoter element in antisense orientation, for use in therapy.

Further the invention provides the use of a DNA molecule having a sequence which comprises in a 5' to 3' direction (i) a Pol II promoter element in sense orientation, (ii) a gene of interest or a portion thereof comprising 40 (preferably 100) nucleotides or more, and (iii) a Pol II promoter element in antisense orientation, for the manufacture of a medicament for the treatment of cancer or a trinucleotide repeat disorder or a viral infection. Any type of cancer may be treated as discussed above, and in particular those cancers which have a translocation event as an early indicator. Similarly any viral infection may be treated, for example HIV, herpes, CMV. The preferred features of the DNA molecules mentioned above also apply to this method and use. The type of therapy or disease to be treated or prevented will depend on the selected gene of interest. For example, the gene of interest may encode a protein, for example interferon protein, insulin, a growth hormone, a clotting factor, a viral antigen, an antibody or an enzyme, which has a known effect on a particular disease and the expression of said gene may be decreased or prevented in accordance with the invention in order to prevent or treat said disease. In particular, the invention may have an application in gene therapy of a patient where it is desired to decrease or prevent the expression of a particular protein which is produced excessively by the patient. This technology could also be useful in gene therapy of a patient where it is desired to decrease or prevent the expression of a particular protein where excessive production of certain protein (gene expression) is unwanted in the cell. Examples of this approach are discussed above.

The inventive approach could also be used for treatment of other trinucleotide repeat disorders in a corresponding manner, for example myotonic dystrophy, myoclonic epilepsy, spinocerebellar ataxias or Friedreich's ataxia.

For example, this technology could be used is in the treatment of Huntington's chorea. This neurodegenerative genetic disorder affects muscle coordination and leads to cognitive decline and dementia. The Huntingtin gene normally encodes a protein called Huntingtin. The mutation of the Huntingtin gene leads to the production of different forms of the protein, which cause gradual damage to specific areas of the brain. The Huntingtin gene contains repeated sections called a trinucleotide repeat (...(CAG)_n...). These repeats vary in length. When the length of this repeated section reaches a certain threshold, a mutant form of the protein is produced. The normal number of trinucleotide repeats is less than 28. An intermediate phenotype is caused by 28-35 repeats and full disease is developed in cells with more than 40 repeats. The present invention could be used to target excessive repeats by incorporating 15-100 trinucleotide repeats as the gene of interest and so decrease effect of the disease.

Other trinucleotide repeat disorders which may be treated in accordance with the invention are described by Orr and Zoghbi (Annual Review of Neuroscience Vol. 30: 575-621), and are set out in the tables below (split into polyQ and non-polyQ disorders). In each case the treatment strategy would be similar to that discussed above for Huntington's chorea and the skilled person would readily be able to determine such a suitable strategy. For example the strategy may be to include 15-200 trinucleotide repeats specific for a particular trinucleotide repeat disorder, preferably 15-100 trinucleotide repeats, as the gene of interest.

Polyglutamine (PolyQ) Diseases

Non-Polyglutamine Diseases

Normal/

Type Gene Codon Pathogenic wildtype

FRAXA (Fragile X FMR1, on the X-

CGG 6 - 53 230+ syndrome) chromosome

FXTAS (Fragile X- FMR1, on the X-

CGG 6 - 53 55-200 associated tremor/ataxia chromosome syndrome)

FRAXE (Fragile XE AFF2 or FMR2, on the

GCC 6 - 35 200+ mental retardation) X-chromosome

FRDA (Friedreich's

FXN or X25, (frataxin) GAA 7 - 34 100+ ataxia)

DM (Myotonic

DMPK CTG 5 - 37 50+ dystrophy)

SCA8 (Spinocerebellar

OSCA or SCA8 CTG 16 - 37 110 - 250 ataxia Type 8)

nnn

SCA12 (Spinocerebellar

PPP2R2B or SCA12 On 5^* 7 - 28 66 - 78 ataxia Type 12)

end

Where it is desired to reduce or prevent the expression of a gene in order to treat a disease, such as those discussed above, then the invention also provides such treatment by integrating a Pol II promoter element in antisense orientation downstream (3') of the endogenous gene and inducing expression from the antisense promoter. The Pol II promoter is preferably a regulatable promoter or one that is only active in particular tissues. For example an inducible system may be used using e.g. a highly active viral or composite promoter, such as the CMV promoter or SV40 late promoter, in combination with one of tetracycline (tet) regulatable, Doxycycline or oestrogen.

In a further aspect the invention provides an in vitro method of preparing a pool of siR A molecules specific for a gene of interest comprising:

- providing a DNA molecule having a sequence which comprises in a 5 ' to 3 ' direction (i) a Pol II promoter element in sense orientation, (ii) the gene of interest or a portion thereof comprising 40 nucleotides or more, and (iii) a Pol II promoter element in antisense orientation,

- transcribing the gene of interest in the presence of Dicer to produce siRNA molecules.

The description and preferred features of the DNA molecule set out hereinabove apply also to this aspect of the invention. The method may be carried out in the presence of a nuclear extract or purified RNA polymerase II to provide the required transcription. If purified Pol II is used then the skilled person will know to include all essential components for efficient transcriptions, like transcription factors, dNTPs etc. Dicer present in nuclear extract may be sufficient but it is preferable to use additional recombinant Dicer to provide for a more efficient reaction. Reaction conditions can be readily determined by those of skill in the art.

Transcription of RNA from the gene of interest will result in the production dsRNA molecules. Since transcription occurs from both the sense and antisense promoters, sense and antisense RNA molecules are produced which then hybridise together to form double-stranded RNA (dsRNA) molecules. These dsRNA molecules are then cleaved by Dicer, an endoribonuclease in the RNase III family, to form a pool of short double-stranded RNA fragments called small interfering RNA (siRNA) molecules, each about 20-25 nucleotides in length.

These siRNA molecules may be used to reduce or prevent the expression of a gene of interest in a mammalian cell. The siRNA molecules may first be isolated, for example by standard RNA isolation methods (e.g. phenol/chloroform or Trizol). The cell may then be transfected with the siRNA molecules using any technique known to the skilled person, for example Lipofectamine 2000. In another aspect the invention provides siRNA molecules produced according to the above method, for use in therapy, which may be prophylactic or therapeutic. The invention also provides the use of siRNA molecules produced according to the above method, for the manufacture of a medicament for the treatment of cancer or a trinucleotide repeat disorder or a viral infection. The details discussed above in relation to therapies, including the diseases that may be treated and the genes of interest, are relevant to these aspects also.

Examples

The following examples are illustrative of the products and methods of making the same falling within the scope of the present invention. They are not to be considered in any way limitative of the invention. Changes and modifications can be made with respect to the invention. That is, the skilled person will recognise many possible variations in these examples covering a wide range of compositions, ingredients, processing methods, and mixtures, and can adjust the naturally occurring levels of the compounds of the invention for a variety of applications.

The examples follow a description of the figures which are referred to in which the examples.

Figure 1 shows:

A. Plasmid design. pYL16 (Werner Bio Agents), containing resistance gene for nourseothricin, was used as vector. Further ura4 or ade6 genes with endogenous promoters and terminators sequences were inserted into pYL16 single (S). Additionally ura4 and ade6 were cloned into the vector in tandem (T) or convergent orientation (C) as depicted.

B. Pol II ChIP analysis of A2 S. pombe (ura4-, ade6-) transformed with plasmids containing ura4 and/or ade6 genes in single (S), tandem (T) or convergent (C) orientation. Cells containing empty plasmid, called vector (V), were used as controls. First graph shows the levels of Pol II detected over ura4. Note that vector cells show background levels, as there is no endogenous or episomal ura4 gene. Second and third graphs show Pol II levels over ade6 and actl genes. Error bars represent standard deviation and were calculated from three independent repeats.

C. qRT-PCR analysis of total RNA isolated from samples as in B. Levels of sense and antisense ura4 mR A are shown as fold over cells containing empty vector.

D. H3K9me3 ChIP analysis. Levels of H3K9me3 analysed in samples (as in B) and probed in ura4 (first graph), ade6 (second graph) and centromeric dg (third graph). Bars represent the levels of H3K9me3 as % of input. Error bars as in B.

E. Growth of transformed A2 cells onto non- and selective plates. Cells were diluted serially and spotted on plates from left to right. V empty vector, S single ura4 or ade6, T ura4 and ade6 in tandem orientation, C ura4 and ade6 in convergent orientation. Darker color on -ade plates represent less of adenine in cells. Figure 2 shows:

A. Pol II ChIP analysis of 972- S. pombe (ura4+, ade6+) transformed with plasmids containing ura4 and/or ade6 genes in single, tandem or convergent orientation, as in Figure IB. B. qRT-PCR analysis of total RNA isolated from samples as in A. Levels of sense and antisense ura4 mRNA are shown as fold over cells containing empty vector.

C. 3' RACE analysis of mRNA isolated from cells as in A. Reverse transcription was performed using phased oligo(dT) with linker sequence, followed with PCR using forward primer positioned in ORF of tested gene (ura4, ade6 or act!) and reverse primer specific to linker. PCR products were analysed on 1.2% agarose gels.

D. H3K9me3 ChIP analysis. Levels of H3K9me3 analysed in samples as in A and probed in ura4 (left graph), ade6 (centre graph) and centromeric dg (right graph). Bars represent the levels of H3K9me3 as % of input. Error bars as in Figure ID.

E. Growth of transformed 972- cells on non- and selective plates. Cells were diluted serially and spotted on plates from left to right. V- empty vector, S - single ura4 or ade6, T - ura4 and ade6 in tandem orientation, C - ura4 and ade6 in convergent orientation.

F. As in E. C plasmid was transformed into Adcrl and Aagol strains. Wt C and wt V plasmids were used as controls.

Figure 3 shows:

A. Design of convergent transcription cassette (CT). pYL16 vector was used as plasmid backbone, nmtl promoter was cloned in sense and antisense orientation within the multicloning site. ORFs of ura4 or ade6 were cloned between the convergent nmtl promoters as depicted in diagram.

B. qRT-PCR analysis of total RNA isolated from 972- cells transformed with pYL16 vector, ura4 or ade6 single genes in pYL16 vector, and ura4 or ade6 in CT cassette. Cells were growing in EMM (CT cassette on) or YES (CT cassette off) media until log phase. PCR performed done with primers specific for ura4 or ade6.

C. Growth of transformed 972- cells on non- and selective plates. Cells were diluted serially and spotted on plates from left to right. V empty vector, S single ura4 or ade6, T ura4 and ade6 in tandem orientation, C ura4 and ade6 in convergent orientation and CT ura4 or ade6 sequence in CT cassette.

Figure 4 shows:

A. qRT-PCR analysis of total RNA isolated from 972- cells transformed with pYL16 vector or rad21 in CT cassette. Cells were growing in EMM (CT cassette on) or YES (CT cassette off) media until log phase. RT primers were specific for sense (left graph) and antisense (right graph) rad21 mRNA. PCR was carried out with primers specific to rad21.

B. qRT-PCR analysis as in A. RT primers were specific for sense (left graph) and antisense (right graph) esol mRNA. PCR performed with primers specific for esol.

C. Comparison of cell growth on YES and EMM plates. Cells were serially diluted and spotted on plates from top to bottom. Plates were incubated at 25°C and 37°C, as shown in photographs. Cells containing rad21 or esol sequence in the CT cassette were compared to wt, rad21 and esol temperature sensitive mutants, as indicated.

Figure 5 shows:

A. Schematic representation of yACTl in pCI expression plasmid transcribed by sense and antisense CMV promoters. Gene structure of yACTl is shown below with grey blocks representing exons and lines introns. Black line below gene diagram corresponds to primer probes used in ChIP and RT-PCR experiments.

B. Pol II ChIP: chromatin was isolated from HeLa cells transfected with either pCI plasmid alone (vector) or with pCI containing yACTl gene fragment (exon 4) between CMV promoters (CTyACTl). Untransfected HeLa cells were used as a control. PCR was performed using yACTl exon 4 primers and primers for GAPDH.

C. qRT-PCR analysis. Total RNA was isolated from the HeLa cells as in B. RT was performed using oligo(dT). Probe in yACTl exon 4 was amplified and quantified. Bars represent levels of PCR signal normalized to signals from untransfected cells.

D. H3K9me3 ChIP. As in B.

E. Western blot analysis of yACTl protein levels in cells transfected as in B. Total proteins were extracted from samples taken 24, 48 and 72 hours after transfection, separated on denaturing gels and probed with anti- yACTl antibody (Sigma). Signals on films were quantified using AIDA software and normalized to untransfected cells, set as 100%.

F. H3K9me2 ChIP as in D. Chromatin was isolated from cells 96, 168 and 264 hours after transfection.

Figure 6 shows: A. qRT-PCR analysis using yACTl reverse primer for reverse transcription, probing for yACTl sense mRNA levels (left graph). Total RNA was isolated from HeLa cells transfected with siRNA (specific for yACTl gene, ThermoFishers), shRNA (specific for yACTl, Sigma) or a yACTl fragment convergently expressed from pCI plasmid (CTyACTl), 48, 72 and 96 hours after transfection. Untransfected cells were used as a control. Levels of sense mRNA signals were normalized to untransfected cells, which were set as 1. Levels of GAPDH mRNA were used as a control (right graph).

B. qRT-PCR analysis as in A, except using yACTl forward primer for reverse transcription to detect yACTl antisense transcripts (left graph) and GAPDH forward primer detecting antisense potential transcripts from GAPDH gene (right graph).

C. Western blot analysis of yACTl protein levels as in Fig 5E. Total proteins were extracted from cells 48, 72 and 96 hours after transfection. Tubulin was used as aloading control (upper band on the gel).

D. H3K9me3 ChIP as in Fig 5D. Chromatin was isolated from cells transfected with siRNA, shRNA or convergently expressed yACTl (CTyACTl), 48 and 96 hours after transfection.

Figure 7 shows:

Model of induced transcriptional silencing in trans. Episomal sense and antisense transcription from gene of interest, leads to production of dsRNA. This is recognized and processed by Dicer, a component of the RNAi machinery, resulting in the formation of siRNA complementary to the gene of interest. siRNA allows the RNAi machinery to recognize an endogenous allele of the gene of interest which leads to its transcriptional silencing in trans, by heterochromatin establishment.

Figure 8 shows:

Convergent transcription (CT) plasmids induce TGS of endogenous γ-actin gene.

A. Diagram depicting CT plasmid containing -ACTl cDNA (exons 2-6).

B. Pol II (N20X antibody) chromatin immunoprecipitation (ChIP) analysis of γ-

ACT1 or GAPDH control using amplicons specific for endogenous gene.

Chromatin isolated from HeLa cells was transfected with CT vector alone (V),

CT yACTl or untransfected (UN). C. Measurement of mRNA levels using oligo dT primed qRT/PCR by from HeLa cells as in B.

D. H3K9me3 antibody ChIP as in B.

E. Western blot analysis and quantitation of γ-actin protein levels compared to tubulin and OAS1 following V and CT yACTl transfections for 1-3 days.

F. Effect of CT yACTl transfection of ES cells, wt or DDCR1 on nascent transcript levels from y-ACTl versus GAPDH.

Figure 9 shows:

Generality of CT induced coTGS.

A. qRT/PCR analysis of polyA+ mRNA levels in HeLa or HEK293 cells transfected with CY ACTl versus empty vector (V).

B. Effect of CTyACTl transfection of ES cells, wt or DDCR1 on polyA+ mRNA levels from y-ACTl versus GAPDH.

C. Effect of CTCYPA transfections on TGS of endogenous genes by Pol II ChIP using either exon or intron PCR primers. Note that Pol II levels were generally higher in exons than introns.

D. Effect of CTPGKl transfections on TGS of endogenous genes by Pol II ChIP using either exon or intron PCR primers. Note that Pol II levels were generally higher in exons than introns.

E. Effect of CI GAPDH transfections on TGS of endogenous genes by Pol II ChIP using either exon or intron PCR primers. Note that Pol II levels were generally higher in exons than introns.

F. As for C-E but measuring reduction in polyA+ mRNA levels.

Figure 10 shows:

In vitro analysis of nuclear dicer effects.

A. Western blot analysis of whole cell extract (WE) versus nuclear extract (NE) with or without dicer immunodepletion (NE-dicer). Hsp70 is only detected in WE confirming purity of NE.

B. Co-immunoprecipitation of dicer and Pol II shown by western blots revealing that two proteins cross interact. Pol IIO has phospho CTD while Pol IIA has non-phospho CTD. C. NE dependent in vitro transcription of CTyACTl Ex4 and vector alone plus control run off template (yielding 363 nt R A). ³²P labelled RNAs were treated with either SI (single strand specific) or VI (double strand specific) nucleases and RNAs fractionated. Separate panel shows that dicer depleted NE still yields CT derives dsRNA.

D. Fractionation of small RNAs isolated from templates as indicated. S denotes single promoter construct (making just sense transcript while CT denotes TyACTl Ex2-6. CTT is as in Fig. 11D. Only CT and CTT yield detectible siRNAs (denoted by arrow) but not in dicer depleted NE.

Figure 11 shows:

Specificity of coTGS induction by CTyACTl trans fection.

A. H3K9me2 ChIP was performed as in Fig. 8 but using additional y-ACTl (Ex2-6) expression constructs: sense (S) or antisense (AS) alone y-ACTl transcription plasmids or S and AS plasmids cotransfected. Note only CT plasmid is effective in inducing full levels of heterochromatm and maximum reduction in mRNA levels.

B. Actin mRNA qRT/PCR was performed as in Fig. 8 but using additional y-ACTl (Ex2-6) expression constructs: sense (S) or antisense (AS) alone y-ACTl transcription plasmids or S and AS plasmids cotransfected. Note only CT plasmid is effective in inducing full levels of heterochromatm and maximum reduction in mRNA levels.

C. CT constructs containing y-ACTl intron 3 or exon 4 only were compared for TGS effects versus the full CT yACT!Ex2-6 sequence measuring polyA+ mRNA and nascent (intronic) transcript.

D. CT yACT!Ex2-6 plasmid was modified by positioning PAS derived from SV40 at either end of the y-ACTl sequence so that both S and AS transcripts are polyadenylated. The coTGS effects were then determined for both nascent and steady state y-ACTl transcripts by qRT/PCR analysis.

Figure 12 shows:

CoTGS of TDP-43. A. TDP-43 gene map is depicted with exons as boxes and introns as lines. Positions of PCR amlpicons used are indicated. Diagram depicting TDP-43 insert into CT plasmid is also presented.

B. Pol II Chip analysis over endogenous TDP-43 exon-intron boundaries following CTTDP43 transfection is shown versus GAPDH control.

C. Western blot analysis of TDP-43 protein levels versus tubulin and OAS1 is presented for Hela cell transfected with CTTDP43 versus empty vector (V) control. Both gels and quantitative analysis are presented.

D. Levels of polyA+ mR A from HeLa cells transfected with CTTDP43 versus empty vector (V) control determined by qRT/PCR using PCR amplicons indicated.

Figure 13 shows:

Extent of coTGS effects.

A. The extent of CTTDP43 induced gene silencing across TDP-43 was measured by the extent of H3K9me2 ChIP analysis. This showed that heterochromatin spreading was restricted to region directly targeted by the CT construct. TDP-43 gene map is shown below with exons as boxes and introns as lines. Positions of PCR amlpicons used are indicated.

B. coTGS effects (Pol II and H3K9me2 ChIP) were measured over longer time courses for CTgACTl trans fections.

C. Model depicting the mechanism of nuclear coTGS. Chromosome and plasmid transcription are indicated. Active Pol II (black ellipses) transcription is depicted by arrows in CT construct and inactive transcription is depicted in chromosome (cross through Pol II). Nuclear RNAi apparatus (in grey ellipse) and nuclear dicer (scissors) are indicated.

Figure 14 shows:

Diagrams showing different CT gene constructs used in this study. Name, insert sequence and size of insert are tabulated. Arrows denote either CMV promoter or SV40 polyA signal (PAS) positions and orientations. Inserted exons or introns are indicated in boxes. Example 1 - Convergent genes in plasmids induce transcriptional cis silencing.

It has previously been shown that transcription of endogenous convergent genes (CG) in S. pombe fails to terminate following the proximal polyA site in the Gl phase of the cell cycle. Consequently, R A polymerase II (Pol II) reads through the downstream convergent gene resulting in the production of long complementary transcripts, which form dsR A. This is recognized by Dicer and processed into siR A. Activated RNA interference then leads to heterochromatin establishment in Gl, which is removed following the S phase. Significantly in G2 cohesin that has been recruited to Gl phase CG heterochromatin blocks read-through transcription. Thus CGs remain euchromatic through the most of the G2 phase until the loss of cohesin in M phase allows renewed CG read-through transcription in the next Gl phase of the cell cycle (Gullerova and Proudfoot, 2008).

To test whether CG heterochromatin is dependent on chromosomal structure or is a more general feature of convergent transcription, the pYL16 expression plasmid was employed, carrying a resistance gene to nourseothricin antibiotic. ura4 and ade6 genes with their own promoter and terminator regions were inserted into pYL16 as single or double gene constructs. In the later case, ura4 and ade6 genes, were positioned in either tandem or convergent orientation (Figure 1A). These plasmids were then transformed into S. pombe A2 strain, which lacks both ura4 and ade6. Positive transformants were selected on plates with nourseothricin.

First, Pol II levels over ura4 or ade6 in A2 cells carrying either vector only (V), plasmids with single ura4 or ade6 (S), tandem ura4-ade6 (T) or convergent ura4-ade6 (C) gene pairs were analysed (Figure 1A). V cells were used as a negative control, since they lack endogenous ura4 and ade6. Pol II signals at ura4 or ade6 in V cells were at background levels. Cells containing plasmids with single ura4 or ade6 (S) gave significant Pol II signal over either ura4 or ade6 genes, confirming the plasmid based expression of these genes. Also tandem arranged ura4 and ade6 on plasmid (T) result in similar Pol II levels over ura4 and ade6. This suggests that the expression of tandem arranged genes is efficient. Significantly, Pol II levels over convergent ura4 and ade6 were significantly decreased (C) (Figure IB). This result suggests that transcription of CGs shows similar behaviour regardless of their endogenous or episomal context. It should be noted that CGs on plasmids appear not to be regulated during the cell cycle, as the Pol II ChIP experiments described here were performed on cycling cells, which are predominantly in G2 phase of the cell cycle. Presumably cohesin is not recruited to convergent genes on plasmids so that they can produce dsR A throughout the cell cycle. This will result in stable heterochromatin formation. Levels of Pol II over endogenous actl were also analyzed and as predicted are unaffected by V, S, T or C plasmids transformation. qRT-PCR was performed using reverse or forward ura4 primers for reverse transcription, to detect steady state levels of either sense or antisense ura4 transcripts (Figure 1C). This experiment revealed high levels of ura4 sense mR A in S and T transformed cells, but only low levels in C transformed cells. Cells carrying V were used as a negative control. These data correlate with the tested Pol II levels (Figure IB) and show decreased sense ura4 mRNA production, when this gene is in convergent arrangement. Significantly, antisense ura4 RNA can be detected only in cells transformed with C plasmid, confirming that transcription of the convergent downstream ade6 reads through ura4. These data resemble the previous analysis of endogenous CGs in the Gl phase of cell cycle (Gullerova and Proudfoot, 2008). This previously led to the demonstration that sense and antisense transcription of CGs form dsRNA and consequent RNAi activation followed by heterochromatin establishment. Therefore H3K9me3 ChIP analysis on V, S, T and C transformed cells was performed and high H3K9me3 levels were detected only in C cells, indicative of heterochromatin formation. A centromeric dg repeat probe was used as a positive control (Figure ID). The effect of decreased ura4 and ade6 mRNAs on cell growth in selective media lacking uracil or adenine was finally tested. Since A2 cells are deficient in endogenous ura4 or ade6, in selective media, they rely on plasmid genes expression. Serial dilutions of V, S, T and C cells were prepared and spotted on non-selective, ura- selective or ade- selective plates. Cells were incubated at 32°C for three days (Figure IE). All strains grew on non-selective media. However, C containing cells clearly grew less well on selective media as compared to S or T strains. This suggests that lower levels of ura4 or ade6 mRNAs and heterochromatin marks respectively, reflect down regulation of these genes and so have an effect on cell growth in selective media. Taken together, the data presented in Figure 1, demonstrate CGs transcriptional read- through in this plasmid context resulting in RNAi activation and heterochromatin formation in cis. Such silencing effects are clearly dependent on gene orientation, as ura4 and ade6 lack this phenotype when arranged in tandem.

Example 2 - Plasmid derived convergent genes induce transcriptional trans silencing.

The fact that CGs on plasmids generate the dsRNA that activate the RNAi pathway, led to considering the possibility that siRNAs generated from plasmid CGs (ura4 or ade6), could also lead to silencing of the endogenous gene in trans. The V, S, T and C constructs were therefore used to transform 972- strain, which is a prototroph and so contains endogenous ura4 and ade6 genes. Pol II ChIP analysis was again performed using probes in either ura4 or ade6 ORFs (Figure 2A). Significant Pol II levels were detected in V cells, which reflect Pol II occupancy at endogenous ura4 or ade6. Pol II levels in S or T cells were somewhat higher, corresponding to the sum of Pol II at the endogenous alleles plus additional copies of plasmid ura4 or ade6. Surprisingly, a substantial reduction in Pol II occupancy was observed with C cells, when compared to V cells. Pol II levels over actl were also analyzed, and no significant level changes were observed between V, S, T and C cells. This suggests that the effect of plasmid expression is specific to the tested genes and has no general effect on transcription. These data predict that siRNAs produced from plasmids have a trans effect on expression of their endogenous allele.

Next, qRT-PCR was performed to analyze steady state ura4 sense and antisense mRNA levels (Figure 2B). As expected, sense ura4 mRNA was detected in V, S and T cells, corresponding to transcription of ura4 from both endogenous and plasmid. However, sense ura4 mRNA was substantially reduced in C strain, which correlates with above tested Pol 11 levels. In contrast, ura4 antisense mRNA could be detected only in C cells. This result again suggests a direct link between gene orientation and antisense transcription. To confirm that observed antisense transcription derives from read-through transcription of plasmid CGs, 3 'RACE analysis was performed (Figure 2C). Phased oligo d(T) with a linker sequence was used to prime transcription, followed be PCR using a reverse primer specific for the linker sequence and a forward primer in each ORF. By this approach all polyadenylated mRNAs produced could be detected. 3 'RACE analysis using total RNA isolated from V, S and T cells, resulted in only one major band, corresponding to mRNA with proximal polyA site usage for both tested ura4 and ade6. However, larger bands were detected when analyzing RNA from C cells. These correspond to long polyadenylated mRNAs, using more distal cryptic polyA sites, presumably derived from the plasmid CGs. actl was also tested as a control and only one band was observed in all analyzed samples corresponding to proximal polyA site usage. These results confirm that transcriptional read-through occurs on both DNA strands when the genes are in convergent arrangement. Further, it was tested whether CG induced trans silencing occurs at the transcriptional or mRNA level. H3K9me3 ChIP was performed on chromatin isolated from V, S, T and C strains, probing in either ura4 or ade6 ORFs (Figure 2D). As expected no positive H3K9me3 signals were detected on ura4 and ade6 (either endogenous or exogenous) with V, S or T chromatin. However, clear H3K9me3 signal was obtained for both ura4 and ade6 chromatin isolated from C cells. As a positive control centromeric dg repeat chromatin was tested, which as expected gave equivalent H3K9me3 signals on all tested chromatin samples. It is concluded from these data that plasmid CG expression induces heterochromatin establishment over the endogenous alleles of the same genes, when transformed plasmids contain convergent gene conformation.

Heterochromatin and gene silencing should induce a growth defect phenotype, when cells are grown on selective plates. Serial dilutions of V, S, T and C strain cells were therefore performed on either non-selective media (to test viability) or on media lacking uracil or adenine (Figure 2E). All strains grew similarly on non-selective plates. In contrast, C cells show a clear growth defect on plates without uracil or adenine. It is concluded that expression of ura4 and ade6 as plasmid CGs inhibits endogenous gene expression in trans through heterochromatin induction. This results in a defective growth phenotype, when the cells are grown in selective media.

It is inferred from our above experiments that CG transcription induces gene silencing both in cis and trans through R Ai. The involvement of major R Ai factors in this process was therefore tested. Cells lacking Dicer (dcrl) or Argonaute (agol) were transformed with C plasmid to determine if a growth defect occurred on uracil lacking media (Figure 2F). Wt cells transformed with V and C were used as controls. As shown a growth defect was again observed in C cells. However, this phenotype was lost in strains lacking dcrl or agol. These results confirm the requirement of RNAi factors in C induced gene silencing.

In summary, it was observed that plasmid expression of ura4 or ade6 has no effect on endogenous transcription, when these genes are present singly or in tandem arrangement. However, plasmid expression of ura4 and ade6 in convergent orientation blocks the expression of the endogenous alleles of these two genes. It was concluded that transcriptional read-through on both DNA strands leads to production of long dsRNA molecules, which is processed by Dicer into siRNAs. This induces RNAi and leads to silencing of endogenous copies of tested genes through heterochromatin formation in trans. Such gene silencing leads to less ura4 or ade6 mRNA, causing a growth defect, when C cells are grown on selective media.

Example 3 - Convergent transcription induces trans silencing

It was examined whether the observed effects of CGs on transcription are dependent on the convergent gene arrangement per se, or a more general feature of convergent transcription. Different pYL16 plasmids were designed by cloning the inducible/repressible nmtl promoter sequence in both sense and antisense orientation into this plasmid, with multi cloning site (MCS) placed between them (CT cassette = Convergent Transcription cassette; Figure 3A). This allowed cloning of the gene sequence of interest into the MCS and the simultaneous induction of its sense and antisense transcription. To test this system, ura4 or ade6 sequences inserted into the CT cassette were used and the resultant plasmids were transformed into 972- cells. First, qRT-PCR analysis was performed on steady state ura4 or ade6 mRNA levels under conditions where convergent transcription (CT) is switched on or off (Figure 3B). V, S or C strains were grown in either EMM (CT on) or YES medium (CT off). Total RNA was isolated and reverse transcribed followed by PCR amplification to detect sense ura4 or ade6 mRNA levels. All tested strains yielded positive sense mRNA signals when grown in YES medium. V and S cells showed similar sense mRNA levels for both ura4 and ade6, when cells were grown in EMM medium with the CT cassette switched on. In contrast, ura4 or ade6 sense mRNA levels were clearly substantially reduced in CT cells when CT cassette was switched on, suggesting that convergent transcription from the CT plasmid leads to endogenous gene silencing.

Next, the growth phenotype of previously tested V, S, T and C strains was analyzed as compared to CT (Figure 3C). Serial dilutions were plated on nonselective plates and incubated for three to four days as before. No visible growth defects were observed confirming the viability of all the tested strains. As shown above, V, S and T cells also grow on plates lacking uracil or adenine without visible growth defect. In contrast C cells were growth restricted due to the reduction of ura4 or ade6 mRNA levels. Interestingly, CT cells grew even less well than C cells on selective plates, suggesting that convergent transcription itself is efficient in promoting silencing in trans. It should be noted that the nmtl promoter is stronger than ura4 or ade6 promoters and will therefore induce a higher transcriptional rate in sense and antisense direction. This may cause the production of more siRNA from CT than C plasmids and so be more efficient and faster in the induction of transcriptional gene silencing in trans. Example 4 - Trans silencing of essential genes induced by convergent transcription

The analysis of essential genes in S. pombe is often an essential aspect of basic research. One commonly used approach is the depletion of essential proteins from the cell by use of temperature sensitive mutants (ts). However, the temperature shift itself, can be stressful to a cell and so induces heat shock phenotypes independently of the gene under test. For example, it has been shown previously that higher temperature perturbs the RNAi pathway and siRNA production. The methodology of siRNA induced knockdowns, used in higher eukaryotes, has proved unsuccessful in S. pombe. Transcriptional silencing in trans, induced by plasmid convergent transcription, as shown in our studies could therefore offer a new experimental system to knock down essential genes. Essential rad21 was therefore cloned into above described CT cassette. Rad21 is a subunit of the cohesin protein complex, which is involved in many regulatory mechanisms, such as gene expression, morphogenesis, cell proliferation, DNA repair and chromosome segregation. CTrad21 cells were grown in EMM and YES media. Total RNA was isolated and analysed by qRT-PCR (Figure 4A). Similar levels of sense rad21 mRNA were detected in V and CTrad21 cells, when the cassette was switched off. However when switched on a strong reduction (10% of V signal) of rad21 sense mRNA levels in CTrad21 cells was observed. The same approach was used to detect the antisense rad21 mRNA levels. As expected, no increase above background levels was observed in V and CTrad21 cells, when grown in YES medium. Instead a positive antisense rad21 mRNA signal in CTrad21 cells was detectd when the CT cassette is on. These results suggest that rad21 in the CT cassette results in rad21 sense and antisense transcription, causing a strong negative effect on total sense rad21 mRNA levels. Next, the same experiment was performed using the essential esol as another test gene (Figure 4B). V and CTesol cells show similar esol sense mRNA levels, when cells were grown in YES medium. However a strong reduction in sense esol mRNA signal was observed when CTesol cassette is switched on. Positive levels of antisense esol mRNA were again detect under these conditions. Overall these results confirm the data in Examples 1 to 3 by use of two other essential genes: rad21 and esol. Also, using the CT cassette it was possible to manipulate mRNA levels of these two essential genes, by simply growing cells in either YES or EMM media, to switch the CT cassette either on or off. To finally confirm the above data by a physiological assay, the growth phenotype of CTrad21 and CTesol cells was looked at. The CT cassette efficiency was also compared to commonly used ts mutants of rad21 and esol. Serial dilutions of wt, CTrad21, rad21ts, CTesol and esolts cells were plated on either YES or EMM plates and incubated at permissive temperature (25°C) for four days (Figure 4C). Additional plates containing serially diluted drops of wt, rad21ts and esolts cells were incubated at restrictive temperature (37°C). All tested strains were viable on YES plate (CT cassette is off) at 25°C. However, only CTrad21 and CTesol cells show visible growth defects on EMM plate, when the CT cassette is switched on. This effect of the induced CT cassette is comparable to ts mutation at restrictive temperature. These results confirm the effect of convergent transcription in transcriptional silencing in trans. The CT cassette offers a new and efficient system to decrease mR A levels of essential genes. This can be achieved without the necessity to grow cells at higher temperatures that could otherwise induce confusing side effects.

Example 5 - Convergent transcription induces trans silencing of mammalian genes

The above studies on gene silencing induced by convergent transcription were all carried out in S. pombe. Next the effect of a CT cassette in mammalian cells was tested. The pCI mammalian expression vector with its CMV promoter was employed, and another CMV promoter was added in antisense orientation following the MCS (Figure 5A), creating a mammalian CT cassette plasmid. Furthermore, a fragment of the γ actin gene (yACTl) was cloned into the MCS to test for potential yACTl silencing. A standard transfection of HeLa cells with CTyACTl plasmid was then performed.

Pol II ChIP was performed on chromatin isolated from untransfected HeLa cells (UN), HeLa cells transfected with empty vector (V) or cells transfected with CTyACTl (CT). PCR primers were designed to amplify either yACTl exon 4 or GAPDH as a control, similar Pol II occupancy was observed over in yACTl in UN and V cells, but only low Pol II levels in CT cells (Figure 5B). Pol II levels over GAPDH were similar in all three samples, excluding a general effect of CT on transcription. Next, qRT-PCR was used to detect yACTl and GAPDH mRNA levels in UN, V and CT cells (Figure 5C). The same PCR primers (with the antisense primer also used for RT) were employed as in Figure 5B. UN, V and CT cells show similar GAPDH mRNA levels, confirming the Pol II ChIP results. yACTl mRNA levels were similar in UN and V cells, but substantially decreased (10 fold) in CT cells. From the above S. pombe data, it was determined that CT plasmids induce gene silencing by establishment of heterochromatin over the endogenous allele of the test gene. It was therefore next tested whether heterochromatin was similarly induced by CT plasmids in HeLa cells. H3K9me3 ChIP analysis on UN, V and CT cells with yACTl and GAPDH primers revealed no positive H3K9me3 signals for the GAPDH gene in all tested samples. Similarly no signal was detected for yACTl in UN and V cells. However, a strong positive signal was demonstrated over yACTl in CT cells (Figure 5D). γ actin protein level reduction following CT transfection was finally investigated. Total proteins were isolated from UN, V and CT cells after 24, 48 and 72 hours of transfection and analysed by Western blot using a specific γ actin antibody (Sigma). The blot was visualized by radioautography and quantified using ImageQuant software. Signals were normalized to the UN signal, set at 100%. Levels of γ actin in UN and V cells were similar at all time points. However with CT transfected cells, a decrease in γ actin protein levels was observed after 24 hours of transfection, with further reductions at 48 and 72 hours post transfection (Figure 5E).

It should be noted that γ actin is an essential gene. Transfection of HeLa cells with CT construct leads to silencing of the endogenous yACTl allele and consequent cell death. It is therefore difficult to collect cells at later time points. Even so it was desired to define how long the CT induced heterochromatin mark persists and so transfected HeLa cells with CT cassette but maintained their survival by splitting them every three days. Aliquots of cells were selected at 96, 168 and 264 hours post transfection and subjected to ChIP analysis with antibody to H3K9me2. High levels of H3K9me3 were detected 96 hours post transfection. This suggests that persistance of the heterochromatin mark is independent of plasmid in the cell, since transient transfections are known to loose non-selected plasmids by about 48 hours post transfection. Low H3K9me2 levels were still visible even 168 hours post transfection, but only H3K9me2 background levels at 264 hours post transfection (Figure 5F). Since transfection efficiency is usually between 60-80%, splitting cells after transfection leads to dilution of positive transfectants. Transfected cells will be disadvantaged as compared to the untransfected cell population. Even so we can conclude that the CT cassette in HeLa cells leads to heterochromatin establishment and consequent transcriptional silencing. Down-regulation of Pol II, mRNA and protein levels of the essential yACTl gene has a strong effect on cell viability leading to cell death within 100 hours post transfection. Also the heterochromatin mark established by the CT cassette has a longer persistence than transient transfection. Example 6 - Transcriptional silencing induced by the CT cassette leads to more efficient inhibition of test gene expression than siRNA or shRNA treatment

Knock down of mammalian gene expression is a commonly used procedure. The usual approach is to design specific siRNA or shRNA and transfect these reagents into mammalian cells. The efficiency of such knock down experiments varies, depending on the expression of the tested gene. Usually a "two hit" transfection is experiment is necessary to achieve a significant knock down effect.

In the next series of experiments it was desired to compare the efficiency of the CT cassette to siRNA or shRNA in the down-regulation of the test gene. To this end commercial yACTl specific siRNA and shRNA was obtained. For transfection of HeLa cells 10 μg of each (siRNA, shRNA and CT) was used per 10 cm³ plate with cells at 60-70% confluency and Lipofectamine was then employed in a one hit transfection protocol. Aliquots were taken at 48, 72 and 96 hours post transfection, total RNA was isolated using Trizol and then analyzed by qRT-PCR. First, sense mRNA levels were tested from the tested yACTl and GAPDH (control) genes. The signals obtained were normalized to mRNA levels from untransfected cells, set as 1 (Figure 6A). A significant decrease of sense yACTl mRNA levels was observed in cells treated with siRNA or shRNA after 72 hours. This effect was lost at the next 96 hours time point, confirming only a transient effect. Interestingly, CT transfected cells show lower levels of yACTl mRNA 48 hours after transfection and even lower levels at the later time points of 72 and 96 hours post transfection (Figure 6A, left graph). GAPDH mRNA signals were similar in all samples at all time points, excluding a general non-specific silencing effect (Figure 6A, right graph). These data show that the silencing effect induced by CT is greater and longer lasting than transient siRNA and shRNA induced down-regulation of mRNA levels. Antisense yACTl transcription was also analyzed (Figure 6B). GAPDH antisense RNA was at background levels in all samples at all time points (Figure 6B, right graph). Similarly only background antisense yACTl transcript levels were detected in siRNA and shRNA treated cells (Figure 6B, left graph). In contrast, positive antisense yACTl transcripts were detected in cells treated with CT, at all time points, reaching the highest level at 72 hours post transfection (Figure 6B, left graph). These results show that only the CT cassette construct generates antisense transcripts of the tested gene.

Next, the down-regulation effect on γ actin protein levels was analyzed (Figure 6C). The same transfection approach was used as in the previous experiment, but isolated total protein from cells at 48, 72 and 96 hours post transfection, followed by Western blot with anti-actin antibody (as in Figure 5E) or anti-tubulin antibody to provide loading controls. Gels were quantified using ImageQuant software and signals were normalized to those obtained from UN cells, set at 100%. Tubulin levels were constant in all samples at all time points, confirming equal loading (Figure 6C, upper band). Levels of γ actin show a moderate, but visible decrease in cells treated with siRNA and shRNA (lower bands, second and third lanes on gel and graph in Figure 6C). Protein levels of γ actin (last lane on each gel) in CT treated cells were significantly reduced as compared to all other samples, suggesting that CT has a stronger effect on gene silencing than siRNA and shRNA. Finally, H3K9me3 levels were tested in cells treated with siRNA, shRNA or CT (Figure 6D). ChIP experiments were performed using chromatin isolated from cells 48 and 96 hours post transfection. Significantly a positive H3K9me3 signal was detect only in cells treated with CT at both time points. Comparison of CT to siRNA and shRNA reveals that CT induced yACTl gene silencing is significantly more efficient. This therefore defines a new approach to stably down-regulate mammalian gene expression. Overall these studies establish that convergent transcription leads to the induction of transcriptional gene silencing in S. pombe, through heterochromatin establishment with consequent Pol II reduction and decreased mRNA levels. A simple plasmid based cassette has been created, containing two promoters in convergent orientation with a gene of interest cloned between them, thereby stimulating sense and antisense transcription of the gene of interest. Expression of this gene cassette leads to silencing in cis and trans (Figure 7). This system has many applications. In particular, two essential genes, rad21 and esol, were tested and levels of down-regulation were achieved comparable to more commonly used temperature sensitive mutants. The major advantage of this system is that it provides a new tool for gene silencing that will work under physiological conditions.

It is also shown that transcriptional gene silencing induced by a convergent gene cassette works in higher eukaryotes; human HeLa cells. These data demonstrate that transfection of a CT cassette, containing gene sequence of interest, leads to its transcriptional silencing in trans through R Ai and heterochromatin establishment. To date, experimental mammalian gene silencing has relied on post transcriptional effects through mRNA inhibition or degradation by use of transfected siRNA or RNA hairpin gene constructs. Transcriptional gene silencing has remained an enigmatic process that is thought to be less experimentally tractable. Herein is described a simple and versatile method to achieve more permanent transcriptional gene silencing in mammalian cells.

Example 7 - Convergent transcription induces trans silencing of mammalian genes

The experimental system is to insert test gene fragments between convergent CMV promoters so generating convergent transcription (CT) plasmids (Fig. 8A). Following CT transient transfection into either HeLa or HEK293 cells we observed reproducible TGS (Fig. 8 and 9). Since transient transfections cannot transfect all cells in the population the levels of transcriptional gene silencing (TGS) observed are considered to be highly significant. With a CT plasmid containing γ-actin gene sequence, (CTyACTl) endogenous gene expression was reduced about 3 fold in terms of levels of Pol II on the gene and mRNA output (Fig. 8B, C). Note that the CT plasmids employed in these experiments (except in Fig 11D) do not contain polyA signals (PAS) so mRNA levels detected can only derive from endogenous genes. These results proved general as CT constructs containing 3 other gene sequences (from CYPA, PGK1 and GAPDH) each gave similar endogenous gene silencing effects at both Pol II and mRNA levels (Fig. 9C-F). Importantly robust heterochromatic marks (H3K9me3) were also detected over endogenous y-ACTl following CT transfection (Fig. 8D). Note that primers used in this ChIP analysis will not detect γ-ACTl sequence on the CT plasmid. A further 2 fold reduction in γ-actin protein levels was detected after 3 days transfection (Fig. 8E). Since γ-actin is a stable and essential protein, a larger effect was not anticipated. It was finally wished to establish that the gene silencing effect is RNAi associated. CTyACTl was therefore transfected into ES cells lacking dicer expression. Although transfection efficiencies for these cells is lower than for HeLa a 2 fold γ-ACTl gene silencing effect of both nascent and steady state RNA was still detected. Significantly dicer knock out ES cells (ADCRl) lost this silencing effect (Fig. 8F and 9B). It is concluded from the above data that CT plasmid expression induces TGS of endogenous target genes through an RNAi mechanism.

Example 8 - use of CT plasmid transfection can induce effective coTGS

The above experiments imply that CT expression induces TGS in mammalian cells. However studies to date on mammalian gene silencing by RNAi have invariably involved either endogenous microRNAs or exogenous siRNA, both acting through post transcriptional gene silencing (PTGS) mechanisms that target cytoplasmic mRNA. These have underlined the cytoplasmic nature of dicer action, even though several reports point towards additional nuclear roles. Indeed long dsRNA as produced by our CT constructs would be predicted to induce a cytoplasmic interferon response that ultimately kills the cell. We therefore measured the levels of 2'5'oligoadenylate synthase (OAS1), a known inducible component of the interferon response pathway. Following CTy ACT 1 transfection into HeLa cells no change in OAS1 levels was detected 72 hrs post-transfection (Fig. 8E). These results imply that CT dsRNA formation is nuclear restricted and so escapes the interferon response. To further investigate the molecular basis of this nuclear RNAi process we show that simple immuno-depletion of dicer from HeLa nuclear extracts also results in reduction of Pol II levels (Fig. 10A). It was confirmed that dicer and Pol II interact with each other by showing that they co-immunoprecipitate (Fig. 10B). Interestingly, dicer immunoprecipites only active phospho-CTD, Pol 110 implying that it interacts with Pol II co-transcriptionally. It was next tested if CT ACT 1 can generate dsRNA and consequent dicer dependent siRNA formation by in vitro transcription in nuclear extracts with added ³²P UTP. A control template that yields a 350 nt single strand RNA following in vitro transcription was also employed (+) as was empty CT vector (V). RNA isolated from untreated extracts gave heterogeneous RNA species for CT and V templates and a single RNA product for +. Single strand specific SI nuclease treated extracts degraded most transcripts implying that they are predominantly single stranded. However some dsRNA was observed with the CTy_^4C77Ex4 construct (Fig. IOC). This effect was controlled by use of VI nuclease which is dsRNA specific. No enrichment of yACTl RNA was observed over the heterogenous RNAs. Finally it was shown that CTy ACT 1 in vitro transcription in nuclear extracts also yielded detectible amounts of siRNA (Fig. 10D). Dicer depletion of the nuclear extracts caused a loss of siRNA production but did not affect levels of dsRNA. Overall these results demonstrate that HeLa cell nuclear extracts contain active dicer associated RNAi activity that recognizes co-transcriptionally formed dsRNA and converts this into siRNA like molecules. It is therefore inferred that HeLa nuclei possess active RNAi apparatus and that this may account for the observed TGS effect seen with the various transfected CT gene plasmids. Indeed the in vitro experiments emphasise the fact that the TGS effects observed occur ^transcriptionally. We therefore refer to this process as coTGS

The above results argue that use of CT plasmid transfection can induce effective coTGS. From a practical view point it was wished to evaluate the relative efficiency of CT induced gene silencing over more orthodox silencing strategies using either exogenous siRNA or plasmids that express short hairpin transcripts that in turn generate siRNAs. As described (Fig. 6), transfections of siRNA, shRNA or CTyACTl against γ-ACTl gave comparable reduction of γ-ACTl mRNA after 72 hours treatment. However siRNA and shRNA affects were lost after 96 hours while CTyACTl induced mRNA reduction was even stronger at this later time point. These results were also confirmed at the protein level. The stable silencing by CTyACTl was again shown to correlate with TGS as H3K9me3 chromatin marks were confirmed for this treatment but were not seen with siRNA or shRNA which induce PTGS. It is concluded that CT induced coTGS may have considerable utility as an alternative gene silencing method to siRNA treatment.

Example 9 - the specificity of CT induced coTGS

To further define the specificity of CT induced coTGS following HeLa cell transfection, additional γ-actin gene (exons 2-6) expression plasmids were constructed with a single CMV promoter driving either sense (S) or antisense (AS) transcription. HeLa cells were transiently transfected with V, S, AS or mixed S and AS plasmids as well as CTy ACT Endogenous y-ACTl was analysed for heterochromatin marks as well as mR A expression levels. Significantly H3K9me3 accumulated over endogenous γ- ACT1 with time for CTyACr transfected cells and to a lesser degree with S+AS constructs. Consistent with heterochromatin formation, γ-ACTl mRNA was also only effectively reduced by coTGS with CTyACZY transfected cells. Presumably the weaker TGS effect of S+AS versus CT reflects the fact that separately synthesized complementary RNAs do not anneal as effectively as co -transcribed transcripts. The even weaker TGS effects seen with AS alone transfected cells may reflect low level recognition of endogenous γ-ACTl mRNA by this antisense transcript. It was also wished to determine whether coTGS effects can be induced by CT plasmids containing only intron sequence. Notably intron sequence was as effective as exon sequence in reducing levels of either nascent γ-ACTl transcript or mRNA (Fig. 11D). This result further emphasizes the fact that CT plasmid transfection induces coTGS since introns are nuclear restricted. Finally we tested whether placing PAS at the end of each convergent transcript (CTT) had a greater effect than CT plasmid induced gene silencing. Transcripts produced from CTT should be stabilised by polyadenylation and possible cytoplasmic export. However it is evident (Fig. 11D) that CTT versus CT plasmid transfection did not enhance coTGS effects on endogenous γ-ACTl, either at nascent or steady state RNA level. It was finally investigated whether CT induced coTGS causes only localized heterochromatin formation or rather results in spreading of heterochromatin outside the target area. A CT construct designed to target the pre-mRNA splicing associated TDP- 43 gene was employed. Using a CT construct containing TDP-43 cDNA exons 2-6, a clear coTGS effect was again shown; reduced Pol II occupancy over the endogenous gene with commensurate reduction in gene expression at the mRNA and protein levels (Fig. 12). The profile of induced heterochromatin marks across TDP-43 was also investigated (Fig. 13 A). While exonic regions cloned into the CT vector showed substantial H3K9me3 marks, above the vector only transfection control, adjacent intronic sequence showed reduced heterochromatin marks. These results indicate that heterochromatin spreading is locally restricted. Presumably the lack of RNA dependent RNA polymerse in mammals, an activity know to cause spreading of heterochromatin marks in plants, may explain this effect. Finally the temporal extent of coTGS induced by CT transfection was tested (Fig 13B). Based on reduced Pol II levels and elevated H3K9me3 on TDP-43 chromatin, it is apparent that TGS effects remain in place for a week following initial CTTDP-43 transient transfection. Beyond this time the coTGS effect diminished, presumably because the small fraction of untransfected HeLa cells eventually outgrow transfected cells.

Overall these results demonstrate that convergent transcription of a variety of test gene fragments transfected into mammalian cell lines can induce clear coTGS effects through a nuclear R Ai pathway (Fig. 13C). These studies provide a new tool for inducing gene silencing in mammalian cells that offers several advantages over existing siR A procedures; longer term silencing using simple lab constructed CT plasmids. Furthermore these results underline the capacity of mammalian cells like other eukaryotes (plants, drosophila) to employ both nuclear coTGS as well as cytoplasmic PTGS.

Experimental Procedures

Yeast strains and plasmids

S. pombe A2 and 972 K was used in this study as wt. Growth conditions and all genetic manipulations were carried out as described previously (Moreno et al. 1991). Plasmid pYL16 containing nourseothricin resistance gene, was used as backbone for experiments in fission yeast. Single gene ura4 or ade6 were cloned into rnulti cloning site. Furthermore, both genes ura4 and ade,6 were cloned into pYI.,16 in tandem or convergent orientation. CT plasmid was constructed by cloning nmtl promoter sequence in sense and antisense orientation, with, multi cloning site in the middle. Ura4, ade6, radii or eso genes were inserted in muiti cloning site on CT plasmid.

Similarly, CMV promoter in antisense orientation, replaced SV40 poiyA signal in pCJ plasmid. Gamma actin gene was cloned in. multi cloning site between sense and antisense CMV promoters on pCI. Transformation and transfection

Transformation of fission yeast was done using Li-acetate method.

Transfection of HeLa, HEK239 and mouse ES cells was performed using

Lipofectamine 2000 (Invitrogen) following the manufacturer's instructions. Growth assay

Exponentially growing cells were serially diluted and dropped onto selective -ura and - ade plates. Growth of cells was checked on EMM complete plates.

Chromatin immunoprecipitation (ChIP)

Cells were grown to OD₆oo 0.5, cross-linked with 1% formaldehyde and incubated for 10 min at 25°C with gentle shaking. These were chilled on ice for 30 min with occasional shaking and harvested by centrifugation at 1000 g for 5 min at 4°C. Pellets were washed 4 times with ice cold buffer I (50 mM Hepes/KOH pH 7.5, 140 mM NaCl, 1 mM EDTA pH 7.5, 1% Triton X-100, 0.1% sodium deoxycholate) and resuspended in 500 μΐ of buffer I containing protease inhibitors (Roche).

The same volume of acid- washed glass beads was added and vortexed for 30 s and chilled on ice for 1 min. This step was repeated four times. Supernatant was transferred into 1.5 ml microfuge tubes and sonicated for 30 s, followed by 45 s interval on ice and repeated five times. Suspension was centrifuged at 14,000 rpm for 15 min at 4°C and supernatant was transferred into 1.5 ml tubes.

Antibodies (Pol II, H3K9me3 and H3K9me2, all from Abeam) were added to the whole cell extracts and incubated over night at 4°C on a rotating wheel. Agarose beads were added to cell extracts and incubated by rotation at 4°C for 2 hr. Beads were washed twice in ice cold buffer I, once in ice cold buffer II (50 mM Hepes/KOH pH 7.5, 500 mM NaCl, 1 mM EDTA pH 7.5, 1% Triton X-100, 0.1% sodium deoxycholate) and once in ice cold buffer III (10 mM Tris-HCl pH 8.0, 250 mM LiCl, 1 mM EDTA pH 7.5, 0.5%) Nonidet P-40, 0.5%> sodium deoxycholate). Beads were resuspended in 100 μΐ of TE buffer (lOmM Tris-Cl pH 7.6, 1 mM EDTA pH 7.5) containing 10 μg/ml RNase A and incubated for 15 min at 37°C. Cross-link reversal was performed in presence of 100 μg Proteinase K, over night at 65°C. Cross-linked chromatin was purified using columns from PCR purification kit (Quiagen). PCR amplification of chromatin derived DNA employed primers specific to ORFs of ura4, and ade6 genes.

Chromatin immunoprecipitation (ChIP); Hela cells.

Transfected HeLa cells were collected from 10 cm plates. Formaldehyde was added directly, at 20°C to tissue culture medium at 1%>: 250ul of 40%>w/v, followed by incubation lOmin 20°C on gently shaking platform. Formaldehyde was inactivated by adding glycine to a final concentration of 0.125M. Medium was aspirated and cells washed twice with 5ml ice cold PBS, containing protease and phosphatase inhibitors and scraped into 2ml tubes. Samples were centrifuged for 4min, 700xg (2800rpm) at 4°C. Cells were gently resuspended in 300ul of cell lysis buffer [5mM PIPES, pH8.0; 85mM KC1; 0.5% nonidet P-40; lmM PMSF; lug/ml pepstatin A; lug/ml leupeptin; 5mM sodium butyrate] and incubated on ice for lOmin. Nuclei were collected by centrifugation at 550g/2400rpm at 4°C and resuspended in ice-cold 400ul nuclear lysis buffer (1% SDS, lOmM EDTA, 50mM Tris-HCl, pH8.0, 0.5mM PMSF, 0.8ug/ml pepstatin A, lug/ml leupeptin, 5mM sodium pyruvate), followed by incubation on ice for lOmin.

Samples were sonicated to an average length of 300-500bp, keeping samples on ice- 12 X 2 Watts (15sec sonicate, 20sec rest) and spun for lOmin, 13000rpm, 4°C to remove cell debris. Supernatant was diluted by adding IP dilution buffer (0.01% SDS, 1,1% Triton X100, 1.2mM EDTA, 16.7mM Tris-HCl pH 8.1, 167mN NaCl, 0.5mM PMSF, 0.8ug/ml pepstatin A, lug/ml leupeptin, 5mM sodium butyrate) and aliquoted into various IP samples. Antibodies were added to samples and incubated o/n at 4°C on rotating wheel. Anti-Pol II (N20X) from Santa Cruz; H3K9me3 and H3K9me2 from Abeam. Immune complexes were pulled down with 60ul of 50 protein A -agarose pre- blocked with salmon sperm DNA/ProtA (#16-157) from Upstate, followed by extensive washes with buffers A-D:

A: 0.1% SDS, 1% Triton X-100, 2mM EDTA, 20mM Tris-HCl pH8.0, 150mM NaCl. B: 0.1% SDS, 1% Triton X-100, 2mM EDTA, 20mM Tris-HCl pH8.0, 500mM NaCl C: 0.25M LiCl, 1% NP40, 1% sodium dexoycholate, lmM EDTA, lOmM Tris-HCl pH8.0.

D: 10: 1 TE pH 8.0.

Immune complexes were eluted with 250ul IP elution buffer (1% SDS, 0.1M NaHC03) and spun down 3min, 13000rpm. Reversal of cross links was performed by initially adding to 3ug/ml RNaseA, 0.3M NaCl, at 65°C 4-5hours, followed by addition of lOul of 0.5M EDTA, 20ul of 1M Tris-HCl, pH 6.5, 2ul lOmg/ml proteinase K and incubation at 45. C for 2 hours.

DNA was purified by Qiagen PCR clean up columns and eluted by lOOul of elution buffer. RNA isolation and RT-PCR

Total cell RNA from S. pombe, HeLa, HEK293 and mouse ES cells was isolated using phenol/chloroform or Trizol (Gibco). RNA was dissolved in sterile water and treated with RNase-free DNase (Promega) for 30 min at 37°C. 100 ng of total RNA was reverse transcribed using Superscript III system (Invitrogen) with oligo dTi₅ or specific primers. Phased oligodT was used for 3 'RACE experiment. cDNA was diluted to 100 μΐ in TE buffer and 10 μΐ was used for PCR. Genomic DNA was employed as a positive control for primer pair efficiency. Results were quantified by real-time PCR with SYBR Green dye and Rotor-Gene 6 software.

Protein extraction and Western blot (fission yeast).

Cells were grown to the exponential phase and washed by 1.2 M Sorbitol. Pellet was resuspended in 0.1 ml of 20% TCA. Cells were broken using glass beads and two rounds of homogenizing at 4°C. After short incubation on ice, 0.9 ml of 5% TCA was added and revortex again. 0.8 ml of protein extract was centrifuged for 10 min at 3000 rpm, to collect precipitated proteins. Pellets were resuspended in 0.25 ml of lx Laemmli protein dye (50 mM Tris-HCl pH 6.8, 100 mM DTT, 2% SDS, 0.1% bromo phenol blue, 10%> glycerol). Samples were boiled for 3 min and spun down for 10 min at 3000 rpm. Equal amounts of proteins were separated on precast bis-tris acrylamide gradient 4%-12% gels (Invitrogen). Western blots employed anti-gamma actin antibody.

Protein extraction (Hela cells).

Cells were collected and pellet was resuspended in RIPA buffer (50 mM Tris-HCl, pH7.5, 150 mM NaCl, 5 mM EDTA, 1% NP-40, 0.25% DOC, 0.1% SDS), followed by 10 min incubation on ice. Cells debris were separated by centrifugation for 5 min, 13k rpm at 4°C. Protein loading buffer was added to the supernatant and samples were boiled prior to the loading on precast bis-tris acrylamide gradient 4%-12% gels (Invitrogen). Western blots employed the following antibodies: anti-yactin (Sigma) anti-tubulin (Sigma), anti-OASl (Abeam), anti-TDP43 (gift from Ashish Dhir), anti- Dicer (Abeam), anti-Pol II 8WG16 (Abeam) and anti-Hsp70 (Abeam). Immunodepletion and immunoprecipitation

Nuclear extract was incubated with anti-Dicer and anti-Pol II antibodies (Abeam) in presence of Protein G beads at 4°C for 90 minutes. Beads with bound protein complexes were washed with 1 ml of washing buffer (50mM Tris-HCl, pH 7.5, 200mM NaCl, 0.1% TritonX, 0.025% NP40).

In vitro transcription and RNA fractionation

DNA templates + (linearized fragment under CMV promoter, 363 nt long), V (CT cassette with no insert) and CT (CTyACTlEx4, 390 nt long) were incubated with nuclear extract and Dicer depleted nuclear extract in presence of ³²P-UTP for 1 hour at 30°C. Total RNA was isolated by phenol/chloroform. Low molecular weight RNA were separated from high molecular weigh RNA by 20%PEG8000/2M NaCl on ice for 30 minutes. Long RNAs were treated with SI or VI nucleases for 10 minutes at r.t. and separated on 6% PAGE gel. Small RNAs were visualized on 20% PAGE gel.

References

Alibu, V.P., Storm, L., Haile, S., Clayton, C, and Horn, D. (2005). A doubly inducible system for RNA interference and rapid RNAi plasmid construction in Trypanosoma brucei. Mol Biochem Parasitol 139, 75-82.

Alio, M., Buggiano, V., Fededa, J.P., Petrillo, E., Schor, I., de la Mata, M., Agirre, E., Plass, M., Eyras, E., Elela, S.A., et al. (2009). Control of alternative splicing through siRNA-mediated transcriptional gene silencing. Nat Struct Mol Biol 16, 717-724.

Barrel, D.P. (2009). MicroRNAs: target recognition and regulatory functions. Cell 136, 215-233.

Baulcombe, D. (2004). RNA silencing in plants. Nature 431, 356-363.

Buhler, M., Verdel, A., and Moazed, D. (2006). Tethering RITS to a nascent transcript initiates RNAi- and heterochromatin-dependent gene silencing. Cell 125, 873-886.

Carthew, R.W., and Sontheimer, E.J. (2009). Origins and Mechanisms of miRNAs and siRNAs. Cell 756, 642-655.

Clemens, M.J. (1997). PKR~a protein kinase regulated by double-stranded RNA. Int J

Biochem Cell Biol 29, 945-949.

Cole TIBS 32, 57 (2007)

Fire, A., Xu, S., Montgomery, M.K., Kostas, S.A., Driver, S.E., and Mello, C.C. (1998). Potent and specific genetic interference by double-stranded RNA in Caenorhabditis elegans. Nature 391, 806-811.

Giordano, E., Rendina, R., Peluso, I., and Furia, M. (2002). RNAi triggered by symmetrically transcribed transgenes in Drosophila melanogaster. Genetics 160, 637- 648.

Grewal, S.I. (2010). RNAi-dependent formation of heterochromatin and its diverse functions. Curr Opin Genet Dev 20, 134-141.

Gullerova, M., and Proudfoot, N.J. (2008). Cohesin complex promotes transcriptional termination between convergent genes in S. pombe. Cell 132, 983-995.

Haussecker, D., and Proudfoot, N.J. (2005). Dicer-dependent turnover of intergenic transcripts from the human beta-globin gene cluster. Mol Cell Biol 25, 9724-9733. Hong, E.J., Villen, J., Gerace, E.L., Gygi, S.P., and Moazed, D. (2005). A cullin E3 ubiquitin ligase complex associates with Rikl and the Clr4 histone H3-K9 methyltransferase and is required for RNAi-mediated heterochromatin formation. RNA Biol 2, 106-111. Iida, T., Nakayama, J., and Moazed, D. (2008). siRNA-mediated heterochromatin establishment requires HP1 and is associated with antisense transcription. Mol Cell 31, 178-189.

Lee, Y., Han, J., Yeom, K.H., Jin, H., and Kim, V.N. (2006). Drosha in primary microRNA processing. Cold Spring Harb Symp Quant Biol 71, 51-57.

Mercer, T.R., Dinger, M.E., and Mattick, J.S. (2009). Long non-coding RNAs: insights into functions. Nat Rev Genet 10, 155-159.

Mello and Conte, Nature 431, 338 (2004)

Moazed, D., Buhler, M., Buker, S.M., Colmenares, S.U., Gerace, EX., Gerber, S.A., Hong, E.J., Motamedi, M.R., Verdel, A., Villen, J., et al. (2006). Studies on the mechanism of RNAi-dependent heterochromatin assembly. Cold Spring Harb Symp Quant Biol 77, 461-471.

Motamedi, M.R., Verdel, A., Colmenares, S.U., Gerber, S.A., Gygi, S.P., and Moazed, D. (2004). Two RNAi complexes, RITS and RDRC, physically interact and localize to noncoding centromeric RNAs. Cell 119, 789-802.

Shi, H., Djikeng, A., Mark, T., Wirtz, E., Tschudi, C, and Ullu, E. (2000). Genetic interference in Trypanosoma brucei by heritable and inducible double-stranded RNA. RNA 6, 1069-1076.

Simmer, F., Buscaino, A., Kos-Braun, I.C., Kagansky, A., Boukaba, A., Urano, T., Kerr, A.R., and Allshire, R.C. Hairpin RNA induces secondary small interfering RNA synthesis and silencing in trans in fission yeast. EMBO Rep 11, 1 12-118.

Tran, N., Cairns, M.J., Dawes, I.W., and Arndt, G.M. (2003). Expressing functional siRNAs in mammalian cells using convergent transcription. BMC Biotechnol J, 21.

Turner, A.M., and Morris, K.V. (2010). Controlling transcription with noncoding RNAs in mammalian cells. Biotechniques 48, ix-xvi.

Verdel, A., Vavasseur, A., Le Gorrec, M., and Touat-Todeschini, L. (2009). Common themes in siRNA-mediated epigenetic silencing pathways. Int J Dev Biol 53, 245-257.

Wang, Z., Morris, J.C., Drew, M.E., and Englund, P.T. (2000). Inhibition of

Trypanosoma brucei gene expression by RNA interference using an integratable vector with opposing T7 promoters. J Biol Chem 275, 40174-40179.

Weinberg et al RNA 12, 256 (2006)

Claims

1. A method of reducing or preventing the expression of a gene of interest comprising providing a DNA molecule having a sequence which comprises in a 5' to 3' direction (i) a Pol II promoter element in sense orientation, (ii) the gene of interest or a portion thereof comprising 40 nucleotides or more, and (iii) a Pol II promoter element in antisense orientation, and expressing the gene of interest incorporated into the DNA molecule in a mammalian cell.

2. A method according to claim 1 wherein the Pol II promoter elements are selected from a highly active viral or composite promoter, a regulatable promoter and an inducible promoter, for example CMV or SV40 late.

3. A method according to claim 1 or 2 wherein the gene of interest or the portion thereof comprises 50 nucleotides or more, preferably 60, 70, 80, 90, 100, 150, 200 or

500 nucleotides or more, and most preferably the full sequence of the gene.

4. A method according to any preceding claim wherein the mammalian cell is in a mammal or in a mammalian expression system.

5. A method according to any preceding claim wherein the expression system is selected from human and mammalian tissues and cell lines, for example HeLa cells, 293T cells, CHO cells, HEP G2 cells, HEK293, cos7 cells.

6. A method according to any preceding claim wherein the mammal is a human, sheep, cow or pig.

7. A method according to any preceding claim wherein expression of the gene of interest is reduced or prevented for at least about 72 hours, more preferably at least about 96 hours or at least about 120 hours.

8. A method according to any preceding claim wherein the gene of interest comprises one or more genes, or portions thereof, selected from viral genes, for example from HIV, herpes or CMV, such as env, gag, pol, tat, rev, nef or vif; genes required for the exosome to function efficiently, such as yeast Rrp6 or any homolog thereof, or yeast Rrp44 or any homolog thereof; genes required for the proteosome to function efficiently, such as ubiquitine activating enzyme, ubiquitine conjugating enzymes or ubiquitine ligase; a region of translocation associated with a cancer, for example all or part of the c-myc fusion protein on chromosome 8; 15-200 trinucleotide repeats specific for a particular trinucleotide repeat disorder, preferably 15-100 trinucleotide repeats, for example 15-100 repeats of the sequence -CAG-.

9. A method according to any preceding claim wherein expression of the gene of interest is decreased at least 10-fold.

10. A method according to any preceding claim wherein the DNA molecule does not contain termination sequences associated with the gene of interest in either the sense or antisense orientations.

11. A DNA molecule having a sequence which comprises in a 5' to 3 ' direction (i) a Pol II promoter element in sense orientation, (ii) a gene of interest or a portion thereof comprising 40 nucleotides or more, and (iii) a Pol II promoter element in antisense orientation, for use in reducing or preventing the expression of the gene of interest in a mammalian cell.

12. Use of a DNA molecule having a sequence which comprises in a 5' to 3' direction (i) a Pol II promoter element in sense orientation, (ii) a gene of interest or a portion thereof comprising 40 nucleotides or more, and (iii) a Pol II promoter element in antisense orientation, for reducing or preventing the expression of the gene of interest in a mammalian cell.

13. A DNA molecule according to claim 11 or use according to claim 12 wherein the DNA molecule does not contain termination sequences associated with the gene of interest in either the sense or antisense orientations.

14. A DNA molecule having a sequence which comprises in a 5' to 3 ' direction (i) a Pol II promoter element in sense orientation, (ii) a gene of interest or a portion thereof comprising 40 nucleotides or more, and (iii) a Pol II promoter element in antisense orientation, for use in therapy.

15. Use of a DNA molecule having a sequence which comprises in a 5' to 3' direction (i) a Pol II promoter element in sense orientation, (ii) a gene of interest or a portion thereof comprising 40 nucleotides or more, and (iii) a Pol II promoter element in antisense orientation, for the manufacture of a medicament for the treatment of cancer or a trinucleotide repeat disorder or a viral infection.

16. A DNA molecule according to claim 14 or use according to claim 15 where the gene of interest comprises one or more genes, or portions thereof, selected from viral genes, for example from HIV, herpes or CMV, such as env, gag, pol, tat, rev, nef or vif; genes required for the exosome to function efficiently, such as yeast Rrp6 or any homolog thereof, or yeast Rrp44 or any homolog thereof; genes required for the proteosome to function efficiently, such as ubiquitine activating enzyme, ubiquitine conjugating enzymes or ubiquitine ligase; a region of translocation associated with a cancer, for example all or part of the c-myc fusion protein on chromosome 8; 15-200 trinucleotide repeats specific for a particular trinucleotide repeat disorder, preferably 15-100 trinucleotide repeats, for example 15-100 repeats of the sequence -CAG-.

17. A DNA molecule according to claim 14 or 16 or use according to claim 15 or 16 wherein the DNA molecule does not contain termination sequences associated with the gene of interest in either the sense or antisense orientations.

18. A method of reducing or preventing the expression of a gene of interest in a mammalian cell comprising integrating a Pol II promoter element in antisense orientation downstream (3') of the gene of interest and inducing expression from the antisense promoter.

19. A method according to claim 18 wherein the Pol II promoter element is an inducible promoter.

20. An in vitro method of preparing a pool of siR A molecules specific for a gene of interest comprising:

- providing a DNA molecule having a sequence which comprises in a 5' to 3' direction (i) a Pol II promoter element in sense orientation, (ii) the gene of interest or a portion thereof comprising 100 nucleotides or more, and (iii) a Pol II promoter element in antisense orientation,

21. A method according to claim 20 where the gene of interest comprises one or more genes, or portions thereof, selected from viral genes, for example from HIV, herpes or CMV, such as env, gag, pol, tat, rev, nef or vif; genes required for the exosome to function efficiently, such as yeast Rrp6 or any homolog thereof, or yeast Rrp44 or any homolog thereof; genes required for the proteosome to function efficiently, such as ubiquitine activating enzyme, ubiquitine conjugating enzymes or ubiquitine ligase; a region of translocation associated with a cancer, for example all or part of the c-myc fusion protein on chromosome 8; 15-200 trinucleotide repeats specific for a particular trinucleotide repeat disorder, preferably 15- 100 trinucleotide repeats, for example 15- 100 repeats of the sequence -CAG-.

22. A method of reducing or preventing the expression of a gene of interest in a mammalian cell comprising transfecting the cell with siRNA molecules produced according to the method of claim 20 or 21.

23. siRNA molecules produced according to the method of claim 20 or 21 , for use in therapy.

24. Use of siRNA molecules produced according to the method of claim 20 or 21 , for the manufacture of a medicament for the treatment of cancer or a trinucleotide repeat disorder or a viral infection. 54

20. An in vitro method of preparing a pool of siRNA molecules specific for a gene of interest comprising:

21. A method according to claim 20 where the gene of interest comprises one or more genes, or portions thereof, selected from viral genes, for example from HIV, herpes or CMV, such as env, gag, pol, tat, rev, nef or vif; genes required for the exosome to function efficiently, such as yeast Rrp6 or any homolog thereof, or yeast Rrp44 or any homolog thereof; genes required for the proteosome to function efficiently, such as ubiquitine activating enzyme, ubiquitine conjugating enzymes or ubiquitine ligase; a region of translocation associated with a cancer, for example all or part of the c-myc fusion protein on chromosome 8; 15-200 trinucleotide repeats specific for a particular trinucleotide repeat disorder, preferablyl5-100 trinucleotide repeats, for example 15-100 repeats of the sequence -CAG-.

22. A method of reducing or preventing the expression of a gene of interest in a mammalian cell comprising trans fecting the cell with siRNA molecules produced according to the method of claim 20 or 21.

22. siRNA molecules produced according to the method of claim 20 or 21, for use in therapy.

23. Use of siRNA molecules produced according to the method of claim 20 or 21, for the manufacture of a medicament for the treatment of cancer or a trinucleotide repeat disorder or a viral infection.