"Gene Transfer Vehicle and Use Therefor"

Field of the invention

The present invention relates to methods for producing proteins, in particular in vivo methods for producing proteins in transgenic animals and animal products, and gene transfer vehicles for use in the same.

Introduction

The large scale production of proteins has wide ranging uses and numerous applications, such as disease treatment and the generation of prophylactic products. Protein provided by such methods include, for example, hormones and growth factors, and antibodies in immunisation protocols.

Chemical synthesis of proteins, although efficient, can generate secondary and/or tertiary structures which differ to those of naturally generated proteins. This becomes particularly evident in the production of large, complex or multi-component proteins. Additionally, proteins produced in this way may also not be subject to specific post-translation modifications, such modifications being difficult to produce in large scale production set up.

Proteins produced in the milk of transgenic animals produce high yields, with the proteins being correctly folded and carrying the expected and correct glycosylation and phosphorylation patterns. However, the generation of large transgenic animals, for example sheep or cattle, is expensive and the process can encounter numerous technical problems leading to a very low level of efficiency.

Firstly, large scale production of transgenic animals incurs high expense with low success rates. Currently, two methods of transgenesis are used: 1 ) transgene constructs are injected directly into fertilised oocytes, or 2) a transgene can be transfected into embryonic stem (ES) cells and cells containing the transgene can be selected. ES cells are then microinjected into blastocysts and continue to all tissues of the developing animal. Microinjection is inefficient but theoretically available to all species. ES cell transfection is limited to use in a small number of species, and is also inefficient.

Foreign DNA can also be transported into eukaryotic cells by other physical, chemical or biological methods. Particle bombardment (Yang et al., (1990) Proc. Natl .Acad. Sci USA 87:9568-9527; Furth et al., (1995) Hybridoma 14:1459-152; Kerr et al., (1996) Animal Biotechn 7:33-45), liposomes (Hens et al., (2000) Biochemica et BioPhysica Acta (BBA) - General Subjects 1523:161 -171 ), retroviral (Mehigh et al., (1993) J. Animal Sci. 71 :687-693; Archer et al., (1994) PNAS 91 :6840-6844) and adenovirus vectors (Yang et al., (1995) Cancer Letters 98:9-17; Jeng et al., (1995) Endocrinology 139:2916-2925; Fan et al., (2002) J. Dairy Sci. 85:1709-1716) have been used to insert foreign DNA. However, stable integration of the foreign DNA will only occur in a very small proportion of the cells which have taken up the DNA. Integration of the foreign DNA is mediated by cellular DNA repair enzymes and occurs at random sites of the host genome.

Secondly, unpackaged foreign DNA is usually inserted at random positions in the genome. This integration is catalysed by cellular DNA repair enzymes. However, only a small proportion of DNA is integrated, whereas most transferred DNA is lost. Retrovirus vectors utilise the step of recombination which leads to the insertion of the virus into the host genome. However, also this integration occurs at random. Very rarely

foreign DNA which carries long uninterrupted DNA segments with homology to endogenous host sequences is inserted site specifically by homologous recombination. Most of the sites at which transgenes are inserted in the host genome will be unsuitable for expression of the foreign DNA.

Thirdly, integrated transgenes are often silenced by the neighbouring gene into which they have integrated, which can have disastrous consequences for the animal and its viability. In addition, the integration event can interfere with gene expression at the integration site.

Establishment of a transgenic line which stably expresses the transgene over successive generations is therefore a difficult, expensive and time consuming enterprise.

Summary of the invention

The present invention provides an efficient transgenesis method wherein insertion of a transgene is targeted to a highly expressed gene locus.

In accordance with a first aspect of the invention there is provided a gene transfer vehicle, said vehicle including a polynucleotide comprising: i) at least one gene (a "transgene"); and ii) a first site specific recombinase site and a second site specific recombinase site,

wherein the gene is located between the first and second site specific recombinase sites and wherein the first and second site specific recombinase sites cannot recombine together.

Optionally, the gene transfer vehicle further includes a polynucleotide encoding a recombinase able to cause recombination to occur between the first site specific recombinase site and a corresponding homologous sequence, and also to cause recombination to occur at the second site specific recombinase site and a corresponding homologous sequence.

The term "homologous" as used above refers to a sequence which is sufficiently similar to the site specific recombinase site in question to undergo recombination with it. The homologous sequence can be identical to the site specific recombinase site, but this is not essential. Homologous sequences can exhibit 95% or greater sequence identity to the site specific recombinase site, for example 96%, 97%, 98% or 99% sequence identity.

In one embodiment of the invention the gene transfer vehicle is a liposome. A liposome gene transfer vehicle can include the polynucleotide in the interior of the liposome, for example within a lipid bilayer. In this embodiment, the recombinase can be provided as an expressible gene in a nucleotide sequence or alternatively may be present as a protein within the interior of the lipid bilayer.

In one embodiment the gene transfer vehicle is a virus or viral vector. Exemplary viruses include but are not limited to a retrovirus or an adenovirus. In this embodiment, the recombinase can be provided as an expressible gene in a nucleotide sequence or alternatively may be present as a protein within the interior of the viral capsid. Conveniently, the viral vector permits insertion of the transgene into the host cell genome at low multiplicity of infection (MOI). Preferably, the viral vector permits insertion of the transgene into the host cell genome at an MOI of 0.1 , 0.2, 0.3, 0.4,

0.5, 1 , 1.5, 2, 2.5, 3 or more. In one embodiment the gene transfer vehicle is a transfection vector, for example a plasmid such as a DNA plasmid.

Optionally, the gene transfer vehicle includes at least one insulator element. Preferably, the at least one insulator element increases the recombination efficiency of the transgene. The insulator element is a nucleotide sequence and can space the two expression cassettes (i.e. the transgene expression and the recombinase expression cassette) apart from each other.

Insulator elements are sequences which either block the activity of transcriptional enhancers on eukaryotic promoters when placed physically between the enhancer and the promoter and/or act as barriers between distinct transcriptional domains (West and Fraser, 2005, Hum MoI Genet, 14 Spec No. 1 : R101 -11 ). Enhancer blocking activity is dependent on binding sites for the DNA binding protein CTCF, whereas barrier function requires binding of the transcription factors USF1 and 2. Insulators are found preferentially at borders between transcriptionally independent domains. They often also define boundaries between areas of differential DNA methylation. They are thought to represent interaction points at which DNA is tethered to the nuclear protein matrix and thereby define independent domains of gene regulation and in a physical sense individual loops of chromatin (Dorman et al., 2007, Semin Cell Dev Biol 18:682-90). An insulator element with a total sequence of 2.4kb (two copies of the chicken b-globin insulator) were used in our experiments. Thus, the insulator element can be from 1 kb to 5kb in size, for example from 2kb to 3kb.

In accordance with a second aspect of the invention, there is provided a method of producing a transgenic animal, the method including the steps of:

(i) modifying the genome of an embryonic stem cell to incorporate a first site specific target recombinase site and a second site specific target recombinase site wherein the first and second site specific target recombinase sites cannot recombine together; (ii) selecting embryonic stem cells so modified and cultivating said selected cells; and

(iii) culturing the cells from step (ii) to produce a transgenic animal.

In this aspect of the invention, the transgenic animal produced includes within its genome first and second site specific target recombinase sites which can be utilised for the easy and targeted insertion of any transgene between the first and second site specific target recombinase sites. In one embodiment the first and second site specific target recombinase sites are inserted into the genome at a location suitable for expression of the transgene.

Optionally, the method of producing a transgenic animal further includes the step of introducing at least one transgene into the transgenic animal by means of site specific recombination. Conveniently the transgene can be introduced using the gene transfer vehicle of the present invention which comprises a first site specific recombinase site and a second site specific recombinase site, wherein said first and second sites are homologous to the first site specific target recombinase site and second site specific target recombinase site respectively, these target recombinase sites having been previously introduced into the animal's genome.

In this embodiment, said method further includes the steps of:

(i) introducing a gene transfer vehicle of the invention into the transgenic animal; and (ii) allowing recombination to occur between the polynucleotide of the gene transfer vehicle and the genome of the animal.

In one embodiment, the first and second site specific target recombinase sites are introduced into the genome of the embryonic stem cell at a location which is under the control of promoter and/or enhancer elements.

In one embodiment, the first and second target recombinase sites are introduced into the embryonic stem cell genome at a location where an endogenous gene is expressed or is to be expressed. In exemplary embodiments the first and second site specific target recombinase sites flank an endogenous gene. Preferably, the endogenous gene selected is expressed or can be expressed within at least one cell type of the transgenic animal.

In one embodiment at least one transgene is replaced by administration of a second gene delivery vehicle including a polynucleotide having a different transgene sequence (i.e. a second transgene).

In accordance with a third aspect of the invention there is provided a method for insertion of a transgene to a specific locus within the genome of an animal, said method including the steps of: i) providing an animal with a first target sequence located between first and second site specific target recombinase sites within the genome of at least one cell in the animal,

wherein the first and second site specific target recombinase sites cannot recombine together; ii) providing a gene transfer vehicle including: a) a polynucleotide comprising at least one transgene and first and second site specific recombinase sites, said transgene being located between first and second site specific recombinase sites, wherein the first and second site specific recombinase sites cannot recombine together and wherein the first site specific recombinase site is homologous to the first site specific target recombinase site in the genome of said animal and wherein the second site specific recombinase site is homologous to the second site specific target recombinase site in the genome of said animal cell; and b) a gene encoding a site specific recombinase able to cause recombination between the first site specific recombinase site of the gene transfer vehicle and the first site specific target recombinase site of the animal genome and between the second site specific recombinase site of the gene transfer vehicle and the second site specific target recombinase site of the genome of said animal cell; and iii) introducing the gene transfer vehicle to said animal cell of step i) and allowing recombination to occur.

In a fourth aspect of the invention there is provided a method of producing a modified embryonic stem cell culture (optionally capable of producing a transgenic animal) able to express a transgene, said method comprising the steps of: i) genetically modifying flanking sequences of an expressible gene of an embryonic stem cell to introduce first and second

site specific target recombinase sites wherein the first and second target recombinase sites cannot recombine together; ii) introducing a gene transfer vehicle into the modified embryonic stem cell, the gene transfer vehicle including first and second site specific recombinase sites and a transgene sequence, wherein the first and second site specific recombinase sites of the expression cassette cannot recombine together, wherein said first site specific recombinase site is homologous to the first site specific target recombinase site of said modified embryonic stem cell and wherein said second site specific recombinase site is homologous to the second site specific target recombinase site of said modified embryonic stem cell; iii) allowing site specific recombination to occur between the first site specific recombinase site and the first site specific target recombinase site in the modified embryonic stem cell and allowing site specific recombination to occur between the second site specific recombinase site and the second site specific target recombinase site in the modified embryonic stem cell; and iv) selecting modified embryonic stem cells having the transgene in its genome.

Optionally, the methods of the invention include the step of purifying the transgene product from the cell or the animal or animal product. Protocols for the isolation and purification of proteins are well known to those skilled in the art.

As an example, in step i) the 5' flanking sequence of the expressible gene can be modified to include the first site specific target recombinase site

and the 3' flanking sequence can be modified to include the second site specific target recombinase site. Flanking sequences are the sequences upstream or downstream (as appropriate) of the selected expressible gene and could be upstream/downstream of elements such as promoters, enhancers or other regulatory elements.

In some embodiments of the methods of the invention, expression of the transgene is inducible, for example by hormones such as but not limited to lactogenic hormones. In other tissues other suitable hormones can be used to induce transgene expression.

In a fifth aspect, the current invention provides a transgenic animal produced by the above method.

Detailed description of the invention

As described above the invention concerns an in vivo trangenesis method enabling a transgene to be targeted to a highly expressed gene locus in an animal or a cell. The gene locus targeted has been pre-selected as containing the elements needed to enable abundant expression of the protein. The method combines viral vector and/or liposome technology with site specific recombination techniques to integrate the transgene at an active and expressed gene locus. In one embodiment, the gene locus is a milk protein locus so that expression of the transgene can occur in the lactating mammary gland.

In one embodiment of the invention the transgene is present as a single copy thereby avoiding epigenetic silencing mechanisms which can be induced by multi-copy transgene arrays. The transgene is inserted into the animal genome so that the transcriptional and translational capacity of

the replaced genomic DNA is utilised to express the transgene. Preferably, the transgene is inserted in a location on the animal genome or cell genome which enables or supports expression of the transgene in body fluids. Optionally, the transgene replaces a gene encoding a milk protein. Where the transgenic protein is expressed in milk, purification of the transgene protein from the transgenic animal is simplified.

Using the methods of the invention, the cost of producing transgenic animals is reduced by using somatic gene therapy techniques to introduce a gene transfer vehicle including the transgene into an animal wherein the animal's genome comprises site specific integration sites generated by modifying the genomic DNA of the animal to include first and second site specific target recombination sites located on each side of a specific position of interest. For example, where the transgene is to be expressed in the milk of an animal, the integration sites are introduced so that the transgene is expressible via the expression machinery of a protein expressed in milk, such as (but not limited to) beta-casein.

A method of producing transgenic animals using the in vivo Pangenesis method of the invention has also been developed. The method is proposed to be of particular use in the generation of large transgenic animals, for example (but not limited to) cattle, sheep, goats and pigs. Such animals can produce large quantities of milk per day and have the capacity to produce large quantities of transgenic protein in their milk. For example, if the transgene is site-specifically integrated into one allele of the bovine beta-casein gene, protein expression levels of up to 5mg/ml of milk can be achieved on a consistent basis. Such a concentration represents a significant improvement over existing methods for the production of exogenous proteins in the milk of transgenic animals.

Fig. 1 is a schematic representation of the in vivo method of the invention for generation of a transgenic animal into which any transgene of choice may be inserted at a site specific point engineered into the transgenic animal's genome. The engineered sites, as explained below, allow site specific recombination, preferably recombinase mediated cassette exchange (RCME), to occur between the gene transfer vehicle introducing the transgene into genome. The gene transfer vehicle is represented in Fig. 1 as a liposome or retroviral vector, but is not limited to these. Additionally, as explained below, the methods of the invention are not limited to use of the lox/Cre system known to the skilled man.

The use of a site specific recombination system to introduce a transgene present in a gene transfer vehicle of the invention into in vitro cell cultures enables stable expression at high levels of the transgene. Cell cultures include, but are not limited to, embryonic stem cells (e.g. mouse embryonic stem cells), mouse mammary gland cells, fusions of embryonic and mammary gland cells and HE293K cells.

The gene transfer vehicles for use in the in vivo production of a protein of interest include a nucleotide sequence encoding the protein of interest, i.e. the transgene, and a site specific recombinase expression cassette. The transgene is located between first and second site specific recombinase sites which cannot recombine together. The presence of the site specific recombinase sites allows recombination, mediated by a recombinase compatible with the site specific recombinase sites, to occur between the site specific recombinase sites of the gene transfer vehicle and the site specific target recombinase sites of the host genome to insert the transgene into the host genome, as explained below. Orientation of the target sites, as explained below, is such that the transgene is inserted into

the host genome in the correct orientation to allow expression of the transgene.

Preferably the gene transfer vehicle is a viral vector or a liposome. Where the gene transfer vehicle is a viral vector it may also include nucleotide sequences encoding the recombinase. Alternatively, where the gene transfer vehicle is a viral vector or a liposome, it may include the recombinase enzyme itself. Optionally, the gene transfer vehicle provides recombinase in a high relative molar ratio with respect to the transgene cassette to catalyse recombination. Thus, the recombinase can be present in a molar excess to the transgene DNA. Optionally, the gene transfer vehicle includes insulator elements to enhance the efficiency of transgene insertion into the host cell genome. Preferably, the viral vector is an adenovirus or retrovirus, but any other suitable virus may be used.

As used herein, the term "nucleotide sequence" refers to a length of a number of nucleotides and may refer to DNA, cDNA, or RNA. The nucleotide sequences may be constructed from natural nucleic acid bases, synthetic bases, or mixtures thereof.

The term "gene" has been used herein to refer to a polynucleotide comprising a nucleotide sequence which encodes a protein. A "transgene" is a gene which is inserted or intended for insertion into a foreign location (particularly a genome) by genetic manipulation.

As used herein, the term "protein" relates to peptidal molecules of any size and includes polypetides and peptides. References to "proteins" should therefore be taken to include references to polypetides and peptides.

An "expression cassette" as referred to herein means a polynucleotide including all elements necessary for transcription of the protein encoded therein. Such elements generally comprise at least one or more promoter sequences, regulatory elements, an origin of replication, a nucleotide coding sequence (or gene), and a stop sequence. Such sequences are well known to the person skilled in the art.

As used herein, a first or second site specific target recombinase site refers to the first and second site specific recombinase sites present within the genome of the embryonic stem cell, animal cell or the animal. Thus the first and second site specific target recombinase sites form a pair of target site specific recombinase sites. This pair of target site specific recombinase sites cannot recombine with each other, i.e. they form a pair of incompatible site specific target recombinase sites. Use of the term "target" indicates that the site specific recombinase sites of the cell or animal genome are in a specific location targeted for location of a transgene.

As used herein, a first or second site specific recombinase site (as opposed to a site specific target recombinase site) is used to refer to a site specific recombinase site present within a gene transfer vehicle. Thus the recombinase sites within the gene transfer vehicle form a pair of site specific recombinase sites. This pair of site specific recombinase sites cannot recombine together and thus form a pair of incompatible site specific recombinase sites.

Site specific recombination

The current invention employs site specific recombination, preferably but not limited to recombinase-mediated cassette exchange (RCME) which

relies on the presence of a pair of incompatible recombinase target sites (referred to as site 1 and site 2). The two sites, which are not able to recombine with each other, are placed in the genome of the host cell. If the same or suitably homolgous pair of sites is present in a transgene construct, then site 1 of the transgene construct and site 1 in the host genome can recombine and site 2 in the transgene construct and site 2 in the host genome can recombine. The net result is the exchange of the cassette flanked by the target sites in the host genome for the transgene flanked by the identical pair of target sites. (See Waterhouse et al., (1993) Nucleic Acids Research 21 :2265-2266; Bouhassira et al., (1997) Blood 90:3332-3334; Bethke & Sauer (1997) Nucleic Acids Research 25:2828- 2834). RCME therefore ensures DNA sequences (or cassettes) flanked by a pair of incompatible recombinase target sites can be exchanged for other DNA cassettes flanked by the same or suitably homologous pair of recombinase target sites.

In RCME recombination of the transgene at the site specific recombinase sites is catalysed by a recombinase enzyme. This speeds up the rate of recombination and the accuracy of recombination into the host genome. The recombinase enzymes recognise specific target sites of between 20 and 150 nucleotides in length and cause double stranded DNA exchange to occur at these sites.

It would be understood by the person skilled in the art that any site specific recombinase site could be utilised in the current invention. One preferred recombinase site which can be used in the current invention includes (but is not limited to) the locus of recombination, lox, e.g. loxP and loxM, elements which are recognised by Cre recombinase. Additionally, FIp recombinase which also operates in mammalian cells may also be used in the methods of the current invention. Other recombinases which can

mediate DNA integration would also be suitable for use in the methods of the present invention, e.g. øC31 , Z-resolvases.

Cre (causes recombination) is a 38kDa Type I topoisomerase protein from bacteriophage P1 which mediates intramolecular recombination (i.e. excision or inversion) and intermolecular recombination (i.e. integration). The enzyme recognises the specific recombinase target sites loxP (locus of crossing over of P1 ) and causes recombination between these pairs of sites.

Lox P comprises two 13 base pair repeats separated by an asymmetric 8bp spacer region. One molecule of Cre binds to each repeat sequence with the 8bp spacer region acting as the site of crossing over:

ATAACTTCGTATA ATGTATGC TATACGAAGTTAT SEQ ID NO 1

inverted repeat spacer region inverted repeat

Directionality to the loxP element is provided by the asymmetricity of the loxP site, which also determines which type of recombination event occurs: two inverted loxP elements result in the inversion of the flanked DNA sequence, whereas the presence of two elements in the same orientation leads to the excision of the flanked sequence, leaving one loxP element behind.

It is thus possible to use RCME to insert a transgene in the correct orientation in a specific genomic site. Methods have been described in Bethke and Sauer, (1997) Nucleic Acids Research 25:2828-2834.

Other recombinase enzymes use different catalytic mechanisms. The skilled person will appreciate that the design of the integration site and gene transfer vehicle should take account of these differences.

In the present invention, the site-specific recombinase target sites flanking the target gene and the site specific recombinase sites flanking the sequence coding for the protein of interest (the transgene) are preferably oriented in the same direction such that the two nucleotide coding sequences are excised, and the nucleotide sequence coding for the protein of interest is inserted into the region vacated by the target gene.

In the present invention, site-specific replacement of the target gene is preferably by recombinase-mediated cassette exchange, most preferably using the Cre/lox system. However, the site-specific recombinase system used may be any such system known in the art.

The nucleotide sequence of the first site specific recombinase site in the gene transfer vehicle and the first site specific target recombinase site on the cell or animal genome have sufficient homology that recombination can occur between the two. Preferably, the nucleotide sequence of the first site specific recombinase site in the gene transfer vehicle and the first site specific target recombinase site on the cell or animal genome are identical. Similarly, the nucleotide sequence of the second site specific recombinase site in the gene transfer vehicle and the second site specific target recombinase site on the cell or animal genome have sufficient homology that recombination can occur between the two. Preferably, the nucleotide sequence of the second site specific recombinase site in the gene transfer vehicle and the second site specific target recombinase site on the cell or animal genome are identical. The first and second site specific recombinase sites should differ sufficiently from one another to

ensure that recombination between these two sites does not occur. The first and second site specific target recombinase sites should differ sufficiently from one another to ensure that recombination between these two sites does not occur. Site specific recombinase sites which fulfill these criteria are described in Lee (1998; Gene, 216, 55-65) and the skilled man would appreciate that such sites and other equivalent sites are suitable for use in the current invention.

Cell lines, animal tissues and animals

Suitable cell lines for use in some of the methods of the invention are selected based on the gene locus into which the transgene is to be inserted. For example, where the transgene is inserted into the casein locus, suitable cells are those which can express casein proteins or can be modified to express casein proteins. Whilst embryonic stem cells do not express casein genes, somatic cell fusion with differentiated mammary gland cells enable expression of casein genes in the resulting fused cells in the presence of lactogenic hormones as shown in Example 3.

In the methods of the present invention it is also possible to use a cell fusion system in which gene-targeted embryonic stem cells are fused with differentiated mammary gland cells. The resulting fusion cells derived from the targeted ES cells faithfully impart the natural regulation of the casein gene into a reporter gene embedded in the gene locus, as shown in Example 3.

The gene transfer vehicle can be introduced into many tissues of the animal comprising the specifically introduced site specific recombination sites, including mammary tissue, bone marrow or other blood producing

tissues, salivary gland tissue, or ovarian tissue. In accordance with this, the animal product in which the transgene is expressed may be milk or any other mammary secretion, blood or any secretion from a blood producing tissue, saliva or ova. Preferably, the animal product is milk or other mammary gland secretion.

Any animal capable of being generated using transgenics may be generated by the methods of the current invention. The term "animal" as used herein includes a non-human organism of the kingdom animalia and may be a non-human or an avian. In one embodiment, the animal is a mammal, preferably a ruminant. In one embodiment the animal is a murine, ovine, caprine, porcine or bovine.

Transgenic animals produced by in vivo methods of the invention

As explained above, the in vivo method for producing transgenic animals enables any transgene to be delivered via a gene transfer vehicle. Using the methods of the current invention, stem cells modified to include first and second site specific target recombinase sites are grown in culture or into animals. The first and second site specific target recombinase sites are engineered into the host cell genome such that the transgene, when inserted into the genome of the transgenic animal, comes under the control of a host promoter which enables expression. Expression is ideally responsive to a external stimuli, e.g. hormones, so that expression of the transgene can be induced. Preferably, the transgene is inserted under the control of a promoter which exhibits tissue specific expression so that the transgene can be purified from a given animal product. These animals may then have a gene transfer vehicle delivered into the specific tissue in which expression occurs. For example, where the site specific target recombinase sites are introduced at the casein gene locus, the gene

delivery vehicle can be administered to the animal by injection into the mammary tissue of the animal. The transgene will then be taken up by the mammary tissue cells and via RCME will be inserted into the host genome at the site of the engineered site specific target recombinase sites.

Transgenic proteins produced by the method of the invention

Many proteins can be produced by the methods of the invention, including human proteins. Non-human proteins are also within the scope of the invention.

The person skilled in the art would appreciate that the proteins produced by the methods of the invention may include fusion proteins where the transgene product is fused to an endogenous protein expressed by the host genome. The transgene product can be purified from the host product using known methods available to the person skilled in the art and optionally cleaved from the fusion product without undue burden and utilising known methods for example hydrolysis, enzyme degradation.

Additionally, the fusion product may be engineered to enhance the ease with which the transgene can be purified from the animal product. For example, the expression cassette may comprise a further nucleotide sequence encoding a ligand such as avidin, streptavidin, antibodies and the like. By fusing the nucleotides sequences encoding the transgene and the further nucleotide sequence encoding the ligand or the like, the transgene produced is expressed as a fusion product which can be purified using known methods available to the skilled man, for example, chromatography columns, immunoprecipitation, gel purification and the like.

Following purification, a protein product may be isolated and/or purified and may be formulated into a composition comprising at least one additional component. Such a composition may comprise a pharmaceutically acceptable excipient, vehicle or carrier.

Milk casein proteins

The caseins are serine rich phosphoproteins which are almost exclusively expressed in the lactating mammary gland. The casein proteins together make up to 80% of total milk protein and also 80% of total mammary gland mRNA (Rijnkels, (2002) J. Mammary Gland Biology and Neoplasia 7:327- 345). They are members of a large family of serine rich phosphoproteins.

The casein gene locus harbours 4 or 5 genes depending on the mammalian species (Rijnkels, (2002) J. Mammary Gland Biology and Neoplasia 7:327-345). The genes are expressed exclusively in the mammary gland and are co-ordinately up-regulated during pregnancy and lactation such that 80% of the mRNA present in a lactating mammary epithelial cell is casein mRNA (Rijnkels et al., (1995) Biochem Journal 311 :927-937). Casein proteins include alpha-casein, beta-casein and kappa-casein.

2+

The casein proteins are able to complex Ca and form micelles which are secreted into milk (Rijnkels et al., (1995) Biochem Journal 311 :927-937). Expression of the casein genes is regulated at the transcriptional and translational level (Groner, (2002) Domest. Anim. Endocrinol. 23:25-32). Initiation of casein gene transcription is critically dependent on induction by prolactin and glucocorticoid hormones and cell-matrix interactions mediated by integrins. The effects are mediated by STAT5 and the glucocorticoid receptor which interact synergistically with the casein

promoter regions via STAT5 binding sites and adjacent GR half sites (Lechner et al., (1997) J. Biol. Chem 272:20954-60; Stoecklin et al., (1999) J. Steroid Biochem. MoI. Biol. 69-195-204; Stoecklin et al., (1997) MoI. Cell. Biol. 17:6708-6716; Streuli et al., (1995) J. Biol. Chem 270: 21639- 21644). However, when reporter genes under the control of β-casein promoter elements are transiently transfected into mammary gland cell lines they are often not inducible by prolactin. In stably transfected cells, however, the β-casein promoter becomes responsive to prolactin treatment suggesting that embedding into cellular chromatin and hence interaction with chromatin constituents is required for prolactin induction (Myers et al., (1998) MoI. Cell. Biol 18:2184-95; Schmitt-Ney et al., (1991 ) MoI. Cell Biol. 11 :3745-55; Pfitzner et al., (1998) MoI. Endocrinol. 12: 1582-1593; Kabotyanski et al., (2006) MoI Endocrinol 20:235-68).

The precise function of the casein proteins is not clear. It is likely that they are required for transporting calcium to the offspring to provide sufficient building material for bone development. Inactivation of the beta-casein gene has shown to have little effect on growth of the offspring in the mouse, as growth is delayed by about 10% (Kumar et al., (1994) Proc. Natl. Acad. Sci USA 89:6943-7).

Deletion of the k-casein gene in mice completely abrogates milk production (Shekar et al., (2006) Proc. Natl. Acad. Sci. USA 103:8000-5). The precise reason for this is unclear as the precise structure of casein micelles is unknown. The available results, however, support a model in which k-casein (the casein protein which is present in milk in least abundance) plays a critical role in ensuring transport and solubility of casein micelles. Deficiency for a-casein in goats, which occurs naturally in genotype O, also decreases total milk yield (Chanat et al., (1999) J. Cell Sci:3399-412).

In the mouse alpha-casein and beta-casein are the most prominent milk protein genes expressed at roughly equal levels (Kumar et al., (1994) Proc. Natl. Acad. Sci USA 89:6943-7; McClenaghan et al., (1995) Biochem. J. 310).

Many known techniques and protocols for manipulation of nucleic acid, for example in the preparation of nucleic acid constructs, introduction of DNA into cells, homologous recombination techniques, gene expression, and analysis of proteins, are described in detail in Current Protocols in Molecular Biology, Ausubel et al. eds., Wiley, 2005, ISBN 047150338X as well as in Molecular Cloning: A Laboratory Manual: 3 ^rd edition, Sambrook et al., 2001 , Cold Spring Harbor Laboratory Press. The disclosures of Sambrook et al., and Ausubel et al., are therefore incorporated herein by reference.

Preferred and alternative features of each aspect of the invention are as for each of the other aspects mutatis mutandis.

The present invention will now be described by way of example only and with reference to the accompanying drawings wherein:

Fig. 1 is a diagrammatic representation of the system of the invention;

Fig. 2a is a diagrammatic representation of the mouse casein gene locus; the site of insertion of a PGK-HPRT selection marker expression cassette into the β-casein gene is indicated;

Fig. 2b shows the sequence of the two incompatible lox sites: loxP and Iox2272. The diverging nucleotides in the unidirectional spacer segment of the Iox2272 site are underlined;

Fig. 3a is a diagrammatic representation of the murine β-casein gene and the corresponding gene targeting construct used to modify the gene;

Fig. 3b is a diagrammatic representation of the murine β-casein gene and its derivatives after homologous recombination and recombinase mediated cassette exchange (RMCE);

Fig. 3c shows PCR analysis of genomic DNA isolated from the cell clones RMCE2272-98, GH1 , GH3 and the parental HM1 embryonic stem cells using the primer combination bcasiO, bcas3, HPRT2 and bgalint.1 ;

Fig. 3d is a Southern blot analysis of cell pools derived from a PEG-fusion of HC11 cells and the cell clone HM1 GH1 after selection in medium containing 50, 100, 150 or 200μg/ml hygromycin B (HyB);

Fig. 4 is a photograph of a representative chromosome spread of a fusion cell clone;

Fig. 5 is a graph showing cell survival rates of HC11 cells and pooled cell clones derived from a somatic cell fusion of HC11 cells with HM1 GH1 cells (fusion), in the presence of increasing concentrations of Gancyclovir;

Fig. 6 shows X-gal staining of 11 different cell clones derived after fusion of HC11 cells with HM1 GH1 cells, selection in medium containing 75μg/ml HyB and induction with lactogenic hormones ([+] LH); control cells, which were not induced with lactogenic hormones ([-]LH);

Fig. 7a represents characterisation of the chimeric β-casein/β- galactosidase transcript and a schematic representation of the expected splice variants of the chimeric β-casein/β-galactosidase gene;

Fig. 7b shows the results of RT-PCR amplification of RNA derived from fusion cell clone F13 using the primer pair bcasi 9/bgalint.1 ;

Fig. 7c shows the results of quantitative PCR of cDNA derived from untreated fusion cells (F13 [-]LH) or fusion cells induced by lactogenic hormones (F13 [+]LH) was used as template;

Fig. 8 is a graphical representation of quantitative PCR analysis of β- casein expression in parental HC11 cells and ten fusion cell clones in the absence ([-] LH) or presence ([+] LH) of lactogenic hormones;

Fig. 9 shows expression of the endogenous β-casein gene and the chimeric β-casein/β-galactosidase allele in the presence ([+] LH) and absence ([-] LH) of lactogenic hormones;

Fig. 10a is a diagrammatic representation of the murine casein locus; the selection marker gene cassettes used to modify the β and γ-casein genes are indicated;

Fig. 10b shows the sequence of the two incompatible lox sites: loxP and Iox2272. The diverging nucleotides in the unidirectional spacer segment of the Iox2272 site are underlined;

Fig. 11 a is a diagrammatic representation of the unmodified and targeted murine γ-casein gene;

Fig. 11 b shows the results of PCR analysis of genomic DNA isolated from the cell clone HM1 -γ-#185 and the parental HM1 embryonic stem cells using the primer combination gcasδ, gcas9 and PGK3 (lanes 1 and 2) and gcasδ and PGK3 (lane 3);

Fig. 11 c is a Southern blot analysis of HM1 cells and the γ-casein targeted cell clones #155 and #185;

Fig. 12a is a schematic representation of the murine casein locus targeted at the γ-casein and β-casein genes;

Fig. 12b is a PCR analysis of genomic DNA isolated from cell clones HM1 - γ/β-#57, the parental HM1 embryonic stem cells and six representative RMCE clones using the primer combination bcas3, bcasiO, bgalint.3, HPRT2 (upper panel) and PGK5, bgalint.1 , pBKpA2 (lower panel);

Fig. 12c shows expression of the chimeric β-casein/β-galactosidase allele in fusion cell clones G1 , Hy9, Hy11 and Hy12 in the presence ([+] LH) and absence ([-] LH) of lactogenic hormones;

Fig. 13a is a diagrammatic representation of the mouse casein gene locus; the site of insertion of a PGK-hytk selection marker expression cassette into the α-casein gene is indicated;

Fig. 13b shows the sequence of the two incompatible lox sites: loxP and Iox2272. The diverging nucleotides in the unidirectional spacer segment of the Iox2272 site are underlined;

Fig. 14 is a diagrammatic representation of the murine β-casein gene and the corresponding gene targeting construct used to modify the gene;

Fig. 15 is a diagrammatic representation of the mouse α-casein gene and the expected protein products derived from the original and modified α- casein genes. Constructs for targeting the α-casein gene are illustrated

together with expression cassettes encoding the hytk fusion protein under the control of a PGK promoter;

Fig. 16a is a diagrammatic representation of the unmodified and the targeted α-casein gene;

Fig. 16b shows the results of a PCR analysis of 3 DNA samples derived from transfected embryonic stem cells and a no template control ([-]);

Fig. 17a is a diagrammatic representation of the endogenous and the targeted a-casein genes and the α-casein gene modified by gene targeting and recombinase mediated cassette exchange (RMCE);

Fig. 17b PCR analysis to identify α-casein alleles modified by homologous and site-specific recombination;

Fig. 18a is a diagrammatic representation of the endogenous and the targeted α-casein genes and the α-casein gene modified by gene targeting and recombinase mediated cassette exchange (RMCE);

Fig. 18b is a Southern blot analysis of cells lines 4B4-2E6 (RCME modified), 4B4-2A5 (RCME), 4B4-4A8 (RCME), HM#139 (targeted), HM1 #6299-5 (targeted), HM1 #6299-12 (RCME) and non-transfected embryonic stem cells (HM1 );

Fig. 19a is a diagrammatic representation of the mouse α-casein gene modified by gene targeting and RMCE and the mRNAs derived from the modified and unmodified allele;

Fig. 19b shows the results of PCR analysis of cDNA derived from the lactating mammary tissue of control mouse and two transgenic mice carrying the modified alpha-casein allele;

Fig. 19c shows the result of quantitative analysis of the cDNA derived from the modified and unmodified alleles;

Fig. 20 shows the results of milk from two lines of transgenic animals carrying a modified a-casein gene and one control mouse separated on a 10% polyacrylamide gel alongside (human) IgG preparations of known concentration and stained with Coomassie Blue;

Fig. 21 shows the results of analysis of milk from two lines of transgenic animals carrying a modified a-casein gene and one control mouse was separated on a 10% polyacrylamide gel alongside (human) IgG preparations of known concentration and blotted to nitrocellulose;

Fig. 22a is a schematic representation of the murine β-casein gene and its derivatives after homologous recombination and recombinase mediated cassette exchange (RMCE);

Fig. 22b shows PCR analysis of genomic DNA isolated from the cell clones HM1 RMCE2272-gal/hytk [GH1], and the cell clones HM1 N1 -2272 to HM1 N7-2272 derived from it;

Fig. 23a is a schematic representation of the plasmids pShuttle-H6 and pBK2272-HPRT. The neomycin resistance marker genes (neo) are indicated as solid arrows. The Cre open reading frame is indicated as a shaded arrow. The PGK and the tk promoter are also indicated as arrows. The HPRT selection marker gene is indicated as a striped arrow. Also

indicated are: the origin of replication, the left and right arms of homology (RAH, LAH) and the inverted terminal repeats (ITR) in the pShuttle vector backbone. Important restriction sites and recombination target sites (loxP and Iox2272, in italics) are shown;

Fig. 23b is a linear representation of the substrate plasmids and the expected recombination product. Features are marked as in Panel A. The primer binding sites (horizontal arrows) used for genotyping and the sizes of the expected PCR products are indicated;

Fig. 23c shows PCR analysis of DNA isolated from HEK 293 cells transfected with the indicated plasmids;

Fig. 24a is a schematic representation of the plasmid pShuttle-G5. Features are as in Fig. 23. The copies of the b-globin insulator element (INS) are indicated as vertically striped boxes;

Fig. 24b shows PCR analysis using the primer combination bcas3, bcasi O and neoint.4 on DNA isolated from representative cell clones derived after electroporation of HM1 RMCE2272-98 cells with the plasmid pShuttle-G5 or pB2272-neo and selection of the transfected cells in medium containing 200μg/ml of G418;

Fig. 25a is a schematic representation of the linear form of the plasmid pBlox1/2-hyg/luc2 and the Cre recombination product derived from it. The luciferase open reading frame (luc) is indicated as a striped arrow, the amplicillin resistance gene (ampR) is indicated as a solid arrow and the hygromycin phosphotransferase gene (hyR) is indicated as a shaded arrow. The primer binding sites (horizontal arrows) used for genotyping and the sizes of the expected PCR products are indicated;

Fig. 25b shows PCR analysis of DNA isolated from HEK 293 and BHK cells transfected with pBlox1/2-hyg/luc2. The cells were either co- transfected with the indicated plasmids or infected with the indicated virus vectors 24h post transfection;

Fig. 26a is a schematic representation of the linear plasmids pShuttle-G5, pBK2272-HPRT and the expected recombination product. The neomycin resistance marker genes (neo) are indicated as solid arrows. The Cre open reading frame is indicated as a shaded arrow. The PGK and the tk promoter are also indicated as arrows. The copies of the b-globin insulator element (INS) are indicated as vertically striped boxes. Also indicated are: the origin of replication, the left and right arms of homology (RAH, LAH) and the inverted terminal repeats (ITR) in the pShuttle vector backbone. The positions of the Iox2272 and loxP sites are marked by vertical arrows;

Fig. 26b shows PCR analysis of DNA isolated from HEK 293 and BHK cells transfected/infected with pBK2272-HPRT plus the indicated plasmids and virus vectors;

Fig. 27a is a schematic representation of the plasmid pB-bcas7-2272-hytk, the insert of the virus vector G5 and the resulting RMCE product. Exons of the β-casein gene are indicated as solid boxes, the neomycin and hytk selection marker genes are indicated as solid and shaded arrows, respectively and the Cre ORF (Cre) is indicated as a solid arrow. The PGK promoter elements directing expression of the selection marker genes are indicated as black arrowheads. The positions of the Iox2272 and loxP sites are marked by vertical arrows. The primer binding sites (horizontal arrows) used for genotyping and the sizes of the expected PCR products are indicated;

Fig. 27b shows PCR analysis of DNA isolated from HEK 293 cells transiently or stably transfected with pB-bcas7-2272-hytk; and

Fig. 27c shows analysis of the 5' end of the RMCE reaction using the primer combination bcas6/PGK5/hytk2. Successful recombination at the Iox2272 site is indicated by the presence of the 215bp PCR product. Non- recombined DNA yields a PCR product of 936bp. Phage λ DNA digested with Hindlll and EcoRI and the NEB PCR marker (New England Biolabs) were used as molecular weight markers.

Examples

Example 1 : Generation of a Targeting construct for the beta-casein gene

Targeting constructs for the β-casein gene (Fig. 3a, Fig. 3b) were generated using a short arm of homology of 841 bp corresponding to nucleotides 509 to 1350 downstream of the transcriptional start site (which was isolated as a Scal/Xmnl fragment) and a long arm of homology of 3886bp corresponding to nucleotides 1351 to 5236 downstream of the transcriptional start site (isolated as a Xmnl/EcoRI fragment) of the beta- casein gene (Fig. 3a). The targeting constructs carried either an HPRT selection marker cassette flanked by a pair of incompatible lox sites as described by KoIb et al., 2001 (pB-cas-6; Fig. 3a.), or an HPRT selection marker cassette and a β-galactosidase marker gene separated by a loxP site (pB-cas-7; Fig. 12a).

Exons of the β-casein gene are indicated in Fig. 3a as black boxes, the HPRT and hytk selection marker genes are indicated as cross hatched arrows, respectively and the β-galactosidase gene (b-gal) is indicated as a

black arrow. The PGK promoter elements directing expression of the selection marker genes (HPRT and b-gal) are indicated as black arrowheads. The relative positions of the Hindlll sites (H), the Seal site (S), the Southern blot probe, sizes of hybridising DNA fragments and the primer binding sites (horizontal arrows) used for genotyping are indicated. The positions of the Iox2272 and loxP sites and the translational start codon (ATG) are marked by vertical arrow. Exons are shown as "e" with numerical reference

Example 2: Targeted Modification of the beta-casein Gene Locus

The β-casein gene was modified in HM1 embryonic stem cells by homologous recombination. HM1 embryonic stem (ES) cells (Magin et al., 1992) were cultivated as described previously (Kumar et al., 1994). The b- casein targeting construct (300 mg) was linearized by Kpn\ digestion and transfected into ES cells by electroporation in Hepes-buffered saline (HBS) using a Bio-Rad Gene Pulser at 800 V and 3 mF, as described previously (Kumar et al., 1994). Cells were selected in complete medium containing 500 mg/ml of G418 (Boehhnger Mannheim, Lewes, UK) for 7 days and in complete medium containing 500 mg/ml G418 and 2 mM Gancyclovir (Syntex, Aachen, Germany) for another 7 days. Cell clones carrying the desired integration into one of the b-casein alleles were identified by PCR. Electroporations for the site-specific recombination event were done in HBS at 250 V and 960 mF using 25 μg of the plasmid pBlox1/2-hyg/luc2 and 50 μg of the plasmid pMC1 -Cre (both in circular form). The cells were subsequently grown for 3 weeks in complete medium supplemented with 75 mg/ml of Hygromycin B (Hy; Boehringer Mannheim, Lewes, UK). Individual clones were analysed for the site- specific recombination by PCR and Southern blotting.

The HPRT targeting construct described above and flanked by a pair of incompatible Cre recombinase target sites (loxP and Iox2272) in tandem orintation was introduced into the second exon of the β-casein gene (Fig. 3a, third panel). This insertion removed the translational start codon (ATG) of the β-casein. After transfection and selection in HAT medium, cells carrying the targeted allele were identified by PCR with the primer combination bcas10/bcas3/HPRT2. ES cells carrying the HPRT gene inserted into the β-casein gene by homologous recombination yield a 1093bp PCR product (primed by oligo-nucleotides HPRT2 and bcasi O) in addition to the 1317bp PCR product derived from the unmodified β-casein allele (primed by oligo-nucleotides bcas3 and bcasiO). One of the targeted cell clones (RMCE2272-98), which carries the targeted b-casein allele, is indicated by the presence of the 1093bp band is represented in the PCR in Fig. 3c.

A β-galactosidase reporter gene was subsequently inserted into the second exon of the β-casein gene using Cre-recombinase mediated cassette exchange (RMCE; Fig. 3a bottom panel). RMCE2272-98 cells were transfected with the plasmid pB2272-galhytk containing a β- galactosidase ORF and a hygromycin-phospho-transferase/thymidin- kinase (hytk) expression cassette and a Cre-expression plasmid. Since first and second site specific recombinase sites flanked both the HPRT cassette present in the beta-casein gene and the insert of the transfected plasmid, Cre will recombine the two Iox2272 and the two loxP sites independently of each other resulting in an exchanged iof the HPRT cassette for the plasmid insert cassette (harbouring the beta-galactosidase and hytk genes) (Fig. 3a). Transfected cells were selected in medium containing hygromycin B (HyB) and the resulting cell clones were analysed by PCR using the primer combination bcas3/bcas10/bgalint.1. In addition to the 1317bp product derived from the endogenous allele a

1204bp product is detectable in clones that have been successfully modified by insertion of the β-galactosidase ORF into the β-casein gene (exemplified by cell clones GH1 and GH3 in Fig. 3b). Replacement of the HPRT gene with hytk and b-gal is indicated by the generation of a 1204bp product and concomitant loss of the 1093bp band.

Southern blotting confirmed the above data (Fig. 3c). A fragment of 3.9kb indicative of the endogenous β-casein allele was detected in all samples. An additional band of 6.3kb was detectable in cell clones carrying the β- galactosidase ORF inserted into the β-casein gene. Location of Southern blot probes relative to the modified b-casein locus is indicated in Fig. 3a. The modified ES cells did not express the inserted β-galactosidase gene. However, it has been shown that HM 1 ES cells are able to generate a limited response to lactogenic hormones when a beta-casein driven reporter gene is transfected transiently together with expression vectors encoding the prolactin receptor, the glucocorticoid receptor and STATδa encoding the prolactin receptor and STATδa (A. KoIb, unpublished data).

Example 3: Cell fusion

In order to overcome the limitations associated with embryonic stem cells and murine mammary gland cells and in order to assess gene transgene expression in the correct chromatin context, a fusion cell system was developed wherein gene targeted embryonic stem cells are fused with differentiated mammary gland cell lines.

The modified ES cells derived from Example 1 were seeded into one gelatinised well of a 6-well plate together with an equal cell number of HC11 mammary epithelial cells (Ball et al., 1988) (5x10 ⁵ cells of each cell type) and incubated overnight in the cell culture medium regularly used for

HM 1 ES cells (see above). 24h later the cells were treated with 1 ml of with polyethylenglycol (PEG, Sigma) for 90 sec, washed 4 times with PBS and incubated in fresh (ES cell) medium over night. After another 24h- incubation period the cells were split into non-gelatinised 10cm dishes and selected in RPMI 1640 medium supplemented with 10% foetal calf serum, 10 U/ml of penicillin/streptomycin, 2 mM glutamine (all from Life Technologies), 5 mg/ml of bovine insulin, 10 ng/ml of murine EGF (SIGMA) and 100μg/ml of hygromycin B (HyB) for 14 days. The RPMI cell culture medium used at this stage favours growth of HC11 cells with the absence of a gelatine layer limiting growth of ES cells. Selection in HyB solely enabled growth of HC11 cells which had taken up at least a part of the ES-cell-dehved casein gene locus carrying the integrated hygromycin- resistance marker gene.

Selection in medium containing 50, 100, 150 and 200μg/ml HyB gave rise to cell clones carrying at least one modified β-casein allele including the β- galactosidase and hytk marker genes as evidenced by Southern blot analysis of cell pools derived under the different concentrations of the selective agent (Fig. 3c). Densitomethc scanning demonstrated that the non-modified and modified allele are present in the cells in approximately a 1 :1 ratio, indicating that selection at 50μg/ml HyB is sufficient to reliably derive HC11 cell clones which have taken up the ES cell derived chromosome. Metaphase chromosome spreads of HC11 cells, embryonic stem cells and fusion cell clones were subsequently analysed for their respective chromosome numbers. As expected HC11 cells, which are known to be tetraploid (Pauloin et al., (2002) Gene 283:155-162) yielded an average number of 80 chromosomes, whereas the diploid HM1 ES cells showed an average number of 40 chromosomes. The fusion cell clones yielded an average number of 81.5 chromosomes indicating that only one or two chromosomes had been taken up by the HC11 cells, at

least one of them being chromosome 5 carrying the modified β-casein gene. An example of a chromosome spread is shown in Fig. 4.

The hytk selection marker gene encodes a fusion protein of the hygromycin-phosphotranferase and the Herpes Simplex Virus (HSV) thymidine kinase. Stable incorporation of the gene renders a cell resistant to the drug Hygromycin B and sensitive to the nucleotide analogue Gancyclovir. Representative clones of cells derived from the cell fusion were subsequently pooled and incubated in medium containing varying amounts of Gancyclovir to determine whether the incorporation of the hytk selection marker gene would be stable under these conditions or whether the chromosome transferred from ES cells to the HC11 cells would be lost rapidly. Cells were counted 10 days after the onset of selection. Cell clones derived from cell fusion were sensitive to the selection, whereas the parental HC11 cells were not (Fig. 5). As shown in Fig. 5, non modified HC11 cells, increasing concentrations of gancyclovir resulted in increased cell death, whereas fused cells were unaffected by gancylovir. This indicates that the incorporated chromosome is indeed stably transmitted through cell divisions.

Example 4: Analysis of hormone responsiveness

Hygromycin resistant cell clones were seeded into 24 well plates and analysed for their responsiveness to lactogenic hormones. HC11 cells are grown to confluence and then induced with RPMI 1640 medium supplemented with 10% FCS, 2mM glutamine, 5μg/ml insulin, 5μg/ml prolactin, 1 μg/ml hydroxicortisone and antibiotics, transcription of α-, β and γ-casein genes is activated (KoIb, 2002). In order to determine whether the hygromycin resistant fusion cell clones would also activate the b-galactosidase marker gene inserted into the β-casein gene,

representative fusion cell clones were seeded into 24-well plates and grown to confluence for 56h before the induction medium was added, b- galactosidase expression was analysed 48h post induction using X-gal staining. Cells in the hormone-stimulated and control plates were fixed in 2% paraformaldehyde and stained with X-gal as described in Price et al., 1987. All clones showed some degree of staining after induction with lactogenic hormones, whereas non-induced cells showed very little or no X-gal positive cells (Fig. 6). The observation that in none of the cell clones all cells were X-gal positive is consistent with published data which demonstrated that only around 15% of HC11 cells express β-casein protein as detected by immuno-histochemistry (Ball et al., (1988) EMBO J. 7:2089-2095). The induction of the β-galactosidase protein synthesis was also confirmed by analysis of β-galactosidase protein activity using a commercially available kit (Promega, data not shown).

Example 5: Characterisation of the chimeric b-casein/b-gal transcript

In order to confirm that transcription of the β-galactosidase gene was derived from the β-casein promoter an RT-PCR analysis was performed using a primer binding within the first exon of the β-casein gene and a primer binding within the β-galactosidase open reading frame. The β- galactosidase gene was isolated from the vector pCMV-β (Clontech) carrying an intron sequence derived from SV40, which includes two alternative splice acceptor signals. Splice donors are marked as "D" and splice acceptors as "A" in Fig. 7. Exons incorporated into possible splice variants are depicted as green boxes and sizes of predicted transcripts are indicated. Several PCR products were observed in cDNA derived from fusion cells induced with lactogenic hormones (Fig. 7b), corresponding to three of the four theoretically possible splice variants.

Quantitative PCR reactions in which cDNA derived from untreated fusion cells (F13 [-]LH) or fusion cells induced by lactogenic hormones (F13 [+] LH) were used as templates to show a prevalence of the 233bp product (Fig. 7c). Sequence analysis of the cloned PCR products confirmed that they correspond to the expected splice variants (data not shown).

The expression of the endogenous β-casein gene and the chimeric β- casein/β-galactosidase gene was then analysed in hormone stimulated cells and control cells using quantitative PCR. Expression and induction of the endogenous β-casein gene was highly variable in the different cell clones derived after cell fusion (Fig. 8). Induction with lactogenic hormone is indicated by [+], and non induction by [-]. The varying cell clones also displayed different morphological phenotypes (cf. Fig. 6), confirming previous observations that HC11 cells possess stem-cell like properties and can be used to isolate various mammary cell types (Deugnier et al., (1999) J. Cell Sci. 112:1035-1044).

Expression rates and response to hormone treatment of the β- galactosidase gene in the different cell clones was directly correlated with the expression and stimulation of the endogenous β-casein gene. Induction of the β-galactosidase gene was less pronounced than induction of the endogenous gene. E.g. a maximum rate of 330 fold induction of the endogenous β-casein gene was observed in fusion clone 13, whereas the induction of the chimeric gene was 67 fold (Fig. 9, Table 4). On average the induction of the endogenous β-casein gene was around 4 to 5 times more efficient than that of the modified β-casein gene (Table 4).

In absolute terms the expression of the chimeric gene (6pg β-galactoidase per pg actin) is similar to that of the endogenous gene (3pg β-casein per pg actin in the induced state). However, expression of the non-induced β-

galactosidase gene is significantly higher than that of the endogenous β- casein gene (Fig. 9).

Example 6: Long range RMCE at the beta and gamma casein gene loci

In order to establish a system in which fragments of the β-casein promoter can be analysed for their contribution to the activity and hormone responsiveness of the β-casein promoter in the context of the entire casein locus, HM1 cells were modified by two homologous recombination events.

Both the β-casein and the γ-casein gene were targeted introducing a loxP site into the β-casein gene and a Iox2272 site into the γ-casein gene (Fig. 10a), making the entire intervening sequence of 75kb (75145bp between the transcriptional initiation sites) accessible to modification by site-specific recombination, as indicated in Fig. 10. Modifications introduced by gene targeting are shown beneath the demonstrated EST (expressed sequence tag) sites. A loxP site flanked by a PGK-HPRT expression cassette and a promotorless b-gal gene is inserted into the b-casein gene. A Lox2272 site flanked by a neomycin-phosphotransferase gene is inserted into the gamma-casein gene. The sequence of the two incompatible lox sites is shown on Fig. 10b.

The γ-casein targeting construct was generated in analogy to the β-casein targeting vector using sequences derived from the first intron (for the short arm of homology) and sequences downstream of exon 2 (for the long arm of homology) (Fig. 11 a). EcoRI sites are shown as E, EcoRV sites as EV and Sphl site as S. Southern blot probes, sizes of hybridizing DNA fragments and primer binding sites (horizontal arrows) are also indicated. Position of the lox 2272 site and the translation start codon are marked by

vertical arrows. Correct targeting of the γ-casein gene was assayed by PCR analysis (Fig. 11 b) and confirmed by Southern blotting (Fig. 11 c). PCR analysis results (Fig. 11 b) assessed genomic DNA isolated from cell clone HM1 -gamma-#185 and parental HM1 embryonic stem cells using primer combination gcas9, gcase9 and PGK3 (lanes 1 and 2) and gcas 8 and pPGK3 (lane 3). A 1483bp band represents the unmodified gamma- casein allele and 1391 bp band is indicative of the targeted gamma-casein allele.

Southern blot analysis of HM1 cells, and gamma casein targeted cell clones #155 and #185 is shown in Fig. 11c. DNA was digested with EcoRI, separated on a 1 % agarose gel, blotted onto Hybond N+ nylon membrane and hybridised with the probe indicate din Fig. 11 a corresponding to a 919bp Sphl/EcoRV fragment encompassing gamma- casein exon 1 and parts of intron 1.

The targeted cells were used as basis for a second gene targeting approach using a new β-casein targeting vector (pB-cas-7) introducing a selectable marker gene (HPRT) and the β-galactosidase ORF into the β- casein gene (Fig. 12a). The HPRT selection marker gene was joined to the 5' end of the β-galactosidase gene and the two genes separated by a loxP site. A Cre-mediated cassette exchange event removes the HPRT gene from the β-casein insertion site, replacing it with the incoming cassette. Successful modification of the ES cells was analysed by PCR and Southern blotting (data not shown).

Analysis of the double targeted cells was undertaken to assess if the minimal 380bp β-casein promoter, when placed in the context of the casein gene locus in the absence of further β-casein promoter regions was able to mediate a full response to lactogenic hormones. This fragment

contained the only two consensus STAT5 binding sites present in the β- casein promoter.

The promoter fragment present in the plasmid pB2272-hytk-INS-βcas was inserted between the v- and β-casein genes together with an insulator element ("INS") and a PGK-hytk expression cassette (Fig. 12a). The insulator was introduced in order to block the influence of the upstream PGK promoter on the expression of the β-galactosidase marker gene. The construct was transfected into the doubly targeted ES cells together with a Cre expression plasmid and cell clones were selected using medium containing HyB. Cell clones were analysed for the correct integration by PCR. When analysed with the primer combination bcas10/bcas3/bgalint.3/HPRT2, a 1317bp product (derived from amplification of the unmodified β-casein allele with bcas3/bcas10) and a 983bp product (derived from the targeted β-casein allele with bcas10/HPRT2) are detected in the parental cell clone β/γ-#57 (Fig. 12b). Successful integration of the RMCE plasmid which leads to the removal of the HPRT selection marker gene and the concomitant insertion of the 380bp β-casein promoter fragment results in the formation of an indicative 1093bp fragment, derived from amplification with the primer pair bcas10/bgalint.3 (recombination via the loxP site; Fig. 12b).

When analysed with the primer combination PGK5/pBKpA2/bgalint.1 , a 437bp product (derived from amplification of the targeted β-casein allele with PGK5 and bgalint.1 ) is detected in the parental cell clone β/γ-#57. Successful integration of the RMCE plasmid leading to the removal of the HPRT selection marker gene and concomitant insertion of the 380bp β- casein promoter fragment resulted in the formation of an indicative 205bp fragment, derived from amplification with the primer pair PGK5 and

pBKpA2 (recombination via the Iox2272 site; Fig. 12b). The PCR results were again confirmed by Southern blot analysis (data not shown). 5 of 14 selected clones carried the desired mutation after HyB selection (4 of the clones are represented in Fig. 12b) and were also sensitive to selection with HAT medium and resistant to selection with medium containing 6TG (6-thio-guanine). This indicated that the targeted β- and v- casein genes reside on the same chromosome, as an RMCE process involving two chromosomes would have retained the HPRT selection marker gene in the genome.

Modified ES cells clones were fused with HC11 cells and selected in medium containing hygromycin B or G418 (both selections are possible, as a functional neomycin resistance cassette is contained within the modified γ-casein gene). Four resulting fusion cell clones (Hy9, Hy11 and Hy12 selected in HyB and G1 selected in G418) were analysed for expression of β-casein and β-galactosidase mRNA. The endogenous β- casein gene was induced between 7 and 316 fold (Table 4) indicating that the derived cell clones display varying degrees of hormone responsiveness. In contrast, the induction of the β-galactosidase gene from the β-casein promoter was similar in all clones and in the range of induction which is usually obtained after transfection of a reporter gene constructs carrying this segment of the β-casein promoter (Table 4). In absolute terms expression from the minimal promoter is one order of magnitude lower (i.e. 3 to 6 fg b-gal/pg actin) than expression from complete promoter in cell clone F13 (6 pg b-gal/pg actin).

In a control experiment an empty vector cassette (plasmid pB2272, Fig. 12a) was introduced into the modified casein gene locus and fusion cell clones were derived. Residual expression of β-galactosidase was detectable (presumably directed by the PGK promoter which is adjacent to

the β-galactosidase gene after the RMCE event) and in the fg/pg actin range. No significant hormone stimulation of β-galactosidase reporter gene expression was observed in these cells (cf. exemplary fusion clone E3; Table 4), whereas the endogenous β-casein gene was induced by 250 fold in the same cells. This confirms that the presence of the STAT5 sites is required for induction by lactogenic hormones but insufficient to mediate full hormone responsiveness of the modified locus.

Example 8: Gene targeting of the murine alpha casein gene

A construct including a short arm of homology derived from the first intron of the a-casein gene of 684bp (isolated as an EcoRV/BssHII fragment) and a long arm of homology encompassing 6696bp including sequences from intron 2 to intron 12 (isolated as a Stul/EcoRV fragment) was generated (Fig. 14). The targeting construct carried an expression cassette encoding a hygromycin-phosphotranferase-thymidine kinase (hytk) fusion protein under control of a PGK promoter inserted in place of the second exon of the a-casein gene (Fig. 14). A homologous recombination event between the α-casein gene and the construct fully removed the second exon of the casein gene plus 146bp of the first intron and 331 bp of the second intron (Fig. 15).

The hytk selection marker gene is flanked by a pair of directly repeated incompatible lox sites (loxP/lox2272). These sites can be utilised as a platform for the stable integration of cassettes flanked by an identical pair of lox sites using the Cre recombinase. In Figs. 14 and 15 exons 1 to 34 are represented as horizontal lines and marked e1 and e34. Genomic regions used for the establishment of the a-casein targeting construct (a- cas-TC) and the resulting genomic structure of the targeted gene is shown. The hygromycin-phosphotransferase-thymidine kinase fusion gene

is represented as a black box. The lox sites flanking the selection marker gene are indicated by vertical arrows. Open reading frames (ORF) which could be utilised after the targeting event are indicated at the bottom of Fig. 12. A short ORF can be generated from an out of frame methionine codon in exon 3. Usage of an in-frame methionine codon in exon 6 can lead to the generated of truncated a-casein protein which lacks both the signal peptide and the serine cluster.

The targeting construct successfully modified the α-casein gene in embryonic stem cells with a frequency of 5% (5 targeted cell clones in 96 cell clones analysed) and 1 % in (4 targeted cell clones in 425 cell clones analysed, Table 5) in two consecutive experiments. A targeting experiment using the HPRT selection marker gene instead of the hytk gene resulted in a targeting frequency of 3.7% (9 targeted cell clones in 240 cell clones analysed, Table 5). The cell clones were genotyped by PCR with a combination of three oligo-nucleotides (acasθ, acas7, pBK-pA2) as shown in Fig. 16a. The unmodified α-casein gene yields a PCR product of 1566bp, whereas the targeted allele yielded a product of 848bp (Fig. 16b for exemplary clone 4B4).

Example 9: Insertion of the transgene into the casein locus at the loxP sites

Targeted cells were used as basis for a recombinase mediated cassette exchange (RMCE) protocol. The targeted cells were electroporated with 20μg of the plasmid pB2272-A1 L-PGKneo and 40μg of the plasmid pMC1 - Cre. The transgene selected was AIL corresponding to the light chain of an antiviral antibody. The transfected cells were incubated in a selection medium containing 200μg/ml G418. Resistant cell clones were grown up and analysed for successful recombinase mediated cassette exchange

using the primer combination acasθ, acas7, pBKpA2 and A1 Lsp.2 (Fig. 17a). DNA derived from cell clones which had undergone the cassette exchange displayed a 1141 bp band (corresponding to the modified a- casein allele) in addition to the 1566bp band derived from the unmodified allele (Fig 17b). Concomitantly the 848bp band generated by the primer combination pBKpA2/acas6 on the targeted allele is lost. Between 20% and 50% of the G418 resistant colonies showed the indicative 1141 bp band (Table 5).

The DNA of targeted and RMCE modified ES cells was analysed by Southern blot. Two probes corresponding to the 5' and the 3' ends of the a-casein gene were utilized. As schematically shown in Fig. 18a the unmodified locus yields a 7.5kb band when hybridized with either probe. The 5' end of the targeted locus will yield a 4.3kb band with the 5' probe (#4478), whereas the 5' end of the RMCE modified a-casein gene yields a 8.8kb band when hybridized with the same probe. The 3' probe detects a 3.3kb fragment in targeted cells and a 4.2kb band in RMCE modified cells. Bands consistent with this expectation were found after hybridization with the 5' probe (#4478, Fig. 18b). Fig. 15b shows the results of Southern blot analysis of cell lines 4B4-2E6 (RMCE modified), 4B4-2A5 (RMCE), 4B4- 4A8 (RMCE), HM1 #139 (targeted), HM1 #6299-5 (targeted), HM1 #6299- 12 (RMCE) and non-transfected embryonic stem cells (HM1 ). 10μg of genomic DNA was digested with EcoRI separated on a 1 % agarose gel and blotted to Hybond N+ membrane. The membrane was then hybridized with probe #4478 and exposed on an X-ray film. The positions of the bands indicative of the unmodified, the targeted and the RMCE modified allele are indicated by arrows.

Example 10: Generation of Mice including loxP sites at the Casein locus

Transgenic mice were derived from the RMCE modified ES cells and analysed for expression of the modified a-casein gene (now carrying the A1 L coding region).

Expression was analysed by RT-PCR using the primer combination acas9a/acas5 (to detect mRNA derived from the unmodified allele) and the primer combination acas9a/A1 Lsp.2 (to detect expression of the chimeric mRNA encompassing exon 1 and parts of exon 2 of the a-casein gene and the complete coding region of the A1 L gene, Fig. 19a).

Fig. 19a is a schematic representation of the mouse a-casein gene modified by gene targeting and RMCE and the mRNAs derived from the modified and unmodified allele. The unmodified a-casein gene gives rise to a 1443nt mRNA. The a-casein exons are indicated as e1 and the exons derived from the A1 L gene are shown asAei . The primer pair acas9a/acas5 amplifies an 881 bp PCR product from the corresponding cDNA. The primer acas9a binds within the first exon of the a-casein gene, primer acasδ in the 32nd exon. The primer A1 Lsp.2 binds within the variable region of the light chain gene A1 L, which constitutes the 3rd exon of the chimeric mRNA derived from the modified a-casein allele. The formation of a 165bp PCR product using cDNA from transgenic mice as a template indicates successful expression directed by the a-casein promoter.

PCR analysis of cDNA derived from the lactating mammary tissue of a control mouse and two transgenic mice carrying the modified a-casein allele (AKO2.2.22, AKO2.2.54) are shown in Fig. 19b. The position of the

PCR products specific for the unmodified and the modified a-casein genes are indicated by arrows, a-casein specific RNA was detected in two strains of transgenic mice and in control mice as evidenced by the occurrence of the indicative 881 bp PCR product generated by the primer pair acas9a/acas5. In addition the 165bp product indicative of the chimeric mRNA could be detected in the two strains of transgenic mice but not in control mice. The PCR products were cloned and sequenced. The sequence exactly matched the predicted sequence (data not shown).

Quantitative analysis of the cDNAs derived from the modified and unmodified alleles is shown in Fig. 19c. 1 :10 and 1 :100 dilutions of cDNA reactions derived from lactating mammary tissue were analysed for expression of GAPDH, a-casein and a-casein-A1 L specific cDNAs. Expression of the a-casein and a-casein-A1 L cDNAs is correlated with GAPDH expression [pg/pg]. Expression of the a-casein gene was reduced to around 40% in the two transgenic strains consistent with the expectation that expression in these mice can only occur from one allele (yielding 0.25 to 0.4pg of a-casein specific cDNA per pg of GAPDH cDNA), but from both alleles in control mice (yielding 0.75pg of a-casein/pg GAPDH cDNA, Fig. 19c). Expression of the transgene was (as expected) only detected in transgenic mice and was at around 0.15 to 0.27 pg of transgene specific cDNA per pg of GAPDH cDNA. This shows that expression from the modified allele is similar to that of the corresponding unmodified allele in transgenic mice.

Milk protein was analysed for the presence of the A1 L light chain protein by Coomassie staining and Western blot analysis. The staining of milk samples shows the presence of an additional milk protein at around 25kDa in milk derived from transgenic mice (A1 L, Fig. 20). Milk from two lines of transgenic animals carrying the modified a-

casein gene and one control animal was separated on a 10% polyacrylaminde gel alongside human IgG preparations ok know concentrations and stained with Coomassie blue. The positions of the heavy and light chains of AIL are indicated.

The presence of the A1 L protein can also demonstrated directly by Western blotting. The antibody carries constant regions of human origin. These can be detected with a human IgG specific antiserum which does not cross react with mouse IgG. In comparison to a control human IgG preparation the intensity of the milk derived A1 L protein is similar to a 10ng/μl solution. The milk samples were diluted 1 in 10, which suggests that the concentration of the light chain protein in milk is around 100ng/μl (Fig. 21 ).

Example 11 : Recombinase mediated cassette exchange into genomic targets using an adenovirus vector

Human embryonic kidney cells (HEK293; ECACC No: 85120602) were cultivated in high glucose DMEM medium supplemented with 10% foetal calf serum, Penicillin/Streptomycin and 2mM Glutamine. Baby hamster kidney cells (BHK 21 ; ECACC No: 85011433) were cultivated in MEM medium supplemented with 10% foetal calf serum, Penicillin/Streptomycin, 2mM Glutamine, 0.1 mM non-essential amino acids and 1.0 mM sodium pyruvate. HM1 mouse embryonic stem cells (Magin, T.M et al., (1992) Nudeic Acids Res. 20, 3795-3796) and HC11 mouse mammary gland cells (Ball et al., (1998) EMBO J. 7, 2089-2095) were cultivated as described (KoIb et al., (1999) Gene 227, 21 -31 ). The cell line HM1 RMCE2272-98 has been described before and contains an HPRT selection marker gene inserted into the murine b-casein gene by homologous recombination (KoIb et al., (2001 ) Anal Biochem, 290, 260-271 ). Cell culture reagents

were purchased from Invitrogen, foetal calf serum was purchased from Sigma.

HEK293 and HC11 cells were transfected using the Gene Juice Reagent (Novagen) according to the manufacturer's recommendations. HM1 ES cells were electroporated using a Bio-Rad Gene Pulser using various conditions as indicated in Tables 7 and 8. HC11 cells were electroprated in a 4cm cuvette at 250V and 950μF.

All transfections were carried out using identical amounts of plasmid DNA. In control reactions the plasmid pCMV-bgal (Clontech) was co-transfected.

The plasmid pBK2272-HPRT was derived from the plasmid pPGK-HPRT (KoIb et al., (1999) Gene 227, 21 -31 ) from which the 2.9kb EcoRI insert was excised and inserted into the plasmid pBK-CMV-2272, which carries a pair of incompatible lox sites (loxP and Iox2272) at the fringes of its multiple cloning site. The loxP site is placed between the Kpnl and Xhol restriction sites, and the Iox2272 site is placed between the Xbal and Notl sites. pB2272-neo is based in the plasmid pB2272 (KoIb et al., (2001 ) Anal Biochem 290, 260-271 ) and carries the same pair of incompatible lox sites at the fringes of its multiple cloning site. The PGK-neo expression cassette is inserted as EcoRI/Hindlll 1.8kb fragment derived from the plasmid pB- PGKneo. The plasmid pB-lox1/2-hyg/luc2 was described previously (KoIb & Siddell (1997) Gene 203, 209-216).

The plasmids used for the generation of adenovirus vectors are based on the Stratagene Ad Easy system and the system was used as recommended by the supplier. The plasmid pShuttle was used as basis for the generation of vectors carrying (1 ) a Cre-expression cassette derived from the plasmid pMC1 -Cre (GU et al., (1973) Cell 73 1155-1164), (2) a

PGK-neo expression cassette derived from pB2272-neo and (3; in case of the plasmid pShuttle-G5) 2 copies of the chicken b-globin insulator element (a gift of Gary Felsenfeld; NIH, Bethesda). Average virus titres were in the range of 10 ⁶ pfu/ml. Infections were performed at an MOI of 1.

The vector pB-cas7-2272-hytk carries a PGK-hytk (hygromycin- phosphotransferase thymidine kinase fusion gene) expression cassette flanked by a pair of incompatible lox sites (Iox2272 and loxP) embedded into the backbone of a mouse b-casein gene. The construct was derived from the plasmids pB-bcas6 and pB2272-hytk (KoIb et al., (2001 ) Anal Biochem 290, 260-271 ) by in vitro Cre mediated recombination. In order to generate stable cell clones an 8kb Nhel/Sfil fragment was excised from the vector and electroporated into HEK 293 cells (1 μg of DNA in 1 x 10 ⁷ cells in a 0.2cm cuvette at 110V).

PCR amplifications were done using Taq Polymerase from various suppliers. Oligonucleotides were purchased from MWG or Sigma- Genosys. Primer sequences, amplicon size and annealing temperatures are given in Table 6. Template DNA for PCR analyses was isolated as described (KoIb et al., (1999) Gene 227, 21-31 ).

The Cre recombinase system can be used to insert genes at predefined sites in the mammalian genome which have been tagged with a lox recognition site. This has been demonstrated in a number of cell types and also (albeit at reduced efficiency) in fertilised mouse oocytes (Waterhouse et al., (1973) Nucleic Acids Res. 21 , 2265-2266;Bonhassira et al., (1997) Blood 90, 3332-3344; Schmerling et al., (2005) Genesis 42, 229-235). We the present inventors utilised the Cre recombinase to insert genes at the murine b-casein gene with a view to expressing these genes in the milk of transgenic animals (KoIb (2002) Cloning Stem Cells 4, 65-80; KoIb (2002)

Mammary Gland Blood Neoplasia 7, 123-134; KoIb (2001 ) BiotechnolAdv 19, 299-316). The present inventors chose the b-casein gene as a target site which is able to equip an inserted transgene with all required regulatory elements to allow for abundant expression of the transgene in the lactating mammary gland (Coates et al., (2005) Trends Biotechnol 23, 407-419).

Electroporation conditions determine the frequency of site-specific recombination

A pair of incompatible lox sites (loxP and Iox2272) in conjunction with a b- galactosidase open reading frame and a hytk (hygromycin- phosphotransferase thymidine kinase fusion gene) selection marker expression cassette was inserted into the second exon of the b-casein gene in HM1 embryonic stem cells using sequential homologous and site- specific recombination (KoIb (2001 ) Anal Biochem 290, 260-271 ). The ATG of the b-casein gene was deleted during that process, such that the ATG of a b-galactosidase is the first translation start codon in a chimeric mRNA initiating at the 1 ^st exon of the b-casein gene (Fig. 22). As shown in Fig. 22a, In Fig. 22, exons of the β-casein gene are indicated as solid boxes, the neomycin (neo) and hytk selection marker genes are indicated as solid arrows, respectively and the b-galactosidase gene (b-gal) is indicated as a hatched arrow. The PGK promoter elements directing expression of the selection marker genes are indicated as black arrowheads. The positions of the Iox2272 and loxP sites and the translational start codon (ATG) are marked by vertical arrows. The primer binding sites (horizontal arrows) used for genotyping and the sizes of the expected PCR products are indicated.

The resulting cell line HM1 RMCE2272-gal/hytk [GH1] was then transfected with the plasmid pB2272-neo together with a two fold excess of a Cre expression vector. The cells were then selected in medium containing 200μg/ml G418. Under these conditions cells which have incorporated the selection marker gene at random sites or by site-specific recombination will survive. The present inventors found that, surprisingly, the frequency of site-specific integration was heavily dependent on the electroporation conditions used (Table 7). At 250V and 960μF almost all of the colonies selected carried a site-specific insertion of the neo gene at the b-casein locus. In contrast transfections carried out at 800V and 3μF, although generating the same number of resistant colonies, showed no site-specific insertion of the neo gene into the predefined b-casein target site. Fig. 22b illustrates the results of the PCR analysis of genomic DNA isolated from the cell clones HM1 RMCE2272-gal/hytk [GH1], and the cell clones HM1 N1 -2272 to HM1 N7-2272 derived from it. A 1317bp band is detected in all samples and represents the unmodified β-casein allele. HM 1 RMCE2272-gal/hytk [GH 1] cells carry an insertion of a b- galactosidase open reading frame and a PGK-hytk expression cassette at one of the b-casein alleles as indicated by the occurrence of a 1203bp PCR product. Cell clones HM1 N1 -2272 to HM1 N7-2272 were derived after an RMCE event which exchanged the b-gal and hytk genes for the neo selection marker gene. The correct modification is indicated by the generation of a 1023bp PCR product and the concomitant loss of the 1203bp band. Phage λ DNA digested with Hindlll and EcoRI was used as molecular weight marker.

In order to define the parameters which determine the frequency of site specific integration, the present inventors conducted a number of electroporations under different conditions. The results do not single out one parameter as being crucial. Although for most transfections an

increase in capacitance leads to an increase in site-specific integration frequency, electroporations carried out a low voltage and high capacitance failed to generate colonies which carried a site-specifically integrated selection marker gene [Table 8]. Likewise transfection efficiency (as indicated by numbers of selected cell clones) itself does not seem to be the sole determinant of site-specific integration frequency.

Liposome mediated gene transfer supports RMCE

Next the present inventors determined whether a chemical (rather than a physical) method for DNA transfer would also be able to mediate site- specific insertion of the transgene cassette. As HM1 ES cells do not display high transfection frequencies with liposome reagents, this experiment was carried out in a derivative of the mouse mammary gland cell line HC11 cells termed HC11 -bcas-F9 (Robinson & KoIb, unpublished data). The cells carry the same modification as the HM1 ES cells used above and were transduced with the plasmids pB2272-neo and pMC1 -Cre using the Novagen Gene Juice reagent and, in parallel, by electroporation. 17% of the selected cell colonies carried the correct site-specific integration of the marker gene into the b-casein locus when the cells were transfected using Gene Juice. Electroporation with the same construct lead to 28% of the cells carrying the correct integration. The absolute number of resistant cell colonies was similar for both techniques (Table 9), however, addition of Cre to the transfection mixture decreased the number of resulting selection resistant colonies when the cells were transfected with Gene Juice and increased the number of resistant colonies when the cells were electroporated. This experiment demonstrates that successful RMCE is not dependent on electroporation as a gene transfer methodology.

The recombinase expression cassette and the transgene cassette can be delivered in one plasmid

The present inventors considered if it would also possible to supply the recombinase and the transgene in the context of a biological gene transfer vehicle e.g. a viral vector. Transfections aimed at achieving recombinase mediated cassette exchange utilise an excess of Cre expression plasmid over the plasmid carrying the transgene construct (KoIb (2001 ) Anal Biochem, 290, 260-271 ; Araki et al., (1995) Proc Natl Acad Sci VSA 92, 160-164; Araki et al (1997) Nucleic Acids Res 25, 868-872) In the experiments described above a two or three fold excess of Cre expression construct was used. In the context of a viral vector carrying both the transgene and the recombinase gene, the genes would be present at equimolar ratio. In order to test whether such a construct would support site-specific recombination of a transgene, a plasmid carrying a neomycin resistance marker gene flanked by a pair of incompatible lox sites was generated. This DNA segment was joined to a Cre expression cassette in the context of the plasmid pShuttle (Stratagene) (Fig. 23a).

Firstly, the construct was co-transfected into BHK cells together with an acceptor plasmid (pBK-2272-HPRT) carrying the same pair of lox sites (Fig. 23b). As shown in Fig. 23c the plasmid (termed pShuttle-H6) supported site-specific recombination with the acceptor plasmid in the presence or absence of any additional Cre expression vector (see lanes H6 and H6 + pMC1 -Cre) as demonstrated by the occurrence of the indicative 421 bp PCR product. In Fig. 23c the 2941 bp product is generated from the non-recombined pBK2272-HPRT plasmid. The 421 bp PCR product is indicative of a recombinase mediated cassette exchange between the PGKneo cassette and the PGK-HPRT cassette. Phage λ

DNA digested with Hindlll and EcoRI was used as molecular weight marker.

The construct pShuttle-H6 was then transfected into HM1 -2272-#98 cells (KoIb (2001 ) Anal Biochem 290, 260-271 ). These derivatives of HM1 cells carry a modified b-casein locus incorporating an HPRT selection marker gene flanked by the loxP/lox2272 pair of Cre target sites. Interestingly, none of the G418 resistant cell clones derived from this transfection carried a site-specific insertion of the neo cassette at the b-casein gene irrespective of the transfection conditions (data not shown). In the plasmid pShuttle-H6 the neo expression cassette and the Cre expression cassette are placed in tandem. A corresponding plasmid in which the orientation of the neo expression cassette was reversed also failed to yield any G418 resistant cell clones which carried a site-specific insertion after transfection into HM 1 RMCE2272-98 cells (data not shown). The present inventors have previously shown that expression cassettes in close vicinity can interfere with each others expression (KoIb & Siddel (1997) Gene 203, 209-216). Therefore the vector was modified by the insertion of two copies of the chicken b-globin insulator between the Cre and neo expression cassettes (Fig. 24a). The resulting vector (pShuttle-G5) was again transfected into HM1 #98 cells and G418 resistant colonies were derived. This time the vast majority of stable cells (85%) had taken up the neo cassette by site-specific recombination (as shown for representative clones in Fig. 24b). Fig. 25b the 1317bp band is detected in all samples and represents the unmodified β-casein allele. Cell clones modified by an RMCE event, which has inserted the neo selection marker gene, display an additional 1023bp PCR product. Phage λ DNA digested with Hindlll and EcoRI was used as molecular weight marker.

This data confirms that it is possible to provide the Cre expression cassette and the transgene which is to be integrated into the host genome in a single contiguous DNA segment.

An adenovirus vector can mediate successful RMCE

The present inventors established an adenovirus vector based on the plasmid pShuttle-G5 and tested its ability to support site-specific recombination in HEK 293 and BHK cells. HEK 293 cells carry segments of the adenovirus genome which are deleted from the recombinant virus and complement the viral vector to allow virus replication. BHK cells in contrast do not support viral replication. The cells were first transfected with an indicator plasmid (pBlox1/2-hyg-luc2 ) (KoIb & Siddle (1997) Gene 203, 209-216) and 24h later infected with the G5 adenovirus vector and a control adenovirus vector (expressing b-galactosidase). As control the cells were co-transfected with a Cre expression plasmid (pMC1 -Cre). DNA was isolated from the cells 24h post infection and analysed using the primer combination hytk1/lucint.2 (Fig. 25a). A PCR product indicative of successful site-specific recombination (519bp) could be detected in HEK 293 cells transfected with the plasmids pMC1 -Cre or infected with the virus G5 (Fig. 25b). In Fig. 25b, successful recombination is indicated by the presence of the 519bp PCR product (indicated). Non-recombined plasmid yields a PCR product of 3426bp. Phage λ DNA digested with Hindlll and EcoRI was used as molecular weight marker.

As can be seen, no recombination could be detected in cells transfected with the plasmid pCMV-βgal or infected with the control virus. In these reactions the primer pair detects a PCR product of around 3kb which is indicative of the non-recombined plasmid (Fig. 25a). In BHK cells recombination of the indicator plasmid can only be detected in cells co-

transfected with the positive control plasmid pMC1 -Cre. No recombination can be detected in cells infected with the adenovirus vector G5 indicating that in the absence of viral replication insufficient amounts of Cre protein are generated to support recombination.

The present inventors subsequently analysed whether the transgene cassette present in the adenovirus vectors can be mobilised such that it integrates site-specifically into an acceptor plasmid. HEK 293 and BHK cells were transfected with the acceptor plasmid pBK2272-HPRT and subsequently infected with the adenovirus vector G5 or the control virus. As control for the recombinase activity the cells were also transfected with the plasmid pMC1 -Cre. To control for the availability of the mobilisable transgene cassette the cells were also transfected with the plasmid pB2272-neo, which corresponds exactly to the cassette present in the viral vector. DNA was isolated from the transfected and infected cells 24h post infection and analysed by PCR with the primer combination: CMVseq.1/PGK5/pBKpA. In the absence of recombination a PCR product of 622 bp is expected, whereas successful recombination is evidenced by the occurrence of a 421 bp product (Fig. 26a).

As shown in Fig. 26b transfection of HEK 293 cells with the acceptor plasmid pBK2272-HPRT, the donor plasmid pB2272-neo and the Cre expression plasmid pMC1 -Cre leads to the generation of a 421 bp product indicative of a successful site-specific recombination. The same product is also generated when the cells are infected with the adenovirus vector G5 irrespective of whether the donor plasmid pB2272-neo was co-transfected. The 622bp PCR product is generated from the non-recombined pBK2272- HPRT plasmid. The 421 bp PCR product is indicative of a recombinase mediated cassette exchange between the PGKneo cassette and the PGK-

HPRT cassette. Phage λ DNA digested with Hindlll and EcoRI was used as molecular weight marker.

This indicates that in this assay format the G5 virus is capable of expressing sufficient amounts of Cre and to act as donor of the 2272-neo cassette. The PCR analysis shown in Fig. 5b analyses the 3' end of the recombination event (Fig. 26a). Analysis of the 5' end of the same fusion confirms these results (data not shown).

The present inventors subsequently assessed whether the adenoviral vector would also be able to support integration of the neo-selection marker cassette from the virus into a target site embedded into the genome. A plasmid carrying the modified b-casein gene (pB-cas7-2272- hytk) was generated and transfected transiently into HEK 293 cells. In addition, pB-cas7-2272-hytk was also stably transfected into HEK 293 cells (Fig. 27b).

Both the transiently and stably transfected cells were subsequently infected with the virus G5 or the control virus. Genomic DNA was isolated 24h post infection and analysed by PCR. The primer combination bcas6/hytk2/PGK5 was used to assess recombination at the 5' end of the integrated cassette. A 936bp PCR product (amplified by the primer pair bcas6/hytk2) is detected in the unmodified b-casein gene, whereas a 215bp product (amplified by the primer pair bcas6/PGK5) is indicative of a b-casein gene modified by RMCE. The primer combination bcas3/hytk1/neoint.4 was used to detect recombination at the 3' end of the integrated cassette. A 1147bp PCR product (amplified by the primer pair bcas3/hytk1 ) is detected in the unmodified b-casein gene, whereas a 1028bp product (amplified by the primer pair bcas3/neoint.4) is indicative of a b-casein gene modified by RMCE. Recombination was readily

detected in HEK 293 cells transiently or stably transfected with the pB- cas7-2272-hytk plasmid indicating that the G5 virus is able to serve as both transgene donor and source of Cre expression mediating stable insertion of a transgene cassette into a predefined genomic site in HEK 293 cells.

The goal for genome modification in gene therapy and transgenesis is that genome alterations can be introduced rapidly and accurately. Cre recombinase has proved a useful tool for genome modifications due to its high activity in mammalian cells (Andreas et al. (2002) Nucleic Acids Res 30, 2299-2306) and its accuracy. The present inventors have shown that Cre mediated cassette exchange is highly dependent on electroporation conditions and vector design. The present inventors results confirm that a single contiguous segment of DNA which carries both, the Cre expression cassette and a mobilisable transgene cassette is able to mediate stable integration of the transgene into a genomic target. The presence of an insulator element assists efficient recombination.

An adenovirus vector which acts both as transgene donor and source of Cre expression is able to mediate transgene integration into extra- chromosomal and chromosomal target sites.

Table 1 Oligonucleotides used for quantitative real-time PCT

Table 2 Primer combinations used for PCR Genotyping

Table 3: Frequencies of homologous recombination and recombinase mediated cassette exchange (RCME) reactions

Table 4: Induction of the endogenous and modified β-casein genes in fusion cell clones

Table 5 Frequencies of homologous and site-specific recombination

Table 6: Primer combinations used for PCR analysis

Table 7: Recombinase mediated cassette exchange efficiency is dependent on electroporation conditions. Colony number are average values from 4 cell culture dishes (+/- standard deviation). Site-specific integration was assessed by PCR.

Table 8: Recombinase mediated cassette exchange efficiency is dependent on electroporation conditions

Table 9: Recombinase mediated cassette exchange efficiency after liposome transfection and electroporation. Where indicated, pMC1 -Cre was co-transfected with the plasmid pB2272-neo in a 3:1 ratio.