AU622870B2 - Vector for integration site independent gene expression in mammalian host cells - Google Patents

Description

WORLD INTELLECTUAL PROPERTY ORGANIZATION Pcr Internadial Bureau INTERNATIONAL APPLICATION PUBLISHED UNDER THE PATENT COOPERATION TREATY -PCT) Pl) International Patent Classification 4~ q C12N 15/00, 5/00 Al (11) International Publication Number: WO 89/ 01517 (43) International Publication Date: 23 February 1989 (23.02.89) (21) International Application Number PCT/GR.88/O0655 (22) International Filing Date: (31) Priority Application Number: (32) Priority Date: S August 1,988 (08.08.88) 8718779 7 Augiist 1987 (07.08.87) (33) Priority Country: GB (71)472) Applicant and Inventor: GROSVELD, Franklin, Gerardus [GB/09]; 70 Salcombe Gardens, Mill Hill, London NW 7 (GB).

(74) Agent: HALLYBONE, Huw, George; Carpmaels and Ransford, 43 Bloomsbury Square, London WCIA 2RA (GB).

(81) Designated States: AT (Eliropean patent), AU, 13 (European patent), CH (European patent), DE (European patent), FR. (European patent), GB (European patent), IT (iEuropean patent), JP, LU (European patent), NL (European patent), SE (European patent), US.

Published Withi internotionaI search report.

622870 ALLJ.P. 2 AP§ i~uf

AUSTRALIAN

9 MR 1989 PATENI OFFICE~ (54) Title: VECTOR FOR INTEGRATION SITE INDEPENDENT GENE EXPRESSION IN MAMMALIAN HOST

CELLS

Eco At 9AM li HindM 3.0 t0.2 0.55 6,3 4.3 7.0 1,15 S..Z 3.2 3.3 02 i5.0 2.0 14.6 2.2 4A4 1.1?,09? 2,2 4,9 16.0 ,2.4 l 0 0,46 017 Ecoltigil Barillwti comd~R (57) Abstract Vectors for the integration of a gene into the genetic mxaterial of a ma~mmalian host cell such that the gene may be expressed by the host cell comprise a promoter and the said gene aind includ,- a dominant activator sequence capable of eliciting host 'zell.-type restricted, integration site independint, copy number dependent expression of the said gene.

W 9/01517 PCT/GB88/00655 Vector for integration site independent gene expression in mammalian host cells.

Field of the Invention The present invention relates to recombinant DNA technology and.in particular to a vector useful for transfecting mammalian cells in vivo and in vitro to obtain expression of a desired structural gene. The invention relates also to the use of such sequences in gene therapy and hetarologous gene expression.

Background to the Invention There is a continuing need for improved expression vectors exhibiting high levels of expression. In particular expression vectors for use in mammalian cell lines are of increasing importance both for the industrial production of desired polypeptides and for the development of therapies for genetic disorders.

There are many known examples of characterised structural genes, which together with appropriate control sequence! may be inserted into suitable vectors and used t' transform host cells. A significant problem With the integration of such a structural gene and control regions into the genome of a mammalian cell is that expression has been shown to be highly dependent upon the position of the inserted sequence in the genome. This results in a wide variation in the expression level and only very rarely in a high expression level. The problem of integration site dependence is solved by thi present Invention and arises from the discovery of sequences referred to herein as "dominant activator sequences" (or "dominant control regions"(DCRs)) which have the property of conferring a cell-type restricted, integration site independent, copy number dependent expression characteristic on a linked gene system.

WO 89/01517 PCf/GB88/006 -2- The human alike globin genes are a cluster of five genes in the order comprising approximately 60kb of DNA on the short arm of chromosome 11. The different genes are expressed in a developmentally and tissue-specific manner, i.e. the embryonic E-gene is expressed in the yolk sac, the foetal S and A- genes, primarily in the foetal liver, and the adult and 3-genes primarily in bone marrow (for a review, see Maniatis et al., Ann.

Rev. Genet., (1981) 14, 145-178). Mutations in this gene family constitute the most widespread family of genetic diseases and a large number of these have been characteriTed, ranging from simple amino acid changes by point mutations to complete deletions of the locus (for a review, see Collins et al, Prog. Nucl. Acid.

Res. Mol. Biol., (1984), 31, 315-462). These often lead to severe clinical problems and early death.

Conventional treatment of these diseases (transfusions) are costly, risky and, in many cases, inadequate. The present invention provides a therapy for such disorders for example by gene therapy (Hock et al, Nature, (1986), 320, 275-277).

The DNA sequences which regulate the human e-globin gene are located both 5' and 3' to the translation nitiation site (Wright et al., Cell, (1984), 38, 265-273; Charnay et al., Cell, (1984), 38, 251-263).

Using murine erythroleukaemia cells (MEL) and K562 cells at least four separate regulatory elements required for the appropriate expression of the human ,-globin gene have been identified; a positively acting globin specific promoter element and a putatively negative regulatory promoter element and two downstream regulatory sequences (enhancers), one located in the gene and one approximately 800bp downstream from the I c* SWO 89/01517 PCT/GB88/00655 -3gene (Antoniou et al, EMBO (1988), 377-384).

A similar enhancer has also been identified downstream the chicken 3-globin gene using cultured chicken erythroid cells (Hesse et al., PNAS USA, (1986), 83, 4312-4316, Choi et al, Nature, (1986), 323, 731-734).

The downstream enhancer has been shown to be a developmental stage specific enhancer using transgenic mice (Kollias et al, Cell, (1986), 46, 89-94 and NAR (1987), 15, 5739-5747; Behringer et al., PNAS USA, (1987), 84, 7056-7060. All of these results therefore indicate that multiple development specific control regions of the 1-globin are located immediately inside and 3' to the 1-globin gene.

However, when a 3-globin gene containing all of these control regions is introduced in transgenic mice, the gene is not expressed at the same level as the mouse P-globin gene and exhibits integration site position effects. This is characterized by a highly variable expression of the transgene that is not correlated with the copy number of the injected gene in the mouse genome. The same phenomenon has been observed in almost all the genes that have been studied in transgenic mice (Palmiter et al, Ann. Rev. Senet., (1986), 465-499). Moreover, the level of expression of each injected gene in the case of 8-globin is, at best, an order of magnitude below that of the endogenous mouse gene (Magram et al., Nature, (1985), 315, 338-340; Townes et al., EMBO (1985), 4 1715-1723; Kollias et al, Cell, (1986), 46, 89-94). A similar problem is observed when the 3-globin or other genes are introduced into cultured cells by transfection or retroviral infection. This poses a big problem when considering gene therapy by gene addition in stem celt'. It is also a major problem for the WO 89/01517 PCT/GB88/609 r, -4expression of recombinant DNA products from tissue cells. Extensive screening for highly producing clones is necessary to identify cell-lines in which the vector is optimally expressed and selection for vector amplification is generally required to achieve e "'ression levels comparable to those of the naturally 'urring genes, such as for example 1-globin genes in Srythroleukaemic cell-lines.

The study of the 3-globin system has been assisted by analysis of haemoglobinpathies such as thalassaemias (van der Ploeg et al, Nature, (1980), 283, 637-642; Cu-tin et al, In: Haemoglobin Switching: Fifth Conference on Haemoglobin Switching, Washington Ed. G.

Stamatoyannopoulos, Alan R. Liss Inc., New York 1987).

The Dutch thalassaemia is heterozygous for a large deletion which removes 100kb upstream of the 3-globin gene, but leaves the 3-globin geres, including all of th" control regions described above intact (Kioussis et al Nature, (1983), 306, 662-666; Wright et al Cell, (1984), 38, 265-273; Taramelli et al, NAR, (1986), 14 7017-7029), Since the patient is heterozygous and transcribes the normal locus in the same nucleus, it indicates that some control mechanism may be overriding the functioning of the control sequences immediately surrounding the genes in the mutant locus. In the case of the Dutch -thalassaemia there are two possible explanations the observed effects; either a cis acting positive element has been removed or there has been insertion of a negative-acting element in an inactive chromatin configuration and behaves like a classical position effect (Kioussis et al, Nature, (1983), 0 6, 6(2-666).

WO 89/01517 PCT/GIB8/00655 In a chromosome, the genetic material is packaged into a DNA/protein complex called chromatin which has the effect of limiting the availability of DNA for functional purposes. It has been established that many gene systems (including the 3-globin system) possess so-called DNase hypersensitive sites. Such sites represent putative regulatory regions, where the normal chromatin structure is altered to allow interaction with trans-acting regulatory regions, Regions upstream from the epsilon-globin gene and downstream from the 3-globin gene which contain a number of "super" hypersensitive sites have been identified. These sites are more sensitive to DNase I digestion in nuclei than the sites found in and around the individual genes when they are expressed (Tuan et al PNAS USA, (1985), 32, 6384-6388; Groudine et al, PNAS USA, (1983), 80, 7551-7555, In addition, they are erythroid cel-1 specific and they are present when any one of the globin genes is expressed.

Tuan et al describe the broad mapping of four major DNase I hypersensitive sites in the 5' boundary area of the "3-like" globin gene. The authors note that certain sequence features of these sites are also found in many transcriptional enhancers and suggest that the sites might also possess enhancer functions and be recognised by erythroid specific cellular factors.

The present invention arose from the discovery that the complete 3-globin locus with intact 5' and 3' boundary regions does not exhibit an integration site position dependence. The regions of the locus responsible for this significant characteristic have been determined WO 89/01517 PCT/GB88/00655i r -6and shown to correspond to the DNase I super hypersensitive sites. These dominant activator regions are quite distinct from enhancers, exhibiting properties such as integration site independence not exhibited by the known enhancers. The dominant activator sequence used in conjunction with the known promoter/enhancer elements reconstitute full expression of the natural gene. The connection made as part of the present invention between DNase I super hypersensitive sites and dominant activator regions allows the invention to be extended to gene systems other than the 3-globin gene system. It has been shown, for example, that a human 3-globin gene can be functionally expressed in transgenic mice in an integratio. site independent manner.

Human T-lymphocytes (T-cells) are produced in bone marrow and mature in the thymus where they develop their immunological characteristic of responding to foreign antigens in the body. T-cells are produced in a highly tissue specific manner. It is recognised that T-cells carry specific cellular markers (Bernard et al "Leukocyte Typing", Ed. Bernard et al, p41, Springer Verlag, Berlin and New York, 1984). One such marker is the E-rosette receptor, -nazw, by the designation CD2.

The structure of the CD2 marker and the CD2 genes has been the subject of considerable study (Brown et al, Meth. in Enzym., (1987), 150, 536-547 and Lang, G. et al, EMBO J, (1988), 1675-1682), In the latter paper, published after the priority date of the present invention, the property of a large fragment including the CD2 gene to exhibit copy number dependence is noted and a suggestion is made that such a fragment contains a locus forming sequence analogous r i i Fx_ 1 wd89/01517 PCT/GB88/00655 -7to those found in 3-globin. DNase super hypersensitive sites have now been identified in the fragment and shown to be dominant activator sequences of the present invention.

Finally, the present invention is applicable to the production of transgenic animals and the techniques for producing such are now widely known. For a review, see Jaenisch, Science, (1988), 240, 1468-1474.

The present invention provides a solution to the problem of integration site dependence of expression making possible the insertion of functionally active gene systems into mammalian genomes both in vitro and in vivo. Specific sequences providing advantagous increases in transcription levels have also been identified.

Summary of the Invention According to the present invention there is provided a vector for the integration of a gene into the genetic material of a mammalian host cell such that the gene may be expressed by the host cell, the vector comprising a promoter and the said gene characterised in that the vector includes a dominant activator sequence capable of eliciting host cell-type restricted, integration site independent, copy number dependent expression of the said gene.

As used herein the term "dominant activator sequence" means a sequence of DNA capable of conferring upon a linked gene expression system the property of host cell-type restricted, integration site independent, copy number dependent expression when integrated into the genome of a host cell compatible with the dominant WO 89/01517 PCT/GB88/0065 fr -8activator sequence. Such a dominant activator sequence retains this property when fully reconstituted within the chromosome of the host cell. The ability to direct efficient host cell-type restricted expression is retained even when fully reformed in a heterologous background such as a different part of the homologous chromosome (eg chromosome 11 for 3-globin) or indeed a different chromosome altogether.

It is hypothesised that the dominant activator sequences of the invention may open the chromatin structure of the DNA, making it more accessible and thus may act as a locus organiser.

The dominant activator sequence may be a single contiguous sequence corresponding to, or derived from a naturally occurring gene system or may consist of two or more such sequences linked together with or without intervening polynuclectides.

The dominant activator sequence may be derived by recombinant DNA techniques from a naturally occurring gene system or may correspond to a naturally ocurring gene system in the sense of being manufactured using known techniques of polynucleotide synthesis from sequence data relating to a naturally occurring gene system. Alterations of the sequence may be made which do not alter the function of the dominant activator sequence.

Preferably, the naturally occurring gene system from which the dominant activator sequence is derived or with which it corresponds is a system which exhibits a highly host cell-type restricted expression SWd 89/01517 PCT/GB88/00655 -9characteristic preferably at a high level. Specific examples of such systems are the haemoglobin systems such as 8-globin system and lymphocyte systems such as the CD2 system.

The dominant activator sequence may consist of, be derived from, or correspond to one or more DNase I super hypersensitive site, preferably of any gene system capable of cell specific expression. Other sequences might however exhibit the functional characteristics of a dominant activator sequence.

Where the naturally occurring dominant activator sequence comprises two or more subsequences separated by an intervening polynucleotide sequence or sequences the dominant activator sequence may comprise two or more of the subsequences linked in the absence of all or a part of one or more of the intervening sequences.

Thus, if the dominant activator sequence of a naturally occurring gene locus comprises two or more discrete subsequences separated by intervening non functional sequences, (for example, two or more super hypersensitive sites) the vector of the invention may comprise a dominant activator sequence comprising two or more of the subsequences linked together with all or part of the intervening sequences removed.

The term "vector" as used herein connotes in its broadest sense any recombinant DNA material capable of transferring DNA from one cell to another.

The vector may be a single piece of DNA in linear or circular form and may, in addition to the essential functional elements of the invention, include such WO 89/01517 PCT/GB8/006 K other sequences as are necesssary for particular applications. For example, the vector may contain additional features such as a selectable marker gene or genes, and/or features which assist translation or other aspects of the production of a cloned product.

The vector suitable for integration consists of an isolated DNA sequence comprising a dominant activator sequence and an independent structural gene expression system.

The DNA sequence is not linked at either end to other substantial DNA sequences. The isolated DNA sequence may however be provided with linkers for ligation into a vector for replication purposes or may be provided with sequences at one or both ends to assist integration into a genome.

The vector defines a "maini locus" which can be integrated into a mammalian host cell; where it is capable of reconstructing itself in a host cell-type restricted, integration site independent, copy number dependent manner.

The invention also provides a transfer vector, suitably in the form of a plasmid, comprising a dominant activator sequence, Such vectors are useful in the construction of vector for integration.

The term "gene" as used herein connotes a DNA sequence, preferably a structural gene encoding a polypeptide The polypeptide may be a commercially useful polypeptide, such as a pharmaceutical and may be Mffff WO, 89/01517 r PCT/GB88/00655 -11entirely heterologous to the host cell. Alternatively the gene may encode a polypeptide which is deficient absent or mutated in the host cell.

The mammalian host cell may be any mammalian host cell susceptible to uptake of the vector of the invention.

The vector DNA may be transferred to the mammalian host cell be transfection, infection, microinjection, cell fusion, or protoplast fusion.

The host cell may be a cell of a living human or animal. In particular, the host cell may be a cell of a transgenic animal such as a mouse. The host cell may be a human stem cell such as bone marrow cell, The host cell may be derived from tissue in which tes dominant activator sequence is functional, such as erythroid cells in the case of the 2-globin locus.

The promoter may be any promoter capable of functioning in the hopt cell and may be for example a mammalian or viral promoter. Optionally, the promoter may be homologoqo with the gene locus of the dominant activato sequence (for example the 3-globin promoter) and may be present in tandem with another promoter (for example the 3-globin promoter and a viral tk or promoter and may include one or more enhancer elements.

In one embodiment of the invention to be described below, the dominant activator sequence is derived from the -globin gene locus, As discussed above, the 3-globin locus contains a number of DNase I super hypersensitive sites which constitute the dominant activator regions (see Tuan et al loc cit), Preferably W) 89/01517 PCT/GB88/00655 -12the dominant activator sequence contains one or more of the DNase I super hypersensitive sites identified within the 3-globin locus. Preferably thes?. from the 5' boundary of the locus, optionally with tne 3' boundary sequences. The dominant activator sequence is within a fragment of 21kb from -lkb ClaI to -22kb B91II immediately upstream of the epsilon-globin gene in the 3-globin locus. This region contains four DNase I super hypersensitive sites with intervening polynucleotides (five distinct sites of which two are very close together), Preferably some or all the intervening nucleotides are removed using known techniques such as digestion with exonuclease, A reduced form of the B-globin locus dominant activator sequence has been produced which exhibits a significantly increased level of expression of a linked gene expression system. This was produced by ligating the following four fragments 2.1kb XbaI-XbaI 1.9kb HindIII-HindIII 1. 5kb KpRnI-BqlII l.lkb partial Sacl fragment This dominant activator sequence, referred to hereid as a "micro locus" is a 6.5kb fragment which may be used as a "cassette" to activate a specific gene expression system.

In a second embodiment of the invention to be described below, the dominant activator sequence is derived frc CD2 gene locus. The CD2 gene locus contains three super hypersensitive sites, one at the 5' boundary of WO 89/01517 PCT/GB88100655 r -13the locus and two at the 3' boundary of the locus.

Preferably the dominant activator contains one or more of the DNase I super hypersensitive sites within the CD2 locus. Most preferably, the dominant activator sequence contains both the super hypersensitive sites from the 3' boundary of the locus, optionally with all or a part of any intervening sequence deleted The dominant activator sequence is contained within a BamHI to .aI fragment 3' to the CD2 gene.

In a further aspect of the invention there is provided a method of producing a polypeptide comprising culturing a host cell transformed with a vector of the invention.

The method may be applied in vitro to produce a desired polypeptide. In addition, the method may be applied in vivo to produce a polypeptide having no therapeutic value to the animal, Such a method of producing a polypeptide is not to be considered as a method of treating the human or animal body.

In a further aspect of the invention the vector may be used in a method of treatment of the human or animal body by replacing or supplementing a defective mammalian gene.

Many diseases of the human or animal body result from deficiencies in the production of certain gene products. The characterising features of the vectors of this invention make them amply suited to the treatment of deficiencies by gene therapy in vivo.

A method of gene therapy is provided comprising i WO 89/01517 PCT/GB88/00655, removing stem cells from the body of a human or an animal, killing stem cells remaining in the body, transforming the removed stem cells with a vector of the invention containing a gene deficient, or absent, in the human or animal body, and replacing the transformed stem cells in the human or animal body, This method can be used to replace or supplement a gene deficient in a human or animal. For example, anindividual suffering from a haemoglobinopathy such as 1-thalassemia could be treated to insert an active highly expressed 3-globin gene locus to make up the deficient E-haemoglobin. Alternatively a deficiency in the immune system could be treated with a CD2 locus vector. This method could be used to treat deficiencies in adenosine deaminase (ADA syndrome) or severe combined immune deficiency syndrome (SCID syndrome).

Bone marrow is a suitable source of stem cells and is advantageous in that it contains the precursors of both lymphocytes and erythroid cells. Alternatively other tissue could be removed from the body, transfected or otherwise provided with a vector of the invention and implanted back into the body.

The invention is now described by way of example with reference to the accompanying drawings.

Brief Dcription of the Drawings Fig 1 hows the mapping of DNase I super hypersensitive sites 5' to the epsilon globin gene. Panels A to C are representations of nitrocellulose filters illustrating the mapping experiment and Panel D control regions"(DCRs)) which have the property of conferring a cell-type restricted, integration site independent, copy number dependent expression characteristic on a linked gene system.

W( 89/01517 PCT/GB88/00655 shows a restriction map of the a-globin locus upstream of the epsilon globin gene for the three restriction enzymes used in these experiments. Probe l.ocations and super hypersensitive sites are marked, Fig 2 shows the corresponding experiment to that in Fig. 1 conducted to map the super hypersensitive sites 3' of the -globin gone.

Fig 3 shows the construction J a human 13-globin "mini" locus Fig 4 shows a structural analysis of the human 3-globin transgene (Fig 4 comprises 4A, 4B, SC).

Fig 5 shows an expression analysis of the human 3-globin transgene in mouse (Fig 5 comprises 5C, Fig 6 shows the production of human 3-globin DNA in various mouse tissues (ys yolk sac, iv Foetal liver and br brain).

Fig 7 shows the refor.nation of DNase I hypersensitive sites on the transgenic mini 3-globin locus.

"ig 8 shows the construction of plasmids GSE1364, 1365 and 1366.

Fig.9 shows the construction of plasmids GSE1359, 1400, 1401 and 1,357.

shows human 3-globin expression, WO 89/01517 PCT/GB8/0065 16- Fig.ll shows the mapping of CD2 hypersensitive sites.

Fig.12 shows a Thy-i CD2 construct.

Fig.13 shows and brain of Fig.14 shows construct.

the distribution of Thy-i in the thymus transgenic mice.

the construction of a 3-globin CD2 Fig. 15 shows the genomic DNA of transgenic mice carrying the construct of Fig. 14.

Example 1 Experiments were conducted to establish whether dominant control regions are present ie the flanking regions of the 3-globin locus by including such on regions on a 3-globin construct. If such sequences were insensitive to integration site position effects, it would be expected: that th., level of expression of the construct is directly related to its copy number in the transgenic mice, that the level of expression per gene copy is the same in each mouse and possibly that the introduced gene is expressed at a level comparable to that of the mouse genes. Prior to testing the 5' and 3' sequences, it i i SWP 89/01517 PCTIGB88/00655 -17was important carefully to map the positions of the hypersensitive sites since their exact position, especially at the 3' side, has not been published in detail (Tuan et al., 1985).

Mapping DNase hypersensitive sites flanking the qlobin locus Very limited DNaseI digestion of nuclei of erythoid cells reveals sites extremely sensitive to DNase upstream of the epsilon-globin gene and downstream of the 8-globin gele (Tuan et al., 1985). We fine mapped the DNaseI hypersensitive sites relative to restriction sites mapped in our cosmids (Taramelli et al., 1986).

Fig. 1 shows a time course of DNaseI digestion on nuclei of HEL (Martin and Papayannopoulou, 1982) and PUTKQ (Klein et al., 1980) cells recut with various restriction enzymes, Southern blotted and hybridized with a number of probes upstream of the epsilon-globin gene (Fig Id). Four hypersensitive sites are present at 18kb (No. 15kb (No. 12.0 and 11.5kb (No.

3a,b) and 5.4kb (No. 4) upstream of the epsilon-globin gene. Nos. 1, 3a and 4 sites are strong hypersensitive sites as judged by their early appearance during the time course of digestion. Nos. 2 and 3b sites are weaker sites and appear later during digestion. In Fig. 1, the experimental details were as follows: Panel A Nuclei of HEL cells were incubated with DNase pg/ml) at 370C and aliquots removed from the reaction at the times indicated (in minutes) above each track.

c I WO 89/01517 PCT/GB88/00654 -18- Aliquots of approximately 107 HEL nuclei were incubated in the absence of DNaseI (0 enz.) or presence of 30 units of Alul or 60 units of HinfI for 15 min. at 37°C. DNA was purified after proteinase K digestion, recut with Asp71S, fractioned on a 0.6% agarose gel transferred to nitrocellulose and probed with an epsilon-globin gene probe (Bam-Eco 1.4kb).

Panel B The DNaseI digested samples of Panel A were recut with BglII and probed with a 3.3kb EcoRI fragment which in addition to 12.0 and 5.8kb bands also detects a cross hybridizing band of 6.8kb after washing filters at a stringency of 0.3xSSC 650C, which is not observed on washing at 0.1xSSC 650C (not shown).

Panel C Nuclei of PUTKO cells were digested with DNaseI at 37 0 C and samples removed from the reaction after 1,2,4 and 8 minutes. An aliquot of approximately 107 nuclei was incubated for 8 minutes at 370C in the absence of DNaseI (lane 0 enz.).

After recutting with BamHI gel electrophoresis and transfer to nitrocellulose, the filter was hybridized with a 0.46kb EcoRI-BglII probe and washed to a final stringency of 0.3xSSC at 650C.

Panel D Shows a restriction map of che 3-globin locus upstream of the epsilon-globin gene for the three restriction enzymes used in these experiments. Locations of probes and deduced positions of DNaseI hypersensitive sites in HEL and PUTKO chromatin are marked.

SWO 89/01517 PCT/GB88/00655 -19- The exact location of the DNaseI hypersensitive site 3' of the adult 3-globin gene was determined using two single copy DNA probes and several restrictions enzyme digests of DNaseI digested HEL nuclei. The data summarised in Fig. 2 show that there is a single DNaseI hypersensitive site between the 2.3kb BglII fragment and the 2.4kb HindIII fragment approximately 3' of the adult i-globin gene (Fig. 2E). (Similar results were obtained with DNaseI digested PUTKO nuclei (data not shown)) In Fig 2 the experimental details were as follows: Panels A co D show DNA from DNaseI digested nuclei of HEL cells recut with the indicated restriction enzymes and probed with either a 1.15kb EcoRI or a l.lkb HindIII-Bam fragment. Both DNA probes contain some human repetitive DNA which necessitated the use of sheared human DNA per ml in prehybridization and hybridization buffers. Filters were washed to a final stringency of 0.3xSSC at 650C.

Panel E shows a restriction map of the 8-globin locus 3' of the 3-globin gene. The position of the two probes used and the 3' HSS site are marked. The position of HSS 5 was confirmed by probing HindIII and BglII recut DNA samples with the 1.15 EcoRI probe (data not shown).

Construction of the 3-globin "mini" locus The globin "mini" locus (Fig. 3) was constructed from three regions (see "Materials and Methods" below for the precise description); a 21kb region immediately upstream of the epsilon-globin gene (-1kb Clal to -22kb I L: :1 I-- WO 89/01517 PCT/GB88/00655 BglII) containing all four hypersensitive sites (Fig.

a 4.7kb (BglII) fragment containing the 1-globin gene and all the known regulatory sequences immediately in and around the gene and a 12kb regioi downstream from the F-globin gene (+12kb KpnI to +24kb BamHI) containing the downstream hypersensitive site (Fig 2).

These regions contained or were provided with unique linkers to create a series of unique restriction sites (SalI, XhoI, Clal, KmnI, PvuI) and cloned into the cosmid pTCF (Grosveld et al., 1982). The entire 38kb insert was excised from the cosmid as a Sail fragment, purified from low melting agarose and injected into fertilized mouse eggs (Kollias et al., 1986).

Referring to Figure 3, the human 8-globin locus is illustrated on top and the sizes are shown in kilobases The second line shows the -globin "mini" locus, with a number of unique restriction sites; the connecting lines indicate the original position of the fragments in the 8-globin locus. The restriction enzyme fragments in the "mini" locus for EcoRI, BamHI and HindIII are shown in kb. The hybridization probes for the 5' flanking region (EcoBgI 0.46). the 3-globin gene (BamEco 0.7 IVS II) and the 3' flanking region (Eco 1.15) are shown at the bottom.

Transcenic mice containing the M-qloin "mini" locus Initial injection of mouse eggs was carried out with a DNA concentration of 2gg/ml, the eggs were transplanted back into the mice, but no offspring were obtained (for unknown reasons). Several series of. injections were then carried out with a DNA concentration of 0.5-1 pg/ml and foetuses were collected after either 12.5 or 13.5 days of gestation. DNA was prepared from SWO 89/01517 ,1 PCTGB88/00655 -21individual placentas, Southern blotted and hybridized with a human 8-globin probe to determine the number of transgenic mice (as described by Kollias et al., 1986). The transgenic DNAS were subsequently digested with EcoRI, Southern blotted and hybridized with a human 3-globin probe and a mouse Thy-i probe (single copy) to determine the number of genes integrated in each mouse (Fig.4A and Table I see page These values varied from a single insert (mouse 33 and 38) up to a hundred copies per cell (mouse 21). It was then determined whether the inserted DNA was integrated intact or whether deletions and/or rearrangements had taken place after injection. Total DNA was digested with BamHI or HindIII, Southern blotted and probed for the presence of the complete construct. In particular a 0.46kb BglII probe (Fig. 3) was used to detect the BamHI fragment at the 5' end of the construct (Fig. 4B) and a 1.15kb EcoRI probe (Fig. 3) to detect 'cne 14.6kb BamHI or 18kb HindIII fragment at the 3' end of the construct (Fig. 4C). All but one of the transgenic mice appeared to have intact "mini" R-globin loci which are integrated in tandem arrays (not shown) as judged by the presence of the correctly sized fragments. It should be noted however that the BamHI and HindIII sites at the 3' end of the locus (Figs. 4 and 6) are very resistant to cleavage, even though the mixed in control DNA has been fully digested (see Methods). This results in partial cleavage bands which are larger than fully digested BamHI (14.6kb) and HindIII (18kb) fragments. Transgenic mouse 38 is the only mouse that has not retained a full copy of the 13-globin "mini" locus. The single copy present in this mouse has lost the BamHI and HindIII sites upstream of the 8-globin gene, although the EcoRI site is still present. This indicates that all the 5' flanking DNA including the whole hypersensitive site region has been deleted in this mouse.

WO 89/01517 PCT/GB88/0065 -22- In Figure 4, the experimental details were as follows: Panel A Approximately 5pg of placental or yolk sac DNA was digested with EcoRI, Southern blotted after gel electrophoresis and hybridized to either the human -globin Bam-Eco IVS II probe or a mouse Thy-i probe.

The gels were quantitated by densitometer scanning of different exposures. The DNA of mouse 17 was partially degraded, therefore, a lower M.W. Thy-1 cross hybridizing band was used to quantitate this sample.

The inset (a high exposure of the globin hybridization) shows the low copy signals.

Panel B Placental or yolk sac DNA was digested with BamHI, Southern blotted and hybridized to a 5' flanking probe (Eco-Bgl 0.46, Fig. 3).

Panel C Placental or yolk sac DNA was digested with BamHI or HindIII, Southern blotted and hybridized to a 3' flanking probe (Eco 1.15, Fig, 3).

In all panels a mixture of marker DNA was used from a DNA digested with BstEII or HindIII. The transgenic mice are indicated at the top, empty lanes contain DNA samples from non-transgenic mice. Sizes are in kilobases Some of the DNA samples are partially digested, although the mixed in DNA control was i WO 89/01517 PCr/GB88/006S5 -23completely digested (not shown). This result in bands which are larger than the 14.6kb BamHI (14.6kb 3.2kb 3.3kb) or the 18kb HindIII fragments (18kb 2.4kb 3.4kb) by cleavage at the next BamHI or HindIII site in the tandem arrays.

The human 3-clobin "mini" locus is fully active To measure the activity of the human 3-globin gene, RNA was prepared from all the transgenic and non-transgenic mice (liver, blood and brain) and analyzed by Sl nuclease protection analysis. Fig. 5A shows the analysis of liver RNA with a 5' human ,-globin, while Fig, 5B shows an analysis of the 3' end of the human 8-globin RNA in the presence of a 3' mouse 3-globin and mouse a-globin probes as internal controls. These results show that the human 8-globin gene is correctly initiated and polyadenylated and that the levels of RNA are proportional to the number of copies of the "mini" locus in each mouse, (except 24 and 38, see below).

Moreover, no expression is found in non erythroid tissue (not shown). The level of expression is clearly very high when the ratie of the human 3- to mouse 3-globin RNA levels (Table I see page is compared to that of the Hull ells control (Zavodny et al., 1983).

The latter cell line is a mouse erythroleukaemia (MEL) cell line containing the human R-globin locus on a human X-1l hybrid chromosome, To carefully quantitate the levels of expression in these mice, RNA levels in mice 33 and 12 which have one and two intact copies of the 3-globin locus were analysed, and the two exceptions mice 24 and 38 with mouse and human 1-globin probes of equal specific activity. Fig. 5C shows that the signal of the human 8-globin gene is approximately half that of the mouse 3-globin diploid in mouse 33 and

I

WO 89/01517 PCT/GB88/00655i -24equal in mouse 12. This is confirmed at the protein level when the red cells of mouse 12 and an age matched control are lysed and the proteins run on a Triton-acid urea gel (Alter et al., 1980). It is clear that the transgenic mouse 12 expresses equal levels of mouse and human 1-globin proteins (Fig. 5D). From all these data it was concluded that each human -globin gene in these mice is fully active, i,e. at a level equivalent to that of the mouse ,-globin genes, while the expression in mice 24 and 38 is suppressed. A peculiar situation is found in mice 21 and 17, The mouse 1-globin RNA levels are very low in these mice, presumably as a result of the very high copy number of the huian 8-globin gene and, as a consequence, very high mRNA (and protein) levels. The high copy number possibly results in a competition for transcription factors (mouse 1-globin RNA is decreased more than mouse a-globi4 RNA, Fig, 5B), while the high levels of 3-globin protein definitely result in an anaemia (a-thalassaemia). Mice 21 and 17 had small colourless livers consistent with the destruction of mature red cells and a concomitant loss of mRNA.

Two mice (38 and 24) form an exception because they express relatively lower amounts of human (3-globin (Fig. 5A,B,C, Table 1 see page The situation in mouse 38 is explained by the fact that this is the only transgenic mouse which has a deletion of the 5' end of the locus (Fig,4). As a consequence, the human 3-globin gene is still expressed, but in a position dependent manner and at low levels, just as found when small 3-globin fragments are introduced in mice (Kollias t al., 1986; Magram et al,, 1985; Townes s._ al 1985).

WO 89/01517 PCT/GB88/00655

I

Mouse 24 has no apparent defect in the locus and the possibility was investigated that this mouse is a true exception or whether it is either a chimaeric mouse or has the 8-globin locus integrated on an inactive X-chromosome. DNA was therefore isolated from the yolk sac, brain and foetal liver (the "erythroid" tissue) and hybridized the EcoRI digested DNAs on Southern blots with a human 3-globin probe and a murine Thy-1 probe to detect the 8-globin transgene versus an endogenous single copy gene, Fig. 6 clearly shows that lOpg of foetal liver DNA contains even less than half of the globin signal that 5pg of yolk sac DNA, i.e.

much less than 25% of the liver cells contain the 3-globin gene. This therefore a chimaeric mouse, in which the 0-globin mini locus is under-represented in the erythroid tissue (foetal liver) which explains the lower levels of human 3-globin mRNA found in these cells, It may be concluded from this that provided the locus has not been affected, there a complete correlation between expression levels and DNA copy numbers and that vach gene is fully expressed in every transgenic mouse, In Figure 5, the experimental details were as follows: Panel A: of NA from 12.5 or 13.5 day livers from different mice (numbers on top) was analyzed by Sl nuclease protection as described previously (Kollias et al., 1986), The protected fragment of the 5' human 3-globin probe (525bp AccI from an IVS I lacking R-globin gene) is 160nt glo), Input probe is shown as ip, The marker is pBR322 DNA digested with Hinfl and the sizes are indicated in nucleotides.

C

WO 89/01517 PCT/GB88/00655 -26- Panel B of RNA was analyzed as in panel A, but using two probes; a 3' probe of the human 3-globin gene (760bp EcoRI-MspI) resulting in a 210nt protected fragment (H beta) and a mouse a-globin probe (260bp BamHI second exon probe), resulting in a 170nt protected fragment (M alpha). The positive control is RNA from the hybrid MEL cell line Hull. Marker is pBR322 x Hinf.

Panel C of RNA was analyzed as in panel B, but in addition, a mouse 0-maj-globin pobe (a 730bp second exon probe) was added resulting in a 100nt protected fragment (Mbeta). The probes were of eqal specific activity, The activity of the probes was assessed by quantitative agarose, gel electrophoresis and autoradiography of the individual probes before and after mixing for the S! niclease analysis (not shown), The positive control is Huiia RNA Each probe is individually tested with RNA from Hull (hu) or non transgenic mouse foetal liver RIA (ml).

panel D: Samples of foetal liver (approximately 20pg) or human red cells were lyzed in 404i of H20 and the globn proteins separated as described (Alter et al., 1980), The gel was stained to visualize the globin protein chains. Mouse 9 is an age matched non-transgenic (litter) mate of mouse 12r The positions of mouse a-, 0- and human P-globin are indioated.

WO 89,'01517 PCT/GB88/00655 -27~- Mapin~D~aeI ypesenitive sites in the transcfenic O-glbin locus The DNaseI hyper-sensitive sites flanking the 3-globin loctig are found in all erythroid tissues stud. -d regardless of which a3-like globin is ex~pressed. To 4,etez.,mine i41 the DNaseI hypersensitive sites are reformed in the mini 3-~locus, of transgenic mice eXpreSsing h~igh leveiv of human, 1-globin mRNA, a DNasel fadeout was performed on a small sample of the liver and brain of mouse 21 (which hi~s the highest copy number of the a3 gene min. locus). Limited DIaseI digestion was carried out. in the presence of carrier foetal Itver or brain nuclei, purified nuclear DNA was recut with BamHI, 'Eractioned, Southern blotted and fragments zpanning the whole construct were used as hybridization probes. The 1.15 EcoRI probe which was used to detect the. 3' hypersensitive site in UEL and PUTKO cells detects a 17.8kb Baml fragment which is duo to partial digestion of the BamHI sitA at the 3' end of the injected fragment, The 17.8kb BamHI fragment is completely insensitive to limitied DNasei digestion in liv, er nuclei of mouse 21 (Fig. 7E). The same filter rep,,okbed with a GO0bp AccI 5' human 1-globin probe shows that the 2.Qkb BamHj fragment encompassing the 5V end (to -1460bp) of the highly transcribed human 13-globin gone is slightly sensitive to DNaseI digeastiont revealing a DNaseI sub-band oDf approximately 1kb (Figo 7D), consistent with a DNasel hypersensitive- site within the 3-globin gene promoter (Groudine et al., 1983). Reprobing the same fil1ter with the 0.46 EcoRI GblII probe (Fig. 1,C) shows that the 2.5kb BamHjI fragment encompassing hypersensitive sites 2, 3a., and 4 is exceedingly senritive to DNasel digestion Inx foetal liver nuclel,! and sub-bands corresponding to cutting at sites 3 and 4 are WO 89/01517 PCT/GB88/0065, -28clearly visible. The same 15kb BamHI fragment of ,ne transgene mini 1 locus is insensitive to DNaseI digestion in brain nuclei of the same animal (not shown). The 3.3kb BamHI fragment (containing hypersensitive site no. 1) of the transgenic mini locus is also very sensitive to DNaseI digestion in liver, but not brain nuclei of mouse 21 (Fig. 7A,B).

In summary, the 5' flanking region is extremely sensitive to DNaseI digestion and has retained the superhypersensitive sites in a tissue specific fashion. Surprisingly, the 3' flanking region is insensitive to DNase digestion and the super hypersensitive site is not retained. The normal hypersensitivity in the promoter is present (Groudine et al., 1983).

In Figure 7, the experimental details were as follows: Nuclei of mouse 21 tissues were incubated with DNase per ml for 0,1,2,4 and 8 minutes at 370C (liver nuclei), or 0,1,2,5 and 10 minutes (brain nuclei).

Aliquots of nuclei were incubated for 8 or 10 minutes in the absence of DNase (lanes 0 enz.). Purified DNA was recut with BamHI fractioned on a 0.6% agarose gel, transferred to nitrocelluldse and probed with DNA fragments spanning the transgenic human B-globin mini locus (Panels A-E).

The 14.0kb SamHI fragment detected by probe D in panel E is a junction fragment of the tandem array of the mini 3-globin locus. The 17.8kb band is the result of incomplete BamHI digestion.

Panel F shows a schematic representation of a repeat .1 89/01517 W9 8901517PCT/GB88/06-455 -29unit of the mini locus tandem array n mouse 21 on which DNase hypersensitive sites are marked. Site 5 is not reformed in foetal liver nuclei (Panel DNaseI sub bands corresponding to HSS 1 and 2 are not seen in DNaseI fadeout experiments, but the BamHI fragments in which they reside are sensitive to DNaseI digestion in liver (Panel~s A and C but not brain (Panel B and data not shown).

TABLE 1 Levels of human a-globin RNA and DNA' Mouse No 9 /a RAZ .D~TA 2

R.NA-/DNA

12 1. 0 17 42 .0 4

N.D.

4 0.3 10.2 10.1 1 7,9 I1 0 47 .2 >100 chimaera 5 11.8 1, 0 0.9 N. D.

N. D.

0.85 1.15 <0.25 1. 1 8.7 0.4 7.2 Hull 0.3 0.506 0.6 WO 89/01517 PCT/GB88/00655L lAll values were normalized to mouse 12 as 1.0 (eg.

see Figs. 4D and 5C) after densitometer scanning of autoradiographs exposed for different times of several experiments.

2 The ratios were obtained by dividing the human 8-globin RNA and DNA signals by those obtained for the mouse a-jlobin RNA and Thy-1 DNA, respectively. Mouse 34 RNA was measured in only one experiment and not included.

3 The DNA of mouse 17 was partially degraded (see Fig.

We therefore used a low M.W. cross hybridizing band on the Thy-1 blot to quantitate the copy number (rather than the degraded 8kb EcoRI Thy-1 band).

4 The ratio of human 3-globin to mouse globin RNA was not determined accurately in this mouse, due to the very heavy high human to mouse RNA signals (see text).

Mouse 24 is chimaeric for the human 3-globin gene (Fig. 6) and could therefore not be quantitated.

MATERIALS AND METHODS Transqenic Mice The 38kb Sall fragment was purified from cosmid vector sequences by gel electrophoresis and prepared for injection as described (Kollias et al., 1986; Brinster et al., 1985). Transgenic foetuses were identified by Southern blot analysis (Southern, 1975) of placental

DNA.

SWO 89/01517 PCT/GB88/00655 -31- RNA analysis RNA extraction of different tissues and S1 nuclease protection analysis were carried out as described previously (Kollias et al., 1986).

DNaseI sensitivity DNaseI sensitivity assays were carried out on isolated nuclei as described by Enver et al. (1984), except that ImM EGTA was included in all buffers until DNaseI digestion in buffer A plus ImM CaCl2 was performed.

In the case of transgenic liver and brain, four livers and brains from age matched normal foetuses were added as carriers to provide sufficient material to carry out the analysis.

Construction of the -qclobin mini locus A 12kb BglII fragment (10-22kb 5' of 3-globin) was cloned into the BamHI site of cosmid pTCF (Grosveld et al., 1982) without a Clal site cosHG B12.

cosHG B12 was cleaved with Asp718 (13.5kb 5' of epsilon-globin) and a unique XhoI linker was inserted cos HG B12X.

cosHG 14.2 (Taramelli et al., 1986) containing the entire epsilon-globin upstream region was also cleaved with DpnI and XhoI linker inserted and packaged cos HG 14.2X.

A Clal linker was inserted in the Xbal site of pUC18 ->pUC18C.

unu cUUea positions of DNaseI hypersensitive sites in HEL and PUTKO chromatin are marked.

WO 89/01517 PCT/GB88/00659, -32- The 4.7kb BgIII fragment containing the 3-globin gene was inserted in the BamHI site of pUCl8C pUCI8C

-BG.

A 12kb HpaI-BamHI fragment (12-24kb 3' of 1-globin) was cloned into HpaI-BamHI of cosmid pTCF lacking a Cla site cos HG HB12.

The final cosmid construct was put together as a four fragment ligation and packaging, a 16.5kb PvuI-XhoI fragment from cos HG bl2X containing the cosmid and part of the upstream sequences, an llkb XhoI-ClaI fragment from cos HG 14.2X containing the reinainder of the upstream sequences, a 4.7kb ClaI-KpnI fragment from pUC18C-BG containing the 3-globin gene and a 19kb KpnI-HpaI fragment from cos HG HBl2 containing the 3' flanking region and the second cosmid for packaging purposes cos HG "mini" locus (Fig. 3).

Discussion of Example 1 Position effect independent expression The introduction of the human 1-globin gene flanked by the upstream and downstream regions in transgenic mice has resulted in an expression pattern that is position independent and directly related to copy number, with levels of human 1-globin mRNA as high mouse 13-globin mRNA. This is different from that observed with previous transgenic mice using the 3-globin gene alone (Magram et al., Chada eatl., 1985; Townes et al., 1985; Kollias et al., 1986, 1987), or other genes such as the immunoglobulin genes (Grosschedl et al., 1984), albumin genes (Pinkert et al., 1987) Thy-1 gene (Kollias et al., 1987) AFP gene (Hammer et al., 1987) 1 I: I I~ .i W 89/01517 PCT/GB88/00655 -33and many others. As a result of position effects in all transgenic mice experiments to date, the level of transgenic expression has been suboptimal and has varied between different mice, i.e. there has been no strict correlation between the copy number of the transgene and the level of expression. Therefore, it has been very difficult to quantitate any experiments accurately and, as a result, it has been impossible, for example, to examine the co-operation between regulatory elements in any gene. The results described in this example indicate that this problem can be overcome for the 8-globin gene. It is clear that the expression levels found in mice containing the intact mini locus is directly proportional to the number of genes that have been incorporated (Table In addition, the expression level per human 3-globin gene in all of those mice is very similar to that observed for the endogenous mouse 3-globin gene (Fig It is therefore concluded that the 3-globin locus flanking sequences confer position independent expression on the 1-globin gene which itself is erythroid specific. In addition it is concluded that all of the regulatory sequences involved in 3-globin gene expressions are included in this construct.

Gene therapy Perhaps the most interesting application for the properties of the dominant control region(s) is that they might allow completely regulated expression of the 0-globin (and possibly other genes) in retroviral vector or transfection systems. Gene expression in such systems is presently position dependent and inefficient. In particular, retroviral vector systems which are presently the only efficient way to transfer x yWUaiLon. DNA was prepared from WO 89/01517 PCT/GB88/00655 -34genes into hematopoietic (or other) stem cells, are very sensitive to the integration site of the retroviral vector (Cone et al., 1987). Inclusion of the globin dominant control regions may solve this problem and allow the efficient transfer of a fully active, single copy human globin gene to hematopoietic stem cells. This would form the basis for gene therapy by gene addition in the case of thalassaemias.

Example 2 A 13-globin "micro locus" was constructed including all four of the five 5' DNaseI superhypersensitive sites (HSS). The method described below is general and the micro locus produced shows an advantageous increase in expression levels. The method described below may therefore be applied to the construction of other other micro locii from dominant activator regions of other genes.

Methods for the construction of a 8-qlobin micro locus To facilitate DNA manipulations required to place all four 5' DNaseI hypersensitive sites and the human 3-globin gene in a vector containing the tk neo gene, a series of polylinker vectors were constructed by inserting oligonucleotides between the Aat II and Pvu II sites of the plasmid vector pUC 18 to generate plasmids GSE 1364, GSE 1365, and GSE 1366 (see Figure The required DNA fragments were cloned into the unique restriction sites created in the different polylinker vectors. In particular: 1. A 2.1kb XbaI-XbaI fragment encompassing DNase HS1 was cloned into the XbaI site of GSE 1364, A' kI-i ka 4 _LL _nc including the whole hypersensitive site region has been deleted in this mouse.

mmmmmmmmmumauMM SWO 89/01517 PCT/GB88/00655 2. A 1.9kb Hind III-Hind III fragment was cloned into the Hind III site of the plasmid carrying HS1 to give a plasmid carrying HSS 1 and 2 in the same orientation as that found in the normal human 1-globin locus, 3. A 1.5kb Asp718-SalI fragment encompassing DNaseI HSS 3 was made blunt-ended by treatment with Klenow fragment of DNA polymerase deoxynucleoside triphosphates and cloned into the HindII site of GSE 1365, 4. A 1.lkb partial SacI fragment encompassing DNaseI HSS4 was similarly made, blunt-ended and cloned into the Smal site of the plasmid carrying HSS3 to give a plasmid with HSS3 and HSS4 in the same orientation ar that found in the normal human 3-globin locus, The 4.9kb BglII fragment containing a normal copy of the human 3-globin gene with 1.56kb of 5' flanking sequence and 1.84kb of 3' flanking sequence was cloned into the BamHI site of GSE 1366, 6. A 2.0kb partial NarI fragment which includes the Herpes Simplex Virus (HSV) thymidine kinase (tk) promoter driving the expression of the Tn5 neomycin resistance gene and 3' polyadenylated sequences of the HSV tk gene was cloned into the NarI site of the plasmid carrying the human 3-globin gene and, 7. The four HSS, the human 3-globin gene and the tkneo gene were combined in a three fragment ligation using a PvuI-BstEII fragment of HSS 1 and 2 with a BstEII Clal fragment of HSS 3 and 4 and a ClaI-PvuI fragment of the plasmid containing the 1-globin and tkneo genes.

WO 89/01517 PCT/GB88/00655 -36- The final construct was shown to direct high-level expression of the human 8-globin gene in mouse er'throleukemic cells (MEL) independent of the position of integration of the DNA in the host cell chromosome.

The level of B-globin expression was related to the gene copy number.

Example 3 To test the effect of changing the position of the DNaseI HSS relative to the B-globin gene, the DNaseI HSS fragments described in Example 2 above were cloned so as to give a 6.5kb NotI restriction fragment. This DNA fragment was cloned in both orientations into a unique NotI site in two different plasmids containing the tkneo gene as a 2.7kb partial EcoRI fragment and the human ,i3globin gene as a 4.9kb BglII fragment as described above.

The two tkneo B-globin plasmids differed in the orientation of the 3-globin gene and cloning of the NST I fragment in both orientations generated the four plasmids GSE 1359, GSE 1400, GSE 1401 and GSE 1357 shown in Figure 9.

The DNA of the four plasmids shown in Figure 9 was linearised by digestion with the restriction enzyme PvuI. The DNAs were recovered by ethanol precipitation and introduced into MEL cells growing exponentially in alpha MEM medium supplemented with 10% foetal calf serum and antibiotics.

MEL cells (approx, 107) were collected by centrifugation (1000 g for 10 minutes) washed once with ESB Buffer (20mM HEPES pH 7.0, 140mM NaC1, 5mM KCl) and resuspended in 1ml of ESB containing CI WO 89/0151 P /GB88/O SWO 89/01517 PCT/GB88/00655 -37- 100g of linearised plasmid DNA. After incubation on ice for 10 minutes, cells were placed in a well of a 24 well tissue culture plate (Nunc) and a 10ms shock was delivered using a Haeffer "Pro genitor" electroshock apparatus at a voltage gradient of 250 volts/cm (Antonious et al EMBO J, (1988), 1, 377-384).

Cells were recovered into 40ml of aMEM and 10% foetal calfserum at 370C for 24 hours and G418 selection was applied by changing the tissue culture medium for aMEM with 10% foetal calf serum and 800 mg G418 (Gibco BRL) per ml. After 11 days in selection G418 resistant cell populations were harvested by centrifugation. RNA and DNA was prepared and part of the culture was induced to undergo erythroid differentiation by growth in aMEM with 10% foetal calf serum and 2% dimethyl sulphoxide (DMSO-BDH). After four days, populations were harvested and RNA prepared by homogenising cells in 3MLiCl 6M urea.

After sonication and incubation at 400 overnight, RNA was precipitated and analysed by Northern blot analysis, RNA from uninduced or induced MEL cell populations (10mg) was denatured with formamide 15% formaldehyde in 1 x MOPS buffer heated at for 10 minutes, chilled on ice and loaded on to a 1% agarose gel. The gel was run at 70 volts/cm for 8 hours in 1 x MOPS buffer (20mM MOPS pH7, Sm sodium acetate 20mM EDTA). The gel was soaked in 20 x S!eC (3 M NaC10.15M trisodium citrate) fcr 30 minutes then blotted directly to a nitrocellulo se (Schleicher and Schuller-0.1pM) overnight. filter was baked for 2 hours at 8000C then hybridlsed to DNA probes labelled to high specific activities by Nick translation (Rigby et al, J. Mol Biol. (1977) 113 mug WO 89/01517 W089/157Pcr/GB88/0065.5 -38- 237-251).

Hybridisation was performed in buffer containi,ig formamide, 6 x SSC, 10 x Denhardts 1% SDS and 2.OOpg salmon sperm DNA per ml (Lang et al EMBO J (1988), 1675-1682), After overnight hybridisation at 420C, filters were washed to a final stringency of 0.1 x SS0 0.1% SDS at 5500. Filters were exposed, to x ray film with intensifying screens at 7000.

Specific DNA probes used in the exper4iments of Figure were: Human 13-clobin A 800 bp EcoRI- MspI fragment which includes the 3rd exon of the human 1-globin gene.

Mouse a-cilobin A 300 bp BamilI fragment of the mouse ta-globin gene including sequences from the 2nd exono tkneo0 A 2.0kb partial Nar 1. fragment encompassing the tk promoter and the Neo resistance gene, Mouse Histone H4 gene A, 300 bp Hinfl fragment of a mo'use histone H4 gene.

The experiment described above shows that: 1. Human i-globin RNA expression in induced MEL cells is as high or higher than the expression of end~ogenous mouse a-globin genes, WO 89/01517

I

PCT/GB88/00655 -39- 2. 0-globin expression using 8 micro locus constructs is higher than that obtained with a 3 mini locus in MEL cells, 3, The dominant activator sequences increase transcription from the tkneo gene approximately 100 fold compared to plasmids that do not contain the dominant activator sequences and 4. The dominant activator sequence works on the human 1-globin gene in both orientations upstream or downstream of the 3-gene, Example 4 Experiments were conducted to identify any dominant activator sequences in the human CD2 gene and to test their ability to direct integration site independent expression, Two DNasel hypersensitive sites were found 3' of the human CD2 gene which is expressed at a high level in the ktiryA of transgenic mice. A 5.5kb BamHI Xbal fragment of CD2 3' sequence including the DNaseI HSS was clone next to a 4.8 kb BglII human 3-globin gene fragment (Example 5 below) or a 9.5kb mouse/human hybrid Thy-1 gene fragment, DNA fragments prepared from 3CD2 and Thy-i CD2 constructs were introduced into fertilized mouse eggs. Transgenic mice with a range of copy numbers were obtained and analysis showed high -level expression of human 0-globin (example 5) and mouse/human hybrid Thy-i gene MRNA in thymus (this example) which was related to copy number and independent of the site of integration into the mouse genome.

The experiments described in this Example and in WO 89/01517 WO 9/0517PCT/GB88/00655, Example 5 demonstrate unequivqcaJlly the presence of a dominant activator sequence in the F 5kb Ba'mHI -XbaI fragment 3' of the human CD2 gene.

I Mapping DNase HSS in the C2 genp Transgenic mice carrying the 2.8kb KpNI fragment of the human CD2 gene were sacrificea and nuclei prepared from thymus and liver of mouse lines CD2-4 and CD2-l (Lang et al, EMBO J, (1988), V,6j), 1675-1682). DNaseI digestions were p~erformed as describe)d herein (see also Grosveld et al, Cell, (1988), 51, 975-985).

DNA samples were recut with the restsriction enzymes indicated in F'igure 11 and probed with 5' or 3' human CD2 probes. strong DNAs FISS were present 5' of the gene and two 3' sites are seen 3' of the CD)2 gene only it in transgenic thymus not in liver nuclei., Fi~nure 12 shows a Thy-1,1 CD2 cohnttruct, The stippled bar denotes human Thy-i sequences and the two arrows show the position of the DNaseI HS$ in the CD2 fragment,, A 5.5kb BamHI Xbal fragment 3' of thei human CD2 gene shown to contain the two 3' D~ase I IiSS was cloned between, BamfII and XbaI sites of the blue script vector 1XS+ (Promega). A 9.5kb EcoRI fragment______ containing a mouse-human hybrid gene (.14ollias et a!, cellf (1987), 51t 21-23), was cloned into the EcoRl site next to the S. 5kb Bamlfl Xb-J'I fragment to give the plasmid Thy'-l CD2 A shown in Fiqure 12, Plasmid sequences were removed by digestion with restricti.on endonucleases. Sall and Not I and the 15kb Thy-i cD2 insert was purifiod by gel eloctrophoresis and passage over and Elutip D Column (Schleicher and Schueli) W W6 89/01517 PCT/GB88/00655 -41- The purified DNA was introduced into the nuclei of fertilized CBA x C57 BL-10 mouse eggs by microinjection (Lang ge. i, EMBO J (1988), 1, 1675-1682). Injected eggs were transplanted into the oviduct of pseudo pregnant foster mothers by standard techniques (Hogan et al, "Manipulating the mouse embryo" Cold Spring ,arbour publications (1986)).

Transgenic mice were identified by Southern blot analysis of tail DNA as described previously (Lang et al Loc cit) and tissues were dissected for RNA preparation when animals were 3 to 4 weeks old.

FIperring to Figure 13 RNA from thymus or brain of transgenic mice 1,2,3 5,6,7 and 8 (51g) was hybridized with a 5' end labelled Tth II Sac I mouse Thy-1 IVth exon probe and a 3' end labelled BglII Neo I human Thy-1 probe at 520C for 12 hours after melting the probes or RNA at 850C for 15 minutes.

RNA DNA hybrids were digested with Nuclease S1 (Boehringer Mannheim) as described previously Kollias et al Cell, (1988), 51, 21-31) and protected fragments run out on a 5% polyacrylamide 8M urea sequencing gel.

The positions of protected fragments for endogenous mouse and transqenic mouse humr. Thy-1 RNAs are indicated.

The level of expression per gene copy of the hybrid mouse human Thy-l gene in transgenic thymus is as high as that for t, e endogenous Thy-1 gene and is related to the gene copy number (mouse 1 has approximately copies of the transgene-mice 2 has 10 copies of the transgene.

I, WO 89/01517 PCT/GB88/00655 -42- This example refers to Figure 14 which shows the constsruction of a 3-globin CD2 plasmid.

3-globin CD2 expression in transgenic mice was brought about vsing dominant activator sequences of the invention.

The inserts of plasmid 8CD2-A and 3-CD2- were isolated by agarose gel electrophoresis after digestion with restriction enzymes Nat I and Sal I. DNA was injected into nuclei of fertilized mouse eggs as described above (see also Grosveld et al, cell (1987), 51 975-985.

Transgenic mice were identified by Southern blot analysis of tail DNA using a human 8-globin indicator probe (900bp BamHI-EcoRI fragment).

Figure 15 shows 5mg of genomic DNA from transgenic mice carrying either the 3CD2-A or 3CD2-8 construct applied to a nitro cellulose filter using a slot blot manifold (Schleicher and Scheull), The filter was hybridised to a nick-translated 3-globin 2nd intron probe and washed to a final stringency of O,lxSSC 0.1% SDS at The strength of the hybridisation signal is a direct measure of the number of gene copies on each mouse.

Figure 15B shows an Sl Nuclease alalysis of thymus RNA prepared from transgenic mice. RNA (Img) was hybridized with a 5' end labelled 52.5 bp AccI fragment of a 3-globin CDNA construct (Grosvold et al, cell, (1987), 1, 975-985 and a 300bp Hinf I fragment of a mouse histcne H4 gene as an internal control, After hybridisation at 52°C for 12 hours, samples were digested with Nuclease Sl and protected fragments WO 89/01517 PCT/GB88/00655 -43analysed on a 5% polyacrylmide 8M urea gel. The positions have protected fragments for endogenous histone H4 and human 3-globin mRNA are indicated. The intensity of the human 3-globin-protected fragment is a direct measure of the amount of human 3-globin mRNA transcribed in the thymus. For comparison 0.1 mg of foetal liver RNA from a transgenic mouse carrying 4 copies of the 3-micro locus was used in the track labelled 33 FL.

Referring to Figure 14, the '5.5kb BamHI-XbaI fragment of human CD2 gene was cloned between the BamHI and Xbal sites of the blue script vector KS+ (promega). Tihe 4.9kb BglIl fragment of the human 2-globin gene was cloned into the BamHI site of this plasmid in both orientations (3CD2A and 3CD2-3). The vertical arrows indicate positions of CD2 DNaseI HSS. The black bars donate exons of the human 3-globin gene and horizontal arrows show the orientation of the gene fragments used.

It will be understood that the invention is described by way of example only and modifications may be made within the scope of the invention.

WO 89/01517 PCT/GB88/0065I -44~

IFERENCES

Alter, B. Goff, S. Ef remor, G. Gravley, M. and Huisman, T. (1 k 8O) Globin chain electrophoresis: a new approach to the determinaition of the G /jratio in foetal. haemoglobin and to the studies of globin synthesis. Brit. J. Haematol. 44, 527-534.

Angel, Imagawa, Chin, Stain, Imb'ra, Rabxnsdorf, M., 'Jonat, Herrlich, P. and Karin, M. (1987) Phorbol ester-inducible genes contain a common cis element recognized by a TPA-mediated transacting- factor. Cell 49, 729-739.

Behringer, Hemmer, R. Brinster, R. Palmiter, R. D. and Townes, T. M. (1987) Two 3' sequences direct erythroid-soecific expression of human P-globin genes in transgenic mice. ?roc. Natl.

Acad. Sci. U.S.A. (in press) Brinster, R. Chan, H. Trumbauer, M. Yagle, IA. Z. and Palmitar, R. D. (1985) Factors affecting the efficiency of introducing foreign DNA into mice by microinjectin(U eggs, Proc. Natl. Acad. Sci.

U.S.A. 82, 4438-4442, Chada, Magram, J. Raphael, K. Radice, G. Lacy, E. and Costantilni, F.

(1985) Specific expression of a foreign P-globin gene in eryjthroid cells of transgenic mice. Nature 314, 377-330.

Charnay, Treisman, Mellon, Chao, Axel, R. and Maniatis, (1984) Differences in human a- and P-globin gene expression in mocuse erythroleukemia cells: the role of intragenic sequences. Cell 38, 251-263.

'W

4 6 89/01517 PCT/GB88/00655 Choi; 0-R. and Engel, J. D. (1986) A 3' enhancer is required for temporal and tissue specific transcriptional activation of the chicken adult beta-g.obin gene. Nature 323, 731-734.

Cockerhill, P. and Garrard, W. (1986) Chomosomal loop anchorage of the kappa Ig gene locus next to the enhancer in a region containing topoisomerase II sites. Cell 273-282.

Collins, F. S. and Weissman, S. M. (1984) The molecular genetics of human hemoglobin. Frog. Nucl. Acid Res. Mol. Biol. 31., 315-462.

Cone, R1. Benarous, W. W. Baorto, D. and Mulligan, R. C. (1987) Regulate.? expression of a complete human P-globin gene encoded by a transmissible retrovirus vector. Mol. Cell. Biol. 887t-697.

Curtin, P. and Kan, Y. W. (1966) A normal P-globin gene is inactivated b7 a large distant deletion. In:, Hemoglobin Switching. Fifth Conference on Hemoglobin Switching, Washington. Ed, G. Stamatoyannopoulos, I Alan R. Liss Inc., New York, 1987, in press.

nver, Brewer, A. C and Patient, R. (1985) Simian Viris cis induction of the Xenonus P-globin DNaseI hypersensitive site, Nature 318 680-683.

Gasser, S. and Laemmli, U. (1986) Cohabitation of scaffold binding regions with upstream/ enhancer elements of three de'v elonoment ally regul ated gen.es of D.melanogaster. Cell 46., 521-530.

Grosschedl, Weaver, D. Baltimore, D. and Costantini, R. (1984) Introduction of a t Ig gene into mouse gerMlinje- pcfcexrs ni lymphoid cells and synthesis of functional antibody. Cell 38 647-658,

NOR

WO 39/01517 -44PCT/GB88/0064 Grosveld, F. Lund, Murray, E. Mellor, A. Dahl, H.H.M. and Flavell, R. A. (1982) Thie construction of cosmid libraries which can be used tc' transform eukaryotic cells. Nucl. Acid Res. 10,1 6715-6732.

Groad-i-.io, Kohwi-Shigematsu, Gelinas, Stamat~tyannppoulos, G.

and Panayannopoulou, T. (1983) Human fetal to adult hemoglobin switching: changes in chromatin structure of the P-globin locus.

Proc. Natl. Acad. Sci. U.S.A. 80, 7551-7555.

Hamnmer, Krumlauf, Oampen, Brinster, R. and Tilghman, S.

(1987) Diversity of aL-fetoprotein gene expression in mice is generated by a combination of separate enhancer elements. Science 235, 53-58.

Hesse, J. Nickcl, J. Lieber, M. R. and Felsenfeld, G. (1986) Regulated gene expression in transfected primary chi.c'en erythrocytes.

Proc. NYazl. Acad. Sci. U.S.A. 83, 4312-4316.

Fock, R. A. and Miller, A. D. (1986) Retroviruas-mediated transfer and expression of drug resistance genes in human haematop iei poenitor cells. Natie 320, 275-277.

Xioussis, Vanin, deLange, Flavell, R. A. and Grosveld, F. G.

(1983) P-globin gene inactivration by DNA translocation i.n yp-tha.assaemia. Nature 306, 662-666.

Klein, G. Zeuthen, J. Eriksson, Tekasaki, P. Bernoco, MA., Rosen, A.

Ma~aucci, Povey, S. and Ber, R. (1980) Hybrid-- zat.i n of a =7feloid 3,eukaemia derived human cell line (1562) with a Burkitt -s Lymphoma line.

J. Nati. Canc. Ins'.. 64., 725-738.

ThKlJ-i4s, Hurst, de~oer, E. and Grosveld, F. (1987) A tissue and developmental speciw"it' enhancer is located downstream from the human P-globin gene. Nucl. Acids 'Res. 15, 5739-5747., WO4 89/01517 PCT/GB88/00655 Kollias, G. Wrighton, Hurst, J. and Grosveld, F. (1986) Regulated expression of human and hybrid yp3-glo'bin genes in transgenic mice: Manipulation of the developmental expression patterns. Cell .,89-94.

Lawson, Knoll, March, Woo, Tsai, M. and O'Mal~ley, B. (1982) Definition of 5' and 3' structural boundaries of the chromatin domain containing the ovalbumin multigene g~amily. J. Biol. Chem. 257, 1501-1507.

Lee, W. Mitchell, P. and Tj ian, R. (1987) Purified transcription :eactor AP-1 interacts withx TPA-inducible enhancer elements. Cell 44, 273-282.

Magraxn, J. Chada, K. and Costantini, F. (1985) Developmental regulation ofacoe dl -globin gene in lzransgenic mice. Nature 315 338-340.

Mani at is, T. Fri tsch, E. Lauer, J. and Lawn, R. M. (1981 Molecular genetics of human hemoglobins. Ann. Rev. Genet. 14, 145-178.

{I .Martin, P. and Papayannopoulou, T. (1982) HEL cells: A new human erythroleukemia cell line with spontaneous and induced globin expression. Science 216 1233-1235.

Nicholls, Fischel-Ghodsian, N. and Higgs, D. (1987) Recombination at the human a-globin gene clusrer, seaqnce features an-d topological constraint, Cell 49, 399378.

Palmiter, R. D. and Brinster, R. L. (1986) Germ-line tranisformation of mice. Ann. Rev. Genet. 20, 465-499.

WO 89/01517 4PCT/GB88/00655 A, Pinkert, C. Ornitz, Brinster, R. and Falmiter, R. (1987) An enhancer located 10kb upstream functions along with its promoter to direct efficient, liver-specific expression in transgenic mice. Genes and Devel. 268-278.

Southern, E. M. (1975) Detection of specific sequences among DNA fragments separated by gel electrophoresis. J. Mol. Biol. 9-8, 503-517.

Taramelli, Zioussis, Vanin, Bartram, Groffen, Hurst, J.

and Grosveld, F. G. (1986) T-f-thalassaemias 1 and 2 are the result of a lO0kbp deletion in the human O-globin cluster. Nuol. Acids Res. 14, 7017-7029.

ii Townes, T. Lingrel, J. Chen, H. Brinster, R. L. and ?=lmiter, R. D. (1985) Eryhroid specific expression of human P-globin g-enes in transgenic mica. EY-O J. 1715-1723.

Tuian, Solomon, Qiliang, L, S. and Irving, M. L. (1985) The "P-like-globin" gene domain in human erythroid cells. Proc. Nati1. Acad.

Sci,. U.S.A. 2, 6384-6388.

van der Ploeg, L. H. Konings, Qort, Roos,,D., Bernini, L. and FPlavell, R. A. (1980) yo-thalassaemia: deletion of the and influ'-ences O-globin gene expression in man. Nat.urs 283, 637-642.

rihS., Rosenthal, Flavell., R. A. a-rd Grosveld, F. (1984) -'A secuences required for regulated expression of O-glooiii genes in raurine ery-throleukemi.a cells. Cell 3.j, 265-273.

Zavodny, P. Roginsky, R. and Skoultchi, A. ,1983) In: Globin geaie expression and hematopoietic differentlatio, Fls. Stamatoya.-'.opoulas, G. and Nienhuis, Alani R. Liss Inc., No~w York, pp. 53-62.

Claims (12)

1. A vector for the integration of a gene into the genetic material of a mammalian host cell such that the gene may be expressed by the host cell, the vector comprising a promoter and the said gene, characterised in that the vector includes a dominant activator sequence capable of eliciting host cell-type restricted, integration site independent, copy number dependent expression of the said gene.

2. A vector according to claim 1 wherein the dominant activator sequence comprises a DNase I super hypersensitive site of a mammalian gene system.

3. A vector according to claim 2 wherein the dominant activator sequence comprises a 3-globin DNase I super hypersensitive site.

4. A vector according to claim 3 wherein the dominant activator sequence is within a 21kb fragment from -1kb Clal to -22kb B~1II immediately upstream of the epsilon-globin gene in the 3-globin locus. A vector according to claim 3 wherein the dominant activator sequence comprises one or more of the following fragments: 2.1kb XbaI-Xba 1.9kb }HindIII HindIIl KpnI-BqlII,and l.lkb partial SacI fragment as herein defined. ,Lk."-yj-y e 1L i Moi Blo: (1977) 113; 1 i PCT/GB88/00655 WO 89/01517

6. A vector according to claim 2 wherein the dominant activator sequence comprises a CD2 DNase I super hypersensitive site.

7. A vector according to claim 6 wherein the dominant activator sequence is within a 5.5kb BamHI to Xbal fragment 3' to the CD2 gene.

8. A mammalian host cell transformed with a vector according to any one of the preceding claims.

9. A transgenic mammal transformed with a vector according to any one of claims 1 to 7. A method of producing a polypeptide comprising culturising a mammalian host cell transformed with a vector according to any one of claims 1 to 7.

11. A method of gene therapy comprising the steps of i) ii) iii) removing stem cells from the body of a mammal, killing stem cells remaining in the body, transforming the removed stem cells with a vector as defined in any one of claims 1 to 7 containing a gene deficient or absent in the body, and replacing the transformed stem cells in the body.

12. A method of gene therapy comprising the steps of i) removing cells from the body of a mammal, ii) transforming the removed cells with a vector as defined in any one of claims 1 to 7 containing a gene deficient or absent in the body, and iii) replacing the transformed cells in the body. 1 r i i 3t~~/ I I I 51 12. A vector according to any one of claims 1 to 7 substantially as hereinbefore described,

13. A mammalian host cell according to claim 8 substantially as hereinbefore described.

14. A transgenic animal according to claim 9 substantially as hereinbefore described. A method according to any one of claims 10 to 12 substantially as hereinbefore described. DATED this 31 July 1991 10 CARTER SMITH BEADLE Fellows Institute of Patent Attorneys of Australia i Patent Attorneys for the Applicant: FRANKLIN GERARDUS GCROSVEID S9 4 mwspe#3293 91 7 31