WO2005059149A2

WO2005059149A2 - Improved protein production

Info

Publication number: WO2005059149A2
Application number: PCT/EP2004/053378
Authority: WO
Inventors: Arie Pieter Otte; Theodorus Hendrikus Jacobus Kwaks; Richard George Antonius Bernardus Sewalt; Henricus Johannes Maria Van Blokland; Abraham Bout
Original assignee: Chromagenics B.V.; Crucell Holland B.V.
Priority date: 2003-12-12
Filing date: 2004-12-09
Publication date: 2005-06-30
Also published as: WO2005059149A3

Abstract

The invention provides E1-immortalized retina cells comprising a recombinant protein expression unit which unit comprises a promoter functionally linked to an open reading frame encoding at least one protein of interest, characterized in that said expression unit comprises at least one element improving expression, wherein said element improving expression is chosen from the group consisting of: a) a stabilizing anti-repressor (STAR) sequence; b) a Transcription Pause (TRAP) sequence; and c) a binding site for a member of a chromatin modification system for rendering chromatin more accessible for transcription (opener), wherein said opener is present in said cell. Methods for producing at least one recombinant rotein in such cells are also provided.

Description

Title: Improved protein production

The invention relates to the fields of medicine and cellular biology. More specifically, the invention relates to means and methods for regulating gene expression, and production of recombinant proteins.

Recombinant proteins can typically be produced from several different host cells. The choice for a particular host cell is usually based on various aspects, including the desired post-translational modifications of the recombinant protein, as well as economic considerations, such as yields, scalability, stability, costs of culture media and culturing facilities, ease of manipulation, time to establish producing clones, and the like. One class of particularly suitable host cells for the recombinant production of a variety of proteins, are El-immortalized retina cells, such as PER.C6™ cells. These cells were created as convenient packaging cells for El-deficient adenoviral vectors (US patent

5,994,128). Recently it was surprisingly found that such cells are capable of expressing high yields of recombinant proteins, including antibodies (international patent application WO 00/63403; Jones et al., 2003). These cells differ in many respects compared to other cells that are frequently used for protein production, such as for instance CHO cells. The immortalization by El sequences, as well as the retina origin of the cells, may influence these differences. For instance, glycoproteins produced recombinantly in El-immortalized retina cells have a characteristic glycosylation pattern (see for instance WO 03/038100) . Furthermore it has been observed that production levels of recombinant proteins in El- immortalized retina cells are dependent on the promoter used, and that the strength of certain promoters differs between these cells and other cell lines . Furhtermore, it has been observed that El-immortalized cells differ from CHO cells in many parameters, including their behaviour with respect to nutrient limitation, obtainable cell densities, consumption of waste products, and population doubling times, which may influence expression of recombinant proteins. It is therefore not a priori known whether certain measures taken in other cell lines will be applicable in El -immortalized retina cells and vice versa. It is one object of the present invention to further improve the process of producing recombinant proteins, including antibodies, in El-immortalized retina cells. A variety of so-called STAR (Stabilizing Anti- Repression) sequences have been described in WO 03/004704. These sequences are capable of performing at least one of the following functions: (a) inhibiting the effect of gene transcription repression elements, (b) at least in part blocking chromatin-associated repression, (c) at least in part blocking the activity of an enhancer, (d) conferring at least one of the following effects upon an operably linked nucleic acid comprising a transcription unit compared to the same nucleic acid alone: (d(i)) a higher predictability of transcription, (d(ii) ) a higher transcription, and/or (d(iii)) a higher stability of transcription over time. Hence, STAR- sequences can be used to improve predictability, yield and/or stability of protein production using expression constructs in at least some host cells, including CHO cells (WO 03/004704; Kwaks et al, 2003) . The efficiency of the different known STAR elements varies (WO 03/004704; Kwaks et al, 2003) . Furthermore, at least some of the STAR sequences function in host cells of different origin (WO 03/004704; Kwaks et al, 2003) , whereas some STAR-ele ents may be cell type specific (Kwaks et al, 2003; international patent application no. PCT/NL03/00410) . Hence, an optimal result may require the tailoring of a combination of STAR element with the host cell line. International patent application no. PCT/N 03/ 00410, filed 30 May 2003, describes the use of STAR sequences for expressing nucleic acid encoding recombinant proteins in several cell lines, but not El-immortalized retina cells . International patent application no. PCT/NL03/00432, filed 13 June 2003, describes the use of STAR sequences for expressing nucleic acid encoding multi eric proteins, such as antibodies, but not in El-immortalized retina cells . European patent application no. 02080347.4, filed 18 December 2002, describes the use of so-called TRAP (Transcription Pause) sequences, optionally in combination with STAR sequences, for expressing nucleic acid encoding recombinant proteins in certain cell lines, but not El-immortalized retina cells. European patent application no. 03075089.7, filed 13 January 2003, describes the use of so-called openers, which are proteins that render chromatin more accessible for transcription, which are used optionally in combination with STAR sequences, to aid in expressing nucleic acid encoding recombinant proteins in certain cell lines, but not El-immortalized retina cells. There is a need in the art for improved protein production in recombinant host cell lines.

Summary of the invention

The present invention provides an El-immortalized retina cell comprising a recombinant protein expression unit which unit comprises a promoter functionally linked to an open reading frame encoding at least one protein of interest, characterized in that said expression unit comprises at least one element improving expression, wherein said element improving expression is chosen from the group consisting of: a) a stabilizing anti-repressor (STAR) sequence; b) a Transcription Pause (TRAP) sequence, and wherein said TRAP sequence is located: i) downstream of the coding sequence of said protein in an orientation that can at least in part prevent formation of antisense RNA of said coding sequence; or ii) upstream of said promoter and in an orientation that can at least in part prevent transcription to enter said protein expression unit; and c) a binding site for a member of a chromatin modification system for rendering chromatin more accessible for transcription (opener) , wherein said opener is present in said cell. In certain embodiments, said element improving expression is a STAR sequence. In certain embodiments, said STAR sequence is chosen from the group consisting of STAR 4 (SEQ. ID. NO. 1), STAR 6 (SEQ. ID. NO. 2); STAR 7 (SEQ. ID. NO. 3), STAR 12 (SEQ. ID. NO. 4), STAR 18 (SEQ. ID. NO. 5), STAR 35 (SEQ. ID. NO. 6) and STAR 40 (SEQ. ID. NO. 7), and a functional fragment or derivative thereof. In preferred embodiments, said STAR sequence is STAR 4,

STAR 6, or STAR 7, or a functional fragment or derivative thereof. In other embodiments, said element improving expression is a Transcription Pause (TRAP) sequence, and wherein said TRAP sequence is located: a) downstream of the coding sequence of said protein in an orientation that can at least in part prevent formation of antisense RNA of said coding sequence; or b) upstream of said promoter and in an orientation that can at least in part prevent transcription to enter said protein expression unit. In certain embodiments hereof, said TRAP sequence is chosen from the group consisting of lambda fragment 35711-38103 (SEQ. ID. NO. 8), a synthetic polyA sequence, for instance as identified by (SEQ. ID. NO. 9 or SEQ. ID. NO. 14), a 92 bp pausing signal from the human μ2 globin gene (SEQ. ID. NO. 10), a combined synthetic polyA sequence and a pausing signal from the human μ2 globin gene (SEQ. ID. NO. 11 or SEQ. ID. NO. 15), the inter histon H3FA-H4F fragment (SEQ. ID. NO. 12) and the inter histone H1F4-H2FB fragment (SEQ. ID. NO. 13) , and a functional fragment or derivative of any of these. In certain embodiments, said TRAP sequence is a combined synthetic polyA sequence and a pausing signal from the human μ2 globin gene (SEQ. ID. NO. 11 or SEQ. ID. NO. 15), or a functional fragment or derivative thereof. In certain embodiments the expression unit comprises at least one STAR and at least one TRAP sequenc . In yet other embodiments, said element improving expression comprises at least one binding site for a member of a chromatin modification system for rendering chromatin more accessible for transcription (opener) , wherein said opener is present in said cell . In certain embodiments hereof, said binding site comprises a lexA or a GAL4 binding site, and said opener is reσombinantly expressed in said cell as a fusion protein comprising a lexA binding domain or GAL4 binding domain rendering it capable of binding to said binding site, and a domain that renders chromatin more accessible for transcription. In certain embodiments, the expression unit comprises at least one TRAP sequence and at least one binding site for a member of a chromatin modification system for rendering chromatin more accessible for transcription (opener) , wherein said opener is present in said cell, and optionally said expression unit further comprises at least one STAR sequence. In certain embodiments, said recombinant protein is an immunoglobulin, such as an antibody. In other embodiments, said recombinant protein is erythropoietin. The invention further provides a culture of cells, wherein the cells are cells according to the invention. In preferred embodiments, an El-immortalized retina cell according to the invention is a cell such as deposited at the ECACC under number 96022940. The invention further provides a method for producing at least one recombinant protein in a cell, said method comprising culturing of said cell and expressing said recombinant protein, characterized in that said cell is a cell according to the invention. In a preferred embodiment, said recombinant protein is collected, either from the cells or from the culture medium or from both.

It is another aspect of the invention to provide a recombinant nucleic acid in the form of an expression unit for recombinant expression of an immunoglobulin, characterized in that said recombinant nucleic acid comprises in operable association in the 5' to 3' direction: at least one STAR element - promoter - coding sequence for an immunoglobulin heavy or light chain - polyadenylation signal - at least one STAR element. Said expression unit may optionally further contain at least one TRAP sequence and/or one binding site for a member of a chromatin modification system for rendering chromatin more accessible for transcription (opener) . In another aspect the invention provides a cell comprising said recombinant nucleic acid. In yet another aspect, the invention provides a method for recombinantly producing an immunoglobulin, the method comprising the steps of: a) providing a cell comprising recombinant nucleic acid in the form of an expression unit for recombinant expression of an immunoglobulin, characterized in that said recombinant nucleic acid comprises in operable association in the 5' to 3' direction: at least one STAR element - promoter - coding sequence for an immunoglobulin heavy or light chain - polyadenylation signal - at least one STAR element; b)' culturing said cell to express said recombinant nucleic acid. In a preferred embodiment, the thus produced recombinant immunoglobulin is collected.

Detailed description of the invention STAR sequences STAR sequences (also referred to as STAR elements herein) suitable for the present invention, as well as methods to obtain and identify them, and their use, have been disclosed in WO 03/004704, incorporated herein in its entirety by reference. As a non-limiting example, Fig. 26 of WO 03/004704 provides the sequences of 65 human STAR elements (STAR 1-65) , each of which is encompassed by the term STAR sequence' in the present invention. Said STAR sequence has to be operably linked to the nucleic acid encoding the recombinant protein in order to be effective. In one embodiment of the invention one STAR element is used. Preferably however, more than 1 STAR element is used. In a particularly preferred embodiment the nucleic acid encoding the recombinant protein is provided with two STAR sequences; one STAR sequence at the 5 ' side of the coding sequence encoding said recombinant protein on the nucleic acid and one STAR sequence at the 3' side of said coding sequence on said nucleic acid. The present invention discloses that at least the following STAR elements function in El-immortalized retina cells: STAR 4 (SEQ. ID. NO. 1), STAR 6 <SEQ. ID. NO. 2); STAR 7 (SEQ. ID. NO. 3), STAR 12 (SEQ. ID. NO. 4), STAR 18 (SEQ. ID. NO. 5), STAR 35 (SEQ. ID. NO. 6) and STAR 40 (SEQ. ID. NO. 7) . Preferably, STAR 4, STAR 6 or STAR 7 is used, as these give the best results in the cells according to the invention. According to the invention, a functional fragment or derivative of a given STAR element is considered equivalent to said STAR element, when it has STAR activity. Such activity should be similar in kind, not necessarily in amount. Functional fragments or derivatives can easily be obtained by a person skilled in the art of molecular biology, by starting with a given STAR sequence, and making deletions, additions, substitutions, inversions and the like. The activity of such functional fragment or derivative can be checked with assays as described in WO 03/004704, to test for STAR activity. Sequences comprising STAR activity were identified in stretches of 400 bases. However, it is expected that not all of these 400 bases are required to retain STAR activity. Methods to delimit the sequences that confer a certain property to a fragment of between 400 and 5000 bases are well known. The minimal sequence length of a fragment comprising STAR activity is estimated to be about 50 bases. One suitable method for testing whether a sequence has STAR activity in a cell comprises (see WO 03/004704) : providing said cell with a vector comprising i) an element with gene-transcription repressing quality and ii) a promoter directing transcription of a reporter gene, wherein said element with gene-transcription repressing quality represses transcription of the reporter gene by said promoter, said vector further comprising the sequence to be tested for STAR activity, preferably placed in between said elements i) and ii) , the method further comprising testing whether the transcription levels from said reporter gene are higher in the presence of said sequence to be tested than in the absence thereof. When the transcription levels are higher in the presence of the tested sequence, the tested sequence has STAR activity. It has been described that STAR elements have at least one of three consequences for production of (heterologous) proteinaceous molecule (also referred to as heterologous or recombinant protein herein) : (1) they increase the predictability of identifying host cell lines that express a proteinaceous molecule at industrially acceptable levels; (2) they result in host cell lines with increased protein yields; and/or (3) they result in host cell lines that exhibit more stable protein production during prolonged cultivation. Each of these attributes is discussed in more detail below: (1) Increased predictability: Integration of transgene expression units can occur at random positions throughout the host cell genome. However, much of the genome is transcriptionally silent heterochromatin. When the expression units include STAR elements flanking the transgene, the position of integration has a reduced effect on expression. The STAR elements impair the ability of adjacent heterochromatin to silence the transgene. Consequently, the proportion of transgene- containing host cells with acceptable expression levels is increased. (2) Yield: The levels of protein expression in primary populations of recombinant host cells, directly after transgene integration, have been surveyed. The expression level of individuals in the populations varies. However, when the transgenes are protected by STAR elements, the variability is reduced. This reduced variability is most conspicuous in that fewer clones are recovered that have low levels of expression. Furthermore, the populations with STAR elements commonly have individuals with strikingly high expression. These high-yielding individuals are favourable for production of proteinaceous molecules. (3) Increased stability: STAR elements increase the stability of transgenes in recombinant host cell lines by ensuring that the transgenes are not transcriptionally silenced during prolonged cultivation. Comparative trials show that, under conditions in which transgenes that are not protected by STAR elements are progressively silenced (5 - 25 passages in cultivation) , STAR element-protected transgenes continue to be expressed at high levels. This is an advantage during industrial production of proteinaceous molecules, during which cell cultivation continues for prolonged periods, from a few weeks to many months .

Hence, a STAR sequence or a functional fragment or derivative thereof can enhance expression of a heterologous proteinaceous molecule. A STAR sequence can exert its activity in a directional way, i.e. more to one side of the fragment containing it than to the other. Moreover, STAR activity can be amplified in amount by increasing the number of STAR elements . Promiscuous STAR elements are able to function in more than one host cell line. For example, STAR6 increases the predictability, yield, and stability of a transgene in both the U-2 OS human osteosarcoma cell line and in CHO-Kl (Chinese hamster ovary) cells. Other STAR elements are species-specific and/or cell type-specific; for example STAR8 increases the predictability, yield, and stability of transgenes in U-2 OS cells, but not in CHO-Kl cells . The STAR elements used in the present invention work in El-immortalized retina cells.

A cell line of the invention is particularly suitable for production of a proteinaceous molecule of interest, because said STAR sequence can enhance expression of a gene of interest (higher yield of a proteinaceous molecule, higher proportion of host cells with acceptable expression levels, and/or higher stability of a gene expression level) . Methods for generating a cell line are known in the art and many techniques are known to provide a cell with a nucleic acid of interest. Hence, a use of a cell line of the invention for the production of a proteinaceous molecule is also herewith provided.

In yet another aspect the invention provides a method for selecting a cell suitable for producing a proteinaceous molecule comprising:

- providing a host cell with a nucleic acid comprising a STAR sequence;

- selecting a cell with enhanced expression of a proteinaceous molecule; wherein said host cell is an El-immortalized retina cell. Siad STAR sequence is preferably chosen from the group consisting of STAR 4 (SEQ. ID. NO. 1), STAR 6 (SEQ. ID. NO. 2); STAR 7 (SEQ. ID. NO. 3), STAR 12 (SEQ. ID. NO. 4), STAR 18 (SEQ. ID. NO. 5), STAR 35 (SEQ. ID. NO. 6) and STAR 40 (SEQ. ID. NO. 7), and a functional fragment or derivative thereof.

Cells according to the invention

The cells according to the invention are El-immortalized retina cells. They have been derived from retina cells, by immortalization with adenovirus El sequences. The cells according to the invention comprise in their genome at least adenovirus ElA and preferably also ElB sequences. Preferably said adenovirus sequences encode all El proteins but lack sequences encoding pIX, or a part thereof. The ElA sequences may be under influence of their endogenous adenovirus ElA promoter, but may also be controlled by a heterologous promoter, such as for instance a PGK promoter. Preferably, the cells according to the invention are derived from primary cells . They may be cells of any origin, and are preferably of human origin. In one preferred aspect, the cells are derived from primary human retina cells. Immortalization of such cells with adenoviral El sequences has for instance been described in US patent 5,994,128. Accordingly, an embryonic retina cell that has been immortalized with El sequences from an adenovirus can be obtained by that method. Other cells expressing ElA and ElB of an adenovirus can be prepared accordingly. A cell according to the invention expresses the ElA and ElB region of an adenovirus. ElA protein has transforming activity, while ElB protein has anti-apoptotic activities. Furthermore, ElA may aid in increasing expression levels from the cells. In particularly preferred embodiments, El- immortalized retina cells according to the invention are cells such as deposited at the ECACC on 29 February 1996 under number 96022940. One El-immortalized cell line useful for the invention, and having the characteristics of the cells deposited at the ECACC under number 96022940, is marketed under the trade name PER.C6™ by Crucell Holland B.V. Such cells expressing the desired protein according to the invention can be obtained by introduction of nucleic acid encoding a recombinant protein, which nucleic acid further comprises at least one STAR sequence, into such El-immortalized retina cells. Preferably said cells are from a stable clone that can be selected and propagated according to standard procedures known to the person skilled in the art. A culture of such a clone is capable of producing recombinant protein of interest. Cells according to the invention preferably are able to grow in suspension culture in serum-free medium. It has previously been shown that El-immortalized retina cells can express intact human IgG and erythropoietin (WO 00/63403, incorporated herein in its entirety by reference) , that such IgGs have human-type glycans and said cells can be grown at large scale (Jones et al, 2003; Nichols et al, 2002) . Moreover, it has been demonstrated that El-immortalized retina cells are capable of expressing other immunoglobulin formats, such as IgM (international patent application no. PCT/EP03/50194, incorporated herein by reference) and IgA (European patent application no. EP 03102598.4, incorporated herein by reference) . Furthermore, a fed- batch process for the production of recombinant proteins has been developed (international patent application no. PCT/EP03/50390, incorporated herein by reference), resulting in yields of more than 1 g recombinant IgG per liter of culture. A protein of interest according to the invention can be any protein and non-limiting examples are enzymes, immunoglobulin chains, therapeutic proteins like anti- cancer proteins or diagnostic proteins. In certain embodiments, a protein of interest in the present invention is an immunoglobulin, such as an antibody. In other embodiments, a protein of interest is erythropoietin . The presence of STAR elements in the expression units according to the present invention results in improved expression from the cells according to the invention.

Expression units An expression unit is a nucleic acid sequence comprising at least a promoter functionally linked to an open reading frame encoding a protein of interest. An expression unit may further contain transcription termination and polyadenylation sequences . Other regulatory sequences such as enhancers may also be included. The expression units according to the invention further comprise at least one STAR sequence, or at least one TRAP sequence (see below) , or at least one binding site for a member of a chromatin modification system for rendering chromatin more accessible for transcription (opener) , wherein said opener is present in said cell (see below) . To obtain expression of nucleic acid sequences encoding recombinant protein, it is well known to those skilled in the art that sequences capable of driving such expression can be functionally linked to the nucleic acid sequences encoding the protein, resulting in recombinant nucleic acid molecules encoding a recombinant protein in expressible format. Functionally linked is meant to describe that the nucleic acid sequences encoding the protein or precursors thereof are linked to the sequences capable of driving expression such that these sequences can drive expression of the protein or precursors thereof. Useful expression vectors are available in the art, e.g. the pcDNA vector series of Invitrogen. Where the sequence encoding the polypeptide of interest is properly inserted with reference to sequences governing the transcription and translation of the encoded polypeptide, the resulting expression cassette is useful to produce the protein of interest, referred to as expression. Sequences driving expression may include promoters, enhancers and the like, and combinations thereof . These should be capable of functioning in the host cell, thereby driving expression of the nucleic acid sequences that are functionally linked to them. Promoters can be constitutive or regulated, and can be obtained from various sources, including viruses, prokaryotic, or eukaryotic sources, or artificially designed. Expression of nucleic acids of interest may be from the natural promoter or derivative thereof or from an entirely heterologous promoter. Some well-known and much used promoters for expression in eukaryotic cells comprise promoters derived from viruses, such as adenovirus, e.g. the ElA promoter, promoters derived from cytomegalovirus (CMV) , such as the CMV immediate early (IE) promoter, promoters derived from Simian Virus 40 (SV40) , and the like. Suitable promoters can also be derived from eucaryotic cells, such as methallothionein (MT) promoters, elongation factor lot (EF-lα) promoter, actin promoter, an immunoglobulin promoter, heat shock promoters, and the like. In one embodiment the sequence capable of driving expression comprises a region from a CMV promoter, preferably the region comprising nucleotides -735 to +95 of the CMV immediate early gene enhancer/promoter. This gives particularly high expression levels in cells expressing ElA of an adenovirus, such as the cells according to the invention.

Culturing a cell is done to enable it to metabolize, and/or grow and/or divide and/or produce recombinant proteins of interest. This can be accomplished by methods well known to persons skilled in the art, and includes but is not limited to providing nutrients for the cell. The methods comprise growth adhering to surfaces, growth in suspension, or combinations thereof. Several culturing conditions can be optimized by methods well known in the art to optimize protein production yields. Culturing can be done for instance in dishes, roller bottles or in bioreactors, using batch, fed-batch, continuous systems, hollow fiber, and the like. In order to achieve large scale (continuous) production of recombinant proteins through cell culture it is preferred in the art to have cells capable of growing in suspension, and it is preferred to have cells capable of being cultured in the absence of animal- or human-derived serum or animal- or human-derived serum components. Thus purification is easier and safety is enhanced due to the absence of additional animal or human proteins derived from the culture medium, while the system is also very reliable as synthetic media are the best in reproducibility.

The conditions for growing or multiplying cells (see e.g. Tissue Culture, Academic Press, Kruse and Paterson, editors (1973)) and the conditions for expression of the recombinant product may differ somewhat, and optimization of the process is usually performed to increase the product yields and/or growth of the cells with respect to each other, according to methods generally known to the person skilled in the art. In general, principles, protocols, and practical techniques for maximizing the productivity of mammalian cell cultures can be found in Mammalian Cell Biotechnology: a Practical Approach (M. Butler, ed., IRL Press, 1991). Introduction of the nucleic acid that is to be expressed in a cell, can be done by one of several methods, all known to the person skilled in the art, also dependent on the format of the nucleic acid to be introduced. Said methods include but are not limited to transfection, infection, injection, transformation, and the like.

TRAP sequences In certain embodiments the present invention uses a TRAnscription Pause (TRAP) sequence to enhance a protein expression characteristic of a protein expression unit. Such sequences and their use are described in co-pending European patent application no. 02080347.4, incorporated herein in its entirety by reference . Such a TRAP sequence may for instance be placed downstream of the coding sequence of said protein in an orientation that can at least in part prevent formation of antisense RNA of said coding sequence. It is thought that such a TRAP sequence at least in part prevents formation of antisense RNA or to at least in part prevent transcription to enter said protein expression unit from the 3' direction. Without wishing to be bound by theory, it is believed that high copy-numbers signify large amounts of (inverted) repeat sequences as well as numerous possibilities for transcriptional read-through to occur, this could lead to RNAi and gene silencing. It is believed that the counter-intuitive blocking of transcription according to the invention by TRAP sequences leads to stable transcription of a transgene. Due to the blocking no antisense RNA is formed and hence the formation of (double strand) dsRNA is inhibited. This could lead to a reduction or complete prevention of so- called RNAi, which involves the formation of dsRNAs of 21 to 23 basepairs, and is thought to be involved in gene silencing. One way of function of the present invention could be that such RNAi induced silencing is at least in part prevented, by the presence of a TRAP sequence downstream of the coding sequence. Usually DNA sequences such as the SV40 polyadenylation signal are used to terminate transcription by placing the SV40 polyadenylation signal immediately downstream of a gene that is expressed. In other words, transcription should be prevented from continuing downstream of the gene. In the present invention transcription blockers (TRAP) are preferably placed both upstream and downstream of the entire expression units, in such a manner that they prevent transcription to enter the expression units, this coming from upstream or downstream of the expression units. The orientation of TRAP when placed downstream is opposite of the usual orientation of the SV40 polyadenylation signals that are placed downstream of genes. In one embodiment, the invention provides a method for expression (or producing) of at least one protein of interest in a cell according to the invention comprising providing said cell with at least one protein expression unit which unit comprises a promoter functionally linked to an open reading frame encoding said at least one protein of interest, characterised in that said protein expression unit further comprises at least one TRAnscription Pause (TRAP) sequence and wherein said TRAP sequence is functionally located downstream of said open reading frame and at least in part prevents formation of antisense RNA. Preferably, said at least one TRAP sequence is in a 3 '-5' orientation (in relation to said coding region) . Preferably, said TRAP sequence reduces the formation of antisense RNA to a non-detectable level. Due to the presence of said TRAP the formation of antisense RNA is at least in part prevented and hence the amount of dsRNA is decreased. As a consequence, the level of small dsRNAs of 21 to 23 basepairs (RNAi) is also decreased and the corresponding (full length) RNA encoding a protein of interest will not be degraded. Hence, translation of said corresponding RNA results in (increased) expression of a protein of interest. Surprisingly, the use of TRAP sequences improves stability of expression. In the above-outlined embodiment, the TRAP sequence can for example be a terminator/polyadenylation signal sequence, but in an orientation which differs from a normally used terminator sequence behind an open reading frame in said protein expression unit. However, it is entirely possible that there are TRAP sequences that are bi-directional . Thus in the above embodiment it is only necessary that said TRAP comprises a TRAP function in the reverse orientation. That it might also encompass TRAP function in the normal orientation is not relevant to the observed e fect. Furthermore, the coding sequence may be followed by a transcription terminator in its normally used configuration, which is then followed by a TRAP sequence, which may for instance be a transcription terminator in reverse orientation. In another embodiment, the the TRAP sequence is located upstream of said promoter and at least in part prevents transcription to enter said protein expression unit. Preferably, said at least one TRAP sequence is in a 5" -3' orientation (in relation to said coding region). Again, a TRAP sequence used in a the latter embodiment can be a terminator/polyadenylation signal sequence, but this time the TRAP sequence is in an unusual position with regard to the open reading frame, because said TRAP is located upstream of the promoter that drives expression of said open reading frame . In this embodiment, the presence of a TRAP sequence at least in part prevents transcription from a promoter sequence located outside a protein expression unit. Hence, the RNA from the protein expression unit does not have to compete with other RNA and hence a more efficient protein production system is provided. The use of a TRAP to at least in part prevent formation of antisense RNA or to at least in part prevent transcription to enter said protein expression unit isolates said protein expression unit from negative effect from outside said unit. A TRAP sequence is herein functionally defined as a sequence capable of at least in part prevent formation of antisense RNA or to at least in part prevent transcription to enter said protein expression unit. One possible method for assaying whether a sequence has TRAP activity is provided in Example 1, and it will be clear that the person skilled in the art may vary some aspects of such a method to find out whether a sequence is a TRAP sequence according to the invention. As disclosed herein within the experimental part, a TRAP sequence can for example be a polyadenylation site and/or a pausing site, where the RNA polymerase II stalls. A TRAP can be derived from any source, as long as efficient termination of transcription is achieved. In one embodiment a TRAP is identified based on its ability to at least in part prevent formation of antisense RNA or to at least in part prevent transcription to enter said protein expression unit. Example 1 provides a method to test the effect of putative TRAPs on transcription. It is shown that STAR elements 7, 17 and 40 are poor in blocking transcription. On the other hand, certain regions of phage λ as well as a synthetic polyA sequence fulfil the criteria of a TRAP, since they are all potent blockers of transcription . In certain embodiments, the protein expression unit comprises at least two TRAP sequences. It is preferred that said at least two TRAP sequences are arranged such that said TRAP sequences are flanking the combination ormed by said promoter and said open reading frame . Fig 2A shows yet another arrangement. However, when multiple protein expression units are present on one molecule containing genetic information it is also possible to at least partly inhibit or block transcription from one protein expression unit into another protein expression unit. In this case TRAP sequences are located between (possible different) protein expression units. In the situation outlined in Fig 2A a TRAP sequence is placed between the terminator of the bicistronic gene and the SV40 promoter. It will be clear to the person skilled in the art that according to the invention STAR sequences and TRAP sequences may also be combined to improve expression in El-immortalized retina cells. Use of STARS and TRAPs is thought to prevent silencing of transgene expression by combined action of keeping chromatin-associated repression out (STAR elements) and by simultaneously creating domains from which aberrant and harmful transcription is kept out (TRAPs) . Preferably, at least two STAR sequences are arranged such that said STAR sequences are flanking the combination formed by said promoter and said open reading frame (as outlined in Fig 2A) . Even more preferably, said at least two TRAP sequences and said at least two STAR are arranged such that a first 5 ' TRAP sequence is upstream of a first STAR sequence and that a second 3 ' TRAP sequence is downstream of a second STAR sequence. Fig 2 provides a, non-limiting, schematic representation of one of the embodiments of this part of the invention. This is the configuration of the DNA elements of the expression units in the plasmid as well as after integration into the genome. Expression unit one is shown in Fig 2A. It contains an open reading frame for a transgene (a reporter gene, Genel) . This is upstream of the attenuated EMCV IRES (Martinez-Salas et al 1999; Mizuguchi et al 2000; Rees et al 1996), and of the open reading frame encoding the zeocin resistance selectable marker protein (zeo) . The gene has the SV40 transcriptional terminator at it 3^f end (t) . This bicistronic transgene is transcribed at high levels from the CMV promoter. Next to this is the puromycin resistance selectable marker (puro) , transcribed as a monocistronic mRNA from the SV40 promoter. The gene has the SV40 transcriptional terminator at its 3' end (t) . STAR elements flank the expression units. The entire cassette with multiple genes plus STARs is flanked by TRAPs in such an orientation that transcription can be prevented to enter the expression units on the plasmid or a TRAP is orientated such that anti-sense RNA is not or hardly formed. To further improve expression a TRAP sequence and/or a STAR sequence is placed between said bicistronic and said monoσistronic gene. In Fig 2B another configuration of expression units is depicted. The construct consists of two transgenes (two reporter genes or the open reading frames for two subunits of a heterodimeriσ protein (Gene 1 and Gene 2) of which Gene 1 is upstream of the attenuated EMCV IRES and the puromycin resistance protein (puro) and Gene 2 is upstream of the EMCV IRES and the zeocin resistance protein (zeo) . These bicistronic transgenes are transcribed at high levels from the CMV promoter, which are directed in different orientations to prevent transcriptional interference. Both bicistronic genes have the SV40 transcriptional terminator at their 3^r ends (t) . STAR elements flank the expression units. The entire cassette with multiple genes plus STARs is flanked by TRAPs in such an orientation that transcription can be prevented to enter the expression units on the plasmid or orientated such that anti-sense RNA is not or hardly not formed. It is clear to a person skilled in the art that the sequence in which the TRAPs and STARs are placed to flank the expression units can vary. In the given example the STARs are placed between the expression unit and the TRAPs, however it is also possible to place the TRAPs between the expression unit and the STAR element. It is also clear to a person skilled in the art that other selection markers and other combinations of selection markers are possible. Examples of possible antibiotic combinations are provided herein. The one antibiotic that is particularly advantageous is zeocin, because the zeocin-resistance protein (zeocin-R) acts by binding the drug and rendering it harmless . Therefore it is easy to titrate the amount of drug that kills cells with low levels of zeocin-R expression, while allowing the high-expressors to survive. Many other antibiotic- resistance proteins in common use are enzymes, and thus act catalytically (not 1:1 with the drug). When a two- step selection is performed it is therefore advantageous to use an antibiotic resistance protein with this 1:1 binding mode of action. Hence, the antibiotic zeocin is a preferred selection marker. For convenience the zeocin antibiotic is in a two-step selection method combined with puromycin-R or blastiσidin-R in the second bicistronic gene, and puromycin-R or hygromycin-R in the monocistronic gene. It is also clear to a person skilled in the art that different promoters can be used as long as they are functional in the used cell. The CMV promoter is considered the strongest available, so it is preferably chosen for the bicistronic gene in order to obtain the highest possible product yield. Other examples of suitable promoters are e.g. mammalian promoters for EF1- alpha or ubiquitin C promoter. The good expression and stability of the SV40 promoter makes it well suited for expression of the monocistronic gene; enough selection marker protein (for example the antibiotic resistance protein puromycin-R in the example cited herein) is made to confer high expression of said selection marker.

Hence, said SV40 promoter is preferentially used as a promoter driving the expression of the selection marker^'. As outlined above, a method according to the invention can comprise at least two TRAP sequences and at least two STAR sequences. In certain embodiments, said at least two TRAP sequences are essentially identical. In certain embodiments, said at least two STAR sequences are essentially identical. Essentially identical TRAP and/or STAR sequences are defined herein as TRAP and/or STAR sequences which are identical in their important domains (the domains that confer the transcription stabilizing or enhancing quality) , but which may vary within their less important domains, for example a point mutation, deletion or insertion at a less important position within the TRAP and/or STAR sequence. Preferentially said essentially identical TRAP and/or STAR sequences provide equal amounts of transcription stabilizing or enhancing activity. Examples of suitable TRAP and/or STAR sequences are outlined in the experimental part herein. Example 1 provides a method for identifying a TRAP sequence. Such a method may comprise the steps of a) providing a cell with a plasmid that comprises i) a promoter sequence ii) an intervening sequence (IV) downstream of said promoter iii) a putative TRAP sequence located in said IV iv) a sequence whose product is detectable and which sequence is located downstream of said IV b) determining the amount of said detectable product, and c) compare said amount with the amount of product obtained in a cell that is provided with a control plasmid without said putative TRAP sequence . The cloning of the putative TRAP sequence is performed in the intervening sequence to avoid the possibility that addition of a sequence results in enhanced RNA instability. This would also result in a lower RNA signal on a blot, but this would have nothing to do with blocking of transcription. Placing the to be tested sequence in intervening sequences results in the transcription of this sequence into RNA, but it is subsequently spliced out, so a functional, in this particular case codA, mRNA is formed. This happens irrespective whether there was an extra sequence within the intervening sequence or not. Any decline in the mRNA signal is therefore not due to loss of RNA stability, but a direct consequence of transcription termination due to the TRAPs sequence . The method for identifying a TRAnscription Pause (TRAP) sequence may also comprise the steps of a) providing a cell with a plasmid that comprise i) a promoter sequence ii) an intervening sequence (IV) downstream of said promoter iii) a putative TRAP sequence located in said IV iv) a sequence whose product is detectable and which sequence is located downstream of said IV v) said plasmid further comprises a selection marker located outside the combination of said promoter, IV, putative TRAP and said sequence whose product is detectable b) selecting a cell via said selection marker of said plasmid, thereby obtaining a cell that comprises said plasmid, c) determining the amount of said detectable product, and d) compare said amount with the amount of product obtained in a cell that is provided with a control plasmid without said putative TRAP sequenc . The term "selection marker or selectable marker" is typically used to refer to a gene and/or protein whose presence can be detected directly or indirectly in a cell, for example a gene and/or a protein that inactivates a selection agent and protects the host cell from the agent's lethal or growth-inhibitory effects (e.g. an antibiotic resistance gene and/or protein) . Another possibility is that said selection marker induces fluorescence or a color deposit (e.g. green fluorescent protein and derivatives, luciferase, or alkaline phosphatase) . The term "selection agent" is typically defined as a chemical compound that is able to kill or retard the growth of host cells (e.g. an antibiotic) . The term "selection" is typically defined as the process of using a selection marker/selectable marker and a selection agent to identify host cells with specific genetic properties (e.g. that the host cell contains a transgene integrated into its genome) . Preferably, the invention provides a method wherein said sequence whose product is detectable is a suicide gene. A suicide gene is typically defined as a gene which product is capable of killing, either directly or indirectly, a cell. More preferably, said suicide gene is codA or codA::upp. Even more preferably, said detectable product is mRNA. However, it is clear to a person skilled in the art that also a protein can be used as a detectable product. In this case amounts/levels of protein are determined by for example Western blotting or by detecting the protein directly, for example GFP or by performing an enzymatic (colour) reaction based on the properties of the corresponding protein. Use of a suicide gene, for example codA or codA: :upp is particularly advantageous for the screening of a library of sequences . When sequences of a library are cloned in the intervening sequence (IV) , a TRAP sequence is easily identified because the suicide gene is not transcribed and translated and hence the lethal product of said suicide gene is not produced and the cell that comprises said TRAP sequence survives . When the cloned sequence is not a TRAP sequence, the cell dies because of the lethal formed product. It is clear that different types of suicide genes can be used. The codA gene encodes the enzyme cytosine deaminase which enzyme converts cytosine to uracil . CodA can be used as a metabolic suicide gene in combination with the prodrug 5- fluorocytosine . The enzyme is able to convert the non- toxic prodrug into 5-fluorouracil-mono phosphate which kills the cells by disrupting DNA synthesis, thereby triggering apoptosis. CodA: :upp is a fusion between a cytosine deaminase gene and an uracil phosphoribosyl transferease gene. Both enzymes act synergistically to convert 5-fluorocytosine into fluoro racil-mono phosphate, a toxic compound. The fusion of the genes leads to a more efficient system. Another example of a suicide gene and a non-toxic prodrug is thymidine kinase and ganciclovir. However, it is clear that it is also possible to use a suicide gene which is not dependent on the presence of a prodrug. Hence, a method for identifying a TRAP sequence may also comprise the steps of a) providing a cell with a plasmid that comprises i) a promoter sequence ii) an intervening sequence (IV) downstream of said promoter iii) a putative TRAP sequence located in said IV iv) a sequence encoding a suicide product and which sequence is located downstream of said IV b) determining whether said cell survives. Preferably, said putative TRAP sequence is derived from a library. In yet another preferred embodiment, a prodrug is used that is converted into a toxic compound by the product of said suicide gene. In example 1, the codA: :upρ open reading frame is used as a sequence whose product is detectable and the amount of RNA is determined. Examples of said TRAP sequences are outlined in the experimental part and in Table 1 (SEQ. ID. NOs . 8-15), but may furter be identified and/or tested by methods such as described above. In certain embodiments, said TRAP sequence comprises the lambda 35711-38103 sequence (SEQ. ID. NO. 8), or a functional fragment or derivative thereof. In another embodiment, said TRAP sequence comprises a synthetic polyA (SPA) sequence or a functional fragment or derivative thereof, such as for example provided by SEQ. ID. NO. 9 or SEQ. ID. NO. 14 (these two SEQ. IDs. provide two alternative versions of an SPA sequence, which work equally well) . In yet another embodiment, said TRAP sequence comprises a combination of an SPA and the human μ2 globin gene pause signal or a functional fragment or derivative thereof, for example a combination of a SPA and the human μ2 globin gene pause signal as provided by SEQ. ID. NO. 11 or SEQ. ID. NO. 15 (these two sequences comprise the two alternative versions of the SPA sequence, that work equally well) . The person skilled in the art will be aware that the sequences according to the invention can be obtained by various methods, including the cloning from the human genome or from the genome of another organism, or by for instance amplifying sequences directly from such a genome by using the knowledge of the sequences, e.g. by PCR, or can in part or wholly be synthesised. A functional fragment and/or derivative of a sequence described by the SEQ. ID. NOs. provided in Table 1 or 2 is a sequence derived with the information given by the SEQ. ID. NOs. in Table 1 or 2. For instance, a sequence that can be derived from such a sequence by deleting, modifying and/or inserting bases in or from a sequence mentioned in Table 1 or 2, wherein said derived sequence comprises the same activity in kind, not necessarily in amount, of a sequence as provided by the SEQ. ID. NOs. in Table 1 or 2. A functional fragment or derivative is further a sequence comprising a part from two or more sequence provided by the SEQ. ID. NOs. in Table 1 or 2, while still having the same activity in kind, not necessarily in amount. A functional fragment of a sequence mentioned in Table 1 or 2 can for example be obtained be deletions from the 5^r end or the 3' end or from inside of said sequences or any combination thereof, wherein said derived sequence comprises the same activity in kind, not necessarily in amount. A functional fragment or derivative also comprises orthologs from other species, which can be found using the known sequences by methods known by the person skilled in the art, and tested for activity, which should be the same in kind, not necessarily in amount. A TRAP sequence suitable for the invention may for instance be selected from the sequences mentioned in Table 1. Preferably said TRAP sequence is combined with a STAR sequence, such as those mentioned in Table 2. Use according to the invention is particular advantageous when applied to expression of at least one protein of interest . In a preferred embodiment a protein expression unit according to the invention is provided, wherein said protein of interest is an immunoglobulin heavy chain. In yet another preferred embodiment a protein expression unit according to the invention is provided, wherein said protein of interest is an immunoglobulin light chain. When these two protein expression units are present within the same (host) cell a multimeric protein and more specifically an antibody is assembled.

Openers The present invention provides means and methods for improving characteristics of protein production in a cell . It has among others been found that chromatin modification systems for rendering chromatin more accessible for transcription have a pronounced effect on expression characteristics of protein expression when allowed to act thereon. In one embodiment the invention therefore provides a method for providing a cell with a protein expression unit comprising providing a nucleic acid comprising said unit with a nucleic acid encoding a binding site for a member of a chromatin modification system for rendering chromatin more accessible for transcription (opener) , wherein said opener is present in said cell, said method further comprising providing said expression unit to said cell and culturing said cell to allow expression of said protein expression unit. Histone modification systems have been shown to encompass proteins capable of rendering chromatin more accessible for transcription. In certain embodiments, an opener of the invention is therefore a histone modification enzyme, preferably capable of modifying a N- ter inal histone tail. Histone modification plays an important role in both chromatin-associated repression and chromatin-associated activation of gene expression. For instance, acetylation of specific lysines in histone H3 and H4 tails is an important parameter. Normally histones are very basic proteins that bind tightly to the acid DNA strands. Addition of an acetyl group to the histone tails converts the basic histones into more neutrally charged proteins. This results in a less tight interaction between the basic histones and the acid DNA strands . Acetylation is therefore associated with making the chromatin more open or accessible for transcription factors. Histone acetyltransferases (HATs) that add acetyl groups to the histone tails are therefore openers according to the present invention. Embodiments of HAT openers are p300/CBP, P/CAF (Yang et al 1996), Gcn5 (Brownell et al, 1996) protein and/or CBP (Bannister and Kouzarides 1996) or a functional part or derivative thereof. However, even today more HAT proteins, comprising similar function are identified. Such HAT proteins can of course also be used as an opener according to the invention. Also specific methylated histone tails have activity in opening chromatin according to the invention. Some types of methylation are associated with rendering chromatin more accessible for transcription, whereas other types of methylation are associated with rendering chromatin less accessible. The Ashl protein (Nakamura et al 2000) is a trithorax group protein that acts as a positive regulator of gene expression. Ashl has methyltransferase activity and adds a methyl group to at least lysine K4 of histone H3 (Beisel et al 2002) . In certain embodiments, a methyltransferase capable of adding a methyl group to at least lysine K4 of histone H3 is thus an opener according to the invention. Preferably, said opener comprises Ashl protein or a functional part or derivative thereof. The different histone methyltransferases have a structural protein motif, the SET domain in common. The SET domain (for Su(var)39, E(z) and trx, the three proteins in which the domain was first identified) is essential for histone methyltransferase activity to take place. It follows that targeting an activating histone methyltransferase or its functional part, the SET domain, can have a beneficial effect on gene expression by interfering at the level of chromatin structure . Besides the acetylation and methylation of histone tails also phosphate groups and ubiquitin groups can be added. Also these events can influence the order in which either acetylation or methylation of histone tails can take place. Collectively this complex interplay between histone modifications is referred to as the ^Λhistone code" that is considered as the most fundamental mechanistic explanation for both repressing and activating epigenetic gene regulation mechanisms. Hence, in certain embodiments an opener according to the invention may be a histone phopsphorylating or ubiquinating enzyme, when this can render chromatin more accessible to transcription. Other embodiments of openers according to the invention are comprised in the class of σhromatin- remodelling proteins such as Tritorax group (TrxG) proteins, CHRAC proteins, ACF group proteins, and a NURF group protein. Specific TrxG proteins are trithorax (Mazo et al, 1990; Petruk et al, 2001), trithorax-like (Farkas et al. 1994), Brahma (Tamku et al 1992), ISWI (Elfring et al, 1994), Ashl (Nakamura et al, 2000; Beisel et al, 2002) , moira (Crosby et al, 1999) , and osa (Treisman et al, 1997) . One TrxG protein is Brahma (Tamkun et al 1992), which protein is part of a multimeric protein complex that operates as a so-called chromatin-remodelling complex. Chromatin-remodelling has been defined as the

ATPase-dependent disruption of nucleosomes to facilitate binding of transcription factors to the chromatin. The chromatin becomes more open or accessible for transcription factors and thus transcription. Other chromatin-remodelling complexes have been defined, such as CHRAC (Varga-Weisz et al; 1997) and NURF (Tsukiyama and Wu 1995) . A comprehensive overview is given by Fyodorov and Kadonaga, 2001. Also these complexes operate in an ATPase dependent fashion. Thus in this embodiment the opener comprises a chromatin-remodelling protein and preferably the Trithorax group protein Brahma, a CHRAC group protein, a NURF group protein, ACF group proteins, (for ATP-utilizing chromatin assembly and remodeling factor) (Ito et al 1997) or a functional part or derivative thereof. Purified ACF fractions contain Imitation SWI (ISWI) protein (Elfring et al, 1994) . Three other proteins co-purify with this complex termed p47, pl70 and pl85 referring to their apparent molecular weight . In certain embodiments said chromatin-remodelling protein comprises an ISWI protein or a Brahma protein or a functional part or derivative thereof. Trithorax group proteins have miscellaneous effects on chromatin, however, at least some proteins of the group are capable of rendering chromatin more accessible to transcription factors. Thus in certain embodiments said opener comprises a protein of the trithorax group and may comprises an ISWI protein or a trithorax protein, a trithorax-like protein, a Brahma protein, an Ash protein, a oira protein, an osa protein or a functional part or derivative thereof.

A functional part or derivative of an opener of the invention comprises the same activity in kind not necessarily in amount as an opener mentioned. This activity being a sequence specific nucleic acid binding activity specific for said binding site and a chromatin modification activity rendering chromatin more accessible for transcription. Fragments or derivatives may be tested in a method of the invention for functionality as an opener. Often parts of a protein can be identified that can be manipulated to at least some extent without affecting the kind of function of the protein. Such openers comprising such modifications are of course within the present invention. For openers that comprise the so-called SET domain, the functional part typically comprises this SET domain. Derivatives may be generated by for instance conservative a ino acid substitutions. These typically retain the same function in kind.

Suitable parts may be generated by mutation, deletion and/or insertions of the opener. Obviously, homologs or orthologs from other species (e.g. human, mouse, rat, hamster, etc) may also be used, and are regarded as derivatives according to the invention. Hence suitable analogues may be found in other than the mentioned species . Such analogues can for instance be selected by amino acid and/or nucleic acid homology. For instance ISWI2 has in human the homologues BRG1 and hbrm. ISWI2 is homologues to Brahma, whereas BAF170 and BAF155 are SWI3 homologues. Another non-limiting example of suitable homologues are BAF170, BAF155 and SWI3 which are homologues of moira. Such homologues are of course also part of the invention.

As described above, many different proteins can act as openers in the present invention. The opener may act directly on the accessibility of chromatin or indirectly via the association with a complex present in the cell, wherein the complex is instrumental in the accessibility of the chromatin. An essential component of the opener of the present invention is the sequence specific association thereof with the binding site on the nucleic acid comprising the protein expression unit. The binding site may be a normal binding site for an opener. Alternatively, a binding specificity for said binding site is provided to an otherwise operable opener. In yet another embodiment a sequence specific nucleic acid binding specificity for said binding site is provided to a protein thereby resulting in an opener of the present invention. It is possible that proteins that when provided with a binding specificity for said binding site do not have a sequence specific binding specificity by themselves (i.e. prior to being provided with such specificity) . Such proteins (further termed pre-openers) can be provided to the cell to achieve a generalized effect on chromatin re-modelling. This is another aspect of the present invention. The invention thus provides a cell comprising a protein expression unit, wherein said cell is provided with a pre-opener of the present invention. Such cells can through the generalized effect on chromatin re-modelling display favourable expression characteristics. This can for instance be due to a shift in the balance between activating and repressing complexes .

An opener may be expressed by the cell prior to providing the cell with the protein expression construct, for instance in case the cell naturally expresses said opener. Alternatively, the opener may be provided to the cell, for instance as a nucleic acid encoding the opener. When openers are used that have been provided with a specific nucleic acid binding activity toward the binding site, it is often appropriate to provide the cell with the opener. However, cell lines may be created already expressing such opener. Such cell lines can then subsequently be used to introduce protein expression unit of the invention at will. Cell lines provided with a nucleic acid comprising an opener provided with a new sequence specific binding activity are therefore also part of the invention. Preferred openers for such cell lines comprises HAT proteins provided with a new sequence specific binding activity. Preferably said HAT proteins comprise p300/CBP protein, a P/CAF protein, a Gcn5 protein and/or a CBP protein or a functional part, derivative and/or analogue thereof. The new sequence specific binding activity preferably comprises a nucleic acid binding domain of a sequence specific DNA binding protein. Non-limiting examples are the GAL4 or the LexA DNA binding domains. However, many other sequence specific binding proteins can be used. A person skilled in the art can use DNA binding domains of a large number of different proteins and generate an opener of the invention. The many examples of fusions of DNA binding domains to other functional proteins may be taken for guidance. It is for instance entirely possible to modify two-hybrid systems such that upon association of the two parts of the hybrid system, an opener of the present invention is generated. In a preferred embodiment said opener is a fusion protein comprising at least a functional part of a mentioned opener, and a sequence specific nucleic acid binding domain. Preferably the opener comprises at least a functional part of a histone- acetyltransferase, a histone methyltransferase or a chromatin-remodelling protein. Preferably, said histone- acetyltransferase comprises a p300/CBP protein, a P/CAF protein, a Gcn5 protein, or a CBP protein or a functional part, derivative and/or analogue thereof. Preferably, said histone methyltransferase comprises an Ashl protein or a functional part, derivative and/or analogue thereof. Said chromatin-remodelling protein preferably comprises a trithorax group protein, a CHRAC group protein, a NURF group protein, ACF group proteins or a functional part, derivative and/or analogue thereof. The mentioned openers may be fused to the DNA specific binding domain of a zinc-finger protein, a bacterial DNA binding protein, a yeast or fungus DNA binding protein. Preferably, said DNA binding protein is LexA or Gal4 or functional part, derivative and/or analogue thereof. Besides protection against chromatin-silencing by means of STAR elements and/or TRAP sequences, the present invention in addition creates means and methods to convert the chromatin of a transgene in a more open state, thus further facilitating the predictability, yield and stability of transgenic protein expression. To achieve this goal, chromatin-remodelling proteins, histone acetyltransferase or histone methyltransferase proteins, herein collectively referred to as 'openers', can be targeted to the promoter of the transgene. The invention thus prevents silencing of transgene expression by the combined action of keeping repression out and by simultaneously keeping chromatin in an open state . By combining STAR elements and/or TRAP sequences with chromatin opening factors the present invention employs two or more different types of DNA elements or proteins that synergistically reinforce each other to create novel (host) cells/cell lines that efficiently and stably express proteins. The invention further comprises the use of an opener for stabilizing expression of an expression unit and the use of an opener for increasing the number of clones expressing a certain amount of protein after genetic modification. Also provided is the use of an opener for increasing transcript levels produced by an expression unit. The targeting of chromatin openers to a transgene or a promoter of a transgene is used to achieve predictable, high yields and stable transcription of a transgene. In the present invention HAT proteins such as p300, CBP, a Gcn5 protein, and/or P/CAF, HMTase proteins such as Ashl or the Brahma protein or functional relevant parts of these proteins are produced as fusion protein with the LexA protein (Bunker and Kingston 1994) . These fusion proteins are placed under control of an inducible or constitutive promoter such as the SV40 promoter (Fig. 8) . The expression unit for these fusion proteins are present on the same plasmid as the expression unit that contains the gene that encode the protein of interest (Gene 1) (Fig 8) . Gene 1 is placed under control of the CMV promoter. Upstream of the CMV promoter binding sites are cloned to which the LexA-HAT, LexA-HMTase or LexA-Brahma proteins are targeted (Fig 8A) . Thus these fusion proteins are targeted to the vicinity of the promoter to keep open the chromatin structure of the promoter in order to facilitate the assessability of the promoter for transcription factors. It is also possible to create one plasmid containing three expression units that encode respectively Gene 1, Gene 2 and LexA-HAT, LexA-HMTase or LexA-Brahma (Fig 8B) . The expression units encoding Gene 1 and Gene 2 are oriented divergent in such a manner that the two CMV promoters are adjacent although differently oriented. Between the two promoters LexA binding sites are placed to which the LexA-fusion protein is targeted. In this manner chromatin openers are targeted to both expression units.

It will also be clear to a person skilled in the art that it is not essential that the LexA fusion proteins or the Brahma protein are expressed from the same plasmid that contains the expression unit with the gene of interest. The LexA fusion proteins or Brahma protein can also be produced from a separate plasmid. Therefore, the invention provides in one embodiment, a method for obtaining a cell according to the invention which expresses one or more proteins comprising providing said cell with one or more protein expression units encoding said one or more proteins, characterised in that at least one but preferably at least two of said protein expression units comprises at least one chromatin opener and/or one STAR sequence and/or one TRAP sequence. In certain embodiments Gene 1 and Gene 2 encode the light and heavy chain of a multimeric immunoglobin protein. Chromatin openers or simply openers (the terms are used interchangeably herein) are involved in opening chromatin structure, through chromatin-remodelling proteins and their complexes such as the Ashl protein, the Brahma protein, other trxG proteins or components of the CHRAC NURF and ACF group chromatin-remodeling complexes. Alternatively, chromatin openers are histone modifiers such as HAT proteins or functional relevant parts of such proteins that are still able to add acetyl groups to histone tails which has the consequence that the tight association between the basic histones and the acid DNA is loosened. Yet another class of Chromatin openers consists of specific histone methyltransferase such as the Ashl protein that add a methyl group to at least lysine 4 (K4) of histone H3 and that also results in opening of chromatin or making it more accessible to the general transcription machinery. Chromatin openers, these being chromatin-remodelling factors, specific HATs or HMTases or even other histone modifiers thus have in common that they facilitate the binding of transcription factors to the promoter and hence increase the possibilities for transcription. Fig 8 provides a, non-limiting, schematic representation of one of the embodiments of this part of the invention .

This is the configuration of the DNA elements of the expression units in the plasmid as well as after integration into the genome. Expression unit one is shown in Fig 8A. It contains an open reading frame for a transgene (a reporter gene, Genel) . This is upstream of the attenuated EMCV IRES (Martinez-Salas et al 1999; Mizuguchi et al 2000; Rees et al 1996), and of the open reading frame encoding the zeocin resistance selectable marker protein (zeo) . The gene cassette has the SV40 transcriptional terminator at their 3' ends (t) . This bicistronic transgene is transcribed at high levels from the CMV promoter. Upstream of the CMV promoter are four LexA binding sites (LexA-BS) . Next to this is the monocistronic gene encoding a fusion protein between the LexA protein and a histone acetyltransferase (HAT) or a functional part of a HAT that is still able to transfer acetyl groups to histone tails (LexA-HAT) . Alternatively a fusion protein is encoded between the LexA protein and a histone methyltransferase (HMTase) or a functional part of a HMTase protein that is still able to transfer a methyl group to at least lysine K4 of histone H3. Alternatively a fusion protein is encoded between LexA and the Brahma protein. Either one of these monocistronic transcription units is transcribed from the SV40 promoter. The genes have the SV40 transcriptional terminator at their 3' ends (t) . This entire cassette with multiple genes is flanked by STAR elements.

Fig 8B is similar as Fig 8A, but one plasmid contains now three expression units that encode respectively Gene 1, Gene 2 and LexA-HAT, LexA-HMTase or LexA-Brahma. The expression units encoding Gene 1 and Gene 2 are oriented divergent in such a manner that the two CMV promoters are adjacent although differently oriented. Between the two promoters LexA binding sites are placed to which the LexA fusion proteins are targeted. In this manner chromatin openers are targeted to both expression units.

It is clear to a person skilled in the art that in these examples more possible combinations can be made. For instance the expression units can be made in such a manner that Genel, Gene 2 and the LexA-HAT, LexA-HMTase, LexA-Brahma or Brahma each are located on separate plasmids. Also STAR elements can be omitted from these constructs and still the expression of genel can be benefited form the presence of chromatin openers .

The invention is further explained in the following examples . The examples do not limit the invention in any way. They merely serve to clarify the invention.

Examples

Example 1: Identification of TRAnscription Pause (TRAP) sequence One object of this invention is to identify DNA elements that act as TRAnscription Pause (TRAP) sequence. In this example a method is provided for identifying TRAP sequences .

Materials and Methods

Plasmids

For constructing pIRES-6, pIRES (Clontech) was taken as a starting plasmid. pIRES was cut with Bglll and Dral, and the CMV promoter, intervening sequence (IV) , IRES and SV40 polyadenylation signal ligated into pBSKS (Stratagene) , cut with BamHI and EcoRV to create pIRES-1. Oligos STOP 1 (CTAGCTAAGTAAGTAAGCTTGG) and STOP 2 (AATTCCAAGCTTACTTA CTTAG) were ligated into Nhel-Xhol cut pIRESl, creating pIRES-2. This results in three stop codons, in three different reading frames, in front of the IRES. Oligos BamHI-Bςrlll-Asσl (TTAAGGATCCAGATCTGGCGCGCC) and Ascl-BglII-BamHI (TTAAGGCGCGCCAGATCTGGATCC) were annealed and ligated into Bsal cut pIRES-2, creating pIRES-3. In this way, Ascl, BamHI and Bgi.II sites were ceated in the intervening sequences. pORFCODA: :UPP (InvitroGen) was cut with ΪVcoI- Nheϊ , filled in with Klenow, and ligated 3' of the IRES, into the Sraal of pIRES-3, creating pIRES-4. The cloning was performed in the intervening sequence to avoid the possibility that addition of a sequence results in enhanced RNA instability. This would also result in a lower RNA signal on a blot, but this would have nothing to do with blocking of transcription. Placing the to be tested sequence in intervening sequences results in the transcription of this sequence into RNA, but it is subsequently spliced out, so a functional codA mRNA is formed. This happens irrespective whether there was an extra sequence within the intervening sequence or not. Any decline in the mRNA signal is therefore not due to loss of RNA stability, but a direct consequence of transcription termination due to the TRAPs sequence. Next, pIRES-4 was cut with Xhol-Xba I, filled in with Klenow, and ligated into Sma I cut pBSKS, creating pIRES-5. pIRES 5 was cut with Sail, and ligated into pPURO partially digested with Sail, creating pIRES-6. Thus the whole cassette consists of the CMV promoter, the IV, IRES, codA: : upp and SV40 polyadenylation signal in a pREP4 (Invitrogen) backbone. In this backbone a hygromycin resistance gene is present, to allow selection of transformants on hygromycin. To create pIRES-31, the IRES and codA: : upp in pIRES- 6 was replaced by codA: : upp only. Oligos WotI-BelI-EV (GGCCGCTGATCAGATATCGCGG) and JVΛel-EcoRV-BcII (CTAGCCGCGATATCTGATCAGC) were annealed and ligated to

Notl-Nhel digested pIRES 6, which releases the IRES and codA: : upp . This created pIRES-30. The CodA: :upp ORF as a BamHI fragment was then ligated into Bell cut pIRES-30, creating pIRES-31 (Fig 3) .

Transfection and culture of U2- OS cells

Transfection and culture of U-2 OS cells with pIRES-6 and pIRES-31 plasmids: The human osteosarcoma U-2 OS cell line (ATCC #HTB-96) was cultured in Dulbecco ' s Modified Eagle Medium + 10% Fetal Calf Serum containing glutamine, penicillin, and streptomycin (supra) at 37°C/5% C02. Cells were transfected with the pIRES-6 and -31 vector containing putative TRAPs in MCSI using SuperFeσt. Hygromycin selection was complete in 2 weeks, after which time a pool of hygromycin resistant U-2 OS clones were isolated and RNA was isolated using conventional protocols (Sambrook et al 1989) .

Results

Several constructs were transfeσted to U-2 OS cells 1) The empty control vector as shown in FIG 3. 2) A 2400 bp long DNA of phage μ (bp 35711-38103) in 5' -3' orientation 3) A 2400 bp long DNA of phage μ (bp 35711-38103) in 3' -5' orientation 4) The 60 bp long MAZ DNA sequence 5) STAR7 6) STAR 40 7) The empty control vector as shown in FIG 3 8) A synthetic poly A (SPA) sequence in 5' -3' orientation 9) A combination of the SPA sequence and a 92 bp long μ2 globin gene pause signal in orientation 5' -3' 10) A synthetic poly A sequence in 3' -5' orientation 11) A combination of the SPA sequence and a 92 bp long μ2 globin gene pause signal in 3' -5' orientation

After transfection selection was performed by hygromycin. After three weeks the entire pool of cells was harvested and mRNA was isolated. The entire pool of cells was used and no individual colonies since the Fig 3 vector replicates episomally which prevents position effects that would occur when the vectors stably integrate .

As shown in FIG 4, the lambda (bp 35711-37230) (lane 2) fragment efficiently blocked transcription of the codA gene driven by the CMV promoter, as compared to the empty control (lane 1 and 7) . Also a synthetic polyA (SPA) sequence (Levitt et al 1989) (AATAAAAGATCCTTATTTTCAC TAGTTCTGTGTGTTGGTTTTTTGTGTG) (SEQ. ID. NO. 9, we also have tested an alternative sequence which worked equally well (SEQ. ID. NO. 14)) either alone (lane 8) or in combination with a 92 bp pausing signal from the human μ2 globin gene (AACATACGCTCTCCATCAAAACAAAACGAAACAAAACAAACTAGCAAAATAGGCTG TCCCCAGTGCAAGTGCAGGTGCCAGAACATTTCTCT) (Enriquez-Harris et al 1991) (lane 9) potently blocked transcription of the codA gene driven by the CMV promoter. The hygromycin resistance gene was used as internal control, indicating the number of copies of the plasmids. The 60 bp long MAZ sequence has been reported to be a powerful transcription blocker (Ashfield et al 1994) . However, in our test system, the MAZ sequence (lane 4) did not prevent transcription. Also STAR 7 (lane 5) and STAR 40 (lane 6) did not prevent transcription of the codA gene driven by the CMV promoter. Quantification of the signals using a phosphoimager showed that the phage lambda (bp 35711- 38103) , the SPA sequence and the SPA/Pause combination blocked 95% of the CMV promoter driven transcription. We conclude that the phage lambda (bp 35711-38103) fragment and the SPA, SPA/Pause sequences (Table 1) serve as TRAP.

Example 2 : TRAP sequences block transcription in a directional fashion

Materials and Methods The experiments of Example 1 are referred to .

Results

As shown in FIG 4, the orientation of the TRAP is an essential parameter in the action of TRAP sequences. The phage lambda (bp 35711-38103) served only as TRAP in one, 5' -3' orientation (lane 2). When tested in the opposite, 3' -5' orientation (lane 3) no blocking of CMV driven transcription was found at all. Similarly, the SPA and the combined SPA/Pause sequence blocked transcription only in the 5' -3' orientation (lanes 8 and 9) and not in the 3' -5' orientation (lanes 10 and 11). The orientation dependency of TRAP sequences is of importance for the orientation in which they can be used when flanking transgenes .

Example 3: TRAPs improve the effects of STAR elements on the expression level of transgenes One object of this invention is to improve transgene expression for heterologous protein production, thus increasing the yield of the heterologous protein.

Materials and Methods

Plasmids The construction of the pPlug&Play-d2EGFP-ires-Zeo (PP) vector is described below. Plasmid pd2EGFP (Clontech 6010-1) was modified by insertion of a linker at the BsiWI site to yield pd2EGFP-link. The linker (made by annealing oligonucleotides GTACGGATATCAGATCTTTAATTAAG and GTACCTTAATT AAAGATCTGATAT) introduced sites for the Pad, Bgrlll, and EσoRV restriction endonucleases . This created the multiple cloning site MCSII for insertion of STAR elements. Then primers (GATCAGATCTGGCGCGCCATTTAAATCGTC TCGCGCGTTTCGGTGATGACGG) and (AGGCGGATCCGAATGTATTTAGA AAAATAAACAAATAGGGG) were used to amplify a region of 0.37 kb from pd2EGFP, which was inserted into the Bg ll site of pIRES (Clontech 6028-1) to yield pIRES-stuf . This introduced sites for the Asσl and Swal restriction endonucleases at MCSI, and acts as a "stuffer fragment" to avoid potential interference between STAR elements and adjacent promoters. pIRES-stuf was digested with Bglll and Fspl to liberate a DNA fragment composed of the stuffer fragment, the CMV promoter, the IRES element (flanked by multiple cloning sites MCS A and MCS B) , and the SV40 polyadenylation signal. This fragment was ligated with the vector backbone of pd2EGFP-link produced by digestion with BaΛiHI and Stul, to yield pd2IRES-link. The open reading frames of the zeocin-resistance genes was inserted into the BamHI/IvOtl sites of MCS B in pd2lRES-link as follows: the zeocin-resistance ORF was amplified by PCR with primers (GATCGGATCCTTC GAAATGGCCAAGTTGACCAGTGC) and (AGGCGCGGCCGCAATTCTCAG TCCTGCTCCTC) from plasmid pEM7/zeo (Invitrogen) , digested with BamHI and NotX, and ligated with BamHI/iVotl-digested pd2IRES-link to yield pd2IRES-link-zeo. The SEAP reporter ORF was introduced into pd2IRES- link-zeo by PCR amplification of pSEAP2-basic with primers (GATCGAATTCTCGCGACTTCG CCCACCATGC) and (AGGCGAATTCACCGGTGTTTAAACTCATGTCTGCTC GAAGCGGCCGG) , and insertion of the SσoRl-digested SEAP cassette into the EcoRl sites in MCS A of the plasmids pd2IRES-link-zeo (to yield plasmid PP2) . PP2 was cut with BσoRI and Mluϊ to remove the SEAP gene and p2EGFP was introduced with primers (GATCGAATTCATGGTGAGCAAGGGCGAGGAG) and ( GGCACGCGTGTTAACCTACACATTGATCCTAGCAGAAGC) . Ascl STAR fragments were cloned in to the Ascl site of MCS I of ppd2EGFP. A 2.4 kb lambda DNA fragment (TRAP) was amplified using primers (GATCATTTAAATGT CGACCTGAATTGCTATGTTTAGTGAGTTG) and (GATCGTCGACGTTTGG

CTGATCGGC) , and cloned as a Sail fragment in MCS I, 5' to the STAR. STAR and TRAP were then amplified using primers (GATCTTAATTAACCAAGCTTGCATGCCTGCAG) and (AGGCGATATCGCG CGAGACGATTTAAATGG) , cut with .EcoRV and Pad , and ligated into the same vector, cut with EcoRV and Pad, from which they were amplified.

Transfection and culture of CHO cells The Chinese Hamster Ovary cell line CHO-Kl (ATCC CCL-61) was cultured in HAMS-F12 medium + 10% Fetal Calf Serum containing 2 mM glutamine, 100 U/ml penicillin, and 100 micrograms/ml streptomcyin at 37°C/5% CO2. Cells were transfected with the indicated plasmids using SuperFect (QIAGEN) as described by the manufacturer. Briefly, cells were seeded to culture vessels and grown overnight to 70- 90% confluence. SuperFect reagent was combined with plasmid DNA at a ratio of 6 microliters per microgram (e.g. for a 10 cm Petri dish, 20 micrograms DNA and 120 microliters SuperFect) and added to the cells. After another overnight incubation zeocin was added to a concentration of 50 μg/ml and the cells were cultured further. After another three days the medium was replaced by fresh medium containing zeocin (100 μg/ml) and cultured further. When individual colonies become visible (approximately ten days after transfection) medium was removed and replaced with fresh medium without zeocin. Individual clones were isolated and transferred to 24- well plates in medium without zeocin. One day after isolation of the colonies zeocin was added to the medium. Expression of the GFP reporter gene was assessed approximately 3 weeks after transfection.

The tested constructs basically consist of a bicistronic gene with the GFP gene, an IRES and the

Zeocin resistance gene under control of the CMV promoter and a monocistronic gene encoding the puromycin resistance gene under control of the SV40 promoter (FIG. 2A) . Diversity in the constructs was created by the addition of the 2400 bp lambda (bp 35711-38103) TRAP (SEQ. ID. NO. 8) to the 5' and 3' ends, STAR 40 (SEQ. ID. NO. 7) to the 5' and 3' ends or the combination of STAR 40 and the lambda (bp 35711-38103) TRAP to the 5' and 3' ^' ends. The constructs were transfected to CHO-Kl cells. Stable colonies were expanded before the GFP signal was determined on a XL-MCL Beckman Coulter flowcytometer. The mean of the GFP signal was taken as measure for the level of GFP expression and this is plotted in Fig. 5.

Results Fig. 5 shows that flanking the entire GFP-IRES-Zeo construct (Fig 2A) with the lambda (bp 35711-38103) TRAP in the 5' -3' orientation (see Fig 4) did not result in stable CHO colonies that express significantly higher levels of GFP protein, as compared to the "empty" control without the TRAP sequences (Control) . However, flanking the entire cassette with the combined lambda (bp 35711- 38103) TRAP (5' -3' orientation) and STAR 40 resulted in significantly higher GFP signals (approximately 200%) as compared to the highest GFP signals that were obtained with a construct that is flanked by STAR 40 elements alone . It was therefore concluded that the lambda (bp 35711-38103) TRAP potentiates the ability of STAR elements to convey higher expression levels to a transgene.

Example : The in luence of a TRAP on protein yield is orientation-dependent. Materials and Methods

The experiments of Example 3 are ref rred to .

Results

As shown in Fig 6, the orientation of the TRAP is an essential parameter in the action of TRAP sequences . The phage lambda (bp 35711-37230) TRAP served only as TRAP in the 5' -3' orientation (FIG 3, lane 2). When tested in the 3' -5' orientation, which did not convey transcription blocking (FIG 3, lane 3) , also no effect on the expression levels of the GFP protein was observed. No effect of the lambda sequence itself was observed and no effect when combined with STAR 40 (FIG 6) . The orientation dependency of TRAP sequences is of importance for the orientation in which they can be used when flanking transgenes .

Example 5: The stability of transgene expression is improved by TRAPs During cultivation of recombinant host cells, it is common practice to maintain antibiotic selection. This is intended to prevent transcriptional silencing of the transgene, or loss of the transgene from the genome by processes such as recombination. However it is undesirable for production of proteins, for a number of reasons. First, the antibiotics that are used are quite expensive, and contribute significantly to the unit cost of the product. Second, for biopharmaceutical use, the protein must be demonstrably pure, with no traces of the antibiotic in the product. One advantage of STARs and TRAPs for heterologous protein production is that they confer stable expression on transgenes during prolonged cultivation, even in the absence of antibiotic selection; this "property is demonstrated in this example.

Materials and Methods GFP expression levels in the colonies that were described in Example 3 were measured after periods of one week. After the initial three weeks after transfection when the first GFP measurements were performed, the colonies were cultured in medium without zeocin or other antibiotics . This continued for the remainder of the experiment.

Results

FIG 7 shows the data on GFP expression of colonies that were stably transfected with the GFP construct that is flanked by the combined lambda {bp 35711-38103) TRAP in the 5' -3' orientation and STAR 40. The colonies with the highest GFP expression levels in Fig 5 were chosen for analysis of stability of expression over time in the absence of selection pressure by antibiotics. The expression of the reporter GFP protein remained stable in the CHO cells in three time points. The first time point represents the start of the experiment when the selection pressure was removed. Measurements were performed after one, two and three weeks, which signifies approximately 10, 20 and 30 cell cycles respectively. Colonies containing the combined Lambda TRAP and STAR 40 were stable in the absence of antibiotics. This demonstrates that application of the ability of a combination of TRAPs and STAR elements protected transgenes from silencing during prolonged cultivation. It also demonstrates that this property was independent of antibiotic selection.

Example 6 : Expression of LexA-HAT and LexA-Brahma and Brahma proteins in CHO cells One object of this invention is to apply chromatin openers to improve the predictability, yield and stability of transgenes in mammalian cell lines. Here we introduce several chromatin openers into CHO cells and we describe the construction of the various opener constructs . Materials and Methods Plasmids Plasmid PP2 was described in example 3. PP2 was cut with EcoRI and Mlul to remove the SEAP gene and p2EGFP was introduced with primers (GATCGAATTCATGGTGAGCAAGGG

CGAGGAG) and (AGGCACGCGTGTTAACCTACACATTGATCCTAGCA GAAGC) . This vector was used as a basis vector to construct PP-LexA (PPL) , PP-LexA-Brm (PPLBrm) , PP-LexA-PCAF (PPLPCAF) , PP-LexA-p300HAT (PPLp300) and PP-LexA- AshlHMTase (PPLHuAshl) .

Br coding sequence was pcr-amplified from plasmid pSVhSNF-μ (Chiba et al 1994) using primers Brm-alF-H3- Agel (GATCAAGCTTACCGG TATGTCCACGCCCACAGACCCTGGTG C) and Brm-al572R-XbaI (AGGCTCTAGAATCACTCATCATCCGTCCCACTT CCTTC) and cloned into pPur (BD biosciences #6156-1) using Hindlll and Xbal to create pPur-Brm. LexA binding sites (LBS) were amplified from plasmid pREP4-HSF-Luc+ (van der Vlag et al, 2000) using primers LBS-for-Sall (AGGCGTCGACGTTTCG ACTCCCAAGCTTTG) and LBS-rev-AscI (GATCGGCGCGCCGGTACC ATAGCGGCCGCGAC) and cloned upstream of the CMV promoter in PP using Sail and Ascl to create PPLbs . LexA was amplified from plasmid pEG202 (Bennetzen and Hall, 1982) using primers LexA-for-H3 (GATCAAGCTTA TGAAGACGTTAACGGCCAGGC) and LexA-rev-Agel (AGGCACCGGTCAG

CCAGTCGCCGTTGCGAATAACC) and cloned downstream of the SV40 promoter in plasmid pPur using Hindlll and Agel creating pPur-LexA. Oligo's Link-for-Bsu (GATCTCCCCTGAGGAAGTGC ACAACCTGAGGCC) and Link-rev-Bsu (GATCTGGCCTCAGGTTGTGCACT TCCTCAGGGG) were ligated into the BamHI site of pPur-LexA to create pPur-LexA-linker. The control vector PPlbs-lexA (PPL) was created by removing the puro coding sequence from pPur-LexA using Agel and Xbal followed by a transfer of the LexA cassette (ApaLI x EcoRI, blunted) into the EcoRV site of PPlbs The Brm pcr-product (primers Brm-alF-H3-AgeI and Brm- al572R-XbaI) was cloned into pPur-LexA using Agel and Xbal to create pPur-LexA-Brm. The P/CAF coding sequence was pσr-amplified from plasmid pCX-P/CAF (Martinez-Balbas et al, 2000) using primers PCAF-alF-h3-AgeI (GATCAAGCTTAC CGGTATGTCCGAGGCTGGCGGGGCCG) and PCAF-a833R-XbaI (AGGCTC TAGAATCACTTGTCAATTAATCCAGCTTCC) and cloned into pPur- LexA-linker using Agel and Xbal to create pPur-LexA-PCAF.

The LexA-Brm cassette was cut from pPur-LexA-Brm using

ApaLI and EcoRI and blunted into the EcoRV site of PPLbs creating PPLBrm. P/CAF was cut from pPur-LexA-PCAF and cloned into PPLBrm using^' Agel and ApaLI/PacI creating

PPLPCAF. The HAT domain of human p300 was pcr-amplified from plasmid pCMVμ-p300 (Martinez-Balbas et al, 2000) using primers p300-a934F-AgeI (GATCACCGGTCAGCCTGCAACTCCACTTTCC CAGCC) and p300-al652R- Nhel (AGGCGCTAGCCTACATGGTGGACCACT GGGCTCTTCGG) and cloned into PPLBrm using Agel and Nhel/Xbal to create PPLp300 (Fig 8A) . The HMTase domain of human Ashl was PCR amplified using primers HuAshl, aal787-For

(GATCACCGGTACAAGCAGCTGTTCCCCCCATC ATATC) and HuAshl, aa2393-Rev (AGGCGCTAGCTCATAATGATGCTGAGT

GAATATTATCAC) and cloned into Agel and Nhel digested

PPLBrm to create PPLHuAshl. 5' STARs were cloned into the Sail site of the various PPL constructs. 3' STARs were cloned either into the Pad site (PPL, PPLBrm and PPLp300) or the Bsu36I site (PPLPCAF) .

Transfection and culture of CHO cells These methods are similar to those described in example 3, except that after transfection and overnight incubation the transfection mixture was replaced with fresh medium, the transfected cells were incubated further, and after overnight cultivation, cells were trypsinized and seeded into fresh culture vessels with fresh medium, and zeocin was added after another overnight incubation .

Example 7 : Chromatin openers improve the level of transgene expression One object of this invention is to improve both the predictability and the levels of transgene expression for heterologous protein production, thus increasing the yield of the heterologous protein and reducing the number of colonies that have to be analysed to obtain a high producer colony. Materials and Methods The tested construct consists of a bicistronic gene with the GFP gene, an IRES and the Zeocin resistance gene under control of the CMV promoter and a monocistronic gene encoding LexA-P/CAF under control of the SV40 promoter, but no STAR elements to flank the entire construct. The construct was transfected to CHO-Kl cells as in Example 6. Stable colonies were expanded before the GFP signal was determined on a XL-MCL Beckman Coulter flowcytometer. The mean of the GFP signal was taken as measure for the level of GFP expression and this is plotted in Fig. 9. The results are compared to colonies that were transfected with a construct containing no LexA-P/CAF gene (Control) and a construct that is flanked with STAR 40 elements (STAR40-shielded) at both the 5' and 3' end, but that contains no LexA-P/CAF.

Results Fig 9 shows that targeting LexA-P/CAF to LexA binding sites upstream of the CMV promoter resulted in a number of CHO colonies that express significantly higher levels of GFP protein, as compared to the "empty" control without LexA-P/CAF. The GFP signal in the colonies with the highest signals was comparable to the highest GFP signal that were obtained with a construct that had flanking STAR 40 elements, but no LexA-P/CAF. However, similar to the distribution of the GFP signals amongst the various colonies, most colonies did not express GFP or at a low level. This indicates that the predictability of the protein expression was not significantly altered as compared to the "empty" control construct . When compared to the GFP signals in colonies transfected with a STAR-shielded construct, these STAR elements convey a higher degree of predictability. The highest GFP expression level in STAR-shielded colonies was of the same order as the GFP expression level in LexA-P/CAF colonies . However, there were significantly more STAR- shielded colonies that showed a high GFP expression level. It is therefore concluded that the LexA-P/CAF opener is able to convey higher expression levels to a transgene, but that they do not convey a higher predictability of transgene expression. Higher predictability is better achieved when STAR elements are added to a construct .

Example 8: The combination of chromatin openers and STAR elements improves predictability and yields of transgene expression Openers were combined with STAR elements as described in Fig 8 and tested were the predictability and yield of transgene expression in stably transfected, individual colonies.

Materials and Methods The tested construct consists of a bicistronic gene with the GFP gene, an IRES and the Zeocin resistance gene under control of the CMV promoter and a monocistronic gene encoding LexA-P/CAF under control of the SV40 promoter. The entire construct is flanked by STAR 40 (Fig 8A) . The construct was transfected to CHO-Kl cells as in Example 6. Stable colonies were expanded before the GFP signal was determined on a XL-MCL Beckman Coulter lowcytometer. The mean of the GFP signal was taken as measure for the level of GFP expression and this was plotted in Fig 10. The results are compared to colonies that are transfected with a construct containing no LexA- P/CAF gene and no STAR elements ("empty" control) and a construct that contains no LexA-P/CAF gene, but that is flanked with STAR 40 at both the 5' and 3' end.

Results Fig 10 shows that the construct in which LexA-P/CAF is targeted to the CMV promoter and that is flanked by STAR elements conveyed high GFP expression levels. The highest GFP expression level was more than three-fold higher than the highest levels in the "empty" control. Moreover, a high degree of predictability of GFP expression levels was found amongst various colonies. In contrast to colonies that expressed a construct with LexA-P/CAF alone (Fig 9) , more colonies that contained the construct with LexA-P/CAF and STAR40 elements had a high level of GFP expression. It is therefore concluded that the combination of STAR elements and an opener conveys both high protein expression levels and a high degree of predictability of expression.

Example 9. STAR-shielded genes that reside on multiple vectors are expressed simultaneously in CHO cells

This example is copied from international patent application no. PCT/NL03/00432, filed 13 June 2003, which application is incorporated herein by reference . STAR elements function to block the effect of transcriptional repression influences on transgene expression units. One of the benefits of STAR elements for heterologous protein production is the increased predictability of finding high-expressing primary recombinant host cells. This feature allows for the simultaneous expression of different genes that reside on multiple, distinct vectors. In this example we use two different STAR7-shielded genes, GFP and RED, which are located on two di ferent vectors . When these two vectors were transfected simultaneously to Chinese hamster ovary (CHO) cells, both were expressed, whereas the corresponding, but unprotected GFP and RED genes, showed hardly such simultaneous expression.

Material and Methods The STAR7 element was tested in the ppGIZ-STAR7 and ppRIP-STAR7 vectors (Fig. 11) . The construction of the pPlug&Play (ppGIZ and ppRIP) vectors is described below. Plasmid pGFP (Clontech 6010-1) was modified by insertion of a linker at the BsiWI site to yield pGFP-link. The linker (made by annealing oligonucleotides 5 " GTACGGATATCAGATCTTTAATTAAG3 ' and 5'GTACCTTAATTAAAGATCTGATATCC3 ' ) introduces sites for the Pad, Bglll, and EcoRV restriction endonucleases. This creates the multiple cloning site MCSII for insertion of STAR elements . Then primers (5 ' GATCAGATCTGGCGCGCCATTTAAATCGTCTCGCGCGTTTCGGTGATGACGG3 ' ) and (5* AGGCGGATCCGAATGTATTTAGAAAAATAAACAAATAGGGG3 ' ) were used to amplify a region of 0.37 kb from pGFP, which was inserted into the Bglll site of pIRES (Clontech 6028- 1) to yield pIRES-stuf . This introduces sites for the Ascl and Swal restriction endonucleases at MCSI, and acts as a "stuffer fragment" to avoid potential interference between STAR elements and adjacent promoters. pIRES-stuf was digested with Bglll and Fspl to liberate a DNA fragment composed of the stuffer fragment, the CMV promoter, the IRES element (flanked by multiple cloning sites MCS A and MCS B) , and the SV40 polyadenylation signal . This fragment was ligated with the vector backbone of pGFP-link produced by digestion with BamHI and Stul, to yield pIRES-link. The open reading frames of the zeocin-resistance gene was inserted into the BamHI /NotJ sites of MCS B in pIRES-link as follows: the zeocin-resistance ORF was amplified by PCR with primers 5 ' GATCGGATCCTTCGAAATGGCCAAGTTGACCAGTGC3 ' and 5*AGGCGCGGCCGCAATTCTCAGTCCTGCTCCTC3' from plasmid pEM7/zeo, digested with BamHI and Notl, and ligated with BamHI/Wotl-digested pIRES-link to yield pIRES-link-zeo. The GFP reporter ORF was introduced into pIRES-link-zeo by amplification of phr-GFP-1 with primers 5 ' GATCGAATTCTCGCGAATGGTGAGCAAGCAGATCCTGAAG3 * and

S'AGGCGAATTCACCGGTGTTTAAACTTACACCCACTCGTGCAGGCTGCCCAGGS • , and insertion of the ScoRI-digested GFP cassette into the .EcoRI site in MCS A of the pIRES-link-zeo plasmid. This created the ppGIZ (for ppGFP-IRES-zeo) . 5' STAR7 was cloned into the Sail site and 3' STAR7 was cloned into the Pad site . The puromycin-resistance ORF was amplified by PCR with primers 5 * GATCGGATCCTTCGAAATGACCGAGTACAAGCCCACG3 ' and 5ΑGGCGCGGCCGCTCAGGCACCGGGCTTGCGGGTC3' from plasmid pBabe-Puro (Morgenstern & Land, 1990), digested with BajnHI and Notl , and ligated with BamHI/iVofcl-digested pIRES-link to yield pIRES-link-puro . The RED gene was amplified by PCR with primers

5 ' GATCTCTAGATCGCGAATGGCCTCCTCCGAGAACGTCATC3 ' and S'AGGCACGCGTTCGCGACTACAGGAACAGGTGGTGGCGS* from plasmid pDsRed2 (Clontech 6943-1), digested with Xbal and Λtlul and ligated to Nhel-MluT digested pIRES-link-puro to yield ppRIP (for ppRED-IRES-puro) . 5' STAR7 was cloned into the Sail site and 3' STAR7 was cloned into the Pad site.

Transfection and culture of CHO cells The Chinese Hamster Ovary cell line CHO-Kl (ATCC CCL-61) was cultured in HAMS-F12 medium + 10% Fetal Calf Serum containing 2 mM glutamine, 100 U/ml penicillin, and 100 micrograms/ml streptomycin at 37° C/5% CO2. Cells were transfected with the plasmids using Lipofectamine 2000 (Invitrogen) as described by the manufacturer. Briefly, cells were seeded to culture vessels and grown overnight to 70-90% confluence. Lipofectamine reagent was combined with plasmid DNA at a ratio of 7.5 microliters per 3 microgram (e.g. for a 10 cm Petri dish, 20 micrograms DNA and 120 microliters Lipofectamine) and added after a 30 minutes incubation at 25°C to the cells. After a 6 hour incubation the transfection mixture was replaced with fresh medium, and the transfected cells were incubated further. After overnight cultivation, cells were trypsinized and seeded into fresh petri dishes with fresh medium with zeocin added to a concentration of 100 μg/ml and the cells were cultured further. When individual colonies became visible (approximately ten days after transfection) medium was removed and replaced with fresh medium (puromycin) . Individual colonies were isolated and transferred to 24-well plates in medium with zeocin. Expression of the GFP and RED reporter genes was assessed approximately 3 weeks after transfection. One tested construct consists of a monocistronic gene with the GFP gene, an IRES and the Zeocin resistance gene under control of the CMV promoter, but either with or without STAR7 element to flank the entire construct (Fig. 11) . The other construct consists of a monocistronic gene with the RED gene, an IRES and the puromycin resistance gene under control of the CMV promoter, but either with or without STAR7 element to flank the entire construct (Fig. 11) . The constructs were transfected to CHO-Kl cells. Stable colonies that were resistant for both zeocin and puromycin (puromycin was used in a concentration of 2.5 μg/ml) were expanded before the GFP and RED signals were determined on a XL-MCL Beckman Coulter flowcytometer. The percentage of cells in one colony that were double positive for both GFP and RED signals is taken as measure for simultaneous expression of both proteins and this is plotted in Fig. 11. Results Fig. 11 shows that simultaneous expression in independent zeocin and puromycin resistant CHO colonies of GFP and a RED reporter genes that are flanked by a STAR element resulted in a higher number of cells that express both GFP and RED proteins, as compared to the control vectors without STAR7 element. The STAR7 element therefore conveys a higher degree of predictability of transgene expression in CHO cells. In the STAR-less colonies at most 9 out of 20 colonies contained double GFP/RED positive cells. The percentage of double positive cells ranged between 10 and 40%. The remaining 11 out of 20 colonies had less than 10% GFP/RED positive cells. In contrast, in 19 out of 20 colonies that contained the STAR-shielded GFP and RED genes, the percentage GFP/RED double positive cells ranged between 25 and 75%. In 15 out of these 19 double positive colonies the percentage GFP/RED double positive cells was higher than 40%. This result shows that it is more likely that simultaneous expression of two genes is achieved when these genes are flanked with STAR elements .

Example 10. Expression of a functional antibody f om two separate plasmids is easier obtained when STAR elements flank the genes encoding the heavy and light chains . Due to the ability of STAR elements to convey higher predictability to protein expression two genes can be expressed simultaneously from distinct vectors. This is shown in example 9 for two reporter genes, GFP and RED. Now the simultaneous expression of a light and a heavy antibody chain is tested. In this example, STAR7-shielded light and heavy antibody cDNAs that reside on distinct vectors were simultaneously transfected to Chinese hamster ovary cells . This resulted in the production of functional antibody, indicating that both heavy and light chains were expressed simultaneously. In contrast, the simultaneous transfection of unprotected light and heavy antibody cDNAs showed hardly expression of functional antibody.

Materials and Methods The tested constructs were the same as described in Example 9, except that the GFP gene was replaced by the gene encoding the light chain of the RINGl antibody (Hamer et al., 2002) and the RED gene was replaced by the gene encoding the heavy chain of the RINGl antibody. The light chain was amplified from the RINGl hybridoma (Hamer et al., 2002) by RT-PCR using the primers S'CAAGAATTCAATGGATTTTCAAGTGCAGS* and 5'CAAGCGG

CCGCTTTGTCTCTAACACTCATTCC3 * . The PCR product was cloned into pcDNA3 after restriction digestion with BcoRI and iVofcl and sequenced to detect potential frame shifts in the sequence. The cDNA was excised with SσoRI and NotT, blunted and cloned in ppGIZ plasmid. The heavy chain was amplified from the RINGl hybridoma (Hamer et al . , 2002) by RT-PCR using the primers 5 'ACAGAATTCTTACCATGGATTTTGGGCTG3 ' and 5ΑCAGCGGCCGCTCATTTACCAGGAGAGTGGG3' . The PCR product was cloned into pcDNA3 after restriction digestion with .EcoRI and NotT. and sequenced to detect potential frame shifts in the sequence . The cDNA was excised with SσoRI and NotX , blunted and cloned in ppRIP plasmid.

Results CHO colonies were simultaneously transfected with the RINGl Light Chain (LC) and RINGl Heavy Chain (HC) cDNAs that reside on two distinct vectors. The Light Chain was coupled to the zeocin resistance gene through an IRES, the Heavy Chain was coupled to the puromycin resistance gene through an IRES. Fig. 12 shows that simultaneous transfection to CHO cells of the heavy and light chain encoding cDNAs resulted in the establishment of independent zeocin and puromycin resistant colonies. When the constructs were flanked by the STAR7 element this resulted in a higher production of functional RINGl antibody, as compared to the control vectors without STAR7 element. The STAR7 element therefore conveyed a higher degree of predictability of antibody expression in CHO cells. In the STAR-less colonies only 1 out of 12 colonies expressed detectable antibody. In contrast, in 7 out of 12 colonies that contained the STAR-shielded Light and Heavy Chain genes, produced functional RINGl antibody that detected the RINGl antigen in an ELISA assay. Significantly, all these 7 colonies produced higher levels of RINGl antibody than the highest control colony (arbitrarily set at 100%) . This result shows that it is more likely that simultaneous expression of two genes encoding two antibody chains is achieved when these genes are flanked with STAR elements.

Example 11. The p300HAT opener improves the level of CMV- driven expression in stably transfected PER.C6 clones, but only for a limited period. In this example we tested whether the P300HAT Opener is able to induce stability of gene expression over an extended period of time in El-immortalized retina cells.

Materials and Methods Plasmids Plasmids CMV-d2EGFP-ires-Zeo (CMV Control) and the CMV-d2EGFP-ires-Zeo—LexA-P300HAT (CMV-p300HAT) were used (Fig. 13), and were constructed as follows. The open reading frame of the zeocin-resistance gene was inserted into BamHI/Ivotl sites downstream of the pIRES as follows: the zeocin-resistance ORF was amplified by PCR with primers GATCGGATCCTTCGAAATGGC CAAGTTGACCAGTGC and AGGCGCGGCCGCAATTCTCAGTCCT GCTCCTC from plasmid pEM7/zeo, digested with BamHI and NotX , and ligated with BamHI/iVotl-digested pIRES-link to yield pIRES-link-zeo. The d2EGFP reporter ORF was introduced into pIRES-link- zeo by amplification of (Clontech 6010-1) with primers GATCGAATTCTCGCGAATGGTGAGCAAGCAG ATCCTGAAG and AGGCGAATTCACCGGTGTTTAAACTTACACCCACTC GTGCAGGCTGCCCAGG, and insertion of the BcoRI-digested d2EGFP cassette into the EcoRX site in the pIRES-link-zeo plasmid. This created the CMV Control (CMV-d2EGFP-IRES-Zeo) . The LexA-P300HAT opener in this plasmid is situated differently when compared to the plasmids in Fig. 8A. In Fig. 8A the SV40-Lex-Opener unit is placed downstream from the other expression unit that encompasses the CMV- driven GFP reporter gene. Transcription of both units is then in the same direction. In the novel plasmid (Fig. 13) , the transcription of the CMV-driven d2EGFP reporter gene is directed away from the transcription of the SV40- driven LexA-P300HAT opener. In this configuration the CMV and SV40 promoters are in close proximity. Between these two promoters LexA binding sites are present. Hence the LexA-P300HAT will influence the expression status of both the CMV-driven reporter gene and the SV40-driven Opener. LexA binding sites (LBS) were amplified from plasmid pREP4-HSF-Luc+ (van der Vlag et al, 2000) using primers AGGCGTCGACGTTTCG ACTCCCAAGCTTTG and GATCGGCGCGCCGGTACC ATAGCGGCCGCGAC and cloned between the CMV and SV40 promoters in PP using Sail and Ascl . LexA was amplified from plasmid pEG202 (Bennetzen and Hall, 1982) using primers GATCAAGCTTATGAAGACGTTAACGGCCAGGC and

AGGCACCGGTCAGCCAGTCGCCGTTGCGAATAACC and cloned downstream of the SV40 promoter in plasmid pPur (BD biosciences #6156-1) using Hindlll and Agel creating pPur-LexA. Oligo's GATCTCCCCTGAGGAAGTGCACAACCTGA GGCC and GATCTGGCCTCAGGTTGTGCACT TCCTCAGGGG were ligated into the BamHI site of pPur-LexA to create pPur-LexA-linker. The HAT domain of human p300 (aa934-1652) was pcr-amplified from plasmid pCMVμ-p300 (Martinez-Balbas et al, 2000) using primers GATCACCGGTCAGCCT GCAACTCCACTTTCCCAGCC and AGGCGCTAGCCTACATGGTGG ACCACTGGGCTCTTCGG and cloned into pPur-LexA-linker using Agel and Nhel/Xbal creating pPur- LexA-P300-HAT. The entire SV40-LexA-P300-HAT transcription unit was cloned downstream of the LexA binding sites, to create —CMV-d2EGFP-ires-Zeo—LexA- P300HAT (CMV-p300HAT) (Fig. 13) .

Transfection, culturing and analysis of PER.C6 cells

PER.C6™ cells were cultured in DMEM medium + pyridoxine + 9% Foetal Bovine Serum (Non-Heat Inactivated) , 8.9 mM MgCl2100 U/ml penicillin, and 100 micrograms/ml streptomycin at 37°C/10% CO?.. Cells were transfected with the plasmids using Lipofectamine 2000 (Invitrogen) as described by the manufacturer. Briefly, cells were seeded to 6-wells and grown overnight to 70-90% confluence. Lipofectamine reagent was combined with plasmid DNA at a ratio of 15 microliters per 3 microgram (e.g. for a 10 cm Petri dish, 20 micrograms DNA and 120 microliters Lipofectamine) and added after 30 minutes incubation at 25°C to the cells. After 6-hour incubation the transfection mixture was replaced with fresh medium, and the transfected cells were incubated further. After overnight cultivation, cells were trypsinized and seeded (1:15, 1:30, 1:60, 1:120 dilutions) into fresh petri dishes (90 mm) with fresh medium with zeocin added to a concentration of 100 μg/ml and the cells were cultured further. When colonies became visible, individual clones were isolated by scraping and transferred to 24-well plates in medium with zeocin. When grown to ~70% confluence, cells were transferred to 6-wells plates. Stable colonies were expanded for 2 weeks in 6-well plates before the GFP signal was determined on a XL-MCL Beckman Coulter flowcytometer. The mean of the GFP signal was taken as measure for the level of GFP expression . Colonies were measured for a second time after 2 weeks . Thereafter colonies were further cultured in the absence of zeocin and the GFP signals were measured after 2 and 4 more weeks .

Results

The results are shown in Fig. 13. It can be seen that targeting LexA~P300HAT to LexA binding sites upstream of the CMV promoter resulted in a number of colonies that express higher levels of d2EGFP protein, as compared to the "empty" control without LexA-P300HAT. It is concluded that an Opener is functional in El-immortalized retina cells.

However, when followed for an extended period of time, expression levels of both plasmids dropped. We conclude that the LexA-p300HAT conveys higher expression levels in comparison with the plasmid without Opener, but only for a limited period. With the CMV promoter the higher expression levels induced by the LexA-P300HAT Opener is therefore limited in time, at least when the cells were cultured in the absence of antibiotic selection pressure.

Example 12. STAR elements improve stability over time of the p300 HAT-mediated increased gene expression levels in El-immortalized retina cells. In this example we tested whether the combination of STAR elements and the LexA-P300HAT is able to promote long-term stability of gene expression in PER.C6 cells.

Plasmids In CMV-p300HAT, a 5' STAR sequence was cloned into the Sail site and a 3' STAR sequence was cloned into the Pad site of the CMV-p300HAT construct to create CMV- p300HAT-STAR. For these experiments STAR 4, STAR 6, STAR 7, STAR 12, STAR 18, STAR 35 and STAR 40 were used as STAR sequences. The constructs therefore allowed a comparison of the effect of these individual STAR elements with respect to each other in PER.C6 cells. In addition, upstream of the 5' STAR and downstream of the 3' STAR a TRAP sequence consisting of a synthetic poly A combined with the 92 bp pausing signal from the human μ2 globin gene (see example 1; the TRAP defined by SEQ. ID. NO. 15 was used) was cloned. The resulting plasmids (see Fig. 14) , comprising TRAP sequences, an opener binding site and a coding sequence for an opener capable of binding to the opener binding site, the dEGFP gene under control of a CMV promoter and coupled to the Zeo marker coding sequence via an IRES sequence, and the STAR sequences were compared to the CMV control plasmid for expression levels . The constructs were tested in PER.C6 cells by methods as described in example 11.

Results The results for the plasmids containing STAR 4, STAR 6 and STAR 7 are shown in Fig. 15. When comparing Fig. 13 (example 11: no STAR sequences present) with Fig. 15, it can be seen that the presence of STAR 6 and STAR7 in the constructs improves the stabililty of expression over time. The presence of the TRAP sequence may also contribute to this effect. This demonstrates that application of a combination of Openers and STAR elements protect transgenes from silencing during prolonged cultivation. It also demonstrates that this property is independent of antibiotic selection.

A graphic representation of the expression levels in 9 clones with the highest expression levels after 4 weeks without selection in the medium is shown in Fig. 16 for the control plasmid (A in Fig. 14) and the plasmids comprising STAR 4, STAR 6 and STAR 7 (C, E and D, respectively in Fig. 14) . In Fig. 17 the effect of the opener (construct B versus A in Fig. 14) in El- immortalized retina cells can be seen under the same conditions. Clearly the opener gives some improvement over the control, but this effect is less than when combined with STAR and TRAP sequences. In Fig. 18 the results for the other tested STAR sequences are given. It can be seen that STAR 12, STAR 18, STAR 35 and STAR 40 combined with TRAPs and openers also contribute to improved expression in El-immortalized retina cells. It can be concluded that STAR sequences are functional in El-immortalized retina cells. Of the tested STAR sequences, STAR 4, STAR 6 and STAR 7 gave the highest improvement in expression levels in PER.C6 cells. Hence, in certain preferred embodiments according to the invention, the used STAR sequences are chosen from the group consisting of STAR 4, STAR 6 and STAR 7, and unctional fragments or derivatives thereo . Upon analysis of all available data it appeared that STAR 7 thus far gave the best results in El-immortalized retina cells. Of the tested constructs, STAR 12 appeared to be the least active in these cells .

Example 13. The effect of STAR elements on expression of antibodies in El-immortalized retina cells

It was shown above that STAR elements improve the expression of a marker gene (d2EGFP) in PER.C6 cells. It is further tested whether STAR elements, optionally in combination with Opener and TRAP sequences, improve the expression of antibodies in El-immortalized retina cells. Antibody expression in such cells has been described (WO 00/63403; Jones et al, 2003) . Expression units as described in those publications are provided with STAR sequences, TRAPS and/or openers as described supra, and the effect of these elements on expression is evaluated, in established clones obtained according to methods known to the person skilled in the art (Jones et al, 2003) . In such embodiments, the heavy and light chain may be present as separate expression units on a single plasmid molecule before transfection into the cells (e.g constructs similar to that used by (Jones et al, 2003) . A suitable expression vector, pcDNA3002 (Neo) , which has been described in WO 03/051927, has been deposited at the European Collection of Cell Cultures (ECACC) under number 01121318. This vector has two CMV promoters behind which the heavy and light chain can be cloned, respectively, resulting in a single plasmid encoding both chains of an antibody . Alternatively, the expression units for the heavy and light chain of the antibody may be present on two separate plasmid molecules (see e.g. example 10 above) . Preferably, each expression unit comprises two STAR sequences as described above. For the experiments, STAR 7 is used, as this has given the best results thus far in El-immortalized cells. In these examples, the expression units can for instance encode the anti-EpCAM antibody, as was also used in the studies of (Jones et al, 2003) . The isolation of the DNA encoding the antigen-binding region of this antibody from a scFv phage display library has been described (Huls et al, 1999) . A leader sequence and constant regions were added as described in Boel et al, 2000. It is expected that also for the expression of antibodies the STAR and/or TRAP and/or opener elements lead to an increase in either predictability of expression (lower number of colonies have to be screened to obtain good producers) , or absolute expression levels, or stability of expression, or a combination thereof, when compared to the controls where these elements are not present. TABLES Table 1. TRAP sequences used

Table 2. STAR sequences used

DESCRIPTION OF FIGURES

FIG 1. Schematic diagram of some expression units that can be silenced. FIG 1A shows a single expression unit gene 1 under the control of the CMV promoter on one plasmid. This plasmid has integrated as multiple copies into the genome, in such an orientation that the transcription is convergent. Consequently there will be read-through transcription from copy one into copy two and vice versa. This will result in the formation of dsRNA. This plasmid suffers from silencing of the genes. Fig IB shows two expression units, gene 1 and gene 2, both under the control of the CMV promoter and located in a divergent orientation on one plasmid. This plasmid has integrated as multiple copies into the genome. No matter what the orientation, there will always be read- through transcription from one gene on copy one into another gene on copy two. This results in the formation of dsRNA. This plasmid suffers from silencing of the genes . FIG 1C shows a single expression unit gene 1 under the control of the CMV promoter on one plasmid. This plasmid has integrated as a single copy into the genome. When integration is adjacent of a promoter that is oriented in a convergent manner relative to the plasmid, there will be read-through transcription from that promoter into gene 1 of the plasmid. This will result in the formation of dsRNA. This plasmid suffers from silencing of the genes.

FIG 2. Schematic diagram of an aspect of the invention (TRAP) FIG 2A shows the first expression unit. It is flanked by TRAPs and STAR elements, and comprises a bicistronic gene containing (from 5' to 3' ) a transgene (encoding for example a reporter gene or one subunit of a multimeric protein; Gene) , an IRES, and a selectable marker (zeo, conferring zeocin resistance) under control of the CMV promoter. A monocistronic selectable marker (puro) under control of the SV40 promoter is included. Both genes have the SV40 transcriptional terminator at their 3' ends (t) . The TRAPs are drawn as an arrow indicating that in this particular orientation transcription driven by any promoter outside the expression unit does not enter the expression unit. FIG 2B shows two expression units on one plasmid. Both Gene 1 and Gene 2 are both part of a bicistronic gene containing (from 5' to 3') a transgene (Genel), an IRES, and a selectable marker (zeo with Gene 1 and puro with Gene 2) under control of the CMV promoter and the SV40 transcriptional terminator (t) . The entire cassette is surrounded by STAR elements and TRAPs the latter are oriented in such a manner that transcription is kept out of the cassette and STAR elements.

FIG 3. The pcodA plasmid to identify and test putative TRAPs. The pIRES-31 plasmid contains the CMV promoter upstream of an Intervening Sequence IV (Clontech) that contains a multiple cloning site in which putative TRAPs are cloned. Downstream is the codA: : upp suicide gene. The plasmid further comprises the hygromycin resistance gene that is under control of the SV40 promoter. The plasmid also has an origin of replication (ori) and ampicillin resistance gene (amp^R) for propagation in Escherichia coll and the EBNA-1 nuclear antigen for high copy episomal replication. The TRAP is drawn as an arrow indicating that the TRAP blocks transcription driven by the CMV promoter in this particular orientation. This is of importance for the orientation of TRAPs in FIG 2, which are also drawn as arrows to indicate the specific orientation of the TRAPs to prevent transcription driven by any promoter outside the expression unit to enter this expression unit .

FIG 4. TRAPs e iciently block CMV promoter driven transcription Indicated constructs with potential TRAPs that are located between the CMV promoter and the codA gene are transfected to U-2 OS cells. 1) The empty control vector without sequences between the CMV promoter and the codA gene (FIG 3) 2) A 2400 bp long DNA of phage μ (bp 35711-38103) in 5' -3' orientation 3) A 2400 bp long DNA of phage μ (bp 35711-38103) in 3' -5' orientation 4) The 60 bp long MAZ DNA sequence (Ashfield et al 1994) 5) ST R7 6) STAR 40 7) The empty control vector as shown in FIG 3 8) A synthetic poly A (SPA) sequence (Levitt et al 1989) in 5' -3' orientation 9) A combination of the SPA sequence and a 92 bp long μ2 globin gene pause signal in 5' -3' orientation 10) A synthetic poly A sequence in 3' -5' orientation 11) A combination of the SPA sequence and a 92 bp long μ2 globin gene pause signal in 3' -5' orientation

After the transfection selection by hygromycin mRNA is isolated and blotted. The blot is incubated with a radioactive labelled probe encompassing the codA gene. As a loading control the blot is also incubated with a radioactive probe encompassing the hygromycin resistance gene. The lambda (bp 35711-38103) fragment efficiently blocks transcription of the codA gene driven by the CMV promoter in the 5' -3' orientation (lane 2), but not in the 3' -5' orientation (lane 3) . Also a synthetic polyA (SPA) sequence either alone (lane 8) or in combination with a 92 bp long μ globin pausing signal (lane 9) efficiently block transcription in the 5' -3' orientation, but not in the 3' -5' orientation (lane 10 and 11) . Neither MAZ sequence (lane 4) , STAR 7 (lane 5) nor STAR 40 (lane 6) prevent transcription of the codA gene driven by the CMV promoter. All signals are compared to the control vector that contains no putative TRAP sequence (lanes 1 and 7) .

FIG 5. TRAPs improve the effects of STAR elements on transgene expression

Constructs that are flanked with the lambda (bp 35711- 38103) TRAP in the A orientation, STAR 40 or the combined lambda (bp 35711-38103) TRAP/ STAR40 are transfected to CHO-Kl cells. The 5' -3' orientation of the lambda (bp 35711-38103) TRAP results in transcription blocking (FIG 3) and the TRAPs are placed to flank to entire construct such that transcription can not enter the expression units. Stable colonies (14 of each construct) are expanded and the GFP signal is determined on a XL-MCL Beckman Coulter flowcytometer . For each independent colony the mean of the GFP signal is plotted. This is taken as measure for the level of GFP expression. The results are compared to colonies that are transfected with a construct containing neither lambda (bp 35711- 38103) TRAP nor STAR element (Control) .

FIG 6. TRAPs act in an orientation-dependent manner

Constructs that are flanked with the lambda (bp 35711- 38103) TRAP in the 3' -5' orientation, STAR 40 or the combined lambda (bp 35711-38103) TRAP/ STAR40 are transfected to CHO-Kl cells. The 3' -5' orientation of the lambda (bp 35711-38103) TRAP does not result in transcription blocking (FIG 3) . Analysis of stable colonies is as in FIG 5.

FIG 7. TRAPs and STARs improve the stability of transgene expression

Stably transfected CHO-Kl colonies that contain either a lambda (bp 35711-38103) TRAP/ STAR-less (Control) GFP construct or the GFP construct that is flanked by the combined lambda (bp 35711-38103) TRAP/ STAR 40 are expanded. Of the both categories four colonies are chosen with the highest GFP levels (see FIG 5) . These colonies are further cultured without the antibiotic (zeocin) and the GFP signal is determined with intervals of one week, which represent approximately 10 cell cycles. The mean of the GFP signal is plotted as in FIG 3. The first bar of each colony represents the GFP' signal at the moment that the antibiotic selection pressure is removed. The adjacent three bars represent the GFP signal that is measured after one, two and three weeks.

FIG 8. Schematic diagram of an aspect of the invention (opener) FIG 8A shows two expression units on one plasmid. Expression unit one comprises a bicistronic gene containing (from 5' to 3' ) a transgene (encoding for example one subunit of a ultimeric protein; Genel) , an IRES, and a selectable marker (zeo, conferring zeocin resistance) under control of the CMV promoter. Upstream of the CMV promoter are four LexA binding sites (LexA- BS) . The expression unit has the SV40 transcriptional terminator at its 3' end (t) . Next is a monocistronic gene encoding a fusion protein between the LexA protein and either (i) a histone acetyltransferase (HAT) or a functional part of a HAT that is still able to transfer acetyl groups to histone tails (LexA-HAT) , (ii) a histone methyltransferase (HMTase) or a functional part (SET domain) of a HMTase that is still able to transfer methyl groups to at least lysine K4 of the histone H3 tail or

(iii) the trithorax group protein Brahma. These genes are under control of the SV40 promoter. The expression unit has the SV40 transcriptional terminator at its 3' end (t) . The entire cassette with the two expression units is flanked by STAR elements.

FIG 8B is similar to FIG 8A, but there are now three expression units on one plasmid. Expression unit one comprises a bicistronic gene containing a transgene

Genel, an IRES, and a selectable marker zeo under control of the CMV promoter. The transcription orientation of this first expression unit is directed upstream. Expression unit two comprises a bicistronic gene containing a transgene Gene2, an IRES, and a selectable marker puro (puromycin resistance gene) under control of the CMV promoter. The transcription orientation of this first expression unit is directed downstream. Between the two CMV promoter of the two expression units are four LexA binding sites (LexA-BS) . The monocistronic gene encodes the same LexA fusion proteins as in FIG 8A. The entire constellation of three expression units is flanked by STAR elements.

FIG 9. Chromatin openers improve CMV driven GFP expression in CHO cells

The constructs that contain the gene encoding LexA-P/CAF are transfected to CHO-Kl cells. Stable colonies (14 of each construct) are expanded and the GFP signal is determined on a XL-MCL Beckman Coulter flowcytometer. For each independent colony the mean of the GFP signal is plotted. This is taken as measure for the level of GFP expression. The results are compared to colonies that are transfected with a construct containing no LexA-P/CAF gene (Control) and a construct that is flanked with STAR 40 elements (STAR40-shielded) at both the 5' and 3' end.

FIG 10. The combination of chromatin openers and STARs enhances CMV promoter driven GFP expression in CHO cells The construct that is flanked by STAR 40 and that contains the gene encoding LexA-P/CAF (see FIG 8) was transfected to CHO-Kl cells. Stable colonies (14 of each construct) are expanded, the GFP signal is determined and the mean of the GFP signal is plotted as in FIG 9. The results are compared to colonies that are transfected with a construct containing no LexA-P/CAF or STAR 40 elements (Control) and a construct that is flanked with STAR 40 elements (STAR40) at both the 5' and 3' end.

FIG 11. STAR elements allow efficient and simultaneous expression of two genes from two distinct vectors.

The ppGIZ, ppGIZ-STAR7, ppRIP and ppRIP-STAR7 vectors used for testing simultaneous expression of respectively GFP and RED are shown. The expression unit comprises (from 5' to 3') genes encoding the GFP or RED proteins, an IRES, and a selectable marker (zeo, conferring zeocin resistance or respectively puro, puromycin resistance gene) under control of the CMV promoter. The expression unit has the SV40 transcriptional terminator at its 3' end (t) . The cassettes with the GFP and RED expression units are either flanked by STAR7 elements (STAR7- shielded) or not (Control) . The two control constructs or the two STAR7-shielded vectors were simultaneously transfected to CHO-Kl cells. Stable colonies that were resistant to both zeocin and puromycin were expanded and the GFP and RED signals were determined on a XL-MCL Beckman Coulter lowcytometer. The percentage of cells in one colony that are double positive for both GFP and RED signals is taken as measure for simultaneous expression of both proteins and this is plotted. FIG 12. STAR elements improve expression of a functional antibody in CHO cells.

The different vectors containing the Light and Heavy Chain of the RINGl antibody are shown. The constructs were simultaneously transfected to CHO cells. Stable colonies that were resistant to both zeocin and puromycin were expanded. The cell culture medium of these colonies was tested for the detection of functional RINGl antibody in an ELISA with RINGl protein as antigen. The values were divided by the number of cells in the colony. The highest value detected in the STAR-less control is arbitrarily set at 100%.

Fig. 13. The p300HAT opener improves the level of CMV- driven expression in stably transfected clones , but only for a limited period.

Two different constructs were transfected to PER.C6 cells. The construct used in panel A is the control construct, and is schematically represented as construct A in Fig. 14. The construct used in panel B is schematically represented as construct B in Fig. 14. An indicated number of stable colonies were expanded and after different, indicated time periods, the d2EGFP signal was determined on a XL-MCL Beckman Coulter lowcytometer. For each independent colony the mean of the d2EGFP signal is plotted. This is taken as measure for the level of d2EGFP expression.

Fig. 14. Schematic representation of expression constructs used in example 12. All constructs comprise an expression unit comprising a bicistronic gene containing (from 5' to 3') the d2EGFP reporter gene, an IRES, and a selectable marker (zeo, conferring zeocin resistance) under control of the CMV promoter. The cassette has the SV40 transcriptional terminator at its 3' end (t) . Upstream of the CMV promoter, constructs B-l further contain four LexA- binding site (LA-BS) and downstream thereof is a monocistronic gene encoding a fusion protein between the LexA protein and the functional p300 histone acetyltransferase domain (LexA-p300) under control of the SV40 promoter. This expression unit also has the SV40 transcriptional terminator at its 3' end (t) . Transcription of both expression units is directed opposite. Hence the LexA binding sites are placed between and will act upon both expression units. Constructs C-I further contain the indicated STAR sequences, and at the outer ends they contain a transcriptional pause sequence consisting of a synthetic poly A combined with the 92 bp pausing signal from the human μ2 globin gene (TRAP) .

Fig. 15. The increased gene expression levels due to the combined action of STAR elements and the p300 HAT opener is highly stable over time. Constructs D and E of Fig. 14 were tested, and the results are given in panels A and B, respectively. An indicated number of stable colonies are expanded and after different, indicated time periods, the d2EGFP signal is determined on a XL-MCL Beckman Coulter flowcytometer. For each independent colony the mean of the d2EGFP signal is plotted. This is taken as measure for the level of d2EGFP expression.

Fig. 16. Increased gene expression levels due to the combined action of STAR elements 4, 6 and 7 and the p300 HAT opener. The tested constructs described in Fig 14 (CMV control, CMV-p300HAT-STAR4, CMV-p300HAT-STAR6 and CMV- p300HAT-STAR7 are constructs A, C, E and D, respectively in Fig. 14) were transfected to PER.C6 cells and an indicated number of stable colonies were expanded. After 95 days in culture the d2EGFP signal was determined on a XL-MCL Beckman Coulter flowcytometer. For each independent colony the mean of the d2EGFP signal is plotted. This is taken as measure for the level of d2EGFP expression.

Fig. 17. Increased gene expression levels due to the action of the p300 HAT opener. The tested constructs described in Fig 14 (CMV control and CMV-p300HAT are constructs A and B, respectively in Fig. 14) were transfected to PER.C6 cells and an indicated number of stable colonies were expanded. After 95 days in culture the d2EGFP signal was determined on a XL-MCL Beckman Coulter flowcytometer. For each independent colony the mean of the d2EGFP signal is plotted. This is taken as measure for the level of d2EGFP expression.

Fig. 18. Increased gene expression levels due to the combined action of STAR elements 12 , 18, 35 and 40 and the p300 HAT opener. The tested constructs described in Fig 14 (CMV control, CMV-p300HAT-STAR12, CMV-p300HAT-STAR18, CMV- p300HAT-STAR353 and CMV-p300HAT-STAR40 are constructs A, F, G, H and I, respectively in Fig. 14) were transfected to PER.C6 cells and an indicated number of stable colonies were expanded. After 95 days in culture the d2EGFP signal was determined on a XL-MCL Beckman Coulter flowcytometer. For each independent colony the mean of the d2EGFP signal is plotted. This is taken as measure for the level of d2EGFP expression. References

Ashfield, R, Patel, AJ, Bossone, SA, Brown, H, Campbell, RD, Marcu, KB, and Proudfoot, NJ. MAZ -dependent termination between closely spaced human complement genes. EMBO J 13, 5656-5667.

Bannister AJ, and Kouzarides, T (1996) The CBP co -activator is a histone acetyltransferase. Nature 384, 641-643. Beisel, C, Imhof, A, Greene, J, Kremmer, E, and Sauer, F (2002) Histone methylation by the Drosophila epigenetic transcriptional regulator Ashl. Nature 419, 857-862.

Boel, E., Verlaan, S., Poppelier, M.J., Westerdaal, N.A., Van Strijp, J.A., Logtenberg, T. (2000). Functional human monoclonal antibodies of all isotypes constructed from phage display library-derived single-chain Fv antibody fragments. J Immunol Methods. 239, 153-66.

Brownell JE, Zhou J, Ranalli T, Kobayashi R, Ξdmondson DG, Roth SY, Allis CD. (1996) Tetrahymena histone acetyltransferase A: a homolog to yeast Gcn5p linking histone acetylation to gene activation. Cell 84, 843-851.

Bunker, CA, and Kingston, RE. (1994) Transcriptional repression by Drosophila and mammalian Polycoiab group proteins in transfected mammalian cells. Mol Cell Biol 14, 1721-1732.

Chiba, H, Muramatsu, M, Nomato, A, and Kato, H. (1994) Two human homoloques of Saccharomyces cerevisiae SWI2/SNF2 and Drosophila brahma are transcriptional coactivators cooperating with the estrogen receptor and retinoic acid receptor. Nucl Acids Res 22, 1815-1820.

Crosby, M. A., Crosby MA, Miller C, Alon T, Watson KL, Verrijzer CP, Goldman-Levi R, Zak NB. (1999) . The trithorax group gene moira encodes a brahma-associated putative chro atin-rerαodeling factor in Drosophila elanogaster. Mol. Cell Biol.19, 1159-1170.

Elfring LK, Deuring R, McCallum CM, Peterson CL, Tamkun JW. (1994) . Identification and characterization of Drosophila relatives of the yeast transcriptional activator SNF2/SWI2. Mol Cell Biol 14,2225-34

Enriquez-Harris P, Levitt N, Briggs D, and Proudfoot NJ. (1991) A pause site for RNA polymerase II is associated with termination of transcription. EMBO J. 10, 1833-184

Farkas, G., Gausz, J., Galloni, M., Reuter, G. , Gyurkovics, H. and Karch, F. (1994) . The Tri hor -li e gene encodes the Drosophila GAGA factor. Nature 371, 806-808

Fyodorov, DV, and Kadonaga, JT. (2001) The many faces of chromatin remodeling : Switching beyond transcription. Cell 106, 523-525.

Hamer, CM, Sewalt, RGAB, Den Blaauwen, JL, Hendrix, M, Satijn, DPE, and Otte, AP. (2002) . A panel of monoclonal antibodies against human Polycomb group proteins . Hybridoma and Hybrido ics 21, 245-52.

Huls GA, Heijnen IAFM, Cuomo ME, Koningsberger JC, Wiegman L, Boel E, van der Vuurst-de Vries A-R, Loyson SAJ, Helfrich W, van Berge Henegouwen GP, van Meijer M, de Kruif J, Logtenberg T. (1999) . A recombinant, fully human monoclonal antibody with antitumor activity constructed from phage-displayed antibody fragments. Nat Bioteσhnol. 17, 276-281.

Ito T, Bulger M, Pazin MJ, Kobayashi R, Kadonaga JT. (1997) ACF, an ISWI-containing and ATP-utilizing chromatin assembly and remodeling factor. Cell 90,145-155.

Jones, D., Kroos, N. , Anema, R., Van Montfort, B., Vooys, A., Van Der Kraats, S., Van Der Helm, E., Smits, S., Schouten, J. , Brouwer, K., Lagerwerf, F., Van Berkel, P., Opstelten, D-J., Logtenberg, T., Bout, A. (2003). High-level expression of recombinant IgG in the human cell line PER.C6™. Biotechnol Prog. 19, 163-8. Levitt N, Briggs D, Gil A, and Proudfoot NJ. (1989) Definition of an efficient synthetic poly (A) site. Genes & Dev 3, 1019- 1025.

Martinez-Balbas MA, Bauer UM, Nielsen SJ, Brehm A, Kouzarides T. (2000) Regulation of E2F1 activity by acetylation. EMBO J. 19 : 662-671

Martinez-Salas, E. (1999) Internal ribosome entry site biology and its use in expression vectors Curr Opin Biotechnol 10, 458-64.

Mazo AM, Huang DH, Mozer BA, Dawid IB. (1990) . The trithorax gene, a trans-acting regulator of the bithorax complex in

Drosophila, encodes a protein with zinc-binding domains. Proc Natl Acad Sci USA 87, 2112-2116.

Mizuguchi, H, Xu, Z, Ishii-Watabe, A, Uchida, E, and Hayakawa, T. (2000) IRES-dependent second gene expression is significantly lower than cap- dependent first gene expression in a bicistronic vector Mol Ther 1, 376-82.

Nakamura, T, Blechman, J, Tada, S, Rozovskaia, T, Itoyama, T, Bullrich, F, Mazo, A, Croce, CM, Geiger, B, and Canaani, E. (2000) huASHl protein, a putative transcription factor encoded by a human homologue of the Drosophila ashl gene, localizes to both nuclei and cell -cell tight junctions. Proc. Natl. Acad. Sci USA 97, 7284-7289.

Petruk, S, Sedkov, Y, Smith, S, Tillib, S, Kraesvski, V, Nakamura, , Canaani, E, Croce, CM, and Mazo, A (2001) Trithorax and dCBP acting in a complex to maintain expression of a homeotic gene. Science 294, 1331-1334. Rees, S, Coote, J, Stables, J, Goodson, S, Harris, S, and Lee, MG. (1996) Bicistronic vector for the creation of stable mammalian cell lines that predisposes all antibiotic -resistant cells to express recombinant protein Biotechniques 20, 102-4, 106, 108-10.

Sambrook, J, Fritsch, EF, and Maniatis, T (1989) Molecular Cloning: A Laboratory Manual , Second ed. , Cold Spring Harbor Laboratory Press, Plainview NY.

Tamkun, JW, Deuring, R, Scott, MP, Kissinger, Pattatucci, AM, Kaufman, TC, and Kennison, JA. (1992) brahma: A regulator of Drosophila homeotic genes structurally related to the yeast transcriptional activator SNF2/SWI2. Cell 68, 561-572.

Trelsman, J. E., Luk, A., Rubin, G. M. and Heberlein, U. (1997) . eyelid antagonizes wingless signaling during Drosophila development and has homology to the Bright family of DNA-binding proteins. Genes Dev. 11, 1949-1962

Van der Vlag, J, den Blaauwen, JL, Sewalt, RG, van Driel, R, and Otte, AP. (2000) Transcriptional repression mediated by polycomb group proteins and other chromatin-associated repressors is selectively blocked by insulators J Blol Chem 275, 697-704.

Varga-Weisz, PD, Wilm, M, Bonte, E, Dumas, K, Mann, M, and Becker, PB (1997) Chromatin-remodelling factor CHRAC contains the ATPases ISWI and topoiosmerase II. Nature 388, 598-602. Yang, X-J, Ogryzko, W, Nishikawa, J, Howard, BH, and

Nakatani, Y (1996) A p300/CBP-associated factor that competes with the adenoviral oncoprotein ElA. Nature 382, 319-324.

Claims

l.An El-immortalized retina cell comprising a recombinant protein expression unit which unit comprises a promoter functionally linked to an open reading frame encoding at least one protein of interest, characterized in that said expression unit comprises at least one element improving expression, wherein said element improving expression is chosen from the group consisting of: a) a stabilizing anti-repressor (STAR) sequence; b) a Transcription Pause (TRAP) sequence, and wherein said TRAP sequence is located: i) downstream of the coding sequence of said protein in an orientation that can at least in part prevent formation of antisense RNA of said coding sequence; or ii) upstream of said promoter and in an orientation that can at least in part prevent transcription to enter said protein expression unit; and c) a binding site for a member of a chromatin modification system for rendering chromatin more accessible for transcription (opener) , wherein said opener is present in said cell. 2. A cell according to claim 1, wherein said element improving expression is a STAR sequence. 3. A cell according to claim 2, wherein said STAR sequence is chosen from the group consisting of STAR 4 (SEQ. ID. NO. 1), STAR 6 (SEQ. ID. NO.

2) STAR 7 (SEQ. ID. NO.

3), STAR 12 (SEQ. ID. NO. 4), STAR 18 (SEQ^ ID. NO. 5), STAR 35 (SEQ. ID. NO. 6) and STAR 40 (SEQ. ID. NO. 7), and a functional fragment or derivative of any of these.

4. A cell according to claim 3, wherein said STAR sequence is STAR 4 (SEQ. ID. NO. 1), or a functional fragment or derivative thereof.

5. A cell according to claim 3, wherein said STAR sequence is STAR 6 (SEQ. ID. NO. 2), or a functional fragment or derivative thereof.

6. A cell according to claim 3, wherein said STAR sequence is STAR 7 (SEQ. ID. NO. 3), or a functional fragment or derivative thereof.

7. A cell according to any one of claims 2-6, further characterized in that said recombinant expression unit further comprises at least one Transcription Pause (TRAP) sequence, and wherein said TRAP sequence is located: i) downstream of the coding sequence of said protein in an orientation that can at least in part prevent formation of antisense RNA of said coding sequence; or ii) upstream of said promoter and in an orientation that can at least in part prevent transcription to enter said protein expression unit.

8. A cell according to claim 7, wherein said TRAP sequence is chosen from the group consisting of lambda fragment 35711-38103 (SEQ. ID. NO. 8), a synthetic polyA sequence as identified by (SEQ. ID. NO. 9) or (SEQ. ID. NO. 14) , a 92 bp pausing signal from the human μ2 globin gene (SEQ. ID. NO. 10), a combined synthetic polyA sequence and a pausing signal from the human μ2 globin gene (SEQ. ID. NO. 11 or SEQ. ID. NO. 15), the inter histon H3FA-H4F fragment (SEQ. ID. NO. 12) and the inter histone H1F4-H2FB fragment (SEQ. ID. NO. 13), and a functional fragment or derivative of any of these.

9. A cell according to claim 8, wherein said TRAP sequence is a combined synthetic polyA sequence and a pausing signal from the human μ2 globin gene (SEQ. ID. NO. 11 or SEQ. ID. NO. 15), or a functional fragment or derivative thereof.

10. A cell according to any one of claims 2-9, further characterized in that said recombinant expression unit comprises at least one binding site for a member of a chromatin modification system for rendering chromatin more accessible for transcription (opener) , wherein said opener is present in said cell.

11. A cell according to claim 1, wherein said element improving expression is a TRAP sequence, and wherein said TRAP sequence is located: i) downstream of the coding sequence of said protein in an orientation that can at least in part prevent formation of antisense RNA of said coding sequence; or ii) upstream of said promoter and in an orientation that can at least in part prevent transcription to enter said protein expression unit.

12. A cell according to claim 11, wherein said TRAP sequence is chosen from the group consisting of lambda fragment 35711-38103 (SEQ. ID. NO. 8), a synthetic polyA sequence as identified by (SEQ. ID. NO. 9) or (SEQ. ID. NO. 14), a 92 bp pausing signal from the human μ2 globin gene (SEQ. ID. NO. 10), a combined synthetic polyA sequence and a pausing signal from the human μ2 globin gene (SEQ. ID. NO. 11 or SEQ. ID. NO. 15), the inter histon H3FA-H4F fragment (SEQ. ID. NO. 12) and the inter histone H1F4-H2FB fragment (SEQ. ID. NO. 13) , and a functional fragment or derivative of any of these.

13. A cell according to claim 12, wherein said TRAP sequence is a combined synthetic polyA sequence and a pausing signal from the human μ2 globin gene (SEQ. ID. NO. 11 or SEQ. ID. NO. 15), or a functional fragment or derivative thereof.

14. A cell according to claim 1, wherein said element improving expression is a binding site for a member of a chromatin modification system for rendering chromatin more accessible for transcription (opener) , wherein said opener is present in said cell.

15. A cell according to claim 14, wherein said binding site comprises a lexA or a GAL4 binding site, and wherein said opener is reσombinantly expressed in said cell as a fusion protein comprising a lexA binding domain or GAL4 binding domain rendering it capable of binding to said binding site, and a domain that renders chromatin more accessible for transcription.

16. A cell according to claim 14 or 15, wherein said opener comprises a domain with histone acetyl transferase activity.

17. A cell according to claim 16, wherein said opener comprises a domain of p300 histone acetyl transferase.

18. A cell according to any one of the preceding claims, wherein said protein of interest comprises at least one subunit of an immunoglobulin.

19. A cell according to any one of the preceding claims, wherein said cell is a cell such as deposited at the ECACC under number 96022940.

20. A cell culture comprising a culture medium and a multitude of cells according to any one of claims 1-19.

21. A method for producing at least one recombinant protein in a cell, said method comprising culturing of said cell and expressing said recombinant protein, characterized in that said cell is a cell according to any one of claims 1-19.

22. A method according to claim 21, further comprising the step of collecting said recombinant protein.

23. A method according to claim 21 or 22, wherein said at least one recombinant protein comprises an immunoglobuli .