HK1026000B - Method for integrating genes at specific sites in mammalian cells via homologous recombination and vectors for accomplishing the same - Google Patents

Description

Method for integrating gene into specific site of mammalian cell and vector used in the method

Technical Field

The present invention relates to a method for targeted integration of a desired exogenous DNA into a specific location within the genome of a mammalian cell, and more particularly describes a novel method for identifying a transcriptionally active target site ("hot spot") in the genome of a mammalian cell and inserting the desired DNA into this site via homologous recombination. The present invention also optionally provides for the ability to amplify the gene for the desired DNA at this location by co-integrating an amplifiable selectable marker, such as DHFR, with the foreign DNA. The invention also describes the construction of novel vectors suitable for achieving the above objects and also provides mammalian cell lines produced by the above method, which contain the desired exogenous DNA integrated into the target hot spot.

Background

Techniques for expressing recombinant proteins in prokaryotic and eukaryotic organisms have been developed. Mammalian cells are significantly superior to bacteria or yeast in protein production because they are able to properly assemble, glycosylate, and post-translationally modify recombinantly expressed proteins. Following transfection into a host cell, the recombinant expression construct may be present as an extrachromosomal element, or integrated into the host cell genome. The generation of stably transfected mammalian cell lines typically comprises the following steps: a DNA construct encoding the desired gene, as well as a drug resistance gene (a dominant selectable marker), is introduced into host cells, followed by culture in the presence of a drug to select for cells that have successfully integrated the exogenous DNA. In many cases, the desired gene is linked to a drug resistance selectable marker which can then be amplified by the gene. The gene encoding dihydrofolate reductase (DHFR) is the most commonly used marker gene. Culturing cells in the presence of methotrexate, a competitive inhibitor of DHFR, results in increased production of DHFR due to the amplification of the DHFR gene. Since the flanking regions of the DNA may also be amplified, co-amplification of the DHFR-linked gene in the infected cell line results in increased protein production and thus high expression of the desired gene.

Although this approach has proven successful, there are still many problems in this system due to the randomness of the integration events. These problems arise because the local genetic environment at the locus greatly affects expression levels, a phenomenon well documented in the literature and generally referred to as "position effects" (see, for example, Al-Shawi et Al, molecular cell biology, 10: 1192-1198 (1990); Yoshimura et Al, molecular cell biology, 7: 1296-1299 (1987)). Since most mammalian DNAs are in a state of no transcriptional activity, the random integration method cannot control the transcriptional state of the integrated DNA, and as a result, the expression level of the integrated gene varies greatly depending on the integration site. For example, integration of foreign DNA into an inactive or transcriptionally "silent" region of the genome results in little or no expression. In contrast, integration into the transcriptionally active site results in high expression.

Thus, when the goal of the work is to obtain high levels of gene expression, which is also typically a desired outcome of genetic engineering methods, it is generally necessary to screen large numbers of transfectants to find such high-yielding clones. In addition, random integration of foreign DNA into the genome can in some cases disrupt important cellular genes, resulting in phenotypic changes. These factors complicate and laborious the generation of highly expressed stable mammalian cell lines.

Recently, our laboratory has described the use of DNA vectors containing dominant selectable markers of impaired translation for mammalian gene expression (published in U.S. serial No. 08/147,696 filed on 3.11.1993, recently granted).

These vectors contain a translation-impaired neomycin phosphotransferase (neo) gene as a dominant selectable marker, which has been artificially engineered to contain an intron into which the DHFR gene and one or more desired genes have been inserted. We have found that the use of these vectors as expression constructs significantly reduces the total number of drug resistant colonies generated, relative to conventional mammalian expression vectors, thereby facilitating screening. In addition, a higher proportion of clones obtained using this system are high expressing clones. These results are clearly attributable to the modifications made to the neo selectable marker. Due to impaired translation of the neo gene, transfected cells are unable to produce enough neo protein to survive drug selection, thereby reducing the total number of drug resistant colonies. In addition, a high percentage of surviving clones contain expression vectors integrated into the genomic locus where basal transcription levels are high, leading to overproduction of neo, thus enabling cells to overcome the problems associated with damaged neo genes. At the same time, the desired gene linked to neo will similarly increase the level of transcription. The result of generating artificial introns within neo makes this same advantage true; survival is dependent on the synthesis of a functional neo gene, which in turn is dependent on the correct and efficient splicing of the neo intron. In addition, these criteria are more likely to be met if the vector DNA is integrated into a region already having high transcriptional activity.

After integration of the vector into the transcription active region, gene amplification was performed by DHFR gene selection. Using this system, clones selected with low levels of methotrexate (50nM) can be obtained that contain several (< 10) copies of the vector secreting high levels (> 55 pg/cell/day) of protein. In addition, the above clones can be obtained in a relatively short time. However, the success of amplification is variable. Some transcriptionally active sites cannot be amplified, and therefore, the frequency and extent of amplification from a particular site is unpredictable.

Overall, the approach of using these translation-impaired vectors is significantly improved over other approaches that incorporate randomly. However, as mentioned above, the problem of lack of control over the integration site remains significant.

One approach to overcoming the problem of random integration is to utilize gene targeting whereby the foreign DNA is targeted to a specific locus within the host genome. The foreign DNA is inserted by means of homologous recombination taking place between the DNA sequence in the expression vector and the corresponding homologous sequence in the genome. However, while the frequency of such recombination occurring naturally is high in yeast and other fungal organisms, the frequency of such recombination is extremely low in higher eukaryotic organisms. Homologous versus non-homologous (random integration) recombination has been reported to have a frequency in the range of 1/100 to 1/5000 in mammalian cells (see, e.g., Capecchi, science, 244: 1288-Bush 1292 (1989); Morrow and Kucherlapati, Curr. Op. Biotech., 4: 577-Bush 582 (1993)).

One of the earliest reports describing homologous recombination in mammalian cells included an artificial system generated in mouse fibroblasts (Thomas et al, cell 44: 419-428 (1986)). A cell line is generated that contains a mutated, non-functional form of the neo gene integrated into the host genome, which is then targeted with a second, non-functional copy of the neo gene containing a different mutation. Functional neo gene reconstitution can only be performed by gene targeting. Homologous recombinants were identified by selection of G418 resistant cells, which were further confirmed by analysis of genomic DNA isolated from resistant clones.

Recently, the use of homologous recombination to replace heavy and light chain immunoglobulin genes at endogenous loci in antibody-secreting cells has been reported (U.S. Pat. No. 5,202,238, Fell et al, 1993). However, this particular method has not been widely used because the method is limited to the production of immunoglobulins in cells that endogenously express immunoglobulins, such as B cells and myeloma cells. Expression was also limited to single copy gene levels due to failure to include co-amplification following homologous recombination. The production of functional immunoglobulins requires two separate integration events: the fact that one is used for the light chain gene and one is then used for the heavy chain gene further complicates the method.

Another example of this type of system is reported in NS/0 cells, where recombinant immunoglobulins are expressed by homologous recombination within the immunoglobulin gamma 2A locus (Hollis et al, International patent application # PCT/IB95 (00014)). The expression level from this site was very high, with a single copy of the integrant of approximately 20 pg/cell/day. However, as in the above example, expression is limited to this level because the amplifiable gene is not co-integrated into this system. It has been reported by others that glycosylation of recombinant proteins expressed in NS/0 cells is aberrant (see, for example, Flesher et al, Biotech. andBioeng., 48: 399-.

Recently, the cre-loxP recombination system of bacteriophage P1 was engineered and used as a tool for gene targeting in eukaryotic cells. In particular, the site-specific integration of foreign DNA into the genome of Chinese Hamster Ovary (CHO) cells using cre recombinase and a series of lox-containing vectors is described (Fukushige and Sauer, Proc. Natl. Acad. Sci. USA, 89: 7905-. The attractiveness of this system is that it provides reproducible expression at the same chromosomal location. However, no effort was made to identify the chromosomal site most suitable for gene expression, and, as in the above example, expression in this system was limited to a single copy level. The problem is also complicated by the fact that a functional recombinase needs to be expressed in mammalian cells.

The use of homologous recombination between the introduced DNA sequence and its endogenous chromosomal locus has been reported to provide a useful tool for genetic manipulation in mammalian cells as well as yeast cells (see, e.g., Bradley et al, meth.enzymol., 223: 855-. To date, most mammalian gene targeting studies have been directed to gene disruption ("knockout") or site-specific mutagenesis of a selected target gene locus in mouse Embryonic Stem (ES) cells. The generation of these "knockout" mouse models has enabled scientists to examine specific structure-function relationships and examine the biological importance of a large number of mouse genes. This field of research is also of great relevance for potential gene therapy applications.

Recently, Celltech (Kent, U.K.) also reported vectors that are said to target the transcriptionally active sites of NSO cells that do not require gene amplification (Peakman et al, hum. antibody. hybrids, 5: 65-74 (1994)). However, immunoglobulin secretion levels in these unexpanded cells were reported to be no more than 20 pg/cell/day, whereas in expanded CHO cells levels were as high as 100 pg/cell/day (supra).

It is highly desirable to develop a gene targeting system that will reproducibly integrate foreign DNA into a predetermined site in the genome that is known to be transcriptionally active. It is also necessary to have such gene targeting systems further facilitate co-amplification of inserted DNA after integration. The design of such a system should allow reproducible and high level expression of any desired cloned gene in mammalian cells, which is of no doubt of great interest to many researchers.

In this application, we provide a novel mammalian expression system based on homologous recombination occurring between two artificial substrates (substrates) contained in two different vectors. In particular, this system uses two novel mammalian expression vectors in combination, which are referred to as "marker" vectors and "targeting vectors".

Essentially, the marker vector can identify and mark sites in the mammalian genome that have transcriptional activity, i.e., sites that have high levels of gene expression. This site can be considered a "hot spot" in the genome. Following integration of the marker vector, the expression system of the present application utilizes homologous recombination between DNA sequences common to both vectors to integrate another DNA into this site, the targeting vector. This system is significantly superior to other homologous recombination systems.

Unlike most other homologous systems used in mammalian cells, this system does not exhibit background, and thus, cells that are only randomly integrated by the vector do not survive selection, and any desired gene cloned into the targeting plasmid begins at a high level of expression from the hot spot of the marker. Thus, the gene expression method of the present invention substantially or completely eliminates the problems inherent in the random integration system discussed in detail above. In addition, this system enables reproducible and high-level expression of any recombinant protein at the same transcriptionally active site in the genome of a mammal. By making the amplifiable dominant selection marker (e.g., DHFR) part of the marker vector, gene amplification can be achieved at this particular transcriptional active site.

Object of the Invention

It is therefore an object of the present invention to provide an improved method for targeting a desired DNA to a specific site in a mammalian cell.

It is another specific object of the invention to provide a novel method to target a desired DNA to a specific site of a mammalian cell via homologous recombination.

It is another specific object of the present invention to provide a novel vector for site-specific integration of a desired DNA into mammalian cells.

It is another specific object of the invention to provide a novel mammalian cell line containing the desired DNA integrated into a predetermined site providing high expression.

Another specific object of the present invention is to provide a novel method for site-specific integration of desired DNA in Chinese Hamster Ovary (CHO) cells.

It is another specific object of the invention to provide a novel method for integrating immunoglobulin genes, or any other gene, into a predetermined chromosomal site in mammalian cells that provides high expression.

It is another specific object of the invention to provide novel vectors and vector combinations suitable for integrating immunoglobulin genes into predetermined sites in mammalian cells that provide high expression.

It is another specific object of the invention to provide a novel mammalian cell line comprising immunoglobulin genes integrated into predetermined sites providing high expression.

It is another specific object of the present invention to provide a novel method for integrating immunoglobulin genes into CHO cells providing high expression, and to provide novel vectors and vector combinations for integrating immunoglobulin genes into CHO cells.

In addition, it is another specific object of the present invention to provide a novel CHO cell line containing immunoglobulin genes integrated into predetermined sites providing high expression, which has been selectively amplified by methotrexate to secrete more functional immunoglobulins.

In particular, the invention relates to:

1) a method of inserting a desired DNA at a target site in the genome of an isolated mammalian cell, the method comprising the steps of:

(i) transfecting or transforming an isolated mammalian cell with a marker plasmid, said plasmid comprising the following sequences:

(a) a DNA region that is heterologous to the mammalian cell genome and provides a unique site for homologous recombination when integrated into the mammalian cell genome;

(b) a first portion of a DNA segment encoding a first selectable marker protein portion; and

(c) at least one other selectable marker DNA which aids in the selection of mammalian cells successfully integrated with the marker plasmid;

(ii) selecting cells containing a marker plasmid, which has integrated into the genome of the cells, by screening for at least one other selectable marker;

(iii) transfecting or transforming the selected cells with a targeting plasmid, the plasmid comprising the following sequences:

(a) a DNA region that is identical to or has sufficient homology with a unique region in the marker plasmid such that the DNA region can recombine with the unique region via homologous recombination;

(b) a second portion of a DNA fragment encoding the same portion of the selectable marker as contained in the marker plasmid, wherein the first selectable marker protein is produced only when the second portion is expressed in combination with the first portion of the first selectable marker protein; and

(iv) cells containing the targeting plasmid are selected by screening for expression of the first selectable marker protein, which integrates into the target site.

2) The method of 1) above, wherein the DNA segment encoding the first selectable marker segment is an exon of a dominant selectable marker.

3) The method of 2) above, wherein the targeting plasmid contains the remaining exons of said first selectable marker.

4) The method of 3) above, wherein at least one DNA encoding the desired protein is inserted between exons of the first selectable marker contained in the targeting plasmid.

5) The method of 4) above, wherein a DNA encoding a dominant selectable marker is also inserted between exons of the first selectable marker contained in the targeting plasmid to co-amplify the DNA encoding the desired protein.

6) The method of 3) above, wherein the first dominant selectable marker is selected from the group consisting of neomycin phosphotransferase, histidinol dehydrogenase, dihydrofolate reductase, hygromycin phosphotransferase, herpes simplex virus thymidine kinase, adenosine deaminase, glutamine synthetase and hypoxanthine guanine phosphoribosyl transferase.

7) The method of the above 4), wherein the desired protein is a mammalian protein.

8) The method of 7) above, wherein the protein is an immunoglobulin.

9) The method of 1) above, further comprising determining the level of RNA in the DNA of step (i) (c) of claim 1 contained in the marker plasmid prior to integrating the targeting vector.

10) The method of the above 9), wherein the other selectable marker contained in the marker plasmid is a dominant selectable marker selected from the group consisting of histidinol dehydrogenase, herpes simplex virus thymidine kinase, hygromycin phosphotransferase, adenosine deaminase and glutamine synthetase.

11) The method of 1) above, wherein the mammalian cell is selected from the group consisting of Chinese hamster ovary cells, myeloma cells, baby hamster kidney cells, COS cells, NSO cells, HeLa cells and NIH3T3 cells.

12) The method of 11) above, wherein the cell is a CHO cell.

13) The method of 1) above, wherein the marker plasmid contains the third exon of the neomycin phosphotransferase gene and the targeting plasmid contains the first two exons of the neomycin phosphotransferase gene.

14) The method of 1) above, wherein the marker plasmid further comprises an unusual restriction endonuclease sequence inserted into the region of homology.

15) The method of 1) above, wherein the unique DNA region providing for homologous recombination is bacterial DNA, viral DNA or synthetic DNA.

16) The method of 1) above, wherein the unique DNA region providing homologous recombination is at least 300 nucleotides.

17) The method of 16) above, wherein the unique DNA region ranges in size from about 300 nucleotides to 20 kilobases.

18) The method of 17) above, wherein the size of the unique DNA region ranges from 2 to 10 kilobases.

19) The method of 1) above, wherein the first selectable marker DNA is divided into at least 3 exons.

20) The method of 1) above, wherein the unique DNA region providing homologous recombination is bacterial DNA, insect DNA, viral DNA or synthetic DNA.

21) The method of 20) above, wherein the unique DNA region does not contain any functional gene.

22) A vector system for inserting a desired DNA into a genomic target site of a mammalian cell, the system comprising at least the following:

(i) a marker plasmid comprising at least the following sequence:

(a) a DNA region that is heterologous to the mammalian cell genome and provides a unique site for homologous recombination when integrated into the mammalian cell genome;

(b) a DNA segment encoding a portion of a first selectable marker protein; and

(c) at least one other selectable marker DNA which aids in the selection of mammalian cells successfully integrated with the marker plasmid; and

(ii) a targeting plasmid comprising at least the following sequence:

(a) a DNA region which is identical to or has sufficient homology with a unique region in the marker plasmid to enable said DNA region to recombine with said unique region via homologous recombination;

(b) a DNA fragment encoding the same portion of the selectable marker as contained in the marker plasmid, wherein an active selectable marker protein encoded by said DNA is produced only when said fragment is expressed in combination with said selectable marker DNA fragment contained in the marker plasmid.

23) The vector system of 22) above, wherein the DNA segment encoding a portion of the first selectable marker protein is an exon of the dominant selectable marker.

24) The vector system of 23) above, wherein the targeting plasmid comprises the remaining exons of said first selectable marker.

25) The vector system of 24) above, wherein at least one DNA encoding a desired protein is inserted between exons of the first selectable marker contained in the targeting plasmid.

26) The vector system of 24) above, wherein a DNA encoding a dominant selectable marker is also inserted between exons of the first selectable marker contained in the targeting plasmid to co-amplify the DNA encoding the desired protein.

27) The vector system of 24) above, wherein the first dominant selectable marker is selected from the group consisting of neomycin phosphotransferase, histidinol dehydrogenase, dihydrofolate reductase, hygromycin phosphotransferase, herpes simplex virus thymidine kinase, adenosine deaminase, glutamine synthetase and hypoxanthine guanine phosphoribosyl transferase.

28) The vector system of 25) above, wherein the desired protein is a mammalian protein.

29) The vector system of 28) above, wherein the protein is an immunoglobulin.

30) The vector system of 22) above, wherein the other selectable marker contained in the marker plasmid is a dominant selectable marker selected from the group consisting of histidinol dehydrogenase, herpes simplex virus thymidine kinase, hygromycin phosphotransferase, adenosine deaminase and glutamine synthetase.

31) The vector system of 22) above, which facilitates insertion of a desired cell into a target site of a genome of a mammalian cell selected from the group consisting of chinese hamster ovary cells, myeloma cells, baby hamster kidney cells, COS cells, NSO cells, HeLa cells and NIH3T3 cells.

32) The vector system of 31) above, wherein the mammalian cell is a CHO cell.

33) The vector system of 22) above, wherein the marker plasmid contains the third exon of the neomycin phosphotransferase gene and the targeting plasmid contains the first two exons of the neomycin phosphotransferase gene.

34) The vector system of 22) above, wherein the marker plasmid further comprises an unusual restriction endonuclease sequence, which is inserted into the region of homology.

35) The vector system of 22) above, wherein the unique DNA region providing for homologous recombination is bacterial DNA, viral DNA or synthetic DNA.

36) The vector system of 22) above, wherein the unique DNA region of (i) (a) in claim 1 that provides homologous recombination contained in the tagged plasmid vector system is at least 300 nucleotides.

37) The vector system of 36) above, wherein the unique DNA region ranges in size from about 300 nucleotides to about 20 kilobases.

38) The vector system of 37) above, wherein the unique DNA region ranges in size from 2 to 10 kilobases.

39) The vector system of 22) above, wherein the first selectable marker DNA is divided into at least 3 exons.

40) The vector system of 22) above, wherein the unique DNA region providing for homologous recombination is bacterial DNA, insect DNA, viral DNA or synthetic DNA.

41) The vector system of 40) above, wherein the unique DNA region does not contain any functional gene.

Brief Description of Drawings

FIG. 1 shows a map of a marker plasmid referred to as Desmond in the present invention. The plasmids are represented in circular form (1A) and in linearized form (1B) for transfection.

FIG. 2(A) shows a map of a targeting plasmid called "Molly". Illustratively, Molly encodes an anti-CD 20 immunoglobulin gene, the expression of which is described in example 1.

FIG. 2(B) shows the linearized version of Molly after digestion with the restriction enzymes KpnI and PacI. This linearized form can be used for transfection.

FIG. 3 shows a potential alignment between the Desmond sequence integrated into the CHO genome and the incoming targeted Molly sequence. One possible arrangement of Molly integrated into Desmond following homologous recombination is also shown.

FIG. 4 shows a Southern analysis of single copy Desmond clones. The samples were as follows:

lane 1: lambda HindIII DNA size marker

Lane 2: desmond clone 10F3

Lane 3: desmond clone 10C12

Lane 4: desmond clone 15C9

Lane 5: desmond clone 14B5

Lane 6: desmond clone 9B2

FIG. 5 shows Northern analysis of single copy Desmond clones. The samples were as follows: a series: northern probed with CAD and DHFR probes as shown. B series: as shown, duplicate Northern probed with CAD and HisD probes. The RNA samples loaded in the A and B series were as follows: lane 1: clone 9B2, lane 2: clone 10C12, lane 3: clone 14B5, lane 4: clone 15C9, lane 5: control RNA from CHO transfected with plasmids containing HisD and DHFR, lane 6: untransfected CHO.

FIG. 6 shows Southern analysis of clones obtained by homologous integration of Molly into Desmond. The samples were as follows:

lane 1: lambda HindIII DNA size marker, lane 2: 20F4, lane 3: 5F9, lane 4: 21C7, lane 5: 24G2, lane 6: 25E1, lane 7: 28C9, lane 8: 29F9, lane 9: 39G11, lane 10: 42F9, lane 11: 50G10, lane 12: molly plasmid DNA linearized with BglII (top most band) and cut with BglII and KpnI (bottom band), lane 13: untransfected Desmond.

FIGS. 7A to 7X are sequence listings for Desmond.

Fig. 8A to 8X are sequence listing of Molly-containing anti-CD 20.

FIG. 9 is a map of the targeting plasmid "Mandy", which, as a graphical representation, encodes the anti-CD 23 gene, the expression of which is described in example 5.

FIGS. 10A to 10U are the "Mandy" sequence listing containing the anti-CD 23 gene described in example 5.

Detailed Description

The present invention provides a novel method for integrating a desired exogenous DNA into a target site within the genome of a mammalian cell via homologous recombination. The invention also provides novel vectors capable of site-specific integration of DNA into a target site in the genome of a mammalian cell.

More specifically, the cloning method of the present invention allows the site-specific integration of desired DNA into mammalian cells by transfecting the cells with a "marker plasmid" comprising a unique sequence foreign to the genome of the mammalian cells and providing a substrate for homologous recombination, followed by a "targeting plasmid" comprising a sequence homologous recombination with the unique sequence contained in the marker plasmid and further comprising the desired DNA to be integrated into the mammalian cells. The integrated DNA typically encodes a desired protein, such as an immunoglobulin or other secreted mammalian glycoprotein.

An exemplary homologous recombination system uses the neomycin phosphotransferase gene as a dominant selectable marker. The use of this particular marker is based on the following known observations:

(i) capable of targeting and restoring the function of the mutated form of the neo gene (described above), and

(ii) we developed a translation-impaired expression vector in which the neo gene was artificially prepared as two exons, and the desired gene was inserted into an intervening intron; the neo exon is correctly spliced and translated in vivo to produce a functional protein, thereby conferring G418 resistance to the resulting cell population. In the present application, the neo gene is divided into 3 exons. The third neo exon is present on the "marker" plasmid, which is subsequently integrated into the host cell genome by integration of the marker plasmid into the mammalian cell. Exons 1 and 2 are present in the targeting plasmid and are separated by an intervening intron into which at least one desired gene is cloned. Homologous recombination between the targeting plasmid and the integrated marker plasmid allows the correct splicing of all 3 exons of the neo gene, allowing the expression of a functional neo protein (as determined by selection of G418 resistant colonies). Prior to the design of the expression system of the present application, we have experimentally tested the functionality of this triple spliced neo construct in mammalian cells. The results of this control experiment indicate that all 3 neo exons were correctly spliced, thus demonstrating the feasibility of the present invention.

However, while the present invention exemplifies neo genes, and more particularly, a gene that is divided into three parts, the general methodology is also effective for other dominant selectable markers.

As described in detail below, the present invention provides many advantages to conventional methods of gene expression, including random integration and gene targeting methods. In particular, the invention provides methods for reproducibly site-specific integration of a desired DNA into a transcriptionally active region of a mammalian cell. In addition, the methods of the invention introduce artificial "homology" regions that can serve as unique substrates for homologous recombination and insertion of desired DNA, and thus, the efficacy of the invention does not require that the cell endogenously contain or express the particular DNA. The method is generally applicable to all mammalian cells and can be used to express recombinant proteins in any form.

The use of triple-spliced selection markers, such as the exemplified triple-spliced neo construct, ensures that all G418 resistant colonies generated are generated by homologous recombination events (random integrants do not produce a functional neo gene and therefore do not survive G418 selection). Thus, the present invention facilitates the screening for desired homologous recombination events. In addition, the frequency of further random integration in cells which have undergone homologous recombination is low.

As described above, a significant advantage of the present invention is apparent in that the number of colonies that need to be screened to identify highly productive clones, i.e. cell lines containing the desired DNA secreting the corresponding protein at high levels, is significantly reduced. At a typical level, clones containing the integrated desired DNA are identified by screening about 5 to 20 colonies (thousands of colonies are screened using standard random integration techniques and hundreds of colonies are screened using the intron insertion vectors described above). In addition, since the integration site is preselected and contains a transcriptionally active region, all exogenous DNA expressed at this site should be capable of producing a substantial, i.e., high level, of the desired protein.

In addition, the present invention is advantageous in that an amplifiable gene is inserted by integration of a marker vector. Therefore, when the desired gene is targeted to this site via homologous recombination, the present invention further enhances gene expression by gene amplification. In this regard, it has been reported in the literature that the gene amplification capacity differs for different genomic sites (Meinkoth et al, molecular cell biology, 7: 1415-1424 (1987)). Therefore, this technique is also advantageous in that the desired gene can be placed at a specific site which is transcriptionally active and easily amplified, which should significantly reduce the time required to isolate such a high producing cell line.

In particular, while the conventional method of constructing a high expression mammalian cell line takes 6 to 9 months, the isolation of such clones of the present invention typically takes only about 3 to 6 months. This can be attributed to the fact that conventionally isolated clones generally must undergo at least three rounds of drug resistance gene amplification to achieve satisfactory gene expression levels. Since the clones produced by homologous recombination are generated from the preselected high expression sites, only a few rounds of amplification are required to achieve satisfactory levels of production.

The present invention also enables reproducible selection of high-producing clones which have incorporated therein a vector with a low copy number, typically a single copy number. This has the advantage that the stability of the clones is enhanced and other potentially harmful side effects associated with high copy numbers can be avoided. As described above, the homologous recombination system of the present invention uses a "marker plasmid" and a "targeting plasmid" in combination, which are described in detail below.

The "marker plasmid" used for marking and identifying transcription hot spots contains at least the following sequences:

(i) a DNA region that is heterologous or unique with respect to the mammalian cell genome, which DNA region can serve as a source of homology, allowing homologous recombination (with DNA contained in a second targeting plasmid) to occur. More specifically, the unique DNA region (i) typically contains bacterial, viral, yeast synthesized or other DNA not normally found in the genome of a mammalian cell that does not have significant homology or sequence identity to DNA contained in the genome of a mammalian cell. In essence, this sequence should be as different from mammalian DNA as possible so that it does not significantly recombine with the host cell genome via homologous recombination. The size of this unique DNA is generally at least about 2 to 10 kilobases, or more, and more preferably at least about 10kb, since several other researchers have suggested that the frequency of targeted recombination increases as the size of the homologous region increases (Capecchi, science 244: 1288-1292 (1989)).

The unique DNA described above can be used as a site for homologous recombination with sequences in a second targeting vector, the upper limit on size being largely constrained by potential stability constraints (if the DNA is too large, it cannot be easily integrated into the chromosome) and difficulties in handling very large DNA.

(ii) DNA comprising a selectable marker DNA segment, typically an exon of a dominant selectable marker gene. The only essential feature of this DNA is that it does not encode a functional selectable marker protein, except for its expression in association with the sequences contained in the targeting plasmid. Targeting plasmids typically contain the remaining exons of the dominant selectable marker gene (those not contained in the "targeting" plasmid). Essentially, a functional selectable marker is produced if homologous recombination occurs (resulting in the association and expression of the (i) sequence of this marker DNA with the portion of the selectable marker DNA fragment contained in the targeting plasmid).

As described above, the present invention exemplifies the use of a neomycin phosphotransferase gene "divided" into two vectors as a dominant selectable marker. However, other selectable markers are also suitable, such as the Salmonella histidinol dehydrogenase gene, hygromycin phosphotransferase gene, herpes simplex virus thymidine kinase gene, adenosine deaminase gene, glutamine synthetase gene and hypoxanthine guanine phosphoribosyl transferase gene.

(iii) (iii) DNA encoding a functional selectable marker protein, which selectable marker is different from selectable marker DNA (ii). The selectable marker allows the successful selection of mammalian cells in which the marker plasmid has successfully integrated into the cellular DNA. More preferably, the marker plasmid contains two such dominant selectable marker DNAs, which are placed at opposite ends of the vector. Advantageously, different selection agents can be used to select integrants, and cells containing the entire vector can also be selected. Alternatively, one marker may be an amplifiable marker as described above to facilitate gene amplification. (ii) Any of the dominant selectable markers listed in (a) can be used as well as other selectable markers known in the art.

In addition, the marker plasmid optionally contains an unusual endonuclease restriction site. This may be necessary because it facilitates lysis. Such unusual restriction sites, if present, should be placed near the middle of the unique region that serves as a substrate for homologous recombination. Preferably, the sequence is at least about 12 nucleotides. It has been reported that introduction of double-stranded breaks by similar methodologies enhances the frequency of homologous recombination (Choulika et al, molecular cell biology, 15: 1968-1973 (1995)). However, the presence of this sequence is not essential.

A "targeting plasmid" contains at least the following sequences:

(1) a unique region of DNA which is identical to or has sufficient homology or sequence identity with a unique region of DNA contained in the marker plasmid to allow said DNA to be combined by homologous recombination with the unique region (i) of the marker plasmid. Suitable types of DNA are described above in the description of region (i) unique to DNA in the marker plasmid.

(2) The remaining exons of the dominant selection marker, one of which is identical to (ii) in the marker plasmid described above. The essential feature of this DNA fragment is that a functional (selectable) marker protein is produced only when the targeting plasmid is integrated via homologous recombination, which results in the integration of this DNA with other selectable marker DNA fragments contained in the marker plasmid, and that, in addition, the desired foreign DNA may be inserted into this DNA fragment. This DNA typically contains the remaining exons of the selectable marker DNA separated by introns. For example, the DNA may contain the first two exons of the neo gene, and the marker plasmid may contain the third exon (the third exon following the neo gene).

(3) The targeting plasmid may also contain a desired DNA, such as DNA encoding a desired polypeptide, preferably inserted into a selectable marker DNA fragment contained in the plasmid. The DNA is typically inserted into introns contained between exons of the selectable marker DNA. This ensures that the desired DNA is integrated when homologous recombination occurs between the targeting plasmid and the marker plasmid. The intron can be native or engineered into a dominant selectable marker DNA segment.

The DNA may encode any desired protein, preferably a protein having a drug or other desired property. Most preferably, the DNA encodes a mammalian protein, in the examples described herein an immunoglobulin or immunoadhesin. However, the present invention is not limited to the production of immunoglobulins.

As mentioned above, the cloning method of the present invention is applicable to any mammalian cell, as its potency does not require the presence of any specific mammalian sequences. Such mammalian cells generally include cells commonly used for protein expression, such as CHO cells, myeloma cells, COS cells, BHK cells, Sp2/0 cells, NIH3T3 cells, and HeLa cells. In the following examples, CHO cells were used. Advantages include the availability of appropriate media, the ability to grow efficiently and achieve high culture densities, and the ability to express mammalian proteins, such as immunoglobulins, in biologically active forms.

In addition, CHO cells were selected largely because the inventors previously used such cells to express immunoglobulins (using the above-described vectors, which contain a dominant selection marker with impaired translation). Thus, the present laboratory has a great deal of experience with the use of such cells for expression. However, it is reasonably expected that similar results will be obtained with other mammalian cells, according to the following examples.

Generally, transformation or transfection of mammalian cells of the invention is accomplished according to conventional methods. In order to better understand the present invention, the construction of exemplary vectors and their use to generate integrants is described in the examples below.

Example 1

Design and preparation of tagged and targeted plasmid DNA vectors

The marker plasmid, referred to herein as "Desmond", is assembled from the following DNA elements:

(a) mouse dihydrofolate reductase gene (DHFR)Incorporated into the transcription cassette and containing the mouse beta globin promoter 5 'to the DHFR start site and the bovine growth hormone polyadenylation signal 3' to the DHFR stop codon. The DHFR transcription cassette was isolated from the expression vector TCAE6 previously constructed in this laboratory (Newman et al, 1992, Biotechnology, 10: 1455-1460).

(b) E.coli beta-galactosidase gene, commercially available from Promega under the trade name pSV-b-galactosidase control vector, Cat # E1081.

(c) Baculovirus DNA, commercially available from Clontech under the trade name pBAKPAK8, catalog # 6145-1.

(d) Cassettes containing cytomegalovirus and SV40 viral promoter and enhancer elements were generated by PCR using derivatives of the expression vector TCAE8 (Reff et al, blood, 83: 435-445 (1994)). The enhancer cassette is inserted into a baculovirus sequence which has been previously modified by insertion of a multiple cloning site.

(e) The E.coli GUS (glucuronidase) gene, commercially available from Clontech under the trade name pB101, catalog # 6017-1.

(f) The firefly luciferase gene, commercially available from Promega, is under the trade name pGEM-Luc (Cat # E1541).

(g) The Salmonella typhimurium histidinol dehydrogenase gene (HisD), which was originally a gift (Donahue et al, Gene 18: 47-59(1982)), was subsequently incorporated into a transcription cassette with the mouse beta globin major promoter at the 5 'end and the SV40 polyadenylation signal at the 3' end.

(a) The DNA elements described in (g) were constructed in the plasmid backbone derived from pBR, resulting in a contiguous DNA sequence of 7.7kb, referred to as "homology" in the drawing. Homology in this sense refers to DNA sequences that are part of the non-mammalian genome and that can be used to promote homologous recombination between transfected plasmids sharing the same DNA sequence of homology.

(h) The neomycin phosphotransferase gene of TN5 (Davis and Smith, Ann. Rev. Micro., 32: 469-518 (1978)). The complete neo gene was subcloned into pBluescript SK- (Stratagene Cat #212205) to facilitate gene manipulation. The synthetic linker was then inserted into a unique PstI site located at neo amino acid codons 51 and 52. This linker encodes the DNA elements necessary to create the artificial splice donor site, the intervening intron, and the splice acceptor site within the neo gene, thus creating two separate exons, referred to herein as neo exons 1 and 2. Exon 1 of neo encodes the first 51 amino acids of neo, while exon 2 encodes the remaining 203 amino acids plus the stop codon of the protein. An NotI cloning site is also created within the intron.

neo exon 2 was further subdivided, yielding neo exons 2 and 3. The method comprises the following steps: a set of PCR primers was designed to amplify a DNA region encoding the first 1112/3 amino acids of neo exon 1, intron and exon 2. The 3 'PCR primer can introduce a new 5' splice site immediately after the second nucleotide of exon 2 amino acid 111 codon, thereby generating a new smaller exon 2. The DNA fragments encoding the original exon 1, intron and new exon 2 were then subcloned into a pBR-based vector and propagated therein. The remainder of the original exon 2 was used as a template for another round of PCR amplification that produced "exon 3". The 5 'primer of this round of PCR amplification introduced a new splice acceptor site 5' to the newly generated exon 3, i.e., the last nucleotide of the codon at amino acid 111. The resulting 3 neo exons encode the following information: the first 51 amino acids of exons 1-neo; exon 2-the next 1112/3 amino acids; exon 3-the last 911/3 amino acids plus the translation stop codon of the neo gene.

neo exon 3 is incorporated into the marker plasmid "Desmond" along with the DNA elements described above. neo exons 1 and 2 are incorporated into the targeting plasmid "Molly". The NotI cloning site generated within the intron between exons 1 and 2 is used in a subsequent cloning step to insert the desired gene into the targeting plasmid.

A second targeting plasmid, "Mandy", was also generated, which is nearly identical to "Molly" (some of the restriction sites on the vector have been changed), except that the original HisD and DHFR genes contained in "Molly" were inactivated. The reason for incorporating these changes is that the Desmond cell line is no longer cultured in the presence of histidinol and therefore does not seem to necessarily include a second copy of the HisD gene. In addition, the DHFR gene is inactivated to ensure that only a single DHFR gene, i.e. the one present at the Desmond marker site, is amplified in any final cell line. "Mandy" was derived from "Molly" by the following modifications:

(i) a synthetic linker was inserted in the middle of the DHFR coding region, which created a stop codon and removed the remainder of the DHFR coding region in reading frame, thereby rendering the gene non-functional.

(ii) Part of the HisD gene was deleted and replaced by a PCR generated HisD fragment which lacked the promoter and start codon of the gene.

FIG. 1 shows the arrangement of these DNA elements in the marker plasmid "Desmond", FIG. 2 shows the arrangement of these elements in the first targeting plasmid "Molly", and FIG. 3 shows the possible arrangement of multiple DNA elements in the CHO genome after Molly DNA targeting and integration into Desmond-tagged CHO cells. FIG. 9 shows the targeting plasmid "Mandy".

Marker and targeting plasmids can be constructed from the above-described DNA elements according to conventional cloning techniques (see, for example, molecular cloning, A laboratory Manual, J.Sambrook et al, 1987, Cold spring harbor laboratory Press, latest methods in molecular biology, ed. F.M. Ausubel et al, 1987, John Wiley and Sons). All plasmids were propagated and maintained in E.coli XLI blue (Stratagene, Cat # 200236). A large number of plasmid preparations were prepared according to the manufacturer's instructions using the Promega Wizard Maxiprep DNA purification System.

Example 2

Construction of a labeled CHO cell line

1. Cell culture and transfection method for generating labeled CHO cell line

The labeled plasmid DNA was linearized by digestion with Bst1107I overnight at 37 ℃. The linearized vector was precipitated with ethanol and resuspended in sterile TE to a concentration of 1 mg/ml. The linearized vector was introduced into DHFR-Chinese hamster ovary cells (CHO cells), DG44 cells (Urlaub et al, Som.cell and mol.Gen., 12: 555-566(1986)) by electroporation as follows.

Exponentially growing cells were harvested by centrifugation, washed 1 time with ice-cold SBS (sucrose buffer, 272mM sucrose, 7mM sodium phosphate, pH7.4, 1mM magnesium chloride) and resuspended in SBS at a concentration of 10⁷Individual cells/ml. After incubation on ice for 15 minutes, 0.4ml of the cell suspension was mixed with 40. mu.g of linearized DNA in a disposable electrophoresis chamber and the cells were electrically stimulated using a BTX electric cell manipulator (San Diego, Calif.) set at a voltage of 230 volts, a capacitance of 400 microfarads and a resistance of 13 ohms. The electrically stimulated cells were then mixed with 20ml of pre-warmed CHO growth medium (CHO-S-SFMII, Gibco/BRL, Cat # 31033-. 48 hours after electroporation, selection medium was added to the plates (selection medium, CHO-S-SFMII without hypoxanthine or thymidine, to which 2mM histidinol (Sigma Cat # H6647) was added when transfected with Desmond). The plates were maintained in selection medium for 30 days, or until some wells showed cell growth. These cells were then removed from the 96-well culture plate and finally expanded in a 120ml spinner flask, where they were maintained in selective medium throughout.

Example 3

Identification of labeled CHO cell lines

(a) Southern analysis

Genomic DNA was isolated from all stably growing Desmond-labeled CHO cells. DNA was isolated using the Invitrogen Easy _ DNA kit according to the manufacturer's instructions. Genomic DNA was then digested with HindIII overnight at 37 ℃ and subjected to Southern analysis using a probe generated by PCR, labeled with digoxigenin, and specific for the DHFR gene. Hybridization and washing were performed according to the manufacturer's instructions using Boehringer Mannheim's DIG easy hyb (catalog # 1603558) and a DIG wash and blocking buffer pack (catalog # 1585762). A DNA sample putatively containing a band hybridizing to the DHFR probe was a Desmond clone, which was generated from a single cell integrated with a single copy of the plasmid. These clones were retained for further analysis, and of the total 45 HisD resistant cell lines isolated, only 5 were single copy integrants. FIG. 4 shows a Southern blot containing these 5 single copy Desmond clones, the names of which are provided in the figure description.

(b) Northern analysis

Total RNA was isolated from all single copy Desmond clones using TRIzol reagent (Gibco/BRL Cat # 15596-. 10-20. mu.g of RNA from each clone was analyzed on a two-fold formaldehyde gel, and the resulting blots were probed with probes generated by PCR, labeled with digoxigenin, and specific for (i) DHFR messenger RNA, (ii) HisD messenger RNA, and (iii) CAD messenger RNA. CAD is a triple-functional protein involved in uridine biosynthesis (Wahl et al, J. Biochem., 254, 17: 8679-8689(1979)), which is equally expressed in all cell types. This is used as an internal control in the present invention to help quantify the RNA loading. The results of Northern analysis using the Boehringer Mannheim reagent hybridization and washing are shown in FIG. 5, from which the single copy Desmond clone, which shows the highest levels of HisD and DHFR messenger RNA, is clone 15C9, lane 4 in both series. This clone was termed a "labelled cell line" and was used for targeting experiments in CHO, examples of which are provided in the following section.

Example 4

Expression of anti-CD 20 antibody in Desmond-labeled CHO cells

C2B8 is a chimeric antibody recognizing the B cell surface antigen CD20 that we have previously cloned and expressed in our laboratories (Reff et al, blood, 83: 434-45 (1994)). The 4.1kb DNA fragment containing the C2B8 light and heavy chain genes and the necessary regulatory elements (eukaryotic promoter and polyadenylation signal) was inserted into an artificial intron located between exons 1 and 2 of the neo gene contained in the pBR-derived cloning vector. The newly generated 5kb DNA fragment (containing neo exon 1, C2B8 and neo exon 2) was excised and used to assemble the targeting plasmid Molly. The other DNA elements used to construct Molly are the same as those used to construct the marker plasmid Desmond identified above. The complete map of Molly is shown in FIG. 2.

Before transfection, the targeting plasmid Molly was linearized by digestion with KpnI and PacI, and the linearized plasmid was precipitated with ethanol and resuspended in sterile TE to a concentration of 1.5 mg/ml. Linearized plasmids were introduced into logarithmic growth Desmond-tagged cells essentially as described above, except that 80. mu.g of DNA was used in each electroporation. 48 hours after electroporation, the selection medium-CHO-SSFMII supplemented with 400. mu.g/ml Geneticin (G418, Gibco/BRL Cat #10131-019) was added to the 96 well plate. The plates were maintained in selection medium for 30 days, or until cell growth occurred in some wells. It is presumed that this growth is the result of clonal expansion of a single G418 resistant cell. The production of C2B8 in the supernatant of all G418 resistant wells was determined by standard ELISA techniques and finally all production clones were transferred to 120ml rotating culture flasks and further analyzed.

Identification of antibody-secreting target cells

In this experiment a total of 50 electroporation runs with the Molly targeting plasmid, each plated in a separate 96-well plate, resulting in a total of 10 viable anti-CD 20 antibody secreting clones that were transferred to 120ml rotating culture flasks. Genomic DNA was isolated from all clones, followed by Southern analysis to determine whether the clones represent a single homologous recombination event or whether additional random integration has occurred in the same cells. Methods for DNA isolation and Southern hybridization are described in the above section. Genomic DNA was digested with EcoRI and probed with a probe generated by PCR, labeled with digoxigenin, and specific for the CD20 heavy chain constant region segment. The results of this Southern analysis are shown in FIG. 6, where it can be seen that 8 of the 10 clones showed a band hybridizing to the CD20 probe, indicating that a single homologous recombination event occurred in these cells. 2 out of 10, clones 24G2 and 28C9 showed the presence of additional bands, indicating that additional random integrations occurred elsewhere in the genome.

We examined the expression levels of anti-CD 20 antibody in all 10 clones, the data of which are shown in table 1 below.

Table 1: expression levels of homologous integrants secreting anti-CD 20

Cloning anti-CD 20, pg/c/d

20F4 3.5

25E1 2.4

42F9 1.8

39G11 1.5

21C7 1.3

50G10 0.9

29F9 0.8

5F9 0.3

28C9* 4.5

24G2* 2.1

These clones contained other randomly integrated copies of anti-CD 20, so the expression levels of these clones reflected contributions from homologous and random sites.

Expression levels are expressed as picograms per cell secreted per day (pg/c/d) for each clone and represent the average level obtained from 3 independent ELISAs from samples taken from 120ml spinner flasks.

As can be seen from the above data, the secreted antibody differed approximately 10-fold between the most productive and least productive clones. To some extent, this was not expected, and we expected similar expression levels from all clones due to the fact that the anti-CD 20 gene was integrated into the same Desmond marker site. However, the observed changes in expression are very small compared to any traditional random integration method or our translation-impaired vector system.

The highest yielding single copy integrant, clone 20F4, was selected for further study. Table 2 (below) describes the ELISA and cell culture data for the 7 day production of this clone.

Table 2: 7 day production data for 20F4

Survival number of% survival/ml T × 2(hr) mg/l pg/c/d

(×10⁵)

1 96 3.4 31 1.3 4.9

2 94 6 29 2.5 3.4

3 94 9.9 33 4.7 3.2

4 90 17.4 30 6.8 3

5 73 14 8.3

6 17 3.5 9.5

On day 0, 2X 10⁵Clone 20F4 was inoculated into a 120ml spinner flask and, over the next 6 days, cell counts were performed, cell doubling times were counted, and 1ml of supernatant samples were removed from the flask and analyzed by ELISA for secreted anti-CD 20.

According to ELISA data, this clone secreted on average 3-5pg antibody/cell/day at the same level as we used other high expressing single copy clones with previously developed translation impaired random integration vectors. This result illustrates that:

(1) the site marked by the Desmond marker vector in the CHO genome has high transcriptional activity and is therefore an excellent site for the expression of recombinant proteins, and

(2) the use of the vectors of the invention allows the targeting by homologous recombination to occur at a frequency sufficiently high to make this system a viable and desirable alternative to the random integration approach.

To further elucidate the efficacy of this system, we also demonstrated that this site is amplifiable, leading to even higher levels of gene expression and protein secretion. By setting the density to 2.5X 10⁴20F4 cells per ml were serially diluted and plated onto 96-well tissue culture plates, and the cells were cultured in medium supplemented with 5, 10, 15 or 20nM methotrexate (CHO-SSFMII) for amplification. Clones secreting antibody were screened using standard ELISA techniques and the highest producing clones were amplified for further analysis. The summary of this amplification experiment is shown in Table 3 below.

Table 3: 20F4 amplification Profile

number of wells for nM MTX assay Table amplified number of wells in 96 wells spinner flask

# expression level of mg/l # pg/c/d

10 56 3-13 4 10-15

15 27 2-14 3 15-18

20 17 4-11 1 ND

Methotrexate amplification was performed on 20F4 using the concentrations indicated above, as described above. Supernatants from all surviving 96-well colonies were analyzed by ELISA and the range of anti-CD 20 expressed by these clones is shown in column 3. According to these results, the highest producing clones were transferred to 120ml spinner flasks and several ELISAs were performed on the supernatants of the spinner flasks to determine pg/cell/day expression levels, the results are shown in column 5.

This data clearly shows that: in the presence of methotrexate, this site can be amplified. Clones amplified with 10 and 15nM methotrexate achieved about 15-20 pg/cell/day.

The 15nM clone, designated 20F4-15A5, from a 96-well plate in which only 22 wells were grown was selected as the highest expressing cell line, and was therefore presumed to be produced by a single cell. This clone was subjected to a further round of methotrexate amplification. Serial dilutions of the culture were plated in 96-well plates and cultured in CHO-SS-FMII medium supplemented with 200, 300 or 400nM methotrexate as described above. Surviving clones were screened by ELISA, several highly productive clones were transferred to spinner flasks for culture and further analyzed. The profile of the second amplification experiment is shown in table 4.

Table 4: 20F4-15A5 amplification Profile

number of wells for nM MTX assay expression amplification in 96 wells Table in spinner flasks

# level mg/l # to level pg/c/d

200 67 23-70 1 50-60

250 86 21-70 4 55-60

300 81 15-75 3 40-50

Methotrexate amplification was established and analyzed as described above for 20F4-15A 5. The highest producing cells (the number of which is shown in column 4) were transferred to 120ml spinner flasks and the expression levels of the cell lines from these wells were expressed as pg/cell/day and the results are shown in column 5.

The highest producing clone was from 250nM methotrexate amplification, 250nM clone 20F4-15A5-250A6 from which only? Wells were grown in 96-well plates, so it was assumed that it was produced by single cells. The data of tables 3 and 4 are clearly illustrated together: two rounds of methotrexate amplification were sufficient to achieve expression levels of 60 pg/cell/day, approaching the maximum capacity of mammalian cells to secrete immunoglobulins (Reff, M.E.Curr.Opin.Biotech., 4: 573-576 (1993)). The ability to achieve this level of secretion with only two amplification steps further enhances the utility of this homologous recombination system. Random integration methods generally require more than two amplification steps to achieve this level, and are difficult to amplify. Therefore, homologous systems are a more efficient and time-saving method of obtaining high levels of gene expression in mammalian cells.

Example 5

Expression of anti-human CD23 antibody in Desmond-labeled CHO cells

CD23 is a low affinity IgE receptor that mediates the binding of IgE to B and T lymphocytes (Sutton, B.J., and Gould, H.J., Nature 366: 421-428 (1993)). Anti-human CD23 monoclonal antibody 5E8 is the human gamma-1 monoclonal antibody recently cloned and expressed in our laboratory. This antibody is disclosed in commonly assigned serial No. 08/803,085 filed on 20/2/1997.

Cloning of the heavy and light chain genes of 5E8 into the mammalian expression vector N5KG1 (derived from the vector NEOSPLA (Barnett et al, antibody expression and engineering, eds. H.Y.Yang and T.Imanaka, p27-40(1995)), according to recent observations, the immunoglobulin light chain secreted in other expression constructs in our laboratory was somewhat more abundant than the heavy chain (Reff et al, 1997, unpublished results), in an attempt to compensate for this deficit, we altered the 5E8 heavy chain gene by adding a strong promoter/enhancer element immediately upstream of the initiation site, in a subsequent step, a2.9 kb DNA fragment containing the 5E8 modified light and heavy chain genes was isolated from the N5KG1 vector and inserted into the targeting vector Mandy Molly containing 5E8 was prepared essentially as described in the previous section and electroporated into Desmond 15C9 CHO cells.

One modification to the above method is the type of medium used. Desmond-labeled CHO cells were cultured in protein-free CD-CHO medium (Gibco-BRL, Cat # AS21206) supplemented with 3mg/L recombinant insulin (3mg/ml stock solution, Gibco-BRL, Cat # AS22057) and 8mM L-glutamine (200mM stock solution, Gibco-BRL, Cat # 25030-081). Subsequently, transfected cells were selected in the above medium supplemented with 400. mu.g/ml geneticin. In this experiment, 20 electroporation were performed and plated in 96 well tissue culture plates. A total of 68 wells showed cell growth and secretion of anti-CD 23, which were presumed to be clones from a single G418 cell. 12 of these were transferred to 120ml spinner flasks for further analysis. We believe that a greater number of clones could be isolated in this experiment (68, as described in example 4, anti-CD 20 clones were only 10) due to the higher cloning efficiency and survival of cells grown in CD-CHO medium relative to CHO-SS-FMII medium. The expression levels of the spinner clones analyzed ranged from 0.5-3pg/c/d, very close to the expression level seen in the anti-CD 20 clone. The most highly productive anti-CD 23 clone, designated 4H12, was methotrexate amplified to increase its expression level. This amplification was established in a manner analogous to the anti-CD 20 clone in example 4. Logarithmic growth of 4H12 cells were serially diluted and plated in 96-well tissue culture plates and cultured in CD-CHO medium supplemented with 3mg/l insulin, 8mM glutamine and 30, 35 or 40nM methotrexate. The summary of this amplification experiment is shown in Table 5.

Table 5: 2H12 amplification Profile

Expression in Table-amplified well number spinner flasks in 96 wells of nM MTX assay

# to level mg/l # level pg/c/d

30 100 6-24 8 10-25

35 64 4-27 2 10-15

40 96 4-20 1 ND

The highest expressing clone obtained was a 30nM clone isolated from a plate in which only 22 wells had cells grown. This clone, designated 4H12-30G5, reproducibly secretes 18-22pg antibody/cell/day. This was in the same range of expression levels as the first amplification of anti-CD 20 clone 20F4 (clone 20F4-15A5 produced 15-18pg/c/d as described in example 4). This data further supports the observation that amplification at this marker site in CHO is reproducible and efficient. A second expansion of the 30nM cell line was recently expected to reach saturation expression levels of anti-CD 23 antibody with only two amplification steps, as was the case for anti-CD 20.

Example 6

Expression of immunoadhesins in Desmond-labeled CHO cells

CTLA-4 is a member of the Ig superfamily, which is located on the surface of T lymphocytes and is thought to play a role in antigen-specific T cell activation (Dariavach et al, European J Immunol, 18: 1901-; and Linsley et al, journal of experimental medicine, 174: 561-569(1991)). To further investigate the precise role of the CTLA-4 molecule in the activation pathway, a soluble fusion protein was generated containing the CTLA-4 extracellular domain linked to a truncated form of the human IgG1 constant region (Linsley et al (supra), we recently expressed the CTLA-4 Ig fusion protein in the mammalian expression vector BLECH1 (derived from plasmid nespla (Barnett et al, antibody expression and engineering, h.y.yang and t.imaka eds., p27-40 (1995)).

CTLA-4-Molly was prepared and electroporated into Desmond clone 15C9 CHO cells as described in the related example above for anti-CD 20. 12 electroporation were performed as described above and plated into 96 well tissue culture plates. 18 wells expressing CTLA-4 were isolated from 96 well plates and transferred to 120ml spinner flasks for culture. Genomic DNA isolated from each clone was then subjected to Southern analysis to determine how many homologous clones additionally contained random integrants. Genomic DNA was digested with BglII and probed with a probe generated by PCR, labeled with digoxigenin, and specific for the constant region of human IgG 1. The analysis results showed that 85% of the CTLA-4 clones were only homologous integrants; the remaining 15% additionally contained a random integrant. This result corroborates the findings discussed above in the expression of anti-CD 20, of which 80% of the clones were single homologous integrants. Thus, we can conclude that this expression system reproducibly produces a single targeted homologous integrant in at least 80% of the clones produced.

The expression levels of the cognate CTLA4-Ig clones ranged from 8-12 pg/cell/day, somewhat higher than the anti-CD 20 and anti-CD 23 antibody clones discussed above. However, we have previously observed that expression of this molecule using an intron inserted vector system also results in significantly higher expression levels than immunoglobulins. At present, we have not been able to provide an explanation of this observation.

Example 7

Targeting anti-CD 20 to other Desmond-tagged CHO cell lines

As discussed in the above section, we obtained 5 single copy Desmond-tagged CHO cell lines (see fig. 4 and 5). To illustrate that the success of the targeting strategy was not due to some unique property of Desmond clone 15C9, nor was the success limited to this clone, we introduced anti-CD 20Molly into Desmond clone 9B2 (Lane 6 in FIG. 4, Lane 1 in FIG. 5). Molly DNA was prepared and electroporated into Desmond9B2 as described in the related examples of anti-CD 20, supra. A homozygote was obtained from this experiment and this clone was transferred to a 120ml spinner flask, which yielded an average of 1.2pg anti-CD 20/cell/day. This is a considerably lower expression level compared to targeting Molly to Desmond 15C 9. However, this was the expected result based on Northern analysis of Desmond clones. As can be seen in FIG. 5, the mRNA level of clone 9B2 was much lower than that of 15C9, indicating that the site of this clone was not as high as the transcriptional activity in 15C 9. Thus, this experiment not only illustrates the reproducibility of this system-presuming that any tagged Desmond site can be targeted by Molly-but also confirms the Northern data that the site in Desmond 15C9 is most transcriptionally active.

In view of the foregoing disclosure, it will be appreciated that, although specific embodiments of the invention have been described herein for purposes of illustration, various modifications may be made without deviating from the scope of the invention. Accordingly, the invention is not limited except as by the appended claims.

Claims (41)

1. A method of inserting a desired DNA at a target site in the genome of an isolated mammalian cell, the method comprising the steps of:

(i) transfecting or transforming an isolated mammalian cell with a marker plasmid, said plasmid comprising the following sequences:

(a) a DNA region that is heterologous to the mammalian cell genome and provides a unique site for homologous recombination when integrated into the mammalian cell genome;

(b) a DNA segment encoding a first portion of a first selectable marker protein portion; and

(c) at least one other selectable marker DNA which aids in the selection of mammalian cells successfully integrated with the marker plasmid;

(ii) selecting cells containing a marker plasmid, which has integrated into the genome of the cells, by screening for at least one other selectable marker;

(iii) transfecting or transforming the selected cells with a targeting plasmid, the plasmid comprising the following sequences:

(a) a DNA region that is identical to or has sufficient homology with a unique region in the marker plasmid such that the DNA region can recombine with the unique region via homologous recombination;

(b) a DNA segment encoding a second portion of the same first selectable marker contained in a marker plasmid, wherein the first selectable marker protein is produced only when said second portion is expressed in combination with the first portion of the first selectable marker protein; and

(iv) cells containing the targeting plasmid are selected by screening for expression of the first selectable marker protein, which integrates into the target site.

2. The method of claim 1, wherein the DNA segment encoding the first selectable marker segment is an exon of a dominant selectable marker.

3. The method of claim 2, wherein the targeting plasmid contains the remaining exons of said first selectable marker.

4. The method of claim 3, wherein at least one DNA encoding a desired protein is inserted between exons of the first selectable marker contained in the targeting plasmid.

5. The method of claim 4, wherein DNA encoding a dominant selectable marker is also inserted between exons of the first selectable marker contained in the targeting plasmid to co-amplify DNA encoding the desired protein.

6. The method of claim 3, wherein the first dominant selectable marker is selected from the group consisting of neomycin phosphotransferase, histidinol dehydrogenase, dihydrofolate reductase, hygromycin phosphotransferase, herpes simplex virus thymidine kinase, adenosine deaminase, glutamine synthetase and hypoxanthine guanine phosphoribosyl transferase.

7. The method of claim 4, wherein the desired protein is a mammalian protein.

8. The method of claim 7, wherein the protein is an immunoglobulin.

9. The method of claim 1, further comprising determining the level of RNA in the DNA of step (i) (c) of claim 1 contained in the marker plasmid prior to integrating the targeting vector.

10. The method of claim 9, wherein the additional selectable marker contained in the marker plasmid is a dominant selectable marker selected from the group consisting of histidinol dehydrogenase, herpes simplex virus thymidine kinase, hygromycin phosphotransferase, adenosine deaminase, and glutamine synthetase.

11. The method of claim 1, wherein the mammalian cell is selected from the group consisting of chinese hamster ovary cells, myeloma cells, baby hamster kidney cells, COS cells, NSO cells, HeLa cells, and NIH3T3 cells.

12. The method of claim 11, wherein the cell is a CHO cell.

13. The method of claim 1, wherein the marker plasmid contains the third exon of the neomycin phosphotransferase gene and the targeting plasmid contains the first two exons of the neomycin phosphotransferase gene.

14. The method of claim 1, wherein the marker plasmid further comprises an unusual restriction endonuclease sequence inserted into the region of homology.

15. The method of claim 1, wherein the unique DNA region that provides for homologous recombination is bacterial DNA, viral DNA, or synthetic DNA.

16. The method of claim 1, wherein the unique DNA region that provides for homologous recombination is at least 300 nucleotides.

17. The method of claim 16, wherein the unique DNA region ranges in size from 300 nucleotides to 20 kilobases.

18. The method of claim 17, wherein the unique DNA region ranges in size from 2 to 10 kilobases.

19. The method of claim 1, wherein the first selectable marker DNA is divided into at least 3 exons.

20. The method of claim 1, wherein the unique DNA region that provides for homologous recombination is bacterial DNA, insect DNA, viral DNA, or synthetic DNA.

21. The method of claim 20, wherein the unique DNA region does not contain any functional genes.

22. A vector system for inserting a desired DNA into a genomic target site of a mammalian cell, the system comprising at least the following:

(i) a marker plasmid comprising at least the following sequence:

(a) a DNA region that is heterologous to the mammalian cell genome and which, when integrated into the mammalian cell genome, provides a unique site for homologous recombination;

(b) a DNA segment encoding a portion of a first selectable marker protein; and

(c) at least one other selectable marker DNA which aids in the selection of mammalian cells successfully integrated with the marker plasmid; and

(ii) a targeting plasmid comprising at least the following sequence:

(a) a DNA region which is identical to or has sufficient homology with a unique region in the marker plasmid to enable said DNA region to recombine with said unique region via homologous recombination;

(b) a DNA segment encoding a portion of the same first selectable marker contained in the marker plasmid, wherein an active selectable marker protein encoded by said DNA is produced only when said segment is expressed in combination with said selectable marker DNA segment contained in the marker plasmid.

23. The vector system of claim 22, wherein the DNA segment encoding a portion of the first selectable marker protein is an exon of a dominant selectable marker.

24. The vector system of claim 23, wherein the targeting plasmid contains the remaining exons of said first selectable marker.

25. The vector system of claim 24, wherein at least one DNA encoding a desired protein is inserted between exons of a first selectable marker contained in a targeting plasmid.

26. The vector system of claim 24, wherein DNA encoding a dominant selectable marker is also inserted between exons of the first selectable marker contained in the targeting plasmid to co-amplify the DNA encoding the desired protein.

27. The vector system of claim 24, wherein the first dominant selectable marker is selected from the group consisting of neomycin phosphotransferase, histidinol dehydrogenase, dihydrofolate reductase, hygromycin phosphotransferase, herpes simplex virus thymidine kinase, adenosine deaminase, glutamine synthetase and hypoxanthine guanine phosphoribosyl transferase.

28. The vector system of claim 25, wherein the desired protein is a mammalian protein.

29. The vector system of claim 28, wherein the protein is an immunoglobulin.

30. The vector system of claim 22, wherein the additional selectable marker contained in the marker plasmid is a dominant selectable marker selected from the group consisting of histidinol dehydrogenase, herpes simplex virus thymidine kinase, hygromycin phosphotransferase, adenosine deaminase, and glutamine synthetase.

31. The vector system of claim 22, which facilitates insertion of a desired cell into a target site of the genome of a mammalian cell selected from the group consisting of chinese hamster ovary cells, myeloma cells, baby hamster kidney cells, COS cells, NSO cells, HeLa cells, and NIH3T3 cells.

32. The vector system of claim 31, wherein the mammalian cell is a CHO cell.

33. The vector system of claim 22, wherein the marker plasmid contains the third exon of the neomycin phosphotransferase gene and the targeting plasmid contains the first two exons of the neomycin phosphotransferase gene.

34. The vector system of claim 22, wherein the marker plasmid further comprises an unusual restriction endonuclease sequence inserted into the region of homology.

35. The vector system of claim 22, wherein the unique DNA region that provides for homologous recombination is bacterial DNA, viral DNA, or synthetic DNA.

36. The vector system of claim 22, wherein the unique DNA region providing for homologous recombination contained in the marker plasmid vector system is at least 300 nucleotides.

37. The vector system of claim 36, wherein the unique DNA region ranges in size from 300 nucleotides to 20 kilobases.

38. The vector system of claim 37, wherein the unique DNA region ranges in size from 2 to 10 kilobases.

39. The vector system of claim 22, wherein the first selectable marker DNA is divided into at least 3 exons.

40. The vector system of claim 22, wherein the unique DNA region providing for homologous recombination is bacterial DNA, insect DNA, viral DNA, or synthetic DNA.

41. The vector system of claim 40, wherein the unique DNA region does not contain any functional genes.