EP0593631A1

EP0593631A1 - Interstitial deletion in chromosomal dna

Info

Publication number: EP0593631A1
Application number: EP92915315A
Authority: EP
Inventors: Raju Kucherlapati
Original assignee: Albert Einstein College of Medicine
Current assignee: Albert Einstein College of Medicine
Priority date: 1991-07-01
Filing date: 1992-06-26
Publication date: 1994-04-27
Also published as: AU2305892A; WO1993001292A1; EP0593631A4; JPH06508756A; CA2112673A1

Abstract

Méthodes et compositions permettant de réaliser des délétions, additions ou substitutions dans un ADN chromosomique. En particulier, une séquence présentant une homologie avec une séquence cible est employée, conjointement à une séquence présentant une homologie par rapport à une séquence répétitive. L'intégration de la construction dans un chromosome permet d'obtenir une séquence emboîtée de délétions en rapport avec le nombre et la distance des séquences répétitives par rapport à la séquence cible.Methods and compositions for carrying out deletions, additions or substitutions in chromosomal DNA. In particular, a sequence having homology with a target sequence is used, in conjunction with a sequence having homology with respect to a repeating sequence. The integration of the construction in a chromosome makes it possible to obtain a nested sequence of deletions in relation to the number and the distance of the repetitive sequences compared to the target sequence.

Description

INTERSTITIAL DELETION IN CHROMOSOMAL DNA

INTRODUCTION

Technical Field

The field of this invention is the directed modification of chromosomal DNA.

Background

The large size of the mammalian chromosome makes it a daunting task to determine the organization of the individual chromosomes. The human genome program, which has been inaugurated, has chosen to try to determine the entire sequence of the human genome. Part of this program is directed to employing the genomes of other eukaryotic organisms, which have simple genomes, such as yeast. In attempting to define the human genome and the spatial relationship of structural genes, numerous tools have been developed and will continue to be developed to aid in this major task.

Besides the interest in elucidating the mammalian chromosomal structure, there is also substantial interest in providing techniques for manipulating and modifying chromosomes. There has been much interest in homologous recombination ("HR") , which allows for modifications of the DNA at a predetermined site. These modifications may be any of a wide variety of alterations, such as insertions, deletions, substitutions, and combinations thereof. Thus, by homologous recombination one may knock out a particular gene, so as to prevent its expression, one may modify the gene, for example, by substituting for a mutation to provide for expression of a wild-type product, one may insert or exchange DNA, so as to substitute a constitutive promoter for a regulated promoter, or one may introduce an enhancer adjacent to a structural gene to increase the expression of the gene.

In addition, it has been observed that the frequency of homologous recombination in mammalian cells is highly variable, often dependent upon the particular locus in a particular cell type. Thus, it would be useful to vary one of the two homologous ends of the targeting construct, as introduction of a repetitive element, in effect does, in order to permit the targeting sequences to search for a recombinogenic site. These various modifications and others may be employed for genetic therapy, to modify cells for expression of desired products, to modify cells for use as therapeutic agents, as viable drugs, and the like. There is, therefore, substantial interest in providing techniques which will allow for controlled modification of chromosomes to achieve a particular purpose.

Relevant Literature

The introduction of YACs into mammalian cells has been described by Pavan et a_l. , (1990) Mol. Cell. Biol. JL0:4163-4169 and Pachnis et a_l. , (1990) Proc. Natl. Acad. Sci. USA. 2:5109-5113. The homology among different LI elements is reported by Larman et al,. , (1983) Ibid 80: 3966- 3970 and Adams et aL. , (1980) Nucleic Acids Res.. 8_.:6113- 6128. Thomas and Capecchi (1987) Cell, 5_1:503-512 describe Site-Directed Mutagenesis by Gene Targeting in Mouse Embryo- Derived Stem Cells. Song et al . , Proc. Natl. Acad. Sci. USA, £54.:6820-6824 (1987), describe Homologous Recombination in Human Cells by a Two Step Process. Rubnitz and Subramani (1984) Mol. Cell. Biol. 4_.:2253-2258, describe the minimum amount of homology required for homologous recombination in mammalian cells. SUMMARY OF THE INVENTION

Methods and compositions are provided for modifying mammalian chromosomes, particularly introduction of deletions, by employing constructs having a sequence homologous to a genomic target sequence and at least one sequence homologous to a repetitive sequence. By providing for homologous recombination with chromosomal DNA, clones may be obtained which have DNA deleted between the target sequence and a repetitive sequence in the chromosome. Conveniently, the lesion may be introduced in a YAC in yeast, and the YAC transformed into a mammalian cell for homologous recombination and replacement of the chromosomal

DNA with the modified YAC DNA.

DESCRIPTION OF THE SPECIFIC EMBODIMENTS DNA compositions, methods of using the compositions, and cellular compositions are provided for modifying chromosomal DNA. The DNA compositions are constructs which comprise a sequence homologous to a target sequence, a sequence homologous to a naturally repetitive sequence, and usually a bridge between the two sequences. The construct is used to modify chromosomal DNA in a target cellular host, particularly by introducing deletions, although other mutations may also be effected. The modification may occur in two steps. In a first step, a yeast artificial chromosome, ("YAC") comprising a large chromosomal DNA fragment is modified by recombination with the DNA construct. In a second step, the YAC or modified fragment is transformed into the target host and the resulting cells screened for the presence of the desired lesion.

The subject method finds use with any eukaryotic host cell, particularly with mammalian hosts, more particularly with primate hosts, e.g., man. The particular host is not critical to this invention, although the procedure may vary to some degree depending upon the target host, the complexity of the genome, the nature of repeats, and the like. As already indicated, the DNA construct will normally comprise at least two parts, more usually three parts: the homologous target DNA sequence; the homologous repetitive sequence; and a bridge. The homologous target DNA sequence may be to any target gene, or sequences that flank it, which gene is to be modified, particularly deleted. The target sequence may be a coding or non-coding sequence, may be an intron, exon, 5'- or 3'-untranslated region, or the like. The homologous target sequence may lie close to the gene, usually it will be within about 250 kb or more of the target gene and may be upstream or downstream, usually upstream or within the target gene. Thus, the target region will usually comprise the target gene and flanking sequences of about 250 kb or more. The target sequence of the construct will usually comprise at least a portion of the target region and may extend substantially beyond the target region.

The homologous target sequence will usually have at least about 70% homology, more usually at least about 80% homology, and may be 90% or more homologous. The homologous portion will comprise at least 15 nucleotides, more usually at least about 30 nucleotides, and may be 0.5 kb or more, usually being not more than about 5 kb.

The homologous repetitive sequence will usually be at least about 15 bp, more usually at least about 30 bp, and usually less than about 1 kbp, more usually less than about

0.5 kbp. The homology will usually be at least about 60%, more usually at least about 70%, and may be 90% or higher to a particular repetitive sequence. However, since the repetitive sequences vary, the homology of the construct sequence will also vary depending upon the sequence of the repetitive sequences in the area of the target region. The construct may comprise one or more sequences homologous to repetitive sequences, where the repetitive sequences may be the same or different. Usually, the construct will include not more than about five repetitive sequences which may be repeats or may be homologous to the different repetitive sequences encountered in the host genome. Repeti ive sequences of interest include the Alu,

LI, α-satellite, telomeric or subtelomeric repeat sequences found in man; and similar sequences found in other mammals.

The bridge, if present, may serve a variety of purposes. It may serve to separate the target homologous sequence and the repetitive homologous sequence. It may also include one or more markers, so that one may detect integration of the sequence by selecting for the marker. A wide variety of markers exist, which provide for resistance to various toxic agents, such as antibiotics, heavy metals, or the like; providing prototropy to an auxotrophic host; providing for various functional sequences, such as excision sequences, primer sites, origins of replication, etc. Where

,a circular construct is used, markers may be present on one or both bridges^' joining the homologous sequences. Where a linear construct is employed for transformation, usually the markers will be between the homologous sequences.

The construct may be part of a larger vector, which may include replication origins for cloning, viral functional sequences, which allow for transfection, integration, autonomous replication (ars) , segregation, and other chromosomal functions.

The construct which is employed for transformation, will usually be at least about 0.25 kbp, more usually at least about 0.5 kbp, and may be 10 kbp or greater, usually not more than about 20 kbp.

A wide variety of target genes may be of interest, depending upon the purpose of the construct. For example, target genes may include oncogenes, members of the immunoglobulin gene superfamily, particularly constant and joining regions of the immunoglobulin gene family, surface membrane receptors, major histocompatibility complex antigens, /3₂-inicroglobulin, enzymes, such as kinases, lipases, etc. or they may be non-coding parts of genes such as upstream regulatory elements.

The construct will be prepared in accordance with conventional ways. A vector may be employed which is capable of replication in a prokaryotic host and has the appropriate replication sequences and one or more markers for selection. Usually, the vector will also include polylinkers, for insertion of the various fragments of interest. The fragments will be obtained by any convenient means, such as restriction digestion from a natural source, synthesis, polymerase chain reaction ("PCR") , etc. and where the termini are inappropriate for insertion into the vector, may be modified by ligation with appropriate linkers, blunt end ligation, or the like. After each insertion, the resulting construct may be cloned and analyzed by restriction enzyme analysis, sequencing, or other convenient means. Once the construct is formed, it may be cloned and expanded and then modified as appropriate, frequently removing major portions of the prokaryotic DNA. Usually, the construct will be used in linear form, where the construct may form an Ω-vector, or a 0-vector. Thus, the homologous sequences when hybridized may have the bridge sequence between the 3'-terminus of the upstream homologous sequence and the 5'-terminus of the downstream homologous sequence or, alternatively, may be joined to the 5'-terminus of the upstream homologous sequence and the 3'-terminus of the downstream homologous sequence.

Depending upon whether a single- or double-step approach is employed, there will be one or two transformation steps. (By "transformation" is intended any means for introducing exogenous DNA into a viable cell while retaining viability. Thus, transformation includes conjugation, transformation, transfection, e.g., calcium phosphate coprecipitation, cell fusion, protoplast or spheroplast fusion, biolistics using DNA coated particles, lipofection, electroporation, microinjection, or the like.) The DNA may be single or double stranded, linear or circular. While direct introduction of the construct into the target host may be employed, by transforming the target host with the construct employing conventional ways, preferably, a two-step process will be employed. In the two-step process, target DNA will be cloned in a YAC in an appropriate yeast host. The clone may be an individual clone known to contain the target DNA, a select group of clones, or a library, as appropriate. Usually, the target DNA will be at least about 50 kbp, more usually at least about 200 kbp, and usually less than about 1000 kbp. Various yeast hosts may be employed, such as . cerevisiae, j3. po bi. etc. Conveniently, the host will be auxotrophic, in an least one metabolic pathway and may be auxotrophic for two or more metabolic pathways. At least one of the metabolic pathways will be complemented by a gene present on the YAC, so as to ensure maintenance of the YAC in the host. The remaining metabolic dependencies may be complemented upon insertion of the targeting vector into the YAC or providing for the necessary nutrients in the medium.

The source of the DNA, as the primary cell or DNA in the YAC, may be any mammalian cell of interest, more particularly primate cells, especially human cells, where the human cells may be normal cells, including embryonic, or neoplastic cells, particularly normal cells. Various cell types may be employed as the primary cells, including fibroblasts, particularly diploid skin fibroblasts, keratinocytes, myoblasts, lymphocytes, glia, epithelial cells, neurons, endothelial cells, or other somatic cells, or germ cells. Of particular interest are skin fibroblasts, myoblasts, T-cells, retinal pigmented epithelial cells, etc. , which can be readily propagated to provide for large numbers of normal cells, embryonic kidney cells, and the like. These cells may or may not be expressing the gene of interest. In those instances where the target gene is inducible or only expressed in certain differentiated cells, one may select cells in which the target gene is expressed, which may require immortalized cells capable of growth in culture.

The YACs are prepared in accordance with conventional ways. Genomic DNA is cleaved enzymatically, mechanically, or by other means to provide fragments which will usually have at least about 50 kbp, more usually at least about 100 kbp, conveniently at least about 200 kbp and usually not more than about 2,000 kbp, more usually not more than about 1,000 kbp. The genomic DNA is inserted into a YAC and then screened using appropriate probes for identifying the presence of the target gene. The presence of the YAC may be verified by a selective medium for the markers present on the YAC. Yeast cells containing a YAC or YAC library may be characterized by hybridization analyses or by the polymerase chain reaction (PCR) using primers. The identified YAC(s) may then be used for manipulation.

The YAC will normally be transferred from the original yeast host to a different yeast host which is convenient for manipulation. The new host will be a haploid or diploid strain having a plurality, usually at least 2, and may have 5 or more mutations in different genes which allow for selection by complementation. Total yeast DNA from the original yeast host may be transformed into yeast cells or spheroplasts yielding cells, which may serve as the hosts for the manipulations. The resulting transformants may be plated on selective media which selects against transformants lacking the complementation markers present on the YAC.

Alternatively, one may transfer the YAC by a genetic cross. The recipient host for manipulation will normally be either a haploid or diploid host having a genetic defect which will be complemented by the genotype of the original yeast strain. If diploid, the recipient host is sporulated and ascospores are mixed with the original yeast host on an appropriate medium to allow mating. If haploid and of opposite mating type to the original host strain, cells may be mated directly. Hybrid diploids are selected on selective media, where only cross-hybrids grow due to complementation between the non-allelic auxotrophic markers. Hybrids may then be sporulated and either random spores selected, for example, using expression of the heterozygous recessive drug-resistance marker, canl. to select for haploid meiotic products, or tetrads dissected using a micromanipulator. The meiotic products may then be analyzed genetically for the presence of the YAC markers, as well as the genetic markers present in the recipient strain. The presence of the YAC may be confirmed by hybridization or PCR analyses.

It may be desirable to increase the number of copies of the YAC per yeast cell in order to increase the efficiency of the transfer into mammalian cells. One may use a YAC with its appropriate host strain that allows a multi-fold amplification of the YAC. See, for example, Smith et al., PNAS (1990) 87:8242-8246. The YAC may be manipulated, as appropriate, to provide for appropriate markers for introduction of the construct into the amplifiable YAC. The amplifiable YAC, when amplified, may also find use to improve the efficiency of gene targeting and homologous recombination.

The construct may be introduced into the yeast host conveniently by fusion with the cloning host or any other of the common means employed for introducing exogenous DNA into a yeast host. Where the yeast host is employed, it may be desirable to include a marker in the construct which provides for complementation of a metabolic mutation in the yeast. LEU, HIS, TRP, URA, LYS, ADE, and the like are common mutations. In this way, the presence of a construct integrated into the YAC can be selected for. The presence of the marker should not interfere with the intended purpose of the construct and should be acceptable by the target host. If desired, one can excise the yeast marker by using homologous recombination, using a construct which results in deletion of the yeast marker. One may also remove the yeast marker by appropriate restriction enzyme digestion, followed by isolation of the linearized DNA and ligation to recreate the construct absent the yeast marker.

The modified YAC may be purified by any convenient means, such as gel electrophoresis, gradient density or velocity centrifugation, chromosome sorting, or the like. The result of the recombination of the construct with the DNA will be a nested series of deletions, depending upon the distance from the target sequence to the repetitive sequences. Where recombination is employed directly with the target host, one will usually obtain a number of colonies, each of which will contain a varying sized deletion. The cells may then be cloned by limit dilution and the target region analyzed for the size of the deletion. One may then select for the clone with the desired sized deletion or in many instances it may be sufficient that there has been the desired knock-out, regardless of the size of the deletion.

With the YAC, the YAC may now be transformed into the target mammalian cell by any convenient means, including fusion, electroporation, or the like. By having a mammalian selection marker present in the construct which allows for selection of integrants, one can select for those mammalian cells in which the construct has been integrated, at the homologous locus resulting in a deletion. Deletions of interest will usually be at least about 1 kbp, more usually at least about 5 kbp, may exceed 20 kbp, being 50 kbp or more.

In the presence of a marker, the transformed target cells are grown in selective medium containing, for the DHFR gene about 0.01-0.5 μM methotrexate or GHT-minus media with dialyzed serum and, where another marker is present, e.g., the neo gene, the medium may contain from about 0.1-1 mg/ l G418. The resistant colonies are isolated and may then be analyzed for the presence of the construct. The resulting cellular compositions will be mixtures of nested deletions, where depending upon the spacing of genes, one may be able to delete not only the target gene, but additional genes as well. One may also be able to modify the repertoire of the genes of the immunoglobulin locus, by removing groups of variable, constant or joining regions. In particular, one may be able to create extensive deletions of a particular target locus, for example, deletion of the MHC Class I and/or Class II regions for the purpose of creating universal donor cells for transplantation therapy. The cellular composition may be divided by limiting dilution and cloned, so as to have cells having the same sized deletions. The cells may be used as appropriate. The following examples are offered by way of illustration and not by way of limitation.

EXPERIMENTAL Yeast strains and propagation. A Saccharomyceε cerevisxae AB1380 strain carries a pYAC4-based 650 kb yeast artificial chromosome (HYA32G5) containing the human Factor IX (F9) gene approximately 40 kb from the telomeric end of the long arm. (Wada et al . , Am. J. Human Genet. 46:95-106 (1990) HYA32G5 was mated to YPH252 (Sikorski and Hieter (1989) Genetics 122:19-27) , diploids sporulated, and tetrads dissected. A his3A200 haploid segregant was identified

(designated YPH599) to enable manipulations with HIS3 vectors. This segregant containing the 650 kb YAC had the following genotype, MATa ura3-52 lys2-801 ade2-101 trplAl his3A200. Transformations were accomplished using the lithium acetate procedure (Ito et al. (1983) J. Bacteriol. 153, 163-168) using 3 μg of Notl linearized plasmid for the interstitial deletions and 6 μg of Sail linearized plasmid for the terminal deletions. Transformants were selected on minimally supplemented SD plates (Rose et al., Methods in Yeast Genetics, Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY) lacking histidine, colony purified and subsequently tested for the ability to grow in the absence of tryptophan or uracil.

Vector Construction. The starting vector for construction of the interstitial deletion vector was pRS303 (Sikorski and Heter (1989) supra) . This vector contains the gene HIS3 and a Bluescript derived polylinker. A 1.2 kb Sall/Xhol fragment from PMClneopolA (Thomas and Capecchi (1987) Cell 51, 503-512) containing all the sequence necessary for expression of the bacterial neomycin phosphotransferase (Nptll) gene in mammalian cells was cloned into the unique Aatll site in pRS303 by blunt end ligation. The resulting plasmid is referred to as pRSN303. A 4.0 kb Sacl fragment of a cloned LI element from pLl.lA (a 6 kb full length^' LINE 1 element cloned from its insertion site in a factor VIII gene) was inserted into the Sacl site of the polylinker in both orientations. The resulting plasmids are referred to as pllla and plllb. The F9 sequences were generated from a 5.4 kb EcoRI fragment of the plasmid pTM6 containing genomic DNA including exons 7 and 8 of the F9 gene (Yoshitake et al. (1985) Biochemistry 24, 3736-3750) . Synthetic oligonucleotides were constructed to amplify by polymerase chain reaction a 3.1 kb fragment from position 120 to position 3178 (within exon 7) of the fragment. This was blunt end ligated into the Xhol and Clal sites of pILla and pILlb such that when lineratized with Not I the 3' end of exon 7 is exposed. The resulting interstitial deletion vectors are referred to as pF9Lla and pF9Llb. The terminal deletion vectors pBPHO and ill contained the same LI element cloned in two orientations.

Pulsed Field gel electrophoresis (PFGE) and restriction analysis. DNA for conventional gels and high molecular weight DNA for PFGE were prepared by established procedures (Davis et al. (1980) Meth. Enzymol 65, 404-411; Schwartz and Cantor (1984) Cell 37, 67-75). Electrophoretic karyotypes were examined using a contour-clamped homogenous electric field (CHEF) apparatus. The following probes were used for hybridization: a 1.2 kb Sall/Xhol neo fragment from pMClneopolA (Thomas and Capecchi (1987) supra) ; a 2 kb EcoRI/BamHI HIS3 fragment from pBM483 (containing the entire coding region of the His3 gene) ; a 270 bp BamHI fragment isolated from Blur8 (Deininger et al. (1981) J. Mol. Biol. 151, 17-33); a 1.6 kb Xhol CEN4 fragment from YRpl4/ARSI/CEN4 (6.0) (Hieter et al. (1983) Cell 40, 381- 392) ; and a 600 bp BamHI F9 fragment from pF9Lla. Radiolabeled probes were prepared by random primer extension (Feinberg and Vogelstein (1983) Anal. Biochem. 132, 6-13) and hybridized to Gene-Screen Plus (Dupont NEN) or Zeta- Probe (Bio-Rad) nylon membranes.

Gene Targeting. The target YAC, HYA32C5 present in yeast strain YPH599, contained a 650 kb fragment of human DNA derived from the Xq27 region (Wada et al. (1990) Am. J. Human Genet. 4.6:95-106. This YAC was shown to be stable and contains the complete copy of the gene for clotting factor IX (F9) , an anonymous DNA marker DXS102 and part of a gene mcf-2. Homologous recombination of pF9Ll series plasmids with their homologous sequences in the YAC is expected to yield a series of deletions. His⁺ colonies derived from transformation of YPH599 were characterized for the presence of markers on the YAC and to determine if deletions had occurred.

When YPH 599 was transformed with a F9 targeting vector containing 5.3 kb of homology (pF9DV) and the pF9Ll series, His⁺ colonies were obtained. The average transformation efficiency of pF9DV was 175 colonies/μg of DNA while the two interstitial deletion vectors yielded 25 and 15 colonies/μg respectively. From several transformations, a total of 7000 colonies with pF9DV were obtained, 750 colonies from pF9Lla and 180 colonies from pF9Llb. A number of these colonies were isolated and tested for their ability to grow in the absence of uracil and tryptophan. 147/160 or 92% of colonies scored positive for both markers indicating that the YAC is present in each of them.

To ascertain if the yeast transformed with pF9Lla and pF9Llb contain modified YACs, individual colonies were grown in YPD media, the cells embedded in low melt agarose, treated to release chromosomal size DNA, and fractionated by PFGE. YPH599 contained a YAC which is 650 kb. When this strain is modified by pF9DV, an Ω-type recombination is expected to yield a YAC which is 5.6 kb larger.

We examined the nature of the YACs harbored by the transformants obtained from the pF9Ll plasmids. Of a total of 147 Trp⁺, His⁺, Ura⁺ transformants tested, 123 (84%) had a YAC which migrated to a different position than that seen in YPH599. To better understand the structure of the modified YACs, the DNA in the gels revealing the chromosome karyotypes was transferred to nylon membranes and the filters hybridized with a neo specific probe. This detailed examination was conducted on 88 clones. Of these, 9 (10%) were unchanged, 14 (16%) were larger than 650 kb, 55 (62%) were smaller, 4 (5%) did not have detectable YACs and 6 (7%) contained two different sized YACs. The larger YACs could result from integration of the circular version of the input plasmid (either singly or in tandem) into any of its homologous sequences (LI, F9 or plasmid sequences) . Since the majority of the derivative YACs are smaller, they must have undergone deletions. Because they carry the terminal genetic markers, TRP1 and URA3, the deletions must be interstitial, and this class was analyzed further. The deleted YACs are of different sizes. The largest of the deleted YACs is 630 kb and the smallest 150 kb. In many cases, more than one member of each class of deleted YACs was obtained suggesting that these molecules are the result of specific rather than random events.

Molecular analysis of the deletions: If the deletions are to be useful genetic tools, it is necessary that they be nested and generated by homologous recombination. We first ascertained if a HR event occurred at the F9 end. The input plasmid contained in EcoRI site in the neo gene. Homologous recombination of F9 using pF9Lla and pF9Llb are expected to yield an 8.6 kb EcoRI band. In addition, digestion with EcoRI is expected to yield 4.8 kb and 3.6 kb internal bands, respectively. DNA from a representative set of transformants was digested with EcoRI and blot hybridized using the neo probe. 13/23 deletion bearing pF9Lla transformants and 21/33 pF9Llb transformants yielded bands expected from homologous recombination at the F9 end. We continued our studies on the YACs which have undergone HR at the F9.

We also wished to establish if the deletions are the result of HR at the LI end. An unambiguous determination of this feature requires that we know the restriction enzyme maps around each of the LI elements. Such a map of the entire YAC is not available. Fortunately, much of the F9 gene itself was sequenced and an LI element was identified at the 5' end of the gene. The restriction enzyme map around this region is known (Yoshitake et al. (1985) supra) . If this element is targeted by recombination, a 630 kb YAC is expected. When pF9Llb was used, we indeed obtained several independent transformants which contained a 630 kb YAC. We examined whether these are the result of targeting the LI element at the 5' end of the F9 gene. Homologous recombination (HR) at this site is expected to yield a 7.5 kb SphI band. All of the independently derived cell lines of this class which were tested had such a band.

The LI sequence present in the deletion plasmids and the target LI at the 5' end of the F9 gene have at least two differences in the restriction enzyme sites. The LI in the input plasmid had an .EcoRI site which is not shared by the target. The target contained a StuI site which is not present in the plasmid borne LI. Therefore, the site of crossover can be localized with respect to these restriction enzyme site differences. If a crossover occurred at the 5'-region of the LI sequence (A), it would yield a 4.8 kb EcoRI fragment. If crossover occurred at the 3'-region of the LI sequence (B) , it would result in an .EcoRI band of 8.4 kb. Analysis of three of the cell lines from the 630 kb class revealed that two of them are the result of a crossover at A and one resulted from crossover at site B. Crossover to the left of the StuI site on the target will yield a 6.2 kb fragment and a 4.5 kb band will result from crossover to the right of the StuI site. Results indicate that two members of the 630 kb class have undergone a crossover to the left of the StuI site and one to the right of the StuI site. Combining these results we can conclude that the YAC in Llb-7 is the result of a crossover in the interval defined by the StuI and .EcoRI polymorphisms. These results clearly demonstrate that the 20 kb deletions are the result of homologous^* recombination involving F9 at one end and an LI element at the other. It is therefore reasonable to extrapolate that at least a majority of the rest of the deletions have resulted from similar events involving other LI elements.

The repetitive elements present in human DNA could also be used to ascertain if a nested series of deletions has been generated (Gusella et a_l. (1982) Proc. Natl. Acad. Sci. USA. 79_.:7804-7808. DNA from YPH599 was digested with a number of different restriction enzymes, and blot- hybridized using the Alu repetitive element probe (Blurδ, Deininger et til. (1981) supra) . Each of the restriction enzymes yielded a different pattern. The pattern that is generated by each of the enzymes is a fingerprint of the YAC. To determine if this fingerprint was altered as a result of deletions, representative YACs bearing different size deletions were digested with .EcoRI and blot hybridized with labeled Blurδ. The complexity of the banding pattern decreased with progressively larger deletions and the patterns observed were consistent with sequential loss of Alu-containing restriction fragments.

Constructing a restriction map of the YAC: If the deletion series are indeed a nested series, it should be possible to use them to construct a restriction enzyme map of the YAC which should correspond to a map generated by conventional methods. We digested the DNA of the series of deleted YACs with Sail, separated the fragments on pulse field gels and hybridized with total human DNA. The 650 kb intact YAC showed bands corresponding to 230, 150, 100, 90 and an unresolved doublet at about 40 kb. A 20 kb deletion proximal to F9 resulted in the disappearance of the 150 kb band and replacement by a 130 kb band. A 60 kb deletion results in a corresponding reduction in the size of this band. Based upon this kind of analysis, which orders fragments progressively from the F9 sequence, it was possible to deduce that the YAC contained the Sail sites at approximately 30, 260, 360, 450 and 610 kb away from the centromere. This information corresponds to the Sail map of HYA32G5. Choice of LI elements targeted by HR: The LI elements that are targeted by the interstitial deletion vectors were determined. We observed that each of the elements was not targeted at equal frequency. This non- random distribution in the choice of Lls for targeting might be based upon the homology of the LI elements in the target to the input LI or it could be a function of the distance from the anchor site at F9. To distinguish between these possibilities, we used a terminal deletion vector (Vollrath et al. (1988) Proc. Natl. Acad. Sci. USA 85:6027-6031: Pavan et al. (1990) ibid. 7:1300-1304) containing the same LI element as that in the pF9Ll series. Hiε⁺ colonies were generated with a frequency of 19 colonies/μg with pBPlll.

, 100 colonies derived from each transformation were tested for phenotype. " For pBPllO and pBPlll derived transformants, 90% and 89% were found to be His⁺, Trp⁺, Ura^", respectively, confirming that the targeted YAC was present. 32 independent transformants from pBPllO and 24 independent transformants from pBPlll were further characterized by pulsed field gel and Southern analysis. 30/32 and 23/24 were found to contain recombination products of several different size classes (verified by probing with CEN4) . The data in the following table summarizes the positions of the LI elements targeted by different vectors and the number of times each class was recovered. Although some of the LI elements are targeted by both classes of vectors, others were targeted by only one class of vectors.

Table

The counting is in the direction of the Trp to Ura genes where the F9 exon 7 is between the LI sequences at 590 and 640. The numbers are in kb from exon 7 of the Factor IX gene.

The results demonstrate that homologous recombination can occur, by having a known target sequence for homology at one site and a repetitive sequence at another site on a vector, so that one may obtain a nested group of deletions of varying size based on the separation of the repetitive sites. In this manner, one can knock out a variety of genes without knowing the entire sequence of the gene. The deletions that are generated by the interstitial and terminal deletion methods have many uses. Since the deletion endpoints are close to vector sequences, it would be possible to rescue the unique DNA sequences at the sites of deletions. Such unique sequences could serve as sequence tagged sites (STS) used as landmarks in human gene mapping. By introducing the YACs into mammalian cells, the resulting gene transfer would permit study of putative regulatory elements that have the capacity to act at long distances. The interstitial deletions would permit removal of sequences at any location of the YAC, thereby generating substrates which could be used for gene structure-function analysis and long-range genetic interactions. The results observed demonstrate that the repetitive elements may be targeted by homologous sequences in either orientation. While the targeting of the repetitive elements is preferentially at repetitive elements relatively close to the targeted sequence, one does obtain a reasonable frequency of targeting to distant repetitive elements. The data indicate that both homology and distance between the repetitive element and the targeted sequence play a role in generating the interstitial deletions. Thus, besides providing for knock-out of specific genes, the subject methodology may be used for gene mapping.

All publications and patent applications mentioned in this specification are herein incorporated by reference to the same extent as if each individual publication or patent application was specifically and individually indicated to be incorporated by reference.

The invention now being fully described, it will be apparent to one of ordinary skill in the art that many changes and modifications can be made thereto without departing from the spirit or scope of the appended claims.

Claims

WHAT IS CLAIMED IS:

1. A method of introducing a mutation at a chromosomal target site in target host cells, said method comprising: transforming viable cells with a construct comprising a DNA domain homologous with DNA at a target site joined to a DNA domain homologous to a repetitive sequence of chromosomal DNA, whereby homologous recombination occurs in a target chromosome of a proportion of said cells resulting in integration of said construct and mutation of said chromosome, with the proviso that the transforming may be carried out in a first step by transforming a YAC containing yeast host with said construct, wherein said YAC comprises target chromosomal DNA of at least 50kbp, resulting in mutation of said chromosomal DNA, and transforming said target host cells with said mutated chromosomal DNA, whereby said target chromosome becomes mutated in a proportion of said target host cells, and selecting for target host cells comprising said mutation.

2. A method according to Claim 1, wherein said mutation is a deletion.

3. A method according to Claim 1, wherein said mutation results in a nested set of deletions, and including the additional step of separating at least one of the deletion mutations from the other mutated cells.

4. A method according to Claim 1, wherein said construct comprises a marker for selection.

5. A method according to Claim 1, wherein said construct comprises a gene complementary to a yeast mutation.

6. A method according to Claim 1, wherein said target host cell is a human cell and said repetitive sequence is the LI sequence, Alu sequence, alpha-satellite sequence, telomeric or sub-telomeric sequence.

7. A* method accordng to Claim 1, wherein said target cell is a rodent cell.

8. A method according to Claim 1, wherein said target cell is an embryonic cell.

9. A method of introducing a mutation at a chromosomal target site in target host cells, said method comprising: transforming yeast cells comprising a YAC with a construct comprising a DNA domain homologous with chromosomal DNA at a target site joined to a DNA domain homologous to a repetitive sequence of chromosomal DNA, wherein said YAC comprises a chromosomal DNA fragment of at least about 50kbp from a mammalian host, wherein said construct is integrated into said fragment resulting in a mutation of said fragment; transforming target host mammalian cells with said mutated fragment containing said construct; and selecting for target host cells comprising said mutation.

10. A method according to Claim 9, wherein said mutation is a deletion.

11. A method according to Claim 10, wherein said target host cell is a^' human cell and said repetitive sequence is the LI sequence, Alu sequence, alpha-satellite sequence, telomeric or sub-telomeric sequence.

12. A DNA sequence comprising a domain homologous to a structural gene of a chromosome of a mammalian host and a domain homologous to a chromosomal repeat sequence of said mammalian host, joined by other than the native sequence joining said two domains in said mammalian host.