AU2451202A - Method for stably cloning large repeating DNA sequences - Google Patents

Description

P00011 Regulation 3.2 Revised 2/98

AUSTRALIA

Patents Act, 1990

ORIGINAL

COMPLETE SPECIFICATION STANDARD PATENT TO BE COMPLETED BY THE APPLICANT NAME OF APPLICANT: ACTUAL INVENTORS: ADDRESS FOR SERVICE: Case Western Reserve University John J Harrington; Gil B Van Bokkelen and Huntington F Willard Peter Maxwell Associates Level 6 Pitt Street SYDNEY NSW 2000 METHOD FOR STABLY CLONING LARGE REPEATING DNA

SEQUENCES

Divisional of Australian Patent Application No. 743,150 (44,294/97) filed on 22 September 1997 INVENTION TITLE: DETAILS OF RELATED

APPLICATION:

The following statement is a full description of this invention including the best method of performing it known to me:m:\docs\971237\004405.doc -la- Method For Stably Cloning Large Repeating DNA Sequences Field of the Invention This invention relates to the field of gene therapy and gene therapy vector 10 technology. It relates to the development and use of synthetic human chromosomes, or synthetic chromosomes in related primates or other mammals.

It generally relates to the development and use of large arrays of repetitive DNA.

Background of the Invention The ability to clone large, repetitive DNA is an important step toward the development and construction of artificial chromosomes and gene therapy i vehicles. In addition, stable cloning of repetitive DNA in microorganisms will be important for generating high resolution physical maps of mammalian chromosomes.

A variety of cloning systems have been developed to facilitate the cloning and propagation of foreign DNA in micro-organisms. Plasmids, bacteriophage, and yeast artificial chromosomes (YACs) have been used successfully to clone many mammalian DNA sequences. However, some types of repetitive DNA appear to be unstable in these vectors (Schalkwyk et al., Curr. Opin. Biotechnol.

6(1):37-43 (1995); Brutlag, D. et al., Cell 10:509-519 (1977)). This results in gaps in physical genomic maps and precludes the use of these vectors as a means of propagating repetitive DNA, and especially highly repetitive mammalian centromeric DNA.

Bacterial Artificial Chromosomes (BACs) have been constructed to allow the cloning of large DNA fragments in E. coli (O'Conner et al., Science 244 (4910):1307-12 (1989); Shizuyaet al., Proc. Natl. Acad. Sci. USA 89 (18):8794-7 (1992); Hosoda et al., Nucleic Acids Res. 18(13):3863-9 (1990)). While this system appears to be capable of stably propagating mammalian DNA up to at least 300 kb, relatively few independent mammalian DNA fragments have been analyzed (Shizuya et al., Proc. Natl. Acad. Sci. USA 89 (18):8794-7 (1992)). In addition, the few fragments that have been tested for stability in the BAC vector, have not been extensively characterized with respect to the types of sequences 10 present in each fragment. Thus, it is unknown whether these fragments contain repetitive DNA elements. In particular, it is clear, based on the restriction site •and Southern analysis, that these fragments do not contain alpha satellite DNA.

Many mammalian DNA sequences appear stable in Yeast Artificial Chromosome (YAC) vectors, and yet certain repetitive elements of similar length 15 are not (Neil et al., Nucleic Acids Res. 18(6):1421-8 (1990)). Knowledge of DNA properties derived from the YAC system thus suggests that large arrays of repeating units are inherently unstable, even under conditions where similar sized DNA composed of non-repeating DNA is stable.

Thus, the stability of large (greater than 20-100 kb) arrays of repeating 20 units, such as alpha satellite DNA in a BAC vector, prior to the present invention, was not predictable with any reasonable certainty. In addition, even if some alpha satellite arrays were stable in the BAC vector, a priori, it was not clear whether arrays of sufficient size and sequence composition to facilitate centromere function could have been stably propagated in this vector.

The difficulties pertaining to cloning and propagating repeated DNA have been recognized in the literature. In addition, the literature clearly describes a correlation between the number of repeats and plasmid instability. Thus, when considering the relevance of a reference to the invention described and claimed herein, one must take into account the type of DNA repeat, its sequence composition, the size of the repeat unit, and the overall size of the repetitive array.

Hofer et al., Eur. J. Biochem. 167:307-313 (1987), describe the cloning of a tandem array of up to 6 copies of a 69 bp repeat. Tandem arrays of 2 and 4 repeats were found to be stable. But arrays of 6 repeats were unstable in the plasmid used in this study. In addition to establishing a correlation between the size of the array and stability, the authors note that "once the hexameric gene has survived the obviously crucial phase of transformation, it is stably replicated in recA, rec BC, and rec+ hosts." This indicates that in addition to instability caused by repetitive DNAper se in the plasmid vector, the transformation into E. coli of plasmids containing direct repeats also results in instability of the tandem arrays.

10 Prior to the inventors' disclosure herein, there were no definitive data •E regarding plasmid instability resulting from transformation of large tandem arrays of sizes similar to those described in this application. Therefore, it was impossible to predict how unstable such arrays would be during the transformation process. Thus, aside from the lack of evidence pertaining to the 15 feasibility of cloning large arrays into plasmid vectors, the study described in Hofer et al. suggests that regardless of the vector used, it would not have been possible to introduce intact large tandem arrays into E. coli without promoting recombination.

Sinden et al., Genetics 129:991-1005 (1991), discuss the structural 20 instability of plasmids containing indirect repeats. As with direct repeats, this study and others show that there is a correlation between the size of the indirect repeat and the degree of structural instability. Sinden et al. note that "it is difficult to maintain inverted repeats greater than 150 bp in length in plasmid DNA in E. coli" and that "the inability to clone long inverted repeats and the genetic instability associated with inverted repeats have been reported by a large number of investigators." Thus, these studies show that different classes of repetitive DNA have different stability properties, and that the properties correlate with the size of the repetitive array (and the number of repeat unit copies). In each case, instability was observed with tandem array sizes far smaller than those described in this application.

Leonhardt et al., Gene 103:107-111 (1991), analyze the stability of a plasmid that contains direct and indirect repeats. However, in contrast to the repetitive arrays disclosed and claimed herein, the direct repeats described were small (on the order of 7 bp) and present in only 2 copies. These repeats were not located in tandem on the plasmid. They were separated by intervening plasmid DNA sequences, often several kb in length. As with the direct repeats, the indirect repeats were separated by a large amount of intervening plasmid 10 sequence. Thus, the plasmids used in this study do not provide any indication that large repetitive arrays could be cloned using plasmids.

Summary of the Invention Therefore, it is an object of this invention to describe a method for constructing uniform or hybrid synthetic arrays of repeating DNA, and especially, alpha satellite DNA. It is a further object of the invention to describe a method for the cloning, propagation, and stable recombinant production of repeating DNA, and especially, naturally occurring or synthetic alpha satellite arrays.

Accordingly, the invention is directed to a method for stably cloning introducing and amplifying) large repeating DNA sequences, vectors containing the above arrays, hosts transformed with such vectors, and compositions containing the amplified vectors and/or sequences.

BriefDescription of the Figures Figure 1 is a schematic diagram of the method of the invention. The numbers "1-16" represent 1-16 copies of monomeric units (of approximately 171 bp) of alpha satellite DNA, tandemly aligned in a linear array. represents a desired restriction enzyme site in the backbone of the vector carrying the array during its expansion to a desired size.

Figure 2 is a graphical representation of the correlation of the percent recombinants (after 50 generations) and the array size (kb).

Detailed Description of the Preferred Embodiments By "stably transformed" is meant that the cloned DNA array containing the repeating units is capable of being propagated in the desired host cell for at least 50 generations of growth with a recombination frequency of less than 0.6% per generation (for 174 kb arrays) and a recombination frequency of less than 0.2% (for 130 kb arrays). Arrays smaller than 130 kb exhibit little or no recombination when cloned by the method of the invention.

By "higher order repeat" is meant a repeating unit that is itself :composed of smaller (monomeric) repeating units. The basic organizational unit of alpha satellite arrays is the approximately 171 bp alphoid monomer. Monomers are organized into chromosome-specific higher order repeating units, which are also tandemly repetitive. The number of constituent monomers in a given higher order repeat varies, from as little as two (for example, in human chromosome 1) to greater than 30 (human Y chromosome). Constituent monomers exhibit varying degrees of homology to one another, from approximately 60% to virtual sequence identity. However, higher order repeats retain a high degree of homology throughout most of a given alphoid array.

The method of the invention provides a method of stably cloning repetitive, and especially highly repetitive regions of DNA, in micro-organisms.

Arrays of defined length, composition, orientation and proper phasing are possible. By "proper phasing" is meant that the precise length and orientation of any given higher order repeat in the array is not altered from that in the naturallyoccurring sequence by the construction of the array, and also that there is no nonrepeating DNA at the junction of the repeating units. For alphoid sequences, for example, the length and orientation of the higher order repeating unit is exactly the same as the naturally-occurring higher order repeat, and there is no nonalphoid sequences present at the junction of the higher order repeats, except for the bases modified to create the restriction sites.

Accordingly, a method for cloning repeating tandem arrays of DNA is provided, wherein a first DNA unit is prepared such that the opposing ends of the DNA unit contain complementary, but non-isoschizomeric restriction sites. This DNA is ligated into a vector, and the vector is linearized at one of the restriction sites. A second DNA unit, prepared as in step is then ligated in tandem with 10 said first unit, so as to form a directional, repeating array. This array is transformed into a host cell, and especially a bacterial host cell, and stable clones containing said array are selected. Starting with the vector linearization, these steps are repeated until a desired array size is reached.

The directional cloning scheme of the invention is illustrated in Figure 1, 15 wherein the cloning of alpha satellite DNA higher order repeats is illustrated. As shown in the figure, the method of the invention utilizes a "build-up" approach, wherein shorter units, preferably higher order repeats, are added together to create the longer tandem array of repeating units. The units are added to each other in a manner that results in a defined orientation, which is established by two S: 20 different restriction sites one at each end of each repeating unit. Preferably, the repeating unit, and especially the higher order repeating unit, that is thebasis of the tandem array contains complementary, but non-isoschizomeric restriction sites at opposing ends. Such ends may be designed into the unit using methods such as polymerase chain reaction (PCR). In the method of the invention, by complementary ends, it is intended to include both complementary overhanging ends and blunt ends.

In a preferred embodiment, the DNA array is alphoid DNA. As shown in Example 1, polymerase chain reaction (PCR) can be used to amplify a single 2.7 kb DNA alphoid unit (actual length 2.712 kb) such that complementary restriction sites (BamH I and Bgl II) are created at opposing ends of the higher order register of alpha satellite repeat from human chromosome 17. The ends of the higher order repeats can be modified using polymerase chain reaction mediated site directed mutagenesis so that complementary restriction sites are created at opposite ends of each repeat. The modified higher order repeats are then cloned into the mini-F cloning vector pBACl08L (Shizuya, H. et al., Proc. Natl. Acad, Sci. USA 89:8794-8797 (1992), incorporated herein by reference). These complementary restriction sites are used in conjunction with non-complementary flanking restriction sites to directly clone synthetic arrays of alpha satellite DNA derived from a single modified higher order repeat (Figure Synthetic arrays 10 can be created from any higher order alphoid repeat, including such alphoid DNA derived from chromosome 17, the Y chromosome, or other chromosome. In addition, hybrid arrays consisting of higher order repeats from both chromosomes can be prepared. In a preferred embodiment, the DNA is human DNA.

Arrays up to 200-215 kb are stable in the vector and hosts of the 15 invention. In one embodiment, an array of 87 kb 215 kb in length is constructed.

In a preferred embodiment, an array of at least 100 kb in length is constructed. In a highly preferred embodiment, an array of at least 140 kb, and especially at least 174 kb in length is constructed; arrays of 174 kb exceed the minimum known observed length of a functional alpha satellite array.

20 Examples of useful complementary, but non-isoschizomeric restriction enzymes that are useful in creating such sites include: Sal I and Xho I; Mun I and EcoR I; Afl III and Nco I/Sty I (isoschizomers: either one can be the partner for the non-isoschizomer partner); Nhe I and Xba I and Sty I/Avr II (isoschizomers) and Spe I (any combination); Cla I and BstB I and Acc I (any combination); Mlu I/Afl III (isoschizomers) and BssH II and Asc I; and Not I and Eag I. Bcl I is a complementary/non-isoschizomer of both BamH I and Bgl II.

The amplified DNA is then digested using, for example, BamH I Sfi I, or Bgl II Sfi I. Following separation of the bands in the digested DNA using physical methods capable of separating such DNA, such as, for example, gel electrophoresis, the DNA band from one digest is excised and ligated to the excised DNA band from the other digest. In the above example, since Bgl II and BamH I generate compatible overhangs, and Sfi I generates an asymmetric overhang that can only religate in a particular orientation, DNA flanked by these sites ligates to the vector DNA to create a tandem dimer arrayed in head to tail fashion. This DNA can then be transformed into a microorganism.

In a second variation of this strategy, blunt cutting restriction enzymes can be substituted for BamH I and Bgl I. For example, Sma I and EcoR V can be substituted for BamH I and Bgl II, respectively. The digestion and fragment isolation are then carried out as above. An important feature of this strategy, 10 common to both the blunt and complementary/nonisoschizomeric variations, is that the physiologic phasing of the arrays can be precisely maintained, if desired.

Examples of additional blunt cutters that can be used include Ssp I, Stu I, Sca I, Pml I, Pvu II, Ecll 36 II, Nae I, Ehe I, Hinc II, Hpa I, SnaB I, Nru I, Fsp I, Dra I, Msc I, Bstl 07 I, Alu I, Asp 700/Xmn I, Avi II, BbrP I, Bstl 107 I, Eco47 III, Dpn S 15 I, Hae III, HindIl, Nam I, MluMI, Mvn I, Rsa I, Swa I, Bsh12361, Eco72I, PalI, and SrfI.

Stability of the cloned plasmids containing large tandem arrays of synthetic alpha satellite DNA can be determined using simple growth and dilution experiments as described in Example 2. For example, stability can be determined S 20 by passage for 50 generations and subsequent analysis of plasmid DNA for structural integrity. Plasmid structure can be analyzed by restriction analysis, and agarose gel electrophoresis. Little or no recombination was observed for these clones, indicating that the directional cloning scheme can be employed to construct and propagate synthetic alpha satellite arrays in the context of a mini-F cloning vector (such as the pBAC108L vector) and a suitable E. coli host.

The methods of the invention can be used to construct any desired repeating DNA unit or combination of units. For example, alpha satellite DNA from any eukaryotic chromosome, especially human DNA, can be cloned. Other examples of large tandem arrays of highly repetitive DNA include the immunoglobulin DNA loci, regions of heterochromatic repeats and telomeres.

The directional cloning method of the invention does not require the presence of polymorphic restriction sites, as required when cloning endogenous (nonmodified) arrays. Even when present, these sites do not permit control of the exact size of the array. Furthermore, by using a single higher order repeat as used in the method of the invention, (see Figure and sequentially doubling its size, the exact sequence of the entire array is known, since the sequence of the original higher order repeat is known.

When cloning endogenous arrays, such as, for example, endogenous alpha satellite arrays, one cannot be confident of the precise composition of a given 10 array. In particular, interruptions in the arrays by non-repetitive DNA may have significant effects on stability in E. coli. In addition, to be suitable as a vector for gene therapy, one must know the exact sequence of the vector being provided to •the recipient of such therapy. The method of the invention obviates that concern S"and allows the artisan to bypass sequencing of native alpha satellite arrays, in 15 favor of constructing a useful array, de novo, from known repeating sequences.

Structurally diverged higher order repeats generally exhibit increased structural stability in E. coli relative to more homogeneous arrays. Thus, by e utilizing homogeneous synthetic arrays according to the method of the invention, an accurate determination of the minimal stability of the repeating DNA in the S 20 vector can be obtained. However, it may be desired to construct heterogeneous arrays. In this case, the advantages of the invention apply to the construction of heterogeneous units as well. The particular number and order of the units can be controlled.

Any desired bacterial host in which the vector is stably maintained may be used as the host. Especially E. coli is useful when utilizing BAC vectors and the BAC system.

Synthetic alpha satellite arrays can be utilized in the construction of synthetic human chromosomes in the following manner: by transfection of episomes containing synthetic alpha satellite arrays into a human or other mammalian cell line; by transfection of episomes containing synthetic alpha satellite arrays in conjunction with randomly cloned human DNA or specific DNA fragments, into a human or other mammalian cell line; by co-transfection into a human or other mammalian cell line of synthetic alpha satellite arrays with unlinked specific chromosomal components, such as telomeric DNA, matrix attachment regions, and/or other chromosomal loci that enhance the mitotic stability of alpha satellite containing episomal DNA.

Co-transfection of these components in unlinked form allows the transfected cell line to construct an infinite number of structural permutations, permitting the most stable forms to be retained, while the unstable forms are lost, over time.

.10 Stable conformations can subsequently be harvested utilizing standard methods and procedures. Those constructs that exhibit episomal stability in the absence of selective pressure can be isolated and subsequently utilized in the preparation of gene therapy vectors containing one or more therapeutically useful entities such as genes, ribozymes, or antisense transcripts.

15 In a further aspect of the invention, the inventors have discovered that large arrays of repetitive DNA can be stably amplified in a plasmid vector. This is unexpected in view of previous results in the art. The inventors have *discovered that there is an inverse relationship between the insert size and the plasmid copy number for stable cloning of large repetitive inserts in plasmids.

20 Therefore, in an alternative embodiment of the invention, the repetitive DNA is directly inserted in a plasmid vector, without a restriction on the flanking restriction enzyme sites. The DNA can be the native sequence (for example, a genomic repetitive sequence produced by digesting genomic DNA), an array produced by the "buildup" approach described above and exemplified herein, a synthetic array, or combinations of these.

Thus, any repetitive nucleotide sequence is potentially available for cloning into a plasmid vector. While highly preferred embodiments of the invention are directed to repetitive DNA consisting of tandem direct repeats, the invention encompasses other forms of repetitive DNA. These include, but are not limited to, interspersed repeated sequences and inverted repeats. The invention -11also encompasses not only the same repetitive unit, but also heterogeneous repetitive units, for example, divergent repetitive sequences. In addition, any combination of these types of repetitive sequence is also encompassed by the invention.

Further, the invention encompasses native or synthetic repetitive sequences.

The invention also encompasses repetitive sequences having any and all biological functions, although in preferred embodiments, the centromeric function co-localization with the kinetochore) is encompassed. These include, but are 10 not limited to, centromeric sequences from any organism, and particularly from mammalian organisms, including alpha-satellite DNA from human and primate sources, minor satellite DNA from Mus musculus, 69 and 79 bp repeats from Mus caroli, satellite III, beta-satellite, gamma-satellite, major satellite, and all submonomeric repeats of each of these classes of tandem repeats.

15 Although the invention is useful and encompasses any type of repetitive DNA, preferred embodiments of repetitive DNA include those that contain submonomeric sequences. Most preferably, tandemly repeated sequences are Sencompassed by the invention. The invention is particularly directed to unit repeat lengths of over around 170 bp. In the most highly preferred embodiment, S 20 the repetitive DNA is human alpha-satellite DNA.

The minimum size range for the repetitive DNA insert is at least approximately 20 kb. However, the invention encompasses any range from to over 140 kb, for example, 40, 60, 80, 100, 120, and over 200 kb.

Any plasmid vector is potentially available to practice the invention as long as the plasmid can be maintained at a copy number sufficiently low to stably amplify the specific repetitive DNA insert. Examples include, but are not limited to, bacterial artificial chromosomes (BAC) and other F factor-based plasmids and P1 vectors. In preferred embodiments, the plasmid is an F factor-based plasmid.

In highly preferred embodiments, the plasmid is a bacterial artificial chromosome, such as pBAC108L.

-12- It is, of course, understood that a plasmid need not be derived from bacterial/prokaryotic origin. While all bacterial plasmids are potentially encompassed by the invention, plasmids derived from eukaryotic sources may also be encompassed. Further, viral plasmids are also useful. Such plasmids are known to those of ordinary skill in the art. See, for example, ATCC Recombinant DNA Materials, ATCC Bacteria and Phages, and Cloning Vectors, Pouwels, ed., all of which are incorporated herein by reference for their relevant teachings.

Routine assays exist to enable the person of ordinary skill in the art to test whether a given plasmid can stably amplify a given repetitive DNA sequence of 10 a desired length. This is described herein in the methods and examples sections.

Whereas certain plasmids are inherently maintained at a low copy number by means of elements endogenous to the plasmids, the person of ordinary skill would be aware of ways to potentially reduce the copy number of a plasmid.

SThese include, but are not limited to, mutating the origin of replication, mutating S 15 genes that encode proteins responsible for plasmid replication, and growth of the plasmid-containing cells in minimal media or at higher or lower culture temperatures.

Having now generally described the invention, the same will be more readily understood through reference to the following examples which are provided by way of illustration, and are not intended to be limiting of the present invention, unless specified.

Examples Example 1 Construction of Chromosome 17Alpha Satellite Vectors In order to clone and propagate alpha satellite DNA in E. coli using BAC vectors, a series of tandem alpha satellite arrays of various sizes were constructed.

-13- The structure of the higher order alpha satellite repeat from human chromosome 17 has been described previously (Waye and Willard, Mol. Cell.

Biol. 6(9):3156-65 (1986)). The predominant higher order repeat of human chromosome 17 is 2.7 kb in length, and consists of 16 alphoid monomers flanked by EcoR I sites. A discrete structural unit of alpha satellite DNA, the higher order repeat, derived from human chromosome 17 or the human Y chromosome was cloned into the plasmid cloning vector pACYC 184 (New England Biolabs) by digesting human genomic DNA with the restriction enzyme EcoR I. The nucleotide sequence of the cloned higher order repeats was verified by DNA 10 sequence analysis.

Polymerase chain reaction (PCR) was used to amplify a single 2.7 kb higher order repeat monomer unit such that complementary restriction sites (BamH I and Bgl II) were created at opposing ends of the higher order register.

The precise length of the higher order repeat was maintained. The.primer pair 15 used to amplify this fragment was VB100 5' gggcgggagatctcagaaaattctttgggatgattgagttg and VB101 5' gggcgggatcccttctgtcttctttttataggaagttattt.

S: The modified higher order repeat was cloned into the BAC vector pBAC108L (a gift from Bruce Birren, California Institute of Technology, Pasadena, CA) by 20 digesting the amplified fragment with BamH I, gel purifying the insert DNA, and ligating into vector DNA which had been digested with BamH I and Hpa I and gel purified. The resulting plasmid was designated pBAC-17a 1.

To construct a synthetic dimer of alpha satellite DNA, aliquots of pBAC-17al were digested separately with either: BamH I Sfi I, or Bgl II Sfi I. Following gel electrophoresis, the 2.7 kb alpha satellite band from the Bgl II/Sfi I digest was excised and ligated to the excised pBAC-17a BamH I/Sfi I fragment. Since Bgl II and BamH I generate compatible overhangs, and Sfi I generates an asymmetric overhang that can only religate in a particular orientation, the 2.7 kb fragment ligates to the vector DNA to create a tandem dimer arrayed in head to tail fashion. Ligation products were transformed into the -14bacterial strain DHIOB by electroporation. Clones were analyzed by restriction analysis, and those that contained a tandemly arrayed modified dimer of the chromosome 17 alpha satellite were designated pBAC-17a2 (Figure This strategy was repeated to create extended alpha satellite arrays consisting of 4, 8, 16, 32, 48, or 64 higher order alpha satellite repeats. A similar strategy was utilized to construct synthetic arrays of higher order repeats from other human chromosomes, such as the Y chromosome.

The construction of BAC vector containing 174 kb of alpha satellite DNA represents the largest amount of this class of DNA to be cloned and propagated in E. coli to date. Previous experiments have successfully cloned approximately 40 kb in E. coli using cosmids (Willard et al., Prog. Clin. Biol. Res. 318:9-18 (1989)). Others have used medium copy number plasmids to clone arrays ranging in size from the about 171 bp alphoid monomer to 40 kb (Waye and Willard, Nucleic Acids Res. 15(18):7549-69 (1987)). In the studies reported in the art, a S. 15 high frequency of recombination was observed in the plasmids that contained the largest alphoid arrays. In contrast, during the propagation of the pBAC-17a64, we observed little evidence of recombination products utilizing standard methods of plasmid purification. Since instability of alpha satellite DNA has been shown in the context of these cloning vectors, we analyzed our preparations of pBAC- 20 17a64 for the presence of obviously rearranged arrays utilizing gel electrophoresis. By this assay, the presence of significant levels of rearranged plasmid was not detectable.

Example 2 Stability assay Although no evidence for high levels of recombination and/or deletion in the synthetic alpha satellite arrays from Example 1 was observed, it was possible that recombination was occurring at relatively high levels, but the large number of different deletion products prevented any one product from being detectable.

Therefore, the rate of recombination of these constructs utilizing a highly sensitive assay was determined. In addition, because larger arrays might be expected to be less stable than smaller ones, several different size constructs were examined. The stability of these constructs was examined below.

Stability assays were carried out using three different alpha satellite array sizes which have been cloned into pBAC108L, as described above. These constructs, called pBAC-17a32, pBAC-17a48, and pBAC-17a64, contain 87 kb, 130 kb, and 174 kb of alpha satellite DNA, respectively. Following transformation (electroporation) into the E. coli strain, DHIOB (GIBCO BRL), 10 single clones were picked and analyzed. Because it is possible that the transformation process itself may lead to DNA rearrangements, only clones containing predominantly full-length constructs, as judged by restriction digest and electrophoresis, were saved as glycerol stocks.

To begin the stability assay, cells from a glycerol stock of each construct were streaked onto LB plates containing 12.5 pg/ml chloramphenicol. Eight of the resulting colonies were picked and grown individually to saturation in 5 ml of LB containing 12.5 jig/ml chloramphenicol (approximately 20 generations).

The plasmid DNA from each clone was then purified, digested with BamH I, and separated by pulse field gel electrophoresis. Clones that contained any full length 20 plasmid were said to have full length plasmid at the single cell stage. Clones which did not contain any detectable full-length plasmid were said to be the result of a recombination event prior to restreaking a rearrangement that occurred during production of the glycerol stock) and were excluded. Of the clones that contained some full-length plasmid, 10 cultures were picked at random, diluted 1 to one million into fresh LB containing 12.5 Ig/ml chloramphenicol, and grown to saturation (approximately 30 generations). From a single cell to this final saturated culture, approximately 50 generations of growth have occurred.

In order to determine the percent of plasmids that rearranged during these generations, each saturated culture was streaked onto LB plates containing 12.5 gg/ml chloramphenicol. Individual colonies were then grown to saturation -16in 1.5 ml LB containing 12.5 (tg/ml chloramphenicol. Following growth, the DNA was purified and analyzed by restriction digest (BamH I) and PFGE. Any clone that contained detectable full length plasmid was scored as unrearranged during the 50 generation experiment. Conversely, any clone which did not contain any detectable full-length plasmid was scored as rearranged during the generation experiment.

To calculate the average rearrangement frequency per generation for each construct, the fraction of rearranged clones was determined after 50 generations.

One minus this value is equal to the fraction of unrearranged clones (after generations). The fraction of clones that rearrange after one generation is 1 minus the 5 0 th root of the fraction of unrearranged clones after fifty generations. This is summarized in the following equation: X 1

O

where X is the fraction of clones which rearrange per generation and Y is the fraction of rearranged clones after 50 generations of growth.

^Using this strategy, the recombination frequency of three alpha satellite containing constructs was determined. After 50 generations of growth, 0% (n=9) of the pBAC-17a32 clones recombined to truncated forms. Recombinants were detected for the pBAC-17m48 and pBAC-17a64 at a level of 8.5% (n=59) and 20 25% respectively. This corresponds to a per generation'recombination •frequency of 0.18% for pBAC-17a48 and 0.57% for pBAC-17a64. Thus, this recombination frequency is significantly lower than that reported for other cloning vectors containing far less alphoid DNA (for example, 40 kb) and which are grown for less generations (for example, 30 generations).

The results show that alpha satellite arrays up to at least 174 kb in size can be stably propagated in E. coli using BAC vectors in the methods of the invention. The 174 kb and 130 kb arrays recombine at a frequency of 0.57% and 0.18% per generation, respectively. Thus, using pBAC-17a64 as an example, following 50 generations of growth from a single cell, approximately 1,000 liters of saturated bacterial culture can be produced from a single cell and at least -17of the cells will contain full-length pBAC-17a64, on average. This degree of rearrangement falls within the expected acceptable range for the large scale production of alpha satellite containing human artificial chromosomes for use in gene therapy.

In addition to determining the frequency of alpha satellite DNA rearrangement in pBAC108L, a correlation between the size of a highly repetitive alpha satellite array and its stability in this vector was established. Based on the recombination frequencies determined above, the minimum upper size limit estimate of homogenous alpha satellite DNA in BAC vectors (assuming 50% full S 10 length clones after 50 generations to be acceptable) is conservatively between 200 and 215 kb (Figure This was determined by extrapolation using the computer program Cricket Graph. This is a minimum estimate of alphoid capacity, as other lines are found that fit the data and produce larger estimates than those stated above. From this correlation, it is estimated that, when using 200-215 kb arrays, S 15 and propagation in bacterial strain DB1OB, greater than 50% of the plasmid will be full length after 50 generations. It is likely that this upper size figure could be extended by utilizing the BAC vector in conjunction with specialized recombination defective bacterial strains. Furthermore, this estimate is based on maximally homogeneous arrays. The stability of diverged arrays including .oeo.i S 20 certain natural alpha satellite arrays should exceed this estimate.

The experiment described here represents the first stable cloning and propagation of an alpha satellite array larger than 50 kb in E. coli. Previously, alpha satellite DNA has been cloned in E. coli using cosmids (Haaf et al., Cell 70(4):681-96 (1992). In addition to the relatively small size of these arrays (equal to or less than 40 kb), the integrity and stability of these arrays was not analyzed, Alpha satellite DNA has also been cloned using YAC's (Neil et al., Nucleic Acids Res. 18(6):1421-8 (1990)). In these studies, the instability of the alpha satellite arrays was noted, and additional manipulations such as agarose gel purification was required to obtain preparations containing predominantly full-length arrays.

In addition, there are certain disadvantages to using YAC's to propagate alpha satellite DNA. Perhaps the most important of these relates to the topology of YAC's. In general, YAC's are linear DNA molecules, and therefore, simple alkaline lysis purification methodology can not be used to purify the alpha satellite construct away from contaminating yeast chromosomes. Instead, pulsed field gel electrophoresis (PFGE), a separation method which is not amenable to scale-up, must be used. Finally, the linear topology of YAC's renders them particularly susceptible to shearing during purification. Here, we have demonstrated that the alpha satellite containing BAC's can be harvested and purified away from E. coli chromosomal DNA without substantial shearing.

10 Previous studies suggest that alpha satellite DNA is an important component of the functional human centromere. Naturally occurring alpha satellite arrays range in size from 230 kb to several megabases in length (Oakey and Tyler, Genomics 7(3):325-30 (1990)); (Wevrick and Willard, Proc. Natl.

Acad. Sci. USA 86(23):9394-8 (1989)). However, recent studies suggest that as 15 little as 140 kb of alpha satellite DNA is sufficient to confer centromere function .0 in human cells (Brown et al., Human Molecular Genetics 3(8):1227-1237 (1994)). The alpha satellite array constructs described herein are the first that allow the large scale, stable production of alpha satellite arrays which are large enough to satisfy the alpha satellite requirements of a functional human 20 centromere. These constructs can serve as the backbone of synthetic human chromosomes.

Example 3 Construction of Y Chromosome Alpha Satellite Vectors Construction was as in Example 1, except that DNA from the human male Y chromosome cell line GM07033 was used; DNA from any normal human male cell line would be equivalent. The predominant higher order alphoid repeat on the human Y chromosome is 5.7 kb in length, and is demarcated by flanking EcoR I sites. The 5.7 kb higher order repeat from the Y chromosome alphoid array was cloned into a standard E. coli cloning vector, pACYC 184 (New England Biolabs).

The ends of the higher order repeat were then modified using PCR to create a BamH I site at one end and a Bgl II site at the other, replacing the existing EcoR I site. The modified higher order repeats were cloned into the pBAC108L cloning vector as above.

Example 4 Construction of Hybrid Alpha Satellite Vectors Construction was as in Example 1, except that alphoid DNA from both .o S* human chromosome 17 and the human Y chromosome was used. Two types of 10 arrays were constructed. One array was a simple alternating repeat wherein one higher order repeat unit of chromosome 17 alphoid DNA alternated with one higher order repeat unit from the Y chromosome alphoid DNA. The second type of array that was constructed alternated a dimeric unit of the chromosome 17 o higher order repeat of alphoid DNa with one unit of the chromosome Y repeat of alphoid DNA. In each case, as with the above examples, the proper phasing of the .higher order repeats derived from each chromosome was retained at the junction of the synthetic hybrid.

Although the foregoing refers to particular preferred embodiments, it will be understood that the present invention is not so limited. It will occur to those ordinarily skilled in the art that various modifications may be made to the disclosed embodiments and that such modifications are intended to be within the scope of the present invention.

-19.1- SEQUENCE LISTING GENERAL INFORMATION: APPLICANT: CASE WESTERN RESERVE UNIVERSITY 10900 EUCLID AVENUE CLEVELAND, OHIO 44106 APPLICANTS/INVENTORS: HARRINGTON, J.J.

VAN BOKKELEN, G.B.

WILLARD, H.F.

(ii) TITLE OF INVENTION: METHOD FOR STABLY CLONING LARGE REPEATING DNA SEQUENCES (iii) NUMBER OF SEQUENCES: 2 (iv) CORRESPONDENCE ADDRESS: ADDRESSEE: STERNE, KESSLER, GOLDSTEIN FOX P.L.L.C.

STREET: 1100 NEW YORK AVENUE, SUITE 600 CITY: WASHINGTON STATE: DC COUNTRY: USA ZIP: 20005-3934 COMPUTER READABLE FORM: MEDIUM TYPE: Floppy disk COMPUTER: IBM PC compatible OPERATING SYSTEM: PC-DOS/MS-DOS SOFTWARE: PatentIn Release Version #1.30 (vi) CURRENT APPLICATION DATA: APPLICATION NUMBER: To be assigned FILING DATE: Herewith

CLASSIFICATION:

(vii) PRIOR APPLICATION DATA: APPLICATION NUMBER: US 08/717,117 FILING DATE: 20-SEP-1996 (viii) ATTORNEY/AGENT INFORMATION: NAME: CIMBALA, MICHELE A.

REGISTRATION NUMBER: 33,851 REFERENCE/DOCKET NUMBER: 1522.001PC03 (ix) TELECOMMUNICATION INFORMATION: TELEPHONE: (202) 371-2600 TELEFAX: (202) 371-2540' INFORMATION FOR SEQ ID NO:1: SEQUENCE CHARACTERISTICS: LENGTH: 41 base pairs TYPE: nucleic acid STRANDEDNESS: single TOPOLOGY: linear (ii) MOLECULE TYPE: cDNA -19.2- (xi) SEQUENCE DESCRIPTION: SEQ ID NO:1: GGGCGGGAGA TCTCAGAAAA TTCTTTGGGA TGATTGAGTT G 41 INFORMATION FOR SEQ ID NO:2: SEQUENCE CHARACTERISTICS: LENGTH: 41 base pairs TYPE: nucleic acid STRANDEDNESS: single TOPOLOGY: linear (ii) MOLECULE TYPE: cDNA (xi) SEQUENCE DESCRIPTION: SEQ ID NO:2: GGGCGGGATC CCTTCTGTCT TCTTTTTATA GGAAGTTATT T 41

**C

e oo C

*C

o o

C

Claims (12)

1. An amplified plasmid vector, said vector comprising a sequence consisting of a repeating DNA array, said array comprising repeating DNA units, wherein said array is greater than approximately 20 kilobases in length.

2. The vector of claim 1, wherein said DNA is centromeric DNA.

3. The vector of claim 2, wherein said centromeric DNA is alpha satellite DNA.

4. The vector of claim 3, wherein said array of said alpha satellite DNA is greater than 100 kb in length.

5. The vector of claim 4, wherein said array of said alpha satellite DNA is greater than 140 kb in length.

6. The vector of claim 2, wherein said alpha satellite DNA is human alpha satellite DNA.

7. claims 1-6. A host cell, stably transformed with the vector of any one of

8. The host cell of claim 7, wherein said host cell is a prokaryotic cell.

9. The host cell of claim 8, wherein said prokaryotic cell is E. coli. factor based.

The plasmid vector of claim 1, wherein said plasmid vector is F- -21-

11. The plasmid vector of claim 10, wherein said plasmid vector is a BAC vector.

12. A method for amplifying the plasmid vector of any of claims 1-6, and 11, comprising: introducing said vector into a host cell; and allowing said vector to replicate in said host cell. Dated this 5th day of March 2002 9 Case Western Reserve University Patent Attorneys for the Applicant PETER MAXWELL ASSOCIATES 9* 9* 9 .p