BACKGROUND OF THE INVENTION The present invention relates to cDNA libraries, the individual clones of which are representative of substantially all the genes in a well-defined chromosome region, such as a single band, that are expressed in a source of RNA such as a cell line or a tissue. The invention further relates to methods for making the libraries and for using them. Identification of the genes associated with genetic abnormalities has been an important avenue to determining the etiological basis of many inherited diseases. See Mapping and Sequencing the Human Genome, National Academy Press, Washington, D.C. Currently, approximately 700 of more than 4,000 inherited diseases have been associated with specific regions in human chromosomes. It has been possible only in some of these cases to isolate a gene associated with a disorder by conventional cloning techniques. In fact, when chromosomal location is all that is known about a gene, current methods do not provide any general method to clone the gene or the specific region in which it is located. Thus, even when a heritable trait has been mapped to a chromosomal region, considerable ingenuity and effort are required to clone a set of DNA fragments that encompasses the region and to identify the sequences in the fragments that might encode the gene or genes responsible for the trait.

Current methods suffer, in particular, from limitations on the sizes of DNA that can be cloned, difficulties in constructing ordered sets of overlapping fragments that span a chromosomal region, and the difficulties of identifying coding sequences in the large DNA regions associated with cytologically defined chromosome abnormalities. For instance, presently available vectors cannot accommodate all the DNA of a chromosome region identified by cytological techniques. Plasmid and phage-derived

cloning vehicles can accommodate fragments only as long as 40 kilobase pairs, far less than the several million base pairs that typically characterize cytologically defined chromosomal features. Furthermore, creating an ordered set of clones covering a sequence of this size by chromosome walking would be a prohibitive undertaking, even without the attendant difficulties caused by spurious hybridization of repeat sequences.

Even YAC cloning vectors cannot accommodate DNA long enough to span a cytologically defined chromosomal region; these vectors can accommodate fragments of DNA several hundred thousand base pairs long, not several million base pairs long. Thus, even with YAC vectors, chromosome walking is required to generate a library that contains, uniquely, the DNA mapping to a particular chromosomal region.

Cloning by microdissection has proven useful in some cases to generate libraries of genomic DNA derived from a chromosomal locus. See, for instance, Kas et al. (1991) Proc. Nat ^« l Acad. Sci., U.S.A. 88:1844-48. Genomic microdissection cloning is difficult, however, since the method requires manual dissection and only a very small amount of clonable DNA is obtained. It is practical at most to obtain only a few samples of a chromosome region, and, therefore, only a relatively small number of DNA molecules are available for cloning. Even for polytene chromosomes, which are particularly well suited to this method, it is possible by microdissection to obtain only, perhaps, 10,000 molecules for cloning. Inevitably, therefore, cloning reactions in this method must be carried out at relatively high DNA dilution, far from optimal conditions. Accordingly, the reactions are relatively inefficient and it is very difficult to produce representative libraries. Moreover, to maintain DNA concentrations that support even relatively inefficient reactions, it is necessary to perform the reactions in nanoliter volumes, further complicating use of the method.

and high dilution. Micro-methods also have been employed to raise DNA concentration to improve cloning efficiency. Still, the method has met with only limited success and contemporary cloning techniques cannot routinely provide a library that contains, uniquely, all the DNA mapping to a cytologically defined chromosomal location.

Moreover, the genomic cloning methods described above inherently cannot create libraries of expressed sequences, even if they can be used to prepare libraries that uniquely represent a cytologically defined locus, such as a chromosome abnormality associated with a genetic disorder. Even when overlapping segments of DNA are obtained that encompass a cytologically defined locus, such as a chromosome abnormality, it is difficult or impossible to identify within these segments sequences that encode the genes of interest, such as the particular genes involved in a genetic disorder associated with the locus.

Features common to eukaryotic genes have been used as guide-posts to identify genes in such uncharacterized DNA, but these methods have limited utility. For instance, the dinucleotide CpG is unusually rare in eukaryotic DNA, except in the CpG islands that occur at the 5• ends of many genes. A high frequency of this dinucleotide in a region of uncharacterized DNA, therefore, suggests the presence of a eukaryotic promoter and a nearby coding region.

CpG frequency can be determined absolutely only by sequencing, which is impractical for sequences of a million base pairs or more. More l imited information regarding the location of CpG islands can be obtained using restriction enzymes that cleave sequences containing the CpG dinucleotide. This method has limited utility, however. The most CpG-sensitive of these enzymes, such as Mspl or Haelll, recognize a four base pair sequence, which occurs once every 250 base pairs, on average. Mapping several million base pairs with enzymes of this type is impractical, because the fragment

resulting from cleavage cannot readily be resolved. Moreover, the presence of CpG islands merely suggests that a gene is present, but does not provide conclusive information in this regard. Zoo blotting is another approach to identifying genes in large uncharacterized segments of DNA that is limited by difficulties similar to the problems that beset using CpG islands to locate genes in eukaryotic DNA.

Zoo blotting relies upon evolutionary conservation to identify functional DNA sequences. The approach is simple. DNA from numerous species is isolated, and each DNA is then fragmented by restriction enzymes. The restricted DNAs are sized on agarose gels and blotted. Sub-clones of the DNAs that encompass the chromosomal region of interest are used as probes of the blots. A sub-clone that hybridizes to fragment(s) in the DNAs of most or all of the different species in the blot is likely to contain a conserved sequence, such as a protein- coding region. Zoo blotting is tedious. It requires hundreds of sub-clones and blots to screen a million base pairs of DNA - just to identify regions of about 10,000 base pairs that contain conserved sequences. Further rounds of screening and sub-cloning are required to more particularly define the regions encoding a gene. Moreover, false positives often result when conserved middle-repetitive sequences hybridize across species. Certainty that a sub-clone encodes a gene depends on sequencing, northern blotting, and, ultimately detecting an expressed protein. In sum, it isn't possible using zoo blots efficiently to screen the very long stretches of DNA required to identify genes within a cytologically defined chromosomal locus.

Exon-trapping is yet another method for identifying genes in eukaryotic DNA. In this method, the presence in an uncharacterized genomic DNA of eukaryotic mRNA splicing signals serves as a signpost for the presence of protein coding regions. According to the method, a

specialized vector is used to detect splice site donor signals in the DNA. The vector contains a promoter, a cloning site, a splice site acceptor sequence, and a sensing gene arranged so that the sensing gene is not expressed unless a donor splice site is juxtaposed between the promoter and the acceptor site. Thus, the sensing gene is expressed only if a DNA inserted into the cloning site contains a splice site donor signal. Expression of the sensing gene, therefore, indicates the presence of a donor site, and the likely presence of a gene in the cloned fragment.

Although elegant, the method suffers from the necessity to sub-clone large DNAs. It is impractical to screen several million base pairs of DNA by this method because of the large number of sub-clones that must be examined. Furthermore, even when the likely presence of a gene in a sub-clone is indicated by expression of the sensing gene, additional characterization must be carried out to verify the presence in the clone of a protein coding region.

Furthermore, as with the other methods discussed above, even when a putative protein-coding region in a genomic clone is detected, a cDNA still must be obtained to express and characterize, and, in most instances, to commercially exploit the gene product. While such cDNA cloning is certainly practical for one or a few cDNAs, it is not feasible to obtain by this method all the cDNAs that map to a cytological location.

In sum, none of the techniques presently known to the art is useful to identify the gene-encoding sequences in regions of DNA more than several ten-thousand base pairs long. Moreover, even for this purpose, the available techniques are inconvenient, insensitive and exhibit poor selectivity. It is impractical or impossible, therefore, to identify the genes in a continuous, cytologically defined region of a chromosome using these methods.

Accordingly, there is a need for a method to generate region-specific DNA libraries and to identify the

protein-encoding sequences in the chromosome region. In particular there is a need for a general method of making chromosome region-specific cDNA libraries that contain only the cDNAs that hybridize to a chromosome region of interest, such as a region associated with a genetic dysfunction of interest.

This problem is not limited to studies of human molecular genetics. Rather, it is a general problem in molecular biology research which is encountered in studying the molecular genetics of other eukaryotic organisms, including, for instance, C. elegans, D. melanogaster, mice, rats, pigs and sheep, to name just a few well-studied animals, and arabidopsis , corn, wheat and tomato, to name just a few better studied plants. In fact, there is more complete orphogenetic and chromosome mapping information in a number of these organisms than in man, and a method for obtaining cDNAs associated with cytologically defined chromosomal loci would be a particularly powerful tool to advance our understanding of molecular genetics in these organisms, as well as in humans.

SUMMARY OF THE INVENTION It is therefore an object of the present invention to provide cDNA libraries consisting essentially of cDNAs that are derived from the same specific region of a chromosome.

It is a further object of the present invention to provide a method to obtain cDNA libraries that consist essentially of cDNAs that come from the same specific region of a chromosome.

In accomplishing the foregoing objects, there has been provided, according to one aspect of the present invention, a method of making a chromosome region- specific DNA library consisting essentially of a plurality of clones comprising DNAs that hybridize under stringent conditions to the same region of a chromosome, the method comprising the steps of: (A) providing a probe DNA library for hybridization composed of a

plurality of DNA clones; then (B) hybridizing the DNA library to chromosomes in situ under stringent conditions; then (C) microdissecting the chromosomes to obtain specific sub-chromosomal regions to which clones of the plurality are hybridized; thereafter, (D) amplifying DNAs of the clones; and then (E) cloning the amplified DNAs.

There has also been provided in accordance with the invention a method of making a chromosome region-specific cDNA library consisting essentially of a plurality of clones comprising cDNAs that hybridize under stringent conditions to the same region of a chromosome, comprising the steps of: (A) providing a probe cDNA library for hybridization composed of a plurality of cDNA clones; then (B) hybridizing the cDNA library to chromosomes in situ under stringent conditions; then (C) microdissecting the chromosomes to obtain specific sub-chromosomal regions to which clones of the plurality are hybridized; thereafter (D) amplifying cDNAs of the clones; and then (E) cloning the amplified cDNAs.

There has also been provided chromosome region- specific cDNA libraries that consist essentially of a plurality of clones that comprise substantially all the different cDNAs in a probe cDNA library that hybridize to the same region of a chromosome.

In particular, in this regard, the invention provides region-specific cDNA libraries wherein the region is defined by banding or size. Thus, the invention provides region specific libraries wherein the region is approximately 100,000,000 to 500,000,000 base pairs, approximately 10,000,000 to 100,000,000 base pairs, or approximately 1,000,000 to 10,000,000 base pairs.

Furthermore, the invention provides chromosome region-specific cDNA libraries made using probe cDNA libraries that are normalized cDNA libraries, tissue specific cDNA libraries, mixtures of tissue specific

cDNA libraries, and developmental stage-specific cDNA libraries.

Other objects, features, and advantages of the present invention will become apparent from the following detailed description. It should be understood, however, that the detailed description and the specific examples, while indicating preferred embodiments of the invention, are given by way of illustration only, since various changes and modifications within the spirit and scope of the invention will become apparent to those skilled in the art from this detailed description.

BRIEF DESCRIPTION OF THE DRAWINGS FIGURE 1 schematically depicts an embodiment of the present invention. GLOSSARY cDNA LIBRARY: A collection of clones, which may be cell colonies or infected cells (e.g., phage plaques), inter alia , each of which propagates an individual cDNA as an insert in a plasmid or a phage-derived vector. A complete or representative library, as the term is used herein, is a library that has a high probability ("P") of containing cDNAs corresponding to each different mRNA that was present in the mRNA population from which the cDNA was made. Ordinarily, a high probability in this respect means, approximately, P = .90 (90% chance) or better. Preferably, P for any cDNA in a library will be .95. Even more preferably, P will be .98.

It will be appreciated that P for an mRNA will depend on the frequency of the mRNA in the original mRNA population, the relative efficiency of cloning the mRNA, how the library is amplified, how it is propagated, and how often, inter alia . cDNA libraries usually are made using mRNA from a specific tissue or a particular type of cell. Moreover, the tissue or cell-type generally will be obtained from an organism at a particular stage of development. Accordingly, cDNA libraries made from mRNA obtained from

a sample from an organism generally will be cell-type and developmental stage-specific. cDNA libraries made from cell lines may best be considered cell specific. CHROMOSOME REGION-SPECIFIC: Mapping to (as by hybridization) or located in a specific part of a chromosome, i.e., a sub-chromosomal region. Notably a minority of cDNAS will map to more than one specific chromosome region, on more than one chromosome.

The region may be defined cytologically, as by banding patterns, being a band or interband, for instance. Other methods may be used to define a region of interest, such as distance from the centromere or telomere, or other reproducible morphological features. It will be appreciated that the degree of specificity is determined by the reproducible ability to identify subchromosomal regions and to microdissect the regions reproducibly for amplification of hybridized sequences. Resolution will be affected by the type of chromosome, its stage in the cell-cycle, methods of fixation and staining, as well as the resolving power of the microdissection technique. Nonetheless, the invention can be practiced with virtually all techniques for obtaining, fixing, staining and microdissecting chromosomes, provided only that the techniques do not interfere with hybridization of the probe cDNA libraries or amplification of hybridized cDNA after microdissection. PCR: Polymerase chain reaction.

DETAILED DESCRIPTION OF THE INVENTION The present invention overcomes many of the disadvantages of conventional methodology by providing chromosome region-specific cDNA libraries that contain essentially all of the cDNAs in a probe cDNA library that hybridize to a specific region of a chromosome. Essentially, in its most highly preferred embodiment, the method provides chromosome region-specific cDNA libraries by: (1) providing chromosome spreads suitable for hybridization and micro-dissection; (2) probe cDNA

libraries having vector sequences flanking the cloning site that can be used for amplification; (3) hybridizing the cDNA library probe to the chromosome spreads under conditions that maintain the spreads in a condition suitable for subsequent microdissection; (4) removing unhybridized and spuriously hybridized probe following hybridization; (5) microdissecting the chromosome(s) to obtain specific regions, including the cDNAs hybridized to the region; (6) carrying out the PCR, or another amplification procedure, to specifically amplify the hybridized cDNAs; (7) recloning the amplification products to propagate the chromosome region-specific cDNA library.

Each of these steps can be elaborated in a variety of ways, as set forth below, and in previously cited references, and additional steps can be added. For instance, the chromosomes can be stained to visualize banding patterns to facilitate microdissection, either before or after hybridization, as set forth hereinbelow. Furthermore, the method can be used to obtain other types of chromosome region-specific libraries, in addition to cDNA libraries, which also is set forth hereinbelow. Probes

Practically any library can be used in accordance with the invention. In particular, any cDNA library can be used for hybridization in accordance with the invention. Libraries useful in the invention can be made by techniques well known in the art, such as those described in Maniatis et al . , MOLECULAR CLONING, A LABORATORY MANUAL (Cold Spring Harbor Laboratory, 1982) and Sambrook et al . , MOLECULAR CLONING, A LABORATORY MANUAL, Second Edition, Vol. 1-3 (Cold Spring Harbor Laboratory, 1989) , for instance.

In addition, commercially available libraries, such as those from the American Type Culture Collection and "CLONTECH," for instance, can be used in accordance with the invention. In fact, any library can be used in the

invention to isolate from the library a sub-population of DNAs hybridizing to a specific sub-chromosomal region.

In accordance with one preferred embodiment of the invention a normalized cDNA library is used for in situ hybridization. See Patanjali et al. (1991) Proc . Nat ' l . Acad . Sci . , U.S .A. .88. 1943-1947. Such libraries are useful to even-out the representation of different mRNAs in the cDNA library so that the copy numbers of the different cDNAs in the library are more nearly equal. When the copy numbers are approximately the same for different cDNAs, as in normalized libraries, they hybridize to complementary sequences in the chromosomes with approximately the same kinetics. This helps to insure that all the different cDNAs in the library have hybridized to a complementary sequence under conditions which insure that the bulk of the cDNA has hybridized to the chromosomes.

It will be appreciated by one of ordinary skill in the art that the mitotic chromosome pairs of a diploid organism contain a total of four copies of each single- copy gene. These four copies constitute the entire target set of sequences for a given clone of the probing library. Hybridization of substantially all the cDNAs in a cDNA library to substantially all the complementary genes in the chromosomes can be facilitated by using a library which minimizes the kinetic differences between the different cDNAs, such as a normalized library.

Another preferred embodiment of the invention provides chromosome region-specific cDNA libraries that are tissue-specific, i.e. that contain cDNAs of genes expressed exclusively in particular tissues. Such preferred libraries can be made by using for hybridization a cDNA library made from the mRNA from a purified cell-type or tissue. More refined chromosome- region specific libraries can be made by using for hybridization a library consisting of cΘNAs expressed more exclusively in a tissue or cell type of interest, such as those made by hybridization-subtraction

techniques. See, for instance, Sargent, T. D. "[46] Isolation of Differentially Expressed Gene," Methods in Enzymology 152: 423 (1987) . In many instances, in addition, such libraries will also be normalized libraries.

Another preferred embodiment of the invention provides subchromosome region-specific cDNA libraries that are developmentally specific. In accordance with the invention, developmentally specific libraries can be made by using for in situ hybridization a developmental stage-specific cDNA library. Libraries of this type can be made from mRNA isolated from developmentally synchronized embryos at a particular stage of development. Analogously to tissue-specific libraries, developmental stage-specific libraries also can be made by subtraction techniques, and such libraries can, in addition, be normalized libraries. Thus, developmental stage and tissue-specific libraries can be made and used as probes to derive chromosome region specific libraries that are developmental stage and/or tissue-specific.

In another application of the invention, chromosome region-specific cDNA expression libraries can be made by probing chromosomes in accordance with the invention with cDNA libraries carried in expression vectors, such as vectors that can direct expression of a cloned DNA sequence in bacterial, yeast, insect or mammalian cells, among others. Probing with libraries of this type facilitates direct expression of the proteins produced by the genes in a specific chromosome region. It will be appreciated that using expression vectors in the invention in this way will facilitate identification and characterization of proteins encoded in a specific chromosomal region by physical, enzymatic and immunological techniques. Of course, the hybridizing cDNAs always can be cloned into an expression vector after amplification.

In addition, other types of DNA libraries can be used in accordance with the invention method. In fact, any

library that can be hybridized stably to the DNA of an identifiable region of a chromosome can be used as a probe in the invention, provided the inserts can be amplified. Thus, genomic DNA libraries also can be used as probes in the invention. It will be appreciated that many of the above-discussed considerations concerning cDNA libraries also apply to other libraries that can be used in the invention. Thus, normalization, the use of expression vectors, and the use of selected sub-libraries all can be applied to non-cDNA library probes.

For instance, linking libraries may be used as probes in the invention to obtain a sub-library that hybridizes to a specific chromosome region. In accordance with the invention, a linking library is hybridized to chromosome spreads, specific regions of the chromosomes are obtained by microdissection, and the linking libraries that hybridize to each region are amplified and subcloned. The resulting sublibraries have reduced complexity, compared to the total probe linking library. They can therefore more readily be assembled into an ordered library. Similarly, the sublibrary will have increased utility for ordering other libraries, such as YAC, cosmid and phage libraries.

In the same vein, libraries produced by repeat sequence-specific amplification primers can be reduced in complexity to correspond to individual chromosome bands and regions by means of the present invention. Vectors

As is well known in the art, and in accordance with the present invention, cDNA's in practically any vector can be hybridized to chromosomes for micro-dissection. The cDNAs hybridizing to a chromosome region obtained by micro-dissection in accordance with the invention are amplified, for instance, by the PCR. When the PCR is used primer pairs of the PCR will hybridize to regions in the vector flanking the cDNA inserts, so that generally the cDNA inserts will be amplified and not the vector portion of the clones. Of course it may be desirable in

some instances to co-amplify certain vector sequences, such as multicloning sites and expression control elements.

Suitable flanking sequences are present in all cloning vectors, and oligonucleotides for priming the PCR can be designed based on the DNA sequences of the flanking regions, which may be published or determined by

DNA sequencing methods well known in the art.

Of course, primers are sold that hybridize to the sequences flanking the cloning sites in many commercially available vectors. They typically are used to determine the sequences of insert fragments, but will serve quite well also to carry out the PCR and may be used in accordance with the invention. Probe Preparation

Libraries may be used for hybridization, in accordance with the invention, as follows. DNA from the libraries is first prepared. Methods for preparing total DNA from a plasmid or phage cDNA library are well known to the art and are described in Maniatis, supra, among other references.

The cDNA inserts in the cDNA library preferably are amplified prior to hybridization; although a whole library DNA may be employed without amplification. Amplification generally is carried out by the PCR using primer pairs that flank the cloning sites in the vector.

When phage cDNA libraries, such as lambda gtlO or gtll cDNA libraries are used, most conveniently, purified phage can be lysed in water and amplification carried out on the liberated DNA, upon addition of nucleotide triphosphates, buffer components, DNA polymerase, and co- factors.

Amplification can be assessed by comparing the amplification products with appropriate controls by gel electrophoresis followed by ethidium bromide staining. Generally, when inserts in a high quality library have been successfully amplified, the pattern of amplification products will resemble the pattern of unamplified inserts

from the original library, and the pattern of mRNAs in the tissue from which the library was prepared. Chromosomes

Chromosomes from practically any source can be prepared for hybridization by a variety of well known cytological methods in accordance with the invention.

It will be appreciated that the invention can be applied to all organisms from which mitotic, meiotic or interphase chromosomes having morphology suitable to micro-dissection can be prepared. Suitable cytogenetic preparations can be made from practically any eukaryotic organisms including, for instance, Caenorhabditis , Drosophila , chickens, mice, rat, pig, sheep, monkeys, chimpanzees and humans, among others, and from practically any plant, including, for instance, Arabidopsis , and important crop plants such as wheat, corn, soybeans, rice, alfalfa, tobacco, oats, rapeseed, barley and grapes, to name just a few.

Techniques for making suitable preparations are described in numerous publications available to the skilled artisan. For instance, to name just four useful compendia of such techniques, Brown, W. V., TEXTBOOK OF CYTOGENETICS, The C. V. Mosby Co. (1972), DuPraw, E. J. , DNA AND CHROMOSOMES, Holt, Rinehart and Winston, Inc. (1980), Bernard et al . , CHROMOSOME HIERARCHY, AN INTRODUCTION TO THE BIOLOGY OF CHROMOSOMES, Oxford University Press, Ely House, London (1975) , and Therman, E., HUMAN CHROMOSOMES; STRUCTURE, BEHAVIOR, EFFECTS, Springer-Verlag (1980) describe general cytological methods for making such preparations from cells and tissues of a wide variety of organisms. It will be appreciated that the technology described in these references, and elsewhere largely is compatible with the other methods required to practice the present invention. Typically, metaphase chromosomes will be preferred for micro-dissection, but the techniques used for metaphase chromosomes also can be adapted to chromosomes in earlier stages of mitosis, particularly prophase,

prometaphase, and early metaphase. Chromosomes in these stages of mitosis offer the possibly of higher resolution microdissection than typical mitotic spreads. The extended morphology of chromosomes during these earlier stages of mitosis results in more highly resolved banding patterns that can be dissected more precisely. Thus, in accordance with the present invention, chromosomes in these stages of mitosis offer the possibility of creating cDNA libraries that map to more finely defined chromosome regions, such as chromosome sub-bands.

In addition, as stated hereinabove, other types of cytogenetic preparation suitable for microdissection may also be used in the invention, including preparations of meiotic or interphase chromosomes. Meiotic chromosomes of many mammalian species, including the mouse, for instance, can be fixed, stained, hybridized and microdissected as described herein for mitotic chromosomes, using similar techniques. Likewise, polytene chromosomes of interphase cells of a variety of organisms, particularly Dipterans such as Drosophila melanogaster, can be fixed, stained, hybridized and microdissected in accordance with the present invention.

In fact, in Drosophila the interphase (polytene) chromosomes in some cells present a very well characterized banding pattern that can serve as a guidepost for reproducibly microdissecting the chromosomes into small regions that may contain only a few genes, or even one gene. Thus, by means of the invention using these cells it may be possible efficiently to isolate a unique coding sequence within a cytologically distinct region of a chromosome.

Cytological preparations for use in the present invention can be made from a wide variety of tissues and cell lines, inter alia . In some applications, cell lines derived from blood or other tissue are preferred as the starting material for cytogenetic preparations for in situ hybridization. Peripheral blood is especially

preferred in some instances, because it can be obtained without harm to most mammals, and because standard techniques have been developed for use in clinical hematology for making, routinely, very high quality blood-derived chromosome preparations suitable for use in the present invention. In plants the root tissue may be preferred.

In sum, a wide variety of techniques are well known to the art for preparing from many different eukaryotes, including animals and plants, cytogenetic preparations that can be used for hybridization in accordance with the invention. These techniques can all be usefully employed in the present invention to prepare chromosomes for hybridization to a probe library. Often it will be particularly convenient to use cells cultured in vitro for preparing chromosomes for micro¬ dissection. To reduce the probability of chromosome rearrangement primary cultures are preferred. Human peripheral lymphocytes are particularly favorable for human chromosomes. Primary fibroblast cultures are highly preferred for mouse chromosomes. Where a cDNA mapping to a chromosome abnormality is desired cell lines that carry the abnormality are preferred. Where a chromosome region has been identified in a cell line as being of interest that cell will be most preferred.

In some instances it will be advantageous to use reagents that arrest cells in metaphase, thereby increasing the ease of preparing metaphase chromosomes for hybridization and high-resolution dissection. A number of compounds well known to the art can be used to accumulate cells in the mitotic phase of the cell cycle for cytogenetic analysis. These substances include Colcemid and colchicine compounds, which are useful in this regard for both animal and plant cells. Typically, cells growing in culture (isolated from an organism of interest) are allowed to proceed through the cell cycle at mitosis. After the cells are blocked in mitosis they are treated with a hypotonic solution,

such as .75 mM potassium chloride, that causes them to swell but not burst. Following hypotonic treatment, the cells typically are fixed by exposure to a non-aqueous fixative solution. A preferred fixative is 3:1 (v:v) methanol:glacial acetic acid. Fixation may be repeated to insure proper spreading of the chromosomes. Repeating the treatment three times often gives satisfactory results. Droplets of the fixed cells, suspended in the fixative, are applied to clean glass slides or coverslips and allowed to dry. As a result of this procedure the chromosomes are spread suitably for hybridization chromosome banding and microdissection.

In some cases it may be desirable to stain the chromosome spreads before hybridization of probe. This is preferred when it is desirable to inspect the integrity of the spreads before hybridization. This will be preferred when using a probe that is difficult to obtain in large amounts and when the chromosomes are difficult to prepare, for instance. Thus, chromosomes may be stained to visualize subchromosomal features as an aid to assessing their integrity and quality, and to aid microdissection. A variety of well-known cytogenetic techniques are useful in this regard. Particularly useful and preferred in the invention are Wright's and other dyes that produce well- characterized banding patterns.

The banding patterns produced by the stains serve as reproducible maps that identify individual chromosomes, and distinguish particular segments of each chromosome. The quality of the banding pattern in a preparation provides an indication of the quality of the stained chromosomes. The banding pattern, moreover, serves as an aid to dissecting chromosomes reproducibly, so that the same region can be obtained from several chromosomes for pooling in a single amplification reaction.

Whether staining is carried out before or after hybridization, Wright's stain may be used for this

purpose, as may many other stains. A typical useful staining procedure is as follows.

Wight's stain is mixed 1:3 with .06 M phosphate pH 6.8 and then applied to the chromosome spreads on a slide or coverslip for 5 to 10 minutes at room temperature. The spreads are then washed briefly with water and air dried. This treatment generally yields satisfactory banding patterns.

Also preferred in the invention is Giemsa staining, without or with pretreatment with protease. Fluorescent dyes, such as DAPI can also be used to stain chromosomes to visualize banding patterns to assess quality or guide microdissection in accordance with the invention. Removing RNA, Denaturation and Hybridization A wide variety of techniques are known for hybridizing a polynucleotide probe to chromosomes in situ . Most of these techniques, if not all of them, can be used in accordance with the present invention. It will be appreciated that each chromosome contains only one copy of each DNA strand of a genetic target sequence.

Moreover, the chromosomes are immobile. The extremely low copy number and diffusive inertness of target sequences in the chromosomes means that the rate of hybridization and the extent of a reaction upon arrest will be determined largely by the concentration of DNAs in the probe library, and the hybridization conditions.

Generally, hybridization will be carried out under stringent conditions of salt concentration, pH and temperature. High concentrations of probe DNAs will favor more rapid and complete reaction. It will be appreciated that the kinetics of hybridization of individual DNAs in a library will be proportional to their relative copy number. Rare DNAs will hybridize the slowest, abundant DNAs the fastest. As reactions are prolonged, however, the possibility of spurious hybridization increases. Thus, as is well understood in the art, a tradeoff may be required between completeness and spurious hybridization. Minimizing spurious

hybridization will increase the risk that relatively rare DNAs will not stably hybridize to their cognate genes in the chromosomes. Prolonging a reaction to insure that all DNAs in the probe hybridize to cognate sequences in the chromosomes, increases the risk of high background hybridization. These tradeoffs are well known in the art, and it is generally possible to select conditions that allow substantially all the different DNAs in a probe library to hybridize to complementary chromosomal sequences, even though it is never possible to be sure that absolutely all the different DNAs have done so. Methods are available to estimate the conditions required to achieve a given level of confidence of hybridization of a probe population. Moreover, specialized libraries can be employed, such as normalized libraries to increase the likelihood that substantially all the DNAs in a library have hybridized to a chromosomal spread.

Accordingly, it will be appreciated that hybridization conditions well known in the art, properly adjusted, provide efficient hybridization between probe sequences and chromosomes. Thus, in the context of the present invention, hybridization of a probe library to chromosomes in situ , under stringent conditions and vast probe excess, taking into account the need to minimize spurious hybridization, results in hybridization of substantially all the DNAs in the probe library to the chromosomes. This definition applies to all types of probe libraries.

Many procedures for hybridizing a probe to a chromosome spread are compatible with the invention. The following typical procedure is provided for illustrative purposes only.

Chromosome spreads prepared as described above, are treated with RNase A to remove RNA that would serve as a spurious target for hybridization of the cDNA probe. Typically, to remove effectively all the RNA, mitotic chromosomes on a slide are incubated with 100 ng/ml RNase

in 2X SSC for about one hour at 37 °C under a coverslip in a moist chamber.

After RNase treatment the chromosome preparations are dehydrated, generally by immersing the chromosomes in a series of aqueous ethanol solutions of increasing concentration, typically beginning with 70% ethanol and ending with 95% ethanol. After the last ethanol wash, the slides are dried by allowing the ethanol to evaporate. DNA in the chromosomes is then denatured. Preferably, the chromosomes are incubated in 70% formamide, 2 X SSC for approximately five minutes at 75°C. Immediately upon completion of the incubation at 70°C the slides are immersed in ice cold 70% ethanol, and then ice cold 95% ethanol. It will be readily appreciated that low temperature ethanol treatment arrests the reassociation process, keeping the DNA denatured.

Hybridization typically is carried out in a 9:1 mixture of hybridization buffer. 9 mL of hybridization buffer may be prepared by mixing: 5 mL of nucleic acid grade, redistilled 100% formamide, 0.2 L 50X Denhardt's solution, 0.8 mL 2OX SSC, 0.4 mL 1 M phosphate buffer, pH 6.8, 2.5 mL 50% dextran sulfate, and 0.1 L 10% SDS. One volume of probe is generally diluted into nine volumes of hybridization buffer for hybridization. Approximately 50 microliters of this probe mix is then added to each slide and covered by a coverslip. Hybridization then is carried out overnight at 37°C in a moist chamber.

Following hybridization the chromosomes are washed twice with 50% formamide, 2X SSC at 42°C for approximately 10 minutes. The chromosomes are then washed at least two times with 2X SSC at room temperature. Removal of unhybridized probe is critical, and thorough washing at this stage is important. Therefore, to ameliorate problems caused by spurious hybridization, a high stringency wash may be employed at

this stage to melt and remove partially matched hybridization products.

After washing, the chromosomes are dehydrated by immersion in 70% and then 90% ethanol for several minutes at room temperature prior to microdissection. Micro-dissection

Following fixing, staining and hybridization, as described above, the cytogenetic preparations are microdissected. Techniques suitable for carrying out microdissection in accordance with the invention are well known in the art.

Typically the chromosomes are on slides or coverslips, which may be mounted in conventional manner on a standard microscope stage. Visualization likewise can be accomplished by microscopes typically employed for cytogenetic experiments. It will be appreciated that the highest resolution microscopes will be the most advantageous in the present invention because they provide the clearest images of the chromosomes and thereby facilitate the highest resolution microdissection.

Microdissection itself requiresmicromanipulators and other microinstruments suitable for chromosome microdissection, which are well known to the art. Micromanipulators useful in the invention include mechanical, hydraulic and computerized stepper-motor driven micromanipulators. Microtools for dissecting the chromosomes can be formed using any suitable microforge, such as the De Fronbourne microforge. Drawn glass microdissection needles that are made on the microforge are mounted in the micromanipulator and then applied to the hybridized and stained chromosomes under high magnification to obtain chromosome fragments from specific segments of each chromosome. The process is observed microscopically and the banding pattern on chromosomes is used as a guide during (or before) microdissection to identify the chromosome^region of each

segment as it is removed during the microdissection process.

In addition to mechanical microdissection, a number of other techniques may be used to obtain specific regions of chromosomes.

One alternative- technique employs droplets of a denaturing agent to melt probe away from specific chromosomal regions. The droplets are applied to the region using conventional micromanipulation methods. Thus, the technique proceeds in much the same fashion as mechanical microdissection, using similar apparatus. The denaturing solution is applied to the chromosome using a micropipet that can be made from capillary tubing by methods well known to the art. After an appropriate period of incubation, which melts hybridized probe away from the chromosomal region and into the microdrop, the droplet is removed either by the same pipet or by another pipet. After neutralization the cDNA in each microdrop can be amplified and subcloned, as described below. In yet another embodiment of the invention, laser beams may be employed to cut or vaporize the chromosomes. Thus, lasers can be used to physically separate specific segments of a chromosome by "cutting" between them. Alternatively, lasers can be used to ablate completely all of a chromosome except a specific region, thus, providing that region only for further processing.

In light of the foregoing, it will be appreciated that any method for reproducibly obtaining specific chromosome regions can be employed in accordance with the present invention. It will moreover be appreciated that the resolution of a microdissection technique will determine the size of the region represented by the libraries of the present invention, and, therefore, the average number of clones in region-specific libraries made by a particular technique.

When the highest resolution techniques are applied to the best defined regions of a chromosome it is possible reproducibly to resolve fragments as small as

1,000,000 base pairs. Under slightly sub-optimal conditions, whether caused by sub-optimal equipment or imperfect chromosome morphology, or both, chromosome regions of approximately 10,000,000 to 100,000,000. More routinely, it is possible to resolve chromosome regions of approximately 100,000,0000 to 500,000,000 base pairs. It will be appreciated readily by those of ordinary skill that libraries made under these conditions in accordance with the present invention will be specific, respectively, to chromosomal regions of approximately 1,000,000 to 10,000,000 base pairs, approximately 10,000,000 to 100,000,000 base pairs, and approximately 100,000,0000 to 500,000,000 base pairs. From this it will be clear that the chromosome region specificity of a library depends on the specificity of the microdissection techniques.

Typically, each chromosome arm can be divided reproducibly into between one and fifty or more fragments. The number depends on the size of the arms and the separation between bands, as well as the resolution of the microdissection technique.

As each fragment is obtained it is transferred to a microcentrifuge tube, or other suitable container for carrying out subsequent steps in the procedure. After the fragments are collected, and those from the same region are pooled, the hybridized probe in each region is amplified, currently by the polymerase chain reaction (PCR) . Of course, other amplification techniques also may be used to amplify the inserts in the clones that hybridized to the fragment. Amplification

The cDNAs in the micro-dissected chromosomal DNAs are then amplified. Presently the most highly preferred way to amplify the cDNA is by the PCR. Accordingly, primer pairs that hybridize solely to the library vector are hybridized to micro-dissected material at sequences flanking the cDNA inserts. The PCR is then carried out, resulting in amplification of the cDNA inserts. Primers

that do not spuriously hybridize to chromosomal DNA are used. The primer binding sites in the cDNA library vector preferably are proximal to the cDNA inserts, but flank any multi-cloning site sequences adjacent the inserts. This provides for amplification of the multicloning sites with the insert, facilitating subsequent recloning of the amplification product. When a commercially available vector is used, often sequencing and reverse-sequencing primers will be available that may be used for amplification. The requisite techniques for amplification and subcloning are well known in the art, as set forth for instance in Maniatis and Sambrook, cited above. Sequencing can also be carried out using these primers, which will provide "sequence tags" for each clone. Regarding sequence tags, see Adams et al . Science 252 1651 (1991) , for instance. Subcloninσ

Following amplification, generally it will be desirable in accordance with the invention to subclone the amplified cDNAs. Practically any vector may be used for this purpose. Also, any of a wide variety of sub¬ cloning techniques can be used. Generally, the multicloning sequence or linker, as the case may be, can be cleaved with a restriction enzyme and the ends generated by the cleavage can be annealed to complementary ends generated in the vector. The annealed ends then can be ligated, and the ligated DNA transformed into a desired host. The transformed cells then can be spread, most preferably at low density for amplification, archival or further experimental use, inter alia.

The human genome contains 24 different chromosomes, which vary considerably in size and morphology. Present micro-dissection techniques reliably can provide an average of 10 or more substantially independent sub- chromosomal regions per chromosome, i.e., more than 240 for the genome as a whole. It has been estimated that the entire human mRNA complement is about 100,000 mRNAs.

Thus, it can be calculated that the typical chromosome region-specific library of cDNAs will contain less than 400 cDNAs. This estimate does not take into account the differences between transcription in different types of cells. Since particular cell types express only a subset of all the genes in the genome, the average number of genes in a chromosome region-specific library will be much less than all the genes in a region when the region- specific cDNA library is made with a tissue-specific cDNA library probe.

The present invention is further described by reference to the following, illustrative examples. EXAMPLE 1: Amplification of cDNAs for in situ hybridization The cDNAs in human and mouse lambda gtlO and lambda gtll libraries were amplified using the PCR. Phage were lysed by diluting them into 250 microliters of H ₂ 0 and then freezing and thawing them twice. Lambda gtlO or gtll forward and reverse insert screening primers were added to the lysed phage. 50 microliters of the reaction contained 25 microliters of lysed phage and 1.25 units of Tag polymerase in 10 mM Tris pH 8.3, 1.5 mM MgCl ₂ , 50 mM KCl, 0.2 μM each of ATP, CTP, GTP and TTP and each primer at 1 μM. Thermal cycling was then carried out using an automated thermal cycler designed for performing the PCR.

To determine the fidelity of amplification, the amplification products were examined by electrophoresis through 0.9% agarose, followed by ethidium staining.

Typically, successful amplification products were seen as a broad smear ranging in size from 300 bp to 2.3 kb, as seen under typical conditions of UV illumination. A primer-dimer also was seen, in varying amounts.

Negative control reactions were carried out in the same way using gtlO and gtll vector (i .e . , DNAs without inserts) . No cDNA smear resulted from these reactions. Positive controls also were carried out. Individual insert-containing lambda gtlO and gtll clones were picked from agar plates, lysed and amplified as above. Discrete

bands were seen for each of the different clones, corresponding to the insert sizes expected.

Although satisfactory results generally were achieved with a standard protocol, conditions also were varied to optimize the amplification of different libraries. Primer concentration was varied between 0.1 and 1 μM, magnesium ion concentration was varied between 1.5 and 3 mM, and library DNA concentration was varied between 1 and 10 microliters of the original library. In addition, the extension time was varied from 1.25 to 1.75 minutes, in accordance the average size of the cDNA inserts in the library. The conditions were adjusted to maximize cDNA product and minimize primer-dimer. EXAMPLE 2: Hybridization and microdissection of a cytogenetic preparation

Mitotic chromosomes of human and mouse cells were prepared and spread on slides. The spreads were treated with RNAse A for one hour to remove RNA, and then dehydrated in 70%, followed by 95% ethanol. The spreads were then thermally denatured at 75°C for five minutes in 70% formamide, 2X SSC, followed by immediate immersion in ice cold 70% and then 95% ethanol. After chilling dehydration the spreads were deproteinized by treatment with proteinase K for about five to ten minutes, and then they were again dehydrated in 70% and 95% ethanol. After the final dehydration the spreads were air-dried using an air-jet, prior to hybridization.

About 200 nanograms to one microgram of probe cDNA, prepared according to EXAMPLE 1, together with 500 nanograms to 10 micrograms of carrier, salmon sperm, DNA in 50ul of hybridization mix was placed over each chromosome spread and covered with a cover slip, as described herein above. Hybridization was carried out overnight at 37°C in moist chamber. Control hybridizations containing only carrier DNA, or no DNA at all, also were carried out.

After overnight incubation, the cover-slips were removed, and the spreads are washed twice at 42°C for ten

minutes with 50% formamide, 2X SSC. After this the spreads were washed twice for 10 minutes in 2X SSC, and then dehydrated in 70%, followed by 95% ethanol. Finally, the spreads were air-dried using an air-jet. After hybridization, but prior to microdissection, the spreads were stained. For staining, Wright's Stain was diluted 1:3 in 0.06 M phosphate buffer pH 6.8, and the spreads were incubated in the diluted stain for 5 to 10 minutes at room temperature. The spreads were then washed thoroughly with water and air dried.

Microdissection was carried out using a standard cell micromanipulation station. One or two individual nuclei were placed into reaction tubes for carrying out the PCR. The PCR was then carried out as described in EXAMPLE 1. The PCR products were visualized as described in EXAMPLE l. When amplified gtlO or gtll cDNA libraries were used for hybridization the PCR products from the nuclei following hybridization and microdissection gave the same agarose pattern as the probe library. When only carrier DNA or no DNA was used in the hybridization, no amplification product was obtained. In sum, the amplified cDNAs hybridized to the mitotic chromosomes and were seen to be re-amplified with high fidelity. EXAMPLE 3: 17q25 specific cDNA library Band 17q25 at the distal end of chromosome 17 is identified by clinical analysis and cytological experiments as containing a disease-associated gene. To clone the disease associated DNA's, cDNAs which hybridize to the 17q25 region are cloned from cDNA libraries made from cells likely to express the disease-associated gene and also from libraries made from cells that likely do not express the gene. Differential expression provides important information helpful to the isolation of the disease-associated gene(s) . Human mitotic chromosomes are prepared. The chromosomes are fixed and prepared for in situ hybridization. Each of several cDNA libraries is prepared essentially as set forth in EXAMPLE 1 and

hybridized to several individual morphologically intact, well-spread human mitotic chromosome preparations. Following hybridization, the chromosomes are microdissected to obtain several 17q25-specific fragments from each hybridization.

The PCR is carried out essentially as described in EXAMPLE 2. PCR primers are added to the fragments. The primers hybridize to regions in the probe vectors flanking the inserts and multicloning sequences (MCS) . The PCR is then carried out using a commercial thermal cycling apparatus.

DNA amplified by the PCR and derived from the cDNA inserts in the probe vectors is digested with a restriction enzyme that cuts infrequently in random sequence DNA but cuts at a site in the MCS. The restricted DNA is then sub-cloned into a secondary vector. 50 clones are obtained, which when hybridized en mass to mitotic spreads give signals only over 17q25. Of these 50 only two are differentially expressed. Further study reveals that the two clones are derived from the same gene, which encodes a membrane-bound GTPase that is dysfunctional in the disease associated with 17q25 abnormalities. EXAMPLE 4: Genetic linkage analysis is used to narrow the region of the genome associated with a disease to a multiband region of the genome, which is nearly the entire long arm of chromosome 16, I6q. Further experiments are carried out to develop polymorphic markers and to further define the region of 16q associated with the disease phenotype.

Chromosomes are prepared for hybridization as described hereinabove. At the same time a representative genomic library is prepared as a probe to hybridize to the chromosomes. C0T1 DNA is included in the hybridization solution to reduce signal to noise problems caused by repetitive genomic elements that rapidly hybridize throughout chromosomal DNA. Following hybridization the chromosomes are microdissected. In

particular, the long arm of chromosome 16 individual bands of 16q likewise are obtained and, individually, introduced into containers for subsequent experimentation. Some 16q arms are pooled, but for the most part, 16q segments fragments are processed individually.

The microdissected DNA's are amplified. Primers that anneal to regions flanking the insertion site in the vector, including the MCS, are added to the DNAs. Nucleotide triphosphates and a thermal-stable DNA polymerase are added, and the PCR is carried out using an automated thermal cycling apparatus.

Following the last cycle of the PCR, amplification products are analyzed by gel-electrophoresis as described in Examples 1 and 2. Analysis of the 16q arm products reveals a smear of fragments derived from the 1,000 or so clones that hybridized to the chromosome arm. The average size of the amplified microdissected cDNAs is similar to the average insert size in the probe library, as expected. Amplification products from smaller chromosomal fragments exhibit less complexity, also as expected, since, in general fewer clones hybridize to smaller regions.

The amplification products are subjected to restriction for sub-cloning. A vector for sub-cloning is digested with an enzyme that generates compatible ends. The vector and the digested fragments are ligated together and then transformed into a host cell. The transformed cells are plated at low density to form a chromosome band-specific sub-library for each hybridization- icrodissection experiment.

Individual clones from each library are selected at random, cultured, and DNA is prepared from them. These DNAs are used to probe YAC libraries to identify YAC clones that map to 16q and particular bands in 16q.