12,106-115

GENOMICS

(1992)

Rapid Physical Mapping of Cloned DNA on Banded Mouse Chromosomes by Fluorescence in Situ Hybridization ANN L. BOYLE,* DAVID M. FELTQutTE,t NICHOLAS C. DRACOPOLI, t,$ DAVID

E. HOUSMAN,t AND DAVID C. WARD*,§

Departments of *Molecular Biophysics & Biochemistry and §Genetics, Yale University School of Medicine, 333 Cedar Street, New Haven, Connecticut 06510; and t Center for Cancer Research and $ Department of Biology, Massachusetts Institute of Technology, 77 Massachusetts Avenue, Cambridge, Massachusetts 02139 Received

June 25, 1991;revisedAugust19,1991

ping has relied most heavily on genetic analysis, including interspecific backcross data, as well as some physical mapping by hybridization to somatic cell hybrid panels. The available hybrid panels usually allow the identification of the chromosome but not further regional localization. In contrast to human chromosomes, which can be subdivided into seven groups, A through G, based on size and centromere position, mouse chromosomes are all telocentric, are similar in size, and are virtually impossible to identify as unbanded chromosomes. Thus, in situ hybridization has not been extended as readily to the mouse genome mapping effort because banding methods that are both compatible with fluorescence in situ hybridization techniques and of high enough resolution to allow for easy band identification have been unavailable. However, we recently demonstrated that the mouse long interspersed sequence -elements, Llmd, reside in Giemsa-positive bands, while the B2 sequences, a family of short interspersed repetitive sequences, are found in Giemsa-negative bands (Boyle et al., 1990). Furthermore, the banding pattern obtained by Llmd~hybridization is of high quality and is amenable to karyotype analysis. Here we report that the combination of hybridization banding and DAPI banding provides a relatively straightforward method for the karyotype analysis of mouse chromosomes and we document its utility by mapping 31 clones to cytogenetically banded chromosomes. We have also identified marker clones for each mouse chromosome that can be used for identification in the absence of banding.

Physical mapping of DNA clones by nonisotopic in situ hybridization has greatly facilitated the human genome mapping effort. Here we combine a variety of in situ hybridization techniques that make the physical mapping of DNA clones to mouse chromosomes much easier. Hybridization of probes containing the mouse long interspersed repetitive element to metaphase chromosomes produces a Giemsa-like banding pattern which can be used to identify individual Mus musculus,Mus spretus, and Mus castaneus chromosomes. The DNA binding fluorophore, DAPI, gives quinacrine-like bands that can complement the hybridization banding data. Simultaneous hybridization of a differentially labeled clone of interest with the banding probe allows the assignment of a mouse clone to a specific cytogenetic band. These methods were validated by first mapping four known genes, Cpa, Ly-2, Cck, and Igh-6, on banded chromosomes. Twenty-seven additional clones, including twenty anonymous cosmids, were then mapped in a similar fashion. Known marker clones and fractional length measurements can also provide information about chromosome assignment and clone order without the necessity of recognizing banding patterns. Clones hybridizing to each murine chromosome have been identified, thus providing a panel of marker probes to assist in chromosome identification.

0 1992

Academic

Press, Inc.

INTRODUCTION

Efforts to map the human genome have greatly intensified in recent years and have relied on both genetic and physical mapping strategies. Advances in nonisotopic in situ hybridization technology make this method an extremely rapid and direct means of assigning clones to specific human chromosomes and localizing the clone within a discrete chromosomal subregion (for examples, see Lichter et al., 1990; Lawrence et al., 1990; Cherif et al., 1990; Fan et al., 1990; Baldini and Ward, 1991). Mouse genome map0888-7543/92 $3.00 Copyright 0 1992 by Academic Press, All rights of reproduction in any form

MATERIALS

METHODS

Chromosome Preparations Metaphase chromosome spreads from female B6C3Fl/J, male Mus spretus, and male Mus castu106

Inc. reserved.

AND

PHYSICAL

MAPPING

neus spleen cultures were prepared as described previously (Boyle et aZ., 1990).

DNA

Probes

The Llmd hybridization banding probe, KSlSA, was provided by Thomas Fanning (Armed Forces Institute of Pathology) (Fanning, 1983) and contains a 1.3-kb EcoRI fragment from the middle of the mouse Ll sequence. Cosmid clone ~47 contains the entire tissue plasminogen activator (Plat) gene and was provided by Sidney Strickland (Stony Brook) (Rajput et aZ., 1987). Bam5X74, a plasmid with a 13-kb insert containing the cholecystokinin (Cck) gene (Friedman et al., 1989), and a plasmid containing 12 kb of the gene for carboxypeptidase A (CPU) were obtained from Jeffrey Friedman (Rockefeller University). Xwt#38, a X clone containing 15 kb from the downstream region on chromosome 10 (Shawlot et al., 1989), was obtained from Paul Overbeek (Baylor College of Medicine). Nancy Maizels (Yale University) provided the plasmid pVECp (Ballard and Bothwell, 1986), containing 10.5 kb of the constant region coding sequence from the immunoglobulin heavy chain (Igh-6) gene. The cmyc plasmid containing the Myc gene was provided by John Sedivy (Yale University). Cosmid pWeLyt-2, from Glen Evans (Salk Institute), contains the lymphocyte antigen 2 gene (Ly-2). Carol Readhead (California Institute of Technology) provided the cosmid clone ~0~138, containing the myelin basic protein gene (Readhead et al., 1987). A14V,V,Cy, containing approximately 18 kb of the T-cell y gene, was the gift of Adrian Hayday (Yale University) (Hayday et al., 1985). 68-36, a 4.2-kb repeat sequence on X, was from Christine Disteche (University of Washington) (Disteche et al., 1985), and the plasmid pY353, containing a Y-specific repeat sequence, was provided by C. Bishop (Bishop and Hatat, 1987). The anonymous H series mouse cosmid clones were derived from the human-mouse hybrid cell line A9/ 1492 (Dracopoli et al., 1988). The genomic DNA was prepared by partial digestion with Mb01 and size-selected on a sucrose gradient, and fractions containing 35-45 kb genomic fragments were ligated into pWel5 (Wahl et al., 1987), which had been prepared by digestion with BamHI and dephosphorylated with calf intestinal phosphatase (CIP) (Sambrook et uZ., 1989). Only 2% of the cosmid clones hybridized to radiolabeled human Cot-l DNA. The cosmids used in this study did not hybridize to human Cot-l DNA and were given the designations Hl-H50. Miniprep DNA for use in the nick-translation reactions was prepared using a standard alkaline lysis protocol (Sambrook et al., 1989).

OF

MOUSE

CLONES

107

Probe Labeling

The cosmid clones H6, H8, Hll, H13, H14, H16, H17, H18, H22, H28, H30, H33, H37, H38, H40, H41, H43, H44, H46, H49, as well as the known clones listed above, were labeled by nick translation with biotin-ll-dUTP (Brigati et al., 1983). The clones KSl3A and ~47 were labeled by nick translation using a mixture of digoxigenin-11-dUTP and dTTP in a ratio of 1:3 (Lichter et al., 1990). To ensure a size range of lo@-500 nucleotides, an aliquot of each nicktranslated probe was heat-denatured and run on a 1.2% agarose gel (nondenaturing). Unincorporated nucleotides were removed using a Sephadex G-50 medium spin column equilibrated with 10 n&f Tris . HCl/l mM EDTA/O.l% SDS, pH 8.0. In Situ Hybridization

Biotinylated cosmid clone (100 ng), mouse genomic competitor DNA (2 pg), and salmon sperm DNA (7 pg) were ethanol-precipitated together and redissolved in 2.5 ~1 of deionized formamide, followed by 2.5 ~1 of hybridization mix (4x SSC, 20% dextran sulfate). Digoxigenin-labeledKS13A (300 ng) wasprecipitated in a separate tube and redissolved as above. After 5 min of denaturation at 75”C, the KS13A was placed on ice and the cosmid clone was placed at 37°C for 10 min. The contents of the two tubes were mixed and immediately applied to denatured metaphase chromosome spreads. Slides were denatured in 70% deionized formamide, 2~ SSC at 70°C for 2 min and then dehydrated through cold 70,90, and 100% ethanol, 5 min each. The 10 ~1 of probe mix was covered with an 18 X 18-mm coverslip, sealed with rubber cement, and incubated in a moist chamber overnight at 37°C. Posthybridization washes and blocking step were performed as described by Lichter and colleagues (1988). The banding probe was detected using antidigoxigenin Fab fragments conjugated with fluorescein (Boehringer Mannheim) diluted to 1.3 .hg/ml in 4X SSC, 1% BSA, 0.1% Tween 20 (dilution buffer). The biotinylated probes were detected using either Texas Red Avidin D (12.5 pg/ml) (Vector Laboratories) or rhodamine ExtrAvidin (Sigma) diIuted in dilution buffer according to manufacturer’s instructions. The chromosomes were counterstained with 4,6-diamidino-2-phenylindole (DAPI) (200 rig/ml), which was added directly to the antifade mounting solution (Johnson et al, 1982). When multiple biotinylated clones were hybridized to unbanded chromosomes, the probes were precipitated together along with the salmon sperm and mouse genomic DNA, resuspended in 5 ~1 of formamide and 5 yl of hybridization mixture, denatured for 5 min at 75”C, and partly reannealed for 10 min at 37°C. In some cases, 6 pg of mouse Cot-l DNA (gift

108

BOYLE ET AL.

from Paul Nisson, Life Technologies, Catalog No. 8440SA) was substituted for the genomic competitor DNA. If the probes were biotinylated, they were detected with FITC-conjugated avidinDCS (5 fig/ml) (Vector Laboratories) and the chromosomes were counterstained with propidium iodide (200 ngjml), which was added to the antifade mounting solution. If both digoxigeninand biotin-labeled probes were used, the biotin was detected with rhodamine avidin and the digoxigenin with antidigoxigenin Fab fragments conjugated with fluorescein as described above. DAPI was used as the counterstain in place of propidium iodide. Imaging

Fluorescent signals were observed using a Zeiss Axioskop (63X 1.2 numerical aperture Plan Neofluar oil immersion objective) and imaged using a cooled CCD camera (Photometric CH220). The microscope was equipped with precision bandpass filters to minimize the image displacement to less than 1 pixel when filter cubes were switched. All image acquisition and processing were done on a Macintosh IIci computer. Each fluorophore (FITC banding, rhodamine signal, and DAPI counterstain) was imaged sequentially and the B-bit source images were thresholded using the Enhance image processing program (MicroFrontiers, Des Moines IA). The images were merged and 24-bit pseudocolored using custom software developed in this laboratory (Tim Rand, unpublished). The DAPI counterstain was used in two different ways-either as a general DNA stain to outline the length of the chromosomes when Ll hybridization banding was used or as an alternative banding method in the absence of Ll banding. In cases where all three images were merged together, both the Ll banding image (FITC) and the DAPI image were pseudocolored green and the cosmid signal image (rhodamine or Texas Red) was pseudocolored red. It was necessary to reduce the intensity of the DAPI image to 40% of its original value so that it would not compete with the Ll banding image, but would provide a light staining in those regions of the chromosomes deficient in Ll sequences. DAPI preferentially stains AT-rich regions and so can produce a Giemsa-like banding pattern. In the absence of Ll banding, this DAPI banding was improved using the image processing software Enhance. The sharpened image was pseudocolored blue and displayed at full intensity. In cases where multiple biotinylated probes were hybridized to unbanded chromosomes, the FITC signals were pseudocolored green and the propidium iodidecounterstained chromosomes were displayed red. When both digoxigeninand biotin-labeled probes were hybridized to unbanded chromosomes, the FITC

and rhodamine signals were pseudocolored green and red, respectively, and the DAPI counterstain was pseudocolored blue. Individual chromosomes from the pseudocolored, merged images were cut and pasted to prepare karyotypes and composite images using the program Pixel Paint Professional. The final images were photographed directly from the computer screen. Fractional length measurements were made from merged images on the computer screen using the software Segmented Ruler (Tim Rand, unpublished). Using the mouse, the ratio of the distance between the centromeric end of the chromosome and the signal to the total length of the chromosome was determined. RESULTS To establish the validity of the double hybridization method for gene mapping to banded mouse chromosomes, we mapped four genes previously localized by a variety of other methods. Band assignments for all imaged chromosomes were made using the standard mouse idiogram by Evans (1989) and reproduced in Fig. IA with permission. Two representative chromosomes from each are presented in Fig. 2A. The cosmid clone pWeLyt-2, containing the Ly-2 gene, was mapped to chromosome 6, band C. Genetic analysis established the location of Ly-2 with respect to a number of genes and loci in the mid region of chromosome 6 (Rueff-Juy et al., 1988) but no cytogenetic mapping data have been published. The carboxypeptidase A gene was localized to chromosome 6, band Bl. This is in agreement with published data that Cpa is -5 CM proximal of Tcrb, which was mapped by isotopic in situ hybridization to 6B (Bucan et al., 1986). Both genetic analysis and somatic cell hybrid panels were used to place Cck to distal 9 (Friedman et al., 1989). Again, our assignment to chromosome 9, band F4, correlates with existing data. pVECr, containing the Igh-6 gene, was mapped to chromosome 12, band Fl2. Although both intra- and interspecific backcrosses have been used to order Igh-6 with respect to many other probes on chromosome 12 (Seldin et al., 1989), the only physical mapping data indicate that Igh-6 is located on distal 12 (D’Eustachio et al., 1980). Each of the 20 anonymous cosmid clones was mapped with respect to Ll hybridization bands. When the slides were scored for imageable chromosomes, preference was given to longer prometaphase chromosomes, which would display a higher resolution banding, and to chromosomes with two chromatid signals that were perpendicular to the long axis of the chromosome. Chromosomes that showed only a single chromatid signal or two chromatid signals in different focal planes were not imaged because it is not possible to determine whether such chromosomes

PHYSICAL

MAPPING

OF

MOUSE

CLONES

FIG. 1. (A) Idiogram of the banding patterns in mouse chromosomes used for assignment of probes to cytogenetic prepared by Evans (14) and is reproduced by permission of the Oxford University Press. (B) XX mouse karyotype stained chromosome on the right and the same chromosome banded with Ll on the left.

bands. The figure was comparing the DAPI-

110

PHYSICAL

MAPPING

are twisted. As in conventional cytogenetic analysis, the quality of the banding varies on different areas of the slide, and high-quality metaphase preparations greatly reduce the time spent analyzing the results. In general, the probe was hybridized to an 18 X 18-mm area of the slide and less than two-thirds of this area was needed to find a sufficient number of good-quality spreads. Lichter and colleagues (1990) reported that four cosmid signals per metaphase spread normally were seen on approximately 80-90% of human spreads. We have found that when we use mouse genomic clones of 10 kb or greater, approximately 80-90% of the spreads have three or more signals. When very short chromosomes are hybridized, approximately 80-90% of the spreads have all four signals. This difference is likely due to the fact that the longer chromosomes, which are preferred for high-resolution mapping, have a greater tendency to be twisted, which can obscure the signal. An example of a merged, pseudocolored image of a complete metaphase spread is shown in Fig. 2B. In each case, entire metaphase spreads were imaged and then the chromosomes of interest were cut and pasted into a composite using the software program Pixel Paint Professional. Very often it is helpful to compare the labeled chromosome with the banding pattern of the other chromosomes in the spread to confirm the chromosome identity. At least eight representative chromosomes were imaged for each clone. Three representative chromosome images after hybridization of each clone are shown in Fig. 2C, and the chromosome and band assignment of each clone is given in Table 1. Figure 2C is a composite prepared from approximately 60 merged images. Variations in probe intensity are due to the differing genetic complexities of the clones and to variability in the chromosome preparations. Labeling efficiency can also be affected by the purity of the DNA used in the nick-translation reaction. We have previously shown that the Ll repetitive element probe does not hybridize to the centromeric

OF

MOUSE

111

CLONES

TABLE Mapping Clone H18 H46 H49 H37 H43 H28 H16 H44 H30 H33 H13 H17 H6 H8 Hll H14 H38 H41 H22 H40

1

Murine DNA Clones to Cytogenetically Banded Metaphase Chromosomes Chromosome 1 1 1 1 2 2 2 3 4 4 5 5 5 7 a 9 11 14 16 17

Band

assignment lE3-4 lH2-4 lH3-5 lA3-5 2H2-4 2B 2H2-4 3B 4D (mid) 4c3 5G2 5A2-3 5E5-F 7F3 8El 9E3 11B (mid) 14A3 16B2-4 17El

heterochromatin (Boyle et al., 1990). Furthermore, many mouse chromosomes have G-negative bands at their distal ends, which are deficient in Ll sequences. However, general DNA counterstains, such as propidium iodide or DAPI, stain the entire length of every chromosome andstain the centromeric heterochromatin of murine chromosomes very brightly. Thus, the informational content of the hybridization banding is maximized if both the banding and the counterstain signals are recorded. When such digital images are merged together, for example, as in Fig. 2B, the intensity of the counterstain image must usually be reduced so as not to compete with the banding image (see Materials and Methods). DAPI, which preferentially binds to A-T-rich re-

FIG. 2. (A) In situ hybridization mapping of four known genes to mouse chromosomes. The lymphocyte antigen 2 gene maps to chromosome 6, band C, the carboxypeptidase A gene to 6B1, the cholecystokinin gene to 9F4, and the immunoglobulin heavy-chain constant region gene (marked “CZ.?‘) to 12Fl-2. Unless otherwise noted, the genes were labeled with biotin, detected by rhodamine or Texas Bed conjugated to avidin, and displayed as red on the computer screen. The Ll banding probe was labeled with digoxigenin, detected with antidigoxigenin Fab fragments conjugated with FITC, and displayed as green. DAPI counterstain was also displayed green (see text). (B) A full metaphaee spread banded with Ll probe and cohybridized with H43. Insert compares the same pair of chromosome 2 homologa-the merged Ll/DAPI image and the DAPI image alone. (C) Physical mapping of 20 anonymous cosmid clones. Three of the 8-12 chromosomes imaged for each one are shown. Detection and colors are as described for A. (D) Chromosome 11 pair hybridized with clone H38. Left pair is banded by Ll hybridization. The image on the right is the same chromosome pair stained with DAPI. (E) The biotinylated anonymous clone Hll was hybridized to metaphase chromosomes along with two marker clones, c47 and pVECa. The cosmid clone ~47, containing the Plut gene, is known to map to proximal chromosome 8 and was labeled with digoxigenin. pVECp was shown in A to map to chromosome 12, band Fl-2, and was labeled with biotin. The biotinylated probes were detected with FITC-avidin and pseudocolored green. The digoxigenin-labeled c47 was detected with rhodamine conjugated to antidigoxigenin Fab fragments and was displayed as red. The chromosomes were counterstained with DAPI (blue). Using this approach, Hll was confirmed to map to distal chromosome 8. (F) Biotinylated H18, H46, and H49 hybridized to propidium iodide-stained chromosomes and detected with FITC-avidin. Mouse Cot-l DNA was used in place of genomic DNA to suppress signal from interspersed repetitive sequences in the three clones.

112

BOYLE

gions, can also give a banding pattern that resembles quinacrine or G bands. In some cases, this DAPI banding can supplement the information provided by Ll hybridization banding. For example, in the insert of Fig. 2B, the end of chromosome 2 is difficult to determine on the Ll merged image, but is quite clear on the DAPI image. In the case of chromosome 11, which has very few distal Ll sequences, DAPI actually allows easier assignment of the clone to a band. Figure 2D compares the DAPI and Ll images of a chromosome pair cohybridized with clone H38. Bands C and El are more evident in the DAPI image and it is possible to see hints of minor low-intensity subbands in B. Figure 1B compares an Ll and DAPI karyotype. The banding patterns produced by the two very different methods are similar; however, the intensities of individual bands do vary. Note as well the intense staining of the centromere regions by DAPI that are not detected by the Ll probe. In general, Ll hybridization provides bands that are sharper and are of higher contrast. It is usually possible to obtain readable banding from Ll hybridization even if the chromosome preparation is not excellent. This is not true for DAPI banding, which is generally better on longer, more decondensed chromosomes. Highly condensed chromosomes show little evidence of DAPI banding. It should be noted that the metaphase spread in Fig. 1B was selected because it represents very good DAPI banding; however, its Ll banding is merely average. DAPI banding varies tremendously with the quality of the chromosome preparation used. Although the parameters for DAPI banding have not been fully optimized, chromosomes that were dropped onto slides in a humid environment (relative humidity 40-50%) and then allowed to age in a desiccator at room temperature for at least 1 month routinely give good banding in our hands. The quality of the deionized formamide used for denaturing the chromosomes for in situ hybridization is critical because inferior grades of formamide cause chromosomes to overdenature, resulting in a loss of chromatin and in fuzzy bands. High-quality formamide usually has an initial conductivity of 200-300 pmho and can be readily deionized to 10 pmho or less using a mixed bead resin. It is always advisable to denature some test slides when testing new formamide batches. The chromosomes should maintain their morphology after denaturation and there should be no significant decrease in the number of metaphase spreads on the slide. In cases where it was difficult to positively identify the labeled chromosome on the basis of the Ll banding pattern alone, a second experiment hybridizing the unknown clone and marker clones for the suspected chromosomes on unbanded chromosomes was performed. Ll banding indicated that Hll mapped to either chromosome 8 or 12, two chromosomes that

ET

AL. Fractional Length 1

A

0

23 5

C

1.;

1.3 3

4

D E

5

-1.13 2.1 2.3 2’2 3 4

T E -?-a 2.2 ;:A

0.6 ] Hi8

0.8 ] H46

H

34 5 ------s

7 H49

1.0

FIG. 3. Idiogram of mouse chromosome 1 [Ref. (14)], with the fractional length scale indicating the positions of the clones H18, H46, and H49.

can be confused, especially on lower quality chromosome preparations. Two marker clones were used in conjunction with the unknown: digoxigenin-labeled ~47, a cosmid clone for the tissue plasminogen activator gene shown previously to map to the proximal end of chromosome 8, and biotin-labeled pVECp (Igh-6), which was shown above to map to chromosome 12Fl2. Using this approach, Hll was mapped to the distal end of chromosome 8 without relying on any banding methods (Fig. 2E). Since Plut and Igh-6 map to opposite ends of their respective chromosomes, the same result could be obtained as easily using a single-color labeling/detection scheme. Similarly, a clone for Cck was used (Friedman et al., 1989) to confirm the localization of H14 to chromosome 9 (data not shown), which on the basis of banding can often be confused with chromosome 13 (see Fig. 1B). In Fig. 2F three of the clones that mapped to chromosome 1, H18, H46, and H49, are cohybridized to an unbanded metaphase chromosome spread. In this case, the location of each clone had been determined previously by hybridizing each individually. The figure clearly shows the physical separation between the clones. Thus, in addition to being used to assign individual clones to cytogenetic bands, in situ hybridization methods can be used to rapidly order clones along an unbanded chromosome, as has been described for human chromosomes (Lichter et al., 1990). The ranges of fractional lengths measured for each clone are displayed on the chromosome 1 idiogram in Fig. 3. The 4 known genes and the 20 anonymous cosmids hybridized to more than half of the murine chromosome complement. Since marker clones can be ex-

PHYSICAL

TABLE Additional Clone” Cpa plasmid Xwt#38 PVEQ A14V,V,Cy” cmyc plasmid cos138 H42’ 68-36 pY353

Marker

Chromosome 6 10 12 13 15 18 19 X Y

MAPPING

2 Clones Band

assignment

Bl B3 Fl-F2 A3 D2-D3 E3-E4 C/D border A3d Entire chromosome

D Clones are described under Materials and Methods. b This clone also gives a weaker signal on chromosome 12, band B, on some of the metaphase spreads examined. ’ H42 is an additional anonymous cosmid clone not included in the group of 20. d Disteche et al. (11).

tremely useful in identifying chromosomes when banding is difficult to interpret, we sought clones to hybridize to the unrepresented chromosomes. These clones are listed in Table 2. The chromosome location of each was confirmed by Ll or DAPI banding and an approximate band assignment determined. In addition, each clone was checked to ensure that it gave a bright reproducible signal evident on the majority of nuclei and spreads examined. Together these clones provide a set of reference clones allowing rapid identification of mouse chromosomes without banding. These clones will be made available to other interested parties upon written request. All of the clones described here can be seen by eye using a conventional epifluorescence microscope. If enhanced signals are required for purposes of photomicroscopy, signal amplification methods can be employed (Pinkel et al., 1986). Because interspecific backcross analysis is an important tool for mouse genetics (reviewed by Avner et al., 1988), the extension of physical mapping on M. spretus and/or M. castaneus can be anticipated. For this reason, the M. musculus Ll probe, KS13A, was hybridized to M. spretus and M. castaneus chromosome spreads. As can be seen in Fig. 4, hybridization of KS13A to M. spretus chromosomes results in a Giemsa-like banding pattern which is similar to the M. musculus pattern. M. castaneus gives similar results (data not shown). Thus, all of the in situ hybridization mapping techniques, including the banding protocols, can be extended to map clones to M. spretus and M. castaneus in addition to M. musculus. DISCUSSION

We have illustrated that nonisotopic in situ hybridization is a useful physical mapping technique suit-

OF

MOUSE

CLONES

113

able for the analysis of the mouse genome. Although in situ hybridization is the only method for assigning a clone to a cytogenetic band, the method can also easily provide chromosome identification and clone ordering information without the need to learn the intricacies of mouse chromosome banding patterns. The multipronged approach to physical mapping of mouse genes outlined above includes Ll hybridization banding, DAPI staining, the use of marker clones, and fractional length ordering of clones. Ll hybridization produces sharp, high-contrast bands that facilitate the identification of individual mouse chromosomes. It is totally compatible with the hybridization of any clone of interest and does not require any additional time. However, maximum information is obtained when the Ll image is merged with a general counterstain which allows for easier orientation of the chromosomes. In most cases when high-quality chromosome preparations are used, the DAPI counterstain alone may produce a banding pattern sufficient for chromosome identification. DAPI banding offers several advantages over quinacrine banding, the fluorescent banding method applied most frequently. DAPI, unlike quinacrine, does not photobleach significantly; the banding can be observed at length under the microscope and it is stable when the slides are stored in the dark at 4°C. Also, DAPI, with an emission maximum of 455 nm, does not overlap with FITC’s emission spectrum (max = 515 nm), whereas quinacrine (emission maximum 503 nm) does, thus making it more difficult to resolve the FITC signal and the quinacrine counterstain. Although these methods alleviate some of the difficulties associated with mouse chromosome karyotyping, knowledge of the banding patterns is still required. Identification of some mouse chromosomes can be quite troublesome and can be exacerbated by inadequate chromosome preparations. For those cases where a cytogenetic band assignment is not needed, cohybridization of the clone of interest with marker probes can provide identification of the chromosome without the need for karyotype familiarity or highquality chromosome preparations. Fractional length measurements provide a rapid method for ordering clones along a chromosome. We have compared the order of some random genomic probes on chromosome X obtained by fractional length measurements with the published recombinational distances from interspecific mouse crosses (Amar et al., 1985, and our unpublished data) and have found them to be in agreement. Although the fractional length ranges for three of the unknowns are indicated along a chromosome 1 idiogram, and these particular values are in agreement with our cytogenetic data, one cannot make a definite band assignment by this method because idiograms are not normalized relative to the fractional length of their chromosomes.

114

BOYLE

ET AL.

karyotype of Mus spretus (XY) from a single metaphase spread. Biotinylated KS13A was hybridized to male M. FIG. 4. Ll hybridization spretuc i chromosome preparations and detected with FITC-avidin. The FITC image and the DAPI counterstain image were merged , and the chrome osomes

arranged

in a karyotype

as described

under

Materials

To date, in contrast to the number of human probes, there have not been very many mouse probes mapped to cytogenetic bands. Disteche et al. (1989) compared the recombination map for clones of the mouse X chromosome with a physical map generated by isotopic in situ hybridization combined with quinacrine banding. They noted that there was good agreement on the order of the clones. Nonisotopic methods are considerably faster, show greater spatial resolution, and are as sensitive as isotopic methods. With the improved methods for identifying mouse chromosomes described here, it should be possible to rapidly map DNA clones to cytogenetic bands and to extend the comparisons of the recombinational and cytogenetic maps. For example, by mapping Cpu to 6Bl and Q-2 to mid 6C, we can predict that all of the probes that have been genetically mapped to between 15 and 32 CM (the recombination values for Cpa and Ly-2, respectively) on chromosome 6 will physically map to the 6B-6C interval. Such studies should indi-

and Methods.

cate regions of recombinational hot spots and suppression that prevent a direct correlation of genetic and physical data. Furthermore, clones that are found by physical methods to lie quite close to the distal end of the chromosome (e.g., H46, H49, H43, H13, HS, and Hll) can be used to help define the ends of the genetic map. The tools are now in place to begin to develop the cytogenetic map of the mouse and to correlate this information with the existing genetic maps.

ACKNOWLEDGMENTS We thank the many people listed under Materials and Methods who provided clones, Martin Ferguson for assistance with the computer software, and Joan Menninger for helpful discussions on chromosome banding. This work was supported by grants from the National Institutes of Health to D.C.W. (HG-00246, HG-00272, and GM-40115), D.E.H. (HG00299), and N.C.D. (CA44176 and HG00198).

PHYSICAL

MAPPING

REFERENCES 1. AMAR, L. C., ARNAUD, D., CANBFUXJ,J., GUENJW J. L., AND AVNER, P. R. (1985). Mapping of the mouse X chromosome using random genomic probes and an interspecific mouse cross. EMBO J. 4: 3695-3700. 2. ABNER, P., AMAR, L., DANDOLO, L., AND GUENET, J. L. (1988). Genetic analysis of the mouse using interspecific crosses. Trenda Genet. 4: 18-23. 3. BALDINI, A., AND WARD, D. C. (1991). In situ hybridization banding of human chromosomes with Alu-PCR products: A simultaneous karyotype for gene mapping studies. Genomics 9: 770-774. A. (1986). Mutational anal4. BALLARD, D. W., AND B-u, ysis of the immunoglobulin heavy chain promoter region. Proc. Natl. Acad. Sci. USA 83: 9626-9630. 5. BISHOP, C. E., AND HATAT, D. (1987). Molecular cloning and sequence analysis of a mouse Y chromosome RNA transcript expressed in the testis. Nucleic Acids Res. 16: 2959-2969. 6. BOYLE, A. L., BALLARD, S. G., AND WAFQ D. C. (1999). Differential distribution of long and short interspersed elements in the mouse genome: Chromosome karyotyping by fluorescence in situ hybridization. Proc. Natl. Acad. Sci. USA 87: 7751-7761. I. BRIGAT~, D. J., ML~YERSON, D., LEARY, J. J., SPALHOLZ, B., TRAVIS, S. Z., FONG, C. K. Y., HSIUNG, G. D., AND WARD, D. C. (1983). Detection of viral genomes in cultured cells and paraffin-embedded tissue sections using biotin-labeled hybridization probes. Virology 128: 32-50. T., COLBERG-POLEY, A. M., WOL8. BUCAN, M., YANG-FENG, GEMUTH,

GUENET,

J. L., FRANCKE,

U., AND LEHRACH,

H.

(1986). Genetic and cytogenetic localisation of the homeobox containing genes on mouse chromosome 6 and human chromosome 7. EMBO J. 6: 2899-2905. 9. CHERIF, D., JULIER, C., DELATTRE, O., Dmm.6, J., LATHR~P, G. M., AND BEROER, R. (1990). Simultaneous localization of cosmids and chromosome R-banding by fluorescence microecopy: Application to regional mapping of human chromosome 11. Proc. Natl. Acad. Sci USA 87: 6639-6643. 10. D’EUSTACHIO, P., PRA~TCHEVA, D., Mmcu, K., AND RUDDLE, F. H. (1980). Chromoeomai location of the structural gene cluster encoding immunoglobulin heavy chains. J. Exp. Med. 151: 1545-1560. 11. DISTECHE, C. M., TANTRAVAHI, U., GANDY, S., EISENHARD, M., ADLER, D., AND KUNKEL, L. M. (1985). Isolationandcharacterization of 2 repetitive DNA fragments located near the centromere of the mouse X chromosome. Cytogenet. Cell Genet. 39: 262-268. 12. DISTECHE, C. M., MCCONNELL, G. K., GRANT, S. G., STEPHENSON, D. A., CHAPMAN, V. M., CANDY, S., AND ADLER, D. A. (1989). Comparison of the physical and recombinational maps of the mouse X chromosome. Genomics 5: 177184. 13. DRACOPOLI, N. C., STANGER, B. Z., Im, C. Y., C&L, K. M., LINCOLN, 8. E., LANDER, E. S., AND HOUSMAN, D. E. (1988). A genetic linkage map of 27 loci from PND to FY on the short arm of human chromosome 1. Am. J. Hum. Gerzet. 43: 462-

OF MOUSE

15. FAN, Y., DAVIS, L. M., AND SHOWS, T. B. (1990). Mapping small DNA sequences by Buorescence in situ hybridization directly on banded metaphase chromosomes. Proc. Natl. Acad. Sci. USA 87: 6223-6227. 16. FANNING, T. (1983). Size and structure of the highly repetitive BamHI element in mice. Nucleic Acids Res. 11: 50735091. 17. FRIEDMAN, J. M., SCHNEIDER, B. S., BARTON, D. E., AND FRANC=, U. (1989). Level of expression and chromosome mapping of the mouse cholecystokmin gene: Implications for murine models of genetic obesity. Genomics 5: 463-469. 18. HAYDAY, A. C., SAITO, H., GILLIES, S. D., KRANZ, D. M., TANIGAWA, G., EISEN, H. N., AND TONEGAWA, S. (1985). Structure, organization, and somatic rearrangement of T cell gamma genes. Cell 40: 259-269. 19. JOHNSON, G. D., DAVIDSON, R. S., McNm K. C., RusSELL, G., GOODWIN, D., AND HO-ROW, E. J. (1982). Fading of immunofluorescence during microscopy: A study of the phenomenon and its remedy. J. Zmmurwl. Methods 55: 231-242. 20. LAWRENCE, J., SINGER, R. H., AND MCNEXL, J. A- (1990). Interphase and metaphase resolution of difIerent distances within the human dystropbin gene. Science 249: 928-932. 21. LIGHTER, P., TANG, C. C., CALL, K., HERMAN~ON. G., EVANS, G. A., HOUSMAN, D., AND WARD, D. C. (1999). High resolution mapping of human chromosome 11 by in situ hybridixation with cosmid clones. Science 247: 6469. 22. LIGHTER, P., C!REME& T., BORDEN, J., -IS, L., AND WARD, D. C. (1988). Delineation of individual chromosomes in metaphase and interphase cells by in situ suppression hybridization using recombinant DNA libraries. Hum. Genet. 80: 224-234. 23. PINKEL, D., STRAUME, T., AND GFZAY, J. W. (1988). Cytogenetic analysis using quantitative, high sensitivity, Buorescence in situ hybridixation. Proc. Natl. Acad. Sei. USA 83: 2934-2938.

24. RAJPUT, B., MARSHALL, A., K-Y, NAYU)R,

E. P. (198B) Standard ideogram. In “Genetic Variant8andStrainsof the LaboratoryMouse” (M. F. Lyons and A. G.Srarle,Ed~.),2nded.,pp. 578477, OxfordUniv. Preen,

Oxford.

S. L., BELIN,

D., RICKLE~,

A. M., w, P. A., R. J., AND STRICKLAND,

S. (1987). ChromosomaI assignment of genes for tissue pIasminogen activator and urokinase in mouse. Somatic Cell Mol. Genet. 13: 581-588. 25. READHEAD, VEDRA,

26. 27. 28.

29,

470.

14. EVANS,

115

CLONES

30.

C., POPKO, R. A., SIDMAN,

B., Tw I, N., SHINE, H. D., SAAR. L., AND HOOD, L. (1987). Expres-

sion of a myelin basic protein gene in transgenic shiverer mice: Correction of the dysmyelinating phenotype. CeU 48: 703-712. RUEFF-JUY, D., DRAPDZR, A., AND CAZENAVE, P. (1988). Mapping of Igk-V genes using backcrossed laboratory and wild mice. Zmmunogenetica 28: 233-239. SAMEWOK, J., FRITWH, E. F., AND MANIATUB, T. (1989). “Molecular Cloning: A Laboratory Manual,” Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY. SELDIN, M. F., HOWARD, T. A., AND D’EUSTACHIO, P. (1989). Comparison of linkage maps of mouse chromosome 12 derived from laboratory strain intraspecific and iifus spretus interspecific backcrosses. Genomics Pe 24-28. SHAWLOT,W., SICILIANO, M. J., STALLINGS, R. L., AND OVERBEEK, P. A. (1989). Insertional inactivation of the downless gene in a family of transgenic mice. Mol. Biol. Med. 8: 299-307. WAHL, G.M., LEWIS, K., Rum, J. C., RQTHENRURG, B., ZHAO, J., AND EVANS, G. (1987). Cosmid vectors for rapid genomic

walking, restriction mapping, and gene transfer. Pnx. N&L Acad. Sci. USA Sk 2169-2164.