sign in sign up
<<
Home , Genomic library

Proc. Natl. Acad. Sci. USA Vol. 86, pp. 5898-5902, August 1989 Genetics Rapid screening of a human genomic in artificial for single-copy sequences (colony hybridization/X- linked disorders/doning/shuttle vector) CHRISTOPHER N. TRAVER*, SUE KLAPHOLZ, RICHARD W. HYMANt, AND RONALD W. DAVIS Department of Biochemistry, Stanford University School of Medicine, Stanford, CA 94305 Contributed by Ronald W. Davis, May 8, 1989

ABSTRACT A yeast artificial chromosome (YAC) library the vector includes a drug-resistance marker selectable in in Saccharomyces cerevisiae consisting of30,000 clones with an mammalian cells. This permits shuttling of YAC DNA from average insert size of 0.1 megabase pair of human DNA has yeast to mammalian cells, a key step for studies of genetic been generated from primary fibroblast DNA. A YAC vector complementation and gene expression with cloned genomic was modified to enable the recovery of both ends of a human DNA. DNA insert in in Escherichia coi and to confer G418 resistance to mammalian cells. A rapid method for yeast colony MATERIALS AND METHODS hybridization was used that exploits the ability of yeast sphero- plasts to regenerate in a thin layer of calcium aginate. This Strains, DNA, and Lines. Saccharomyces cerevisiae method permits direct replica plating and processing of colo- strain AB1380 (MATa, ura3, trpl, ade2-1, can1-100, lys2-1, nies from the primary transformation plate to nitrocellulose hisS) and the vector pYAC4 were kindly provided by David filters. Yeast colony hybridization conditions have been estab- Burke (1). The following cloned probes from the Xq28 region lished to identify, within a YAC library of human genomic were used for hybridization: St14-1 (DXS52) (5), DX13 DNA, artificial chromosomes with homology to human DNA (DXS15) (6), MN12 (DXS33) (7), hs7 (8), and p625.8, located probes of unique single-copy sequence. An artificial chromo- 30 kilobase pairs (kbp) from the 3' end of the human factor some with a 0.1-megabase-pair insert from the human Xq28 VIII gene (J. Gitschier, personal communication; ref. 9). The region has been identified by hybridization to a DNA probe that MN12 probe is a 2.8-kbp EcoRI/Msp I restriction fragment detects a unique sequence near the 3' end of the factor VIII and the p625.8 probe is a 3.15-kbp EcoRI restriction frag- gene. ment; both are free of highly repetitive DNA sequences. The cloned probe pNN415 is a repetitive-DNA-free 3.5-kbp Sal I The ability to clone and manipulate large pieces of DNA is restriction fragment derived from the ampicillin-resistant essential for many aspects of mammalian . (AmpR) end of the hX1 YAC (see Results). Mouse 3T3 and Key applications of this technology include (i) physical primary human fibroblast MRC-5 cells were obtained from the American Type Culture Collection and were grown isolation ofgenes involved in genetic disease from genetically according to their recommendations. linked restriction fragment length polymorphisms, (ii) anal- Media and Storage of Yeast Transformants. All media were ysis of the large transcription units of some mammalian prepared as described (10). For long-term storage of the genes, and (iii) physical mapping of ordered fragments from library, a synthetic complete medium lacking uracil and large . The introduction of the yeast artificial chro- tryptophan (SCUT) replica plate of each primary transfor- mosome (YAC) as a vehicle for large pieces of DNA mation plate was further replica-plated after 1-2 days at 30'C has opened these areas for investigation (1). YACs have been to a nitrocellulose disc (Schleicher & Schuell, BA85) on a used to physically map the relatively small of Cae- yeast extract/peptone/dextrose (YPD) plate, which was then norhabditis elegans (2) and to clone centromeric sequences incubated for 12 hr at 30°C. Discs were then placed colony- from Schizosaccharomyces pombe (3). In addition, YACs side up on Whatman 3MM paper saturated with YPD/ 30o with human DNA inserts have been screened for moderately (vol/vol) glycerol. A fresh nitrocellulose disc, saturated as repetitive DNA elements (4). Several obstacles need to be described above, was overlaid onto the first disc, forming a overcome to apply YAC technology to the comprehensive sandwich. The two discs were placed between dry Whatman analysis ofmammalian genomes. In particular, large numbers 3MM paper and pressure was applied briefly with a steel of colonies must be screened to identify any unique single- replica block. Discs were wrapped in foil, placed in plastic copy sequence (1). YACs are usually present at one copy per storage bags, and stored at -70°C until use. Colonies were yeast cell. To find a single-copy human sequence in a YAC recovered by separating the two nitrocellulose discs, placing library with an average insert size of0.1 megabase pair (Mbp) each colony-side up on a YPD plate, and incubating at 30°C requires detection, on average, of 1 colony of 30,000. This 1-2 days, at which point one disc was refrozen as described paper presents a rapid method for screening a human ge- further nomic YAC library for single-copy human sequences, elim- above, and the other was used for experiments. inating the need to pick each colony from within top agar to arrays. Two YACs containing human DNA inserts with RESULTS homology to probes from the Xq28 region have been identi- Modifications to the YAC Vector. A YAC can be con- fied by this technique. A modified YAC has structed from a large DNA restriction fragment by ligating been constructed. This vector was designed to facilitate the the the elements to walking within a library by allowing recovery ofboth ends of onto ends of fragment necessary a given insert within plasmids in . In addition, Abbreviations: YAC, yeast artificial chromosome; FIGE, field- inversion . The publication costs of this article were defrayed in part by page charge *To whom reprint requests should be addressed. payment. This article must therefore be hereby marked "advertisement" tPresent address: Cancer and Developmental Biology, Syntex Re- in accordance with 18 U.S.C. §1734 solely to indicate this fact. search, 3401 Hillview Avenue, Palo Alto, CA 94304. 5898 Downloaded by guest on September 29, 2021 Genetics: Traver et al. Proc. Natl. Acad. Sci. USA 86 (1989) 5899 stabilize a linear yeast chromosome: a , telo- insert of the artificial chromosome, DNA probes from each meres, an , and a means of selection (11). vector arm ofthe YAC hybridized to the same-sized 400- and The YAC vector, pYAC4 (1), as originally described, is 450-kbp, respectively, Xho I restriction fragments (data not sufficient for cloning large DNA fragments in S. cerevisiae. shown). Hence, DNA has been shuttled from a single-copy For genetic complementation studies with human genomic YAC in yeast to mammalian cells by spheroplast fusion and DNA, a vector that enables shuttling of these large DNA G418 selection. An unintegrated artificial chromosome in fragments from yeast to mammalian cells is necessary. In mammalian cells was not detected, but the particular artificial addition, a vector that provides a means to isolate each end chromosome used in these experiments may have been ofan insert for bidirectional walking from a given clone would defective for mammalian replication, segregation, or telo- be useful. Both of these goals were accomplished by modi- meric function. The presence of a mammalian drug- fying the pYAC4 vector with the addition of the neomycin- resistance marker on a YAC will allow for testing of various resistance (Neo) gene and the ColEl origin of replication cloned sequences for these functions. from pSV2neo (12) (Fig. 1). The Neo gene can confer Library Construction. DNA for the human genomic library kanamycin resistance to E. coli and G418 resistance to was derived from 46,XY primary human fibroblast MRC-5 mammalian cells. To determine whether transfection of a cells passaged 25 generations before harvest. MRC-5 cells YAC generated from pYACneo into mammalian cells would were chosen because (i) these cells are available to other result in G418 resistance, the following experiment was researchers and (it) they are not known to undergo somatic performed. A 75-kbp YAC containing inserted mouse DNA rearrangements. DNA was prepared as described (1) with the was transferred from yeast into mouse 3T3 cells by sphero- following modifications: after partial EcoRI digestion, DNA plast fusion and subsequent G418 selection (13). From 106 was fractionated by velocity sedimentation to eliminate frag- mouse cells fused, three independent G418-resistant colonies ments <50 kbp. Gradient fractions were collected, assayed were recovered; no resistant colonies were recovered from by field-inversion gel electrophoresis (FIGE) (15), dialyzed, control fusions between 3T3 cells and the yeast host strain and concentrated. This DNA then was ligated to pYACneo AB1380. DNA was prepared in agarose blocks from the three vector DNA prepared as described (1) (Fig. 1). Large DNA resistant clones and analyzed by contour-clamped homoge- ligation products were fractionated from unligated vector neous electric field pulse-field gel electrophoresis (14). By arms by velocity sedimentation and were purified as de- Southern blot hybridization, all three clones appeared to have scribed above. DNA was transformed into S. cerevisiae incorporated the Neo gene into high molecular weight DNA strain AB1380 as described (1), with the following modifica- >2 Mb. High molecular weight DNA from two of the clones tion. Yeast spheroplasts regenerate in a thin layer of calcium also hybridized to a pBR322-derived probe specific for the alginate with colonies appearing on the surface of the plate opposite end of the YAC. After digestion ofDNA from these (16), thus eliminating the need to pick each colony from two clones with Xho I, which does not cut within the 60-kbp within top agar. During the transformation procedure, 2 vol

Eco RI, Cloning site

NeoR

BamHl, EcoRl, phosphatase & size-fractionated partial EcoRl digested MRC-5 DNA DNA ligase

MRC-5 insert DNA (65-150 kb) Xhol R EcoRI EcoRI NeoR Xhol I I TEL TRP1 CEN4 URA3 TEL ARS1 FIG. 1. Vector construction. For construction of the vector pYACneo (pNN415), the YAC vector pYAC4 (1) was modified as follows: a 4.2-kbp Pvu I/BamHI fragment containing the neomycin-resistance gene and the ColE1 origin of replication from pSV2neo (12) was cloned into the unique Sal I site of pYAC4 by filling in the ends of the Pvu I/BamHI fragment with the Klenow fragment of E. coli DNA polymerase I, ligating to Xho I linkers, cleaving with Xho I endonuclease, and ligating to Sal I-cleaved pYAC4. This regenerated a BamHI site, which was converted to a Sal I site by partial cleavage with BamHI, blunt-end repair, and ligation to unphosphorylated Sal I linkers. Open boxes, S. cerevisiae DNA; solid boxes, Tetrahymena ; stippled box, neomycin-resistance gene from pSV2neo. Downloaded by guest on September 29, 2021 5900 Genetics: Traver et al. Proc. Natl. Acad. Sci. USA 86 (1989)

of 1 M sorbitol were added to spheroplasts after incubation A in spheroplast recovery medium (17). Spheroplasts were pelleted for 4 min at 250 x g in a Sorval GLC2B rotor and resuspended in the absence of calcium ion. For each 150-mm transformation plate, spheroplasts were resuspended in 75 ILI 6 of 1 M sorbitol to which an equal volume of 4% sodium alginate (Sigma; alginic acid, low viscosity, sodium salt) in 1 M sorbitol was added. Spheroplasts were spread rapidly in a thin layer onto a plate of synthetic complete medium lacking uracil containing 1 M sorbitol and 0.025 M CaCI2. Plates were incubated at 30'C for 3-4 days before processing. Transfor- mation with this protocol was of equal efficiency to top agar regeneration: 300 transformants per jug of insert DNA. To determine the size of the YACs, genomic DNA in agarose plugs was prepared from 20 randomly selected clones and analyzed by FIGE. The artificial chromosomes had an C 1 2 3 average insert size of 0.1 Mbp of human DNA, with a range of 0.07-0.15 Mbp. In the library there was no detectable background of transformation by religated vector arms as determined by color assay (18). On 50 large plates, =30,000 independently transformed yeast colonies were generated, comprising approximately one genome equivalent of human insert DNA. Library Screening. Colonies were replica plated directly -220 kb from each primary transformation plate to three SCUT - 120 kb plates, two ofwhich were further replica plated after 1-2 days at 300C to a total of six YPD plates overlaid with nitrocellu- lose discs. The discs were processed for colony hybridization (10). Frozen nitrocellulose replicas for storage were gener- FIG. 2. Yeast colony hybridization of a single-copy human DNA ated from the third synthetic plate (see Materials and Meth- sequence to an artificial chromosome. (A) Autoradiogram of yeast ods). After >6 months storage at -70'C, accurate replicas of colony hybridization. One colony, AB1380/hX2 (YNN375), from yeast colonies that retain the selectable markers for the among 30,000 colonies in a human genomic DNA YAC library artificial chromosomes have been recovered. hybridized specifically to probe p625.8 in a reaction mixture con- taining four DNA probes. Nitrocellulose discs were hybridized as The human genomic library was screened with six different recommended by the manufacturers (Schleicher & Schuell) in a total probes. Pooled probes were used to increase the chances of vol of 400 ml with 5 x 106 cpm/ml of random hexamer-primed finding a positive signal in an estimated one genome equiv- 32P-labeled DNA (21). Discs were subjected to autoradiography with alent of human 46,XY DNA and to minimize the number of intensifying screens for 1 day at -700C. (B) Visualization of YACs. rounds of screening per set of discs. In an effort to identify Undigested genomic DNA from S. cerevisiae was fractionated by sequences in the region assigned to the X-chromosome linked FIGE (switching-interval ramp, 0.3 sec, with a constant 3:1 ratio bipolar disorder locus (Xq28) (19, 20), the library was between forward and reverse intervals for 22 hr; forward migration interval, 9 sec) and stained with ethidium bromide. Lanes: 1, DNA screened initially with three unique sequence probes (MN12, from a control colony of strain AB1380 with a 160-kbp artificial DX13, and hs7) and one probe that detects a sequence chromosome (AB1380/hCl); 2, DNA from a colony with a 150-kbp present in several copies within a relatively small (<100 kbp) artificial chromosome (AB1380/hXl) that hybridized to the MN12 interval (Stl4) (5, 7). Hybridization conditions were as rec- DNA probe; 3, DNA from a colony with a 120-kbp artificial chro- ommended by the nitrocellulose disc manufacturer (Schle- mosome (AB1380/hX2) that hybridized to the p625.8 DNA probe. icher & Schuell). After 5 days of autoradiography with Sizes of the YACs and the smallest endogenous yeast chromosome intensifying screens at -70°C, 1 positive colony was detected (220 kbp) were determined by contour-clamped homogeneous elec- among the 30,000 screened. Overall background hybridiza- tric field gel electrophoresis (14) with a pulse interval of 12 sec or by tion was low. FIGE using A ladders (FMC BioProducts) as size standards. (C) Autoradiogram of DNA blot hybridization to p625.8 ofthe gel shown Undigested genomic DNA from this colony was prepared in B. Lanes are the same as in B. The 120-kbp artificial chromosome in agarose and separated by FIGE. Southern blot hybridiza- of AB1380/hX2 (lane 3) hybridizes specifically to probe p625.8, tion revealed that this yeast colony, AB1380/hX1 (YNN374), confirming the positive signal for colony hybridization in A. contains a 150-kbp YAC to which probe MN12 showed specific hybridization (data not shown). Subsequent analysis human genomic library, YACs that hybridize to unique revealed that hX1 contains a sequence homologous, but not single-copy human DNA sequences. identical, to MN12 (S.K., unpublished data). A subclone Recovering Ends of YAC Inserts as Plasmids in E. coli. In (pNN415) containing 3.5 kbp of single-copy DNA sequence addition to conferring G418 resistance to mammalian cells, was derived from one end of this YAC insert (see below). the presence of the Neo gene in pYACneo facilitates the This probe was used to rescreen the library together with recovery in E. coli of plasmids containing each end of a YAC three Xq28 sequences: St14, hs7, and p625.8. This second insert. An attempt was made to recover in plasmids the ends screen identified two positive colonies with equally intense of the human DNA inserts from artificial chromosomes hX1 hybridization signals. One, which served as an internal and hX2. DNA from the 150-kbp hX1 artificial chromosome control for hybridization, was the original colony from which was separated from yeast chromosomal DNA by preparative the subcloned single-copy probe pNN415 was derived. The FIGE. The hX1 DNA was digested in agarose with Xho I, second positive yeast colony, AB1380/hX2 (YNN375) (Fig. extracted from the gel, and concentrated using an Elutip-d 2A), was shown by Southern blot analysis to contain a column (Schleicher & Schuell), ligated overnight at 15°C and 120-kbp artificial chromosome that hybridized specifically to used to transform E. coli to either ampicillin or kanamycin probe p625.8 (Fig. 2 B and C). Thus, conditions have been resistance. Only ampicillin-resistant colonies were recov- determined by which one can identify rapidly, within a total ered; these colonies contained a 13-kbp with an Downloaded by guest on September 29, 2021 Genetics: Traver et al. Proc. Natl. Acad. Sci. USA 86 (1989) 5901 8-kbp EcoRI/Xho I human DNA insert. The hX1 YAC was It may be possible to modify the pYACneo vector to develop shown by Southern blot analysis using a vector-specific an efficient yeast-mammalian shuttle system for large frag- probe to contain a second Xho I site located -50 kbp from the ments of DNA. This would allow genetic complementation end of the Neo-containing arm. Thus, this end fragment was with entire cloned genes-in particular, large transcription probably too large to recover efficiently as a plasmid in E. coli units such as the factor VIII gene (9)-and could aid in using the DNA recovery, ligation, and transformation meth- generating the proper levels of expression. One could study, ods described above. Similarly for the 120-kbp artificial in a more natural context, effects of cis-acting regulatory chromosome hX2, neither end of the human insert was elements operating at a great distance or mechanisms such as recovered when Xho I cleavage, ligation, and E. coli trans- alternative splicing (25). Mutations could be introduced into formation were used. In this instance, the two Xho I sites cloned human genes through in nearest to the ends of the hX2 insert were located -40 kbp yeast (26), and the effects of these mutations could be from the telomeric Xho I sites. One end ofthe hX2 insert was analyzed by shuttling recombinant artificial chromosomes recovered successfully as a 1.9-kbp human DNA fragment in into mammalian cells. In addition, the pYACneo artificial a plasmid (pNN416) using EcoRV digestion ofyeast miniprep chromosome vector may be a good host for the addition and DNA (see Fig. 1), ligation, and E. coli transformation with testing ofstructural chromosomal elements for their ability to kanamycin selection. stabilize a linear artificial chromosome in human cells. For The recovery of only one of the three insert ends using example, human alpha satellite repeats are found at centro- Xho I digestion was not surprising since Xho I cleaves meric regions and are candidates for being the functional infrequently in human DNA, with =1 site per 100 kbp (22). The infrequency is caused both by the underrepresentation of human centromere (27). Two percent of the YACs in this CG dinucleotides in human DNA (22) and by CG methylation human genomic library have been shown to contain alpha (which is not expected to occur when human DNA is intro- satellite DNA (C.N.T., unpublished data), reflecting the duced into S. cerevisiae) (23). Additional vector modifica- percentage of alpha satellite found in total human DNA (27). tions clearly are needed to facilitate the recovery of terminal The ability of an alpha satellite array to confer mitotic fragments to allow rescreening a library for overlapping stability to an artificial chromosome transferred from yeast to clones. Possible approaches include incorporating more ver- human cells could be tested in this shuttle system. satile restriction sites into the vector arms between the telomeric and the We thank Mary Smith for excellent technical assistance, Jane sequences drug-resistance genes, and add- Gitschier and Douglas Vollrath for gifts of plasmids, Sue Kenwrick ing T3 and T7 promoters adjacent to the cloning site for the and Jane Gitschier for communicating unpublished results, and Pat generation of short RNA probes (24). Brown for critically reading the manuscript. This work was sup- ported by the Medical Scientist Training Program at Stanford Uni- DISCUSSION versity (C.N.T.), National Institute of Mental Health Postdoctoral Fellowship Grant 1F32 MH09784-01 (S.K.), a John Simon Guggen- An artificial chromosome library in yeast has been generated heim Memorial Foundation Fellowship (R.W.H.), and National from primary human fibroblast DNA. Approximately 30,000 Institutes of Health Grant 5 R37 GM21891 and National Institute of human DNA fragments, averaging 0.1 Mbp, have been Aging Grant P01 AG02908 to R.W.D. R.W.H. was on sabbatical cloned in a modified YAC vector. This vector, pYACneo, leave from the Pennsylvania State University College of Medicine. enables recovery within plasmids ofboth ends of an insert for rescreening a library to identify overlapping clones. The Neo 1. Burke, D. T., Carle, G. F. & Olson, M. V. (1987) Science 236, gene from this vector is able to confer G418 resistance to 806-812. mammalian cells fusion with har- 2. Coulson, A., Waterston, R., Kiff, J., Sulston, J. & Kohara, Y. upon yeast spheroplasts (1988) Nature (London) 335, 184-186. boring a YAC. A rapid method for yeast colony hybridization 3. Hahnenberger, K. M., Baum, M. P., Polizzi, C. M., Carbon, J. uses yeast spheroplast regeneration in a thin layer of calcium & Clarke, L. (1989) Proc. Natl. Acad. Sci. USA 86, 577-581. alginate for direct replica plating and processing of colonies 4. Little, R. D., Porta, G., Carle, G. F., Schlessinger, D. & from the primary transformation plate to nitrocellulose discs. D'Urso, M. (1989) Proc. Natl. Acad. Sci. USA 86, 1598-1602. This transformation protocol has an efficiency equal to that 5. Oberle, I., Drayna, D., Camerino, G., White, R. & Mandel, of top agar regeneration. A tremendous amount of time is J.-L. (1985) Proc. Natl. Acad. Sci. USA 82, 2824-2828. saved by eliminating the need to pick colonies from within top 6. Drayna, D., Davies, K. E., Hartley, D. A., Mandel, J.-L. agar. In addition, the storage method described in this paper Camerino, G., Williamson, R. & White, R. L. (1984) Proc. allows for long-term maintenance of accurate replicas of the Natl. Acad. Sci. USA 81, 2836-2839. library using minimal space. 7. Patterson, M., Kenwrick, S., Thibodeau, S., Faulk, K., Mattei, This library contains approximately one genome equiva- M. & Davies, K. (1987) Nucleic Acids Res. 15, 2639-2650. lent of human DNA. with five different 8. Nathans, J., Thomas, D. & Hogness, D. (1986) Science 232, 46,XY Screening 193-202. probes from the Xq28 region revealed two positive colonies, 9. Gitschier, J., Drayna, D., Tuddenham, E., White, R. & Laun, one of which contained DNA from the Xq28 region (hX2). R. (1985) Nature (London) 314, 738-740. The other colony contained a sequence with homology to 10. Sherman, F., Fink, G. R. & Hicks, J. B. (1986) Laboratory Xq28 probe MN12 (hXl) but was not derived from the MN12 Course Manual for Methods in Yeast Genetics (Cold Spring region of the X chromosome. Since an estimated 0.5 genome Harbor Lab., Cold Spring Harbor, NY). equivalents of X chromosome were present in this library, a 11. Murray, A. W. & Szostak, J. W. (1983) Nature (London) 305, more comprehensive library must be generated to increase 189-193. the probability of finding any unique sequence of interest. 12. Southern, P. & Berg, P. (1982) J. Mol. Appl. Genet. 1, 327-341. Simultaneous hybridization with four probes to screen one 13. Ward, M., Scott, R. J., Davey, M. R., Clothier, R. H., Cock- set of discs produced a strong signal with little background, ing, E. C. & Balls, M. (1986) Somatic Cell Mol. Genet. 12, suggesting that a substantially larger number of probes could 101-109. be 14. Chu, G., Vollrath, D. & Davis, R. W. (1986) Science 234, pooled during hybridization. Filters also can be reused for 1582-1585. repeated rounds of hybridization. 15. Carle, G. F., Frank, M. & Olson, M. V. (1986) Science 232, Transfer of an artificial chromosome constructed in 65-68. pYACneo from yeast to mammalian cells can confer G418 16. Lambie, E. J. & Roeder, G. S. (1986) Genetics 114, 769-789. resistance to mammalian cells. In the three resistant clones 17. Burgers, P. M. J. & Percival, K. J. (1987) Anal. Biochem. 163, studied, the DNA had integrated into a mouse chromosome. 391-397. Downloaded by guest on September 29, 2021 5902 Genetics: Traver et al. Proc. Nati. Acad. Sci. USA 86 (1989)

18. Hieter, P., Mann, C., Snyder, M. & Davis, R. W. (1985) Cell 23. Proffitt, J. H., Davie, J. R., Swinton, D. A. & Hattman, S. 40, 381-392. (1984) Mol. Cell. Biol. 4, 985-988. 19. Baron, M., Risch, N., Hamburger, R., Mandel, B., Kushner, 24. Wahl, G. M., Lewis, K. A., Ruiz, J. C., Rothenberg, B., Zhao, S., Newman, M., Drumer, D. & Belmaker, R. (1987) Nature J. & Evans, G. A. (1987) Proc. Nail. Acad. Sci. USA 84, (London) 326, 289-292. 2160-2164. 20. Mendlewicz, J., Simon, P., Sevy, S., Charon, F., Brocas, H., 25. Saga, Y., Tung, J. S,. Shen, F. W., Pancoast, T. C. & Boyse, Legros, S. & Vassart, G. (1987) Lancet i, 1230-1232. E. A. (1988) Mol. Cell. Biol. 8, 4889-4895. 21. Feinberg, A. P. & Vogelstein, B. (1983) Anal. Biochem. 132, 26. Scherer, S. & Davis, R. W. (1979) Proc. Nadl. Acad. Sci. USA 6-13. 76, 4951-4955. 22. Nelson, M. & McClelland, M. (1987) Nucleic Acids Res. 15, 27. Waye, J. S. & Willard, H. F. (1987) Nucleic Acids Res. 15, 219-230. 7549-7569. Downloaded by guest on September 29, 2021