Proc. Natl. Acad. Sci. USAVol. 88, pp. 8106-8110, September 1991Cell Biology

Centromere formation in mouse cells cotransformed with humanDNA and a dominant marker gene

(functional centromere/cotransfection/centromere-linked marker/i o c /in situ hybridization)

GYULA HADLACZKY*t, TUNDE PRAZNOVSZKY*, IMRE CSERPAN*, JUDIT KERESO*, MIKL6S PNTERFY*,ILDIK6 KELEMEN*, EROL ATALAY*, ANNA SZELESt, J6ZSEF SZELEI§, VILMOS TUBAK*,AND KORNtL BURG**Institute of Genetics Biological Research Center, Hungarian Academy of Sciences, 6701 Szeged, P.O. Box 521, Hungary; tDepartment of Genetics, J6zsefAttila University, Szeged, Hungary; and 1lnstitute of Biochemistry, Biological Research Center, Hungarian Academy of Sciences, Szeged, Hungary

Communicated by Robert L. Sinsheimer, April 18, 1991 (receivedfor review February 6, 1991)

ABSTRACT A 13,863-base-pair (bp) putative centromericDNA fragment has been isolated from a human genomic libraryby using a probe obtained from metaphase chromosomes ofhuman colon carcinoma cells. The abundance of this DNA wasestimated to be 16-32 copies per genome. Cotransfection ofmouse cells with this sequence and a selectable marker gene(aminoglycoside 3'-phosphotransferase type II, APH-II) re-sulted in a transformed cell line carrying an additional cen-tromere in a dicentric chromosome. This centromere wascapable of binding an anti-centromere antibody. In situ hy-bridization demonstrated that the human DNA sequence aswell as the APH-II gene and vector DNA sequences werelocated only in the additional centromere of the dicentricchromosome. The extra centromere separated from the dicen-tric chromosome, forming a stable minichromosome. Thisfunctional centromere linked to a dominant selectable markermay be a step toward the construction of an artificial mam-malian chromosome.

The centromere is a specialized region of the eukaryoticchromosome that is the site of kinetochore formation, astructure that allows the precise segregation ofchromosomesduring cell division and may play a role in the higher-orderorganization of eukaryotic chromosomes (1). The isolationand cloning ofcentromeres is essential not only to understandtheir molecular structure and function but also to constructstable artificial chromosomes. By taking advantage of cen-tromere-linked genes, functional centromeres of lower eu-karyotes (yeast) were isolated (for reviews, see refs. 2 and 3)and then combined with telomeres, which stabilize the chro-mosome ends, to permit the construction of yeast artificialchromosomes (4, 5) and initiate studies of chromosomefunction and genetic manipulation.

Unlike yeast, higher eukaryotic cells contain repetitiveDNA sequences that form a boundary at both sides of thecentromere. This highly repetitive DNA interacts with cer-tain proteins, especially in animal chromosomes, to create agenetically inactive zone around the centromere. This peri-centric heterochromatin keeps any selectable marker gene ata considerable distance, and thus repetitive DNA preventsthe isolation of centromeric sequences by chromosome"walking."

In this report we describe the isolation ofa 14-kilobase (kb)DNA fragmentll obtained from a human genomic library. Theprobe used for library screening was isolated from mechan-ically fragmented human metaphase chromosomes by theimmunoprecipitation with anti-centromere antibodies. Bycotransfection of the 14-kb fragment of putative centromeric

DNA and a selectable marker [aminoglycoside 3'-phospho-transferase gene (APH-TI gene)], a mouse cell line wasproduced that carries a functioning additional centromereeither in a dicentric chromosome or in a minichromosome.The newly formed centromere was capable of binding mousecentromere proteins, as shown by anti-centromere antibodybinding. Furthermore, the human DNA, the APH-II gene,and vectorDNA sequences were colocalized to the additionalcentromere. This transformed cell line, carrying a cen-tromere-linked selectable marker, provides a tool to isolatecomponents of a functional centromere and to build anartificial mammalian chromosome.

MATERIALS AND METHODSCell Lines. The human colon carcinoma cell line COLO 320

was grown in suspension in RPMI 1640 medium supple-mented with 10%6 (vol/vol) fetal calf serum. Mouse fibroblastcells (LMTK-) were maintained as a monolayer in F12medium supplemented with 10%6 fetal calf serum.Chromosome Isolation and Immunoprecipitation of Chro-

mosome Fragments. Metaphase chromosomes of COLO 320cells were isolated by our standard method (6). Isolatedmetaphase chromosomes were resuspended in 1 ml of buffer(150mM NaCl/50mM Tris-HCl, pH 7.5/10mM MgCl2/5 mM2-mercaptoethanol) at a DNA concentration of 1 mg/ml anddigested with 500 units ofEcoRI restriction endonuclease for1 h. The suspension was diluted with 4 ml of IPP buffer (500mM NaCI/10 mM Tris-HCl, pH 8.0/0.1% Nonidet P40),sonicated for five 30-s bursts with an MSE 5-70 sonicator, andthen centrifuged at 1500 x g for 10 min to remove unbrokenchromosomes and large chromosome fragments. The super-natant contained only small (<1 um) chromosome fragmentsas judged by light microscopy.

Protein A-Sepharose CL4B (Pharmacia; 250 mg) wasswollen in IPP buffer and incubated with 500 u1 of humananti-centromere serum LU851 (7) diluted 1:20 with IPPbuffer. A suspension of sonicated chromosome fragments (5ml) was mixed with anti-centromere-Sepharose (1 ml) andincubated at room temperature for 2 h with gentle rolling.After three subsequent 25-ml washes with IPP buffer, theSepharose was centrifuged at 200 x g for 10 min.DNA Techniques. DNA was isolated from the immunopre-

cipitate by proteinase K treatment (Merck, 100 l.g/ml) in 10mM Tris-HCI, pH 8.0/2.5 mM EDTA/1% SDS at 500Covernight, followed by repeated phenol extractions, andprecipitation with isopropanol. All general DNA manipula-

Abbreviation: APH-II, aminoglycoside 3'-phosphotransferase typeIItTo whom reprint requests should be addressed.'The sequence reported in this paper has been deposited in theGenBank data base (accession no. M64093).

8106

The publication costs of this article were defrayed in part by page chargepayment. This article must therefore be hereby marked "advertisement"in accordance with 18 U.S.C. §1734 solely to indicate this fact.

Dow

nloa

ded

by g

uest

on

Janu

ary

23, 2

021

Proc. Natl. Acad. Sci. USA 88 (1991) 8107

tions were done according to ref. 8. DNA probes were labeledfor library screening and for Southern blot analysis byrandom oligonucleotide priming (9). DNA sequencing wasperformed by the dideoxynucleotide method (10, 11).

Isolation of A Clones. A A Charon 4A human genomiclibrary (12) was screened with 32P-labeled (at 2 x 108 cpm/j,specific activity) high molecular weight DNA isolated fromthe immunoprecipitated fragments. Eight strongly hybridiz-ing clones were selected from 4 x 10- plaques.

Transfection of Mouse Cells. The calcium phosphatemethod (13), with 40 ,ug of ACM8 and 40 ,ug of AgtWESneoDNA per Petri dish (80 mm) and a 2-min glycerol shock, wasused for transfection. Transformed cells were selected ongrowth medium containing G418 (Geneticin, Sigma; 400,ug/ml).In Situ Hybridization. In situ DNA-DNA hybridization to

metaphase chromosomes was done with [3H]thymidine-labeled probes as described (14) or with biotin-labeled probesas described (15).

Indirect Immunofluorescence. Indirect immunofluores-cence of mouse metaphase cells was done as described (7).When indirect immunofluorescence and in situ hybridizationwere performed on the same metaphases, mitotic cells wereresuspended in a glycine/hexylene glycol buffer (7), swollenat 370C for 10 min, cytocentrifuged, and fixed with methanolat -20'C. After the standard immunostaining (7), metaphaseswere photographed, then coverslips were removed withphosphate-buffered saline, and slides were fixed in ice-coldmethanol/acetic acid [3:1 (vol/vol)], air-dried, and used forin situ hybridization.Microscopy. In the in situ hybridization experiments with

biotin-labeled probes, chromosomes were counterstainedwith propidium iodide (14), whereas in indirect immunoflu-orescence the DNA binding dye Hoechst 33258 was used.

RESULTSIsolation of Putative Centromeric DNA. Metaphase chro-

mosomes isolated from a human colon carcinoma cell linewere mechanically fragmented by sonication after a limitedendonuclease digestion. The suspension containing smallchromosome fragments was incubated with human anti-centromere serum coupled to protein A-Sepharose CL-4B.DNA was isolated from the chromosome fragments bound toanti-centromere-Sepharose. The centromeric DNA is in thestructurally most-stable region of mammalian chromosomes(16). Therefore, this DNA is the most resistant to enzymaticdigestion and mechanical shearing. Results of electrophore-sis of immunoprecipitated and supernatant DNA supportthese observations (Fig. 1). The bulk ofDNA from chromo-some fragments that did not bind to the anti-centromere-Sepharose (supernatant) ranged from several hundred basepairs to 5 kb (Fig. 1, lanes A and B) whereas chromosomefragments bound to the anti-centromere-Sepharose con-tained an additional population of high molecular weight(9-20 kb) DNA fragments (Fig. 1, lanes C and D). This highmolecular weight DNA was isolated from the agarose gel andused as a probe to screen a A Charon 4A human genomiclibrary (12), and 8 of the most strongly hybridizing cloneswere selected from 4 x 105 plaques. Restriction enzymeanalysis of the selected clones proved that these clones wereidentical, probably due to amplification of the genomic li-brary. One of these clones (CM8) containing a 14-kb humanDNA insert was used for further experiments. The restrictionmap of the 14-kb human DNA insert for some relevantrestriction endonucleases is shown in Fig. 2.

Southern blots of human lymphocyte DNA digested withvarious restriction endonucleases and probed with subfrag-ments of the CM8 insert suggested that the 14-kb insertrepresents an uninterrupted part of the human genome (T.P.,

FIG. 1. Agarose gel electrophoresisof DNA fragments obtained by immu-noprecipitation. Lanes: A and B, DNAisolated from chromosome fragmentsremaining unbound to anti-centromere-Sepharose; C and D, DNA isolatedfrom chromosome fragments bound toanti-centromere-Sepharose (note thepresence of a population of high mo-lecular weight DNA fragments); B andD, samples treated with RNase A (100,ug/ml) prior to electrophoresis; M,

A B C D M HindIII-digested A markers.

J.K., and G.H., unpublished results). Thus, extreme left andright EcoRI subfragments of the CM8 insert form parts of10-kb and 13-kb genomic EcoRI fragments, respectively, andthe inner subfragments of the insert are the same size ascorresponding fragments ofgenomic DNA (data not shown).

In Southern blot experiments by simultaneously hybridiz-ing a single-copy DNA probe (XV2c) (17) and the central XhoI-EcoRI fragment of the CM8 insert (Fig. 2) with seriallydiluted human peripheral lymphocyte DNA, the copy num-ber of CM8 was estimated to be 16-32 per genome (data notshown).In situ hybridization of [3H]thymidine-labeled CM8 DNA

to human (COLO 320) metaphase chromosomes resulted inthe preferential localization of silver grains to the cen-tromeres (Fig. 3). We found a centromeric signal on repre-sentatives of all chromosome groups, and no preferentialhybridization to any individual human chromosome wasdetected.

In situ hybridization using biotin-labeled subfragments orthe whole CM8 insert was not detected by the standardmethod (15). Furthermore, if a hybridization method that issuitable for single-copy-gene detection with a biotin-labeledprobe was used (18), apart from the typical R-band-like Aluhybridization pattern (19), no specific hybridization signalwas detected on any of the chromosomes with the whole14-kb CM8 insert. A likely explanation for this negative resultis that centromeric sequences are not accessible to thehybridization probe due to their compact packing.The sequence of the 13,863-base-pair human CM8 clone

showed no homology to any known sequence when com-pared with a complete nucleic acid data bank [MicroGenie

A CM8 (13.863 bp)

pCM21 pCMl pKI620 pCMIll pcM6.1. .1.

3350 440 4366

ApuE E E IBgX

.1 * a a la 'b -

'9 91 iiI IfrII W 1P pprsvrP PP P

pCMl 6 EcoRI

1 subdones1859 808 3046bp

HE KEIH-EBp1a I Ip

prrpp

E=EcoRI Bg=Bgl II. X=Xho I. H=Hind Ill,K=Kpn I. B-BaHl, P=Pst I, Al=Ahu repeats

Bg E. P pI'll I r4

wppp P

1 kb

FIG. 2. Restriction endonuclease map of the human genomicDNA insert of CM8 A Charon 4A clone. Arrows show the positionsof a 300-base-pair Alu repeat deficient in the flanking direct repeatsequences.

Cell Biology: Hadlaczky et al.

Dow

nloa

ded

by g

uest

on

Janu

ary

23, 2

021

8108 Cell Biology: Hadlaczky et al.

a

I ':'

4^

_ _w 19. _ 11.-21

U-m-rn- 61

A

B

FIG. 3. In situ hybridization of [3H]thymidine-labeled CM8 DNAto human metaphase chromosomes. (A) Preferential localization ofsilver grains to the centromeres of human chromosomes (arrow-heads). (B) Distribution of silver grains (o) on 131 metacentricchromosomes. Numbers indicate the frequency of silver grain local-ization to certain regions of the chromosomes.

sequence software (MG-1M-5.0), Beckman]. However, a300-base-pair Alu repeat, deficient in the flanking directrepeat sequences, was found in the 2.5-kb EcoRI-Xho Ifragment (Fig. 2), which explains the Alu-type in situ hybrid-ization pattern obtained with the CM8 insert.Human DNA and the Dominant Selectable Marker Form a

Centromere. To detect any in vivo centromere function ofthehuman putative centromeric DNA, the CM8 DNA and theselectable APH-II gene were introduced into mouse LMTK-cells by the calcium phosphate coprecipitation method (13).The pAG60 plasmid (20) containing the APH-II gene wascloned into a AgtWES (21) bacteriophage vector. We decidedto introduce the whole ACM8 and AgtWESneo constructionsinto mouse cells for two reasons. (i) By separating the markergene from the putative centromeric sequences, we hoped thatit would be possible to avoid inactivating the APH-II gene, a

process that may occur during centromere formation. (it) ADNA is capable of forming long tandem arrays of DNAmolecules by concatamerization, and in Schizosaccharomy-ces pombe 4-15 copies of conserved sequence motifs formcentromeres (22). Thus a multiplication of the putative cen-tromeric DNA by concatamerization might increase thechance of centromere formation.Transformed cells were selected on growth medium con-

taining G418 (400 ,ug/ml). Individual G418-resistant cloneswere analyzed as follows. The presence ofhuman sequencesin the transformed clones was monitored using Southernblots probed with subfragments of the CM8 insert. Excesscentromeres were screened for by indirect immunofluores-cence using human anti-centromere serum LU851 (7). Thechromosomal localization of "foreign" DNA sequences wasdetermined by in situ hybridization with biotin-labeledprobes.

Eight transformed clones have been analyzed. All of theclones contained human DNA sequences integrated intomouse chromosomes. However, only two clones (EC5/6 andEC3/7) showed the regular presence of dicentric chromo-somes. Individual cells of clone EC5/6 carrying di-, tri-, andmulticentromeric chromosomes exhibited extreme instabil-ity. In more than 60% of the cells of this cell line, thechromosomal localization of the integrated DNA sequencesvaried from cell to cell. Due to this instability, clone EC5/6proved to be unsuitable for the present study. However, cellsof clone EC3/7 were stable, carrying either a dicentric (85%)or a minichromosome (10%6). Centromeres of dicentric chro-mosomes and minichromosomes were indistinguishable fromthe mouse centromeres by immunostaining with anti-centromere antibodies (Fig. 4 A and B). To determinewhether newly introduced DNA was contributing to cen-tromere formation, in situ hybridization with biotin-labeledCM8 and A phage DNA was carried out. Without exceptionthese probes hybridized to the same spots: either the distalcentromere of the dicentric chromosome (Fig. 4C) or thecentromere ofthe minichromosome (Fig. 4D). In less than 5%of the EC3/7 cells was an alternative localization of thehybridization signal found. These included more than one

FIG. 4. Detection of dicentric and minichromosome of the EC3/7 cells by indirect immunofluorescence with anti-centromere antibodies (Aand B) and by in situ hybridization with biotin-labeled CM8 probes (C) or with A (vector) DNA sequences (D). A and C are the same field andB and D are the same field. Counterstaining of chromosomes was with propidium iodide. Arrowheads point to dicentric and minichromosomes.

Proc. Natl. Acad. Sci. USA 88 (1991)

Dow

nloa

ded

by g

uest

on

Janu

ary

23, 2

021

Proc. Natl. Acad. Sci. USA 88 (1991) 8109

integration site, the lack of a detectable signal, or hybridiza-tion to chromosomes other than the dicentric chromosome.To demonstrate the integration of the human sequence and

the APH-II gene in the centromeric region, immunostainingof centromeres with anti-centromere antibodies followed byin situ hybridization with CM8 and APH-II sequences wascarried out on the same metaphase plates ofEC3/7 cells. Thein situ hybridization signals with both biotin-labeled probescolocalized with the immunostained centromeric region ofthe chromosomes carrying additional centromeres (Fig. 5).Extra Centromeres Are Functioning. Extra centromeres of

dicentric or minichromosomes bound the anti-centromereantibodies ofthe LU851 serum. By using this anti-centromereserum on a multicentromeric chromosome of a mouse cellline, only one centromere was labeled by the antibodies (datanot shown). In hydroxyurea/caffeine-treated cells, where thechromosomes were fragmented in vivo (23), our anti-centromere serum recognized only the active centromeres(Kdroly Fdtyol, I.C., T.P., V.T., F.K., and G.H., unpub-lished results).The regular breakage of the dicentric chromosome and the

formation ofthe minichromosome strongly suggested that theextra centromeres were capable of binding mitotic spindles.When the mouse centromere and the extra centromere arepulled in opposite directions, the extra centromere couldbreak-off.By subcloning EC3/7 cells, we have established two single-

cell-derived stable cell lines (EC3/7C5 and EC3/7C6) thatcarry the extra centromere on minichromosomes. Indirectimmunofluorescence with anti-centromere antibodies andsubsequent in situ hybridization experiments proved that theminichromosomes were derived from the dicentric chromo-some. In both cell lines, the majority ofthe hybridization signalwas found on the minichromosomes but traces of CM8 or Asequences were detected at the end of a monocentric chro-mosome (Fig. 6). In EC3/7C5 and EC3/7C6 cell lines (140 and128 metaphases, respectively), no dicentric chromosome wasfound and minichromosomes were detected in 97.2% and98.1% of cells, respectively. These cell lines provided directevidence that the extra centromere was functioning and wascapable of maintaining the minichromosomes.

Stability of Transformed Cell Lines. Immunofluorescenceanalysis ofEC3/7 cells (103 metaphases) cultured for 46 daysin nonselective medium showed that 80.6% of the cellscontained either a dicentric chromosome (60.2%) or a mini-chromosome (20.4%). Subsequent hybridization with biotin-labeled probes proved the presence of "foreign" DNA in theadditional centromeres (data not shown). These results indi-cated that no serious loss or inactivation of the extra cen-

FIG. 6. Indirect immunofluorescence with anti-centromere se-rum (A) and in situ hybridization of biotin-labeled CM8 sequences (B)to metaphase of EC3/7C5 cell line carrying stable minichromosome.Arrowheads show the minichromosome and traces of CM8 se-quences left at the end of the formerly dicentric chromosome.Counterstaining was with propidium iodide.

tromeres occurred during this period of culture under non-selective conditions.To estimate the copy number ofthe integrated human DNA

in the extra centromere, DNA of the EC3/7 cell line andhuman lymphocyte DNA were digested with restriction en-donucleases and probed with subfragments of the CM8 insertin a Southern blot hybridization experiment. By comparingthe intensity of the hybridization signal on EC3/7 DNA tothat of the human DNA, the minimum number of integratedhuman sequences in the additional centromere was estimatedto be >30.

FIG. 5. Colocalization of the integrated DNA sequences with the centromeric region detected by immunostaining with anti-centromereserum. (A) DNA staining with propidium iodide. (B) Indirect immunofluorescence with anti-centromere serum. (C) Subsequent in situhybridization with biotin-labeled APH-II probe to the same dicentric chromosome of the EC3/7 cells. Arrowheads show the extra centromereof the dicentric chromosome.

Cell Biology: Hadlaczky et al.

Dow

nloa

ded

by g

uest

on

Janu

ary

23, 2

021

8110 Cell Biology: Hadlaczky et al.

DISCUSSIONWe have described a transformed mouse cell line (EC3/7)that carries a dicentric chromosome or, less frequently, a

minichromosome. The additional centromere of the dicentricchromosome and the centromere of the minichromosomewere formed in vivo after the cotransfection of mouse cellswith human putative centromeric DNA and a dominantselectable marker (APH-II). The presence of human DNA,the APH-II gene, and the vector sequences in the additionalcentromeres was demonstrated by in situ hybridization.These sequences colocalized with the immunostained cen-tromeric region, as visualized by indirect immunofluores-cence with anti-centromere antibodies. As a rule, one of thecentromeres of dicentric chromosomes is functionally inac-tive (24). In spite of this general phenomenon, three lines ofevidence suggest that the extra centromeres in the EC3/7 cellline are functioning. (i) The extra centromeres bind anti-centromere antibodies from LU851 serum. This anti-centromere serum, which contains four anti-centromere an-tibodies (7), recognizes only active centromeres. This is inagreement with the previous finding (25) that centromereproteins can only be detected by immunostaining at activecentromeres. (ii) Minichromosomes containing the newlyformed centromere regularly appeared in EC3/7 cells. Thefunctioning additional centromere probably leads to chromo-somal breakage during mitosis and, therefore, minichromo-somes are the break-off products caused by the extra func-tional centromere. (iii) Two stable cell lines were isolated bysubcloning EC3/7 cells. In these cell lines, the extra cen-tromere was separated from the mouse chromosome andformed a stable minichromosome that was maintained by thefunctioning extra centromere of EC3/7 origin.

It has been suggested that the inactivation of centromeresmight occur by an alteration in chromatin conformation (25)such as heterochromatinization. However, this process forEC3/7 cells carrying the centromere-linked APH-II genewould be suicidal under selective conditions. In this respect,the proximity of the APH-II gene to the centromere safe-guards the maintenance of the functional centromere.

Calculations made on the basis of the sizes of ACM8 andAgtWESneo constructions and their estimated copy numberssuggest that the integrated DNA may exceed a million basepairs in length. However, rearrangements within the se-quences used for transfection may alter this expected size. Infact, hybridization of the extreme ends of the CM8 insert toEC3/7 DNA suggests a significant rearrangement of thesesequences in the extra centromere (unpublished data). At thistime we cannot rule out the possibility that mouse DNAsequences might be involved in the formation of the addi-tional centromeres. Consequently, we think that it is prema-ture to conclude that the isolated human DNA fragmentpresent in the additional centromere has the sole capacity forthe formation of a functional centromere. In spite of this, thecell line carrying a centromere-linked dominant selectivemarker offers a unique opportunity to isolate the componentsrequired in a functional centromere of a mammalian chro-mosome. Furthermore, these cell lines may provide a steptoward the construction of an artificial mammalian (human)chromosome.

We are grateful to Prof. Bob Williamson for providing the XV2cprobe; to Miss Kati Holland for her critical reading ofthe manuscript;and to Drs. G. B. Kiss, A. A. Szalay, and P. Venetianer for theiruseful suggestions. We thank G. Nollenburg and B. Dusha for the artwork. This work was supported in parts by grants from OKKFT/Tt/286/86 and OTKA 1988/287 (to G.H.). M.P. is on a trainingcourse from the Institute of Biochemistry, Agricultural and Biotech-nological Research Center, H-2101 Goddlld, POB 170, Hungar-y. E.A. was a fellow of the International Training Course of theBiological Research Center in 1987.

1. Hadlaczky, Gy. (1985) Int. Rev. Cytol. 94, 57-76.2. Blackburn, E. H. & Szostak, J. W. (1984) Annu. Rev. Bio-

chem. 53, 163-194.3. Clarke, L. & Carbon, J. (1985) Annu. Rev. Genet. 19, 29-56.4. Murray, A. W. & Szostak, J. W. (1983) Nature (London) 305,

189-193.5. Burke, D. T., Carle, G. F. & Olson, M. V. (1987) Science 236,

806-812.6. Hadlaczky, Gy., Praznovszky, T. & Bisztray, G. (1982) Chro-

mosoma 86, 643-659.7. Hadlaczky, Gy., Praznovszky, T., Rasko, I. & Kereso, J.

(1989) Chromosoma 97, 282-288.8. Maniatis, T., Fritsch, E. F. & Sambrook, J. (1982) Molecular

Cloning:A Laboratory Manual (Cold Spring Harbor Lab., ColdSpring Harbor, NY).

9. Feinberg, A. P. & Vogelstein, B. (1983) Anal. Biochem. 132,6-13.

10. Sanger, F., Coulson, A. R., Barrel, B. G., Smith, A. J. H. &Roe, B. A. (1980) J. Mol. Biol. 143, 161-178.

11. Biggin, M. D., Gibson, T. J. & Houg, G. F. (1983) Proc. Natl.Acad. Sci. USA 80, 3963-3965.

12. Maniatis, T., Hardison, R. C., Lacy, E., Lauer, J., O'Connell,C., Quon, D., Sim, D. K. & Efstratiadis, A. (1978) Cell 15,687-701.

13. Harper, M. E. & Saunders, G. F. (1981) Chromosoma 83,431-439.

14. Pinkel, D., Straume, T. & Gray, J. W. (1986) Proc. Natl. Acad.Sci. USA 83, 2934-2938.

15. Graham, F. L. & Van Der Eb, A. J. (1973) Virology 52,456-467.

16. Hadlaczky, Gy., Sumner, A. T. & Ross, A. (1981) Chromo-soma 81, 557-567.

17. Estivill, X., Farral, M., Scambler, P. J., Bell, G. N., Hawley,K. M. F., Leuch, N. J., Bates, G. P., Kruyer, H. C., Freder-ick, P. A., Stanier, P., Watson, E. K., Williamson, R. &Wainwright, B. J. (1987) Nature (London) 326, 840-845.

18. Lawrence, J. B., Villnave, C. A. & Singer, R. H. (1988) Cell52, 51-61.

19. Korenberg, J. R. & Rykowski, M. C. (1988) Cell 53, 391-400.20. Colb6re-Garapin, F., Horodniceanu, F., Kourilsky, P. & Ga-

rapin, A. C. (1981) J. Mol. Biol. 150, 1-14.21. Leder, P., Tiemeier, D. & Enquist, L. (1977) Science 196,

175-177.22. Chikashige, Y., Kinoshita, N., Nakaseko, Y., Matsumoto, T.,

Murakami, S., Niwa, 0. & Yanagida, N. (1989) Cell 57,739-751.

23. Brinkley, B. R., Zinkowski, R. P., Mollon, L. W., Davis,F. M., Pisegna, M. A., Pershouse, M. & Rao, P. N. (1988)Nature (London) 336, 251-254.

24. Therman, E., Sarto, G. E. & Patau, K. (1974) Am. J. Hum.Genet. 26, 83-92.

25. Earnshaw, W. C. & Migeon, B. R. (1985) Chromosoma 92,290-296.

Proc. Natl. Acad Sci. USA 88 (1991)

Dow

nloa

ded

by g

uest

on

Janu

ary

23, 2

021