Proc. Nati. Acad. Sci. USA Vol. 89, pp. 4923-4927, June 1992 Medical Sciences Molecular cloning of the breakpoints of a complex Philadelphia chromosome translocation: Identification of a repeated region on chromosome 17 (fluorescence in situ hybridization/chronic myelogenous leukemia/primate evolution) TIMOTHY W. MCKEITHAN*tt, LARRY WARSHAWSKYt, RAFAEL ESPINOSA lilt, AND MICHELLE M. LEBEAUt *Department of Pathology and tSection of Hematology/Oncology, Department of Medicine, University of Chicago, Chicago, IL 60637 Communicated by Hewson Swift, February 18, 1992 (received for review August 28, 1991)
ABSTRACT Complex translocations in chronic myeloge- regard, it is notable that many of these breakpoints occur in nous leukemia involve various chromosomes, in addition to bands that are known to contain fragile sites, which are chromosomes 9 and 22, in a nonrandom fashion. We have chromosomal sites at which aberrations, including breaks, analyzed the DNA from leukemia cells characterized by a gaps, and structural rearrangements, preferentially occur in complex translocation, t(9;22;10;17)(q34;qll;p13;q21), by us- cells cultured under certain conditions. ing the techniques of Southern blot hybridization, in situ An alternative possibility is that the sites ofthe breakpoints hybridization, and molecular cloning; one of the breakpoints is in the third chromosome are specific for the t(9;22), perhaps at 17q21, a band that is frequently involved in complex 9;22 as a result of sequence similarity to either ofthe genes located translocations. All of the breakpoint junctions and the corre- at the breakpoints on chromosomes 9 or 22 (ABL and BCR, sponding normal sequences from the four involved chromo- respectively). Molecular analysis of standard two-way 9;22 somes have been molecularly cloned. Restriction mapping is translocations has, so far, failed to show sequence similarities consistent with a simple concerted exchange of chromosomal between the regions ofABL and BCR at which the two breaks material among the four chromosomes, except that additional occur although Alu repetitive sequences are sometimes in- changes appeared to have occurred within the chromosome 17 volved (9). sequences. The cloned sequences on chromosome 17 at band To gain some insight into the mechanisms responsible for q21 were found to be repeated in normal cells. By fluorescence the nonrandom involvement of additional chromosomes in in situ hybridization, a strong signal is seen at 17q21, but a complex 9;22 translocations, we have begun to analyze such weaker signal is also present at 17q23. By comparison with cases at the molecular level. We identified a patient with other primate species, an inversion in chromosome 17 during CML who had a complex four-way translocation, evolution appears to be responsible for the splitting of the t(9;22;10;17)(q34;qll;pl3;q21), which is illustrated in Fig. 1. cluster of repeat units in normal human cells. Chromosome band 17q21 is one of the bands most frequently involved in complex 9;22 translocations. In addition, the Chronic myelogenous leukemia (CML) is characterized cy- band has been shown to contain a common fragile site (10). togenetically by a reciprocal translocation involving chromo- Here we report the molecular cloning of the breakpoints of somes 9 and 22, t(9;22)(q34;qll), giving rise to the Philadel- this translocation and the corresponding normal sequences. phia chromosome, which contains a chimeric gene composed of the 5' portion of the BCR gene on chromosome 22 and the 3' portion of the ABL gene on chromosome 9 (1). The MATERIALS AND METHODS BCR-ABL gene encodes a fusion protein with enhanced Patient Sample and Cell Lines. The leukemia cells from tyrosine kinase activity (2). This protein appears to play a role patient 1 were characterized by a t(9;22;10;17)(q34;qll; in the transformation of hematopoietic stem cells (3, 4). pl3;q21) (case number 21 in ref. 11). Lymphoblastoid cell In most CML patients, the Philadelphia chromosome re- lines from orangutan and baboon were obtained from the sults from a standard t(9;22). The remaining 4-8% (5, 6) of American Type Culture Collection [PUTI line (CRL 1850) Philadelphia chromosome-positive patients have complex and 26 CB-1 line (CRL 1495), respectively]. translocations involving three or more chromosomes, two of Southern Blot Analysis. DNA extraction and Southern blot which are chromosomes 9 and 22. The karyotypes of >300 analysis were performed as described (12). The 3' breakpoint patients with CML characterized by a complex translocation cluster region (bcr) probe [a 1.2-kilobase-pair (kb) HindIIl- have been described (5-8). With the exception of the Y EcoRI probe containing bcr exon 4 (13)] was provided by chromosome, every chromosome has been found to be Carol Westbrook (University of Chicago). Additional details involved in these variant translocations; however, certain are described in Fig. 2. For Southern blot hybridization to chromosomal bands show a significantly higher number of one of the probes (Fig. 3a, probe 4), the oligolabeled probe breaks than that expected if the breaks were randomly was preannealed with human Cot-1 DNA (GIBCO/BRL), by distributed along all of the chromosomes. using the method recommended by the supplier. No clinical differences have been noted between patients Molecular Cloning. Three bacteriophage libraries were who have the standard two-way 9;22 translocations and those prepared. The first, used for cloning the 9;22 and the 10;22 who have complex translocations (5). This fact suggests that junctions, was obtained by complete HindIII digestion of the break in the third chromosome offers no additional DNA extracted from the leukemia cells of patient 1. The selective advantage to the malignant cell. Thus, the involve- DNA was size-fractionated on a low-gelling-temperature ment of the third chromosome may reflect an intrinsic agarose gel (SeaKem LGT agarose, FMC), and DNA was tendency to break that is unaffected by selection. In this Abbreviations: CML, chronic myelogenous leukemia; bcr, break- The publication costs of this article were defrayed in part by page charge point cluster region on chromosome 22; FISH, fluorescence in situ payment. This article must therefore be hereby marked "advertisement" hybridization. in accordance with 18 U.S.C. §1734 solely to indicate this fact. 1To whom reprint requests should be addressed.
4923 Downloaded by guest on September 24, 2021 4924 Medical Sciences: McKeithan et al. Proc. Natl. Acad Sci. USA 89 (1992) RESULTS 0E 9 der(1 7) der(9) Cloning the Breakpoint Junctions and the Corresponding - _ 17 Normal Sequences. In the four-way translocation ofpatient 1,
K A distal 9q moved to chromosome 22; distal 22q moved to chromosome 10; distal 10p moved to chromosome 17; and mu distal 17q moved to chromosome 9 (Fig. 1). If the DNA rearrangements corresponded to the karyotypic changes in a straightforward manner, complete molecular characteriza- tion would require the cloning of four breakpoint junctions 22 ! der(22) rib and the corresponding normal sequences on chromosomes 9, 10 (Ph') 10, 17, and 22. The t(9;22) breakpoints on chromosome 9 are der(l0) quite variable (17), but the breaks on chromosome 22 occur almost exclusively within a 5.8-kb segment of the BCR gene FiG. 1. Model of the four-way translocation t(9;22;10;17)(q34; known as the bcr. Thus, the strategy was to begin at chro- qll;p13;q21) of patient 1. On the left, small arrows indicate the mosome 22 and, by chromosomal walking, to clone sequen- breakpoints. The relative movements of the distal chromosomal tially the various breakpointjunctions and normal sequences. fragments before rejoining are also indicated. Ph', Philadelphia Cloning of the 9;22 and 10;22 junctions was followed by chromosome. cloning of the normal chromosome 9 and 10 sequences isolated from bands containing fragments of 9-21 kb by spanning the chromosomal breakpoints. These sequences in successive phenol and chloroform extractions. Bacterio- turn allowed the 9;17 and 10;17 junctions to be cloned, phage ADASH DNA (Stratagene) was digested with Xho I followed, finally, by isolation of the normal chromosome 17 and HindIII and differentially precipitated. Ligated vector sequences. A Southern blot prepared from the patient's leukemic cells and human DNA was packaged using Gigapack Gold II (Stratagene). For the remaining cloning steps, "random" was hybridized to a probe containing bcr exon 4. Two libraries were prepared from patient 1 and from tumor-cell rearranged bands were seen in addition to a germ-line-sized DNA from a patient with a non-Hodgkin lymphoma who did band (Fig. 2 a and b). Both bands shown in Fig. 2a were not have abnormalities of the relevant chromosomal bands. cloned, beginning the sequential process of cloning the var- The libraries were prepared similarly, except that size selec- ious breakpoint junctions and normal sequences. At each tion (16-23 kb) followed partial digestion with Sau3A, and the step, the chromosomal origin ofthe various molecular clones vector was digested with Sac I and BamHI; the library from was confirmed by in situ hybridization to metaphase chro- mosomes using either biotin-labeled or 3H-labeled probes patient 1 was prepared in the vector ADASH II (Stratagene). Library screening and bacteriophage mapping were as de- (data not shown). scribed (12). In general, for each cloning step, repeat-free Fig. 3 illustrates the restriction maps ofthe cloned regions. DNA Hybridization to Southern blots prepared from patient and fragments were identified by hybridization ofplacental control placental DNA samples was used to confirm that the to Southern blots ofdigested clones. Subclones were made in sizes predicted for the normal and rearranged fragments, on Bluescript vectors (Stratagene). To isolate clones homolo- the basis ofthe restriction maps, were correct. For example, gous to the chromosome 17 sequences adjacent to the 9;17 in Fig. 2c, hybridization of a chromosome 10 probe (Fig. 3f, junction present on the der(9) chromosome, a fragment probe 2) spanning the breakpoint shows two rearranged containing the entire chromosome 17 portion of one of the bands. One represents the 10;22junction and comigrates with der(9) clones was radioactively labeled, and repetitive se- quences were blocked by preannealing to sonicated human a b c d e placental DNA (14).
In Situ Hybridization. For in situ chromosomal hybridiza- chr.22 chr.1 ° 17 chr b (10:17) (9:1_-2 Kb tion, human metaphase cells were prepared from phytohem- - agglutinin-stimulated peripheral blood lymphocytes of nor- _I 231 mal individuals, from bone marrow cells from patient 1 that - t _ were cultured without mitogens, or from orangutan or ba- 494.. boon lymphoblastoid cell lines. Radiolabeled probe was prepared by nick-translation of the entire plasmid with all 6.6 four 3H-labeled deoxyribonucleoside triphosphates to a spe- cific activity of 5 x 107 to 1 x 108 dpm/,ug. In situ hybrid- ization of3H-labeled probes was performed as described (15). Metaphase cells were hybridized at 5, 10, 20, or40 ng ofprobe C P C P C P C P C P per ml of hybridization mixture (15). Autoradiographs were HindIII Bgn exposed for 11 days. The procedure used for fluorescence in FiG. 2. Hybridization of DNA from the leukemia cells of patient situ hybridization (FISH) is as described (16). Biotin-labeled 1 (lanes P) and control placental DNA (lanes C). A single Southern probes were prepared by nick-translation with biotin-labeled blot was hybridized to probes from chromosomes (chr) 22, 10, and dUTP (Bio-11-dUTP, Enzo Diagnostics). Hybridized probes 17. The positions ofthe probes used are shown in Fig. 3 as bars below were detected with fluorescein-conjugated avidin (Vector the restriction maps. Arrowheads indicate the bands resulting from Laboratories). Most of the probes used were whole bacteri- a rearrangement. Lanes: a, a HindIII digest; b-e, a Bgl II digest; a ophage clones. For the in situ hybridizations shown in Fig. 4, and b, hybridization to probe 1 (Fig. 3d), a 1.4-kb HindIII-EcoRI the entire insert from each of two chromosome 17 bacterio- fragment containing bcr exon 4; c, hybridization to probe 2 (Fig. 3/), a 0.4-kb Rsa 10; d, hybrid- phage clones showing relatively little overlap was subcloned I-BamHI fragment from chromosome ization to probe 3 (Fig. 3g), a 1.2-kb Rsa I probe from chromosome a was into Bluescript, and mixture of the two plasmids used, 17 adjoining the 10;17junction (in shorter exposures the upper band comprising -33 kb of distinct human sequences. is seen as a doublet, so that a total offive bands can be distinguished Somatic Cell Hybrid Analysis. Southern blots prepared in the control); e, hybridization to probe 4 (Fig. 3a), a 6-kb fragment from HindIII digests of 18 hamster x human somatic cell containing chromosome 17 sequences adjacent to the 9;17 junction hybrids (Bios, New Haven, CT) were hybridized to probe 3. and a minute amount of chromosome 9. Downloaded by guest on September 24, 2021 Medical Sciences: McKeithan et al. Proc. Natl. Acad. Sci. USA 89 (1992) 4925
a junction 9& 17 4 -
b chr. 9 11
C junction 22 & 9 4
d chr. 22 + FIG. 3. (a-h) Restriction maps ofbacteriophage e junction clones containing the breakpoint junctions of the 10 & 22 t(9;22;10;17) and the corresponding normal se- quences. In some cases, the maps are composites made from overlapping clones. The four chromo- f chr. 10 2 H somes are distinguished by shading; vertical bars indicate uncertainty in the position of the break- g junction STT m T. ; fl >X?vA _7 T~v Y?7es v Ts pointjunctions. Restriction sites are symbolized as 17 & 10 indicated at the bottom. The order of cloning is shown on the left. The probes used for Southern V h chr17 T~§ 4VT blot analysis (Fig. 2) are shown as bars below the line. The single-copy sequences from the 10;17 junction were not precisely localized. The restric- T Smal Y BgAI t SaA 10 kb tion pattern of a single representative clone from Sinai ? EcoRi ... chromosome 17 is illustrated; some clones have T BanH FY Sad T HindIll slightly different maps. a rearranged band in Fig. 2b, whereas the other represents the Rearrangements within the chromosome 17 sequences are 10;17 junction and hybridizes with the rearranged band in presumably responsible for the restriction pattern seen in the Fig. 2d, which shows the hybridization pattern using a 9;17junction clones, which is different from that ofany ofthe low-copy fragment (Fig. 3g, probe 3) from the chromosome normal chromosome 17 clones isolated. The restriction maps 17 portion of a der(17) clone. In each case, the sizes of the ofthe various chromosome 17 clones are very similar but not hybridizing bands matched that expected from the maps of all are completely consistent with each other. It is likely that the bacteriophage clones. the differences are due to their being derived from different Identification of a Repeated Region on Chromosome 17. copies ofthe DNA repeats. The similarity in restriction maps Surprisingly, probe 3, from chromosome 17, hybridized to suggests that the copies have diverged relatively little in multiple bands on the various digests-five bands were seen sequence, as supported also by the strong hybridization of in the Bgl II digest; fewer bands were seen with other digests, the various bands to the chromosome 17 probes (Fig. 2). The presumably due to comigration of bands. It is likely that this map of a single representative chromosome 17 clone is pattern ofhybridization reflects amplification ofthe sequence illustrated in Fig. 3; it is not known whether this clone is during evolution. Upon Southern blot analysis using a panel derived from the specific repeat affected by the translocation. of somatic cell hybrids (data not shown), only those lanes A single restriction site difference was noted between the prepared from lines containing human chromosome 17 relevant portions of this clone and the 10;17 clones (Fig. 3 g showed hybridization to probe 3. Thus, all of the repeats and h). appear to lie on chromosome 17. These results are consistent Initial Southern blot analysis of 11 DNA samples (from with those obtained using FISH, described below. normal individuals and CML patients without evidence for Chromosome 17 sequences were also cloned in the oppo- involvement of chromosome 17) showed three restriction site "direction." The map of the chromosome 17 sequences fragment length polymorphisms (unpublished results). It is present near the 9;17 junction does not match that expected not known whether these result from simple base-pair dif- from the map ofthe normal chromosome 17. In particular, the ferences or, more interestingly, from differences in the num- maps showed different patterns with Nco I. The apparently ber or organization of the repeats. The size of the repeat unit appropriate position of the Sac I sites (Fig. 3) was fortuitous has not been determined. Mapping of the clones over 25 kb since, as determined later, the bands of similar size from the does not give any indication of a repetitive pattern; presum- der(9) and the normal chromosome 17 clones did not cross- ably, then, the repeat unit is greater than 20 kb. hybridize. Because of the differences in the map, we sus- FISH to normal metaphase chromosomes using DNA from pected that a deletion had occurred within chromosome 17 a chromosome 17 clone gave a strong signal on chromosome sequences. However, the maps of the clones isolated using 17 at band q21 (Fig. 4 a and b). Unexpectedly, a fainter signal probe 4 (Fig. 3a) from the 9;17junction overlapped with those was also seen on chromosome 17 at bands q23-q24. Specific of the earlier set of chromosome 17 clones, and most of the hybridization to other chromosomes was not observed. The bands hybridizing to probe 4 and to probe 3 (originating results suggest that multiple copies of sequences homologous adjacent to the 10;17junction) were the same (Fig. 2 d and e). to the probe are present at 17q21 and that one or a few copies Also unexpectedly, probe 4 hybridized to the clones con- are present at 17q23-q24. taining the 10;17 junction. There are two possible explana- In situ hybridization to interphase nuclei (Fig. 4c) showed tions for this result. Concomitant with the translocation, clusters of multiple hybridization signals. In some cells, four there may have been duplication of chromosome 17 se- clusters were observed, presumably corresponding to the quences at the two breakpoint junctions; alternatively, if the two hybridizing regions on each of the two chromosome 17 repeats are in tandem, almost all of one copy may have been homologs. The number of hybridization signals has not yet deleted, leaving homologous sequences adjacent to the two been accurately determined, due to the presence of hybrid- chromosomal junctions. ization signals at different levels of focus. Downloaded by guest on September 24, 2021 4926 Medical Sciences: McKeithan et al. Proc. Nad. Acad. Sci. USA 89 (1992)
FIG. 4. Localization ofchromosome 17 sequences by FISH of a biotin-labeled chromosome 17 probe to metaphase cells or interphase cells. In a, c, and d, the metaphase cells counterstained with 4',6-diamidino-2-phenylindole dihydrochloride are to the left and detection ofthe probe with fluorescein isothiocyanate-conjugated avidin is to the right. Arrows and arrowheads in panels on the left show the positions at which significant signals are seen in the corresponding panels to the right. (a) Normal human metaphase cell. Strong signals can be seen at 17q21 (arrows) and weaker signals can be seen distally at 17q23 (arrowheads). (Inset) Partial karyotype illustrating the hybridization signals in another metaphase cell. (b) Normal human interphase cell from phytohemagglutinin-stimulated lymphocytes, illustrating clusters of multiple signals (arrows). (c) Baboon metaphase cell. (d) Orangutan metaphase cell. In c and d, single strong signals (arrows) are seen distally at the band corresponding to the human 17q23. The karyotypes of both of these cell lines were normal for the respective species. The results of recent studies suggest that multiple probes undergone rearrangements in addition to this paracentric can be resolved in interphase nuclei by FISH only if they are inversion. separated by at least 25 kb (18, 19). Given the size ofthe probe The results of the in situ hybridization studies could be (%33 kb), the observation of multiple well-separated signals explained if one of the breaks in the paracentric inversion in interphase nuclei suggests that the size of the repeat unit occurred within the cluster ofrepeats that we have identified. is at least 50 kb and probably significantly larger. To study this further, metaphase chromosomes of lympho- Involvement of Chromosome 17 Sequences in a Paracentric blastoid cell lines from orangutan and baboon were analyzed Chromosomal Inversion During Primate Evolution. Compar- (Fig. 4 c and d). In both cases, hybridization of a chromo- ison ofhigh-resolution karyotypes from humans and the great some 17 clone was detected at a position corresponding to the apes has shown that a moderate number of inversions and distal weaker signal on human chromosomes (q23-q24); in translocations are responsible for most of the differences in these species no secondary signal was identified. These chromosomal organization among these species (20). For results suggest that the repeated sequences we have identi- most of the chromosomes, it was possible to deduce the fied were initially present at the position corresponding to the structure of the ancestral chromosome and the nature of the human band 17q23.1. The inversion that occurred in a rearrangements that occurred in each evolutionary line. In common progenitor of man, chimpanzee, and gorilla had a particular, the results of comparison of the chromosome 17 breakpoint near the telomeric end of the cluster and resulted homologs in the human, chimpanzee, gorilla, orangutan, and in movement of most, but not all, of the cluster to band baboon were compatible with a model in which the baboon q21.31. chromosome resembles the ancestral chromosome 17 whereas different inversions occurred in an orangutan pro- genitor and in a common progenitor ofman, chimpanzee, and DISCUSSION gorilla. The orangutan chromosome resulted from a pericen- Two models have been proposed for complex translocations. tric inversion whereas the human chromosome resulted from In the first model, multiple chromosomal breaks occur within a paracentric inversion between bands q21.31 and q23.1. The a cell at the same time, and the broken chromosomes are then chromosome 17 homologs in chimpanzees and gorillas have improperly rejoined (21). The second postulates that the Downloaded by guest on September 24, 2021 Medical Sciences: McKeithan et al. Proc. Natl. Acad. Sci. USA 89 (1992) 4927
complex translocation results from sequential two-way trans- We appreciate the valuable comments of Drs. Janet Rowley, locations that occur in different cell generations (22). To Feyruz Rassool, and Manuel Diaz and the able technical assistance explain the relatively high frequency of complex transloca- of Elizabeth van Melle and Steven Marsten. This research was supported by a grant from the National Institutes of Health tions, the latter model seems to require some form of insta- (CA41644) to M.M.L. and T.W.M., who are both Scholars of the bility of one of the junctions formed during the initial trans- Leukemia Society of America. location. This is not implausible, given the fact that certain DNA rearrangements-namely, insertion of foreign DNA- 1. Stam, K., Heisterkamp, N., Grosveld, G., de Klein, A., have been reported to be associated occasionally with an Verma, R. S., Coleman, M., Dosik, H. & Groffen, J. (1985) N. induced chromosomal instability (23). Engl. J. Med. 313, 1429-1433. The present data are more consistent with the first model: 2. Konopka, J. B., Watanabe, S. M. & Witte, 0. N. (1984) Cell this translocation appears to result from a concerted effec- 37, 1035-1042. tively simultaneous exchange of material among the four 3. McLaughlin, J., Chianese, E. & Witte, 0. N. (1987) Proc. Nat!. chromosomes. With the exception ofthat of Acad. Sci. USA 84, 6558-6562. chromosome 17, 4. Daley, G. Q., Van Etten, R. A. & Baltimore, D. (1990) Science the restriction maps of the various normal and rearranged 247, 824-830. clones can be lined up in perfect register. Internal rearrange- 5. De Braekeleer, M. (1987) Cytogenet. Cell Genet. 44, 215-222. ments appear to have occurred within chromosome 17 se- 6. Rowley, J. D. & Testa, J. R. (1982) Adv. Cancer Res. 36, quences. 103-148 1982. The alternative model would require three consecutive 7. Heim, S., Billstrom, R., Kristoffersson, U., Mandahl, N., translocations. For the maps to match as perfectly as they do, Strombeck, B. & Mitelman, F. (1985) Cancer Genet. Cytoge- the successive breaks, at least on chromosomes 9, 10, and 22, net. 18, 215-227. must have occurred within a few hundred nucleotides ofeach 8. Bernstein, R. (1988) Semin. Hematol. 25, 20-34. other. There is no known or speculative mechanism apparent 9. de Klein, A., van Agthoven, T., Groffen, C., Heisterkamp, N., Groffen, J. & Grosveld, G. (1986) Nucleic Acids Res. 14, to us that could easily explain such a precise molecular 7071-7082. "memory. " 10. Yunis, J. J., Soreng, A. L. & Bowe, A. E. (1987) Oncogene 1, An argument against the first model is that it may seem 59-69. unlikely that a single cell would have four double-stranded 11. Dub6, I., Dixon, J., Beckett, T., Grossman, A., Weinstein, M., DNA breaks at the same time. There are a number of ways, Benn, P., McKeithan, T., Norman, C. & Pinkerton, P. (1989) however, in which a cell could sustain nearly lethal damage Genes Chromosomes Cancer 1, 106-111. that might result in multiple DNA breaks. A particularly 12. McKeithan, T. W., Ohno, H. & Diaz, M. 0. (1990) Genes appealing agent is ionizing radiation, some forms ofwhich are Chromosomes Cancer 1, 247-255. capable of producing large quantities of free radicals within 13. Rubin, C. M., Westbrook, C. A., Smith, S. D., Hooberman, A. L., Colowich, A., Geiger, T. A., Steele, M. M. & Rowley, a relatively small volume that could cause multiple DNA J. D. (1987) in Recent Advances in Leukemia and Lymphoma breaks (24). In this regard, it is intriguing that radiation is (Liss, New York), pp. 125-131. known to be a predisposing factor for various forms of 14. Litt, M. & White, R. L. (1985) Proc. Nat!. Acad. Sci. USA 82, leukemia, including CML (25). 6206-6210. As mentioned above, chromosome 17 at band q21 is one of 15. Le Beau, M. M., Westbrook, C. A., Diaz, M. 0. & Rowley, the sites most frequently involved in complex 9;22 translo- J. D. (1984) Nature (London) 312, 70-71. cations. It is intriguing that a cluster of repeats in this band 16. Rowley, J. D., Diaz, M. O., Espinosa, R., Patel, Y. D., van appear to have been involved not only in the particular Melle, E., Ziemin, S., Taillon-Miller, P., Lichter, P., Evans, translocation here but also in an G. A., Kersey, J. H., Ward, D. C., Domer, P. H. & Le Beau, complex analyzed inversion M. M. (1990) Proc. Natl. Acad. Sci. USA 87, 9358-9362. that occurred during primate evolution. This result raises the 17. Westbrook, C. A. (1991) in Chronic Myelogenous Leukemia: possibility that this region is unusually prone to chromosome Molecular Approaches to Research and Therapy, eds. Deisse- rearrangement. roth, A. & Arlinghaus, R. B. (Dekker, New York), pp. 179- It has been suggested that chromosomal rearrangements 207. during evolution may often involve fragile sites (26). A fragile 18. Lawrence, J. B., Singer, R. H. & McNeil, J. A. (1990) Science site at 17q21 has been identified in human, gorilla, and 249, 928-932. chimpanzee (10, 27); however, the human fragile site is 19. Trask, B. J., Massa, H., Kenrick, S. & Gitschier, J. (1991) Am. expressed at relatively low levels and only with certain J. Hum. Genet. 48, 1-15. agents. This has made it difficult to determine whether the 20. Yunis, J. J. & Prakash, 0. (1982) Science 215, 1525-1530. 21. Morris, C., Kennedy, M., Heisterkamp, N., Columbano- sequences we have identified at 17q21 correspond to the Green, L., Romeril, K., Romain, D., Hallman, D., Testa, J., fragile site within that band. Groffen, J. & Fitzgerald, P. (1991) Genes Chromosomes Cancer It is often assumed that chromosomal breaks occur ran- 3, 263-271. domly and that the pattern ofrearrangements actually seen in 22. Ishihara, T. & Minamihisamatsu, M. (1988) Cancer Genet. cancer cells results solely from selection for those rare Cytogenet. 32, 75-92. changes that give the cell a proliferative advantage. An 23. Murnane, J. P. & Young, B. R. (1989) Gene 84, 201-205. alternative possibility is that the frequency of the various 24. Holmberg, M. (1978) in Mutagen-induced Chromosome Dam- chromosomal abnormalities seen in cancer is influenced by age in Man, eds. Evans, H. J. & Lloyd, D. C. (University variability in the likelihood that the Press, Edinburgh), pp. 14-21. particular rearrangement 25. Mole, R. H. (1990) in Leukemia, eds. Henderson, E. S. & would take place. This alternative possibility is supported by Lister, T. A. (Saunders, Philadelphia), pp. 253-269. the nonrandom involvement of chromosomal bands in com- 26. Miro, R., Clemente, I. C., Fuster, C. & Egozcue, J. (1987) plex rearrangements. Determining the special properties of Hum. Genet. 75, 345-349. such loci should aid in elucidating the mechanism of chro- 27. Smeets, D. F. C. M. & van de Klundert, F. A. J. M. (1990) mosomal translocations. Cytogenet. Cell Genet. 53, 8-14. Downloaded by guest on September 24, 2021