GENES, CHROMOSOMES 8 CANCER 4~241-249(1992)

Chromosomal Sublocalization of the 2; I3 Translocation Breakpoint in Alveolar Rhabdomyosarcoma David N. Shapiro, Marc B. Valentine, Jack E. Sublett, Anne E. Sinclair, Alan M. Tereba, Hans Scheffer, Charles H.C.M. Buys, and A. Thomas Look Departments of Hematology-Oncology (D.N.S., M.B.V., J.E.S., A.E.S., A.T.L.) and Virology and Molecular Biology (A.M.T.), St. Jude Children’s Research Hospital, the Department of Pediatrics, University of Tennessee Memphis College of Medicine (D.N.S., A.T.L.). Memphis, Tennessee, and the Department of Human Genetics, State University of Groningen, Groningen, The Netherlands (H.S., C.H.C.M.B.)

A characteristic balanced reciprocal chromosomal translocation [t(2; I3)(q35;q I4)] has been identified in more than 50% of alveolar rhabdomyosarcomas.As the first step in characterization of the genes involved in this translocation, we constructed somatic cell hybrids that retained either the derivative chromosome 2 or the derivative chromosome I 3 without a normal chromosome I 3 homologue. Ten linked DNA probes known to be located within bands 13q 13-q I 4 were mapped relative to the breakpoint on chromosome 13, allowing localization of the breakpoint region between two loci separated by 5.5 cM. A long-range restriction map extending approximately 2,300 kb around these loci failed to provide evidence of rearrangement. Additionally, we confirmed that the FMS-like tyrosine kinase gene (FLT), previously localized to 13q I 2 by in situ hybridization, is located proximal to the breakpoint, and we demonstrated that FLT is not a target for disruption by this tumor-specific translocation. Genes Chrom Cancer 4:24 1-249 ( I 992). IC 1992 WiIey-Lisr. tnc. INTRODUCTION

Recurrent, nonrandom chromosomal translocations have been reported in certain solid tumors, including Ewing’s sarcoma, synovial sarcoma, leiomyoma, liposarcoma, and rhabdomyosarcoma (Aurias et al., 1983; Douglass et al., 1987;Reeves et al., 1989; Eneroth et al., 1990; Pandis et al., 1990). However, the identification of the involved cognate genes has proved difficult because of the absence of known loci near enough to specific solid tumor chromosomal breakpoints to allow their direct isolation. A characteristic t(2;13)(q35;q14)has been noted in the tumor cells of approximately one-ha1f of the patients with alveolar rhabdomyosarcoma, a malignant tumor of skeletal muscle with characteristic histologic and clinical features (Douglass et al., 1987; Rowe et al., 1987; WangWuu et al., 1988). The t(2;13) usually occurs as a balanced reciprocal translocation, although the derivative chromosome 13 is reportedly absent in approximately 25% of the cases analyzed. The occasional apparent absence of the derivative chromosome 13 suggests that a gene critical for cellular transformation or tumor progression is on the derivative chromosome 2. As an initial step toward characterizationand isolation of the genes involved in t(2;13), we previously examined the breakpoint region on chromosome 13 using in situ hybridization (Valentine et al., 1989).The results demonstrated that the translocation breakpoint is delimited by two anonymous DNA segments separated by a mean recombination distance of at 0 1992 WILEY-LISS, INC.

least 32 CM (Bowcock et al., 1990). In this study, we have used somatic cell hybrids that have retained either the derivative 2 or derivative 13 without a normal 13 homologue so that we could significantly refine the boundaries of the breakpoint region. Additionally, we have constructed a long-range physical map around the locus distal to the breakpoint, and we show that neither the proximal nor the distal locus is close enough to the breakpoint to allow direct identification of the genes involved in this translocation. Finally, we show that the FLT tyrosine kinase, a member of the FMS oncogene family previously shown by in situ hybridization to map to 13q12, is proximal to the translocation breakpoint and is not disrupted by the t(2;13) in alveolar rhabdomyosarcoma. M A T E R I A L S AND M E T H O D S

Cell Lines

The rhabdomyosarcoma cell line Rh30 was derived from the bone marrow of a patient with metastatic alveolar rhabdomyosarcoma. Previous karyotypic analysis demonstrated that this cell line has the characteristic reciprocal translocation between the long arms of chromosomes2 and 13[t(2;13)(q35;q14)](Douglass et a]., 1987).The embryonal rhabdomyosarcoma cell line BG was derived from the bone marrow of a Received June 11, 1991; accepted October 2, 1991 Address reprint requests to Dr. David N. Shapiro. Department of Hematology-Oncology,St. Jude Children’sResearch Ilospital, 332 N. Lauderdale, Memphis, TN 38105.

242

SHAPJRO ET A L

patient with metastatic rhabdomyosarcoma without the t(2;13). The hypoxanthine-guanine phosphoribosyltransferase-deficientChinese hamster cell line E36, derived from the V79 lung fibroblast line, has been described elsewhere (Gillin et al., 1972).All cell lines were maintained as adherent cultures in RPMI-1640 medium (Whittaker Bioproducts, Walkersville, MD) supplemented with 10% fetal bovine serum (Whittaker), 2 mM glutamine, and 50 pg/ml gentamicin sulfate (Whittaker) at 37°C in a humidified 59’0 C02 atmosphere. Where indicated, rhabdomyosarcoma cell lines were incubated with 3 pM 5-azacytidine (Sigma Chemical Co., St. Louis, MO) and fed weekly over 14 days (approximately 4 doublings) before harvesting. Hamster-Human Somatic Cell Hybrids

Somatic cell hybrids were produced by fusion of E36 hamster cells to the human rhabdomyosarcoma cell line Rh30, with polyethylene glycol used as described (Hunt et al., 1990).Hybrid cells were selected for and then propagated in complete medium containing hypoxanthine (100 pM), aminopterin (0.4 pM), thymidine (16 pM), and ouabain (1 pM). Cytogenetic Analysis

Somatic cell hybrid clones were incubated in standard growth medium supplemented with 0.1 pg/ml of Colcemid for 2 h. Metaphase cells were prepared and, after methano1:acetic acid (3:1 vol/vol) fixation, stained with Giemsa and trypsin-Giemsa (Douglass et al., 1987). D N A Isolation, Digestion, and Gel Electrophoresis

For conventional electrophoresis, DNA was extracted with phenol followed by chloroform from the indicated somatic cell hybrid lines (Wallace, 1987). DNA samples were digested with 3-10 U of restriction endonuclease/pg for 8-12 h and electrophoresed in 0.8% agarose gels in Tris-acetate buffer. For contour-clamped homogeneous field electrophoresis (CHEF),cultured cells were embedded in 1YO low melt agarose (Marine Colloids, Rockland, ME) and DNA prepared from the embedded cells as described (Chu et al., 1986). One-half of a gel slice containing approximately 3 pg of embedded DNA was incubated with 30 U of restriction endonuclease 3 h under conditions appropriate for each enzyme. This was followed by the exchange of fresh buffer and incubation for an additional 3 h with 30 U of fresh enzyme. Except for Sfl, all of the enzymes used were known to be sensitive to the presence of 5-methylcytosine residues in the recognition sequence (McClelland and Nelson, 1986).

Digested gel slabs were loaded onto a 1YOagarose gel and electrophoresed with a CHEF-DR I1 apparatus (Biorad, Richmond, CA). For resolution of restriction fragments larger than 1 megabase, samples were run at 15°C in 0.5X TBE for 120 h at 85 V with linear ramping from 120 to 1200 sec. For resolution of fragments smaller than 1 megabase, electrophoresis was performed for 36 h at 150 V with linear ramping from 120 to 180 sec. Molecular weights were estimated by comparison with ethidium bromide-stained lambda concatamers or Sacchuromyces cerevisiae chromosomes which were included with all gels. Estimates of fragment lengths have an error of approximately 10%. Both Southern and CHEF gels were exposed for 150 sec to a 254 nM W light source before denaturation, neutralization, and standard capillary transfer to nylon membranes (Duralon-W, Stratagene, La Jolla, CA). After transfer, DNA was cross-linked to the membranes by exposure for 20 sec to a 330 nM W light source. Membranes were hybridized for 18 h at 42°C in a solution composed of 50% formamide, 5X SSC, 5X Denhardt’s solution, 10% dextran sulfate, 0.5% SDS, and 100 pg/ml of sheared salmon sperm DNA. Radiolabeled DNA probes were added at a concentration of 1 x l o 6 cpm/ml. Final post-hybridization washes were performed in 0.1% SSC,0.1% SDS at 55°C before autoradiography. D N A Probe Preparation

The plasmids containing cDNA or genomic inserts for the genes examined in this study are shown in Table 1. Isolated inserts were labeled by the random hexadeoxyribonucleotide primer technique to a specific activity of 1-2 x lo9 cpm/pg (Feinberg and Vogelstein, 1983). The probe designated TUBB, consisting of a 240 bp insert from the 3’-untranslated region of the 21p-tubulin pseudogene, has been shown to hybridize to sequences dispersed on chromosomes 6,8, and 13(Floyd-Smithet al., 1986).Although the 2.0 kb Hind111 fragment identified by TUBB is specific for chromosome 13 sequences (termed TUBBPZ)and was useful for Southern blot analysis of somatic cell hybrids, cross-hybridization with sequences on chromosomes 6 and 8 made it unsuitable for CHEF analysis. To isolate a chromosome 13-specific TUBBPZ probe, we constructed a genomic library in the lambda phage vector EMBL3 by partial MboI digestion of DNA from the human-hamster somatic cell hybrid ICD. This hybrid contains 13pter-q14 as the only human genetic material (Buys et al., 1985). A unique 900 bp SalI fragment (TUBBPZP),specific for chromosome 13 and located 750 bp from the 2.0 kb

ALVEOLAR RHABDOMYOSARCOMA BREAKPOINT

TABLE I. C h r o m o s o m e 13 Probes Locus

Probe

Reference

D 1356 D I3533

pHUlO p C R I359 p7F I 2 p2 I P-3’UT pTH I62 pL28-5 pG24E2.4 pG 14E I .9 p7D2 pESD- 14. I. I p4.7R p4.4 p3-9

Dryja e t al., I984 B o w c o c k e t al., I990 Cavenee e t al., 1984 Floyd-Smith e t al., I986 Leppert e t al., I987 Bowcock e t al., I987 Scheffer e t al., I986 Scheffer e t al., 1986 Cavenee e t al.. I984 Squire e t al., I986 Friend e t al., I986 Matsushima e t al., I987 Shibuya e t al., I990

D13SI TUBBPZ D I3537 D I3S62 D13S21 D I3522 D13SlO ESD RBI FLT

Hind111 pseudogene sequence, was isolated and used for all CHEF analyses. Extension of the long-range physical map around the TUBBPZ locus was also facilitated by the characterization of a TUBBP2-Pcontaining cosmid clone from a pJB8 human placental cosmid library. A 2.5 kb EcoRI fragment (TUBBP2-C) was isolated from this cosmid and used for these studies.

243

nated E8), were chosen for further analysis. Linked DNA probes known to map to 13q13-ql4 (Table 1) were hybridized to genomic DNA from hybrid clones D1 and E8. Because hybrid D1 contains a der(l3) without a normal chromosome 13, probes mapping centromeric to the t(2;13) breakpoint on chromosome 13 hybridize with D1, whereas those mapping distally fail to hybridize. On the other hand, DNA from hybrid E8, which contains a der(2) without a normal chromosome 13, hybridizes with probes mapping telomeric to the breakpoint on chromosome 13. The results of these studies (summarized in Figs. 1 and 2) demonstrate that D13S6, D13S33, and D13S1 lie proximal to the breakpoint, whereas ESD, D13S10, D13S21, D13S22, D13S62, D13S37, and TUBBPZ lie distal to it. Available linkage data suggest that D13S1 is the proximal probe closest to the breakpoint and that TUBBPZ is the closest distal probe (Fig. 2). The mean, sex-averaged recombination distance between these two probes is 5.5 cM, equivalent to a physical distance of approximately 6 megabases; thus this significantly refines the previous localization provided by in situ hybridization.

RESULTS Characterization of Somatic Cell Hybrids

We initially sought to sublocalize the breakpoint region on chromosome 13 by constructing and analyzing somatic cell hybrids that contained the relevant derivative chromosomes. In previous in situ hybridization studies, the breakpoint region on chromosome 13 had been localized between the anonymous DNA probes D13S1 (7F12) and D13S10 (7D2) (Valentine et al., 1989). Linkage data suggest that these loci are separated by a mean recombination distance of at least 32 cM, and that D13S1 is proximal to the translocation breakpoint whereas D13S10 lies distal to it (Bowcock et al., 1990). To delimit the breakpoint region on chromosome 13 more closely, we derived hybrid clones by fusion of the alveolar rhabdomyosarcoma cell line Rh30 to hamster E36 cells. Hybrids were initially scored for the presence or absence of the proximal probe D13S1 and the distal probe D13S10 by Southern blot analysis. Several hybrids that contained D13S1 without D13S10 as well as D13S10 without D13S1 were identified. Two hybrids, one containing a derivative chromosome 13 without a cytogenetically evident normal chromosome 13 homologue (designated D1) and the other containing a derivative chromosome 2 without a normal chromosome 13 homologue (desig-

Figure I. Southern blot analysis of Hindlll-digested DNA from a karyotypically normal EBV-transformed lymphoblastoid line (NI), a human-hamster somatic cell hybrid containing I3pter-q I4 as the only human genetic material (ICD), an HGPRT-deficient hamster cell line used as the fusion partner (E36), der(2) containing somatic cell hybrid (E8), and der( 13) containing somatic cell hybrid (D I)probed with TUBB (A) and D 13s I (6).The TUBB restriction fragment of 16.5 kb has been assigned t o chromosome 8, the I I kb fragment t o chromosome 6, and the 2 kb fragment t o chromosome 13. Fragments of 6 and 4 kb have not been assigned t o specific chromosomes (Floyd-Smith et al., 1986). For D 135 I, Hindlll digestion revealed the presence of a nonpolymorphic 5.5 kb fragment in those cell lines containing chromosome 13q 14. Fragment sizes are indicated in kilobases (kb).

der( 13)

cen

8

Probe

Locus

C1

D1

07

der(2)

m A7

€8

ICD

0

0

-pHUlO

D13S6

-pCR1359 -p7F12

D13S33

0

0

0

0

0

0

D13S1

0

0

0

0

0

0

-p21fi3'UT

TUBBP2

0

0

0

0

0

0

-pTH162

D13S37

0

0

0

0

0

0

-pL28-5 -pG14E1.9

D13S62 D13S22

0 0

0

-pG24E2.4 7p7D2

D13S21 D13S10

0 0 0

0 0 0

0

0 0 0 0

0 0 0 0

-pESD14.1.1

ESD

0

0.07

0.03 0.055

0.15

0.075

0.015 0.021 0.001

0

0.075

0

0

tel Figure 2. Segregation of linked markers on chromosome I 3 between somatic cell hybrids containing either der(2) or der( 13) without normal I 3 homologues. The sex-averaged meiotic recombination distance (0) is given in centimorgans and is based on the linkage data of Bowcock ec al. (1990) as modified by Higgins et ai. ( I 990).

A.

Non - translocated

B.

Translocated

Figure 3. CHEF analysis of D N A from the non-translocated rhabdomyosarcoma cell line, BG (A) and the translocated cell line, Rh30 (B), probed with TUBBP2-P after digestion with the indicated restriction enzymes. For both cell lines (A and B), D N A was digested with Mlul (lane I), Not1 (lane Z),Nrul (lane 3). BssHll (lane 4), or SfiI (lane 5). For Rh30 (B). double digests were performed with Mlul/Notl (lane a), MluvNrul (lane 7),Mlul/BssHII (lane 8), Notl/BssHII (lane 9), NotllSfiI (lane 10). or BssHII/SfiI (lane I I). Fragment sizes are indicated in kilobases (kb).

ALVEOLAR RHABDOMYOSARCOMA BREAKPOINT

CHEF Analysis of TUBBPZ and D 13s I Loci

We next sought to extend the physical map around TUBBPZ and D13S1 by using CHEF analysis. DNA from the Rh30 and BG cell lines was digested in situ with several rare-cutting restriction enzymes, including BssHII, MuI, NotI, NruI, and S$I, used alone and in combination. Under the conditions employed, DNA fragments ranging from 50 to 1,600 kb were resolved. With the chromosome 13-specificprobe TUBBP2-P, DNA from the rhabdomyosarcoma cell line BG was digested individually with several restriction enzymes; each generated unique fragments with each enzyme, ranging in size from 300 kb ( S f l )to 1300 kb (A4ZuI)(Fig. 3A, lanes 1-5). When DNA from the Rh30 cell line containing the t(2;13) was digested with the same enzymes and probed with TUBBP2-P, an additional, larger fragment was present with BssHII (1,100 kb), Mu1 (1,600kb), and NotI (1,600 kb) (Fig. 3B, lanes 1, 2, 4). Fragments common to both cell lines were observed after digestion with S f l (300 kb) and NmI (1,000 kb) (Fig. 3A, B, lanes 3 and 5). These data are summarized in Table 2. We wished to determine whether the heterologous fragments observed after BssHII, MuI, and Not1 digestion of DNA from the Rh30 cell line were caused by differences in methylation between the normal and translocated chromosomes or represented altered fragments due to a translocation event. A double digest performed with BssHII and Mu1 showed that the 550 and 1,100 kb BssHII fragments were contained within an unrearranged 1,300 kb Ma1 fragment common to both the translocated and non-translocatedcell lines (Fig. 3B, lanes 1 and 8), suggesting that the translocation was not responsible for the altered BssHII fragment. We confirmed that the heterogenTABLE 2. Pulsed-Field Restriction Fragment Sizes (kb) With TUBBPZ-P Probe Cell Linea

Enzyme

BG

Mlul Not1 Nrul BssHll SfiI MlullNotl MluIlNrul Mlul/BssHII NotlIBss HII NotllSfiI BssHIl/SfiI

1.300 900 1,000 550 300 ND ND ND ND ND ND

'ND, not done.

Rh 30 [t(2;13)] 1,6001 I ,300 I ,600/900

1,000 I,I00/550 300 I,200/580 1,000 I, IOOl550 I,I00/550 300 300

245

eous fragments generated by digestion with BssHII, MuI, and NotI in the translocated cell line were caused by differences in methylation between the normal and translocated chromosomes by growing the Rh30 cell line in the presence of 5-azacytidinebefore digestion with these enzymes. This treatment resulted in the collapse of the larger 1,600 kb Mu1 fragment, the 1,600 kb NotI fragment, and the 1,100 kb BssHII fragment to fragments of 1,300kb, 900 kb, and 550 kb, respectively, in agreement with the sizes obtained with the non-translocated BG cell line (Fig. 4). These data, combined with the results of the use of a 2.5 kb EcoRI probe (TUBBP2-C)derived after the mapping of a contiguous cosmid clone (data not shown) and a series of additional double digests (Fig. 3B), allowed the construction of a 2,300 kb map around the TUBBP2 locus (Fig. 5). A similar, although less extensive analysis was carried out around the D13S1 locus. These studies showed no differences in fragment sizes, obtained with multiple enzymes, between the BG and Rh30 rhabdomyosarcoma cell lines, including a common 2,500 kb Mu1 band (Table 3). Our results suggest that although the D13S1 locus flanks the breakpoint region on chromosome 13, it is not close enough to the breakpoint to be useful in the characterization of the t(2;13),nor is there any evidence of physical linkage to TUBBP2. Additionally, because of the relatively weak genetic linkage data which we used to establish the order of loci in this area (Table 2), the adjacent proximal (D13S33) and distal (D13S37) loci at the breakpoint were examined for evidence of rearrangement in translocated cell lines. Digestion with multiple, rare-cutting restriction enzymes followed by CHEF analysis failed to demonstrate an altered restriction fragment with probes for either of these loci (data not shown). Analysis of FLT Locus

Specific translocations between cellular oncogenes and a variety of other important genes that are presumably critical to neoplastic transformation have been identified in several human tumors. Although the FMS-like tyrosine kinase (FLT) gene has been localized to chromosome segment 13q12 by in situ hybridization and thus should be significantly proximal to the breakpoint region on chromosome 13, we wished to obtain independent confirmation of its lack of involvement in this translocation (Matsushima et al., 1987). Southern analysis of the somatic cell hybrids containing either the der(2) or der(l3) showed that an unrearranged 2.0 kb Hind11 genomic fragment was present in D1 (data not shown), confirming that the FLT locus is proximal to the breakpoint.

H

H

1600 kb -

1000 kb 600 kb -

300 kb -

+ - + -

+ - +-

Figure 4. CHEF analysis of demethylated Rh30 D N A ( + ) and untreated BG D N A ( - ) hybridized with the TUBBPZ-P probe. The Rh30 cell line was treated with 5-azacytidine and D N A was prepared as described in Materials and Methods before electrophoresis. Fragment sizes are indicated in kilobases (kb).

1

U 1 Kb

I

M

R

TUBBPP-P

I

p21B'3UT

B N

S

[B]is

[N]S

-

I))(5R

B M

N

t

H

100 Kb

I I U 2 Kb

B

ES I ,

L

TUBBPZ-P

E

SMRE \I

I

L

TUBBPZ-C

Figure 5. Long-range restriction map around the TUBBPZ locus and restriction maps of the TUBBPZ-P phage clone (top) and TUBBPZ-C cosmid clone (bottom). The following abbreviations represent identified restriction sites: A, BornHI; B, BssHII; E. EcoRI; H, Hindlll; L. Sall; M, Mlul; N, Notl; R, Nrul: S, SfiI. Restriction sites enclosed in brackets are methylated in the Rh30 cell line.

ALVEOLAR RHABDOMYOSARCOMA BREAKPOINT

TABLE 3. Pulsed-Field Restriction Fragment Sizes (kb) With D13SI Probe Cell Line Enzyme

Mlul Not1 Nrul BssHll SfiI

BG

Rh 30 [t(2;13)]

2,500 535 2,500f600

2,500/600

700 400

700 400

2,500

53s

Additionally, CHEF analysis of the translocated and non-translocated rhabdomyosarcoma cell lines digested with several rare-cutting restriction enzymes did not reveal any differences in fragment sizes with use of a 3.6 kb cDNA probe that spans portions of the extracellular, transmembrane, and kinase domains of the FLT gene (data not shown) (Shibuya et al., 1990). Thus, it is unlikely that the FLT oncogene is a target for disruption by the t(2;13) in alveolar rhabdomyosarcoma. DISCUSSION

Rhabdomyosarcomas are highly malignant tumors arising from undifferentiated mesenchyme and resembling developing skeletal muscle. Although two distinct histologic variants of rhabdomyosarcoma have long been recognized, the molecular events underlying cellular transformation and tumor heterogeneity remain poorly understood. Embryonal rhabdomyosarcomas are characterized by the allelic loss of a putative tumor suppressor locus at llp15.5, a region where a number of important myogenic genes are physically clustered (Scrable et al., 1987, 1990). By contrast, the majority of alveolar rhabdomyosarcomas have a reciprocal translocation between the long arms of chromosomes 2 and 13 (Douglas et al., 1987; Wang-Wuu et al., 1988). No rhabdomyosarcomas have been shown to have both of these genetic alterations, suggesting that these genotypic differences correlate with the previously observed phenotypic variation in clinical and histopathologic characteristics. As an initial step in the characterization of the genes involved in the alveolar rhabdomyosarcomaspecific t(2;13), we constructed and analyzed somatic cell hybrids retaining the relevant derivative chromosomes 2 and 13.This analysis, together with available linkage data, allowed us to refine the limits of the breakpoint region on chromosome 13 significantly. Additionally, we were able to construct a physical map around these flanking loci and to show that the breakpoint region is not identified with probes for

247

either of these loci. Finally, we showed that the FLT oncogene is proximal to the breakpoint region on chromosome 13 and that it is not disrupted by this translocation. We sought to refine the limits of the breakpoint region on chromosome 13rather than on chromosome 2 because higher resolution linkage data were available for the 13q14 subchromosomal region (Leppert et al., 1987; Bowcock et al., 1988, 1990; Higgins et al., 1990). This analysis localized the breakpoint region on chromosome 13 between the anonymous locus D13S1 proximally and the TUBBPZ locus distally, representing a refinement of the localization recently reported by Mitchell and coworkers (Mitchell et al., 1991). The sex-averaged recombination distance between these two loci is 5.5 cM, suggesting a physical distance of approximately 6 megabases between these two segments (Ott, 1986).However, it should be noted that the genetic distance may significantly exceed that determined by long-range restriction mapping (Ott, 1986; Arveiler et al., 1989 Fulton et al., 1989 Higgins et al., 1990). This is because crossover frequency and the degree of interference are not constant throughout the genome, and significant differences between sexes are seen in the rate of crossing over (Ott, 1986). Caution must also be exercised when the results of multipoint linkage analysis are interpreted, especially when closely linked markers with limited polymorphisms are utilized. For example, Higgins and coworkers (Higgins et al., 1990) provided physical evidence that the order of two loci in 13q14 distal to the breakpoint region is actually reversed, although this order was estimated to be 35,000 times less likely than that favored by three-locus linkage analysis (Leppert et al., 1987; Bowcock et al., 1988; Higgins et al., 1990). In this regard, the odds that the proximal flanking locus, D13S1, is distal to its adjacent locus, D13S33, are estimated to be 16:l (Fig. 2). Similarly, the odds that the distal flanking locus, TUBBPZ, is proximal to its adjacent locus, D13S37, are calculated to be 100:1 by multilocus linkage analysis (Bowcock et al., 1990). Thus, constraints imposed by the uncertainties of linkage analysis are inherent in the delineation of the chromosome 13breakpoint boundaries.This situation could potentially be resolved by direct physical linkage of adjacent loci or by significant refinement of the linkage map through the isolation of additional informative probes. Under the conditions employed, we were unable to establish physical linkage between our flanking probes and their adjacent loci. We were able, however, to demonstrate physical linkage between D13S62 and D13S22, which are present on a common 2,500 kb MluI fragment (data not shown), as well as

248

SHAPIRO ET AL.

to confirm the linkage between D13S10 and D13S21, which are present on a common 900 kb Not1 fragment (Higgins et al., 1990). These loci are located distal to the breakpoint region on chromosome 13and show no rearrangement in the translocated cell line after analysis with multiple restriction enzymes (data not shown). Our mapping of the region around the TUBBPZ locus was facilitated in part by “natural” partial digestion products that occurred exclusively in the translocated cell line. Differences in restriction fragment patterns resulting from cell line- or tissuespecific differential methylation, and, rarely, restriction-site polymorphisms have been observed during the construction of long-rangemaps uulier and White, 1988 Gessler and Bruns, 1989 Reilly et al., 1989; Gardner et al., 1990). The possibility that TUBBPZ identified the breakpoint region on chromosome 13 was not supported by additional restriction mapping and the demethylation of DNA from the translocated cell line grown in 5-azacytidine,resulting in a restriction fragment pattern identical to that of the nontranslocated cell line (Figs. 2,3). Allele-specificmethylation of the der(2) chromosome TUBBPZ locus was suggested by a restriction pattern resembling the parental translocated cell line that was obtained with a Not1 digest of the der(2) containing hybrid E8 (data not shown).Whether this allelic difference in methylation is a consequence of cell culture or an inherent characteristic of the TUBBPZ locus in alveolar rhabdomyosarcoma remains to be determined (Silva and White, 1988). The long-range map around the TUBBPZ locus revealed a clustering of rare-cutting restriction sites approximately midway through the mapped 2,300 kb segment. This cluster is a reasonable candidate for a CpG island because it contained BssHII, MuI, and NruI sites within a 45 kb stretch of contiguous DNA. Consistent with these observations, TUBBP2-C detected a primary transcript of approximately 4 kb in Rh30 poly A RNA analyzed by Northern blotting (data not shown). There is good evidence that CpG islands mark the start of many genes (Gardiner-Gardern and Frommer, 1987; Lindsay and Bird, 1987).This CpG island may therefore identify genes closely linked to the TUBBPZ locus that could prove useful in further characterization of this chromosomal region. Finally, it was of interest to investigate whether there was any evidence of structural alterations around the FLT gene, a member of the FMS tyrosine kinase family, which was previously mapped to 13q12. Expression of FMS-related gene products has been noted in cultured skeletal muscle cell lines, and +

FLT transcripts have been detected in rodent skeletal muscle (Claycomband Lanson, 1987;Leibovitch et al., 1988 Shibuya et al., 1990).In agreement with previously reported data, our studies demonstrate that FLT is located proximal to the breakpoint, as shown by specific hybridization with a der(l3)xontaininghybrid. CHEF analysis failed to demonstrate differences in restriction fragment sizes between the translocated and non-translocated cell lines, and Northern blot analysis did not demonstrate any significant ddferences in transcript size between these tumors (data not shown). Our data are in agreement with those of Barr and coworkers, who also found no evidence of tumor-specific rearrangements with several 13q probes, including FLT (Barr et al., 1991). When combined with the results of linkage analysis of multiple probes from this region, our studies provide significant refinement of the localization of the t(2;13) breakpoint region on chromosome 13. The next steps in the investigation of the t(2;13) will require the isolation of probes internal to the flanking loci that we have identified and in close enough proximity to the chromosome 13 breakpoint to provide physical evidence of altered restriction fragments that result from this translocation. This will allow the eventual cloning of the t(2;13)breakpoint region on chromosome 13 and the identification of the involved genes. ACKNOWLEDGMENTS

We gratefully acknowledge Dr. Anne Bowcock for linkage data as well as for supplying the pL28-5probe and Drs. Christopher Mitchell and John Cowell for helpful discussions. We thank John Gilbert for editorial review, Jean Laughter for manuscript preparation, and Sandra Farmer and Sharon Nooner for expert technical assistance. This work was supported in part by Cancer Center Support (CORE) grant (CA-21765), Solid Tumor Program Project grant (CA-23099),and the American Lebanese Syrian Associated Charities (ALSAC). REFERENCES Arveiler B, Vincent A, Mandel J-L (1989) Toward a physical map of the Xq28 region in man: Linking color vision, GGPD, and coagulation factor VIII genes to an X-Y homology region. Genomics 4:460-471. Aurias A, Rimbaut C, Buffe D, Dubousset J, Mazabrand A (1983) Chromosomal translocation in Ewing’s sarcoma. N Engl J Med 309:4!3& 497. Barr FG, Biegel JA, Sellinger B, Womer RB, Emanuel BS (1991) Molecular and cytogenetic analysis of chromosomal arms 2q and 13q in alveolar rhabdomyosarcoma. Genes Chrom Cancer 3:153-161. Bowcock AM, Farrer LA, Hebert Jh4, Kidd JA, Lee W-H, Kidd KK, Cavalli-SforzaLL (1987) Refining the chromosome 13q14 map (HGM 9). Cytogenet Cell Genet 46585. Bowcock AM, Farrer LA, Hebert JM. Agger M, Sternlieb I. Scheinberg M,Buys CHCM, Scheffer H. Frydman M. Chajek-Saul T, Bonne-

ALVEOLAR RHABDOMYOSARCOMA BREAKPOINT

Tamir B, Cavalli-Sforza LL (1988) Eight closely linked loci place the Wilson disease locus within 13q14-21. Am J Hum Genet 43664474. Bowcock AM, Farrer LA, Herbert JM, Agger M, Bale AE,Buys CHM, James D, Donis-Keller H, Cavalli-Sforza LL (1990) A fine structure linkage map for chromosome 13. Cytogenet Cell Genet 51:9&967. Buys CHCM, Scheffer H, Pearson PL, Aanstoot GH. Goor N, Nienhaus AJ (1985) Construction and application of a hybrid panel for regional localization of chromosome 13-specific probes. Cytogenet Cell Genet 40:597. Cavenee W, Leach R, Mohandas T, Pearson P, White R (1984) Isolation and regional localization of DNA segments revealing polymorphic loci for chromosome 13. Am J Hum Genet 3610-24. Chu G, Volrath D, Davis RW (1986) Separation of large DNA molecules by contour-clamped homogeneous electric fields. Science 2341582-1585. Claycomb C, Lanson NAJr (1987) Proto-oncogene expression in proliferating and differentiating cardiac and skeletal muscle. Biochem J 247701-706. Douglass EC, Valentine M, Etcubanas E, Parham D, Webber BL, Houghton PJ. Green A (1987) A specific chromosomal abnormality in rhabdomyasarcoma. Cytogenet Cell Genet 45:1&155. Dryja TP, Rapaport JM, Weichselbaum R, Bruns GAP (1984) Chromosome 13 restriction fragment length polymorphisms. Hum Genet 65:32@324. Eneroth M, Mandahl N, Heim S, Willen H, Rydholm A, Alberts KA. Mitelman F (1993) Localization of the chromosomal breakpoints of the t(1216) in liposarcoma to subbands 12q13.3 and 16~11.2.Cancer Genet Cytogenet 48:lOl-107. Feinberg AP, Vogelstein B (1983) A technique for radiolabeling DNA restriction endonuclease fragments to high specific activity. Anal Biochem 1326-13. Floyd-Smith G, de Martinville B, Francke U (1986) An expressed P-tubulin gene, TUBB, is located on the short arm of human chromosome 6 and two related sequences are dispersed on chromosome 8 and 13. Exp Cell Res 163539-548. Friend SH, Bernard SR, Rogelj S, Weinberg RA, Rapaport JM, Albert DM, Dryja TP (1986) A human DNA segment with properties of the gene that predisposes to retinoblastoma and osteosarcoma. Nature 323643446. Fulton TR, Bowcock AM, Smith DR, Daneshva L, Green P, CavalliSforza LL. Donis-Keller H (1989)A 12 membase restriction mau of the cystic fibrosis locus. Nucleic Acids R& 17271-2&1. Gardiner-Gardern M, Frommer M (1987) CpG islands in vertebrate genomes. J Mol Biol 196:261-282. Gardner K, Horisberger M, Kraus J, Tantravahi U, Korenber J, Rao V, Reddy S, Patterson D (1990) Analysis of human chromosome 21: Correlation of physical and cytogenetic maps; gene and CpG island distributions. EMBO J 92534. Gessler M, Bruns GAP (1989) A physical map around the WAGR complex on the short arm of chromosome 11. Genomics 54S55. Gillin FD, Roufa DJ, Beaudet AL, Caskey CT (1972) 8Azaguanine resistance in mammalian cells I. Hypoxanthine-guanine phosphoribosyltransferase. Genetics 72:23%252. Higgins MJ, Tummel C, Nooland J, Neumann P, Lalande M (1990) Construction of the physical map for three loci in chromosome band 13q14: Comparison to the genetic map. Proc Natl Acad Sci USA 873415-3419. Hunt JD,Valentine M, Tereba A (1990) Expression of N-myc from chromosome 2 in human neuroblastoma cells containing amplified N-myc sequences. Mol Cell Biol 10:82>829. Julier C, White R (1988) Detection of a Not1 polymorphism with the pMeth probe by pulsed-field gel electrophoresis. Am J Hum Genet 424548 Leibovitch SA, Leibovitch M-P, Guillier M, Hare1 J (1988) Expression

249

of two c-fms related gene products in rat muscular stem cells. Oncogene 229>298. Leppert M, Cavenee W, Callahan P, Holm T, OConnell P, Thompson K, Lathrop GM, Lalouel J-M, White R (1987) A primary genetic map of chromosome 13q. Am J Hum Genet 39425433. Lindsay S, Bird AF' (1987) Use of restriction enzymes to detect potential gene sequences in mammalian DNA. Nature 327336-338. Matsushima HM, Yoshida MC, Sasaki M, Shibuya M (1987) A possible new member of the tyrosine kinase family, human frt sequence is highly conserved in vertebrates and located on human chromosome 13. Jpn J Cancer Res 78655-661. McClelland M, Nelson M (1986) The effect of site-specific methylation on restriction endonuclease digestion. Nucleic Acids Res 1 3 2 0 207. Mitchell CD, Ventris JA, Warr TJ, Cowell JK (1991) Molecular definition in a somatic cell hybrid of a specific 213 translocation breakpoint in childhood rhabdomyosarcoma. Oncogene 689-92. Ott J (1986) A short guide to linkage analysis. In Davies KE (ed): Human Genetic Diseases: A Practical Approach. Oxford IRL Press, pp 19-32. Pandis N, Heim S, Bardi G, Mandahl N, Mitelman F (1990) High resolution mapping of consistent leiomyoma breakpoints in chromosome 12 and 14 to 12q15 and 14q24.1. Genes Chrom Cancer 2227230. Reeves BR. Smith S, Fisher C, Warren W, Knight J, Martin C, Chan AML, Gusterson BA, Westbury G, Cooper CS (1989) Characterization of the translocation between chromosomes X and 18 in human synovial sarcomas. Oncogene 4375-378. Reilly DS, Sosnoski DM, Nussbaum RL (1989) Detection of translocation breakpoints by pulsed field gel analysis: Practical considerations. Nucleic Acids Res 135414. Rowe D, Gerrard M, Gibbons B, Malpas JS (1987) Two further cases of t(2;13) in alveolar rhabdomyosarcoma including a review of published chromosome breakpoints. Br J Cancer 56379-380. Scheffer H, van der Lelie D, Aanstoot GH, Goor N, Nienhaus AJ, van der Hout AH, Pearson PL, Buys CHCM (1986) A straightforward approach to isolate DNA sequences with potential linkage to the retinoblastoma locus. Hum Genet 7424%255. Scrable HJ, Witte DP, Lampkin BC, Cavenee WK (1987) Chromosomal localization of the human rhabdomyosarcoma locus by mitotic recombination mapping. Nature 329:645-647. %able HJ, Johnson DK, Rinchik EM, Cavenee WK (1990) Rhabdomyosarcoma-associated locus and MyoDl are syntenic but separate loci on the short arm of chromosome 11. Proc Natl Acad Sci USA 872182-2186. Shibuya M, Yamaguchi S, Yamane A, Ikeda T, Tojo A, Matsushime H, Sat0 M (1990) Nucleotide sequence and expression of a novel human receptor-type tyrosine kinase gene (Rt) closely related to the fms familv. Oncocrene 5519-524. Silva AJ, White R (1988)hheritance of allelic blueprints for methylahon patterns. Cell 51:145-152. Squire J, Dryja TP, Dunn J, Goddard A, Hofmann T, Musarella M, Willard HF. Becker Al, Gallie BL. Phillips RA (1986) Cloning of the esterase D gene: A polymorphic gene probe closely linked to the retinoblastoma locus on chromosome 13. Proc Natl Acad Sci USA 83:6575-6577. Valentine M, Douglass EC, Look AT (1989) Closely linked Ioci on the long arm of chromosome 13 flank a specific 213 translocation breakpoint in childhood rhabdomyosarcoma. Cytogenet Cell Genet 52128-132. Wallace DM (1987) Large- and small-scale phenol extractions. Methods Enzymol 1523M1. Wang-Wuu S, Soukup S, Ballard E, Gotwals B, Lampkin B (1988) Chromosomal analysis of sixteen human rhabdomyosarcomas. Cancer Res 4&983-987.