Oncogene (2000) 19, 69 ± 74 ã 2000 Macmillan Publishers Ltd All rights reserved 0950 ± 9232/00 $15.00 www.nature.com/onc Nuclear localization and transactivating capacities of the papillary renal cell carcinoma-associated TFE3 and PRCC (fusion)

Marian AJ Weterman*,1, Jan JM van Groningen1, Aswin Jansen1 and Ad Geurts van Kessel1

1Department of Human Genetics 417, University Hospital Nijmegen, PO Box 9101, 6500 HB Nijmegen, The Netherlands

The papillary renal cell carcinoma-associated 1 (Weterman et al., 1996a,b; Sidhar et al., t(X;1)(p11;q21) leads to fusion of the transcription factor 1996). As a consequence, two reciprocal fusion , TFE3 on the X-chromosome to a novel gene, TFE3PRCC and PRCCTFE3, are formed encoding PRCC, on . As a result, two putative two putative fusion proteins. TFE3 is a ubiquitously fusion proteins are formed: PRCCTFE3, which contains expressed transcription factor that was originally all known domains for DNA binding, dimerization, and identi®ed as a binding to the mE3 element of transactivation of the TFE3 protein, and the reciprocal the immunoglobulin heavy chain enhancer (Beckmann product TFE3PRCC. Upon transfection into COS cells, et al., 1990) which is thought to be crucial for the both wild type and fusion proteins were found to be activation of immunoglobulin heavy chain genes located in the nucleus. When comparing the transactivat- (Banerji et al., 1983; Gillies et al., 1983; Neuberger ing capacities of these (fusion) proteins, signi®cant 1983). The TFE3 protein belongs to a superfamily of di€erences were noted. PRCCTFE3 acted as a threefold transcription factors, all of which contain a basic better transactivator than wild type TFE3 both in a helix ± loop ± helix region followed by a leucine zipper TFE3-speci®c and in a general (Zebra) reporter assay. (bHLHzip). Both domains are essential for DNA In addition, PRCC and the two fusion proteins were binding and dimerization (Beckmann and Kadesch, found to be potent transactivators in the Zebra reporter 1991; Roman et al., 1992). The PRCCTFE3 fusion assay. We propose that, as a result of the (X;1) protein still contains these domains, in addition to the translocation, fusion of the N-terminal PRCC sequences N-terminal 156 amino acids of PRCC. The reciprocal to TFE3 alters the transactivation capacity of the translocation product, TFE3PRCC, contains the N- transcription factor thus leading to aberrant gene terminal 178 amino acids of TFE3 harboring none of regulation and, ultimately, tumor formation. Oncogene the known functional domains of TFE3, and the C- (2000) 19, 69 ± 74. terminal 335 amino acids of PRCC. The PRCC gene exhibits no signi®cant homologies Keywords: TFE3; PRCC; renal cell carcinoma; t(X;1) to other known genes. It is also ubiquitously expressed translocation and its predicted protein shows a relatively high content of prolines and glycines. Within the N- terminal translocated part, the relative content of prolines is 27%. PRCC contains three P-X-X-P motifs Introduction which have been encountered in proteins that interact with SH3 domains of other proteins (Sudol, 1998; Ren Based on their histologic appearance, renal cell cancers et al., 1993). Together with the presence of several can be divided into several groups. Chromophilic potential sites for kinase phosphorylation, among tumors which mostly show a papillary growth which one for tyrosine phosphorylation, it may be pattern, and are thus commonly referred to as speculated that PRCC is involved in cellular signaling. papillary renal cell cancers, constitute 10 ± 15% of the Homopolymeric stretches of prolines and glycines have cases. Within this group, a subset is encountered with also been encountered in transactivation domains chromosomal translocations involving the Xp11region, (Gerber et al., 1994), thus pointing at a role in mostly a (X;1)(p11;q21) translocation. There is transcription regulation. Indeed, the amino-acid accumulating evidence that these tumors de®ne a composition of the N-terminal part of PRCC is distinct subgroup of renal cell cancers with a relatively remarkably similar to that of other known transactiva- early age of onset and a lower male-female preponder- tion domains such as the one in Hox4.2 (Artandi et al., ance as compared to other renal cell cancers (Tonk et 1995). al., 1995; Dijkhuizen et al., 1998). Tumor-speci®c chromosomal translocations fre- In two cases, the (X;1)(p11;q21) translocation has quently lead to the fusion of genes encoding been reported as the sole cytogenetic abnormality transcription factors (Rabbitts, 1994). For example, present (Meloni et al., 1993; de Jong et al., 1986), in t(11;22)-positive Ewing sarcomas the transactivation indicating that the genes involved may play a pivotal domain of EWS is fused to the DNA-binding domain role in tumorigenesis. Positional cloning of a genomic of FLI1, a member of the Ets family of transcription breakpoint fragment revealed that as a result of the factors (Delattre et al., 1992). Such fusions may lead to translocation the transcription factor TFE3 gene on the inappropriate expression of target genes (Sorensen and X-chromosome is fused to a novel gene, PRCC,on Triche, 1996). In addition, it has been demonstrated that expression of EWS/FLI1 leads to transcription of at least a partially distinct set of genes, as compared to *Correspondence: MAJ Weterman Received 2 August 1999; revised 24 September 1999; accepted 29 cells expressing wild type FLI1 (Braun et al., 1995). September 1999 Alternatively, the subcellular localization of the fusion Localization and transactivation of TFE3 and PRCC MAJ Weterman et al 70 protein may be altered in such a way that it interferes corresponding cDNAs were cloned in frame with green with the ability of the transcription factor to exert its ¯uorescent protein (GFP) in the mammalian expression normal function. For example, in t(1;11)-positive acute vector pEGFP-C, followed by transient transfection leukemias, a fusion between the cytoplasmic protein into green monkey kidney epithelial cells (COS). Both eps15 and the transcription factor MLL/HRX/ALL1 is living cells and ®xed cells were analysed. DAPI staining observed. The HRXeps15 fusion protein is also located was performed for identi®cation of the position of the in the nucleus, but in distinct nuclear subdomains nucleus. (Rogaia et al., 1997). Typical patterns are shown in Figure 1. In the Therefore, we sought to establish the subcellular majority of cells, normal TFE3 was found to be localization of TFE3, PRCC, and the respective fusion located in the nucleus, as might be expected for a proteins PRCCTFE3 and TFE3PRCC. In addition, we transcription factor. However, in a minority of cells have determined the transactivational properties of cytoplasmic ¯uorescence was also observed. In a few these (fusion) proteins using general (Zebra) and cases, the nucleus even appeared to be void of TFE3-speci®c reporter systems. ¯uorescence (Figure 1a). The patterns within the nucleus varied from relatively small dots dispersed throughout the nucleus, except for a few negative areas which may correspond to the nucleoli (Figure 1c), to Results irregular patches (Figure 1a). Plating of the cells at di€erent levels of con¯uency before electroporation did Nuclear localization of the TFE3 and PRCC (fusion) not a€ect the variability in the patterns observed. proteins Likewise, variation in culturing period between For direct visualization of the cellular location of transfection and harvesting did not have any e€ect TFE3, PRCC and the derived fusion proteins, the on the ratio of these patterns.

Figure 1 Cellular localization of GFP-fusions of TFE3, PRCC, and their derived fusion products (green) in COS cells. Typical results are shown. (a,b,c) TFE3 in ®xed (a) and living cells (b,c); (d,e,f) PRCC in living (d) and in ®xed cells (e,f); g,h: TFE3PRCC in living (g) and ®xed cells (h); i,j,k: PRCCTFE3 in ®xed cells. (a,e,f,h,i,j) DAPI staining in blue. The GFP ¯uorescence and DAPI staining are shown separately in j and k respectively

Oncogene Localization and transactivation of TFE3 and PRCC MAJ Weterman et al 71

a

b

Figure 2 Schematic picture of reporter and expression (e€ector) constructs. TATA: minimal promoter; AD: activation domain; bHLHZIP: basic region, helix ± loop ± helix, followed by a Leucine zipper. DB: DNA-binding domain; DZ: dimerization domain; ZB: Zebra-binding domain. In case of the pZd constructs, cDNAs were cloned in frame with the N-terminal end of the Zebra coding region

Figure 4 (a) Transactivation in a general transactivation system (Zebra). Test constructs were all Zd-fusion proteins. Values given (fold activation) are averages of six individual experiments. Transactivation is expressed relative to that of the control/empty vector pZd1, re¯ecting the basal activity of the minimal promoter. (b) Western blot analysis of the pellet fraction of the COS cell lysates which were also used for the reporter assays. The pZd1 vector (pZd) and Zd-fusions of the respective proteins (TFE3PRCC, PRCCTFE3, PRCC, TFE3) were analysed. The size of the Zd-TFE3 and Zd-PRCCTFE3 proteins is indicated by an arrow

Figure 3 Transactivation in a TFE3-speci®c reporter system. Values given (fold activation) are averages of six individual experiments. Transactivation is expressed relative to that of a Transactivation by TFE3 and PRCCTFE3 using a control/empty vector pEGFPN3 re¯ecting the basal activity of the mE3-luciferase reporter minimal promoter To determine whether the transactivating capacity of TFE3 is altered as a result of the fusion to PRCC, we Normal PRCC also exhibited a nuclear localization made use of a luciferase reporter construct in which (Figure 1d ± f). Cytoplasmic ¯uorescence was rarely mE3 elements are coupled to a minimal promoter seen, if present at all. Again, irregular ¯uorescence (TATA) and a luciferase reporter gene (Figure 2). The patterns were observed within the nucleus. In some DNA binding and dimerization domains present in the cases, larger ¯uorescent areas were noted (Figure 1e,f). TFE3 and PRCCTFE3 proteins will enable binding to When comparing these localization patterns to those this reporter construct. Expression constructs were of the TFE3 and PRCC derived fusion proteins, no transfected in ®vefold excess to ensure maximum apparent di€erences could be detected. PRCCTFE3 binding to the reporter construct. If transactivation localization was nuclear in the majority of cells (Figure domains are present, the luciferase gene will be 1i, j,k), but cytoplasmic ¯uorescence was also observed activated. Since the TFE3 cDNA we recently in a minority of the cells. Likewise, PRCC and described (Weterman et al., 1996a) is longer than the TFE3PRCC showed similar nuclear patterns (Figure mouse cDNA reported previously (TFE3-L) (Roman et 1g,h). No apparent di€erences were observed between al., 1991) we decided to include this latter construct for living cells and cells ®xed with formaldehyde. reference. A mutant form of TFE3 (TFE3-S) which Expression of all transfected proteins was optimal lacks one of the two known transactivation domains between 24 ± 48 h after transfection. Despite the and is known to cause a threefold decrease in variation in patterns, no obvious di€erences were transactivation upon comparison to TFE3-L (Roman noted in expression levels or stability when comparing et al., 1991) was also included. In addition, the the ¯uorescence of wild type and fusion proteins. reciprocal fusion protein TFE3PRCC and wild type

Oncogene Localization and transactivation of TFE3 and PRCC MAJ Weterman et al 72 PRCC were tested, although they both lack the TFE3PRCC. On the other hand, it is known that domains needed for DNA-binding and dimerization. proteins without a consensus NLS may be directed to As expected, no transactivation above background was the nucleus through strong interactions with other observed for these proteins (results not shown). nuclear proteins or through the presence of as yet As can be seen in Figure 3, activation resulting from unidenti®ed nuclear localization signals (Weis, 1998). transfection of TFE3-S is 2 ± 3-fold lower than that of There was no apparent change in the localization of the longer products TFE3-L and full length TFE3. No the fusion proteins as compared to normal TFE3 and signi®cant di€erences were observed between the latter PRCC. To con®rm that the ¯uorescence was indeed two proteins. However, the activation caused by caused by the GFP-fusion constructs, a polyclonal PRCCTFE3 was signi®cantly elevated (threefold) anti-PRCC antiserum was also used which showed the compared to that caused by the normal TFE3 protein. same nuclear localization patterns (not shown). In addition, no apparent di€erences were noted upon comparison of the localization patterns in living and Transactivation by TFE3 and PRCC (fusion) proteins formaldehyde ®xed cells, indicating that the patterns using the Zebra-luciferase reporter observed were genuine and not due to ®xation Since the PRCCTFE3 fusion protein showed increased artefacts. transactivation as compared to TFE3 in the TFE3- When comparing the transactivating capacity of speci®c reporter assay, we examined the transactivating TFE3 with that of PRCCTFE3, activation was capacities of wild type PRCC and the corresponding increased threefold through the addition of N-terminal fusion protein TFE3PRCC, next to wild type TFE3 and PRCC sequences. This di€erence was observed in both PRCCTFE3, using a general transactivation system reporter systems used. In addition, it was found that (Zebra). All wild type and fusion proteins were cloned PRCC itself can function as a transactivating protein. into a pZd vector as depicted in Figure 2. Zd codes for a We conclude that the N-terminal fragment of PRCC mutant Zebra protein which contains dimerization and adds a transactivating domain to the PRCCTFE3 DNA-binding domains but lacks the transactivation protein, resulting in increased transactivation of the domain. As a reporter, a construct was used containing reporter gene. Future experiments will be aimed at Zebra-responsive elements upstream of a minimal delineation of the exact location of the transactivation promoter coupled to the luciferase gene. In this system, domain within this region. TFE3-S showed a lower both PRCC and the fusion proteins TFE3PRCC and transactivation level (2 ± 3-fold) than either one of the PRCCTFE3 turned out to be more potent transactiva- longer proteins (TFE3 and TFE3-L) which is consis- tors than TFE3 (2 ± 6-fold; Figure 4a). The increase in tent with the results published by Roman et al. (1991). transactivating capacity of PRCCTFE3 as compared to Since there was no signi®cant di€erence in transactiva- TFE3 was similar to that observed in the mE3 reporter tion between TFE3 and TFE3-L, it can be concluded system (threefold). As a control, Western blot analysis that the extra N-terminal sequences of the full length was performed demonstrating that equal amounts of TFE3 do not contain transactivation domains. In the TFE3 and PRCCTFE3 proteins were present in the process of cloning the TFE3 product, we identi®ed a lysates (Figure 4b). minor alternative transcript present in RT ± PCR products of RNA from both normal and tumor tissues, lacking 65 nt within the coding region. This Discussion deletion results in a frameshift causing a premature stop, thus leading to the formation of a short protein Exogeneously expressed TFE3, PRCC and correspond- of 110 amino acids. This construct was not included in ing fusion proteins all exhibited nuclear localization our assays. patterns. In each case there was some variation both in The reciprocal product TFE3PRCC which lacks the intensity and in the extent and size of the positive N-terminal portion of PRCC can still act as a areas. However, similar patterns and intensities were transactivator. Its transactivating capacity does not noted when comparing the wild type and fusion signi®cantly di€er from that of wild type PRCC as proteins. In the majority of cells, TFE3 is located in measured in the Zebra reporter assays. In this system, the nucleus, as might be expected for a transcription TFE3 was a less potent transactivator. It can be factor. A nuclear localization signal (NLS) has been concluded that wild type TFE3 and PRCC both identi®ed which is also present in other members of the contain domains capable of transactivation. However, TFE3 family of proteins (TFE3, TFEB, TFEC, and the the most striking e€ect of the translocation is the microphtalmia factor mi). Accordingly, mutations in elevated transactivation capacity of the PRCCTFE3 this sequence led to aberrant localization of the fusion product. Recently, two variants of the t(X;1) microphtalmia factor mi (Takebayashi et al., 1996). translocation were reported in which TFE3 is fused to This NLS is located at the C-terminal side of the fusion the splicing factor genes PSF and NonO resulting in junction. Within the PRCC sequence, another NLS fusion of almost the entire splicing factor protein to the could be identi®ed (KPKKRK at position 137) at the C-terminal part of TFE3 containing the DNA-binding N-terminal side of the fusion. Consequently, domain (Clark et al., 1997). This suggests that indeed PRCCTFE3 contains both NLS sequences, whereas the formation of the PRCCTFE3 fusion product is the the TFE3PRCC fusion protein lacks these NLS. most critical for tumor formation. The alteration in Nevertheless, this protein is still located in the transactivating capacity may lead to aberrant expres- nucleus. When examining the full length sequence of sion of TFE3 target genes. Alternatively, non-target TFE3, a potential alternative NLS (RRERRE) was genes may be activated through interactions of N- detected at position 130, N-terminal to the breakpoint. terminal PRCC sequences with other nuclear proteins This may explain the observed nuclear localization of and/or transcription factors. Previously, TFE3 has

Oncogene Localization and transactivation of TFE3 and PRCC MAJ Weterman et al 73 been reported to interact with other transcription and washed with phosphate bu€ered saline (PBS) prior to factors (Rao et al., 1997). In non-lymphoid cells it analysis under a Zeiss Axiophot epi¯uorescence microscope was shown, using a combination of in vitro and in vivo equipped with appropriate ®lters. Alternatively, cells were assays, that a three-protein-DNA complex can be ®xed in 2% acid-free formaldehyde for 30 min at room formed on a minimal B-cell-speci®c m enhancer. temperature, washed with PBS, and stained with 0.5 mg/ml of DAPI (Serva, Heidelberg, Germany) for 15 ± 20 min at room PU.1, a macrophage-speci®c factor, and Ets, both temperature, washed twice with PBS-Tween, once with PBS, members of the ETS family of proteins, can bind to the followed by rinsing in water (26), dehydration in methanol, mA and mB sites within this enhancer. Binding of TFE3 and mounting in Mowiol (Hoechst, Frankfurt am Main, to the adjacent mE3 element leads to optimal Germany). Digital images were recorded using Oncor-image transactivation of the enhancer (Rao et al., 1997). software (Gaithersburgh, MD, USA) with a cooled CCD Recently, TFE3 and Smad3, an intermediate in TGFb camera (Sensys, Photometrics, Tucson, AZ, USA). signaling, were shown to synergistically activate the PE2 promoter, an E-box containing fragment of the Transactivation assays plasminogen activator inhibitor PAI-1 in mutant HT1080 ®brosarcoma cells. Binding of both these COS7 cells were transiently transfected with 2.5 mgofa bGAL vector, 4 mg of the reporter vector containing proteins was found to be indispensable for TGFb- luciferase and 20 mg of the e€ector vector (expression inducible activation of the PE2 promoter (Hua et al., constructs of TFE3, PRCC and derived translocation 1998). Taken together, TFE3 may function in concert products) as described above. Cells were allowed to grow with other more tissue-restricted transcription factors for 2 days after electroporation before harvesting. After two and/or transactivating proteins to drive the expression washes with PBS, cells were harvested by scraping, washed of target genes. The addition of PRCC sequences to with PBS, collected by centrifugation, and resuspended in the C-terminal part of TFE3 may lead to novel lysis bu€er (25 mM Bicine pH 7.8, 0.05% Tween 20, and protein-protein interactions which in turn may inter- 0.05% Tween 80). Lysis was allowed to proceed for 20 min at fere with the normal regulation of gene expression, room temperature. Lysates were centrifuged for 15 min at ultimately resulting in tumorigenesis. 48C at 19 000 g. The resulting supernatant was stored at 7808C until use. The amount of protein was determined using the Biorad protein assay (Biorad, Hercules, CA, USA), and bGAL assays were performed to correct for di€erences in Materials and methods transformation eciency (Edlund et al., 1985). Twenty-®ve or 50 mg of protein corrected for the bGAL values, was used to Plasmid constructs perform the luciferase assay according to the instructions of the manufacturer (Promega, Madison, WI, USA). Luciferase Full length cDNAs for TFE3, PRCC, TFE3PRCC, and activity was measured in a luminometer (Lumat EG&G PRCCTFE3 were cloned in frame with green ¯uorescent Berthold). protein (GFP) in pEGFP-C vectors for subcellular localiza- All assays were performed six times, using several tion studies, or out of frame to the GFP in pEGFP-N vectors independently isolated DNA batches of the e€ector plasmids. (Clontech, Palo Alto, CA, USA) for transactivation studies. Reporter constructs containing mE3 elements coupled to a luciferase reporter gene were generously provided by Dr Western blot analysis Calame (Department Microbiology, Columbia University, The pellet fraction of the same lysates used for the reporter NY, USA), as were expression constructs containing the assays was boiled for 5 min in SDS-sample bu€er before previously described TFE3-L, and TFE3-S cDNAs (Roman loading aliquots onto SDS-polyacrylamide gels. Immunoblot- et al., 1991). For use of the Zebra transactivation assay, full ting was performed according to the method of Towbin et al. length cDNAs encoding TFE3, PRCC, and their respective (1979) using nitrocellulose membranes (Schleicher and fusion proteins were cloned in frame to the Zd protein into Schuell, Germany). After blocking overnight in 5% non fat the pZd vector (Askovic and Baumann, 1997). For our use, dry milk/Phosphate Bu€ered Saline containing 0,05% Tween- the HindIIIBamHI restriction fragment containing seven 20 (PBST) the blot was incubated with the primary antibody Zebra-responsive elements and the minimal E4 promoter of in a 1 : 500 dilution. The R4 polyclonal antibody against EBV of the described reporter construct was isolated, ®lled in TFE3 was generously provided by Dr Calame. After washes to create blunt ends, and cloned into the SmaI site of the with PBST the blot was subsequently incubated with a swine pGL2basic vector (Promega, Madison, WI, USA) which anti-rabbit coupled horse radish peroxidase (DAKO, Den- contains the luciferase reporter gene. mark). Peroxidase activity was detected using a chemolumi- nescent reaction (ECL kit, Amersham, UK). Cellular localization assays Green monkey kidney COS1 and COS7 cells were grown in DMEM with 10% fetal calf serum until subcon¯uent. Approximately 2 ± 56107 cells were used per electropora- Acknowledgements tion. Electroporations were performed using a Genepulser The TFE3 cDNA expression constructs (TFE3-L, TFE3-S), and capacitance extender (Biorad, Hercules, CA, USA), in and mE3 reporter constructs were generously provided by 0.4 cm electroporation cuvettes (Eurogentec, Luik, Belgium) Dr Calame. The Zebra and Zd constructs were a kind gift at 0.26 kV, 960 mF. Subsequently, cells were plated onto of Dr Askovic. This work was supported by the Dutch poly-L-Lysine coated microscope slides, grown for 24 ± 48 h, Cancer Society, grant 98-1804.

References

Artandi SE, Merrell K, Avitahl N, Wong KK and Calame K. Banerji J, Olson L and Scha€ner W. (1983). Cell, 33, 729 ± (1995). Nucleic Acids Res., 23, 3865 ± 3871. 740. Askovic S and Baumann R. (1997). BioTechniques., 22, 948 ± Beckmann H, Su LK and Kadesch T. (1990). Genes Dev., 4, 951. 167 ± 179.

Oncogene Localization and transactivation of TFE3 and PRCC MAJ Weterman et al 74 Beckmann H and Kadesch T. (1991). Genes Dev., 5, 1057 ± Ren R, Mayer BJ, Cicchetti P and Baltimore D. (1993). 1066. Science, 259, 1157 ± 1161. BraunBS,FriedenR,LessnickSL,MayWAandDennyCT. Rogaia D, Grignani F, Carbone R, Riganelli D, Lococo F, (1995). Mol. Cell. Biol., 15, 4623 ± 4630. Nakamura T, Croce CM, Di®ore PP and Pelicci PG. Clark J, Lu Y-L, Sidhar SK, Parker C, Gill S, Smedley D, (1997). Cancer Res., 57, 799 ± 802. Hamoudi R, Linehan WM, Shipley J and Cooper CS. Roman C, Cohn L and Calame K. (1991). Science, 254, 94 ± (1997). Oncogene, 15, 2233 ± 2239. 97. de Jong B, Molenaar IM, Leeuw JA, Idenburg VIS, and RomanC,MateraAG,CooperC,ArtandiS,BlainS,Ward Oosterhuis JW. (1986). Cancer Genet. Cytogenet., 21, DC and Calame K. (1992). Mol. Cell. Biol., 12, 817 ± 827. 165 ± 169. Sidhar SK, Clark J, Gill S, Hamoundi R, Crew AJ, Gwilliam DelattreO,ZucmanJ,PloustagelB,DesmazeC,MelotT, R, Ross M, Linehan WM, Birdsall S, Shipley J and PeterM,KovarH,JoubertI,deJongP,RouleauG, Cooper CS. (1996). Hum. Mol. Genet., 5, 1333 ± 1338. Aurias A and Thomas G. (1992). Nature, 359, 162 ± 165. Sorensen PHB and Triche TJ. (1996). Semin. Cancer Biol., 7, DijkhuizenT,vandenBergE,StoerkelS,TerpeH-J, 3±14. Buerger H and de Jong B. (1998). Cancer Genet. Sudol M. (1998). Oncogene, 17, 1469 ± 1474. Cytogenet., 104, 74 ± 76. Takebayashi K, Chida K, Tsukamoto I, Morii E, Munakata Edlund T, Walker MD, Barr PJ and Rutter WJ. (1985). H, Arnheiter H, Kuroki T, Kitamura Y and Nomura S. Science, 230, 912 ± 916. (1996). Mol. Cell. Biol., 16, 1203 ± 1211. Gerber H-P, Seipel K, Georgiev O, Hoe€erer M, Hug M, Tonk V, Wilson KS, Timmons CF, Schneider NR and Rusconi S and Scha€ner W. (1994). Science, 263, 808 ± Tomlinson GE. (1995). Cancer Genet. Cytogenet., 81, 72 ± 811. 75. Gillies SD, Morrison SL, Oi VT and Tonegawa S. (1983). Towbin H, Staehelin T and Gordon J. (1979). Proc. Natl. Cell, 33, 717 ± 728. Acad.Sci.USA,76, 4350 ± 4354. HuaXX,LiuXD,AnsariDOandLodishHF.(1998).Genes Weis K. (1998). Trends Biochem. Sci., 23, 185 ± 189. Dev., 12, 3084 ± 3095. Weterman MAJ, Wilbrink M, Janssen I, Janssen HAP, van Meloni AM, Dobbs RMJ, Pontes JE and Sandberg AA. den Berg E, Fisher SE, Craig I and Geurts van Kessel A. (1993). Cancer Genet. Cytogenet., 65, 1±6. (1996a). Cytogenet. Cell. Genet., 75, 2±6. Neuberger MS. (1983). EMBO J., 2, 1373 ± 1378. Weterman MAJ, Wilbrink M and Geurts van Kessel A. Rabbitts TH. (1994). Nature, 372, 143 ± 149. (1996b). Proc. Natl. Acad. Sci. USA, 93, 15294 ± 15298. Rao E, Dang W, Tian G and Sen RJ. (1997). J. Biol. Chem., 272, 6722 ± 6732.

Oncogene