myc protooncogene linked to retroviral promoter, but not to ... - NCBI

Proc. Nati. Acad. Sci. USA Vol. 85, pp. 2924-2928, May 1988 Biochemistry

myc protooncogene linked to retroviral promoter, but not to enhancer, transforms embryo cells (tranfection/rerovralprotomyc hybrid genes/conversion of proto-myc to viral myc/viral lymphomas)

R.-P. ZHOU AND P. H. DUESBERG Department of Molecular Biology, University of California, Berkeley, CA 94720

Contributed by P. H. Duesberg, December 16, 1987

ABSTRACT To define conditions under which the chicken protooncogene p-myc is converted to a viral and possibly to a cellular transforming gene, we assayed tansforming function of hybrid genes put together from cloned retroviral and p-myc elements and of p-myc genes Isolated from spontaneous viral lymphomas. Transforming function was measured in quail embryo cells transfected with cloned myc genes. We found that only myc genes with a promoter of a retroviral long terminal repeat (LTR) located between the native p-myc promoter and the second p-myc exon have transforming function. Transforming efficiencies decreased with increasing lengths of unspliced sequences between the LTR and p-myc exon 2. p-myc DNAs with LTRs downstream of the coding region or upstream but in the opposite transcriptional orientation failed to transform embryo cells. Likewise, only those retroviral-p-myc combinations from chicken B-cell lymphomas with a LTR positioned as promoter up m of p-myc exon 2 had transforming function. We conclude that substitution of a retroviral LTR for the promoter and for as yet poorly defined, untranscribed regulatory elements of p-myc is sufficient to convert chicken p-myc to a transforming gene. However, retroviral LTRs can only convert p-myc genes to embryo-cell-transforming genes from a limited number of positions, and not as position-independent enhancers. Further, we deduce that there are two classes of viral chicken B-cell lymphomas, those with and those without embryo-cell-transforming p-myc genes.

that deletion of this exon may be necessary for transforming function. However, one study has reported that a complete human p-myc linked to the long terminal repeat (LTR) of an avian retrovirus transforms quail embryo cells (8). Nevertheless, because exon 1 of human p-myc is totally unrelated to the avian counterpart, the transforming function of this construct in quail cells may have reflected an incompatibility between possible negative avian regulators and the human p-myc exon 1 amounting to a functional truncation. Thus, conversion of p-myc to a transforming myc gene appears to depend either on truncation of p-myc elements, or on the substitution of the p-myc promoter by the viral promoter, or on both. Whereas many viral B-cell lymphomas of chicken contain p-myc genes with retroviruses integrated between p-myc exons 1 and 2, others contain retroviruses integrated further upstream or downstream of p-myc (10-13). Such p-myc genes were proposed to be functionally equivalent to viral myc genes due to transcriptional activation from retroviral promoters (10)-the downstream-promotion hypothesis (3, 4)-or due to position-independent enhancers that are also thought to be encoded in the viral LTR (11). The most persuasive argument in favor of the downstream-promotion, and possibly of the enhancer, hypothesis is that transcription of p-myc is elevated 5- to 30-fold in about 75% of retroviral lymphomas (10, 14). However, p-myc DNAs from lymphomas with retroviruses integrated nearby have been reported not to transform cells in culture, whereas the proviral DNAs of myc-containing viruses do (3, 15). Moreover, there is little evidence that the LTR-myc hybrid RNAs of lymphomas are indeed translated into myc protein. Only one report describes 2- to 7-fold elevated levels of myc protein in certain lymphoma cell lines in which myc RNA levels are elevated 50- to 100-fold over normal (16). Further, it has been pointed out that if integration within a few kilobases of p-myc DNA were sufficient to convert p-myc to a virus-like transforming gene, more than 1 in every 106 infected cells should be transformed, because chicken cells contain about 106 kilobases (kb) of DNA and retrovirus integration is not site-specific (17). By contrast, it has been estimated that only 1 in 1011 infected cells in an animal becomes transformed and leads to a clonal leukemia (17). Moreover, no transformation has ever been described among (>106) cells infected in vitro. Thus, the downstreampromotion hypothesis must be more limited than originally proposed. Here we ask which structural alterations are necessary to convert normal p-myc to a transforming gene and whether p-myc genes from viral lymphomas have transforming functions.

A nucleotide sequence, now termed myc, was originally identified as a specific element of the transforming gene of avian carcinoma/leukemia virus MC29 (1). A closely related myc sequence is present in three other carcinoma and leukemia viruses [MH2, OK10, and CMII (2)]. These viruses cause carcinomas, leukemias, and sarcomas and also transform embryo cells in vitro (3, 4). The myc sequence of these viruses is transduced from exons 2 and 3 of the chicken protooncogene p-myc (2, 3). Structural and functional analyses have shown that most of these myc sequences are necessary for transforming function (3-6). It is not clear which of the several structural differences between p-myc and viral myc genes are essential to convert p-myc to a viral transforming gene. The viral myc genes differ from the cellular p-myc genes in (i) the substitution of the cellular promoter by a retroviral promoter, (ii) the truncation of p-myc exon 1, (iii) scattered point mutations in the shared coding regions of myc, and (iv) the addition of a retroviral gag leader to the coding region of p-myc exons 2 and 3. Previous studies have suggested that point mutations (7, 8) and gag leader sequences (6, 8, 9) are not essential for transforming function of myc-encoded proteins. Since all viral myc genes lack p-myc exon 1 (2), it appeared plausible The publication costs of this article were defrayed in part by page charge payment. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. §1734 solely to indicate this fact.

Abbreviations: p-myc, myc protooncogene; LTR, long terminal repeat.

2924

=.X

Proc. Natl. Acad. Sci. USA 85 (1988)

Biochemistry: Zhou and Duesberg MATERIALS AND METHODS

Viral, Cellular, and Hybrid myc Clones. Refer to Fig. LA. Plasmid pMCV3 contains in a pBR322 vector a complete chicken p-myc gene within a 10-kb BamHI fragment from the -

A

*.1-

< z:

X>

cc

-

-'

'7

..-~4

.-

a_

e I..

,

i

-

-:'

R

.-rlvc EN:"?

M

SE3

,j ~

pBA3

H.. i

r,- 7:R

-,

pBE2

I

r.S mEZ

S5D r-arr Ad

SD-Pmiyc

pleuk

0

2

4

6

8

lOkb

Structure of cloned chicken p-myc, viral myc, and hybrid (A) myc-related clones constructed in vitro as described in Materials and Methods. (B) p-myc clones isolated from retroviral lymphomas. Open boxes indicate viral sequences; stippled boxes indicate p-myc exons E1-E3; hatched boxes indicate viral myc sequences; boxes with stippled arrowheads stand for viral LTRs, with the arrowheads pointing in the transcriptional direction; solid lines represent other sequences; and broken lines represent deletions. SD, splice donor; p, p-myc promoter. FIG. 1.

myc genes.

2925

chicken genome (18). pMH2 carries a complete, unpermuted MH2 provirus in pBR325 (5). dmht-SH (5), a cloned deletion mutant of MH2 that lacks the mht sequence but retains the wild-type myc gene of MH2, was used as a standard for assays measuring transforming function of myc genes by transfection in quail embryo cells. The MH2-p-myc hybrids pSE3 and pBE3 were constructed by substituting the myc coding region of MH2 from upstream of their common Not I restriction site by that of p-myc including p-myc exon 1. For this purpose pMH2 was digested with either Sac I (pSE3) or BstEII (pBE3) restriction endonuclease and the ends were blunted with bacteriophage T4 polymerase (19). The DNA was then digested with Not I. The vector fragment with the 5' LTR and the 3' half of MH2 virus was purified by electrophoresis in a 1% low-melting-point agarose gel (19) and ligated to a Dra I-Not I fragment from p-myc, which contains p-myc exon 1 and part of exon 2. To construct pSE2 (Sac I site was used) and pBE2 (BstEII site was used), the above vector fragment with the 5' LTR and 3' half of MH2 was ligated to a Sma I-Not I fragment of p-myc. This fragment contains 297 base pairs (bp) of intron 1 and exon 2 up to the conserved Not I site. Clones pSE3, pBE3, pSE2, and pBE2 all have a LTR 5' of the myc coding region to act as promoter, and a 3' LTR as terminator of transcription. pLTR-myc was constructed by replacing the Not I-BamHI fragment of pSE3 (which includes the 3' portion of myc, the 3' LTR, and part of the tetracycline-resistance gene of pBR322) with the 3' Not I-BamHI p-myc fragment of pMCV3. pINE2 is a derivative of pSE3 from which p-myc exon 1 was deleted. For this purpose pSE3 was first digested with BstEII. The BstEII site was then filled in with Klenow fragment of Escherichia coli DNA polymerase I. The DNA was then partially digested-with Sma I and religated, and a clone was selected that lacked exon 1 but retained intron 1 up to the 5'-most Sma I site. pmyc-EN32 and pmyc-EN31 were constructed by replacing the 5' portion of MH2, from a HindIII site in the pBR325 region of pMH2 to the Not I site in myc, with the 5' portion of p-myc extending from a BstEII site 653 bp upstream of p-myc exon 1 (pmyc-EN32) or an Apa I site 83 bp upstream of exon 1 (pmyc-EN31) to the Not I site. Both clones retain the putative promoter of p-myc defined by nuclease S1 mapping and primer extension (20, 21) but do not have a LTR at the 5' site of the myc gene. pmyc-lAB was constructed by replacing the BstEII-Not I fragment of pMH2 with a BstEII-Not I fragment of p-myc. The 5' LTR is 653 bp upstream of exon 1. pILTR-myc was constructed in two steps. First, the 5' LTR of pMH2 was inverted by rotating it between a 5' HindIII and a 3' Bgl II site. For this purpose the Bgl II site was first converted to a HindIII site with an appropriate linker. The inverted 5' LTR was then cut out of MH2 from between a 5' HindIII site and a 3' Apa I site and inserted into the Apa I site of p-myc. The LTR of this clone is 83 bp upstream of exon 1 and about 63 bp upstream of the presumed "TATA box" (20, 22). pBA3 was constructed by replacing the BstEII-Not I fragment of pMH2 by the Apa I-Not I fragment of p-myc. SD' pmyc-E2 was constructed by replacing the Sma I-Not I fragment of pMH2 by the Sma I-Not I region of p-myc containing 296 bp of intron 1 and most of exon 2. The MH2-derived region 5' of p-myc includes the viral splice donor (SD). To construct SD - pmyc-E2, the splice donor of MH2 was inactivated by in vitro mutagenesis. For this purpose a 2.5-kb Xho I region was first deleted from MH2. Then a Sac I-Xba I fragment of the deleted MH2 that contains the splice donor in gag was cloned into the Sac I and Xba I sites of M13 phage. The splice donor sequence AAAG/GTGA was then mutagenized with the commercially synthesized oligodeoxynucleotide primer 5' CGGACGAAATCAGCTGTATGACGGC 3' to ACAG/CTGA, according to Zoller and Smith (23). The mutation was confirmed by a diagnostic Pvu II site. The Sac I-Xba I

2926

Biochemistry: Zhou and Duesberg

fragment of wild-type MH2 was then replaced by the mutated equivalent to create an MH2 provirus without splice donor (SD-MH2), which was used to construct SD-pmyc-E2 as described for the construction of SD'pmyc-E2. Refer to Fig. 1B. LL6 is a p-myc DNA, cloned from a chicken B-cell lymphoma, that has a viral LTR at the 3' end of p-myc (11, 24). LL4 is a p-myc DNA cloned from a chicken lymphoma with a single LTR inserted in intron 1 of p-myc, 470 bp upstream of p-myc exon 2 (25). pleuk-1c is a lymphoma DNA with a LTR inserted 1 bp upstream of the splice donor of exon 1. This clone was constructed from clone tjflO (26) by adding back the missing part of the retroviral LTR and the poly(A) site of p-myc. For this purpose a Cla I-EcoRI fragment containing the lost 5' region of the viral LTR from MH2 provirus (5) was cloned into the Cla I-EcoRI sites of pBR322. The resulting clone was called pRl-LTR. Then the EcoRI fragment from clone tjflO containing the 3' half of a viral LTR and the p-myc region without the poly(A) site was inserted into the EcoRI site of pR1-LTR. The missing poly(A) site was restored by replacing the Not I-Sca I region, which includes the 3' half of p-myc and part of pBR of pR1-LTR, with the corresponding Not I-Sca I region containing the native poly(A) site of p-myc from clone pMCV3.

RESULTS Transcriptional Control from a Retroviral Promoter Is Sufficient to Convert p-myc to a Transforming Gene. To determine whether the retroviral LTR can convert p-myc to a transforming gene by acting as a position-independent enhancer, a LTR was placed 5' of the p-myc promoter but in the opposite transcriptional orientation. The resulting construct, pILTR-myc (Fig. LA), did not transform quail embryo cells under conditions where our cloned viral myc standard generated 40 foci per ,ug of DNA (Table 1). The standard, dmht-SH, is a cloned mht deletion mutant of MH2 provirus that has about the same transforming efficiency as a cloned provirus of wild-type MH2 (5). In addition, a possible enhancer function of the LTR was examined by joining a LTR to the 3' end of p-myc. The resulting hybrid myc genes, pmyc-EN31 and -32 (Fig. LA), contain the native promoter of p-myc and different lengths of cellular sequence upstream. The myc coding region of these and several other constructs described below was assembled by joining a 5' region derived from p-myc to a 3' region of MH2 virus at the common Not I site (Fig. lA). These clones also failed to transform quail cells (Table 1). To determine whether truncation of exon 1 from the chicken p-myc gene is necessary for transforming function, a LTR promoter was linked to p-myc upstream of exon 1 at several different positions and the constructs were tested for Table 1. Transformation efficiency of cloned viral, cellular, and hybrid myc genes Clone Foci per kug Foci per ,ug Clone pINE2 40 0 dmht-SH 0 pBE2 6 p-myc 7 0 pSE2 pILTR-myc 1 0 pmyc-EN31 SD-pmyc-E2 0 44 pmyc-EN32 SD+pmyc-E2 From lymphomas 10 pBE3 14 20 LL4 pSE3 30 0 pBA3 pleuk-lc LL6 0 0 pmyc-1AB Experimental DNAs were transferred onto quail embryo cells according to Zhou et al. (5). The transforming efficiencies, measured in focus-forming units per microgram of DNA (foci per ,ug), are the average of at least two independent assays.

Proc. Natl. Acad. Sci. USA 85

(1988)

transforming function by transfection into quail embryo cells. Five such constructs, termed pBE3, pSE3, pLTRmyc, pBA3, and pmyc-lAB, were tested (Fig. 1A). Three of these, pSE3, pBE3, and pLTR-myc, which carry the LTR downstream of the putative p-myc promoter (p in Fig. 1) and upstream of the putative 5' border of p-myc exon 1, transformed quail cells with efficiencies of 20, 10, and 16 foci per ,ug compared to 40 foci per .g of the dmht-SH standard (Table 1). However, pBA3 and pmyc-lAB, which carry the LTR upstream of the p-myc promoter, failed to transform fibroblasts (Table 1). It is conceivable that the constructs pSE3, pBE3, and pLTR-myc transformed because p-myc exon 1 was eliminated due to deletion or aberrant splicing, although the viral splice donor of the gag gene was not present in these constructs. Therefore, RNA made in cells transformed by pSE3 and pBE3 was examined for the presence of p-myc exon 1. Individual foci of transformed quail cells were picked and grown into mass cultures for isolation of poly(A)+ RNA. The RNA was then subjected to gel electrophoresis, transferred onto a nitrocellulose filter, and tested for hybridization with a complete p-myc gene (10-kb BamHI region of p-myc clone pMCV3) and with a probe specific for p-myc exon 1 (1-kb Sma I region of p-myc, Fig. lA). Both the p-myc exon 1 probe and the complete p-myc gene hybridized with the same RNA species (Fig. 2). The sizes of these RNAs were as expected for mRNAs of those constructs (Fig. LA), namely 3.3 kb for pSE3 and 3.1 kb for pBE3. It was also found that pSE3 and pBE3 express the known 58-kDa myc protein (p58) at much higher levels than normal cells (Fig. 3). It is concluded that truncation of p-myc exon 1 is not necessary for transforming function. Nevertheless, the presence of exon 1 decreases the transforming efficiency of hybrid myc genes by a factor >2 compared to the viral myc standard (Table 1). However, the position of the retroviral LTR relative to p-myc appears to be critical, as only the clones with a LTR very close to p-myc exon 1 transformed quail cells. The failure of pBA3 and pmyc-lAB to transform A. Complete probe

p-myc

1 2 3

B. Exon-1 specific probe 1 2 3

I

4.4 kb->

4.4 kb->

2.4 kb-

2.4 kb --h>

1 35 kb->

4

1.35 kb

FIG. 2. Electrophoretic analysis of cytoplasmic mRNAs from quail embryo cells transformed by viral myc and chicken p-myc under the control of retroviral promoters. Gel electrophoresis of 2 ,ug of poly(A)+ RNA and blot hybridization analysis were done as described (19, 27). (A) myc RNA detected by the 32P-labeled 10-kb BamHI fragment of p-myc DNA (Fig. 1A). Lane 1: mRNA from cells transformed by clone pBE3, which contains chicken p-myc under the control of a retroviral LTR. Lanes 2 and 3: mRNA from cells transformed by clone pSE3, which contains chicken p-myc under the control of a retroviral LTR. Lane 4: mRNA from cells transformed by dmht-SH, a clone of the myc gene of MH2 virus. (B) myc RNA detected by the 32P-labeled 1-kb Sma I fragment of p-myc that contains p-myc exon 1. The 32P-labeled probe used in A was washed off and the blot was then rehybridized as forA with the exon 1 probe.

Proc. Natl. Acad. Sci. USA 85 (1988)

Biochemistry: Zhou and Duesberg B

A 2

3

2

4

3 p.O

,O

6

97 kD -* 97 kD

--



4* 45 kD -4

.,.

-

p58

FIG. 3. Electrophoretic analysis of proteins encoded by cellular and viral myc genes under the control of retroviral promoters. The proteins were extracted from transformed, [35S]methionine-labeled quail cells, immunoprecipitated with antiserum to myc protein, and subjected to electrophoresis in a NaDodSO4/7.5% polyacrylamide gel as described (28, 29). The myc proteins were identified by autoradiography of dried gels and estimated to be 58-61 kDa based on the positions of the protein size markers (97, 66, and 45 kDa) indicated. (A) Lane 1: myc proteins from normal quail cells. Lane 2: myc proteins from cells transformed by the dmht-SH clone, containing the myc gene of MH2 virus. Lanes 3 and 4: myc proteins of cells transformed by clone LL4, containing p-myc under the control of a viral LTR from a chicken lymphoma. (B) Lane 1: myc proteins from cells transformed by clone pSE3, which contains p-myc under the control of a retroviral LTR. Lanes 2 and 3: myc proteins from cells transformed by clone pBE3, which contains p-myc under the control of a retroviral LTR.

quail fibroblasts could then be due to the presence in these constructs of either transcription- or translation-inhibitory, normally untranscribed sequences from upstream of p-myc exon 1. In an effort to define further the effect of the position of a retroviral LTR promoter in converting p-myc to a fibroblasttransforming gene, several more constructs were analyzed that carried a LTR in p-myc intron 1 at different positions relative to the myc coding region. One of these, pINE2, which carries the LTR 708 bp upstream of the 5' border of exon 2 (Fig. LA), failed to transform quail embryo cells (Table 1). pSE2 and pBE2 also carry a LTR in the first p-myc intron, 297 bp upstream of the p-myc coding region (Fig. LA). These two clones, which differ from each other only in viral sequences upstream of the p-myc insert, transformed quail embryo cells with efficiencies of 6 and 7 compared to 40 foci per ,ug DNA of the viral myc gene standard (Table 1). Two other constructs, SD(-pmyc-E2 and SD-pmyc-E2 (Fig. LA), contained 266 bp from the viral gag region including the viral splice donor and 297 bp of intron 1 upstream of the myc coding region. In SD-pmyc-E2 the splice donor was destroyed by site-specific mutagenesis. It was found that SD+pmyc-E2 had almost the same transforming efficiency as the viral myc standard, but SD-pmyc-E2 had barely detectable transforming activity (Table 1). p-myc Genes from Viral Lymphomas in Chicken: Only Those with a Viral LTR Promoter Transform Quail Embryo Cells. To test whether p-myc DNAs from viral lymphomas with retroviruses integrated within or nearby p-myc transform quail cells like viral transforming genes or like the above constructs, cloned p-myc DNAs from several lymphomas were analyzed. One of these, LL4, carries a viral LTR about 470 bp upstream of p-myc exon 2 (24, 25) (Fig. 1B). This LTR has the same transcriptional direction as the p-myc gene. It is probably derived from a complete retrovi-

2927

rus that was deleted by homologous recombination between the two LTRs. This clone transformed quail embryo cells, like proviral DNAs or the above constructs (Table 1; Fig. 4). It expresses a mRNA from which the residual intron 1 region is removed by splicing, since an oligonucleotide complementary to the intron 1 sequence of LL4 did not hybridize to the mRNA from LL4-transformed cells (data not shown). This LL4 mRNA was probably spliced from the same potential splice donor site that generated the myc sequence of MC29 virus 461 bp upstream of p-myc exon 2 (22, 30). LL4transformed quail cells expressed two myc proteins, of about 58 and 63 kDa (Fig. 3). Another lymphoma-derived clone, pleuk-ic, in which a viral LTR was integrated near the end of p-myc exon 1 and in the same transcriptional orientation as p-myc (26), also transformed quail embryo cells as efficiently as the proviral standard dmht-SH (Fig. 1B; Table 1). Since the LTR is inserted into p-myc exon 1, it is probable that the intron sequence is spliced out of the mRNA of this clone. Finally, a p-myc clone from a viral lymphoma, LL6 (11, 25), that carries a viral LTR at the 3' end of p-Myc was tested for transforming function. The structure of this p-myc retroviral hybrid is consistent with the model that postulates conversion of p-myc to a transforming gene by an enhancer that is independent of the position or transcriptional orientation (8, 11). As expected from the results with the constructs (Fig. lA), this lymphoma-derived p-myc gene did not transfo-m quail fibroblasts (Table 1).

DISCUSSION Retroviral Promoter Converts p-myc to a Transforming Gene. Here we have demonstrated that replacement of the p-myc promoter and structural or functional elimination of as yet poorly defined untranscribed regulatory sequences (30) by a retroviral LTR is sufficient to convert p-myc to a viral or a virus-like transforming gene. The primary consequence of this promoter substitution is an enhanced expression of p-myc that is estimated to be at least 20 times greater than the norm. However, noncoding sequences between the substitute LTR promoter and the myc coding region, either from upstream of p-myc exon 1 or from p-myc intron 1 or from a retrovirus, can decrease or eliminate the transforming efficiency of a LTR-promoted p-myc gene (Fig. LA and Table 1). It would follow that a qualitative change in gene structure rather than changes in gene products is the essential step by which retroviruses or retroviral LTRs convert p-myc genes to transforming genes. This change necessitates an alteration of the germ-line configuration of p-myc, but not of the p-myc coding sequence. Our results show that truncation of p-myc exon 1, as observed in all four known retroviruses with myc genes, is not necessary for transforming function., A probable answer to the question why p-myc exon 1 is not found in viral myc genes is that it either was not transduced or was transduced and then eliminated because it is not essential for transformB

C

ENI;%

FIG. 4. Foci of quail embryo cells transformed by cellular and viral myc genes. (A) Foci formed by a cloned p-myc gene artificially linked to a retroviral LTR (pSE3). (B) Foci formed by a p-myc gene under transcriptional control of a retroviral LTR cloned from a spontaneous viral lymphoma (LU). (C) Foci formed by the cloned myc gene of MH2 virus (dmht-SH). (x 16.)

2928

Biochemistry: Zhou and Duesberg

ing function. Further, since exon 1 was observed to decrease the half-life of p-myc mRNA (31), its elimination would generate a more stable virus. In addition, our data confirm and extend previous evidence that point mutations (7, 8) and gag leader sequences (6, 8, 9) are not essential for transforming function and argue against the proposals that point mutations activate p-myc to transforming genes (25, 32). Role of Retrovirus Integration Within p-myc in B-Cell Lymphomgagenesis. The retrovirus-p-myc hybrid DNAs from viral B-cell lymphomas analyzed here fall into two classes: those that transform primary quail embryo cells upon transfection and those that do not. p-myc of the lymphoma DNAs with transforming function (LL4 and pleuk-ic) is under the control of a retroviral promoter, like its viral homolog in retroviral myc genes. p-myc of a lymphoma that failed to transform embryo cells (LL6) contains a viral LTR downstream of p-myc or upstream in the opposite transcriptional orientation where it cannot function as a promoter (Fig. 18). These results indicate that a viral LTR must function as a promoter in order to convert p-myc to a gene capable of transforming quail embryo cells. Thus, the proposed enhancing effect of viral LTRs (8, 11) is not sufficient to transform quail embryo cells, probably because the mRNA levels expressed by such enhancer-p-myc conjugations are too low to achieve transformation. We cannot exclude that LTR enhancers activate p-myc sufficiently to achieve B-cell transformation even if fibroblasts and other embryo cells are not transformed in vitro. Nevertheless, there are no known viral myc gene variants that are leukemogenic but fail to transform embryo cells in vitro to support this argument. Our data, showing that only discrete regions and orientations of retroviral integration sites convert p-myc to a transforming DNA, provide then at least a partial answer to a question raised by the downstream-promotion hypothesis, why not all cells with retroviruses integrated within several kilobases of p-myc are transformed. All those hybrid genes in which the retroviral promoter is joined with p-myc in the opposite transcriptional orientation or downstream of p-myc do not transform embryo cells. Further, we deduce, from the constructs described above, that not all p-myc genes linked with a retroviral LTR in a promoter orientation would transform. This is because aberrant translation initiation codons or translation-inhibiting RNA elements in sequences between the LTR and the coding region of myc would render some LTR-p-myc hybrid mRNAs untranslatable, unless such sequences are eliminated by aberrant splicing. Thus, not all LTR-p-myc hybrid mRNAs that are expressed at high levels in lymphomas (10) must be translatable, as has been shown (16) or suggested (12) previously. Goodenow and Hayward (14) found that most, if not all, retroviral-p-myc hybrid DNAs from viral lymphomas contain only defective prQviral DNAs, so that, if viral defectiveness is necessary for transforming function, the probability that an infected cell will become transformed is reduced even further. Since retroviral-p-myc hybrid genes from some viral lymphomas transform embryo cells but those from others do not, the role of such hybrid structures in viral lymphomagenesis remains unresolved. Clearly, p-myc DNAs that can transform embryo cells could be the cause of viral lymphomas. However, there is no evidence that p-myc DNAs from lymphomas that do not transform embryo cells are transforming genes. It is possible that other mechanisms of transformation are responsible for such B-cell lymphomas. Therefore, it is important to test whether the lymphoma DNAs that do not transform embryo cells can transform chicken B cells and express high levels of myc protein. Since the lymphomas from which transforming p-myc DNAs can be isolated have not been distinguished pathologically from lymphomas that lack transforming DNAs, it is also possible that all viral lymphomas are caused by trans-

Proc. Natl. Acad. Sci. USA 85

(1988)

forming genes unrelated to p-myc. Finally, our results predict that heterologous cellular promoters might convert p-myc genes to transforming genes, if they were joined with p-myc by appropriate rearrangements and if they were as efficient as retroviral LTRs. We are grateful to K. Cichutek and G. S. Martin for critical reviews of this manuscript. We thank W. F. Lee and H. Varmus (San Francisco) for the lymphoma DNA clones pLL4 and pLL6, W. Hayward (New York) for the lymphoma DNA clone tjfl0, and S. I. Evans (Cambridge, England) for antiserum to myc protein. This work was supported by OIG Grant 5 R35 CA39915, and CTR Grant 1547AR1. 1. Duesberg, P. H., Bister, K. & Vogt, P. K. (1977) Proc. Natl. Acad. Sci. USA 74, 4320-4324. 2. Duesberg, P. H. (1985) Science 228, 669-677. 3. Weiss, R., Teich, N., Varmus, H. & Coffin, J. (1985) RNA Tumor Viruses: Molecular Biology of Tumor Viruses (Cold Spring Harbor Laboratory, Cold Spring Harbor, NY), 2nd Ed., Vol. 2/supplements and appendices. 4. Weiss, R., Teich, N., Varmus, H. & Coffin, J. (1982) RNA Tumor Viruses: Molecular Biology of Tumor Viruses (Cold Spring Harbor Laboratory, Cold Spring Harbor, NY). 5. Zhou, R.-P., Kan, N., Papas, T. & Duesberg, P. H. (1985) Proc. Natl. Acad. Sci. USA 82, 6389-6393. 6. Patschinsky, T., Jansen, H. W., Blocker, H., Frank, R. & Bister, K. (1986) J. Virol. 59, 341-353. 7. Hayflick, J., Seeburg, P. H., Ohlsson, R., Pfeifer-Ohlsson, S., Papas, T. & Duesberg, P. H. (1985) Proc. Natl. Acad. Sci. USA 82, 2718-2722. 8. Martin, P., Henry, C., Ferre, F., Duterque-Coquilland, M., Lagrou, C., Ghysdael, J., Bebuire, B., Stehelin, D. & Saule, S. (1986) EMBO J. 5, 1529-1533. 9. Shaw, J., Hayman, M. J. & Enrietto, P. J. (1985) J. Virol. 56, 943-950. 10. Hayward, W. S., Neel, B. G. & Astrin, S. M. (1981) Nature (London) 290, 475-480. 11. Payne, G. S., Bishop, J. M. & Varmus, H. E. (1982) Nature (London) 295, 209-214. 12. Shih, C., Linial, M., Goodenow, M. & Hayward, W. S. (1984) Proc. Natl. Acad. Sci. USA 81, 4697-4701. 13. Robinson, H. L. & Gagnon, G. C. (1986) J. Virol. 57, 28-36. 14. Goodenow, M. M. & Hayward, W. S. (1987) J. Virol. 61, 24892498. 15. Cooper, G. M. & Neiman, P. E. (1981) Nature (London) 292,

857-858.

16. Hahn, S. R., Abrams, H. D., Rohrschneider, L. R. & Eisenman, R. N. (1983) Cell 34, 789-798. 17. Duesberg, P. H. (1987) Cancer Res. 47, 1199-1220. 18. Robins, T., Bister, K., Garon, C., Papas, T. & Duesberg, P. (1982) J. Virol. 41, 635-642. 19. Maniatis, T., Fritsch, E. F. & Sambrook, J. (1982) Molecular Cloning: A Laboratory Manual (Cold Spring Harbor Laboratory, Cold Spring Harbor, NY). 20. Linial, M. & Groudine, M. (1985) Proc. Natl. Acad. Sci. USA 82, 53-57. 21. Nottenburg, C. & Varmus, H. (1986) Mol. Cell. Biol. 6, 2800-2806. 22. Papas, T. S., Kan, N. K., Watson, D. K., Flordellis, C. S., Psallidopoulos, M. C., Lautenberger, J., Samuel, K. P. & Duesberg, P. (1984) in Cancer Cells 2/Oncogenes and Viral Genes (Cold Spring Harbor Laboratory, Cold Spring Harbor, NY), pp. 153-163. 23. Zoller, M. J. & Smith, M. (1983) Methods Enzymol. 100, 468-500. 24. Lee, W. M. F., Schwab, M., Westaway, D. & Varmus, H. E. (1985) Mol. Cell. Biol. 5, 3345-3356. 25. Westaway, D., Payne, G. & Varmus, H. E. (1984) Proc. Natl. Acad. Sci. USA 81, 843-847. 26. Neel, B. G., Gasic, G. P., Rogler, C. E., Skalka, A. M., Ju, G., Hishinuma, G. F., Papas, T., Astrin, S. M. & Hayward, W. S. (1982) J. Virol. 44, 158-166. 27. Thomas, A. S. (1980) Proc. Natl. Acad. Sci. USA 77, 5201-5205. 28. Brugge, J. S. & Erikson, R. L. (1977) Nature (London) 269, 346-348. 29. Laemnlli, U. K. (1970) Nature (London) 227, 680-685. 30. Watson, D. K., Reddy, E. P., Duesberg, P. H. & Papas, T. S. (1983) Proc. Natl. Acad. Sci. USA 80, 2146-2150. 31. Piechaczyk, M., Yang, J.-Q., Blanchard, J.-M., Jeanteur, P. & Marcu, K.-B. (1985) Cell 42, 549-598. 32. Rabbitts, T. H., Hamlyn, P. H. & Baer, R. (1983) Nature (London)

306, 760-765.