Proc. NatL Acad. Sa. USAVol. 78, No. 11, pp. 6789-6792, November 1981Biochemistry

DNA sequences homologous to vertebrate oncogenes are conservedin Drosophila melanogaster

(acute leukemia viruses/evolution)

BEN-ZION SHILO* AND ROBERT A. WEINBERGCenter for Cancer Research and Department of Biology, Massachusetts Institute of Technology, Cambridge, Massachusetts 02139

Communicated by David Baltimore, August 4, 1981

ABSTRACT Sequences homologous to the oncogene sequencesof acute RNA tumor viruses have been shown to be highly con-served within vertebrates. In the present work, eight differentoncogene DNA sequences have been used as probes to search forhomologous sequences in the DNA of organisms of other phyla.Five of these probes hybridized to the DNA ofDrosophila melan-ogaster. Abelson leukemia virus probe detected a single homolo-gous DNA fragment in Drosphila DNA. In contrast, probes pre-pared from the genomes of Harvey, avian, and feline sarcomaviruses and avian myelocytomatosis virus hybridized with multiplehomologous sequences in Drosophila DNA. The identification ofsequences homologous to vertebrate oncogenes in invertebratesdemonstrates both a high degree of conservation of these genesand a wide distribution among divergent species. It seems likelythat sequences homologous to vertebrate oncogenes play a crucialrole in metazoan metabolism.

The acutely oncogenic retroviruses constitute a class of 15-20virus types that are able to induce rapidly a variety ofleukemiasand sarcomas in appropriate hosts (1-3). These strains aroseupon passage of slowly leukemogenic viruses through rodents,cats, and chickens. The new viruses that emerged differ fromthe original, slowly leukemogenic parent viruses in two re-spects. First, most of these viruses lack portions of the geneticinformation of the parental virus and, hence, have become de-fective in replication. Second, in place of the deleted geneticsequences, these viruses have acquired new genetic informa-tion which has, in many cases, been identified as being re-sponsible for their acutely oncogenic properties (1-9).The source of the new genetic information has been traced

in many instances to the genome of the host animal throughwhich the parent virus was originally passed (4-6, 9-14). Thus,each of the transforming genes that have been acquired by thevirus can be associated with a counterpart sequence in the ge-nome of the host. These transforming genes have been termed"oncogenes," whereas their normal cellular counterparts havebeen referred to as "proto-oncogenes" (6).

In one well-studied case, it has been shown that the proteinencoded by the proto-oncogene appears similar to the inducedoncogene protein (14-16). Such a result raises questions re-garding the mechanism oftransformation by oncogenes and therole of their gene products in normal cellular metabolism. Onepossibility is that the association ofthe proto-oncogene with thegenome of a retrovirus results in transformation due to higherlevels of expression of the gene (14-20). It remains possible aswell that changes in structural sequences of the proto-onco-genes give rise to transforming capacity in the resultingoncogene.

A variety of observations have shown that DNA and proteinsequences of some well-characterized proto-oncogenes arehighly conserved within vertebrates (4, 5, 14, 15, 21). There-fore, it appears that proto-oncogenes, like many other cellulargenes, fulfill functions that are essential in vertebrates at thelevel of cellular metabolism or tissue differentiation. A clue tothe function of proto-oncogenes was provided by experimentsshowing that the products of certain proto-oncogenes may in-teract with one another in a cascade pathway of protein kinases(22, 23), whose function may be related to regulation ofcellularenergy metabolism.We wished to get a broader view of the evolution of proto-

oncogenes by looking for the most primitive organism in whichthese genes could be detected. The presence of these genes inorganisms very dissimilar from vertebrates would suggest a roleof these genes in cellular metabolism rather than in vertebrate-specific tissue differentiation. Moreover, the genetic systemsof these less complex organisms could be exploited to gain newinsights into the function of proto-oncogenes.

MATERIALS AND METHODSSources and Length of the Oncogene Specific Probes. The

oncogene specific probes were kindly provided to us by the fol-lowing scientists: (i) Abelson leukemia virus probe (clonepABsub9: 3-kilobase pair (kbp) insert; J. Wang and D. Balti-more); (ii) avian myelocytomatosis virus (MC29) (clone MyC3-Pst: 1.5-kbp insert; J. M. Bishop); (iii) avian sarcoma virus (ASV)(Pvu II E fragment: 800-bp insert; J. M. Bishop); (iv) ST/felinesarcoma virus (FeSV) (clone Pst-3: 700-bp insert; C. Sherr); (v)Moloney murine sarcoma virus (clone p440: 440-bp insert; S.Goffand D. Baltimore); (vi) Kirsten murine sarcoma virus (cloneHiHi-3: 1-kbp insert; R. Ellis and E. Scolnick); (vii) Avian er-ythroblastosis virus (clone pAE Pvu II: 2.5-kbp insert; J. M.Bishop); and (viii) Harvey murine sarcoma virus (HaSV) (cloneBS-9: 450-bp insert; R. Ellis and E. Scolnick). This last-namedprobe is not reactive with rat 30S sequences (R. Ellis, personalcommunication).The clones contain either a fragment of the oncogene se-

quence or the entire oncogene and are not reactive with theleukemia virus sequences. All probes were cloned and propa-gated in the bacterial plasmid pBR322.

Cellular DNA. Mouse DNA was extracted from NIH3T3cells. The other DNAs were kindly provided by the followinginvestigators: Drosophila mekanogaster fly DNA, M. Hoffinan;A4E6 Drosophila cell line DNA, L. Cherbas; Lytechinus pictussea urchin DNA, J. Ruderman; Caenorhabditis elegans ne-

Abbreviations: kb, kilobase(s); kbp, kilobase pair(s); HaSV, Harveymurine sarcoma virus; MC29, avian myelocytomatosis virus; ASV, aviansarcoma virus; FeSV, feline sarcoma virus.* Present address: Dept. of Virology, Weizmann Institute, Rehovot,Israel.

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

6789

Dow

nloa

ded

by g

uest

on

Dec

embe

r 7,

202

0

6790 Biochemistry: Shilo and Weinberg

matode DNA, V. Ambros; Saccharomyces cerevisiae yeastDNA, N. Neff; and Dictyostelium discoideum slime mold DNA,R. Kessin.

Nucleic Acid Hybridization. DNA was digested with restric-tion endonucleases, fractionated by agarose gel electrophoresis,and transferred to nitrocellulose filter paper by the Southernprocedure (24). The nitrocellulose filters to which DNA wasbound were prehybridized for 3-6 hr at 40'C in 35% (vol/vol)formamide/0.75 M NaCl/75 mM sodium citrate/65 mMKH2PO4/5 mM EDTA/0.1% polyvinyl pyrrolidone/0.1%Ficoll/1% bovine serum albumin containing 500 pug of boiledsalmon sperm DNA per ml. Hybridization was done at 40'C for12-24 hr in 35% formamide/0.75 M NaCl/75 mM sodium ci-trate/65 mM KH2PO4/5 mM EDTA/0.02% polyvinyl pyrroli-done/0.02% Ficoll/0.2% bovine serum albumin containing 100gg ofsalmon sperm DNA per ml, 10% (wt/vol) Dextran sulfate,and nick translated, 32P-labeled probe (25) (5 x 106 cpm; spe-cific activity of about 1 x 108 cpm of DNA per ,g). After hy-bridization, the filters were washed in 0.30 M NaCl/30 mMsodium citrate/26 mM KH2POJ1 mM EDTA/0.1% NaDod-S04. Initially the filters were washed at 55°C and then exposedto x-ray film. The filter was then incubated in a wash solutionheld at 60°C and reexposed to film. The filter was then re-washed at 68°C and exposed a final time to x-ray film. The filterswere washed each time for 1-2 hr in 0.30 M NaCl/30 mM so-dium citrate/26 mM KH2PO41 mM EDTA/0. 1% NaDodSO4.In general, the background noise disappeared after washes atthe higher temperatures, whereas the specific hybridizationsignals were not significantly reduced.

RESULTSWe wished to use retrovirus-derived onc probes to search forhomologous sequences in the DNA of a variety of species usingthe Southern blot procedure (24). The use of such probes todetect distantly related sequences is successful only if a limiteddegree of sequence divergence has occurred during the evo-lution of these genes. Excessive sequence divergence resultsin less stable hybrids, and consequently a reduction in hybrid-ization signal may occur. To compensate for such divergence,we incubated probes with filter-bound DNAs at a formamideconcentration lower than normally used, in order to favor for-mation of less stable hybrids.A second factor compensated as well for a decrease in hy-

bridization signal. The lower sequence complexity ofmany non-vertebrate genomes made it possible to apply a greater numberof genome equivalents ofDNA to each gel channel used in theSouthern analysis. This resulted in turn in a proportionate en-hancement of any signal detected by hybridization.

Cloned, sequence-specific probes for eight different verte-brate oncogenes were provided by several investigators. TheDNA of plasmids containing these sequences was labeled bynick translation (25) and used initially to probe nitrocellulosefilters carrying DNA extracted from mouse or Drosophila me-lanogaster cells. These DNAs had been cleaved with variousrestriction enzymes prior to electrophoresis and gel-filter trans-fer. Each probe consisted only of oncogene sequences and,therefore, carried no DNA sequence of a parental leukemiavirus. Probes derived from Moloney murine sarcoma virus,Kirsten sarcoma virus, and avian erythroblastosis virus detectedhomologous fragments in mouse DNA, but no hybridizationwith Drosophila DNA was detected (data not shown). However,the other five probes hybridized not only with mouse DNA butalso with DNA extracted from Drosophila.The simplest pattern of hybridization with Drosophila DNA

was observed when using the oncogene probe of Abelson mu-rine leukemia virus. A single band was observed after digestion

of the Drosophila DNA with BamHI, EcoRI, or HindIII (Fig.1). The single band that appeared after BamHI digestion cor-responds to a DNA fragment of only 2 kbp, even though theprobe contained 3 kbp of Abelson oncogene sequences. Thus,it is likely that only a portion of the Abelson virus oncogene isconserved in Drosophila DNA. However, within the sequencesthat are conserved, the degree of mismatch to the probe is lim-ited because hybridization could be readily detected even aftera subsequent high-temperature wash of the filter. The Abelsonproto-oncogene in mice is highly spliced and has a size ofat least30 kilobase (kb) (26). The homologous gene in Drosophila iseither unspliced or contains only small introns.

In contrast to the Abelson virus probe, the probes preparedfrom HaSV, MC29, ASV, and FeSV detected multiple DNAfragments in Drosophila DNA (Figs. 2, 3). The number ofbandsthat appeared varied from 3 to 10, depending upon the probeand restriction enzyme used. Because the oncogene sequencein each ofthese four probes is less than 1.5 kbp, the multiplicityof bands cannot be accounted for solely by cleavage of a singlestretch of homologous sequence into several DNA fragments.Rather, each one of these four oncogene probes appears to de-tect a family of genes present in Drosophila DNA. The patternof the bands is different for each probe, indicating that thesefour different gene families are distinct from one another.Within a given gene family, differences in the intensity of hy-bridization are observed. For example, an EcoRI digest ofDro-sophila DNA was probed with a 450-bp HaSV-specific se-quence, and three bands were resolved (Fig. 2). A fragment of6 kbp annealed intensely. An intermediate degree of hybrid-ization to a fragment of 9 kbp was detected, and faint hybrid-ization appeared with a DNA fragment of5 kbp. The differencesin intensity are likely to be a reflection of the degree of diver-gence of the different members of the family from the mam-malian oncogene sequence. The intensity of the darker bandsin Drosophila DNA was greater than that of the major band

a bc d e

9.8-6.7-4.5-

2.5-2.2-

FIG. 1. Hybridization of Abelson leukemia virus DNA probe withDrosophila and mouse DNAs. Drosophila melanogaster DNA was ex-tractedfrom flies anddigested with different restriction enzymes. Fourmicrograms was loaded on each channel and fractionated by agarosegel electrophoresis. The DNA was transferred to nitrocellulose filterand hybridized with 32P-labeled Abelson virus DNA. DNA was di-gested by endonucleaseBamHI (channel a), EcoRI (channel b), HindIU(channel c), orPst I (channel d). DNA from NIH3T3 mouse cell (10 gkg)was digested with BamHI (channel e). Size markers are in kbp. Thisexposure was obtained after a filter wash at 680C.

Proc. Nad Acad. Sci. USA 78 (1981)

Dow

nloa

ded

by g

uest

on

Dec

embe

r 7,

202

0

Proc. NatL Acad. Sci. USA 78 (1981) 6791

a b cd e f. W";

The ability to detect sequences in Drosophila DNA that arehomologous to five different vertebrate oncogenes encouragedus to search for homologies in the DNA oforganisms from otherphyla and kingdoms including sea urchin, nematode, yeast, andslime mold. No homologies were detected when yeast or slimemold DNA was probed with these five oncogenes. With ne-matode DNA, however, one faint band was observed after hy-bridization to HaSV probe and to Abelson virus probe (data notshown), three bands were detected as a result of hybridizationto MC29 probe (Fig. 3), and no hybridization to the FeSV probewas observed. Sea urchin DNA demonstrated only faint hy-bridization to the Abelson virus probe (data not shown) and nohybridization to the other probes. The failure to detect se-quences in sea urchin is likely a consequence of its genomiccomplexity, which is 1 order of magnitude higher than that ofDrosophila DNA. This reduces the number of genome equiv-alents that can be applied to the gel and, consequently, lowersthe intensity of the signal observable upon Southern analysis.

DISCUSSIONFIG. 2. Hybridization of HaSV DNA pio

mouse DNAs. DNAs were analyzed as in Fifbridized with DNA extracted from D. melanowith restriction enzymes. Four micrograms ofchannel: a,BamHI digest; b, EcoRP; c, Hindil]was incubated with DNA extracted from A4Ein culture. Four micrograms of A4E6 DNAchannel; e, BamHI digest; f, EcoRI; g, Himcleaved mouse NIH3T3 cell DNA. This exposfilter wash at 6800.

which appeared in mouse DNA (Fig. 2), E

was derived from rat DNA. This is dueDrosophila gene is at a 10-fold higher cochannel because of the lower complexigenome.

The pattern observed with each of theidentical for Drosophila DNA extracted eiA4E6 Drosophila cells grown in cultureThus, the gene families appear to be stathese two different cellular environmentto other repeated gene families whose ci

differ greatly in cell culture and fly strair

23-9.8-6.7-

abcd eA X I

4.5-

FIG. 3. Hybridization of MC29, FeSV, andDrosophila and nematode DNAs. DNAs wereMC29 probe was incubated with DNA extractedFour micrograms was loaded on each channeEcoRI; c, HindIi; d, Pst I; e, the same probe hC. elegans nematode DNA that had been digestprobe was hybridized with Drosophila DNA(channel f) or with Pst I (channel g). ASV pro}DrosophilaDNA cleaved withEcoRl (channel h]nel i). These exposures were obtained after filteing temperatures: 680C (channels a-d) and 60C

obe with Drosophila and In this work we have shown that five different vertebrate on-g. 1. HaSV probe was hy- cogene probes hybridized with homologous sequences in Dro-Vaster flies and digested sophila mwlanogaster DNA. The results cannot be explainedfDNAwas loadedon each by the reactivity of the pBR322 plasmid DNA that

6Drosophila cellsprobe are present in all of the labeled probes used in the hybridiza-

.was applied to each gel tions. Various results point to this conclusion. For example,dIll; h, 10 jug of EcoRI- analysis of a single preparation of EcoRI-cleaved Drosophilaure was obtained after a DNA yielded a different pattern of hybridization with five of

the probes used and no hybridization with the other threeprobes. Moreover, when one ofthe filters was used sequentially

even though the probe with two different probes, distinct patterns of hybridizationto the fact that each were observed. Therefore, the observed hybridization was spe-

ancentration in the gel cific to the oncogene sequences in each probe.nnt

of the Drosophila The DNA sequences of several vertebrate oncogenes haveity of the Drosophila been shown to be present within mammals and birds (4, 5, 21).different probes was Such findings implied that some of these genes evolved more

ither from flies or from than 300 million years ago and may fulfill functions essential for(for example, Fig. 2). vertebrate cell metabolism or differentiation. In this work, we

h Fig. show that five vertebrate oncogene sequences can be detected

ts. This is in contrast in Drosophila DNA. We conclude that the common precursorstTromosomal locations ofthese genes were already evolved 800 million to a billion yearsas (27). ago at a period prior to the divergence of the Annelid-Arthopodfrom the Echinoderm-Chordate superphylum. These se-

quences must play roles that are crucial to metazoan cellular or

f i organismal physiology. Four of the oncogenes which we foundto be present in Drosophila DNA are known to encode kinasesin vertebrate cells (28-31). Some of these gene products havealso been reported to be members of a cascade of interactingkinases which may be important in regulating energy metabo-lism (22, 23). This would suggest that those oncogenes and theirnonchordate homologs code for centrally important functionsof cellular metabolism which may be found in all eukaryotic

x * ~~~~~cells.We have been unable to detect sequences that are homol-

ogous to those of vertebrate oncogenes in unicellular eukaryoticorganisms such as yeast by using the Southern technique. Such

ASV DNA probes with findings would have indicated that these sequences play a roleanalyzed as in Fig. 1 in essential metabolic functions unrelated to processes of dif-d from Drosophila flies' ferentiation occurring in multicellular organisms.,1: a, BamHI digest; b) The ability to detect sequences in Drosophila that are ho-.ybridized with 5 Ag of mologous to vertebrate oncogenes suggests that they should beted with Hindml. FeSV conserved in all organisms ofthe Echinoderm-Chordate and thecleaved with Hindu Annelid-Arthropoid superphyla. Even though urchin is

)e was hybridized wih olutionarily closer to vertebrates than are arthropods, oncogene)orwithHindill (chan-

,r washes at the follow- homologs were not detected in sea urchin DNA by the SouthernDC (channels e-i). blotting procedure. This is likely due to the higher sequence

23-9.8-6.7-4.5-

2.5-2.2-

Biochemistry: Shilo and Weinberg

......:..

'G.."..

Dow

nloa

ded

by g

uest

on

Dec

embe

r 7,

202

0

6792 Biochemistry: Shilo and Weinberg

complexity of the sea urchin genome. Nevertheless, it may bepossible to study the oncogene homologs of organisms havinghigh genomic complexity by isolating these genes in the formof molecular clones from phage vector libraries. Such isolationwould make possible a detailed study of these sequences whoserole as functional genes is at present only speculative. Thus, thespectrum of organisms in which proto-oncogene structure andfunction could be studied would be significantly broadened.

We thank J. Wang, D. Baltimore, E. Scolnick, J. M. Bishop, C.Sherr, S. Goff, R. Ellis, and D. Lowy for providing oncogene DNAclones; L. Parada for help in plasmid preparations; M. Hoffmann, L.Cherbas, J. Ruderman, V. Ambros, N. Neff, and R. Kessin for providingcellular DNA; and J. Toole for critical reading ofthe manuscript. B. -Z. S.was a Chaim Weizmann Postdoctoral Fellow. This work was supportedby National Cancer Institute Grant 26717 to R.A.W. and Grant 14051to S. E. Luria.

1. Shih, T. Y. & Scolnick, E. M. (1980) in Viral Oncology, ed. KleinG. (Raven, New York), pp. 135-160.

2. Viral Oncogenes (1979) Cold Spring Harbor Symp. Quant. Biol44.

3. Andersson P. (1980) Adv. Cancer Res. 33, 109-172.4. Stehelin, D., Varmus, H. E., Bishop, J. M. & Vogt, P. K. (1976)

Nature (London) 260, 170-173.5. Frankel, A. E. & Fischinger, P. J. (1976) Proc. Natl Acad. Sci.

USA 73, 3705-3709.6. Bishop, J. M., Courtneidge, S. A., Levinson, A. D., Opper-

mann, H., Quintrell, N., Sheiness, D. K., Weiss, S. R. & Var-mus, H. E. (1979) Cold Spring Harbor Symp. Quant. Biol 44,919-930.

7. Andersson, P., Goldfarb, M. P. & Weinberg, R. A. (1979) Cell16, 63-75.

8. Canaani, E., Robbins, K. E. & Aaronson, S. A. (1979) Nature(London) 282, 378-383.

9. Chang, E. H., Maryak, J. M., Wei, C. M., Shih, T. Y., Shober,R., Cheung, H. L., Ellis, R. W., Hager, G.L., Scolnick, E. M.& Lowy, D. R. (1980) J. Virot 35, 76-92.

10. Sheiness, D. K. & Bishop, J. M. (1979) J. Virot 31, 514-521.

11. Frankel, A. E., Gilbert, J. H., Porzig, K. J., Scolnick, E. M. &Aaronson, S. A. (1979) J. Virol. 30, 821-827.

12. Yoshida, M., Kawai, S. & Toyoshima, K. (1980) Nature (London)287, 653-654.

13. Roussel, M., Saule, S., Lagrou, C., Rommens, C., Berg, H.,Graf, T. & Stehelin, D. (1979) Nature (London) 281, 452-455.

14. Oppermann, H., Levinson, A., Varmus, H., Levintow, L. &Bishop, J. M. (1979) Proc. Natl, Acad. Sci. USA 76, 1804-1808.

15. Collett, M. S., Erikson, E., Purchio, A. F., Brugge, J. S. & Er-ikson, R. L. (1979) Proc. Nati Acad. Sci. USA 76, 3159-3163.

16. Sefton, B. M., Hunter, T. & Beemon, K. (1980) Proc. Natl. Acad.Sci. USA 77, 2059-2063.

17. Witte, 0. N., Rosenberg, N. E. & Baltimore, D. (1979) Nature(London) 281, 396-398.

18. Langebeheim, H., Shih, T. Y. & Scolnick, E. M. (1981) Virology,in press.

19. Karess, R. E., Hayward, W. S. & Hanfusa, H. (1979) Proc. NattAcad. Sci. USA 76, 3154-3158.

20. Oskarsson, M., McClements, W. L., Blair, D. G., Maizel, J. V.& Vande Woude, G. F. (1980) Science 207, 1222-1224.

21. Spector, D. H., Varmus, H. E. & Bishop, J. M. (1978) Proc. NatlAcad. Sci. USA 75, 4102-4106.

22. Spector, M., O'Neal, S. & Racker, E. (1981) J. Biol Chem. 256,4219-4227.

23. Spector, M., Pepinsky, R. B., Vogt, V. M. & Racker, E. (1981)Cell 25, 9-22.

24. Southern, E. M. (1975) J. Mol Biol 98, 503-517.25. Rigby, P. W. J., Dieckmann, M., Rhodes, C. & Berg, P. (1977)

J. Mol Biol 113, 237-251.26. Goff, S. P., Gilboa, E., Witte, 0. N. & Baltimore, D. (1980) Cell

22, 777-785.27. Potter, S. S., Brorein, W. J., Dunsmuir, P. & Rubin, G. M.

(1979) Cell 17, 415-427.28. Witte, 0. N., Dasgupta, A. & Baltimore, D. (1980) Nature (Lon-

don) 286, 826-831.29. Collett, M. S., Purchio, A. F. & Erikson, R. L. (1980) Nature

(London) 285, 167-169.30. Hunter, T. & Sefton, B. M. (1980) Proc. Natl Acad. Sci. USA 77,

1311-1315.31. Van de Ven, W. J. M., Reynolds, F. H. & Stephenson, J. R.

(1980) Virology 101, 185-197.

Proc. Nad Acad. Sci. USA 78 (1981)

Dow

nloa

ded

by g

uest

on

Dec

embe

r 7,

202

0