Molecular Biology Reports 16: 17-25, 1992. 9 1992 Kluwer Academic Publishers. Printed in Belgium.

17

Development and characterization of a panel of monoclonal antibodies against the catalytic domain of the human fes proto-oncogene product Adrie van Bokhoven ~, Hans L.P. van Duijnhoven 1, Manfred J0cker 2, Anton J.M. Roebroek 3 & Wim J.M. van de Ven 1,3

1Molecular Oncology Section; Department of Biochemistry, Universityof Nijmegen, Nijmegen, The Netherlands; 2Medizinische Klinik, UniversitgitKrln, Germany; 3Laboratoryfor Molecular Oncology, Centerfor Human Genetics, Universityof Leuven, Herestraat 49, B-3000 Leuven, Belgium Received 12 April 1991; accepted in revised form 20 June 1991

Key words:fes proto-oncogen, tyrosine-specific protein kinase, monoclonal antibody

Abstract

In developing monoclonal antibodies (Moabs) against the human fes proto-oncogene product, recombinant DNA technology was used to target reactivity of the Moabs towards the catalytic domain of it. Therefore, sequences of human fes exons 15-19 encoding amino acid residues 612 to 822 which harbor the catalytic domain except the presumed ATP-binding region, were fused in phase to the bacterial trp E gene which encodes anthranilate synthase. After partial purification of it, the bacterially produced hybrid product of this trp E-Afes fusion gene was used as immunogen. A series of twelve mouse Moabs was obtained which recognized the human p92 fes protein and the viral oncogene product p85 gag-fes enc o d e d by the Snyder-Theilen strain of feline sarcoma virus. Reactivity appeared to be directed towards the catalytic domain of the human fes proto-oncogene product. This was demonstrated by in vitro transcription and translation experiments using human fes coding sequences from exons 16-19. Upon testing their functional activity in divers immunological techniques, the whole panel of Moabs appeared to be useful in immunoprecipitation, Western blot and immunohistochemical analysis. Immunocytochemical analysis indicated that p85 gag'fes is predominantly a cytoplasmic protein.

Introduction

The human fes proto-oncogene encodes a tyrosine-specific protein kinase of about 92 kDa and, although neither its function nor its precise role in neoplastic transformation is known at present, the gene is thought to be involved in both normal and neoplastic hematopoiesis [6]. This assumption is based on the expression pattern of

the fes proto-oncogene, which is found to be largely though not entirely restricted to normal and leukemic myeloid cells [4, 7, 17, 20]. Transforming viral oncogenes of a number of independent feline and avian sarcoma virus isolates contain genetic sequences that apparently were derived from the felinefes and chicken fps protooncogene [6]. Mobilization of the transforming potential of the proto-oncogene was in all these

18 cases characterized by the fusion of protooncogene to retroviral gag gene sequences. Although the various sarcoma virus isolates had acquired different portions of the proto-oncogene, their products all exhibited protein kinase activity with specificity for tyrosine residues [6]. This seems to implicate the catalytic domain as an important component of these transforming fes proteins. For studies to elucidate of the physiological role that the fes proto-oncogene product may play in myeloid cell proliferation and/or differentiation, the availability of a panel of monoclonal antibodies which are specifically raised against the catalytic domain of the protein and which are functional in divers immunological applications could be instrumental. The aim of this study is to generate antibodies with such specifications, since these are presently not available. Initially, antisera that recognized the viral v-fes oncogene products were obtained from rats bearing tumors produced by Snyder-Theilen feline sarcoma virus (ST-FeSV) [2, 8]. Using a similar approach, rat monoclonal antibodies with specificity for v-fes products were developed [25]. Thereafter, mouse monoclonal antibodies that were raised against fes synthetic peptides became available (NCI). Later on, rabbit antisera raised against bacterial trpE-c-fes-encoded hybrid proteins were described [5, 10]. One of these was specifically raised against the human p92 fes region from amino acid residue 406 to 451 [5], which is located just upstream of the SH2 domain of this proto-oncogene product [12]. We here report about the production in bacteria of a hybrid protein (trpE-Afes) that contains the human p92 fes catalytic domain (residues 612-822) and its use in the development of Moabs that specifically react with epitopes within this domain of the human proto-oncogene product. Characterization of these newly developed Moabs indicated that they are functional in a variety of immunological applications. By immunocytochemical analysis, the subcellular localization of the transforming p85 gag-fes polyprotein encoded by ST-FeSV was determined.

Materials and methods

Cell culture In addition to the human promyelocytic cell line HL60, fetal mink lung cell line, CCL64, obtained from the American Culture Collection, Rockville, Maryland, a subclone of CCL64 nonproductively transformed by the Snyder-Theilen strain of feline sarcoma virus (ST-FeSV), designated ST-FeSVCCL64, and a subclone of CCL64 productively transformed by ST-FeSV(4070A), designated ST-FeSV-4070A-CCL64, were used in this study. Cells were grown in Dulbecco modified Eagle medium supplemented with 10% fetal bovine serum, penicillin, streptomycin and L-glutamin.

Molecular cloning of fes coding sequences into prokaryotic expression vectors As prokaryotic expression vectors, the pATH [22] and pUR [ 16] series were selected for use in this study. Of the various vectors of these systems, pATH 11 and pUR292 appeared to be instrumental. As bacterial host for recombinant DNA constructs of pATH 11, bacterial strain RRt was used, and for those ofpUR292, strain JM 109. As coding sequences for the catalytic domain of the fes proto-oncogene, an 0.95 kbp cDNA clone (EcoRI/EcoRI fragment), isolated from a cDNA library constructed with KG-1 mRNA [ 13-15], was used. This cDNA clone contained coding sequences ofexons 15-19 [13, 14]. Starting at the third codon (of the authentic fes open reading frame) within exon 15 and ending at the authentic stop codon in exon 19, the fes coding sequences included a major portion of the fesprotein kinase domain. The fes cDNA fragment was cloned into the unique EcoRI site of pATH 11. The resulting recombinant DNA construct was designated pJG44. In that way, a trp E-Afes fusion gene was generated in which the fes open reading frame was fused in phase to the trp E coding sequences. For construction of a lac Z-Afes fusion gene, thefes DNA fragment was cloned into the HindlII site of pUR292 after

19

the lac Z-Afes fusion gene was obtained upon induction by IPTG (isopropyl-fl-D-thio-galactoside). Bacteria transformed with pJG44 were grown in tryptophan-free M9 medium and synthesis of the hybrid protein encoded by the trp EAfes fusion gene was obtained as described by Spindler et al. [22].

treatment of the fes c D N A fragment with D N A polymerase I and adding a HindlII linker, as indicated in Fig. 1. The resulting recombinant D N A construct was designated pJG19. The nucleotide sequence of the Afes insert was checked by nucleotide sequence analysis [9, 18]. Bacteria transformed with pJG19 were grown in TY medium and synthesis of the hybrid protein encoded by

A E

K K '

sK

L.J 1

~

P

1'so I

I

I

i

I 2 3

K

SS

: I

PPP P i~-''$

B SK K

BB

X

S

H

I

I

'~

1

pp I I

9 I 45

S

]5 9

I I

II I I I 678910

'

p p

'

Xh

I I

9 I

I I

11 12 13

'

p~

SE p

p

p /

I l l l l 9 II II

i

'i

i Xh

id

14 1516 17 18 ii

p

19

i

I~-(A)

n

c-DNAfes (o95k-'~

AATAAA I

B SaP

Hc

I

sapH E EcoR! fragment

f s

1.0 klAo

~

Eco RI

Eco R I

'I

T4 Ligase

1

I

~p'~r~\\

~. DNA polymerase I Hind i]]

!' ~acZ

;/

linker

I H?m

Hind []] [ IT4

Ligase

Fig. 1. (A) Schematic restriction map and genetic organization of human fes. The positions of the noncoding exon sequences, represented by open boxes, and the coding exons sequences, represented by black boxes, are given relative to the restriction map. The start codon in exon 2 is indicated by a black triangle and the stop codon in exon 19 by an asterisk. The 0.95 human fes cDNA clone harboring coding sequences ofexons 15-19 is schematically indicated. B: BamHI; Bg: BglII; E: EcoRI; H: HindllI; K: KpnI; P: PstI; S: SstI; X: XbaI; Xh: Xhol. (B) Scheme for the molecular cloning of the 0.95 kb human fes c D N A fragment into expression vectors pATH11 and pUR292. The cDNA is represented as an EcoRl fragment.

20

Generation of monoclonal antibodies For the generation of Moabs, mice were immunized with the bacterially produced hybrid protein encoded by the trp E-Afes fusion gene in pJG44. The trpE-Afes (trpE = trp E-encoded polypeptide anthranilate-synthase; Afes=polypeptide encoded by the 0.95 kb fes cDNA sequences) hybrid protein encoded by pJG44 was purified according tot the procedure described by Adam et al. [1]. Hybridomas were obtained as described earlier [24]. Antibodies produced by hybridomas were analyzed by ELISA and Western blot analysis for reactivity with bacterial lysates containing the hybrid proteins trpE-Afes and flgal-Afes (flgal = ]~-galactosidase). As controis in these experiments, bacterial lysates containing the trpE polypeptide (about 37 kDa) or /~gal (about 116 kDa) were used.

ELISA and Western blotting analys& For ELISA experiments, complete bacterial lysates containing approximately 1 mg/ml trpE, trpE-Afes, flgal, and flgal-Afes, were diluted 1:250 in 0.05 M sodium carbonate buffer (pH 9.6) and 100#1 was coated o/n at 4 ~ on microtiterplates (Nunc). After blocking with PBS/I~o gelatine (PBS-G) for 2 h at 37 ~ hybridoma culture supernatant was applied for 1 h at 37 ~ As controls, MON-63 (anti-flgal) and a polyclonal mouse anti-trpE antiserum were used. Antibody binding was detected after incubation with peroxidase-conjugated rabbit anti-mouse antiserum (DAKO) diluted 1:250 in PBS-G for 1 h at 37 ~ using 5'-amino-salicylic acid as a substrate. After 30 min, the extinction was measured at 450 nm. For Western blot analysis of cell lysates, containing the viral polyprotein p 8 5 gag'fes o r the human p92 fes, or bacterial lysates containing the hybrid proteins mentioned above, samples of such lysates were prepared as described before [24] and, upon S D S-PAGE, protein s were transferred onto 0.45 #m nitrocellulose (Schleicher & Schuell). SDS-PAGE was performed on 12.5~

poly-acrylamide gels under reducing conditions. To prevent non-specific protein binding, the nitrocellulose sheets were incubated for 1 h at room temperature in PBS/0.05 ~o Tween 20 (PBS-T). Subsequently, blots were incubated for 2 h at room temperature in PBS/0.05~o Tween20 (PB S-T). Thereafter, blots were incubated for 2 h at room temperature with hybridoma culture supernatant containing an anti-fes Moab or with SP2/O-Agl4 culture supernatant. In each case, supernatants were diluted 1:40 in PBS-T. The blots were subsequently incubated for 1 h at room temperature with peroxidase-conjugated rabbit anti-mouse antiserum (DAKO), diluted 1:250 in PBS-T. After washing with PBS-T (3 x 5 min), blots were incubated with 4-chloro-l-naphtol.

Immunocytochemistry and isotype determination Subconfluent monolayers of ST-FeSV-CCL64 and CCL64 cells were fixed with cold acetone/ methanol for 5 sec and, thereafter, rinsed with PBS-T and dried. The cells were incubated with one of the anti-Afes antibodies for 2 h at room temperature. Subsequently, cells were rinsed twice with PBS-T for 10 min and incubated with rabbit anti-mouse FITC-Ig (1:50 dilution). Thereafter, cells were rinsed twice with PBS-T for Table 1. Isotypes of M o a b s and immunoreactivity pattern.

MON-9.1 MON-10 M O N - 11 MON-14 MON-16,5 MON-17.3 M O N - 18 MON-23 MON-30 MON-34 MON-36 MON-37

Isotype

p92 r~

p85 g~g-fe~

IFA

IgG1 IgG2a IgG 1 IgG2b IgG1 IgG1 lgG 1 IgG1 lgG1 IgG1 lgG1 IgG1

+ + + + + + + + + + + +

nt nt nt nt + + + + nt nt + + nt + + + +

+ + nt nt nt + + + + nt nt nt nt nt nt

+

+ + + + + +

Detection of p92 fes and p85 s~g-fes is based on Western blot analysis. + + : positive staining; + : weak staining; nt: not tested

21 10 min for 5 min and coverslipped. In control experiments, MON-100 was used, which is a monoclonal antibody with specificity for the neuroendocrine 7B2 protein [21]. The isotype of the Moabs was determined using a dipstick isotyping kit (Holland Biotechnology BV, Leiden, The Netherlands).

In vitro transcription and translation analysis The transcription vector pSP64 was used for in vitro transcription of an 0.75 kbp Pvull/EcoRI fes cDNA according to the guidelines of the manufacturer (Promega). The obtained RNA was translated in an in vitro translation experiment using a rabbit reticulocyte lysate according to the

guidelines of the manufacturer (Amersham). 358methionine labeled in vitro translation products were analyzed using standard immunoprecipitation and SDS-PAGE procedures [23].

Results and discussion

The objective of these studies was to develop a panel of monoclonal antibodies that specifically recognize the catalytic domain of the human fes proto-oncogene product. To target immunoreactivity towards the selected fes protein domain, we opted for a bacterially produced hybrid protein as immunogen. The pATH and pUR prokaryotic expression systems were selected for our studies; the pATH system to produce a hybrid protein for

Fig. 2. SDS-PAGE analysis of bacterially produced hybrid proteins. Proteins are stained with Coomassie Blue. (A) Lysates of bacteria transformed with pATH11 (lanes 1 and 2) or with pJG44 (lanes 3 and 4); protein synthesis was either induced (lanes 2 and 3) or not (lanes 1 and 4) as described in Materials and methods. In lane 5, partially purified hybrid protein trpE-z~ffes is shown. (B) Lysates of bacteria transformed with pU292 (lanes 1 and 2) or with pJG19 (lanes 3 and 4); protein synthesis was either induced by IPTG (Danes 2 and 4) or not (lanes 1 and 3). In lane 5, partially purified hybrid protein flgal-Afes is shown.

22 immunization and the pUR system to facilitate hybridoma screening. An 0.95 kb fes cDNA[ 14] comprising the protein kinase features such as the phosphotransfer motif, the tyrosine residue (y713) which serves as a phosphoacceptor site during autophosphorylation, but not the presumed ATP binding region [ 16], was used in the construction of the trpE-Afes and flgal-Afes hybrid proteins. Details of the construction strategy is outlined in Fig. 1. It should be noted that in this scheme only those vectors of the two expression systems are shown that were directly involved in the generation of the correct fusion genes. Nucleotide sequences analysis of the Afes insert confirmed its nature and composition (data not shown). Expression studies of the trp E-Afes and lac ZAfes fusion genes in bacteria, revealed that upon their induction, relatively large quantities of the hybrid proteins were produced. In Fig. 2A (lanes 1 and 2), proteins synthesized by bacteria transformed with pATH 11 DNA are shown after SDS-PAGE and Coomassie Blue staining. It is

clear that, upon induction (lane 2), relatively large quantities of an additional protein are synthesized. Similar high levels of an apparently induced protein (Fig. 2A, lane 3) were produced by bacteria transformed with pJG44 (Fig. 2A, lanes 3 and 4). From these results, we concluded that the induced proteins represented trpE (lane 2) and the hybrid protein trpE-Afes (lane 3). In similar experiments, bacterial production of flgal (Fig. 2B, lane 2) and the hybrid protein flgal-Afes (Fig. 2B, lane 4) was demonstrated. Based upon the electrophoretic mobility of the hybrid proteins during SDS-PAGE, the molecular weights of trpE-Afes (Fig. 2A, lane 3) and flgal-Afes (Fig. 2B, lane4) could be calculated as 55kDa and 135kDa, respectively. This is in accordance with the expected molecular weight of the hybrid proteins. The Afes-encoded polypeptide, starting at amino acid residue K 612 and containing 211 residues, has a calculated molecular weight of about 23 kDa. In the hybrid proteins, the trpE polypeptide portion is about 37 kDa and

Fig. 3. Detection by Western blot analysis of p92 fes in lysates of the promyelocytic cell line HL60. Proteins in HL60 cell lysates were subjected to SDS-PAGE and, thereafter, electroblotted onto nitrocellulose. Blots were analyzed with monoclonal antibody MON-17.3 (lane 1), MON-30 (lane 2), MON-27 (lane 3), or SP2/O-Agl4 cell culture supernatant as a control (lane 4). The p92 fes

23 that of flgal, 116 kDa. flgal-Afes seemed less stable than trpE-Afes. Both hybrid proteins could be obtained in a partially purified form (Fig. 2A, lane 5 and Fig. 2B, lane 5), however, trpE-Afes gave consistently higher yields in the purification procedure. For the development of Moabs directed against the catalytic domain of the fes proto-oncogene, female Balb/c mice were immunized intraperitoneally with partially purified trpE-Afes and after completion of the immunization scheme, spleen cells were fused to Sp2/O-Ag14 myeloma cells. Antibodies produced by hybrodomas were classifted initially according to their reactivity pattern in ELISA and Western blot analysis (Materials and methods). A total of 39 hybridomas reacted positively in the initial screening. Of these, twelve were selected for further analysis and they are listed in Table 1. To further define these twelve Moabs, it was tested whether they recognized the product of the fes proto-oncogene, p92 reS. Therefore, Western blot analysis was performed of proteins synthesized by HL60 cells. Based upon Northern blot analysis (data not shown), this human promyelocytic cell line is known to express the fes proto-oncogene. In Fig. 3, the results with Moabs MON-17.3 (lane 1), MON-30 (lane 2), and MON-37 (lane 3) are shown. As can be seen in Table 1, all Moabs recognized the p92 fes protein. In similar Western blot experiments, five Moabs were tested for their reactivity towards the viral polyprotein p85 gag-fes encoded by the Snyder-Theilen strain of feline sarcoma virus. From these studies, it appeared that they all recognized the polyprotein p 8 5 gag-fes (data not shown). It should be noted that some of the Moabs recognized a protein of about 60 kDa in addition to the p85 gag'fes polyprotein. The nature of this protein is not clear at the moment. It is possible that it represents a proteolytic cleavage product of the polyprotein, for instance a product that is generated after cleaving off the gag-geneencoded portion (p 15-p 12) from the polyprotein. To establish that the reactivity of the Moabs was indeed directed towards the catalytic domain of the human fes proto-oncogene product, in vitro transcription and translation experiments were

carried out and the proteins obtained in this way were analyzed by immunoprecipitation analysis. Experimental details are described under Materials and methods. In Fig. 4 (lanes 1 and 2), it is shown that the in vitro produced fes RNA directs the synthesis of a 17 kDa protein. The molecular weight of this protein corresponds to the molecular weight calculated on the basis of human fes nucleotide sequence data. Immunoprecipitation analysis using anti-fes monoclonal antibody MON-14 indicated that the 17 kDa protein is recognized by this Moab. Other Moabs of the panel also recognized the 17 kDa protein (data not shown) but not control SP2/O-Agl4 cell culture

Fig. 4. In vitro transcription and translation analysis of fes cDNA sequences. An 0.75 kb Pvull/EcoRI fragment [ 14] of the 0.95 kb fes cDNA clone was molecularly cloned into pSP64 and corresponding RNA was obtained in an in vitro transcription reaction. Using SDS-PAGE analysis, the in vitro translation products obtained with this RNA (lane 1) were compared to a control experiment where no RNA was added (lane 2). Thereafter, immunoprecipitation analysis of the proteins was performed using anti-fes monoclonal antibody MON-14 (lane 3) or control SP2/O-Agl4 cell culture supernatant (lane 4). The 17 kDa fes product is marked by arrows. Molecular weight markers are indicated.

24 supernatant (Fig. 4, lane 4). These results clearly establish that the epitope that is recognized by MON-14 is located within the last 154 amino acid residues of the fes-encoded protein. The question remains whether all Moabs recognize the same epitope in the protein kinase domain of the fes proto-oncogene product or that there are divers epitopes in that region of 154 amino acid residues. To establish whether the panel Moabs could also be used in immunocytochemical studies, CCL64 cells transformed by ST-FeSV were studied. In Fig. 5, the staining patterns of ST-FeSVCCL64cells with MON-16.5 (Fig. 5A), MON17.3 (Fig. 5B) and MON-9 (Fig. 5C) and control MON-100 (Fig. 5D) are given. All three Moabs gave a cytoplasmic staining under the experimen-

tal conditions used, although the pattern obtained with MON-9 was not as homogeneous as with the other two. The subcellular localization of the viral fes product has not yet clearly been established. Studies have indicated that the ST-FeSVencoded polyprotein is tightly associated with plasma membranes [21]. Myristylation of the amino-terminus of the FeSV gag-encoded protein might be responsible for this property [ 19]. However, biochemical fractionation and immunofluorescence studies [3, 26] suggest that a portion of the viral fps gene product (fps is the avain counterpart of mammalian fes) is associated with the plasma membrane but that the majority is located in the cytoplasm. From another fractionation study, the contrary was concluded [ 11 ]. Our resuits obtained by direct immunocytochemical

Fig. 5. Immunocytochemical analysis of ST-FeSV-CCL64 cells with the anti-fes Moabs MON-16.5 (A), MON-17.3 (B), MON9.1, (C) or MON-100 as a control.

25 analysis seem to support the observation that the ST-FeSV-encoded protein is predominantly a cytoplasmic protein. Finally, we conclude from our studies that we have generated a panel of well defined monoclonal antibodies, which specifically react with the catalytic domain of the human fes protooncogene product and which are functional in a variety of immunological applications. This opens new perspective for further studies on the human p92 fes protein. At present, the human fes protooncogene product has not yet been purified to homogeneity for further studies on its structure and enzyme activity. A major problem in this context is that physiological expression levels of the enzyme are very low. The newly developed monoclonal antibodies are presently used to purify relatively large quantities of the human p92 fes protein by affinity chromatography. As soon as the purified p92 fes protein is available, experiments will be performed to find out whether the newly developed antibodies may inhibit the kinase activity of the enzyme. At present, they are the best candidates for such future studies.

Acknowledgements We acknowledge the excellent technical assistance of J.J.M. van Groningen. This work was supported by the Inter-University Network for Fundamental Research, which is sponsored by the Belgian Government and by the Dutch Cancer Society (KWF).

References 1. Adam SA, Nakagawa T, Swanson MS, WoodruffTK & Dreyfuss G (1986) Mol. Cell. Biol. 6:2932-2943 2. Barbacid M, Beemon K & Devare SG (1980) Proc. Natl. Acad. Sci. USA 77:5158-5162

3. Feldman RA, Wang E & Hanafusa H (1983) J. Virol. 45: 782-791 4. Feldman R, Gabrilove J, Tam J, Moore M & Hanafusa H (1985) Proc. Natl. Acad. Sci. USA 82:2379-2383 5. Greer P, Maltby V, Rossant J, Bernstein A & Pawson T (1990) Mol. Cell. Biol. 10:2521-2527 6. Hanafusa H (1988) In: Reddy EP, Skalka AM & Curran T (Eds) The Oncogene Handbook (pp 39-57) Elsevier, Amsterdam 7. MacDonald I, Levy J & Pawson T (1985) Mol. Cell. Biol. 5:2543-2551 8. Mathey-Prevot B, Hanafusa H & Kawai S (1982) Cell 28: 897-906 9. Messing J & Vieira J (1982) Gene 19:269-276 10. Moran MF, Koch CA, Sadowski I & Pawson T (1988) Oncogene 3:665-672 11. Moss P, Radke K, Carter VC, Young J, Gilmore T & Martin GS (1984) J. Virol. 52:557-565 12. Pawson T (1988) Oncogene 3:491-495 13. Roebroek AJM, Schalken JA, Verbeek JS, Van den Ouweland AMW, Onnekink C, Bloemers HPJ & Van de Ven WJM (1985) EMBO J. 4:2897-2903 14. Roebroek AJM, Schalken JA, Bussemakers, MJG, Van Heerikhuizen H, Onnekink C, Debruyne FMJ, Bloemers HPJ & Van de Ven WJM (1986) Mol. Biol. Rep. 11: 117-125 15. Roebroek AJM, Schalken JA, Leunissen JAM, Onnekink C, Bloemers HPJ & Van de Ven WJM (1986) EMBO J. 5:2197-2202 16. Rtither U & Muller-Hill B (1983) EMBO J. 2:1791-1794 17. Samarut J, Mathey-Prevot B & Hanafusa H (1985) Mol. Cell. Biol. 5:1067-1072 18. Sanger F, Nicklen S & Coulson AR (1977) Proc. Natl. Acad. Sci. USA 74:5463-5467 19. Schulz A & Oroszlan S (1984) Virology 133:431-437 20. Smithgall T, Yu G & Glazer R (1988) J. Biol. Chem. 263: 15050-15055 21. Snyder HW & Singhal MC (1983) Cancer Invest. 1: 225232 22. Spindler KR, Rosser DSE & Berk AJ (1984) J. Virol. 49: 132-141 23. Van de Ven WJM, Reynolds FH & Stephenson JR (1980) Virology 101:185-197 24. Van Duijnhoven JLP, Ayoubi TAY, Timmer EDJ, Braks AAM, Roebroek AJM, Martens GJM & Van de Ven WJM (1989) FEBS Let. 255:372-376 25. Veronese F, Kelloff GJ, Reynolds FH, Hill RW & Stephenson JR (1982) J. Virol. 43:896-904 26. Woolford J & Beemon K (1984) Virology 135:168-180