Analysis of the Catalytic Domain of Phosphotransferase Activity of Two Avian Sarcoma Virus-transforming Proteins*

Proteolytic digestion of the transforming protein of Rous sarcoma virus (pp60"") with trypsin, chymotrypsin, or thermolysin generated a 29,000-dalton frag- ment representing the carboxyl half of this molecule. This proteolytic fragment was able to phosphorylate pp60*'"-specific immunoglobulin as well as exogenous substrates such as angiotensin, casein, and tubulin. When quantitated on a molar basis, the protease-re-sistant fragment of pp60"" had a greater specific activity than the intact enzyme. Digestion of pp9OYes, the transforming protein of Y73 sarcoma virus with these proteases yielded a peptide of similar molecular weight which was capable of autophosphorylation as well as the phosphorylation of exogenous substrates. The pro- teolytic fragments of both pp60"'" and pp90Y"displayed the same strict specificity for phosphorylation of tyrosine as the intact enzymes. These results indicate that the 29,000-dalton carboxyl end of pp60"'" and pp9OYe" can function independently as phosphotransferases and indicate that the catalytic domains of these molecules have a conformation which confers protection against limited conditions of proteolysis. assays were performed using 50 p1 of alternating column fraction, 4 pl of 50 mM angiotensin or 2 pl of TBR serum, and 10 p1 of 50 mM Tris, pH 7.5, 5 mM MgC12, and 5 pCi of (Y-~~P)ATP for 20 min at room temperature. The angiotensin phosphorylation reaction was terminated by heating at 90 "C for 5 min and assayed by spotting 5 pl of the reaction mix on Whatman 3MM paper. The samples were electrophoresed at 4000 V for 1.5 h. Phosphorylated angiotensin was visualized by autoradiography and quantitated by liquid scintillation spectrometry after excision from the paper.

' The abbreviations used are: RSV, Rous sarcoma virus; TPCK, L-1-tosylamido-2-phenylethyl chloromethyl ketone; TBR, tumor-bearing rabbit; NaDodS04, sodium dodecyl sulfate; PTR, protease-resistant fragment. ~~ ~~ ~~~ ~~~~ which encode tyrosine kinases. In addition, the biochemical properties of these phosphotransferases are very similar and there is considerable overlap in the protein substrates which are phosphorylated on tyrosine in vivo after transformation by viruses encoding these related gene products (7,14,15). In addition, normal cells contain highly conserved genes which are homologous to these viral transforming genes, and it is believed that the viral genes were acquired from the cellular genome by recombination (7). All of these lines of evidence suggest that the tyrosine kinase-transforming genes and their cellular homologues represent a multigene family, possibly derived from a single ancestral gene. The greatest degree of homology among these tyrosine kinases is present in one domain of these proteins which is found in the carboxyl half of all of the viral tyrosine-specific phosphotransferases except the ab1 gene product (where this domain is NH,-terminal). Previous genetic and biochemical studies have provided evidence which suggests that this highly conserved domain of these transforming proteins possesses the catalytic site of phosphotransferase activity. Firstly, mutations which inactivate the protein kinase activity of these proteins map in this domain (16)(17)(18), and secondly, a proteolytic fragment from this domain was shown to phosphorylate IgG in the immune complex protein kinase assay (19). In this report, we further probe this domain of pp60"" and the related gene product pp9oy'" encoded by the Y73 sarcoma virus. It is shown that a 29,000-dalton-fragment from the carboxyl domain of pp60"" and pp90YeS is highly resistant to digestion by a variety of proteases and can function as a tyrosine-specific protein kinase. These results suggest that sequences in the NH,-terminal half of pp60"" and pp90ye" are not necessary for phosphotransferase activity and that structural and functional properties of the catalytic domain may be shared among different tyrosine kinase-transforming proteins.
Analysis of Proteolytic Fragments of pp60"" and pp9Wes 3T3 cells were obtained by transformation of mouse Balb 3T3 cells with the SR-D strain of RSV using polyethylene glycol as described by Kawai (20). Serum TBR serum was prepared from rabbits bearing tumors induced by the SRD strain of RSV from L. Rohrschneider as described previously (1). Monoclonal antibody to ppl9"g was provided by D. Boettiger. Whole medium from the hybridoma cells were used for immunoprecipitation. Monoclonal antibodies to pp60"" were prepared from hybridoma cells obtained by fusion of myeloma cells with spleen cells from a mouse immunized with pp60"" produced in Escherichia coli (21). Purified IgG was used for the immunoprecipitation experiments. Details of the preparation of these reagents are described elsewhere (22). Antiserum to mouse immunoglobulin was obtained from Miles.

Immunoprecipitation and Sample Analysis
Cells were washed three times in 0.15 M NaCI, 0.01 M Tris hydrochloride (pH 7.2) and lysed a t 4 "C in modified RIPA buffer (0.15 M NaCI, 1% sodium deoxycholate, 1% Triton X-lOO,O.l% NaDodSO,, 0.01 ethylene glycol bis(8-aminoethy1ether)-N,N,N',N'-tetraacetic acid), and the lysate was clarified a t 40,000 X g for 30 min. The lysate was incubated with antibody for 60 min, and the immune complexes were adsorbed to formalin-fixed Staphylococcus aureus bacteria (23) for 15 min. In experiments using monoclonal antibodies to pp6OSr", anti-mouse IgG was included in the reaction to enhance binding of the monoclonal antibody to the immunoadsorbent. The bacteria were washed three times with RIPA and either incubated with sample buffer for direct analysis or treated further as described in the figure legends. The samples were analyzed on 7.5 or 10% NaDodS04polyacrylamide gels as described previously (I). The fixed and stained gels were dried and the radiolabeled proteins visualized with the aid of DuPont Lightning Plus intensifying screens using Kodak XAR film or incubated with 2,5-diphenyloxazole-dimethyl sulfoxide for fluorographic enhancement of 3H as described (24). The dried gels were then exposed to XAR film.
Digestion with Proteolytic Enzymes I) Immunoprecipitates-After three washes with RIPA buffer, the immune complexes were further washed with 100 mM NaC1, IO mM Tris, pH 7.5 (NaC1:Tris). The pellet was then resuspended in either 50 p1 of 50 mM ammonium carbonate containing TPCK-treated trypsin or chymotrypsin, or 10 mM Tris, pH 7.2,2 mM CaCI, containing thermolysin at the concentration indicated for 10 min at room temperature. The reaction was terminated with sample buffer or aprotinin as indicated.
2) Immunoaffinity Column-A pool of nine monoclonal antibodies directed against pp60"" (22) was covalently cross-linked to protein A-Sepharose as described (25). pp60"" was bound to the 200 pl of affinity beads by incubation of a RIPA buffer lysate from 1 X lo8 cells with the beads for 45 min a t 4 "C. The beads were washed with 30 ml of RIPA, 10 ml of 100 mM NaC1, 10 mM Tris, pH 7.2, and incubated in solution with 400 GI of ammonium carbonate and 20 pg,/ml of TPCKtrypsin for 10 min a t room temperature. The protease-released material was removed after sedimentation of the beads a t 1000 X g for 5 min. The column was washed with 50 mM diethylamine, pH 11, to remove bound antigens and neutralized with 100 mM sodium phosphate.

Partial Proteolysis
The protein to be analyzed was excised from a dried gel, rehydrated, and re-electrophoresed on a 12.5% NaDodSO,-polyacrylamide gel in the presence of the Staphylococcus V 8 protease as described (5) without the 60-min incubation between the stacking and separating gels.
21 TBR-IgG-Phosphorylation of IgG was performed by incubating the immune complexes with 5 pCi of (T-~'P)ATP in 10 mM Tris, pH 7.5, 5 mM MgC12 for 10 min a t 0 "C as described (2).

Gel Filtration
Three hundred microliters of trypsin-treated pp60"" (described above) were loaded onto a 100-cm column containing Bio-Gel P-60, fine grade, equilibrated with 10% glycerol. 0.5-ml fractions were collected a t a flow rate of 2.5 ml/h. The column was calibrated using 300 pg of bovine serum albumin, ovalbumin, chymotrypsinogen, and cytochrome c. The elution of these proteins was monitored by adsorbance at Azm. Phosphotransferase assays were performed using 50 p1 of alternating column fraction, 4 pl of 50 mM angiotensin or 2 pl of TBR serum, and 10 p1 of 50 mM Tris, pH 7.5, 5 mM MgC12, and 5 pCi of ( Y -~~P ) A T P for 20 min a t room temperature. The angiotensin phosphorylation reaction was terminated by heating a t 90 "C for 5 min and assayed by spotting 5 p l of the reaction mix on Whatman 3MM paper. The samples were electrophoresed at 4000 V for 1.5 h.
Phosphorylated angiotensin was visualized by autoradiography and quantitated by liquid scintillation spectrometry after excision from the paper.

RESULTS
Proteolytic Cleavage of pp60""" Fig. 1 displays the products generated by enzymatic cleavage of the proteins immunoprecipitated from 32P-labeled RSV-transformed mouse cell lysates with serum from a rabbit bearing a tumor induced by Rous sarcoma virus (TBR serum). The major protein immunoprecipitated by TBR serum is pp60"", the product of the RSV src gene. Other minor phosphoproteins precipitated by this serum included Pr76, the viral gag-gene translation product, and pp52, a proteolytic product of pp60"" generated during cell lysis. Incubation of this immune complex with trypsin resulted in the digestion of all of these protein species ( Fig. l A , lunes 2-7). A single polypeptide of M , = 29,000 was resistant to further cleavage even at concentrations of 2 mg/ ml. This polypeptide fragment appeared to be derived from pp60"'" since the level of 32P incorporation into this species was greater than that of the minor phosphoproteins precipitated by TBR serum. In addition, this PTR fragment was not generated after digestion of proteins precipitated nonspecifically with normal rabbit serum; and preincubation of TBR serum with disrupted virus to block the precipitation of viral structural proteins (in this case Pr76) did not prevent the production of the 29-kDa PTR fragment (data not shown).
Enzymatic digestion with either chymotrypsin ( species which migrated slightly faster than the trypsin-generated PTR fragment (lunes 4-6). These results suggest that a 29,000 fragment of immunoprecipitated pp60"'" is protected from digestion by these three proteolytic enzymes.
In Fig. lA, the products of proteolytic digestion which were released from the immunoadsorbent (lunes 2, 4, 6,8,10,12) were separated from the proteins which remained bound to the bacterial adsorbent (lunes I, 3,5,7,9,11,13). Digestion with trypsin released the majority of the PTR fragment from the immune complex whereas only 30-50% of the PTR was released by chymotrypsin. Examination of the Coomassie blue-stained gel from this experiment indicated that the After the addition of the hacterial immunoadsorbent, the immune complexes were washed three times as described under "Experimental Procedures" and incubated with the enzymes designated above for 10 min at room temperature. One pl of aprotinin was then added to each reaction and the immune complex "bound" proteins separated from the protease "released" peptides by sedimentation of the bacterial immunoadsorbents ( A ) . In H, this separation was not performed. Each sample was incubated with sample buffer, boiled for 60 s, and electrophoresed on a 7.5% gel, autoradiography 24 h. amount of the PTR fragment found in the supernatant after protease treatment was proportional to the amount of IgG dissociated from the immunoadsorbent. (No reduction in the mobility of the IgG heavy chain was detectable after proteolytic digestion.) We have found that approximately half of the PTR fragment sediments with IgG on a glycerol gradient, indicating that some of the PTR molecules are still bound to IgG after release from the immunoadsorbent (data not shown). Protein Kinase Actioity-In order to determine whether the protease-resistant fragment described in Fig. 1 is active as a protein kinase, we examined the ability of the trypsin-treated p p 6 P fragments to phosphorylate TBR-IgG. It has been shown previously that pp60"" can phosphorylate the heavy chain of IgG from TBR serum when bound in an immune complex (2,4). Fig. 2R displays the phosphorylation of TBR-IgG after treatment of immunoprecipitated pp60"" with trypsin. In order to approximate the per cent of pp60"" converted to the PTR fragment, the cells were labeled for 16 h with ['HI lysine. After immunoprecipitation by normal rabbit serum or TBR and incubation in the presence or absence of trypsin, IgG in the trypsin-treated immunoprecipitates (lane 4 ) was approximately double that found in the untreated sample (lane 3). This result suggests that the trypsin treatment of pp60"" increases its phosphotransferase activity using TBR-IgG as substrate.
T o determine whether the increase in pp60"" activity correlated with the generation of the PTR fragment, we examined IgG phosphorylation after incubation with various concentrations of trypsin (Fig. 3). The fate of pp60"" was followed by polyacrylamide gel electrophoresis of ["HJleucine-labeled pp60"". Treatment with 2.5 pg/ml of trypsin resulted in a loss of 75% of intact pp60"" and no 60-kDa protein was detectable after treatment with 5 or 10 pg/ml of trypsin. An intermediate 47-kDa fragment of pp60"" was detectable in immunoprecipitates incubated with 2.5-5 pg of trypsin. The 29-kDa PTR fragment was detectable in samples treated with 5 pg/ml of trypsin and the levels of this peptide increased with higher doses of trypsin. The increased incorporation of:"P into TBR-IgG closely correlated with the appearance of the PTR fragment. Under conditions which allowed the highest level of :'?P incorporation, there was no detectable 60-kDa or 47-kDa form the samples were treated with the protease inhibitor aprotinin of pp60"". and then divided into two fractions. One-half of the sample Since TBR-IgG is tightly associated with pp60"", this phoswas analyzed directly (Fig. ZA), and the other half was incu-phorylation reaction represents an unusual enzyme-substrate bated with ['"PJATP and M e (Fig. 2B). The relative inten-interaction. In order to examine the phosphorylation of exsity of the pp60"" band was compared to that of the PTR ogenous substrates after trypsin treatment of pp60"", we fragment band by densitometric tracing of the autoradiogram assayed the phosphorylation of the [Val']angiotensin 11. Anin Fig. 2 A . The densitometric quantitation revealed that ap-giotensin has been shown to serve as a substrate for pp60"" proximately 20% of pp60"" is resistant to proteolysis in this in in uitro reactions (27,28). In these assays, pp60"" was assay. Fig. 2R demonstrates that the incorporation of '*P into immunoprecipitated by a monoclonal antibody to pp6OnrC. This  lanes 3 and 4 ) . After the addition of the bacterial immunoadsorbent, the immune complexes were washed three times with RIPA buffer and one time with NaC1:Tris. The samples were then incubated with 50 pl of 50 mM ammonium carbonate in the presence (lunes 2 and 4 ) or absence (lunes 1 and 2 ) of 20 pg/ml of trypsin as described under "Experimental Procedures." After the addition of aprotinin, sample buffer was added to one-half of each sample ( A ) , and the other half of each sample was adjusted to 10 mM Tris, pH 7.2, 5 mM MgC12, and incubated for 10 min of 0 "C with 5 pCi of (y-"P)ATP ( R ) . The samples were analyzed on 10% SDSpolyacrylamide gels. Exposure time: A, 21 days; R, 2 h. The IgG band from lanes 3 and 4 was excised and counted by liquid scintillation spectroscopy (lane 3 , 5600 cpm; lune 4, 114,000 cpm). The intensity of the pp60"" hand from lane 3 was found to he 10-fold greater than the PTR fragment from lune 4 by densitometric analysis. Since lysine residues are distributed evenly throughout pp60"", this indicates that approximately 2 0 5 of pp60"" was resistant to proteolysis. monoclonal antibody does not interfere with the phosphorylation of angiotensin, and the PTR fragment generated after trypsin digestion of immune complex-bound pp60"" appears to be identical to that found after treatment of TBR-precipitated pp60"". Table I shows the incorporation of 32P into angiotensin after incubation of immunoprecipitated pp60"" in the presence or absence of trypsin. The trypsin-treated immunoprecipitate phosphorylated angiotensin to 5-fold higher levels than the untreated immunoprecipitated sample. Phosphoamino acid analysis of both samples revealed that phosphotyrosine was the only phosphorylated amino acid. All of the negative control samples had background levels of '"P binding to the filters.
In both the TBR-I& and angiotensin phosphotransferase reactions, the activity of pp60 was undiminished or stimulated 2-to 5-fold by proteolytic digestion under conditions in which intact pp60 was not detectable. This result suggests that the Immunoprecipitates containing pp60"" were prepared and treated with trypsin as described in Fig. 2 except that either 0,2.5,5, 10, or 20 pg/ml of enzyme was used in each of five reactions. One-half of each sample was directly analyzed on a NaDodS04-polyacrylamide gel to follow digestion of pp60"", and one-half was incubated with (y-"P)ATP to assay IgG phosphorylation. IgG phosphorylation and levels of pp60"", pp47'", and PTR quantitated by excision of the protein band from the gel and liquid scintillation spectroscopy.

TABLE I
Phosphorylation of angiotensin after trypsin treatment of pp60"" A lysate from SRD-3T3 cells ( 5 X 10' cells) was prepared as described in Fig. 5 and with anti-mouse IgG (1P:Control) or 40 pl of IgG from hybridoma cell line 273 with anti-mouse IgG (IP:anti-pp60n" A and €3). The washed immunoprecipitates were incubated in 0.05 M ammonium carbonate in the presence (+) or absence of trypsin (20 pglml) as described under "Experimental Procedures." After the addition of 1 pl of aprotinin to each sample, the reaction was adjusted to 10 mM Tris, pH 7.2, 5 mM MgCI, and incubated with 5 pCi of [y-R'P]ATP in the presence (+) or absence (-) of 3 mM [Val']angiotensin I1 as described under "Experimental Procedures." The reaction was then adjusted to 4% trichloroacetic acid. The '"P-labeled acid-soluble peptide was assayed by the filter-binding procedure descrihed under "Experimental Procedures." IP:anti-ppGV"-B represents a different experiment in which only monoclonal 273 was used for immunoprecipitation. In order to ensure that the counts per min detected in this assay did not represent phosphorylation of proteins present in the immunoprecipitate which were released from the immunoadsorbent during the phosphorylation reaction and were not precipitated by trichloroacetic acid, the samples were also electrophoresed on a polyacrylamide gel. All of the radioactivity moved with the dye front indicating that counts detected in this assay represented phosphorylation of angiotensin. protease-resistant fragment(s) of pp60 has a higher specific activity than the intact molecule. Gel Filtration of Trypsin-treated p p 6 P " I n each of the above phosphotransferase assays the reactions were carried out on the total digestion products of the immunoprecipitates. In order to distinguish which trypsin-generated fragment was responsible for the activity detected in these assays, the trypsin-treated fraction was fractionated on a Bio-gel P-60 column (Fig. 4). The sample was prepared for fractionation by trypsin treatment of pp60"" bound to an immunoaffinity column containing covalently bound monoclonal antibodies directed against pp60"". This procedure generated proteolytic fragments which were free of immunoglobulin molecules. The only phosphate-containing fragment which is released by trypsin comigrates was the 29-kDa PTR fragment. Fig. 4 shows the elution profile of the phosphotransferase activity present in this trypsin eluate using either TBR-IgG or angiotensin I1 as substrates. The peak fractions of activity using either substrate eluted a few fractions before chymotrypsinogen (Mr = 25,7001, indicating that the active fragment has a molecular weight of approximately 29,000-30,000. Since this is the estimated molecular weight of the only proteaseresistant fragment detectable on NaDodS0,-polyacrylamide gels ( Figs. 1 and 2), this suggests that this PTR fragment is responsible for the enzymatic activity detected after proteolysis. Using the PTR fragment fractionated on this column we have shown that this fragment of pp60"'" can also phosphorylate tubulin and casein and retains the strict specificity for phosphorylation of tyrosine using these proteins as substrates (data not shown). Trypsin Resistance of pp9OY'""The transforming protein of Yamaguchi 73 (Y73) sarcoma virus also displays tyrosinespecific protein kinase activity. Although the viral gene encoding this protein was derived from a cellular gene distinct from the cellular homologue of the u-src gene, the amino acid sequence of pp90ye" shares considerable amino acid homology with pp60"" (12). In order to determine whether p p 9 P contains a protease-resistant domain similar to the PTR fragment of pp60"", cell lysates of 32P-labeled Y73 transformed cells were immunoprecipitated with monoclonal antibody to the gag gene-encoded protein pp19. p p 9 P is precipitated by anti-ppl9 since the Y73-transforming protein is a chimeric   FIG. 4. Gel filtration analysis of trypsin-digested pp60"". SRD-3T3 cells (3 X 10' cells) were lysed in RIPA with 5 mM 8-mercaptoethanol, and the clarified lysate bound to an immunoaffinity column containing covalently linked monoclonal antibody to pp6O"''. A trypsin digest of the bound pp60"'" was prepared as described under "Experimental Procedures" and loaded onto a Bio-Gel P-60 column equilibrated with P-60 buffer. 50 p1 from alternating fractions were collected and assayed for phosphorylation of TBR-IgG (0---0) or angiotensin (M) as described under "Experimental Procedures." -30 protein encoded by a fused gene containing a portion of the viral gag gene-linked to yes gene-specific sequences. pp90yes was the only detectable protein precipitated with this antibody (Fig. 5 ) . Incubation of pp90Yes with trypsin resulted in the generation of a polypeptide which comigrated with the PTR fragment from pp60"" (lanes 2 and 4). This suggests that the conformation of this domain of pp90ye* is very similar to that of pp60"" in that both domains are protected from proteolytic digestion. Since pp90Ye8 was immunoprecipitated with antibody to p19, it is clear that antibody to the transformation-specific regions of pp90Yes is not necessary for the protease protection demonstrated in these assays. In order to determine whether the PTR fragment from pp90y"" was able to phosphorylate exogenous substrates, angiotensin I1 phosphorylation was assayed. Table I1 shows that the trypsintreated sample phosphorylated angiotensin to 2-fold greater levels than the untreated samples. The percentage of pp9OY' " resistant to proteolysis was approximately 10-20% (data not shown). Thus, trypsin treatment of pp9OY' " increases the phosphotransferase of this molecule similarly to that found for pp60"". These results suggest that the conformation of pp60"'" responsible for protease resistance is conserved in pp90ye" and that this region of each molecule can function independent of the remainder of the molecule as a protein kinase.
Peptide Mapping of the PTR Fragments- Fig. 5B shows the partial proteolytic digestion pattern of intact pp60"" and pp90"" and their respective PTR fragments with Staphylococcus V8 protease (V8). It has been shown previously that reelectrophoresis of pp60"" with low concentrations of V8 protease produces a single cut in pp60"" which generates a 34-kDa phosphoserine-containing fragment derived from the amino end of pp60"" and a 26-kDa phosphotyrosine-containing fragment from the carboxyl end of pp60"" (30,31). Incubation of the pp6P-derived PTR with V8 protease generated a fragment which comigrated with the COOH-derived 26-kDa V8 fragment of pp60"'" (Fig. 5 , lane I). This result suggests that the PTR fragment is derived from the COOH domain of pp60"" (9). Incubation of pp90 with V8 protease produces a fragment which comigrates with the 26-kDa fragment of pp60"" (lane 4). This fragment would be produced if the V8 protease cleaved p p 9 P a t the Glu residue corresponding to  3T3 cells (A, lanes I and 2 ) or Y73-transformed cells (A, lanes 3 and 4 ) were labeled with "P as described in Fig. 1, lysed with RIPA, and the clarified lysate incubated with monoclonal antibody to pp60"" (273) (SRD-3T3 cells) or monoclonal antibody to p19 (Y73-chicken cells). The washed immunoprecipitates were incubated with 0.05 M ammonium carbonate in the presence (lanes 2 and 4 ) or  absence (lanes I and 3)

TABLE I1
Phosphorylation of angiotensin after trypsin treatment of p p 9 P A lysate from Y73-transformed chicken cells was incubated with 500 pl of media from a nonproducer hybridoma cell line (IPControl) or from hybridoma cells secreting monoclonal antibody to pplSW (IP:anti-pplSRaR). Angiotensin phosphorylation was assayed as in Table I. IPanti:pp19RaR-B represents a different experiment in which only anti-pp19RaR media was used for the immunoprecipitation. " __ G1u331, the proposed V8 cleavage site in pp60"" (9). Incubation of the pp9oy'" PTR fragment (lunes 2 and 3) with V8 protease generated a fragment which comigrated with the COOHderived 26-kDa fragment of pp60"". Since the 26-kDa peptides from the PTR fragments as well as those derived from intact pp60'" and pp90"" proteins comigrated, it would appear that few, if any, amino acids were cleaved from the carboxyl end of these molecules during the trypsin digestion. The carboxyl origin of the PTR fragments is further demonstrated in Fig.  6 in which pp60"" and pp9Wen were phosphorylated in vitro after immunoprecipitation with monoclonal antibody to pp60'" or pigRaK, respectively. Partial peptide mapping of these proteins with V8 protease revealed that the 26-kDa peptide was specifically labeled in each protein (Fig. 6, lanes 7 and   11). This is consistent with previous reports that autophosphorylation of pp60"" and pp9oy'" takes place in the carboxyl domain of these molecules (29, 32, 33). Digestion of the in vitro labeled immunoprecipitated pp60"" and pp9Dy" proteins (lanes 2 and 3) with trypsin-generated PTR fragments which comigrate with the in vivo labeled PTR fragments and which were cleaved to the 26-kDa carboxyl peptide when re-electrophoresed with V8 protease (lanes 8 and 12).  (lanes 3 and 4 ) were lysed in RIPA, and the clarified lysate incubated with monoclonal antibody to pp60"" (273) (lanes I and 2 ) or to p19W (lanes 3 and 4). After washing the immunoadsorbent-bound immune complexes, they were incubated with (-y-"P)ATP as described under "Experimental Procedures. Phosphorylation by proteolytic fragments of pp60'" and pp90ya. pp60"" and pp9oY" were immunoprecipitated from 5 X 10' SRD-3T3 cells (lanes 1 and 2 ) or YS3-transformed chicken cells (lanes 3 and 4 ) as described in Fig. 6. The washed immunoprecipitates were incubated with 0.05 M ammonium carbonate alone (lanes 1 and  3 ) or with 20 pg/ml of trypsin (lanes 2 and 4 ) as described under "Experimental Procedures." After the addition of aprotinin, ( y -R'P)ATP and 10 mM Tris, pH 7.2, 5 mM MgC12 or 5 mM MnCI2 was added and the reaction performed as described under "Experimental Procedures." Sample buffer was added to terminate the reaction and the samples were electrophoresed as described in Fig. 6. The pp90 hand from lune 3 and the PTR hand from lane 4 were re-electrophoresed on a 12.5% gel in the presence (lanes 6 and 8 ) or absence (lanes 5 and 7) of 50 ng of V8 protease. Lanes 5 and 6, pp90; lanes 7 and 8, PTR. tectable in Fig. 5, which contains the proteins phosphorylated in vitro before cleavage with trypsin. Since the PTR fragment appears to be active as a protein kinase, it is possible that trypsin treatment has rendered some proteins susceptible to phosphorylation by the trypsin-treated protein kinases. The major polypeptide phosphorylated in the trypsin-treated pp90'" comigrated with the in vivo labeled PTR fragment. This band was excised from the gel and re-electrophoresed in the presence of V8 protease. The peptide map of this band (Fig. 7, lanes 7 and 8) is identical to that of the PTR fragment of pp90"" labeled in vitro before trypsin treatment (Fig. 6).
This suggests that the PTR fragment was autophosphorylated in this reaction. A fainter band which migrates slightly slower than the PTRye'" was detectable in the trypsin-treated pp60"" immunoprecipitate. The relationship of this protein and other proteins phosphorylated in this reaction to pp60"" is under investigation.

DISCUSSION
The experiments in this report extend previous analyses of the phosphotransferase activity of the carboxyl domain of pp60"" and another viral tyrosine-specific protein kinase, pp9@ve". It is shown that the COOH-terminal 29,000 daltons of pp60"" can be released as an active phosphotransferase by limited proteolysis with trypsin. This fragment can phosphorylate IgG from serum of animals bearing tumors induced by RSV as well as exogenous substrates such as angiotensin, tubulin, and casein. Proteolysis of pp60"" does not alter the strict specificity of this enzyme for phosphorylation of tyrosine. In each immune complex-bound phosphotransferase assay, proteolytic digestion of pp60"" resulted in higher levels of substrate phosphorylation despite an apparent 5-fold difference in the molar amounts of the PTR fragment relative to intact pp60"". These results suggest that the PTR fragment has a higher specific activity than intact pp60""' and thus raises the possibility that structural constraints on this domain imposed by the NH,-terminal half of pp60"" could regulate the phosphotransferase activity of the carboxyl domain. Indeed, Purchio and co-workers have evidence which suggests that autophosphorylation within the NH2-terminal domain ofpp60"" might increase the specific activity of pp60"' The configuration within the carboxyl half of pp60"" confers protection against limited proteolytic digestion. Analysis of the amino acid sequences in this domain does not reveal any obvious structural features which would account for this behavior. I t is c!ear that the protease resistance of this domain is not dependent on antibody binding since pp90 was immunoprecipitated by antibody to the gag portion of this chimeric protein. In addition, the PTR fragment is generated after precipitation of pp60 with monoclonal antibodies which recognize different epitopes on pp60"" (data not shown). It will be of interest to determine whether loss of protease resistance correlates with the loss of the functional integrity of this domain of pp60"", i.e. do mutations which inactivate the phosphotransferase activity of pp60"" alter the protease-resistant configuration of this substructural domain?
The transforming protein of Y73 sarcoma virus was shown to contain a similar protease-resistant phosphotransferaseactive domain, suggesting that this structural feature is conserved in other tyrosine kinases. In addition, Weinmester and co-workers have reported that limited proteolysis of the Fuginami sarcoma virus pp140KaK-f1'" protein releases COOH-ter-minal29-kDa and 45-kDa peptides which are phosphorylated upon addition of ATP and Mg (38). Within this family of related viral tyrosine kinase-transforming proteins there is extensive homology in the amino acid sequences corresponding to the PTR domains of pp60"" and pp90yee" while sequences outside of this domain show little or no homology with large differences in size. This suggests that the PTR domain of these enzymes might have evolved as separate genetic entities which have become linked with different genetic elements. It is possible that the non-PTR domain of these enzymes could regulate the functional activity of the phosphotransferase domain, possibly affecting the substrate specificity and/or the specific activity of the enzyme. The genetic linkage of the catalytic domain of tyrosine kinases with different regulatory elements would allow for functional diversification of tyrosine-specific protein kinases. For instance, sequences encoding the tyrosine kinase domain of cellular growth hormone receptors might have become genetically linked to sequences which confer hormone responsiveness to the kinase domain. This would be analogous to the family of cellular dehydrogenases which share dinucleotide binding domains and differ in the domains which confer substrate specificity (for review see Ref. 39). In many other enzyme systems, limited proteolytic digestion has been used to separate different domains of activity. In the case of rabbit muscle phosphorylase, subtilisin cleaves the molecule into two fragments, a 30-kDa NH,terminal peptide which carries the regulatory binding site for allosteric effectors and a 70-kDa COOH-terminal fragment containing the catalytic site (36,37).

Analysis of Proteolytic Frag
The isolation of this domain of kinase activity provides the means to investigate various aspects of the regulation of the enzymatic activity of the tyrosine kinase-transforming proteins.