The Purified Product of the Transforming Gene of Avian Sarcoma Virus Phosphorylates Tyrosine*

The product of the avian sarcoma virus transforming gene (sm) is a phosphoprotein of 60,000 daltons ( ~ ~ 6 0 " ~ ) which is responsible for the oncogenic potential of the virus. Recent findings indicate that this protein pos- sesses an affiliated protein kinase activity. We have determined by hydrodynamic measurements and gel filtration that this kinase activity tracks with a highly asymmetric molecule of 60,000 daltons, strengthening the idea that pp6OSm alone (as opposed to a complex) possesses the enzymatic activity. To more fully characterize the properties of this kinase activity, we undertook its purification by two independent methods. In each case, a protein related to pp6OSm was extensively purified from contaminating cellular proteins. The yields from one of the procedures were sufficient to induce high titer monospecific anti- bodies against pp60"" in mice. We have shown that purified ~ ~ 6 0 " ~ is able to phosphorylate several protein substrates other than IgG. The conclusion that pp6OSm possesses the responsible enzymatic activity was strengthened by demonstrating that a temperature-sensitive conditional mutation in src affected the ther- mal stability of the purified protein. It has recently been shown that the protein kinase activity affiliated with pp60am phosphorylates tyrosine residues on IgG. We have examined the target specific- ity of the purified protein on several substrates other than IgG, and show that in every case, the phosphoryl- ation occurs exclusively at a tyrosine residue; it therefore appears that tyrosine phosphorylation is not an artifact of phosphorylation in the immunoprecipitate,

The Purified Product of the Transforming Gene of Avian Sarcoma Virus Phosphorylates Tyrosine* (Received for publication, May 28, 1980) Arthur D. Levinson The product of the avian sarcoma virus transforming gene (sm) is a phosphoprotein of 60,000 daltons (~~6 0 "~) which is responsible for the oncogenic potential of the virus. Recent findings indicate that this protein possesses an affiliated protein kinase activity. We have determined by hydrodynamic measurements and gel filtration that this kinase activity tracks with a highly asymmetric molecule of 60,000 daltons, strengthening the idea that pp6OSm alone (as opposed to a complex) possesses the enzymatic activity.
To more fully characterize the properties of this kinase activity, we undertook its purification by two independent methods. In each case, a protein related to pp6OSm was extensively purified from contaminating cellular proteins. The yields from one of the procedures were sufficient to induce high titer monospecific antibodies against pp60"" in mice. We have shown that purified ~~6 0 "~ is able to phosphorylate several protein substrates other than IgG. The conclusion that pp6OSm possesses the responsible enzymatic activity was strengthened by demonstrating that a temperaturesensitive conditional mutation in src affected the thermal stability of the purified protein.
It has recently been shown that the protein kinase activity affiliated with pp60am phosphorylates tyrosine residues on IgG. We have examined the target specificity of the purified protein on several substrates other than IgG, and show that in every case, the phosphorylation occurs exclusively at a tyrosine residue; it therefore appears that tyrosine phosphorylation is not an artifact of phosphorylation in the immunoprecipitate, but instead represents the general substrate specificity of pp6OSm.
Genetic evidence indicates that the oncogenic potential of avian sarcoma virus resides in a single viral gene termed src (Vogt, 1977;Hanafusa, 1977). The product of this gene is a 60,000-dalton phosphoprotein (Brugge and Erikson, 1977;Pur-chi0 et al., 1978;Levinson et al., 1978; which is located in the plasma membrane of ASV' infected cells (Willingham et al., 1979;Krueger et al., 1980;Courtneidge et al., 1980). A closely related protein, pp60Pr0f"~s", occurs in uninfected cells from a wide variety of vertebrates Oppermann et al., 1979;Karess et al., 1979; and is presumed to be encoded by a cellular gene (sarc or proto-src).
Recent work has prompted the conclusion that pp60"" may function as a protein kinase: immunoprecipitates prepared with extracts of ASV-transformed cells and anti-pp60"" serum catalyzed the transfer of labeled phosphate from [y-:"P]ATP to the heavy chain of IgG Levinson et al., 1978). Chemical analysis of the IgG phosphorylated in the immune complex kinase reaction has shown the site of phosphorylation to be a tyrosine residue (Hunter and Sefton, 1980), an unexpected result in view of the apparently low incidence of such a modification (Taborsky, 1974;Uy and Wold, 1977). Moreover, cells transformed by ASV contain 7fold more phosphotyrosine than do their untransformed counterparts (Hunter and Sefton, 1980). As a result of these observations, it was suggested that pp60""' may function in vivo as a phosphotyrosine kinase (Hunter and Sefton, 1980). However, since phosphorylation of IgG in the LC. assay is an in vitro reaction occurring under unusual conditions, we cannot be certain that the newly described phosphorylation of tyrosine accurately reflects the activity of pp6OSrc in infected cells. Indeed, because the LC. kinase displays properties unlike those of known protein kinases (Richert et al., 1979), it has been suggested that the observed activity may not reflect the action of an authentic protein kinase, but rather may represent an in vitro anomaly which obscures a more fundamental function (Richert et al., 1979). Furthermore, while the above experiments clearly demonstrate a relationship between I.C. kinase and the presence of pp60""' in the immunoprecipitate, it has not been conclusively established that the activity is an inherent property of the pp60"" molecule itself, as opposed perhaps to such an activity binding tightly and specifically to pp60"".
In an attempt to address these issues, we determined by hydrodynamic measurements and column chromatography that the kinase is associated with a highly asymmetric molecule of 60,000 daltons. Furthermore, we undertook the purification of pp60"" by two separate and independent methods. In each case, the protein was extensively purified. It could be demonstrated that the LC. kinase activity copurified with a protein related to pp6Os", a result also reported recently by Erikson et al. (1979). The purified protein could phosphorylate several exogenously added substrates while in solution, and in each instance, phosphorylation occurred exclusively at tyrosine residues.

EXPERIMENTAL PROCEDURES
General Procedures-In two previous publications (Levinson et al., 1978;Oppermann et a[., 1979), we have described our procedures for the propagation and isotopic labeling of cultured cells, the preparation of antisera from rabbits bearing tumors induced by the 11973 Schmidt-Ruppin strain of ASV (TBR serum), the immunoprecipitation of virus-specific proteins, the assay of pp60""-associated immune complex kinase activity, and the fractionation of proteins by XISpolyacrylamide gel electrophoresis. T2 NRK cells (rat kidney fibroblasts transformed by SR-ASV) were kindly provided by L. Turek (National Cancer Institute). NRK cells transformed by SR-A tsNY68 ASV (Kawai and Hanafusa, 1971) and its parent were kindly provided by S. Kawai (Institute of Medical Science, University of Tokyo). All reagents were from Sigma unless otherwise indicated.
The lysate was applied to 15 to 20% linear sucrose gradient in Buffer B (Buffer A without NP40) and centrifuged for 60 h at 36,000 rpm in an SW41 rotor at -2°C. Twenty-one fractions were collected and analyzed for LC. kinase activity. Marker proteins were visualized by SDS-polyacrylamide gel electrophoresis of samples from each fraction.
Determination of Stokes Radius-Stokes radius was determined by gel filtration using Sephedex G-150 superfine (Pharmacia). A clarified lysate of Tz NRK cells was prepared in the same way as for determination of sedimentation coefficients. Standards used included catalase (52 A), bovine serum albumin (35 A), ovalbumin (28 A), and cytochrome c (17 A) (Lindblad et al., 1977). The extract was applied to and eluted from a packed column (1 X 60 cm) of G-150 superfine Sephadex equilibrated in Buffer B. Thirty-five fractions were collected and alternate fractions assayed for I.C. kinase activity and for markers by SDS-polyacrylamide gel electrophoresis. The elution of calalase was monitored by direct enzyme assay using hydrogen peroxide.
Purification of I.C. Kinase Activity by Ion Exchange Chromatography-Tz NRK cells (-1 X 10'") from 50 roller bottles were collected, washed with phosphate-buffered saline (0.15 M KC1, 10 mM NaP04, pH 7.2), and frozen at -7OOC until needed. Extracts were prepared by lysis of cell packs in sufficient Buffer C (0.25 M KCI, 1 mM MgCI, 0.1 mM EDTA, 0.1% hexamethylphosphoramide, 0.1 mM dithiothreitol, 1% NP40, 25 mM 2-(N-morpholino)ethanesulfonic acid (Calbiochem), pH 6.2) to maintain protein at -15 mg/ml. Insoluble material was removed by centrifugation at 20,000 X g for 1 h at 2°C. The clarified lysate was batch-processed on DEAE-52 (Whatman) (equilibrated in Buffer C). Material not binding to the resin was adjusted to 10% ethylene glycol and loaded onto a hydroxylapatite (Clarkson Chemical Co.) column (2 X 20 cm) preequilibrated with Buffer c . Following loading, the column was washed with Buffer C (100 ml) and eluted with a 400-ml linear salt gradient (0 M to 0.13 M NaHP04, pH 7.2) in Buffer C at a flow rate of 12 ml/h. Fractions containing the peak of LC. kinase activity were pooled and dialyzed overnight against Buffer D (0.04 M KCl, 1 mM EDTA, 1 mM MgCl, 0.1% hexamethylphosphoramide, 0.1 mM dithiothreitol, 1% NP40, 25 mM 2-(N-morpho1ino)ethanesulfonic acid, pH 6.2). This material was then chromatographed on a carboxylmethyl cellulose 50 (Pharmacia) column (1 X 15 cm) preequilibrated with Buffer D. Elution was performed with a 160-ml gradient (0.04 to 0.5 M KC1) in Buffer D. Fractions containing the peak of I.C. kinase activity were pooled and dialyzed overnight against Buffer D. This material was then chromatographed on a Cibacron blue agarose (Bethesda Research Laboratories) column (1 X 5 cm) preequilibrated with Buffer D. Elution was performed with a 100-ml gradient (0.04 to 0.45 M KC1) in Buffer D. All chromatographic steps were performed at 4°C. Selected fractions from all 3 columns were monitored for conductivity. Protein was monitored by the Bio-Rad protein assay (Bio-Rad, Richmond, Cal.). Proteins were iodinated from appropriate fractions using iodogen as described (Devare and Stephenson, 1977).
Immunization of Mice with p52"Approximately 5 pg of purified p52 obtained from the Cibacron blue agarose column were administered 3 times at 2-week intervals to 6 Balb/c mice. The first injection was introduced subcutaneously in complete Freunds. The second injection was administered intraperitoneally in incomplete Freunds, and the final injection was administered subcutaneously in incomplete Freunds. Mice were bled from the tail at various times for analysis of serum content.
Purification of pp60"' by Antibody Affinity Chromatography-IgG from normal and TBR serum was purified from serum (preheated for 15 min at 50°C) by precipitation (twice) with 45% ammonium sulfate, followed by chromatography on DEAE-cellulose and dialysis against coupling buffer (0.5 M NaCI, 0.1 M borate, pH 8.4). The IgG was coupled to CNBr-activated Sepharose 4B (Pharmacia) at a concentration of 2 m g / d of resin in coupling buffer. The reaction was terminated by addition of neutralized ethanolamine to 0.5 M and the coupled resin was extensively washed before use. The purification of pp60"" was achieved by applying clarified cellular lysates to the column, followed by extensive washing with 0.3 M KCI, 0.1 mM EDTA, 1% NP40, 0.05% deoxycholate, 20 mM NaP04, pH 7.2. Specifically bound material was eluted with 0.8 M NaSCN in 20 mM Tris, pH 7.4, and assayed immediately following dilution of the NaSCN to below 10 mM. No appreciable loss of activity attends this brief exposure to NaSCN. All manipulations involving pp60"" were performed at 4°C.
Assay Procedures-LC. kinase activities were measured as described (Levinson et al., 1978), using limiting amounts of pp60'". Under these conditions, incorporation of radio-labeled phosphate into IgG is linear with respect to pp60""' concentration (Richert et al., 1979). For soluble protein kinase assays with purified pp6ObrC, substrates (usually 1 to 10 fig) were incubated in 0.1 ml of reaction mixture (0.1 M KCI, 10 mM MgCl, 1 PM of ATP, 20 mM buffer (piperazine-N,N"bis(2-ethanesulfonic acid)), pH 6.8) containing purified pp60"" and 1 X lofi cpm of (y"P]ATP (40 Ci/mmol) (New England Nuclear). The reaction was typically terminated by dilution into SDS sample buffer (Laemmli, 1970) and then heated to 90°C for 10 min prior to analysis by SDS-gel electrophoresis. Tubulin (chicken), kindly provided by D. Cleveland, was heated before use to 80°C for 5 min to eliminate endogenous protein kinase activity.
Identification of Phosphorylated Amino Acids-Proteins to be analyzed were typically obtained from SDS-gels and processed as described (Beeman and Hunter, 1978). Amino acids were separated by thin layer chromatography at pH 3.5 (Hunter and Sefton, 1980). Phosphoserine and phosphothreonine (Sigma) were run as internal standards and visualized by ninhydrin staining. We hydrolyzed "Plabeled IgG of TBR serum phosphorylated with pp60"' in order to obtain phosphotyrosine as an additional standard (Hunter and Sefton, 1980). prepared from ASV4ransformed rat cells and subjected to rate-zonal gradient centrifugation or gel filtration. I.C. kinase activity was monitored as described under "Experimental Procedures." A, rate-zonal centrifugation. Lysate was layered on a 5 to 20% sucrose gradient and centrifuged. IgG (7.0 S), bovine serum albumin @SA) (4.3 S), ovalbumin (Oua) (3.6 s), and cytochrome c (Cyt. c) (1.9 S) were included as internal standards. Fractions were collected and assayed for LC. kinase activity. The bands of phosphorylated heavy chain of IgG were excised from the gel and counted in a scintillation spectrometer. B, gel filtration on Sephadex G-150 superfine. Lysate was applied to a column (30 x 0.7 cm) of Sephadex. Catalase (Cat), bovine serum albumin, ovalbumin, and cytochrome c were included as internal standards. Fractions were collected and assayed for I.C. kinase activity.

Properties
weight of the protein to be 62,000 (Siegal and Monty, 1966).
The frictional coefficient was calculated similarly ( f/fo = I.@), indicating the protein to be highly asymmetric. Because the calculated molecular weight of the I.C. kinase (62,000) agrees well with the molecular weight of pp60"" (60,000 as determined by independent methods), we attribute the I.C. kinase activity to pp60" alone, rather than to a complex between pp60"" and a cellular enzyme. In addition, it is clear that the enzymatically active pp60"" behaves largely as a monomer in solution, in agreement with the data obtained by Maness et al. (1979).
Purification of the I.C. Kinase Activity by Zon Exchange Chromatography-Our purpose in extensively purifying the I.C. kinase activity was 2-fold. We fist sought to characterize the protein(s) which copurified with the kinase. Second, we required a sufficient supply of purified material for more thorough in vitro analysis of this enzymatic activity.
We chose as starting material a continuous line of ASVtransformed T2 NRK cells. Extracts were prepared from 50 g of these cells. Since the I.C. kinase assay is the most sensitive and simplest method of detection of pp60"" (Levinson et al., 1978), the purification was monitored with this procedure. The extract was fractionated according to the scheme outlined under "Experimental Procedures": the results of sequential chromatography on hydroxylapatite, carboxymethyl cellulose, and Cibacron blue agarose are shown in Fig. 2. Column chromatography fractions were analyzed for total protein and I.C. kinase activity. We succeeded in purifying the I.C. kinase activity -650-fold (this calculation was based on the increase of specific activity of the enzyme during purification and could not allow for any presently unrecognized activation or inactivation that might have occurred).
Fractions from the Cibacron blue agarose column were iodinated to visualize the proteins by SDS-polyacrylamide gel electrophoresis. Fig. 3A illustrates that while a heterogenous collection of proteins was applied to the Cibacron blue agarose column, a prominent protein of 52,000 daltons (p52) was specifically bound. To determine if this protein bore any relationship to pp6OSrc, the bound fraction was immunoprecipitated with control serum, serum directed against ASV structural proteins, and serum from two rabbits bearing ASVinduced tumors (TBR serum). Fig. 3B demonstrates that at least a portion of '"I-labeled p52 is immunoprecipitated by TBR serum (Lanes c and d), but not by either normal or anti-ASV serum (Lanes a and b).
Because of the specific reactivity of p52 with anti-pp6Os" sera, this 52,000-dalton protein seemed likely to represent a proteolytic digestion product of pp60"". To test this hypothesis, 6 BALB/c mice were immunized with the pooled fractions containing p52 from the Cibacon blue agarose column (see "Experimental Procedures"). Preimmune and immune sera were then tested for their ability to react with pp60"". To optimize the detection of antibodies directed against contaminating proteins, we tested the antisera with labeled extracts prepared from the same cell line as was used for the purification of the I.C. kinase activity. Serum from mice immunized with the purified protein was capable of specifically reacting with a protein of 60,000 daltons (Fig. 4, Lanes c, d, and e) that co-migrated with pp60""' immunoprecipitated with conventionally prepared TBR serum (Fig. 4, Lane g). In addition, the titer of the mouse sera to this 60,000-dalton protein increased dramatically following multiple injections of the purified protein (Fig. 4, Lanes a to c).
To establish unambiguously the nature of the protein precipitated by the mouse serum, we compared its peptide map with that of pp60" by limited proteolysis using Staphylococcus aureus V8 protease (Cleveland et al., 1978). trates that the 60,000-dalton protein immunoprecipitated by the mouse sera is identical with pp60"'. Moreover, the V8 protease map of total p52 was indistinguishable from that of the immunoprecipitated material (data not shown). We are thus confident that at least a major portion of p52 is indeed related to pp6OsrC, generated most likely from the latter by proteolysis. Because p52 co-purifies with I.C. kinase activity through fractionations on 4 columns, it appears that the kinase activity resides on p52"".
Purification of pp60"" by Affinity Chromatography-We have repeatedly been unable to achieve the large scale purification of intact pp60"" by conventional methods, even when any of a wide variety of protease inhibitors were included throughout the manipulations.* While the purification of p52"" addresses several fundamental issues (see "Discussion"), the in vitro properties of p52"" may not faithfully reflect those of pp60"". We therefore sought an alternate method of purification which would minimize proteolytic degradation of pp60"".
We attempted this purification of pp60"" by immunoaffinity * A. D. Levinson  " B I labeled protein could be immunoprecipitated by TBR, but not by normal serum (Fig. 6, Lanes e and f ) . We thus conclude that immunoaffinity chromatography provides a one-step procedure resulting in extensive purification of pp60""'. Presumably because of the rapidity of this method, little degradation of pp60src to p52""' occurs. Phosphoprotein Kinase Actioity of Purified pp60"rv-To determine whether purified pp60""' can function as a protein kinase in a soluble assay system, the kinase activity of affinity- Immunoprecipitation of pp60"" by serum from mice immunized with p52. Clarified lysates were prepared from ASVtransformed rat cells labeled with [ "Slmethionine as described under "Experimental Procedures." Samples were then immunoprecipitated with preimmune serum from mouse 5 ( a ) , serum from mouse 5 obtained after first immunization ( h ) , serum from mouse 5 after second immunization (c), serum from mouse 4 obtained after second immunization ( d ) . serum from mouse 6 obtained after second immunization ( e ) , normal rabbit serum ( f l , and T R H serum (g). Immunoprecipitated proteins were then visualized by SDS-polyacrylamide gel electrophoresis and autoradiography.

26K-
chromatography. This method is rapid and typically results in extensive purification. We therefore prepared an IgG-Sepharose resin, using TRR serum as the source of IgG. As Fig. 6 illustrates, when extracts of cells labeled with either ['""SImethionine (Lane C ) or :"€' (lane d) were passed over the column, a phosphoprotein of 60,000 daltons was retained which could be eluted with 0.8 M NaSCN. When an extract labeled with ["%]methionine was passed over a column containing normal IgG coupled to Sepharose, no such protein bound to the resin (Fig. 6, Lane h). The affinity purified ,"?P- purified pp60"" was tested with a variety of protein substrates. Phosphorylation of TBR IgG was observed, but not normal IgG (Fig. 7 A ) . In contrast to the immunoprecipitated activity observed with crude extracts , however, the purified enzyme was now able to phosphorylate exogenously added substrates, presumably because pp60""' was not immobilized by antibody. Of 14 proteins tested, we found only 3 that served as substrates for this kinase activity. The most active acceptor among the proteins tested was tubulin, a cytoplasmic protein which is the major structural element of microtubules. Fig. 7A illustrates that while both the a and p subunits of tubulin served as substrates, the p subunit was preferred. Also effective, although less so, was casein. Proteins which did not detectably serve as substrates included bovine serum albumin, actin, tropomyosin, myosin, cytochrome c, ovalbumin, phosvitin, and any of the histones (data not shown).
T o determine whether phosphorylation of substrate proceeded with the kinetics expected of a conventional enzymesubstrate reaction, we determined the time course of phosphorylation of tubulin under standard conditions using purified pp60"". In contrast to the virtually instantaneous reaction kinetics observed during LC. phosphorylation (Levinson et al., 1978;Rubsamen et al., 1979), phosphorylation of tubulin proceeded in a time-dependent fashion for periods up to 1 h (Fig. 7B). Analysis of this reaction has permitted determination of kinetic parameters: a K,, for ATP of 12 p~ and a V,,,,, of 40 nmol of tubulin phosphorylated/min/mg of pp60""' were observed. The K , approximates a previous value obtained with the LC. kinase assay (Rubsamen et al., 1979). By contrast, the V,,., is appreciably lower than might be expected from the properties of other, better characterized protein kinases (Kemp et al., 1977;Bylund and Krebs, 1975).
The phosphotransfer activity in the LC. reaction can use both ribo-and deoxyribonucleoside triphosphates as donors of phosphate (Richert et al., 1979). Our purified enzyme displayed the same unusual property, transferring phosphate from dATP with an efficiency comparable to that found with ATP (data not shown). The activity of the purified protein was unaffected by CAMP, in agreement with the report of Erikson et al. (1979).
To more directly implicate pp60""' in the phosphorylation of soluble substrates, immunoaffinity purified pp60""' was prepared from clones of NRK cells transformed with either wild type virus (SR-A) or tsNY68 SR-A virus. The latter contains a mutation in src which has multiple effects on the structure and function of pp60""': transformation of infected cells is temperature-sensitive according to all biological parameters examined (Kawai and Hanafusa, 1971); phosphorylation of pp60""' in oiuo is abnormally thermolabile (Levinson et al., 1978;; and the I.C. kinase activity of the mutant pp60""' is inactivated by heat far more rapidly than the activity of the wild type protein Levinson et al., 1978). We found that purified pp60""' derived from tsNY68-transformed NRK cells was much more thermolabile in its ability to phosphorylate tubulin than was pp60""' purified from cells transformed by wild type ASV (Fig.  8). Since the only identified lesion in tsNY68 is located in src, our findings further implicate pp60""' in the kinase reaction.

Identification of the Amino Acid
Phosphorylated by pp60""'"Following the demonstration that IgG becomes phosphorylated at a t-yrosine residue in the LC. kinase reaction, the suggestion was made that pp60""' may function in I 1t~'o . as a phosphotyrosine kinase (Hunter and Sefton, 1980), although such an enzymatic activity has no known precedent (Taborsky, 1974;Uy and Wold, 1977). It was therefore of obvious interest to determine the substrate specificity of pp60""' when  (1 p g ) , ( c ) I'BH IgG (1 p g ) , ( f ) casein (10 p g ) , and ( h ) tubulin (10 p g ) . Lanes d, e, and g are as in Lanes c, f. and h, respectively, without the inclusion of pp60'"'. Lane i is tubulin, stained with Coomassie blue in order to visualize the a (56 K) and p (53 K) subunits. Lanes e and f (casein substrate) were exposed for 12 h; all other lanes were exposed for 2 h. B, time course of tubulin phosphorylation. Purified pp60'" was incubated in the standard reaction mixture with tubulin for ( a ) 0 min, ( h ) 2 min, (c) 6 min, (d) 18 min. or ( e ) 54 min, at 30OC. Samples were then analyzed for radioactivity by SDS-polyacrylamide gel electrophoresis.

FIG. 8. Heat inactivation of pp60"" kinase activity. pp60'"
was purified (by immunoadsorbtion) from SH-ASV-transformed rat cells and from tsNY68-transformed rat cells, and then heat inactivated at 4OoC for (a) 0 min, ( h ) 1 min, ( c ) 2 min, ( d ) 4 min, ( e ) 8 min, or ( f ) I6 min. Ten micrograms of substrate (tubulin) were then added in the standard reaction mixture and the reaction was allowed to proceed at 30°C for 60 min. Samples were then withdrawn and analyzed by SDS-polyacrylamide gel electrophoresis for phosphate incorporation into tubulin. unencumbered by antibody immobilization. T o test this, we acid-hydrolyzed proteins that had been phosphorylated by purified pp60""' and then fractionated the resulting free amino acids by thin layer chromatography (Hunter and Sefton, 1980). Using IgG labeled in the LC. kinase reaction as the source of phosphotyrosine for reference ( Fig.   9, Lane h) (Hunter and Sefton, 1980), we found that purified pp60""' phosphorylated TBR IgG exclusively at tyrosine residues (Lane c). Lanes d to f in Fig. 9 illustrate that all other exogenously added substrates of pp60"" which were tested also become phosphorylated at tyrosine residues. In addition, we confirmed the recent report that pp60""' labeled in rliclo with "'P contains both phosphoserine and phosphotyrosine residues (Lane a ) (Hunter and Sefton, 1980). Pr76, the polyprotein precurser of ASV structural proteins, apparently is phosphorylated in vivo only a t serine residues (Lane g). It thus appears that phosphorylation at tyrosine residues is not an artifact of the LC. kinase assay but rather is representative of the substrate specificity displayed by pp6O""',in vitro.
As mentioned above, pp60"" is itself phosphorylated at tyrosine and serine residues in vivo (Hunter and Sefton, 1980); moreover, mutations in src that affect its kinase activity also affect phosphorylation of the tyrosine residue Oppermann et al., 1979;Hunter and Sefton, 1980). It has therefore been proposed that pp60""' may phosphorylate itself  9. Identification of the amino acid phosphorylated by pp60"". Purified pp60"' was incubated under standard reaction conditions with a variety of substrates. The reaction was terminated and analyzed by SDS-polyacrylamide gel electrophoresis. Appropriate protein bands were then excised and processed as described under "Experimental Procedures" for analysis of phosphorylated amino acids by thin layer chromatography. a, pp60""-labeled in oiuo with ."I>; h, IgG phosphorylated in the LC. kinase assay; c, IgG phosphorylated by purified pp60'" (soluble assay); d, casein phosphorylated by purified pp60'". e, a-tubulin phosphorylated by purified pp60""', f, 8-tubulin phosphorylated by purified pp60'", g, 1'1-76 labeled with in V~L J O . P-serine and P-threonine markers were included as internal standards and are indicated as such.
by means of either an intra-or intermolecular reaction (Erikson et al., 1979 Hunter andSefton, 1980). Using partially purified pp6OSr", Erikson et al. (1979) have observed phosphorylation of pp60""' which they believe to be autocatalytic. We, by contrast, have failed to detect self-phosphorylation of pp6OS"', either in the absence of exogenous substrates (Fig. 7, Lane a ) or in their presence (Lanes 6, f, and h). We cannot presently account for these conflicting results.

I s the I.C.
Kinase Activity an Inherent Property of pp60""'?-Previous evidence has implicated pp60"" in the I.C. kinase reaction Levinson et al., 1978, but the possibility remained that pp60"". tightly and specifically bound a cellular enzyme responsible for phosphorylation of IgG. If this latter possibility were correct, the active complex between pp60""' and the cellular enzyme would display an aggregate molecular weight greater than 60,000. Since the size of proteins can be accurately determined from a knowledge of their sedimentation coefficient and Stokes radius (Siegal and Monty, 1966), we determined these values for the I.C. kinase activity by rate-zonal centrifugation and gel fitration. In both cases, the activity behaved as a sharply migrating species (Fig. l ) , allowing accurate determination of the necessary parameters. From this knowledge, a M,. = 62,000 was calculated for the I.C. kinase activity. We therefore conclude that pp60""' is alone responsible for I.C. phosphorylation. This conclusion is consistent with data obtained by purification of the I.C. kinase activity.
Purification by ion exchange chromatography yields a protein of 52,000 daltons (p52) (Fig. 3). Much of p52 seems to be related to pp60"" by several criteria. First, it could be immunoprecipitated by TBR sera, but not by control serum. Anti-ASV serum also fails to precipitate p52, ruling out the possibility that p52 bears any relationship to virion structural proteins. Second, the V8 protease maps of immunoprecipitated and total p52 demonstrated a chemical relationship of these proteins to pp60"". Third, the injection of as little as 5 pg of p52 into mice resulted in the production of apparently monospecific antibodies which immunoprecipitated pp60""' (Figs. 4 and 5). There is no evidence in any of the immunized mice for production of antibody against proteins unrelated to pp60"" (Fig. 4). It should be emphasized that this experiment was performed using p52 purified from rat cells, serum from mice injected with p52, and "'S-labeled extracts prepared from the rat cells used originally for purification of p52. This experimental design maximizes the opportunity to detect antibodies against proteins unrelated to pp60""' that might contaminate the purified material.
By peptide map analysis, we have recently established that p52 is derived from pp60""', most likely by proteolytic removal of 8000 daltons of protein from the NHr terminus of pp60""'.:' Since p52 can be generated by prolonged incubation of cell lysates (Courtneidge et al., 1980). it appears that the bulk of purified p52 results from proteolysis of pp60""' during the course of purification.
An apparently analogous protein is usually recovered from extracts of infected cells by immunoprecipitation with TBR serum.:' The recoveries of p52 in this manner are erratic and often quite low. We therefore suspect that the presence of p52 in both crude and purified preparations is an artifact, although the protein might also be an intermediate in the physiological breakdown of pp60""'.
Purification by immunoaffinity chromatography yields pp60""' which is active in the I.C. kinase assay (Figs. 6 and 7), a result also reported recently by . Presumably because purification in this fashion is much less time consuming than purification by ion exchange chromatography, preparations of pp60""' can be obtained which contain little or no ~52""'. Since pp60""' purified in this manner is not significantly contaminated with cellular proteins (Fig. 6). this provides an independent demonstration of the affiiation of pp60""' with I.C. kinase activity.
It should be noted that the availability of purified pp60""' (or ~5 2 '~' ) in quantities sufficient for immunization allows the preparation of monospecific antibodies directed against pp60""' in a variety of animals. T h e use of mice for the immunizations will facilitate the production of hybridoma clones producing antibody specific for pp60"". Monoclonal antisera should prove to be invaluable reagents for detailed analysis of the structure and function of pp60""'.
Purified pp60""' Phosphorylates Substrate Proteins at Tyrosine Residues-Recent evidence from Hunter and Sefton (1980) has established that the phosphorylation of TBR IgG by extracts containing pp60""' occurs exclusively at tyrosine residues. This observation was surprising, since protein kinases typically phosphorylate proteins at the hydroxyl groups of serine and threonine residues (Rubin and Rosen, 1975;Krebs and Beavo, 1979). One could conclude from this result that either pp60"' is a novel protein kinase, or that I.C. phosphorylation, representing a highly unusual reaction, was failing to faithfully reflect the in uivo properties of 60""'. We have shown that purified preparations containing pp60""' are able to phosphorylate a variety of protein substrates, and that this phosphorylation invariably occurs a t tyrosine residues (Figs. 7 and 9).
T o more directly implicate pp60"" in these reactions, pp60"" was purified from cells infected with wild type ASV, and from cells infected with a strain of ASV carrying a temperaturesensitive lesion in pp60""' (tsNY68 ASV). It is clear that the phosphorylation of substrate (tubulin) is much more thermolabile when pp60"" is encoded by tsNY68 ASV than when it is encoded by wild type ASV. Since the only identified lesion in tsNY68 ASV is located in src, these findings strongly implicate pp60""' in the kinase reaction. This conclusion, in turn, lends support to the claim that pp60""' may be functioning in vivo as a phosphotyrosine kinase. It should be emphasized, however, that the possibility the protein possesses ad-'' H. Opperman, A. D. Levinson ditional enzymatic properties, or displays additional specificities, has not been excluded.
We have compared published amino acid sequences of proteins (Dayhoff, 1979) with their ability to serve as substrates for pp60"". In our experience to date, no direct relationship seems to exist between tyrosine content and acceptor ability. In general, acidic proteins seem to serve as more efficient substrates than basic proteins. A more thorough analysis of various acceptors will be required to extend these generalizations.
Implications of in Vitro Phosphorylation by pp60"""The observation that denatured tubulin serves as a relatively efficient substrate for src kinase may have little or no physiological significance. In accord with previous reports (Sloboda et al., 1975), our tubulin contains an endogenous protein kinase activity which must be heat-inactivated before use as substrate. Since proteins are often more active as substrates for protein kinases following their denaturation (Bylund and Krebs, 1975), it is not clear at this time whether native tubulin would even serve as substrate for pp60"". For example, we have failed to detect acceptor activity in other cytoskeletal elements such as actin, myosin, and tropinin2 Since protein kinases are typically somewhat indiscriminant in their substrate specificity in uitro, attempts to predict specific in vivo targets from results in uitro must be regarded with caution.
The phosphorylation of exogenous substrates by pp60"" is linear with respect to time for periods of at last 1 h (Fig. 7), as expected for a conventional enzyme-substrate reaction. Our results indicate that, as a kinase, pp60"" is much less efficient at phosphorylating substrates than are other protein kinases. The significance of this is presently uncertain; perhaps pp60"" has an unusually strict substrate specificity.
Our initial search for kinase activity associated with pp60"" was prompted by the hypothesis that pp60"" was phosphorylating itself in a crude in vitro assay system (Levinson et al., 1978). We have, however, consistently failed to observe any phosphorylation of pp60" in soluble kinase assays involving the purified protein, in contrast to the results obtained by Erikson et al. (1979). We cannot presently account for this discrepancy.
The protein kinase activity associated with pp60"" has greatly enhanced our ability to analyze the nature and function of the protein. We (Levinson et al., 1978) and  have suggested that phosphorylation of cellular proteins might be responsible for neoplastic transformation by ASV. The present report and the results of others Hunter and Sefton, 1980) are in accord with this proposal, but by no means constitute definitive evidence for the hypothesis. Moreover, the number and nature of in vivo targets of pp60"" remain enigmatic, although recent work indicates that a cellular protein of 36,000 daltons (pp36) may be phosphorylated by pp60"" (Radke and Martin, 1979).
Although pp60"" is affiliated with the plasma membrane in infected cells (Willingham et al., 1979;Kreuger et al., 1980;Courtneidge et al., 1980), its targets in vivo may not be so associated. Our findings indicate that pp60"" is a highly elongated molecule (see "Results") which is apparently anchored to the plasma membrane through an NH2-terminal domain, the remainder of the molecule being exposed to the cytoplasm: The exposed domain seems to represent most of p52, which we have purified and shown to have protein kinase activity. Thus, the otherwise unwanted breakdown of pp60"" during large-scale purification has fortuitously provided us with evidence that the enzymatically active domain of pp60"" may function by affecting substrates located in the cytoplasm s. Courtneidge and A. D. Levinson, unpublished observations. rather than within the substance of the plasma membrane.
We conclude that pp60"" has the properties expected of a cyclic nucleotide-independent phosphotyrosine kinase, although its activity on the arbitrary substrates which we have used does not approach conventional levels for protein kinases. The kinase activity of pp60"" is unusual in at least two regards: both ribo-and deoxyribonucleoside triphosphates can serve as donors of phosphate; and the recipient of the phosphate is tyrosine. Phosphorylation of tyrosine has also been attributed to proteins encoded by the transforming genes of at least two other tumor viruses: Abelson murine leukemia virus (Witte et al., 1980) and polyoma virus (Eckhart et al., 1979). These findings raise the possibility that neoplastic transformation by different viruses may follow a common design. Rigorous characterization of the enzymatic properties of purified pp6OS" should assist in the test of this hypothesis.