Synthetic peptide substrates for a tyrosine protein kinase.

Immunoprecipitates containing the transforming protein of the avian sarcoma virus, Y73, together with its associated tyrosine-specific protein kinase, have an activity which will phosphorylate the synthetic peptide Lys-Leu-Ile-Glu-Asp-Asn-Glu-Tyr-Thr-Ala-Arg at the tyrosine residue. This peptide corresponds to 10 out of 11 amino acids surrounding the phosphorylated tyrosine in both pp60src and P90, the transforming proteins of Rous sarcoma virus and Y73 virus, respectively. The apparent Km for phosphorylation of the peptide was about 5 mM. A second peptide with the sequence Lys-Leu-Ile-Asp-Asn-Glu-Tyr-Thr-ala-Arg differing from the first peptide only by the absence of the glutamic acid 4 residues from the tyrosine was also phosphorylated, but the apparent Km for the reaction was 40 mM. Several sites of tyrosine phosphorylation in viral transforming proteins have been found to have one or more glutamic acids close to the phosphorylated tyrosine on the NH2-terminal side. Taken together with our in vitro phosphorylation studies, this suggests that the primary sequence surrounding target tyrosines may play a role in recognition of substrates by tyrosine protein kinases, and in particular, that glutamic acid residues on the NH2-terminal side may be important.


Several sites of tyrosine phosphorylation in viral transforming proteins have been found to have one or more glutamic acids close to the phosphorylated tyrosine on the NHz-terminal side. Taken together with our
in vitro phosphorylation studies, this suggests that the primary sequence surrounding target tyrosines may play a role in recognition of substrates by tyrosine protein kinases, and in particular, that glutamic acid residues on the NHz-terminal side may be important.
Tyrosine phosphorylation is a rare protein modification in normal cells: phosphotyrosine accounts for only about 0.05% of total acid-stable phosphoamino acid in protein (1,2). Nevertheless, modification of prot'eins by phosphorylation of tyrosine has been implicated in growth control (3,4), and in regulation of cell shape (5). Furthermore, transformation by a group of RNA tumor viruses appears to involve a perturbation in cellular phosphotyrosine metabolism (see Ref. 6 for review). A common feature of these viruses is that their genomes contain some sequences derived from a normal cell and that these cellular sequences are expressed as part of their transforming proteins. Five classes of virus of this type have been identified containing different cellular sequences. In every case, the viral transforming protein itself has been found to be associated with a tyrosine protein kinase activity. For RSV' there is formal proof that the transforming protein, pp6Obrc, possesses an intrinsic ability to transfer phosphate from ATP to tyrosine in proteins (7, 8). It seems likely that the transforming proteins of the other viruses will prove to be tyrosine protein kinases as well, and that the cellular se-* This investigation was supported by Grants CA 14195, CA 17096, and CA 24845 awarded by the National Cancer Institute. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "aduertisemeat"in accordance with 18  The abbreviations used are: RSV, Rous sarcoma virus; Boc, tbutoxycarbonyl; PIPES, 1,4-piperazinediethanesulfonic acid, quences in these viruses will turn out to be genes which code for tyrosine protein kinases in normal cells. A number of cellular substrates for the viral tyrosine protein kinases have been identified (see Ref. 9 for review). In addition, the viral transforming proteins themselves all have at least one site of tyrosine phosphorylation. The basis on which tyrosine protein kinases select their substrates is therefore of particular interest because of its relevance to viral transformation.
The CAMP-dependent protein kinase and the casein protein kinases appear to recognize their substrates at least partly on the basis of the primary amino acid sequence surrounding the target residue. The serines phosphorylated by the CAMPdependent protein kinase usually have one or two basic amino acids placed close by on the NH2-terminal side (see Ref. 10 for review), In contrast, the serines and threonines phosphorylated by the casein protein kinases often have neighboring acidic residues which are on the NH2-terminal side for casein kinase I and on the COOH-terminal side for casein kinase I1 (see Ref. 11 for review). The identification of a new type of protein kinase-phosphorylating tyrosine has prompted investigation of whether there are primary sequence requirements in the recognition of target tyrosine residues. By partial sequence analysis, we and others have found that there are striking homologies between several of the tyrosine phosphorylation sites in the viral transforming proteins (12-15). They are characterized by the presence of a lysine or arginine 7 residues to the NH2-terminal side of the phosphorylated tyrosine and one or more acidic residues, most commonly glutamic acid, in the intervening distance. Sites of tyrosine phosphorylation in cellular proteins are less evidently homologous to those in the viral transforming proteins, but most of them appear to have a glutamic acid close to the tyrosine on the NHn-terminal side (15). While it is almost certain that there will be secondary and tertiary structure requirements, these primary sequence features may be determinants in the recognition of specific tyrosines by tyrosine protein kinases.
The primary sequence requirements for phosphorylation by the CAMP-dependent protein kinase were assessed directly by using as substrates for the purified enzyme synthetic peptides corresponding in sequence to known phosphorylation sites (16)(17)(18)(19). We have begun to use a similar approach for the tyrosine protein kinases. The only site of tyrosine phosphorylation for which the complete sequence is known is that of pp60"" (13, 15). Accordingly, we have synthesized the peptide Lys-Leu-Ile-Glu-Asp-Asn-Tyr-Thr-Ala-Arg which, with the substitution of a lysine for an arginine in the NH2-terminal position, corresponds to this tyrosine phosphorylation site (20). We have found that the tyrosine protein kinase associated with the transforming protein of the avian sarcoma virus, Y73 (21,22), is able to phosphorylate the tyrosine in this peptide. We have also tested a number of related peptides as substrates to determine the influence of the amino acid sequence embedding the tyrosine on this phosphorylation reaction.

4844
Peptide Substrates for a

EXPERIMENTAL PROCEDURES
Synthesis and Purification of Peptides-The peptide Lys-Leu-Ile-

Glu-Asp-Asn-Glu-Tyr-Thr-Ala-Arg was synthesized using a Beckman
Synthesizer Model 990 as described (23). Boc-alanine, Boc-aspartic acid (benzyl), Boc-asparagine (xanthyl), Boc-glutamic acid (benzyl), Boc-leucine, Boc-lysine (2-chlorocarbobenzoxycarbonyl), Boc-threonine (benzyl), and Boc-tyrosine (2,6-dichlorobenzyl) were used together with Boc-arginine (tosyl) resin ester. Cleavage from the resin and deprotection were performed with HF. Analysis of the peptide material by thin layer electrophoresis and chromatography showed two major ninhydrin-staining components. To purify these two peptides, 200 mg of crude material dissolved in 50 ml of H20 and adjusted to pH 8.9 was applied to a Sephadex QAE-A25 column (35 X 1.5 cm) in the HC03-form equilibrated with 0.01 M NH4HC03, pH 8.9. The peptides were eluted with a linear gradient from 0.01 M NH4HC03, pH 8.9, to 0.5 M NKHC03, pH 8.9, using 500 ml of each buffer. Two main peaks of material absorbing at 280 nm were detected eluting at 0.34 M (I) and 0.25 M (11) salt. A minor peak was observed at 0.23 M (111). These three peaks were pooled separately and lyophilized several times from H20 until a constant weight was attained. Amino acid analysis of these peaks gave the following amino acid compositions normalized to alanine. I: Ala,  (14) followed by high performance thin layer chromatography analysis of the phenylthiohydantoin (25). The peptide, src 11, lacked a glutamic acid residue. Its sequence was shown to be Lys-Leu-Ile-Asp-Asn-Glu-Tyr-Thr-Ala-Arg by automated Edman analysis. The position of the glutamic acid in src I1 was confrrmed by digestion of the peptide with Staphylococcus aureus V8 protease. The digest contained two peptides, one of which was found to be Tyr-Thr-Ala-Arg. The peptide, src 111, lacked both isoleucine and glutamic acid. Digestion of src I11 with S. aureus protease also yielded the peptide Tyr-Thr-Ala-Arg. Therefore src I11 has the sequence Lys-Leu-Asp-Asn-Glu-Tyr-Thr-Ala-Arg.
Several other peptides were derived from src I as follows. The peptide Leu-Ile-Asp-Asn-Glu-Tyr-Thr-Ala-Arg (src IV) was prepared by digestion of 1 mg of src I in 500 p1 of 0.05 M NH4HC03 with 150 pg of L-1-tosylamido-2-phenylethyl chloromethyl ketone-treated trypsin (Worthington, TRTPCK) for 16 h at 23 "C. Following two cycles of lyophilization the digest was separated by electrophoresis at pH 4.7 at 1 kV for 27 min on a 100-pm cellulose thin layer plate (26). Peptides were detected by staining marker strips. The released lysine was well resolved from the other major peptide product which was eluted. Upon acid hydrolysis and analysis of the hydrolysate by twodimensional electrophoresis and chromatography this peptide was found to contain Ala, Arg, Asp, Glu, Ile, Leu, Thr, and Tyr. The peptide Lys-Leu-Ile-Glu-Asp-Asn-Glu-Tyr-Th-Ala (src V) was prepared by digestion of 1 mg of src I in 500 p1 of 0.05 M NH4HC03 with 8 pg of carboxypeptidase B (Sigma, 135 units/mg) for 15 min at 30 "C. These conditions were chosen so that arginine was the only amino acid released. Following boiling to inactivate the enzyme and lyophilization, the digest was resolved by electrophoresis at pH 1.9 at 1.5 kV for 15 min on a 100-pm cellulose thin layer plate (1). The released arginine was well separated from the other major peptide product which was eluted. Upon acid hydrolysis this peptide was ound to contain Ala, Asp, Glu, Ile, Leu, Lys, Thr, and Tyr. The peptide Lys-Leu-Ile-Glu-Asp-Asn-Glu-Tyr (src VI) was pre-50 pg of a-chymotrypsin (Worthington, CD1) for 12 h at 23 "C. After lyophilization, the digest was separated by electrophoresis at pH 5.3 in a buffer containing 5% 1-butanol, 2.5% acetic acid, 5% pyridine for 20 min at 1 kV on a 100-pm cellulose thin layer plate. Peptides were detected by staining marker strips and a peptide with the expected mobility was eluted. Upon acid hydrolysis this peptide was found to contain Asp, Glu, Ile, Leu, Lys, and Tyr.
The peptide Tyr-Thr-Ala-Arg (src VII) was prepared by digesting 500 pg of src I in 200 pl of 0.05 M NH4HC03 with 50 pg of S. aureus V8 protease (Miles Laboratories, Inc., Elkhart, IN) for 12 h at 23 "C. After lyophilization, the digest was resolved by electrophoresis at pH 8.9 at 1 kV for 20 min on a 100-pm cellulose thin layer plate (1). Peptides were detected by staining marker strips, and a peptide with Tyrosine Protein Kinase the appropriate mobility which stained positively for arginine was eluted. Upon acid hydrolysis this peptide was found to contain Ala, . .

"
The DeDtides were all dissolved at 10 m d m l in Hz0 and the pH _ _ adjusted to 7. Their concentrations were checked by measurement of absorbance at 278 nm ( E M = 1300).
Cell Culture ana' Immunoprecipitation-Chick cells transformed with Y73 virus were grown as described previously (14). The Y73 virus-transforming protein, P90, was isolated from lysates of Y73 virus-transformed chick cells by immunoprecipitation with a rabbit antiserum raised against disrupted virions of the Prague strain of RSV (the kind gift of Jim Neil, the University of Southern California) (14). The titer of this serum had been determined, and the immunoprecipitates were prepared in antigen excess with about 5 X lo5 cells/ pl of serum. The immune complexes were collected by absorption to S. aureus (Pansorbin, Calbiochem) using 500 pg/pl of serum and washed according to our usual procedure at 4 "C (27). After a final wash in Tris-buffered saline, the S. aureus pellet was resuspended in a buffer containing 20 m~ PIPES, pH 7.0, 10 mM MnC12, using 10 pl for each microliter of precipitating serum and stored in liquid N2.
Peptide Phosphorylation-Each reaction contained 5 pl of immunoprecipitate in the form of the bacterial suspension described above (equivalent to 2.5 X lo5 cells), 2 p1 of peptide in H20, and 1 pl of [y-32P]ATP (5 mCi/ml, 40 pM; Amersham/Searle). The mixtures were set up on ice and the reaction started by addition of the ATP followed by incubation at 30 "C for 10 min. The reaction was stopped by transfer to an ice bath followed rapidly by centrifugation for 2 min in a microcentrifuge. The supernatants were aspirated and 1 pl of each sample was spotted immediately onto a 100-pm cellulose thin layer plate using a center origin. The samples were electrophoresed at pH 3.5 in a buffer containing 0.5% pyridine, 5% acetic acid (src I, src 11, src 111, src VII), or pH 4.0 in a buffer containing 1.5% pyridine, 5% acetic acid (src IV, src V, src VI) for 40 min at 1 kV. The phosphorylated peptides migrated toward the cathode. The plates were dried and exposed to Kodak XAR-5 film with a fluorescent screen at -70 "C for 12 h. In each series a control incubation without added peptide was included. Areas of cellulose corresponding to the phosphorylated peptides and a similar region from the control were aspirated from the plate. The radioactivity was eluted with pH 4.7 buffer and counted with an aqueous scintillator. Phosphorylated peptides were eluted in a similar fashion for phosphoamino acid analysis and automated sequencing which were performed as described (5,15).

RESULTS
To test whether the synthetic peptide Lys-Leu-Ile-Glu-Asp-Asn-Glu-Tyr-Thr-Ala-Arg (src I) is a substrate for tyrosine protein kinases, ideally we would have liked to use a purified protein kinase known to phosphorylate this sequence in a native protein. In the absence of such a purified enzyme, we chose to use preparations of P90, the transforming protein of the avian sarcoma virus Y73. When isolated by immunoprecipitation from Y73 virus-transformed chick cells, P90 has an associated tyrosine protein kinase activity (22) which wiU phosphorylate a site in P90 itself with a sequence identical with that in pp60" (12, 14, 15). There is reason to believe, therefore, that the synthetic peptide might be a proper substrate for the P9O-associated protein kinase activity.
P90 is a chimeric protein containing virally specified sequences at its NH2 terminus joined to sequences encoded by the cellular information in the Y73 virus genome (22). The immunoprecipitates were made with antisera directed against the NH2-terminal viral portion of P90. Since the cellular COOH-terminal domain is believed to specify the tyrosine protein kinase activity associated with P90, we thought that the active site of P90 would be available to phosphorylate exogenous substrates in this type of preparation. Such immunoprecipitates contain the enzyme activity bound to antibody molecules which in turn are complexed with protein A molecules on the surface of the S. aureus cells. It is possible that the immobilized state of the enzyme in this type of preparation may affect the absolute kinetic parameters of the reaction (see "Discussion"). For this reason, the terms apparent K,,, and apparent V,,, have been used throughout. Never- RIM src I was electrophoresed at pH 3.5 as described under "Experimental Procedures." The phosphorylated peptide was located by autoradiography and eluted from the plate. The eluate was lyophilized and hydrolyzed in 6 N HCl at 110 "C for 1 h. Half the hydrolysate (2,000 cpm) was resolved in two dimensions on a cellulose thin layer plate by electrophoresis at pH 1.9 for 20 min at 1.5 kV (horizontal direction) followed by electrophoresis at pH 3.5 for 16 min at 1 kV (vertical direction). The plate was exposed to film with a fluorescent screen for 10 h. The positions of the origin (O), the phosphoserine

( P S E R ) and phosphothreonine (P.THR) markers (dotted circles)
and phosphotyrosine (P.TYR) are indicated. theless, it should be valid to compare the kinetic constants determined for different synthetic peptides. Another drawback with the use of an immunoprecipitate as the source of protein kinase is that the amount of enzyme is not defined. One advantage of the system is that the enzyme can readily be removed from the reaction mixture.
To assay for the phosphorylation of the peptide src I, we used incubation conditions which are suitable for the phosphorylation of PSO itself, namely 10 mM MnC12,20 m~ PIPES,

Tyrosine Protein
Kinase 4845 " e " _ " Five jd of a phosphorylation reaction containing 1 mM src I was electrophoresed at pH 3.5 as described under "Experimental Procedures." The phosphorylated peptide was located by autoradiography and eluted from the plate. The eluate (21,000 cpm) was subjected to automated sequence analysis. The radioactivity released at each cycle was counted directly using Cerenkov radiation.
FIG. 4. Initial rates of phosphorylation of src I and src 11. A double reciprocal plot of u" against [peptide]" is given. Because the precise amount of enzyme in each assay was unknown the initial rates were determined simply as the radioactivity incorporated into peptide at each concentration. pH 7.0. When added to PSO-containing immunoprecipitates under these conditions in the presence of 5 /.LM [y-:"P]ATP, the peptide was phosphorylated. Electrophoresis at pH 3.5 proved to be the best system for separating the phosphorylated src I from the unincorporated ATP (see Fig. l), although for some of the more anionic peptides tested subsequently, electrophoresis a t pH 4.0 was employed. Initial experiments showed that incorporation into phosphorylated peptides was linear at the concentrations used here for 10 min at 30 "C under standard conditions. This incubation time was used for all experiments. In addition to the tyrosine residue, src I also contains a threonine which could potentially be phosphorylated. Twodimensional separation of an acid hydrolysate of "'P-labeled phosphorylated src I, however, showed that phosphotyrosine was the only phosphoamino acid present (Fig. 2). This was c o n f i i e d by automated sequence analysis of phosphorylated src I which resulted in the release of the phenylthiohydantoin of phosphotyrosine a t cycle 8 as expected (Fig. 3).
By varying the concentration of peptide in the phosphorylation reaction, the apparent K , for src I was found to be approximately 5 m~ (Fig. 4, Table I). As pointed out above, this value should not be construed as a true K,,,. Moreover, because of the limited solubility of the peptide a value this high is difficult to determine with precision. Using the same assay, we have measured approximate apparent K,,, values for two other synthetic peptides related to src I which were by- Kinetic constants were determined as described in the legend to Fig. 4 and under "Experimental Procedures." The values were calculated from at least six initial rate measurements in each case. The apparent V,,, values are expressed relative to the value for src I which has been given the value of 1. Likewise, the relative rates of reaction at a peptide concentration of 2 m~ are expressed relative to the rate fur src I. nd indicates that the values were not determined for the reasons given in the text.  Table 1. The two peptides, src I1 and src 111, which lack the glutamic acid at position 4, have much higher apparent K,,, values than src I. The two peptides, src VI and src VII, in which the tyrosine is terminal were poor substrates. Src V, the peptide lacking the COOHterminal arginine, has an apparent K,,, slightly lower than src I. We encountered some difficulty testing src IV, the peptide without the NHz-terminal lysine. The presence of this peptide in the assay causes a major fraction of the radioactivity to adhere to the bacteria. Although this could be eluted by adjusting the pH of the reaction to 4, comparison with the other peptides may be vitiated.

Peptide Substrates for a Tyrosine Protein Kinase
Since an unknown amount of enzyme was present in the reaction, absolute VmaX values cannot be determined. The relative apparent V,,, for src I, however, is lower than for src I1 (Table I). Relative apparent V,,, values are also listed for src 111 to V in Table I. The rates of phosphorylation of src VI and src VI1 were too low to estimate an apparent V,,,. Instead some indication of the efficiency of phosphorylation of these peptides relative to src I is given in Table I as defined by the incorporation of 32P at a peptide concentration of 2 m~ under our standard assay conditions. We have not determined the K, for inhibition of phosphorylation of protein substrates by any of the peptides nor have we measured the K , for ATP. We have tested a variety of other tyrosine-containing peptides in our system. For example, the peptide Glu-Glu-Glu-Glu-Tyr-Met-Pro-Met-Glu, corresponding to the sequence around tyrosine 315 in the polyoma virus middle T antigen (28), which has been reported to be the site of tyrosine phosphorylation of the middle T antigen in vitro (29), was phosphorylated with an apparent K,,, of 3 m~. Luteinizing hormone release factor (Pyr-His-Trp-Ser-Tyr-Gly-Leu-Arg-Pro-GlyNH2) was phosphorylated on tyrosine, but its K , was unmeasurably high. The peptide Arg-Gly-Tyr-Ala-Leu-Gly was also phosphorylated but extremely poorly. Free tyrosine was not phosphorylated.
We have also assessed the ability of the tyrosine protein kinases associated with a number of other viral transforming proteins to phosphorylate src I. Immunoprecipitates of pp60"" of RSV made with a serum directed against the COOHterminal hexapeptide of pp6OSrc, in which the active site of pp60"" is likely to be exposed were able to phosphorylate src I. Surprisingly, this was also true of immunoprecipitates of pp60"" made with a rabbit antitumor serum in which pp60"" molecules are able to phosphorylate the immunoglobulin heavy chain but not exogenous protein substrates (30). Immunoprecipitates containing pp60""", the cellular homologue of pp60"", made with antitumor serum were likewise able to phosphorylate src I. The tyrosine protein kinases associated with the transforming proteins of PRCII virus (31) and Abelson virus (32) and the middle T antigen of polyoma virus (33) phosphorylated src I as well, although the efficiency of this reaction varied widely. The epidermal growth factor-stimulated tyrosine protein kinase associated with the plasma membrane of A431 human tumor cells (3) was also active in the phosphorylation of src I.

DISCUSSION
Through the use of synthetic peptides we have begun to assess to what extent particular features of the primary sequence of the tyrosine phosphorylation site in pp6OSrc influence the phosphorylation of this tyrosine by tyrosine-specific protein kinases. Our preliminary results indicate that primary sequence does indeed play some role in the recognition of this tyrosine. A synthetic peptide (src I) corresponding to 10 out of 11 residues around the phosphorylated tyrosine in pp60"" and P90 serves as a substrate for the tyrosine protein kinase activity associated with the transforming protein of Y73 virus.
The apparent K, for the peptide in the phosphorylation reaction is about 5 mM. This value is about 2 orders of magnitude higher than that obtained for the best synthetic peptide substrates of the CAMP-dependent protein kinases (18,19). This difference might reflect the unusual nature of an assay in which the enzyme is immobilized. However, the phosphorylation of a synthetic peptide very similar to src I in a soluble enzyme system has been found to exhibit a similarly high apparent K,,, value (34). Furthermore, in the parallel situation offered by serine protein kinases which are associated with membranes and thereby immobilized, phosphorylation of exogenous substrates displays normal kinetics (35). Since many of the tyrosine protein kinases are themselves membrane-affiliated (36, 37), it might even be misleading to attempt to determine kinetic parameters in solution. The worry that the measured kinetic values might be the result of a situation in which the "autophosphorylation" of P90 is competing with peptide phosphorylation appears to be unfounded, The phosphorylation of P90 is complete within a few seconds of incubation at 30 "C and the extent of this reaction is less than 1% of that of peptide phosphorylation.
If the apparent K , value determined here were real, the high value for phosphorylation of src I might indicate that the primary structure of the phosphorylation site does not provide such an important recognition determinant as tertiary structure. Before concluding this, however, it should be noted that for technical reasons src I was synthesized with a lysine at its NHz terminus rather than the arginine found in the sequence of pp60"".
The possibility that lysine does not substitute effectively for this arginine should be borne in mind, since in the case of the peptide substrates for the CAMP-dependent protein kinase replacement of arginine by lysine causes a 10to 100-fold increase in the apparent K , (18,19). The studies of Casnellie et al. (34), however, showed that a peptide with an arginine in this position still had a K , value as high as the one determined here for src I. The nature of the protein kinase phosphorylating the peptide is another consideration. The decay curve for thermal inactivation of the peptide-phosphorylating activity in P9O-containing immunoprecipitates shows only a single component.' This means that a single enzyme is responsible for phosphorylation of the peptide. Presumably, this activity is the same as that which phosphorylates P90 itself at a sequence identical in 10 out of 11 positions with that of src I. Therefore, it seems likely that the peptide is an appropriate substrate for this enzyme. It would be desirable, however, to repeat these experiments with a purified soluble tyrosine protein kinase.
Regardless of the potential problems with interpreting the values of the kinetic constants determined in this assay, comparison of the kinetic parameters for the phosphorylation of different peptides would seem to be justified. As a result of this comparison, there do appear to be elements of the primary sequence surrounding the target tyrosine which are recognized. Clearly, the tyrosine is inefficiently phosphorylated if it is either the NHP or COOH terminus of the peptide. Removal of the COOH-terminal arginine from src I led to a decrease in apparent K,. Whether this reflects an inhibiting influence of a basic residue so close to the tyrosine on the COOH-terminal side is not clear. The effect is not simply due to the loss of a positive charge from the peptide, since the removal of the NHz-terminal lysine did not lead to such a change. Other synthetic peptides with longer COOH-terminal extensions in which the arginine is replaced by neutral amino acids should be tested.
The removal of the NHz-terminal lysine did not alter the properties of the peptide as a substrate substantially. In some ways, this is surprising since a lysine or arginine 7 residues to the NHZ-terminal side of the phosphorylated tyrosine is a widespread feature of the sites of tyrosine phosphorylation in viral transforming proteins (15). Several possibilities could account for this. The basic amino acid found at this site may not play a role in substrate recognition in short peptides, for instance because it is the NHz-terminal residue. Alternatively, if a basic residue were important, lysine may not be able to substitute for arginine at this site. Finally, a requirement for a positively charged residue at this position might be fulfiied by the a-amino group of the NH2-terminal leucine in src IV.
Perhaps the most striking result obtained with the analogues of src I is the 10-fold increase in apparent K,,, observed for those peptides lacking the glutamic acid 4 residues upstream of the tyrosine. This probably does not simply reflect a decrease in the anionic character of the peptide since src I does not have a higher apparent K, than src IV. Moreover, the effect is not due to the decreased distance between the lysine and the tyrosine, since in the case of src I1 removal of * M. Houslay and T. Hunter, unpublished observations.

Tyrosine Protein
Kinase 4847 the lysine did not alter the apparent K,,, (data not shown). Rather the increased apparent K , values of the peptides lacking glutamic acid at position 4 taken together with the conservation of a glutamic acid at this position in at least three sites of phosphorylation in viral transforming proteins and the presence of glutamic acid(s) on the NHB-terminal side of several other phosphorylated tyrosines suggest that this glutamic acid may be a crucial element in the recognition site. This idea is not unreasonable since the casein protein kinases appear to recognize acidic residues in the vicinity of the target serine or threonine. For casein kinase I, these are on the NH2terminal side. In addition, such sites often contain multiple acidic residues. The peptides src I1 and src I11 retain 2 out of the 3 acidic residues on the NHp-terminal side of the tyrosine. It will be interesting to test other peptides in which these amino acids are replaced by neutral residues and to determine whether glutamic acid and aspartic acid are equivalent in their influence. The optimal spacing of these acidic residues from the tyrosine should also be investigated. The tyrosine protein kinases associated with a number of other viral transforming proteins were able to phosphorylate src I. The exact identity of the tyrosine protein kinases involved in each case is at present unknown, so that it is unclear how many of the tyrosine protein kinases will be able to phosphorylate this amino acid sequence. In this regard, hcwever, it is noteworthy that Casnellie et al. (34) have recerkly reported an uncharacterized tyrosine protein kinase associz-ted with plasma membrane of LSTRA cells which can phosphorylate a synthetic peptide similar to src I based on the tyrosine phosphorylation site in pp60"".