Specificity of protein phosphotyrosine phosphatases. Comparison with mammalian alkaline phosphatase using polypeptide substrates.

The specificity of cytosolic protein phosphotyrosine (PPT) phosphatases was investigated using different peptides and proteins that were phosphorylated on tyrosine residues by the EGF receptor kinase. The acidic phosphoproteins, serum albumin, casein, and myosin light chains, were dephosphorylated by the PPT phosphatases with apparent Km values of 1.2 to 12.5 microM and apparent velocities of 0.2 to 18 mumol/min/mg. In contrast, [Tyr(32P)]histone and the phosphotyrosine peptides [Val5]angiotensin and RR-src, a peptide with sequence Arg-Arg-Leu-Ile-Glu-Asp-Ala-Glu-Tyr-Ala-Ala-Arg-Gly, were unreactive with the PPT phosphatases. However, each of these unreactive phosphopolypeptides was dephosphorylated under the same conditions by calf-intestine alkaline phosphatase. The data reveal how PPT phosphatase activity has been ascribed to different cellular enzymes. When acidic phosphotyrosine proteins were used as substrates in assays for PPT phosphatase activity the cytosolic enzymes were isolated, whereas when phosphotyrosine histones were used as substrates only the membrane-bound alkaline phosphatase was detected. Apparently the protein tyrosine kinase and the protein tyrosine phosphatases do not have the same specificity, so substrates such as histone, angiotensin, or RR-src are phosphorylated but not hydrolyzed. Therefore, these polypeptides would be ideal for the characterization of protein tyrosine kinases in cellular extracts.

Reversible phosphorylation of cellular proteins on tyrosine residues is believed to play a key role in the transformation of cells by a variety of oncogenes, in the action of growth factors, and in the action of insulin (for reviews see Refs. [1][2][3]. The transforming proteins and the growth factor receptors display protein tyrosine kinase activity. These kinases act as their own substrates and thus can be labeled on tyrosine residues using [T-~~PJATP. Analysis o f the sequence o f amino acids adjacent to the phosphory~ated tyrosines revealed that two or three acidic residues typically preceded the tyrosine (4). A variety of synthetic peptides, including analogues of * This work was supported by American Cancer Society Grant NP-384 and National Institutes of Health Grant AM31374. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. __ angiotensin, have been used to investigate the role primary structure plays in substrate recognition by the kinases (5)(6)(7)(8).
Recently, this laboratory has purified two distinct forms of protein phosphotyrosine (PPT') phosphatase from rabbit kidney (9). In this study we compared the reactivity of different Tyr(P) proteins and Tyr(P) peptides with the phosphatases to determine whether or not their specificity paralleled that of the kinases. We found that PPT phosphatases catalyzed the dephosphorylation of the acidic proteins serum albumin, casein, and myosin light chains from smooth muscle. In contrast, the basic protein histone and the peptides [Val6] angiotensin I1 and RR-src were not dephosphorylated by the PPT phosphatases but were dephosphorylated by mammalian alkaline phosphatase. Thus, our results show that experiments using Tyr(P) histone or Tyr(P) peptides as substrate would be expected to measure only the activity of alkaline phosphatase, not PPT phosphatases.

Methods
Preparation of TyrfP) Proteins-Bovine serum albumin was reduced and alkylated by the method of Shriner and Brautigan (9). Casein was dissolved in dilute alkaline solution and precipitated by the addition of 100% trichloroacetic acid solution to a final concentration of 10%. The precipitated casein was collected by centrifugation at 10,000 X g for 15 min, resuspended in 100 mM ammonia, and dialyzed exhaustively against the same solution before lyophilization. Smooth muscle myosin from chicken gizzards was purified by the method of Sobieszek (10). Myosin light chains were prepared using a modified version of the method of Holt and Lowey (11). Briefly, crude 2 r n~ EDTA, 0.3 M KCI, 2 mM dithiothreitol, and 50 mM Tris/HCl, myosin was incubated overnight in a buffer containing 5 M guanidine, pH 7.9. An equal volume of cold distilled water was added and the mixture was allowed to stand for 30 min. This step precipitated the myosin heavy chains which were pelleted by centrifugation at 10,000 X g for 15 min. The soluble myosin light chains were precipitated by slowly adding cold ethanol to the supernatant to a final concentration

_~-_
residues by the EGF receptor kinase in membranes from A431 cells as described previously (9). The total 3zP present in the recovered phosphoprotein was determined by spotting 2-pl aliquots in triplicate onto squares of Whatman No. 50 paper and counting them in a Packard Model 300C Tri-Carb scintillation counter. The amount of 32P incorporated into proteins was calculated from the specific radioactivity of a 10 mM solution of [T-~'P]ATP.
Phosphorylation of Peptide Substrates-The peptides [Val'langiotensin I1 and RR-src were phosphorylated on tyrosine residues by the EGF receptor kinase in the same manner as for the protein substrates. The peptides were soluble in trichloroacetic acid, SO they were isolated from the reaction mixture by ion-exchange chromatography on Bio-Rad AG 1-X2 acetate resin. After a 16-h incubation, the reaction mixture was acidified with an equal volume of 1 N acetic acid, applied to a 3-ml column and allowed to sit for a few minutes. Subsequently, the phosphopeptide was eluted with 1 N acetic acid whereas other reactants were retained on the resin. The Tyr(P) peptides were lyophilized and dissolved in either 25 mM HEPES, pH 7.0, 1 mM EDTA for reaction with PPT phosphatase I, or in 50 mM MES, pH 5.0, 1 mM EDTA for reaction with PPT phosphatase 11. Determination of PPT Phosphatase Actiuity-Dephosphorylation of [Tyr(32P)]serum albumin, [T~r(~'P)]casein, or [Tyr(32P)]myosin light chains was measured by the release of 32Pi from the protein that was precipitated with trichloroacetic acid as described previously (9).

Dephosphorylation of [T~r(~~P)]histone, [Tyr(32P)]
[Va16]-angiotensin 11, or [T~r(~'P)1RR-src was measured by loss of 32Pi. Aliquots of the reaction mixture were pipetted onto 1-cm squares of Whatman P81 phosphocellulose paper which bound the polypeptides. The papers were immediately spotted with an equal volume of 30% acetic acid and washed in 400 ml of 30% acetic acid for 15 min and in 300 ml of 15% acetic acid for 10 min and finally were rinsed with 100% ethanol for 5 min prior to drying under a heat lamp (12). The papers were placed in 5 ml of scintillation fluid and counted in a Packard Model 300C Tri-Carb scintillation counter.
Dephosphorylation of Acidic Tyrr'P)-labeled Proteins by PPT Phosphatase I-Reaction mixtures of 50 pl contained T~r(~'P)-labeled protein suspended in 25 pl of 25 mM HEPES, pH 7.0, 1 mM EDTA. The reaction was initiated by the addition of purified phosphatase suspended in 20 p1 of 25 mM HEPES, pH 7.0, 1 mM EDTA, 50 mM 2-mercaptoethanol, 1 mg/ml of serum albumin, and 5 p1 of H20. Activities were calculated relative to samples containing no enzyme but where 20 pl of the buffer was added to bring the final volume of the reaction mixture up to 50 pl. Reactions were incubated for 10 min at 30 "C and were terminated by the addition of 0.1 ml of 25 mg/ml of serum albumin immediately followed by the addition of 0.85 ml of cold 15% trichloroacetic acid. The reaction mixtures were placed on ice for 15 min and centrifuged for 5 min in a Fisher Model 235B microcentrifuge. The amount of radioactivity released was determined by dissolving 0.5 ml of the supernatants in 5.0 ml of scintillation fluid and counting as stated above. Initial T~r (~' p ) concentration was determined by dissolving the precipitate of the control samples in 1 ml of 0.1 N NaOH and counting 0.5-ml aliquots as stated above.
Dephosphorylation of Acidic Tyr?'P)-labeled Proteins with PPT Phosphatase ZZ-Each reaction contained 25 pl of substrate as described above and was initiated by the addition of enzyme suspended in 20 pl of 100 mM MES, pH 5.0, 1 mM EDTA, 50 mM 2-mercaptoethanol, 1 mg/ml of serum albumin, and 5 pl of HzO. Substrate concentrations were determined and controls were run as described in the above paragraph.
Hydrolysis of p-Nitrophenylphosphote (p-NPP)-Both the PPT phosphatases as well as alkaline phosphatase catalyzed the hydrolysis of p-NPP. We used p-NPP phosphatase activities to compare the relative amounts of the different enzymes added to each experiment. To measure p-NPP hydrolysis, 90 pl of phosphatase solution and 10 p1 of 100 mM p-NPP were mixed at 30 "C. When the solution turned pale yellow, the reaction was terminated by the addition of 900 p1 of 1 M NaZC03. The concentration of p-nitrophenol was calculated from the absorbance of the solution at 410 nm, using a millimolar extinction coefficient of 17.5. The amount of enzyme that produced 1 nmol of p-nitrophenollmin at 30 "C at the optimum pH was defined as 1 unit. were prepared. Each of the proteins and peptides used were phosphorylated on tyrosine residues by the EGF receptor kinase in membranes from A431 cells. As is well known, in most cases the extent of phosphorylation of added proteins is low, approximately 0.01 to 0.1 mol of 3 2~/ m~l of protein.

Characterization
Amino acid analyses of the labeled albumin and histone by thin layer electrophoresis revealed that more than 80% of the radioactivity was present as Tyr(P) (data not shown). The substrates were phosphorylated to different extents as follows: alkylated serum albumin, 1 nmol of 32P/mg; histone, 1 nmol of 32P/mg; casein, 0.08 nmol/mg; myosin mixed light chains, Dephosphorylation of Acidic Proteins by PPT Phosphatase, Types Z and ZZ-Rabbit kidney cytosolic PPT phosphatases typq's I and 11, have parallel substrate specificity. PPT phosphatase, type I, displays optimal activity at pH 7.0 and requires mercaptans for activity. In contrast, PPT phosphatase, type 11, exhibits optimal activity at pH 5.0 and does not require mercaptans for activity. As shown in Table I  12% of the 32P was released from the protein (Fig. 1). The lack of reactivity with histone presumably reflects the specificity of the PPT phosphatases.
In contrast, when calf intestine alkaline phosphatase was incubated with [T~r(~~P)]histone at pH 8.5 the reaction appeared complete within 30 min and 70% of the 32P was released from the protein, as shown in Fig. 1  incubation was +.lo% of the initial 200,000 cpm. Phosphatase activity in the enzyme samples added to the peptides was confirmed by the hydrolysis of p-nitrophenylphosphate.
In contrast to the PPT phosphatases, calf intestine alkaline phosphatase dephosphorylated both angiotensin and RR-src.
More than 70% of the 32P was released from [TY~(~'P)]RRsrc when incubated with alkaline phosphatase at 30 "C, pH 7.0. Similarly, when [TY~(~*P)] [Vals]-angiotensin was incubated with alkaline phosphatase under the same conditions as those used for RR-src, over 65% of the 32P was released.
These results show that the peptides were enzymatically dephosphorylated by alkaline phosphatase, but not by PPT phosphatases.

DISCUSSION
Our results show that PPT phosphatases only dephosphorylate certain acidic Tyr(P) protein substrates which is indicative of their relatively high degree of specificity. The PPT phosphatases do not parallel the specificity of the EGF receptor kinase, which was used to phosphorylate all the polypeptides tested as substrates. Regardless of the net charge of the polypeptides, protein tyrosine kinases phosphorylated tyrosine residues that were preceded by several acidic side chains (5-8, 13). Thus, it appears that the primary structure surrounding the phosphorylated tyrosine residue, and not the overall charge of the substrate, plays a principal role in determining kinase specificity. In contrast to the kinase, PPT phosphatases I and I1 reacted solely with the acidic phosphoproteins. Even though they are distinct enzymes, the two cytosolic phosphatases appear to have parallel specificity. Of the proteins tested, T~r(~'P)-alkylated serum albumin was the best substrate for both PPT phosphatases, types I and 11. The data indicate that [T~r(~~P)]histone was not a substrate for either of the PPT phosphatases. The lack of reactivity with histone is intriguing because it helps to clarify conflicting reports in the literature. S w a p et al. (15) used Tyr(P) histone, phosphorylated by the EGF-receptor kinase, to monitor protein Tyr(P) phosphatase activity in homogenate5 of TCRC-2 cells. They found about 10-fold higher specific activity in the particulate uersus the soluble fraction. A membranebound glycoprotein was recovered and characterized as an alkaline-type phosphatase. The authors concluded "that alkaline phosphatases may be involved in the dephosphorylation of membrane-bound phosphotyrosine-containing proteins." Our results confirm that Tyr(P) histone is indeed a substrate for mammalian alkaline phosphatase. However, this phosphoprotein is not dephosphorylated by cytosolic PPT phosphatases. Using acidic Tyr(P) proteins rather than Tyr(P) histone to measure PPT phosphatase activity, Horlein et al. (16), Foulkes et al. (171,and Shriner and Brautigan (9) have characterized cytosolic enzymes that are distinct from alkaline phosphatase. Based on these results, we believe that if one uses Tyr(P) histone to measure PPT phosphatase activity, only the membrane-bound alkaline phosphatase will be detected.
In addition to phosphomonoesters such as p-nitrophenylphosphate alkaline phosphatases react with Ser(P) and Tyr(P) residues in proteins (14,18) and exhibit some specificity for Tyr(P) relative to Ser(P) (14). Nonetheless, PPT phosphatases appear to be far more specific for Tyr(P) proteins than alkaline phosphatases. Relative to intestine alkaline phosphatase, Foulkes et al. (17) have shown that cytosolic PPT phosphatase was nearly 3000 times more active with Tyr(P) casein than withp-nitrophenylphosphate as substrate. We have been unable to detect hydrolysis of Ser(P) in proteins with PPT phosphatases. Furthermore, PPT phosphatases, types I or 11, were unreactive with Tyr(P) peptides, RR-src, or [Va16]angiotensin, though alkaline phosphatase dephosphorylated each of these phosphopeptides under nearly the same conditions. We conclude that the cytosolic PPT phosphatases prefer substrates with a net negative charge with a more extensive tertiary structure than afforded by a tridecapeptide such as RR-src.
Although the in vivo substrates of protein tyrosine kinases and PPT phosphatases continue to elude identification, in vitro studies employing artificial substrates and synthetic peptides have proven useful in the characterization of these enzymes. An important conclusion of this report, with direct practical applications, is that neither Tyr(P) histone, nor Tyr(P) peptides are dephosphorylated by cytosolic PPT phosphatases. Thus, histone IIA would be a preferred substrate for measuring protein tyrosine kinase activity in cell extracts. Removal of the particulate fraction containing membranebound alkaline phosphatase should eliminate nearly all of the Tyr(P) histone phosphatase activity. Moreover, use of histone as substrate in a reaction mixture containing orthovanadate, a potent and effective PPT phosphatase inhibitor (19), would be ideal for investigating the kinetics and specificity of protein tyrosine kinases.
Finally, differences in substrate specificity between cytosolic PPT phosphatase and membrane-bound alkaline phosphatase may yet prove to be of physiological significance. It will be interesting to discover whether there are both acidic and basic amino acid sequences surrounding Tyr(P) in the proteins phosphorylated in vivo.