ANALYTICAL

BIOCHEMISTRY

190,233-237

(1990)

An Alternative Method for a Fast Separation of Phosphotyrosine G. Mufioz Laboratorio

Received

and S. H. Marshall de Gene’tica Molecular,

November

Instituto

de Biologia,

Universidad

Press.

Inc.

Phosphorylation of amino acids in proteins plays a central role in signal transduction pathways which regulate several biological processes including cellular proliferation (1). Under normal conditions, phosphorylation generally occurs on serine and threonine residues and more rarely on tyrosine. Phosphotyrosine represents less than 1% of the phosphorylated residues in bulk phosphoprotein (2). Lately, different tyrosine-specific protein kinases have been isolated and some of them seem to be critically involved in cellular transformation (3-5). Increased tyrosine-specific phosphorylation of proteins has been found to be a marker for some human malignant cell lines (6,7). Nevertheless, it is technically difficult to resolve the phosphorylated amino acids from each other especially after in vivo and in vitro labeling experiments. The analysis of protein hydrolysates can be rendered erroneous by the possible presence of either mononucleotides, ATP, ribose phosphate, or inorganic orthophosphate, which represent phosphorylated contaminants comigrating with phosphoamino acids (8,9). The 0003-2697/90 Copyright All rights

Casilla 4059, Valparaiso,

Chile

2, 1989

A simple and rapid procedure is described for fully separating phosphotyrosine from phosphoserine and phosphothreonine through one-dimensional thin-layer chromatography. The migration properties of these phosphoamino acids are compared with those of CMP, UMP, ATP, ribose phosphate, and inorganic orthophosphate, considered the most frequent comigrating products derived from 32P-labeling experiments. We demonstrate that R, values for the three phosphoamino acids differ from those displayed by the mentioned contaminating compounds. One of the most relevant advantages of this procedure is that a complete separation of phosphotyrosine can be achieved in only 90 min. o 1990 Academic

Cato’lica de Valparako,

$3.00 0 1990 by Academic Press, of reproduction in any form

available procedures for the separation of phosphoamino acids most frequently used include high-pressure liquid chromatography (HPLC) (10) and high-voltage electrophoresis alone or in combination with TLC (2,11-13), as well as a number of combinations of these methods (14). Nevertheless, all of them have a share of disadvantages. HPLC, although a rapid procedure, is not practical when several samples need to be analyzed simultaneously. On the other hand, the resolving power of one-dimensional chromatography or that of electrophoresis alone is generally insufficient for a clear-cut separation. Finally, although there are procedures yielding a complete resolution of the three phosphoamino acids, these necessarily require at least two successive dimensions (2,8, 9, 14). In this report we describe a simple and fast one-dimensional TLC method that clearly separates the three most common phosphoamino acids from each other, and from 32P-labeled contaminants. The procedure is based on ascending chromatography on silica gel thinlayer plates using a mixture of absolute ethanol and 25% ammonia solution as a solvent. Under this condition, P-Tyrl displays the highest Rf value, and can be completely resolved in the time span the solvent reaches the top of the plate (one cycle). Nevertheless, in order to fully resolve P-Thr from P-Ser, successive cycles are required.

MATERIALS

AND

METHODS

Phosphoamino acids. Phosphorylated amino acids O-phospho-DL-serine, 0-phospho-DL-threonine, Ophospho-D-tyrosine, and ribose phosphate, were purchased from Sigma Chemical Co. (St. Louis, MO).

1 Abbreviations used: P-Ser, phothreonine; P-Tyr, phosphotyrosine; phenylmethylsuifonyl fluoride.

phosphoserine; P-Thr, NP-40, Nonidet P-40;

phosPMSF,

233 Inc. reserved.

234

MUNOZ

AND MARSHALL

B

A FRONT

C

,-

0

0

!

e

0

-

P-tyr

-

P-Thr

-P-Ser

1=

-

FIG. 1. Phosphoamino acid analysis using silica gel TLC plates and a solvent mixture of absolute ethanol:25% ammonia solution, in a ratio of 3.5:1.6. Identical aliquots of a model system mixture containing P-Ser, P-Thr, and P-Tyr were spotted on three plates, and resolved after successive cycles under the same solvent. Phosphoamino acids were visualized after ninhydrin staining. A, one cycle; B, two cycles; C, three cycles. Samples were spotted at 0.5 cm from the bottom edge. The arrow in A shows the resolution of P-Tyr after one cycle. Markers in A and B, correspond to the relative mobilities of the three phosphoamino acids, as indicated in C. Distance between origin and front: 10.5 cm.

Nuclcoside monophosphates. 5’-ATP disodium salt, 2’,3’-CMP disodium salt, 2’,3’-UMP disodium salt were also from Sigma Chemical Co. Chemicals. Absolute ethanol, n-butanol, and acetic acid were purchased from Merck Laboratories (Darmstadt, West Germany). Ammonia solution (25%) and ninhydrin were from Prolabo (France). Chromatography. All chromatographic plates were from Merck Laboratories. Thin-layer chromatography

of phosphoamino acids was performed on precoated aluminum sheets (0.2 mm layer thickness, 11 cm height) of either cellulose or silica gel plates. Nucleoside monophosphates and ribose phosphate were resolved on silica gel plates (0.1 mm layer thickness, 11 cm height) with fluorescent indicator. Ascending chromatography was carried out with a solvent mixture containing absolute ethanol:25% ammonia solution, in a ratio of 3.5:1.6.

TABLE TABLE

1

R, Values for P-Ser, P-Thr,

R, Values and P-Tyr

2

Obtained for CMP, UMP, ATP, Orthophosphate, Ribose Phosphate, and P-Tyr Absolute ethanok25% ammonia solution (3.51.6)

Absolute ethanol:25% ammonia solution (3.5:1.6) Phosphoamino acid

Rf (1 cycle)

4 (2 cycles)

(3 cycles)

Compound

(1 cycle)

(2 cycles)

(3 cycles)

P-Ser P-Thr P-Tyr

0.05 0.07 0.12

0.12 0.15 0.26

0.17 0.20 0.32

2’-, 3’-CMP

0.14 0.17 0.15 0.18 0.02 0.00 0.40 0.12

0.30 0.37 0.30 0.34 0.03 0.00 0.56 0.26

0.42 0.47 0.40 0.45 0.04 0.00 0.77 0.32

P-Ser P-Tyr

0.14

Rf

0.15

Rf

2'-, 3’-UMP ATP pi P-ribose P-Tyr

4

4

CHROMATOGRAPHIC

SEPARATION

OF

235

PHOSPHOTYROSINE

A FRONT

B

-

FRONT-

”

ORlOIN

.

0

.

-

P,Ty,

0

9

-

P-Thr Pmser

ORlOIN

-

1

2

-

9

-

PmTyr

.

-

P,Thr P-SW

I

1

2

FIG. 2. Resolution of phosphoamino acids from hydrolysates on TLC plates. (A) Ninhydrin staining after two successive solvent cycles of: lane 1, a mixture of standard phosphoamino acids; lane 2, the mixture of phosphoamino acids after acid hydrolysis. (B) Autoradiogram of a a2P-labeled mouse liver protein hydrolysate after three cycles on TLC plates: lane 1, origin included, lane 2, origin removed before autoradiograproduct which might run close to the origin. Phosphoamino acid markers were visualized after phy to avoid obscuring any 32P-labeled ninhydrin staining. Distance between origin and front: 10.5 cm.

Migration of standards was completed after 90 min at room temperature, which, under our conditions, defines one cycle. In between cycles, plates were air-dried followed by new cycles under the original solvent. Detection of phosphoamino acids, nucleotides, and ribose phosphate on TLC plates. Phosphoamino acids were detected by spraying dried plates with a 0.3% ninhydrin solution in n-butanol with 3% acetic acid, and visualized after the plate was baked for 10 min at 60°C. Nucleotides and ribose phosphate were visualized directly under a uv light. Phosphoamino acids were photographed using Kodalith film. Preparation of crude extracts. A fresh mouse liver was homogenized in buffer A (20 mM Hepes, pH 7.2,lO InM MgCl,, 1.0 mM CaCl,, 50 PM Na,VO,, 0.5% NP-40, 1 InM PMSF) at 4°C. Vanadate was included to inhibit endogenous protein-tyrosine phosphatases (15). An aliquot of the crude liver extract was incubated for 2 h at 37°C in the presence of 0.1 mCi/ml of [32P]orthophosphoric acid (activity, 2.1 mCi/ml; code P-032020 from Comision Chilena de Energia Nuclear) in a final volume of 0.7 ml. PHA-stimulated peripheral human blood lymphocytes. Peripheral human blood lymphocytes (3.5 ml) stimulated with phytohemmaglutinin A were prepared

as described (16) and incubated in the presence of 1.0 mCi of [32P]orthophosphoric acid for 2 h at 37°C. Phosphoamino acid analysis. Proteins were precipitated with 8 vol of cold acetone. After 30 min at -3O”C, the proteins were recovered by centrifugation. The pellet was resuspended in water and an equal volume of 12 N HCl was added. Partial acid hydrolysis was performed for 3 h at 110°C and the products thus liberated, were analyzed by TLC and autoradiography. ATP and [32P]orthophosphate were spotted on the same chromatographic plate. As additional controls, the following samples were submitted to acid hydrolysis: a mixture of the three phosphoamino acids, [32P]orthophosphoric acid, and a mixture of the phosphoamino acids plus [32P]orthophosphoric acid. RESULTS AND DISCUSSION Although phosphotyrosine is the less abundant of the three predominant phosphoamino acids in Go, we decided to use as a model system a solution containing the three phosphoamino acids in identical concentrations in order to test their resolution under the new solvent. Initial experiments showed that silica gel plates resolved these phosphoamino acids much better than cellulose plates did. Thus, all the data presented in this

236

MUfiOZ

AND

FRONT-

.

ORIGIN

-

-

P-Tyr

-

P,Thr P-ser

1

FIG. 3.

Autoradiogram of 32P-labeled human lymphocyte hydrolysate on TLC plates. Plates were autoradiographed after two solvent cycles. Phosphoamino acid markers are indicated. Distance between origin and front: 10.5 cm.

report will be based on the development of phosphoamino acids on the former type of plates. Figure 1 shows the differential resolution obtained for P-Tyr, P-Thr, and P-Ser after one, two, and three successive cycles respectively. It is worth noticing that after one cycle, P-Tyr is already resolved (arrow in Fig. 1, lane A). The separation is achieved in only 90 min, representing a clear advantage over classical chromatographic and/or electrophoretic separation methods described in the literature, which take from 120 min (2) to several hours (14). Nevertheless, a complete resolution of P-Thr from P-Ser requires up to three cycles (Fig. 1, lanes B and C). Notwithstanding, P-Tyr is resolved as the phosphoamino acid which displays the highest Rf value, a condition that eases its clear and unambigous identification on the plate, in spite of the fact that the solvent mixture proposed in this report is not significantly different from that used by others (2,14). In any case, the recovery of phosphotyrosine is clearly attained irrespective of its recovery over P-Ser and P-Thr. In contrast, most procedures based on one-dimensional separation of phosphoamino acids yield P-Tyr with the lowest mobility of all (2). Table 1 shows the average R, values observed for the three phosphoamino acids, under the conditions described in Fig. 1, as well as the difference in R, values between P-Tyr and P-Ser. Several problems are encountered when purifying phosphoamino acids after 32P-labeling experiments. The most commonly used technique is partial acid hy-

MARSHALL

drolysis of proteins. Under these conditions, [32P]ATP or [32P]orthophosphate, which are normally used in protein phosphorylation, render additional phosphorylated compounds that comigrate with phosphoamino acid standards. Of all mononucleotides, 3’-UMP and 3’CMP generated from RNA hydrolysis, are the most commonly found (17). In order to test these possibilities, different putative contaminants were spotted on the same TLC plate and resolved in one and successive cycles of ascending chromatography. The results are summarized in Table 2. It is important to note from this table that ATP and especially Pi remains at the origin, and that neither 2’- or 3’-UMP and/or 2’- or 3’-CMP comigrate with any of the three most commonly found phosphoamino acids. Also, ribose phosphate shows the highest Rf of all. Since the conditions used for partial acid hydrolysis of proteins might modify the migration properties of phosphoamino acids, the mixture used as a model system was hydrolyzed as if it was a mixture of proteins, and the resulting products resolved in TLC plates. As shown in Fig. 2, the mobility of the three phosphoamino acids was not altered at all after hydrolysis (lane 2). When compared to the untreated mixture (lane l), some release of the free amino acids could have occurred as seen by the appearance of at least one new nynhydrinpositive spot near the front solvent. Indeed this appears to be the case, for we have verified this possibility experimentally (data not shown). In addition, a mouse liver hydrolysate supplemented with the mixture of commercial phophoamino acids, also displayed unaltered Rf values when resolved on TLC plates (data not shown), suggesting that the separation of P-Tyr according to the proposed method can be readily achieved irrespective of the quality of the sample that contains it. In order to double check this possibility we ran a mixture of some of the putative contaminants, 2’- and 3’-UMP, 2’- and 3’CMP, and ribose phosphate, altogether with P-Tyr. The results confirmed that the heterogeneity of sample composition does not affect the resolution of P-Tyr. Moreover, exposure of purified yeast tRNA to identical partial acid hydrolysis conditions did not render any spot comigrating with P-Tyr. Finally, and in order to make certain that the conditions described also worked for in vitro as well as for in uiuo labeled samples, we carried two different kind of experiments. First, a protein hydrolysate obtained after 32P-labeling of a mouse liver extract was resolved with three successive cycles of ascending chromatography. Since P-Tyr has been reported in mouse liver (18), and to further enrich for this phosphoamino acid, the incubation was carried out in the presence of sodium vanadate, an inhibitor of protein-tyrosine phosphatases (15). Figure 2B shows the autoradiogram of the resolved mixture (lane 1 with the origin included, lane 2 with the origin removed). Two radioactive spots are clearly de-

CHROMATOGRAPHIC

SEPARATION

tected comigrating with the supplemented commercial P-Tyr and P-Thr, respectively. No radioactive P-Ser was observed in this experiment. One possible explanation is that the extract might be enriched with a protein-serine phosphatase activity. Most importantly, there are no contaminants detectable as major spots on the plate masking the resolution of P-Tyr. Second, a PHA-stimulated human lymphocyte culture incubated in the presence of [32P]orthophosphoric acid, was the source of a bulk hydrolysate resolved with two cycles of ascending chromatography. As seen in Fig. 3, P-Ser was the major product resolved on the plate and although P-Tyr was not preferentially labeled, the position of its expected mobility does not show any radioactive signal that could indicate a putative comigrating product resulting from an in uiuo labeling. In conclusion, a simple, rapid, one-dimensional TLC procedure is presented which resolves P-Tyr from PThr and P-Ser, as well as from all other common 32P-labeled contaminants found after in vitro and in viuo labeling experiments. Moreover, the technique allows further separation of any of the three phosphoamino acids, after successive cycles of ascending chromatography using the same solvent mixture.

OF

237

PHOSPHOTYROSINE

REFERENCES 1. Sibley, (1988)

D. R., Benovic, J. L., Caron, Endow. Rev. 9,38-56.

M. G., and Lefkowitz,

2. Hunter,

T., and Sefton, B. M. (1980) Proc. Natl. Acad. 77,1311-1315. Bishop, J. M. (1983) Annu. Reu. Biochem. 62,301-355.

3. 4. Hunter,

T.,

and

Cooper,

J. A. (1985)

Annu.

R. J. Sci. USA

Reu. Biockm.

54,

897-930.

5. Edelman, 6.

A. M., Blumenthal, D. K., and Krebs, E. G. (1987) Annu. Rev. Biochem. 56,567-613. Giordano, S., Di Renzo, M. F., Cirillo, D., Naldini, L., Chado-Piat, L., and Comoglio, P. M. (1987) Znt. J. Cancer 39, 482-487.

7. Comoglio, P. M., Di Renzo, F., and Gaudino, G. (1987) Ann. N. Y. Acad. Ski. 511,256-261. 8. Rothberg, P. G., Harris, T. J. R., Nomoto, A., and Wimmer, E. (1978) Proc. Natl. Acad. Sci. USA 76,4868-4872. 9. Sefton,

B. M., Hunter,

T., Beemon,

K., and Eckhard,

W. (1980)

Cell20,807-816. 10. Robert, J. C., Soumarmon, togr. 338, 315-324.

A., and Lewin,

M. J. (1985)

J. Chroma-

11. Eckart, W., Hutchinson, M. A., and Hunter, T. (1979) Cell 18, 925-933. 12. Bitte, L., and Kabat, D. (1974) in Methods in Enzymology (Grossman, L., and Moldave, K., Eds.), Vol. 30, pp. 563-590, Academic Press, New York. 13. Mitchell,

H. K., and Lunan,

K. D. (1964)

Arch.

Biochem.

Biophys.

106,219-222.

ACKNOWLEDGMENTS We thank Dr. Ema Navarrete and Ms. Christa Seebach for providing PHA-stimulated human blood lymphocytes, Mr. C. Gonzalez for the art work, and Mr. G. Sierralta for assistance with the editing process. We are deeply indebted to Dr. Tony Hunter from the Salk Institute (La Jolla, CA) for his various suggestions and critical review of the manuscript. This research was supported by Direction General de Investigation, Universidad Catolica de Valparaiso, Chile.

14. Manai,

M., and Cozzone,

A. J. (1982)

15. Tonks, N. K., Diltz, C. D., and Fischer, 263,6722-6730. 16. Bird, B. A., and Forrester, F. P. (1981) 17.

Health Hunter,

and Human T. (1989),

Services personal

18. Wong, T. W., and Goldberg, USA 80, 2529-2533.

Anal.

124,12-18.

B&hem.

E. H. (1988)

J. Biol.

in Cell Culture, CDC, Atlanta, GA. communication. A, R. (1983)

hoc.

Natl.

Chem.

U.S. Dept.

Acad.

Sci.