Protein-Tyrosine Kinases

ANNUAL REVIEWS

Further

Quick links to online content Ann. Rev. Biochem. 1985.54:897-930 Copyright © 1985 by Annual Reviews Inc. All rights reserved

Annu. Rev. Biochem. 1985.54:897-930. Downloaded from www.annualreviews.org Access provided by University of California - San Diego on 03/13/15. For personal use only.

PROTEIN-TYROSINE KINASES 1 Tony Hunter and Jonathan A. Cooper Molecular Biology and Virology Laboratory, The Salk Institute, Post Office Box

85800, San Diego, California 92138

CONTENTS PERSPECTIVES AND SUMMARy. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

897

TWO TYPES OF PROTEIN·TYROSINE KINASE . . . . . .. . ... . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . .

899

RETROVIRAL PROTEIN·TYROSINE KINASES AND THEIR CELLULAR HOMO· LOGS . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ...... . . . . . ...... . . . . . . . . . . . . . . . . . . ..... pp60v-src GROWTH FACTOR RECEPTOR-ASSOCIATED PROTEIN·TYROSINE KINASES

90 I

.

• • • . . . • • • • • • • • • • . • • • • • • • • • • • • • . • . . • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • . • • • • • • • • • • • • • • • • •

The EGF receptor ...... . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ... ... . . .

899

910 912

OTHER PROTEIN·TYROSINE KINASES ....................... ............. . . . . ..... . . . . . . .

918

GENERAL PROPERTIES O F PROTEIN·TYROSINE KINASES . . . . . . . . . . .... . . . . . . . . . . . . .

919 920

. .

Substrate Selection................................................................................ Inhibitors of Protein-Tyrosine Kinases........................................................

922

LIPID PHOSPHORYLATION BY PROTEIN·TYROSINE KINASES . . . . . . . ... . . . .. . . . . . .

922

CONCLUSIONS AND PROSPECTS . ... . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

923

PERSPECTIVES AND SUMMARY Phosphate tightly associated with protein has been known since the late nineteenth century. Early hints that this phosphate might be covalently linked were glean,ed in protein until

1906, but the first phospho amino acid was not isolated from 1933. Since then a variety of covalent linkages of phosphate to

proteins have been found. The most common involve esterification of phos· phate to serine and threonine, with smaller amounts being linked to lysine, arginine, histidine, aspartic acid, glutamic acid, and cysteine. Even though the existence of phosphodiester bonds linking tyrosine to mono- or polyribonu'Present address: Fred Hutchinson Cancer Research Center, 1124 Columbia Street, Seattle, Washington 98104

897 0066-4154/85/0701-0897$02.00

898

HUNTER & COOPER

cleotides had set a precedent, reports of orthophosphate esterified to tyrosine did not emerge until 1979 (1). Phosphotyrosine (P.Tyr) was first detected in hydrolysates of viral transforming proteins labeled by incubation of immuno precipitates with radioactive ATP (1-3) . Initially it was suggested that this might either be an aberrant in vitro reaction or represent a phosphoenzyme intermediate, but when suitable separation techniques for P.Tyr were de


veloped it became clear that all animal cells do il)deed contain low levels of P.Tyr in protein (2, 4). The previous failure to detect P.Tyr upon conventional analysis of protein hydrolysates had been due to its masking by the much more abundant phosphothreonine (P.Thr). The occurrence of phosphorylated tyrosine as a stable yet reversible mod ification of proteins implies the existence of one or more protein kinases capable of phosphorylating tyrosine and also of protein phosphatases capable of hydrolyzing P. Tyr. The last five years have seen a burgeoning of systems in which tyrosine phosphorylation has been implicated, including viral trans formation and growth control. Already six cellular genes have been identified that encode tyrosine-specific protein kinases, and there are indications of several more. Some of these protein-tyrosine kinases (PTK) have been purified and extensively characterized. The first PTK activities to be detected were associated with viral transform ing proteins, particularly those of acutely oncogenic retroviruses. Although virally coded, the provenance of these transforming proteins is ultimately cellular, since the transforming genes of these viruses are recognizable as coopted cellular genes [for review see (5)]. There are at least five cellular genes

of this kind, and they all appear to encode proteins with PTK activity. The second type of PTK activity was detected in association with .growth factor receptors (6). The receptor PTKs are stimulated by their respective ligands, and there are at least four enzymes of this sort. At the outset there was some concern whether the PTK activities of the two types of protein were intrinsic or due to associated enzymes. It is now clear, however, that these proteins do indeed all have a catalytic domain capable of transferring phosphate to proteins. The most persuasive evidence comes in those cases where primary amino acid sequences have been predicted from molecularly cloned genes. A stretch of about 260 amino acids within each of these proteins has obvious homology with the sequence of the catalytic subunit of the serine-specific cyclic AMP-dependent protein kinase (A-kinase). De spite this homology the PTKs all show an exquisitely strict specificity for tyrosine in protein. Genes corresponding to the PTK genes are readily recognizable in Drosophi

la (7, 8 ) . From the fact that PTK activity can be detected in insects (8) and in even simpler eukaryotic species (9), it seems likely that these genes have the same function as in higher eukaryotes. The ancient evolutionary origin and the

899

PROTEIN-TYROSINE KINASES

high degree of cConservation of these genes suggest that tyrosine phosphoryla tion plays an important role in eukaryotic cellular physiology. By analogy with the documented effects of phosphorylation on serine and threonine, Ithe reversible phosphorylation of proteins on tyrosine would be expected to modulate protein function. Protein phosphatases specific for P. Tyr have been reported [for review see (10)] fulfilling a critical requirement for a regulatory system based on tyrosine phosphorylation. Nevertheless, although


some progn!ss has been made in identifying intracellular substrates for PTKs, in no case has the phosphorylation been of physiological significance. While there is little doubt that the enzymes in question can and do phos phorylate proteins on tyrosine, there are recent suggestions that they may also phosphorylate nonprotein substrates, such as phosphatidylinositol (12,

13). If

these activities prove to be intrinsic, the true function of PTKs will have to be reassessed. The aim of this review is to describe our current knowledge of PTKs with emphasis on the enzymes themselves rather than their substrates.

TWO TYPES OF PROTEIN-TYROSINE KINASE Initially it appeared that one could classify the PTKs based on the distinction between those whose genes have become part of retroviruses and those associ ated with growth factor receptors. Recently, however, it has become apparent that there is overlap between these two types (11). Nevertheless it is still useful to group the growth factor receptor PTKs separately because their regulation is a distinct feature (Table 1). For the purposes of this review we have selected a representative of each type for detailed consideration; these are pp60v-src and the epidenn al growth factor (EGF) receptor respectively. pp60v-src, the 60,000dalton product of the Rous sarcoma virus (RSV)

sre

gene, is the prototypic

retroviral PTK, while the EGF receptor is by far the best understood growth factor receptor PTK. The growing number of other PTKs that do not fit naturally into either category will be discussed separately.

RETROVIRAL PROTEIN-TYROSINE KINASES AND THEIR CELLULAR HOMOLOGS The products of the v-sre (2, 14, 15), v-yes (16), vjgr (17), v-fps (18-20), v

jes (21), v-abl (3, 22),

and v-ros (23) retroviral oncogenes (v-one genes) all

have PTK activities (Table 1). The v-sre, v-yes, v-jps, and v-ros genes are the oncogenes .of different chicken sarcoma viruses, the v1gr and v-fes genes are the oncogenes of cat sarcoma viruses, and the v-abl gene is the oncogene of a mouse lymphoma virus. These v-one genes originated from cellular genes (c-onc genes) of the species in question. The nucleotide sequences of all the

900

HUNTER & COOPER

Table 1

Protein-tyrosine kinases·

Protein

Gene Virus-related


c -src c-yes c-fgr c-fps c-fes c-ab/ c-ros

c

pp60c-sr

:fj>s p98c p92c:fes bl p150c-a

Intracellular location

Membrane

Soluble Cytoplasm

Gene

v-src

Protein c

pp60v-sr

r P70"ag:fg :fps "ag 40 P1 :fes P8ssag P120"ag-abl ros P68gag-

v-erb-B

v-er gp68174

Growth factor receEtors

c-erb-B

EGF receptor PDGF receptor Insulin receptor IGF-l receptor

Others

p5 6)stra p75

Plasma membrane Plasma membrane Plasma membrane Plasma membrane

Membrane/cytoskeleton Membrane Cytoplasm Membrane/soluble Cytoplasm Membrane Membrane

P90"ag-yes

v-yes v-fgr v-fps v-fes v-ab/ v-ros

Intracellular location

b-B

Membrane

Membrane Cytoplasm

'The genetic origins (references in text) and subcellular localizations [(57-63,200,202,219,221, reviewed in 218)] of known cellular (left columns) and their corresponding viral (right columns) protein-tyrosine kinases are

listed.

viral genes and some of their cellular counterparts have been determined from molecular clones (24-34). Several points emerge from the predicted protein sequences. From their very high degree of conservation (31, 32) and other evidence (35), it appears likely that the c1ps and c-fes genes are homologous chicken and cat genes respectively, leaving us with six distinct cellular genes encoding PTKs of this sort. Although the complete sequences of most of the cellular gene products are not available : we can deduce a great deal from the sequences of the viral proteins. The different proteins are all related over a continuous stretch of about 260 amino acids, but are more or less divergent over the rest of the protein (Figure 1). For pp60v-src this region comprises its COOH-terminal half, the part of the protein shown to be sufficient for phos phate transfer. This is also the region of pp60v-src which is homologous to the A-kinase catalytic subunit (36, 37). A comparable 'catalytic domain' is found in various positions within the other viral transforming proteins. The sequences


PROTEIN-TYROSINE KINASES

901

outside th e catalyti c domain are pres umed to determine in large part the unique properties of the ind iv id ual members of this gene fam ily. The prod ucts of the v-src, v-yes, v-fgr, v-fps, v-fes, v-abl, and v-ros genes were all sh own to have PTK activity (2,14-23) by assay in immunoprecipitates (38,3 9) . In s uch 'solid state ' reactions e ither the heavy chain of the combining antibody or else the transforming protein itself is phosphorylated . While this technique serves as a rapid diagnosis for whether a given oncogene prod uct is likely to b e a PTK , it is not very useful for studying the protein's enzymati c properties. This requires conventional purifi cation of the protein in a sol uble form. Unfortunately, in contrast to other protein k inases, which are us ually isol ated from tissues,it is hard to obtain large amo unts of v irally transformed cells ,and none of the viral transforming proteins is very ab undant. An alternate and potentially copio us so urce of protein wo uld be bacteria expressing on cogenes from recombinant plasmids. The v-src, v-abl, and v-fps prote ins have all been expressed in bacteria and shown to have PTK activity ,providing strong s upport for the idea that this activity is innate (40-42). It has been hard, however, to obtain the expressed protei ns in a soluble state ,although progress has been made in p urifying an enzymati cally active fragment of the v-abl protein synthesized in E. coli (43). Despite the frequency of on cogenes en coding PTKs ,it should be stressed that not all retroviral on cogene prod ucts have this activity. Indeed the predi cted sequen ces of the v-sis, v-ras, v-myc, v-myb, v-fos, v-ski, v-rei, v-erb-A, and v-ets oncogenes have little or no relationship to those of the PTK family,nor do their prod ucts have dete ctable protein kinase activity. Intriguing ly ,however, the products of the v-mil (orv-mht) (44), v-raf(45),v-fms(46),v-mas(47),and v-erb-B (48) on cogenes have se quen ce homology with the catalyti c domain of the protein kinases ,and yet the se proteins do not have dete ctable PTK activity in immun opre cipitates . Nevertheless , since the v-erh-B gene prod uct has turned out to be a fragment of the EGF re ceptor PTK (11) , one sho uld be cautio us i n dismissing the possibility that the v-mil, v-raj, v -fms, and v-mas proteins are PTKs with covert activity or phosphotransferases with another specificity. We may simply lack the appropriate assay for demonstrating s uch activities. pp60v-src AND SYNTHESIS OF pp60v-src The complete sequen ces of the v-src genes from two strains of RSV and the chicken c-src gene are known (24-28) . pp60v-src is a protein of 526 amino acids which is synthesized on sol uble ri bosomes , but whi ch rapidly associates with membranes following synthesis ( 49). The mature protein is indisting uishable in size from the primary translation product (50,51),whi ch implies that association of pp60v-src w ith STRUCTURE


PhK -Y cGPK cAPK

( 1 J TRDAALRl SHS'I'IIG FYENYEPKEILGroVSSWRRCIHKEP'l'CKE • • • • • YAVKII OV'll:lGG SFSAEEVOELREATr.KEVDILRKV� UPN IrQ Lf(J)T 3l8/DSFKHLIGG LDDVSNKAYEDAEAKAKYEAEAAFFAN LKLSDFNIIDTLGVGGFGRVELVQLKSEESKT • • • • • FAMKILKKRHI • • • • • • VDTRQQEH IRSEKQIMQGAH.SDFIVRLYRT (1) G NMAKKG SEQESVKEFLAKAKEDFLKKW ENPAQNTAU LOOFERI KTLG'It;SFG RVMLVKHME'It; NH • • • • • • YAMKI LDKOKV • • • • • • VKLKQ lEU TLNE KRILQAVN .FPFLVKL EFS l.G G V E E l *A n * ** L •

tG G FG V G � 226/LVAYYSKHADGLCHRLANVCPTSKPQ"l';lGLAf(J)AW E IPRESLRLEAKLGQGC FGEVWIIG'lWNDTTR • .a • • • • • GRYSSESDYWSFGILLWETFSLGASPliPNLSNQQTREFVEI