489
Biochimica et Biophysica Acta, 1052 (1990) 489-498 Elsevier BBAMCR 12687
Activation of the EGF receptor tyrosine kinase by divalent metal ions: comparison of holoreceptor and isolated kinase domain properties J o h n G. K o l a n d a n d R i c h a r d A. C e r i o n e Department of Pharmacology, NYS College of Veterinary Medicine, Cornell University, Ithaca, N Y (U.S.A.) (Received 22 September 1989)
Key words: Epidermal growth factor; EGF receptor; Tyrosine kinas¢; Receptor activation; Metal ion; Holoreceptor
The activation of the epidermal growth factor (EGF) receptor tyrosine kinase activity is thought to represent a key initial step in EGF-mediated mitogenesis. The mechanisms underlying the regulation of the EGF receptor tyrosine kinase activity were examined through comparisonsof the holoreceptor, purified from human placenta, and a soluble 42 kDa tyrosine kinase domain (TKD), generated by the limited trypsin proteolysis of the holoreceptor. The results of these studies highlight the importance of divalent metal ions (Me2+), i.e., Mnz+ and Mg 2+, as activators of the tyrosine kinase activity. Manganese is an extremely effective activator of the holoreceptor tyrosine kinase, and under some conditions (low ionic strength) it completely alleviates the need for EGF to stimulate activity. In contrast, Mg 2+ only weakly stimulates the holoreceptor tyrosine kinase activity in the absence of EGF, but promotes essentially full activity in the presence of the growth factor. Like the holoreceptor, the soluble TKD is highly active in the presence of Mn 2+. However, the isolated TKD is completely inactive in the presence of Mg 2+, and, in fact, Mg 2+ inhibits the Mn 2+-stimulated tyrosine kinase activity. The differences in the effects of Mnz + and Mg z+ on the isolated TKD were further demonstrated by monitoring the effects of Me 2+ on the modification of a reactive cysteine residue(s) on the TKD. While Mn2+ potentiates the inhibition by cysteine-directed reagents of the tyrosine kinase activity, Mg 2+ has no effect on either the rate or the extent of the inhibition. Both the regulation by Mn2+ of the kinase activity of the TKD and the potentiation by Mn2+ of the cysteine reactivity of the TKD occur over a millimolar concentration range, which implicates a direct binding interaction by the metal ion. Overall, these results demonstrate that there are two key activator sites on the EGF receptor, i.e., the EGF binding site on the extracellular domain and a Me z + binding site on the cytoplasmic TKD. Me z+ interactions with the cytoplasmic kinase domain apparently result in conformational changes which regulate the levels of tyrosine kinase activity, influence the degree to which this activity is responsive to EGF, and probably account for the effects of Me 2+ on the aggregation state of the receptor (Carraway, ILL., III, Koland, J.G. and Cerlone, R.A. (1989) J. Biol. Chem. 264, 8699-8707). In general, MgZ+-induced conformation changes prime the receptor for activation by EGF, while Mn2+ can fully activate the receptor tyrosine kinase and thereby short-circuit growth factor control.
Introduction The diverse responses to EGF are mediated by a transmembrane receptor protein and its associated Abbreviations: EGF, epidermal growth factor; TKD, the 42 kDa tyrosine kinase domain generated by limited trypsin proteolysis of the EGF holoreceptor; GAT, a random copolymer of glutamate, alanine and tyrosine (see Materials and Methods); SDS-PAGE, sodium dodecyl sulfate-polyacrylamide gel electrophoresis; Hepes, 4-(2-hydroxyethyl)-l-piperazine-ethanestdfonic acid; NEM, N-ethylmaleimide; and MIANS, 2-(4'-maleimidylanilino)naphthalene-6-sulfonic acid.
tyrosine kinase activity [1-3]. The EGF receptor is comprised of a single 170 kDa glycosylated polypeptide, which forms an extracellular growth factor binding do. main, an intracellular tyrosine kinase domain, and a connecting transmembrane sequence [2,4]. Upon the binding of EGF to the extracellular domain, the tyrosine kinase activity of the cytoplasmic domain is stimulated. The mechanism of this allosteric control of the receptor tyrosine kinase activity is not fully understood: both intermolecular [5-7] and intramolecular [8,9] mechanisms have been proposed. The central premise of any intermolecular activation mechanism is that EGFinduced changes in the aggregation of extracellular do-
016%4889/90/$03.50 © 1990 Elsevier Science Publishers B.V. (Biomedical Division)
490 mains can control those interactions between cytoplasmic tyrosine kinase domains which modulate tyrosine kinase activity. For example, one such model proposes that the binding of EGF induces the aggregation of inactive receptor monomers to form an active dimeric receptor [2,6]. Recently, we have reported [10], based on fluorescence resonance energy transfer studies, that the EGF receptors on the surfaces of isolated A431 cell membranes aggregate under conditions where these receptors are active tyrosine kinases. However, we also have found that the monomeric form of the purified EGF receptor from human placenta, in detergent solution, is capable of tyrosine kinase activity [8]. Taken together, these results suggest that receptor aggregation is not prerequisite for kinase activity, although it may play an important role in the regulation of this activity, or it may be necessary for other aspects of the EGF-signaling pathway. During the course of our studies both with the purified EGF receptor in detergent solution [8], and in intact membranes [10], divalent metal ions were found to be necessary for tyrosine kinase activity and for receptor aggregation. A primary aim of the present study was to better clarify the roles of divalent metal ions in the activation of the EGF receptor tyrosine kinase. Our approach involved steady-state kinetic studies of both the purified holoreceptor from human placenta and a soluble form of the tyrosine kinase domain (designated below as the TKD), which was generated by limited proteolysis of the holoreceptor [11]. This isolated kinase domain, molecular mass approx. 42 kDa, is free from the regulatory influence of the growth factor binding site. The results of these studies suggest the occurrence of a direct binding interaction between divalent metal ions and the cytoplasmic domain of the EGF receptor. These metal ion interactions, which appear to occur independently of the binding of either EGF or ATP, have important regulatory effects on the tyrosine kinase activities of both the EGF holoreceptor and the soluble TKD. They also are likely to be responsible for the induction of EGF receptor aggregation in membranes [10]. Whereas both Mn2÷ and Mg 2+ appear to be capable of binding to the cytoplasmic domain of the holoreceptor, as well as directly to the soluble TKD, these metal ions show striking differences in their ability to stimulate tyrosine kinase activity and to support growth factor control of this activity. Materials and Methods
Materials Ammonium sulfate (grade III), ATP (vanadium-free), N-ethylmaleimide (NEM), Triton X-100, diphenylcarbamyl chloride-treated trypsin, soybean trypsin inhibitor, and (Glu : Ala : Tyr)6 : 3:1 (GAT), a random
copolymer of glutamate, alanine and tyrosine in the indicated mole ratios and average molecular mass of 43 kDa were purchased from Sigma. Murine EGF was supplied by Boehringer-Mannheim Biochemicals. 2-(4'M a l e i m i d y l a n i l i n o ) n a p h t h a l e n e - 6 - s u l f o n i c acid (MIANS) was acquired from Molecular Probes, and [y-32p]ATP was supplied by Dupont-New England Nuclear.
Purification of the EGF receptor and isolation of the EGF receptor TKD The human placental EGF receptor was solubilized in Triton X-100 and purified as previously described [8]. The 42 kDa TKD fragment of the receptor was generated by limited trypsin proteolysis as described by Basu et al. [11]. Purified receptor (5-20 pmol) was digested with 0.5-20 /~g/ml trypsin in buffer A (40 mM Hepes/Na, 0.05% Triton X-100, 10% (v/v) glycerol (pH 7.4)) for 30 rain at room temperature. The digestion was stopped by the adidtion of excess soybean trypsin inhibitor, and the extent of tryptic digestion was monitored by autophosphorylation with [y-32p]ATP, SDSPAGE, and autoradiography [8]. The kinase domain was isolated by gel-filtration chromatography on an Ultrogel AcA-34 (IBF Biotechnics) column equilil~rated and eluted with 20 mM Hepes/Na, 0.05% Triton X-I00, 25 mM NaC1, 10% (v/v) glycerol (pH 7.4). Column fractions containing the peak of GAT phosphorylation activity (see below) were pooled and frozen at - 7 0 °C until use. Tyrosine kinase activity assays The initial velocities of the tyrosine kinase reactions catalyzed by the EGF holoreceptor and the isolated TKD were determined by assaying the incorporation of 32Pi into GAT which occurred in a 5 min period at room temperature as previously described [8]. The phosphorylation reaction was determined to be linear over this period. Concentrations of [y-32P]ATP (5000-10000 cpm/pmol), MnC12 and MgCI 2 were varied in an assay medium which consisted of buffer A supplemented with 0.1 mg/ml bovine serum albumin. The assay medium contained 0.4 m g / m l GAT and 10 /~M [y-32p]ATP unless otherwise indicated in the text. The K M for the substrate GAT under the conditions of these assays was determined to be - 10/xg/ml. Activities are expressed per ml of receptor or TKD preparation. Steady-state kinetics analyses The equilibrium random mechanism for a two substrate enzyme reaction yields a steady-state rate equation in the form of Eqn. 1 (see Results), and the rate equation corresponding to the general two substrate mechanism can be similarly derived [12]. These equations have been used to describe the Me 2+ and ATP dependencies of the initial velocity of Me2+-activated
491 enzymes which use ATP as a substrate [13]. These rate equations were fit to the data of Fig. 5 by a nonlinear least-squares algorithm [14], which yielded the best fit kinetic constants quoted in the text. The general and random rate equations fit the data equally well, although Eqn. 1 includes one less parameter. The family of theoretical curves shown in Fig. 5 was generated from Eqn. 1, using the determined kinetic constants. Although ATP in solution is known to exist in various protonation states and also as a complex in the presence of Mn 2÷, the bulk ATP and Mn 2÷ concentrations were used in these rate equations. Under the conditions of the assays of Fig. 5 the predominant nucleotide form was M n A T P 2-, with less than 5% existing as M n H A T P and uncomplexed nucleotide species [15]. Since Mn 2÷ when added was in a large excess relative to ATP, the Mn 2 + concentration terms reflect the effects of the free form of this metal ion. Chemical modification of the TKD by N E M and M I A N S Purified T K D (12 fmol) was diluted into buffer A supplemented as indicated in the text and preincubated for 5 min at room temperature. The cysteine-modification reactions were initiated either by the addition of N E M from a 1.0 mM stock solution (prepared immediately before use) to a concentration of 200 #M, or by the serial dilution of a 10 m M MIANS stock solution to a concentration of 20/~M. The 25 /xl reactions were allowed to proceed for the indicated periods and then were quenched by the addition of 11 #1 Buffer A supplemented with 20 m M DTT, 40 m M MnCI 2 and 0.4 m g / m l bovine serum albumin. After a 5 rain incubation, the G A T phosphorylation assay (see above) was initiated by the addition of [y-32p]ATP and G A T to final concentrations of 10 # M and 0.4 m g / m l , respectively (final volume 45 #1). The time course of inhibition of the T K D activity by N E M was accurately fit by the equation A ( t ) = ( A o - A ~ ) . e x p ( - k ' t ) + A~, where k ' is the pseudo-first-order rate constant for the inhibition reaction. The values of k', A 0 and Aoo were determined by a nonlinear least-squares analysis [14]. A~o was determined to be 0.12-A 0. Hence the activity remaining after a prolonged exposure to 200 # M N E M was a small but finite fraction of the initial activity. The dependence of k ' on Mn 2+ concentration was fit by nonlinear least-squares [14] to a hyperbolic binding equation which yielded the K d value for the Mn 2+ - T K D interaction. Results
Generation of an active 42 kDa TKD by trypsin proteolysis of the purified human placental E G F receptor It was previously shown by Basu et al. [11] that the limited proteolysis of the E G F receptor purified from A431 carcinoma cells yields a 42 kDa fragment with
[Trypsin]
0 0.5
2
5
20 0
(l.tg/ml)
170
42
kDa -~
kD a
--~
Fig. 1. Limited trypsin protcolysis of the EGF receptor. Purified placental EGF receptor (1 pmol) was subjected to limited proteolysis with the indicated concentration of trypsin as described in Materials and Methods. Samples were then autophosphorylated for 10 min at room temperature in the presence of 10 mM MnCl 2 and 10 /~M [-t-22p]ATP (see Materials and Methods). Autophosphorylation was quenched by the addition of SDS-PAGE sample buffer and the extent of proteolysis and formation of the TKD was analyzed by SDS-PAGE and autoradiography [8]. Because the major C-terminal sites of tyrosine autophosphorylation are absent from the TKD fragment, the yield of this domain is greatly underestimated by autoradiography [16].
tyrosine kinase actiyity. The tryptic digestion of the detergent-solubilized E G F receptor from h u m a n placenta also yielded an active 42 kDa tyrosine kinase (see Fig. 1, lane 3). This 42 kDa polypeptide represents an internal portion of the cytosolic tyrosine kinase domain (TKD); the major autophosphorylation sites [16] were removed by an initial C-terminal clip, which converted the receptor from the native 170 kDa form (Fig. 1, lane 1) to an - 1 5 5 kDa form (lane 2). The N-terminal growth factor binding domain was subsequently removed by a cleavage on the cytosolic side of the transmembrane sequence. The exact limits of this soluble T K D are not known. However, some estimates can be made based on biochemical and immunological comparisons of the T K D with the holoreceptor. Specifically, the amino terminus of the T K D is likely to be downstream from the primary phosphorylation site for protein kinase C, i.e., threonine 654 [17,18], since purified preparations of protein kinase C did not phosphorylate-the trypsin-generated T K D (data not shown). The
492 carboxyl terminus of the TKD is likely to extend to at least the phenylalanine residue at position 944 in the holoreceptor. This is based on the finding that a polyclonal antibody, which appears to react within a region bordered by residues 943 and 1011 as assessed by Western analyses of various recombinant fragments of the EGF receptor generated in E. coli, also reacts with the TKD (data not shown). The fact that the TKD lacks the major autophosphorylation sites led to a significant underestimation of the yield of the active TKD as assessed by autoradiographic analysis of the autophosphorylation activity. The TKD is capable of only a weak autophosphorylation reaction which was less than 1/10th as effective as the autophosphorylation of the intact holoreceptor (see Fig. 1). This may reflect the weak phosphorylation of a conserved tyrosine residue which is present within the core of the postulated tyrosine kinase active site (i.e., at position 845). However, the TKD was able to effectively phosphorylate exogenous substrates. Thus it was subsequently isolated by gel-filtration chromatography (see Materials and Methods) with an overall yield of 30-50%, as determined by assaying the Mn2+-stimulated phos-
phorylation activity with the tyrosine-containing synthetic peptide substrate (Glu : Ala : Tyr)6: 3:] (GAT), The TKD appeared to be fully soluble. Unlike the holoreceptor, the TKD could not be functionally inserted into phospholipid bilayer vesicles and did not associate with phospholipid vesicles during sucrose density gradient sedimentation (data not shown).
Concentration dependence of the TKD activity The role of receptor-receptor interactions in the activation of the EGF receptor tyrosine kinase activity has been controversial (cf. Refs. 5-9). In order to examine whether the activation of the isolated TKD by Mn2+ resulted from the reversible self-aggregation of the protein, the isolated TKD was assayed as its concentration was varied over a 75-fold range (Fig. 2). The data presented in Fig. 2 illustrate that the tyrosine kinase activity of the TKD was not stimulated by increasing the protein concentration. Specifically, the activity was essentially linearly dependent on protein concentration, at the lower levels of protein, and then was attenuated at the higher protein concentrations. The latter effect was probably due to substrate depletion which occurred at the higher levels of the TKD. A concentration-dependent activation of the tyrosine kinase activity (i.e., an upward curvature in Fig. 2) would have indicated that the aggregation of the TKD was prerequisite for kinase activity. Thus these results suggested that full activation of the tyrosine kinase activity occurs within the monomeric TKD.
A
~3
E m
o
E Q. 3
:Q2 < Q b3
m
I"
Biochimica et Biophysica Acta, 1052 (1990) 489-498 Elsevier BBAMCR 12687
Activation of the EGF receptor tyrosine kinase by divalent metal ions: comparison of holoreceptor and isolated kinase domain properties J o h n G. K o l a n d a n d R i c h a r d A. C e r i o n e Department of Pharmacology, NYS College of Veterinary Medicine, Cornell University, Ithaca, N Y (U.S.A.) (Received 22 September 1989)
Key words: Epidermal growth factor; EGF receptor; Tyrosine kinas¢; Receptor activation; Metal ion; Holoreceptor
The activation of the epidermal growth factor (EGF) receptor tyrosine kinase activity is thought to represent a key initial step in EGF-mediated mitogenesis. The mechanisms underlying the regulation of the EGF receptor tyrosine kinase activity were examined through comparisonsof the holoreceptor, purified from human placenta, and a soluble 42 kDa tyrosine kinase domain (TKD), generated by the limited trypsin proteolysis of the holoreceptor. The results of these studies highlight the importance of divalent metal ions (Me2+), i.e., Mnz+ and Mg 2+, as activators of the tyrosine kinase activity. Manganese is an extremely effective activator of the holoreceptor tyrosine kinase, and under some conditions (low ionic strength) it completely alleviates the need for EGF to stimulate activity. In contrast, Mg 2+ only weakly stimulates the holoreceptor tyrosine kinase activity in the absence of EGF, but promotes essentially full activity in the presence of the growth factor. Like the holoreceptor, the soluble TKD is highly active in the presence of Mn 2+. However, the isolated TKD is completely inactive in the presence of Mg 2+, and, in fact, Mg 2+ inhibits the Mn 2+-stimulated tyrosine kinase activity. The differences in the effects of Mnz + and Mg z+ on the isolated TKD were further demonstrated by monitoring the effects of Me 2+ on the modification of a reactive cysteine residue(s) on the TKD. While Mn2+ potentiates the inhibition by cysteine-directed reagents of the tyrosine kinase activity, Mg 2+ has no effect on either the rate or the extent of the inhibition. Both the regulation by Mn2+ of the kinase activity of the TKD and the potentiation by Mn2+ of the cysteine reactivity of the TKD occur over a millimolar concentration range, which implicates a direct binding interaction by the metal ion. Overall, these results demonstrate that there are two key activator sites on the EGF receptor, i.e., the EGF binding site on the extracellular domain and a Me z + binding site on the cytoplasmic TKD. Me z+ interactions with the cytoplasmic kinase domain apparently result in conformational changes which regulate the levels of tyrosine kinase activity, influence the degree to which this activity is responsive to EGF, and probably account for the effects of Me 2+ on the aggregation state of the receptor (Carraway, ILL., III, Koland, J.G. and Cerlone, R.A. (1989) J. Biol. Chem. 264, 8699-8707). In general, MgZ+-induced conformation changes prime the receptor for activation by EGF, while Mn2+ can fully activate the receptor tyrosine kinase and thereby short-circuit growth factor control.
Introduction The diverse responses to EGF are mediated by a transmembrane receptor protein and its associated Abbreviations: EGF, epidermal growth factor; TKD, the 42 kDa tyrosine kinase domain generated by limited trypsin proteolysis of the EGF holoreceptor; GAT, a random copolymer of glutamate, alanine and tyrosine (see Materials and Methods); SDS-PAGE, sodium dodecyl sulfate-polyacrylamide gel electrophoresis; Hepes, 4-(2-hydroxyethyl)-l-piperazine-ethanestdfonic acid; NEM, N-ethylmaleimide; and MIANS, 2-(4'-maleimidylanilino)naphthalene-6-sulfonic acid.
tyrosine kinase activity [1-3]. The EGF receptor is comprised of a single 170 kDa glycosylated polypeptide, which forms an extracellular growth factor binding do. main, an intracellular tyrosine kinase domain, and a connecting transmembrane sequence [2,4]. Upon the binding of EGF to the extracellular domain, the tyrosine kinase activity of the cytoplasmic domain is stimulated. The mechanism of this allosteric control of the receptor tyrosine kinase activity is not fully understood: both intermolecular [5-7] and intramolecular [8,9] mechanisms have been proposed. The central premise of any intermolecular activation mechanism is that EGFinduced changes in the aggregation of extracellular do-
016%4889/90/$03.50 © 1990 Elsevier Science Publishers B.V. (Biomedical Division)
490 mains can control those interactions between cytoplasmic tyrosine kinase domains which modulate tyrosine kinase activity. For example, one such model proposes that the binding of EGF induces the aggregation of inactive receptor monomers to form an active dimeric receptor [2,6]. Recently, we have reported [10], based on fluorescence resonance energy transfer studies, that the EGF receptors on the surfaces of isolated A431 cell membranes aggregate under conditions where these receptors are active tyrosine kinases. However, we also have found that the monomeric form of the purified EGF receptor from human placenta, in detergent solution, is capable of tyrosine kinase activity [8]. Taken together, these results suggest that receptor aggregation is not prerequisite for kinase activity, although it may play an important role in the regulation of this activity, or it may be necessary for other aspects of the EGF-signaling pathway. During the course of our studies both with the purified EGF receptor in detergent solution [8], and in intact membranes [10], divalent metal ions were found to be necessary for tyrosine kinase activity and for receptor aggregation. A primary aim of the present study was to better clarify the roles of divalent metal ions in the activation of the EGF receptor tyrosine kinase. Our approach involved steady-state kinetic studies of both the purified holoreceptor from human placenta and a soluble form of the tyrosine kinase domain (designated below as the TKD), which was generated by limited proteolysis of the holoreceptor [11]. This isolated kinase domain, molecular mass approx. 42 kDa, is free from the regulatory influence of the growth factor binding site. The results of these studies suggest the occurrence of a direct binding interaction between divalent metal ions and the cytoplasmic domain of the EGF receptor. These metal ion interactions, which appear to occur independently of the binding of either EGF or ATP, have important regulatory effects on the tyrosine kinase activities of both the EGF holoreceptor and the soluble TKD. They also are likely to be responsible for the induction of EGF receptor aggregation in membranes [10]. Whereas both Mn2÷ and Mg 2+ appear to be capable of binding to the cytoplasmic domain of the holoreceptor, as well as directly to the soluble TKD, these metal ions show striking differences in their ability to stimulate tyrosine kinase activity and to support growth factor control of this activity. Materials and Methods
Materials Ammonium sulfate (grade III), ATP (vanadium-free), N-ethylmaleimide (NEM), Triton X-100, diphenylcarbamyl chloride-treated trypsin, soybean trypsin inhibitor, and (Glu : Ala : Tyr)6 : 3:1 (GAT), a random
copolymer of glutamate, alanine and tyrosine in the indicated mole ratios and average molecular mass of 43 kDa were purchased from Sigma. Murine EGF was supplied by Boehringer-Mannheim Biochemicals. 2-(4'M a l e i m i d y l a n i l i n o ) n a p h t h a l e n e - 6 - s u l f o n i c acid (MIANS) was acquired from Molecular Probes, and [y-32p]ATP was supplied by Dupont-New England Nuclear.
Purification of the EGF receptor and isolation of the EGF receptor TKD The human placental EGF receptor was solubilized in Triton X-100 and purified as previously described [8]. The 42 kDa TKD fragment of the receptor was generated by limited trypsin proteolysis as described by Basu et al. [11]. Purified receptor (5-20 pmol) was digested with 0.5-20 /~g/ml trypsin in buffer A (40 mM Hepes/Na, 0.05% Triton X-100, 10% (v/v) glycerol (pH 7.4)) for 30 rain at room temperature. The digestion was stopped by the adidtion of excess soybean trypsin inhibitor, and the extent of tryptic digestion was monitored by autophosphorylation with [y-32p]ATP, SDSPAGE, and autoradiography [8]. The kinase domain was isolated by gel-filtration chromatography on an Ultrogel AcA-34 (IBF Biotechnics) column equilil~rated and eluted with 20 mM Hepes/Na, 0.05% Triton X-I00, 25 mM NaC1, 10% (v/v) glycerol (pH 7.4). Column fractions containing the peak of GAT phosphorylation activity (see below) were pooled and frozen at - 7 0 °C until use. Tyrosine kinase activity assays The initial velocities of the tyrosine kinase reactions catalyzed by the EGF holoreceptor and the isolated TKD were determined by assaying the incorporation of 32Pi into GAT which occurred in a 5 min period at room temperature as previously described [8]. The phosphorylation reaction was determined to be linear over this period. Concentrations of [y-32P]ATP (5000-10000 cpm/pmol), MnC12 and MgCI 2 were varied in an assay medium which consisted of buffer A supplemented with 0.1 mg/ml bovine serum albumin. The assay medium contained 0.4 m g / m l GAT and 10 /~M [y-32p]ATP unless otherwise indicated in the text. The K M for the substrate GAT under the conditions of these assays was determined to be - 10/xg/ml. Activities are expressed per ml of receptor or TKD preparation. Steady-state kinetics analyses The equilibrium random mechanism for a two substrate enzyme reaction yields a steady-state rate equation in the form of Eqn. 1 (see Results), and the rate equation corresponding to the general two substrate mechanism can be similarly derived [12]. These equations have been used to describe the Me 2+ and ATP dependencies of the initial velocity of Me2+-activated
491 enzymes which use ATP as a substrate [13]. These rate equations were fit to the data of Fig. 5 by a nonlinear least-squares algorithm [14], which yielded the best fit kinetic constants quoted in the text. The general and random rate equations fit the data equally well, although Eqn. 1 includes one less parameter. The family of theoretical curves shown in Fig. 5 was generated from Eqn. 1, using the determined kinetic constants. Although ATP in solution is known to exist in various protonation states and also as a complex in the presence of Mn 2÷, the bulk ATP and Mn 2÷ concentrations were used in these rate equations. Under the conditions of the assays of Fig. 5 the predominant nucleotide form was M n A T P 2-, with less than 5% existing as M n H A T P and uncomplexed nucleotide species [15]. Since Mn 2÷ when added was in a large excess relative to ATP, the Mn 2 + concentration terms reflect the effects of the free form of this metal ion. Chemical modification of the TKD by N E M and M I A N S Purified T K D (12 fmol) was diluted into buffer A supplemented as indicated in the text and preincubated for 5 min at room temperature. The cysteine-modification reactions were initiated either by the addition of N E M from a 1.0 mM stock solution (prepared immediately before use) to a concentration of 200 #M, or by the serial dilution of a 10 m M MIANS stock solution to a concentration of 20/~M. The 25 /xl reactions were allowed to proceed for the indicated periods and then were quenched by the addition of 11 #1 Buffer A supplemented with 20 m M DTT, 40 m M MnCI 2 and 0.4 m g / m l bovine serum albumin. After a 5 rain incubation, the G A T phosphorylation assay (see above) was initiated by the addition of [y-32p]ATP and G A T to final concentrations of 10 # M and 0.4 m g / m l , respectively (final volume 45 #1). The time course of inhibition of the T K D activity by N E M was accurately fit by the equation A ( t ) = ( A o - A ~ ) . e x p ( - k ' t ) + A~, where k ' is the pseudo-first-order rate constant for the inhibition reaction. The values of k', A 0 and Aoo were determined by a nonlinear least-squares analysis [14]. A~o was determined to be 0.12-A 0. Hence the activity remaining after a prolonged exposure to 200 # M N E M was a small but finite fraction of the initial activity. The dependence of k ' on Mn 2+ concentration was fit by nonlinear least-squares [14] to a hyperbolic binding equation which yielded the K d value for the Mn 2+ - T K D interaction. Results
Generation of an active 42 kDa TKD by trypsin proteolysis of the purified human placental E G F receptor It was previously shown by Basu et al. [11] that the limited proteolysis of the E G F receptor purified from A431 carcinoma cells yields a 42 kDa fragment with
[Trypsin]
0 0.5
2
5
20 0
(l.tg/ml)
170
42
kDa -~
kD a
--~
Fig. 1. Limited trypsin protcolysis of the EGF receptor. Purified placental EGF receptor (1 pmol) was subjected to limited proteolysis with the indicated concentration of trypsin as described in Materials and Methods. Samples were then autophosphorylated for 10 min at room temperature in the presence of 10 mM MnCl 2 and 10 /~M [-t-22p]ATP (see Materials and Methods). Autophosphorylation was quenched by the addition of SDS-PAGE sample buffer and the extent of proteolysis and formation of the TKD was analyzed by SDS-PAGE and autoradiography [8]. Because the major C-terminal sites of tyrosine autophosphorylation are absent from the TKD fragment, the yield of this domain is greatly underestimated by autoradiography [16].
tyrosine kinase actiyity. The tryptic digestion of the detergent-solubilized E G F receptor from h u m a n placenta also yielded an active 42 kDa tyrosine kinase (see Fig. 1, lane 3). This 42 kDa polypeptide represents an internal portion of the cytosolic tyrosine kinase domain (TKD); the major autophosphorylation sites [16] were removed by an initial C-terminal clip, which converted the receptor from the native 170 kDa form (Fig. 1, lane 1) to an - 1 5 5 kDa form (lane 2). The N-terminal growth factor binding domain was subsequently removed by a cleavage on the cytosolic side of the transmembrane sequence. The exact limits of this soluble T K D are not known. However, some estimates can be made based on biochemical and immunological comparisons of the T K D with the holoreceptor. Specifically, the amino terminus of the T K D is likely to be downstream from the primary phosphorylation site for protein kinase C, i.e., threonine 654 [17,18], since purified preparations of protein kinase C did not phosphorylate-the trypsin-generated T K D (data not shown). The
492 carboxyl terminus of the TKD is likely to extend to at least the phenylalanine residue at position 944 in the holoreceptor. This is based on the finding that a polyclonal antibody, which appears to react within a region bordered by residues 943 and 1011 as assessed by Western analyses of various recombinant fragments of the EGF receptor generated in E. coli, also reacts with the TKD (data not shown). The fact that the TKD lacks the major autophosphorylation sites led to a significant underestimation of the yield of the active TKD as assessed by autoradiographic analysis of the autophosphorylation activity. The TKD is capable of only a weak autophosphorylation reaction which was less than 1/10th as effective as the autophosphorylation of the intact holoreceptor (see Fig. 1). This may reflect the weak phosphorylation of a conserved tyrosine residue which is present within the core of the postulated tyrosine kinase active site (i.e., at position 845). However, the TKD was able to effectively phosphorylate exogenous substrates. Thus it was subsequently isolated by gel-filtration chromatography (see Materials and Methods) with an overall yield of 30-50%, as determined by assaying the Mn2+-stimulated phos-
phorylation activity with the tyrosine-containing synthetic peptide substrate (Glu : Ala : Tyr)6: 3:] (GAT), The TKD appeared to be fully soluble. Unlike the holoreceptor, the TKD could not be functionally inserted into phospholipid bilayer vesicles and did not associate with phospholipid vesicles during sucrose density gradient sedimentation (data not shown).
Concentration dependence of the TKD activity The role of receptor-receptor interactions in the activation of the EGF receptor tyrosine kinase activity has been controversial (cf. Refs. 5-9). In order to examine whether the activation of the isolated TKD by Mn2+ resulted from the reversible self-aggregation of the protein, the isolated TKD was assayed as its concentration was varied over a 75-fold range (Fig. 2). The data presented in Fig. 2 illustrate that the tyrosine kinase activity of the TKD was not stimulated by increasing the protein concentration. Specifically, the activity was essentially linearly dependent on protein concentration, at the lower levels of protein, and then was attenuated at the higher protein concentrations. The latter effect was probably due to substrate depletion which occurred at the higher levels of the TKD. A concentration-dependent activation of the tyrosine kinase activity (i.e., an upward curvature in Fig. 2) would have indicated that the aggregation of the TKD was prerequisite for kinase activity. Thus these results suggested that full activation of the tyrosine kinase activity occurs within the monomeric TKD.
A
~3
E m
o
E Q. 3
:Q2 < Q b3
m
I"
