THE JOURNAL OF BIOLOGICAL CHEMISTRY
© 2002 by The American Society for Biochemistry and Molecular Biology, Inc.
Vol. 277, No. 2, Issue of January 11, pp. 1576 –1585, 2002
Printed in U.S.A.
Comparison of the Biochemical and Kinetic Properties of the Type 1
Receptor Tyrosine Kinase Intracellular Domains
DEMONSTRATION OF DIFFERENTIAL SENSITIVITY TO KINASE INHIBITORS*
Received for publication, June 25, 2001, and in revised form, November 2, 2001
Published, JBC Papers in Press, November 5, 2001, DOI 10.1074/jbc.M105907200
Perry S. Brignola‡, Karen Lackey§, Sue H. Kadwell‡, Christine Hoffman‡, Earnest Horne‡,
H. Luke Carter‡, J. Darren Stuart¶, Kevin Blackburn储, Mary B. Moyer储, Krystal J. Alligood**,
Wilson B. Knight¶, and Edgar R. Wood¶‡‡
From the Departments of ¶Systems Research, ‡Gene Expression and Protein Biochemistry, §High Throughput Chemistry,
储Pathway Genomics, and **Cancer Biology, GlaxoSmithKline Inc., Research Triangle Park, North Carolina 27709
Epidermal growth factor receptor (EGFR), ErbB-2,
and ErbB-4 are members of the type 1 receptor tyrosine
kinase family. Overexpression of these receptors, especially ErbB-2 and EGFR, has been implicated in multiple
forms of cancer. Inhibitors of EGFR tyrosine kinase activity are being evaluated clinically for cancer therapy.
The potency and selectivity of these inhibitors may affect the efficacy and toxicity of therapy. Here we describe the expression, purification, and biochemical
comparison of EGFR, ErbB-2, and ErbB-4 intracellular
domains. Despite their high degree of sequence homology, the three enzymes have significantly different catalytic properties and substrate kinetics. For example,
the catalytic activity of ErbB-2 is less stable than that of
EGFR. ErbB-2 uses ATP-Mg as a substrate inefficiently
compared with EGFR and ErbB-4. The three enzymes
have very similar substrate preferences for three optimized peptide substrates, but differences in substrate
synergies were observed. We have used the biochemical
and kinetic parameters determined from these studies
to develop an assay system that accurately measures
inhibitor potency and selectivity between the type 1
receptor family. We report that the selectivity profile of
molecules in the 4-anilinoquinazoline series can be modified through specific aniline substitutions. Moreover,
these compounds have activity in whole cells that reflect the potency and selectivity of target inhibition determined with this assay system.
The type 1 receptor tyrosine kinase family contains four
members, epidermal growth factor receptor (EGFR1 or ErbB1), ErbB-2, ErbB-3, and ErbB-4 (reviewed in Ref. 1). These
receptors consist of an extracellular ligand binding domain, a
single transmembrane domain, an intracellular tyrosine kinase catalytic domain, and a tyrosine-rich cytoplasmic tail. The
* 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.
‡‡ To whom correspondence should be addressed: Dept. of Systems
Research, GlaxoSmithKline, Inc., 5 Moore Dr., Research Triangle Park,
NC 27709. Tel.: 919-483-3910; Fax: 919-483-3895; E-mail: erw39216@
gsk.com.
1
The abbreviations used are: EGFR, epidermal growth factor receptor;
HA, hydroxyapatite; HIC, hydrophobic interaction chromatography;
VEGFR, vascular endothelial growth factor receptor; PKC␨, protein kinase C␨; MOPS, 4-morpholinepropanesulfonic acid; Ab, antibody; BisTris,
2-[bis(2-hydroxyethyl)amino]-2-(hydroxymethyl)propane-1,3-diol; MES,
4-morpholineethanesulfonic acid.
biological activity of these receptors is mediated through the
following signal transduction mechanism. A complex pattern of
ligand/receptor interactions results in the formation of homoand heterodimers among type 1 receptor family members.
Dimerization activates the catalytic domains and allows one
receptor monomer to phosphorylate tyrosine residues on the
cytoplasmic tail of the other monomer in the pair. Phosphorylation of tyrosine residues on the cytoplasmic tail creates specific binding sites for Src homology 2 or phosphotyrosine binding domain containing proteins. The recruitment of these
proteins to the receptor activates signal transduction pathways
that produce the response of the cell to the ligand. EGFR,
ErbB-2, and ErbB-4 all possess active tyrosine kinase catalytic
domains. ErbB-3 is catalytically impaired (2), but it is biologically active because it can form heterodimers with the other
family members.
Members of the type 1 family of receptors have been implicated in cancer. Specifically, overexpression of ErbB-2 and
EGFR is correlated with poor prognosis and reduced overall
survival (3). The role of ErbB-3 and ErbB-4 in cancer is not as
well described. In pre-clinical models of tumor formation, however, co-expression of these receptors with EGFR or ErbB-2
results in a more aggressive transformed phenotype (4, 5). The
therapeutic potential of the type 1 receptors as targets has been
clinically proven for ErbB-2 through the use of HerceptinTM, an
anti-ErbB-2 antibody (6). Inhibitors of the tyrosine kinase activity of EGFR have demonstrated pre-clinical activity in animal tumor models and are being evaluated clinically (7).
The design and discovery of future generations of kinase
inhibitors directed at the type 1 receptor family are progressing
at a rapid rate (8). The catalytic domains of these receptors
have very similar amino acid sequences, and it is likely that
some inhibitors will affect more than one member of the family.
A clear understanding of how these molecules affect the different type 1 receptors is necessary to interpret biological effects
and develop the structure-activity relationships of inhibitor
enzyme selectivity. Biochemical characteristics of the enzyme
assay system such as substrate kinetics and concentration,
peptide substrate sequence, and enzyme activation state can
significantly influence the apparent affinity of a kinase inhibitor. For this reason it is important to have well characterized
enzyme and assay systems for inhibitor analysis.
We describe the expression and purification of the intracellular domains of EGFR, ErbB-2, and ErbB-4 using a baculovirus system. We have directly compared several biochemical
properties of the enzymes including steady-state substrate kinetics, metal preferences, and peptide substrate selectivity. We
also report that the three enzymes are significantly different in
1576
This paper is available on line at http://www.jbc.org
Comparison of Type 1 Receptor Tyrosine Kinases
their sensitivity to kinase inhibitors in the 4-anilinoquinoline
class. The structure-activity relationship of a subset of these
compounds with respect to type 1 receptor potency and selectivity is presented.
EXPERIMENTAL PROCEDURES
Cloning—Recombinant baculoviruses expressing the intracellular
domains for each family member were made from full-length human
clones by PCR. During PCR, oligonucleotides were used that introduced
restriction sites for subcloning as follows: for EGFR (GenBankTM accession number X00588), amino acids 671–1210 with StuI, XbaI ends; for
ErbB-2 (GenBankTM accession number X03363), amino acids 691–1255
with XbaI, HindIII ends; and for ErbB-4 (GenBankTM accession number
L07868), amino acids 690 –1309 with NotI, XbaI ends. All constructs
were fused to an N-terminal His6 tag (MKKGHHHHHHG) and subcloned into pFastBac (Invitrogen). Each construct was verified by DNA
sequence analysis. The recombinant baculoviruses were made following
Bac-To-BacTM baculovirus expression system manual (Invitrogen),
which utilizes the method developed at Monsanto/Searle (9).
Fermentation—Large scale (2-liter) virus preparations for fermentation were made by infecting Spodoptera frugiperda (Sf-9) cells in 6-liter
shake flasks at 27.5 °C and 120 rpm at a multiplicity of infection ⫽ 0.1.
The virus-containing culture supernatants were harvested at 72 h
post-infection via centrifugation at 2500 rpm for 20 min. Viral titers
were determined via enzyme-linked immunosorbent assay. A 36-liter
stirred bioreactor (University Research Glassware vessel outfitted with
overhead stirrer, internal dip tubes, and heat-transfer coil) was inoculated with Trichoplusia ni (T. ni) cells (kindly obtained from JRH
Biosciences, Woodland, CA) at ⬃0.5 ⫻ 106 cells per ml. The culture was
grown in Excell 405TM insect cell medium (JRH Biosciences) containing
50 ␮g/ml gentamicin (Invitrogen). The fermentor was maintained at
27.5 °C, 30 rpm using a paddle-type impeller and 50% dissolved oxygen
via sparging. Cells were allowed to double overnight, and the culture
was infected at a density of ⬃1 ⫻ 106 cells per ml at a multiplicity of
infection ⫽ 1. The culture was monitored daily for pH, glucose, lactate,
glutamine, cell count, and viability. Infection was allowed to proceed at
the above parameters. Cells were harvested 48 h post-infection with a
Centritech威 100 continuous flow centrifuge (DuPont). Concentrated
cells were centrifuged at 2000 rpm for 20 min and washed with protease
inhibitor buffer (1⫻ Dulbecco’s phosphate-buffered saline (Invitrogen),
1 mM EDTA (Sigma), 1 mM p-aminobenzamidine (Sigma), 1 ␮g/ml
aprotinin (Roche Molecular Biochemicals), 1 ␮g/ml leupeptin (Roche
Molecular Biochemicals)). Cells were spun again at 2000 rpm for 20 min
and quick frozen in dry ice/ethanol after discarding the supernatant.
Purification—The buffers used in this purification include the chelating Sepharose buffers as follows: buffer A, 25 mM HEPES, pH 7.5,
750 mM NaCl, 10% glycerol; buffer B, buffer A ⫹ 0.5 M imidazole. The
following homogenization buffer was used: buffer A ⫹ 25 mM imidazole
(5% buffer B). The following hydroxyapatite buffers (HA) were used:
buffer C, 20 mM HEPES, 20 mM NaH2PO4, pH 6.8, 10% glycerol; buffer
D, buffer C ⫹ 200 mM NaH2PO4, pH 6.8. The following hydrophobic
interaction chromatography buffers (HIC) were used: buffer E, 20 mM
Tris-HCl, pH 7.5, 0.6 M (NH4)2SO4; buffer F, 20 mM Tris-HCl, pH 7.5,
10% glycerol.
Frozen cell paste was thawed by resuspending the cells in homogenization buffer supplemented with a protease inhibitor mixture (Sigma), 1 mM MgCl2, and 5 ␮g/ml of DNase I and RNase. Cells were lysed
using a Polytron homogenizer (Brinkmann Instruments), and the resulting lysate was spun at 30,000 ⫻ g for 1 h to remove insoluble
proteins and cell debris. The supernatant was filtered through 0.45-␮m
filters and applied to a chelating Sepharose (Amersham Biosciences)
column previously equilibrated in homogenization buffer. After loading,
the column was rinsed with homogenization buffer until the A280
reached base line. The column was then washed with 10% buffer B to
remove any nonspecifically bound proteins and eluted with a 20-column
volume linear gradient to 100% buffer B. Fractions were tested for
peptide substrate phosphorylation as described below. Active fractions
were pooled and diluted 8-fold in buffer C and applied to a Ceramic HA
(Bio-Rad) column previously equilibrated in buffer C. After washing,
the column was eluted using a linear gradient to 100% buffer D. Active
fractions were pooled and brought to 0.6 M (NH4)2SO4 by addition of a
2.5 M (NH4)2SO4 stock solution, and the sample was applied to an HIC
column previously equilibrated in buffer E. A reverse linear gradient to
100% buffer F was used to elute protein. Active fractions were pooled
and stored at ⫺80 °C.
Type 1 Receptor Tyrosine Kinase Assays—Reactions were performed
in 96-well polystyrene round-bottom plates in a final volume of 45 ␮M.
1577
Reaction mixtures contained 50 mM MOPS (pH 7.5), 2 mM MnCl2, 10 ␮M
ATP, 1.0 ␮Ci of [␥-33P]ATP per reaction, 50 ␮M peptide substrate, and
1 mM dithiothreitol. We have used three peptide substrates. Peptide A
(biotin-(aminohexanoic acid)-EEEEYFELVAKKK-CONH2) was optimized for EGFR phosphorylation using a synthetic library approach
(10). Peptide B (biotin-(aminohexanoic acid)-GGMEDIYFEFMGGKKKCONH2) was optimized for ErbB-2 phosphorylation in a similar fashion
(11). Peptide C (biotin-(aminohexanoic acid)-RAHEEIYHFFFAKKKCONH2) was optimized for phosphorylation by ErbB-2 (12). All peptides
were synthesized by Quality Controlled Biochemicals Inc., Hopkinton,
MA. The concentrations of dissolved peptide stocks were determined by
amino acid analysis. Enzymatic reactions were initiated by adding 1
pmol (20 nM) per reaction of the indicated type 1 receptor intracellular
domain unless otherwise indicated. Reactions were terminated after 10
min at 23 °C by adding 45 ␮l of 0.5% phosphoric acid in water. 75 ␮l of
the terminated reaction mix was then transferred to MAPH phosphocellulose filter plates (Millipore, Marlborough, MA). The plates were
filtered and washed three times with 200 ␮l of 0.5% phosphoric acid. 50
␮l of scintillation mixture (Optiphase by Wallac, Turku, Finland) was
added to each well, and the assay was quantified by counting in a
Packard Topcount (Packard Instrument Co.).
Kinase Selectivity Assays—The catalytic domains of vascular endothelial growth factor receptor 2 (VEGFR2), c-Fms, c-Src, and protein
kinase C␨ (PKC␨) were expressed and purified using methods similar to
those described for ErbB-2, ErbB-4, and EGFR. Kinase assays were
performed as described above with the following modifications. c-Fms
assays contained 10 nM enzyme, 50 mM MOPS, pH 7.5, 10 mM MgCl2, 20
␮M ATP, and 200 ␮M peptide A. The reaction was allowed to proceed for
30 min, terminated, and quantified as described above. VEGFR2 assays
contained 10 nM enzyme, 100 mM HEPES, pH 7.5, 0.1 mg/ml bovine
serum albumin, 0.1 mM dithiothreitol, 360 nM peptide A, 75 ␮M ATP,
and 5 mM MgCl2. The reaction was allowed to proceed for 40 min.
Product was detected using a homogeneous time-resolved fluorescence
procedure (13). Briefly, the reactions were quenched by adding 100 ␮l of
100 mM HEPES, pH 7.5, 100 mM EDTA, 45 nM streptavidin-linked
allophycocyanin (Molecular Probes, Eugene, OR), and 3 nM europiumconjugated anti-phosphotyrosine antibody (Wallac, Turku, Finland).
The product was detected using a Victor plate reader (Wallac, Turku,
Finland) with a time delay at 665 nm. c-Src assays contained 0.4 nM
enzyme, 100 mM HEPES, and 100 nM peptide substrate (biotin-(6aminocaproic acid)-AAAQQIYGQI-NH2). The reaction was allowed to
proceed for 40 min, and the product was detected using the homogeneous time-resolved fluorescence procedure. PKC␨ assays contained 5 nM
enzyme, 50 mM MOPS, pH 7.3, 10 mM MgCl2, 1 ␮M ATP, 0.125 ␮Ci/well
[␥-33P]ATP, 5 ␮M peptide (biotin-(aminohexanoic acid)-FKLKRKGSFKKFA-CONH2). The reaction was allowed to proceed for 40 min,
terminated, and quantified using a scintillation proximity assay procedure as described (14).
Cellular Inhibition Assays—HN5 cells were cultured in low glucose
Dulbecco’s modified Eagle’s medium containing 10% (v/v) fetal bovine
serum. BT474 cells were cultured in RPMI containing 10% (v/v) fetal
bovine serum. Test compounds were prepared as 5 mM stocks in Me2SO,
and stocks were diluted to 1000⫻ test concentrations in Me2SO. Cells
were dosed with compounds diluted in the appropriate growth medium.
After incubating with compounds for 6 h, cells were rinsed with ice-cold
phosphate-buffered saline and were lysed in RIPA (150 mM NaCl, 50
mM Tris-HCl, pH 7.5, 0.25% deoxycholate, 0.1% Nonidet P-40) containing protease inhibitor mixture (Roche Molecular Biochemicals) and 1
mM sodium orthovanadate. Cells were scraped and transferred to microcentrifuge tubes for sonication and microcentrifugation. EGFR was
immunoprecipitated from 0.25 mg of HN5 lysate using 1 ␮g of antiEGFR Ab-13 (Lab Vision). ErbB-2 was immunoprecipitated from 0.25
mg of BT474 lysate using 0.5 ␮g of anti-c-Neu Ab-3 (Oncogene Research
Products). Immune complexes were precipitated using 50 ␮l of protein
G Plus/protein A-agarose suspension (Calbiochem) and were resuspended in Laemmli sample buffer. Equal volumes of each sample were
loaded onto duplicate NOVEX gels (Invitrogen) for electrophoresis and
Western blot analysis. Phosphorylated receptors were detected using
anti-Tyr(P) clone PT66 (Sigma) diluted 1:5000 in TBST (150 mM NaCl,
10 mM Tris-HCl, pH 7.5, 0.1% (v/v) Tween 20) containing 4% (w/v)
bovine serum albumin (Blocking Buffer). Total EGFR was detected
using anti-EGFR Ab-12 (Lab Vision) diluted 1:5000 in Blocking Buffer,
and total ErbB-2 was detected using anti-c-Neu Ab-3 (Oncogene Research Products) diluted 1:5000. SuperSignal West Femto Maximum
Sensitivity Substrate (Pierce) was used for detection. Films were
scanned on a Bio-Rad Fluor-S Multi-Imager, and bands were quantified
by densitometric analysis.
Data Analysis—Initial rates were determined by measuring the
1578
Comparison of Type 1 Receptor Tyrosine Kinases
FIG. 1. Purification of type 1 receptor intracellular domains. GelCode BlueTM-stained (Pierce) NuPAGE威 10% BisTris gels (Invitrogen)
were run with MES buffer. A, 2 ␮g of ErbB-2. B, 1 ␮g of EGFR. C, 1.5 ␮g of ErbB-4.
amount of phosphorylated peptide formed in 10 min, since this time
period produced adequate amounts of product for all three enzymes and
was within the linear portion of the reaction progress curve. Three
experiments were performed for all kinetic studies, and the average
data were globally fit to the equations described under results. Nonlinear regressions were conducted using the program Sigma Plot威 (Jandel
Scientific, San Rafael, CA).
RESULTS
Enzyme Expression and Purification—The intracellular domains of EGFR, ErbB-2, and ErbB-4 were cloned into a baculovirus expression system (see “Experimental Procedures”).
The first purification step for each enzyme was nickel chelate
affinity chromatography. Further purification was performed
using ceramic hydroxyapatite chromatography (HA) followed
by HIC (see under “Experimental Procedures”). The protein
eluting from HIC was analyzed for purity by SDS-PAGE (Fig.
1). Each protein migrated as predicted from its molecular
weight. We estimate the purity of these samples to be (ⱖ95%).
The identity of each enzyme was confirmed by N-terminal
sequencing (data not shown).
For each enzyme, analysis of the activity following chelating
Sepharose revealed it was devoid of any contaminating kinase
activities and had kinetic characteristics identical to the pure
enzyme (data not shown). Because additional purification steps
resulted in a significant loss in yield, we used enzyme purified
through the chelating Sepharose stage for the remainder of our
studies. Typical protein yields from fermentation after chelating Sepharose were 4 mg/liter for ErbB-2, 6 mg/liter for ErbB-4,
and 10 mg/liter for EGFR.
Time Course of Substrate Phosphorylation—The kinase activity of the purified type 1 receptor intracellular domains was
evaluated by measuring the phosphorylation of an exogenous
peptide substrate (see “Experimental Procedures”). The production of product as a function of time is shown in Fig. 2. In
this experiment the amount of product formed is less than 10%
of the total substrate available for all three enzymes. Therefore, the observed decrease in the rate of the reaction as a
function of time is not due to substrate consumption. The
experiment was analyzed by fitting the data to a model of
single exponential enzyme decay (Equation 1).
P ⫽ v共e⫺kt兲 䡠 t
(Eq. 1)
In Equation 1, v represents the initial rate of product (P)
formation that decays at a rate (k) over a period of time (t).
FIG. 2. Time course of peptide substrate phosphorylation. Peptide substrate phosphorylation reactions were performed using peptide
C and terminated at the indicated time using ErbB-4 (ƒ), ErbB-2 (䡺),
and EGFR (E). Data were fit to Equation 1.
EGFR was linear over the course of the experiment (v ⫽ 0.5
pmol/min, k ⫽ 0). ErbB-2 and ErbB-4 had higher initial rates
but demonstrated significant decay as follows: ErbB-2, v ⫽ 0.85
pmol/min, k ⫽ 0.014/min; ErbB-4, v ⫽ 1.25 pmol/min, k ⫽
0.007/min. Approximately 10% of the total enzyme has lost
activity in 10 min at the calculated rates of ErbB-2 and ErbB-4
decay. All subsequent experiments, therefore, were terminated
at 10 min to obtain a good estimate of the initial rate.
We have found that purified ErbB-2 is highly stable when
maintained in a concentrated solution, even at room temperature. However, the enzyme rapidly loses activity upon dilution.
These observations suggest that enzyme inactivation after dilution results from alterations in the tertiary structure of the
enzyme or alterations in its aggregation state. Our observations are consistent with previous results demonstrating that
aggregation of the EGFR catalytic domain significantly affects
the rate of catalysis (15, 16).
Divalent Cation Preferences—The interaction between protein kinases and divalent cations is complex. All phosphotransferases require a divalent cation that coordinates the phos-
Comparison of Type 1 Receptor Tyrosine Kinases
1579
FIG. 3. Divalent cation dependence. Peptide phosphorylation reactions were performed using peptide C and the indicated type 1 receptor catalytic domain. For each panel, ErbB-4 (ƒ), ErbB-2 (䡺), and EGFR
(E) were compared. A, MgCl2 concentration was varied. B, MnCl2 concentration was varied.
phate groups of the nucleotide triphosphate substrate (17).
Kinases have a similar requirement, but they also have an
additional site of interaction that can be distinguished from the
primary site by affinity. Kinases can be activated or inactivated
by binding of a cation to this second site depending upon the
specific kinase and identity of the cation (18, 19). The effect of
divalent cation concentration on the rate of peptide substrate
phosphorylation catalyzed by the type 1 receptor kinases is
shown in Fig. 3. The type 1 receptor kinases are different with
respect to their ability to utilize Mg2⫹. ErbB-2 is almost inactive in the presence of Mg2⫹, whereas EGFR and ErbB-4 can
use this cation efficiently (Fig. 3A). Another key difference is
the optimum concentration of Mn2⫹ used by each enzyme (Fig.
3B). In this experiment ATP was held constant at 10 ␮M. Under
these conditions any concentration of cation above 10 ␮M is
essentially free from bound nucleotide. Peak activity for
ErbB-2 was observed at 2 mM Mn2⫹. ErbB-4, on the other hand,
exhibits optimal activity at 0.2 mM Mn2⫹, and EGFR exhibits
optimal activity at 0.6 mM Mn2⫹. Presumably these differences
reflect different affinities of Mn2⫹ for binding to the second
cation site.
To evaluate further the interaction between the type 1 receptor kinases, ATP, and divalent cations, we determined the
FIG. 4. ATP-metal substrate kinetics. Peptide substrate phosphorylation reactions were performed using peptide C and the indicated
enzyme (50 ␮M). Plots of reciprocal velocity versus reciprocal concentration of ATP-metal are shown using ATP-Mg (E) or ATP-Mn (●). Ten
concentrations of ATP-metal were evaluated ranging from 4 to 210 ␮M.
Data were fit to Equation 2, and the kinetic parameters from this fit are
summarized in Table I. Additional free MgCl2 or MnCl2 was added to
the reaction mix at the optimal concentration determined from the
experiment shown in Fig. 3 and listed in Table I. A, ErbB-2, B, EGFR,
C, ErbB-4.
apparent Km and Vmax values for ATP-Mg and ATP-Mn at a
fixed concentration of peptide C (Fig. 4). These experiments
were performed at a concentration of free Mg2⫹ or Mn2⫹ equivalent to the optimum concentration determined in the experiment shown in Fig. 3. ATP-Mg and ATP-Mn titrations were
performed in parallel, and the reactions were initiated with the
same diluted enzyme mixture to ensure that the results are
directly comparable. The substrate kinetic parameters were
determined by nonlinear least squares analysis of the averaged
1580
Comparison of Type 1 Receptor Tyrosine Kinases
TABLE I
Kinetic parameters for ATP-metal
Peptide substrate phosphorylation reactions were performed using the indicated enzyme and peptide C (50 ␮M). MgCl2 and MnCl2 were added
at the optimum concentration as indicated. Three experiments were performed for each enzyme substrate combination using the concentration of
ATP-metal described for Fig. 3. The steady-state kinetic parameters were determined by fitting the averaged initial velocity data to Equation 2 (see
“Results”). The standard errors of the parameter estimates were all less than 10% of the value shown.
ATP-Mg
ATP-Mn
Enzyme
ErbB-2
EGFR
ErbB-4
MgCl2
kcat
Km(app)
mM
s⫺1
␮M
10
10
10
0.008
0.020
0.045
27
17
37
kcat/Km
M
⫺1
s⫺1
300
1180
1220
MnCl2
kcat
Km(app)
mM
s⫺1
␮M
1.60
0.62
0.20
0.060
0.022
0.093
13
5
21
kcat/Km
M
⫺1
s⫺1
4600
4400
4430
initial velocity data fitting to the Henri-Michaelis-Menten
equation (Equation 2).
v ⫽ VA/共Km共app兲 ⫹ A兲
(Eq. 2)
In this equation, v is the measured initial velocity; V is the
maximum velocity; A is the concentration of ATP-metal; and
Km(app) is the apparent Michaelis constant for ATP-metal. Similar experiments were performed for EGFR and ErbB-4, and
the results are summarized in Table I. The kcat values were
calculated by dividing Vmax by the total enzyme concentration.
It is likely that this procedure significantly underestimates kcat
since the fraction of active enzyme is unknown (11). For all
three enzymes, Km(app) was slightly lower for ATP-Mn. For
EGFR the kcat value was similar for either form of the ATP
substrate. However, for ErbB-2 the kcat value was 7.5-fold
higher for ATP-Mn. The relative efficiency of various substrates can be determined by comparing the ratio of the turnover number (kcat) and Km(app). For ErbB-2 the kcat/Km(app)
value for ATP-Mn (4615 M⫺1 s⫺1) is 16 times greater than that
for ATP-Mg (296 M⫺1 s⫺1). EGFR and ErbB-4 have a modest
preference for ATP-Mn.
Steady-state Substrate Kinetics—We evaluated the steadystate kinetic parameters of three different peptide substrates
for each of the type 1 receptor kinases. Peptide phosphorylation
was measured at 8 fixed concentrations of peptide and 12
varied concentrations of ATP-Mn. Autophosphorylation contributes to the overall velocity to a varying extent depending
upon the enzyme-peptide substrate pair under evaluation. The
contribution of autophosphorylation was eliminated from our
analysis by subtracting the amount of product formed in the
absence of peptide substrate from the amount formed in the
presence of peptide substrate at each concentration of ATP.
The results of ErbB-2 phosphorylation of peptide A are shown
in Fig. 5A. Reciprocal velocity as a function of reciprocal
ATP-Mn concentration was plotted for each fixed peptide concentration. The data were globally fit to the equation for a
sequential two-substrate reaction (Equation 3) (20).
v ⫽ VAB/KiaKmb ⫹ KmbA ⫹ KmaB ⫹ AB
(Eq. 3)
In this equation, v is the measured initial velocity; V is the
maximum velocity; A is the concentration of ATP-Mn; B is the
concentration of peptide; Kma and Kmb are the Michaelis constants for ATP-Mn and peptide, respectively; and Kia is the
dissociation constant for ATP-Mn.
The kinetic parameters for EGFR phosphorylation of peptide
A were determined from the data shown in Fig. 5B. These
experiments were performed in parallel with the ErbB-2 phosphorylation of peptide A, but the pattern of the fit of the data
reveals significant differences. In the plot of reciprocal velocity
as a function of reciprocal ATP-Mn concentration, the different
fixed peptide concentrations result in lines that intersect to the
left of the vertical axis and well below the horizontal axis. This
pattern is indicative of significant negative substrate synergy
FIG. 5. Two substrate steady-state kinetics. Peptide substrate
phosphorylation reactions were performed using peptide A and the
indicated enzyme. Twelve concentrations of ATP-Mn were used ranging
from 1 to 150 ␮M. Seven concentrations of peptide were used ranging
from 8 to 500 ␮M. The entire data set was globally fit to Equation 3, and
the kinetic parameters from this fit are summarized in Table II. Reciprocal velocity versus reciprocal concentration of ATP-Mn is plotted for
different concentrations of peptide A as follows: ●, 500 ␮M; E, 250 ␮M;
, 125 ␮M; ƒ, 62.5 ␮M; f, 31.25 ␮M. Only a subset of the total data set
was plotted for clarity. A, ErbB-2. B, EGFR.
(21). A similar pattern was obtained when the reciprocal velocity was plotted as a function of reciprocal peptide concentration
(data not shown).
The negative substrate synergy discerned from the Lineweaver-Burke analysis using peptide A is readily apparent
from the calculated kinetic parameters (Table II). The ratio of
Kia/Kma is an estimate of the degree of binding synergy between
substrates in a two-substrate kinase reaction (22). If binding of
one substrate has no effect on the Km of the other substrate,
Comparison of Type 1 Receptor Tyrosine Kinases
1581
TABLE II
ATP and peptide substrate steady-state kinetic parameters
Peptide substrate phosphorylation reactions were performed using the indicated enzyme and peptide substrate. Substrate concentrations were
varied as described for Fig. 4. Three experiments were performed for each enzyme substrate combination. The steady-state kinetic parameters were
determined by fitting the averaged initial velocity data to Equation 3 (see “Results”). The standard errors of the parameter estimates are shown.
kcat
s
ErbB-2
Peptide
Peptide
Peptide
EGFR
Peptide
Peptide
Peptide
ErbB-4
Peptide
Peptide
Peptide
Kma
Kmb
⫺1
Kia
␮M
kcat/Kmb
s
⫺1
M
Kia/Kma
⫺1
A
B
C
0.005
0.037
0.024
19 ⫾ 2
84 ⫾ 17
50 ⫾ 8
53 ⫾ 8
157 ⫾ 40
58 ⫾ 11
25 ⫾ 8
31 ⫾ 17
65 ⫾ 17
100
235
414
1.3
0.6
1.3
A
B
C
0.023
0.218
0.057
34 ⫾ 4
76 ⫾ 19
21 ⫾ 3
316 ⫾ 38
899 ⫾ 200
128 ⫾ 16
1.9 ⫾ 1.4
8⫾5
0.5 ⫾ 1.2
73
242
445
0.06
0.10
0.02
A
B
C
0.031
0.076
0.105
44 ⫾ 7
29 ⫾ 4
42 ⫾ 3
270 ⫾ 38
144 ⫾ 22
42 ⫾ 3
13 ⫾ 4
9⫾5
1.1 ⫾ 2.1
114
528
2500
0.30
0.31
0.03
then Kia/Kma is equal to 1. The EGFR Kia/Kma values range
from 0.020 (peptide B) to 0.10 (peptide C). By comparison, the
ErbB-2 Kia/Kma values range from 0.6 (peptide B) to 1.3 (peptides A and C). This lack of negative substrate synergy for
ErbB-2 is graphically demonstrated by lines that intersect to
the left of the vertical axis and near the horizontal axis (Fig.
5A).
Posner et al. (23) observed similar negative substrate synergy for EGFR, and Jan et al. (11) observed no substrate synergy for ErbB-2. These two studies are not directly comparable
with each other, however, because dissimilar enzyme constructs and different peptide substrates were used. In our
work, we have used comparable intracellular domain constructs and identical peptide substrates for EGFR and ErbB-2.
Our results confirm the substrate synergy observations made
by Posner et al. (23) and Jan et al. (11) and suggest that the
presence or absence of negative substrate synergy is an inherent property of the particular type 1 receptor intracellular
domain.
Peptide Substrate Selectivity—The catalytic efficiency of an
enzyme substrate combination is described by the ratio kcat/Km.
This ratio is the best measure for comparison of the efficiency of
product formation catalyzed by related enzymes. It also captures
the effects of substrate molecular structure on either kinetic
constant. Thus, the ratio is a good measure of different substrate
selectivity for the same enzyme (24). The kinetic parameters for
all three enzymes and all three peptides were determined, and
the kcat/Kmb ratios were calculated (Table II). Peptide A was
optimized for phosphorylation by EGFR (10). Peptide B was
optimized for phosphorylation by ErbB-2 (11). Peptide C was also
optimized for phosphorylation by ErbB-2 (12).
All three enzymes phosphorylated peptide A with similar
efficiency (kcat/Kmb ⬵100 M⫺1 s⫺1). Peptide B was a better
substrate for EGFR, ErbB-2, and ErbB-4 even though it was
optimized for phosphorylation by ErbB-2. Jan et al. (11) analyzed the kinetics of ErbB-2 phosphorylation of peptide B at a
single fixed concentration of ATP (50 ␮M) and found a kcat/
Kmb(app) of 309 M⫺1 s⫺1 . We varied each substrate at several
fixed concentrations of second substrate and obtained kinetic
parameters by globally fitting the entire data set. We determined kcat/Kmb to be 235 M⫺1 s⫺1 for the ErbB-2 peptide B
combination. Our result is similar to that reported by Jan et al.
(11) because ErbB-2 exhibits very little substrate binding synergy. The kcat/Kmb for peptide C was better than peptides A or
B for all three enzymes. This effect was primarily
Kmb-mediated.
The three peptide substrates are used with vastly different
efficiencies ranging from the EGFR peptide A combination
(kcat/Kmb ⫽ 73) to the ErbB-4 peptide C combination (kcat/Kmb
⫽ 2500). However, all three kinases select the three peptide
substrates with the same rank order of kcat/Kmb; peptide A ⬍
peptide B ⬍ peptide C. By this criterion, the different members
of the type 1 receptor family have very similar downstream
substrate preferences.
Inhibitor Potency and Selectivity—Members of the 4-anilinoquinazoline class of compounds have been shown to be potent
tyrosine kinase inhibitors (25). Many molecules in this class
have been shown to inhibit EGFR (26). These molecules bind in
the ATP-binding pocket and compete with this substrate.
EGFR, ErbB-2, and ErbB-4 have very similar catalytic domains, so members of this class of kinase inhibitors may affect
more than one member of the type 1 receptor family.
A representative set of compounds was selected to demonstrate key structure-activity relationships in the optimized assay system for the type 1 receptor kinases (Table III). The
design and synthesis of these compounds has been fully described elsewhere (27). The aniline modifications to a core
6,7-dimethoxyquinazoline were previously shown to affect dramatically the tyrosine kinase inhibition profile (26). Of the type
1 receptor family members, the inhibition of EGFR appeared to
be the least affected by the aniline changes. The range of EGFR
IC50 values was less than 5-fold, despite the broad size of the
aniline substitutions. The 4-(3-hydroxyanilino)-substituted
quinazolines (compounds 1 and 2) showed general tyrosine
kinase inhibition. When the hydroxyl of compound 1 was replaced with a 3-methoxyl group (compound 3), a significant loss
in inhibition of VEGFR2 was observed, but the compound retained a similar type 1 receptor inhibition profile. AG1478
(compound 4) demonstrated the reported profile with a marked
potency against EGFR (28). Bulky aniline substituents, represented by compounds 5– 8, confer improved potency against
ErbB-2 while retaining EGFR inhibition. The benzyloxyaniline
(compound 5) and halogenated benzyloxyaniline (compound 6)
substituents appeared to confer significant selectivity for dual
inhibition of EGFR and ErbB-2 compared with the other kinase
enzymes tested.
Intracellular Inhibition of EGFR and ErbB-2—The ability of
potent dual ErbB-2/EGFR inhibitors to block receptor autophosphorylation in whole cells was analyzed in two cell lines,
HN5 and BT474 (Fig. 6). HN5 cells overexpress EGFR (29), and
BT474 cells overexpress ErbB-2 (30). Logarithmically growing
cells were incubated with various concentrations of the indicated compound for 6 h. The cells were lysed and the indicated
receptor was captured by immunoprecipitation. The relative
receptor content was determined by immunoblot with a separate anti-receptor antibody, and the state of phosphorylation
1582
Comparison of Type 1 Receptor Tyrosine Kinases
TABLE III
Structure-activity relationship of quinazoline aniline substitutions
Peptide substrate phosphorylation reactions were performed using peptide A and the indicated enzyme. The indicated compound was serially
diluted 1 to 3 and added to the reactions to produce an 11-point dose-response curve ranging from 0.0001 to 10 ␮M. The IC50 values were estimated
from data fit to the equation y ⫽ Vmax 䡠 (1 ⫺ (x/(K ⫹ x)). cFms, VEGFR2, cSrc, and PKC␨ assays were performed as described under “Experimental
Procedures.” n.d., not determined.
was analyzed by immunoblot with an anti-phosphotyrosine
antibody. Treatment of the HN5 cell line with all three compounds did not affect the total EGFR content, but there was a
dose-dependent decrease in tyrosine-phosphorylated EGFR
(Fig. 6A). The approximate IC50 value for inhibition of EGFR
for these compounds ranges from 0.5 to 5 ␮M. Similar results
were obtained for inhibition of ErbB-2 autophosphorylation in
the BT474 cell line (Fig. 6B).
DISCUSSION
The discovery and clinical development of inhibitors of EGFR
catalytic activity is progressing at a rapid rate (7). These molecules are promising for the treatment of tumors that overexpress this receptor. The discovery of dual inhibitors of ErbB-2
and EGFR may be the next step in the progression of this
therapeutic approach (31, 32). The catalytic domains of the
active members of the type 1 receptor family are very similar.
Comparison of Type 1 Receptor Tyrosine Kinases
1583
FIG. 6. Intracellular inhibition of EGFR and ErbB-2. A, EGFR overexpressing HN5 cells were treated for 6 h with variable concentrations
of the indicated compound. The cells were lysed, and EGFR was isolated by immunoprecipitation. A, lower, total EGFR in the sample was
determined by Western blot using anti-EGFR antibody. A, upper, tyrosine-phosphorylated EGFR was determined by Western blotting using
anti-phosphotyrosine. B, ErbB-2 overexpressing BT474 cells were treated as described for A. Total ErbB-2 and tyrosine-phosphorylated ErbB-2
were detected by Western blot in a similar fashion as in A.
The amino acid sequences of the catalytic domains of EGFR,
ErbB-2, and ErbB-4 are 84 – 86% identical to one another depending upon the specific comparison. Because of this high
degree of similarity, it is possible that compounds designed to
inhibit one family member will have some effect on the other
receptors. The interpretation of the biological effects of type 1
receptor kinase inhibitors will be enhanced by a detailed understanding of the potency profile of the inhibitor versus all
three catalytically active enzymes. The potency of a kinase
inhibitor in a biological system depends upon several biochemical properties. These properties include inhibitor Kd, substrate
identity, enzyme substrate kinetic parameters, enzyme activation state, enzyme reaction mechanism, and inhibitor mode of
action.
The evaluation of the biochemical and kinetic properties of
members of the type 1 receptor family has been greatly facilitated by the expression and purification of defined domains
using baculovirus expression systems. EGFR has been most
extensively studied (15, 33–36). The catalytic domain is subject
to a regulatory mechanism that is driven by the formation of
higher order complexes (15, 36). The divalent metal ion requirement seems to reflect aspects of EGFR catalytic activity
associated with the activation state (37). Autophosphorylation
of EGFR does not seem to affect catalytic activity (36). Biochemical and kinetic properties of human ErbB-2 and the rat
homologue, c-Neu, have been characterized using purified
baculovirus-expressed proteins (11, 38 – 41). ErbB-2 seems to
be similar to EGFR in some respects. A higher order structure
such as dimerization was correlated with increased enzyme
activity (40, 41), and divalent metal requirement seems to
reflect the activation state (38). ErbB-3 has been expressed and
purified from baculovirus and shown to have impaired tyrosine
kinase activity (2). Purified ErbB-4 has not been characterized
in this manner, and little is known about the biochemical and
kinetic properties of this protein. Differences in the biochemical
properties of the type 1 receptor kinases that influence inhibitor potency and selectivity are hard to determine from the
studies described above due to differences in protein construct
and experimental design. We have used comparable constructs
and identical experimental design to accurately assess similarities and differences between these enzymes.
ATP-Metal Substrate Preferences—Protein kinases have diverse responses to divalent metal concentration. cAMP-
dependent protein kinase binds ATP in association with one or
two Mg2⫹ ions. The first ion binds with high affinity and
coordinates the ␤- and ␥-phosphate of ATP. The second Mg2⫹
ion binds with lower affinity and coordinates ␣- and ␥-phosphates of ATP (42). All protein kinases require binding by the
first metal ion for activity. The consequences of second metal
ion binding vary. The kcat of cAMP-dependent protein kinase is
reduced by second ion binding (43), but the Km value for ATP is
decreased. c-Src and C-terminal Src kinase have increased kcat
values in response to second site binding, but the Km value for
ATP is unchanged (19). ErbB-2, EGFR, and ErbB-4 all have
optimal Mg2⫹ concentrations far above the concentration
bound to ATP (Fig. 3A). This suggests that Mg2⫹ occupancy of
a second site is required for optimal activity. We cannot determine from these studies, however, if the effect is Km- or
kcat-mediated.
Several reports (11, 33, 34, 38) indicate that EGFR and
ErbB-2 intracellular domains exhibit a preference for ATP-Mn
relative to ATP-Mg. We observe a similar ATP-Mn preference,
but the optimal free concentration of Mn2⫹ is different for each
enzyme (Fig. 3B). For example, ErbB-4 appears to have a
10-fold higher affinity for Mn2⫹ compared with ErbB-2. This
experiment and other published cation preference experiments
were conducted at a fixed concentration of ATP, so the kinetic
nature of the preference cannot be established. We conducted
ATP-metal substrate kinetic studies in the presence of a fixed
concentration of free divalent cation equal to the optimal concentration (Fig. 4). Presumably, the optimal concentration represents equal binding site occupancy for each metal-enzyme
pair. This way the true effect of metal identity can be determined rather than a concentration-dependent phenomenon associated with differential metal site occupancy. The kcat/Km(app)
is a good overall measure of substrate preference. ErbB-4 and
EGFR have very similar kcat/Km(app) values for ATP-Mg and
4-fold higher kcat/Km(app) values for ATP-Mn. The kinetic nature of the preference for ATP-Mn for the enzymes is different,
however. EGFR has a similar kcat value for either cation, but
ATP-Mn has a significantly lower Km(app) value. The increased
ATP-Mn kcat/Km(app) value for ErbB-4 results from an effect on
both kinetic parameters. ErbB-2 demonstrates the greatest
preference for ATP-Mn. The kcat/Km(app) value is 16-fold higher
than for ATP-Mg. This effect is almost entirely kcat-mediated.
It is remarkable that two enzymes as closely related as
1584
Comparison of Type 1 Receptor Tyrosine Kinases
EGFR and ErbB-2 would exhibit such dramatic differences in
ATP-metal substrate kinetic parameters. The biggest kinetic
contributor to this difference is the very low kcat value for
ErbB-2 and ATP-Mg. ATP-Mg is believed to be the natural
biological substrate for protein kinases. The inability of ErbB-2
to utilize this substrate suggests that ATP-Mn may mimic a
physiologically relevant activation mechanism. Higher order
enzyme structure may be one such mechanism that is affected
by ATP-Mn (15).
Steady-state Kinetics—We determined the steady-state kinetic parameters of the type 1 receptor kinases by measuring
initial velocities at varied concentrations of both substrates.
The plot of reciprocal velocity versus reciprocal ATP-Mn concentration for ErbB-2 at different fixed concentrations of peptide (Fig. 4A) reveals lines that intersect to the left of the
vertical axis near the horizontal axis. This relationship is consistent with a sequential Bi-Bi steady state ordered or a sequential Bi-Bi rapid equilibrium random mechanism. Our results do not distinguish between these two possibilities, but a
sequential Bi-Bi rapid equilibrium random mechanism was
demonstrated previously for EGFR (23). An analysis of burstphase kinetics for ErbB-2 demonstrated that chemical catalysis
occurs 3-fold faster than the rate-limiting step of the reaction.
The rate-limiting step appeared to be ADP release (11).
A striking feature of our analysis of EGFR is the presence of
a strong negative effect of one substrate binding on the affinity
of the other. A similar effect was observed in an analysis of
EGFR holoenzyme (23). EGF binding to the holoenzyme significantly reduced the negative substrate binding synergy. Posner
et al. (23) hypothesized that ligand binding causes a conformational change that increases the stability of the tertiary enzyme-ATP-peptide complex. Our work was conducted using
just the intracellular domain of EGFR. The fact that we observed such significant negative substrate binding synergy
suggests that the intracellular domain adopts a confirmation
similar to the nonactivated holoenzyme.
Peptide Selectivity—Receptor protein tyrosine kinases have
unique peptide substrate specificity, and the specificity of the
kinase has been proposed to influence downstream signal
transduction events (10). Substrate specificity is determined by
the amino acid sequence within 5 residues on either side of the
phosphorylated tyrosine and corresponding side chains in the
catalytic domain of the kinase. Single amino acid substitutions
within the peptide-binding region of the catalytic domain have
been found to effect significantly the peptide substrate amino
acid sequence preferences (44, 45). The type 1 receptor kinases
have a number of amino acid sequence differences in the Cterminal lobe that could affect peptide substrate selection.2
We evaluated the peptide substrate steady-state kinetic parameters for each of the type 1 receptor kinases using three
different peptide substrates. One peptide was optimized for
phosphorylation by EGFR (10), and two peptides were optimized for phosphorylation by ErbB-2 (11, 12). It is difficult to
compare peptide substrate selectivity for two different enzymes
by comparing the Km or kcat values for a single substrate
because the intrinsic catalytic capability of the enzymes may
differ. Comparison of the kcat/Kmb for the three peptides reveals
that each enzyme selects the three different peptides with the
same rank order: peptide A ⬍ peptide B ⬍ peptide C. The three
different type 1 receptor tyrosine kinases, therefore, have very
similar peptide-substrate preferences for these three optimized
substrates. Other amino acid sequence substitutions around
the phosphorylation site or additional protein-protein interac-
2
D. Vanderwall, personal communication.
tions may still allow differential downstream signaling by
these enzymes.
Sensitivity to Kinase Inhibitors and Biological Activity—We
have used the purified type 1 receptor intracellular domains,
substrates, and kinetic parameters determined above to design
an assay system that will allow us to assess accurately the
structure-activity relationship of inhibitor potency and selectivity. A variety of inhibition profiles can be obtained within
the 4-anilinoquinazoline template. Potent relatively nonselective inhibitors, selective EGFR inhibitors, and selective dual
ErbB-2/EGFR inhibitors were characterized. EGFR seems to
be the most promiscuous member of the family because relatively small or bulky aniline substitutions retain potent activity. Dual ErbB-2/EGFR inhibition is conferred by bulky aniline
substitutions, but these molecules are not as potent versus
ErbB-4.
We have used purified intracellular domains to create our
type 1 receptor enzyme assay system for practical reasons such
as expression level and ease of purification. The biologically
relevant forms of the receptors also contain the transmembrane and extracellular domains, and the conformation of these
forms might be influenced by ligand binding. We evaluated the
effect of kinase inhibitors that we characterized using our
assay system in whole cells to determine whether the potency
and selectivity that we obtain in the enzyme studies translates
to a true biological effect. Inhibitors in this class are competitive with ATP, and our enzyme studies are conducted using
relatively low ATP concentrations (10 ␮M) to approximate the
compound Ki. Therefore, the IC50 of the inhibitor in whole cells
will differ from the enzyme IC50 due to the higher intracellular
concentration of ATP and other unknown factors such as membrane permeability. Three potent dual ErbB-2/EGFR inhibitors
were characterized. All three were very effective at specifically
reducing both ErbB-2 and EGFR autophosphorylation without
affecting receptor levels (Fig. 6). The enzyme assay system
therefore faithfully reflects dual activity in a biological system
for ErbB-2 and EGFR inhibitors.
We have characterized the biological activity of a subset of
these molecules more completely (31, 32). The inhibition of
receptor autophosphorylation that we observe in tissue culture
translates into inhibition of tumor cell proliferation and growth
in mouse models. Future compound design and inhibitor analysis may lead to the discovery of different compounds with a
variety of type 1 receptor inhibition profiles. The biological
evaluation of such molecules should lead to a better understanding of the role of these enzymes in normal and tumor cell
growth, and ultimately to improved therapies for patients with
cancers that are linked to the type 1 receptor family.
Acknowledgments—We thank Warren Rocque and Kurt Weaver for
technical suggestions; Janet Warner for help in obtaining the baculoviruses; Bill Holmes and Byron Ellis for purifying VEGFR2, c-Fms,
c-Src, and PKC␨; and Wendy Liu, Brad McDonald, Stephanie
Schweiker, and Anne Truesdale for VEGFR2, c-Fms, c-Src, and PKC␨
selectivity screening. We also thank Gaochao Tian and Anne Truesdale
for critically reading this manuscript and Brent D. Abbey for converting
the graphics into EPS files.
REFERENCES
1. Olayioye, M. A., Neve, R. M., Lane, H. A., and Hynes, N. E. (2000) EMBO J. 19,
3159 –3167
2. Guy, P. M., Platko, J. V., Cantley, L. C., Cerione, R. A., and Carraway, K. L.,
III (1994) Proc. Natl. Acad. Sci. U. S. A. 91, 8132– 8136
3. Slamon, D. S., Drandt, R., Ciardiello, F., and Normanno, N. (1995) Crit. Rev.
Oncol. Hematol. 19, 183–232
4. Alroy, I., and Yarden, Y. (1997) FEBS Lett. 410, 83– 86
5. Cohen, B. D., Kiener, P. A., Green, J. M., Foy, L., Fell, H. P., and Zhang, K.
(1996) J. Biol. Chem. 271, 30897–30903
6. Pegram, M. D., Lipton, A., Hayes, D. F., Weber, B. L., Baselga, J. M., Tripathy,
D., Baly, D., Baughman, S. A., Twaddell, T., Glaspy, J. A., and Slamon, D. J.
(1998) J. Clin. Oncol. 16, 2659 –2671
7. Woodburn, J. R. (1999) Pharmacol. Ther. 82, 241–250
8. Lawrence, D. S., and Niu, J. (1998) Pharmacol. Ther. 77, 81–114
Comparison of Type 1 Receptor Tyrosine Kinases
9. Luckow, V. A., Lee, S. C., Barry, G. F., and Olins, P. O. (1993) J. Virol. 67,
4566 – 4579
10. Songyang, Z., Carraway, K. L., Eck, M. J., Harrison, S. C., Feldman, R. A.,
Mohammadi, M., Schlessinger, J., Hubbard, S. R., Smith, D. P., Eng, C.,
Lorenzo, M. J., Ponder, B. A. J., Mayer, B. J., Cantley, L. C. (1995) Nature
373, 536 –539
11. Jan, A. Y., Johnson, E. F., Diamonti, J., Carraway, K. L., and Anderson, K. S.
(2000) Biochemistry 39, 9786 –9803
12. Stuart, J. D., Blackburn, K., Moyer, M., and Bramson, H. N. (May 30, 2001)
U. S. Provisional Patent Application No. 60/294365, filed May 30, 2001
13. Park, Y.-W., Cummings, R. T., Wu, L., Zheng, S., Cameron, P. M., Woods, A.,
Zaller, D. M., Marcy, A. I., and Hermes, J. D. (1999) Anal. Biochem. 269,
94 –104
14. McDonald, O. B., Antonsson, B., Arkinstal, S., Marshall, C. J., and Wood, E. R.
(1999) Anal. Biochem. 268, 318 –329
15. Mohammadi, M., Honegger, A., Sorokin, A., Ullrich, A., Schlessinger, J., and
Hurwitz, D. R. (1993) Biochemistry 32, 8742– 8748
16. Gregoriou, M., Jones, P. F., Timmse, J. F., Yang, J. J., Radford, S. E., and Rees,
A. R. (1995) Biochem. J. 306, 667– 678
17. Knowles, J. R. (1980) Annu. Rev. Biochem. 54, 897–930
18. Armstrong, R. N., Kondo, H., Granot, J., Kaiser, E. T., and Mildvan, A. S.
(1979) Biochemistry 18, 1230 –1238
19. Sun, G., and Budde, R. J. A. (1997) Biochemistry 36, 2139 –2146
20. Cleland, W. W. (1979) Methods Enzymol. 63, 103–138
21. Segal, I. H. (1975) Enzyme Kinetics, pp. 274 –283, John Wiley & Sons, Inc.,
New York
22. Knight, W. B., and Cleland, W. W. (1989) Biochemistry 28, 5728 –5734
23. Posner, I., Engel, M., and Levitzki, A. (1992) J. Biol. Chem. 267, 20638 –20647
24. Copeland, R. A. (2000) Enzymes. A Practical Introduction to Structure, Mechanism, and Data Analysis, pp. 122–123, Wiley-VCH, New York
25. Traxler, P. (1998) Expert Opin. Ther. Patents 8, 1599 –1625
26. Denny, W. A., Rewcastle, G. W., Bridges, A. J., Fry, D. W., and Kraker, A. J.
(1996) Clin. Exp. Pharmacol. Physiol. 23, 424 – 427
27. Bridges, A. J., Zhou, H., Cody, D. R., Rewcastle, G. W., McMichael, A.,
Showalter, H. D., Fry, D. W., Kraker, A. J., and Denny, W. A. (1996) J. Med.
Chem. 39, 267–276
1585
28. Partik, G., Hochegger, K., Schorkhuber, M., and Marian, B. (1999) J. Cancer
Res. Clin. Oncol. 125, 379 –388
29. Modjtahedi, H., Affleck, K., Stubberfield, C., and Dean, C. (1998) Int. J. Oncol.
13, 335–342
30. Pasleau, F., Grooteclaes, M., and Gol-Winkler, R. (1993) Oncogene 8, 849 – 854
31. Rusnak, D. W., Affleck, K., Cockerill, S. G., Stubberfield, C., Harris, R., Page,
M., Smith, K. J., Guntrip, S. B., Carter, M. C., Shaw, R. J., Jowett, A.,
Stables, J., Topley, P., Wood, E. R., Brignola, P. S., Reep, B. R., Kadwell,
S. H., Mullin, R. J., Alligood, K. J., Keith, B. R., Crosby, R. M., Murray,
D. M., Knight, W. B., Gilmer, T. M., and Lackey, K. (2001) Cancer Res. 61,
7196 –7203
32. Cockerill, S., Stubberfield, C., Stables, J., Carter, M., Guntrip, S., Smith, K.,
McKeown, S., Shaw, R., Topley, P., Thomsen, L., Affleck, K., Jowett, A.,
Hayes, D., Willson, M., Woollard, P., and Spalding, D. (2001) Bioorg. Med.
Chem. Lett. 11, 1401–1405
33. Wedegaertner, P. B., and Gill, G. N. (1989) J. Biol. Chem. 264, 11346 –11353
34. Hsu, C. J., Mohammadi, M., Nathan, M., Honnegger, A., Ullrich, A.,
Schlessinger, J., and Hurwitz, D. R. (1990) Cell Growth Differ. 1, 191–200
35. Hsu, C. J., Hurwitz, D. R., Mervic, M., and Zilberstein, A. (1991) J. Biol. Chem.
266, 603– 608
36. Cadena, D. L., Chan, C., and Gill, G. N. (1994) J. Biol. Chem. 269, 260 –265
37. Koland, J. G., and Cerione, R. A. (1990) Biochim. Biophys. Acta 1052, 489 – 498
38. Guy, P. M., Carraway, K. L., III, and Cerione, R. A. (1992) J. Biol. Chem. 267,
13851–13856
39. Myers, J. N., Levea, C. M., Smith, J. E., Kallen, R. G., Tung, L., and Greene,
M. I. (1992) Receptor 2, 1–16
40. LeVea, C. M., Myers, J. N., Dougall, W. C., Wian, X., and Greene, M. I. (1993)
Receptor 3, 293–309
41. Samanta, A., LeVea, C. M., Dougall, W. C., Qian, X., Greene, M. I. (1994) Proc.
Natl. Acad. Sci. U. S. A. 91, 1711–1715
42. Zheng, J., Knighton, D. R., Ten Eyck, L. F., Karlsson, R., Xuong, N., Taylor,
S. S., and Sowadski, J. M. (1993) Biochemistry 32, 2154 –2161
43. Cook, P. F., Neville, M. E. J., Vrana, K. E., Hartl, F. T., and Roskoski, R. (1982)
Biochemistry 21, 5794 –5799
44. Till, J. H., Chan, P. M., and Miller, W. T. (1999) J. Biol. Chem. 274, 4995–5003
45. Konkol, L., Hirai, J. T., and Adams, J. A. (2000) Biochemistry 39, 255–262