Tyrosine kinase inhibitors. 11. Soluble analogues of pyrrolo- and pyrazoloquinazolines as epidermal growth factor receptor inhibitors: Synthesis, biological evaluation, and modeling of the mode of binding

Article abstract of DOI:10.1021/jm960789h

A new route to N-1-substituted pyrazolo- and pyrroloquinazolines has been developed from the known quinazolones 19 and 23, via conversion to the corresponding thiones, S-methylation to the thioethers, N-1-alkylation, and coupling with 3-bromoaniline. C-3-

Full text of DOI:10.1021/jm960789h

J . Med. Chem. 1997, 40, 1519-1529
1519
Tyr osin e Kin a se In h ibitor s. 11. Solu ble An a logu es of P yr r olo- a n d
P yr a zoloqu in a zolin es a s Ep id er m a l Gr ow th F a ctor Recep tor In h ibitor s:
Syn th esis, Biologica l Eva lu a tion , a n d Mod elin g of th e Mod e of Bin d in g
†
‡,§
‡
‡
‡
Brian D. Palmer, Susanne Trumpp-Kallmeyer, David W. Fry, J ames M. Nelson, H. D. Hollis Showalter, and
,
†
William A. Denny*
Cancer Society Research Laboratory, Faculty of Medicine and Health Science, The University of Auckland School of Medicine,
Private Bag 92019, Auckland 1000, New Zealand, and Parke-Davis Pharmaceutical Research, Division of Warner-Lambert
Company, 2800 Plymouth Road, Ann Arbor, Michigan 48106-1047
X
Received November 12, 1996
A new route to N-1-substituted pyrazolo- and pyrroloquinazolines has been developed from
the known quinazolones 19 and 23, via conversion to the corresponding thiones, S-methylation
to the thioethers, N-1-alkylation, and coupling with 3-bromoaniline. C-3-Substituted pyrrolo-
quinazolines were prepared by Mannich base chemistry. A series of compounds bearing
solubilizing side chains at these positions has been prepared and evaluated for inhibition of
the tyrosine kinase activity of the isolated epidermal growth factor receptor (EGFR) and of its
autophosphorylation in EGF-stimulated A431 cells. Several analogues, particularly C-3-
substituted pyrroloquinazolines, retained high potency in both assays. A model for the binding
of the general class of 4-anilinoquinazolines to the EGFR was constructed from structural
information (particularly for the catalytic subunit of the cAMP-dependent protein kinase) and
structure-activity relationships (SAR) in the series. In this model, the pyrrole ring in
pyrroloquinazolines (and the 6- and 7-positions of quinazoline and related pyridopyrimidine
inhibitors) occupies the entrance of the ATP binding pocket of the enzyme, with the pyrrole
nitrogen located at the bottom of the cleft and the pyrrole C-3 position pointing toward a pocket
corresponding to the ribose binding site of ATP. This allows considerable bulk tolerance for
C-3 substituents and lesser but still significant bulk tolerance for N-1 substituents. The
observed high selectivity of these compounds for binding to EGFR over other similar tyrosine
kinases is attributed to the 4-anilino ring binding in an adjacent hydrophobic pocket which
has an amino acid composition unique to the EGFR. The SAR seen for inhibition of the isolated
enzyme by the pyrazolo- and pyrroloquinazolines discussed here is fully consistent with this
binding model. For the N-1-substituted compounds, inhibition of autophosphorylation in A431
cells correlates well with inhibition of the isolated enzyme, as seen previously for related
pyridopyrimidines. However, the C-3-substituted pyrroloquinazolines show unexpectedly high
potencies in the autophosphorylation assay, making them of particular interest.
The epidermal growth factor receptor (EGFR) is
known to be overexpressed in a large percentage of
values of 0.44 nM for inhibition of the isolated enzyme.10
However, the poor aqueous solubility of these com-
pounds is a major drawback to their further develop-
ment.
1
-3
clinical cancers of various types
and to be associated
with poor prognosis.4 Compounds which inhibit the
ability of EGFR to transmit a phosphorylation signal
following binding of its cognate ligand EGF are therefore
of potential interest as anticancer drugs, and this is an
,5
6
-16
active field of drug development. A number of reports
have shown that the broad class of 4-anilinoquinazolines
are potent and highly selective inhibitors of EGFR
phosphorylation, resulting from competitive binding at
6
the ATP site. We have previously demonstrated struc-
In the present paper we discuss a possible binding
mode for these inhibitors on the EGFR enzyme. This
uses a molecular model of the EGFR TK constructed
from the refined structure of the catalytic subunit of the
ture-activity relationships (SAR) for compounds related
to the parent compound 1, showing the utility of small
lipophilic groups in the 3′-position7 and of electron-
9
donating substituents in the 6- or 7-position. In a later
1
7
cAMP-dependent protein kinase and structure-activ-
ity relationships for competitive inhibition by a series
of quinazolines, pyrido[d]pyrimidines, and fused tricyclic
1
0
paper, we showed that linear tricyclic derivatives
which incorporated electron-donating amino substitu-
ents into a fused 5- or 6-membered aromatic ring were
also potent inhibitors. Thus the pyrazolo- and pyrrolo-
quinazoline compounds 2a and 3a both showed IC50
8
-11,16
quinazolines.
Partly on the basis of this model,
several more soluble analogues of both the 2 and 3
tricyclic series (2b-d ,f,h , 3b-h , and 4-9) were pre-
pared and evaluated.
†
The University of Auckland School of Medicine.
Parke-Davis Pharmaceutical Research.
Address correspondence regarding molecular modeling to this
‡
Ch em istr y
§
The acetic acid derivatives 2h and 3h were prepared
author.
X
10
Abstract published in Advance ACS Abstracts, April 1, 1997.
by direct alkylation of the known parent heterocycles
S0022-2623(96)00789-3 CCC: $14.00 © 1997 American Chemical Society

1
520 J ournal of Medicinal Chemistry, 1997, Vol. 40, No. 10
Palmer et al.
Sch em e 1a
Sch em e 3a
a
(i) RCl‚HCl/CsCO3/4A molecular sieves/Me2CO/reflux/24-48
h.
Sch em e 4a
a
(i) NaH/20 °C/5 min, then BrCH2COOEt/20 °C/1.5 h; (ii) 2 N
NaOH/MeOH/30 min.
Sch em e 2a
a
(i) R1R2NH/HCHO/AcOH/50 °C/1-3 h.
and 3a . The previously reported10 route to these
compounds via the 4-chloroquinazolines gave much
lower yields, due largely to the very low solubility of 19
and 23.
Preparation of the tricycles bearing amino side chain
substituents at N-1 was best achieved (albeit in only
moderate yields) by treating the parent heterocycle with
the hydrochloride salt of the appropriate chloroalkyl-
amine in the presence of cesium carbonate and pow-
dered 4A molecular sieves (Scheme 3). Finally, intro-
duction of solubilizing substituents at the 3-position of
the pyrroloquinazoline was achieved by a Mannich
reaction between the parent pyrroloquinazoline 3a ,
formaldehyde, and the appropriate secondary amine, in
acetic acid at 50 °C (Scheme 4). Although quite stable
at room temperature, these 3-substituted products
decomposed at temperatures above ca. 130 °C, regen-
erating the starting tricycle in a reverse Mannich
reaction.
a
(
i) P2S5/pyridine/reflux/18 h; (ii) MeI/1 N NaOH/20 °C/2 h; (iii)
NaH/20 °C/5 min, then MeI/20 °C/30 min; (iv) 3-bromoaniline/
concentrated HCl/120 °C/4 h; (v) NaH/20 °C/5 min, then
ClCH2CH(OH)CH2OH/45 °C/30 min.
P r op osed Mod el for 4-(P h en yla m in o)-
qu in a zolin e Bin d in g to EGF R
2
a and 3a , using ethyl bromoacetate on the 1-sodio
derivatives followed by base hydrolysis of the resulting
esters (Scheme 1). Attempts to introduce 1-methyl and
Str u ctu r e-Activity Rela tion sh ip Stu d ies. Some
information identifying a possible binding mode on the
enzyme comes from the SAR of a large number of
4-(phenylamino)quinazolines for inhibition of the phos-
phorylation by EGFR of a fragment of phospholipase
Cγ1 (Table 1). The core 4-(phenylamino)quinazoline
pharmacophore 1 has an IC50 of 346 nM in this assay,
with a clear benefit resulting from small lipophilic
electron-withdrawing groups at the 3-position of the
1
-(dihydroxypropyl) substituents in an analogous man-
ner were unsuccessful, due primarily to the formation
of inseparable mixtures of products resulting from
alkylation in the anilinoquinazoline ring in addition to
the desired N-1 position. The required compounds were
eventually synthesized as shown in Scheme 2. The
known1
0,15
quinazolones 19 and 23 were converted into
8
the corresponding thiones 20 and 24, which were
methylated to form the thioethers 21 and 25 in good
yield. Alkylation of the thioethers with methyl iodide
or 1-chloropropane-2,3-diol proceeded cleanly at the N-1
position, and the resulting products (22 and 26-28)
were reacted with 3-bromoaniline at 120 °C to give the
desired 5-anilino derivatives. This also proved a sig-
nificantly superior route to the parent compounds 2a
anilino moiety. Thus, substitution of 1 with 3′-Br to
give 10 increases inhibitory potency more than 10-fold
(IC50 27 nM). However, substitution of the anilino side
chain at other positions, or with larger groups, greatly
reduces activity, suggesting limited bulk tolerance.16
The requirement for the quinazoline structure was
established by studying alterations in the nitrogen
substitution pattern within the bicyclic ring system.8

Tyrosine Kinase Inhibitors
J ournal of Medicinal Chemistry, 1997, Vol. 40, No. 10 1521
Ta ble 1. Structure-Activity Relationships for Inhibition of
EGFR by 4-(Phenylamino)quinazolines
Leu-820 on strand 7 and Thr-830 on strand 8. The
ribose and the triphosphate moieties extend toward the
opening of the cleft where the phosphate transfer occurs.
The ribose 2′-OH can form a hydrogen bond with Cys-
7
73, either directly (as in cAMP kinase) or indirectly
through water (as in casein kinase). The same binding
mode has been observed in several recently published
protein kinase structures in agreement with the high
degree of sequence conservation present in the catalytic
no.
formula
R
X
Y
Z
IC50 (nM)a
ref
18-21
core of protein kinases.
Despite the high degree of
1
A
A
A
A
A
A
A
A
B
B
H
H
H
H
H
N
N
N
C
N
N
N
N
N
N
C
N
N
N
N
N
346
27
>100000
5500
0.006
0.025
>10000
>10000
0.44
8
8
8
8
9
9
9
9
10
10
homology, there are some significant amino acid differ-
ences in the ATP binding site of EGFR TK compared to
other tyrosine kinases, which presumably account for
the high selectivity of this class of compounds for the
EGFR TK. The entrance of the binding pocket, which
includes the Cys-773 residue, is considerably more
hydrophobic than in PDGFR (Asp), FGFR (Asn), or c-src
1
1
1
1
1
1
1
2
3
0
1
2
3
4
5
6
a
a
Br
Br
Br
Br
Br
Br
Br
6,7-diOEt
6,7-diOMe
2-Me
8-OMe
N
(Ser). An investigation of the ATP binding pocket for
CH
0.44
the EGFR TK shows three additional sulfur-containing
amino acid residues (Cys-751, Met-769, and Met-742).
In particular, the Cys-751 residue on strand 4 is only
present in EGFR TK and is part of a hydrophobic pocket
adjacent to the ATP binding pocket. This pocket is
present in the other tyrosine kinases as well but is more
shallow due to the exchange of Cys-751 for Val or Ile.
The ATP competitive inhibitor isopentyladenine (17)
has been cocrystallized with the similar enzyme CDK2
and forms a homologous bidentate hydrogen bond with
the backbone of the corresponding amino acid residues
on the extended coil stretch. The crystal structure of
CDK2 complexed with the inhibitor olomucine (18) also
a
IC50, concentration of drug (nM) to inhibit the phosphorylation
of a 14-residue fragment of phospholipase Cγ1; see references.
Substitution of the N-1 in 1 by carbon to give 11 resulted
in a >3700-fold loss of activity, whereas similar substi-
tution of N-3 (to give 12) produced a 200-fold loss. This
indicates the importance of both nitrogens in the
interaction with the enzyme (potentially as H-bond
acceptors). Substitution of C-2 in the quinazoline ring
system of 1, even with small substituents such as Me
9
(
(
e.g., 15), abolishes activity, while substituents at C-8
e.g., 16) are not favorable.8 This face of the pharma-
cophore therefore also appears to bind to an area of
restricted size. In sharp contrast, steric bulk can be
readily tolerated at the 6- and 7- positions, as exempli-
fied by the 4-(3-bromoanilino)-6,7-dialkoxyquinazolines.
The diethoxy compound 13 shows similar potency (IC50
9
0
.006 nM) to the corresponding dimethoxy analogue 14.
Moreover, as noted above, a third ring can be fused to
the bicyclic quinazoline ring system without loss of
binding affinity (e.g., 2a and 3a both have IC50s of 0.44
nM).10 Generally, where comparisons have been made,
the pyrido[d]pyrimidines also show similar SAR to the
quinazolines. In the parent compounds, aza substitu-
tion in the ring at the 5-7-positions has little effect,
but at the 8-position, this substitution is strongly
detrimental.
shows a bidentate hydrogen bond between the inhibitor
and the extended coil stretch, although one of the
hydrogen bonds is formed with the backbone of a
different amino acid residue, located closer to the
1
1
2
1
entrance of the ATP binding pocket. These structural
data also suggest that hydrogen bonding is essential for
inhibitor binding, although the same amino acids on the
extended coil stretch are not necessarily involved. The
above hydrogen bond donor-acceptor is not present in
the 4-(phenylamino)quinazolines. Nevertheless, the
SAR studies (discussed above) indicate an essential role
for the quinazoline N-1 and an important role for N-3,
suggesting that both nitrogens might act as hydrogen
bond acceptors. The 3D model of the EGFR TK suggests
that these hydrogen bonds are formed with amino acid
residues from the extended coil stretch at the bottom
of the ATP binding site cleft.
Str u ctu r a l Da ta fr om Oth er P r otein Kin a ses.
The ATP binding site is a common feature of the entire
PK family and is defined by two â-sheets that form a
cleft between the N- and C-terminal lobes. In the 3D
model of EGFR TK constructed from the refined struc-
ture of the catalytic subunit of the cAMP-dependent
1
7
protein kinase, the adenine base is anchored by a
bidentate donor-acceptor hydrogen-bonding motif. The
hydrogen bonds between the N-6 hydrogen and the
backbone carbonyl of Gln-767 and between N-1 and the
backbone amide of Met-769 fix the purine onto the
extended coil stretch that connects the N-terminal lobe
with the C-terminal lobe of the enzyme. A third
hydrogen bond between N-7 and the hydroxyl group of
Thr-830 located on strand 8 preceding the activation
loop may be less important because it is not conserved
in all protein kinases.
Resu lts a n d Discu ssion
The structures and physicochemical properties of the
new tricyclic derivatives prepared are given in Table 2.
All the analogues were evaluated for their ability to
inhibit tyrosine phosphorylation of a random copolymer
of tyrosine and glutamic acid (Sigma) by EGF-stimu-
lated full-length EGFR enzyme isolated from A431
Nonpolar interactions with one face of the purine ring
occur with Val-702 on strand 3 and Ala-719 and Leu-
7
68 which are part of the conserved glycine-rich flap.
6
Nonpolar interactions from the opposite face come from
cells. Full dose-response curves were determined for

1
522 J ournal of Medicinal Chemistry, 1997, Vol. 40, No. 10
Palmer et al.
Ta ble 2. Structural and Biological Properties of Soluble Substituted 5-[(3-Bromophenyl)amino]pyrrolo[3,2-g]quinazolines and
-pyrazolo[4,3-g]quinazolines
IC50 (nM)a
no.
R
mp or ref
Pyrazolo Series
ref 10
227-229
261
262-264
216-220
252 dec
sol (mM)
EGFR
autophos
2
2
2
2
2
2
a
b
c
d
f
H
Me
0.11
0.33
0.10
36
0.44
0.37
12
40
3.7
53
20
13
318
262
192
1196
CH2CH(OH)CH2OH
(CH2)2NMe2
(CH2)2Nmorpholide
CH2COOH
b
c
h
0.26
Pyrrolo N-Substituted Series
ref 10
3
3
3
3
3
3
3
3
a
b
c
d
e
f
H
Me
1.5
0.49
9.7
0.44
0.80
1.6
41
22
22
49
145
230
16
77
1780
178-180
CH2CH(OH)CH2OH
(CH2)2NMe2
(CH2)3NMe2
(CH2)2Nmorpholide
(CH2)3Nmorpholide
CH2COOH
184-187
150-152
171-172
11
21
114-116
2.5
2.4
11
3.7
8.8
5.1
g
h
123-126
>170 dec
Pyrrolo 3-Substituted Series
240-245 dec
220 dec
4
5
6
7
8
9
CH2N(CH2CH2OH)2
CH2NMe2
CH2Nmorpholide
CH2N(Me)(CH2)2NMe2
CH2N(Me)CH2COOMe
CH2N(Me)CH2COOH
3.5
2.6
4.8
7.5
3.4
0.72
135
9.7
9.6
81
9.6
584
20
11
27
229 dec
139-141 dec
227-230 dec
240 dec
1.5
a
IC50, concentration of drug (nM) to inhibit the phosphorylation of a random polyglutamic acid/tyrosine copolymer by EGFR, prepared
from human A431 carcinoma cell vesicles by immunoaffinity chromatography; see the Experimental Section for details. Values are the
averages from at least two independent dose-response curves; variation was generally (15%. b HCl salt. c DiHCl salt.
each compound, and the resulting IC50s listed in Table
are the average of at least two such determinations.
activity. However, in this case the larger propanediol
substituent had lesser effect on potency, with 3c also
being only 4-fold less potent than 3a , with an IC50 of
1.6 nM. Again, the strongly basic (CH2)2NMe2 side
chain resulted in a considerable loss of potency for 3d ,
while the longer chain homologue 3e was slightly more
effective. The more weakly basic ethylmorpholide 3f
was the best of the basic analogues, being only 8-fold
less potent than 3a , but both morpholide analogues
(3f,g) were much less soluble than the more strongly
basic compounds. The anionic derivative 3h also re-
tained moderate activity (IC50 5.1 nM) and good solubil-
ity. Finally, the 3-substituted pyrrolo series (4-9) was
overall the most effective, with IC50s averaging about 4
nM. No clear trend was present, with the anionic
derivative 9 in this case being the most effective, with
an IC50 of 0.72 nM.
To see whether the EGFR binding model proposed
above was consistent with these results, manual docking
studies were performed to explore the fit of these
compounds to the enzyme. Different initial orientations
of the aromatic rings were used for selected bicyclic
quinazolines and tricyclic pyrrolo- and pyrazoloquinazo-
lines. Amino acid side chains were reoriented to relieve
unfavorable steric interactions, and the complexes were
subsequently energy minimized. Only SAR data were
considered as criteria in choosing the most plausible
complex because of the inability of current force fields
to reproduce the large induced fit (and therefore correct
enzyme-inhibitor interaction energies) of these en-
zymes upon ligand binding.
1
As noted above, both previous SAR for the general class
of 4-(phenylamino)quinazolines and molecular modeling
studies suggested significant bulk tolerance at the 6/7-
positions, allowing compounds such as the 6,7-dieth-
oxyquinazoline 13 and the tricyclic derivatives 2a and
3
a to retain high potency. In the present study, the
N-substituted pyrazolo and pyrrolo compounds listed in
Table 2 were prepared to probe the extent of this bulk
tolerance and to provide more water-soluble analogues.
The solubilities of selected compounds were determined
by HPLC following sonication for 30 min at 20 °C.
Lactate buffer (0.05 M) was used for compounds with
neutral side chains and water for hydrochloride salts
of amines and sodium salts of acids.
In the N-substituted pyrazolo series, the NMe ana-
logue 2b had similar potency to 2a , showing that an
H-bond donation here is not critical. The diol 2c was
prepared to retain a noncharged side chain but did not
show improved solubility. As expected, the strongly
basic (CH2)2NMe2 side chain in 2d did provide much
more solubility (36 mM), but the compound was about
1
00-fold less potent than 2a . The weakly basic mor-
pholide side chain analogue 2f was much more active,
with an IC50 of 3.7 nM. Finally, the anionic CH2COOH
compound 2h was much less effective.
In the pyrrolo series (3a -h ), the parent compound
a was appreciably more soluble than the corresponding
3
pyrazolo analogue 2a . Substitution of the pyrrole
nitrogen with Me in 3b again resulted in little loss of

Tyrosine Kinase Inhibitors
J ournal of Medicinal Chemistry, 1997, Vol. 40, No. 10 1523
F igu r e 1. Superposition of compound 10 (green) and ATP (cyan) in the EGFR TK. In particular, the 4-[(3-bromophenyl)amino]
substituent binds to a hydrophobic pocket deep in the binding cleft, which is not used by ATP.
Only one binding mode was found that satisfied all
the SAR data described above and that had no major
unfavorable steric interactions. In this binding mode
substituent is located in a deep, moderately sized pocket
adjacent to the adenine binding pocket. The bottom of
this inhibitor pocket is made up of Thr-766 (strand 5),
Cys-751 (strand 4), Thr-830, and Met-742 (helix RC).
The top of this pocket is more hydrophilic with Glu-738
(helix RC), Lys-721 (strand 3), and Asp-831 (strand 8).
The 4-(phenylamino)quinazolines, but not ATP, in-
teract specifically with this pocket. The 4-(3-bromo-
anilino) substituent is adjacent to Cys-751 and Met-742
and can form favorable sulfur-aromatic interactions.
The amino acid composition of the bottom of this pocket
is unique to EGFR TK. In particular, a cysteine residue
(Cys-751) is present only in the EGFR TK, which might
explain part of the selectivity of this class of compounds
for EGFR. The fact that there are very limited sites
for favorable substitution off the 4-(3-bromoanilino) ring
is in agreement with the restricted size of this pocket.
The additional binding energy provided by the 4-(3-
bromoanilino) substituent in this deep pocket may be
the reason for the extremely high binding affinity of
these small molecules. In comparison, the much less
potent (micromolar range) competitive inhibitors iso-
(illustrated with 10 in Figure 1), the bicyclic (or tricyclic)
chromophore binds to a narrow hydrophobic pocket in
the N-terminal domain of EGFR TK, which is the
binding site of the adenine base of ATP. The N-1 of the
quinazoline ring forms a hydrogen bond with the NH
backbone of Met-769. The NH backbone of this amino
acid residue normally hydrogen bonds to N-1 of the
adenine ring in ATP and helps to fix it in the binding
pocket. This hydrogen bond is crucial for orienting the
quinazoline ring in the pocket as demonstrated by more
than a 3700-fold loss in inhibitory potency upon sub-
stitution of N-1 for a carbon atom. A second hydrogen
bond can be formed with N-3 of the quinazoline ring
and the side chain of Thr-766, which is located at the
beginning of the extended coil stretch deep in the
binding cleft. Removal of the N-3 of the quinazoline ring
results in a moderate loss of affinity (a 200-fold loss in
inhibitory potency) indicating that this hydrogen bond
is probably less important. The 4-(3-bromoanilino)

1
524 J ournal of Medicinal Chemistry, 1997, Vol. 40, No. 10
Palmer et al.
F igu r e 2. Compound 9 docked into the ATP binding site of the EGFR TK. The N-1 of the quinazoline ring forms a hydrogen
bond with the NH backbone of Met-769. A second hydrogen bond can be formed with N-3 of the quinazoline ring and the side
chain of Thr-766, located at the end of strand 5 deep in the binding cleft. The C-3 substituent of the pyrrole ring is located in a
pocket corresponding to the ribose binding site of ATP and can form electrostatic interactions with Arg-817. In N-substituted
pyrroloquinazolines, the substituents extend out of the binding pocket and do not participate in specific interactions with the
enzyme.
pentyladenine (17) and olomucine (18) do not occupy the
corresponding hydrophobic pocket when complexed with
CDK2, binding closer to the entrance of the ATP binding
site and making fewer hydrophobic contacts.
pyrazoloquinazolines can be tolerated without a major
loss of affinity.
This is consistent with the present results (Table 2),
which show that some N-1- and C-3-substituted pyr-
roloquinazolines in particular (e.g., 3c,f, 5, and 9) retain
high potency against the enzyme. This is consistent
with an enzyme-inhibitor complex in which the 3-sub-
stituted pyrrole ring is occupying the entrance of the
ATP binding pocket, with the pyrrolo NH located at the
bottom of the cleft close to the extended coil stretch. The
C-3 side chain in compound 9 is thus located in a pocket
corresponding to the ribose binding site of ATP and can
form electrostatic interactions with Arg-817 located on
the catalytic loop. Additional interactions might occur
with Cys-773 on the extended coil stretch (Figure 2).
On the other hand, N-1-substituted pyrroloquinazolines
are on average somewhat less active than the parent
3a , with only the smallest substituent (methyl) being
tolerated without loss of activity. The loss of potency
The model predicts that substitution of H-2 or H-8 in
the quinazoline ring will be detrimental to binding, as
shown by previous SAR data.9 Both of the correspond-
ing carbon atoms are located close to the peptide
backbone of the extended coil stretch, and their substi-
tution would prevent proper hydrogen bonding of N-1
and N-3 of the quinazoline ring with the extended coil
stretch (Figure 1). Loss of potency upon substitution
1
1
of C-8 with aza is also explained by this binding mode,
where the aza is located close to the carbonyl oxygen of
Met-769 at the bottom of the cleft, leading to unfavor-
able repulsive electrostatic interactions (Figure 1). The
model also predicts that large substituents at C-6 and
C-7 of the quinazoline and pyrido[d]pyrimidines and at
C-3 and (to some extent) N-1 of the tricyclic pyrrolo- and

Tyrosine Kinase Inhibitors
J ournal of Medicinal Chemistry, 1997, Vol. 40, No. 10 1525
with longer, bulkier alkyl side chains might be due to
unfavorable steric interactions with the bottom of the
pocket. This causes a slight reorientation of the mol-
ecule in the binding site as is apparent from docking
studies. In the enzyme-inhibitor complex, long alkyl
side chains extend out of the binding pocket and do not
participate in specific interactions with the enzyme.
The compounds were also evaluated for their ability
to inhibit the autophosphorylation of EGFR in EGF-
stimulated A431 cells in culture, quantitating the
response by Western blotting. Both parent compounds
This allows considerable bulk tolerance for C-3 substit-
uents and lesser but still significant bulk tolerance for
N-1 substituents. The data for inhibition of substrate
phosphorylation by EGFR are consistent with this
binding model. A similar approach has recently been
2
3
reported to construct a binding model for the dianili-
nophthalimide EGFR inhibitor CGP 52411.
This binding model is different to one published
recently by the Ciba-Geigy group for a series of related
24
4
-(phenylamino)pyrrolopyrimidines.
In their model,
the pyrrolopyrimidine chromophore is rotated through
80°, placing the 4-(phenylamino) substituent at the
2
2
a and 3a showed reasonable potency in this assay (ca.
0 nM), and the pyrrolo C-2 morpholide 3f, which was
1
entrance of the ATP binding pocket, occupying the
ribose binding site. The extensive SAR data presented
in the current paper clearly exclude this binding mode.
In particular, the N-substituted pyrroloquinazolines are
unable to bind in their proposed binding mode, due to
steric interference of the N-1 substituents with strands
also the most potent against the isolated enzyme, was
slightly superior with an IC50 of 16 nM. Overall, for
the N-substituted pyrazolo and pyrrolo compounds 2
and 3, there was a good correlation between potency for
inhibition of autophosphorylation in cells and inhibition
of phosphorylation by the isolated enzyme (eq 1). The
4
and 5 and the extended coil stretch. The correlation
between inhibition of isolated enzyme and inhibition of
autophosphorylation for the N-substituted tricyclic com-
pounds is similar to that seen previously for pyridopy-
rimidines. This suggests that the two classes of com-
pounds function similarly and that similar solubilizing
functions might be acceptable in both classes. However,
the unexpectedly high potencies of the C-3-substituted
compounds in the autophosphorylation assay show that
activity against the isolated enzyme does not always
predict well for activity in cells and, therefore, make
these compounds of particular interest.
log(IC (autophos)) ) 0.67((0.12)
5
0
log(IC (enzyme)) + 1.47((0.11)
5
0
n ) 13, r ) 0.87, F2,11 ) 33, p ) 0.001
(1)
only exception to this was the pyrrolo acid derivative
h , which was less active than predicted in the auto-
3
phosphorylation assay and was not used in the deriva-
tion of eq 1. Previous studies22 with acid derivatives
have also shown that they have poorer cellular activity
than predicted by their inhibition of the isolated en-
zyme, presumably due to low uptake into cells. This
equation agrees well with one derived previously for
related pyridopyrimidines:11
Exp er im en ta l Section
Analyses were carried out in the Microchemical Laboratory,
University of Otago, Dunedin, NZ. Melting points were deter-
mined on an Electrothermal 9200 melting point apparatus.
NMR spectra were obtained on a Bruker AC-200 or AM-400
log(IC (autophos)) ) 0.98 log(IC (enzyme)) + 0.81
5
0
50
(2)
spectrometer and are referenced to Me
and sodium 3-(trimethylsilyl)propanesulfonate for D
4
Si for organic solutions
O solu-
2
However, the autophosphorylating activity of the 3-sub-
stituted pyrrolo compounds 4-9 was not well fitted by
eq 1 and even within this series was not well correlated
with enzyme IC50. The strong base analogue 5, the
weak base 6, and the neutral ester 8 all showed
autophosphorylation activity superior to that of the
parent compound 3a , with IC50s below 10 nM.
tions. Thin-layer chromatography was carried out on alumi-
num-backed silica gel plates (Merck 60 F254). Flash column
chromatography was carried out on Merck silica gel (230-400
mesh). Petroleum ether refers to the fraction boiling at 40-
60 °C. Mass spectra were recorded on a Varian VG 7070
spectrometer at nominal 5000 resolution.
5-[(3-Br om op h en yl)a m in o]p yr r olo[3,2-g]qu in a zolin e-
1
0,15
1
-a cetic Acid (3h ): Sch em e 1. A solution of the known
Con clu sion s
pyrroloquinazoline 3a (0.10 g, 0.29 mmol) in DMF (1 mL) was
added dropwise under nitrogen to a suspension of NaH (14.1
mg of a 60% dispersion in oil, 0.35 mmol) in THF (1 mL). After
An efficient synthetic route has been developed for
N-1-substituted pyrroloquinazolines, via alkylation of
the thioether 21. C-3 Mannich base analogues can also
be readily prepared from the preformed compound 3a .
This allowed the preparation of a series of analogues
bearing quite bulky solubilizing side chains at these
positions. Some of these, particularly the C-3 ana-
logues, retained high potency against the isolated
enzyme. A molecular model of the EGFR TK was
constructed from SAR data for inhibition of the EGFR
by 4-(phenylamino)quinazolines and analogues and from
structural information (particularly for the catalytic
subunit of the cAMP-dependent protein kinase ). Us-
ing this model, a possible binding mode for this class of
tricyclic inhibitors was identified, where the pyrrolo or
pyrazolo ring occupies the entrance of the ATP binding
pocket of the enzyme, with the nitrogen located at the
bottom of the cleft and the C-3 position pointing toward
a pocket corresponding to the ribose binding site of ATP.
5
min, ethyl bromoacetate (36 µL, 0.32 mmol) was added, and
the solution was stirred at 20 °C for 1.5 h. Water was added,
and the mixture was extracted with EtOAc and worked up to
give an oil which was chromatographed on silica gel. Elution
with EtOAc/petroleum ether (1:4) gave a fraction enriched in
the desired product, which was further purified by recrystal-
lization from EtOAc/petroleum ether to give ethyl 5-[(3-
bromophenyl)amino]pyrrolo[3,2-g]quinazoline-1-acetate (3i) (40
1
mg, 32%) as a cream solid: mp 194-196 °C; H NMR (CDCl
3
)
δ 8.70 (br s, 1 H, NH), 8.12 (br s, 2 H), 7.71 (m, 2 H), 7.32 (d,
J ) 3.4 Hz, 1 H), 7.28-7.26 (m, 3 H), 6.69 (d, J ) 3.4 Hz, 1
H), 4.90 (s, 2 H), 4.24 (q, J ) 7.1 Hz, 2 H), 1.29 (t, J ) 7.1 Hz,
3 H); 13C NMR δ 168.11 (s), 152.69 (d), 145.15 (s), 140.63 (s),
1
7
1
(
1
40.10 (s), 133.52 (s), 130.25 (d), 130.07 (s), 126.88 (d), 124.27
d), 122.62 (s), 119.87 (d), 111.82 (d), 109.97 (s), 105.67 (d),
02.90 (d), 62.05 (t), 47.85 (t), 14.13 (q). Anal. (C20
C, H, N.
4 2
H17BrN O )
A solution of 3i (40 mg, 0.09 mmol) in MeOH (5 mL) was
treated with 2 N NaOH (1 mL), and the mixture was stirred
at 20 °C for 30 min. The MeOH was removed under reduced

1
526 J ournal of Medicinal Chemistry, 1997, Vol. 40, No. 10
Palmer et al.
pressure, and the remaining solution was carefully neutralized
with 1 N HCl. The precipitated product was collected and
washed well with water to give 5-[(3-bromophenyl)amino]-
130.05 (s), 117.20 (s), 114.22 (d), 105.48 (d), 101.24 (d), 32.79
(q), 12.03 (q). Anal. (C12
11 3
H N S) C, H, N.
A solution of thioether 22 (0.10 g, 0.44 mmol), 3-bromo-
aniline (0.10 mL, 0.96 mmol), and concentrated HCl (39 µL,
0.48 mmol) in N-methylpyrrolidinone (20 mL) was warmed at
pyrrolo[3,2-g]quinazoline-1-acetic acid (3h ) (28 mg, 78%): mp
1
(
1
MeOH) 175 °C dec; H NMR [(CD
3
)
2
SO] δ 13.80 (br, 1 H),
0.63 (br, 1 H), 9.02 (s, 1 H), 8.73 (s, 1 H), 8.25 (s, 1 H), 7.92
2
120 °C for 4 h, then cooled, and poured into Et O (200 mL).
(
dt, J ) 6.9, 1.8 Hz, 1 H), 7.84 (s, 1 H), 7.78 (d, J ) 3.4 Hz, 1
After chilling at -20 °C overnight, the precipitated solid was
H), 7.43-7.38 (m, 2 H), 6.84 (d, J ) 3.4 Hz, 1 H), 5.22 (s, 2 H);
filtered off and dissolved in hot MeOH, and the solution was
1
3
C NMR δ 169.97 (s), 158.88 (s), 150.73 (d), 140.41 (s), 140.28
s), 135.59 (d), 130.40 (d), 129.48 (s), 126.89 (d), 125.19 (d),
21.66 (d), 121.11 (s), 115.11 (d), 108.37 (s), 102.10 (d), 102.00
3
basified with concentrated aqueous NH and then concentrated
(
1
to dryness. The residue was partitioned between EtOAc and
water, and the organic portion was worked up and chromato-
graphed on alumina. EtOAc/petroleum ether (1:1) gave fore-
runs, while EtOAc eluted 5-[(3-bromophenyl)amino]-1-meth-
4 2
(d), 47.35 (t). Anal. (C18H13BrN O ) C, H, N.
Similar alkylation of the known10 5-[(3-bromophenyl)amino]-
H-pyrazolo[4,3-g]quinazoline (2a ) with NaH and ethyl bro-
ylpyrrolo[3,2-g]quinazoline (3b) (97 mg, 62%): mp (EtOAc/
1
1
petroleum ether) 178-180 °C; H NMR [(CD
3
)
2
SO] δ 9.81 (s,
moacetate gave ethyl 5-[(3-bromophenyl)amino]-1H-pyrazolo-
4,3-g]quinazoline-1-acetate (2i) (58%): mp (EtOAc/petroleum
1
H), 8.83 (s, 1 H), 8.56 (s, 1 H), 8.34 (t, J ) 1.9 Hz, 1 H), 7.98
[
1
(dd, J ) 7.2, 1.5 Hz, 1 H), 7.85 (s, 1 H), 7.70 (d, J ) 3.2 Hz, 1
ether) 191 °C; H NMR (CDCl
8
7
3
) δ 8.71 (s, 1 H), 8.29 (s, 2 H),
.22 (s, 1 H), 8.08 (br s, 1 H), 7.77 (br s, 1 H), 7.72 (br d, J )
.2 Hz, 1 H), 7.31-7.29 (m, 2 H), 5.24 (s, 2 H), 4.26 (q, J ) 7.1
H), 7.36 (dd, J ) 7.9, 7.2 Hz, 1 H), 7.28 (dd, J ) 7.9, 1.5 Hz,
1
H), 6.75 (d, J ) 3.2 Hz, 1 H), 3.91 (s, 3 H); 13C NMR δ 157.82
1
3
(s), 151.75 (d), 144.60 (s), 141.56 (s), 140.29 (s), 134.93 (d),
30.30 (d), 128.93 (s), 125.35 (d), 123.71 (d), 121.14 (s), 120.26
d), 113.69 (d), 109.32 (s), 105.18 (d), 100.88 (d), 32.72 (q). Anal.
‚H O) C, H, N.
-[(3-Br om op h en yl)a m in o]-1-m et h ylp yr a zolo[4,3-g]-
Hz, 2 H), 1.29 (t, J ) 7.1 Hz, 3 H); C NMR δ 167.82 (s), 158.34
s), 154.72 (d), 142.01 (s), 139.51 (s), 135.47 (d), 130.31 (d),
27.53 (d), 124.80 (d), 124.60 (s), 122.63 (s), 120.41 (d), 114.05
d), 111.01 (s), 105.08 (d), 62.16 (t), 50.37 (t), 14.12 (q). Anal.
‚0.5H O) C, H.
Hydrolysis of 2i as above gave 5-[(3-bromophenyl)amino]-
pyrazolo[4,3-g]quinazoline-1-acetic acid (2h ) (91%): mp (EtOAc)
1
(
1
(
(
C
17
H
13BrN
4
2
(
(
5
C
19
H
16BrN
5
O
2
2
1
0
qu in a zolin e (2b). Similar reaction of the known 1H-
pyrazolo[4,3-g]quinazolinone 23 with P gave 1H-pyrazolo-
[4,3-g]quinazoline-5(6H)-thione (24) (85%): mp >400 °C; H
NMR [(CD SO] δ 13.52 (br, 2 H), 9.18 (d, J ) 0.7 Hz, 1 H),
.52 (d, J ) 0.7 Hz, 1 H), 8.15 (s, 1 H), 7.79 (s, 1 H); C NMR
δ 187.20 (s), 142.49 (s), 142.23 (d), 141.14 (s), 136.14 (d), 124.30
s), 124.10 (d), 122.99 (s), 105.45 (d); DEIMS found M
02.0310, C S requires 202.0313.
Similar methylation of 24 gave 5-(methylthio)-1H-pyrazolo-
2
S
5
1
1
2
52 °C dec; H NMR [(CD
3 2
) SO] δ 13.10 (br, 1 H), 10.12 (br, 1
3 2
)
H), 9.17 (s, 1 H), 8.61 (s, 1 H), 8.56 (s, 1 H), 8.32 (br s, 1 H),
1
3
8
8
5
1
.04 (s, 1 H), 7.97 (d, J ) 8.1 Hz, 1 H), 7.40-7.25 (m, 2 H),
13
.42 (s, 2 H); C NMR δ 158.71 (s), 153.91 (d), 146.66 (s),
+
(
2
41.82 (s), 141.05 (s), 136.62 (s), 135.24 (d), 130.35 (d), 125.95
9 6 4
H N
(d), 124.17 (d), 123.85 (s), 121.13 (s), 120.70 (d), 116.71 (d),
+
1
4
10.70 (s), 104.82 (d), 49.98 (t); FABMS found [M + H]
00.0230, 398.0203, C17 requires 400.0203, 398.0253.
12BrN ‚0.5H O) C, H.
-[(3-Br om op h e n yl)a m in o]-1-m e t h ylp yr r olo[3,2-g]-
1
[
3 2
4,3-g]quinazoline (25) (58%): mp 380 °C dec; H NMR [(CD ) -
H
5 2
12BrN O
SO] δ 13.53 (s, 1 H), 8.91 (s, 1 H), 8.75 (s, 1 H), 8.55 (s, 1 H),
Anal. (C17
H
5
O
2
2
7
1
.99 (s, 1 H), 2.73 (s, 3 H); 13C NMR δ 172.72 (s), 152.05 (d),
43.97 (s), 141.81 (s), 135.64 (d), 124.14 (s), 117.81 (d), 117.05
5
1
5
qu in a zolin e (3b): Sch em e 2. A suspension of the known
pyrrolo[3,2-g]quinazolone 19 (1.55 g, 8.37 mmol) and P S (3.72
2 5
g, 16.74 mmol) in pyridine (80 mL) was heated under reflux
for 18 h and then concentrated to dryness. The residue was
slurried in boiling water and filtered. The solid was dissolved
in 1 N NaOH and filtered and the filtrate carefully neutralized
with saturated aqueous NH Cl. The precipitated solid was
4
collected and dried under vacuum to give 1H-pyrrolo[3,2-g]-
quinazoline-5(6H)-thione (20) (1.54 g, 91%): mp >370 °C; H
NMR [(CD
8 4
(s), 105.31 (d), 12.13 (q). Anal. (C10H N S) C, H, N, S.
Similar reaction of 25 with NaH and MeI afforded 1-methyl-
5
2
-(methylthio)pyrazolo[4,3-g]quinazoline (26) (96%): mp (EtOAc)
1
02-203 °C; H NMR (CDCl
3
) δ 8.93 (s, 1 H), 8.62 (d, J ) 0.9
Hz, 1 H), 8.30 (d, J ) 0.9 Hz, 1 H), 7.89 (s, 1 H), 4.19 (s, 3 H),
2
1
13
.76 (s, 3 H); C NMR δ 173.81 (s), 152.83 (d), 144.72 (s),
41.89 (s), 134.41 (d), 124.86 (s), 118.90 (s), 117.53 (d), 104.80
(
d), 35.81 (q), 12.69 (q). Anal. (C11
Coupling of 26 with 3-bromoaniline as above gave 5-[(3-
bromophenyl)amino]-1-methylpyrazolo[4,3-g]quinazoline (2b)
10 4
H N S) C, H, N.
1
3
)
2
2
SO] δ 13.50 (br, 1 H, exch with D O), 11.80 (br s,
1
7
H, exch with D
2
O), 8.96 (s, 1 H), 8.08 (s, 1 H), 7.74 (s, 1 H),
1
(
[
66%): mp (EtOAc/petroleum ether) 227-229 °C; H NMR
.73 (dd, J ) 3.1, 2.4 Hz, 1 H), 7.40 (t, J ) 50.8 Hz, 1 H), 6.80
(CD SO] δ 10.05 (s, 1 H), 9.14 (s, 1 H), 8.61 (s, 1 H), 8.51 (d,
3 2
)
1
3
(
br s, 1 H); C NMR δ 185.81 (s), 140.95 (s), 140.36 (d), 138.93
J ) 0.9 Hz, 1 H), 8.31 (t, J ) 1.9 Hz, 1 H), 8.00 (d, J ) 0.9 Hz,
1
(s), 130.74 (d), 129.73 (s), 121.98 (s), 121.41 (d), 107.33 (d),
H), 7.97 (br d, J ) 8.0 Hz, 1 H), 7.38 (dd, J ) 8.0, 7.9 Hz, 1
+
1
2
02.46 (d); DEIMS found M 201.0367, C10
01.0361.
H
7
N
3
S requires
13
H), 7.31 (br d, J ) 7.9 Hz, 1 H), 4.16 (s, 3 H); C NMR δ 158.69
s), 153.82 (d), 146.51 (s), 141.23 (s), 141.01 (s), 134.14 (d),
30.31 (d), 125.89 (d), 124.13 (d), 123.63 (s), 121.11 (s), 120.65
d), 116.58 (d), 110.48 (s), 104.46 (d), 35.48 (q). Anal. (C16
(
1
(
Methyl iodide (0.42 mL, 6.78 mmol) was added to a solution
of 20 (1.30 g, 6.46 mmol) and 1 N NaOH (9.00 mL, 9.00 mmol)
in MeOH/water (1:1, 80 mL). The mixture was stirred at room
temperature for 2 h, and the precipitate was filtered off and
washed well with water to give 5-(methylthio)-1H-pyrrolo[3,2-
g]quinazoline (21) (0.84 g, 60%): mp (MeOH) 278 °C dec; H
NMR [(CD
H
12
-
5
BrN ) C, H, N.
5
-[(3-Br om op h en yl)a m in o]-1-(2,3-d ih yd r oxyp r op yl)-
p yr r olo[3,2-g]qu in a zolin e (3c). A solution of the pyrrolo-
quinazoline 21 (0.20 g, 0.93 mmol) in DMF (2 mL) was added
dropwise under nitrogen to a suspension of NaH (0.11 g of a
1
3
)
2
SO] δ 11.63 (br s, 1 H, exch with D
2
O), 8.84 (s, 1
H), 8.37 (s, 1 H), 7.91 (s, 1 H), 7.80 (dd, J ) 3.0, 2.4 Hz, 1 H),
6
1
0% dispersion in mineral oil, 2.79 mmol) in THF (2 mL). After
0 min, 1-chloropropane-2,3-diol (85 µL, 1.02 mmol) was added,
1
3
6
(
1
.80 (dd, J ) 3.0, 0.9 Hz, 1 H), 2.71 (s, 3 H); C NMR δ 170.08
s), 150.36 (s), 142.09 (s), 140.59 (s), 131.96 (d), 129.90 (s),
17.26 (s), 113.73 (d), 106.85 (d), 101.79 (d), 11.97 (t). Anal.
S) C, H, N, S.
A solution of 21 (0.10 g, 0.46 mmol) in DMF (1 mL) was
and the solution was warmed at 45 °C for 30 min. Water was
added, and the mixture was extracted into EtOAc, worked up,
and chromatographed on silica gel. EtOAc eluted starting
material (90 mg, 45%), while MeOH/EtOAc (1:9) gave 1-(2,3-
11 9 3
(C H N
added under nitrogen to a suspension of NaH (22 mg of a 60%
dispersion in oil, 0.57 mmol) in THF (2 mL). After 5 min,
methyl iodide (32 µL, 0.51 mmol) was added, and the solution
was stirred at 20 °C for 30 min and then partitioned between
EtOAc and water. Workup of the organic layer gave 1-methyl-
dihydroxypropyl)-5-(methylthio)pyrrolo[3,2-g]quinazoline (27)
1
(70 mg, 26%): mp (trituration with Me
NMR [(CD
2
CO) 198-200 °C; H
3
)
2
SO] δ 8.85 (s, 1 H), 8.36 (s, 1 H), 8.03 (s, 1 H),
7.75 (d, J ) 3.2 Hz, 1 H), 6.80 (d, J ) 3.2 Hz, 1 H), 5.06 (d, J
) 5.2 Hz, 1 H, OH), 4.86 (t, J ) 5.6 Hz, 1 H, OH), 4.44 (dd, J
) 14.4, 3.6 Hz, 1 H), 4.20 (dd, J ) 14.4, 7.3 Hz, 1 H), 3.86 (m,
5
-(methylthio)pyrrolo[3,2-g]quinazoline (22) (0.10 g, 94%): mp
1
13
(
8
EtOAc/petroleum ether) 198-200 °C; H NMR [(CD
3
)
2
SO] δ
1 H), 3.41 (m, 2 H), 2.71 (s, 3 H); C NMR δ 170.09 (s), 150.50
.86 (s, 1 H), 8.37 (s, 1 H), 7.97 (s, 1 H), 7.76 (d, J ) 3.3 Hz, 1
(d), 142.01 (s), 140.81 (s), 136.12 (d), 130.09 (s), 117.13 (s),
113.98 (d), 105.94 (d), 101.12 (d), 70.60 (d), 63.22 (t), 49.19 (t),
1
3
H), 6.80 (d, J ) 3.3 Hz, 1 H), 3.92 (s, 3 H), 2.71 (s, 3 H);
C
NMR δ 170.29 (s), 150.60 (d), 142.11 (s), 140.90 (s), 136.06 (d),
11.98 (q). Anal. (C14
H
15
N
3
O
2
2
S‚0.5H O) C, H, N, S.

Tyrosine Kinase Inhibitors
J ournal of Medicinal Chemistry, 1997, Vol. 40, No. 10 1527
A solution of diol 27 (0.15 g, 0.52 mmol), 3-bromoaniline
6 H), 1.93 (m, 2 H); 13C NMR δ 157.74 (s), 151.72 (d), 144.59
(s), 141.57 (s), 139.67 (s), 134.02 (d), 130.25 (d), 128.96 (s),
125.28 (d), 123.64 (d), 121.12 (s), 120.18 (d), 113.73 (d), 109.32
(s), 105.22 (d), 101.11 (d), 55.80 (t), 45.06 (q), 43.55 (t), 27.47
(
0.12 mL, 1.14 mmol), and concentrated HCl (51 µL, 0.57
mmol) in N-methylpyrrolidinone (8 mL) was warmed at 120
C for 15 h, then cooled, and poured into Et O (300 mL). The
resulting precipitate was filtered off, dissolved in hot propan-
-ol/water (1:1), and basified with concentrated NH The
°
2
5
(t). Anal. (C21H22BrN ) C, H, N.
2
3
.
5-[(3-Br om op h en yl)a m in o]-1-[2-(4-m or p h olin o)eth yl]-
1
resulting solution was concentrated to dryness, and the residue
was triturated with water, dissolved in EtOAc/MeOH (9:1),
and chromatographed on silica gel. Elution with EtOAc gave
foreruns, while EtOAc/MeOH (19:1) eluted 5-[(3-bromophenyl)-
amino]-1-(2,3-dihydroxypropyl)pyrrolo[3,2-g]quinazoline (3c)
as a yellow solid (84 mg, 39%). This was converted im-
mediately into the hydrochloride salt by dissolution in MeOH
p yr r olo[3,2-g]qu in a zolin e (3f) (19%): mp 114-116 °C; H
NMR (CDCl ) δ 8.74 (s, 1 H), 8.17 (s, 1 H), 8.14 (br s, 1 H),
3
7.87 (s, 1 H), 7.73 (m, 1 H), 7.58 (br, 1 H), 7.45 (d, J ) 3.4 Hz,
1 H), 7.28 (m, 2 H), 6.70 (d, J ) 3.4 Hz, 1 H), 4.34 (t, J ) 6.7
Hz, 2 H), 2.50 (t, J ) 4.6 Hz, 4 H); 13C NMR δ 157.49 (s), 152.66
(d), 145.02 (s), 140.86 (s), 140.13 (s), 133.50 (d), 130.26 (d),
129.82 (s), 126.85 (d), 124.20 (d), 122.65 (s), 119.77 (d), 111.38
(d), 109.66 (s), 105.97 (d), 101.67 (d), 66.95 (t), 57.67 (t), 53.82
(
1 mL), addition of concentrated HCl (1 drop), and cooling at
-
20 °C for 48 h. The salt was filtered off and washed well
(t), 44.31 (t). Anal. (C22
5-[(3-Br om oph en yl)am in o]-1-[3-(4-m or ph olin o)pr opyl]-
H
22BrN
5
2
‚0.5H O) C, H, N.
1
with Et
H), 11.44 (s, 1 H), 9.15 (s, 1 H), 8.94 (s, 1 H), 8.12 (t, J ) 1.9
Hz, 1 H), 7.97 (s, 1 H), 7.87 (d, J ) 3.4 Hz, 1 H), 7.81 (br d, J
8.2 Hz, 1 H), 7.53 (br d, J ) 8.1 Hz, 1 H), 7.48 (dd, J ) 8.2,
.1 Hz, 1 H), 6.92 (d, J ) 3.4 Hz, 1 H), 5.10 (br, 2 H, OH), 4.46
dd, J )14.5, 2.6 Hz, 1 H), 4.25 (dd, J ) 14.5, 7.2 Hz, 1 H),
2
O: mp 184-187 °C; H NMR [(CD
3
)
2
SO] δ 14.89 (br,
1
1
p yr r olo[3,2-g]qu in a zolin e (3g) (34%): mp 147-149 °C; H
NMR (CDCl ) δ 8.73 (s, 1 H), 8.17 (s, 1 H), 8.13 (br s, 1 H),
3
)
8
(
7.89 (s, 1 H), 7.72 (m, 1 H), 7.62 (br, 1 H), 7.39 (d, J ) 3.3 Hz,
1 H), 7.28 (m, 2 H), 6.67 (d, J ) 3.3 Hz, 1 H), 4.31 (t, J ) 6.7
Hz, 2 H), 3.72 (t, J ) 4.6 Hz, 4 H), 2.38 (t, J ) 4.6 Hz, 4 H),
1
3
3
6
1
.87 (m, 1 H), 3.40 (dd, J )11.0, 5.2 Hz, 1 H), 3.31 (dd, J )11.0,
2.27 (t, J ) 6.9 Hz, 2 H), 2.05 (dt, J ) 6.9, 6.7 Hz, 2 H);
C
1
3
.5 Hz, 1 H); C NMR δ 160.26 (s), 149.41 (d), 140.59 (s),
NMR δ 157.51 (s), 152.58 (d), 144.92 (s), 140.24 (s), 140.16 (s),
133.37 (d), 103.25 (d), 129.82 (s), 126.81 (d), 124.21 (d), 122.64
(s), 119.79 (d), 111.36 (d), 109.58 (s), 106.17 (d), 101.54 (d),
66.97 (t), 55.16 (t), 53.55 (t), 44.23 (t), 26.43 (t). Anal. (C H24-
38.71 (s), 136.60 (d), 130.62 (d), 130.04 (s), 128.80 (d), 126.99
(
d), 123.36 (d), 121.13 (s), 116.42 (d), 106.60 (s), 102.11 (d),
8.35 (d), 70.65 (d), 63.01 (t), 49.37 (t). Anal. (C19 17BrN4O
HCl) C, H, N, S.
9
H
2
‚
2
3
BrN O‚0.5H O) C, H, N.
5
2
Similar alkylation of the pyrazoloquinazoline 25 with
5-[(3-Br om op h en yl)a m in o]-1-[2-(N,N-d im eth yla m in o)-
1
-chloropropane-2,3-diol followed by reaction of the resultant
eth yl]p yr a zolo[4,3-g]qu in a zolin e (2d ) (12%): converted
crude thioether 28 with 3-bromoaniline afforded 5-[(3-bro-
mophenyl)amino]-1-(2,3-dihydroxypropyl)pyrazolo[4,3-g]quinazo-
into the hydrochloride salt and crystallized from MeOH/Et O;
2
1
mp 262-264 °C; H NMR [(CD ) SO] δ 10.89 (br, 1 H), 9.95
3
2
1
line (2c): mp (EtOAc) 175-180 °C dec; H NMR [(CD
3
)
2
SO] δ
(s, 1 H), 9.07 (s, 1 H), 8.65 (s, 1 H), 8.64 (s, 1 H), 8.37 (s, 1 H),
8.01 (d, J ) 8.4 Hz, 1 H), 7.36 (dd, J ) 8.4, 7.9 Hz, 1 H), 7.29
(br d, J ) 7.9 Hz, 1 H), 4.77 (t, J ) 6.7 Hz, 2 H), 3.46 (br, 2 H),
2.73 (s, 6 H); 13C NMR δ 158.08 (s), 152.46 (d), 149.20 (d),
145.39 (s), 142.97 (s), 141.26 (s), 138.14 (s), 130.29 (d), 125.63
(d), 123.86 (d), 121.12 (s), 120.41 (d), 112.78 (d), 111.12 (s),
106.68 (d), 54.98 (t), 42.91 (q), 39.96 (t). Anal. (C H BrN ‚-
1
1
1.36 (s, 1 H), 9.12 (s, 1 H), 8.61 (s, 1 H), 8.53 (d, J ) 0.9 Hz,
H), 8.32 (t, J ) 1.8 Hz, 1 H), 8.00 (d, J ) 0.9 Hz, 1 H), 7.95
(
d, J ) 8.0 Hz, 1 H), 7.40 (dd, J ) 8.0, 7.9 Hz, 1 H), 7.30 (d, J
7.9 Hz, 1 H), 5.05 (d, J ) 5.1 Hz, 1 H, OH), 4.89 (t, J ) 5.6
Hz, OH), 4.45 (dd, J ) 14.6, 2.5 Hz, 1 H), 4.25 (dd, J ) 14.6,
.1 Hz, 1 H), 3.87 (m, 1 H), 3.35 (m, 2 H). Anal. (C18
BrN ‚0.5H O) C, H.
-[(3-Br om op h en yl)a m in o]-1-[2-(N,N-d im eth yla m in o)-
)
7
16
H -
1
9
19
6
5
O
2
2
HCl‚H O) C, H, N, Cl.
2
5
5-[(3-Br om op h en yl)a m in o]-1-[2-(4-m or p h olin o)eth yl]-
p yr a zolo[4,3-g]qu in a zolin e (2f) (21%): converted into the
hydrochloride salt; mp (EtOAc/petroleum ether) 216-220 °C;
eth yl]p yr r olo[3,2-g]qu in a zolin e (3d ): Gen er a l Exa m p le
of Sch em e 3. A mixture of the pyrroloquinazoline 3a (0.10
g, 0.29 mmol), 2-(dimethylamino)ethyl chloride hydrochloride
1
H NMR [(CD ) SO] δ 10.07 (br s, 1 H), 9.15 (s, 1 H), 8.61 (s,
3 2
(
4
51 mg, 0.35 mmol), CsCO
A molecular sieves (0.20 g) in dry Me
under reflux for 6 h. Additional equivalents of alkylating
agent and CsCO were added, and the mixture was heated
3
(0.13 g, 0.41 mmol), and powdered
1 H), 8.53 (s, 1 H), 8.31 (br s, 1 H), 8.07 (s, 1 H), 7.96 (d, J )
7.3 Hz, 1 H), 7.33 (m, 2 H), 4.68 (t, J ) 6.4 Hz, 2 H), 3.50 (t,
J ) 4.2 Hz, 4 H), 2.87 (t, J ) 6.4 Hz, 2 H), 2.50 (t, J ) 4.2 Hz,
4 H); 13C NMR δ 153.81 (d), 141.23 (s), 134.59 (d), 130.61 (d),
130.35 (d), 125.96 (d), 124.20 (d), 123.69 (s), 121.13 (s), 120.72
(d), 117.70 (s), 116.66 (s), 115.78 (s), 112.63 (d), 104.65 (d),
2
CO (10 mL) was heated
3
under reflux overnight. After concentration under reduced
presssure, the residue was partitioned between EtOAc and
water and then filtered. The organic layer was worked up and
chromatographed on silica gel. EtOAc eluted unreacted 3a
+
65.95 (t), 57.02 (t), 53.08 (t), 45.71 (t); DEIMS found M
454.0965, 452.0969, C H BrN O requires 454.0940, 452.0960.
2
1
21
6
(
11 mg, 11%), while EtOAc/MeOH/concentrated NH
3
(9:1:trace)
Anal. (C H BrN O‚2HCl‚H O) C, H.
2
1
21
6
2
1
1
eluted material which was rechromatographed on alumina.
Elution with EtOAc/petroleum ether (2:3) gave pure 5-[(3-
bromophenyl)amino]-1-[2-(N,N-dimethylamino)ethyl]pyrrolo-
5-[(3-Br om op h en yl)a m in o]-3-[(N ,N ,N-tr im eth yleth yl-
en ed ia m in o)m eth yl]p yr r olo[3,2-g]qu in a zolin e (7): Gen -
er a l Exa m p le of Sch em e 4. A solution of pyrroloquinazoline
3a (0.14 g, 0.41 mmol), formaldehyde (24 µL of a 40% aqueous
solution, 0.45 mmol), and N,N,N′-trimethylethylenediamine
(57 µL, 0.45 mmol) was warmed at 50 °C for 1 h. A further
equivalent each of formaldehyde and amine was added, and
warming was continued for a further 2 h. One drop of
[
3,2-g]quinazoline (3d ) (23 mg, 19%): mp (EtOAc/petroleum
1
ether) 150-152 °C; H NMR (CDCl
s, 2 H), 7.43 (d, J ) 3.2 Hz, 1 H), 7.28 (m, 2 H), 6.68 (d, J )
.2 Hz, 1 H), 4.31 (t, J ) 6.9 Hz, 2 H), 2.76 (t, J ) 6.9 Hz, 2
3
) δ 8.72 (s, 1 H), 7.72 (br
3
1
3
H), 2.31 (s, 6 H); C NMR δ 152.58 (s), 146.32 (d), 140.20 (s),
33.83 (s), 133.45 (d), 130.25 (d), 129.84 (s), 126.81 (d), 124.22
d), 122.64 (s), 119.81 (d), 111.50 (d), 105.85 (d), 101.72 (d),
8.43 (t), 45.62 (q), 44.88 (t). Anal. (C18 ‚0.5H O)
C, H; N: found, 15.9; calcd, 16.5.
1
(
5
concentrated NH OH was added, and the solution was con-
4
centrated to dryness onto alumina and chromatographed
directly. EtOAc eluted a trace of starting material, while
MeOH/EtOAc (1:19) eluted 5-[(3-bromophenyl)amino]-3-
H
16BrN
5
O
2
2
1
1
The following 1-substituted pyrrolo- and pyrazoloquinazo-
lines were synthesized in a similar manner. Purification was
best achieved by sequential chromatography on silica gel and
alumina as described above followed by recrystallization from
EtOAc/petroleum ether.
[(N ,N ,N-trimethylethylenediamino)methyl]pyrrolo[3,2-g]-
quinazoline (7) as a yellow oil (77 mg, 41%). Dissolution in
EtOAc followed by precipitation with petroleum ether afforded
1
crystalline methanol solvate: mp 139-141 °C dec; H NMR
[(CD ) SO] δ 11.33 (s, 1 H), 9.81 (s, 1 H), 8.81 (s, 1 H), 8.51 (s,
3 2
5
-[(3-Br om op h en yl)a m in o]-1-[3-(N,N-d im eth yla m in o)-
1 H), 8.27 (t, J ) 1.8 Hz, 1 H), 7.98 (d, J ) 8.2 Hz, 1 H), 7.74
(s, 1 H), 7.58 (d, J ) 2.0 Hz, 1 H), 7.37 (dd, J ) 8.2, 8.1 Hz, 1
H), 7.28 (br d, J ) 8.1 Hz, 1 H), 3.82 (s, 2 H), 2.51 (t, J ) 7.6
Hz, 2 H), 2.40 (t, J ) 7.6 Hz, 2 H), 2.25 (s, 3 H), 2.11 (s, 6 H);
p r op yl]p yr r olo[3,2-g]qu in a zolin e (3e) (26%): mp 171-172
°
1
7
Hz, 1 H), 7.27 (br d, J ) 8.0 Hz, 1 H), 6.76 (d, J ) 3.3 Hz, 1
H), 4.33 (t, J ) 6.8 Hz, 2 H), 2.17 (t, J ) 6.9 Hz, 2 H), 2.12 (s,
1
C; H NMR [(CD
3
)
2
SO] δ 9.81 (s, 1 H), 8.83 (s, 1 H), 8.56 (s,
H), 8.36 (t, J ) 1.8 Hz, 1 H), 8.00 (br d, J ) 8.0 Hz, 1 H),
.89 (s, 1 H), 7.74 (d, J ) 3.3 Hz, 1 H), 7.36 (dd, J ) 8.0, 8.0
1
3
C NMR δ 157.78 (s), 151.46 (d), 144.62 (s), 141.51 (s), 140.26
(s), 130.24 (d), 129.44 (d), 129.10 (s), 125.39 (d), 123.95 (d),
121.08 (s), 120.54 (d), 112.08 (d), 112.00 (s), 109.00 (s), 106.53

1
528 J ournal of Medicinal Chemistry, 1997, Vol. 40, No. 10
Palmer et al.
(
d), 56.98 (t), 54.46 (t), 52.42 (t), 45.46 (q), 42.16 (q). Anal.
‚MeOH) C, H, N.
The following 3-substituted pyrrolo[3,2-g]quinazolines were
Deter m in a tion of Aqu eou s Solu bility. Stock solutions
of drugs were made up in methanol or DMSO and used to
calibrate the HPLC (peak area in nanomoles, assuming a
linear response). Accurately weighed amounts (to give ap-
proximately a 50 mM solution) were then sonicated for 30 min
in 0.05 M sodium lactate buffer, pH 4.0 (neutral compounds,
hydrochloride salts of amines), or in water (sodium salts of
acids). After standing for an additional 30 min, the samples
were centrifuged at 13 000 rpm for 3 min, and the concentra-
tion of drug in the supernatant was determined by HPLC,
using the calibration determined previously.
22 6
(C H25BrN
prepared in a similar manner by reaction of the parent
heterocycle with formaldehyde and the appropriate secondary
amine.
5
-[(3-Br om op h en yl)a m in o]-3-[(N,N-d im et h yla m in o)-
m eth yl]p yr r olo[3,2-g]qu in a zolin e (5) (34%): mp (EtOAc/
petroleum ether) 220 °C dec; H NMR [(CD ) SO] δ 11.35 (s, 1
3 2
H), 9.86 (s, 1 H), 8.83 (s, 1 H), 8.53 (s, 1 H), 8.28 (t, J ) 1.6
Hz, 1 H), 7.98 (br d, J ) 8.5 Hz, 1 H), 7.76 (s, 1 H), 7.59 (d, J
)
8
1
1
Molecu la r Mod elin g. The 3D model of EGFR TK was
constructed using the homology modeling module implemented
in LOOK25 with cAMP kinase as the template. The basis for
the 3D model was the sequence alignment published recently
2.0 Hz, 1 H), 7.37 (dd, J ) 8.5, 8.2 Hz, 1 H), 7.29 (br d, J )
.2 Hz, 1 H), 3.73 (s, 2 H), 2.23 (s, 6 H); 1 C NMR δ 157.81 (s),
3
51.49 (d), 144.66 (s), 141.50 (s), 140.34 (s), 140.23 (s), 130.21
(
d), 129.53 (d), 129.00 (s), 125.43 (d), 124.09 (d), 121.07 (s),
20.70 (d), 112.14 (d), 109.03 (s), 106.56 (d), 53.83 (t), 44.87
q). Anal. (C19 ) C, H, N.
-[(3-Br om op h en yl)a m in o]-3-[(4-m or p h olin o)m eth yl]-
p yr r olo[3,2-g]qu in a zolin e (6) (37%): mp (EtOAc/petroleum
2
6
by Knighton et al. The EGFR TK inhibitors were constructed
in SYBYL 6.2.27 The charges for the small molecule inhibitors
were derived from the semiempirical molecular orbital package
1
(
H18BrN
5
5
28
MOPAC using the MNDO approach. The inhibitors were
docked manually into the ATP binding site, and geometry was
optimized with the Tripos force field implemented in SYBYL
1
ether) 229 °C dec; H NMR [(CD
(
7
3
)
2
SO] δ 11.38 (s, 1 H), 9.82
s, 1 H), 8.84 (s, 1 H), 8.52 (s, 1 H), 8.27 (t, J ) 1.7 Hz, 1 H),
.98 (d, J ) 8.3 Hz, 1 H), 7.75 (s, 1 H), 7.61 (d, J ) 2.1 Hz, 1
6
.2 to relieve unfavorable steric contacts.
En zym e Assa y. Epidermal growth factor receptor was
H), 7.37 (dd, J ) 8.3, 8.3 Hz, 1 H), 7.29 (br d, J ) 8.3 Hz, 1 H),
.81 (s, 2 H), 3.59 (t, J ) 4.4 Hz, 4 H), 2.48 (t, J ) 4.4 Hz, 4
isolated from human A431 carcinoma cell shed membrane
vesicles by immunoaffinity chromatography as previously
3
1
3
H); C NMR δ 157.80 (s), 151.54 (d), 144.64 (s), 141.44 (s),
40.25 (s), 130.25 (d), 129.85 (d), 129.17 (s), 125.52 (d), 124.16
d), 121.10 (s), 120.78 (d), 111.94 (d), 110.83 (s), 109.08 (s),
06.62 (d), 66.17 (t), 53.02 (t), 52.83 (t). Anal. (C21 O)
C, H, N.
-[(3-Br om op h e n yl)a m in o]-3-[[N ,N -b is(2-h yd r oxy-
29
described. The enzyme assay was performed in 96-well filter
1
(
1
plates (Millipore MADVN6550). The total volume was 0.1 mL
containing 20 mM HEPES, pH 7.4, 50 µM sodium vanadate,
H20BrN
5
4
[
0 mM magnesium chloride, 10 µM ATP containing 0.5 µCi of
P]ATP, 20 µg of polyglutamic acid/tyrosine (Sigma Chemical
32
5
Co., St. Louis, MO), 1 ng of EGFR tyrosine kinase, and
appropriate dilutions of inhibitor. All components except the
ATP were added to the well, and the plate was incubated with
shaking for 10 min at 25 °C. The reaction was terminated by
addition of 0.1 mL of 20% trichloroacetic acid (TCA) and the
plate kept at 4 °C for at least 15 min to allow the substrate to
precipitate. The wells were then washed five times with 0.125
et h yl)a m in o]m et h yl]p yr r olo[3,2-g]qu in a zolin e (4). In
this case the product was retained strongly on alumina, and
purification was best achieved by chromatography on silica
gel, eluting with MeOH/EtOAc (1:9) to give a yellow solid
(
1
18%): mp (EtOAc) 240-245 °C dec; H NMR [(CD
3
)
2
SO] δ
1
1.30 (s, 1 H), 9.69 (s, 1 H), 8.84 (s, 1 H), 8.52 (s, 1 H), 8.30 (t,
J ) 1.9 Hz, 1 H), 7.97 (br d, J ) 7.7 Hz, 1 H), 7.73 (s, 1 H),
.63 (d, J ) 1.9 Hz, 1 H), 7.37 (dd, J ) 7.7, 7.6 Hz, 1 H), 7.29
32
mL of 10% TCA, and P incorporation was determined with
7
a Wallac beta plate counter. Control activity (no drug) gave
a count of approximately 100 000 cpm. At least two indepen-
dent dose-response curves were done and the IC50 values
computed. The reported values are averages; variation was
generally (15%.
(
2
1
br d, J ) 7.6 Hz, 1 H), 3.98 (s, 2 H), 3.53 (t, J ) 6.0 Hz, 4 H),
.65 (t, J ) 6.0 Hz, 4 H); 13C NMR δ 157.73 (s), 151.42 (d),
44.57 (s), 141.47 (s), 140.30 (s), 130.26 (d), 129.02 (d), 125.37
d), 123.78 (d), 121.13 (s), 120.41 (d), 112.72 (s), 112.04 (d),
(
1
08.85 (s), 106.43 (d), 59.14 (t), 56.18 (t), 49.72 (t); FABMS
EGF Recep tor Au top h osp h or yla tion in A431 Hu m a n
Ep id er m oid Ca r cin om a Cells. Cells were grown to conflu-
ency in 6-well plates (35 mm diameter) and exposed to serum-
free medium for 18 h. They were then treated with 8 for 2 h
and with EGF (100 ng/mL) for 5 min. The monolayers were
lyzed in 0.2 mL of boiling Laemlli buffer (2% sodium dodecyl
sulfate, 5% â-mercaptoethanol, 10% glycerol, and 50 mM Tris,
pH 6.8), and the lysates were heated to 100 °C for 5 min.
Proteins in the lysate were separated by polyacrylamide gel
electrophoresis and electrophoretically transferred to nitrocel-
lulose. The membrane was washed once in 10 mM Tris, pH
7.2, 150 mM NaCl, and 0.01% azide (TNA) and blocked
overnight in TNA containing 5% bovine serum albumin and
1% ovalbumin. The membrane was blotted for 2 h with
antiphosphotyrosine antibody (UBI, 1 mg/mL in blocking
buffer) and then washed twice in TNA, once in TNA containing
0.05% Tween-20 and 0.05% nonidet P-40, and twice in TNA.
The membranes were then incubated for 2 h in blocking buffer
containing 0.1 mCi/mL [1 I]protein A and then washed again
as above. After the blots were dry they were loaded into a
film cassette and exposed to X-AR X-ray film for 1-7 days.
Band intensities were determined with a Molecular Dynamics
laser densitometer.
+
found [M + H] 458.1027, 456.1017, C21
58.1015, 456.1035. Anal. (C21
-[(3-Br om op h en yl)a m in o]-3-[[N-(ca r b oxym et h yl)-N-
H
23BrN
5
O
2
requires
4
H
23BrN
5
O
2
‚H
2
O) C, H.
5
m eth yla m in o]m eth yl]p yr r olo[3,2-g]qu in a zolin e eth yl es-
ter (8) (37%): mp (ETOAc/petroleum ether) 227-230 °C dec;
1
3 2
H NMR [(CD ) SO] δ 11.35 (br s, 1 H), 9.78 (s, 1 H), 8.82 (s,
H), 8.53 (s, 1 H), 8.27 (br s, 1 H), 7.98 (d, J ) 8.0 Hz, 1 H),
.75 (s, 1 H), 7.59 (d, J ) 2.0 Hz, 1 H), 7.37 (dd, J ) 8.0, 8.0
1
7
Hz, 1 H), 7.28 (br d, J ) 8.0 Hz, 1 H), 4.07 (q, J ) 7.0 Hz, 2
H), 3.98 (s, 2 H), 3.34 (s, 3 H), 2.40 (s, 2 H), 1.16 (t, J ) 7.0
Hz, 3 H); C NMR δ 170.57 (s), 157.78 (s), 151.51 (d), 144.68
(
1
(
(
1
3
s), 141.46 (s), 140.33 (s), 130.22 (d), 129.66 (d), 128.91 (s),
25.41 (d), 123.98 (d), 121.07 (s), 120.57 (d), 112.12 (d), 111.61
s), 109.05 (s), 106.59 (d), 59.63 (t), 56.71 (t), 50.89 (t), 41.67
+
q), 14.01 (q); DEIMS found M 469.0917, 467.0929, C22
requires 469.0936, 467.0957.
22
H -
5 2
BrN O
A solution of the ester 8 (0.10 g, 0.22 mmol) in MeOH (10
mL) was stirred with 1 N NaOH (1 mL, 1.0 mmol) at room
temperature for 1 h. Careful acidification with 1 N HCl gave
25
5
-[(3-bromophenyl)amino]-3-[[N-(carboxymethyl)-N-methylami-
no]methyl]pyrrolo[3,2-g]quinazoline hydrochloride (9) (64 mg,
6
1
1
3%): mp (MeOH/Et
2 3 2
O) 240 °C dec; H NMR [(CD ) SO] δ
2.41 (d, J ) 1.9 Hz, 1 H), 11.91 (br, 1 H), 10.23 (s, 1 H), 8.97
Ack n ow led gm en t. This work was partially sup-
ported by the Auckland Division of the Cancer Society
of New Zealand.
(
s, 1 H), 8.30 (dd, J ) 1.9, 1.6 Hz, 1 H), 8.11 (d, J ) 2.4 Hz, 1
H), 8.07 (s, 1 H), 8.05 (br d, J ) 8.1 Hz, 1 H), 7.51 (dd, J ) 8.2,
.6 Hz, 1 H), 7.46 (dd, J ) 8.2, 8.1 Hz, 1 H), 4.73 (s, 2 H), 4.21
1
1
3
(s, 2 H), 2.95 (s, 3 H); C NMR δ 167.33 (s), 160.24 (s), 149.69
(d), 140.17, 138.92 (s), 135.57 (d), 130.46 (d), 129.85 (s), 128.54
(d), 126.59 (d), 123.09 (d), 121.08 (s), 119.47 (s), 116.19 (d),
Refer en ces
(
1) El-Zayat, A. A. E.; Pingree, T. F.; Mock, P. M.; Clark, G. M.;
Otto, R. A.; Von Hoff, D. D. Epidermal growth factor receptor
amplification in head and neck cancer. Cancer J . 1991, 4, 375-
380.
1
07.50 (s), 103.87 (s), 100.21 (d), 53.90 (t), 50.21 (t), 39.91 (q);
+
5 2
FABMS found [M + H] 442.0685, 440.0734, C20H19BrN O
requires 442.0702, 440.0722.

Tyrosine Kinase Inhibitors
J ournal of Medicinal Chemistry, 1997, Vol. 40, No. 10 1529
(
2) Morishige, K.; Kurachi, H.; Amemiya, K.; Fujita, Y.; Yamamoto,
T.; Miyake, A.; Tanizawa, O. Evidence for the involvement of
transforming growth factor R and epidermal growth factor
receptor autocrine growth mechanism in primary human ovarian
cancers in vitro. Cancer Res. 1991, 51, 5322-5328.
(15) Showalter, H. D. H.; Sun, L.; Sercel, A. D.; Winters, R. T.; Denny,
W. A.; Palmer, B. D. Concise syntheses of the novel 1H-pyrrolo-
[3,2-g]quinazoline ring system and its [2,3-f] angular isomer. J .
Org. Chem. 1996, 61, 1155-1158.
(
16) Thompson, A. M.; Bridges, A. J .; Fry, D. W.; Kraker, A. J .; Denny,
W. A. Tyrosine kinase inhibitors. 7. 7-Amino-4-(phenylamino)-
and 7-amino-4-[(phenylmethyl)amino]pyrido[4,3-d]pyrimidines;
a new class of inhibitors of the tyrosine kinase activity of the
epidermal growth factor receptor. J . Med. Chem. 1995, 38, 3780-
(
3) J ardines, L.; Weiss, M.; Fowble, B.; Greene, M. Neu (c-erbB-2/
HER2) and the epidermal growth factor receptor (EGFR) in
breast cancer. Pathobiology 1993, 61, 268-282.
(
4) Hickey, K.; Grehan, D.; Reid, I. M.; O’Brian, S.; Walsh, T. N.;
Hennessy, T. P. J . Expression of epidermal growth factor
receptor and proliferating cell nuclear antigen predicts response
of esophageal squamous cell carcinoma to chemoradiotherapy.
Cancer 1994, 74, 1693-1698.
3
788.
(
17) Zheng, J .; Trafny, E. A.; Knighton, D. R.; Xuong, N.-H.; Taylor,
S. S.; Ten Eyck, L. F.; Sowadsky, J . M. 2.2 Å refined crystal
structure of the catalytic subunit of cAMP-dependent protein
kinase complexed with MnATP and a peptide inhibitor. Acta
Crystallogr. 1993, D49, 362-365.
(
5) Delarue, J . C.; Terrier, P.; Terrierlacombe, M. J .; Mouriesse, H.;
Gotteland, M.; Maylevin, F. Combined overexpression of c-erbB-2
protein and epidermal growth factor receptor (EGF-r) could be
predictive of early and long-term outcome in human breast
cancer: a pilot study. Bull. Cancer 1994, 81, 1067-1077.
6) Fry, D. W.; Kraker, A. J .; McMichael, A.; Ambroso, L. A.; Nelson,
J . M.; Leopold, W. R.; Connors, R. W.; Bridges, A. J . A specific
inhibitor of the epidermal growth factor receptor tyrosine kinase.
Science 1994, 265, 1093-1095.
(18) De Bondt, H. L.; Rosenblatt, J .; J ancarik, J .; J ones, H. D.;
Morgan, D. O.; Kim, S. H. Crystal structure of cyclin-dependent
kinase 2. Nature 1993, 363, 595-602.
(
(
(
19) Zang, F.; Strand, A.; Robbins, D.; Cobb, M. H.; Goldsmith, E. J .
Atomic structure of the MAP kinase ERK2 at 2.3 A resolution.
Nature 1994, 367, 704-711.
7) Ward, W. H. J .; Cook, P. N.; Slater, A. M.; Davies, D. H.;
Holdgate, G. A.; Green, L. R. Epidermal growth factor receptor
tyrosine kinase. Investigation of catalytic mechanism, structure-
based searching and discovery of a potent inhibitor. Biochem.
Pharmacol. 1994, 48, 659-666.
(
20) Owen, D. J .; Noble, M. E. M.; Garman, E. F.; Papageorgiou, A.
C.; J ohnson, L. N. Two structures of the catalytic domain of
phosphorylase kinase: an active protein kinase complexed with
substrate analogue and product. Structure 1995, 3, 467-482.
21) Schulze-Gahmen, U.; Brandsen, J .; J ones, H. D.; Morgan, D. O.;
Meijer, L.; Vesely, J .; Kim, S.-H. Multiple modes of ligand recog-
nition: crysral structures of cyclin-dependent protein kinase 2
in complex with ATP and two inhibitors, olomucine and isopen-
tyladenine. Protein 1995, 22, 378-391.
(
(
8) Rewcastle, G. W.; Denny, W. A.; Bridges, A. J .; Zhou, H.; Cody,
D. R.; McMichael, A.; Fry, D. W. Tyrosine kinase inhibitors. 5.
Synthesis and structure-activity relationships for 4-[(phenyl-
methyl)amino]- and 4-(phenylamino)quinazolines as potent ade-
onosine-5′-triphosphate binding site inhibitors of the tyrosine
kinase domain of the epidermal growth factor receptor. J . Med.
Chem. 1995, 38, 3482-3487.
(22) Chen, H.; Bioziau, J .; Parker, F.; Maroun, R.; Tocque, B.; Roques,
B.; Garbay-J aureguiberry, C. Synthesis and structure-activity
studies of a series of [(hydroxybenzyl)amino]salicylates as inhibi-
tors of EGF receptor-associated tyrosione kinase activity. J . Med.
Chem. 1993, 36, 4094-4098.
(
9) Bridges, A. J .; Zhou, H.; Cody, D. R.; Rewcastle, G. W.; Mc-
Michael, A.; Showalter, H. D. H.; Fry, D. W.; Kraker, A. J .;
Denny, W. A. Tyrosine kinase inhibitors. 8. An unusually steep
structure activity relationship for analogues of 4-(3-bromo-
anilino)-6,7-dimethoxyquinazoline (PD 153035), a potent inhibi-
tor of the epidermal growth factor receptor. J . Med. Chem. 1996,
(
23) Furet, P.; Caravatti, G.; Lydon, N.; Priestle, J . P.; Sowadski, J .
M.; Trinks, U.; Traxler, P. Modelling study of protein kinase
inhibitors - binding mode of staurosporine and origin of the
selectivity of CGP 52411. J . Comput.-Aided Mol. Des. 1995, 9,
3
9, 267-276.
(
10) Rewcastle, G. W.; Palmer, B. D.; Bridges, A. J .; Showalter, H.
D. H.; Nelson, J .; McPherson, A.; Kraker, A. J .; Fry, D. W.;
Denny, W. A. Tyrosine kinase inhibitors. 9. Synthesis and
evaluation of fused tricyclic quinazoline analogues as ATP site
inhibitors of the tyrosine kinase activity of the epidermal growth
factor receptor. J . Med. Chem. 1996, 39, 918-928.
4
65-472.
(
24) Traxler, P. M.; Furet, P.; Mett, H.; Buchdunger, E.; Meyer, T.;
Lydon, N. 4-(Phenylamino)pyrrolopyrimidines: potent and selec-
tive, ATP site directed inhibitors of the EGF-receptor protein
tyrosine kinase. J . Med. Chem. 1996, 39, 2285-2292.
(
11) Rewcastle, G. W.; Palmer, B. D.; Thompson, A. M.; Bridges, A.
J .; Fry, D. J .; Kraker, A. J .; Denny, W. A. Tyrosine kinase
inhibitors. 10. Isomeric 4-[(3-bromophenyl)amino]pyrido[d]py-
rimidines are potent ATP binding site inhibitors of the tyrosine
kinase function of the epidermal growth factor receptor. J . Med.
Chem. 1996, 39, 1823-1835.
(25) LOOK, Molecular Applications Group, Suite T, 445 Sherman
Ave, Palo Alto, CA 94306.
(
26) Knighton, D. R.; Cadena, D. L.; Zheng, J .; Ten Eyck, L. F.;
Taylor, S. S.; Sowadski, J . M.; Gill, G. N. Structural features
that specify tyrosine kinase activity deduced from homology
modeling of the epidermal growth factor receptor. Proc. Natl.
Acad. Sci. 1993, 90, 5001-5005.
(
12) Barker, A. J . Tricyclic derivatives and their use as anti-cancer
agents. Patent Application EPO 0 635 507 A1 to Zeneca Ltd.,
J an. 21, 1995; Chem. Abstr. 1995, 122, 31545y.
(27) SYBYL molecular modeling software, Tripos E & S, 6548 Clayton
Rd, St. Louis, MO 62117.
(
13) Himmelsbach, F.; von Ruden, T.; Dahmann, G.; Metz, T.
Pyrimido[5,4-d]pyrimidines, drugs containing these compounds,
their use, and process for preparing them. Patent Application
PCT/EP95/03482 to Dr. Karl Thomae, GMBH, Sept. 5, 1995;
Chem. Abstr. 1996, 124, 303748, 343326.
(
28) Dewar, M. J . S.; Thiel, W. Ground states of molecules. 38. The
MNDO method. Approximation and parameters. J . Am. Chem.
Soc. 1977, 99, 4907.
(29) Gill, G. N.; Weber, W. Purification of functionally active epider-
mal growth factor receptor protein using a competitive antago-
nist monoclonal antibody and competitive elution with epidermal
growth factor. Methods Enzymol. 1987, 146, 82-88.
(
14) Hudson, A. T.; Vile, S.; Barraclough, P.; Franzmann, K. W.;
McKeown, S. C.; Page, M. J . Substituted heteroaromatic com-
pounds and their use in medicine. Patent Application PCT/GB95/
0
2202 to The Wellcome Foundation, Sept. 18, 1995; Chem. Abstr.
996, 125, 114665.
1
J M960789H