Development of natural product-derived receptor tyrosine kinase inhibitors based on conservation of protein domain fold

Article abstract of DOI:10.1021/jm0307943

Receptor tyrosine kinases (RTKs) such as Tie-2, IGF1R, Her-2/Neu, EGFR, and VEGFR1-3 play crucial roles in the control of cell growth and differentiation. Inhibition of such RTKs has become a major focus of current anticancer drug development, and therefore the discovery of new classes of inhibitors for these signal-transducing proteins is of prime importance. We have recently proposed a novel concept for improving the hit-finding process by employing natural products as biologically validated starting points in structural space for compound library development. In this concept, natural products are regarded as evolutionary chosen ligands for protein domains which are structurally conserved yet genetically mobile. Here we report on the discovery of novel and highly selective VEGFR-2 and -3, Tie-2, and IGF1R inhibitors derived from the naturally occurring Her-2/Neu kinase inhibitor nakijiquinone C and developed on the basis of this concept. Based on the structure of the natural product, a small library (74 members) was synthesized and investigated for inhibition of kinases with highly similar ATP-binding domains. The library yielded inhibitors with IC₅₀s in the low micromolar range with high frequency (7 out of 74). In particular, four inhibitors of Tie-2 were found, a kinase critically involved in the formation of new blood vessels from preexisting ones (angiogenesis) and believed to be a new promising target in antitumor therapy. These results support the "domain concept". To advance the development of improved inhibitors, extensive molecular modeling studies were undertaken, including the construction of new homology models for VEGFR-2 and Tie-2. These studies revealed residues in the kinase structure which are crucial to the development of tailor-made receptor tyrosine kinase inhibitors.

Full text of DOI:10.1021/jm0307943

J . Med. Chem. 2003, 46, 2917-2931
2917
Develop m en t of Na tu r a l P r od u ct-Der ived Recep tor Tyr osin e Kin a se In h ibitor s
Ba sed on Con ser va tion of P r otein Dom a in F old
Lars Kissau,† Petra Stahl,†,§ Ralph Mazitschek,‡ Athannasios Giannis,*,‡ and Herbert Waldmann*,†
Max-Planck-Institut fu¨r molekulare Physiologie, Abteilung Chemische Biologie, Otto-Hahn-Strasse 11, 44227 Dortmund,
Germany and Universita¨t Dortmund, Fachbereich 3, Organische Chemie, 44221 Dortmund, Germany, and Institut fu¨r
Organische Chemie, Universita¨t Leipzig, J ohannisallee 29, 04103 Leipzig, Germany
Received J anuary 31, 2003
Receptor tyrosine kinases (RTKs) such as Tie-2, IGF1R, Her-2/Neu, EGFR, and VEGFR1-3
play crucial roles in the control of cell growth and differentiation. Inhibition of such RTKs has
become a major focus of current anticancer drug development, and therefore the discovery of
new classes of inhibitors for these signal-transducing proteins is of prime importance. We have
recently proposed a novel concept for improving the hit-finding process by employing natural
products as biologically validated starting points in structural space for compound library
development. In this concept, natural products are regarded as evolutionary chosen ligands
for protein domains which are structurally conserved yet genetically mobile. Here we report
on the discovery of novel and highly selective VEGFR-2 and -3, Tie-2, and IGF1R inhibitors
derived from the naturally occurring Her-2/Neu kinase inhibitor nakijiquinone C and developed
on the basis of this concept. Based on the structure of the natural product, a small library (74
members) was synthesized and investigated for inhibition of kinases with highly similar ATP-
binding domains. The library yielded inhibitors with IC50s in the low micromolar range with
high frequency (7 out of 74). In particular, four inhibitors of Tie-2 were found, a kinase critically
involved in the formation of new blood vessels from preexisting ones (angiogenesis) and believed
to be a new promising target in antitumor therapy. These results support the “domain concept”.
To advance the development of improved inhibitors, extensive molecular modeling studies were
undertaken, including the construction of new homology models for VEGFR-2 and Tie-2. These
studies revealed residues in the kinase structure which are crucial to the development of tailor-
made receptor tyrosine kinase inhibitors.
In tr od u ction
well as in numerous tumors of epithelial origin including
breast and oesophageal tumors.12
Receptor tyrosine kinases play crucial roles in the
transduction of signals across the plasma membrane.1
For instance, angiogenesis, the development of new
blood vessels from preexisting ones, depends on endo-
thelium specific receptor tyrosine kinases, in particular
the vascular endothelial growth factor receptors
(VEGFR1-3) and the Tie-2 receptor.2 All VEGFRs and
Tie-2 have been implicated in tumor angiogenesis,3-7
and aberrant angiogenesis is considered as a key step
in tumor growth, invasion, and metastasis.8,9
Given their biological relevance, RTKs are among the
most important drug targets, and several RTK inhibi-
tors are in advanced stages of clinical investigation or
have already reached the market.13,14 Therefore, the
development of new classes of RTK inhibitors which
may be employed as tools in biological studies or as new
guiding structures for the development of drug candi-
dates is of paramount importance to clinical biology and
medicinal chemistry.
We have advertised a novel concept for enhancing the
efficiency of the hit-finding process.15 Within this con-
cept, biologically active natural products that bind to
structurally conserved yet genetically mobile protein
domains are regarded as evolutionary selected and
biologically validated starting points in structural space
for compound library development. Because of their
biological prevalidation and the structural conservatism
employed by Nature in the repeated use of a limited
number of protein domains with very similar structure,
compound libraries designed on the basis of a given
natural product should yield better hits at a higher rate
than classical compound libraries designed on the basis
of chemical accessibility alone.
The insulin-like growth-factor 1 receptor (IGF1R)
exerts mitogenic, cell survival, and insulin-like activi-
ties, is involved in postnatal growth physiology, and is
connected to proliferative disorders such as breast
cancer.10 The Her-2/Neu protooncogene belongs to the
EGF-family of receptor tyrosine kinases. It is overex-
pressed in approximately 30% of all primary breast,
ovary, and stomach cancers.11 The EGFR (epidermal
growth factor receptor, ErbB-1) has been implicated in
human tumorigenesis, for example, of glioblastoma as
* Corresponding authors. H.W.: Phone: 49-231-133-2400, Fax: 49-
231-133-2499. E-mail: herbert.waldmann@mpi-dortmund.mpg.de.
A.G: E-mail: giannis@chemie.uni-leipzig.de.
† Max-Planck-Institut fu¨r Molekulare Physiologie and Universita¨t
Dortmund.
In this paper we provide proof of this concept,
employing the naturally occurring receptor tyrosine
kinase inhibitor nakijiquinone C 1c as a representative
example.16 Nakijiquinones C and D were isolated by
‡ Universita¨t Leipzig.
§ Present address: Aventis Pharma AG, Chemical Research, Build-
ing G 838, 65926 Frankfurt am Main, Germany.
10.1021/jm0307943 CCC: $25.00 © 2003 American Chemical Society
Published on Web 06/05/2003

2918 J ournal of Medicinal Chemistry, 2003, Vol. 46, No. 14
Kissau et al.
Sch em e 1. Modular Composition of the Nakijiquinone
Library
to the ATP binding site.19 Also the quinone-type building
block was varied which should contribute essential
hydrogen bonds to the hinge region of receptor tyrosine
kinases and therefore serve to orient the inhibitor
within the binding site. Finally, size and stereochem-
istry of the amino acid were changed, which both might
enhance binding and improve selectivity via formation
of hydrogen bonds and side chain interaction with the
protein. Consequently, in the second series of library
compounds the diterpene moiety was substituted with
simple hydrophobic structures such as benzyl, n-butyl,
and cyclohexyl groups or a trans-decalin. For the amino
acid, the hydrophilic serine and D- and L-threonine, the
hydrophobic valine, and glycine were considered. To
expand the SAR data, the type and number of H-bond
donors and acceptors in each library member were
varied, and thus the compounds contained either one
or two amino acid, hydroxyl, or methoxy groups or
combinations thereof. The core p-quinone was left
unchanged.
Commercially available guaiacol 8 was protected with
2-bromopropane and then brominated to afford com-
pound 1020 (Scheme 2). Lithiation using n-BuLi and
subsequent reaction with benzyl bromide yielded C-
benzylated aromatic compound 11, from which the
isopropyl ether was selectively removed to give 12. Aside
from the desired lithiation and reaction with benzyl
bromide, the addition of the n-butyl group (13 and 14)
from n-BuLi was encountered. The reaction with benzyl
bromide was carried out in the presence of CuI to form
the cuprate. Without the addition of this copper salt,
the desired coupling product was not formed.
The side reaction with n-BuLi could be suppressed
by the use of sec-BuLi, but the yield of the desired
benzyl-substituted quinone was not improved. Com-
pounds 11 and 13 were not separated at this stage but
immediately subjected to deprotection, employing BCl3
to yield 12 and 14, which could be separated. Oxidation
with oxygen and salcomin21 as a catalyst converted
these phenols to the corresponding quinones 15 and 16.
For the synthesis of compounds containing a cyclo-
hexylmethyl substituent on the quinone, a different
procedure was followed since a substitution protocol
analogous to the procedure used for the benzyl ana-
logues failed. Attempts to generate cyclohexylvinyl
bromide and subject it to a Suzuki coupling with boronic
acid 2422 resulted in an unsatisfactory yield of 15%,
probably due to the instability of the vinyl bromide.
Thus, aryl bromide 10 was lithiated and treated with
cyclohexanal 17. The resulting alcohol 18 was converted
into the corresponding oxalic acid ester 19 and deoxy-
genated using tributyltin hydride to give 20.23 After
deprotection, oxidation with salcomin in DMF converted
F igu r e 1. Structures of the nakijiquinones and structurally
closely related analogues.
J . Kobayashi et al. from a marine sponge and subse-
quently shown to exhibit inhibitory activity against
EGFR, c-erbB2, and PKC as well as cytotoxic activity
against L1210 and KB tumor cell lines.17 Notably,
nakijiquinone C is the first naturally occurring inhibitor
of the Her-2/Neu receptor tyrosine kinase.
Resu lts
For the synthesis of a nakijiquinone analogue library,
initially a solid-phase approach was considered. Attach-
ment to the solid support via the amino acid part of the
molecule appeared to be advantageous for variation of
the diterpene, the quinoid, and the amino acid parts of
the parent compound. Unfortunately, the bond between
the amino acid and the quinone is very base labile, a
fact learned from initial failures encountered in the
synthesis of the natural product itself.18 Thus, we
proceeded to construct a library via solution-phase
synthesis. The general design of the library of nak-
ijiquinones analogues was guided by the modular com-
position of these natural products.
In the first series of compounds, analogues were
prepared whose structure closely resembles the nak-
ijiquinones (Figure 1). Thus, the decalin core was not
changed entirely, but modified and analogues with
altered configuration at C-2 and with an exocyclic
double bond instead of an endocyclic one were synthe-
sized. Also, different amino acids were introduced, or
omitted, and finally the p-quinoid system was replaced
by an o-quinoid structure or an aromatic group. Rep-
resentative members of this library which were subse-
quently evaluated for biological activity are shown in
Figure 1. In total 18 compounds (including nakijiquino-
nes C and D) were prepared.18
In the second series of analogues (Scheme 1), the
deviation from the natural product was more pro-
nounced. In particular, variations were carried out on
the hydrophobic diterpene unit, which according to the
accepted binding mode of kinase inhibitors to the ATP-
binding domain may occupy a hydrophobic pocket close

Receptor Tyrosine Kinase Inhibitors
J ournal of Medicinal Chemistry, 2003, Vol. 46, No. 14 2919
Sch em e 2. Synthesis of Nakijiquinone Analogues with
Cyclohexyl, n-Butyl, or Benzyl Substituentsa
Sch em e 3. Synthesis of Nakijiquinone Analogues with
trans-Decalin Subsitutenta
a
Reagents and conditions: (a) Ph3PCH2Br, THF, KOtBu; (b)
Pd(PPh3)4, 24, EtOH, Na2CO3; (c) BCl3, CH2Cl2; (d) MeOH, THF,
Rh/Al2O3, H2, to give 28 (not shown) then BCl3, CH2Cl2; (e)
salcomine, O2, DMF.
Sch em e 4. Synthesis of Amino Acid-Substituted
Alkylquinonesa
a
Reagents and conditions: (a) i-PrBr Yields 9 (not shown); (b)
Br2, see ref 19; (c) n-BuLi, THF -78 °C, CuI, R1Br to give 11 and
13 (not shown); (d) BCl3, CH2Cl2; (e) salcomine, O2, DMF; (f) 10,
n-BuLi, THF, TMEDA, then 17; (h) iPr2Net, CH2Cl2, methyl oxalyl
chloride, DMAP; (h) toluene, AIBN, n-Bu3SnH, reflux; (i) BCl3,
CH2Cl2; (j) salcomine, O2, DMF.
20 into quinone 21 (Scheme 2). The deoxygenation of
benzyl alcohol 18 was also feasible via conversion to the
corresponding thioester followed by radical deoxygen-
ation with tributyltin hydride and AIBN as initiator.24
a
Reagents and conditions: (a) amino acid, NaHCO3, EtOH.
Finally, a trans-decalin was attached to the quinone
core either via a CH2 or CH bridge. Applying the
established chemistry, these analogues were synthe-
sized using again 2-hydroxy-3-methoxyphenyl bromide
9 as the starting material. Intermediate 10 was con-
verted to boronic acid 24 which was subjected to Suzuki
coupling with vinyl bromide 23 to yield olefin 25
(Scheme 3). After deprotection to give 26 oxidation
yielded quinone 27. Initial plans to synthesize 23 from
trans-1-decalone using a Takai reaction25 failed, and a
Wittig reaction was employed instead. Thus, 23 was
obtained in 63% yield from trans-1-decalone 22. Only
pressure, employing a rhodium (5% supported on alu-
mina) catalyst.27
An inseparable mixture of diastereomers of 28 (3:1,
determined via GC-analysis) was obtained. It was
deprotected with BCl3 to give phenol 29 (mixture of
diastereomers) which was then oxidized to afford quino-
ne 30.
The 2-methoxy-6-alkylquinones 15, 16, 21, 27, and
30 were converted to the amino acid-substituted deri-
vates by stirring a solution of the quinone in ethanol in
the presence of 4 equiv of NaHCO3 and the amino acid
to yield the target compounds (Scheme 4). Thus, com-
pounds 31a -e and 32a -e were obtained by treatment
of 15 and 16 with the corresponding amino acids.
Following the same synthesis protocol, 33a -e were
synthesized from 21, while 27 served as starting mate-
rial for the synthesis of 34a -e. Finally, 35a -e were
synthesized employing 30 as starting material (see
Scheme 4).
1
the (E)-alkene could be detected by H and 13C NMR
spectroscopy. The stereochemical assignment is based
on data recorded for compounds of similar structure, for
which the protons of the (Z)- and (E)-alkene differ
significantly in their chemical shifts.26 To obtain the
methylene-bridged analogues, intermediate 25 was
hydrogenated at room temperature and atmospheric

2920 J ournal of Medicinal Chemistry, 2003, Vol. 46, No. 14
Kissau et al.
Ta ble 1. IC50 Values (µM) of Compounds Displaying Activity toward One or More Kinasesa
compound
ErbB2
EGF
IGF1R
VEGFR-2
VEGFR-3
Tie-2
5
b
b
b
b
b
b
b
b
20 ( 3.0
b
b
b
b
b
15
27
36
37
38b
59c
b
18 ( 2.7
14 ( 2.1
b
>100
>100
b
b
44 ( 6.6
> 100
b
b
29 ( 4.4
>100
>50
6 ( 0.9
0.5 ( 0.015
8 ( 1.2
b
>50
b
b
b
b
3 ( 0.5
b
5 ( 0.8
9 ( 1.4
a
IC50 values were only determined for compounds displaying promising activity in the initial screening. b Compound showed no inhibition
in an initial screen and was not investigated. Each value represents the mean standard deviation of four independent reading points.
Sch em e 5. Synthesis of Nakijiqiuinone Analogues,
after hydrogenation with Rh/Al2O3, compound 52 was
obtained. Deprotection of 45, 50, and 52 was achieved
with BCl3 in methylene chloride to furnish 46, 52, and
53. However, the oxidation of these phenols with sal-
comin proceeded with only moderate yields. We there-
fore evaluated other oxidation methods and, finally,
found that the use of Fremy’s salt (ON(SO3K)2) in the
presence of a phase-transfer catalyst was superior.29 The
yield over two steps (deprotection of the isopropyl ether
and oxidation) was improved to 82% in the case of
compound 47. The dimethoxyalkyl-p-quinones 47, 54,
and 57 derived from 46, 52, and 53 via oxidation were
treated with KOH in methanol/water to saponify one
of the vinylogous esters. This procedure had already
proven successful in the total synthesis of nakijiquinone
C. Finally, compounds 48, 55, and 58 were reacted with
the desired amino acids to yield the target compounds
49a -e, 56a -e, and 59a -c, respectively.
Lacking an Alkyl Substituenta
a
Reagents and conditions: (a) salcomine, O2, DMF; (b) L-Ser,
EtOH, NaHCO3; (c) amino acid, EtOH, NaHCO3.
Biologica l Eva lu a tion of th e Libr a r y
The introduction of the amino acid at the C-2 position
probably proceeds via an addition/elimination mecha-
nism, while attachment of the second amino acid
involves a 1,4 addition at C-6 followed by an oxidation.
Since these reactions were carried out under an argon
atmosphere, the quinone itself is believed to be the
oxidant. This also explains the moderate yields deter-
mined for some of these conversions. In most cases, the
diastereomers resulting from the reduction of the CdC
double bond with Rh/Al2O3 could not be separated.
However, the diastereomers of compound 35b could be
separated via reversed phase HPLC. Analogues lacking
an alkyl substituent on the quinone were synthesized
starting from commercially available guaiacol 8 (Scheme
5). The reaction sequence of oxidation to 36 and subse-
quent conversion to 37 and 38 a -c followed the estab-
lished protocol of stirring the quinone with the desired
amino acid in ethanol in the presence of excess NaHCO3.
For analogues of the natural products carrying one
amino acid substituent and a hydroxyl group on the
quinone, a different synthetic route was applied (Scheme
6). The commercially available catechol 39 was mono-
protected as isopropyl ether to give 40. After bromina-
tion of 40 ortho to the hydroxyl group to yield 41,28
oxidation to the p-quinone 42 was carried out with
salcomin and oxygen.
To obtain a preliminary picture of the biochemical
properties of the members of the nakijiquinone analogue
library, the ability of the analogues to inhibit several
different receptor tyrosine kinases were investigated.
On the basis of the known biological activity of the
parent nakijiquinones (see above), VEGFR-2 (KDR, flk-
1), VEGFR-3 (flt-4), Tie-2, Her-2/Neu, EGFR (ErbB-1),
ErbB-2, and IGF1R were chosen.
In the assay, the kinase-catalyzed phosphorylation of
poly(Glu-Tyr) in the presence of varying concentrations
of inhibitor was determined. The kinases were employed
as fusion proteins of glutathione-S-transferase (GST)
and the respective kinase domain. Kinase activity was
determined by means of an anti-phosphotyrosine anti-
body conjugated to horseradish peroxidase (POD). The
chemiluminescence caused by the reaction catalyzed by
POD immobilized after antibody binding to phosphoty-
rosine residues was measured. Most of the library
members showed no activity in this initial screen. Those
compounds with noticeable inhibition, shown in Table
1, were then reevaluated in a second screening and the
IC50 values were determined.
Of the 74 compounds subjected to the kinase assay
seven compounds displayed inhibitory activity in the low
micromolar range. The library of the nakijiquinone
analogues did not embody an inhibitor of Her-2/Neu,
the EGFR, and ErbB-2 that warranted further detailed
investigation. Most remarkably, however, four of the
compounds investigated turned out to be inhibitors of
the Tie-2 receptor (Table 1).
Reduction with NaBH4 and subsequent hydroxyl
protection as methyl ether (43) was followed either by
lithiation or by conversion to the boronic acid 44 (see
Experimental Section) which was converted to 45 fol-
lowing the methodology described above. Using the
method described above, 45 was then reacted with vinyl
bromide 23 to give the unsaturated compound 50, and
Two of these (compounds 15 and 27) proved to be
selective for this kinase, i.e., the other kinases in the
assay were not inhibited more than 50% at concentra-

Receptor Tyrosine Kinase Inhibitors
J ournal of Medicinal Chemistry, 2003, Vol. 46, No. 14 2921
Sch em e 6. Synthesis of Analogues with One Amino Acid and an OH Group on the Quinoid Corea
a
Reagents and conditions: (a) i-PrBr, KOH, DMSO; (b) Br2,tert-butylamine, toluene; (c) salcomine, O2, DMF; (d) NaBH4, EtOH, then
(CH3O)2SO2, EtOH, H2O, NaHCO3; (e) n-BuLi, THF, BnBr, CuI; (f) BCl3, CH2Cl2; (g) n-BuLi, B(OMe)3, to yield 44 (not shown) then 23,
Pd(PPh3)4, Na2CO3; (h) ON(SO3K)2 Na2CO3, Aliquat 336, C6H6, H2O; (i) Rh/Al2O3, H2, MeOH, THF, yields 52 (not shown); (j) KOH, MeOH,
H2O; (k) amino acid, NaHCO3, EtOH.
tions up to 70 µM. The most potent inhibitor identified
in this study is 38b. Not only does it inhibit the Tie-2
receptor with an IC50 value of 5 µM, but it is also
similarly efficient in blocking the VEGFR-3 and the
IGF1R. Kinetic studies revealed, that both 38b and 59c
are ATP competitive inhibitors of the receptor tyrosine
kinases evaluated.
not well suited as a basis for computer-aided ligand
design. RTKs dimerize upon binding an extracellular
effector followed by the phosphorylation of residues in
the activation loop.1 Since the downstream effect of the
RTKs is exerted via their kinase domain, homology
models of the complete kinase domains of Tie-2 and
VEGFR-2 were constructed to be used in the analysis
of the biological activity of the receptor tyrosine kinase
inhibitors identified in this study.30
Molecu la r Mod elin g Stu d ies
The receptor tyrosine kinases attracting the most
interest in this study as a consequence of the biological
assays were Tie-2, IGF1R, VEGFR2, and VEGFR-3.
Crystal structures of the kinase domains of the first
three proteins have been determined; however, some of
the structures show the kinases in an unliganded form.
Since RTKs undergo significant structural changes upon
ligand binding,1 crystal structures without ligands are
The crystal structures of the FGF1 receptor31 were
chosen as templates on the basis of the high sequence
homology within the kinase domain of FGF1R with
Tie-2 (∼45%) and VEGFR-2 (∼42%). These values are
well above the 30% usually considered to be the limit
for a good homology model.32 Aside from the high
sequence homology, these structures contain the RTK
in complex with a ligand, i.e., in an activated or closed

2922 J ournal of Medicinal Chemistry, 2003, Vol. 46, No. 14
Kissau et al.
choice of model depending to a certain extent on the lead
structure used. Our work further suggests that the exact
conformation of the nucleotide binding loop of Tie-2,
VEGFR-2, and IGF1R depends partially on the shape
and size of the bound inhibitor, a result similar to the
phenomenon described for the FGF1 receptor in complex
with different inhibitors. To design high affinity ligands,
it is advantageous to build different models and evaluate
the interaction of a given inhibitor with each structure.
Parallel to the variability of the nucleotide binding loop,
the activation loop, another highly flexible region, must
be scrutinized as well. Especially the position of the
amino acids in the vicinity of the conserved DFG motif
may be crucial for selective ligand binding. As evidenced
by an overlay of several kinase structures, the exact
position of the so-called C-helix varies considerably. This
flexibility must also be taken into account when ratio-
nalizing experimental results or designing novel inhibi-
tors.
F igu r e 2. Overlay of the crystal structure of the kinase
domain of insulin receptor (red) with homology models of the
kinase domains for VEGFR-2 (black) and Tie-2 (yellow).35
form. For model development it was assumed, that even
though the Tie-2 crystal structure shows some signifi-
cant structural differences to other RTKs in their
unliganded forms, the conformation of the activated
Tie-2 RTK, especially the ATP binding site, would be
highly similar to other activated RTKs, given that the
mode of ATP binding is nearly identical in these
enzymes (Figure 2).33
The importance of the “induced fit” concept and the
limitations of the rigid receptor hypothesis have been
discussed repeatedly.34 M. Mohammadi et al. have
demonstrated, that certain inhibitors cause significant
changes in the conformations of loops in the kinase
domain.35 The two crystal structures of the FGF1
receptor containing two different types of inhibitors
demonstrate the applicability of the induced fit concept
to this class of enzymes. During our analysis of the
nakijiquinone-type RTK inhibitors, it was found that
applying this knowledge to our modeling studies re-
sulted in improved binding modes and superior expla-
nations for the experimental SAR data.16
The amino acid sequences for Tie-2 and VEGFR-2
were retrieved from the Swissprot Database. The pri-
mary accession numbers are P35968 (VEGFR-2) and
Q02763 (Tie-2). The alignment used in the construction
of these homology models was initially generated using
DIALIGN36 and manually corrected during the process
of optimizing the models. To determine the final align-
ment, the highly conserved residues in the G-rich loop
as well as the HRDLAARN motif were considered fixed,
and the amino acids between these fixed points were
shifted to yield a maximum identity or similarity. Se-
quence gaps were positioned in alternative positions
and the resulting 3D models evaluated. The 3D model
was generated using MODELLER,37 a computer pro-
gram that models protein 3D structure by satisfaction
of spatial restraints. Using this software and a template
structure with more than 40% sequence identity to the
target protein, the model is likely to have about 90% of
the main-chain atoms modeled with an RMS deviation
from the X-ray structure of 1 Å.38 The positions shown
in Figure 3 gave the best results. During the optimiza-
tion work, the routines implemented in the
WHATCHECK software39 were employed to analyze
the structures generated. The structures were subse-
Thus, for the construction of the homology models of
VEGFR-2 and Tie-2, both FGF 1 receptor structures
were used as templates. We believe that these homology
models will be useful in the design of inhibitors, the
F igu r e 3. Alignment of the primary sequences of FGF1-R, Tie-2 (TEK), and VEGFR-2 (KDR).42 Identical residues are shown
in green; similarity is indicated by yellow color. Consensus sequence is indicated by red background color. Figure 3 was created
using BOXSHADE with the interface provided at http://www-bioweb.pasteur.fr/seqanal/interfaces/boxshade.html.

Receptor Tyrosine Kinase Inhibitors
J ournal of Medicinal Chemistry, 2003, Vol. 46, No. 14 2923
F igu r e 4. Ramachandran plots: (a) Tie-2 model B. (b) VEGFR-2 model B.
Ta ble 2. Residue Substitutions in the Hydrophobic Pocket
quently imported into WitnotP.40 Residues in highly
unlikely conformations were adjusted and all hydrogens
added. For further work, the residues of the kinase
insert domain were deleted from the structures since
their alignment was rather poor and their influence on
the design of ATP competitive inhibitors is believed to
be minimal. The C-termini were changed to the N-
methylamides, and acetamides were added to the N-
termini. The resulting structures were subjected to a
CHARMM41 minimization, keeping the positions of the
backbone atoms fixed.
residue in IGF1R
residue in VEGFR-2
residue in Tie-2
Leu 1002
Met 1076
Val 1060
Val 1065
Phe 1054
Met 1051
Val 1050
Met 1139
Gly 1149
Leu 840
Val 916
Val 899
Val 914
Ile 892
Leu 889
Ile 888
Leu 1035
Cys 1045
Ile 830
Ile 902
Ile 886
Leu 900
Leu 879
Leu 876
Val 875
Leu 971
Ala 981
the substrate, and this route does not appear to be
promising.
In addition, the conformational alterations in the
activation loop of kinases are manifold.1 To assign a
conformation of this loop via homology modeling would
be very speculative.
Full minimization of the proteins with no restraints
resulted in a collapse of the ATP binding site if no
inhibitor was present. It was thus chosen to employ the
structures resulting from the minimization of side
chains for the initial manual docking work. During the
final minimization of the protein-ligand complexes,
remaining bond and angle deviations were fixed.
The exact structures of the kinase insert domains
were not predicted accurately by MODELLER, as these
regions usually resulted in a random coil and unlikely
ψ-ψ combinations. Since our models were intended to
be employed in the design of competitive inhibitors of
ATP, this region of the protein was deleted from the
models to remove any unwanted strain on the position
of important residues. The Ramachandran plots of the
FGF1R-derived homology models are shown in Figure
4. More than 95% of the residues are found in allowed
regions of the plots. The nonglycine residues found
outside expected areas belong to the activation loop or
highly flexible regions which are disordered in most
crystal structures. None of the residues which we have
termed important for the design of selective inhibitors
is found outside allowed regions, however. Aside from
the DFG motif, the conformation of which we believe to
be correct, the prediction of the conformation of the
activation loop cannot be satisfactorily achieved by the
methods employed. No further optimization of this loop
was carried out, since the influence on the shape of the
ATP binding pocket is thought to be marginal. The
correct conformation of the activation loop would be
required for the design of bisubstrate inhibitors. How-
ever, none of the kinase inhibitors currently under
development or clinical investigation attempts to mimic
In Figures 5 a-d, the side chain atoms of those
residues believed to be important for achieving selectiv-
ity between kinases are displayed. An overlay of the
kinase domain models provides a good basis for the
design of selective inhibitors for each enzyme. Several
substitutions alter the shape and size of the hydro-
phobic pocket and the region where the purine moiety
of ATP would be bound. Theses substitutions are listed
in Table 2.
Aside from the substitutions in the hydrophobic
pocket, variations in the nucleotide binding loop cannot
be disregarded. Next to the conserved phenylalanine
(F835 in Tie-2, F845 in VEGFR-2, and F980 in IGF1R),
A844 in VEGFR-2 is replaced by N834 in Tie-2 and S979
in IGF1R. As shown previously,17 the asparagine in
Tie-2 may contribute to the hydrogen bonding to an
inhibitor by Tie-2 and via this interaction favors a
conformation of the nucleotide binding loop as shown
in Figure 5b. This in turn may synergistically advance
lipophilic interactions between the phenylalanine and
the hydrophobic residues of potential inhibitors. The
differences listed here provided a good explanation for
the patterns of selectivity observed for nakijiquinone
analogues and in the future can be exploited for fine-
tuning of the selectivity of optimized RTK inhibitors.
Static homology models have their limitations with
respect to being able to account for all detected struc-

2924 J ournal of Medicinal Chemistry, 2003, Vol. 46, No. 14
Kissau et al.
F igu r e 5. Developed models for Tie-2 and VEGFR-2: (a) Tie-2 model B. (b) Tie-2 model A. (c) VEGFR-2 model A. (d) VEGFR-2
model B.
tural characteristics. For instance, the role of the
exchange Phe f Tyr in the hinge region can at present
not be explained. None of the known inhibitors of these
kinases exhibits a marked interaction with the phenolic
OH. Likewise, the full effect of the Pro at the beginning
of the KDR nucleotide binding loop cannot be elucidated
easily. In addition, more information from further
crystal structures of these kinases in complex with
inhibitors is needed to fully understand the possible in
vivo conformations of the activation loop. Likewise, the
significant motion of the C-helix can only be properly
estimated by analysis of the structures in complex with
inhibitors of varying morphology or by supportive
experimental data.
With the completed homology models in hand, we set
out to determine and explain the molecular basis for
the biological activity of the nakijiquinone analogues.
In addition to the previously evaluated compounds, the
interactions of compounds 38b and 59c with VEGFR-
2, Tie-2, and IGF1R were explored. The structural
differences in the hydrophobic pocket as described above
provided a useful rationalization for the experimental
observations. The fact that VEGFR-2 is not inhibited
by 59c can be explained by a steric clash of the decalin
with Cys1045, which corresponds to Ala981 and Gly1122
in Tie-2 and IGF1R, respectively. The 18-fold higher
F igu r e 6. Proposed binding mode of 59c to IGF1R.
inhibition of IGF1R compared to Tie-2 is, according to
our analysis, caused by the stronger hydrophobic inter-
action of the trans-decalin with the (smaller) hydropho-
bic pocket in IGF1R. The proposed binding mode of 59c
to IGF1R is shown in Figure 6.

Receptor Tyrosine Kinase Inhibitors
J ournal of Medicinal Chemistry, 2003, Vol. 46, No. 14 2925
Discu ssion
This article describes the synthesis of a small library
of analogues of the naturally occurring tyrosine kinase
inhibitor nakijiquinone C and its evaluation with re-
spect to the inhibition of other kinases. The investiga-
tion is based on the recently proposed concept that
natural products are biologically validated starting
points in structural space for compound library develop-
ment. Within this concept they are regarded as molec-
ular entities selected by evolution for binding to struc-
turally conserved yet genetically mobile protein domains.
The proteins investigated in this study were chosen on
one hand for their biological relevance (i.e., involvement
in tumorigenesis, angiogenesis, or lymphangiogenesis)
and on the other hand due to the high homology in their
ATP binding kinase domains (>45%). According to the
above-mentioned “domain concept”, this high homology
suggests that a library synthesized around a natural
product binding to such domains should yield ligands
for the similar domains in other proteins with a high
hit rate. An overlay of the crystal structures of FGF1R,
IGF1R, and the homology models generated for VEG-
FR-2 and Tie-2 (similar to Figure 2) and comparison
with the homology to the ATP binding domain in the
natural target Her-2/Neu of the underlying nakijiquino-
ne C (34% homology to FGF1R) suggests that a nak-
ijiquinone-derived library should fulfill the demands
raised above.
In this paper we focus on a natural product with
proven biological activity as a biologically prevalidated
starting point in structural space for compound library
development. We would like to stress, however, that the
requirement of biological prevalidation is not limited to
naturally occurring compounds. Of course nonnatural
compound classes that have been found to be biologically
relevant can be employed as structural starting points
within the reasoning of our proposal as well.
After the development of the first ATP competitive
kinase inhibitors and the determination of the first
kinase structure (PKA),43 the successful development
of selective ATP-competitive inhibitors was thought to
be highly unlikely given the high degree of conservation
of residues involved in ATP binding and the similarity
of the kinase domains.13 Our work and the success
encountered in the kinase field during the past decade
support our proposal that the similarity of protein
domains should be employed for the development of
selective inhibitors based on an already biologically
validated compound.
F igu r e 7. The solvent-accessible volume of the hydrophobic
pocket of IGF1R (green), KDR (yellow), and Tie-2 (red)
We note that compound 59c was obtained as an
inseparable mixture of diastereomers whose composition
could not be determined by NMR spectroscopy. Molec-
ular modeling suggests, however, that the diastereomer
depicted in Figure 6 is expected to be the active
component, since the other diastereomer would not fit
well into the binding site in the homology model.
This result, combined with our previously published
analyses, provides a good indication about future fine-
tuning of selectivity toward a particular kinase. Explor-
ing the differences in the hydrophobic pocket, for
example, by adding different substituents at varying
positions of the decalin moiety, should allow the use of
nakijiquinone analogues as a molecular tool kit for the
study of the biological effects caused by inhibition of a
certain receptor tyrosine kinase. The explanation for the
selectivity of the bis-threonine derivative 38b follows
the same lines. It too is an ATP competitive inhibitor.
The selectivity between VEGFR-2 and VEGFR-3 is
somewhat intriguing, given their homology, especially
within the ATP binding site. Further variation of this
structure should provide sufficient data to fully eluci-
date the binding mode of receptor tyrosine kinase
inhibitors belonging to this compound class.
We note, however, that the structure of 38b does not
fit the amino acid-quinone-decalin pattern displayed
by 59c. Since the positioning of a carboxylate in the
hydrophobic pocket is improbable, 38b must bind in a
different mode. Our molecular modeling work indicates
that a binding mode where 38b blocks the triphosphate
binding region without interacting with residues in the
hinge region could serve as an explanation. Such a bind-
ing mode would provide hints to the cause of the intrigu-
ing selectivity between VEGFR-2 and VEGFR-3.
The data described above show that a relatively small
library designed around a naturally occurring kinase
inhibitor may yield potent inhibitors for other kinases
with a high hit rate (7 inhibitors out of 74 compounds
investigated). Starting from the natural product nak-
ijiquinone C, a known inhibitor of the Her-2/Neu kinase,
structurally related analogues were found to inhibit Tie-
2, VEGFR2/3, and IGF1R. Several of the hits clustered
at the kinase Tie-2. While from a 74-member library a
statistically relevant statement cannot be deduced, this
result clearly supports the validity of the domain
concept for library development.
Further proof of the importance of exploiting struc-
tural differences in the hydrophobic pocket to attain
selectivity can be derived from comparing the size and
shape of the hydrophobic pockets in our homology
models.42
Figure 7 demonstrates, the possibility of achieving
selectivity through variation of the hydrophobic part of
an inhibitor, while leaving the part which contributes
essential hydrogen bonds to the hinge region untouched.
The results recorded in this study also convincingly
demonstrate that the trains of thought starting from
protein domains and biologically active natural products

2926 J ournal of Medicinal Chemistry, 2003, Vol. 46, No. 14
Kissau et al.
high-resolution mass spectra (HRMS) were measured on a
Finnigan MAT MS70 (for EI and FAB) or a Finnigan LCQ (for
ESI, all samples are solved in H2O/MeOH/5%HCOOH) spec-
trometer. Specific rotations were measured with a Perkin-
Elmer 241 polarimeter. Flash chromatography was performed
using Baker silica gel (230-400 mesh ASTM). Reaction
progress was monitored by TLC using Merck silica gel 60F254
aluminum sheets (Merck, Darmstadt, Germany).
High-pressure liquid chromatography (HPLC) was per-
formed with a Rainin-Dynamax (model SD-300) instrument.
A gradient composed of I (0.1% TFA in water) and II (0.1%TFA
in acetonitrile) was used for semipreparative and preparative
HPLC; method A: Time 0, 20% II. Time 15 min 40% II. Time
30 min 100% II. Time 35 100% II. Method B: Time 0, 50% II.
Time 15 min 75% II. Time 30 min 100% II. Time 35 100% II.
Molecu la r Mod elin g. The molecular modeling studies
were carried out on a SGI O2 work station (RS10000, 300
MHz). The CHARMM version used for the minimizations of
protein-ligand complexes was c27b. Input files for CHARMM
were created within WitnotP. Minimizations were performed
on an Origin 2000 (16 x RS12000 400 MHz). The homology
models were constructed using MODELLER version 4.0, using
the standard settings. The Ramachandran plots of the homol-
ogy models were created using WHATCHECK. The volumes
and shapes of the hydrophobic pocket were calculated by filling
the voids using the PASS algorithm. The front boundary was
arbitrarily defined by a plane constructed using F1007-CR,
F1007-C, M1076-CR, and M1076-N (numbers referring to the
IGF1R sequence) and the corresponding residues in VEGFR-2
and Tie-2. The pseudo-atoms placed by PASS were covered
with an SA-surface (probe radius 1.4 Å). Volumes were
calculated using WitnotP.
The homology model employed in this study differs from the
model used in our previous publication.17c The homology model
for the VEGFR-2 active site used in the previous publication
was derived by replacing corresponding residues in the FGF1R
structure, neglecting all residues located more than 5 Å from
the ATP binding site. To determine if differences in the
primary sequences in more remote parts of the kinase domain
could influence the shape and size of the ATP binding pocket,
a full model based on the FGF1R template was created using
MODELLER. The new model shows the nucleotide binding
loop in a slightly different form than in the previous model.
call for the synthesis of compound libraries. The use of
a given natural product alone is not sufficient. Thus,
the original nakijiquinone was not active on any of the
other kinases investigated.
In line with the domain concept, natural products and
domain structure therefore can only be regarded as the
initial starting points in the hit finding process. The
varying amino acid sequences found in domains with
very similar fold call for the generation of chemical
diversity in the library that matches the diversity in
the precise architecture of domains with conserved fold.
In a sense, the similarity of protein structures provides
the basis for the choice of the underlying framework of
the library members, and then the differences in the
similar protein structures are exploited to obtain selec-
tivity. We would like to point out that the Tie-2 and
VEGFR-3 inhibitors identified in this study might be
used advantageously for the development of a new class
of agents to interfere in the processes of angiogenesis
(Tie-2) and lymphangiogenesis. Given its different pat-
tern of H-bond donors/acceptors when compared to other
active members of the library, this may be particularly
true for the bis(threonine) derivative 38b. At the out-
set of the study inhibitors of Tie-2 and VEGFR-3 were
not known (in the meantime Tie-2 inhibitors were
identified).44-49 This underscores the potential harbored
by the domains/natural product approach. The inhibi-
tion of IGF1R by 59c and its selectivity also warrant
further investigation.
The similarity of the domains facilitates the use of
modern methods of molecular modeling to rationalize
the experimental findings and to further stimulate the
development of improved inhibitors. In this context, the
development of convincing homology models becomes
important, in particular when extensive structural data
(X-ray, NMR structures) of the domains are missing.
In this study, two homology models for the receptor
tyrosine kinases Tie-2 (TEK) and VEGFR-2 (KDR)50
were developed and employed in the molecular modeling
studies. In agreement with established models for RTK
inhibitors, analysis of these models has shown that
selective inhibitors cannot be designed by attempting
to form different pattern of hydrogen bonds alone.
Rather, differences in the hydrophobic pockets have to
be exploited. Instead, the hydrogen bonds serve to orient
the compounds within the active site. These findings
are consistent with results previously reported for
FGF1-R, IGF1R, and VEGFR-2.51 Furthermore, we
believe these enzymes to be a convincing demonstration
of the limitations of the rigid receptor hypothesis.
Rather than fitting molecules into a rigid receptor
structure, it appears to be crucial to build several models
to accommodate inhibitors of varying shapes and sizes.
Finally, we have employed these models to elucidate the
SAR data collected from our nakijiquinone analogue
library. These results can now be employed for the
design of more active and selective inhibitors of the
receptor tyrosine kinases Tie-2 and VEGFR-2.
3-Meth oxy-2-isop r op oxyp h en yl Br om id e (10). To 85 mL
of DMSO was added KOH (8.90 g, 0.159 mol). After the
mixture was stirred for 10 min, 2-hydroxy-3-methoxyphenyl
bromide 9 (8.70 g, 42.9 mmol) was added, followed after 10
min by 2-bromopropane (8.1 mL, 86.3 mmol). Stirring was
continued for 1 h after which the KOH was filtered and the
filtrate was diluted with 100 mL CH2Cl2. The organic layer
was washed with a saturated solution of NH4Cl (2×), NaCl
(2×), dried over Na2SO4, and evaporated to dryness under
reduced pressure. The crude product was purified by flash
chromatography on silica gel (95:5 hexane/EtOAc) to give 10.1
g (96%) of pure 4 as a colorless oil. Rf: 0.64 (9:1 hexane/EtOAc);
3
4
1H NMR (250 MHz, CDCl3): δ 7.12 (dd, J ) 7.5 Hz, J ) 2.0
Hz, 1H, H-6), 6.90-6.80 (m, 2H, H-4, H-5), 4.59-4.50 (m, 1H,
3
CH(CH3)2) 3.81 (s, 3H, OCH3), 1.32 (d, J ) 6.2 Hz, 6H, CH-
(CH3)2); 13C NMR (62.9 MHz, CDCl3): δ 154.1 (C-3), 144.7 (C-
2), 125.0 (C-6), 124.2 (C-5), 118.7 (C-4), 111.5 (C-1), 75.8
(CH(CH3)2), 55.9 (OCH3), 22.5 (CH(CH3)2); MS (EI, 70 eV): m/z
(%) 246/244 (17/17) [M+], 204/202 (98/100) [M+ - C(CH3)2], 189/
187 (29/29) [M+ - CH3 - C(CH3)2], 124 (10), 109/107 (6/5), 79
(7); HRMS (70 eV): calcd for C10H13O281Br: 246.0079, found:
246.0076.
6-Ben zyl-2-m eth oxyp h en ol (12) a n d 6-Bu tyl-2-m eth -
oxyp h en ol (14). To a solution of 3-methoxy-2-isopropoxyphe-
nyl bromide 10 (300 mg, 4.90 mmol) in 3.2 mL THF at -78 °C
was added dropwise n-BuLi (2.5 M in hexane) (539 µL, 1.35
mmol). After the mixture was stirred at this temperature for
30 min, CuI (285 mg, 1.50 mmol) was added. The solution was
slowly warmed to -40 °C, then cooled to -78 °C, and at this
temperature a solution of benzyl bromide (217 µL, 1.83 mmol)
in 0.5 mL of THF was added. The mixture was warmed to room
Exp er im en ta l Section
Gen er a l. Melting points were determined in open capillar-
1
ies using a Bu¨chi 535 apparatus and are uncorrected. H and
13C NMR spectra were recorded on a Bruker AC 250 or DRX
500 spectrometer at room temperature. IR spectra were
recorded on a Bruker IFS 88 spectrometer. Mass spectra and

Receptor Tyrosine Kinase Inhibitors
J ournal of Medicinal Chemistry, 2003, Vol. 46, No. 14 2927
temperature, diluted with 20 mL ether and 20 mL water, and
extracted with ether (1 × 20 mL). The combined organic layers
were washed with a saturated solution of Na2CO3 until the
aqueous layer was colorless, dried over Na2SO4, and evapo-
rated under reduced pressure. The crude product was purified
by flash chromatography on silica gel (95:5 hexane/EtOAc) to
give 111 mg of products 11 and 13 as a mixture.
EtOAc) to give 208 mg (91%) of pure 18 as a colorless oil. Rf.:
0.19 (9:1 hexane/EtOAc); 1H NMR (500 MHz, CDCl3): δ 7.04-
6.91 (m, 2H, H-6, H-7), 6.81-6.77 (m, 1H, H-5), 4.74-4.71 (m,
1H, H-1), 4.65-4.55 (m, 1H, OCH(CH3)2), 3.83 (s, 3H, OCH3),
3.47-3.40 (m, 1H, OH), 2.39-2.28 (m, 1H, H-1′), 2.17-0.83
3
(m, 10H, cyclohexane), 1.35 (d, J ) 6.1 Hz, 3H, OCH(CH3)2),
3
1.22 (d, J ) 6.1 Hz, 3H, OCH(CH3)2); 13C NMR (125.6 MHz,
CDCl3): δ 152.3 (C-4), 144.0 (C-3), 137.7 (C-2), 123.2 (C-7),
119.1 (C-6), 110.8 (C-5), 81.2 (C-1), 74.3/74.0 (OCH(CH3)2), 55.5
(OCH3), 44.4/43.7 (C-1′), 29.4/29.3 (C-4′), 28.9/28.6 (C-3′), 26.7/
26.5 (C-3′), 26.2/26.0 (C-2′), 25.7/25.5 (C-2′), 22.9/22.3 (OCH-
(CH3)2), 21.5/21.0 (OCH(CH3)2); MS (EI, 70 eV): m/z (%) 278
(11) [M+], 218 (7) [M+ - C3H7 - OH], 195 (24) 153 (100) [M+
- C3H7 - C6H10], 137 (14) [M+ - OC3H7 - C6H10], 93 (7);
HRMS (70 eV): calcd for C17H26O3: 278.1882, found: 278.1873.
To a solution of the mixture of 11 and 13 in 10 mL of
CH2Cl2 at 0 °C was added BCl3 (1 M in CH2Cl2, 510 µL, 0.51
mmol). After the mixture was stirred at 0 °C for 1 h, water
was added, and the mixture was extracted with CH2Cl2 (3 ×
15 mL). The combined organic layers were dried over Na2SO4
and evaporated under reduced pressure. The crude product
was purified by flash chromatography on silica gel (95:5
hexane/EtOAc) to give 44.9 mg (17%, 2 steps) of pure 12 as a
colorless oil and 46.8 mg (21%, 2 steps) of pure 14 as a colorless
oil.
1-Cycloh exyl-4-m eth oxy-3-isop r op oxyben zylm eth oxy-
oxa lyl Ester (19). DIPEA (198.0 µL, 1.13 mmol), with a
catalytic amount of DMAP, followed by methyl oxalyl chloride
(104 µL, 1.13 mmol) were added to a solution of alcohol 18 (50
mg, 0.18 mmol) in dry CH2Cl2. The reaction mixture was
stirred at room temperature for 24 h. The reaction was diluted
with 10 mL of CH2Cl2 and washed with a saturated solution
of Na2CO3 and brine. The water solution was extracted with
CH2Cl2 (2 × 15 mL). The combined organic layers were dried
over Na2SO4 and evaporated under reduced pressure. The
crude product was purified by flash chromatography on silica
gel (95:5 hexane/EtOAc) to give 44.4 mg (68%) of pure 19 as a
colorless oil. Rf: 0.38 (9:1 hexane/EtOAc); 1H NMR (500 MHz,
Compound 12: Rf: 0.54 (9:1 hexane/EtOAc); 1H NMR (250
MHz, CDCl3): δ 7.26-7.14 (m, 5H, H-2′(2), H-3′ (2), H-4′),
4
6.77-6.68 (m, 3 H, H-3, H-4, H-5), 5.73 (d, J ) 1.53 Hz, 1H,
OH), 4.00 (s, 2H, H-7), 3.85 (s, 3H, OCH3); 13C NMR (62.9 MHz,
CDCl3): δ 146.5 (C-2), 143.6 (C-1′), 140.9 (C-1), 129.0, 128.4,
127.1 (C-6, C-2′(2), C-3′(2)), 126.0 (C-4′), 122.9 (C-4), 119.5 (C-
5), 108.8 (C-3), 56.1 (OCH3), 35.6 (C-7); MS (EI, 70 eV): m/z
(%) 214 (100) [M+], 197 (7) [M+ - OH], 181 (19), 153 (12), 136
(23), 107 (6), 91(5); HRMS (70 eV): calcd for
C14H14O2:
214.0994, found: 214.0989.
1
Compound 14: Rf: 0.6 (9:1 hexane/EtOAc); H NMR (250
MHz, CDCl3): δ 6.80-6.67 (m, 3 H, H-3, H-4, H-5), 5.68 (d, 4J
) 1.83 Hz, 1H, OH), 3.82 (s, 3H, OCH3), 2.67-2.61 (m, 2H,
(CH2CH2CH2CH3)), 1.66-1.54 (m, 2H, (CH2CH2CH2CH3)),
1.45-1.28 (m, 2H, (CH2CH2CH2CH3)), 0.98-0.87 (m, 3H, (CH2-
CH2CH2CH3); 13C NMR (62.9 MHz, CDCl3): δ 146.2 (C-2),
143.5 (C-1), 128.8 (C-6), 122.4 (C-5), 119.2 (C-4), 108.2 (C-3),
56.0 (OCH3), 31.8 (CH2CH2CH2CH3), 29.5 (CH2CH2CH2CH3),
23.0 (CH2CH2CH2CH3), 14.1 (CH2CH2CH2CH3); MS (EI, 70
eV): m/z (%) 180 (53) [M+], 137 (100) [M+ - CH2CH2CH3], 106
(8) [M+ - CH2CH2CH3 - OCH3], 77 (7), 41 (3); HRMS (70 eV):
calcd for C11H16O2: 180.1150, found: 180.1118.
3
3
CDCl3): δ 7.01 (ψt, J ) 7.9 Hz, 1H, H-7), 6.94 (dd, J ) 7.9
4
3
4
Hz, J ) 1.4 Hz, 1H, H-6), 6.82 (dd, J ) 8.0 Hz, J ) 1.4 Hz,
1H, H-5), 6.18 (d, 3J ) 7.0 Hz, 1H, H-1), 4.69-4.64 (m, 1H,
OCH(CH3)2), 3.83 (s, 3H, OCH3), 2.35-2.28 (m, 1H, H-1′),
1.95-0.92 (m, 10H, cyclohexane), 1.39 (d, 3J ) 6.1 Hz, 3H,
CH(CH3)2), 1.25 (d, 3J ) 6.1 Hz, 3H, OCH(CH3)2); 13C NMR
(125.6 MHz, CDCl3): δ 158.6 (CdO), 157.1 (CdO), 152.4
(C-4), 144.1 (C-3), 132.8 (C-2), 123.2 (C-7), 118.7 (C-6), 111.4
(C-5), 78.4 (C-1), 74.9 (OCH(CH3)2), 55.5/53.4 (OCH3), 43.6/
43.1 ((CO)2OCH3), 41.2/41.1 (C-1′), 29.2/29.1 (C-4′), 28.8/28.3
(C-3′(2)), 26.2/26.0 (C-2′(2)), 22.9, 22.3 (OCH(CH3)2); MS
(EI, 70 eV): m/z (%) 364 (12) [M+], 219 (16) [M+ - C3H6-
CH3O(CO)2O], 218 (100) [M+ - C3H7 - CH3O(CO)2O], 138(10),
137 (59) [M+ - C3H6 - C6H10 - CH3O(CO)2O]; HRMS (70 eV):
calcd for C20H28O6: 364.1886, found: 364.1892.
6-Ben zyl-2-m eth oxy-2,5-cycloh exa d ien e-1,4-d ion e (15).
To a solution of 6-benzyl-2-methoxyphenol 12 (50.9 g, 0.237
mmol) in 1.5 mL of DMF was added salcomine (N,N′-bis-
(salicylidene)ethylendiamine-cobalt(II) hydrate) (8.7 mg, 26
µmol). The solution was stirred at room temperature under
an oxygen atmosphere for 12 h, after which the solution was
diluted with 20 mL of ether and 10 mL of water, and extracted
with ether (3 × 15 mL). The combined organic layers were
dried over Na2SO4 and evaporated under reduced pressure.
The crude product was purified by flash chromatography on
silica gel (4:1 hexane/EtOAc) to give 38 mg (70%) of pure 15
as a yellow solid. Rf: 0.26 (3:1 hexane/EtOAc); mp 127 °C; 1H
NMR (500 MHz, CDCl3): δ 7.33-7.18 (m, 5H, H-2′(2), H-3′-
(2), H-4′), 6.30 (dd, 4J ) 1.5 Hz, 4J ) 2.3 Hz, 1H, H-5), 5.86 (d,
4J ) 2.3 Hz, 1H, H-3), 3.81 (s, 3H, OCH3), 3.75 (d, 4J ) 1.5
Hz, 2H, H-7); 13C NMR (125.6 MHz, CDCl3): δ 187.4 (C-1),
181.8 (C-4), 158.8 (C-2), 146.7 (C-6), 136.4 (C-1′), 133.8 (C-5),
129.4 (C-2′(2C)), 128.3 (C-3′(2C)), 127.0 (C-4′), 107.2 (C-3), 56.3
(OCH3), 34.9 (C-7); MS (EI, 70 eV): m/z (%) 228 (100) [M+],
213 (45) [M+ - CH3], 197 (5) [M+ - OCH3], 196 (18), 168 (5),
157 (8), 143 (12), 128(8), 115 (34); HRMS (70 eV): calcd for
6-(Cycloh exylm eth yl)-2-m eth oxyp h en ol (20). To a solu-
tion of methoxyoxalyl ester 19 (44.4 mg, 0.122 mmol) in 8.7
mL of dry toluene were added tributyltin hydride (66 µL, 0.249
mmol) and a catalytic amount of AIBN. The reaction mixture
was heated at reflux for 2 h. The solvent was removed in vacuo,
and the residue was purified by flash chromatography on silica
gel (95:5 hexane/EtOAc). To a solution of the deoxygenated
product in 6.3 mL of CH2Cl2 at 0 °C was added BCl3 (1 M in
CH2Cl2) (135 µL, 0.135 mmol). After the mixture was stirred
at 0 °C for 1 h, water was added and extracted with CH2Cl2 (3
× 10 mL). The combined organic layers were dried over
Na2SO4 and evaporated under reduced pressure. The crude
product was purified by flash chromatography on silica gel
(95:5 hexane/EtOAc) to give 9.8 mg (37%, two steps) of pure
20 as a white solid; Rf: 0.48 (9:1 hexane/EtOAc); mp 52 °C;
1H NMR (500 MHz, CDCl3): δ 6.77-6.69 (m, 3H, H-3, H-4,
4
H-5), 5.64 (d, J ) 3.0 Hz, 1H, OH), 3.88 (s, 3H, OCH3), 2.55
(d, 3J ) 7.1 Hz, 2H, H-7), 1.70-0.90 (m, 11H, cyclohexane);
13C NMR (125.6 MHz, CDCl3): δ 146.3 (C-2), 143.7 (C-1), 127.2
(C-6), 123.4 (C-5), 118.8 (C-4), 108.0 (C-3), 55.9 (OCH3), 38.2
(C-7), 37.5 (C-1′), 33.3 (C-2′(2)), 26.6 (C-4′), 26.4 (C-3′(2)); MS
(EI, 70 eV): m/z (%) 220 (100) [M+], 138 (88) [M+ - C6H10],
137 (76) [M+ - C6H11], 123 (12) [M+ - C6H10 - CH3], 122 (6)
[M+ - C6H11 - CH3], 106 (6), 83 (5); HRMS (70 eV): calcd for
C
14H12O3: 228.0786, found: 228.0791.
1-Cycloh exyl-4-m eth oxy-3-isop r op oxyben zyl Alcoh ol
(18). To a solution 3-methoxy-2-isopropoxyphenyl bromide 9
of (200 mg, 0.817 mmol) in 5.9 mL of THF was added dropwise
1.3 mL of TMEDA. The mixture was cooled to -78 °C, and
n-BuLi (2.5 M in hexane, 2.2 mL, 5.5 mmol) was added
dropwise. The solution was slowly warmed to -40 °C, then
cooled to -78 °C, and to this solution was added a solution of
cyclohexanecarbaldehyde 17 (316 µL, 4.10 mmol) in 3.3 mL of
THF. The mixture was warmed to room temperature and
diluted with 15 mL of ether. The organic layer was washed
with 2 N HCl (2 × 15 mL) and brine, dried over Na2SO4, and
evaporated under reduced pressure. The crude product was
purified by flash chromatography on silica gel (95:5 hexane/
C
14H20O2: 220.1463, found: 220.1469.
(E)-2-(tr a n s-Deca h yd r o-1-n a p h th yl)br om oeth en e (23).
To a suspension of bromomethyl(triphenyl)phosphonium bro-
mide (3.5 g, 8.0 mmol) in 13 mL of dry THF at -78 °C was
added KOtBu (885 mg, 8.0 mmol). The yellow reaction mixture
was stirred 30 min at this temperature after which a solution
of trans-1-decalone (500 mg, 3.25 mmol) in 2.5 mL of THF was

2928 J ournal of Medicinal Chemistry, 2003, Vol. 46, No. 14
Kissau et al.
added. The mixture was allowed to warm to room temperature,
then stirred for 2 h, and diluted with water and ether. The
aqueous solution was extracted with ether (3 × 20 mL). The
combined organic layers were dried over Na2SO4 and evapo-
rated under reduced pressure. The crude product was purified
by flash chromatography on silica gel (95:5 hexane/EtOAc) to
give 548 mg (73%) of pure 23 as a yellowish oil. Rf: 0.75 (9:1
hexane/EtOAc; 1H NMR (250 MHz, CDCl3): δ 5.76 (s, 1H, Cd
CHBr), 3.04-2.99 (m, 1H, H-2), 1.86-1.03 (m, 15H, decalin);
13C NMR (62.9 MHz, CDCl3): δ 148.4 (C-1), 97.8 (CdCHBr),
48.8 (C-8a′), 44.6 (C-4a′), 34.4, 34.4, 32.6, 28.6 (C-2′, C-4′, C-5′,
C-8′), 26.6, 26.3, 26.2 (C-3′, C-6′, C-7′); MS (EI, 70 eV): m/z
NMR (500 MHz, CDCl3): δ 6.74-6.63 (m, 3H, H-3, H-4, H-5),
4
6.02 (s, 1H, H-7), 5.57 (d, J ) 1.5 Hz, 1H, OH), 3.81 (s, 3H,
OCH3), 2.71-2.66 (m, 1H, H-2′), 1.94-1.05 (m, 15H, Decalin);
13C NMR (125.6 MHz, CDCl3): δ 148.0 (C-2), 146.0 (C-1′), 143.0
(C-1), 125.4 (C-6), 122.8, (C-4), 118.9 (C-5), 113.5 (C-7), 108.7
(C-3), 55.9 (OCH3), 48.4 (C-8a′), 45.0 (C-4a′), 35.0, 34.7 (C-4′,
C-5′), 31.0, 29.0, (C-2′, C-8′), 27.7, 26.5, 26.3 (C-3′, C-6′, C-7′);
MS (EI, 70 eV): m/z (%) 272 (42) [M+], 176 (10), 144 (7), 135
(100) [M+ - C8O2H9], 93 (6); HRMS (70 eV): calcd for
C
18H24O2: 272.1776, found: 272.1771.
3-[(tr a n s-Deca h yd r o-1-n a p h th yl)m eth yl]-1-m eth oxy-2-
isop r op oxyben zen e (28). A solution of olefin 25 (129 mg,
0.411 mmol) in a 4:1 mixture of methanol-THF (9 mL) was
hydrogenated for 48 h at room temperature and atmospheric
pressure in the presence of 5% rhodium supported on alumina
(27.0 mg, 0.132 mmol). The catalyst was filtered off, and the
solvent was removed under reduced pressure. The crude
product was purified by flash chromatography on silica gel
(95:5 hexane/EtOAc) to give 122 mg (94%) of a 3:1 isomer
mixture of 28 as a colorless oil; Rf.: 0.58 (9:1 hexane/EtOAc);
1H NMR (500 MHz, CDCl3): δ 6.93-6.89 (m, 1H, H-4), 6.74-
6.68 (m, 2H, H-5, H-6), 4.51-4.46 (m, 1H, OCH(CH3)2), 3.81
(%) 230/228 (13/14) [M+], 149 (100) [M+ - Br], 135 (8) [M+
Br - CH2], 107 (14), 96 (11); HRMS (70 eV): calcd for C11H17
Br: 228.0514, found: 228.0516.
-
79
-
3-Meth oxy-2-isop r op oxyp h en ylylbor on ic Acid (24). To
a solution of aromatic compound 10 (1.20 g, 4.90 mmol) in 16
mL of THF at -78 °C was added dropwise n-BuLi (2.2 mL,
5.5 mmol) (2.5 M in hexane). The solution was slowly warmed
to -40 °C, then cooled to -78 °C, and at this temperature
B(OMe)3 (2.4 mL, 25.2 mmol) was added. The mixture was
warmed to room temperature and cooled to 0 °C, and 5%
aqueous HCl was added until a pH of 6.5 was reached. The
reaction mixture was diluted with 20 mL of ether and 20 mL
of water and extracted with ether (3 × 20 mL). The combined
organic layers were dried over Na2SO4 and evaporated under
reduced pressure. The crude product was purified by flash
chromatography on silica gel (4:1 hexane/EtOAc) to give 237
mg (23%) of pure 24 as a white solid. Rf: 0.36 (3:1 hexane/
2
4
(s, 3H, OCH3), 3.10 (dd, J ) 13.0 Hz, J ) 3.6 Hz, 1H, H-7),
2
4
2.72 (dd, J ) 13.0 Hz, J ) 3.7 Hz, 1H, H-7), 2.62-2.57 (m,
1H, H-1′), 2.17-2.14 (m, 1H, H-2′), 2.08-2.03 (m, 1H, H-2′),
1.93-0.81 (m, 14H, decalin), 1.29-1.25 (m, 6H, OCH(CH3)2);
13C NMR (125.6 MHz, CDCl3): δ 152.8 (C-1), 145.1/145.0 (C-
2), 137.2/136.6 (C-3), 123.0 (C-4), 122.8/122.6 (C-5), 109.5/109.4
(C-6), 74.2 (OCH(CH3)2), 55.5 (OCH3), 48.5/46.8 (C-8a′), 43.1/
43.0 (C-4a′), 39.2/36.3 (C-7), 35.2/35.0 (C-1′), 34.6/34.6, 32.2/
31.2, 30.4 (C-4′, C-5′, C-8′), 28.9/27.2 (C-2′), 27.1/27.0, 26.7/
26.6, 25.9 (C-3′, C-6′, C-7′), 22.8/22.6, 22.4/20.6 (OCH(CH3)2);
MS (EI, 70 eV): m/z %) ) 316 (24) [M+], 274 (47) [M+ - C3H6],
254 (14), 222 (10), 206 (21), 180 (42), 149(30), 138 (67), 137
(100) [M+ - C11O2H15], 95 (31), 81 (54); HRMS (70 eV): calcd
for C21H32O2: 316.2402, found: 316.2396.
3
EtOAc); mp 81 °C; 1H NMR (250 MHz, CDCl3): δ 7.40 (dd, J
4
) 7.2 Hz, J ) 1.7 Hz, 1H, H-6), 7.13-7.00 (m, 2H, H-4, H-5),
6.25 (brs, 2H, OH), 4.84-4.69 (m, 1H, OCH(CH3)2), 3.86 (s,
3H, OCH3), 1.31 (d, 3J ) 6.2 Hz, 6H, OCH(CH3)2); 13C NMR
(62.9 MHz, CDCl3): δ 151.6 (C-3, C-2), 127.3 (C-6), 124.0 (C-
5), 123.9 (C-1), 115.6 (C-4), 75.3 (OCH(CH3)2), 55.7 (OCH3),
22.3 (OCH(CH3)2); MS (EI, 70 eV): m/z (%) 210 (28) [M+], 168
(70) [M+ - C(CH3)2], 167 (15) [M+ - CH(CH3)2], 150 (100) [M+
- H - OCH(CH3)2], 149 (27), 135 (12), 107 (10); HRMS (70
eV): calcd for C10H15O4B: 210.1063; found: 210.1060.
N-(5-Ben zyl-3,6-d ioxo-4-N-L-ser in e-1,4-cycloh exa d i-
en yl)-L-ser in e (31a ). A mixture of quinone 15 (15.0 mg, 65.8
µmol) and L-serine (10.2 mg, 97.1 µmol) in 7.8 mL of EtOH
was stirred at 40 °C for 24 h in the presence of NaHCO3 (22.2
mg, 0.264 mmol). After filtration, the filtrate was evaporated
under reduced pressure, and the residue was was purified by
reversed-phase HPLC (VP 250/21 Nucleosil 100-7 C-18 HD,
flow rate 13 mL/min; method A; UV detection at 300 nm; tR
3-[(1E)-(tr a n s-Deca h yd r o-1-n a p h t h yl)m et h ylen e]-1-
m eth oxy-2-isop r op oxyben zen e (25). To a solution of vinyl
bromide 23 (77.8 mg, 0.340 mmol) in 1.1 mL of toluene was
added Pd(PPh3)4 (10.6 mg, 9.19 µmol). The solution was stirred
for 10 min, and a solution of boronic acid 24 (78.6 mg, 0.375
mmol) in 380 µL of ethanol and 228 µL of a 2 M solution of
Na2CO3 were added. The solution was warmed to 90 °C for
24 h, diluted with ether and water, and extracted with ether
(3 × 15 mL). The combined organic layers were dried over
Na2SO4 and evaporated under reduced pressure. The crude
product was purified by flash chromatography on silica gel
(95:5 hexane/EtOAc) to give 93.7 mg (88%) of pure 25 as a
colorless oil. Rf: 0.48 (9:1 hexane/EtOAc); 1H NMR (500 MHz,
CDCl3): δ 6.97-6.94 (m, 1H, H-4), 6.78-6.73 (m, 2 H, H-5,
H-6), 6.17 (s, 1H, H-7), 4.30-4.25 (m, 1H, OCH(CH3)2), 3.85
(s, 3H, OCH3), 2.84-2.81 (m, 1H, H-2′), 1.96-1.94 (m, 1H,
17.5 min) to afford 17a (4.7 mg, 24%), as a dark red solid; [R]20
D
) -23.5° (c ) 0.085, MeOH), 1H NMR (500 MHz, CD3OD):
7.27-7.24 (m, 2H, H-2′(2)), 7.17-7.14 (m, 3H, H-3′(2), H-4′),
5.39 (s, 1H, H-2), 4.48 (s, 1H, R-CH-4-Ser), 4.23 (s, 1H, R-CH-
1-Ser), 4.02-3.94 (m, 3H, â-CH2-4-Ser, â-CH-1-Ser), 3.86-3.81
3
(m, 2H, â-CH-1-Ser, H-7), 3.61 (d, J ) 8.7 Hz, 1H, H-7); MS
(ESI): m/z: calcd for C19H20N2O8: 404,37 found: 405 [M +
H].
N-(5-Ben zyl-3,6-dioxo-4-N-D-valin e-1,4-cycloh exadien yl)-
D-va lin e (31d ). A mixture of quinone 15 (17.0 mg, 74.5 mmol)
and D-valine (13.4 mg, 0.114 mmol) in 8.8 mL of EtOH was
stirred at 30 °C for 24 h in the presence of NaHCO3 (98 mg,
1.17 mmol). After filtration, the filtrate was evaporated under
reduced pressure, and the residue was was purified by
reversed-phase HPLC (VP 250/21 Nucleosil 100-7 C-18 HD,
flow rate 13 mL/min; method B, UV detection at 300 nm; tR
3
H-2′), 1.87-0.84 (m, 14H, decalin), 1.24 (d, J ) 6.2 Hz, 6H,
OCH(CH3)2); 13C NMR (125.6 MHz, CDCl3): δ 153.1 (C-1),
145.9 (C-2), 145.3 (C-1′), 134.3 (C-5), 122.8, 122.7 (C-1, C-4),
115.9 (C-7), 110.2 (C-6), 75.2 (OCH(CH3)2), 55.7 (OCH3), 48.4
(C-8a′), 44.9 (C-4a′), 35.0, 34.7 (C-4′, C-5′), 30.8, 29.1, (C-2′,
C-8′), 27.5, 26.6 (C-3′, C-6′, C-7′), 22.7, 22.2 (OCH(CH3)2); MS
(EI, 70 eV): m/z (%) 314 (9) [M+], 242 (18), 226 (10), 200 (100),
185 (16), 135 (25), 130 (32); HRMS (70 eV): calcd for
4.2 min) to afford 31d (18.4 mg, 75%), as a red solid; [R]20
)
D
+10.2° (c ) 0.92, MeOH); 1H NMR (500 MHz, CD3OD): 7.30-
7.12 (m, 5H, H-2′(2), H-3′(2), H-4′), 5.42 (s, 1H, H-2), 4.26-
4.21 (m, 1H, R-CH-4-Val), 3.97 (d, 3J ) 5.4 Hz, 1H, R-CH-1-
Val), 3.80 (s, 2H, H-7), 2.34-2.27 (m, 2H, â-CH-Val), 1.01 (d,
3J ) 6.9 Hz, 3H, γ-CH3-4-Val), 0.98 (d, 3J ) 6.9 Hz, 3H, γ-CH3-
C
21H30O2: 314.2246, found: 314.2252.
6-[(1E)-(tr a n s-Deca h yd r o-1-n a p h t h yl)m et h ylen e]-2-
m eth oxyp h en ol (26). To a solution of compound 25 (86.6 mg,
0.292 mmol) in 6.3 mL of CH2Cl2 at 0 °C was added BCl3 (1 M
in CH2Cl2) (325 µL, 0.32 mmol). After the mixture was stirred
at 0 °C for 1 h, water was added, and the mixture was
extracted with CH2Cl2 (3 × 15 mL). The combined organic
layers were dried over Na2SO4 and evaporated under reduced
pressure. The crude product was purified by flash chromatog-
raphy on silica gel (95:5 hexane/EtOAc) to give 73.9 mg (93%)
3
3
4-Val), 077 (d, J ) 6.9 Hz, 3H, γ-CH3-1-Val), 071 (d, J ) 6.9
Hz, 3H, γ-CH3-1-Val); MS (ESI): m/z: calcd for C23H28N2O6:
428,48 found: 429 [M + H].
2-Isop r op oxyp h en ol (40). To 270 mL of DMSO was added
KOH (28.3 g, 0.505 mol). After the mixture was stirred for 10
min catechol 39 (15.0 g, 0.136 mol) was added, followed after
10 min by 2-bromopropane (12.8 mL, 0.136 mol). Stirring was
continued for 30 min after which the KOH was filtered, and
1
of pure 26 as a colorless oil; Rf: 0.38 (9:1 hexane/EtOAc); H

Receptor Tyrosine Kinase Inhibitors
J ournal of Medicinal Chemistry, 2003, Vol. 46, No. 14 2929
the filtrate was diluted with 200 mL of CH2Cl2. The organic
layer was washed with saturated solution of NH4Cl (2×) and
NaCl (2×), dried over Na2SO4, and evaporated under reduced
pressure. The crude product was purified by flash chromatog-
raphy on silica gel (95:5 hexane/EtOAc) to give 11.0 g (53%)
200 mL), dried over MgSO4, and evaporated under reduced
pressure. The crude product was purified by flash chromatog-
raphy on silica gel (95:5 hexane/EtOAc) to give 2.9 g (66%,
two steps) of pure 43 as a colorless oil. Rf.: 0.44 (9:1 hexane/
1
4
EtOAc);. H NMR (250 MHz, CDCl3): δ 6.61 (d, J ) 2.8 Hz,
1H, H-6), 6.41 (d, 4J ) 2.8 Hz, 1H, H-4), 4.56-4.50 (m, 1H,
OCH(CH3)2), 3.77 (s, 3H, C5-OCH3), 3.72 (s, 3H, C2-OCH3),
1.34 (d, 3J ) 6.1 Hz, 6H, CH(CH3)2); 13C NMR (62.9 MHz,
CDCl3): δ 156.2 (C-5), 152.2 (C-3), 141.9 (C-2), 117.6 (C-1),
108.4 (C-6), 103.0 (C-4), 71.5 (OCH(CH3)2), 60.4 (C-5-OCH3),
55.6 (C-2-OCH3), 21.9 (OCH(CH3)2); MS (EI, 70 eV): m/z (%)
276/274 (41/42) [M+], 246/244 (20/20) [M+ - 2CH3], 234/232
(49/50) [M+ - C(CH3)2], 219/217 (97/100), 204/202 (64/65) [M+
- C(CH3)2 - 2CH3], 191 (12), 189/187 (49/38), 154 (7), 139 (11);
HRMS (70 eV): calcd for C11H15O379Br: 274.0205, found:
274.0199.
1
of pure 7 as a colorless oil. Rf: 0.42 (9:1 hexane/EtOAc); H
NMR (250 MHz, CDCl3): δ 6.95-6.91 (m, 1H, H-4), 6.87-6.78
(m, 3H, H-3, H-5, H-6), 5.81 (brs, 1H, OH), 4.59-4.50 (m, 1H,
3
OCH(CH3)2), 1.34 (d, J ) 6.1 Hz, 6H, OCH(CH3)2); 13C NMR
(62.9 MHz, CDCl3): δ 146.6 (C-2), 144.6 (C-1), 121.4 (C-4),
120.0 (C-5), 114.6 (C-6), 113.4 (C-3), 71.5 (OCH(CH3)2), 22.1
(OCH(CH3)2); MS (EI, 70 eV): m/z (%) 152 (13) [M+], 110 (100)
[M+ - C(CH3)2]; HRMS (70 eV): calcd for C9H12O2: 152.0837,
found: 152.0846.
2-Hyd r oxy-3-isop r op oxyp h en yl Br om id e (41). Br2 (3.66
mL, 71.3 mmol) was added dropwise to a solution of tert-
butylamine (15.3 mL, 72.7 mmol) in 134 mL of dry toluene at
-30 °C. After being stirred for 0.5 h, the reaction mixture was
cooled to -60 °C. A solution of 2-isopropoxyphenol 40 (11.0 g,
72.3 mol) in 9 mL of CH2Cl2 was added dropwise to the reaction
mixture, and the mixture was warmed to room temperature
over a period of 5 h. The reaction mixture was treated with
10% Na2S2O3, washed with brine, and dried over Na2SO4. The
crude product was purified by flash chromatography on silica
gel (95:5 hexane/EtOAc) to give 6.5 g (39%) of pure 41 as a
colorless oil. Rf: 0.63 (3:1 hexane/EtOAc); 1H NMR (250 MHz,
6-Ben zyl-1,4-d im eth oxy-2-isop r op oxyben zen e (45). To
a solution of 2,5-dimethoxy-3-isopropoxyphenyl bromide 43
(200 mg, 0.728 mmol) in 1.9 mL of THF at -78 °C was added
dropwise n-BuLi (2.5 M in hexane) (320 µL, 0.8 mmol). After
the mixture was stirred at this temperature for 30 min, CuI
(169 mg, 0.887 mmol) was added. The solution was slowly
warmed to -40 °C, then cooled to -78 °C, and a solution of
benzyl bromide (115 µL, 0.97 mmol) in 0.6 mL of THF was
added. The mixture was warmed to room temperature, diluted
with 20 mL of ether and 20 mL of water, and extracted with
ether (1 × 20 mL). The combined organic layers were washed
with a saturated solution of Na2CO3 until the aqueous layer
was not blue, dried over Na2SO4, and evaporated under
reduced pressure. The crude product was purified by flash
chromatography on silica gel (95:5 hexane/EtOAc) to give 114
mg (55%) of 45 as a colorless oil; 1H NMR (250 MHz, CDCl3):
δ 7.27-7.16 (m, 5H, H-2′(2), H-3′(2), H-4′), 6.37 (d, 4J ) 2.8
Hz, 1H, H-5), 6.23 (d, 4J ) 3.0 Hz, 1H, H-3), 4.56-4.46 (m,
1H, OCH(CH3)2), 3.95 (s, 2H, H-7), 3.70 (s, 3H, C-1-OCH3),
3.66 (s, 3H, C-4-OCH3), 1.36 (d, 3J ) 6.1 Hz, 6H, OCH(CH3)2);
13C NMR (62.9 MHz, CDCl3): δ 155.6 (C-4), 151.5 (C-2), 141.0
(C-1′), 135.2 (C-1), 128.9, 128.6, 128.6 (C-6, C-2′(2), C-3′(2)),
125.9 (C-4′), 106.5 (C-5), 101.0 (C-3), 70.7 (OCH(CH3)2), 60.4
(C-1-OCH3), 55.4 (C-4-OCH3), 36.2 (C-7), 22.1 (OCH(CH3)2);
MS (EI, 70 eV): m/z (%) 286 (93) [M+], 244 (87) [M+ - C3H6],
229 (100) [M+ - C4H9], 183 (9), 153 (8), 141 (8), 91(16); HRMS
(70 eV): calcd for C18H22O3: 286.1569, found: 286.1573.
3
4
CDCl3): δ 7.06 (dd, J ) 8.1 Hz, J ) 1.3 Hz, 1H, H-6), 6.79
3
4
3
(dd, J ) 8.1 Hz, J ) 1.2 Hz, 1H, H-4), 6.70 (ψt, J ) 8.1 Hz,
1H, H-5), 6.02 (s, 1H, OH), 4.59-4.53 (m, 1H, OCH(CH3)2),
1.36 (d, 3J ) 6.1 Hz, 6H, OCH(CH3)2); 13C NMR (62.9 MHz,
CDCl3): δ 145.3 (C-3), 144.2 (C-2), 124.8 (C-6), 120.5 (C-5),
112.4 (C-4), 108.3 (C-1), 72.3 (OCH(CH3)2), 22.1 (OCH(CH3)2);
MS (EI, 70 eV): m/z (%) 232/230 (21/22) [M+], 190/188 (97/
100) [M+ - C(CH3)2], 108 (5); HRMS (70 eV): calcd for
C9H11O281Br: 231.9922, found: 231.9920.
2-Br om o-6-isop r op oxy-p-ben zoqu in on e (42). To a solu-
tion of 2-hydroxy-3-isopropoxyphenyl bromide 8 (5.0 g, 22.0
mmol) in 138 mL of DMF was added 80.6 mg (0.24 mmol) of
salcomine (N,N′-bis(salicylidene)ethylendiamine cobalt(II) salt
hydrate). The solution was stirred at room temperature under
an oxygen atmosphere for 16 h, diluted with 200 mL of ether
and 100 mL of water, and extracted with ether (3 × 150 mL).
The combined organic layers were dried over Na2SO4 and
evaporated under reduced pressure. The crude product was
purified by flash chromatography on silica gel (9:1 hexane/
EtOAc) to give 3.9 g (73%) of pure 42 as a yellow oil. Rf: 0.47
(3:1 hexane/EtOAc); 1H NMR (250 MHz, CDCl3): δ 7.19 (d, 4J
) 2.1 Hz, 1H, H-3), 5.93 (d,4J ) 2.1 Hz, 1H, H-5), 4.58-4.48
(m, 1H, OCH(CH3)2), 1.36 (d, 3J ) 6.1 Hz, 6H, OCH(CH3)2);
13C NMR (62.9 MHz, CDCl3): δ 184.7 (C-1), 174.5 (C-4), 162.3
(C-6), 138.0 (C-3), 126.5 (C-2), 107.8 (C-5), 72.5 (OCH(CH3)2),
21.7 (OCH(CH3)2); MS (EI, 70 eV): m/z (%) 246/244 (3/3) [M+],
232/230 (19/20), 190/188 (97/100), 108 (5), 79 (5); HRMS (70
eV): calcd for C9H9O379Br: 243.9735, found: 243.9741.
2,5-Dim eth oxy-3-isop r op oxyp h en yl Br om id e (43). To
a suspension of 2-bromo-6-isopropoxy-p-benzoquinone 42 (3.9
g, 16.1 mmol) in 42 mL of ethanol at 0 °C was added NaBH4
(1.7 g, 45.4 mmol) in portions. After addition was complete,
the reaction mixture was allowed to stir for 4 h at 0 °C. The
excess of NaBH4 was destroyed by careful addition of 1 M HCl.
The reaction mixture was diluted with 20 mL of water and 50
mL of ether and extracted with ether (2 × 50 mL). The
combined organic layers were washed with water (2 × 30 mL),
dried over Na2SO4, and evaporated under reduced pressure
to give 3.3 g of the hydroquinone, which was subjected to the
O-methylation without further purification.
3-Be n zyl-2,5-d im e t h oxy-2,5-cycloh e xa d ie n e -1,4-d i-
on e (47). To a solution of Fremy’s salt (ON(SO3K)2, 77.4 mg,
0.289 mmol) in 1.8 mL of Na2CO3 (15% solution in water) were
added Aliquat 336 (tricaprylmethylammonium chloride) (51.6
mg) in 710 µL benzene and phenol 46 (30.0 mg, 0.115 mmol)
in 1.4 mL of benzene. The reaction mixture was stirred at room
temperature for 3 h and diluted with ether and water. The
water layer was extracted with ether (3 × 10 mL). The
combined organic layers were dried over Na2SO4 and evapo-
rated under reduced pressure. The crude product was purified
by flash chromatography on silica gel (5:1 hexane/EtOAc) to
give 27.4 mg (92%) of 47 as a yellow oil; Rf: 0.22 (3:1 hexane/
EtOAc); 1H NMR (250 MHz, CDCl3): δ 7.26-7.15 (m, 5H, H-2′-
(2), H-3′ (2), H-4′), 5.71 (s, 1H, H-6), 4.08 (s, 3H, C-2-OCH3),
3.79 (s, 3H, C-5-OCH3), 3.77 (s, 2H, H-7); 13C NMR (62.9 MHz,
CDCl3): δ 183.6 (C-4), 181.9 (C-1), 158.7 (C-5), 155.7 (C-2),
138.9 (C-1′), 129.0, 128.5, 128.4 (C-2′(2), C-3′(2), C-4′), 126.3
(C-1), 105.4 (C-6), 61.3 (C-5-OCH3), 56.4 (C-2-OCH3), 28.7
(C-7); MS (EI, 70 eV): m/z (%) 258 (45) [M+], 227 (100) [M+
OCH3], 215 (66), 197 (5), 183 (10), 167 (6), 115 (14), 91 (51);
HRMS (70 eV): calcd for C15H14O4: 258.0892, found: 258.0896.
-
3-Ben zyl-2-h yd r oxy-5-m eth oxy-2,5-cycloh exa d ien e-1,4-
d ion e (48) To a solution of quinone 47 (27.4 mg, 0.106 mmol)
in 5.8 mL of MeOH and 3.4 mL of H2O was added water until
the starting material was dissolved. To the reaction mixture
was added 9 drops of 1 N KOH, and the dark red solution was
stirred at room temperature for 12 h. The reaction mixture
was diluted with water and ether, and solid NH4Cl was added.
The aqueous solution was extracted with ether until decolori-
zation. The combined organic layers were dried over Na2SO4
and evaporated under reduced pressure. The crude product
To a stirred mixture of 2-bromo-6-isopropoxyhydroquinone
(3.3 g) in 6 mL of ethanol, dimethyl sulfate (8.25 mL, 86.5
mmol), and NaHSO3 (0.23 g, 2.18 mmol) at 0 °C was added a
solution of NaOH (3.3 g, 82.5 mmol) in 18.5 mL of water over
30 min. The mixture was warmed to 70-80 °C for 4 h and
diluted with water and ether. The aqueous solution was
extracted with ether (3 × 80 mL). The combined organic layers
were washed with 2 N NaOH (2 × 200 mL) and water (2 ×

2930 J ournal of Medicinal Chemistry, 2003, Vol. 46, No. 14
Kissau et al.
(7) Skobe, M.; Hawighorst, T.; J ackson, D. G.; Prevo, R.; J anes, L.;
Velasco, P.; Riccardi, L.; Alitalo, K.; Claffey, K.; Detmar, M.;
Induction of tumor lymphangiogenesis by VEGF-C promotes
breast cancer metastasis. Nat. Med. 2001, 7, 192-198.
(8) Giannis, A.; Ru¨bsam, F. Integrin antagonists and other low
molecular weight compounds as inhibitors of angiogenesis: New
drugs in cancer therapy. Angew. Chem. 1997, 109, 606-609;
Angew. Chem., Int. Ed. Engl. 1997, 36, 588-590 and refs
therein.
(9) Carmeliet, P.; J ain, R. K. Angiogenesis in cancer and other
diseases. Nature 2000, 407, 249-257 and refs therein.
(10) O’Connor, R.; Fennely, C.; Krause, D. Regulation of survival
signals from the insulin-like growth factor-I receptor. Biochem.
Soc. Trans. 2000, 28, 47-51.
(11) Hudziak, R. M.; Ullrich, A. Cell transformation potential of a
HER2 transmembrane domain deletion mutant retained in the
endoplasmic reticulum. J . Biol. Chem. 1991, 266, 24109-14115.
(12) Heldin, C.-H.; Ro¨nnstrand, L. Growth factor receptor in cell
transformation. In Oncogenes and Tumor Suppressors; Peters,
G., Vousden, K., Eds.; Oxford University Press, Cary, NC, 1997;
pp 55-85.
was purified by flash chromatography on silica gel (100:1
CHCl3/MeOH) to give 17.6 mg (68%) of 48 as a yellow solid
(during the contact with silicagel compound 48 turns purple).
1
Rf: 0.06 (3:1 hexane/EtOAc); mp 116 °C; H NMR (250 MHz,
CDCl3): δ 7.36-7.16 (m, 5H, H-2′(2), H-3′(2), H-4′), 5.83 (s,
1H, H-6), 3.82 (s, 3H, C-5-OCH3), 3.78 (s, 2H, H-7); 13C NMR
(62.9 MHz, CDCl3): δ 182.7 (C-4), 181.5 (C-1), 161.8 (C-5),
151.4 (C-2), 138.6 (C-1′), 129.0, 128.3, (C-2′(2), C-3′(2)), 126.2
(C-4′), 117.8 (C-3), 102.2 (C-6), 56.7 (C-5-OCH3), 28.3 (C-7);
MS (EI, 70 eV): m/z (%) 244 (100) [M+], 243 (42) [M+ - H],
229 (9) [M+ - CH3], 215 (10), 183 (23), 131 (8), 167 (6), 103
(5), 91 (29); HRMS (70 eV): calcd for C14H12O4: 244.0736,
found: 244.0733.
Tyr osin e Kin a se Assa y. The catalytic sections of the
corresponding tyrosine kinases were expressed as GST fusion
proteins (Klinik fu¨r Tumorbiologie, Freiburg, Germany).
In brief, the assays were performed in white 96 well
microtiter plates (Greiner) which had been coated at 4 °C for
16 h with 12.5 µg polyGlu-Tyr 4:1 sodium salt (Sigma) in 125
µL of PBS (phosphate buffered saline, Gibco) per well. The
coating solution was flicked off and the plates were washed
twice with PBST (PBS + 0.5% Tween-20). Kinases were
diluted in kinase buffer (100 mM HEPES pH 7.4, 100 mM
NaCl) and added at 20 ng protein in 50 µL of buffer. Stock
solutions of the test compounds were made in DMSO and
diluted to a final concentration of 5% DMSO with water
(working solution). Twenty five microliters of the working
solutions was added to each well. The kinase reaction was
started by addition of 25 µL of ATP (100 µM in 40 mM MnCl2)
using a multichannel pipet. Negative controls received 40 mM
MnCl2 without ATP. The plates were incubated at RT for 30
min. The microtiter plates were washed three times with
PBST. One hundred microliters of anti phosphotyrosine an-
tibody (PY20, Calbiochem) conjugated to horseradish peroxi-
dase (POD) was added at 1:5000 dilution in PBST + 0.1% BSA
(bovine serum albumin, Calbiochem) and incubated for 1 h at
RT. Plates were washed five times with PBST. For detection,
100 µL of POD chemiluminescence substrate (Roche) was
added. The resulting light emission was detected with an Orion
luminescence reader (Berthold).
(13) Cohen, P. Protein kinases- the major drug targets of the twenty-
first century? Nat. Rev. Cancer 2002, 1, 309-315.
(14) Bridges, A. J . Chemical Inhibitors of Protein Kinases. Chem.
Rev. 2001, 101, 2541-2571.
(15) Breinbauer, R.; Vetter, I.; Waldmann, H. From Protein Domains
to Drug Candidates - Natural Products as Guiding Principles
in the Design and Synthesis of Compound Libraries. Angew.
Chem., Int. Ed. 2002, 41, 2878-2890.
(16) Part of this work was published in preliminary form: Stahl, P.;
Kissau, L.; Mazitschek, R.; Giannis, A.; Waldmann, H. Natural
Product Derived Receptor Tyrosine Kinase Inhibitors: Identi-
fication of IGF1R, Tie-2 and VEGFR-3 Inhibitors. Angew. Chem.,
Int. Ed. 2002, 41,1174-1178.
(17) Kobayashi, J .; Madono, T.; Shigemori, H. Nakijiquinones C and
D, New Sesquiterpenoid Quinones with a Hydroxy Amino Acid
Residue from
a Marine Sponge Inhibiting c-erbB-2 Kinase.
Tetrahedron 1995, 51, 10867-10874.
(18) Stahl, P.; Kissau, L.; Mazitschek, R.; Huwe, A.; Furet, P.;
Giannis, A.; Waldmann, H. Total Synthesis and Biological
Evaluation of the Nakijiquinones. J . Am. Chem. Soc. 2001, 123,
11586-11593.
(19) See e. g. Bold, G.; Altmann, K.-H.; Frei, J .; Marc, L.; Manley, P.
W.; Traxler, P.; Wietfeld, B.; Bru¨ggen, J .; Buchdunger, E.;
Cozens, R.; Ferrari, S.; Furet, P.; Hofmann, F.; Martiny-Baron,
G.; Mestan, J .; Ro¨sel, J .; Sills, M.; Stover, D.; Acemoglu, F.; Boss,
E.; Emmenegger, R.; La¨sser, L.; Masso, E.; Roth, R.; Schlachter,
C.; Vetterli, W.; Wyss, D.; Wood, J . M. New Anilinophthalazines
as Potent and Orally Well Absorbed Inhibitors of the VEGF
Receptor Tyrosine Kinases Useful as Antagonists of Tumor-
Driven Angiogenesis. J . Med. Chem. 2000, 43, 2310-2323 and
refs therein.
(20) Ishizaki, M.; Ozaki, K.; Kanematsu, A.; Isoda, T. Hoshino,
Synthetic approaches toward spiro[2,3-dihydro-4H-1-benzopy-
ran-4,1′-cyclohexan]-2-one derivatives via radical reactions: total
synthesis of (()-lycoramine. O. J . Org. Chem. 1993, 58, 3877-
3885.
(21) Wakamatsu, T.; Nishi, T.; Ohnuma, T.; Ban, Y. A convenient
synthesis of juglone via neutral salcomine oxidation. Synth.
Commun. 1984, 14, 1167-1173.
(22) Miyaura, N.; Suzuki, A. Palladium-Catalyzed Cross-Coupling
Reactions of Organoboron Compounds. Chem. Rev. 1995, 95,
2457-2483.
Ack n ow led gm en t. This research was supported by
the Fonds der Chemischen Industrie. L.K. wishes to
acknowledge the support by the Studienstiftung des
deutschen Volkes (Promotionsstipendium), the Fonds
der Chemischen Industrie (Kekule-Stipendium), and
e-fellows.net (Online-Stipendium). R.M. is grateful to
the Land Baden-Wu¨rttemberg for a sholarship from the
Landesgraduiertenfo¨rderung.
Su p p or tin g In for m a tion Ava ila ble: Further experimen-
tal details are available free of charge via the Internet at http://
pubs.acs.org.
(23) Dolan, S. C.; Millan, J . M. A New Method for the Deoxygenation
of Tertiary and Secondary Alcohols. J . Chem. Soc., Chem.
Commun. 1985, 22, 1588-1589.
(24) Genot, A.; Florent, J .-C.; Monneret, C. Anthracyclinone. 3. Chiral
pool synthesis of anthracyclinones via tetralin intermediates.
J . Org. Chem. 1987, 52, 1057-1063.
(25) Takai, K.; Nitta, K.; Utimoto, K. Simple and selective method
for aldehydes (RCHO) fwdarw. (E)-haloalkenes (RCH: CHX)
conversion by means of a haloform-chromous chloride system.
J . Am. Chem. Soc. 1986, 108, 7408-7410.
(26) Harada, T.; Katsuhira, T.; Hattori, K.; Oku, A. Stereochemistry
in carbenoid formation by bromine/lithium and bromine/zinc
exchange reactions of gem-dibromo compounds. Tetrahedron
1994, 50, 7987-8002.
Refer en ces
(1) Hubbard, S. R.; Till, J . H. Protein Tyrosine Kinases: Structure
and Function. Annu. Rev. Biochem. 2000, 69, 373-98.
(2) Yancopoulos, G. D.; Davis, S.; Gale, N. W.; Rudge, J . S.; Wiegand,
S. J .; Holash, J . Vascular-specific growth factors and blood vessel
formation. Nature 2000, 407, 242-248.
(3) Hiratsuka, S.; Maru, Y.; Okada, A.; Seiki, M.; Noda, T.; Shibuya,
M. Involvement of Flt-1 tyrosine kinase (vascular endothelial
growth factor receptor-1) in pathological angiogenesis. Cancer
Res. 2001, 61, 1207-1213.
(4) Kubo, H.; Fujiwara, T.; J ussila, L.; Hashi, H.; Ogawa, M.;
Shimizu, K.; Awane, M.; Sakai, Y.; Takabayashi, A.; Alitalo, K.;
Yamaoka, Y.; Nishikawa, S. I. Involvement of vascular endot-
helial growth factor receptor-3 in maintenance of integrity of
endothelial cell lining during tumor angiogenesis. Blood 2000,
96, 546-553.
(5) Stratmann, A.; Acker, T.; Burger, A. M.; Amann, K.; Risau, W.;
Plate, K. H. Differential inhibition of tumor angiogenesis by
TIE2 and vascular endothelial growth factor receptor-2 dominant-
negative receptor mutants. Int. J . Cancer 2001, 91, 273-282.
(6) Stacker, S. A.; Caesar, C.; Baldwin, M. E.; Thornton, G. E.;
Williams, R. A.; Prevo, R.; J ackson, D. G.; Nishikawa, S.; Kubo,
H.; Achen, M. G. VEGF-D promotes the metastatic spread of
tumor cells via the lymphatics. Nat. Med. 2001, 7, 186-191.
(27) Kaufman, T. S. Synthesis and mass spectral data of four
potential biomarkers related to the C19 tricyclanes found
Australian oils and puget sound sediments. Synth. Commun.
1995, 25, 1205-1221.
(28) See ref 1 and refs cited therein.
(29) Olson, G. L.; Cheung, H.-C.; Morgan, K.; Saucy, G. A new
synthesis of alpha.-tocopherol. J . Org. Chem. 1980, 45, 803-
805.
(30) A homology model of the ATP binding site of KDR has been used
to evaluate the anilinophthalazine-class of RTK Inhibitors. See
ref 18.

Receptor Tyrosine Kinase Inhibitors
J ournal of Medicinal Chemistry, 2003, Vol. 46, No. 14 2931
(31) PDB entries 1FGI and 2FGI
(44) Arnold, L. D.; Calderwood, D. J .; Dixon, R. W.; J ohnston, D. N.;
Kamens, J . S.; Munschauer, R.; Rafferty, P.; Ratnofsky, S. E.
Pyrrolo[2,3-d]pyrimidines Containing an Extended 5-Substituent
as Potent and Selective Inhibitors of lck I. Bioorg. Med. Chem.
Lett. 2000, 10, 2167-2170.
(45) Burchat, A. F.; Calderwood, D. J .; Hirst, G. C.; Holman, N. J .;
J ohnston, D. N..; Munschauer, R.; Rafferty, P.; Tometzki, G. B.
Pyrrolo[2,3-d]pyrimidines Containing an Extended 5-Substituent
as Potent and Selective Inhibitors of lck. Bioorg. Med. Chem.
Lett. 2000, 10, 2171-2174.
(46) Calderwood, D. J .; J ohnston, D. N.; Munschauer, R.; Rafferty,
P. Pyrrolo[2,3-d]pyrimidines Containing Diverse N-7 Substitu-
ents as Potent Inhibitors of Lck. Bioorg. Med. Chem. Lett. 2002,
12, 1683-1686.
(47) Burchat, A. F.; Calderwood, D. J .; Friedman, M. M.; Hirst, G.
C.; Li, B.; Rafferty, P.; Ritter, K.; Skinner, B. S. Pyrazolo[3,4-
d]pyrimidines Containing an Extended 3-Substituent as Potent
Inhibitors of Lck - a Selective Insight. Bioorg. Med. Chem. Lett.
2002, 12, 1687-1690.
(48) Stieber, F.; Mazitschek, R.; Soric, N.; Giannis, A.; Waldmann,
H. Traceless Solid-Phase Synthesis of 2-Aminothiazoles: Recep-
tor Tyrosine Kinase Inhibitors with Dual Selectivity for Tie-2
and VEGFR-2. Angew. Chem., Int. Ed. 2002, 41, 4757-4761.
(49) Zhou, B.-N.; J ohnson, R. K.; Mattern, M. R.; Fisher, P. W.;
Kingston, D. G. I.; The First Naturally Occurring Tie 2 Inhibitor.
Org. Lett. 2001, 3, 4047-4049.
(50) A homology model of the IGF1R kinase domain was also created.
The conclusions drawn with respect to the binding modes of the
nakijiquinone analogues were essentially the same as those
derived from the available crystal structures.
(32) Marti-Renom, M. A.; Stuart, A. C.; Fiser, A.; Sanchez, R.; Melo,
F.; Sali, A. Comparative Protein Structure Modeling of Genes
and Genomes. Annu. Rev. Biophys. Biomol. Struct. 2000, 29,
291-325.
(33) This assumption would have also permitted the choice of the
IGF1R or IRK structure as a templates. Indeed, this possibility
was exploited, and the resulting 3D models were also used in
the evaluation of the nakijiquinone analogues. However, the
sequence homology to the VEGFs and Tie-2 is lower.
(34) Davis, A. M.; Teague, S. J . Hydrogen Bonding, Hydrophobic
Interactions, and Failure of the Rigid Receptor Hypothesis.
Angew. Chem., Int. Ed. 1999, 38, 736-749.
(35) Mohammadi, M.; McMahon, G.; Sun, L.; Tang, C.; Hirth, P.; Yeh,
B. K.; Hubbard, S. R.; Schlessinger, J . Structure of the Tyrosine
Kinase Domain of Fibroblast Growth Receptor in Complex with
Inhibitors. Science 1997, 276, 955-960.
(36) Morgenstern, B. Improvement of the segment-to-segment ap-
proach to multiple sequence alignment. Bioinformatics 1999, 15,
211-218.
(37) Sali, A.; Potterton, L.; Yuan, F.; van Vlijmen, H.; Karplus, M.
Evaluation of Comparative Protein Modeling by MODELLER.
Proteins 1995, 23, 318-326.
(38) Modeller 4.0 Manual. For additional references and information
please refer to Sali@rockvax.rockefeller.edu.
(39) Hooft, R. W. W.; Vriend, G.; Sander, C.; Abola, E. E. Errors in
protein structure. Nature 1996, 381, 272.
(40) WitnotP is a molecular modeling software developed by and
licensed from Novartis AG, Basel, Switzerland. For further
information and licensing conditions please contact Dr. Armin
Widmer at Novartis AG.
(41) Brooks, B. R.; Bruccoleri, R. E.; Olafson, B. D.; States, D. J .;
Swaminathan, S.; Karplus, M. CHARMM: A Program for
Macromolecular Energy, Minimization, and Dynamics Calcula-
tions. J . Comput. Chem 1983, 4, 187-217.
(42) For a detailed description of how this figure was created, see
Experimental Section.
(51) McTigue, M. A.; Wickersham, J . A.; Pinko, C.; Showalter, R. E.;
Parast, C. V.; Tempczyk-Russell, A.; Gehring, M. R.; Mrocz-
kowski, B.; Kan, C.-C.; Villafranca, J . E.; Appelt, K. Crystal
structure of the kinase domain of human vascular endothelial
(43) Knighton, D. R.; Zheng, J . H.; Ten Eyck, L. F.; Ashford, V. A.;
Xuong, N. H.; Taylor; S. S.; Sowadski, J . M. Crystal Structure
of the Catalytic Subunit of Cyclic Adenosine Monophosphate-
Dependent Protein Kinase. Science 1991, 253, 407-414.
growth factor receptor 2: a key enzyme in angiogenesis.
Structure 1999, 7, 319-330.
J M0307943