Discovery of quinolinediones exhibiting a heat shock response and angiogenesis inhibition

Article abstract of DOI:10.1021/jm7014099

A series of substituted quinoline-5,8-diones were synthesized and evaluated as inhibitors of the chaperone protein Hsp90 using two assays: competition for binding to C-terminal ATP-binding site and competition for binding to N-terminal ATP-binding site. I

Full text of DOI:10.1021/jm7014099

2492
J. Med. Chem. 2008, 51, 2492–2501
Discovery of Quinolinediones Exhibiting a Heat Shock Response and Angiogenesis Inhibition
Robert H. J. Hargreaves,† Cynthia L. David,‡ Luke J. Whitesell,§ Daniel V. LaBarbera,† Akmal Jamil,† Jean C. Chapuis,† and
Edward B. Skibo*,†
Department of Chemistry and Biochemistry, Arizona State UniVersity, Tempe, Arizona 85287-1604, Steele Memorial Children’s Research
Center, UniVersity of Arizona, Tucson, Arizona 85724, and Whitehead Institute, Nine Cambridge Center, Cambridge, Massachusetts 02142
ReceiVed NoVember 8, 2007
A series of substituted quinoline-5,8-diones were synthesized and evaluated as inhibitors of the chaperone
protein Hsp90 using two assays: competition for binding to C-terminal ATP-binding site and competition
for binding to N-terminal ATP-binding site. In addition, the ability of the compounds to induce the heat
shock response was determined using a reporter fibroblast cell line. Of all the compounds assayed, only
6-aziridinyl-2-biphenylquinoline-5,8-dione induced a heat shock response and did so without interacting at
the ATP binding sites of Hsp90. COMPARE analysis was carried out on quinoline-5,8-diones active in the
National Cancer Institute’s 60-cell line screen with the goal of discovering quinoline-5,8-dione structures
that interact with other cellular targets (molecular targets) important for cancer chemotherapy. COMPARE
analysis led to the discovery of a combretastatin-like quinoline-5,8-dione structure that, in fact, inhibited
angiogenesis.
Chart 1
Introduction
Heat shock protein 90 (Hsp90a) is emerging as an important
target for cancer chemotherapy because of its role in chaperoning
proteins involved in signal transduction pathways that control
tumor cell proliferation and survival.1–3 The prototype inhibitor
of Hsp90 is geldanamycin shown in the inset of Chart 1. This
complex natural product has been the subject of total synthesis4,5
and analogue development.6 In addition, the synthesis and
screening of purine libraries has afforded small molecules
capable of geldanamycin-like binding to Hsp90.7–9 This labora-
tory has approached Hsp90 inhibitor design by molecular
modeling and library synthesis.10 As illustrated in Chart 1, the
substituted quinoline quinone ring system has some of the
structural elements of geldanamycin. In fact, these studies
revealed that some quinoline quinones possess geldanamycin-
like cytostatic and cytotoxic activity, based on COMPARE
analysis,11,12 although none actually interact with Hsp90. This
report further investigates the ability of diverse quinoline
quinones (1 and 2) in Chart 1 to interact with Hsp90 and
influence protein maturation. One analog of 2 was found to
induce the heat shock response consistent with an ability to
impair protein homeostasis in cells without directly interacting
with the known drug-binding sites on Hsp90.
of 1 and 2, COMPARE analysis 11,12 was carried using National
Cancer Institute 60-cell line mean graph data for each com-
pound. Thus, the pattern of cytostatic and cytotoxic parameters,
GI50 (growth inhibition 50%), TGI (total growth inhibition), and
LC50 (lethal concentration 50%) on the mean graphs of 1 and
2 was compared with those of over 38000 compounds in the
National Cancer Institute’s archives. The goals were to find a
compound of known mechanism of action and perhaps a
molecular target important for cancer chemotherapy. As a result
of this screening effort, several new lead compounds directed
toward novel molecular targets were discovered. We followed
up on one of these leads and document an angiogenesis inhibitor
based on the quinoline quinone ring system.
The quinoline quinone system is also a component in a
number of antitumor agents, most notably lavendamycim and
streptonigrin shown in Chart 2. Therefore, substituted quinoline
quinones could interact with targets important for cancer
chemotherapy other than Hsp90. To assess other cellular targets
* To whom correspondence should be addressed. Phone: 480-965-3581.
Fax: 480-965-2747. E-mail: eskibo@asu.edu.
†
Arizona State University.
University of Arizona.
Whitehead Institute.
‡
§
a Abbreviations: Hsp90, heat shock protein; GI50, growth inhibition 50%;
TGI, total growth inhibition; LC50, lethal concentration 50%; gHMBC,
gradient heteronuclear multiple bond coherence; DMSO, dimethyl sulfoxide;
ATP, adenosine triphosphate; EGFP, enhanced green fluorescent protein;
mRNA, messenger ribonucleic acid; FGF, fibroblast growth factor; HUVEC,
human umbilical vein endothelial cell; NSC, Cancer Chemotherapy National
Service Center; FGFRs, fibroblast growth factor receptors; TLC, thin layer
chromatography.
Results and Discussion
Synthesis. The preparation of the target compounds was
initiated with the Friedlander quinoline synthesis, Scheme 1.
10.1021/jm7014099 CCC: $40.75
2008 American Chemical Society
Published on Web 03/26/2008

Quinolinediones Exhibiting a Heat Shock Response
Journal of Medicinal Chemistry, 2008, Vol. 51, No. 8 2493
Table 1. Aminoquinolinedione Derivatives
Chart 2
cmpd
2-substituent
phenyl
phenyl
phenyl
phenyl
phenyl
6-substituent
7-substituent
1a
1b
1c
1d
1e
N-pyrrolidino
benzylamino
H
allylamino
3-methoxy
propylamino
H
H
H
benzylamino
H
H
1f
phenyl
3-methoxy
propylamino
H
1g
1h
1i
3,4,5-trimethoxy
phenyl
3,4,5-trimethoxy
phenyl
benzylamino
H
benzylamino
H
phenyl
3,4,5-trimethoxy
benzylamino
H
1j
phenyl
3,4,5-trimethoxy
benzylamino
H
1k
3,4,5-trimethoxy
phenyl
methoxy
Table 2. Aziridinyl Quinolinedione Derivatives
cmpd
2-substituent
methyl
methyl
phenyl
3,4,5-trimethoxy
phenyl
3-substituent 6-substituent 7-substituent
2a
2b
2c
2d
H
aziridinyl
H
Scheme 1
methoxy
methoxy
H
H
H
aziridinyl
aziridinyl
H
aziridinyl
2e
2f
2g
2h
biphenyl
phenyl
phenyl
H
aziridinyl
aziridinyl
aziridinyl
H
H
H
H
Methyl
H
H
phenyl
aziridinyl
Hsp90 Inhibition Assays. Assays were carried out on 1 and
2 to assess their ability to interact with previously described
drug-binding sites in the C-terminus and N-terminus of Hsp90
and their ability to induce a cellular heat shock response in a
manner similar to that of known Hsp90 inhibitors.
Competition for Binding to C-Terminal ATP (Adeno-
sine Triphosphate) Binding Site on Hsp90 Using Immobi-
lized Novobiocin.14 Compound series 1 and 2 were formulated
in DMSO (dimethyl sulfoxide) at 35 mM immediately prior to
experiments. Compound solutions were added to whole cell
lysates at a final nominal concentration of 100 µM. After drug
addition, lysates were incubated on ice for 30 min and then
sepharose beads, on which novobiocin had been immobilized,
were added followed by incubation at 4 °C with rocking for 60
Thus, the reaction of 2-amino-3,6-dimethoxybenzaldehyde13
with the appropriate ketones afforded substituted-5,8-dimethox-
yquinolines (3). Oxidative demethylation of 3, to produce
substituted quinoline-5,8-diones (4), was carried out in high
yields with ceric ammonium nitrate, Scheme 1. Treatment of
derivatives 4 without a 3-substituent with primary amines or
aziridine afforded a mixture of the 6- and 7-amino derivatives
1 and 2, Tables 1 and 2. In all instances, the isolated yields of
the 6-isomer were higher than that of the 7-isomer. We observed
that an electron-releasing substituent (methyl and methoxy)
at the 3-position directed amination predominately to the
7-position. The 3-substituent releases electrons to the 8-carbonyl,
so the Michael-type addition of the amine occurred at the
7-position with negative charge development occurring at the
5-carbonyl. Both the 6- and the 7-isomers of 1 and 2 were
evaluated as Hsp90 inhibitors and as cytotoxic and cytostatic
agents. Confirmation of the structures of the 6- and 7-isomers
was readily carried out using gHMBC (Gradient Heteronuclear
Multiple Bond Coherence), as illustrated in Figure 1.
Figure 1. gHMBC spectra of the 6- and 7-isomers of aziridinyl
quinolinediones 2g and 2h, respectively.

2494 Journal of Medicinal Chemistry, 2008, Vol. 51, No. 8
HargreaVes et al.
Figure 2. Heat shock response in 3T3Y9/B12 cells. Positive controls consisted of geldanamycin at 1 and 5 µM. Final drug concentrations were
5 µM.
Table 3. In Vitro Assay Results of Aminoquinolinedione Derivatives
min. The novobiocin beads were then washed, and the proteins
mid-range
mid-range
mid-range
were eluted and run on gels, followed by blotting for Hsp90. A
positive result consisted of the disappearance of the Hsp90 band,
indicating that the drug was able to compete for Hsp90 binding
with novobiocin. None of the compounds displayed detectable
competition for C-terminal binding of Hsp90. Authentic novo-
biocin was used at a concentration of 1 mM (partial inhibition
of binding) and 10 mM (complete inhibition) as positive
controls. It was not possible to test the candidate compounds at
these concentrations due to limitation in their aqueous solubility.
Thus, it was not possible to compare the candidate compounds
to novobiocin directly, but it is safe to say that they demonstrated
no inhibitory activity at a concentration far above that which
resulted in biological effects in cells.
cmpd
GI50 (range)
TGI and range
LC50 and range
1a
1b
1c
1d
1e
1f
1g
1h
1i
-4.35 (1.77)
-5.77 (3.92)
-4.48 (3.01)
-5.72 (3.72)
-5.45 (2.67)
-5.26 (2.42)
inactive
-5.57 (2.95)
-5.63 (1.96)
-4.24 (1.87)
inactive
-4.81 (3.08)
inactive
-4.85 (4.00)
-4.72 (2.08)
-4.54 (2.22)
inactive
-4.20 (2.15)
-4.81 (1.99)
inactive
inactive
-4.18 (1.84)
inactive
-4.36 (1.77))
-4.18 (1.38)
-4.15 (1.58)
inactive
inactive
-4.18 (1.34)
inactive
1j
Table 4. In Vitro Assay Results of Aziridinylquinolinedione Derivatives
mid-range
mid-range
mid-range
Competition for Binding to N-Terminal ATP-Binding
Site on Hsp90 Using Immobilized Geldanamycin.15 Candidate
compounds were formulated in DMSO immediately prior to
experiment at 35 mM. Compound solutions were added to whole
cell lysates at a final nominal concentration of 100 µM. This
assay was carried out as described above for the C-terminal
ATP binding assay, except that geldanamycin was derivatized
and immobilized on agarose beads as previously described. A
positive result again consisted of disappearance of the Hsp90
band, indicating that the drug was able to compete effectively
with the immobilized geldanamycin for Hsp90 binding. None
of the compounds displayed N-terminal binding at the ATP
binding site of Hsp90 in this assay. Because the candidate
compounds were tested at a 6-fold higher concentration than
the positive control (geldanamycin), it is safe to conclude that
the compounds are at least 6-fold less active than geldanamycin
and demonstrated no inhibitory activity at a concentration far
above that which resulted in biological effects in cells.
Heat Shock Response in Fibroblasts Stably Transfected
with GFP Reporter Construct.16 Candidate compounds were
freshly formulated in DMSO. Compound solutions were added
to 3T3Y9/B12 reporter cells that are stably transfected with a
heat shock-responsive EGFP (enhanced green fluorescent pro-
tein) construct and incubated overnight. Final compound
concentrations were 5 µM. Treated wells were evaluated the
following day using fluorescence microscopy and a microplate
fluorometer. The positive controls consisting of geldanamycin
at 1 and 5 µM showed fluorescence exceeding control values
(Figure 2). Compound 2e showed modest activity in this assay
although this compound did not bind to Hsp90. This result is
most consistent with a secondary or nonspecific effect unrelated
to a direct action on the chaperone machinery.
cmpd
GI50 (range)
TGI and range
LC50 and range
2a
2b
2c
2d
2e
2g
2h
-6.29 (2.14)
-6.56 (1.47)
-6.28 (2.30)
-6.23 (2.15)
-5.10 (2.61)
-7.56 (2.07)
-7.07 (3.21)
-5.69 (1.29)
-6.08 (1.41)
-5.78 (1.88)
-5.75 (1.26)
-4.57 (2.32)
-6.49 (2.47)
-5.48 (4.00)
-5.10 (1.74)
-5.36 (2.27)
-5.31 (2.01)
-5.26 (2.22)
-4.19 (1.22)
-5.69 (4.00)
-4.75 (2.78)
GI50 and TGI, which are the concentrations of drug required
for 50% growth inhibition and total growth inhibition, respec-
tively. The cytotoxic parameter is the LC50, which is the
concentration required for 50% cell kill. These in vitro data
were obtained under the In Vitro Cell Line Screening Project
at the National Cancer Institute.11,17–20 The log mean values
for GI50, TGI, and LC50 in 60 cell lines are provided in Tables
3 and 4 along with the log range (the maximum difference
between the least-sensitive and the most-sensitive cell lines).
The log range parameter has been used to gain insights into the
selectivity of antitumor agents.21 The “inactive” category
pertains to compounds with log mean values of -4 with log
range values of 0 to 1.
COMPARE analysis11,12 provided insights into what might
be the molecular target of the active analogues of 1 and 2
exhibiting high selectivity (compounds with bold borders in
Tables 3 and 4). The criteria for choosing active compounds
for COMPARE analysis include: log mean values <-4 with
log range values of >2, and with high selectivity for 10 or more
cell lines and/or selectivity for specific histologies. The mean
graphs of compounds meeting these criteria have distinctive
patterns amenable to meaningful COMPARE analysis. In
contrast, mean graphs with low log range values are essentially
flat and will correlate with other such mean graphs without
meaningful target information.
Cytotoxic and Cytostatic Activity, COMPARE Anal-
ysis. The mean cytostatic and cytotoxic parameters of 1 and 2
are listed in Tables 3 and 4. The cytostatic parameters include

Quinolinediones Exhibiting a Heat Shock Response
Journal of Medicinal Chemistry, 2008, Vol. 51, No. 8 2495
Chart 4
Chart 3
The pattern of cytostatic and cytotoxic parameters (GI50, TGI,
and LC50 mean graphs) of selected analogues of 1 and 2 were
compared with those of over 38000 compounds in the National
Cancer Institute’s archives. The goal was to find a compound
of known mechanism of action and perhaps a molecular target
that correlated well with the mean graphs of the selective
analogues of 1 and 2. Thus, these analogues were used as a
seed compound in the following NCI databases: Standard Agent
database, Synthetic Agent database, and the Molecular Targets
Database. The Standard Agent Database consists of 175 drugs
with cancer treatment applications as well as new compounds
with a high level of interest.17 Molecular target levels (e.g.,
topoisomerase II, DT-diaphorase, proteins of unknown function,
etc.) have been measured in the 60-cell line panel using mRNA
(messenger ribonucleic acid) expression and enzyme activity
levels.22
The results of COMPARE analysis are discussed below for
compounds 1e, 1h, and 2g,h. The rest of the active compounds
did not exhibit significant correlations with known molecular
targets. It is noteworthy that COMPARE analysis reveled that
none of the aziridinyl quinone derivatives were dependent on
DT-diaphorase for activity. This two-electron reducing enzyme
typically activates aziridinyl quinones as cytotoxic alkylating
agents.
Compound 1e, and to a lesser degree the isomer 1f, correlated
well with the naphthoquinone natural product 5 (NSC 679186;23
NSC, Cancer Chemotherapy National Service Center) as well
as with fibroblast growth factor (FGF-2), Chart 3. The natural
product has been reported to be cytotoxic and had a moderate
correlation with the dual specificity protein phosphatase CDC
25A.24 There are obvious structural similarities between 1e and
5, particularly the oxygen-containing side chain of both
compounds. The FGF-2 molecular target plays an important role
in cell growth and has been the target of antitumor agent
design.25–27 The fibroblast growth factors interact with the
receptor tyrosine kinases (FGFRs) in conjunction with heparin-
binding coreceptors. Indeed, reported FGF-2 inhibitors are
heterocycles tethered to glucuronide derivatives.25 Perhaps the
oxygenated side chain of 1e plays a role in the FGF-2 inhibition
through mimicry of a carbohydrate residue.
our COMPARE analyses. Significantly, this molecular target
is involved in mitosis29 and, therefore, 1h is mechanistically
similar to 6. In addition, the structures of both compounds are
nearly superimposable, except for the benzylamino group of
1h. This additional substituent was considered responsible for
the lower potency of 1h compared to 6. In the following section,
we provide evidence that validates these findings.
Compounds 2g,h are the only active aziridinyl derivatives
that correlate with well-characterized molecular targets, the
protein tyrosine phosphatases.10 In contrast, compound 2c
exhibits high correlations with MAP7 gene-expressed proteins.22
The high specificity of 2c for leukemia and renal cancers, as
well as the possibility of exploiting a new molecular target, make
this analogue worthy of further investigation.
Antiangiogenesis Assays. We were intrigued by the high
correlation of compound 1h with both the antimitotic agent 6
and erythrocyte binding protein. To validate this result, we
assayed 1h, 1g, and 1k in the HUVEC inhibition and angio-
genesis assays. Compound 6 exerts its antitumor activity by
inhibiting angiogenesis, the development of new blood vessels
that permits tumor growth. Compounds 1h and 1g were assayed
to determine the possible role of the benzylamino group in
inhibiting angiogenesis. A model obtained by superimposing 6
onto the quinolinedione nucleus suggested that the 6-methoxy
derivative 1k would be a better combretastatin mimic than 1h
and 1g, Figure 3.
The HUVEC assay utilizes human umbilical vein endothelial
cells to evaluate the inhibition of cell growth as well as the
inhibition of cord and junction formation on a basement
membrane matrix. The control picture in Figure 4 shows
untreated cells forming cords and junctions in the matrix. Both
1h and 1g show a significant decrease in cord length and
junction formation at 10 µg/mL along with 23% growth
inhibition. Parallel inhibition assays[LW1] against HUVEC,
NCI-H460 (nonsmall-cell lung cancer), and MCF-7 (breast
cancer) revealed that both 1h and 1g inhibited the proliferation
of cancer cells more than the endothelial cells. These results
indicate that 1h and 1g possess some antiangiogenesis activity,
but that another mechanism may be responsible for the stronger
cancer cell inhibition. In contrast, compound 1k showed
complete inhibition of cord and junction formation at 10 µg/
mL[LW2]. Furthermore, both the endothelial cells and the
MCF-7 cancer cells are equally sensitive to 1k. In fact, the
Compound 1h correlated well with both the antimitotic agent
combretastatin A-4,28 compound 6 (NSC 64001) in Chart 4,
and the molecular target erythrocyte binding protein. The
correlation coefficient of 0.8 was the highest we observed in

2496 Journal of Medicinal Chemistry, 2008, Vol. 51, No. 8
HargreaVes et al.
Figure 3. Superimposed minimized structures of the combretastatin 6 (red) and the quinolinedione 1k (yellow).
Figure 4. Endothelial cells in the HUVEC angiogenesis assay either untreated or treated with 10 µg/mL of 1h, 1g, or 1k.
combretastatin A-4-treated endothelial cells are identical in
appearance to those treated with 1k. However, compound 6 is
about a 1000-fold more potent than 1k. We believe that the
altered position of the trimethoxyphenyl group of 1k is the main
reason for the decreased potency, see Figure 3. Although
compound 1k is not as potent as compound 6, it may still prove
to be a useful antitumor agent if it possesses low toxicity and
suitable solubility. Lastly, the mechanism of antiangiogenesis

Quinolinediones Exhibiting a Heat Shock Response
Journal of Medicinal Chemistry, 2008, Vol. 51, No. 8 2497
Analogues of 1 and 2 shown in Tables 3 and 4 were screened in
the National Cancer Institute’s 60-cell line screen.11,17–20 Mean
graph results were then analyzed with COMPARE.11,12
exhibited by 1k is unknown at this time and a future publication
will address this point.
2-(3,4,5-Trimethoxyphenyl)-5,8-dimethoxyquinoline (3c). A
solution of 2-amino-3,6-dimethoxybenzaldehyde (0.65 g, 3.59
mmol), 3,4,5-trimethoxyacetophenone (0.83 g, 3.95 mmol), and
10% ethanolic KOH (2 mL) was refluxed under nitrogen for 21 h.
The reaction mixture was cooled and the solvent removed in vacuo
to yield a yellow oil, which was purified by flash chromatography
(SiO2, ethyl acetate/hexanes (1:1)) to yield colorless crystals: 0.84 g
(66% yield); mp 110.5–111.5 °C; TLC (ethyl acetate/hexanes [1:1])
Rf ) 0.27; 1H NMR (300 MHz, CDCl3) δ 8.59 (d, J ) 9 Hz, 1H),
7.85 (d, J ) 9 Hz, 1H), 7.40 (s, 2H), 6.96 (d, J ) 8.7 Hz, 1H),
6.75 (d, J ) 8.7 Hz, 1H), 4.06 (s, 3H), 4.00 (s, 6H), 3.98 (s, 3H),
3.91 (s, 3H); 13C NMR (100 MHz, CDCl3) δ 156.27, 153.44,
149.45, 148.67, 140.30, 139.37, 135.46, 131.65, 120.32, 118.40,
107.70, 104.99, 103.48, 60.88, 56.24, 56.17, 55.69; MS 355 (M+),
340, 326, 310, 282, 252, 196, 177.
Conclusions
The Hsp90 screening and the COMPARE analysis of the
quinolinedione libraries 1 and 2 resulted in four leads.
Compound 2e modestly induced the heat shock response in
fibroblasts although it does not compete with ATP for binding
at either the C-terminal or N-terminal sites of Hsp90. The
biological significance of this result is as yet unknown.
Compound 1e mean graph data correlated well with the
fibroblast growth factor (FGF-2) molecular target and there are
some structural similarities with reported FGF-2 inhibitors. It
may be possible to design other such inhibitors by tethering
oxygenated side chains to heterocyclic quinones.
Compound 1h is considered to be an analogue of compound
6 based on COMPARE analysis and structural similarities. We
verified that analogues 1h, 1g, and 1k are inhibitors of
angiogenesis. Similarly, combretastatin A-4 exerts its antitumor
activity by inhibiting angiogenesis.
Compounds 2g,h are the only active aziridinyl derivatives
that correlated with well-characterized molecular targets, the
protein tyrosine phosphatases.10 Ongoing analogue development,
where the aziridinyl group of 2g,h is substituted for other
alkylating moieties, has yet to yield equally active compounds.
Compound 2c displayed some histological selectivity (leu-
kemia and renal) without correlating with a known cancer
chemotherapeutic molecular target. This finding suggests that
2c may be a novel antitumor agent directed toward an as-yet-
uncharacterized molecular target and, therefore, should be the
focus of further analogue development.
Finally, this study illustrates the value of COMPARE analysis
in mining a library of compounds for novel structures that
correlate with molecular targets useful for cancer chemotherapy.
The present study provides criteria for selecting compounds for
COMPARE analysis. The goal of the present effort was to
discover other molecular targets for the quinolinediones besides
the intended target, Hsp90. In future studies, the quinolinediones
selected from the COMPARE analysis (2e, 1e, 1h, 2g, 2h, and
2c) will be lead compounds for novel drug development. A
recent example of this means of drug discovery was the
discovery of a novel IMP dehydrogenase inhibitor structure
among a series of DNA cleaving agents.30 Molecular modeling
and analogue development has validated this COMPARE result.
2-(4-Biphenyl)-5,8-dimethoxyquinoline(3d). A solution of
2-amino-3,6-dimethoxybenzaldehyde (0.75 g, 4.14 mmol), 4-acetyl-
biphenyl (0.9 g, 4.58 mmol) and 10% ethanolic KOH (2 mL) in
ethanol (50 mL) was refluxed under nitrogen for 20 h. The solution
was cooled in an ice/water bath and the crystals filtered, washed
with ice cold ethanol, and dried: 1.20 g (85% yield); mp 162–162.5
1
°C; TLC (ethyl acetate/hexanes [1:1]) Rf ) 0.69; H NMR (500
MHz, CDCl3) δ 8.61 (d, J ) 8.5 Hz, 1H), 8.29 (d, J ) 8 Hz, 2H),
7.94 (d, J ) 8.5 Hz, 1H), 7.74 (d, J ) 8 Hz, 2H), 7.67 (d, J ) 7
Hz, 2H), 7.46 (t, J ) 7.5, 2H), 7.37 (t, J ) 7.5 Hz, 1H), 6.96 (d,
J ) 9 Hz, 1H), 6.74 (d, J ) 9 Hz, 1H), 4.07 (s, 3H), 3.97 (s, 3H);
13C NMR (125 MHz, CDCl3) δ 156.06, 149.67, 148.78, 141.89,
140.64, 140.59, 138.50, 131.68, 128.78, 128.04, 127.48, 127.10,
120.48, 118.34, 107.72, 103.52, 56.40, 55.76; MS 341 (M+), 326,
312, 297, 283, 268, 254, 178.
3,5,8-Trimethoxy-2-methylquinoline(3e). A solution of 2-amino-
3,6-dimethoxybenzaldehyde (1 g, 5.52 mmol), methoxyacetone (1
g, 1.04 mL, 11.4 mmol), and ethanolic KOH (10%, 2 mL) in ethanol
(50 mL) was refluxed under nitrogen for 30 min. The reaction
mixture was cooled and the solvent removed in vacuo. The resulting
oil was purified by flash column chromatography (SiO2, chloroform/
methanol [98:2]) to yield a white solid (1.14 g, 89%); mp 117–8
1
°C; TLC (ethyl acetate/hexanes [1:1]) Rf ) 0.28; H NMR (300
MHz, CDCl3) δ 7.67 (s, 1H), 6.77 (d, J ) 8.5 Hz, 1H), 6.71 (d, J
) 8.5 Hz, 1H), 4.02 (s, 3H), 3.97 (s, 3H), 3.96 (s, 3H), 2.69 (s,
3H); 13C NMR (100 MHz, CDCl3) δ 152.12, 151.83, 148.93,
147.63, 134.20, 121.45, 106.61, 103.82, 103.45, 55.90, 55.66, 55.30,
20.75; MS 233 (M+), 218, 204, 175, 160, 83.
3,5,8-Trimethoxy-2-phenylquinoline (3f). A solution of 2-amino-
3,6-dimethoxybenzaldehyde (1 g, 5.52 mmol), 2-methoxyacetophe-
none (0.83 g, 5.53 mmol), and ethanolic KOH (10%, 2 mL) in
ethanol (50 mL) was refluxed under nitrogen for 1 h. The reaction
mixture was cooled, and the solvent was removed in vacuo. The
resulting oil was purified using flash column chromatography (SiO2,
dichloromethane) to yield a white solid (1.17 g, 72%); mp 124–5
Experimental Section
All analytically pure compounds were dried under high vacuum
in a drying pistol over refluxing methanol. Many of the compounds
crystallized from reaction mixtures or elution solvents upon
concentration and therefore no recrystallization solvent is specified.
Elemental analyses of final products that were assayed were run at
Atlantic Microlab, Inc., Norcross, Ga. Melting points and decom-
position points were determined with a Mel-Temp apparatus. All
TLCs were performed on silica gel plates using a variety of solvent
systems and a fluorescent indicator for visualization. IR spectra
were taken as KBr pellets and only the strongest absorbances were
1
°C; TLC (ethyl acetate/hexanes [1:1]) Rf ) 0.61; H NMR (300
MHz, CDCl3) δ 8.02 (m, 2H), 7.84 (s, 1H), 7.43 (m, 3H), 6.77 (d,
J ) 8.4 Hz, 1H), 6.70 (d, J ) 8.4 Hz, 1H), 3.99 (s, 3H), 3.94 (s,
6H); 13C NMR (75 MHz, CDCl3) δ 151.71, 150.39, 149.74, 147.60,
137.86, 134.99, 129.94, 128.52, 127.77, 121.84, 108.32, 104.57,
104.11, 56.06, 55.62, 55.41; MS 295 (M+), 280, 266, 250, 236.
5,8-Dimethoxy-3-methyl-2-phenylquinoline (3g). A solution of
2-amino-3,6-dimethoxybenzaldehyde (1 g, 5.52 mmol), propiophe-
none (0.73 mL, 0.75 g, 5.60 mmol), and 10% ethanolic KOH (2
mL) was refluxed under nitrogen for 19 h. The solution was cooled
and the solvent removed in vacuo to yield an oil, which was purified
by flash chromatography (SiO2, ethyl acetate/hexanes [1:1]) to
afford a white solid; 1.06 g (69% yield); mp 145.5–146.5 °C; TLC
(ethyl acetate/hexanes [1:1]) Rf ) 0.59; 1H NMR (300 MHz, CDCl3)
δ 8.39 (s, 1H), 7.62 (m, 2H), 7.42 (m, 3H), 6.87 (d, J ) 8.2 Hz,
1H), 6.72 (d, J ) 8.2 Hz, 1H), 4.00 (s, 3H), 3.96 (s, 3H), 2.47 (s,
3H); 13C NMR (125 MHz, CDCl3) δ 159.50, 149.67, 148.14,
1
reported. H NMR spectra were obtained with a 300 or 500 MHz
spectrometer. All chemical shifts are reported relative to TMS.
The following heat shock protein-related assays were carried out
as previously described: Competition for binding to C-terminal
ATP-binding site on Hsp90 using immobilized novobiocin,14
competition for binding to N-terminal ATP-binding site on Hsp90
using immobilized geldanamycin,15 and heat shock response in
fibroblasts stably transfected with GFP reporter construct.16 All
stock solutions of 1 and 2 were prepared fresh in DMSO.

2498 Journal of Medicinal Chemistry, 2008, Vol. 51, No. 8
HargreaVes et al.
141.02, 138.98, 131.65, 129.27, 128.96, 128.02, 127.97, 120.85,
106.20, 103.53, 56.08, 55.75, 20.74; MS 279 (M+), 264, 250, 235,
220, 191.
128.77, 127.78, 126.71, 122.18, 106.08, 50.88, 26.59, 23.78; HRMS
(EI) m/z, 304.1207; requires, 304.1212.
6-Benzylamino-2-phenylquinoline-5,8-dione (1b) and 7-Ben-
zylamino-2-phenylquinoline-5,8-dione (1c). To a stirred solution
of 4a (200 mg, 0.85 mmol) in ethanol (15 mL) was added
benzylamine (186 mL, 182 mg, 1.7 mmol). After 4 h, the reaction
mixture was filtered and the red solid was washed with ice-cold
ethanol and dried. This solid was purified by flash chromatography
(SiO2, ethyl acetate/hexanes [2:3]) to give 1b and 1c.
2-(3,4,5-Trimethoxy)quinoline-5,8-dione (4c). To a solution of
3c (350 mg, 0.99 mmol) in acetonitrile (30 mL) was added dropwise
a solution of ceric ammonium nitrate (1.35 g, 2.46 mmol) in water
(30 mL). After one hour, the resulting precipitate was filtered,
washed with water, and dried to yield a dark red solid: 177 mg
(55% yield); mp 166–7 °C; TLC (ethyl acetate/hexanes [1:1]) Rf
1
) 0.17; H NMR (300 MHz, CDCl3) δ 8.45 (d, J ) 8.2 Hz, 1H),
Compound 1b was recrystallized from chloroform/hexanes as
orange crystals: 46 mg (16% yield); mp 208–209 °C; TLC
(chloroform/methanol [49:1]) Rf ) 0.76; 1H NMR (500 MHz,
CDCl3) δ 8.46 (d, J ) 8.5 Hz, 1H), 8.14 (m, 2H), 8.07 (d, J ) 8.5
Hz, 1H), 7.32–7.54 (m, 8H), 6.41 (t, J ) 5.5 Hz, 1H), 5.85 (s,
1H), 4.43 (d, J ) 5.5 Hz, 2H); 13C NMR (125 MHz, CDCl3) δ
182.12, 180.41, 160.83, 148.36, 146.82, 137.88, 135.83, 135.39,
130.56, 129.34, 129.32, 129.21, 128.46, 127.84, 127.81, 125.02,
101.50, 47.20. Anal. (C22H16N2O2 ·0.3H2O) C, H, N.
8.06 (d, J ) 8.2 Hz, 1H), 7.41 (s, 1H), 7.16 (d, J ) 10.5 Hz, 1H),
7.07 (d, J ) 10.5 Hz, 1H), 4.00 (s, 6H), 3.93 (s, 3H); MS 325
(M+), 310, 282, 252, 196, 140.
2-(4-Biphenyl)-quinoline-5,8-dione (4d). To a solution of 3d
(0.55 g, 1.61 mmol) in acetonitrile (25 mL) was added dropwise a
solution of ceric ammonium nitrate (2 g, 3.65 mmol) in water (25
mL). After 40 min., the resulting yellow precipitate was filtered,
washed with water and dried: 425 mg (85% yield); mp 221–222
°C; TLC (ethyl acetate/hexanes [1:1]) Rf ) 0.45; H NMR (500
MHz, CDCl3) δ 8.46 (d, J ) 8 Hz, 1H), 8.27 (d, J ) 8.5 Hz, 2H),
8.15 (d, J ) 8 Hz, 1H), 7.76 (d, J ) 8.5 Hz, 2H), 7.67 (d, J ) 7.5
Hz, 2H), 7.48 (t, J ) 7.5 Hz, 2H), 7.40 (t, J ) 7.5 Hz, 1H), 7.16
(d, J ) 10.5 Hz, 1H), 7.06 (d, J ) 10.5 Hz, 1H); 13C NMR (125
MHz, CDCl3) δ 184.48, 183.30, 161.60, 147.39, 143.55, 140.03,
139.07, 137.94, 136.06, 135.36, 128.90, 128.22, 127.93, 127.67,
127.47, 127.13, 123.89; MS 311 (M+), 283, 254, 228, 178, 84.
3-Methoxy-2-methylquinoline-5,8-dione (4e). To a stirred solu-
tion 3e (0.5 g, 2.15 mmol) in acetonitrile (25 mL) was added
dropwise a solution of ceric ammonium nitrate (2.93 g, 5.35 mmol)
in water (25 mL). After 1 h, the reaction mixture was extracted
with dichloromethane (3×75 mL), the organic layers were combined
and dried (Na2SO4), and the solvent was removed in vacuo to yield
a yellow solid. This was crystallized from chloroform/hexanes to
yield yellow crystals: 355 mg (81% yield); mp 156–7 °C dec; TLC
(ethyl acetate/hexanes [1:1]) Rf ) 0.19; 1H NMR (300 MHz, CDCl3)
δ 7.62 (s, 1H), 7.05 (d, J ) 10.5 Hz, 1H), 6.97 (d, J ) 10.5 Hz,
1H), 4.02 (s, 3H), 2.66 (s, 3H); MS 203 (M+), 203, 188, 175, 160,
132.
3-Methoxy-2-phenylquinolin-5,8-dione (4f). To a stirred solu-
tion of 3f (0.5 g, 1.69 mmol) in acetonitrile (30 mL) was added
dropwise a solution of ceric ammonium nitrate (2.32 g, 4.23 mmol)
in water (30 mL). After one hour, the resulting precipitate was
filtered, washed with water, and dried to afford an orange solid:
410 mg (91% yield); mp 197–8 °C dec; TLC (ethyl acetate/hexanes
[1:1]) Rf ) 0.35; 1H NMR (300 MHz, CDCl3) δ 8.01 (m, 2H),
7.81 (s, 1H), 7.46 (m, 3H), 7.09 (d, J ) 10.5 Hz, 1H), 7.01 (d, J
) 10.5 Hz, 1H), 4.04 (s, 3H); 13C NMR (75 MHz, CDCl3) δ 185.08,
182.39, 156.73, 153.47, 139.29, 137.37, 135.84, 129.94, 129.79,
129.29, 128.10, 113.86, 56.34; MS 265 (M+), 264, 250, 236, 207,
140.
Compound 1c was recrystallized from chloroform/hexanes as red
crystals: 80 mg (28% yield); mp 168–170 °C; TLC (chloroform/
1
methanol [49:1]) Rf ) 0.67; H NMR (500 MHz, CDCl3) δ 8.41
(d, J ) 8.5 Hz, 1H), 8.19 (m, 2H), 7.99 (d, J ) 8.5 Hz, 1H),
7.31–7.53 (m, 8H), 6.21 (t, J ) 5.5 Hz, 1H), 6.00 (s, 1H), 4.41 (d,
J ) 5.5 Hz, 2H); 13C NMR (125 MHz, CDCl3) δ 181.42, 181.36,
162.41, 149.22, 147.28, 137.61, 135.57, 135.05, 130.63, 129.06,
128.91, 128.24, 127.86, 127.67, 125.73, 122.49, 102.88, 46.87.
Anal. (C22H16N2O2) C, H, N.
6-Allylamino-2-phenylquinoline-5,8-dione (1d). To a stirred
solution of 4a (200 mg, 0.85 mmol) in ethanol (15 mL) was added
allylamine (191 µL, 145 mg, 2.55 mmol). After 2 h, the reaction
mixture was filtered and the red solid was washed with ice-cold
ethanol and dried. This solid was purified by flash chromatography
(SiO2, ethyl acetate/hexanes [2:3]) to give 1d as the major isomer
that was recrystallized from dichloromethane/hexanes as a red solid:
61 mg (25% yield); mp 187–188 °C dec; TLC (chloroform/methanol
[49:1]) Rf ) 0.37; 1H NMR (300 MHz, CDCl3) δ 8.41 (d, J ) 8.1
Hz, 1H), 8.19 (m, 2H), 8.00 (d, J ) 8.1 Hz, 1H), 7.52 (m, 3H),
6.02 (m, 1H), 5.97 (s, 1H), 5.89 (m, 1H), 5.29–5.38 (m, 2H), 3.88
(m, 2H); 13C NMR (125 MHz, CDCl3) 181.40, 181.34, 162.38,
149.21, 147.32, 137.60, 135.04, 131.33, 130.62, 128.90, 127.84,
125.71, 122.46, 118.48, 104.73, 102.70, 44.94. Anal.
(C18H14N2O2 ·0.2H2O) C, H, N.
6-(3-Methoxypropylamino)-2-phenylquinoline-5,8-dione (1e)
and 7-(3-Methoxypropylamino)-2-phenylquinoline-5,8-dione
(1f). To a stirred solution of 4a (200 mg, 0.85 mmol) in ethanol
(15 mL) was added 3-methoxypropylamine (174 µL, 1.71 mmol).
After 2 h, the solvent was removed in vacuo, and the residue was
purified by flash chromatography (SiO2, ethyl acetate/hexanes [1:1])
to give 1e and 1f as products.
Compound 1e was recrystallized from dichloromethane/hexanes
as a red solid: 102 mg (37% yield); mp 165–166 °C; TLC
(chloroform/methanol [49:1]) Rf ) 0.37; 1H NMR (300 MHz,
CDCl3) δ 8.39 (d, J ) 8.1 Hz, 1H), 8.20 (m, 2H), 7.98 (d, J ) 8.1
Hz, 1H), 7.50 (m, 3H), 6.49 (m, 1H), 5.95 (s, 1H), 3.55 (t, J ) 5.4
Hz, 2H), 3.40 (s, 3H), 3.34 (q, J ) 6 Hz, 2H), 1.98 (quint, J ) 6
Hz, 2H). Anal. (C19H18N2O3) C, H, N.
Compound 1f was recrystallized from dichloromethane/hexanes
as a red solid: 11 mg (4% yield); 1H NMR (300 MHz, CDCl3)
8.47 (d, J ) 8.1 Hz, 1H), 8.14 (m, 2H), 8.06 (d, J ) 8.1 Hz, 1H),
7.49 (m, 3H), 6.66 (m, 1H), 5.79 (s, 1H), 3.55 (t, J ) 6 Hz, 2H),
3.40 (s, 3H), 3.35 (q, J ) 6 Hz, 2H), 1.99 (pent, J ) 6 Hz, 2H).
The low yield precluded elemental analysis; spectroscopic analysis
was consistent with the structure 1f.
3-Methyl-2-phenylquinolin-5,8-dione (4g). To a stirred solution
of 3g (0.5 g, 1.79 mmol) in acetonitrile (30 mL) was added dropwise
a solution of ceric ammonium nitrate (2.45 g, 4.47 mmol) in water
(20 mL). After 30 min, water (50 mL) was added and the solution
stirred for a further 5 min. The mixture was then filtered, and the
pale yellow solid was washed well with water and dried: 373 mg
(84% yield); mp > 210 °C dec; TLC (ethyl acetate/hexanes [1:1])
1
Rf ) 0.32; H NMR (300 MHz, CDCl3) δ 8.28 (s, 1H), 7.58 (m,
2H), 7.48 (m, 3H), 7.13 (d, J ) 10.5 Hz, 1H), 7.04 (d, J ) 10.5
Hz, 1H), 2.53 (s, 3H); MS 248 (M+ - H) 220, 191, 166, 139, 115.
2-Phenyl-6-pyrrolidinylquinoline-5,8-dione (1a). To a stirred
solution of 4a (200 mg, 0.85 mmol) in ethanol (15 mL) was added
pyrrolidine (142 µL, 121 mg, 1.7 mmol). After 4 h, the reaction
mixture was filtered and the red solid was washed with ice-cold
ethanol and dried. This solid was crystallized from chloroform/
hexanes to afford 1a as red crystals: 187 mg (72% yield); mp
182–183 °C dec; TLC (chloroform/methanol [49:1]) Rf ) 0.40; 1H
NMR (400 MHz, CDCl3) δ 8.36 (d, J ) 8.4 Hz, 1H), 8.20 (m,
2H), 7.96 (d, J ) 8.4 Hz, 1H), 7.49 (m, 3H), 5.95 (s, 1H), 3.99
(bs, 2H), 3.38 (bs, 2H), 2.01 (m, 4H); 13C NMR (100 MHz, CDCl3)
δ 182.50, 180.64, 161.71, 148.79, 148.25, 137.73, 135.23, 130.34,
6-Benzylamino-2-(3,4,5-trimethoxyphenyl)-quinoline-5,8-di-
one (1g) and 7-Benzylamino-2-(3,4,5-trimethoxyphenyl)-quino-
line-5,8-dione (1 h). To a stirred solution of 4c (200 mg, 0.62
mmol) in ethanol (15 mL) was added benzylamine (134 µL, 1.23
mmol). After 2 h, the reaction mixture was filtered and the red
solid was washed with ice-cold ethanol and dried. This solid was
purified by flash chromatography (SiO2, ethyl acetate/hexane [1:1])
to afford 1g and 1h.

Quinolinediones Exhibiting a Heat Shock Response
Journal of Medicinal Chemistry, 2008, Vol. 51, No. 8 2499
Compound 1g was crystallized from chloroform/hexanes to afford
of pale orange crystals (70% yield): mp, dec >235 °C; TLC (ethyl
acetate) Rf ) 0.40; FTIR (KBr pellet): 808, 1124, 1242, 1342, 1464,
an orange crystalline product: 119 mg (45% yield); mp 221–222
1
1
°C dec; TLC (chloroform/methanol [49:1]) Rf ) 0.51; H NMR
1581, 1678, 2841, 2941, 2993, 3065 cm-1; H NMR (CDCl3) δ
(500 MHz, CDCl3) 8.38 (d, J ) 8 Hz, 1H), 7.93 (d, J ) 8 Hz,
1H), 7.41 (s, 2H), 7.40–7.30 (m, 5H), 6.22 (t, J ) 5 Hz, 1H), 5.99
(s, 1H), 4.41 (d, J ) 5 Hz, 2H), 3.99 (s, 6H), 3.92 (s, 3H); 13C
NMR (125 MHz, CDCl3) 181.31, 181.26, 162.01, 153.58, 149.09,
147.31, 140.66, 135.56, 134.98, 133.23, 129.06, 128.26, 127.72,
125.53, 122.29, 105.26, 102.79, 60.97, 56.38, 46.89. Anal.
(C25H22N2O5 ·0.3H2O) C, H, N.
Compound 1h was crystallized from chloroform/hexanes to
afford an orange crystalline product: 46 mg (17% yield); mp
203–204 °C; TLC (chloroform/methanol [49:1]) Rf ) 0.64; 1H
NMR (500 MHz, CDCl3) 8.45 (d, J ) 8.1 Hz, 1H), 8.01 (d, J )
8.1 Hz, 1H), 7.39–7.31 (m, 7H), 6.37 (t, J ) 6 Hz, 1H), 5.85 (s,
1H), 4.44 (d, J ) 6 Hz, 2H), 3.99 (s, 6H), 3.92 (s, 3H); 13C NMR
(125 MHz, CDCl3) δ 182.05, 180.30, 160.44, 153.89, 146.29,
146.65, 140.55, 135.79, 135.32, 133.43, 129.30, 129.16, 128.46,
127.79, 124.79, 123.00, 105.15, 101.47, 61.20, 56.56, 47.16. Anal.
(C25H22N2O5 ·0.3H2O) C, H, N.
2-Phenyl-6-(3,4,5-trimethoxybenzylamino)quinoline-5,8-one
(1i) and 2-Phenyl-7-(3,4,5-trimethoxybenzylamino)quinoline-
5,8-one (1j). To a stirred solution of 4a (380 mg, 1.62 mmol) in
ethanol (25 mL) was added 3,4,5-trimethoxybenzylamine (0.55 mL,
3.2 mmol). After 4 h, the reaction mixture was filtered and the
orange solid washed with ice-cold ethanol and dried. This solid
was purified by flash chromatography (SiO2, ethyl acetate/hexanes
[1:1]) to give 1i and 1j.
Compound 1i was crystallized from chloroform/hexanes as a dark
orange crystals: 264 mg (38% yield); mp 198–199 °C dec; TLC
(chloroform/methanol [49:1]) Rf ) 0.56; 1H NMR (500 MHz,
CDCl3) δ 8.42 (d, J ) 8.4 Hz, 1H), 8.19 (m, 2H), 8.01 (d, J ) 8.4
Hz, 1H), 7.51 (m, 3H), 6.54 (s, 2H), 6.16 (t, J ) 5.4 Hz, 1H), 6.02
(s, 1H), 4.34 (d, J ) 5.4 Hz, 2H), 3.87 (s, 6H), 3.86 (s, 3H); 13C
NMR (125 MHz, CDCl3) 181.45, 181.35, 162.47, 153.74, 149.20,
147.25, 137.90, 137.57, 135.07, 131.19, 130.69, 128.93, 127.86,
125.72, 122.53, 104.77, 102.91, 60.86, 56.21, 47.26. Anal.
(C25H22N2O5 ·0.6H2O) C, H, N.
8.49 (1H, d, J ) 8.4), 8.02 (1H, d, J ) 8.4), 7.41 (2H, s), 6.36
(1H, s), 3.99 (6H, s), 3.95 (3H, s), 3.92 (3H, s); 2D NMR HMBC
(see Supporting Information); MS m/z. Anal. (C19H17NO6 ·0.3H2O)
C, H, N.
6-Aziridinyl-2-methylquinoline-5,8-dione (2a). To a stirred
solution of 4b (200 mg, 1.16 mmol) in ethanol (15 mL) was added
aziridine (184 µL, 3.46 mmol). After 2 h, the resulting precipitate
was filtered, washed with ice-cold ethanol, and dried to yield a
yellow solid. This was purified by flash chromatography (SiO2,
chloroform/methanol [97:3]) to yield yellow crystals: 54 mg (22%
yield); mp 184–185 °C dec; TLC (chloroform/methanol [19:1]) Rf
1
) 0.63; H NMR (500 MHz, CDCl3) δ 8.30 (d, J ) 8 Hz, 1H),
7.49 (d, J ) 8 Hz, 1H), 6.42 (s, 1H), 2.77 (s, 3H), 2.32 (s, 4H);
13C NMR (125 MHz, CDCl3) 183.29, 181.26, 165.12, 157.23,
147.45, 134.63, 126.94, 126.38, 119.76, 27.74, 25.34; MS 214
(M+), 185, 159, 131, 120, 103, 94. Anal. (C12H10N2O2) C, H, N.
7-Aziridinyl-3-methoxy-2-methylquinolin-5,8-dione (2b). To
a solution of 4e (250 mg, 1.23 mmol) in ethanol (20 mL) was added
aziridine (300 µl, 5.65 mmol) and the mixture was stirred for 2 h.
The resulting precipitate was filtered, washed with ice-cold ethanol,
and dried to yield an orange solid. This solid was purified by column
chromatography (SiO2, chloroform/ethyl acetate [8:2]) and crystal-
lized from chloroform/hexanes to yield yellow crystals (42 mg,
14%); mp > 210 °C dec; TLC (chloroform/methanol [19:1]) Rf )
1
0.73; H NMR (300 MHz, CDCl3) δ 7.62 (s, 1H), 6.26 (s, 1H),
4.00 (s, 3H), 2.63 (s, 3H), 2.32 (s, 4H); 13C NMR (125 MHz,
CDCl3) 184.31, 179.14, 158.33, 157.37, 155.03, 139.14, 129.56,
117.85, 111.36, 56.12, 27.83, 20.04; MS 244 (M+), 229, 217, 202,
189, 146, 104. Anal. (C13H12N2O3) C, H, N.
7-Aziridinyl-3-methoxy-2-phenylquinolin-5,8-dione (2c). To
a stirred solution of 4f (215 mg, 0.81 mmol) in ethanol (25 mL)
was added aziridine (129 µl, 2.43 mmol). After 3 h, the resulting
precipitate was filtered, washed with ice-cold ethanol, and dried to
yield a yellow solid. This was purified by column chromatography
(SiO2, chloroform/ethyl acetate [8:2]) and crystallized from chlo-
roform/hexanes to yield yellow crystals: 115 mg (46% yield); mp
197 °C dec; TLC (chloroform/methanol [19:1]) Rf ) 0.81; 1H NMR
(500 MHz, CDCl3) δ 8.03 (m, 2H), 7.82 (s, 1H), 7.46 (m, 3H),
6.31 (s, 1H), 4.04 (s, 3H), 2.34 (s, 4H); 13C NMR (125 MHz,
CDCl3) 184.02, 178.81, 158.65, 156.92, 152.46, 139.75, 136.04,
129.92, 129.69, 129.63, 128.08, 118.06, 113.93, 56.33, 27.87; MS
306 (M+), 305, 291, 279, 263, 140. Anal. (C18H14N2O3 ·0.25H2O)
C, H, N.
6-Aziridinyl-2-(3,4,5-trimethoxyphenyl)quinoline-5,8-dione
(2d). To a stirred solution of 4c (140 mg, 0.43 mmol) in ethanol
(20 mL) was added aziridine (200 µL, 3.77 mmol). After 4 h, the
resulting precipitate was filtered, washed with ice-cold ethanol, and
dried to yield an orange solid. This solid was purified by column
chromatography (SiO2, chloroform/ethyl acetate [7:3]) and crystal-
lized from chloroform/hexanes to yield yellow crystals: 77 mg (49%
yield); mp 187–8 °C; TLC (chloroform/methanol [19:1]) Rf ) 0.73;
1H NMR (500 MHz, CDCl3) 8.44 (d, J ) 8 Hz, 1H), 7.99 (d, J )
8 Hz, 1H), 7.40 (s, 2H), 6.46 (s, 1H), 3.99 (s, 6H), 3.92 (s, 3H),
2.35 (s, 4H); 13C NMR (125 MHz, CDCl3) 182.91, 181.10, 161.64,
157.35, 153.65, 147.77, 140.69, 135.26, 133.05, 126.73, 123.09,
120.03, 105.21, 60.99, 56.39, 27.78; MS 366 (M+), 351, 323, 293,
265, 209. Anal. (C20H13N2O5 ·0.1H2O) C, H, N.
Compound 1j was crystallized from chloroform/hexanes as an
orange solid: 110 mg (16% yield); mp 215–216 °C dec; TLC
(chloroform/methanol [49:1]) Rf ) 0.66; 1H NMR (500 MHz,
CDCl3) δ 8.47 (d, J ) 8.4 Hz, 1H), 8.12–8.16 (m, 2H), 8.08 (d, J
) 8.4 Hz, 1H), 7.54–7.48 (m, 3H), 6.54 (s, 2H), 6.36 (t, J ) 5.7
Hz, 1H), 5.85 (s, 1H), 4.36 (d, J ) 5.7 Hz, 2H), 3.86 (2s, 9H); 13
C
NMR (125 MHz, CDCl3) 181.88, 180.20, 160.63, 153.76, 148.04,
146.54, 137.86, 137.57, 135.17, 131.19, 130.36, 129.08, 128.98,
127.54, 124.80, 104.54, 101.33, 60.86, 56.19, 47.25. Anal.
(C25H22N2O5 ·0.3H2O) C, H, N.
2-(3,4,5-Trimethoxyphyenyl)-6-methoxy-5,8-quinolinedione
(1k). Compound 1k was prepared by the two-step synthesis
described below.
To 100 mg (0.310 mmoles) of quinolinedione 4c in 10 mL of
ethanol at room temperature was added 0.34 mmols of piperidine,
and the reaction was stirred at room temperature. After the
disappearance of starting material by TLC (ethyl acetate), the
resulting red precipitate was filtered and washed with cold ethanol
and dried under high vacuum to yield 78 mg (62% yield) of
analytically pure 2-(3,4,5-trimethoxyphyenyl)-6-piperidino-5,8-
quinolinedione: mp, dec >170 °C; TLC (ethyl acetate) Rf ) 0.39;
FTIR (KBr pellet): 804, 1004, 1122, 1236, 1342, 1452, 1566, 1672,
1
2839, 2937, 3454, cm-1; H NMR (CDCl3) δ 8.23 (1H, d), 8.19
6-Aziridinyl-2-(4-biphenyl)quinoline-5,8-dione (2e). To a stirred
solution of 4d (275 mg, 0.88 mmol) in ethanol (30 mL) was added
aziridine (141 µL, 2.65 mmol). After 4 h, the resulting precipitate
was filtered, washed with ice-cold ethanol, and dried to yield an
orange solid. This solid was purified by column chromatography
(SiO2, chloroform/ethyl acetate [8:2]) and crystallized from chlo-
roform/hexanes to give orange crystals: 157 mg (51% yield); mp
(1H, d), 6.01 (1H, s), 3.79 (6H, s), 3.63 (3H, s), 3.40 (3H, m), 3.2
(3H, s), 1.57 (4H, s).
A mixture containing 50 mg (0.123 mmols) of 2-(3,4,5-
trimethoxyphyenyl)-6-piperidino-5,8-quinolinedione, 2 mL of metha-
nol, and 92 µL of sulfuric acid were refluxed for two hours. The
reaction was cooled and concentrated in vacuo. The resulting residue
was neutralized with saturated sodium bicarbonate solution and
extracted with ethyl acetate. The organic extracts were dried with
sodium sulfate and concentrated. The resulting pale orange solid
was recrystallized with ethyl acetate and hexanes to afford 30 mg
1
205–7 °C; TLC (ethyl acetate/hexanes [1:1]) Rf ) 0.20; H NMR
(400 MHz, CDCl3) δ 8.47 (d, J ) 8 Hz, 1H), 8.28 (d, J ) 8.4 Hz,
2H), 8.09 (d, J ) 8 Hz, 1H), 7.75 (d, J ) 8.4 Hz, 2H), 7.67 (d, J
) 7.6 Hz, 2H), 7.48 (t, J ) 7.6 Hz, 2H), 7.39 (t, J ) 7.6 Hz, 1H),

2500 Journal of Medicinal Chemistry, 2008, Vol. 51, No. 8
HargreaVes et al.
6.48 (s, 1H), 2.35 (s, 4H); 13C NMR (100 MHz, CDCl3) δ 183.05,
181.14, 161.56, 157.37, 147.97, 143.46, 140.14, 136.25, 135.34,
128.90, 128.26, 127.90, 127.64, 127.16, 126.90, 123.09, 120.12,
27.78; MS (M+). Anal. (C23H16N2O2 ·0.2H2O) C, H, N. Anal. Calcd
(C23H16N2O2.0.2H2O): C, 77.60; H, 4.64; N, 7.87. Found: C, 77.37;
H, 4.52; N, 7.81.
(6) Nimmanapalli, R.; OBryan, E.; Bhalla, K. Geldanamycin and its
analogue 17-allylamino-17-demethoxygeldanamycin lowers Bcr-Abl
levels and induces apoptosis and differentiation of Bcr-Abl-positive
human leukemic blasts. Cancer Res. 2001, 61, 1799–1804.
(7) Chiosis, G.; Lucas, B.; Shtil, A.; Huezo, H.; Rosen, N. Development
of a purine-scaffold novel class of Hsp90 binders that inhibit the
proliferation of cancer cells and induce the degradation of Her2
tyrosine kinase. Bioorg. Med. Chem. 2002, 10, 3555–3564.
(8) Chiosis, G.; Timaul, M. N.; Lucas, B.; Munster, P. N.; Zheng, F. F.;
Sepp-Lorenzino, L.; Rosen, N. A small molecule designed to bind to
the adenine nucleotide pocket of Hsp90 causes Her2 degradation and
the growth arrest and differentiation of breast cancer cells. Chem. Biol.
2001, 8, 289–299.
(9) Lucas, B; Rosen, N; Chiosis, G. Facile synthesis of a library of 9-alkyl-
8-benzyl-9H-purin-6-ylamine derivatives. J. Comb. Chem. 2001, 3,
518–20.
(10) Hargreaves, R.; David, C. L.; Whitesell, L.; Skibo, E. B. Design of
quinolinedione-based geldanamycin analogues. Bioorg. Med. Chem.
Lett. 2003, 13, 3075–3078.
(11) Paull, D. K.; Shoemaker, R. H.; Hodes, L.; Monks, A.; Scudiero, D. A.;
Rubinstein, L.; Plowman, J.; Boyd, M. R. Display and analysis of
differential activity of drugs against human tumor cell lines: Develop-
ment of mean graph and COMPARE algorithm. J. Natl. Cancer Inst.
1989, 81, 1088–1092.
(12) Sausville, E. A.; Zaharevitz, D.; Gussio, R.; Meijer, L.; Louarn-Leost,
M.; Kunick, C.; Schultz, R.; Lahusen, T.; Headlee, D.; Stinson, S.;
Arbuck, S. G.; Senderowicz, A. Cyclin-dependent kinases: Initial
approaches to exploit a novel therapeutic target. Pharmacol. Ther.
1999, 82, 285–292.
(13) Showalter, H. D. H.; Pohlmann, G. An improved synthesis of 4,7-
dimethoxy-1H-indole. Org. Prep. Proced. Int. Briefs 1992, 24, 484–
488.
(14) Marcu, M. G.; Schulte, T. W.; Neckers, L. J. Novobiocin and related
coumarins and depletion of heat shock protein 90-dependent signaling
proteins. J. Natl. Cancer Inst. 2000, 92, 242–248.
(15) Whitesell, L.; Mimnaugh, E. G.; DeCosta, B.; Myers, C. E.; Neckers,
L. M. Inhibition of heat shock protein HSP90-pp60v-src heteroprotein
complex formation by benzoquinone ansamycins: Essential role for
stress proteins in oncogenic transformation. Proc. Natl. Acad. Sci.
U.S.A. 1994, 91, 8324–8328.
(16) Bagatell, R.; Paine-Murrieta, G. D.; Taylor, C. W.; Pulcini, E. J.;
Akinaga, S.; Benjamin, I. J.; Whitesell, L. Induction of a heat shock
factor 1-dependent stress response alters the cytotoxic activity of
Hsp90-binding agents. Clin. Cancer Res. 2000, 6, 3312–3318.
(17) Boyd, M. R. Principles and Practices of Oncology (PPO updates);
J.B. Lippincott: Phildelphia, PA, 1989; pp 1–12.
(18) Lanza, D. L.; Yost, G. S. Selective dehydrogenation/oxygenation of
3-methylindole by cytochrome P450 enzymes. Drug Metab. Dispos.
2001, 29, 950–953.
(19) Bassett, S.; Urrabaz, R.; Sun, D. Cellular response and molecular
mechanism of antitumor activity by leinamycin in MiaPaCa human
pancreatic cancer cells. Anti-Cancer Drugs 2004, 15, 689–696.
(20) Thompson, J. E.; Thompson, C. B. Putting the rap on akt. J. Clin.
Oncol. 2004, 22, 4217–4226.
6-Aziridinyl-3-methyl-2-phenylquinolin-5,8-dione (2f). To a
stirred solution of 4g (300 mg, 1.20 mmol) in ethanol (25 mL)
was added aziridine (190 µL, 3.6 mmol). After 1 h, the reaction
mixture was filtered, and the precipitate was washed with ice-cold
ethanol and dried to yield a yellow solid. The solid was purified
by flash chromatography (SiO2, chloroform/ethyl acetate [8:2]) and
crystallized from chloroform/hexanes to give yellow crystals: 170
mg (49% yield); mp 193 °C dec; TLC (chloroform/methanol [19:
1
1]) Rf ) 0.78; H NMR (400 MHz, CDCl3) δ 8.28 (s, 1H), 7.58
(m, 2H), 7.46 (m, 3H), 6.44 (s, 1H), 2.52 (s, 3H), 2.33 (s, 3H); 13
C
NMR (100 MHz, CDCl3) δ 183.02, 181.56, 164.04, 157.21, 145.58,
138.98, 136.35, 135.74, 129.23, 129.02, 128.31, 126.88, 120.22,
27.76, 20.61; MS 290 (M+), 262, 248, 234, 220, 205, 191, 166,
139, 115. Anal. (C18H14N2O2) C, H, N.
Antiangiogenesis Assays. Growth inhibition and cord formation
assays were conducted using human umbilical vein endothelial cells
(HUVEC) purchased from BD Biosciences. HUVEC cells were
grown in EGM-2 medium (Cambrex).
Growth Inhibition Assay. The standard sulforhodamine B assay
was used to assess growth inhibition. Briefly, cells were inoculated
in 96-well plates and incubated for 24 h. Serial dilutions of the
compounds were then added. After 48 h, the plates were fixed with
trichloroacetic acid, stained with sulforhodamine B, and read with
an automated plate reader. A growth inhibition of 50% (GI50 or
the drug concentration causing a 50% reduction in the net protein
increase) was calculated from optical density data with Immunosoft
software.
Cord Formation Assay. Matrigel, a basement membrane matrix,
was purchased from BD Biosciences. An aliquot of 60 µL was
placed in each well of an ice-cold 96-well plate. The plates were
left 15 min at room temperature and then incubated for 30 min at
37 °C to permit the material to gel. Meanwhile, HUVEC were
harvested and diluted to a concentration of 2 × 105 cells/mL. Cell
suspension (100 µL) was added to each well. The compounds to
be tested were then added in an additional 100 µL of culture
medium. After 24 h of incubation, images were acquired using an
inverted Nikon Diaphot 300 and DXM1250F Nikon digital camera.
Drug effects were assessed and compared with untreated controls
by measuring the length of cords formed and the number of
junctions.
(21) Xing, C.; Skibo, E. B.; Dorr, R. T. Aziridinyl quinone antitumor agents
based on indoles and cyclopent[b]indoles: Structure-activity relation-
ships for cytotoxicity and antitumor activity. J. Med. Chem. 2001,
44, 3545–3562.
(22) Ross, D. T.; Scherf, U.; Eisen, M. B.; Perou, C. M.; Rees, C.; Spellman,
P.; Iyer, V.; Jeffrey, S. S.; Van De Rijn, M.; Waltham, M.;
Pergamenschikov, A.; Lee, J. C. E.; Lashkari, D.; Shalon, D.; Myers,
T. G.; Weinstein, J. N.; Botstein, D.; Brown, P. O. Systematic variation
in gene expression patterns in human cancer cell lines. Nat. Genet.
2000, 24, 227–235.
(23) Itokawa, H.; Matsumoto, K.; Morita, H.; Takeya, K. Cytotoxic
naphthoquinones from Mansoa alliacea. Phytochemistry 1992, 31,
1061–2.
(24) Turowski, P.; Franckhauser, C.; Morris, M. C.; Vaglio, P.; Fernandez,
A.; Lamb, N. J. Functional cdc25C dual-specificity phosphatase is
required for S-phase entry in human cells. Mol. Biol. Cell 2003, 14,
2984–2998.
(25) Murphy, P. V.; Pitt, N.; O’Brien, A.; Enright, P. M.; Dunne, A.;
Wilson, S. J.; Duane, R. M.; O’Boyle, K. M. Identification of novel
inhibitors of fibroblast growth factor (FGF-2) binding to heparin and
endothelial cell survival from a structurally diverse carbohybrid library.
Bioorg. Med. Chem. Lett. 2002, 12, 3287–3290.
(26) Aviezer, D.; Seddon, A. P.; Wildey, M. J.; Bohlen, P.; Yayon, A.
Development of a high-throughput screening assay for inhibitors of
fibroblast growth factor-receptor-heparin interactions. J. Biomol.
Screen. 2001, 6, 171–177.
Acknowledgment. We thank the Arizona Disease Control
Research Commission for their generous support.
Supporting Information Available: Elemental analyses, 2D
NMR HMBC of 1k, and detailed antiangiogenesis assay results
are included. This material is available free of charge via the Internet
at http://pubs.acs.org.
References
(1) Vinogradov, S.; Roig, V.; Sergueeva, Z.; Nguyen, C. H.; Arimondo,
P.; Thuong, N. T.; Bisagni, E.; Sun, J. S.; Helene, C.; Asseline, U.
Synthesis and binding properties of oligo-2′-deoxyribonucleotides
conjugated with triple-helix-specific intercalators: Benzo[e] and ben-
zo(G] pyridoindoles. Bioconjugate Chem. 2003, 14, 120–135.
(2) Neckers, L. Hsp90 inhibitors as novel cancer chemotherapeutic agents.
Trends Mol. Med. 2002, 8, S55–S61.
(3) Neckers, L.; Schulte, T. W.; Mimnaugh, E. Geldanamycin as a potential
anti-cancer agent: Its molecular target and biochemical activity. InVest.
New Drugs 1999, 17, 361–373.
(4) Andrus, M. B.; Meredith, E. L.; Simmons, B. L.; Sekhar, B. B. V. S.;
Hicken, E. J. Total synthesis of (+)-geldanamycin and (-)-o-
quinogeldanamycin with use of asymmetric anti- and syn-glycolate
aldol reactions. Org. Lett. 2002, 4, 3549–3552.
(5) Andrus, M. B.; Meredith, E. L.; Sekhar, B. B. V. S. Synthesis of the
left-hand portion of geldanamycin using an antiglycolate aldol reaction.
Org. Lett. 2001, 3, 259–262.
(27) Manetti, F.; Corelli, F.; Botta, M. Fibroblast growth factors and their
inhibitors. Curr. Pharm. Des. 2000, 6, 1897–1924.

Quinolinediones Exhibiting a Heat Shock Response
Journal of Medicinal Chemistry, 2008, Vol. 51, No. 8 2501
(28) Nam, N. H. Combretastatin A-4 analogues as antimitotic antitumor
agents. Curr. Med. Chem. 2003, 10, 1697–1722.
(29) Loneragan, G. H.; Gould, D. H.; Mason, G. L.; Garry, F. B.; Yost,
G. S.; Lanza, D. L.; Miles, D. G.; Hoffman, B. W.; Mills, L. J.
Association of 3-methyleneindolenine, a toxic metabolite of 3-meth-
ylindole, with acute interstitial pneumonia in feedlot cattle. Am. J.
Vet. Res. 2001, 62, 1525–1530.
(30) Ghodousi, A.; Huang, X.; Cheng, Z.; Skibo, E. B. Pyrrolobenzimi-
dazoles (PBIs) linked to heterocycles and peptides. Design of DNA
base pair specific phosphate hydrolyzing agents and novel cytotoxic
agents. J. Med. Chem. 2004, 47, 90100.
JM7014099