Natural and semisynthetic azaphilones as a new scaffold for Hsp90 inhibitors

Article abstract of DOI:10.1016/j.bmc.2010.06.068

A series of mold metabolites of Ascomycetes, structurally belonging to the class of azaphilones, were found to inhibit the heat shock protein Hsp90. In particular, bulgarialactone B was tested for its binding to Hsp90 using surface plasmon resonance and l

Full text of DOI:10.1016/j.bmc.2010.06.068

Bioorganic & Medicinal Chemistry 18 (2010) 6031–6043
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry
journal homepage: www.elsevier.com/locate/bmc
Natural and semisynthetic azaphilones as a new scaffold for Hsp90 inhibitors
a
b
b
c
Loana Musso a, Sabrina Dallavalle a, , Lucio Merlini , Adriana Bava , Gianluca Nasini , Sergio Penco ,
Giuseppe Giannini c, Chiara Giommarelli d, Andrea De Cesare d, Valentina Zuco d, Loredana Vesci c,
Claudio Pisano c, Fabrizio Dal Piaz e, Nunziatina De Tommasi e, Franco Zunino d
*
a Dipartimento di Scienze Molecolari Agroalimentari, Università di Milano, Via Celoria 2, 20133 Milano, Italy
b Dipartimento di Chimica, Materiali ed Ingegneria Chimica ‘Giulio Natta’ del Politecnico, CNR-Istituto di Chimica del Riconoscimento Molecolare, Sezione ‘Adolfo Quilico’;via
Mancinelli 7, 20131 Milano, Italy
c Sigma-tau, R&D, Via Pontina km 30,400, 00040 Pomezia, Italy
d Unità di Chemioterapia e Farmacologia Antitumorale Preclinica, Dipartimento di Oncologia Sperimentale e Laboratori, Fondazione IRCCS Istituto Nazionale dei Tumori, Via Venezian
1, 20133 Milano, Italy
e Dipartimento di Scienze Farmaceutiche, Università di Salerno, Italy
a r t i c l e i n f o
a b s t r a c t
Article history:
A series of mold metabolites of Ascomycetes, structurally belonging to the class of azaphilones, were
found to inhibit the heat shock protein Hsp90. In particular, bulgarialactone B was tested for its binding
to Hsp90 using surface plasmon resonance and limited proteolysis assays and for its effects on Hsp90 cli-
ent proteins expression in a series of human tumor cell lines. This compound showed high affinity for
Hsp90, interacting with the 90–280 region of the N-terminal domain and down-regulated the Hsp90 cli-
ent proteins Raf-1, survivin, Cdk4, Akt, and EGFR. Bulgarialactone B and other natural azaphilones
showed antiproliferative activity in a panel of human tumor cell lines; their conversion into semisyn-
thetic derivatives by reaction with primary amines increased the antiproliferative activity. Preliminary
results indicated in vivo activity of bulgarialactone B against an ascitic ovarian carcinoma xenograft, thus
supporting the therapeutic potential of this novel series of Hsp90 inhibitors.
Received 8 February 2010
Revised 16 June 2010
Accepted 20 June 2010
Available online 26 June 2010
Keywords:
Azaphilones
Cancer
Antitumour
Hsp90
Ó 2010 Elsevier Ltd. All rights reserved.
Chaperones
Cytotoxicity
Plasmon surface resonance
Limited proteolysis
p53
Natural products
1. Introduction
cells to tolerate the imbalanced signaling that such oncoproteins
create. Thus, targeting Hsp90 may have the potential advantage
Interest in the heat shock protein 90 (Hsp90) molecular chaper-
one as a therapeutic target is related to its central role in correct
folding and stabilization of proteins involved in malignant behav-
ior and tumor progression.1 Multiple signal transduction pathways
implicated in the regulation of cell proliferation and survival are
dependent on Hsp90.2 Several Hsp90 client proteins are involved
in critical processes including cell-cycle regulation and apoptosis.
The heat shock proteins are often overexpressed in tumor cells,
and this supports their ability to survive under unfavorable stress
conditions (e.g., hypoxia and acidosis). The essential chaperoning
function of Hsp90 is subverted during oncogenesis to make malig-
nant transformation possible and to facilitate rapid somatic evolu-
tion. Functioning as a biochemical buffer for the numerous genetic
lesions that are present within tumors, Hsp90 allows mutant
proteins to retain or even gain function while permitting cancer
of simultaneously blocking multiple oncogenic pathways.
Hsp90 exists as a homodimer made up of three domains.1,3 The
N-terminal domain contains an ATP-binding site that binds the
natural products geldanamycin and radicicol, and the analogue
17-AAG (Chart 1). The middle domain is highly charged and has
high affinity for cochaperones and client proteins. A second ATP-
binding site is located in the C-terminus of Hsp90. This C-terminal
nucleotide binding pocket has been shown to bind not only ATP,
but cisplatin, novobiocin, epigallocatechin-3-gallate (EGCG) and
taxol.4
A number of highly specific Hsp90 inhibitors have been identi-
fied. They redirect Hsp90 chaperoning activity and decrease cellu-
lar levels of its numerous cancer-related client proteins.1 Such
inhibitors exhibit promising antitumor activity as single agents
or in combination with other cytotoxic agents,5–7 and some of
them are in clinical development.8,9
During a target-oriented screening of natural compounds aimed
at stabilizing and enhancing the function of p53, a transcription
* Corresponding author. Tel.: +39 2 50316818; fax: +39 2 50316801.
E-mail address: sabrina.dallavalle@unimi.it (S. Dallavalle).
0968-0896/$ - see front matter Ó 2010 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmc.2010.06.068

6032
L. Musso et al. / Bioorg. Med. Chem. 18 (2010) 6031–6043
of Bulgaria inquinans by Stadler et al.10 In order to increase the pro-
O
O
O
H
N
duction of compounds for biological tests and chemical derivatiza-
tion, we cultured B. inquinans ICRM-184 on corn-steep agar. Under
these conditions, we obtained bulgarialactones A and B in high
yields and, in addition, we isolated a new azaphilone, named bul-
garialactone D (4), whose structure was established on the basis
of spectroscopic data (Chart 2).
OMe
OMe
O
O
N
N
H
H
O
HO
OMe
OMe
OMe
Catalytic hydrogenation of 2 saturated the triene system of the
side-chain, giving compound 5 (Scheme 1).
O
O
NH2
NH2
17-AAG
Among the rich class of azaphilones, the metabolites of the asco-
mycete Monascus purpureus are well known and easily available, due
to the use of the mold to ferment edible rice in East Asia.20 From a
culture of this fungus on RA medium, we isolated the known anka-
flavin (6),21 monascin (7),22 monascorubrin (8),23 and rubropuncta-
tin (9)24 (Chart 3), all possessing the peculiar core of azaphilones.
These compounds possess different side-chains in position 3 and 6
compared to bulgarialactone B.
Finally, we investigated a different series of azaphilones, deflec-
tins25 (10–13), that we reisolated from cultures of a strain of Asper-
gillus deflectus CBS 109.55 on RA medium (Chart 4). These
compounds, differently from the Monascus metabolites, have an
angular lactone fused to a 3,7-dimethyl-7,8-dihydroisochromen-6-
one.
O
O
Geldanamycin (GDA)
O
O
O
Cl
O
HO
OH
Radicicol
Chart 1.
To gain insight into the structural requirements affecting the
biological activity of these natural compounds, we exploited the
known reactivity of azaphilones toward amines, consisting in an
exchange of the pyranyl oxygen for nitrogen to give vinylogous
factor which regulates the cell cycle and thus functions as a tumor
suppressor, we found that a fungal metabolite, named bulgarialac-
tone B (2),10 did not change the wild-type p53 protein level but in-
duced partial down-regulation of the mutant p53, which is known
to be stabilized by Hsp90. This result prompted us to investigate
the activity of bulgarialactone B as a potential Hsp90 inhibitor.
Structurally, bulgarialactone B features a highly oxygenated tri-
cyclic core linked to a 13-C unsaturated chain containing a b-dike-
tone. It is a member of a large group of fungal pigments known as
azaphilones, mainly isolated from perfect and imperfect stages of
Ascomycetes, such as Aspergillus, Penicillium, Hypoxylon, and Mon-
ascus spp.11
Azaphilones have been reported to exhibit a wide range of bio-
logical activities, including monoamine oxidase inhibition,12 tumor
promotion inhibition,13–15 cytotoxicity,16,17 gp120-CD4 binding
inhibition,18 and sphingosine kinase inhibition.19
The promising results obtained by testing compound 2 prompted
us to investigate the activity of bulgarialactone B and other azaphil-
ones as potential novel Hsp90 inhibitors.
c
-pyridones.11
Thus, we prepared derivatives of the three series with differ-
ently substituted primary amines (Schemes 2 and 3).
The reaction of bulgarialactone B and of its hexahydroderivative
5 with amines proceeded via Michael addition of the nucleophilic
amines to the electrophilic C-8 carbon. This resulted in the forma-
tion of carbinolamines which underwent C–O bond cleavage to
generate b-enaminones 14a–e (Scheme 2). This behavior had pre-
viously been observed in epicocconone26 and wortmannin.27
The reactions proceeded smoothly with primary amines,
whereas treatment with secondary amines did not produce the de-
sired enaminones. A possible explanation for this different reactiv-
ity could be the higher stability of enaminones 14 derived from
primary amines due to hydrogen bonding between the enamine
and the carbonyl at position 7. In fact, the chemical shift of the
amine hydrogens, which is always in the range of 10–11 ppm, is
indicative of strong hydrogen bonding26 (Scheme 2).
2. Chemistry
When 5 was reacted with an excess of benzylamine for a pro-
longed time, compound 15 was obtained due to a further nucleo-
philic attack of the amine on the carbonyl in position 30 (Scheme 3).
Bulgarialactone B (2), together with bulgarialactone A (1) and
bulgarialactone C (3), was first isolated in low yields from a strain
O
O
OH
OH
1'
4
5
8
3
6
OH
2
R
O
O
O
O
O
H3C
O
H3C
9
O
O
R = H C5-C6 single bond Bulgarialactone A (1)
R = OH C5-C6 single bond Bulgarialactone B (2)
R = H C5-C6 double bond Bulgarialactone C (3)
Bulgarialactone D (4)
Chart 2.

L. Musso et al. / Bioorg. Med. Chem. 18 (2010) 6031–6043
6033
nique.28 SPR allows to verify the affinity of compounds toward the
protein, and to assess how they associate with and dissociate from
the protein in real time, giving a more detailed view of their interac-
tion with Hsp90. SPR experiments were performed using radicicol
(Chart 1), a well-known Hsp90 inhibitor, as a positive control. Our
data demonstrated the high affinity of 2 for Hsp90, comparable with
that measured for radicicol (Table 2), both in terms of thermody-
namic (KD) and kinetic (kd) dissociation constants.
O
O
3
OH
6
OH
OH
OH
a
O
O
O
O
O
H3C
O
H3C
O
O
The limited proteolysis approach is based on the evidence that
exposed, weakly structured, and flexible regions of a protein can be
recognized by a proteolytic enzyme. Trypsin, chymotrypsin, and
endoprotease V8 were selected as conformational probes in order
to provide structural information on different regions of Hsp90. The
fragments released from the protein were identified by MALDI-TOF/
MS analyses, leading to the assignment of cleavage sites (Table 2).
The differences in the proteolytic patterns were analyzed to identify
the protein regions involved in the molecular interactions.29
In Figure 2A, the preferentially hydrolyzed positions observed on
isolated Hsp90 are shown: widespread cleavage sites within the en-
tire protein structure have been identified. Comparing our data with
the predicted secondary structure organization of Hsp90,30 it can be
deduced that most of the identified cleavage sites are located in
unstructured portions, thus confirming the efficiency of the ap-
proach. To further validate the method, the radicicol/Hsp90 complex
was investigated, since the radicicol binding region has been deeply
characterized by crystallography.31 The results achieved in the rad-
icicol/Hsp90 complex are shown in Figure 2B. Comparison between
the proteolysis patterns observed on Hsp90 with or without radici-
col indicates significant protection of the N-terminal domain from
enzymatic hydrolysis. This result is in agreement with the radici-
col/Hsp90 interaction involving the ATPase located in the N-termi-
nal domain of Hsp90.31
2
5
Scheme 1. Reagents and conditions: (a) H2/Pd/C 10%, AcOEt, rt, 78%.
O
R
R
O
6
3
O
O
O
O
O
O
H3C
H3C
O
O
R = C7H15 Monascorubrin (8)
R = C5H11 Rubropunctatin (9)
R = C7H15 Ankaflavin (6)
R = C5H11 Monascin (7)
Chart 3.
Me
O
R = C6H13
R = C8H17
R = C8H17
R' = H
R' = H
R' = CH3 (12)
(10)
(11)
Me
O
O
R = C10H21 R' = CH3 (13)
O
O
H
R'
R
Limited proteolysis analysis of the compound 2/Hsp90 complex
produced the fragmentation pattern summarized in Figure 2C.
These data demonstrate that most of the N-terminal domain sites
as well as those located in the protein portion interconnecting this
domain to the middle domain were protected following complex
formation, thus suggesting an interaction involving the 90–280 re-
gion of Hsp90.
Chart 4.
Treatment of compound 14a with sodium periodate afforded
the tricyclic derivative 16 possessing the 7-nitrogen substituted
core, without isolation of the intermediate aldehyde (Scheme 4).
By contrast, azaphilones containing a pyranyl ring, such as 6
and 14, reacted with amines to give directly the corresponding cyc-
lic derivatives 17–19. The reaction apparently proceeds via an
open-ring intermediate containing an enaminone and a carbonyl
group, which immediately cyclizes to the vinylogous pyridone11
(Schemes 5 and 6).
In contrast to the effects of ansamycin inhibitors of Hsp90 (e.g.,
17AAG), compound 2 did not induce up-regulation of Hsp70. The
lack of modulation of Hsp70 could reflect a different binding mode
of 2 to Hsp90, which did not involve the shift of client association
from Hsp90 to Hsp70 prior to proteasome-mediated degradation.
In A431 cells treated with 2 for 4 h, a coimmunoprecipitation
experiment performed with anti-Raf-1 antibody revealed a re-
duced amount of Hsp90 bound to the client protein, as could be ex-
pected from an Hsp90 inhibitor (Fig. 3A). The effect was
comparable to that of 17AAG at equitoxic concentrations. How-
ever, this short treatment did not induce a shift of the binding of
Raf-1 from Hsp90 to Hsp70, because the level of coprecipitated
Hsp70 with Raf-1 was undetectable (Fig. 3B). The lack of up-regu-
lation of Hsp70 in treated cells has been observed with other puta-
tive inhibitors of Hsp90 (e.g., epigallocatechin32 and curcumin33).
This finding would suggest a mechanism of interaction different
from that of geldanamycins, that target the ATP-binding site of
Hsp90. This interpretation was also supported by proteolysis
experiments (Fig. 2). If interaction with other cochaperones con-
tributes to the modulation of Hsp90 function by the novel agents
remains to be explored.
3. Results and discussion
As a first step of this study, we investigated the cellular and bio-
chemical profile of compound 2.
The antiproliferative assays were performed on a panel of hu-
man tumor cell lines following 72 h exposure. The IC50 values of
2, reported in Table 1, indicated variable antiproliferative effects
on the tested cell lines in a range of micromolar concentrations.
Successively, compound 2 was tested for its effects on Hsp90 cli-
ent protein expression in the same cell lines used in the antiprolifer-
ative assay (Fig. 1). Following 24-h exposure to a concentration
corresponding to IC80 (around 2 ꢀ IC50; i.e., ca. 10
lM), the exam-
ined client proteins (Raf-1, survivin, Cdk4, Akt, and EGFR) were
down-regulated. The effect was marginal in melanoma JR8 cells. In
contrast, protein depletion was almost complete in the ovarian car-
cinoma cell line, IGROV-1. The p53 protein was partially down-reg-
ulated only in A431 cells carrying a mutant p53 protein. This effect
was consistent with the stabilization of the mutant form by Hsp90.
In an effort to confirm the Hsp90/compound 2 binding and to
identify the binding site, we used surface plasmon resonance (SPR)
analyses and the limited proteolysis–mass spectrometry tech-
In the above experiments, aimed at a biochemical analysis of
protein expression modulation (Fig. 1), the exposure time was only
24 h, because prolonged exposure resulted in the partial detach-
ment of cells. Indeed, at 48 h, marked cleavage of poly(ADP-
ribose)polymerase and activation of caspase-3 were found in cells
treated with 2, supporting apoptosis induction (Fig. 4).

6034
L. Musso et al. / Bioorg. Med. Chem. 18 (2010) 6031–6043
R 1'
O
R
4
1''
R
5
OH
OH
OH
O
O
6
a
O
O
OH
1'''
N
O
O
O
H3C
7
H3C
H3C
+H2N
O
O
O-
R'
R'
H
14
2,5
14a R' = -CH2Ph
14b R' = -CH2CH2CH2OH
O
HO
NHCOCH3
R =
COOH
O
14c R' =
14d R' =
N
O
HO
14e R' = -CH2Ph
R =
Scheme 2. Reagents and conditions: (a) R0NH2, MeOH or CH2Cl2, 2–24 h, rt, N2.
Ph
NH
6
O
O
6
OH
OH
a
OH
O
O
OH
O
H3C
O
O
H3C
O
HN
O
CH2Ph
5
15
Scheme 3. Reagents and conditions: (a) PhCH2NH2 (4 mequiv), MeOH, rt, 72 h, 30%.
R 3
R
4
5
8
6
OH
O
O
N
OH
a
Me
Ph
Me
O
H3C
O
O
O
7
H3C
O
O
HN
a
Me
O
N
Me
O
O
14a
16
O
O
O
O
H
H
Me
Me
O
HO
C10H21
C10H21
R =
19
14
Scheme 6. Reagents and conditions: (a) PhCH2NH2, THF, rt, 10 min, 62%.
Scheme 4. Reagents and conditions: (a) NaIO4, CH3OH/CH2Cl2 1:1, H2O, 0 °C,
45 min, then rt 72 h, 55%.
O
C7H15
O
C7H15
C7H15
O
O
a
O
O
O
H
N
O
O
O
H3C
O
H3C
N
R
R
H3C
O
17-18
17
18
O
O
6
R = -CH2-Ph
OH
H
O
R =
OH
OH
HO
Scheme 5. Reagents and conditions: (a) for 17: RNH2, THF, rt, 1 h, 77%; for 18: RNH2, H2O/MeOH 1:1, phosphate buffer pH 7, rt, 51%.

L. Musso et al. / Bioorg. Med. Chem. 18 (2010) 6031–6043
6035
Table 1
Antiproliferative activity in different cell lines, ATPase activity inhibition and binding to Hsp90 (IC50
,
l
M
SD) of azaphilones 1–13
Compd
Antiproliferative activity (
l
M)
Hsp90-ATPase activity inhibition
Binding Hsp90 (FP)
Cell lines
NCI-H460
A431
JR8
IGROV-1
1
2
4
5
6
7
8
9
10
11
12
13
18.7 0.4
50
>50
21.3 3.1
30 0.2
>50
>50
>50
>50
—
6.87 0.82
5.53 0.71
17.60 1.9
1.20 0.11
7.05 0.89
17.77 2.5
130.89
28.2 3.4
>28
26 3.7
>25
11.45 1.71
6.42 0.87
—
5.70 0.95
5.64 0.79
30.64 3.9
130.89
>30
4.12 0.53
4.65 0.64
—
0.80 0.11
5.08 0.64
>30
102.88
28.2 3.5
>28
26 3.5
—
>50
9.87 0.71
>50
14.5 0.93
26.3 3.5
>50
22.9 3.3
9.2 1.2
>50
>100
61.7 0.85
>100
35.4 1.3
>100
>100
0.27 0.01
0.040 0.001
>100
>100
>100
>28
26 3.2
—
>50
>50
>50
>50
>50
>25
—
—
>100
JR8
A431
STO
IGROV-1
C
2
C
2
C
2
C
2
Raf-1
Survivin
A431
C
2
Cdk4
Akt
EGFR
Tubulin
Actin
Figure 1. Analysis of Hsp90 client protein levels (survivin, Cdk4, Akt, Raf-1) in tumor cells treated with (2). Total cellular extracts were obtained 24 h after treatment (11 lM,
IC80). Actin is shown as a control for protein loading. EGFR levels were analyzed only in A431 cells and tubulin is shown as a control for protein loading.
Hsp90 inhibitors. We therefore tested the natural compounds
1–13 for their cytotoxic activity against a variety of human tumor
cell lines; they were also tested for their ability to inhibit Hsp90 ATP-
ase activity and for binding affinity in the FP assay (Table 1).
The most potent compounds of this series were 1, 2, and 5, char-
acterized by IC50 values 610 lM against several cell lines. The
potency of these compounds was substantially lower than that of
geldanamycin or its analogue 17-AAG (IC50 in A431 cells,
Table 2
Study of the interaction with Hsp90
kd (sꢁ1
)
Protected sitesb
a
a
Compd
KD (nM)
Radicicol
2
3.04 0.82
1.30 0.53
0.011 0.003
0.017 0.82
75, 94, 215
94, 215, 245, 277
a
SPR data.
Limited proteolysis data.
b
0.2 0.04 and 0.07 0.01 lM, respectively). In the case of
compound 2, the reduced potency was consistent with a low affinity
The promising results obtained by testing compound 2 prompted
for Hsp90 in the competitive binding assay (IC50 61.7 0.85 vs
us to investigate the activity of other azaphilones as potential
A
Hsp90
HSP82
245
245
277
277
94
462
557
75
215
490 539 560
462 557
643
643
643
B
Hsp90/radicicol
HSP82+Radicicol
490
462
560
557
C
Hsp90/compound 2
HSP82+2
413
75
368
560
Figure 2. Schematic representation of the results obtained from limited proteolysis experiments on recombinant Hsp90. The preferential cleavage sites detected performing
enzymatic digestions on recombinant Hsp90 (A), on the Hsp90/radicicol complex (B), or on the Hsp90/compound 2 complex (C), are in dark gray. The Hsp90 N-terminal
domain is highlighted in light gray, the middle domain is boxed and the C-terminal domain is highlighted in gray.

6036
L. Musso et al. / Bioorg. Med. Chem. 18 (2010) 6031–6043
A
B
C
17AAG
2
C
17AAG
2
IP: Raf-1
HSP70
WB: HSP90
EGFR
Akt
WB: Raf-1
WB: HSP70
Raf-1
Actin
Figure 3. (A) Analysis of Hsp90 client protein levels and Hsp70 in A431 cells. Total cellular extracts were obtained 24 h after treatment (11
control for protein loading. EGFR levels were analyzed only in A431 cells and tubulin is shown as a control for protein loading. (B) Coimmunoprecipitation of Raf-1/Hsp90 4 h
after 17-AAG or (2) treatment, in A431 cells. Cells were treated with equitoxic (IC80) concentrations of 17AAG (0.1 M) or 2 (11 M). Cell lysates were than harvested and
immunoprecipitated with anti-Raf-1 rabbit polyclonal antibody. Immunoprecipitates were immunoblotted for Hsp90 Blots were stripped and probed for Raf-1.
lM, IC80). Actin is shown as a
l
l
anti-Raf-1 antibody, performed after 4-h exposure, revealed
reduced binding of Raf-1 to Hsp90 (not shown). However, the
depletion was dose dependent and the variable effects on the client
protein levels were likely determined by the relative sensitivity of
the various cell lines (Fig. 6).
24 h
48 h
C
2
C
2
cleaved Caspase-3
PARP-1
Actin
Compound 17 shows a strong binding to the enzyme (FP test),
confirmed by the SPR analysis (KD 47.8 0.5 nM). However, the
limited proteolysis experiment for 17 (Fig. 7) indicates a pattern
similar to that of control, that is, no protection of the sensitive re-
gions. It is therefore possible that 17 interacts with the protein in
rather rigid regions (beta leaflets or, more likely, helical regions),
about which the limited proteolysis approach cannot give enough
information. Again, the deflectin derivative 19 showed no appre-
ciable Hsp90 binding, like the parent natural compounds, thus con-
firming the detrimental effect of an angular structure.
IGROV-1
Figure 4. Biochemical analysis of apoptosis-related factors in IGROV-1 cells.
Cleaved Caspase-3 and PARP-1 expression levels were analyzed by Western blot
and total cellular extracts, used for this analysis, were obtained both 24 h or 48 h
after treatment with 2 (11 lM).
The lack of close correlation between affinity for Hsp90, as
determined in the FP assay, and cellular effects, in terms of cell
growth inhibition and of client protein depletion, could reflect an
unfavorable cellular pharmacokinetics or interaction with other
proteins which may limit the Hsp90 binding inside the cell. Indeed,
the compounds related to bulgarialactone (e.g., 9 and 17) are char-
acterized by a reactive conjugate system that could react with
amines and thiols, present at high concentrations inside the cell
(e.g., GSH).
Preliminary in vivo antitumor activity studies supported the
therapeutic potential of this new series of compounds. In particu-
lar, intraperitoneal administration of 2 to athymic nude mice bear-
ing ovarian carcinoma xenografts growing intraperitoneally as
ascitic tumors resulted in a substantial increase in survival (Fig. 8).
In conclusion, the available evidence based on the biochemical,
pharmacodynamic and cellular effects indicated that the com-
pounds described in this paper may represent a novel series of
Hsp90 inhibitors.
It is evident that a major limitation of this class of natural com-
pounds is the low potency in target inhibition, which requires
micromolar concentrations to inhibit cell proliferation. Also con-
sidering the chemical features of these agents, we cannot rule
out the possibility of interactions with other cellular components.
If this is the case, these interactions may have a detrimental effect
by limiting an effective inhibition of the primary target. However,
our approaches aimed at modifying the basic structure of these
agents provide evidence that specific variations may improve the
Hsp90 inhibitory effect. In particular, compounds 16 and 17, con-
taining a lipophilic chain linked to the azaphilone tricyclic ring sys-
tem, brought about substantially increased binding to the enzyme,
accompanied by relevant cytotoxic effects. Finally, we have docu-
mented the pharmacological interest of a representative com-
pound of the novel series. Current efforts are directed toward
optimization of the in vivo properties of selected analogues.
1.095 0.05 for 17AAG). Interestingly, the two pyranyl compounds
8 and 9, with a saturated side-chain, exhibited strong binding to
Hsp90. This was paralleled by the twofold increase of binding going
from 2 to 5. However, this strong binding was not reflected in the
antiproliferative potency. A plausible explanation for this discrep-
ancy could be the unfavorable cellular pharmacokinetics.
The series of natural deflectins (10–13), characterized by an
angular lactone fused to a dihydroisochromenone ring, were found
to be almost inactive.
In an attempt to identify the structural requirements involved
in the Hsp90 inhibitory activity, the semisynthetic compounds
14–19 were prepared exploiting the well-known reactivity of aza-
philones with amines. The IC50 values of compounds 14–19,
reported in Table 3, indicated that these agents were generally
more potent than compounds of the previous series (1–13).
Among the open-ring analogues (14a–e), compound 14a, con-
taining a benzyl substituent on the nitrogen, was the most potent
in the antiproliferative assay. These compounds showed moderate
binding to Hsp90, comparable to that of 2. On the contrary, com-
pounds 16 and 17, containing the same tricyclic system and the
same benzyl substituent on N-7, exhibited increased Hsp90 bind-
ing, the highest value (IC50 4 nM) being shown by the monascoru-
brin derivative 17.
Figure 6 shows the analysis of client protein expression modu-
lation. The effects on the examined proteins were variable among
cell lines. Compounds 14a and 14d caused marked depletion of
Raf-1 and Akt in IGROV-1 and JR8 cells (Fig. 5A). In contrast to
the marginal effect of 2, compounds 14a, 14d, 15, and 17 were very
effective in causing depletion of client proteins in JR8 cells (Fig. 5A
and B). Compound 14d caused almost complete depletion of EGFR
in A431 (Fig. 5A). The effects of 14a and 16 on EGFR expression in
A431 cells were less marked at equitoxic concentrations (IC80
)
(Fig. 5A and C). For compound 14a, coimmunoprecipitation with

L. Musso et al. / Bioorg. Med. Chem. 18 (2010) 6031–6043
6037
Table 3
Antiproliferative activity on different cell lines, ATPase activity inhibition and binding to Hsp90 (IC50
,
l
M
SD) of compounds 14–19
Compd
Antiproliferative activity (
l
M)
Hsp90 ATPase activity inhibition
Binding Hsp90 (FP)
Cell lines
NCI-H460
A431
JR8
IGROV-1
STO
14a
14b
14c
14d
14e
15
16
17
18
19
4.3 0.1
11.0 4.2
17.3 0.3
10.2 1.7
30 7.5
21.3 3.1
>20
3.4 0.06
—
6.7 0.2
0.30
3.30
4.48
1.54
4.24
1.20
3.98
2.67
>18
2
4.80
6.00
15.46
8.59
9.02
5.70
5.16
4.0
0.20
1.90
2.32
1.72
2.86
0.80
2.92
3.46
>18
—
3.57
9.47
5.77
2.06
4.0
4.36
2.92
2.28
>10
4.5
5.75 0.75
13.11 0.87
—
6.10 0.68
19.9 1.24
14.5 0.93
8.9 0.5
—
27.7 0.65
2.29 0.06
6.83 0.04
27.4 0.38
37.5 0.59
35.4 1.3
0.28 0.06
0.004 0.0001
>50
—
—
>50
>50
>100
A431
14a 14d
STO
IGROV-1
14a 14d
JR8
C
C
14a 14d
C
C
14a 14d
A
B
EGFR
Raf-1
Akt
Cdk4
Actin
JR8
A431
16
STO
16
IGROV-1
C 16
JR8
C
16
C
C
C
17 19
C
EGFR
Raf-1
Raf-1
Akt
Survi
vin
survivin
p53
Akt
Cdk4
Actin
Cdk4
Actin
Figure 5. Comparison of biochemical effects on Hsp90 client proteins. (A) Hsp90 client expression levels after 24 h treatment with (14a) or (14d). Total cellular extracts were
obtained 24 h after treatment with equitoxic (IC80) doses of (14a, 2.7 M for A431 cells; 10.7 M for JR8 cells, 2.3 M for IGROV-1 and 11.6 M for STO cells) or (14d, 8.6 M for
A431; IGROV-1 and STO cells; 16 M for JR8 cells). Actin is shown as a control for protein loading. (B) Hsp90 client expression levels after 24 h treatment with (17) or (5) in JR8
cells. Total cellular extracts were obtained 24 hafter treatment withequitoxic (IC80) doses of (17, 6.36 M) or (5, 9.3 M). Actin is shown as a control for protein loading. (C) Hsp90
client expression levels after 24 h treatment with (16). Total cellular extracts were obtained 24 h after treatment (16, 9.8 M, IC80). Actin is shown as a control for protein loading.
l
l
l
l
l
l
l
l
l
285.34 and A. deflectus CBS 109.55, were purchased from CBS col-
lection (Centraal Bureau voor Schimmelcultures, Baarn). The mas-
ter fungal strains were preserved as mycelium (Bulgaria and
Monascus) or spore (A. deflectus) plugs, suspended in physiological
solution (0.9% NaCl), at 4 °C. Stock cultures for experimental trials
were grown in slants on agar medium preserved at 4 °C until the
use and sub-cultured once every three weeks (M. purpureus) or
two months (B. inquinans and A. deflectus).
4. Experimental
4.1. General experimental procedures
Mass spectra were obtained with a Finnigan-MATT-TSQ 70ev, a
Bruker Esquire 3000 and for HR-MS with a Bruker APEX-QZT ICR
spectrometers. The 1H NMR spectra were carried out on a Bruker
DMX 500 or ARX 400 instruments at the temperature of 305 K.
HPLC analyses were performed using a LiChroCARTcolumn RP-18
250-4 (Merck) on an Agilent 1100 instrument; mobile phase:
H2O/MeCN; flow rate = 0.5 mL minꢁ1. Thin and preparative layer
chromatography (TLC and PLC) were performed on precoated
Merck Silica Gel 60 F254 plates. FC (flash chromatography) was per-
formed with Merck silica gel (0.04–0.06 mm).
4.3. Solid-state fermentation and extraction procedures
Large-scale cultures were made in Roux flasks (750 mL), con-
taining 100 mL of productive RA (composition in g/L: rice 90, agar
20) medium or CSA (composition in g/L: corn-steep liquor 10, su-
crose 100, glucose 90, yeast extract 5, K2HPO4 2, agar 20) medium.
The flasks were inoculated with 5 mL of a mycelium or spore sus-
pension of a preculture of fungus in slants containing 40 mL OA
(Oatmeal agar Merck) medium, PDA (Potato Dextrose agar Merck)
or CSA medium. One slant was used to inoculate two flasks. After
4.2. Microorganisms
The Ascomycetes used in the present study, B. inquinans (ICRM-
184, from the Collection of CNR-ICRM Institute), M. purpureus CBS

6038
L. Musso et al. / Bioorg. Med. Chem. 18 (2010) 6031–6043
EGFR
Akt
Cdk4
survivin
EC50 11 µM
EC50 16.8 µM
EC50 8.49 µM
EC50 11 µM
EGFR
Akt
Cdk4
EC50 3.57 µM
EC50 8.93 µM
EC50 4.93 µM
EC50 2.32 µM
survivin
100
80
60
40
20
0
100
80
60
40
20
0
0.1
1
10
10
(2) µM
(14a) µM
Figure 6. Dose–response of the biochemical effect of (2) and (14a) on Hsp90 client proteins in A431 cells. Total cellular extracts were obtained 24 h after treatment with (2)
(4.4, 11, 22 M) or (14a) (0.3, 2.32, 8.93 M). Protein expression levels of EGFR, Akt, Cdk4 and survivin were analyzed by Western blot. The quantification of Western blot
l
l
analyses was obtained with the Image Quant 5.2 software and results (relative expression normalized to the tubulin, as loading control) are presented as % of protein
expression, as compared to the control. IC50 values of the down-regulation of each protein level are reported in the plots.
Figure 7. Effect of 2 on survival of mice bearing ip ovarian carcinoma tumor IGROV-1. (s) Control mice; (d) mice treated ip daily with 2 (25 mg/kg) for 4 days/week for
2 weeks.
4.4. Isolation and purification of bulgarialactones A (1), B (2)
and D (4) from cultures of B. inquinans ICRM-184
100
80
After preliminary screening on several media either in surface
60
or submerged culture, the best results were obtained when the
fungus was grown on CSA medium. B. inquinans large-scale cul-
tures were made in 40 Roux flasks inoculated with 5 mL of a myce-
40
20
0
lium suspension prepared from 15 days old preculture in the same
0
10
20
30
40
50
60
medium CSA, maintained at 24 °C in the dark and extracted three
times after 30–35 days of growth. The dark oil crude extract
(32 g) was partially purified by solvents: after washing with hex-
ane, the solid residue was dissolved in CH2Cl2, until discoloration
of the solvent. The CH2Cl2 fraction (22 g), containing the mixture
of Bulgarialactones, was subjected first to FC in a column precondi-
tioned with hexane/AcOEt 2:1, eluting with a mixture of hexane/
AcOEt/MeOH. The fractions eluted with hexane/AcOEt 1:1 afforded
bulgarialactone A (1, 450 mg). Elution with hexane/AcOEt from 1:2
to AcOEt 100% gave pure bulgarialactone B (2, 5.3 g, yield 16.5%),
while elution in AcOEt/MeOH 100:1 gave bulgarialactone D (3) as
a mixture with bulgarialactone B (4.8 g). Final purification of
metabolites was reached by means of further PLC or FC.
Days after tumor cell inoculation
Figure 8.
incubation in thermostat at 24 °C, the cultures were extracted two
times with a mixture of AcOEt/MeOH (100:1). The organic fraction
was dried with Na2SO4 and evaporated under vacuum. Separation
and purification of the metabolites were performed by column
chromatography on silica gel. Further purification was performed
on preparative chromatography plates (PLC), when needed. Finally,
the purified compounds were checked for their identity by NMR,
MS and HPLC and consistency with other spectroscopic data
reported for reference compounds.

L. Musso et al. / Bioorg. Med. Chem. 18 (2010) 6031–6043
6039
Bulgarialactone B (2) was separated from bulgarialactone D (3) first
by a FC stepwise elution with CH2Cl2/MeOH from 50:1 to 10:1 and
then on PLC preparative plate eluting with toluene/acetone 8:2 to
give 1.4 g of bulgarialactone B (yield 4.4%) and 240 mg of bulgari-
alactone D (yield 0.75%). The final yield of the main metabolite,
bulgarialactone B, was 6.7 g (ca. 21% of crude total extract). The
spectroscopic data (1H NMR and MS-ESI) for bulgarialactones A
and B matched with those reported in the literature.10
medium and stored as spores, at 4 °C. The fungus was inoculated in
10 Roux flasks on RA medium with a 10 days old preculture grown
on PDA medium and incubated at 24 °C. After 30 days, the cultures
were extracted to give a brown oil. The crude residue (5.5 g) was
purified by FC on silica gel with a stepwise elution with hexane/
AcOEt from 7:3 to 4:6. A further purification by reverse phase
(RP-18) column chromatography, with a stepwise elution with
H2O/acetone from 4:6 to 1:9, afforded 350 mg (6.4 %), 600 mg
(ca. 11%), and 75 mg (1.4 %) of deflectins 10–12, respectively, and
550 mg (10 %) of deflectin 13.
Bulgarialactone D (4): Brown solid; 1H NMR (300 MHz, CDCl3): d
0.86 (t, J = 7.25 Hz, 3H, H-120), 1.01 (d, J = 6.87, 3H, C100CH3), 1.28–
1.42 (m, 2H, H-110), 1.70 (s, 3H, C9aCH3), 2.09–2.27 (m, 2H, H-
100 + H-5), 2.44–2.59 (m, 1H, H-5), 3.45–3.74 (m, 4H, CH2OH + H-
6 + H-4), 4.15–4.27 (m, 1H, H-4), 4.59–4.76 (m, 2H, H-8), 5.87 (dd,
J = 7.6, 15.3 Hz, 1H, H-90), 6.07 (d, J = 14.9 Hz, 1H, H-40), 6.14 (dd,
J = 10.7, 15.26 Hz, 1H, –CH@CH–), 6.27 (dd, J = 11.4, 14 Hz, 1H,
–CH@CH–), 6.60 (dd, J = 10.7, 14.9 Hz, 1H, –CH@CH–), 6.72 (s, 1H,
H-20), 7.35 (dd, J = 11.4, 14.9 Hz, 1H, –CH@CH–), 15.84 (br s, 1H,
OH). 13C NMR (75 MHz, CDCl3): d 189.62, 185.48, 177.05, 169.21,
167.69, 150.63, 147.54, 143.27, 143.00, 129.55, 128.74, 128.62,
126.03, 120.61, 101.19, 86.06, 73.66, 66.91, 63.84, 39.02, 32.62,
31.73, 29.63, 23.36, 19.81, 11.89. ESI-MS: m/z 455 [M+H+] and 477
[M+Na]; negative ion: m/z 453 [MꢁH+]. Anal. (C26H30O7) C, H.
4.8. 6-Hydroxymethyl-3-(1-hydroxy-10-methyl-3-oxo-dodec-1-
enyl)-9a-methyl-5,6-dihydro-9aH-furo[3,2-g]isochromene-2,9-
dione (5)
Compound 2 (350 mg, 0.77 mmol) dissolved in AcOEt (15 mL)
was hydrogenated at room temperature with 10% Pd/C (25 mg).
The reaction was monitored frequently on TLC in hexane/acetone
2:1 and stopped when the orange pigment was completely con-
verted into the yellow pigment, before the formation of more polar
products. After filtration on Celite and evaporation of the solvent,
the crude residue was purified by FC eluting with hexane/i-PrOH
3:1 to afford the title compound (350 mg, 78%). HPLC: retention
time (tR) 14.37 min, solvent CH3CN–H2O, 85:15; 1H NMR
(300 MHz, CDCl3): d 0.75–0.88 (m, 6H, H-120, C100CH3), 0.99–1.16
(m, 2H, H-110), 1.16–1.39 (m, 8H, 4CH2), 1.51–1.68 (m, 2H, CH2),
1.72 (s, 3H, C9aCH3), 1.87–2.04 (m, 1H, H-100), 2.43 (t, J = 8.0 Hz,
1H, CH2CO), 2.73–2.96 (m, 2H, H-5), 3.83 (dd, J = 5.0 Hz, 12.2 Hz,
1H, CH2OH), 3.96 (dd, J = 3.4, 12.2 Hz, 1H, CH2OH),4.34–4.44 (m,
1H, H-6), 6.70 (s, 1H, H-20), 7.05 (s, 1H, H-4), 7.84 (s, 1H, H-8),
15.60 (br s, 1H, OH). ESI-MS: m/z 459 [M+H+] and 481 [M+Na+].
Anal. (C26H34 O7) C, H.
4.5. Isolation and purification of azaphilones 6–9 from cultures
of M. purpureus CBS 285.34
A preliminary screening on small-scale (15 flasks) was made in
order to isolate and characterize the azaphilones produced from
the strain of M. purpureus. The orange pigments monascorubrin (8)
and rubropunctatin (9), were isolated only from cultures grown on
RA medium and at short-time incubation (max 10 days, 28 °C); the
yellow pigments ankaflavin (6) and monascin (7), and the red pig-
ments monascorubramin, and rubropunctamin were isolated from
cultures of the fungus when grown on RA medium or CSA medium,
but at longer incubationtime (30–35 days, 28 °C). The crude extracts
were separated and purified first by means of FC on silica gel with a
stepwise elution with hexane/AcOEt from 8:2 to 0:1, to afford yellow
and orange pigments or CH2Cl2/acetone from 50:1 to 1:1 for the red
pigments. The pure pigments were obtained by chromatography on
reverse phase (RP-18): yellow pigments ankaflavin and monascin
and orange pigments monascorubrin and rubropunctatin were sep-
arated on preparative plates (PLC) using H2O/CH3CN 3:7 as eluent,
while the separation of red pigments was performed eluting with
H2O/CH3CN 1:1. The spectroscopic data (1H NMR and MS) for com-
pounds 6–9 were identical with those reported in lit.14,34
4.9. 6-(Benzylaminomethylene)-5-(2,3-dihydroxypropyl)-3-(1-
hydroxy-10-methyl-3-oxo-dodeca-1,4,6,8-tetraenyl)-7a-methyl-
6H,7aH-benzofuran-2,7-dione (14a)
To a solution of 2 (300 mg, 0.66 mmol) in methanol (8 mL) ben-
zylamine (72 lL, 0.66 mmol) was added under nitrogen atmo-
sphere. The solution was stirred 24 h at room temperature, then
the solvent was removed under reduced pressure. The crude was
purified by gel permeation chromatography on Sephadex LH-20
using CH2Cl2 as eluent to obtain the title compound as a red solid
(120 mg, 52%). 1H NMR (300 MHz, CDCl3): d 0.88 (t, J = 7.3 Hz, 3H,
H-120), 1.00 (d, J = 6.7, 3H, C100CH3), 1.26–1.41 (m, 2H, –H-110),
1.70 (s, 3H, C7aCH3), 2.06–2.21 (m, 1H, H-100), 2.56–2.66 (m, 2H,
H-100), 3.53 (dd, J = 6.3, 11.16 Hz, 1H, H-300), 3.71 (dd, J = 3.7,
11.2 Hz, 1H, H-300), 3.85–3.95 (m, 1H, H-200), 4.55 (d, J = 6.3 Hz,
2H, –CH2Ph), 5.84 (dd, J = 7.8, 15.26 Hz, 1H, H-90), 6.05 (d,
J = 14.8 Hz, 1H, H-40), 6.13 (dd, J = 10.8, 15.26 Hz, –CH@CH–), 6.25
(dd, J = 11.3, 14.9 Hz, 1H, –CH=CH–), 6.54 (dd, J = 10.8, 14.89 Hz,
1H, –CH@CH–), 6.74 (s, 1H, H-20), 6.93 (s, 1H, H-4), 7.21–7.41 (m,
6H, Ph + –CH@CH–), 7.70 (d, J = 13.4 Hz, 1H, H-1000), 10.51–10.65
(m, 1H, NH), 16.00 (br s, 1H, OH). ESI-MS: m/z 560 [M+H+]; Anal.
(C33H37NO7) C, H, N.
4.6. Large-scale production of monascorubrin 8 and
rubropunctatin 9 from M. purpureus
For the preparation of semisynthetic amino-derivatives of mon-
ascorubrin 8, the strain of M. purpureus was grown in batches of 40
flasks on RA medium (100 mL). Each flask was inoculated with a
suspension of the fungus prepared from 7 days old cultures grown
in slants on OA medium (28 °C). The crude extract fraction, soluble
in CH2Cl2 (4.3 g), was purified by FC on silica gel eluting with hex-
ane/AcOEt at increasing polarity to give a mixture of yellow and or-
ange pigments. The final purification of these pigments was
performed by chromatography on reverse phase (RP-18), eluting
with H2O/acetone from 1:1 to 2:8 to give 470 mg of compound 8
(11%) and 300 mg of compound 9 (7%).
4.10. 5-(2,3-Dihydroxypropyl)-3-(1-hydroxy-10-methyl-3-oxo-
dodeca-1,4,6,8-tetraenyl)-6-[(3-hydroxypropylamino)-methyl
ene]-7a-methyl-6H,7aH-benzofuran-2,7-dione (14b)
Under
nitrogen
atmosphere
3-aminopropanol
(2 lL,
4.7. Isolation and purification of deflectins 10–13 from cultures
0.02 mmol) was added to a solution of 2 (10 mg, 0.02 mmol) in
CH2Cl2 (1 mL). The solution was stirred for 2 h at room tempera-
ture, then the solvent was removed under reduced pressure. The
crude was purified by PLC using CH2Cl2/CH3OH 95:5 as eluent to
of A. deflectus
The strain of A. deflectus was maintained in slants on HA (com-
position in g/L: yeast extract 4, malt extract 10, glucose 4, agar 15)

6040
L. Musso et al. / Bioorg. Med. Chem. 18 (2010) 6031–6043
obtain the title compound as a red oil (4 mg, 38%). 1H NMR
(300 MHz, CDCl3):
0.85 (t, J = 7.2 Hz, 3H, H-120), 1.00 (d,
(25 mg, 0.12 mmol). The mixture was stirred 45 min at 0 °C and
then 3 days at rt. The solvent was removed under reduced pressure
and the crude was extracted with CH2Cl2. The combined organic
fractions were washed with saturated NaCl solution, dried over
Na2SO4, and concentrated in vacuo. The crude was purified by flash
column chromatography on deactivated silica gel (110 g of SiO2 in
400 mL of a 4% aqueous solution of KH2PO4 evaporated under re-
duced pressure and dried at 110 °C overnight) using CH2Cl2/ace-
tone 19.5:0.5 as eluent, to give the title compound (25 mg,
55%).1H NMR (300 MHz, CDCl3): d 0.85 (t, J = 7.5 Hz, 3H, H-120),
1.00 (d, J = 6.5 Hz, 3H, C-100CH3), 1.28–1.41 (m, 2H, H-110), 1.69
(s, 3H, H-7a), 2.05–2.20 (m, 1H, H-100), 4.89 (s, 2H, CH2Ph), 5.77
(dd, J = 8.1, 15.3 Hz, 1H, CH@CH), 6.05 (d, J = 14.9 Hz, 1H, CH@CH),
6.12 (dd, J = 10.4, 15.3 Hz, 1H, CH@CH), 6.25 (dd, J = 11.4, 14.9 Hz,
1H, CH@CH), 6.49 (dd, J = 10.4, 14.9 Hz, 1H, CH@CH), 6.57 (d,
J = 7.5 Hz, 1H, CH@CHN), 6.70 (s, 1H, H-20), 6.75 (s, 1H, H-4), 6.91
(d, J = 7.5, 1H, CH@CHN), 7.39–7.49 (m, 6H, CH@CH + Ph), 16.81
(br s, 1H, C10-OH). ESI-MS: m/z 510 [M+H+] and 532 [M+Na+]. Anal.
(C32H31NO5) C, H, N.
d
J = 6.9 Hz, 3H, C100CH3), 1.28–1.42 (m, 2H, H-110), 1.66 (s, 3H, H-
7a), 1.77–1.94 (m, 4H, H-4000 + 2OH), 2.05–2.21 (m, 1H, H-100),
2.56–2.76 (m, 2H, H-100), 3.44–3.80 (m, 6H, H-3000 + H-5000 + H-300),
3.86–4.00 (m, 1H, H-200), 5.81 (dd, J = 7.6, 15.26 Hz, 1H, H-90),
6.05 (d, J = 15.3 Hz, 1H, H-40), 6.15 (dd, J = 11.4, 15.64 Hz, 1H,
–CH@CH–), 6.25 (dd, J = 11.8, 14.50 Hz, 1H, –CH@CH–), 6.54 (dd,
J = 10.7, 14.50 Hz, 1H, –CH@CH–), 6.71 (s, 1H, H-20), 6.90 (s, 1H,
H-4), 7.27 (dd, J = 11.4, 15.6 Hz, 1H, CH@CH), 7.74 (d, J = 14.1 Hz,
1H, H-1000), 10.28–10.46 (m, 1H, –NH). ESI-MS: m/z 528 [M+H+];
Anal. (C29H37NO8) C, H, N.
4.11. 2-Acetylamino-6-{[5-(2,3-dihydroxypropyl)-3-(1-hydroxy-
10-methyl-3-oxo-dodeca-1,4,6,8-tetraenyl)-7a-methyl-2,7-
dioxo-7,7a-dihydro-2H-benzofuran-6-ylidenemethyl]-amino}-
hexanoic acid (14c)
A solution of N -acetyl-
L-lysine (124 mg, 0.66 mmol) in water
a
(400
lL) was added to a solution of 2 (50 mg, 0.11 mmol) in CH3OH
(5 mL). The resulting mixture was stirred 2 h at room temperature.
The solvent was removed under reduced pressure and the crude was
4.14. 7-Benzyl-9a-methyl-3-octanoyl-6-propenyl-7H,9aH-furo-
[3,2-g]isoquinoline-2,9-dione (17)
washed with CH2Cl2/CH3OH 95:5; the excess of N -acetyl-L-lysine
a
was filtered off under vacuum and washed with CH2Cl2/CH3OH
5:1. The solution was then evaporated and the crude was purified
by preparative chromatography using CH2Cl2/CH3OH 85:15 as elu-
ent, to give the title compound as a red solid (28 mg, 39%). 1H
NMR (300 MHz, DMSO): 0.83 (t, J = 7.07 Hz, 3H, H-120), 0.98 (d,
J = 7.1 Hz, 3H, C100CH3), 1.19–1.41 (m, 4H, H-110 + CH2Lys), 1.44–
1.74 (m, 4H, CH2Lys), 1.82 (s, 3H, C7aCH3), 2.05–2.42 (m, 1H, H-
100), 2.50 (m, 2H, CH2Lys), 2.54–2.79 (m, 2H, H-100), 3.30–3.50 (m,
2H, H-300), 3.56–3.70 (m, 1H, H-200), 3.98–4.18 (m, 1H, –CHCOOH),
5.88 (dd, J = 7.8, 15.3 Hz, 1H, H-9), 6.08–6.30 (m, 2H, CH@CH),
6.31–6.47 (m, 1H, CH@CH), 6.63 (s, 1H, H-20), 6.71 (dd, J = 10.4,
14.5 Hz, 1H, CH@CH), 6.77 (s, 1H, H-4), 7.22 (dd, J = 11.5, 14.9 Hz,
1H, CH@CH), 7.86–8.06 (m, 2H, H-1000 + CONH), 10.22–10.42 (m,
1H, NH). ESI-MS: m/z 641 [M+H+]. Anal. (C34H44N2O10) C, H, N.
To a stirred solution of compound 6 (50 mg, 1 mequiv) in THF
(50 mL), benzylamine (21 lL, 1.5 mequiv) was added. The solu-
tion changed from orange to red. The reaction was stirred at room
temperature for 1 h. After evaporation of the solvent, the residue
was taken up with water and the aqueous phase was extracted
with AcOEt (100 mL). The organic phase was concentrated in va-
cuo to give a residue that was purified by PLC using hexane/
AcOEt 1:1 as eluent, to afford compound 17 (48 mg, 77%) as a
red solid. 1H NMR (300 MHz, CDCl3): d 0.87 (t, J = 7.1 Hz, 3H, H-
70), 1.22–1.42 (m, 10H, CH2), 1.62 (s, 3H, C9aCH3), 1.86 (dd, 3H,
J = 1.5, 6.3 Hz, C@CHCH3), 2.74–2.80 (m, 2H, H-20), 5.49 (s, 2H,
–CH2Ph), 6.51 (d, J = 16.0 Hz, 1H, –CH@CHCH3), 6.64 (m, 1H,
–CH@CHCH3), 6.68 (s, 1H, H-5), 7.04 (s, 1H, H-4), 7.27–7.48 (m,
5H, Ph), 8.23 (s, 1H, H-8). ESI-MS: m/z 494 [M+Na+]. Anal.
(C30H33NO4) C, H, N.
4.12. 5-(2,3-Dihydroxypropyl)-3-(1-hydroxy-10-methyl-3-oxo-
dodeca-1,4,6,8-tetraenyl)-7a-methyl-6-[(2-morpholin-4-yl-
ethylamino)-methylene]-6H,7aH-benzofuran-2,7-dione (14d)
4.15. 9a-Methyl-3-octanoyl-6-propenyl-7-(3,4,5-trihydroxy-6-
hydroxymethyl-tetrahydro-pyran-2-yl)-7H,9aH-furo[3,2-g]iso-
quinoline-2,9-dione (18)
Under nitrogen atmosphere 4-(2-aminoethyl) morpholine
(16 lL, 0.12 mmol) was added to a solution of 2 (56 mg,
0.12 mmol) in CH3OH (2 mL). The solution was stirred 23 h at room
temperature. The solvent was removed under reduced pressure
and the crude was purified by PLC using CH2Cl2/CH3OH 95:5 as elu-
ent to obtain the title compound as a red oil. (11 mg, 15%). 1H NMR
(300 MHz, CDCl3): d 0.88 (t, J = 7.07 Hz, 3H, H-120), 1.03 (d,
J = 6.70 Hz, 3H, C100CH3), 1.30–1.45 (m, 2H, H-110), 1.70 (s, 3H,
C7aCH3), 2.07–2.25 (m, 1H, H-100), 2.44–2.58 (m, 4H, H-
100 + NCH2), 2.59–2.76 (m, 4H, H-4000 + NCH2), 3.41–3.54 (m, 2H,
H-3000), 3.59 (dd, J = 6.3, 10.79 Hz, H-300), 3.69–3.85 (m, 5H,OCH2-
morpholine + H-300), 3.94–4.05 (m, 1H, H-200), 5.84 (dd, J = 8.2,
15.6 Hz, H-90), 6.03–6.36 (m, 3H, CH@CH), 6.57 (dd, J = 10.8,
14.1 Hz, 1H, CH@CH), 6.79 (s, 1H, H-20), 6.94 (s, 1H, H-4), 7.21–
7.39 (m, 1H, CH@CH), 7.73 (d, J = 14.5 Hz, 1H, H-1000), 10.38–10.55
(m, 1H, NH). ESI-MS: m/z 583 [M+H+]. Anal. (C32H42N2O8) C, H, N.
This compound was synthesized following the procedure de-
scribed by Lin et al.35 The 1H NMR matched with that reported
for the natural product.36
4.16. 2-Benzyl-3,6a-dimethyl-9-(2-methyl-dodecanoyl)-2H,6aH-
furo[2,3-h]isoquinoline-6,8-dione (19)
To a stirred solution of deflectin 13 (25 mg, 1 mequiv) in THF
(5 mL) benzylamine (8 lL, 1.2 mequiv) was added. The mixture
was maintained at room temperature for 10 min. After evaporation
of the solvent, the residue was dissolved in CH2Cl2 and purified by
PLC using hexane/acetone 7:3 as eluent to afford the single isomer
19 as a red solid (19 mg, 62%).1H NMR (400 MHz, acetone-d6): d
0.88 (3H, m, Me-80), 0.96 (3H, d, J = 7.0, Me-12), 1.1–1.5 (12H, m,
H2-20 and H2-70), 1.61 (3H, s, Me-7), 1.35 and 1.77 (2H, m, H2-10),
2.29 (3H, s, Me-3), 3.58 (1H, tq, J = 6.5, 7.0, H-12), 4.85 (1H, s, H-
5), 5.29 and 5.33 (2H, d, J = 16.3, CH2-Ph), 6.40 (1H, s, H-4), 7.38–
7.45 (5H, m, Ph), 8.73 (1H, s, H-1). ESI-MS: m/z 538 [M+Na+]. Anal.
(C33H41NO4) C, H, N.
4.13. 7-Benzyl-3-(1-hydroxy-10-methyl-3-oxo-dodeca-1,4,6,8-
tetraenyl)-9a-methyl-7H,9aH-furo[3,2-g]isoquinoline-2,9-dione
(16)
To a solution of 14a (50 mg, 0.09 mmol) in CH3OH/CH2Cl2 1:1
(2 mL) cooled at 0 °C were added water (0.4 mL) and NaIO4

L. Musso et al. / Bioorg. Med. Chem. 18 (2010) 6031–6043
6041
4.17. 6-(Benzylaminomethylene)-5-(2,3-dihydroxypropyl)-3-(1-
hydroxy-10-methyl-3-oxo-dodec-1-enoyl)-7a-methyl-6H,
7aH-benzofuran-2,7-dione (14e) and 3-(3-benzylamino-10-
methyldodec-2-enoyl)-6-(benzylaminomethylene)-5-(2,3-
dihydroxy-propyl)-7a-methyl-6H,7aH-benzofuran-2,7-dione (15)
Hsp90/inhibitor were formed by incubating the protein with a 10:1
molar excess of the individual inhibitor at 37 °C for 15 min prior to
proteolytic enzyme addition. Each complex was digested using a
1:100 (w/w) enzyme-to-substrate ratio. The extent of the reactions
was monitored on a time-course basis by sampling the incubation
mixture after 5, 15, and 30 min of digestion. Samples were desalted
by ziptip C4 (Millipore) and the proteolytic fragments were ana-
lyzed by MALDI-TOF/MS using a MALDI-MX micro (Waters). In
order to optimize sensitivity and accuracy of the mass measure-
ments, three different m/z ranges were explored for each sample:
a first range, from m/z 500 to 3500 was analyzed in reflector mode;
the other two ranges, from m/z 3500 to 20,000 and from m/z 20,000
to 95,000 were analyzed in linear mode. Each m/z range was cali-
brated using a suitable peptide or protein mixture. Mass data were
elaborated using the Masslynx software (Waters).
Preferential hydrolysis sites on Hsp90 under different condi-
tions were identified on the basis of the fragments released during
the enzymatic digestions. When comparative experiments were
carried out on Hsp90 in the presence or in the absence of inhibi-
tors, differences in the susceptibility of specific cleavage sites were
detected, from which protein regions involved in the conforma-
tional changes could be inferred.28
A solution of compound 5 (15 mg) in MeOH (1 mL) was stirred
at room temperature and benzylamine (12 lL) was added. The
mono-adduct 14e was isolated stopping the reaction after
30 min. The residue was purified by means of PLC eluting in hex-
ane/acetone 2:1, to give 16 mg (85%) of compound 14e. The bis-ad-
duct 15 was obtained by reaction of 14e (60 mg, 1 mequiv) with
benzylamine (57 lL, 4 mequiv) dissolved in MeOH (10 mL). The
mixture was stirred at room temperature for 72 h. After evapora-
tion of the solvent, the residue was purified first by FC with hex-
ane/acetone 6:4 as eluent and then by PLC eluting in CH2Cl2/
MeOH 40:1 to afford 26 mg (30%) of compound 15.
Compound 14e: 1H NMR (300 MHz, CDCl3): d 0.78–0.93 (m, 6H,
H-120 + C100CH3), 0.97–1.18 (m, 2H, H-110), 1.19–1.43 (m, 8H,
4CH2), 1.50–1.67 (m, 2H, CH2), 1.70 (s, 3H, C7aCH3), 1.94–2.11
(m, 1H, H-100), 2.42 (t, J = 7.2 Hz, CH2CO), 2.54–2.70 (m, 2H, H-
100), 3.55 (dd, J = 6.2, 11.0 Hz, 1H, H-300), 3.74 (dd, J = 3.3, 11.0 Hz,
1H, H-300), 3.87–3.97 (m, 1H, H-200), 4.75 (m, J = 5.5 Hz, 2H, –CH2Ph),
6.70 (s, 1H, H-20), 6.87 (s, 1H, H-4), 7.32–7.47 (m, 5H, Ph), 7.69 (d,
J = 13.24 Hz, 1H, H-1000), 10.53–10.71 (m, 1H, NH), 15.78 (br s, 1H,
OH). ESI-MS: m/z 566 [M+H+]. Anal. (C33H43NO7) C, H, N.
4.19. Cell lines and culture conditions
The human ovarian carcinoma cell line IGROV-1, the human
melanoma cell line JR8, and the human epithelial carcinoma cell
line A431 were routinely grown in RPMI 1640 (Lonza, Vierviers,
Belgium), supplemented with 10% (v/v) heat-inactivated fetal bo-
vine serum (GIBCO, Invitrogen, Paisley, UK). The human peritoneal
mesothelioma cell line STO was kindly provided by Dr. N. Zaffaroni
(IRCCS Istituto Nazionale Tumori, Milan) and grown in 50/50
DMEM-Ham’s F12 medium (Lonza, Vierviers, Belgium), supple-
mented with 10% (v/v) heat-inactivated fetal bovine serum and
Compound 15: 1H NMR (300 MHz, CDCl3): d 0.76–0.89 (m, 6H,
H-120 + C100CH3), 0.98–1.15 (m, 2H, H-110), 1.15–1.37 (m, 8H,
4CH2), 1.46–1.57 (m, 2H, CH2), 1.60 (s, 3H, C7aCH3), 2.22–2.33
(m, 2H, CH2CO), 2.45–2.63 (m, 2H, H-100), 3.38–3.54 (m, 1H, H-
300), 3.54–3.69 (m, 1H, H-300), 3.75–3.99 (m, 1H, H-200), 4.41–4.56
(m, 4H, 2CH2Ph), 6.24 (s, 1H, H-20), 6.94 (s, 1H, H-4), 7.14–7.39
(m, 10H, 2Ph), 7.61 (d, J = 12.97 Hz, H-1000), 10.41–10.55 (m, NH,
H-2000), 11.79–11.90 (m, NH, C30NH). ESI-MS: m/z 655 [M+H+] and
677 [M+Na+]. Anal. (C40H50N2O6) C, H, N.
Surface plasmon resonance analyses: SPR analyses were con-
ducted using a Biacore 3000 optical biosensor equipped with re-
search-grade CM5 sensor chips (Biacore AB). Using this platform,
two separate recombinant Hsp90 surfaces, a BSA surface and an
unmodified reference surface, were prepared for simultaneous
2 mM L-glutamine (L-Gln). All cell lines were maintained at 37 °C
in a 5%/95% CO2/air atmosphere.
4.20. Drugs and treatment conditions
All the compounds tested were freshly prepared dissolving
powder in dimethylsulfoxide (DMSO) (BDH Prolabo, Milan, Italy)
and following dilution in culture medium. 17-AAG was a gift from
Dr. N. Zaffaroni and it was provided as a stock solution (5 mM) that
we freshly diluted in culture medium before each experiment.
Cells were incubated with drugs for 6, 24 or 72 h (depending on
the assay), at 37 °C in culture medium supplemented with 10%
(v/v) heat-inactivated fetal bovine serum; drug concentrations
used for each assay are reported in Legends to Figures.
analyses. Proteins (100
were immobilized on individual sensor chip surfaces at a flow rate
of 5
L/min using standard amine-coupling protocols37 to obtain
densities of 8–12 kRU. Compound 2 and radicicol were dissolved
in 100% DMSO to obtain 4 mM solutions, and diluted 1:200 (v/v)
in PBS (10 mM NaH2PO4, 150 mM NaCl, pH 7.4) to a final DMSO
concentration of 0.5%. Compounds concentration series were pre-
pared as twofold dilutions into running buffer: for each sample
the complete binding study was performed using a six points con-
lg/mL in 10 mM sodium acetate, pH 5.0)
l
centration series, typically spanning 0.05–10 lM, and triplicate
4.21. Cell growth inhibition
aliquots of each compound concentration were dispensed into
single-use vials. Included in each analysis were multiple blank
samples of running buffer alone.38 Binding experiments were
Cell sensitivity to all the tested compounds was determined by
a growth inhibition assay. Briefly, cells were seeded in 12-well
plates (50,000 cells/well), 24 h before experiments. Cells were ex-
posed to the drugs for 72 h to allow approximately three replica-
tions in the control cells. Adherent cells were trypsinized and
counted with a cell counter (Coulter Electronics, Luton, UK). IC50
values, derived from dose–response curves, were defined as drug
concentrations required for 50% inhibition of cell growth.
performed at 25 °C, using a flow rate of 50
itoring of association and 200 s monitoring of dissociation. Simple
interactions were adequately fit to single-site bimolecular
lL/min, with 60 s mon-
a
interaction model (A + B = AB), yielding a single KD. Sensorgram
elaborations were performed using the Biaevaluation software pro-
vided by Biacore AB.39
4.18. Limited proteolysis
4.22. Western blot analysis
Limited proteolysis experiments were performed at 37 °C, PBS
Twenty-four hour after seeding (100,000 cells/mL) on Petri
dishes, cells were treated with drugs for 24 h and then processed
to obtain whole-cell extracts. Cells were rinsed twice with ice-cold
0.1% DMSO, at a 3
trypsin, chymotrypsin, and endoprotease V8 as proteolytic agents;
30 L of solution were used radical experiment. Binary complexes
lM recombinant Hsp90 concentration using
l

6042
L. Musso et al. / Bioorg. Med. Chem. 18 (2010) 6031–6043
PBS supplemented with 0.1 mM sodium orthovanadate and then
lysed in hot sample buffer.40 After determination of the protein
concentration by BCA Protein Assay (Thermo Scientific, Rockford,
4.25. Fluorescence polarization assay
GM-FITC, supplied by Invivogen (06C23-MT, California 92192,
USA), was previously dissolved in DMSO to obtain 10 mM stock
solutions and kept at ꢁ20 °C until use. HSP90, purchased from
Stressgen (cat. No. SPP-776, Victoria BC, Canada), was previously
IL, USA), cellular extracts (40 lg) were separated by sodium dode-
cylsulfate–polyacrylamide gel electrophoresis (SDS–PAGE) and
transferred onto nitrocellulose membranes. Immunoreactive bands
were revealed by enhanced chemiluminescence detection (Amer-
sham Biosciences, Rockford, IL, USA) using anti-Raf-1 and anti-
Cdk4 (Santa Cruz Biotechnology Inc., CA, USA), anti-EGFR (Upstate
Biotechnology, Millipore Corporate, Billerica, MA, USA), anti-Survi-
vin (Abcam, Cambridge, UK), anti-Akt (Transduction Laboratories,
Lexington, USA), anti-PARP-1 (Calbiochem, Merck Chemicals Ltd,
Nottingham, UK), anti-Cleaved Caspase-3 (Asp175) (Cell Signaling
Technology, Inc., MA, USA), anti-p53 (Dako, Glostrup, Denmark),
anti-Actin and anti-Tubulin antibodies (Sigma Chemical Co., St.
Louis, MO, USA).
dissolved in assay buffer (HFB) to form 2.2 lM stock solutions
and kept at ꢁ80 °C until use. The compounds were previously dis-
solved in DMSO to obtain stock solutions and kept at ꢁ20 °C. The
day of experiment, the compounds were prepared by serial dilu-
tions in assay buffer (HFB) containing 20 mM HEPES (K) pH 7.3,
50 mM KCl, 5 mM MgCl2, 20 mM Na2MoO4, and 0.01% NP40. Before
each use, 0.1 mg/mL bovine gamma globulin and 2 mM DTT were
freshly added.
Fluorescence polarization (FP) was performed in Opti-Plate™-
96F well plates (Perkin Elmer, Zaventem, Belgium) using a plate
reader (Wallac Envision 2101 multilabel reader, Perkin Elmer,
Zaventem, Belgium) To evaluate the binding affinity of the mole-
4.23. Coimmunoprecipitation and immunoblot analysis
cules, 50
of HSP90 in the presence of 5
l
L of the GM-FTC solution (5 nM) were added to 30 nM
L of the test compounds at increas-
l
Following the designated treatments, cells were lysed in NET
buffer [50 mmol/L Tris (pH 7.4), 150 mmol/L sodium chloride,
0.1% NP40, 1 mmol/L phenylmethylsulfonyl fluoride, 1 mmol/L
ing concentrations. The plate was mixed on a shaker at 4 °C for 4 h,
and the FP values in mP (millipolarization units) were recorded.
The IC50 values were calculated as the inhibitor concentration
where 50% of the tracer is displaced; each data point is the result
of the average of triplicate wells, and was determined from a plot
using nonlinear least-squares analysis. Curve fitting was per-
formed using Prism GraphPad software program (GraphPad soft-
ware Inc., San Diego, CA).
EDTA, 2.5 lg/mL leupeptin, 5 lg/mL aprotinin] first for 30 min on
ice and then for 30 min at 4 °C in rotation on a wheel. Nuclear
and cellular debris were cleared by centrifugation (10,000g,
10 min, 4 °C). Total cellular proteins were then quantified using
the BCA protein assay. Cell lysates (500 lg) were incubated with
the Raf-1-specific antibody (Santa Cruz Biotechnology Inc., CA,
USA) for 2 h at 4 °C. Protein A-sepharose (Sigma–Aldrich,
Germany) (90 lL), previously prepared in TNT solution (20 mM
4.26. Yeast Hsp82 protein expression and purification
Tris–HCl pH 7.4, 150 mM NaCl, 1% Triton X-100) was added to
the lysates and incubated overnight at 4 °C. The immunoprecipi-
tates (separated by centrifugation 15,000g, 2 min, 4 °C) were
washed twice in NET buffer and twice in PBS with 1% aprotinin
and 1 mmol/L PMSF. Proteins were eluted with the SDS sample
loading buffer before the immunoblot analyses with specific anti-
bodies against Raf-1, Hsp90 or Hsp70 (Santa Cruz Biotechnology
Inc., CA, USA).
Escherichia coli strain BL21-CodonPlus(DE3)-RIL (Stratagene; La
Jolla, CA) was transformed with the plasmid pET28-Hsp82/NdeI
containing the full-length Hsp82 gene of Saccharomyces cerevisi-
ae, with an N-terminal His6 tag for purification. Transformed cells
were selected on LB agar, containing 50
lg/mL kanamycin and
40 g/mL chloramphenicol, and a preliminary analysis on individ-
l
ual colonies was performed to evaluate their overexpressing
ability.
A 4 L culture of the clone with the highest Hsp82 expression
was grown in the presence of kanamycin until A600 reached 1.2,
and then expression of His-tagged Hsp82 was induced by the addi-
4.24. Determination of yeast Hsp90 ATPase activity
Inhibition of the intrinsic ATPase activity of full-length recombi-
nant yeast Hsp82 was measured by a modified enzymatically cou-
pled ATPase assay,41 in which the amount of inorganic phosphate
released by HSP90-dependent ATP hydrolysis is utilized by maltose
phosphorylase, which phosphorylates maltose and then cleaves
the disaccharide to produce glucose and glucose-1-phosphate. Glu-
cose is a substrate for glucose oxidase, which uses atmospheric
oxygen to produce gluconolactone and hydrogen peroxide. Finally,
horseradish peroxidase consumes hydrogen peroxide by the oxida-
tion of the colorless N-acetyl-3,7-dihydroxyphenoxazine (Amplex
Red) to provide the Resorufin product, which has a unique absorp-
tion spectra at 563 nm. For every molecule of ATP, that is, hydro-
lyzed by yeast Hsp90, one molecule of Amplex Red is ultimately
oxidized to Resorufin.
tion of 1 mM isopropyl-1-thio-b-D-galactopyranoside. Cells were
then grown for 3 h and harvested by centrifugation at 7000 rpm
for 30 min, at +4 °C. Harvested cells were washed in ice-cold
1ꢀ PBS, weighed and then stored at ꢁ70 °C. Frozen cells were
thawed and then re-suspended and lysed in B-PERÒ Bacterial Pro-
tein Extraction Reagent (Pierce; Rockford, IL), containing phenyl-
methylsulfonyl fluoride (PMSF). The lysate was centrifuged at
27,000g for 15 min at +4 °C and supernatant was ultra-centrifuged
at 100,000g for 45 min at +4 °C. Supernatant was then loaded, at a
flow rate of 1 mL/min, onto a Ni–NTA column (HisTrap™ HP Col-
umns, GE Healthcare) prepacked with NiSepharose and activated
according to manufacturer’s instructions. The column was washed
with potassium phosphate buffer (pH 7.4) containing 30 mM imid-
azole. A linear gradient of imidazole in potassium phosphate buffer
was finally applied to the column for Hsp82 elution. The collected
fractions were pooled, then subjected to gel filtration on Superdex
200 resin (GE Healthcare) in 20 mM Tris–HCl, pH 7.5, containing
300 mM KCl and 1 mM EDTA, and finally stored at ꢁ80 °C in the
presence of glycerol (10% final concentration).
Inhibition of yeast Hsp82 ATPase activity by test compounds
was tested, following 3 h incubation in the presence of 500 lM
ATP, in 96-well cell culture cluster clear plates by using a commer-
cially available kit (PiPer™ Phosphate Assay kit; Molecular Probes,
Eugene, OR), according to manufacturer’s instructions. To deter-
mine IC50 values, several appropriate concentrations of test com-
pounds were evaluated depending on their relative potency.
Each assay plate was run with a positive control which con-
sisted of three wells containing the well-known Hsp90 inhibitor
4.27. Animals
Experiments were carried out using female athymic Swiss nude
mice, 8–10 weeks old (Charles River, Calco, Italy). Mice were main-
17-DMAG, at a final concentration of 20 lM.

L. Musso et al. / Bioorg. Med. Chem. 18 (2010) 6031–6043
6043
8. Taldone, T.; Gozman, A.; Maharaj, R.; Chiosis, G. Curr. Opin. Pharmacol. 2008, 8,
370.
9. Biamonte, M. A.; Van de Water, R.; Arndt, J. W.; Scannevin, R. H.; Perret, D.; Lee,
W.-C. J. Med. Chem. 2010, 53, 3.
10. Stadler, M.; Anke, H.; Dekermendjian, K.; Reiss, R.; Sterner, O.; Witt, R. Nat.
Prod. Lett. 1995, 7, 7.
11. Powell, A. D. G.; Robertson, A.; Whalley, W. B. Chem. Soc. Spec. Publ. 1956, 5, 27.
12. Michael, A. P.; Grace, E. J.; Kotiw, M.; Barrow, R. A. J. Chem. 2003, 56, 13.
13. Yasukawa, K.; Takahashi, M.; Natori, S.; Kawai, K.; Yamazaki, M.; Takeuchi, M.;
Takido, M. Oncology 1994, 51, 108.
tained in laminar flow rooms keeping temperature and humidity
constant. Mice had free access to food and water. Experiments
were approved by the Ethics Committee for Animal Experimenta-
tion of the Istituto Nazionale Tumori of Milan according to institu-
tional guidelines.
4.28. Intraperitoneally growing tumor
14. Akihisa, T.; Tokuda, H.; Ukiwa, M.; Kivota, A.; Yasukawa, K.; Sakamoto, N.;
Kimura, Y.; Suzuki, T.; Takavasu, J. Chem. Biodivers. 2005, 2, 1305.
15. Akihisa, T.; Tokuda, H.; Yasukawa, K.; Ukiwa, M.; Kivota, A.; Sakamoto, N.;
Suzuki, T.; Tanabe, N.; Nishino, H.; Kavasu, J. J. Agric. Food Chem. 2005, 53, 562.
16. Su, N.-W.; Lin, Y.-L.; Lee, M.-H.; Ho, C.-H. J. Agric Food Chem. 2005, 53, 1949.
17. Knecht, A.; Humpf, H.-U. Mol. Nutr. Food Res. 2006, 50, 406.
18. Omura, S.; Tanaka, H.; Ikead, H.; Masuma, R. J. Antibiot. 1993, 46, 1908.
19. Kono, K.; Tanaka, M.; Ono, Y.; Hosoya, T.; Ogita, T.; Kohama, T. J. Antibiot. 2001,
54, 415.
20. Li, C. L.; Zhu, Y.; Wang, Y. Y.; Zhu, J. S.; Chang, J.; Kritchevsky, D. Nutr. Res. 1998,
18, 71.
21. Manchand, P. S.; Whalley, W. B.; Chen, F.-C. Phytochemistry 1973, 12, 2531.
22. Fielding, B. C.; Holker, J. S. E.; Jones, D. F.; Powell, A. D. G.; Richmond, K. W.;
Robertson, A.; Whalley, W. B. J. Chem. Soc. 1961, 4579.
23. Fielding, B. C.; Haws, E. J.; Holker, J. S. E.; Powell, A. D. G.; Robertson, A.;
Stanway, D. N.; Whalley, W. B. Tetrahedron Lett. 1960, 24.
24. Haws, E. J.; Holker, J. S. E.; Kelly, A.; Powell, A. D. G.; Robertson, A. J. Chem. Soc.
1959, 3598.
25. Anke, H.; Kemmer, T.; Hofle, G. J. Antibiot. 1981, 34, 923.
26. Coghlan, D. R.; Mackintosh, J. A.; Karuso, P. Org. Lett. 2005, 7, 2401.
27. Yuan, H.; Barnes, K. R.; Weissleder, R.; Cantley, L.; Josephson, L. Chem. Biol.
2007, 14, 321.
28. Birolo, L.; Dal Piaz, F.; Pucci, P.; Marino, G. J. Biol. Chem. 2002, 277, 17428.
29. Atkinson, R. A.; Joseph, C.; Dal Piaz, F.; Birolo, L.; Stier, G.; Pucci, P.; Pastore, A.
Biochemistry 2000, 39, 5255.
30. Dehner, A.; Furrer, J.; Richter, K.; Schuster, I.; Buchner, J.; Kessler, H.
ChemBioChem 2003, 4, 870.
The human IGROV-1 ovarian carcinoma was adapted to grow ip
and maintained by serial ip passages of ascitic cells into healthy
mice. Ascitic tumor cells 2.5 ꢀ 106 in 0.2 mL of saline were ip in-
jected. Mice developed hemorrhagic ascites and diffuse peritoneal
carcinomatosis and died by 15–30 days. Survival was the end point
of the study and the median day of death (median survival time:
MST) was calculated for each group. Antitumor activity was as-
sessed as T/C%, that is, the ratio of MST in treated over control
mice ꢀ 100. Treatment started the day after tumor cell injection,
delivering 2 ip every day for four days a week for two weeks
(qdx4/wx2w) at a dose of 25 mg/kg.
4.29. Statistical analysis
Curves reporting the percentage of survivors over time were
estimated by the Kaplan–Meyer product limit method and com-
pared with the log-rank test. Analyses were done with Graph Pad
Prism (GraphPad Software, Inc., San Diego, CA). The test was
two-sided.
31. Roe, S. M.; Prodromou, C.; O’Brien, R.; Ladbury, J. E.; Piper, P. W.; Pearl, L. H. J.
Med. Chem. 1999, 42, 260.
32. Giommarelli, C.; Zuco, V.; Favini, E.; Pisano, C.; Dal Piaz, F.; De Tommasi, N.;
Zunino, F. Cell. Mol. Life Sci. 2010, 67, 995.
Supplementary data
Supplementary data associated with this article can be found, in
the online version, at doi:10.1016/j.bmc.2010.06.068. These data
include MOL files and InChiKeys of the most important compounds
described in this article.
33. Li, Y.; Zhang, T.; Jiang, Y.; Lee, H.-F.; Schwartz, S. J.; Sun, D. Mol. Pharmacol.
2009, 6, 1152.
34. Akihisa, T.; Tokuda, H.; Yasukawa, K.; Ukiya, M.; Kiyota, A.; Sakamoto, N.;
Suzuki, T.; Tanabe, N.; Nishino, H. J. Agric. Food Chem. 2005, 53, 562.
35. Lin, T. F.; Yakushijin, K.; Büchi, G. H.; Demain, A. L. J. Ind. Microbiol. 1992, 9, 173.
36. Hajjkaj, H.; Klaébé, A.; Loret, M. O.; Tzédakis, T.; Goma, G.; Blanc, P. J. Appl.
Environ. Microbiol. 1997, 63, 2671.
References and notes
37. Johnsson, B.; Löfås, S.; Lindquist, G. Anal. Biochem. 1991, 198, 268.
38. Myszka, D. G. J. Mol. Recognit. 1999, 12, 279.
1. Whitesell, L.; Lindquist, S. L. Nat. Rev. Cancer 2005, 5, 761.
2. Morimoto, R. I.; Kline, M. P.; Bimston, D. N.; Cotto, J. J. Essays Biochem. 1997, 32,
17.
3. Prodromou, C.; Roe, S. M.; O’Brien, R.; Ladbury, J. E.; Piper, P. W.; Pearl, L. Cell
1997, 90, 65.
4. Donnelly, A.; Blagg, B. S. J. Curr. Med. Chem. 2008, 15, 2702.
5. Neckers, L. Trends Mol. Med. 2002, 8, S55.
6. Janin, Y. L. J. Med. Chem. 2005, 48, 7503.
39. Papalia, G. A.; Leavitt, S.; Bynum, M. A.; Katsamba, P. S.; Wilton, R.; Qiu, H.;
Steukers, M.; Wang, S.; Bindu, L.; Phogat, S.; Giannetti, M.; Ryan, T. E.; Pudlak,
V. A.; Matusiewicz, K.; Michelson, K. M.; Nowakowski, A.; Pahm-Baginski, A.;
Brooks, J.; Tieman, B. C.; Bruce, B. D. Anal. Biochem. 2006, 359, 94.
40. Zuco, V.; Zanchi, C.; Cassinelli, G.; Lanzi, C.; Supino, R.; Pisano, C.; Zanier, R.;
Giordano, V.; Garattini, E.; Zunino, F. Cell Death Differ. 2004, 11, 280.
41. Avila, C.; Hadden, M. K.; Ma, Z.; Kornilayev, B. Z.; Ye, Q.-Z.; Bòagg, B. S. J. Bioorg.
Med. Chem. Lett. 2006, 16, 3005.
7. Drysdale, M. J.; Brough, P. A. Drug Discovery Today 2008, 13, 38.