Inhibition of Hsp90 with resorcylic acid macrolactones: Synthesis and binding studies

Article abstract of DOI:10.1002/chem.201001119

A series of resorcylic acid macrolactones, analogues of the natural product radicicol has been prepared by chemical synthesis, and evaluated as inhibitors of heat shock protein 90 (Hsp90), an emerging attractive target for novel cancer therapeutic agents.

Full text of DOI:10.1002/chem.201001119

DOI: 10.1002/chem.201001119
Inhibition of Hsp90 with Resorcylic Acid Macrolactones:
Synthesis and Binding Studies
James E. H. Day,[a] Swee Y. Sharp,[b] Martin G. Rowlands,[b] Wynne Aherne,[b]
William Lewis,[a] S. Mark Roe,[c] Chrisostomos Prodromou,[c] Laurence H. Pearl,[c]
Paul Workman,[b] and Christopher J. Moody*[a]
Abstract: A series of resorcylic acid
macrolactones, analogues of the natural
product radicicol has been prepared by
chemical synthesis, and evaluated as in-
hibitors of heat shock protein 90
(Hsp90), an emerging attractive target
for novel cancer therapeutic agents.
The synthesis involves acylation of an
ortho-toluic acid dianion, esterification,
followed by a ring-closing metathesis
to form the macrocycle. Subsequent
manipulation of the protected hydroxy-
methyl side chain allows access to a
range of new analogues following de-
protection of the two phenolic groups.
Co-crystallization of one of the new
macrolactones with the N-terminal
domain of yeast Hsp90 confirms that it
binds in a similar way to the natural
product radicicol and to our previous
synthetic analogues, but that the intro-
duction of the additional hydroxymeth-
yl substituent appears to result in an
unexpected change in conformation of
the macrocyclic ring. As a result of this
conformational change, the compounds
bound less favorably to Hsp90.
Keywords: anticancer
agents
·
enzyme inhibitors · macrocycles ·
macrocyclic compounds
· meta-
thesis · proteins
Introduction
folded conformations that are rapidly ubiquinated and sub-
ject to proteasomal degradation, and thereby disrupt multi-
ple cancer causing pathways simultaneously. The approach
has been successfully validated by the entry of at least 12
Hsp90 inhibitors into clinical trial.[11,12]
One of the most promising cancer therapeutic strategies to
emerge in recent years is the inhibition of heat shock pro-
tein 90 (Hsp90),[1–9] an abundant ATP-dependent chaperone
that is central to the regulation and maintenance of a range
of proteins, its so-called clients proteins.[5,10] Although many
of these clients are essential to normal cellular processes,
others are known oncogenic proteins such as CRAF, BRAF,
ERBB2, AKT, telomerase and mutant p53. Hence inhibition
of Hsp90 can cause client proteins to adopt abnormally
As with many other areas of medicinal chemistry, initial
leads came from the natural world. Two natural products,
geldanamycin (1) and radicicol (2) (see below) were discov-
ered to be potent inhibitors of Hsp90, binding to the ATP-
binding domain in the N-terminal region of the protein.[13–15]
Although subsequent work has led to the discovery of fur-
ther ansamycin geldanamycin analogues, three of which are
now in clinical trial,[12] the early promise of radicicol, a
member of the resorcylic acid lactone family,[16–19] against
Hsp90 has yet to translate into a drug candidate molecule
based on the molecular framework of the natural product.
Whilst remaining one of the most potent Hsp90 inhibitors in
vitro,[20] radicicol has little or no activity in vivo,[21,22] pre-
sumably as a result of its reactive epoxide and dienone moi-
eties. Nevertheless, the compound continues to attract con-
siderable attention, and a number of naturally derived and
synthetic analogues, some of which do show in vivo activity,
have been described.[23–28]
[a] Dr. J. E. H. Day, Dr. W. Lewis, Prof. Dr. C. J. Moody
School of Chemistry, University of Nottingham
University Park, Nottingham, NG7 2RD (UK)
Fax : (+44)115-951-3564
E-mail: c.j.moody@nottingham.ac.uk
[b] Dr. S. Y. Sharp, Dr. M. G. Rowlands, Dr. W. Aherne,
Prof. Dr. P. Workman
Cancer Research UK Centre for Cancer Therapeutics
The Institute of Cancer Research, Haddow Laboratories
15 Cotswold Road, Sutton, Surrey, SM2 5NG (UK)
[c] Dr. S. M. Roe, Dr. C. Prodromou, Prof. Dr. L. H. Pearl
Section of Structural Biology, The Institute of Cancer Research
Chester Beatty Laboratories, 237 Fulham Road
London SW3 6JB (UK)
Our own work in this area has involved the synthesis and
biological evaluation of a series of novel resorcylic acid mac-
rolactones of varying ring size and conformation.[29] One of
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/chem.201001119.
10366
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 10366 – 10372

FULL PAPER
Scheme 1. Retrosynthetic analysis of the resorcylic acid macrolactones 4.
these analogues, NP261 (3), exhibited potent Hsp90 binding,
with an IC50 =~40 nm as measured by a fluorescence polari-
zation (FP) binding assay, and in addition showed the estab-
lished molecular signature of Hsp90 inhibitors, that is, deple-
tion of client proteins with upregulation of Hsp70 in treated
cancer cells. X-Ray crystallography studies revealed that 3
also was bound to the N-terminal ATP site of Hsp90 in a
comparable way to the structurally more complex natural
product.[29] The molecule adopts a similar folded conforma-
tion with the same key hydrogen-bonding interactions be-
tween the salicylate ester and phenolic groups and the pro-
tein (the carboxylate side-chain of Asp79, the main-chain
amide group of Gly83, the hydroxyl side-chain of Thr171,
the main-chain carbonyl of Leu34, and conserved water
molecules; yeast Hsp90 numbering) (Figure 2C), although it
obviously lacks the interaction with the e-amino side-chain
of Lys44, present through the epoxide oxygen in the natural
product. In an attempt to restore this potentially useful H-
bonding interaction, we now report the synthesis of a new
series of resorcylic acid macrolactones incorporating heter-
oatoms, with the potential to bind to the side chain of
Lys44, thereby increasing the potential potency of the com-
pounds.[30]
(PMB) group to enable subsequent manipulation of this
substituent. In common with our recent approach based on
isocoumarins,[30] it also incorporates the chlorine atom from
the start to obviate the need for late stage chlorination of
every compound. Hence the problem reduces to the synthe-
sis of the benzylic ketone 5, and aside of the aforementioned
approach based on isocoumarins, used by Lett et al. in the
original synthesis of radicicol,[31,32] and Danishefskyꢁs elegant
use of the Diels–Alder reaction to construct the benzene
ring,[33] most routes disconnect at the benzylic bond by some
sort of acylation (or equivalent) reaction. Thus, benzylic
chlorides 6 (Y=Cl) have been coupled with 1,3-dithiane
anions, equivalent to an acyl anion derived from carbonyl
compound 7 (Z=H),[29,34] or toluate anions derived from 6
(Y=H) have been acylated with, for example, Weinreb
amides 7 (Z=NMeOMe). This latter approach, extensively
employed by Winssinger and colleagues,[26,28,35–37] has proved
successful for a range of resorcylic acid lactone derivatives,
and therefore seemed ideally suited to our needs, requiring
the known toluate ester 8 featuring ethoxymethyl protecting
groups for the two phenols,[37] and Weinreb amide 9 as start-
ing materials.
The Weinreb amide 9 could be prepared in three steps as
shown in Scheme 2. 3-Chloro-2-(chloromethyl)prop-1-ene
was reacted with the sodium salt of p-methoxybenzyl alco-
hol in THF to give the known 3-chloro-2-(4-methoxybenzy-
loxy)methylpropene (10).[38] Copper-catalyzed reaction[39] of
the allyl chloride 10 and the organozinc derived from ethyl
iodobutyrate[40] gave ester 11 in good yield. Direct conver-
sion into the Weinreb amide 9 was achieved in excellent
yield using the isopropylmagnesium chloride mediated reac-
tion with N,O-dimethylhydroxylamine hydrochloride devel-
oped by Williams et al.[41]
Results and Discussion
It was clear from our earlier studies that an additional sub-
stituent on the macrolactone ring was required to form an
H-bond to the side-chain of Lys44,[29,30] and examination of
the Hsp90 bound structure of 3 suggested that compounds
of the general structure 4 (X=OR, NR2) might possess the
desired characteristics, and therefore attention was focused
on their synthesis. Many recent routes to resorcylic acid
macrolactones, including our own,[29] rely on a ring-closing
metathesis (RCM) reaction to form the macrocyclic ring,
and given the reliability of this approach we saw no reason
to change our strategy. Our approach to the required 14-
membered macrolactone is outlined in Scheme 1, and incor-
porates an allylic alcohol protected with a p-methoxybenzyl
The known ortho-toluate ester 8[37] in THF was treated
with LDA (2.5 equiv) at À788C, and the deep red colored
solution was immediately treated with the Weinreb amide 9.
However, after work-up none of the desired acylated prod-
uct was isolated. When the corresponding isopropyl ester
was used, a mixture of decomposition products was ob-
served. Previous studies with toluate anions have shown that
Chem. Eur. J. 2010, 16, 10366 – 10372
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
10367

C. J. Moody et al.
The PMB-protecting group could be removed oxidatively
from macrocycle 14 using cerium(IV) ammonium nitrate
(CAN) to give allyl alcohol 15 in reasonable yield
(Scheme 4), ready for conversion into other substituents.
Scheme 2. a) NaH, 4-MeOC6H4CH2OH, THF, 08C to RT (56%); b)
EtO2CCH2CH2CH2ZnI (from the iodide, Zn in DMA), DMA/THF,
CuCN, 708C (68%); c) i) MeONHMe.HCl, iPrMgCl, THF, À20 to À58C;
ii) saturated aq. NH4Cl, À58C to RT (92%).
by-products such as those resulting from self condensation
are common in such reactions.[42] In order to circumvent the
formation of decomposition products, it was reasoned that
use of the dianion from an ortho-toluic acid instead of the
ester-derived monoanion might be preferable, since such
dianions have been successfully used in synthesis previous-
ly.[43–46] Thus the corresponding toluic acid 12[47] was treated
with 2.2 equivalents of sec-butyllithium. Quenching the re-
sulting dilithiated toluate with the Weinreb amide 9 resulted
in a mixture of starting acid 12 and acylated material after
work-up. The crude mixture was reacted with 3-butenol
under Mitsunobu conditions [diisopropyl azodicarboxylate
(DIAD), Ph3P and THF], and product could be easily sepa-
rated to give the required ester 13 in 41% yield over two
steps. Ring-closing metathesis using Grubbsꢁ second-genera-
Scheme 4. a) CAN, MeCN/H2O, 0 8C (54%); b) Ac2O, NEt3, CH2Cl2
(75%); c) i) MsCl, NEt3, CH2Cl2, 08C; ii) LiBr, THF/CH2Cl2, 08C; iii)
morpholine, NaI, iPr2NEt, DMF (75% over three steps); d) i) MsCl,
ꢀ
NEt3, CH2Cl2, 08C; ii) NaN3, DMF; iii) EtO2CC CH, CuSO4, Na-ascor-
bate, tBuOH (20% over three steps).
tion catalyst {benzylidene
[1,3-bis(2,4,6,-trimethylphenyl)-2-
Thus, allyl alcohol 15 could be acetylated using acetic anhy-
dride to give acetate 16, or mesylated and then reacted with
lithium bromide in a mixture of tetrahydrofuran and di-
chloromethane. The resulting allyl bromide was reacted with
morpholine in the presence of sodium iodide and Hunigꢁs
base to give tertiary amine 17 in good yield over three steps.
In addition, the mesylate was reacted with sodium azide and
the product subjected to the copper-catalyzed Huisgen cy-
cloaddition[48] with ethyl propiolate to give triazole 18 in
20% over three steps.
imidazolidinylidene]dichloro(tricyclohexylphosphine)ruthe-
nium} proceeded in excellent yield to give the desired mac-
rocycle 14, unfortunately as an inseparable mixture (48:52)
of double bond isomers (Scheme 3). The reaction was most
conveniently carried out on about 50 mg scale since increas-
ing the scale of the reaction (>200 mg) resulted in much
longer reaction times and reduced yields.
Removal of the EOM-protecting groups from compounds
14–18 using TFA in dichloromethane or HCl in dioxane was
readily achieved in moderate to excellent yield (26–98%) to
give a range of novel radicicol analogues 19–23 (Scheme 5).
After deprotection of PMB-ether 14, both Z and E isomers
of compound 19 could be separated by chromatography, the
double-bond assignment being based on NOE spectroscopy.
Thus, in the Z isomer an NOE enhancement was seen be-
tween the alkene proton at C5 and the methylene at C7.
Also, the E isomer of allyl alcohol 20 was successfully re-
crystallized (from the E/Z mixture) and its structure con-
firmed by X-ray crystallography (Figure 1).
The macrocycle 20 was co-crystallized (as its Z isomer
from the E/Z mixture) with the N-terminal domain of yeast
Hsp90, and the structure of the resulting complex solved by
molecular replacement. Comparison with the previously de-
termined structure of Hsp90-bound NP261 3[29] showed that
Scheme 3. For compound 12, a) i) sBuLi, THF, À788C; ii) 9, THF,
À788C; b) DIAD, PPh3, THF, HOCH2CH2CHCH2, 08C to RT (41%
over two steps); c) Grubbsꢁ II catalyst, CH2Cl2, 458C (92%).
10368
www.chemeurj.org
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 10366 – 10372

Hsp90 Inhibitors
FULL PAPER
Table 1. Biological activity of the macrocyclic lactones 19–23 compared
to radicicol 2 and NP261 3.
Entry Compound
R
TR-Fret
FP IC50 HCT116 SRB
IC50 [mm]
[mm]
GI50 [mm]
1
2
3
2
3
–
–
0.011
0.35
18% @
10 mm
7.70
0.0043
0.038
0.85
0.00061
7.6
9.4
(E)-19
OPMB
4
5
6
7
(Z)-19
20
21
OPMB
OH
OAc
1.65
0.65
1.90
4.80
8.2
45
10
34
Scheme 5. a) TFA, CH2Cl2; b) HCl, dioxane.
0.90
4.20
22
morpholinyl 32.5% @
10 mm
8
23
triazolyl
9.60
3.85
43
the macrocyclic ring results in a significant reduction in
binding to Hsp90 compared to the quite potent (IC50 ~40 nm
in FP assay) NP261 3. However, compounds 19–23 are only
marginally less potent than 3 in the SRB growth inhibition
assay. The loss of potency in inhibition of Hsp90 is presuma-
bly a result of the change in conformation of the macrolac-
tone ring upon introduction of the substituent at C6, some-
thing that was not foreseen in our original analysis.
Figure 1. X-ray crystal structure of the E isomer of resorcyclic acid mac-
rolactone 20.
Conclusion
the compounds bind with overall similarity, except for the
macrolactone ring between carbons atoms C6 to C9 (same
numbering as Figure 1), where a distinctly different confor-
mation is adopted (Figure 2B). Unfortunately, the electron
density for macrocycle 20 in this region of the macrolactone
ring was weak relative to that for the structure of 3, and
therefore the H-bonding interactions with the protein are
less easy to define with certainty. Figure 2C shows all the
possible bonding interactions for the most populated confor-
mation of 20, and illustrates that our planned H-bond be-
tween the side chain of Lys44 and the hydroxymethyl sub-
stituent at C6 is indeed present, notwithstanding the fact
that the macrocyclic ring now adopts an unexpected confor-
mation. The glycerol molecule, which was found in complex
with 20, probably originates from the cryoprotection of the
crystals and forms a number of hydrogen bounds with com-
pound 20. It is likely that these bonds were also previously
present, but mediated via a water molecule, that may have
then been displaced by the incoming glycerol molecule.
The macrocyclic lactones 19–23 were evaluated for bind-
ing to the ATP site of the N-terminal domain of human
Hsp90b in two Hsp90 assays: the fluorescence polarization
(FP) assay,[49,50] and the TR-FRET assay.[51] Their growth in-
hibitory activity against a human colon cancer cell line
(HCT116) was also measured by the SRB assay. As can be
seen in Table 1, introduction of groups into the 6-position of
In summary, Hsp90 is an important drug target for molecu-
lar cancer therapeutics, and we have previously synthesized
and evaluated a simple analogue of the natural product radi-
cicol known as NP261 3. In an attempt to increase potency
by introducing additional H-bonding substituents, we decid-
ed to incorporate heteroatom groups at the C6-position of
the NP261 macrocycle. Unfortunately, as demonstrated by
X-ray crystallography, this induced a significant and unex-
pected change in conformation of the macrocyclic ring, re-
sulting in compounds that bound less favorably than the un-
substituted ring. The unexpected change in ring conforma-
tion suggests an alternative approach to the design of ana-
logues of macrolactone 3 for Hsp90 inhibition.
Experimental Section
Full experimental procedures and spectroscopic characterization data are
given in the Supporting Information. Data for the compounds that were
evaluated in biological assays are given below.
(Z) and (E)-13-Chloro-14,16-dihydroxy-6-(4-methoxybenzyloxy)methyl-
3,4,7,8,9,10-hexahydro-1H-benzo[c][1]oxacyclotetradecine-1,11
ACHTUNGTRENNUNG(12H)-
dione (19)
Z Isomer: colorless solid; m.p. 142–1448C; 1H NMR (400 MHz, CDCl3):
d=11.72 (s, 1H, OH), 7.20 (d, 2H, J = 8.4 Hz, ArH), 6.86 (d, 2H, J =
8.4 Hz, ArH), 6.57 (s, 1H, ArH), 6.20 (s, 1H, OH), 5.41 (t, 1H, J =
6.9 Hz, =CH), 4.41 (t, 2H, J = 5.6 Hz, CH2), 4.37 (s, 2H, CH2), 4.22 (s,
Chem. Eur. J. 2010, 16, 10366 – 10372
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
10369

C. J. Moody et al.
E Isomer: colorless solid; m.p. 164–
1668C; 1H NMR (400 MHz, CDCl3):
d=7.26 (d, 2H, J
= 8.6 Hz, ArH),
6.88 (d, 2H, J = 8.6 Hz, ArH), 6.58 (s,
1H, ArH), 6.16 (s, 1H, OH), 5.51 (t,
1H,
J
=
6.8 Hz, =CH), 4.44 (s,
5.6 Hz,
2H,CH2), 4.42 (t, 2H,
J
=
CH2), 4.30 (s, 2H, CH2), 3.94 (s, 2H,
CH2), 3.80 (s, 3H, CH3), 2.63–2.55 (m,
4H, 2ꢂCH2), 2.21 (t, 2H, J = 6.4 Hz,
CH2), 1.73 (quin, 2H,
J = 6.6 Hz,
CH2), 1.58–1.56 ppm (m, 2H, CH2);
phenolic OH not observed; 13C NMR
(125 MHz, CDCl3): d=206.5 (C),
170.5 (C), 163.3 (C), 159.2 (C), 156.3
(C), 138.9 (C), 136.1 (C), 130.2 (CH),
129.3 (CH), 125.9 (CH), 114.7 (C),
113.8 (CH), 107.3 (C), 103.5 (CH),
72.3 (CH2), 67.3 (CH2), 66.2 (CH2),
55.3 (CH3), 46.8 (CH2), 40.8 (CH2),
35.0 (CH2), 27.0 (CH2), 23.2 (CH2),
22.0 ppm (CH2); IR (dichlorome-
thane):
nmax =3684,
3512,
1720,
1657 cmÀ1; MS (ESI): m/z (%): 513/
511 (37/100) [M+Na]+, 508/506 (4/2);
m/z:
calcd
for
C26H2935ClNaO7:
511.1494; found: 511.1482 [M+Na]+.
13-Chloro-14,16-dihydroxy-6-hydroxy-
methyl-3,4,7,8,9,10-hexahydro-1H-
benzo[c][1]oxacyclotetradecine-1,11-
AHCTUNGTREG(NNUN 12H)-dione (20)
Mixture of double bond isomers 49:51;
colorless solid; m.p. 203–2068C;
1H NMR (400 MHz, [D6]acetone):
mixture of isomers d=9.72 (brs, 1H,
OH), 6.53 (s, 1H, ArH), 6.52 (s, 1H,
ArH), 5.54 (t, 1H, J = 6.9 Hz, =CH),
5.35 (t, 1H, J = 6.9 Hz, =CH), 4.44 (t,
4H, J
= 5.6 Hz, CH2), 4.40 (s, 2H,
CH2), 4.33 (s, 2H, CH2), 4.09 (s, 2H,
CH2), 4.05 (s, 2H, CH2), 2.66–2.56 (m,
8H, 3ꢂCH2), 2.24–2.19 (m, 4H, 2ꢂ
CH2), 1.73 (quin, 2H,
J = 6.6 Hz,
Figure 2. A) Cartoon showing NP261 3 (left) and macrocycle 20 (right) bound in the N-terminal domain of
yeast Hsp90. B) Two orthogonal views of the superimposition of NP261 3 (green) and 20 (yellow), showing
that both compounds bind with overall similarity, except for the macrolactone ring of compound 20 between
carbons atoms C6 to C9 (same numbering as Figure 1). The structures for both compounds were obtained at
2.2 ꢃ resolution. For atomic coordinates and structure factors, see PDB codes 2CGF and 2XD6, respectively.
C) PyMOL diagram showing binding interactions of NP261 3 (left) and 20 (right). The electron density for
macrocycle 20 from carbon 6 to 9 of the macrolactone ring was very weak relative to that for NP261 structure.
However, we have shown all the possible bonding interactions for the most populated conformation of 20. It
should be noted that bond interactions to the disordered section of 20 are therefore weak. Dotted blue lines,
hydrogen bonds; green, amino acid residues involved in H-bonding; cyan-colored spheres, water molecules;
cyan residues, residues solely in van der Waals contact; magenta, glycerol molecule.
CH2), 1.67–1.55 ppm (m, 6H, 3ꢂCH2);
13C NMR (125 MHz, [D6]acetone):
mixture of isomers d=205.0 (C), 171.4
(C), 170.8 (C), 163.1 (C), 162.8 (C),
158.0 (C), 157.99 (C), 141.6 (C), 141.5
(C), 137.5 (C), 137.3 (C), 123.6 (CH),
122.4 (CH), 115.0 (C), 107.0 (CH),
106.6 (CH), 102.8 (CH), 102.7 (CH),
67.2 (CH2), 66.4 (CH2), 65.7 (CH2),
59.3 (CH3), 46.1 (CH2), 45.7 (CH2),
40.9 (CH2), 40.1 (CH2), 33.5 (CH2),
26.6 (CH2), 26.4 (CH2), 25.8 (CH2),
25.0 (CH2), 23.0 (CH2), 22.0 (CH2),
21.5 ppm (CH2); IR (dichlorome-
thane): nmax =3552, 3491, 1692, 1645 cmÀ1; MS (ESI): m/z (%): 393/391
(34/100) [M+Na]+; m/z: calcd for C18H2135ClNaO6: 391.0919; found:
391.0923 [M+Na]+.
2H, CH2), 3.87 (s, 2H, CH2), 3.81 (s, 3H, CH3), 2.54–2.49 (m, 4H, 2ꢂ
CH2), 2.20–2.17 (m, 2H, CH2), 1.68 (quin, 2H, J = 6.2 Hz, CH2), 1.59–
1.57 ppm (m, 2H, CH2); 13C NMR (125 MHz, CDCl3): d=206.3 (C),
171.1 (C), 163.8 (C), 159.3 (C), 156.2 (C), 138.7 (C), 136.0 (C), 130.3 (C),
129.4 (CH), 125.5 (CH), 114.6 (C), 113.8 (CH), 107.3 (C), 103.6 (CH),
73.9 (CH2), 72.0 (CH2), 67.0 (CH2), 55.3 (CH3), 46.2 (CH2), 41.6 (CH2),
26.9 (CH2), 26.3 (CH2), 25.0 (CH2), 21.8 ppm (CH2); IR (dichlorome-
thane): nmax =3514, 1721, 1660, 1611, 1590, 1513 cmÀ1; MS (ESI): m/z
(%): 513/511 (35/100) [M+Na]+, 508/506 (7/21) [M+H4N]+; m/z: calcd
for C26H2935ClNaO7: 511.1494; found: 511.1474 [M+Na]+.
(13-Chloro-14,16-dihydroxy-1,11-dioxo-3,4,7,8,9,10,11,12-octahydro-1H-
benzo[c][1]oxacyclotetradecin-6-yl)methyl acetate (21)
1
Colorless oil; H NMR (400 MHz, CDCl3): mixture of isomers d=6.61 (s,
1H, ArH), 6.60 (s, 1H, ArH), 6.16 (m, 2H, OH), 5.52–5.45 (m, 2H, =
CH), 4.55 (s, 4H, CH2), 4.46 (q, 4H, J = 5.5 Hz, 2ꢂCH2), 4.31 (s, 2H,
CH2), 4.26 (s, 2H, CH2), 2.63–2.58 (m, 6H, 3ꢂCH2), 2.52 (t, 2H, J =
6.4 Hz, CH2), 2.22–2.16 (m, 4H, 2ꢂCH2), 2.10 (s, 3H, CH3), 2.02 (s, 3H,
CH3), 1.82 (m, 2H,CH2), 1.74 (m, 2H,CH2), 1.66–1.61 ppm (m, 2H,CH2);
10370
www.chemeurj.org
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 10366 – 10372

Hsp90 Inhibitors
FULL PAPER
13C NMR (125 MHz, CDCl3): mixture of isomers d=206.5 (C), 206.2 (C),
171.1 (C), 171.0 (C), 170.9 (C), 170.4 (C), 163.8 (C), 163.7 (C), 156.37
(C), 156.36 (C), 136.5 (C), 136.4 (C), 136.0 (C), 135.9 (C), 127.2 (CH),
126.8 (CH), 114.66 (C), 114.63 (C), 107.18 (C), 107.14 (C), 103.6 (CH),
68.0 (CH2), 66.7 (CH2), 66.0 (CH2), 62.2 (CH2), 46.8 (CH2), 46.2 (CH2),
41.6 (CH2), 40.5 (CH2), 34.2 (CH2), 27.0 (CH2), 26.9 (CH2), 26.5 (CH2),
25.0 (CH2), 22.9 (CH2), 22.0 (CH2), 21.7 (CH2), 21.0 (CH2), 20.9 ppm
built using SKETCHER. The inhibitor molecule and the waters were
added in the final stages.
Biology: Compounds were assayed for their ability to bind to Hsp90
using two assays.
TR-Fret assay: A highly robust time-resolved fluorescence energy trans-
fer (TR-FRET) assay was used to measure the binding of biotinylated
geldanamycin (600 nm) to the full length human Hsp90 His-tagged pro-
tein (40 nm), as described previously.[51] Briefly, the europium labelled
anti-His-tagged protein antibody (Perkin–Elmer Prod. No. AD0110) was
added at 1 nm and the streptavidin Sureligh APC (Perkin–Elmer Prod.
No. AD0201) at 90 nm. Compounds were tested across a ten point con-
centration range up to 10 mm and the IC50 determined.
(CH2); IR (dichloromethane): nmax =3686, 3512, 1732, 1659, 1606 cmÀ1
;
MS (ESI): m/z (%): 435/433 (35/100) [M+Na]+, 430/428 (7/21)
[M+H4N]+; m/z: calcd for C20H2335ClNaO7: 433.1025; found: 433.1024
[M+Na]+.
13-Chloro-14,16-dihydroxy-6-(morpholinomethyl)-3,4,7,8,9,10-hexahydro-
1H-benzo[c][1]oxacyclotetradecine-1,11ACTHNUGRTENUNG(12H)-dione (22)
FP assay: This is a measurement of binding competition with a fluores-
Off white foam; m.p. 85–898C; 1H NMR (400 MHz, CDCl3): mixture of
isomers d=11.72 (s, 1H, OH), 11.67 (s, 1H, OH), 6.62 (s, 1H, ArH), 6.60
(s, 1H, ArH), 5.49–5.37 (m, 2H, =CH), 4.42 (t, 4H, J = 4.9 Hz, CH2),
4.31 (s, 2H, CH2), 4.27 (s, 2H, CH2), 3.67 (brs, 8H, 4ꢂCH2), 2.90 (brs,
2H, CH2), 2.84 (brs, 2H, CH2), 2.58 (m, 6H, 3ꢂCH2), 2.48 (t, 2H, J =
6.4 Hz, CH2), 2.36 (brs, 8H, 4ꢂCH2), 2.20 (m, 4H, 2ꢂCH2), 1.76–
1.59 ppm (m, 16H, 8ꢂCH2); 13C NMR (125 MHz, CDCl3): mixture of iso-
mers d=206.94 (C), 206.86 (C), 171.1 (C), 170.6 (C), 163.7 (C), 163.5 (C),
157.0 (C), 138.1 (C), 137.6 (C), 136.0 (C), 136.0 (C), 126.8 (CH), 126.0
(CH), 115.1 (C), 115.0 (C), 106.9 (C), 106.7 (C), 103.62 (CH), 103.59
(CH), 66.9 (CH2), 66.7 (CH2), 66.2 (CH2), 65.4 (CH2), 57.4 (CH2), 53.7
(CH2), 53.6 (CH3), 46.7 (CH2), 46.2 (CH2), 41.1 (CH2), 40.7 (CH2), 35.0
(CH2), 27.1 (CH2), 26.93 (CH2), 26.91 (CH2), 25.1 (CH2), 23.1 (CH2), 22.0
(CH2), 21.6 ppm (CH2); IR (dichloromethane): nmax =3627, 3512, 1720,
1659, 1592 cmÀ1; MS (ESI): m/z (%): 440/438 (35/100) [M+H]+; m/z:
calcd for C22H2935ClNO6: 438.1683; found: 438.1704 [M+H]+.
cent probe as described previously.[49,50]
Growth inhibition assay: The colorimetric sulforhodamine B assay (SRB)
was used to measure growth inhibition studies as described previously.[53]
The IC50 was calculated as the drug concentration that inhibits cell
growth by 50% compared with control growth.
Acknowledgements
This work was supported by Cancer Research UK (CUK) grant numbers
C215A7606 (C.J.M.), CA309A8274 (P.W.), and NHS funding to the
NIHR Biomedical Research Centre. P.W. is a Cancer Research UK Life
Fellow.
Ethyl 1-((13-chloro-14,16-dihydroxy-1,11-dioxo-3,4,7,8,9,10,11,12-octahy-
dro-1H-benzo[c][1]oxacyclotetradecin-6-yl)methyl)-1H-1,2,3-triazole-4-
carboxylate (23)
[1] P. Workman, F. Burrows, L. Neckers, N. Rosen, Ann. N. Y. Acad.
Sci. 2007, 1113, 202.
[2] D. B. Solit, G. Chiosis, Drug Discovery Today 2008, 13, 38.
[3] T. Taldone, A. Gozman, R. Maharaj, G. Chiosis, Curr. Opin. Phar-
macol. 2008, 8, 370.
[4] M. J. Drysdale, P. A. Brough, Curr. Top. Med. Chem. 2008, 8, 859.
[5] L. H. Pearl, C. Prodromou, P. Workman, Biochem. J. 2008, 410, 439.
[6] D. Mahalingam, R. Swords, J. S. Carew, S. T. Nawrocki, K. Bhalla,
F. J. Giles, Br. J. Cancer 2009, 100, 1523.
[7] T. Taldone, W. Sun, G. Chiosis, Bioorg. Med. Chem. 2009, 17, 2225.
[8] J. S. Hahn, BMB Rep. 2009, 42, 623.
[9] Y. L. Janin, Drug Discovery Today 2010, 15, 342.
[10] L. H. Pearl, C. Prodromou, Annu. Rev. Biochem. 2006, 75, 271.
[11] Y. S. Kim, S. V. Alarcon, S. Lee, M. J. Lee, G. Giaccone, L. Neckers,
J. B. Trepel, Curr. Top. Med. Chem. 2009, 9, 1479.
[12] M. A. Biamonte, R. Van de Water, J. W. Arndt, R. H. Scannevin, D.
Perret, W. C. Lee, J. Med. Chem. 2010, 53, 3.
[13] C. Prodromou, S. M. Roe, R. OꢁBrien, J. E. Ladbury, P. W. Piper,
L. H. Pearl, Cell 1997, 90, 65.
[14] S. M. Roe, C. Prodromou, R. OꢁBrien, J. E. Ladbury, P. W. Piper,
L. H. Pearl, J. Med. Chem. 1999, 42, 260.
[15] M. M. U. Ali, S. M. Roe, C. K. Vaughan, P. Meyer, B. Panaretou,
P. W. Piper, C. Prodromou, L. H. Pearl, Nature 2006, 440, 1013.
[16] N. Winssinger, S. Barluenga, J. Chem. Soc. Chem. Commun. 2007,
22.
[17] S. Barluenga, P. Y. Dakas, M. Boulifa, E. Moulin, N. Winssinger, C.
R. Chim. 2008, 11, 1306.
[18] T. Hofmann, K. H. Altmann, C. R. Chim. 2008, 11, 1318.
[19] N. Winssinger, J. G. Fontaine, S. Barluenga, Curr. Top. Med. Chem.
2009, 9, 1419.
[20] S. V. Sharma, T. Agatsuma, H. Nakano, Oncogene 1998, 16, 2639.
[21] S. Soga, L. M. Neckers, T. W. Schulte, Y. Shiotsu, K. Akasaka, H.
Narumi, T. Agatsuma, Y. Ikuina, C. Murakata, T. Tamaoki, S. Aki-
naga, Cancer Res. 1999, 59, 2931.
Mixture of double bond isomers 47:53; colorless solid; m.p. 212–2148C;
1H NMR (400 MHz, [D6]DMSO): mixture of isomers d=10.59 (brs, 1H,
OH), 10.33 (brs, 1H, OH), 8.74 (s, 2H, 2ꢂtriazole), 6.53 (s, 1H, ArH),
6.50 (s, 1H, ArH), 5.53 (t, 1H, J = 6.8 Hz, =CH), 5.45 (t, 1H, J =
6.9 Hz, =CH), 5.02 (s, 2H, CH2), 5.00 (s, 2H, CH2), 4.33–4.27 (m, 8H, 4ꢂ
CH2), 4.03 (s, 2H, CH2), 3.84 (s, 2H, CH2), 2.67–2.62 (m, 2H, CH2), 2.38
(t, 2H, J = 6.0 Hz, CH2), 1.95 (t, 2H, J = 6.8 Hz, CH2), 1.87 (m, 2H,
CH2), 1.55 (quin, 2H, J = 6.5 Hz, CH2), 1.45 (brs, 4H, 2ꢂCH2), 1.38
(quin, 2H, J
= 6.5 Hz, CH2), 1.29 ppm (t, 6H, J =7.0 Hz, 2ꢂCH3);
13C NMR (125 MHz, CDCl3): mixture of isomers d=206.1 (C), 206.0 (C),
169.0 (C), 168.3 (C), 160.7 (C), 158.0 (C), 156.7 (C), 156.4 (C), 156.0 (C),
139.3 (C), 139.2 (C), 135.7 (C), 134.6 (C), 133.6 (C), 129.8 (CH), 129.2
(CH), 128.63 (CH), 113.7 (C), 113.6 (C), 113.0 (C), 111.5 (C), 102.9
(CH), 102.8 (CH), 65.9 (CH2), 65.5 (CH2), 61.0 (CH2), 55.6 (CH2), 48.8
(CH2), 45.6 (CH2), 45.3 (CH2), 32.9 (CH2), 27.8 (CH2), 26.8 (CH2), 26.2
(CH2), 25.2 (CH2), 23.2 (CH2), 22.4 (CH2), 21.7 (CH2), 14.6 ppm (CH3);
IR (dichloromethane): nmax =3687, 1722, 1660, 1557 cmÀ1; MS (ESI): m/z
(%): 516/514 (35/100) [M+Na]+, 494/492 (6/17) [M+H]+; m/z: calcd for
C23H2635ClN3NaO7: 514.1357; found: 514.1343 [M+Na]+.
Protein crystallography: The expression, purification and crystallization
of the N-terminal domain of yeast Hsp90 has been previously de-
scribed.[52] Co-crystallizations were conducted by dissolving the inhibitor
in 100% DMSO at 50 mm and adding 5 mL to 1 mL of Hsp90 N-terminal
domain at 4 mgmLÀ1 in 20 mm Tris pH 7.5 and 1 mm EDTA. The com-
plex was then concentrated to 200 mL (20 mgmLÀ1) and crystallized as
previously described.[52] Single crystals appeared overnight of approxi-
mate dimensions 0.3ꢂ0.2ꢂ0.2 mm. These were flash frozen after stepwise
addition of glycerol to 30% and data collected on station ID23.1 and
ID23.2 at the ESRF. The data were integrated using MOSFLM and
scaled and merged using SCALA in CCP4.
The complex was initially solved by isomorphous replacement using a
previously determined N-terminal structure (PDB ID 1AH6) in the usual
space group, P4322. The model was refined in REFMAC5 in CCP4 and
rebuilt using COOT. The FreeR did not refine below 30% and so other
space groups were investigated. Refinement proceeded satisfactorily in
C2, with four molecules in the asymmetric unit. The inhibitor library was
[22] S. Soga, Y. Shiotsu, S. Akinaga, S. V. Sharma, Curr. Cancer Drug
Targets 2003, 3, 359.
[23] T. Agatsuma, H. Ogawa, K. Akasaka, A. Asai, Y. Yamashita, T.
Mizukami, S. Akinaga, Y. Saitoh, Bioorg. Med. Chem. 2002, 10,
3445.
Chem. Eur. J. 2010, 16, 10366 – 10372
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
10371

C. J. Moody et al.
[24] X. D. Geng, Z. Q. Yang, S. J. Danishefsky, Synlett 2004, 1325.
[25] X. G. Lei, S. J. Danishefsky, Adv. Synth. Catal. 2008, 350, 1677.
[26] E. Moulin, V. Zoete, S. Barluenga, M. Karplus, N. Winssinger, J.
Am. Chem. Soc. 2005, 127, 6999.
[27] S. Barluenga, C. H. Wang, J. G. Fontaine, K. Aouadi, K. Beebe, S.
Tsutsumi, L. Neckers, N. Winssinger, Angew. Chem. 2008, 120, 4504;
Angew. Chem. Int. Ed. 2008, 47, 4432.
[28] S. Barluenga, J. G. Fontaine, C. H. Wang, K. Aouadi, R. H. Chen, K.
Beebe, L. Neckers, N. Winssinger, ChemBioChem 2009, 10, 2753.
[29] N. Proisy, S. Y. Sharp, K. Boxall, S. Connelly, S. M. Roe, C. Prodro-
mou, A. M. Z. Slawin, L. H. Pearl, P. Workman, C. J. Moody, Chem.
Biol. 2006, 13, 1203.
[30] J. E. H. Day, S. Y. Sharp, M. G. Rowlands, W. Aherne, P. Workman,
C. J. Moody, Chem. Eur. J. 2010, 16, 2758.
[31] M. Lampilas, R. Lett, Tetrahedron Lett. 1992, 33, 773.
[32] M. Lampilas, R. Lett, Tetrahedron Lett. 1992, 33, 777.
[33] Z. Q. Yang, X. D. Geng, D. Solit, C. A. Pratilas, N. Rosen, S. J. Dani-
shefsky, J. Am. Chem. Soc. 2004, 126, 7881.
[34] R. M. Garbaccio, S. J. Stachel, D. K. Baeschlin, S. J. Danishefsky, J.
Am. Chem. Soc. 2001, 123, 10903.
[35] S. Barluenga, P. Lopez, E. Moulin, N. Winssinger, Angew. Chem.
2004, 116, 3549; Angew. Chem. Int. Ed. 2004, 43, 3467.
[36] S. Barluenga, E. Moulin, P. Lopez, N. Winssinger, Chem. Eur. J.
2005, 11, 4935.
[41] J. M. Williams, R. B. Jobson, N. Yasuda, G. Marchesini, U.-H. Dol-
ling, E. J. J. Grabowski, Tetrahedron Lett. 1995, 36, 5461.
[42] T. A. Carpenter, G. E. Evans, F. J. Leeper, J. Staunton, M. R. Wilkin-
son, J. Chem. Soc. Perkin Trans. 1 1984, 1043.
[43] F. M. Hauser, R. P. Rhee, J. Org. Chem. 1977, 42, 4155.
[44] F. M. Hauser, R. Rhee, J. Am. Chem. Soc. 1977, 99, 4533.
[45] H. R. Tsou, M. Otteng, T. Tran, M. B. Floyd, M. Reich, G. Birnberg,
K. Kutterer, S. Ayral-Kaloustian, M. Ravi, R. Nilakantan, M. Grillo,
J. P. McGinnis, S. K. Rabindran, J. Med. Chem. 2008, 51, 3507.
[46] H. R. Tsou, X. Liu, G. Birnberg, J. Kaplan, M. Otteng, T. Tran, K.
Kutterer, Z. L. Tang, R. Suayan, A. Zask, M. Ravi, A. Bretz, M.
Grillo, J. P. McGinnis, S. K. Rabindran, S. Ayral-Kaloustian, T. S.
Mansour, J. Med. Chem. 2009, 52, 2289.
[47] N. Winssinger, S. Barluenga, M. Karplus, WO2008021213A1, 2008.
[48] M. Meldal, C. W. Tornoe, Chem. Rev. 2008, 108, 2952.
[49] B. W. Dymock, X. Barril, P. A. Brough, J. E. Cansfield, A. Massey,
E. McDonald, R. E. Hubbard, A. Surgenor, S. D. Roughley, P.
Webb, P. Workman, L. Wright, M. J. Drysdale, J. Med. Chem. 2005,
48, 4212.
[50] R. Howes, X. Barril, B. W. Dymock, K. Grant, C. J. Northfield,
A. G. S. Robertson, A. Surgenor, J. Wayne, L. Wright, K. James, T.
Matthews, K. M. Cheung, E. McDonald, P. Workman, M. J. Dry-
sdale, Anal. Biochem. 2006, 350, 202.
[51] V. Zhou, S. L. Han, A. Brinker, H. Klock, J. Caldwell, X. J. Gu,
Anal. Biochem. 2004, 331, 349.
[52] C. Prodromou, P. W. Piper, L. H. Pearl, Proteins 1996, 25, 517.
[53] S. Y. Sharp, L. R. Kelland, M. R. Valenti, L. A. Brunton, S. Hobbs,
P. Workman, Mol. Pharmacol. 2000, 58, 1146.
[37] E. Moulin, S. Barluenga, F. Totzke, N. Winssinger, Chem. Eur. J.
2006, 12, 8819.
[38] S. L. Boulet, L. A. Paquette, Synthesis 2002, 895.
[39] Z. Geng, B. Chen, P. Chiu, Angew. Chem. 2006, 118, 6343; Angew.
Chem. Int. Ed. 2006, 45, 6197.
Received: April 28, 2010
[40] C. M. Thompson, J. A. Frick, J. Org. Chem. 1989, 54, 890.
Published online: July 26, 2010
10372
www.chemeurj.org
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 10366 – 10372