Development of synthetic routes to macrocyclic compounds based on the HSP90 inhibitor radicicol

Article abstract of DOI:10.1016/j.tetlet.2006.01.116

Short routes are reported to novel macrolides (e.g., 9, 12, 20) related to the HSP90 inhibitor radicicol.

Full text of DOI:10.1016/j.tetlet.2006.01.116

Tetrahedron Letters 47 (2006) 2237–2240
Development of synthetic routes to macrocyclic compounds
based on the HSP90 inhibitor radicicol
Butrus Atrash, Tracey S. Cooper, Peter Sheldrake,
Paul Workman and Edward McDonald*
Cancer Research UK Centre for Cancer Therapeutics, The Institute of Cancer Research, 15 Cotswold Road, Sutton,
Surrey SM2 5NG, UK
Received 21 November 2005; revised 4 January 2006; accepted 18 January 2006
Available online 10 February 2006
Abstract—Short routes are reported to novel macrolides (e.g., 9, 12, 20) related to the HSP90 inhibitor radicicol.
Ó 2006 Elsevier Ltd. All rights reserved.
Radicicol 1,1 a fungal natural product from Monospo-
rium bonorden, is an inhibitor of the ATPase activity
of the heat shock protein HSP90.2 In cells HSP90 acts
as a chaperone ensuring the correct folding of many cli-
ent proteins including several that are implicated in can-
cer.3 Inhibitors of HSP90 bring about depression of
cellular levels of several oncogenic proteins and 17-
AAG, a derivative of geldanamycin, is currently in
Phase II trials for treatment of cancer.4
cells.6 Here we describe an alternative approach to hit
generation based on the natural product radicicol.7
Radicicol 1 inhibits yeast HSP902 with KD = 19 nM: it is
active in cells but not in vivo. Radicicol 1 has structural
features, which are undesirable in a clinical drug candi-
date. The epoxide and dienone functionalities have the
potential to react with nucleophiles in vivo resulting in
rapid reduction of drug levels in tissues and potential
toxicity due to modification of essential biopolymers.
The phenols are likely to be metabolised by conjugation
also leading to high clearance in vivo. We therefore
decided to design short synthetic routes that would give
access to collections of compounds resembling radicicol.
The compounds should be devoid of electrophilic func-
tionality, and the designs should allow preparation of
analogues to enhance biological profile and confer
drug-like properties including aqueous solubility.
H
O
O
O
O
Et
O
H
HO
HO
O
Cl
Me
N
OH
OH
NH
1
2
We therefore initiated research aimed at the identifica-
tion of small molecule inhibitors of HSP90 that might
be developed as totally synthetic drugs. One of these
studies resulted in the discovery, by high throughput
screening, of a series of resorcinylic pyrazoles,5 for
example, 2 and has led to compounds with high potency
both in vitro and in inhibiting the growth of cancer
Our initial approach, outlined in Scheme 1, would afford
a collection of acetylenic lactone (X = O) and lactam
(X = NR2) macrocycles 6 with differing ring sizes. Com-
binatorial variation of the three starting materials would
allow replacement of the resorcinol ring by a wide range
of substituted phenyl or heteroaryl rings, and the versa-
tile alkyne could be converted to various functional
groups including (by hydration) the benzyl ketone of
radicicol. The design includes a second ring heteroatom
to afford an additional site for the preparation of ana-
logues (NR1) and for enhancing water solubility. Final-
ly, fragments 4 and 5 could carry substituents, for
example, alkyl, aryl on any of their carbon chain atoms.
*
Corresponding author. Tel.: +44 208 722 4294; fax: +44 208 722
4047; e-mail: ted.mcdonald@icr.ac.uk
0040-4039/$ - see front matter Ó 2006 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tetlet.2006.01.116

2238
B. Atrash et al. / Tetrahedron Letters 47 (2006) 2237–2240
O
HX
NHR1
O
m
OMe
I
4
X
NR1
n
m
R3
R3
OH
3
n
6
5
X = O, NR2
m, n = 0, 1, 2
Scheme 1.
Sonogashira8 coupling of methyl 2-iodo-benzoate (3,
R3 = H) and pentyn-5-ol (5, n = 1) in degassed piper-
idine9 mediated by tetrakis(triphenylphosphine)palla-
dium(0) and copper iodide gave the desired product 7
in 86% yield. The alcohol function of 7 was converted
to either the triflate or the more stable 4-nitrobenzene-
sulfonate10 and these activated substrates reacted with
a variety of amino-alcohols followed by acetylation of
the resultant secondary amine to give the hydroxy-esters
8 in yields of 60–86% (Scheme 2).
Table 1. Cyclisation yields for conversion hydroxy-esters 8 to lactones
9
m
R
Ring size
Yield % of 9
a
b
c
d
e
f
0
1
2
3
3
4
Me
H
H
H
Me
H
12
13
14
15
15
16
No product
63
28
32
30
32
Modest yields in the foregoing cyclisation steps (0–63%
for lactones, zero for lactams) were attributed to ring
strain imposed by the linear character of the acetylene
bond. The formation of macrolides should be more fac-
ile if the acetylene were replaced with a cis double bond
but conversion of acetylenic precursors to intermediates
with cis double bonds proved difficult.16 Therefore,
attention was switched to a new design (outlined in
Scheme 4) incorporating an ortho-substituted aryl group
in place of the acetylene. These target compounds 13
have an aryl fused macrocyclic structure that is not pres-
ent in radicicol. However, the biaryl motif is found in
the resorcinylic pyrazole series5,6 of HSP90 inhibitors,
for example 2, and so these hybrid structures were of
interest for increasing the diversity of our screening set.
Liberation of the hydroxy-acid for cyclisation to the
macrolide, through hydrolysis of esters 8, proved to be
problematic due to the formation of isocoumarin 11.13
However, careful monitoring of the hydrolysis reaction
performed with LiOH in methanol at room temperature
gave the desired hydroxy-acid and the isocoumarin as a
2:1 mixture, readily separable by flash chromatography.
R
O
O
( )m
N
OH
Ac
11
Boronic ester 14 was synthesised by ortho lithiation of
ethyl benzoate with lithium 2,2,6,6-tetramethylpiperi-
dide (LiTMP) followed by in situ reaction with triiso-
propyl borate; boronic ester exchange with 2,2-
dimethylpropane-1,3-diol17 then afforded 14 in 84%
yield. Palladium mediated cross-coupling of 14 with iod-
ides 15 and 16 afforded the biaryl compounds 17 and 18
in 74% and 54% yield, respectively (Scheme 5).
For the crucial step of macrocyclisation, success was
achieved only with Yamaguchi’s reagent.11,12 Five pro-
totype lactones were produced with 13–16-membered
rings (Table 1).
Cyclisation of a parallel series of amino acids 10 was
attempted using a number of reagents.14 However, only
one 13-membered macrolactam 12 was synthesised by
closing the ring at a different position according to
Scheme 3.
Aldehyde 17 was reductively aminated with 6-amino-
hexan-1-ol and sodium triacetoxyborohydride, while
R
O
R
O
O
HX
m
a
b
O
NAc
OMe
OMe
NAc
m
OH
9a-f
7
8a-f X = O
R = H, Me
m = 0 - 4
10a-f X = NH
Scheme 2. Reagents and conditions: (a) (i) Tf2O, Et3N, CH2Cl2, 0 °C, 30 min or 4-NO2C6H4SO2Cl, Et3N, CH2Cl2, rt, 12 h; (ii)
H2NCH2(CH2)mCH(R)OH, Et3N, CH2Cl2, 12 h; (iii) Ac2O, Et3N, CH2Cl2, 12 h; (b) (i) LiOH (2 equiv), MeOH, rt, 4 h; (ii) Yamaguchi’s reagent.11
C6H6, DMAP, Et3N, reflux, 16 h.

B. Atrash et al. / Tetrahedron Letters 47 (2006) 2237–2240
2239
O
CO2Et
OH
O
O
NH
O
NH
CO2H
I
a,b
c,d
12
Scheme 3. Reagents and conditions: (a) H2NCH2CH2CO2EtÆHCl, EDC,15 HOBt (90%); (b) 4-pentyn-1-ol, CuI, Pd(PPh3)4, piperidine (95%); (c)
LiOH, MeOH (78%); (d) Yamaguchi’s reagent11 C6H6, DMAP, reflux, 24 h (22%).
O
HO
H2N
m
O
m
CO2Et
B(OR)2
N
n
I
X
R
OH
13
X = CH=CH, or heteroatom
n
X
Scheme 4. Proposed synthetic route to biaryl macrolides.
O
O
EtO2C
I
Ac
b,c
a
N
d,e,f
CHO
CHO
17
EtO2C
O
B
15
19
O
14
EtO2C
S
O
O
I
g,h,c
d,e,f
N
Ac
a
OH
S
OH
S
16
18
20
Scheme 5. Reagents and conditions: (a) Pd(PPh3)4, K3PO4, dioxane, 100 °C; (b) H2N(CH2)6OH; (c) NaBH(OAc)3; (d) Ac2O, py, CH2Cl2; (e) LiOH,
MeOH, rt; (f) Yamaguchi’s reagent, C6H6, DMAP, Et3N, reflux; (g) Dess–Martin; (h) H2N(CH2)5OH.
alcohol 18 was oxidised (Dess–Martin periodinane) and
the aldehyde product was reacted with 5-aminopentan-
1-ol/sodium triacetoxyborohydride. The resultant sec-
ondary amines were acetylated and the required hydr-
oxy acid precursors were generated by base catalysed
hydrolysis of the ethyl esters. Macrolactonisation using
the Yamaguchi reagent yielded the corresponding 14-
membered lactones 19 and 20 in yields of 60% and
32%, respectively.
oxazole 21 in moderate yield. This was coupled to
BocNH(CH2)6NH2 via the acid chloride followed by
hydrolysis of the methyl ester with bis(tributyltin)oxide
(BBTO) in refluxing toluene.18,19 Removal of the Boc
group with HCl/dioxane gave the required amino acid,
which was then cyclised with pentafluorophenyl diph-
enylphosphinate (FDPP)20 to give 14-membered macro-
cycle 22 in 31% yield.
In summary, we have developed short modular synthe-
ses to three classes of macrocycles that are topologically
similar to radicicol but do not have the undesirable
functionality found in the natural product.21
One further macrocycle was synthesised as shown in
Scheme 6. Reaction of phthalic anhydride with
methyl isocyanoacetate in the presence of DBU gave
H
C
N
CO2Me
N
a,c
CO2H
O
a
O
O
O
H
O
O
CO2Me
N
N
N
O
21
22
Scheme 6. Reagents and conditions: (a) DBU, THF, rt, 24 h (56%); (b) (i) (COCl)2, Et3N, (ii) BocNH(CH2)6NH2 (76%); (c) (i) BBTO, PhMe, reflux,
24 h, (ii) HCl, dioxane, (iii) FDPP, DMF, rt, 38 h (31%).

2240
B. Atrash et al. / Tetrahedron Letters 47 (2006) 2237–2240
Zoete, V.; Barluenga, S.; Karplus, M.; Winssinger, N. J.
Acknowledgements
Am. Chem. Soc. 2005, 127, 6999–7004.
8. Sonogashira, K.; Tohda, Y.; Hagihara, N. Tetrahedron
Lett. 1975, 4467–4470.
9. Greatly reduced yields were observed when the solvent
was not degassed.
10. The corresponding mesylate was found to be unstable.
11. Inaga, J.; Hirata, K.; Sacki, H.; Katsuki, T.; Yamaguchi,
M. Bull. Chem. Soc. Jpn. 1979, 52, 1989.
This work was supported by Cancer Research UK
[CUK] programme grant number C309/A2187. Paul
Workman is a Cancer Research UK Life Fellow. We
thank Dr. Amin Mirza and Dr. Bernard Nutley for their
assistance with NMR and mass spectrometry.
12. No cyclisation was observed with dipyridyldisulfide/tri-
phosphine in acetonitrile or xylene at reflux. Corey, E. J.;
Nicolaou, K. C. J. Am. Chem. Soc. 1974, 96, 5614–5616.
13. Similar observations have been reported recently by other
workers: Bellina, F.; Ciucci, D.; Vergamini, P.; Rossi, R.
Tetrahedron 2000, 56, 2533–2545.
References and notes
1. Delmotte, P.; Delmotte-Plaquee, J. Nature 1953, 171, 344–
345.
2. Roe, S. M.; Prodromou, C.; O’Brien, R.; Ladbury, J. E.;
Piper, P. W.; Pearl, L. H. J. Med. Chem. 1999, 42, 260–
266.
3. Workman, P. Cancer Lett. 2004, 26, 149–157.
4. Banerji, U.; O’Donnell, A.; Scurr, M.; Pacey, S.; Staple-
ton, S.; Asad, Y.; Simmons, L.; Maloney, A.; Raynaud,
F.; Campbell, M.; Walton, M.; Lakhani, S.; Kaye, S.;
Workman, P.; Judson, I. J. Clin. Oncol. 2005, 23, 4152–
4161.
5. Cheung, K.-M.; Matthews, T.; James, K.; Rowlands, M.;
Boxall, K.; Sharp, S.; Maloney, A.; Roe, S. M.; Prodro-
mou, C.; Pearl, L.; Aherne, G. W.; McDonald, E.;
Workman, P. Bioorg. Med. Chem. Lett. 2005, 15, 3338–
3343.
6. Dymock, B.; Barril, X.; Brough, P.; Cansfield, J.; Massey,
A.; McDonald, E.; Hubbard, R.; Surgenor, A.; Roughley,
S.; Webb, P.; Workman, P.; Wright, L.; Drysdale, M. J.
Med. Chem. 2005, 48, 4212–4215.
14. Dipyridyldisulfide, diethylcyanophosphonate and O-(7-
azabenzotriazol-l-yl)-N,N,N1,N1-tetramethyluronium hexa-
fluorophosphate (HATU) gave no evidence of cyclisa-
tion.
15. 1-Ethyl-3-(3-dimethylaminopropyl)-carbodiimide.
16. 1-3-Dipolar cycloaddition reactions, Lindlar hydrogena-
tion, hydroboration and hydration of the alkyne bond
were attempted without success.
17. Kritensen, J.; Lysen, M.; Vedso, P.; Begtrup, M. Org.
Lett. 2001, 3, 1435–1437.
18. Moriya, O.; Takenaka, H.; Lyoda, M.; Urata, Y.; Endo,
T. J. Chem. Soc., Perkin Trans. 1 1994, 413–418.
19. Attempted hydrolysis of the methyl ester using LiOH,
HC1 and acetic acid, or KOSiMe3 resulted in unidentifi-
able product mixtures.
20. Deng, J.; Hamada, Y.; Shiori, T. J. Am. Chem. Soc. 1995,
117, 7824–7825.
21. These prototype compounds did not inhibit HSP90, but
related compounds incorporating phenolic functionality
have been shown to inhibit HSP90 ATPase. See following
paper: Tetrahedron Lett. 2006, 47, doi:10.1016/j.tetlet.
2006.01.099.
7. Several synthetic analogues of radicicol have been
reported including cycloproparadicicol: Geng, X.; Yang,
Z. Q.; Danishefsky, S. J. Synlett 2004, 1325–1333, and
references cited therein; and pochonin D Moulin, E.;