Novel echinocandin antifungals. Part 2: Optimization of the side chain of the natural product FR901379. Discovery of micafungin

Article abstract of DOI:10.1016/j.bmcl.2008.03.093

Further optimization of the potent antifungal activity of side chain analogs of the natural product FR901379 led to the discovery of compound 8 with an excellent, well-balanced profile. Potent compounds with reduced hemolytic potential were designed based upon a disruption of the linearity of the terphenyl lipophilic side chain. The optimized compound (8, FK463, micafungin) displayed the best balance and was selected as the clinical candidate.

Full text of DOI:10.1016/j.bmcl.2008.03.093

Available online at www.sciencedirect.com
Bioorganic & Medicinal Chemistry Letters 18 (2008) 2886–2890
Novel echinocandin antifungals. Part 2: Optimization
of the side chain of the natural product FR901379.
Discovery of micafungin
Masaki Tomishima,a,* Hidenori Ohki,a Akira Yamada,a
Katsuyuki Makib and Fumiaki Ikedab
aMedicinal Chemistry Research Laboratories, Astellas Pharma Inc., 2-1-6 Kashima, Yodogawa-ku, Osaka 532-8514, Japan
bMedicinal Biology Research Laboratories, Astellas Pharma Inc., 2-1-6 Kashima, Yodogawa-ku, Osaka 532-8514, Japan
Received 5 March 2008; revised 25 March 2008; accepted 29 March 2008
Available online 8 April 2008
Abstract—Further optimization of the potent antifungal activity of side chain analogs of the natural product FR901379 led to the
discovery of compound 8 with an excellent, well-balanced profile. Potent compounds with reduced hemolytic potential were
designed based upon a disruption of the linearity of the terphenyl lipophilic side chain. The optimized compound (8, FK463, mica-
fungin) displayed the best balance and was selected as the clinical candidate.
Ó 2008 Elsevier Ltd. All rights reserved.
In earlier publications from these laboratories, we have
described our research on the chemical modification of
natural products with potent antifungal activity.1–3 Seri-
ous fungal infections in immunocompromised individu-
als are contributing to an increase in the incidence of
deep-seated, disseminated mycoses, and new agents are
needed.4,5 Our efforts have been directed at the discov-
ery of novel antifungal agents that are safer and more
efficacious than known drugs.
less-linear, but rigid structures into the side chain
moiety.
In the work described herein, we have further optimized
the antifungal activity and reduced the hemolytic activ-
ity of the derivatives by the introduction of a series of
five- and six-membered heterocyclic moieties in place
of the central phenyl ring of the terphenyl analog 3,
which was the most potent analog reported in the previ-
ous paper.3 The basic design concept employed is out-
lined in Figure 1. We speculated that the reason for
the reduced hemolysis of the compounds reported in
the previous paper3 was due to the lipophilic side chain
having a less-linear, but a rigid structure, leading to re-
duced interaction with the membrane structure of red
blood cells. It is known that the membrane of mamma-
lian cells and fungal cells contains different amounts of
unsaturated phospholipids,8 thus suggesting that less-
linear side chains may only interact weakly with the
mammalian cell membrane, leading to reduced potential
for hemolysis of red blood cells. Accordingly, we de-
signed a series of analogs containing saturated central
rings and heteroatoms, as well as five-membered ring
heteroaromatic derivatives.
In previous reports,2,3 we have described the chemical
modification of the side chain of the echinocandin-type
1,3-b-glucan synthase inhibitor FR901379, the only nat-
ural echinocandin with high intrinsic water solubility.6,7
These efforts led to the discovery of a number of analogs
with potent antifungal activity, and in particular, we
established a relationship between the calculated logP
(ClogP) of the side chain moiety (calculated as the
methylamine amide) and antifungal activity, as well as
data which suggested that the key to reducing the strong
hemolytic activity associated with the natural product
and a number of the earlier derivatives was to introduce
Keywords: Antifungal; Natural product; FR901379; FR13153; FK463;
Micafungin; Echinocandin; Candida albicans; Aspergillus fumigatus;
Isoxazole.
The compounds prepared in this work were synthesized
as shown in Scheme 1, and as indicated in the earlier
publications. Deacylated hexapeptide nucleus (1) was
*
Corresponding author. Tel.: +81
masaki.tomishima@jp.astellas.com
6
6390 1285; e-mail:
0960-894X/$ - see front matter Ó 2008 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2008.03.093

M. Tomishima et al. / Bioorg. Med. Chem. Lett. 18 (2008) 2886–2890
2887
HO
OH
O
O
O
HO
Introduction of heterocycles
for disruption of linearity
NH
NH
H3C
H2N
N
O
OH
R
O
HN
O
HO
NH
O
CH3
OH
R
O
HO
N
H
N
Potent antifungal activity
& weak hemolytic activity
Potent antifungal activity
& potent hemolytic activity
O
OH
NaO3SO
3
HO
Figure 1. Side chain design for potent antifungal activity and weak hemolytic activity.
O
HO
R
HOBt
WSC HCl
CH2Cl2
80 - 100%
HO
OH
NH2
HO
OH
O
O
O
HO
HO
NH
NH
NH
R
N
N
H3C
H2N
H3C
H2N
O
R
N
N
O
O
N
O
O
HN
OH
O
HN
OH
O
O
HO
HO
NH
O
CH3
OH
NH
O
CH3
OH
DMAP - DMF
O
N
O
N
H
N
H
N
HO
HO
O
O
OH
OH
HO3SO
NaO3SO
1
4 - 10
HO
HO
O(CH2)7CH3
R=
2
N
7
8
9
R=
R=
R=
77 %
75 %
64 %
77 %
O(CH2)4CH3
O(CH2)4CH3
O(CH2)4CH3
O(CH2)5CH3
(FR131535)
S
3
O(CH2)3CH3
R=
N
O
N
O
N
4-6
N
Y
O(CH2)nCH3
R=
X
4
5
6
X = CH Y = N n = 5
X = N Y = N n = 6
X = CH Y = CH n = 5
61 %
66 %
76 %
S
10 R=
N
Scheme 1. Synthesis of FR131535 and novel tri-cyclic analogs.
prepared by enzymatic deacylation of FR901379.9 Re-
acylation of compound 1 with the respective HOBT acti-
vated esters, prepared conveniently from the corre-
sponding carboxylic acids, led to the crude final
products as DMAP salts after trituration of the reaction
mixture with ethyl acetate. Conversion to the sodium
salts and ODS column chromatography eluting with
acetonitrile-water mixtures and lyophilization afforded
the final products as amorphous powders.
hydroxyphenyl)piperazine under standard conditions.
Acid 25 was prepared by the reaction of the Grignard re-
agent prepared from 4-bromo-n-hexyloxybenzene with
tert-butyl 4-oxo-1-piperidinecarboxylate, elimination of
water and deprotection of the t-boc group with TFA,
followed by coupling of the resulting piperidine deriva-
tive with ethyl 4-fluorobenzoate and alkaline hydrolysis.
Biological evaluation of the derivatives containing the
novel six-membered heterocyclic ring was performed,
and the results are summarized in Table 1. As shown
in Table 1, the replacement of the central phenyl ring
in the terphenyl side chain of compound 3 by saturated
six-membered heterocyclic ring led to compounds with
significantly reduced hemolysis, as indicated by the per-
centage of hemolysis at a drug concentration of 1 mg/ml.
The introduction of nitrogen atoms to the central ring,
as well as the disruption of linearity by including satu-
rated, rather than aromatic rings, meant that the
The synthesis of the novel six-membered heterocyclic
ring-containing side-chain carboxylic acids is as outlined
in Scheme 2. Acids 17 and 19 were obtained by hydroly-
sis of the corresponding ethyl or methyl ester with so-
dium hydroxide in THF–EtOH or THF–MeOH. The
esters were obtained by coupling of the respective piper-
azine derivatives with ethyl 4-fluorobenzoate or methyl
4-chloronicotinate. The piperazines were prepared by
alkylation and amide hydrolysis of 1-acetyl-4-(4-

2888
M. Tomishima et al. / Bioorg. Med. Chem. Lett. 18 (2008) 2886–2890
O
O
a
b
N
N
OH
N
N
O(CH2)nCH3
HN
N
O(CH2)nCH3
12 n = 5: 62 %
13 n = 6: 59 %
14 n = 5: 85 %
15 n = 6: 87 %
11
R
O
O
R
O
O
c
N
N
O(CH2)6CH3
N
N
O(CH2)5CH3
N
16 R = Et: 62 %
17 R = H: 89 %
18 R = Me: 72 %
d
d
19 R = H: 91 %
O
O
OH
e
f
Br
O(CH2)5CH3
N
O(CH2)5CH3
HN
O(CH2)5CH3
20
21 100 %
22 76 %
R
O
O
R
O
O
c
g
N
O(CH2)5CH3
N
O(CH2)5CH3
24 R = Et: 89 %
25 R = H: 96%
23 20 %
d
Scheme 2. Reagents and condition: (a) alkyl bromide, K2CO3, DMF; (b) concd HCl; (c) ethyl 4-fluorobenzoate or methyl nicotinate, K2CO3,
DMSO; (d) 1 N NaOH, corresponding alcohol, THF; (e) Mg, ether, then 1-t-boc-4-piperidone; (f) TFA, CH2Cl2; (g) H2, Pd/C, THF.
Table 1. Antifungal and hemolytic activities of compounds containing six-membered heterocyclic ring
Compound
R
ClogP
MICa (lg/ml)
Hemolysis (%, 1 mg/ml)
C. albicans
FP633
A. fumigatus
FP1305
RPMIb
Serumc
RPMIb,d
Serumc
3
4
6.14
5.17
0.025
0.1
1.56
0.002
0.002
1.56
1.56
79
O(CH2)3CH3
O(CH2)5CH3
3.13
<20
N
N
N
N
N
O(CH2)6CH3
O(CH2)5CH3
5
6
5.30
5.87
0.2
3.13
1.56
0.005
0.003
6.25
1.56
<20
33
N
0.05
a Minimum inhibitory concentration.
b In RPMI media.
c In mouse serum.
d IC50 values, 50% inhibitory concentration of hyphal growth.
ClogP10 was lower, hence the ClogP was adjusted be-
tween 5 and 6 by extending the length of the alkyl chain
at the terminus of the side chain. Compounds 4–6 dis-
played comparable in vitro antifungal activity against
Candida albicans and Aspergillus fumigatus to the ter-
phenyl analog 3. In particular, compound 4 displayed
equivalent, potent activity to 3, but very low hemolytic
potential, and was selected for further evaluation.
Encouraged by the results outlined in Table 1, we next
turned our attention to five-membered heterocyclic
rings, as their intrinsic nature means that the structures
are somewhat bent away from linearity.
Acid 29 was prepared from the a-bromoketone derived
from ethyl 4-acetylbenzoate by reaction with pyridinium
tribromide in HBr–acetic acid, followed by thiazole ring
formation with the corresponding thioamide, and basic
hydrolysis. The isoxazole acid 33 was prepared from
4-n-pentyloxyacetophenone by coupling with methyl
4-formylbenzoate using titanium tetrachloride and
triethylamine, followed by the reaction of the resulting
enone with hydroxylamine and oxidation with MnO2
and hydrolysis of the ester group. Thiadiazole acid 37
was prepared by hydrolysis of the ester derived by cycli-
zation of a diacyl hydrazide in the presence of P2S5 in
THF. The side chain acid for compound 8 was prepared
as outlined in our preliminary report, involving a 1,3-
dipolar cycloaddition reaction between an acetylene
The novel, five-membered ring-containing side-chain
carboxylic acids were prepared as shown in Scheme 3.

M. Tomishima et al. / Bioorg. Med. Chem. Lett. 18 (2008) 2886–2890
2889
S
O
O(CH2)4CH3
a
b
O
O
O
O
Br
N
R
O
O
O
28 R = Et 60%
29 R = H 99 %
26
27 95 %
c
O(CH2)4CH3
N
O
d
e
O
O
O
O
O(CH2)4CH3
c
O(CH2)4CH3
O
R
O
30
34
31 20 %
32 R = Et 24%
33 R = H 90%
N
N
O
f
g
H2N NH
O
O
O
HN NH
O(CH2)4CH3
O(CH2)4CH3
O(CH2)4CH3
S
R
O
O
O
36 R = Et 93%
37 R = H 90%
c
35 100 %
Scheme 3. Reagents: (a) pyridium tribromide, HBr–AcOH; (b) 4-pentyloxybenzthioamide, THF; (c) NaOH aq, corresponding alcohol, THF; (d)
methyl 4-formylbenzoate, TiCl4, Et3N, CH2Cl2; (e) hydroxyamine hydrochloride, EtOH then MnO2, dichloroethane; (f) 4-methoxycarbonylbenzoyl
chloride, pyridine, THF; (g) P2S5, THF.
Table 2. Antifungal and hemolytic activities of compounds containing aromatic five-membered ring
Compound
R
ClogP
MICa (lg/ml)
C. albicans FP633 A. fumigatus
FP1305
RPMIb,d
Hemolysis (%, 1 mg/ml)
RPMIb
0.025
Serumc
1.56
Serumc
1.56
3
7
6.14
5.68
5.32
5.32
5.74
0.002
NT
79
O(CH2)3CH3
O(CH2)4CH3
O(CH2)4CH3
O(CH2)4CH3
O(CH2)5CH3
N
0.1
6.25
1.56
1.56
0.78
3.13
1.56
0.78
0.78
<20
<20
38
S
8
0.05
0.05
0.05
0.002
0.001
0.002
N
O
N
O
9
N
S
10
42
N
a Minimum inhibitory concentration.
b In RPMI media.
c In mouse serum.
d IC50 values, 50% inhibitory concentration of hyphal growth.
Table 3. MIC in mouse serum, in vivo antifungal effect and PK profiles of selected compounds
Compound
C. albicans FP633
A. fumigatus FP1305
PK (mice, 5 mg/kg, iv)
Serum MICa (lg/ml) ED50 (mg/kg) Serum MICa (lg/ml) ED50 (mg/kg) C30min (lg/ml) t1/2 (h) AUC0ꢀ1 (lg ml/ml)
b
b
2
3
4
8
25
1.56
3.71
12.5
1.56
4.31
NTc
30.3
NTc
18.8
10.1
2.0
109
NTc
105
78
0.447
0.395
0.325
NTc
3.5
3.13
1.56
1.56
1.56
0.531
0.228
5.1
a Minimum inhibitory concentration in mouse serum.
b Based on survival at 2 weeks after infection.
c Not tested.

2890
M. Tomishima et al. / Bioorg. Med. Chem. Lett. 18 (2008) 2886–2890
and a nitrile oxide.11 Biological evaluation of the deriv-
atives containing the novel five-membered heterocyclic
ring was performed, and results are summarized in Ta-
ble 2. As indicated in Table 2, thiazole 7, isoxazole 8,
isoxazole isomer 9, and thiadiazole 10 all displayed re-
duced hemolytic potential compared to the terphenyl
compound 3. In particular, compound 8 displayed po-
tent in vitro antifungal activity and low hemolytic po-
tential and was selected for further evaluation.
Acknowledgment
We thank Dr. David Barrett (Medicinal Chemistry Re-
search Laboratories) for kind help and advice.
Supplementary data
Representative synthetic procedure and spectrum data
for micafungin (8) and determination method for MICs
in mouse serum are available. Supplementary data asso-
ciated with this article can be found, in the online ver-
sion, at doi:10.1016/j.bmcl.2008.03.093.
Table 3 indicates the in vivo antifungal effects of the se-
lected compounds (4, 8) against the principal fungal spe-
cies causing life-threatening infections in humans, C.
albicans and A. fumigatus, along with pharmacokinetic
(PK) data in mice. It can be clearly seen from the data
in this table that a more potent serum MIC (minimum
inhibitory concentration) gives superior in vivo efficacy
for C. albicans infections. As indicated in the earlier pa-
per2 FR131535 displays comparable in vivo efficacy to
fluconazole against C. albicans. Compounds 4 and 8
showed about 10 times superior in vivo efficacy com-
pared to FR131535. Against A. fumigatus systemic
infection, compound 8, in particular showed excellent
efficacy, displaying an ED50 value of 0.228 mg/kg. The
data in Table 3 suggest that the potent efficacy in the
Candida model results from the combination of good
serum MIC and a high AUC in the PK study, whereas
for good efficacy in the Aspergillus model, it is necessary
to have a good balance of MIC, elimination half-life.
Compound 8 was selected for further evaluation as a
clinical candidate.
References and notes
1. Barrett, D. Biochim. Biophys. Acta 2002, 1587, 224.
2. Fujie, A.; Iwamoto, T.; Sato, B.; Muramatsu, H.; Kasa-
hara, C.; Furuta, T.; Hori, Y.; Hino, M.; Hashimoto, S.
Bioorg. Med. Chem. Lett. 2001, 11, 399.
3. Tomishima, M.; Ohki, H.; Yamada, A.; Maki, K.; Ikeda,
F. Bioorg. Med. Chem. Lett. 2008, 18, 1474.
4. Dismukes, W. E. Clin. Infect. Dis. 2006, 42, 1289.
5. (a) Patterson, T. F. Lancet 2005, 366, 1013; (b) De Pauw,
B. E. Surg. Infect. 2006, 7, S93.
6. Iwamoto, T.; Fujie, A.; Sakamoto, K.; Tsurumi, Y.;
Shigematsu, N.; Yamashita, M.; Hashimoto, S.; Okuhara,
M.; Kohsaka, M. J. Antibiot. 1994, 47, 1084.
7. Iwamoto, T.; Fujie, A.; Nitta, K.; Hashimoto, S.; Oku-
hara, M.; Kohsaka, M. J. Antibiot. 1994, 47, 1092.
8. (a) Avery, S. V.; Smith, S. L.; Ghazi, A. M.; Hptroff, M. J.
Appl. Environ. Microbiol. 1999, 65, 1191; (b) Wessels, J.
M.; Veerkamp, J. H. Biochim. Biophys. Acta 1973, 291,
190.
9. (a) Boeck, L. D.; Fukuda, D. S.; Abbott, B. J.; Debono,
M. J. Antibiot. 1989, 42, 382; (b) Debono, M.; Abbott, B.
J.; Fukuda, D. S.; Barnhart, M.; Willard, K. E.; Molloy,
R. M.; Michel, K. H.; Turner, J. R.; Butler, T. F.; Hunt,
A. H. J. Antibiot. 1989, 42, 389.
10. ClogP values were calculated using CLOGP program
(Biobyte Corp., Version 4.3).
11. Tomishima, M.; Ohki, H.; Yamada, A.; Takasugi, H.;
Maki, K.; Tawara, S.; Tanaka, H. J. Antibiot. 1999, 52,
674.
In summary, further optimization of the acyl side chain
of a series of analogs of the water-soluble 1,3-b-glucan
synthase inhibitor FR901379 led to the discovery of an
analog (8) with potent in vivo activity against both C.
albicans and A. fumigatus. This analog (FK463, mica-
fungin) was selected as a clinical candidate and detailed
in vivo and clinical trial results have been disclosed. The
key to success was the introduction of various heterocy-
cles in place of the central phenyl ring of the terphenyl
analog 3 as such modifications resulted in potent anti-
fungal activity and significantly reduced hemolysis.