Design and synthesis of novel benzofurans as a new class of antifungal agents targeting fungal N-myristoyltransferase. Part 1

Article abstract of DOI:10.1016/S0960-894X(01)00319-5

Potent and selective Candida albicans N-myristoyltransferase (CaNmt) inhibitors have been identified through optimization of a lead compound 1 discovered by random screening. The inhibitor design is based on the crystal structure of the CaNmt complex with compound (S)-3 and structure-activity relationships (SARs) have been clarified. Modification of the C-4 side chain of 1 has led to the discovery of a potent and selective CaNmt inhibitor 11 (RO-09-4609), which exhibits antifungal activity against C. albicans in vitro.

Full text of DOI:10.1016/S0960-894X(01)00319-5

Bioorganic & Medicinal Chemistry Letters 11 (2001) 1833–1837
Design and Synthesis of Novel Benzofurans as a New Class of
Antifungal Agents Targeting Fungal N-Myristoyltransferase. Part 1
Miyako Masubuchi, Ken-ichi Kawasaki, Hirosato Ebiike, Yoshihiko Ikeda, Shinji Tsujii,
Satoshi Sogabe, Toshihiko Fujii, Kiyoaki Sakata, Yasuhiko Shiratori, Yuko Aoki,
Tatsuo Ohtsuka*and Nobuo Shimma
Nippon Roche Research Center, 200 Kajiwara, Kamakura, Kanagawa 247-8530, Japan
Received 16 February 2001; accepted 28 April 2001
Abstract—Potent and selective Candida albicans N-myristoyltransferase (CaNmt) inhibitors have been identified through optimi-
zation of a lead compound 1 discovered by random screening. The inhibitor design is based on the crystal structure of the CaNmt
complex with compound (S)-3 and structure–activity relationships (SARs) have been clarified. Modification of the C-4 side chain of
1 has led to the discovery of a potent and selective CaNmt inhibitor 11 (RO-09-4609), which exhibits antifungal activity against C.
albicans in vitro. # 2001 Elsevier Science Ltd. All rights reserved.
N-Myristoyltransferase (Nmt) is an enzyme that trans-
fers the myristoyl group of myristoyl CoA to the N-
terminal glycine of eukaryotic cellular proteins. N-Myr-
istoylation of several G-proteins, Gpa1, Arf1, Arf2 and
Vps15, which are essential for fungal growth, has been
reported to be indispensable for their function in Sac-
charomyces cerevisiae.1À4 Nmt has also been proven to
be essential for the viability of fungi, including medi-
cally important pathogenic fungi such as Candida albi-
cans (C. albicans) and Cryptococcus neoformans.5,6
Therefore, Nmt is a good target for the development of
novel fungicidal drugs with a new mode of action.
a highly potent and safe antifungal agent by improving
the enzyme inhibitory activity while eliminating the
b-blocking action (IC50: 0.098 mM).
Synthesis
The general synthesis of benzofuran derivatives is out-
lined in Scheme 1.11 Acylresorcinol derivatives I were
prepared by (1) ortho lithiation of 1,3-bis(methoxy-
methyl)resorcinol followed by acylation with acyl chlo-
ride, R1COCl, or DMF and (2) removal of the
methoxymethyl groups by acid hydrolysis. 4-Hydro-
xybenzofuran derivatives II were prepared by the same
method reported by Atkinson et al., which includes
mono-alkylation of intermediate I with ethyl bromo-
acetate and base catalyzed cyclization of the alkylated
product.12 Alkylation of intermediates II with excess
alkylene dibromide gave bromoalkyl derivatives III in
good yields. Amination of III with various amines at
70 ꢀC gave the desired amino derivatives IV without
affecting the ester group.
Although several peptidomimetic inhibitors of C. albi-
cans Nmt (CaNmt) have been reported, their antifungal
activity is only marginal.7À9 Therefore, we set out to
seek novel non-peptidic Nmt inhibitors having strong
antifungal activity. We discovered a lead compound 1
that competitively inhibits CaNmt (IC50: 0.98 mM) with
high selectivity over human Nmt (IC50: 194 mM), by a
random screening of the Roche chemical libraries.
However, 1 did not show potent antifungal activity in
vitro (IC50 against C. albicans CY1002: 390 mM).
Compound 1 was originally reported by Grinev et al. to
have antiarrhythmic and b-blocking action.10 Thus we
initiated the chemical modification study of 1 to identify
Results and Discussion
Since it is well known that an aromatic compound hav-
ing a b-aminoalcohol moiety shows a b-blocking action,
we removed the hydroxy group of 1 to eliminate this
activity. As we expected, removing the hydroxy group
lowered the b-adrenoceptor blockade significantly
*Corresponding author. Tel.: +81-647-45-4356; fax: +81-467-45-
6824; e-mail: tatsuo.ohtsuka@roche.com
0960-894X/01/$ - see front matter # 2001 Elsevier Science Ltd. All rights reserved.
PII: S0960-894X(01)00319-5

1834
M. Masubuchi et al. / Bioorg. Med. Chem. Lett. 11 (2001) 1833–1837
without affecting its CaNmt enzyme inhibitory activity
(Table 1).13À15 The IC50s of 2 and 4 in the b-adreno-
ceptor blockade assay were 38 and 25 mM, respectively,
whereas those of the corresponding hydroxy derivatives
1 and 3 were 0.098 and 0.0043 mM, respectively.16 To
find the optimal chain length of the side chain, com-
pounds 5–7 were prepared and their enzyme inhibitory
activity was compared with that of compound 4 (Table 2).
The result indicated that trimethylene (4) was optimal.
For further optimization, we analyzed the X-ray crystal
structure (3.5 A) of a binary complex of CaNmt and (S)-
3 that was obtained by soaking experiments17 (Fig. 1).
The details of the X-ray crystallographic analysis will be
reported in a separate paper.18 The important interac-
tions between (S)-3 and CaNmt are summarized in
Figure 1. The C-4 substituent on the benzofuran ring
extends to a C-terminal leucine (Leu 451) of the poly-
peptide chain. The C-2 substituent is surrounded by
Table 2. Optimization of the C-4 substituent
Compd
R
Enzyme inhibition13 Antifungal activity14
CaNmt
50
HsNmt
C. albicans CY1002
5
>630
590
4
6
7
1.7
4.4
15
200
>580
>550
300
190
53
Scheme 1. Reagents and conditions: (a) n-BuLi, acyl chloride, THF,
À78 ꢀC to rt, 1 h, 25–68%, or n-BuLi, DMF, 0 ꢀC to rt, 17 h, 51%; (b)
4 N HCl, MeOH/1,4-dioxane=1:1, 50 ꢀC, 2–4 h, 77–93%; (c) ethyl
bromoacetate, K2CO3, acetone, 60 ꢀC, 15 h, or rt, 15 h, 38–73%; (d)
NaOEt, EtOH, 0 ꢀC, 20 h, 9–76%; (e) alkylene dibromide, K2CO3,
DMF, rt, 2–4 h, 43–88%; (f) amine, EtOH, 70 ꢀC, 12–48 h, 29–100%.
IC50: mM.
Table 1. Enzyme inhibitory activity, antifungal activity and b-adrenoceptor blockade of benzofuran derivatives
Compd
R
Enzyme inhibition13
Antifungal activity14
b-Adrenoceptor blockade15
CaNmt
HsNmt
194
C. albicans CY1002
1
0.98
390
0.098
2
3
4
4.4
1.2
1.7
380
470
200
470
210
300
38
0.0043
25
Alprenolol
NT
NT
NT
0.0023
NT, not tested; IC50: mM.

M. Masubuchi et al. / Bioorg. Med. Chem. Lett. 11 (2001) 1833–1837
1835
three phenylalanine residues, Phe 115, Phe 240 and Phe
339. The amino group in the C-4 side chain interacts
with the carboxyl group of Leu 451. This interaction
should be essential for the enzyme inhibitory activity
because the C-terminal carboxylate was reported to play
an important role in the catalytic mechanism.19 The
tert-butyl amino group in the C-4 side chain is sur-
rounded by hydrophobic aromatic amino acid residues,
Tyr 107, Tyr 119 and Phe 176. This structural analysis
suggested that the replacement of the tert-butyl group
by another hydrophobic group would increase the
enzyme inhibitory activity. Thus, we designed and syn-
thesized compounds 8–11 having a hydrophobic group
on the amino group.
The assay results are shown in Table 3. Among them,
the (pyridin-3-ylmethyl)amino derivative 11 (RO-09-
4609) showed the strongest enzyme inhibitory activity
Figure 1. X-ray crystal structure of a binary complex of CaNmt and (S)-3 (only the binding site is shown).
Table 3. Optimization of the amino substituent in the C-4 side chain
Compd
R
Enzyme inhibition13
Antifungal activity14
C. albicans CY1002
300
CaNmt
1.7
HsNmt
200
4
8
1.6
>570
3.3
170
>570
530
85
>280
10
9
10
11 (RO-09-4609)
12
0.1
>540
>540
1.6
>540
>270
IC50: mM.

1836
M. Masubuchi et al. / Bioorg. Med. Chem. Lett. 11 (2001) 1833–1837
against CaNmt with extremely high selectivity (>5000-
fold) over human Nmt. The high selectivity can be
explained by the differences between CaNmt and
HsNmt of the amino acid sequence alignment. The
identity of the amino acid sequence alignment of the
two Nmts is only 48%, which makes 26 amino acid
residues around the benzofuran binding site different
between the two Nmts (data not shown). It also showed
moderate antifungal activity against C. albicans CY1002
in vitro. The increased enzyme inhibitory activity of 11
over the benzyl derivative 10 can be explained by an
additional favorable hydrogen bonding between the
pyridine nitrogen and the Tyr 119 hydroxy group (Fig.
2). This interpretation was supported by the finding that
the inhibitory activity of 11 against a mutant CaNmt
having alanine for Tyr 119 was 14 times weaker than
that against the wild CaNmt.20
ionic interaction between the carboxylate and the basic
amino group of the inhibitor.21
The crystal structure revealed that the benzofuran ring
occupied the hydrophobic pocket of the binding site and
little space was left for an extra substituent at the C-6
and C-7 positions. This observation was consistent with
the SAR finding that mono-methylation at the C-7
position lowered CaNmt inhibitory activity by a factor
of 252.
Since there seemed to be some space around the C-3
position, the methyl group at the C-3 position of 11 was
replaced by a larger substituent. The assay results of the
C-3 derivatives are shown in Table 4. However, the
result indicated that methyl was still the best among
them. The X-ray crystallographic analysis did not
clearly explain the reason why neither ethyl derivative
13 nor cyclopropyl derivative 14 showed strong inhibi-
tory activity against CaNmt, though there appeared to
be enough space for them.
Since aniline derivative 9 and pyridin-3-ylmethoxy deri-
vative 12 showed no activity at all, it is suggested that
(1) the basicity of the aniline is too weak to interact with
the C-terminal carboxylate (Leu 451) and (2) strong
basicity of the aliphatic amino group is essential for
Compound 11 did not show antifungal activity in a
mouse systemic candidiasis model, although the anti-
fungal activity in vitro was nearly as strong as that of
fluconazole.
In summary, the structural information of the CaNmt
complex with compound (S)-3 was effectively utilized
for designing more potent inhibitors and led to the
identification of 11 (RO-09-4609), a new class of anti-
fungal agents, which showed strong competitive CaNmt
inhibitory activity (Ki=0.27 mM), extremely high selec-
tivity as compared to HsNmt and moderate antifungal
activity against C. albicans in vitro. The optimal sub-
stituents on C-3 and C-4 were revealed to be methyl and 3-
[(pyridin-3-ylmethyl)amino]propoxy, respectively. Com-
pound 11 was selected as a lead compound for further
optimization. The mechanism of the antifungal activity
is being investigated in detail.
Figure 2. Hypothetical interaction between 11 and Tyr 119 of CaNmt.
Table 4. Optimization of the C-3 substituent
Compd
R
Enzyme inhibition13
Antifungal activity14
CaNmt
HsNmt
C. albicans CY1002
11 (RO-09-4609)
CH3
CH2CH3
Cyclopropyl
Isopropyl
H
0.1
10
4.4
83
>540
280
91
260
>560
1.6
12
11
NT
NT
13
14
15
16
79
SC-5827222
Fluconazole
0.83
NT
>140
NT
200
0.72
NT, not tested; IC50: mM.

M. Masubuchi et al. / Bioorg. Med. Chem. Lett. 11 (2001) 1833–1837
1837
References and Notes
istoyl–CoA at 0.5 mM. Potency against Human Nmt as asses-
sed by IC50 using substrate peptide GNAASARR-NH2 (0.5
mM) and myristoyl–CoA at 0.5 mM. NT, not tested.
14. Broth dilution method; medium: YNBPB (=YNB+1%
glucose+0.25% K2HPO4), pH 7.0, inoculum size; 1Â104 cfu/
mL, incubation: 1 day at 27 ꢀC.
1. Gordon, J. I.; Duronio, R. J.; Rudnick, D. A.; Adams,
S. P.; Gokel, G. W. J. Biol. Chem. 1991, 266, 8647.
2. Johnson, D. R.; Bhatnagar, R. S.; Knoll, L. J.; Gordon, J. I.
Annu. Rev. Biochem. 1994, 63, 869.
15. Adrenergic, beta, non-selective assay, performed by
NOVA SCREEN.
3. Stearns, T.; Kahn, R. A.; Botstein, D.; Hoyt, M. A. Mol.
Cell Biol. 1990, 10, 6690.
16. Treatment of 3-methyl-4-oxiranylmethoxy-benzofuran-2-
carboxylic acid ethyl ester (Sangwan, N. K.; Rastogi, S. N.;
Kar, K. Eur. J. Med. Chem. 1987, 22, 153) with tert-butyl-
amine in ethanol at 70 ꢀC for 2 h afforded 3.
17. Compound (S)-3 was the more active isomer (IC50 against
CaNmt: 1.1 mM) and (R)-3 was the less active isomer (IC50
against CaNmt: 150 mM).
4. Herman, P. K.; Stack, J. H.; DeModena, J. A.; Emr, S. D.
Cell 1991, 64, 425.
5. Weinberg, R. A.; McWherter, C. A.; Freeman, S. K.;
Wood, D. C.; Gordon, J. I.; Lee, S. C. Mol. Microbiol. 1995,
16, 241.
6. Lodge, J. K.; Jackson-Machelski, E.; Toffaletti, D. L.; Per-
fect, J. R.; Gordon, J. I. Proc. Natl. Acad. Sci. U.S.A. 1994,
91, 12008.
7. Nagarajan, S. R.; Gordon, J. I.; Sikorski, J. A. J. Med.
Chem. 1997, 40, 1422.
8. Devadas, B.; Gordon, J. I.; Sikorski, J. A. J. Med. Chem.
1997, 40, 2609.
9. Devadas, B.; Gordon, J. I.; Sikorski, J. A. J. Med. Chem.
1998, 41, 996.
10. Grinev, A. N.; Mashkovskii, M. D.; Kaverina, N. V.;
Lyubchanskaya, V. M.; Uretskaya, G. Ya.; Zaitseva, K. A.;
Shvarts, G. Ya.; Muratov, M. A.; Senova, Z. P.; Lyskovtsev,
V. V. Russian Patent SU 77-2494901, 1986.
11. Aoki, Y.; Ebiike, H.; Fujii, T.; Kawasaki, K.; Liu, P.;
Masubuchi, M.; Ohtsuka, T.; Tsujii, S. Eur. Patent 09763, 1999.
12. Atkinson, J. G.; Guindon, Y.; Lau, C. K. Eur. Patent
146243, 1985.
18. Manuscript in preparation. The coordinates of the protein
will be deposited with PDB.
19. Bhatnagar, R. S.; Futterer, K.; Farazi, T. A.; Korolev, S.;
Murray, C. L.; Jackson-MacHelski, E.; Gokel, G. W.; Gor-
don, J. I.; Waksman, G. Nat. Struct. Biol. 1998, 5, 1091.
20. There are two tyrosine residues (Tyr 119 and Tyr 107)
surrounding the pyridyl group of compound 11. Alanine sub-
stitution of the tyrosine residues indicated that only Tyr 119
interacts with 11. For wild type CaNmt, the Ki value of com-
pound 11 was 0.27 mM. For Tyr119Ala and Tyr107Ala, the Ki
values of compound 11 are 3.9 and 0.33 mM, respectively.
21. Compound 12 was prepared by treatment of compound
III (R1=CH3) with 3-pyridinemethanol and NaH in DMF at
room temperature for 19 h in 11% yield.
22. A dipeptide inhibitor of CaNmt, N-[[4-[4-(2-methyl-1H-
imidazol-1-yl)butyl]phenyl]acetyl]-l-seryl-N-(2-cyclohexylethyl)-
l-lysinamide, reported in refs 8, 9, and 19.
13. Potency against C. albicans Nmt as assessed by IC50 using
substrate peptide GLTISKLFRR-NH2 (0.5 mM) and myr-