Identification of a broad-spectrum azasordarin with improved pharmacokinetic properties

Article abstract of DOI:10.1016/S0960-894X(03)00161-6

The synthesis and antifungal activity of 5′- and 5′-6′-substituted azasordarin derivatives are described. Modification of the 5′-position led to the discovery of the spirocyclopentyl analogue 7g, which is the first azasordarin to register single-digit MIC values versus Aspergillus spp. Further investigation identified the 5′-i-Pr derivative 7b, which displays superior pharmacokinetic properties compared to other azasordarins.

Full text of DOI:10.1016/S0960-894X(03)00161-6

Bioorganic & Medicinal Chemistry Letters 13 (2003) 1419–1423
Identification of a Broad-Spectrum Azasordarin with Improved
Pharmacokinetic Properties
Michael H. Serrano-Wu,a,* Denis R. St. Laurent,a Tina M. Carroll,a Marco Dodier,b
Qi Gao,a Patrice Gill,b Claude Quesnelle,b Anne Marinier,b Charles E. Mazzucco,a
Alicia Regueiro-Ren,a Terry M. Stickle,a Dedong Wu,a Hyekyung Yang,a Zheng Yang,a
Ming Zheng,a Mary E. Zoeckler,a Dolatrai M. Vyasa and Balu N. Balasubramaniana
aBristol-Myers Squibb Pharmaceutical Research Institute, 5 Research Parkway, Wallingford, CT 06492, USA
´
Bristol-Myers Squibb Pharmaceutical Research Institute, 100, boul. de l’Industrie, Candiac, Quebec, Canada J5R 1J1
b
Received 13 December 2002; revised 10 February 2003; accepted 10 February 2003
Abstract—The synthesis and antifungal activity of 50- and 50-60-substituted azasordarin derivatives are described. Modification of
the 50-position led to the discovery of the spirocyclopentyl analogue 7g, which is the first azasordarin to register single-digit MIC
values versus Aspergillus spp. Further investigation identified the 50-i-Pr derivative 7b, which displays superior pharmacokinetic
properties compared to other azasordarins.
# 2003 Elsevier Science Ltd. All rights reserved.
Invasive fungi have in recent decades emerged as
important pathogens, in particular among the growing
and increasingly diverse population of immunocompro-
mised patients. Specifically, infections due to Candida
spp. constitute the fourth leading cause of nosocomial
bloodstream infections, while the incidence and mortal-
ity rates associated with the fungal pathogens Aspergil-
lus spp. and Cryptoccocus neoformans are rapidly on the
rise.1 Despite the growing medical need, however, few
existing therapies with suitable safety and broad spec-
trum profiles are available for the treatment of systemic
fungal infections. Moreover, the emergence of fungal
pathogens resistant to current chemotherapies further
increases the demand for the discovery and development
of novel antifungal agents.
Figure 1. Sordarin (1) and azasordarin (2).
EF-2 (85%), sordarin is able to stabilize the post-trans-
lational complex through high-affinity binding to the
fungal ribosome/EF-2 complex and block translocation.5
The selective inhibition of fungal protein synthesis thus
became an attractive target for the development of new
antifungal agents that were mechanistically distinct
from existing therapies.
The natural product sordarin (1, Fig. 1) was discovered
in 1971 as a metabolite of Sordaria araneosa and identi-
fied as a potent antifungal compound.2 Interest in this
diterpene glycoside has recently been sparked by the
identification of its cellular target in fungal organisms,
the transcription elongation factor EF-2.3,4 Despite the
high sequence homology between fungal and human
Researchers at Glaxo have recently disclosed aza-
sordarin derivatives such as 2 in which the 60-deoxy-
glycoside found in the natural product has been
replaced by a morpholine ring.6,7 While compound 2
extends the antifungal spectrum of the sordarin class
to include Candida glabrata and C. parapsilosis, no
inhibitory activity was observed against C . neoformans
or the filamentous fungi Aspergillus spp. Moreover,
*Corresponding author. Tel.:+1-203-677-7019; fax:+1-203-677-7702;
e-mail: michael.serrano-wu@bms.com
0960-894X/03/$ - see front matter # 2003 Elsevier Science Ltd. All rights reserved.
doi:10.1016/S0960-894X(03)00161-6

1420
M. H. Serrano-Wu et al. / Bioorg. Med. Chem. Lett. 13 (2003) 1419–1423
pharmacokinetic analysis demonstrated that 2 dis-
played a short plasma half-life in mice, and that b.i.d.
(or higher) dosing frequency was required to demon-
strate efficacy.8 Having identified these limitations in
spectrum and bioavailability, our goal at the outset of
our discovery efforts was to identify novel azasordarins
with improved antifungal and metabolic properties.
Our strategy was to explore diverse substitution pat-
terns on the morpholine ring in order to expand the
SAR of the azasordarins and identify potential inter-
actions within the active sites of less-susceptible fungal
pathogens.
Figure 2. Impact of C5 stereochemistry on glycosidic linkage.
and ester protecting groups were removed to reveal the
morpholino acid 6, which was N-alkylated under stan-
dard conditions (R-X, NaHCO3, EtOH, 80 ꢀC) to afford
the final azasordarin targets 7a–g.12
In order to prepare azasordarin derivatives with novel
ring substitution patterns, we anticipated the synthesis
of morpholino lactols and subsequent coupling with the
sordarin aglycone, known as sordaricin. A wide range
of enantiopure amino alcohols9 were thus condensed
with p-anisaldehyde, and then subjected to a second
reductive amination event with dimethoxyacetaldehyde
(Scheme 1). Acidic hydrolysis of the dimethyl ketal
intermediate then afforded the ring-closed morpholino
lactols 4 in 20–40% overall yield from amino alcohol.
The in vitro activity of the novel 5-substituted morpho-
lin-2-yl azasordarins was determined against a diverse
panel of fungal pathogens (Table 1). A measurable
improvement in spectrum was observed in transposing
the methyl substituent from the 60- to 50-position of the
morpholino ring. Namely, the 60-Me analogue 2 was
devoid of activity versus the pathogens C. neoformans
and Aspergillus fumigatus, whereas the 50-Me analogue
7a registered MIC values of <0.06 and 32 mg/mL,
respectively. Polar functionality was not tolerated at the
50-position (7c, R=CH2OH), which is consistent with
structure–activity relationships identified in other hetero-
cyclic series.7b
To construct the anomeric bond, we treated the morpho-
lino lactols 4 with NaH in THF, and then alkylated with
a suitably protected sordaricin triflate derivative as has
been previously reported.10 The diastereoselectivity of
this lactol alkylation appeared to depend on the resident
chirality in the morpholino ring, as the lactol derived
from (R)-valinol afforded the b-anomeric configuration
found in the natural product as the major adduct,
whereas the (S)-derived diastereomer gave rise to the
unnatural a-configuration. While we do not fully
understand the origins of this stereoselectivity, some
insight may be gleaned by examining the low-energy
chair conformations of the glycosidated products (Fig.
2). The observed major products allow the diequatorial
disposition of both the 5-i-Pr and the glycosidic sub-
stituents, and presumably this is lower in energy than a
transition-state assembly that situates either of these
groups axial.11 Following lactol alkylation, the amine
The 5,6-disubstituted morpholine derivatives 7e
(R5=R6=Me) and 7f [R5, R6=–(CH2)3–] confirmed
our observation that substitution of the 50-position
broadened the antifungal spectrum of the azasordarins,
with single digit values observed for C. parapsilosis and
Aspergillus flavus and values of <0.25 mg/mL versus C.
albicans, C. glabrata, and Cryptococcus neoformans.13
Excellent potency was also observed with the i-Pr-sub-
stituted derivative 7b, with MIC values less than 0.06
mg/mL observed versus C. albicans, C. glabrata, and C.
neoformans. The geminally-substituted spiro-cyclopentyl
morpholin-2-yl azasordarin 7g exhibited the best MIC
profile of all the derivatives listed in Table 1, and is also
the only compound to register a single digit (8 mg/mL)
MIC versus A. fumigatus.14 Collectively, the high degree
of lipophilicity and steric bulk tolerated at the 50-posi-
tion of the morpholine ring suggest a possible binding
mode for the azasordarins within a hydrophobic pocket
of the fungal protein synthesis machinery.
We next embarked on identifying the optimal nitrogen
substitution pattern on the morpholine ring. Synthetic
considerations prompted us to investigate this position
in the 50-Me series, as analogues with bulkier 50-sub-
stituents (i.e., i-Pr) were either unreactive or afforded
mixtures of amine and carboxylate alkylation products
under our reaction conditions. An extremely stringent
SAR emerged in this 50-Me azasordarin series (Table 2).
Namely, N-allyl substitution appeared to be required in
order to observe antifungal activity that included C.
parapsilosis, C. neoformans, and Asp. spp. Within the
N-allyl derivatives, halogenation at the 2-position was
preferred over 3-halo substitution, while surprisingly the
2-cyano derivative 8e was less active than electronically-
similar halogenated analogues. The importance of the
sidechain olefin is underscored by the disparate MICs
Scheme 1. Synthesis of azasordarin derivatives. (a) p-anisaldehyde,
polymer-supported BHꢁ4 , MeOH; (b) dimethoxyacetaldehyde, ZnCl2,
NaBH3CN, MeOH; (c) 6 N HCl, 80 ꢀC; (d) Bn sordaricin triflate,
NaH, THF; (e) H2, Pd(OH)2 on C, MeOH; (f) R4-X, NaHCO3, NaI,
EtOH, 80 ꢀC; (g) RCHO, NaBH3CN, CH3CN.

M. H. Serrano-Wu et al. / Bioorg. Med. Chem. Lett. 13 (2003) 1419–1423
1421
observed with the 2-methallyl derivative 8d and the
corresponding saturated analogue 8n, which completely
lacks measurable activity versus C. parapsilosis and Asp.
spp., and diminished potency versus C. neoformans.
However, N-benzyl (i.e., 8o and 8p) or heteroaromatic
derivatives (8q) are not suitable replacements for the
allyl moiety.
A known liability of previously disclosed azasordarin
analogues is their short plasma half-lives in mice,
Table 1. Antifungal activity of sordarin morpholino derivatives 2 and 7a–g
Compd
R5
R6
MICa (mg/mL)
CP
CA
CN
CG
AFa
ALb
2
—
Me
i-Pr
CH2OH
CF3
Me
Me
—
—
—
—
Me
8
1
<0.06
<0.06
<0.06
0.5
<0.06
<0.06
<0.06
<0.008
>128
0.125
0.125
<0.06
1
>128
32
32
>128
>128
64
4
0.5
2
7a
7b
7c
7d
7e
7f
7g
<0.06
<0.06
2
2
128
>128
8
>128
>128
>128
1
0.5
0.25
<0.008
0.25
0.125
0.015
1
4
Cyclopentyl
Spirocyclopentyl
4
0.5
64
8
—
0.125
aMIC value defined as the lowest drug concentration required to inhibit 90–100% visible growth over 48 h relative to controls. CP=C. parapsilosis
CA=C. albicans CN=Cryptococcus neoformans CG=C. glabrata AF=Aspergillus fumigatus AL=Aspergillus flavus.
bMIC determined after 24 h.
Table 2. Variation of N-substitution pattern
Compd
R4
MIC (mg/mL)
CP
CA
CN
CG
64
0.25
0.125
0.125
0.125
1
AF
AL
8a
8b
7a
8c
8d
8e
8f
8g
8h
8i
8j
8k
8l
8m
8n
8o
8p
8q
8r
8s
H
Allyl
2-Chloroallyl
2-Bromoallyl
2-Methallyl
>128
64
8
>128
128
<0.06
<0.06
0.125
4
1
128
>128
16
64
>128
>128
>128
32
>128
>128
32
>128
>128
0.125
<0.06
<0.06
<0.06
0.25
0.125
0.125
<0.06
<0.06
0.125
0.25
1
4
8
0.5
4
8
128
>128
>128
>128
>128
>128
>128
>128
>128
>128
>128
>128
>128
>128
>128
>128
>128
2-Cyanoallyl
>128
>128
>128
>128
32
128
2-Chloroallyloxy
cis-3-Chloroallyl
cis-3-Bromoallyl
3,3-Dichloroallyl
3,3-Difluoroallyl
3,3-Dimethylallyl
Cyclohex-2-enyl
Cyclopropylmethyl
i-Bu
4-MeO benzyl
2-Cl benzyl
2-(5-Chlorothiophenyl)
Acroyl
Cyclopropylsulfonyl
2
1
>128
>128
>128
>128
>128
>128
>128
>128
>128
>128
>128
>128
>128
>128
1
<0.06
1
8
8
8
64
>128
>128
>128
>128
4
>128
>128
>128
>128
2
2
<0.06
<0.06
<0.06
<0.06
8
1
0.5
4
2
<0.06
1
0.125
64
128
>128
32
128

1422
M. H. Serrano-Wu et al. / Bioorg. Med. Chem. Lett. 13 (2003) 1419–1423
necessitating dose frequencies of b.i.d. or higher to
demonstrate efficacy.8 Indeed, pharmacokinetic profil-
ing of the 50-Me substituted analogue 7a revealed rapid
clearance from mice plasma following iv administration
(total body clearance rate=95 mL/min/kg, Table 3).15
In vitro biotransformation studies identified N-deal-
kylation of azasordarin analogues as a significant lia-
bility that limited in vivo exposure. While the stringent
SAR requirements outlined above prohibited drastic
structural variation on the N-substituent, we discovered
that increasing steric bulk at the 50-position of the mor-
pholine slowed metabolic N-dealkylation while main-
taining broad spectrum antifungal activity. For
example, the rate of disappearance of i-Pr-substituted
7b following incubation with human (rate=0.092 nmol/
min/mg) and mouse (0.082 nmol/min/mg) liver micro-
somes was considerably lower than for Me-substituted
analogue 7a (human: 0.214 nmol/min/mg; mouse: 0.154
nmol/min/mg).
be quite versatile towards improving potency against
less-susceptible pathogens (i.e., A. flavus) as well as
reducing hepatic clearance, and led to the discovery of
the most potent sordarin derivative (7g) disclosed to
date. The in vivo efficacy of 7g and other analogues such
as 7b in systemic fungal infection models (i.e., Crypto-
coccus) thought to be resistant to the sordarins will be
reported in due course.
Acknowledgements
The authors would like to thank Dr. Thomas Tully and
colleagues for supplying bulk quantities of sordarin. We
are also grateful to Dr. Gene Dubowchik for helpful
suggestions to this manuscript.
References and Notes
The improved metabolic stability of 7b versus 7a was
validated in vivo by measuring serum concentrations of
each compound in mice following iv and oral dosing
(Table 3). The total body clearance of 7b was sig-
nificantly reduced to 20 mL/min/kg, which correlates
well with the slower rate of metabolism observed for
this analogue. With respect to 2, compound 7b dis-
played a longer iv half-life (T1/2=1.6 h), which may be
attributed to the greater volume of distribution for 7b.
Compound 7b appears therefore to have favorable
pharmacokinetic properties compared to other aza-
sordarins with respect to metabolic stability and plasma
half-life, suggesting that further structural modifications
to enhance the bioavailability of other azasordarin
derivatives should be possible.
1. Stevens, D. A.; Holmberg, K. Current Opin. in Anti-infec-
tive Invest. Drugs 1999, 1, 306.
2. Hauser, D.; Sigg, H. P. Helvetica Chim Acta 1971, 54, 1178.
3. (a) Dominguez, J. M.; Martin, J. J. Antimicrob. Agents
Chemotherapy 1998, 42, 2274. (b) Capa, L.; Mendoza, A.;
Lavandra, J. L.; Gomez de Las, H. F.; Garcia-Bustos, J. F.
Antimicrob. Agents Chemotherapy 1998, 42, 2279. (c) Justice,
M. C.; Hsu, M. J.; Tse, B.; Ku, T.; Balkovec, J.; Schmatz, D.;
Nielsen, J. J. Biol. Chem. 1998, 273, 3148.
4. (a) For reviews, see: Odds, F. C. Exp. Opin. Ther. Pat.
2001, 11, 283. (b) Ziegelbauer, K.; Spaltmann, F. Drugs Fut.
2000, 25, 63. (c) Gargolla-Viola, D. Current Opin. in Anti-
infective Invest. Drugs 1999, 1, 297.
5. Dominguez, J. M.; Gomez-Lorenzo, M. G.; Martin, J. J.
J. Biol. Chem. 1999, 274, 22423.
6. Herreros, E.; Almela, M. J.; Lozano, S.; Gomez de las
Heras, F.; Gargallo-Viola, D. Antimicrob. Agents Chemother.
2001, 45, 3132.
7. (a) For other heterocyclic replacements of the tetrahydro-
pyranyl sidechain, see: Kaneko, S.; Arai, M.; Uchida, T.;
Harasaki, T.; Fukuoka, T.; Konosu, T. Bioorg. Med. Chem.
Lett. 2002, 12, 1705. (b) Serrano-Wu, M. H.; St. Laurent,
D. R.; Chen, Y.; Huang, S.; Lam, K.; Matson, J. A.; Mazzucco,
C. E.; Stickle, T. M.; Tully, T. P.; Wong, H. S.; Vyas, D. M.;
Balasubramanian, B. N. Bioorg. Med. Chem. Lett. 2002, 12,
2757. (c) Arai, M.; Harasaki, T.; Fukuoka, T.; Kaneko, S.;
Konosu, T. Bioorg. Med. Chem. Lett. 2002, 12, 2733.
8. Aviles, P.; Pateman, A.; San Roman, R.; Guillen, M.;
Gargolla-Viola, D. 40th Meeting of the Intersciences Con-
ference of Antimicrobial Agents and Chemotherapy (2000),
Abstract 1690.
In conclusion, we have identified a series of 50-sub-
stituted morpholin-2-yl azasordarin derivatives with
superior antifungal and pharmacokinetic properties.
The 50-position of the morpholino sidechain proved to
Table 3. Mouse pharmacokinetic parameters of azasordarin analo-
gues 2, 7a, and 7b following iv (5 mg/kg) and oral (20 mg/kg) admin-
istration (n=3)
9. (a) Amino alcohols were commercially available or pre-
pared from the corresponding epoxide. For example, see:
Schaus, S. E.; Larrow, J. F.; Jacobsen, E. N. J. Org. Chem.
1997, 62, 4197. (b) Katagiri, T.; Takahashi, M.; Fujiwara, Y.;
Ihara, H.; Uneyama, K. J. Org. Chem. 1999, 64, 7323.
10. (a) Coteron, J. M.; Chiara, J. L.; Fernandez-Mayoralas,
A.; Fiandor, J. M.; Valle, N. Tetrahedron Lett. 2000, 41, 4373.
(b) Bueno, J. M.; Chicharro Gonzalo, J.; Coteron, J.; Cuevas,
J. C.; Fiandor, J. M.; Mallo, A. ; WO 9958512; Chem. Abstr.
1999, 131, 337207.
27a
7b
Cmax (ng/mL)
Total CL (mL/min/kg)
Vss (L/kg)
iv T1/2 (h)
Oral T1/2 (h)
F(%)
N/A
13a
0.49a
0.44a
N/A
N/A
5946
95
3882
20
11. The greater nucleophilicity of b-alkoxide pyranose anions
has been elegantly demonstrated by Schmidt; see: Schmidt, R.
Angew. Chem., Int. Ed. Engl. 1986, 25, 212. We are grateful to
a referee for bringing this reference to our attention.
12. The purity and identity of all new compounds were
7.1
1.8
2.1
1.6
1.6
3.1
100
83
1
aData obtained from ref 8.
established by H NMR and HRMS. For instance, analytical

M. H. Serrano-Wu et al. / Bioorg. Med. Chem. Lett. 13 (2003) 1419–1423
1423
1
data for 7b: H NMR (CDCl3, 500 MHz) d 9.80 (s, 1H), 6.06
not be susceptible to sordarin derivatives. See: Shastry, M.;
Nielsen, J.; Ku, T.; Hsu, M.; Liberator, P.; Anderson, J.;
Schmatz, D.; Justice, M. C. Microbiology 2001, 147, 383. (b)
Santos, C.; Ballesta, J. P. G. Molec. Microb. 2002, 43, 227.
14. The anomeric configuration of 7g was established by
X-ray crystallography of the corresponding morpholine pre-
cursor 6. The a-anomer of compound 7g was considerably less
active against all pathogens tested (C. albicans MIC=0.5 mg/
mL; C. glabrata MIC=4 mg/mL; all others>16 mg/mL).
15. Although 7a exhibited high total body clearance after iv
administration, the oral bioavailability was 100%, which may
be attributed to the rapid absorption of 7a and the resulting
saturation of first-pass liver metabolism.
(m, 1H), 5.42 (s, 1H), 5.34 (s, 1H), 4.43 (dd, J=7.0, 2.7 Hz,
1H), 4.23 (d, J=9.5 Hz, 1H), 4.12 (app q, J=7.0 Hz, 2H), 3.88
(dd, J=11.9, 3.4 Hz, 1H), 3.59 (d, J=15.0 Hz, 1H), 3.44 (dd,
J=11.6, 7.3 Hz, 1H), 3.40 (d, J=9.5 Hz, 1H), 3.02 (d, J=14.6
Hz, 1H), 3.00 (dd, J=11.9, 4.0 Hz, 1H), 2.55 (m, 1H), 2.33 (m,
1H), 1.75-2.24 (m, 8H), 1.26 (m, 3H) 1.03 (d, J=7.0 Hz, 3H),
0.99 (d, J=6.7 Hz, 3H), 0.93 (d, J=6.7 Hz, 3H), 0.90 (d, J=7.0
Hz, 3H), 0.81 (d, J=7.0 Hz, 3H). HRMS (ES) exact mass calcd
for C30H44ClNO5 (M+H+): 534.2986; found 534.2987.
13. (a) The exceptional MIC observed with several analogues
versus C. neoformans and Asp. fumigatus is interesting in light
of recent studies which suggest that EF-2 in these fungi should