Probing the role of the mycosamine C2′-OH on the activity of amphotericin B

Article abstract of DOI:10.1021/ol2000765

A synthetic route to a mycosamine donor was designed and provided access to a set of AmB derivatives targeted to probe the effect of the C2′-OH. It was determined that the configuration of the C2′-position is inconsequential but that O-methylation of this alcohol was deleterious to its mode of action. Additionally, the analog incorporating a mycosamine derivative from the enantiomeric series was devoid of activity.

Full text of DOI:10.1021/ol2000765

ORGANIC
LETTERS
2011
Vol. 13, No. 6
1390–1393
Probing the Role of the Mycosamine
C20-OH on the Activity of Amphotericin B
Mitchell P. Croatt and Erick M. Carreira*
Laboratorium fu€r Organische Chemie, ETH Zu€rich, CH-8093 Zu€rich, Switzerland
carreira@org.chem.ethz.ch
Received January 11, 2011
ABSTRACT
A synthetic route to a mycosamine donor was designed and provided access to a set of AmB derivatives targeted to probe the effect of the C20-OH.
It was determined that the configuration of the C20-position is inconsequential but that O-methylation of this alcohol was deleterious to its mode of
action. Additionally, the analog incorporating a mycosamine derivative from the enantiomeric series was devoid of activity.
The drug of choice for hospital patients with systemic
fungal infections is amphotericin B (AmB, 1, Figure 1).
Yet, the full details concerning its mode of action remain
elusive despite efforts over four decades.1 Indeed, there is a
plenum of questions unanswered concerning the remark-
able ability of this small molecule to undergo self-assembly
and recognition in lipid bilayers. This becomes even more
significant given the fact that AmB has been found to have
activity in treating Alzheimers, Creutzfeldt-Jakob, and
HIV infections, wherein it has been suggested that AmB/
protein interactions may be relevant. A number of deriva-
tives and closely related structures have been isolated,2
biosynthesized,3 and accessed by total synthesis.4,5 These
have been used to probe the role of the various functional
groups in the mode of action, including the hydroxyls
along the C1-C12 spine, the polyene,6 and carboxylic
acid.7 Carbohydrates appended to natural products are
increasingly being recognized as critical to their activity.8
Yet, despite the fact that mycosamine is a highly conserved
moiety in the family of polyene macrolide antibiotics, no
studies have been directed at understanding its effect on
yeast at a resolution that focuses on the individual func-
tional groups. Herein we document an initial investigation
of the role of the mycosamine C20-OH on the fungicidal
activity of AmB. In the course of our studies, we have
(1) (a) Volmer, A. A.; Szpilman, A. M.; Carreira, E. M. Nat. Prod.
Rep. 2010, 27, 1329. (b) Cereghetti, D. M.; Carreira, E. M. Synthesis
2006, 914.
(2) (a) Zotchev, S. B. Curr. Med. Chem. 2003, 10, 211. (b) Rychnovsky,
S. D. Chem. Rev. 1995, 2021.
(3) (a) Aparicio, J. F.; Caffrey, P.; Gil, J. A.; Zotchev, S. B. Appl.
Microbiol. Biotechnol. 2003, 61, 179. (b) Caffrey, P.; Lynch, S.; Flood,
E.; Finnan, S.; Oliynyk, M. Chem. Biol. 2001, 8, 713. (c) McNamara,
C. M.; Box, S.; Crawforth, J. M.; Hickman, B. S.; Norwood, T. J.;
Rawlings, B. J. J. Chem. Soc., Perkin Trans. 1 1998, 83.
(4) (a) Szpilman, A. M.; Cereghetti, D. M.; Wurtz, N. R.; Manthorpe,
J. M.; Carreira, E. M. Angew. Chem., Int. Ed. 2008, 47, 4335. (b) Szpilman,
A. M.; Manthorpe, J. M.; Carreira, E. M. Angew. Chem., Int. Ed. 2008, 47,
4339.
(6) Brautaset, T.; Sletta, H.; Nedal, A.; Borgos, S. E. F.; Degnes,
K. F.; Bakke, I.; Volokhan, O.; Sekurova, O. N.; Treshalin, I. D.;
Mirchink, E. P.; Dikiy, A.; Ellingsen, T. E.; Zotchev, S. B. Chem. Biol.
2008, 15, 1198.
(7) (a) Carmody, M.; Murphy, B.; Byrne, B.; Power, P.; Rai, D.;
Rawlings, B.; Caffrey, P. J. Biol. Chem. 2005, 280, 34420. (b) Palacios,
D. S.; Anderson, T. M.; Burke, M. D. J. Am. Chem. Soc. 2007, 129,
13804.
(8) (a) For a recent review, see: Ernst, B.; Magnani, J. L. Nat. Rev.
Drug Discovery 2009, 8, 661. (b) For significant examples, see: Silva,
D. J.; Kahne, D. E. J. Am. Chem. Soc. 1993, 115, 7962. (c) Mergott, D. J.;
Frank, S. A.; Roush, W. R. Proc. Natl. Acad. Sci. U.S.A. 2004, 101,
11955. (d) Wang, P.; Zhu, J.; Yuan, Y.; Danishefsky, S. J. J. Am. Chem.
Soc. 2009, 131, 16669.
(5) Szpilman, A. M.; Carreira, E. M. Angew. Chem., Int. Ed. 2010, 49,
9592.
r
10.1021/ol2000765
Published on Web 02/15/2011
2011 American Chemical Society

Scheme 1. Synthesis of Mycosamine Donors
Figure 1. Structure of amphotericin B and synthetic analogs.
intriguing aspects of this model the role of C20-OH has
not been experimentally explicitly tested with derivatives
thatdonot extensivelyalter the mycosamine orits relation-
ship to the macrocycle. We therefore designed analogues
that would directly probe the putative H-bonding interac-
tion between mycosamine and ergosterol and further map
out subtleties of the mycosamine subunit on the activity of
the polyene macrolide antibiotic AmB. In approaching the
problem, we decided to take a course of action that
minimized the perturbations of the mycosamine itself.
Thus it was our intention to maintain the overall polar
and structural characteristics of the parent sugar.
observed that the relative configuration at C20 is not a
determinant for activity, and yet the corresponding C20-
methyl ether analog is 50-fold less effective (Figure 1).
Interestingly, the preparation of AmB that incorporates
the ent-C20-epi-mycosamine displays 100-fold decreased
activity in yeast. These results implicate a key role for
mycosamine in AmB wherein the aminosugar is not merely
a polar headgroup and, importantly, whose role is more
significant than merely forming a hydrogen bond to sterols
residing nearby.
The early literature of AmB describes simple derivatives
at the C-16-CO2H and mycosamine-NH2 that provide
some indication of their importance.1 For example, AmB
methyl ester (AmE, 2) is equipotent to AmB in yeast
(AmB = 0.30 μM and AmE = 0.25 μM), but mycosamine
N-acetamides are inactive, leading to the suggestion that a
basic amine at C30 is necessary.9 Caffrey has determined
that the aglycone of AmB is inactive, thus underscoring the
relevance of the aminosugar itself.7 Moreover, we have
shown that amphotericin B incorporating N,N-bis(amino-
propylene) derived mycosamines display enhanced fun-
gicidal activity and possess an improved therapeutic
window.10
Aninteraction involvingmycosamine C20 and ergosterol
hydroxyl groups was suggested some time ago to be critical
in enabling the sterol to stabilize the putative barrel-stave
pore in fungal membranes.11 More recently, studies have
been conducted with conformationally rigidified analogs
in which short hydrophobic tethers linked the mycosamine
amine with the C-16 carboxylic acid.12 It is well worth
noting that the calculated conformation for the three
structures prepared and analyzed did not correlate to the
most stable conformation relating mycosamine and
the macrocycle, discussed in detail below. Despite the
We have described an improved glycosidation method
for the AmB aglycone in the context of a synthesis of C-35
deoxyAmB.13,14 The approach relies on the use of the C20-
mycosamine epimer as a donor because of exquisite
anomeric control that ensues in the coupling reaction.
Thus, the route enables rapid access to C20-derivatives
and provides a point of entry for studies at this locus.
Consequently, our investigation was designed to develop a
general approach to the mycosamine platform (Scheme 1).
The synthesis commences with the conversion of 2-ace-
tylfuran 3 to allylic alcohol 4 in five steps following the
route developed by O’Doherty.15 This resulting alcohol
was engaged in a directed epoxidation reaction and subse-
quently protected to afford epoxide 5. The synthesis of the
mycosamine donor analog was completed by epoxide
opening with azide, acylation of the resultant alcohol,
and generation of the corresponding trichloroacetimidate
(8), whose configuration was established by comparison to
previously synthesized donors.4 The synthesis route is
shorter than any previously described14a and, more im-
portantly, allows for facile manipulation of different
groups of the sugar.
A semisynthetic route that commences with the natural
product was relied upon to access a suitably protected
aglycone (Scheme 2).7b AmB was subjected to amine
(9) Cheron, M.; Cybulska, B.; Mazerski, J.; Grzybowska, J.;
Czerwinski, A.; Borowski, E. Biochem. Pharmacol. 1988, 37, 827.
(10) Paquet, V.; Carreira, E. M. Org. Lett. 2006, 8, 1807.
(11) (a) Baginsky, M.; Resat, H.; Borowski, E. Biochim. Biophys.
Acta 2002, 1567, 63. (b) Baran, M.; Mazerski, J. Biophys. Chem. 2002, 95,
125. (c) Langlet, J.; Berges, J.; Caillet, J.; Demaret, J. P. Biochim.
Biophys. Acta 1994, 1191, 79.
(13) Nicolaou, K. C.; Daines, R. A.; Ogawa, Y.; Chakraborty, T. K.
J. Am. Chem. Soc. 1988, 110, 4696.
(14) (a) Manthorpe, J. M.; Szpilman, A. M.; Carreira, E. M. Synthe-
sis 2005, 3380. (b) Szpilman, A. M.; Carreira, E. M. Org. Lett. 2009, 11,
1305.
(12) Matsumori, N.; Sawada, Y.; Murata, M. J. Am. Chem. Soc.
2005, 127, 10667.
(15) Guo, H.; O’Doherty, G. A. Angew. Chem., Int. Ed. 2007, 46,
5206.
Org. Lett., Vol. 13, No. 6, 2011
1391

Scheme 2. Divergent Synthesis of AmB Analogs
Figure 2. Biological activity of the AmB analogs.
protection (Fmoc); esterification (C16-CO2H f C16-
CO2Me); and methyl ketal formation to yield 9. Subse-
quent silylation of the nine alcohols, oxidative cleavage of
the sugar, and diastereoselective CdO reduction complete
the synthesis of alcohol 10.
Mycosamine donor 8 was coupled to aglycone 10 to
produce a mixture of the glycosylated product (Scheme 2),
along with an orthoester, which is a well-known coproduct
of the glycosidation of the hindered secondary alcohol.
These are conveniently separated by chromatography on
silica gel, following alkaline cleavage of the C20-acetate.
The glycosylated products were subjected to global depro-
tection to provide AmB analog 11 (C20-epi-AmE), incor-
porating the C20-epimer of the mycosamine. Alternatively,
the C20-OH was subjected to O-methylation after acetate
cleavage, which following deprotection provided 12 (C20-
epi-C20-OMe-AmE). Coupling of the aglycone to (()-8
provided access to 13 (C20-epi-ent-mycosamine-AmE),
which is the C20-epi isomer derived from the unnatural
ent-mycosamine.
With analogs 11-13 in hand, their growth inhibition
activity against Saccharomyces cerevisiae (BY4741) was
assayed (Figure 2). The C20-epimer (11) displayed activity
equivalent to AmE 2 (0.30 μM for 11 and 0.25 μM for
AmE). To further understand analog 11, it was subjected
to a liposome assay which we have utilized to measure
potassium efflux. In the assay 11 was indistinguishable
from AmE (2) in the presence of cholesterol and ergosterol
(see Supporting Information). Thus, the configuration at
C20 is not a major determinant for activity in either yeast or
liposomes. This result is significant because it questions the
models in which the C20-alcohol forms a hydrogen bond
with the ergosterol that is critical for channel formation
Figure 3. Models for the association of ergosterol with AmB.
(Figure 3).11,16 It is well worth noting that the only experi-
mental study examining this hypothesis relied on cova-
lently tethered, conformationally restricted derivatives of
AmB whose activities in yeast and liposomes failed to
correlate altogether.12 Additionally, in these models the
calculated minima for the conformationally restricted
analogs do not correspond to the low energy minima
calculated by Borowski.11 In the low energy conformer,
the C(10)-O-C(19)-C(18) dihedral of -60° minimizes
the unfavorable steric interactions between the sugar and
polyene backbone, and the C(20)-C(10)-O-C(19) dihedral
subtends an angle of 180° as a consequence of the exo-
anomeric effect (Figure 3).
Analysis of the model in light of the results with C20-epi-
mycosamine is revealing. The structure illustrated for
mycosamine in Figure 3 corresponds to the lowest energy
conformer. In the model that is commonly invoked, the
axially disposed C20-OH (X in Figure 3) is engaged in
H-bonding to the hydroxyl group of ergosterol. This inter-
action is proposed to bring the steroid in proximity to the
macrolide heptaene, thereby stabilizing the integrity of the
channel. The fact that the C20-epi diastereomer (Y = OH in
Figure 3) displays similar activity cannot be reconciled with
this model, because the corresponding interaction in the
(16) The lack of importance of the C20-OH configuration is consis-
tent with reports in which AmB has been observed to form channels in
sterol-free membranes, a fact that would itself cast doubt on the
necessity for H-bonding as a key stabilizing feature.
1392
Org. Lett., Vol. 13, No. 6, 2011

gauche pentane interactions between C20-OH and C-20 of
the macrocycle, in addition to positioning of the polar amine
into the lipophilic membrane of the putatively formed pore.
In summary, a synthetic route to mycosamine donor 8
providedaccesstoa set of mycosamine derivatives targeted
to probe the effect of C20-OH on the biological properties
of AmB. A key observation in this study is that the
configuration of the C20-OH is inconsequential to the
activity of the polyene macrolide in yeast and liposomes;
however, O-methylation of this alcohol was significantly
deleterious. Additionally, the analog incorporating a my-
cosamine derivative from the enantiomeric series was
devoid of activity. Collectively, the results underscore the
fact that the mycosamine subunit is not merely a polar
group and that the theorized lowest energy conformation
by Borowski is critical for full expression of the biological
activity in yeast. The active conformation is one which
incorporates an exoanomeric stabilization and minimiza-
tion of steric effects between the mycosamine and the
polyene backbone. Further structure-activity relation-
ships involving the mycosamine are currently under in-
vestigation. In general any attempts at exploring
structure-activity relationships ultimately must contend
with the prospect that the fundamental structure has been
altered; in a broader sense the work we have delineated
underscores the possibilities in utilizing stereochemical
probes to gain insight into the biological mode of action
with relatively minimal structural perturbation.
Figure 4. Modeling of C20-epi-analogs 11 and 13 overlaying
C18-20.
C20-epi series would necessarily move the steroid substan-
tively further away.
We next proceeded to examine the C20-methyl ether
derivative (12). It was found to be approximately 50-fold
less active than 2 and 11. The methyl ether group precludes
the possibility for the C20-position to act as a hydrogen-
bond donor altogether. This observation is intriguing in
light of the results indicating the insignificance of the
configuration at C20. We subsequently examined the ana-
log incorporating the mycosamine derivative from the
enantiomeric series. Analog 13 was essentially inactive
(MIC95 > 20 μM). This is interesting because the equator-
ial disposed substituents of the mycosamine in analogs 11
and 13 occupy similar spatial dispositions relative to the
pyran ring (Figure 2). However, further scrutiny reveals
significant differences in the conformational profile of the
two diastereomers. Analysis of models of 11 and 13 suggests
a key difference with respect to the spatial disposition of the
enantiomeric mycosamine derivative relative to the macro-
cycle (Figure 4 and Supporting Information). In the current
study, the lack of activity of 13 when contrasted to 11 is
consistent with conformer 110 as critical for activity. Anal-
ysis of a conformer of 13 that maps onto 110 (see 1300 in
Figure 4) reveals that it suffers from destabilizing double
Acknowledgment. This research was funded by a grant
from the Swiss National Science Foundation. Support in
the form of an ETH Fellowship (M.P.C.) is gratefully
acknowledged.
Supporting Information Available. Experimental pro-
cedures and characterization for all new compounds. This
material is available free of charge via the Internet at
http://pubs.acs.org.
Org. Lett., Vol. 13, No. 6, 2011
1393