Synthesis and conformational analysis of fusidic acid side chain derivatives in relation to antibacterial activity

Article abstract of DOI:10.1021/jm010899a

Novel fusidic acid type antibiotics having flexible side chains are described. Saturation of the double bond between C-17 and C-20 of fusidic acid produces four stereoisomers differing in the configuration at C-17 and C-20. The structure-activity relation

Full text of DOI:10.1021/jm010899a

J . Med. Chem. 2001, 44, 3125-3131
3125
Syn th esis a n d Con for m a tion a l An a lysis of F u sid ic Acid Sid e Ch a in Der iva tives
in Rela tion to An tiba cter ia l Activity
Tore Duvold,* Morten Dahl Sørensen, Fredrik Bjo¨rkling, Anne S. Henriksen, and Niels Rastrup-Andersen
Leo Pharmaceutical Products, Industriparken 55, DK-2750 Ballerup, Denmark
Received April 17, 2001
Novel fusidic acid type antibiotics having flexible side chains are described. Saturation of the
double bond between C-17 and C-20 of fusidic acid produces four stereoisomers differing in
the configuration at C-17 and C-20. The structure-activity relationship of the stereoisomers
was studied using computer-assisted analyses of low-energy conformations and crystallographic
data. Only one of the four stereoisomers showed potent antibiotic activity comparable with
that of fusidic acid, whereas the other three stereoisomers retained little or no activity. The
orientation of the side chain is crucial, and there is only a limited space for bioactive side
chain conformations. This investigation demonstrates the essential role of the side chain
conformations in relation to antibacterial activity and contradicts earlier assumptions that
the ∆17(20) bond is an essential feature in the molecule.
In tr od u ction
Fusidic acid belongs to a family of naturally occurring
antibiotics, the fusidanes, having in common a tetracy-
clic ring system with the unique chair-boat-chair
conformation separating them from steroids (Figure 1).
These compounds also have in common the same
carboxylic acid bearing side chain linked to the ring
system at C-17 via a double bond and an acetate group
at C-16. They show a high degree of antibacterial
activity and have a similar spectrum. Fusidic acid, the
most potent of the fusidanes, was first isolated from
Fusidium coccineum in 19601-3 and has since 1962 been
used clinically in the treatment of both topical and
systemic infections caused by staphylococci. Although
fusidic acid is commonly used against staphylococci, it
is also efficient against several other Gram-positive
species.4 The clinical value of fusidic acid is also due to
its excellent distribution in various tissues, low degree
of toxicity and allergic reactions, and the absence of
cross-resistance with other clinically used antibiotics.5,6
The structure-activity relationship (SAR) of fusidic acid
has been extensively studied (Figure 1),7,8 and a large
number of analogues have been prepared. However, only
a few of these analogues showed activities comparable
with that of fusidic acid. Most of these have a similar
antibacterial spectrum and are cross-resistant. Despite
the extensive SAR studies, the potential of side chain
modifications has not earlier been explored.9 As part of
our renewed interest in improving the antibacterial and
pharmacokinetic properties of fusidic acid type antibiot-
ics, we decided to further investigate the relatively
unexplored side chain. In particular, we wanted to
investigate the importance of the ∆17(20) double bond
which has been assumed to be essential for antibacterial
activity.10,11 In our earlier work, only two of the four
possible epimers of analogues with a single bond
F igu r e 1. An overview of structure-activity relationships of
fusidic acids showing important and essential structural and
functional features.
between C-17 and C-20 have been prepared. These
analogues have in common the 17R configuration but
are epimers at C-20. They retain little or no antibacte-
rial activity. We suggested that the ∆17,20 double bond
is not an essential feature for the biological activity of
the molecule and that the loss of activity of the 17R
analogues previously synthesized is due primarily to an
unfavorable configuration of the side chain. In the
following, we present the synthesis of new 17,20-
dihydrofusidic acid analogues, the elucidation of their
absolute configurations, conformational analyses of the
side chains by means of molecular modeling, and their
in vitro antibacterial activity.
Ch em istr y
The tetrahydrofusidic acid analogues 2 and 3 with the
17R configuration had been prepared previously by
means of catalytic hydrogenation of fusidic acid and its
∆17(20) isomer, lumifusidic acid, successively.12 For the
preparation of the 17S analogues, we took advantage
of the NaBH4 reduction of the corresponding fusidic acid
lactone proceeding solely with the attack from the
R-face12 of the molecule yielding the desired 17S, 20R
* To whom correspondence should be addressed. Tore Duvold,
Department of Medicinal Chemistry, Leo Pharmaceutical Products,
Industriparken 55, DK-2750 Ballerup, Denmark. Tel: +45-44 92 38
00. Fax: +45-44 94 55 10. E-mail: tore.duvold@leo-pharma.com.
10.1021/jm010899a CCC: $20.00 © 2001 American Chemical Society
Published on Web 08/21/2001

3126 J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 19
Duvold et al.
Sch em e 1a
droxy group of lactones 6a and 6b with a tert-butyldi-
methylsilyl (TBDMS) group and reduced the resulting
protected lactones 7a and 7b to the corresponding diols
8a and 8b with LiAlH4. The C-21 primary hydroxy
group of 8a and 8b were selectively protected with the
bulky and moderately stable diphenylmethyl silyl
(DPMS) group, and the 16-hydroxy group was thereafter
acetylated with acetic anhydride and excess of pyridine.
The diphenylmethylsilyl group in 10a and 10b was
cleaved with TBA+F- buffered with acetic acid, and the
resulting primary hydroxy group of 11a and 11b was
oxidized in a two-step manner to the corresponding
carboxylic acid, first to the aldehyde with Dess-Martin
reagent and further to the acid with sodium chlorite.
The TBDMS group of 12a and 12b was finally cleaved
with aqueous HF yielding 17,20-dihydrofusidic acid 4
and 5, respectively.
Str u ctu r e Elu cid a tion . The chemical structure of
compound 4 and 5 could be derived unambiguous from
the synthetic route and from NMR and MS data. Total
1
assignment of the H and 13C data were performed for
key compounds, and in addition, the stereochemistry in
positions 17 and 20 could be derived from NOESY
experiments (Figure 5). In 6a (20R) and 6b (20S), H-17
clearly showed NOEs to H-13 as well as H-16 proving
the configuration of position 17 to be 17S. A strong NOE
between H-20 and H-17 was observed in 6a . In contrast
to this observation, no NOE was observed between these
two protons in 6b. Instead, a strong NOE between H-20
and the H-18 protons was observed. These observations
proved the stereochemistry in position 20 to be 20R for
6a and 20S for 6b, respectively.
Molecu la r Mod elin g. Conformational analysis of the
four side chain analogues 2-5 were carried out using
computer-assisted molecular modeling and compared
with fusidic acid. The conformation of the fusidane ring
system was in all calculations kept in the conformation
found in the crystal structure of fusidic acid methyl ester
3-p-bromobenzoate13 and was assumed to be relatively
unaffected by side chain modifications. This assumption
was later confirmed by the obtained crystal structure
of 17S,20S-dihydrofusidic acid, 5.14 Thus, in the calcula-
tions of low-energy conformations, only the side chain
and the C-16 acetoxy group were allowed to change. The
schematic overview of fusidic acid SAR previously
established (Figure 1) explains that the carboxyl group
in the side chain is essential for activity, the C-16
acetoxy group is important but can be exchanged with
other functional groups, and the ∆24(25) bond is of little
or no importance. Superposition of the lowest energy
conformations (global minimum) of fusidic acid, 17R,-
20S-tetrahydrofusidic acid (2), 17R,20R-tetrahydro-
fusidic acid (3), 17S,20R-dihydrofusidic acid (4), and
17S,20S-dihydrofusidic acid (5) is shown in Figure 4.
As suspected, the C-16 acetoxy group is located in
approximately the same position in all five structures.
It should be noted, however, that the position of the
acetate group is slightly different in the 17R compounds
2 and 3 as compared to fusidic acid and the 17S
compounds (Figure 1 and Table 1). In contrast, the
positions of the carboxyl group in the global minimum
conformations of 3 and 4 are very different from the
position found for fusidic acid, whereas the position of
the carboxyl group of 2 is more similar to that of fusidic
a
(a) (i) aq. NaOH in EtOH, reflux, (96%), (ii) NaBH4 in MeOH/
water, rt, 3 h, (94%); (b) aq. 28% NaOH, EtOH, reflux, 1 h, (93%);
(c) TBDMSCl, imidazole in DMF, rt, overnight, (96%); (d) LiAlH4,
THF, reflux, (quant); (e) DPMSCl, Et3N, CH2Cl2, -20 °C, (quant);
(f) Ac2O/pyridine, (90%); (g) TBA+F-, AcOH, THF, (90%); (h) (i)
Dess-Martin periodinane, CH2Cl2, 0 °C, 3 h, (quant); (ii) NaClO2,
t-BuOH, (81%); (i) aq. HF in THF, rt, 24 h (87%).
stereoisomer (Scheme 1). Inversion of the C-20 config-
uration was obtained by refluxing the saturated lactone
6a in the presence of concentrated aqueous NaOH. This
resulted in quantitative C-20 epimerization to lactone
6b with the 17S,20S configuration. Synthesis of 17S,-
20R- and 17S,20S-dihydrofusidic acid were then carried
out starting from lactones 6a and 6b, respectively. We
were unable to isolate the open form of the saturated
lactones, either as the corresponding free acids or
trapped in situ as corresponding carboxylic esters.
Instead, we employed a stepwise synthetic strategy to
restore the free C-21 carboxyl group and the C-16
acetoxy group (Scheme 1). We first protected the 3-hy-

Conformational Analysis of Fusidic Acid Derivatives
J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 19 3127
Ta ble 1. Comparison of the Positions of the C16 Acetoxy and
COOH Groups in the Four Side Chain Analogues 2-5 and
Fusidic Acid
acetoxy
COOH
COOH
compound
RMS (Å)a
RMS (Å)a
plane angleb
fusidic acid (1)
0.35 ( 0.06
0.89 ( 0.09
0.87 ( 0.26
0.25 ( 0.10
0.18 ( 0.05
0.37 ( 0.16
2.39 ( 1.30
3.61 ( 0.81
2.85 ( 1.00
1.10 ( 0.65
11.1 ( 9.5
18.9 ( 7.0
19.1 ( 11.0
61.6 ( 1.7
20.1 ( 7.4
2
3
4
5
a
The C16 acetoxy and COOH RMS values are measured with
respect to the crystal structure of fusidic acid methyl ester 3-p-
bromobenzoate. All conformations within 3 kcal/mol of the global
b
minimum were included. The plane angle of the COOH group
was only calculated for those conformations in the 0-3 kcal/mol
energy window which have a COOH RMS less than 2 Å. The plane
angle is measured with respect to the global minimum conforma-
tion of fusidic acid (see text). The carbon atoms of the cyclohexyl
rings of the tetracyclic ring system was used for superposition of
the modeled conformations to the crystral structure of fusidic acid
methyl ester 3-p-bromobenzoate prior to all calculations.
F igu r e 3. Superposition of the global minimum conformations
of fusidic acid and the four side chain analogues on the crystal
structure of fusidic acid methyl ester 3-p-bromobenzoate. The
carbon atoms of the cyclohexyl rings of the tetracyclic ring
system were superimposed. The crystal structure is shown in
atom colors, 1 is shown in green, 2 is shown in yellow, 3 is
shown in cyan, 4 shown in orange, and 5 is shown in magenta.
The lipophilic part of the side chain of all conformations has
been omitted for clarity. The 3-p-bromobenzoate moiety of the
crystal structure has also been omitted.
value of the carboxyl group is less than 2 Å. To further
analyze these conformations, we determined the angle
between the plane defined by the carboxyl group of the
conformation in question and the plane defined by the
carboxyl group of the global minimum conformation of
fusidic acid. The crystal structure of fusidic acid methyl
ester 3-p-bromobenzoate was not used for this analysis
because it contains a methyl ester instead of a carboxyl
group. The extra methyl probably affects the plane angle
slightly. The results show that the carboxyl groups of
fusidic acid (1) and compounds 2, 3, and 5 are presented
in a manner similar to the interacting elongation factor
G (EF-G) receptor,11,15 whereas the carboxyl group of 4
is presented significantly different (Table 1). Finally, we
wanted to examine the position of the lipophilic moiety
of the side chain. The position of this moiety of the side
chain in all low energy conformations where the RMS
value of the carboxyl group is less than 2 Å is shown in
Figure 4. Clearly, only for compound 5 does the lipo-
philic moiety occupy the same conformational space as
fusidic acid. For compound 3, the side chain can also
attain conformations similar to that of fusidic acid.
Discu ssion
Extensive SAR work has been made on fusidic acid,
including a limited number of side chain modifications.
In this early work, only two of four possible isomers of
tetrahydrofusidic acid were obtained by means of cata-
lytic hydrogenation of fusidic acid and its ∆17(20)
isomer, lumifusidic acid. However, these analogues have
no or little antibiotic activity. It was assumed that the
altered conformations, resulting from the more flexible
saturated side chain and the absence of a conjugated
carboxylic acid, were responsible for the loss of activity.
In our present conformational analysis of tetrahydro-
fusidic acid and dihydrofusidic acid stereoisomers, we
found a common orientation of the carboxyl group in
compounds 2 (17R,20S), 3 (17R,20R), and 5 (17S,20S)
in comparison with fusidic acid. However, only in
compound 5 do both the carboxyl group and the lipo-
philic moiety of the side chain cover the conformational
F igu r e 2. Fusidic acid and derivatives.
acid, and a high resemblance to fusidic acid is observed
for 5 (Figure 3). A comparison of the positions of the
carboxyl group and the corresponding group in the
crystal structure of fusidic acid methyl ester 3-p-
bromobenzoate was carried out for all conformations
within 3 kcal/mol of the global minimum of compounds
1-5. The RMS values obtained from this analysis are
listed in Table 1. It is obvious when looking at all
conformations in this energy window that the position
of the carboxyl group in fusidic acid is generally best
emulated in the conformations found for compound 5.
However, all four analogues have several conformations
within 3 kcal/mol of the global minimum where the RMS

3128 J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 19
Duvold et al.
F igu r e 5. Assignment of stereochemistry by nuclear over-
hauser effect obtained in 2D NOESY and 2D TROESY experi-
ments. s, NOESY cross-peak with high intensity; m, NOESY
cross-peak with medium intensity.
terial activities of the four fusidic acid analogues. The
inactivity of compound 4 is most likely due to different
orientation of the carboxyl group. Furthermore, it is
interesting to note that the lipophilic moiety of the side
chain is positioned significantly different in compound
2 as compared to fusidic acid and compound 5 (Figure
4). Thus, the low activity of compound 2 as compared
to the high activity of both fusidic acid and compound
5 suggests that the lipophilic moiety of the side chain
is positioned above the ring plane in a bioactive confor-
mation. In conclusion, the correlation between low
energy conformations and antibacterial activity dem-
onstrates that the double bond between C-17 and C-20
is not essential for the antibacterial activity of the
molecule but rather that the carboxyl group and the
lipophilic moiety of the side chain must be oriented in
a constrained conformational space similar to that of
fusidic acid. Only one of the side chain analogues, with
the best conformational fit to fusidic acid, 17S,20S-
dihydrofusidic acid (5), was found equipotent to fusidic
acid. However, it is not known whether the side chains
of 5 and fusidic acid have the optimal bioactive confor-
mations. These studies demonstrate the importance of
both the carboxyl group and the lipophilic moiety for
the antibacterial activity of fusidic acid. Further inves-
tigations for the search of more active side chains are
in progress.
F igu r e 4. Superposition of all conformations within 3 kcal/
mol of the global minimum where the RMS value of the
carboxyl group is less than 2 Å (see text). Only the lipophilic
part of the side chain is shown. The color scheme and atoms
superimposed are the same as in Figure 3.
space found for fusidic acid (Figure 4). The carboxyl
group in low-energy conformations of compound 4 (17S,-
20R) could not be oriented similarly to fusidic acid.
Therefore, the importance of a ∆17(20) double bond for
antibacterial activity was further investigated. The two
isomers, 17R,20R and 17R,20S, previously prepared,
have in common the 17R configuration and are epimers
at C-20. We were now able to prepare the remaining
isomers, compounds 4 and 5, by means of NaBH4
reduction of fusidic acid lactone yielding the saturated
17S,20R lactone (6a ). The corresponding 20S epimer
(6b) was obtained by base catalyzed epimerization of
lactone 6a . The C-21 carboxyl group and the C-16
acetoxy group could then be reintroduced in a stepwise
manner.
In good correlation with our conformational modeling
studies, the antibacterial assays showed potent activity
of compound 5 and no activity of compound 4 (Table 2).
We could confirm previous reports showing low activity
for compound 2. The antibacterial spectrum and potency
of compound 5 was found to be similar to fusidic acid.
Our conformational modeling studies provide a rational
explanation for the observed differences in the antibac-
Exp er im en ta l Section
Molecu la r Mod elin g. All calculations were performed on
a Silicon Graphics O2 R10000 workstation. The conformational
analyses of the C-16 acetoxy and the C-17 side chains of the
fusidic acid analogues were carried out using the Monte Carlo
(Mcrlo) routine of MacroModel 7.0 (Schro¨dinger Inc.). The
conformation of the tetracyclic ring system was kept fixed in
the conformation observed in the crystal structure of fusidic
acid methyl ester 3-p-brombenzoate.13 It should be noted that
the torsion angles C5-C6-C7-C8 and C5-C10-C9-C8
describing the conformation of the boat-shaped B-ring are -30°
and -4°, respectively. These two torsion angles were con-
strained using the FXTA command, the force constant being
the default 1000 kJ /mol. The structures in the conformational
analyses were energy-minimized with the truncated Newton

Conformational Analysis of Fusidic Acid Derivatives
J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 19 3129
MIC (µg/mL)
Ta ble 2. Antimicrobial Activity of Fusidic Acid and Derivativesa
organism/strain
fusidic acid (1)
2
4
5
Staphylococcus aureus ATCC 2977
Staphylococcus aureus CJ 232 (MRSA)
Staphylococcus aureus CJ 234 (R) (MRSA)
Staphylococcus aureus CJ 234 (F) (MRSA)
Staphylococcus aureus ATCC 6538P
Staphylococcus aureus ATCC 29213
Staphylococcus epidermidis ATCC 12228
Propionibacterium acnes NCTC 737
Corynebacterium xerosis NCTC 9755
Streptococcus faecium EI 119 (P)
Streptococcus sp. EF 6
0.006 (0.0007-0.05)
4
>125
>125
>125
>125
>125
>125
>125
4
>125
>125
>125
>125
>125
>125
>125
>125
>125
>125
>125
>125
>125
64
0.03 (0.008-0.128)
0.015 (0.005-0.05)
4 (2.5-5.5)
0.9 (0.4-2)
>125
4
2 (0.6-7)
3.8 (2.8-5)
4
0.03 (0.01-0.06)
0.006 (0.002-0.017)
0.03 (0.02-0.04)
>125
>125
0.007 (0.001-0.036)
0.012 (0.005-0.03)
0.007 (0.003-0.016)
0.01 (0.007-0.02)
0.0007 (0.00004-0.01)
0.01 (0.003-0.03)
0.003 (0.0009-0.007)
0.03 (0.015-0.05)
0.01 (0.001-0.1)
4
16
0.03 (0.007-0.1)
0.01 (0.001-0.1)
0.025 (0.008-0.08)
4
4
4
16
4
16
>125
>125
>125
>125
>125
>125
>125
>125
>125
>125
>125
16
4
16
16
16
4
16
Streptococcus pyogenes EC
Streptococcus thermophilus EG 5
Streptococcus zooepidermidis ED (gr.C)
Streptococcus salivarius EG 7 (gr.A)
Clostridium perfringens KT 13
Micrococcus luteus ATCC 9341
Bacillus cereus ATCC 10876
Escherichia coli HA 44
Saccharomyces cerevisiae ZZ 7
Candida albicans ATCC 10231
Aspergillus niger ATCC 16404
1 (0.3-3)
4
>125
4
16
16
1 (0.4-2.5)
1
>125
4
64
64
a
Numbers in brackets represent the concentration interval containing the real MIC value with 95% confidence. MRSA ) meticilline
resistant S. aureus, R ) rifampicin resistant, P ) penicillin resistant, F ) fus. resistant.
conjugate gradient (TNGG) method using MMFF94s force
field, until the default derivative convergence criterion of 0.05
kJ /mol/Å were met. All structures converged within 300 cycles
of minimization. The solvent was set to water using the
SLVNT command. The resulting structures were visualized
with Sybyl 6.6 (Tripos Inc.).
All solvents and reagents were of highest available quality
and used as such. Reactions were monitored by TLC analyses
using 0.25-mm glass-coated silica plates (E. Merck 60 F254).
Chromatography was performed on silica gel 60, 230-400
mesh (E. Merck) using mixtures of ethyl acetate and low
boiling petroleum ether as eluant. For chromatography of
compounds having a carboxylic acid group, the eluants con-
tained 0.5-1% of formic acid. Anhydrous solvents were
prepared by storing analytical grade solvents over 4 Å molec-
ular sieves a few days prior to use. The water content was
measured before use on a Carl Fisher apparatus (typical water
content: 5-12 ppm for THF, dichloromethan and 20-40 ppm
for DMF).
17S,20R-Dih yd r ofu sid ic Acid La cton e (6a ). Fusidic acid
(6 g, 11.3 mmol) was dissolved in EtOH (90 mL) and 2 N NaOH
(30 mL) was added. The resulting reaction mixture was
refluxed for 1 h, cooled to room temperature, and acidified with
diluted HCl causing crystallization. The crystals were collected
by filtration and recrystallized from MeOH to yield 5.1 g (96%)
of colorless crystals, mp 154-155 °C (Lit. 158.5-159.5 °C).17
MS Calc. for C29H44O4 m/z 456.3240, observed m/z 456.3238.
Anal. (C29H44O4) C, H.
(ii) 16-Deacetylfusidic acid lactone from (i) (5.0 g, 10.9 mmol)
was dissolved in MeOH (100 mL), and a solution of NaBH4
(0.83 g, 22.0 mmol) in water (20 mL) was added dropwise. The
resulting reaction mixture was vigorously stirred for 2 h at
room temperature. The reaction was quenched by diluted HCl
and concentrated under reduced pressure. The mixture was
suspended in water and extracted with EtOAc (2 × 100 mL).
The combined organic extracts were washed with brine (2 ×
50 mL), dried (Na2SO4), and concentrated to a white powder
which was recrystallized from MeOH to yield 4.7 g (94%) of
colorless crystals, mp 160-161 °C. (Lit. 168-171 °C)12
1H NMR (CDCl3): 0.90 (s, 3H), 0.92 (d, J ) 6.9 Hz, 3H),
0.96 (s, 3H), 1.34 (s, 3H), 1.63 (bs, 3H), 1.70 (bs, 3H), 2.66 (m,
1H), 2.79 (m, 1H), 3.15 (m, 1H), 3.76 (m, 1H), 4.30 (m, 1H),
4.99 (dd, 1H), 5.09 (bt, 1H). MS (EI+): m/z 458, 376, 389.
17S,20S-Dih yd r ofu sid ic Acid La cton e (6b). Lactone 3
(2 g, 4.4 mmol) was dissolved in ethanol (10 mL) and 28%
aqueous NaOH (5 mL). The resulting yellow solution was
heated at 60 °C for 1 h. The reaction mixture was allowed to
cool, and the mixture was acidified to pH 4 with concentrated
acetic acid resulting an almost colorless solution. Water was
added under continuous stirring until precipitation of colorless
crystals. Crystals were collected by filtration yielding 1.95 g
of 16-deacetyl-17S,20S-dihydrofusidic acid lactone (6b). Re-
An tim icr obia l Activity. Minimum inhibitory concentra-
tions (MICs) were estimated using an agar cup assay.16
Bacterial strains were obtained from the American Type
Culture Collection or from our own collection of clinical
isolates. Colonies from fresh overnight culture were resus-
pended in saline water to 0.5 MacFarland corresponding to
108 CFU/mL. A total of 200 mL Mueller Hinton agar (Oxoid)
at 48 °C was inoculated at a concentration of 106 CFU/mL and
poured into square Petri dishes (245 × 245 mm). Holes were
made in the inoculated plates and 200 µL of each sample was
disposed into each hole. Dilution series for fusidic acid and
fusidic acid analogues contained six dilutions between 0.25 and
125 µg/mL. For Streptococci, Mueller Hinton agar was supple-
mented with 5% sheep blood. Plates were appropriately
incubated and zone diameters of growth inhibition were
measured using an electronic caliper. MICs were estimated
using a linear regression curve between the zone diameter of
growth inhibition and the log2 of the sample concentration. A
95% confidence interval was estimated when possible using
Statgraphics Softwear -Plus for Windows 4.1, (Statistical
Graphics Corp.).
Gen er a l Exp er im en ta l P r oced u r es. Melting points were
measured on a Bu¨chi 535 apparatus and are uncorrected.
NMR spectra were recorded at 300° K on either a Bruker
ARX300 or a Bruker DRX500 spectrometer equipped with a 5
mm qnp and a 5 mm broadband inverse probe, respectively.
In most cases, CDCl3 was used as solvent. All chemical shifts
are given in ppm δ scale using either tetramethylsilane
(unsilylated compounds, TMS δ ) 0.00 ppm), or chloroform
(silylated compounds 1H δ ) 7.25 ppm, 13C δ ) 76.81 ppm,
respectively) as internal reference.
1
Conventional H, 13C, and DEPT135 spectra were obtained
on all compounds. In addition HMQC, HMBC, COSY, HH-
TOCSY, and CHTOCSY experiments were performed on key
compounds to make total assignments.
The stereochemistry was elucidated by comparing NOESY
and TROESY experiments. Mass spectra were recorded on
either a Micromass LC-QuattroII or a high-resolution Micro-
mass AutoSpec sector instrument. Elementary analyses were
obtained on a home built combustion equipment.

3130 J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 19
Duvold et al.
crystallization from methanol-water afforded 1.85 g (93%),
mp 167-169 °C.
(50 mL), water (50 mL), and brine (50 mL). The organic
solution was dried (Na2SO4) and solvents were evaporated
under reduced pressure yielding 1.96 g (quantitative) of a
colorless foam of 3-O-TBDMS-21-O-diphenylmethylsilyl-17S,-
20R-methanofusidin-3,11,16,21-tetrol (9a ).
NMR (CDCl3): 0.01 (s, 3H), 0.02 (s, 3H), 0.67 (s, 3H), 0.80
(d J ) 6.9 Hz, 3H), 0.89 (s, 9H), 0.93 (s, 3H), 1.12 (s, 3H), 1.31
(s, 3H), 1.49 (bs, 3H), 1.65 (bs, 3H), 2.38 (m, 1H), 2.64 (m, 1H),
3.41 (bd, 1H), 3.67 (m, 1H), 3.69 (m, 1H), 3.93 (dd, 1H), 4.26
(m, 1H), 4.53 (dt, 1H), 5.04 (bt, 1H), 7.32-7.46 (m, 6H), 7.54-
7.62 (m, 4H).
3-O-TBDMS-21-O-Dip h en ylm et h ylsilyl-17S,20S-d ih y-
d r ofu sid in -3,11,16,21-tetr ol (9b). The reaction was carried
out as described for the preparation of 9a . NMR (CDCl3): 0.01
(s, 3H), 0.02 (s, 3H), 0.66 (s, 3H), 0.81 (d J ) 6.9 Hz, 3H), 0.90
(s, 9H), 0.93 (s, 3H), 1.05 (s, 3H), 1.33 (s, 3H), 1.50 (bs, 3H),
1.64 (bs, 3H), 2.57 (m, 1H) 3.61 (dd, 1H), 3.68 (bs, 1H), 3.86
(dd, 1H), 3.98 (d, 1H), 4.25 (bs, 1H), 4.40 (m, 1H), 5.00 (bt,
1H), 7.39 (m, 6H), 7.57 (m, 4H).
3-O-TBDMS-21-O-Diph en ylm eth ylsilyl-16â-acetoxy-17S,-
20R-d ih yd r ofu sid in -3,11,16,21-tetr ol (10a ). Compound 9a
acetylated by dissolving in pyridine (10 mL) and acetic
anhydride (5 mL). The resulting mixture was stirred overnight
at room temperature in a stoppered bottle. After this time,
the reaction mixture was concentrated under reduced pressure
yielding a pale yellow oil. Essentially pure title compound
(10a ), 1.85 g (90%), was obtained as a white foam after column
chromatography using a mixture of EtOAc and low boiling
petroleum ether as eluant.
1H NMR (CDCl3): 0.88 (s, 3H), 0.92 (d J ) 6.9 Hz, 3H), 0.96
(s, 3H), 1.36 (s, 3H), 1.61 (bs, 3H), 1.69 (bs, 3H), 2.52 (m, 1H),
2.65 (m, 1H), 2.70 (m, 1H), 3.75 (m, 1H), 4.36 (m, 1H), 5.00 (t,
1H), 5.08 (bt, 1H). MS: Calc. for C29H46O4 m/z 458.3396,
observed: m/z 458.3396. Anal. Calcd for C29H46O4: C, 75.94;
H, 10.11. Found: C, 75.38; H, 10.11.
3-O-TBDMS-17S,20R-Dih ydr ofu sidic Acid Lacton e (7a).
16-Deacetyl-17S,20R-dihydrofusidic acid lactone (6a ) (2.0 g,
4.4 mmol) was dissolved in anhydrous DMF (10 mL). To the
solution was added imidazole (0.6 mg, 8.8 mmol) and TBDM-
SCl (1.3 g, 8.8 mmol), and the reaction mixture was stirred
overnight under an atmosphere of argon. Water (50 mL) was
added to the reaction mixture followed by extraction with
EtOAc (2 × 50 mL), and the combined organic layers were
washed successively with water and brine. The organic layer
was dried (Na2SO4) and concentrated under reduced pressure
resulting in a colorless solid. Recrystallization from methanol
yielded 2.4 g (96%) of 7a as a colorless powder, melting point
138.5-140 °C.
NMR (CDCl3): 0.01 (s,3H), 0.02 (s, 3H), 0.79 (d J ) 6.9 Hz,
3H), 0.88 (s, 3H), 0.89 (s, 9H), 0.93 (s, 3H), 1.31 (s, 3H), 1.62
(bs, 3H), 1.69 (bs, 3H), 2.64 (m, 1H), 2.78 (m, 1H), 3.13 (m,
1H), 3.69 (m, 1H), 4.27 (m, 1H), 4.97 (dd, 1H), 5.08 (m, 1H).
MS: (EI+) m/z 572, 497, 423.
3-O-TBDMS-17S,20S-Dih ydr ofu sidic Acid Lacton e (7b).
The reaction was carried out as described for the preparation
1
of 7a . Mp 163-163.5 °C. H NMR (CDCl3): 0.01 (s, 3H), 0.02
NMR (CDCl3): 0.065 (s, 3H), 0.023 (s, 3H), 0.63 (s, 3H), 0.79
(d J ) 6.9 Hz), 0.89 (s, 9H), 0.92 (s, 3H), 0.94 (s, 3H), 1.29 (s,
3H), 1.50 (bs, 3H), 1.64 (bs, 3H), 1.96 (s, 3H), 2.19 (m,2H), 2.57
(m, 2H), 3.68 (bs, 1H), 3.73 (m, 2H), 4.15 (bs, 1H), 4.97 (bt,
1H), 5.34 (bt, 1H), 7.37 (m, 6H), 7.56 (m, 4H). MS (EI+) m/z
814, 736, 522.
3-O-TBDMS-21-O-Diph en ylm eth ylsilyl-16â-acetoxy-17S,-
20S-d ih yd r ofu sid in -3,11,16,21-tetr ol (10b). The reaction
was carried out as described for the preparation of 10a . NMR
(CDCl3): 0.005 (s,3H), 0.022 (s, 3H), 0.61 (s, 3H), 0.79 (d J )
6.5 Hz, 3H), 0.90 (s, 9H), 0.93 (s, 3H), 1.29 (s, 3H), 1.56 (bs,
3H), 1.68 (bs, 3H), 1,80 (s, 3H), 2.17 (m, 2H), 2.57 (m, 2H),
3.66 (dd, 1H), 3.69 (bs, 1H), 3.70 (dd, 1H), 4.20 (bs, 1H), 5.08
(bt, 1H), 5,25 (bt, 1H), 7.56 (m, 4H), 7.36 (m, 6H).
3-O-TBDMS-16â-Acet oxy-17S,20R-d ih yd r ofu sid in -3,-
11,16,21-tetr ol (11a ). Compound 10a (1.9 g, 2.3 mmol) was
dissolved in tetrahydrofuran (30 mL) and glacial acetic acid
(0.25 mL). To this solution was added TBA+F- (1.17 g, 4.6
mmol), and the reaction mixture was stirred at room temper-
ature for 15 min. The reaction mixture was diluted with EtOAc
(100 mL) and the organic solution was washed with water (2
× 25 mL) and brine (2 × 25 mL). The organic solution was
dried (Na2SO4) and concentrated under reduced pressure
yielding a colorless syrup. Pure title compound (11a ), 1.3 g
(90%), was obtained as a colorless foam after column chroma-
tography using a mixture of EtOAc and low boiling petroleum
ether as eluant.
(s, 3H), 0.80 (d J ) 6.5 Hz, 3H), 0.87 (s, 3H), 0.89 (s, 9H), 0.93
(s, 3H), 1.33 (s, 3H), 1.59 (bs, 3H), 1.67 (bs, 3H), 2.52 (m, 1H),
2.61 (m, 1H), 2.67 (m, 1H), 3.69 (m, 1H), 4.34 (m, 1H), 4.99 (t,
1H), 5.07 (m, 1H).
3-O -T B D M S -17S ,20R -D i h y d r o fu s i d i n -3,11,16,21-
tetr ol (8a ). Lithium aluminum hydride (0.4 g, 100 mmol) was
suspended in anhydrous THF (30 mL) under argon in an oven-
dried two-necked round-bottom flask fitted with a condenser.
To the stirred suspension was added a solution of lactone 7a
(1.75 g, 3.0 mmol) in anhydrous THF (10 mL) in such a rate
causing gentle reflux. The reaction mixture was refluxed under
vigorous stirring for 3 h and then allowed to attain room
temperature. Excess lithium aluminum hydride was destroyed
with EtOAc, and water was then added slowly. The resulting
suspension was acidified with diluted hydrochloric acid to pH
5. The mixture was extracted with EtOAc (3 × 50 mL), and
the combined organic layers were washed with brine (2 × 50
mL). The organic layer was dried (Na2SO4) and concentrated
under reduced pressure yielding 1.76 g (quantitative) of
essentially pure title compound diol (5) as a colorless white
powder. An analytically pure sample was obtained by recrys-
tallization from methanol, mp 147-150 °C.
NMR (CDCl3): 0.018 (s, 3H), 0.002 (s,3H), 0.80 (d J ) 6.8
Hz, 3H), 0.89 (s, 9H), 0.94 (s, 3H), 1.11 (s, 3H), 1.31 (s, 3H),
1.60 (bs, 3H), 1.67 (bs, 3H), 2.42 (m, 1H), 2.64 (m, 1H), 3.67
(m, 1H), 3.68 (m, 1H), 3.92 (dd, 1H), 4.29 (m, 1H), 4.53 (bt,
1H), 5.12 (bt, 1H). MS: (EI+) m/z 540, 522, 409.
NMR (C6D6): 0.08 (s, 3H), 0.09 (s, 3H), 0.91 (d J ) 6 Hz,
3H), 0.91 (s, 3H), 1.03 (s, 3H), 1.06 (s, 9H), 1.39 (s, 3H), 1.62
(bs, 3H), 1.68 (bs, 3H), 1.79 (s, 3H), 3.55 (m, 1H), 3.57 (dd,
1H), 3.66 (dd, 1H), 4.08 (m, 1H), 5.25 (bt, 1H), 5.52 (t, 1H).
3-O-TBDMS-16â-Acetoxy-17S,20S-d ih yd r ofu sid in -3,11,-
16,21-tetr ol (11b). The reaction was carried out as described
for the preparation of 11a . NMR (CDCl3): 0.001 (s, 3H), 0.017
(s, 3H), 0.79 (d J ) 6.5 Hz, 3H), 0.89 (s, 9H), 0.93 (s, 3H), 1.00-
(s, 3H), 1.32 (s,3H), 1.60 (bs, 3H), 1.67 (bs, 3H), 2.02 (s, 3H),
2.54 (m, 1H), 2.66 (m, 1H), 3.59 (dd, 1H), 3.66 (bs, 1H), 3.67
(dd, 1H), 4.30 (m, 1H), 5.09 (bt, 1H), 5.33 (dt, 1H).
3-O-TBDMS-17S,20R-Dih yd r ofu sid ic Acid (12a ). (i)
Dess-Martin periodinane (0.89 g, 2.1 mmol) was added
portionwise to a solution of compound 11a (1.26 g, 2.0 mmol)
in anhydrous THF (20 mL) under argon at 0 °C. The resulting
reaction mixture was stirred for 3 h at 0 °C. The reaction was
stopped by adding a solution of saturated NaHCO3 (50 mL)
and 1 N sodium thiosulfate (50 mL), and the resulting two
3-O-TBDMS-17S,20S-Dih yd r ofu sid in -3,11,16,21-tetr ol
(8b). The reaction was carried out as described for the
preparation of 8a . Mp 122-124 °C. NMR (CDCl3): 0.004 (s,
3H), 0.02 (s, 3H), 0.80 (d J ) 6.8 Hz, 3H), 0.89 (s, 9H), 0.93 (s,
3H), 1.60 (bs, 3H), 1.67 (bs, 3H), 2.61 (m, 1H), 3.59 (dd, 1H),
3.68 (m, 1H), 3.74 (dd, 1H) 4.29 (m, 1H), 4.48 (dt, 1H), 5.09
(bt, 1H).
3-O-TBDMS-21-O-Dip h en ylm et h ylsilyl-17S,20R-d ih y-
d r ofu sid in -3,11,16,21-tetr ol (9a ). Diol (8a ) (1.5 g, 2.5 mmol)
was dissolved in anhydrous dichloromethane (20 mL) and
triethylamine (0.7 mL, 5 mmol) under argon and cooled at -25
°C. To the cooled solution was added over a period of 15 min
a solution of diphenylmethylchlorosilane (0.57 mL, 2.75 mmol)
in anhydrous dichloromethane (5 mL) so that the temperature
did not exceed -20 °C and stirring was continued for 15 min.
Water (50 mL) was added to the reaction mixture followed by
extraction with EtOAc (2 × 100 mL). The combined extracts
were washed successively with a solution of saturated NaHCO3

Conformational Analysis of Fusidic Acid Derivatives
J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 19 3131
layers were vigorously stirred for 30 min. The mixture was
extracted with EtOAc (2 × 100 mL), and the combined organic
extracts were washed with saturated NaHCO3 (50 mL) and
brine (50 mL). The organic solution was dried (Na2SO4) and
concentrated under reduced pressure yielding 1.25 g (quanti-
tative) of a colorless syrup.
NMR (CDCl3): 0.0002 (s, 3H), 0.012 (s, 3H), 0.79 (d J ) 6.5
Hz, 3H), 0.88 (s, 9H), 0.90 (s, 3H), 1.28 (s, 3H), 1.55 (bs, 3H),
1.65 (bs, 3H), 2.03 (s, 3H), 2.53 (m, 1H), 2.67 (m, 1H), 2.69 (m,
1H),3.67 (m, 1H), 4.25 (bs, 1H), 4.99 (bt, 1H), 5.43 (t, 1H), 9.50
(d, J ) 5.0, 1H).
17S,20S-Dih yd r ofu sid ic Acid La cton e (6b). Anal. (C29
46O4) C,H: C: calcd. 75.94; found 75.38; H: calcd. 10.11;
found 10.11.
-
H
3-O-TBDMS-17S,20R-Dih yd r ofu sid ic Acid (12a ). Anal.
(C43H74O8Si) C,H; C: Calcd. 70.21; found, 70.05; H: calcd.
10.19; found, 10.19.
17S,20R-Dih yd r ofu sid ic Acid (4). Anal. (C31H50O6, 2H2O)
C,H; C: Calcd. 67.12; found, 67.35; H: calcd. 9.81; found, 9.52.
17S,20S-Dih yd r ofu sid ic Acid (5). Anal. (C31H50O6, 2H2O)
C,H; C: Calcd. 71.78; found 70.90; H: calcd. 9.72; found, 9.66.
(ii) Aldehyde (1.25 g, 2.0 mmol) from (i) was used without
purification dissolved in tert-butanol (15 mL). To this solution
was added 2-methyl-2-butene (0.48 mL, 5.4 mmol), 1 N sodium
dihydrogenphosphate (5.5), and sodium chlorite (0.46 g, 5.0
mmol) in water (20 mL), and the resulting reaction mixture
was stirred vigorously overnight at room temperature. The
reaction mixture was acidified to pH 4 with acetic acid and
transferred to a separatory funnel with EtOAc. The two layers
were shaken and separated. The aqueous layer was re-
extracted twice with EtOAc. The combined organic extracts
were washed twice with brine, dried (Na2SO4), and concen-
trated under reduced pressure yielding 1.2 g of a pale yellow
foam. Purification by column chromatography using a mixture
of EtOAc, low boiling petroleum ether, and a trace of formic
acid as eluant yielded 1.05 g (81% from 11a) of pure acid 12a ,
the title compound, as a semicrystalline compound.
NMR (CDCl3): 0.002 (s, 3H), 0.007 (s, 3H), 0.79 (d J ) 6.9
Hz), 0.88 (s, 9H), 0.91 (s, 3H), 1.003 (s, 3H), 1.28 (s, 3H), 1.56
(bs, 3H), 1.65 (bs, 3H), 1.01 (s, 3H), 2.47 (m, 1H), 2,66 (m, 1H),
2.71 (m, 1H), 3.67 (bs, 1H), 4.28 (bs, 1H), 5,05 (bt, 1H), 5.39
(bt, 1H). MS (ES-): 631. Anal. (C43H74O8Si) C, H.
3-O-TBDMS-17S,20S-Dih yd r ofu sid ic Acid (12b). The
reaction was carried out as described for the preparation of
12a . Mp 210-212 °C. NMR (CDCl3): 0.007 (s, 3H), 0.009 (s,
3H), 079 (d J ) 6.8 Hz, 3H), 0.88(s, 9H), 0.93 (s, 3H), 0.98 (s,
3H), 1.29 (s, 3H), 1.57 (bs, 3H), 1.66 (s, 3H), 1.92 (s, 3H), 3.67
(bs, 1H), 4.31 (bs, 1H), 5.05 (bt, 1H), 5.30 (bt, 1H).
17S,20R-Dih yd r ofu sid ic Acid (4). 3-O-TBS-17S,20R-di-
hydrofusidic acid (12a ) (1.0 g, 1.55 mmol) was dissolved in THF
(12 mL) and 40% aqueous HF (3 mL) in a round-bottom Teflon
flask, and the resulting mixture was stirred vigorously at room
temperature for 24 h. After this time, water (25 mL) was added
and the pH of the mixture was adjusted to 5 with 2 N NaOH.
The mixture was extracted with EtOAc (3 × 50 mL). The
combined organic extracts were washed with brine (3 × 25
mL), dried (Na2SO4), and concentrated under reduced pressure
yielding 1.1 g of compound 4 as a colorless solid. Recrystalli-
zation from methanol-water yielded 0.72 g (87%) of a dihy-
drate as colorless crystals, mp 243-245 °C.
Ack n ow led gm en t. We thank Dr. W. von Daehne
and Dr. W. O. Godtfredsen for inspiring discussions and
useful suggestions and Mrs. A. G. Moller for excellent
technical assistance. This project was supported by the
EU (Inhibition and Transitions in Protein Synthesis,
Protein Structure Approach to Antibiotic Development,
Contract PL922188).
Refer en ces
(1) Godtfredsen, W. O.; J ahnsen, S.; Lorck, H.; Roholt, K.; Tybring,
L. Fusidic acid: A new antibiotic. Nature 1962, 987.
(2) Godtfredsen, W. O.; Roholt, K.; Tybring, L. Fucidin A new orally
active antibiotic. Lancet 1962, 928-931.
(3) Barber, M.; Waterworth, P. M. Antibacterial activity in vitro of
fucidin. Lancet 1962, 931.
(4) Golledge, C. Fusidic acid in other infections. Int. J . Antimicrob.
Agents 1999, 12, S11-S15.
(5) Christiansen, K. Fusidic acid adverse drug reactions. Int. J .
Antimicrob. Agents 1999, 12, S3-S9.
(6) Turnidge, J . Fusidic acid pharmacology, pharmacokinetics and
pharmacodynamics. Int. J . Antimicrob. Agents 1999, 12, S23-
S9.
(7) Godtfredsen, W. O.; Albrethsen, C.; von Daehne, W.; Tybring,
L.; Vangedal, S. Transformations of fusidic acid and the rela-
tionship between structure and antibacterial activity. Antimi-
crob. Agents Chemother. 1965, 132-137.
(8) von Daehne, W.; Godtfredsen, W. O.; Rasmussen, P. R. Structure
activity relationships in fusidic acid-type antibiotics. Adv. Appl.
Microbiol. 1979, 25, 95-146.
(9) J anssen, G.; Vanderhaeghe, H. Modification of the side chain of
fusidic acid (Ramycin). J . Med. Chem. 1966, 10, 205-208.
(10) Bodley, J . W.; Godtfredsen, W. O. Studies on translocation XI:
Structure-function relationships of the fusidane-type antibiot-
ics. Biochem. Biophys Res. Commun. 1972, 46, 871-877.
(11) Willie, G. R.; Richman, N.; Godtfredsen, W. O.; Bodley, J . W.
Some characteristics of and structural requirements for the
interaction of 24,25-dihydrofusidic acid with ribosome elongation
factor G complexes. Biochemistry 1975, 14, 1713-1718.
(12) Godtfredsen, W. O.; von Daehne, W.; Tybring, L.; Vangedal, S.
Fusidic acid derivatives. I. Relationship between structure and
antibacterial activity. J . Med. Chem. 1966, 9, 15-22.
(13) Cooper, A.; Hodgkin, D. C. The crystal structure and absolute
configuration of fusidic acid methyl ester 3-p-bromobenzoate.
Tetrahedron 1968, 24, 909-922.
(14) Sotofte, I.; Duvold, T. A new potent fusidic acid analogue. Acta
Cryst. 2001, E57, 829-831.
(15) Laurberg, M.; Kristensen, O.; Martemyanov, K.; Gudkov, A. T.;
Nagaev, I.; Hughes, D.; Liljas, A. Structure of a mutant EF-G
reveals domain III and possibly the fusidic acid binding site. J .
Mol. Biol. 2000, 303, 593-603.
NMR (CDCl3): 0.91 (d J ) 6.8 Hz 3H), 0.97 (s, 3H), 1.01 (s,
3H), 1.31 (s, 3H), 1.57 (bs, 3H), 1.66 (bs, 3H), 2.01 (s, 3H), 2.51
(m, 1H), 2,68 (m, 2H), 3.75 (bs, 1H), 4.24 (bs, 1H), 5.05 (bt,
1H), 5.39 (bt, 1H). MS: Calc. for C31H46O4 (M-2xH2O) m/z
482.3396, Observed m/z 482.3400. Anal. (C31H50O6, 2H2O) C,
H.
17S,20S-Dih yd r ofu sid ic Acid (5). The reaction was car-
ried out as described for the preparation of 4. Mp 195-195.5
°C. NMR (CDCl3): 0.91 (d, J ) 6.8 Hz, 1H), 0.97 (s, 3H), 1.01
(s, 3H), 1.32 (s, 3H), 1.58 (bs, 3H), 1.68 (bs, 3H), 1.95 (s, 3H),
2.60 (m, 1H), 2.71 (m, 2H), 3.75 (bs, 1H), 4.34 (bs, 1H), 5.06
(bt, 1H), 5.32 (bs, 1H). MS: Calc. for C31H46O4 (M-2H2O) m/z
482.3396, Observed m/z 482.3391 Anal. (C31H50O6, 2H2O) C,
H.
(16) Hewitt, W.; Vincent, S. The agar diffusion assay. In theory and
application of microbiological assay. Academic Press: San Diego,
1988; pp 38-79.
(17) Godtfredsen, W. O.; von Daehne, W.; Vangedal, S.; Marquet, A.;
Arigoni, D.; Melera, A. The stereochemistry of fusidic acid.
Tetrahedron 1965, 21, 3505-3530.
17S,20R-Dih yd r ofu sid ic Acid La cton e (6a ). Anal. (C29
44O4) C,H; C: Calcd: 76.27; found: 76.20; H: calcd. 9.71;
found, 9.77.
-
H
J M010899A