Synthesis and biological evaluation of photoaffinity labeled fusidic acid analogues

Article abstract of DOI:10.1021/jm050583t

Novel photoaffinity labeled fusidic acid analogues were obtained by a synthetic sequence employing a Wittig reaction between a fusidic acid aldehyde and benzyl bromides in the key step. Three commonly used photoreactive groups, benzophenone, trifluoromethyldiazirine, and aryl azide, were used. The photoaffinity labeled fusidic acid analogues demonstrated a potent antibacterial activity (MIC 0.016-4 μg/mL) and therefore represent a potential tool for the elucidation of the interactions between fusidic acid and its receptor EF-G.

Full text of DOI:10.1021/jm050583t

J. Med. Chem. 2006, 49, 1503-1505
1503
Synthesis and Biological Evaluation of
Photoaffinity Labeled Fusidic Acid Analogues
Ditte Riber,† Musturi Venkataramana,‡ Suparna Sanyal,‡ and
Tore Duvold*,†
LEO Pharma, Industriparken 55, 2750 Ballerup, Denmark, and
Department of Cell and Molecular Biology, Uppsala UniVersity,
Box-596, BMC, Uppsala, Sweden
ReceiVed June 19, 2005
Figure 1. The structure of fusidic acid.
Abstract: Novel photoaffinity labeled fusidic acid analogues were
obtained by a synthetic sequence employing a Wittig reaction between
a fusidic acid aldehyde and benzyl bromides in the key step. Three
commonly used photoreactive groups, benzophenone, trifluorometh-
yldiazirine, and aryl azide, were used. The photoaffinity labeled fusidic
acid analogues demonstrated a potent antibacterial activity (MIC
0.016-4 µg/mL) and therefore represent a potential tool for the
elucidation of the interactions between fusidic acid and its receptor
EF-G.
Fusidic acid (Figure 1), 1, is a unique antibiotic with a potent
activity against Staphlococcus aureus, including strains resistant
to other classes of antibiotics. Fusidic acid inhibits the bacterial
protein synthesis by interference with elongation factor G (EF-
G) in the translocation step,1 the process by which the ribosome
moves relative to mRNA.2 In the presence of fusidic acid, EF-G
remains bound to the ribosome after GTP hydrolysis and
translocation, preventing further protein synthesis. Although the
overall role of fusidic acid in bacterial protein synthesis is well
described, little is known about the details in the interaction
with EF-G. A putative binding site has recently been proposed
by analyzing a number of fusidic acid resistant bacteria with
mutations in EF-G.3 The location of a mutation cluster confer-
ring resistance to fusidic acid was identified and used to suggest
a possible binding site. However, the exact binding mode of
fusidic acid to EF-G and the ribosome remains unknown.
A detailed binding model should represent a useful tool for
the rational design of new fusidic acid analogues with improved
antibacterial properties and resistance profile. In the absence
of a cocrystal structure of fusidic acid and EF-G, photoaffinity
labeling studies offer a useful approach in identifying the amino
acid residues involved in ligand binding. We herein describe
the first example of an efficient synthetic sequence for the
preparation of photoaffinity labeled fusidic acid with three
different types of photoreactive groups in the side chain.
The structure-activity relationship (SAR) of fusidic acid has
been extensively studied, and a large number of analogues have
been prepared and tested for antibacterial activity.4-6 These
results demonstrate the crucial importance of the carboxylic acid
and the acetate as well as of the hydroxy group at C-3. In a
recent paper studying the role of the side chain, we showed
that the side chain can only occupy a limited conformational
space with regard to both the carboxylic acid group and the
lipophilic moiety.7 However, lipophilic substituents can be
linked to the side chain without causing loss of antibacterial
potency.8 The SAR data clearly suggest that the lipophilic part
of the side chain represents an ideal location for aromatic
Figure 2. Photoaffinity labeling the side chain in fusidic acid.
Figure 3. The selected photoaffinity groups.
photoaffinity moieties (Figure 2), and thus retention of potency
and minimum changes in the ligand-receptor interactions could
be expected.
Three photoreactive groups, benzophenone 3, trifluorometh-
yldiazirine 4, and azide 5, were selected (Figure 3), as they are
considered to be the most reliable and have demonstrated utility
in several ligand-receptor studies.9 The phenyl azide group was
chosen instead of the more reactive tetrafluorophenyl azide since
we were not able to prepare the necessary triphenylphosphine
building block according to the protocol outlined in Scheme 3.
The selected photoaffinity groups react with the receptor through
different reactive intermediates: a radical (benzophenone), a
carbene (trifluoromethyldiazirine), and a nitrene (aromatic
azide).9
The use of different groups increases the chance of successful
cross-linking to the target protein, since it cannot be predicted
in advance which group will be most readily attached to the
receptor. Upon RNase treatment and protein digestion, the
fragments are to be analyzed by MS-MS to determine the
peptide sequence to which fusidic acid is bound. Correlating
this information to the structural data available, this could
subsequently reveal the location of fusidic acid in complex with
EF-G and the ribosome.
* Corresponding author. Tel.: +4572262225; Fax: +4544669720;
e-mail: tore.duvold@leo-pharma.com.
† LEO Pharma.
Aldehyde derivative, 6, of fusidic acid was synthesized to
enable introduction of photoaffinity groups via a Wittig reaction
‡ Uppsala University.
10.1021/jm050583t CCC: $33.50 © 2006 American Chemical Society
Published on Web 02/15/2006

1504 Journal of Medicinal Chemistry, 2006, Vol. 49, No. 5
Scheme 1. Synthesis of 6a
Letters
Scheme 3. Synthesis of 16a,b + 15ca
a Reagents: (a) (CH3)3CO2CH2Cl, TEA, DMF, 90%; (b) O3, CH2Cl2,
-78 °C, 64%.
Scheme 2. Synthesis of the Diazirine Coupling Compound 12ba
a Reagents: (a) PPh3, toluene; (b) t-BuOK, toluene; (c) 7, a: 75%, b:
68%, c: 85% (for three steps); (d) H2, Pd/C, MeOH, a: 100%, b: 57%;
(e) K2CO3, MeOH, a: 85%, b: 69%.
Scheme 4. Synthesis of 18a
a Reagents: (a) N-Trifluoroacetylpiperidine, n-butyllithium., 83%; (b)
hydroxylamine, pyridine, ethanol, 93% (not purified); (c) p-toluenesulfonyl
chloride, pyridine, 82%; (d) NH3, 76%; (e) NBS, AIBN, 59%.
by coupling with benzyl bromides 12a-c. The aldehyde was
prepared by protecting the carboxylic acid as a pivaloyloxy-
methyl ester followed by selective ozonolysis of the ∆C24(25)
double bond (Scheme 1).
4-(Bromomethyl)benzophenone, 12a, and 4-nitrobenzyl bro-
mide, 12c, were commercially available, whereas bromotri-
fluoromethyldiazirine, 12b, was synthesized by slightly modify-
ing a reported procedure (Scheme 2).10 Starting from readily
available 4-bromotoluene, 7, reaction with N-trifluoroacetylpi-
peridine gave trifluoroacetyltoluene 8, which was converted to
the hydroxyl oxime 9 by condensation with hydroxylamine in
pyridine. Tosylation with p-toluenesulfonyl chloride gave 10,
which was treated with ammonia under pressure to yield
diaziridine 11. Bromination at the benzylic position with
N-bromosuccinimide (NBS) using a catalytic amount of 2,2′-
azobisisobutyronitrile (AIBN) also promoted formation of the
diazirine moiety to give 12b. The reaction most likely takes
place by in situ formation of Br2 which is capable of oxidizing
the diaziridine.
Coupling between fusidic acid aldehyde, 6, and benzyl
bromides, 12a-c, was smoothly achieved via a Wittig reaction
(Scheme 3). Synthesis of the benzophenone labeled fusidic acid
derivative 16a was prepared through in situ formation of
phosphonium salt 13a with triphenylphosphine in toluene at
reflux, subsequent deprotonation with potassium tert-butoxide
to give 14a, and reaction with aldehyde 7 to give coupling
product 15a. Selective hydrogenation of the double bond was
possible, but prolonged reaction time led to complete reduction
of the carbonyl in the benzophenone moiety as well. Carboxylic
acid 16a was obtained by deprotection with potassium carbonate
in methanol.
a Reagents: (a) H2, Pd/C, MeOH, 62%; (b) TfN3, TEA, CuSO4, H2O,
MeOH, 74% (c) K2CO3, MeOH, 86%.
heating a mixture of the diazirine and triphenylphosphine. The
reaction was carried out at 80 °C, since the diazirine was found
to decompose at temperatures above 90 °C. The phosphonium
salt was in this case isolated and purified before treatment with
t-BuOK to give ylid, 14b. The coupling product, 15b, was
obtained by condensation with aldehyde 6, and the double bond
was then carefully hydrogenated without reducing the labile
diazirine functionality. The desired product 16b was obtained
by deprotection of the acid with potassium carbonate in
methanol.
In the case of coupling with 4-nitrobenzyl bromide, 12c, the
Wittig reaction was carried out at room temperature to give 15c.
Hydrogenation of both the double bond and the aromatic nitro
group gave 17 (Scheme 4), in which the aromatic amine was
converted into an azide by treatment with triflyl azide, triethyl-
amine, and CuSO4 as catalyst.11 The desired fusidic acid
analogue was then obtained by deprotection of the acid with
potassium carbonate to give 18 (Scheme 4).
To facilitate the identification of cross-linked fragments after
protein digestion, additional radiolabeling was required. The
synthesis of the tritium labeled analogues have been successfully
obtained by selective tritiation of the ∆24(25) double bond in
For the preparation of the diazirine photolabeled fusidic acid
analogue, phosphonium salt 13b was synthesized by carefully

Letters
Journal of Medicinal Chemistry, 2006, Vol. 49, No. 5 1505
Table 1. MIC Values of Photoaffinity Labeled Fusidic Acid Analogues
16a, 16b, and 18
successfully cross-linked to EF-G. Localization of the exact
binding site will be a valuable tool in our efforts to design
derivatives with improved antibacterial and pharmacokinetic
properties.
MIC (µg/mL)a
benzophenone
fusidic acid
(16a)
diazirine
fusidic acid
(16b)
azide
fusidic
acid (18)
fusidic
acid (1)
organism/strain
Acknowledgment. We thank Dr. Hanne Aae Theilgaard and
Ms. Merete Henriksen for providing the antimicrobial data. This
project was funded by the EU (New Antimicrobials Targeting
Translation in Bacteria and Fungi, Contract QLK2-CT-2002-
00892) and by Swedish National Research Council (VR) grant
to S.S. (621-20002-4747).
S. aureus, CJ 247
0.063
1
-
-
S. aureus, CJ 234 R <0.063 1-2
1-4
1-2
1-2
1-2
1-2
4
S. aureus, CJ 200
S. aureus, CJ 251
S. aureus, CJ 288
S. Epidermis, CK 5
C. xerosis, FF
0.063
4
2-4
<0.063 1-2
1-4
<0.063 1-2
1-4
0.063
0.016
1-4
0.016-0.063
1-4
0.063-0.125
0.016
Supporting Information Available: General procedure for the
Wittig reaction, biological procedure for antimicrobial testing,
analytical and spectral chacterization data for compounds 16a, 16b,
and 18, and binding of fusidic acid and its photoaffinity labeled
analogues to EF-G-GDP on the ribosome. This material is available
free of charge via the Internet at http://pubs.acs.org
a The MIC values were obtained from quadruple testing of the labeled
fusidic acid analogues and are within the stated intervals.
References
(1) (a) Yamaki, H. Inhibition of protein synthesis by fusidic and
helvolinic acids, steroidal antibiotics J. Antibiot. 1965, 18, 228. (b)
Harvey, C. L.; Knight, S. G.; Sih, C. L. On the mode of action of
fusidic acid. Biochemistry 1966, 5, 3320. (c) Tanaka, N.; Yamaki,
H.; Lin, Y.; Umezawa, H. Further studies on inhibition of protein
synthesis by fusidic and helvolinic acids. J. Antibiot. 1967, 20, 156-
161.
(2) Burns, K.; Cannon, M.; Cundliffe, E. A resolution of conflicting
reports concerning the mode of action of fusidic acid. FEBS Lett.
1974, 40, 219-223.
(3) 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.
(4) Godfredsen, W. O.; Albrethsen, C.; von Daehne, W.; Tybring, L.;
Vangedal, S. Transformations of fusidic acid and the relationship
between structure and antibacterial activity. Antimicrob. Agents
Chemother. 1965, 132-137.
(5) von Daehne, W.; Godtfredsen, W. O.; Rasmussen, P. R. Structure
activity relationships in fusidic acid-type antibiotics. AdV. Appl.
Microbiol. 1979, 25, 95-146.
Figure 4. Autoradiogram of the SDS-PAGE of UV-treated 70S
ribosome-EF-G-GDP-3H-labeled fusidic acid azide analogue complex
reveals cross-linking only on EF-G (lane A), similar treatment with
only 70S ribosome shows no cross-linking (lane B). Lane M, 14C-labeled
molecular weight markers.
derivatives 15a, 15b, and 15b. The preparation of the tritium
labeled analogues will be published elsewhere.
The three compounds revealed potent antibacterial activity
with minimum inhibitory concentration (MIC) values ranging
from 0.016 to 4 µg/mL (Table 1). The retained antibacterial
activity of the photoaffinity labeled fusidic acid analogues
suggested that these could also bind to EF-G on the ribosome
in the same way as fusidic acid, and they should therefore be
suitable for cross-linking experiments.
(6) Duvold, T.; Sørensen, M. D.; Bjo¨rkling, F.; Henriksen, A. S.; Rastrup-
Andersen, N. Synthesis and conformational analysis of fusidic acid
side chain derivatives in relation to antibacterial activity. J. Med.
Chem. 2001, 44, 3125-3131.
(7) Duvold, T.; Jørgensen, A.; Andersen, N. R.; Henriksen, A. S.;
Sørensen, g, F. 17S,20S-Methanofusidic acid, a new potent semi-
synthetic fusidane antibiotic. Bioorg. Med. Chem. Lett. 2002, 12,
3569-3572.
(8) Duvold et al. unpublished results.
We have tested their binding to EF-G by nitrocellulose filter-
binding assay following the protocol of Bodley et al.12 The data
clearly demonstrate that the three photoaffinity labeled fusidic
acid analogues are equally potent in the formation and stabiliza-
tion of 70S ribosome-EF-G-GDP complex as fusidic acid.
Furthermore, UV cross-linking with the 3H-labeled azide
analogue to such complex confirms covalent binding only on
EF-G when autoradiographed after separation of the complex
in a SDS-PAGE (Figure 4). Detection of the cross-linking sites
on EF-G is currently ongoing.13 Successful localization of the
fusidic acid binding site should become a valuable tool in the
efforts to design new analogues.
In summary, an efficient synthetic route for a series of
photoaffinity labeled fusidic acid analogues has been developed
employing a Wittig reaction between aldehyde 6 and benzyl
bromides 12a-c in the key step. Antimicrobial evaluation
showed that the three analogues retained a potent activity against
fusidic acid-sensitive bacteria and that azide fusidic acid 18 was
(9) (a) Kotzyba-Hilbert, F.; Kapfer, I.; Goeldner, M. Recent trends in
photoaffinity labelling. Angew. Chem., Int. Ed. Engl. 1995, 34, 1296-
1312. (b) Dorma´n, G.; Prestwich, G. D. Using photolabile ligands
in drug discovery and development. Trends Biotechnol. 2000, 18,
64-77. (c) Hatanaka, Y.; Sadakane, Y. Photoaffinity labelling in
drug discovery and developments: Chemical gateway for entering
proteomic frontier. Curr. Topics Med. Chem. 2002, 2, 271-288. (d)
Fleming, S. A. Chemical reagents in photoaffinity labelling. Tetra-
hedron 1995, 51, 12479-12520.
(10) Strømgaard, K.; Saito, D. R.; Shindou, H.; Ishii, S.; Shimizu, T.;
Nakanishi, K. Ginkgolide derivatives for photolabeling studies:
Preparation and pharmacological Evaluation. J. Med. Chem. 2002,
45, 4038-4046.
(11) Liu, Q.; Tor, Y. Simple conversion of aromatic amines into azides.
Org. Lett. 2003, 5, 2571-2572.
(12) Bodley, J. W.; Zieve, F. J.; Lin, L.; Zieve, S. T. Studies on
translocation.3. Conditions necessary for the formation and detection
of a stable ribosome-G factor-guanosine diphosphate complex in the
presence of fusidic acid. J. Biol. Chem. 1970, 245, 5656-5661.
(13) These studies are being carried out by Dr. Sanyal at Uppsala
University.
JM050583T