Synthesis and biological investigation of new 4″-malonyl tethered derivatives of erythromycin and clarithromycin

Article abstract of DOI:10.1016/j.bmcl.2005.12.033

A new approach to 4″-substituted derivatives of erythromycin and clarithromycin was developed by converting them into corresponding 4″-malonic monoesters. Subsequent carbodiimide coupling with alcohols and amines provided new macrolide derivatives that are capable of binding to 50S ribosomal subunits and inhibiting protein synthesis in cell-free system.

Full text of DOI:10.1016/j.bmcl.2005.12.033

Bioorganic & Medicinal Chemistry Letters 16 (2006) 1506–1509
Synthesis and biological investigation of new 400-malonyl
tethered derivatives of erythromycin and clarithromycin
Daniel Sherman,a Liqun Xiong,b Alexander S. Mankinb and Artem Melmana,*
aDepartment of Organic Chemistry, Hebrew University of Jerusalem, Jerusalem 91904, Israel
bCenter for Pharmaceutical Biotechnology, University of Illinois, Chicago, IL 60607, USA
Received 25 August 2005; accepted 12 December 2005
Available online 4 January 2006
Abstract—A new approach to 400-substituted derivatives of erythromycin and clarithromycin was developed by converting them into
corresponding 400-malonic monoesters. Subsequent carbodiimide coupling with alcohols and amines provided new macrolide deriv-
atives that are capable of binding to 50S ribosomal subunits and inhibiting protein synthesis in cell-free system.
Ó 2005 Elsevier Ltd. All rights reserved.
Erythromycin 1a and other macrolide antibiotics
X
R1
(Scheme 1) have been used for the treatment of a variety
of bacterial infections for the past 50 years. Low toxicity
and cost as well as low incidence of side effects resulted
in extensive use of erythromycin both for treatment of
infections and prophylactic purposes.1 The active use
of the drug resulted in a wide spread of macrolide resis-
tance in a number of pathogenic strains. As an example,
in Asia the majority (up to 80% in Hong Kong) of Strep-
tococcus pneumoniae strains carry resistance to macro-
lide antibiotics.2 Attempts to improve relatively poor
bioavailability of erythromycin, increase its spectrum
of activity, and overcome bacterial resistance resulted
in synthesis and investigation of a number of semisyn-
thetic macrolide derivatives such as clarithromycin 1b,
azithromycin 1c, and roxithromycin 1d. The most recent
generation of macrolide antibiotics, such as telithromy-
cin (KETEKÒ) 1e and ABT-773 1f,3 possesses enhanced
biological activity; however, their synthesis is much
more complicated.
OR
R2
HO NMe2
2'
O
O
3
R4
O
O
R3
1a R=H, R1=R2=OH, R3=H, R4= cladinose, X= C=O
1b R=Me, R1=R2=OH, R3=H, R4= cladinose, X= C=O
1c R=H, R1=R2=OH, R3=H, R4= cladinose, X= N(Me)CH2
1d R=H, R1=R2=OH, R3=H, R4= cladinose,
X= C=NOCH2OCH2CH2OMe
1e R=Me, R1,R2= -OC(O)N(CH2)3
N
3-Pyridyl,
R3, R4=O, X= C=O
N
1f R=CH2CH=CH-3-Quinolinyl, R1,R2= -OC(O)NH- ,
R3, R4=O, X= C=O
Scheme 1. Erythromycin and semisynthetic macrolide antibiotics.
growth.4a–i X-ray structures showed that desosamine su-
gar at position 5 of the macrolide lactone ring provides
very important contribution to the overall binding ener-
gy of the drug.4a,i,5 Indeed, acylation of 20-hydroxy
group of desosamine drastically decreases antibacterial
activity.6 Newer macrolide derivatives such as telithro-
mycin, ABT-733 or bridged ketolides possess additional
heterocyclic arm which is most likely responsible for
their enhanced antibacterial activity.3
All macrolide antibiotics share a similar mechanism of
action which involves selective binding to the 50S sub-
unit of bacterial ribosome resulting in inhibition of pro-
tein synthesis. Genetic, biochemical, and more recent
crystallographic studies of ribosome–macrolide com-
plexes have indicated that macrolides bind in nascent
peptide exit tunnel thus blocking polypeptide chain
The extent of involvement of the cladinose sugar in ribo-
some binding is less clear. While a simple removal of the
cladinose moiety was found to substantially decrease the
activity of erythromycin,7 a number of highly active
macrolide derivatives, such as acylides8 and ketolides,9
Keywords: Macrolides; Ribosome; Antibiotics; Tethering.
*
Corresponding author. Tel.: +97226585279; fax: +97226585345;
e-mail: amelman@chem.ch.huji.ac.il
0960-894X/$ - see front matter Ó 2005 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2005.12.033

D. Sherman et al. / Bioorg. Med. Chem. Lett. 16 (2006) 1506–1509
1507
do not contain cladinose. There were relatively few
attempts to modify cladinose residues to increase the
bioavailability or activity of erythromycin. Attempts to
introduce simple alkanoyl6,10 and polyunsaturated alke-
noyl11 substituents into 400 position of erythromycin
involved either diacylation at 20 and 400 positions fol-
lowed by a selective hydrolysis, or selective formylation
of 20-position followed by the acylation of 400-position
and selective hydrolysis of 20 formyl group.6 Resultant
compounds exhibited lower10 or substantially lower12
than erythromycin antibacterial activity,13 most proba-
bly due to decreased bioavailability. High nucleotide
content of ribosome gives rise to a number of possible
interactions such as hydrogen bonding, p-stacking as
well as electrostatic interactions. Erythromycin and
clarithromycin derivatives carrying cladinose sugar
modified with functional groups capable of such types
of interaction have a potential to exhibit increased affin-
ity toward ribosomes including mutant ribosomes of
resistant bacteria. Indeed, Fernandes and co-authors
demonstrated substantial enhancing of antibacterial
activity through alteration of 400-position of erythromy-
cin oxime derivatives with heteroatom-functionalized
side chains.14 This type of derivatization, however, is
relatively difficult and laborious to perform by previous-
ly described methods.6
our synthetic approach to 400-tethered macrolide deriva-
tives of type 5 (Scheme 2) involved the use of one
common intermediate of type 4.
In line with this approach, 20-acetyl erythromycin 2a
was prepared by a treatment of 1a with acetic anhydride
in the absence of acylation catalysts.16 No acylation of
other hydroxyl groups was detected even in the presence
of a large excess of acetic anhydride. This is possible due
to the vicinity of NMe2 group in desosamine residue
which substantially increases rates of both introduction
and removal of 20-acyl group.6 In contrast, acylation of
1a with an excess of acetic anhydride in the presence of
DMAP yielded a complex mixture in which products
possessing one to four acetyl groups were detected by
ESI-MS. Compound 2a was then acylated by DCC cou-
pling through non-symmetric malonate method15,17
using tert-butyldiphenylsilyl and benzyl malonates.
The reaction of 20-acetyl erythromycin 2a with tert-
butyldiphenylsilyl malonate and DCC was found to
provide a mixture of starting material, mono-, and dia-
cylation products. In contrast, acylation of 2a with
benzyl malonate/DCC system proceeded smoothly to
give benzyl ester 3b without diacylation. Reasons for
such a different reactivity are not clear since tert-butyldi-
phenylsilyl malonate was found to be less reactive than
benzyl malonate in acylation of other sterically hindered
alcohols. The structure of 3b was verified by the
ESI MS–MS technique. Fragmentation of MH+ of 3a
produced MH+-cladinose-C(O)CH2CO2Bn cations. No
traces of MH+-cladinose ions that would indicate the
presence of isomers other than 400-benzylmalonylery-
thromycin were observed.
We recently reported on a new one-pot method for teth-
ering of organic molecules through the formation of
non-symmetric malonate derivatives.15 This method en-
ables efficient binding of alcohols and amines under very
mild conditions in the absence of acylation catalysts.
Since malonate linker is biocompatible and reasonably
stable, we were attracted by the possibility to derivatize
cladinose sugar of macrolide antibiotics by introducing
functional groups that could influence binding of the
drug to the ribosome. In contrast to all previous
attempts for selective 400-derivatization of macrolides,
We initially planned to remove the 20-acyl group by
methanolysis of ester 3b. However, there was a consider-
able decrease in the methanolysis rates of malonyl deriv-
ative 3b in comparison to 20-acetyl erythromycin 2a
O
O
O
R1
O
R1
O
OR
O
OR
O
OR
O
HO
HO
HO
1a
1b
c
a
b
O
NMe2
HO NMe2 d
O
NMe2
OH
OH
OH
2'
2'
2'
O
O
O
3
3
3
O
O
O
O
O
O
O
O
O
OMe
OMe
OMe
O
OR2
O
OH
O
O
4''
4''
O
O
4''
O
O
O
OH
2a: R = H, R1= Me
3a: R = H, R1= Me, R2=SiPh2tBu
3b: R = H, R1= Me, R2=CH2Ph
3c: R = H, R1= CH2OMe, R2=CH2Ph
4a: R = H
4b: R = Me
2b: R = H, R1= CH2OMe
2c: R = Me, R1= CH2OMe
3d: R = Me, R1= CH2OMe, R2=CH2Ph
O
OR
O
HO
HO NMe2
OH
2'
O
3
O
O
O
5a: R = H, XR3=OCH2Ph
OMe
5b: R = CH3, XR3=OCH2Ph
O
5c: R = CH3, XR3=OCH2C6H4p-OMe
5d: R = CH3, XR3=OCH2C6H44-N3
5e: R = H, XR3=N(Me)CH2CH2NMe2
4''
O
O
XR3
O
Scheme 2. Preparation of cladinose tethered derivatives of erythromycin and clarithromycin. Reagents and conditions: (a) (R1CO)2O, DCM, rt;
(b) R2OCOCH2CO2H, DCC, rt; (c) MeOH, rt, 24 h, then H2/Pd/MeOH/acetate buffer; (d) HXR3, DCC, DCM, rt.

1508
D. Sherman et al. / Bioorg. Med. Chem. Lett. 16 (2006) 1506–1509
which resulted in competing methanolysis of 400-ben-
zylmalonyl functions thus contaminating the desired
20-deacetylated product 4a with erythromycin. A num-
ber of alternative 20-protecting groups were evaluated.
Acylation of erythromycin with anhydrides of chloro-
acetic and dichloroacetic acids as well as HCO2COMe
failed to provide the desired 20-acylation products. In
contrast, reaction with methoxyacetic anhydride pre-
pared by the treatment of methoxyacetic acid with
DCC was found to proceed smoothly producing 20-
methoxyacetyl erythromycin 2b. As expected, 400-acyla-
tion of 2b also proceeded selectively thus providing dies-
ter 3c that after subsequent methanolysis and catalytic
hydrogenation provided acid 4a without impurities of
erythromycin. The same synthetic sequence was success-
fully applied to clarithromycin to afford acid 4b.
significantly reduced uptake of the compounds 4a,b by
the cell reflected in their high MICs.
Interaction of several new compounds with the ribo-
some was further studied using RNA footprinting tech-
nique.4g Ribosomes were complexed with erythromycin
or telithromycin (as controls) or with one of the newly
synthesized compounds: 4a, 4b or 5e, and rRNA was
modified with dimethyl sulfate (DMS). Desosamine su-
gar of active macrolide antibiotics is known to protect
adenines at positions 2058 and 2059 of 23S rRNA
from DMS modification.4g Both of the control drugs,
erythromycin and telithromycin, protected these rRNA
positions from DMS modification. Similarly, the tested
compounds with derivatized cladinose sugar effectively
protected A2058 and A2059 from modification thus
confirming binding of the new derivatives at the macro-
lide site. Thus, increased bulkiness of the substitution
at position 3 of lactone ring does not prevent new
derivatives from binding to the ‘conventional’ macro-
lide site in the ribosome. In E. coli ribosome, the
extended side chain of telithromycin reaches across
the nascent peptide exit tunnel and interacts with the
loop of helix 35 of 23S rRNA where it protects the
adenine residue at position 752 of 23S rRNA from
modification with DMS. In contrast, binding of eryth-
romycin or clarithromycin slightly enhances accessibil-
ity of A752 to modification with DMS.5,19 Neither of
the three tested compounds, 4a,b or 5e, protected
A752. Instead, a slight enhancement of its modification
with DMS, similar to that seen in the presence of
erythromycin, was observed (data not shown). No
other protection or enhancement in modification of
23S rRNA nucleotides was seen. Thus, we conclude
that in contrast to the side chain of ketolide antibiotics,
the side chains at the cladinose sugar do not reach po-
sition 752, suggesting that its orientation is different
from that of the ketolide’s alkyl-aryl extension. Given
that in the currently available crystallographic struc-
tures of the ribosome–macrolide complexes, the cladi-
nose sugar points in the general direction of the
peptidyl transferase center, we anticipate that upon fur-
ther growth of the side chain it will eventually reach
positions 2610 or 2555 located at the peptidyl transfer-
ase/exit tunnel doorway.
Since carboxylic groups of malonates 4a,b, as those of
other malonate monoesters, are capable to form ketene
intermediates upon the treatment with carbodiimides,
several new macrolide derivatives were prepared by car-
bodiimide-mediated coupling with compounds possess-
ing hydroxy and/or amino groups. DCC-mediated
coupling of 4a,b with benzyl, 4-methoxybenzyl, and 4-
azidobenzyl alcohols provided corresponding non-sym-
metric malonates 5a–d. A reaction with trimethylethyl
enediamine was found to proceed much slower and the
use of two equivalents of the amine and DCC was essen-
tial to achieve 35% isolated yield.
Compounds 4a,b and 5a,b,e were tested for their ability to
inhibit the growth of a macrolide-sensitive Escherichia
coli strain HN818 that lacks AcrAB transporter.18
Minimal inhibitory concentrations of control drugs,
erythromycin and clarithromycin, were 2 lg/ml. MIC of
5a,b, and e were somewhat higher—8 lg/ml, whereas
MIC of 4a,b were significantly higher—128 lg/ml.
Activity of compounds 4a,b, 5a,b,e as protein synthesis
inhibitors was tested in E. coli cell-free translation sys-
tem. Though inhibitory activity of compounds 4a and
4b was reduced compared to that of erythromycin
(IC50 3.7 and 3.9 lM vs 0.9 lM), inhibition of transla-
tion afforded by compounds 5a, 5b, and 5e (1.1, 1.2,
and 1.1 lM correspondingly) was comparable to that
of erythromycin. Given efficient inhibition of cell-free
protein synthesis by compounds 5a,b, and e, we can con-
clude that their somewhat reduced antibacterial activity
(MIC 8 lg/ml vs 2 lg/ml) should be attributed to their
decreased uptake by the cell, possibly due to increased
molecular weight. High in vitro activity of compounds
5a,b possessing bulky and relatively rigid benzyl groups
indicates that the ribosome can accommodate fairly
large substitutions at this position—an observation
compatible with the available crystal structures.
In conclusion, we developed a new common precursor
approach to the synthesis of 400-derivatized macrolide
derivatives and tested the resultant compound in several
biological assays. Introduction of non-polar moieties
into 400-position was found to retain ribosome affinity
but decrease bioavailability, while the introduction of
acidic groups substantially decreases both of them.
Acknowledgments
Since steric effects do not appear to play substantial role
in ribosome binding, the lower activity of acids 4a,b
should be attributed to either hydrophobic environment
in binding region or, possibly, the presence of a phos-
phate group in a close vicinity to the carboxylate func-
tion in the ribosome complexes with either 4a or 4b.
The presence of a negatively charged group obviously
This research was supported by Israel Science Founda-
tion (Grant No. 176/02-1) (to A.M.) and NIH Grant
U19 AI56575 (to A.S.M.). We thank Teva Pharmaceu-
tical for generous donation of clarithromycin and Aven-
tis Pharmaceuticals for telithromycin.

D. Sherman et al. / Bioorg. Med. Chem. Lett. 16 (2006) 1506–1509
1509
9. Kashimura, M.; Asaka, T.; Misawa, Y.; Matsumoto, K.;
Morimoto, S. J. Antibiot. 2001, 54, 664.
References and notes
10. Martin, Y. C.; Jones, P. H.; Perun, T. J.; Grudy, W. E.; Bell,
S.; Bower, R. R.; Shipkowitz, N. L. J. Med. Chem. 1972, 15,
634.
11. Philippe, M.; Campos, A.; Sebag, H.; Rougier, A.;
Dupuis, D.; Saint-Leger, D.; Vanlerberghe, G. Chem.
Pharm. Bull. 1990, 38, 1672.
1. Bryskier, A. Antibiotic therapy: past, present and future.
Presse Med. 1995, 24, 55.
2. Ip, M.; Lyon, D. J.; Yung, R. W. H.; Chan, C.; Cheng, A.
F. B. Antimicrob. Agents Chemother. 2001, 45, 1578.
3. Gaynor, M.; Mankin, A. S. Curr. Top. Med. Chem. 2003,
3, 949.
12. Philippe, M.; Campos, A.; Sebag, H.; Rougier, A.;
Dupuis, D.; Saint-Leger, D.; Vanlerberghe, G. Chem.
Pharm. Bull. 1990, 38, 1672.
4. (a) Schlunzen, F.; Harms, J. M.; Franceschi, F.; Hansen,
H. A.; Bartels, H.; Zarivach, R.; Yonath, A. Structure
2003, 11, 329; (b) Schlunzen, F.; Zarivach, R.; Harms, J.;
Bashan, A.; Tocilj, A.; Albrecht, R.; Yonath, A.; France-
schi, F. Nature 2001, 413, 814; (c) Berisio, R.; Schluenzen,
F.; Harms, J.; Bashan, A.; Auerbach, T.; Baram, D.;
Yonath, A. Nat. Struct. Biol. 2003, 10, 366; (d) Berisio, R.;
Harms, J.; Schluenzen, F.; Zarivach, R.; Hansen, H. A.;
Fucini, P.; Yonath, A. J. Bacteriol. 2003, 185, 4276; (e)
Hansen, J.; Ippolito, J. A.; Ban, N.; Nissen, P.; Moore, P.
B.; Steitz, T. A. Mol. Cell 2002, 10, 117; (f) Sigmund, C.
D.; Morgan, E. A. Proc. Natl. Acad. Sci. U.S.A. 1982, 79,
5602; (g) Moazed, D.; Noller, H. F. Biochimie 1987, 69,
879; (h) Horinouchi, S.; Weisblum, B. Proc. Natl. Acad.
Sci. U.S.A. 1980, 77, 7079; (i) Tu, D.; Blaha, G.; Moore,
P. B.; Steitz, T. A. Cell 2005, 121, 257.
13. In other macrolide antibiotics such as tylosin and deltamy-
cin, there was an increase of antibacterial activity, see: (a)
Okamoto, R.; Nomura, H.; Tsunekawa, H.; Fukumoto, T.;
Inui, T.; Sawa, T.; Taskeuchi, T.; Umezawa, H. J. Antibiot.
1979, 542; (b) Okamoto, R.; Nomura, H.; Tsunekawa, H.;
Fukumoto, T.; Inui, T.; Sawa, T.; Taskeuchi, T.; Umezawa,
H. J. Antibiot. 1980, 1309; (c) Okamoto, R.; Nomura, H.;
Tsunekawa, H.; Fukumoto, T.; Inui, T.; Sawa, T.; Taskeu-
chi, T.; Umezawa, H. J. Antibiot. 1990, 284.
14. Fernandes, P. B.; Baker, W. R.; Freiberg, L. A.; Hardy, D.
J.; McDonald, E. D. Antimicrob. Agents Chemother. 1989,
33, 78.
15. Sherman, D.; Shelkov, R.; Melman, A. Tetrahedron Lett.
2005, 46, 4901.
5. Hansen, L. H.; Mauvais, P.; Douthwaite, S. Mol. Micro-
biol. 1999, 31, 623.
6. Jones, P. H.; Perun, T. J.; Rowley, E. K.; Baker, E. J.
J. Med. Chem. 1972, 15, 631.
16. All compounds were characterized by 1H, 13C NMR
spectroscopy. Compounds 4a,b and 5a,b,e were also
characterized by HRMS.
17. Shelkov, R.; Nahmany, M.; Melman, A. J. Org. Chem.
2002, 67, 8975.
7. LeMahieu, R. A.; Carson, M. C.; Kierstead, R. W.
J. Med. Chem. 1974, 17, 953.
18. Xiong, L.; Kloss, P.; Douthwaite, S.; Andersen, N. M.;
Swaney, S.; Shinabarger, D. L.; Mankin, A. S. J.
Bacteriol. 2000, 182, 5325.
19. (a) Xiong, L.; Shah, S.; Mauvais, P.; Mankin, A. S. Mol.
Microbiol. 1999, 31, 633; (b) Garza-Ramos, G.; Xiong, L.;
Zhong, P.; Mankin, A. J. Bacteriol. 2002, 183, 6898.
8. (a) Tanikawa, T.; Asaka, T.; Kashimura, M.; Suzuki, K.;
Sugiyama, H.; Sato, M.; Kameo, K.; Morimoto, S.;
Nishida, A. J. Med. Chem. 2003, 46, 2706; (a) Tanikawa,
T.; Asaka, T.; Kashimura, M.; Misawa, Y.; Suzuki, K.;
Sato, M.; Kameo, K.; Morimoto, S.; Nishida, A. J. Med.
Chem. 2001, 44, 4027.