Hybrids of macrolides and nucleobases or nucleosides

Article abstract of DOI:10.1016/S0040-4039(00)00417-2

A few examples of hybrids/conjugates/chimeras of erythromycin A derivatives and nucleobases (uracil and thymine) or thymidine-derived nucleosides are reported. Linkers and reaction conditions have been investigated to avoid the degradation of the macrolide moiety (glycoside hydrolysis, ring cleavage, dehydration, etc.). (C) 2000 Elsevier Science Ltd.

Full text of DOI:10.1016/S0040-4039(00)00417-2

TETRAHEDRON
LETTERS
Pergamon
Tetrahedron Letters 41 (2000) 3371–3375
Hybrids of macrolides and nucleobases or nucleosides
Anna M. Costa ∗ and Jaume Vilarrasa ∗
Departament de Química Orgànica, Facultat de Química, Universitat de Barcelona, 08028 Barcelona, Catalonia, Spain
Received 1 January 2000; accepted 9 March 2000
Abstract
A few examples of hybrids/conjugates/chimeras of erythromycin A derivatives and nucleobases (uracil and
thymine) or thymidine-derived nucleosides are reported. Linkers and reaction conditions have been investigated
to avoid the degradation of the macrolide moiety (glycoside hydrolysis, ring cleavage, dehydration, etc.). © 2000
Elsevier Science Ltd. All rights reserved.
Keywords: antibiotics; macrolides; uracils; uridines.
The need for more potent and selective therapeutic agents and the fight against new antibiotic- or
antiviral-resistant strains are always forcing pharmaceutical companies to search for novel types of
bioactive molecular structures. Coupling of two or more different natural products, aimed at obtaining
‘unnatural’ molecules (hybrids, chimeras or conjugates) with a possible cooperative effect, is an approach
which has been investigated by many research groups, often with success.1,2 In this connection, the
recognition ability of nucleobases has been used to design hybrids with a great variety of molecules
such as porphyrins,3 β-lactam antibiotics,4 steroids,5 or peptides;6 some of these conjugates have found
applications in fields such as molecular recognition, inhibition of gene expression, or anti-cancer and
anti-virus therapies.
Taking into account the remarkable usefulness, nowadays, of the marketed erythromycin A-derived
macrolides, viz. azithromycin, clarithromycin, and roxithromycin,7 we started a project aimed at pre-
paring hybrids of erythromycins and nucleosides.8 In this preliminary communication, syntheses and
features of compounds 1–4 are reported.
∗
Corresponding authors. Tel: +34 93 4021258; fax: +34 93 3397878; e-mail: vilarrasa@qo.ub.es (J. Vilarrasa)
0040-4039/00/$ - see front matter © 2000 Elsevier Science Ltd. All rights reserved.
PII: S0040-4039(00)00417-2
tetl 16697

3372
Owing to the well-known instability of erythromycin derivatives in acidic media (formation of internal
hemiacetals or acetals if the CO group at C9 has not been modified, quick hydrolysis of the L-cladinose
moiety followed by a slower release of D-desosamine, translactonisation, etc.) and to basic media
(ring opening, retro-aldol reaction, etc.),9 the essential point was to link the macrolide and nucleoside
substructures by means of protocols that avoided deprotection steps or workups involving strong acidic
or basic treatments.
Macrolide components of hybrids 1 and 3: Reduction of erythromycin A oxime (prepared from
+
erythromycin A, EA, in 83% yield by using a very large excess of HONH3 Cl− in pyridine, at rt for
30 h) with TiCl3 and NaBH3CN, according to the procedure described by Leeds and Kirst,10 gave (9S)-
erythromycylamine A (5) in 65% yield. Treatment with acrolein (propenal) followed by in situ reduction
with NaBH4, as reported by Ryden et al.,11 afforded (9S)-9-N-(3-hydroxypropyl)erythromycylamine (6)
in 46% yield.
Macrolide components of hybrids 2 and 4: In two synthetic steps from EA oxime (Beckmann-like
rearrangement followed by reduction), according to the procedure of Djokic et al.,12 we prepared in
45% overall yield 9a-aza-9-deoxo-9a-homoerythromycin A (7), the synthetic precursor of the antibiotic
azitromycin (N-Me-7). Conversion of 7 into its N-3-hydroxypropyl derivative (8) was first tried by
reaction with acrolein and reduction in situ, as in the previous case, without success, even when using
a large excess of acrolein or by heating. On the other hand, reaction of 7 with methyl acrylate (methyl
propenoate), used as the solvent, at 60°C for 30 h gave 84% of the conjugate (hetero-Michael) adduct.
Reduction of the ester group with an excess of DIBALH gave only trace amounts of 8. The best results
were obtained with LiAlH4 (3–4 mmol per mmol of substrate) in THF at 0°C for 1 h (ca. 65% of 8
and 10% of 7, arising from a retro-addition in the basic medium, with 6–10% of the ester recovered);
longer reaction times or larger amounts of LiAlH4 were detrimental, with the appearance of more polar
byproducts.
Protected nucleobases as components of hybrids 1 and 2: We decided to use thymine and uracil
protected at N3 with Mocvinyl groups,13 because of their easy spectroscopic characterization and because
this protecting group can be removed afterwards under very controlled conditions. Reaction of these
nucleobases (10 mmol) with methyl propiolate (methyl propynoate, 22 mmol) and DMAP (5 mmol),

3373
in CH3CN (80 mL) at rt for 2 h, gave the bis-Mocvinyl derivatives 9a and 9b, respectively, in almost
quantitative yields. To remove the group at N1, as pyrrolidine13 turned out to be too reactive and
unselective, controlled amounts of morpholine (2 mmol per mmol of 9a) or of morpholine (1 mmol) plus
DBU (0.1 mmol) were employed under dilute conditions (0.01–0.05 M in CH3CN) at rt; thus, 10a14 and
10b were isolated in ca. 60% yields by column chromatography, 9a and 9b being recovered in 20–25%.
Thymidine derivatives as components of hybrids 3 and 4: At the beginning we decided to study AZT-
derived substrates as nucleoside components. Starting from 50-O-tert-butyldimethylsilyl-AZT (11), it
was envisaged that either the corresponding amine (12), isocyanate (13), or a synthetic equivalent (14)
might be appropriate. Isocyanate 13 could not be prepared from amine 12 (the symmetrical urea was
the major product); reaction of 11 with Me3P/THF in the presence of an excess of CO2 succeeded,
but 13 was too unstable to be purified and stored. On the other hand, reaction of amine 12 with N,N0-
carbonyldiimidazole in CH2Cl2 at 0°C gave 14 in 91% yield; this compound is stable and easy to purify.
Coupling of components: The Mitsunobu reaction between 6 and 10a (2 equiv), in dioxan at rt for 1
h, afforded the desired Mocvinyl-protected 1a in 87% yield.15 It was supposed that secondary hydroxy
groups of L-cladinose and D-desosamine moieties would not significantly compete with the primary
alcohol of 6, as found. Deprotection to 1a was accomplished in 83% isolated yield with 4 equiv. of
pyrrolidine in CH3CN, at rt for 24 h.16 Similarly, 6 and 10b gave 1b, 8 and 10a yielded 2a, and 8 and
10b gave 2b.
Coupling of 6 with 14, and that of 8 with 14, which would give carbamate-linked hybrids, did not
proceed at rt, while in refluxing CH3CN complex mixtures were formed.17 On the other hand, coupling
of amine 5 with 14 in CH2Cl2–DMF, at rt for 2 h, gave the urea in 80% yield, the TBS ether of which was
cleaved with HF/pyridine in THF at 0°C to afford hybrid 3 in 75% isolated yield. Moreover, coupling of
amine 7 with 14, in CH3CN–DMF at 50°C for 2 h, afforded, also after deprotection with HF/pyridine,
the desired urea 4,18 in 50% overall yield.
Concluding remarks: The antibiotic activity of 1–4 has been screened against Bacillus subtilis
ATCC6633, with a negative outcome (only between 4 and 13% of the azithromycin activity). Com-
pounds 1a,b and 2a,b were checked for their anti-mycobacterial activity (at TAACF, Birmingham, AL),
showing 80–83% of inhibition against Mycobacterium tuberculosis H37Rv, and as antitumorals (at NCI,
Bethesda), but none was active enough to pass to the next step. Compound 4 showed no activity against
the HIV-1 (NL4-3) ‘wild-type’ strain, nor against the HIV-1 AZT-resistant strain. Other samples (final
products and reaction intermediates) are under evaluation.
In summary, we have found suitable reaction conditions to link nucleobases and nucleosides to erythro-
mycin A derivatives, without causing undesired reactions in the macrolide substructures. Although the
preliminary results reported here are not exciting as far as the bioactivity is concerned, they are important
from a synthetic point of view, as the feasibility of different routes is now established. Syntheses of
alternative series of macrolide–nucleoside chimeras with longer linkers are in course in our lab, with the
hope that some of these structures exhibit a dual-action behaviour, taking into account the features of
their parent systems.

3374
Acknowledgements
This work has been supported by the Ministerio de Educación y Cultura (PM95-0061) and, in part,
by the Generalitat de Catalunya (1996SGR00120). We are also indebted to the Generalitat de Catalunya
for a doctorate grant to A.M.C. (1994–97), which allowed us to start this project. Thanks are due to Dra.
M. Carmen Cruzado and Pilar García (Fyse-Erkimia, Aranjuez) for screening the antibiotic properties of
1–4, and to Dr. José Esté (Hospital Germans Trias i Pujol, Badalona) for the anti-AIDS evaluation of 4
and related thymidine derivatives.
References
1. For representative instances, by semisynthesis, see: (a) Wang, J.; De Clercq, P. J. Angew. Chem., Int. Ed. Engl. 1995,
34, 1749. (b) Depew, K. M.; Zeman, M.; Boyer, S. H.; Denhart, D. J.; Ikemoto, N.; Crothers, D. M.; Danishefsky, S. J.
Angew. Chem., Int. Ed. Engl. 1996, 35, 2797. (c) Tietze, L. F.; Schneider, G.; Wölfling, J.; Nöbel, T.; Wulff, C.; Schubert,
I.; Rübeling, A. Angew. Chem., Int. Ed. Engl. 1998, 37, 2469. (d) Dräger, G.; Garming, A.; Maul, C.; Noltemeyer, M.;
Thiericke, R.; Zerlin, M.; Kirschning, A. Chem. Eur. J. 1998, 4, 1324, and references cited therein.
2. For reviews, by genetic modification of microorganisms, see: (a) Pieper, R.; Kao, C.; Khosla, C.; Luo, G.; Cane, D. E.
Chem. Soc. Rev. 1996, 25, 297. (b) Katz, L. Chem. Rev. 1997, 97, 2557. (c) Khosla, C. Chem. Rev. 1997, 97, 2577.
3. (a) Cornia, M.; Binacchi, S.; Del Soldato, T.; Zanardi, F.; Casiraghi, G. J. Org. Chem. 1995, 60, 4964, and references cited
therein. (b) Jiang, X.; Pandey, R. K.; Smith, K. M. Tetrahedron Lett. 1995, 36, 365, and references cited therein.
4. Dömling, A.; Starnecker, M.; Ugi, I. Angew. Chem., Int. Ed. Engl. 1995, 34, 2238.
5. Manoharan, M.; Tivel, K. L.; Cook, P. D. Tetrahedron Lett. 1995, 36, 3651.
6. (a) Uhlmann, E.; Peyman, A.; Breipohl, G.; Will, D. W. Angew. Chem., Int. Ed. Engl. 1998, 37, 2796. (b) Nielsen, P. E.;
Haaima, G. Chem. Soc. Rev. 1997, 26, 73.
7. (a) Kirst, H. A. In Recent Progress in the Chemical Synthesis of Antibiotics; Lukacs, G.; Ohno, M., Eds.; Springer: Berlin,
1990; Vol. 1, p. 39. (b) Lartey, P. A.; Faghih, R. In Recent Progress in the Chemical Synthesis of Antibiotics and Related
Microbial Products; Lukacs, G., Ed.; Springer: Berlin, 1993; Vol. 2, p. 121. (c) Lartey, P. A.; Perun, T. J. Studies in Natural
Products Chemistry; Atta-ur-Rahman, Ed.; Elsevier: 1993; Vol. 13, p. 155.
8. For a pioneering work on the subject (conjugates of erythromycylamine and aminopyrimidines, which are more active
against Gram-negative bacteria and show stronger affinity for ribosomes than erythromycin A), see: Werner, R. G.; Teraoka,
H.; Nierhaus, K. H. Biochem. Biophys. Res. Commun. 1978, 83, 1147.
9. For a classical review, see: (a) Macrolide Antibiotics: Chemistry, Biology, and Practice; Omura, S., Ed.; Academic Press:
Orlando, 1984. For more recent, representative instances of the reactivity of erythromycins, see: (b) Sakan, K.; Babirad,
S. A.; Smith, D. A.; Houk, K. N. Tetrahedron Lett. 1990, 31, 3687. (c) Bartra, M.; Urpí, F.; Vilarrasa, J. Tetrahedron Lett.
1992, 33, 3669. (d) Waddell, S. T.; Blizzard, T. A. Tetrahedron Lett. 1992, 33, 7827. (e) Pariza, R. J.; Freiberg, L. S. Pure
Appl. Chem. 1994, 66, 2365. (f) Faghih, R.; Edwards, C. M.; Freiberg, L. A.; Nellans, H. N. Synlett 1998, 751.
10. Leeds, J. P.; Kirst, H. A. Synth. Commun. 1988, 18, 777.
11. Ryden, R.; Timms, G. H.; Prime, D. M.; Wildsmith, E. J. Med. Chem. 1973, 16, 1059.
12. Djokic, S.; Kobrehel, G.; Lazarevski, G.; Lopotar, N.; Tamburasev, Z.; Kamenar, B.; Nagl, A.; Vickovic, I. J. Chem. Soc.,
Perkin Trans. 1 1986, 1881.
13. (a) Faja, M.; Ariza, X.; Gálvez, C.; Vilarrasa, J. Tetrahedron Lett. 1995, 36, 3261. (b) Costa, A. M.; Faja, M.; Farràs, J.;
Vilarrasa, J. Tetrahedron Lett. 1998, 39, 1835.
1
14. Spectral data of 3-[(E)-2-(methoxycarbonyl)vinyl]thymine (10a): H NMR (DMSO-d6, 200 MHz) δ 1.80 (d, J=1.0, 3H,
C6-Me), 3.69 (s, 3H, COOMe), 6.95 (d, J=14.7, 1H, CHCOOMe), 7.41 (dq, J=5.1, J=1.2, 1H, H6), 8.18 (d, J=14.6, 1H,
CH_CHCOOMe), 11.29 (d, J=5.1, 1H, NH); 13C NMR (DMSO-d6, 50.3 MHz): δ 12.6 (CH3), 51.8 (CH3), 107.4 (C),
110.7 (CH), 135.2 (CH), 137.6 (CH), 150.4 (CO), 163.2 (CO), 167.5 (CO); IR (KBr) ν 1750, 1715, 1650; CIMS m/z 228
[M+NH4]+.
15. To a suspension of 6 (71 mg, 0.090 mmol), Ph3P (50 mg, 0.19 mmol), and 10a (38 mg, 0.18 mmol) in dioxan (1–2 mL) was
added dropwise DEAD (28 µL, 31 mg, 0.18 mmol) in 1 mL of dioxan. After stirring for 1 h, the solvent was evaporated and
the products were separated by flash chromatography (from 95:5 CH2Cl2:MeOH to 95:5:2 CH2Cl2:MeOH:conc. ammonia)
to give (9S)-9-N-{3-[3-(E)-2-(methoxycarbonyl)vinylthymin-1-yl]propyl}erythromycylamine A (76 mg, 87%).

3375
16. Pyrrolidine (13 µL, 11 mg, 0.16 mmol) was added to a solution of Mocvinyl-protected 1a (39 mg, 0.04 mmol) in CH3CN
(0.4 mL). After stirring for 24 h, the solvent was removed and the residue was taken with CH2Cl2 and water. The aqueous
layer was acidified to pH 5.0, the layers were separated, and the aqueous phase was rinsed with CH2Cl2. Afterwards, the
aqueous solution was basified to pH 9.0 by addition of 2 M NaOH and was extracted three times with CH2Cl2. These
organic extracts were collected and dried over anhydrous Na2SO4. Filtration and removal of the solvent in vacuo afforded
chromatographically and spectroscopically pure (9S)-9-N-[3-(thymin-1-yl)propyl]erythromycylamine A, 1a (30 mg, 83%):
1H NMR (CDCl3, 300 MHz) δ 0.88 (t, J=7.5, 3H, Me-15), 1.01 (d, J=6.5, 3H, Me-8), 1.10 (d, J=6.5, 3H, Me-4), 1.11 (s,
3H, Me-12), 1.16 (d, J=7.0, 3H, Me-10), 1.20 (d, J=7.0, 3H, Me-2), 1.22 (d, J=6.0, 3H, Me-50), 1.24 (s, 3H, Me-6 or Me-
300), 1.26 (s, 3H, Me-300 or Me-6), 1.32 (d, J=6.5, 3H, Me-500), 1.46–1.58 (m, 2H, H14a, H7b), 1.58 (dd, J=15.3, J=5.1,
1H, H200a), 1.67 (ddd, J=12.6, J=3.8, J=2.1, 1H, H40b), 1.92 (d, J=1.0, 3H, Me-thym.), 1.87–1.99 (m, 4H, H4, H14b,
CH2CH2CH2), 2.10–2.27 (m, 4H, H8, H9, H10, 400-OH), 2.29 (s, 6H, NMe2), 2.38 (d, J=15.0, 1H, H200b), 2.44–2.55 (m,
2H, H30, NHCHaHb), 2.71–2.79 (m, 1H, NHCHaHb), 2.89 (dq, J=7.9, J=7.0, 1H, H2), 3.03 (t, J=8.7, 1H, H400), 3.25
(dd, J=10.3, J=7.1, 1H, H20), 3.32 (s, 3H, MeO-300), 3.50 (dqd, J=10.6, J=5.7, J=1.8, 1H, H50), 3.58 (d, J=7.5, 1H, H5),
3.67–3.83 (m, 2H, CH2-thym.), 3.85 (br s, 1H, H11), 4.05 (dq, J=9.1, J=6.3, 1H, H500), 4.20 (d, J=7.5, 1H, H3), 4.46 (d,
J=7.5, 1H, H10), 4.68 (dd, J=10.2, J=2.3, 1H, H13), 4.97 (d, J=4.5, 1H, H100), 7.11 (q, J=1.5, 1H, H6-thym.); 13C NMR
(CDCl3, 75.4 MHz): δ 9.3 (CH3), 11.1 (CH3), 12.3 (CH3), 15.2 (CH3), 16.6 (CH3), 16.8 (CH3), 18.5 (CH3), 21.4 (CH3),
21.5 (CH2), 21.5 (CH3), 21.7 (CH3), 26.6 (CH3), 28.6 (CH2), 29.3 (CH2), 29.6 (CH), 31.2 (CH), 35.0 (CH2), 36.5 (CH2),
39.9 (CH), 40.2 (CH3), 44.9 (CH), 45.9 (CH2), 46.2 (CH2), 49.4 (CH3), 65.4 (CH), 65.5 (CH), 69.0 (CH), 70.4 (CH),
70.9 (CH), 71.1 (CH), 72.6 (C), 74.0 (C), 76.1 (C), 77.9 (CH), 78.1 (CH), 79.6 (CH), 83.8 (CH), 96.2 (CH), 103.2 (CH),
110.7 (C), 140.4 (CH), 150.8 (CO), 164.1 (CO), 177.7 (CO); HRFABMS, calcd for C45H81N4O14: 901.5749 [M+1]; found:
901.5726.
17. Including the desired carbamate (FABMS 1176.7 [M+1]) but as a minor product. Reaction of 6 with crude isocyanate 13
gave similar mixtures.
18. Spectral data of 9a-aza-9a-[30-deoxythymidin-30-yl)aminocarbonyl]-9-deoxo-9a-homoerythromycin A (4): 1H NMR
(CD3OD, 500 MHz) δ 0.90 (t, J=7.3, 3H, Me-15), 1.00–1.02 (m, 6H, Me-8, Me-4), 1.19–1.32 (m, 23H, H40a, H7a,
Me-2, Me-6, Me-10, Me-12, Me-50, Me-300, Me-500), 1.42–1.54 (m, 2H, H14a, H7b), 1.58 (dd, J=15.0, J=5.0, 1H, H200a),
1.79–1.86 (m, 2H, H4, H40b), 1.85 (dqd, J=14.3, J=7.8, J=2.3, 1H, H14b), 1.89 (d, J=1.0, 3H, Me-thym.), 2.21 (br s, 1H,
H8), 2.29–2.39 (m, 2H, H20-thym., H200-thym.), 2.42 (d, J=16.0, 1H, H200b), 2.44 (br s, 6H, NMe2), 2.80 (quint, J=7.6,
1H, H2), 2.92 (t, J=9.9, 1H, H30), 3.04 (d, J=9.5, 1H, H400), 3.12 (br s, 1H, H9b or H10), 3.32 (s, 3H, MeO-300), 3.49 (d,
J=7.0, 1H, H5), 3.59 (br s, 1H, H11), 3.71 (dq, J=9.4, J=5.9, 1H, H50), 3.81 (dd, J=12.5, J=3.5, 1H, H50-thym.), 3.86 (dd,
J=12.0, J=2.5, 1H, H500-thym.), 4.00 (br s, 1H, H40-thym.), 4.08 (d, J=8.5, 1H, H3), 4.14 (dq, J =9.5, J=6.2, 1H, H500),
4.29 (q, J=6.3, 1H, H30-thym.), 4.51 (d, J=7.0, 1H, H10), 4.90 (d, J=5.0, 1H, H100), 5.14 (d, J=8.5, 1H, H13), 6.29 (t, J=6.0,
1H, H10-thym.), 7.88 (br s, 1H, H6-thym.); 13C NMR (CD3OD, 75.4 MHz): δ 10.3 (CH3), 11.8 (CH3), 12.5 (CH3), 13.4
(CH3), 16.4 (CH3), 18.7 (CH3), 19.1 (CH3), 20.8 (CH3), 21.6 (CH3), 21.8 (CH3), 23.4 (CH2), 29.0 (CH), 30.8 (CH2), 31.7
(CH2), 36.0 (CH2), 39.1 (CH2), 40.4 (CH3), 41.9 (CH), 46.7 (CH), 50.0 (CH3), 52.5 (CH), 53.6 (CH), 63.0 (CH), 63.0
(CH2), 65.6 (CH), 66.7 (CH), 68.9 (CH), 72.2 (CH), 74.3 (C), 76.1 (C), 76.4 (CH), 77.4 (C), 77.7 (CH), 79.2 (CH), 80.8
(CH), 85.9 (CH), 87.1 (CH), 87.2 (CH), 97.5 (CH), 104.3 (CH), 111.7 (C), 138.1 (CH), 153.2 (C), 160.8 (CO), 166.4 (CO),
178.1 (CO); HRFABMS, calcd for C48H85N5O17: 1003.5940 [M+1]; found: 1003.5900, calcd for C48H84N5O17: 1002.5862
[M]; found: 1002.5859.