Solid-phase synthesis of hydroxy-acids leading to macrolactones

Article abstract of DOI:10.1016/S0040-4039(00)00711-5

A sequence of five chemical steps on solid support yields hydroxy-acids in a very straightforward and efficient manner. The resulting compounds can be cyclized in solution phase by forming an ester bond to produce macrolactones. (C) 2000 Published by Elsevier Science Ltd.

Full text of DOI:10.1016/S0040-4039(00)00711-5

Tetrahedron Letters 41 (2000) 4751±4755
Solid-phase synthesis of hydroxy-acids leading to
macrolactones
Rene Gagnon, Yves L. Dory* and Pierre Deslongchamps*
Á
Â
Laboratoire de Synthese Organique, Departement de Chimie, Institut de Pharmacologie de Sherbrooke,
e
 Â
Universite de Sherbrooke, 3001, 12 Avenue Nord, Sherbrooke, Quebec J1H 5N4, Canada
Received 26 January 2000; revised 17 April 2000; accepted 18 April 2000
Abstract
A sequence of ®ve chemical steps on solid support yields hydroxy-acids in a very straightforward and
ecient manner. The resulting compounds can be cyclized in solution phase by forming an ester bond to
produce macrolactones. # 2000 Published by Elsevier Science Ltd.
Introduction. In our previous e€orts1 to develop a powerful methodology to build libraries of
macrocycles on solid support, the crucial macrocyclization step was carried out on solid support
and by forming a carbon±carbon bond. However, since original results indicated that the high
dilution conditions usually necessary to macrocyclize could not be met on solid support,2 then a
parallel research project was started to solve that problem. The alternative route consists in pre-
paring the cyclization precursors on solid support and then cyclizing them in solution; the
description of this particular work is the topic of this letter.
Design of the sequence. Macrolactones were chosen as targets for this project because they have
potential biological activity as shown by many members of the macrolide family,3 It was decided
to macrocyclize by forming an ester bond instead of a carbon±carbon bond as was usually done
in all our previous work.1,4 Since the macrolactonization was to be carried out in solution, it
ensues that hydroxy-acids had to be prepared on solid support and in as little chemical steps as
possible. To be able to accomplish this, it was necessary to reduce the number of functional group
deprotection steps to the minimum. In fact, beside the cleavage of the resin (which, to some
extend can be considered as functional group deprotection) there remained only one such step.
The formation of an hydroxy-acid requires ®ve steps on solid support as summarized in Scheme 1.
The ®rst step is the hooking of the ®rst building block to the Wang resin by esteri®cation. Then
the second building block is directly attached to the growing chain by SN2 displacement of a
* Corresponding authors. Fax: (819) 820-6823; e-mail: yves.dory@courrier.usherb.ca
0040-4039/00/$ - see front matter # 2000 Published by Elsevier Science Ltd.
PII: S0040-4039(00)00711-5

4752
Scheme 1.
benzylic chloride by means of a malonate anion.5 An alcohol is then set free and coupled to the
last building block by Mitsunobu coupling.6 This leaves only the cleavage of the hydroxy-acid
o€ the resin and eventually the lactonization of the resulting compound in solution. The chemistry
was ®ne-tuned in solution using 4-benzyloxy benzyl alcohol as a resin model, and the best conditions
found were adapted to solid-phase chemistry to produce two lactones 13 and 14 (Scheme 2).
Chemistry. For the solution-phase chemistry the Wang resin was replaced by p-benzyloxy
benzyl alcohol. The ®rst step consisted of the esteri®cation of that alcohol with the acid chloride
1, the ®rst building block, and with a yield of 93%. The resulting benzyl chloride in 2 was treated
with the anion of the malonate 3,5 the second building block. The SN2 displacement reaction
produced the desired THP ether 4 with a yield of 86%. The THP ether was then cleaved at step 3
with PTSA (yield: 79%) to produce the corresponding alcohol 5 necessary for the subsequent
Mitsunobu coupling. Thus, at step 4, the tosylamide 77 was added to the allylic alcohol 5 as well
as DEAD and triphenylphosphine. However, the desired compound 9 was only obtained with the
very disappointing yield of 40%. Because of this low yield, this reaction was never tried on solid
support. In order to get 9 with better results and to ®nd a safer route for solid-phase chemistry,
the allylic alcohol 5 was transformed into the allylic chloride 6 by means of N-chlorosuccinimide
and triphenylphosphine (yield: 70%). The allylic chlorine atom of 6 was then displaced in an SN2
fashion by the anion of the tosylamide 7 and with a good yield of 83%. Even in this alternative
route, the swiftness sought for was lost because the sequence was then one step longer than
originally planned. Consequently, the only parameter which was left for possible modi®cations is
the nature of the sulfonamide Xs on compound 7; the acidity of the NH proton was increased by
replacing the tosyl group by a nosyl (o-nitrobenzene sulfonyl) group.8 The resulting nosylamide
87 underwent Mitsunobu coupling with the alcohol 5 to produce 10 with the excellent yield of
95%. It was therefore unnecessary in the case of 8 to investigate the alternative route via the
allylic chloride 6. The next step allowed deprotection of both the alcohol and the carboxylic
acid functionalities. The benzyloxy benzyl ester group hydrolysis in solution was equivalent to the
cleavage from the Wang resin. Tri¯uoroacetic acid was used for that double purpose. The desired
hydroxy-acids 119 and 1210 were obtained together with various amounts of the corresponding
tri¯uoroacetates; hence the subsequent treatment of the reaction mixtures with methanol under
neutral bu€ered conditions to get them as pure compounds with yields of 90% in both cases. The
two hydroxy-acids 11 and 12 were then treated with DEAD and triphenylphosphine to provoke
their lactonization. The 19-membered lactones 1311 and 1412 were obtained with the respective
yields of 45 and 28% under these conditions. Thus, the hydroxy-acid 11 was obtained in ®ve steps
(Mitsunobu route) with an overall yield of 23%, while the alternative six-step route (SN2 via
allylic chloride 6) yielded the same compound 11 with a better overall yield of 33%. On the other
hand, the other ®ve-step route leading to 12 was much more successful since the overall yield was
54%.

4753
Scheme 2.
The same sequences, six steps in the case of 11 and ®ve steps for 12, were carried out on solid
support using the same conditions. The overall yields were 23 and 50% after ¯ash chromato-
graphy of the crudes. Therefore, the solid-phase routes were lower yielded than the solution-phase
routes by as much as 10% in the case of 11. The overall yield were evaluated from the original
loading of the Wang resin.
Conclusion. The chemistry described here is very ecient at producing hydroxy-acids. The method
seems rather robust so that it should be general and could also be applied to more elaborate
hydroxy-acids and even to amino acids by varying the nature of the third building block. However,

4754
the tosylamide group is not as good at coupling with alcohols under Mitsunobu conditions as
previously observed.1 In the current case the tosylamide is prepared from an isolated primary
amine; whereas in the past projects the tosylamides were always originating from a-amino-acids.
In that particular case, the a-carbonyl should obviously pull the electrons from the NH group
and its proton should be more acidic, and consequently the Mitsunobu coupling should be eased.
This positive e€ect cannot take place in the present work and the Mitsunobu coupling goes
poorly with the tosylamide group. The nosylamide group solves the problem completely since the
nitro group produces the same favorable electro-withdrawing e€ect to an even larger extent; it is
therefore the best choice as a sulfonamide if a library of hydroxy-acids like 11 and 12 was to be
prepared according to the chemistry shown. On the other hand the sequence may not be well
suited for the preparation of arrays of lactones like 13 and 14 because the extra step in solution
might be detrimental to ecient automation.
Acknowledgements
Financial support from Wyeth-Ayerst Research Laboratories and NSERC (Canada) through a
University±Industry research grant is greatly appreciated. We wish to thank Dr. Magid Abou-
Gharbia and Dr. John Ellingboe for stimulating discussions.
References
1. Ramaseshan, M.; Ellingboe, J. W.; Dory, Y. L.; Deslongchamps, P. Tetrahedron Lett. 2000, 41, 4743.
2. Crowley, J. I.; Rapoport, H. J. Am. Chem. Soc. 1970, 92, 6363.
3. Norcross, R. D.; Paterson, I. Chem. Rev. 1995, 95, 2041.
4. Deslongchamps, P. Aldrichimica Acta 1991, 24, 43; Pure Appl. Chem. 1992, 64, 1831.
5. Deslongchamps, P.; Lamothe, S.; Lin, H.-S. Can. J. Chem. 1987, 65, 56.
6. Mitsunobu, O. Synthesis 1981, 1.
7. 6-Aminohexan-1-ol was treated with TsCl and 2N NaOH in ethyl acetate and water to yield the corresponding
tosylamide (85%) whose alcohol was protected as a THP ether (PTSA and DHP) to give 7 in 68% yield.
Compound 8 was obtained in the same manner from 6-Aminohexan-1-ol and NsCl with a yield of 54% for the
two reactions.
8. Yang, L.; Chiu, K. Tetrahedron Lett. 1997, 38, 7307.
9. Hydroxy-acid 11 analytical data: 1H NMR ꢀ (CDCl3): 1.25 (4H, m, CH2CH2CH2CH2N), 1.49 (4H, m,
CH2CH2CH2CH2CH2N), 2.43 (3H, s, CH3-Ph-SO2), 2.46 (2H, d, J=5 Hz, CH2CHˆCHCH2N), 3.05 (2H, m,
CH2CH2N), 3.29 (2H, s, CH2ArCO2), 3.60 (2H, t, J=6 Hz, CH2OH), 3.72 (8H, br s, CHˆCHCH2N and OCH3),
5.44 (2H, m, CHˆCH), 7.11 (2H, d, J=8 Hz, Ar), 7.28 (2H, d, J=8 Hz, Ar), 7.65 (2H, d, J=8 Hz, Ar), 7.98 (2H,
d, J=8 Hz, Ar). 13C NMR ꢀ (CDCl3): 21.4, 25.1, 26.2, 28.2, 30.0, 32.3, 38.3, 44.7, 47.6, 52.7, 58.3, 62.5, 126.0,
127.0, 128.4, 129.2, 129.6, 129.9, 130.1, 136.8, 141.8, 143.2, 170.7. MS (CI, NH3): 557 [M^MeOH]+.
10. Hydroxy-acid 12 analytical data: 1H NMR ꢀ (CDCl3): 1.27 (4H, m, CH2CH2CH2CH2N), 1.48 (4H, m,
CH2CH2CH2CH2CH2N), 2.51 (2H, d, J=5 Hz, CH2CHˆCHCH2N), 3.23 (2H, m, CH2CH2N), 3.32 (2H, s, CH
2ArCO2), 3.61 (2H, t, J=6 Hz, CH2OH), 3.74 (6H, s, OCH3), 3.87 (2H, d, J=5 Hz, CHˆCHCH2N), 5.51 (2H, m,
CHˆCH), 7.15 (2H, d, J=8 Hz, Ar), 7.6±7.7 (4H, d, Ar), 8.00 (2H, d, J=8 Hz, Ar). 13C NMR ꢀ (CDCl3): 25.2,
26.2, 27.9, 30.2, 32.3, 38.5, 44.1, 47.4, 52.7, 58.4, 62.6, 124.2, 127.0, 128.5, 128.8, 129.9, 130.3, 130.6, 130.8, 131.7,
133.5, 141.8, 147.9, 170.8. MS (EI, 70 eV): 603 [M^OH]+.
11. Macrocycle 13 analytical data: 1H NMR ꢀ (CDCl3): 1.40 (4H, m, CH2CH2CH2CH2N), 1.50 (2H, m, CH2CH2N),
1.74 (2H, m, CH2CH2O), 2.21 (2H, br s, CH2CHˆCHCH2N), 2.41 (3H, s, CH3±Ph±SO2), 3.01 (2H, m,
CH2CH2N), 3.25 (2H, br s, CHˆCHCH2N), 3.36 (2H, s, CH2ArCO2), 3.81 (6H, s, OCH3), 4.40 (2H, m, CH2O),
5.17 (2H, m, CHˆCH), 7.07 (2H, d, J=8 Hz, Ar), 7.25 (2H, d, J=8 Hz, Ar), 7.62 (2H, d, J=8 Hz, Ar), 7.88 (2H,

4755
d, J=8 Hz, Ar). 13C NMR ꢀ (CDCl3): 21.5, 26.7, 28.0, 29.1, 29.5, 37.9, 46.0, 48.8, 53.0, 56.9, 66.3, 124.9, 127.1,
129.5, 129.8, 137.9, 140.9, 143.0, 165.8, 171.2. MS (EI, 70 eV): 571 [M]+, 540 [M^MeO]+. HRMS: calcd [M]+:
571.2240; found: 571.2233.
12. Macrocycle 14 analytical data: 1H NMR ꢀ (CDCl3): 1.3±1.55 (6H, m, CH2CH2CH2CH2N), 1.74 (2H, m, CH2CH2
O), 2.19 (2H, br s, CH2CHˆCHCH2N), 3.18 (2H, m, CH2CH2N), 3.36 (2H, br s, CHˆCHCH2N), 3.37 (2H, s,
CH2ArCO2), 3.81 (6H, s, OCH3), 4.42 (2H, t, J=5 Hz, CH2O), 5.19 (2H, m, CHˆCH), 7.09 (2H, d, J=8 Hz, Ar),
7.25 (2H, d, J=8 Hz, Ar), 7.5±7.7 (2H, m, Ar), 7.92 (2H, d, J=8 Hz, Ar). MS (EI, 70 eV): 602 [M]+. HRMS: calcd
[M]+: 602.1934; found: 602.1939.