Design and synthesis of macro-heterocycles structurally related to tirofiban

Article abstract of DOI:10.1016/S0040-4039(00)00709-7

Sequences of nine and four steps were designed to yield two tirobifan related macrocycles by SN2 macrocyclization. The cyclic compounds contain the amino acid tyrosine, one insaturation and are either 18- or 20-membered rings. All the reaction conditions are very mild and the overall yields indicate that the results depend a lot on the structure of the targets and on the length of the reaction sequence. (C) 2000 Published by Elsevier Science Ltd.

Full text of DOI:10.1016/S0040-4039(00)00709-7

Tetrahedron Letters 41 (2000) 4737±4742
Design and synthesis of macro-heterocycles structurally
related to tiro®ban
Mahesh Ramaseshan,a Martin Robitaille,a John W. Ellingboe,b
Yves L. Dorya,* and Pierre Deslongchampsa,*
a
Á
Â
Laboratoire de Synthese Organique, Departement de Chimie, Institut de Pharmacologie de Sherbrooke,
e
Â
Â
Universite de Sherbrooke. 3001, 12 Avenue Nord, Sherbrooke, Quebec, J1H 5N4, Canada
bWyeth-Ayerst Research, Division of Chemical Sciences, Pearl River, NY 10965, USA
Received 26 January 2000; revised 17 April 2000; accepted 18 April 2000
Abstract
Sequences of nine and four steps were designed to yield two tirobifan related macrocycles by SN2
macrocyclization. The cyclic compounds contain the amino acid tyrosine, one insaturation and are either
18- or 20-membered rings. All the reaction conditions are very mild and the overall yields indicate that the
results depend a lot on the structure of the targets and on the length of the reaction sequence. # 2000
Published by Elsevier Science Ltd.
1. Introduction
A class of organic structures with outstanding pharmaceutical activity has been termed the
macrocycle family. Compounds like Epothilone,1 Erythromycin, Neocarzinostatin, Rifampin and
Amphotericin are under clinical study or already marketed drugs and belong to this important
family. Although very diverse2 in terms of chemical nature, the macrocycles share a very important
property: restricted conformational mobility. Once bioactivity is found, the search for the active
conformation is eased. With that important aspect in mind we became interested in the design
and synthesis of speci®c macrocycles that could be used to ®nd very useful conformational
information on very ¯exible lead compounds or even marketed drugs. Some of our past research
dealt precisely with the construction of macro-carbocycles3 from various sets of building blocks.
In principle this type of building block approach should be easily adapted to SPOC (solid phase
organic chemistry) synthesis and potentially useful macrocyclic sca€olds could be obtained in a
combinatorial fashion.
N-(n-Butanesulfonyl)-O-(4-(4-piperidinyl)-butyl)-(S)-tyrosine 1 (Scheme 1) is a potent glyco-
protein (GP)IIb/IIIa blocker4 also known as Tiro®ban and marketed under the name Aggrastat.
* 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)00709-7

4738
It is an anti-platelet aggregation agent used in various vascular pathologies. The construction of
conformationally more rigid analogs of this very ¯exible molecule could permit the determination
of the pharmacophoric conformation and eventually lead to even more potent lead compounds.
This letter describes the synthesis of macro-heterocycles 2 and 3 built around the trifunctional
a-amino tyrosine (Scheme 2) and that share many common features with Tiro®ban 1.
Scheme 1.
Scheme 2.
Use of tyrosine. The malonate unit was successfully used as a ®rst connector in all our previous
investigation series.5 However, we looked for alternative ®rst connectors that bear at least three
functional groups. This could allow straightforward hooking to a resin (e.g. Wang), and sequential
alkylation or acylation of the two remaining sites. Several amino acids ®ll that bill and have the
advantage of being part of numerous natural ligands. For the present work, dealing with the
synthesis of macrocycles related to Tiro®ban, tyrosine was chosen and was protected in a suitable
way (4)6 (Scheme 3).
Solution chemistry. The two target macrocycles 2 and 3 were synthesized during this study.
Mild N- and O-alkylation methodologies (Mitsunobu7 couplings) were used to assemble the
building units. The macrocycles were formed at a late stage by malonate anion reaction on allylic
chlorides.8 Macrocycles 2 and 3 were both prepared according to two sequences: a long sequence
requiring nine steps and a short four step sequence as summarized in Scheme 2 and detailed in
Schemes 3±5.
Long multistep synthesis. The protected tyrosine 46 was subjected to Mitsunobu coupling with
the alcohol 58 to obtain 6 (75%) (Scheme 3). After the deprotection of the phenol with PTSA and
methanol in dichloromethane (71%), 7 was subjected to another Mitsunobu reaction with both
the diols 89 and 9;10 the yields of 10 and 11 were again high (about 80%). After THP cleavage
(step 4) (around 75%), the alcohols 12 and 13 were converted to the corresponding iodides 14 and
15 with DEAD, Ph3P and MeI in toluene (66% and 75%). With these iodides in hand, the next

4739
Scheme 3.
step 6 involved the formation of the malonates 16 and 17 under standard conditions11 and in
good yields (85%). After deprotection of the silyl group by means of TBAF (step 7) (82 and
71%), the resulting alcohols 18 and 19 were chlorinated by means of HCA and Ph3P in good
yields (around 70%) to give the precursors of cyclization 20 and 21.
Short multistep synthesis. A new approach to build the two allylic chlorides 20 and 21 already
synthesized was investigated with the optic of reducing the length of the sequence (i.e. three steps
only, rather than eight) (Scheme 4). This was made possible by removing non-productive steps

4740
Scheme 4.
Scheme 5.
like group deprotections and also by performing four instead of ®ve connections. The latter
modi®cation was a consequence of the decrease in the number of building blocks (the new third
building block being in fact a fusion of the third and fourth blocks used at ®rst). Since recourse to
anion chemistry was not necessary before the SN2 macrocyclization step, the allylic chloride
functionality could be introduced at an earlier stage than before by means of the very mild
Mitsunobu alkylation conditions. This usually highly reactive allylic chloride could then survive
all subsequent reactions until the ring formation step due to the general mildness of the whole
sequence. Thus, trans-1-chloro-2-penten-5-ol 2212 was coupled to the tosylate 4 under Mitsunobu
conditions with DIAD and with a yield of 77% (step 1). Step 2 consisted in the cleavage of the
THP ether by means of PTSA in a 1:1 mixture of dichloromethane and methanol (77%). The
resulting phenol 24 was transformed into the same two cyclization precursors 20 and 21 as in the

4741
®rst approach and by means of Mitsunobu couplings with the alcohols 2513 and 2614 (yields of 62
and 66%, respectively).
Macrocyclization. The macrocyclization (step 9 for the long sequence and step 4 for the short
sequence) was carried out at 75ꢀC by simply adding Cs2CO3 in acetonitrile and THF to the allylic
chloride precursors 20 and 21 and produced the macrocycles 2 and 3 with yields of 5 and 28%.
These low yields indicate that the SN2 transition structures leading to the macrocycles 2 and 3 are
strained despite the presence of many insaturations. It can be observed that some tension is
relieved when the macrocycle becomes larger since the 18-membered ring 2 is almost not present
(the reaction mixture is very complex and 2 could only be detected by LC-MS) whereas the 20-
membered ring compound 315 is obtained in substantial amount.
Solid phase chemistry. The protected tyrosine 276 (see Scheme 3) was coupled to the hydroxy-
methyl resin by means of DIC and DMAP in dichloromethane. All the solution phase chemistry
described above (long and short sequences) was then repeated on solid support and the cleavage
step was carried out by means of sodium methoxide. The long sequence yielded none of the
expected products and 2 was not obtained either through the short sequence. On the other hand
the 20-membered macrocycle 3 was formed with an overall yield of 5% with the short sequence
protocole.
Conclusion. The aim of this work was twofold: Our primary goal was to develop a simple and
versatile method to build macro-heterocycles useful to investigate the bioactive conformation of
very ¯exible leads or drugs like Tiro®ban. We were also interested in studying both the e€ect of
the length of the reaction sequences and of the geometrical characteristics of the target macro-
cycles on the overall yields. Two macrocycles 2 and 3 of 18- and 20-members had been selected
for this work. They proved to be dicult cases because only the larger less strained macrocycle 3
could be obtained in a very clean way both in solution and also on solid support. On the whole,
the technology that we have developed to build macro-heterocycles of biological interest is very
e€ective because it takes only four steps to build those molecules. Moreover, it allows rapid
construction of many macrocyclic derivatives around a particular theme, because it takes
advantage of the fast assembly of three building blocks. Consequently, the construction of
libraries of macrocyclic materials in an automated way and on solid support is a feasible task. In
the present work we showed that tyrosine can introduce a lot of strain even inside rather large
ring systems when the amine and the phenol groups are used as tethers. In similar cases it would
be advisable to build libraries of larger macrocycles or to modify the adverse trans geometry of
the alkene building block 5 and 22.
Acknowledgements
Financial support from Wyeth-Ayerst Research Laboratories and NSERC (Canada) through a
University-Industry research grant is greatly appreciated.
References
1. Nicolaou, K. C.; Roschangar, F.; Vourloumis, D. Angew. Chem. Int. Ed. 1998, 37, 2014.
2. DeWitt, S. H. H.; Kiely, J. K.; Stankovic, C. J.; Schroeder, M. C.; Cody, D. M. R.; Pavia, M. R. Proc. Natl. Acad.
Sci. USA 1993, 90, 6909.
3. Deslongchamps, P. Aldrichimica Acta 1991, 24, 43; Pure Appl. Chem. 1992, 64, 1831.

4742
4. Egbertson, M.; Chang, C. T.-C.; Duggan, M. E.; Gould, R. J.; Halczenko, W.; Hartman, G. D.; Laswell, W. L.;
Lynch Jr, J. J.; Lynch, R. J.; Manno, P. D.; Naylor, A. M.; Prugh, J. D.; Ramjit, D. R.; Sitko, G. R.; Smith, R. S.;
Turchi, L. M.; Zhang, G. J. Med. Chem. 1994, 37, 2537.
5. Ndibwami, A.; Lamothe, S.; Guay, D.; Plante, R.; Soucy, P.; Golstein, S. Deslongchamps, P. Can. J. Chem. 1993,
71, 695.
6. Tyrosine was treated with tosyl chloride and 1 M NaOH in dioxane, then NaOH in re¯uxing ethanol to yield
Ts-Tyr-OH (57%). Diazomethane treatment in ether gave the corresponding methyl ester quantitatively. The phenol
group was protected with DHP and PPTS in dichloromethane (98%) to give 4. Finally the methyl ester was
hydrolyzed with 1 M NaOH to yield 27 (99%).
7. Mitsunobu, O. Synthesis 1981, 1.
8. 1,3-Propanediol was monoprotected with DHP and PTSA (96%); the remaining alcohol was oxidized to the
aldehyde using Swern conditions (96%). The resulting aldehyde was treated with the anion of triethyl
phosphonoacetate to give the corresponding trans acrylate (99%) whose ester was reduced with DIBAH (84%).
The resulting alcohol was protected by means of TBDPSCl and imidazole in THF (98%) and the THP ether was
cleaved with PPTS in isopropanol at room temperature to give the alcohol 5 (98%).
9. 1,4-Butanediol was monoprotected with DHP and PTSA (73%) to give 8.
10. 1,6-Hexanediol was monoprotected with DHP and PTSA to give 9 (61%).
11. Brillon, D.; Deslongchamps, P. Tetrahedron Lett. 1986, 27, 1131.
12. The allylic alcohol obtained after DIBAH reduction (see Ref 8) was treated with hexachloroacetone, Ph3P and 2,6-
lutidine in THF to yield the corresponding allylic chloride (87%). The THP ether was cleaved by means of PPTS
in dichloromethane and methanol (1:1) to give 22 (91%).
13. The alcohol 8 was treated with MsCl and Et3N in dichloromethane to give the corresponding mesylate
quantitatively. The mesylate was transformed into the corresponding iodide by means of NaI in re¯uxing
acetone (90% yield). The iodide was treated with sodium dimethyl malonate in THF to yield the corresponding
substituted malonate (98%) whose THP ether was subsequently cleaved with PPTS in dichloromethane and
methanol (3:1) to give 25 in 96% yield.
14. The alcohol 9 was transformed into the corresponding iodide with DEAD, Ph3P and MeI in toluene (84%).
Subsequent treatment of that iodide with sodium dimethyl malonate in THF yielded quantitatively the
corresponding substituted malonate whose THP ether was cleaved with PPTS in dichloromethane and methanol
(3:1) to give 26 in 85% yield.
1
15. Macrocycle 3 analytical data: H NMR (CDCl3): ꢀ 1.03 (m, 2H), 1.25 (m, 2H), 1.42 (m, 2H), 1.65 (m, 4H), 2.43
(bs, 3H), 2.51 (m, 4H), 2.92 (m, 1H), 3.25 (m, 3H), 3.63 (s, 3H), 3.68 (s, 3H), 4.07 (t, 3H, J=6.2 Hz), 4.32 (m, 1H),
4.85 (m, 1H), 4.94 (m, 1H), 6.78 (d, 2H, J=8.79 Hz), 7.06 (d, 2H, J=8.51 Hz), 7.27 (m, 2H), 7.74 (d, 2H, J=8.24
Hz). 13C NMR (CDCl3): ꢀ 21.92, 23.23, 24.93, 28.52, 28.69, 31.30, 35.29, 36.66, 48.94, 52.55, 52.70, 57.26, 63.13,
67.38, 115.23, 125.99, 127.91, 128.95, 129.58, 130.53, 131.02, 171.20, 171.88. MS (EI, 70 eV): 630 [M+H]+.