Efficient access to ATP mimics, potential FGF receptor tyrosine kinase inhibitors

Article abstract of DOI:10.1016/S0040-4039(02)01923-8

Nucleophilic opening of bis-epoxide 1 derived from D-mannitol by organolithium species in the presence of boron trifluoride etherate allows the synthesis of enantiopure polyhydroxy tetrahydrofurane skeletons with aromatic moieties in the pseudo-anomeric p

Full text of DOI:10.1016/S0040-4039(02)01923-8

TETRAHEDRON
LETTERS
Pergamon
Tetrahedron Letters 43 (2002) 8129–8132
Efficient access to ATP mimics, potential FGF receptor tyrosine
kinase inhibitors
Se´verinne Rigolet, Isabelle McCort and Yves Le Merrer*
Universite´ Rene´ Descartes, Laboratoire de Chimie et Biochimie Pharmacologiques et Toxicologiques, UMR 8601,
45 rue des Saints-Pe`res, 75270 Paris Cedex 06, France
Received 24 July 2002; accepted 6 September 2002
Abstract—Nucleophilic opening of bis-epoxide 1 derived from
D-mannitol by organolithium species in the presence of boron
trifluoride etherate allows the synthesis of enantiopure polyhydroxy tetrahydrofurane skeletons with aromatic moieties in the
pseudo-anomeric position. ortho-Lithiation and transmetalation are the most appropriate conditions for the preparation of the
organometallic derivatives where halogen–metal exchange is unfavorable. © 2002 Elsevier Science Ltd. All rights reserved.
Fibroblast growth factor receptors (FGFRs) belong to
a large family of transmembrane proteins with intrinsic
tyrosine kinase activity. These receptors (FGFR1
through FGFR4) play key roles in the control of many
cellular processes including cell proliferation and differ-
entiation. Germinal missense mutations in FGFR3
which have been demonstrated to induce constitutive
activation of the receptor have been associated with
inherited human skeletal dwarfism.1 Recently certain
forms of cancer, in particular carcinomas (bladder and
cervix) have been ascribed to somatic mutations of
FGFR3 affecting the same amino acids as germinal
mutations.2
D-ribose of ATP and on the other hand to position the
aromatic group via a non-hydrolyzable bond. Further-
more, the primary alcohol function of the hydroxy-
methyl group in 5-position of the tetrahydrofurane
skeleton could be subsequently used to introduce a
carbon chain to bestow tissue specificity.
The retrosynthesis of the target compounds 2 is out-
lined in Fig. 1 and involves a one-pot tandem alkyla-
tion–cyclization of C2-symmetrical
derived from -mannitol by action of nucleophiles.
L
-ido bis-epoxide 1
D
We have previously reported, for the synthesis of gly-
cosidase inhibitors, the nucleophilic opening of bis-
epoxide 1 by amines leading to polyhydroxylated
azepane or piperidine by N-heterocyclization.4 In order
to favor the 5-exo-tet O-cyclization 2 (path a), with
regard to the 6-endo-tet O-cyclization 3 (path b) or
eventually to the formation of the acyclic 4 compound,
we investigated the opening of bis-epoxide 1 by hetero-
and carbon nucleophiles in aprotic medium.
To date, a number of tyrosine kinase inhibitors have
been synthesized,3 but none are reported to be efficient
at counteracting the effects of activated FGFR3 in
human disorders. Our goal was to design a new class of
tyrosine kinase inhibitors able to mimic ATP and to
specifically block the constitutive activation of FGFR3
in vitro and in vivo.
We report here preliminary results concerning the syn-
thesis of enantiopure polyhydroxy tetrahydrofurane
skeletons with an aromatic moiety in the pseudo-
anomeric position in order to occupy the adenine site.
In this context, the polyhydroxy tetrahydrofurane is
equivalent to a scaffold able on one hand to mimic the
Obtention of the halogen-substituted derivative
2
(Nu=Br) is attractive since it is the precursor towards
further transformations (Scheme 1). Dilithium
tetrabromonickelate5 reacted slowly on 1 at room tem-
perature, yielding mainly the tetrahydropyrane 3a ([h]D20
+8 (c 1.0, CH2Cl2)), whereas lithium bromide afforded
regiospecific ring cyclization promoted by a stoichio-
metric amount of BF3·Et2O leading to the desired tetra-
hydrofurane 2a ([h]2D0 +62 (c 1.04, CH2Cl2)).
Keywords: bis-epoxide; boron trifluoride etherate; lithiation;
transmetalation; tyrosine kinase.
At the same time, we explored the possibility to intro-
duce in one step an aromatic group at the pseudo-
* Corresponding author. Fax: (33) 01 42 86 83 87; e-mail: yves.
le-merrer@biomedicale.univ-paris5.fr
0040-4039/02/$ - see front matter © 2002 Elsevier Science Ltd. All rights reserved.
PII: S0040-4039(02)01923-8

8130
S. Rigolet et al. / Tetrahedron Letters 43 (2002) 8129–8132
Figure 1.
halogen–metal exchange with t-BuLi (2.3 equiv.),10 tet-
rahydrofurane 2c ([h]2D0 +63 (c 1.0, CH2Cl2)) was
obtained along with bromo derivative 2a in a 1:1
proportion.
Thus, the 5-exo-tet O-cyclization is always preferred
with organolithium reagents in the presence of boron
trifluoride etherate at low temperature, but the presence
of bromide in the reaction mixture must be completely
avoided. In our hands, different assays to eliminate
halogens from reaction medium either by decantation
or recrystallization of the organolithium failed.11 There-
fore to introduce various substituents at the pseudo-
anomeric position, we studied ortho-lithiation and
transmetalation from the corresponding organotin
derivatives.
Scheme 1.
anomeric position. Only symmetrical substitution
occurred with the p-methoxyphenyl organocuprous6
leading to 4a, when the reaction was carried out with
either 1 or 2 equiv. of nucleophile with 45 and 73%
yields, respectively.
Anisole was easily converted to the corresponding
ortho-lithiate by n-BuLi10b in the presence of
N,N,N%,N%-tetramethylethylenediamine (Scheme 3).12 In
these conditions the reaction with 1 was carried out
with an excess of BF3·Et2O to avoid total complexation
In order to circumvent the bis-opening, we turned our
attention towards organolithium species catalyzed by
boron trifluoride etherate (1.5 equiv.)7 which both
increase steric hindrance around the epoxide and inhibit
activation of a second addition. Thus, highly selective
mono-opening was disclosed from C2-symmetric bis-
epoxide derived from pentane,8 for which evolution by
cyclization was unfavorable according to the Baldwin’s
rules.9 These conditions seemed very attractive to us
since BF3·Et2O did not prevent heterocyclization as
shown before for 2a. Indeed, from bis-epoxide 1, hete-
rocyclization always proceeds faster than a second
addition of the nucleophile and bis-opening leading to
the acyclic compound 4 has never been observed.
Commercial phenyl lithium reacted quickly with bis-
epoxide 1 at −78°C yielding to tetrahydrofurane 2b
([h]2D0 +67 (c 1.0, CH2Cl2)) (Scheme 2). When organo-
lithium was prepared from p-bromoanisole by classical
Scheme 3.
Scheme 2.

S. Rigolet et al. / Tetrahedron Letters 43 (2002) 8129–8132
8131
Scheme 4.
by TMEDA and led to 2d ([h]2D0 +56 (c 1.0, CH2Cl2))
with good yield.13 To our knowledge, this is the first
example showing the compatibility of these two
reagents.
Research. The authors are grateful to Drs. L. Legeai-
Mallet, J. Bonaventure and F. Radvanyi for fruitful
discussions concerning FGFR3.
After a-lithiation of N-methyl indole (3 equiv.) with
n-BuLi,14 addition to bis-epoxide 1 led to 2e ([h]2D0 +79
(c 1.0, CH2Cl2)) with 50% yield which has been
improved to 85% by a second addition of a-lithiate
solution (2 equiv.).
References
1. Bonaventure, J.; Rousseau, F.; Legeai-Mallet, L.; Le
Merrer, M.; Munnich, A.; Maroteaux, P. Am. J. Med.
Genet. 1996, 63, 148–154.
2. Cappellen, D.; de Oliveira, C.; Ricol, D.; de Medina, S.;
Bourdin, J.; Sastre-Garau, X.; Chopin, D.; Thiery, J.-P.;
Radvanyi, F. Nature Genet. 1999, 23, 18–20.
Transmetalation of commercial tributylphenyltin by n-
BuLi in the presence of TMEDA15 followed by addi-
tion on 1 afforded 2b in 79% yield (Scheme 4).16 We
thus applied this method to other organotin species
which were obtained from the corresponding bromo
aromatic derivatives17 and easily purified by flash
chromatography18 preventing any contamination of by-
product 2a during the opening of bis-epoxide 1.
3. (a) Levitski, A.; Gazit, A. Science 1995, 1782–1788; (b)
Groundwater, P. W.; Solomons, K. R. H.; Drewe, J. A.;
Munawar, M. A. Progr. Med. Chem. 1996, 25, 233–329;
(c) Boschelli, D. H. Drugs Fut. 1999, 24, 515–537.
4. (a) Le Merrer, Y.; Poitout, L.; Depezay, J.-C.; Dosbaa,
I.; Geoffroy, S.; Foglietti, J.-J. Bioorg. Med. Chem. 1997,
5, 519–533; (b) Le Merrer, Y.; Sanie`re, M.; McCort, I.;
Dupuy, C.; Depezay, J.-C. Tetrahedron Lett. 2001, 42,
2661–2663.
5. Dawe, R. D.; Molinski, T. F.; Turner, J. V. Tetrahedron
Lett. 1984, 25, 2061–2064.
6. Ghosh, A. K.; McKee, S. P.; Thompson, W. J. Tetra-
hedron Lett. 1991, 32, 5729–5732.
7. Anhydrous, redistilled BF3·Et2O was purchased from
Aldrich.
Whereas total deprotection of hydroxyl groups (Scheme
5) was achieved for 2b, 2c and 2d with hydrogen/Pd
black in AcOH furnishing 5b–5d with good yields (82–
94%), deprotection of 2e in these conditions failed.
However, catalytic transfer hydrogenation with ammo-
nium formate in the presence of Pd/C 10% in MeOH
successfully furnished 5e with 70% yield.19
8. Rychnovsky, S. D.; Griesgraber, G.; Zeller, S.; Skalitzky,
D. J. J. Org. Chem. 1991, 56, 5161–5169.
9. Baldwin, J. E. J. Chem. Soc., Chem. Commun. 1976,
734–736.
10. (a) Oshiki, T.; Imamoto, T. Bull. Chem. Soc. Chim. Jpn.
1990, 63, 3719–3721; (b) n-BuLi and t-BuLi were always
used as a pentane solution.
Scheme 5. Reagents and conditions: (i) From 2b–2d: H2, Pd
black, AcOH, 15 h; from 2e: Pd/C 10%, HCO2NH4, MeOH,
reflux, 5 h.
11. Schlosser, M.; Ladenberger, V. J. Organomet. Chem.
1967, 8, 193–197.
In conclusion, we have determined the optimal condi-
tions to obtain enantiopure polyhydroxy tetra-
hydrofurane skeletons with aromatic moieties in the
pseudo-anomeric position by nucleophilic opening of
12. (a) Slocum, D. W.; Moon, R.; Thompson, D. S.; Li, J.
D.; Slocum, M. G.; Siegel, A.; Gayton-Garcia, R. Tetra-
hedron Lett. 1994, 35, 385–388; (b) Rennels, R. A.; Mali-
akal, A. J.; Collum, D. B. J. Am. Chem. Soc. 1998, 120,
421–422.
13. Typical experimental procedure for ortho-lithiation: To a
solution of n-BuLi (360 mL, 1.7 M in pentane, 0.612
mmol) in ether (1 mL) and TMEDA (93.3 mL, 0.612
mmol) under argon was added anisole (33.6 mL, 0.306
mmol). After stirring for 30 min, the yellow reaction
mixture was added via cannula to a solution of bis-epox-
ide (50 mg, 0.153 mmol) in THF (1 mL) and BF3·Et2O
(120 mL, 0.947 mmol) at −55°C. After stirring for 5 min,
the reaction was quenched with saturated NaHCO3 and
the solution was warmed to room temperature. The
mixture was then extracted with EtOAc, dried (Na2SO4)
bis-epoxide 1 derived from
D-mannitol by organo-
lithium derivatives followed by hydrogenolysis with
yields up to 50% for the two steps. We are currently
applying the methodology to provide a flexible route to
various aromatic and heteroaromatic substituents to
obtain FGFR3 tyrosine kinase inhibitors.
Acknowledgements
This work is supported by the ACI program ‘Mole´cules
et Cibles The´rapeutiques’ of the French Ministry of

8132
S. Rigolet et al. / Tetrahedron Letters 43 (2002) 8129–8132
18. Purification of organotin species by flash chromatogra-
and concentrated under reduce pressure. Flash chro-
phy using cyclohexane/EtOAc as solvent required the
matography (cyclohexane/EtOAc, 75:25, Rf=0.2),
afforded 2d (58.2 mg, 87%) as a colorless oil: H NMR
(250 MHz, CDCl3) l 2.38 (brs, 1H, OH), 3.0–3.2 (m, 2H,
H1), 3.6–3.9 (m containing s at 3.78, 6H, CH3, H3, H6),
1
presence of 2% Et3N.
19. Physical data of final compounds: 5b ([h]2D0 +24 (c 1.0,
1
MeOH); H NMR (250 MHz, MeOD) l 2.89 (dd, J1,1%
=
13.6 Hz, J1,2=7.0 Hz, 1H, H1), 3.00 (dd, J1%,1=13.6 Hz,
1%,2=7.0 Hz, 1H, H1%), 3.6–3.8 (m, 4H, H3, H5, H6), 3.99
4.0–4.1 (m, 2H, H4, H5), 4.32 (dt, J2,3=3.5 Hz, J2,1
=
J
J2,1%=7.0 Hz, 1H, H2), 4.39, 4.58 (AB, JAB=11.5 Hz, 2H,
CH2Ph), 4.48, 4.51 (A%B%, JA%B%=11.7 Hz, 2H, CH2Ph),
6.8–7.0 (m, 2H, Harom), 7.2–7.4 (m, 12H, Harom); CIMS
(NH3) m/z: 452 (M+18, 100%).
(brs, 1H, H4), 4.15 (dt, J2,3=3.0 Hz, J2,1=J2,1%=7.0 Hz,
1H, H2) 7.1–7.4 (m, 5H, Harom); 13C NMR (63 MHz,
MeOD) l 37.8 (C1), 63.6 (C6), 78.3, 80.4, 84.1, 87.7 (C2,
C3, C4, C5), 127.1, 129.3, 130.2 (CHarom), 140.1 (Cqarom);
HRMS calcd for C12H17O4 (MH+) 225.1127, found
14. Labadie, S. S.; Teng, E. J. Org. Chem. 1994, 59, 4250–
4254.
15. (a) Peterson, D. J. J. Am. Chem. Soc. 1971, 4027–4030;
(b) Rychnovsky, S. D. J. Org. Chem. 1989, 54, 4982–
4984.
225.1124. 5c ([h]2D0 +26.5 (c 1.13, MeOH); H NMR (250
1
MHz, MeOD) l 2.83 (dd, J1%,1=13.7 Hz, J1,2=6.8 Hz,
1H, H1), 2.94 (dd, J1,1%=13.7 Hz, J1%,2=7.3 Hz, 1H, H1%),
3.6–3.8 (m containing s at 3.74, 7H, CH3, H3, H5, H6),
3.98 (brs, 1H, H4), 4.10 (dt, J2,3=3.0 Hz, J2,1=J2,1%=7.1
16. Typical experimental procedure for transmetalation: To a
solution of n-BuLi (360 mL, 1.7 M in pentane, 0.306
mmol) in THF (1 mL) and TMEDA (47 mL, 0.31 mmol)
under argon was added tributylphenyltin (155 mL, 0.46
mmol) at −78°C. After stirring for 30 min at −20°C, the
reaction mixture was added via cannula to a solution of
bis-epoxide (50 mg, 0.153 mmol) in THF (2 mL) and
BF3·Et2O (150 mL, 1.184 mmol) at 55°C. After stirring
for 5 min, the reaction was quenched with saturated
NaHCO3 and the solution was warmed to room tempera-
ture. The mixture was then extracted with EtOAc, dried
(Na2SO4) and concentrated under reduce pressure. Flash
chromatography (toluene/EtOAc, from 1:0 to 4:1, Rf=
0.3), afforded 2b (49 mg, 79%) as a colorless oil: 1H
NMR (250 MHz, CDCl3+D2O) l 3.06 (d, J1%,2=7.0 Hz,
2H, H1), 3.67 (dd, J6,6%=11.6 Hz, J6,5=3.8 Hz, 1H, H6),
3.74 (d, J3,2=3.2 Hz, 1H, H3), 3.81 (dd, J6%,6=11.6 Hz,
3
Hz, 1H, H2), 6.81, 7.20 (2d, J=8.5 Hz, 4H, Harom); 13C
NMR (63 MHz, MeOD) l 34.4 (C1), 55.3 (CH2), 63.2
(C6), 77.8, 80.1, 83.9, 87.3 (C2, C3, C4, C5), 114.4, 130.9
(CHarom), 131.7, 159.2 (Cqarom); HRMS calcd for
C13H19O5 (MH+) 255.1232, found 255.1231. 5d ([h]2D0
+37.5 (c 0.32, MeOH); 1H NMR (250 MHz, MeOD) l
2.94 (dd, J1,1%=13.4 Hz, J1,2=6.7 Hz, 1H, H1), 2.98 (dd,
J1%,1=13.4 Hz, J1%,2=7.0 Hz, 1H, H1%), 3.6–3.9 (m contain-
ing s at 3.81, 7H, CH3, H3, H5, H6), 3.98 (brs, 1H, H4),
4.20 (dt, J2,3=2.9 Hz, J2,1=J2,1%=6.8 Hz, 1H, H2), 6.8–
7.0 (m, 2H, Harom), 7.1–7.3 (m, 2H, Harom); 13C NMR (63
MHz, MeOH) l 29.8 (C1), 55.5 (CH3), 63.3 (C6), 78.1,
80.1, 82.1, 87.3 (C2, C3, C4, C5), 111.1, 121.2, 128.4, 131.8
(CHarom), 127.7, 158.7 (Cqarom); HRMS calcd for
C13H19O5 (MH+) 255.1232, found 255.1237. 5e ([h]2D0 +32
1
(c 1.0, MeOH); H NMR (250 MHz, MeOD) l 3.05 (dd,
J
J
J
6%,5=3.0 Hz, 1H, H6%), 4.0–4.1 (m, 2H, H4, H5), 4.27 (dt,
2,3=3.2 Hz, J2,1=7.1 Hz, 1H, H2), 4.38, 4.58 (AB,
AB=11.5 Hz, 2H, CH2Ph), 4.51 (s, 2H, CH2Ph), 7.2–7.4
J1,1%=15.4 Hz, J1,2=7.0 Hz, 1H, H1), 3.11 (dd, J1%,1=13.6
Hz, J1%,2=7.0 Hz, 1H, H1), 3.64–3.74 (m, 2H, H6), 3.68 (s,
3H, CH3), 3.79–3.83 (m, 1H, H5), 3.87 (d, J3,2=2.4 Hz,
1H, H3), 4.02 (brs, 1H, H4), 4.51 (ddd, J2,3=2.4 Hz,
(m, 15H, Harom); 13C NMR (63 MHz, CDCl3) l 34.6
(C1), 63.2 (C6), 71.3, 71.8 (CH2Ph) 82.2, 82.5, 82.7, 84.7
(C2, C3, C4, C5), 126.2, 127.5, 127.9, 128.3, 128.5, 129.1
(CHarom), 137.3, 137.5, 138.4 (Cqarom); CIMS (NH3) m/z:
422 (M+18, 100%).
J2,1=7.0 Hz, J2,1%=6.7 Hz, 1H, H2), 6.8 (s, 1H, Harom),
7.2–7.4 (m, 4H, Harom); 13C NMR (63 MHz, MeOD) l
26.9 (C1), 29.8 (CH3), 63.6 (C6), 78.5, (C3), 80.4 (C4), 82.3
(C2), 87.7 (C5), 109.8, 120.0, 120.5, 121.5 (CHarom), 129.4,
138.8, 139.4 (Cqarom); HRMS calcd for C15H20NO4
(MH+) 278.1392, found 278.1397.
17. Azizian, H.; Eaborn, C.; Pidcock, A. J. Organomet.
Chem. 1981, 215, 49–58.