(3 mM)9 of 14 in THF at reflux, with periodic addition of
more LiBH4 and MeOH over several days to give 91% of
the desired amino alcohol 16, 2% recovered 14a, and no 8
or pyrrolidine.7,10 The absence of the pyrrolidine is probably
due to the electron-withdrawing effect of the bromine group,
which favors cleavage of the tetrahedral intermediate to the
aldehyde and retards formation of the imine that is the
pyrrolidine precursor.
Scheme 3. Synthesis of Triamine Ester 22
p-Bromoaniline was used to optimize methoxycarbonyl-
ation conditions. Reaction of p-bromoacetanilide, Pd(OAc)2
(4%), PPh3 (16%), CO (60 psi), and NaOAc (5 equiv) in
MeOH at 120-130 °C for 24 h gives 80% of the desired
methyl ester. A similar reaction of p-bromotrifluoroacet-
anilide affords 95% of p-bromoaniline and <1% of the
desired methyl ester, indicating that the trifluoroacetamide
is not stable to carbonylation conditions. p-Bromoaniline is
also not methoxycarbonylated under these conditions. We
therefore attempted to acetylate amino alcohol 16.
Reaction of 16 with AcCl (3 equiv) and Et3N (5 equiv) in
CH2Cl2 only acetylates the alcohol. However, reaction of
16, AcCl (10 equiv), and Et3N (20 equiv) in CH2Cl2 affords
79% of acetoxy acetoacetamide 17 as a mixture of keto and
enol tautomers. The structure of 17 was verified by aceto-
acetylation of the acetate of 16 with diketene in CH2Cl2,
which yields the identical material in lower yield. Reaction
of AcCl and Et3N is known to give diketene, and products
of addition of diketene have occasionally been isolated from
reactions with AcCl and Et3N.11 The aniline nitrogen of 16
is very hindered and does not react with acetyl chloride but
is acylated with the unhindered diketene. This result is similar
to the observation that the side chain of taxol can be
introduced more easily by acylation with a â-lactam than
an acid chloride.12
Reaction of 17, Pd(OAc)2 (10%), PPh3 (40%), CO (60
psi), and NaOAc (5 equiv) in dry MeOH at 120-130 °C
for 3 d provides a 3:1 mixture of amino alcohol 18 and the
corresponding acetate. Hydrolysis of the acetate is completed
by heating the crude mixture and 5 equiv of NaOAc in dry
MeOH at 120 °C for 24 h to give amino alcohol 18 in 72%
yield (see Scheme 3). Conversion of the bromide of 17 to a
methyl ester makes the aniline a better leaving group, so
NaOAc cleaves the acetoacetamide under the reaction
conditions after carbonylation.
MsCl and pyridine in CH2Cl2 affords the crude mesylate,
which is treated with 10 equiv of NaN3 in dry DMF to give
54% of crude trifluoroacetamide azide 20. Hydrolysis of
trifluoroacetamide 20 with NaOMe in dry CH2Cl2 provides
pure aniline 21 in 44% yield from 19. Hydrogenation and
hydrogenolysis of 21 over Pd(OH)2 under H2 (1 atm) in 20:1
MeOH/concentrated HCl gives 84% of the hydrochloride salt
of 22 that is 90-95% pure.13 The synthesis of fully
functionalized tricyclic triamine 22 was accomplished in 11
steps in 5% overall yield from methyl 2-amino-5-bromo-
benzoate (10).
Introduction of the guanidine onto the hindered secondary
amine of 22 is difficult, and the standard methods fail. In
his synthesis of martinellic acid, Ma developed a novel
AgNO3-catalyzed reaction with N-(Boc)-N-(3-methyl-2-
butenyl)-S-methylisothiourea.3 We recently reported a general
new method for the preparation of hindered guanidines.14
Treatment of a hindered amine with cyanogen bromide and
NaHCO3 in EtOH gives the cyanamide in virtually quantita-
tive yield. Reaction with a second amine in the polar, non-
nucleophilic solvent hexafluoro-2-propanol in a sealed tube
at 90-120 °C forms the guanidine in >80% yield.
Reaction of amino alcohol 18, TFAA, and Et3N in CH2-
Cl2 as in the protection of 8 provides trifluoroacetamide 19
in low yield presumably due to ester hydrolysis catalyzed
by Et3N. Reaction of 18, 10 equiv of TFAA, and 20 equiv
of the weaker base pyridine in CH2Cl2 for 12 h yields 74%
of trifluoroacetamide alcohol 19 after hydrolysis of the
trifluoroacetate ester during workup. Treatment of 19 with
This sequence works well on triamine 22. Reaction with
NaHCO3 (25 equiv) and cyanogen bromide (2.2 equiv) in
MeOH for 1 h at 0 °C yields bis cyanamide 23 quantitatively
(see Scheme 4). As expected the hindered aniline does not
react. Heating 23 (2 × 10-2 M) and 3-methyl-2-buten-1-
amine15 (2 equiv) in hexafluoro-2-propanol at 120 °C for
32 h provides crude methyl martinellate (24). Hydrolysis of
the methyl ester cannot be carried out under acidic conditions
since the prenyl double bonds are reactive.14 Hydrolysis in
(9) Both 8 and the pyrrolidine are formed at higher concentrations.
(10) Reduction with LiNH2BH3 was less effective: Myers, A. G.; Yang,
B. H.; Kopecky D. J. Tetrahedron Lett. 1996, 37, 3623-3626.
(11) Maujean, A.; Chuche, J. Tetrahedron Lett. 1976, 2905-2908.
(12) (a) Holton, R. A.; Kim, H.-B.; Somoza, C.; Liang, F.; Biediger, R.
J.; Boatman, P. D.; Shindo, M.; Smith, C. C.; Kim, S.; Nadizadeh, H.;
Suzuki, Y.; Tao. C.; Vu, P.; Tang, S.; Zhang, P.; Murthi, K. K.; Gentile, L.
N.; Liu, J. H. J. Am. Chem. Soc. 1994, 116, 1599-1600. (b) Ojima, I.;
Sun, C. M.; Zucco, M.; Park, Y. H.; Duclos, O.; Kuduk, S. Tetrahedron
Lett. 1993, 34, 4149-4152.
(13) The free triamino ester is unstable and readily forms polymeric
amides. This compound is best stored and characterized as the hydrochloride
salt.
(14) Snider, B. B.; O’Hare, S. M. Tetrahedron Lett. 2001, 42, 2455-
2458.
(15) Jordis, U.; Grohmann, F.; Ku¨enburg, B. Org. Prep. Proced. Int.
1997, 29, 549-560.
Org. Lett., Vol. 3, No. 26, 2001
4219