Treatment of N,N-dibenzylamino alcohols with sulfonyl chloride leads to rearranged beta-chloro amines, precursors to beta-amino acids, and not to tetrahydroisoquinolines.

Article abstract of DOI:10.1021/ol991402i

[reaction: see text] Very recently, the unexpected formation of 3-substituted 1,2,3,4-tetrahydroisoquinolines starting from N,N-dibenzyl-protected beta-amino alcohols was reported. The authors claimed that treatment with tosyl chlorides induced intramolec

Full text of DOI:10.1021/ol991402i

ORGANIC
LETTERS
2000
Vol. 2, No. 5
647-649
Treatment of N,N-Dibenzylamino
Alcohols with Sulfonyl Chloride Leads
to Rearranged â-Chloro Amines,
Precursors to â-Amino Acids, and Not
to Tetrahydroisoquinolines
Klaus Weber, Stephan Kuklinski, and Peter Gmeiner*
Department of Medicinal Chemistry, Emil Fischer Center, Friedrich-Alexander
UniVersity, Schuhstrasse 19, D-91052 Erlangen, Germany
gmeiner@pharmazie.uni-erlangen.de
Received December 25, 1999
ABSTRACT
Very recently, the unexpected formation of 3-substituted 1,2,3,4-tetrahydroisoquinolines starting from N,N-dibenzyl-protected â-amino alcohols
was reported. The authors claimed that treatment with tosyl chlorides induced intramolecular Friedel−Crafts alkylation. Reexamination of the
reactions in our laboratory clearly proved rearranged chloro amines instead of the initially assumed tetrahydoisoquinoline structures. The
chloro amines investigated can be employed as highly useful intermediates for an EPC synthesis of â-amino nitriles and â-amino acids.
N,N-Dibenzyl-protected â-amino alcohols are known as
important intermediates for the synthesis of configurationally
stable R-amino aldehydes which caused considerable impact
on the development of modern ex-chiral pool synthesis.
Pioneering work in this field was performed by M. T. Reetz.1
Furthermore, stereoselective lithiation and C-substitution of
carbamate derivatives was reported by D. Hoppe.2 In a series
of papers, we have shown that O-activation of N,N-dibenzyl-
protected â-amino alcohols, prepared from asparagine or
aspartic acid, by MsCl, Ms2O, or Tf2O or by using Mitsunobu
conditions and subsequent displacement reactions gives
practical access to precursors of enantiomerically pure
â-amino acids,3 diamines,4 amino lactams,5 natural product
derivatives,6 and dopaminergic tetrahydroindolizines.7 In
some cases, the â-amino sulfonate intermediates were
unstable. Thus, we observed 1,2-migration of the N,N-
dibenzylamino functionality to give products of type 4
(Scheme 1). Obviously, this reaction sequence proceeds via
an aziridinium sulfonate or chloride (3).6 We assume that
the preference of the secondary halide (or pseudohalide)
structure compared to the primary substituted isomers is
thermodynamically driven. In the presence of a nitrile or a
second sulfonate as a further functional group, the migration
was followed by â-elimination 8 or pyrrolidinium formation,
respectively.9 Although we examined these reactions very
(4) (a) Gmeiner, P.; Hummel, E. Synthesis 1994, 1026. (b) Gmeiner, P.;
Hummel, E.; Haubmann, C.; Ho¨fner, G. Arch. Pharm (Weinheim) 1995,
328, 265.
(5) Weber, K.; Gmeiner, P. Synlett 1998, 885.
(6) Gmeiner, P.; Junge, D.; Ka¨rtner, A. J. Org. Chem. 1994, 59, 6766.
(7) Gmeiner, P.; Lerche, H. Heterocycles 1990, 31, 9.
(8) Gmeiner, P. Arch. Pharm. (Weinheim) 1991, 324, 551.
(9) (a) Gmeiner, P.; Orecher, F.; Thomas, C.; Weber, K. Tetrahedron
Lett. 1995, 36, 381. (b) Thomas, C.; Orecher, F.; Gmeiner, P. Synthesis
1998, 1491.
(1) Reetz, M. T. Angew. Chem., Int. Ed. Engl. 1991, 30, 1531.
(2) Guarnieri, W.; Grehl, M.; Hoppe, D. Angew. Chem., Int. Ed. Engl.
1994, 33, 1734.
(3) (a) Gmeiner, P. Tetrahedron Lett. 1990, 31, 5717. (b) Gmeiner, P.
Liebigs Ann. Chem. 1991, 501.
10.1021/ol991402i CCC: $19.00 © 2000 American Chemical Society
Published on Web 02/09/2000

Scheme 1
Scheme 2
Besides this, EIMS gave a (M - Cl)+ peak of 368.3
erroneously identified as (M + H)+ by Chandrasekhar et al.10
Furthermore, 13C NMR spectroscopy in combination with a
CH-COSY experiment clearly indicated two magnetically
equivalent phenyl substituents without substituents in the
ortho positions. The migration of the amino group and the
secondary alkyl chloride substructure was elucidated by 1H
NMR and mass spectroscopy when the chemical shifts of
the vicinally positioned proton signals (2.72/2.93 ppm for
CH2N and 3.90 ppm for CHCl) and a characteristic R-cleav-
carefully, the synthesis of 1,2,3,4-tetrahydroisoquinolines (2)
by an intramolecular Friedel-Crafts reaction was not
detected. Thus, we were surprised by a very recent com-
munication from S. Chandrasekhar and co-workers describing
such a cyclization for serine- and phenylglycine-derived N,N-
dibenzyl-protected â-amino tosylates.10 Starting from N,N-
dibenzylserinol, the authors also reported the formation of a
tetracyclic 5a-azanaphthacene system involving the alkylation
+
1
age giving Bn2NCH2 fragmentation in the HRMS were
diagnostic.
of both benzyl groups. Our comparison of the H NMR
available as Supporting Information of ref 10, with those of
the rearranged â-chloro amines of type 4 showed a high
similarity and, thus, lead us to the assumption that the authors
might have been deceived. This prompted us to reexamine
major experiments of the work.
Starting from the O-benzyl-protected amino alcohol 1b13
and from the serinol derivative 1c, reaction with tosyl
chloride under identical conditions was also investigated. We
observed formation of the chloro amines 5b and 5c,
respectively, when 5c showed O-tosylation at the second
hydroxyl function. Again, analytical data confirmed the
chloro amine structure and excluded tetrahydroisoquinoline
Thus, we reacted the O-TBDMS-protected serinol 1a11
with TsCl as described. After extraction and chromatography,
1
we isolated a pure product in 41% yield which gave a H
1
formation.14 Whereas the H NMR data reported for the
NMR spectrum basically identical to that published for the
potential 2-benzyl-3-silyloxymethyl-1,2,3,4-tetrahydroiso-
quinoline. However, careful interpretation of our analytical
data proved the chloro amine structure 5a without any doubt
(Scheme 2).12
benzyloxy derivative corresponded well to our spectra for
5b, this was not the case for 5c. Here, the formation of an
azanaphthacene through bis-Friedel-Crafts alkylation was
(13) Beaulieu, P. L.; Wernic, D. J. Org. Chem. 1996, 61, 3635.
(14) 5b: 1H NMR (360 MHz CDCl3) δ ) 2.75 (dd, J ) 13.7, 6.2 Hz,
1H, NCH2CH), 2.91 (dd, J ) 13.7, 7.5 Hz, 1H, NCH2CH), 3.47 (dd, J )
10.5, 6.3 Hz, 1H, OCH2CH), 3.57 (d, J ) 13.4 Hz, 2H, NCH2Ph), 3.65 (d,
J ) 13.4 Hz, 2H, NCH2Ph), 3.71 (dd, J ) 10.5, 6.3 Hz, 1H, OCH2CH),
3.95-4.05 (m, 1H, CHCl), 4.44 (d, J ) 12.0 Hz, 1H, OCH2Ph), 4.51 (d, J
) 12.0 Hz, 1H, OCH2Ph ), 7.22-7.39 (m, 15H, ar). C24H26ClNO: HRMS
(EI). Calcd 379.1703 (M+), 210.1282 (Bn2NCH2+). Found: 379.1712 (M+),
210.1273 (Bn2NCH2+). 5c: 1H NMR (360 MHz CDCl3) δ ) 2.44 (s, 3H,
CH3), 2.72 (dd, J ) 13.7, 5.5 Hz, 1H, NCH2CH), 2.80 (dd, J ) 13.7, 8.2
Hz, 1H, NCH2CH), 3.51 (d, J ) 13.4 Hz, 2H, NCH2Ph), 3.62 (d, J ) 13.4
Hz, 2H, NCH2Ph), 3.82-3.95 (m, 1H, CHCl), 3.91 (dd, J ) 10.0, 7.3 Hz,
1H, OCH2CH), 4.30 (dd, J ) 10.0, 6.9 Hz, 1H, OCH2CH), 7.22-7.35 (m,
12H, ar), 7.72-7.78 (m, 2H, ar). C24H26ClNO3S: HRMS (EI). Calcd
443.1322 (M+), 210.1282 (Bn2NCH2+). Found: 443.1327 (M+) 210.1293
(Bn2NCH2+). 5d: 1H NMR (360 MHz, CDCl3) δ ) 2.80 (dd, J ) 13.8,
5.8 Hz, 1H, NCH2CH), 2.98 (dd, J ) 13.8, 8.1 Hz, 1H, NCH2CH), 3.62
(d, J ) 13.4 Hz, 2H, NCH2Ph), 3.65 (dd, J ) 11.9, 5.8 Hz, 1H, ClCH2-
CH), 3.68 (d, J ) 13.4 Hz, 2H, NCH2Ph), 3.78 (dd, J ) 11.9, 4.5 Hz, 1H,
ClCH2CH), 3.88-4.00 (m, 1H, CHCl), 7.23-7.40 (m, 10H, ar). C17H19-
Cl2N: HRMS (EI). Calcd 307.0895 (M+), 210.1282 (Bn2NCH2+). Found:
307.0902 (M+), 210.1280 (Bn2NCH2+).
In detail, the experimentally determined molecular com-
position employing both high-resolution mass spectroscopy
and combustion analysis (including the measurement of CHN
and Cl) corresponded very well to the theoretical values.
(10) Chandrasekhar, S.; Mohanty, P. K.; Harikishan, K.; Sasmal, P. K.
Org. Lett. 1999, 1, 877.
(11) Laib, T.; Chastanet, J.; Zhu, J. J. Org. Chem. 1998, 63, 1709.
(12) 5a: 1H NMR (360 MHz, CDCl3) δ ) 0.01 (s, 6H, SiCH3), 0.86 (s,
9H, C(CH3)3), 2.72 (dd, J ) 13.7, 6.3 Hz, 1H, NCH2CH), 2.93 (dd, J )
13.7, 6.4 Hz, 1H, NCH2CH), 3.62 (s, 4H, NCH2Ph), 3.66 (dd, J ) 10.6,
5.3 Hz, 1H, OCH2), 3.76 (dd, J ) 10.6, 5.4 Hz, 1H, OCH2), 3.90 (dddd, J
) 6.4, 6.3, 5.4, 5.3 Hz, 1H, CHCl), 7.21-7.38 (m, 10H, ar); 13C NMR (63
MHz, CDCl3) δ ) -5.3 (SiCH3) 18.3 (C(CH3)3), 25.8 (CCH3), 57.3 (NCH2-
CH), 59.0 (NCH2Ph), 60.7 (CHCl), 65.8 (OCH2), 127.1 (CH-ar), 128.2 (CH-
ar), 128.9 (CH-ar), 139.1 (C-ar) (the interpretation of the data is based on
CH-COSY experiments). Anal. Calcd for C23H34ClNOSi: C, 68.37; H, 8.48;
Cl, 8.77; N, 3.47. Found: C, 68.40; H, 8.53; Cl, 8.72; N, 3.43. HRMS
(EI). Calcd 403.2098 (M+), 210.1282 (Bn2NCH2+). Found: 403.2091 (M+),
210.1288 (Bn2NCH2+).
648
Org. Lett., Vol. 2, No. 5, 2000

reported. However, after displacement of the tosylate function
of 5c with LiCl in DMF giving the dichloride 5d the spectra
showed high correspondence. Therefore, we assume that the
experiments of S. Chandrasekhar et al. resulted in further
chloride displacement during the reaction or brine extraction.
Thus, dichloride 5d appeared to show the azanaphthacene
formation expected by the authors of ref 10.
Upon reaction of the secondary chloride 5a with Bu4NCN,
the nitrile 6 was formed, giving analytical data (1H NMR,
optical rotation) identical to those of a sample which we
synthesized from D-asparagine by benzylation, dehydration
of the carboxamide, reduction, and subsequent silylation of
the primary alcohol.15
rearranged chloride 8b, exclusively.19 On the other hand,1H
NMR based analysis of crude product obtained from 7a gave
a mixture of both the primary and the regioisomeric
secondary chloride. However, after chromatograpy and
storage, the thermodynamic situation was achieved when
only the secondary chloride 8a could be detected. Reaction
of 8a,b with NaCN proceeded again through an aziridinium
intermediate which is regioselectively opened at the sterically
less demanding CH2 position to give the amino nitriles 10a,b
21
(kinetic control).20,
Transformation into the protected
â-homophenyl alanine 9a and the â-homo tyrosine 9b could
be realized by treatment of 10a,b with hydrochloric acid in
78 and 84% yields, respectively.
Starting from N,N-dibenzyl-protected amino alcohols,
which can be readily obtained from natural amino acids, the
combination of O-activation and subsequent introduction of
a cyano substituent can be used for the synthesis of â-amino
nitriles.16 Further hydrolysis gives access to suitably protected
â-amino acids and, thus, demonstrates a highly practical
homologation sequence for amino acids.17 To demonstrate
this, we reacted the phenylalanine- and tyrosine-derived
â-amino alcohols 7a and 7b13,18 with MsCl using Et3N as a
base Scheme 3). This combination of reagents is more
The configurational integrity of the reaction sequence was
established by comparing optical rotation data of the
protected homophenylalanine 9a with those reported for a
differentially synthesized product8 and by esterification,
N-deprotection, and derivatization of the homotyrosine 9b
with enantiomerically pure 1-phenylethyl isocyanate.22 Sub-
sequent 1H NMR studies indicated that the synthetic material
was isomerically pure.
Acknowledgment. This work was supported by the FCI.
Dr. R. Waibel is acknowledged for HRMS measurements.
OL991402I
(18) (a) Reetz, M. T.; Drewes, M. W.; Schmitz, A. Angew. Chem., Int.
Ed. Engl. 1987, 26, 1141. O’Brien, P.; Powell, H. R.; Raithby, P. R. Warren,
S. J. Chem. Soc., Perkin Trans. 1 1997, 1031.
Scheme 3
(19) 8b: 1H NMR (200 MHz, CDCl3) δ ) 2.56 (dd, J ) 14.4, 9.0 Hz,
1H, NCH2CH), 2.80 (d, J ) 7.0 Hz, 2H, CHCH2Ar), 3.23 (dd, J ) 14.4,
4.2 Hz, 1H, NCH2CH), 3.58 (d, J ) 13.6 Hz, 2H, NCH2Ph), 3.72 (d, J )
13.6 Hz, 2H, NCH2Ph), 3.92-4.02 (m, 1H, CHCl), 5.04 (s, 2H, CH2O),
6.85-6.89 (m, 2H, ar), 6.97-7.01 (m, 2H, ar), 7.19-7.46 (m, 15H, ar).
Anal. Calcd for C30H30ClNO: C, 79.02; H, 6.63; N, 3.07. Found: C, 78.51;
H, 6.47; N, 3.74.
(20) 8a: 1H NMR (400 MHz, CDCl3) δ ) 2.59 (dd, J ) 14.4, 9.1 Hz,
1H, NCH2CH); 2.78-2.84 (m, 2H, CHCH2Ph), 3.30 (dd, J ) 14.4, 4.0
Hz, 1H, NCH2CH), 3.59 (d, J ) 13.6 Hz, 2H, NCH2Ph), 3.71 (d, J ) 13.6
Hz, 2H, NCH2Ph), 3.99-4.06 (m, 1H, CHCl), 7.00-7.10 (m, 2H, ar), 7.18-
7.40 (m, 13H, ar). Anal. Calcd for C23H24ClN: C, 78.95; H, 6.92; N, 4.00.
Found: C, 78.80; H, 6.73; N, 4.34. 1-Chloro-2-dibenzylamino-3-phenyl-
propane (recorded together with 8a from crude product): 1H NMR (400
MHz, CDCl3) δ ) 2.78-2.84 (m, 1H, CH2Ph), 3.01 (dd, J ) 13.8, 6.7 Hz,
1H, CH2Ph), 3.16-3.24 (m, 1H, CHN), 3.53 (dd, J ) 11.5, 5.5 Hz, 1H,
CH2Cl), 3.69 (dd, J ) 11.5, 6.8 Hz, 1H, CH2Cl), 3.72 (d, J ) 13.8 Hz, 2H,
NCH2Ph), 3.78 (d, J ) 13.8 Hz, 2H, NCH2Ph), 7.00-7.10 (m, 2H, ar),
7.18-7.40 (m, 13H, ar).
(21) 10a: 1H NMR and optical rotation data were consistent with those
reported in ref 8 for a differentially synthesized product. 10b: 1H NMR
(200 MHz, CDCl3) δ ) 2.41 (dd, J ) 10.5, 8.3 Hz, 1H, CH2CN), 2.50-
2.63 (m, 2H, CH2CN, CH2Ar), 3.08 (dd, J ) 13.6, 5.5 Hz, 1H, CH2Ar),
3.17-3.31 (m, 1H, CHN), 3.66 (d, J ) 13.7 Hz, 2H, NCH2), 3.78 (d, J )
13.7 Hz, 2H, NCH2), 5.05 (s, 2H, OCH2), 6.88-6-92 (m, 4H, ar), 7.20-
7.46 (m, 15H, ar). Anal. Calcd for C31H30N2O: C, 83.37; H, 6.77; N, 6.27.
Found: C, 83.05; H, 6.85; N, 6.46.
(22) A solution of 9b in MeOH and concentrated aqueous HCl was heated
for 12 h at 80 °C. After extraction under basic conditions and evaporation,
the residue was dissolved in MeOH and stirred under H2 in the presence of
catalytic amounts of Pd(OH)2. After 15 h the mixture was filtrated and
evaporated. The residue was dissolved in THF and coupled with (S)-1-
phenylethylisocyanate (1 h, rt). After evaporation the crude urea was
investigated by 1H NMR spectroscopy (400 MHz). The procedure was
repeated with racemic 1-phenylethyl isocyanate. Diagnostic signals: δ )
3.61 ppm (OCH3) for the (3S,2′S)-configured urea; δ ) 3.66 ppm (OCH3)
for the (3S,2′R)-configured urea. The synthetic material proved isomerically
pure.
suitable than TsCl/pyridine since the product separation from
the excess of sulfonyl chloride is more convenient and the
yields are generally higher, at least in our hands. Activation
of the amino alcohol 7b resulted in formation of the
(15) Gmeiner, P.; Hummel, E.; Haubmann, C. Liebigs Ann. Chem. 1995,
1987.
(16) For a prevoius synthesis of â-amino nitriles, see: Reetz, M. T.;
Kayser, F.; Harms, K. Tetrahedron Lett. 1994, 35, 8769.
(17) For previous homologations of amino acids, see: Podlech, J.;
Seebach, D. Liebigs Ann. Chem. 1995, 1217. Caputo, R.; Cassano, E.;
Longobardo, L. Palumbo, G. Tetrahedron 1995, 51, 12337 and references
therein.
Org. Lett., Vol. 2, No. 5, 2000
649