Rearrangement of β-chloro N-oxides to hydroxylamines: Opening of the oxazetidinium intermediate by different nucleophiles

Article abstract of DOI:10.1021/jo900031e

The rearrangement of β-chloro N-oxides to hydroxylamines is stereospecific in accord with the presence of a cyclic oxazetidinium intermediate. The latter opens with a range of nucleophiles (carboxylates, cyanide, azide, and thiols)

Full text of DOI:10.1021/jo900031e

This is opened by a nucleophile derived from the oxidation
conditions, e.g., benzoate (Figure 1). Intermediate 4 had been
suggested by Owari, but no investigation of the proposed
stereospecifity of the reaction has been carried out. In 1960
Ishidate5 claimed to have found an example of intermediate 4
in the urine of dogs after the corresponding ꢀ-amino chloride
had been given intravenously. The Owari-rearrangement can
be seen in connection to the allylic [2,3]-Meisenheimer rear-
rangement. However, the mechanism for the formation of 4 is
distinctly different from the [2,3]-Meisenheimer rearrangement,
in which the N-oxide attacks an sp2-carbon and which proceeds
via a five-membered transition state.6
Rearrangement of ꢀ-Chloro N-Oxides to
Hydroxylamines: Opening of the Oxazetidinium
Intermediate by Different Nucleophiles†
Ulrike K. Wefelscheid and Simon Woodward*
School of Chemistry, UniVersity of Nottingham, UniVersity
Park, Nottingham, NG7 2RD, United Kingdom
simon.woodward@nottingham.ac.uk
ReceiVed January 7, 2009
FIGURE 1. Four-membered-ring oxazetidinium intermediate in the
Owari-rearrangement.
The rearrangement of ꢀ-chloro N-oxides to hydroxylamines
is stereospecific in accord with the presence of a cyclic
oxazetidinium intermediate. The latter opens with a range
of nucleophiles (carboxylates, cyanide, azide, and thiols).
To investigate the stereospecificity of the Owari-rearrange-
ment we synthesized enantiomerically pure ꢀ-chloroamines
starting from dibenzylated alaninol, phenylalaninol, valinol, and
methioninol, respectively. After treatment with mesylchloride
for 16 h, the crude material contained mixtures of the corre-
sponding primary and secondary chlorides (ratios 2:1 to 4:1).
Equilibration in chloroform at 50 °C for 18-25 h yielded the
required secondary ꢀ-chloroamines 6, 8, 10, and 12 in 77-88%
(Scheme 2).7 These contained only traces of the corresponding
primary chloride at a level (0-3%) that did not interfere with
our subsequent Owari-rearrangement trials.
During investigations on efficient routes toward enamine
N-oxides 1, we recently delineated a novel type of rearrangement
for N-oxides of the ꢀ-chloroamines 2 to alkoxyamines 3
(Scheme 1).1 This kind of rearrangement was briefly described
by Owari2 in 1953 and has only subsequently been observed
serendipitously by Denny3 in the synthesis of the anticancer
agent chlorambucil and by Morimoto4 during the synthesis of
erythromycin derivatives. It offers high potential to allow very
significant increases in molecular diversity, if a single-pot
reaction and a range of nucleophiles could be used compared
to multistep synthesis.
SCHEME 2. Synthesis of ꢀ-Chloroamines
SCHEME 1. Enamine N-Oxides and Hydroxylamines from
ꢀ-Chloroamines
Using ꢀ-chloroamine 6 we optimized the reaction conditions
for the Owari-rearrangement in order to make it a one-pot
procedure and to increase the yield (Scheme 3). We chose
triethylamine (10 equiv) instead of potassium carbonate as base,
so that it was not necessary to change the solvent after the
oxidation step. Triethylamine also reacted as a trap for any
excess oxidant. We found that with 1.5 equiv of mCPBA
reaction times of 30-40 min at 0 °C are optimal for the
oxidation to the N-oxide. For the subsequent rearrangement
2.5-3 h at 0 °C is sufficient for complete conversion. It is
important to add the ꢀ-chloroamine to a solution of the peracid
Mechanistically the “Owari-rearrangement” should proceed
via stereospecific ring closure to the four-membered cycle 4.
† Dedicated to Prof. Hans-Ulrich Reissig on the occasion of his 60th birthday.
(1) Bernier, D.; Blake, A.; Woodward, S. J. Org. Chem. 2008, 73, 4229–
4232.
(5) Ishidate, M.; Tsukagoshi, S. Chem. Pharm. Bull. 1960, 8, 87–89.
(6) Kleinschmidt, R. F.; Cope, A. C. J. Am. Chem. Soc. 1944, 66, 1929–
1933. For a review on reactions of amine N-oxides see: Albini, A. Synthesis
1993, 263–277.
(2) Owari, S. Chem. Pharm. Bull. 1953, 1, 353–357.
(3) Tercel, M.; Wilson, W. R.; Denny, W. A. J. Med. Chem. 1995, 38, 1247–
1252.
(4) Morimoto, S.; Adachi, T.; Watanabe, Y.; Omura, S. Heterocycles 1990,
31, 305–319.
(7) The transformation from 7 to 8 is known to proceed in a stereospecific
manner: Weber, K.; Kuklinski, S.; Gmeiner, P. Org. Lett. 2000, 2, 647–649.
2254 J. Org. Chem. 2009, 74, 2254–2256
10.1021/jo900031e CCC: $40.75 2009 American Chemical Society
Published on Web 02/09/2009

SCHEME 3. Owari-Rearrangement of Enantiomerically
Pure ꢀ-Chloroamines
SCHEME 4. Owari-Rearrangements with Non-Benzoate
Nucleophiles
to ensure that an excess of acid is present, protonating all of
the N-oxide formed. In the reverse addition mode, the ꢀ-chlo-
roamine itself can act as a base promoting rearrangement and
thus exposing the product to N-oxidation. Using the optimal
procedure we isolated the hydroxylamine 13 in enantiomerically
pure form in 77% yield (Scheme 3). The chloride reacted as a
competing nucleophile to the m-chlorobenzoate and 21% of the
corresponding byproduct was formed. Deliberate addition of
excess m-chlorobenzoic acid (mCBA, 3 equiv) led to 83% of
benzoate 13 and 13% of the corresponding chloride. The
formation of the chloride product could be suppressed by the
addition of AgNO3 (1.05 equiv) and 13 was isolated in 93%
yield. The equivalent Owari-rearrangement of phenylalanine
derived ꢀ-chloroamine 8 also proceeded smoothly. With ꢀ-chlo-
roamine 10, which has a substituent in the R-position to the
chloride, the reaction was more problematic, presumably because
of steric hindrance. Compound 15 was formed in only 48% yield
and was difficult to purify. Methionine derived ꢀ-chloroamine
12 could easily be converted to the triply oxidized and
rearranged compound 16 in 89% yield. This represents an
interesting transformation, as three new functionalities (benzoate,
hydroxylamine, and sulfone) are generated in a single step.
We have used our improved procedure to allow the incor-
poration of other, non-benzoate, nucleophiles. As the oxidation
is carried out with mCPBA, the new nucleophiles have to be
superior to m-chlorobenzoate. Tetrabutylammonium salts of
cyanide and azide were selected because of their high nucleo-
philicity and their good solubility in dichloromethane. The
addition of 3 equiv of tetrabutylammonium cyanide gave the
corresponding cyanide 17 in 84% yield (Scheme 4). The
benzoate 13 was found as a byproduct in 7% yield. An excess
of tetrabutylammonium azide (3 equiv) yielded 72% of ꢀ-azi-
doalkoxyamine 18 and 15% of benzoate 13. With thiophenol
and 2-thioethanol as nucleophiles, no benzoate 13 was detected
as a byproduct. Phenylthioether 19 was isolated in 72% yield,
whereas thioether 20 was easily obtained in 92% yield. When
we used peracetic acid as oxidant, the time for the oxidation
step was longer, presumably due to the more acidic reaction
mixture: we employed 28 wt % AcOOH in AcOH, whereas
the mCPBA in the previous described reactions was 75 wt %
in mCBA.8 ꢀ-Acetoxyalkoxyamine 21 was isolated in 92%
yield. We could prove the enantiopurity of compounds 13-18,
20, and 21 by chiral HPLC,9 indicating for the first time that
SCHEME 5. Synthesis of oxazane 24 via
Owari-rearrangement.
the chemistry of Figure 1 is completely stereospecific. An
attempt to observe 4 by following the reaction from the N-oxide
of 6 to 13 at -60 °C via 13C NMR spectroscopy was
unsuccessful. We saw only signals of the N-oxide of 6 and
smooth formation of 13.
As significant ring strain might be expected in the formation
of 4 we were interested to probe such effects by attempted
formation of a bicyclic analogue. A suitable precursor was
prepared in one step from benzylated prolinol (22) according
to the procedure described in Scheme 2 (Scheme 5). Here no
equilibration was necessary, as the proton NMR spectrum of
the crude compound 23 did not show the presence of any isomer.
The oxidation of 23 yielded 24 as a single isomer, and the
configuration could be proven by NOE analyses of the crude
compound. For the Owari-rearrangement it was necessary to
(8) The purity of the peracids were determined according to a procedure
described in the following: Woodward, S. in Transition Metals in Organic
SynthesissA Practical Approach; Gibson, S. E., Ed.; Oxford University Press:
Oxford, UK, 1997; p 7.
(9) For phenylthioether 19 we could not find conditions for complete baseline
separation of the racemic compound. However, the enantiomeric ratio is very
high, as we do not see any shoulder of the minor enantiomer.
J. Org. Chem. Vol. 74, No. 5, 2009 2255

Hz, 1 H, 2-H), 7.26-7.32 (m, 2 H, Ar), 7.34-7.44 (m, 8 H, Ar)
ppm; 13C NMR (101 MHz, CDCl3) δ 23.1 (q, C-3), 55.5 (d, C-2),
59.1 (t, CH2Ph), 62.0 (t, C-1), 127.1, 128.2, 128.9 (3 d, Ar), 139.1
(s, Ar) ppm; IR (ATR) ν 3060-2810 (CdCH, CsH), 1495, 1450
(CdC) cm-1; HRMS (pos. ESI) C17H20ClN·H+ calcd 274.1363,
found 274.1353. Anal. calcd for C17H20ClN (273.8): C 74.57, H
7.36, N 5.12. Found: C 74.45, H 7.38, N 5.18.
heat the reaction mixture for several hours at 75-85 °C in
dichloroethane, or at 85 °C (internal sensor temperature) for
45 min in CH2Cl2 in a microwave. We isolated oxazane 26 in
40% and 52% yields respectively from these two approaches
both with an enantiomeric ratio of 95:5. Neither the benzylated
prolinol 22 nor the chloropiperidine 23 employed showed
enantiomeric ratios higher than 95:5. Therefore we can conclude
that the Owari-rearrangment is stereospecific even when the
intermediate 4 suffers considerable ring strain, such as 25. The
lower yield might be explained by the unfavored boat-like
transition state shown in Scheme 5. Intermediate 26 is a
surrogate of 5,6-dihydroxyhexamine that has found use in target
synthesis.10
Procedure for the Owari-Rearrangement: (2S)-2-(Dibenzy-
laminooxy)propyl 3-Chlorobenzoate (13). To a solution of
mCPBA (75 wt %, 93 mg, 0.41 mmol) in CH2Cl2 (1.5 mL) was
added a solution of ꢀ-chloroamine 6 (75 mg, 0.27 mmol) in CH2Cl2
(2.5 mL) at 0 °C. After 30 min AgNO3 (48 mg, 0.28 mmol) and
NEt3 (0.37 mL, 2.7 mmol) were added. Stirring was continued for
3 h at 0 °C, then the solvent was removed under reduced pressure.
Flash-chromatography (silica gel, petrol 40-60 °C/NEt3 20:0.04)
In conclusion we have optimized the conditions for Owari-
type rearrangement with a variety of external nucleophiles. In
addition we have for the first time proved the stereospecifity of
this rearrangement and developed a new access to enantiomeri-
cally pure ꢀ-cyano-, ꢀ-azido-, ꢀ-thio, and ꢀ-aceto hydroxylamines.
afforded 103 mg (93%) of compound 13 as a colorless oil. [R]22
D
1
-27.8 (c 1.54, CHCl3); H NMR (400 MHz, CDCl3) δ 0.97 (d, J
) 6.5 Hz, 3 H, 3-H), 3.66 (ddq, J ) 4.2, 4.5, 6.5 Hz, 1 H, 2-H),
AB-system (δA ) 3.85, δB ) 3.88, JAB ) 13.0 Hz, 4 H, CH2Ph),
ABX-system (δA ) 4.07, δB ) 4.10, JAB ) 11.5 Hz, JAX ) 4.5
Hz, JBX ) 4.2 Hz, 2 H, 1-H), 7.21-7.31 (m, 6 H, Ar), 7.33-7.38
(m, 5 H, Ar), 7.52 (ddd, J ) 1.1, 2.1, 7.9 Hz, 1 H, Ar), 7.86 (ddd,
J ) 1.1, 1.6, 7.9 Hz, 1 H, Ar), 7.95 (ddd, J ) 0.4, 1.6, 2.1 Hz, 1
H, Ar) ppm; 13C NMR (101 MHz, CDCl3) δ 16.0 (q, C-3), 62.8 (t,
CH2Ph), 66.7 (t, C-1), 75.2 (d, C-2), 127.4, 127.7, 128.2, 129.55,
129.66, 129.72 (6 d, Ar), 132.0 (s, Ar), 132.9 (d, Ar), 134.4, 137.5
(2 s, Ar), 165.1 (s, CdO) ppm; IR (ATR) ν 3030-2840 (CdCH,
CsH), 1725 (CdO), 1575, 1495 (CdC) cm-1; HRMS (pos. ESI)
C24H24ClNO3 ·H+ calcd 410.1523, found 410.1515; enantiomeric
ratio >99:1 (determined by chiral HPLC, column OD, hexane/
iPrOH 99:1, flow 0.5 mL/min, minor enantiomer: 16.1 min, major
enantiomer: 17.3 min). Anal. calcd for C24H24ClNO3 (409.9): C
70.32, H 5.90, N 3.42. Found: C 70.12, H 5.91, N 3.19.
Experimental Section
Procedure for the Preparation of ꢀ-Chloroamines:
(2R)-N,N-Dibenzyl-2-chloropropan-1-amine (6). To a solution of
the ꢀ-amino alcohol 6 (1.15 g, 4.52 mmol), NEt3 (1.88 mL, 13.6
mmol), and a catalytic amount of DMAP (14 mg, 0.11 mmol) in
CH2Cl2 (10 mL) was added MsCl (0.70 mL, 9.0 mmol) at 0 °C.
Stirring was continued for 6 h at this temperature and then at rt
overnight. The reaction mixture was diluted with EtOAc then
washed with brine and the solvent was removed in vacuo. The
proton-NMR spectrum of the crude material showed a 3.6:1 mixture
of the desired secondary chloride and the corresponding primary
chloride. Equilibration in CHCl3 (10 mL) at 50 °C for 20 h and
subsequent flash-chromatography (silica gel, petrol 40-60 °C/NEt3
20:0.4) yielded a 97:3 mixture of compound 6 and the corresponding
primary chloride as a colorless oil, which crystallized after 1 h in
Acknowledgment. We thank Dr. David Bernier for useful
discussions of this work. We thank Dr. A. K. Forrest for
identifying ref 4. We thank the EPSRC for support through grant
EP/E030092/1.
the fridge (1.13 g, 91%, compound 6: 88%). Mp 47-48 °C; [R]27
D
1
-18.1 (c 1.00, CHCl3); H NMR (400 MHz, CDCl3) δ 1.41 (d, J
Supporting Information Available: Experimental and ana-
) 6.7 Hz, 3 H, 3-H), 2.63 (dd, J ) 7.6, 13.3, 1 H, 1-H), 2.77 (dd,
J ) 6.2, 13.3 Hz, 1 H, 1-H), 3.56 (d, J ) 13.7 Hz, 2 H, CH2Ph),
3.68 (d, J ) 13.6 Hz, 2 H, CH2Ph), 4.00 (qdd, J ≈ 6.5, 6.5, 7.6
1
lytical details and copies of H and 13C NMR spectra of all
new compounds. This material is available free of charge via
the Internet at http://pubs.acs.org.
(10) For example: Takahata, H.; Ihara, K.; Kubota, M.; Momose, T.
Heterocycles 1997, 46, 349–356.
JO900031E
2256 J. Org. Chem. Vol. 74, No. 5, 2009