A novel, one-step method for the conversion of primary alcohols into carbamate-protected amines

Article abstract of DOI:10.1016/S0040-4039(02)00659-7

A novel process for the one-step conversion of primary alcohols into carbamate-protected amines has been developed using a modified Burgess reagent. Although this letter mainly focuses on the conversion of alcohols into the corresponding Cbz-protected amines, the potential for extending this process to a wide range of carbamates has also been demonstrated. A detailed catalytic cycle has been proposed. While exploring the scope of this new reagent, an N-aryl piperidine to an N-aryl pyrrolidine rearrangement has been observed and rationalized.

Full text of DOI:10.1016/S0040-4039(02)00659-7

TETRAHEDRON
LETTERS
Pergamon
Tetrahedron Letters 43 (2002) 3887–3890
A novel, one-step method for the conversion of primary alcohols
into carbamate-protected amines
Michael R. Wood,* June Y. Kim and Kathy M. Books
Department of Medicinal Chemistry, Merck Research Laboratories, PO Box 4, West Point, PA 19486, USA
Received 29 March 2002; accepted 3 April 2002
Abstract—A novel process for the one-step conversion of primary alcohols into carbamate-protected amines has been developed
using a modified Burgess reagent. Although this letter mainly focuses on the conversion of alcohols into the corresponding
Cbz-protected amines, the potential for extending this process to a wide range of carbamates has also been demonstrated. A
detailed catalytic cycle has been proposed. While exploring the scope of this new reagent, an N-aryl piperidine to an N-aryl
pyrrolidine rearrangement has been observed and rationalized. © 2002 Elsevier Science Ltd. All rights reserved.
The interconversion of functional groups during syn-
thetic sequences1 is often necessitated by limitations of
reagent chemoselectivity, functional group incompati-
bility, and the finite number of available starting mate-
rials. Frequently, the nascent functional group requires
protection to allow subsequent transformations at other
sites of the molecule. For example, the classical trans-
formation of a primary alcohol into a protected amine
typically requires a four-step process of mesylation,
azide displacement, azide reduction, and amine protec-
tion. In such cases, it would be particularly attractive to
install the desired functionality already conveniently
protected. Methods of converting alcohols directly into
protected amines often utilize bis-protected ammonia
equivalents and require triphenylphosphine and azadi-
carboxylate reagents.2 Presented here is a novel class of
reagents, based on the Burgess reagent, for the one-step
conversion of alcohols into a variety of carbamate-pro-
tected amines.
tion to form the corresponding methyl carbamates.
Although the direct transformation of a primary alco-
hol into a protected amine was achieved, this reaction
has seen limited use due to the harsh conditions
required to remove the methyl carbamate in order to
access the free amine.4 However, if the novel ‘benzyl
Burgess reagent’ (2) is utilized, the resulting primary
amines are obtained bearing the convenient Cbz pro-
tecting group. Reagent 2 can be prepared by substitut-
ing benzyl alcohol for methanol (Scheme 1) with only
minor modifications to the original procedure.5
According to Scheme 2, reagent 2 was allowed to react
with primary alcohols at elevated temperatures to
provide good yields of the Cbz-protected amines shown
in Table 1. Various functional groups were found to be
compatible with reagent 2, including esters, nitriles,
chlorides and aromatic nitro groups. Entry 6 illustrates
the ability of this new reagent to form differentially
Burgess reagent (1, Scheme 1),3 prepared from
methanol, chlorosulfonyl isocyanate and triethylamine,
has been shown to be an efficient reagent for the
stereospecific cis-dehydration of secondary and tertiary
alcohols to provide olefins. Also described in that initial
publication, albeit in less detail, was the observation
that primary alcohols do not undergo elimination due
to a competing (and predominant) displacement reac-
Scheme 1.
Scheme 2.
Keywords: Burgess reagent; functional group interconversion; Cbz;
carbamate; primary alcohol.
* Corresponding author. Tel.: 1-215-652-9451; fax: 1-215-652-3971;
e-mail: michael wood2@merck.com
–
0040-4039/02/$ - see front matter © 2002 Elsevier Science Ltd. All rights reserved.
PII: S0040-4039(02)00659-7

3888
M. R. Wood et al. / Tetrahedron Letters 43 (2002) 3887–3890
Table 1. Conversion of primary alcohols to Cbz-protected
amines
pound
3
enters into equilibrium with the
reagent-derived triethylamine to form the anion 4 and
its triethylammonium counter ion. Although the pKa
for 3 is not known, it is reasonable to propose that this
critical equilibrium mostly likely favors the formation
of the nucleophilic compound 4 to a great extent. Upon
heating to 85°C in a sealed vessel, compound 4 now
reacts with the small amount of 3 present to form
intermediate 6 and the leaving group N-carbobenzyl-
oxysulfamate 5. The required presence of both compo-
nents 3 and 4 is supported by the observation that
addition of excess base, such as triethylamine, N-
methyl morpholine or pyridine, shifts this equilibrium
further to the right and slows or even halts reaction
progression. If a strong base (NaHMDS, 1 equiv.) is
added prior to heating, the reaction coordinate does
not progress beyond the initial adduct 3, as determined
by LC/MS. Since it is unlikely that compound 5 can
serve as a nucleophile,7 and if this were the major
pathway for the formation of 6, then the maximum
possible yield would be 50%. However, once a small
amount of intermediate 6 is formed, it enters into a
catalytic cycle whereby compound 4, once again acting
as the nucleophile, reacts with 6 to form an equivalent
of 7 concomitant with the regeneration of compound 6.
In this manner, all of compound 4 is eventually fun-
neled to compound 7 through reaction with 6. Neces-
sarily,
a catalytic amount of 6 remains at the
completion of the reaction, and is detectable by LC/MS
prior to work-up. Compound 7 accumulates during the
reaction and has been isolated after basic work-up (1 M
NaOH) where R=(CH2)3Ph(4-NO2) as its insoluble
sodium salt, 8.8 To isolate the desired products directly,
an acidic work-up is preferred, although acids weaker
than HCl were not explored. Other observations that
support this bimolecular mechanism are that reaction
rates and yields increase with higher reaction concen-
trations. It would also appear that significant positive
charge develops at the electrophilic carbon during the
transition state because a b-heteroatom prevents the
displacement, while switching to a more polar solvent
(THF or MeCN) slightly increases yields and rates, and
permits reasonable reaction rates at lower temperatures
(70–75°C).
protected diamines. Benzylic alcohols such as entry 7,
under a variety of conditions, tended to give lower
yields, presumably because of their higher reactivity
(vida infra).
As proposed in the initial publication,3 primary alco-
hols react with this class of reagents (1, 2) via an SN2
type mechanism. In a typical reaction,6 a solution of 2
and the primary alcohol is heated to 50°C for 1 h,
allowing complete formation of the alkyl N-carboben-
zyloxysulfamate ester (3, Scheme 3). Once formed, com-
Scheme 3. Proposed mechanism for the conversion of primary alcohols to Cbz-protected amines by ‘benzyl Burgess reagent’.

M. R. Wood et al. / Tetrahedron Letters 43 (2002) 3887–3890
3889
Scheme 4. Proposed mechanism for the rearrangement of the N-aryl piperidine to the N-aryl pyrrolidine.
alternate carbamate-forming, Burgess-type reagents.
For example, the tert-butyl Burgess reagent (12,
Scheme 5) was prepared and allowed to react with
alcohol 13 under slightly milder reaction conditions
(70°C, 20 h, benzene). This single, unoptimized attempt
yielded 41% of the desired Boc-protected amine,
demonstrating the potential for even these very hin-
dered and acid labile Burgess reagents.
This mechanism can be used to help rationalize the
routinely observed lower yields and faster reaction rates
of benzylic alcohols (Table 1, entry 7). While it is
possible that more side reactions occur due to greater
ionization at the benzylic carbon, this ionization might
also facilitate reaction of compound 3. If greater than
50% of the material is funneled from 3 plus 4 to 6, then
there will be insufficient 4 remaining to complete the
reaction cycle, thereby lowering yields. By contrast, the
reaction of benzyl alcohol described by Burgess3
afforded an 80% yield of the desired methyl carbamate.
The higher yield observed by Burgess may be a result of
performing the reaction neat at a higher temperature,
or the result of fewer steric interactions when forming a
methyl carbamate compared to a benzylic carbamate.
In summary, a novel method for the conversion of
primary alcohols into Cbz-protected amines has been
described. Mechanistic proposals for the overall reac-
tion, as well as an unexpected N-aryl piperidine to
N-aryl pyrrolidine rearrangement have been proposed
based on experimental observations. The complete
scope of carbamates that can be formed remains to be
explored.9
Returning to Table 1, entry 8, this unexpected rear-
rangement of an N-aryl piperidine to an N-aryl pyrro-
lidine during the course of the reaction was confirmed
by multidimensional 1H/13C NMR spectroscopy. In
addition to the 73% isolated yield of the pyrrolidine
product, a minor amount (ca. 7%) of the desired pipe-
ridine isomer was observed by LC/MS. The initial
adduct 9 (Scheme 4) forms smoothly and completely as
determined by LC/MS. Although this compound
should enter into an equilibrium in analogy to that
proposed for compound 3 (Scheme 3), an additional
equilibration between compounds 9 and 10 must also
be proposed. While some of compound 9 reacts as
expected to provide the piperidine product (not shown),
the majority of compound 9 is slowly converted to
compound 10, and eventually to the observed product,
through the intermediacy of the [2.2.1] bicyclic aryl
ammonium cation (11) and its presumably tightly asso-
ciated anion (5). If ions 11 and 5 were not closely
associated and nucleophilic attack on 11 by a nitrogen
nucleophile was the actual mechanism, it would be
difficult to rationalize the approximately 10:1 selectivity
seen for the pyrrolidine product over the piperidine
product. The tendency for the equilibrium to favor 10
over 9 might stem from several causes. As a result of
geometric constraints, k1 might be faster than k−2, while
k2 might be faster than k−1 for purely stochastic reasons
(two equivalent reaction sites for k2 versus only one
reaction site for k−1). In addition, compound 10 might
enjoy faster conversion to product due to reduced steric
demands on the nucleophile (absence of b-branching).
Although not shown for reasons of clarity, a similar
equilibration might take place through intermediates
analogous to compound 6 (Scheme 3).
Scheme 5. Formation of a Boc-protected amine.
Acknowledgements
The authors would like to thank Professor David A.
Evans, Dr. Mark G. Bock, Dr. Scott D. Kuduk and
Dr. Annette S. Kim for helpful discussions and Joan S.
Murphy for the detailed structural determination of
Table 1, entry 8.
References
1. Larock, R. C. Comprehensive Organic Transformations—A
Guide to Functional Group Preparations, 2nd ed.; John
Wiley & Sons: New York, 1999.
2. (a) Wada, M.; Mitsunobu, O. Tetrahedron Lett. 1972, 13,
1279–1282; (b) Lee, B. H.; Miller, M. J.; Prody, C. A.;
Neilands, J. B. J. Med. Chem. 1985, 28, 317–325; (c)
Dodd, D. S.; Kozikowski, A. P. Tetrahedron Lett. 1994,
35, 977–980.
3. Burgess, E. M.; Penton, H. R., Jr.; Taylor, E. A. J. Org.
Chem. 1973, 38, 26–31.
4. Greene, T. W.; Wuts, P. G. M. Protective Groups in
Organic Synthesis, 2nd ed.; John Wiley & Sons: New
York, 1991; pp. 317–318.
Other alcohols may be substituted in place of methanol
according to Scheme 1, resulting in the formation of

3890
M. R. Wood et al. / Tetrahedron Letters 43 (2002) 3887–3890
5. Typical procedure for the preparation of reagent 2: To a
stirred solution of chlorosulfonyl isocyanate (1.30 mL,
14.9 mmol) in dry benzene (35 mL) under nitrogen in a
cool water bath was added anhydrous benzyl alcohol
(Aldrich, 1.54 mL, 14.9 mmol) dropwise over 30 min. The
water bath was then removed and the solution allowed to
warm to ambient temperature for 20 min. The above
solution was then transferred via cannula into a rapidly
stirred solution of triethylamine (4.20 mL, 30.1 mmol) in
dry benzene (17 mL) over the course of 1 h under nitrogen
at ambient temperature. A slight exotherm and the forma-
tion of solid triethylamine hydrochloride were observed.
After an additional 40 min, the contents of the reaction
vessel, along with as much precipitate as possible, was
transferred via cannula into a dry Schlenk filtration
apparatus (medium porosity) to remove the triethylamine
hydrochloride. The resulting clear, colorless, benzene solu-
tion of 2 thus prepared was estimated to have a final
concentration of 0.25 M. This solution could be stored in
the freezer for at least a month with no deleterious effect
on reaction yields, although a slight yellow color and a
small amount of precipitate can develop over time.
Attempts to isolate reagent 2 as a crystalline solid, similar
to Burgess reagent, were unsuccessful.
mixture was removed to check for the formation of the
initial adduct 3 (LC/MS generally shows M++18(H2O):
426). The reaction was stirred for 12–16 h and then
periodically monitored, by LC/MS, for the disappearance
of 3. (CAUTION: Remove the reaction vessel from the
heating bath and allow its temperature to drop below the
boiling point of the solvent prior to removing an analytical
sample.) Frequently, LC/MS spectra obtained prior to
work-up show numerous side products that disappear after
work-up. After complete disappearance of 3, the reaction
was cooled to ambient temperature, benzene was removed
in vacuo, and the residue was partitioned between EtOAc
and 0.5 M HCl. The organic layer was washed with 5%
sodium bicarbonate and brine, then dried over sodium
sulfate. Filtration, solvent removal and silica gel chro-
matography (10–60% EtOAc in hexanes, linear gradient)
provided 318 mg (77%) of a white solid. 1H NMR (400
MHz, DMSO-d6): l 8.14 (d, J=8.8 Hz, 2H), 7.48 (d,
J=8.8 Hz, 2H), 7.34 (m, 5H), 7.26 (t, J=6.0 Hz, 1H,
NH), 5.00 (s, 2H), 3.02 (dt, J=6, 6 Hz, 2H), 2.71 (t,
J=7.6 Hz, 2H), 1.59 (m, 2H), 1.42 (m, 2H); HRMS calcd
+
for C18H20N2O4+NH4 346.1761, found: 346.1791.
7. The formation of a nucleophilic di-anion of 5, under these
reaction conditions, does not seem reasonable.
6. Typical procedure for the conversion of a primary alcohol
into a Cbz-protected amine (Table 1, entry 1): To a flame-
dried reaction vessel equipped with a stir bar and septum
was added 4-(4-nitrophenyl)butan-1-ol (0.21 mL, 1.25
mmol) followed by a benzene solution of 2 (5.0 mL, 1.25
mmol, 0.25 M). The septum was then removed and quickly
replaced with a Teflon cap prior to the reaction being
placed into a 50°C bath. After 1 h, the bath temperature
was increased to 85°C and a small aliquot of the reaction
8. This compound was characterized by LC/MS and 1H
NMR in DMSO-d6, and was stoichiometrically converted
into the desired product under acidic conditions.
9. After completion of this work, we became aware of a
report using similarly modified Burgess reagents: Nico-
laou, K. C.; Huang, X.; Snyder, S. A.; Rao, P. B.; Bella,
M.; Reddy, M. V. Angew. Chem., Intl. Ed. Engl. 2002, 41,
834–838.