Nitrogen Pronucleophiles in the Phosphine-Catalyzed γ-Addition Reaction

Full text of DOI:10.1021/jo970848e

5670
J . Org. Chem. 1997, 62, 5670-5671
Nitr ogen P r on u cleop h iles in th e
P h osp h in e-Ca ta lyzed γ-Ad d ition Rea ction
Barry M. Trost* and Gregory R. Dake
Department of Chemistry, Stanford University,
Stanford, California 94305-5080
Received May 13, 1997
The conjugate addition of nucleophilic species to the
â-carbon of R,â-unsaturated systems is a fundamental
concept in synthetic organic chemistry.1 A significant
improvement in synthetic design would occur if we could
alter the reactivity of Michael acceptors so that 1,4
addition could be circumvented in favor of other useful
transformations. Our discovery of phosphine’s ability to
induce addition of carbon and oxygen pronucleophiles to
the 4-position of alkynoates led us to test this new
reactivity paradigm with nitrogen-based nucleophiles (eq
1).2,3 Despite our previous successes, a fear existed that
the excellent donor properties of nitrogen in conjugate
additions would result in undesired Michael addition
products (path a), as opposed to γ-addition mediated by
the phosphine-catalyzed process (path b). However, we
have found that under our phosphine-catalyzed condi-
tions Michael addition processes are entirely subverted
in favor of the desired manifold with a variety of nitrogen
nucleophiles, including hydroxamic acid esters, providing
an entry into tripeptide structural mimics.
The competition between Michael addition versus
phosphine-controlled γ-addition was examined further
using 3-butyn-2-one (4), a more reactive substrate (eq 3).
Reaction between 4 and p-toluenesulfonamide using
either tpp or dppp gave only polymeric material. The
condensation between phthalimide and 4 using either
10% tpp or 15% dppp gave the adduct 5a in 48% yield.
The more nucleophilic imide 3 combines with 4 yielding
product 5b in 69% with tpp as the catalyst.
Satisfied that competitive Michael addition was not a
significant problem, we shifted our attention to the
nucleophilic partner in these reactions. Although p-
toluenesulfonamide and phthalimide function well under
our conditions, we sought a nitrogen pronucleophile
which would allow us more synthetic flexibility. The
esters of hydroxamic acids are an interesting class of
nitrogen pronucleophiles which could meet the criteria
of our phosphine-catalyzed addition reaction. Not only
are these compounds conveniently made using standard
amino acid coupling technology, they represented a set
of nitrogen acids with approprate pKa combined with
small steric constraints.5
To this end, we tested the reaction between N-meth-
oxypentanamide (6a ) and 1 using 15% dppp with our
acetic acid/sodium acetate buffer system in toluene at 85
°C (eq 4). Gratifyingly, the adduct 7a was produced in
60% yield. Because our interest lay in nucleophiles
containing a branch at the R-position of the hydroxamic
acid ester, we submitted N-methoxyisobutyramide (6b)
to these same conditions. A 33% yield of the desired
product 7b was recovered. At this same time, our
interest in nucleophiles derived from amino acids had led
us to examine the alanine derivative 6c as well. Under
similar conditions as described above, a 39% yield of the
desired compound 7c was observed.
In order to test the feasibility of these processes,
methyl 2-butynoate (1) was reacted with a number of
nitrogen pronucleophiles using our phosphine catalysis
system, which also involves a general acid-base catalyst
(eq 2). For example, an equimolar mixture of 1 with
p-toluenesulfonamide with 50% acetic acid and 50%
sodium acetate using 10% triphenylphosphine (tpp) in
toluene at 90 °C produced the adduct 2a in 72% yield.4
The structure of compound 2a is clearly established by
the presence of the new olefinic resonances in the 1H
NMR spectrum [(δ 6.75 (dt, J ) 15.7, 5.2 Hz, 1H); 5.94
(dt, J ) 15.7, 1.86 Hz, 1H)]. The use of 5% of the
bidentate phosphines bis(diphenylphosphino)methane
(dppm) or 1,2-bis(diphenylphosphino)ethane (dppe) as
catalyst led to a much lower recovery of 2, 28% and 39%,
respectively. The use of a larger amount (15%) of 1,3-
bis(diphenylphosphino)propane (dppp) with acetic acid-
sodium acetate catalyzed the condensation of 1 with
phthalimide or tetrahydrophthalimide 3 yielding com-
pounds 2b and 2c in 88% and 57% (81% brsm), respec-
tively.
(1) Perlmutter, P. Conjugate Addition Reactions in Organic Syn-
thesis; Pergamon Press: Oxford, 1992. J ung, M. E. In Comprehensive
Organic Synthesis; Trost, B. M., Fleming, I., Semmelhack, M. F., Eds.;
Pergamon Press: Oxford, 1991; Vol. 4, Chapter 1.1, pp 1-67.
(2) Trost, B. M.; Li, C.-J . J . Am. Chem. Soc. 1994, 116, 3167.
(3) Trost, B. M.; Li, C.-J . J . Am. Chem. Soc. 1994, 116, 10819.
(4) This compound has been fully characterized spectroscopically and
its elemental composition established by combustion analysis or high-
resolution mass spectrometry.
(5) The pKa of N-methoxyacetamide was established to be between
16.9 and 17.1 (depending on the indicator) in DMSO. See Bordwell, F.
G.; Fried, H. E.; Hughes, D. L.; Lynch, T.-Y.; Satish, A. V.; Whang, Y.
E. J . Org. Chem. 1990, 55, 3330.
S0022-3263(97)00848-7 CCC: $14.00 © 1997 American Chemical Society

Communications
Sch em e 1. Mech a n istic Ra tion a le of th e
J . Org. Chem., Vol. 62, No. 17, 1997 5671
tryptophan derivative 8b also combines easily with 1
using tpp at 110 °C to give the desired product 9b in 76%
yield. The divalent sulfur moiety contained in 8c also
poses no problem with these same conditions, providing
the substituted amide 9c in 72% yield.
P h osp h in e-Ca ta lyzed Ad d ition Rea ction
Our efforts to improve the efficiency of this process is
outlined below (eq 5, Table 1). Using the simpler
The cleavage of the N-O bond in these compounds
gives entry into vinylogous glycine derivatives. Vinylo-
gous amino acids have been used as structural mimics
of peptidal compounds, serving as tripeptide equivalents.6
The cleavage of the N-O bond is readily accomplished
using titanium(III) chloride in combination with a proton
source such as water. As an illustration, compounds 9a
and 9c were subjected to TiCl3 in wet THF at room
temperature to give the demethoxylated compounds 10a
and 10c in 80% and 88% yields, respectively (eq 7).
Ta ble 1
entry NucH phosphinea
acid sourceb
T (°C) t (h) % yield
1
2
3
4
5
6
7
8
9
6b
6b
6b
6b
6b
6c
6c
6c
6c
6c
6c
10% dppp HOAc-NaOAc
15% dppp HOAc-NaOAc
20% dppp HOAc-NaOAc
10% PPh3 HOAc-NaOAc
10% PPh3 HOAc-NaOAc
10% dppp HOAc-NaOAc
10% dppp PhOH
10% dppp PhOH-NaOPh
10% dppp HOAc-TMG
10% dppba HOAc-NaOAc
10% PPh3 HOAc-NaOAc
90
90
90
85
110 10
85 18
85 18
85 18
85 18
85 18
110 18
5.5
2
2
35
33
42
21
61
39
25
14
-
3.5
-
10
11
-
The ability of phosphines to act as a “nucleophilic
trigger” and modify the reactive manifold of ynoate
systems is truly remarkable. The mechanistic rationale
for this process is outlined in Scheme 1. Although
vinylphosphonium intermediate I could, after protona-
tion, serve as a Michael type acceptor for an appropriate
nucleophile, the thermodynamically more stable inter-
mediate II is invariably the dominant reactive intermedi-
ate using methyl 2-butynoate as the acceptor. After
nucleophilic addition to vinylphosphonium species II (or
its enol), and proton transfer, the phosphine is ejected
to reinititate a new catalytic cycle.
The synthetic power of these phosphine-catalyzed
reactions is wondrous, especially when their simplicity
is also considered. The use of an activating methoxy
substituent on an amide nitrogen to provide a nucleo-
philic partner for the formation of these 1:1 adducts
reveal the utility of this atom economical process.
68
a
b
dppba ) 2-(diphenylphosphino)benzoic acid. 50 mol % each
was used for each reagant. TMG ) tetramethylguanidine.
nucleophile 6b, neither the variation of the amount of
the phosphine catalyst (entries 1-3), nor the use of a
mono- or bidentate phosphine (entries 1 and 4) ap-
preciably affected the outcome of the reaction. The
combination of the use of tpp as catalyst at a higher
temperature did have a significant effect on the reaction,
providing 7b in 61% yield. Concurrently, alternative acid
sources in the reaction between 1 and 6c were found not
to dramatically affect the course of the reaction (entries
6-9), except the use of tetramethylguanidine (TMG) as
cobase in entry 9 resulted in a complete inhibition of
reaction. The inclusion of an internal acid source in the
phosphine as in 2-(diphenylphosphino)benzoic acid
(dppba) resulted in no reaction (entry 10). However, the
use of the triarylphosphine catalyst in combination with
a higher reaction temperature as in entry 5 gave the
desired adduct 7c in 68%, a very acceptable result (entry
11). The yields obtained are reflective of the conversion
of these processes, and not byproduct formation. Specif-
ically, the formation of Michael addition products was
not observed.
Ack n ow led gm en t. We thank the National Science
Foundation and the National Institutes of Health. Mass
spectra were graciously provided by the Mass Spec-
trometry Facility, University of California, San Fran-
cisco, supported by the NIH Division of Research
Resources.
With these new conditions available, this γ-addition
process was tested using other hydroxamic acid ester
derivatives of amino acids with 1 (eq 6). The sterically
encumbered valine derivative 8a , like hydroxamic acid
ester 6c, smoothly reacts with 1 (10% tpp, 100 °C) to give
the desired compound 9a in 67% yield. Using 15% dppp
at 85 °C, as before, yielded 9a in only 23% yield. The
Su p p or tin g In for m a tion Ava ila ble: Sample experimen-
tal procedures and spectral data for 2a -c, 5a ,b, 7a -c, 9a -c,
10a ,c (5 pages).
J O970848E
(6) Hagihara, M.; Anthony, N. J .; Stout, T. J .; Clardy, J .; Schreiber,
S. L. J . Am. Chem. Soc. 1992, 114, 6568.