Making imines without making water - exploiting a recognition-mediated Aza-Wittig reaction

Article abstract of DOI:10.1021/ol8024499

(Figure Presented) A recognition-mediated aza-Wittig reaction permits the efficient formation of an imine in dry CDCl₃ from an iminophosphorane and an aldehyde. The kinetic and thermodynamic parameters for this reaction are compared with those

Full text of DOI:10.1021/ol8024499

ORGANIC
LETTERS
2009
Vol. 11, No. 2
301-304
Making Imines Without Making
Water-Exploiting a
Recognition-Mediated Aza-Wittig
Reaction
Vicente del Amo and Douglas Philp*
EaStCHEM and Centre for Biomolecular Sciences, School of Chemistry, UniVersity of
St. Andrews, North Haugh, St. Andrews, Fife KY16 9ST, United Kingdom
d.philp@st-andrews.ac.uk
Received October 23, 2008
ABSTRACT
A recognition-mediated aza-Wittig reaction permits the efficient formation of an imine in dry CDCl3 from an iminophosphorane and an aldehyde.
The kinetic and thermodynamic parameters for this reaction are compared with those obtained from a condensation reaction between an
aldehyde and an amine that forms the same product.
Dynamic covalent chemistry1 (DCC) presents an op-
portunity to develop the chemistry of complex systems
through the creation of collections of compounds, which
interact and interconvert with each other through the breaking
and reforming of covalent bonds and whose composition2
is regulated by thermodynamics. Thus, the development of
protocols based on DCC that establish and manage replica-
tion, organization, and evolution within synthetic molecular
and supramolecular assemblies offer a programmed approach
to predetermined dynamic behavior s so-called3 systems
chemistry. Among the various covalent bonds known to have
a dynamic behavior, imines have attracted4 special interest.
Imines are formed typically by the reversible acid-catalyzed
condensation of an amine with an aldehyde with extrusion
of a molecule of water. Under some conditions, for example,
in nonpolar solvents, the equilibrium lies far on the side of
the starting materials and, to drive the CdN bond formation
to completion, water must be removed through azeotropic
distillation or by employing a chemical drying agent in the
reaction mixture. These conditions limit the utility of this
reaction, particularly in respect of exploiting molecular
recognition associated with hydrogen bonds. Since recogni-
tion events are a prerequisite for directing events within
dynamic libraries, there is a clear need to develop methods
for the formation of imines under conditions that are
compatible with recognition motifs used commonly in
nonpolar solvents.
As part of our ongoing studies of self-replicating systems,
we have been exploring5 the consequences of recognition
processes on the acceleration of cycloaddition reactions.
Recently, our focus has shifted toward exploiting the
implications of replication events in dynamic and reactive
chemical systems and networks. As a first step, we reported6
a dynamic replicating system based on imine formation using
a classical condensation approach. However, given the
limitations of imine formation in nonpolar solvents, we
wished to develop an alternative method of imine formation.
(1) For general reviews on DCC see: (a) Lehn, J.-M. Chem.-Eur. J.
1999, 5, 2455. (b) Rowan, S. J.; Cantrill, S. J.; Cousins, G. R. L.; Sanders,
J. K. M.; Stoddart, J. F. Angew. Chem., Int. Ed. 2002, 41, 898. (c) Corbett,
P. T.; Leclaire, J.; Vial, L.; West, K. R.; Wietor, J.-L.; Sanders, J. K. M.;
Otto, S. Chem. ReV. 2006, 106, 3652. (d) Otto, S.; Severin, K. Top. Curr.
Chem. 2007, 277, 267.
10.1021/ol8024499 CCC: $40.75
Published on Web 12/15/2008
2009 American Chemical Society

Although much less explored than its carbon-based cousin,
the aza-Wittig reaction7 offers an attractive method for imine
formation. Reaction between an iminophosphorane8 (gener-
ated from the Staudinger reaction9 of the correponding azide
with a triarylphosphine) and a carbonyl compond leads to
the formation of a CdN bond in a similar manner to CdC
bond formation by phosphonium ylids. In general, however,
the reaction is rather inefficient when it is performed10 in
an intermolecular sense, although the intramolecular aza-
Wittig reaction has attracted considerable attention as a result
of its potential for the synthesis of nitrogen-containing
heterocycles. In this communication, we compare and
contrast the recognition-mediated formation of an imine in
CDCl3 through a traditional condensation-based approach and
through an aza-Wittig reaction. We demonstrate that the
recognition-mediated aza-Wittig methodology leads to ef-
ficient and high conversion to imine, thus establishing a new
method for the formation of the dynamic imine bond in
CDCl3.
Scheme 1
The traditional approach to the synthesis of imine 3 involves
the reaction of amine 1 with aldehyde 2 in CDCl3 at 25 °C
(Scheme 1) with the removal of water. The formation of imine
3 occurs through bimolecular reaction of the amine and the
aldehyde (k2N/k-2N, Scheme 1). However, the location of
(2) For some recent examples of DCC, see: (a) Cheeseman, J. D.;
Corbett, A. D.; Shu, R.; Croteau, J.; Gleason, J. L.; Kazlauskas, R. J. J. Am.
Chem. Soc. 2002, 124, 5692. (b) Otto, S.; Kubik, S. J. Am. Chem. Soc.
2003, 125, 7804. (c) Stulz, E.; Scott, S. M.; Bond, A. D.; Teat, S. J.; Sanders,
J. K. M. Chem.-Eur. J. 2003, 9, 6039. (d) Gonzalez-Alvarez, A.; Alfonso,
I.; Lopez-Ortiz, F.; Aguirre, A.; Garcia-Granda, S.; Gotor, V. Eur. J. Org.
Chem. 2004, 1117. (e) Ramstro¨m, O.; Lohmann, S.; Bunyapaiboonsri, T.;
Lehn, J.-M. Chem.-Eur. J. 2004, 10, 1711. (f) Chichak, K. S.; Cantrill,
S. J.; Pease, A. R.; Chiu, S.-H.; Cave, G. W. V.; Atwood, J. L.; Stoddart,
J. F. Science 2004, 304, 1308. (g) Buryak, A.; Severin, K. Angew. Chem.,
Int. Ed. 2005, 44, 7935. (h) Corbett, P. T.; Sanders, J. K. M.; Otto, S. J. Am.
Chem. Soc. 2005, 127, 9390. (i) Milanesi, L.; Hunter, C. A.; Sedelinkova,
S. E.; Waltho, J. P. Chem.-Eur. J. 2006, 12, 1081. (j) Shi, B.; Stevenson,
R.; Campopiano, D. J.; Greaney, M. F. J. Am. Chem. Soc. 2006, 128, 8459.
(k) Bulos, F.; Roberts, S. L.; Furlan, R. E. L.; Sanders, J. K. M. Chem.
Commun. 2007, 3092. (l) Ludlow, R. F.; Liu, J.; Li, H.; Roberts, S. L.;
Sanders, J. K. M.; Otto, S. Angew. Chem., Int. Ed. 2007, 46, 5762. (m)
Haussmann, P. C.; Khan, S. I.; Stoddart, J. F. J. Org. Chem. 2007, 72,
6708. (n) Schultz, D.; Nitschke, J. R. Chem.-Eur. J. 2007, 13, 3660.
(3) (a) Kindermann, M.; Stahl, I.; Reimold, M.; Pankau, W. M.; von
Kiedrowski, G. Angew. Chem., Int. Ed. 2005, 44, 6750. (b) Stankiewicz,
J.; Eckardt, L. H. Angew. Chem., Int. Ed. 2006, 45, 342. (c) Corbett, P. T.;
Sanders, J. K. M.; Otto, S. Angew. Chem., Int. Ed. 2007, 46, 8858. (d)
Sarma, R. J.; Nitschke, J. R. Angew. Chem., Int. Ed. 2008, 47, 377. For a
general review on Systems Chemistry see: (e) Ludlow, R. F.; Otto, S. Chem.
Soc. ReV. 2008, 37, 101.
complementary recognition sites on amine 1 (amidopyridine)
and aldehyde 2 (carboxylic acid), allows the association of these
reactive partners in a binary complex [1•2] opening an alterna-
tive reaction channel leading to the formation of imine 3 (k1N/
k-1N, Scheme 1). The formation of this complex arranges the
reactive groups for a pseudointramolecular reaction within the
[1•2] complex that should be accelerated strongly. In order to
establish a baseline for comparison, we performed a control
reaction between amine 1 and m-tolualdehyde in the presence
of 4-bromophenylacetic acid ([1] ) [m-tolualdehyde] ) [4-bro-
mophenylacetic acid] ) 15 mM) in dry11 CDCl3 at 25 °C to
form imine 4. The extent of the reaction was assayed12 at regular
intervals by 500 MHz 1H NMR spectroscopy and a concentra-
tion-time profile was constructed (Figure 1, open circles).
Kinetic simulation and fitting of this data to the appropriate
bimolecular model (k2N and k-2N, Scheme 1) afforded the
(4) Meyer, C. D.; Joiner, C. S.; Stoddart, J. F. Chem. Soc. ReV. 2007,
36, 1705.
(5) For some recent examples, see: (a) Pearson, R. J.; Kassianidis, E.;
Slawin, A. M. Z.; Philp, D. Chem.-Eur. J. 2006, 12, 6829. (b) Kassianidis,
E.; Philp, D. Angew. Chem., Int. Ed. 2006, 45, 6334. (c) Kassianidis, E.;
Pearson, R. J.; Philp, D. Chem.-Eur. J. 2006, 12, 8789. (d) Bennes, R.;
Philp, D. Org. Lett. 2006, 8, 3651. (e) Turega, S. M.; Philp, D. Chem.
Commun. 2006, 3684.
(11) CDCl3 (99.8% atom D) was purchased from Aldrich as 100 g
bottles. Bottles were opened under a positive pressure of Ar and the contents
were stored over preactivated 4 Å molecular sieves. The water content was
determined to be 14 ppm using Karl Fischer titration (Mettler Toledo DL32
Coulometer). Samples of aldehyde 2 and 4-bromophenylacetic acid contain
one molecule of water of crystallization. Therefore, the solutions discussed
here, which are prepared from these compounds, will have H2O concentra-
tions of 15 mM.
(6) del Amo, V.; Slawin, A. M. Z.; Philp, D. Org. Lett. 2008, 10, 4589.
(7) Reviews: (a) Palacios, F.; Alonso, C.; Aparicio, D.; Rubiales, G.;
de los Santos, J. M. Tetrahedron 2007, 63, 523. (b) Eguchi, S. ARKIVOC
2005, 2, 98. (c) Fresneda, P. M.; Molina, P. Synlett 2004, 1. (d) Eguchi, S.;
Okano, T.; Okawa, T. Rec. Res. Dep. Org. Chem. 1997, 1, 337. (e) Molina,
P.; Vilaplana, M. J. Synthesis 1994, 197.
(8) (a) Staudinger, H.; Meyer, J. HelV. Chim. Act. 1919, 2, 635. (b)
Patonay-Peli, E.; Litkei, G. Synthesis 1990, 511. (c) Cambon, F. A. Synth.
Commun. 1994, 24, 2653. (d) Shalev, D. E.; Chiacchiera, J. Org. Chem.
1996, 61, 1689.
(12) The time course of each reaction was followed by 500 MHz 1H
NMR spectroscopy. The disappearance of the resonances arising from the
aldehyde protons in 2 or m-tolualdehyde (δ ) 9.9-10.1) and the
simultaneous appearance of resonances arising from the imine protons in 3
or 4 (δ ) 8.4-8.5) were monitored. Concentration vs time profiles were
constructed by deconvolution of these resonances.
(9) For a general review on the chemistry and properties of azides see:
Bra¨se, S.; Gil, C.; Knepper, K.; Zimmermann, V. Angew. Chem., Int. Ed.
2005, 44, 5188, and references therein.
(13) von Kiedrowski, G. SimFit, A program for the analysis of kinetic
data, Version 1.0; Ruhr-Universita¨t Bochum: Germany, 1994.
(10) Palacios, F.; Vicario, J.; Aparicio, D. J. Org. Chem. 2006, 71, 7690.
302
Org. Lett., Vol. 11, No. 2, 2009

6-31G(d,p)) on the possible conformations of imine product
3. The results of these calculations are summarized in Figure
2a.
Figure 1. Rate profiles at 25 °C in dry CDCl3 for the formation of
imine 3 (b) or imine 4 (O) from amine 1 and aldehyde 2 or
m-tolualdehyde, respectively. 4-Bromophenylacetic acid (15 mM)
was added to the reaction involving m-tolualdehyde. The starting
concentrations of amine and aldehyde were 15 mM. In both cases,
the lines represent the best fit of the experimental data to the
appropriate kinetic model (Scheme 1) using SimFit.15 See Sup-
porting Information for details.
best fit values for these rate constants: k2N ) 1.27 × 10-3
M-1 s-1 and k-2N ) 1.09 × 10-3 M-1 s-1. Therefore, as
expected, the equilibrium constant for imine formation in
CDCl3 is only 1.15. When the same reaction was performed
using aldehyde 2 ([1] ) [2] ) 15 mM, dry CDCl3, 25 °C),
the concentration-time profile for the formation of imine 3
is dramatically different (filled circles, Figure 1). As ex-
pected, imine 3 is formed very rapidly as a result of fast,
pseudointramolecular reaction through the [1•2] complex.
However, imine formation stalls after only two hours at 60%
conversion ([3] ) 9 mM).
Figure 2. Ball and stick representations of the calculated (B3LYP/
6-31G(d,p)) structures of (a) three conformations of imine 3 and
(b) the 1,3,2-λ5-oxazaphosphazetidine intermediate accessed by the
[2•6] complex during the aza-Wittig reaction. Carbon atoms are
colored green, oxygen atoms are colored red, nitrogen atoms are
colored blue, hydrogen atoms are colored white and phosphorus
atoms are colored mauve. Hydrogen bonds are represented by
dashed lines. Most hydrogen atoms have been removed for clarity.
The introduction of recognition has had two effects on
the reaction. First, the rate of reaction between the amine
and the aldehyde is accelerated strongly through the [1•2]
complex s initial rate of imine formation is accelerated by
almost 30× (initial rate ) 20.4 mM h-1 vs 0.73 mM h-1 for
the control reaction between 1 and m-tolualdehyde). Second,
the position of equilibrium has been shifted - the overall
equilibrium constant for formation of 3 is now 6.00.
However, this change in equilibrium constant only represents
a stabilization of the system by ∼ 4 kJ mol-1 in the presence
of recognition. We found this result surprising, as one might
expect the noncovalent interactions used to assemble [1•2]
to live on13 in imine 3, stabilizing it signficantly. Thus, the
results of introducing recognition elements in order to guide
imine formation in this system are somewhat unsatisfactory.
Although we do accelerate the formation of imine 3, the
reaction stalls as a result of the issues associated with
condensation reactions performed in nonpolar solvents such
as CDCl3 s in particular the formation of water as the
reaction progresses.
In conformation 2HB, both hydrogen bonds used to
assemble [1•2] are still present in imine 3. However, this
conformation is 20.0 kJ mol-1 less stable than conformation
1HB which possesses one hydrogen bond only between the
carboxylic acid proton and the amidopyridine ring nitrogen
atom. Conformation 1HB is, in turn, 13.4 kJ mol-1 less stable
than the Open conformation. The hydrogen bonded confor-
mations s although they contain stabilizing noncovalent
interactions - incorporate significant distortion around the
CdN unit. This distortion is energetically unfavorable and
outweighs the stabilizing noncovalent interactions. These
results help to explain why the recognition-mediated con-
densation reaction is relatively unsuccessful at shifting the
equilibrium position for the formation of imine 3. Although
the complex [1•2] is relatively successful at accelerating the
addition of the amine to the aldehyde, no stablizing hydrogen
bonds can be sustained in the product.
A potential solution to our inability to alter the equilibrium
position for the formation of imine 3 in CDCl3 is to exploit
a recognition-mediated irreversible process for imine forma-
tion. In this respect, the aza-Wittig reaction is ideal since it
also affords an imine, however, the other product of the
To shed some light on the apparently poor stabilization
of imine 3 by intramolecular hydrogen bonding, we per-
formed a series of electronic structure calculations (B3LYP/
Org. Lett., Vol. 11, No. 2, 2009
303

reaction is triphenylphosphine oxide (TPPO), not water, and
the reaction is irreversible. Thus, the location of comple-
mentary recognition sites in iminophosphorane 6 and alde-
hyde 2, should allow the association of these reactive partners
in a binary complex [2•6]. If, as a result of formation of this
complex, the reactive groups are then oriented in a mutually
suitable geometry for the aza-Wittig reaction to occur, the
formation of imine 3 should be accelerated strongly. Thus,
we performed a series of electronic structure calculations
(B3LYP/6-31G(d,p)) on the transition states leading, from
[2•6] to the corresponding imine 3. As expected, the [2•6]
complex can access a transition state leading to the four-
membered ring 1,3,2-λ5-oxazaphosphazetidine (Figure 2b)
that is an accepted intermediate14 in the aza-Wittig reaction.
This species can then decompose, with loss of TPPO, to form
imine 3.
Figure 3. Rate profiles at 25 °C in dry CDCl3 for the formation of
imine 3 (b) or imine 4 (O) from iminophosphorane 6 and aldehyde
2 or m-tolualdehyde), respectively. 4-Bromophenyl acetic acid (15
mM) was added to the reaction involving m-tolualdehyde. The
starting concentrations of iminophosphorane and aldehyde were 15
mM. In both cases, the solid lines represent the best fit of the
experimental data to the appropriate kinetic model (Scheme 1) using
SimFit.13 See Supporting Information for details.
We next sought to verify these computational predic-
tions by comparing the aza-Wittig reaction between
aldehyde 2 and iminophosphorane 6 with the condensation
reaction between amine 1 and aldehyde 2. In all experi-
ments, iminophosphorane 6 was prepared freshly by
reacting azide 5 with an slight excess of PPh3 in dry CDCl3
at 40 °C under an Ar atmosphere. Formation of 6
1
10× (initial rate ) 6.26 mM h-1 vs 0.59 mM h-1 for the
control reaction between 6 and m-tolualdehyde). The kinetic
data allows us to estimate the kinetic effective molarity
(kEM) generated in the [2•6] complex s the ratio k1AW/k2AW
s is 1.2 M. The overall conversion to imine is 92% and the
equilbrium constant for imine formation is now 132. Despite
the fact that water is present in the CDCl3 solution at the
start of the reaction, no further water is produced in the
formation of imine 3.
In conclusion, we have demonstrated that with a careful
design and location of recognition elements on the reaction
partners, it is possible to exploit an aza-Wittig reaction to
facilitate the synthesis of an imine under conditions that are
normally unfavorable for direct condensation. The ability to
form a reversible imine bond under kinetic control is
especially significant for our long-term goal of constructing,
with high efficiency and fidelity, chemical subsystems that
can be deployed within chemical reaction networks. Experi-
ments directed toward realization of this goal are currently
under way in our laboratory.
proceeded quantitatively in about 3 h as assessed by H
and 31P NMR spectroscopy.
To determine the magnitude of any rate acceleration arising
from the formation of the binary complex [2•6], the rate of
the recognition-mediated reaction was compared to the rate
of the corresponding bimolecular aza-Wittig reaction between
imino phosphorane 6 and m-tolualdehyde. The control aza-
Wittig reaction was performed in dry CDCl3 at 25 °C ([2]
) [m-tolualdehyde] ) [4-bromophenylacetic acid] ) 15 mM)
forming imine 4. As before, the extent of the reaction was
assayed at regular intervals by 500 MHz 1H NMR spectros-
copy and reaction progress was assessed as described
previously. The concentration-time profile reveals (Figure
3, open circles) that the initial rate of formation of imine 4
through the bimolecular aza-Wittig pathway is similar (0.59
mM h-1) to the condensation reaction between 2 and
m-tolualdehyde (0.73 mM h-1). However, in contrast to the
condensation reaction, the irreversible nature of the aza-
Wittig reaction ensures that the conversion to imine 4 is
higher.
When the reaction was repeated with the recognition-
capable aldehyde 2 in place of the control aldehyde ([2] )
[6] ) 15 mM; dry CDCl3, 25 °C), the rate of formation of
imine 3 increased dramatically (filled circles, Figure 3).
After 16 h, the conversion to imine 3 is now over 90%.
The initial rate of imine formation is accelerated by over
Acknowledgment. We thank the EPSRC for financial
support (Grant EP/E017851/1). We thank Prof. G. von
Kiedrowski (Ruhr-Universita¨t Bochum, Germany) for pro-
viding us with his SimFit program.
Supporting Information Available: Synthetic procedures
and characterization for compounds 1 to 6. Experimental
details of kinetic experiments and kinetic simulation and
fitting. Details of electronic structure calculations. This
material is available free of charge via the Internet at
http://pubs.acs.org.
(14) We have demonstrated previously that recognition processes can
be exploited in stabilizing otherwise unfavorable products and, hence, make
unfavorable reaction processes proceed in a productive sense. See, for
example: (a) Bennes, R.; Philp, D.; Spencer, N.; Kariuki, B. M.; Harris,
K. D. M. Org. Lett. 1999, 1, 1087. (b) Bennes, R.; Babiloni, M. S.; Hayes,
W.; Philp, D. Tetrahedron Lett. 2001, 42, 2377.
(15) Cossio, F. P.; Alonso, C.; Lecea, B.; Ayerbe, M.; Rubiales, G.;
Palacios, F. J. Org. Chem. 2006, 71, 2839.
OL8024499
304
Org. Lett., Vol. 11, No. 2, 2009