Kinetic stabilization of N, N-dimethyl-2-propyn-1-amine N-oxide by encapsulation

Article abstract of DOI:10.1021/ol402236z

Thermally and in nonprotic media N,N-dimethyl-2-propyn-1-amine N-oxide undergoes two consecutive sigmatropic rearrangements affording propenal. Two molecular containers capable of the quantitative inclusion/encapsulation of the N-oxide are described. The

Full text of DOI:10.1021/ol402236z

ORGANIC
LETTERS
2013
Vol. 15, No. 19
4976–4979
Kinetic Stabilization of
N,N‑Dimethyl-2-propyn-1-amine
N‑Oxide by Encapsulation
†
Albano Galan, Guzman Gil-Ramırez, and Pablo Ballester*
†
,†,‡
ꢀ
ꢀ
´
Institute of Chemical Research of Catalonia (ICIQ), Avgda. Paısos Catalans 16,
¨
43007 Tarragona, Spain, and Catalan Institution for Research and Advanced Studies
´
(ICREA), Passeig Lluıs Companys, 23, 08018 Barcelona, Spain
pballester@iciq.es
Received August 6, 2013
ABSTRACT
Thermally and in nonprotic media N,N-dimethyl-2-propyn-1-amine N-oxide undergoes two consecutive sigmatropic rearrangements affording
propenal. Two molecular containers capable of the quantitative inclusion/encapsulation of the N-oxide are described. The N-oxide becomes
kinetically stabilized when included in the containers. The relationship between the observed kinetic stabilization of the N-oxide and the
thermodynamic and kinetic stability of its inclusion complexes is explained and modeled.
In recent years, different types of synthetic molecular
containers have been developed as mimics of protein
binding sites, antibodies, and enzymes.1À3 Molecular con-
tainers are molecules or supramolecules able to reproduce
some of the properties exhibited by their biological coun-
terparts. These properties include the reversible inclusion/
encapsulation of molecular guests within their internal
cavities, the isolation of the included guests from the bulk
solvent, and the selection of guests on the basis of hostÀ
guest complementarity in size, shape, and functional groups.
In relevant examples, synthetic molecular containers can
efficiently promote chemical reactions of encapsulated4,5
and coencapsulated substrates,6,7 altering the typical
regio- and/or stereochemical outcome of the reactions free
in solution.8,10 They can also kinetically stabilize highly
reactive intermediates9 or substrates,11 bind guests in high
energy conformations,12À15 as well as modify the catalytic
properties of encapsulated organometallic complexes.16,17
In this communication, we describe a significant kinetic
stabilization of the N,N-dimethyl-2-propyn-1-amine
N-oxide, 1a, achieved by supramolecular inclusion in two
different molecular containers. We also present a simple
kinetic model that quantitatively relates the observed
kinetic stabilization with the thermodynamic/kinetic sta-
bility of the inclusion complexes (Scheme 1). The reported
(8) Hastings, C. J.; Backlund, M. P.; Bergman, R. G.; Raymond,
K. N. Angew. Chem., Int. Ed. 2011, 50, 10570–10573.
(9) Restorp, P.; Rebek, J., Jr. J. Am. Chem. Soc. 2008, 130, 11850–
11851.
(10) Yoshizawa, M.; Tamura, M.; Fujita, M. Science 2006, 312, 251–
254.
(11) Mal, P.; Breiner, B.; Rissanen, K.; Nitschke, J. R. Science 2009,
324, 1697–1699.
(12) Breiner, B.; Clegg, J. K.; Nitschke, J. R. Chem. Sci. 2011, 2,
51–56.
(13) Takezawa, H.; Murase, T.; Fujita, M. J. Am. Chem. Soc. 2012,
134, 17420–17423.
† Institute of Chemical Research of Catalonia.
‡ Catalan Institution for Research and Advanced Studies.
(1) Yoshizawa, M.; Klosterman, J. K.; Fujita, M. Angew. Chem., Int.
Ed. 2009, 48, 3418–3438.
(2) Breiner, B.; Clegg, J. K.; Nitschke, J. R. Chem. Sci. 2011, 2, 51–56.
(3) Wiester, M. J.; Ulmann, P. A.; Mirkin, C. A. Angew. Chem., Int.
Ed. 2011, 50, 114–137.
(4) Fiedler, D.; van Halbeek, H.; Bergman, R. G.; Raymond, K. N.
J. Am. Chem. Soc. 2006, 128, 10240–10252.
(5) Hooley, R. J.; Rebek, J., Jr. Org. Biomol. Chem. 2007, 5, 3631–
3636.
(14) Korner, S. K.; Tucci, F. C.; Rudkevich, D. M.; Heinz, T.; Rebek,
J. Chem.;Eur. J. 2000, 6, 187–195.
(15) Porel, M.; Jayaraj, N.; Raghothama, S.; Ramamurthy, V. Org.
Lett. 2010, 12, 4544–4547.
(6) Horiuchi, S.; Murase, T.; Fujita, M. Chem.;Asian J. 2011, 6,
1839–1847.
(7) Kang, J. M.; Hilmersson, G.; Santamaria, J.; Rebek, J., Jr. J. Am.
Chem. Soc. 1998, 120, 3650–3656.
(16) Wang, Z. J.; Clary, K. N.; Bergman, R. G.; Raymond, K. N.;
Toste, F. D. Nat. Chem. 2013, 5, 100–103.
(17) Horiuchi, S.; Murase, T.; Fujita, M. Angew. Chem., Int. Ed.
2012, 51, 12029–12031.
r
10.1021/ol402236z
Published on Web 09/19/2013
2013 American Chemical Society

results constitute unique examples of the use of molecular
containers in the kinetic stabilization of reactive species,
which would otherwise fragment through sigmatropic
rearrangements.
Scheme 1. Proposed Mechanism for the Thermal Decomposition
of Tertiary Prop-2-ynylic N-Oxides 1
Thermally and in nonprotic media, tertiary propargyla-
mine N-oxides undergo a concerted [2,3]-sigmatropic re-
arrangement to yield O-allenyl ethers 2.18,19 In turn, these
O-allenyl ethers are kinetically unstable and experience a
rapid elimination, akin to a [1,5]-hydrogen shift, generat-
ing propenal 3 and the corresponding Schiff base 4.
The kinetic stability of the intermediate O-allenyl ether 2
depends on the substituents on the nitrogen atom.20
We hypothesized that the confinement of 1a in properly
inner functionalized molecular containers could reduce or
even eliminate its tendency to decompose by increasing
the energy difference between bound 1a and the transition
state (TS) leading to the O-allenyl ether 2a, compared to
that of the reaction occurring free in solution. Both the
formation of hydrogen bonds with the oxygen atom of the
N-oxide 1a (reduce oxygen nucleophilicity) and the induc-
tion of steric strain in the cis-conformation of 1a (required
in the TS) seemed to be good starting points to achieve the
goal. Due to the extensive use of aromatic panels to shape
the concave cavities of uni- and supramolecular contain-
ers, it is synthetically challenging to place polar groups in
their interiors.21,22 In addition, the lack of inner functio-
nalizationdisfavors theinclusion of polarguests, rendering
the selectivity of the encapsulation’s process mainly deter-
mined by size and shape complementarity. In trying
to overcome these limitations, we and others have used
aryl-extended calix[4]pyrroles as privileged scaffolds for
the construction of unimolecular23 and supramolecular
containers with polar interiors.24
Figure 1. Molecular structures of N-oxides 1a and 7, containers
5 and 6, and reference compounds 8 and 9. Top right: Side and
top views of the energy-minimized complex 1a⊂5. Note that 1a
must adopt the trans-conformation to fit in the aryl-extended
calix[4]pyrrole binding pocket of 5.
that are structurally related to tetranitro 5 (Figure 1).25 We
demonstrated that N-oxides can be encapsulated alone or
coencapsulated with a CHCl3 molecule in the mechanically
locked capsule 6 (Figure 1).26 Molecular modeling studies
suggested a good match between the trans-conformation of
1a and the size, shape, and functionality of the inner cavities
of the calix[4]pyrrole units of containers 5 and 6. Conver-
sely, the cis-conformer of 1a is not a good fit. Interestingly,
trans-1a forms four hydrogen bonds with the endohedrally
directed pyrrole NHs of the calixpyrrole units of the con-
tainers. Thus, the included/encapsulated trans-1a displays
the features planned above to alter its inherent reactivity
(Figure 1).
The N-oxide of N,N-dimetyl-2-propyn-1-amine, 1a, was
obtained as a white solid in 80% yield from reacting
equimolar amounts of m-chloroperbenzoic acid and the
amine in chloroform solution at 25 °C followed by column
chromatography through alumina.20 N-Oxide 1a was
stable for months when stored as a solid in a refrigerator.
1
In contrast, the H NMR analysis of a 60 mM dichlor-
omethane solution of 1a at 298 K provided evidence for
the decomposition of the N-oxide with the appearance of
proton signals characteristic for propenal 3. The intensity
of the proton signals of 3 grew at the expense of those
assigned to 1a. The observation of the proton signals
corresponding to 3 indicated (a) that the O-allenyl ether
intermediate 2a is not kinetically stable and (b) that the
N-methylenemethanamine 4a, a byproduct of the formal
[1,5]-hydrogen shift reaction, decomposes rapidly.27
1H NMR spectroscopy was used to monitor the decom-
position kinetics of a freshly prepared CD2Cl2 solution of
We previously reported the efficient inclusion of N-oxides
in water-soluble “four-wall” aryl-extended calix[4]pyrroles
(18) Craig, J. C.; Ekwuribe, N. N.; Gruenke, L. D. Tetrahedron Lett.
1979, 20, 4025–4028.
(19) Khuthier, A. H.; Aliraqi, M. A.; Hallstrom, G.; Lindeke, B.
J. Chem. Soc., Chem. Commun. 1979, 9–10.
(20) Hallstrom, G.; Lindeke, B.; Khuthier, A. H.; Aliraqi, M. A.
Tetrahedron Lett. 1980, 21, 667–670.
(21) For recent reviews of inner functionalization of molecular
containers, see: (a) Adriaenssens, L.; Ballester, P. Chem. Soc. Rev.
2013, 42, 3261–3277. (b) Kubik, S. Molecular Cages and Capsules with
Functionalized Inner Surfaces. In Chemistry of Nanocontainers;
Albrecht, M., Hahn, E., Eds.; Springer: Berlin, 2011; Vol. 319, pp 1À34.
(22) Young, M. C.; Johnson, A. M.; Gamboa, A. S.; Hooley, R. J.
Chem. Commun. 2013, 49, 1627–1629.
(23) Park, I. W.; Kim, S. K.; Lee, M. J.; Lynch, V. M.; Sessler, J. L.;
Lee, C. H. Chem.;Asian J. 2011, 6, 2911–2915.
(24) Adriaenssens, L.; Ballester, P. Chem. Soc. Rev. 2013, 42, 3261–
3277.
(25) Verdejo, B.; Gil-Ramirez, G.; Ballester, P. J. Am. Chem. Soc.
2009, 131, 3178–3179.
(26) Chas, M.; Ballester, P. Chem. Sci. 2012, 3, 186–191.
(27) Cattoen, X.; Miqueu, K.; Gornitzka, H.; Bourissou, D.;
Bertrand, G. J. Am. Chem. Soc. 2005, 127, 3292–3293.
Org. Lett., Vol. 15, No. 19, 2013
4977

1
1a (Figure 2a,b). The H NMR spectrum of the solution
recorded after 3 h showed the expected proton signals for
3 (δ = 9.6 ppm, CHO, and δ = 6.5 ppm, HCdCH2)
(Figure 2b). We followed the change in the integral values
for both the proton signals of N-oxide 1a and propenal
3 during two half-lives. The decomposition process was
investigated at three different initial concentrations of 1a
(26, 60, and 120 mM). The experimental kinetic data were
fitted, using global multivariate factor analysis and the
differential kinetics module implement in SPECFIT,28 to a
reaction model considering the direct conversion of 1a to 3.
We obtained very good fits of the experimental data, for
both the decomposition and formation reactions, at any of
the concentrationsused(see Supporting Information). The
averaged value of the rate constant returned from the fits
was kobs = 5.0 ( 0.3 Â 10À5 sÀ1. The half-life of N-oxide 1a
free insolutionis3.8( 0.2 h and, as expectedfor first-order
reactions (t1/2 = 0.693/kobs), independent of its initial
amount. The analysis of the kinetic data assumes that the
rate-determining step of the reaction is the [2,3]-sigmatropic
rearrangement of 1a affording the O-allenyl ether 2. This
hypothesis is substantiated by the following evidence: (a)
the intermediate O-allenyl ether 2 is not detected experi-
1
Figure 2. Selected regions of the H NMR spectra (500 MHz,
298 K) of (a) a freshly prepared 26 mM CD2Cl2 solution of 1a;
(b) a solution of 1a after 3 h; (c) a solution of 1a in the presence of
1 equiv of 5 after 13 h; (d) a freshly prepared equimolar solution
of 1a and 5. See Scheme 1 and Figure 1 for proton assignment.
Primed numbers indicate signals of protons bound.
1
mentally in the H NMR kinetic experiments and (b) the
results are obtained from the theoretical investigation of the
reaction pathways.29 The addition of an equimolar amount
of 5 to a freshly prepared 2.5 mM CD2Cl2 solution of 1a
had a strong impact on the decomposition kinetics. The
first half-life of 1a in the presence of 5 was estimated by
simple visual extrapolation of the curves of disappearance
of 1a⊂5 and formation of 3 as t1/2 > 50 h. This represents
an almost 18-fold increase compared to the kinetic stability
of 1a free in solution. The 1H NMR analysis of the freshly
prepared equimolar mixture of 1a and 5 reveals the quanti-
tative assembly of the 1a⊂5 inclusion complex (Figure 2d).
We sought to determine the relationship between the
observed kinetic stabilization of 1a and the thermodynamic
and kinetic properties of the inclusion complex 1a⊂5.30
To this end, the kinetic data for the decomposition of 1a in
the presence of 5 were fitted, using the differential kinetics
module of SPECFIT, to a model that also considers the
reversible formation of the inclusion complex 1a⊂5
(Scheme 2). This kinetic model implicitly assumes that the
decomposition of 1a occurs exclusively when the species
is free in solution. In order to minimize the number of
variables to fit, the value of the rate constant k1 for the
decomposition of free 1a was fixed to 5.0 Â 10À5 sÀ1. This is
the value calculated in the absence of molecular container 5
(vide supra).
Moreover, complex 7⊂5 was employed as an ideal system
to derive reference values for the kinetic and thermo-
1
dynamic properties of 1a⊂5. The 1D H NMR spectrum
of a CD2Cl2 solution of 5 containing 0.5 equiv of N-oxide 7
showed separated proton signals for free and bound tetra-
nitro host 5. In addition, half of the total amount of
container 5 is bound, and no proton signals for the free
N-oxide 7 are detected. These observations indicated that
the binding process is slow on the 1H NMR time scale and
that the stability constant of the inclusion complex can be
estimated at K(7⊂5) >104 MÀ1. Weperformeda2D EXSY
experiment on the mixture at 298 K. Using the integration
values of the diagonal and cross-peaks, we determined the
rate constant for the magnetization exchange between
bound and free states of 5 as kÀ1mag = 10 sÀ1. This value
is directly related to the dissociation rate constant of the
7⊂5 complex koff = kÀ1mag = 10 sÀ1 and was used to set the
value of dissociation rate constant for the 1a⊂5 complex in
the data fitting.
Scheme 2. Kinetic Model Involving the Formation of the
Stabilizing Inclusion Complex 1a⊂5
(28) SPECFIT/32 from Spectrum Software Associates.
(29) The geometries of the transition states (TS) for each one of the
two sequential rearrangements were determined using DFT calculations
at the M06-2X/aug-cc-pVDZ level of theory and using a continuum
solvation model for CH2Cl2. The free energy differences (ΔG*) with
respect to trans-1a for TS1[2,3] and TS2[1,5] are 24.3 and 22.0 kcal/mol,
respectively. TS1 is the rate-determining step, and the rate law obeys
Àk1[1]. The sensible agreement that exists between the theoretical and
experimental (23.1 kcal/mol) values determined for ΔG* supports the
Next, the kinetic data of the supramolecular system 1a/5
(changes in concentrations of 3, 5, and 1a⊂5 species vs
time) were analyzed with the kinetic model in Scheme 2,
and k1 and koff were fixed to the above values. The process
provided good fits to the experimental data and returned
the value of the second-order rate constant for complex
identity of k1 = kobs
.
(30) The half-life of a substrate depends on its initial concentration
for all reactions not following first-order kinetics.
4978
Org. Lett., Vol. 15, No. 19, 2013

formation, kon = 1.2 Â 105 MÀ1 sÀ1, as the only variable
to refine. The ratio of formation and dissociation rate
constants for the 1a⊂5 complex affords its thermodynamic
stability constant as K(1a⊂5) = 1.2 Â 104 MÀ1. The
calculated value is in very good agreement with the one
measured using ITC experiments for the 7⊂5 model com-
plex (K(7⊂5) = 1.3 ( 0.2 Â 104 MÀ1). This result gives
clear indication of the quality and fit of the kinetic data
analysis used. It also demonstrates that the kinetic stabi-
lization of 1a in the presence of 5 is not due to the release
of 1a from the container being rate-limiting, but simply to
the significant reduction of the concentration of free 1a in
solution.
Aslong asthe release of 1afrom the container isnotrate-
determining, the kinetic model quantitatively predicts
the relationship between the apparent kinetic stabiliza-
tion of 1a and the thermodynamic/kinetic stability of the
complexes inhibiting decomposition. Thus, control com-
pounds 8 and 9 with K(7⊂8 or 7⊂9) < 50 MÀ1 proved to
be completely inefficient in the kinetic stabilization of 1a
(Supporting Information). Next, we simulated the ex-
pected kinetic speciation for propenal 3 from a 0.5 mM
equimolar mixture of 1a and a putative container forming
anencapsulation complexwith a stability constant 2orders
of magnitude larger than that for 1a⊂5 (Supporting
Information).31 The first half-life of 1a was determined
as 45 days using curve interpolation of the simulated
changes in the concentration of 3 versus time (Supporting
Information). This represents a significant kinetic stabiliza-
tion of 1a when compared to the experimentally measured
first half-lives of 3.8 h for 2.5 mM solutions of free 1a or
>50 h in the presence of 1 equiv of 5. Inshort, after45 days,
only 50% of 1a in solution should have decomposed
to propenal 3 if stored (encapsulated) in such molecular
container.
earmarks for quantitative formation of the encapsulation
complex 1a⊂6 (Figure S2a). The pyrrole NH protons
moved downfield (δ = 10.2 ppm; Δδ = 1.6 ppm) as a
result of hydrogen bonding interaction between the NHs
and the oxygen atom of the N-oxide. The protons of the
methyl groups for the bound guest appear upfield shifted
and resonate as two diastereotopic signals (δ = 0.82 and
0.74 ppm) due to the chiral nature of the container. Even
after 2 months, the weekly 1H NMR spectroscopy analysis
of the 0.5 mM CD2Cl2 solution containing the 1a⊂6
complex did not reveal the presence of signals for the
protons of 3.33 This observation suggests that K(1a⊂6)
should be more than 2 orders of magnitude larger than
1
K(1a⊂5) (Figure S12). The series of H NMR spectra
acquired during this time period indicated that the encap-
sulation complex 1a⊂6 remained unaltered in solution
(Figure S4). Finally, the decomposition of 1a into 3
was induced by competitive release from the container
(Supporting Information).
In conclusion, we demonstrated that the inherent reac-
tivity of N-oxide 1a affording 3 can be modulated by
inclusion/encapsulation in molecular containers. The con-
tainers used feature concave, polar interiors based on aryl-
extended calix[4]pyrrole scaffolds. N-Oxide 1a is encapsu-
lated in trans-conformation and forms four hydrogen
bonds between its oxygen atom and the pyrrolic NH
protons of the container. We derived a kinetic model that
quantitatively explains the relationship between the appar-
ent kinetic stabilization of 1a and the thermodynamic/
kinetic stability of the complexes formed with the mole-
cular flasks. The kinetic stabilization does not result from
a rate-determining dissociation of 1a from the container
but simply from the significant reduction of its concentra-
tion free in solution in the presence of the molecular
containers.
N-Oxides 1a and 7 are also nicely encapsulated in the
calix[4]pyrrole hemisphere of the mechanically locked
capsule 6, while the calix[4]arene hemisphere is occupied
by one molecule of chloroform (Figure S3).26,32 The value
of K(7⊂6) was estimated using ITC experiments as >1.0 Â
106 MÀ1, supporting the suitability of 6 asa molecularflask
for extending the first half-life of 1a in the liquid state. The
1H NMR spectrum of a 0.5 mM CD2Cl2 solution contain-
ing equimolaramountsof1aand 6 shows the characteristic
Acknowledgment. We are grateful for funding from
~
Gobierno de Espana MINECO (CTQ2011-23014), Gen-
eralitat de Catalunya DURSI (2009SGR6868), and ICIQ
Foundation. A.G. thanks MINECO for a FPU fellowship.
Supporting Information Available. Experimental pro-
1
cedures, H NMR experiments, fits of the kinetic data
simulations, and ITC experiments. This material is avail-
able free of charge via the Internet at http://pubs.acs.org.
(31) The speciation simulations provide identical results indepen-
dently of the absolute values used for kon and koff, as long as its ratio is
maintained constant and equal to 2.0 Â 106 MÀ1 and koff > 5 Â 10À5 sÀ1
.
(33) We estimate that the lower detection limit of our analysis for 3 is
0.2 mM.
(32) In dichloromethane solution, the chemical exchange between the
encapsulated solvent molecule and the bulk solvent molecules is fast on
the 1H NMR time scale even at low temperatures.
The authors declare no competing financial interest.
Org. Lett., Vol. 15, No. 19, 2013
4979