Using polarization effects to alter chemical reactivity: A simple host which enhances amine nucleophilicity

Article abstract of DOI:10.1021/ol005604m

(Matrix Presented) The rational design of a bis(phosphine oxide) host which is capable of binding a benzylic amine is presented. The ability of this host to increase the rate of addition of 4-fluorobenzylamine to N-phenylmaleimide is rationalized in terms of the enhancement of the nucleophilicity of the benzylic amine.

Full text of DOI:10.1021/ol005604m

ORGANIC
LETTERS
2
000
Vol. 2, No. 10
365-1368
Using Polarization Effects to Alter
Chemical Reactivity: A Simple Host
Which Enhances Amine Nucleophilicity
1
Peter R. Ashton, Patrizia Calcagno, Neil Spencer, Kenneth D. M. Harris,* and
Douglas Philp*
School of Chemistry, UniVersity of Birmingham, Edgbaston,
Birmingham B15 2TT, United Kingdom
d.philp@bham.ac.uk
Received February 1, 2000
ABSTRACT
The rational design of a bis(phosphine oxide) host which is capable of binding a benzylic amine is presented. The ability of this host to
increase the rate of addition of 4-fluorobenzylamine to N-phenylmaleimide is rationalized in terms of the enhancement of the nucleophilicity
of the benzylic amine.
Enzymes are capable of catalyzing a large range of reactions
with complete regio- and stereoselectivity and a variety of
mechanisms exist to facilitate catalysis. Many enzymes use
function. A key question in the design of efficient, synthetic
catalysts is the role of polarization effects in enzyme
catalysis. Recently, we described the formation of an
1
4
strong hydrogen-bond acceptors or donors to polarize
substrates toward reaction or to stabilize charge in reaction
intermediates or transition states. In many cases, these
interactions cause polarization of the substrate thus altering
exceptionally short CsH‚‚‚O hydrogen bond between a
terminal alkyne and a water molecule. The formation of this
hydrogen bond was facilitated by the presence, in the solid
state, of two hydrogen bonds between the protons of the
2
its chemical reactivity. A long-term aim for the organic
Chem. Soc. 1991, 113, 382. Kelly, T. R.; Ekkundi, V. S.; Meghani, P.
Tetrahedron Lett. 1990, 31, 3381. Mock, W. L.; Irra, T. A.; Wepsiec, J. P.;
Adhya, M. J. Org. Chem. 1989, 54, 5302. Mock, W. L.; Irra, T. A.; Wepsiec,
J. P.; Manimaran, T. L. J. Org. Chem. 1983, 48, 3619.
chemist remains the emulation of some of the specificity and
catalytic efficiency of enzymes through the synthesis of
3
artificial receptors designed to function as mimics of enzyme
(4) Kariuki, B. M.; Harris, K. D. M.; Philp, D.; Robinson, J. M. A. J.
Am. Chem. Soc. 1997, 119, 12679.
(
1) Bugg, T. An Introduction to Enzyme and Coenzyme Chemistry;
(5) This polarization requires that the effects of the hydrogen bond
donation are transmitted effectively throughout the system. This may not
be true for certain systems. See: Philp, D.; Robinson, J. M. A. J. Chem.
Soc., Perkin Trans. 2 1998, 1643.
Blackwell Science: Oxford, 1997. Fersht, A. Structure and Mechanism in
Protein Science; W. H. Freeman: New York, 1999.
(2) Breslow, R. Acc. Chem. Res. 1995, 28, 146. Murakami, Y.; Kikuchi,
J.; Hisaeda, Y.; Hayashida, O. Chem. ReV. 1996, 96, 721. Kirby, A. J.
(6) Schrader has described similar bis(phosphonate) and bis(phosphate)
hosts which bind several classes of compounds which bear amino groups.
See, for example: Schrader, T. J. Org. Chem. 1998, 63, 264-272. Schrader,
T. J. Am. Chem. Soc. 1998, 120, 11816-11817. Herm, M.; Schrader, T.
Chem. Eur. J. 2000, 6, 47-53.
(7) All calculations were performed using SPARTAN (Version 5.1.3,
Wavefunction Inc., Irvine, CA, 1999). Trimethylphosphine oxide was used
in place of triphenylphosphine oxide to reduce computation time. An initial
guess structure for the 2:1 complex between trimethylphosphine oxide and
methylamine was generated from the crystal structure reported in ref 4.
Angew. Chem., Int Ed. Engl. 1996, 35, 707.
(3) Hosseini, M. W.; Lehn, J. M.; Jones, K. C.; Plute, K. E.; Mertes, K.
B.; Mertes, M. P. J. Am. Chem. Soc. 1989, 111, 6330. Huc, I.; Pieters, J.;
Rebek, J., Jr. J. Am. Chem. Soc. 1994, 116, 10296. Tecilla, P.; Jubian, V.;
Hamilton, A. D. Tetrahedron 1995, 51, 435. Kang, J.; Hilmersson, G.;
Santamaria, J.; Rebek, J., Jr. J. Am. Chem. Soc. 1998, 120, 7389. Kang, J.;
Hilmersson, G.; Santamaria, J.; Rebek, J., Jr. J. Am. Chem. Soc. 1998, 120,
3
650. Marty, M.; Clyde-Watson, Z.; Twyman, L. J.; Nakash, M.; Sanders,
J. K. M. Chem. Commun. 1998, 2265. Hamilton, A. D.; Hirst, S. C. J. Am.
1
0.1021/ol005604m CCC: $19.00 © 2000 American Chemical Society
Published on Web 04/19/2000

7
water molecule and the oxygen atoms of two triphenylphos-
pine oxide (TPPO) molecules, as shown schematically in
Figure 1a. In this aggregate, the water molecule donates two
performed calculations on methylamine and a methylamine:
trimethylphosphine oxide aggregate at the HF/6-31G(d,p)
level of theory. The results (Figure 2) of these calculations
Figure 2. Values of the electrostatic potential at nitrogen in
methylamine and in the [MeNH
the HF/6-31G(d,p) level of theory (in kcal mol ).
2
2
‚TMPO ] aggregate calculated at
-
1
suggest that there is a large increase in the magnitude of the
most negative value of the electrostatic potential at nitrogen
on complexation of methylamine by two trimethylphosphine
oxide (TMPO) molecules. We were therefore confident that
hosts such as mPO should enhance the nucleophilicity of
bound amines significantly.
Figure 1. (a) A triphenyl phospine oxide:water aggregate facilitates
a short CsH‚‚‚O hydrogen bond through polarization. (b) A bis-
8
The hosts mPO and pPO were synthesized in two steps
(
phosphine oxide) receptor enhances amine nucleophilicity through
polarization.
starting from the appropriate xylylene dibromide. Lithio-
diphenylphosphide was prepared by reaction of chlorodiphen-
ylphosphine with lithium metal in dry THF. Addition of the
appropriate xylylene dibromide to the solution of lithio-
diphenylphosphide afforded the corresponding diphenylphos-
phines which were oxidized directly by treatment with
hydrogen bonds, one to each of the TPPO acceptors. This
double donation increases the partial negative charge on the
oxygen atom of the water molecule thus improving its ability
5
to accept a further hydrogen bond from the alkyne. Ab initio
quantum mechanics calculations suggest that the negative
electrostatic potential on the oxygen atom increased by more
than 40% as a result of the presence of the two hydrogen
bonds to the two TPPO molecules. We reasoned that this
polarization effect could also be exploited to enhance the
nucleophilicity of a primary amine (Figure 1b) through the
synthesis of a bis(phosphine oxide) host.
2 2
mCPBA in CH Cl to afford mPO and pPO in overall yields
of 22% and 17%, respectively. An analogous route, starting
from benzyl bromide, was used to synthesize the control
compound bDPPO.
Unfortunately, the receptor pPO was insufficiently soluble
in nonpolar solvents, in which hydrogen bonding to primary
amines would be expected to be strongest, to permit any
systematic studies of its ability to bind primary amines.
However, immediate evidence of the ability of mPO to bind
primary amines came from LSI mass spectrometry. When
Data from our previous X-ray diffraction studies suggested
that the optimum P‚‚‚P distance required in a host designed
to bind to primary amines was between 5.8 and 7.0 Å.
3
CHCl solutions of 1:1 mixtures of mPO and a selection of
6
Therefore, we designed the bis(phosphine oxide) hosts mPO
primary amines were allowed to evaporate and the residues
9
and pPO in which two diphenylphosphine oxide groups are
bridged by a xylene-derived spacer. Molecular modeling
suggested that the P‚‚‚P distances are 5.9 Å in the case of
mPO and 7.0 Å in the case of pPO. Additionally, we planned
to use benzyldiphenylphosphine oxide bDPPO as a com-
parison compound for the two bis(phosphine oxides).
analyzed by LSI mass spectrometry, signals (Table 1) arising
from 1:1 complexes between mPO and the amines were
evident.
Further evidence for the association of mPO with a variety
1
of primary amines was evident in the H NMR spectra of
1:1 mixtures of mPO and primary amines recorded in CDCl
solution at room temperature. Strong downfield shifts (∼1.5
3
10
+
(
8) Spectroscopic data for mPO: mp 58.8-65.7 °C; m/z (FAB ) 507
+
2
(
1
[MH ], 100%), 306 (5), 201 (9); δH (300 MHz; CDCl3) 3.53 (d, JPH )
3.6, ArCH2, 4H), 6.85-7.05 (m, ArH, 4H), 7.31-7.72 (m, ArH, 20H);
1
3
δC (75 MHz; CDCl3) 37.8 (d, JPC ) 66.2 Hz), 128.4 (d, JPC ) 11.8 Hz),
2
3
3
1
)
28.4 (d, JPC ) 11.8 Hz), 129.6, 131.1 (d, JPC ) 9.2 Hz), 131.5 (d, JPC
1
2
9.6 Hz), 131.8, 131.9 (d, JPC ) 99.4 Hz), 132.1, 132.2 (d, JPC ) 22.0
Hz).
(
9) Liquid secondary ion mass spectrometry (LSIMS) data were obtained
Ab initio quantum mechanics calculations were used in
an attempt to quantify the effect of the formation of two
hydrogen bonds to the amine on the electrostatic potential
surface around the amine nitrogen atom. Accordingly, we
+
using a VG Zabspec instrument equipped with a Cs ion source and utilizing
a m-nitrobenzyl alcohol matrix.
(10) The chemical shift of the amine NH2 resonance is concentration
dependent. The values for chemical shift changes quoted in the text are
corrected for this concentration dependence.
1366
Org. Lett., Vol. 2, No. 10, 2000

Amine 1 and maleimide 2 were allowed to react in CDCl
3
solution at 40 °C from starting concentrations of 50 mM for
both 1 and 2. The progress of the reaction was monitored
by 400 MHz H NMR spectroscopy and the concentrations
Table 1. Mass Spectrometric Data for the Complexation of
Primary Amines by mPO
1
of products 3 and 4 were determined from these data. The
kinetic data (Figure 3) indicate that the reaction is relatively
slow under these conditions and that 3 and 4 are formed in
approximately equal amounts. When the reaction was
repeated under identical conditions in the presence of 1 equiv
of the control phosphine oxide bDPPO, no difference in the
rates of formation of 3 and 4 was observed. By contrast,
when the same reaction was performed under identical
conditions in the presence of 1 equiv of the receptor mPO,
the kinetic data (Figure 3) indicate that the addition of mPO
a
Mass of expected molecular ion for a 1:1 complex between mPO and
b
amine shown. Mass of observed molecular ion and intensity relative to
base peak in %.
ppm) of the resonances arising from the primary amino group
were observed in all cases. Much smaller chemical shift
1
changes were observed in other H resonances and in the
3
1
P resonance of mPO. All of these data were used to
determine an approximate10 association constant for the
-1
complex between mPO and 4-fluorobenzylamine of 50 M .
To assess the effect of complexation on the nucleophilicity
of the amine nitrogen lone pair, we decided to study the
reaction between 4-fluorobenzylamine 1 and N-phenyl-
maleimide 2 (Scheme 1). This reaction can give rise to two
Scheme 1
Figure 3. Kinetic data obtained for reactions between 1 and 2 in
CDCl at 40 °C in the absence of mPO (open symbols) and in the
3
presence of 1 equiv of mPO (filled symbols). In both cases, the
starting concentrations of 1 and 2 were 50 mM. Circles represent
the concentration of 3 and triangles represent the concentration of
4.
dramatically increases the rate of reaction between 1 and 2.
Additionally, the production of bisamide 3sthe product of
direct addition to the imide carbonylsis enhanced signifi-
cantly over the production of aminosuccinimide 4sthe
conjugate addition product. This observation suggests that,
in addition to increasing the reaction rate through an
enhancement of the partial negative charge at nitrogen, the
amine nucleophile has also become harder and now prefers
to react with the harder electrophilic center, namely the
carbonyl group. The reaction between 1 and 2 was also
performed in the presence of 0.2 equiv of the receptor mPO.
In this case, the kinetic data were almost identical to those
obtained when the reaction was performed in the presence
of 1 equiv of mPO.
products (Scheme 1): bisamide 3, which is the result of direct
addition to one of the carbonyl groups of 2, and the
aminosuccinimide 4, which is the result of conjugate addition
to the terminus of the R,â unsaturated system present in 2.
(11) Kinetic simulation and fitting was performed with Gepasi (Version
.21). Mendes, P. Comput. Appl. Biosci. 1993, 9, 563. Mendes, P. Trends
3
Biochem. Sci. 1997, 22, 361. Mendes, P.; Kell, D. B. Bioinformatics 1998,
14, 869.
Org. Lett., Vol. 2, No. 10, 2000
1367

presence and in the absence of mPO. These results are
summarized in Table 2.
Table 2. Results of Kinetic Simulation and Fitting of the
Kinetic Data for the Reaction between 1 and 2 in the Presence
and Absence of mPO
It is clear from the data presented in Table 2 that the
production of 3 and 4 are accelerated significantly when
4
-fluorobenzylamine is bound to the receptor mPO. In
best fit value,
rate constanta
10-5 M-1 s-1 b
particular, the rate constant for the formation of 3 via the
catalyzed pathway is between six and seven times higher
than that in the uncatalyzed pathway.
k3(uncat)
k4(uncat)
k3(cat)
5.11 ( 0.08
5.09 ( 0.03
33.8 ( 0.7
15.2 ( 0.9
In conclusion, we have demonstrated the rational design
of a host which is capable of accelerating the nucleophilic
addition of an amine to a maleimide through a reagent
polarization effect. We are currently engaged in extending
this methodology to other reaction classes and in the design
of more complex host systems.
k4(cat)
a
The rate constant descriptors refer to Scheme 1. b Simulation and fitting
were carried out by first obtaining values for k3(uncat) and k4(uncat) from
the data collected in the absence of mPO. These values were then used in
the derivation of best fit values for k3(cat) and k4(cat) from fitting of the
kinetic data obtained in the presence of 1 equivalent of mPO to the model
shown in Scheme 1. Errors in rate constants are quoted at the 2σ level, on
the basis of the standard deviations obtained from the kinetic simulation
and fitting procedure.
Acknowledgment. This research was supported by the
University of Birmingham (postgraduate studentship for P.C.)
and the Biotechnology and Biological Sciences Research
Council (grants 6/B03840 and 6/B04100).
Kinetic simulation and fitting11 of the reaction profiles to
the model shown in Scheme 1 allowed us to extract values
for the rate constants for the addition of 1 to 2 both in the
OL005604M
1368
Org. Lett., Vol. 2, No. 10, 2000