Evidence for the importance of polarizability in biomimetic catalysis involving cyclophane receptors

Article abstract of DOI:10.1021/jo960521y

Cyclophanes 1-6 catalyze the nucleophilic dealkylation of a simple sulfonium compound by potassium iodide. The cation-π interaction is important in substrate binding, but the primarily electrostatic nature of this effect does not explain all observations concerning catalysis. As a series of substituents are placed on the cyclophane framework, a systematic variation in catalyst effectiveness is seen, such that more polarizable substituents produce more potent catalysts. This provides support for the notion that transition states are especially polarizable, and catalysis can be enhanced by maximizing London dispersion forces. The reactions studied here are very similar to the broad class of biological methylations mediated by S-adenosylmethionine, and the biological catalysts may use forces similar to those described here.

Full text of DOI:10.1021/jo960521y

J . Org. Chem. 1996, 61, 4355-4360
4355
Evid en ce for th e Im p or ta n ce of P ola r iza bility in Biom im etic
Ca ta lysis In volvin g Cyclop h a n e Recep tor s
†
Sarah M. Ngola and Dennis A. Dougherty*
Arnold and Mabel Beckman Laboratories of Chemical Synthesis, Division of Chemistry and Chemical
Engineering, California Institute of Technology, Pasadena, California, 91125
Received March 18, 1996X
Cyclophanes 1-6 catalyze the nucleophilic dealkylation of a simple sulfonium compound by
potassium iodide. The cation-π interaction is important in substrate binding, but the primarily
electrostatic nature of this effect does not explain all observations concerning catalysis. As a series
of substituents are placed on the cyclophane framework, a systematic variation in catalyst
effectiveness is seen, such that more polarizable substituents produce more potent catalysts. This
provides support for the notion that transition states are especially polarizable, and catalysis can
be enhanced by maximizing London dispersion forces. The reactions studied here are very similar
to the broad class of biological methylations mediated by S-adenosylmethionine, and the biological
catalysts may use forces similar to those described here.
In tr od u ction
special stabilization afforded the transition state by the
microenvironment of the host.
Cyclophane 1 is a general receptor for organic cations
such as quaternary ammonium, imminium, and dialky-
larylsulfonium compounds.1 Complexation arises in part
from the cation-π interaction, in which the positive
charge of the guest is stabilized through its interaction
2
with the π face of simple aromatic systems. We have
shown that, to a considerable extent, binding through the
cation-π interaction can be thought of as primarily
electrostatic in nature, involving the quadrupole moment
of the aromatic and the charge of the ion.2
-4
Of course,
a full, quantitative model of the cation-π interaction
would require consideration of additional terms, such as
the interaction of the ion with induced dipoles in the
aromatics.
Both an alkylation reaction involving formal creation
of positive charge in the transition state (eq 1) and a
dealkylation reaction in which positive charge is formally
destroyed (eq 2) are accelerated by the host. This
establishes that a transition state with only a partial
positive charge can be stabilized more effectively than
Given the general cation-binding ability of host 1, it
was not surprising that it is also a catalyst for S 2
N
reactions in which the transition state has a partial
positive charge (eqs 1 and 2).5 This catalysis by the
host, carried out in aqueous media, is biomimetic, in that
it involves prior binding of a substrate followed by
reaction in the binding cavity. In contrast to most
biomimetic systems, however, there is no contribution
from proximity effects. Both the catalyzed and uncata-
lyzed reactions are bimolecular reactions, and there is
no evidence for interactions between the host and the
nucleophile to form a ternary complex with the substrate.
All rate acceleration is achieved solely through transition-
state stabilization due to interactions with the host.
While this may not be the best way to maximize the rate
enhancements seen, it does allow an evaluation of any
5
,6
either fully charged or completely neutral substrates.
,6
The first-order electrostatic analysis of the cation-π
interaction cannot account for such behavior. Instead,
in our initial analysis we proposed a prominent role for
5,6
London dispersion forces in the catalysis by host 1.
London dispersion forces arise from induced-dipole-
induced-dipole interactions between molecules, and the
strength of an induced-dipole is dependent on the polar-
izability of the molecule. Since transition states have
long, weak bonds, they are expected to be more polariz-
able than ground-state substrates or products.7 The role
of the host is to surround the transition state with an
electron-rich π-system that is polarizable and in a
relatively fixed orientation, so that induced dipoles in
both the host and the transition state are suitably
aligned. Water is much less polarizable than benzene
†
Recipient of a General Electric fellowship and an NSF predoctoral
fellowship.
(Table 1), and so the host provides a more suitable
X
Abstract published in Advance ACS Abstracts, J une 15, 1996.
environment for London dispersion interactions than an
aqueous medium.8 Note that in this model, it is not
sufficient that polarizability contributes to binding. Po-
larizability must be more important for binding transition
states than for binding ground states. Only in this way
can it enhance catalysis.
(1) Petti, M. A.; Shepodd, T. J .; Barrans, R. E., J r.; Dougherty, D.
A. J . Am. Chem. Soc. 1988, 110, 6825-6840. Kearney, P. C.; Mizoue,
L. S.; Kumpf, R. A.; Forman, J . E.; McCurdy, A.; Dougherty, D. A. J .
Am. Chem. Soc. 1993, 115, 9907-9919.
(
(
(
2) Dougherty, D. A. Science 1996, 271, 163-168.
3) Kumpf, R. A.; Dougherty, D. A. Science 1993, 261, 1708-1710.
4) Meccozzi, S.; West, A. P.; Dougherty, D. A. J . Am. Chem. Soc.
1
996, 118, 2307-8; Proc. Natl. Acad. Sci. U.S.A., in press.
5) Stauffer, D. A.; Barrans, J r., R. E.; Dougherty, D. A. Angew.
Chem., Int. Ed. Engl. 1990, 29, 915-918.
6) McCurdy, A.; J imenez, L.; Stauffer, D. A.; Dougherty, D. A. J .
Am. Chem. Soc. 1992, 114, 10314-10321.
(
(7) Blake, J . F.; Lim, D.; J orgensen, W. L. J . Org. Chem. 1994, 59,
803-805. J orgensen, W. L.; Blake, J . F.; Lim, D.; Severance, D. L. J .
Chem. Soc., Faraday Trans. 1994, 90, 1727-1732.
(
S0022-3263(96)00521-X CCC: $12.00 © 1996 American Chemical Society

4
356 J . Org. Chem., Vol. 61, No. 13, 1996
Ngola and Dougherty
Ch a r t 1
Ta ble 1. P ola r iza bilitiesa of Selected Com p ou n d s
polarizability
-
24
-3
compound
water
(10
cm
)
1.45
benzene
toluene
anisole
chlorobenzene
bromobenzene
10.32
12.3
13.1
14.1
14.7
a
CRC Handbook of Chemistry and Physics, 71st ed.; CRC
Press: Boca Raton, FL, 1990; pp 10-198-10-207.
tion patterns appear to position the substituents in
favorable locations for interacting with the transition
state. We find that the substituents exert small but
significant effects on the catalytic effectiveness of host
1
. Intriguingly, the trend in substituent effects on
catalysis does indeed parallel the expected changes in
polarizability.
Resu lts a n d Discu ssion
Hosts 1-6 were used to catalyze the dealkylation of
sulfonium salt 7 with iodide ion acting as the nucleophile
(eq 2). Significant rate enhancements are seen, and
different hosts produce measurably different rate en-
hancements, as shown in Figure 2.
The kinetics of the host-catalyzed dealkylation can be
F igu r e 1. CPK representations of the S,S,S,S enantiomer of
host 2 (top) and 6 (bottom) in rhomboid binding conformations.
The host structure was obtained from the X-ray structure of
described by a modified Michaelis-Menten scheme (Fig-
ure 3).5
,6
s p
K and K are known quantities that can be
9
1
10
the tetraester of host 1 with the methyl groups deleted for
obtained by NMR or CD binding studies. The uncata-
lyzed rate constant kun is obtained under pseudo first-
order conditions by using a large excess of the nucleo-
phile. Therefore, the only unknown is kcat, which can be
obtained by recording the change in substrate concentra-
tion over time in the presence of host and using this data
to numerically solve the rate equations for the value of
clarity. Bromine atoms were then placed in standard positions.
The oxygen atoms of the molecule are cross-hatched and the
bromine atoms are shaded.
In order to seek support for this model, we have studied
hosts 2-6 to evaluate potential substituent effects.
Special emphasis was placed on the extent to which the
substituents might alter the polarizability of the host
environment. Table 1 shows the experimental polariz-
abilities of several reference compounds, which we will
use to order the polarizabilities of the host substituents.
In host 2, the bromines are in fixed positions on the
ethenoanthracene units of the host, while in 3-6 the
substituents are attached to the more flexible “linker”
region of the host. As shown in Figure 1, both substitu-
k
cat. A typical fit is shown in Figure 4. As in previous
work,5 5-nitroquinoline (8) was a competitive inhibitor
,6
(
8) Pearson, R. G.; Sobel, H.; Songstad, J . J . Am. Chem. Soc. 1968,
0, 319. Maryott, A. A.; Buckley, F. U. S. National Bureau of Standards
Circular No. 537, 1953.
9) Forman, J . E.; Marsh, R. E.; Schaefer, W. P.; Dougherty, D. A.
Acta Crystallogr. 1993, B49, 892-896.
10) Forman, J . E.; Barrans, R. E., J r.; Dougherty, D. A. J . Am.
Chem. Soc. 1995, 117, 9213-9228.
9
(
(

Polarizability in Biomimetic Catalysis
J . Org. Chem., Vol. 61, No. 13, 1996 4357
�
F igu r e 2. Disappearance of substrate with time for selected
reactions.
F igu r e 4. The experimental rise in product with host 3 as
catalyst plotted with the simulated curve obtained from the
numerical solution of the kinetic scheme of Figure 3.
Ta ble 2. Bin d in g a n d Ca ta lysis Da ta
q
-
∆Gs
-∆∆G
a
kcat/kunb
b
host
(kcal/mol)
(kcal/mol)
1
2
3
4
5
6
5.7
6.1
5.2
4.6
5.8
5.6
4.9
5.9
6.4
9.7
9.8
1.0
1.1
1.2
1.4
1.4
1.6
F igu r e 3. Kinetic scheme for the reactions considered here
(H ) host, S ) substrate, P ) product, X ) nucleophile).
12.0
a
At 298 K. b At 319 K.
stabilized relative to substrate upon moving from water
into the host cavity.
It should be appreciated that, in one sense, it is
surprising that adding substituents should increase the
reaction rate. Since S 2 transition states are more
N
crowded than the substrate alone, it might have been
expected that adding steric bulk in the region where the
nucleophile must approach would slow the reaction. It
is possible that such adverse interactions do influence
the rate, but that they are overwhelmed by the electronic
influence of the substituents.
of the host-catalyzed dealkylation, blocking the host
binding site and effectively preventing catalysis from
taking place. This rules out catalysis by association of
the guest with a host aggregate rather than actual
binding of the guest in the host cavity.
It is also important to emphasize that all the hosts
appear to be binding the guest in a similar orientation.
A method for determining binding constants using cir-
cular dichroism (CD) has been developed recently in our
group,10 and certain features of the CD spectrum have
implications for the conformation of the host and the
binding orientations for the guest. Previous studies using
cyclophane 1 indicate that there are two major binding
The results from the study of the host-catalyzed de-
alkylation reactions are summarized in Table 2. As
discussed above, the rate enhancements (kcat/kun) in this
system are not large. Still, we believe that not only the
rate enhancements but also the differences in rate
enhancements are meaningful. Figure 2 clearly shows
that there are significant differences between the rates
of the uncatalyzed reaction and the two host-catalyzed
reactions indicated. A valid comparison between the two
host-catalyzed reactions can be made since the initial
percentage of the guest bound is the same in both cases.
This is a well-behaved kinetic system with a simple rate
law. The value of kun is the same for all systems, so the
only variables are the binding affinity of the substrate,
which can be determined with good accuracy, and the
observed rate of disappearance of substrate, which is
easily monitored by HPLC and found to be reproducible.
orientations of the host: rhomboid and toroid. The C
symmetric rhomboid form binds flat aromatic guests such
as 9, while the D -symmetric toroid form binds more
2
-
2
nearly spherical guests such as 10. This information was
used to design a study to establish the host binding
orientation in the presence of the sulfonium guest 7. CD
spectra of each host were taken in the presence of 7, 9,
and 10 as guests. These spectra were then compared,
to infer the adopted binding conformation of each host
with each of three guests. For example, with host 3 in
the presence of 10, there is a peak shift from 220 to 235
nm which is not observed in the presence of either 7 or
9 (Figure 5), and this appears to be characteristic of a
toroid binding conformation. The CD spectra of this host
in the presence of both 7 and 9 are very similar and
support a rhomboid binding conformation (as shown in
q
The quantity ∆∆G () RT ln(kcat/kun)) represents the
extent to which the transition state is preferentially

4
358 J . Org. Chem., Vol. 61, No. 13, 1996
Ngola and Dougherty
a
F igu r e 6. Host 5 with guest 7 aligned in the cavity at the
midpoint of the angle range that is calculated to produce
induced CD of the same sign as that observed experimentally.
Ta ble 3. Va lu es of th e Up field 1H NMR Sh ifts (p p m ) of
th e P r oton s of Gu est 7 Sca led to th e High est Sh ift
host
methyl protons
aromatic protons
b
1
3
4
5
6
0.58, 0.72
0.56, 0.80
0.47, 0.76
0.33, 0.52
0.30, 0.45
0.83, 1.00
0.68, 1.00
0.98, 1.00
0.90, 1.00
0.91, 1.00
compared to the induced CD observed experimentally to
infer the binding orientation of the guest as a range of
possible angles of rotation of the guest about a set axis
in the host coordinate system.10 In Figure 6, the guest 7
is shown positioned at an angle midway between the
range predicted by the coupled oscillator calculations. The
positively-charged sulfur atom is positioned for maximal
interaction with one of the aromatic rings of the host,
while one of the methyl groups is placed in close proxim-
ity to one of the chlorine atoms.
c
The binding orientation of the guest can also be
1
partially inferred from H NMR shifts seen on binding.
1
The upfield H NMR shifts of the guest protons reflect
how close a proton is to the shielding face of the aromatic
rings of the host cavity. The overall pattern of shifts seen
for guest 7 is similar for all the hosts (Table 3), with only
slight variations.
We thus conclude that host 1 has provided a useful
platform for evaluating substituent effects on catalysis.
All spectroscopic evidence suggests that the structural
perturbations of substitution are minimal. A large
amount of structural information places the substituent
very near the reacting center. As in any study of
substituent effects, one can never be certain that unan-
ticipated, and otherwise undetectable, structural changes
are occurring with substitution, and that these are
responsible for the rate variation. At present, however,
we feel all available evidence argues against this view.
F igu r e 5. (a) Best fit CD spectra of host (R,R,R,R)-4 and its
complex with guest 7 in aqueous buffer (pH 9). (b) Best fit CD
spectra of host (R,R,R,R)-4 and its complex with guest 9 in
aqueous borate buffer (pH 9). (c) Best fit CD spectra of host
For purposes of comparison, the series of hosts 3-6,
along with the parent 1, is preferable. In host 2 the
substantially different steric environment makes com-
parison more difficult. We presume that the much larger
rate enhancement seen with tetrabromo host 6 compared
to isomeric tetrabromo host 2 is due to a more favorable
positioning of the bromine atoms in 6, and so the series
(R,R,R,R)-4 and its complex with guest 10 in aqueous borate
buffer (pH 9).
Figure 1). The changes in the CD spectrum of host 3
upon binding each of the three guests are representative
of the whole series of hosts. In addition, there is induced
CD with the binding of 7 which occurs at 254 nm (Figure
3
-6 is preferable for comparative studies.
5
). In each host, the induced CD is of the same sign,
supporting similar binding conformations of the hosts in
the presence of 7.
The observation of induced CD in a guest allowed the
calculation of binding orientations using the coupled
oscillator approach to calculating induced CD developed
previously.10 This calculated induced CD was then
The data of Table 2 lead immediately to one important
conclusion. Clearly, there is no correlation between
q
binding energies (-∆G ) and catalysis (∆∆G ). That is,
s
the substituent effects seen in catalysis are not just
the result of making the host a generically better binding
site. In fact, no typical substituent parameter (electro-

Polarizability in Biomimetic Catalysis
J . Org. Chem., Vol. 61, No. 13, 1996 4359
negativity, hydrophobicity, steric size, π donor/acceptor
ability ...) correlates with the kcat/kun data.
In all macrocyclizations, enantiomerically pure ethenoan-
thracenes were coupled, although both R,R and S,S forms were
used. Workup of the macrocyclic products differed slightly
from that previously reported for host 1. After the macro-
cyclizations were complete, the reactions were filtered and the
In contrast, the correlation of kcat/kun with the polar-
izability data of Table 1 is very good. Adding substitu-
ents near a reacting center leads to a consistent rate
effect that correlates with the extent to which the
substituent is polarizable. As such, we consider the
current findings to support the model that increasing the
polarizability of the microenvironment in which a reac-
tion is occurring leads to a rate increase. That is, London
dispersion forces, working in concert with electrostatics
3
DMF or CH CN removed. The residues were then flash
chromatographed over silica eluting first with methylene
chloride and then ether in order to separate the macrocyclic
compounds from baseline impurities. The macrocycles were
then isolated from oligomers using preparative centrifugal
thin-layer chromatography (silica plates, 0-5% ether in
CH
2 2
Cl ).
1
Host 3, Tetr a m eth yl Ester . Yield 6%; H NMR (CDCl
3
)
(the cation-π interaction), hydrophobic interactions, and
δ 7.08 (d, J ) 8, 4H), 6.98 (s, 4H), 6.90 (d, J ) 2, 4H), 6.38 (dd,
other forces can be an important contributor to catalysis.
We have previously noted the similarity of the dealky-
lation reaction studied here to the important class of
J ) 2, 8, 4H), 5.23 (s, 4H), 4.98 (s, 8H), 3.78 (s, 12H), 2.14 (s,
+
+
12H); FAB-MS m/ e 964 (M
calcd for C60 12Li: 971.3619.
3
Host 5, Tetr a m eth yl Ester . Yield 8%; H NMR (CDCl )
); HRMS of M - Li 971.3621,
52
H O
1
biological methylation reactions involving the sulfonium
compound S-adenosylmethionine (SAM).11 We also specu-
δ 7.40 (s, 4H), 7.13 (d, J ) 8, 4H), 6.93 (d, J ) 2, 4H), 6.43 (dd,
J ) 2, 8, 4H), 5.25 (s, 4H), 5.04 (AB, J ) 14, ∆υ ) 52 Hz, 8H),
lated that perhaps cation-π interactions would be promi-
nent in the active sites of enzymes that mediate such
transformations. Recently, this speculation has received
some support from an X-ray structure of a DNA methy-
lase enzyme.12 The sulfonium of SAM is positioned in
van der Waals contact with a tryptophan from the
+
3
1
.78 (s, 12H); FAB-MS m/ e M - Li cluster 1050-1060 (1053
+
00 integral % within cluster); HRMS of M - Li 1051.1422,
35
calcd for C56
40 4
H O12 Cl Li: 1051.1434.
1
3
Host 6, Tetr a m eth yl Ester . Yield 10%; H NMR (CDCl )
δ 7.56 (s, 4H), 7.13 (d, J ) 8, 4H), 6.94 (d, J ) 2, 4H), 6.44 (dd,
J ) 2, 8, 4H), 5.26 (s, 4H), 4.99 (AB, J ) 14, ∆υ ) 39 Hz, 8H),
2
3.78 (s, 12H); FAB-MS m/ e M cluster 1220-1230 (1224 100
+
enzyme, in an ideal arrangement for cation-π interac-
+
integral % within cluster); HRMS of M - Li 1226.9399, calcd
tions.
7
9
for C56
40 4
H O12 Br Li: 1226.9413.
In a more general context, it will always be true that
any organic structure will be more polarizable than
water. Thus, it seems safe to say that any enzyme active
site will be more polarizable than water and so better
able to employ London dispersion forces for catalysis. Of
course, there is considerable variability in the polariz-
abilities of amino acid side chains, and it will be interest-
ing to consider which reaction types are best suited to
strong influence by London dispersion interactions.
Ester Hyd r olysis. All tetraacid macrocycles were prepared
from the corresponding tetramethyl esters using the following
hydrolytic procedure. The tetraesters were dissolved in 1-2
mL of DMSO. Thirty equivalents of aqueous CsOH (1.0 M
solution) were added, which caused a white precipitate to form.
Water (1-2 mL) was then added, and the solution was allowed
to stir for 24 h. The solution was then frozen and lyophilized.
The resulting residue was dissolved in a minimum amount of
water and loaded onto a cation exchange column (neutral pH,
+
Dowex 50 × 4, NH
4
form). The material was eluted with
water that had been passed through a Milli-Q purification
system. The fractions containing host were identified by their
UV activity on TLC silica gel plates. The appropriate fractions
were then combined and lyophilized to give the acid com-
pounds. Standard aqueous solutions of these host compounds
were prepared by dissolving them in borate-d buffer (50-70%
yields).
Con clu sion s
The current studies establish that precisely positioning
substituents within putative van der Waals contact with
a transition state can enhance catalysis. Interestingly,
the effectiveness of a given substituent correlates well
with the polarizability of the substituent, suggesting a
special role for London dispersion interactions in transi-
tion state stabilization. This is a potentially general
effect that should have significant implications for a
number of biological catalysts.
1
Host 3, Tetr a a cid .
3
H NMR (10% CD CN/90% borate,
referenced to internal DMG δ 1.09) δ 7.21 (d, J ) 8, 4H), 7.17
(s, 4H), 7.07 (bs, 4H), 6.63 (d, J ) 8, 4H), 5.19 (s, 4H), 5.05
(
AB, J ) 13, ∆υ ) 37 Hz, 8H), 2.20 (s, 12H).
1
Host 5, Tetr a a cid . H NMR (10% CD
3
CN/90% borate,
referenced to internal DMG δ 1.09) δ 7.57 (s, 4H), 7.24 (d, J )
8
, 4H), 7.09 (bs, 4H), 6.65 (d, J ) 8, 4H), 5.22 (s, 4H), 5.13
Exp er im en ta l Section
(AB, J ) 14, ∆υ ) 48 Hz, 8H).
Host 6, Tetr a a cid . 1H NMR (10% CD
CN/90% borate,
Gen er a l Meth od s. Instrumental and analytical methods
3
1
,10,13
were as in previous work from these laboratories.
referenced to internal DMG δ 1.09) δ 7.71 (s, 4H), 7.25 (d, J )
8, 4H), 7.10 (bs, 4H), 6.64 (d, J ) 8, 4H), 5.22 (s, 4H), 5.09
(AB, J ) 14, ∆υ ) 52 Hz, 8H).
Hosts 1, 2, and 4 were synthesized according to literature
1
procedures. Compound 8 was commercially available. Guests
9
and 10 were prepared by exhaustive alkylation of the
Kin etics of Su lfon iu m Sa lt Dea lk yla tion . Stock solu-
tions of hosts 1-6 were prepared in borate-d buffer. Stock
solutions of KI, KSCN, KHP (internal integration standard,
potassium hydrogen phthalate), 7, and 8 for the HPLC studies
were made by weighing each solid and dissolving it in the
nondeuterated borate buffer. The reaction rates were moni-
tored by integration of substrate and internal standard peak
areas from an HPLC trace using a Waters Baseline 810
software package. Each kinetic run was performed at least
twice. A representative reaction mixture consisted of 316 µM
host 1, 338 µM guest 7, 0.01160 M KI, and 513 µM KHP for a
total volume of 500 µL in borate buffer. The ratio of host to
guest was varied with the binding constant to keep the
percentage of guest bound in the range of 60 to 80%.
appropriate quinoline and amine with iodomethane. Guest 7
was prepared by refluxing trimethyloxonium tetrafluoroborate
with the corresponding sulfide.
Ma cr ocycliza tion s. The tetramethyl ester of hosts 3, 5,
and 6 were prepared by a condensation of the appropriate
ethenoanthracene with the corresponding bis(halomethyl)-
benzene in a suspension of cesium carbonate in anhydrous
CH
CN or DMF following the procedure developed for host 1.1
3
(11) Cantoni, G. L. Annu. Rev. Biochem. 1975, 49, 435; Maw, G. A.
The Chemistry of the Sulphonium Group; Wiley: New York, 1981;
Chapter 17. Lederer, E. Q. Rev. Chem. Soc. 1969, 23, 453. Walsh, C.
Enzyme Reaction Mechanisms; Freeman: New York, 1979; pp 851-
8
63.
(
For each experiment, the reaction solution is prepared in
an Eppendorf tube without the nucleophile. The solution of
nucleophile is then added, and the tube is shaken vigourously
just prior to the first injection of sample for the first time point.
12) Cheng, X.; Kumar, S.; Posfai, J .; Pflugrath, J . W.; Roberts, R.
J . Cell 1993, 74, 299-307.
13) Barrans, R. E., J r.; Dougherty, D. A. Supramol. Chem. 1994,
, 121-130.
(
4

4
360 J . Org. Chem., Vol. 61, No. 13, 1996
Ngola and Dougherty
The tube is then placed in an oil bath maintained at 46 °C by
a Thermo Watch. At each time point, a 20 µL aliquot is
removed from the sample which is then promptly returned to
the oil bath. This 20 µL aliquot is neutralized with 20 µL of
phosphate buffer (pH ) 7) and the 40 µL sample injected onto
the column. A gradient elution was used to separate the
calibration curve consisting of measuring relative peak areas
of five samples of various sulfonium salt concentrations and a
fixed KHP concentration was used to convert peak areas to
concentrations.
Ack n ow led gm en t . We thank the Office of Naval
Research for support of this work. We also thank Dr.
Alison McCurdy for assistance with initial experiments
and many helpful discussions, and Dr. J onathan For-
man for assistance with the CD measurements and
calculations.
reaction solution components. Solvent A was 0.1% TFA/H
2
O
(
v/v) and solvent B was 0.1% TFA/CH CN (v/v). Elution was
3
performed at 1.8 mL/min with 80% A from 0 to 8 min, a linear
gradient to 100% B from 8 to 13 min, maintained at 100% B
from 13 to 19 min, brought back linearly to 80% A from 19 to
2
5 min, and allowed to equilibrate at 80% A from 25 to 42
min. Eluting compounds were detected at 254 and 230 nm. A
J O960521Y