Spectroscopic determination of pressure-induced shifts in inclusion complexation equilibria

Article abstract of DOI:10.1021/ac961271o

The effect of modest hydrostatic pressure (<350 bar) on condensed-phase equilibrium processes has been largely overlooked, due in large part to the small compressibility of these phases relative to gases or supercritical fluids. Although the bulk properties of condensed phases are not significantly modified by pressure in this modest regime, the solvation processes driving inclusion complexation may be appreciably affected. In this paper, we examine this hypothesis using steady-state fluorescence spectroscopy to determine the pressure dependence of association constants. The widely used host molecule, β-cyclodextrin, provides an incompressible hydrophobic cavity into which structurally analogous fluorescent probes are encapsulated. By comparing the unique pressure dependencies of these equilibria, the importance of local site solvation and rim interactions in influencing the pressure dependence is demonstrated. The structurally analogous complexes chosen for these studies are expected to have similar pressure-dependent behavior based on comparable solvation structures. However, pressure-induced changes in the association constant for these two analogs are quite distinct, with differences in K_c ranging from clearly pressure dependent (-14%) to pressure independent over 338 bar. Additional solvation perturbations are observed in the pressure dependence of the quantum efficiency for both complexes (-7.3% and -9.4%). Thus, pressure-induced perturbation in the fluorescence properties of the complex need not be accompanied by simultaneous changes in the complexation equilibrium. Finally, these pressure-induced changes in complexation selectivity are important for all measurements conducted under variable pressure conditions, including liquid chromatography and process monitoring.

Full text of DOI:10.1021/ac961271o

Anal. Chem. 1997, 69, 2136-2142
Spectroscopic Determination of Pressure-Induced
Shifts in Inclusion Complexation Equilibria
Shirley M. Hoenigman and Christine E. Evans*
Department of Chemistry, University of Michigan, Ann Arbor, Michigan 48109-1055
has a hydrophilic exterior and cavity rim due to the sugars’
hydroxyls, and these allow for increased solubility of organics in
water as well as chiral sites for additional host-guest interactions.
Moreover, the rigid cavity of these oligosaccharides makes them
ideal models of the hard binding sites in protein-substrate
interactions.1 Excellent reviews addressing both the fundamental
processes driving the complexation of these host molecules and
their many analytical applications have appeared in the literature.2,3
The pressure-dependent behavior of these important inclusion
complexes has, however, been largely overlooked. The role of
pressure in discerning and controlling reaction pathways is well
known and has been clearly demonstrated in the past several
decades.4,5 Elegant experiments in both the solid6,7 and solution
states8,9 have shown the power of pressure as a well-defined
perturbation parameter. However, studies assessing the influence
of pressure on inclusion complexation have been very limited.1,5,10
It may be that the low compressibility of polar liquid phases has
led to the common misconception that no significant perturbations
in solute interaction are expected to occur at pressures below 500
bar. While it is true that the bulk properties are not affected
appreciably in this pressure regime,11 the impact on the local site
solvation environment can be significant.12,13 Indeed, pressure-
induced solvation perturbations may become even more pro-
nounced for the mixed solvation and confined geometry environ-
ments of inclusion complexes, where local interactions play a key
role in the complexation process.
The effect of modest hydrostatic pressure (<3 5 0 bar) on
condensed-phase equilibrium processes has been largely
overlooked, due in large part to the small compressibility
of these phases relative to gases or supercritical fluids.
Although the bulk properties of condensed phases are not
significantly modified by pressure in this modest regime,
the solvation processes driving inclusion complexation
may be appreciably affected. In this paper, we examine
this hypothesis using steady-state fluorescence spectros-
copy to determine the pressure dependence of association
constants. The widely used host molecule, â-cyclodextrin,
provides an incompressible hydrophobic cavity into which
structurally analogous fluorescent probes are encapsu-
lated. By comparing the unique pressure dependencies
of these equilibria, the importance of local site solvation
and rim interactions in influencing the pressure depen-
dence is demonstrated. The structurally analogous com-
plexes chosen for these studies are expected to have
similar pressure-dependent behavior based on compa-
rable solvation structures. However, pressure-induced
changes in the association constant for these two analogs
are quite distinct, with differences in Kc ranging from
clearly pressure dependent (-1 4 %) to pressure indepen-
dent over 3 3 8 bar. Additional solvation perturbations are
observed in the pressure dependence of the quantum
efficiency for both complexes (-7 .3 % and -9 .4 %). Thus,
pressure-induced perturbation in the fluorescence prop-
erties of the complex need not be accompanied by
simultaneous changes in the complexation equilibrium.
Finally, these pressure-induced changes in complexation
selectivity are important for all measurements conducted
under variable pressure conditions, including liquid chro-
matography and process monitoring.
In contrast with temperature, where both kinetic energy and
volume of the system are affected, pressure acts only to reduce
the system volume. In the case of chemical reactions and
interactions, the equilibrium position of the reaction is shifted with
pressure to minimize the total volume occupied by all solvated
species. The magnitude of this equilibrium position shift is given
by the difference in the partial molar volumes of the products
(1) Torgerson, P. M.; Drickamer, H. G.; Weber, G. Biochemistry 1 9 7 9 , 18,
3079-3083.
(2) Szejtli, J. Cyclodextrins and Their Inclusion Compounds; Akade´miai Kiado´:
Complexation and reversible binding interactions play an
essential role in a wide range of important biological and
technological systems. Among these processes, inclusion com-
plexation involving so-called host-guest interactions are utilized
in applications ranging from analytical separations to drug delivery.
These host-guest complexes are driven by weaker forces such
as hydrogen-bonding and van der Waals interactions, and depend-
ing on the species involved, they can exhibit a high degree of
selective interaction. One of the most commonly utilized host
molecules is â-cyclodextrin (âCD), which is a toroidally shaped
Budapest, 1982.
(3) Armstrong, D. W.; DeMond, W. J. Chromatogr. Sci. 1 9 8 4 , 22, 411-415.
(4) Tietze, L. F.; Hubsch, T.; Voss, E.; Buback, M.; Tost, W. J. Am. Chem. Soc.
1 9 8 8 , 110, 4065-4066.
(5) Taniguchi, Y.; Makimoto, S.; Suzuki, K. J. Phys. Chem. 1 9 8 1 , 85, 3469-
3472.
(6) Drickamer, H. G.; Bray, K. L. Int. Rev. Phys. Chem. 1 9 8 9 , 8, 41-64.
(7) Gu, Y.; Xu, Z. J. Phys.: Condens. Matter 1 9 9 1 , 3, 6553-6558.
(8) Zakin, M. R.; Grubb, S. G.; King, H. E.; Herschbach, D. R. Physica 1 9 8 6 ,
139, 530-532.
(9) Nordmeier, E. J. Phys. Chem. 1 9 9 2 , 96, 1494-1501.
(10) Sueishi, Y.; Nishimura, N.; Hirata, K.; Kuwata, K. J. Phys. Chem. 1 9 9 1 , 95,
5359-5361.
molecule composed of seven
D-glucopyranose monomers. The
(11) Hamann, S. D. Physico-Chemical Effects of Pressure; Butterworths Scientific
cyclodextrin cavity is relatively hydrophobic and is capable of
accommodating naphthyl moieties with a good fit, provided other
substituents do not preclude encapsulation. The âCD molecule
Publications: London, 1957.
(12) Weber, G.; Drickamer, H. G. Q. Rev. Biophys. 1 9 8 3 , 16, 89-112.
(13) Rogers, K. K.; Sligar, S. G. J. Mol. Biol. 1 9 9 1 , 221, 1453-1460.
2136 Analytical Chemistry, Vol. 69, No. 11, June 1, 1997
S0003-2700(96)01271-1 CCC: $14.00 © 1997 American Chemical Society

Figure 1. Model solutes 2-(N-methylanilino)naphthalene-6-sulfonic
acid (2,6-MANS) and 2-anilinonaphthalene-6-sulfonic acid (2,6-ANS).
and the reactants, ∆V (vide infra). For reactions where ∆V is
positive, pressure-induced dissociation is expected, whereas a
negative ∆V will result in a pressure-induced increase in associa-
tion. Thus, based on the pressure dependence of the complex-
ation constant, fundamental information about solvation properties
may be ascertained. In addition, pressure-induced perturbations
of solute complexation are important from a pragmatic perspective.
This is especially true for analytical applications such as liquid
chromatography, where pressure is inherent to the separation
process.14
In this study, we utilize steady-state fluorescence to investigate
the affect of pressure on inclusion complexes formed with
â-cyclodextrin (âCD). In addition to its many applications, âCD
also provides a uniquely incompressible environment, so that no
structural changes in the âCD cavity are expected in the moderate
pressure regime used here.1,10 Two structurally analogous anili-
nonaphthalenesulfonate dyes commonly utilized as fluorescent
probes for biological studies15,16 have been chosen as model
solutes. These dyes have the distinct advantage of intense
fluorescence when bound within the cyclodextrin cavity. Encap-
sulation of the naphthyl moiety within the âCD cavity results in
a greatly enhanced quantum efficiency and an emission maximum
shift of more than 20 nm. This combination of properties makes
these host-guest pairs well suited for the direct measurement of
complexation constants by steady-state fluorescence under el-
evated pressure conditions.
Figure 2. Schematic of high-pressure fluorescence spectroscopy
apparatus.
host-guest complexation and to eliminate any self-association.
After thorough mixing, the probe-cyclodextrin solutions were
allowed to equilibrate for 60-90 min prior to fluorescence
measurements.
High-pressure fluorescence measurements were conducted
using the apparatus depicted in Figure 2. Sample pressurization
was accomplished in a 110-cm length of fused-silica capillary
(Polymicro Technologies; 320 µm i.d. × 440 µm o.d.) connected
between a pressurization valve (High Pressure Equipment; Model
11-11AF1) and an injection valve (Valco; Model C6WY). This
configuration creates a system that can be operated in both flowing
and static modes. The sample was introduced into the capillary
using a syringe pump (Applied Biosystems; Model 140A) together
with a 1-mL injection loop. After the capillary was completely filled
with sample, the pressurization valve was closed, and the system
slowly attained the desired pressure. Using ultrapure water as
the pressure transmitting fluid, pressures up to 350 bar were
readily achieved. Measurements at multiple pressures for a single
injection of equilibrated sample solution were identical to those
performed using an individual injection at each pressure. As
expected, the sequence of pressure values during each experiment
had no affect on the resultant measurement.
In this configuration, the determination of steady-state fluo-
rescence in the high-pressure regime was facilitated by the optical
transparency of the fused-silica capillary. A window was made in
the capillary 60 cm from the injection valve by removing the
polyimide coating. This window region was then mounted in a
light-tight enclosure and fitted with Teflon sleeves to ensure
capillary stability during pressurization. Fluorescence excitation
was accomplished using a mercury/ xenon arc lamp (Xenon Corp.)
and a 10-nm band-pass filter centered at 313 nm. Excitation
radiation was transmitted to the capillary window using a 500-µm
optical fiber. Emission radiation was collected in a 90° coplanar
geometry with another 500-µm optical fiber, filtered, and focused
onto a PMT (Hamamatsu; Model R1463). For 2,6-ANS measure-
ments, a 10-nm band-pass filter centered at 450 nm was used, while
a 10-nm band-pass filter at 470 nm was used for 2,6-MANS
measurements. The resulting photocurrent was then digitized
(National Instruments; Model AT-MIO-16-L-9), and the signal was
EXPERIMENTAL SECTION
Materials and Methods. The fluorescent solutes utilized in
this study, 2-anilinonaphthalene-6-sulfonic acid (2,6-ANS) and 2-(N-
methylanilino)naphthalene-6-sulfonic acid sodium salt (2,6-MANS),
were obtained from Molecular Probes and are illustrated in Figure
1. Underivatized â-cyclodextrin (Janssen; >99%) served as the
host molecule in all cases, and solutions were freshly prepared
in ultrapure deionized water (Millipore; >18 MΩ). All materials
were used without further purification. The fluorophore concen-
tration was maintained constant at 1.00 × 10-5 M, while the âCD
concentration range was chosen in accordance with Benesi-
Hildebrand constraints (vide infra). Both the solute and host
concentrations were systematically controlled to ensure only 1:1
(14) Ringo, M. C.; Evans, C. E. Anal. Chem. 1 9 9 7 , 69, 643-649.
(15) Ruiz Silva, B. E.; Burtnick, L. D.; Turro, N. J. Biochem. Int. 1 9 9 1 , 23, 905-
913.
(16) Mock, D.; Langford, G.; Dubois, D.; Criscimagna, N.; Horowitz, P. Anal.
Biochem. 1 9 8 5 , 151, 178-181.
Analytical Chemistry, Vol. 69, No. 11, June 1, 1997 2137

averaged over 50 s (rate ) 20 Hz) using the LabView software
program (National Instruments; Version 3.1.1). Exposure times
of up to 10 min showed no measurable photodegradation.
Data Analysis. Complex formation constants were deter-
mined using the method of Benesi-Hildebrand, the derivation
of which has been described in detail elsewhere.17,18 In brief, this
method uses simple mass balance arguments to determine the
molar complexation constant (Kc) based on signal measurements
under systematically varying cyclodextrin concentration condi-
tions. For conditions where the analytical concentration of
cyclodextrin (CâCD) is very much greater than the equilibrium
concentration of the complex ([probe‚âCD]), the Benesi-Hilde-
brand relationship is given as
Experimental Design. In discerning the role of pressure on
inclusion complexation, careful experimental design is required
in order to isolate pressure from all other parameters. With this
goal in mind, a capillary fluorescence cell is utilized to maintain
constant temperature conditions by allowing rapid heat dissipation
and simultaneously eliminating both inner and outer filter effects.
The band-pass filters (fwhm ) 10 nm) are incorporated to ensure
that the analytical signal will be unaffected by small frequency
shifts in the excitation or emission maxima induced by elevated
pressures (vide infra), while concurrently minimizing the contri-
bution of unbound fluorophore intensity to the overall signal.
Judicious choice of guest and host species also plays a
significant role in the experimental design. Anilinonaphthalene-
sulfonate dyes (Figure 1) have been chosen as model solutes on
the basis of their unique fluorescence properties. In the unbound
state, these fluorophores are weakly fluorescent in aqueous
solution, with quantum yields of 0.090 for 2,6-ANS and 0.006 for
2,6-MANS.19 Upon encapsulation within the hydrophobic âCD
cavity, the excitation maxima are unchanged, the emission maxima
are blue-shifted by 20 nm for 2,6-ANS and by 26 nm for 2,6-MANS,
and the quantum yield is dramatically increased. The fluorescence
intensity contributions of the unbound fluorophore are, thus,
minimized or eliminated, and the measured analytical signal
originates predominantly from the complexed species. In contrast
with some ANS analogs,20 the 2,6-ANS‚âCD and 2,6-MANS‚âCD
complexes show no dependence of the emission maxima on the
âCD concentration. Under these conditions, association constants
for complexation may be determined directly using the Benesi-
Hildebrand approach (eq 1).
Cprobe
-1
) (KckiQcomplexCâCD
)
-1 + (kiQcomplex
)
(1)
Fcomplex
When the fluorescence intensity of the complex (Fcomplex) is
measured as a function of the analytical cyclodextrin concentra-
tion, the complexation constant is determined from the intercept-
slope ratio of a double-reciprocal plot. Furthermore, the product
of the instrumental constant (ki) and the quantum efficiency of
the complex (Qcomplex) can be assessed directly from the intercept.
A modest and constant analytical concentration of the fluorescent
probe (Cprobe) is utilized throughout these studies to maintain 1:1
complexation conditions.
It is important to note that small contributions to the measured
fluorescence intensity from the unbound fluorophore may intro-
duce significant errors in the resulting complexation constant.18
Even in cases where the contribution of unbound fluorophore to
the measured signal is less than 5%, a significant error is
introduced because the equilibrium concentration of free fluoro-
phore varies systematically with the analytical concentration of
host species. That is, the concentration of unbound fluorophore
is at a minimum at high âCD concentrations and at a maximum
at the lowest âCD concentration. Thus, a systematic error is
introduced in the slope of Benesi-Hildebrand plot, leading to an
overestimation of the complexation constant. Using calibration
curves, this error may be corrected by subtracting the fluores-
cence intensity attributable to the unbound probe from the overall
measured intensity of the complex. In this way, all measured
fluorescence intensities in the 2,6-ANS study were corrected for
the intensity contribution of the unbound fluorophore. No
contribution from the free fluorophore was measured in the 2,6-
MANS studies, due to the large binding constant and the low
quantum yield of the free fluorophore.
After iterative correction for the unbound fluorophore intensity,
the Benesi-Hildebrand relationship was utilized to experimentally
determine complexation constants from the measured fluores-
cence intensity as a function of the âCD concentration (eq 1).
This method of data analysis yields not only the association
constant but also the product of the instrumental constant and
the quantum efficiency of the complex. In these studies, the effect
of pressure on both the complexation constant and the complex
quantum efficiency are evaluated at six pressures between 7 and
345 bar. From this information, the effects of uniform hydrostatic
pressure on the equilibrium position of the inclusion reaction as
well as on the fluorescence properties of the complex are
determined.
In addition to pragmatic measurement considerations, these
two ANS analogs have been chosen for their structural and
chemical similarities. Both solutes contain no ionizable groups
in the pH range from 2 to 11, eliminating pressure-induced
ionization effects. Moreover, no solution buffering is required,
which ensures that neither competing pressure-dependent equi-
libria nor quenching by buffer constituents will be a factor.
Identical positional substitution on the naphthyl moiety of these
probes allows both molecules to fit similarly inside the â-cyclo-
dextrin cavity. Finally, the complex structure is such that the
anilino nitrogen on the probe molecules is in proximity to the
hydroxyls on the cyclodextrin rim.20,21 Substitution at the anilino
nitrogen, therefore, is expected to yield insight into the effect of
pressure on the site-specific rim interactions of these inclusion
complexes.
The host molecule, â-cyclodextrin, has been chosen for the
well-defined cavity geometry that remains fixed over the pressure
regime from 1 to 350 bar.1,10 In addition to its many analytical
applications, this incompressible host is also commonly utilized
to emulate hard-site binding in proteins. Consistent with the
chosen probe molecules, the âCD host is not ionizable over pH
2-12. Furthermore, because âCD is nonfluorescent, it does not
interfere with the analytical signal arising from the probe-
cyclodextrin complex. Thus, this combination of analogous probe
(19) Seliskar, C. J.; Brand, L. J. Am. Chem. Soc. 1 9 7 1 , 93, 5414-5420.
(20) Nishijo, J.; Nagai, M.; Ysauda, M.; Ohno, E.; Ushiroda Y. J. Pharm. Sci.
1 9 9 5 , 84, 1420-1426.
(17) Benesi, H. A.; Hildebrand, J. H. J. Am. Chem. Soc. 1 9 4 9 , 71, 2703-2707.
(18) Hoenigman, S. M.; Evans, C. E. Anal. Chem. 1 9 9 6 , 68, 3274-3276.
(21) Crescenzi, V.; Gamini, A.; Palleschi, A.; Rizzo, R. Gazz. Chim. Ital. 1 9 8 6 ,
116, 435-440.
2138 Analytical Chemistry, Vol. 69, No. 11, June 1, 1997

molecules and âCD provides a well-defined model system for the
examination of pressure-induced perturbations on inclusion com-
plexation.
Table 1. Association Constantsa (M-1) of ANS‚âCD
Complexes at Various Pressures
pressure (bar)
2,6-ANS‚âCDb
2,6-MANS‚âCDc
6.89
68.9
138
207
2160 ( 60
2130 ( 60
2070 ( 50
1970 ( 60
1910 ( 30
1830 ( 40
10.5 ( 2.0
16 500 ( 300
16 700 ( 400
16 100 ( 500
16 100 ( 300
16 000 ( 300
15 500 ( 300
1.2 ( 2.4
RESULTS AND DISCUSSION
P ressure Dependence of Association Constants. The
pressure dependence of equilibrium constants is described by the
fundamental thermodynamic parameter, ∆V, and is based on
volumetric considerations. For the complexation reaction,
276
345
∆V (cm3/ mol)d
probe + âCD S probe‚âCD
(2)
a
Association constants are reported for a single representative trial.
The uncertainty associated with each binding constant was calculated
using the variances determined from a least-squares fit of the Benesi-
Hildebrand plot for each pressure. b Experiment conducted at 292.5
K. c Experiment conducted at 292.3 K. d The overall ∆V was determined
from three separate measurements of ∆V for each complex.
a shift in the equilibrium position is expected if a difference exists
between the partial molar volumes of products and reactants.11
In solution, the partial molar volume depends upon the size of
the molecules as well as the intermolecular forces that comprise
the solvation environment. It is these interactions that define the
solvation volume of the species involved in the reaction and thus
the magnitude and directionality of ∆V.
complex dissociation is favored with pressure, and this behavior
is independent of the absolute value of the association constant.
In other studies,5,10 ∆V is approximated using the sum of the
individual volume changes occurring during solute complexation,
including desolvation of the probe molecule, desolvation of the
âCD cavity, solute inclusion inside the cavity, and any conforma-
tional changes in the âCD upon complexation. In summing these
contributions, the magnitude of the increase in ∆V predicted for
the desolvation processes is expected to dominate the negative
∆V accompanying the inclusion process. Since âCD is structurally
rigid, no conformational changes upon complexation are expected
to contribute to the overall ∆V. Thus, both of these approaches
predict a positive change in molar volume upon complexation and
the concomitant decrease in the complexation constant with
pressure.
The few experimental studies in the literature appear to be in
general agreement with these predictions. High-pressure spec-
troscopic studies of 2-naphthyl acetate inclusion into âCD yielded
a ∆V of +10 ( 2 cm3/ mol.5 Studies with poly-â-cyclodextrin (poly-
âCD) and the 1,8-substituted ANS probe resulted in a measured
∆V of +9.3 cm3/ mol.1 Based on the comparability in probe
structure, the latter value may be used to predict the expected
change in the current study. For the pressure range of 1-345
bar and a temperature of 292.5 K, this molar volume change
corresponds to a 14% decrease in association constant.
For the reactions under investigation here, the molar equilib-
rium constant, Kc, is defined as
[probe‚âCD]
[probe][âCD]
Kc )
(3)
where [probe‚âCD], [probe], and [âCD] refer to the equilibrium
molar concentrations of the complex, fluorophore, and â-cyclo-
dextrin, respectively. To describe the expected pressure depen-
dence of Kc, it is useful to proceed through Gibbs free energy,
∆G,
∆G ) -RT ln Kc
(4)
where R is the gas constant and T represents the absolute
temperature. The pressure dependence of ∆G is based on the
fundamental thermodynamic equation of state and is defined as
∂∆G
∂P
) ∆V
T
(5)
(
)
Under the modest pressure range of interest, the partial differential
may be accurately expressed as a simple difference,11 and
substitution of eq 4 into eq 5 yields
∆ ln K
-RT
c ) ∆V - ∆nRTκS
(6)
∆P
where the term ∆nRTκS accounts for any increase in molar
concentration with pressure and relates directly to the bulk solvent
compressibility, κS. The difference in the stoichiometric coef-
ficients of products over reactants is ∆n, and the overall term is
approximately 1.1 cm3/ mol in these studies.
Experimentally measured Kc values under elevated pressure
conditions are shown in Table 1, with representative double-
reciprocal plots illustrated in Figure 3. The association constant
determined for 2,6-ANS‚âCD at 7 bar is 2160 M-1 and is in general
agreement with earlier studies conducted at atmospheric pres-
sure.18,22 Literature values for the 2,6-MANS‚âCD complex forma-
tion constant at ambient pressure vary from 20 000 (ref 23) to
7360 M-1 (ref 22), while the value in this study was found to be
16 500 ( 300 M-1. Although the 2,6-ANS and 2,6-MANS fluoro-
phores are almost structurally identical, the association constants
at standard temperature and pressure are dramatically different.
The fluorophores differ only by the presence of a methyl group
on the anilino nitrogen atom, and this portion of the probe
molecules is situated near the wider rim of the â-cyclodextrin
cavity.20,21 A larger association constant for the 2,6-MANS‚âCD
complex thus suggests that specific rim interactions play a key
For inclusion complexation reactions involving âCD, the ∆V
is expected to be positive, because the host molecule is rigid and
incompressible over 10 kbar.1,10 In contrast with many associated
complexes, the âCD host molecule cannot change conformation
to achieve minimum energy solvation of the guest, which would
serve to decrease the volume of the complex. Of the limited
studies investigating the effect of pressure on inclusion complex-
ation equilibria, several hypotheses regarding the observed ∆V
have been proposed. Torgerson et al.1 suggest that, in subjecting
an incompressible host such as â-cyclodextrin to elevated pres-
sure, the observed ∆V may be approximated as the difference in
the compressibilities of the guest molecule and the volume of
water occupying the unassociated cyclodextrin cavity. Because
the aromatic guest molecules are more compressible than water,
(22) Catena, G. C.; Bright, F. V. Anal. Chem. 1 9 8 9 , 61, 905-909.
(23) Seliskar, C. J.; Brand, L. Science 1 9 7 1 , 171, 799-800.
Analytical Chemistry, Vol. 69, No. 11, June 1, 1997 2139

the pressure perturbation of the complexation equilibrium position
is not strongly dependent on complexation geometry. For the
positional isomers, 2,6-ANS is completely encapsulated inside the
cyclodextrin cavity, while 1,8-ANS exhibits only modest entry into
the cavity.21,24,25 In accordance with previous studies, the pressure-
dependent behavior would appear to be indifferent to the ther-
modynamic impetus that led to complex formation.1,5,10
In direct contrast with the behavior of the 2,6-ANS analog, the
association constant of the 2,6-MANS‚âCD complex exhibits no
significant pressure dependence (Table 1). As evidenced by a
∆V of +1.2 ( 2.4 cm3/ mol, the expected dissociation of this
structurally analogous complex with pressure is not realized. This
lack of pressure dependence does not appear to be in agreement
with predictions based on simple volumetric considerations. The
nearly identical structures of the 2,6-ANS‚âCD and 2,6-MANS‚âCD
complexes, together with the analogous structures of the unbound
probe molecules (Figure 1), lead to the expectation of comparable
∆V values based on eq 6. Thus, the assumption of complex molar
volume comparability based exclusively on analogous structural
considerations appears to be insufficient. That is, volume changes
associated with the solute desolvation are expected to be similar
for these analogs, and the desolvation of the âCD cavity is clearly
identical in both cases. Based on the previous model, the disparity
in measured ∆V between these analogous solutes may be
attributed to a difference in the volume upon inclusion. These
analogs differ only in anilino substituents that have been directly
implicated in hydrogen bonding with the rim of the âCD. As a
result, the apparent difference in the molar volume change upon
inclusion for these two molecules is consistent with a greater
influence of local site interactions on the molar volume than
previously appreciated. High-pressure NMR studies are presently
underway to evaluate this local site perturbation with pressure
for these analogous inclusion complexes. Furthermore, in the
above models, the volume change associated with the inclusion
process often considers the resultant complex as a single
structure. However, previous observations of a distribution of
emitting states for 2,6-ANS‚âCD and 1,8-ANS‚âCD1,26-28 may
indicate the presence of a range of complex structures. That is,
the complex may be comprised of a range of structures with
varying probe positions inside the cavity. Since these complexes
may no longer be discrete structures, the range of inclusion
complex geometries and, therefore, complex molar volumes may
differ for 2,6-ANS and 2,6-MANS. Unfortunately, evaluation of this
contribution to the change in molar volume upon inclusion is not
presently feasible, as no studies to our knowledge have assessed
the role of pressure on the distribution of complex structures.
Thus, local solvation and the detailed distribution of the probe
within the cyclodextrin cavity may play a larger role in dictating
the pressure dependent behavior for inclusion complexation than
previously appreciated.
Figure 3. Benesi-Hildebrand data for the 2,6-ANS‚âCD (top) and
2,6-MANS‚âCD (bottom) complexes.
role in differentiating the complexation constants of these struc-
turally similar analogs.
As the hydrostatic pressure is increased, both complexes are
well behaved, with double-reciprocal plots showing excellent
linearity at all pressures (Figure 3). The resultant association
constants as a function of pressure determined from these plots
are shown in Table 1. For the 2,6-ANS‚âCD complex, a 14%
decrease in the association constant is measured for an ac-
companying pressure increase of 338 bar. This result is in good
qualitative agreement with theoretical predictions for the pressure
dependence of inclusion complexes based on the complexation
of 1,8-ANS with an incompressible host like poly-â-cyclodextrin.
While the association constants for these two isomers are
significantly different (K1,8-ANS‚poly-âCD ) 50 M-1;1 K2,6-ANS‚âCD ) 2160
M-1), the pressure-dependent behavior appears to be nearly
identical over this pressure range. Because the volumetric
considerations are essentially identical for the ANS‚âCD com-
plexes, and because âCD cannot contract about the guest, the
solvated complex occupies a greater volume than the dissociated
species solvated by water. The measured ∆V for the 2,6-ANS‚âCD
complex is +10.5 ( 2.0 cm3/ mol, which is fully consistent with
previous studies of 1,8-ANS‚poly-âCD complexation,1 where the
∆V was measured to be +9.3 cm3/ mol for this analogous complex.
This general agreement with earlier investigations suggests that
Because the ∆V of a complexation reaction measures thermo-
dynamic properties, detailed mechanistic information is not gained
from these studies. The ∆V corresponds to perturbations in the
initial and final states of the system, which is important to
(24) Schneider, H.-J.; Blatter, T.; Simova, S. J. Am. Chem. Soc. 1 9 9 1 , 113, 1996-
2000.
(25) Nishijo, J.; Nagai, M. J. Pharm. Sci. 1 9 9 1 , 80, 58-62.
(26) Bright, F. V.; Catena, G. C.; Huang, J. J. Am. Chem. Soc. 1 9 9 0 , 112, 1343-
1346.
(27) Huang, J.; Bright, F. V. J. Phys. Chem. 1 9 9 0 , 94, 8457-8463.
(28) Nakamura, A.; Saitoh, K.; Toda, F. Chem. Phys. Lett. 1 9 9 1 , 187, 110-115.
2140 Analytical Chemistry, Vol. 69, No. 11, June 1, 1997

be an accurate measure of Qcomplexki, the contribution of the
unbound fluorophore must be eliminated (vide supra), and the
quantum efficiency must be distinguished from fluorescence
intensity decrease arising from the dissociation of the complex.
Clearly, the fluorescence intensity decreases as the complex
dissociates with pressure, thus increasing the number of unbound
fluorophore molecules. The total fluorescence intensity is given
by the sum of the intensities of the emitting species,
Table 2. Qcomplexkia over 338 bar for ANS‚âCD
Complexes
pressure (bar)
2,6-ANS‚âCD
2,6-MANS‚âCD
6.89
68.9
138
207
276
345
532 ( 10
521 ( 10
513 ( 9
508 ( 10
498 ( 6
493 ( 8
226 ( 1
221 ( 1
218 ( 1
213 ( 1
209 ( 1
205 ( 1
Ftotal ) Fcomplex + Fprobe
The fluorescence of each component is given by
Fcomplex ) Qcomplexki[complex]
and
(7)
(8)
(9)
% decrease
7.33
9.37 ( 0.25b
Qcomplexki values were divided by 103. b The overall percent decrease
is averaged from three trials.
a
Fprobe ) Qprobeki[probe]
understanding general solvation changes induced by pressure.
Since the fluorescent probes chosen for these studies are very
sensitive to their solvation environment, further insight can be
gained into the role of local solvation. The study undertaken here
provides a unique opportunity to investigate any pressure-induced
change in the fluorophore quantum efficiency.
Because a correction for any contribution of unbound fluorophore
to the total intensity has been applied,18 the intensity used to
determine the association constants and quantum yield is attribut-
able solely to that of the complex. Using the Benesi-Hildebrand
approach, the association constant is effectively isolated from the
complex quantum efficiency (eq 1). As a result, pressure-induced
effects in the quantum efficiency are determined independent of
any pressure perturbation in the complexation constant. Thus,
in these experiments, the overall decrease in fluorescence intensity
observed with pressure is the result of both the decrease in the
equilibrium concentration of the complex and the decrease in the
complex quantum efficiency.
Based on these observations, it is clear that a solvation
environment perturbation occurs for both complexes, which is
manifest in the decrease in quantum efficiency. The 2,6-ANS‚âCD
complex shows a substantial decrease in both the quantum
efficiency and the equilibrium constant, which suggests that a
significant change in solvation has occurred over 338 bar. For
the 2,6-MANS‚âCD complex, a similar decrease in quantum
efficiency is not accompanied by a shift in the equilibrium position
of the complexation reaction. Clearly, a pressure-induced per-
turbation in the fluorescence properties of the complex can be
independent of the ∆V, so that not all of the solvation information
is contained exclusively in the ∆V of a reaction. The steady-state
fluorescence measurements, thus, are useful to distinguish
between two different pressure-induced solvation perturbations,
although it is not clear if the mechanism by which the quantum
efficiency decreases is the same for both complexes, or if this
mechanism is directly correlated to the ∆V of the 2,6-ANS‚âCD
complex.
P ressure Dependence of Complex Quantum Efficiency.
Solvent-sensitive probe molecules have been widely used in
chemistry and biological studies. Fluorescence properties of
emission frequency and quantum efficiency are both commonly
utilized to assess perturbations in the solvation environment.29,30
For the probes of interest in these studies, significant blue shifts
are observed on complexation; however, the bound and unbound
2,6-ANS exhibit no pressure-induced change in the emission
frequency. This result is fully consistent with previous studies
involving the positional isomer 1,8-ANS.31,32 As a result, no
significant shifts in the fluorescence spectra of the free and
complexed fluorophores are anticipated in the modest pressure
regime. Moreover, the use of 10-nm band-pass filters for excita-
tion and emission ensures that the analytical signal will be
unaffected by any small frequency shifts induced by elevated
pressure.
The quantum efficiencies of the unbound and complexed
solutes, however, may exhibit a pressure dependence. Using the
present experimental configuration (Figure 2), the unbound
solutes show no statistically significant change in fluorescence
intensity over the pressure range of interest here. Pressure-
induced perturbations on the quantum efficiency of the complexed
solutes are assessed using the inverse of the intercept (kiQcomplex
)
for the Benesi-Hildebrand plot (eq 1). Measurements of the
intercept as a function of pressure for 2,6-ANS‚âCD and 2,6-
MANS‚âCD are shown in Table 2. Since the instrumental
constant is independent of pressure, any pressure-induced changes
in this parameter provide a direct measure of the pressure
dependence of the complex quantum efficiency. As shown in
Table 2, the 2,6-ANS‚âCD and 2,6-MANS‚âCD complexes exhibit
a decrease of 7.3% and 9.4% in the quantum efficiency over 338
bar, respectively.
Unfortunately, the specific mechanism by which the complex
quantum efficiency decreases with pressure cannot be deduced
from steady-state fluorescence measurements. Positing mecha-
nisms is further impeded by the lack of detailed characterization
of the excited states of the anilinonaphthalenesulfonic acids in
neat solvents,33,34 much less in a mixed solvation environment
where the naphthyl moiety is encapsulated in the âCD cavity.
Thus, although it is possible to detect a clear, pressure-induced
effect on the solvation environment, the exact origin of this
perturbation requires more detailed mechanistic studies.
Studies are presently underway in our laboratory to determine
the localized perturbations in solvation structure induced in this
To confirm the validity of these determinations, it is important
to isolate pressure-induced perturbations in the quantum efficiency
of the complex from other possible factors. For the intercept to
(29) Turner, D. C.; Brand, L. Biochemistry 1 9 6 8 , 10, 3381-3390.
(30) Stryer, L. Science 1 9 6 8 , 162, 526-533.
(31) Rollinson, A. M.; Drickamer, H. G. J. Chem. Phys. 1 9 8 0 , 73, 5981-5996.
(32) Bismuto, E.; Sirangelo, I.; Irace, G.; Gratton, E. Biochemistry 1 9 9 6 , 35,
1173-1178.
(33) Robinson, G. W.; Robbins, R. J.; Fleming, G. R.; Morris, J. M.; Knight, A. E.
W.; Morrison, R. J. S. J. Am. Chem. Soc. 1 9 7 8 , 100, 7145-7150.
(34) Kosower, E. M.; Dodiuk, H. J. Phys. Chem. 1 9 7 8 , 82, 2012-2015.
Analytical Chemistry, Vol. 69, No. 11, June 1, 1997 2141

modest pressure regime. Among analytical techniques, NMR is
well suited to provide the localized environment information
required to discern pressure-induced changes within the unbound
guest and host as well as the associated species. Previous studies
with high-pressure NMR have demonstrated its capabilities in
monitoring perturbations for a wide range of molecules and
environments.35-37 Controlled-pressure NMR applied to this
important class of inclusion complexes is expected to yield insight
into those site-specific changes in solvation that have a direct
impact on binding interactions. Moreover, mechanistic informa-
tion regarding the role of the hydrophilic rim interactions may
be more clearly delineated.
ecules. However, the location of the positional substitution in
these analogous probe molecules is consistent with pressure-
induced perturbation in site-specific interactions with the hydro-
philic rim of the cyclodextrin cavity. High-pressure NMR studies
are presently underway to investigate the role of local site solvation
on this pressure dependence of association.
In addition to fundamental considerations, this pressure-
dependent behavior has a pragmatic impact on many measure-
ments of analytical importance. As demonstrated here, pressure
can act nonuniformly on inclusion complexation, yielding pressure-
induced changes in interaction selectivity. Such pressure pertur-
bations in selectivity are expected to have a direct effect on liquid
chromatographic separations using inclusion complexation mech-
anisms.14 This important class of separations has pressure as an
inherent parameter utilized to drive flow through the chromato-
graphic column. In this case, solute capacity factor and selectivity
may be affected by pressure-induced changes in solute complex-
ation. Thus, it would seem prudent to begin to monitor and
control pressure more carefully under these separation conditions.
Although demonstrated here for positionally substituted solutes,
studies are underway in our laboratory to assess the effect of
pressure on chiral separations performed by inclusion complex-
ation.38
CONCLUSIONS
Commonly ignored in polar condensed phases, pressure is
demonstrated here to have a significant impact on the host-guest
interactions that form the basis for many analytical techniques.
For two structurally analogous host-guest pairs, significantly
differing pressure-dependent behaviors are observed. The 2,6-
ANS‚âCD complex shows a 14% decrease in association constant
for a concomitant pressure increase from 7 to 345 bar, whereas
no measurable effect on the complexation equilibrium is observed
for the 2,6-MANS‚âCD complex. This result is somewhat surpris-
ing, based on theoretical predictions of pressure-induced dissocia-
tion when â-cyclodextrin is utilized as the host molecule. Simple
volumetric additivity is insufficient to describe the pressure-
dependent behavior of these structurally analogous probe mol-
ACKNOWLEDGMENT
We gratefully acknowledge the General Electric Co. and the
Eli Lilly Corp. for partial support of this research.
(35) Jonas, J.; Jonas, A. Annu. Rev. Biophys. Biomol. Struct. 1 9 9 4 , 23, 287-318.
(36) Wallen, S. L.; Palmer, B. J.; Garrett, B. C.; Yonker, C. R. J. Phys. Chem.
1 9 9 6 , 100, 3959-3964.
Received for review December 16, 1996. Accepted March
19, 1997.X
AC961271O
(37) Yonker, C. R.; Wallen, S. L.; Linehan, J. C. J. Supercrit. Fluids 1 9 9 5 , 8,
250-254.
(38) Ringo, M. C.; Evans, C. E. J. Phys. Chem., submitted.
X Abstract published in Advance ACS Abstracts, May 1, 1997.
2142 Analytical Chemistry, Vol. 69, No. 11, June 1, 1997