Pairwise Gibbs energies of interaction involving N-alkyl-2- pyrrolidinones and related compounds in aqueous solution obtained from kinetic medium effects

Article abstract of DOI:10.1021/jo991262q

Kinetic solvent effects of N-alkyl-2-pyrrolidinones and structurally related compounds on the water-catalyzed hydrolysis reactions of p- methoxyphenyl dichloroacetate (MPDA), 1-benzoyl-3-phenyl-1,2,4-triazole (BPhT), and 1-benzoyl-1,2,4-triazole (BT) in highly dilute aqueous solutions at pH 4 and 298.15 K have been determined by UV/vis spectroscopy. Using a thermodynamic description of solute-solute interactions in aqueous solutions, the kinetic results have been analyzed in terms of pairwise Gibbs energy interaction parameters: G(c) values. These are negative, indicating that hydrophobic interactions in the initial state dominate the medium effects. The interaction parameters increase in the order MPDABT>BPhT. However, when differences in reactivity and transition state effects are taken into account, it appears that BPhT is more successful in establishing hydrophobic interactions with the cosolutes than are MPDA and BT. Using the SWAG-approach for additivity of group interactions, additivity is observed for the first three consecutive CH₂ groups in the cosolute in all three hydrolysis reactions. Larger alkyl substituents cause larger retardations than anticipated on basis of this additivity. The results are explained by intramolecular destructive overlap of the polar hydration shell of the amide functionality and the apolar (hydrophobic) hydration shell of the alkyl group, which extends to the third CH₂ group in the N-alkyl group of the cosolute molecule. The inner apolar groups, therefore, have a reduced apparent hydrophobicity. More remote CH₂ groups develop independent hydrophobic hydration shells. The effect of the position of a CH₂ group in the cosolute molecule is also considered. Kinetic solvent effects with structurally related esters show that amide-amide, ester-ester, and amide- ester group interactions affect the transition state in different ways. Finally, the effects of PVP polymers on the three hydrolysis reactions have been examined. The data presented enhance the understanding of pairwise hydrophobic interactions in aqueous solutions. In addition the results provide insights into the interactions between hydrophobic and hydrophilic hydration shells as well as into the energetics of amide hydration and interactions involving amides in aqueous solution, both playing important roles in protein stabilization.

Full text of DOI:10.1021/jo991262q

J . Org. Chem. 2000, 65, 411-418
411
P a ir w ise Gibbs En er gies of In ter a ction In volvin g
N-Alk yl-2-p yr r olid in on es a n d Rela ted Com p ou n d s in Aqu eou s
Solu tion Obta in ed fr om Kin etic Med iu m Effects
†
†
,†
‡
J oke J . Apperloo, Lisette Streefland, J an B. F. N. Engberts,* and Michael J . Blandamer
Department of Organic and Molecular Inorganic Chemistry, University of Groningen, Nijenborgh 4,
9
747 AG Groningen, The Netherlands, and the Department of Chemistry, University of Leicester,
Leicester LE1 7RH, England
Received August 9, 1999
Kinetic solvent effects of N-alkyl-2-pyrrolidinones and structurally related compounds on the water-
catalyzed hydrolysis reactions of p-methoxyphenyl dichloroacetate (MP DA), 1-benzoyl-3-phenyl-
1
4
,2,4-triazole (BP h T), and 1-benzoyl-1,2,4-triazole (BT) in highly dilute aqueous solutions at pH
and 298.15 K have been determined by UV/vis spectroscopy. Using a thermodynamic description
of solute-solute interactions in aqueous solutions, the kinetic results have been analyzed in terms
of pairwise Gibbs energy interaction parameters: G(c) values. These are negative, indicating that
hydrophobic interactions in the initial state dominate the medium effects. The interaction
parameters increase in the order MP DA<BT<BP h T, suggesting increasing hydrophobic stabiliza-
tion in the order of MP DA>BT>BP h T. However, when differences in reactivity and transition
state effects are taken into account, it appears that BP h T is more successful in establishing
hydrophobic interactions with the cosolutes than are MP DA and BT. Using the SWAG-approach
2
for additivity of group interactions, additivity is observed for the first three consecutive CH groups
in the cosolute in all three hydrolysis reactions. Larger alkyl substituents cause larger retardations
than anticipated on basis of this additivity. The results are explained by intramolecular destructive
overlap of the polar hydration shell of the amide functionality and the apolar (hydrophobic) hydration
shell of the alkyl group, which extends to the third CH
molecule. The inner apolar groups, therefore, have a reduced apparent hydrophobicity. More remote
CH groups develop independent hydrophobic hydration shells. The effect of the position of a CH
2
group in the N-alkyl group of the cosolute
2
2
group in the cosolute molecule is also considered. Kinetic solvent effects with structurally related
esters show that amide-amide, ester-ester, and amide-ester group interactions affect the
transition state in different ways. Finally, the effects of PVP polymers on the three hydrolysis
reactions have been examined. The data presented enhance the understanding of pairwise
hydrophobic interactions in aqueous solutions. In addition the results provide insights into the
interactions between hydrophobic and hydrophilic hydration shells as well as into the energetics
of amide hydration and interactions involving amides in aqueous solution, both playing important
roles in protein stabilization.
In tr od u ction
decrease or increase the chemical potentials of the initial
state and the activated complex, depending on whether
the interactions are favorable (stabilizing) or not favor-
able (destabilizing), and consequently affect the Gibbs
energy of activation.
Kinetic medium effects on hydrolysis reactions in dilute
aqueous solution provide an excellent method for inves-
tigating solute hydration and solute-solute interactions
in aqueous solution, since the activation parameters are
strongly affected by changes in the structural properties
of water in the hydration shell of the reactants during
the activation process.1 Changes in solvent composition,
brought about by addition of small amounts of cosolutes,
affect the hydration characteristics of both reactant and
activated complex due to interactions with the cosolute
About a decade ago, we developed a theory with which
kinetic medium effects of solvolysis reactions can be
analyzed quantitatively in terms of thermodynamic
,2
4,5
interaction parameters. The Gibbs energy of interaction
stems from the nonideal part, the excess Gibbs energy
E
(
G ), of the total solution which is determined by the
chemical potentials of the solutes. This nonideal part can
3
via overlap of their hydration cospheres.
Transition-state theory provides a basis of understand-
ing kinetic medium effects. The interacting cosolutes can
be ascribed completely to solute-solute interactions.
E
Usually, G is expressed by a molality expansion, using
virial coefficients. These coefficients obtain physical
6
significance by the theory of McMillan and Mayer. In
†
E
University of Groningen.
University of Leicester.
1) Blandamer, M. J .; Burgess, J .; Engberts, J . B. F. N. Faraday
sufficiently dilute solutions, G is determined by pairwise
‡
solute-solute interactions only. The thermodynamics
(
Discuss. Chem. Soc. 1989, 85, 309. For a brief review, see: Engberts,
J . B. F. N.; Blandamer, M. J . J . Phys. Org. Chem. 1998, 11, 841.
(4) Blokzijl, W.; Engberts, J . B. F. N.; J ager, J .; Blandamer, M. J .
J . Phys. Chem. 1987, 91, 6022.
(5) Blokzijl, W.; J ager, J .; Engberts, J . B. F. N.; Blandamer, M. J .
J . Am. Chem. Soc. 1986, 108, 6411.
(
2) (a) Blandamer, M. J .; Burgess, J . Pure Appl. Chem. 1983, 55,
5. (b) Blandamer, M. J .; Burgess, J . Pure Appl. Chem. 1990, 62, 9.
3) Gurney, R. W. Ionic Processes in Solution; McGraw-Hill: New
York, 1953.
5
(
(6) McMillan, W. G.; Mayer, J . E. J . Phys. Chem. 1945, 13, 276.
1
0.1021/jo991262q CCC: $19.00 © 2000 American Chemical Society
Published on Web 12/30/1999

4
12 J . Org. Chem., Vol. 65, No. 2, 2000
Apperloo et al.
Sch em e 2. N-Alk yl-2-p yr r olid in on es
Sch em e 1. Rea ction Mech a n ism for th e
Wa ter -Ca ta lyzed Hyd r olysis of
1
-Acyl-3-su bstitu ted -1,2,4-tr ia zolesa
initial state and the activated complex is responsible for
the marked changes in rate constants that are observed
when hydrophobic cosolutes are added. In the past, we
have studied an extensive range of cosolutes using kinetic
medium effects, including mono-, di-, and polyhydric
5
,8
4,9
alcohols, (alkylated) urea(s), mono- and disaccha-
rides,1
0,11
carboxamides, sulfonamides, sulfones and sul-
a
For 1-benzoyl-1,2,4-triazole (BT) R1 ) C6H5 and R ) H; for
2
9
,12
13
1
2
foxides,
sodium n-alkyl sulfates, n-alkylated ammo-
1
-benzoyl-3-phenyl-1,2,4,-triazole (BPhT) R ) R ) C6H5. The
hydrolysis of MPDA follows the same mechanism.
nium bromides,14 and R-amino acids.
given us profound insights into pairwise hydrophobic
interactions in aqueous solution. The results did, in
15,16
The results have
were then linked with kinetics through transition-state
theory, yielding the following equation for a water-
catalyzed hydrolysis reaction4,5
17
most cases, not allow an analysis in terms of additivity
of group contributions toward the medium effect using
the Savage and Wood additivity of groups (SWAG)
1
8
k(mc)
2
approach for solute-solute interactions in aqueous solu-
tion, due to the influence of the polar group hydration.
In this study, we have focused on a different class of
cosolutes, the N-alkyl-2-pyrrolidinones, cyclic amides
with alkyl substituents at the nitrogen atom (Scheme 2).
Two reasons account for this choice. First, these solutes
are highly soluble in water (even N-cyclohexyl-2-pyrro-
lidinone is miscible with water in all proportions).
Therefore, they are suitable cosolutes for studying hy-
drophobic interactions in aqueous solution. Second, we
are interested whether the obtained G(c) values can be
analyzed in terms of additivity of pairwise group interac-
tions, using the SWAG approach. The results allow a
comparison with the results obtained for substituted
acyclic amides9 and enhance our understanding of the
energetics of amide hydration and amide-amide interac-
tions in aqueous solution, which are important in protein
stability and still under debate.19
ln
)
2(gc-IS - gc-AC)m - nφM m
c w
c
k (m ) 0)
0
c
RTm0
(
1)
where k is the pseudo-first-order rate constant for reac-
tion in the aqueous solution containing the cosolute c, k
0
the pseudo-first-order rate constant for reaction in the
absence of cosolute, R the gas constant, T the tempera-
-
1
0 c
ture, m the standard state (1 mol kg ), m the molality
of the cosolute, n the number of water molecules involved
in the activated complex of the hydrolysis reaction (n )
2
M
in the hydrolysis reactions investigated in this study),
the molar mass of water, and φ the practical osmotic
w
coefficient (which equals unity in dilute aqueous solu-
tions). The term (gc-IS - gc-AC) is referred to as the G(c)
value, the pairwise Gibbs energy interaction parameter,
which is the difference in pairwise interactions of the
cosolute (c) with the initial state (IS) and the activated
complex (AC). The second half of the equation reflects
the effect of the cosolute on the reactivity of water, since
water is solvent as well as reactant. Thus, G(c) represents
the overall effect of the cosolute on the Gibbs energy of
activation for the hydrolytic process. G(c) is obtained from
,12
Resu lts a n d Discu ssion
N-Alk yl-2-p yr r olid in on es. We measured the kinetic
solvent effects of five N-alkyl-2-pyrrolidinones (Scheme
the slope of a plot of ln(k/k
cosolute.
0
) versus the molality of the
(8) Blokzijl, W.; Engberts, J . B. F. N.; Blandamer, M. J . J . Am.
Chem. Soc. 1990, 112, 1197.
(9) Kerstholt, R. P. V.; Engberts, J . B. F. N.; Blandamer, M. J . J .
Previously, we applied this quantitative treatment of
rate constants in the analysis of kinetic medium effects
on several hydrolysis reactions in dilute aqueous media.
In particular, the neutral (i.e., water-catalyzed) hydroly-
ses of 1-benzoyl-(3-phenyl)-1,2,4-triazole (BT and BP h T)
and p-methoxyphenyl dichloroacetate (MP DA) have been
investigated in depth. In the pH range where only water
acts as a general base (usually between pH 3-5 for these
hydrolyses), the water-catalyzed hydrolysis reaction pro-
ceeds via a dipolar activated complex in which two water
molecules are involved, with three protons “in flight”7
Chem. Soc., Perkin Trans. 2 1993, 49.
(10) Galema, S. A.; Blandamer, M. J .; Engberts, J . B. F. N. J . Am.
Chem. Soc. 1990, 112, 9665.
(11) Galema, S. A.; Blandamer, M. J .; Engberts, J . B. F. N. J . Org.
Chem. 1992, 57, 1995.
(12) Engberts, J . B. F. N.; Kerstholt, R. P. V.; Blandamer, M. J . J .
Chem. Soc., Chem. Commun. 1991, 1230.
(13) Noordman, W. H.; Blokzijl, W.; Blandamer, M. J .; Engberts, J .
B. F. N. J . Org. Chem. 1993, 58, 7111.
(14) Hol, P.; Streefland, L.; Blandamer, M. J .; Engberts, J . B. F. N.
J . Chem. Soc., Perkin Trans. 2 1997, 485.
(15) Streefland, L.; Blandamer, M. J .; Engberts, J . B. F. N. J . Phys.
Chem. 1995, 99, 5769.
(16) Streefland, L.; Blandamer, M. J .; Engberts, J . B. F. N. J . Am.
Chem. Soc. 1996, 118, 9539.
(Scheme 1). The difference in hydrophobicity between the
(
17) For a review on hydrophobic interactions, see: Blokzijl, W.;
Engberts, J . B. F. N. Angew. Chem., Int. Ed. Engl. 1993, 32, 1545.
(18) Savage, J . J .; Wood, R. H. J . Solution Chem. 1976, 5, 733.
(19) (a) Williams, D. H. Aldrichim. Acta 1991, 24, 71. (b) Doig, A.
J .; Williams, D. H. J . Am. Chem. Soc. 1992, 114, 338.
(
7) (a) Karzijn, W.; Engberts, J . B. F. N. Tetrahedron Lett. 1978,
1
787. (b) Mooij, H. J .; Engberts, J . B. F. N.; Charton, M. Recl. Trav.
Chim. Pays-Bas 1988, 107, 185.

Pairwise Gibbs Energies of Interaction
J . Org. Chem., Vol. 65, No. 2, 2000 413
F igu r e 1. Solvent effects on the hydrolysis of BP h T in the
presence of NMP (9), NEP (O), NiPP (2), and NnBP (1).
F igu r e 2. G(c) values (J kg mol-2) versus the number of CH
2
groups in the N-alkyl chain of N-alkyl-2-pyrrolidinones for the
kinetic probes MP DA (9), BP h T (2), and BT (O).
Ta ble 1. G(c) Va lu esa (J k g m ol-2) for th e Differ en t
Cosolu te-P r obe Com bin a tion s
It is remarkable that the MP DA hydrolysis is retarded
to a much larger extent than the hydrolyses of both BT
and BP h T. In other words, the ester hydrolysis is more
sensitive toward solvent effects than the amide hydroly-
probe
cosolute
MPDA
BPhT
BT
NMP
NEP
NiPP
NnBP
NCHP
-925(10)
-1176(18)
-1407(10)
-1989(16)
-2980(23)
-133(4)
-157(10)
-206(15)
-467(21)
-1000(40)
-292(10)
-354(11)
-354(4)
-600(15)
-600(41)
9
ses. This is in accordance with earlier findings. This
increase in Gibbs energy of activation for the ester
hydrolysis is unlikely to be governed by increased sta-
bilization of the initial state, because MP DA is less
hydrophobic than BP h T and BT. Stabilization of the
initial state by amide-amide or amide-ester H-bonding
interactions is out of the question since both reactants
a
Errors in parentheses.
2
) on the hydrolyses of MP DA, BP h T, and BT at
different cosolute molalities and up to several molal at
98.15 K and pH 4. N-Cyclohexyl-2-pyrrolidinone is in
fact not part of the series of homologues shown in Scheme
but is interesting in view of its unlimited solubility in
(ester and tertiary amides) and cosolutes (tertiary amides)
are hydrogen bond acceptors only. Moreover, it has been
2
suggested that amides form stronger H-bonds with water
than with other amides;2
2,23
i.e., the solute-solvent
2
interactions dominate the solute-solute interactions, and
the same is anticipated for the ester functionality. We
therefore contend that the larger solvent effects for the
ester hydrolysis find their origin in a transition-state
effect; i.e., the transition state of the amide hydrolysis is
stabilized more than the transition state of the ester
hydrolysis in the presence of the N-alkyl-2-pyrrolidino-
nes. The transition state of the ester hydrolysis is less
polar than that of the amide hydrolyses (i.e., the differ-
ence in polarity between IS and TS is smaller for the ester
water and furthermore can provide information about
differences in hydrophobicity of cyclic and acyclic alkyl
substituents. As an example, the solvent effects of the
series in Scheme 2 on the hydrolysis of BP h T are shown
in Figure 1 up to 1 m of added cosolute. Generally, linear
relationships between ln(k/k
0
) and the molality of the
cosolute have been obtained up to approximately 0.75
2
0
m, consistent with pairwise interactions between coso-
lute and kinetic probe. From the slopes of the correlations
in Figure 1, G(c) values were obtained, using eq 1. The
results, together with the G(c) values obtained for the
hydrolyses of BT and MP DA, are shown in Table 1. To
illustrate the dependence of G(c) on the structure of the
cosolute, G(c) values have also been plotted versus the
0
hydrolysis, as is reflected by a larger k ). There is an
increase in polarity of the solvent when the rather polar
N-alkyl-2-pyrrolidinones are added to the medium, which
would explain the larger stabilizing effect on the transi-
tion state of the triazole hydrolysis reactions due to polar
interactions. Considering these transition-state effects,
an interpretation of the results in Table 1 solely in terms
of hydrophobic effects is inadequate. Although MP DA
appears to be able to interact more strongly with the
cosolutes via hydrophobic interactions than BT and
BP h T, this is in fact not true. Table 2, which contains
data similar to that in Table 1, but now re-expressed with
number of CH
2
groups in the N-alkyl substituent (Figure
2
). In all cases, N-alkyl-2-pyrrolidinones cause a retarda-
tion of the hydrolysis reactions, expressed by negative
G(c) values. These retardations can be largely attributed
to favorable interactions between apolar moieties in the
cosolute and the reactant molecule, i.e., an initial state
stabilization due to hydrophobic interactions.2
1
0
the NMP retardations as a reference rather than k ,
(
20) With the exception of NCHP as a cosolute. Solvent effects of
NCHP on the three hydrolysis reactions show linear behavior up to
.2-0.3 m. Above these concentrations higher-order interactions come
reflects this fallacy. The pattern emerging from Table 2
is visualized in Figure 3. By using the converted data in
0
into play. Collection of kinetic data up to 3 m yielded S-shaped curves,
which could be perfectly analyzed with the Menger-Portnoy model
for micellar catalysis, indicating the formation of NCHP aggregates.
This was confirmed by fluorescence spectrophotometric measurements
using pyrene as a probe that is sensitive to the solvent microenviron-
ment.
(21) Benak, H.; Engberts, J . B. F. N.; Blandamer, M. J . J . Chem.
Soc., Perkin Trans. 2 1992, 2035.
(22) Nilar, S. H.; Pluta, T. S. J . Am. Chem. Soc. 1995, 117, 12603.
(23) Eberhardt, E. S.; Raines, R. T. J . Am. Chem. Soc. 1994, 116,
2149.

4
14 J . Org. Chem., Vol. 65, No. 2, 2000
Apperloo et al.
Ta ble 3. G(CH2) Va lu es in J k g Mol- for Sever a l
2
Ta ble 2. G(c)/G(NMP ) Va lu es for th e Differ en t
Cosolu te-P r obe Com bin a tion s
Cosolu te-P r obe Com bin a tion s
probe
cosolutes
cosolute
MPDA
BPhT
BT
probe
cyclic amidesa
acyclic amides9
alcohols32,8
NMP
NEP
NiPP
NnBP
NCHP
1
1
1
MPDA
BPhT
BT
-241
-37
-142
-51
n.d.b
-136
1.27
1.52
2.15
3.22
1.18
1.55
3.60
7.52
1.21
1.21
2.05
2.05
-31
n.d.b
-90
a
This study (first three CH2 groups included only). b Not
determined.
for the acyclic amides. This pattern is not easily explain-
able in terms of pairwise group interactions. The different
noncovalent interactions playing a role in both the initial
state and the transition state are difficult to unravel.
Moreover, since the N-alkyl-2-pyrrolidinones are tertiary
amides, a comparison is probably not justified. The
2
G(CH ) values for the cyclic amides are about one-third
of the values obtained for short-chain alcohols for the
hydrolyses of BP h T and BT (Table 3). This indicates that
the amide group is more extensively hydrated than the
alcohol functionality (presumably in terms of hydrogen
bonding), and the apparent hydrophobicity is lower for
2
CH groups directly attached to an amide functionality.
Or, differently formulated, the hydration shells of the
polar and apolar parts of the cosolutes are more incom-
patible in the case of the cyclic amides.
F igu r e 3. G(c)/G(NMP) versus the number of CH
the N-alkyl chain of N-alkyl-2-pyrrolidinones for the hydrolysis
of MP DA (9), BP h T (2), and BT (O).
2
-groups in
The intercepts in Figure 2 represent the contribution
of the pyrrolidinone unit (p) toward the medium effect.
The G(p) values are -567 (MP DA), -74 (BP h T), and
-
2
-
25 (BT) J kg mol . The fact that these values are
Table 2, we correct for differences in transition-state
stabilization and reactivity, thereby allowing a better
comparison of the probes, in which initial state effects
substantially different for the different kinetic probes
suggests again that there are specific cosolute-probe
interactions cooperative, most likely in the initial as well
as in the transition state. An extrapolation toward a
contribution for the amide (CONH) unit only (i.e.,
(hydrophobic effects) might show up more clearly. As
anticipated and as shown in Figure 3, hydrolysis of the
more hydrophobic probe (BP h T) is then relatively most
sensitive to the hydrophobicity of the cosolute. Although
the absolute values of the medium effects are not solely
governed by hydrophobic interactions, the relative values
reveal that these interactions play an important role in
the recognition process between the cosolutes and the
kinetic probes.
n(CH
contribution to the G(c) of the three CH
a ring system is not obvious in this case. It was shown
previously that CH units in a ring can either reduce the
apparent hydrophobicity or increase the apparent hy-
2
)
total ) 0) is not so straightforward either, since the
2
units located in
2
5
9
2
drophobicity relative to CH groups that are not joined
in a ring system. In addition, a modified G(CONH) would
be obtained in this case; its solvent effect is influenced
by the attached alkyl groups. A rough estimate, however,
would imply a positive contribution toward G(c) for the
amide functionality. This is not in agreement with
The next topic of interest is whether the medium
effects, as expressed in G(c) values, can be analyzed in
terms of additivity of pairwise group interactions, as was
described by Savage and Wood.18 Does each CH
group
2
in the cosolute molecule, irrespective of its position, add
a constant increment to the medium effect? Figure 2
suggests that this is not the case throughout the whole
series of N-alkyl-2-pyrrolidinones, since there is no linear
9
previous findings, which suggest a negative contribution.
The above discussion is based on the CH
additivity that is observed for the short-chained N-alkyl-
-pyrrolidinones. Upon extending the chain to n-butyl
2
group
2
relationship between G(c) and the number of CH
It seems that the longer the alkyl chain, the larger the
contribution of a CH unit toward the medium effect. This
2
groups.
and cyclohexyl, deviation from additivity develops. These
cosolutes cause a solvent effect much larger than ex-
pected when the observed additivity for the short-chained
solutes is extrapolated. A similar pattern is observed for
other cosolutes bearing hydrophilic and hydrophobic
2
pattern is particularly pronounced for solutes with apolar
groups larger than isopropyl. It appears that for the
shorter alkyl chains additivity of CH
be applied. When the G(c) values for NMP, NEP, and
NiPP are plotted against n(CH ) on a larger scale (not
shown), additivity is reasonably good, with slopes rep-
resenting the G(CH ) (Table 3). Thus, there is group
additivity for the first 3.5 CH units, but the longer alkyl
chains deviate from this additivity pattern. The G(CH
2
interactions may
14,24
groups.
The explanation for this phenomenon is that the
smaller alkyl substituents are entirely located within that
part of the hydration sphere of the amide functionality
where the hydration water is strongly affected by hydro-
gen bonding with the amide group. This position in the
hydration sphere prevents complete development of
hydrophobic hydration shells for these alkyl moieties that
2
2
2
2
)
value for MP DA is more negative than that obtained for
a series of acyclic primary, secondary, and tertiary
9
amides for the same kinetic probe. On the contrary, for
(24) Streefland, L.; Blandamer, M. J .; Engberts, J . B. F. N. J . Chem.
the hydrolysis of BP h T a more negative value is found
Soc., Perkin Trans. 2 1997, 769.

Pairwise Gibbs Energies of Interaction
J . Org. Chem., Vol. 65, No. 2, 2000 415
Sch em e 3. Ad d ition of a CH Gr ou p To,
are incompatible with the hydrophilic amide hydration
shell. Consequently, their availability to interact with
hydrophobic groups in the substrate via hydrophobic
interactions is reduced. The effect of the cosolute amide
group on the hydrogen-bonding interactions is sensed
particularly in the first hydration shell. Since a break in
2
Resp ectively, th e N-Alk yl Ch a in a n d th e Rin g
System of NMP
2
additivity is observed after three consecutive CH groups,
it seems that this intramolecular hydration shell overlap
stretches over one layer of water molecules in the
hydration shell.
Outside the influencing effect of the amide hydration
sphere, CH
but with a more negative value for G(CH
CH groups are less influenced in their interactions with
the kinetic probe. Unfortunately, insufficient data are
available to support this view.
At first sight it is rather surprising that additivity is
found for NMP, NEP, and NiPP. A gradually decreasing
influence of the amide functionality would be expected.
But when elaborating on the assumption that the part
of the amide hydration shell which affects the hydropho-
bic hydration of attached alkyl groups contains only one
layer of water molecules, the Me, Et, and i-Pr groups are
indeed situated in this layer because the effective diam-
eter of a water molecule26 is ca. 2.75 Å and the distance
between the nitrogen atom and the â carbon atom in the
N-alkyl substituent is 2.39 Å. This assumption is sup-
ported by calculations27 on the interaction between
2
-group additivity is likely to occur as well,25
2
), since these
2
J kg mol- . G(c) values for NMP and NEP are -925 and
2
-2
-1176 J kg mol , respectively (Table 2). In other words,
-
2
a CH
2
group in the ring system adds -170 J kg mol to
group in the N-alkyl
the medium effect, while a CH
2
-2
chain contributes -251 J kg mol . The SWAG approach
is clearly not applicable to these cases. The obvious
explanation for the lower apparent hydrophobicity of the
2
CH group in the ring system is the reduced hydrophobic
surface area. The part of the apolar surface inside the
ring is not exposed to the solvent, and interactions of that
part of the molecule with (the hydration shell of) the
kinetic probe do not take place. In addition, the confor-
mational differences between a five- and six-membered
ring may have an impact on their hydrophobicity. This
would be particularly important when there is a prefer-
ential site of interaction with the kinetic probe.
>
CH- and >NH in water, which spans 3 Å. Therefore,
deviation from additivity is likely to occur already for
N-(n-propyl)-2-pyrrolidinone, which contains a γ-C atom
Ch a n gin g th e F u n ction a l Gr ou p : Str u ctu r a lly
Rela ted Ester Cosolu tes. It was noted above that
hydrophobic interactions do not solely determine the
medium effects of N-alkyl-2-pyrolidinones on the hy-
drolysis reactions of MP DA, BP h T, and BT. Since these
reactions involve the hydrolysis of an ester and two
amides, it would be particularly relevant to study the
effects of ester cosolutes of similar structure as the
N-alkyl-2-pyrrolidinones. In this way, information about
amide-amide, amide-ester, and ester-ester interactions
in the activated complex can be obtained, providing a
more complete picture of the interplay of the several
noncovalent interactions governing the medium effect.
In view of this discussion, the kinetic results pertinent
to the structurally related esters and amides shown in
Scheme 4 have been compared. The G(c) values are
compiled in Table 4. For a valuable comparison, the
results for the ester/amide cosolute pairs shown above
are also given relative to G(c) for amide cosolutes (Table
5). From Table 5, it appears that the Gibbs energies of
pairwise interactions of MP DA are rather similar for
both amide and ester cosolutes, whereas for BP h T and
BT, pairwise Gibbs energy interaction parameters for
ester cosolutes are, respectively, 3.5 and 1.4 times more
negative than those for the amide cosolutes.
(G(c) not determined), and the deviation is clear-cut in
the case of NnBP. For these cosolutes, the outer CH
2
moieties are more available for hydrophobic interactions
with the kinetic probes, i.e., for the kinetic probe they
have larger apparent hydrophobicities. NCHP has one
2
more CH moiety than NnBP, but they have similar
distances between the N-atom and the most remote
C-atom. For both the MP DA and BP h T hydrolyses, |G(c)|
doubles with an enormous decrease of 1000 and 500 J
-
2
kg mol , respectively, but, interestingly, for BT there is
no difference in the G(c) value for NnBP and NCHP. In
this case, interactions do not seem to be of a hydrophobic
nature but instead may be purely determined by effects
caused by interactions involving the amide functionality
which are modified by the size of the substituent.
Effect of Ad d in g a CH
2
Moiety Eith er to th e
N-Alk yl Ch a in or to th e P yr r olid in on e Rin g. One of
the assumptions of the SWAG approach is that group
interactions are independent of the group’s position in
the molecule. To check the validity of this assumption,
we measured the kinetic solvent effect of N-methyl-2-
piperidone (2PIP), a six-membered cyclic amide. The
results can be compared to the data already available
for NMP and NEP (Scheme 3). In the case of perfect
pairwise group additivity, NEP and 2PIP should cause
the same solvent effect. Due to overlap of the UV
absorption spectra, medium effects on the hydrolysis of
BP h T and BT could not be determined, and G(c) values
have only been obtained for the hydrolysis of MP DA. The
G(c) value for 2PIP for the hydrolysis of MPDA is -1095
Clearly, the ester probe does not “distinguish” between
either ester or amide solutes, whereas the amide probes
do. This pattern is illustrated in Figure 4, where the
solvent effects caused by the amide cosolutes have been
plotted versus those caused by ester cosolutes. The data
points for BP h T in particular seem to deviate from equal
cosolute effects. How can these results be explained? The
differences in retardation caused by amide and ester
cosolutes for the hydrolysis of BP h T and BT do not seem
to reflect the hydrophobicity of the cosolute. In terms of
2
(25) We observed additivity of the CH group contribution outside
the influence of the polar ammonium hydration shell in the case of
alkylated ammonium bromides (see ref 14).
(26) Pierotti, R. J . J . Phys. Chem. 1965, 69, 281.
(27) Pettitt, B. M.; Karplus, M. J . Phys. Chem. 1988, 92, 3994.
number of CH
2
moieties, the amide cosolutes are more

4
16 J . Org. Chem., Vol. 65, No. 2, 2000
Apperloo et al.
Sch em e 4. N-Alk yl-2-p yr r olid in on es a n d
Str u ctu r a lly Sim ila r Ester Cosolu tes
F igu r e 4. |G(c)| values of amide cosolutes plotted versus the
|
G(c)| values of ester cosolutes for the hydrolysis of MP DA (9),
BP h T (2), and BT (O). Dotted line represents equal solvent
effects for both amide and ester cosolutes (G(c) values in J kg
-
2
mol ).
acceptors, they can stabilize the polarized transition state
of the hydrolysis reactions to a greater extent than the
ester cosolutes can. This larger stabilization of the
transition state would explain the smaller kinetic me-
dium effects caused by the amide cosolutes for the
hydrolysis of BT and BP h T, but not for MP DA. On the
whole, the G(c) values for the effects of the ester and
amide cosolutes on the hydrolysis of MP DA are remark-
ably similar (see Table 4). Despite the fact that the
MP DA hydrolysis is the most sensitive to changes in the
medium, it is obviously not particularly sensitive to
changes in functional groups (i.e., amide vs ester func-
tionality) or to the exact structure of the cosolute. This
observation could lead to the conclusion that medium
effects on the hydrolysis of MP DA do reflect the hydro-
phobicity of the cosolute better than those for BT and
BP h T. However, a clear correlation between the G(c) and
Ta ble 4. G(c) Va lu esa (J k g m ol-2) for th e Th r ee P a ir s of
Ester /Am id e Cosolu tes (See Sch em e 4) for th e Hyd r olysis
of MP DA, BP h T, a n d BT
probe
cosolute
MPDA
BPhT
BT
BuLac
NMP
MAc
DMA
ValLac
-836(10)
-925(10)
-857(15)
-841(13)
-964(18)
-1095(12)
-435(9)
-133(4)
-395(11)
-292(10)
-408(14)
-301(14)
-498(13)
-139(15)
n.d.b
n.d.
n.d.
2
PIP
n.d.
a
Errors in parentheses. b Not possible to determine.
Ta ble 5. G(c) Va lu es (J k g m ol-2) for Ea ch Ester /Am id e
Cosolu te Com bin a tion Rela tive to th e Am id e Cosolu te
2
n(CH ) does not appear to exist. Presumably, the insen-
sitivity to the polar group is fortuitous and caused by a
number of counteracting contributions to the Gibbs
energy of interaction, leading to comparable retardations
for ester and amide cosolutes.
probe
cosolute
MPDA
BPhT
BT
NMP/BuLac
DMA/MAc
PIP/ValLac
1:0.9
1:1
1:0.9
1:3.3
1:3.6
n.d.
1:1.4
1:1.4
n.d.
a
P VP a s a Cosolu te. We also investigated the effects
of poly(vinylpyrrolidinone) (PVP) on the hydrolysis of
MP DA, BP h T, and BT. In this way, the kinetic effects
of a concentrated number (cluster) of amide bonds can
be studied as a simple model for a protein backbone. PVP
2
a
Not possible to determine.
hydrophobic than the ester cosolutes, since they contain
an additional CH group. On the other hand, the hydra-
3
30
has a broad variety of applications, due to its nontoxic
tion of the polar functional group can influence the
hydrophobicity of the apolar groups, so it might be
erroneous to assume that the amide cosolutes are more
hydrophobic. Probably their hydrophobicities are similar
and thus do not explain the observed effects. We note
that amides are more polar than esters. For example, IR
wavenumbers for CdO stretching vibrations are 1775
and 1700 cm for ValLac and NMP, respectively. Also,
the strengths of H-bond interactions with phenol as an
H-bond donor show that amides are more polar than
esters.29 Since the amide cosolutes are better H-bond
nature and the fact that it is very soluble in both water
and a large number of organic solvents. PVP interacts
both with hydrophobic and with hydrophilic groups. The
30
solvent effects of PVP have been investigated for two
molecular weights: MW 8100 and 57 500. The results
have been analyzed in terms of retardation caused per
monomer molality of PVP, to make a comparison with
NMP, which is equivalent to the PVP monomer in terms
of number of CH groups. Clearly, a pairwise Gibbs energy
interaction parameter cannot be determined, since there
are many kinetic probe molecules interacting with one
polymer molecule. In Figure 5 the solvent effects of the
high molecular weight polymer have been plotted as a
-1
28
(28) Silverstein, M.; Clayton Bussler, G.; Morrill, T. C. Spectroscopic
Identification of Organic Compounds, 3rd ed.; Wiley: New York, 1974;
p 102.
(29) Murthy, A. S. N.; Rao, C. N. R. Appl. Spectrosc. Rev. 1968, part
(30) Encyclopedia of Polymer Science and Technology; Mark, H. F.,
2
, 69.
Gaylord, N. G., Eds.; Wiley: New York, 1971; Vol. 14, p 243.

Pairwise Gibbs Energies of Interaction
J . Org. Chem., Vol. 65, No. 2, 2000 417
MP DA, BP h T, and BT can be analyzed in terms of
pairwise Gibbs energy interaction parameters. These are
negative, reflecting the dominant stabilization of the
initial state by hydrophobic interactions. However, hy-
drophobic interactions do not solely govern the solvent
effects, as was particularly revealed by comparison of the
results for the different kinetic probes. Noncovalent
interactions involving the polar amide and ester func-
tional groups in kinetic probe and cosolute affect the
medium effects as well and play a role in the stabilization
of the activated complexes.
Regarding additivity of functional group interactions:
2
there is no unique rate-retarding effect by a CH group,
but its effect in the molecule depends on the distance to
the amide functionality. As a result of the intramolecular
destructive overlap of the amide (hydrophilic) and alkyl
(hydrophobic) hydration shells of the cosolute molecules,
F igu r e 5. Rate effects caused by PVP (MW 57 500) on the
hydrolyses of MP DA (9), BP h T (2), and BT (O), relative to
the rate constant in water (k ).
0
reduced contributions of CH interactions to the medium
2
effects are observed for the first three consecutive carbon
atoms on the amide nitrogen atom. These contributions
are additive. However, the hydrophobic effect becomes
increasingly more pronounced for hydrophobic moieties
which are more than three consecutive carbon atoms
away from the amide functionality. Additivity of these
function of monomer molality for the three kinetic probes.
Within the experimental error, the results for the low
molecular PVP are similar to those shown in Figure 5.
The retardations of the hydrolyses of MP DA and BT
caused by PVP are slightly less than those caused by
similar concentrations of NMP. Since the retardations
of NMP and PVP are similar, we contend that the
polymer monomeric units behave similarly compared
with NMP in their interactions with the two probes.
Remarkably, PVP slightly increases the reaction rate of
the hydrolysis of (the most hydrophobic probe) BP h T.
Note that the effect of NMP on the hydrolysis of BP h T
was also small, although retarding. Thus, for BP h T at
least, NMP has only little hydrophobic character.
The differences observed for the solvent effects caused
by PVP and NMP find their origin in the differing
hydration properties of PVP and NMP. First, PVP
presumably adopts a random coil configuration in aque-
ous solution, whereby inner residues are less hydrated
and also less accessible for interactions with the kinetic
probes. The size of the probes, and therefore steric
aspects, may be of importance as well. Second, in the PVP
molecule, the amide groups are sufficiently close for the
hydration shells to overlap and it is unknown how this
overlap affects the polar and apolar hydration within the
molecule. A space-filling model of PVP shows that the
nitrogen atom of the amide bond is buried between the
CH-groups of both the backbone and ring system. Hence,
polar stabilizing interactions with the activated complex
take place via a carbonyl rather than via an amide group.
Considering these differences, it is clear that the relative
accelerations caused by PVP cannot be ascribed to a
single structural or hydration property. Overall, PVP
does not significantly disturb the 3-D structure of water;
the hydrophobic and hydrophilic parts reduce each other’s
effects on water-water interactions, and its presence is
only just “noticed” by the kinetic probes. The kinetic
solvent effects clearly reflect the Gibbs interaction ener-
gies of both the initial state and the transition state with
PVP.
CH groups is anticipated but was not investigated for
2
the N-alkyl-2-pyrrolidinones.
A kinetic solvent effect study involving cyclic esters
that are structurally related to the N-alkyl-2-pyrrolidi-
nones showed how subtle the balance of noncovalent
interactions can be.
The effects of PVP polymers on the three hydrolysis
reactions are small and suggest that the hydrophilic and
hydrophobic interactions nearly cancel. For BP h T, a
slight rate acceleration is observed in the presence of
PVP, indicating the more hydrophilic character of the
cosolute.
The results contribute to a better understanding of
noncovalent interactions in dilute aqueous media. In
particular, the energetics of hydrophobic interactions and
interactions involving the amide functionality, which are
of crucial importance for the rather marginal stability of
the folded protein, have been considered in detail.
Exp er im en ta l Section
Mater ials. p-Methoxyphenyl dichloroacetate (MP DA), 1-ben-
zoyl-1,2,4-triazole (BT), and 1-benzoyl-3-phenyl-1,2,4-triazole
7
,31
(BP h T) were prepared according to literature procedures.
Cosolutes were commercially available (DMA, MAc from
J anssen, NiPP, 2PIP from Aldrich, NEP, ValLac from Fluka,
NMP, NCHP, BuLac, PVP from GAF, NnBP from Tokyo
Kassei, J apan). All cosolutes, except for PVP, were purified
by distillation in vacuo prior to immediate use.
Kin etic Mea su r em en ts. Aqueous solutions for the kinetic
measurements were prepared by weight immediately before
use. Water was distilled twice in an all-quartz distillation unit.
The pH of the solution was carefully adjusted with an aqueous
HCl solution with an Orion pH-meter. Between 5 and 8 µL of
-4
a stock solution containing the kinetic probe (ca. 10 M) in
acetonitrile (P.A. quality) were injected into 2.5 mL of reaction
medium in a quartz cuvette and placed in a thermostated cell
compartment (25.0 ( 0.05 °C) of a Perkin-Elmer λ5 or λ2
spectrophotometer. Pseudo-first-order rate constants were
determined by following changes in absorbance at 288 nm
Con clu sion s
(
31) Engbersen, J . F. J .; Engberts, J . B. F. N. J . Am. Chem. Soc.
975, 97, 1563.
32) Blokzijl, W.; Blandamer, M. J .; Engberts, J . B. F. N. J . Org.
Chem. 1991, 56,1832.
1
At low molalities, kinetic medium effects of N-alkyl-
-pyrrolidinones on the water-catalyzed hydrolyses of
(
2

4
18 J . Org. Chem., Vol. 65, No. 2, 2000
Apperloo et al.
(
MP DA), 273 nm (BP h T), or 250 nm (BT) using a Perkin-
kinetics were observed. Data were converted to rate constants
using a commercially available data station. Rate constants
were obtained in triplicate and generally reproducible to within
1%. Rate constants for PVP were reproducible to within 3%.
Rate constants were determined for at least five different
molalities. G(c) values were obtained by a linear regression
program. Errors in G(c) values are standard deviations.
Elmer λ2 or λ5 UV/vis spectrophotometer. Rate constants for
-
3
-1
the reaction in the absence of cosolute were 3.0 × 10
s ,
-
3
-1
-3 -1
1
.24 × 10
s
, and 2.07 × 10
s , respectively, in good
4
,5
agreement with literature values. The half-lives of these
hydrolysis reactions are such that any possible hydrolysis of
cosolutes can be safely neglected.
In general, the absorbance did not exceed 0.7. The reactions
were followed for about 10 half-lives, and excellent first-order
J O991262Q