NMR On-Line Monitoring of Esterification
Catalyzed by Cutinase
Catherine Sarazin,* Frangoise Ergan *t Jean-Paul Seguin, Gerard
Goethals,* Marie-Dorninique Legoy,h and Jean-Noel Barbotin
Laboratoire de Genie Cellulaire, Universite de Picardie Jules Verne, 33 rue
Saint-Leu, 80039 Amiens Cedex, France
Received October 19, 1995/Accepted March 19, 1996
A nuclear magnetic resonance (NMR) method has been
developed to monitor on-line Iipase-catalyzed esterifica-
tion reactions without the need to sample the reaction
medium. The technique, through 'H NMR, measures the
concentrations of alcohol, ester, hydroxylic hydrogens
in the organic phase, and hydroxylic hydrogens in the
aqueous phase, if any. Also, the chemical shift evolution
of the two types of hydroxylic hydrogens has been fol-
lowed, providing information on water content of the
organic phase and on the appearance of a distinct aque-
ous phase. As far as 13C NMR is concerned, it has been
possible to measure, first the acid and the ester concentra-
tions in the carbonyl region, and second, the alcohol and
the ester concentrations in the methylene region. All 'H
and 13C results are in agreement with one another. Fur-
thermore, NMR allows for the choice of detection zone.
Preliminary studies on the solid phase proved the pres-
ence of much more water in the solid phase than in the
organic phase, and also gave evidence of the existence
of two types of esters, one in the organic phase, mainly
associated with the acid, and the other one not associated
with the acid, most probably entrapped within the solid
enzyme. 0 1996 John Wiley & Sons, Inc.
Key words: NMR studies esterification cutinase lipase
INTRODUCTION
Enzymatic catalysis in organic media has emerged as
a powerful technology creating new opportunities for
enzymes as practical catalysts. Lipases are of growing
interest as catalysts in fat technology and in organic
synthesis. However, to take full advantage of this tech-
nology, one must understand the physical basis for varia-
tion of enzyme properties with the medium. In the last
few years, many investigations have been carried out
into the effect of water content and nature of organic
solvents on catalytic behavior of enzymes suspended in
low-water media (Dordick, 1989; Gupta, 1992; Kliba-
* To whom all correspondence should be addressed. Telephone:
(33) 22-82-74-71; fax: (33) 22-82-7595,
I' Present address: Institut Universitaire de Technologie du Mans,
DCpartement Biologie AppliquCe, 53020 Lava1 Cedex, France.
Present address: Laboratoire de Chimie Organique et CinCtique,
FacultC des Sciences, UniversitC de Picardie Jules Verne, 80039
Amiens Cedex, France.
5 Present address: Laboratoire de Genie ProtCique et Cellulaire,
UniversitC de La Rochelle P61e Sciences et Technologie, 17042 La
Rochelle, France.
Biotechnology and Bioengineering, Vol. 51, Pp. 636-644 (1996)
0 1996 John Wiley & Sons, Inc.
nov, 1989). Heterogeneous enzymatic catalysis assigns
a fundamental role to the microenvironment: a minimal
amount of water in the substrates largely reduces the
lag time (Kanerva et al., 1990; Kitaguchi et al., 1990;
Zaks and Klibanov, 1988). It is well established that,
for almost all enzymes, a layer of adsorbed water is
essential to promote their catalytic activity in organic
solvents (Zaks and Klibanov, 1988).
Some NMR methods have already been described
in enzymology to determine the structure of products
obtained (Rabiller et al., 1992) to analyze the composi-
tion of the medium on samples withdrawn and extracted
at preset time intervals (O'Connor et al., 1992; Rabiller
and MazC, 1989; Roberts, 1990). Kinetic aspects have
also been studied (Bhamidipati and Hamilton, 1993;
Brosio et al., 1993;Mendz et al., 1993). Solid-state NMR
was used to investigate solvent dependence of enzyme
dynamics (Burke et al., 1992). A 2H NMR technique
was described to measure total water bound to subtilisin
in the microgram range through the analysis of the hy-
droxyl signal (Parker et al., 1992). In a previous work
(Decagny et al., 1995),we have investigated water activ-
ity through the use of 'H NMR spectroscopy in biphasic
media. Furthermore, NMR spectroscopy provides infor-
mation on substrate and product structure and dynamics
through chemical shift evolution. Changes in association
between them can be monitored as a function of reaction
time and partition phenomena can be observed (Rob-
erts, 1990).
The present study describes a new approach to moni-
tor the enzyme reaction in multiphase medium. NMR
offers, in theory, the possibility of measuring exactly
and quickly both structure and quantity of molecules
without the need for previous separations. Therefore,
it may be used on the whole reaction medium. The
major problem concerns the sensitivity of the method
(Pollesello et al., 1993).
In this work, the enzymatic esterification of butan-l-
01 with caprylic acid catalyzed by a cutinase (Martinez
et al., 1992) is carried out directly in the NMR tube in
the probe of the spectrometer. Proton and carbon spec-
tra are recorded automatically until the end of the reac-
tion. Concentrations of all species including water are
determined simultaneously in situ, and chemical shifts
are analyzed in terms of associations.
CCC 0006-3592/96/060636-09
MATERIALS AND METHODS
Chemicals
Cutinase (purity 60%) from Fusariurn sofani expressed
in E. coli (Martinez et al., 1992)was a generous gift from
Corvas International NV (Gent, Belgium). Caprylic acid
and butan-1-01 were purchased from Sigma. All other
reagents were of analytical grade.
Enzyme Reaction
The chosen system was the synthesis of n-butyl capry-
late from caprylic acid and butan-1-01, as described
below:
8CH3-7CH2-6CH2-5CH2-4CH2-3CH2-2CH2-1COOH
caprylic acid
cutinase
sCH~-'CH~-6CH~-5CH~-4CH~-3CH~-aCH~-COO-a'CH~-2fCH~-3rCH~-4rCH~
n-butyl caprylate
The reaction took place inside an NMR tube with a
diameter of 5 mm. The enzyme (6 mg) was transferred
into the NMR tube and equilibrated with MgC12 salt
for 2 days under vacuum to obtain a water activity of
0.35. The esterification reaction starts by the addition
of 300 pL of an equimolar mixture of caprylic acid and
butan-1-01. Thus, the organic phase was solely composed
of the substrates. The desired water activity of the
substrates (0.35) was previously obtained by adding 1%
(w/w) of water in the organic mixture. The tube was then
placed inside the NMR probe. Water activity, measured
with a Novasina thermoconstanter, was used to charac-
terize the water in the solid phase and in the organic
phase for multiphase media (Halling, 1994).
NMR Spectra
NMR spectra were recorded continuously at 30"C, with
a spin rate of 20 Hz (1200 rpm), until the end of the
reaction, with a Bruker AM-300 WB spectrometer
equipped with an Aspect 3000 data system and using a
5-mm lH/13C dual probe. Proton spectra were recorded
at 300 MHz with a spectral width of 6000 Hz; free induc-
tion decays (FIDs) were sampled over 8K and zero
filled to 16 K datapoints to enhance digital resolution.
Relaxation delay was set to 3 s and eight scans were
collected following a 30" excitation pulse. The proton-
decoupled carbon spectra were recorded at 75 MHz
with a spectral width of 18,000 Hz; FIDs were sampled
SARAZIN ET AL.: NMR STUDIES OF ESTERlFlCATlON CATALYZED BY CUTINASE
over 16K and zero filled to 32K datapoints. Care has
been taken when using integrated 13Csignals for quanti-
tative work to ensure carbon relaxation; the relaxation
delay time was set to 15 s and 20 scans were collected
following an 80" excitation pulse. The carbon FID imme-
diately followed by the proton FID was recorded auto-
matically, then a 10-min delay was allowed before the
next set of carbon and proton FIDs. Carbon and proton
FIDs were normalized before the Fourier transforma-
tion. An exponential multiplication of 3 Hz was applied
on the carbon FID. The deuterium of C6D6 (20 p L ) was
used as the field frequency lock and the chemical shifts
were referenced to TMS. No isotopic exchange was ob-
served in presence of C6D6.
+ 4CH3-3CH2-ZCH2-'CH20H
butan-1-01
-k H20
Quantification of Substrates and Products
Concentrations (butan-1-01, water, hydroxylic hydro-
gens) were quantified using equations taking into ac-
count the number of hydrogens resonating at each fre-
quency and integrated area of the concerned signals.
For example, the following equation was used to deter-
mine the butanol concentration:
integrated area of butan-1-01
[butan-1-01] =
integrated area of (ester + butan-1-01)
X [butan-l-olIinitial
The extent of synthesis can be expressed as:
[ester]
% esterification = x 100
[butan-l-olIinitial
From 'H spectra, the integrated area of ester is that of
the methylene "'CH2 peak and the integrated area of
butan-1-01 is that of the methylene 'CH2 peak.
From 13Cspectra, the ester amount was assessed with
the same equation using the area of the ester signal
(01'CH2) and the area of the butan-1-01 signal ('CH,).
Different mixtures of known concentrations were
NMR analyzed under the same analytical conditions.
The quantification was in agreement with true values:
NOE and relaxation effects were similar on carbon
atoms used.
637
RESULTS
Peak Assignment of NMR Spectra
Figure 1 presents the spectra of an equimolar mixture
of butan-1-01and caprylic acid obtained by 'H (Fig. la)
and 13C (Fig. lb) NMR. The values of the chemicals
shifts and their assignments are presented for pure
butan-1-01, pure caprylic acid, their equimolar mix-
ture, and pure ester n-butyl caprylate in Table I, in
agreement with literature data (Bretmaier and Voelter,
1987). It can be observed (Fig. la) that a single signal,
at 8.00 ppm, is obtained for all the hydroxylic hydrogens
of acid, alcohol, and water initially present in the mix-
ture: this is due to the fast exchange of hydroxylic hydro-
gens between them. The chemical shift of this peak is
not the average value of caprylic acid and butan-1-01
chemical shifts, because the water added to the equimo-
lar mixture to reach an a , of 0.35 participates in the
average chemical shift of this peak. Table I also shows
that the 'H signal for the *CH2 of the ester overlaps
with the 'H signal for the 2CH2of the acid.
The chemical shift to the carbonyl peak of ca-
prylic acid switches from 180.65 ppm for acid alone to
178.34 ppm when the acid is mixed with butan-1-01 (Ta-
ble I). It is suggested that addition of butan-1-01 in
caprylic acid modifies the dimerization balance of ca-
I I
12 0 10 0
I ' I ' 1 "
180 160 140
Figure 1. NMR spectra of equimolar mixtures of caprylic acid and butan-1-01: (a) 'H, and
(b) 13C.
638 BIOTECHNOLOGY AND BIOENGINEERING, VOL. 51, NO. 6, SEPTEMBER 20, 1996
prylic acid to the benefit of alcohol-acid association.
This fact also explains the small variations observed for
'CH2 and 2CH2signals of butan-1-01 in the mixture.
'H as well as I3C spectra show peaks with distinct
resonances for the alcohol (lCH2) and the ester (*'CH2).
This will allow for the monitoring of alcohol consump-
tion and ester production during esterification reactions.
Kinetic Study by 'H NMR Spectroscopy
The proton spectra evolution during the esterification
reaction is shown on Figure 2. The spectrum at the
bottom (reaction medim at time 0) is identical to the
spectrum of equimolar mixture of substrates without
enzyme (Fig. la). The peaks below 3 ppm present too
much overlapping, preventing their use for quantifica-
tion. However, the overlapping of 'CH2 of caprylic acid
and "CH2of the ester gives a single signal. This signal can
be used to quantify acid and ester together. Therefore,
Figure 2 represents only the spectra above 3 ppm, and
allows the monitoring of ester production with alco-
hol consumption.
An upfield shift of the peak of exchangeable hydrox-
ylic hydrogens of acid, alcohol, and initial water is ob-
served with time, due to the consumption of the two
substrates and the production of water by the reaction.
In fact, as the reaction proceeds, when one molecule of
I I 1 I 1
so 60 40 20 00
PPM
' ~ " ' I " ' I ' I
120 100 80 60 40 20 0
PPM
 Table I. Chemical shifts (ppm) and peak attribution of 'H and I3CNMR spectra of butanol,
caprylic acid, equimolar mixture of butan-1-01 and caprylic acid, and n-butyl caprylate.
Equimolar n-Butyl
Peak assignment Butan-1-01 Caprylic acid mixture caprylate

'H
OH (acid + alcohol) 5.19 12.39 8.00
'CHI 3.55 3.58
"'CH2 4.00
'CH2 2.28 2.24
"CH2 2.20
TH2 1.58 1.56
'CH2 1.52 1.52
%H2 1.37 1.31 1.55
"CH2 1.55
"CH2 1.35
45,6,7~~ 1.29 1.28
4CH3 0.92 0.90
"CH3 0.89 0.88 0.91
"'CH3 0.88

13C
co 180.65 178.34 172.53
"'CH2 63.79
'CH2 61.96 62.19
'CH2 35.31 34.92
'CH2 34.57 34.50
"CH2 34.38
TH2 32.32 32.27 32.27
"CH2 31.41
4 ~ ~ 2 29.67 29.63 29.64
%H2 29.60 29.55 29.54
TH2 25.24 25.37 25.46
'CH2 23.16 23.12 23.10
3'C~2 19.64
3CH2 19.53 19.41
"CH3 14.26 14.24 14.24
4'CH3 13.90
4CH3 14.10 14.04

alcohol and one molecule of acid are consumed, one

molecule of water is produced. Therefore, one hydrox-
ylic hydrogen of butan-1-01 (5.19 ppm) and one of ca-
OH prylic acid (12.39 ppm) are replaced by the two protons
8.6 h
L of one water molecule (4.80 pprn). It is thus expected
that the chemical shift of this hydroxylic hydrogen peak
decreases from the initial value of 8.00 ppm to 6.50 ppm
at the end of the reaction.
A second important observation arising from these
4.0 h h spectra is the appearance of one new broad signal at
4.80 ppm. This peak, very faint at 1.6 h, is confirmed
for later times. The peak picking for the peak centered
1.6h --Jx on 4.80 ppm is possible only after 1.6 h of reaction
(esterification yield 20%). The intensity of this peak
increases with time. This peak represents water as a
Time 0 distinct phase, consisting in droplets, nonmiscible with
r - , I . 1 * I . I . I
the organic phase. Its chemical shift has been compared
8.0 7.0 6.0 5.0 4.0 3.0
with that of pure water. Since it remains constant for
PPM the whole reaction, it represents water alone and the
Figure 2. Proton spectra evolution as a function of time for an equi-
area of the signal can be used to quantify the amount
molar mixture of caprylic acid and butan-1-01 (300 pL) and 6 mg of of water not solubilized in the organic phase.
cutinase at 30°C. Figure 3 represents the time course of the concentra-

SARAZIN ET AL.: NMR STUDIES OF ESTERlFlCATlON CATALYZED BY CUTINASE 639

 91
?-J
.-c0r 5
y 4
cr
c produced water does not participate in that peak. This
a 3
0
E 2
0 1
0 starts and ester and water are produced (esterification
0 1 2 3 4 5 6 7
Time ( h )
Figure 3. Concentrations ( M ) as a function of time during esterifica-
tion of an equimolar mixture of caprylic acid and butan-1-01 (300 gL)
catalyzed by 6 mg of cutinase at 30°C measured on 'Hspectra. Butyl
caprylate (0);butan-1-01 (0);acid + ester signal (A); nonmiscible
water (0); exchangeable hydroxylic hydrogens (m).
tions obtained through 'H NMR: butan-1-01; n-butyl
caprylate; acid + ester; nonmiscible water; and ex-
changeable hydroxylic hydrogens. Butan-1-01 and n-
butyl caprylate concentrations are provided by the inte-
gration of their methylene triplet peaks at 3.58 ppm and
4.00 ppm, respectively. The overlap of the methylene
proton signals close to 2.20 ppm of the acid and ester
(Table I) provides a constant integration of this signal,
confirming that the acid consumption is equivalent to
the ester production and therefore proving the accuracy
of the quantification method. The symmetry between
the alcohol concentration decrease and the ester con-
centration increase is related to the same facts. The
curve for nonmiscible water corresponds to the integra-
tion of the peak at 4.80 ppm. The curve, with an initial
value of 8 M, does not represent the concentration of
any one species, but instead represents the concentra-
tion of exchangeable hydroxylic hydrogens of acid, alco-
hol, and miscible water (initially present in the reaction
medium and produced by the esterification reaction).
The initial concentration of each substrate is 3.75 M and
the initial water concentration in the organic phase (1%
w/w) is 0.45 M: the corresponding initial concentration
in hydroxylic hydrogens is thus 8.40 M. Up to 1 h of
reaction, there is no reaction and the addition of all
exchangeable hydrogens remains constant. As soon as
water appears as a distinct phase (1.6 h), this concentra-
tion decreases with a slope twice that of butan-1-01, due
to the consumption of both acid and alcohol.
Chemical Shift Study by 'H NMR Spectroscopy
Chemical shifts of the hydroxylic hydrogens are repre-
sented as a function of ester concentration in the reac-
640 BIOTECHNOLOGY AND BIOENGINEERING, VOL. 51, NO. 6, SEPTEMBER 20. 1996
1 tion medium (Fig. 4). As indicated previously, two sig-
nals characterize hydroxylic hydrogens in the reaction
medium. At the beginning of the reaction, one single
signal is observed, corresponding to the hydrogen atoms
of hydroxylic groups of acid and alcohol as well as those
of the water soluble in the organic phase. Up to 2 h,
the chemical shift remains constant, indicating that the
acid and alcohol remain in equimolar amounts and that
makes sense for the first hour of reaction, because no
synthesis occurs (Fig. 3). This lag period is due to en-
zyme and substrate water activities which are not at
optimum. However, between 1 and 2 h, the reaction
6 9
yield 28%).This steady chemical shift indicates that the
water produced does not participate in the chemical
shift. Water produced remains with the enzyme or con-
stitutes droplets. After 2 h, the chemical shift decreases
with time, up to 6 h.
A shielding effect on the chemical shift of the hydrox-
ylic hydrogen peak is observed for increasing times; i.e.,
increasing amounts of ester in the acid-alcohol equimo-
lar mixture. In agreement with preliminary experiments
carried out on acetic acid-ethyl acetate mixtures (data
not shown), this chemical shift evolution is due to the
increase in the proportion of acid complexed with the
ester. The same evolution of the chemical shift was
observed in presence of ethanol in acetic acid-ethyl
acetate mixtures (equimolar mixtures of acid, alcohol,
and increasing amounts of ester). This hypothesis will
be confirmed with the 13C analyses of chemical shifts.
After 6 h, no evolution of the chemical shift is observed:
this is in agreement with the kinetic study (Fig. 3 ) , which
shows that ester oroduction levels off at 6 h.
A 8.5 I
E
a 8.0
Q
W
7.5
cr
0
-.__
I..
......
.-u 6.0
"I
a 5.5
II:
7 . 1
0 10 20 30 40 50 60 70 80 90 100
Es t e r i f i c a t i o n (%)
Figure 4. Chemical shifts of exchangeable hydroxylic hydrogens as
a function of the esterification yield. Reaction media (-) are com-
pared to reconstituted media without enzyme (------).Hydroxylic hy-
drogens in the organic phase (m), and hydroxylic hydrogens in the
aqueous phase (0).The aqueous phase in the reconstituted media
appears only for the 70% esterification yield (m).
 The chemical shift of the peak at 4.80 ppm, observed
after 1.6 h of reaction, remains constant until the end
of the reaction. The shape and chemical shift of this peak
are characteristic of water as a distinct phase, neither
miscible with the organic phase nor exchanging any
hydrogen with it.
The evolution of the chemical shift of exchangeable
hydroxylic hydrogens has been compared to the chemi-
cal shift obtained in mixtures of given compositions in
acid, alcohol, water, and ester, simulating reaction me-
dia at different times (Fig. 4). The organic phase in the
reaction medium solubilizes less water than the organic
phase of the medium with the same composition, but
containing no enzyme. Indeed, the chemical shift of
exchangeable hydroxylic hydrogens decreases faster
than in the reaction medium, probably due to a larger
contribution of water.
Confirming this hypothesis, the nonmiscible water in
reconstituted media without enzyme is detected as a
distinct phase for an ester yield of 70%. In the reaction
medium, this distinct aqueous phase appears at about
20% esterification. This difference is due to the presence
of cutinase. The presence of cutinase also explains the
steady chemical shift of hydroxylic hydrogens observed
in the reaction medium and not in reconstituted media.
The enzyme retains some of the produced water, which
constitutes the distinct phase detected at 4.80 ppm.
Moreover, the hydroxylic hydrogen signals for the
aqueous distinct phase of the reaction medium and the
reconstituted media show different shapes. In the latter,
the signal is a sharp peak with all ester concentrations,
whereas the signal shape changes during the reaction.
This change in shape might be attributed to the presence
of cutinase, which could slow down the rate of hydrogen
exchange. Once esterification is completed, the peak
becomes sharp again.
Kinetic Study by 13C NMR Spectroscopy
13Cspectra are obtained under the same reaction condi-
tions as previously described for proton spectra. Figure
5 shows the spectrum of the medium at initial time and
its evolution with time, on the one hand for carbonyl
signals (Fig. 5a) and, on the other hand, for methylene
signals (Fig. 5b).
During the reaction, the signal intensity of the car-
bony1 of the acid decreases and an upfield shift of the
peak from 178.34 ppm to 177.14 pprn is observed.
Around 173 ppm, which corresponds to the carbonyl of
the ester, two peaks, increasing with time, are observed.
The chemical shift of the upfield signal remains constant
during the reaction, while the other one shifts in a similar
manner to the acid carbonyl peak.
The same phenomenon is observed in the methylene
region: the intensity of the methylene peak of butan-l-
01 at 62.12 ppm decreases. Two increasing peaks, around
64 ppm, are registered in the ester methylene zone: one
SARAZIN ET AL.: NMR STUDIES OF ESTERlFlCATlON CATALYZED BY CUTINASE
of constant chemical shift at 64.00 ppm, and one with
a chemical shift varying from 64.50 ppm to 64.15 ppm.
To quantify the extent of synthesis using 13Cspectra,
the integrated areas of these two peaks were added, thus
allowing the recovery of the synthesis yield determined
from proton spectra. The evolution of the ester concen-
tration obtained from 13Cspectra is compared with the
evolution calculated from 'H spectra in Figure 6. The
increase of the downfield signal stops at around 20%
of esterification. The peak with constant chemical shift
appears at 1.25 h and increases similarly to the peak
obtained through 'H NMR. The same data analyses are
required for the two carbonyl signals in the 173-ppm
region and a similar behavior is observed.
These results suggest the hypothesis of two different
environments for the produced ester. In mixtures pre-
pared with the same liquid phase composition, but with-
out cutinase, only one peak is observed for the ester,
in the carbonyl region as well as in the methylene region.
Therefore, it can be suggested that the splitting of the
ester peak is related to the presence of cutinase.
Because the splitting of the ester peak is observed
only in the presence of cutinase, two experiments were
designed to force the enzyme to stay either within the
detection coils or outside the detection coils (Fig. 7)
and to study separately the two phases of the system:
i.e., the enzyme solid phase and the organic liquid phase,
respectively. In experiment A, glass beads have been
put above the cutinase powder, to avoid the rising of
the cutinase; this is done to keep the enzyme under the
detection coil region; only the organic phase is detected.
In experiment B, the cutinase is kept in the volume
detected by the receiver coil by putting glass beads at
the bottom of the NMR tubes; cutinase occupies the
whole detected volume. In the first case, the water signal
intensity is weak and 13Cester peaks of constant chemi-
cal shift are not observed. In the second case, the non-
miscible water is detected and the greatest 13C ester
peaks are those of constant chemical shifts. These two
experiments lead to the conclusion that water produced
by the reaction is mainly retained in the cutinase area.
Moreover, they also show that the ester of constant
chemical shift is also localized within the cutinase.
Partition of the ester is only observed during enzy-
matic esterification and when the cutinase is in the detec-
tion coil zone. When cutinase is added to a mixture of
ester and water or mixtures of different compositions
in acid, alcohol, ester, and water, only one kind of ester
is present-the one associated with the acid. The cutin-
ase can only retain the ester synthesized during the
reaction. It is important to note that the ester partition
is linked to the evolution of the distinct aqueous phase.
Chemical Shift Study by 13CNMR Spectroscopy
The chemical shift study is carried out on the carbonyl
signals as opposed to the kinetic study carried out with
641
 a
ACID 1 ESTER

-
1.6 h
-hh-- 2CHzacid -I-aCH2
0.25 h 1

0
1 I I 1 8 1 I
65 6 0 - 55 50 45 40 35

PPM
Figure 5. Carbon spectra evolution as a function of time for an equimolar mixture of caprylic acid
and butan-1-01 (300 pL) and 6 mg of cutinase at 30°C: (a) carbonyl region, and (b) methylene region.

the methylene signals. Methylene signals could have
been used for this study even though variations are
weaker: in fact, in the molecule, methylene groups are
further from the association centers than the carbonyl
groups which participate directly in the association.
However, ‘H NMR of methylene carbons cannot be
used for this study: no effect is observed for methylene
protons, because the effect of associations on densities
of electronic populations are too weak.
Figure 8 shows the evolution of the chemical shift of
the carbonyl signals of ester and acid in reaction media
and in reconstituted media, as a function of esterification
yield. For example, a reconstituted medium at 10% es-
terification yield contains acid and alcohol at a concen-
tration of 3.375 M and ester and water at 0.375 M. The
0% yield corresponds to 3.75 M for acid and alcohol.
The study of the chemical shifts of media containing
increasing equimolar amounts of ester and water and
decreasing equimolar amounts of acid and alcohol, with-
out enzyme, shows an association between the acid and
the ester (Sarazin et al., 1992). Indeed, the upfield shift
of the carbonyl peak of the acid is due to the breakage
of hydrogen bonds of the auto-association of acid and/
or of its association with alcohol, to form association
with the ester.
The chemical shift values of the carbonyl of the ester
are higher than the values for pure ester (172.53 ppm),
because the association with acid induces a downfield

642 BIOTECHNOLOGY AND BIOENGINEERING, VOL. 51, NO. 6, SEPTEMBER 20, 1996
n
w
W 4

c
.-0
4
W
.-+-0 .- 175
.-L t
a,

::g
174
a,
4
cn 1
W 0
0 173
0 - _.
U I
0 1 2 3 4 5 6 7 0 10 20 30 40 50 60 70 80 90 100

Time (h) Es t e rif ic a t io n (%)

Figure 6. Esterification as a function of time of an equimolar mixture
of caprylic acid and butan-1-01(300 pL) catalyzed by 6 mg of cutinase.
Ester yields are measured on 13C spectra. CHI of variable chemical
shift (H); CH2 at 64 ppm (0);sum of the two CH2 peaks (0). The
results are compared with the ester yield previously obtained from
'H spectra (0).
Figure 8. Chemical shifts of carbonyl carbons as a function of the
esterification yield. Reaction media (-) are compared with recon-
stituted media without enzyme (------).Ester carbonyl carbon of vari-
able chemical shift (H); ester carbonyl carbon of constant chemical
shift (0);acid carbonyl carbon (0).

carbonyl carbons are related to an increase in the molar

shift. The first detectable carbonyl signal of the ester is
registered at 174.33 ppm. For low molar ratios of ester/
acid, the ester is fully associated with the acid; the elec-
tron density of the carbonyl carbon is lower than for
the pure ester, thus explaining the observed deshielding
effect. We have previously shown that complexing ace-
tophenone to phenol leads to a deshielding of the car-
bony1 carbon (SCguin et al., 1989). As the content of
ester increases, the proportion of ester associated with
the acid decreases. The ester chemical shift value de-
creases, moving closer to the chemical shift value of
pure ester. In agreement with preliminary experiments
carried out on acetic acid-ethanol-ethyl acetate mix-
tures, the chemical shift decreases of acid and ester
ratio of ethyl acetate.
The same phenomenon is observed in reaction mix-
tures (Fig. 8). However, the slope of the graphs is weaker
for reaction media than for reconstituted media: for
example, as far as the carbonyl signal of the acid is
concerned, a chemical shift of 177.75 pprn is obtained
with a reconstituted medium corresponding to 42% es-
terification, whereas 70% esterification is required in the
reaction medium. It seems that the acid of the reaction
medium is not completely in contact with the ester pro-
duced by the reaction. This observation confirms that
the ester produced is not fully recovered in the organic
phase. This is in agreement with the presence of a second
ester peak at 173 ppm. Moreover, the same behavior is

I I ' " L A A
equimolar
I equimolar
substrates I
COILS substrates 4.8 ppm COILS 300pL 4.8 ppm
I

Glass
beads

t
173 ppm 173 ppm

Figure 7. Variation of 'H and 13C spectra by changing the detection zone. (a) Experiment A The cutinase is kept under the detection coil
region; a weak water signal is recorded on proton spectra; and a single peak is observed for the ester carbonyl carbon. (b) Experiment B: The
cutinase is in the detection zone; the second ester carbonyl form, at 173 ppm, is observed and the nonmiscible water signal is important.

SARAZIN ET AL.: NMR STUDIES OF ESTERlFlCATlON CATALYZED BY CUTINASE 643

observed for the carbonyl signal of the ester. A chemical
shift of 174 ppm corresponds to an esterification yield
of 48% in reconstituted media and of 72% in reaction
media. Part of the ester of the reaction medium seems
to be “hidden” and not in contact with the acid.
Conclusion
A new method, based on NMR, has been developed
for use as a quantitative tool in monitoring esterification
reactions of fatty acids and alcohols. The technique has
been shown to be quick, easy, and reproducible. Fur-
thermore, considerable time may be saved over conven-
tional analyses in that the reaction can be monitored
without the need for sampling the reaction medium.
NMR studies bring forth some new information as
compared to the more classical means of investigation.
As far as water is concerned, ‘H NMR indicates that
the water produced at the beginning of the reaction is
retained by the cutinase. Later during the reaction, this
synthesized water forms a distinct aqueous phase. It
should be noted that the same kinetic evolution has been
observed using GC analysis, suggesting that agitation in
the NMR spectrometer is the same as in a “classical”
tube. Regarding the ester, 13Cstudies show the presence
of two ester forms. The chemical shift values, correlated
with changes in associations between substrates and
products, show the existence of an acid-ester complex
(‘H and 13C)as well as the existence of an ester propor-
tion not accessible to the substrates (”C).
In the future, the stability of the cutinase could also
be explored through NMR, especially in terms of water
amounts and mobility. A similar study has already been
published using 1 7 0 NMR to relate the stability of P-
galactosidase and water mobility (Yoshioka et al., 1993).
The authors thank Sally Percy for linguistic advice.
References
Bhamidipati, S. P., Hamilton, J. A. 1993. NMR studies of phospholi-
pase C hydrolysis of phosphatidylcholine in model membranes.
J. Biol. Chem. 268.2431-2434.
Bretmaier, E., Voelter, W. 1987. Carbon-13 NMR spectroscopy. 3rd
edition. VCH Publishers, New York.
Brosio, E., Delfini, M., Veglia, G., Pizzoferrato, L. 1994. Kinetics of
the phosphate groups enzymatic hydrolysisin myo-inositol hexakis-
phosphate studied by 31P spectroscopy. Q. Magn. Res. Biol. Med.
1135-140.
Burke, P. A., Griffin, R. G., Klibanov, A.M. 1992. Solid-state NMR
assessment of enzyme active center structure under non aqueous
conditions. J. Biol. Chem. 267 20057-20064.
644 BIOTECHNOLOGY AND BIOENGINEERING, VOL. 51, NO. 6, SEPTEMBER 20, 1996
Decagny, B., Ergan, F., Sarazin, C., Barbotin, J. N., SBguin, J. P. 1996.
Water activity by ‘H NMR spectroscopy: application to the study
of water exchange in biphasic media. Anal. Biochem. 234:142-148.
Dordick, J. S. 1989. Enzymatic catalysis in monophasic organic sol-
vents. Enzyme Microb. Technol. 11 194-211.
Gupta, N. M. 1992. Enzyme function in organic solvents. Eur. J. Bio-
chem. 203 25-32.
Halling, P. J. 1994. Thermodynamic predictions for biocatalysis in
nonconventional media: theory, tests and recommendations for ex-
perimental design and analysis. Enzyme Microb. Technol. 16
178-206.
Kanerva, L. T., Vihanto, J., Halme, M. H., Loponen, J. M., Euranto,
E. K. 1990. Solvent effects in lipase-catalyzed transesterification
reactions. Acta Chem. Scand. 44: 1032-1035.
Kitaguchi, H., Itoh, I., Ono, M. 1990. Effects of water and water-
mimicking solvents’on the lipase catalyzed esterificationin an apolar
solvent. Chem. Lett. 7 1203-1206.
Klibanov, A. M. 1989. Enzymatic catalysis in anhydrous organic sol-
vents. Trends Biochem. Sci. 14: 141-144.
Martinez, C., De Geus, P., Lauwereys, M., Matthyssens, G., Cambillau,
C. 1992. Fusarium solani cutinase lipolytic enzyme with a catalytic
serine accessible to solvent. Nature 356 615-618.
Mendz, G. L., Lim, T. N., Hazell, S. L. 1993. Fluorinated probes to
measure carboxylesterase activities using ’% NMR spectroscopy:
application to erythrocytes and Helicobacterpylori. Arch. Biochem.
Biophys. 305: 252-260.
O’Connor, C. J., Petricevic, S. F., Coddington, J. M., Stanley, R. A.
1992. An NMR assay for quantitating lipase activity in biphasic
macroemulsions. J. Am. Oil Chem. SOC.69: 295-300.
Parker, M. C., Moore, B. D., Blacker, A. J. 1992. Measuring enzyme
hydration in nonpolar organic solvents using NMR. Biotechnol.
Bioeng. 46:452-458.
Pollesello, P., Toffanin, R., Eriksson, O., Kilpelainen, I., Hynninen,
P. H., Paoletti, S., Leo Saris, N.-E. 1993. Analysis in crude extracts
by 13Cnuclear magnetic resonance. Anal. Biochem. 214 238-244.
Rabiller, C., MazC, F. 1989. Quantitative analysis and determination
of the enantiomeric purity of glycerides by I3C NMR spectroscopy.
Application to the lipase catalysed transesterificationof triacylglyc-
erides. Magn. Reson. Chem. 27 582-584.
Rabiller, C., Heisler, A., Hagele, G. 1992.pp. 283-290. In: J. Tramper,
M. H. Vermiie, H. H. Beeftink, and U. von Stokar (eds.), vol. 8.
Elsevier, Amsterdam.
Roberts, M. F. 1990. Nuclear magnetic resonance spectroscopy to
follow phospholipase kinetics and products. Meth. Enzymol. 197:
31-48.
Sarazin, C., Goethals, G., SCguin, J. P., Legoy, M. D., Barbotin, J. N.
1992.Etude par spectroscopieRMN ‘H et 13Cde la synthkse d’esters
en milieu organique catalysCe par une lipase: influence de l’eau.
J. Chim. Phys. 8 9 541-547.
SCguin, J. P., Goethals, G., Moreau-Descoings, M. C., Sarazin, C.,
Doucet, J. P. 1989.Recherche des sites d’associationprivilBgiCs dans
des moltcules polyfonctionnelles par RMN du carbone 13. Bull.
SOC.Chim. Belg. 98: 547-555.
Yoshioka, S., Aso, Y., Izutsu, K.-I., Terao, T. 1993. Stability of 0-
galactosidase, a model protein drug, is related to water mobility as
measured by 1 7 0 nuclear magnetic resonance (NMR). Pharm. Res.
10: 103-108.
Zaks, A., Klibanov, A. M. 1988. The effect of water on enzyme action
in organic media. J. Biol. Chem. 263 8017-8021.