Mechanistic studies on the enzymatic transesterification of polyesters

Article abstract of DOI:10.1016/j.tet.2003.10.104

The effect of solvents on the enzymatic transesterification of aliphatic polyesters has been investigated. It has been shown that the hydrophobicity and the polarity of the solvent have little effect on the reaction, however, the molecular weight of the product depends on the solubility of the product in the medium. Above a certain molecular weight, when the product is no longer soluble in the medium, transesterification does not occur. Deuterium NMR studies have shown that transesterification tends to take place at the ends of the polymer chain rather than at random along the polymer chain.

Full text of DOI:10.1016/j.tet.2003.10.104

Tetrahedron 60 (2004) 765–770
Mechanistic studies on the enzymatic transesterification
of polyesters
*
R. W. McCabe and A. Taylor
Centre for Materials Science, University of Central Lancashire, Preston PR1 2HE, UK
Received 11 July 2003; revised 13 August 2003; accepted 17 October 2003
Abstract—The effect of solvents on the enzymatic transesterification of aliphatic polyesters has been investigated. It has been shown that the
hydrophobicity and the polarity of the solvent have little effect on the reaction, however, the molecular weight of the product depends on the
solubility of the product in the medium. Above a certain molecular weight, when the product is no longer soluble in the medium,
transesterification does not occur. Deuterium NMR studies have shown that transesterification tends to take place at the ends of the polymer
chain rather than at random along the polymer chain.
q 2003 Elsevier Ltd. All rights reserved.
1. Introduction
the presence of toluene then not only was the dispersity
higher but that the molecular weight obtainable was
reduced. The conclusion being that transesterification
occurs in the presence of toluene as solvent causing scission
of the growing polymer. This conclusion was confirmed
when the synthon, 1,6-hexanedioic acid di(4-hydroxybutyl)
ester (BAB) was shown not to react with itself in the
absence of solvent, whereas in toluene the enzyme catalysed
the transesterification and a number of higher molecular
weight oligomers were formed.7
In recent years the enzymatic synthesis of polyesters has
aroused considerable interest because of the possibility of
producing novel polyesters with unusual properties. The
early work was done using activated acids and or diols, such
as divinyl adipate and halogenated alcohols in non-aqueous
solvents,1 the resulting polyesters were oligomers2 of rela-
tively low molecular weight of approximately 2000 Da. It
was shown subsequently that using the enzyme Candida
antarctica lipase B supported on acrylic beads (Novozyme
435TM) it was possible to synthesise much higher molecular
weight polyesters in the absence of solvent.3 A number of
workers have also studied the enzymatic catalysis of
polyesters by the transesterification of both activated and
unactivated diesters with diols4 and by the ring opening
transesterification of lactones.5
Lipase catalysed transesterification reactions in organic
media with monoesters are well known and have been used
to separate racemic mixtures of alcohols and carboxylic
acids or to select a specific ester or alcohol group within a
molecule as substrate.8 Therefore, the original question of
why the enzyme appears to catalyse the transesterification of
polyesters in some circumstances and not in others becomes
more interesting, because one would expect it to catalyse
transesterification in all circumstances. It had been con-
sidered initially that the enzyme might adopt a different
configuration in solvent thereby affecting the reaction and
the resulting products. However, it has been shown, using
synchrotron CD spectroscopy that no major changes in the
structure of the protein take place when used in solvents
such as hexane and toluene.9 The possibility that solvent
molecules were being absorbed onto the hydrophobic
regions in or around the active site was considered, as
even one or two bulky solvent molecules such as toluene
absorbed in a critical area could easily affect the rate at
which the substrate could diffuse into the active site of the
enzyme.
When the polyesters prepared by enzymatic synthesis in the
absence of solvent were studied it was found that they had
significantly different properties compared to those prepared
by the conventional high temperature process. It was found
that not only was it possible to obtain higher molecular
weight polymers but they had a lower dispersity, that is a
narrower molecular weight distribution than the conven-
tional polyesters with the result that products made from
these materials have superior physical properties.6 These
results were explained by the apparent absence of a
transesterification reaction during the synthesis.2 It was
subsequently shown that if the reaction was carried out in
Keywords: Transesterification; Polyesters; Novozyme 435.
*
Corresponding author. E-mail address: ataylor3@uclan.ac.uk;
rwmccabe@uclan.ac.uk
In all our earlier work, the acid-diol esterification reaction
0040–4020/$ - see front matter q 2003 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tet.2003.10.104

766
R. W. McCabe, A. Taylor / Tetrahedron 60 (2004) 765–770
and any transesterification reaction between ester groups,
leading to the formation of the polyester would be taking
place simultaneously, with the final composition containing
the products of both. Therefore, in order to study the
transesterification reaction, it was necessary to develop a
method whereby the transesterification reaction could be
studied independently of the esterification reaction. If the
reaction was being affected by solvent absorbed at the active
site, then the effects of the size, shape and physical
properties of the solvent on the transesterification reaction
were of interest.
The differences between these results are not considered to
be significant. It appears that small additions of solvent do
not affect the degree of transesterification. It is unlikely,
therefore, that absorption of solvent into the hydrophobic
areas of the enzyme takes place, as it would be expected that
the hydrophobic attraction of the lipase would extract
toluene from such a polar medium as 1,4-butanediol, even at
these low concentrations.
The effect of the shape of the solvent molecule was then
investigated. The standard transesterification experiment
was carried out with a number of solvents of different shapes
and hydrophobicity. The results are as shown in Table 3
below.
2. Results
Table 3. Effect of solvent configuration on transesterification
2.1. Transesterification reactions
Solvent
Mw
Mn
Dispersity
C log P
The hydroxyl group of the added diol involved in the
transesterification reaction causes scission of the high
molecular weight polyester, the extent of transesterification
being proportional to the reduction in the average molecular
weight of the polyester. The results obtained are shown in
Table 1.
Toluene control
n-Butylbenzene
iso-Butylbenzene
tert-Butylbenzene
4-Chlorotoluene
Hexane
5800
5200
5900
5900
5900
18,000
2900
2800
3000
2900
2900
8600
2.0
1.9
2.0
2.0
2.0
2.0
2.7
4.0
4.0
4.1
3.3
3.8
Table 1. Effect of solvent on the transesterification of polyester
The GPC profiles for all the experiments listed in Table 3
were essentially the same for all the aromatic solvents.
Solvent
Mw
Mn
Dispersity
PTMEG 650
1,4-butanediol
Toluene
36,000
31,000
5800
32,000
26,000
2900
1.1
1.2
2.0
1.8
It is evident from these results that the degree of
transesterification as measured by chain scission is not
affected by the geometry of the solvent molecule. The
differences seen in the experiments with the aromatic
solvents are not considered significant; the only significant
difference being between the aromatic solvents and the
aliphatic hexane. It does not appear, however, that the
hydrophobicity as measured by C log P is the cause of
the difference; as there is significant transesterification at a
C log P of 20.4 in dioxane and also at C log P of 4.1 in tert-
butylbenzene.
Dioxane
1400
750
It appears from these results that the amount of transesterifica-
tion is independent of hydroxyl concentration. The greatest
breakdown of the high molecular weight polyester occurred
with toluene and dioxane and not with 1,4-butanediol or
polytetramethylene ether glycol of molecular weight 650
(PTMEG 650), which have much higher concentrations of
hydroxyl groups. This observation appears to contradict the
conclusion of Kumar and Gross,10 who proposed that there is
slower transesterification in higher molecular weight poly-
caprolactones because the higher the molecular weight, the
fewer terminal hydroxyls being present to take part in the
transesterification reaction.
The transesterification reaction was repeated using solvents
that were substantially more polar in order to determine if
the reaction is affected by the polarity of the solvent.
Propylene carbonate, triethylene glycol methyl ether and
tetraethylene glycol methyl ether were chosen as the added
solvent. In all cases extensive transesterification took
place and the molecular weight fell to around 3000 Da as
measured by GPC. There did not appear to be any signi-
ficant difference in the amount of transesterification in
relation to the polarity of the solvent.
It had been found by Harffey11 that the addition of as little as
6% w/w of toluene to the reaction medium gave the same
result as when the reaction was carried out in toluene as the
reaction solvent. The possibility that toluene was being
absorbed from the medium onto the hydrophobic areas of
the protein, thereby affecting the mechanism of the reaction,
was considered. Therefore, the effect of low concentrations
of toluene in 1,4-butanediol on the degree of transesterifica-
tion was determined. The results are shown in Table 2.
The only two media in which transesterification does not
occur are 1,4-butanediol and polytetramethylene ether
glycol both of these are very poor solvents for the high
molecular weight polyester. It would appear that the
transesterification reaction takes place in any solvent in
which the higher molecular weight polyester is soluble.
Table 2. Effect of toluene concentration on transesterification
Principal solvent
[Toluene] (mM)
Mw
Mn
Dispersity
2.2. NMR studies
1,4-BD
1,4-BD
1,4-BD
1,4-BD
0.11 (0.001%)
1.1 (0.01%)
67 (0.625%)
200 (1.87%)
26,000
28,000
27,000
28,000
14,000
19,000
17,000
18,000
1.9
1.5
1.6
1.6
In order to elucidate the mechanism, the transesterification
reaction was studied using deuterium NMR. The insertion of
deuterated 1,4-butanediol into the polyester being followed

R. W. McCabe, A. Taylor / Tetrahedron 60 (2004) 765–770
767
Figure 1. 2H NMR spectrum of deuterated 1,4-butanediol.
by observing the downfield shift due to the presence of
adjacent ester carbonyl groups.
4.18 ppm. The C2H2–C2H2 peak had also moved downfield
to 2.33 ppm, which proved that there had been significant
incorporation of the deuterated diol into the polyester.
A 1% solution of d8-1,4-butanediol in 1,4-butanediol was
2
prepared and the H NMR spectrum obtained (Fig. 1).
The initial experiments without added toluene were
repeated using a polyhexane adipate polyester of 2000 Da.
The sample taken at 24 h showed that the HO–C2H2 peak at
4.11 ppm had split equally with the COO–C2H2 peak at
4.68 ppm. Similarly, the diol C2H2–C2H2 peak at 2.21 ppm
had also split equally with the polyester C2H2–C2H2 peak at
2.12 ppm. This showed that after 24 h, significant chain
scission caused by transesterification with the added diol
had taken place and after 48 h even more transesterification
was found.
The two peaks of the deuterated diol were at 4.11 ppm,
corresponding to the 1,10- and 0 4,40-deuteriums and
2.14 ppm, corresponding to the 2,2 - and 3,30-deuteriums.
The signal-to-noise ratio was 150:1, therefore we were
confident that using this method we would be able to see
whether transesterification had taken place, by looking for
the insertion of the deuterated 1,4-butanediol into the
polyester.
In order to maximise the visibility of the deuterated diol in
the polyester, the concentration of the deuterated diol was
increased to 10%. Samples were taken at 2, 5 and 21 h and
the 2H NMR spectra obtained, no difference could be seen in
any of the spectra, thus indicating that in the 1,4-butanediol
medium no observable transesterification had taken place.
After 24 h a very small peak at 4.68 ppm was starting to
show, indicating that a small amount of deuterated diol had
been incorporated into the polyester. After 48 h a somewhat
larger peak was observed at 4.68 ppm, which indicated that
transesterification does take place in 1,4-butanediol, but that
it is very slow.
The polyester with the lowest and most uniform molecular
weight is the simple oligomer 1,6-hexanedioic acid di(4-
hydroxybutyl) ester (BAB). In order to see if this was also
susceptible to transesterification in the absence of solvent, a
sample of BAB was prepared starting from 6-carboxy-11-
hydroxy-7-oxaundecanoic acid (AB) synthesised by the
method of Harffey.11 AB was reacted with a two-fold molar
excess of 1,4-butanediol using Novozyme 435 as catalyst.
After 24 h 1H NMR spectroscopy showed that 100%
conversion to BAB had taken place. The enzyme and
solvent were removed and 10% deuterated diol and
Novozyme 435 were added and the mixture heated at
60 8C for 24 h. 2H NMR spectroscopy on the purified
sample showed a very small, deuterated ester peak at
4.68 ppm, this indicated that some but not very much
transesterification had occurred.
This experiment was repeated after adding 6 ml of toluene
2
and stirring at 60 8C a H NMR spectrum with excellent
signal-to-noise ratio (2000 scans) was obtained. After 24 h,
there was a small but distinct peak at 4.68 ppm due to the
increased chemical shift when one end of the diol is
incorporated into an ester group and the peak at 2.12 ppm
was starting to split, with a pronounced shoulder at
2.21 ppm, as a result of the d8-1,4-butanediol now forming
a significant part of the ester groups. This indicated that the
d8-1,4-butanediol had been incorporated into the polyester
and that transesterification had occurred (see Fig. 2A).
Samples from the experiments above were analysed by GPC
to see if the effect of the transesterification could be
determined. Most interestingly, the GPC of the high
molecular weight polyester transesterified with deuterated
1,4-butanediol in the absence of toluene showed that the
peak molecular weight, Mw 36,000, had declined very little
but a number of low molecular weight oligomers had
appeared (Fig. 3), whereas the 2000 Mw polyester had
reduced to 980 Mw in the same time.
After 48 h it could be seen that both of the deuterium
resonances had split and the new peaks had moved
downfield (see Fig. 2B). The OC2H2 peak had moved in
its entirety to 4.73 ppm, leaving only a small peak at
This result shows that the scission is not taking place at
random along the polyester backbone, but rather that it takes

768
R. W. McCabe, A. Taylor / Tetrahedron 60 (2004) 765–770
2
2
Figure 2. H NMR spectrum of polyesterþ H8-1,4-butanediolþtoluene A after 24 h and B after 48 h.
place at the ester groups near to the end of the polymer
chain. This is unlikely to be due to any property of the
enzyme because each ester group and its environs are
identical and any could fit into the pocket of the enzyme. It
is more likely to be a property of the polyester. It is possible
that it forms a tight coil in these media and it is only the ends
of the chain that are available to the enzyme.
It appears from these results, that the rate of transesterifica-
tion reaction, measured by the amount of insertion of
2H8-1,4-butanediol into the polyester, whilst occurring in
1,4-butanediol is very much faster in toluene. The amount of
insertion into the high molecular weight polyester and the
2000 molecular weight polyester appear to be approxi-
mately the same, however, the results of the transesterifica-
tion are quite different. In the case of the high molecular
weight polyester there is nominal reduction in the molecular
weight as the chain scission only occurs at the ends of the
molecule. In the 2000 Da polyester the chain scission occurs
at random along the chain with the result that there is a
significant reduction in molecular weight.
The transesterification of the oligomer BAB appears to be
quite rapid, however, BAB cannot polymerise in the
presence of excess diol, therefore the deuterated diol can
only react at one of its two hydroxyls and this is reflected in
the relative size of the peak shifted downfield.
Figure 3. High molecular weight polyester after transesterification with
deuterated 1,4-butanediol.
The conclusion from these experiments seemed to be that

R. W. McCabe, A. Taylor / Tetrahedron 60 (2004) 765–770
769
the reason for the lack of transesterification in the absence of
solvent is the insolubility of the high molecular weight
polyester in 1,4-butanediol or PTMEG 650. It is proposed
that in the absence of solvent the growing polymer chain is
held in the proximity of the enzyme active site by hydrogen
bonding to the surface of the enzyme, a mechanism known
in the synthesis of biopolymers such as RNA.12 It is only
when the polymer is in solution that the visceral ester groups
are available to the enzyme for transesterification to take
place. It might be that as the polyester reaches a critical
molecular weight it starts to drop out of solution in the diol,
thus limiting the transesterification reaction.
stopped and the acid number and hydroxyl number
determined by titration. The bound enzyme was filtered
off and the residual enzyme deactivated by heating the
polyester at 200 8C for 15 min. This polyester had a Mn of
17,500, a Mw of 37,000 and dispersity 1.9. The acid number
was measured as 2 mg KOH g21 and the hydroxyl number
was 13 mg KOH g21. This polyester sample became the
standard for all our subsequent transesterification experi-
ments. The molecular weights for this and all subsequent
reaction products were determined by using a Waters HPLC
with a model 510 pump, model 410 refractive index detector
and the Waters 717 autosampler. The column was packed
˚
with Polymer Labs. 1000 A polystyrene copolymer packing
Modelling using Sybyl 6.0 and Sculpt 3.0 produced a very
similar result, in both cases the growing polyester partially
wrapped around the surface of the enzyme, bound by
multiple hydrogen bonds to the surface of the enzyme. The
non-bound part of the polymer chain adopted a tightly
coiled configuration held together by intramolecular hydro-
gen bonds. If this is an accurate simulation then in a medium
in which the polyester is not soluble it is only at or near to
the ends of the molecule that the ester groups are available
for transesterification. This would explain the fact that in the
absence of solvent much higher molecular weight products
are obtained, there being no transesterification to cause
scission of the growing polymer.
and the eluent used was THF stabilised with 250 ppm
of butylated hydroxytoluene (BHT) at a flow rate of
1 ml min21. The sample concentration for all experiments
was 0.5% w/vol with an injection volume of 40 ml. The data
were analysed using the Millennium 32 GPC software. The
GPC trace for the polyhexane adipate polyester is shown in
Figure 4.
4.2. Procedure for transesterification
In order to study the effects of different solvents on the
transesterification reaction a standard transesterification
reaction was developed.
2 g of high molecular weight polyester plus 8 ml of solvent,
0.1 g of 1,6-hexanediol and 0.1 g of Novozyme 435 were
added to a stirred cell reactor and heated at 60 8C for 24 h.
The reaction was then stopped by filtering off the bound
enzyme and cooling rapidly to 20 8C. In order to simulate
the conditions that had existed in earlier syntheses, the
transesterification was carried out in the presence of the
solvents toluene and dioxane in the presence of 1,4-
butanediol and polytetramethylene ether glycol (PTMEG)
650 to simulate the excess monomer and oligomer that is
present in the early stages of the solvent free polymerisation.
n-butylbenzene, iso-butylbenzene, tert-butylbenzene, 4-
chlorotoluene, hexane, propylene carbonate, triethylenegly-
col methyl ether and tetraethyleneglycol methyl ether were
used to investigate the effect of hydrophobicity, polarity and
size of the solvent molecule on the transesterification
3. Conclusions
Transesterification reactions catalysed by Novozyme 435
have been carried out between high molecular weight
polyester and the diols 1,4-butanediol, 1,6-hexanediol and
PTMEG 650 in the presence of a number of different
solvents. It appears that the transesterification reaction
depends predominantly on the solubility of the polyester in
the medium and that excess hydroxyl content does not
influence the transesterification if the polyester is not
soluble in the diol. The hydrophobicity, polarity and shape
of the solvent molecules do not influence the nature of
the transesterification reaction. It appears that unless the
polyester is soluble in the medium the growth of the
polymer is not limited by the transesterification reaction and
high molecular weight polyester is formed. The limited
amount of transesterification occurs preferentially at the
ends of the polymer rather than at random along the chain.
4. Experimental
4.1. Synthesis of polyhexane adipate polyester
R1-½OCH2ðCH2Þ4CH2OCOðCH2Þ4COOꢀn-R2
A high molecular weight polyhexane adipate polyester was
synthesised using Novozyme 435 as the catalyst. A 1:1 molar
ratio of 1,6-hexanediol and adipic acid was stirred at 60 8C
for 1 h, 0.5% w/w Novozyme 435 added and a vacuum of
100 mb applied for 24 h. The vacuum was increased to
10 mb and the stirring and heating continued for a further
24 h at which point the vacuum was increased to 3 mb for a
further 24 h. When the polyester had reached a molecular
weight of 36,000 as measured by GPC the reaction was
Figure 4. GPC of the high molecular weight polyhexane adipate used as
substrate for transesterification.

770
R. W. McCabe, A. Taylor / Tetrahedron 60 (2004) 765–770
reaction. All materials were purchased from Sigma-Aldrich
and were standard reagent grade purity.
cals Ltd, while the Novozyme 435 was a gift from
Novozyme AS Denmark.
4.3. NMR studies
The toluene and 1,4-butanediol experiments were repeated
using 1,10,2,20,3,30,4,40-octadeutero-1,4-butanediol in place
of the 1,6-hexanediol, as the transesterification agent. A
10% solution of the d8-1,4-butanediol in 1,4-butanediol was
added to 5 g of the polyester, and 0.1 g of Novozyme 435.
The mixture was stirred in a cell reactor at 60 8C and
sampled at 24 and 48 h. Filtering off the Novozyme stopped
the reaction and the residual 1,4-butanediol was stripped off
using a Kugelrohr evaporator prior to observing the NMR
spectra.
References and notes
1. Wallace, J. S.; Morrow, C. J. J. Polym. Sci.: Part A, Polym.
Chem. 1989, 27, 3271–3284.
2. Margolin, A. F.; Crenne, J. Y.; Klibanov, A. M. Tetrahedron
Lett. 1987, 1607–1609.
3. Binns, F.; Harffey, P.; Roberts, S. M.; Taylor, A. J. Polym. Sci.
A: Polym. Chem. 1998, 36, 2069–2080.
4. Chaudhary, A. K.; Lopez, J.; Beckman, E. J.; Russell, A. J.
Biotechnol. Prog. 1997, 13, 318–321.
5. Cordova, A.; Iversen, T.; Hult, K.; Martinelle, M. Polymer
1998, 39(25), 6519–6524.
4.4. Molecular modelling
6. Taylor, A.; Waugh, J. UK Patent WO97/40083, 1997..
7. Binns, F.; Harffey, P.; Roberts, S. M.; Taylor, A. J. Chem.
Soc., Perkin Trans. 1 1999, 2671–2676.
The molecular modelling was done using both Sculpt 3.013
and Sybyl 6.014 on a SGI Octane UNIX work station. The
enzyme was modelled using the pdb file 1tca. A 2000 Da
polybutane adipate polyester was constructed and docked
manually to the Ser 105 residue of the active site. The Sculpt
model was energy minimised using MM3 with both van der
Waals and electrostatic interactions. The Kollman all Atom
forcefield was used for the Sybyl model, which was then
energy minimised using the Powell method within the Sybyl
program.
8. Andersen, E. M.; Larsson, K. M.; Kirk, O. Biocatalysis
Biotransformations 1998, 16, 181–204.
9. McCabe, R. W.; Rodgers, A.; Taylor, A. In press.
10. Kumar, A.; Gross, R. A. J. Am. Chem. Soc. 2000, 122,
11767–11770.
11. Harffey, P. Ph.D. Thesis, University of Liverpool, 1998..
12. Fersht, A. R. Biochemistry 1987, 26, 8031–8036.
13. www.mdli.com/products/sculpt.html.
14. www.tripos.com.
Acknowledgements
This work was supported financially by Baxenden Chemi-