Size selectivity in lipase catalysed tetrol acylation

Article abstract of DOI:10.1016/j.molcatb.2014.08.004

Size selectivity of Candida antarctica lipase B (CAL-B) was examined in the acylation of pentaerythritol with oleic acid. Biolubricant mixtures consisting of mono-, di-, tri-, and tetraoleates were expected in variable excess. Enzymatic tetraoleate format

Full text of DOI:10.1016/j.molcatb.2014.08.004

Journal of Molecular Catalysis B: Enzymatic 109 (2014) 40–46
Contents lists available at ScienceDirect
Journal of Molecular Catalysis B: Enzymatic
journal homepage: www.elsevier.com/locate/molcatb
Size selectivity in lipase catalysed tetrol acylation
Manuel Happe, Martial Kouadio, Christopher Treanor, Jan-Phillip Sawall,
∗
Antoine Fornage, Marc Sugnaux, Fabian Fischer
Institute of Life Technologies, University of Applied Sciences and Arts Western Switzerland, HES-SO Valais-Wallis, Route du Rawyl 64, CH-1950 Sion 2,
Switzerland
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 20 June 2014
Received in revised form 6 August 2014
Accepted 11 August 2014
Available online 19 August 2014
Size selectivity of Candida antarctica lipase B (CAL-B) was examined in the acylation of pentaerythritol
with oleic acid. Biolubricant mixtures consisting of mono-, di-, tri-, and tetraoleates were expected in
variable excess. Enzymatic tetraoleate formation was suppressed under solvent conditions; however, this
size selectivity was lost without solvent and tetra-acylated pentaerythritol accumulated in up to 93%. The
◦
lipase caused size selectivity persisted over a broad temperature range from 35 to 95 C. A Fischer–Speier
esterification showed that substrate bulkiness was only a minor contributor to observed size selectiv-
ity. All in all, switch on/off size selectivity using CAL-B allowed to vary pentaerythritol biolubricant
compositions in an unprecedented manner.
Keywords:
Lipase
Biolubricant
Esterification
Size-selectivity
Solvent effect
©
2014 Elsevier B.V. All rights reserved.
1
. Introduction
Size selectivity is a phenomenon with consequences for stereo
and regio selectivity. For example the asymmetrisation of cyclic
meso-1,2-dicarboxylates using porcine liver esterase showed
exceptional size selectivity effects [5]. The size of the meso ring
influenced R/S meso-trick hydrolysis selectivity as it was reversed
from S to R with increasing ring size. Another example is lactone
hydrolysis by CAL-B where increasing lactone ring sizes reversed
R and S selectivity in ring opening polymerisation [6]. Also other
polycondensations showed size selectivity using lipases [7]. Size
selectivity is also reported for proteases. Serine prolyl oligopepti-
dase hydrolysed peptide chains after proline in terminal protein
sections with up 30 amino acids [8].
Lipase catalysed acylation of symmetric pentaerythritol with
fatty acids potentially yields mono-, di-, and triacylated esters,
and eventually the less obvious tetraester. Native lipases catalyse
mono-, di-, and triglycerolester hydrolysis and by inverse hydrol-
ysis its formation; a fourth acylation would lead to an unusual
large fat molecule. Such tetraester formation is possible with achi-
ral pentaerythritol, which is a tetrol and acylated once more than
the triol glycerol. The lipase mediated synthesis of pentaerythri-
tol tetraesters is therefore a good platform to test lipase controlled
size selectivity. Pentaerythritol tetraester hydrolysis was examined
before with rat pancreatic lipase and no hydrolysis occurred [1].
Every lipase is different and in this work, Candida antarctica lipase B
Lipase catalysed esterifications in organic solvents is widely
described [9]. Hydrophobic organic solvents with a log P around 2
is used for its potential in size selective catalysis. The entry channel
support well water free (0.01% H O content) lipase catalysis [10,11].
2
with native CAL-B is rather narrow, 10 A˚ × 4 A˚ wide and 12 A˚ deep
Solvent-free lipase catalysis is another possibility and an attractive
option for industrial application. CAL-B showed before size selectiv-
ity with trimethylolpropane, which is in this work further explored
[12].
To examine size selectivity effects in pentaerythritol oleate syn-
thesis a fast and reliable analytical method was required. Gas
chromatography (GC) and high pressure liquid chromatography
[
2]. In view of this small entry to the catalytic site it is well pos-
sible that larger pentaerythritol tetraolates cannot be synthesised.
This assumption is in line with observations that short chain fatty
acids are esterified faster than longer ones [3]. The hydrolysis rate
of common triglycerides was even further reduced by engineering
a more restricted entry channel which enhanced size selectivity [4].
(HPLC) are not well suited as oleic acid mixtures contain also
other fatty acids (>10%). This multiplies product peaks with each
additional acylation on a tetrol. Any chromatographic method is
therefore of limited use. Pentaerythritol tetraesters are widely
described but it is not excluded that in many instances they
∗ Corresponding author. Tel.: +41 27 606 86 58; fax: +41 27 606 86 15.
E-mail address: fabian.fischer@hevs.ch (F. Fischer).
http://dx.doi.org/10.1016/j.molcatb.2014.08.004
1
381-1177/© 2014 Elsevier B.V. All rights reserved.

M. Happe et al. / Journal of Molecular Catalysis B: Enzymatic 109 (2014) 40–46
41
were less pure than thought due to the absence of a simple ana-
2.5. Sulphuric acid catalysis
lytical method to assay its esterification degree using bio and
chemical catalysis. Recently, an alternative convenient 1H NMR
method was developed and proven successful in the analysis of
trimethylolpropan oleate biolubricant mixtures [12]. A 400 MHz
A series of 25 mL vials were filled with substrates as in the
general procedure but without solvent. 5% concentrated sulphuric
acid (96%) was added (based on pentaerythritol weight). The
vials were sealed and fixed on a rotating (8 rpm) cylinder in a
NMR spectrometer in combination with THF-d separated the three
8
◦
possible methylene singlets H-C(2) of the three acylation degrees
and allowed product distribution analyses. Such chemical shift sep-
aration was not observed with other deuterated solvents such as
chloroform (CDCl ) [13] or methanol (CD OD). It was hypothesised
100 C preheated hybridisation oven. At defined times vials were
removed and aliquots of reaction mixture were analysed by 1
H
NMR/THF-d8.
3
3
that the four possible pentaerythritol oleates can equally be iden-
tified using this novel and convenient 1H NMR method.
2.6. Conversion, yield and product distribution by 1
H NMR
Pentaerythritol esters are high performance lubricants in air-
craft engines [14] and various industrial machinery, but are also
used as consistency wax in cosmetics [15]. It is also used, for exam-
ple, as lubricant in new green refrigeration systems because it
dissolves carbon dioxide that is nowadays used instead of halogen
alkanes [16].
In the following pentaerythritol is acylated with oleic acid by
CAL-B catalysis (Fig. 1). Size selectivity and product distribution
will be influenced by various process parameters, such as pres-
ence and absence of solvent, stoichiometry, and temperature. A
Fischer–Speier esterification shall distinguish substrate bulkiness
from size selectivity by C. antarctica lipase B.
A single 1H NMR experiment provided complete reaction
progress information. The solvent of any reaction mixture was
1
evaporated and aliquots thereof dissolved in 0.5 mL THF-d and a H
8
NMR recorded on a 400 MHz NMR spectrometer (Bruker Avance).
This NMR-solvent separated the chemical shifts for all four pos-
sible products 1–4 and also oleic acid conversion was assayed at
the same time. The pentaerythritol-oleates 1–4 distribution was
determined by integrating H-C(2) singlets found at: 4.05 ppm (1),
4
.06 (2) 4.08 ppm (3), and 4.11 ppm (4). Their integrams were
divided by related acylation degrees (1, 2, 3, 4) and the results
were normalized to 100%. The oleic acid conversion was quanti-
ꢀ
fied by the H-C(2 ) triplet that shifted from 2.19 to 2.29 ppm upon
esterification.
2
. Experimental
2.1. Material
2.7. Analytical and preparative chromatography
C. antarctica Lipase B, Novozym® 435 (10,000 PLU/g) was
High pressure liquid chromatography (HPLC) was realised
with an Agilent PN 993967-906 and Eclipse XDB-C8 column,
4.6 × 150 mm, 5 ␮m; and a Charged Aerosol Detector, CAD-Ultra
ESA Dionex. Fractions from selected peaks were collected and
obtained from Novozymes A/S [Bagsværd, Denmark]. 3 A˚ Molec-
ular sieve from Merck KGaA, [Darmstadt, Germany] was activated
before use. Oleic acid (90%) and pentaerythritol were purchased at
Sigma-Aldrich Chemie [Taufkirchen, Germany]. Tert-butanol (99%)
was from Fluka Analytical [Buchs, Switzerland], and THF-d8 from
Armar AG [Döttingen, Switzerland].
eluted a second time. Isolated compounds were then analysed
by 1H NMR for purity. Also preparative product isolation using
silica gel normal pressure chromatography with EtOAc–heptane
(1:2) was realised [12]. The separation quality was controlled by
2.2. General procedure for lipase catalysis in tert-butanol
thin layer chromatography (TLC) and two separation runs were
required.
0.27 g (2 mmol) pentaerythritol and 0.58 g (2 mmol) oleic acid
were added (1:1 stoichiometry) to a 25 mL vial. Also other oleic
2
2
.8. Analytical data
acid quantities were used: 1.15 g (4 mmol), 1.70 g (6 mmol), and
®
2
.26 g (8 mmol). 0.14 g Novozym 435 (CAL-B) (50% weight of pen-
.8.1. Pentaerythritol monooleate (1)
taerythritol) was employed with all substrate ratios. 0.8 g molecular
sieve 3 A˚ was added to adsorb condensing water; and finally, 2 mL
tert-butanol was filled in. The sealed vial was fixed vertically on a
ꢀ
ꢀ
H (400 MHz, THF-d ): 5.33 (m, 2H, HC CH, H-9 , H-10 ); 4.05
␦
8
(s, 2H, OCH , H -2); 3.61 (t, 3H, OH, H-2); 3.51 (d, 6H, OCH , Ha-2);
2
b1
2
ꢀ ꢀ ꢀ
.27 (t, 2H, COCH , H-2 ); 2.04 (m, 4H, CH . H-8 , H-11 ); 1.57 (t, 2H,
2 2
2
rotating cylinder (8 rpm) in a hybridisation oven and processed at
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
◦
CH2, H-3 ); 1.32–1.29 (bm, 20H,CH2, H-4 -7 , H-12 -17 ), 0.89 (t, 3H,
CH3, H-18 ).
6
5 C for up to 15 days. For reaction control an aliquot of reaction
ꢀ
1
mixture was taken, tert-butanol evaporated, and a H NMR/THF-d
8
ꢀ
ꢀ
C (100 MHz, THF-d ): 173.23 (C O, C-1 ); 130.38 (C C, C-9 , C-
␦
8
recorded to determine conversion, yield and product (1–4) distri-
bution.
ꢀ
0 ); 63.34 (C(O)OCH , C-2); 61.82 (HOCH , C-2); 45.15 (C(CH ) ,
2 2 2 4
1
ꢀ
C-1); 34.49 (C(O)CH , C-2 ); 32.75, 30.60, 30.57, 30.37, 30.16, 30.13,
3
2
2
ꢀ ꢀ ꢀ ꢀ ꢀ ꢀ
0.09, 29.99, 29.96, CH , C-4 -7 , C-12 -16 ; 27.9 (CH , C-8 ,C-11 );
2 2
2.3. Solvent screening for optimal lipase catalysis
ꢀ
ꢀ
ꢀ
3.44 (CH , C-3 ,C-17 ); 14.32 (CH , C-18 ).
2
3
A
1:4 (pentaerythritol/oleic acid) substrate mixture was
screened using various organic solvents (2 mL) according to the
general method. After processing the immobilised lipase and
molecular sieves were filtered off and the solvent evaporated with
a Rotavap at 25 mbar or on a high vacuum stand. An aliquot of the
2.8.2. Pentaerythritol dioleate (2)
H (400 MHz, THF-d ): 5.32 (m, 4H, HC CH, H-9 , H-10 ); 4.06 (s,
ꢀ
ꢀ
␦
8
2H, OCH , H -2); 3.53 (m, 4H, OCH , H -2); 2.28 (t, 4H, COCH , H-
a
2
b2
2
2
ꢀ
ꢀ
ꢀ
ꢀ
2 ); 2.03 (m, 8H, CH , H-8 , H-11 ); 1.57 (t, 4H, CH , H-3 ); 1.33–1.30
(bm, 40H, CH , H-4 -7 , H-12 -17 ); 0.89 (t, 6H, CH , H-18 ).
2
2
1
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
raw reaction mixture was analysed by H NMR.
2
3
ꢀ
ꢀ
C (100 MHz, THF-d ): 173.23 (C O, C-1 ); 130.38 (C C, C-9 , C-
␦
8
ꢀ
0 ); 63.34 (C(O)OCH , C-2); 61.82 (HOCH , C-2); 45.15 (C(CH ) ,
2 2 2 4
1
2.4. Solvent-free lipase catalysis
ꢀ
C-1); 34.49 (C(O)CH , C-2 ); 32.75, 30.6, 30.57, 30.37, 30.16, 30.13,
2
ꢀ ꢀ ꢀ ꢀ ꢀ ꢀ
0.09, 29.99, 29.96 (CH , C-4 -7 , C-12 -16 ); 27.9 (CH , C-8 ,C-11 );
2 2
3
2
The same preparation and process conditions, without solvent,
according to the general procedure.
ꢀ
ꢀ
ꢀ
3.44 (CH , C-3 ,C-17 ); 14.32 (CH , C-18 ).
2
3

4
2
M. Happe et al. / Journal of Molecular Catalysis B: Enzymatic 109 (2014) 40–46
Fig. 1. Lipase mediated acylation of pentaerythritol (PE) with oleic acid. Using tert-butanol as solvent intermediate 3 accumulated and 4 was hardly formed. Size selectivity
was lost under solvent-free conditions and tetraoleate 4 formed in excess.
ꢀ
2
.8.3. Pentaerythritol trioleate (3)
C-1); 34.41 (C(O)CH , C-2 ); 32.75, 30.6, 30.58, 30.55, 30.16, 30.13,
2
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
H (400 MHz, THF-d ): 5.33 (t, 6H, HC CH, H-9 , H-10 ); 4.08 (s,
30.09, 29.99, 29.96 (CH , C-4 -7 , C-12 -16 ); 27.91 (CH , C-8 ,C-11 );
␦
8
2
2
ꢀ ꢀ ꢀ
23.44 (CH , C-3 , C-17 ); 14.32 (CH , C-18 ).
2 3
6
2
H, OCH , H -2); 3.53 (m, 2H, OCH , Ha-2); 2.28 (t, 2H, COCH , H-
2
b3
2
2
ꢀ
ꢀ
ꢀ
ꢀ
); 2.04 (s, 12H, CH . H-8 , H-11 ); 1.57 (m, 6H, CH , H-3 ); 1.32–1.29
2
ꢀ
2
ꢀ
ꢀ
ꢀ
ꢀ
(
bm, 60H, CH , H-4 -7 , H-12 -17 ); 0.89 (t, 6H, CH , H-18 ).
2.8.4. Pentaerythritol tetraoleate (4)
H (400 MHz, THF-d ): 5.33 (m, 8H, HC CH, H-9 , H-10 ); 4.11 (s,
2
3
ꢀ
ꢀ
ꢀ
ꢀ
C (100 MHz, THF-d ): 172.98 (C O, C-1 ); 130.38 (C C, C-9 ,
␦
8
␦
8
ꢀ
ꢀ ꢀ
8H, OCH , H -2); 2.29 (t, 8H, COCH , H-2 ); 2.04 (m, 16H, CH . H-8 ,
b4
2 2 2
C-10 ); 62.97 (C(O)OCH , C-2); 61.01 (HOCH , C-2); 44.2 (C(CH ) ,
2
2
2
4
Fig. 2. (A) Lipase mediated catalysis of pentaerythritol with various oleic acid ratios in tert-butanol. Monooleate 1 (blue) in high proportion, dioleate 2 (red) transformed
swiftly into trioleate 3 (green), while tetraoleate 4 (violet) formation was limited due to size selectivity. (B) Size selectivity (A) was lost with solvent-free conditions. A binary
product mixture of tri- and tetraoleates 3–4 resulted and tetraoleate 4 (violet) accumulated over time up to 93% (For interpretation of the color information in this figure
legend, the reader is referred to the web version of the article.).

M. Happe et al. / Journal of Molecular Catalysis B: Enzymatic 109 (2014) 40–46
43
Table 1
Oleic acid conversion in enzymatic acylation of pentaerythritol using CAL-B to screen
for a suitable organic solvent.
Solvent
Oleic acid conversion [%]
Pentane
Toluene
tert-Butanol
Dioxane
Acetone
Hexane
THF
Cyclohexane
Diphenylether
41.7
40.8
39.5
33.8
30.9
21.9
21.0
17.9
15.9
Fig. 3. Fischer–Speier esterification of pentaerythritol with oleic acid (1:1, n/n).
Monooleate 1 (green) was formed rapidly, dioleate 2 (red) was readily transformed
into trioleate 3 (blue) and the tetraoleate 4 (violet) fraction increased constantly.
The trioleate’s 3 bulkiness retarded tetraoleate 4 formation to some degree (For
interpretation of the color information in this figure legend, the reader is referred
to the web version of the article.).
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
H-11 ); 1.59 (m, 8H, CH , H-3 ); 1.32-1.29 (bm, 80H, CH , H-4 -7 ,
2
2
ꢀ
ꢀ
H-12 -17 ); 0.89 (t, 12H, CH , H-18 ).
3
ꢀ
ꢀ
C (100 MHz, THF-d ): 172.98 (C O, C-1 ); 130.38 (C C, C-9 -
␦
8
ꢀ
0 ); 62.97 (C(O)OCH , C-2); 44.2 (C(CH ) , C-1); 34.41 (C(O)CH ,
2 2 4 2
1
ꢀ
C-2 ); 32.75, 30.6, 30.58, 30.55, 30.16, 30.13, 30.09, 29.99, 29.96
ꢀ ꢀ ꢀ ꢀ ꢀ ꢀ
CH , C-4 -7 , C-12 -16 ); 27.91 (CH , C-8 , C-11 ); 23.44 (CH , C-
2 2 2
(
3
size selectivity limits conversions trioleate 3 is not transformed
into tetraoleate 4. With prolonged process time, up to 15 days,
the oleic acid conversions hardly improved (Fig. 2A). As a con-
sequence, product mixtures were obtained in all transformations
with any substrate ratios. This is a usual occurrence with multi-
ple reactions on the same molecule and was observed before with
similar trimethylolpropane-oleate biolubricants [18]. Product mix-
tures after multiple reactions on the same molecule are also known
in chemical catalyses of star and wheel like structures [19].
ꢀ
ꢀ ꢀ
,C-17 ); 14.32 (CH , C-18 ).
3
3
. Results and discussion
.1. Lipase catalysed tetrol acylation in tert-butanol
Size selectivity induced pentaerythritol trioleate 3 accumula-
tion was observed and at any time no tetraoleate 4 was formed
in the first 24 hours (Fig. 2A). A solvent screening was realised
to ensure optimal substrate solubility, pentane, toluene and tert-
butanol were well suited for this biotransformation (Table 1) [17].
Tert-butanol was chosen for all experiments as the most convenient
and greenest solvent of the available options. The biotransforma-
3
1
The H NMR of obtained product mixtures showed that mono
acylated pentaerythritol 1 was rapidly generated with all substrate
ratios followed by dioleate 2 formation. The dioleate 2 was swiftly
transformed into trioleate 3, which accumulated, and tetraoleate
4 formation was largely suppressed. Nevertheless in some cases
minor quantities of 4 were recovered (1–12%) (Fig. 2B). The fourth
acylation was sluggish due to diffusion-like kinetics as observed
before with trimethylolpropane which is structurally close to pen-
taerythtitol [12]. It was therefore not excluded that bulkiness of 3
tion of oleic acid (1:1, n:n) with pentaerythritol was quantitative
98.5%) at 65 C/three days. The same lipase quantity was also used
with increasing oleic acid ratios (Fig. 2A) and conversions decreased
slowly to 71%, which is close to expected 75% using a 1:4 ratio. If
◦
(
Fig. 4. Temperature influence on size selectivity of CAL-B using tert-butanol/24 h. X-axes pentaerythritol/oleic acid ratios (1:1 to 1:4). Monooleate 1 (blue), dioleate 2 (red),
trioleate 3 (green). At all temperatures tetraoleate 4 (violet) formation was largely suppressed due to size selectivity (For interpretation of the color information in this figure
legend, the reader is referred to the web version of the article.).

4
4
M. Happe et al. / Journal of Molecular Catalysis B: Enzymatic 109 (2014) 40–46
contributed to observed size selectivity. This led to the question
if the observed size selectivity was purely caused by C. antarctica
lipase B or if it was enhanced by substrate bulkiness. Native lipases
hydrolyse and synthesise mono-, di- and triglycerides but not
pseudo tetraglycerides as the investigated tetraoleate 4. The cat-
alytic active site with CAL-B is located at the end of a narrow access
port, which is a 10 A˚ × 4 A˚ wide and 12 A˚ deep channel [2]. Conse-
quently, there is limited space to transform bulky trioleate 3 into
tetraoleate 4.
3.2. Solvent-free lipase catalysis
Solvent-free conditions caused size selectivity loss yielding
mainly tetraoleate 4 and no monooleate 1 and dioleate 2. Over
prolonged reaction time, the tetraoleate 4 fraction increased up
to 93% and the rest was trioleate 3 (Fig. 2B). The poor solubility
of pentaerythritol in oleic acid was the origin of the narrow prod-
uct distributions in solvent-free conditions (Fig. 2B). The absence
of mono and diacylated products 1 and 2 was explained by poor
pentraerythritol dissolution. The first acylation (PE → 1) was there-
fore slow. The second (1 → 2) and third (2 → 3) acylations were,
by contrast, fast as the intermediates 1 and 2 were dissolved and
became further acylated. The trioleate 3 accumulated due to CAL-
B’s size exclusion effect. A further reaction (3 → 4) to tetraoleate 4
was observed and size selectivity was lost most likely as a result of
different kinetics. Whether the size limiting channel in CAL-B was
altered by solvent-free conditions is not determined and further
work is needed. All in all, solvent-free conditions led to tetraoleate
4
excess with only one of three other esters. This was not observed
under solvent conditions.
3.3. Fischer–Speier esterification
Acid catalysed esterification was conducted to estimate the
size selectivity effect caused by substrate bulkiness, in particular
of intermediate 3. A catalysing proton is very small in compar-
ison to steric C. antartica lipase B and any substrate bulkiness
effect should become apparent with acid catalysis. The solvent-free
◦
Fischer–Speier esterification was realised at 100 C with 5% sul-
phuric acid [20]. Resulting product distribution was much broader
than in solvent-free CAL-B catalysis. With sulphuric acid all four
possible esters 1–4 were formed (Figs. 2B and 3). Initially, trioleate
3
accumulated with acid catalysis as under solvent conditions using
CAL-B. But the accumulation of 3 was temporary, and then its con-
centration dropped. The initially observed trioleate 3 excess was
explained by the expected substrate bulkiness effect determining
the reaction rate for the fourth esterification (3 → 4). All in all, acid
catalysis was non-selective in comparison to lipase catalysis (Fig. 3).
In CAL-B catalysis no tetraoleate 4 was formed after a processing
times that was 6 times longer (Fig. 2A; see 1:1 ratio). In final con-
clusion, substrate bulkiness was a minor cause in observed size
selectivity with CAL-B.
Fig. 5. (A) Oleic acid conversions under solvent conditions. (B) Solvent-free trans-
formation resulted in low oleic acid conversions. (C) Temperature dependence of
oleic acid conversions with tert-butanol/24 h.
improved from 2.9% to 73% and declined to 66% above the opti-
◦
mal temperature with 95 C/24 h. The size selectivity persisted up
◦
to 65 C as no tetraoleate 4 was detected (Fig. 4). However, when
the temperature rose above 75 C a minor tetraester 4 formation
◦
was observed. The trioleate 3 excess was reduced at higher tem-
peratures, but this was also the case with lower temperatures
3.4. Temperature variation and selectivity
◦
<
65 C. The cause of this non-linear temperature influence was
not established. Data recording from experiments realized at lower
temperature was less reliable as conversions were low and signals
Catalyst selectivity improves in general by lowering process
temperature [21]. This is based on the theory about the activated
◦
#
#
#
week what explains the absence of the dioleate 2 at 45 C/1:4 ratio
Fig. 4). Finally, the temperature variation experiments showed that
complex described by the Eyring equation ꢀG = ꢀH − TꢀS that
(
allows to separate enthalpic from entropic influence. The entropy
#
there was persisting size selectivity with CAL-B from low to high
reaction temperatures.
ꢀS
contribution in the transition state becomes less impor-
tant with lower temperature and consequently selectivity should
improve [22]. In the process examined here, four consecutive reac-
tion steps occurred (PE → 1 → 2 → 3 → 4). The well observable size
selectivity concerned the 3 → 4 transformation. Starting with the
3.5. Comparison of persistence and absence in size selectivity
◦
lowest possible reaction temperature, 35 C/24 h, oleic acid con-
The size selectivity of CAL-B depended on the use or absence of
tert-butanol. This solvent dissolved the substrates at the same time
verted sluggishly (Fig. 4). With raising temperature, conversions

M. Happe et al. / Journal of Molecular Catalysis B: Enzymatic 109 (2014) 40–46
45
Fig. 6. 1H NMR (400 MHz, THF-d8) of Hx-C(2) product singlets (x = a, b1, b2, b3 and b4) in biolubricant mixtures. (A) All four possible products 1–4 (right). Conditions: CAL-B
◦
◦
in tert-butanol/3d/65 C with 1:1 PE/OA ratio. (B) Two product mixture of 3 and 4 (20:80). Conditions: Solvent-free CAL-B catalysis with 1:4 PE/OA ratio after 15 days/65 C.
1
Right: proton labeling for Hx-C(2) assignment in H NMR spectra A and B for 1–4. Ha protons appeared all at ıHa = 3.53 ppm and are not visible in A and B.
and conserved CAL-B activity due to its preferable hydrophobicity
log P = 0.8 [23], 0.35 [24,25]. This lipase catalysed mono-, di- and
triglyceride formation, however, the synthesis of pseudo tetraglyc-
eride 4 remained well restricted. The available space in the channel
toward the catalytic site in CAL-B is small and caused most likely
observed size selectivity. This hypothesis was tested by examining
if trioleates 3 bulkiness reduces the reaction rate of the fourth acy-
lation to a significant degree. For this purpose, sulphuric acid was
used as there a small proton catalyses instead of sterically hindered
CAL-B. Also the temperature was varied. Due to the robustness of
four singlets with the available 400 MHz NMR spectrometer. As
aforementioned, the four individual Hb(1–4)-C(2) singlets appeared
in a narrow perimeter at: ıHb1 = 4.05 ppm (1), ıHb2 = 4.06 ppm (2),
ıHb3 = 4.08 ppm (3) and at ıHb4 = 4.11 ppm (4) (see labelling in
Fig. 6). The proton integrams were divided by the corresponding
acylation degrees and normalised to 100%. Oleic acid conversion
ꢀ
was determined from H-C(2 ) triplet that shifted from 2.19 to
2.29 ppm upon esterification.
HPLC analysis was considered too, but found less convenient
than 1H NMR analysis because of the use of impure oleic acid
(90%) that contained also other fatty acids. In HPLC multitudes
of peaks were recorded. Fractions of HPLC-peaks recovered in a
first separation run were eluted a second time and the recovered
fractions were again analysed by 1H NMR. The spectra showed
that even after two chromatographic separations impurities had
to be expected. In some contrast, thin layer chromatography (TLC)
indicated simpler mixtures and suggested easier separation. But
following preparative chromatography on silica-gel showed that
product purification remained challenging as experienced with
HPLC. At least two separation runs were needed to isolate pure
products. All in all, the convenient 1H NMR/THF-d8 method facili-
tated size selectivity examinations by far.
®
immobilised CAL-B (Novozym 435) it was possible to conduct
◦
lipase mediated esterification close to 100 C, which allowed to
compare sulphuric acid with enzyme catalysis. The catalyst com-
parison showed that substrates bulkiness is a negligible cause for
observed trioleate 3 accumulation. Nevertheless, trioleate 3 excess
was observed using sulphuric acid, but tetraoleate 4 formation set
in immediately. This contrasted with CAL-B use where hardly any
◦
tetraoleate 4 was formed, and this up to 95 C with a 6-fold longer
reaction time.
Why did solvent-free CAL-B catalysis yield tetraoleate 4 in
excess (93%) and monooleate 1 and dioleate 2 were absent? The
data suggested diffusion-like kinetics at the beginning of the reac-
tion (Fig. 5A–C). The rate determining step was pentaerythritol
solubility in the solvent-free reaction mixture. The further acylation
of intermediate 1 and 2 was fast and led to a kinetically controlled
accumulation of the tetraoleate 4 fraction. All in all, CAL-B allowed
to produce tri- and tetraoleates (3 and 4) in excess.
4. Conclusions
The lipase mediated acylation of pentaerythritol with oleic
acid showed size selectivity effects. Process parameters were var-
ied to cause excessive generation of one of the four possible
products. C. antarctica lipase B (CAL-B) exercised size selectivity
using tert-butanol and pentaerythritol-trioleate 3 accumulated,
while the corresponding tetraoleate 4 was not formed. This con-
trasted sharply with solvent-free conditions where tetraoleate 4
was obtained in up to 93% purity. Also the temperature influ-
3.6. Analysis
1H NMR analysis of reaction mixtures allowed to quantify
substrate conversion, yield and product distribution at once.
Deuterated tetrahydrofuran (THF-d ) separated all four possible
8
◦
Hb(1–4)-C(2) singlets of the products 1–4 (Fig. 6). Other deuterated
ence on the lipase’s size selectivity was examined. Below 65 C,
solvents such as CDCl3 [26,27] or CD OD did not separate these
no tetraoleate was detected after 3 days, above this temperature
3

4
6
M. Happe et al. / Journal of Molecular Catalysis B: Enzymatic 109 (2014) 40–46
some minor tetraoleate formation was observed, but even with
[9] P. Adlercreutz, Chem. Soc. Rev. 42 (2013) 6406–6436.
[10] C. Laane, S. Boeren, K. Vos, C. Veeger, Biotechnol. Bioeng. 30 (1987) 78–81.
◦
9
5 C CAL-B size selectivity persisted rather well. The obtained four
[
11] F. Fischer, M. Happe, J. Emery, A. Fornage, R. Schütz, J. Mol. Catal. B: Enzym. 90
2013) 98–106.
biolubricants 1–4 are envisioned for direct use and also as interme-
diates for further esterification of remaining hydroxyl groups with
specific substituents. All in all, lipase mediated esterification using
CAL-B with and without solvent enabled size selectivity variation
in the acylation of pentaerythritol with oleic acid.
(
[12] M. Happe, P. Grand, S. Farquet, S. Aeby, J.-C. Héritier, F. Corthay, E. Mabillard,
R. Marti, E. Vanoli, A.-F. Grogg, S. Nussbaum, A. Roduit, F. Tièche, S. Salem, C.
Constantin, E. Schmitt, S. Zahno, C. Ellert, A. Habib, J. Wyss, F. Fischer, Green
Chem. 14 (2012) 2337–2345.
[13] K. Kamalakar, A.K. Rajak, R.B.N. Prasad, M.S.L. Karuna, Ind. Crop Prod. 51 (2013)
249–257.
[
[
[
14] M. Kohler, N.V. Heeb, J. Chromatogr. A 926 (2001) 161–165.
15] K. Hill, Pure Appl. Chem. 79 (2007) 1999–2011.
Acknowledgements
16] O. Fandi n˜ o, E.R. López, L. Lugo, M. Teodirescu, A.M. Mainar, J. Fernãndez, J. Chem.
Eng. Data 53 (2008) 1854–1861.
17] Z.-Q. Duan, W. Du, D.-H. Liu, Bioresour. Technol. 101 (2010) 2568–2571.
18] R.E. Gawley, A. Jeffrey, Principles of Asymmetric Synthesis, Elsevier, Amster-
dam, 2012.
This work was partly supported by the HES-SO Call project no.
[
[
2
3506 and the School of Engineering of the HES-SO Valais-Wallis
at Sion.
[
19] Y. Yu, A.D. Bond, P.W. Leonard, K.P.C. Vollhardt, G.D. Whitener, Angew. Chem.
Int. Ed. 45 (2006) 1794–1799.
References
[
20] J.J. Li, Name Reactions Fifth Edition A Collection of Detailed Mechanisms and
Synthetic Applications, Springer, Cham, 2014, pp. 252.
[
[
[
1] F.H. Mattson, R.A. Volpenhein, J. Lipid Res. 13 (1972) 325–328.
2] J. Uppenberg, M.T. Hansen, S. Patkar, T.A. Jones, Structure 2 (1994) 293–308.
3] N.R. Pedersen, R. Wimmer, J. Emmersen, P. Degn, L.H. Pedersena, Carbohyd. Res.
[
[
21] K. Faber, Biotransformations in Organic Chemistry, Springer, Berlin, 2011.
22] C.O. Åkerman, A.E.V. Hagström, M.A. Mollaahmad, S. Karlsson, R. Hatti-Kaul,
Proc. Biochem. 46 (2011) 2225–2231.
3
37 (2002) 1179–1184.
[
[
[
23] G. Carrea, G. Ottolina, S. Riva, Trends Biotechnol. 13 (1995) 63–70.
24] N. Bridiau, N. Issaoui, T. Maugard, Biotechnol. Prog. 26 (2010) 1278–1289.
25] D. Yazbeck, A. Derrick, M. Panesar, A. Deese, A. Gujral, J. Tao, Org. Process Res.
Dev. 10 (2006) 655–660.
[4] H. Brundiek, S.K. Padhi, R. Kourist, A. Evitt, U.T. Bornscheuer, Eur. J. Lipid Sci.
Technol. 114 (2012) 1148–1153.
[
[
5] G. Sabbioni, J.B. Jones, J. Org. Chem. 52 (1987) 4565–4570.
6] J. van Buijtenen, B.A.C. van As, M. Verbruggen, L. Roumen, J.A.J.M. Vekemans, K.
Pieterse, P.A.J. Hilbers, L.A. Hulshof, A.R.A. Palmans, E.W. Meijer, J. Am. Chem.
Soc. 129 (2007) 7393–7398.
[
[
26] K.V. Padmaja, B.V.S.K. Rao, R.K. Reddy, P.S. Bhaskar, A.K. Singh, R.B.N. Prasad,
Ind. Crop Prod. 35 (2012) 237–240.
27] D.-S. Yao, B.-Y. Zhang, Y.-H. Li, W.-Q. Xiao, Tetrahedron Lett. 45 (2004)
[
[
7] L. Martino, M. Scandola, Z. Jiang, Polymer 53 (2012) 1839–1848.
8] D. Rea, V. Fülöp, CNS Neurol. Disord. Drug Targets 10 (2011) 306–310.
8953–8956.