chemija. 2014. vol. 25. No. 3. P.

185–194
© lietuvos mokslų akademija, 2014

Enzymatic phenethyl octanoate synthesis: lipase

selection and reaction optimization by response surface
methodology

Dovilė Šinkūnienė, Phenethyl octanoate (PEO) is a natural flavour compound found in rum, spirits and wine.
Six commercial and non-commercial lipase preparations were screened for 2-phenethyl
Simas Kazlauskas, octanoate synthesis activity, the most active enzymes were Lipozyme® RM IM and Pala­
tase® 20000 L (immobilized and soluble Rhizomucor miehei lipases). Three substrates acyl-
Vida Bendikienė* donors (octanoic acid, glyceryl trioctanoate and coconut oil) and two solvents (hexane and
tert-butanol) were compared in the synthesis reactions. Higher conversions and reaction
Department of Biochemistry rates were observed in hexane using immobilized enzyme. The optimal reaction conditions
and Molecular Biology, determined by means of Response Surface Methodology (RSM) were as follows: reaction
Vilnius University, time 120 minutes, temperature of 30 °C, 0.8 M and 0.267 M concentrations of 2-phenyl­
M. K. Čiurlionio St. 21, ethanol and glyceryl trioctanoate, respectively, 7% concentration of enzyme Lipozyme®
LT-03101 Vilnius, RM IM, resulting in 80% reaction conversion. The phenylethyl ester synthesis reaction was
Lithuania transferred to a solvent-free system using a natural substrate coconut oil.

Key words: phenethyl octanoate, flavor esters, lipase, Rhizomucor miehei, response surface
methodology (RSM)

INTRODUCTION most of the flavouring compounds are produced via chemi-

cal synthesis or by extraction from natural materials. Even
A large variety of compounds determine the flavour of food though flavour compounds often occur naturally in plant
products, including alcohols, aldehydes, esters, dicarbonyls, and animal tissues, the extraction process is costly due to low
short to medium-chain free fatty acids, methyl ketones, lac- quantities in the tissues [1].
tones, phenolic compounds and sulphur compounds. Since Flavour esters are usually produced by conventional
early times humankind has extracted flavour compounds chemical synthesis which seldom is environmental-friendly,
from natural sources. Nowadays flavour esters are widely because polluting and corrosive inorganic catalysts are com-
used in food, fragrance and cosmetic industry. Flavours rep- monly used. In addition, it lacks specificity and selectivity,
resent over a quarter of the world market for food additives, which may cause the formation of undesirable racemic mix-
tures or other by-products, thus reducing process efficiency
* Corresponding author. E-mail: vida.bendikiene@gf.vu.lt and increasing downstream costs [1, 2].
186 Dovilė Šinkūnienė, Simas Kazlauskas, Vida Bendikienė

Consumer preferences to the products labelled “natural” EXPERIMENTAL

and “ecological” increased the demand for naturally produced
flavour compounds. Scientific interest was drawn to the bio- Materials
technological, especially, enzymatic synthesis of flavour com- Lipases Palatase® 20000 L (further in the text Palatase®),
pounds [3], as mild reaction conditions with high specificity Lipozyme® RM IM, Lipoclean®, Lipopan® F BG were kindly
minimize side-reactions and allow easier and cheaper separa- donated by JSC Biopolis, Novozymes representative in Lithu-
tion of reaction products. Enzymatic synthesis is an attractive ania. Lipase from Enterobacter aerogenes-13 (190 U/ml) was
alternative especially when food, flavour and fragrance com- provided by JSC BIOCENTRAS, Vilnius, Lithuania. Lipase
pounds are in scope, because while avoiding harsh reaction from Pseudomonas mendocina 3121-1 (70 U/ml) was pro-
conditions, high yields are achieved and natural compounds vided by the Institute of Biochemistry, Vilnius, Lithuania. The
are synthesized from natural and renewable sources [4]. Per- reagents of analytical grade or >=99% were purchased from
fumery and flavour grade esters need to be free of inorganic Sigma-Aldrich (St. Louis, MO, USA): 2-phenylethanol (2-PE),
impurities; reduced reaction temperature and increased yield phenethyl octanoate (PEO), octanoic acid (OA), glyceryl trioc-
reduce the process costs, therefore enzymatic catalysis is an al- tanoate (GTO), coconut oil, hexane, tert-butanol, diethyl ether,
ternative, especially when food-grade enzymes are employed. petroleum ether, acetic acid. Pre-coated thin-layer chromato-
Phenethyl octanoate (synonyms 2-phenylethyl octanoate, graphic (TLC) plates DC-Fertigplatten SIL G-25 UV254 were
phenethyl caprylate) is a naturally occurring flavour com- purchased from Macherey-Nagel (Düren, Germany).
pound in wines and spirits [5], okra (Abelmoschus esculen-
tus) seed coat [6]. It has a mild, fruity, wine-like odour and PEO synthesis
flavour and can be used in beer, rum, sherry, wine, grape and The ester synthesis was performed in 2 ml capped test tubes
other fruit flavour applications. Some species of ants like at least in duplicate, response surface design (RSM) experi-
Camponotus arminius and Crematogaster clarithorax use it ments were performed according to Table 1, in equimolar
in their chemical communication system [7]. ratios of 2-PE and OA (taking into consideration that GTO
Attempts have been made to synthesize phenethyl esters stoichiometrically contains 3 acyl-moieties). In a standard
enzymatically from natural sources of octanoic acid, espe- procedure the initial concentrations of substrates were 1 M,
cially coconut oil [8, 9]. Coconut fat is rich in short-chain reaction was started by adding an appropriate mass or vol-
fatty acid of C8 (octanoic acid, 7.5% w/w) and coconut cream ume of lipase to the preheated substrate solution in a solvent
is produced from coconut milk after centrifugation and con- (hexane or tert-butanol). Enzyme concentration was 5% v/v
tains about 25% (w/v) of fat. Coconut milk is cheap and read- or m/v for screening reactions and 4.2% for other reactions,
ily available in many countries of the world. Both coconut fat if not stated otherwise. The total volume of the reaction mix-
and coconut cream may be attractive substrates for enzymat- ture was 1 ml.
ic production of natural esters [10, 11]. Samples were incubated in a Biosan Thermoshaker (TS-
In order to obtain high yields of a desired product, en- 100) at 30–60 °C, 1 400 rpm, along with controls. For further
zyme and reaction conditions have to be selected carefully. analysis aliquots of 20 µl, taken from the samples, were mixed
The purpose of this study was to select the appropriate with equal volumes of cold diethyl ether and frozen at –20 °C.
enzyme catalyst and reaction conditions for phenethyl oc-
tanoate synthesis from 2-phenyl ethanol (phenethyl alcohol) Analysis
and various acyl-donor substrates: octanoic acid (caprylic Quantitative thin layer chromatography (TLC). Analysis of
acid), glyceryl trioctanoate (reaction scheme in Fig. 1) and the reaction products was carried out on TLC plates (5 × 10 cm
coconut oil (a natural source of octanoic acid). and 10 × 10 cm) pre-coated with 0.25 mm Silica Gel 60 UV254

Fig. 1. Enzymatic phenethyloctanoate synthesis shceme

 Enzymatic phenethyl octanoate synthesis: lipase selection and reaction optimization by response surface methodology 187

Tab l e 1 . Central composite design parameter values used for the phenethyl octanoate synthesis reaction optimization by RSM
Factor A: Enzyme Factor B: Substrate Response: Phenethyl­
Run Factor C, Temperature, °C Factor D, Time, min
concentration, % concentration, mM octanoate yield, %
1. 7 0.8 40 60 64.9
2. 7 0.8 40 120 72.6
3. 4.5 1.4 35 90 59.1
4. 9.5 1.4 35 90 65.3
5. 4.5 1.4 35 90 55.4
6. 0.1 1.4 35 90 21.5
7. 2 2.0 40 120 42.4
8. 2 2.0 30 60 36.2
9. 4.5 2.6 35 90 49.6
10. 7 2.0 30 120 53.8
11. 4.5 1.4 25 90 49.3
12. 2 0.8 30 120 62.4
13. 4.5 1.4 35 30 50.0
14. 4.5 1.4 35 90 57.4
15. 4.5 1.4 35 90 53.1
16. 2 2.0 40 60 32.0
17. 7 0.8 30 60 77.1
18. 4.5 0.2 35 90 79.3
19. 4.5 1.4 35 90 56.8
20. 2 0.8 30 60 52.1
21. 7 2.0 40 60 48.1
22. 7 0.8 30 120 82.2
23. 2 0.8 40 60 47.7
24. 7 2.0 30 60 52.8
25. 2 2.0 30 120 44.3
26. 4.5 1.4 45 90 56.0
27. 2 0.8 40 120 52.5
28. 7 2.0 40 120 54.3
29. 4.5 1.4 35 90 55.6
30. 4.5 1.4 35 150 65.5

(Merck). The sample volumes of 2 µl were applied at a distance uct – phenethyl octanoate (NIST database was used). Hy-
of 10 mm from the bottom edge of the plate ant 10 mm between drogen was used as a carrier gas, detector temperature was
each spot using a 5 µl Hamilton syringe. The plate was air-dried 350 °C, and starting temperature of a column was 50 °C with
and chromatograms were developed in a covered TLC tank a gradient of 10 °C/min increased to 310 °C.
using a mobile phase of light petroleum ether (b. p. 40–60 °C):
diethyl ether : acetic acid (85:15:2, v/v) (a modified method of Response surface methodology
Sinkuniene et al. [12]). The tank was pre-saturated with mo- Experimental design
bile phase vapour before use. The developed TLC plates were In order to determine the optimal reaction conditions a five-
air-dried for about 10–15 min. 2-PE and PEO spots were visu- level-four-factor central composite design (CCD) was em-
alised under UV light (254 nm), all spots (including OA and ployed, resulting in 30 experimental runs. Each numeric fac-
GTO) were visualised after spraying the TLC plates with 0.2% tor is varied over 5 levels: plus and minus alpha (axial points,
2,7-dichloroflouresceine solution in 2-propanol and drying on alpha = 1.76, a total of eight points), plus and minus 1 (facto-
30 °C surface for one hour. Pure 2-PE, OA, GTO and PEO solu- rial points, a total of 16 points) and the centre point (centre
tions in hexane were used as standards. points had six replicates).
Quantitative analysis of reaction products separated by The variables selected for the GTO transesterification
TLC was performed using the principles of densitometry with 2-phenylethanol were four numeric factors: enzyme
by the UVItec Cambridge Fire-reader imaging system and concentration (2–7%), substrate concentration (0.8–2 M),
UVItec Fire-reader 15.10 software considering the spot area time (60–120 min), temperature (30–40 °C). The experiment
and colour intensity (spot volume). The calculations are based design in actual values is shown in Table 2.
on the standard curves of compounds specific absorbance. The experimental design, data analysis and regression
Gas chromatography-mass spectrometry by Perkin­ model building were performed using the Design Expert
Elmer GC-MS was used to identify the reaction prod­ software (version 8.0.1.0, Stat-Ease Inc., Minneapolis, MN).
188 Dovilė Šinkūnienė, Simas Kazlauskas, Vida Bendikienė

Data analysis composition) and 2-PE [8], but additional data concerning
The experimental data were analyzed using the Design Ex- the phenethyl octanoate yield or optimized synthesis con-
pert 8.01 version and then interpreted. Analysis of variance ditions was necessary for further implementation. Li et al.
(ANOVA), a regression analysis and the plotting of response synthesized phenethyl-esters from butter oil and have shown
surface were performed to establish optimum conditions for that phenethyl octanoate synthesis can be effectively cata-
the transesterification reaction yield. ANOVA was used to test lyzed by Lipozyme TL IM [15]. Here we attempted to define a
adequacy and fitness of the responses for linear, 2 function suitable enzyme preparation and optimal reaction conditions
interaction (2fi) and quadratic functions of the variables. for 2-PEO synthesis.
A model with P-values (P > F) less than 0.05 was regarded Enzymes selected for screening were the commercial
as significant. The lack-of-fit test was used to compare the lipase preparations Palatase® 20000 L, Lipozyme® RM IM,
residual and pure errors at the replicated design points. The Lipoclean®, Lipopan® F BG, Lipolase® L EX, and the non-
highest-order significant polynomial with not significant commercial soluble lipase preparations from Enterobacter
lack of fit was selected. The predicted residual sum of the aerogenes and Pseudomonas mendocina.
squares (PRESS) was used as a measure of fit of the model to Immobilized (Lipozyme® RM IM) and soluble (Palatase®)
the points in the design. After predicting the optimal condi- preparations of Rhizomucor miehei lipase (produced by sub-
tions for synthesis reaction, the experiment was repeated in merged fermentation of a genetically modified Aspergillus
triplicate to check the reliability of the predicted values and oryzae microorganism) were selected to determine the in-
experimental data. fluence of the catalyst physical state [16]. Other two lipase
preparations of Novozymes were selected with expectation
RESULTS AND DISCUSSION to have a wide range of substrate specificity, as suggested by
their standard application: Lipoclean® is used for laundry de-
Enzyme screening and selection tergents while Lipopan® for baking industry [16]. Lipolase®
Possibility to synthesize a desired product in a good yield de- L EX was chosen because of an outstanding esterification and
pends not only on a successful selection of an enzyme catalyst transesterification activities in biodiesel synthesis reactions,
but also on a careful choice of reaction conditions. All lipases as shown by our group (not published data). Enterobacter
naturally catalyze hydrolysis of triglyceride ester bonds, but aerogenes lipase was selected for screening, because it has the
the specificity and selectivity for certain substrates vary a highest activity in hydrolysis of medium chain-length fatty
lot [13]. Usually the reason for different enzyme selectivity acid esters, in particular octanoic acid nitrophenyl ester [17]
lies behind the structure of an active site, but knowledge of and Pseudomonas mendocina lipase was chosen because of
the active site structure does not yet permit to forecast the its property to have a higher relative activity for natural sub-
enzymes specificity or selectivity. This is also true for lipases strate coconut oil than longer-fatty acids containing triolein
[14]. Since different lipases from varied sources have distinct or olive oil [18].
substrate selectivity and specificity, it is valuable to know From the screening results, shown in Fig. 2, 2-phenyl­
their ability to produce different esters. ethyl ester formation is obvious after two hours of reaction
It was previously shown by Tan et al. that commercial time when reaction is catalyzed by Palatase® (lane 3) and
preparations of fungal lipases from Rhizomucor miehei, Can- Lipozyme® RM IM (lane 4). A much lower product quantity is
dida rugosa and Aspergillus niger are capable to catalyze PEO observed when Lipolase L EX is used. There was no esterifica-
synthesis from coconut milk (containing OA in triglyceride tion activity in cases of Lipopan F BG or lipases from E. aero-

Fig. 2. Lipase screening for 2-phenethyl octanoate (PEO) synthesis reaction: esterification of octanoic acid (OA)
with 2-phenylethanol (2-PE). Reaction time: 2 hours, 30 °C. Lanes: standards of OA, PE, PEO; reaction catalyzed
by Lipopan® F BG (1), Lipoclean® (2), Palatase® (3), Lipozyme® RM IM (4), Pseudomonas mendocina lipase (5),
Enterobacter aerogenes lipase (6), Lipolase® 100 L EX (7). Ester products are shown in a box
 Enzymatic phenethyl octanoate synthesis: lipase selection and reaction optimization by response surface methodology 189

Tab l e 2 . Lipase screening results for 2-phenethyl octanoate synthesis reaction (OA + PEO). Only active lipases are presented
Ester yield* in screening reactions, %
Enzyme
2h 24 h
Palatase® 33.5 67.3
Lipozyme® RM IM 63.8 70.8
Lipolase® 100 L EX 4.1 62.9
* Standard deviation is 1.51.

genes and Ps. mendocina. The screening reaction yields after found not only in coconut oil, but also in dairy fat (butter),
2 and 24 hours are represented in Table 2. which can be used as a complex substrate [8].
In two hours high conversion was obtained using both Both enzyme preparations were more active in hexane
soluble and immobilized Rhizomucor miehei lipase, therefore compared to tert-butanol (Fig. 3). Although higher ester
the preparations were selected for further experiments to de- yields in 6 hours are obtained in hexane (63% Lipozyme RM
termine the optimal reaction conditions. IM), initial reaction rates depend more on the type of the
catalyst used, and are higher for immobilized enzyme prepa-
Influence of solvent on PEO yield ration both in hexane and tert-butanol. The reaction with the
Several conventional organic solvents were chosen, because immobilized enzyme reaches plateau in 1–3 hours, whereas
the substrates and products of PEO synthesis reaction are reactions catalyzed by Palatase® do not seem to reach equilib-
scarcely soluble or insoluble in water. In addition, aqueous rium after 6 hours in both solvents.
media could result in an increased hydrolysis of the reaction The same tendencies are observed for GTO substrate:
product. Lipase activity and specificity usually depends on a lower reaction yield in 24 hours is obtained in tert-butanol
physicochemical properties of solvents, especially hydropho- (Fig. 4). Ester yield is especially low when using a soluble en-
bicity [19, 20]. Therefore two solvents with different polarities zyme (Palatase®) in tert-butanol, in this case the conversion
were chosen: n-hexane being a non-polar solvent ant tert- is only 4.3% after 24 hours. The highest yield is observed with
butanol a polar solvent with 1-octanol / water partition coef- Lipozyme® RM IM in hexane, and reaches about 65% in six
ficient Log P values of 4.00 and 0.35, respectively [21]. It was hours and does not increase further. If Lipozyme® RM IM in
previously shown that esterification of undecanoic acid with tert-butanol is used, the reaction also reaches equilibrium af-
glycerol catalyzed by Rhizomucor miehei lipase (Lipozyme® ter one hour, but the conversion is only 40%.
IM20) resulted in the greatest ester yields in solvents of the If coconut oil substrate is used, the trend of much faster
highest log P (4.0–4.5) [22]. reaction with Lipozyme® RM IM and higher yields in hexane
We have chosen three different substrates for PEO syn- remains the same. The main difference in using a coconut oil
thesis: OA, GTO and coconut oil. OA serves as a substrate substrate for transesterification with 2-PE is that the concen­
for esterification reaction and the latter two are substrates tration increases even after 6 hours of reaction in all cases and
for transesterification reactions. Coconut oil is a natural tri- reaches over 80% in the case of Lipozyme® RM IM in hexane
glyceride substrate with a mixture of fatty acids of various (Fig. 5). Different curve pattern could be observed due to the
lengths and saturations [23], therefore GTO was chosen as a different enzyme specificity to different fatty acids present in
model substrate to demonstrate that the targeted OA is incor- coconut oil [23]. Rhizomuor miehei lipase similarly to other
porated into PEO. Triglycerides containing octanoic acid are lipases has different selectivity for various chain-length fatty

Fig. 3. Esterification of octanoic acid with 2-phenyl-

ethanol, catalyzed by Palatase® and Lipozyme® RM IM
enzymes at 30 °C, in hexane and tert-butanol
190 Dovilė Šinkūnienė, Simas Kazlauskas, Vida Bendikienė

Fig. 4. Transesterification of gliceryltrioctanoate with

2-phenylethanol, catalyzed by Palatase® and Lipo-
zyme® RM IM enzymes at 30 °C, in hexane and tert-
butanol

Fig. 5. Transtesterification of coconut oil with PE, cata-

lyzed by Palatase® and Lipozyme® RM IM enzymes at
30 °C, in hexane and tert-butanol

acid glyceryl esters present in coconut oil (most of them have is commonly optimized by RSM [26–30], in most cases es-
a longer aliphatic chain compared to octanoic acid). If the en- ters are synthesized via direct esterification reaction. There
zyme had a higher selectivity towards shorter (including oc- are some examples of transesterification reactions involving
tanoic) acid esters, the synthesis of phenethyl octanoate from 2-PE. Rose aromatic ester (2-phenylethyl acetate) was syn-
coconut oil in shorter reaction times would be preferred due thesized enzymatically by transesterification of vinyl ac-
to higher proportion of reacted favourable fatty acid. More etate with 2-PE and was optimized by RSM, the yields have
experimental data is needed to make conclusions. reached 85.4% in 79 minutes with 4% Candida antarctica en-
zyme (Novozyme® 435) for 50 mM substrate concentrations
Optimal reaction conditions determined by response sur- in hexane [31]. Because of the importance to have a method
face methodology for “natural” ester synthesis, for optimization reaction we
Response surface methodology (RSM) consists of a group of have chosen a triglyceride GTO substrate which resembles
mathematical and statistical techniques used to develop an the reaction where natural fats are used for the synthesis.
adequate functional relationship between a response of inter- According to the data presented in 3.1, immobilized li-
est and a number of associated controllable variables. Deter- pase Lipozyme® RM IM and hexane solvent were chosen. The
mining the relationships between several variables using the influence of four reaction setup factors (enzyme and sub-
RSM means varying several parameters at a time, according strate concentrations, temperature and time) on PEO yield
to the statistically built-up design, which can be represented was determined.
by a matrix [24]. In contrast to the classical technique which The experiment design and response in actual values is
varies only one parameter at a time, is time consuming and shown in Table 2. The regression calculations were done by
cannot predict the parameter interactions, RSM is a faster the Design Expert program to fit the polynomial models to
and less expensive technique which provides sufficient infor- the selected response. A quadratic model was suggested as
mation for statistically acceptable results [25]. the most appropriate and the model was reduced by remov-
Simple fruit flavour ester synthesis (n-butyl acetate, n- ing the insignificant terms (with p values of “Prob > F” great-
propyl acetate, isoamyl butyrate, ethyl butyrate and others) er than 0.1000, indicating model terms are insignificant),
 Enzymatic phenethyl octanoate synthesis: lipase selection and reaction optimization by response surface methodology 191

the term BC was close to the marginal value therefore it was The mutual effect of parameters can be better understood
left in the model. The ANOVA is presented in Table 3. The by examining the response surface charts. In a three-dimen-
model equation in actual factors is presented in Table 4. The sional chart two variables can be explored, the others have to
model is highly significant with a p-value of <0.0001 and the be set to constant values. In Fig. 6 the influence of enzyme
adjusted multiple correlation coefficient R2 = 0.9167. and substrate concentrations to the reaction yield is plotted:

Tab l e 3 . ANOVA for the response surface quadratic model

p-value
Source Sum of squares df Mean square F value
Prob > F
Model 4 718.05 8 589.76 40.89 <0.0001 Significant
A-Enzyme concentration 2 164.95 1 2 164.95 150.09 <0.0001
B-Substrate concentration 1 788.02 1 1 788.02 123.96 <0.0001
C-Temperature 45.15 1 45.15 3.13 0.0914
D-Time 297.53 1 297.53 20.63 0.0002
AB 48.51 1 48.51 3.36 0.0809
BC 40.25 1 40.25 2.79 0.1097
A2 292.03 1 292.03 20.25 0.0002
B2 144.09 1 144.09 9.99 0.0047
Residual 302.90 21 14.42
Lack of fit 281.89 16 17.62 4.19 0.0604 Not significant
Pure error 21.01 5 4.20
Cor total 5 020.95 29

Tab l e 4 . Model coefficients for equation in terms of actual factors

Term Model coefficient
Intercept +76.85148
Enzyme concentration +10.60914
Substrate concentration –45.14129
Temperature –1.01448
Time +0.11737
Enzyme conc.·Substrate conc. –1.16082
Substrate conc.·Temperature +0.52868
Enzyme conc.2 –0.56571
Substrate conc.2 +6.24123

Fig. 6. Response surface plot showing the mutual effect of substrate and enzyme concentrations on
the reaction yield (conversion, %), temperature and time are set to mean values (i. e. 35 °C and 90 min)
192 Dovilė Šinkūnienė, Simas Kazlauskas, Vida Bendikienė

the higher the enzyme concentration and the lower the sub- ethanol and 2 mg/ml of each enzyme (Penicillium roqueforti,
strate concentration, the higher is the reaction yield. It is also Candida rugosa and Aspergillus niger) in hexane [32]. In our
observed in Table 4 that the influence of enzyme concentra- case the conversion was higher after ten-fold shorter reaction
tion is positive and the influence of substrate concen­tration is time. Similarly longer reaction time was also applied in the
negative, that could be due to the enzyme saturation by sub- study of Li et al. which have synthesized phenethyl octanoate
strate. In the temperature of 35 °C and 90 minutes of the re- from 2-phenylethanol (125 mM) and octanoic acid (150 mM)
action time the yields vary from 37% (enzyme concen­tration hexane and achieved 84% yield in 24 hours catalyzed by 5%
2%, substrate concentration 2 M) to 75% (enzyme concen­ w/w (reactants) of Lipozyme TL IM [15]. Reaction con-
tration 7%, substrate concentration 0.8 M). ditions optimized in our study offer high productivity in
The effect of the reaction temperature is also negative, a shorter reaction time, 80.0% conversion can be obtained in
which means catalysis works well in lower temperatures. It is as short as two hours.
not surprising that the effect of the reaction time is positive,
but as it can be observed from Fig. 7, even when approaching Transesterification of coconut oil with 2-phenyl ethanol
120 min, the yield is still increasing. Therefore it should be Rhizomucor miehei lipase was tested to carry out the natu-
considered that longer reaction times, if they are acceptable, ral substrate – coconut oil – transesterification reaction in
might increase the PEO yield. solvent-free conditions. Coconut oil is one of the most popu-
lar natural sources rich in octanoic acid. The possibility to
Optimal conditions avoid conventional solvent usage (substrates act as solvents
The optimal conditions of Lipozyme® RM IM-catalyzed themselves) provides easier purification because fewer com-
PEO synthesis were predicted using the Design Expert nu- ponents are present in the final reaction mixture. The process
merical optimization tool. The target was to maximize PEO does not require solvents therefore it is more favourable for
yield by choosing the optimal variable values from all re- the environmental-friendly industrial application. The reac-
sponse surface design range. The optimal conditions for tion product is a mixture of different acid chain length esters
GTO transesterification with PE were as follows: enzyme and can be used not only after further purification but also
concentration = 7%, substrate concentration = 0.8 M, tem- directly in food products, such as cheese [15].
perature = 30 °C, time = 120 min, the predicted yield was A liquid form of lipase Palatase® was tested out for this re-
81.1%. To validate the model the experiment was performed action because a better distribution of enzyme was expected
in the given conditions and the obtained yield was 80.0%. The in a relatively viscous reaction mixture.
results correlate well with the predicted data. The reaction was conducted by a step-wise addition of
Jungbae et al. have synthesized various flavour esters en- alcohol (three times 24 hours apart in equal portions with
zymatically, using enzyme mixtures. The PEO yield was 64% the final concentration of two alcohol molecules per one
after 20 hours using 10 mM octanoic acid, 20 mM phenyl­ acyl moiety). The 2-phenylethyl esters yield reached 37%

Fig. 7. Response surface plot showing the mutual effect of reaction temperature and time on the reaction yield
(conversion, %), enzyme and substrate concentrations are set to mean values (i. e. 4.52% and 1.4 M, respectively)
 Enzymatic phenethyl octanoate synthesis: lipase selection and reaction optimization by response surface methodology 193

Fig. 8. Transesterification of coconut oil by Palatase® (4.2%) in a solvent-free medium at 30 °C

and 54% of oil was transesterified in 72 hours (Fig. 8). The when 7% of lipase preparation is used. Coconut oil, contain-
reaction conversion was not as high as for trioctanoin sub- ing octanoic acid in the triglyceride composition, can be
strate in a solvent. One probable cause for a lower yield is transesterified in a solvent free medium, and even though the
a slower mass-transfer due to higher viscosity, other causes conversion is lower compared to the solvent-system, there is
are enzyme stability and phase limitations. Li et al. have used a high potential to increase it by choosing appropriate reac-
enzymatic synthesis to produce a mixture of 2-phenethyl es- tion conditions.
ters in solvent-free conditions using butter-oil as a substrate.
Under the optimised conditions of the immobilized enzyme ACKNOWLEDGEMENTS
Lipozyme TL IM-catalysed reaction, the total yield of C4–18
2-phenethyl esters was about 817 mg per gram of butter oil. This work was funded by the European Social Fund under the
The optimal reaction conditions led to an 80.0% conversion National Integrated Programme Biotechnology and Biophar-
of 2-phenethyl alcohol [15]. Herein with high probability macy, Grant VP1-3.1-SMM-08-K01-005.
after selection of proper reaction conditions the immobi- Biopolis Ltd (Vilnius, Lihtuania), distributor of Novo-
lized Rhizomucor miehei lipase Lipozyme RM IM could also zymes, is also gratefully acknowledged for a kind supply of
successfully be used for coconut oil transesterification in a the enzymes.
solvent-free medium.
Received 1 April 2014
CONCLUSIONS Accepted 7 May 2014

Increasing attention for the synthesis of “natural” com- References

pounds used in flavour, fragrance and food industry makes
biotechnological processes more attractive, especially when 1. M. A. Longo, M. A. Sanromán, Food Technol. Biotech., 44,
food-grade enzymes can be used. Phenethyl octanoate is a 335 (2006).
fragrance compound that is found and can be used in pro- 2. W.-D. Chiang, S.-W. Chang, C.-J. Shieh, Process Biochem.,
duction of alcoholic beverages and fragrance applications. 38, 1193 (2003).
3. M. T. Batte, J. Beaverson, N. H. Hooker, T. Haab, Food
The synthesis of phenethyl octanoate has not yet been thor-
Policy, 32, 145 (2007).
oughly researched and optimized; therefore selection of ap-
4. B. M. Dhake, K. P. Thakare, D. D. Bhanage, Flavour Fragr. J.,
plicable enzymes and reaction conditions was done.
28, 71 (2013).
PEO can successfully be synthesized enzymatically using 5. Y. Shmuel (ed.), Dictionary of Food Compounds, 2nd edn.,
an immobilized lipase from Rhizomucor miehei (Novozymes® CRC Press (2010).
Lipozyme® RM IM) in a hexane solvent. Soluble lipase (Pal- 6. P. Arapitsas, Food Chem., 110, 1041 (2008).
atase®) is also active. PEO can be synthesized by direct esteri- 7. The Pherobase [http://www.pherobase.com/].
fication of oleic acid or by transesterification of gliceryltrioc- 8. H. S. G. Tan, B. Yu, P. Curran, S. Q. Liu, Food Chem., 124,
tanoate, which was further optimized using response surface 80 (2011).
methodology. Reaction yield is higher in lower temperatures 9. J. Sun, B. Yu, P. Curran, S.-Q. Liu, Food Chem., 135, 2714
(30 °C), 80% yields can be expected in as short as two hours (2012).
194 Dovilė Šinkūnienė, Simas Kazlauskas, Vida Bendikienė

10. M. Lubary, J. H. Ter Horst, G. W. Hofland, P. J. Jansens, Dovilė Šinkūnienė, Simas Kazlauskas, Vida Bendikienė
J. Agr. Food Chem., 57, 116 (2008). FERMENTINĖ FENETILOKTANOATO SINTEZĖ:
11. S. Benjamin, A. Pandey, Acta Biotechnol., 17, 241 (1997).
LIPAZIŲ ATRANKA IR REAKCIJŲ OPTIMIZAVIMAS
12. D. Šinkūnienė, V. Bendikienė, B. Juodka, Rom. Biotech.
ATSAKO PAVIRŠIAUS METODOLOGIJA
Lett., 16, 1 (2011).
13. F. Hasan, A. A. Shah, A. Hameed, A. A. Shan, Biotechnol. Santrauka
Adv., 27, 782 (2009). Nuolat didėjantis susidomėjimas natūraliais junginiais, naudoja-
14. M. Kapoor, M. N. Gupta, Process Biochem., 47, 555 (2012). mais kvepalų ir maisto pramonėje, lemia vis didėjantį biotechnolo-
15. C. Li, J. Sun, T. Li, S.-Q. Liu, D. Huang, Food Chem., 154, ginio gamybos proceso populiarumą. Fenetiloktanoatas – kvapusis
205 (2014). esteris, aptinkamas spirituotame vyne ir kai kuriuose augaluose,
16. Novozymes [www.novozymes.com]. gali būti naudojamas alkoholinių gėrimų ir kvepalų gamyboje. Fe-
17. D. Šinkūnienė, V. Kiriliauskaitė, V. Bendikienė, B. Juodka, netiloktanoato fermentinė sintezė iki šiol nebuvo detaliai ištirta ir
Rom. Biotech. Lett., 13, 74 (2008).
optimizuota, todėl atrinkome tinkamus fermentus bei optimalias
18. U. K. Jinwal, U. Roy, A. R. Chowdhury, A. P. Bhaduri,
reakcijos sąlygas. Iš tirtųjų fenetiloktanoato sintezei atrinkti fer-
P. K. Roy, Bioorgan. Med. Chem., 11, 1041 (2003).
mentiniai preparatai Palatase® ir Lipozyme® RM IM bei tirpiklis
19. F. Secundo, G. Carrea, J. Mol. Catal. B: Enzym., 19–20, 93
heksanas. Tirta galimybė sintetinti fenetiloktanoatą iš skirtingų
(2002).
20. A. R. M. Yahya, W. A. Anderson, M. Moo-Young, Enzyme substratų: oktano (kaprilo) rūgšties (esterinimas), trikaprilino
Microb. Tech., 23, 438 (1998). (TK) bei kokosų aliejaus (peresterinimas). Po 6 reakcijos valandų
21. D. R. Lide, in: D. R. Lide (ed.), Handbook of Chemistry and pasiektas 75 % konversijos laipsnis naudojant Lipozyme RM IM ir
Physics, 85th edn., CRC Press (2004). TK substratą heksane; visais atvejais reakcija geriau vyksta heksane.
22. S. J. Kuo, K. L. Parkin, J. Am. Oil Chem. Soc., 73, 1427 Imobilizuotas preparatas Lipozyme® RM IM peresterinimo reakcijų
(1996). metu yra žymiai efektyvesnis už Palatase®, o šis pakankamai akty-
23. J. Bezard, M. Bugaut, L. De Physiologie, J. Am. Oil Chem. vus (konversijos laipsnis didesnis nei 60 %) tik vykdant esterinimo
Soc., 48, 134 (1971). reakciją heksane. Taikant atsako paviršiaus metodologiją nustatyta
24. A. I. Khuri, S. Mukhopadhyay, Wiley Interdiscip. Rev. trikaprilino peresterinimo 2-feniletanoliu konversijos priklauso-
Comput. Stat., 2, 128 (2010). mybė nuo reakcijos trukmės, temperatūros, alkoholio bei fermento
25. D. Montgomery, Design and Analysis of Experiments, 6th koncentracijos. Optimaliomis sąlygomis (trikaprilino koncentraci-
edn., John Wiley & Sons, New York (2005). ja – 0,8 M, temperatūra – 30 °C, reakcijos trukmė – 2 val., fermento
26. P. Mahapatra, A. Kumari, V. Kumar Garlapati, R. Banerjee, koncentracija – 7 %) gauta fenetiloktanoato išeiga – 80 %. Nauja
A. Nag, J. Mol. Catal. B:Enzym., 60, 57 (2009).
ir originalu, kad, palyginti su literatūroje aptinkamais duomenimis,
27. D. R. Hamsaveni, S. G. Prapulla, S. Divakar, Process Bio­
tokia didelė išeiga pasiekiama per trumpą reakcijos laiką.
chem., 36, 1103 (2001).
28. S. H. Krishna, B. Manohar, S. Divakar, N. G. Karanth,
J. Am. Oil Chem. Soc., 76, 1483 (1999).
29. J. M. Rodriguez-Nogales, E. Roura, E. Contreras, Process
Biochem., 40, 63 (2005).
30. V. C. Aragão, M. R. A. Porto, C. A. V. Burkert, S. J. Kalil,
J. F. de M. Burkert, Food Technol. Biotech., 49, 103 (2011).
31. C. Kuo, S. Chiang, H. Ju, Y. Chen, M. Liao, Y. Liu, C. Shieh,
J. Sci. Food Agric., 92, 2141 (2012).
32. J. Kim, D. H. Altreuter, D. S. Clark, J. S. Dordick, J. Am. Oil
Chem. Soc., 75, 1109 (1998).