[Link].

com European Journal of Lipid Science and Technology

Research Article
Accept e d Article
Regioselectivity and fatty acid specificity of crude lipase extracts from

Pseudozyma tsukubaensis, Geotrichum candidum and Candida rugosa†

Mickaël Laguerre1*, Madeleine Nlandu Mputu2, Benoît Brïys2, Michel Lopez2, Pierre Villeneuve3,

Eric Dubreucq1

1
Montpellier SupAgro, UMR IATE, F-34060 Montpellier, France
2
Ets. J Soufflet, Quai Sarrail, 10400 Nogent-sur-Seine, France
3
CIRAD Persyst, UMR IATE, Montpellier F-34060, France

*
Corresponding author: mickaelaguerre@[Link] (Laguerre M)

Running title: Hydrolytic performances of microbial lipases

Keywords: Microbial lipases, crude lipases, substrate specificity, chain length, regioselectivity,

hydrolysis

†
This article has been accepted for publication and undergone full peer review but has not been
through the copyediting, typesetting, pagination and proofreading process, which may lead to
differences between this version and the Version of Record. Please cite this article as doi:
[10.1002/ejlt.201600302].

© 2017 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

Received: July 19, 2016 / Revised: November 10, 2016 / Accepted: December 16, 2016

1
 [Link] European Journal of Lipid Science and Technology

Abstract

Microbial lipases are of particular interest for biotechnological applications because of their high
Accept e d Article
specificity and their ability to perform reactions at room temperature. Here are characterized

extracellular and cell-bound crude lipase extracts from the yeasts Pseudozyma tsukubaensis CBS

422.96, Candida rugosa CBS 5213, and Geotrichum candidum NRRL Y-552. Enzyme preparations

were optimally active at neutral pH and ambient temperature, and displayed different substrate

specificities. Unlike P. tsukubaensis and C. rugosa extracts, crude lipases from G. candidum were

unable to hydrolyze ethyl esters of saturated fatty acids (C8:0-C18:0). The activity of extracts from P.

tsukubaensis was maximal for the octyl chain and then decreases as the acyl chain was lengthened, while

cell-bound lipases from C. rugosa hydrolyzed any saturated fatty acid ethyl esters from C8:0 to C18:0.

Preparations from C. rugosa and P. tsukubaensis displayed a sn-1,3-regioselectivity on triolein, while

cell-bound lipases from G. candidum were non-regioselective. The extracellular crude lipases from G.

candidum exhibited a small preference for the internal position both on triolein and on six edible oils,

releasing a high proportion of DAG-1,3 (~ 50 %) compared to DAG-1,2(2,3). To the contrary, extracts

from P. tsukubaensis and C. rugosa catalyzed the formation of almost 100 % DAG-1,2(2,3) from oils.

Practical applications

Oils with a high DAG-1,3 content have attracted considerable attention as healthful foods capable of

lowering plasmatic level of TAG. Our results suggest that the crude enzymatic extracts from the wild-

type, non-genetically modified G. candidum strain might be particularly well adapted for partial

hydrolysis of oils to produce such nutraceutical diacylglycerols. In non-optimized condition, about

50/50 % of DAG-1,2(2,3)/DAG-1,3 was already obtained, especially for rapeseed oil.

2
 [Link] European Journal of Lipid Science and Technology

1. Introduction
Accept e d Article
Lipases (triacylglycerol acylhydrolase, E.C.[Link].) catalyze the hydrolysis of water-insoluble

triacylglycerols to generate free fatty acids, mono and diacylglycerols and glycerol. Beside their natural

lipolytic function, lipases can catalyze esterification and transesterification reactions and some of them

can resolve racemics [1-2]. Therefore, they have been widely used for biotechnological applications in

the food and detergent industries [3], oil processing and biodiesel production [4, 5], synthesis of

surfactants and lipophilized active molecules such as antioxidants [6], and for the preparation of

enantiomerically pure pharmaceuticals [7]. Lipases are prominent in the world enzyme market, since

they are considered the third greatest enzyme group in sale volume after proteases and amylases.

Microbial lipases have useful characteristics, such as action under mild conditions, stability in organic

solvents, high substrate specificity and regio- and enantioselectivity. Among them, cell-bound or

whole-cell lipases are attracting more and more attention due to the fact that they are naturally

immobilized [8]. As such, these lipases can be directly used in industrial cost effective processes with

improved stability, eliminating the isolation, purification, and immobilization procedures, thus

minimizing any possible loss of lipase activity. This also constitutes an efficient way to obtain and use

enzymes from non-GMO, Generally Recognized as Safe (GRAS) microorganisms.

The determination of substrate specificities of microbial lipases and optimal reactional conditions are

very helpful to find the best match between a specific lipase and a given reaction. Here, the investigation

is focused on lipase extracts obtained from three yeasts: Pseudozyma tsukubaensis CBS 422.96, C.

rugosa CBS 5213, and G. candidum NRRL Y-552.

The two latter were selected in this study because several of these lipases are commercially-available

and have been broadly documented in literature. Indeed, it is well established that G. candidum species

produces lipases that are considerably more active on unsaturated fatty acids than saturated ones [9].

Moreover, G. candidum (ATCC 34614) secretes four isoenzymes with different positional selectivities:

lipases I and II were found to be non-regioselective with triolein, whereas the lipases III and IV

hydrolyzed oleoyl esters at sn-2 position of triolein twice faster than those positioned at sn-1(3), thus

displaying an apparent sn-2-regioselectivity [10]. Regarding C. rugosa, it is known to produce multiple

3
 [Link] European Journal of Lipid Science and Technology

lipase isoenzymes with distinct differences in the selectivity toward the fatty acyl chain length [11].

According to Schmitt et al. [12], lipases from C. rugosa are globally unspecific towards a broad range
Accept e d Article
of fatty acid chains but have lower activity toward long unsaturated fatty acid. Several works have also

been dedicated to the study of the regioselectivity of C. rugosa lipases, and different results have been

observed. Ota et al. [13] found that these lipases were apparently sn-2-regioselectives—which is

probably due to their acyltransferase activity—, whereas, another group showed that C. rugosa crude

lipases do not have any positional specificity [14].

In contrast to G. candidum and C. rugosa species, there is virtually no data on the lipases produced by

Pseudozyma tsukubaensis (formerly known as Candida tsukubaensis or Cryptococcus tsukubaensis).

This yeast species from the Basidiomycota phylum was first isolated from Perilla frutescens, a flower

grown on Tsukuba Mountain in Japan. Its strains are mainly studied for their ability to produce

biosurfactants such as mannosylerythritol alkyl esters [15, 16]. However, apart from a US Patent giving

little information on a lipase [17], very few is known on the specificity of these biocatalysts. This study

thus aims at characterizing the substrate specificity (regio- and typoselectivity) of the crude lipase

extracts from P. tsukubaensis and to compare them with strains from the two well-known G. candidum

and C. rugosa species. We have also studied the effect of temperature and pH on lipase activity as well

as the regio-distribution of the diacylglycerols generated by partial hydrolysis of common edible oils.

2. Materials and methods

2.1. Chemicals and strains

The yeast strains used were Pseudozyma tsukubaensis CBS 422.96 and Candida rugosa CBS 5213 from

the Centraalbureau voor Schimmelcultures (Delft, The Netherlands), and Geotrichum candidum NRRL

Y-552 from ARS Culture (NRRL) Collection (USDA, USA).

Peptone, yeast extract, anhydrous glucose, yeast nitrogen base (YNB with amino-acids) and malt extract

were purchased from Difco Laboratories (Detroit, MI, USA). Sulphuric acid, pyridine, ethanol, and

hexane (all of analytical grade) were from Carlo Erba (Italy). Ethyl esters of oleic (C18:1), octanoic

4
 [Link] European Journal of Lipid Science and Technology

(C8:0), decanoic (C10:0), dodecanoic (C12:0), tetradecanoic (C14:0), hexadecanoic (C16:0), and

octadecanoic (C18:0) acids (all of them being > 99 % purity) were obtained from Sigma-Aldrich (Saint
Accept e d Article
Quentin, France). The corresponding free fatty acids were from the same supplier, as well as triolein

(purity grade > 98 %), N-methyl-N-tri(methylsilyl)fluoroacetamide (MSTFA) and standards for GC

analysis (methyl pentadecanoate, ethyl pentadecanoate, and acid pentadecanoic). Polyvinyl alcohol was

from Merck (München, Germany). Ultrapure water was used for culture media and buffers. Refined

rapeseed, peanut, sunflower, and soybean oils were purchased in local supermarket, as well as olive and

cod liver oils.

2.2. Strain conservation

P. tsukubaensis, G. candidum, and C. rugosa were maintained on yeast malt peptone glucose (YMPG)

conservation medium with 20 % agar. The tubes were stored at 4 °C for no more than three months. For

P. tsukubaensis, YMPG should be refreshed every month. YMPG medium consisted of 10 g/L glucose,

20 g/L yeast extract, 20 g/L peptone, 20 g/L agar in pure water. For longer storage, P. tsukubaensis, G.

candidum, and C. rugosa were suspended in glycerol:water medium (40:60 % ; w:w) and conserved at

− 80 °C in cryotubes.

2.3. Culture conditions

Three successive pre-cultures of 24 h were performed before inoculating cells on the final culture

medium. Pre-cultures and cultures were performed in Erlenmeyer flasks filled to 10 % of their volume,

covered with cotton plugs, incubated at 28 °C, and shaken at 80 oscillations per min (amplitude 7 cm)

on an agitation table. For the first pre-culture, a loopful of cells was transferred to a 100 mL Erlenmeyer

flask containing 10 mL YPG. The second one was obtained by adding 1 mL of the first pre-culture into

10 mL YNB (13.7 g/L) supplemented with 20 g/L glucose in a 100 mL Erlenmeyer flask. The third pre-

culture was obtained by adding 10 mL of the second one into 100 mL of the same medium in a 1 L

Erlenmeyer flask.

5
 [Link] European Journal of Lipid Science and Technology

The culture medium was YNB (13.7 g/L) buffered and adjusted to pH 5.5 with 100 mM

NaH2PO4/NaHPO4 buffer. The carbon source was rapeseed oil (5 g/L) sterilized by autoclaving at 120
Accept e d Article
°C for 20 min. One hundred mL of this culture medium were inoculated with 10 mL of the third pre-

culture. This final culture was performed in 1 L Erlenmeyer flasks filled to l/10 of their volume.

2.4. Preparation of crude lipase extracts

After 16, 20, and 70 h of cultivation for G. candidum, P. tsukubaensis, and C. rugosa, respectively,

fermentation musts were centrifuged (8,000 x g, 10 min). The supernatant was recovered and defatted

three times with three volume of hexane (3 x 100 mL), then used as extracellular crude lipase extract.

The cell pellets were washed three times with 33 mL of 50 mM sodium phosphate buffer, pH 6.5, for

30 sec under ultrasonic agitation, and then centrifuged at 10,000 x g for 10 min. The supernatants

obtained from this washed cell pellets were pooled, then defatted three times with hexane (3 x 100 mL).

The aqueous phase constituted the cell-bound crude lipase extract. Both extracts were stored at - 20 °C

until use.

2.5. Determination of lipase activity

Hydrolysis activity was determined by measuring the initial rates of fatty acid production (μmol/min).

One unit of activity (U) is the amount of enzyme releasing 1 µmol of free fatty acid in 1 min under the

standard assay conditions described below. Lipase activity was expressed as units/mL (U/mL) of crude

lipase extract.

Experiments were conducted as described by Vaysse et al. [18] and modified as follows. Lipid substrates

were prepared as lipid emulsion consisting of 100 mM ethyl oleate dispersed by sonication (Branson

Sonifier 250) in an aqueous solution of 20 g/L poly(vinyl alcohol). For reactions (total volume 1 mL in

glass tubes), 500 μL of crude lipase sample were added to 400 μL of 50 mM sodium phosphate buffer,

pH 6.5. Reaction was triggered by the addition of 100 μL of the lipid substrate emulsion (10 mM final

concentration) and carried out at 30 °C for 15 min then stopped by the addition of 950 μL of an

6
 [Link] European Journal of Lipid Science and Technology

ethanol/sulphuric acid (100:0.8, v/v) mixture. After the addition of 50 μL of internal standards (ethanolic

solution of pentadecanoic acid and its ethyl esters, 0.04 M), lipid-soluble compounds such as free fatty
Accept e d Article
acids and ethyl oleate were extracted with 1 mL hexane.

Esters and free fatty acids were then quantified by gas chromatography (GC). Two hundred microliters

of the hexane phase were dispensed into a GC vial equipped with a 300 µL glass insert. Fatty acids were

derivatized by silylation by adding 25 μL pyridine and 25 μL N-methyl-N-tri(methylsilyl)-

trifluoroacetamide (MSTFA) to the vial then heating at 50 °C for 20 min before analysis. A Hewlett-

Packard 5890 GC system equipped with a flame ionization detector, an automatic sampler (sample

volume 0.5 μL) and a split/splitless injector was used for analysis. The capillary column was a DB-5ht

(15 m × 0.25 mm, J&W Scientific, Massy, France). The helium carrier flow was 2 mL/min and the split

ratio was 1:18. Temperature conditions were: injector 280 °C, detector 290 °C, oven 180–195 °C at 10

°C/min.

2.6. Determination of temperature and pH effects on lipolytic activity

The pH effect on lipase activity was measured using buffers of different pH (4–9) at 30 ºC. The

employed buffer solution were citrate (pH 4–6), sodium phosphate (pH 5–9) and glycine (pH 8-10)

buffers at a final concentration of 100 mM.

Temperature influence on lipase activity was determined by carrying out the enzyme assay at different

temperatures (4, 10, 20, 30, 40, 50 and 60 °C) at pH 6.5.

2.7. Determination of fatty acyl chain length specificity on ethyl monoesters

Specificity constants 1/α [19] were determined using an equimolar mixture of six saturated ethyl esters

as substrates (C8:0, C10:0, C12:0, C14:0, C16:0, and C18:0; 1.6 μmol each in 1 mL of reaction medium)

as previously described by Vaysse et al. [18]. For reactions with saturated C8:0–C18:0 esters in

competition, GC temperatures were: injector 340 °C, detector 370 °C, oven 60–260 °C at 25 °C/min.

7
 [Link] European Journal of Lipid Science and Technology

Calibration curves were realized using lipid dispersion of mixtures of free fatty acids and ethyl

monoesters prepared according to the same protocol, without enzyme.

Accept e d Article
2.8. Determination of the DAG regio-distribution after partial hydrolysis of TAG

In a test tube, 500 µL of crude lipase sample were added to 400 µL of sodium phosphate buffer (50 mM,

pH 6.5).

(i) For regioselectivity study, lipid substrates were prepared as emulsions consisting of 100 mM triolein

(purity grade > 98 %) dispersed by sonication (Branson Sonifier 250) in an aqueous solution of 20 g/L

polyvinyl alcohol.

(ii) For the study of the regiodistribution of DAGs generated by partial hydrolysis of oils, commercially

available edible oils (rapeseed, peanut, sunflower, olive, soybean, and cod liver oils) were prepared at

0.1 g/mL in an aqueous solution of 20 g/L polyvinyl alcohol. Lipid dispersions of edible oils were

prepared as described with triolein.

Reactions were triggered by the addition of 100 μL of the lipid dispersion (10 mM final concentration

for triolein or 0.01 g/mL for edible oils). They were carried out at 30 °C for 15 min and stopped by the

addition of 1 mL of an ethanol/sulphuric acid (100:0.8, v/v) mixture. Lipid-soluble compounds such as

free fatty acids and ethyl oleate were then extracted with 2 mL hexane. Around 1.5 mL of hexane phase

was dispensed into 2 mL amber vials (stored at - 20 °C for a short period of time if necessary until use).

It is worth noting that the content of diacylglycerols (DAG) evolves with time, with a significant

conversion of DAG-1,2(2,3) into DAG-1,3 over a storage period of 6 month at - 20 °C. Hexanic phase

should thus be analyzed extemporaneously by thin layer chromatography to avoid acyl migration and to

minimize bias in the interpretation of the lipase regioselectivity.

Three µL of the hexane phases were deposited on a HPTLC silica gel plate 60 (Merck, Darmstadt,

Germany) then developed with a mobile phase of hexane:diethyl ether:formic acid ([Link]).

Chromatographic separation was performed in an Automatic Development Chamber (Camag, Muttenz,

Switzerland). Spots were revealed by dipping the HPTLC plate into an aqueous solution containing

copper sulphate (10 %), H3PO4 (8 %), and methanol (5 %), and charring at 180 °C for 10 min. Spot

8
 [Link] European Journal of Lipid Science and Technology

quantitation by densitometry was then performed using a TLC Scanner 3 (Camag) at 325 nm in

reflectance mode.
Accept e d Article
3. Results and discussion

Preliminary experiments using rhodamine B plates [20] confirmed that the P. tsukubaensis, G. candidum

and C. rugosa strains chosen were lipase producers. Indeed, hydrolytic products have been detected

from medium chain triacylglycerols (miglyol 810 N; mainly C8-C10), peanut, rapeseed, palm kernel

and coconut oils used as carbon sources (data not shown). Then, various enzymatic extracts were

produced from the cultivation of the three yeasts in 1 L Erlenmeyer flasks. Supernatants containing

extracellular lipases were found to have little or no activity for C. rugosa. On the contrary, for G.

candidum and P. tsukubaensis, extracellular lipases exhibited significant hydrolysis activity with ethyl

oleate. The cell pellets were washed with 50 mM sodium phosphate buffer pH 6.5 under ultrasonic

agitation, then centrifuged. For sake of clarity, the terms extracellular crude lipase extract are referred

below to as the first supernatant and cell-bound crude lipase extract to the second one. Both extracts

were studied for G. candidum and P. tsukubaensis, whereas only cell-bound crude lipases were analyzed

for C. rugosa.

The effects of temperature and pH on lipase activity, along with their substrate specificity in terms of

chain length preference, were performed on ethyl oleate as a model substrate. Triolein was used for the

characterization of the global regio-selectivity of the extracts. The regio-distribution of diacylglycerols

generated after partial hydrolysis of six edible oils was also determined.

3.1. Effect of temperature on lipase activity on ethyl monoester model

Figure 1 shows a similar hydrolytic behavior as a function of the temperature for the crude lipase

extracts. For the three yeasts, the apparent optimal activity was observed at 20-40 °C, with a modest

peak at 30 °C. Beyond 40 °C, hydrolysis activity collapsed and reached a near zero-value at 50 °C.

Except for the extracellular crude extract from P. tsukubaensis, no activity has been detected at 60 °C.

9
 [Link] European Journal of Lipid Science and Technology

This profile is usual for microbial lipases. Noteworthy, all our extracts were significantly active at 10

°C, and some at 4 °C. Namely, extracellular crude lipases from C. rugosa and G. candidum, along with
Accept e d Article
cell-bound crude lipases from G. candidum, can be regarded as cold active lipases keeping more than

40 % of their activity at 4 °C.

3.2. Effect of pH on lipase activity on ethyl monoester model

The hydrolysis activity increased with pH in the range of pH 4-5 to ~ 7, above which it sharply declined

over pH 8-9 for C. rugosa, and pH 10 for G. candidum. An optimum at neutral pH has been reported in

many studies, such as for Mucor sp. [21], Aspergillus niger [22], Candida rugosa [23], and Candida

deformans [24] lipases. For instance, Benjamin and Pandey [25] found that among the three isoenzyme

lipases purified from C. rugosa DSM-2031, lipases A and C had an optimal activity at pH 7 (at 37 °C),

while lipase B was more active at pH 7.5-8.5.

The influence of pH seems to be different between the two P. tsukubaensis crude lipase preparations.

For cell-bound lipase extracts, the pH effect is similar to that of C. rugosa and G. candidum extracts,

while for extracellular lipases, the maximal activity seems to be obtained in range of pH 4 to 6. This was

confirmed on the same extracts after 3 months of storage at - 20 °C.

3.3. Effect of the saturated alkyl chain length on ethyl monoester model

Chain-length selectivity of the crude lipase extracts was determined in aqueous biphasic medium.

Substrates were saturated C8:0–C18:0 ethyl esters, so that only carbon chain-length specificity was

studied. Selectivity was determined by multicompetitive reactions according to Rangheard et al. [19].

Strong differences between typo-selectivity profiles can be observed in Figure 3.

Both extracellular and cell-bound crude lipase extracts from P. tsukubaensis display an esterase activity

with a preference for ethyl caprylate (C8:0 EE). For extracellular ones, 1/α drastically decreased as acyl

chain is lengthened from C8:0 to C10:0 and C12:0, then slightly increased for C14:0, C16:0, and C18:0

(Figure 3). For cell-bound lipases, 1/α decreased with an increase of the saturated chain length.

10
 [Link] European Journal of Lipid Science and Technology

The specificity constant of the cell-bound crude lipases from C. rugosa was higher than 0.6 for almost

every chain lengths, corresponding to a low chain-length selectivity of the extract. Crude extracts from
Accept e d Article
this strain can thus be regarded as versatile enzymatic tools as they hydrolyze any saturated fatty acid

ethyl esters from C8:0 to C18:0. C. rugosa is known to produce multiple lipase isoenzymes with distinct

differences in substrate specificity, in particular with regard to selectivity toward the fatty acyl chain

length [11]. Because the cell-bound crude lipase extract studied here probably contains several

isoenzymes, it is not surprising to observe in Figure 3 the absence of specificity for a given chain length:

the addition of all the chain length specificities of all the lipase isoenzymes may finally lead to a flat

profile regarding this parameter. Moreover, according to Schmitt et al. [12], lipases from C. rugosa are

globally unspecific lipases towards a broad range of fatty acid chains but have lower activity toward

long unsaturated fatty acid.

Extracts from G. candidum did not display any activity with the investigated C8:0-C18:0 substrates. The

reaction duration was lengthened up to 60 min, without any change. Increasing the amount of enzyme

extract (extracellular and cell-bound) neither allowed any detectable hydrolysis. To the contrary, these

extracts were active on oleyl chains whether esterified to ethanol or glycerol (Figures 1-2, hydrolysis

of ethyl oleate, and Figure 4, hydrolysis of triolein). This suggests a strong selectivity of both

extracellular and cell-bound crude lipases for unsaturated vs. saturated fatty acyl chains. Indeed, as early

as 1961, Alford and Pierce [9] found that a lipase from G. candidum was considerably more active on

oleic and linoleic acids than on any of the other fatty acids in emulsions of corn oil for instance.

However, since the lipolytic activities were determined at temperatures below 0 °C, selectivity for

unsaturated lipids may have been due to higher crystallization of saturated lipids compared to

unsaturated ones. Baillargeon and Sonnet [26] showed that crude lipases from G. candidum NRRL Y-

553—which is close to the strain studied here—exhibited selectivity for unsaturated lipids (oleic,

linoleic, and linolenic acids) compared to palmitic acid in the hydrolysis of emulsified soybean oil.

However, results obtained with lipases purified from different G. candidum strains and various culture

and preparation modes, greatly differed in their substrate specificity. Using crude lipases from G.

candidum 4013, Brabcova et al., [8] did not find any selectivity for unsaturated chains with cell-bound

lipases, whereas such selectivity—although modest—was measured with extracellular lipases. When

11
 [Link] European Journal of Lipid Science and Technology

lipases from G. candidum CMICC 335426 were purified, it was shown that lipase B exhibited a

remarkable specificity for unsaturated substrates with a 9(cis) double bond, while lipase A was not
Accept e d Article
specific [27]. To complicate matter further, the same authors found that lipase II from G. candidum

ATCC 34614—that has the same native molecular masse and pI that lipase B from CMICC 335426

strain—did not exert marked selectivity for unsaturation as with lipase B. This exemplifies that the

strain, the nature of isoenzyme, the medium conditions and the extract preparation mode dramatically

impact the substrate specificity of pure or crude lipases from G. candidum. It should be kept in mind

that lipases are only comparable for strictly given conditions, i.e. specificity values cannot be considered

as universally applicable for all G. candidum lipases and reaction media.

3.4. Lipase regioselectivity on TAG model

Regioselectivity is an important parameter to properly implement lipase-catalyzed bioconversion of

lipids through a broad array of reaction pathways, the simplest being hydrolysis reaction. Thus, it has

been tested on triolein for the various crude lipase extracts.

Regarding P. tsukubaensis, a modest difference can be observed between extracellular and cell-bound

lipase extracts. The distribution between DAG-1,2(2,3) and DAG-1,3 was 88:12% for extracellular and

80:20 for cell-bound ones. This shows an intermediate behavior between sn-1,3-regioselectivity and

non-regioselectivity, extracellular lipases catalyzing more sn-1,3-hydrolysis than cell-bound lipases. As

already hypothesized, these crude extracts may contain several isoenzymes having different

regioselectivity, resulting in a global behaviour of lower selectivity than the individual enzymes. It

should also be mentioned that the only available publication on the positional specificity of a lipase from

P. tsukubaensis [17] reported a surprising 48 % of DAG-1,3 using triolein as substrate, which should

have been interpreted as showing sn-2-regioselectivity although the authors concluded that the enzyme

was non-specific. However, these data have to be interpreted with caution because along with P.

tsukubaensis CBS 6389, Ishii [17] also found a level of DAG-1,3 of 51, 45, and 50 %, which is extremely

high and unusual, with lipases from Candida humicola, Candida foliorum and Candida auricularia,

12
 [Link] European Journal of Lipid Science and Technology

respectively. It can be supposed that a migration of acyl moieties from central to external position has

probably occurred in the samples of this study.

Accept e d Article
For cell-bound crude lipase extract from C. rugosa, results depicted in Figure 4 show a net sn-1,3-

regioselectivity on triolein with a DAG-1,2(2,3)/DAG-1,3 repartition of 98:2 %. Several works have

been dedicated to the study of the regioselectivity of C. rugosa lipases, and different results have been

observed. For instance, using triolein, Ota et al. [13] found that lipases from this yeast were apparently

sn-2-regioselectives with a PSI of – 14.6, such a value coming likely from an acyltransferase activity of

the lipase. Other groups, however, showed on synthetic triacylglycerols that C. rugosa crude lipases do

not have any positional specificity [14]. Using 1,2-dipalmitoyl-3-oleylglycerol and 1,2-dioleoyl-3-

palmitoylglycerol, these authors obtained hydrolysis levels of 66-64 % in the sn-1 and sn-2 positions

and 34-36 % in the sn-3 position. Using a third triacylglycerol having palmityl chain in the central

position (1,3-dioleoyl-2-palmitoylglycerol), they measured a 32 %-hydrolysis in sn-2 position and 68 %

in both sn-1 and sn-3 positions. These two examples—seemingly inconsistent—demonstrate the

variability of the positional specificity of lipase extracts and the difficulty to extrapolate from case to

case. In this context, our result showing that C. rugosa crude lipases, prepared in the conditions of our

study, may be strongly sn-1,3-regioselective is valuable for chemists since the many articles, reviews

and books dedicated to this topic most always report that C. rugosa lipases are non regioselective.

For G. candidum extracts, cell-bound crude lipases are positionally non-specific with a DAG-

1,2(2,3)/DAG-1,3 ratio of 67/33 %. Extracellular crude extracts exhibited an hydrolysis profile between

non-regioselectivity and apparent slight sn-2-regioselectivity with a DAG-1,2(2,3)/DAG-1,3 ratio of

60/40 %. Regarding these latter extract, several putative mechanisms can explain why G. candidum

extracellular lipases are able to remove the fatty acyl moiety from the central position of triolein: (i) sn-

2-hydrolysis; (ii) random acyl migration moving chain from central to external position followed by

classical hydrolysis; (iii) lipase exhibiting acyltransferase activity [28], etc. Regardless of the exact

mechanism of action, however, the practical conclusion is that extracellular crude lipase extract from

this strain is able to produce relatively high level of DAG-1,3 (40 %) compared to DAG-1,2(2,3). In

1994, Sugihara et al. [10] demonstrated the presence of four isoenzymes for G. candidum ATCC 34614:

one major lipase (coined as I) followed by three minor lipases (coined as II, III, and IV). If their

13
 [Link] European Journal of Lipid Science and Technology

molecular masses were similar, their positional selectivity greatly varied. Lipases I and II were found to

be non-regioselective with triolein, whereas the lipases III and IV hydrolyzed oleoyl esters at sn-2
Accept e d Article
position of triolein twice faster than those positioned at sn-1(3), thus displaying an apparent sn-2-

regioselectivity.

3.5. Regiodistribution of DAG generated by partial hydrolysis of commercially available oils

Edible oils with a high DAG-1,3 content have attracted considerable attention as healthful food. Animal

and human studies suggest that ingestion of DAG-1,3 may indeed lead to a lower plasmatic TAG level

compared to TAG ingestion, with the same composition, digestibility, and energetic level [29, 30, 31,

32]. DAG-1,3 are hydrolyzed by sn-1,3-regioselective lipases in the intestinal lumen to give MAG-1

which cannot be further esterified into TAG as it is the case with MAG-2 coming from DAG-1,2(2,3).

For industrial production, however, DAG-1,3 cannot be obtained with high yield by direct chemical

synthesis because of a lack of positional selectivity. Therefore, DAG-1,3 enzymatic synthesis has

attracted considerable interest in virtue of its milder, simpler, and greener process, and higher selectivity.

Overall, there are two enzymatic strategies to produce them. (i) sn-1,3-regioselective lipases can be used

to synthesize DAG-1,3 via esterification of free fatty acids onto glycerol, via glycerolysis between

glycerol and TAGs, or via acyl transfer between TAGs and the free hydroxy groups of MAGs. (ii) The

second route is carried out with lipases that are able to produce a high proportion of DAG-1,3 from

TAGs. As such, they can be directly utilized in partial hydrolysis involving oil and water as reaction

medium. To know which crude lipase was better adapted to the first or the second strategy, we assessed

the regiodistribution (external vs. central) of fatty acyl moieties on DAGs produced by partial hydrolysis

of six common edible oils, i.e. olive oil, cod liver oil, and refined rapeseed, peanut, sunflower, and

soybean oils.

In Figure 5, it appears that P. tsukubaensis crude extracts (both extracellular and cell-bound lipases) are

able to form high levels of external DAGs like many other microbial lipases. Likewise, cell-bound

lipases from C. rugosa produce the same regiodistribution with a DAG-1,2(2,3)/DAG-1,3 ratio of 100/0

14
 [Link] European Journal of Lipid Science and Technology

% for all the investigated oils. In contrast, both extracellular and cell-bound lipase extracts from

G. candidum lead to the formation of 55 and 45 % of DAG-1,2(2,3) and DAG-1,3, respectively.

Accept e d Article
Overall, these results confirm the trends already observed on homogenous lipid substrate (i.e. pure

triolein), and provide valuable information to know which strategy is to be carried out with this or that

crude lipase extract. For example, crude lipases from C. rugosa and P. tsukubaensis seem to be able to

catalyze the esterification of fatty acyl moieties at external position, leading to the formation of high

levels of DAG-1,3. Further studies should thus be conducted to compare these crude lipase preparations

with well-known sn-1,3-regioselective lipases that have been purified and used with this purpose,

especially those from Chromobacterium viscosum, Rhizopus delemar, Rhizomucor miehei [33], as well

as commercially available recombinant phospholipases/lipases such as Lecitase® Ultra [34]. Our results

also suggest that crude enzymatic extracts from G. candidum are particularly well adapted for partial

hydrolysis of oils since in non-optimized conditions, about 50/50 % of DAG-1,2(2,3)/DAG-1,3 were

obtained, especially for rapeseed oil (Figure 5).

4. Conclusion

Crude lipase extracts studied here are optimally active at neutral pH and ambient temperatures. Lipolytic

extracts from P. tsukubaensis CBS 422.96 and C. rugosa CBS 5213 were able to hydrolyze esters of

saturated and unsaturated fatty acids on both ethyl ester and triacylglycerol models. They also exhibited

hydrolysis preference for external positions (sn-1,3) on pure triolein and edible oils. In contrast, due to

their high selectivity for unsaturated acyl esters, the lipase extracts from G. candidum NRRL Y-552 can

be used in their crude form as tools for the isolation of pure unsaturated fatty acids and for the production

of novel triacylglycerols by transesterification reactions. On the other hand, these lipases display

particular regioselectivity enabling to envision the preparation of high quantities of DAG-1,3 through

partial hydrolysis of common edible oils. Finally, crude lipase extracts prepared from these three wild-

type yeasts appear as versatile enzymatic tools for various biotechnological applications in lipid

modification.

15
 [Link] European Journal of Lipid Science and Technology

The authors have declared no conflict and competing financial interest.

Accept e d Article
5. References

[1] Zorn, K., Oroz-Guinea, I., Brundiek, H., Bornscheuer, U.T., Engineering and application of enzymes

for lipid modification, an update. Prog. Lipid Res. 2016, 63, 153-164.

[2] Bornscheuer, U.T. Enzymes in lipid modification: past achievements and current trends. Eur. J. Lipid

Sci. Technol. 2014, 116, 1322-1331.

[3] Sanchez, M., Prim, N., Rendez-Gil, F., Pastor, F. I. J., Diaz, P., Engineering of baker’s yeasts,

Escherichia coli and Bacillus hosts for the production of B. subtilis lipase A. Biotechnol Bioeng. 2002,

78, 339-345.

[4] Shimada, Y., Watanabe, Y., Sugihara, A., Tominaga, Y., Enzymatic alcoholysis for biodiesel fuel

production and application of the reaction to oil processing. J. Mol. Catal. B: Enzym. 2002, 17, 133–

142.

[5] Li, L., Du, W., Liu, D., Wang, L., Li, Z., Lipase-catalyzed transesterification of rapeseed oils for

biodiesel production with a novel organic solvent as the reaction medium. J. Mol. Catal. B: Enzym.

2006, 43, 58–62.

[6] López Giraldo, L. J., Laguerre, M., Lecomte, J., Figueroa-Espinoza, M. C., Barouh, N., Baréa, B.,

Villeneuve, P., Lipase-catalyzed synthesis of chlorogenate fatty esters in solvent-free medium. Enzyme

Microb. Technol. 2007, 41, 721-726.

16
 [Link] European Journal of Lipid Science and Technology

[7] Mustranta, A., Use of lipases in the resolution of racemic ibuprofen. Appl. Microbiol. Biotechnol.

1992, 38, 61-66.

Accept e d Article
[8] Brabcova, J., Zarevucka, M., Mackova, M., Differences in hydrolytic abilities of two crude lipases

from Geotrichum candidum 4013. Yeast 2010, 27, 1029-1038.

[9] Alford, J. A., Pierce, D. A., Lipolytic activity of microorganisms at low and intermediate

temperatures. III. Activity of microbial lipases at temperatures below 0 °C. J. Food Sci. 1961, 26, 518-

524.

[10] Sugihara, A., Shimada, Y., Nakamura, M., Nagao, T., Tominaga, Y., Positional and fatty acid

specificities of Geotrichum candidum lipases. Protein Eng. 1994, 7, 585-588.

[11] Brocca, S., Secundo, F., Ossola, M., Alberghina, L., Carrea, G., Lotti, M., Sequence of the lid

affects activity and specificity of Candida rugosa lipase isoenzymes. Protein Sci. 2003, 12, 2312-2319.

[12] Schmitt, J., Brocca, S., Schmid, R. D., Pleiss, J., Blocking the tunnel: engineering of Candida

rugosa lipase mutants with short chain length specificity. Prot. Engineering 2002, 15, 595-601.

[13] Ota, Y., Itabashi, Y., Hasuo, M., Measurement of positional specificity index of microbial lipases

by chiral phase high-pressure liquid chromatography. Biosc. Biotechnol. Biochem. 1996, 60, 145-146.

[14] Benzonana, G., Esposito, S., On the positional and chain specificities of Candida cylindracea

lipase. Biochim. Biophys. Acta 1971, 231, 15-22.

[15] Fukuoka, T., Morita, T., Konishi, M., Imura, T., Kitamoto, D., A basidiomycetous yeast,

Pseudozyma tsukubaensis, efficiently produces a novel glycolipid biosurfactant. The identification of a

new diastereomer of mannosylerythritol lipid-B. Carbohydrate Res. 2008, 343, 555-560.

17
 [Link] European Journal of Lipid Science and Technology

[16] Morita, T., Takashima, M., Fukuoka, T., Konishi, M., Imura, T., Kitamoto, D., Isolation of
Accept e d Article
basidiomycetous yeast Pseudozyma tsukubaensis and production of glycolipid biosurfactant, a

diastereoisomer type of mannosylerythritol lipid-B. Appl. Microbiol. Biotechnol. 2010, 88, 679-688.

[17] Ishii, M., Thermally stable and positionally non-specific lipase isolated from Candida. US Patent,

1993, N° 5,273,898.

[18] Vaysse, L., Ly, A., Moulin, G., Dubreucq, E., Chain-length selectivity of various lipases during

hydrolysis, esterification and alcoholysis in biphasic aqueous medium. Enzyme Microb. Technol. 2002,

31, 648-655.

[19] Rangheard, M. S., Langrand, G., Triantaphylides, C., Baratti, J. C., Multi-competitive enzymatic

reactions in organic media: a simple test for the determination of lipase fatty acid specificity. Biochim.

Biophys. Acta 1989, 1004, 20–28.

[20] Kouker, G., Jaeger, K. E., Specific and sensitive plate assay for bacterial lipases. Appl. Environ.

Microbiol. 1987, 53, 211-213.

[21] Abbas, H., Hiol, A., Deyris, V., Comeau, L., Isolation and characterization of an extracellular lipase

from Mucor strain isolated from palm fruit. Enzyme Microb. Technol. 2002, 31, 968-978.

[22] Kamini, N. R., Mala, J. G. S., Puvanakrishnan, R., Lipase production from Aspergillus niger by

solid-state fermentation using gingelly oil cake. Process Biochem. 1998, 33, 505-511.

[23] Fadiloglu, S., Soylemez, Z., Kinetics of lipase-catalyzed hydrolysis of olive oil. Food Res. Int.

1997, 30, 171-175.

18
 [Link] European Journal of Lipid Science and Technology

[24] Boutur, O., Dubreucq, E., Galzy, P., Factors influencing ester synthesis catalyzed in aqueous media

by the lipase from Candida deformans (Zach) Langeron and Guerra. J. Biotechnol. 1995, 42, 23-33.
Accept e d Article
[25] Benjamin, S., Pandey, A., Isolation and characterization of three distinct forms of lipases from

Candida rugosa produced in solid state fermentation. Braz. Arch. Biol. Technol. 2001, 44, 213-221.

[26] Baillargeon, M. W., Sonnet, P., Selective lipid hydrolysis by Geotrichum candidum NRRL Y-553

lipase. Biotechnol. Lett. 1991, 13, 871-874.

[27] Sidebottom, C. M., Charton, E., Dunn, P. P. J., Mycock, G., Davies, C., Sutton, J. L., Macrae, A.

R., Slabas, A. R., Geotrichum candidum produces several lipases with markedly different substrate

specificities. Eur. J. Biochem. 1991, 202, 485-491.

[28] Briand, D., Dubreucq, E., Galzy, P., Functioning and regioselectivity of the lipase of Candida

parapsilosis (Ashford) Langeron and Talice in aqueous medium. New interpretation of regioselectivity

taking acyl migration into account. Eur. J. Biochem. 1995, 228, 169-175.

[29] Murata, M., Hara, K., Ide, T., Alteration by diacylglycerols of the transport and fatty acid

composition of lymph chylomicron in rats. Biosci. Biotechnol. Biochem. 1994, 58, 1416-1419.

[30] Taguchi, H., Watanabe, H., Onizawa, K., Nagao, T., Gotoh, N., Yasukawa, T., Tsushima, R.,

Shimasaki, H., Itakura, H., Double-blind controlled study on the effects of the dietary diacylglycerol on

postprandial serum and chylomicron triacylglycerol responses in healthy humans. J. Am. Coll. Nutr.

2000, 19, 789-796.

[31] Taguchi, H., Nagao, T., Watanabe, H., Onizawa, K., Matsuo, N., Tokimitsu, I., Itakura, H., Energy

value and digestibility of dietary oil containing mainly 1,3-diayclglycerol are similar to those of

triacylglycerol. Lipids 2001, 36, 379-382.

19
 [Link] European Journal of Lipid Science and Technology

[32] Takase, H., Shoji, K., Hase, T., Tokimitsu, I., Effect of diacylglycerol on postprandial lipid
Accept e d Article
metabolism in non-diabetic subjects with and without insulin resistance. Atherosclerosis 2005, 180, 197-

204.

[33] Berger, M., Laumen, K., Schneider, M. P., Enzymatic esterification of glycerol I. Lipase-catalyzed

synthesis of regioisomerically pure 1,3-sn-diacylglycerols. J. Am. Oil Chem. Soc. 1992, 69, 955-960.

[34] Liu, N., Wang, Y., Zhao, Q., Zhang, Q., Zhao, M., Fast synthesis of 1,3-DAG by Lecitase® Ultra-

catalyzed esterification in solvent-free system. Eur. J. Lipid Sci. Technol. 2011, 113, 973–979.

20
 [Link] European Journal of Lipid Science and Technology

Figure legends
Accept e d Article
Figure 1. Effect of the reaction medium temperature on the hydrolysis activity of crude lipase

extracts from P. tsukubaensis (A), C. rugosa (B), and G. candidum (C). Reactions were

performed during 15 min at 30 °C, pH 6.5 in the presence of 10 mM ethyl oleate emulsified

with a polyvinylic alcohol solution (2 %, v). Data points and error bars represent means (n=3)

± standard deviations. Error bars lie within data points.

Figure 2. Effect of the reaction medium pH on hydrolysis activity of crude lipase extracts from

P. tsukubaensis, C. rugosa, and G. candidum. The three buffer solutions used were (in final

concentration), 100 mM citrate buffer pH 4-6, 100 mM phosphate buffer solution pH 4-9; and

100 mM glycine/NaOH pH 8-10. Reactions were performed during 15 min at 30 °C, pH 6.5 in

the presence of 10 mM ethyl oleate emulsified with a polyvinyl alcohol solution (2 %, v). Data

points and error bars represent means (n=3) ± standard deviations. Error bars lie within data

points.

Figure 3. Relative specificity constants 1/α for hydrolysis of ethyl esters of saturated fatty acids

C8–C18 catalyzed by P. tsukubaensis CBS 422.96 (A) and C. rugosa CBS 5213 (B) crude

lipases. Reactions were performed during 15 min at 30 °C, pH 6.5 in the presence of 10 mM

ethyl esters mixture (1.6 mM each ester) emulsified with a polyvinyl alcohol solution (2 %, v).

Data points and error bars represent means (n=3) ± standard deviations.

Figure 4. Regioselectivity (on triolein) of crude lipases from P. tsukubaensis, G. candidum,

and C. rugosa. Reactions were performed during 15 min at 30 °C, pH 6.5 in the presence of

10 mM triolein emulsified with a polyvinylic alcohol solution (2 %, v). Data points and error

bars represent means (n=3) ± standard deviations. Error bars lie within data points.

21
 [Link] European Journal of Lipid Science and Technology
Accept e d Article
Figure 5. Regiodistribution of diacylglycerol generated by partial hydrolysis of six edible oils

by crude lipases prepared from P. tsukubaensis, G. candidum, and C. rugosa. Reactions were

performed during 15 min at 30 °C, pH 6.5 in the presence of 0.1 g/mL oil emulsified with a

polyvinylic alcohol solution (2 %, v). Data points and error bars represent means (n=3) ±

standard deviations. Error bars lie within data points.

22
 [Link] European Journal of Lipid Science and Technology
Accept e d Article
(A) P. tsukubaensis extracts (B) C. rugosa extracts (C) G. candidum extracts
100 100 100
Relative activity (%)

80 80 80

60 60 60

40 40 40

20 Extracellular lipases 20 20 Extracellular lipases

Cell-bound lipases Cell-bound lipases Cell-bound lipases
0 0 0
0 10 20 30 40 50 60 0 10 20 30 40 50 60 0 10 20 30 40 50 60
Temperature (°C) Temperature (°C) Temperature (°C)

Figure 1.

23
 [Link] European Journal of Lipid Science and Technology

(A) P. tsukubaensis (extracellular) (B) P. tsukubaensis (cell-bound)

Accept e d Article
100 100
Relative activity (%)

Citrate Citrate
80 Phosphate 80 Phosphate
Glycine Glycine
60 60

40 40

20 20

0 0
3 4 5 6 7 8 9 10 11 3 4 5 6 7 8 9 10 11

pH pH
(C) G. candidum (extracellular) (D) G. candidum (cell-bound)

100 100
Relative activity (%)

80 80

60 60

40 40
Citrate Citrate

20 Phosphate 20
Phosphate
Glycine Glycine
0 0
3 4 5 6 7 8 9 10 11 3 4 5 6 7 8 9 10 11

pH pH
(E) C. rugosa (cell-bound)

100 Citrate
Relative activity (%)

Phosphate
80 Glycine

60

40

20

0
3 4 5 6 7 8 9 10 11

pH

Figure 2.

24
 [Link] European Journal of Lipid Science and Technology

(A) P. tsukubaensis extract Extracellular crude lipases

Cell-bound crude lipases
Accept e d Article
1.0

0.8

0.6

0.4
Specificity constant (1/α)

0.2

0.0

(B) C. rugosa extract Cell-bound crude lipases

1.0

0.8

0.6

0.4

0.2

0.0
C8 C10 C12 C14 C16 C18

Figure 3. (new file as requested)

25
 [Link] European Journal of Lipid Science and Technology

100 98
DAG-1,2(2,3)
88
Diacylglycerol regiodistribution (%)
Accept e d Article
DAG-1,3
80
80

67

60
60

40
40
33

20
20
12

2
0
P. tsukubaensis P. tsukubaensis G. candidum G. candidum C. rugosa
(extracellular) (cell-bound) (extracellular) (cell-bound) (cell-bound)

Figure 4.

26
 [Link] European Journal of Lipid Science and Technology

(A) P. tsukubaensis (extracellular) (B) P. tsukubaensis (cell-bound)

DAG regiodistribution (%)
Accept e d Article
DAG-1,2(2,3) DAG-1,2(2,3)
DAG-1,3 DAG-1,3
100 100 100 100 100
100 100 100 100 100 100 95
100 100

80 80

60 60

40 40

20 20 5
0 0 0 0 0 0 0 0 0 0 0
0 0

(C) G. candidum (extracellular) (D) G. candidum (cell-bound)

DAG regiodistribution (%)

100 100
DAG 1,2 (1,3) DAG-1,2(2,3)
80 DAG 1,3 80 DAG-1,3
59
55 56 55 55 55 54 54
60 49 51 50 50 60 53 52
45 47 45 46 46 48
45 44 45
41
40 40

20 20

0 0
DAG regiodistribution (%)

(E) C. rugosa (cell-bound)

DAG-1,2(2,3)
DAG-1,3
100 100 100 100 100
100 85
80
60
40
15
20
0 0 0 0 0
0

Figure 5. (new file as requested)

27