Revealing the Roles of Subdomains in the Catalytic Behavior of Lipases/Acyltransferases Homologous to CpLIP2 through Rational Design of Chimeric Enzymes

Article abstract of DOI:10.1002/cbic.201600672

The lipases/acyltransferases homologous to CpLIP2 of Candida parapsilosis efficiently catalyze acyltransfer reactions in lipid/water media with high water activity (a_W>0.9). Two new enzymes of this family, CduLAc from Candida dubliniensis and CalLAc8 from Candida albicans, were characterized. Despite 82 % sequence identity, the two enzymes have significant differences in their catalytic behaviors. In order to understand the roles played by the different subdomains of these proteins (main core, cap and C-terminal flap), chimeric enzymes were designed by rational exchange of cap and C-terminal flap, between CduLAc and CalLAc8. The results show that the cap region plays a significant role in substrate specificity; the main core was found to be the most important part of the protein for acyltransfer ability. Similar exchanges were made with CAL-A from Candida antarctica, but only the C-terminal exchange was successful. Yet, the role of this domain was not clearly elucidated, other than that it is essential for activity.

Full text of DOI:10.1002/cbic.201600672

Accepted Article
Title: Enlightening the role of subdomains in the catalytic behavior of
lipases/acyltransferases homologous to CpLIP2 through rational
design of chimeric enzymes
Authors: Anne-Hélène Jan, Eric Dubreucq, and Maeva Subileau
This manuscript has been accepted after peer review and appears as an
Accepted Article online prior to editing, proofing, and formal publication
of the final Version of Record (VoR). This work is currently citable by
using the Digital Object Identifier (DOI) given below. The VoR will be
published online in Early View as soon as possible and may be different
to this Accepted Article as a result of editing. Readers should obtain
the VoR from the journal website shown below when it is published
to ensure accuracy of information. The authors are responsible for the
content of this Accepted Article.
To be cited as: ChemBioChem 10.1002/cbic.201600672
Link to VoR: http://dx.doi.org/10.1002/cbic.201600672
A Journal of
www.chembiochem.org

10.1002/cbic.201600672
ChemBioChem
FULL PAPER
Enlightening the role of subdomains in the catalytic behavior of
lipases/acyltransferases homologous to CpLIP2 through rational
design of chimeric enzymes
Anne-Hélène Jan,[a] Éric Dubreucq,[a] and Maeva Subileau*[a]
Abstract: The lipases/acyltransferases homologous to CpLIP2 of
Candida parapsilosis catalyze efficiently acyltransfer reactions in
rates in greener conditions [4]. Within this aim our research team
has evidenced a group of lipases/acyltransferases homologous
to CpLIP2 from Candida parapsilosis [4a, 5]. CpLIP2 was the first
lipase/ acyltransferase described [6], on the basis of which we
have produced and characterized several mutants and natural
variants [4a, 7]. Recently we have described a simple methodology
to assess the acyltransferase potential of lipases that allowed
lipid/water media with high water activity (aW
> 0.9). The
characterization of two new enzymes of this family, CduLAc from
Candida dubliniensis and CalLAc8 from Candida albicans, was
performed. Despite their 82% identity, these 2 enzymes have
significant differences in their catalytic behavior. To understand the
role played by the different subdomains of these proteins (the main
core, the cap and the C-term flap), chimeric enzymes were designed
with rational exchanges of cap and C-term flap, between CduLAc
consistent sorting of the tested enzymes in three classes, with
[4a]
high,
medium
and
low
acyltransferase
ability
.
Lipases/acyltransferases display medium to high acyltransferase
and CalLAc8. Results showed that the cap region played
a
activity and are phylogenetically related to the Candida albicans
lipase-like family of the lipase engineering database (LED)
with more than 55 % of identity in their primary sequences
(lipases with high sequence similarity are assigned to a single
homologous family in the LED). They also share low, but
[8]
significant role in the substrate specificity and the main core was
found to be the most important part of the protein in the acyltransfer
ability. Similar exchanges were made with CAL-A from
Moesziomyces (Candida) antarctica but only the C-term exchange
was successful. Yet, the role of the C-term was not clearly
evidenced, except that it is essential for the activity.
significant sequence similarity with the LED superfamily of
[8]
Moesziomyces (Candida) antarctica lipase
A
(CAL-A)
,
although CAL-A was shown to display only low acyltransferase
ability [4a]. Apart from their exceptional property to catalyze the
transesterification reaction at higher rate than hydrolysis in
aqueous media (aw > 0.9), the lipases/acyltransferases display
various substrate specificities: medium or long chain selectivity
for the acyl-donor substrate, a preference for unbranched
primary alcohols as acyl-acceptor [9]. In addition, CpLIP2 was
Introduction
The development of sustainable processes for the production of
fuels and chemicals has become a challenge for research and
industry. This is particularly true in oleochemistry where
traditional technologies require abundant organic solvent usage
and harsh reaction conditions [1]. To overcome these customary
routes, the utilization of lipases offers interesting alternative
Lipases (E.C. 3.1.1.3) naturally catalyze the hydrolysis of
triacylglycerols (TAG) into glycerol and free fatty acids (FFA) and
can be used for synthesis reactions, mostly in conditions where
hydrolysis is limited (in organic solvent). Lipases can exhibit
wide substrate spectrum, good chemo-, regio- and
stereoselectivity and eventually high stability in organic media
also shown to be able to catalyze the acyltransfer reaction to
[10]
other nucleophiles than alcohols, such as peptide
or
[11]
hydroxylamine
.
Like
other
standard
lipases,
[2]
.
lipases/acyltransferases could thus be implemented in a wide
range of applications.
Lipases of the CAL-A LED superfamily[8] exhibit no significant
sequence similarity with other LED superfamilies and are
predicted to share with CAL-A its structural peculiarities[7a, 12]
:
CAL-A is a globular class Y α/β serine hydrolase of 436 amino-
acids, and possesses, in addition to the main core of the protein,
two distinct sub-domains called the “cap” and the “flap” [12-13]. In
comparison to other α/β hydrolases, CAL-A and its homologs
have very divergent sequences, particularly in the cap region
[3]
.
These extensive properties make them valuable biocatalyst for
industrial applications.
Lipase engineering first aimed at improving their properties
toward tough reaction conditions mimicking the traditional
oleochemistry processes [1]. More recently, the possibility to use
peculiar lipases that preferentially catalyze transesterification
reactions in water as solvent and in mild-conditions has
appeared very promising to reach higher reaction yields and
which is exceptionally long (about 100 residues) [12]
.
The CAL-A cap was described to comprise a hydrophobic
tunnel-like binding site for the acyl moiety, comparable to that of
the Candida rugosa lipases (CRL) [12], and mutations in the cap
domain indeed allowed modification in the substrate specificity
profiles[4b]. The CAL-A flap is composed of 2 β-strands in the
[12]
C-terminal region and was also compared to the lid in CRL
.
[a]
Dr. A.-H. Jan, Prof. E. Dubreucq, Dr. M. Subileau
Montpellier Supagro, UMR 1208 IATE
This flap was recently proposed to play a “lid-like” role in the
interfacial activation of the enzyme [13b]
.
2 place Viala, 34060 Montpellier cedex 2, France
E-mail: maeva.subileau@supagro.fr
To better understand the role and importance of these
2 structural subdomains in substrate and reaction specificities of
the lipases and lipases/acyltransferases of the CAL-A LED
Supporting information for this article is given via a link at the end of
the document.
This article is protected by copyright. All rights reserved.

10.1002/cbic.201600672
ChemBioChem
FULL PAPER
superfamily [8], we decided to perform “subdomains exchanges”
by rational design of chimeric genes. To generate these
enzymes, the main challenge was to find biocatalysts presenting
differences in their catalytic behaviour but also sharing enough
identity in their primary sequences to allow the production of
Recently, 2 new lipases/acyltransferases were identified by our
[7b]
group: CduLAc from Candida dubliniensis
and CalLAc8 from
Candida albicans [4a]. These lipases share 82 % identity in their
primary sequence and 56 % and 58 % with CpLIP2, for CalLAc8
and CduLAc respectively. Like CpLIP2, CalLAc8 and CduLAc
3D models have a structure similar to that of CAL-A, with a α/β
hydrolase fold that holds a catalytic triad (serine, histidine,
aspartic acid), a cap and a flap subdomains.
functional
chimeric
enzymes.
Searching
in
the
lipases/acyltransferases family, we have identified two new
enzymes meeting these requirements: CduLAc from
C. dubliniensis and CalLAc8 from C. albicans. A similar strategy
has already been successfully conducted in lipases from
The high identity in their sequences and structures conjugated
with differences of reactivity and specificity that will be presented
in the following characterization, make of these two biocatalysts
models of great interest for the elucidation of structural
determinants involved in some of the lipases/acyltransferases
peculiarities, and to generate the chimeric enzymes with rational
sub-domains exchanges.
[14]
[15]
C. rugosa (CRL)
,
Bacillus subtilis (lipase A)
, and
M. (C.) antarctica (lipase B) [16]. Compared to the wild enzymes,
when functional, these chimeric biocatalysts exhibited changes
in specific activity and substrate specificity, confirming the
important role of the lid in the stability and the selectivity of the
enzymes [15]. In these studies the swapped sequence comprised
less than 30 amino-acids. Here we describe functional
exchanges of about 10 to 100 residues sequences that allowed
the modulation of the stability and substrate specificity of the
enzymes. Our results also give interesting insights into the
putative respective functions of the cap and flap in this original
group of lipases.
Design of chimeric enzymes to assess the role of the
subdomains in the lipases/acyltransferases
Referring to the structure of CAL-A (2VEOA.PDB) and CpLIP2
3D model [7a], sequence alignment and 3D modelling suggest
that, within the proteins of the LED superfamily[8], 3 subdomains
[7a,
can be identified: the main core, the cap and the C-term flap
12-13]. The cap comprises about 100 residues organized in
α helices (6 in CAL-A) and the flap domain consists of a 10 to 25
amino-acids C-terminal sequence. Within the variants of the
CAL-A LED superfamily, the most variable regions are those of
these putative caps and flaps (Table S1). Besides, as said
earlier, the CAL-A flap was compared to the lid of C. rugosa
Results and Discussion
Choosing the candidate enzymes for rational exchanges of
sub-domains: CduLAc and CalLAc8
Figure 1 Design of chimeric enzymes based on the wild-type CalLAc8, CduLAc and CAL-A.
* no lipase activity detected in the screening conditions
This article is protected by copyright. All rights reserved.

10.1002/cbic.201600672
ChemBioChem
FULL PAPER
lipases, which covers the active site and prevents direct access
Characterisation of the wild and chimeric enzymes to
assess structure/function relationships of subdomains.
to the substrate binding site[12]
.
In the CAL-A LED superfamily, the C-term flaps are variable in
composition and length (Table S1, Figure S1). In the
lipases/acyltransferases (CpLIP2, CtroL4, CalLAc8, CduLAc and
CalLAc5) the C-term is 10 to 15 residues longer than in the
classical lipases such as CAL-A or AflaL0[9] and is not
structurally well defined. It is somewhat reminiscent of the
mobile lid-like α-helix of the CAL-B family that displays low
similarity within the family, and only few contacts with the rest of
the structure[16]. We hypothesized that the main core, cap and
flap of the studied enzymes could play different roles in the
substrate and reaction specificities.
Transesterification activity in aqueous medium
125
CduLAc
CalLAc8_cap
CpLIP2
CalLAc8
CalLAc8_ct
25
CalLAc8
_cap_ct
5
III
II
I
1
CAL-A
To investigate how each sub-domain is involved in the catalytic
behavior of the lipases/acyltransferases, the chosen strategy
was to use CduLAc and CalLAc8 as models, because they are
very similar in sequence but exhibit significant differences in
their acyltransfer ability and chain-length selectivity. Based on
their sequences and 3D models, we created chimeric enzymes
with rational exchanges of sub-domains (caps and flaps)
(figure 1). CAL-A was added to the study for the creation of
additional chimeric enzymes (figure 1). CAL-A served as
reference to evaluate the transferability of the chosen strategy to
enzymes more variable in sequence. It also appeared
particularly interesting because the catalytic behavior of CAL-A
is significantly different from those of the lipases/
acyltransferases: its acyltransfer ability is low, its substrate
specificity is centered on myristic acid (C14:0), and it exhibits
exceptional thermostability. The overall strategy could allow the
production of new biocatalysts with interesting specificities.
Comparison of the catalytic characteristics of the wild and
chimeric enzymes could facilitate the elucidation of
structure/function relationships within the subdomains and the
identification of hot spots that could be targeted to rationally
modify the catalytic specificities.
More precisely, regarding the C-term flap, its beginning was
determined based on sequences alignments of the proteins of
the family (Figure S1). We selected the most variable region that
corresponded to the last 29 amino acids in CduLAc. To observe
the influence of the C-term flap and its length, we designed a
mutant of CduLAc “without C-term”, lacking the 29 amino acids
of the C terminal end (CduLAc_woC). We also substituted the C-
term of CalLAc8 and CAL-A with the one of CduLAc
(CalLAc8_ct and CAL-A_ct, respectively).
The cap was defined based on the description of the structure of
CAL-A by Widmann et al [12]. To study the role played by the cap
in the substrate specificity, we designed two chimers: the cap of
CduLAc was grafted on the main cores of CalLAc8 and CAL-A
(CalLAc8_cap and CAL-A_cap, respectively). Unfortunately, no
lipase activity was detected for CAL-A_cap, perhaps because
these two subdomains were not compatible, the caps of CAL-A
and CduLAc being too different (only 12% id). At last, to
evaluate the conjugated effect of both C-term flap and cap, we
built a mutant with the main core of CalLAc8 and the cap and
C-term flap of CduLAc (CalLAc8_cap_ct). All the mutants were
subcloned in Komagataella (Pichia) pastoris X33 and functional
enzymes were produced and characterized, except for
CAL-A_cap.
CAL-A_ct
0,2
0,04
CduLAc_woC
1000
0
500
1500
(kt/kh)app
Figure 2. Efficiency of the enzymes: rt max of the lipases (log scale) in function
of (kt/kh)app I, II and III areas correspond to high, medium and low
acyltransferase ability. Detailed results are given in table 1. The most efficient
biocatalysts for acyltransfer activity shall exhibit high rt maxt and high (kt/kh)app
.
.
First, we evaluated the influence of the change of subdomains
on the acyltransfer ability of the enzymes (figure 2, table 1). The
acyltransfer activity of CalLAc8 and the chimeric enzymes was
assessed precisely with ethyl oleate in aqueous emulsion in the
presence of various methanol concentrations between 0 and
4.5 M at 30°C, and compared with those of CduLAc and CAL-A
previously published [7b, 9]. Each enzyme was characterized by (i)
the transfer ratio (TR) at the various methanol concentrations
tested, that corresponds to the relative initial rate of alcoholysis
compared to the sum of initial rates of alcoholysis plus
competitive hydrolysis, (ii) the alcohol concentration required to
reach 50 % of TR ([MeOH]TR50) and (iii) the ratio of the apparent
rates constant of acyltransfer versus hydrolysis (kt/kh)app
Figure 2 represents the efficiency of the enzymes to catalyze
acyltransfer. The higher the rt (maximum specific transfer
[4a, 17]
.
max
rate) and (kt/kh)app are, the more efficient the lipase is. As shown
on figure 2, despite 82 % identity in their primary sequences,
CalLAc8 is less efficient to catalyze acyltransfer than CduLAc
with ethyl oleate as a substrate. Indeed, its [MeOH]TR50 is
5 times higher and its rt max 3 times lower than those of CduLAc.
CalLAc8 belongs to the class II lipases/ acyltransferases [4a], that
are characterized by medium acyltransferase ability with
(kt/kh)app between 100 and 1000 and [MeOH]TR50 between 0.05
and 0.5 M, like the lipase/acyltransferase of reference CpLIP2.
In comparison, CduLAc belongs to class I (high acyltransferase
ability), exhibiting (kt/kh)app superior to 1000 and [MeOH]TR50
inferior to 0.05 M. Moreover, as methanol concentration
increased up to 1 M, the specific activity of CduLAc was
enhanced, while for CalLAc8 the maximal activity was reached
with no methanol (figure S2) and decreased when methanol was
added.
This article is protected by copyright. All rights reserved.

10.1002/cbic.201600672
ChemBioChem
FULL PAPER
and the modification of amino acids composition did not have
much impact. In comparison, we observed that the mutant
CAL-A_ct was characterised by rt max in the same range as that
of CAL-A, but that the (kt/kh)app of this chimeric biocatalyst was
1.5 times higher than that of the wild-type enzyme. It thus
appeared that, in the case of CAL-A_ct the change in
composition and length of the C-term flap (14 additional residues
compared to CAL-A) had a significant impact on the acyltransfer
ability. Therefore it appears that the C-term somewhat plays a
role in the acyltransfer ability and that the overall size of this
subdomain might be more important than the amino-acid
composition. In addition, although the exchanges of C-term
modified numerous amino-acids, its effect was nothing
comparable to those previously described with only punctual
modifications of the amino acids comprised in the nucleophile
Table 1. Results of the experimental determination of kinetic constants and
acyltransfer abilities of the enzymes: (i) the ratios of the apparent rate
constants of acyltransfer versus hydrolysis (kt/kh)app and (kt/kh)app’ are given
(respectively related the nucleophiles concentrations and thermodynamic
activities), as well as (ii) methanol concentrations and (iii) corresponding
thermodynamic activities of methanol (aMeOH) and water (aW) at TR50 and (iv)
maximum specific transfer rate (rt max). Based on the (kt/kh)app and [MeOH]TR50
,
the enzymes are graded in categories of transfer ability: class
3
I
lipases/acyltransferases with [MeOH]TR50 inferior to 0.05 M (in green), class II
lipases/acyltransferases with [MeOH]TR50 between 0.05 and 0.5M (in blue),
class III lipases with [MeOH]TR50 superior to 0.5M (in red). Reactions were
conducted in phosphate buffer (pH 6.5, aW> 0.9) at 30°C, in the presence of 10
µmol.mL-1 ethyl oleate and 0.01-4.5 M methanol.
at TR50
Class
Enzyme
rt max
(kt/kh) (kt/kh) [MeOH]
aMeOH aW
0.002 0.999
0.003 0.999
0.005 0.998
0.008 0.997
0.008 0.996
0.008 0.996
(µmol.min-1.mg-1
)
(M)
app
app'
pocket [7c]
.
1357
607
0.04
0.08
0.13
0.19
0.20
0.20
I
CduLAc
121
0.2
60.4
77
To evaluate the influence of the cap on the acyltransfer activity,
we studied the mutant CalLAc8_cap. While the value of (kt/kh)app
was similar for CalLAc8 and CalLAc8_cap, the rt max was 2 times
higher. Looking at the specific activities as a function of the
methanol concentration, we observed that the increase of
activities was proportional for hydrolysis and transesterification,
with comparable profile (Figure S2). The difference of
composition between CalLAc8 and CalLAc8_cap might thus not
affect the nucleophile attack (and the acyltransfer ability) but the
formation of the acyl-enzyme (and the total specific activity),
suggesting that the lipid substrate binding site is in the cap.
Conjugated effects of the cap and the C-term were observed
with CalLAc8_cap_ct. Its rt max value was 3 times lower and
(kt/kh)app similar to the one of CalLAc8. The mutations of both
cap and C-term flap might have destabilized the structure of the
enzyme but no significant effect on the acyltransfer behavior
could be evidenced.
CduLAc_
woC
710
422
263
275
303
315
189
136
123
117
II
II
II
II
II
CpLIP2
CalLAc8_
cap
CalLAc8
45
CalLAc8_
cap_ct
13
CalLAc8_
ct
263
118
0.22
0.009 0.996
II
30
40
27
20
13
1.20
1.92
0.047 0.978 III
0.074 0.966 III
CAL-A_ct
CAL-A
0.5
0.8
It is to be noted that none of the CalLAc8 chimeric enzymes
were activated in the presence of methanol, unlike what was
observed with CduLAc[7b]. This activation phenomenon could
thus not be transferred with these sub-domains exchanges.
Therefore, it seems that the cap does not play a significant role
in the acyltransfer ability of the enzymes while the effect of the
C-term flap remains unclear. Our results tend to show that it is
the main core of the protein that probably has the greatest
influence on the specificity of reaction and also on the
activation/inhibition phenomenon. This observation is to be
related to the nucleophile pocket recently described by Jan et al
[7c]. Indeed, although very similar, the nucleophile pockets of
CduLAc and CalLAc8 differ by 18% of their residues, and
notably by the key residue A366 of CduLAc [7b], which is a
threonine in CalLAc8.
The 18% differences in the two enzymes primary sequences are
therefore responsible for significant differences in their
acyltransfer ability and their level of activity in the presence of
methanol (increase or decrease).
Regarding the role of the C-term flap, the rt max of CduLAc_woC
was 615 times lower than that of the wild enzyme, meaning that
the presence of the C-term is crucial for the enzymatic activity of
CduLAc and probably for the overall protein structuration.
Moreover, its (kt/kh)app was 2 times lower than the one of CduLAc
showing a significant decrease of the acyltransfer ability. When
comparing the 3D models of CduLAc and CduLAc_woC, we
observed that the nucleophile pocket previously identified in the
main core of the protein[7c] appears obstructed by the C-term flap
in the wild enzyme. By removing the 29 last amino acids, in
addition to a general destabilisation, we might have created an
easier and larger entrance to the active site, facilitating its
accessibility for the lipid substrate (and the water), through the
nucleophile pocket. The decrease of specific activity might be
linked to the steric hindrance of the lipid substrate inside the
nucleophile pocket, limiting the accessibility for the acyl acceptor
(docking experiments, not shown). About the swapping of the
C-term flaps, the mutant CalLAc8_ct had comparable catalytic
behaviour to its wild-type enzyme CalLAc8 (figure 2, table 1). In
this case, the length of the C-term was the same as the wild type
Effect on the nature of the accepting alcohol on
transesterification
As the ability of the enzymes to catalyse acyltransfer can also
depends on the nature of the nucleophile, their behavior in the
presence of various alcohols was evaluated. The competition
between water and alcohol as acyl acceptor is modulated by
both the chemical reactivity and the thermodynamic activity of
the reactants and by intrinsic properties of the biocatalysts that
will preferentially catalyze either hydrolysis or alcoholysis
reactions in the deacylation step of the catalytic mechanism.
This article is protected by copyright. All rights reserved.

10.1002/cbic.201600672
ChemBioChem
FULL PAPER
Thus, to evaluate only the enzyme-substrate interactions and
catalytic features, the tested alcohols were added in the reaction
mixture at the same relative thermodynamic activities (aalcohol/(aW
+ aalcohol)). The relations between the alcohol concentrations and
thermodynamic activities were estimated by the UNIFAC group
contribution method in the biphasic medium, using liquid-liquid
equilibria parameters (table 2). For a same thermodynamic
activity, the alcohol concentration can be very different (e.g. 7
fold variation between ethanol and butanol).
increase of alcohol concentration on the activity, CduLAc and
CalLAc8 showed comparable preference for the acyl acceptor
alcohol. For example, with the two enzymes the higher TR were
always obtained with butan-1-ol and the lower with propan-2-ol.
Therefore, although their structural differences play a significant
role in the enzyme selectivity toward alcohol or water, resulting
in medium or high acyltransferase character, they do not seem
to influence peculiar selectivity for one alcohol or another.
The experiments conducted with various alcohols as
nucleophiles did not show clear distinction between the enzymes,
wild type and chimeric. Indeed, the specificity profiles of acyl
acceptors were the same for all the enzymes (Figure S3).
Previous works led on CpLIP2 evidenced that some amino acids
of the cap, in contact with the main core of the protein and in the
vicinity of the active site had an impact on the acyl acceptor
specificity. For example, the mutation I231F in CpLIP2 caused a
decrease of specificity for the secondary alcohols[7a]. Yet, as the
Table 2. Relative thermodynamic activity of alcohols aalcohol/(aalcohol+aW) and
the corresponding alcohols concentrations. Relation between thermodynamic
activities asnd concentrations were estimated using UNIFAC- LLE taking into
account the presence of alcohol, water and ester in a biphasic system at 30°C.
relative thermodynamic
activity of alcohol (%)
alcohol concentration
(M)
alcohol
Ethanol
0.079
0.24
1.0%
2.9%
1.0%
2.9%
1.0%
2.9%
1.0%
2.9%
1.0%
2.9%
1.0%
2.9%
1,2
1
a.1
78%
48%
63%
75%
35%
47%
0.029
0.087
0,029
0,087
0,012
0,034
0,012
0,034
0,012
0,034
Propan-1-ol
Propan-2-ol
Butan-1-ol
Butan-2-ol
Isobutanol
0,8
0,6
0,4
0,2
0
a.2
b.1
b.2
1,2
1
70%
60%
69%
81%
0,8
0,6
0,4
0,2
0
89%
89%
5
4,5
4
3,5
3
2,5
2
1,5
1
93%
91%
86%
77%
80%
69%
The transesterification of methyl oleate (C18:1 ME) in emulsion
was performed in the presence of primary alcohols (ethanol,
propan-1-ol, butan-1-ol and isobutanol) and secondary alcohols
(propan-2-ol and butan-2-ol) at an estimated relative
thermodynamic activity of alcohol of 1 % and 2.9 % (Table 2).
Figure 3 shows that the two wild enzymes, CalLAc8 and CduLAc
displayed a transesterification activity toward all primary and
secondary alcohols. For a similar relative thermodynamic activity
of alcohol, the transesterification ratio increased with the length
of the alkyl chain. The transesterification ratio also depended on
the structural isomer tested. For both enzymes, TR was higher
for primary than secondary alcohols. Comparing the effect of the
increase of the relative thermodynamic activity of alcohol (1 %
and 2.9 %), the hydrolysis activity decreased while the
transesterification increased. Similar results were obtained with
CpLIP2 by Neang et al [9]. As observed with the increase of
methanol concentration in the first experiment of this study and
in previous studies [7b], the total activities of CalLAc8 and
CduLAc seemed to be sensitive to the increase of alcohol
concentration. While the total activity of CalLAc8 decreased, the
one of CduLAc was higher at 2.9 % than at 1 % of relative
thermodynamic activity of alcohol. Apart from the differences
evidenced in transesterification ratios and in the influence of the
0,5
0
5
4,5
4
3,5
3
97%
88%
89%
93%
96%
84%
2,5
2
1,5
1
0,5
0
Figure 3 Transesterification ( ) and competitive hydrolysis (
) activities
catalyzed by (a) CalLAc8 and (b) CduLAc with various alcohols. Reactions
were performed at 30°C, pH 6.5 in the presence of 10 mM of C18:1 ME and a
relative thermodynamic activity of alcohol of 1 % (i.1) or 2.92 % (i.2). For each
enzyme, values of transesterification and hydrolysis activity are expressed
relatively to their specific activity in condition without alcohol (hydrolysis, not
shown) and normalized to 1: (a) 15.96 ±0.23, (b) 11.8 ±0.44 µmol.min-1.mg-1
This article is protected by copyright. All rights reserved.

10.1002/cbic.201600672
ChemBioChem
FULL PAPER
amino acids of the cap in contact with the main core of CalLAc8
are identical or at least share similar properties than the ones
present in CduLAc, no difference in the alcohol specificity was
observed in relation with this cap exchange. The only notable
differences between wild-type and chimers were (i) the increase
of TR for CAL-A_ct compared to CAL-A and (ii) the decrease of
total activity for CduLAc_woC in comparison to CduLAc (Figure
S3).
at pH 6.5 and 30°C in the presence of 2.2 M methanol. Figure 4
[18]
shows specificity constants 1/α
for total activity
(transesterification plus competitive hydrolysis) catalyzed by the
studied enzymes.
In these conditions, no residual hydrolysis activity was detected
with CduLAc and CalLAc8, confirming the acyltransferase
character of the two wild enzymes. Compared to the activity with
ethyl oleate as a substrate, the activity on saturated monoesters
was very low. Indeed, the total specific activities in the presence
of 2.2 M methanol and ethyl oleate were of 61 U.mg-1 and
20 U.mg-1, 74 and 32 times higher than with saturated
monoesters for CduLAc and CalLAc8 respectively. CduLAc
specificity constants distribution was bimodal, with C16:0 EE as
the best long-chain substrate and another optimum for C8:0 EE.
CalLAc8 discriminated strongly between its preferred substrate,
C8:0 EE, and the other monoesters, notably with a very low
affinity for C12:0 EE (no activity detected in the tested
conditions).
To bring insights into the understanding of these different
substrate specificities, docking experiments were conducted with
the 3D models of both wild lipases/acyltransferases. For each
substrate, 150 possible docking positions were obtained as the
result of 15 runs, each retaining the 10 best positions. The
supposed mechanism of reaction is a nucleophilic attack by the
oxygen atom of the hydroxyl group of the catalytic serine on the
carbon atom of the carboxyl group of the ester. The docking was
considered correct when the distance between these two atoms
was inferior to 4 Å. In addition to the distance, the number of
correct positioning for each molecule within the 150 results was
considered to be an indicator of the probability of formation of
the acyl-enzyme. It is to be noted that the docking calculations
have to be considered cautiously, as they do not take into
account the mobility of the molecules and represent only one
putative arrangement of the residues. Yet, they can be used to
support our experimental observations. In CduLAc, two tunnels
inside the cap allowed the positioning of the substrates (Figure
5). The first tunnel (number 1 on the figure 5) is equivalent to the
49 % 57 %
1
57 %
50 %
52 %
0,8
CAL-A
58 %
55 %
CAL-A_ct
0,6
0,4
0,2
0
57 %
52 %
66 %
42 %
34 %
C8
92 %
C10
C12
C14
C16
C18
1
0,8
0,6
0,4
0,2
0
89 %
CduLAc
42 %
100 %
CduLAc_woC
63 %
100 %
C8
C10
C12
C14
C16
C18
1
0,8
0,6
0,4
0,2
0
CalLAc8
CalLAc8_ct
CalLAc8_cap_ct
CalLAc8_cap
tunnel binding-site of CAL-A, in which was trapped
a
polyethylene glycol molecule in the CAL-A crystallized structure
[13a]
and that was compared to the CRL binding-site [12]. In this
tunnel, C16:0 EE and C8:0 EE were in direct competition, with
distances to the catalytic serine of 3.5 Å and 3.3 Å respectively.
C12:0 EE was also docked in this tunnel at a distance of 3.5 Å. It
has to be noted that the number of correct dockings obtained for
the three substrates in the tunnel 1 can be correlated to the
specificity constants. Indeed, more molecules of C8:0 EE and
C16:0 EE were docked correctly, compared to C12:0 EE. In the
second position inside the cap, referred as tunnel number 2 on
figure 5, docking was only possible for C12:0 EE and C16:0 EE.
No docking of C8:0 EE was correct in this tunnel. However, a
third position (tunnel number 3 on figure 5) was identified in the
cavity previously described as the nucleophile pocket [7c], in
which only C8:0 EE was docked at a distance of 3.3 Å of the
catalytic serine.
C8
C10
C12
C14
C16
C18
Figure 4 Relative specificity constants 1/α for total activity (transesterification
plus competitive hydrolysis of ethyl esters of saturated fatty acids C8-C18 in
presence of methanol) catalyzed by CAL-A, CduLAc, CalLAc8 and the
chimeric enzymes (CAL-A_ct, CduLAc_woC, CalLAc8_cap, CalLAc8_ct and
CalLAc8_cap_ct). Transesterification ratios are stated above the bars if not
100%. Reactions were performed during 15 min at 30°C pH 6.5 in the
presence of 2.2 M of methanol and 10 mM of ethyl esters mixture (1.6 mM of
each ester).
Influence of the acyl chain of the donor ester on the activity with
saturated monoesters of various carbon chain length
The substrate specificity was assessed with an equimolar
mixture of six saturated monoesters (from 8 to 18 carbon atoms)
This article is protected by copyright. All rights reserved.

10.1002/cbic.201600672
ChemBioChem
FULL PAPER
lying under the catalytic serine and previously described as the
“nucleophile pocket” [7c] (data not shown). Concerning the mutant
CAL-A_ct, the monomodal distribution was slightly modified,
centered on C12 instead of C14 for the wild enzyme. Moreover,
the 1/α of C8 to C12 were higher for CAL-A_ct than CAL-A and
the 1/α of C14 to C18 decreased. With the chimeric CaLAc8_ct,
no major change of specificity profile was observed. Indeed, only
a minor decrease of 1/α for the non-preferred substrates (C10 to
C18) was evidenced. The C-term of CduLAc seems therefore to
be slightly more favorable to the short chain substrates. Indeed
the increase of specificity for shorter substrates was related to
an increase of specific activities for these substrates, while the
ones for the longer substrates were more constant between the
chimers and the wild-types. Although different in length and
amino acids composition, the effect observed on the substrate
specificity was marginal when swapping the C-term flaps, like for
the reaction specificity. This favours the hypothesis that the long
C-term flap has low physical contact with the rest of the protein
and may not play a direct role in the enzyme selectivity.
Nevertheless, our results tend to show that its presence is
crucial as its deletion was deleterious for the overall activity of
the lipase/acyltransferase CduLAc. Interestingly, Wikmark et
al[13b] showed that its deletion in CAL-A was not so damaging,
but their experimental conditions were quite different from ours
and the truncated sequence was slightly shorter.
Regarding the influence of the cap, the mutant CalLAc8_cap,
had a specificity profile similar to the one of CduLAc. The
exchange of sub-domain caused a significant increase of the
specific activity for the long-chain fatty esters, particularly for
C16:0, but the specific activity towards C8:0 remained the same.
This confirms that the cap of CduLAc accommodates better the
long chain substrates than the one of CalLAc8. This is
consistent with the previous docking observations on the 3D
models of the wild enzymes and shows that the acyl-donor
specificity can be transferred by cap “exchange”.
However, the mutant CalLAc8_cap_ct presented an increase of
1/α for the C8:0 to C12:0 EE and a decrease for the C14:0 to
C18:0 EE. The specific activities for the C8:0 and C16:0 EE
(0.350 and 0.160 U.mg-1) were equivalent to the ones of
CalLAc8 (0.306 U.mg-1 and 0.147 U.mg-1), while the specific
activities for C10:0 and C12:0 EE were significantly increased
(3.4 times for C10:0 EE). Therefore, it seems that combined
effect of the C-term and cap is different from that of the cap only
and more difficult to predict.
1
2
3
Figure 5. Docking of C8:0 (pink), C12:0 (green) and C16:0 (gold) in CduLAc.
The catalytic triad is colored in orange and the cap and C-term flap in darker
blue. The identified tunnels for the positioning of lipidic substrates are
numbered 1 to 3 and highlighted in yellow.
On the contrary, only the first tunnel allowed the correct
positioning of the substrates in CalLAc8 (data not shown). The
dockings of C8:0 EE and C16:0 EE in tunnel 1 in CalLAc8
seemed to be similar to those in CduLAc but differed for C12:0
EE. Indeed while this substrate was successfully docked in
CduLAc, no correct positioning for the reaction was obtained
with CalLAc8. Moreover, the correct positioning of C8:0 EE was
obtained with a higher probability than the one of C16:0 EE.
Interestingly, the total specific activity of CalLAc8 is twice higher
for C8:0 EE than C16:0 EE (0.31 U/mg and 0.15 U/mg), which
matches the docking experiments. In comparison, the specific
activities of CduLAc for the same substrates were 0.19 U.mg-1
(C8:0 EE) and 0.28 U.mg-1 (C16:0 EE). Regarding tunnel 2, it
was only evidenced in CduLAc where it appeared to favor C16:0
EE. The substrate specificities of CduLAc and CalLAc8
observed experimentally were thus supported by the docking
experiments in the 3D model developed. The cap appears to be
the crucial sub-domain for the substrate specificity (lipid acyl-
donor), as it was observed previously in CpLIP2, in CAL-A and
in C.rugosa lipases [7a, 13a, 19]
.
Then, we evaluated the acyl donor specificity of the chimeric
enzymes. The deletion of the C-term flap in CduLAc led again to
a drastic decrease of total specific activity, from 0.82 U.mg-1 to
0.04 U.mg-1 respectively for CduLAc and CduLAc_woC. As
observed on figure 4, the specificity constants for CduLAc_woC
were higher for monoesters from C8:0 to C14:0. Based on the
hypothesis that a new access to the active site for the lipid
substrate is created, the absence of C-term flap might have
increased the accessibility for short length substrates, so the
competition between the esters is higher than in the wild type
enzyme. Docking experiments conducted with C8:0 EE, C12:0
EE and C16:0 EE on CduLAc and CduLAc_woC showed that if
the dockings in the cap were identical, there was possibly a
larger access for the substrate to the active site. Indeed, in
CduLAc_woC the absence of C-term shielding the active site
suggest a possible binding of the lipidic substrate in a activity
Influence of temperature on activity
Finally, we have evaluated the hydrolysis activities of CduLAc,
CalLAc8, at various temperatures, in a range of 5 to 80°C, with
ethyl oleate as substrate. To assess the influence of the
subdomain change on the enzyme behavior towards
temperature increase, same experiments were realized on the
chimeric enzymes. Reactions were performed in buffered
aqueous emulsions at pH 6.5. The experimental determination
of the specific activities at each temperature allowed the
calculation of the enzymes’ activation energies EA in these
conditions, using the Arrhenius model.
This article is protected by copyright. All rights reserved.

10.1002/cbic.201600672
ChemBioChem
FULL PAPER
CalLAc8_ct were respectively of 17 kJ.mol-1 and 16 kJ.mol-1, a
same order of magnitude as the one of CalLAc8 (15 kJ.mol-1). At
80°C, the activity of CAL-A_ct was still increasing, as observed
with CAL-A. Regarding CduLAc_woC, the depletion of the C-
term in CduLAc did not have an impact on the profile of activity
at various temperatures and no change of EA was observed.
Thus, it seems that the C-term plays a crucial role in the activity
of the enzyme, regardless of the temperature.
40
35
30
25
20
15
10
5
1,8
1,6
1,4
1,2
1
0,8
0,6
0,4
0,2
0
Surprisingly, the conjugated effect of the cap and C-term flap
was unfavorable to the cold active property of CalLAc8
(CalLAc8_cap_ct). Indeed, as presented in table 3, the value of
EA was 31 kJ.mol-1 for CalLAc8_cap_ct. Thus CalLAc8_cap_ct
had a mesophilic profile with medium EA, like CduLAc. Some
interactions might exist between the C-term and the cap of
CduLAc increasing the rigidity of the chimer CalLAc8_cap_ct
0
0
20
40
60
80
Temperature (°C)
that could explain the loss of the cold active property[20]
.
Figure 6 Influence of temperature on the hydrolysis activity of CduLAc (□),
CalLAc8 (●) and CAL-A ( ). Reactions were performed during 15 min at
temperatures in the range 5 – 80°C, pH 6.5 in the presence of 10 mM of ethyl
oleate in PVA emulsion.
Apart from the differences observed for CalLAc8_cap_ct, the
exchange of subdomains did not change the enzymes’ activity
profile towards temperature. When functional, the chimeric
enzymes therefore appear quite stable. It is not very surprising
for CalLAc8_cap which in fact has 96% identity with CalLAc8. As
the flap exchanges did not have an impact on the activity profile
towards temperature, it again tends to confirm that the flap has
only few contacts with the rest of the protein but its role in the
enzyme activity remains unclear.
Table 3. Activation energies of CalLAc8, CduLAc, CAL-A and the chimeric
enzymes for the hydrolysis of ethyl oleate at pH 6.5. Enzymes in green are
non cold-active, in blue cold active and in orange thermophilic. Optimum
temperatures are also provided.
Conclusions
Enzyme
CduLAc
EA (kJ.mol-1)
Optimum temperature
34
40°C
Despite the high level of identity (82 %) between lipases/
acyltransferases CduLAc and CalLAc8, several differences in
their catalytic behavior were enlightened. Indeed CduLAc
belongs to the class I of lipases/acyltransferases while CalLAc8
is a less efficient acyltransferase and is related to class II. Apart
from the difference in the reaction specificity, we showed that
the acyl donor preference of CduLAc is characterized by a
bimodal distribution (C8 and C16 being the preferred substrates),
while the one of CalLAc8 is very specific of C8. By docking
experiments, we managed to identify three binding pockets for
the lipidic substrate in CduLAc (two of them buried in the cap),
while there were only two in CalLAc8, suggesting that the cap
plays an important role in the accommodation of the substrate.
To investigate further the role of the cap and C-term flap
subdomains in the catalytic behavior of the lipases/
acyltransferases, we generated chimeric enzymes based on
CduLAc and CalLAc8 on one hand and CduLAc and CAL-A on
the other hand. The exchanges of cap and C-term modified the
activity (rt max) but did not have a major impact on the acyltransfer
34
31
40°C
20°C
CduLac_woC
CalLAc8_cap_ct
CalLAc8
15
15
14
34
29
20°C
20°C
CalLAc8_cap
CalLAc8_ct
CAL-A
20°C
> 80°C
> 80°C
CAL-A_ct
As presented in figure 6, CalLAc8 and CduLAc exhibited heat
lability with a significant decrease of activity above 20°C and
40°C respectively, to reach 4% of their maximal activity at 80°C
after 15 min of reaction. As observed before for the
characterized lipases/acyltransferases CpLIP2 and CtroL4, the
enzymes of this peculiar family do not seem to share the
exceptional thermostability of CAL-A [5]. As shown in table 3,
despite their high level of identity, CduLAc has a mesophilic
profile, with an activation energy of 34 kJ.mol-1 while CalLAc8 is
a cold-active enzyme, with EA of 15 kJ.mol-1, in the same range
ability ([MeOH]TR50 and (kt/kh)app
. It is therefore the few
differences observed in the amino acids composition of the
nucleophile pocket that are probably crucial for the preference
for the alcohol compared to water. The main core of the proteins,
comprising the nucleophile pocket enlightened by Jan et al.[7c]
was thus indirectly proven to be the most important part of the
protein in the acyltransfer ability. On the other hand, the
increase of rt max observed for CalLAc8_cap thus seems to be
conditioned by the accommodation of the lipid substrate. This
favors the hypothesis that the formation of the acyl-enzyme is
the rate-limiting step of the reaction. The acyl donor specificity
as the one of CpLIP2, described by Neang et al [5]
.
The exchanges of cap and C-term flap did not have an impact
on the cold active property of CalLAc8 (CalLAc8_cap and
CalLAc8_ct) nor on the thermophilic characteristic of CAL-A
(CAL-A_ct). Indeed, the activation energies of CalLAc8_cap and
This article is protected by copyright. All rights reserved.

10.1002/cbic.201600672
ChemBioChem
FULL PAPER
Competent cells of Escherichia coli XL1-Blue MRF’ (Stratagene la Jolla,
CA, USA) were used for DNA propagation. K. pastoris X-33, pPICZαB
vector and zeocin were from Invitrogen (Life Technologies SAS, Saint
Aubin, France). SacI restriction enzyme and RNaseA were purchased
from Roche Diagnostic AG (Rotkreuz, Switzerland). All lipase substrates
and reagents were purchased from Sigma-Aldrich (Lyon, France) and
were of analytical grade.
seems to rely more on the cap. Indeed, the specificity of the
mutant CalLAc8_cap differed from that of CalLAc8 and was
more similar to the one of CduLAc. This result shows that cap
engineering is a promising tool to modify that acyl-donor
substrate specificity.
Further investigations have to be conducted on the role of the
C-term. Indeed, if we evidenced its implication in the activity of
the enzyme, we could not point out the role of this subdomain
precisely. Interactions might be possible between the C-term
and the cap and could be favored when both have the same
origin (e.g. those of CduLAc). More interactions could rigidify the
structure, limit the access to the active site to longer substrate
(C16), and increase the stability (CalLAc8, CalLAc8_cap,
CalLAc8_ct are cold-active while CalLAc8_cap_ct is not). Yet,
this is not totally true for the wild CduLAc which favors the C16
substrate but is not cold active, so it is difficult to conclude. Only
the comparison of real structures would allow clear elucidation of
the complex structure/function relationships of such chimeric
enzymes. Wikmark et al[13b] suggested that the C-term might
have a role in the interfacial activation of CAL-A, as the
truncated variant of this enzyme followed a Mickaelis-Menten
kinetic, but their use of amphiphilic substrate up to the critical
micellar concentration is not really comparable to our conditions
with water/lipid interface. To show the impact of the C-term and
its length on the interfacial activation, further studies are
required on CduLAc and CduLAc_woC on one hand and CAL-A
and CAL-A_ct on the other hand.
Production of recombinant lipases
The protein sequence of CAL-A from M. (C.) antarctica was obtained
from NCBI databases under accession number 2VEO_A. The sequences
of CalLAC8 from C. albicans and CduLAc from C. dubliniensis,
accessible under numbers XP_711685 and XP_002421466.1,
respectively, were from Subileau et al [4a] and Jan et al [7b], with the signal
peptide suppressed. DNA was synthesized and subcloned by Life
Technologies (Regensburg, Germany) in pPICZαB in fusion with the
signal peptide of the alpha-mating factor of Saccharomyces cerevisiae
provided in the plasmid, as described by Neang et al [9]. To produce
mutants CalLAc8_cap, CalLAc8_ct, CalLAc8_cap_ct, CduLAc_woC,
CaL-A_ct and CAL-A_cap, site directed mutagenesis was performed by
Life Technologies on the plasmids containing the gene of CalLAc8 (for
CalLAc8_cap, CalLAc8_ct and CalLAc8_cap_ct), the gene of CduLAc
(for CduLAc_woC) or the gene of CAL-A (for CAL-A_ct and CAL-A_cap).
Clones for the expression of the different enzymes were produced as
described by Neang et al [9]. Recombinant lipases were obtained from
culture supernatants of transformed K.pastoris as described by Brunel et
al [30] and modified by Neang et al [9]
.
Enzyme extract preparation
The final 800 mL culture medium was centrifuged at 4°C for 10 min at
10 000 xg to eliminate the cells. The supernatant (approximately 700 mL)
was successively filtered with 5.0, 0.8 and 0.45 µm pores diameter filters
(Millipore, Molheim, France) before its concentration to 90 mL by hollow-
fiber tangential flow ultrafiltration using a Quixstand apparatus (GE
Healthcare, Chalfont St Giles, UK) equipped with a 30 KDa cut-off
module. Diafiltration with 600 mL of ultrapure water was then performed
in the same ultrafiltration system. Enzymes were produced at multigram
per liter scale.
Experimental Section
Sequences analyses
Sequence visualization and multisequence alignments were generated
[21]
using Seaview
and ClustalX [22]. Protein sequence identities were
obtained using the SIM alignment tool for protein sequences program [23]
.
Evaluation of thermodynamic activities
Protein analysis
Thermodynamic activities of water and others alcohols at 30°C in the
presence of ethyl oleate (for methanol) or methyl oleate (for ethanol,
propan-1-ol, propan-2-ol, butan-1-ol, butan-2-ol and isobutanol) in a two-
phases system were estimated using the UNIFAC group contribution
Protein concentrations in enzymatic extracts were determined by the
[31]
Bradford method
using pure, lyophilized CpLIP2 as standard. Protein
concentrations are therefore given in mg equivalent CpLIP2 protein per
[32]
mL. SDS-PAGE was performed according to method of Laemmli
method [24] using the liquid liquid equilibrium parameters [25]
.
using 10% acrylamide gels. Proteins were stained with Coomassie
Brilliant Blue R-250. Molecular weight were evaluated as described by
Construction and analysis of the 3D models
Subileau et al [4a]
.
The 3D models of CduLAc and CalLAc8 were designed, based on the
crystallographic structure of CAL-A (PDB ID: 2VEO, 3GUU)[13a] and the
3D model of CpLIP2 described by Subileau et al [7a], using Modeller
9.14[26] via UCSF Chimera 1.10.2[27]. The cavities inside the proteins
Enzymatic assays
Hydrolysis and transesterification (alcoholysis) activities were
respectively determined by measuring the initial rates of fatty acid and
ester production (µmol.min-1.mL-1). The transfer ratio (TR in %)
corresponds to the relative initial rate of alcoholysis compared to the total
activity (initial rates of alcoholysis plus competitive hydrolysis). Specific
activity is the total activity (µmol.min-1.mL-1) divided by protein
concentration (mg.mL-1). Results are expressed as the mean of 3
independent repetitions ± standard deviation.
were determined using the online program CASTp[28]
. Docking
experiments were conducted with ethyl caprylate, ethyl laurate and ethyl
palmitate using AutodockVina[29] via UCSF Chimera. The score attributed
to the docked molecules in Chimera can be assimilated to a binding
energy. Therefore, using the Arrhenius equation, we calculated the
dissociation constants (KD) of each docked molecule of lipidic substrates.
Plasmids, strains and reagents
The amount of enzyme was adapted for each biocatalyst so as to be in
conditions where the initial rate of hydrolysis of the lipidic substrate was
proportional to protein concentration, in the presence of saturated or
This article is protected by copyright. All rights reserved.

10.1002/cbic.201600672
ChemBioChem
FULL PAPER
[6] D. Briand, E. Dubreucq, P. Galzy, Biotechnol Lett 1994, 16,
813-818.
unsaturated esters. The transesterification reaction was realized using
ethyl oleate as a substrate when methanol was the acyl acceptor and
using methyl oleate when acyl acceptors were other alcohols (ethanol,
propan-1-ol, propan-2-ol, butan-1-ol, butan-2-ol and isobutanol).
Specificity constants 1/α [18] 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 [18]. The influence of temperature on the enzymatic
activity, was studied between 5°C and 80°C. Experiments were
[7] a) M. Subileau, A. H. Jan, H. Nozac'h, M. Pérez-Gordo, V.
Perrier, E. Dubreucq, BBA Proteins Proteom 2015, 1854, 1400-
1411; b) A. H. Jan, M. Subileau, C. Deyrieux, V. Perrier, E.
Dubreucq, BBA Proteins Proteom 2016, 1864, 187-194; c) A. H.
Jan, D. E., M. Subileau, submitted in ASC Catal. 2016, cs-
2016-02883v.
[8] M. Fischer, J. Pleiss, Nucleic Acids Res 2003, 31, 319-321.
[9] P. M. Neang, M. Subileau, V. Perrier, E. Dubreucq, J Mol Catal
B: Enzym 2013, 94, 36-46.
conducted as described by Neang et al [9] and modified by Jan et al [7b]
.
[10] E. Husson, C. Humeau, C. Harscoat, X. Framboisier, C. Paris,
E. Dubreucq, I. Marc, I. Chevalot, Process Biochem. 2011, 46,
945-952.
As described by Subileau et al [4a], the following parameters were
determined from experiments with various methanol concentrations:
[11] L. Vaysse, E. Dubreucq, J.-L. Pirat, P. Galzy, J Biotechnol 1997,
- The maximum specific transfer rates rt
(it has to be noted that
max
53, 41-46.
because the enzymes studied were in-lab produced in the same
[12] M. Widmann, P. B. Juhl, J. Pleiss, BMC Genomics 2010, 11,
123.
[13] a) D. J. Ericsson, A. Kasrayan, P. Johansson, T. Bergfors, A. G.
Sandstrom, J. E. Backvall, S. L. Mowbray, J Mol Biol 2008, 376,
109-119; b) Y. Wikmark, K. Engelmark Cassimjee, R. Lihammar,
J. E. Backvall, Chembiochem 2016, 17, 141-145.
[14] a) S. Brocca, F. Secundo, M. Ossola, L. Alberghina, G. Carrea,
M. Lotti, Protein Sci 2003, 12, 2312-2319; b) C. C. Akoh, G. C.
Lee, J. F. Shaw, Lipids 2004, 39, 513-526.
system and exhibit comparable mass, the rt
can be used to
max
compare the efficiency of the enzymes).
- The alcohol concentration required to reach 50% of transfer ratio (TR)
[MeOH]TR50
,
- The ratio of the apparent rate constants for the acyltransfer to
methanol and for hydrolysis (kt/kh)app, determined from the initial rates
for acyltransfer to methanol (rt) and for hydrolysis (rh) and nucleophiles
concentrations ([H2O] and [MeOH]) with the following equation [17]
:
[15] F. Secundo, G. Carrea, C. Tarabiono, P. Gatti-Lafranconi, S.
Brocca, M. Lotti, K. E. Jaeger, M. Puls, T. Eggert, J Mol Catal
B-Enzym 2006, 39, 166-170.
[16] M. Skjot, L. De Maria, R. Chatterjee, A. Svendsen, S. A. Patkar,
P. R. Ostergaard, J. Brask, Chembiochem 2009, 10, 520-527.
[17] V. Kasche, Enzyme Microb Tech 1986, 8, 4-16.
[18] L. Vaysse, A. Ly, G. Moulin, E. Dubreucq, Enzyme Microb Tech
2002, 31, 648-655.
푘t
푟
[H2O]
t
( )
2 (
)
app
=
∙
푘h
푟h [MeOH]
A
linear relationship was observed between the rt/rh and the
[MeOH]/[H2O] ratios within the range of methanol concentrations tested,
with (kt/kh)app as the slope. (kt/kh)app’ is also the ratio of the apparent rate
constants of acyltransfer versus hydrolysis but determined as the slope
of the initial rates ratio rt/rh as
a function of the ratio of the
[19] a) M. Norin, F. Haeffner, A. Achour, T. Norin, K. Hult, Protein
Sci 1994, 3, 1493-1503; b) M. Holmquist, Chem Phys Lipids
1998, 93, 57-66; c) J. Pleiss, M. Fischer, R. D. Schmid, Chem
Phys Lipids 1998, 93, 67-80; d) P. Grochulski, F. Bouthillier, R.
J. Kazlauskas, A. N. Serreqi, J. D. Schrag, E. Ziomek, M.
Cygler, Biochemistry 1994, 33, 3494-3500.
[20] C. Struvay, G. Feller, Int J Mol Sci 2012, 13, 11643-11665.
[21] N. Galtier, M. Gouy, C. Gautier, Comput Appl Biosci 1996, 12,
543-548.
[22] M. A. Larkin, G. Blackshields, N. P. Brown, R. Chenna, P. A.
McGettigan, H. McWilliam, F. Valentin, I. M. Wallace, A. Wilm,
R. Lopez, J. D. Thompson, T. J. Gibson, D. G. Higgins,
Bioinformatics 2007, 23, 2947-2948.
corresponding methanol versus water thermodynamic activities
aMeOH/aW.
Acknowledgements
Authors thank the University of Montpellier and Montpellier
SupAgro for the PhD granting of Ms. AH Jan.
GC analyses were performed using LiPolGreen platform
(http://www.supagro.fr/plantlippol-green/).
K. pastoris cultivations were conducted on the UMR IATE
fermentation platform, a member of the I-FERM consortium
(http://www.i-ferm.com/en).
[23] X. Huang, Adv Appl Math 1991, 12, 337-357.
[24] A. Fredenslund, R. L. Jones, J. M. Prausnitz, AIChE J 1975, 21,
1086-1099.
[25] a) T. Magnussen, P. Rasmussen, A. Fredenslund, Ind Eng
Chem Process Des Dev 1981, 20, 331-339; b) P. Oracz, M.
Goral, J Chem Eng Data 2011, 56, 4853-4861.
[26] a) A. Fiser, R. K. Do, A. Sali, Protein Sci 2000, 9, 1753-1773; b)
M. A. Marti-Renom, A. C. Stuart, A. Fiser, R. Sanchez, F. Melo,
A. Sali, Annu Rev Biophys Biomol Struct 2000, 29, 291-325; c)
A. Sali, T. L. Blundell, J Mol Biol 1993, 234, 779-815.
[27] E. F. Pettersen, T. D. Goddard, C. C. Huang, G. S. Couch, D. M.
Greenblatt, E. C. Meng, T. E. Ferrin, J Comput Chem 2004, 25,
1605-1612.
Keywords: lipases/acyltransferases • synthesis de novo •
structure/function • cap • chimeric enzymes
References
[1] U. Schörken, P. Kempers, Eur J Lipid Sci Technol 2009, 111,
627-645.
[2] a) F. Hasan, A. A. Shah, A. Hameed, Enzyme Microb Tech
2006, 39, 235-251; b) P. Villeneuve, Biotechnol Adv 2007, 25,
512-536.
[3] L. Casas-Godoy, S. Duquesne, F. Bordes, G. Sandoval, A.
Marty, Methods Mol Biol 2012, 861, 3-30.
[4] a) M. Subileau, A. H. Jan, J. Drone, C. Rutyna, V. Perrier, E.
Dubreucq, submitted in RSC Catal. Sci. Tech. 2016, CY-ART-
08-2016-001805; b) K. Zorn, I. Oroz-Guinea, H. Brundiek, U. T.
Bornscheuer, Prog Lip Res 2016, 63, 153-164.
[28] J. Dundas, Z. Ouyang, J. Tseng, A. Binkowski, Y. Turpaz, J.
Liang, Nucleic Acids Res 2006, 34, W116-118.
[29] O. Trott, A. J. Olson, J Comput Chem 2010, 31, 455-461.
[30] L. Brunel, V. Neugnot, L. Landucci, H. Boze, G. Moulin, F.
Bigey, E. Dubreucq, J Biotechnol 2004, 111, 41-50.
[31] M. M. Bradford, Anal Biochem 1976, 72, 248-254.
[32] U. K. Laemmli, Nature 1970, 227, 680-685.
[5] P. M. Neang, M. Subileau, V. Perrier, E. Dubreucq, Appl
Microbiol Biotechnol 2014, 98, 8927-8936.
This article is protected by copyright. All rights reserved.

10.1002/cbic.201600672
ChemBioChem
FULL PAPER
FULL PAPER
Design of chimeric enzymes: Despite
82% identity, the 2 recently
A-H. Jan, E. Dubreucq, M. Subileau*
CalLAC8
CduLAc
Page No. – Page No.
discovered lipases/acyltransferases,
CalLAc8 and CduLAc displayed
different catalytic behavior. To
Enlightening the role of subdomains
in the catalytic behavior of
elucidate the role of the subdomains
(main core, cap and C-term flap) of
these enzymes, chimeric enzymes
were designed and characterized.
This evidenced the implication of the
cap in the substrate specificity and the
main core in the acyltransfer ability.
lipases/acyltransferases homologous
to CpLIP2 through rational design of
chimeric enzymes
This article is protected by copyright. All rights reserved.